remove firmware files, merge more uncritical stuff from 3.0.101

271 files changed, 17182 insertions, 44506 deletions

diff --git a/drivers/gpu/drm/i2c/ch7006_drv.c b/drivers/gpu/drm/i2c/ch7006_drv.c
index d3f2e87..08792a7 100644
--- a/drivers/gpu/drm/i2c/ch7006_drv.c
+++ b/drivers/gpu/drm/i2c/ch7006_drv.c
@@ -24,8 +24,6 @@
  *
  */
 
-#include <linux/module.h>
-
 #include "ch7006_priv.h"
 
 /* DRM encoder functions */
@@ -252,7 +250,10 @@ static int ch7006_encoder_create_resources(struct drm_encoder *encoder,
 
 	drm_mode_create_tv_properties(dev, NUM_TV_NORMS, ch7006_tv_norm_names);
 
-	priv->scale_property = drm_property_create_range(dev, 0, "scale", 0, 2);
+	priv->scale_property = drm_property_create(dev, DRM_MODE_PROP_RANGE,
+						   "scale", 2);
+	priv->scale_property->values[0] = 0;
+	priv->scale_property->values[1] = 2;
 
 	drm_connector_attach_property(connector, conf->tv_select_subconnector_property,
 				      priv->select_subconnector);
diff --git a/drivers/gpu/drm/i2c/sil164_drv.c b/drivers/gpu/drm/i2c/sil164_drv.c
index b7d45ab..0b67732 100644
--- a/drivers/gpu/drm/i2c/sil164_drv.c
+++ b/drivers/gpu/drm/i2c/sil164_drv.c
@@ -24,8 +24,6 @@
  *
  */
 
-#include <linux/module.h>
-
 #include "drmP.h"
 #include "drm_crtc_helper.h"
 #include "drm_encoder_slave.h"
diff --git a/drivers/gpu/drm/i810/i810_dma.c b/drivers/gpu/drm/i810/i810_dma.c
index 2c8a60c..8f371e8 100644
--- a/drivers/gpu/drm/i810/i810_dma.c
+++ b/drivers/gpu/drm/i810/i810_dma.c
@@ -99,6 +99,7 @@ static int i810_mmap_buffers(struct file *filp, struct vm_area_struct *vma)
 	buf_priv = buf->dev_private;
 
 	vma->vm_flags |= (VM_IO | VM_DONTCOPY);
+	vma->vm_file = filp;
 
 	buf_priv->currently_mapped = I810_BUF_MAPPED;
 
@@ -221,6 +222,8 @@ static int i810_dma_cleanup(struct drm_device *dev)
 			pci_free_consistent(dev->pdev, PAGE_SIZE,
 					    dev_priv->hw_status_page,
 					    dev_priv->dma_status_page);
+			/* Need to rewrite hardware status page */
+			I810_WRITE(0x02080, 0x1ffff000);
 		}
 		kfree(dev->dev_private);
 		dev->dev_private = NULL;
@@ -1207,8 +1210,6 @@ int i810_driver_load(struct drm_device *dev, unsigned long flags)
 	dev->types[8] = _DRM_STAT_SECONDARY;
 	dev->types[9] = _DRM_STAT_DMA;
 
-	pci_set_master(dev->pdev);
-
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/i810/i810_drv.c b/drivers/gpu/drm/i810/i810_drv.c
index ec12f7d..6f98d05 100644
--- a/drivers/gpu/drm/i810/i810_drv.c
+++ b/drivers/gpu/drm/i810/i810_drv.c
@@ -30,8 +30,6 @@
  *    Gareth Hughes <gareth@valinux.com>
  */
 
-#include <linux/module.h>
-
 #include "drmP.h"
 #include "drm.h"
 #include "i810_drm.h"
@@ -43,17 +41,6 @@ static struct pci_device_id pciidlist[] = {
 	i810_PCI_IDS
 };
 
-static const struct file_operations i810_driver_fops = {
-	.owner = THIS_MODULE,
-	.open = drm_open,
-	.release = drm_release,
-	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_mmap,
-	.poll = drm_poll,
-	.fasync = drm_fasync,
-	.llseek = noop_llseek,
-};
-
 static struct drm_driver driver = {
 	.driver_features =
 	    DRIVER_USE_AGP | DRIVER_REQUIRE_AGP | DRIVER_USE_MTRR |
@@ -66,7 +53,17 @@ static struct drm_driver driver = {
 	.reclaim_buffers_locked = i810_driver_reclaim_buffers_locked,
 	.dma_quiescent = i810_driver_dma_quiescent,
 	.ioctls = i810_ioctls,
-	.fops = &i810_driver_fops,
+	.fops = {
+		 .owner = THIS_MODULE,
+		 .open = drm_open,
+		 .release = drm_release,
+		 .unlocked_ioctl = drm_ioctl,
+		 .mmap = drm_mmap,
+		 .poll = drm_poll,
+		 .fasync = drm_fasync,
+		 .llseek = noop_llseek,
+	},
+
 	.name = DRIVER_NAME,
 	.desc = DRIVER_DESC,
 	.date = DRIVER_DATE,
diff --git a/drivers/gpu/drm/i915/Makefile b/drivers/gpu/drm/i915/Makefile
index ce7fc77..0ae6a7c 100644
--- a/drivers/gpu/drm/i915/Makefile
+++ b/drivers/gpu/drm/i915/Makefile
@@ -3,7 +3,7 @@
 # Direct Rendering Infrastructure (DRI) in XFree86 4.1.0 and higher.
 
 ccflags-y := -Iinclude/drm
-i915-y := i915_drv.o i915_dma.o i915_irq.o \
+i915-y := i915_drv.o i915_dma.o i915_irq.o i915_mem.o \
 	  i915_debugfs.o \
           i915_suspend.o \
 	  i915_gem.o \
@@ -28,7 +28,6 @@ i915-y := i915_drv.o i915_dma.o i915_irq.o \
 	  intel_dvo.o \
 	  intel_ringbuffer.o \
 	  intel_overlay.o \
-	  intel_sprite.o \
 	  intel_opregion.o \
 	  dvo_ch7xxx.o \
 	  dvo_ch7017.o \
diff --git a/drivers/gpu/drm/i915/dvo_ch7017.c b/drivers/gpu/drm/i915/dvo_ch7017.c
index 1ca799a..d3e8c54 100644
--- a/drivers/gpu/drm/i915/dvo_ch7017.c
+++ b/drivers/gpu/drm/i915/dvo_ch7017.c
@@ -227,7 +227,7 @@ static bool ch7017_init(struct intel_dvo_device *dvo,
 	default:
 		DRM_DEBUG_KMS("ch701x not detected, got %d: from %s "
 			      "slave %d.\n",
-			      val, adapter->name, dvo->slave_addr);
+			      val, adapter->name,dvo->slave_addr);
 		goto fail;
 	}
 
diff --git a/drivers/gpu/drm/i915/dvo_ch7xxx.c b/drivers/gpu/drm/i915/dvo_ch7xxx.c
index 4a03660..7eaa94e 100644
--- a/drivers/gpu/drm/i915/dvo_ch7xxx.c
+++ b/drivers/gpu/drm/i915/dvo_ch7xxx.c
@@ -111,7 +111,7 @@ static char *ch7xxx_get_id(uint8_t vid)
 /** Reads an 8 bit register */
 static bool ch7xxx_readb(struct intel_dvo_device *dvo, int addr, uint8_t *ch)
 {
-	struct ch7xxx_priv *ch7xxx = dvo->dev_priv;
+	struct ch7xxx_priv *ch7xxx= dvo->dev_priv;
 	struct i2c_adapter *adapter = dvo->i2c_bus;
 	u8 out_buf[2];
 	u8 in_buf[2];
@@ -303,7 +303,7 @@ static void ch7xxx_dump_regs(struct intel_dvo_device *dvo)
 
 	for (i = 0; i < CH7xxx_NUM_REGS; i++) {
 		uint8_t val;
-		if ((i % 8) == 0)
+		if ((i % 8) == 0 )
 			DRM_LOG_KMS("\n %02X: ", i);
 		ch7xxx_readb(dvo, i, &val);
 		DRM_LOG_KMS("%02X ", val);
diff --git a/drivers/gpu/drm/i915/dvo_ivch.c b/drivers/gpu/drm/i915/dvo_ivch.c
index 04f2893..a12ed94 100644
--- a/drivers/gpu/drm/i915/dvo_ivch.c
+++ b/drivers/gpu/drm/i915/dvo_ivch.c
@@ -344,8 +344,8 @@ static void ivch_mode_set(struct intel_dvo_device *dvo,
 			   (adjusted_mode->hdisplay - 1)) >> 2;
 		y_ratio = (((mode->vdisplay - 1) << 16) /
 			   (adjusted_mode->vdisplay - 1)) >> 2;
-		ivch_write(dvo, VR42, x_ratio);
-		ivch_write(dvo, VR41, y_ratio);
+		ivch_write (dvo, VR42, x_ratio);
+		ivch_write (dvo, VR41, y_ratio);
 	} else {
 		vr01 &= ~VR01_PANEL_FIT_ENABLE;
 		vr40 &= ~VR40_CLOCK_GATING_ENABLE;
@@ -410,7 +410,7 @@ static void ivch_destroy(struct intel_dvo_device *dvo)
 	}
 }
 
-struct intel_dvo_dev_ops ivch_ops = {
+struct intel_dvo_dev_ops ivch_ops= {
 	.init = ivch_init,
 	.dpms = ivch_dpms,
 	.mode_valid = ivch_mode_valid,
diff --git a/drivers/gpu/drm/i915/dvo_sil164.c b/drivers/gpu/drm/i915/dvo_sil164.c
index a0b13a6..e4b4091 100644
--- a/drivers/gpu/drm/i915/dvo_sil164.c
+++ b/drivers/gpu/drm/i915/dvo_sil164.c
@@ -104,7 +104,7 @@ static bool sil164_readb(struct intel_dvo_device *dvo, int addr, uint8_t *ch)
 
 static bool sil164_writeb(struct intel_dvo_device *dvo, int addr, uint8_t ch)
 {
-	struct sil164_priv *sil = dvo->dev_priv;
+	struct sil164_priv *sil= dvo->dev_priv;
 	struct i2c_adapter *adapter = dvo->i2c_bus;
 	uint8_t out_buf[2];
 	struct i2c_msg msg = {
diff --git a/drivers/gpu/drm/i915/dvo_tfp410.c b/drivers/gpu/drm/i915/dvo_tfp410.c
index aa2cd3e..8ab2855 100644
--- a/drivers/gpu/drm/i915/dvo_tfp410.c
+++ b/drivers/gpu/drm/i915/dvo_tfp410.c
@@ -56,7 +56,7 @@
 #define TFP410_CTL_2_MDI	(1<<0)
 
 #define TFP410_CTL_3		0x0A
-#define TFP410_CTL_3_DK_MASK	(0x7<<5)
+#define TFP410_CTL_3_DK_MASK 	(0x7<<5)
 #define TFP410_CTL_3_DK		(1<<5)
 #define TFP410_CTL_3_DKEN	(1<<4)
 #define TFP410_CTL_3_CTL_MASK	(0x7<<1)
@@ -225,12 +225,12 @@ static void tfp410_mode_set(struct intel_dvo_device *dvo,
 			    struct drm_display_mode *mode,
 			    struct drm_display_mode *adjusted_mode)
 {
-	/* As long as the basics are set up, since we don't have clock dependencies
-	* in the mode setup, we can just leave the registers alone and everything
-	* will work fine.
-	*/
-	/* don't do much */
-	return;
+    /* As long as the basics are set up, since we don't have clock dependencies
+     * in the mode setup, we can just leave the registers alone and everything
+     * will work fine.
+     */
+    /* don't do much */
+    return;
 }
 
 /* set the tfp410 power state */
diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index 7dd8806..19bab81 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -61,7 +61,6 @@ static int i915_capabilities(struct seq_file *m, void *data)
 	const struct intel_device_info *info = INTEL_INFO(dev);
 
 	seq_printf(m, "gen: %d\n", info->gen);
-	seq_printf(m, "pch: %d\n", INTEL_PCH_TYPE(dev));
 #define B(x) seq_printf(m, #x ": %s\n", yesno(info->x))
 	B(is_mobile);
 	B(is_i85x);
@@ -82,7 +81,6 @@ static int i915_capabilities(struct seq_file *m, void *data)
 	B(supports_tv);
 	B(has_bsd_ring);
 	B(has_blt_ring);
-	B(has_llc);
 #undef B
 
 	return 0;
@@ -100,12 +98,12 @@ static const char *get_pin_flag(struct drm_i915_gem_object *obj)
 
 static const char *get_tiling_flag(struct drm_i915_gem_object *obj)
 {
-	switch (obj->tiling_mode) {
-	default:
-	case I915_TILING_NONE: return " ";
-	case I915_TILING_X: return "X";
-	case I915_TILING_Y: return "Y";
-	}
+    switch (obj->tiling_mode) {
+    default:
+    case I915_TILING_NONE: return " ";
+    case I915_TILING_X: return "X";
+    case I915_TILING_Y: return "Y";
+    }
 }
 
 static const char *cache_level_str(int type)
@@ -125,7 +123,7 @@ describe_obj(struct seq_file *m, struct drm_i915_gem_object *obj)
 		   &obj->base,
 		   get_pin_flag(obj),
 		   get_tiling_flag(obj),
-		   obj->base.size / 1024,
+		   obj->base.size,
 		   obj->base.read_domains,
 		   obj->base.write_domain,
 		   obj->last_rendering_seqno,
@@ -219,7 +217,7 @@ static int i915_gem_object_list_info(struct seq_file *m, void *data)
 			++mappable_count; \
 		} \
 	} \
-} while (0)
+} while(0)
 
 static int i915_gem_object_info(struct seq_file *m, void* data)
 {
@@ -501,7 +499,7 @@ static int i915_interrupt_info(struct seq_file *m, void *data)
 	seq_printf(m, "Interrupts received: %d\n",
 		   atomic_read(&dev_priv->irq_received));
 	for (i = 0; i < I915_NUM_RINGS; i++) {
-		if (IS_GEN6(dev) || IS_GEN7(dev)) {
+		if (IS_GEN6(dev)) {
 			seq_printf(m, "Graphics Interrupt mask (%s):	%08x\n",
 				   dev_priv->ring[i].name,
 				   I915_READ_IMR(&dev_priv->ring[i]));
@@ -563,6 +561,45 @@ static int i915_hws_info(struct seq_file *m, void *data)
 	return 0;
 }
 
+static void i915_dump_object(struct seq_file *m,
+			     struct io_mapping *mapping,
+			     struct drm_i915_gem_object *obj)
+{
+	int page, page_count, i;
+
+	page_count = obj->base.size / PAGE_SIZE;
+	for (page = 0; page < page_count; page++) {
+		u32 *mem = io_mapping_map_wc(mapping,
+					     obj->gtt_offset + page * PAGE_SIZE);
+		for (i = 0; i < PAGE_SIZE; i += 4)
+			seq_printf(m, "%08x :  %08x\n", i, mem[i / 4]);
+		io_mapping_unmap(mem);
+	}
+}
+
+static int i915_batchbuffer_info(struct seq_file *m, void *data)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	drm_i915_private_t *dev_priv = dev->dev_private;
+	struct drm_i915_gem_object *obj;
+	int ret;
+
+	ret = mutex_lock_interruptible(&dev->struct_mutex);
+	if (ret)
+		return ret;
+
+	list_for_each_entry(obj, &dev_priv->mm.active_list, mm_list) {
+		if (obj->base.read_domains & I915_GEM_DOMAIN_COMMAND) {
+		    seq_printf(m, "--- gtt_offset = 0x%08x\n", obj->gtt_offset);
+		    i915_dump_object(m, dev_priv->mm.gtt_mapping, obj);
+		}
+	}
+
+	mutex_unlock(&dev->struct_mutex);
+	return 0;
+}
+
 static int i915_ringbuffer_data(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
@@ -598,40 +635,33 @@ static int i915_ringbuffer_info(struct seq_file *m, void *data)
 	struct drm_device *dev = node->minor->dev;
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	struct intel_ring_buffer *ring;
-	int ret;
 
 	ring = &dev_priv->ring[(uintptr_t)node->info_ent->data];
 	if (ring->size == 0)
 		return 0;
 
-	ret = mutex_lock_interruptible(&dev->struct_mutex);
-	if (ret)
-		return ret;
-
 	seq_printf(m, "Ring %s:\n", ring->name);
 	seq_printf(m, "  Head :    %08x\n", I915_READ_HEAD(ring) & HEAD_ADDR);
 	seq_printf(m, "  Tail :    %08x\n", I915_READ_TAIL(ring) & TAIL_ADDR);
 	seq_printf(m, "  Size :    %08x\n", ring->size);
 	seq_printf(m, "  Active :  %08x\n", intel_ring_get_active_head(ring));
 	seq_printf(m, "  NOPID :   %08x\n", I915_READ_NOPID(ring));
-	if (IS_GEN6(dev) || IS_GEN7(dev)) {
+	if (IS_GEN6(dev)) {
 		seq_printf(m, "  Sync 0 :   %08x\n", I915_READ_SYNC_0(ring));
 		seq_printf(m, "  Sync 1 :   %08x\n", I915_READ_SYNC_1(ring));
 	}
 	seq_printf(m, "  Control : %08x\n", I915_READ_CTL(ring));
 	seq_printf(m, "  Start :   %08x\n", I915_READ_START(ring));
 
-	mutex_unlock(&dev->struct_mutex);
-
 	return 0;
 }
 
 static const char *ring_str(int ring)
 {
 	switch (ring) {
-	case RCS: return "render";
-	case VCS: return "bsd";
-	case BCS: return "blt";
+	case RING_RENDER: return " render";
+	case RING_BSD: return " bsd";
+	case RING_BLT: return " blt";
 	default: return "";
 	}
 }
@@ -674,7 +704,7 @@ static void print_error_buffers(struct seq_file *m,
 	seq_printf(m, "%s [%d]:\n", name, count);
 
 	while (count--) {
-		seq_printf(m, "  %08x %8u %04x %04x %08x%s%s%s%s%s%s%s",
+		seq_printf(m, "  %08x %8u %04x %04x %08x%s%s%s%s%s%s",
 			   err->gtt_offset,
 			   err->size,
 			   err->read_domains,
@@ -684,7 +714,6 @@ static void print_error_buffers(struct seq_file *m,
 			   tiling_flag(err->tiling),
 			   dirty_flag(err->dirty),
 			   purgeable_flag(err->purgeable),
-			   err->ring != -1 ? " " : "",
 			   ring_str(err->ring),
 			   cache_level_str(err->cache_level));
 
@@ -698,38 +727,6 @@ static void print_error_buffers(struct seq_file *m,
 	}
 }
 
-static void i915_ring_error_state(struct seq_file *m,
-				  struct drm_device *dev,
-				  struct drm_i915_error_state *error,
-				  unsigned ring)
-{
-	seq_printf(m, "%s command stream:\n", ring_str(ring));
-	seq_printf(m, "  HEAD: 0x%08x\n", error->head[ring]);
-	seq_printf(m, "  TAIL: 0x%08x\n", error->tail[ring]);
-	seq_printf(m, "  ACTHD: 0x%08x\n", error->acthd[ring]);
-	seq_printf(m, "  IPEIR: 0x%08x\n", error->ipeir[ring]);
-	seq_printf(m, "  IPEHR: 0x%08x\n", error->ipehr[ring]);
-	seq_printf(m, "  INSTDONE: 0x%08x\n", error->instdone[ring]);
-	if (ring == RCS && INTEL_INFO(dev)->gen >= 4) {
-		seq_printf(m, "  INSTDONE1: 0x%08x\n", error->instdone1);
-		seq_printf(m, "  BBADDR: 0x%08llx\n", error->bbaddr);
-	}
-	if (INTEL_INFO(dev)->gen >= 4)
-		seq_printf(m, "  INSTPS: 0x%08x\n", error->instps[ring]);
-	seq_printf(m, "  INSTPM: 0x%08x\n", error->instpm[ring]);
-	if (INTEL_INFO(dev)->gen >= 6) {
-		seq_printf(m, "  FADDR: 0x%08x\n", error->faddr[ring]);
-		seq_printf(m, "  FAULT_REG: 0x%08x\n", error->fault_reg[ring]);
-		seq_printf(m, "  SYNC_0: 0x%08x\n",
-			   error->semaphore_mboxes[ring][0]);
-		seq_printf(m, "  SYNC_1: 0x%08x\n",
-			   error->semaphore_mboxes[ring][1]);
-	}
-	seq_printf(m, "  seqno: 0x%08x\n", error->seqno[ring]);
-	seq_printf(m, "  ring->head: 0x%08x\n", error->cpu_ring_head[ring]);
-	seq_printf(m, "  ring->tail: 0x%08x\n", error->cpu_ring_tail[ring]);
-}
-
 static int i915_error_state(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
@@ -737,7 +734,7 @@ static int i915_error_state(struct seq_file *m, void *unused)
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	struct drm_i915_error_state *error;
 	unsigned long flags;
-	int i, j, page, offset, elt;
+	int i, page, offset, elt;
 
 	spin_lock_irqsave(&dev_priv->error_lock, flags);
 	if (!dev_priv->first_error) {
@@ -752,20 +749,35 @@ static int i915_error_state(struct seq_file *m, void *unused)
 	seq_printf(m, "PCI ID: 0x%04x\n", dev->pci_device);
 	seq_printf(m, "EIR: 0x%08x\n", error->eir);
 	seq_printf(m, "PGTBL_ER: 0x%08x\n", error->pgtbl_er);
-
-	for (i = 0; i < dev_priv->num_fence_regs; i++)
-		seq_printf(m, "  fence[%d] = %08llx\n", i, error->fence[i]);
-
 	if (INTEL_INFO(dev)->gen >= 6) {
 		seq_printf(m, "ERROR: 0x%08x\n", error->error);
-		seq_printf(m, "DONE_REG: 0x%08x\n", error->done_reg);
+		seq_printf(m, "Blitter command stream:\n");
+		seq_printf(m, "  ACTHD:    0x%08x\n", error->bcs_acthd);
+		seq_printf(m, "  IPEIR:    0x%08x\n", error->bcs_ipeir);
+		seq_printf(m, "  IPEHR:    0x%08x\n", error->bcs_ipehr);
+		seq_printf(m, "  INSTDONE: 0x%08x\n", error->bcs_instdone);
+		seq_printf(m, "  seqno:    0x%08x\n", error->bcs_seqno);
+		seq_printf(m, "Video (BSD) command stream:\n");
+		seq_printf(m, "  ACTHD:    0x%08x\n", error->vcs_acthd);
+		seq_printf(m, "  IPEIR:    0x%08x\n", error->vcs_ipeir);
+		seq_printf(m, "  IPEHR:    0x%08x\n", error->vcs_ipehr);
+		seq_printf(m, "  INSTDONE: 0x%08x\n", error->vcs_instdone);
+		seq_printf(m, "  seqno:    0x%08x\n", error->vcs_seqno);
 	}
+	seq_printf(m, "Render command stream:\n");
+	seq_printf(m, "  ACTHD: 0x%08x\n", error->acthd);
+	seq_printf(m, "  IPEIR: 0x%08x\n", error->ipeir);
+	seq_printf(m, "  IPEHR: 0x%08x\n", error->ipehr);
+	seq_printf(m, "  INSTDONE: 0x%08x\n", error->instdone);
+	if (INTEL_INFO(dev)->gen >= 4) {
+		seq_printf(m, "  INSTDONE1: 0x%08x\n", error->instdone1);
+		seq_printf(m, "  INSTPS: 0x%08x\n", error->instps);
+	}
+	seq_printf(m, "  INSTPM: 0x%08x\n", error->instpm);
+	seq_printf(m, "  seqno: 0x%08x\n", error->seqno);
 
-	i915_ring_error_state(m, dev, error, RCS);
-	if (HAS_BLT(dev))
-		i915_ring_error_state(m, dev, error, BCS);
-	if (HAS_BSD(dev))
-		i915_ring_error_state(m, dev, error, VCS);
+	for (i = 0; i < dev_priv->num_fence_regs; i++)
+		seq_printf(m, "  fence[%d] = %08llx\n", i, error->fence[i]);
 
 	if (error->active_bo)
 		print_error_buffers(m, "Active",
@@ -777,10 +789,10 @@ static int i915_error_state(struct seq_file *m, void *unused)
 				    error->pinned_bo,
 				    error->pinned_bo_count);
 
-	for (i = 0; i < ARRAY_SIZE(error->ring); i++) {
-		struct drm_i915_error_object *obj;
+	for (i = 0; i < ARRAY_SIZE(error->batchbuffer); i++) {
+		if (error->batchbuffer[i]) {
+			struct drm_i915_error_object *obj = error->batchbuffer[i];
 
-		if ((obj = error->ring[i].batchbuffer)) {
 			seq_printf(m, "%s --- gtt_offset = 0x%08x\n",
 				   dev_priv->ring[i].name,
 				   obj->gtt_offset);
@@ -792,20 +804,11 @@ static int i915_error_state(struct seq_file *m, void *unused)
 				}
 			}
 		}
+	}
 
-		if (error->ring[i].num_requests) {
-			seq_printf(m, "%s --- %d requests\n",
-				   dev_priv->ring[i].name,
-				   error->ring[i].num_requests);
-			for (j = 0; j < error->ring[i].num_requests; j++) {
-				seq_printf(m, "  seqno 0x%08x, emitted %ld, tail 0x%08x\n",
-					   error->ring[i].requests[j].seqno,
-					   error->ring[i].requests[j].jiffies,
-					   error->ring[i].requests[j].tail);
-			}
-		}
-
-		if ((obj = error->ring[i].ringbuffer)) {
+	for (i = 0; i < ARRAY_SIZE(error->ringbuffer); i++) {
+		if (error->ringbuffer[i]) {
+			struct drm_i915_error_object *obj = error->ringbuffer[i];
 			seq_printf(m, "%s --- ringbuffer = 0x%08x\n",
 				   dev_priv->ring[i].name,
 				   obj->gtt_offset);
@@ -838,16 +841,7 @@ static int i915_rstdby_delays(struct seq_file *m, void *unused)
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
 	struct drm_device *dev = node->minor->dev;
 	drm_i915_private_t *dev_priv = dev->dev_private;
-	u16 crstanddelay;
-	int ret;
-
-	ret = mutex_lock_interruptible(&dev->struct_mutex);
-	if (ret)
-		return ret;
-
-	crstanddelay = I915_READ16(CRSTANDVID);
-
-	mutex_unlock(&dev->struct_mutex);
+	u16 crstanddelay = I915_READ16(CRSTANDVID);
 
 	seq_printf(m, "w/ctx: %d, w/o ctx: %d\n", (crstanddelay >> 8) & 0x3f, (crstanddelay & 0x3f));
 
@@ -945,11 +939,7 @@ static int i915_delayfreq_table(struct seq_file *m, void *unused)
 	struct drm_device *dev = node->minor->dev;
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	u32 delayfreq;
-	int ret, i;
-
-	ret = mutex_lock_interruptible(&dev->struct_mutex);
-	if (ret)
-		return ret;
+	int i;
 
 	for (i = 0; i < 16; i++) {
 		delayfreq = I915_READ(PXVFREQ_BASE + i * 4);
@@ -957,8 +947,6 @@ static int i915_delayfreq_table(struct seq_file *m, void *unused)
 			   (delayfreq & PXVFREQ_PX_MASK) >> PXVFREQ_PX_SHIFT);
 	}
 
-	mutex_unlock(&dev->struct_mutex);
-
 	return 0;
 }
 
@@ -973,40 +961,24 @@ static int i915_inttoext_table(struct seq_file *m, void *unused)
 	struct drm_device *dev = node->minor->dev;
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	u32 inttoext;
-	int ret, i;
-
-	ret = mutex_lock_interruptible(&dev->struct_mutex);
-	if (ret)
-		return ret;
+	int i;
 
 	for (i = 1; i <= 32; i++) {
 		inttoext = I915_READ(INTTOEXT_BASE_ILK + i * 4);
 		seq_printf(m, "INTTOEXT%02d: 0x%08x\n", i, inttoext);
 	}
 
-	mutex_unlock(&dev->struct_mutex);
-
 	return 0;
 }
 
-static int ironlake_drpc_info(struct seq_file *m)
+static int i915_drpc_info(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
 	struct drm_device *dev = node->minor->dev;
 	drm_i915_private_t *dev_priv = dev->dev_private;
-	u32 rgvmodectl, rstdbyctl;
-	u16 crstandvid;
-	int ret;
-
-	ret = mutex_lock_interruptible(&dev->struct_mutex);
-	if (ret)
-		return ret;
-
-	rgvmodectl = I915_READ(MEMMODECTL);
-	rstdbyctl = I915_READ(RSTDBYCTL);
-	crstandvid = I915_READ16(CRSTANDVID);
-
-	mutex_unlock(&dev->struct_mutex);
+	u32 rgvmodectl = I915_READ(MEMMODECTL);
+	u32 rstdbyctl = I915_READ(RSTDBYCTL);
+	u16 crstandvid = I915_READ16(CRSTANDVID);
 
 	seq_printf(m, "HD boost: %s\n", (rgvmodectl & MEMMODE_BOOST_EN) ?
 		   "yes" : "no");
@@ -1056,95 +1028,6 @@ static int ironlake_drpc_info(struct seq_file *m)
 	return 0;
 }
 
-static int gen6_drpc_info(struct seq_file *m)
-{
-
-	struct drm_info_node *node = (struct drm_info_node *) m->private;
-	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 rpmodectl1, gt_core_status, rcctl1;
-	unsigned forcewake_count;
-	int count=0, ret;
-
-
-	ret = mutex_lock_interruptible(&dev->struct_mutex);
-	if (ret)
-		return ret;
-
-	spin_lock_irq(&dev_priv->gt_lock);
-	forcewake_count = dev_priv->forcewake_count;
-	spin_unlock_irq(&dev_priv->gt_lock);
-
-	if (forcewake_count) {
-		seq_printf(m, "RC information inaccurate because somebody "
-			      "holds a forcewake reference \n");
-	} else {
-		/* NB: we cannot use forcewake, else we read the wrong values */
-		while (count++ < 50 && (I915_READ_NOTRACE(FORCEWAKE_ACK) & 1))
-			udelay(10);
-		seq_printf(m, "RC information accurate: %s\n", yesno(count < 51));
-	}
-
-	gt_core_status = readl(dev_priv->regs + GEN6_GT_CORE_STATUS);
-	trace_i915_reg_rw(false, GEN6_GT_CORE_STATUS, gt_core_status, 4);
-
-	rpmodectl1 = I915_READ(GEN6_RP_CONTROL);
-	rcctl1 = I915_READ(GEN6_RC_CONTROL);
-	mutex_unlock(&dev->struct_mutex);
-
-	seq_printf(m, "Video Turbo Mode: %s\n",
-		   yesno(rpmodectl1 & GEN6_RP_MEDIA_TURBO));
-	seq_printf(m, "HW control enabled: %s\n",
-		   yesno(rpmodectl1 & GEN6_RP_ENABLE));
-	seq_printf(m, "SW control enabled: %s\n",
-		   yesno((rpmodectl1 & GEN6_RP_MEDIA_MODE_MASK) ==
-			  GEN6_RP_MEDIA_SW_MODE));
-	seq_printf(m, "RC1e Enabled: %s\n",
-		   yesno(rcctl1 & GEN6_RC_CTL_RC1e_ENABLE));
-	seq_printf(m, "RC6 Enabled: %s\n",
-		   yesno(rcctl1 & GEN6_RC_CTL_RC6_ENABLE));
-	seq_printf(m, "Deep RC6 Enabled: %s\n",
-		   yesno(rcctl1 & GEN6_RC_CTL_RC6p_ENABLE));
-	seq_printf(m, "Deepest RC6 Enabled: %s\n",
-		   yesno(rcctl1 & GEN6_RC_CTL_RC6pp_ENABLE));
-	seq_printf(m, "Current RC state: ");
-	switch (gt_core_status & GEN6_RCn_MASK) {
-	case GEN6_RC0:
-		if (gt_core_status & GEN6_CORE_CPD_STATE_MASK)
-			seq_printf(m, "Core Power Down\n");
-		else
-			seq_printf(m, "on\n");
-		break;
-	case GEN6_RC3:
-		seq_printf(m, "RC3\n");
-		break;
-	case GEN6_RC6:
-		seq_printf(m, "RC6\n");
-		break;
-	case GEN6_RC7:
-		seq_printf(m, "RC7\n");
-		break;
-	default:
-		seq_printf(m, "Unknown\n");
-		break;
-	}
-
-	seq_printf(m, "Core Power Down: %s\n",
-		   yesno(gt_core_status & GEN6_CORE_CPD_STATE_MASK));
-	return 0;
-}
-
-static int i915_drpc_info(struct seq_file *m, void *unused)
-{
-	struct drm_info_node *node = (struct drm_info_node *) m->private;
-	struct drm_device *dev = node->minor->dev;
-
-	if (IS_GEN6(dev) || IS_GEN7(dev))
-		return gen6_drpc_info(m);
-	else
-		return ironlake_drpc_info(m);
-}
-
 static int i915_fbc_status(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
@@ -1240,59 +1123,14 @@ static int i915_emon_status(struct seq_file *m, void *unused)
 	return 0;
 }
 
-static int i915_ring_freq_table(struct seq_file *m, void *unused)
-{
-	struct drm_info_node *node = (struct drm_info_node *) m->private;
-	struct drm_device *dev = node->minor->dev;
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	int ret;
-	int gpu_freq, ia_freq;
-
-	if (!(IS_GEN6(dev) || IS_GEN7(dev))) {
-		seq_printf(m, "unsupported on this chipset\n");
-		return 0;
-	}
-
-	ret = mutex_lock_interruptible(&dev->struct_mutex);
-	if (ret)
-		return ret;
-
-	seq_printf(m, "GPU freq (MHz)\tEffective CPU freq (MHz)\n");
-
-	for (gpu_freq = dev_priv->min_delay; gpu_freq <= dev_priv->max_delay;
-	     gpu_freq++) {
-		I915_WRITE(GEN6_PCODE_DATA, gpu_freq);
-		I915_WRITE(GEN6_PCODE_MAILBOX, GEN6_PCODE_READY |
-			   GEN6_PCODE_READ_MIN_FREQ_TABLE);
-		if (wait_for((I915_READ(GEN6_PCODE_MAILBOX) &
-			      GEN6_PCODE_READY) == 0, 10)) {
-			DRM_ERROR("pcode read of freq table timed out\n");
-			continue;
-		}
-		ia_freq = I915_READ(GEN6_PCODE_DATA);
-		seq_printf(m, "%d\t\t%d\n", gpu_freq * 50, ia_freq * 100);
-	}
-
-	mutex_unlock(&dev->struct_mutex);
-
-	return 0;
-}
-
 static int i915_gfxec(struct seq_file *m, void *unused)
 {
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
 	struct drm_device *dev = node->minor->dev;
 	drm_i915_private_t *dev_priv = dev->dev_private;
-	int ret;
-
-	ret = mutex_lock_interruptible(&dev->struct_mutex);
-	if (ret)
-		return ret;
 
 	seq_printf(m, "GFXEC: %ld\n", (unsigned long)I915_READ(0x112f4));
 
-	mutex_unlock(&dev->struct_mutex);
-
 	return 0;
 }
 
@@ -1391,119 +1229,16 @@ static int i915_gen6_forcewake_count_info(struct seq_file *m, void *data)
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
 	struct drm_device *dev = node->minor->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	unsigned forcewake_count;
 
-	spin_lock_irq(&dev_priv->gt_lock);
-	forcewake_count = dev_priv->forcewake_count;
-	spin_unlock_irq(&dev_priv->gt_lock);
-
-	seq_printf(m, "forcewake count = %u\n", forcewake_count);
-
-	return 0;
-}
-
-static const char *swizzle_string(unsigned swizzle)
-{
-	switch(swizzle) {
-	case I915_BIT_6_SWIZZLE_NONE:
-		return "none";
-	case I915_BIT_6_SWIZZLE_9:
-		return "bit9";
-	case I915_BIT_6_SWIZZLE_9_10:
-		return "bit9/bit10";
-	case I915_BIT_6_SWIZZLE_9_11:
-		return "bit9/bit11";
-	case I915_BIT_6_SWIZZLE_9_10_11:
-		return "bit9/bit10/bit11";
-	case I915_BIT_6_SWIZZLE_9_17:
-		return "bit9/bit17";
-	case I915_BIT_6_SWIZZLE_9_10_17:
-		return "bit9/bit10/bit17";
-	case I915_BIT_6_SWIZZLE_UNKNOWN:
-		return "unkown";
-	}
-
-	return "bug";
-}
-
-static int i915_swizzle_info(struct seq_file *m, void *data)
-{
-	struct drm_info_node *node = (struct drm_info_node *) m->private;
-	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	mutex_lock(&dev->struct_mutex);
-	seq_printf(m, "bit6 swizzle for X-tiling = %s\n",
-		   swizzle_string(dev_priv->mm.bit_6_swizzle_x));
-	seq_printf(m, "bit6 swizzle for Y-tiling = %s\n",
-		   swizzle_string(dev_priv->mm.bit_6_swizzle_y));
-
-	if (IS_GEN3(dev) || IS_GEN4(dev)) {
-		seq_printf(m, "DDC = 0x%08x\n",
-			   I915_READ(DCC));
-		seq_printf(m, "C0DRB3 = 0x%04x\n",
-			   I915_READ16(C0DRB3));
-		seq_printf(m, "C1DRB3 = 0x%04x\n",
-			   I915_READ16(C1DRB3));
-	} else if (IS_GEN6(dev) || IS_GEN7(dev)) {
-		seq_printf(m, "MAD_DIMM_C0 = 0x%08x\n",
-			   I915_READ(MAD_DIMM_C0));
-		seq_printf(m, "MAD_DIMM_C1 = 0x%08x\n",
-			   I915_READ(MAD_DIMM_C1));
-		seq_printf(m, "MAD_DIMM_C2 = 0x%08x\n",
-			   I915_READ(MAD_DIMM_C2));
-		seq_printf(m, "TILECTL = 0x%08x\n",
-			   I915_READ(TILECTL));
-		seq_printf(m, "ARB_MODE = 0x%08x\n",
-			   I915_READ(ARB_MODE));
-		seq_printf(m, "DISP_ARB_CTL = 0x%08x\n",
-			   I915_READ(DISP_ARB_CTL));
-	}
-	mutex_unlock(&dev->struct_mutex);
-
-	return 0;
-}
-
-static int i915_ppgtt_info(struct seq_file *m, void *data)
-{
-	struct drm_info_node *node = (struct drm_info_node *) m->private;
-	struct drm_device *dev = node->minor->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_ring_buffer *ring;
-	int i, ret;
-
-
-	ret = mutex_lock_interruptible(&dev->struct_mutex);
-	if (ret)
-		return ret;
-	if (INTEL_INFO(dev)->gen == 6)
-		seq_printf(m, "GFX_MODE: 0x%08x\n", I915_READ(GFX_MODE));
-
-	for (i = 0; i < I915_NUM_RINGS; i++) {
-		ring = &dev_priv->ring[i];
-
-		seq_printf(m, "%s\n", ring->name);
-		if (INTEL_INFO(dev)->gen == 7)
-			seq_printf(m, "GFX_MODE: 0x%08x\n", I915_READ(RING_MODE_GEN7(ring)));
-		seq_printf(m, "PP_DIR_BASE: 0x%08x\n", I915_READ(RING_PP_DIR_BASE(ring)));
-		seq_printf(m, "PP_DIR_BASE_READ: 0x%08x\n", I915_READ(RING_PP_DIR_BASE_READ(ring)));
-		seq_printf(m, "PP_DIR_DCLV: 0x%08x\n", I915_READ(RING_PP_DIR_DCLV(ring)));
-	}
-	if (dev_priv->mm.aliasing_ppgtt) {
-		struct i915_hw_ppgtt *ppgtt = dev_priv->mm.aliasing_ppgtt;
-
-		seq_printf(m, "aliasing PPGTT:\n");
-		seq_printf(m, "pd gtt offset: 0x%08x\n", ppgtt->pd_offset);
-	}
-	seq_printf(m, "ECOCHK: 0x%08x\n", I915_READ(GAM_ECOCHK));
-	mutex_unlock(&dev->struct_mutex);
+	seq_printf(m, "forcewake count = %d\n",
+		   atomic_read(&dev_priv->forcewake_count));
 
 	return 0;
 }
 
 static int
-i915_debugfs_common_open(struct inode *inode,
-			 struct file *filp)
+i915_wedged_open(struct inode *inode,
+		 struct file *filp)
 {
 	filp->private_data = inode->i_private;
 	return 0;
@@ -1520,12 +1255,12 @@ i915_wedged_read(struct file *filp,
 	char buf[80];
 	int len;
 
-	len = snprintf(buf, sizeof(buf),
+	len = snprintf(buf, sizeof (buf),
 		       "wedged :  %d\n",
 		       atomic_read(&dev_priv->mm.wedged));
 
-	if (len > sizeof(buf))
-		len = sizeof(buf);
+	if (len > sizeof (buf))
+		len = sizeof (buf);
 
 	return simple_read_from_buffer(ubuf, max, ppos, buf, len);
 }
@@ -1541,7 +1276,7 @@ i915_wedged_write(struct file *filp,
 	int val = 1;
 
 	if (cnt > 0) {
-		if (cnt > sizeof(buf) - 1)
+		if (cnt > sizeof (buf) - 1)
 			return -EINVAL;
 
 		if (copy_from_user(buf, ubuf, cnt))
@@ -1559,145 +1294,12 @@ i915_wedged_write(struct file *filp,
 
 static const struct file_operations i915_wedged_fops = {
 	.owner = THIS_MODULE,
-	.open = i915_debugfs_common_open,
+	.open = i915_wedged_open,
 	.read = i915_wedged_read,
 	.write = i915_wedged_write,
 	.llseek = default_llseek,
 };
 
-static ssize_t
-i915_max_freq_read(struct file *filp,
-		   char __user *ubuf,
-		   size_t max,
-		   loff_t *ppos)
-{
-	struct drm_device *dev = filp->private_data;
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	char buf[80];
-	int len;
-
-	len = snprintf(buf, sizeof(buf),
-		       "max freq: %d\n", dev_priv->max_delay * 50);
-
-	if (len > sizeof(buf))
-		len = sizeof(buf);
-
-	return simple_read_from_buffer(ubuf, max, ppos, buf, len);
-}
-
-static ssize_t
-i915_max_freq_write(struct file *filp,
-		  const char __user *ubuf,
-		  size_t cnt,
-		  loff_t *ppos)
-{
-	struct drm_device *dev = filp->private_data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	char buf[20];
-	int val = 1;
-
-	if (cnt > 0) {
-		if (cnt > sizeof(buf) - 1)
-			return -EINVAL;
-
-		if (copy_from_user(buf, ubuf, cnt))
-			return -EFAULT;
-		buf[cnt] = 0;
-
-		val = simple_strtoul(buf, NULL, 0);
-	}
-
-	DRM_DEBUG_DRIVER("Manually setting max freq to %d\n", val);
-
-	/*
-	 * Turbo will still be enabled, but won't go above the set value.
-	 */
-	dev_priv->max_delay = val / 50;
-
-	gen6_set_rps(dev, val / 50);
-
-	return cnt;
-}
-
-static const struct file_operations i915_max_freq_fops = {
-	.owner = THIS_MODULE,
-	.open = i915_debugfs_common_open,
-	.read = i915_max_freq_read,
-	.write = i915_max_freq_write,
-	.llseek = default_llseek,
-};
-
-static ssize_t
-i915_cache_sharing_read(struct file *filp,
-		   char __user *ubuf,
-		   size_t max,
-		   loff_t *ppos)
-{
-	struct drm_device *dev = filp->private_data;
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	char buf[80];
-	u32 snpcr;
-	int len;
-
-	mutex_lock(&dev_priv->dev->struct_mutex);
-	snpcr = I915_READ(GEN6_MBCUNIT_SNPCR);
-	mutex_unlock(&dev_priv->dev->struct_mutex);
-
-	len = snprintf(buf, sizeof(buf),
-		       "%d\n", (snpcr & GEN6_MBC_SNPCR_MASK) >>
-		       GEN6_MBC_SNPCR_SHIFT);
-
-	if (len > sizeof(buf))
-		len = sizeof(buf);
-
-	return simple_read_from_buffer(ubuf, max, ppos, buf, len);
-}
-
-static ssize_t
-i915_cache_sharing_write(struct file *filp,
-		  const char __user *ubuf,
-		  size_t cnt,
-		  loff_t *ppos)
-{
-	struct drm_device *dev = filp->private_data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	char buf[20];
-	u32 snpcr;
-	int val = 1;
-
-	if (cnt > 0) {
-		if (cnt > sizeof(buf) - 1)
-			return -EINVAL;
-
-		if (copy_from_user(buf, ubuf, cnt))
-			return -EFAULT;
-		buf[cnt] = 0;
-
-		val = simple_strtoul(buf, NULL, 0);
-	}
-
-	if (val < 0 || val > 3)
-		return -EINVAL;
-
-	DRM_DEBUG_DRIVER("Manually setting uncore sharing to %d\n", val);
-
-	/* Update the cache sharing policy here as well */
-	snpcr = I915_READ(GEN6_MBCUNIT_SNPCR);
-	snpcr &= ~GEN6_MBC_SNPCR_MASK;
-	snpcr |= (val << GEN6_MBC_SNPCR_SHIFT);
-	I915_WRITE(GEN6_MBCUNIT_SNPCR, snpcr);
-
-	return cnt;
-}
-
-static const struct file_operations i915_cache_sharing_fops = {
-	.owner = THIS_MODULE,
-	.open = i915_debugfs_common_open,
-	.read = i915_cache_sharing_read,
-	.write = i915_cache_sharing_write,
-	.llseek = default_llseek,
-};
-
 /* As the drm_debugfs_init() routines are called before dev->dev_private is
  * allocated we need to hook into the minor for release. */
 static int
@@ -1716,21 +1318,33 @@ drm_add_fake_info_node(struct drm_minor *minor,
 	node->minor = minor;
 	node->dent = ent;
 	node->info_ent = (void *) key;
-
-	mutex_lock(&minor->debugfs_lock);
-	list_add(&node->list, &minor->debugfs_list);
-	mutex_unlock(&minor->debugfs_lock);
+	list_add(&node->list, &minor->debugfs_nodes.list);
 
 	return 0;
 }
 
+static int i915_wedged_create(struct dentry *root, struct drm_minor *minor)
+{
+	struct drm_device *dev = minor->dev;
+	struct dentry *ent;
+
+	ent = debugfs_create_file("i915_wedged",
+				  S_IRUGO | S_IWUSR,
+				  root, dev,
+				  &i915_wedged_fops);
+	if (IS_ERR(ent))
+		return PTR_ERR(ent);
+
+	return drm_add_fake_info_node(minor, ent, &i915_wedged_fops);
+}
+
 static int i915_forcewake_open(struct inode *inode, struct file *file)
 {
 	struct drm_device *dev = inode->i_private;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret;
 
-	if (INTEL_INFO(dev)->gen < 6)
+	if (!IS_GEN6(dev))
 		return 0;
 
 	ret = mutex_lock_interruptible(&dev->struct_mutex);
@@ -1747,7 +1361,7 @@ int i915_forcewake_release(struct inode *inode, struct file *file)
 	struct drm_device *dev = inode->i_private;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	if (INTEL_INFO(dev)->gen < 6)
+	if (!IS_GEN6(dev))
 		return 0;
 
 	/*
@@ -1785,24 +1399,6 @@ static int i915_forcewake_create(struct dentry *root, struct drm_minor *minor)
 	return drm_add_fake_info_node(minor, ent, &i915_forcewake_fops);
 }
 
-static int i915_debugfs_create(struct dentry *root,
-			       struct drm_minor *minor,
-			       const char *name,
-			       const struct file_operations *fops)
-{
-	struct drm_device *dev = minor->dev;
-	struct dentry *ent;
-
-	ent = debugfs_create_file(name,
-				  S_IRUGO | S_IWUSR,
-				  root, dev,
-				  fops);
-	if (IS_ERR(ent))
-		return PTR_ERR(ent);
-
-	return drm_add_fake_info_node(minor, ent, fops);
-}
-
 static struct drm_info_list i915_debugfs_list[] = {
 	{"i915_capabilities", i915_capabilities, 0},
 	{"i915_gem_objects", i915_gem_object_info, 0},
@@ -1826,6 +1422,7 @@ static struct drm_info_list i915_debugfs_list[] = {
 	{"i915_bsd_ringbuffer_info", i915_ringbuffer_info, 0, (void *)VCS},
 	{"i915_blt_ringbuffer_data", i915_ringbuffer_data, 0, (void *)BCS},
 	{"i915_blt_ringbuffer_info", i915_ringbuffer_info, 0, (void *)BCS},
+	{"i915_batchbuffers", i915_batchbuffer_info, 0},
 	{"i915_error_state", i915_error_state, 0},
 	{"i915_rstdby_delays", i915_rstdby_delays, 0},
 	{"i915_cur_delayinfo", i915_cur_delayinfo, 0},
@@ -1833,7 +1430,6 @@ static struct drm_info_list i915_debugfs_list[] = {
 	{"i915_inttoext_table", i915_inttoext_table, 0},
 	{"i915_drpc_info", i915_drpc_info, 0},
 	{"i915_emon_status", i915_emon_status, 0},
-	{"i915_ring_freq_table", i915_ring_freq_table, 0},
 	{"i915_gfxec", i915_gfxec, 0},
 	{"i915_fbc_status", i915_fbc_status, 0},
 	{"i915_sr_status", i915_sr_status, 0},
@@ -1841,8 +1437,6 @@ static struct drm_info_list i915_debugfs_list[] = {
 	{"i915_gem_framebuffer", i915_gem_framebuffer_info, 0},
 	{"i915_context_status", i915_context_status, 0},
 	{"i915_gen6_forcewake_count", i915_gen6_forcewake_count_info, 0},
-	{"i915_swizzle_info", i915_swizzle_info, 0},
-	{"i915_ppgtt_info", i915_ppgtt_info, 0},
 };
 #define I915_DEBUGFS_ENTRIES ARRAY_SIZE(i915_debugfs_list)
 
@@ -1850,9 +1444,7 @@ int i915_debugfs_init(struct drm_minor *minor)
 {
 	int ret;
 
-	ret = i915_debugfs_create(minor->debugfs_root, minor,
-				  "i915_wedged",
-				  &i915_wedged_fops);
+	ret = i915_wedged_create(minor->debugfs_root, minor);
 	if (ret)
 		return ret;
 
@@ -1860,18 +1452,6 @@ int i915_debugfs_init(struct drm_minor *minor)
 	if (ret)
 		return ret;
 
-	ret = i915_debugfs_create(minor->debugfs_root, minor,
-				  "i915_max_freq",
-				  &i915_max_freq_fops);
-	if (ret)
-		return ret;
-
-	ret = i915_debugfs_create(minor->debugfs_root, minor,
-				  "i915_cache_sharing",
-				  &i915_cache_sharing_fops);
-	if (ret)
-		return ret;
-
 	return drm_debugfs_create_files(i915_debugfs_list,
 					I915_DEBUGFS_ENTRIES,
 					minor->debugfs_root, minor);
@@ -1885,10 +1465,6 @@ void i915_debugfs_cleanup(struct drm_minor *minor)
 				 1, minor);
 	drm_debugfs_remove_files((struct drm_info_list *) &i915_wedged_fops,
 				 1, minor);
-	drm_debugfs_remove_files((struct drm_info_list *) &i915_max_freq_fops,
-				 1, minor);
-	drm_debugfs_remove_files((struct drm_info_list *) &i915_cache_sharing_fops,
-				 1, minor);
 }
 
 #endif /* CONFIG_DEBUG_FS */
diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
index 9341eb8..ef16443 100644
--- a/drivers/gpu/drm/i915/i915_dma.c
+++ b/drivers/gpu/drm/i915/i915_dma.c
@@ -41,7 +41,6 @@
 #include <linux/pnp.h>
 #include <linux/vga_switcheroo.h>
 #include <linux/slab.h>
-#include <linux/module.h>
 #include <acpi/video.h>
 
 static void i915_write_hws_pga(struct drm_device *dev)
@@ -781,12 +780,6 @@ static int i915_getparam(struct drm_device *dev, void *data,
 	case I915_PARAM_HAS_RELAXED_DELTA:
 		value = 1;
 		break;
-	case I915_PARAM_HAS_GEN7_SOL_RESET:
-		value = 1;
-		break;
-	case I915_PARAM_HAS_LLC:
-		value = HAS_LLC(dev);
-		break;
 	default:
 		DRM_DEBUG_DRIVER("Unknown parameter %d\n",
 				 param->param);
@@ -891,7 +884,7 @@ static int i915_get_bridge_dev(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	dev_priv->bridge_dev = pci_get_bus_and_slot(0, PCI_DEVFN(0, 0));
+	dev_priv->bridge_dev = pci_get_bus_and_slot(0, PCI_DEVFN(0,0));
 	if (!dev_priv->bridge_dev) {
 		DRM_ERROR("bridge device not found\n");
 		return -1;
@@ -1078,9 +1071,6 @@ static void i915_setup_compression(struct drm_device *dev, int size)
 	unsigned long cfb_base;
 	unsigned long ll_base = 0;
 
-	/* Just in case the BIOS is doing something questionable. */
-	intel_disable_fbc(dev);
-
 	compressed_fb = drm_mm_search_free(&dev_priv->mm.stolen, size, 4096, 0);
 	if (compressed_fb)
 		compressed_fb = drm_mm_get_block(compressed_fb, size, 4096);
@@ -1107,6 +1097,7 @@ static void i915_setup_compression(struct drm_device *dev, int size)
 
 	dev_priv->cfb_size = size;
 
+	intel_disable_fbc(dev);
 	dev_priv->compressed_fb = compressed_fb;
 	if (HAS_PCH_SPLIT(dev))
 		I915_WRITE(ILK_DPFC_CB_BASE, compressed_fb->start);
@@ -1196,39 +1187,22 @@ static int i915_load_gem_init(struct drm_device *dev)
 	/* Basic memrange allocator for stolen space */
 	drm_mm_init(&dev_priv->mm.stolen, 0, prealloc_size);
 
-	mutex_lock(&dev->struct_mutex);
-	if (i915_enable_ppgtt && HAS_ALIASING_PPGTT(dev)) {
-		/* PPGTT pdes are stolen from global gtt ptes, so shrink the
-		 * aperture accordingly when using aliasing ppgtt. */
-		gtt_size -= I915_PPGTT_PD_ENTRIES*PAGE_SIZE;
-		/* For paranoia keep the guard page in between. */
-		gtt_size -= PAGE_SIZE;
-
-		i915_gem_do_init(dev, 0, mappable_size, gtt_size);
-
-		ret = i915_gem_init_aliasing_ppgtt(dev);
-		if (ret)
-			return ret;
-	} else {
-		/* Let GEM Manage all of the aperture.
-		 *
-		 * However, leave one page at the end still bound to the scratch
-		 * page.  There are a number of places where the hardware
-		 * apparently prefetches past the end of the object, and we've
-		 * seen multiple hangs with the GPU head pointer stuck in a
-		 * batchbuffer bound at the last page of the aperture.  One page
-		 * should be enough to keep any prefetching inside of the
-		 * aperture.
-		 */
-		i915_gem_do_init(dev, 0, mappable_size, gtt_size - PAGE_SIZE);
-	}
+	/* Let GEM Manage all of the aperture.
+	 *
+	 * However, leave one page at the end still bound to the scratch page.
+	 * There are a number of places where the hardware apparently
+	 * prefetches past the end of the object, and we've seen multiple
+	 * hangs with the GPU head pointer stuck in a batchbuffer bound
+	 * at the last page of the aperture.  One page should be enough to
+	 * keep any prefetching inside of the aperture.
+	 */
+	i915_gem_do_init(dev, 0, mappable_size, gtt_size - PAGE_SIZE);
 
-	ret = i915_gem_init_hw(dev);
+	mutex_lock(&dev->struct_mutex);
+	ret = i915_gem_init_ringbuffer(dev);
 	mutex_unlock(&dev->struct_mutex);
-	if (ret) {
-		i915_gem_cleanup_aliasing_ppgtt(dev);
+	if (ret)
 		return ret;
-	}
 
 	/* Try to set up FBC with a reasonable compressed buffer size */
 	if (I915_HAS_FBC(dev) && i915_powersave) {
@@ -1315,7 +1289,6 @@ cleanup_gem:
 	mutex_lock(&dev->struct_mutex);
 	i915_gem_cleanup_ringbuffer(dev);
 	mutex_unlock(&dev->struct_mutex);
-	i915_gem_cleanup_aliasing_ppgtt(dev);
 cleanup_vga_switcheroo:
 	vga_switcheroo_unregister_client(dev->pdev);
 cleanup_vga_client:
@@ -1765,10 +1738,10 @@ static DEFINE_SPINLOCK(mchdev_lock);
  */
 unsigned long i915_read_mch_val(void)
 {
-	struct drm_i915_private *dev_priv;
+  	struct drm_i915_private *dev_priv;
 	unsigned long chipset_val, graphics_val, ret = 0;
 
-	spin_lock(&mchdev_lock);
+  	spin_lock(&mchdev_lock);
 	if (!i915_mch_dev)
 		goto out_unlock;
 	dev_priv = i915_mch_dev;
@@ -1779,9 +1752,9 @@ unsigned long i915_read_mch_val(void)
 	ret = chipset_val + graphics_val;
 
 out_unlock:
-	spin_unlock(&mchdev_lock);
+  	spin_unlock(&mchdev_lock);
 
-	return ret;
+  	return ret;
 }
 EXPORT_SYMBOL_GPL(i915_read_mch_val);
 
@@ -1792,10 +1765,10 @@ EXPORT_SYMBOL_GPL(i915_read_mch_val);
  */
 bool i915_gpu_raise(void)
 {
-	struct drm_i915_private *dev_priv;
+  	struct drm_i915_private *dev_priv;
 	bool ret = true;
 
-	spin_lock(&mchdev_lock);
+  	spin_lock(&mchdev_lock);
 	if (!i915_mch_dev) {
 		ret = false;
 		goto out_unlock;
@@ -1806,9 +1779,9 @@ bool i915_gpu_raise(void)
 		dev_priv->max_delay--;
 
 out_unlock:
-	spin_unlock(&mchdev_lock);
+  	spin_unlock(&mchdev_lock);
 
-	return ret;
+  	return ret;
 }
 EXPORT_SYMBOL_GPL(i915_gpu_raise);
 
@@ -1820,10 +1793,10 @@ EXPORT_SYMBOL_GPL(i915_gpu_raise);
  */
 bool i915_gpu_lower(void)
 {
-	struct drm_i915_private *dev_priv;
+  	struct drm_i915_private *dev_priv;
 	bool ret = true;
 
-	spin_lock(&mchdev_lock);
+  	spin_lock(&mchdev_lock);
 	if (!i915_mch_dev) {
 		ret = false;
 		goto out_unlock;
@@ -1834,9 +1807,9 @@ bool i915_gpu_lower(void)
 		dev_priv->max_delay++;
 
 out_unlock:
-	spin_unlock(&mchdev_lock);
+  	spin_unlock(&mchdev_lock);
 
-	return ret;
+  	return ret;
 }
 EXPORT_SYMBOL_GPL(i915_gpu_lower);
 
@@ -1847,10 +1820,10 @@ EXPORT_SYMBOL_GPL(i915_gpu_lower);
  */
 bool i915_gpu_busy(void)
 {
-	struct drm_i915_private *dev_priv;
+  	struct drm_i915_private *dev_priv;
 	bool ret = false;
 
-	spin_lock(&mchdev_lock);
+  	spin_lock(&mchdev_lock);
 	if (!i915_mch_dev)
 		goto out_unlock;
 	dev_priv = i915_mch_dev;
@@ -1858,9 +1831,9 @@ bool i915_gpu_busy(void)
 	ret = dev_priv->busy;
 
 out_unlock:
-	spin_unlock(&mchdev_lock);
+  	spin_unlock(&mchdev_lock);
 
-	return ret;
+  	return ret;
 }
 EXPORT_SYMBOL_GPL(i915_gpu_busy);
 
@@ -1872,10 +1845,10 @@ EXPORT_SYMBOL_GPL(i915_gpu_busy);
  */
 bool i915_gpu_turbo_disable(void)
 {
-	struct drm_i915_private *dev_priv;
+  	struct drm_i915_private *dev_priv;
 	bool ret = true;
 
-	spin_lock(&mchdev_lock);
+  	spin_lock(&mchdev_lock);
 	if (!i915_mch_dev) {
 		ret = false;
 		goto out_unlock;
@@ -1888,9 +1861,9 @@ bool i915_gpu_turbo_disable(void)
 		ret = false;
 
 out_unlock:
-	spin_unlock(&mchdev_lock);
+  	spin_unlock(&mchdev_lock);
 
-	return ret;
+  	return ret;
 }
 EXPORT_SYMBOL_GPL(i915_gpu_turbo_disable);
 
@@ -1951,8 +1924,6 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 		goto free_priv;
 	}
 
-	pci_set_master(dev->pdev);
-
 	/* overlay on gen2 is broken and can't address above 1G */
 	if (IS_GEN2(dev))
 		dma_set_coherent_mask(&dev->pdev->dev, DMA_BIT_MASK(30));
@@ -1985,7 +1956,7 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 
 	agp_size = dev_priv->mm.gtt->gtt_mappable_entries << PAGE_SHIFT;
 
-	dev_priv->mm.gtt_mapping =
+        dev_priv->mm.gtt_mapping =
 		io_mapping_create_wc(dev->agp->base, agp_size);
 	if (dev_priv->mm.gtt_mapping == NULL) {
 		ret = -EIO;
@@ -2068,14 +2039,11 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
 	if (!IS_I945G(dev) && !IS_I945GM(dev))
 		pci_enable_msi(dev->pdev);
 
-	spin_lock_init(&dev_priv->gt_lock);
 	spin_lock_init(&dev_priv->irq_lock);
 	spin_lock_init(&dev_priv->error_lock);
 	spin_lock_init(&dev_priv->rps_lock);
 
-	if (IS_IVYBRIDGE(dev))
-		dev_priv->num_pipe = 3;
-	else if (IS_MOBILE(dev) || !IS_GEN2(dev))
+	if (IS_MOBILE(dev) || !IS_GEN2(dev))
 		dev_priv->num_pipe = 2;
 	else
 		dev_priv->num_pipe = 1;
@@ -2152,7 +2120,7 @@ int i915_driver_unload(struct drm_device *dev)
 		unregister_shrinker(&dev_priv->mm.inactive_shrinker);
 
 	mutex_lock(&dev->struct_mutex);
-	ret = i915_gpu_idle(dev, true);
+	ret = i915_gpu_idle(dev);
 	if (ret)
 		DRM_ERROR("failed to idle hardware: %d\n", ret);
 	mutex_unlock(&dev->struct_mutex);
@@ -2205,7 +2173,6 @@ int i915_driver_unload(struct drm_device *dev)
 		i915_gem_free_all_phys_object(dev);
 		i915_gem_cleanup_ringbuffer(dev);
 		mutex_unlock(&dev->struct_mutex);
-		i915_gem_cleanup_aliasing_ppgtt(dev);
 		if (I915_HAS_FBC(dev) && i915_powersave)
 			i915_cleanup_compression(dev);
 		drm_mm_takedown(&dev_priv->mm.stolen);
@@ -2271,12 +2238,18 @@ void i915_driver_lastclose(struct drm_device * dev)
 
 	i915_gem_lastclose(dev);
 
+	if (dev_priv->agp_heap)
+		i915_mem_takedown(&(dev_priv->agp_heap));
+
 	i915_dma_cleanup(dev);
 }
 
 void i915_driver_preclose(struct drm_device * dev, struct drm_file *file_priv)
 {
+	drm_i915_private_t *dev_priv = dev->dev_private;
 	i915_gem_release(dev, file_priv);
+	if (!drm_core_check_feature(dev, DRIVER_MODESET))
+		i915_mem_release(dev, file_priv, dev_priv->agp_heap);
 }
 
 void i915_driver_postclose(struct drm_device *dev, struct drm_file *file)
@@ -2295,11 +2268,11 @@ struct drm_ioctl_desc i915_ioctls[] = {
 	DRM_IOCTL_DEF_DRV(I915_IRQ_WAIT, i915_irq_wait, DRM_AUTH),
 	DRM_IOCTL_DEF_DRV(I915_GETPARAM, i915_getparam, DRM_AUTH),
 	DRM_IOCTL_DEF_DRV(I915_SETPARAM, i915_setparam, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
-	DRM_IOCTL_DEF_DRV(I915_ALLOC, drm_noop, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_FREE, drm_noop, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_INIT_HEAP, drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_ALLOC, i915_mem_alloc, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_FREE, i915_mem_free, DRM_AUTH),
+	DRM_IOCTL_DEF_DRV(I915_INIT_HEAP, i915_mem_init_heap, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
 	DRM_IOCTL_DEF_DRV(I915_CMDBUFFER, i915_cmdbuffer, DRM_AUTH),
-	DRM_IOCTL_DEF_DRV(I915_DESTROY_HEAP,  drm_noop, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
+	DRM_IOCTL_DEF_DRV(I915_DESTROY_HEAP,  i915_mem_destroy_heap, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
 	DRM_IOCTL_DEF_DRV(I915_SET_VBLANK_PIPE,  i915_vblank_pipe_set, DRM_AUTH|DRM_MASTER|DRM_ROOT_ONLY),
 	DRM_IOCTL_DEF_DRV(I915_GET_VBLANK_PIPE,  i915_vblank_pipe_get, DRM_AUTH),
 	DRM_IOCTL_DEF_DRV(I915_VBLANK_SWAP, i915_vblank_swap, DRM_AUTH),
@@ -2327,8 +2300,6 @@ struct drm_ioctl_desc i915_ioctls[] = {
 	DRM_IOCTL_DEF_DRV(I915_GEM_MADVISE, i915_gem_madvise_ioctl, DRM_UNLOCKED),
 	DRM_IOCTL_DEF_DRV(I915_OVERLAY_PUT_IMAGE, intel_overlay_put_image, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
 	DRM_IOCTL_DEF_DRV(I915_OVERLAY_ATTRS, intel_overlay_attrs, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF_DRV(I915_SET_SPRITE_COLORKEY, intel_sprite_set_colorkey, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
-	DRM_IOCTL_DEF_DRV(I915_GET_SPRITE_COLORKEY, intel_sprite_get_colorkey, DRM_MASTER|DRM_CONTROL_ALLOW|DRM_UNLOCKED),
 };
 
 int i915_max_ioctl = DRM_ARRAY_SIZE(i915_ioctls);
diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index 2b8c1f2..111686a 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -35,84 +35,46 @@
 #include "intel_drv.h"
 
 #include <linux/console.h>
-#include <linux/module.h>
 #include "drm_crtc_helper.h"
 
-static int i915_modeset __read_mostly = -1;
+static int i915_modeset = -1;
 module_param_named(modeset, i915_modeset, int, 0400);
-MODULE_PARM_DESC(modeset,
-		"Use kernel modesetting [KMS] (0=DRM_I915_KMS from .config, "
-		"1=on, -1=force vga console preference [default])");
 
-unsigned int i915_fbpercrtc __always_unused = 0;
+unsigned int i915_fbpercrtc = 0;
 module_param_named(fbpercrtc, i915_fbpercrtc, int, 0400);
 
-int i915_panel_ignore_lid __read_mostly = 0;
+int i915_panel_ignore_lid = 0;
 module_param_named(panel_ignore_lid, i915_panel_ignore_lid, int, 0600);
-MODULE_PARM_DESC(panel_ignore_lid,
-		"Override lid status (0=autodetect [default], 1=lid open, "
-		"-1=lid closed)");
 
-unsigned int i915_powersave __read_mostly = 1;
+unsigned int i915_powersave = 1;
 module_param_named(powersave, i915_powersave, int, 0600);
-MODULE_PARM_DESC(powersave,
-		"Enable powersavings, fbc, downclocking, etc. (default: true)");
 
-int i915_semaphores __read_mostly = -1;
+unsigned int i915_semaphores = 0;
 module_param_named(semaphores, i915_semaphores, int, 0600);
-MODULE_PARM_DESC(semaphores,
-		"Use semaphores for inter-ring sync (default: -1 (use per-chip defaults))");
 
-int i915_enable_rc6 __read_mostly = -1;
+unsigned int i915_enable_rc6 = 0;
 module_param_named(i915_enable_rc6, i915_enable_rc6, int, 0600);
-MODULE_PARM_DESC(i915_enable_rc6,
-		"Enable power-saving render C-state 6 (default: -1 (use per-chip default)");
 
-int i915_enable_fbc __read_mostly = -1;
+unsigned int i915_enable_fbc = 0;
 module_param_named(i915_enable_fbc, i915_enable_fbc, int, 0600);
-MODULE_PARM_DESC(i915_enable_fbc,
-		"Enable frame buffer compression for power savings "
-		"(default: -1 (use per-chip default))");
 
-unsigned int i915_lvds_downclock __read_mostly = 0;
+unsigned int i915_lvds_downclock = 0;
 module_param_named(lvds_downclock, i915_lvds_downclock, int, 0400);
-MODULE_PARM_DESC(lvds_downclock,
-		"Use panel (LVDS/eDP) downclocking for power savings "
-		"(default: false)");
 
-int i915_panel_use_ssc __read_mostly = -1;
+unsigned int i915_panel_use_ssc = 1;
 module_param_named(lvds_use_ssc, i915_panel_use_ssc, int, 0600);
-MODULE_PARM_DESC(lvds_use_ssc,
-		"Use Spread Spectrum Clock with panels [LVDS/eDP] "
-		"(default: auto from VBT)");
 
-int i915_vbt_sdvo_panel_type __read_mostly = -1;
+int i915_vbt_sdvo_panel_type = -1;
 module_param_named(vbt_sdvo_panel_type, i915_vbt_sdvo_panel_type, int, 0600);
-MODULE_PARM_DESC(vbt_sdvo_panel_type,
-		"Override selection of SDVO panel mode in the VBT "
-		"(default: auto)");
 
-static bool i915_try_reset __read_mostly = true;
+static bool i915_try_reset = true;
 module_param_named(reset, i915_try_reset, bool, 0600);
-MODULE_PARM_DESC(reset, "Attempt GPU resets (default: true)");
-
-bool i915_enable_hangcheck __read_mostly = true;
-module_param_named(enable_hangcheck, i915_enable_hangcheck, bool, 0644);
-MODULE_PARM_DESC(enable_hangcheck,
-		"Periodically check GPU activity for detecting hangs. "
-		"WARNING: Disabling this can cause system wide hangs. "
-		"(default: true)");
-
-bool i915_enable_ppgtt __read_mostly = 1;
-module_param_named(i915_enable_ppgtt, i915_enable_ppgtt, bool, 0600);
-MODULE_PARM_DESC(i915_enable_ppgtt,
-		"Enable PPGTT (default: true)");
 
 static struct drm_driver driver;
 extern int intel_agp_enabled;
 
 #define INTEL_VGA_DEVICE(id, info) {		\
-	.class = PCI_BASE_CLASS_DISPLAY << 16,	\
+	.class = PCI_CLASS_DISPLAY_VGA << 8,	\
 	.class_mask = 0xff0000,			\
 	.vendor = 0x8086,			\
 	.device = id,				\
@@ -203,7 +165,7 @@ static const struct intel_device_info intel_pineview_info = {
 
 static const struct intel_device_info intel_ironlake_d_info = {
 	.gen = 5,
-	.need_gfx_hws = 1, .has_hotplug = 1,
+	.need_gfx_hws = 1, .has_pipe_cxsr = 1, .has_hotplug = 1,
 	.has_bsd_ring = 1,
 };
 
@@ -219,7 +181,6 @@ static const struct intel_device_info intel_sandybridge_d_info = {
 	.need_gfx_hws = 1, .has_hotplug = 1,
 	.has_bsd_ring = 1,
 	.has_blt_ring = 1,
-	.has_llc = 1,
 };
 
 static const struct intel_device_info intel_sandybridge_m_info = {
@@ -228,7 +189,6 @@ static const struct intel_device_info intel_sandybridge_m_info = {
 	.has_fbc = 1,
 	.has_bsd_ring = 1,
 	.has_blt_ring = 1,
-	.has_llc = 1,
 };
 
 static const struct intel_device_info intel_ivybridge_d_info = {
@@ -236,7 +196,6 @@ static const struct intel_device_info intel_ivybridge_d_info = {
 	.need_gfx_hws = 1, .has_hotplug = 1,
 	.has_bsd_ring = 1,
 	.has_blt_ring = 1,
-	.has_llc = 1,
 };
 
 static const struct intel_device_info intel_ivybridge_m_info = {
@@ -245,7 +204,6 @@ static const struct intel_device_info intel_ivybridge_m_info = {
 	.has_fbc = 0,	/* FBC is not enabled on Ivybridge mobile yet */
 	.has_bsd_ring = 1,
 	.has_blt_ring = 1,
-	.has_llc = 1,
 };
 
 static const struct pci_device_id pciidlist[] = {		/* aka */
@@ -304,7 +262,7 @@ MODULE_DEVICE_TABLE(pci, pciidlist);
 #define INTEL_PCH_CPT_DEVICE_ID_TYPE	0x1c00
 #define INTEL_PCH_PPT_DEVICE_ID_TYPE	0x1e00
 
-void intel_detect_pch(struct drm_device *dev)
+void intel_detect_pch (struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct pci_dev *pch;
@@ -337,7 +295,7 @@ void intel_detect_pch(struct drm_device *dev)
 	}
 }
 
-void __gen6_gt_force_wake_get(struct drm_i915_private *dev_priv)
+static void __gen6_gt_force_wake_get(struct drm_i915_private *dev_priv)
 {
 	int count;
 
@@ -353,22 +311,6 @@ void __gen6_gt_force_wake_get(struct drm_i915_private *dev_priv)
 		udelay(10);
 }
 
-void __gen6_gt_force_wake_mt_get(struct drm_i915_private *dev_priv)
-{
-	int count;
-
-	count = 0;
-	while (count++ < 50 && (I915_READ_NOTRACE(FORCEWAKE_MT_ACK) & 1))
-		udelay(10);
-
-	I915_WRITE_NOTRACE(FORCEWAKE_MT, (1<<16) | 1);
-	POSTING_READ(FORCEWAKE_MT);
-
-	count = 0;
-	while (count++ < 50 && (I915_READ_NOTRACE(FORCEWAKE_MT_ACK) & 1) == 0)
-		udelay(10);
-}
-
 /*
  * Generally this is called implicitly by the register read function. However,
  * if some sequence requires the GT to not power down then this function should
@@ -377,35 +319,17 @@ void __gen6_gt_force_wake_mt_get(struct drm_i915_private *dev_priv)
  */
 void gen6_gt_force_wake_get(struct drm_i915_private *dev_priv)
 {
-	unsigned long irqflags;
-
-	spin_lock_irqsave(&dev_priv->gt_lock, irqflags);
-	if (dev_priv->forcewake_count++ == 0)
-		dev_priv->display.force_wake_get(dev_priv);
-	spin_unlock_irqrestore(&dev_priv->gt_lock, irqflags);
-}
+	WARN_ON(!mutex_is_locked(&dev_priv->dev->struct_mutex));
 
-static void gen6_gt_check_fifodbg(struct drm_i915_private *dev_priv)
-{
-	u32 gtfifodbg;
-	gtfifodbg = I915_READ_NOTRACE(GTFIFODBG);
-	if (WARN(gtfifodbg & GT_FIFO_CPU_ERROR_MASK,
-	     "MMIO read or write has been dropped %x\n", gtfifodbg))
-		I915_WRITE_NOTRACE(GTFIFODBG, GT_FIFO_CPU_ERROR_MASK);
+	/* Forcewake is atomic in case we get in here without the lock */
+	if (atomic_add_return(1, &dev_priv->forcewake_count) == 1)
+		__gen6_gt_force_wake_get(dev_priv);
 }
 
-void __gen6_gt_force_wake_put(struct drm_i915_private *dev_priv)
+static void __gen6_gt_force_wake_put(struct drm_i915_private *dev_priv)
 {
 	I915_WRITE_NOTRACE(FORCEWAKE, 0);
-	/* The below doubles as a POSTING_READ */
-	gen6_gt_check_fifodbg(dev_priv);
-}
-
-void __gen6_gt_force_wake_mt_put(struct drm_i915_private *dev_priv)
-{
-	I915_WRITE_NOTRACE(FORCEWAKE_MT, (1<<16) | 0);
-	/* The below doubles as a POSTING_READ */
-	gen6_gt_check_fifodbg(dev_priv);
+	POSTING_READ(FORCEWAKE);
 }
 
 /*
@@ -413,32 +337,20 @@ void __gen6_gt_force_wake_mt_put(struct drm_i915_private *dev_priv)
  */
 void gen6_gt_force_wake_put(struct drm_i915_private *dev_priv)
 {
-	unsigned long irqflags;
+	WARN_ON(!mutex_is_locked(&dev_priv->dev->struct_mutex));
 
-	spin_lock_irqsave(&dev_priv->gt_lock, irqflags);
-	if (--dev_priv->forcewake_count == 0)
-		dev_priv->display.force_wake_put(dev_priv);
-	spin_unlock_irqrestore(&dev_priv->gt_lock, irqflags);
+	if (atomic_dec_and_test(&dev_priv->forcewake_count))
+		__gen6_gt_force_wake_put(dev_priv);
 }
 
-int __gen6_gt_wait_for_fifo(struct drm_i915_private *dev_priv)
+void __gen6_gt_wait_for_fifo(struct drm_i915_private *dev_priv)
 {
-	int ret = 0;
-
-	if (dev_priv->gt_fifo_count < GT_FIFO_NUM_RESERVED_ENTRIES) {
-		int loop = 500;
-		u32 fifo = I915_READ_NOTRACE(GT_FIFO_FREE_ENTRIES);
-		while (fifo <= GT_FIFO_NUM_RESERVED_ENTRIES && loop--) {
-			udelay(10);
-			fifo = I915_READ_NOTRACE(GT_FIFO_FREE_ENTRIES);
-		}
-		if (WARN_ON(loop < 0 && fifo <= GT_FIFO_NUM_RESERVED_ENTRIES))
-			++ret;
-		dev_priv->gt_fifo_count = fifo;
+	int loop = 500;
+	u32 fifo = I915_READ_NOTRACE(GT_FIFO_FREE_ENTRIES);
+	while (fifo < 20 && loop--) {
+		udelay(10);
+		fifo = I915_READ_NOTRACE(GT_FIFO_FREE_ENTRIES);
 	}
-	dev_priv->gt_fifo_count--;
-
-	return ret;
 }
 
 static int i915_drm_freeze(struct drm_device *dev)
@@ -523,12 +435,9 @@ static int i915_drm_thaw(struct drm_device *dev)
 		mutex_lock(&dev->struct_mutex);
 		dev_priv->mm.suspended = 0;
 
-		error = i915_gem_init_hw(dev);
+		error = i915_gem_init_ringbuffer(dev);
 		mutex_unlock(&dev->struct_mutex);
 
-		if (HAS_PCH_SPLIT(dev))
-			ironlake_init_pch_refclk(dev);
-
 		drm_mode_config_reset(dev);
 		drm_irq_install(dev);
 
@@ -634,40 +543,13 @@ static int ironlake_do_reset(struct drm_device *dev, u8 flags)
 static int gen6_do_reset(struct drm_device *dev, u8 flags)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	int	ret;
-	unsigned long irqflags;
-
-	/* Hold gt_lock across reset to prevent any register access
-	 * with forcewake not set correctly
-	 */
-	spin_lock_irqsave(&dev_priv->gt_lock, irqflags);
-
-	/* Reset the chip */
-
-	/* GEN6_GDRST is not in the gt power well, no need to check
-	 * for fifo space for the write or forcewake the chip for
-	 * the read
-	 */
-	I915_WRITE_NOTRACE(GEN6_GDRST, GEN6_GRDOM_FULL);
-
-	/* Spin waiting for the device to ack the reset request */
-	ret = wait_for((I915_READ_NOTRACE(GEN6_GDRST) & GEN6_GRDOM_FULL) == 0, 500);
-
-	/* If reset with a user forcewake, try to restore, otherwise turn it off */
-	if (dev_priv->forcewake_count)
-		dev_priv->display.force_wake_get(dev_priv);
-	else
-		dev_priv->display.force_wake_put(dev_priv);
 
-	/* Restore fifo count */
-	dev_priv->gt_fifo_count = I915_READ_NOTRACE(GT_FIFO_FREE_ENTRIES);
-
-	spin_unlock_irqrestore(&dev_priv->gt_lock, irqflags);
-	return ret;
+	I915_WRITE(GEN6_GDRST, GEN6_GRDOM_FULL);
+	return wait_for((I915_READ(GEN6_GDRST) & GEN6_GRDOM_FULL) == 0, 500);
 }
 
 /**
- * i915_reset - reset chip after a hang
+ * i965_reset - reset chip after a hang
  * @dev: drm device to reset
  * @flags: reset domains
  *
@@ -707,6 +589,9 @@ int i915_reset(struct drm_device *dev, u8 flags)
 	case 7:
 	case 6:
 		ret = gen6_do_reset(dev, flags);
+		/* If reset with a user forcewake, try to restore */
+		if (atomic_read(&dev_priv->forcewake_count))
+			__gen6_gt_force_wake_get(dev_priv);
 		break;
 	case 5:
 		ret = ironlake_do_reset(dev, flags);
@@ -743,16 +628,12 @@ int i915_reset(struct drm_device *dev, u8 flags)
 			!dev_priv->mm.suspended) {
 		dev_priv->mm.suspended = 0;
 
-		i915_gem_init_swizzling(dev);
-
 		dev_priv->ring[RCS].init(&dev_priv->ring[RCS]);
 		if (HAS_BSD(dev))
 		    dev_priv->ring[VCS].init(&dev_priv->ring[VCS]);
 		if (HAS_BLT(dev))
 		    dev_priv->ring[BCS].init(&dev_priv->ring[BCS]);
 
-		i915_gem_init_ppgtt(dev);
-
 		mutex_unlock(&dev->struct_mutex);
 		drm_irq_uninstall(dev);
 		drm_mode_config_reset(dev);
@@ -861,12 +742,12 @@ static int i915_pm_poweroff(struct device *dev)
 }
 
 static const struct dev_pm_ops i915_pm_ops = {
-	.suspend = i915_pm_suspend,
-	.resume = i915_pm_resume,
-	.freeze = i915_pm_freeze,
-	.thaw = i915_pm_thaw,
-	.poweroff = i915_pm_poweroff,
-	.restore = i915_pm_resume,
+     .suspend = i915_pm_suspend,
+     .resume = i915_pm_resume,
+     .freeze = i915_pm_freeze,
+     .thaw = i915_pm_thaw,
+     .poweroff = i915_pm_poweroff,
+     .restore = i915_pm_resume,
 };
 
 static struct vm_operations_struct i915_gem_vm_ops = {
@@ -875,24 +756,9 @@ static struct vm_operations_struct i915_gem_vm_ops = {
 	.close = drm_gem_vm_close,
 };
 
-static const struct file_operations i915_driver_fops = {
-	.owner = THIS_MODULE,
-	.open = drm_open,
-	.release = drm_release,
-	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_gem_mmap,
-	.poll = drm_poll,
-	.fasync = drm_fasync,
-	.read = drm_read,
-#ifdef CONFIG_COMPAT
-	.compat_ioctl = i915_compat_ioctl,
-#endif
-	.llseek = noop_llseek,
-};
-
 static struct drm_driver driver = {
-	/* Don't use MTRRs here; the Xserver or userspace app should
-	 * deal with them for Intel hardware.
+	/* don't use mtrr's here, the Xserver or user space app should
+	 * deal with them for intel hardware.
 	 */
 	.driver_features =
 	    DRIVER_USE_AGP | DRIVER_REQUIRE_AGP | /* DRIVER_USE_MTRR |*/
@@ -923,7 +789,21 @@ static struct drm_driver driver = {
 	.dumb_map_offset = i915_gem_mmap_gtt,
 	.dumb_destroy = i915_gem_dumb_destroy,
 	.ioctls = i915_ioctls,
-	.fops = &i915_driver_fops,
+	.fops = {
+		 .owner = THIS_MODULE,
+		 .open = drm_open,
+		 .release = drm_release,
+		 .unlocked_ioctl = drm_ioctl,
+		 .mmap = drm_gem_mmap,
+		 .poll = drm_poll,
+		 .fasync = drm_fasync,
+		 .read = drm_read,
+#ifdef CONFIG_COMPAT
+		 .compat_ioctl = i915_compat_ioctl,
+#endif
+		 .llseek = noop_llseek,
+	},
+
 	.name = DRIVER_NAME,
 	.desc = DRIVER_DESC,
 	.date = DRIVER_DATE,
@@ -987,46 +867,3 @@ module_exit(i915_exit);
 MODULE_AUTHOR(DRIVER_AUTHOR);
 MODULE_DESCRIPTION(DRIVER_DESC);
 MODULE_LICENSE("GPL and additional rights");
-
-#define __i915_read(x, y) \
-u##x i915_read##x(struct drm_i915_private *dev_priv, u32 reg) { \
-	u##x val = 0; \
-	if (NEEDS_FORCE_WAKE((dev_priv), (reg))) { \
-		unsigned long irqflags; \
-		spin_lock_irqsave(&dev_priv->gt_lock, irqflags); \
-		if (dev_priv->forcewake_count == 0) \
-			dev_priv->display.force_wake_get(dev_priv); \
-		val = read##y(dev_priv->regs + reg); \
-		if (dev_priv->forcewake_count == 0) \
-			dev_priv->display.force_wake_put(dev_priv); \
-		spin_unlock_irqrestore(&dev_priv->gt_lock, irqflags); \
-	} else { \
-		val = read##y(dev_priv->regs + reg); \
-	} \
-	trace_i915_reg_rw(false, reg, val, sizeof(val)); \
-	return val; \
-}
-
-__i915_read(8, b)
-__i915_read(16, w)
-__i915_read(32, l)
-__i915_read(64, q)
-#undef __i915_read
-
-#define __i915_write(x, y) \
-void i915_write##x(struct drm_i915_private *dev_priv, u32 reg, u##x val) { \
-	u32 __fifo_ret = 0; \
-	trace_i915_reg_rw(true, reg, val, sizeof(val)); \
-	if (NEEDS_FORCE_WAKE((dev_priv), (reg))) { \
-		__fifo_ret = __gen6_gt_wait_for_fifo(dev_priv); \
-	} \
-	write##y(val, dev_priv->regs + reg); \
-	if (unlikely(__fifo_ret)) { \
-		gen6_gt_check_fifodbg(dev_priv); \
-	} \
-}
-__i915_write(8, b)
-__i915_write(16, w)
-__i915_write(32, l)
-__i915_write(64, q)
-#undef __i915_write
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index c0f19f5..b570415 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -35,9 +35,7 @@
 #include "intel_ringbuffer.h"
 #include <linux/io-mapping.h>
 #include <linux/i2c.h>
-#include <linux/i2c-algo-bit.h>
 #include <drm/intel-gtt.h>
-#include <linux/backlight.h>
 
 /* General customization:
  */
@@ -108,7 +106,6 @@ struct opregion_header;
 struct opregion_acpi;
 struct opregion_swsci;
 struct opregion_asle;
-struct drm_i915_private;
 
 struct intel_opregion {
 	struct opregion_header *header;
@@ -128,15 +125,11 @@ struct drm_i915_master_private {
 	struct _drm_i915_sarea *sarea_priv;
 };
 #define I915_FENCE_REG_NONE -1
-#define I915_MAX_NUM_FENCES 16
-/* 16 fences + sign bit for FENCE_REG_NONE */
-#define I915_MAX_NUM_FENCE_BITS 5
 
 struct drm_i915_fence_reg {
 	struct list_head lru_list;
 	struct drm_i915_gem_object *obj;
 	uint32_t setup_seqno;
-	int pin_count;
 };
 
 struct sdvo_device_mapping {
@@ -145,6 +138,7 @@ struct sdvo_device_mapping {
 	u8 slave_addr;
 	u8 dvo_wiring;
 	u8 i2c_pin;
+	u8 i2c_speed;
 	u8 ddc_pin;
 };
 
@@ -154,40 +148,33 @@ struct drm_i915_error_state {
 	u32 eir;
 	u32 pgtbl_er;
 	u32 pipestat[I915_MAX_PIPES];
-	u32 tail[I915_NUM_RINGS];
-	u32 head[I915_NUM_RINGS];
-	u32 ipeir[I915_NUM_RINGS];
-	u32 ipehr[I915_NUM_RINGS];
-	u32 instdone[I915_NUM_RINGS];
-	u32 acthd[I915_NUM_RINGS];
-	u32 semaphore_mboxes[I915_NUM_RINGS][I915_NUM_RINGS - 1];
-	/* our own tracking of ring head and tail */
-	u32 cpu_ring_head[I915_NUM_RINGS];
-	u32 cpu_ring_tail[I915_NUM_RINGS];
+	u32 ipeir;
+	u32 ipehr;
+	u32 instdone;
+	u32 acthd;
 	u32 error; /* gen6+ */
-	u32 instpm[I915_NUM_RINGS];
-	u32 instps[I915_NUM_RINGS];
+	u32 bcs_acthd; /* gen6+ blt engine */
+	u32 bcs_ipehr;
+	u32 bcs_ipeir;
+	u32 bcs_instdone;
+	u32 bcs_seqno;
+	u32 vcs_acthd; /* gen6+ bsd engine */
+	u32 vcs_ipehr;
+	u32 vcs_ipeir;
+	u32 vcs_instdone;
+	u32 vcs_seqno;
+	u32 instpm;
+	u32 instps;
 	u32 instdone1;
-	u32 seqno[I915_NUM_RINGS];
+	u32 seqno;
 	u64 bbaddr;
-	u32 fault_reg[I915_NUM_RINGS];
-	u32 done_reg;
-	u32 faddr[I915_NUM_RINGS];
-	u64 fence[I915_MAX_NUM_FENCES];
+	u64 fence[16];
 	struct timeval time;
-	struct drm_i915_error_ring {
-		struct drm_i915_error_object {
-			int page_count;
-			u32 gtt_offset;
-			u32 *pages[0];
-		} *ringbuffer, *batchbuffer;
-		struct drm_i915_error_request {
-			long jiffies;
-			u32 seqno;
-			u32 tail;
-		} *requests;
-		int num_requests;
-	} ring[I915_NUM_RINGS];
+	struct drm_i915_error_object {
+		int page_count;
+		u32 gtt_offset;
+		u32 *pages[0];
+	} *ringbuffer[I915_NUM_RINGS], *batchbuffer[I915_NUM_RINGS];
 	struct drm_i915_error_buffer {
 		u32 size;
 		u32 name;
@@ -195,12 +182,12 @@ struct drm_i915_error_state {
 		u32 gtt_offset;
 		u32 read_domains;
 		u32 write_domain;
-		s32 fence_reg:I915_MAX_NUM_FENCE_BITS;
+		s32 fence_reg:5;
 		s32 pinned:2;
 		u32 tiling:2;
 		u32 dirty:1;
 		u32 purgeable:1;
-		s32 ring:4;
+		u32 ring:4;
 		u32 cache_level:2;
 	} *active_bo, *pinned_bo;
 	u32 active_bo_count, pinned_bo_count;
@@ -216,25 +203,17 @@ struct drm_i915_display_funcs {
 	int (*get_display_clock_speed)(struct drm_device *dev);
 	int (*get_fifo_size)(struct drm_device *dev, int plane);
 	void (*update_wm)(struct drm_device *dev);
-	void (*update_sprite_wm)(struct drm_device *dev, int pipe,
-				 uint32_t sprite_width, int pixel_size);
 	int (*crtc_mode_set)(struct drm_crtc *crtc,
 			     struct drm_display_mode *mode,
 			     struct drm_display_mode *adjusted_mode,
 			     int x, int y,
 			     struct drm_framebuffer *old_fb);
-	void (*write_eld)(struct drm_connector *connector,
-			  struct drm_crtc *crtc);
 	void (*fdi_link_train)(struct drm_crtc *crtc);
 	void (*init_clock_gating)(struct drm_device *dev);
 	void (*init_pch_clock_gating)(struct drm_device *dev);
 	int (*queue_flip)(struct drm_device *dev, struct drm_crtc *crtc,
 			  struct drm_framebuffer *fb,
 			  struct drm_i915_gem_object *obj);
-	int (*update_plane)(struct drm_crtc *crtc, struct drm_framebuffer *fb,
-			    int x, int y);
-	void (*force_wake_get)(struct drm_i915_private *dev_priv);
-	void (*force_wake_put)(struct drm_i915_private *dev_priv);
 	/* clock updates for mode set */
 	/* cursor updates */
 	/* render clock increase/decrease */
@@ -244,37 +223,26 @@ struct drm_i915_display_funcs {
 
 struct intel_device_info {
 	u8 gen;
-	u8 is_mobile:1;
-	u8 is_i85x:1;
-	u8 is_i915g:1;
-	u8 is_i945gm:1;
-	u8 is_g33:1;
-	u8 need_gfx_hws:1;
-	u8 is_g4x:1;
-	u8 is_pineview:1;
-	u8 is_broadwater:1;
-	u8 is_crestline:1;
-	u8 is_ivybridge:1;
-	u8 has_fbc:1;
-	u8 has_pipe_cxsr:1;
-	u8 has_hotplug:1;
-	u8 cursor_needs_physical:1;
-	u8 has_overlay:1;
-	u8 overlay_needs_physical:1;
-	u8 supports_tv:1;
-	u8 has_bsd_ring:1;
-	u8 has_blt_ring:1;
-	u8 has_llc:1;
-};
-
-#define I915_PPGTT_PD_ENTRIES 512
-#define I915_PPGTT_PT_ENTRIES 1024
-struct i915_hw_ppgtt {
-	unsigned num_pd_entries;
-	struct page **pt_pages;
-	uint32_t pd_offset;
-	dma_addr_t *pt_dma_addr;
-	dma_addr_t scratch_page_dma_addr;
+	u8 is_mobile : 1;
+	u8 is_i85x : 1;
+	u8 is_i915g : 1;
+	u8 is_i945gm : 1;
+	u8 is_g33 : 1;
+	u8 need_gfx_hws : 1;
+	u8 is_g4x : 1;
+	u8 is_pineview : 1;
+	u8 is_broadwater : 1;
+	u8 is_crestline : 1;
+	u8 is_ivybridge : 1;
+	u8 has_fbc : 1;
+	u8 has_pipe_cxsr : 1;
+	u8 has_hotplug : 1;
+	u8 cursor_needs_physical : 1;
+	u8 has_overlay : 1;
+	u8 overlay_needs_physical : 1;
+	u8 supports_tv : 1;
+	u8 has_bsd_ring : 1;
+	u8 has_blt_ring : 1;
 };
 
 enum no_fbc_reason {
@@ -297,17 +265,6 @@ enum intel_pch {
 #define QUIRK_LVDS_SSC_DISABLE (1<<1)
 
 struct intel_fbdev;
-struct intel_fbc_work;
-
-struct intel_gmbus {
-	struct i2c_adapter adapter;
-	bool force_bit;
-	bool has_gpio;
-	u32 reg0;
-	u32 gpio_reg;
-	struct i2c_algo_bit_data bit_algo;
-	struct drm_i915_private *dev_priv;
-};
 
 typedef struct drm_i915_private {
 	struct drm_device *dev;
@@ -318,19 +275,12 @@ typedef struct drm_i915_private {
 	int relative_constants_mode;
 
 	void __iomem *regs;
-	/** gt_fifo_count and the subsequent register write are synchronized
-	 * with dev->struct_mutex. */
-	unsigned gt_fifo_count;
-	/** forcewake_count is protected by gt_lock */
-	unsigned forcewake_count;
-	/** gt_lock is also taken in irq contexts. */
-	struct spinlock gt_lock;
-
-	struct intel_gmbus *gmbus;
 
-	/** gmbus_mutex protects against concurrent usage of the single hw gmbus
-	 * controller on different i2c buses. */
-	struct mutex gmbus_mutex;
+	struct intel_gmbus {
+		struct i2c_adapter adapter;
+		struct i2c_adapter *force_bit;
+		u32 reg0;
+	} *gmbus;
 
 	struct pci_dev *bridge_dev;
 	struct intel_ring_buffer ring[I915_NUM_RINGS];
@@ -365,6 +315,7 @@ typedef struct drm_i915_private {
 
 	int tex_lru_log_granularity;
 	int allow_batchbuffer;
+	struct mem_block *agp_heap;
 	unsigned int sr01, adpa, ppcr, dvob, dvoc, lvds;
 	int vblank_pipe;
 	int num_pipe;
@@ -380,20 +331,21 @@ typedef struct drm_i915_private {
 	uint32_t last_instdone1;
 
 	unsigned long cfb_size;
-	unsigned int cfb_fb;
-	enum plane cfb_plane;
+	unsigned long cfb_pitch;
+	unsigned long cfb_offset;
+	int cfb_fence;
+	int cfb_plane;
 	int cfb_y;
-	struct intel_fbc_work *fbc_work;
 
 	struct intel_opregion opregion;
 
 	/* overlay */
 	struct intel_overlay *overlay;
-	bool sprite_scaling_enabled;
 
 	/* LVDS info */
 	int backlight_level;  /* restore backlight to this value */
 	bool backlight_enabled;
+	struct drm_display_mode *panel_fixed_mode;
 	struct drm_display_mode *lfp_lvds_vbt_mode; /* if any */
 	struct drm_display_mode *sdvo_lvds_vbt_mode; /* if any */
 
@@ -403,7 +355,6 @@ typedef struct drm_i915_private {
 	unsigned int lvds_vbt:1;
 	unsigned int int_crt_support:1;
 	unsigned int lvds_use_ssc:1;
-	unsigned int display_clock_mode:1;
 	int lvds_ssc_freq;
 	struct {
 		int rate;
@@ -421,7 +372,7 @@ typedef struct drm_i915_private {
 	struct notifier_block lid_notifier;
 
 	int crt_ddc_pin;
-	struct drm_i915_fence_reg fence_regs[I915_MAX_NUM_FENCES]; /* assume 965 */
+	struct drm_i915_fence_reg fence_regs[16]; /* assume 965 */
 	int fence_reg_start; /* 4 if userland hasn't ioctl'd us yet */
 	int num_fence_regs; /* 8 on pre-965, 16 otherwise */
 
@@ -552,7 +503,7 @@ typedef struct drm_i915_private {
 	u8 saveAR[21];
 	u8 saveDACMASK;
 	u8 saveCR[37];
-	uint64_t saveFENCE[I915_MAX_NUM_FENCES];
+	uint64_t saveFENCE[16];
 	u32 saveCURACNTR;
 	u32 saveCURAPOS;
 	u32 saveCURABASE;
@@ -613,9 +564,6 @@ typedef struct drm_i915_private {
 		struct io_mapping *gtt_mapping;
 		int gtt_mtrr;
 
-		/** PPGTT used for aliasing the PPGTT with the GTT */
-		struct i915_hw_ppgtt *aliasing_ppgtt;
-
 		struct shrinker inactive_shrinker;
 
 		/**
@@ -722,9 +670,10 @@ typedef struct drm_i915_private {
 	unsigned int lvds_border_bits;
 	/* Panel fitter placement and size for Ironlake+ */
 	u32 pch_pf_pos, pch_pf_size;
+	int panel_t3, panel_t12;
 
-	struct drm_crtc *plane_to_crtc_mapping[3];
-	struct drm_crtc *pipe_to_crtc_mapping[3];
+	struct drm_crtc *plane_to_crtc_mapping[2];
+	struct drm_crtc *pipe_to_crtc_mapping[2];
 	wait_queue_head_t pending_flip_queue;
 	bool flip_pending_is_done;
 
@@ -740,7 +689,6 @@ typedef struct drm_i915_private {
 	int child_dev_num;
 	struct child_device_config *child_dev;
 	struct drm_connector *int_lvds_connector;
-	struct drm_connector *int_edp_connector;
 
 	bool mchbar_need_disable;
 
@@ -775,18 +723,11 @@ typedef struct drm_i915_private {
 	/* list of fbdev register on this device */
 	struct intel_fbdev *fbdev;
 
-	struct backlight_device *backlight;
-
 	struct drm_property *broadcast_rgb_property;
 	struct drm_property *force_audio_property;
-} drm_i915_private_t;
 
-enum hdmi_force_audio {
-	HDMI_AUDIO_OFF_DVI = -2,	/* no aux data for HDMI-DVI converter */
-	HDMI_AUDIO_OFF,			/* force turn off HDMI audio */
-	HDMI_AUDIO_AUTO,		/* trust EDID */
-	HDMI_AUDIO_ON,			/* force turn on HDMI audio */
-};
+	atomic_t forcewake_count;
+} drm_i915_private_t;
 
 enum i915_cache_level {
 	I915_CACHE_NONE,
@@ -814,37 +755,39 @@ struct drm_i915_gem_object {
 	 * (has pending rendering), and is not set if it's on inactive (ready
 	 * to be unbound).
 	 */
-	unsigned int active:1;
+	unsigned int active : 1;
 
 	/**
 	 * This is set if the object has been written to since last bound
 	 * to the GTT
 	 */
-	unsigned int dirty:1;
+	unsigned int dirty : 1;
 
 	/**
 	 * This is set if the object has been written to since the last
 	 * GPU flush.
 	 */
-	unsigned int pending_gpu_write:1;
+	unsigned int pending_gpu_write : 1;
 
 	/**
 	 * Fence register bits (if any) for this object.  Will be set
 	 * as needed when mapped into the GTT.
 	 * Protected by dev->struct_mutex.
+	 *
+	 * Size: 4 bits for 16 fences + sign (for FENCE_REG_NONE)
 	 */
-	signed int fence_reg:I915_MAX_NUM_FENCE_BITS;
+	signed int fence_reg : 5;
 
 	/**
 	 * Advice: are the backing pages purgeable?
 	 */
-	unsigned int madv:2;
+	unsigned int madv : 2;
 
 	/**
 	 * Current tiling mode for the object.
 	 */
-	unsigned int tiling_mode:2;
-	unsigned int tiling_changed:1;
+	unsigned int tiling_mode : 2;
+	unsigned int tiling_changed : 1;
 
 	/** How many users have pinned this object in GTT space. The following
 	 * users can each hold at most one reference: pwrite/pread, pin_ioctl
@@ -855,22 +798,22 @@ struct drm_i915_gem_object {
 	 *
 	 * In the worst case this is 1 + 1 + 1 + 2*2 = 7. That would fit into 3
 	 * bits with absolutely no headroom. So use 4 bits. */
-	unsigned int pin_count:4;
+	unsigned int pin_count : 4;
 #define DRM_I915_GEM_OBJECT_MAX_PIN_COUNT 0xf
 
 	/**
 	 * Is the object at the current location in the gtt mappable and
 	 * fenceable? Used to avoid costly recalculations.
 	 */
-	unsigned int map_and_fenceable:1;
+	unsigned int map_and_fenceable : 1;
 
 	/**
 	 * Whether the current gtt mapping needs to be mappable (and isn't just
 	 * mappable by accident). Track pin and fault separate for a more
 	 * accurate mappable working set.
 	 */
-	unsigned int fault_mappable:1;
-	unsigned int pin_mappable:1;
+	unsigned int fault_mappable : 1;
+	unsigned int pin_mappable : 1;
 
 	/*
 	 * Is the GPU currently using a fence to access this buffer,
@@ -880,8 +823,6 @@ struct drm_i915_gem_object {
 
 	unsigned int cache_level:2;
 
-	unsigned int has_aliasing_ppgtt_mapping:1;
-
 	struct page **pages;
 
 	/**
@@ -959,9 +900,6 @@ struct drm_i915_gem_request {
 	/** GEM sequence number associated with this request. */
 	uint32_t seqno;
 
-	/** Postion in the ringbuffer of the end of the request */
-	u32 tail;
-
 	/** Time at which this request was emitted, in jiffies. */
 	unsigned long emitted_jiffies;
 
@@ -1018,11 +956,8 @@ struct drm_i915_file_private {
 
 #define HAS_BSD(dev)            (INTEL_INFO(dev)->has_bsd_ring)
 #define HAS_BLT(dev)            (INTEL_INFO(dev)->has_blt_ring)
-#define HAS_LLC(dev)            (INTEL_INFO(dev)->has_llc)
 #define I915_NEED_GFX_HWS(dev)	(INTEL_INFO(dev)->need_gfx_hws)
 
-#define HAS_ALIASING_PPGTT(dev)	(INTEL_INFO(dev)->gen >=6)
-
 #define HAS_OVERLAY(dev)		(INTEL_INFO(dev)->has_overlay)
 #define OVERLAY_NEEDS_PHYSICAL(dev)	(INTEL_INFO(dev)->overlay_needs_physical)
 
@@ -1055,17 +990,15 @@ struct drm_i915_file_private {
 
 extern struct drm_ioctl_desc i915_ioctls[];
 extern int i915_max_ioctl;
-extern unsigned int i915_fbpercrtc __always_unused;
-extern int i915_panel_ignore_lid __read_mostly;
-extern unsigned int i915_powersave __read_mostly;
-extern int i915_semaphores __read_mostly;
-extern unsigned int i915_lvds_downclock __read_mostly;
-extern int i915_panel_use_ssc __read_mostly;
-extern int i915_vbt_sdvo_panel_type __read_mostly;
-extern int i915_enable_rc6 __read_mostly;
-extern int i915_enable_fbc __read_mostly;
-extern bool i915_enable_hangcheck __read_mostly;
-extern bool i915_enable_ppgtt __read_mostly;
+extern unsigned int i915_fbpercrtc;
+extern int i915_panel_ignore_lid;
+extern unsigned int i915_powersave;
+extern unsigned int i915_semaphores;
+extern unsigned int i915_lvds_downclock;
+extern unsigned int i915_panel_use_ssc;
+extern int i915_vbt_sdvo_panel_type;
+extern unsigned int i915_enable_rc6;
+extern unsigned int i915_enable_fbc;
 
 extern int i915_suspend(struct drm_device *dev, pm_message_t state);
 extern int i915_resume(struct drm_device *dev);
@@ -1118,7 +1051,7 @@ i915_enable_pipestat(drm_i915_private_t *dev_priv, int pipe, u32 mask);
 void
 i915_disable_pipestat(drm_i915_private_t *dev_priv, int pipe, u32 mask);
 
-void intel_enable_asle(struct drm_device *dev);
+void intel_enable_asle (struct drm_device *dev);
 
 #ifdef CONFIG_DEBUG_FS
 extern void i915_destroy_error_state(struct drm_device *dev);
@@ -1127,6 +1060,18 @@ extern void i915_destroy_error_state(struct drm_device *dev);
 #endif
 
 
+/* i915_mem.c */
+extern int i915_mem_alloc(struct drm_device *dev, void *data,
+			  struct drm_file *file_priv);
+extern int i915_mem_free(struct drm_device *dev, void *data,
+			 struct drm_file *file_priv);
+extern int i915_mem_init_heap(struct drm_device *dev, void *data,
+			      struct drm_file *file_priv);
+extern int i915_mem_destroy_heap(struct drm_device *dev, void *data,
+				 struct drm_file *file_priv);
+extern void i915_mem_takedown(struct mem_block **heap);
+extern void i915_mem_release(struct drm_device * dev,
+			     struct drm_file *file_priv, struct mem_block *heap);
 /* i915_gem.c */
 int i915_gem_init_ioctl(struct drm_device *dev, void *data,
 			struct drm_file *file_priv);
@@ -1196,7 +1141,7 @@ int i915_gem_dumb_create(struct drm_file *file_priv,
 int i915_gem_mmap_gtt(struct drm_file *file_priv, struct drm_device *dev,
 		      uint32_t handle, uint64_t *offset);
 int i915_gem_dumb_destroy(struct drm_file *file_priv, struct drm_device *dev,
-			  uint32_t handle);
+			  uint32_t handle);			  
 /**
  * Returns true if seq1 is later than seq2.
  */
@@ -1206,62 +1151,43 @@ i915_seqno_passed(uint32_t seq1, uint32_t seq2)
 	return (int32_t)(seq1 - seq2) >= 0;
 }
 
-u32 i915_gem_next_request_seqno(struct intel_ring_buffer *ring);
+static inline u32
+i915_gem_next_request_seqno(struct intel_ring_buffer *ring)
+{
+	drm_i915_private_t *dev_priv = ring->dev->dev_private;
+	return ring->outstanding_lazy_request = dev_priv->next_seqno;
+}
 
 int __must_check i915_gem_object_get_fence(struct drm_i915_gem_object *obj,
 					   struct intel_ring_buffer *pipelined);
 int __must_check i915_gem_object_put_fence(struct drm_i915_gem_object *obj);
 
-static inline void
-i915_gem_object_pin_fence(struct drm_i915_gem_object *obj)
-{
-	if (obj->fence_reg != I915_FENCE_REG_NONE) {
-		struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
-		dev_priv->fence_regs[obj->fence_reg].pin_count++;
-	}
-}
-
-static inline void
-i915_gem_object_unpin_fence(struct drm_i915_gem_object *obj)
-{
-	if (obj->fence_reg != I915_FENCE_REG_NONE) {
-		struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
-		dev_priv->fence_regs[obj->fence_reg].pin_count--;
-	}
-}
-
 void i915_gem_retire_requests(struct drm_device *dev);
-void i915_gem_retire_requests_ring(struct intel_ring_buffer *ring);
-
 void i915_gem_reset(struct drm_device *dev);
 void i915_gem_clflush_object(struct drm_i915_gem_object *obj);
 int __must_check i915_gem_object_set_domain(struct drm_i915_gem_object *obj,
 					    uint32_t read_domains,
 					    uint32_t write_domain);
-int __must_check i915_gem_object_finish_gpu(struct drm_i915_gem_object *obj);
-int __must_check i915_gem_init_hw(struct drm_device *dev);
-void i915_gem_init_swizzling(struct drm_device *dev);
-void i915_gem_init_ppgtt(struct drm_device *dev);
+int __must_check i915_gem_object_flush_gpu(struct drm_i915_gem_object *obj);
+int __must_check i915_gem_init_ringbuffer(struct drm_device *dev);
 void i915_gem_cleanup_ringbuffer(struct drm_device *dev);
 void i915_gem_do_init(struct drm_device *dev,
 		      unsigned long start,
 		      unsigned long mappable_end,
 		      unsigned long end);
-int __must_check i915_gpu_idle(struct drm_device *dev, bool do_retire);
+int __must_check i915_gpu_idle(struct drm_device *dev);
 int __must_check i915_gem_idle(struct drm_device *dev);
 int __must_check i915_add_request(struct intel_ring_buffer *ring,
 				  struct drm_file *file,
 				  struct drm_i915_gem_request *request);
 int __must_check i915_wait_request(struct intel_ring_buffer *ring,
-				   uint32_t seqno,
-				   bool do_retire);
+				   uint32_t seqno);
 int i915_gem_fault(struct vm_area_struct *vma, struct vm_fault *vmf);
 int __must_check
 i915_gem_object_set_to_gtt_domain(struct drm_i915_gem_object *obj,
 				  bool write);
 int __must_check
-i915_gem_object_pin_to_display_plane(struct drm_i915_gem_object *obj,
-				     u32 alignment,
+i915_gem_object_set_to_display_plane(struct drm_i915_gem_object *obj,
 				     struct intel_ring_buffer *pipelined);
 int i915_gem_attach_phys_object(struct drm_device *dev,
 				struct drm_i915_gem_object *obj,
@@ -1277,22 +1203,9 @@ i915_gem_get_unfenced_gtt_alignment(struct drm_device *dev,
 				    uint32_t size,
 				    int tiling_mode);
 
-int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj,
-				    enum i915_cache_level cache_level);
-
 /* i915_gem_gtt.c */
-int __must_check i915_gem_init_aliasing_ppgtt(struct drm_device *dev);
-void i915_gem_cleanup_aliasing_ppgtt(struct drm_device *dev);
-void i915_ppgtt_bind_object(struct i915_hw_ppgtt *ppgtt,
-			    struct drm_i915_gem_object *obj,
-			    enum i915_cache_level cache_level);
-void i915_ppgtt_unbind_object(struct i915_hw_ppgtt *ppgtt,
-			      struct drm_i915_gem_object *obj);
-
 void i915_gem_restore_gtt_mappings(struct drm_device *dev);
 int __must_check i915_gem_gtt_bind_object(struct drm_i915_gem_object *obj);
-void i915_gem_gtt_rebind_object(struct drm_i915_gem_object *obj,
-				enum i915_cache_level cache_level);
 void i915_gem_gtt_unbind_object(struct drm_i915_gem_object *obj);
 
 /* i915_gem_evict.c */
@@ -1374,19 +1287,17 @@ extern void intel_modeset_init(struct drm_device *dev);
 extern void intel_modeset_gem_init(struct drm_device *dev);
 extern void intel_modeset_cleanup(struct drm_device *dev);
 extern int intel_modeset_vga_set_state(struct drm_device *dev, bool state);
-extern bool intel_fbc_enabled(struct drm_device *dev);
+extern void i8xx_disable_fbc(struct drm_device *dev);
+extern void g4x_disable_fbc(struct drm_device *dev);
+extern void ironlake_disable_fbc(struct drm_device *dev);
 extern void intel_disable_fbc(struct drm_device *dev);
+extern void intel_enable_fbc(struct drm_crtc *crtc, unsigned long interval);
+extern bool intel_fbc_enabled(struct drm_device *dev);
 extern bool ironlake_set_drps(struct drm_device *dev, u8 val);
-extern void ironlake_init_pch_refclk(struct drm_device *dev);
 extern void ironlake_enable_rc6(struct drm_device *dev);
 extern void gen6_set_rps(struct drm_device *dev, u8 val);
-extern void intel_detect_pch(struct drm_device *dev);
-extern int intel_trans_dp_port_sel(struct drm_crtc *crtc);
-
-extern void __gen6_gt_force_wake_get(struct drm_i915_private *dev_priv);
-extern void __gen6_gt_force_wake_mt_get(struct drm_i915_private *dev_priv);
-extern void __gen6_gt_force_wake_put(struct drm_i915_private *dev_priv);
-extern void __gen6_gt_force_wake_mt_put(struct drm_i915_private *dev_priv);
+extern void intel_detect_pch (struct drm_device *dev);
+extern int intel_trans_dp_port_sel (struct drm_crtc *crtc);
 
 /* overlay */
 #ifdef CONFIG_DEBUG_FS
@@ -1427,16 +1338,27 @@ extern void intel_display_print_error_state(struct seq_file *m,
  */
 void gen6_gt_force_wake_get(struct drm_i915_private *dev_priv);
 void gen6_gt_force_wake_put(struct drm_i915_private *dev_priv);
-int __gen6_gt_wait_for_fifo(struct drm_i915_private *dev_priv);
+void __gen6_gt_wait_for_fifo(struct drm_i915_private *dev_priv);
 
 /* We give fast paths for the really cool registers */
 #define NEEDS_FORCE_WAKE(dev_priv, reg) \
 	(((dev_priv)->info->gen >= 6) && \
-	 ((reg) < 0x40000) &&		 \
-	 ((reg) != FORCEWAKE))
+	((reg) < 0x40000) && \
+	((reg) != FORCEWAKE))
 
 #define __i915_read(x, y) \
-	u##x i915_read##x(struct drm_i915_private *dev_priv, u32 reg);
+static inline u##x i915_read##x(struct drm_i915_private *dev_priv, u32 reg) { \
+	u##x val = 0; \
+	if (NEEDS_FORCE_WAKE((dev_priv), (reg))) { \
+		gen6_gt_force_wake_get(dev_priv); \
+		val = read##y(dev_priv->regs + reg); \
+		gen6_gt_force_wake_put(dev_priv); \
+	} else { \
+		val = read##y(dev_priv->regs + reg); \
+	} \
+	trace_i915_reg_rw(false, reg, val, sizeof(val)); \
+	return val; \
+}
 
 __i915_read(8, b)
 __i915_read(16, w)
@@ -1445,8 +1367,13 @@ __i915_read(64, q)
 #undef __i915_read
 
 #define __i915_write(x, y) \
-	void i915_write##x(struct drm_i915_private *dev_priv, u32 reg, u##x val);
-
+static inline void i915_write##x(struct drm_i915_private *dev_priv, u32 reg, u##x val) { \
+	trace_i915_reg_rw(true, reg, val, sizeof(val)); \
+	if (NEEDS_FORCE_WAKE((dev_priv), (reg))) { \
+		__gen6_gt_wait_for_fifo(dev_priv); \
+	} \
+	write##y(val, dev_priv->regs + reg); \
+}
 __i915_write(8, b)
 __i915_write(16, w)
 __i915_write(32, l)
diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index 723d1ae..46e04a1 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -58,7 +58,6 @@ static void i915_gem_free_object_tail(struct drm_i915_gem_object *obj);
 
 static int i915_gem_inactive_shrink(struct shrinker *shrinker,
 				    struct shrink_control *sc);
-static void i915_gem_object_truncate(struct drm_i915_gem_object *obj);
 
 /* some bookkeeping */
 static void i915_gem_info_add_obj(struct drm_i915_private *dev_priv,
@@ -180,7 +179,7 @@ i915_gem_get_aperture_ioctl(struct drm_device *dev, void *data,
 	mutex_unlock(&dev->struct_mutex);
 
 	args->aper_size = dev_priv->mm.gtt_total;
-	args->aper_available_size = args->aper_size - pinned;
+	args->aper_available_size = args->aper_size -pinned;
 
 	return 0;
 }
@@ -196,8 +195,6 @@ i915_gem_create(struct drm_file *file,
 	u32 handle;
 
 	size = roundup(size, PAGE_SIZE);
-	if (size == 0)
-		return -EINVAL;
 
 	/* Allocate the new object */
 	obj = i915_gem_alloc_object(dev, size);
@@ -259,6 +256,73 @@ static int i915_gem_object_needs_bit17_swizzle(struct drm_i915_gem_object *obj)
 		obj->tiling_mode != I915_TILING_NONE;
 }
 
+static inline void
+slow_shmem_copy(struct page *dst_page,
+		int dst_offset,
+		struct page *src_page,
+		int src_offset,
+		int length)
+{
+	char *dst_vaddr, *src_vaddr;
+
+	dst_vaddr = kmap(dst_page);
+	src_vaddr = kmap(src_page);
+
+	memcpy(dst_vaddr + dst_offset, src_vaddr + src_offset, length);
+
+	kunmap(src_page);
+	kunmap(dst_page);
+}
+
+static inline void
+slow_shmem_bit17_copy(struct page *gpu_page,
+		      int gpu_offset,
+		      struct page *cpu_page,
+		      int cpu_offset,
+		      int length,
+		      int is_read)
+{
+	char *gpu_vaddr, *cpu_vaddr;
+
+	/* Use the unswizzled path if this page isn't affected. */
+	if ((page_to_phys(gpu_page) & (1 << 17)) == 0) {
+		if (is_read)
+			return slow_shmem_copy(cpu_page, cpu_offset,
+					       gpu_page, gpu_offset, length);
+		else
+			return slow_shmem_copy(gpu_page, gpu_offset,
+					       cpu_page, cpu_offset, length);
+	}
+
+	gpu_vaddr = kmap(gpu_page);
+	cpu_vaddr = kmap(cpu_page);
+
+	/* Copy the data, XORing A6 with A17 (1). The user already knows he's
+	 * XORing with the other bits (A9 for Y, A9 and A10 for X)
+	 */
+	while (length > 0) {
+		int cacheline_end = ALIGN(gpu_offset + 1, 64);
+		int this_length = min(cacheline_end - gpu_offset, length);
+		int swizzled_gpu_offset = gpu_offset ^ 64;
+
+		if (is_read) {
+			memcpy(cpu_vaddr + cpu_offset,
+			       gpu_vaddr + swizzled_gpu_offset,
+			       this_length);
+		} else {
+			memcpy(gpu_vaddr + swizzled_gpu_offset,
+			       cpu_vaddr + cpu_offset,
+			       this_length);
+		}
+		cpu_offset += this_length;
+		gpu_offset += this_length;
+		length -= this_length;
+	}
+
+	kunmap(cpu_page);
+	kunmap(gpu_page);
+}
+
 /**
  * This is the fast shmem pread path, which attempts to copy_from_user directly
  * from the backing pages of the object to the user's address space.  On a
@@ -319,58 +383,6 @@ i915_gem_shmem_pread_fast(struct drm_device *dev,
 	return 0;
 }
 
-static inline int
-__copy_to_user_swizzled(char __user *cpu_vaddr,
-			const char *gpu_vaddr, int gpu_offset,
-			int length)
-{
-	int ret, cpu_offset = 0;
-
-	while (length > 0) {
-		int cacheline_end = ALIGN(gpu_offset + 1, 64);
-		int this_length = min(cacheline_end - gpu_offset, length);
-		int swizzled_gpu_offset = gpu_offset ^ 64;
-
-		ret = __copy_to_user(cpu_vaddr + cpu_offset,
-				     gpu_vaddr + swizzled_gpu_offset,
-				     this_length);
-		if (ret)
-			return ret + length;
-
-		cpu_offset += this_length;
-		gpu_offset += this_length;
-		length -= this_length;
-	}
-
-	return 0;
-}
-
-static inline int
-__copy_from_user_swizzled(char __user *gpu_vaddr, int gpu_offset,
-			  const char *cpu_vaddr,
-			  int length)
-{
-	int ret, cpu_offset = 0;
-
-	while (length > 0) {
-		int cacheline_end = ALIGN(gpu_offset + 1, 64);
-		int this_length = min(cacheline_end - gpu_offset, length);
-		int swizzled_gpu_offset = gpu_offset ^ 64;
-
-		ret = __copy_from_user(gpu_vaddr + swizzled_gpu_offset,
-				       cpu_vaddr + cpu_offset,
-				       this_length);
-		if (ret)
-			return ret + length;
-
-		cpu_offset += this_length;
-		gpu_offset += this_length;
-		length -= this_length;
-	}
-
-	return 0;
-}
-
 /**
  * This is the fallback shmem pread path, which allocates temporary storage
  * in kernel space to copy_to_user into outside of the struct_mutex, so we
@@ -384,34 +396,72 @@ i915_gem_shmem_pread_slow(struct drm_device *dev,
 			  struct drm_file *file)
 {
 	struct address_space *mapping = obj->base.filp->f_path.dentry->d_inode->i_mapping;
-	char __user *user_data;
+	struct mm_struct *mm = current->mm;
+	struct page **user_pages;
 	ssize_t remain;
-	loff_t offset;
-	int shmem_page_offset, page_length, ret;
-	int obj_do_bit17_swizzling, page_do_bit17_swizzling;
+	loff_t offset, pinned_pages, i;
+	loff_t first_data_page, last_data_page, num_pages;
+	int shmem_page_offset;
+	int data_page_index, data_page_offset;
+	int page_length;
+	int ret;
+	uint64_t data_ptr = args->data_ptr;
+	int do_bit17_swizzling;
 
-	user_data = (char __user *) (uintptr_t) args->data_ptr;
 	remain = args->size;
 
-	obj_do_bit17_swizzling = i915_gem_object_needs_bit17_swizzle(obj);
+	/* Pin the user pages containing the data.  We can't fault while
+	 * holding the struct mutex, yet we want to hold it while
+	 * dereferencing the user data.
+	 */
+	first_data_page = data_ptr / PAGE_SIZE;
+	last_data_page = (data_ptr + args->size - 1) / PAGE_SIZE;
+	num_pages = last_data_page - first_data_page + 1;
 
-	offset = args->offset;
+	user_pages = drm_malloc_ab(num_pages, sizeof(struct page *));
+	if (user_pages == NULL)
+		return -ENOMEM;
 
 	mutex_unlock(&dev->struct_mutex);
+	down_read(&mm->mmap_sem);
+	pinned_pages = get_user_pages(current, mm, (uintptr_t)args->data_ptr,
+				      num_pages, 1, 0, user_pages, NULL);
+	up_read(&mm->mmap_sem);
+	mutex_lock(&dev->struct_mutex);
+	if (pinned_pages < num_pages) {
+		ret = -EFAULT;
+		goto out;
+	}
+
+	ret = i915_gem_object_set_cpu_read_domain_range(obj,
+							args->offset,
+							args->size);
+	if (ret)
+		goto out;
+
+	do_bit17_swizzling = i915_gem_object_needs_bit17_swizzle(obj);
+
+	offset = args->offset;
 
 	while (remain > 0) {
 		struct page *page;
-		char *vaddr;
 
 		/* Operation in this page
 		 *
 		 * shmem_page_offset = offset within page in shmem file
+		 * data_page_index = page number in get_user_pages return
+		 * data_page_offset = offset with data_page_index page.
 		 * page_length = bytes to copy for this page
 		 */
 		shmem_page_offset = offset_in_page(offset);
+		data_page_index = data_ptr / PAGE_SIZE - first_data_page;
+		data_page_offset = offset_in_page(data_ptr);
+
 		page_length = remain;
 		if ((shmem_page_offset + page_length) > PAGE_SIZE)
 			page_length = PAGE_SIZE - shmem_page_offset;
+		if ((data_page_offset + page_length) > PAGE_SIZE)
+			page_length = PAGE_SIZE - data_page_offset;
 
 		page = shmem_read_mapping_page(mapping, offset >> PAGE_SHIFT);
 		if (IS_ERR(page)) {
@@ -419,38 +469,36 @@ i915_gem_shmem_pread_slow(struct drm_device *dev,
 			goto out;
 		}
 
-		page_do_bit17_swizzling = obj_do_bit17_swizzling &&
-			(page_to_phys(page) & (1 << 17)) != 0;
-
-		vaddr = kmap(page);
-		if (page_do_bit17_swizzling)
-			ret = __copy_to_user_swizzled(user_data,
-						      vaddr, shmem_page_offset,
-						      page_length);
-		else
-			ret = __copy_to_user(user_data,
-					     vaddr + shmem_page_offset,
-					     page_length);
-		kunmap(page);
+		if (do_bit17_swizzling) {
+			slow_shmem_bit17_copy(page,
+					      shmem_page_offset,
+					      user_pages[data_page_index],
+					      data_page_offset,
+					      page_length,
+					      1);
+		} else {
+			slow_shmem_copy(user_pages[data_page_index],
+					data_page_offset,
+					page,
+					shmem_page_offset,
+					page_length);
+		}
 
 		mark_page_accessed(page);
 		page_cache_release(page);
 
-		if (ret) {
-			ret = -EFAULT;
-			goto out;
-		}
-
 		remain -= page_length;
-		user_data += page_length;
+		data_ptr += page_length;
 		offset += page_length;
 	}
 
 out:
-	mutex_lock(&dev->struct_mutex);
-	/* Fixup: Kill any reinstated backing storage pages */
-	if (obj->madv == __I915_MADV_PURGED)
-		i915_gem_object_truncate(obj);
+	for (i = 0; i < pinned_pages; i++) {
+		SetPageDirty(user_pages[i]);
+		mark_page_accessed(user_pages[i]);
+		page_cache_release(user_pages[i]);
+	}
+	drm_free_large(user_pages);
 
 	return ret;
 }
@@ -752,11 +800,11 @@ i915_gem_shmem_pwrite_fast(struct drm_device *dev,
 		if (IS_ERR(page))
 			return PTR_ERR(page);
 
-		vaddr = kmap_atomic(page);
+		vaddr = kmap_atomic(page, KM_USER0);
 		ret = __copy_from_user_inatomic(vaddr + page_offset,
 						user_data,
 						page_length);
-		kunmap_atomic(vaddr);
+		kunmap_atomic(vaddr, KM_USER0);
 
 		set_page_dirty(page);
 		mark_page_accessed(page);
@@ -791,36 +839,71 @@ i915_gem_shmem_pwrite_slow(struct drm_device *dev,
 			   struct drm_file *file)
 {
 	struct address_space *mapping = obj->base.filp->f_path.dentry->d_inode->i_mapping;
+	struct mm_struct *mm = current->mm;
+	struct page **user_pages;
 	ssize_t remain;
-	loff_t offset;
-	char __user *user_data;
-	int shmem_page_offset, page_length, ret;
-	int obj_do_bit17_swizzling, page_do_bit17_swizzling;
+	loff_t offset, pinned_pages, i;
+	loff_t first_data_page, last_data_page, num_pages;
+	int shmem_page_offset;
+	int data_page_index,  data_page_offset;
+	int page_length;
+	int ret;
+	uint64_t data_ptr = args->data_ptr;
+	int do_bit17_swizzling;
 
-	user_data = (char __user *) (uintptr_t) args->data_ptr;
 	remain = args->size;
 
-	obj_do_bit17_swizzling = i915_gem_object_needs_bit17_swizzle(obj);
+	/* Pin the user pages containing the data.  We can't fault while
+	 * holding the struct mutex, and all of the pwrite implementations
+	 * want to hold it while dereferencing the user data.
+	 */
+	first_data_page = data_ptr / PAGE_SIZE;
+	last_data_page = (data_ptr + args->size - 1) / PAGE_SIZE;
+	num_pages = last_data_page - first_data_page + 1;
 
-	offset = args->offset;
-	obj->dirty = 1;
+	user_pages = drm_malloc_ab(num_pages, sizeof(struct page *));
+	if (user_pages == NULL)
+		return -ENOMEM;
 
 	mutex_unlock(&dev->struct_mutex);
+	down_read(&mm->mmap_sem);
+	pinned_pages = get_user_pages(current, mm, (uintptr_t)args->data_ptr,
+				      num_pages, 0, 0, user_pages, NULL);
+	up_read(&mm->mmap_sem);
+	mutex_lock(&dev->struct_mutex);
+	if (pinned_pages < num_pages) {
+		ret = -EFAULT;
+		goto out;
+	}
+
+	ret = i915_gem_object_set_to_cpu_domain(obj, 1);
+	if (ret)
+		goto out;
+
+	do_bit17_swizzling = i915_gem_object_needs_bit17_swizzle(obj);
+
+	offset = args->offset;
+	obj->dirty = 1;
 
 	while (remain > 0) {
 		struct page *page;
-		char *vaddr;
 
 		/* Operation in this page
 		 *
 		 * shmem_page_offset = offset within page in shmem file
+		 * data_page_index = page number in get_user_pages return
+		 * data_page_offset = offset with data_page_index page.
 		 * page_length = bytes to copy for this page
 		 */
 		shmem_page_offset = offset_in_page(offset);
+		data_page_index = data_ptr / PAGE_SIZE - first_data_page;
+		data_page_offset = offset_in_page(data_ptr);
 
 		page_length = remain;
 		if ((shmem_page_offset + page_length) > PAGE_SIZE)
 			page_length = PAGE_SIZE - shmem_page_offset;
+		if ((data_page_offset + page_length) > PAGE_SIZE)
+			page_length = PAGE_SIZE - data_page_offset;
 
 		page = shmem_read_mapping_page(mapping, offset >> PAGE_SHIFT);
 		if (IS_ERR(page)) {
@@ -828,45 +911,34 @@ i915_gem_shmem_pwrite_slow(struct drm_device *dev,
 			goto out;
 		}
 
-		page_do_bit17_swizzling = obj_do_bit17_swizzling &&
-			(page_to_phys(page) & (1 << 17)) != 0;
-
-		vaddr = kmap(page);
-		if (page_do_bit17_swizzling)
-			ret = __copy_from_user_swizzled(vaddr, shmem_page_offset,
-							user_data,
-							page_length);
-		else
-			ret = __copy_from_user(vaddr + shmem_page_offset,
-					       user_data,
-					       page_length);
-		kunmap(page);
+		if (do_bit17_swizzling) {
+			slow_shmem_bit17_copy(page,
+					      shmem_page_offset,
+					      user_pages[data_page_index],
+					      data_page_offset,
+					      page_length,
+					      0);
+		} else {
+			slow_shmem_copy(page,
+					shmem_page_offset,
+					user_pages[data_page_index],
+					data_page_offset,
+					page_length);
+		}
 
 		set_page_dirty(page);
 		mark_page_accessed(page);
 		page_cache_release(page);
 
-		if (ret) {
-			ret = -EFAULT;
-			goto out;
-		}
-
 		remain -= page_length;
-		user_data += page_length;
+		data_ptr += page_length;
 		offset += page_length;
 	}
 
 out:
-	mutex_lock(&dev->struct_mutex);
-	/* Fixup: Kill any reinstated backing storage pages */
-	if (obj->madv == __I915_MADV_PURGED)
-		i915_gem_object_truncate(obj);
-	/* and flush dirty cachelines in case the object isn't in the cpu write
-	 * domain anymore. */
-	if (obj->base.write_domain != I915_GEM_DOMAIN_CPU) {
-		i915_gem_clflush_object(obj);
-		intel_gtt_chipset_flush();
-	}
+	for (i = 0; i < pinned_pages; i++)
+		page_cache_release(user_pages[i]);
+	drm_free_large(user_pages);
 
 	return ret;
 }
@@ -922,13 +994,10 @@ i915_gem_pwrite_ioctl(struct drm_device *dev, void *data,
 	 * pread/pwrite currently are reading and writing from the CPU
 	 * perspective, requiring manual detiling by the client.
 	 */
-	if (obj->phys_obj) {
+	if (obj->phys_obj)
 		ret = i915_gem_phys_pwrite(dev, obj, args, file);
-		goto out;
-	}
-
-	if (obj->gtt_space &&
-	    obj->base.write_domain != I915_GEM_DOMAIN_CPU) {
+	else if (obj->gtt_space &&
+		 obj->base.write_domain != I915_GEM_DOMAIN_CPU) {
 		ret = i915_gem_object_pin(obj, 0, true);
 		if (ret)
 			goto out;
@@ -947,23 +1016,17 @@ i915_gem_pwrite_ioctl(struct drm_device *dev, void *data,
 
 out_unpin:
 		i915_gem_object_unpin(obj);
-
-		if (ret != -EFAULT)
+	} else {
+		ret = i915_gem_object_set_to_cpu_domain(obj, 1);
+		if (ret)
 			goto out;
-		/* Fall through to the shmfs paths because the gtt paths might
-		 * fail with non-page-backed user pointers (e.g. gtt mappings
-		 * when moving data between textures). */
-	}
-
-	ret = i915_gem_object_set_to_cpu_domain(obj, 1);
-	if (ret)
-		goto out;
 
-	ret = -EFAULT;
-	if (!i915_gem_object_needs_bit17_swizzle(obj))
-		ret = i915_gem_shmem_pwrite_fast(dev, obj, args, file);
-	if (ret == -EFAULT)
-		ret = i915_gem_shmem_pwrite_slow(dev, obj, args, file);
+		ret = -EFAULT;
+		if (!i915_gem_object_needs_bit17_swizzle(obj))
+			ret = i915_gem_shmem_pwrite_fast(dev, obj, args, file);
+		if (ret == -EFAULT)
+			ret = i915_gem_shmem_pwrite_slow(dev, obj, args, file);
+	}
 
 out:
 	drm_gem_object_unreference(&obj->base);
@@ -1076,6 +1139,7 @@ int
 i915_gem_mmap_ioctl(struct drm_device *dev, void *data,
 		    struct drm_file *file)
 {
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_i915_gem_mmap *args = data;
 	struct drm_gem_object *obj;
 	unsigned long addr;
@@ -1087,6 +1151,11 @@ i915_gem_mmap_ioctl(struct drm_device *dev, void *data,
 	if (obj == NULL)
 		return -ENOENT;
 
+	if (obj->size > dev_priv->mm.gtt_mappable_end) {
+		drm_gem_object_unreference_unlocked(obj);
+		return -E2BIG;
+	}
+
 	down_write(&current->mm->mmap_sem);
 	addr = do_mmap(obj->filp, 0, args->size,
 		       PROT_READ | PROT_WRITE, MAP_SHARED,
@@ -1201,6 +1270,74 @@ out:
 }
 
 /**
+ * i915_gem_create_mmap_offset - create a fake mmap offset for an object
+ * @obj: obj in question
+ *
+ * GEM memory mapping works by handing back to userspace a fake mmap offset
+ * it can use in a subsequent mmap(2) call.  The DRM core code then looks
+ * up the object based on the offset and sets up the various memory mapping
+ * structures.
+ *
+ * This routine allocates and attaches a fake offset for @obj.
+ */
+static int
+i915_gem_create_mmap_offset(struct drm_i915_gem_object *obj)
+{
+	struct drm_device *dev = obj->base.dev;
+	struct drm_gem_mm *mm = dev->mm_private;
+	struct drm_map_list *list;
+	struct drm_local_map *map;
+	int ret = 0;
+
+	/* Set the object up for mmap'ing */
+	list = &obj->base.map_list;
+	list->map = kzalloc(sizeof(struct drm_map_list), GFP_KERNEL);
+	if (!list->map)
+		return -ENOMEM;
+
+	map = list->map;
+	map->type = _DRM_GEM;
+	map->size = obj->base.size;
+	map->handle = obj;
+
+	/* Get a DRM GEM mmap offset allocated... */
+	list->file_offset_node = drm_mm_search_free(&mm->offset_manager,
+						    obj->base.size / PAGE_SIZE,
+						    0, 0);
+	if (!list->file_offset_node) {
+		DRM_ERROR("failed to allocate offset for bo %d\n",
+			  obj->base.name);
+		ret = -ENOSPC;
+		goto out_free_list;
+	}
+
+	list->file_offset_node = drm_mm_get_block(list->file_offset_node,
+						  obj->base.size / PAGE_SIZE,
+						  0);
+	if (!list->file_offset_node) {
+		ret = -ENOMEM;
+		goto out_free_list;
+	}
+
+	list->hash.key = list->file_offset_node->start;
+	ret = drm_ht_insert_item(&mm->offset_hash, &list->hash);
+	if (ret) {
+		DRM_ERROR("failed to add to map hash\n");
+		goto out_free_mm;
+	}
+
+	return 0;
+
+out_free_mm:
+	drm_mm_put_block(list->file_offset_node);
+out_free_list:
+	kfree(list->map);
+	list->map = NULL;
+
+	return ret;
+}
+
+/**
  * i915_gem_release_mmap - remove physical page mappings
  * @obj: obj in question
  *
@@ -1228,6 +1365,19 @@ i915_gem_release_mmap(struct drm_i915_gem_object *obj)
 	obj->fault_mappable = false;
 }
 
+static void
+i915_gem_free_mmap_offset(struct drm_i915_gem_object *obj)
+{
+	struct drm_device *dev = obj->base.dev;
+	struct drm_gem_mm *mm = dev->mm_private;
+	struct drm_map_list *list = &obj->base.map_list;
+
+	drm_ht_remove_item(&mm->offset_hash, &list->hash);
+	drm_mm_put_block(list->file_offset_node);
+	kfree(list->map);
+	list->map = NULL;
+}
+
 static uint32_t
 i915_gem_get_gtt_size(struct drm_device *dev, uint32_t size, int tiling_mode)
 {
@@ -1340,7 +1490,7 @@ i915_gem_mmap_gtt(struct drm_file *file,
 	}
 
 	if (!obj->base.map_list.map) {
-		ret = drm_gem_create_mmap_offset(&obj->base);
+		ret = i915_gem_create_mmap_offset(obj);
 		if (ret)
 			goto out;
 	}
@@ -1412,7 +1562,7 @@ i915_gem_object_get_pages_gtt(struct drm_i915_gem_object *obj,
 		obj->pages[i] = page;
 	}
 
-	if (i915_gem_object_needs_bit17_swizzle(obj))
+	if (obj->tiling_mode != I915_TILING_NONE)
 		i915_gem_object_do_bit_17_swizzle(obj);
 
 	return 0;
@@ -1434,7 +1584,7 @@ i915_gem_object_put_pages_gtt(struct drm_i915_gem_object *obj)
 
 	BUG_ON(obj->madv == __I915_MADV_PURGED);
 
-	if (i915_gem_object_needs_bit17_swizzle(obj))
+	if (obj->tiling_mode != I915_TILING_NONE)
 		i915_gem_object_save_bit_17_swizzle(obj);
 
 	if (obj->madv == I915_MADV_DONTNEED)
@@ -1581,28 +1731,6 @@ i915_gem_process_flushing_list(struct intel_ring_buffer *ring,
 	}
 }
 
-static u32
-i915_gem_get_seqno(struct drm_device *dev)
-{
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	u32 seqno = dev_priv->next_seqno;
-
-	/* reserve 0 for non-seqno */
-	if (++dev_priv->next_seqno == 0)
-		dev_priv->next_seqno = 1;
-
-	return seqno;
-}
-
-u32
-i915_gem_next_request_seqno(struct intel_ring_buffer *ring)
-{
-	if (ring->outstanding_lazy_request == 0)
-		ring->outstanding_lazy_request = i915_gem_get_seqno(ring->dev);
-
-	return ring->outstanding_lazy_request;
-}
-
 int
 i915_add_request(struct intel_ring_buffer *ring,
 		 struct drm_file *file,
@@ -1610,19 +1738,10 @@ i915_add_request(struct intel_ring_buffer *ring,
 {
 	drm_i915_private_t *dev_priv = ring->dev->dev_private;
 	uint32_t seqno;
-	u32 request_ring_position;
 	int was_empty;
 	int ret;
 
 	BUG_ON(request == NULL);
-	seqno = i915_gem_next_request_seqno(ring);
-
-	/* Record the position of the start of the request so that
-	 * should we detect the updated seqno part-way through the
-	 * GPU processing the request, we never over-estimate the
-	 * position of the head.
-	 */
-	request_ring_position = intel_ring_get_tail(ring);
 
 	ret = ring->add_request(ring, &seqno);
 	if (ret)
@@ -1632,7 +1751,6 @@ i915_add_request(struct intel_ring_buffer *ring,
 
 	request->seqno = seqno;
 	request->ring = ring;
-	request->tail = request_ring_position;
 	request->emitted_jiffies = jiffies;
 	was_empty = list_empty(&ring->request_list);
 	list_add_tail(&request->list, &ring->request_list);
@@ -1647,14 +1765,11 @@ i915_add_request(struct intel_ring_buffer *ring,
 		spin_unlock(&file_priv->mm.lock);
 	}
 
-	ring->outstanding_lazy_request = 0;
+	ring->outstanding_lazy_request = false;
 
 	if (!dev_priv->mm.suspended) {
-		if (i915_enable_hangcheck) {
-			mod_timer(&dev_priv->hangcheck_timer,
-				  jiffies +
-				  msecs_to_jiffies(DRM_I915_HANGCHECK_PERIOD));
-		}
+		mod_timer(&dev_priv->hangcheck_timer,
+			  jiffies + msecs_to_jiffies(DRM_I915_HANGCHECK_PERIOD));
 		if (was_empty)
 			queue_delayed_work(dev_priv->wq,
 					   &dev_priv->mm.retire_work, HZ);
@@ -1711,7 +1826,7 @@ static void i915_gem_reset_fences(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int i;
 
-	for (i = 0; i < dev_priv->num_fence_regs; i++) {
+	for (i = 0; i < 16; i++) {
 		struct drm_i915_fence_reg *reg = &dev_priv->fence_regs[i];
 		struct drm_i915_gem_object *obj = reg->obj;
 
@@ -1743,7 +1858,7 @@ void i915_gem_reset(struct drm_device *dev)
 	 * lost bo to the inactive list.
 	 */
 	while (!list_empty(&dev_priv->mm.flushing_list)) {
-		obj = list_first_entry(&dev_priv->mm.flushing_list,
+		obj= list_first_entry(&dev_priv->mm.flushing_list,
 				      struct drm_i915_gem_object,
 				      mm_list);
 
@@ -1769,7 +1884,7 @@ void i915_gem_reset(struct drm_device *dev)
 /**
  * This function clears the request list as sequence numbers are passed.
  */
-void
+static void
 i915_gem_retire_requests_ring(struct intel_ring_buffer *ring)
 {
 	uint32_t seqno;
@@ -1797,12 +1912,6 @@ i915_gem_retire_requests_ring(struct intel_ring_buffer *ring)
 			break;
 
 		trace_i915_gem_request_retire(ring, request->seqno);
-		/* We know the GPU must have read the request to have
-		 * sent us the seqno + interrupt, so use the position
-		 * of tail of the request to update the last known position
-		 * of the GPU head.
-		 */
-		ring->last_retired_head = request->tail;
 
 		list_del(&request->list);
 		i915_gem_request_remove_from_client(request);
@@ -1815,7 +1924,7 @@ i915_gem_retire_requests_ring(struct intel_ring_buffer *ring)
 	while (!list_empty(&ring->active_list)) {
 		struct drm_i915_gem_object *obj;
 
-		obj = list_first_entry(&ring->active_list,
+		obj= list_first_entry(&ring->active_list,
 				      struct drm_i915_gem_object,
 				      ring_list);
 
@@ -1915,8 +2024,7 @@ i915_gem_retire_work_handler(struct work_struct *work)
  */
 int
 i915_wait_request(struct intel_ring_buffer *ring,
-		  uint32_t seqno,
-		  bool do_retire)
+		  uint32_t seqno)
 {
 	drm_i915_private_t *dev_priv = ring->dev->dev_private;
 	u32 ier;
@@ -1979,9 +2087,9 @@ i915_wait_request(struct intel_ring_buffer *ring,
 					   || atomic_read(&dev_priv->mm.wedged));
 
 			ring->irq_put(ring);
-		} else if (wait_for_atomic(i915_seqno_passed(ring->get_seqno(ring),
-							     seqno) ||
-					   atomic_read(&dev_priv->mm.wedged), 3000))
+		} else if (wait_for(i915_seqno_passed(ring->get_seqno(ring),
+						      seqno) ||
+				    atomic_read(&dev_priv->mm.wedged), 3000))
 			ret = -EBUSY;
 		ring->waiting_seqno = 0;
 
@@ -1990,12 +2098,17 @@ i915_wait_request(struct intel_ring_buffer *ring,
 	if (atomic_read(&dev_priv->mm.wedged))
 		ret = -EAGAIN;
 
+	if (ret && ret != -ERESTARTSYS)
+		DRM_ERROR("%s returns %d (awaiting %d at %d, next %d)\n",
+			  __func__, ret, seqno, ring->get_seqno(ring),
+			  dev_priv->next_seqno);
+
 	/* Directly dispatch request retiring.  While we have the work queue
 	 * to handle this, the waiter on a request often wants an associated
 	 * buffer to have made it to the inactive list, and we would need
 	 * a separate wait queue to handle that.
 	 */
-	if (ret == 0 && do_retire)
+	if (ret == 0)
 		i915_gem_retire_requests_ring(ring);
 
 	return ret;
@@ -2019,8 +2132,7 @@ i915_gem_object_wait_rendering(struct drm_i915_gem_object *obj)
 	 * it.
 	 */
 	if (obj->active) {
-		ret = i915_wait_request(obj->ring, obj->last_rendering_seqno,
-					true);
+		ret = i915_wait_request(obj->ring, obj->last_rendering_seqno);
 		if (ret)
 			return ret;
 	}
@@ -2028,37 +2140,12 @@ i915_gem_object_wait_rendering(struct drm_i915_gem_object *obj)
 	return 0;
 }
 
-static void i915_gem_object_finish_gtt(struct drm_i915_gem_object *obj)
-{
-	u32 old_write_domain, old_read_domains;
-
-	/* Act a barrier for all accesses through the GTT */
-	mb();
-
-	/* Force a pagefault for domain tracking on next user access */
-	i915_gem_release_mmap(obj);
-
-	if ((obj->base.read_domains & I915_GEM_DOMAIN_GTT) == 0)
-		return;
-
-	old_read_domains = obj->base.read_domains;
-	old_write_domain = obj->base.write_domain;
-
-	obj->base.read_domains &= ~I915_GEM_DOMAIN_GTT;
-	obj->base.write_domain &= ~I915_GEM_DOMAIN_GTT;
-
-	trace_i915_gem_object_change_domain(obj,
-					    old_read_domains,
-					    old_write_domain);
-}
-
 /**
  * Unbinds an object from the GTT aperture.
  */
 int
 i915_gem_object_unbind(struct drm_i915_gem_object *obj)
 {
-	drm_i915_private_t *dev_priv = obj->base.dev->dev_private;
 	int ret = 0;
 
 	if (obj->gtt_space == NULL)
@@ -2069,28 +2156,23 @@ i915_gem_object_unbind(struct drm_i915_gem_object *obj)
 		return -EINVAL;
 	}
 
-	ret = i915_gem_object_finish_gpu(obj);
-	if (ret == -ERESTARTSYS)
-		return ret;
-	/* Continue on if we fail due to EIO, the GPU is hung so we
-	 * should be safe and we need to cleanup or else we might
-	 * cause memory corruption through use-after-free.
-	 */
-
-	i915_gem_object_finish_gtt(obj);
+	/* blow away mappings if mapped through GTT */
+	i915_gem_release_mmap(obj);
 
 	/* Move the object to the CPU domain to ensure that
 	 * any possible CPU writes while it's not in the GTT
-	 * are flushed when we go to remap it.
+	 * are flushed when we go to remap it. This will
+	 * also ensure that all pending GPU writes are finished
+	 * before we unbind.
 	 */
-	if (ret == 0)
-		ret = i915_gem_object_set_to_cpu_domain(obj, 1);
+	ret = i915_gem_object_set_to_cpu_domain(obj, 1);
 	if (ret == -ERESTARTSYS)
 		return ret;
+	/* Continue on if we fail due to EIO, the GPU is hung so we
+	 * should be safe and we need to cleanup or else we might
+	 * cause memory corruption through use-after-free.
+	 */
 	if (ret) {
-		/* In the event of a disaster, abandon all caches and
-		 * hope for the best.
-		 */
 		i915_gem_clflush_object(obj);
 		obj->base.read_domains = obj->base.write_domain = I915_GEM_DOMAIN_CPU;
 	}
@@ -2103,11 +2185,6 @@ i915_gem_object_unbind(struct drm_i915_gem_object *obj)
 	trace_i915_gem_object_unbind(obj);
 
 	i915_gem_gtt_unbind_object(obj);
-	if (obj->has_aliasing_ppgtt_mapping) {
-		i915_ppgtt_unbind_object(dev_priv->mm.aliasing_ppgtt, obj);
-		obj->has_aliasing_ppgtt_mapping = 0;
-	}
-
 	i915_gem_object_put_pages_gtt(obj);
 
 	list_del_init(&obj->gtt_list);
@@ -2147,7 +2224,7 @@ i915_gem_flush_ring(struct intel_ring_buffer *ring,
 	return 0;
 }
 
-static int i915_ring_idle(struct intel_ring_buffer *ring, bool do_retire)
+static int i915_ring_idle(struct intel_ring_buffer *ring)
 {
 	int ret;
 
@@ -2161,18 +2238,24 @@ static int i915_ring_idle(struct intel_ring_buffer *ring, bool do_retire)
 			return ret;
 	}
 
-	return i915_wait_request(ring, i915_gem_next_request_seqno(ring),
-				 do_retire);
+	return i915_wait_request(ring, i915_gem_next_request_seqno(ring));
 }
 
-int i915_gpu_idle(struct drm_device *dev, bool do_retire)
+int
+i915_gpu_idle(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = dev->dev_private;
+	bool lists_empty;
 	int ret, i;
 
+	lists_empty = (list_empty(&dev_priv->mm.flushing_list) &&
+		       list_empty(&dev_priv->mm.active_list));
+	if (lists_empty)
+		return 0;
+
 	/* Flush everything onto the inactive list. */
 	for (i = 0; i < I915_NUM_RINGS; i++) {
-		ret = i915_ring_idle(&dev_priv->ring[i], do_retire);
+		ret = i915_ring_idle(&dev_priv->ring[i]);
 		if (ret)
 			return ret;
 	}
@@ -2375,8 +2458,7 @@ i915_gem_object_flush_fence(struct drm_i915_gem_object *obj,
 		if (!ring_passed_seqno(obj->last_fenced_ring,
 				       obj->last_fenced_seqno)) {
 			ret = i915_wait_request(obj->last_fenced_ring,
-						obj->last_fenced_seqno,
-						true);
+						obj->last_fenced_seqno);
 			if (ret)
 				return ret;
 		}
@@ -2408,8 +2490,6 @@ i915_gem_object_put_fence(struct drm_i915_gem_object *obj)
 
 	if (obj->fence_reg != I915_FENCE_REG_NONE) {
 		struct drm_i915_private *dev_priv = obj->base.dev->dev_private;
-
-		WARN_ON(dev_priv->fence_regs[obj->fence_reg].pin_count);
 		i915_gem_clear_fence_reg(obj->base.dev,
 					 &dev_priv->fence_regs[obj->fence_reg]);
 
@@ -2434,7 +2514,7 @@ i915_find_fence_reg(struct drm_device *dev,
 		if (!reg->obj)
 			return reg;
 
-		if (!reg->pin_count)
+		if (!reg->obj->pin_count)
 			avail = reg;
 	}
 
@@ -2444,7 +2524,7 @@ i915_find_fence_reg(struct drm_device *dev,
 	/* None available, try to steal one or wait for a user to finish */
 	avail = first = NULL;
 	list_for_each_entry(reg, &dev_priv->mm.fence_list, lru_list) {
-		if (reg->pin_count)
+		if (reg->obj->pin_count)
 			continue;
 
 		if (first == NULL)
@@ -2519,8 +2599,7 @@ i915_gem_object_get_fence(struct drm_i915_gem_object *obj,
 				if (!ring_passed_seqno(obj->last_fenced_ring,
 						       reg->setup_seqno)) {
 					ret = i915_wait_request(obj->last_fenced_ring,
-								reg->setup_seqno,
-								true);
+								reg->setup_seqno);
 					if (ret)
 						return ret;
 				}
@@ -2539,7 +2618,7 @@ i915_gem_object_get_fence(struct drm_i915_gem_object *obj,
 
 	reg = i915_find_fence_reg(dev, pipelined);
 	if (reg == NULL)
-		return -EDEADLK;
+		return -ENOSPC;
 
 	ret = i915_gem_object_flush_fence(obj, pipelined);
 	if (ret)
@@ -2639,7 +2718,6 @@ i915_gem_clear_fence_reg(struct drm_device *dev,
 	list_del_init(&reg->lru_list);
 	reg->obj = NULL;
 	reg->setup_seqno = 0;
-	reg->pin_count = 0;
 }
 
 /**
@@ -2777,7 +2855,7 @@ i915_gem_object_bind_to_gtt(struct drm_i915_gem_object *obj,
 
 	fenceable =
 		obj->gtt_space->size == fence_size &&
-		(obj->gtt_space->start & (fence_alignment - 1)) == 0;
+		(obj->gtt_space->start & (fence_alignment -1)) == 0;
 
 	mappable =
 		obj->gtt_offset + obj->base.size <= dev_priv->mm.gtt_mappable_end;
@@ -2923,144 +3001,51 @@ i915_gem_object_set_to_gtt_domain(struct drm_i915_gem_object *obj, bool write)
 	return 0;
 }
 
-int i915_gem_object_set_cache_level(struct drm_i915_gem_object *obj,
-				    enum i915_cache_level cache_level)
-{
-	struct drm_device *dev = obj->base.dev;
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	int ret;
-
-	if (obj->cache_level == cache_level)
-		return 0;
-
-	if (obj->pin_count) {
-		DRM_DEBUG("can not change the cache level of pinned objects\n");
-		return -EBUSY;
-	}
-
-	if (obj->gtt_space) {
-		ret = i915_gem_object_finish_gpu(obj);
-		if (ret)
-			return ret;
-
-		i915_gem_object_finish_gtt(obj);
-
-		/* Before SandyBridge, you could not use tiling or fence
-		 * registers with snooped memory, so relinquish any fences
-		 * currently pointing to our region in the aperture.
-		 */
-		if (INTEL_INFO(obj->base.dev)->gen < 6) {
-			ret = i915_gem_object_put_fence(obj);
-			if (ret)
-				return ret;
-		}
-
-		i915_gem_gtt_rebind_object(obj, cache_level);
-		if (obj->has_aliasing_ppgtt_mapping)
-			i915_ppgtt_bind_object(dev_priv->mm.aliasing_ppgtt,
-					       obj, cache_level);
-	}
-
-	if (cache_level == I915_CACHE_NONE) {
-		u32 old_read_domains, old_write_domain;
-
-		/* If we're coming from LLC cached, then we haven't
-		 * actually been tracking whether the data is in the
-		 * CPU cache or not, since we only allow one bit set
-		 * in obj->write_domain and have been skipping the clflushes.
-		 * Just set it to the CPU cache for now.
-		 */
-		WARN_ON(obj->base.write_domain & ~I915_GEM_DOMAIN_CPU);
-		WARN_ON(obj->base.read_domains & ~I915_GEM_DOMAIN_CPU);
-
-		old_read_domains = obj->base.read_domains;
-		old_write_domain = obj->base.write_domain;
-
-		obj->base.read_domains = I915_GEM_DOMAIN_CPU;
-		obj->base.write_domain = I915_GEM_DOMAIN_CPU;
-
-		trace_i915_gem_object_change_domain(obj,
-						    old_read_domains,
-						    old_write_domain);
-	}
-
-	obj->cache_level = cache_level;
-	return 0;
-}
-
 /*
- * Prepare buffer for display plane (scanout, cursors, etc).
- * Can be called from an uninterruptible phase (modesetting) and allows
- * any flushes to be pipelined (for pageflips).
- *
- * For the display plane, we want to be in the GTT but out of any write
- * domains. So in many ways this looks like set_to_gtt_domain() apart from the
- * ability to pipeline the waits, pinning and any additional subtleties
- * that may differentiate the display plane from ordinary buffers.
+ * Prepare buffer for display plane. Use uninterruptible for possible flush
+ * wait, as in modesetting process we're not supposed to be interrupted.
  */
 int
-i915_gem_object_pin_to_display_plane(struct drm_i915_gem_object *obj,
-				     u32 alignment,
+i915_gem_object_set_to_display_plane(struct drm_i915_gem_object *obj,
 				     struct intel_ring_buffer *pipelined)
 {
-	u32 old_read_domains, old_write_domain;
+	uint32_t old_read_domains;
 	int ret;
 
+	/* Not valid to be called on unbound objects. */
+	if (obj->gtt_space == NULL)
+		return -EINVAL;
+
 	ret = i915_gem_object_flush_gpu_write_domain(obj);
 	if (ret)
 		return ret;
 
+
+	/* Currently, we are always called from an non-interruptible context. */
 	if (pipelined != obj->ring) {
 		ret = i915_gem_object_wait_rendering(obj);
-		if (ret == -ERESTARTSYS)
+		if (ret)
 			return ret;
 	}
 
-	/* The display engine is not coherent with the LLC cache on gen6.  As
-	 * a result, we make sure that the pinning that is about to occur is
-	 * done with uncached PTEs. This is lowest common denominator for all
-	 * chipsets.
-	 *
-	 * However for gen6+, we could do better by using the GFDT bit instead
-	 * of uncaching, which would allow us to flush all the LLC-cached data
-	 * with that bit in the PTE to main memory with just one PIPE_CONTROL.
-	 */
-	ret = i915_gem_object_set_cache_level(obj, I915_CACHE_NONE);
-	if (ret)
-		return ret;
-
-	/* As the user may map the buffer once pinned in the display plane
-	 * (e.g. libkms for the bootup splash), we have to ensure that we
-	 * always use map_and_fenceable for all scanout buffers.
-	 */
-	ret = i915_gem_object_pin(obj, alignment, true);
-	if (ret)
-		return ret;
-
 	i915_gem_object_flush_cpu_write_domain(obj);
 
-	old_write_domain = obj->base.write_domain;
 	old_read_domains = obj->base.read_domains;
-
-	/* It should now be out of any other write domains, and we can update
-	 * the domain values for our changes.
-	 */
-	BUG_ON((obj->base.write_domain & ~I915_GEM_DOMAIN_GTT) != 0);
 	obj->base.read_domains |= I915_GEM_DOMAIN_GTT;
 
 	trace_i915_gem_object_change_domain(obj,
 					    old_read_domains,
-					    old_write_domain);
+					    obj->base.write_domain);
 
 	return 0;
 }
 
 int
-i915_gem_object_finish_gpu(struct drm_i915_gem_object *obj)
+i915_gem_object_flush_gpu(struct drm_i915_gem_object *obj)
 {
 	int ret;
 
-	if ((obj->base.read_domains & I915_GEM_GPU_DOMAINS) == 0)
+	if (!obj->active)
 		return 0;
 
 	if (obj->base.write_domain & I915_GEM_GPU_DOMAINS) {
@@ -3069,13 +3054,7 @@ i915_gem_object_finish_gpu(struct drm_i915_gem_object *obj)
 			return ret;
 	}
 
-	ret = i915_gem_object_wait_rendering(obj);
-	if (ret)
-		return ret;
-
-	/* Ensure that we invalidate the GPU's caches and TLBs. */
-	obj->base.read_domains &= ~I915_GEM_GPU_DOMAINS;
-	return 0;
+	return i915_gem_object_wait_rendering(obj);
 }
 
 /**
@@ -3297,10 +3276,6 @@ i915_gem_ring_throttle(struct drm_device *dev, struct drm_file *file)
 
 			if (ret == 0 && atomic_read(&dev_priv->mm.wedged))
 				ret = -EIO;
-		} else if (wait_for_atomic(i915_seqno_passed(ring->get_seqno(ring),
-							     seqno) ||
-				    atomic_read(&dev_priv->mm.wedged), 3000)) {
-			ret = -EBUSY;
 		}
 	}
 
@@ -3504,11 +3479,9 @@ i915_gem_busy_ioctl(struct drm_device *dev, void *data,
 			 * so emit a request to do so.
 			 */
 			request = kzalloc(sizeof(*request), GFP_KERNEL);
-			if (request) {
-				ret = i915_add_request(obj->ring, NULL, request);
-				if (ret)
-					kfree(request);
-			} else
+			if (request)
+				ret = i915_add_request(obj->ring, NULL,request);
+			else
 				ret = -ENOMEM;
 		}
 
@@ -3532,7 +3505,7 @@ int
 i915_gem_throttle_ioctl(struct drm_device *dev, void *data,
 			struct drm_file *file_priv)
 {
-	return i915_gem_ring_throttle(dev, file_priv);
+    return i915_gem_ring_throttle(dev, file_priv);
 }
 
 int
@@ -3607,23 +3580,7 @@ struct drm_i915_gem_object *i915_gem_alloc_object(struct drm_device *dev,
 	obj->base.write_domain = I915_GEM_DOMAIN_CPU;
 	obj->base.read_domains = I915_GEM_DOMAIN_CPU;
 
-	if (HAS_LLC(dev)) {
-		/* On some devices, we can have the GPU use the LLC (the CPU
-		 * cache) for about a 10% performance improvement
-		 * compared to uncached.  Graphics requests other than
-		 * display scanout are coherent with the CPU in
-		 * accessing this cache.  This means in this mode we
-		 * don't need to clflush on the CPU side, and on the
-		 * GPU side we only need to flush internal caches to
-		 * get data visible to the CPU.
-		 *
-		 * However, we maintain the display planes as UC, and so
-		 * need to rebind when first used as such.
-		 */
-		obj->cache_level = I915_CACHE_LLC;
-	} else
-		obj->cache_level = I915_CACHE_NONE;
-
+	obj->cache_level = I915_CACHE_NONE;
 	obj->base.driver_private = NULL;
 	obj->fence_reg = I915_FENCE_REG_NONE;
 	INIT_LIST_HEAD(&obj->mm_list);
@@ -3661,7 +3618,7 @@ static void i915_gem_free_object_tail(struct drm_i915_gem_object *obj)
 	trace_i915_gem_object_destroy(obj);
 
 	if (obj->base.map_list.map)
-		drm_gem_free_mmap_offset(&obj->base);
+		i915_gem_free_mmap_offset(obj);
 
 	drm_gem_object_release(&obj->base);
 	i915_gem_info_remove_obj(dev_priv, obj->base.size);
@@ -3698,7 +3655,7 @@ i915_gem_idle(struct drm_device *dev)
 		return 0;
 	}
 
-	ret = i915_gpu_idle(dev, true);
+	ret = i915_gpu_idle(dev);
 	if (ret) {
 		mutex_unlock(&dev->struct_mutex);
 		return ret;
@@ -3733,71 +3690,12 @@ i915_gem_idle(struct drm_device *dev)
 	return 0;
 }
 
-void i915_gem_init_swizzling(struct drm_device *dev)
-{
-	drm_i915_private_t *dev_priv = dev->dev_private;
-
-	if (INTEL_INFO(dev)->gen < 5 ||
-	    dev_priv->mm.bit_6_swizzle_x == I915_BIT_6_SWIZZLE_NONE)
-		return;
-
-	I915_WRITE(DISP_ARB_CTL, I915_READ(DISP_ARB_CTL) |
-				 DISP_TILE_SURFACE_SWIZZLING);
-
-	if (IS_GEN5(dev))
-		return;
-
-	I915_WRITE(TILECTL, I915_READ(TILECTL) | TILECTL_SWZCTL);
-	if (IS_GEN6(dev))
-		I915_WRITE(ARB_MODE, ARB_MODE_ENABLE(ARB_MODE_SWIZZLE_SNB));
-	else
-		I915_WRITE(ARB_MODE, ARB_MODE_ENABLE(ARB_MODE_SWIZZLE_IVB));
-}
-
-void i915_gem_init_ppgtt(struct drm_device *dev)
-{
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	uint32_t pd_offset;
-	struct intel_ring_buffer *ring;
-	int i;
-
-	if (!dev_priv->mm.aliasing_ppgtt)
-		return;
-
-	pd_offset = dev_priv->mm.aliasing_ppgtt->pd_offset;
-	pd_offset /= 64; /* in cachelines, */
-	pd_offset <<= 16;
-
-	if (INTEL_INFO(dev)->gen == 6) {
-		uint32_t ecochk = I915_READ(GAM_ECOCHK);
-		I915_WRITE(GAM_ECOCHK, ecochk | ECOCHK_SNB_BIT |
-				       ECOCHK_PPGTT_CACHE64B);
-		I915_WRITE(GFX_MODE, GFX_MODE_ENABLE(GFX_PPGTT_ENABLE));
-	} else if (INTEL_INFO(dev)->gen >= 7) {
-		I915_WRITE(GAM_ECOCHK, ECOCHK_PPGTT_CACHE64B);
-		/* GFX_MODE is per-ring on gen7+ */
-	}
-
-	for (i = 0; i < I915_NUM_RINGS; i++) {
-		ring = &dev_priv->ring[i];
-
-		if (INTEL_INFO(dev)->gen >= 7)
-			I915_WRITE(RING_MODE_GEN7(ring),
-				   GFX_MODE_ENABLE(GFX_PPGTT_ENABLE));
-
-		I915_WRITE(RING_PP_DIR_DCLV(ring), PP_DIR_DCLV_2G);
-		I915_WRITE(RING_PP_DIR_BASE(ring), pd_offset);
-	}
-}
-
 int
-i915_gem_init_hw(struct drm_device *dev)
+i915_gem_init_ringbuffer(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = dev->dev_private;
 	int ret;
 
-	i915_gem_init_swizzling(dev);
-
 	ret = intel_init_render_ring_buffer(dev);
 	if (ret)
 		return ret;
@@ -3816,8 +3714,6 @@ i915_gem_init_hw(struct drm_device *dev)
 
 	dev_priv->next_seqno = 1;
 
-	i915_gem_init_ppgtt(dev);
-
 	return 0;
 
 cleanup_bsd_ring:
@@ -3855,7 +3751,7 @@ i915_gem_entervt_ioctl(struct drm_device *dev, void *data,
 	mutex_lock(&dev->struct_mutex);
 	dev_priv->mm.suspended = 0;
 
-	ret = i915_gem_init_hw(dev);
+	ret = i915_gem_init_ringbuffer(dev);
 	if (ret != 0) {
 		mutex_unlock(&dev->struct_mutex);
 		return ret;
@@ -3932,7 +3828,7 @@ i915_gem_load(struct drm_device *dev)
 	INIT_LIST_HEAD(&dev_priv->mm.gtt_list);
 	for (i = 0; i < I915_NUM_RINGS; i++)
 		init_ring_lists(&dev_priv->ring[i]);
-	for (i = 0; i < I915_MAX_NUM_FENCES; i++)
+	for (i = 0; i < 16; i++)
 		INIT_LIST_HEAD(&dev_priv->fence_regs[i].lru_list);
 	INIT_DELAYED_WORK(&dev_priv->mm.retire_work,
 			  i915_gem_retire_work_handler);
@@ -4250,7 +4146,7 @@ rescan:
 		 * This has a dramatic impact to reduce the number of
 		 * OOM-killer events whilst running the GPU aggressively.
 		 */
-		if (i915_gpu_idle(dev, true) == 0)
+		if (i915_gpu_idle(dev) == 0)
 			goto rescan;
 	}
 	mutex_unlock(&dev->struct_mutex);
diff --git a/drivers/gpu/drm/i915/i915_gem_debug.c b/drivers/gpu/drm/i915/i915_gem_debug.c
index cc93cac..8da1899 100644
--- a/drivers/gpu/drm/i915/i915_gem_debug.c
+++ b/drivers/gpu/drm/i915/i915_gem_debug.c
@@ -72,7 +72,7 @@ i915_verify_lists(struct drm_device *dev)
 			break;
 		} else if (!obj->active ||
 			   (obj->base.write_domain & I915_GEM_GPU_DOMAINS) == 0 ||
-			   list_empty(&obj->gpu_write_list)) {
+			   list_empty(&obj->gpu_write_list)){
 			DRM_ERROR("invalid flushing %p (a %d w %x gwl %d)\n",
 				  obj,
 				  obj->active,
@@ -157,7 +157,7 @@ i915_gem_object_check_coherency(struct drm_i915_gem_object *obj, int handle)
 	for (page = 0; page < obj->size / PAGE_SIZE; page++) {
 		int i;
 
-		backing_map = kmap_atomic(obj->pages[page]);
+		backing_map = kmap_atomic(obj->pages[page], KM_USER0);
 
 		if (backing_map == NULL) {
 			DRM_ERROR("failed to map backing page\n");
@@ -181,13 +181,13 @@ i915_gem_object_check_coherency(struct drm_i915_gem_object *obj, int handle)
 				}
 			}
 		}
-		kunmap_atomic(backing_map);
+		kunmap_atomic(backing_map, KM_USER0);
 		backing_map = NULL;
 	}
 
  out:
 	if (backing_map != NULL)
-		kunmap_atomic(backing_map);
+		kunmap_atomic(backing_map, KM_USER0);
 	iounmap(gtt_mapping);
 
 	/* give syslog time to catch up */
diff --git a/drivers/gpu/drm/i915/i915_gem_evict.c b/drivers/gpu/drm/i915/i915_gem_evict.c
index 21a8271..da05a26 100644
--- a/drivers/gpu/drm/i915/i915_gem_evict.c
+++ b/drivers/gpu/drm/i915/i915_gem_evict.c
@@ -36,6 +36,7 @@ static bool
 mark_free(struct drm_i915_gem_object *obj, struct list_head *unwind)
 {
 	list_add(&obj->exec_list, unwind);
+	drm_gem_object_reference(&obj->base);
 	return drm_mm_scan_add_block(obj->gtt_space);
 }
 
@@ -48,6 +49,21 @@ i915_gem_evict_something(struct drm_device *dev, int min_size,
 	struct drm_i915_gem_object *obj;
 	int ret = 0;
 
+	i915_gem_retire_requests(dev);
+
+	/* Re-check for free space after retiring requests */
+	if (mappable) {
+		if (drm_mm_search_free_in_range(&dev_priv->mm.gtt_space,
+						min_size, alignment, 0,
+						dev_priv->mm.gtt_mappable_end,
+						0))
+			return 0;
+	} else {
+		if (drm_mm_search_free(&dev_priv->mm.gtt_space,
+				       min_size, alignment, 0))
+			return 0;
+	}
+
 	trace_i915_gem_evict(dev, min_size, alignment, mappable);
 
 	/*
@@ -106,7 +122,7 @@ i915_gem_evict_something(struct drm_device *dev, int min_size,
 			goto found;
 	}
 	list_for_each_entry(obj, &dev_priv->mm.active_list, mm_list) {
-		if (!obj->base.write_domain || obj->pin_count)
+		if (! obj->base.write_domain || obj->pin_count)
 			continue;
 
 		if (mark_free(obj, &unwind_list))
@@ -123,6 +139,7 @@ i915_gem_evict_something(struct drm_device *dev, int min_size,
 		BUG_ON(ret);
 
 		list_del_init(&obj->exec_list);
+		drm_gem_object_unreference(&obj->base);
 	}
 
 	/* We expect the caller to unpin, evict all and try again, or give up.
@@ -141,10 +158,10 @@ found:
 				       exec_list);
 		if (drm_mm_scan_remove_block(obj->gtt_space)) {
 			list_move(&obj->exec_list, &eviction_list);
-			drm_gem_object_reference(&obj->base);
 			continue;
 		}
 		list_del_init(&obj->exec_list);
+		drm_gem_object_unreference(&obj->base);
 	}
 
 	/* Unbinding will emit any required flushes */
@@ -178,7 +195,7 @@ i915_gem_evict_everything(struct drm_device *dev, bool purgeable_only)
 	trace_i915_gem_evict_everything(dev, purgeable_only);
 
 	/* Flush everything (on to the inactive lists) and evict */
-	ret = i915_gpu_idle(dev, true);
+	ret = i915_gpu_idle(dev);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 97b47c5..1ca53ff 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -32,7 +32,6 @@
 #include "i915_drv.h"
 #include "i915_trace.h"
 #include "intel_drv.h"
-#include <linux/dma_remapping.h>
 
 struct change_domains {
 	uint32_t invalidate_domains;
@@ -203,9 +202,9 @@ i915_gem_object_set_to_gpu_domain(struct drm_i915_gem_object *obj,
 	cd->invalidate_domains |= invalidate_domains;
 	cd->flush_domains |= flush_domains;
 	if (flush_domains & I915_GEM_GPU_DOMAINS)
-		cd->flush_rings |= intel_ring_flag(obj->ring);
+		cd->flush_rings |= obj->ring->id;
 	if (invalidate_domains & I915_GEM_GPU_DOMAINS)
-		cd->flush_rings |= intel_ring_flag(ring);
+		cd->flush_rings |= ring->id;
 }
 
 struct eb_objects {
@@ -287,14 +286,14 @@ i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj,
 	 * exec_object list, so it should have a GTT space bound by now.
 	 */
 	if (unlikely(target_offset == 0)) {
-		DRM_DEBUG("No GTT space found for object %d\n",
+		DRM_ERROR("No GTT space found for object %d\n",
 			  reloc->target_handle);
 		return ret;
 	}
 
 	/* Validate that the target is in a valid r/w GPU domain */
 	if (unlikely(reloc->write_domain & (reloc->write_domain - 1))) {
-		DRM_DEBUG("reloc with multiple write domains: "
+		DRM_ERROR("reloc with multiple write domains: "
 			  "obj %p target %d offset %d "
 			  "read %08x write %08x",
 			  obj, reloc->target_handle,
@@ -303,9 +302,8 @@ i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj,
 			  reloc->write_domain);
 		return ret;
 	}
-	if (unlikely((reloc->write_domain | reloc->read_domains)
-		     & ~I915_GEM_GPU_DOMAINS)) {
-		DRM_DEBUG("reloc with read/write non-GPU domains: "
+	if (unlikely((reloc->write_domain | reloc->read_domains) & I915_GEM_DOMAIN_CPU)) {
+		DRM_ERROR("reloc with read/write CPU domains: "
 			  "obj %p target %d offset %d "
 			  "read %08x write %08x",
 			  obj, reloc->target_handle,
@@ -316,7 +314,7 @@ i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj,
 	}
 	if (unlikely(reloc->write_domain && target_obj->pending_write_domain &&
 		     reloc->write_domain != target_obj->pending_write_domain)) {
-		DRM_DEBUG("Write domain conflict: "
+		DRM_ERROR("Write domain conflict: "
 			  "obj %p target %d offset %d "
 			  "new %08x old %08x\n",
 			  obj, reloc->target_handle,
@@ -337,7 +335,7 @@ i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj,
 
 	/* Check that the relocation address is valid... */
 	if (unlikely(reloc->offset > obj->base.size - 4)) {
-		DRM_DEBUG("Relocation beyond object bounds: "
+		DRM_ERROR("Relocation beyond object bounds: "
 			  "obj %p target %d offset %d size %d.\n",
 			  obj, reloc->target_handle,
 			  (int) reloc->offset,
@@ -345,7 +343,7 @@ i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj,
 		return ret;
 	}
 	if (unlikely(reloc->offset & 3)) {
-		DRM_DEBUG("Relocation not 4-byte aligned: "
+		DRM_ERROR("Relocation not 4-byte aligned: "
 			  "obj %p target %d offset %d.\n",
 			  obj, reloc->target_handle,
 			  (int) reloc->offset);
@@ -462,60 +460,11 @@ i915_gem_execbuffer_relocate(struct drm_device *dev,
 	return ret;
 }
 
-#define  __EXEC_OBJECT_HAS_FENCE (1<<31)
-
-static int
-pin_and_fence_object(struct drm_i915_gem_object *obj,
-		     struct intel_ring_buffer *ring)
-{
-	struct drm_i915_gem_exec_object2 *entry = obj->exec_entry;
-	bool has_fenced_gpu_access = INTEL_INFO(ring->dev)->gen < 4;
-	bool need_fence, need_mappable;
-	int ret;
-
-	need_fence =
-		has_fenced_gpu_access &&
-		entry->flags & EXEC_OBJECT_NEEDS_FENCE &&
-		obj->tiling_mode != I915_TILING_NONE;
-	need_mappable =
-		entry->relocation_count ? true : need_fence;
-
-	ret = i915_gem_object_pin(obj, entry->alignment, need_mappable);
-	if (ret)
-		return ret;
-
-	if (has_fenced_gpu_access) {
-		if (entry->flags & EXEC_OBJECT_NEEDS_FENCE) {
-			if (obj->tiling_mode) {
-				ret = i915_gem_object_get_fence(obj, ring);
-				if (ret)
-					goto err_unpin;
-
-				entry->flags |= __EXEC_OBJECT_HAS_FENCE;
-				i915_gem_object_pin_fence(obj);
-			} else {
-				ret = i915_gem_object_put_fence(obj);
-				if (ret)
-					goto err_unpin;
-			}
-		}
-		obj->pending_fenced_gpu_access = need_fence;
-	}
-
-	entry->offset = obj->gtt_offset;
-	return 0;
-
-err_unpin:
-	i915_gem_object_unpin(obj);
-	return ret;
-}
-
 static int
 i915_gem_execbuffer_reserve(struct intel_ring_buffer *ring,
 			    struct drm_file *file,
 			    struct list_head *objects)
 {
-	drm_i915_private_t *dev_priv = ring->dev->dev_private;
 	struct drm_i915_gem_object *obj;
 	int ret, retry;
 	bool has_fenced_gpu_access = INTEL_INFO(ring->dev)->gen < 4;
@@ -568,7 +517,6 @@ i915_gem_execbuffer_reserve(struct intel_ring_buffer *ring,
 		list_for_each_entry(obj, objects, exec_list) {
 			struct drm_i915_gem_exec_object2 *entry = obj->exec_entry;
 			bool need_fence, need_mappable;
-
 			if (!obj->gtt_space)
 				continue;
 
@@ -583,55 +531,58 @@ i915_gem_execbuffer_reserve(struct intel_ring_buffer *ring,
 			    (need_mappable && !obj->map_and_fenceable))
 				ret = i915_gem_object_unbind(obj);
 			else
-				ret = pin_and_fence_object(obj, ring);
+				ret = i915_gem_object_pin(obj,
+							  entry->alignment,
+							  need_mappable);
 			if (ret)
 				goto err;
+
+			entry++;
 		}
 
 		/* Bind fresh objects */
 		list_for_each_entry(obj, objects, exec_list) {
-			if (obj->gtt_space)
-				continue;
-
-			ret = pin_and_fence_object(obj, ring);
-			if (ret) {
-				int ret_ignore;
-
-				/* This can potentially raise a harmless
-				 * -EINVAL if we failed to bind in the above
-				 * call. It cannot raise -EINTR since we know
-				 * that the bo is freshly bound and so will
-				 * not need to be flushed or waited upon.
-				 */
-				ret_ignore = i915_gem_object_unbind(obj);
-				(void)ret_ignore;
-				WARN_ON(obj->gtt_space);
-				break;
-			}
-		}
+			struct drm_i915_gem_exec_object2 *entry = obj->exec_entry;
+			bool need_fence;
 
-		/* Decrement pin count for bound objects */
-		list_for_each_entry(obj, objects, exec_list) {
-			struct drm_i915_gem_exec_object2 *entry;
+			need_fence =
+				has_fenced_gpu_access &&
+				entry->flags & EXEC_OBJECT_NEEDS_FENCE &&
+				obj->tiling_mode != I915_TILING_NONE;
 
-			if (!obj->gtt_space)
-				continue;
+			if (!obj->gtt_space) {
+				bool need_mappable =
+					entry->relocation_count ? true : need_fence;
 
-			entry = obj->exec_entry;
-			if (entry->flags & __EXEC_OBJECT_HAS_FENCE) {
-				i915_gem_object_unpin_fence(obj);
-				entry->flags &= ~__EXEC_OBJECT_HAS_FENCE;
+				ret = i915_gem_object_pin(obj,
+							  entry->alignment,
+							  need_mappable);
+				if (ret)
+					break;
 			}
 
-			i915_gem_object_unpin(obj);
+			if (has_fenced_gpu_access) {
+				if (need_fence) {
+					ret = i915_gem_object_get_fence(obj, ring);
+					if (ret)
+						break;
+				} else if (entry->flags & EXEC_OBJECT_NEEDS_FENCE &&
+					   obj->tiling_mode == I915_TILING_NONE) {
+					/* XXX pipelined! */
+					ret = i915_gem_object_put_fence(obj);
+					if (ret)
+						break;
+				}
+				obj->pending_fenced_gpu_access = need_fence;
+			}
 
-			/* ... and ensure ppgtt mapping exist if needed. */
-			if (dev_priv->mm.aliasing_ppgtt && !obj->has_aliasing_ppgtt_mapping) {
-				i915_ppgtt_bind_object(dev_priv->mm.aliasing_ppgtt,
-						       obj, obj->cache_level);
+			entry->offset = obj->gtt_offset;
+		}
 
-				obj->has_aliasing_ppgtt_mapping = 1;
-			}
+		/* Decrement pin count for bound objects */
+		list_for_each_entry(obj, objects, exec_list) {
+			if (obj->gtt_space)
+				i915_gem_object_unpin(obj);
 		}
 
 		if (ret != -ENOSPC || retry > 1)
@@ -648,19 +599,16 @@ i915_gem_execbuffer_reserve(struct intel_ring_buffer *ring,
 	} while (1);
 
 err:
-	list_for_each_entry_continue_reverse(obj, objects, exec_list) {
-		struct drm_i915_gem_exec_object2 *entry;
-
-		if (!obj->gtt_space)
-			continue;
-
-		entry = obj->exec_entry;
-		if (entry->flags & __EXEC_OBJECT_HAS_FENCE) {
-			i915_gem_object_unpin_fence(obj);
-			entry->flags &= ~__EXEC_OBJECT_HAS_FENCE;
-		}
+	obj = list_entry(obj->exec_list.prev,
+			 struct drm_i915_gem_object,
+			 exec_list);
+	while (objects != &obj->exec_list) {
+		if (obj->gtt_space)
+			i915_gem_object_unpin(obj);
 
-		i915_gem_object_unpin(obj);
+		obj = list_entry(obj->exec_list.prev,
+				 struct drm_i915_gem_object,
+				 exec_list);
 	}
 
 	return ret;
@@ -754,7 +702,7 @@ i915_gem_execbuffer_relocate_slow(struct drm_device *dev,
 		obj = to_intel_bo(drm_gem_object_lookup(dev, file,
 							exec[i].handle));
 		if (&obj->base == NULL) {
-			DRM_DEBUG("Invalid object handle %d at index %d\n",
+			DRM_ERROR("Invalid object handle %d at index %d\n",
 				   exec[i].handle, i);
 			ret = -ENOENT;
 			goto err;
@@ -819,22 +767,6 @@ i915_gem_execbuffer_flush(struct drm_device *dev,
 	return 0;
 }
 
-static bool
-intel_enable_semaphores(struct drm_device *dev)
-{
-	if (INTEL_INFO(dev)->gen < 6)
-		return 0;
-
-	if (i915_semaphores >= 0)
-		return i915_semaphores;
-
-	/* Disable semaphores on SNB */
-	if (INTEL_INFO(dev)->gen == 6)
-		return 0;
-
-	return 1;
-}
-
 static int
 i915_gem_execbuffer_sync_rings(struct drm_i915_gem_object *obj,
 			       struct intel_ring_buffer *to)
@@ -847,7 +779,7 @@ i915_gem_execbuffer_sync_rings(struct drm_i915_gem_object *obj,
 		return 0;
 
 	/* XXX gpu semaphores are implicated in various hard hangs on SNB */
-	if (!intel_enable_semaphores(obj->base.dev))
+	if (INTEL_INFO(obj->base.dev)->gen < 6 || !i915_semaphores)
 		return i915_gem_object_wait_rendering(obj);
 
 	idx = intel_ring_sync_index(from, to);
@@ -873,8 +805,7 @@ i915_gem_execbuffer_sync_rings(struct drm_i915_gem_object *obj,
 	}
 
 	from->sync_seqno[idx] = seqno;
-
-	return to->sync_to(to, from, seqno - 1);
+	return intel_ring_sync(to, from, seqno - 1);
 }
 
 static int
@@ -1048,31 +979,6 @@ i915_gem_execbuffer_retire_commands(struct drm_device *dev,
 }
 
 static int
-i915_reset_gen7_sol_offsets(struct drm_device *dev,
-			    struct intel_ring_buffer *ring)
-{
-	drm_i915_private_t *dev_priv = dev->dev_private;
-	int ret, i;
-
-	if (!IS_GEN7(dev) || ring != &dev_priv->ring[RCS])
-		return 0;
-
-	ret = intel_ring_begin(ring, 4 * 3);
-	if (ret)
-		return ret;
-
-	for (i = 0; i < 4; i++) {
-		intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(1));
-		intel_ring_emit(ring, GEN7_SO_WRITE_OFFSET(i));
-		intel_ring_emit(ring, 0);
-	}
-
-	intel_ring_advance(ring);
-
-	return 0;
-}
-
-static int
 i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 		       struct drm_file *file,
 		       struct drm_i915_gem_execbuffer2 *args,
@@ -1086,11 +992,10 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 	struct intel_ring_buffer *ring;
 	u32 exec_start, exec_len;
 	u32 seqno;
-	u32 mask;
 	int ret, mode, i;
 
 	if (!i915_gem_check_execbuffer(args)) {
-		DRM_DEBUG("execbuf with invalid offset/length\n");
+		DRM_ERROR("execbuf with invalid offset/length\n");
 		return -EINVAL;
 	}
 
@@ -1105,26 +1010,25 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 		break;
 	case I915_EXEC_BSD:
 		if (!HAS_BSD(dev)) {
-			DRM_DEBUG("execbuf with invalid ring (BSD)\n");
+			DRM_ERROR("execbuf with invalid ring (BSD)\n");
 			return -EINVAL;
 		}
 		ring = &dev_priv->ring[VCS];
 		break;
 	case I915_EXEC_BLT:
 		if (!HAS_BLT(dev)) {
-			DRM_DEBUG("execbuf with invalid ring (BLT)\n");
+			DRM_ERROR("execbuf with invalid ring (BLT)\n");
 			return -EINVAL;
 		}
 		ring = &dev_priv->ring[BCS];
 		break;
 	default:
-		DRM_DEBUG("execbuf with unknown ring: %d\n",
+		DRM_ERROR("execbuf with unknown ring: %d\n",
 			  (int)(args->flags & I915_EXEC_RING_MASK));
 		return -EINVAL;
 	}
 
 	mode = args->flags & I915_EXEC_CONSTANTS_MASK;
-	mask = I915_EXEC_CONSTANTS_MASK;
 	switch (mode) {
 	case I915_EXEC_CONSTANTS_REL_GENERAL:
 	case I915_EXEC_CONSTANTS_ABSOLUTE:
@@ -1138,24 +1042,33 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 			    mode == I915_EXEC_CONSTANTS_REL_SURFACE)
 				return -EINVAL;
 
-			/* The HW changed the meaning on this bit on gen6 */
-			if (INTEL_INFO(dev)->gen >= 6)
-				mask &= ~I915_EXEC_CONSTANTS_REL_SURFACE;
+			ret = intel_ring_begin(ring, 4);
+			if (ret)
+				return ret;
+
+			intel_ring_emit(ring, MI_NOOP);
+			intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(1));
+			intel_ring_emit(ring, INSTPM);
+			intel_ring_emit(ring,
+					I915_EXEC_CONSTANTS_MASK << 16 | mode);
+			intel_ring_advance(ring);
+
+			dev_priv->relative_constants_mode = mode;
 		}
 		break;
 	default:
-		DRM_DEBUG("execbuf with unknown constants: %d\n", mode);
+		DRM_ERROR("execbuf with unknown constants: %d\n", mode);
 		return -EINVAL;
 	}
 
 	if (args->buffer_count < 1) {
-		DRM_DEBUG("execbuf with %d buffers\n", args->buffer_count);
+		DRM_ERROR("execbuf with %d buffers\n", args->buffer_count);
 		return -EINVAL;
 	}
 
 	if (args->num_cliprects != 0) {
 		if (ring != &dev_priv->ring[RCS]) {
-			DRM_DEBUG("clip rectangles are only valid with the render ring\n");
+			DRM_ERROR("clip rectangles are only valid with the render ring\n");
 			return -EINVAL;
 		}
 
@@ -1205,7 +1118,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 		obj = to_intel_bo(drm_gem_object_lookup(dev, file,
 							exec[i].handle));
 		if (&obj->base == NULL) {
-			DRM_DEBUG("Invalid object handle %d at index %d\n",
+			DRM_ERROR("Invalid object handle %d at index %d\n",
 				   exec[i].handle, i);
 			/* prevent error path from reading uninitialized data */
 			ret = -ENOENT;
@@ -1213,7 +1126,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 		}
 
 		if (!list_empty(&obj->exec_list)) {
-			DRM_DEBUG("Object %p [handle %d, index %d] appears more than once in object list\n",
+			DRM_ERROR("Object %p [handle %d, index %d] appears more than once in object list\n",
 				   obj, exec[i].handle, i);
 			ret = -EINVAL;
 			goto err;
@@ -1251,7 +1164,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 
 	/* Set the pending read domains for the batch buffer to COMMAND */
 	if (batch_obj->base.pending_write_domain) {
-		DRM_DEBUG("Attempting to use self-modifying batch buffer\n");
+		DRM_ERROR("Attempting to use self-modifying batch buffer\n");
 		ret = -EINVAL;
 		goto err;
 	}
@@ -1268,7 +1181,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 			 * so every billion or so execbuffers, we need to stall
 			 * the GPU in order to reset the counters.
 			 */
-			ret = i915_gpu_idle(dev, true);
+			ret = i915_gpu_idle(dev);
 			if (ret)
 				goto err;
 
@@ -1276,27 +1189,6 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 		}
 	}
 
-	if (ring == &dev_priv->ring[RCS] &&
-	    mode != dev_priv->relative_constants_mode) {
-		ret = intel_ring_begin(ring, 4);
-		if (ret)
-				goto err;
-
-		intel_ring_emit(ring, MI_NOOP);
-		intel_ring_emit(ring, MI_LOAD_REGISTER_IMM(1));
-		intel_ring_emit(ring, INSTPM);
-		intel_ring_emit(ring, mask << 16 | mode);
-		intel_ring_advance(ring);
-
-		dev_priv->relative_constants_mode = mode;
-	}
-
-	if (args->flags & I915_EXEC_GEN7_SOL_RESET) {
-		ret = i915_reset_gen7_sol_offsets(dev, ring);
-		if (ret)
-			goto err;
-	}
-
 	trace_i915_gem_ring_dispatch(ring, seqno);
 
 	exec_start = batch_obj->gtt_offset + args->batch_start_offset;
@@ -1356,7 +1248,7 @@ i915_gem_execbuffer(struct drm_device *dev, void *data,
 	int ret, i;
 
 	if (args->buffer_count < 1) {
-		DRM_DEBUG("execbuf with %d buffers\n", args->buffer_count);
+		DRM_ERROR("execbuf with %d buffers\n", args->buffer_count);
 		return -EINVAL;
 	}
 
@@ -1364,7 +1256,7 @@ i915_gem_execbuffer(struct drm_device *dev, void *data,
 	exec_list = drm_malloc_ab(sizeof(*exec_list), args->buffer_count);
 	exec2_list = drm_malloc_ab(sizeof(*exec2_list), args->buffer_count);
 	if (exec_list == NULL || exec2_list == NULL) {
-		DRM_DEBUG("Failed to allocate exec list for %d buffers\n",
+		DRM_ERROR("Failed to allocate exec list for %d buffers\n",
 			  args->buffer_count);
 		drm_free_large(exec_list);
 		drm_free_large(exec2_list);
@@ -1375,7 +1267,7 @@ i915_gem_execbuffer(struct drm_device *dev, void *data,
 			     (uintptr_t) args->buffers_ptr,
 			     sizeof(*exec_list) * args->buffer_count);
 	if (ret != 0) {
-		DRM_DEBUG("copy %d exec entries failed %d\n",
+		DRM_ERROR("copy %d exec entries failed %d\n",
 			  args->buffer_count, ret);
 		drm_free_large(exec_list);
 		drm_free_large(exec2_list);
@@ -1416,7 +1308,7 @@ i915_gem_execbuffer(struct drm_device *dev, void *data,
 				   sizeof(*exec_list) * args->buffer_count);
 		if (ret) {
 			ret = -EFAULT;
-			DRM_DEBUG("failed to copy %d exec entries "
+			DRM_ERROR("failed to copy %d exec entries "
 				  "back to user (%d)\n",
 				  args->buffer_count, ret);
 		}
@@ -1437,7 +1329,7 @@ i915_gem_execbuffer2(struct drm_device *dev, void *data,
 
 	if (args->buffer_count < 1 ||
 	    args->buffer_count > UINT_MAX / sizeof(*exec2_list)) {
-		DRM_DEBUG("execbuf2 with %d buffers\n", args->buffer_count);
+		DRM_ERROR("execbuf2 with %d buffers\n", args->buffer_count);
 		return -EINVAL;
 	}
 
@@ -1447,7 +1339,7 @@ i915_gem_execbuffer2(struct drm_device *dev, void *data,
 		exec2_list = drm_malloc_ab(sizeof(*exec2_list),
 					   args->buffer_count);
 	if (exec2_list == NULL) {
-		DRM_DEBUG("Failed to allocate exec list for %d buffers\n",
+		DRM_ERROR("Failed to allocate exec list for %d buffers\n",
 			  args->buffer_count);
 		return -ENOMEM;
 	}
@@ -1456,7 +1348,7 @@ i915_gem_execbuffer2(struct drm_device *dev, void *data,
 			     (uintptr_t) args->buffers_ptr,
 			     sizeof(*exec2_list) * args->buffer_count);
 	if (ret != 0) {
-		DRM_DEBUG("copy %d exec entries failed %d\n",
+		DRM_ERROR("copy %d exec entries failed %d\n",
 			  args->buffer_count, ret);
 		drm_free_large(exec2_list);
 		return -EFAULT;
@@ -1471,7 +1363,7 @@ i915_gem_execbuffer2(struct drm_device *dev, void *data,
 				   sizeof(*exec2_list) * args->buffer_count);
 		if (ret) {
 			ret = -EFAULT;
-			DRM_DEBUG("failed to copy %d exec entries "
+			DRM_ERROR("failed to copy %d exec entries "
 				  "back to user (%d)\n",
 				  args->buffer_count, ret);
 		}
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 2eacd78..e46b645 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -29,279 +29,6 @@
 #include "i915_trace.h"
 #include "intel_drv.h"
 
-/* PPGTT support for Sandybdrige/Gen6 and later */
-static void i915_ppgtt_clear_range(struct i915_hw_ppgtt *ppgtt,
-				   unsigned first_entry,
-				   unsigned num_entries)
-{
-	uint32_t *pt_vaddr;
-	uint32_t scratch_pte;
-	unsigned act_pd = first_entry / I915_PPGTT_PT_ENTRIES;
-	unsigned first_pte = first_entry % I915_PPGTT_PT_ENTRIES;
-	unsigned last_pte, i;
-
-	scratch_pte = GEN6_PTE_ADDR_ENCODE(ppgtt->scratch_page_dma_addr);
-	scratch_pte |= GEN6_PTE_VALID | GEN6_PTE_CACHE_LLC;
-
-	while (num_entries) {
-		last_pte = first_pte + num_entries;
-		if (last_pte > I915_PPGTT_PT_ENTRIES)
-			last_pte = I915_PPGTT_PT_ENTRIES;
-
-		pt_vaddr = kmap_atomic(ppgtt->pt_pages[act_pd]);
-
-		for (i = first_pte; i < last_pte; i++)
-			pt_vaddr[i] = scratch_pte;
-
-		kunmap_atomic(pt_vaddr);
-
-		num_entries -= last_pte - first_pte;
-		first_pte = 0;
-		act_pd++;
-	}
-}
-
-int i915_gem_init_aliasing_ppgtt(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct i915_hw_ppgtt *ppgtt;
-	uint32_t pd_entry;
-	unsigned first_pd_entry_in_global_pt;
-	uint32_t __iomem *pd_addr;
-	int i;
-	int ret = -ENOMEM;
-
-	/* ppgtt PDEs reside in the global gtt pagetable, which has 512*1024
-	 * entries. For aliasing ppgtt support we just steal them at the end for
-	 * now. */
-	first_pd_entry_in_global_pt = 512*1024 - I915_PPGTT_PD_ENTRIES;
-
-	ppgtt = kzalloc(sizeof(*ppgtt), GFP_KERNEL);
-	if (!ppgtt)
-		return ret;
-
-	ppgtt->num_pd_entries = I915_PPGTT_PD_ENTRIES;
-	ppgtt->pt_pages = kzalloc(sizeof(struct page *)*ppgtt->num_pd_entries,
-				  GFP_KERNEL);
-	if (!ppgtt->pt_pages)
-		goto err_ppgtt;
-
-	for (i = 0; i < ppgtt->num_pd_entries; i++) {
-		ppgtt->pt_pages[i] = alloc_page(GFP_KERNEL);
-		if (!ppgtt->pt_pages[i])
-			goto err_pt_alloc;
-	}
-
-	if (dev_priv->mm.gtt->needs_dmar) {
-		ppgtt->pt_dma_addr = kzalloc(sizeof(dma_addr_t)
-						*ppgtt->num_pd_entries,
-					     GFP_KERNEL);
-		if (!ppgtt->pt_dma_addr)
-			goto err_pt_alloc;
-	}
-
-	pd_addr = dev_priv->mm.gtt->gtt + first_pd_entry_in_global_pt;
-	for (i = 0; i < ppgtt->num_pd_entries; i++) {
-		dma_addr_t pt_addr;
-		if (dev_priv->mm.gtt->needs_dmar) {
-			pt_addr = pci_map_page(dev->pdev, ppgtt->pt_pages[i],
-					       0, 4096,
-					       PCI_DMA_BIDIRECTIONAL);
-
-			if (pci_dma_mapping_error(dev->pdev,
-						  pt_addr)) {
-				ret = -EIO;
-				goto err_pd_pin;
-
-			}
-			ppgtt->pt_dma_addr[i] = pt_addr;
-		} else
-			pt_addr = page_to_phys(ppgtt->pt_pages[i]);
-
-		pd_entry = GEN6_PDE_ADDR_ENCODE(pt_addr);
-		pd_entry |= GEN6_PDE_VALID;
-
-		writel(pd_entry, pd_addr + i);
-	}
-	readl(pd_addr);
-
-	ppgtt->scratch_page_dma_addr = dev_priv->mm.gtt->scratch_page_dma;
-
-	i915_ppgtt_clear_range(ppgtt, 0,
-			       ppgtt->num_pd_entries*I915_PPGTT_PT_ENTRIES);
-
-	ppgtt->pd_offset = (first_pd_entry_in_global_pt)*sizeof(uint32_t);
-
-	dev_priv->mm.aliasing_ppgtt = ppgtt;
-
-	return 0;
-
-err_pd_pin:
-	if (ppgtt->pt_dma_addr) {
-		for (i--; i >= 0; i--)
-			pci_unmap_page(dev->pdev, ppgtt->pt_dma_addr[i],
-				       4096, PCI_DMA_BIDIRECTIONAL);
-	}
-err_pt_alloc:
-	kfree(ppgtt->pt_dma_addr);
-	for (i = 0; i < ppgtt->num_pd_entries; i++) {
-		if (ppgtt->pt_pages[i])
-			__free_page(ppgtt->pt_pages[i]);
-	}
-	kfree(ppgtt->pt_pages);
-err_ppgtt:
-	kfree(ppgtt);
-
-	return ret;
-}
-
-void i915_gem_cleanup_aliasing_ppgtt(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct i915_hw_ppgtt *ppgtt = dev_priv->mm.aliasing_ppgtt;
-	int i;
-
-	if (!ppgtt)
-		return;
-
-	if (ppgtt->pt_dma_addr) {
-		for (i = 0; i < ppgtt->num_pd_entries; i++)
-			pci_unmap_page(dev->pdev, ppgtt->pt_dma_addr[i],
-				       4096, PCI_DMA_BIDIRECTIONAL);
-	}
-
-	kfree(ppgtt->pt_dma_addr);
-	for (i = 0; i < ppgtt->num_pd_entries; i++)
-		__free_page(ppgtt->pt_pages[i]);
-	kfree(ppgtt->pt_pages);
-	kfree(ppgtt);
-}
-
-static void i915_ppgtt_insert_sg_entries(struct i915_hw_ppgtt *ppgtt,
-					 struct scatterlist *sg_list,
-					 unsigned sg_len,
-					 unsigned first_entry,
-					 uint32_t pte_flags)
-{
-	uint32_t *pt_vaddr, pte;
-	unsigned act_pd = first_entry / I915_PPGTT_PT_ENTRIES;
-	unsigned first_pte = first_entry % I915_PPGTT_PT_ENTRIES;
-	unsigned i, j, m, segment_len;
-	dma_addr_t page_addr;
-	struct scatterlist *sg;
-
-	/* init sg walking */
-	sg = sg_list;
-	i = 0;
-	segment_len = sg_dma_len(sg) >> PAGE_SHIFT;
-	m = 0;
-
-	while (i < sg_len) {
-		pt_vaddr = kmap_atomic(ppgtt->pt_pages[act_pd]);
-
-		for (j = first_pte; j < I915_PPGTT_PT_ENTRIES; j++) {
-			page_addr = sg_dma_address(sg) + (m << PAGE_SHIFT);
-			pte = GEN6_PTE_ADDR_ENCODE(page_addr);
-			pt_vaddr[j] = pte | pte_flags;
-
-			/* grab the next page */
-			m++;
-			if (m == segment_len) {
-				sg = sg_next(sg);
-				i++;
-				if (i == sg_len)
-					break;
-
-				segment_len = sg_dma_len(sg) >> PAGE_SHIFT;
-				m = 0;
-			}
-		}
-
-		kunmap_atomic(pt_vaddr);
-
-		first_pte = 0;
-		act_pd++;
-	}
-}
-
-static void i915_ppgtt_insert_pages(struct i915_hw_ppgtt *ppgtt,
-				    unsigned first_entry, unsigned num_entries,
-				    struct page **pages, uint32_t pte_flags)
-{
-	uint32_t *pt_vaddr, pte;
-	unsigned act_pd = first_entry / I915_PPGTT_PT_ENTRIES;
-	unsigned first_pte = first_entry % I915_PPGTT_PT_ENTRIES;
-	unsigned last_pte, i;
-	dma_addr_t page_addr;
-
-	while (num_entries) {
-		last_pte = first_pte + num_entries;
-		last_pte = min_t(unsigned, last_pte, I915_PPGTT_PT_ENTRIES);
-
-		pt_vaddr = kmap_atomic(ppgtt->pt_pages[act_pd]);
-
-		for (i = first_pte; i < last_pte; i++) {
-			page_addr = page_to_phys(*pages);
-			pte = GEN6_PTE_ADDR_ENCODE(page_addr);
-			pt_vaddr[i] = pte | pte_flags;
-
-			pages++;
-		}
-
-		kunmap_atomic(pt_vaddr);
-
-		num_entries -= last_pte - first_pte;
-		first_pte = 0;
-		act_pd++;
-	}
-}
-
-void i915_ppgtt_bind_object(struct i915_hw_ppgtt *ppgtt,
-			    struct drm_i915_gem_object *obj,
-			    enum i915_cache_level cache_level)
-{
-	struct drm_device *dev = obj->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	uint32_t pte_flags = GEN6_PTE_VALID;
-
-	switch (cache_level) {
-	case I915_CACHE_LLC_MLC:
-		pte_flags |= GEN6_PTE_CACHE_LLC_MLC;
-		break;
-	case I915_CACHE_LLC:
-		pte_flags |= GEN6_PTE_CACHE_LLC;
-		break;
-	case I915_CACHE_NONE:
-		pte_flags |= GEN6_PTE_UNCACHED;
-		break;
-	default:
-		BUG();
-	}
-
-	if (dev_priv->mm.gtt->needs_dmar) {
-		BUG_ON(!obj->sg_list);
-
-		i915_ppgtt_insert_sg_entries(ppgtt,
-					     obj->sg_list,
-					     obj->num_sg,
-					     obj->gtt_space->start >> PAGE_SHIFT,
-					     pte_flags);
-	} else
-		i915_ppgtt_insert_pages(ppgtt,
-					obj->gtt_space->start >> PAGE_SHIFT,
-					obj->base.size >> PAGE_SHIFT,
-					obj->pages,
-					pte_flags);
-}
-
-void i915_ppgtt_unbind_object(struct i915_hw_ppgtt *ppgtt,
-			      struct drm_i915_gem_object *obj)
-{
-	i915_ppgtt_clear_range(ppgtt,
-			       obj->gtt_space->start >> PAGE_SHIFT,
-			       obj->base.size >> PAGE_SHIFT);
-}
-
 /* XXX kill agp_type! */
 static unsigned int cache_level_to_agp_type(struct drm_device *dev,
 					    enum i915_cache_level cache_level)
@@ -322,28 +49,6 @@ static unsigned int cache_level_to_agp_type(struct drm_device *dev,
 	}
 }
 
-static bool do_idling(struct drm_i915_private *dev_priv)
-{
-	bool ret = dev_priv->mm.interruptible;
-
-	if (unlikely(dev_priv->mm.gtt->do_idle_maps)) {
-		dev_priv->mm.interruptible = false;
-		if (i915_gpu_idle(dev_priv->dev, false)) {
-			DRM_ERROR("Couldn't idle GPU\n");
-			/* Wait a bit, in hopes it avoids the hang */
-			udelay(10);
-		}
-	}
-
-	return ret;
-}
-
-static void undo_idling(struct drm_i915_private *dev_priv, bool interruptible)
-{
-	if (unlikely(dev_priv->mm.gtt->do_idle_maps))
-		dev_priv->mm.interruptible = interruptible;
-}
-
 void i915_gem_restore_gtt_mappings(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -354,8 +59,24 @@ void i915_gem_restore_gtt_mappings(struct drm_device *dev)
 			      (dev_priv->mm.gtt_end - dev_priv->mm.gtt_start) / PAGE_SIZE);
 
 	list_for_each_entry(obj, &dev_priv->mm.gtt_list, gtt_list) {
+		unsigned int agp_type =
+			cache_level_to_agp_type(dev, obj->cache_level);
+
 		i915_gem_clflush_object(obj);
-		i915_gem_gtt_rebind_object(obj, obj->cache_level);
+
+		if (dev_priv->mm.gtt->needs_dmar) {
+			BUG_ON(!obj->sg_list);
+
+			intel_gtt_insert_sg_entries(obj->sg_list,
+						    obj->num_sg,
+						    obj->gtt_space->start >> PAGE_SHIFT,
+						    agp_type);
+		} else
+			intel_gtt_insert_pages(obj->gtt_space->start
+						   >> PAGE_SHIFT,
+					       obj->base.size >> PAGE_SHIFT,
+					       obj->pages,
+					       agp_type);
 	}
 
 	intel_gtt_chipset_flush();
@@ -389,35 +110,8 @@ int i915_gem_gtt_bind_object(struct drm_i915_gem_object *obj)
 	return 0;
 }
 
-void i915_gem_gtt_rebind_object(struct drm_i915_gem_object *obj,
-				enum i915_cache_level cache_level)
-{
-	struct drm_device *dev = obj->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	unsigned int agp_type = cache_level_to_agp_type(dev, cache_level);
-
-	if (dev_priv->mm.gtt->needs_dmar) {
-		BUG_ON(!obj->sg_list);
-
-		intel_gtt_insert_sg_entries(obj->sg_list,
-					    obj->num_sg,
-					    obj->gtt_space->start >> PAGE_SHIFT,
-					    agp_type);
-	} else
-		intel_gtt_insert_pages(obj->gtt_space->start >> PAGE_SHIFT,
-				       obj->base.size >> PAGE_SHIFT,
-				       obj->pages,
-				       agp_type);
-}
-
 void i915_gem_gtt_unbind_object(struct drm_i915_gem_object *obj)
 {
-	struct drm_device *dev = obj->base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	bool interruptible;
-
-	interruptible = do_idling(dev_priv);
-
 	intel_gtt_clear_range(obj->gtt_space->start >> PAGE_SHIFT,
 			      obj->base.size >> PAGE_SHIFT);
 
@@ -425,6 +119,4 @@ void i915_gem_gtt_unbind_object(struct drm_i915_gem_object *obj)
 		intel_gtt_unmap_memory(obj->sg_list, obj->num_sg);
 		obj->sg_list = NULL;
 	}
-
-	undo_idling(dev_priv, interruptible);
 }
diff --git a/drivers/gpu/drm/i915/i915_gem_tiling.c b/drivers/gpu/drm/i915/i915_gem_tiling.c
index 1a93066..99c4faa 100644
--- a/drivers/gpu/drm/i915/i915_gem_tiling.c
+++ b/drivers/gpu/drm/i915/i915_gem_tiling.c
@@ -92,25 +92,7 @@ i915_gem_detect_bit_6_swizzle(struct drm_device *dev)
 	uint32_t swizzle_x = I915_BIT_6_SWIZZLE_UNKNOWN;
 	uint32_t swizzle_y = I915_BIT_6_SWIZZLE_UNKNOWN;
 
-	if (INTEL_INFO(dev)->gen >= 6) {
-		uint32_t dimm_c0, dimm_c1;
-		dimm_c0 = I915_READ(MAD_DIMM_C0);
-		dimm_c1 = I915_READ(MAD_DIMM_C1);
-		dimm_c0 &= MAD_DIMM_A_SIZE_MASK | MAD_DIMM_B_SIZE_MASK;
-		dimm_c1 &= MAD_DIMM_A_SIZE_MASK | MAD_DIMM_B_SIZE_MASK;
-		/* Enable swizzling when the channels are populated with
-		 * identically sized dimms. We don't need to check the 3rd
-		 * channel because no cpu with gpu attached ships in that
-		 * configuration. Also, swizzling only makes sense for 2
-		 * channels anyway. */
-		if (dimm_c0 == dimm_c1) {
-			swizzle_x = I915_BIT_6_SWIZZLE_9_10;
-			swizzle_y = I915_BIT_6_SWIZZLE_9;
-		} else {
-			swizzle_x = I915_BIT_6_SWIZZLE_NONE;
-			swizzle_y = I915_BIT_6_SWIZZLE_NONE;
-		}
-	} else if (IS_GEN5(dev)) {
+	if (INTEL_INFO(dev)->gen >= 5) {
 		/* On Ironlake whatever DRAM config, GPU always do
 		 * same swizzling setup.
 		 */
@@ -122,10 +104,10 @@ i915_gem_detect_bit_6_swizzle(struct drm_device *dev)
 		 */
 		swizzle_x = I915_BIT_6_SWIZZLE_NONE;
 		swizzle_y = I915_BIT_6_SWIZZLE_NONE;
-	} else if (IS_MOBILE(dev) || (IS_GEN3(dev) && !IS_G33(dev))) {
+	} else if (IS_MOBILE(dev)) {
 		uint32_t dcc;
 
-		/* On 9xx chipsets, channel interleave by the CPU is
+		/* On mobile 9xx chipsets, channel interleave by the CPU is
 		 * determined by DCC.  For single-channel, neither the CPU
 		 * nor the GPU do swizzling.  For dual channel interleaved,
 		 * the GPU's interleave is bit 9 and 10 for X tiled, and bit
@@ -458,9 +440,14 @@ i915_gem_swizzle_page(struct page *page)
 void
 i915_gem_object_do_bit_17_swizzle(struct drm_i915_gem_object *obj)
 {
+	struct drm_device *dev = obj->base.dev;
+	drm_i915_private_t *dev_priv = dev->dev_private;
 	int page_count = obj->base.size >> PAGE_SHIFT;
 	int i;
 
+	if (dev_priv->mm.bit_6_swizzle_x != I915_BIT_6_SWIZZLE_9_10_17)
+		return;
+
 	if (obj->bit_17 == NULL)
 		return;
 
@@ -477,9 +464,14 @@ i915_gem_object_do_bit_17_swizzle(struct drm_i915_gem_object *obj)
 void
 i915_gem_object_save_bit_17_swizzle(struct drm_i915_gem_object *obj)
 {
+	struct drm_device *dev = obj->base.dev;
+	drm_i915_private_t *dev_priv = dev->dev_private;
 	int page_count = obj->base.size >> PAGE_SHIFT;
 	int i;
 
+	if (dev_priv->mm.bit_6_swizzle_x != I915_BIT_6_SWIZZLE_9_10_17)
+		return;
+
 	if (obj->bit_17 == NULL) {
 		obj->bit_17 = kmalloc(BITS_TO_LONGS(page_count) *
 					   sizeof(long), GFP_KERNEL);
diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index f57e5cf..d05f03c 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -364,12 +364,10 @@ static void notify_ring(struct drm_device *dev,
 
 	ring->irq_seqno = seqno;
 	wake_up_all(&ring->irq_queue);
-	if (i915_enable_hangcheck) {
-		dev_priv->hangcheck_count = 0;
-		mod_timer(&dev_priv->hangcheck_timer,
-			  jiffies +
-			  msecs_to_jiffies(DRM_I915_HANGCHECK_PERIOD));
-	}
+
+	dev_priv->hangcheck_count = 0;
+	mod_timer(&dev_priv->hangcheck_timer,
+		  jiffies + msecs_to_jiffies(DRM_I915_HANGCHECK_PERIOD));
 }
 
 static void gen6_pm_rps_work(struct work_struct *work)
@@ -383,7 +381,6 @@ static void gen6_pm_rps_work(struct work_struct *work)
 	pm_iir = dev_priv->pm_iir;
 	dev_priv->pm_iir = 0;
 	pm_imr = I915_READ(GEN6_PMIMR);
-	I915_WRITE(GEN6_PMIMR, 0);
 	spin_unlock_irq(&dev_priv->rps_lock);
 
 	if (!pm_iir)
@@ -421,6 +418,7 @@ static void gen6_pm_rps_work(struct work_struct *work)
 	 * an *extremely* unlikely race with gen6_rps_enable() that is prevented
 	 * by holding struct_mutex for the duration of the write.
 	 */
+	I915_WRITE(GEN6_PMIMR, pm_imr & ~pm_iir);
 	mutex_unlock(&dev_priv->dev->struct_mutex);
 }
 
@@ -533,9 +531,8 @@ static irqreturn_t ivybridge_irq_handler(DRM_IRQ_ARGS)
 		unsigned long flags;
 		spin_lock_irqsave(&dev_priv->rps_lock, flags);
 		WARN(dev_priv->pm_iir & pm_iir, "Missed a PM interrupt\n");
+		I915_WRITE(GEN6_PMIMR, pm_iir);
 		dev_priv->pm_iir |= pm_iir;
-		I915_WRITE(GEN6_PMIMR, dev_priv->pm_iir);
-		POSTING_READ(GEN6_PMIMR);
 		spin_unlock_irqrestore(&dev_priv->rps_lock, flags);
 		queue_work(dev_priv->wq, &dev_priv->rps_work);
 	}
@@ -647,9 +644,8 @@ static irqreturn_t ironlake_irq_handler(DRM_IRQ_ARGS)
 		unsigned long flags;
 		spin_lock_irqsave(&dev_priv->rps_lock, flags);
 		WARN(dev_priv->pm_iir & pm_iir, "Missed a PM interrupt\n");
+		I915_WRITE(GEN6_PMIMR, pm_iir);
 		dev_priv->pm_iir |= pm_iir;
-		I915_WRITE(GEN6_PMIMR, dev_priv->pm_iir);
-		POSTING_READ(GEN6_PMIMR);
 		spin_unlock_irqrestore(&dev_priv->rps_lock, flags);
 		queue_work(dev_priv->wq, &dev_priv->rps_work);
 	}
@@ -710,13 +706,14 @@ i915_error_object_create(struct drm_i915_private *dev_priv,
 
 	page_count = src->base.size / PAGE_SIZE;
 
-	dst = kmalloc(sizeof(*dst) + page_count * sizeof(u32 *), GFP_ATOMIC);
+	dst = kmalloc(sizeof(*dst) + page_count * sizeof (u32 *), GFP_ATOMIC);
 	if (dst == NULL)
 		return NULL;
 
 	reloc_offset = src->gtt_offset;
 	for (page = 0; page < page_count; page++) {
 		unsigned long flags;
+		void __iomem *s;
 		void *d;
 
 		d = kmalloc(PAGE_SIZE, GFP_ATOMIC);
@@ -724,29 +721,10 @@ i915_error_object_create(struct drm_i915_private *dev_priv,
 			goto unwind;
 
 		local_irq_save(flags);
-		if (reloc_offset < dev_priv->mm.gtt_mappable_end) {
-			void __iomem *s;
-
-			/* Simply ignore tiling or any overlapping fence.
-			 * It's part of the error state, and this hopefully
-			 * captures what the GPU read.
-			 */
-
-			s = io_mapping_map_atomic_wc(dev_priv->mm.gtt_mapping,
-						     reloc_offset);
-			memcpy_fromio(d, s, PAGE_SIZE);
-			io_mapping_unmap_atomic(s);
-		} else {
-			void *s;
-
-			drm_clflush_pages(&src->pages[page], 1);
-
-			s = kmap_atomic(src->pages[page]);
-			memcpy(d, s, PAGE_SIZE);
-			kunmap_atomic(s);
-
-			drm_clflush_pages(&src->pages[page], 1);
-		}
+		s = io_mapping_map_atomic_wc(dev_priv->mm.gtt_mapping,
+					     reloc_offset);
+		memcpy_fromio(d, s, PAGE_SIZE);
+		io_mapping_unmap_atomic(s);
 		local_irq_restore(flags);
 
 		dst->pages[page] = d;
@@ -785,11 +763,11 @@ i915_error_state_free(struct drm_device *dev,
 {
 	int i;
 
-	for (i = 0; i < ARRAY_SIZE(error->ring); i++) {
-		i915_error_object_free(error->ring[i].batchbuffer);
-		i915_error_object_free(error->ring[i].ringbuffer);
-		kfree(error->ring[i].requests);
-	}
+	for (i = 0; i < ARRAY_SIZE(error->batchbuffer); i++)
+		i915_error_object_free(error->batchbuffer[i]);
+
+	for (i = 0; i < ARRAY_SIZE(error->ringbuffer); i++)
+		i915_error_object_free(error->ringbuffer[i]);
 
 	kfree(error->active_bo);
 	kfree(error->overlay);
@@ -819,7 +797,7 @@ static u32 capture_bo_list(struct drm_i915_error_buffer *err,
 		err->tiling = obj->tiling_mode;
 		err->dirty = obj->dirty;
 		err->purgeable = obj->madv != I915_MADV_WILLNEED;
-		err->ring = obj->ring ? obj->ring->id : -1;
+		err->ring = obj->ring ? obj->ring->id : 0;
 		err->cache_level = obj->cache_level;
 
 		if (++i == count)
@@ -891,92 +869,6 @@ i915_error_first_batchbuffer(struct drm_i915_private *dev_priv,
 	return NULL;
 }
 
-static void i915_record_ring_state(struct drm_device *dev,
-				   struct drm_i915_error_state *error,
-				   struct intel_ring_buffer *ring)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	if (INTEL_INFO(dev)->gen >= 6) {
-		error->faddr[ring->id] = I915_READ(RING_DMA_FADD(ring->mmio_base));
-		error->fault_reg[ring->id] = I915_READ(RING_FAULT_REG(ring));
-		error->semaphore_mboxes[ring->id][0]
-			= I915_READ(RING_SYNC_0(ring->mmio_base));
-		error->semaphore_mboxes[ring->id][1]
-			= I915_READ(RING_SYNC_1(ring->mmio_base));
-	}
-
-	if (INTEL_INFO(dev)->gen >= 4) {
-		error->ipeir[ring->id] = I915_READ(RING_IPEIR(ring->mmio_base));
-		error->ipehr[ring->id] = I915_READ(RING_IPEHR(ring->mmio_base));
-		error->instdone[ring->id] = I915_READ(RING_INSTDONE(ring->mmio_base));
-		error->instps[ring->id] = I915_READ(RING_INSTPS(ring->mmio_base));
-		if (ring->id == RCS) {
-			error->instdone1 = I915_READ(INSTDONE1);
-			error->bbaddr = I915_READ64(BB_ADDR);
-		}
-	} else {
-		error->ipeir[ring->id] = I915_READ(IPEIR);
-		error->ipehr[ring->id] = I915_READ(IPEHR);
-		error->instdone[ring->id] = I915_READ(INSTDONE);
-	}
-
-	error->instpm[ring->id] = I915_READ(RING_INSTPM(ring->mmio_base));
-	error->seqno[ring->id] = ring->get_seqno(ring);
-	error->acthd[ring->id] = intel_ring_get_active_head(ring);
-	error->head[ring->id] = I915_READ_HEAD(ring);
-	error->tail[ring->id] = I915_READ_TAIL(ring);
-
-	error->cpu_ring_head[ring->id] = ring->head;
-	error->cpu_ring_tail[ring->id] = ring->tail;
-}
-
-static void i915_gem_record_rings(struct drm_device *dev,
-				  struct drm_i915_error_state *error)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_i915_gem_request *request;
-	int i, count;
-
-	for (i = 0; i < I915_NUM_RINGS; i++) {
-		struct intel_ring_buffer *ring = &dev_priv->ring[i];
-
-		if (ring->obj == NULL)
-			continue;
-
-		i915_record_ring_state(dev, error, ring);
-
-		error->ring[i].batchbuffer =
-			i915_error_first_batchbuffer(dev_priv, ring);
-
-		error->ring[i].ringbuffer =
-			i915_error_object_create(dev_priv, ring->obj);
-
-		count = 0;
-		list_for_each_entry(request, &ring->request_list, list)
-			count++;
-
-		error->ring[i].num_requests = count;
-		error->ring[i].requests =
-			kmalloc(count*sizeof(struct drm_i915_error_request),
-				GFP_ATOMIC);
-		if (error->ring[i].requests == NULL) {
-			error->ring[i].num_requests = 0;
-			continue;
-		}
-
-		count = 0;
-		list_for_each_entry(request, &ring->request_list, list) {
-			struct drm_i915_error_request *erq;
-
-			erq = &error->ring[i].requests[count++];
-			erq->seqno = request->seqno;
-			erq->jiffies = request->emitted_jiffies;
-			erq->tail = request->tail;
-		}
-	}
-}
-
 /**
  * i915_capture_error_state - capture an error record for later analysis
  * @dev: drm device
@@ -1001,7 +893,7 @@ static void i915_capture_error_state(struct drm_device *dev)
 		return;
 
 	/* Account for pipe specific data like PIPE*STAT */
-	error = kzalloc(sizeof(*error), GFP_ATOMIC);
+	error = kmalloc(sizeof(*error), GFP_ATOMIC);
 	if (!error) {
 		DRM_DEBUG_DRIVER("out of memory, not capturing error state\n");
 		return;
@@ -1010,18 +902,59 @@ static void i915_capture_error_state(struct drm_device *dev)
 	DRM_INFO("capturing error event; look for more information in /debug/dri/%d/i915_error_state\n",
 		 dev->primary->index);
 
+	error->seqno = dev_priv->ring[RCS].get_seqno(&dev_priv->ring[RCS]);
 	error->eir = I915_READ(EIR);
 	error->pgtbl_er = I915_READ(PGTBL_ER);
 	for_each_pipe(pipe)
 		error->pipestat[pipe] = I915_READ(PIPESTAT(pipe));
-
+	error->instpm = I915_READ(INSTPM);
+	error->error = 0;
 	if (INTEL_INFO(dev)->gen >= 6) {
 		error->error = I915_READ(ERROR_GEN6);
-		error->done_reg = I915_READ(DONE_REG);
-	}
 
+		error->bcs_acthd = I915_READ(BCS_ACTHD);
+		error->bcs_ipehr = I915_READ(BCS_IPEHR);
+		error->bcs_ipeir = I915_READ(BCS_IPEIR);
+		error->bcs_instdone = I915_READ(BCS_INSTDONE);
+		error->bcs_seqno = 0;
+		if (dev_priv->ring[BCS].get_seqno)
+			error->bcs_seqno = dev_priv->ring[BCS].get_seqno(&dev_priv->ring[BCS]);
+
+		error->vcs_acthd = I915_READ(VCS_ACTHD);
+		error->vcs_ipehr = I915_READ(VCS_IPEHR);
+		error->vcs_ipeir = I915_READ(VCS_IPEIR);
+		error->vcs_instdone = I915_READ(VCS_INSTDONE);
+		error->vcs_seqno = 0;
+		if (dev_priv->ring[VCS].get_seqno)
+			error->vcs_seqno = dev_priv->ring[VCS].get_seqno(&dev_priv->ring[VCS]);
+	}
+	if (INTEL_INFO(dev)->gen >= 4) {
+		error->ipeir = I915_READ(IPEIR_I965);
+		error->ipehr = I915_READ(IPEHR_I965);
+		error->instdone = I915_READ(INSTDONE_I965);
+		error->instps = I915_READ(INSTPS);
+		error->instdone1 = I915_READ(INSTDONE1);
+		error->acthd = I915_READ(ACTHD_I965);
+		error->bbaddr = I915_READ64(BB_ADDR);
+	} else {
+		error->ipeir = I915_READ(IPEIR);
+		error->ipehr = I915_READ(IPEHR);
+		error->instdone = I915_READ(INSTDONE);
+		error->acthd = I915_READ(ACTHD);
+		error->bbaddr = 0;
+	}
 	i915_gem_record_fences(dev, error);
-	i915_gem_record_rings(dev, error);
+
+	/* Record the active batch and ring buffers */
+	for (i = 0; i < I915_NUM_RINGS; i++) {
+		error->batchbuffer[i] =
+			i915_error_first_batchbuffer(dev_priv,
+						     &dev_priv->ring[i]);
+
+		error->ringbuffer[i] =
+			i915_error_object_create(dev_priv,
+						 dev_priv->ring[i].obj);
+	}
 
 	/* Record buffers on the active and pinned lists. */
 	error->active_bo = NULL;
@@ -1077,12 +1010,11 @@ void i915_destroy_error_state(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_i915_error_state *error;
-	unsigned long flags;
 
-	spin_lock_irqsave(&dev_priv->error_lock, flags);
+	spin_lock(&dev_priv->error_lock);
 	error = dev_priv->first_error;
 	dev_priv->first_error = NULL;
-	spin_unlock_irqrestore(&dev_priv->error_lock, flags);
+	spin_unlock(&dev_priv->error_lock);
 
 	if (error)
 		i915_error_state_free(dev, error);
@@ -1266,7 +1198,7 @@ static void i915_pageflip_stall_check(struct drm_device *dev, int pipe)
 	} else {
 		int dspaddr = DSPADDR(intel_crtc->plane);
 		stall_detected = I915_READ(dspaddr) == (obj->gtt_offset +
-							crtc->y * crtc->fb->pitches[0] +
+							crtc->y * crtc->fb->pitch +
 							crtc->x * crtc->fb->bits_per_pixel/8);
 	}
 
@@ -1557,7 +1489,7 @@ static int ironlake_enable_vblank(struct drm_device *dev, int pipe)
 
 	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
 	ironlake_enable_display_irq(dev_priv, (pipe == 0) ?
-				    DE_PIPEA_VBLANK : DE_PIPEB_VBLANK);
+				    DE_PIPEA_VBLANK: DE_PIPEB_VBLANK);
 	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
 
 	return 0;
@@ -1605,7 +1537,7 @@ static void ironlake_disable_vblank(struct drm_device *dev, int pipe)
 
 	spin_lock_irqsave(&dev_priv->irq_lock, irqflags);
 	ironlake_disable_display_irq(dev_priv, (pipe == 0) ?
-				     DE_PIPEA_VBLANK : DE_PIPEB_VBLANK);
+				     DE_PIPEA_VBLANK: DE_PIPEB_VBLANK);
 	spin_unlock_irqrestore(&dev_priv->irq_lock, irqflags);
 }
 
@@ -1710,6 +1642,13 @@ static bool kick_ring(struct intel_ring_buffer *ring)
 		I915_WRITE_CTL(ring, tmp);
 		return true;
 	}
+	if (IS_GEN6(dev) &&
+	    (tmp & RING_WAIT_SEMAPHORE)) {
+		DRM_ERROR("Kicking stuck semaphore on %s\n",
+			  ring->name);
+		I915_WRITE_CTL(ring, tmp);
+		return true;
+	}
 	return false;
 }
 
@@ -1726,9 +1665,6 @@ void i915_hangcheck_elapsed(unsigned long data)
 	uint32_t acthd, instdone, instdone1, acthd_bsd, acthd_blt;
 	bool err = false;
 
-	if (!i915_enable_hangcheck)
-		return;
-
 	/* If all work is done then ACTHD clearly hasn't advanced. */
 	if (i915_hangcheck_ring_idle(&dev_priv->ring[RCS], &err) &&
 	    i915_hangcheck_ring_idle(&dev_priv->ring[VCS], &err) &&
@@ -1759,7 +1695,6 @@ void i915_hangcheck_elapsed(unsigned long data)
 	    dev_priv->last_instdone1 == instdone1) {
 		if (dev_priv->hangcheck_count++ > 1) {
 			DRM_ERROR("Hangcheck timer elapsed... GPU hung\n");
-			i915_handle_error(dev, true);
 
 			if (!IS_GEN2(dev)) {
 				/* Is the chip hanging on a WAIT_FOR_EVENT?
@@ -1767,6 +1702,7 @@ void i915_hangcheck_elapsed(unsigned long data)
 				 * and break the hang. This should work on
 				 * all but the second generation chipsets.
 				 */
+
 				if (kick_ring(&dev_priv->ring[RCS]))
 					goto repeat;
 
@@ -1779,6 +1715,7 @@ void i915_hangcheck_elapsed(unsigned long data)
 					goto repeat;
 			}
 
+			i915_handle_error(dev, true);
 			return;
 		}
 	} else {
@@ -1811,6 +1748,17 @@ static void ironlake_irq_preinstall(struct drm_device *dev)
 		INIT_WORK(&dev_priv->rps_work, gen6_pm_rps_work);
 
 	I915_WRITE(HWSTAM, 0xeffe);
+	if (IS_GEN6(dev) || IS_GEN7(dev)) {
+		/* Workaround stalls observed on Sandy Bridge GPUs by
+		 * making the blitter command streamer generate a
+		 * write to the Hardware Status Page for
+		 * MI_USER_INTERRUPT.  This appears to serialize the
+		 * previous seqno write out before the interrupt
+		 * happens.
+		 */
+		I915_WRITE(GEN6_BLITTER_HWSTAM, ~GEN6_BLITTER_USER_INTERRUPT);
+		I915_WRITE(GEN6_BSD_HWSTAM, ~GEN6_BSD_USER_INTERRUPT);
+	}
 
 	/* XXX hotplug from PCH */
 
@@ -1829,26 +1777,6 @@ static void ironlake_irq_preinstall(struct drm_device *dev)
 	POSTING_READ(SDEIER);
 }
 
-/*
- * Enable digital hotplug on the PCH, and configure the DP short pulse
- * duration to 2ms (which is the minimum in the Display Port spec)
- *
- * This register is the same on all known PCH chips.
- */
-
-static void ironlake_enable_pch_hotplug(struct drm_device *dev)
-{
-	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
-	u32	hotplug;
-
-	hotplug = I915_READ(PCH_PORT_HOTPLUG);
-	hotplug &= ~(PORTD_PULSE_DURATION_MASK|PORTC_PULSE_DURATION_MASK|PORTB_PULSE_DURATION_MASK);
-	hotplug |= PORTD_HOTPLUG_ENABLE | PORTD_PULSE_DURATION_2ms;
-	hotplug |= PORTC_HOTPLUG_ENABLE | PORTC_PULSE_DURATION_2ms;
-	hotplug |= PORTB_HOTPLUG_ENABLE | PORTB_PULSE_DURATION_2ms;
-	I915_WRITE(PCH_PORT_HOTPLUG, hotplug);
-}
-
 static int ironlake_irq_postinstall(struct drm_device *dev)
 {
 	drm_i915_private_t *dev_priv = (drm_i915_private_t *) dev->dev_private;
@@ -1911,8 +1839,6 @@ static int ironlake_irq_postinstall(struct drm_device *dev)
 	I915_WRITE(SDEIER, hotplug_mask);
 	POSTING_READ(SDEIER);
 
-	ironlake_enable_pch_hotplug(dev);
-
 	if (IS_IRONLAKE_M(dev)) {
 		/* Clear & enable PCU event interrupts */
 		I915_WRITE(DEIIR, DE_PCU_EVENT);
@@ -1970,8 +1896,6 @@ static int ivybridge_irq_postinstall(struct drm_device *dev)
 	I915_WRITE(SDEIER, hotplug_mask);
 	POSTING_READ(SDEIER);
 
-	ironlake_enable_pch_hotplug(dev);
-
 	return 0;
 }
 
@@ -2096,10 +2020,6 @@ static void ironlake_irq_uninstall(struct drm_device *dev)
 	I915_WRITE(GTIMR, 0xffffffff);
 	I915_WRITE(GTIER, 0x0);
 	I915_WRITE(GTIIR, I915_READ(GTIIR));
-
-	I915_WRITE(SDEIMR, 0xffffffff);
-	I915_WRITE(SDEIER, 0x0);
-	I915_WRITE(SDEIIR, I915_READ(SDEIIR));
 }
 
 static void i915_driver_irq_uninstall(struct drm_device * dev)
@@ -2138,10 +2058,8 @@ void intel_irq_init(struct drm_device *dev)
 		dev->driver->get_vblank_counter = gm45_get_vblank_counter;
 	}
 
-	if (drm_core_check_feature(dev, DRIVER_MODESET))
-		dev->driver->get_vblank_timestamp = i915_get_vblank_timestamp;
-	else
-		dev->driver->get_vblank_timestamp = NULL;
+
+	dev->driver->get_vblank_timestamp = i915_get_vblank_timestamp;
 	dev->driver->get_scanout_position = i915_get_crtc_scanoutpos;
 
 	if (IS_IVYBRIDGE(dev)) {
diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h
index 4528d3f..5dc3b6d 100644
--- a/drivers/gpu/drm/i915/i915_reg.h
+++ b/drivers/gpu/drm/i915/i915_reg.h
@@ -80,53 +80,12 @@
 #define  GRDOM_RENDER	(1<<2)
 #define  GRDOM_MEDIA	(3<<2)
 
-#define GEN6_MBCUNIT_SNPCR	0x900c /* for LLC config */
-#define   GEN6_MBC_SNPCR_SHIFT	21
-#define   GEN6_MBC_SNPCR_MASK	(3<<21)
-#define   GEN6_MBC_SNPCR_MAX	(0<<21)
-#define   GEN6_MBC_SNPCR_MED	(1<<21)
-#define   GEN6_MBC_SNPCR_LOW	(2<<21)
-#define   GEN6_MBC_SNPCR_MIN	(3<<21) /* only 1/16th of the cache is shared */
-
-#define GEN6_MBCTL		0x0907c
-#define   GEN6_MBCTL_ENABLE_BOOT_FETCH	(1 << 4)
-#define   GEN6_MBCTL_CTX_FETCH_NEEDED	(1 << 3)
-#define   GEN6_MBCTL_BME_UPDATE_ENABLE	(1 << 2)
-#define   GEN6_MBCTL_MAE_UPDATE_ENABLE	(1 << 1)
-#define   GEN6_MBCTL_BOOT_FETCH_MECH	(1 << 0)
-
 #define GEN6_GDRST	0x941c
 #define  GEN6_GRDOM_FULL		(1 << 0)
 #define  GEN6_GRDOM_RENDER		(1 << 1)
 #define  GEN6_GRDOM_MEDIA		(1 << 2)
 #define  GEN6_GRDOM_BLT			(1 << 3)
 
-/* PPGTT stuff */
-#define GEN6_GTT_ADDR_ENCODE(addr)	((addr) | (((addr) >> 28) & 0xff0))
-
-#define GEN6_PDE_VALID			(1 << 0)
-#define GEN6_PDE_LARGE_PAGE		(2 << 0) /* use 32kb pages */
-/* gen6+ has bit 11-4 for physical addr bit 39-32 */
-#define GEN6_PDE_ADDR_ENCODE(addr)	GEN6_GTT_ADDR_ENCODE(addr)
-
-#define GEN6_PTE_VALID			(1 << 0)
-#define GEN6_PTE_UNCACHED		(1 << 1)
-#define GEN6_PTE_CACHE_LLC		(2 << 1)
-#define GEN6_PTE_CACHE_LLC_MLC		(3 << 1)
-#define GEN6_PTE_CACHE_BITS		(3 << 1)
-#define GEN6_PTE_GFDT			(1 << 3)
-#define GEN6_PTE_ADDR_ENCODE(addr)	GEN6_GTT_ADDR_ENCODE(addr)
-
-#define RING_PP_DIR_BASE(ring)		((ring)->mmio_base+0x228)
-#define RING_PP_DIR_BASE_READ(ring)	((ring)->mmio_base+0x518)
-#define RING_PP_DIR_DCLV(ring)		((ring)->mmio_base+0x220)
-#define   PP_DIR_DCLV_2G		0xffffffff
-
-#define GAM_ECOCHK			0x4090
-#define   ECOCHK_SNB_BIT		(1<<10)
-#define   ECOCHK_PPGTT_CACHE64B		(0x3<<3)
-#define   ECOCHK_PPGTT_CACHE4B		(0x0<<3)
-
 /* VGA stuff */
 
 #define VGA_ST01_MDA 0x3ba
@@ -191,7 +150,7 @@
 #define MI_SUSPEND_FLUSH	MI_INSTR(0x0b, 0)
 #define   MI_SUSPEND_FLUSH_EN	(1<<0)
 #define MI_REPORT_HEAD		MI_INSTR(0x07, 0)
-#define MI_OVERLAY_FLIP		MI_INSTR(0x11, 0)
+#define MI_OVERLAY_FLIP		MI_INSTR(0x11,0)
 #define   MI_OVERLAY_CONTINUE	(0x0<<21)
 #define   MI_OVERLAY_ON		(0x1<<21)
 #define   MI_OVERLAY_OFF	(0x2<<21)
@@ -229,13 +188,6 @@
 #define  MI_SEMAPHORE_UPDATE	    (1<<21)
 #define  MI_SEMAPHORE_COMPARE	    (1<<20)
 #define  MI_SEMAPHORE_REGISTER	    (1<<18)
-#define  MI_SEMAPHORE_SYNC_RV	    (2<<16)
-#define  MI_SEMAPHORE_SYNC_RB	    (0<<16)
-#define  MI_SEMAPHORE_SYNC_VR	    (0<<16)
-#define  MI_SEMAPHORE_SYNC_VB	    (2<<16)
-#define  MI_SEMAPHORE_SYNC_BR	    (2<<16)
-#define  MI_SEMAPHORE_SYNC_BV	    (0<<16)
-#define  MI_SEMAPHORE_SYNC_INVALID  (1<<0)
 /*
  * 3D instructions used by the kernel
  */
@@ -277,22 +229,16 @@
 #define   ASYNC_FLIP                (1<<22)
 #define   DISPLAY_PLANE_A           (0<<20)
 #define   DISPLAY_PLANE_B           (1<<20)
-#define GFX_OP_PIPE_CONTROL(len)	((0x3<<29)|(0x3<<27)|(0x2<<24)|(len-2))
-#define   PIPE_CONTROL_CS_STALL				(1<<20)
-#define   PIPE_CONTROL_QW_WRITE				(1<<14)
-#define   PIPE_CONTROL_DEPTH_STALL			(1<<13)
-#define   PIPE_CONTROL_WRITE_FLUSH			(1<<12)
-#define   PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH	(1<<12) /* gen6+ */
-#define   PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE	(1<<11) /* MBZ on Ironlake */
-#define   PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE		(1<<10) /* GM45+ only */
-#define   PIPE_CONTROL_INDIRECT_STATE_DISABLE		(1<<9)
-#define   PIPE_CONTROL_NOTIFY				(1<<8)
-#define   PIPE_CONTROL_VF_CACHE_INVALIDATE		(1<<4)
-#define   PIPE_CONTROL_CONST_CACHE_INVALIDATE		(1<<3)
-#define   PIPE_CONTROL_STATE_CACHE_INVALIDATE		(1<<2)
-#define   PIPE_CONTROL_STALL_AT_SCOREBOARD		(1<<1)
-#define   PIPE_CONTROL_DEPTH_CACHE_FLUSH		(1<<0)
+#define GFX_OP_PIPE_CONTROL	((0x3<<29)|(0x3<<27)|(0x2<<24)|2)
+#define   PIPE_CONTROL_QW_WRITE	(1<<14)
+#define   PIPE_CONTROL_DEPTH_STALL (1<<13)
+#define   PIPE_CONTROL_WC_FLUSH	(1<<12)
+#define   PIPE_CONTROL_IS_FLUSH	(1<<11) /* MBZ on Ironlake */
+#define   PIPE_CONTROL_TC_FLUSH (1<<10) /* GM45+ only */
+#define   PIPE_CONTROL_ISP_DIS	(1<<9)
+#define   PIPE_CONTROL_NOTIFY	(1<<8)
 #define   PIPE_CONTROL_GLOBAL_GTT (1<<2) /* in addr dword */
+#define   PIPE_CONTROL_STALL_EN	(1<<1) /* in addr word, Ironlake+ only */
 
 
 /*
@@ -330,12 +276,6 @@
 #define FENCE_REG_SANDYBRIDGE_0		0x100000
 #define   SANDYBRIDGE_FENCE_PITCH_SHIFT	32
 
-/* control register for cpu gtt access */
-#define TILECTL				0x101000
-#define   TILECTL_SWZCTL			(1 << 0)
-#define   TILECTL_TLB_PREFETCH_DIS	(1 << 2)
-#define   TILECTL_BACKSNOOP_DIS		(1 << 3)
-
 /*
  * Instruction and interrupt control regs
  */
@@ -350,23 +290,10 @@
 #define RING_CTL(base)		((base)+0x3c)
 #define RING_SYNC_0(base)	((base)+0x40)
 #define RING_SYNC_1(base)	((base)+0x44)
-#define GEN6_RVSYNC (RING_SYNC_0(RENDER_RING_BASE))
-#define GEN6_RBSYNC (RING_SYNC_1(RENDER_RING_BASE))
-#define GEN6_VRSYNC (RING_SYNC_1(GEN6_BSD_RING_BASE))
-#define GEN6_VBSYNC (RING_SYNC_0(GEN6_BSD_RING_BASE))
-#define GEN6_BRSYNC (RING_SYNC_0(BLT_RING_BASE))
-#define GEN6_BVSYNC (RING_SYNC_1(BLT_RING_BASE))
 #define RING_MAX_IDLE(base)	((base)+0x54)
 #define RING_HWS_PGA(base)	((base)+0x80)
 #define RING_HWS_PGA_GEN6(base)	((base)+0x2080)
-#define ARB_MODE		0x04030
-#define   ARB_MODE_SWIZZLE_SNB	(1<<4)
-#define   ARB_MODE_SWIZZLE_IVB	(1<<5)
-#define   ARB_MODE_ENABLE(x)	GFX_MODE_ENABLE(x)
-#define   ARB_MODE_DISABLE(x)	GFX_MODE_DISABLE(x)
 #define RENDER_HWS_PGA_GEN7	(0x04080)
-#define RING_FAULT_REG(ring)	(0x4094 + 0x100*(ring)->id)
-#define DONE_REG		0x40b0
 #define BSD_HWS_PGA_GEN7	(0x04180)
 #define BLT_HWS_PGA_GEN7	(0x04280)
 #define RING_ACTHD(base)	((base)+0x74)
@@ -400,12 +327,6 @@
 #define IPEIR_I965	0x02064
 #define IPEHR_I965	0x02068
 #define INSTDONE_I965	0x0206c
-#define RING_IPEIR(base)	((base)+0x64)
-#define RING_IPEHR(base)	((base)+0x68)
-#define RING_INSTDONE(base)	((base)+0x6c)
-#define RING_INSTPS(base)	((base)+0x70)
-#define RING_DMA_FADD(base)	((base)+0x78)
-#define RING_INSTPM(base)	((base)+0xc0)
 #define INSTPS		0x02070 /* 965+ only */
 #define INSTDONE1	0x0207c /* 965+ only */
 #define ACTHD_I965	0x02074
@@ -419,6 +340,14 @@
 #define INSTDONE	0x02090
 #define NOPID		0x02094
 #define HWSTAM		0x02098
+#define VCS_INSTDONE	0x1206C
+#define VCS_IPEIR	0x12064
+#define VCS_IPEHR	0x12068
+#define VCS_ACTHD	0x12074
+#define BCS_INSTDONE	0x2206C
+#define BCS_IPEIR	0x22064
+#define BCS_IPEHR	0x22068
+#define BCS_ACTHD	0x22074
 
 #define ERROR_GEN6	0x040a0
 
@@ -438,11 +367,9 @@
 
 #define MI_MODE		0x0209c
 # define VS_TIMER_DISPATCH				(1 << 6)
-# define MI_FLUSH_ENABLE				(1 << 12)
+# define MI_FLUSH_ENABLE				(1 << 11)
 
 #define GFX_MODE	0x02520
-#define GFX_MODE_GEN7	0x0229c
-#define RING_MODE_GEN7(ring)	((ring)->mmio_base+0x29c)
 #define   GFX_RUN_LIST_ENABLE		(1<<15)
 #define   GFX_TLB_INVALIDATE_ALWAYS	(1<<13)
 #define   GFX_SURFACE_FAULT_ENABLE	(1<<12)
@@ -450,9 +377,6 @@
 #define   GFX_PSMI_GRANULARITY		(1<<10)
 #define   GFX_PPGTT_ENABLE		(1<<9)
 
-#define GFX_MODE_ENABLE(bit) (((bit) << 16) | (bit))
-#define GFX_MODE_DISABLE(bit) (((bit) << 16) | (0))
-
 #define SCPD0		0x0209c /* 915+ only */
 #define IER		0x020a0
 #define IIR		0x020a4
@@ -490,7 +414,6 @@
 #define   INSTPM_AGPBUSY_DIS (1<<11) /* gen3: when disabled, pending interrupts
 					will not assert AGPBUSY# and will only
 					be delivered when out of C3. */
-#define   INSTPM_FORCE_ORDERING				(1<<7) /* GEN6+ */
 #define ACTHD	        0x020c8
 #define FW_BLC		0x020d8
 #define FW_BLC2		0x020dc
@@ -538,7 +461,7 @@
 
 /* Enables non-sequential data reads through arbiter
  */
-#define   MI_ARB_DUAL_DATA_PHASE_DISABLE	(1 << 9)
+#define   MI_ARB_DUAL_DATA_PHASE_DISABLE       	(1 << 9)
 
 /* Disable FSB snooping of cacheable write cycles from binner/render
  * command stream
@@ -674,7 +597,6 @@
 #define   DPFC_CTL_PLANEA	(0<<30)
 #define   DPFC_CTL_PLANEB	(1<<30)
 #define   DPFC_CTL_FENCE_EN	(1<<29)
-#define   DPFC_CTL_PERSISTENT_MODE	(1<<25)
 #define   DPFC_SR_EN		(1<<10)
 #define   DPFC_CTL_LIMIT_1X	(0<<6)
 #define   DPFC_CTL_LIMIT_2X	(1<<6)
@@ -709,7 +631,7 @@
 
 #define ILK_DISPLAY_CHICKEN1	0x42000
 #define   ILK_FBCQ_DIS		(1<<22)
-#define	  ILK_PABSTRETCH_DIS	(1<<21)
+#define   ILK_PABSTRETCH_DIS 	(1<<21)
 
 
 /*
@@ -1100,29 +1022,6 @@
 #define C0DRB3			0x10206
 #define C1DRB3			0x10606
 
-/** snb MCH registers for reading the DRAM channel configuration */
-#define MAD_DIMM_C0			(MCHBAR_MIRROR_BASE_SNB + 0x5004)
-#define MAD_DIMM_C1			(MCHBAR_MIRROR_BASE_SNB + 0x5008)
-#define MAD_DIMM_C2			(MCHBAR_MIRROR_BASE_SNB + 0x500C)
-#define   MAD_DIMM_ECC_MASK		(0x3 << 24)
-#define   MAD_DIMM_ECC_OFF		(0x0 << 24)
-#define   MAD_DIMM_ECC_IO_ON_LOGIC_OFF	(0x1 << 24)
-#define   MAD_DIMM_ECC_IO_OFF_LOGIC_ON	(0x2 << 24)
-#define   MAD_DIMM_ECC_ON		(0x3 << 24)
-#define   MAD_DIMM_ENH_INTERLEAVE	(0x1 << 22)
-#define   MAD_DIMM_RANK_INTERLEAVE	(0x1 << 21)
-#define   MAD_DIMM_B_WIDTH_X16		(0x1 << 20) /* X8 chips if unset */
-#define   MAD_DIMM_A_WIDTH_X16		(0x1 << 19) /* X8 chips if unset */
-#define   MAD_DIMM_B_DUAL_RANK		(0x1 << 18)
-#define   MAD_DIMM_A_DUAL_RANK		(0x1 << 17)
-#define   MAD_DIMM_A_SELECT		(0x1 << 16)
-/* DIMM sizes are in multiples of 256mb. */
-#define   MAD_DIMM_B_SIZE_SHIFT		8
-#define   MAD_DIMM_B_SIZE_MASK		(0xff << MAD_DIMM_B_SIZE_SHIFT)
-#define   MAD_DIMM_A_SIZE_SHIFT		0
-#define   MAD_DIMM_A_SIZE_MASK		(0xff << MAD_DIMM_A_SIZE_SHIFT)
-
-
 /* Clocking configuration register */
 #define CLKCFG			0x10c00
 #define CLKCFG_FSB_400					(5 << 0)	/* hrawclk 100 */
@@ -1402,7 +1301,6 @@
 #define _VSYNC_A		0x60014
 #define _PIPEASRC	0x6001c
 #define _BCLRPAT_A	0x60020
-#define _VSYNCSHIFT_A	0x60028
 
 /* Pipe B timing regs */
 #define _HTOTAL_B	0x61000
@@ -1413,8 +1311,6 @@
 #define _VSYNC_B		0x61014
 #define _PIPEBSRC	0x6101c
 #define _BCLRPAT_B	0x61020
-#define _VSYNCSHIFT_B	0x61028
-
 
 #define HTOTAL(pipe) _PIPE(pipe, _HTOTAL_A, _HTOTAL_B)
 #define HBLANK(pipe) _PIPE(pipe, _HBLANK_A, _HBLANK_B)
@@ -1423,7 +1319,6 @@
 #define VBLANK(pipe) _PIPE(pipe, _VBLANK_A, _VBLANK_B)
 #define VSYNC(pipe) _PIPE(pipe, _VSYNC_A, _VSYNC_B)
 #define BCLRPAT(pipe) _PIPE(pipe, _BCLRPAT_A, _BCLRPAT_B)
-#define VSYNCSHIFT(pipe) _PIPE(pipe, _VSYNCSHIFT_A, _VSYNCSHIFT_B)
 
 /* VGA port control */
 #define ADPA			0x61100
@@ -1432,7 +1327,6 @@
 #define   ADPA_PIPE_SELECT_MASK	(1<<30)
 #define   ADPA_PIPE_A_SELECT	0
 #define   ADPA_PIPE_B_SELECT	(1<<30)
-#define   ADPA_PIPE_SELECT(pipe) ((pipe) << 30)
 #define   ADPA_USE_VGA_HVPOLARITY (1<<15)
 #define   ADPA_SETS_HVPOLARITY	0
 #define   ADPA_VSYNC_CNTL_DISABLE (1<<11)
@@ -1575,7 +1469,6 @@
 /* Selects pipe B for LVDS data.  Must be set on pre-965. */
 #define   LVDS_PIPEB_SELECT		(1 << 30)
 #define   LVDS_PIPE_MASK		(1 << 30)
-#define   LVDS_PIPE(pipe)		((pipe) << 30)
 /* LVDS dithering flag on 965/g4x platform */
 #define   LVDS_ENABLE_DITHER		(1 << 25)
 /* LVDS sync polarity flags. Set to invert (i.e. negative) */
@@ -1615,6 +1508,9 @@
 #define   LVDS_B0B3_POWER_DOWN		(0 << 2)
 #define   LVDS_B0B3_POWER_UP		(3 << 2)
 
+#define LVDS_PIPE_ENABLED(V, P) \
+	(((V) & (LVDS_PIPE_MASK | LVDS_PORT_EN)) == ((P) << 30 | LVDS_PORT_EN))
+
 /* Video Data Island Packet control */
 #define VIDEO_DIP_DATA		0x61178
 #define VIDEO_DIP_CTL		0x61170
@@ -1627,7 +1523,6 @@
 #define   VIDEO_DIP_SELECT_AVI		(0 << 19)
 #define   VIDEO_DIP_SELECT_VENDOR	(1 << 19)
 #define   VIDEO_DIP_SELECT_SPD		(3 << 19)
-#define   VIDEO_DIP_SELECT_MASK		(3 << 19)
 #define   VIDEO_DIP_FREQ_ONCE		(0 << 16)
 #define   VIDEO_DIP_FREQ_VSYNC		(1 << 16)
 #define   VIDEO_DIP_FREQ_2VSYNC		(2 << 16)
@@ -1644,21 +1539,12 @@
  */
 #define   PP_READY		(1 << 30)
 #define   PP_SEQUENCE_NONE	(0 << 28)
-#define   PP_SEQUENCE_POWER_UP	(1 << 28)
-#define   PP_SEQUENCE_POWER_DOWN (2 << 28)
-#define   PP_SEQUENCE_MASK	(3 << 28)
-#define   PP_SEQUENCE_SHIFT	28
+#define   PP_SEQUENCE_ON	(1 << 28)
+#define   PP_SEQUENCE_OFF	(2 << 28)
+#define   PP_SEQUENCE_MASK	0x30000000
 #define   PP_CYCLE_DELAY_ACTIVE	(1 << 27)
+#define   PP_SEQUENCE_STATE_ON_IDLE (1 << 3)
 #define   PP_SEQUENCE_STATE_MASK 0x0000000f
-#define   PP_SEQUENCE_STATE_OFF_IDLE	(0x0 << 0)
-#define   PP_SEQUENCE_STATE_OFF_S0_1	(0x1 << 0)
-#define   PP_SEQUENCE_STATE_OFF_S0_2	(0x2 << 0)
-#define   PP_SEQUENCE_STATE_OFF_S0_3	(0x3 << 0)
-#define   PP_SEQUENCE_STATE_ON_IDLE	(0x8 << 0)
-#define   PP_SEQUENCE_STATE_ON_S1_0	(0x9 << 0)
-#define   PP_SEQUENCE_STATE_ON_S1_2	(0xa << 0)
-#define   PP_SEQUENCE_STATE_ON_S1_3	(0xb << 0)
-#define   PP_SEQUENCE_STATE_RESET	(0xf << 0)
 #define PP_CONTROL	0x61204
 #define   POWER_TARGET_ON	(1 << 0)
 #define PP_ON_DELAYS	0x61208
@@ -2215,6 +2101,9 @@
 #define   DP_PIPEB_SELECT		(1 << 30)
 #define   DP_PIPE_MASK			(1 << 30)
 
+#define DP_PIPE_ENABLED(V, P) \
+	(((V) & (DP_PIPE_MASK | DP_PORT_EN)) == ((P) << 30 | DP_PORT_EN))
+
 /* Link training mode - select a suitable mode for each stage */
 #define   DP_LINK_TRAIN_PAT_1		(0 << 28)
 #define   DP_LINK_TRAIN_PAT_2		(1 << 28)
@@ -2410,21 +2299,9 @@
 #define   PIPECONF_PALETTE	0
 #define   PIPECONF_GAMMA		(1<<24)
 #define   PIPECONF_FORCE_BORDER	(1<<25)
-#define   PIPECONF_INTERLACE_MASK	(7 << 21)
-/* Note that pre-gen3 does not support interlaced display directly. Panel
- * fitting must be disabled on pre-ilk for interlaced. */
-#define   PIPECONF_PROGRESSIVE			(0 << 21)
-#define   PIPECONF_INTERLACE_W_SYNC_SHIFT_PANEL	(4 << 21) /* gen4 only */
-#define   PIPECONF_INTERLACE_W_SYNC_SHIFT	(5 << 21) /* gen4 only */
+#define   PIPECONF_PROGRESSIVE	(0 << 21)
 #define   PIPECONF_INTERLACE_W_FIELD_INDICATION	(6 << 21)
-#define   PIPECONF_INTERLACE_FIELD_0_ONLY	(7 << 21) /* gen3 only */
-/* Ironlake and later have a complete new set of values for interlaced. PFIT
- * means panel fitter required, PF means progressive fetch, DBL means power
- * saving pixel doubling. */
-#define   PIPECONF_PFIT_PF_INTERLACED_ILK	(1 << 21)
-#define   PIPECONF_INTERLACED_ILK		(3 << 21)
-#define   PIPECONF_INTERLACED_DBL_ILK		(4 << 21) /* ilk/snb only */
-#define   PIPECONF_PFIT_PF_INTERLACED_DBL_ILK	(5 << 21) /* ilk/snb only */
+#define   PIPECONF_INTERLACE_FIELD_0_ONLY		(7 << 21)
 #define   PIPECONF_CXSR_DOWNCLOCK	(1<<16)
 #define   PIPECONF_BPP_MASK	(0x000000e0)
 #define   PIPECONF_BPP_8	(0<<5)
@@ -2490,7 +2367,7 @@
 
 #define DSPFW1			0x70034
 #define   DSPFW_SR_SHIFT	23
-#define   DSPFW_SR_MASK		(0x1ff<<23)
+#define   DSPFW_SR_MASK 	(0x1ff<<23)
 #define   DSPFW_CURSORB_SHIFT	16
 #define   DSPFW_CURSORB_MASK	(0x3f<<16)
 #define   DSPFW_PLANEB_SHIFT	8
@@ -2548,7 +2425,6 @@
 #define  WM0_PIPE_CURSOR_MASK	(0x1f)
 
 #define WM0_PIPEB_ILK		0x45104
-#define WM0_PIPEC_IVB		0x45200
 #define WM1_LP_ILK		0x45108
 #define  WM1_LP_SR_EN		(1<<31)
 #define  WM1_LP_LATENCY_SHIFT	24
@@ -2563,8 +2439,6 @@
 #define WM3_LP_ILK		0x45110
 #define  WM3_LP_EN		(1<<31)
 #define WM1S_LP_ILK		0x45120
-#define WM2S_LP_IVB		0x45124
-#define WM3S_LP_IVB		0x45128
 #define  WM1S_LP_EN		(1<<31)
 
 /* Memory latency timer register */
@@ -2781,140 +2655,6 @@
 #define _DSPBSURF		0x7119C
 #define _DSPBTILEOFF		0x711A4
 
-/* Sprite A control */
-#define _DVSACNTR		0x72180
-#define   DVS_ENABLE		(1<<31)
-#define   DVS_GAMMA_ENABLE	(1<<30)
-#define   DVS_PIXFORMAT_MASK	(3<<25)
-#define   DVS_FORMAT_YUV422	(0<<25)
-#define   DVS_FORMAT_RGBX101010	(1<<25)
-#define   DVS_FORMAT_RGBX888	(2<<25)
-#define   DVS_FORMAT_RGBX161616	(3<<25)
-#define   DVS_SOURCE_KEY	(1<<22)
-#define   DVS_RGB_ORDER_RGBX	(1<<20)
-#define   DVS_YUV_BYTE_ORDER_MASK (3<<16)
-#define   DVS_YUV_ORDER_YUYV	(0<<16)
-#define   DVS_YUV_ORDER_UYVY	(1<<16)
-#define   DVS_YUV_ORDER_YVYU	(2<<16)
-#define   DVS_YUV_ORDER_VYUY	(3<<16)
-#define   DVS_DEST_KEY		(1<<2)
-#define   DVS_TRICKLE_FEED_DISABLE (1<<14)
-#define   DVS_TILED		(1<<10)
-#define _DVSALINOFF		0x72184
-#define _DVSASTRIDE		0x72188
-#define _DVSAPOS		0x7218c
-#define _DVSASIZE		0x72190
-#define _DVSAKEYVAL		0x72194
-#define _DVSAKEYMSK		0x72198
-#define _DVSASURF		0x7219c
-#define _DVSAKEYMAXVAL		0x721a0
-#define _DVSATILEOFF		0x721a4
-#define _DVSASURFLIVE		0x721ac
-#define _DVSASCALE		0x72204
-#define   DVS_SCALE_ENABLE	(1<<31)
-#define   DVS_FILTER_MASK	(3<<29)
-#define   DVS_FILTER_MEDIUM	(0<<29)
-#define   DVS_FILTER_ENHANCING	(1<<29)
-#define   DVS_FILTER_SOFTENING	(2<<29)
-#define   DVS_VERTICAL_OFFSET_HALF (1<<28) /* must be enabled below */
-#define   DVS_VERTICAL_OFFSET_ENABLE (1<<27)
-#define _DVSAGAMC		0x72300
-
-#define _DVSBCNTR		0x73180
-#define _DVSBLINOFF		0x73184
-#define _DVSBSTRIDE		0x73188
-#define _DVSBPOS		0x7318c
-#define _DVSBSIZE		0x73190
-#define _DVSBKEYVAL		0x73194
-#define _DVSBKEYMSK		0x73198
-#define _DVSBSURF		0x7319c
-#define _DVSBKEYMAXVAL		0x731a0
-#define _DVSBTILEOFF		0x731a4
-#define _DVSBSURFLIVE		0x731ac
-#define _DVSBSCALE		0x73204
-#define _DVSBGAMC		0x73300
-
-#define DVSCNTR(pipe) _PIPE(pipe, _DVSACNTR, _DVSBCNTR)
-#define DVSLINOFF(pipe) _PIPE(pipe, _DVSALINOFF, _DVSBLINOFF)
-#define DVSSTRIDE(pipe) _PIPE(pipe, _DVSASTRIDE, _DVSBSTRIDE)
-#define DVSPOS(pipe) _PIPE(pipe, _DVSAPOS, _DVSBPOS)
-#define DVSSURF(pipe) _PIPE(pipe, _DVSASURF, _DVSBSURF)
-#define DVSKEYMAX(pipe) _PIPE(pipe, _DVSAKEYMAXVAL, _DVSBKEYMAXVAL)
-#define DVSSIZE(pipe) _PIPE(pipe, _DVSASIZE, _DVSBSIZE)
-#define DVSSCALE(pipe) _PIPE(pipe, _DVSASCALE, _DVSBSCALE)
-#define DVSTILEOFF(pipe) _PIPE(pipe, _DVSATILEOFF, _DVSBTILEOFF)
-#define DVSKEYVAL(pipe) _PIPE(pipe, _DVSAKEYVAL, _DVSBKEYVAL)
-#define DVSKEYMSK(pipe) _PIPE(pipe, _DVSAKEYMSK, _DVSBKEYMSK)
-
-#define _SPRA_CTL		0x70280
-#define   SPRITE_ENABLE			(1<<31)
-#define   SPRITE_GAMMA_ENABLE		(1<<30)
-#define   SPRITE_PIXFORMAT_MASK		(7<<25)
-#define   SPRITE_FORMAT_YUV422		(0<<25)
-#define   SPRITE_FORMAT_RGBX101010	(1<<25)
-#define   SPRITE_FORMAT_RGBX888		(2<<25)
-#define   SPRITE_FORMAT_RGBX161616	(3<<25)
-#define   SPRITE_FORMAT_YUV444		(4<<25)
-#define   SPRITE_FORMAT_XR_BGR101010	(5<<25) /* Extended range */
-#define   SPRITE_CSC_ENABLE		(1<<24)
-#define   SPRITE_SOURCE_KEY		(1<<22)
-#define   SPRITE_RGB_ORDER_RGBX		(1<<20) /* only for 888 and 161616 */
-#define   SPRITE_YUV_TO_RGB_CSC_DISABLE	(1<<19)
-#define   SPRITE_YUV_CSC_FORMAT_BT709	(1<<18) /* 0 is BT601 */
-#define   SPRITE_YUV_BYTE_ORDER_MASK	(3<<16)
-#define   SPRITE_YUV_ORDER_YUYV		(0<<16)
-#define   SPRITE_YUV_ORDER_UYVY		(1<<16)
-#define   SPRITE_YUV_ORDER_YVYU		(2<<16)
-#define   SPRITE_YUV_ORDER_VYUY		(3<<16)
-#define   SPRITE_TRICKLE_FEED_DISABLE	(1<<14)
-#define   SPRITE_INT_GAMMA_ENABLE	(1<<13)
-#define   SPRITE_TILED			(1<<10)
-#define   SPRITE_DEST_KEY		(1<<2)
-#define _SPRA_LINOFF		0x70284
-#define _SPRA_STRIDE		0x70288
-#define _SPRA_POS		0x7028c
-#define _SPRA_SIZE		0x70290
-#define _SPRA_KEYVAL		0x70294
-#define _SPRA_KEYMSK		0x70298
-#define _SPRA_SURF		0x7029c
-#define _SPRA_KEYMAX		0x702a0
-#define _SPRA_TILEOFF		0x702a4
-#define _SPRA_SCALE		0x70304
-#define   SPRITE_SCALE_ENABLE	(1<<31)
-#define   SPRITE_FILTER_MASK	(3<<29)
-#define   SPRITE_FILTER_MEDIUM	(0<<29)
-#define   SPRITE_FILTER_ENHANCING	(1<<29)
-#define   SPRITE_FILTER_SOFTENING	(2<<29)
-#define   SPRITE_VERTICAL_OFFSET_HALF	(1<<28) /* must be enabled below */
-#define   SPRITE_VERTICAL_OFFSET_ENABLE	(1<<27)
-#define _SPRA_GAMC		0x70400
-
-#define _SPRB_CTL		0x71280
-#define _SPRB_LINOFF		0x71284
-#define _SPRB_STRIDE		0x71288
-#define _SPRB_POS		0x7128c
-#define _SPRB_SIZE		0x71290
-#define _SPRB_KEYVAL		0x71294
-#define _SPRB_KEYMSK		0x71298
-#define _SPRB_SURF		0x7129c
-#define _SPRB_KEYMAX		0x712a0
-#define _SPRB_TILEOFF		0x712a4
-#define _SPRB_SCALE		0x71304
-#define _SPRB_GAMC		0x71400
-
-#define SPRCTL(pipe) _PIPE(pipe, _SPRA_CTL, _SPRB_CTL)
-#define SPRLINOFF(pipe) _PIPE(pipe, _SPRA_LINOFF, _SPRB_LINOFF)
-#define SPRSTRIDE(pipe) _PIPE(pipe, _SPRA_STRIDE, _SPRB_STRIDE)
-#define SPRPOS(pipe) _PIPE(pipe, _SPRA_POS, _SPRB_POS)
-#define SPRSIZE(pipe) _PIPE(pipe, _SPRA_SIZE, _SPRB_SIZE)
-#define SPRKEYVAL(pipe) _PIPE(pipe, _SPRA_KEYVAL, _SPRB_KEYVAL)
-#define SPRKEYMSK(pipe) _PIPE(pipe, _SPRA_KEYMSK, _SPRB_KEYMSK)
-#define SPRSURF(pipe) _PIPE(pipe, _SPRA_SURF, _SPRB_SURF)
-#define SPRKEYMAX(pipe) _PIPE(pipe, _SPRA_KEYMAX, _SPRB_KEYMAX)
-#define SPRTILEOFF(pipe) _PIPE(pipe, _SPRA_TILEOFF, _SPRB_TILEOFF)
-#define SPRSCALE(pipe) _PIPE(pipe, _SPRA_SCALE, _SPRB_SCALE)
-#define SPRGAMC(pipe) _PIPE(pipe, _SPRA_GAMC, _SPRB_GAMC)
-
 /* VBIOS regs */
 #define VGACNTRL		0x71400
 # define VGA_DISP_DISABLE			(1 << 31)
@@ -3124,10 +2864,6 @@
 #define   ILK_DPFC_DIS1		(1<<8)
 #define   ILK_DPFC_DIS2		(1<<9)
 
-#define IVB_CHICKEN3	0x4200c
-# define CHICKEN3_DGMG_REQ_OUT_FIX_DISABLE	(1 << 5)
-# define CHICKEN3_DGMG_DONE_FIX_DISABLE		(1 << 2)
-
 #define DISP_ARB_CTL	0x45000
 #define  DISP_TILE_SURFACE_SWIZZLING	(1<<13)
 #define  DISP_FBC_WM_DIS		(1<<15)
@@ -3200,13 +2936,12 @@
 #define SDEIER  0xc400c
 
 /* digital port hotplug */
-#define PCH_PORT_HOTPLUG        0xc4030		/* SHOTPLUG_CTL */
+#define PCH_PORT_HOTPLUG        0xc4030
 #define PORTD_HOTPLUG_ENABLE            (1 << 20)
 #define PORTD_PULSE_DURATION_2ms        (0)
 #define PORTD_PULSE_DURATION_4_5ms      (1 << 18)
 #define PORTD_PULSE_DURATION_6ms        (2 << 18)
 #define PORTD_PULSE_DURATION_100ms      (3 << 18)
-#define PORTD_PULSE_DURATION_MASK	(3 << 18)
 #define PORTD_HOTPLUG_NO_DETECT         (0)
 #define PORTD_HOTPLUG_SHORT_DETECT      (1 << 16)
 #define PORTD_HOTPLUG_LONG_DETECT       (1 << 17)
@@ -3215,7 +2950,6 @@
 #define PORTC_PULSE_DURATION_4_5ms      (1 << 10)
 #define PORTC_PULSE_DURATION_6ms        (2 << 10)
 #define PORTC_PULSE_DURATION_100ms      (3 << 10)
-#define PORTC_PULSE_DURATION_MASK	(3 << 10)
 #define PORTC_HOTPLUG_NO_DETECT         (0)
 #define PORTC_HOTPLUG_SHORT_DETECT      (1 << 8)
 #define PORTC_HOTPLUG_LONG_DETECT       (1 << 9)
@@ -3224,7 +2958,6 @@
 #define PORTB_PULSE_DURATION_4_5ms      (1 << 2)
 #define PORTB_PULSE_DURATION_6ms        (2 << 2)
 #define PORTB_PULSE_DURATION_100ms      (3 << 2)
-#define PORTB_PULSE_DURATION_MASK	(3 << 2)
 #define PORTB_HOTPLUG_NO_DETECT         (0)
 #define PORTB_HOTPLUG_SHORT_DETECT      (1 << 0)
 #define PORTB_HOTPLUG_LONG_DETECT       (1 << 1)
@@ -3245,15 +2978,15 @@
 
 #define _PCH_DPLL_A              0xc6014
 #define _PCH_DPLL_B              0xc6018
-#define PCH_DPLL(pipe) (pipe == 0 ?  _PCH_DPLL_A : _PCH_DPLL_B)
+#define PCH_DPLL(pipe) _PIPE(pipe, _PCH_DPLL_A, _PCH_DPLL_B)
 
 #define _PCH_FPA0                0xc6040
 #define  FP_CB_TUNE		(0x3<<22)
 #define _PCH_FPA1                0xc6044
 #define _PCH_FPB0                0xc6048
 #define _PCH_FPB1                0xc604c
-#define PCH_FP0(pipe) (pipe == 0 ? _PCH_FPA0 : _PCH_FPB0)
-#define PCH_FP1(pipe) (pipe == 0 ? _PCH_FPA1 : _PCH_FPB1)
+#define PCH_FP0(pipe) _PIPE(pipe, _PCH_FPA0, _PCH_FPB0)
+#define PCH_FP1(pipe) _PIPE(pipe, _PCH_FPA1, _PCH_FPB1)
 
 #define PCH_DPLL_TEST           0xc606c
 
@@ -3323,7 +3056,6 @@
 #define _TRANS_VSYNC_A           0xe0014
 #define  TRANS_VSYNC_END_SHIFT  16
 #define  TRANS_VSYNC_START_SHIFT 0
-#define _TRANS_VSYNCSHIFT_A	0xe0028
 
 #define _TRANSA_DATA_M1          0xe0030
 #define _TRANSA_DATA_N1          0xe0034
@@ -3334,27 +3066,12 @@
 #define _TRANSA_DP_LINK_M2       0xe0048
 #define _TRANSA_DP_LINK_N2       0xe004c
 
-/* Per-transcoder DIP controls */
-
-#define _VIDEO_DIP_CTL_A         0xe0200
-#define _VIDEO_DIP_DATA_A        0xe0208
-#define _VIDEO_DIP_GCP_A         0xe0210
-
-#define _VIDEO_DIP_CTL_B         0xe1200
-#define _VIDEO_DIP_DATA_B        0xe1208
-#define _VIDEO_DIP_GCP_B         0xe1210
-
-#define TVIDEO_DIP_CTL(pipe) _PIPE(pipe, _VIDEO_DIP_CTL_A, _VIDEO_DIP_CTL_B)
-#define TVIDEO_DIP_DATA(pipe) _PIPE(pipe, _VIDEO_DIP_DATA_A, _VIDEO_DIP_DATA_B)
-#define TVIDEO_DIP_GCP(pipe) _PIPE(pipe, _VIDEO_DIP_GCP_A, _VIDEO_DIP_GCP_B)
-
 #define _TRANS_HTOTAL_B          0xe1000
 #define _TRANS_HBLANK_B          0xe1004
 #define _TRANS_HSYNC_B           0xe1008
 #define _TRANS_VTOTAL_B          0xe100c
 #define _TRANS_VBLANK_B          0xe1010
 #define _TRANS_VSYNC_B           0xe1014
-#define _TRANS_VSYNCSHIFT_B	 0xe1028
 
 #define TRANS_HTOTAL(pipe) _PIPE(pipe, _TRANS_HTOTAL_A, _TRANS_HTOTAL_B)
 #define TRANS_HBLANK(pipe) _PIPE(pipe, _TRANS_HBLANK_A, _TRANS_HBLANK_B)
@@ -3362,8 +3079,6 @@
 #define TRANS_VTOTAL(pipe) _PIPE(pipe, _TRANS_VTOTAL_A, _TRANS_VTOTAL_B)
 #define TRANS_VBLANK(pipe) _PIPE(pipe, _TRANS_VBLANK_A, _TRANS_VBLANK_B)
 #define TRANS_VSYNC(pipe) _PIPE(pipe, _TRANS_VSYNC_A, _TRANS_VSYNC_B)
-#define TRANS_VSYNCSHIFT(pipe) _PIPE(pipe, _TRANS_VSYNCSHIFT_A, \
-				     _TRANS_VSYNCSHIFT_B)
 
 #define _TRANSB_DATA_M1          0xe1030
 #define _TRANSB_DATA_N1          0xe1034
@@ -3397,10 +3112,7 @@
 #define  TRANS_FSYNC_DELAY_HB4  (3<<27)
 #define  TRANS_DP_AUDIO_ONLY    (1<<26)
 #define  TRANS_DP_VIDEO_AUDIO   (0<<26)
-#define  TRANS_INTERLACE_MASK   (7<<21)
 #define  TRANS_PROGRESSIVE      (0<<21)
-#define  TRANS_INTERLACED       (3<<21)
-#define  TRANS_LEGACY_INTERLACED_ILK (2<<21)
 #define  TRANS_8BPC             (0<<5)
 #define  TRANS_10BPC            (1<<5)
 #define  TRANS_6BPC             (2<<5)
@@ -3562,13 +3274,15 @@
 #define  ADPA_CRT_HOTPLUG_VOLREF_475MV  (1<<17)
 #define  ADPA_CRT_HOTPLUG_FORCE_TRIGGER (1<<16)
 
+#define ADPA_PIPE_ENABLED(V, P) \
+	(((V) & (ADPA_TRANS_SELECT_MASK | ADPA_DAC_ENABLE)) == ((P) << 30 | ADPA_DAC_ENABLE))
+
 /* or SDVOB */
 #define HDMIB   0xe1140
 #define  PORT_ENABLE    (1 << 31)
-#define  TRANSCODER(pipe)       ((pipe) << 30)
-#define  TRANSCODER_CPT(pipe)   ((pipe) << 29)
-#define  TRANSCODER_MASK        (1 << 30)
-#define  TRANSCODER_MASK_CPT    (3 << 29)
+#define  TRANSCODER_A   (0)
+#define  TRANSCODER_B   (1 << 30)
+#define  TRANSCODER_MASK   (1 << 30)
 #define  COLOR_FORMAT_8bpc      (0)
 #define  COLOR_FORMAT_12bpc     (3 << 26)
 #define  SDVOB_HOTPLUG_ENABLE   (1 << 23)
@@ -3584,6 +3298,9 @@
 #define  HSYNC_ACTIVE_HIGH      (1 << 3)
 #define  PORT_DETECTED          (1 << 2)
 
+#define HDMI_PIPE_ENABLED(V, P) \
+	(((V) & (TRANSCODER_MASK | PORT_ENABLE)) == ((P) << 30 | PORT_ENABLE))
+
 /* PCH SDVOB multiplex with HDMIB */
 #define PCH_SDVOB	HDMIB
 
@@ -3611,35 +3328,15 @@
 #define PCH_PP_STATUS		0xc7200
 #define PCH_PP_CONTROL		0xc7204
 #define  PANEL_UNLOCK_REGS	(0xabcd << 16)
-#define  PANEL_UNLOCK_MASK	(0xffff << 16)
 #define  EDP_FORCE_VDD		(1 << 3)
 #define  EDP_BLC_ENABLE		(1 << 2)
 #define  PANEL_POWER_RESET	(1 << 1)
 #define  PANEL_POWER_OFF	(0 << 0)
 #define  PANEL_POWER_ON		(1 << 0)
 #define PCH_PP_ON_DELAYS	0xc7208
-#define  PANEL_PORT_SELECT_MASK	(3 << 30)
-#define  PANEL_PORT_SELECT_LVDS	(0 << 30)
-#define  PANEL_PORT_SELECT_DPA	(1 << 30)
 #define  EDP_PANEL		(1 << 30)
-#define  PANEL_PORT_SELECT_DPC	(2 << 30)
-#define  PANEL_PORT_SELECT_DPD	(3 << 30)
-#define  PANEL_POWER_UP_DELAY_MASK	(0x1fff0000)
-#define  PANEL_POWER_UP_DELAY_SHIFT	16
-#define  PANEL_LIGHT_ON_DELAY_MASK	(0x1fff)
-#define  PANEL_LIGHT_ON_DELAY_SHIFT	0
-
 #define PCH_PP_OFF_DELAYS	0xc720c
-#define  PANEL_POWER_DOWN_DELAY_MASK	(0x1fff0000)
-#define  PANEL_POWER_DOWN_DELAY_SHIFT	16
-#define  PANEL_LIGHT_OFF_DELAY_MASK	(0x1fff)
-#define  PANEL_LIGHT_OFF_DELAY_SHIFT	0
-
 #define PCH_PP_DIVISOR		0xc7210
-#define  PP_REFERENCE_DIVIDER_MASK	(0xffffff00)
-#define  PP_REFERENCE_DIVIDER_SHIFT	8
-#define  PANEL_POWER_CYCLE_DELAY_MASK	(0x1f)
-#define  PANEL_POWER_CYCLE_DELAY_SHIFT	0
 
 #define PCH_DP_B		0xe4100
 #define PCH_DPB_AUX_CH_CTL	0xe4110
@@ -3670,7 +3367,6 @@
 #define  PORT_TRANS_B_SEL_CPT	(1<<29)
 #define  PORT_TRANS_C_SEL_CPT	(2<<29)
 #define  PORT_TRANS_SEL_MASK	(3<<29)
-#define  PORT_TRANS_SEL_CPT(pipe)	((pipe) << 29)
 
 #define TRANS_DP_CTL_A		0xe0300
 #define TRANS_DP_CTL_B		0xe1300
@@ -3709,39 +3405,10 @@
 #define  EDP_LINK_TRAIN_800_1200MV_0DB_SNB_B	(0x38<<22)
 #define  EDP_LINK_TRAIN_VOL_EMP_MASK_SNB	(0x3f<<22)
 
-/* IVB */
-#define EDP_LINK_TRAIN_400MV_0DB_IVB		(0x24 <<22)
-#define EDP_LINK_TRAIN_400MV_3_5DB_IVB		(0x2a <<22)
-#define EDP_LINK_TRAIN_400MV_6DB_IVB		(0x2f <<22)
-#define EDP_LINK_TRAIN_600MV_0DB_IVB		(0x30 <<22)
-#define EDP_LINK_TRAIN_600MV_3_5DB_IVB		(0x36 <<22)
-#define EDP_LINK_TRAIN_800MV_0DB_IVB		(0x38 <<22)
-#define EDP_LINK_TRAIN_800MV_3_5DB_IVB		(0x33 <<22)
-
-/* legacy values */
-#define EDP_LINK_TRAIN_500MV_0DB_IVB		(0x00 <<22)
-#define EDP_LINK_TRAIN_1000MV_0DB_IVB		(0x20 <<22)
-#define EDP_LINK_TRAIN_500MV_3_5DB_IVB		(0x02 <<22)
-#define EDP_LINK_TRAIN_1000MV_3_5DB_IVB		(0x22 <<22)
-#define EDP_LINK_TRAIN_1000MV_6DB_IVB		(0x23 <<22)
-
-#define  EDP_LINK_TRAIN_VOL_EMP_MASK_IVB	(0x3f<<22)
-
 #define  FORCEWAKE				0xA18C
 #define  FORCEWAKE_ACK				0x130090
-#define  FORCEWAKE_MT				0xa188 /* multi-threaded */
-#define  FORCEWAKE_MT_ACK			0x130040
-#define  ECOBUS					0xa180
-#define    FORCEWAKE_MT_ENABLE			(1<<5)
-
-#define  GTFIFODBG				0x120000
-#define    GT_FIFO_CPU_ERROR_MASK		7
-#define    GT_FIFO_OVFERR			(1<<2)
-#define    GT_FIFO_IAWRERR			(1<<1)
-#define    GT_FIFO_IARDERR			(1<<0)
 
 #define  GT_FIFO_FREE_ENTRIES			0x120008
-#define    GT_FIFO_NUM_RESERVED_ENTRIES		20
 
 #define GEN6_UCGCTL2				0x9404
 # define GEN6_RCZUNIT_CLOCK_GATE_DISABLE		(1 << 13)
@@ -3769,11 +3436,7 @@
 #define   GEN6_CAGF_MASK			(0x7f << GEN6_CAGF_SHIFT)
 #define GEN6_RP_CONTROL				0xA024
 #define   GEN6_RP_MEDIA_TURBO			(1<<11)
-#define   GEN6_RP_MEDIA_MODE_MASK		(3<<9)
-#define   GEN6_RP_MEDIA_HW_TURBO_MODE		(3<<9)
-#define   GEN6_RP_MEDIA_HW_NORMAL_MODE		(2<<9)
-#define   GEN6_RP_MEDIA_HW_MODE			(1<<9)
-#define   GEN6_RP_MEDIA_SW_MODE			(0<<9)
+#define   GEN6_RP_USE_NORMAL_FREQ		(1<<9)
 #define   GEN6_RP_MEDIA_IS_GFX			(1<<8)
 #define   GEN6_RP_ENABLE			(1<<7)
 #define   GEN6_RP_UP_IDLE_MIN			(0x1<<3)
@@ -3825,60 +3488,7 @@
 #define GEN6_PCODE_MAILBOX			0x138124
 #define   GEN6_PCODE_READY			(1<<31)
 #define   GEN6_READ_OC_PARAMS			0xc
-#define   GEN6_PCODE_WRITE_MIN_FREQ_TABLE	0x8
-#define   GEN6_PCODE_READ_MIN_FREQ_TABLE	0x9
+#define   GEN6_PCODE_WRITE_MIN_FREQ_TABLE	0x9
 #define GEN6_PCODE_DATA				0x138128
-#define   GEN6_PCODE_FREQ_IA_RATIO_SHIFT	8
-
-#define GEN6_GT_CORE_STATUS		0x138060
-#define   GEN6_CORE_CPD_STATE_MASK	(7<<4)
-#define   GEN6_RCn_MASK			7
-#define   GEN6_RC0			0
-#define   GEN6_RC3			2
-#define   GEN6_RC6			3
-#define   GEN6_RC7			4
-
-#define G4X_AUD_VID_DID			0x62020
-#define INTEL_AUDIO_DEVCL		0x808629FB
-#define INTEL_AUDIO_DEVBLC		0x80862801
-#define INTEL_AUDIO_DEVCTG		0x80862802
-
-#define G4X_AUD_CNTL_ST			0x620B4
-#define G4X_ELDV_DEVCL_DEVBLC		(1 << 13)
-#define G4X_ELDV_DEVCTG			(1 << 14)
-#define G4X_ELD_ADDR			(0xf << 5)
-#define G4X_ELD_ACK			(1 << 4)
-#define G4X_HDMIW_HDMIEDID		0x6210C
-
-#define IBX_HDMIW_HDMIEDID_A		0xE2050
-#define IBX_AUD_CNTL_ST_A		0xE20B4
-#define IBX_ELD_BUFFER_SIZE		(0x1f << 10)
-#define IBX_ELD_ADDRESS			(0x1f << 5)
-#define IBX_ELD_ACK			(1 << 4)
-#define IBX_AUD_CNTL_ST2		0xE20C0
-#define IBX_ELD_VALIDB			(1 << 0)
-#define IBX_CP_READYB			(1 << 1)
-
-#define CPT_HDMIW_HDMIEDID_A		0xE5050
-#define CPT_AUD_CNTL_ST_A		0xE50B4
-#define CPT_AUD_CNTRL_ST2		0xE50C0
-
-/* These are the 4 32-bit write offset registers for each stream
- * output buffer.  It determines the offset from the
- * 3DSTATE_SO_BUFFERs that the next streamed vertex output goes to.
- */
-#define GEN7_SO_WRITE_OFFSET(n)		(0x5280 + (n) * 4)
-
-#define IBX_AUD_CONFIG_A			0xe2000
-#define CPT_AUD_CONFIG_A			0xe5000
-#define   AUD_CONFIG_N_VALUE_INDEX		(1 << 29)
-#define   AUD_CONFIG_N_PROG_ENABLE		(1 << 28)
-#define   AUD_CONFIG_UPPER_N_SHIFT		20
-#define   AUD_CONFIG_UPPER_N_VALUE		(0xff << 20)
-#define   AUD_CONFIG_LOWER_N_SHIFT		4
-#define   AUD_CONFIG_LOWER_N_VALUE		(0xfff << 4)
-#define   AUD_CONFIG_PIXEL_CLOCK_HDMI_SHIFT	16
-#define   AUD_CONFIG_PIXEL_CLOCK_HDMI		(0xf << 16)
-#define   AUD_CONFIG_DISABLE_NCTS		(1 << 3)
 
 #endif /* _I915_REG_H_ */
diff --git a/drivers/gpu/drm/i915/i915_suspend.c b/drivers/gpu/drm/i915/i915_suspend.c
index 0d13778..5ad0b51 100644
--- a/drivers/gpu/drm/i915/i915_suspend.c
+++ b/drivers/gpu/drm/i915/i915_suspend.c
@@ -28,7 +28,6 @@
 #include "drm.h"
 #include "i915_drm.h"
 #include "intel_drv.h"
-#include "i915_reg.h"
 
 static bool i915_pipe_enabled(struct drm_device *dev, enum pipe pipe)
 {
@@ -40,7 +39,7 @@ static bool i915_pipe_enabled(struct drm_device *dev, enum pipe pipe)
 		return false;
 
 	if (HAS_PCH_SPLIT(dev))
-		dpll_reg = PCH_DPLL(pipe);
+		dpll_reg = (pipe == PIPE_A) ? _PCH_DPLL_A : _PCH_DPLL_B;
 	else
 		dpll_reg = (pipe == PIPE_A) ? _DPLL_A : _DPLL_B;
 
@@ -65,7 +64,7 @@ static void i915_save_palette(struct drm_device *dev, enum pipe pipe)
 	else
 		array = dev_priv->save_palette_b;
 
-	for (i = 0; i < 256; i++)
+	for(i = 0; i < 256; i++)
 		array[i] = I915_READ(reg + (i << 2));
 }
 
@@ -87,7 +86,7 @@ static void i915_restore_palette(struct drm_device *dev, enum pipe pipe)
 	else
 		array = dev_priv->save_palette_b;
 
-	for (i = 0; i < 256; i++)
+	for(i = 0; i < 256; i++)
 		I915_WRITE(reg + (i << 2), array[i]);
 }
 
@@ -770,13 +769,15 @@ static void i915_restore_display(struct drm_device *dev)
 	/* FIXME: restore TV & SDVO state */
 
 	/* only restore FBC info on the platform that supports FBC*/
-	intel_disable_fbc(dev);
 	if (I915_HAS_FBC(dev)) {
 		if (HAS_PCH_SPLIT(dev)) {
+			ironlake_disable_fbc(dev);
 			I915_WRITE(ILK_DPFC_CB_BASE, dev_priv->saveDPFC_CB_BASE);
 		} else if (IS_GM45(dev)) {
+			g4x_disable_fbc(dev);
 			I915_WRITE(DPFC_CB_BASE, dev_priv->saveDPFC_CB_BASE);
 		} else {
+			i8xx_disable_fbc(dev);
 			I915_WRITE(FBC_CFB_BASE, dev_priv->saveFBC_CFB_BASE);
 			I915_WRITE(FBC_LL_BASE, dev_priv->saveFBC_LL_BASE);
 			I915_WRITE(FBC_CONTROL2, dev_priv->saveFBC_CONTROL2);
@@ -830,7 +831,7 @@ int i915_save_state(struct drm_device *dev)
 
 	if (IS_IRONLAKE_M(dev))
 		ironlake_disable_drps(dev);
-	if (INTEL_INFO(dev)->gen >= 6)
+	if (IS_GEN6(dev))
 		gen6_disable_rps(dev);
 
 	/* Cache mode state */
@@ -881,26 +882,23 @@ int i915_restore_state(struct drm_device *dev)
 	}
 	mutex_unlock(&dev->struct_mutex);
 
-	if (drm_core_check_feature(dev, DRIVER_MODESET))
-		intel_init_clock_gating(dev);
+	intel_init_clock_gating(dev);
 
 	if (IS_IRONLAKE_M(dev)) {
 		ironlake_enable_drps(dev);
 		intel_init_emon(dev);
 	}
 
-	if (INTEL_INFO(dev)->gen >= 6) {
+	if (IS_GEN6(dev))
 		gen6_enable_rps(dev_priv);
-		gen6_update_ring_freq(dev_priv);
-	}
 
 	mutex_lock(&dev->struct_mutex);
 
 	/* Cache mode state */
-	I915_WRITE(CACHE_MODE_0, dev_priv->saveCACHE_MODE_0 | 0xffff0000);
+	I915_WRITE (CACHE_MODE_0, dev_priv->saveCACHE_MODE_0 | 0xffff0000);
 
 	/* Memory arbitration state */
-	I915_WRITE(MI_ARB_STATE, dev_priv->saveMI_ARB_STATE | 0xffff0000);
+	I915_WRITE (MI_ARB_STATE, dev_priv->saveMI_ARB_STATE | 0xffff0000);
 
 	for (i = 0; i < 16; i++) {
 		I915_WRITE(SWF00 + (i << 2), dev_priv->saveSWF0[i]);
diff --git a/drivers/gpu/drm/i915/i915_trace.h b/drivers/gpu/drm/i915/i915_trace.h
index dac7bba..d623fef 100644
--- a/drivers/gpu/drm/i915/i915_trace.h
+++ b/drivers/gpu/drm/i915/i915_trace.h
@@ -385,29 +385,29 @@ TRACE_EVENT(i915_flip_complete,
 );
 
 TRACE_EVENT(i915_reg_rw,
-	TP_PROTO(bool write, u32 reg, u64 val, int len),
-
-	TP_ARGS(write, reg, val, len),
-
-	TP_STRUCT__entry(
-		__field(u64, val)
-		__field(u32, reg)
-		__field(u16, write)
-		__field(u16, len)
-		),
-
-	TP_fast_assign(
-		__entry->val = (u64)val;
-		__entry->reg = reg;
-		__entry->write = write;
-		__entry->len = len;
-		),
-
-	TP_printk("%s reg=0x%x, len=%d, val=(0x%x, 0x%x)",
-		__entry->write ? "write" : "read",
-		__entry->reg, __entry->len,
-		(u32)(__entry->val & 0xffffffff),
-		(u32)(__entry->val >> 32))
+           TP_PROTO(bool write, u32 reg, u64 val, int len),
+
+           TP_ARGS(write, reg, val, len),
+
+           TP_STRUCT__entry(
+                   __field(u64, val)
+                   __field(u32, reg)
+                   __field(u16, write)
+                   __field(u16, len)
+                   ),
+
+           TP_fast_assign(
+                   __entry->val = (u64)val;
+                   __entry->reg = reg;
+                   __entry->write = write;
+                   __entry->len = len;
+                   ),
+
+           TP_printk("%s reg=0x%x, len=%d, val=(0x%x, 0x%x)",
+                     __entry->write ? "write" : "read",
+		     __entry->reg, __entry->len,
+		     (u32)(__entry->val & 0xffffffff),
+		     (u32)(__entry->val >> 32))
 );
 
 #endif /* _I915_TRACE_H_ */
diff --git a/drivers/gpu/drm/i915/intel_acpi.c b/drivers/gpu/drm/i915/intel_acpi.c
index bae3edf..2cb8e0b 100644
--- a/drivers/gpu/drm/i915/intel_acpi.c
+++ b/drivers/gpu/drm/i915/intel_acpi.c
@@ -64,7 +64,7 @@ static int intel_dsm(acpi_handle handle, int func, int arg)
 
 	case ACPI_TYPE_BUFFER:
 		if (obj->buffer.length == 4) {
-			result = (obj->buffer.pointer[0] |
+			result =(obj->buffer.pointer[0] |
 				(obj->buffer.pointer[1] <<  8) |
 				(obj->buffer.pointer[2] << 16) |
 				(obj->buffer.pointer[3] << 24));
@@ -208,7 +208,7 @@ static bool intel_dsm_pci_probe(struct pci_dev *pdev)
 
 	ret = intel_dsm(dhandle, INTEL_DSM_FN_SUPPORTED_FUNCTIONS, 0);
 	if (ret < 0) {
-		DRM_DEBUG_KMS("failed to get supported _DSM functions\n");
+		DRM_ERROR("failed to get supported _DSM functions\n");
 		return false;
 	}
 
diff --git a/drivers/gpu/drm/i915/intel_bios.c b/drivers/gpu/drm/i915/intel_bios.c
index b48fc2a..e5fa074 100644
--- a/drivers/gpu/drm/i915/intel_bios.c
+++ b/drivers/gpu/drm/i915/intel_bios.c
@@ -1,5 +1,5 @@
 /*
- * Copyright © 2006 Intel Corporation
+ * Copyright © 2006 Intel Corporation
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -75,7 +75,7 @@ get_blocksize(void *p)
 
 static void
 fill_detail_timing_data(struct drm_display_mode *panel_fixed_mode,
-			const struct lvds_dvo_timing *dvo_timing)
+			struct lvds_dvo_timing *dvo_timing)
 {
 	panel_fixed_mode->hdisplay = (dvo_timing->hactive_hi << 8) |
 		dvo_timing->hactive_lo;
@@ -116,75 +116,20 @@ fill_detail_timing_data(struct drm_display_mode *panel_fixed_mode,
 	drm_mode_set_name(panel_fixed_mode);
 }
 
-static bool
-lvds_dvo_timing_equal_size(const struct lvds_dvo_timing *a,
-			   const struct lvds_dvo_timing *b)
-{
-	if (a->hactive_hi != b->hactive_hi ||
-	    a->hactive_lo != b->hactive_lo)
-		return false;
-
-	if (a->hsync_off_hi != b->hsync_off_hi ||
-	    a->hsync_off_lo != b->hsync_off_lo)
-		return false;
-
-	if (a->hsync_pulse_width != b->hsync_pulse_width)
-		return false;
-
-	if (a->hblank_hi != b->hblank_hi ||
-	    a->hblank_lo != b->hblank_lo)
-		return false;
-
-	if (a->vactive_hi != b->vactive_hi ||
-	    a->vactive_lo != b->vactive_lo)
-		return false;
-
-	if (a->vsync_off != b->vsync_off)
-		return false;
-
-	if (a->vsync_pulse_width != b->vsync_pulse_width)
-		return false;
-
-	if (a->vblank_hi != b->vblank_hi ||
-	    a->vblank_lo != b->vblank_lo)
-		return false;
-
-	return true;
-}
-
-static const struct lvds_dvo_timing *
-get_lvds_dvo_timing(const struct bdb_lvds_lfp_data *lvds_lfp_data,
-		    const struct bdb_lvds_lfp_data_ptrs *lvds_lfp_data_ptrs,
-		    int index)
-{
-	/*
-	 * the size of fp_timing varies on the different platform.
-	 * So calculate the DVO timing relative offset in LVDS data
-	 * entry to get the DVO timing entry
-	 */
-
-	int lfp_data_size =
-		lvds_lfp_data_ptrs->ptr[1].dvo_timing_offset -
-		lvds_lfp_data_ptrs->ptr[0].dvo_timing_offset;
-	int dvo_timing_offset =
-		lvds_lfp_data_ptrs->ptr[0].dvo_timing_offset -
-		lvds_lfp_data_ptrs->ptr[0].fp_timing_offset;
-	char *entry = (char *)lvds_lfp_data->data + lfp_data_size * index;
-
-	return (struct lvds_dvo_timing *)(entry + dvo_timing_offset);
-}
-
 /* Try to find integrated panel data */
 static void
 parse_lfp_panel_data(struct drm_i915_private *dev_priv,
 			    struct bdb_header *bdb)
 {
-	const struct bdb_lvds_options *lvds_options;
-	const struct bdb_lvds_lfp_data *lvds_lfp_data;
-	const struct bdb_lvds_lfp_data_ptrs *lvds_lfp_data_ptrs;
-	const struct lvds_dvo_timing *panel_dvo_timing;
+	struct bdb_lvds_options *lvds_options;
+	struct bdb_lvds_lfp_data *lvds_lfp_data;
+	struct bdb_lvds_lfp_data_ptrs *lvds_lfp_data_ptrs;
+	struct bdb_lvds_lfp_data_entry *entry;
+	struct lvds_dvo_timing *dvo_timing;
 	struct drm_display_mode *panel_fixed_mode;
-	int i, downclock;
+	int lfp_data_size, dvo_timing_offset;
+	int i, temp_downclock;
+	struct drm_display_mode *temp_mode;
 
 	lvds_options = find_section(bdb, BDB_LVDS_OPTIONS);
 	if (!lvds_options)
@@ -206,44 +151,75 @@ parse_lfp_panel_data(struct drm_i915_private *dev_priv,
 
 	dev_priv->lvds_vbt = 1;
 
-	panel_dvo_timing = get_lvds_dvo_timing(lvds_lfp_data,
-					       lvds_lfp_data_ptrs,
-					       lvds_options->panel_type);
+	lfp_data_size = lvds_lfp_data_ptrs->ptr[1].dvo_timing_offset -
+		lvds_lfp_data_ptrs->ptr[0].dvo_timing_offset;
+	entry = (struct bdb_lvds_lfp_data_entry *)
+		((uint8_t *)lvds_lfp_data->data + (lfp_data_size *
+						   lvds_options->panel_type));
+	dvo_timing_offset = lvds_lfp_data_ptrs->ptr[0].dvo_timing_offset -
+		lvds_lfp_data_ptrs->ptr[0].fp_timing_offset;
+
+	/*
+	 * the size of fp_timing varies on the different platform.
+	 * So calculate the DVO timing relative offset in LVDS data
+	 * entry to get the DVO timing entry
+	 */
+	dvo_timing = (struct lvds_dvo_timing *)
+			((unsigned char *)entry + dvo_timing_offset);
 
 	panel_fixed_mode = kzalloc(sizeof(*panel_fixed_mode), GFP_KERNEL);
 	if (!panel_fixed_mode)
 		return;
 
-	fill_detail_timing_data(panel_fixed_mode, panel_dvo_timing);
+	fill_detail_timing_data(panel_fixed_mode, dvo_timing);
 
 	dev_priv->lfp_lvds_vbt_mode = panel_fixed_mode;
 
 	DRM_DEBUG_KMS("Found panel mode in BIOS VBT tables:\n");
 	drm_mode_debug_printmodeline(panel_fixed_mode);
 
+	temp_mode = kzalloc(sizeof(*temp_mode), GFP_KERNEL);
+	temp_downclock = panel_fixed_mode->clock;
 	/*
-	 * Iterate over the LVDS panel timing info to find the lowest clock
-	 * for the native resolution.
+	 * enumerate the LVDS panel timing info entry in VBT to check whether
+	 * the LVDS downclock is found.
 	 */
-	downclock = panel_dvo_timing->clock;
 	for (i = 0; i < 16; i++) {
-		const struct lvds_dvo_timing *dvo_timing;
-
-		dvo_timing = get_lvds_dvo_timing(lvds_lfp_data,
-						 lvds_lfp_data_ptrs,
-						 i);
-		if (lvds_dvo_timing_equal_size(dvo_timing, panel_dvo_timing) &&
-		    dvo_timing->clock < downclock)
-			downclock = dvo_timing->clock;
+		entry = (struct bdb_lvds_lfp_data_entry *)
+			((uint8_t *)lvds_lfp_data->data + (lfp_data_size * i));
+		dvo_timing = (struct lvds_dvo_timing *)
+			((unsigned char *)entry + dvo_timing_offset);
+
+		fill_detail_timing_data(temp_mode, dvo_timing);
+
+		if (temp_mode->hdisplay == panel_fixed_mode->hdisplay &&
+		temp_mode->hsync_start == panel_fixed_mode->hsync_start &&
+		temp_mode->hsync_end == panel_fixed_mode->hsync_end &&
+		temp_mode->htotal == panel_fixed_mode->htotal &&
+		temp_mode->vdisplay == panel_fixed_mode->vdisplay &&
+		temp_mode->vsync_start == panel_fixed_mode->vsync_start &&
+		temp_mode->vsync_end == panel_fixed_mode->vsync_end &&
+		temp_mode->vtotal == panel_fixed_mode->vtotal &&
+		temp_mode->clock < temp_downclock) {
+			/*
+			 * downclock is already found. But we expect
+			 * to find the lower downclock.
+			 */
+			temp_downclock = temp_mode->clock;
+		}
+		/* clear it to zero */
+		memset(temp_mode, 0, sizeof(*temp_mode));
 	}
-
-	if (downclock < panel_dvo_timing->clock && i915_lvds_downclock) {
+	kfree(temp_mode);
+	if (temp_downclock < panel_fixed_mode->clock &&
+	    i915_lvds_downclock) {
 		dev_priv->lvds_downclock_avail = 1;
-		dev_priv->lvds_downclock = downclock * 10;
+		dev_priv->lvds_downclock = temp_downclock;
 		DRM_DEBUG_KMS("LVDS downclock is found in VBT. "
 			      "Normal Clock %dKHz, downclock %dKHz\n",
-			      panel_fixed_mode->clock, 10*downclock);
+			      temp_downclock, panel_fixed_mode->clock);
 	}
+	return;
 }
 
 /* Try to find sdvo panel data */
@@ -310,13 +286,6 @@ parse_general_features(struct drm_i915_private *dev_priv,
 		dev_priv->lvds_use_ssc = general->enable_ssc;
 		dev_priv->lvds_ssc_freq =
 			intel_bios_ssc_frequency(dev, general->ssc_freq);
-		dev_priv->display_clock_mode = general->display_clock_mode;
-		DRM_DEBUG_KMS("BDB_GENERAL_FEATURES int_tv_support %d int_crt_support %d lvds_use_ssc %d lvds_ssc_freq %d display_clock_mode %d\n",
-			      dev_priv->int_tv_support,
-			      dev_priv->int_crt_support,
-			      dev_priv->lvds_use_ssc,
-			      dev_priv->lvds_ssc_freq,
-			      dev_priv->display_clock_mode);
 	}
 }
 
@@ -389,7 +358,7 @@ parse_sdvo_device_mapping(struct drm_i915_private *dev_priv,
 		if (p_child->dvo_port != DEVICE_PORT_DVOB &&
 			p_child->dvo_port != DEVICE_PORT_DVOC) {
 			/* skip the incorrect SDVO port */
-			DRM_DEBUG_KMS("Incorrect SDVO port. Skip it\n");
+			DRM_DEBUG_KMS("Incorrect SDVO port. Skip it \n");
 			continue;
 		}
 		DRM_DEBUG_KMS("the SDVO device with slave addr %2x is found on"
@@ -404,13 +373,15 @@ parse_sdvo_device_mapping(struct drm_i915_private *dev_priv,
 			p_mapping->dvo_wiring = p_child->dvo_wiring;
 			p_mapping->ddc_pin = p_child->ddc_pin;
 			p_mapping->i2c_pin = p_child->i2c_pin;
+			p_mapping->i2c_speed = p_child->i2c_speed;
 			p_mapping->initialized = 1;
-			DRM_DEBUG_KMS("SDVO device: dvo=%x, addr=%x, wiring=%d, ddc_pin=%d, i2c_pin=%d\n",
+			DRM_DEBUG_KMS("SDVO device: dvo=%x, addr=%x, wiring=%d, ddc_pin=%d, i2c_pin=%d, i2c_speed=%d\n",
 				      p_mapping->dvo_port,
 				      p_mapping->slave_addr,
 				      p_mapping->dvo_wiring,
 				      p_mapping->ddc_pin,
-				      p_mapping->i2c_pin);
+				      p_mapping->i2c_pin,
+				      p_mapping->i2c_speed);
 		} else {
 			DRM_DEBUG_KMS("Maybe one SDVO port is shared by "
 					 "two SDVO device.\n");
@@ -570,10 +541,10 @@ parse_device_mapping(struct drm_i915_private *dev_priv,
 		count++;
 	}
 	if (!count) {
-		DRM_DEBUG_KMS("no child dev is parsed from VBT\n");
+		DRM_DEBUG_KMS("no child dev is parsed from VBT \n");
 		return;
 	}
-	dev_priv->child_dev = kcalloc(count, sizeof(*p_child), GFP_KERNEL);
+	dev_priv->child_dev = kzalloc(sizeof(*p_child) * count, GFP_KERNEL);
 	if (!dev_priv->child_dev) {
 		DRM_DEBUG_KMS("No memory space for child device\n");
 		return;
@@ -616,7 +587,7 @@ init_vbt_defaults(struct drm_i915_private *dev_priv)
 	/* Default to using SSC */
 	dev_priv->lvds_use_ssc = 1;
 	dev_priv->lvds_ssc_freq = intel_bios_ssc_frequency(dev, 1);
-	DRM_DEBUG_KMS("Set default to SSC at %dMHz\n", dev_priv->lvds_ssc_freq);
+	DRM_DEBUG("Set default to SSC at %dMHz\n", dev_priv->lvds_ssc_freq);
 
 	/* eDP data */
 	dev_priv->edp.bpp = 18;
@@ -665,7 +636,7 @@ intel_parse_bios(struct drm_device *dev)
 	if (!dmi_check_system(intel_no_opregion_vbt) && dev_priv->opregion.vbt) {
 		struct vbt_header *vbt = dev_priv->opregion.vbt;
 		if (memcmp(vbt->signature, "$VBT", 4) == 0) {
-			DRM_DEBUG_KMS("Using VBT from OpRegion: %20s\n",
+			DRM_DEBUG_DRIVER("Using VBT from OpRegion: %20s\n",
 					 vbt->signature);
 			bdb = (struct bdb_header *)((char *)vbt + vbt->bdb_offset);
 		} else
@@ -690,7 +661,7 @@ intel_parse_bios(struct drm_device *dev)
 		}
 
 		if (!vbt) {
-			DRM_DEBUG_DRIVER("VBT signature missing\n");
+			DRM_ERROR("VBT signature missing\n");
 			pci_unmap_rom(pdev, bios);
 			return -1;
 		}
diff --git a/drivers/gpu/drm/i915/intel_bios.h b/drivers/gpu/drm/i915/intel_bios.h
index dbda6e3..5f8e4ed 100644
--- a/drivers/gpu/drm/i915/intel_bios.h
+++ b/drivers/gpu/drm/i915/intel_bios.h
@@ -1,5 +1,5 @@
 /*
- * Copyright © 2006 Intel Corporation
+ * Copyright © 2006 Intel Corporation
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -120,9 +120,7 @@ struct bdb_general_features {
 	u8 ssc_freq:1;
 	u8 enable_lfp_on_override:1;
 	u8 disable_ssc_ddt:1;
-	u8 rsvd7:1;
-	u8 display_clock_mode:1;
-	u8 rsvd8:1; /* finish byte */
+	u8 rsvd8:3; /* finish byte */
 
         /* bits 3 */
 	u8 disable_smooth_vision:1;
@@ -135,10 +133,7 @@ struct bdb_general_features {
         /* bits 5 */
 	u8 int_crt_support:1;
 	u8 int_tv_support:1;
-	u8 int_efp_support:1;
-	u8 dp_ssc_enb:1;	/* PCH attached eDP supports SSC */
-	u8 dp_ssc_freq:1;	/* SSC freq for PCH attached eDP */
-	u8 rsvd11:3; /* finish byte */
+	u8 rsvd11:6; /* finish byte */
 } __attribute__((packed));
 
 /* pre-915 */
@@ -202,7 +197,8 @@ struct bdb_general_features {
 struct child_device_config {
 	u16 handle;
 	u16 device_type;
-	u8  device_id[10]; /* ascii string */
+	u8  i2c_speed;
+	u8  rsvd[9];
 	u16 addin_offset;
 	u8  dvo_port; /* See Device_PORT_* above */
 	u8  i2c_pin;
@@ -244,7 +240,7 @@ struct bdb_general_definitions {
 	 * And the device num is related with the size of general definition
 	 * block. It is obtained by using the following formula:
 	 * number = (block_size - sizeof(bdb_general_definitions))/
-	 *	     sizeof(child_device_config);
+	 * 		sizeof(child_device_config);
 	 */
 	struct child_device_config devices[0];
 } __attribute__((packed));
@@ -450,11 +446,11 @@ struct bdb_driver_features {
 #define EDP_VSWING_1_2V		3
 
 struct edp_power_seq {
-	u16 t1_t3;
-	u16 t8;
+	u16 t3;
+	u16 t7;
 	u16 t9;
 	u16 t10;
-	u16 t11_t12;
+	u16 t12;
 } __attribute__ ((packed));
 
 struct edp_link_params {
@@ -467,12 +463,8 @@ struct edp_link_params {
 struct bdb_edp {
 	struct edp_power_seq power_seqs[16];
 	u32 color_depth;
-	struct edp_link_params link_params[16];
 	u32 sdrrs_msa_timing_delay;
-
-	/* ith bit indicates enabled/disabled for (i+1)th panel */
-	u16 edp_s3d_feature;
-	u16 edp_t3_optimization;
+	struct edp_link_params link_params[16];
 } __attribute__ ((packed));
 
 void intel_setup_bios(struct drm_device *dev);
diff --git a/drivers/gpu/drm/i915/intel_crt.c b/drivers/gpu/drm/i915/intel_crt.c
index 4d3d736..0979d88 100644
--- a/drivers/gpu/drm/i915/intel_crt.c
+++ b/drivers/gpu/drm/i915/intel_crt.c
@@ -24,7 +24,6 @@
  *	Eric Anholt <eric@anholt.net>
  */
 
-#include <linux/dmi.h>
 #include <linux/i2c.h>
 #include <linux/slab.h>
 #include "drmP.h"
@@ -70,7 +69,7 @@ static void intel_crt_dpms(struct drm_encoder *encoder, int mode)
 	temp &= ~(ADPA_HSYNC_CNTL_DISABLE | ADPA_VSYNC_CNTL_DISABLE);
 	temp &= ~ADPA_DAC_ENABLE;
 
-	switch (mode) {
+	switch(mode) {
 	case DRM_MODE_DPMS_ON:
 		temp |= ADPA_DAC_ENABLE;
 		break;
@@ -153,13 +152,17 @@ static void intel_crt_mode_set(struct drm_encoder *encoder,
 	if (adjusted_mode->flags & DRM_MODE_FLAG_PVSYNC)
 		adpa |= ADPA_VSYNC_ACTIVE_HIGH;
 
-	/* For CPT allow 3 pipe config, for others just use A or B */
-	if (HAS_PCH_CPT(dev))
-		adpa |= PORT_TRANS_SEL_CPT(intel_crtc->pipe);
-	else if (intel_crtc->pipe == 0)
-		adpa |= ADPA_PIPE_A_SELECT;
-	else
-		adpa |= ADPA_PIPE_B_SELECT;
+	if (intel_crtc->pipe == 0) {
+		if (HAS_PCH_CPT(dev))
+			adpa |= PORT_TRANS_A_SEL_CPT;
+		else
+			adpa |= ADPA_PIPE_A_SELECT;
+	} else {
+		if (HAS_PCH_CPT(dev))
+			adpa |= PORT_TRANS_B_SEL_CPT;
+		else
+			adpa |= ADPA_PIPE_B_SELECT;
+	}
 
 	if (!HAS_PCH_SPLIT(dev))
 		I915_WRITE(BCLRPAT(intel_crtc->pipe), 0);
@@ -541,24 +544,6 @@ static const struct drm_encoder_funcs intel_crt_enc_funcs = {
 	.destroy = intel_encoder_destroy,
 };
 
-static int __init intel_no_crt_dmi_callback(const struct dmi_system_id *id)
-{
-	DRM_DEBUG_KMS("Skipping CRT initialization for %s\n", id->ident);
-	return 1;
-}
-
-static const struct dmi_system_id intel_no_crt[] = {
-	{
-		.callback = intel_no_crt_dmi_callback,
-		.ident = "ACER ZGB",
-		.matches = {
-			DMI_MATCH(DMI_SYS_VENDOR, "ACER"),
-			DMI_MATCH(DMI_PRODUCT_NAME, "ZGB"),
-		},
-	},
-	{ }
-};
-
 void intel_crt_init(struct drm_device *dev)
 {
 	struct drm_connector *connector;
@@ -566,10 +551,6 @@ void intel_crt_init(struct drm_device *dev)
 	struct intel_connector *intel_connector;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	/* Skip machines without VGA that falsely report hotplug events */
-	if (dmi_check_system(intel_no_crt))
-		return;
-
 	crt = kzalloc(sizeof(struct intel_crt), GFP_KERNEL);
 	if (!crt)
 		return;
@@ -594,10 +575,7 @@ void intel_crt_init(struct drm_device *dev)
 				1 << INTEL_ANALOG_CLONE_BIT |
 				1 << INTEL_SDVO_LVDS_CLONE_BIT);
 	crt->base.crtc_mask = (1 << 0) | (1 << 1);
-	if (IS_GEN2(dev))
-		connector->interlace_allowed = 0;
-	else
-		connector->interlace_allowed = 1;
+	connector->interlace_allowed = 1;
 	connector->doublescan_allowed = 0;
 
 	drm_encoder_helper_add(&crt->base.base, &intel_crt_helper_funcs);
diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
index 1e5790f..2e0c24d 100644
--- a/drivers/gpu/drm/i915/intel_display.c
+++ b/drivers/gpu/drm/i915/intel_display.c
@@ -24,58 +24,56 @@
  *	Eric Anholt <eric@anholt.net>
  */
 
-#include <linux/cpufreq.h>
 #include <linux/module.h>
 #include <linux/input.h>
 #include <linux/i2c.h>
 #include <linux/kernel.h>
 #include <linux/slab.h>
 #include <linux/vgaarb.h>
-#include <drm/drm_edid.h>
 #include "drmP.h"
 #include "intel_drv.h"
 #include "i915_drm.h"
 #include "i915_drv.h"
 #include "i915_trace.h"
 #include "drm_dp_helper.h"
+
 #include "drm_crtc_helper.h"
-#include <linux/dma_remapping.h>
 
 #define HAS_eDP (intel_pipe_has_type(crtc, INTEL_OUTPUT_EDP))
 
-bool intel_pipe_has_type(struct drm_crtc *crtc, int type);
+bool intel_pipe_has_type (struct drm_crtc *crtc, int type);
 static void intel_update_watermarks(struct drm_device *dev);
 static void intel_increase_pllclock(struct drm_crtc *crtc);
 static void intel_crtc_update_cursor(struct drm_crtc *crtc, bool on);
 
 typedef struct {
-	/* given values */
-	int n;
-	int m1, m2;
-	int p1, p2;
-	/* derived values */
-	int	dot;
-	int	vco;
-	int	m;
-	int	p;
+    /* given values */
+    int n;
+    int m1, m2;
+    int p1, p2;
+    /* derived values */
+    int	dot;
+    int	vco;
+    int	m;
+    int	p;
 } intel_clock_t;
 
 typedef struct {
-	int	min, max;
+    int	min, max;
 } intel_range_t;
 
 typedef struct {
-	int	dot_limit;
-	int	p2_slow, p2_fast;
+    int	dot_limit;
+    int	p2_slow, p2_fast;
 } intel_p2_t;
 
 #define INTEL_P2_NUM		      2
 typedef struct intel_limit intel_limit_t;
 struct intel_limit {
-	intel_range_t   dot, vco, n, m, m1, m2, p, p1;
-	intel_p2_t	    p2;
-	bool (* find_pll)(const intel_limit_t *, struct drm_crtc *,
-			int, int, intel_clock_t *, intel_clock_t *);
+    intel_range_t   dot, vco, n, m, m1, m2, p, p1;
+    intel_p2_t	    p2;
+    bool (* find_pll)(const intel_limit_t *, struct drm_crtc *,
+		      int, int, intel_clock_t *);
 };
 
 /* FDI */
@@ -83,21 +81,17 @@ struct intel_limit {
 
 static bool
 intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
-		    int target, int refclk, intel_clock_t *match_clock,
-		    intel_clock_t *best_clock);
+		    int target, int refclk, intel_clock_t *best_clock);
 static bool
 intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
-			int target, int refclk, intel_clock_t *match_clock,
-			intel_clock_t *best_clock);
+			int target, int refclk, intel_clock_t *best_clock);
 
 static bool
 intel_find_pll_g4x_dp(const intel_limit_t *, struct drm_crtc *crtc,
-		      int target, int refclk, intel_clock_t *match_clock,
-		      intel_clock_t *best_clock);
+		      int target, int refclk, intel_clock_t *best_clock);
 static bool
 intel_find_pll_ironlake_dp(const intel_limit_t *, struct drm_crtc *crtc,
-			   int target, int refclk, intel_clock_t *match_clock,
-			   intel_clock_t *best_clock);
+			   int target, int refclk, intel_clock_t *best_clock);
 
 static inline u32 /* units of 100MHz */
 intel_fdi_link_freq(struct drm_device *dev)
@@ -110,56 +104,56 @@ intel_fdi_link_freq(struct drm_device *dev)
 }
 
 static const intel_limit_t intel_limits_i8xx_dvo = {
-	.dot = { .min = 25000, .max = 350000 },
-	.vco = { .min = 930000, .max = 1400000 },
-	.n = { .min = 3, .max = 16 },
-	.m = { .min = 96, .max = 140 },
-	.m1 = { .min = 18, .max = 26 },
-	.m2 = { .min = 6, .max = 16 },
-	.p = { .min = 4, .max = 128 },
-	.p1 = { .min = 2, .max = 33 },
+        .dot = { .min = 25000, .max = 350000 },
+        .vco = { .min = 930000, .max = 1400000 },
+        .n = { .min = 3, .max = 16 },
+        .m = { .min = 96, .max = 140 },
+        .m1 = { .min = 18, .max = 26 },
+        .m2 = { .min = 6, .max = 16 },
+        .p = { .min = 4, .max = 128 },
+        .p1 = { .min = 2, .max = 33 },
 	.p2 = { .dot_limit = 165000,
 		.p2_slow = 4, .p2_fast = 2 },
 	.find_pll = intel_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_i8xx_lvds = {
-	.dot = { .min = 25000, .max = 350000 },
-	.vco = { .min = 930000, .max = 1400000 },
-	.n = { .min = 3, .max = 16 },
-	.m = { .min = 96, .max = 140 },
-	.m1 = { .min = 18, .max = 26 },
-	.m2 = { .min = 6, .max = 16 },
-	.p = { .min = 4, .max = 128 },
-	.p1 = { .min = 1, .max = 6 },
+        .dot = { .min = 25000, .max = 350000 },
+        .vco = { .min = 930000, .max = 1400000 },
+        .n = { .min = 3, .max = 16 },
+        .m = { .min = 96, .max = 140 },
+        .m1 = { .min = 18, .max = 26 },
+        .m2 = { .min = 6, .max = 16 },
+        .p = { .min = 4, .max = 128 },
+        .p1 = { .min = 1, .max = 6 },
 	.p2 = { .dot_limit = 165000,
 		.p2_slow = 14, .p2_fast = 7 },
 	.find_pll = intel_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_i9xx_sdvo = {
-	.dot = { .min = 20000, .max = 400000 },
-	.vco = { .min = 1400000, .max = 2800000 },
-	.n = { .min = 1, .max = 6 },
-	.m = { .min = 70, .max = 120 },
-	.m1 = { .min = 10, .max = 22 },
-	.m2 = { .min = 5, .max = 9 },
-	.p = { .min = 5, .max = 80 },
-	.p1 = { .min = 1, .max = 8 },
+        .dot = { .min = 20000, .max = 400000 },
+        .vco = { .min = 1400000, .max = 2800000 },
+        .n = { .min = 1, .max = 6 },
+        .m = { .min = 70, .max = 120 },
+        .m1 = { .min = 10, .max = 22 },
+        .m2 = { .min = 5, .max = 9 },
+        .p = { .min = 5, .max = 80 },
+        .p1 = { .min = 1, .max = 8 },
 	.p2 = { .dot_limit = 200000,
 		.p2_slow = 10, .p2_fast = 5 },
 	.find_pll = intel_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_i9xx_lvds = {
-	.dot = { .min = 20000, .max = 400000 },
-	.vco = { .min = 1400000, .max = 2800000 },
-	.n = { .min = 1, .max = 6 },
-	.m = { .min = 70, .max = 120 },
-	.m1 = { .min = 10, .max = 22 },
-	.m2 = { .min = 5, .max = 9 },
-	.p = { .min = 7, .max = 98 },
-	.p1 = { .min = 1, .max = 8 },
+        .dot = { .min = 20000, .max = 400000 },
+        .vco = { .min = 1400000, .max = 2800000 },
+        .n = { .min = 1, .max = 6 },
+        .m = { .min = 70, .max = 120 },
+        .m1 = { .min = 10, .max = 22 },
+        .m2 = { .min = 5, .max = 9 },
+        .p = { .min = 7, .max = 98 },
+        .p1 = { .min = 1, .max = 8 },
 	.p2 = { .dot_limit = 112000,
 		.p2_slow = 14, .p2_fast = 7 },
 	.find_pll = intel_find_best_PLL,
@@ -227,44 +221,44 @@ static const intel_limit_t intel_limits_g4x_dual_channel_lvds = {
 };
 
 static const intel_limit_t intel_limits_g4x_display_port = {
-	.dot = { .min = 161670, .max = 227000 },
-	.vco = { .min = 1750000, .max = 3500000},
-	.n = { .min = 1, .max = 2 },
-	.m = { .min = 97, .max = 108 },
-	.m1 = { .min = 0x10, .max = 0x12 },
-	.m2 = { .min = 0x05, .max = 0x06 },
-	.p = { .min = 10, .max = 20 },
-	.p1 = { .min = 1, .max = 2},
-	.p2 = { .dot_limit = 0,
+        .dot = { .min = 161670, .max = 227000 },
+        .vco = { .min = 1750000, .max = 3500000},
+        .n = { .min = 1, .max = 2 },
+        .m = { .min = 97, .max = 108 },
+        .m1 = { .min = 0x10, .max = 0x12 },
+        .m2 = { .min = 0x05, .max = 0x06 },
+        .p = { .min = 10, .max = 20 },
+        .p1 = { .min = 1, .max = 2},
+        .p2 = { .dot_limit = 0,
 		.p2_slow = 10, .p2_fast = 10 },
-	.find_pll = intel_find_pll_g4x_dp,
+        .find_pll = intel_find_pll_g4x_dp,
 };
 
 static const intel_limit_t intel_limits_pineview_sdvo = {
-	.dot = { .min = 20000, .max = 400000},
-	.vco = { .min = 1700000, .max = 3500000 },
+        .dot = { .min = 20000, .max = 400000},
+        .vco = { .min = 1700000, .max = 3500000 },
 	/* Pineview's Ncounter is a ring counter */
-	.n = { .min = 3, .max = 6 },
-	.m = { .min = 2, .max = 256 },
+        .n = { .min = 3, .max = 6 },
+        .m = { .min = 2, .max = 256 },
 	/* Pineview only has one combined m divider, which we treat as m2. */
-	.m1 = { .min = 0, .max = 0 },
-	.m2 = { .min = 0, .max = 254 },
-	.p = { .min = 5, .max = 80 },
-	.p1 = { .min = 1, .max = 8 },
+        .m1 = { .min = 0, .max = 0 },
+        .m2 = { .min = 0, .max = 254 },
+        .p = { .min = 5, .max = 80 },
+        .p1 = { .min = 1, .max = 8 },
 	.p2 = { .dot_limit = 200000,
 		.p2_slow = 10, .p2_fast = 5 },
 	.find_pll = intel_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_pineview_lvds = {
-	.dot = { .min = 20000, .max = 400000 },
-	.vco = { .min = 1700000, .max = 3500000 },
-	.n = { .min = 3, .max = 6 },
-	.m = { .min = 2, .max = 256 },
-	.m1 = { .min = 0, .max = 0 },
-	.m2 = { .min = 0, .max = 254 },
-	.p = { .min = 7, .max = 112 },
-	.p1 = { .min = 1, .max = 8 },
+        .dot = { .min = 20000, .max = 400000 },
+        .vco = { .min = 1700000, .max = 3500000 },
+        .n = { .min = 3, .max = 6 },
+        .m = { .min = 2, .max = 256 },
+        .m1 = { .min = 0, .max = 0 },
+        .m2 = { .min = 0, .max = 254 },
+        .p = { .min = 7, .max = 112 },
+        .p1 = { .min = 1, .max = 8 },
 	.p2 = { .dot_limit = 112000,
 		.p2_slow = 14, .p2_fast = 14 },
 	.find_pll = intel_find_best_PLL,
@@ -326,7 +320,7 @@ static const intel_limit_t intel_limits_ironlake_single_lvds_100m = {
 	.m1 = { .min = 12, .max = 22 },
 	.m2 = { .min = 5, .max = 9 },
 	.p = { .min = 28, .max = 112 },
-	.p1 = { .min = 2, .max = 8 },
+	.p1 = { .min = 2,.max = 8 },
 	.p2 = { .dot_limit = 225000,
 		.p2_slow = 14, .p2_fast = 14 },
 	.find_pll = intel_g4x_find_best_PLL,
@@ -340,24 +334,24 @@ static const intel_limit_t intel_limits_ironlake_dual_lvds_100m = {
 	.m1 = { .min = 12, .max = 22 },
 	.m2 = { .min = 5, .max = 9 },
 	.p = { .min = 14, .max = 42 },
-	.p1 = { .min = 2, .max = 6 },
+	.p1 = { .min = 2,.max = 6 },
 	.p2 = { .dot_limit = 225000,
 		.p2_slow = 7, .p2_fast = 7 },
 	.find_pll = intel_g4x_find_best_PLL,
 };
 
 static const intel_limit_t intel_limits_ironlake_display_port = {
-	.dot = { .min = 25000, .max = 350000 },
-	.vco = { .min = 1760000, .max = 3510000},
-	.n = { .min = 1, .max = 2 },
-	.m = { .min = 81, .max = 90 },
-	.m1 = { .min = 12, .max = 22 },
-	.m2 = { .min = 5, .max = 9 },
-	.p = { .min = 10, .max = 20 },
-	.p1 = { .min = 1, .max = 2},
-	.p2 = { .dot_limit = 0,
+        .dot = { .min = 25000, .max = 350000 },
+        .vco = { .min = 1760000, .max = 3510000},
+        .n = { .min = 1, .max = 2 },
+        .m = { .min = 81, .max = 90 },
+        .m1 = { .min = 12, .max = 22 },
+        .m2 = { .min = 5, .max = 9 },
+        .p = { .min = 10, .max = 20 },
+        .p1 = { .min = 1, .max = 2},
+        .p2 = { .dot_limit = 0,
 		.p2_slow = 10, .p2_fast = 10 },
-	.find_pll = intel_find_pll_ironlake_dp,
+        .find_pll = intel_find_pll_ironlake_dp,
 };
 
 static const intel_limit_t *intel_ironlake_limit(struct drm_crtc *crtc,
@@ -409,7 +403,7 @@ static const intel_limit_t *intel_g4x_limit(struct drm_crtc *crtc)
 		limit = &intel_limits_g4x_hdmi;
 	} else if (intel_pipe_has_type(crtc, INTEL_OUTPUT_SDVO)) {
 		limit = &intel_limits_g4x_sdvo;
-	} else if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT)) {
+	} else if (intel_pipe_has_type (crtc, INTEL_OUTPUT_DISPLAYPORT)) {
 		limit = &intel_limits_g4x_display_port;
 	} else /* The option is for other outputs */
 		limit = &intel_limits_i9xx_sdvo;
@@ -493,34 +487,33 @@ static bool intel_PLL_is_valid(struct drm_device *dev,
 			       const intel_clock_t *clock)
 {
 	if (clock->p1  < limit->p1.min  || limit->p1.max  < clock->p1)
-		INTELPllInvalid("p1 out of range\n");
+		INTELPllInvalid ("p1 out of range\n");
 	if (clock->p   < limit->p.min   || limit->p.max   < clock->p)
-		INTELPllInvalid("p out of range\n");
+		INTELPllInvalid ("p out of range\n");
 	if (clock->m2  < limit->m2.min  || limit->m2.max  < clock->m2)
-		INTELPllInvalid("m2 out of range\n");
+		INTELPllInvalid ("m2 out of range\n");
 	if (clock->m1  < limit->m1.min  || limit->m1.max  < clock->m1)
-		INTELPllInvalid("m1 out of range\n");
+		INTELPllInvalid ("m1 out of range\n");
 	if (clock->m1 <= clock->m2 && !IS_PINEVIEW(dev))
-		INTELPllInvalid("m1 <= m2\n");
+		INTELPllInvalid ("m1 <= m2\n");
 	if (clock->m   < limit->m.min   || limit->m.max   < clock->m)
-		INTELPllInvalid("m out of range\n");
+		INTELPllInvalid ("m out of range\n");
 	if (clock->n   < limit->n.min   || limit->n.max   < clock->n)
-		INTELPllInvalid("n out of range\n");
+		INTELPllInvalid ("n out of range\n");
 	if (clock->vco < limit->vco.min || limit->vco.max < clock->vco)
-		INTELPllInvalid("vco out of range\n");
+		INTELPllInvalid ("vco out of range\n");
 	/* XXX: We may need to be checking "Dot clock" depending on the multiplier,
 	 * connector, etc., rather than just a single range.
 	 */
 	if (clock->dot < limit->dot.min || limit->dot.max < clock->dot)
-		INTELPllInvalid("dot out of range\n");
+		INTELPllInvalid ("dot out of range\n");
 
 	return true;
 }
 
 static bool
 intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
-		    int target, int refclk, intel_clock_t *match_clock,
-		    intel_clock_t *best_clock)
+		    int target, int refclk, intel_clock_t *best_clock)
 
 {
 	struct drm_device *dev = crtc->dev;
@@ -548,7 +541,7 @@ intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 			clock.p2 = limit->p2.p2_fast;
 	}
 
-	memset(best_clock, 0, sizeof(*best_clock));
+	memset (best_clock, 0, sizeof (*best_clock));
 
 	for (clock.m1 = limit->m1.min; clock.m1 <= limit->m1.max;
 	     clock.m1++) {
@@ -567,9 +560,6 @@ intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 					if (!intel_PLL_is_valid(dev, limit,
 								&clock))
 						continue;
-					if (match_clock &&
-					    clock.p != match_clock->p)
-						continue;
 
 					this_err = abs(clock.dot - target);
 					if (this_err < err) {
@@ -586,8 +576,7 @@ intel_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 
 static bool
 intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
-			int target, int refclk, intel_clock_t *match_clock,
-			intel_clock_t *best_clock)
+			int target, int refclk, intel_clock_t *best_clock)
 {
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -634,9 +623,6 @@ intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 					if (!intel_PLL_is_valid(dev, limit,
 								&clock))
 						continue;
-					if (match_clock &&
-					    clock.p != match_clock->p)
-						continue;
 
 					this_err = abs(clock.dot - target);
 					if (this_err < err_most) {
@@ -654,8 +640,7 @@ intel_g4x_find_best_PLL(const intel_limit_t *limit, struct drm_crtc *crtc,
 
 static bool
 intel_find_pll_ironlake_dp(const intel_limit_t *limit, struct drm_crtc *crtc,
-			   int target, int refclk, intel_clock_t *match_clock,
-			   intel_clock_t *best_clock)
+			   int target, int refclk, intel_clock_t *best_clock)
 {
 	struct drm_device *dev = crtc->dev;
 	intel_clock_t clock;
@@ -681,8 +666,7 @@ intel_find_pll_ironlake_dp(const intel_limit_t *limit, struct drm_crtc *crtc,
 /* DisplayPort has only two frequencies, 162MHz and 270MHz */
 static bool
 intel_find_pll_g4x_dp(const intel_limit_t *limit, struct drm_crtc *crtc,
-		      int target, int refclk, intel_clock_t *match_clock,
-		      intel_clock_t *best_clock)
+		      int target, int refclk, intel_clock_t *best_clock)
 {
 	intel_clock_t clock;
 	if (target < 200000) {
@@ -817,19 +801,6 @@ static void assert_pch_pll(struct drm_i915_private *dev_priv,
 	u32 val;
 	bool cur_state;
 
-	if (HAS_PCH_CPT(dev_priv->dev)) {
-		u32 pch_dpll;
-
-		pch_dpll = I915_READ(PCH_DPLL_SEL);
-
-		/* Make sure the selected PLL is enabled to the transcoder */
-		WARN(!((pch_dpll >> (4 * pipe)) & 8),
-		     "transcoder %d PLL not enabled\n", pipe);
-
-		/* Convert the transcoder pipe number to a pll pipe number */
-		pipe = (pch_dpll >> (4 * pipe)) & 1;
-	}
-
 	reg = PCH_DPLL(pipe);
 	val = I915_READ(reg);
 	cur_state = !!(val & DPLL_VCO_ENABLE);
@@ -906,7 +877,7 @@ static void assert_panel_unlocked(struct drm_i915_private *dev_priv,
 	int pp_reg, lvds_reg;
 	u32 val;
 	enum pipe panel_pipe = PIPE_A;
-	bool locked = true;
+	bool locked = locked;
 
 	if (HAS_PCH_SPLIT(dev_priv->dev)) {
 		pp_reg = PCH_PP_CONTROL;
@@ -929,17 +900,13 @@ static void assert_panel_unlocked(struct drm_i915_private *dev_priv,
 	     pipe_name(pipe));
 }
 
-void assert_pipe(struct drm_i915_private *dev_priv,
-		 enum pipe pipe, bool state)
+static void assert_pipe(struct drm_i915_private *dev_priv,
+			enum pipe pipe, bool state)
 {
 	int reg;
 	u32 val;
 	bool cur_state;
 
-	/* if we need the pipe A quirk it must be always on */
-	if (pipe == PIPE_A && dev_priv->quirks & QUIRK_PIPEA_FORCE)
-		state = true;
-
 	reg = PIPECONF(pipe);
 	val = I915_READ(reg);
 	cur_state = !!(val & PIPECONF_ENABLE);
@@ -947,25 +914,22 @@ void assert_pipe(struct drm_i915_private *dev_priv,
 	     "pipe %c assertion failure (expected %s, current %s)\n",
 	     pipe_name(pipe), state_string(state), state_string(cur_state));
 }
+#define assert_pipe_enabled(d, p) assert_pipe(d, p, true)
+#define assert_pipe_disabled(d, p) assert_pipe(d, p, false)
 
-static void assert_plane(struct drm_i915_private *dev_priv,
-			 enum plane plane, bool state)
+static void assert_plane_enabled(struct drm_i915_private *dev_priv,
+				 enum plane plane)
 {
 	int reg;
 	u32 val;
-	bool cur_state;
 
 	reg = DSPCNTR(plane);
 	val = I915_READ(reg);
-	cur_state = !!(val & DISPLAY_PLANE_ENABLE);
-	WARN(cur_state != state,
-	     "plane %c assertion failure (expected %s, current %s)\n",
-	     plane_name(plane), state_string(state), state_string(cur_state));
+	WARN(!(val & DISPLAY_PLANE_ENABLE),
+	     "plane %c assertion failure, should be active but is disabled\n",
+	     plane_name(plane));
 }
 
-#define assert_plane_enabled(d, p) assert_plane(d, p, true)
-#define assert_plane_disabled(d, p) assert_plane(d, p, false)
-
 static void assert_planes_disabled(struct drm_i915_private *dev_priv,
 				   enum pipe pipe)
 {
@@ -974,14 +938,8 @@ static void assert_planes_disabled(struct drm_i915_private *dev_priv,
 	int cur_pipe;
 
 	/* Planes are fixed to pipes on ILK+ */
-	if (HAS_PCH_SPLIT(dev_priv->dev)) {
-		reg = DSPCNTR(pipe);
-		val = I915_READ(reg);
-		WARN((val & DISPLAY_PLANE_ENABLE),
-		     "plane %c assertion failure, should be disabled but not\n",
-		     plane_name(pipe));
+	if (HAS_PCH_SPLIT(dev_priv->dev))
 		return;
-	}
 
 	/* Need to check both planes against the pipe */
 	for (i = 0; i < 2; i++) {
@@ -1021,76 +979,11 @@ static void assert_transcoder_disabled(struct drm_i915_private *dev_priv,
 	     pipe_name(pipe));
 }
 
-static bool dp_pipe_enabled(struct drm_i915_private *dev_priv,
-			    enum pipe pipe, u32 port_sel, u32 val)
-{
-	if ((val & DP_PORT_EN) == 0)
-		return false;
-
-	if (HAS_PCH_CPT(dev_priv->dev)) {
-		u32	trans_dp_ctl_reg = TRANS_DP_CTL(pipe);
-		u32	trans_dp_ctl = I915_READ(trans_dp_ctl_reg);
-		if ((trans_dp_ctl & TRANS_DP_PORT_SEL_MASK) != port_sel)
-			return false;
-	} else {
-		if ((val & DP_PIPE_MASK) != (pipe << 30))
-			return false;
-	}
-	return true;
-}
-
-static bool hdmi_pipe_enabled(struct drm_i915_private *dev_priv,
-			      enum pipe pipe, u32 val)
-{
-	if ((val & PORT_ENABLE) == 0)
-		return false;
-
-	if (HAS_PCH_CPT(dev_priv->dev)) {
-		if ((val & PORT_TRANS_SEL_MASK) != PORT_TRANS_SEL_CPT(pipe))
-			return false;
-	} else {
-		if ((val & TRANSCODER_MASK) != TRANSCODER(pipe))
-			return false;
-	}
-	return true;
-}
-
-static bool lvds_pipe_enabled(struct drm_i915_private *dev_priv,
-			      enum pipe pipe, u32 val)
-{
-	if ((val & LVDS_PORT_EN) == 0)
-		return false;
-
-	if (HAS_PCH_CPT(dev_priv->dev)) {
-		if ((val & PORT_TRANS_SEL_MASK) != PORT_TRANS_SEL_CPT(pipe))
-			return false;
-	} else {
-		if ((val & LVDS_PIPE_MASK) != LVDS_PIPE(pipe))
-			return false;
-	}
-	return true;
-}
-
-static bool adpa_pipe_enabled(struct drm_i915_private *dev_priv,
-			      enum pipe pipe, u32 val)
-{
-	if ((val & ADPA_DAC_ENABLE) == 0)
-		return false;
-	if (HAS_PCH_CPT(dev_priv->dev)) {
-		if ((val & PORT_TRANS_SEL_MASK) != PORT_TRANS_SEL_CPT(pipe))
-			return false;
-	} else {
-		if ((val & ADPA_PIPE_SELECT_MASK) != ADPA_PIPE_SELECT(pipe))
-			return false;
-	}
-	return true;
-}
-
 static void assert_pch_dp_disabled(struct drm_i915_private *dev_priv,
-				   enum pipe pipe, int reg, u32 port_sel)
+				   enum pipe pipe, int reg)
 {
 	u32 val = I915_READ(reg);
-	WARN(dp_pipe_enabled(dev_priv, pipe, port_sel, val),
+	WARN(DP_PIPE_ENABLED(val, pipe),
 	     "PCH DP (0x%08x) enabled on transcoder %c, should be disabled\n",
 	     reg, pipe_name(pipe));
 }
@@ -1099,8 +992,8 @@ static void assert_pch_hdmi_disabled(struct drm_i915_private *dev_priv,
 				     enum pipe pipe, int reg)
 {
 	u32 val = I915_READ(reg);
-	WARN(hdmi_pipe_enabled(dev_priv, val, pipe),
-	     "PCH HDMI (0x%08x) enabled on transcoder %c, should be disabled\n",
+	WARN(HDMI_PIPE_ENABLED(val, pipe),
+	     "PCH DP (0x%08x) enabled on transcoder %c, should be disabled\n",
 	     reg, pipe_name(pipe));
 }
 
@@ -1110,19 +1003,19 @@ static void assert_pch_ports_disabled(struct drm_i915_private *dev_priv,
 	int reg;
 	u32 val;
 
-	assert_pch_dp_disabled(dev_priv, pipe, PCH_DP_B, TRANS_DP_PORT_SEL_B);
-	assert_pch_dp_disabled(dev_priv, pipe, PCH_DP_C, TRANS_DP_PORT_SEL_C);
-	assert_pch_dp_disabled(dev_priv, pipe, PCH_DP_D, TRANS_DP_PORT_SEL_D);
+	assert_pch_dp_disabled(dev_priv, pipe, PCH_DP_B);
+	assert_pch_dp_disabled(dev_priv, pipe, PCH_DP_C);
+	assert_pch_dp_disabled(dev_priv, pipe, PCH_DP_D);
 
 	reg = PCH_ADPA;
 	val = I915_READ(reg);
-	WARN(adpa_pipe_enabled(dev_priv, val, pipe),
+	WARN(ADPA_PIPE_ENABLED(val, pipe),
 	     "PCH VGA enabled on transcoder %c, should be disabled\n",
 	     pipe_name(pipe));
 
 	reg = PCH_LVDS;
 	val = I915_READ(reg);
-	WARN(lvds_pipe_enabled(dev_priv, val, pipe),
+	WARN(LVDS_PIPE_ENABLED(val, pipe),
 	     "PCH LVDS enabled on transcoder %c, should be disabled\n",
 	     pipe_name(pipe));
 
@@ -1212,9 +1105,6 @@ static void intel_enable_pch_pll(struct drm_i915_private *dev_priv,
 	int reg;
 	u32 val;
 
-	if (pipe > 1)
-		return;
-
 	/* PCH only available on ILK+ */
 	BUG_ON(dev_priv->info->gen < 5);
 
@@ -1233,11 +1123,7 @@ static void intel_disable_pch_pll(struct drm_i915_private *dev_priv,
 				  enum pipe pipe)
 {
 	int reg;
-	u32 val, pll_mask = TRANSC_DPLL_ENABLE | TRANSC_DPLLB_SEL,
-		pll_sel = TRANSC_DPLL_ENABLE;
-
-	if (pipe > 1)
-		return;
+	u32 val;
 
 	/* PCH only available on ILK+ */
 	BUG_ON(dev_priv->info->gen < 5);
@@ -1245,15 +1131,6 @@ static void intel_disable_pch_pll(struct drm_i915_private *dev_priv,
 	/* Make sure transcoder isn't still depending on us */
 	assert_transcoder_disabled(dev_priv, pipe);
 
-	if (pipe == 0)
-		pll_sel |= TRANSC_DPLLA_SEL;
-	else if (pipe == 1)
-		pll_sel |= TRANSC_DPLLB_SEL;
-
-
-	if ((I915_READ(PCH_DPLL_SEL) & pll_mask) == pll_sel)
-		return;
-
 	reg = PCH_DPLL(pipe);
 	val = I915_READ(reg);
 	val &= ~DPLL_VCO_ENABLE;
@@ -1266,8 +1143,7 @@ static void intel_enable_transcoder(struct drm_i915_private *dev_priv,
 				    enum pipe pipe)
 {
 	int reg;
-	u32 val, pipeconf_val;
-	struct drm_crtc *crtc = dev_priv->pipe_to_crtc_mapping[pipe];
+	u32 val;
 
 	/* PCH only available on ILK+ */
 	BUG_ON(dev_priv->info->gen < 5);
@@ -1281,27 +1157,12 @@ static void intel_enable_transcoder(struct drm_i915_private *dev_priv,
 
 	reg = TRANSCONF(pipe);
 	val = I915_READ(reg);
-	pipeconf_val = I915_READ(PIPECONF(pipe));
-
-	if (HAS_PCH_IBX(dev_priv->dev)) {
-		/*
-		 * make the BPC in transcoder be consistent with
-		 * that in pipeconf reg.
-		 */
-		val &= ~PIPE_BPC_MASK;
-		val |= pipeconf_val & PIPE_BPC_MASK;
-	}
-
-	val &= ~TRANS_INTERLACE_MASK;
-	if ((pipeconf_val & PIPECONF_INTERLACE_MASK) == PIPECONF_INTERLACED_ILK)
-		if (HAS_PCH_IBX(dev_priv->dev) &&
-		    intel_pipe_has_type(crtc, INTEL_OUTPUT_SDVO))
-			val |= TRANS_LEGACY_INTERLACED_ILK;
-		else
-			val |= TRANS_INTERLACED;
-	else
-		val |= TRANS_PROGRESSIVE;
-
+	/*
+	 * make the BPC in transcoder be consistent with
+	 * that in pipeconf reg.
+	 */
+	val &= ~PIPE_BPC_MASK;
+	val |= I915_READ(PIPECONF(pipe)) & PIPE_BPC_MASK;
 	I915_WRITE(reg, val | TRANS_ENABLE);
 	if (wait_for(I915_READ(reg) & TRANS_STATE_ENABLE, 100))
 		DRM_ERROR("failed to enable transcoder %d\n", pipe);
@@ -1326,7 +1187,7 @@ static void intel_disable_transcoder(struct drm_i915_private *dev_priv,
 	I915_WRITE(reg, val);
 	/* wait for PCH transcoder off, transcoder state */
 	if (wait_for((I915_READ(reg) & TRANS_STATE_ENABLE) == 0, 50))
-		DRM_ERROR("failed to disable transcoder %d\n", pipe);
+		DRM_ERROR("failed to disable transcoder\n");
 }
 
 /**
@@ -1411,17 +1272,6 @@ static void intel_disable_pipe(struct drm_i915_private *dev_priv,
 	intel_wait_for_pipe_off(dev_priv->dev, pipe);
 }
 
-/*
- * Plane regs are double buffered, going from enabled->disabled needs a
- * trigger in order to latch.  The display address reg provides this.
- */
-static void intel_flush_display_plane(struct drm_i915_private *dev_priv,
-				      enum plane plane)
-{
-	I915_WRITE(DSPADDR(plane), I915_READ(DSPADDR(plane)));
-	I915_WRITE(DSPSURF(plane), I915_READ(DSPSURF(plane)));
-}
-
 /**
  * intel_enable_plane - enable a display plane on a given pipe
  * @dev_priv: i915 private structure
@@ -1445,10 +1295,20 @@ static void intel_enable_plane(struct drm_i915_private *dev_priv,
 		return;
 
 	I915_WRITE(reg, val | DISPLAY_PLANE_ENABLE);
-	intel_flush_display_plane(dev_priv, plane);
 	intel_wait_for_vblank(dev_priv->dev, pipe);
 }
 
+/*
+ * Plane regs are double buffered, going from enabled->disabled needs a
+ * trigger in order to latch.  The display address reg provides this.
+ */
+static void intel_flush_display_plane(struct drm_i915_private *dev_priv,
+				      enum plane plane)
+{
+	u32 reg = DSPADDR(plane);
+	I915_WRITE(reg, I915_READ(reg));
+}
+
 /**
  * intel_disable_plane - disable a display plane
  * @dev_priv: i915 private structure
@@ -1474,24 +1334,19 @@ static void intel_disable_plane(struct drm_i915_private *dev_priv,
 }
 
 static void disable_pch_dp(struct drm_i915_private *dev_priv,
-			   enum pipe pipe, int reg, u32 port_sel)
+			   enum pipe pipe, int reg)
 {
 	u32 val = I915_READ(reg);
-	if (dp_pipe_enabled(dev_priv, pipe, port_sel, val)) {
-		DRM_DEBUG_KMS("Disabling pch dp %x on pipe %d\n", reg, pipe);
+	if (DP_PIPE_ENABLED(val, pipe))
 		I915_WRITE(reg, val & ~DP_PORT_EN);
-	}
 }
 
 static void disable_pch_hdmi(struct drm_i915_private *dev_priv,
 			     enum pipe pipe, int reg)
 {
 	u32 val = I915_READ(reg);
-	if (hdmi_pipe_enabled(dev_priv, val, pipe)) {
-		DRM_DEBUG_KMS("Disabling pch HDMI %x on pipe %d\n",
-			      reg, pipe);
+	if (HDMI_PIPE_ENABLED(val, pipe))
 		I915_WRITE(reg, val & ~PORT_ENABLE);
-	}
 }
 
 /* Disable any ports connected to this transcoder */
@@ -1503,19 +1358,18 @@ static void intel_disable_pch_ports(struct drm_i915_private *dev_priv,
 	val = I915_READ(PCH_PP_CONTROL);
 	I915_WRITE(PCH_PP_CONTROL, val | PANEL_UNLOCK_REGS);
 
-	disable_pch_dp(dev_priv, pipe, PCH_DP_B, TRANS_DP_PORT_SEL_B);
-	disable_pch_dp(dev_priv, pipe, PCH_DP_C, TRANS_DP_PORT_SEL_C);
-	disable_pch_dp(dev_priv, pipe, PCH_DP_D, TRANS_DP_PORT_SEL_D);
+	disable_pch_dp(dev_priv, pipe, PCH_DP_B);
+	disable_pch_dp(dev_priv, pipe, PCH_DP_C);
+	disable_pch_dp(dev_priv, pipe, PCH_DP_D);
 
 	reg = PCH_ADPA;
 	val = I915_READ(reg);
-	if (adpa_pipe_enabled(dev_priv, val, pipe))
+	if (ADPA_PIPE_ENABLED(val, pipe))
 		I915_WRITE(reg, val & ~ADPA_DAC_ENABLE);
 
 	reg = PCH_LVDS;
 	val = I915_READ(reg);
-	if (lvds_pipe_enabled(dev_priv, val, pipe)) {
-		DRM_DEBUG_KMS("disable lvds on pipe %d val 0x%08x\n", pipe, val);
+	if (LVDS_PIPE_ENABLED(val, pipe)) {
 		I915_WRITE(reg, val & ~LVDS_PORT_EN);
 		POSTING_READ(reg);
 		udelay(100);
@@ -1526,28 +1380,6 @@ static void intel_disable_pch_ports(struct drm_i915_private *dev_priv,
 	disable_pch_hdmi(dev_priv, pipe, HDMID);
 }
 
-static void i8xx_disable_fbc(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 fbc_ctl;
-
-	/* Disable compression */
-	fbc_ctl = I915_READ(FBC_CONTROL);
-	if ((fbc_ctl & FBC_CTL_EN) == 0)
-		return;
-
-	fbc_ctl &= ~FBC_CTL_EN;
-	I915_WRITE(FBC_CONTROL, fbc_ctl);
-
-	/* Wait for compressing bit to clear */
-	if (wait_for((I915_READ(FBC_STATUS) & FBC_STAT_COMPRESSING) == 0, 10)) {
-		DRM_DEBUG_KMS("FBC idle timed out\n");
-		return;
-	}
-
-	DRM_DEBUG_KMS("disabled FBC\n");
-}
-
 static void i8xx_enable_fbc(struct drm_crtc *crtc, unsigned long interval)
 {
 	struct drm_device *dev = crtc->dev;
@@ -1556,25 +1388,36 @@ static void i8xx_enable_fbc(struct drm_crtc *crtc, unsigned long interval)
 	struct intel_framebuffer *intel_fb = to_intel_framebuffer(fb);
 	struct drm_i915_gem_object *obj = intel_fb->obj;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int cfb_pitch;
 	int plane, i;
 	u32 fbc_ctl, fbc_ctl2;
 
-	cfb_pitch = dev_priv->cfb_size / FBC_LL_SIZE;
-	if (fb->pitches[0] < cfb_pitch)
-		cfb_pitch = fb->pitches[0];
+	if (fb->pitch == dev_priv->cfb_pitch &&
+	    obj->fence_reg == dev_priv->cfb_fence &&
+	    intel_crtc->plane == dev_priv->cfb_plane &&
+	    I915_READ(FBC_CONTROL) & FBC_CTL_EN)
+		return;
+
+	i8xx_disable_fbc(dev);
+
+	dev_priv->cfb_pitch = dev_priv->cfb_size / FBC_LL_SIZE;
+
+	if (fb->pitch < dev_priv->cfb_pitch)
+		dev_priv->cfb_pitch = fb->pitch;
 
 	/* FBC_CTL wants 64B units */
-	cfb_pitch = (cfb_pitch / 64) - 1;
-	plane = intel_crtc->plane == 0 ? FBC_CTL_PLANEA : FBC_CTL_PLANEB;
+	dev_priv->cfb_pitch = (dev_priv->cfb_pitch / 64) - 1;
+	dev_priv->cfb_fence = obj->fence_reg;
+	dev_priv->cfb_plane = intel_crtc->plane;
+	plane = dev_priv->cfb_plane == 0 ? FBC_CTL_PLANEA : FBC_CTL_PLANEB;
 
 	/* Clear old tags */
 	for (i = 0; i < (FBC_LL_SIZE / 32) + 1; i++)
 		I915_WRITE(FBC_TAG + (i * 4), 0);
 
 	/* Set it up... */
-	fbc_ctl2 = FBC_CTL_FENCE_DBL | FBC_CTL_IDLE_IMM | FBC_CTL_CPU_FENCE;
-	fbc_ctl2 |= plane;
+	fbc_ctl2 = FBC_CTL_FENCE_DBL | FBC_CTL_IDLE_IMM | plane;
+	if (obj->tiling_mode != I915_TILING_NONE)
+		fbc_ctl2 |= FBC_CTL_CPU_FENCE;
 	I915_WRITE(FBC_CONTROL2, fbc_ctl2);
 	I915_WRITE(FBC_FENCE_OFF, crtc->y);
 
@@ -1582,13 +1425,36 @@ static void i8xx_enable_fbc(struct drm_crtc *crtc, unsigned long interval)
 	fbc_ctl = FBC_CTL_EN | FBC_CTL_PERIODIC;
 	if (IS_I945GM(dev))
 		fbc_ctl |= FBC_CTL_C3_IDLE; /* 945 needs special SR handling */
-	fbc_ctl |= (cfb_pitch & 0xff) << FBC_CTL_STRIDE_SHIFT;
+	fbc_ctl |= (dev_priv->cfb_pitch & 0xff) << FBC_CTL_STRIDE_SHIFT;
 	fbc_ctl |= (interval & 0x2fff) << FBC_CTL_INTERVAL_SHIFT;
-	fbc_ctl |= obj->fence_reg;
+	if (obj->tiling_mode != I915_TILING_NONE)
+		fbc_ctl |= dev_priv->cfb_fence;
 	I915_WRITE(FBC_CONTROL, fbc_ctl);
 
-	DRM_DEBUG_KMS("enabled FBC, pitch %d, yoff %d, plane %d, ",
-		      cfb_pitch, crtc->y, intel_crtc->plane);
+	DRM_DEBUG_KMS("enabled FBC, pitch %ld, yoff %d, plane %d, ",
+		      dev_priv->cfb_pitch, crtc->y, dev_priv->cfb_plane);
+}
+
+void i8xx_disable_fbc(struct drm_device *dev)
+{
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	u32 fbc_ctl;
+
+	/* Disable compression */
+	fbc_ctl = I915_READ(FBC_CONTROL);
+	if ((fbc_ctl & FBC_CTL_EN) == 0)
+		return;
+
+	fbc_ctl &= ~FBC_CTL_EN;
+	I915_WRITE(FBC_CONTROL, fbc_ctl);
+
+	/* Wait for compressing bit to clear */
+	if (wait_for((I915_READ(FBC_STATUS) & FBC_STAT_COMPRESSING) == 0, 10)) {
+		DRM_DEBUG_KMS("FBC idle timed out\n");
+		return;
+	}
+
+	DRM_DEBUG_KMS("disabled FBC\n");
 }
 
 static bool i8xx_fbc_enabled(struct drm_device *dev)
@@ -1610,9 +1476,30 @@ static void g4x_enable_fbc(struct drm_crtc *crtc, unsigned long interval)
 	unsigned long stall_watermark = 200;
 	u32 dpfc_ctl;
 
+	dpfc_ctl = I915_READ(DPFC_CONTROL);
+	if (dpfc_ctl & DPFC_CTL_EN) {
+		if (dev_priv->cfb_pitch == dev_priv->cfb_pitch / 64 - 1 &&
+		    dev_priv->cfb_fence == obj->fence_reg &&
+		    dev_priv->cfb_plane == intel_crtc->plane &&
+		    dev_priv->cfb_y == crtc->y)
+			return;
+
+		I915_WRITE(DPFC_CONTROL, dpfc_ctl & ~DPFC_CTL_EN);
+		intel_wait_for_vblank(dev, intel_crtc->pipe);
+	}
+
+	dev_priv->cfb_pitch = (dev_priv->cfb_pitch / 64) - 1;
+	dev_priv->cfb_fence = obj->fence_reg;
+	dev_priv->cfb_plane = intel_crtc->plane;
+	dev_priv->cfb_y = crtc->y;
+
 	dpfc_ctl = plane | DPFC_SR_EN | DPFC_CTL_LIMIT_1X;
-	dpfc_ctl |= DPFC_CTL_FENCE_EN | obj->fence_reg;
-	I915_WRITE(DPFC_CHICKEN, DPFC_HT_MODIFY);
+	if (obj->tiling_mode != I915_TILING_NONE) {
+		dpfc_ctl |= DPFC_CTL_FENCE_EN | dev_priv->cfb_fence;
+		I915_WRITE(DPFC_CHICKEN, DPFC_HT_MODIFY);
+	} else {
+		I915_WRITE(DPFC_CHICKEN, ~DPFC_HT_MODIFY);
+	}
 
 	I915_WRITE(DPFC_RECOMP_CTL, DPFC_RECOMP_STALL_EN |
 		   (stall_watermark << DPFC_RECOMP_STALL_WM_SHIFT) |
@@ -1625,7 +1512,7 @@ static void g4x_enable_fbc(struct drm_crtc *crtc, unsigned long interval)
 	DRM_DEBUG_KMS("enabled fbc on plane %d\n", intel_crtc->plane);
 }
 
-static void g4x_disable_fbc(struct drm_device *dev)
+void g4x_disable_fbc(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 dpfc_ctl;
@@ -1680,12 +1567,32 @@ static void ironlake_enable_fbc(struct drm_crtc *crtc, unsigned long interval)
 	u32 dpfc_ctl;
 
 	dpfc_ctl = I915_READ(ILK_DPFC_CONTROL);
+	if (dpfc_ctl & DPFC_CTL_EN) {
+		if (dev_priv->cfb_pitch == dev_priv->cfb_pitch / 64 - 1 &&
+		    dev_priv->cfb_fence == obj->fence_reg &&
+		    dev_priv->cfb_plane == intel_crtc->plane &&
+		    dev_priv->cfb_offset == obj->gtt_offset &&
+		    dev_priv->cfb_y == crtc->y)
+			return;
+
+		I915_WRITE(ILK_DPFC_CONTROL, dpfc_ctl & ~DPFC_CTL_EN);
+		intel_wait_for_vblank(dev, intel_crtc->pipe);
+	}
+
+	dev_priv->cfb_pitch = (dev_priv->cfb_pitch / 64) - 1;
+	dev_priv->cfb_fence = obj->fence_reg;
+	dev_priv->cfb_plane = intel_crtc->plane;
+	dev_priv->cfb_offset = obj->gtt_offset;
+	dev_priv->cfb_y = crtc->y;
+
 	dpfc_ctl &= DPFC_RESERVED;
 	dpfc_ctl |= (plane | DPFC_CTL_LIMIT_1X);
-	/* Set persistent mode for front-buffer rendering, ala X. */
-	dpfc_ctl |= DPFC_CTL_PERSISTENT_MODE;
-	dpfc_ctl |= (DPFC_CTL_FENCE_EN | obj->fence_reg);
-	I915_WRITE(ILK_DPFC_CHICKEN, DPFC_HT_MODIFY);
+	if (obj->tiling_mode != I915_TILING_NONE) {
+		dpfc_ctl |= (DPFC_CTL_FENCE_EN | dev_priv->cfb_fence);
+		I915_WRITE(ILK_DPFC_CHICKEN, DPFC_HT_MODIFY);
+	} else {
+		I915_WRITE(ILK_DPFC_CHICKEN, ~DPFC_HT_MODIFY);
+	}
 
 	I915_WRITE(ILK_DPFC_RECOMP_CTL, DPFC_RECOMP_STALL_EN |
 		   (stall_watermark << DPFC_RECOMP_STALL_WM_SHIFT) |
@@ -1697,7 +1604,7 @@ static void ironlake_enable_fbc(struct drm_crtc *crtc, unsigned long interval)
 
 	if (IS_GEN6(dev)) {
 		I915_WRITE(SNB_DPFC_CTL_SA,
-			   SNB_CPU_FENCE_ENABLE | obj->fence_reg);
+			   SNB_CPU_FENCE_ENABLE | dev_priv->cfb_fence);
 		I915_WRITE(DPFC_CPU_FENCE_OFFSET, crtc->y);
 		sandybridge_blit_fbc_update(dev);
 	}
@@ -1705,7 +1612,7 @@ static void ironlake_enable_fbc(struct drm_crtc *crtc, unsigned long interval)
 	DRM_DEBUG_KMS("enabled fbc on plane %d\n", intel_crtc->plane);
 }
 
-static void ironlake_disable_fbc(struct drm_device *dev)
+void ironlake_disable_fbc(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 dpfc_ctl;
@@ -1737,109 +1644,24 @@ bool intel_fbc_enabled(struct drm_device *dev)
 	return dev_priv->display.fbc_enabled(dev);
 }
 
-static void intel_fbc_work_fn(struct work_struct *__work)
-{
-	struct intel_fbc_work *work =
-		container_of(to_delayed_work(__work),
-			     struct intel_fbc_work, work);
-	struct drm_device *dev = work->crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	mutex_lock(&dev->struct_mutex);
-	if (work == dev_priv->fbc_work) {
-		/* Double check that we haven't switched fb without cancelling
-		 * the prior work.
-		 */
-		if (work->crtc->fb == work->fb) {
-			dev_priv->display.enable_fbc(work->crtc,
-						     work->interval);
-
-			dev_priv->cfb_plane = to_intel_crtc(work->crtc)->plane;
-			dev_priv->cfb_fb = work->crtc->fb->base.id;
-			dev_priv->cfb_y = work->crtc->y;
-		}
-
-		dev_priv->fbc_work = NULL;
-	}
-	mutex_unlock(&dev->struct_mutex);
-
-	kfree(work);
-}
-
-static void intel_cancel_fbc_work(struct drm_i915_private *dev_priv)
-{
-	if (dev_priv->fbc_work == NULL)
-		return;
-
-	DRM_DEBUG_KMS("cancelling pending FBC enable\n");
-
-	/* Synchronisation is provided by struct_mutex and checking of
-	 * dev_priv->fbc_work, so we can perform the cancellation
-	 * entirely asynchronously.
-	 */
-	if (cancel_delayed_work(&dev_priv->fbc_work->work))
-		/* tasklet was killed before being run, clean up */
-		kfree(dev_priv->fbc_work);
-
-	/* Mark the work as no longer wanted so that if it does
-	 * wake-up (because the work was already running and waiting
-	 * for our mutex), it will discover that is no longer
-	 * necessary to run.
-	 */
-	dev_priv->fbc_work = NULL;
-}
-
-static void intel_enable_fbc(struct drm_crtc *crtc, unsigned long interval)
+void intel_enable_fbc(struct drm_crtc *crtc, unsigned long interval)
 {
-	struct intel_fbc_work *work;
-	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
+	struct drm_i915_private *dev_priv = crtc->dev->dev_private;
 
 	if (!dev_priv->display.enable_fbc)
 		return;
 
-	intel_cancel_fbc_work(dev_priv);
-
-	work = kzalloc(sizeof *work, GFP_KERNEL);
-	if (work == NULL) {
-		dev_priv->display.enable_fbc(crtc, interval);
-		return;
-	}
-
-	work->crtc = crtc;
-	work->fb = crtc->fb;
-	work->interval = interval;
-	INIT_DELAYED_WORK(&work->work, intel_fbc_work_fn);
-
-	dev_priv->fbc_work = work;
-
-	DRM_DEBUG_KMS("scheduling delayed FBC enable\n");
-
-	/* Delay the actual enabling to let pageflipping cease and the
-	 * display to settle before starting the compression. Note that
-	 * this delay also serves a second purpose: it allows for a
-	 * vblank to pass after disabling the FBC before we attempt
-	 * to modify the control registers.
-	 *
-	 * A more complicated solution would involve tracking vblanks
-	 * following the termination of the page-flipping sequence
-	 * and indeed performing the enable as a co-routine and not
-	 * waiting synchronously upon the vblank.
-	 */
-	schedule_delayed_work(&work->work, msecs_to_jiffies(50));
+	dev_priv->display.enable_fbc(crtc, interval);
 }
 
 void intel_disable_fbc(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	intel_cancel_fbc_work(dev_priv);
-
 	if (!dev_priv->display.disable_fbc)
 		return;
 
 	dev_priv->display.disable_fbc(dev);
-	dev_priv->cfb_plane = -1;
 }
 
 /**
@@ -1869,7 +1691,6 @@ static void intel_update_fbc(struct drm_device *dev)
 	struct drm_framebuffer *fb;
 	struct intel_framebuffer *intel_fb;
 	struct drm_i915_gem_object *obj;
-	int enable_fbc;
 
 	DRM_DEBUG_KMS("\n");
 
@@ -1910,15 +1731,8 @@ static void intel_update_fbc(struct drm_device *dev)
 	intel_fb = to_intel_framebuffer(fb);
 	obj = intel_fb->obj;
 
-	enable_fbc = i915_enable_fbc;
-	if (enable_fbc < 0) {
-		DRM_DEBUG_KMS("fbc set to per-chip default\n");
-		enable_fbc = 1;
-		if (INTEL_INFO(dev)->gen <= 6)
-			enable_fbc = 0;
-	}
-	if (!enable_fbc) {
-		DRM_DEBUG_KMS("fbc disabled per module param\n");
+	if (!i915_enable_fbc) {
+		DRM_DEBUG_KMS("fbc disabled per module param (default off)\n");
 		dev_priv->no_fbc_reason = FBC_MODULE_PARAM;
 		goto out_disable;
 	}
@@ -1946,13 +1760,8 @@ static void intel_update_fbc(struct drm_device *dev)
 		dev_priv->no_fbc_reason = FBC_BAD_PLANE;
 		goto out_disable;
 	}
-
-	/* The use of a CPU fence is mandatory in order to detect writes
-	 * by the CPU to the scanout and trigger updates to the FBC.
-	 */
-	if (obj->tiling_mode != I915_TILING_X ||
-	    obj->fence_reg == I915_FENCE_REG_NONE) {
-		DRM_DEBUG_KMS("framebuffer not tiled or fenced, disabling compression\n");
+	if (obj->tiling_mode != I915_TILING_X) {
+		DRM_DEBUG_KMS("framebuffer not tiled, disabling compression\n");
 		dev_priv->no_fbc_reason = FBC_NOT_TILED;
 		goto out_disable;
 	}
@@ -1961,44 +1770,6 @@ static void intel_update_fbc(struct drm_device *dev)
 	if (in_dbg_master())
 		goto out_disable;
 
-	/* If the scanout has not changed, don't modify the FBC settings.
-	 * Note that we make the fundamental assumption that the fb->obj
-	 * cannot be unpinned (and have its GTT offset and fence revoked)
-	 * without first being decoupled from the scanout and FBC disabled.
-	 */
-	if (dev_priv->cfb_plane == intel_crtc->plane &&
-	    dev_priv->cfb_fb == fb->base.id &&
-	    dev_priv->cfb_y == crtc->y)
-		return;
-
-	if (intel_fbc_enabled(dev)) {
-		/* We update FBC along two paths, after changing fb/crtc
-		 * configuration (modeswitching) and after page-flipping
-		 * finishes. For the latter, we know that not only did
-		 * we disable the FBC at the start of the page-flip
-		 * sequence, but also more than one vblank has passed.
-		 *
-		 * For the former case of modeswitching, it is possible
-		 * to switch between two FBC valid configurations
-		 * instantaneously so we do need to disable the FBC
-		 * before we can modify its control registers. We also
-		 * have to wait for the next vblank for that to take
-		 * effect. However, since we delay enabling FBC we can
-		 * assume that a vblank has passed since disabling and
-		 * that we can safely alter the registers in the deferred
-		 * callback.
-		 *
-		 * In the scenario that we go from a valid to invalid
-		 * and then back to valid FBC configuration we have
-		 * no strict enforcement that a vblank occurred since
-		 * disabling the FBC. However, along all current pipe
-		 * disabling paths we do need to wait for a vblank at
-		 * some point. And we wait before enabling FBC anyway.
-		 */
-		DRM_DEBUG_KMS("disabling active FBC for update\n");
-		intel_disable_fbc(dev);
-	}
-
 	intel_enable_fbc(crtc, 500);
 	return;
 
@@ -2041,10 +1812,14 @@ intel_pin_and_fence_fb_obj(struct drm_device *dev,
 	}
 
 	dev_priv->mm.interruptible = false;
-	ret = i915_gem_object_pin_to_display_plane(obj, alignment, pipelined);
+	ret = i915_gem_object_pin(obj, alignment, true);
 	if (ret)
 		goto err_interruptible;
 
+	ret = i915_gem_object_set_to_display_plane(obj, pipelined);
+	if (ret)
+		goto err_unpin;
+
 	/* Install a fence for tiled scan-out. Pre-i965 always needs a
 	 * fence, whereas 965+ only requires a fence if using
 	 * framebuffer compression.  For simplicity, we always install
@@ -2054,8 +1829,6 @@ intel_pin_and_fence_fb_obj(struct drm_device *dev,
 		ret = i915_gem_object_get_fence(obj, pipelined);
 		if (ret)
 			goto err_unpin;
-
-		i915_gem_object_pin_fence(obj);
 	}
 
 	dev_priv->mm.interruptible = true;
@@ -2068,14 +1841,10 @@ err_interruptible:
 	return ret;
 }
 
-void intel_unpin_fb_obj(struct drm_i915_gem_object *obj)
-{
-	i915_gem_object_unpin_fence(obj);
-	i915_gem_object_unpin(obj);
-}
-
-static int i9xx_update_plane(struct drm_crtc *crtc, struct drm_framebuffer *fb,
-			     int x, int y)
+/* Assume fb object is pinned & idle & fenced and just update base pointers */
+static int
+intel_pipe_set_base_atomic(struct drm_crtc *crtc, struct drm_framebuffer *fb,
+			   int x, int y, enum mode_set_atomic state)
 {
 	struct drm_device *dev = crtc->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -2118,7 +1887,7 @@ static int i9xx_update_plane(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 		dspcntr |= DISPPLANE_32BPP_NO_ALPHA;
 		break;
 	default:
-		DRM_ERROR("Unknown color depth %d\n", fb->bits_per_pixel);
+		DRM_ERROR("Unknown color depth\n");
 		return -EINVAL;
 	}
 	if (INTEL_INFO(dev)->gen >= 4) {
@@ -2128,14 +1897,18 @@ static int i9xx_update_plane(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 			dspcntr &= ~DISPPLANE_TILED;
 	}
 
+	if (HAS_PCH_SPLIT(dev))
+		/* must disable */
+		dspcntr |= DISPPLANE_TRICKLE_FEED_DISABLE;
+
 	I915_WRITE(reg, dspcntr);
 
 	Start = obj->gtt_offset;
-	Offset = y * fb->pitches[0] + x * (fb->bits_per_pixel / 8);
+	Offset = y * fb->pitch + x * (fb->bits_per_pixel / 8);
 
 	DRM_DEBUG_KMS("Writing base %08lX %08lX %d %d %d\n",
-		      Start, Offset, x, y, fb->pitches[0]);
-	I915_WRITE(DSPSTRIDE(plane), fb->pitches[0]);
+		      Start, Offset, x, y, fb->pitch);
+	I915_WRITE(DSPSTRIDE(plane), fb->pitch);
 	if (INTEL_INFO(dev)->gen >= 4) {
 		I915_WRITE(DSPSURF(plane), Start);
 		I915_WRITE(DSPTILEOFF(plane), (y << 16) | x);
@@ -2144,100 +1917,6 @@ static int i9xx_update_plane(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 		I915_WRITE(DSPADDR(plane), Start + Offset);
 	POSTING_READ(reg);
 
-	return 0;
-}
-
-static int ironlake_update_plane(struct drm_crtc *crtc,
-				 struct drm_framebuffer *fb, int x, int y)
-{
-	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct intel_framebuffer *intel_fb;
-	struct drm_i915_gem_object *obj;
-	int plane = intel_crtc->plane;
-	unsigned long Start, Offset;
-	u32 dspcntr;
-	u32 reg;
-
-	switch (plane) {
-	case 0:
-	case 1:
-	case 2:
-		break;
-	default:
-		DRM_ERROR("Can't update plane %d in SAREA\n", plane);
-		return -EINVAL;
-	}
-
-	intel_fb = to_intel_framebuffer(fb);
-	obj = intel_fb->obj;
-
-	reg = DSPCNTR(plane);
-	dspcntr = I915_READ(reg);
-	/* Mask out pixel format bits in case we change it */
-	dspcntr &= ~DISPPLANE_PIXFORMAT_MASK;
-	switch (fb->bits_per_pixel) {
-	case 8:
-		dspcntr |= DISPPLANE_8BPP;
-		break;
-	case 16:
-		if (fb->depth != 16)
-			return -EINVAL;
-
-		dspcntr |= DISPPLANE_16BPP;
-		break;
-	case 24:
-	case 32:
-		if (fb->depth == 24)
-			dspcntr |= DISPPLANE_32BPP_NO_ALPHA;
-		else if (fb->depth == 30)
-			dspcntr |= DISPPLANE_32BPP_30BIT_NO_ALPHA;
-		else
-			return -EINVAL;
-		break;
-	default:
-		DRM_ERROR("Unknown color depth %d\n", fb->bits_per_pixel);
-		return -EINVAL;
-	}
-
-	if (obj->tiling_mode != I915_TILING_NONE)
-		dspcntr |= DISPPLANE_TILED;
-	else
-		dspcntr &= ~DISPPLANE_TILED;
-
-	/* must disable */
-	dspcntr |= DISPPLANE_TRICKLE_FEED_DISABLE;
-
-	I915_WRITE(reg, dspcntr);
-
-	Start = obj->gtt_offset;
-	Offset = y * fb->pitches[0] + x * (fb->bits_per_pixel / 8);
-
-	DRM_DEBUG_KMS("Writing base %08lX %08lX %d %d %d\n",
-		      Start, Offset, x, y, fb->pitches[0]);
-	I915_WRITE(DSPSTRIDE(plane), fb->pitches[0]);
-	I915_WRITE(DSPSURF(plane), Start);
-	I915_WRITE(DSPTILEOFF(plane), (y << 16) | x);
-	I915_WRITE(DSPADDR(plane), Offset);
-	POSTING_READ(reg);
-
-	return 0;
-}
-
-/* Assume fb object is pinned & idle & fenced and just update base pointers */
-static int
-intel_pipe_set_base_atomic(struct drm_crtc *crtc, struct drm_framebuffer *fb,
-			   int x, int y, enum mode_set_atomic state)
-{
-	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int ret;
-
-	ret = dev_priv->display.update_plane(crtc, fb, x, y);
-	if (ret)
-		return ret;
-
 	intel_update_fbc(dev);
 	intel_increase_pllclock(crtc);
 
@@ -2255,7 +1934,7 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
 
 	/* no fb bound */
 	if (!crtc->fb) {
-		DRM_ERROR("No FB bound\n");
+		DRM_DEBUG_KMS("No FB bound\n");
 		return 0;
 	}
 
@@ -2263,12 +1942,7 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
 	case 0:
 	case 1:
 		break;
-	case 2:
-		if (IS_IVYBRIDGE(dev))
-			break;
-		/* fall through otherwise */
 	default:
-		DRM_ERROR("no plane for crtc\n");
 		return -EINVAL;
 	}
 
@@ -2278,7 +1952,6 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
 					 NULL);
 	if (ret != 0) {
 		mutex_unlock(&dev->struct_mutex);
-		DRM_ERROR("pin & fence failed\n");
 		return ret;
 	}
 
@@ -2298,22 +1971,21 @@ intel_pipe_set_base(struct drm_crtc *crtc, int x, int y,
 		 * This should only fail upon a hung GPU, in which case we
 		 * can safely continue.
 		 */
-		ret = i915_gem_object_finish_gpu(obj);
+		ret = i915_gem_object_flush_gpu(obj);
 		(void) ret;
 	}
 
 	ret = intel_pipe_set_base_atomic(crtc, crtc->fb, x, y,
 					 LEAVE_ATOMIC_MODE_SET);
 	if (ret) {
-		intel_unpin_fb_obj(to_intel_framebuffer(crtc->fb)->obj);
+		i915_gem_object_unpin(to_intel_framebuffer(crtc->fb)->obj);
 		mutex_unlock(&dev->struct_mutex);
-		DRM_ERROR("failed to update base address\n");
 		return ret;
 	}
 
 	if (old_fb) {
 		intel_wait_for_vblank(dev, intel_crtc->pipe);
-		intel_unpin_fb_obj(to_intel_framebuffer(old_fb)->obj);
+		i915_gem_object_unpin(to_intel_framebuffer(old_fb)->obj);
 	}
 
 	mutex_unlock(&dev->struct_mutex);
@@ -2414,18 +2086,6 @@ static void intel_fdi_normal_train(struct drm_crtc *crtc)
 			   FDI_FE_ERRC_ENABLE);
 }
 
-static void cpt_phase_pointer_enable(struct drm_device *dev, int pipe)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 flags = I915_READ(SOUTH_CHICKEN1);
-
-	flags |= FDI_PHASE_SYNC_OVR(pipe);
-	I915_WRITE(SOUTH_CHICKEN1, flags); /* once to unlock... */
-	flags |= FDI_PHASE_SYNC_EN(pipe);
-	I915_WRITE(SOUTH_CHICKEN1, flags); /* then again to enable */
-	POSTING_READ(SOUTH_CHICKEN1);
-}
-
 /* The FDI link training functions for ILK/Ibexpeak. */
 static void ironlake_fdi_link_train(struct drm_crtc *crtc)
 {
@@ -2523,7 +2183,7 @@ static void ironlake_fdi_link_train(struct drm_crtc *crtc)
 
 }
 
-static const int snb_b_fdi_train_param[] = {
+static const int snb_b_fdi_train_param [] = {
 	FDI_LINK_TRAIN_400MV_0DB_SNB_B,
 	FDI_LINK_TRAIN_400MV_6DB_SNB_B,
 	FDI_LINK_TRAIN_600MV_3_5DB_SNB_B,
@@ -2576,10 +2236,7 @@ static void gen6_fdi_link_train(struct drm_crtc *crtc)
 	POSTING_READ(reg);
 	udelay(150);
 
-	if (HAS_PCH_CPT(dev))
-		cpt_phase_pointer_enable(dev, pipe);
-
-	for (i = 0; i < 4; i++) {
+	for (i = 0; i < 4; i++ ) {
 		reg = FDI_TX_CTL(pipe);
 		temp = I915_READ(reg);
 		temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
@@ -2628,7 +2285,7 @@ static void gen6_fdi_link_train(struct drm_crtc *crtc)
 	POSTING_READ(reg);
 	udelay(150);
 
-	for (i = 0; i < 4; i++) {
+	for (i = 0; i < 4; i++ ) {
 		reg = FDI_TX_CTL(pipe);
 		temp = I915_READ(reg);
 		temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
@@ -2697,10 +2354,7 @@ static void ivb_manual_fdi_link_train(struct drm_crtc *crtc)
 	POSTING_READ(reg);
 	udelay(150);
 
-	if (HAS_PCH_CPT(dev))
-		cpt_phase_pointer_enable(dev, pipe);
-
-	for (i = 0; i < 4; i++) {
+	for (i = 0; i < 4; i++ ) {
 		reg = FDI_TX_CTL(pipe);
 		temp = I915_READ(reg);
 		temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
@@ -2742,7 +2396,7 @@ static void ivb_manual_fdi_link_train(struct drm_crtc *crtc)
 	POSTING_READ(reg);
 	udelay(150);
 
-	for (i = 0; i < 4; i++) {
+	for (i = 0; i < 4; i++ ) {
 		reg = FDI_TX_CTL(pipe);
 		temp = I915_READ(reg);
 		temp &= ~FDI_LINK_TRAIN_VOL_EMP_MASK;
@@ -2809,17 +2463,6 @@ static void ironlake_fdi_pll_enable(struct drm_crtc *crtc)
 	}
 }
 
-static void cpt_phase_pointer_disable(struct drm_device *dev, int pipe)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 flags = I915_READ(SOUTH_CHICKEN1);
-
-	flags &= ~(FDI_PHASE_SYNC_EN(pipe));
-	I915_WRITE(SOUTH_CHICKEN1, flags); /* once to disable... */
-	flags &= ~(FDI_PHASE_SYNC_OVR(pipe));
-	I915_WRITE(SOUTH_CHICKEN1, flags); /* then again to lock */
-	POSTING_READ(SOUTH_CHICKEN1);
-}
 static void ironlake_fdi_disable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
@@ -2849,8 +2492,6 @@ static void ironlake_fdi_disable(struct drm_crtc *crtc)
 		I915_WRITE(FDI_RX_CHICKEN(pipe),
 			   I915_READ(FDI_RX_CHICKEN(pipe) &
 				     ~FDI_RX_PHASE_SYNC_POINTER_EN));
-	} else if (HAS_PCH_CPT(dev)) {
-		cpt_phase_pointer_disable(dev, pipe);
 	}
 
 	/* still set train pattern 1 */
@@ -2951,7 +2592,7 @@ static void ironlake_pch_enable(struct drm_crtc *crtc)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	int pipe = intel_crtc->pipe;
-	u32 reg, temp, transc_sel;
+	u32 reg, temp;
 
 	/* For PCH output, training FDI link */
 	dev_priv->display.fdi_link_train(crtc);
@@ -2959,21 +2600,12 @@ static void ironlake_pch_enable(struct drm_crtc *crtc)
 	intel_enable_pch_pll(dev_priv, pipe);
 
 	if (HAS_PCH_CPT(dev)) {
-		transc_sel = intel_crtc->use_pll_a ? TRANSC_DPLLA_SEL :
-			TRANSC_DPLLB_SEL;
-
 		/* Be sure PCH DPLL SEL is set */
 		temp = I915_READ(PCH_DPLL_SEL);
-		if (pipe == 0) {
-			temp &= ~(TRANSA_DPLLB_SEL);
+		if (pipe == 0 && (temp & TRANSA_DPLL_ENABLE) == 0)
 			temp |= (TRANSA_DPLL_ENABLE | TRANSA_DPLLA_SEL);
-		} else if (pipe == 1) {
-			temp &= ~(TRANSB_DPLLB_SEL);
+		else if (pipe == 1 && (temp & TRANSB_DPLL_ENABLE) == 0)
 			temp |= (TRANSB_DPLL_ENABLE | TRANSB_DPLLB_SEL);
-		} else if (pipe == 2) {
-			temp &= ~(TRANSC_DPLLB_SEL);
-			temp |= (TRANSC_DPLL_ENABLE | transc_sel);
-		}
 		I915_WRITE(PCH_DPLL_SEL, temp);
 	}
 
@@ -2986,15 +2618,12 @@ static void ironlake_pch_enable(struct drm_crtc *crtc)
 	I915_WRITE(TRANS_VTOTAL(pipe), I915_READ(VTOTAL(pipe)));
 	I915_WRITE(TRANS_VBLANK(pipe), I915_READ(VBLANK(pipe)));
 	I915_WRITE(TRANS_VSYNC(pipe),  I915_READ(VSYNC(pipe)));
-	I915_WRITE(TRANS_VSYNCSHIFT(pipe),  I915_READ(VSYNCSHIFT(pipe)));
 
 	intel_fdi_normal_train(crtc);
 
 	/* For PCH DP, enable TRANS_DP_CTL */
 	if (HAS_PCH_CPT(dev) &&
-	    (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT) ||
-	     intel_pipe_has_type(crtc, INTEL_OUTPUT_EDP))) {
-		u32 bpc = (I915_READ(PIPECONF(pipe)) & PIPE_BPC_MASK) >> 5;
+	    intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT)) {
 		reg = TRANS_DP_CTL(pipe);
 		temp = I915_READ(reg);
 		temp &= ~(TRANS_DP_PORT_SEL_MASK |
@@ -3002,7 +2631,7 @@ static void ironlake_pch_enable(struct drm_crtc *crtc)
 			  TRANS_DP_BPC_MASK);
 		temp |= (TRANS_DP_OUTPUT_ENABLE |
 			 TRANS_DP_ENH_FRAMING);
-		temp |= bpc << 9; /* same format but at 11:9 */
+		temp |= TRANS_DP_8BPC;
 
 		if (crtc->mode.flags & DRM_MODE_FLAG_PHSYNC)
 			temp |= TRANS_DP_HSYNC_ACTIVE_HIGH;
@@ -3031,24 +2660,6 @@ static void ironlake_pch_enable(struct drm_crtc *crtc)
 	intel_enable_transcoder(dev_priv, pipe);
 }
 
-void intel_cpt_verify_modeset(struct drm_device *dev, int pipe)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int dslreg = PIPEDSL(pipe), tc2reg = TRANS_CHICKEN2(pipe);
-	u32 temp;
-
-	temp = I915_READ(dslreg);
-	udelay(500);
-	if (wait_for(I915_READ(dslreg) != temp, 5)) {
-		/* Without this, mode sets may fail silently on FDI */
-		I915_WRITE(tc2reg, TRANS_AUTOTRAIN_GEN_STALL_DIS);
-		udelay(250);
-		I915_WRITE(tc2reg, 0);
-		if (wait_for(I915_READ(dslreg) != temp, 5))
-			DRM_ERROR("mode set failed: pipe %d stuck\n", pipe);
-	}
-}
-
 static void ironlake_crtc_enable(struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
@@ -3131,8 +2742,9 @@ static void ironlake_crtc_disable(struct drm_crtc *crtc)
 
 	intel_disable_plane(dev_priv, plane, pipe);
 
-	if (dev_priv->cfb_plane == plane)
-		intel_disable_fbc(dev);
+	if (dev_priv->cfb_plane == plane &&
+	    dev_priv->display.disable_fbc)
+		dev_priv->display.disable_fbc(dev);
 
 	intel_disable_pipe(dev_priv, pipe);
 
@@ -3163,13 +2775,13 @@ static void ironlake_crtc_disable(struct drm_crtc *crtc)
 		temp = I915_READ(PCH_DPLL_SEL);
 		switch (pipe) {
 		case 0:
-			temp &= ~(TRANSA_DPLL_ENABLE | TRANSA_DPLLB_SEL);
+			temp &= ~(TRANSA_DPLL_ENABLE | TRANSA_DPLLA_SEL);
 			break;
 		case 1:
 			temp &= ~(TRANSB_DPLL_ENABLE | TRANSB_DPLLB_SEL);
 			break;
 		case 2:
-			/* C shares PLL A or B */
+			/* FIXME: manage transcoder PLLs? */
 			temp &= ~(TRANSC_DPLL_ENABLE | TRANSC_DPLLB_SEL);
 			break;
 		default:
@@ -3179,8 +2791,7 @@ static void ironlake_crtc_disable(struct drm_crtc *crtc)
 	}
 
 	/* disable PCH DPLL */
-	if (!intel_crtc->no_pll)
-		intel_disable_pch_pll(dev_priv, pipe);
+	intel_disable_pch_pll(dev_priv, pipe);
 
 	/* Switch from PCDclk to Rawclk */
 	reg = FDI_RX_CTL(pipe);
@@ -3298,8 +2909,9 @@ static void i9xx_crtc_disable(struct drm_crtc *crtc)
 	intel_crtc_dpms_overlay(intel_crtc, false);
 	intel_crtc_update_cursor(crtc, false);
 
-	if (dev_priv->cfb_plane == plane)
-		intel_disable_fbc(dev);
+	if (dev_priv->cfb_plane == plane &&
+	    dev_priv->display.disable_fbc)
+		dev_priv->display.disable_fbc(dev);
 
 	intel_disable_plane(dev_priv, plane, pipe);
 	intel_disable_pipe(dev_priv, pipe);
@@ -3384,12 +2996,10 @@ static void intel_crtc_disable(struct drm_crtc *crtc)
 	struct drm_device *dev = crtc->dev;
 
 	crtc_funcs->dpms(crtc, DRM_MODE_DPMS_OFF);
-	assert_plane_disabled(dev->dev_private, to_intel_crtc(crtc)->plane);
-	assert_pipe_disabled(dev->dev_private, to_intel_crtc(crtc)->pipe);
 
 	if (crtc->fb) {
 		mutex_lock(&dev->struct_mutex);
-		intel_unpin_fb_obj(to_intel_framebuffer(crtc->fb)->obj);
+		i915_gem_object_unpin(to_intel_framebuffer(crtc->fb)->obj);
 		mutex_unlock(&dev->struct_mutex);
 	}
 }
@@ -3422,25 +3032,18 @@ static void ironlake_crtc_commit(struct drm_crtc *crtc)
 	ironlake_crtc_enable(crtc);
 }
 
-void intel_encoder_prepare(struct drm_encoder *encoder)
+void intel_encoder_prepare (struct drm_encoder *encoder)
 {
 	struct drm_encoder_helper_funcs *encoder_funcs = encoder->helper_private;
 	/* lvds has its own version of prepare see intel_lvds_prepare */
 	encoder_funcs->dpms(encoder, DRM_MODE_DPMS_OFF);
 }
 
-void intel_encoder_commit(struct drm_encoder *encoder)
+void intel_encoder_commit (struct drm_encoder *encoder)
 {
 	struct drm_encoder_helper_funcs *encoder_funcs = encoder->helper_private;
-	struct drm_device *dev = encoder->dev;
-	struct intel_encoder *intel_encoder = to_intel_encoder(encoder);
-	struct intel_crtc *intel_crtc = to_intel_crtc(intel_encoder->base.crtc);
-
 	/* lvds has its own version of commit see intel_lvds_commit */
 	encoder_funcs->dpms(encoder, DRM_MODE_DPMS_ON);
-
-	if (HAS_PCH_CPT(dev))
-		intel_cpt_verify_modeset(dev, intel_crtc->pipe);
 }
 
 void intel_encoder_destroy(struct drm_encoder *encoder)
@@ -3463,8 +3066,11 @@ static bool intel_crtc_mode_fixup(struct drm_crtc *crtc,
 			return false;
 	}
 
-	/* All interlaced capable intel hw wants timings in frames. */
-	drm_mode_set_crtcinfo(adjusted_mode, 0);
+	/* XXX some encoders set the crtcinfo, others don't.
+	 * Obviously we need some form of conflict resolution here...
+	 */
+	if (adjusted_mode->crtc_htotal == 0)
+		drm_mode_set_crtcinfo(adjusted_mode, 0);
 
 	return true;
 }
@@ -4579,11 +4185,10 @@ static void ironlake_update_wm(struct drm_device *dev)
 	 */
 }
 
-void sandybridge_update_wm(struct drm_device *dev)
+static void sandybridge_update_wm(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int latency = SNB_READ_WM0_LATENCY() * 100;	/* In unit 0.1us */
-	u32 val;
 	int fbc_wm, plane_wm, cursor_wm;
 	unsigned int enabled;
 
@@ -4592,10 +4197,8 @@ void sandybridge_update_wm(struct drm_device *dev)
 			    &sandybridge_display_wm_info, latency,
 			    &sandybridge_cursor_wm_info, latency,
 			    &plane_wm, &cursor_wm)) {
-		val = I915_READ(WM0_PIPEA_ILK);
-		val &= ~(WM0_PIPE_PLANE_MASK | WM0_PIPE_CURSOR_MASK);
-		I915_WRITE(WM0_PIPEA_ILK, val |
-			   ((plane_wm << WM0_PIPE_PLANE_SHIFT) | cursor_wm));
+		I915_WRITE(WM0_PIPEA_ILK,
+			   (plane_wm << WM0_PIPE_PLANE_SHIFT) | cursor_wm);
 		DRM_DEBUG_KMS("FIFO watermarks For pipe A -"
 			      " plane %d, " "cursor: %d\n",
 			      plane_wm, cursor_wm);
@@ -4606,32 +4209,14 @@ void sandybridge_update_wm(struct drm_device *dev)
 			    &sandybridge_display_wm_info, latency,
 			    &sandybridge_cursor_wm_info, latency,
 			    &plane_wm, &cursor_wm)) {
-		val = I915_READ(WM0_PIPEB_ILK);
-		val &= ~(WM0_PIPE_PLANE_MASK | WM0_PIPE_CURSOR_MASK);
-		I915_WRITE(WM0_PIPEB_ILK, val |
-			   ((plane_wm << WM0_PIPE_PLANE_SHIFT) | cursor_wm));
+		I915_WRITE(WM0_PIPEB_ILK,
+			   (plane_wm << WM0_PIPE_PLANE_SHIFT) | cursor_wm);
 		DRM_DEBUG_KMS("FIFO watermarks For pipe B -"
 			      " plane %d, cursor: %d\n",
 			      plane_wm, cursor_wm);
 		enabled |= 2;
 	}
 
-	/* IVB has 3 pipes */
-	if (IS_IVYBRIDGE(dev) &&
-	    g4x_compute_wm0(dev, 2,
-			    &sandybridge_display_wm_info, latency,
-			    &sandybridge_cursor_wm_info, latency,
-			    &plane_wm, &cursor_wm)) {
-		val = I915_READ(WM0_PIPEC_IVB);
-		val &= ~(WM0_PIPE_PLANE_MASK | WM0_PIPE_CURSOR_MASK);
-		I915_WRITE(WM0_PIPEC_IVB, val |
-			   ((plane_wm << WM0_PIPE_PLANE_SHIFT) | cursor_wm));
-		DRM_DEBUG_KMS("FIFO watermarks For pipe C -"
-			      " plane %d, cursor: %d\n",
-			      plane_wm, cursor_wm);
-		enabled |= 3;
-	}
-
 	/*
 	 * Calculate and update the self-refresh watermark only when one
 	 * display plane is used.
@@ -4646,8 +4231,7 @@ void sandybridge_update_wm(struct drm_device *dev)
 	I915_WRITE(WM2_LP_ILK, 0);
 	I915_WRITE(WM1_LP_ILK, 0);
 
-	if (!single_plane_enabled(enabled) ||
-	    dev_priv->sprite_scaling_enabled)
+	if (!single_plane_enabled(enabled))
 		return;
 	enabled = ffs(enabled) - 1;
 
@@ -4697,161 +4281,6 @@ void sandybridge_update_wm(struct drm_device *dev)
 		   cursor_wm);
 }
 
-static bool
-sandybridge_compute_sprite_wm(struct drm_device *dev, int plane,
-			      uint32_t sprite_width, int pixel_size,
-			      const struct intel_watermark_params *display,
-			      int display_latency_ns, int *sprite_wm)
-{
-	struct drm_crtc *crtc;
-	int clock;
-	int entries, tlb_miss;
-
-	crtc = intel_get_crtc_for_plane(dev, plane);
-	if (crtc->fb == NULL || !crtc->enabled) {
-		*sprite_wm = display->guard_size;
-		return false;
-	}
-
-	clock = crtc->mode.clock;
-
-	/* Use the small buffer method to calculate the sprite watermark */
-	entries = ((clock * pixel_size / 1000) * display_latency_ns) / 1000;
-	tlb_miss = display->fifo_size*display->cacheline_size -
-		sprite_width * 8;
-	if (tlb_miss > 0)
-		entries += tlb_miss;
-	entries = DIV_ROUND_UP(entries, display->cacheline_size);
-	*sprite_wm = entries + display->guard_size;
-	if (*sprite_wm > (int)display->max_wm)
-		*sprite_wm = display->max_wm;
-
-	return true;
-}
-
-static bool
-sandybridge_compute_sprite_srwm(struct drm_device *dev, int plane,
-				uint32_t sprite_width, int pixel_size,
-				const struct intel_watermark_params *display,
-				int latency_ns, int *sprite_wm)
-{
-	struct drm_crtc *crtc;
-	unsigned long line_time_us;
-	int clock;
-	int line_count, line_size;
-	int small, large;
-	int entries;
-
-	if (!latency_ns) {
-		*sprite_wm = 0;
-		return false;
-	}
-
-	crtc = intel_get_crtc_for_plane(dev, plane);
-	clock = crtc->mode.clock;
-	if (!clock) {
-		*sprite_wm = 0;
-		return false;
-	}
-
-	line_time_us = (sprite_width * 1000) / clock;
-	if (!line_time_us) {
-		*sprite_wm = 0;
-		return false;
-	}
-
-	line_count = (latency_ns / line_time_us + 1000) / 1000;
-	line_size = sprite_width * pixel_size;
-
-	/* Use the minimum of the small and large buffer method for primary */
-	small = ((clock * pixel_size / 1000) * latency_ns) / 1000;
-	large = line_count * line_size;
-
-	entries = DIV_ROUND_UP(min(small, large), display->cacheline_size);
-	*sprite_wm = entries + display->guard_size;
-
-	return *sprite_wm > 0x3ff ? false : true;
-}
-
-static void sandybridge_update_sprite_wm(struct drm_device *dev, int pipe,
-					 uint32_t sprite_width, int pixel_size)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int latency = SNB_READ_WM0_LATENCY() * 100;	/* In unit 0.1us */
-	u32 val;
-	int sprite_wm, reg;
-	int ret;
-
-	switch (pipe) {
-	case 0:
-		reg = WM0_PIPEA_ILK;
-		break;
-	case 1:
-		reg = WM0_PIPEB_ILK;
-		break;
-	case 2:
-		reg = WM0_PIPEC_IVB;
-		break;
-	default:
-		return; /* bad pipe */
-	}
-
-	ret = sandybridge_compute_sprite_wm(dev, pipe, sprite_width, pixel_size,
-					    &sandybridge_display_wm_info,
-					    latency, &sprite_wm);
-	if (!ret) {
-		DRM_DEBUG_KMS("failed to compute sprite wm for pipe %d\n",
-			      pipe);
-		return;
-	}
-
-	val = I915_READ(reg);
-	val &= ~WM0_PIPE_SPRITE_MASK;
-	I915_WRITE(reg, val | (sprite_wm << WM0_PIPE_SPRITE_SHIFT));
-	DRM_DEBUG_KMS("sprite watermarks For pipe %d - %d\n", pipe, sprite_wm);
-
-
-	ret = sandybridge_compute_sprite_srwm(dev, pipe, sprite_width,
-					      pixel_size,
-					      &sandybridge_display_srwm_info,
-					      SNB_READ_WM1_LATENCY() * 500,
-					      &sprite_wm);
-	if (!ret) {
-		DRM_DEBUG_KMS("failed to compute sprite lp1 wm on pipe %d\n",
-			      pipe);
-		return;
-	}
-	I915_WRITE(WM1S_LP_ILK, sprite_wm);
-
-	/* Only IVB has two more LP watermarks for sprite */
-	if (!IS_IVYBRIDGE(dev))
-		return;
-
-	ret = sandybridge_compute_sprite_srwm(dev, pipe, sprite_width,
-					      pixel_size,
-					      &sandybridge_display_srwm_info,
-					      SNB_READ_WM2_LATENCY() * 500,
-					      &sprite_wm);
-	if (!ret) {
-		DRM_DEBUG_KMS("failed to compute sprite lp2 wm on pipe %d\n",
-			      pipe);
-		return;
-	}
-	I915_WRITE(WM2S_LP_IVB, sprite_wm);
-
-	ret = sandybridge_compute_sprite_srwm(dev, pipe, sprite_width,
-					      pixel_size,
-					      &sandybridge_display_srwm_info,
-					      SNB_READ_WM3_LATENCY() * 500,
-					      &sprite_wm);
-	if (!ret) {
-		DRM_DEBUG_KMS("failed to compute sprite lp3 wm on pipe %d\n",
-			      pipe);
-		return;
-	}
-	I915_WRITE(WM3S_LP_IVB, sprite_wm);
-}
-
 /**
  * intel_update_watermarks - update FIFO watermark values based on current modes
  *
@@ -4892,239 +4321,12 @@ static void intel_update_watermarks(struct drm_device *dev)
 		dev_priv->display.update_wm(dev);
 }
 
-void intel_update_sprite_watermarks(struct drm_device *dev, int pipe,
-				    uint32_t sprite_width, int pixel_size)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	if (dev_priv->display.update_sprite_wm)
-		dev_priv->display.update_sprite_wm(dev, pipe, sprite_width,
-						   pixel_size);
-}
-
 static inline bool intel_panel_use_ssc(struct drm_i915_private *dev_priv)
 {
-	if (i915_panel_use_ssc >= 0)
-		return i915_panel_use_ssc != 0;
-	return dev_priv->lvds_use_ssc
+	return dev_priv->lvds_use_ssc && i915_panel_use_ssc
 		&& !(dev_priv->quirks & QUIRK_LVDS_SSC_DISABLE);
 }
 
-/**
- * intel_choose_pipe_bpp_dither - figure out what color depth the pipe should send
- * @crtc: CRTC structure
- * @mode: requested mode
- *
- * A pipe may be connected to one or more outputs.  Based on the depth of the
- * attached framebuffer, choose a good color depth to use on the pipe.
- *
- * If possible, match the pipe depth to the fb depth.  In some cases, this
- * isn't ideal, because the connected output supports a lesser or restricted
- * set of depths.  Resolve that here:
- *    LVDS typically supports only 6bpc, so clamp down in that case
- *    HDMI supports only 8bpc or 12bpc, so clamp to 8bpc with dither for 10bpc
- *    Displays may support a restricted set as well, check EDID and clamp as
- *      appropriate.
- *    DP may want to dither down to 6bpc to fit larger modes
- *
- * RETURNS:
- * Dithering requirement (i.e. false if display bpc and pipe bpc match,
- * true if they don't match).
- */
-static bool intel_choose_pipe_bpp_dither(struct drm_crtc *crtc,
-					 unsigned int *pipe_bpp,
-					 struct drm_display_mode *mode)
-{
-	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_encoder *encoder;
-	struct drm_connector *connector;
-	unsigned int display_bpc = UINT_MAX, bpc;
-
-	/* Walk the encoders & connectors on this crtc, get min bpc */
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
-		struct intel_encoder *intel_encoder = to_intel_encoder(encoder);
-
-		if (encoder->crtc != crtc)
-			continue;
-
-		if (intel_encoder->type == INTEL_OUTPUT_LVDS) {
-			unsigned int lvds_bpc;
-
-			if ((I915_READ(PCH_LVDS) & LVDS_A3_POWER_MASK) ==
-			    LVDS_A3_POWER_UP)
-				lvds_bpc = 8;
-			else
-				lvds_bpc = 6;
-
-			if (lvds_bpc < display_bpc) {
-				DRM_DEBUG_KMS("clamping display bpc (was %d) to LVDS (%d)\n", display_bpc, lvds_bpc);
-				display_bpc = lvds_bpc;
-			}
-			continue;
-		}
-
-		if (intel_encoder->type == INTEL_OUTPUT_EDP) {
-			/* Use VBT settings if we have an eDP panel */
-			unsigned int edp_bpc = dev_priv->edp.bpp / 3;
-
-			if (edp_bpc < display_bpc) {
-				DRM_DEBUG_KMS("clamping display bpc (was %d) to eDP (%d)\n", display_bpc, edp_bpc);
-				display_bpc = edp_bpc;
-			}
-			continue;
-		}
-
-		/* Not one of the known troublemakers, check the EDID */
-		list_for_each_entry(connector, &dev->mode_config.connector_list,
-				    head) {
-			if (connector->encoder != encoder)
-				continue;
-
-			/* Don't use an invalid EDID bpc value */
-			if (connector->display_info.bpc &&
-			    connector->display_info.bpc < display_bpc) {
-				DRM_DEBUG_KMS("clamping display bpc (was %d) to EDID reported max of %d\n", display_bpc, connector->display_info.bpc);
-				display_bpc = connector->display_info.bpc;
-			}
-		}
-
-		/*
-		 * HDMI is either 12 or 8, so if the display lets 10bpc sneak
-		 * through, clamp it down.  (Note: >12bpc will be caught below.)
-		 */
-		if (intel_encoder->type == INTEL_OUTPUT_HDMI) {
-			if (display_bpc > 8 && display_bpc < 12) {
-				DRM_DEBUG_KMS("forcing bpc to 12 for HDMI\n");
-				display_bpc = 12;
-			} else {
-				DRM_DEBUG_KMS("forcing bpc to 8 for HDMI\n");
-				display_bpc = 8;
-			}
-		}
-	}
-
-	if (mode->private_flags & INTEL_MODE_DP_FORCE_6BPC) {
-		DRM_DEBUG_KMS("Dithering DP to 6bpc\n");
-		display_bpc = 6;
-	}
-
-	/*
-	 * We could just drive the pipe at the highest bpc all the time and
-	 * enable dithering as needed, but that costs bandwidth.  So choose
-	 * the minimum value that expresses the full color range of the fb but
-	 * also stays within the max display bpc discovered above.
-	 */
-
-	switch (crtc->fb->depth) {
-	case 8:
-		bpc = 8; /* since we go through a colormap */
-		break;
-	case 15:
-	case 16:
-		bpc = 6; /* min is 18bpp */
-		break;
-	case 24:
-		bpc = 8;
-		break;
-	case 30:
-		bpc = 10;
-		break;
-	case 48:
-		bpc = 12;
-		break;
-	default:
-		DRM_DEBUG("unsupported depth, assuming 24 bits\n");
-		bpc = min((unsigned int)8, display_bpc);
-		break;
-	}
-
-	display_bpc = min(display_bpc, bpc);
-
-	DRM_DEBUG_KMS("setting pipe bpc to %d (max display bpc %d)\n",
-		      bpc, display_bpc);
-
-	*pipe_bpp = display_bpc * 3;
-
-	return display_bpc != bpc;
-}
-
-static int i9xx_get_refclk(struct drm_crtc *crtc, int num_connectors)
-{
-	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	int refclk;
-
-	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS) &&
-	    intel_panel_use_ssc(dev_priv) && num_connectors < 2) {
-		refclk = dev_priv->lvds_ssc_freq * 1000;
-		DRM_DEBUG_KMS("using SSC reference clock of %d MHz\n",
-			      refclk / 1000);
-	} else if (!IS_GEN2(dev)) {
-		refclk = 96000;
-	} else {
-		refclk = 48000;
-	}
-
-	return refclk;
-}
-
-static void i9xx_adjust_sdvo_tv_clock(struct drm_display_mode *adjusted_mode,
-				      intel_clock_t *clock)
-{
-	/* SDVO TV has fixed PLL values depend on its clock range,
-	   this mirrors vbios setting. */
-	if (adjusted_mode->clock >= 100000
-	    && adjusted_mode->clock < 140500) {
-		clock->p1 = 2;
-		clock->p2 = 10;
-		clock->n = 3;
-		clock->m1 = 16;
-		clock->m2 = 8;
-	} else if (adjusted_mode->clock >= 140500
-		   && adjusted_mode->clock <= 200000) {
-		clock->p1 = 1;
-		clock->p2 = 10;
-		clock->n = 6;
-		clock->m1 = 12;
-		clock->m2 = 8;
-	}
-}
-
-static void i9xx_update_pll_dividers(struct drm_crtc *crtc,
-				     intel_clock_t *clock,
-				     intel_clock_t *reduced_clock)
-{
-	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int pipe = intel_crtc->pipe;
-	u32 fp, fp2 = 0;
-
-	if (IS_PINEVIEW(dev)) {
-		fp = (1 << clock->n) << 16 | clock->m1 << 8 | clock->m2;
-		if (reduced_clock)
-			fp2 = (1 << reduced_clock->n) << 16 |
-				reduced_clock->m1 << 8 | reduced_clock->m2;
-	} else {
-		fp = clock->n << 16 | clock->m1 << 8 | clock->m2;
-		if (reduced_clock)
-			fp2 = reduced_clock->n << 16 | reduced_clock->m1 << 8 |
-				reduced_clock->m2;
-	}
-
-	I915_WRITE(FP0(pipe), fp);
-
-	intel_crtc->lowfreq_avail = false;
-	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_LVDS) &&
-	    reduced_clock && i915_powersave) {
-		I915_WRITE(FP1(pipe), fp2);
-		intel_crtc->lowfreq_avail = true;
-	} else {
-		I915_WRITE(FP1(pipe), fp);
-	}
-}
-
 static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 			      struct drm_display_mode *mode,
 			      struct drm_display_mode *adjusted_mode,
@@ -5138,7 +4340,7 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 	int plane = intel_crtc->plane;
 	int refclk, num_connectors = 0;
 	intel_clock_t clock, reduced_clock;
-	u32 dpll, dspcntr, pipeconf, vsyncshift;
+	u32 dpll, fp = 0, fp2 = 0, dspcntr, pipeconf;
 	bool ok, has_reduced_clock = false, is_sdvo = false, is_dvo = false;
 	bool is_crt = false, is_lvds = false, is_tv = false, is_dp = false;
 	struct drm_mode_config *mode_config = &dev->mode_config;
@@ -5179,7 +4381,15 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 		num_connectors++;
 	}
 
-	refclk = i9xx_get_refclk(crtc, num_connectors);
+	if (is_lvds && intel_panel_use_ssc(dev_priv) && num_connectors < 2) {
+		refclk = dev_priv->lvds_ssc_freq * 1000;
+		DRM_DEBUG_KMS("using SSC reference clock of %d MHz\n",
+			      refclk / 1000);
+	} else if (!IS_GEN2(dev)) {
+		refclk = 96000;
+	} else {
+		refclk = 48000;
+	}
 
 	/*
 	 * Returns a set of divisors for the desired target clock with the given
@@ -5187,8 +4397,7 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 	 * reflck * (5 * (m1 + 2) + (m2 + 2)) / (n + 2) / p1 / p2.
 	 */
 	limit = intel_limit(crtc, refclk);
-	ok = limit->find_pll(limit, crtc, adjusted_mode->clock, refclk, NULL,
-			     &clock);
+	ok = limit->find_pll(limit, crtc, adjusted_mode->clock, refclk, &clock);
 	if (!ok) {
 		DRM_ERROR("Couldn't find PLL settings for mode!\n");
 		return -EINVAL;
@@ -5198,24 +4407,53 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 	intel_crtc_update_cursor(crtc, true);
 
 	if (is_lvds && dev_priv->lvds_downclock_avail) {
-		/*
-		 * Ensure we match the reduced clock's P to the target clock.
-		 * If the clocks don't match, we can't switch the display clock
-		 * by using the FP0/FP1. In such case we will disable the LVDS
-		 * downclock feature.
-		*/
 		has_reduced_clock = limit->find_pll(limit, crtc,
 						    dev_priv->lvds_downclock,
 						    refclk,
-						    &clock,
 						    &reduced_clock);
+		if (has_reduced_clock && (clock.p != reduced_clock.p)) {
+			/*
+			 * If the different P is found, it means that we can't
+			 * switch the display clock by using the FP0/FP1.
+			 * In such case we will disable the LVDS downclock
+			 * feature.
+			 */
+			DRM_DEBUG_KMS("Different P is found for "
+				      "LVDS clock/downclock\n");
+			has_reduced_clock = 0;
+		}
+	}
+	/* SDVO TV has fixed PLL values depend on its clock range,
+	   this mirrors vbios setting. */
+	if (is_sdvo && is_tv) {
+		if (adjusted_mode->clock >= 100000
+		    && adjusted_mode->clock < 140500) {
+			clock.p1 = 2;
+			clock.p2 = 10;
+			clock.n = 3;
+			clock.m1 = 16;
+			clock.m2 = 8;
+		} else if (adjusted_mode->clock >= 140500
+			   && adjusted_mode->clock <= 200000) {
+			clock.p1 = 1;
+			clock.p2 = 10;
+			clock.n = 6;
+			clock.m1 = 12;
+			clock.m2 = 8;
+		}
 	}
 
-	if (is_sdvo && is_tv)
-		i9xx_adjust_sdvo_tv_clock(adjusted_mode, &clock);
-
-	i9xx_update_pll_dividers(crtc, &clock, has_reduced_clock ?
-				 &reduced_clock : NULL);
+	if (IS_PINEVIEW(dev)) {
+		fp = (1 << clock.n) << 16 | clock.m1 << 8 | clock.m2;
+		if (has_reduced_clock)
+			fp2 = (1 << reduced_clock.n) << 16 |
+				reduced_clock.m1 << 8 | reduced_clock.m2;
+	} else {
+		fp = clock.n << 16 | clock.m1 << 8 | clock.m2;
+		if (has_reduced_clock)
+			fp2 = reduced_clock.n << 16 | reduced_clock.m1 << 8 |
+				reduced_clock.m2;
+	}
 
 	dpll = DPLL_VGA_MODE_DIS;
 
@@ -5289,6 +4527,8 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 	/* Set up the display plane register */
 	dspcntr = DISPPLANE_GAMMA_ENABLE;
 
+	/* Ironlake's plane is forced to pipe, bit 24 is to
+	   enable color space conversion */
 	if (pipe == 0)
 		dspcntr &= ~DISPPLANE_SEL_PIPE_MASK;
 	else
@@ -5308,21 +4548,12 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 			pipeconf &= ~PIPECONF_DOUBLE_WIDE;
 	}
 
-	/* default to 8bpc */
-	pipeconf &= ~(PIPECONF_BPP_MASK | PIPECONF_DITHER_EN);
-	if (is_dp) {
-		if (mode->private_flags & INTEL_MODE_DP_FORCE_6BPC) {
-			pipeconf |= PIPECONF_BPP_6 |
-				    PIPECONF_DITHER_EN |
-				    PIPECONF_DITHER_TYPE_SP;
-		}
-	}
-
 	dpll |= DPLL_VCO_ENABLE;
 
 	DRM_DEBUG_KMS("Mode for pipe %c:\n", pipe == 0 ? 'A' : 'B');
 	drm_mode_debug_printmodeline(mode);
 
+	I915_WRITE(FP0(pipe), fp);
 	I915_WRITE(DPLL(pipe), dpll & ~DPLL_VCO_ENABLE);
 
 	POSTING_READ(DPLL(pipe));
@@ -5409,32 +4640,33 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 		I915_WRITE(DPLL(pipe), dpll);
 	}
 
-	if (HAS_PIPE_CXSR(dev)) {
-		if (intel_crtc->lowfreq_avail) {
+	intel_crtc->lowfreq_avail = false;
+	if (is_lvds && has_reduced_clock && i915_powersave) {
+		I915_WRITE(FP1(pipe), fp2);
+		intel_crtc->lowfreq_avail = true;
+		if (HAS_PIPE_CXSR(dev)) {
 			DRM_DEBUG_KMS("enabling CxSR downclocking\n");
 			pipeconf |= PIPECONF_CXSR_DOWNCLOCK;
-		} else {
+		}
+	} else {
+		I915_WRITE(FP1(pipe), fp);
+		if (HAS_PIPE_CXSR(dev)) {
 			DRM_DEBUG_KMS("disabling CxSR downclocking\n");
 			pipeconf &= ~PIPECONF_CXSR_DOWNCLOCK;
 		}
 	}
 
-	pipeconf &= ~PIPECONF_INTERLACE_MASK;
-	if (!IS_GEN2(dev) &&
-	    adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) {
+	if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) {
 		pipeconf |= PIPECONF_INTERLACE_W_FIELD_INDICATION;
 		/* the chip adds 2 halflines automatically */
+		adjusted_mode->crtc_vdisplay -= 1;
 		adjusted_mode->crtc_vtotal -= 1;
+		adjusted_mode->crtc_vblank_start -= 1;
 		adjusted_mode->crtc_vblank_end -= 1;
-		vsyncshift = adjusted_mode->crtc_hsync_start
-			     - adjusted_mode->crtc_htotal/2;
-	} else {
-		pipeconf |= PIPECONF_PROGRESSIVE;
-		vsyncshift = 0;
-	}
-
-	if (!IS_GEN3(dev))
-		I915_WRITE(VSYNCSHIFT(pipe), vsyncshift);
+		adjusted_mode->crtc_vsync_end -= 1;
+		adjusted_mode->crtc_vsync_start -= 1;
+	} else
+		pipeconf &= ~PIPECONF_INTERLACE_W_FIELD_INDICATION; /* progressive */
 
 	I915_WRITE(HTOTAL(pipe),
 		   (adjusted_mode->crtc_hdisplay - 1) |
@@ -5483,156 +4715,6 @@ static int i9xx_crtc_mode_set(struct drm_crtc *crtc,
 	return ret;
 }
 
-/*
- * Initialize reference clocks when the driver loads
- */
-void ironlake_init_pch_refclk(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_mode_config *mode_config = &dev->mode_config;
-	struct intel_encoder *encoder;
-	u32 temp;
-	bool has_lvds = false;
-	bool has_cpu_edp = false;
-	bool has_pch_edp = false;
-	bool has_panel = false;
-	bool has_ck505 = false;
-	bool can_ssc = false;
-
-	/* We need to take the global config into account */
-	list_for_each_entry(encoder, &mode_config->encoder_list,
-			    base.head) {
-		switch (encoder->type) {
-		case INTEL_OUTPUT_LVDS:
-			has_panel = true;
-			has_lvds = true;
-			break;
-		case INTEL_OUTPUT_EDP:
-			has_panel = true;
-			if (intel_encoder_is_pch_edp(&encoder->base))
-				has_pch_edp = true;
-			else
-				has_cpu_edp = true;
-			break;
-		}
-	}
-
-	if (HAS_PCH_IBX(dev)) {
-		has_ck505 = dev_priv->display_clock_mode;
-		can_ssc = has_ck505;
-	} else {
-		has_ck505 = false;
-		can_ssc = true;
-	}
-
-	DRM_DEBUG_KMS("has_panel %d has_lvds %d has_pch_edp %d has_cpu_edp %d has_ck505 %d\n",
-		      has_panel, has_lvds, has_pch_edp, has_cpu_edp,
-		      has_ck505);
-
-	/* Ironlake: try to setup display ref clock before DPLL
-	 * enabling. This is only under driver's control after
-	 * PCH B stepping, previous chipset stepping should be
-	 * ignoring this setting.
-	 */
-	temp = I915_READ(PCH_DREF_CONTROL);
-	/* Always enable nonspread source */
-	temp &= ~DREF_NONSPREAD_SOURCE_MASK;
-
-	if (has_ck505)
-		temp |= DREF_NONSPREAD_CK505_ENABLE;
-	else
-		temp |= DREF_NONSPREAD_SOURCE_ENABLE;
-
-	if (has_panel) {
-		temp &= ~DREF_SSC_SOURCE_MASK;
-		temp |= DREF_SSC_SOURCE_ENABLE;
-
-		/* SSC must be turned on before enabling the CPU output  */
-		if (intel_panel_use_ssc(dev_priv) && can_ssc) {
-			DRM_DEBUG_KMS("Using SSC on panel\n");
-			temp |= DREF_SSC1_ENABLE;
-		}
-
-		/* Get SSC going before enabling the outputs */
-		I915_WRITE(PCH_DREF_CONTROL, temp);
-		POSTING_READ(PCH_DREF_CONTROL);
-		udelay(200);
-
-		temp &= ~DREF_CPU_SOURCE_OUTPUT_MASK;
-
-		/* Enable CPU source on CPU attached eDP */
-		if (has_cpu_edp) {
-			if (intel_panel_use_ssc(dev_priv) && can_ssc) {
-				DRM_DEBUG_KMS("Using SSC on eDP\n");
-				temp |= DREF_CPU_SOURCE_OUTPUT_DOWNSPREAD;
-			}
-			else
-				temp |= DREF_CPU_SOURCE_OUTPUT_NONSPREAD;
-		} else
-			temp |= DREF_CPU_SOURCE_OUTPUT_DISABLE;
-
-		I915_WRITE(PCH_DREF_CONTROL, temp);
-		POSTING_READ(PCH_DREF_CONTROL);
-		udelay(200);
-	} else {
-		DRM_DEBUG_KMS("Disabling SSC entirely\n");
-
-		temp &= ~DREF_CPU_SOURCE_OUTPUT_MASK;
-
-		/* Turn off CPU output */
-		temp |= DREF_CPU_SOURCE_OUTPUT_DISABLE;
-
-		I915_WRITE(PCH_DREF_CONTROL, temp);
-		POSTING_READ(PCH_DREF_CONTROL);
-		udelay(200);
-
-		/* Turn off the SSC source */
-		temp &= ~DREF_SSC_SOURCE_MASK;
-		temp |= DREF_SSC_SOURCE_DISABLE;
-
-		/* Turn off SSC1 */
-		temp &= ~ DREF_SSC1_ENABLE;
-
-		I915_WRITE(PCH_DREF_CONTROL, temp);
-		POSTING_READ(PCH_DREF_CONTROL);
-		udelay(200);
-	}
-}
-
-static int ironlake_get_refclk(struct drm_crtc *crtc)
-{
-	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_encoder *encoder;
-	struct drm_mode_config *mode_config = &dev->mode_config;
-	struct intel_encoder *edp_encoder = NULL;
-	int num_connectors = 0;
-	bool is_lvds = false;
-
-	list_for_each_entry(encoder, &mode_config->encoder_list, base.head) {
-		if (encoder->base.crtc != crtc)
-			continue;
-
-		switch (encoder->type) {
-		case INTEL_OUTPUT_LVDS:
-			is_lvds = true;
-			break;
-		case INTEL_OUTPUT_EDP:
-			edp_encoder = encoder;
-			break;
-		}
-		num_connectors++;
-	}
-
-	if (is_lvds && intel_panel_use_ssc(dev_priv) && num_connectors < 2) {
-		DRM_DEBUG_KMS("using SSC reference clock of %d MHz\n",
-			      dev_priv->lvds_ssc_freq);
-		return dev_priv->lvds_ssc_freq * 1000;
-	}
-
-	return 120000;
-}
-
 static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 				  struct drm_display_mode *mode,
 				  struct drm_display_mode *adjusted_mode,
@@ -5657,9 +4739,7 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 	struct fdi_m_n m_n = {0};
 	u32 temp;
 	u32 lvds_sync = 0;
-	int target_clock, pixel_multiplier, lane, link_bw, factor;
-	unsigned int pipe_bpp;
-	bool dither;
+	int target_clock, pixel_multiplier, lane, link_bw, bpp, factor;
 
 	list_for_each_entry(encoder, &mode_config->encoder_list, base.head) {
 		if (encoder->base.crtc != crtc)
@@ -5692,7 +4772,16 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 		num_connectors++;
 	}
 
-	refclk = ironlake_get_refclk(crtc);
+	if (is_lvds && intel_panel_use_ssc(dev_priv) && num_connectors < 2) {
+		refclk = dev_priv->lvds_ssc_freq * 1000;
+		DRM_DEBUG_KMS("using SSC reference clock of %d MHz\n",
+			      refclk / 1000);
+	} else {
+		refclk = 96000;
+		if (!has_edp_encoder ||
+		    intel_encoder_is_pch_edp(&has_edp_encoder->base))
+			refclk = 120000; /* 120Mhz refclk */
+	}
 
 	/*
 	 * Returns a set of divisors for the desired target clock with the given
@@ -5700,8 +4789,7 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 	 * reflck * (5 * (m1 + 2) + (m2 + 2)) / (n + 2) / p1 / p2.
 	 */
 	limit = intel_limit(crtc, refclk);
-	ok = limit->find_pll(limit, crtc, adjusted_mode->clock, refclk, NULL,
-			     &clock);
+	ok = limit->find_pll(limit, crtc, adjusted_mode->clock, refclk, &clock);
 	if (!ok) {
 		DRM_ERROR("Couldn't find PLL settings for mode!\n");
 		return -EINVAL;
@@ -5711,17 +4799,21 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 	intel_crtc_update_cursor(crtc, true);
 
 	if (is_lvds && dev_priv->lvds_downclock_avail) {
-		/*
-		 * Ensure we match the reduced clock's P to the target clock.
-		 * If the clocks don't match, we can't switch the display clock
-		 * by using the FP0/FP1. In such case we will disable the LVDS
-		 * downclock feature.
-		*/
 		has_reduced_clock = limit->find_pll(limit, crtc,
 						    dev_priv->lvds_downclock,
 						    refclk,
-						    &clock,
 						    &reduced_clock);
+		if (has_reduced_clock && (clock.p != reduced_clock.p)) {
+			/*
+			 * If the different P is found, it means that we can't
+			 * switch the display clock by using the FP0/FP1.
+			 * In such case we will disable the LVDS downclock
+			 * feature.
+			 */
+			DRM_DEBUG_KMS("Different P is found for "
+				      "LVDS clock/downclock\n");
+			has_reduced_clock = 0;
+		}
 	}
 	/* SDVO TV has fixed PLL values depend on its clock range,
 	   this mirrors vbios setting. */
@@ -5774,38 +4866,56 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 	/* determine panel color depth */
 	temp = I915_READ(PIPECONF(pipe));
 	temp &= ~PIPE_BPC_MASK;
-	dither = intel_choose_pipe_bpp_dither(crtc, &pipe_bpp, mode);
-	switch (pipe_bpp) {
-	case 18:
-		temp |= PIPE_6BPC;
-		break;
-	case 24:
+	if (is_lvds) {
+		/* the BPC will be 6 if it is 18-bit LVDS panel */
+		if ((I915_READ(PCH_LVDS) & LVDS_A3_POWER_MASK) == LVDS_A3_POWER_UP)
+			temp |= PIPE_8BPC;
+		else
+			temp |= PIPE_6BPC;
+	} else if (has_edp_encoder) {
+		switch (dev_priv->edp.bpp/3) {
+		case 8:
+			temp |= PIPE_8BPC;
+			break;
+		case 10:
+			temp |= PIPE_10BPC;
+			break;
+		case 6:
+			temp |= PIPE_6BPC;
+			break;
+		case 12:
+			temp |= PIPE_12BPC;
+			break;
+		}
+	} else
 		temp |= PIPE_8BPC;
+	I915_WRITE(PIPECONF(pipe), temp);
+
+	switch (temp & PIPE_BPC_MASK) {
+	case PIPE_8BPC:
+		bpp = 24;
 		break;
-	case 30:
-		temp |= PIPE_10BPC;
+	case PIPE_10BPC:
+		bpp = 30;
 		break;
-	case 36:
-		temp |= PIPE_12BPC;
+	case PIPE_6BPC:
+		bpp = 18;
 		break;
-	default:
-		WARN(1, "intel_choose_pipe_bpp returned invalid value %d\n",
-			pipe_bpp);
-		temp |= PIPE_8BPC;
-		pipe_bpp = 24;
+	case PIPE_12BPC:
+		bpp = 36;
 		break;
+	default:
+		DRM_ERROR("unknown pipe bpc value\n");
+		bpp = 24;
 	}
 
-	intel_crtc->bpp = pipe_bpp;
-	I915_WRITE(PIPECONF(pipe), temp);
-
 	if (!lane) {
 		/*
 		 * Account for spread spectrum to avoid
 		 * oversubscribing the link. Max center spread
 		 * is 2.5%; use 5% for safety's sake.
 		 */
-		u32 bps = target_clock * intel_crtc->bpp * 21 / 20;
+		u32 bps = target_clock * bpp * 21 / 20;
 		lane = bps / (link_bw * 8) + 1;
 	}
 
@@ -5813,8 +4923,51 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 
 	if (pixel_multiplier > 1)
 		link_bw *= pixel_multiplier;
-	ironlake_compute_m_n(intel_crtc->bpp, lane, target_clock, link_bw,
-			     &m_n);
+	ironlake_compute_m_n(bpp, lane, target_clock, link_bw, &m_n);
+
+	/* Ironlake: try to setup display ref clock before DPLL
+	 * enabling. This is only under driver's control after
+	 * PCH B stepping, previous chipset stepping should be
+	 * ignoring this setting.
+	 */
+	temp = I915_READ(PCH_DREF_CONTROL);
+	/* Always enable nonspread source */
+	temp &= ~DREF_NONSPREAD_SOURCE_MASK;
+	temp |= DREF_NONSPREAD_SOURCE_ENABLE;
+	temp &= ~DREF_SSC_SOURCE_MASK;
+	temp |= DREF_SSC_SOURCE_ENABLE;
+	I915_WRITE(PCH_DREF_CONTROL, temp);
+
+	POSTING_READ(PCH_DREF_CONTROL);
+	udelay(200);
+
+	if (has_edp_encoder) {
+		if (intel_panel_use_ssc(dev_priv)) {
+			temp |= DREF_SSC1_ENABLE;
+			I915_WRITE(PCH_DREF_CONTROL, temp);
+
+			POSTING_READ(PCH_DREF_CONTROL);
+			udelay(200);
+		}
+		temp &= ~DREF_CPU_SOURCE_OUTPUT_MASK;
+
+		/* Enable CPU source on CPU attached eDP */
+		if (!intel_encoder_is_pch_edp(&has_edp_encoder->base)) {
+			if (intel_panel_use_ssc(dev_priv))
+				temp |= DREF_CPU_SOURCE_OUTPUT_DOWNSPREAD;
+			else
+				temp |= DREF_CPU_SOURCE_OUTPUT_NONSPREAD;
+		} else {
+			/* Enable SSC on PCH eDP if needed */
+			if (intel_panel_use_ssc(dev_priv)) {
+				DRM_ERROR("enabling SSC on PCH\n");
+				temp |= DREF_SUPERSPREAD_SOURCE_ENABLE;
+			}
+		}
+		I915_WRITE(PCH_DREF_CONTROL, temp);
+		POSTING_READ(PCH_DREF_CONTROL);
+		udelay(200);
+	}
 
 	fp = clock.n << 16 | clock.m1 << 8 | clock.m2;
 	if (has_reduced_clock)
@@ -5887,32 +5040,39 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 	/* Set up the display plane register */
 	dspcntr = DISPPLANE_GAMMA_ENABLE;
 
-	DRM_DEBUG_KMS("Mode for pipe %d:\n", pipe);
+	DRM_DEBUG_KMS("Mode for pipe %c:\n", pipe == 0 ? 'A' : 'B');
 	drm_mode_debug_printmodeline(mode);
 
 	/* PCH eDP needs FDI, but CPU eDP does not */
-	if (!intel_crtc->no_pll) {
-		if (!has_edp_encoder ||
-		    intel_encoder_is_pch_edp(&has_edp_encoder->base)) {
-			I915_WRITE(PCH_FP0(pipe), fp);
-			I915_WRITE(PCH_DPLL(pipe), dpll & ~DPLL_VCO_ENABLE);
+	if (!has_edp_encoder || intel_encoder_is_pch_edp(&has_edp_encoder->base)) {
+		I915_WRITE(PCH_FP0(pipe), fp);
+		I915_WRITE(PCH_DPLL(pipe), dpll & ~DPLL_VCO_ENABLE);
 
-			POSTING_READ(PCH_DPLL(pipe));
-			udelay(150);
-		}
-	} else {
-		if (dpll == (I915_READ(PCH_DPLL(0)) & 0x7fffffff) &&
-		    fp == I915_READ(PCH_FP0(0))) {
-			intel_crtc->use_pll_a = true;
-			DRM_DEBUG_KMS("using pipe a dpll\n");
-		} else if (dpll == (I915_READ(PCH_DPLL(1)) & 0x7fffffff) &&
-			   fp == I915_READ(PCH_FP0(1))) {
-			intel_crtc->use_pll_a = false;
-			DRM_DEBUG_KMS("using pipe b dpll\n");
-		} else {
-			DRM_DEBUG_KMS("no matching PLL configuration for pipe 2\n");
-			return -EINVAL;
+		POSTING_READ(PCH_DPLL(pipe));
+		udelay(150);
+	}
+
+	/* enable transcoder DPLL */
+	if (HAS_PCH_CPT(dev)) {
+		temp = I915_READ(PCH_DPLL_SEL);
+		switch (pipe) {
+		case 0:
+			temp |= TRANSA_DPLL_ENABLE | TRANSA_DPLLA_SEL;
+			break;
+		case 1:
+			temp |=	TRANSB_DPLL_ENABLE | TRANSB_DPLLB_SEL;
+			break;
+		case 2:
+			/* FIXME: manage transcoder PLLs? */
+			temp |= TRANSC_DPLL_ENABLE | TRANSC_DPLLB_SEL;
+			break;
+		default:
+			BUG();
 		}
+		I915_WRITE(PCH_DPLL_SEL, temp);
+
+		POSTING_READ(PCH_DPLL_SEL);
+		udelay(150);
 	}
 
 	/* The LVDS pin pair needs to be on before the DPLLs are enabled.
@@ -5922,16 +5082,17 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 	if (is_lvds) {
 		temp = I915_READ(PCH_LVDS);
 		temp |= LVDS_PORT_EN | LVDS_A0A2_CLKA_POWER_UP;
-		if (HAS_PCH_CPT(dev)) {
-			temp &= ~PORT_TRANS_SEL_MASK;
-			temp |= PORT_TRANS_SEL_CPT(pipe);
-		} else {
-			if (pipe == 1)
+		if (pipe == 1) {
+			if (HAS_PCH_CPT(dev))
+				temp |= PORT_TRANS_B_SEL_CPT;
+			else
 				temp |= LVDS_PIPEB_SELECT;
+		} else {
+			if (HAS_PCH_CPT(dev))
+				temp &= ~PORT_TRANS_SEL_MASK;
 			else
 				temp &= ~LVDS_PIPEB_SELECT;
 		}
-
 		/* set the corresponsding LVDS_BORDER bit */
 		temp |= dev_priv->lvds_border_bits;
 		/* Set the B0-B3 data pairs corresponding to whether we're going to
@@ -5965,12 +5126,14 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 		I915_WRITE(PCH_LVDS, temp);
 	}
 
+	/* set the dithering flag and clear for anything other than a panel. */
 	pipeconf &= ~PIPECONF_DITHER_EN;
 	pipeconf &= ~PIPECONF_DITHER_TYPE_MASK;
-	if ((is_lvds && dev_priv->lvds_dither) || dither) {
+	if (dev_priv->lvds_dither && (is_lvds || has_edp_encoder)) {
 		pipeconf |= PIPECONF_DITHER_EN;
-		pipeconf |= PIPECONF_DITHER_TYPE_SP;
+		pipeconf |= PIPECONF_DITHER_TYPE_ST1;
 	}
+
 	if (is_dp || intel_encoder_is_pch_edp(&has_edp_encoder->base)) {
 		intel_dp_set_m_n(crtc, mode, adjusted_mode);
 	} else {
@@ -5981,9 +5144,8 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 		I915_WRITE(TRANSDPLINK_N1(pipe), 0);
 	}
 
-	if (!intel_crtc->no_pll &&
-	    (!has_edp_encoder ||
-	     intel_encoder_is_pch_edp(&has_edp_encoder->base))) {
+	if (!has_edp_encoder ||
+	    intel_encoder_is_pch_edp(&has_edp_encoder->base)) {
 		I915_WRITE(PCH_DPLL(pipe), dpll);
 
 		/* Wait for the clocks to stabilize. */
@@ -5999,36 +5161,32 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 	}
 
 	intel_crtc->lowfreq_avail = false;
-	if (!intel_crtc->no_pll) {
-		if (is_lvds && has_reduced_clock && i915_powersave) {
-			I915_WRITE(PCH_FP1(pipe), fp2);
-			intel_crtc->lowfreq_avail = true;
-			if (HAS_PIPE_CXSR(dev)) {
-				DRM_DEBUG_KMS("enabling CxSR downclocking\n");
-				pipeconf |= PIPECONF_CXSR_DOWNCLOCK;
-			}
-		} else {
-			I915_WRITE(PCH_FP1(pipe), fp);
-			if (HAS_PIPE_CXSR(dev)) {
-				DRM_DEBUG_KMS("disabling CxSR downclocking\n");
-				pipeconf &= ~PIPECONF_CXSR_DOWNCLOCK;
-			}
+	if (is_lvds && has_reduced_clock && i915_powersave) {
+		I915_WRITE(PCH_FP1(pipe), fp2);
+		intel_crtc->lowfreq_avail = true;
+		if (HAS_PIPE_CXSR(dev)) {
+			DRM_DEBUG_KMS("enabling CxSR downclocking\n");
+			pipeconf |= PIPECONF_CXSR_DOWNCLOCK;
+		}
+	} else {
+		I915_WRITE(PCH_FP1(pipe), fp);
+		if (HAS_PIPE_CXSR(dev)) {
+			DRM_DEBUG_KMS("disabling CxSR downclocking\n");
+			pipeconf &= ~PIPECONF_CXSR_DOWNCLOCK;
 		}
 	}
 
-	pipeconf &= ~PIPECONF_INTERLACE_MASK;
 	if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) {
-		pipeconf |= PIPECONF_INTERLACED_ILK;
+		pipeconf |= PIPECONF_INTERLACE_W_FIELD_INDICATION;
 		/* the chip adds 2 halflines automatically */
+		adjusted_mode->crtc_vdisplay -= 1;
 		adjusted_mode->crtc_vtotal -= 1;
+		adjusted_mode->crtc_vblank_start -= 1;
 		adjusted_mode->crtc_vblank_end -= 1;
-		I915_WRITE(VSYNCSHIFT(pipe),
-			   adjusted_mode->crtc_hsync_start
-			   - adjusted_mode->crtc_htotal/2);
-	} else {
-		pipeconf |= PIPECONF_PROGRESSIVE;
-		I915_WRITE(VSYNCSHIFT(pipe), 0);
-	}
+		adjusted_mode->crtc_vsync_end -= 1;
+		adjusted_mode->crtc_vsync_start -= 1;
+	} else
+		pipeconf &= ~PIPECONF_INTERLACE_W_FIELD_INDICATION; /* progressive */
 
 	I915_WRITE(HTOTAL(pipe),
 		   (adjusted_mode->crtc_hdisplay - 1) |
@@ -6071,6 +5229,12 @@ static int ironlake_crtc_mode_set(struct drm_crtc *crtc,
 
 	intel_wait_for_vblank(dev, pipe);
 
+	if (IS_GEN5(dev)) {
+		/* enable address swizzle for tiling buffer */
+		temp = I915_READ(DISP_ARB_CTL);
+		I915_WRITE(DISP_ARB_CTL, temp | DISP_TILE_SURFACE_SWIZZLING);
+	}
+
 	I915_WRITE(DSPCNTR(plane), dspcntr);
 	POSTING_READ(DSPCNTR(plane));
 
@@ -6097,188 +5261,12 @@ static int intel_crtc_mode_set(struct drm_crtc *crtc,
 
 	ret = dev_priv->display.crtc_mode_set(crtc, mode, adjusted_mode,
 					      x, y, old_fb);
-	drm_vblank_post_modeset(dev, pipe);
 
-	if (ret)
-		intel_crtc->dpms_mode = DRM_MODE_DPMS_OFF;
-	else
-		intel_crtc->dpms_mode = DRM_MODE_DPMS_ON;
+	drm_vblank_post_modeset(dev, pipe);
 
 	return ret;
 }
 
-static bool intel_eld_uptodate(struct drm_connector *connector,
-			       int reg_eldv, uint32_t bits_eldv,
-			       int reg_elda, uint32_t bits_elda,
-			       int reg_edid)
-{
-	struct drm_i915_private *dev_priv = connector->dev->dev_private;
-	uint8_t *eld = connector->eld;
-	uint32_t i;
-
-	i = I915_READ(reg_eldv);
-	i &= bits_eldv;
-
-	if (!eld[0])
-		return !i;
-
-	if (!i)
-		return false;
-
-	i = I915_READ(reg_elda);
-	i &= ~bits_elda;
-	I915_WRITE(reg_elda, i);
-
-	for (i = 0; i < eld[2]; i++)
-		if (I915_READ(reg_edid) != *((uint32_t *)eld + i))
-			return false;
-
-	return true;
-}
-
-static void g4x_write_eld(struct drm_connector *connector,
-			  struct drm_crtc *crtc)
-{
-	struct drm_i915_private *dev_priv = connector->dev->dev_private;
-	uint8_t *eld = connector->eld;
-	uint32_t eldv;
-	uint32_t len;
-	uint32_t i;
-
-	i = I915_READ(G4X_AUD_VID_DID);
-
-	if (i == INTEL_AUDIO_DEVBLC || i == INTEL_AUDIO_DEVCL)
-		eldv = G4X_ELDV_DEVCL_DEVBLC;
-	else
-		eldv = G4X_ELDV_DEVCTG;
-
-	if (intel_eld_uptodate(connector,
-			       G4X_AUD_CNTL_ST, eldv,
-			       G4X_AUD_CNTL_ST, G4X_ELD_ADDR,
-			       G4X_HDMIW_HDMIEDID))
-		return;
-
-	i = I915_READ(G4X_AUD_CNTL_ST);
-	i &= ~(eldv | G4X_ELD_ADDR);
-	len = (i >> 9) & 0x1f;		/* ELD buffer size */
-	I915_WRITE(G4X_AUD_CNTL_ST, i);
-
-	if (!eld[0])
-		return;
-
-	len = min_t(uint8_t, eld[2], len);
-	DRM_DEBUG_DRIVER("ELD size %d\n", len);
-	for (i = 0; i < len; i++)
-		I915_WRITE(G4X_HDMIW_HDMIEDID, *((uint32_t *)eld + i));
-
-	i = I915_READ(G4X_AUD_CNTL_ST);
-	i |= eldv;
-	I915_WRITE(G4X_AUD_CNTL_ST, i);
-}
-
-static void ironlake_write_eld(struct drm_connector *connector,
-				     struct drm_crtc *crtc)
-{
-	struct drm_i915_private *dev_priv = connector->dev->dev_private;
-	uint8_t *eld = connector->eld;
-	uint32_t eldv;
-	uint32_t i;
-	int len;
-	int hdmiw_hdmiedid;
-	int aud_config;
-	int aud_cntl_st;
-	int aud_cntrl_st2;
-
-	if (HAS_PCH_IBX(connector->dev)) {
-		hdmiw_hdmiedid = IBX_HDMIW_HDMIEDID_A;
-		aud_config = IBX_AUD_CONFIG_A;
-		aud_cntl_st = IBX_AUD_CNTL_ST_A;
-		aud_cntrl_st2 = IBX_AUD_CNTL_ST2;
-	} else {
-		hdmiw_hdmiedid = CPT_HDMIW_HDMIEDID_A;
-		aud_config = CPT_AUD_CONFIG_A;
-		aud_cntl_st = CPT_AUD_CNTL_ST_A;
-		aud_cntrl_st2 = CPT_AUD_CNTRL_ST2;
-	}
-
-	i = to_intel_crtc(crtc)->pipe;
-	hdmiw_hdmiedid += i * 0x100;
-	aud_cntl_st += i * 0x100;
-	aud_config += i * 0x100;
-
-	DRM_DEBUG_DRIVER("ELD on pipe %c\n", pipe_name(i));
-
-	i = I915_READ(aud_cntl_st);
-	i = (i >> 29) & 0x3;		/* DIP_Port_Select, 0x1 = PortB */
-	if (!i) {
-		DRM_DEBUG_DRIVER("Audio directed to unknown port\n");
-		/* operate blindly on all ports */
-		eldv = IBX_ELD_VALIDB;
-		eldv |= IBX_ELD_VALIDB << 4;
-		eldv |= IBX_ELD_VALIDB << 8;
-	} else {
-		DRM_DEBUG_DRIVER("ELD on port %c\n", 'A' + i);
-		eldv = IBX_ELD_VALIDB << ((i - 1) * 4);
-	}
-
-	if (intel_pipe_has_type(crtc, INTEL_OUTPUT_DISPLAYPORT)) {
-		DRM_DEBUG_DRIVER("ELD: DisplayPort detected\n");
-		eld[5] |= (1 << 2);	/* Conn_Type, 0x1 = DisplayPort */
-		I915_WRITE(aud_config, AUD_CONFIG_N_VALUE_INDEX); /* 0x1 = DP */
-	} else
-		I915_WRITE(aud_config, 0);
-
-	if (intel_eld_uptodate(connector,
-			       aud_cntrl_st2, eldv,
-			       aud_cntl_st, IBX_ELD_ADDRESS,
-			       hdmiw_hdmiedid))
-		return;
-
-	i = I915_READ(aud_cntrl_st2);
-	i &= ~eldv;
-	I915_WRITE(aud_cntrl_st2, i);
-
-	if (!eld[0])
-		return;
-
-	i = I915_READ(aud_cntl_st);
-	i &= ~IBX_ELD_ADDRESS;
-	I915_WRITE(aud_cntl_st, i);
-
-	len = min_t(uint8_t, eld[2], 21);	/* 84 bytes of hw ELD buffer */
-	DRM_DEBUG_DRIVER("ELD size %d\n", len);
-	for (i = 0; i < len; i++)
-		I915_WRITE(hdmiw_hdmiedid, *((uint32_t *)eld + i));
-
-	i = I915_READ(aud_cntrl_st2);
-	i |= eldv;
-	I915_WRITE(aud_cntrl_st2, i);
-}
-
-void intel_write_eld(struct drm_encoder *encoder,
-		     struct drm_display_mode *mode)
-{
-	struct drm_crtc *crtc = encoder->crtc;
-	struct drm_connector *connector;
-	struct drm_device *dev = encoder->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	connector = drm_select_eld(encoder, mode);
-	if (!connector)
-		return;
-
-	DRM_DEBUG_DRIVER("ELD on [CONNECTOR:%d:%s], [ENCODER:%d:%s]\n",
-			 connector->base.id,
-			 drm_get_connector_name(connector),
-			 connector->encoder->base.id,
-			 drm_get_encoder_name(connector->encoder));
-
-	connector->eld[6] = drm_av_sync_delay(connector, mode) / 2;
-
-	if (dev_priv->display.write_eld)
-		dev_priv->display.write_eld(connector, crtc);
-}
-
 /** Loads the palette/gamma unit for the CRTC with the prepared values */
 void intel_crtc_load_lut(struct drm_crtc *crtc)
 {
@@ -6495,15 +5483,21 @@ static int intel_crtc_cursor_set(struct drm_crtc *crtc,
 			goto fail_locked;
 		}
 
-		ret = i915_gem_object_pin_to_display_plane(obj, 0, NULL);
+		ret = i915_gem_object_pin(obj, PAGE_SIZE, true);
 		if (ret) {
-			DRM_ERROR("failed to move cursor bo into the GTT\n");
+			DRM_ERROR("failed to pin cursor bo\n");
 			goto fail_locked;
 		}
 
+		ret = i915_gem_object_set_to_gtt_domain(obj, 0);
+		if (ret) {
+			DRM_ERROR("failed to move cursor bo into the GTT\n");
+			goto fail_unpin;
+		}
+
 		ret = i915_gem_object_put_fence(obj);
 		if (ret) {
-			DRM_ERROR("failed to release fence for cursor");
+			DRM_ERROR("failed to move cursor bo into the GTT\n");
 			goto fail_unpin;
 		}
 
@@ -6622,7 +5616,7 @@ static struct drm_display_mode load_detect_mode = {
 
 static struct drm_framebuffer *
 intel_framebuffer_create(struct drm_device *dev,
-			 struct drm_mode_fb_cmd2 *mode_cmd,
+			 struct drm_mode_fb_cmd *mode_cmd,
 			 struct drm_i915_gem_object *obj)
 {
 	struct intel_framebuffer *intel_fb;
@@ -6664,7 +5658,7 @@ intel_framebuffer_create_for_mode(struct drm_device *dev,
 				  int depth, int bpp)
 {
 	struct drm_i915_gem_object *obj;
-	struct drm_mode_fb_cmd2 mode_cmd;
+	struct drm_mode_fb_cmd mode_cmd;
 
 	obj = i915_gem_alloc_object(dev,
 				    intel_framebuffer_size_for_mode(mode, bpp));
@@ -6673,9 +5667,9 @@ intel_framebuffer_create_for_mode(struct drm_device *dev,
 
 	mode_cmd.width = mode->hdisplay;
 	mode_cmd.height = mode->vdisplay;
-	mode_cmd.pitches[0] = intel_framebuffer_pitch_for_width(mode_cmd.width,
-								bpp);
-	mode_cmd.pixel_format = drm_mode_legacy_fb_format(bpp, depth);
+	mode_cmd.depth = depth;
+	mode_cmd.bpp = bpp;
+	mode_cmd.pitch = intel_framebuffer_pitch_for_width(mode_cmd.width, bpp);
 
 	return intel_framebuffer_create(dev, &mode_cmd, obj);
 }
@@ -6696,11 +5690,11 @@ mode_fits_in_fbdev(struct drm_device *dev,
 		return NULL;
 
 	fb = &dev_priv->fbdev->ifb.base;
-	if (fb->pitches[0] < intel_framebuffer_pitch_for_width(mode->hdisplay,
-							       fb->bits_per_pixel))
+	if (fb->pitch < intel_framebuffer_pitch_for_width(mode->hdisplay,
+							  fb->bits_per_pixel))
 		return NULL;
 
-	if (obj->base.size < mode->vdisplay * fb->pitches[0])
+	if (obj->base.size < mode->vdisplay * fb->pitch)
 		return NULL;
 
 	return fb;
@@ -7032,7 +6026,9 @@ static void intel_increase_pllclock(struct drm_crtc *crtc)
 	if (!HAS_PIPE_CXSR(dev) && (dpll & DISPLAY_RATE_SELECT_FPA1)) {
 		DRM_DEBUG_DRIVER("upclocking LVDS\n");
 
-		assert_panel_unlocked(dev_priv, pipe);
+		/* Unlock panel regs */
+		I915_WRITE(PP_CONTROL,
+			   I915_READ(PP_CONTROL) | PANEL_UNLOCK_REGS);
 
 		dpll &= ~DISPLAY_RATE_SELECT_FPA1;
 		I915_WRITE(dpll_reg, dpll);
@@ -7041,6 +6037,9 @@ static void intel_increase_pllclock(struct drm_crtc *crtc)
 		dpll = I915_READ(dpll_reg);
 		if (dpll & DISPLAY_RATE_SELECT_FPA1)
 			DRM_DEBUG_DRIVER("failed to upclock LVDS!\n");
+
+		/* ...and lock them again */
+		I915_WRITE(PP_CONTROL, I915_READ(PP_CONTROL) & 0x3);
 	}
 
 	/* Schedule downclock */
@@ -7070,7 +6069,9 @@ static void intel_decrease_pllclock(struct drm_crtc *crtc)
 	if (!HAS_PIPE_CXSR(dev) && intel_crtc->lowfreq_avail) {
 		DRM_DEBUG_DRIVER("downclocking LVDS\n");
 
-		assert_panel_unlocked(dev_priv, pipe);
+		/* Unlock panel regs */
+		I915_WRITE(PP_CONTROL, I915_READ(PP_CONTROL) |
+			   PANEL_UNLOCK_REGS);
 
 		dpll |= DISPLAY_RATE_SELECT_FPA1;
 		I915_WRITE(dpll_reg, dpll);
@@ -7078,6 +6079,9 @@ static void intel_decrease_pllclock(struct drm_crtc *crtc)
 		dpll = I915_READ(dpll_reg);
 		if (!(dpll & DISPLAY_RATE_SELECT_FPA1))
 			DRM_DEBUG_DRIVER("failed to downclock LVDS!\n");
+
+		/* ...and lock them again */
+		I915_WRITE(PP_CONTROL, I915_READ(PP_CONTROL) & 0x3);
 	}
 
 }
@@ -7192,11 +6196,10 @@ static void intel_unpin_work_fn(struct work_struct *__work)
 		container_of(__work, struct intel_unpin_work, work);
 
 	mutex_lock(&work->dev->struct_mutex);
-	intel_unpin_fb_obj(work->old_fb_obj);
+	i915_gem_object_unpin(work->old_fb_obj);
 	drm_gem_object_unreference(&work->pending_flip_obj->base);
 	drm_gem_object_unreference(&work->old_fb_obj->base);
 
-	intel_update_fbc(work->dev);
 	mutex_unlock(&work->dev->struct_mutex);
 	kfree(work);
 }
@@ -7323,7 +6326,7 @@ static int intel_gen2_queue_flip(struct drm_device *dev,
 		goto out;
 
 	/* Offset into the new buffer for cases of shared fbs between CRTCs */
-	offset = crtc->y * fb->pitches[0] + crtc->x * fb->bits_per_pixel/8;
+	offset = crtc->y * fb->pitch + crtc->x * fb->bits_per_pixel/8;
 
 	ret = BEGIN_LP_RING(6);
 	if (ret)
@@ -7340,9 +6343,9 @@ static int intel_gen2_queue_flip(struct drm_device *dev,
 	OUT_RING(MI_NOOP);
 	OUT_RING(MI_DISPLAY_FLIP |
 		 MI_DISPLAY_FLIP_PLANE(intel_crtc->plane));
-	OUT_RING(fb->pitches[0]);
+	OUT_RING(fb->pitch);
 	OUT_RING(obj->gtt_offset + offset);
-	OUT_RING(0); /* aux display base address, unused */
+	OUT_RING(MI_NOOP);
 	ADVANCE_LP_RING();
 out:
 	return ret;
@@ -7364,7 +6367,7 @@ static int intel_gen3_queue_flip(struct drm_device *dev,
 		goto out;
 
 	/* Offset into the new buffer for cases of shared fbs between CRTCs */
-	offset = crtc->y * fb->pitches[0] + crtc->x * fb->bits_per_pixel/8;
+	offset = crtc->y * fb->pitch + crtc->x * fb->bits_per_pixel/8;
 
 	ret = BEGIN_LP_RING(6);
 	if (ret)
@@ -7378,7 +6381,7 @@ static int intel_gen3_queue_flip(struct drm_device *dev,
 	OUT_RING(MI_NOOP);
 	OUT_RING(MI_DISPLAY_FLIP_I915 |
 		 MI_DISPLAY_FLIP_PLANE(intel_crtc->plane));
-	OUT_RING(fb->pitches[0]);
+	OUT_RING(fb->pitch);
 	OUT_RING(obj->gtt_offset + offset);
 	OUT_RING(MI_NOOP);
 
@@ -7411,7 +6414,7 @@ static int intel_gen4_queue_flip(struct drm_device *dev,
 	 */
 	OUT_RING(MI_DISPLAY_FLIP |
 		 MI_DISPLAY_FLIP_PLANE(intel_crtc->plane));
-	OUT_RING(fb->pitches[0]);
+	OUT_RING(fb->pitch);
 	OUT_RING(obj->gtt_offset | obj->tiling_mode);
 
 	/* XXX Enabling the panel-fitter across page-flip is so far
@@ -7446,7 +6449,7 @@ static int intel_gen6_queue_flip(struct drm_device *dev,
 
 	OUT_RING(MI_DISPLAY_FLIP |
 		 MI_DISPLAY_FLIP_PLANE(intel_crtc->plane));
-	OUT_RING(fb->pitches[0] | obj->tiling_mode);
+	OUT_RING(fb->pitch | obj->tiling_mode);
 	OUT_RING(obj->gtt_offset);
 
 	pf = I915_READ(PF_CTL(intel_crtc->pipe)) & PF_ENABLE;
@@ -7482,7 +6485,7 @@ static int intel_gen7_queue_flip(struct drm_device *dev,
 		goto out;
 
 	intel_ring_emit(ring, MI_DISPLAY_FLIP_I915 | (intel_crtc->plane << 19));
-	intel_ring_emit(ring, (fb->pitches[0] | obj->tiling_mode));
+	intel_ring_emit(ring, (fb->pitch | obj->tiling_mode));
 	intel_ring_emit(ring, (obj->gtt_offset));
 	intel_ring_emit(ring, (MI_NOOP));
 	intel_ring_advance(ring);
@@ -7558,7 +6561,6 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc,
 	if (ret)
 		goto cleanup_pending;
 
-	intel_disable_fbc(dev);
 	mutex_unlock(&dev->struct_mutex);
 
 	trace_i915_flip_request(intel_crtc->plane, obj);
@@ -7696,11 +6698,8 @@ static void intel_crtc_init(struct drm_device *dev, int pipe)
 
 	intel_crtc_reset(&intel_crtc->base);
 	intel_crtc->active = true; /* force the pipe off on setup_init_config */
-	intel_crtc->bpp = 24; /* default for pre-Ironlake */
 
 	if (HAS_PCH_SPLIT(dev)) {
-		if (pipe == 2 && IS_IVYBRIDGE(dev))
-			intel_crtc->no_pll = true;
 		intel_helper_funcs.prepare = ironlake_crtc_prepare;
 		intel_helper_funcs.commit = ironlake_crtc_commit;
 	} else {
@@ -7780,9 +6779,10 @@ static void intel_setup_outputs(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_encoder *encoder;
 	bool dpd_is_edp = false;
-	bool has_lvds;
+	bool has_lvds = false;
 
-	has_lvds = intel_lvds_init(dev);
+	if (IS_MOBILE(dev) && !IS_I830(dev))
+		has_lvds = intel_lvds_init(dev);
 	if (!has_lvds && !HAS_PCH_SPLIT(dev)) {
 		/* disable the panel fitter on everything but LVDS */
 		I915_WRITE(PFIT_CONTROL, 0);
@@ -7877,11 +6877,10 @@ static void intel_setup_outputs(struct drm_device *dev)
 			intel_encoder_clones(dev, encoder->clone_mask);
 	}
 
+	intel_panel_setup_backlight(dev);
+
 	/* disable all the possible outputs/crtcs before entering KMS mode */
 	drm_helper_disable_unused_functions(dev);
-
-	if (HAS_PCH_SPLIT(dev))
-		ironlake_init_pch_refclk(dev);
 }
 
 static void intel_user_framebuffer_destroy(struct drm_framebuffer *fb)
@@ -7911,7 +6910,7 @@ static const struct drm_framebuffer_funcs intel_fb_funcs = {
 
 int intel_framebuffer_init(struct drm_device *dev,
 			   struct intel_framebuffer *intel_fb,
-			   struct drm_mode_fb_cmd2 *mode_cmd,
+			   struct drm_mode_fb_cmd *mode_cmd,
 			   struct drm_i915_gem_object *obj)
 {
 	int ret;
@@ -7919,26 +6918,16 @@ int intel_framebuffer_init(struct drm_device *dev,
 	if (obj->tiling_mode == I915_TILING_Y)
 		return -EINVAL;
 
-	if (mode_cmd->pitches[0] & 63)
+	if (mode_cmd->pitch & 63)
 		return -EINVAL;
 
-	switch (mode_cmd->pixel_format) {
-	case DRM_FORMAT_RGB332:
-	case DRM_FORMAT_RGB565:
-	case DRM_FORMAT_XRGB8888:
-	case DRM_FORMAT_ARGB8888:
-	case DRM_FORMAT_XRGB2101010:
-	case DRM_FORMAT_ARGB2101010:
-		/* RGB formats are common across chipsets */
-		break;
-	case DRM_FORMAT_YUYV:
-	case DRM_FORMAT_UYVY:
-	case DRM_FORMAT_YVYU:
-	case DRM_FORMAT_VYUY:
+	switch (mode_cmd->bpp) {
+	case 8:
+	case 16:
+	case 24:
+	case 32:
 		break;
 	default:
-		DRM_DEBUG_KMS("unsupported pixel format %u\n",
-				mode_cmd->pixel_format);
 		return -EINVAL;
 	}
 
@@ -7956,12 +6945,11 @@ int intel_framebuffer_init(struct drm_device *dev,
 static struct drm_framebuffer *
 intel_user_framebuffer_create(struct drm_device *dev,
 			      struct drm_file *filp,
-			      struct drm_mode_fb_cmd2 *mode_cmd)
+			      struct drm_mode_fb_cmd *mode_cmd)
 {
 	struct drm_i915_gem_object *obj;
 
-	obj = to_intel_bo(drm_gem_object_lookup(dev, filp,
-						mode_cmd->handles[0]));
+	obj = to_intel_bo(drm_gem_object_lookup(dev, filp, mode_cmd->handle));
 	if (&obj->base == NULL)
 		return ERR_PTR(-ENOENT);
 
@@ -8132,10 +7120,6 @@ void gen6_disable_rps(struct drm_device *dev)
 	I915_WRITE(GEN6_RPNSWREQ, 1 << 31);
 	I915_WRITE(GEN6_PMINTRMSK, 0xffffffff);
 	I915_WRITE(GEN6_PMIER, 0);
-	/* Complete PM interrupt masking here doesn't race with the rps work
-	 * item again unmasking PM interrupts because that is using a different
-	 * register (PMIMR) to mask PM interrupts. The only risk is in leaving
-	 * stale bits in PMIIR and PMIMR which gen6_enable_rps will clean up. */
 
 	spin_lock_irq(&dev_priv->rps_lock);
 	dev_priv->pm_iir = 0;
@@ -8230,37 +7214,11 @@ void intel_init_emon(struct drm_device *dev)
 	dev_priv->corr = (lcfuse & LCFUSE_HIV_MASK);
 }
 
-static bool intel_enable_rc6(struct drm_device *dev)
-{
-	/*
-	 * Respect the kernel parameter if it is set
-	 */
-	if (i915_enable_rc6 >= 0)
-		return i915_enable_rc6;
-
-	/*
-	 * Disable RC6 on Ironlake
-	 */
-	if (INTEL_INFO(dev)->gen == 5)
-		return 0;
-
-	/*
-	 * Disable rc6 on Sandybridge
-	 */
-	if (INTEL_INFO(dev)->gen == 6) {
-		DRM_DEBUG_DRIVER("Sandybridge: RC6 disabled\n");
-		return 0;
-	}
-	DRM_DEBUG_DRIVER("RC6 enabled\n");
-	return 1;
-}
-
 void gen6_enable_rps(struct drm_i915_private *dev_priv)
 {
 	u32 rp_state_cap = I915_READ(GEN6_RP_STATE_CAP);
 	u32 gt_perf_status = I915_READ(GEN6_GT_PERF_STATUS);
 	u32 pcu_mbox, rc6_mask = 0;
-	u32 gtfifodbg;
 	int cur_freq, min_freq, max_freq;
 	int i;
 
@@ -8272,13 +7230,6 @@ void gen6_enable_rps(struct drm_i915_private *dev_priv)
 	 */
 	I915_WRITE(GEN6_RC_STATE, 0);
 	mutex_lock(&dev_priv->dev->struct_mutex);
-
-	/* Clear the DBG now so we don't confuse earlier errors */
-	if ((gtfifodbg = I915_READ(GTFIFODBG))) {
-		DRM_ERROR("GT fifo had a previous error %x\n", gtfifodbg);
-		I915_WRITE(GTFIFODBG, gtfifodbg);
-	}
-
 	gen6_gt_force_wake_get(dev_priv);
 
 	/* disable the counters and set deterministic thresholds */
@@ -8299,9 +7250,9 @@ void gen6_enable_rps(struct drm_i915_private *dev_priv)
 	I915_WRITE(GEN6_RC6p_THRESHOLD, 100000);
 	I915_WRITE(GEN6_RC6pp_THRESHOLD, 64000); /* unused */
 
-	if (intel_enable_rc6(dev_priv->dev))
-		rc6_mask = GEN6_RC_CTL_RC6_ENABLE |
-			((IS_GEN7(dev_priv->dev)) ? GEN6_RC_CTL_RC6p_ENABLE : 0);
+	if (i915_enable_rc6)
+		rc6_mask = GEN6_RC_CTL_RC6p_ENABLE |
+			GEN6_RC_CTL_RC6_ENABLE;
 
 	I915_WRITE(GEN6_RC_CONTROL,
 		   rc6_mask |
@@ -8326,7 +7277,7 @@ void gen6_enable_rps(struct drm_i915_private *dev_priv)
 	I915_WRITE(GEN6_RP_IDLE_HYSTERSIS, 10);
 	I915_WRITE(GEN6_RP_CONTROL,
 		   GEN6_RP_MEDIA_TURBO |
-		   GEN6_RP_MEDIA_HW_MODE |
+		   GEN6_RP_USE_NORMAL_FREQ |
 		   GEN6_RP_MEDIA_IS_GFX |
 		   GEN6_RP_ENABLE |
 		   GEN6_RP_UP_BUSY_AVG |
@@ -8387,59 +7338,6 @@ void gen6_enable_rps(struct drm_i915_private *dev_priv)
 	mutex_unlock(&dev_priv->dev->struct_mutex);
 }
 
-void gen6_update_ring_freq(struct drm_i915_private *dev_priv)
-{
-	int min_freq = 15;
-	int gpu_freq, ia_freq, max_ia_freq;
-	int scaling_factor = 180;
-
-	max_ia_freq = cpufreq_quick_get_max(0);
-	/*
-	 * Default to measured freq if none found, PCU will ensure we don't go
-	 * over
-	 */
-	if (!max_ia_freq)
-		max_ia_freq = tsc_khz;
-
-	/* Convert from kHz to MHz */
-	max_ia_freq /= 1000;
-
-	mutex_lock(&dev_priv->dev->struct_mutex);
-
-	/*
-	 * For each potential GPU frequency, load a ring frequency we'd like
-	 * to use for memory access.  We do this by specifying the IA frequency
-	 * the PCU should use as a reference to determine the ring frequency.
-	 */
-	for (gpu_freq = dev_priv->max_delay; gpu_freq >= dev_priv->min_delay;
-	     gpu_freq--) {
-		int diff = dev_priv->max_delay - gpu_freq;
-
-		/*
-		 * For GPU frequencies less than 750MHz, just use the lowest
-		 * ring freq.
-		 */
-		if (gpu_freq < min_freq)
-			ia_freq = 800;
-		else
-			ia_freq = max_ia_freq - ((diff * scaling_factor) / 2);
-		ia_freq = DIV_ROUND_CLOSEST(ia_freq, 100);
-
-		I915_WRITE(GEN6_PCODE_DATA,
-			   (ia_freq << GEN6_PCODE_FREQ_IA_RATIO_SHIFT) |
-			   gpu_freq);
-		I915_WRITE(GEN6_PCODE_MAILBOX, GEN6_PCODE_READY |
-			   GEN6_PCODE_WRITE_MIN_FREQ_TABLE);
-		if (wait_for((I915_READ(GEN6_PCODE_MAILBOX) &
-			      GEN6_PCODE_READY) == 0, 10)) {
-			DRM_ERROR("pcode write of freq table timed out\n");
-			continue;
-		}
-	}
-
-	mutex_unlock(&dev_priv->dev->struct_mutex);
-}
-
 static void ironlake_init_clock_gating(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -8563,12 +7461,10 @@ static void gen6_init_clock_gating(struct drm_device *dev)
 		   ILK_DPARB_CLK_GATE  |
 		   ILK_DPFD_CLK_GATE);
 
-	for_each_pipe(pipe) {
+	for_each_pipe(pipe)
 		I915_WRITE(DSPCNTR(pipe),
 			   I915_READ(DSPCNTR(pipe)) |
 			   DISPPLANE_TRICKLE_FEED_DISABLE);
-		intel_flush_display_plane(dev_priv, pipe);
-	}
 }
 
 static void gen7_setup_fixed_func_scheduler(struct drm_i915_private *dev_priv)
@@ -8621,8 +7517,6 @@ static void ivybridge_init_clock_gating(struct drm_device *dev)
 		I915_WRITE(DSPCNTR(pipe),
 			   I915_READ(DSPCNTR(pipe)) |
 			   DISPPLANE_TRICKLE_FEED_DISABLE);
-		intel_flush_display_plane(dev_priv, pipe);
-	}
 }
 
 static void g4x_init_clock_gating(struct drm_device *dev)
@@ -8786,7 +7680,7 @@ void ironlake_enable_rc6(struct drm_device *dev)
 	/* rc6 disabled by default due to repeated reports of hanging during
 	 * boot and resume.
 	 */
-	if (!intel_enable_rc6(dev))
+	if (!i915_enable_rc6)
 		return;
 
 	mutex_lock(&dev->struct_mutex);
@@ -8856,11 +7750,9 @@ static void intel_init_display(struct drm_device *dev)
 	if (HAS_PCH_SPLIT(dev)) {
 		dev_priv->display.dpms = ironlake_crtc_dpms;
 		dev_priv->display.crtc_mode_set = ironlake_crtc_mode_set;
-		dev_priv->display.update_plane = ironlake_update_plane;
 	} else {
 		dev_priv->display.dpms = i9xx_crtc_dpms;
 		dev_priv->display.crtc_mode_set = i9xx_crtc_mode_set;
-		dev_priv->display.update_plane = i9xx_update_plane;
 	}
 
 	if (I915_HAS_FBC(dev)) {
@@ -8881,7 +7773,7 @@ static void intel_init_display(struct drm_device *dev)
 	}
 
 	/* Returns the core display clock speed */
-	if (IS_I945G(dev) || (IS_G33(dev) && !IS_PINEVIEW_M(dev)))
+	if (IS_I945G(dev) || (IS_G33(dev) && ! IS_PINEVIEW_M(dev)))
 		dev_priv->display.get_display_clock_speed =
 			i945_get_display_clock_speed;
 	else if (IS_I915G(dev))
@@ -8905,34 +7797,6 @@ static void intel_init_display(struct drm_device *dev)
 
 	/* For FIFO watermark updates */
 	if (HAS_PCH_SPLIT(dev)) {
-		dev_priv->display.force_wake_get = __gen6_gt_force_wake_get;
-		dev_priv->display.force_wake_put = __gen6_gt_force_wake_put;
-
-		/* IVB configs may use multi-threaded forcewake */
-		if (IS_IVYBRIDGE(dev)) {
-			u32	ecobus;
-
-			/* A small trick here - if the bios hasn't configured MT forcewake,
-			 * and if the device is in RC6, then force_wake_mt_get will not wake
-			 * the device and the ECOBUS read will return zero. Which will be
-			 * (correctly) interpreted by the test below as MT forcewake being
-			 * disabled.
-			 */
-			mutex_lock(&dev->struct_mutex);
-			__gen6_gt_force_wake_mt_get(dev_priv);
-			ecobus = I915_READ_NOTRACE(ECOBUS);
-			__gen6_gt_force_wake_mt_put(dev_priv);
-			mutex_unlock(&dev->struct_mutex);
-
-			if (ecobus & FORCEWAKE_MT_ENABLE) {
-				DRM_DEBUG_KMS("Using MT version of forcewake\n");
-				dev_priv->display.force_wake_get =
-					__gen6_gt_force_wake_mt_get;
-				dev_priv->display.force_wake_put =
-					__gen6_gt_force_wake_mt_put;
-			}
-		}
-
 		if (HAS_PCH_IBX(dev))
 			dev_priv->display.init_pch_clock_gating = ibx_init_clock_gating;
 		else if (HAS_PCH_CPT(dev))
@@ -8948,11 +7812,9 @@ static void intel_init_display(struct drm_device *dev)
 			}
 			dev_priv->display.fdi_link_train = ironlake_fdi_link_train;
 			dev_priv->display.init_clock_gating = ironlake_init_clock_gating;
-			dev_priv->display.write_eld = ironlake_write_eld;
 		} else if (IS_GEN6(dev)) {
 			if (SNB_READ_WM0_LATENCY()) {
 				dev_priv->display.update_wm = sandybridge_update_wm;
-				dev_priv->display.update_sprite_wm = sandybridge_update_sprite_wm;
 			} else {
 				DRM_DEBUG_KMS("Failed to read display plane latency. "
 					      "Disable CxSR\n");
@@ -8960,20 +7822,18 @@ static void intel_init_display(struct drm_device *dev)
 			}
 			dev_priv->display.fdi_link_train = gen6_fdi_link_train;
 			dev_priv->display.init_clock_gating = gen6_init_clock_gating;
-			dev_priv->display.write_eld = ironlake_write_eld;
 		} else if (IS_IVYBRIDGE(dev)) {
 			/* FIXME: detect B0+ stepping and use auto training */
 			dev_priv->display.fdi_link_train = ivb_manual_fdi_link_train;
 			if (SNB_READ_WM0_LATENCY()) {
 				dev_priv->display.update_wm = sandybridge_update_wm;
-				dev_priv->display.update_sprite_wm = sandybridge_update_sprite_wm;
 			} else {
 				DRM_DEBUG_KMS("Failed to read display plane latency. "
 					      "Disable CxSR\n");
 				dev_priv->display.update_wm = NULL;
 			}
 			dev_priv->display.init_clock_gating = ivybridge_init_clock_gating;
-			dev_priv->display.write_eld = ironlake_write_eld;
+
 		} else
 			dev_priv->display.update_wm = NULL;
 	} else if (IS_PINEVIEW(dev)) {
@@ -8984,7 +7844,7 @@ static void intel_init_display(struct drm_device *dev)
 			DRM_INFO("failed to find known CxSR latency "
 				 "(found ddr%s fsb freq %d, mem freq %d), "
 				 "disabling CxSR\n",
-				 (dev_priv->is_ddr3 == 1) ? "3" : "2",
+				 (dev_priv->is_ddr3 == 1) ? "3": "2",
 				 dev_priv->fsb_freq, dev_priv->mem_freq);
 			/* Disable CxSR and never update its watermark again */
 			pineview_disable_cxsr(dev);
@@ -8993,7 +7853,6 @@ static void intel_init_display(struct drm_device *dev)
 			dev_priv->display.update_wm = pineview_update_wm;
 		dev_priv->display.init_clock_gating = gen3_init_clock_gating;
 	} else if (IS_G4X(dev)) {
-		dev_priv->display.write_eld = g4x_write_eld;
 		dev_priv->display.update_wm = g4x_update_wm;
 		dev_priv->display.init_clock_gating = g4x_init_clock_gating;
 	} else if (IS_GEN4(dev)) {
@@ -9054,7 +7913,7 @@ static void intel_init_display(struct drm_device *dev)
  * resume, or other times.  This quirk makes sure that's the case for
  * affected systems.
  */
-static void quirk_pipea_force(struct drm_device *dev)
+static void quirk_pipea_force (struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
@@ -9079,8 +7938,10 @@ struct intel_quirk {
 };
 
 struct intel_quirk intel_quirks[] = {
+	/* HP Compaq 2730p needs pipe A force quirk (LP: #291555) */
+	{ 0x2a42, 0x103c, 0x30eb, quirk_pipea_force },
 	/* HP Mini needs pipe A force quirk (LP: #322104) */
-	{ 0x27ae, 0x103c, 0x361a, quirk_pipea_force },
+	{ 0x27ae,0x103c, 0x361a, quirk_pipea_force },
 
 	/* Thinkpad R31 needs pipe A force quirk */
 	{ 0x3577, 0x1014, 0x0505, quirk_pipea_force },
@@ -9100,9 +7961,6 @@ struct intel_quirk intel_quirks[] = {
 
 	/* Lenovo U160 cannot use SSC on LVDS */
 	{ 0x0046, 0x17aa, 0x3920, quirk_ssc_force_disable },
-
-	/* Sony Vaio Y cannot use SSC on LVDS */
-	{ 0x0046, 0x104d, 0x9076, quirk_ssc_force_disable },
 };
 
 static void intel_init_quirks(struct drm_device *dev)
@@ -9148,16 +8006,13 @@ static void i915_disable_vga(struct drm_device *dev)
 void intel_modeset_init(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	int i, ret;
+	int i;
 
 	drm_mode_config_init(dev);
 
 	dev->mode_config.min_width = 0;
 	dev->mode_config.min_height = 0;
 
-	dev->mode_config.preferred_depth = 24;
-	dev->mode_config.prefer_shadow = 1;
-
 	dev->mode_config.funcs = (void *)&intel_mode_funcs;
 
 	intel_init_quirks(dev);
@@ -9181,9 +8036,6 @@ void intel_modeset_init(struct drm_device *dev)
 
 	for (i = 0; i < dev_priv->num_pipe; i++) {
 		intel_crtc_init(dev, i);
-		ret = intel_plane_init(dev, i);
-		if (ret)
-			DRM_DEBUG_KMS("plane %d init failed: %d\n", i, ret);
 	}
 
 	/* Just disable it once at startup */
@@ -9197,10 +8049,8 @@ void intel_modeset_init(struct drm_device *dev)
 		intel_init_emon(dev);
 	}
 
-	if (IS_GEN6(dev) || IS_GEN7(dev)) {
+	if (IS_GEN6(dev) || IS_GEN7(dev))
 		gen6_enable_rps(dev_priv);
-		gen6_update_ring_freq(dev_priv);
-	}
 
 	INIT_WORK(&dev_priv->idle_work, intel_idle_update);
 	setup_timer(&dev_priv->idle_timer, intel_gpu_idle_timer,
@@ -9236,7 +8086,8 @@ void intel_modeset_cleanup(struct drm_device *dev)
 		intel_increase_pllclock(crtc);
 	}
 
-	intel_disable_fbc(dev);
+	if (dev_priv->display.disable_fbc)
+		dev_priv->display.disable_fbc(dev);
 
 	if (IS_IRONLAKE_M(dev))
 		ironlake_disable_drps(dev);
@@ -9252,10 +8103,6 @@ void intel_modeset_cleanup(struct drm_device *dev)
 	 * enqueue unpin/hotplug work. */
 	drm_irq_uninstall(dev);
 	cancel_work_sync(&dev_priv->hotplug_work);
-	cancel_work_sync(&dev_priv->rps_work);
-
-	/* flush any delayed tasks or pending work */
-	flush_scheduled_work();
 
 	/* Shut off idle work before the crtcs get freed. */
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
@@ -9338,7 +8185,7 @@ struct intel_display_error_state {
 struct intel_display_error_state *
 intel_display_capture_error_state(struct drm_device *dev)
 {
-	drm_i915_private_t *dev_priv = dev->dev_private;
+        drm_i915_private_t *dev_priv = dev->dev_private;
 	struct intel_display_error_state *error;
 	int i;
 
@@ -9354,7 +8201,7 @@ intel_display_capture_error_state(struct drm_device *dev)
 		error->plane[i].control = I915_READ(DSPCNTR(i));
 		error->plane[i].stride = I915_READ(DSPSTRIDE(i));
 		error->plane[i].size = I915_READ(DSPSIZE(i));
-		error->plane[i].pos = I915_READ(DSPPOS(i));
+		error->plane[i].pos= I915_READ(DSPPOS(i));
 		error->plane[i].addr = I915_READ(DSPADDR(i));
 		if (INTEL_INFO(dev)->gen >= 4) {
 			error->plane[i].surface = I915_READ(DSPSURF(i));
diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c
index 993ec13..57b3728 100644
--- a/drivers/gpu/drm/i915/intel_dp.c
+++ b/drivers/gpu/drm/i915/intel_dp.c
@@ -36,7 +36,7 @@
 #include "i915_drv.h"
 #include "drm_dp_helper.h"
 
-#define DP_RECEIVER_CAP_SIZE	0xf
+
 #define DP_LINK_STATUS_SIZE	6
 #define DP_LINK_CHECK_TIMEOUT	(10 * 1000)
 
@@ -48,24 +48,16 @@ struct intel_dp {
 	uint32_t DP;
 	uint8_t  link_configuration[DP_LINK_CONFIGURATION_SIZE];
 	bool has_audio;
-	enum hdmi_force_audio force_audio;
+	int force_audio;
 	uint32_t color_range;
-	int dpms_mode;
 	uint8_t link_bw;
 	uint8_t lane_count;
-	uint8_t dpcd[DP_RECEIVER_CAP_SIZE];
+	uint8_t dpcd[4];
 	struct i2c_adapter adapter;
 	struct i2c_algo_dp_aux_data algo;
 	bool is_pch_edp;
 	uint8_t	train_set[4];
-	int panel_power_up_delay;
-	int panel_power_down_delay;
-	int panel_power_cycle_delay;
-	int backlight_on_delay;
-	int backlight_off_delay;
-	struct drm_display_mode *panel_fixed_mode;  /* for eDP */
-	struct delayed_work panel_vdd_work;
-	bool want_panel_vdd;
+	uint8_t link_status[DP_LINK_STATUS_SIZE];
 };
 
 /**
@@ -93,17 +85,6 @@ static bool is_pch_edp(struct intel_dp *intel_dp)
 	return intel_dp->is_pch_edp;
 }
 
-/**
- * is_cpu_edp - is the port on the CPU and attached to an eDP panel?
- * @intel_dp: DP struct
- *
- * Returns true if the given DP struct corresponds to a CPU eDP port.
- */
-static bool is_cpu_edp(struct intel_dp *intel_dp)
-{
-	return is_edp(intel_dp) && !is_pch_edp(intel_dp);
-}
-
 static struct intel_dp *enc_to_intel_dp(struct drm_encoder *encoder)
 {
 	return container_of(encoder, struct intel_dp, base.base);
@@ -139,7 +120,7 @@ static void intel_dp_complete_link_train(struct intel_dp *intel_dp);
 static void intel_dp_link_down(struct intel_dp *intel_dp);
 
 void
-intel_edp_link_config(struct intel_encoder *intel_encoder,
+intel_edp_link_config (struct intel_encoder *intel_encoder,
 		       int *lane_num, int *link_bw)
 {
 	struct intel_dp *intel_dp = container_of(intel_encoder, struct intel_dp, base);
@@ -154,12 +135,16 @@ intel_edp_link_config(struct intel_encoder *intel_encoder,
 static int
 intel_dp_max_lane_count(struct intel_dp *intel_dp)
 {
-	int max_lane_count = intel_dp->dpcd[DP_MAX_LANE_COUNT] & 0x1f;
-	switch (max_lane_count) {
-	case 1: case 2: case 4:
-		break;
-	default:
-		max_lane_count = 4;
+	int max_lane_count = 4;
+
+	if (intel_dp->dpcd[DP_DPCD_REV] >= 0x11) {
+		max_lane_count = intel_dp->dpcd[DP_MAX_LANE_COUNT] & 0x1f;
+		switch (max_lane_count) {
+		case 1: case 2: case 4:
+			break;
+		default:
+			max_lane_count = 4;
+		}
 	}
 	return max_lane_count;
 }
@@ -189,27 +174,16 @@ intel_dp_link_clock(uint8_t link_bw)
 		return 162000;
 }
 
-/*
- * The units on the numbers in the next two are... bizarre.  Examples will
- * make it clearer; this one parallels an example in the eDP spec.
- *
- * intel_dp_max_data_rate for one lane of 2.7GHz evaluates as:
- *
- *     270000 * 1 * 8 / 10 == 216000
- *
- * The actual data capacity of that configuration is 2.16Gbit/s, so the
- * units are decakilobits.  ->clock in a drm_display_mode is in kilohertz -
- * or equivalently, kilopixels per second - so for 1680x1050R it'd be
- * 119000.  At 18bpp that's 2142000 kilobits per second.
- *
- * Thus the strange-looking division by 10 in intel_dp_link_required, to
- * get the result in decakilobits instead of kilobits.
- */
-
+/* I think this is a fiction */
 static int
-intel_dp_link_required(int pixel_clock, int bpp)
+intel_dp_link_required(struct drm_device *dev, struct intel_dp *intel_dp, int pixel_clock)
 {
-	return (pixel_clock * bpp + 9) / 10;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+
+	if (is_edp(intel_dp))
+		return (pixel_clock * dev_priv->edp.bpp + 7) / 8;
+	else
+		return pixel_clock * 3;
 }
 
 static int
@@ -223,28 +197,25 @@ intel_dp_mode_valid(struct drm_connector *connector,
 		    struct drm_display_mode *mode)
 {
 	struct intel_dp *intel_dp = intel_attached_dp(connector);
+	struct drm_device *dev = connector->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	int max_link_clock = intel_dp_link_clock(intel_dp_max_link_bw(intel_dp));
 	int max_lanes = intel_dp_max_lane_count(intel_dp);
-	int max_rate, mode_rate;
 
-	if (is_edp(intel_dp) && intel_dp->panel_fixed_mode) {
-		if (mode->hdisplay > intel_dp->panel_fixed_mode->hdisplay)
+	if (is_edp(intel_dp) && dev_priv->panel_fixed_mode) {
+		if (mode->hdisplay > dev_priv->panel_fixed_mode->hdisplay)
 			return MODE_PANEL;
 
-		if (mode->vdisplay > intel_dp->panel_fixed_mode->vdisplay)
+		if (mode->vdisplay > dev_priv->panel_fixed_mode->vdisplay)
 			return MODE_PANEL;
 	}
 
-	mode_rate = intel_dp_link_required(mode->clock, 24);
-	max_rate = intel_dp_max_data_rate(max_link_clock, max_lanes);
-
-	if (mode_rate > max_rate) {
-			mode_rate = intel_dp_link_required(mode->clock, 18);
-			if (mode_rate > max_rate)
-				return MODE_CLOCK_HIGH;
-			else
-				mode->private_flags |= INTEL_MODE_DP_FORCE_6BPC;
-	}
+	/* only refuse the mode on non eDP since we have seen some weird eDP panels
+	   which are outside spec tolerances but somehow work by magic */
+	if (!is_edp(intel_dp) &&
+	    (intel_dp_link_required(connector->dev, intel_dp, mode->clock)
+	     > intel_dp_max_data_rate(max_link_clock, max_lanes)))
+		return MODE_CLOCK_HIGH;
 
 	if (mode->clock < 10000)
 		return MODE_CLOCK_LOW;
@@ -305,38 +276,6 @@ intel_hrawclk(struct drm_device *dev)
 	}
 }
 
-static bool ironlake_edp_have_panel_power(struct intel_dp *intel_dp)
-{
-	struct drm_device *dev = intel_dp->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	return (I915_READ(PCH_PP_STATUS) & PP_ON) != 0;
-}
-
-static bool ironlake_edp_have_panel_vdd(struct intel_dp *intel_dp)
-{
-	struct drm_device *dev = intel_dp->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	return (I915_READ(PCH_PP_CONTROL) & EDP_FORCE_VDD) != 0;
-}
-
-static void
-intel_dp_check_edp(struct intel_dp *intel_dp)
-{
-	struct drm_device *dev = intel_dp->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	if (!is_edp(intel_dp))
-		return;
-	if (!ironlake_edp_have_panel_power(intel_dp) && !ironlake_edp_have_panel_vdd(intel_dp)) {
-		WARN(1, "eDP powered off while attempting aux channel communication.\n");
-		DRM_DEBUG_KMS("Status 0x%08x Control 0x%08x\n",
-			      I915_READ(PCH_PP_STATUS),
-			      I915_READ(PCH_PP_CONTROL));
-	}
-}
-
 static int
 intel_dp_aux_ch(struct intel_dp *intel_dp,
 		uint8_t *send, int send_bytes,
@@ -351,9 +290,8 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 	int recv_bytes;
 	uint32_t status;
 	uint32_t aux_clock_divider;
-	int try, precharge = 5;
+	int try, precharge;
 
-	intel_dp_check_edp(intel_dp);
 	/* The clock divider is based off the hrawclk,
 	 * and would like to run at 2MHz. So, take the
 	 * hrawclk value and divide by 2 and use that
@@ -361,27 +299,24 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 	 * Note that PCH attached eDP panels should use a 125MHz input
 	 * clock divider.
 	 */
-	if (is_cpu_edp(intel_dp)) {
-		if (IS_GEN6(dev) || IS_GEN7(dev))
-			aux_clock_divider = 200; /* SNB & IVB eDP input clock at 400Mhz */
+	if (is_edp(intel_dp) && !is_pch_edp(intel_dp)) {
+		if (IS_GEN6(dev))
+			aux_clock_divider = 200; /* SNB eDP input clock at 400Mhz */
 		else
 			aux_clock_divider = 225; /* eDP input clock at 450Mhz */
 	} else if (HAS_PCH_SPLIT(dev))
-		aux_clock_divider = 63; /* IRL input clock fixed at 125Mhz */
+		aux_clock_divider = 62; /* IRL input clock fixed at 125Mhz */
 	else
 		aux_clock_divider = intel_hrawclk(dev) / 2;
 
-	/* Try to wait for any previous AUX channel activity */
-	for (try = 0; try < 3; try++) {
-		status = I915_READ(ch_ctl);
-		if ((status & DP_AUX_CH_CTL_SEND_BUSY) == 0)
-			break;
-		msleep(1);
-	}
+	if (IS_GEN6(dev))
+		precharge = 3;
+	else
+		precharge = 5;
 
-	if (try == 3) {
-		WARN(1, "dp_aux_ch not started status 0x%08x\n",
-		     I915_READ(ch_ctl));
+	if (I915_READ(ch_ctl) & DP_AUX_CH_CTL_SEND_BUSY) {
+		DRM_ERROR("dp_aux_ch not started status 0x%08x\n",
+			  I915_READ(ch_ctl));
 		return -EBUSY;
 	}
 
@@ -391,7 +326,7 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 		for (i = 0; i < send_bytes; i += 4)
 			I915_WRITE(ch_data + i,
 				   pack_aux(send + i, send_bytes - i));
-
+	
 		/* Send the command and wait for it to complete */
 		I915_WRITE(ch_ctl,
 			   DP_AUX_CH_CTL_SEND_BUSY |
@@ -408,17 +343,13 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 				break;
 			udelay(100);
 		}
-
+	
 		/* Clear done status and any errors */
 		I915_WRITE(ch_ctl,
 			   status |
 			   DP_AUX_CH_CTL_DONE |
 			   DP_AUX_CH_CTL_TIME_OUT_ERROR |
 			   DP_AUX_CH_CTL_RECEIVE_ERROR);
-
-		if (status & (DP_AUX_CH_CTL_TIME_OUT_ERROR |
-			      DP_AUX_CH_CTL_RECEIVE_ERROR))
-			continue;
 		if (status & DP_AUX_CH_CTL_DONE)
 			break;
 	}
@@ -448,7 +379,7 @@ intel_dp_aux_ch(struct intel_dp *intel_dp,
 		      DP_AUX_CH_CTL_MESSAGE_SIZE_SHIFT);
 	if (recv_bytes > recv_size)
 		recv_bytes = recv_size;
-
+	
 	for (i = 0; i < recv_bytes; i += 4)
 		unpack_aux(I915_READ(ch_data + i),
 			   recv + i, recv_bytes - i);
@@ -466,7 +397,6 @@ intel_dp_aux_native_write(struct intel_dp *intel_dp,
 	int msg_bytes;
 	uint8_t	ack;
 
-	intel_dp_check_edp(intel_dp);
 	if (send_bytes > 16)
 		return -1;
 	msg[0] = AUX_NATIVE_WRITE << 4;
@@ -509,7 +439,6 @@ intel_dp_aux_native_read(struct intel_dp *intel_dp,
 	uint8_t ack;
 	int ret;
 
-	intel_dp_check_edp(intel_dp);
 	msg[0] = AUX_NATIVE_READ << 4;
 	msg[1] = address >> 8;
 	msg[2] = address & 0xff;
@@ -553,7 +482,6 @@ intel_dp_i2c_aux_ch(struct i2c_adapter *adapter, int mode,
 	int reply_bytes;
 	int ret;
 
-	intel_dp_check_edp(intel_dp);
 	/* Set up the command byte */
 	if (mode & MODE_I2C_READ)
 		msg[0] = AUX_I2C_READ << 4;
@@ -645,32 +573,24 @@ intel_dp_i2c_aux_ch(struct i2c_adapter *adapter, int mode,
 	return -EREMOTEIO;
 }
 
-static void ironlake_edp_panel_vdd_on(struct intel_dp *intel_dp);
-static void ironlake_edp_panel_vdd_off(struct intel_dp *intel_dp, bool sync);
-
 static int
 intel_dp_i2c_init(struct intel_dp *intel_dp,
 		  struct intel_connector *intel_connector, const char *name)
 {
-	int	ret;
-
 	DRM_DEBUG_KMS("i2c_init %s\n", name);
 	intel_dp->algo.running = false;
 	intel_dp->algo.address = 0;
 	intel_dp->algo.aux_ch = intel_dp_i2c_aux_ch;
 
-	memset(&intel_dp->adapter, '\0', sizeof(intel_dp->adapter));
+	memset(&intel_dp->adapter, '\0', sizeof (intel_dp->adapter));
 	intel_dp->adapter.owner = THIS_MODULE;
 	intel_dp->adapter.class = I2C_CLASS_DDC;
-	strncpy(intel_dp->adapter.name, name, sizeof(intel_dp->adapter.name) - 1);
+	strncpy (intel_dp->adapter.name, name, sizeof(intel_dp->adapter.name) - 1);
 	intel_dp->adapter.name[sizeof(intel_dp->adapter.name) - 1] = '\0';
 	intel_dp->adapter.algo_data = &intel_dp->algo;
 	intel_dp->adapter.dev.parent = &intel_connector->base.kdev;
 
-	ironlake_edp_panel_vdd_on(intel_dp);
-	ret = i2c_dp_aux_add_bus(&intel_dp->adapter);
-	ironlake_edp_panel_vdd_off(intel_dp, false);
-	return ret;
+	return i2c_dp_aux_add_bus(&intel_dp->adapter);
 }
 
 static bool
@@ -678,29 +598,29 @@ intel_dp_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *mode,
 		    struct drm_display_mode *adjusted_mode)
 {
 	struct drm_device *dev = encoder->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
 	int lane_count, clock;
 	int max_lane_count = intel_dp_max_lane_count(intel_dp);
 	int max_clock = intel_dp_max_link_bw(intel_dp) == DP_LINK_BW_2_7 ? 1 : 0;
-	int bpp = mode->private_flags & INTEL_MODE_DP_FORCE_6BPC ? 18 : 24;
 	static int bws[2] = { DP_LINK_BW_1_62, DP_LINK_BW_2_7 };
 
-	if (is_edp(intel_dp) && intel_dp->panel_fixed_mode) {
-		intel_fixed_panel_mode(intel_dp->panel_fixed_mode, adjusted_mode);
+	if (is_edp(intel_dp) && dev_priv->panel_fixed_mode) {
+		intel_fixed_panel_mode(dev_priv->panel_fixed_mode, adjusted_mode);
 		intel_pch_panel_fitting(dev, DRM_MODE_SCALE_FULLSCREEN,
 					mode, adjusted_mode);
 		/*
 		 * the mode->clock is used to calculate the Data&Link M/N
 		 * of the pipe. For the eDP the fixed clock should be used.
 		 */
-		mode->clock = intel_dp->panel_fixed_mode->clock;
+		mode->clock = dev_priv->panel_fixed_mode->clock;
 	}
 
 	for (lane_count = 1; lane_count <= max_lane_count; lane_count <<= 1) {
 		for (clock = 0; clock <= max_clock; clock++) {
 			int link_avail = intel_dp_max_data_rate(intel_dp_link_clock(bws[clock]), lane_count);
 
-			if (intel_dp_link_required(mode->clock, bpp)
+			if (intel_dp_link_required(encoder->dev, intel_dp, mode->clock)
 					<= link_avail) {
 				intel_dp->link_bw = bws[clock];
 				intel_dp->lane_count = lane_count;
@@ -714,6 +634,19 @@ intel_dp_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *mode,
 		}
 	}
 
+	if (is_edp(intel_dp)) {
+		/* okay we failed just pick the highest */
+		intel_dp->lane_count = max_lane_count;
+		intel_dp->link_bw = bws[max_clock];
+		adjusted_mode->clock = intel_dp_link_clock(intel_dp->link_bw);
+		DRM_DEBUG_KMS("Force picking display port link bw %02x lane "
+			      "count %d clock %d\n",
+			      intel_dp->link_bw, intel_dp->lane_count,
+			      adjusted_mode->clock);
+
+		return true;
+	}
+
 	return false;
 }
 
@@ -759,7 +692,7 @@ intel_dp_set_m_n(struct drm_crtc *crtc, struct drm_display_mode *mode,
 	struct drm_encoder *encoder;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int lane_count = 4;
+	int lane_count = 4, bpp = 24;
 	struct intel_dp_m_n m_n;
 	int pipe = intel_crtc->pipe;
 
@@ -773,11 +706,13 @@ intel_dp_set_m_n(struct drm_crtc *crtc, struct drm_display_mode *mode,
 			continue;
 
 		intel_dp = enc_to_intel_dp(encoder);
-		if (intel_dp->base.type == INTEL_OUTPUT_DISPLAYPORT ||
-		    intel_dp->base.type == INTEL_OUTPUT_EDP)
-		{
+		if (intel_dp->base.type == INTEL_OUTPUT_DISPLAYPORT) {
 			lane_count = intel_dp->lane_count;
 			break;
+		} else if (is_edp(intel_dp)) {
+			lane_count = dev_priv->edp.lanes;
+			bpp = dev_priv->edp.bpp;
+			break;
 		}
 	}
 
@@ -786,7 +721,7 @@ intel_dp_set_m_n(struct drm_crtc *crtc, struct drm_display_mode *mode,
 	 * the number of bytes_per_pixel post-LUT, which we always
 	 * set up for 8-bits of R/G/B, or 3 bytes total.
 	 */
-	intel_dp_compute_m_n(intel_crtc->bpp, lane_count,
+	intel_dp_compute_m_n(bpp, lane_count,
 			     mode->clock, adjusted_mode->clock, &m_n);
 
 	if (HAS_PCH_SPLIT(dev)) {
@@ -806,53 +741,27 @@ intel_dp_set_m_n(struct drm_crtc *crtc, struct drm_display_mode *mode,
 	}
 }
 
-static void ironlake_edp_pll_on(struct drm_encoder *encoder);
-static void ironlake_edp_pll_off(struct drm_encoder *encoder);
-
 static void
 intel_dp_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
 		  struct drm_display_mode *adjusted_mode)
 {
 	struct drm_device *dev = encoder->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
 	struct drm_crtc *crtc = intel_dp->base.base.crtc;
 	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 
-	/* Turn on the eDP PLL if needed */
-	if (is_edp(intel_dp)) {
-		if (!is_pch_edp(intel_dp))
-			ironlake_edp_pll_on(encoder);
-		else
-			ironlake_edp_pll_off(encoder);
-	}
-
-	/*
-	 * There are four kinds of DP registers:
-	 *
-	 * 	IBX PCH
-	 * 	SNB CPU
-	 *	IVB CPU
-	 * 	CPT PCH
-	 *
-	 * IBX PCH and CPU are the same for almost everything,
-	 * except that the CPU DP PLL is configured in this
-	 * register
-	 *
-	 * CPT PCH is quite different, having many bits moved
-	 * to the TRANS_DP_CTL register instead. That
-	 * configuration happens (oddly) in ironlake_pch_enable
-	 */
+	intel_dp->DP = DP_VOLTAGE_0_4 | DP_PRE_EMPHASIS_0;
+	intel_dp->DP |= intel_dp->color_range;
 
-	/* Preserve the BIOS-computed detected bit. This is
-	 * supposed to be read-only.
-	 */
-	intel_dp->DP = I915_READ(intel_dp->output_reg) & DP_DETECTED;
-	intel_dp->DP |=  DP_VOLTAGE_0_4 | DP_PRE_EMPHASIS_0;
+	if (adjusted_mode->flags & DRM_MODE_FLAG_PHSYNC)
+		intel_dp->DP |= DP_SYNC_HS_HIGH;
+	if (adjusted_mode->flags & DRM_MODE_FLAG_PVSYNC)
+		intel_dp->DP |= DP_SYNC_VS_HIGH;
 
-	/* Handle DP bits in common between all three register formats */
-
-	intel_dp->DP |= DP_VOLTAGE_0_4 | DP_PRE_EMPHASIS_0;
+	if (HAS_PCH_CPT(dev) && !is_edp(intel_dp))
+		intel_dp->DP |= DP_LINK_TRAIN_OFF_CPT;
+	else
+		intel_dp->DP |= DP_LINK_TRAIN_OFF;
 
 	switch (intel_dp->lane_count) {
 	case 1:
@@ -865,297 +774,134 @@ intel_dp_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
 		intel_dp->DP |= DP_PORT_WIDTH_4;
 		break;
 	}
-	if (intel_dp->has_audio) {
-		DRM_DEBUG_DRIVER("Enabling DP audio on pipe %c\n",
-				 pipe_name(intel_crtc->pipe));
+	if (intel_dp->has_audio)
 		intel_dp->DP |= DP_AUDIO_OUTPUT_ENABLE;
-		intel_write_eld(encoder, adjusted_mode);
-	}
+
 	memset(intel_dp->link_configuration, 0, DP_LINK_CONFIGURATION_SIZE);
 	intel_dp->link_configuration[0] = intel_dp->link_bw;
 	intel_dp->link_configuration[1] = intel_dp->lane_count;
-	intel_dp->link_configuration[8] = DP_SET_ANSI_8B10B;
+
 	/*
 	 * Check for DPCD version > 1.1 and enhanced framing support
 	 */
 	if (intel_dp->dpcd[DP_DPCD_REV] >= 0x11 &&
 	    (intel_dp->dpcd[DP_MAX_LANE_COUNT] & DP_ENHANCED_FRAME_CAP)) {
 		intel_dp->link_configuration[1] |= DP_LANE_COUNT_ENHANCED_FRAME_EN;
+		intel_dp->DP |= DP_ENHANCED_FRAMING;
 	}
 
-	/* Split out the IBX/CPU vs CPT settings */
-
-	if (is_cpu_edp(intel_dp) && IS_GEN7(dev)) {
-		if (adjusted_mode->flags & DRM_MODE_FLAG_PHSYNC)
-			intel_dp->DP |= DP_SYNC_HS_HIGH;
-		if (adjusted_mode->flags & DRM_MODE_FLAG_PVSYNC)
-			intel_dp->DP |= DP_SYNC_VS_HIGH;
-		intel_dp->DP |= DP_LINK_TRAIN_OFF_CPT;
-
-		if (intel_dp->link_configuration[1] & DP_LANE_COUNT_ENHANCED_FRAME_EN)
-			intel_dp->DP |= DP_ENHANCED_FRAMING;
-
-		intel_dp->DP |= intel_crtc->pipe << 29;
+	/* CPT DP's pipe select is decided in TRANS_DP_CTL */
+	if (intel_crtc->pipe == 1 && !HAS_PCH_CPT(dev))
+		intel_dp->DP |= DP_PIPEB_SELECT;
 
+	if (is_edp(intel_dp) && !is_pch_edp(intel_dp)) {
 		/* don't miss out required setting for eDP */
 		intel_dp->DP |= DP_PLL_ENABLE;
 		if (adjusted_mode->clock < 200000)
 			intel_dp->DP |= DP_PLL_FREQ_160MHZ;
 		else
 			intel_dp->DP |= DP_PLL_FREQ_270MHZ;
-	} else if (!HAS_PCH_CPT(dev) || is_cpu_edp(intel_dp)) {
-		intel_dp->DP |= intel_dp->color_range;
-
-		if (adjusted_mode->flags & DRM_MODE_FLAG_PHSYNC)
-			intel_dp->DP |= DP_SYNC_HS_HIGH;
-		if (adjusted_mode->flags & DRM_MODE_FLAG_PVSYNC)
-			intel_dp->DP |= DP_SYNC_VS_HIGH;
-		intel_dp->DP |= DP_LINK_TRAIN_OFF;
-
-		if (intel_dp->link_configuration[1] & DP_LANE_COUNT_ENHANCED_FRAME_EN)
-			intel_dp->DP |= DP_ENHANCED_FRAMING;
-
-		if (intel_crtc->pipe == 1)
-			intel_dp->DP |= DP_PIPEB_SELECT;
-
-		if (is_cpu_edp(intel_dp)) {
-			/* don't miss out required setting for eDP */
-			intel_dp->DP |= DP_PLL_ENABLE;
-			if (adjusted_mode->clock < 200000)
-				intel_dp->DP |= DP_PLL_FREQ_160MHZ;
-			else
-				intel_dp->DP |= DP_PLL_FREQ_270MHZ;
-		}
-	} else {
-		intel_dp->DP |= DP_LINK_TRAIN_OFF_CPT;
-	}
-}
-
-#define IDLE_ON_MASK		(PP_ON | 0 	  | PP_SEQUENCE_MASK | 0                     | PP_SEQUENCE_STATE_MASK)
-#define IDLE_ON_VALUE   	(PP_ON | 0 	  | PP_SEQUENCE_NONE | 0                     | PP_SEQUENCE_STATE_ON_IDLE)
-
-#define IDLE_OFF_MASK		(PP_ON | 0        | PP_SEQUENCE_MASK | 0                     | PP_SEQUENCE_STATE_MASK)
-#define IDLE_OFF_VALUE		(0     | 0        | PP_SEQUENCE_NONE | 0                     | PP_SEQUENCE_STATE_OFF_IDLE)
-
-#define IDLE_CYCLE_MASK		(PP_ON | 0        | PP_SEQUENCE_MASK | PP_CYCLE_DELAY_ACTIVE | PP_SEQUENCE_STATE_MASK)
-#define IDLE_CYCLE_VALUE	(0     | 0        | PP_SEQUENCE_NONE | 0                     | PP_SEQUENCE_STATE_OFF_IDLE)
-
-static void ironlake_wait_panel_status(struct intel_dp *intel_dp,
-				       u32 mask,
-				       u32 value)
-{
-	struct drm_device *dev = intel_dp->base.base.dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-
-	DRM_DEBUG_KMS("mask %08x value %08x status %08x control %08x\n",
-		      mask, value,
-		      I915_READ(PCH_PP_STATUS),
-		      I915_READ(PCH_PP_CONTROL));
-
-	if (_wait_for((I915_READ(PCH_PP_STATUS) & mask) == value, 5000, 10)) {
-		DRM_ERROR("Panel status timeout: status %08x control %08x\n",
-			  I915_READ(PCH_PP_STATUS),
-			  I915_READ(PCH_PP_CONTROL));
 	}
 }
 
-static void ironlake_wait_panel_on(struct intel_dp *intel_dp)
-{
-	DRM_DEBUG_KMS("Wait for panel power on\n");
-	ironlake_wait_panel_status(intel_dp, IDLE_ON_MASK, IDLE_ON_VALUE);
-}
-
-static void ironlake_wait_panel_off(struct intel_dp *intel_dp)
-{
-	DRM_DEBUG_KMS("Wait for panel power off time\n");
-	ironlake_wait_panel_status(intel_dp, IDLE_OFF_MASK, IDLE_OFF_VALUE);
-}
-
-static void ironlake_wait_panel_power_cycle(struct intel_dp *intel_dp)
-{
-	DRM_DEBUG_KMS("Wait for panel power cycle\n");
-	ironlake_wait_panel_status(intel_dp, IDLE_CYCLE_MASK, IDLE_CYCLE_VALUE);
-}
-
-
-/* Read the current pp_control value, unlocking the register if it
- * is locked
- */
-
-static  u32 ironlake_get_pp_control(struct drm_i915_private *dev_priv)
-{
-	u32	control = I915_READ(PCH_PP_CONTROL);
-
-	control &= ~PANEL_UNLOCK_MASK;
-	control |= PANEL_UNLOCK_REGS;
-	return control;
-}
-
 static void ironlake_edp_panel_vdd_on(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 pp;
 
-	if (!is_edp(intel_dp))
-		return;
-	DRM_DEBUG_KMS("Turn eDP VDD on\n");
-
-	WARN(intel_dp->want_panel_vdd,
-	     "eDP VDD already requested on\n");
-
-	intel_dp->want_panel_vdd = true;
-
-	if (ironlake_edp_have_panel_vdd(intel_dp)) {
-		DRM_DEBUG_KMS("eDP VDD already on\n");
-		return;
-	}
-
-	if (!ironlake_edp_have_panel_power(intel_dp))
-		ironlake_wait_panel_power_cycle(intel_dp);
+	/*
+	 * If the panel wasn't on, make sure there's not a currently
+	 * active PP sequence before enabling AUX VDD.
+	 */
+	if (!(I915_READ(PCH_PP_STATUS) & PP_ON))
+		msleep(dev_priv->panel_t3);
 
-	pp = ironlake_get_pp_control(dev_priv);
+	pp = I915_READ(PCH_PP_CONTROL);
 	pp |= EDP_FORCE_VDD;
 	I915_WRITE(PCH_PP_CONTROL, pp);
 	POSTING_READ(PCH_PP_CONTROL);
-	DRM_DEBUG_KMS("PCH_PP_STATUS: 0x%08x PCH_PP_CONTROL: 0x%08x\n",
-		      I915_READ(PCH_PP_STATUS), I915_READ(PCH_PP_CONTROL));
-
-	/*
-	 * If the panel wasn't on, delay before accessing aux channel
-	 */
-	if (!ironlake_edp_have_panel_power(intel_dp)) {
-		DRM_DEBUG_KMS("eDP was not running\n");
-		msleep(intel_dp->panel_power_up_delay);
-	}
 }
 
-static void ironlake_panel_vdd_off_sync(struct intel_dp *intel_dp)
+static void ironlake_edp_panel_vdd_off(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 pp;
 
-	if (!intel_dp->want_panel_vdd && ironlake_edp_have_panel_vdd(intel_dp)) {
-		pp = ironlake_get_pp_control(dev_priv);
-		pp &= ~EDP_FORCE_VDD;
-		I915_WRITE(PCH_PP_CONTROL, pp);
-		POSTING_READ(PCH_PP_CONTROL);
-
-		/* Make sure sequencer is idle before allowing subsequent activity */
-		DRM_DEBUG_KMS("PCH_PP_STATUS: 0x%08x PCH_PP_CONTROL: 0x%08x\n",
-			      I915_READ(PCH_PP_STATUS), I915_READ(PCH_PP_CONTROL));
-
-		msleep(intel_dp->panel_power_down_delay);
-	}
-}
-
-static void ironlake_panel_vdd_work(struct work_struct *__work)
-{
-	struct intel_dp *intel_dp = container_of(to_delayed_work(__work),
-						 struct intel_dp, panel_vdd_work);
-	struct drm_device *dev = intel_dp->base.base.dev;
-
-	mutex_lock(&dev->mode_config.mutex);
-	ironlake_panel_vdd_off_sync(intel_dp);
-	mutex_unlock(&dev->mode_config.mutex);
-}
-
-static void ironlake_edp_panel_vdd_off(struct intel_dp *intel_dp, bool sync)
-{
-	if (!is_edp(intel_dp))
-		return;
-
-	DRM_DEBUG_KMS("Turn eDP VDD off %d\n", intel_dp->want_panel_vdd);
-	WARN(!intel_dp->want_panel_vdd, "eDP VDD not forced on");
-
-	intel_dp->want_panel_vdd = false;
+	pp = I915_READ(PCH_PP_CONTROL);
+	pp &= ~EDP_FORCE_VDD;
+	I915_WRITE(PCH_PP_CONTROL, pp);
+	POSTING_READ(PCH_PP_CONTROL);
 
-	if (sync) {
-		ironlake_panel_vdd_off_sync(intel_dp);
-	} else {
-		/*
-		 * Queue the timer to fire a long
-		 * time from now (relative to the power down delay)
-		 * to keep the panel power up across a sequence of operations
-		 */
-		schedule_delayed_work(&intel_dp->panel_vdd_work,
-				      msecs_to_jiffies(intel_dp->panel_power_cycle_delay * 5));
-	}
+	/* Make sure sequencer is idle before allowing subsequent activity */
+	msleep(dev_priv->panel_t12);
 }
 
-static void ironlake_edp_panel_on(struct intel_dp *intel_dp)
+/* Returns true if the panel was already on when called */
+static bool ironlake_edp_panel_on (struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 pp;
-
-	if (!is_edp(intel_dp))
-		return;
+	u32 pp, idle_on_mask = PP_ON | PP_SEQUENCE_STATE_ON_IDLE;
 
-	DRM_DEBUG_KMS("Turn eDP power on\n");
+	if (I915_READ(PCH_PP_STATUS) & PP_ON)
+		return true;
 
-	if (ironlake_edp_have_panel_power(intel_dp)) {
-		DRM_DEBUG_KMS("eDP power already on\n");
-		return;
-	}
+	pp = I915_READ(PCH_PP_CONTROL);
 
-	ironlake_wait_panel_power_cycle(intel_dp);
+	/* ILK workaround: disable reset around power sequence */
+	pp &= ~PANEL_POWER_RESET;
+	I915_WRITE(PCH_PP_CONTROL, pp);
+	POSTING_READ(PCH_PP_CONTROL);
 
-	pp = ironlake_get_pp_control(dev_priv);
-	if (IS_GEN5(dev)) {
-		/* ILK workaround: disable reset around power sequence */
-		pp &= ~PANEL_POWER_RESET;
-		I915_WRITE(PCH_PP_CONTROL, pp);
-		POSTING_READ(PCH_PP_CONTROL);
-	}
+	pp |= PANEL_UNLOCK_REGS | POWER_TARGET_ON;
+	I915_WRITE(PCH_PP_CONTROL, pp);
+	POSTING_READ(PCH_PP_CONTROL);
 
-	pp |= POWER_TARGET_ON;
-	if (!IS_GEN5(dev))
-		pp |= PANEL_POWER_RESET;
+	if (wait_for((I915_READ(PCH_PP_STATUS) & idle_on_mask) == idle_on_mask,
+		     5000))
+		DRM_ERROR("panel on wait timed out: 0x%08x\n",
+			  I915_READ(PCH_PP_STATUS));
 
+	pp |= PANEL_POWER_RESET; /* restore panel reset bit */
 	I915_WRITE(PCH_PP_CONTROL, pp);
 	POSTING_READ(PCH_PP_CONTROL);
 
-	ironlake_wait_panel_on(intel_dp);
-
-	if (IS_GEN5(dev)) {
-		pp |= PANEL_POWER_RESET; /* restore panel reset bit */
-		I915_WRITE(PCH_PP_CONTROL, pp);
-		POSTING_READ(PCH_PP_CONTROL);
-	}
+	return false;
 }
 
-static void ironlake_edp_panel_off(struct intel_dp *intel_dp)
+static void ironlake_edp_panel_off (struct drm_device *dev)
 {
-	struct drm_device *dev = intel_dp->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 pp;
+	u32 pp, idle_off_mask = PP_ON | PP_SEQUENCE_MASK |
+		PP_CYCLE_DELAY_ACTIVE | PP_SEQUENCE_STATE_MASK;
 
-	if (!is_edp(intel_dp))
-		return;
-
-	DRM_DEBUG_KMS("Turn eDP power off\n");
+	pp = I915_READ(PCH_PP_CONTROL);
 
-	WARN(intel_dp->want_panel_vdd, "Cannot turn power off while VDD is on\n");
+	/* ILK workaround: disable reset around power sequence */
+	pp &= ~PANEL_POWER_RESET;
+	I915_WRITE(PCH_PP_CONTROL, pp);
+	POSTING_READ(PCH_PP_CONTROL);
 
-	pp = ironlake_get_pp_control(dev_priv);
-	pp &= ~(POWER_TARGET_ON | EDP_FORCE_VDD | PANEL_POWER_RESET | EDP_BLC_ENABLE);
+	pp &= ~POWER_TARGET_ON;
 	I915_WRITE(PCH_PP_CONTROL, pp);
 	POSTING_READ(PCH_PP_CONTROL);
 
-	ironlake_wait_panel_off(intel_dp);
+	if (wait_for((I915_READ(PCH_PP_STATUS) & idle_off_mask) == 0, 5000))
+		DRM_ERROR("panel off wait timed out: 0x%08x\n",
+			  I915_READ(PCH_PP_STATUS));
+
+	pp |= PANEL_POWER_RESET; /* restore panel reset bit */
+	I915_WRITE(PCH_PP_CONTROL, pp);
+	POSTING_READ(PCH_PP_CONTROL);
 }
 
-static void ironlake_edp_backlight_on(struct intel_dp *intel_dp)
+static void ironlake_edp_backlight_on (struct drm_device *dev)
 {
-	struct drm_device *dev = intel_dp->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 pp;
 
-	if (!is_edp(intel_dp))
-		return;
-
 	DRM_DEBUG_KMS("\n");
 	/*
 	 * If we enable the backlight right away following a panel power
@@ -1163,28 +909,21 @@ static void ironlake_edp_backlight_on(struct intel_dp *intel_dp)
 	 * link.  So delay a bit to make sure the image is solid before
 	 * allowing it to appear.
 	 */
-	msleep(intel_dp->backlight_on_delay);
-	pp = ironlake_get_pp_control(dev_priv);
+	msleep(300);
+	pp = I915_READ(PCH_PP_CONTROL);
 	pp |= EDP_BLC_ENABLE;
 	I915_WRITE(PCH_PP_CONTROL, pp);
-	POSTING_READ(PCH_PP_CONTROL);
 }
 
-static void ironlake_edp_backlight_off(struct intel_dp *intel_dp)
+static void ironlake_edp_backlight_off (struct drm_device *dev)
 {
-	struct drm_device *dev = intel_dp->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	u32 pp;
 
-	if (!is_edp(intel_dp))
-		return;
-
 	DRM_DEBUG_KMS("\n");
-	pp = ironlake_get_pp_control(dev_priv);
+	pp = I915_READ(PCH_PP_CONTROL);
 	pp &= ~EDP_BLC_ENABLE;
 	I915_WRITE(PCH_PP_CONTROL, pp);
-	POSTING_READ(PCH_PP_CONTROL);
-	msleep(intel_dp->backlight_off_delay);
 }
 
 static void ironlake_edp_pll_on(struct drm_encoder *encoder)
@@ -1247,39 +986,41 @@ static void intel_dp_sink_dpms(struct intel_dp *intel_dp, int mode)
 static void intel_dp_prepare(struct drm_encoder *encoder)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
-
-	ironlake_edp_backlight_off(intel_dp);
-	ironlake_edp_panel_off(intel_dp);
+	struct drm_device *dev = encoder->dev;
 
 	/* Wake up the sink first */
-	ironlake_edp_panel_vdd_on(intel_dp);
 	intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_ON);
-	intel_dp_link_down(intel_dp);
-	ironlake_edp_panel_vdd_off(intel_dp, false);
 
-	/* Make sure the panel is off before trying to
-	 * change the mode
-	 */
+	if (is_edp(intel_dp)) {
+		ironlake_edp_backlight_off(dev);
+		ironlake_edp_panel_off(dev);
+		if (!is_pch_edp(intel_dp))
+			ironlake_edp_pll_on(encoder);
+		else
+			ironlake_edp_pll_off(encoder);
+	}
+	intel_dp_link_down(intel_dp);
 }
 
 static void intel_dp_commit(struct drm_encoder *encoder)
 {
 	struct intel_dp *intel_dp = enc_to_intel_dp(encoder);
 	struct drm_device *dev = encoder->dev;
-	struct intel_crtc *intel_crtc = to_intel_crtc(intel_dp->base.base.crtc);
 
-	ironlake_edp_panel_vdd_on(intel_dp);
-	intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_ON);
+	if (is_edp(intel_dp))
+		ironlake_edp_panel_vdd_on(intel_dp);
+
 	intel_dp_start_link_train(intel_dp);
-	ironlake_edp_panel_on(intel_dp);
-	ironlake_edp_panel_vdd_off(intel_dp, true);
-	intel_dp_complete_link_train(intel_dp);
-	ironlake_edp_backlight_on(intel_dp);
 
-	intel_dp->dpms_mode = DRM_MODE_DPMS_ON;
+	if (is_edp(intel_dp)) {
+		ironlake_edp_panel_on(intel_dp);
+		ironlake_edp_panel_vdd_off(intel_dp);
+	}
 
-	if (HAS_PCH_CPT(dev))
-		intel_cpt_verify_modeset(dev, intel_crtc->pipe);
+	intel_dp_complete_link_train(intel_dp);
+
+	if (is_edp(intel_dp))
+		ironlake_edp_backlight_on(dev);
 }
 
 static void
@@ -1291,32 +1032,29 @@ intel_dp_dpms(struct drm_encoder *encoder, int mode)
 	uint32_t dp_reg = I915_READ(intel_dp->output_reg);
 
 	if (mode != DRM_MODE_DPMS_ON) {
-		ironlake_edp_backlight_off(intel_dp);
-		ironlake_edp_panel_off(intel_dp);
-
-		ironlake_edp_panel_vdd_on(intel_dp);
+		if (is_edp(intel_dp))
+			ironlake_edp_backlight_off(dev);
 		intel_dp_sink_dpms(intel_dp, mode);
 		intel_dp_link_down(intel_dp);
-		ironlake_edp_panel_vdd_off(intel_dp, false);
-
-		if (is_cpu_edp(intel_dp))
+		if (is_edp(intel_dp))
+			ironlake_edp_panel_off(dev);
+		if (is_edp(intel_dp) && !is_pch_edp(intel_dp))
 			ironlake_edp_pll_off(encoder);
 	} else {
-		if (is_cpu_edp(intel_dp))
-			ironlake_edp_pll_on(encoder);
-
-		ironlake_edp_panel_vdd_on(intel_dp);
+		if (is_edp(intel_dp))
+			ironlake_edp_panel_vdd_on(intel_dp);
 		intel_dp_sink_dpms(intel_dp, mode);
 		if (!(dp_reg & DP_PORT_EN)) {
 			intel_dp_start_link_train(intel_dp);
-			ironlake_edp_panel_on(intel_dp);
-			ironlake_edp_panel_vdd_off(intel_dp, true);
+			if (is_edp(intel_dp)) {
+				ironlake_edp_panel_on(intel_dp);
+				ironlake_edp_panel_vdd_off(intel_dp);
+			}
 			intel_dp_complete_link_train(intel_dp);
-		} else
-			ironlake_edp_panel_vdd_off(intel_dp, false);
-		ironlake_edp_backlight_on(intel_dp);
+		}
+		if (is_edp(intel_dp))
+			ironlake_edp_backlight_on(dev);
 	}
-	intel_dp->dpms_mode = mode;
 }
 
 /*
@@ -1349,11 +1087,11 @@ intel_dp_aux_native_read_retry(struct intel_dp *intel_dp, uint16_t address,
  * link status information
  */
 static bool
-intel_dp_get_link_status(struct intel_dp *intel_dp, uint8_t link_status[DP_LINK_STATUS_SIZE])
+intel_dp_get_link_status(struct intel_dp *intel_dp)
 {
 	return intel_dp_aux_native_read_retry(intel_dp,
 					      DP_LANE0_1_STATUS,
-					      link_status,
+					      intel_dp->link_status,
 					      DP_LINK_STATUS_SIZE);
 }
 
@@ -1365,25 +1103,27 @@ intel_dp_link_status(uint8_t link_status[DP_LINK_STATUS_SIZE],
 }
 
 static uint8_t
-intel_get_adjust_request_voltage(uint8_t adjust_request[2],
+intel_get_adjust_request_voltage(uint8_t link_status[DP_LINK_STATUS_SIZE],
 				 int lane)
 {
+	int	    i = DP_ADJUST_REQUEST_LANE0_1 + (lane >> 1);
 	int	    s = ((lane & 1) ?
 			 DP_ADJUST_VOLTAGE_SWING_LANE1_SHIFT :
 			 DP_ADJUST_VOLTAGE_SWING_LANE0_SHIFT);
-	uint8_t l = adjust_request[lane>>1];
+	uint8_t l = intel_dp_link_status(link_status, i);
 
 	return ((l >> s) & 3) << DP_TRAIN_VOLTAGE_SWING_SHIFT;
 }
 
 static uint8_t
-intel_get_adjust_request_pre_emphasis(uint8_t adjust_request[2],
+intel_get_adjust_request_pre_emphasis(uint8_t link_status[DP_LINK_STATUS_SIZE],
 				      int lane)
 {
+	int	    i = DP_ADJUST_REQUEST_LANE0_1 + (lane >> 1);
 	int	    s = ((lane & 1) ?
 			 DP_ADJUST_PRE_EMPHASIS_LANE1_SHIFT :
 			 DP_ADJUST_PRE_EMPHASIS_LANE0_SHIFT);
-	uint8_t l = adjust_request[lane>>1];
+	uint8_t l = intel_dp_link_status(link_status, i);
 
 	return ((l >> s) & 3) << DP_TRAIN_PRE_EMPHASIS_SHIFT;
 }
@@ -1405,63 +1145,34 @@ static char	*link_train_names[] = {
  * These are source-specific values; current Intel hardware supports
  * a maximum voltage of 800mV and a maximum pre-emphasis of 6dB
  */
+#define I830_DP_VOLTAGE_MAX	    DP_TRAIN_VOLTAGE_SWING_800
 
 static uint8_t
-intel_dp_voltage_max(struct intel_dp *intel_dp)
+intel_dp_pre_emphasis_max(uint8_t voltage_swing)
 {
-	struct drm_device *dev = intel_dp->base.base.dev;
-
-	if (IS_GEN7(dev) && is_cpu_edp(intel_dp))
-		return DP_TRAIN_VOLTAGE_SWING_800;
-	else if (HAS_PCH_CPT(dev) && !is_cpu_edp(intel_dp))
-		return DP_TRAIN_VOLTAGE_SWING_1200;
-	else
-		return DP_TRAIN_VOLTAGE_SWING_800;
-}
-
-static uint8_t
-intel_dp_pre_emphasis_max(struct intel_dp *intel_dp, uint8_t voltage_swing)
-{
-	struct drm_device *dev = intel_dp->base.base.dev;
-
-	if (IS_GEN7(dev) && is_cpu_edp(intel_dp)) {
-		switch (voltage_swing & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
-			return DP_TRAIN_PRE_EMPHASIS_6;
-		case DP_TRAIN_VOLTAGE_SWING_600:
-		case DP_TRAIN_VOLTAGE_SWING_800:
-			return DP_TRAIN_PRE_EMPHASIS_3_5;
-		default:
-			return DP_TRAIN_PRE_EMPHASIS_0;
-		}
-	} else {
-		switch (voltage_swing & DP_TRAIN_VOLTAGE_SWING_MASK) {
-		case DP_TRAIN_VOLTAGE_SWING_400:
-			return DP_TRAIN_PRE_EMPHASIS_6;
-		case DP_TRAIN_VOLTAGE_SWING_600:
-			return DP_TRAIN_PRE_EMPHASIS_6;
-		case DP_TRAIN_VOLTAGE_SWING_800:
-			return DP_TRAIN_PRE_EMPHASIS_3_5;
-		case DP_TRAIN_VOLTAGE_SWING_1200:
-		default:
-			return DP_TRAIN_PRE_EMPHASIS_0;
-		}
+	switch (voltage_swing & DP_TRAIN_VOLTAGE_SWING_MASK) {
+	case DP_TRAIN_VOLTAGE_SWING_400:
+		return DP_TRAIN_PRE_EMPHASIS_6;
+	case DP_TRAIN_VOLTAGE_SWING_600:
+		return DP_TRAIN_PRE_EMPHASIS_6;
+	case DP_TRAIN_VOLTAGE_SWING_800:
+		return DP_TRAIN_PRE_EMPHASIS_3_5;
+	case DP_TRAIN_VOLTAGE_SWING_1200:
+	default:
+		return DP_TRAIN_PRE_EMPHASIS_0;
 	}
 }
 
 static void
-intel_get_adjust_train(struct intel_dp *intel_dp, uint8_t link_status[DP_LINK_STATUS_SIZE])
+intel_get_adjust_train(struct intel_dp *intel_dp)
 {
 	uint8_t v = 0;
 	uint8_t p = 0;
 	int lane;
-	uint8_t	*adjust_request = link_status + (DP_ADJUST_REQUEST_LANE0_1 - DP_LANE0_1_STATUS);
-	uint8_t voltage_max;
-	uint8_t preemph_max;
 
 	for (lane = 0; lane < intel_dp->lane_count; lane++) {
-		uint8_t this_v = intel_get_adjust_request_voltage(adjust_request, lane);
-		uint8_t this_p = intel_get_adjust_request_pre_emphasis(adjust_request, lane);
+		uint8_t this_v = intel_get_adjust_request_voltage(intel_dp->link_status, lane);
+		uint8_t this_p = intel_get_adjust_request_pre_emphasis(intel_dp->link_status, lane);
 
 		if (this_v > v)
 			v = this_v;
@@ -1469,20 +1180,18 @@ intel_get_adjust_train(struct intel_dp *intel_dp, uint8_t link_status[DP_LINK_ST
 			p = this_p;
 	}
 
-	voltage_max = intel_dp_voltage_max(intel_dp);
-	if (v >= voltage_max)
-		v = voltage_max | DP_TRAIN_MAX_SWING_REACHED;
+	if (v >= I830_DP_VOLTAGE_MAX)
+		v = I830_DP_VOLTAGE_MAX | DP_TRAIN_MAX_SWING_REACHED;
 
-	preemph_max = intel_dp_pre_emphasis_max(intel_dp, v);
-	if (p >= preemph_max)
-		p = preemph_max | DP_TRAIN_MAX_PRE_EMPHASIS_REACHED;
+	if (p >= intel_dp_pre_emphasis_max(v))
+		p = intel_dp_pre_emphasis_max(v) | DP_TRAIN_MAX_PRE_EMPHASIS_REACHED;
 
 	for (lane = 0; lane < 4; lane++)
 		intel_dp->train_set[lane] = v | p;
 }
 
 static uint32_t
-intel_dp_signal_levels(uint8_t train_set)
+intel_dp_signal_levels(uint8_t train_set, int lane_count)
 {
 	uint32_t	signal_levels = 0;
 
@@ -1547,43 +1256,13 @@ intel_gen6_edp_signal_levels(uint8_t train_set)
 	}
 }
 
-/* Gen7's DP voltage swing and pre-emphasis control */
-static uint32_t
-intel_gen7_edp_signal_levels(uint8_t train_set)
-{
-	int signal_levels = train_set & (DP_TRAIN_VOLTAGE_SWING_MASK |
-					 DP_TRAIN_PRE_EMPHASIS_MASK);
-	switch (signal_levels) {
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_0:
-		return EDP_LINK_TRAIN_400MV_0DB_IVB;
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_3_5:
-		return EDP_LINK_TRAIN_400MV_3_5DB_IVB;
-	case DP_TRAIN_VOLTAGE_SWING_400 | DP_TRAIN_PRE_EMPHASIS_6:
-		return EDP_LINK_TRAIN_400MV_6DB_IVB;
-
-	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_0:
-		return EDP_LINK_TRAIN_600MV_0DB_IVB;
-	case DP_TRAIN_VOLTAGE_SWING_600 | DP_TRAIN_PRE_EMPHASIS_3_5:
-		return EDP_LINK_TRAIN_600MV_3_5DB_IVB;
-
-	case DP_TRAIN_VOLTAGE_SWING_800 | DP_TRAIN_PRE_EMPHASIS_0:
-		return EDP_LINK_TRAIN_800MV_0DB_IVB;
-	case DP_TRAIN_VOLTAGE_SWING_800 | DP_TRAIN_PRE_EMPHASIS_3_5:
-		return EDP_LINK_TRAIN_800MV_3_5DB_IVB;
-
-	default:
-		DRM_DEBUG_KMS("Unsupported voltage swing/pre-emphasis level:"
-			      "0x%x\n", signal_levels);
-		return EDP_LINK_TRAIN_500MV_0DB_IVB;
-	}
-}
-
 static uint8_t
 intel_get_lane_status(uint8_t link_status[DP_LINK_STATUS_SIZE],
 		      int lane)
 {
+	int i = DP_LANE0_1_STATUS + (lane >> 1);
 	int s = (lane & 1) * 4;
-	uint8_t l = link_status[lane>>1];
+	uint8_t l = intel_dp_link_status(link_status, i);
 
 	return (l >> s) & 0xf;
 }
@@ -1608,18 +1287,18 @@ intel_clock_recovery_ok(uint8_t link_status[DP_LINK_STATUS_SIZE], int lane_count
 			 DP_LANE_CHANNEL_EQ_DONE|\
 			 DP_LANE_SYMBOL_LOCKED)
 static bool
-intel_channel_eq_ok(struct intel_dp *intel_dp, uint8_t link_status[DP_LINK_STATUS_SIZE])
+intel_channel_eq_ok(struct intel_dp *intel_dp)
 {
 	uint8_t lane_align;
 	uint8_t lane_status;
 	int lane;
 
-	lane_align = intel_dp_link_status(link_status,
+	lane_align = intel_dp_link_status(intel_dp->link_status,
 					  DP_LANE_ALIGN_STATUS_UPDATED);
 	if ((lane_align & DP_INTERLANE_ALIGN_DONE) == 0)
 		return false;
 	for (lane = 0; lane < intel_dp->lane_count; lane++) {
-		lane_status = intel_get_lane_status(link_status, lane);
+		lane_status = intel_get_lane_status(intel_dp->link_status, lane);
 		if ((lane_status & CHANNEL_EQ_BITS) != CHANNEL_EQ_BITS)
 			return false;
 	}
@@ -1644,9 +1323,8 @@ intel_dp_set_link_train(struct intel_dp *intel_dp,
 
 	ret = intel_dp_aux_native_write(intel_dp,
 					DP_TRAINING_LANE0_SET,
-					intel_dp->train_set,
-					intel_dp->lane_count);
-	if (ret != intel_dp->lane_count)
+					intel_dp->train_set, 4);
+	if (ret != 4)
 		return false;
 
 	return true;
@@ -1662,20 +1340,14 @@ intel_dp_start_link_train(struct intel_dp *intel_dp)
 	int i;
 	uint8_t voltage;
 	bool clock_recovery = false;
-	int voltage_tries, loop_tries;
+	int tries;
 	u32 reg;
 	uint32_t DP = intel_dp->DP;
 
-	/*
-	 * On CPT we have to enable the port in training pattern 1, which
-	 * will happen below in intel_dp_set_link_train.  Otherwise, enable
-	 * the port and wait for it to become active.
-	 */
-	if (!HAS_PCH_CPT(dev)) {
-		I915_WRITE(intel_dp->output_reg, intel_dp->DP);
-		POSTING_READ(intel_dp->output_reg);
-		intel_wait_for_vblank(dev, intel_crtc->pipe);
-	}
+	/* Enable output, wait for it to become active */
+	I915_WRITE(intel_dp->output_reg, intel_dp->DP);
+	POSTING_READ(intel_dp->output_reg);
+	intel_wait_for_vblank(dev, intel_crtc->pipe);
 
 	/* Write the link configuration data */
 	intel_dp_aux_native_write(intel_dp, DP_LINK_BW_SET,
@@ -1683,53 +1355,40 @@ intel_dp_start_link_train(struct intel_dp *intel_dp)
 				  DP_LINK_CONFIGURATION_SIZE);
 
 	DP |= DP_PORT_EN;
-
-	if (HAS_PCH_CPT(dev) && (IS_GEN7(dev) || !is_cpu_edp(intel_dp)))
+	if (HAS_PCH_CPT(dev) && !is_edp(intel_dp))
 		DP &= ~DP_LINK_TRAIN_MASK_CPT;
 	else
 		DP &= ~DP_LINK_TRAIN_MASK;
 	memset(intel_dp->train_set, 0, 4);
 	voltage = 0xff;
-	voltage_tries = 0;
-	loop_tries = 0;
+	tries = 0;
 	clock_recovery = false;
 	for (;;) {
 		/* Use intel_dp->train_set[0] to set the voltage and pre emphasis values */
-		uint8_t	    link_status[DP_LINK_STATUS_SIZE];
 		uint32_t    signal_levels;
-
-
-		if (IS_GEN7(dev) && is_cpu_edp(intel_dp)) {
-			signal_levels = intel_gen7_edp_signal_levels(intel_dp->train_set[0]);
-			DP = (DP & ~EDP_LINK_TRAIN_VOL_EMP_MASK_IVB) | signal_levels;
-		} else if (IS_GEN6(dev) && is_cpu_edp(intel_dp)) {
+		if (IS_GEN6(dev) && is_edp(intel_dp)) {
 			signal_levels = intel_gen6_edp_signal_levels(intel_dp->train_set[0]);
 			DP = (DP & ~EDP_LINK_TRAIN_VOL_EMP_MASK_SNB) | signal_levels;
 		} else {
-			signal_levels = intel_dp_signal_levels(intel_dp->train_set[0]);
-			DRM_DEBUG_KMS("training pattern 1 signal levels %08x\n", signal_levels);
+			signal_levels = intel_dp_signal_levels(intel_dp->train_set[0], intel_dp->lane_count);
 			DP = (DP & ~(DP_VOLTAGE_MASK|DP_PRE_EMPHASIS_MASK)) | signal_levels;
 		}
 
-		if (HAS_PCH_CPT(dev) && (IS_GEN7(dev) || !is_cpu_edp(intel_dp)))
+		if (HAS_PCH_CPT(dev) && !is_edp(intel_dp))
 			reg = DP | DP_LINK_TRAIN_PAT_1_CPT;
 		else
 			reg = DP | DP_LINK_TRAIN_PAT_1;
 
 		if (!intel_dp_set_link_train(intel_dp, reg,
-					     DP_TRAINING_PATTERN_1 |
-					     DP_LINK_SCRAMBLING_DISABLE))
+					     DP_TRAINING_PATTERN_1))
 			break;
 		/* Set training pattern 1 */
 
 		udelay(100);
-		if (!intel_dp_get_link_status(intel_dp, link_status)) {
-			DRM_ERROR("failed to get link status\n");
+		if (!intel_dp_get_link_status(intel_dp))
 			break;
-		}
 
-		if (intel_clock_recovery_ok(link_status, intel_dp->lane_count)) {
-			DRM_DEBUG_KMS("clock recovery OK\n");
+		if (intel_clock_recovery_ok(intel_dp->link_status, intel_dp->lane_count)) {
 			clock_recovery = true;
 			break;
 		}
@@ -1738,30 +1397,20 @@ intel_dp_start_link_train(struct intel_dp *intel_dp)
 		for (i = 0; i < intel_dp->lane_count; i++)
 			if ((intel_dp->train_set[i] & DP_TRAIN_MAX_SWING_REACHED) == 0)
 				break;
-		if (i == intel_dp->lane_count) {
-			++loop_tries;
-			if (loop_tries == 5) {
-				DRM_DEBUG_KMS("too many full retries, give up\n");
-				break;
-			}
-			memset(intel_dp->train_set, 0, 4);
-			voltage_tries = 0;
-			continue;
-		}
+		if (i == intel_dp->lane_count)
+			break;
 
 		/* Check to see if we've tried the same voltage 5 times */
 		if ((intel_dp->train_set[0] & DP_TRAIN_VOLTAGE_SWING_MASK) == voltage) {
-			++voltage_tries;
-			if (voltage_tries == 5) {
-				DRM_DEBUG_KMS("too many voltage retries, give up\n");
+			++tries;
+			if (tries == 5)
 				break;
-			}
 		} else
-			voltage_tries = 0;
+			tries = 0;
 		voltage = intel_dp->train_set[0] & DP_TRAIN_VOLTAGE_SWING_MASK;
 
 		/* Compute new intel_dp->train_set as requested by target */
-		intel_get_adjust_train(intel_dp, link_status);
+		intel_get_adjust_train(intel_dp);
 	}
 
 	intel_dp->DP = DP;
@@ -1784,7 +1433,6 @@ intel_dp_complete_link_train(struct intel_dp *intel_dp)
 	for (;;) {
 		/* Use intel_dp->train_set[0] to set the voltage and pre emphasis values */
 		uint32_t    signal_levels;
-		uint8_t	    link_status[DP_LINK_STATUS_SIZE];
 
 		if (cr_tries > 5) {
 			DRM_ERROR("failed to train DP, aborting\n");
@@ -1792,40 +1440,36 @@ intel_dp_complete_link_train(struct intel_dp *intel_dp)
 			break;
 		}
 
-		if (IS_GEN7(dev) && is_cpu_edp(intel_dp)) {
-			signal_levels = intel_gen7_edp_signal_levels(intel_dp->train_set[0]);
-			DP = (DP & ~EDP_LINK_TRAIN_VOL_EMP_MASK_IVB) | signal_levels;
-		} else if (IS_GEN6(dev) && is_cpu_edp(intel_dp)) {
+		if (IS_GEN6(dev) && is_edp(intel_dp)) {
 			signal_levels = intel_gen6_edp_signal_levels(intel_dp->train_set[0]);
 			DP = (DP & ~EDP_LINK_TRAIN_VOL_EMP_MASK_SNB) | signal_levels;
 		} else {
-			signal_levels = intel_dp_signal_levels(intel_dp->train_set[0]);
+			signal_levels = intel_dp_signal_levels(intel_dp->train_set[0], intel_dp->lane_count);
 			DP = (DP & ~(DP_VOLTAGE_MASK|DP_PRE_EMPHASIS_MASK)) | signal_levels;
 		}
 
-		if (HAS_PCH_CPT(dev) && (IS_GEN7(dev) || !is_cpu_edp(intel_dp)))
+		if (HAS_PCH_CPT(dev) && !is_edp(intel_dp))
 			reg = DP | DP_LINK_TRAIN_PAT_2_CPT;
 		else
 			reg = DP | DP_LINK_TRAIN_PAT_2;
 
 		/* channel eq pattern */
 		if (!intel_dp_set_link_train(intel_dp, reg,
-					     DP_TRAINING_PATTERN_2 |
-					     DP_LINK_SCRAMBLING_DISABLE))
+					     DP_TRAINING_PATTERN_2))
 			break;
 
 		udelay(400);
-		if (!intel_dp_get_link_status(intel_dp, link_status))
+		if (!intel_dp_get_link_status(intel_dp))
 			break;
 
 		/* Make sure clock is still ok */
-		if (!intel_clock_recovery_ok(link_status, intel_dp->lane_count)) {
+		if (!intel_clock_recovery_ok(intel_dp->link_status, intel_dp->lane_count)) {
 			intel_dp_start_link_train(intel_dp);
 			cr_tries++;
 			continue;
 		}
 
-		if (intel_channel_eq_ok(intel_dp, link_status)) {
+		if (intel_channel_eq_ok(intel_dp)) {
 			channel_eq = true;
 			break;
 		}
@@ -1840,11 +1484,11 @@ intel_dp_complete_link_train(struct intel_dp *intel_dp)
 		}
 
 		/* Compute new intel_dp->train_set as requested by target */
-		intel_get_adjust_train(intel_dp, link_status);
+		intel_get_adjust_train(intel_dp);
 		++tries;
 	}
 
-	if (HAS_PCH_CPT(dev) && (IS_GEN7(dev) || !is_cpu_edp(intel_dp)))
+	if (HAS_PCH_CPT(dev) && !is_edp(intel_dp))
 		reg = DP | DP_LINK_TRAIN_OFF_CPT;
 	else
 		reg = DP | DP_LINK_TRAIN_OFF;
@@ -1874,7 +1518,7 @@ intel_dp_link_down(struct intel_dp *intel_dp)
 		udelay(100);
 	}
 
-	if (HAS_PCH_CPT(dev) && (IS_GEN7(dev) || !is_cpu_edp(intel_dp))) {
+	if (HAS_PCH_CPT(dev) && !is_edp(intel_dp)) {
 		DP &= ~DP_LINK_TRAIN_MASK_CPT;
 		I915_WRITE(intel_dp->output_reg, DP | DP_LINK_TRAIN_PAT_IDLE_CPT);
 	} else {
@@ -1885,12 +1529,8 @@ intel_dp_link_down(struct intel_dp *intel_dp)
 
 	msleep(17);
 
-	if (is_edp(intel_dp)) {
-		if (HAS_PCH_CPT(dev) && (IS_GEN7(dev) || !is_cpu_edp(intel_dp)))
-			DP |= DP_LINK_TRAIN_OFF_CPT;
-		else
-			DP |= DP_LINK_TRAIN_OFF;
-	}
+	if (is_edp(intel_dp))
+		DP |= DP_LINK_TRAIN_OFF;
 
 	if (!HAS_PCH_CPT(dev) &&
 	    I915_READ(intel_dp->output_reg) & DP_PIPEB_SELECT) {
@@ -1928,40 +1568,6 @@ intel_dp_link_down(struct intel_dp *intel_dp)
 	DP &= ~DP_AUDIO_OUTPUT_ENABLE;
 	I915_WRITE(intel_dp->output_reg, DP & ~DP_PORT_EN);
 	POSTING_READ(intel_dp->output_reg);
-	msleep(intel_dp->panel_power_down_delay);
-}
-
-static bool
-intel_dp_get_dpcd(struct intel_dp *intel_dp)
-{
-	if (intel_dp_aux_native_read_retry(intel_dp, 0x000, intel_dp->dpcd,
-					   sizeof(intel_dp->dpcd)) &&
-	    (intel_dp->dpcd[DP_DPCD_REV] != 0)) {
-		return true;
-	}
-
-	return false;
-}
-
-static bool
-intel_dp_get_sink_irq(struct intel_dp *intel_dp, u8 *sink_irq_vector)
-{
-	int ret;
-
-	ret = intel_dp_aux_native_read_retry(intel_dp,
-					     DP_DEVICE_SERVICE_IRQ_VECTOR,
-					     sink_irq_vector, 1);
-	if (!ret)
-		return false;
-
-	return true;
-}
-
-static void
-intel_dp_handle_test_request(struct intel_dp *intel_dp)
-{
-	/* NAK by default */
-	intel_dp_aux_native_write_1(intel_dp, DP_TEST_RESPONSE, DP_TEST_ACK);
 }
 
 /*
@@ -1976,61 +1582,36 @@ intel_dp_handle_test_request(struct intel_dp *intel_dp)
 static void
 intel_dp_check_link_status(struct intel_dp *intel_dp)
 {
-	u8 sink_irq_vector;
-	u8 link_status[DP_LINK_STATUS_SIZE];
-
-	if (intel_dp->dpms_mode != DRM_MODE_DPMS_ON)
-		return;
+	int ret;
 
 	if (!intel_dp->base.base.crtc)
 		return;
 
-	/* Try to read receiver status if the link appears to be up */
-	if (!intel_dp_get_link_status(intel_dp, link_status)) {
+	if (!intel_dp_get_link_status(intel_dp)) {
 		intel_dp_link_down(intel_dp);
 		return;
 	}
 
-	/* Now read the DPCD to see if it's actually running */
-	if (!intel_dp_get_dpcd(intel_dp)) {
+	/* Try to read receiver status if the link appears to be up */
+	ret = intel_dp_aux_native_read(intel_dp,
+				       0x000, intel_dp->dpcd,
+				       sizeof (intel_dp->dpcd));
+	if (ret != sizeof(intel_dp->dpcd)) {
 		intel_dp_link_down(intel_dp);
 		return;
 	}
 
-	/* Try to read the source of the interrupt */
-	if (intel_dp->dpcd[DP_DPCD_REV] >= 0x11 &&
-	    intel_dp_get_sink_irq(intel_dp, &sink_irq_vector)) {
-		/* Clear interrupt source */
-		intel_dp_aux_native_write_1(intel_dp,
-					    DP_DEVICE_SERVICE_IRQ_VECTOR,
-					    sink_irq_vector);
-
-		if (sink_irq_vector & DP_AUTOMATED_TEST_REQUEST)
-			intel_dp_handle_test_request(intel_dp);
-		if (sink_irq_vector & (DP_CP_IRQ | DP_SINK_SPECIFIC_IRQ))
-			DRM_DEBUG_DRIVER("CP or sink specific irq unhandled\n");
-	}
-
-	if (!intel_channel_eq_ok(intel_dp, link_status)) {
-		DRM_DEBUG_KMS("%s: channel EQ not ok, retraining\n",
-			      drm_get_encoder_name(&intel_dp->base.base));
+	if (!intel_channel_eq_ok(intel_dp)) {
 		intel_dp_start_link_train(intel_dp);
 		intel_dp_complete_link_train(intel_dp);
 	}
 }
 
 static enum drm_connector_status
-intel_dp_detect_dpcd(struct intel_dp *intel_dp)
-{
-	if (intel_dp_get_dpcd(intel_dp))
-		return connector_status_connected;
-	return connector_status_disconnected;
-}
-
-static enum drm_connector_status
 ironlake_dp_detect(struct intel_dp *intel_dp)
 {
 	enum drm_connector_status status;
+	bool ret;
 
 	/* Can't disconnect eDP, but you can close the lid... */
 	if (is_edp(intel_dp)) {
@@ -2040,7 +1621,15 @@ ironlake_dp_detect(struct intel_dp *intel_dp)
 		return status;
 	}
 
-	return intel_dp_detect_dpcd(intel_dp);
+	status = connector_status_disconnected;
+	ret = intel_dp_aux_native_read_retry(intel_dp,
+					     0x000, intel_dp->dpcd,
+					     sizeof (intel_dp->dpcd));
+	if (ret && intel_dp->dpcd[DP_DPCD_REV] != 0)
+		status = connector_status_connected;
+	DRM_DEBUG_KMS("DPCD: %hx%hx%hx%hx\n", intel_dp->dpcd[0],
+		      intel_dp->dpcd[1], intel_dp->dpcd[2], intel_dp->dpcd[3]);
+	return status;
 }
 
 static enum drm_connector_status
@@ -2048,6 +1637,7 @@ g4x_dp_detect(struct intel_dp *intel_dp)
 {
 	struct drm_device *dev = intel_dp->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
+	enum drm_connector_status status;
 	uint32_t temp, bit;
 
 	switch (intel_dp->output_reg) {
@@ -2069,7 +1659,15 @@ g4x_dp_detect(struct intel_dp *intel_dp)
 	if ((temp & bit) == 0)
 		return connector_status_disconnected;
 
-	return intel_dp_detect_dpcd(intel_dp);
+	status = connector_status_disconnected;
+	if (intel_dp_aux_native_read(intel_dp, 0x000, intel_dp->dpcd,
+				     sizeof (intel_dp->dpcd)) == sizeof (intel_dp->dpcd))
+	{
+		if (intel_dp->dpcd[DP_DPCD_REV] != 0)
+			status = connector_status_connected;
+	}
+
+	return status;
 }
 
 static struct edid *
@@ -2080,7 +1678,7 @@ intel_dp_get_edid(struct drm_connector *connector, struct i2c_adapter *adapter)
 
 	ironlake_edp_panel_vdd_on(intel_dp);
 	edid = drm_get_edid(connector, adapter);
-	ironlake_edp_panel_vdd_off(intel_dp, false);
+	ironlake_edp_panel_vdd_off(intel_dp);
 	return edid;
 }
 
@@ -2092,7 +1690,7 @@ intel_dp_get_edid_modes(struct drm_connector *connector, struct i2c_adapter *ada
 
 	ironlake_edp_panel_vdd_on(intel_dp);
 	ret = intel_ddc_get_modes(connector, adapter);
-	ironlake_edp_panel_vdd_off(intel_dp, false);
+	ironlake_edp_panel_vdd_off(intel_dp);
 	return ret;
 }
 
@@ -2117,17 +1715,11 @@ intel_dp_detect(struct drm_connector *connector, bool force)
 		status = ironlake_dp_detect(intel_dp);
 	else
 		status = g4x_dp_detect(intel_dp);
-
-	DRM_DEBUG_KMS("DPCD: %02hx%02hx%02hx%02hx%02hx%02hx%02hx%02hx\n",
-		      intel_dp->dpcd[0], intel_dp->dpcd[1], intel_dp->dpcd[2],
-		      intel_dp->dpcd[3], intel_dp->dpcd[4], intel_dp->dpcd[5],
-		      intel_dp->dpcd[6], intel_dp->dpcd[7]);
-
 	if (status != connector_status_connected)
 		return status;
 
-	if (intel_dp->force_audio != HDMI_AUDIO_AUTO) {
-		intel_dp->has_audio = (intel_dp->force_audio == HDMI_AUDIO_ON);
+	if (intel_dp->force_audio) {
+		intel_dp->has_audio = intel_dp->force_audio > 0;
 	} else {
 		edid = intel_dp_get_edid(connector, &intel_dp->adapter);
 		if (edid) {
@@ -2152,34 +1744,26 @@ static int intel_dp_get_modes(struct drm_connector *connector)
 
 	ret = intel_dp_get_edid_modes(connector, &intel_dp->adapter);
 	if (ret) {
-		if (is_edp(intel_dp) && !intel_dp->panel_fixed_mode) {
+		if (is_edp(intel_dp) && !dev_priv->panel_fixed_mode) {
 			struct drm_display_mode *newmode;
 			list_for_each_entry(newmode, &connector->probed_modes,
 					    head) {
-				if ((newmode->type & DRM_MODE_TYPE_PREFERRED)) {
-					intel_dp->panel_fixed_mode =
+				if (newmode->type & DRM_MODE_TYPE_PREFERRED) {
+					dev_priv->panel_fixed_mode =
 						drm_mode_duplicate(dev, newmode);
 					break;
 				}
 			}
 		}
+
 		return ret;
 	}
 
 	/* if eDP has no EDID, try to use fixed panel mode from VBT */
 	if (is_edp(intel_dp)) {
-		/* initialize panel mode from VBT if available for eDP */
-		if (intel_dp->panel_fixed_mode == NULL && dev_priv->lfp_lvds_vbt_mode != NULL) {
-			intel_dp->panel_fixed_mode =
-				drm_mode_duplicate(dev, dev_priv->lfp_lvds_vbt_mode);
-			if (intel_dp->panel_fixed_mode) {
-				intel_dp->panel_fixed_mode->type |=
-					DRM_MODE_TYPE_PREFERRED;
-			}
-		}
-		if (intel_dp->panel_fixed_mode) {
+		if (dev_priv->panel_fixed_mode != NULL) {
 			struct drm_display_mode *mode;
-			mode = drm_mode_duplicate(dev, intel_dp->panel_fixed_mode);
+			mode = drm_mode_duplicate(dev, dev_priv->panel_fixed_mode);
 			drm_mode_probed_add(connector, mode);
 			return 1;
 		}
@@ -2227,10 +1811,10 @@ intel_dp_set_property(struct drm_connector *connector,
 
 		intel_dp->force_audio = i;
 
-		if (i == HDMI_AUDIO_AUTO)
+		if (i == 0)
 			has_audio = intel_dp_detect_audio(connector);
 		else
-			has_audio = (i == HDMI_AUDIO_ON);
+			has_audio = i > 0;
 
 		if (has_audio == intel_dp->has_audio)
 			return 0;
@@ -2261,13 +1845,8 @@ done:
 }
 
 static void
-intel_dp_destroy(struct drm_connector *connector)
+intel_dp_destroy (struct drm_connector *connector)
 {
-	struct drm_device *dev = connector->dev;
-
-	if (intel_dpd_is_edp(dev))
-		intel_panel_destroy_backlight(dev);
-
 	drm_sysfs_connector_remove(connector);
 	drm_connector_cleanup(connector);
 	kfree(connector);
@@ -2279,10 +1858,6 @@ static void intel_dp_encoder_destroy(struct drm_encoder *encoder)
 
 	i2c_del_adapter(&intel_dp->adapter);
 	drm_encoder_cleanup(encoder);
-	if (is_edp(intel_dp)) {
-		cancel_delayed_work_sync(&intel_dp->panel_vdd_work);
-		ironlake_panel_vdd_off_sync(intel_dp);
-	}
 	kfree(intel_dp);
 }
 
@@ -2322,7 +1897,7 @@ intel_dp_hot_plug(struct intel_encoder *intel_encoder)
 
 /* Return which DP Port should be selected for Transcoder DP control */
 int
-intel_trans_dp_port_sel(struct drm_crtc *crtc)
+intel_trans_dp_port_sel (struct drm_crtc *crtc)
 {
 	struct drm_device *dev = crtc->dev;
 	struct drm_mode_config *mode_config = &dev->mode_config;
@@ -2335,8 +1910,7 @@ intel_trans_dp_port_sel(struct drm_crtc *crtc)
 			continue;
 
 		intel_dp = enc_to_intel_dp(encoder);
-		if (intel_dp->base.type == INTEL_OUTPUT_DISPLAYPORT ||
-		    intel_dp->base.type == INTEL_OUTPUT_EDP)
+		if (intel_dp->base.type == INTEL_OUTPUT_DISPLAYPORT)
 			return intel_dp->output_reg;
 	}
 
@@ -2386,7 +1960,6 @@ intel_dp_init(struct drm_device *dev, int output_reg)
 		return;
 
 	intel_dp->output_reg = output_reg;
-	intel_dp->dpms_mode = -1;
 
 	intel_connector = kzalloc(sizeof(struct intel_connector), GFP_KERNEL);
 	if (!intel_connector) {
@@ -2420,13 +1993,10 @@ intel_dp_init(struct drm_device *dev, int output_reg)
 	else if (output_reg == DP_D || output_reg == PCH_DP_D)
 		intel_encoder->clone_mask = (1 << INTEL_DP_D_CLONE_BIT);
 
-	if (is_edp(intel_dp)) {
+	if (is_edp(intel_dp))
 		intel_encoder->clone_mask = (1 << INTEL_EDP_CLONE_BIT);
-		INIT_DELAYED_WORK(&intel_dp->panel_vdd_work,
-				  ironlake_panel_vdd_work);
-	}
 
-	intel_encoder->crtc_mask = (1 << 0) | (1 << 1) | (1 << 2);
+	intel_encoder->crtc_mask = (1 << 0) | (1 << 1);
 	connector->interlace_allowed = true;
 	connector->doublescan_allowed = 0;
 
@@ -2462,60 +2032,28 @@ intel_dp_init(struct drm_device *dev, int output_reg)
 			break;
 	}
 
+	intel_dp_i2c_init(intel_dp, intel_connector, name);
+
 	/* Cache some DPCD data in the eDP case */
 	if (is_edp(intel_dp)) {
-		bool ret;
-		struct edp_power_seq	cur, vbt;
-		u32 pp_on, pp_off, pp_div;
+		int ret;
+		u32 pp_on, pp_div;
 
 		pp_on = I915_READ(PCH_PP_ON_DELAYS);
-		pp_off = I915_READ(PCH_PP_OFF_DELAYS);
 		pp_div = I915_READ(PCH_PP_DIVISOR);
 
-		/* Pull timing values out of registers */
-		cur.t1_t3 = (pp_on & PANEL_POWER_UP_DELAY_MASK) >>
-			PANEL_POWER_UP_DELAY_SHIFT;
-
-		cur.t8 = (pp_on & PANEL_LIGHT_ON_DELAY_MASK) >>
-			PANEL_LIGHT_ON_DELAY_SHIFT;
-
-		cur.t9 = (pp_off & PANEL_LIGHT_OFF_DELAY_MASK) >>
-			PANEL_LIGHT_OFF_DELAY_SHIFT;
-
-		cur.t10 = (pp_off & PANEL_POWER_DOWN_DELAY_MASK) >>
-			PANEL_POWER_DOWN_DELAY_SHIFT;
-
-		cur.t11_t12 = ((pp_div & PANEL_POWER_CYCLE_DELAY_MASK) >>
-			       PANEL_POWER_CYCLE_DELAY_SHIFT) * 1000;
-
-		DRM_DEBUG_KMS("cur t1_t3 %d t8 %d t9 %d t10 %d t11_t12 %d\n",
-			      cur.t1_t3, cur.t8, cur.t9, cur.t10, cur.t11_t12);
-
-		vbt = dev_priv->edp.pps;
-
-		DRM_DEBUG_KMS("vbt t1_t3 %d t8 %d t9 %d t10 %d t11_t12 %d\n",
-			      vbt.t1_t3, vbt.t8, vbt.t9, vbt.t10, vbt.t11_t12);
-
-#define get_delay(field)	((max(cur.field, vbt.field) + 9) / 10)
-
-		intel_dp->panel_power_up_delay = get_delay(t1_t3);
-		intel_dp->backlight_on_delay = get_delay(t8);
-		intel_dp->backlight_off_delay = get_delay(t9);
-		intel_dp->panel_power_down_delay = get_delay(t10);
-		intel_dp->panel_power_cycle_delay = get_delay(t11_t12);
-
-		DRM_DEBUG_KMS("panel power up delay %d, power down delay %d, power cycle delay %d\n",
-			      intel_dp->panel_power_up_delay, intel_dp->panel_power_down_delay,
-			      intel_dp->panel_power_cycle_delay);
-
-		DRM_DEBUG_KMS("backlight on delay %d, off delay %d\n",
-			      intel_dp->backlight_on_delay, intel_dp->backlight_off_delay);
+		/* Get T3 & T12 values (note: VESA not bspec terminology) */
+		dev_priv->panel_t3 = (pp_on & 0x1fff0000) >> 16;
+		dev_priv->panel_t3 /= 10; /* t3 in 100us units */
+		dev_priv->panel_t12 = pp_div & 0xf;
+		dev_priv->panel_t12 *= 100; /* t12 in 100ms units */
 
 		ironlake_edp_panel_vdd_on(intel_dp);
-		ret = intel_dp_get_dpcd(intel_dp);
-		ironlake_edp_panel_vdd_off(intel_dp, false);
-
-		if (ret) {
+		ret = intel_dp_aux_native_read(intel_dp, DP_DPCD_REV,
+					       intel_dp->dpcd,
+					       sizeof(intel_dp->dpcd));
+		ironlake_edp_panel_vdd_off(intel_dp);
+		if (ret == sizeof(intel_dp->dpcd)) {
 			if (intel_dp->dpcd[DP_DPCD_REV] >= 0x11)
 				dev_priv->no_aux_handshake =
 					intel_dp->dpcd[DP_MAX_DOWNSPREAD] &
@@ -2529,13 +2067,18 @@ intel_dp_init(struct drm_device *dev, int output_reg)
 		}
 	}
 
-	intel_dp_i2c_init(intel_dp, intel_connector, name);
-
 	intel_encoder->hot_plug = intel_dp_hot_plug;
 
 	if (is_edp(intel_dp)) {
-		dev_priv->int_edp_connector = connector;
-		intel_panel_setup_backlight(dev);
+		/* initialize panel mode from VBT if available for eDP */
+		if (dev_priv->lfp_lvds_vbt_mode) {
+			dev_priv->panel_fixed_mode =
+				drm_mode_duplicate(dev, dev_priv->lfp_lvds_vbt_mode);
+			if (dev_priv->panel_fixed_mode) {
+				dev_priv->panel_fixed_mode->type |=
+					DRM_MODE_TYPE_PREFERRED;
+			}
+		}
 	}
 
 	intel_dp_add_properties(intel_dp, connector);
diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h
index 7c31daf..58b54ff 100644
--- a/drivers/gpu/drm/i915/intel_drv.h
+++ b/drivers/gpu/drm/i915/intel_drv.h
@@ -26,7 +26,6 @@
 #define __INTEL_DRV_H__
 
 #include <linux/i2c.h>
-#include "i915_drm.h"
 #include "i915_drv.h"
 #include "drm_crtc.h"
 #include "drm_crtc_helper.h"
@@ -35,12 +34,12 @@
 #define _wait_for(COND, MS, W) ({ \
 	unsigned long timeout__ = jiffies + msecs_to_jiffies(MS);	\
 	int ret__ = 0;							\
-	while (!(COND)) {						\
+	while (! (COND)) {						\
 		if (time_after(jiffies, timeout__)) {			\
 			ret__ = -ETIMEDOUT;				\
 			break;						\
 		}							\
-		if (W && drm_can_sleep()) msleep(W);	\
+		if (W && !(in_atomic() || in_dbg_master())) msleep(W);	\
 	}								\
 	ret__;								\
 })
@@ -48,6 +47,13 @@
 #define wait_for(COND, MS) _wait_for(COND, MS, 1)
 #define wait_for_atomic(COND, MS) _wait_for(COND, MS, 0)
 
+#define MSLEEP(x) do { \
+	if (in_dbg_master()) \
+	       	mdelay(x); \
+	else \
+		msleep(x); \
+} while(0)
+
 #define KHz(x) (1000*x)
 #define MHz(x) KHz(1000*x)
 
@@ -104,7 +110,6 @@
 /* drm_display_mode->private_flags */
 #define INTEL_MODE_PIXEL_MULTIPLIER_SHIFT (0x0)
 #define INTEL_MODE_PIXEL_MULTIPLIER_MASK (0xf << INTEL_MODE_PIXEL_MULTIPLIER_SHIFT)
-#define INTEL_MODE_DP_FORCE_6BPC (0x10)
 
 static inline void
 intel_mode_set_pixel_multiplier(struct drm_display_mode *mode,
@@ -165,61 +170,17 @@ struct intel_crtc {
 	int16_t cursor_x, cursor_y;
 	int16_t cursor_width, cursor_height;
 	bool cursor_visible;
-	unsigned int bpp;
-
-	bool no_pll; /* tertiary pipe for IVB */
-	bool use_pll_a;
-};
-
-struct intel_plane {
-	struct drm_plane base;
-	enum pipe pipe;
-	struct drm_i915_gem_object *obj;
-	bool primary_disabled;
-	int max_downscale;
-	u32 lut_r[1024], lut_g[1024], lut_b[1024];
-	void (*update_plane)(struct drm_plane *plane,
-			     struct drm_framebuffer *fb,
-			     struct drm_i915_gem_object *obj,
-			     int crtc_x, int crtc_y,
-			     unsigned int crtc_w, unsigned int crtc_h,
-			     uint32_t x, uint32_t y,
-			     uint32_t src_w, uint32_t src_h);
-	void (*disable_plane)(struct drm_plane *plane);
-	int (*update_colorkey)(struct drm_plane *plane,
-			       struct drm_intel_sprite_colorkey *key);
-	void (*get_colorkey)(struct drm_plane *plane,
-			     struct drm_intel_sprite_colorkey *key);
 };
 
 #define to_intel_crtc(x) container_of(x, struct intel_crtc, base)
 #define to_intel_connector(x) container_of(x, struct intel_connector, base)
 #define to_intel_encoder(x) container_of(x, struct intel_encoder, base)
 #define to_intel_framebuffer(x) container_of(x, struct intel_framebuffer, base)
-#define to_intel_plane(x) container_of(x, struct intel_plane, base)
-
-#define DIP_HEADER_SIZE	5
 
 #define DIP_TYPE_AVI    0x82
 #define DIP_VERSION_AVI 0x2
 #define DIP_LEN_AVI     13
 
-#define DIP_TYPE_SPD	0x83
-#define DIP_VERSION_SPD	0x1
-#define DIP_LEN_SPD	25
-#define DIP_SPD_UNKNOWN	0
-#define DIP_SPD_DSTB	0x1
-#define DIP_SPD_DVDP	0x2
-#define DIP_SPD_DVHS	0x3
-#define DIP_SPD_HDDVR	0x4
-#define DIP_SPD_DVC	0x5
-#define DIP_SPD_DSC	0x6
-#define DIP_SPD_VCD	0x7
-#define DIP_SPD_GAME	0x8
-#define DIP_SPD_PC	0x9
-#define DIP_SPD_BD	0xa
-#define DIP_SPD_SCD	0xb
-
 struct dip_infoframe {
 	uint8_t type;		/* HB0 */
 	uint8_t ver;		/* HB1 */
@@ -272,13 +233,6 @@ struct intel_unpin_work {
 	bool enable_stall_check;
 };
 
-struct intel_fbc_work {
-	struct delayed_work work;
-	struct drm_crtc *crtc;
-	struct drm_framebuffer *fb;
-	int interval;
-};
-
 int intel_ddc_get_modes(struct drm_connector *c, struct i2c_adapter *adapter);
 extern bool intel_ddc_probe(struct intel_encoder *intel_encoder, int ddc_bus);
 
@@ -299,9 +253,8 @@ void
 intel_dp_set_m_n(struct drm_crtc *crtc, struct drm_display_mode *mode,
 		 struct drm_display_mode *adjusted_mode);
 extern bool intel_dpd_is_edp(struct drm_device *dev);
-extern void intel_edp_link_config(struct intel_encoder *, int *, int *);
+extern void intel_edp_link_config (struct intel_encoder *, int *, int *);
 extern bool intel_encoder_is_pch_edp(struct drm_encoder *encoder);
-extern int intel_plane_init(struct drm_device *dev, enum pipe pipe);
 
 /* intel_panel.c */
 extern void intel_fixed_panel_mode(struct drm_display_mode *fixed_mode,
@@ -313,15 +266,14 @@ extern void intel_pch_panel_fitting(struct drm_device *dev,
 extern u32 intel_panel_get_max_backlight(struct drm_device *dev);
 extern u32 intel_panel_get_backlight(struct drm_device *dev);
 extern void intel_panel_set_backlight(struct drm_device *dev, u32 level);
-extern int intel_panel_setup_backlight(struct drm_device *dev);
+extern void intel_panel_setup_backlight(struct drm_device *dev);
 extern void intel_panel_enable_backlight(struct drm_device *dev);
 extern void intel_panel_disable_backlight(struct drm_device *dev);
-extern void intel_panel_destroy_backlight(struct drm_device *dev);
 extern enum drm_connector_status intel_panel_detect(struct drm_device *dev);
 
 extern void intel_crtc_load_lut(struct drm_crtc *crtc);
-extern void intel_encoder_prepare(struct drm_encoder *encoder);
-extern void intel_encoder_commit(struct drm_encoder *encoder);
+extern void intel_encoder_prepare (struct drm_encoder *encoder);
+extern void intel_encoder_commit (struct drm_encoder *encoder);
 extern void intel_encoder_destroy(struct drm_encoder *encoder);
 
 static inline struct intel_encoder *intel_attached_encoder(struct drm_connector *connector)
@@ -353,6 +305,9 @@ extern void intel_release_load_detect_pipe(struct intel_encoder *intel_encoder,
 					   struct drm_connector *connector,
 					   struct intel_load_detect_pipe *old);
 
+extern struct drm_connector* intel_sdvo_find(struct drm_device *dev, int sdvoB);
+extern int intel_sdvo_supports_hotplug(struct drm_connector *connector);
+extern void intel_sdvo_set_hotplug(struct drm_connector *connector, int enable);
 extern void intelfb_restore(void);
 extern void intel_crtc_fb_gamma_set(struct drm_crtc *crtc, u16 red, u16 green,
 				    u16 blue, int regno);
@@ -362,18 +317,16 @@ extern void intel_enable_clock_gating(struct drm_device *dev);
 extern void ironlake_enable_drps(struct drm_device *dev);
 extern void ironlake_disable_drps(struct drm_device *dev);
 extern void gen6_enable_rps(struct drm_i915_private *dev_priv);
-extern void gen6_update_ring_freq(struct drm_i915_private *dev_priv);
 extern void gen6_disable_rps(struct drm_device *dev);
 extern void intel_init_emon(struct drm_device *dev);
 
 extern int intel_pin_and_fence_fb_obj(struct drm_device *dev,
 				      struct drm_i915_gem_object *obj,
 				      struct intel_ring_buffer *pipelined);
-extern void intel_unpin_fb_obj(struct drm_i915_gem_object *obj);
 
 extern int intel_framebuffer_init(struct drm_device *dev,
 				  struct intel_framebuffer *ifb,
-				  struct drm_mode_fb_cmd2 *mode_cmd,
+				  struct drm_mode_fb_cmd *mode_cmd,
 				  struct drm_i915_gem_object *obj);
 extern int intel_fbdev_init(struct drm_device *dev);
 extern void intel_fbdev_fini(struct drm_device *dev);
@@ -393,25 +346,5 @@ extern int intel_overlay_attrs(struct drm_device *dev, void *data,
 extern void intel_fb_output_poll_changed(struct drm_device *dev);
 extern void intel_fb_restore_mode(struct drm_device *dev);
 
-extern void assert_pipe(struct drm_i915_private *dev_priv, enum pipe pipe,
-			bool state);
-#define assert_pipe_enabled(d, p) assert_pipe(d, p, true)
-#define assert_pipe_disabled(d, p) assert_pipe(d, p, false)
-
 extern void intel_init_clock_gating(struct drm_device *dev);
-extern void intel_write_eld(struct drm_encoder *encoder,
-			    struct drm_display_mode *mode);
-extern void intel_cpt_verify_modeset(struct drm_device *dev, int pipe);
-
-/* For use by IVB LP watermark workaround in intel_sprite.c */
-extern void sandybridge_update_wm(struct drm_device *dev);
-extern void intel_update_sprite_watermarks(struct drm_device *dev, int pipe,
-					   uint32_t sprite_width,
-					   int pixel_size);
-
-extern int intel_sprite_set_colorkey(struct drm_device *dev, void *data,
-				     struct drm_file *file_priv);
-extern int intel_sprite_get_colorkey(struct drm_device *dev, void *data,
-				     struct drm_file *file_priv);
-
 #endif /* __INTEL_DRV_H__ */
diff --git a/drivers/gpu/drm/i915/intel_dvo.c b/drivers/gpu/drm/i915/intel_dvo.c
index 69bea56..8ac91b8 100644
--- a/drivers/gpu/drm/i915/intel_dvo.c
+++ b/drivers/gpu/drm/i915/intel_dvo.c
@@ -157,6 +157,7 @@ static bool intel_dvo_mode_fixup(struct drm_encoder *encoder,
 		C(vsync_end);
 		C(vtotal);
 		C(clock);
+		drm_mode_set_crtcinfo(adjusted_mode, CRTC_INTERLACE_HALVE_V);
 #undef C
 	}
 
diff --git a/drivers/gpu/drm/i915/intel_fb.c b/drivers/gpu/drm/i915/intel_fb.c
index 19ecd78..d0ce34b 100644
--- a/drivers/gpu/drm/i915/intel_fb.c
+++ b/drivers/gpu/drm/i915/intel_fb.c
@@ -65,7 +65,7 @@ static int intelfb_create(struct intel_fbdev *ifbdev,
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct fb_info *info;
 	struct drm_framebuffer *fb;
-	struct drm_mode_fb_cmd2 mode_cmd;
+	struct drm_mode_fb_cmd mode_cmd;
 	struct drm_i915_gem_object *obj;
 	struct device *device = &dev->pdev->dev;
 	int size, ret;
@@ -77,12 +77,11 @@ static int intelfb_create(struct intel_fbdev *ifbdev,
 	mode_cmd.width = sizes->surface_width;
 	mode_cmd.height = sizes->surface_height;
 
-	mode_cmd.pitches[0] = ALIGN(mode_cmd.width * ((sizes->surface_bpp + 7) /
-						      8), 64);
-	mode_cmd.pixel_format = drm_mode_legacy_fb_format(sizes->surface_bpp,
-							  sizes->surface_depth);
+	mode_cmd.bpp = sizes->surface_bpp;
+	mode_cmd.pitch = ALIGN(mode_cmd.width * ((mode_cmd.bpp + 7) / 8), 64);
+	mode_cmd.depth = sizes->surface_depth;
 
-	size = mode_cmd.pitches[0] * mode_cmd.height;
+	size = mode_cmd.pitch * mode_cmd.height;
 	size = ALIGN(size, PAGE_SIZE);
 	obj = i915_gem_alloc_object(dev, size);
 	if (!obj) {
@@ -149,10 +148,14 @@ static int intelfb_create(struct intel_fbdev *ifbdev,
 
 //	memset(info->screen_base, 0, size);
 
-	drm_fb_helper_fill_fix(info, fb->pitches[0], fb->depth);
+	drm_fb_helper_fill_fix(info, fb->pitch, fb->depth);
 	drm_fb_helper_fill_var(info, &ifbdev->helper, sizes->fb_width, sizes->fb_height);
 
-	/* Use default scratch pixmap (info->pixmap.flags = FB_PIXMAP_SYSTEM) */
+	info->pixmap.size = 64*1024;
+	info->pixmap.buf_align = 8;
+	info->pixmap.access_align = 32;
+	info->pixmap.flags = FB_PIXMAP_SYSTEM;
+	info->pixmap.scan_align = 1;
 
 	DRM_DEBUG_KMS("allocated %dx%d fb: 0x%08x, bo %p\n",
 		      fb->width, fb->height,
@@ -276,14 +279,8 @@ void intel_fb_restore_mode(struct drm_device *dev)
 {
 	int ret;
 	drm_i915_private_t *dev_priv = dev->dev_private;
-	struct drm_mode_config *config = &dev->mode_config;
-	struct drm_plane *plane;
 
 	ret = drm_fb_helper_restore_fbdev_mode(&dev_priv->fbdev->helper);
 	if (ret)
 		DRM_DEBUG("failed to restore crtc mode\n");
-
-	/* Be sure to shut off any planes that may be active */
-	list_for_each_entry(plane, &config->plane_list, head)
-		plane->funcs->disable_plane(plane);
 }
diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c
index 1239d79..918bac8 100644
--- a/drivers/gpu/drm/i915/intel_hdmi.c
+++ b/drivers/gpu/drm/i915/intel_hdmi.c
@@ -44,9 +44,7 @@ struct intel_hdmi {
 	uint32_t color_range;
 	bool has_hdmi_sink;
 	bool has_audio;
-	enum hdmi_force_audio force_audio;
-	void (*write_infoframe)(struct drm_encoder *encoder,
-				struct dip_infoframe *frame);
+	int force_audio;
 };
 
 static struct intel_hdmi *enc_to_intel_hdmi(struct drm_encoder *encoder)
@@ -60,69 +58,37 @@ static struct intel_hdmi *intel_attached_hdmi(struct drm_connector *connector)
 			    struct intel_hdmi, base);
 }
 
-void intel_dip_infoframe_csum(struct dip_infoframe *frame)
+void intel_dip_infoframe_csum(struct dip_infoframe *avi_if)
 {
-	uint8_t *data = (uint8_t *)frame;
+	uint8_t *data = (uint8_t *)avi_if;
 	uint8_t sum = 0;
 	unsigned i;
 
-	frame->checksum = 0;
-	frame->ecc = 0;
+	avi_if->checksum = 0;
+	avi_if->ecc = 0;
 
-	for (i = 0; i < frame->len + DIP_HEADER_SIZE; i++)
+	for (i = 0; i < sizeof(*avi_if); i++)
 		sum += data[i];
 
-	frame->checksum = 0x100 - sum;
+	avi_if->checksum = 0x100 - sum;
 }
 
-static u32 intel_infoframe_index(struct dip_infoframe *frame)
-{
-	u32 flags = 0;
-
-	switch (frame->type) {
-	case DIP_TYPE_AVI:
-		flags |= VIDEO_DIP_SELECT_AVI;
-		break;
-	case DIP_TYPE_SPD:
-		flags |= VIDEO_DIP_SELECT_SPD;
-		break;
-	default:
-		DRM_DEBUG_DRIVER("unknown info frame type %d\n", frame->type);
-		break;
-	}
-
-	return flags;
-}
-
-static u32 intel_infoframe_flags(struct dip_infoframe *frame)
-{
-	u32 flags = 0;
-
-	switch (frame->type) {
-	case DIP_TYPE_AVI:
-		flags |= VIDEO_DIP_ENABLE_AVI | VIDEO_DIP_FREQ_VSYNC;
-		break;
-	case DIP_TYPE_SPD:
-		flags |= VIDEO_DIP_ENABLE_SPD | VIDEO_DIP_FREQ_VSYNC;
-		break;
-	default:
-		DRM_DEBUG_DRIVER("unknown info frame type %d\n", frame->type);
-		break;
-	}
-
-	return flags;
-}
-
-static void i9xx_write_infoframe(struct drm_encoder *encoder,
-				 struct dip_infoframe *frame)
+static void intel_hdmi_set_avi_infoframe(struct drm_encoder *encoder)
 {
-	uint32_t *data = (uint32_t *)frame;
+	struct dip_infoframe avi_if = {
+		.type = DIP_TYPE_AVI,
+		.ver = DIP_VERSION_AVI,
+		.len = DIP_LEN_AVI,
+	};
+	uint32_t *data = (uint32_t *)&avi_if;
 	struct drm_device *dev = encoder->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
-	u32 port, flags, val = I915_READ(VIDEO_DIP_CTL);
-	unsigned i, len = DIP_HEADER_SIZE + frame->len;
+	u32 port;
+	unsigned i;
 
+	if (!intel_hdmi->has_hdmi_sink)
+		return;
 
 	/* XXX first guess at handling video port, is this corrent? */
 	if (intel_hdmi->sdvox_reg == SDVOB)
@@ -132,87 +98,18 @@ static void i9xx_write_infoframe(struct drm_encoder *encoder,
 	else
 		return;
 
-	flags = intel_infoframe_index(frame);
-
-	val &= ~VIDEO_DIP_SELECT_MASK;
+	I915_WRITE(VIDEO_DIP_CTL, VIDEO_DIP_ENABLE | port |
+		   VIDEO_DIP_SELECT_AVI | VIDEO_DIP_FREQ_VSYNC);
 
-	I915_WRITE(VIDEO_DIP_CTL, val | port | flags);
-
-	for (i = 0; i < len; i += 4) {
+	intel_dip_infoframe_csum(&avi_if);
+	for (i = 0; i < sizeof(avi_if); i += 4) {
 		I915_WRITE(VIDEO_DIP_DATA, *data);
 		data++;
 	}
 
-	flags |= intel_infoframe_flags(frame);
-
-	I915_WRITE(VIDEO_DIP_CTL, VIDEO_DIP_ENABLE | val | port | flags);
-}
-
-static void ironlake_write_infoframe(struct drm_encoder *encoder,
-				     struct dip_infoframe *frame)
-{
-	uint32_t *data = (uint32_t *)frame;
-	struct drm_device *dev = encoder->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_crtc *crtc = encoder->crtc;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int reg = TVIDEO_DIP_CTL(intel_crtc->pipe);
-	unsigned i, len = DIP_HEADER_SIZE + frame->len;
-	u32 flags, val = I915_READ(reg);
-
-	intel_wait_for_vblank(dev, intel_crtc->pipe);
-
-	flags = intel_infoframe_index(frame);
-
-	val &= ~(VIDEO_DIP_SELECT_MASK | 0xf); /* clear DIP data offset */
-
-	I915_WRITE(reg, VIDEO_DIP_ENABLE | val | flags);
-
-	for (i = 0; i < len; i += 4) {
-		I915_WRITE(TVIDEO_DIP_DATA(intel_crtc->pipe), *data);
-		data++;
-	}
-
-	flags |= intel_infoframe_flags(frame);
-
-	I915_WRITE(reg, VIDEO_DIP_ENABLE | val | flags);
-}
-static void intel_set_infoframe(struct drm_encoder *encoder,
-				struct dip_infoframe *frame)
-{
-	struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder);
-
-	if (!intel_hdmi->has_hdmi_sink)
-		return;
-
-	intel_dip_infoframe_csum(frame);
-	intel_hdmi->write_infoframe(encoder, frame);
-}
-
-static void intel_hdmi_set_avi_infoframe(struct drm_encoder *encoder)
-{
-	struct dip_infoframe avi_if = {
-		.type = DIP_TYPE_AVI,
-		.ver = DIP_VERSION_AVI,
-		.len = DIP_LEN_AVI,
-	};
-
-	intel_set_infoframe(encoder, &avi_if);
-}
-
-static void intel_hdmi_set_spd_infoframe(struct drm_encoder *encoder)
-{
-	struct dip_infoframe spd_if;
-
-	memset(&spd_if, 0, sizeof(spd_if));
-	spd_if.type = DIP_TYPE_SPD;
-	spd_if.ver = DIP_VERSION_SPD;
-	spd_if.len = DIP_LEN_SPD;
-	strcpy(spd_if.body.spd.vn, "Intel");
-	strcpy(spd_if.body.spd.pd, "Integrated gfx");
-	spd_if.body.spd.sdi = DIP_SPD_PC;
-
-	intel_set_infoframe(encoder, &spd_if);
+	I915_WRITE(VIDEO_DIP_CTL, VIDEO_DIP_ENABLE | port |
+		   VIDEO_DIP_SELECT_AVI | VIDEO_DIP_FREQ_VSYNC |
+		   VIDEO_DIP_ENABLE_AVI);
 }
 
 static void intel_hdmi_mode_set(struct drm_encoder *encoder,
@@ -227,40 +124,32 @@ static void intel_hdmi_mode_set(struct drm_encoder *encoder,
 	u32 sdvox;
 
 	sdvox = SDVO_ENCODING_HDMI | SDVO_BORDER_ENABLE;
-	if (!HAS_PCH_SPLIT(dev))
-		sdvox |= intel_hdmi->color_range;
+	sdvox |= intel_hdmi->color_range;
 	if (adjusted_mode->flags & DRM_MODE_FLAG_PVSYNC)
 		sdvox |= SDVO_VSYNC_ACTIVE_HIGH;
 	if (adjusted_mode->flags & DRM_MODE_FLAG_PHSYNC)
 		sdvox |= SDVO_HSYNC_ACTIVE_HIGH;
 
-	if (intel_crtc->bpp > 24)
-		sdvox |= COLOR_FORMAT_12bpc;
-	else
-		sdvox |= COLOR_FORMAT_8bpc;
-
 	/* Required on CPT */
 	if (intel_hdmi->has_hdmi_sink && HAS_PCH_CPT(dev))
 		sdvox |= HDMI_MODE_SELECT;
 
 	if (intel_hdmi->has_audio) {
-		DRM_DEBUG_DRIVER("Enabling HDMI audio on pipe %c\n",
-				 pipe_name(intel_crtc->pipe));
 		sdvox |= SDVO_AUDIO_ENABLE;
 		sdvox |= SDVO_NULL_PACKETS_DURING_VSYNC;
-		intel_write_eld(encoder, adjusted_mode);
 	}
 
-	if (HAS_PCH_CPT(dev))
-		sdvox |= PORT_TRANS_SEL_CPT(intel_crtc->pipe);
-	else if (intel_crtc->pipe == 1)
-		sdvox |= SDVO_PIPE_B_SELECT;
+	if (intel_crtc->pipe == 1) {
+		if (HAS_PCH_CPT(dev))
+			sdvox |= PORT_TRANS_B_SEL_CPT;
+		else
+			sdvox |= SDVO_PIPE_B_SELECT;
+	}
 
 	I915_WRITE(intel_hdmi->sdvox_reg, sdvox);
 	POSTING_READ(intel_hdmi->sdvox_reg);
 
 	intel_hdmi_set_avi_infoframe(encoder);
-	intel_hdmi_set_spd_infoframe(encoder);
 }
 
 static void intel_hdmi_dpms(struct drm_encoder *encoder, int mode)
@@ -339,9 +228,7 @@ intel_hdmi_detect(struct drm_connector *connector, bool force)
 	if (edid) {
 		if (edid->input & DRM_EDID_INPUT_DIGITAL) {
 			status = connector_status_connected;
-			if (intel_hdmi->force_audio != HDMI_AUDIO_OFF_DVI)
-				intel_hdmi->has_hdmi_sink =
-						drm_detect_hdmi_monitor(edid);
+			intel_hdmi->has_hdmi_sink = drm_detect_hdmi_monitor(edid);
 			intel_hdmi->has_audio = drm_detect_monitor_audio(edid);
 		}
 		connector->display_info.raw_edid = NULL;
@@ -349,9 +236,8 @@ intel_hdmi_detect(struct drm_connector *connector, bool force)
 	}
 
 	if (status == connector_status_connected) {
-		if (intel_hdmi->force_audio != HDMI_AUDIO_AUTO)
-			intel_hdmi->has_audio =
-				(intel_hdmi->force_audio == HDMI_AUDIO_ON);
+		if (intel_hdmi->force_audio)
+			intel_hdmi->has_audio = intel_hdmi->force_audio > 0;
 	}
 
 	return status;
@@ -405,7 +291,7 @@ intel_hdmi_set_property(struct drm_connector *connector,
 		return ret;
 
 	if (property == dev_priv->force_audio_property) {
-		enum hdmi_force_audio i = val;
+		int i = val;
 		bool has_audio;
 
 		if (i == intel_hdmi->force_audio)
@@ -413,13 +299,13 @@ intel_hdmi_set_property(struct drm_connector *connector,
 
 		intel_hdmi->force_audio = i;
 
-		if (i == HDMI_AUDIO_AUTO)
+		if (i == 0)
 			has_audio = intel_hdmi_detect_audio(connector);
 		else
-			has_audio = (i == HDMI_AUDIO_ON);
+			has_audio = i > 0;
 
-		if (i == HDMI_AUDIO_OFF_DVI)
-			intel_hdmi->has_hdmi_sink = 0;
+		if (has_audio == intel_hdmi->has_audio)
+			return 0;
 
 		intel_hdmi->has_audio = has_audio;
 		goto done;
@@ -493,7 +379,6 @@ void intel_hdmi_init(struct drm_device *dev, int sdvox_reg)
 	struct intel_encoder *intel_encoder;
 	struct intel_connector *intel_connector;
 	struct intel_hdmi *intel_hdmi;
-	int i;
 
 	intel_hdmi = kzalloc(sizeof(struct intel_hdmi), GFP_KERNEL);
 	if (!intel_hdmi)
@@ -517,9 +402,9 @@ void intel_hdmi_init(struct drm_device *dev, int sdvox_reg)
 	intel_encoder->type = INTEL_OUTPUT_HDMI;
 
 	connector->polled = DRM_CONNECTOR_POLL_HPD;
-	connector->interlace_allowed = 1;
+	connector->interlace_allowed = 0;
 	connector->doublescan_allowed = 0;
-	intel_encoder->crtc_mask = (1 << 0) | (1 << 1) | (1 << 2);
+	intel_encoder->crtc_mask = (1 << 0) | (1 << 1);
 
 	/* Set up the DDC bus. */
 	if (sdvox_reg == SDVOB) {
@@ -546,15 +431,6 @@ void intel_hdmi_init(struct drm_device *dev, int sdvox_reg)
 
 	intel_hdmi->sdvox_reg = sdvox_reg;
 
-	if (!HAS_PCH_SPLIT(dev)) {
-		intel_hdmi->write_infoframe = i9xx_write_infoframe;
-		I915_WRITE(VIDEO_DIP_CTL, 0);
-	} else {
-		intel_hdmi->write_infoframe = ironlake_write_infoframe;
-		for_each_pipe(i)
-			I915_WRITE(TVIDEO_DIP_CTL(i), 0);
-	}
-
 	drm_encoder_helper_add(&intel_encoder->base, &intel_hdmi_helper_funcs);
 
 	intel_hdmi_add_properties(intel_hdmi, connector);
diff --git a/drivers/gpu/drm/i915/intel_i2c.c b/drivers/gpu/drm/i915/intel_i2c.c
index 4f40b2b..d98cee6 100644
--- a/drivers/gpu/drm/i915/intel_i2c.c
+++ b/drivers/gpu/drm/i915/intel_i2c.c
@@ -36,7 +36,7 @@
 
 /* Intel GPIO access functions */
 
-#define I2C_RISEFALL_TIME 10
+#define I2C_RISEFALL_TIME 20
 
 static inline struct intel_gmbus *
 to_intel_gmbus(struct i2c_adapter *i2c)
@@ -44,6 +44,13 @@ to_intel_gmbus(struct i2c_adapter *i2c)
 	return container_of(i2c, struct intel_gmbus, adapter);
 }
 
+struct intel_gpio {
+	struct i2c_adapter adapter;
+	struct i2c_algo_bit_data algo;
+	struct drm_i915_private *dev_priv;
+	u32 reg;
+};
+
 void
 intel_i2c_reset(struct drm_device *dev)
 {
@@ -70,15 +77,15 @@ static void intel_i2c_quirk_set(struct drm_i915_private *dev_priv, bool enable)
 	I915_WRITE(DSPCLK_GATE_D, val);
 }
 
-static u32 get_reserved(struct intel_gmbus *bus)
+static u32 get_reserved(struct intel_gpio *gpio)
 {
-	struct drm_i915_private *dev_priv = bus->dev_priv;
+	struct drm_i915_private *dev_priv = gpio->dev_priv;
 	struct drm_device *dev = dev_priv->dev;
 	u32 reserved = 0;
 
 	/* On most chips, these bits must be preserved in software. */
 	if (!IS_I830(dev) && !IS_845G(dev))
-		reserved = I915_READ_NOTRACE(bus->gpio_reg) &
+		reserved = I915_READ_NOTRACE(gpio->reg) &
 					     (GPIO_DATA_PULLUP_DISABLE |
 					      GPIO_CLOCK_PULLUP_DISABLE);
 
@@ -87,29 +94,29 @@ static u32 get_reserved(struct intel_gmbus *bus)
 
 static int get_clock(void *data)
 {
-	struct intel_gmbus *bus = data;
-	struct drm_i915_private *dev_priv = bus->dev_priv;
-	u32 reserved = get_reserved(bus);
-	I915_WRITE_NOTRACE(bus->gpio_reg, reserved | GPIO_CLOCK_DIR_MASK);
-	I915_WRITE_NOTRACE(bus->gpio_reg, reserved);
-	return (I915_READ_NOTRACE(bus->gpio_reg) & GPIO_CLOCK_VAL_IN) != 0;
+	struct intel_gpio *gpio = data;
+	struct drm_i915_private *dev_priv = gpio->dev_priv;
+	u32 reserved = get_reserved(gpio);
+	I915_WRITE_NOTRACE(gpio->reg, reserved | GPIO_CLOCK_DIR_MASK);
+	I915_WRITE_NOTRACE(gpio->reg, reserved);
+	return (I915_READ_NOTRACE(gpio->reg) & GPIO_CLOCK_VAL_IN) != 0;
 }
 
 static int get_data(void *data)
 {
-	struct intel_gmbus *bus = data;
-	struct drm_i915_private *dev_priv = bus->dev_priv;
-	u32 reserved = get_reserved(bus);
-	I915_WRITE_NOTRACE(bus->gpio_reg, reserved | GPIO_DATA_DIR_MASK);
-	I915_WRITE_NOTRACE(bus->gpio_reg, reserved);
-	return (I915_READ_NOTRACE(bus->gpio_reg) & GPIO_DATA_VAL_IN) != 0;
+	struct intel_gpio *gpio = data;
+	struct drm_i915_private *dev_priv = gpio->dev_priv;
+	u32 reserved = get_reserved(gpio);
+	I915_WRITE_NOTRACE(gpio->reg, reserved | GPIO_DATA_DIR_MASK);
+	I915_WRITE_NOTRACE(gpio->reg, reserved);
+	return (I915_READ_NOTRACE(gpio->reg) & GPIO_DATA_VAL_IN) != 0;
 }
 
 static void set_clock(void *data, int state_high)
 {
-	struct intel_gmbus *bus = data;
-	struct drm_i915_private *dev_priv = bus->dev_priv;
-	u32 reserved = get_reserved(bus);
+	struct intel_gpio *gpio = data;
+	struct drm_i915_private *dev_priv = gpio->dev_priv;
+	u32 reserved = get_reserved(gpio);
 	u32 clock_bits;
 
 	if (state_high)
@@ -118,15 +125,15 @@ static void set_clock(void *data, int state_high)
 		clock_bits = GPIO_CLOCK_DIR_OUT | GPIO_CLOCK_DIR_MASK |
 			GPIO_CLOCK_VAL_MASK;
 
-	I915_WRITE_NOTRACE(bus->gpio_reg, reserved | clock_bits);
-	POSTING_READ(bus->gpio_reg);
+	I915_WRITE_NOTRACE(gpio->reg, reserved | clock_bits);
+	POSTING_READ(gpio->reg);
 }
 
 static void set_data(void *data, int state_high)
 {
-	struct intel_gmbus *bus = data;
-	struct drm_i915_private *dev_priv = bus->dev_priv;
-	u32 reserved = get_reserved(bus);
+	struct intel_gpio *gpio = data;
+	struct drm_i915_private *dev_priv = gpio->dev_priv;
+	u32 reserved = get_reserved(gpio);
 	u32 data_bits;
 
 	if (state_high)
@@ -135,14 +142,13 @@ static void set_data(void *data, int state_high)
 		data_bits = GPIO_DATA_DIR_OUT | GPIO_DATA_DIR_MASK |
 			GPIO_DATA_VAL_MASK;
 
-	I915_WRITE_NOTRACE(bus->gpio_reg, reserved | data_bits);
-	POSTING_READ(bus->gpio_reg);
+	I915_WRITE_NOTRACE(gpio->reg, reserved | data_bits);
+	POSTING_READ(gpio->reg);
 }
 
-static bool
-intel_gpio_setup(struct intel_gmbus *bus, u32 pin)
+static struct i2c_adapter *
+intel_gpio_create(struct drm_i915_private *dev_priv, u32 pin)
 {
-	struct drm_i915_private *dev_priv = bus->dev_priv;
 	static const int map_pin_to_reg[] = {
 		0,
 		GPIOB,
@@ -153,48 +159,65 @@ intel_gpio_setup(struct intel_gmbus *bus, u32 pin)
 		0,
 		GPIOF,
 	};
-	struct i2c_algo_bit_data *algo;
+	struct intel_gpio *gpio;
 
 	if (pin >= ARRAY_SIZE(map_pin_to_reg) || !map_pin_to_reg[pin])
-		return false;
+		return NULL;
 
-	algo = &bus->bit_algo;
+	gpio = kzalloc(sizeof(struct intel_gpio), GFP_KERNEL);
+	if (gpio == NULL)
+		return NULL;
 
-	bus->gpio_reg = map_pin_to_reg[pin];
+	gpio->reg = map_pin_to_reg[pin];
 	if (HAS_PCH_SPLIT(dev_priv->dev))
-		bus->gpio_reg += PCH_GPIOA - GPIOA;
-
-	bus->adapter.algo_data = algo;
-	algo->setsda = set_data;
-	algo->setscl = set_clock;
-	algo->getsda = get_data;
-	algo->getscl = get_clock;
-	algo->udelay = I2C_RISEFALL_TIME;
-	algo->timeout = usecs_to_jiffies(2200);
-	algo->data = bus;
-
-	return true;
+		gpio->reg += PCH_GPIOA - GPIOA;
+	gpio->dev_priv = dev_priv;
+
+	snprintf(gpio->adapter.name, sizeof(gpio->adapter.name),
+		 "i915 GPIO%c", "?BACDE?F"[pin]);
+	gpio->adapter.owner = THIS_MODULE;
+	gpio->adapter.algo_data	= &gpio->algo;
+	gpio->adapter.dev.parent = &dev_priv->dev->pdev->dev;
+	gpio->algo.setsda = set_data;
+	gpio->algo.setscl = set_clock;
+	gpio->algo.getsda = get_data;
+	gpio->algo.getscl = get_clock;
+	gpio->algo.udelay = I2C_RISEFALL_TIME;
+	gpio->algo.timeout = usecs_to_jiffies(2200);
+	gpio->algo.data = gpio;
+
+	if (i2c_bit_add_bus(&gpio->adapter))
+		goto out_free;
+
+	return &gpio->adapter;
+
+out_free:
+	kfree(gpio);
+	return NULL;
 }
 
 static int
-intel_i2c_quirk_xfer(struct intel_gmbus *bus,
+intel_i2c_quirk_xfer(struct drm_i915_private *dev_priv,
+		     struct i2c_adapter *adapter,
 		     struct i2c_msg *msgs,
 		     int num)
 {
-	struct drm_i915_private *dev_priv = bus->dev_priv;
+	struct intel_gpio *gpio = container_of(adapter,
+					       struct intel_gpio,
+					       adapter);
 	int ret;
 
 	intel_i2c_reset(dev_priv->dev);
 
 	intel_i2c_quirk_set(dev_priv, true);
-	set_data(bus, 1);
-	set_clock(bus, 1);
+	set_data(gpio, 1);
+	set_clock(gpio, 1);
 	udelay(I2C_RISEFALL_TIME);
 
-	ret = i2c_bit_algo.master_xfer(&bus->adapter, msgs, num);
+	ret = adapter->algo->master_xfer(adapter, msgs, num);
 
-	set_data(bus, 1);
-	set_clock(bus, 1);
+	set_data(gpio, 1);
+	set_clock(gpio, 1);
 	intel_i2c_quirk_set(dev_priv, false);
 
 	return ret;
@@ -208,15 +231,12 @@ gmbus_xfer(struct i2c_adapter *adapter,
 	struct intel_gmbus *bus = container_of(adapter,
 					       struct intel_gmbus,
 					       adapter);
-	struct drm_i915_private *dev_priv = bus->dev_priv;
-	int i, reg_offset, ret;
+	struct drm_i915_private *dev_priv = adapter->algo_data;
+	int i, reg_offset;
 
-	mutex_lock(&dev_priv->gmbus_mutex);
-
-	if (bus->force_bit) {
-		ret = intel_i2c_quirk_xfer(bus, msgs, num);
-		goto out;
-	}
+	if (bus->force_bit)
+		return intel_i2c_quirk_xfer(dev_priv,
+					    bus->force_bit, msgs, num);
 
 	reg_offset = HAS_PCH_SPLIT(dev_priv->dev) ? PCH_GMBUS0 - GMBUS0 : 0;
 
@@ -228,8 +248,7 @@ gmbus_xfer(struct i2c_adapter *adapter,
 
 		if (msgs[i].flags & I2C_M_RD) {
 			I915_WRITE(GMBUS1 + reg_offset,
-				   GMBUS_CYCLE_WAIT |
-				   (i + 1 == num ? GMBUS_CYCLE_STOP : 0) |
+				   GMBUS_CYCLE_WAIT | (i + 1 == num ? GMBUS_CYCLE_STOP : 0) |
 				   (len << GMBUS_BYTE_COUNT_SHIFT) |
 				   (msgs[i].addr << GMBUS_SLAVE_ADDR_SHIFT) |
 				   GMBUS_SLAVE_READ | GMBUS_SW_RDY);
@@ -258,8 +277,7 @@ gmbus_xfer(struct i2c_adapter *adapter,
 
 			I915_WRITE(GMBUS3 + reg_offset, val);
 			I915_WRITE(GMBUS1 + reg_offset,
-				   GMBUS_CYCLE_WAIT |
-				   (i + 1 == num ? GMBUS_CYCLE_STOP : 0) |
+				   (i + 1 == num ? GMBUS_CYCLE_STOP : GMBUS_CYCLE_WAIT) |
 				   (msgs[i].len << GMBUS_BYTE_COUNT_SHIFT) |
 				   (msgs[i].addr << GMBUS_SLAVE_ADDR_SHIFT) |
 				   GMBUS_SLAVE_WRITE | GMBUS_SW_RDY);
@@ -298,15 +316,11 @@ clear_err:
 	I915_WRITE(GMBUS1 + reg_offset, 0);
 
 done:
-	/* Mark the GMBUS interface as disabled after waiting for idle.
-	 * We will re-enable it at the start of the next xfer,
-	 * till then let it sleep.
+	/* Mark the GMBUS interface as disabled. We will re-enable it at the
+	 * start of the next xfer, till then let it sleep.
 	 */
-	if (wait_for((I915_READ(GMBUS2 + reg_offset) & GMBUS_ACTIVE) == 0, 10))
-		DRM_INFO("GMBUS timed out waiting for idle\n");
 	I915_WRITE(GMBUS0 + reg_offset, 0);
-	ret = i;
-	goto out;
+	return i;
 
 timeout:
 	DRM_INFO("GMBUS timed out, falling back to bit banging on pin %d [%s]\n",
@@ -314,21 +328,23 @@ timeout:
 	I915_WRITE(GMBUS0 + reg_offset, 0);
 
 	/* Hardware may not support GMBUS over these pins? Try GPIO bitbanging instead. */
-	if (!bus->has_gpio) {
-		ret = -EIO;
-	} else {
-		bus->force_bit = true;
-		ret = intel_i2c_quirk_xfer(bus, msgs, num);
-	}
-out:
-	mutex_unlock(&dev_priv->gmbus_mutex);
-	return ret;
+	bus->force_bit = intel_gpio_create(dev_priv, bus->reg0 & 0xff);
+	if (!bus->force_bit)
+		return -ENOMEM;
+
+	return intel_i2c_quirk_xfer(dev_priv, bus->force_bit, msgs, num);
 }
 
 static u32 gmbus_func(struct i2c_adapter *adapter)
 {
-	return i2c_bit_algo.functionality(adapter) &
-		(I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL |
+	struct intel_gmbus *bus = container_of(adapter,
+					       struct intel_gmbus,
+					       adapter);
+
+	if (bus->force_bit)
+		bus->force_bit->algo->functionality(bus->force_bit);
+
+	return (I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL |
 		/* I2C_FUNC_10BIT_ADDR | */
 		I2C_FUNC_SMBUS_READ_BLOCK_DATA |
 		I2C_FUNC_SMBUS_BLOCK_PROC_CALL);
@@ -358,13 +374,11 @@ int intel_setup_gmbus(struct drm_device *dev)
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	int ret, i;
 
-	dev_priv->gmbus = kcalloc(GMBUS_NUM_PORTS, sizeof(struct intel_gmbus),
+	dev_priv->gmbus = kcalloc(sizeof(struct intel_gmbus), GMBUS_NUM_PORTS,
 				  GFP_KERNEL);
 	if (dev_priv->gmbus == NULL)
 		return -ENOMEM;
 
-	mutex_init(&dev_priv->gmbus_mutex);
-
 	for (i = 0; i < GMBUS_NUM_PORTS; i++) {
 		struct intel_gmbus *bus = &dev_priv->gmbus[i];
 
@@ -376,7 +390,7 @@ int intel_setup_gmbus(struct drm_device *dev)
 			 names[i]);
 
 		bus->adapter.dev.parent = &dev->pdev->dev;
-		bus->dev_priv = dev_priv;
+		bus->adapter.algo_data	= dev_priv;
 
 		bus->adapter.algo = &gmbus_algorithm;
 		ret = i2c_add_adapter(&bus->adapter);
@@ -386,11 +400,8 @@ int intel_setup_gmbus(struct drm_device *dev)
 		/* By default use a conservative clock rate */
 		bus->reg0 = i | GMBUS_RATE_100KHZ;
 
-		bus->has_gpio = intel_gpio_setup(bus, i);
-
 		/* XXX force bit banging until GMBUS is fully debugged */
-		if (bus->has_gpio && IS_GEN2(dev))
-			bus->force_bit = true;
+		bus->force_bit = intel_gpio_create(dev_priv, i);
 	}
 
 	intel_i2c_reset(dev_priv->dev);
@@ -411,15 +422,32 @@ void intel_gmbus_set_speed(struct i2c_adapter *adapter, int speed)
 {
 	struct intel_gmbus *bus = to_intel_gmbus(adapter);
 
-	bus->reg0 = (bus->reg0 & ~(0x3 << 8)) | speed;
+	/* speed:
+	 * 0x0 = 100 KHz
+	 * 0x1 = 50 KHz
+	 * 0x2 = 400 KHz
+	 * 0x3 = 1000 Khz
+	 */
+	bus->reg0 = (bus->reg0 & ~(0x3 << 8)) | (speed << 8);
 }
 
 void intel_gmbus_force_bit(struct i2c_adapter *adapter, bool force_bit)
 {
 	struct intel_gmbus *bus = to_intel_gmbus(adapter);
 
-	if (bus->has_gpio)
-		bus->force_bit = force_bit;
+	if (force_bit) {
+		if (bus->force_bit == NULL) {
+			struct drm_i915_private *dev_priv = adapter->algo_data;
+			bus->force_bit = intel_gpio_create(dev_priv,
+							   bus->reg0 & 0xff);
+		}
+	} else {
+		if (bus->force_bit) {
+			i2c_del_adapter(bus->force_bit);
+			kfree(bus->force_bit);
+			bus->force_bit = NULL;
+		}
+	}
 }
 
 void intel_teardown_gmbus(struct drm_device *dev)
@@ -432,6 +460,10 @@ void intel_teardown_gmbus(struct drm_device *dev)
 
 	for (i = 0; i < GMBUS_NUM_PORTS; i++) {
 		struct intel_gmbus *bus = &dev_priv->gmbus[i];
+		if (bus->force_bit) {
+			i2c_del_adapter(bus->force_bit);
+			kfree(bus->force_bit);
+		}
 		i2c_del_adapter(&bus->adapter);
 	}
 
diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c
index b19a5cf..09881ac 100644
--- a/drivers/gpu/drm/i915/intel_lvds.c
+++ b/drivers/gpu/drm/i915/intel_lvds.c
@@ -72,16 +72,14 @@ static void intel_lvds_enable(struct intel_lvds *intel_lvds)
 {
 	struct drm_device *dev = intel_lvds->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 ctl_reg, lvds_reg, stat_reg;
+	u32 ctl_reg, lvds_reg;
 
 	if (HAS_PCH_SPLIT(dev)) {
 		ctl_reg = PCH_PP_CONTROL;
 		lvds_reg = PCH_LVDS;
-		stat_reg = PCH_PP_STATUS;
 	} else {
 		ctl_reg = PP_CONTROL;
 		lvds_reg = LVDS;
-		stat_reg = PP_STATUS;
 	}
 
 	I915_WRITE(lvds_reg, I915_READ(lvds_reg) | LVDS_PORT_EN);
@@ -96,16 +94,17 @@ static void intel_lvds_enable(struct intel_lvds *intel_lvds)
 		DRM_DEBUG_KMS("applying panel-fitter: %x, %x\n",
 			      intel_lvds->pfit_control,
 			      intel_lvds->pfit_pgm_ratios);
-
-		I915_WRITE(PFIT_PGM_RATIOS, intel_lvds->pfit_pgm_ratios);
-		I915_WRITE(PFIT_CONTROL, intel_lvds->pfit_control);
-		intel_lvds->pfit_dirty = false;
+		if (wait_for((I915_READ(PP_STATUS) & PP_ON) == 0, 1000)) {
+			DRM_ERROR("timed out waiting for panel to power off\n");
+		} else {
+			I915_WRITE(PFIT_PGM_RATIOS, intel_lvds->pfit_pgm_ratios);
+			I915_WRITE(PFIT_CONTROL, intel_lvds->pfit_control);
+			intel_lvds->pfit_dirty = false;
+		}
 	}
 
 	I915_WRITE(ctl_reg, I915_READ(ctl_reg) | POWER_TARGET_ON);
 	POSTING_READ(lvds_reg);
-	if (wait_for((I915_READ(stat_reg) & PP_ON) != 0, 1000))
-		DRM_ERROR("timed out waiting for panel to power on\n");
 
 	intel_panel_enable_backlight(dev);
 }
@@ -114,25 +113,24 @@ static void intel_lvds_disable(struct intel_lvds *intel_lvds)
 {
 	struct drm_device *dev = intel_lvds->base.base.dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
-	u32 ctl_reg, lvds_reg, stat_reg;
+	u32 ctl_reg, lvds_reg;
 
 	if (HAS_PCH_SPLIT(dev)) {
 		ctl_reg = PCH_PP_CONTROL;
 		lvds_reg = PCH_LVDS;
-		stat_reg = PCH_PP_STATUS;
 	} else {
 		ctl_reg = PP_CONTROL;
 		lvds_reg = LVDS;
-		stat_reg = PP_STATUS;
 	}
 
 	intel_panel_disable_backlight(dev);
 
 	I915_WRITE(ctl_reg, I915_READ(ctl_reg) & ~POWER_TARGET_ON);
-	if (wait_for((I915_READ(stat_reg) & PP_ON) == 0, 1000))
-		DRM_ERROR("timed out waiting for panel to power off\n");
 
 	if (intel_lvds->pfit_control) {
+		if (wait_for((I915_READ(PP_STATUS) & PP_ON) == 0, 1000))
+			DRM_ERROR("timed out waiting for panel to power off\n");
+
 		I915_WRITE(PFIT_CONTROL, 0);
 		intel_lvds->pfit_dirty = true;
 	}
@@ -400,21 +398,53 @@ out:
 
 static void intel_lvds_prepare(struct drm_encoder *encoder)
 {
+	struct drm_device *dev = encoder->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_lvds *intel_lvds = to_intel_lvds(encoder);
 
-	/*
+	/* We try to do the minimum that is necessary in order to unlock
+	 * the registers for mode setting.
+	 *
+	 * On Ironlake, this is quite simple as we just set the unlock key
+	 * and ignore all subtleties. (This may cause some issues...)
+	 *
 	 * Prior to Ironlake, we must disable the pipe if we want to adjust
 	 * the panel fitter. However at all other times we can just reset
 	 * the registers regardless.
 	 */
-	if (!HAS_PCH_SPLIT(encoder->dev) && intel_lvds->pfit_dirty)
-		intel_lvds_disable(intel_lvds);
+
+	if (HAS_PCH_SPLIT(dev)) {
+		I915_WRITE(PCH_PP_CONTROL,
+			   I915_READ(PCH_PP_CONTROL) | PANEL_UNLOCK_REGS);
+	} else if (intel_lvds->pfit_dirty) {
+		I915_WRITE(PP_CONTROL,
+			   (I915_READ(PP_CONTROL) | PANEL_UNLOCK_REGS)
+			   & ~POWER_TARGET_ON);
+	} else {
+		I915_WRITE(PP_CONTROL,
+			   I915_READ(PP_CONTROL) | PANEL_UNLOCK_REGS);
+	}
 }
 
 static void intel_lvds_commit(struct drm_encoder *encoder)
 {
+	struct drm_device *dev = encoder->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct intel_lvds *intel_lvds = to_intel_lvds(encoder);
 
+	/* Undo any unlocking done in prepare to prevent accidental
+	 * adjustment of the registers.
+	 */
+	if (HAS_PCH_SPLIT(dev)) {
+		u32 val = I915_READ(PCH_PP_CONTROL);
+		if ((val & PANEL_UNLOCK_REGS) == PANEL_UNLOCK_REGS)
+			I915_WRITE(PCH_PP_CONTROL, val & 0x3);
+	} else {
+		u32 val = I915_READ(PP_CONTROL);
+		if ((val & PANEL_UNLOCK_REGS) == PANEL_UNLOCK_REGS)
+			I915_WRITE(PP_CONTROL, val & 0x3);
+	}
+
 	/* Always do a full power on as we do not know what state
 	 * we were left in.
 	 */
@@ -552,8 +582,6 @@ static void intel_lvds_destroy(struct drm_connector *connector)
 	struct drm_device *dev = connector->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	intel_panel_destroy_backlight(dev);
-
 	if (dev_priv->lid_notifier.notifier_call)
 		acpi_lid_notifier_unregister(&dev_priv->lid_notifier);
 	drm_sysfs_connector_remove(connector);
@@ -662,14 +690,6 @@ static const struct dmi_system_id intel_no_lvds[] = {
 	},
 	{
 		.callback = intel_no_lvds_dmi_callback,
-		.ident = "Dell OptiPlex FX170",
-		.matches = {
-			DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
-			DMI_MATCH(DMI_PRODUCT_NAME, "OptiPlex FX170"),
-		},
-	},
-	{
-		.callback = intel_no_lvds_dmi_callback,
 		.ident = "AOpen Mini PC",
 		.matches = {
 			DMI_MATCH(DMI_SYS_VENDOR, "AOpen"),
@@ -716,14 +736,6 @@ static const struct dmi_system_id intel_no_lvds[] = {
 		},
 	},
 	{
-                .callback = intel_no_lvds_dmi_callback,
-                .ident = "Clientron E830",
-                .matches = {
-                        DMI_MATCH(DMI_SYS_VENDOR, "Clientron"),
-                        DMI_MATCH(DMI_PRODUCT_NAME, "E830"),
-                },
-        },
-        {
 		.callback = intel_no_lvds_dmi_callback,
 		.ident = "Asus EeeBox PC EB1007",
 		.matches = {
@@ -733,30 +745,6 @@ static const struct dmi_system_id intel_no_lvds[] = {
 	},
 	{
 		.callback = intel_no_lvds_dmi_callback,
-		.ident = "Asus AT5NM10T-I",
-		.matches = {
-			DMI_MATCH(DMI_BOARD_VENDOR, "ASUSTeK Computer INC."),
-			DMI_MATCH(DMI_BOARD_NAME, "AT5NM10T-I"),
-		},
-	},
-	{
-		.callback = intel_no_lvds_dmi_callback,
-		.ident = "Hewlett-Packard t5745",
-		.matches = {
-			DMI_MATCH(DMI_BOARD_VENDOR, "Hewlett-Packard"),
-			DMI_MATCH(DMI_BOARD_NAME, "hp t5745"),
-		},
-	},
-	{
-		.callback = intel_no_lvds_dmi_callback,
-		.ident = "Hewlett-Packard st5747",
-		.matches = {
-			DMI_MATCH(DMI_BOARD_VENDOR, "Hewlett-Packard"),
-			DMI_MATCH(DMI_BOARD_NAME, "hp st5747"),
-		},
-	},
-	{
-		.callback = intel_no_lvds_dmi_callback,
 		.ident = "MSI Wind Box DC500",
 		.matches = {
 			DMI_MATCH(DMI_BOARD_VENDOR, "MICRO-STAR INTERNATIONAL CO., LTD"),
@@ -892,18 +880,6 @@ static bool lvds_is_present_in_vbt(struct drm_device *dev,
 	return false;
 }
 
-static bool intel_lvds_supported(struct drm_device *dev)
-{
-	/* With the introduction of the PCH we gained a dedicated
-	 * LVDS presence pin, use it. */
-	if (HAS_PCH_SPLIT(dev))
-		return true;
-
-	/* Otherwise LVDS was only attached to mobile products,
-	 * except for the inglorious 830gm */
-	return IS_MOBILE(dev) && !IS_I830(dev);
-}
-
 /**
  * intel_lvds_init - setup LVDS connectors on this device
  * @dev: drm device
@@ -925,9 +901,6 @@ bool intel_lvds_init(struct drm_device *dev)
 	int pipe;
 	u8 pin;
 
-	if (!intel_lvds_supported(dev))
-		return false;
-
 	/* Skip init on machines we know falsely report LVDS */
 	if (dmi_check_system(intel_no_lvds))
 		return false;
@@ -975,11 +948,9 @@ bool intel_lvds_init(struct drm_device *dev)
 	intel_encoder->type = INTEL_OUTPUT_LVDS;
 
 	intel_encoder->clone_mask = (1 << INTEL_LVDS_CLONE_BIT);
-	if (HAS_PCH_SPLIT(dev))
-		intel_encoder->crtc_mask = (1 << 0) | (1 << 1) | (1 << 2);
-	else
-		intel_encoder->crtc_mask = (1 << 1);
-
+	intel_encoder->crtc_mask = (1 << 1);
+	if (INTEL_INFO(dev)->gen >= 5)
+		intel_encoder->crtc_mask |= (1 << 0);
 	drm_encoder_helper_add(encoder, &intel_lvds_helper_funcs);
 	drm_connector_helper_add(connector, &intel_lvds_connector_helper_funcs);
 	connector->display_info.subpixel_order = SubPixelHorizontalRGB;
@@ -1101,19 +1072,6 @@ out:
 		pwm = I915_READ(BLC_PWM_PCH_CTL1);
 		pwm |= PWM_PCH_ENABLE;
 		I915_WRITE(BLC_PWM_PCH_CTL1, pwm);
-		/*
-		 * Unlock registers and just
-		 * leave them unlocked
-		 */
-		I915_WRITE(PCH_PP_CONTROL,
-			   I915_READ(PCH_PP_CONTROL) | PANEL_UNLOCK_REGS);
-	} else {
-		/*
-		 * Unlock registers and just
-		 * leave them unlocked
-		 */
-		I915_WRITE(PP_CONTROL,
-			   I915_READ(PP_CONTROL) | PANEL_UNLOCK_REGS);
 	}
 	dev_priv->lid_notifier.notifier_call = intel_lid_notify;
 	if (acpi_lid_notifier_register(&dev_priv->lid_notifier)) {
@@ -1123,9 +1081,6 @@ out:
 	/* keep the LVDS connector */
 	dev_priv->int_lvds_connector = connector;
 	drm_sysfs_connector_add(connector);
-
-	intel_panel_setup_backlight(dev);
-
 	return true;
 
 failed:
diff --git a/drivers/gpu/drm/i915/intel_modes.c b/drivers/gpu/drm/i915/intel_modes.c
index d1928e7..3b26a3b 100644
--- a/drivers/gpu/drm/i915/intel_modes.c
+++ b/drivers/gpu/drm/i915/intel_modes.c
@@ -26,9 +26,7 @@
 #include <linux/slab.h>
 #include <linux/i2c.h>
 #include <linux/fb.h>
-#include <drm/drm_edid.h>
 #include "drmP.h"
-#include "drm_edid.h"
 #include "intel_drv.h"
 #include "i915_drv.h"
 
@@ -43,13 +41,13 @@ bool intel_ddc_probe(struct intel_encoder *intel_encoder, int ddc_bus)
 	u8 buf[2];
 	struct i2c_msg msgs[] = {
 		{
-			.addr = DDC_ADDR,
+			.addr = 0x50,
 			.flags = 0,
 			.len = 1,
 			.buf = out_buf,
 		},
 		{
-			.addr = DDC_ADDR,
+			.addr = 0x50,
 			.flags = I2C_M_RD,
 			.len = 1,
 			.buf = buf,
@@ -76,7 +74,6 @@ int intel_ddc_get_modes(struct drm_connector *connector,
 	if (edid) {
 		drm_mode_connector_update_edid_property(connector, edid);
 		ret = drm_add_edid_modes(connector, edid);
-		drm_edid_to_eld(connector, edid);
 		connector->display_info.raw_edid = NULL;
 		kfree(edid);
 	}
@@ -84,11 +81,10 @@ int intel_ddc_get_modes(struct drm_connector *connector,
 	return ret;
 }
 
-static const struct drm_prop_enum_list force_audio_names[] = {
-	{ HDMI_AUDIO_OFF_DVI, "force-dvi" },
-	{ HDMI_AUDIO_OFF, "off" },
-	{ HDMI_AUDIO_AUTO, "auto" },
-	{ HDMI_AUDIO_ON, "on" },
+static const char *force_audio_names[] = {
+	"off",
+	"auto",
+	"on",
 };
 
 void
@@ -97,24 +93,27 @@ intel_attach_force_audio_property(struct drm_connector *connector)
 	struct drm_device *dev = connector->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_property *prop;
+	int i;
 
 	prop = dev_priv->force_audio_property;
 	if (prop == NULL) {
-		prop = drm_property_create_enum(dev, 0,
+		prop = drm_property_create(dev, DRM_MODE_PROP_ENUM,
 					   "audio",
-					   force_audio_names,
 					   ARRAY_SIZE(force_audio_names));
 		if (prop == NULL)
 			return;
 
+		for (i = 0; i < ARRAY_SIZE(force_audio_names); i++)
+			drm_property_add_enum(prop, i, i-1, force_audio_names[i]);
+
 		dev_priv->force_audio_property = prop;
 	}
 	drm_connector_attach_property(connector, prop, 0);
 }
 
-static const struct drm_prop_enum_list broadcast_rgb_names[] = {
-	{ 0, "Full" },
-	{ 1, "Limited 16:235" },
+static const char *broadcast_rgb_names[] = {
+	"Full",
+	"Limited 16:235",
 };
 
 void
@@ -123,16 +122,19 @@ intel_attach_broadcast_rgb_property(struct drm_connector *connector)
 	struct drm_device *dev = connector->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	struct drm_property *prop;
+	int i;
 
 	prop = dev_priv->broadcast_rgb_property;
 	if (prop == NULL) {
-		prop = drm_property_create_enum(dev, DRM_MODE_PROP_ENUM,
+		prop = drm_property_create(dev, DRM_MODE_PROP_ENUM,
 					   "Broadcast RGB",
-					   broadcast_rgb_names,
 					   ARRAY_SIZE(broadcast_rgb_names));
 		if (prop == NULL)
 			return;
 
+		for (i = 0; i < ARRAY_SIZE(broadcast_rgb_names); i++)
+			drm_property_add_enum(prop, i, i, broadcast_rgb_names[i]);
+
 		dev_priv->broadcast_rgb_property = prop;
 	}
 
diff --git a/drivers/gpu/drm/i915/intel_opregion.c b/drivers/gpu/drm/i915/intel_opregion.c
index cffb007..e7a97b5 100644
--- a/drivers/gpu/drm/i915/intel_opregion.c
+++ b/drivers/gpu/drm/i915/intel_opregion.c
@@ -51,61 +51,61 @@
 #define MBOX_ASLE      (1<<2)
 
 struct opregion_header {
-	u8 signature[16];
-	u32 size;
-	u32 opregion_ver;
-	u8 bios_ver[32];
-	u8 vbios_ver[16];
-	u8 driver_ver[16];
-	u32 mboxes;
-	u8 reserved[164];
+       u8 signature[16];
+       u32 size;
+       u32 opregion_ver;
+       u8 bios_ver[32];
+       u8 vbios_ver[16];
+       u8 driver_ver[16];
+       u32 mboxes;
+       u8 reserved[164];
 } __attribute__((packed));
 
 /* OpRegion mailbox #1: public ACPI methods */
 struct opregion_acpi {
-	u32 drdy;       /* driver readiness */
-	u32 csts;       /* notification status */
-	u32 cevt;       /* current event */
-	u8 rsvd1[20];
-	u32 didl[8];    /* supported display devices ID list */
-	u32 cpdl[8];    /* currently presented display list */
-	u32 cadl[8];    /* currently active display list */
-	u32 nadl[8];    /* next active devices list */
-	u32 aslp;       /* ASL sleep time-out */
-	u32 tidx;       /* toggle table index */
-	u32 chpd;       /* current hotplug enable indicator */
-	u32 clid;       /* current lid state*/
-	u32 cdck;       /* current docking state */
-	u32 sxsw;       /* Sx state resume */
-	u32 evts;       /* ASL supported events */
-	u32 cnot;       /* current OS notification */
-	u32 nrdy;       /* driver status */
-	u8 rsvd2[60];
+       u32 drdy;       /* driver readiness */
+       u32 csts;       /* notification status */
+       u32 cevt;       /* current event */
+       u8 rsvd1[20];
+       u32 didl[8];    /* supported display devices ID list */
+       u32 cpdl[8];    /* currently presented display list */
+       u32 cadl[8];    /* currently active display list */
+       u32 nadl[8];    /* next active devices list */
+       u32 aslp;       /* ASL sleep time-out */
+       u32 tidx;       /* toggle table index */
+       u32 chpd;       /* current hotplug enable indicator */
+       u32 clid;       /* current lid state*/
+       u32 cdck;       /* current docking state */
+       u32 sxsw;       /* Sx state resume */
+       u32 evts;       /* ASL supported events */
+       u32 cnot;       /* current OS notification */
+       u32 nrdy;       /* driver status */
+       u8 rsvd2[60];
 } __attribute__((packed));
 
 /* OpRegion mailbox #2: SWSCI */
 struct opregion_swsci {
-	u32 scic;       /* SWSCI command|status|data */
-	u32 parm;       /* command parameters */
-	u32 dslp;       /* driver sleep time-out */
-	u8 rsvd[244];
+       u32 scic;       /* SWSCI command|status|data */
+       u32 parm;       /* command parameters */
+       u32 dslp;       /* driver sleep time-out */
+       u8 rsvd[244];
 } __attribute__((packed));
 
 /* OpRegion mailbox #3: ASLE */
 struct opregion_asle {
-	u32 ardy;       /* driver readiness */
-	u32 aslc;       /* ASLE interrupt command */
-	u32 tche;       /* technology enabled indicator */
-	u32 alsi;       /* current ALS illuminance reading */
-	u32 bclp;       /* backlight brightness to set */
-	u32 pfit;       /* panel fitting state */
-	u32 cblv;       /* current brightness level */
-	u16 bclm[20];   /* backlight level duty cycle mapping table */
-	u32 cpfm;       /* current panel fitting mode */
-	u32 epfm;       /* enabled panel fitting modes */
-	u8 plut[74];    /* panel LUT and identifier */
-	u32 pfmb;       /* PWM freq and min brightness */
-	u8 rsvd[102];
+       u32 ardy;       /* driver readiness */
+       u32 aslc;       /* ASLE interrupt command */
+       u32 tche;       /* technology enabled indicator */
+       u32 alsi;       /* current ALS illuminance reading */
+       u32 bclp;       /* backlight brightness to set */
+       u32 pfit;       /* panel fitting state */
+       u32 cblv;       /* current brightness level */
+       u16 bclm[20];   /* backlight level duty cycle mapping table */
+       u32 cpfm;       /* current panel fitting mode */
+       u32 epfm;       /* enabled panel fitting modes */
+       u8 plut[74];    /* panel LUT and identifier */
+       u32 pfmb;       /* PWM freq and min brightness */
+       u8 rsvd[102];
 } __attribute__((packed));
 
 /* ASLE irq request bits */
@@ -227,6 +227,7 @@ void intel_opregion_asle_intr(struct drm_device *dev)
 	asle->aslc = asle_stat;
 }
 
+/* Only present on Ironlake+ */
 void intel_opregion_gse_intr(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -296,26 +297,19 @@ static int intel_opregion_video_event(struct notifier_block *nb,
 	/* The only video events relevant to opregion are 0x80. These indicate
 	   either a docking event, lid switch or display switch request. In
 	   Linux, these are handled by the dock, button and video drivers.
-	*/
+	   We might want to fix the video driver to be opregion-aware in
+	   future, but right now we just indicate to the firmware that the
+	   request has been handled */
 
 	struct opregion_acpi *acpi;
-	struct acpi_bus_event *event = data;
-	int ret = NOTIFY_OK;
-
-	if (strcmp(event->device_class, ACPI_VIDEO_CLASS) != 0)
-		return NOTIFY_DONE;
 
 	if (!system_opregion)
 		return NOTIFY_DONE;
 
 	acpi = system_opregion->acpi;
-
-	if (event->type == 0x80 && !(acpi->cevt & 0x1))
-		ret = NOTIFY_BAD;
-
 	acpi->csts = 0;
 
-	return ret;
+	return NOTIFY_OK;
 }
 
 static struct notifier_block intel_opregion_notifier = {
@@ -361,7 +355,7 @@ static void intel_didl_outputs(struct drm_device *dev)
 
 	list_for_each_entry(acpi_cdev, &acpi_video_bus->children, node) {
 		if (i >= 8) {
-			dev_printk(KERN_ERR, &dev->pdev->dev,
+			dev_printk (KERN_ERR, &dev->pdev->dev,
 				    "More than 8 outputs detected\n");
 			return;
 		}
@@ -387,7 +381,7 @@ blind_set:
 	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
 		int output_type = ACPI_OTHER_OUTPUT;
 		if (i >= 8) {
-			dev_printk(KERN_ERR, &dev->pdev->dev,
+			dev_printk (KERN_ERR, &dev->pdev->dev,
 				    "More than 8 outputs detected\n");
 			return;
 		}
diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c
index 5ba5e66..1fe7c07 100644
--- a/drivers/gpu/drm/i915/intel_overlay.c
+++ b/drivers/gpu/drm/i915/intel_overlay.c
@@ -25,6 +25,8 @@
  *
  * Derived from Xorg ddx, xf86-video-intel, src/i830_video.c
  */
+
+#include <linux/seq_file.h>
 #include "drmP.h"
 #include "drm.h"
 #include "i915_drm.h"
@@ -115,57 +117,57 @@
 
 /* memory bufferd overlay registers */
 struct overlay_registers {
-	u32 OBUF_0Y;
-	u32 OBUF_1Y;
-	u32 OBUF_0U;
-	u32 OBUF_0V;
-	u32 OBUF_1U;
-	u32 OBUF_1V;
-	u32 OSTRIDE;
-	u32 YRGB_VPH;
-	u32 UV_VPH;
-	u32 HORZ_PH;
-	u32 INIT_PHS;
-	u32 DWINPOS;
-	u32 DWINSZ;
-	u32 SWIDTH;
-	u32 SWIDTHSW;
-	u32 SHEIGHT;
-	u32 YRGBSCALE;
-	u32 UVSCALE;
-	u32 OCLRC0;
-	u32 OCLRC1;
-	u32 DCLRKV;
-	u32 DCLRKM;
-	u32 SCLRKVH;
-	u32 SCLRKVL;
-	u32 SCLRKEN;
-	u32 OCONFIG;
-	u32 OCMD;
-	u32 RESERVED1; /* 0x6C */
-	u32 OSTART_0Y;
-	u32 OSTART_1Y;
-	u32 OSTART_0U;
-	u32 OSTART_0V;
-	u32 OSTART_1U;
-	u32 OSTART_1V;
-	u32 OTILEOFF_0Y;
-	u32 OTILEOFF_1Y;
-	u32 OTILEOFF_0U;
-	u32 OTILEOFF_0V;
-	u32 OTILEOFF_1U;
-	u32 OTILEOFF_1V;
-	u32 FASTHSCALE; /* 0xA0 */
-	u32 UVSCALEV; /* 0xA4 */
-	u32 RESERVEDC[(0x200 - 0xA8) / 4]; /* 0xA8 - 0x1FC */
-	u16 Y_VCOEFS[N_VERT_Y_TAPS * N_PHASES]; /* 0x200 */
-	u16 RESERVEDD[0x100 / 2 - N_VERT_Y_TAPS * N_PHASES];
-	u16 Y_HCOEFS[N_HORIZ_Y_TAPS * N_PHASES]; /* 0x300 */
-	u16 RESERVEDE[0x200 / 2 - N_HORIZ_Y_TAPS * N_PHASES];
-	u16 UV_VCOEFS[N_VERT_UV_TAPS * N_PHASES]; /* 0x500 */
-	u16 RESERVEDF[0x100 / 2 - N_VERT_UV_TAPS * N_PHASES];
-	u16 UV_HCOEFS[N_HORIZ_UV_TAPS * N_PHASES]; /* 0x600 */
-	u16 RESERVEDG[0x100 / 2 - N_HORIZ_UV_TAPS * N_PHASES];
+    u32 OBUF_0Y;
+    u32 OBUF_1Y;
+    u32 OBUF_0U;
+    u32 OBUF_0V;
+    u32 OBUF_1U;
+    u32 OBUF_1V;
+    u32 OSTRIDE;
+    u32 YRGB_VPH;
+    u32 UV_VPH;
+    u32 HORZ_PH;
+    u32 INIT_PHS;
+    u32 DWINPOS;
+    u32 DWINSZ;
+    u32 SWIDTH;
+    u32 SWIDTHSW;
+    u32 SHEIGHT;
+    u32 YRGBSCALE;
+    u32 UVSCALE;
+    u32 OCLRC0;
+    u32 OCLRC1;
+    u32 DCLRKV;
+    u32 DCLRKM;
+    u32 SCLRKVH;
+    u32 SCLRKVL;
+    u32 SCLRKEN;
+    u32 OCONFIG;
+    u32 OCMD;
+    u32 RESERVED1; /* 0x6C */
+    u32 OSTART_0Y;
+    u32 OSTART_1Y;
+    u32 OSTART_0U;
+    u32 OSTART_0V;
+    u32 OSTART_1U;
+    u32 OSTART_1V;
+    u32 OTILEOFF_0Y;
+    u32 OTILEOFF_1Y;
+    u32 OTILEOFF_0U;
+    u32 OTILEOFF_0V;
+    u32 OTILEOFF_1U;
+    u32 OTILEOFF_1V;
+    u32 FASTHSCALE; /* 0xA0 */
+    u32 UVSCALEV; /* 0xA4 */
+    u32 RESERVEDC[(0x200 - 0xA8) / 4]; /* 0xA8 - 0x1FC */
+    u16 Y_VCOEFS[N_VERT_Y_TAPS * N_PHASES]; /* 0x200 */
+    u16 RESERVEDD[0x100 / 2 - N_VERT_Y_TAPS * N_PHASES];
+    u16 Y_HCOEFS[N_HORIZ_Y_TAPS * N_PHASES]; /* 0x300 */
+    u16 RESERVEDE[0x200 / 2 - N_HORIZ_Y_TAPS * N_PHASES];
+    u16 UV_VCOEFS[N_VERT_UV_TAPS * N_PHASES]; /* 0x500 */
+    u16 RESERVEDF[0x100 / 2 - N_VERT_UV_TAPS * N_PHASES];
+    u16 UV_HCOEFS[N_HORIZ_UV_TAPS * N_PHASES]; /* 0x600 */
+    u16 RESERVEDG[0x100 / 2 - N_HORIZ_UV_TAPS * N_PHASES];
 };
 
 struct intel_overlay {
@@ -190,7 +192,7 @@ struct intel_overlay {
 static struct overlay_registers *
 intel_overlay_map_regs(struct intel_overlay *overlay)
 {
-	drm_i915_private_t *dev_priv = overlay->dev->dev_private;
+        drm_i915_private_t *dev_priv = overlay->dev->dev_private;
 	struct overlay_registers *regs;
 
 	if (OVERLAY_NEEDS_PHYSICAL(overlay->dev))
@@ -225,8 +227,7 @@ static int intel_overlay_do_wait_request(struct intel_overlay *overlay,
 	}
 	overlay->last_flip_req = request->seqno;
 	overlay->flip_tail = tail;
-	ret = i915_wait_request(LP_RING(dev_priv), overlay->last_flip_req,
-				true);
+	ret = i915_wait_request(LP_RING(dev_priv), overlay->last_flip_req);
 	if (ret)
 		return ret;
 
@@ -262,8 +263,8 @@ i830_activate_pipe_a(struct drm_device *dev)
 	DRM_DEBUG_DRIVER("Enabling pipe A in order to enable overlay\n");
 
 	mode = drm_mode_duplicate(dev, &vesa_640x480);
-	drm_mode_set_crtcinfo(mode, 0);
-	if (!drm_crtc_helper_set_mode(&crtc->base, mode,
+	drm_mode_set_crtcinfo(mode, CRTC_INTERLACE_HALVE_V);
+	if(!drm_crtc_helper_set_mode(&crtc->base, mode,
 				       crtc->base.x, crtc->base.y,
 				       crtc->base.fb))
 		return 0;
@@ -331,7 +332,7 @@ static int intel_overlay_continue(struct intel_overlay *overlay,
 				  bool load_polyphase_filter)
 {
 	struct drm_device *dev = overlay->dev;
-	drm_i915_private_t *dev_priv = dev->dev_private;
+        drm_i915_private_t *dev_priv = dev->dev_private;
 	struct drm_i915_gem_request *request;
 	u32 flip_addr = overlay->flip_addr;
 	u32 tmp;
@@ -358,7 +359,7 @@ static int intel_overlay_continue(struct intel_overlay *overlay,
 	}
 	OUT_RING(MI_OVERLAY_FLIP | MI_OVERLAY_CONTINUE);
 	OUT_RING(flip_addr);
-	ADVANCE_LP_RING();
+        ADVANCE_LP_RING();
 
 	ret = i915_add_request(LP_RING(dev_priv), NULL, request);
 	if (ret) {
@@ -455,8 +456,7 @@ static int intel_overlay_recover_from_interrupt(struct intel_overlay *overlay)
 	if (overlay->last_flip_req == 0)
 		return 0;
 
-	ret = i915_wait_request(LP_RING(dev_priv), overlay->last_flip_req,
-				true);
+	ret = i915_wait_request(LP_RING(dev_priv), overlay->last_flip_req);
 	if (ret)
 		return ret;
 
@@ -591,7 +591,7 @@ static u32 calc_swidthsw(struct drm_device *dev, u32 offset, u32 width)
 	ret = ((offset + width + mask) >> shift) - (offset >> shift);
 	if (!IS_GEN2(dev))
 		ret <<= 1;
-	ret -= 1;
+	ret -=1;
 	return ret << 2;
 }
 
@@ -781,10 +781,14 @@ static int intel_overlay_do_put_image(struct intel_overlay *overlay,
 	if (ret != 0)
 		return ret;
 
-	ret = i915_gem_object_pin_to_display_plane(new_bo, 0, NULL);
+	ret = i915_gem_object_pin(new_bo, PAGE_SIZE, true);
 	if (ret != 0)
 		return ret;
 
+	ret = i915_gem_object_set_to_gtt_domain(new_bo, 0);
+	if (ret != 0)
+		goto out_unpin;
+
 	ret = i915_gem_object_put_fence(new_bo);
 	if (ret)
 		goto out_unpin;
@@ -825,7 +829,7 @@ static int intel_overlay_do_put_image(struct intel_overlay *overlay,
 	regs->SWIDTHSW = calc_swidthsw(overlay->dev,
 				       params->offset_Y, tmp_width);
 	regs->SHEIGHT = params->src_h;
-	regs->OBUF_0Y = new_bo->gtt_offset + params->offset_Y;
+	regs->OBUF_0Y = new_bo->gtt_offset + params-> offset_Y;
 	regs->OSTRIDE = params->stride_Y;
 
 	if (params->format & I915_OVERLAY_YUV_PLANAR) {
@@ -925,7 +929,7 @@ static void update_pfit_vscale_ratio(struct intel_overlay *overlay)
 	 * line with the intel documentation for the i965
 	 */
 	if (INTEL_INFO(dev)->gen >= 4) {
-		/* on i965 use the PGM reg to read out the autoscaler values */
+	       	/* on i965 use the PGM reg to read out the autoscaler values */
 		ratio = I915_READ(PFIT_PGM_RATIOS) >> PFIT_VERT_SCALE_SHIFT_965;
 	} else {
 		if (pfit_control & VERT_AUTO_SCALE)
@@ -943,10 +947,10 @@ static int check_overlay_dst(struct intel_overlay *overlay,
 {
 	struct drm_display_mode *mode = &overlay->crtc->base.mode;
 
-	if (rec->dst_x < mode->hdisplay &&
-	    rec->dst_x + rec->dst_width <= mode->hdisplay &&
-	    rec->dst_y < mode->vdisplay &&
-	    rec->dst_y + rec->dst_height <= mode->vdisplay)
+	if (rec->dst_x < mode->crtc_hdisplay &&
+	    rec->dst_x + rec->dst_width <= mode->crtc_hdisplay &&
+	    rec->dst_y < mode->crtc_vdisplay &&
+	    rec->dst_y + rec->dst_height <= mode->crtc_vdisplay)
 		return 0;
 	else
 		return -EINVAL;
@@ -1106,7 +1110,7 @@ static int intel_panel_fitter_pipe(struct drm_device *dev)
 }
 
 int intel_overlay_put_image(struct drm_device *dev, void *data,
-			    struct drm_file *file_priv)
+                            struct drm_file *file_priv)
 {
 	struct drm_intel_overlay_put_image *put_image_rec = data;
 	drm_i915_private_t *dev_priv = dev->dev_private;
@@ -1309,10 +1313,10 @@ static int check_gamma(struct drm_intel_overlay_attrs *attrs)
 }
 
 int intel_overlay_attrs(struct drm_device *dev, void *data,
-			struct drm_file *file_priv)
+                        struct drm_file *file_priv)
 {
 	struct drm_intel_overlay_attrs *attrs = data;
-	drm_i915_private_t *dev_priv = dev->dev_private;
+        drm_i915_private_t *dev_priv = dev->dev_private;
 	struct intel_overlay *overlay;
 	struct overlay_registers *regs;
 	int ret;
@@ -1401,7 +1405,7 @@ out_unlock:
 
 void intel_setup_overlay(struct drm_device *dev)
 {
-	drm_i915_private_t *dev_priv = dev->dev_private;
+        drm_i915_private_t *dev_priv = dev->dev_private;
 	struct intel_overlay *overlay;
 	struct drm_i915_gem_object *reg_bo;
 	struct overlay_registers *regs;
@@ -1429,24 +1433,24 @@ void intel_setup_overlay(struct drm_device *dev)
 		ret = i915_gem_attach_phys_object(dev, reg_bo,
 						  I915_GEM_PHYS_OVERLAY_REGS,
 						  PAGE_SIZE);
-		if (ret) {
-			DRM_ERROR("failed to attach phys overlay regs\n");
-			goto out_free_bo;
-		}
+                if (ret) {
+                        DRM_ERROR("failed to attach phys overlay regs\n");
+                        goto out_free_bo;
+                }
 		overlay->flip_addr = reg_bo->phys_obj->handle->busaddr;
 	} else {
 		ret = i915_gem_object_pin(reg_bo, PAGE_SIZE, true);
 		if (ret) {
-			DRM_ERROR("failed to pin overlay register bo\n");
-			goto out_free_bo;
-		}
+                        DRM_ERROR("failed to pin overlay register bo\n");
+                        goto out_free_bo;
+                }
 		overlay->flip_addr = reg_bo->gtt_offset;
 
 		ret = i915_gem_object_set_to_gtt_domain(reg_bo, true);
 		if (ret) {
-			DRM_ERROR("failed to move overlay register bo into the GTT\n");
-			goto out_unpin_bo;
-		}
+                        DRM_ERROR("failed to move overlay register bo into the GTT\n");
+                        goto out_unpin_bo;
+                }
 	}
 
 	/* init all values */
@@ -1533,7 +1537,7 @@ static void intel_overlay_unmap_regs_atomic(struct intel_overlay *overlay,
 struct intel_overlay_error_state *
 intel_overlay_capture_error_state(struct drm_device *dev)
 {
-	drm_i915_private_t *dev_priv = dev->dev_private;
+        drm_i915_private_t *dev_priv = dev->dev_private;
 	struct intel_overlay *overlay = dev_priv->overlay;
 	struct intel_overlay_error_state *error;
 	struct overlay_registers __iomem *regs;
diff --git a/drivers/gpu/drm/i915/intel_panel.c b/drivers/gpu/drm/i915/intel_panel.c
index 230a141..f8aa821 100644
--- a/drivers/gpu/drm/i915/intel_panel.c
+++ b/drivers/gpu/drm/i915/intel_panel.c
@@ -48,7 +48,7 @@ intel_fixed_panel_mode(struct drm_display_mode *fixed_mode,
 
 	adjusted_mode->clock = fixed_mode->clock;
 
-	drm_mode_set_crtcinfo(adjusted_mode, 0);
+	drm_mode_set_crtcinfo(adjusted_mode, CRTC_INTERLACE_HALVE_V);
 }
 
 /* adjusted_mode has been preset to be the panel's fixed mode */
@@ -84,7 +84,7 @@ intel_pch_panel_fitting(struct drm_device *dev,
 			if (scaled_width > scaled_height) { /* pillar */
 				width = scaled_height / mode->vdisplay;
 				if (width & 1)
-					width++;
+				    	width++;
 				x = (adjusted_mode->hdisplay - width + 1) / 2;
 				y = 0;
 				height = adjusted_mode->vdisplay;
@@ -141,8 +141,8 @@ static u32 i915_read_blc_pwm_ctl(struct drm_i915_private *dev_priv)
 			dev_priv->saveBLC_PWM_CTL2 = val;
 		} else if (val == 0) {
 			I915_WRITE(BLC_PWM_PCH_CTL2,
-				   dev_priv->saveBLC_PWM_CTL2);
-			val = dev_priv->saveBLC_PWM_CTL2;
+				   dev_priv->saveBLC_PWM_CTL);
+			val = dev_priv->saveBLC_PWM_CTL;
 		}
 	} else {
 		val = I915_READ(BLC_PWM_CTL);
@@ -178,10 +178,13 @@ u32 intel_panel_get_max_backlight(struct drm_device *dev)
 	if (HAS_PCH_SPLIT(dev)) {
 		max >>= 16;
 	} else {
-		if (INTEL_INFO(dev)->gen < 4)
+		if (IS_PINEVIEW(dev)) {
 			max >>= 17;
-		else
+		} else {
 			max >>= 16;
+			if (INTEL_INFO(dev)->gen < 4)
+				max &= ~1;
+		}
 
 		if (is_backlight_combination_mode(dev))
 			max *= 0xff;
@@ -200,12 +203,13 @@ u32 intel_panel_get_backlight(struct drm_device *dev)
 		val = I915_READ(BLC_PWM_CPU_CTL) & BACKLIGHT_DUTY_CYCLE_MASK;
 	} else {
 		val = I915_READ(BLC_PWM_CTL) & BACKLIGHT_DUTY_CYCLE_MASK;
-		if (INTEL_INFO(dev)->gen < 4)
+		if (IS_PINEVIEW(dev))
 			val >>= 1;
 
-		if (is_backlight_combination_mode(dev)) {
+		if (is_backlight_combination_mode(dev)){
 			u8 lbpc;
 
+			val &= ~1;
 			pci_read_config_byte(dev->pdev, PCI_LBPC, &lbpc);
 			val *= lbpc;
 		}
@@ -232,7 +236,7 @@ static void intel_panel_actually_set_backlight(struct drm_device *dev, u32 level
 	if (HAS_PCH_SPLIT(dev))
 		return intel_pch_panel_set_backlight(dev, level);
 
-	if (is_backlight_combination_mode(dev)) {
+	if (is_backlight_combination_mode(dev)){
 		u32 max = intel_panel_get_max_backlight(dev);
 		u8 lbpc;
 
@@ -242,9 +246,11 @@ static void intel_panel_actually_set_backlight(struct drm_device *dev, u32 level
 	}
 
 	tmp = I915_READ(BLC_PWM_CTL);
-	if (INTEL_INFO(dev)->gen < 4) 
+	if (IS_PINEVIEW(dev)) {
+		tmp &= ~(BACKLIGHT_DUTY_CYCLE_MASK - 1);
 		level <<= 1;
-	tmp &= ~BACKLIGHT_DUTY_CYCLE_MASK;
+	} else
+		tmp &= ~BACKLIGHT_DUTY_CYCLE_MASK;
 	I915_WRITE(BLC_PWM_CTL, tmp | level);
 }
 
@@ -276,7 +282,7 @@ void intel_panel_enable_backlight(struct drm_device *dev)
 	intel_panel_actually_set_backlight(dev, dev_priv->backlight_level);
 }
 
-static void intel_panel_init_backlight(struct drm_device *dev)
+void intel_panel_setup_backlight(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
@@ -308,74 +314,3 @@ intel_panel_detect(struct drm_device *dev)
 
 	return connector_status_unknown;
 }
-
-#ifdef CONFIG_BACKLIGHT_CLASS_DEVICE
-static int intel_panel_update_status(struct backlight_device *bd)
-{
-	struct drm_device *dev = bl_get_data(bd);
-	intel_panel_set_backlight(dev, bd->props.brightness);
-	return 0;
-}
-
-static int intel_panel_get_brightness(struct backlight_device *bd)
-{
-	struct drm_device *dev = bl_get_data(bd);
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	return dev_priv->backlight_level;
-}
-
-static const struct backlight_ops intel_panel_bl_ops = {
-	.update_status = intel_panel_update_status,
-	.get_brightness = intel_panel_get_brightness,
-};
-
-int intel_panel_setup_backlight(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct backlight_properties props;
-	struct drm_connector *connector;
-
-	intel_panel_init_backlight(dev);
-
-	if (dev_priv->int_lvds_connector)
-		connector = dev_priv->int_lvds_connector;
-	else if (dev_priv->int_edp_connector)
-		connector = dev_priv->int_edp_connector;
-	else
-		return -ENODEV;
-
-	props.type = BACKLIGHT_RAW;
-	props.max_brightness = intel_panel_get_max_backlight(dev);
-	dev_priv->backlight =
-		backlight_device_register("intel_backlight",
-					  &connector->kdev, dev,
-					  &intel_panel_bl_ops, &props);
-
-	if (IS_ERR(dev_priv->backlight)) {
-		DRM_ERROR("Failed to register backlight: %ld\n",
-			  PTR_ERR(dev_priv->backlight));
-		dev_priv->backlight = NULL;
-		return -ENODEV;
-	}
-	dev_priv->backlight->props.brightness = intel_panel_get_backlight(dev);
-	return 0;
-}
-
-void intel_panel_destroy_backlight(struct drm_device *dev)
-{
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	if (dev_priv->backlight)
-		backlight_device_unregister(dev_priv->backlight);
-}
-#else
-int intel_panel_setup_backlight(struct drm_device *dev)
-{
-	intel_panel_init_backlight(dev);
-	return 0;
-}
-
-void intel_panel_destroy_backlight(struct drm_device *dev)
-{
-	return;
-}
-#endif
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c
index d9777a2..3bd85f7 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.c
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.c
@@ -34,16 +34,6 @@
 #include "i915_trace.h"
 #include "intel_drv.h"
 
-/*
- * 965+ support PIPE_CONTROL commands, which provide finer grained control
- * over cache flushing.
- */
-struct pipe_control {
-	struct drm_i915_gem_object *obj;
-	volatile u32 *cpu_page;
-	u32 gtt_offset;
-};
-
 static inline int ring_space(struct intel_ring_buffer *ring)
 {
 	int space = (ring->head & HEAD_ADDR) - (ring->tail + 8);
@@ -52,6 +42,20 @@ static inline int ring_space(struct intel_ring_buffer *ring)
 	return space;
 }
 
+static u32 i915_gem_get_seqno(struct drm_device *dev)
+{
+	drm_i915_private_t *dev_priv = dev->dev_private;
+	u32 seqno;
+
+	seqno = dev_priv->next_seqno;
+
+	/* reserve 0 for non-seqno */
+	if (++dev_priv->next_seqno == 0)
+		dev_priv->next_seqno = 1;
+
+	return seqno;
+}
+
 static int
 render_ring_flush(struct intel_ring_buffer *ring,
 		  u32	invalidate_domains,
@@ -119,118 +123,6 @@ render_ring_flush(struct intel_ring_buffer *ring,
 	return 0;
 }
 
-/**
- * Emits a PIPE_CONTROL with a non-zero post-sync operation, for
- * implementing two workarounds on gen6.  From section 1.4.7.1
- * "PIPE_CONTROL" of the Sandy Bridge PRM volume 2 part 1:
- *
- * [DevSNB-C+{W/A}] Before any depth stall flush (including those
- * produced by non-pipelined state commands), software needs to first
- * send a PIPE_CONTROL with no bits set except Post-Sync Operation !=
- * 0.
- *
- * [Dev-SNB{W/A}]: Before a PIPE_CONTROL with Write Cache Flush Enable
- * =1, a PIPE_CONTROL with any non-zero post-sync-op is required.
- *
- * And the workaround for these two requires this workaround first:
- *
- * [Dev-SNB{W/A}]: Pipe-control with CS-stall bit set must be sent
- * BEFORE the pipe-control with a post-sync op and no write-cache
- * flushes.
- *
- * And this last workaround is tricky because of the requirements on
- * that bit.  From section 1.4.7.2.3 "Stall" of the Sandy Bridge PRM
- * volume 2 part 1:
- *
- *     "1 of the following must also be set:
- *      - Render Target Cache Flush Enable ([12] of DW1)
- *      - Depth Cache Flush Enable ([0] of DW1)
- *      - Stall at Pixel Scoreboard ([1] of DW1)
- *      - Depth Stall ([13] of DW1)
- *      - Post-Sync Operation ([13] of DW1)
- *      - Notify Enable ([8] of DW1)"
- *
- * The cache flushes require the workaround flush that triggered this
- * one, so we can't use it.  Depth stall would trigger the same.
- * Post-sync nonzero is what triggered this second workaround, so we
- * can't use that one either.  Notify enable is IRQs, which aren't
- * really our business.  That leaves only stall at scoreboard.
- */
-static int
-intel_emit_post_sync_nonzero_flush(struct intel_ring_buffer *ring)
-{
-	struct pipe_control *pc = ring->private;
-	u32 scratch_addr = pc->gtt_offset + 128;
-	int ret;
-
-
-	ret = intel_ring_begin(ring, 6);
-	if (ret)
-		return ret;
-
-	intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(5));
-	intel_ring_emit(ring, PIPE_CONTROL_CS_STALL |
-			PIPE_CONTROL_STALL_AT_SCOREBOARD);
-	intel_ring_emit(ring, scratch_addr | PIPE_CONTROL_GLOBAL_GTT); /* address */
-	intel_ring_emit(ring, 0); /* low dword */
-	intel_ring_emit(ring, 0); /* high dword */
-	intel_ring_emit(ring, MI_NOOP);
-	intel_ring_advance(ring);
-
-	ret = intel_ring_begin(ring, 6);
-	if (ret)
-		return ret;
-
-	intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(5));
-	intel_ring_emit(ring, PIPE_CONTROL_QW_WRITE);
-	intel_ring_emit(ring, scratch_addr | PIPE_CONTROL_GLOBAL_GTT); /* address */
-	intel_ring_emit(ring, 0);
-	intel_ring_emit(ring, 0);
-	intel_ring_emit(ring, MI_NOOP);
-	intel_ring_advance(ring);
-
-	return 0;
-}
-
-static int
-gen6_render_ring_flush(struct intel_ring_buffer *ring,
-                         u32 invalidate_domains, u32 flush_domains)
-{
-	u32 flags = 0;
-	struct pipe_control *pc = ring->private;
-	u32 scratch_addr = pc->gtt_offset + 128;
-	int ret;
-
-	/* Force SNB workarounds for PIPE_CONTROL flushes */
-	intel_emit_post_sync_nonzero_flush(ring);
-
-	/* Just flush everything.  Experiments have shown that reducing the
-	 * number of bits based on the write domains has little performance
-	 * impact.
-	 */
-	flags |= PIPE_CONTROL_RENDER_TARGET_CACHE_FLUSH;
-	flags |= PIPE_CONTROL_INSTRUCTION_CACHE_INVALIDATE;
-	flags |= PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE;
-	flags |= PIPE_CONTROL_DEPTH_CACHE_FLUSH;
-	flags |= PIPE_CONTROL_VF_CACHE_INVALIDATE;
-	flags |= PIPE_CONTROL_CONST_CACHE_INVALIDATE;
-	flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE;
-
-	ret = intel_ring_begin(ring, 6);
-	if (ret)
-		return ret;
-
-	intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(5));
-	intel_ring_emit(ring, flags);
-	intel_ring_emit(ring, scratch_addr | PIPE_CONTROL_GLOBAL_GTT);
-	intel_ring_emit(ring, 0); /* lower dword */
-	intel_ring_emit(ring, 0); /* uppwer dword */
-	intel_ring_emit(ring, MI_NOOP);
-	intel_ring_advance(ring);
-
-	return 0;
-}
-
 static void ring_write_tail(struct intel_ring_buffer *ring,
 			    u32 value)
 {
@@ -290,7 +182,7 @@ static int init_ring_common(struct intel_ring_buffer *ring)
 	I915_WRITE_START(ring, obj->gtt_offset);
 	I915_WRITE_CTL(ring,
 			((ring->size - PAGE_SIZE) & RING_NR_PAGES)
-			| RING_VALID);
+			| RING_REPORT_64K | RING_VALID);
 
 	/* If the head is still not zero, the ring is dead */
 	if ((I915_READ_CTL(ring) & RING_VALID) == 0 ||
@@ -317,6 +209,16 @@ static int init_ring_common(struct intel_ring_buffer *ring)
 	return 0;
 }
 
+/*
+ * 965+ support PIPE_CONTROL commands, which provide finer grained control
+ * over cache flushing.
+ */
+struct pipe_control {
+	struct drm_i915_gem_object *obj;
+	volatile u32 *cpu_page;
+	u32 gtt_offset;
+};
+
 static int
 init_pipe_control(struct intel_ring_buffer *ring)
 {
@@ -337,8 +239,7 @@ init_pipe_control(struct intel_ring_buffer *ring)
 		ret = -ENOMEM;
 		goto err;
 	}
-
-	i915_gem_object_set_cache_level(obj, I915_CACHE_LLC);
+	obj->cache_level = I915_CACHE_LLC;
 
 	ret = i915_gem_object_pin(obj, 4096, true);
 	if (ret)
@@ -388,24 +289,18 @@ static int init_render_ring(struct intel_ring_buffer *ring)
 
 	if (INTEL_INFO(dev)->gen > 3) {
 		int mode = VS_TIMER_DISPATCH << 16 | VS_TIMER_DISPATCH;
+		if (IS_GEN6(dev) || IS_GEN7(dev))
+			mode |= MI_FLUSH_ENABLE << 16 | MI_FLUSH_ENABLE;
 		I915_WRITE(MI_MODE, mode);
-		if (IS_GEN7(dev))
-			I915_WRITE(GFX_MODE_GEN7,
-				   GFX_MODE_DISABLE(GFX_TLB_INVALIDATE_ALWAYS) |
-				   GFX_MODE_ENABLE(GFX_REPLAY_MODE));
 	}
 
-	if (INTEL_INFO(dev)->gen >= 5) {
+	if (INTEL_INFO(dev)->gen >= 6) {
+	} else if (IS_GEN5(dev)) {
 		ret = init_pipe_control(ring);
 		if (ret)
 			return ret;
 	}
 
-	if (INTEL_INFO(dev)->gen >= 6) {
-		I915_WRITE(INSTPM,
-			   INSTPM_FORCE_ORDERING << 16 | INSTPM_FORCE_ORDERING);
-	}
-
 	return ret;
 }
 
@@ -418,131 +313,83 @@ static void render_ring_cleanup(struct intel_ring_buffer *ring)
 }
 
 static void
-update_mboxes(struct intel_ring_buffer *ring,
-	    u32 seqno,
-	    u32 mmio_offset)
+update_semaphore(struct intel_ring_buffer *ring, int i, u32 seqno)
 {
-	intel_ring_emit(ring, MI_SEMAPHORE_MBOX |
-			      MI_SEMAPHORE_GLOBAL_GTT |
-			      MI_SEMAPHORE_REGISTER |
-			      MI_SEMAPHORE_UPDATE);
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	int id;
+
+	/*
+	 * cs -> 1 = vcs, 0 = bcs
+	 * vcs -> 1 = bcs, 0 = cs,
+	 * bcs -> 1 = cs, 0 = vcs.
+	 */
+	id = ring - dev_priv->ring;
+	id += 2 - i;
+	id %= 3;
+
+	intel_ring_emit(ring,
+			MI_SEMAPHORE_MBOX |
+			MI_SEMAPHORE_REGISTER |
+			MI_SEMAPHORE_UPDATE);
 	intel_ring_emit(ring, seqno);
-	intel_ring_emit(ring, mmio_offset);
+	intel_ring_emit(ring,
+			RING_SYNC_0(dev_priv->ring[id].mmio_base) + 4*i);
 }
 
-/**
- * gen6_add_request - Update the semaphore mailbox registers
- * 
- * @ring - ring that is adding a request
- * @seqno - return seqno stuck into the ring
- *
- * Update the mailbox registers in the *other* rings with the current seqno.
- * This acts like a signal in the canonical semaphore.
- */
 static int
 gen6_add_request(struct intel_ring_buffer *ring,
-		 u32 *seqno)
+		 u32 *result)
 {
-	u32 mbox1_reg;
-	u32 mbox2_reg;
+	u32 seqno;
 	int ret;
 
 	ret = intel_ring_begin(ring, 10);
 	if (ret)
 		return ret;
 
-	mbox1_reg = ring->signal_mbox[0];
-	mbox2_reg = ring->signal_mbox[1];
+	seqno = i915_gem_get_seqno(ring->dev);
+	update_semaphore(ring, 0, seqno);
+	update_semaphore(ring, 1, seqno);
 
-	*seqno = i915_gem_next_request_seqno(ring);
-
-	update_mboxes(ring, *seqno, mbox1_reg);
-	update_mboxes(ring, *seqno, mbox2_reg);
 	intel_ring_emit(ring, MI_STORE_DWORD_INDEX);
 	intel_ring_emit(ring, I915_GEM_HWS_INDEX << MI_STORE_DWORD_INDEX_SHIFT);
-	intel_ring_emit(ring, *seqno);
+	intel_ring_emit(ring, seqno);
 	intel_ring_emit(ring, MI_USER_INTERRUPT);
 	intel_ring_advance(ring);
 
+	*result = seqno;
 	return 0;
 }
 
-/**
- * intel_ring_sync - sync the waiter to the signaller on seqno
- *
- * @waiter - ring that is waiting
- * @signaller - ring which has, or will signal
- * @seqno - seqno which the waiter will block on
- */
-static int
-intel_ring_sync(struct intel_ring_buffer *waiter,
-		struct intel_ring_buffer *signaller,
-		int ring,
+int
+intel_ring_sync(struct intel_ring_buffer *ring,
+		struct intel_ring_buffer *to,
 		u32 seqno)
 {
 	int ret;
-	u32 dw1 = MI_SEMAPHORE_MBOX |
-		  MI_SEMAPHORE_COMPARE |
-		  MI_SEMAPHORE_REGISTER;
 
-	ret = intel_ring_begin(waiter, 4);
+	ret = intel_ring_begin(ring, 4);
 	if (ret)
 		return ret;
 
-	intel_ring_emit(waiter, dw1 | signaller->semaphore_register[ring]);
-	intel_ring_emit(waiter, seqno);
-	intel_ring_emit(waiter, 0);
-	intel_ring_emit(waiter, MI_NOOP);
-	intel_ring_advance(waiter);
+	intel_ring_emit(ring,
+			MI_SEMAPHORE_MBOX |
+			MI_SEMAPHORE_REGISTER |
+			intel_ring_sync_index(ring, to) << 17 |
+			MI_SEMAPHORE_COMPARE);
+	intel_ring_emit(ring, seqno);
+	intel_ring_emit(ring, 0);
+	intel_ring_emit(ring, MI_NOOP);
+	intel_ring_advance(ring);
 
 	return 0;
 }
 
-/* VCS->RCS (RVSYNC) or BCS->RCS (RBSYNC) */
-int
-render_ring_sync_to(struct intel_ring_buffer *waiter,
-		    struct intel_ring_buffer *signaller,
-		    u32 seqno)
-{
-	WARN_ON(signaller->semaphore_register[RCS] == MI_SEMAPHORE_SYNC_INVALID);
-	return intel_ring_sync(waiter,
-			       signaller,
-			       RCS,
-			       seqno);
-}
-
-/* RCS->VCS (VRSYNC) or BCS->VCS (VBSYNC) */
-int
-gen6_bsd_ring_sync_to(struct intel_ring_buffer *waiter,
-		      struct intel_ring_buffer *signaller,
-		      u32 seqno)
-{
-	WARN_ON(signaller->semaphore_register[VCS] == MI_SEMAPHORE_SYNC_INVALID);
-	return intel_ring_sync(waiter,
-			       signaller,
-			       VCS,
-			       seqno);
-}
-
-/* RCS->BCS (BRSYNC) or VCS->BCS (BVSYNC) */
-int
-gen6_blt_ring_sync_to(struct intel_ring_buffer *waiter,
-		      struct intel_ring_buffer *signaller,
-		      u32 seqno)
-{
-	WARN_ON(signaller->semaphore_register[BCS] == MI_SEMAPHORE_SYNC_INVALID);
-	return intel_ring_sync(waiter,
-			       signaller,
-			       BCS,
-			       seqno);
-}
-
-
-
 #define PIPE_CONTROL_FLUSH(ring__, addr__)					\
 do {									\
-	intel_ring_emit(ring__, GFX_OP_PIPE_CONTROL(4) | PIPE_CONTROL_QW_WRITE |		\
-		 PIPE_CONTROL_DEPTH_STALL);				\
+	intel_ring_emit(ring__, GFX_OP_PIPE_CONTROL | PIPE_CONTROL_QW_WRITE |		\
+		 PIPE_CONTROL_DEPTH_STALL | 2);				\
 	intel_ring_emit(ring__, (addr__) | PIPE_CONTROL_GLOBAL_GTT);			\
 	intel_ring_emit(ring__, 0);							\
 	intel_ring_emit(ring__, 0);							\
@@ -552,7 +399,8 @@ static int
 pc_render_add_request(struct intel_ring_buffer *ring,
 		      u32 *result)
 {
-	u32 seqno = i915_gem_next_request_seqno(ring);
+	struct drm_device *dev = ring->dev;
+	u32 seqno = i915_gem_get_seqno(dev);
 	struct pipe_control *pc = ring->private;
 	u32 scratch_addr = pc->gtt_offset + 128;
 	int ret;
@@ -569,9 +417,8 @@ pc_render_add_request(struct intel_ring_buffer *ring,
 	if (ret)
 		return ret;
 
-	intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(4) | PIPE_CONTROL_QW_WRITE |
-			PIPE_CONTROL_WRITE_FLUSH |
-			PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE);
+	intel_ring_emit(ring, GFX_OP_PIPE_CONTROL | PIPE_CONTROL_QW_WRITE |
+			PIPE_CONTROL_WC_FLUSH | PIPE_CONTROL_TC_FLUSH);
 	intel_ring_emit(ring, pc->gtt_offset | PIPE_CONTROL_GLOBAL_GTT);
 	intel_ring_emit(ring, seqno);
 	intel_ring_emit(ring, 0);
@@ -586,10 +433,8 @@ pc_render_add_request(struct intel_ring_buffer *ring,
 	PIPE_CONTROL_FLUSH(ring, scratch_addr);
 	scratch_addr += 128;
 	PIPE_CONTROL_FLUSH(ring, scratch_addr);
-
-	intel_ring_emit(ring, GFX_OP_PIPE_CONTROL(4) | PIPE_CONTROL_QW_WRITE |
-			PIPE_CONTROL_WRITE_FLUSH |
-			PIPE_CONTROL_TEXTURE_CACHE_INVALIDATE |
+	intel_ring_emit(ring, GFX_OP_PIPE_CONTROL | PIPE_CONTROL_QW_WRITE |
+			PIPE_CONTROL_WC_FLUSH | PIPE_CONTROL_TC_FLUSH |
 			PIPE_CONTROL_NOTIFY);
 	intel_ring_emit(ring, pc->gtt_offset | PIPE_CONTROL_GLOBAL_GTT);
 	intel_ring_emit(ring, seqno);
@@ -604,7 +449,8 @@ static int
 render_ring_add_request(struct intel_ring_buffer *ring,
 			u32 *result)
 {
-	u32 seqno = i915_gem_next_request_seqno(ring);
+	struct drm_device *dev = ring->dev;
+	u32 seqno = i915_gem_get_seqno(dev);
 	int ret;
 
 	ret = intel_ring_begin(ring, 4);
@@ -622,19 +468,6 @@ render_ring_add_request(struct intel_ring_buffer *ring,
 }
 
 static u32
-gen6_ring_get_seqno(struct intel_ring_buffer *ring)
-{
-	struct drm_device *dev = ring->dev;
-
-	/* Workaround to force correct ordering between irq and seqno writes on
-	 * ivb (and maybe also on snb) by reading from a CS register (like
-	 * ACTHD) before reading the status page. */
-	if (IS_GEN7(dev))
-		intel_ring_get_active_head(ring);
-	return intel_read_status_page(ring, I915_GEM_HWS_INDEX);
-}
-
-static u32
 ring_get_seqno(struct intel_ring_buffer *ring)
 {
 	return intel_read_status_page(ring, I915_GEM_HWS_INDEX);
@@ -730,13 +563,13 @@ void intel_ring_setup_status_page(struct intel_ring_buffer *ring)
 	 */
 	if (IS_GEN7(dev)) {
 		switch (ring->id) {
-		case RCS:
+		case RING_RENDER:
 			mmio = RENDER_HWS_PGA_GEN7;
 			break;
-		case BCS:
+		case RING_BLT:
 			mmio = BLT_HWS_PGA_GEN7;
 			break;
-		case VCS:
+		case RING_BSD:
 			mmio = BSD_HWS_PGA_GEN7;
 			break;
 		}
@@ -778,7 +611,7 @@ ring_add_request(struct intel_ring_buffer *ring,
 	if (ret)
 		return ret;
 
-	seqno = i915_gem_next_request_seqno(ring);
+	seqno = i915_gem_get_seqno(ring->dev);
 
 	intel_ring_emit(ring, MI_STORE_DWORD_INDEX);
 	intel_ring_emit(ring, I915_GEM_HWS_INDEX << MI_STORE_DWORD_INDEX_SHIFT);
@@ -799,11 +632,6 @@ gen6_ring_get_irq(struct intel_ring_buffer *ring, u32 gflag, u32 rflag)
 	if (!dev->irq_enabled)
 	       return false;
 
-	/* It looks like we need to prevent the gt from suspending while waiting
-	 * for an notifiy irq, otherwise irqs seem to get lost on at least the
-	 * blt/bsd rings on ivb. */
-	gen6_gt_force_wake_get(dev_priv);
-
 	spin_lock(&ring->irq_lock);
 	if (ring->irq_refcount++ == 0) {
 		ring->irq_mask &= ~rflag;
@@ -828,8 +656,6 @@ gen6_ring_put_irq(struct intel_ring_buffer *ring, u32 gflag, u32 rflag)
 		ironlake_disable_irq(dev_priv, gflag);
 	}
 	spin_unlock(&ring->irq_lock);
-
-	gen6_gt_force_wake_put(dev_priv);
 }
 
 static bool
@@ -953,8 +779,7 @@ static int init_status_page(struct intel_ring_buffer *ring)
 		ret = -ENOMEM;
 		goto err;
 	}
-
-	i915_gem_object_set_cache_level(obj, I915_CACHE_LLC);
+	obj->cache_level = I915_CACHE_LLC;
 
 	ret = i915_gem_object_pin(obj, 4096, true);
 	if (ret != 0) {
@@ -1111,105 +936,26 @@ static int intel_wrap_ring_buffer(struct intel_ring_buffer *ring)
 	return 0;
 }
 
-static int intel_ring_wait_seqno(struct intel_ring_buffer *ring, u32 seqno)
+int intel_wait_ring_buffer(struct intel_ring_buffer *ring, int n)
 {
-	struct drm_i915_private *dev_priv = ring->dev->dev_private;
-	bool was_interruptible;
-	int ret;
+	struct drm_device *dev = ring->dev;
+	struct drm_i915_private *dev_priv = dev->dev_private;
+	unsigned long end;
+	u32 head;
 
-	/* XXX As we have not yet audited all the paths to check that
-	 * they are ready for ERESTARTSYS from intel_ring_begin, do not
-	 * allow us to be interruptible by a signal.
+	/* If the reported head position has wrapped or hasn't advanced,
+	 * fallback to the slow and accurate path.
 	 */
-	was_interruptible = dev_priv->mm.interruptible;
-	dev_priv->mm.interruptible = false;
-
-	ret = i915_wait_request(ring, seqno, true);
-
-	dev_priv->mm.interruptible = was_interruptible;
-
-	return ret;
-}
-
-static int intel_ring_wait_request(struct intel_ring_buffer *ring, int n)
-{
-	struct drm_i915_gem_request *request;
-	u32 seqno = 0;
-	int ret;
-
-	i915_gem_retire_requests_ring(ring);
-
-	if (ring->last_retired_head != -1) {
-		ring->head = ring->last_retired_head;
-		ring->last_retired_head = -1;
+	head = intel_read_status_page(ring, 4);
+	if (head > ring->head) {
+		ring->head = head;
 		ring->space = ring_space(ring);
 		if (ring->space >= n)
 			return 0;
 	}
 
-	list_for_each_entry(request, &ring->request_list, list) {
-		int space;
-
-		if (request->tail == -1)
-			continue;
-
-		space = request->tail - (ring->tail + 8);
-		if (space < 0)
-			space += ring->size;
-		if (space >= n) {
-			seqno = request->seqno;
-			break;
-		}
-
-		/* Consume this request in case we need more space than
-		 * is available and so need to prevent a race between
-		 * updating last_retired_head and direct reads of
-		 * I915_RING_HEAD. It also provides a nice sanity check.
-		 */
-		request->tail = -1;
-	}
-
-	if (seqno == 0)
-		return -ENOSPC;
-
-	ret = intel_ring_wait_seqno(ring, seqno);
-	if (ret)
-		return ret;
-
-	if (WARN_ON(ring->last_retired_head == -1))
-		return -ENOSPC;
-
-	ring->head = ring->last_retired_head;
-	ring->last_retired_head = -1;
-	ring->space = ring_space(ring);
-	if (WARN_ON(ring->space < n))
-		return -ENOSPC;
-
-	return 0;
-}
-
-int intel_wait_ring_buffer(struct intel_ring_buffer *ring, int n)
-{
-	struct drm_device *dev = ring->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	unsigned long end;
-	int ret;
-
-	ret = intel_ring_wait_request(ring, n);
-	if (ret != -ENOSPC)
-		return ret;
-
 	trace_i915_ring_wait_begin(ring);
-	if (drm_core_check_feature(dev, DRIVER_GEM))
-		/* With GEM the hangcheck timer should kick us out of the loop,
-		 * leaving it early runs the risk of corrupting GEM state (due
-		 * to running on almost untested codepaths). But on resume
-		 * timers don't work yet, so prevent a complete hang in that
-		 * case by choosing an insanely large timeout. */
-		end = jiffies + 60 * HZ;
-	else
-		end = jiffies + 3 * HZ;
-
+	end = jiffies + 3 * HZ;
 	do {
 		ring->head = I915_READ_HEAD(ring);
 		ring->space = ring_space(ring);
@@ -1266,7 +1012,7 @@ void intel_ring_advance(struct intel_ring_buffer *ring)
 
 static const struct intel_ring_buffer render_ring = {
 	.name			= "render ring",
-	.id			= RCS,
+	.id			= RING_RENDER,
 	.mmio_base		= RENDER_RING_BASE,
 	.size			= 32 * PAGE_SIZE,
 	.init			= init_render_ring,
@@ -1277,19 +1023,14 @@ static const struct intel_ring_buffer render_ring = {
 	.irq_get		= render_ring_get_irq,
 	.irq_put		= render_ring_put_irq,
 	.dispatch_execbuffer	= render_ring_dispatch_execbuffer,
-	.cleanup		= render_ring_cleanup,
-	.sync_to		= render_ring_sync_to,
-	.semaphore_register	= {MI_SEMAPHORE_SYNC_INVALID,
-				   MI_SEMAPHORE_SYNC_RV,
-				   MI_SEMAPHORE_SYNC_RB},
-	.signal_mbox		= {GEN6_VRSYNC, GEN6_BRSYNC},
+       .cleanup			= render_ring_cleanup,
 };
 
 /* ring buffer for bit-stream decoder */
 
 static const struct intel_ring_buffer bsd_ring = {
 	.name                   = "bsd ring",
-	.id			= VCS,
+	.id			= RING_BSD,
 	.mmio_base		= BSD_RING_BASE,
 	.size			= 32 * PAGE_SIZE,
 	.init			= init_ring_common,
@@ -1306,23 +1047,23 @@ static const struct intel_ring_buffer bsd_ring = {
 static void gen6_bsd_ring_write_tail(struct intel_ring_buffer *ring,
 				     u32 value)
 {
-	drm_i915_private_t *dev_priv = ring->dev->dev_private;
+       drm_i915_private_t *dev_priv = ring->dev->dev_private;
 
        /* Every tail move must follow the sequence below */
-	I915_WRITE(GEN6_BSD_SLEEP_PSMI_CONTROL,
-		GEN6_BSD_SLEEP_PSMI_CONTROL_RC_ILDL_MESSAGE_MODIFY_MASK |
-		GEN6_BSD_SLEEP_PSMI_CONTROL_RC_ILDL_MESSAGE_DISABLE);
-	I915_WRITE(GEN6_BSD_RNCID, 0x0);
-
-	if (wait_for((I915_READ(GEN6_BSD_SLEEP_PSMI_CONTROL) &
-		GEN6_BSD_SLEEP_PSMI_CONTROL_IDLE_INDICATOR) == 0,
-		50))
-	DRM_ERROR("timed out waiting for IDLE Indicator\n");
-
-	I915_WRITE_TAIL(ring, value);
-	I915_WRITE(GEN6_BSD_SLEEP_PSMI_CONTROL,
-		GEN6_BSD_SLEEP_PSMI_CONTROL_RC_ILDL_MESSAGE_MODIFY_MASK |
-		GEN6_BSD_SLEEP_PSMI_CONTROL_RC_ILDL_MESSAGE_ENABLE);
+       I915_WRITE(GEN6_BSD_SLEEP_PSMI_CONTROL,
+	       GEN6_BSD_SLEEP_PSMI_CONTROL_RC_ILDL_MESSAGE_MODIFY_MASK |
+	       GEN6_BSD_SLEEP_PSMI_CONTROL_RC_ILDL_MESSAGE_DISABLE);
+       I915_WRITE(GEN6_BSD_RNCID, 0x0);
+
+       if (wait_for((I915_READ(GEN6_BSD_SLEEP_PSMI_CONTROL) &
+                               GEN6_BSD_SLEEP_PSMI_CONTROL_IDLE_INDICATOR) == 0,
+                       50))
+               DRM_ERROR("timed out waiting for IDLE Indicator\n");
+
+       I915_WRITE_TAIL(ring, value);
+       I915_WRITE(GEN6_BSD_SLEEP_PSMI_CONTROL,
+	       GEN6_BSD_SLEEP_PSMI_CONTROL_RC_ILDL_MESSAGE_MODIFY_MASK |
+	       GEN6_BSD_SLEEP_PSMI_CONTROL_RC_ILDL_MESSAGE_ENABLE);
 }
 
 static int gen6_ring_flush(struct intel_ring_buffer *ring,
@@ -1350,18 +1091,18 @@ static int
 gen6_ring_dispatch_execbuffer(struct intel_ring_buffer *ring,
 			      u32 offset, u32 len)
 {
-	int ret;
+       int ret;
 
-	ret = intel_ring_begin(ring, 2);
-	if (ret)
-		return ret;
+       ret = intel_ring_begin(ring, 2);
+       if (ret)
+	       return ret;
 
-	intel_ring_emit(ring, MI_BATCH_BUFFER_START | MI_BATCH_NON_SECURE_I965);
-	/* bit0-7 is the length on GEN6+ */
-	intel_ring_emit(ring, offset);
-	intel_ring_advance(ring);
+       intel_ring_emit(ring, MI_BATCH_BUFFER_START | MI_BATCH_NON_SECURE_I965);
+       /* bit0-7 is the length on GEN6+ */
+       intel_ring_emit(ring, offset);
+       intel_ring_advance(ring);
 
-	return 0;
+       return 0;
 }
 
 static bool
@@ -1399,22 +1140,17 @@ gen6_bsd_ring_put_irq(struct intel_ring_buffer *ring)
 /* ring buffer for Video Codec for Gen6+ */
 static const struct intel_ring_buffer gen6_bsd_ring = {
 	.name			= "gen6 bsd ring",
-	.id			= VCS,
+	.id			= RING_BSD,
 	.mmio_base		= GEN6_BSD_RING_BASE,
 	.size			= 32 * PAGE_SIZE,
 	.init			= init_ring_common,
 	.write_tail		= gen6_bsd_ring_write_tail,
 	.flush			= gen6_ring_flush,
 	.add_request		= gen6_add_request,
-	.get_seqno		= gen6_ring_get_seqno,
+	.get_seqno		= ring_get_seqno,
 	.irq_get		= gen6_bsd_ring_get_irq,
 	.irq_put		= gen6_bsd_ring_put_irq,
 	.dispatch_execbuffer	= gen6_ring_dispatch_execbuffer,
-	.sync_to		= gen6_bsd_ring_sync_to,
-	.semaphore_register	= {MI_SEMAPHORE_SYNC_VR,
-				   MI_SEMAPHORE_SYNC_INVALID,
-				   MI_SEMAPHORE_SYNC_VB},
-	.signal_mbox		= {GEN6_RVSYNC, GEN6_BVSYNC},
 };
 
 /* Blitter support (SandyBridge+) */
@@ -1435,13 +1171,79 @@ blt_ring_put_irq(struct intel_ring_buffer *ring)
 			  GEN6_BLITTER_USER_INTERRUPT);
 }
 
+
+/* Workaround for some stepping of SNB,
+ * each time when BLT engine ring tail moved,
+ * the first command in the ring to be parsed
+ * should be MI_BATCH_BUFFER_START
+ */
+#define NEED_BLT_WORKAROUND(dev) \
+	(IS_GEN6(dev) && (dev->pdev->revision < 8))
+
+static inline struct drm_i915_gem_object *
+to_blt_workaround(struct intel_ring_buffer *ring)
+{
+	return ring->private;
+}
+
+static int blt_ring_init(struct intel_ring_buffer *ring)
+{
+	if (NEED_BLT_WORKAROUND(ring->dev)) {
+		struct drm_i915_gem_object *obj;
+		u32 *ptr;
+		int ret;
+
+		obj = i915_gem_alloc_object(ring->dev, 4096);
+		if (obj == NULL)
+			return -ENOMEM;
+
+		ret = i915_gem_object_pin(obj, 4096, true);
+		if (ret) {
+			drm_gem_object_unreference(&obj->base);
+			return ret;
+		}
+
+		ptr = kmap(obj->pages[0]);
+		*ptr++ = MI_BATCH_BUFFER_END;
+		*ptr++ = MI_NOOP;
+		kunmap(obj->pages[0]);
+
+		ret = i915_gem_object_set_to_gtt_domain(obj, false);
+		if (ret) {
+			i915_gem_object_unpin(obj);
+			drm_gem_object_unreference(&obj->base);
+			return ret;
+		}
+
+		ring->private = obj;
+	}
+
+	return init_ring_common(ring);
+}
+
+static int blt_ring_begin(struct intel_ring_buffer *ring,
+			  int num_dwords)
+{
+	if (ring->private) {
+		int ret = intel_ring_begin(ring, num_dwords+2);
+		if (ret)
+			return ret;
+
+		intel_ring_emit(ring, MI_BATCH_BUFFER_START);
+		intel_ring_emit(ring, to_blt_workaround(ring)->gtt_offset);
+
+		return 0;
+	} else
+		return intel_ring_begin(ring, 4);
+}
+
 static int blt_ring_flush(struct intel_ring_buffer *ring,
 			  u32 invalidate, u32 flush)
 {
 	uint32_t cmd;
 	int ret;
 
-	ret = intel_ring_begin(ring, 4);
+	ret = blt_ring_begin(ring, 4);
 	if (ret)
 		return ret;
 
@@ -1456,24 +1258,30 @@ static int blt_ring_flush(struct intel_ring_buffer *ring,
 	return 0;
 }
 
+static void blt_ring_cleanup(struct intel_ring_buffer *ring)
+{
+	if (!ring->private)
+		return;
+
+	i915_gem_object_unpin(ring->private);
+	drm_gem_object_unreference(ring->private);
+	ring->private = NULL;
+}
+
 static const struct intel_ring_buffer gen6_blt_ring = {
-	.name			= "blt ring",
-	.id			= BCS,
-	.mmio_base		= BLT_RING_BASE,
-	.size			= 32 * PAGE_SIZE,
-	.init			= init_ring_common,
-	.write_tail		= ring_write_tail,
-	.flush			= blt_ring_flush,
-	.add_request		= gen6_add_request,
-	.get_seqno		= gen6_ring_get_seqno,
-	.irq_get		= blt_ring_get_irq,
-	.irq_put		= blt_ring_put_irq,
-	.dispatch_execbuffer	= gen6_ring_dispatch_execbuffer,
-	.sync_to		= gen6_blt_ring_sync_to,
-	.semaphore_register	= {MI_SEMAPHORE_SYNC_BR,
-				   MI_SEMAPHORE_SYNC_BV,
-				   MI_SEMAPHORE_SYNC_INVALID},
-	.signal_mbox		= {GEN6_RBSYNC, GEN6_VBSYNC},
+       .name			= "blt ring",
+       .id			= RING_BLT,
+       .mmio_base		= BLT_RING_BASE,
+       .size			= 32 * PAGE_SIZE,
+       .init			= blt_ring_init,
+       .write_tail		= ring_write_tail,
+       .flush			= blt_ring_flush,
+       .add_request		= gen6_add_request,
+       .get_seqno		= ring_get_seqno,
+       .irq_get			= blt_ring_get_irq,
+       .irq_put			= blt_ring_put_irq,
+       .dispatch_execbuffer	= gen6_ring_dispatch_execbuffer,
+       .cleanup			= blt_ring_cleanup,
 };
 
 int intel_init_render_ring_buffer(struct drm_device *dev)
@@ -1484,10 +1292,8 @@ int intel_init_render_ring_buffer(struct drm_device *dev)
 	*ring = render_ring;
 	if (INTEL_INFO(dev)->gen >= 6) {
 		ring->add_request = gen6_add_request;
-		ring->flush = gen6_render_ring_flush;
 		ring->irq_get = gen6_render_ring_get_irq;
 		ring->irq_put = gen6_render_ring_put_irq;
-		ring->get_seqno = gen6_ring_get_seqno;
 	} else if (IS_GEN5(dev)) {
 		ring->add_request = pc_render_add_request;
 		ring->get_seqno = pc_render_get_seqno;
diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h
index bc0365b..39ac2b6 100644
--- a/drivers/gpu/drm/i915/intel_ringbuffer.h
+++ b/drivers/gpu/drm/i915/intel_ringbuffer.h
@@ -1,6 +1,13 @@
 #ifndef _INTEL_RINGBUFFER_H_
 #define _INTEL_RINGBUFFER_H_
 
+enum {
+    RCS = 0x0,
+    VCS,
+    BCS,
+    I915_NUM_RINGS,
+};
+
 struct  intel_hw_status_page {
 	u32	__iomem	*page_addr;
 	unsigned int	gfx_addr;
@@ -29,11 +36,10 @@ struct  intel_hw_status_page {
 struct  intel_ring_buffer {
 	const char	*name;
 	enum intel_ring_id {
-		RCS = 0x0,
-		VCS,
-		BCS,
+		RING_RENDER = 0x1,
+		RING_BSD = 0x2,
+		RING_BLT = 0x4,
 	} id;
-#define I915_NUM_RINGS 3
 	u32		mmio_base;
 	void		__iomem *virtual_start;
 	struct		drm_device *dev;
@@ -46,16 +52,6 @@ struct  intel_ring_buffer {
 	int		effective_size;
 	struct intel_hw_status_page status_page;
 
-	/** We track the position of the requests in the ring buffer, and
-	 * when each is retired we increment last_retired_head as the GPU
-	 * must have finished processing the request and so we know we
-	 * can advance the ringbuffer up to that position.
-	 *
-	 * last_retired_head is set to -1 after the value is consumed so
-	 * we can detect new retirements.
-	 */
-	u32		last_retired_head;
-
 	spinlock_t	irq_lock;
 	u32		irq_refcount;
 	u32		irq_mask;
@@ -79,12 +75,7 @@ struct  intel_ring_buffer {
 	int		(*dispatch_execbuffer)(struct intel_ring_buffer *ring,
 					       u32 offset, u32 length);
 	void		(*cleanup)(struct intel_ring_buffer *ring);
-	int		(*sync_to)(struct intel_ring_buffer *ring,
-				   struct intel_ring_buffer *to,
-				   u32 seqno);
 
-	u32		semaphore_register[3]; /*our mbox written by others */
-	u32		signal_mbox[2]; /* mboxes this ring signals to */
 	/**
 	 * List of objects currently involved in rendering from the
 	 * ringbuffer.
@@ -123,12 +114,6 @@ struct  intel_ring_buffer {
 	void *private;
 };
 
-static inline unsigned
-intel_ring_flag(struct intel_ring_buffer *ring)
-{
-	return 1 << ring->id;
-}
-
 static inline u32
 intel_ring_sync_index(struct intel_ring_buffer *ring,
 		      struct intel_ring_buffer *other)
@@ -195,6 +180,9 @@ static inline void intel_ring_emit(struct intel_ring_buffer *ring,
 void intel_ring_advance(struct intel_ring_buffer *ring);
 
 u32 intel_ring_get_seqno(struct intel_ring_buffer *ring);
+int intel_ring_sync(struct intel_ring_buffer *ring,
+		    struct intel_ring_buffer *to,
+		    u32 seqno);
 
 int intel_init_render_ring_buffer(struct drm_device *dev);
 int intel_init_bsd_ring_buffer(struct drm_device *dev);
@@ -203,11 +191,6 @@ int intel_init_blt_ring_buffer(struct drm_device *dev);
 u32 intel_ring_get_active_head(struct intel_ring_buffer *ring);
 void intel_ring_setup_status_page(struct intel_ring_buffer *ring);
 
-static inline u32 intel_ring_get_tail(struct intel_ring_buffer *ring)
-{
-	return ring->tail;
-}
-
 static inline void i915_trace_irq_get(struct intel_ring_buffer *ring, u32 seqno)
 {
 	if (ring->trace_irq_seqno == 0 && ring->irq_get(ring))
diff --git a/drivers/gpu/drm/i915/intel_sdvo.c b/drivers/gpu/drm/i915/intel_sdvo.c
index 0de20ee..06b51ff 100644
--- a/drivers/gpu/drm/i915/intel_sdvo.c
+++ b/drivers/gpu/drm/i915/intel_sdvo.c
@@ -43,13 +43,12 @@
 #define SDVO_TV_MASK   (SDVO_OUTPUT_CVBS0 | SDVO_OUTPUT_SVID0)
 
 #define SDVO_OUTPUT_MASK (SDVO_TMDS_MASK | SDVO_RGB_MASK | SDVO_LVDS_MASK |\
-			SDVO_TV_MASK)
+                         SDVO_TV_MASK)
 
 #define IS_TV(c)	(c->output_flag & SDVO_TV_MASK)
 #define IS_TMDS(c)	(c->output_flag & SDVO_TMDS_MASK)
 #define IS_LVDS(c)	(c->output_flag & SDVO_LVDS_MASK)
 #define IS_TV_OR_LVDS(c) (c->output_flag & (SDVO_TV_MASK | SDVO_LVDS_MASK))
-#define IS_DIGITAL(c) (c->output_flag & (SDVO_TMDS_MASK | SDVO_LVDS_MASK))
 
 
 static const char *tv_format_names[] = {
@@ -93,11 +92,6 @@ struct intel_sdvo {
 	*/
 	uint16_t attached_output;
 
-	/*
-	 * Hotplug activation bits for this device
-	 */
-	uint8_t hotplug_active[2];
-
 	/**
 	 * This is used to select the color range of RBG outputs in HDMI mode.
 	 * It is only valid when using TMDS encoding and 8 bit per color mode.
@@ -147,7 +141,7 @@ struct intel_sdvo_connector {
 	/* Mark the type of connector */
 	uint16_t output_flag;
 
-	enum hdmi_force_audio force_audio;
+	int force_audio;
 
 	/* This contains all current supported TV format */
 	u8 tv_format_supported[TV_FORMAT_NUM];
@@ -289,117 +283,117 @@ static const struct _sdvo_cmd_name {
 	u8 cmd;
 	const char *name;
 } sdvo_cmd_names[] = {
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_RESET),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_DEVICE_CAPS),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_FIRMWARE_REV),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_TRAINED_INPUTS),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_ACTIVE_OUTPUTS),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_ACTIVE_OUTPUTS),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_IN_OUT_MAP),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_IN_OUT_MAP),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_ATTACHED_DISPLAYS),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HOT_PLUG_SUPPORT),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_ACTIVE_HOT_PLUG),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_ACTIVE_HOT_PLUG),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_INTERRUPT_EVENT_SOURCE),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_TARGET_INPUT),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_TARGET_OUTPUT),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_INPUT_TIMINGS_PART1),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_INPUT_TIMINGS_PART2),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_INPUT_TIMINGS_PART1),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_INPUT_TIMINGS_PART2),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_INPUT_TIMINGS_PART1),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_OUTPUT_TIMINGS_PART1),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_OUTPUT_TIMINGS_PART2),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_OUTPUT_TIMINGS_PART1),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_OUTPUT_TIMINGS_PART2),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_CREATE_PREFERRED_INPUT_TIMING),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_PREFERRED_INPUT_TIMING_PART1),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_PREFERRED_INPUT_TIMING_PART2),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_INPUT_PIXEL_CLOCK_RANGE),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_OUTPUT_PIXEL_CLOCK_RANGE),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SUPPORTED_CLOCK_RATE_MULTS),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_CLOCK_RATE_MULT),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_CLOCK_RATE_MULT),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SUPPORTED_TV_FORMATS),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_TV_FORMAT),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_TV_FORMAT),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SUPPORTED_POWER_STATES),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_POWER_STATE),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_ENCODER_POWER_STATE),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_DISPLAY_POWER_STATE),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_CONTROL_BUS_SWITCH),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SDTV_RESOLUTION_SUPPORT),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SCALED_HDTV_RESOLUTION_SUPPORT),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SUPPORTED_ENHANCEMENTS),
-
-	/* Add the op code for SDVO enhancements */
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_HPOS),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HPOS),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_HPOS),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_VPOS),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_VPOS),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_VPOS),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_SATURATION),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SATURATION),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_SATURATION),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_HUE),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HUE),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_HUE),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_CONTRAST),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_CONTRAST),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_CONTRAST),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_BRIGHTNESS),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_BRIGHTNESS),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_BRIGHTNESS),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_OVERSCAN_H),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_OVERSCAN_H),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_OVERSCAN_H),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_OVERSCAN_V),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_OVERSCAN_V),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_OVERSCAN_V),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_FLICKER_FILTER),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_FLICKER_FILTER),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_FLICKER_FILTER),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_FLICKER_FILTER_ADAPTIVE),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_FLICKER_FILTER_ADAPTIVE),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_FLICKER_FILTER_ADAPTIVE),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_FLICKER_FILTER_2D),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_FLICKER_FILTER_2D),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_FLICKER_FILTER_2D),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_SHARPNESS),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SHARPNESS),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_SHARPNESS),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_DOT_CRAWL),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_DOT_CRAWL),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_TV_CHROMA_FILTER),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_TV_CHROMA_FILTER),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_TV_CHROMA_FILTER),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_TV_LUMA_FILTER),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_TV_LUMA_FILTER),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_TV_LUMA_FILTER),
-
-	/* HDMI op code */
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SUPP_ENCODE),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_ENCODE),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_ENCODE),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_PIXEL_REPLI),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_PIXEL_REPLI),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_COLORIMETRY_CAP),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_COLORIMETRY),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_COLORIMETRY),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_AUDIO_ENCRYPT_PREFER),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_AUDIO_STAT),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_AUDIO_STAT),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HBUF_INDEX),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_HBUF_INDEX),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HBUF_INFO),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HBUF_AV_SPLIT),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_HBUF_AV_SPLIT),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HBUF_TXRATE),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_HBUF_TXRATE),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_HBUF_DATA),
-	SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HBUF_DATA),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_RESET),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_DEVICE_CAPS),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_FIRMWARE_REV),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_TRAINED_INPUTS),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_ACTIVE_OUTPUTS),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_ACTIVE_OUTPUTS),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_IN_OUT_MAP),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_IN_OUT_MAP),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_ATTACHED_DISPLAYS),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HOT_PLUG_SUPPORT),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_ACTIVE_HOT_PLUG),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_ACTIVE_HOT_PLUG),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_INTERRUPT_EVENT_SOURCE),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_TARGET_INPUT),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_TARGET_OUTPUT),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_INPUT_TIMINGS_PART1),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_INPUT_TIMINGS_PART2),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_INPUT_TIMINGS_PART1),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_INPUT_TIMINGS_PART2),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_INPUT_TIMINGS_PART1),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_OUTPUT_TIMINGS_PART1),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_OUTPUT_TIMINGS_PART2),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_OUTPUT_TIMINGS_PART1),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_OUTPUT_TIMINGS_PART2),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_CREATE_PREFERRED_INPUT_TIMING),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_PREFERRED_INPUT_TIMING_PART1),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_PREFERRED_INPUT_TIMING_PART2),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_INPUT_PIXEL_CLOCK_RANGE),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_OUTPUT_PIXEL_CLOCK_RANGE),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SUPPORTED_CLOCK_RATE_MULTS),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_CLOCK_RATE_MULT),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_CLOCK_RATE_MULT),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SUPPORTED_TV_FORMATS),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_TV_FORMAT),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_TV_FORMAT),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SUPPORTED_POWER_STATES),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_POWER_STATE),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_ENCODER_POWER_STATE),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_DISPLAY_POWER_STATE),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_CONTROL_BUS_SWITCH),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SDTV_RESOLUTION_SUPPORT),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SCALED_HDTV_RESOLUTION_SUPPORT),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SUPPORTED_ENHANCEMENTS),
+
+    /* Add the op code for SDVO enhancements */
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_HPOS),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HPOS),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_HPOS),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_VPOS),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_VPOS),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_VPOS),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_SATURATION),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SATURATION),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_SATURATION),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_HUE),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HUE),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_HUE),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_CONTRAST),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_CONTRAST),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_CONTRAST),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_BRIGHTNESS),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_BRIGHTNESS),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_BRIGHTNESS),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_OVERSCAN_H),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_OVERSCAN_H),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_OVERSCAN_H),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_OVERSCAN_V),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_OVERSCAN_V),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_OVERSCAN_V),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_FLICKER_FILTER),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_FLICKER_FILTER),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_FLICKER_FILTER),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_FLICKER_FILTER_ADAPTIVE),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_FLICKER_FILTER_ADAPTIVE),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_FLICKER_FILTER_ADAPTIVE),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_FLICKER_FILTER_2D),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_FLICKER_FILTER_2D),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_FLICKER_FILTER_2D),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_SHARPNESS),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SHARPNESS),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_SHARPNESS),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_DOT_CRAWL),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_DOT_CRAWL),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_TV_CHROMA_FILTER),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_TV_CHROMA_FILTER),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_TV_CHROMA_FILTER),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_MAX_TV_LUMA_FILTER),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_TV_LUMA_FILTER),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_TV_LUMA_FILTER),
+
+    /* HDMI op code */
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_SUPP_ENCODE),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_ENCODE),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_ENCODE),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_PIXEL_REPLI),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_PIXEL_REPLI),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_COLORIMETRY_CAP),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_COLORIMETRY),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_COLORIMETRY),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_AUDIO_ENCRYPT_PREFER),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_AUDIO_STAT),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_AUDIO_STAT),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HBUF_INDEX),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_HBUF_INDEX),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HBUF_INFO),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HBUF_AV_SPLIT),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_HBUF_AV_SPLIT),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HBUF_TXRATE),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_HBUF_TXRATE),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_SET_HBUF_DATA),
+    SDVO_CMD_NAME_ENTRY(SDVO_CMD_GET_HBUF_DATA),
 };
 
 #define IS_SDVOB(reg)	(reg == SDVOB || reg == PCH_SDVOB)
@@ -981,6 +975,7 @@ intel_sdvo_set_input_timings_for_mode(struct intel_sdvo *intel_sdvo,
 
 	intel_sdvo_get_mode_from_dtd(adjusted_mode, &intel_sdvo->input_dtd);
 
+	drm_mode_set_crtcinfo(adjusted_mode, 0);
 	return true;
 }
 
@@ -1118,12 +1113,8 @@ static void intel_sdvo_mode_set(struct drm_encoder *encoder,
 		}
 		sdvox |= (9 << 19) | SDVO_BORDER_ENABLE;
 	}
-
-	if (INTEL_PCH_TYPE(dev) >= PCH_CPT)
-		sdvox |= TRANSCODER_CPT(intel_crtc->pipe);
-	else
-		sdvox |= TRANSCODER(intel_crtc->pipe);
-
+	if (intel_crtc->pipe == 1)
+		sdvox |= SDVO_PIPE_B_SELECT;
 	if (intel_sdvo->has_hdmi_audio)
 		sdvox |= SDVO_AUDIO_ENABLE;
 
@@ -1250,26 +1241,81 @@ static bool intel_sdvo_get_capabilities(struct intel_sdvo *intel_sdvo, struct in
 	return true;
 }
 
-static int intel_sdvo_supports_hotplug(struct intel_sdvo *intel_sdvo)
+/* No use! */
+#if 0
+struct drm_connector* intel_sdvo_find(struct drm_device *dev, int sdvoB)
+{
+	struct drm_connector *connector = NULL;
+	struct intel_sdvo *iout = NULL;
+	struct intel_sdvo *sdvo;
+
+	/* find the sdvo connector */
+	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
+		iout = to_intel_sdvo(connector);
+
+		if (iout->type != INTEL_OUTPUT_SDVO)
+			continue;
+
+		sdvo = iout->dev_priv;
+
+		if (sdvo->sdvo_reg == SDVOB && sdvoB)
+			return connector;
+
+		if (sdvo->sdvo_reg == SDVOC && !sdvoB)
+			return connector;
+
+	}
+
+	return NULL;
+}
+
+int intel_sdvo_supports_hotplug(struct drm_connector *connector)
 {
 	u8 response[2];
+	u8 status;
+	struct intel_sdvo *intel_sdvo;
+	DRM_DEBUG_KMS("\n");
+
+	if (!connector)
+		return 0;
+
+	intel_sdvo = to_intel_sdvo(connector);
 
 	return intel_sdvo_get_value(intel_sdvo, SDVO_CMD_GET_HOT_PLUG_SUPPORT,
 				    &response, 2) && response[0];
 }
 
-static void intel_sdvo_enable_hotplug(struct intel_encoder *encoder)
+void intel_sdvo_set_hotplug(struct drm_connector *connector, int on)
 {
-	struct intel_sdvo *intel_sdvo = to_intel_sdvo(&encoder->base);
+	u8 response[2];
+	u8 status;
+	struct intel_sdvo *intel_sdvo = to_intel_sdvo(connector);
 
-	intel_sdvo_write_cmd(intel_sdvo, SDVO_CMD_SET_ACTIVE_HOT_PLUG, &intel_sdvo->hotplug_active, 2);
+	intel_sdvo_write_cmd(intel_sdvo, SDVO_CMD_GET_ACTIVE_HOT_PLUG, NULL, 0);
+	intel_sdvo_read_response(intel_sdvo, &response, 2);
+
+	if (on) {
+		intel_sdvo_write_cmd(intel_sdvo, SDVO_CMD_GET_HOT_PLUG_SUPPORT, NULL, 0);
+		status = intel_sdvo_read_response(intel_sdvo, &response, 2);
+
+		intel_sdvo_write_cmd(intel_sdvo, SDVO_CMD_SET_ACTIVE_HOT_PLUG, &response, 2);
+	} else {
+		response[0] = 0;
+		response[1] = 0;
+		intel_sdvo_write_cmd(intel_sdvo, SDVO_CMD_SET_ACTIVE_HOT_PLUG, &response, 2);
+	}
+
+	intel_sdvo_write_cmd(intel_sdvo, SDVO_CMD_GET_ACTIVE_HOT_PLUG, NULL, 0);
+	intel_sdvo_read_response(intel_sdvo, &response, 2);
 }
+#endif
 
 static bool
 intel_sdvo_multifunc_encoder(struct intel_sdvo *intel_sdvo)
 {
 	/* Is there more than one type of output? */
-	return hweight16(intel_sdvo->caps.output_flags) > 1;
+	int caps = intel_sdvo->caps.output_flags & 0xf;
+	return caps & -caps;
 }
 
 static struct edid *
@@ -1290,7 +1336,7 @@ intel_sdvo_get_analog_edid(struct drm_connector *connector)
 }
 
 enum drm_connector_status
-intel_sdvo_tmds_sink_detect(struct drm_connector *connector)
+intel_sdvo_hdmi_sink_detect(struct drm_connector *connector)
 {
 	struct intel_sdvo *intel_sdvo = intel_attached_sdvo(connector);
 	enum drm_connector_status status;
@@ -1343,25 +1389,13 @@ intel_sdvo_tmds_sink_detect(struct drm_connector *connector)
 
 	if (status == connector_status_connected) {
 		struct intel_sdvo_connector *intel_sdvo_connector = to_intel_sdvo_connector(connector);
-		if (intel_sdvo_connector->force_audio != HDMI_AUDIO_AUTO)
-			intel_sdvo->has_hdmi_audio = (intel_sdvo_connector->force_audio == HDMI_AUDIO_ON);
+		if (intel_sdvo_connector->force_audio)
+			intel_sdvo->has_hdmi_audio = intel_sdvo_connector->force_audio > 0;
 	}
 
 	return status;
 }
 
-static bool
-intel_sdvo_connector_matches_edid(struct intel_sdvo_connector *sdvo,
-				  struct edid *edid)
-{
-	bool monitor_is_digital = !!(edid->input & DRM_EDID_INPUT_DIGITAL);
-	bool connector_is_digital = !!IS_DIGITAL(sdvo);
-
-	DRM_DEBUG_KMS("connector_is_digital? %d, monitor_is_digital? %d\n",
-		      connector_is_digital, monitor_is_digital);
-	return connector_is_digital == monitor_is_digital;
-}
-
 static enum drm_connector_status
 intel_sdvo_detect(struct drm_connector *connector, bool force)
 {
@@ -1397,7 +1431,7 @@ intel_sdvo_detect(struct drm_connector *connector, bool force)
 	if ((intel_sdvo_connector->output_flag & response) == 0)
 		ret = connector_status_disconnected;
 	else if (IS_TMDS(intel_sdvo_connector))
-		ret = intel_sdvo_tmds_sink_detect(connector);
+		ret = intel_sdvo_hdmi_sink_detect(connector);
 	else {
 		struct edid *edid;
 
@@ -1406,12 +1440,10 @@ intel_sdvo_detect(struct drm_connector *connector, bool force)
 		if (edid == NULL)
 			edid = intel_sdvo_get_analog_edid(connector);
 		if (edid != NULL) {
-			if (intel_sdvo_connector_matches_edid(intel_sdvo_connector,
-							      edid))
-				ret = connector_status_connected;
-			else
+			if (edid->input & DRM_EDID_INPUT_DIGITAL)
 				ret = connector_status_disconnected;
-
+			else
+				ret = connector_status_connected;
 			connector->display_info.raw_edid = NULL;
 			kfree(edid);
 		} else
@@ -1452,8 +1484,11 @@ static void intel_sdvo_get_ddc_modes(struct drm_connector *connector)
 		edid = intel_sdvo_get_analog_edid(connector);
 
 	if (edid != NULL) {
-		if (intel_sdvo_connector_matches_edid(to_intel_sdvo_connector(connector),
-						      edid)) {
+		struct intel_sdvo_connector *intel_sdvo_connector = to_intel_sdvo_connector(connector);
+		bool monitor_is_digital = !!(edid->input & DRM_EDID_INPUT_DIGITAL);
+		bool connector_is_digital = !!IS_TMDS(intel_sdvo_connector);
+
+		if (connector_is_digital == monitor_is_digital) {
 			drm_mode_connector_update_edid_property(connector, edid);
 			drm_add_edid_modes(connector, edid);
 		}
@@ -1719,10 +1754,10 @@ intel_sdvo_set_property(struct drm_connector *connector,
 
 		intel_sdvo_connector->force_audio = i;
 
-		if (i == HDMI_AUDIO_AUTO)
+		if (i == 0)
 			has_audio = intel_sdvo_detect_hdmi_audio(connector);
 		else
-			has_audio = (i == HDMI_AUDIO_ON);
+			has_audio = i > 0;
 
 		if (has_audio == intel_sdvo->has_hdmi_audio)
 			return 0;
@@ -1945,7 +1980,7 @@ intel_sdvo_select_i2c_bus(struct drm_i915_private *dev_priv,
 			  struct intel_sdvo *sdvo, u32 reg)
 {
 	struct sdvo_device_mapping *mapping;
-	u8 pin;
+	u8 pin, speed;
 
 	if (IS_SDVOB(reg))
 		mapping = &dev_priv->sdvo_mappings[0];
@@ -1953,16 +1988,18 @@ intel_sdvo_select_i2c_bus(struct drm_i915_private *dev_priv,
 		mapping = &dev_priv->sdvo_mappings[1];
 
 	pin = GMBUS_PORT_DPB;
-	if (mapping->initialized)
+	speed = GMBUS_RATE_1MHZ >> 8;
+	if (mapping->initialized) {
 		pin = mapping->i2c_pin;
+		speed = mapping->i2c_speed;
+	}
 
 	if (pin < GMBUS_NUM_PORTS) {
 		sdvo->i2c = &dev_priv->gmbus[pin].adapter;
-		intel_gmbus_set_speed(sdvo->i2c, GMBUS_RATE_1MHZ);
+		intel_gmbus_set_speed(sdvo->i2c, speed);
 		intel_gmbus_force_bit(sdvo->i2c, true);
-	} else {
+	} else
 		sdvo->i2c = &dev_priv->gmbus[GMBUS_PORT_DPB].adapter;
-	}
 }
 
 static bool
@@ -2020,7 +2057,7 @@ intel_sdvo_connector_init(struct intel_sdvo_connector *connector,
 	drm_connector_helper_add(&connector->base.base,
 				 &intel_sdvo_connector_helper_funcs);
 
-	connector->base.base.interlace_allowed = 1;
+	connector->base.base.interlace_allowed = 0;
 	connector->base.base.doublescan_allowed = 0;
 	connector->base.base.display_info.subpixel_order = SubPixelHorizontalRGB;
 
@@ -2043,7 +2080,6 @@ intel_sdvo_dvi_init(struct intel_sdvo *intel_sdvo, int device)
 {
 	struct drm_encoder *encoder = &intel_sdvo->base.base;
 	struct drm_connector *connector;
-	struct intel_encoder *intel_encoder = to_intel_encoder(encoder);
 	struct intel_connector *intel_connector;
 	struct intel_sdvo_connector *intel_sdvo_connector;
 
@@ -2061,17 +2097,7 @@ intel_sdvo_dvi_init(struct intel_sdvo *intel_sdvo, int device)
 
 	intel_connector = &intel_sdvo_connector->base;
 	connector = &intel_connector->base;
-	if (intel_sdvo_supports_hotplug(intel_sdvo) & (1 << device)) {
-		connector->polled = DRM_CONNECTOR_POLL_HPD;
-		intel_sdvo->hotplug_active[0] |= 1 << device;
-		/* Some SDVO devices have one-shot hotplug interrupts.
-		 * Ensure that they get re-enabled when an interrupt happens.
-		 */
-		intel_encoder->hot_plug = intel_sdvo_enable_hotplug;
-		intel_sdvo_enable_hotplug(intel_encoder);
-	}
-	else
-		connector->polled = DRM_CONNECTOR_POLL_CONNECT | DRM_CONNECTOR_POLL_DISCONNECT;
+	connector->polled = DRM_CONNECTOR_POLL_CONNECT | DRM_CONNECTOR_POLL_DISCONNECT;
 	encoder->encoder_type = DRM_MODE_ENCODER_TMDS;
 	connector->connector_type = DRM_MODE_CONNECTOR_DVID;
 
@@ -2253,7 +2279,7 @@ intel_sdvo_output_setup(struct intel_sdvo *intel_sdvo, uint16_t flags)
 			      bytes[0], bytes[1]);
 		return false;
 	}
-	intel_sdvo->base.crtc_mask = (1 << 0) | (1 << 1) | (1 << 2);
+	intel_sdvo->base.crtc_mask = (1 << 0) | (1 << 1);
 
 	return true;
 }
@@ -2312,15 +2338,17 @@ static bool intel_sdvo_tv_create_property(struct intel_sdvo *intel_sdvo,
 		intel_sdvo_connector->max_##name = data_value[0]; \
 		intel_sdvo_connector->cur_##name = response; \
 		intel_sdvo_connector->name = \
-			drm_property_create_range(dev, 0, #name, 0, data_value[0]); \
+			drm_property_create(dev, DRM_MODE_PROP_RANGE, #name, 2); \
 		if (!intel_sdvo_connector->name) return false; \
+		intel_sdvo_connector->name->values[0] = 0; \
+		intel_sdvo_connector->name->values[1] = data_value[0]; \
 		drm_connector_attach_property(connector, \
 					      intel_sdvo_connector->name, \
 					      intel_sdvo_connector->cur_##name); \
 		DRM_DEBUG_KMS(#name ": max %d, default %d, current %d\n", \
 			      data_value[0], data_value[1], response); \
 	} \
-} while (0)
+} while(0)
 
 static bool
 intel_sdvo_create_enhance_property_tv(struct intel_sdvo *intel_sdvo,
@@ -2347,19 +2375,25 @@ intel_sdvo_create_enhance_property_tv(struct intel_sdvo *intel_sdvo,
 		intel_sdvo_connector->left_margin = data_value[0] - response;
 		intel_sdvo_connector->right_margin = intel_sdvo_connector->left_margin;
 		intel_sdvo_connector->left =
-			drm_property_create_range(dev, 0, "left_margin", 0, data_value[0]);
+			drm_property_create(dev, DRM_MODE_PROP_RANGE,
+					    "left_margin", 2);
 		if (!intel_sdvo_connector->left)
 			return false;
 
+		intel_sdvo_connector->left->values[0] = 0;
+		intel_sdvo_connector->left->values[1] = data_value[0];
 		drm_connector_attach_property(connector,
 					      intel_sdvo_connector->left,
 					      intel_sdvo_connector->left_margin);
 
 		intel_sdvo_connector->right =
-			drm_property_create_range(dev, 0, "right_margin", 0, data_value[0]);
+			drm_property_create(dev, DRM_MODE_PROP_RANGE,
+					    "right_margin", 2);
 		if (!intel_sdvo_connector->right)
 			return false;
 
+		intel_sdvo_connector->right->values[0] = 0;
+		intel_sdvo_connector->right->values[1] = data_value[0];
 		drm_connector_attach_property(connector,
 					      intel_sdvo_connector->right,
 					      intel_sdvo_connector->right_margin);
@@ -2383,21 +2417,25 @@ intel_sdvo_create_enhance_property_tv(struct intel_sdvo *intel_sdvo,
 		intel_sdvo_connector->top_margin = data_value[0] - response;
 		intel_sdvo_connector->bottom_margin = intel_sdvo_connector->top_margin;
 		intel_sdvo_connector->top =
-			drm_property_create_range(dev, 0,
-					    "top_margin", 0, data_value[0]);
+			drm_property_create(dev, DRM_MODE_PROP_RANGE,
+					    "top_margin", 2);
 		if (!intel_sdvo_connector->top)
 			return false;
 
+		intel_sdvo_connector->top->values[0] = 0;
+		intel_sdvo_connector->top->values[1] = data_value[0];
 		drm_connector_attach_property(connector,
 					      intel_sdvo_connector->top,
 					      intel_sdvo_connector->top_margin);
 
 		intel_sdvo_connector->bottom =
-			drm_property_create_range(dev, 0,
-					    "bottom_margin", 0, data_value[0]);
+			drm_property_create(dev, DRM_MODE_PROP_RANGE,
+					    "bottom_margin", 2);
 		if (!intel_sdvo_connector->bottom)
 			return false;
 
+		intel_sdvo_connector->bottom->values[0] = 0;
+		intel_sdvo_connector->bottom->values[1] = data_value[0];
 		drm_connector_attach_property(connector,
 					      intel_sdvo_connector->bottom,
 					      intel_sdvo_connector->bottom_margin);
@@ -2426,10 +2464,12 @@ intel_sdvo_create_enhance_property_tv(struct intel_sdvo *intel_sdvo,
 		intel_sdvo_connector->max_dot_crawl = 1;
 		intel_sdvo_connector->cur_dot_crawl = response & 0x1;
 		intel_sdvo_connector->dot_crawl =
-			drm_property_create_range(dev, 0, "dot_crawl", 0, 1);
+			drm_property_create(dev, DRM_MODE_PROP_RANGE, "dot_crawl", 2);
 		if (!intel_sdvo_connector->dot_crawl)
 			return false;
 
+		intel_sdvo_connector->dot_crawl->values[0] = 0;
+		intel_sdvo_connector->dot_crawl->values[1] = 1;
 		drm_connector_attach_property(connector,
 					      intel_sdvo_connector->dot_crawl,
 					      intel_sdvo_connector->cur_dot_crawl);
@@ -2475,7 +2515,7 @@ static bool intel_sdvo_create_enhance_property(struct intel_sdvo *intel_sdvo,
 
 	if (IS_TV(intel_sdvo_connector))
 		return intel_sdvo_create_enhance_property_tv(intel_sdvo, intel_sdvo_connector, enhancements.reply);
-	else if (IS_LVDS(intel_sdvo_connector))
+	else if(IS_LVDS(intel_sdvo_connector))
 		return intel_sdvo_create_enhance_property_lvds(intel_sdvo, intel_sdvo_connector, enhancements.reply);
 	else
 		return true;
@@ -2564,14 +2604,6 @@ bool intel_sdvo_init(struct drm_device *dev, int sdvo_reg)
 	if (!intel_sdvo_get_capabilities(intel_sdvo, &intel_sdvo->caps))
 		goto err;
 
-	/* Set up hotplug command - note paranoia about contents of reply.
-	 * We assume that the hardware is in a sane state, and only touch
-	 * the bits we think we understand.
-	 */
-	intel_sdvo_get_value(intel_sdvo, SDVO_CMD_GET_ACTIVE_HOT_PLUG,
-			     &intel_sdvo->hotplug_active, 2);
-	intel_sdvo->hotplug_active[0] &= ~0x3;
-
 	if (intel_sdvo_output_setup(intel_sdvo,
 				    intel_sdvo->caps.output_flags) != true) {
 		DRM_DEBUG_KMS("SDVO output failed to setup on SDVO%c\n",
diff --git a/drivers/gpu/drm/i915/intel_sdvo_regs.h b/drivers/gpu/drm/i915/intel_sdvo_regs.h
index 770bdd6..50bebc3 100644
--- a/drivers/gpu/drm/i915/intel_sdvo_regs.h
+++ b/drivers/gpu/drm/i915/intel_sdvo_regs.h
@@ -1,5 +1,5 @@
 /*
- * Copyright © 2006-2007 Intel Corporation
+ * Copyright © 2006-2007 Intel Corporation
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -46,19 +46,19 @@
 #define SDVO_OUTPUT_LAST    (14)
 
 struct intel_sdvo_caps {
-	u8 vendor_id;
-	u8 device_id;
-	u8 device_rev_id;
-	u8 sdvo_version_major;
-	u8 sdvo_version_minor;
-	unsigned int sdvo_inputs_mask:2;
-	unsigned int smooth_scaling:1;
-	unsigned int sharp_scaling:1;
-	unsigned int up_scaling:1;
-	unsigned int down_scaling:1;
-	unsigned int stall_support:1;
-	unsigned int pad:1;
-	u16 output_flags;
+    u8 vendor_id;
+    u8 device_id;
+    u8 device_rev_id;
+    u8 sdvo_version_major;
+    u8 sdvo_version_minor;
+    unsigned int sdvo_inputs_mask:2;
+    unsigned int smooth_scaling:1;
+    unsigned int sharp_scaling:1;
+    unsigned int up_scaling:1;
+    unsigned int down_scaling:1;
+    unsigned int stall_support:1;
+    unsigned int pad:1;
+    u16 output_flags;
 } __attribute__((packed));
 
 /* Note: SDVO detailed timing flags match EDID misc flags. */
@@ -68,46 +68,46 @@ struct intel_sdvo_caps {
 
 /** This matches the EDID DTD structure, more or less */
 struct intel_sdvo_dtd {
-	struct {
-		u16 clock;	/**< pixel clock, in 10kHz units */
-		u8 h_active;	/**< lower 8 bits (pixels) */
-		u8 h_blank;	/**< lower 8 bits (pixels) */
-		u8 h_high;	/**< upper 4 bits each h_active, h_blank */
-		u8 v_active;	/**< lower 8 bits (lines) */
-		u8 v_blank;	/**< lower 8 bits (lines) */
-		u8 v_high;	/**< upper 4 bits each v_active, v_blank */
-	} part1;
-
-	struct {
-		u8 h_sync_off;	/**< lower 8 bits, from hblank start */
-		u8 h_sync_width;	/**< lower 8 bits (pixels) */
-		/** lower 4 bits each vsync offset, vsync width */
-		u8 v_sync_off_width;
-		/**
-		* 2 high bits of hsync offset, 2 high bits of hsync width,
-		* bits 4-5 of vsync offset, and 2 high bits of vsync width.
-		*/
-		u8 sync_off_width_high;
-		u8 dtd_flags;
-		u8 sdvo_flags;
-		/** bits 6-7 of vsync offset at bits 6-7 */
-		u8 v_sync_off_high;
-		u8 reserved;
-	} part2;
+    struct {
+	u16 clock;		/**< pixel clock, in 10kHz units */
+	u8 h_active;		/**< lower 8 bits (pixels) */
+	u8 h_blank;		/**< lower 8 bits (pixels) */
+	u8 h_high;		/**< upper 4 bits each h_active, h_blank */
+	u8 v_active;		/**< lower 8 bits (lines) */
+	u8 v_blank;		/**< lower 8 bits (lines) */
+	u8 v_high;		/**< upper 4 bits each v_active, v_blank */
+    } part1;
+
+    struct {
+	u8 h_sync_off;	/**< lower 8 bits, from hblank start */
+	u8 h_sync_width;	/**< lower 8 bits (pixels) */
+	/** lower 4 bits each vsync offset, vsync width */
+	u8 v_sync_off_width;
+	/**
+	 * 2 high bits of hsync offset, 2 high bits of hsync width,
+	 * bits 4-5 of vsync offset, and 2 high bits of vsync width.
+	 */
+	u8 sync_off_width_high;
+	u8 dtd_flags;
+	u8 sdvo_flags;
+	/** bits 6-7 of vsync offset at bits 6-7 */
+	u8 v_sync_off_high;
+	u8 reserved;
+    } part2;
 } __attribute__((packed));
 
 struct intel_sdvo_pixel_clock_range {
-	u16 min;	/**< pixel clock, in 10kHz units */
-	u16 max;	/**< pixel clock, in 10kHz units */
+    u16 min;			/**< pixel clock, in 10kHz units */
+    u16 max;			/**< pixel clock, in 10kHz units */
 } __attribute__((packed));
 
 struct intel_sdvo_preferred_input_timing_args {
-	u16 clock;
-	u16 width;
-	u16 height;
-	u8	interlace:1;
-	u8	scaled:1;
-	u8	pad:6;
+    u16 clock;
+    u16 width;
+    u16 height;
+    u8	interlace:1;
+    u8	scaled:1;
+    u8	pad:6;
 } __attribute__((packed));
 
 /* I2C registers for SDVO */
@@ -159,9 +159,9 @@ struct intel_sdvo_preferred_input_timing_args {
  */
 #define SDVO_CMD_GET_TRAINED_INPUTS			0x03
 struct intel_sdvo_get_trained_inputs_response {
-	unsigned int input0_trained:1;
-	unsigned int input1_trained:1;
-	unsigned int pad:6;
+    unsigned int input0_trained:1;
+    unsigned int input1_trained:1;
+    unsigned int pad:6;
 } __attribute__((packed));
 
 /** Returns a struct intel_sdvo_output_flags of active outputs. */
@@ -182,7 +182,7 @@ struct intel_sdvo_get_trained_inputs_response {
  */
 #define SDVO_CMD_GET_IN_OUT_MAP				0x06
 struct intel_sdvo_in_out_map {
-	u16 in0, in1;
+    u16 in0, in1;
 };
 
 /**
@@ -215,10 +215,10 @@ struct intel_sdvo_in_out_map {
 
 #define SDVO_CMD_GET_INTERRUPT_EVENT_SOURCE		0x0f
 struct intel_sdvo_get_interrupt_event_source_response {
-	u16 interrupt_status;
-	unsigned int ambient_light_interrupt:1;
-	unsigned int hdmi_audio_encrypt_change:1;
-	unsigned int pad:6;
+    u16 interrupt_status;
+    unsigned int ambient_light_interrupt:1;
+    unsigned int hdmi_audio_encrypt_change:1;
+    unsigned int pad:6;
 } __attribute__((packed));
 
 /**
@@ -230,8 +230,8 @@ struct intel_sdvo_get_interrupt_event_source_response {
  */
 #define SDVO_CMD_SET_TARGET_INPUT			0x10
 struct intel_sdvo_set_target_input_args {
-	unsigned int target_1:1;
-	unsigned int pad:7;
+    unsigned int target_1:1;
+    unsigned int pad:7;
 } __attribute__((packed));
 
 /**
@@ -319,57 +319,57 @@ struct intel_sdvo_set_target_input_args {
 #define SDVO_CMD_GET_SUPPORTED_TV_FORMATS		0x27
 /** 6 bytes of bit flags for TV formats shared by all TV format functions */
 struct intel_sdvo_tv_format {
-	unsigned int ntsc_m:1;
-	unsigned int ntsc_j:1;
-	unsigned int ntsc_443:1;
-	unsigned int pal_b:1;
-	unsigned int pal_d:1;
-	unsigned int pal_g:1;
-	unsigned int pal_h:1;
-	unsigned int pal_i:1;
-
-	unsigned int pal_m:1;
-	unsigned int pal_n:1;
-	unsigned int pal_nc:1;
-	unsigned int pal_60:1;
-	unsigned int secam_b:1;
-	unsigned int secam_d:1;
-	unsigned int secam_g:1;
-	unsigned int secam_k:1;
-
-	unsigned int secam_k1:1;
-	unsigned int secam_l:1;
-	unsigned int secam_60:1;
-	unsigned int hdtv_std_smpte_240m_1080i_59:1;
-	unsigned int hdtv_std_smpte_240m_1080i_60:1;
-	unsigned int hdtv_std_smpte_260m_1080i_59:1;
-	unsigned int hdtv_std_smpte_260m_1080i_60:1;
-	unsigned int hdtv_std_smpte_274m_1080i_50:1;
-
-	unsigned int hdtv_std_smpte_274m_1080i_59:1;
-	unsigned int hdtv_std_smpte_274m_1080i_60:1;
-	unsigned int hdtv_std_smpte_274m_1080p_23:1;
-	unsigned int hdtv_std_smpte_274m_1080p_24:1;
-	unsigned int hdtv_std_smpte_274m_1080p_25:1;
-	unsigned int hdtv_std_smpte_274m_1080p_29:1;
-	unsigned int hdtv_std_smpte_274m_1080p_30:1;
-	unsigned int hdtv_std_smpte_274m_1080p_50:1;
-
-	unsigned int hdtv_std_smpte_274m_1080p_59:1;
-	unsigned int hdtv_std_smpte_274m_1080p_60:1;
-	unsigned int hdtv_std_smpte_295m_1080i_50:1;
-	unsigned int hdtv_std_smpte_295m_1080p_50:1;
-	unsigned int hdtv_std_smpte_296m_720p_59:1;
-	unsigned int hdtv_std_smpte_296m_720p_60:1;
-	unsigned int hdtv_std_smpte_296m_720p_50:1;
-	unsigned int hdtv_std_smpte_293m_480p_59:1;
-
-	unsigned int hdtv_std_smpte_170m_480i_59:1;
-	unsigned int hdtv_std_iturbt601_576i_50:1;
-	unsigned int hdtv_std_iturbt601_576p_50:1;
-	unsigned int hdtv_std_eia_7702a_480i_60:1;
-	unsigned int hdtv_std_eia_7702a_480p_60:1;
-	unsigned int pad:3;
+    unsigned int ntsc_m:1;
+    unsigned int ntsc_j:1;
+    unsigned int ntsc_443:1;
+    unsigned int pal_b:1;
+    unsigned int pal_d:1;
+    unsigned int pal_g:1;
+    unsigned int pal_h:1;
+    unsigned int pal_i:1;
+
+    unsigned int pal_m:1;
+    unsigned int pal_n:1;
+    unsigned int pal_nc:1;
+    unsigned int pal_60:1;
+    unsigned int secam_b:1;
+    unsigned int secam_d:1;
+    unsigned int secam_g:1;
+    unsigned int secam_k:1;
+
+    unsigned int secam_k1:1;
+    unsigned int secam_l:1;
+    unsigned int secam_60:1;
+    unsigned int hdtv_std_smpte_240m_1080i_59:1;
+    unsigned int hdtv_std_smpte_240m_1080i_60:1;
+    unsigned int hdtv_std_smpte_260m_1080i_59:1;
+    unsigned int hdtv_std_smpte_260m_1080i_60:1;
+    unsigned int hdtv_std_smpte_274m_1080i_50:1;
+
+    unsigned int hdtv_std_smpte_274m_1080i_59:1;
+    unsigned int hdtv_std_smpte_274m_1080i_60:1;
+    unsigned int hdtv_std_smpte_274m_1080p_23:1;
+    unsigned int hdtv_std_smpte_274m_1080p_24:1;
+    unsigned int hdtv_std_smpte_274m_1080p_25:1;
+    unsigned int hdtv_std_smpte_274m_1080p_29:1;
+    unsigned int hdtv_std_smpte_274m_1080p_30:1;
+    unsigned int hdtv_std_smpte_274m_1080p_50:1;
+
+    unsigned int hdtv_std_smpte_274m_1080p_59:1;
+    unsigned int hdtv_std_smpte_274m_1080p_60:1;
+    unsigned int hdtv_std_smpte_295m_1080i_50:1;
+    unsigned int hdtv_std_smpte_295m_1080p_50:1;
+    unsigned int hdtv_std_smpte_296m_720p_59:1;
+    unsigned int hdtv_std_smpte_296m_720p_60:1;
+    unsigned int hdtv_std_smpte_296m_720p_50:1;
+    unsigned int hdtv_std_smpte_293m_480p_59:1;
+
+    unsigned int hdtv_std_smpte_170m_480i_59:1;
+    unsigned int hdtv_std_iturbt601_576i_50:1;
+    unsigned int hdtv_std_iturbt601_576p_50:1;
+    unsigned int hdtv_std_eia_7702a_480i_60:1;
+    unsigned int hdtv_std_eia_7702a_480p_60:1;
+    unsigned int pad:3;
 } __attribute__((packed));
 
 #define SDVO_CMD_GET_TV_FORMAT				0x28
@@ -379,53 +379,53 @@ struct intel_sdvo_tv_format {
 /** Returns the resolutiosn that can be used with the given TV format */
 #define SDVO_CMD_GET_SDTV_RESOLUTION_SUPPORT		0x83
 struct intel_sdvo_sdtv_resolution_request {
-	unsigned int ntsc_m:1;
-	unsigned int ntsc_j:1;
-	unsigned int ntsc_443:1;
-	unsigned int pal_b:1;
-	unsigned int pal_d:1;
-	unsigned int pal_g:1;
-	unsigned int pal_h:1;
-	unsigned int pal_i:1;
-
-	unsigned int pal_m:1;
-	unsigned int pal_n:1;
-	unsigned int pal_nc:1;
-	unsigned int pal_60:1;
-	unsigned int secam_b:1;
-	unsigned int secam_d:1;
-	unsigned int secam_g:1;
-	unsigned int secam_k:1;
-
-	unsigned int secam_k1:1;
-	unsigned int secam_l:1;
-	unsigned int secam_60:1;
-	unsigned int pad:5;
+    unsigned int ntsc_m:1;
+    unsigned int ntsc_j:1;
+    unsigned int ntsc_443:1;
+    unsigned int pal_b:1;
+    unsigned int pal_d:1;
+    unsigned int pal_g:1;
+    unsigned int pal_h:1;
+    unsigned int pal_i:1;
+
+    unsigned int pal_m:1;
+    unsigned int pal_n:1;
+    unsigned int pal_nc:1;
+    unsigned int pal_60:1;
+    unsigned int secam_b:1;
+    unsigned int secam_d:1;
+    unsigned int secam_g:1;
+    unsigned int secam_k:1;
+
+    unsigned int secam_k1:1;
+    unsigned int secam_l:1;
+    unsigned int secam_60:1;
+    unsigned int pad:5;
 } __attribute__((packed));
 
 struct intel_sdvo_sdtv_resolution_reply {
-	unsigned int res_320x200:1;
-	unsigned int res_320x240:1;
-	unsigned int res_400x300:1;
-	unsigned int res_640x350:1;
-	unsigned int res_640x400:1;
-	unsigned int res_640x480:1;
-	unsigned int res_704x480:1;
-	unsigned int res_704x576:1;
-
-	unsigned int res_720x350:1;
-	unsigned int res_720x400:1;
-	unsigned int res_720x480:1;
-	unsigned int res_720x540:1;
-	unsigned int res_720x576:1;
-	unsigned int res_768x576:1;
-	unsigned int res_800x600:1;
-	unsigned int res_832x624:1;
-
-	unsigned int res_920x766:1;
-	unsigned int res_1024x768:1;
-	unsigned int res_1280x1024:1;
-	unsigned int pad:5;
+    unsigned int res_320x200:1;
+    unsigned int res_320x240:1;
+    unsigned int res_400x300:1;
+    unsigned int res_640x350:1;
+    unsigned int res_640x400:1;
+    unsigned int res_640x480:1;
+    unsigned int res_704x480:1;
+    unsigned int res_704x576:1;
+
+    unsigned int res_720x350:1;
+    unsigned int res_720x400:1;
+    unsigned int res_720x480:1;
+    unsigned int res_720x540:1;
+    unsigned int res_720x576:1;
+    unsigned int res_768x576:1;
+    unsigned int res_800x600:1;
+    unsigned int res_832x624:1;
+
+    unsigned int res_920x766:1;
+    unsigned int res_1024x768:1;
+    unsigned int res_1280x1024:1;
+    unsigned int pad:5;
 } __attribute__((packed));
 
 /* Get supported resolution with squire pixel aspect ratio that can be
@@ -433,90 +433,90 @@ struct intel_sdvo_sdtv_resolution_reply {
 #define SDVO_CMD_GET_SCALED_HDTV_RESOLUTION_SUPPORT		0x85
 
 struct intel_sdvo_hdtv_resolution_request {
-	unsigned int hdtv_std_smpte_240m_1080i_59:1;
-	unsigned int hdtv_std_smpte_240m_1080i_60:1;
-	unsigned int hdtv_std_smpte_260m_1080i_59:1;
-	unsigned int hdtv_std_smpte_260m_1080i_60:1;
-	unsigned int hdtv_std_smpte_274m_1080i_50:1;
-	unsigned int hdtv_std_smpte_274m_1080i_59:1;
-	unsigned int hdtv_std_smpte_274m_1080i_60:1;
-	unsigned int hdtv_std_smpte_274m_1080p_23:1;
-
-	unsigned int hdtv_std_smpte_274m_1080p_24:1;
-	unsigned int hdtv_std_smpte_274m_1080p_25:1;
-	unsigned int hdtv_std_smpte_274m_1080p_29:1;
-	unsigned int hdtv_std_smpte_274m_1080p_30:1;
-	unsigned int hdtv_std_smpte_274m_1080p_50:1;
-	unsigned int hdtv_std_smpte_274m_1080p_59:1;
-	unsigned int hdtv_std_smpte_274m_1080p_60:1;
-	unsigned int hdtv_std_smpte_295m_1080i_50:1;
-
-	unsigned int hdtv_std_smpte_295m_1080p_50:1;
-	unsigned int hdtv_std_smpte_296m_720p_59:1;
-	unsigned int hdtv_std_smpte_296m_720p_60:1;
-	unsigned int hdtv_std_smpte_296m_720p_50:1;
-	unsigned int hdtv_std_smpte_293m_480p_59:1;
-	unsigned int hdtv_std_smpte_170m_480i_59:1;
-	unsigned int hdtv_std_iturbt601_576i_50:1;
-	unsigned int hdtv_std_iturbt601_576p_50:1;
-
-	unsigned int hdtv_std_eia_7702a_480i_60:1;
-	unsigned int hdtv_std_eia_7702a_480p_60:1;
-	unsigned int pad:6;
+    unsigned int hdtv_std_smpte_240m_1080i_59:1;
+    unsigned int hdtv_std_smpte_240m_1080i_60:1;
+    unsigned int hdtv_std_smpte_260m_1080i_59:1;
+    unsigned int hdtv_std_smpte_260m_1080i_60:1;
+    unsigned int hdtv_std_smpte_274m_1080i_50:1;
+    unsigned int hdtv_std_smpte_274m_1080i_59:1;
+    unsigned int hdtv_std_smpte_274m_1080i_60:1;
+    unsigned int hdtv_std_smpte_274m_1080p_23:1;
+
+    unsigned int hdtv_std_smpte_274m_1080p_24:1;
+    unsigned int hdtv_std_smpte_274m_1080p_25:1;
+    unsigned int hdtv_std_smpte_274m_1080p_29:1;
+    unsigned int hdtv_std_smpte_274m_1080p_30:1;
+    unsigned int hdtv_std_smpte_274m_1080p_50:1;
+    unsigned int hdtv_std_smpte_274m_1080p_59:1;
+    unsigned int hdtv_std_smpte_274m_1080p_60:1;
+    unsigned int hdtv_std_smpte_295m_1080i_50:1;
+
+    unsigned int hdtv_std_smpte_295m_1080p_50:1;
+    unsigned int hdtv_std_smpte_296m_720p_59:1;
+    unsigned int hdtv_std_smpte_296m_720p_60:1;
+    unsigned int hdtv_std_smpte_296m_720p_50:1;
+    unsigned int hdtv_std_smpte_293m_480p_59:1;
+    unsigned int hdtv_std_smpte_170m_480i_59:1;
+    unsigned int hdtv_std_iturbt601_576i_50:1;
+    unsigned int hdtv_std_iturbt601_576p_50:1;
+
+    unsigned int hdtv_std_eia_7702a_480i_60:1;
+    unsigned int hdtv_std_eia_7702a_480p_60:1;
+    unsigned int pad:6;
 } __attribute__((packed));
 
 struct intel_sdvo_hdtv_resolution_reply {
-	unsigned int res_640x480:1;
-	unsigned int res_800x600:1;
-	unsigned int res_1024x768:1;
-	unsigned int res_1280x960:1;
-	unsigned int res_1400x1050:1;
-	unsigned int res_1600x1200:1;
-	unsigned int res_1920x1440:1;
-	unsigned int res_2048x1536:1;
-
-	unsigned int res_2560x1920:1;
-	unsigned int res_3200x2400:1;
-	unsigned int res_3840x2880:1;
-	unsigned int pad1:5;
-
-	unsigned int res_848x480:1;
-	unsigned int res_1064x600:1;
-	unsigned int res_1280x720:1;
-	unsigned int res_1360x768:1;
-	unsigned int res_1704x960:1;
-	unsigned int res_1864x1050:1;
-	unsigned int res_1920x1080:1;
-	unsigned int res_2128x1200:1;
-
-	unsigned int res_2560x1400:1;
-	unsigned int res_2728x1536:1;
-	unsigned int res_3408x1920:1;
-	unsigned int res_4264x2400:1;
-	unsigned int res_5120x2880:1;
-	unsigned int pad2:3;
-
-	unsigned int res_768x480:1;
-	unsigned int res_960x600:1;
-	unsigned int res_1152x720:1;
-	unsigned int res_1124x768:1;
-	unsigned int res_1536x960:1;
-	unsigned int res_1680x1050:1;
-	unsigned int res_1728x1080:1;
-	unsigned int res_1920x1200:1;
-
-	unsigned int res_2304x1440:1;
-	unsigned int res_2456x1536:1;
-	unsigned int res_3072x1920:1;
-	unsigned int res_3840x2400:1;
-	unsigned int res_4608x2880:1;
-	unsigned int pad3:3;
-
-	unsigned int res_1280x1024:1;
-	unsigned int pad4:7;
-
-	unsigned int res_1280x768:1;
-	unsigned int pad5:7;
+    unsigned int res_640x480:1;
+    unsigned int res_800x600:1;
+    unsigned int res_1024x768:1;
+    unsigned int res_1280x960:1;
+    unsigned int res_1400x1050:1;
+    unsigned int res_1600x1200:1;
+    unsigned int res_1920x1440:1;
+    unsigned int res_2048x1536:1;
+
+    unsigned int res_2560x1920:1;
+    unsigned int res_3200x2400:1;
+    unsigned int res_3840x2880:1;
+    unsigned int pad1:5;
+
+    unsigned int res_848x480:1;
+    unsigned int res_1064x600:1;
+    unsigned int res_1280x720:1;
+    unsigned int res_1360x768:1;
+    unsigned int res_1704x960:1;
+    unsigned int res_1864x1050:1;
+    unsigned int res_1920x1080:1;
+    unsigned int res_2128x1200:1;
+
+    unsigned int res_2560x1400:1;
+    unsigned int res_2728x1536:1;
+    unsigned int res_3408x1920:1;
+    unsigned int res_4264x2400:1;
+    unsigned int res_5120x2880:1;
+    unsigned int pad2:3;
+
+    unsigned int res_768x480:1;
+    unsigned int res_960x600:1;
+    unsigned int res_1152x720:1;
+    unsigned int res_1124x768:1;
+    unsigned int res_1536x960:1;
+    unsigned int res_1680x1050:1;
+    unsigned int res_1728x1080:1;
+    unsigned int res_1920x1200:1;
+
+    unsigned int res_2304x1440:1;
+    unsigned int res_2456x1536:1;
+    unsigned int res_3072x1920:1;
+    unsigned int res_3840x2400:1;
+    unsigned int res_4608x2880:1;
+    unsigned int pad3:3;
+
+    unsigned int res_1280x1024:1;
+    unsigned int pad4:7;
+
+    unsigned int res_1280x768:1;
+    unsigned int pad5:7;
 } __attribute__((packed));
 
 /* Get supported power state returns info for encoder and monitor, rely on
@@ -544,25 +544,25 @@ struct intel_sdvo_hdtv_resolution_reply {
  * The high fields are bits 8:9 of the 10-bit values.
  */
 struct sdvo_panel_power_sequencing {
-	u8 t0;
-	u8 t1;
-	u8 t2;
-	u8 t3;
-	u8 t4;
-
-	unsigned int t0_high:2;
-	unsigned int t1_high:2;
-	unsigned int t2_high:2;
-	unsigned int t3_high:2;
-
-	unsigned int t4_high:2;
-	unsigned int pad:6;
+    u8 t0;
+    u8 t1;
+    u8 t2;
+    u8 t3;
+    u8 t4;
+
+    unsigned int t0_high:2;
+    unsigned int t1_high:2;
+    unsigned int t2_high:2;
+    unsigned int t3_high:2;
+
+    unsigned int t4_high:2;
+    unsigned int pad:6;
 } __attribute__((packed));
 
 #define SDVO_CMD_GET_MAX_BACKLIGHT_LEVEL		0x30
 struct sdvo_max_backlight_reply {
-	u8 max_value;
-	u8 default_value;
+    u8 max_value;
+    u8 default_value;
 } __attribute__((packed));
 
 #define SDVO_CMD_GET_BACKLIGHT_LEVEL			0x31
@@ -570,16 +570,16 @@ struct sdvo_max_backlight_reply {
 
 #define SDVO_CMD_GET_AMBIENT_LIGHT			0x33
 struct sdvo_get_ambient_light_reply {
-	u16 trip_low;
-	u16 trip_high;
-	u16 value;
+    u16 trip_low;
+    u16 trip_high;
+    u16 value;
 } __attribute__((packed));
 #define SDVO_CMD_SET_AMBIENT_LIGHT			0x34
 struct sdvo_set_ambient_light_reply {
-	u16 trip_low;
-	u16 trip_high;
-	unsigned int enable:1;
-	unsigned int pad:7;
+    u16 trip_low;
+    u16 trip_high;
+    unsigned int enable:1;
+    unsigned int pad:7;
 } __attribute__((packed));
 
 /* Set display power state */
@@ -591,23 +591,23 @@ struct sdvo_set_ambient_light_reply {
 
 #define SDVO_CMD_GET_SUPPORTED_ENHANCEMENTS		0x84
 struct intel_sdvo_enhancements_reply {
-	unsigned int flicker_filter:1;
-	unsigned int flicker_filter_adaptive:1;
-	unsigned int flicker_filter_2d:1;
-	unsigned int saturation:1;
-	unsigned int hue:1;
-	unsigned int brightness:1;
-	unsigned int contrast:1;
-	unsigned int overscan_h:1;
-
-	unsigned int overscan_v:1;
-	unsigned int hpos:1;
-	unsigned int vpos:1;
-	unsigned int sharpness:1;
-	unsigned int dot_crawl:1;
-	unsigned int dither:1;
-	unsigned int tv_chroma_filter:1;
-	unsigned int tv_luma_filter:1;
+    unsigned int flicker_filter:1;
+    unsigned int flicker_filter_adaptive:1;
+    unsigned int flicker_filter_2d:1;
+    unsigned int saturation:1;
+    unsigned int hue:1;
+    unsigned int brightness:1;
+    unsigned int contrast:1;
+    unsigned int overscan_h:1;
+
+    unsigned int overscan_v:1;
+    unsigned int hpos:1;
+    unsigned int vpos:1;
+    unsigned int sharpness:1;
+    unsigned int dot_crawl:1;
+    unsigned int dither:1;
+    unsigned int tv_chroma_filter:1;
+    unsigned int tv_luma_filter:1;
 } __attribute__((packed));
 
 /* Picture enhancement limits below are dependent on the current TV format,
@@ -628,8 +628,8 @@ struct intel_sdvo_enhancements_reply {
 #define SDVO_CMD_GET_MAX_TV_CHROMA_FILTER		0x74
 #define SDVO_CMD_GET_MAX_TV_LUMA_FILTER			0x77
 struct intel_sdvo_enhancement_limits_reply {
-	u16 max_value;
-	u16 default_value;
+    u16 max_value;
+    u16 default_value;
 } __attribute__((packed));
 
 #define SDVO_CMD_GET_LVDS_PANEL_INFORMATION		0x7f
@@ -670,8 +670,8 @@ struct intel_sdvo_enhancement_limits_reply {
 #define SDVO_CMD_GET_TV_LUMA_FILTER			0x78
 #define SDVO_CMD_SET_TV_LUMA_FILTER			0x79
 struct intel_sdvo_enhancements_arg {
-	u16 value;
-} __attribute__((packed));
+    u16 value;
+}__attribute__((packed));
 
 #define SDVO_CMD_GET_DOT_CRAWL				0x70
 #define SDVO_CMD_SET_DOT_CRAWL				0x71
@@ -724,7 +724,7 @@ struct intel_sdvo_enhancements_arg {
 #define SDVO_CMD_GET_AUDIO_TX_INFO	0x9c
 #define SDVO_NEED_TO_STALL  (1 << 7)
 
-struct intel_sdvo_encode {
-	u8 dvi_rev;
-	u8 hdmi_rev;
+struct intel_sdvo_encode{
+    u8 dvi_rev;
+    u8 hdmi_rev;
 } __attribute__ ((packed));
diff --git a/drivers/gpu/drm/i915/intel_sprite.c b/drivers/gpu/drm/i915/intel_sprite.c
deleted file mode 100644
index 98444ab..0000000
--- a/drivers/gpu/drm/i915/intel_sprite.c
+++ /dev/null
@@ -1,664 +0,0 @@
-/*
- * Copyright © 2011 Intel Corporation
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice (including the next
- * paragraph) shall be included in all copies or substantial portions of the
- * Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
- * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
- * SOFTWARE.
- *
- * Authors:
- *   Jesse Barnes <jbarnes@virtuousgeek.org>
- *
- * New plane/sprite handling.
- *
- * The older chips had a separate interface for programming plane related
- * registers; newer ones are much simpler and we can use the new DRM plane
- * support.
- */
-#include "drmP.h"
-#include "drm_crtc.h"
-#include "drm_fourcc.h"
-#include "intel_drv.h"
-#include "i915_drm.h"
-#include "i915_drv.h"
-
-static void
-ivb_update_plane(struct drm_plane *plane, struct drm_framebuffer *fb,
-		 struct drm_i915_gem_object *obj, int crtc_x, int crtc_y,
-		 unsigned int crtc_w, unsigned int crtc_h,
-		 uint32_t x, uint32_t y,
-		 uint32_t src_w, uint32_t src_h)
-{
-	struct drm_device *dev = plane->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_plane *intel_plane = to_intel_plane(plane);
-	int pipe = intel_plane->pipe;
-	u32 sprctl, sprscale = 0;
-	int pixel_size;
-
-	sprctl = I915_READ(SPRCTL(pipe));
-
-	/* Mask out pixel format bits in case we change it */
-	sprctl &= ~SPRITE_PIXFORMAT_MASK;
-	sprctl &= ~SPRITE_RGB_ORDER_RGBX;
-	sprctl &= ~SPRITE_YUV_BYTE_ORDER_MASK;
-
-	switch (fb->pixel_format) {
-	case DRM_FORMAT_XBGR8888:
-		sprctl |= SPRITE_FORMAT_RGBX888;
-		pixel_size = 4;
-		break;
-	case DRM_FORMAT_XRGB8888:
-		sprctl |= SPRITE_FORMAT_RGBX888 | SPRITE_RGB_ORDER_RGBX;
-		pixel_size = 4;
-		break;
-	case DRM_FORMAT_YUYV:
-		sprctl |= SPRITE_FORMAT_YUV422 | SPRITE_YUV_ORDER_YUYV;
-		pixel_size = 2;
-		break;
-	case DRM_FORMAT_YVYU:
-		sprctl |= SPRITE_FORMAT_YUV422 | SPRITE_YUV_ORDER_YVYU;
-		pixel_size = 2;
-		break;
-	case DRM_FORMAT_UYVY:
-		sprctl |= SPRITE_FORMAT_YUV422 | SPRITE_YUV_ORDER_UYVY;
-		pixel_size = 2;
-		break;
-	case DRM_FORMAT_VYUY:
-		sprctl |= SPRITE_FORMAT_YUV422 | SPRITE_YUV_ORDER_VYUY;
-		pixel_size = 2;
-		break;
-	default:
-		DRM_DEBUG_DRIVER("bad pixel format, assuming RGBX888\n");
-		sprctl |= DVS_FORMAT_RGBX888;
-		pixel_size = 4;
-		break;
-	}
-
-	if (obj->tiling_mode != I915_TILING_NONE)
-		sprctl |= SPRITE_TILED;
-
-	/* must disable */
-	sprctl |= SPRITE_TRICKLE_FEED_DISABLE;
-	sprctl |= SPRITE_ENABLE;
-	sprctl |= SPRITE_DEST_KEY;
-
-	/* Sizes are 0 based */
-	src_w--;
-	src_h--;
-	crtc_w--;
-	crtc_h--;
-
-	intel_update_sprite_watermarks(dev, pipe, crtc_w, pixel_size);
-
-	/*
-	 * IVB workaround: must disable low power watermarks for at least
-	 * one frame before enabling scaling.  LP watermarks can be re-enabled
-	 * when scaling is disabled.
-	 */
-	if (crtc_w != src_w || crtc_h != src_h) {
-		dev_priv->sprite_scaling_enabled = true;
-		sandybridge_update_wm(dev);
-		intel_wait_for_vblank(dev, pipe);
-		sprscale = SPRITE_SCALE_ENABLE | (src_w << 16) | src_h;
-	} else {
-		dev_priv->sprite_scaling_enabled = false;
-		/* potentially re-enable LP watermarks */
-		sandybridge_update_wm(dev);
-	}
-
-	I915_WRITE(SPRSTRIDE(pipe), fb->pitches[0]);
-	I915_WRITE(SPRPOS(pipe), (crtc_y << 16) | crtc_x);
-	if (obj->tiling_mode != I915_TILING_NONE) {
-		I915_WRITE(SPRTILEOFF(pipe), (y << 16) | x);
-	} else {
-		unsigned long offset;
-
-		offset = y * fb->pitches[0] + x * (fb->bits_per_pixel / 8);
-		I915_WRITE(SPRLINOFF(pipe), offset);
-	}
-	I915_WRITE(SPRSIZE(pipe), (crtc_h << 16) | crtc_w);
-	I915_WRITE(SPRSCALE(pipe), sprscale);
-	I915_WRITE(SPRCTL(pipe), sprctl);
-	I915_WRITE(SPRSURF(pipe), obj->gtt_offset);
-	POSTING_READ(SPRSURF(pipe));
-}
-
-static void
-ivb_disable_plane(struct drm_plane *plane)
-{
-	struct drm_device *dev = plane->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_plane *intel_plane = to_intel_plane(plane);
-	int pipe = intel_plane->pipe;
-
-	I915_WRITE(SPRCTL(pipe), I915_READ(SPRCTL(pipe)) & ~SPRITE_ENABLE);
-	/* Can't leave the scaler enabled... */
-	I915_WRITE(SPRSCALE(pipe), 0);
-	/* Activate double buffered register update */
-	I915_WRITE(SPRSURF(pipe), 0);
-	POSTING_READ(SPRSURF(pipe));
-}
-
-static int
-ivb_update_colorkey(struct drm_plane *plane,
-		    struct drm_intel_sprite_colorkey *key)
-{
-	struct drm_device *dev = plane->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_plane *intel_plane;
-	u32 sprctl;
-	int ret = 0;
-
-	intel_plane = to_intel_plane(plane);
-
-	I915_WRITE(SPRKEYVAL(intel_plane->pipe), key->min_value);
-	I915_WRITE(SPRKEYMAX(intel_plane->pipe), key->max_value);
-	I915_WRITE(SPRKEYMSK(intel_plane->pipe), key->channel_mask);
-
-	sprctl = I915_READ(SPRCTL(intel_plane->pipe));
-	sprctl &= ~(SPRITE_SOURCE_KEY | SPRITE_DEST_KEY);
-	if (key->flags & I915_SET_COLORKEY_DESTINATION)
-		sprctl |= SPRITE_DEST_KEY;
-	else if (key->flags & I915_SET_COLORKEY_SOURCE)
-		sprctl |= SPRITE_SOURCE_KEY;
-	I915_WRITE(SPRCTL(intel_plane->pipe), sprctl);
-
-	POSTING_READ(SPRKEYMSK(intel_plane->pipe));
-
-	return ret;
-}
-
-static void
-ivb_get_colorkey(struct drm_plane *plane, struct drm_intel_sprite_colorkey *key)
-{
-	struct drm_device *dev = plane->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_plane *intel_plane;
-	u32 sprctl;
-
-	intel_plane = to_intel_plane(plane);
-
-	key->min_value = I915_READ(SPRKEYVAL(intel_plane->pipe));
-	key->max_value = I915_READ(SPRKEYMAX(intel_plane->pipe));
-	key->channel_mask = I915_READ(SPRKEYMSK(intel_plane->pipe));
-	key->flags = 0;
-
-	sprctl = I915_READ(SPRCTL(intel_plane->pipe));
-
-	if (sprctl & SPRITE_DEST_KEY)
-		key->flags = I915_SET_COLORKEY_DESTINATION;
-	else if (sprctl & SPRITE_SOURCE_KEY)
-		key->flags = I915_SET_COLORKEY_SOURCE;
-	else
-		key->flags = I915_SET_COLORKEY_NONE;
-}
-
-static void
-snb_update_plane(struct drm_plane *plane, struct drm_framebuffer *fb,
-		 struct drm_i915_gem_object *obj, int crtc_x, int crtc_y,
-		 unsigned int crtc_w, unsigned int crtc_h,
-		 uint32_t x, uint32_t y,
-		 uint32_t src_w, uint32_t src_h)
-{
-	struct drm_device *dev = plane->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_plane *intel_plane = to_intel_plane(plane);
-	int pipe = intel_plane->pipe, pixel_size;
-	u32 dvscntr, dvsscale = 0;
-
-	dvscntr = I915_READ(DVSCNTR(pipe));
-
-	/* Mask out pixel format bits in case we change it */
-	dvscntr &= ~DVS_PIXFORMAT_MASK;
-	dvscntr &= ~DVS_RGB_ORDER_RGBX;
-	dvscntr &= ~DVS_YUV_BYTE_ORDER_MASK;
-
-	switch (fb->pixel_format) {
-	case DRM_FORMAT_XBGR8888:
-		dvscntr |= DVS_FORMAT_RGBX888;
-		pixel_size = 4;
-		break;
-	case DRM_FORMAT_XRGB8888:
-		dvscntr |= DVS_FORMAT_RGBX888 | DVS_RGB_ORDER_RGBX;
-		pixel_size = 4;
-		break;
-	case DRM_FORMAT_YUYV:
-		dvscntr |= DVS_FORMAT_YUV422 | DVS_YUV_ORDER_YUYV;
-		pixel_size = 2;
-		break;
-	case DRM_FORMAT_YVYU:
-		dvscntr |= DVS_FORMAT_YUV422 | DVS_YUV_ORDER_YVYU;
-		pixel_size = 2;
-		break;
-	case DRM_FORMAT_UYVY:
-		dvscntr |= DVS_FORMAT_YUV422 | DVS_YUV_ORDER_UYVY;
-		pixel_size = 2;
-		break;
-	case DRM_FORMAT_VYUY:
-		dvscntr |= DVS_FORMAT_YUV422 | DVS_YUV_ORDER_VYUY;
-		pixel_size = 2;
-		break;
-	default:
-		DRM_DEBUG_DRIVER("bad pixel format, assuming RGBX888\n");
-		dvscntr |= DVS_FORMAT_RGBX888;
-		pixel_size = 4;
-		break;
-	}
-
-	if (obj->tiling_mode != I915_TILING_NONE)
-		dvscntr |= DVS_TILED;
-
-	/* must disable */
-	dvscntr |= DVS_TRICKLE_FEED_DISABLE;
-	dvscntr |= DVS_ENABLE;
-
-	/* Sizes are 0 based */
-	src_w--;
-	src_h--;
-	crtc_w--;
-	crtc_h--;
-
-	intel_update_sprite_watermarks(dev, pipe, crtc_w, pixel_size);
-
-	if (crtc_w != src_w || crtc_h != src_h)
-		dvsscale = DVS_SCALE_ENABLE | (src_w << 16) | src_h;
-
-	I915_WRITE(DVSSTRIDE(pipe), fb->pitches[0]);
-	I915_WRITE(DVSPOS(pipe), (crtc_y << 16) | crtc_x);
-	if (obj->tiling_mode != I915_TILING_NONE) {
-		I915_WRITE(DVSTILEOFF(pipe), (y << 16) | x);
-	} else {
-		unsigned long offset;
-
-		offset = y * fb->pitches[0] + x * (fb->bits_per_pixel / 8);
-		I915_WRITE(DVSLINOFF(pipe), offset);
-	}
-	I915_WRITE(DVSSIZE(pipe), (crtc_h << 16) | crtc_w);
-	I915_WRITE(DVSSCALE(pipe), dvsscale);
-	I915_WRITE(DVSCNTR(pipe), dvscntr);
-	I915_WRITE(DVSSURF(pipe), obj->gtt_offset);
-	POSTING_READ(DVSSURF(pipe));
-}
-
-static void
-snb_disable_plane(struct drm_plane *plane)
-{
-	struct drm_device *dev = plane->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_plane *intel_plane = to_intel_plane(plane);
-	int pipe = intel_plane->pipe;
-
-	I915_WRITE(DVSCNTR(pipe), I915_READ(DVSCNTR(pipe)) & ~DVS_ENABLE);
-	/* Disable the scaler */
-	I915_WRITE(DVSSCALE(pipe), 0);
-	/* Flush double buffered register updates */
-	I915_WRITE(DVSSURF(pipe), 0);
-	POSTING_READ(DVSSURF(pipe));
-}
-
-static void
-intel_enable_primary(struct drm_crtc *crtc)
-{
-	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int reg = DSPCNTR(intel_crtc->plane);
-
-	I915_WRITE(reg, I915_READ(reg) | DISPLAY_PLANE_ENABLE);
-}
-
-static void
-intel_disable_primary(struct drm_crtc *crtc)
-{
-	struct drm_device *dev = crtc->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	int reg = DSPCNTR(intel_crtc->plane);
-
-	I915_WRITE(reg, I915_READ(reg) & ~DISPLAY_PLANE_ENABLE);
-}
-
-static int
-snb_update_colorkey(struct drm_plane *plane,
-		    struct drm_intel_sprite_colorkey *key)
-{
-	struct drm_device *dev = plane->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_plane *intel_plane;
-	u32 dvscntr;
-	int ret = 0;
-
-	intel_plane = to_intel_plane(plane);
-
-	I915_WRITE(DVSKEYVAL(intel_plane->pipe), key->min_value);
-	I915_WRITE(DVSKEYMAX(intel_plane->pipe), key->max_value);
-	I915_WRITE(DVSKEYMSK(intel_plane->pipe), key->channel_mask);
-
-	dvscntr = I915_READ(DVSCNTR(intel_plane->pipe));
-	dvscntr &= ~(DVS_SOURCE_KEY | DVS_DEST_KEY);
-	if (key->flags & I915_SET_COLORKEY_DESTINATION)
-		dvscntr |= DVS_DEST_KEY;
-	else if (key->flags & I915_SET_COLORKEY_SOURCE)
-		dvscntr |= DVS_SOURCE_KEY;
-	I915_WRITE(DVSCNTR(intel_plane->pipe), dvscntr);
-
-	POSTING_READ(DVSKEYMSK(intel_plane->pipe));
-
-	return ret;
-}
-
-static void
-snb_get_colorkey(struct drm_plane *plane, struct drm_intel_sprite_colorkey *key)
-{
-	struct drm_device *dev = plane->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_plane *intel_plane;
-	u32 dvscntr;
-
-	intel_plane = to_intel_plane(plane);
-
-	key->min_value = I915_READ(DVSKEYVAL(intel_plane->pipe));
-	key->max_value = I915_READ(DVSKEYMAX(intel_plane->pipe));
-	key->channel_mask = I915_READ(DVSKEYMSK(intel_plane->pipe));
-	key->flags = 0;
-
-	dvscntr = I915_READ(DVSCNTR(intel_plane->pipe));
-
-	if (dvscntr & DVS_DEST_KEY)
-		key->flags = I915_SET_COLORKEY_DESTINATION;
-	else if (dvscntr & DVS_SOURCE_KEY)
-		key->flags = I915_SET_COLORKEY_SOURCE;
-	else
-		key->flags = I915_SET_COLORKEY_NONE;
-}
-
-static int
-intel_update_plane(struct drm_plane *plane, struct drm_crtc *crtc,
-		   struct drm_framebuffer *fb, int crtc_x, int crtc_y,
-		   unsigned int crtc_w, unsigned int crtc_h,
-		   uint32_t src_x, uint32_t src_y,
-		   uint32_t src_w, uint32_t src_h)
-{
-	struct drm_device *dev = plane->dev;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
-	struct intel_plane *intel_plane = to_intel_plane(plane);
-	struct intel_framebuffer *intel_fb;
-	struct drm_i915_gem_object *obj, *old_obj;
-	int pipe = intel_plane->pipe;
-	int ret = 0;
-	int x = src_x >> 16, y = src_y >> 16;
-	int primary_w = crtc->mode.hdisplay, primary_h = crtc->mode.vdisplay;
-	bool disable_primary = false;
-
-	intel_fb = to_intel_framebuffer(fb);
-	obj = intel_fb->obj;
-
-	old_obj = intel_plane->obj;
-
-	/* Pipe must be running... */
-	if (!(I915_READ(PIPECONF(pipe)) & PIPECONF_ENABLE))
-		return -EINVAL;
-
-	if (crtc_x >= primary_w || crtc_y >= primary_h)
-		return -EINVAL;
-
-	/* Don't modify another pipe's plane */
-	if (intel_plane->pipe != intel_crtc->pipe)
-		return -EINVAL;
-
-	/*
-	 * Clamp the width & height into the visible area.  Note we don't
-	 * try to scale the source if part of the visible region is offscreen.
-	 * The caller must handle that by adjusting source offset and size.
-	 */
-	if ((crtc_x < 0) && ((crtc_x + crtc_w) > 0)) {
-		crtc_w += crtc_x;
-		crtc_x = 0;
-	}
-	if ((crtc_x + crtc_w) <= 0) /* Nothing to display */
-		goto out;
-	if ((crtc_x + crtc_w) > primary_w)
-		crtc_w = primary_w - crtc_x;
-
-	if ((crtc_y < 0) && ((crtc_y + crtc_h) > 0)) {
-		crtc_h += crtc_y;
-		crtc_y = 0;
-	}
-	if ((crtc_y + crtc_h) <= 0) /* Nothing to display */
-		goto out;
-	if (crtc_y + crtc_h > primary_h)
-		crtc_h = primary_h - crtc_y;
-
-	if (!crtc_w || !crtc_h) /* Again, nothing to display */
-		goto out;
-
-	/*
-	 * We can take a larger source and scale it down, but
-	 * only so much...  16x is the max on SNB.
-	 */
-	if (((src_w * src_h) / (crtc_w * crtc_h)) > intel_plane->max_downscale)
-		return -EINVAL;
-
-	/*
-	 * If the sprite is completely covering the primary plane,
-	 * we can disable the primary and save power.
-	 */
-	if ((crtc_x == 0) && (crtc_y == 0) &&
-	    (crtc_w == primary_w) && (crtc_h == primary_h))
-		disable_primary = true;
-
-	mutex_lock(&dev->struct_mutex);
-
-	ret = intel_pin_and_fence_fb_obj(dev, obj, NULL);
-	if (ret)
-		goto out_unlock;
-
-	intel_plane->obj = obj;
-
-	/*
-	 * Be sure to re-enable the primary before the sprite is no longer
-	 * covering it fully.
-	 */
-	if (!disable_primary && intel_plane->primary_disabled) {
-		intel_enable_primary(crtc);
-		intel_plane->primary_disabled = false;
-	}
-
-	intel_plane->update_plane(plane, fb, obj, crtc_x, crtc_y,
-				  crtc_w, crtc_h, x, y, src_w, src_h);
-
-	if (disable_primary) {
-		intel_disable_primary(crtc);
-		intel_plane->primary_disabled = true;
-	}
-
-	/* Unpin old obj after new one is active to avoid ugliness */
-	if (old_obj) {
-		/*
-		 * It's fairly common to simply update the position of
-		 * an existing object.  In that case, we don't need to
-		 * wait for vblank to avoid ugliness, we only need to
-		 * do the pin & ref bookkeeping.
-		 */
-		if (old_obj != obj) {
-			mutex_unlock(&dev->struct_mutex);
-			intel_wait_for_vblank(dev, to_intel_crtc(crtc)->pipe);
-			mutex_lock(&dev->struct_mutex);
-		}
-		intel_unpin_fb_obj(old_obj);
-	}
-
-out_unlock:
-	mutex_unlock(&dev->struct_mutex);
-out:
-	return ret;
-}
-
-static int
-intel_disable_plane(struct drm_plane *plane)
-{
-	struct drm_device *dev = plane->dev;
-	struct intel_plane *intel_plane = to_intel_plane(plane);
-	int ret = 0;
-
-	if (intel_plane->primary_disabled) {
-		intel_enable_primary(plane->crtc);
-		intel_plane->primary_disabled = false;
-	}
-
-	intel_plane->disable_plane(plane);
-
-	if (!intel_plane->obj)
-		goto out;
-
-	mutex_lock(&dev->struct_mutex);
-	intel_unpin_fb_obj(intel_plane->obj);
-	intel_plane->obj = NULL;
-	mutex_unlock(&dev->struct_mutex);
-out:
-
-	return ret;
-}
-
-static void intel_destroy_plane(struct drm_plane *plane)
-{
-	struct intel_plane *intel_plane = to_intel_plane(plane);
-	intel_disable_plane(plane);
-	drm_plane_cleanup(plane);
-	kfree(intel_plane);
-}
-
-int intel_sprite_set_colorkey(struct drm_device *dev, void *data,
-			      struct drm_file *file_priv)
-{
-	struct drm_intel_sprite_colorkey *set = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_mode_object *obj;
-	struct drm_plane *plane;
-	struct intel_plane *intel_plane;
-	int ret = 0;
-
-	if (!dev_priv)
-		return -EINVAL;
-
-	/* Make sure we don't try to enable both src & dest simultaneously */
-	if ((set->flags & (I915_SET_COLORKEY_DESTINATION | I915_SET_COLORKEY_SOURCE)) == (I915_SET_COLORKEY_DESTINATION | I915_SET_COLORKEY_SOURCE))
-		return -EINVAL;
-
-	mutex_lock(&dev->mode_config.mutex);
-
-	obj = drm_mode_object_find(dev, set->plane_id, DRM_MODE_OBJECT_PLANE);
-	if (!obj) {
-		ret = -EINVAL;
-		goto out_unlock;
-	}
-
-	plane = obj_to_plane(obj);
-	intel_plane = to_intel_plane(plane);
-	ret = intel_plane->update_colorkey(plane, set);
-
-out_unlock:
-	mutex_unlock(&dev->mode_config.mutex);
-	return ret;
-}
-
-int intel_sprite_get_colorkey(struct drm_device *dev, void *data,
-			      struct drm_file *file_priv)
-{
-	struct drm_intel_sprite_colorkey *get = data;
-	struct drm_i915_private *dev_priv = dev->dev_private;
-	struct drm_mode_object *obj;
-	struct drm_plane *plane;
-	struct intel_plane *intel_plane;
-	int ret = 0;
-
-	if (!dev_priv)
-		return -EINVAL;
-
-	mutex_lock(&dev->mode_config.mutex);
-
-	obj = drm_mode_object_find(dev, get->plane_id, DRM_MODE_OBJECT_PLANE);
-	if (!obj) {
-		ret = -EINVAL;
-		goto out_unlock;
-	}
-
-	plane = obj_to_plane(obj);
-	intel_plane = to_intel_plane(plane);
-	intel_plane->get_colorkey(plane, get);
-
-out_unlock:
-	mutex_unlock(&dev->mode_config.mutex);
-	return ret;
-}
-
-static const struct drm_plane_funcs intel_plane_funcs = {
-	.update_plane = intel_update_plane,
-	.disable_plane = intel_disable_plane,
-	.destroy = intel_destroy_plane,
-};
-
-static uint32_t snb_plane_formats[] = {
-	DRM_FORMAT_XBGR8888,
-	DRM_FORMAT_XRGB8888,
-	DRM_FORMAT_YUYV,
-	DRM_FORMAT_YVYU,
-	DRM_FORMAT_UYVY,
-	DRM_FORMAT_VYUY,
-};
-
-int
-intel_plane_init(struct drm_device *dev, enum pipe pipe)
-{
-	struct intel_plane *intel_plane;
-	unsigned long possible_crtcs;
-	int ret;
-
-	if (!(IS_GEN6(dev) || IS_GEN7(dev)))
-		return -ENODEV;
-
-	intel_plane = kzalloc(sizeof(struct intel_plane), GFP_KERNEL);
-	if (!intel_plane)
-		return -ENOMEM;
-
-	if (IS_GEN6(dev)) {
-		intel_plane->max_downscale = 16;
-		intel_plane->update_plane = snb_update_plane;
-		intel_plane->disable_plane = snb_disable_plane;
-		intel_plane->update_colorkey = snb_update_colorkey;
-		intel_plane->get_colorkey = snb_get_colorkey;
-	} else if (IS_GEN7(dev)) {
-		intel_plane->max_downscale = 2;
-		intel_plane->update_plane = ivb_update_plane;
-		intel_plane->disable_plane = ivb_disable_plane;
-		intel_plane->update_colorkey = ivb_update_colorkey;
-		intel_plane->get_colorkey = ivb_get_colorkey;
-	}
-
-	intel_plane->pipe = pipe;
-	possible_crtcs = (1 << pipe);
-	ret = drm_plane_init(dev, &intel_plane->base, possible_crtcs,
-			     &intel_plane_funcs, snb_plane_formats,
-			     ARRAY_SIZE(snb_plane_formats), false);
-	if (ret)
-		kfree(intel_plane);
-
-	return ret;
-}
-
diff --git a/drivers/gpu/drm/i915/intel_tv.c b/drivers/gpu/drm/i915/intel_tv.c
index 5f332e0..2136e6b 100644
--- a/drivers/gpu/drm/i915/intel_tv.c
+++ b/drivers/gpu/drm/i915/intel_tv.c
@@ -194,10 +194,10 @@ static const u32 filter_table[] = {
  *
  *     if (f >= 1) {
  *         exp = 0x7;
- *	   mant = 1 << 8;
+ * 	   mant = 1 << 8;
  *     } else {
  *         for (exp = 0; exp < 3 && f < 0.5; exp++)
- *	   f *= 2.0;
+ * 	       f *= 2.0;
  *         mant = (f * (1 << 9) + 0.5);
  *         if (mant >= (1 << 9))
  *             mant = (1 << 9) - 1;
@@ -430,7 +430,7 @@ static const struct tv_mode tv_modes[] = {
 		.vsync_start_f1	= 6,		    .vsync_start_f2	= 7,
 		.vsync_len	= 6,
 
-		.veq_ena	= true,		    .veq_start_f1	= 0,
+		.veq_ena	= true,		    .veq_start_f1    	= 0,
 		.veq_start_f2	= 1,		    .veq_len		= 18,
 
 		.vi_end_f1	= 20,		    .vi_end_f2		= 21,
@@ -472,7 +472,7 @@ static const struct tv_mode tv_modes[] = {
 		.vsync_start_f1 = 6,		    .vsync_start_f2	= 7,
 		.vsync_len	= 6,
 
-		.veq_ena	= true,		    .veq_start_f1	= 0,
+		.veq_ena	= true,		    .veq_start_f1    	= 0,
 		.veq_start_f2	= 1,		    .veq_len		= 18,
 
 		.vi_end_f1	= 20,		    .vi_end_f2		= 21,
@@ -515,7 +515,7 @@ static const struct tv_mode tv_modes[] = {
 		.vsync_start_f1	= 6,	    .vsync_start_f2	= 7,
 		.vsync_len	= 6,
 
-		.veq_ena      = true,	    .veq_start_f1	= 0,
+		.veq_ena	= true,		    .veq_start_f1    	= 0,
 		.veq_start_f2 = 1,	    .veq_len		= 18,
 
 		.vi_end_f1	= 20,		    .vi_end_f2		= 21,
@@ -558,7 +558,7 @@ static const struct tv_mode tv_modes[] = {
 		.vsync_start_f1	= 6,		    .vsync_start_f2	= 7,
 		.vsync_len	= 6,
 
-		.veq_ena	= true,		    .veq_start_f1	= 0,
+		.veq_ena	= true,		    .veq_start_f1    	= 0,
 		.veq_start_f2	= 1,		    .veq_len		= 18,
 
 		.vi_end_f1	= 20,		    .vi_end_f2		= 21,
@@ -602,14 +602,14 @@ static const struct tv_mode tv_modes[] = {
 		.vsync_start_f1	= 6,	   .vsync_start_f2	= 7,
 		.vsync_len	= 6,
 
-		.veq_ena	= true,		    .veq_start_f1	= 0,
+		.veq_ena	= true,		    .veq_start_f1    	= 0,
 		.veq_start_f2	= 1,		    .veq_len		= 18,
 
 		.vi_end_f1	= 24,		    .vi_end_f2		= 25,
 		.nbr_end	= 286,
 
 		.burst_ena	= true,
-		.hburst_start = 73,	    .hburst_len		= 34,
+		.hburst_start = 73,	    	    .hburst_len		= 34,
 		.vburst_start_f1 = 8,	    .vburst_end_f1	= 285,
 		.vburst_start_f2 = 8,	    .vburst_end_f2	= 286,
 		.vburst_start_f3 = 9,	    .vburst_end_f3	= 286,
@@ -646,7 +646,7 @@ static const struct tv_mode tv_modes[] = {
 		.vsync_start_f1	= 5,	    .vsync_start_f2	= 6,
 		.vsync_len	= 5,
 
-		.veq_ena	= true,	    .veq_start_f1	= 0,
+		.veq_ena	= true,		    .veq_start_f1    	= 0,
 		.veq_start_f2	= 1,	    .veq_len		= 15,
 
 		.vi_end_f1	= 24,		    .vi_end_f2		= 25,
@@ -674,6 +674,78 @@ static const struct tv_mode tv_modes[] = {
 		.filter_table = filter_table,
 	},
 	{
+		.name       = "480p@59.94Hz",
+		.clock 	= 107520,
+		.refresh	= 59940,
+		.oversample     = TV_OVERSAMPLE_4X,
+		.component_only = 1,
+
+		.hsync_end      = 64,               .hblank_end         = 122,
+		.hblank_start   = 842,              .htotal             = 857,
+
+		.progressive    = true,.trilevel_sync = false,
+
+		.vsync_start_f1 = 12,               .vsync_start_f2     = 12,
+		.vsync_len      = 12,
+
+		.veq_ena        = false,
+
+		.vi_end_f1      = 44,               .vi_end_f2          = 44,
+		.nbr_end        = 479,
+
+		.burst_ena      = false,
+
+		.filter_table = filter_table,
+	},
+	{
+		.name       = "480p@60Hz",
+		.clock 	= 107520,
+		.refresh	= 60000,
+		.oversample     = TV_OVERSAMPLE_4X,
+		.component_only = 1,
+
+		.hsync_end      = 64,               .hblank_end         = 122,
+		.hblank_start   = 842,              .htotal             = 856,
+
+		.progressive    = true,.trilevel_sync = false,
+
+		.vsync_start_f1 = 12,               .vsync_start_f2     = 12,
+		.vsync_len      = 12,
+
+		.veq_ena        = false,
+
+		.vi_end_f1      = 44,               .vi_end_f2          = 44,
+		.nbr_end        = 479,
+
+		.burst_ena      = false,
+
+		.filter_table = filter_table,
+	},
+	{
+		.name       = "576p",
+		.clock 	= 107520,
+		.refresh	= 50000,
+		.oversample     = TV_OVERSAMPLE_4X,
+		.component_only = 1,
+
+		.hsync_end      = 64,               .hblank_end         = 139,
+		.hblank_start   = 859,              .htotal             = 863,
+
+		.progressive    = true,		.trilevel_sync = false,
+
+		.vsync_start_f1 = 10,               .vsync_start_f2     = 10,
+		.vsync_len      = 10,
+
+		.veq_ena        = false,
+
+		.vi_end_f1      = 48,               .vi_end_f2          = 48,
+		.nbr_end        = 575,
+
+		.burst_ena      = false,
+
+		.filter_table = filter_table,
+	},
+	{
 		.name       = "720p@60Hz",
 		.clock		= 148800,
 		.refresh	= 60000,
@@ -683,7 +755,31 @@ static const struct tv_mode tv_modes[] = {
 		.hsync_end      = 80,               .hblank_end         = 300,
 		.hblank_start   = 1580,             .htotal             = 1649,
 
-		.progressive	= true,		    .trilevel_sync = true,
+		.progressive    = true, 	    .trilevel_sync = true,
+
+		.vsync_start_f1 = 10,               .vsync_start_f2     = 10,
+		.vsync_len      = 10,
+
+		.veq_ena        = false,
+
+		.vi_end_f1      = 29,               .vi_end_f2          = 29,
+		.nbr_end        = 719,
+
+		.burst_ena      = false,
+
+		.filter_table = filter_table,
+	},
+	{
+		.name       = "720p@59.94Hz",
+		.clock		= 148800,
+		.refresh	= 59940,
+		.oversample     = TV_OVERSAMPLE_2X,
+		.component_only = 1,
+
+		.hsync_end      = 80,               .hblank_end         = 300,
+		.hblank_start   = 1580,             .htotal             = 1651,
+
+		.progressive    = true, 	    .trilevel_sync = true,
 
 		.vsync_start_f1 = 10,               .vsync_start_f2     = 10,
 		.vsync_len      = 10,
@@ -707,7 +803,7 @@ static const struct tv_mode tv_modes[] = {
 		.hsync_end      = 80,               .hblank_end         = 300,
 		.hblank_start   = 1580,             .htotal             = 1979,
 
-		.progressive	= true,		    .trilevel_sync = true,
+		.progressive    = true, 	        .trilevel_sync = true,
 
 		.vsync_start_f1 = 10,               .vsync_start_f2     = 10,
 		.vsync_len      = 10,
@@ -732,12 +828,12 @@ static const struct tv_mode tv_modes[] = {
 		.hsync_end      = 88,               .hblank_end         = 235,
 		.hblank_start   = 2155,             .htotal             = 2639,
 
-		.progressive	= false,	  .trilevel_sync = true,
+		.progressive    = false, 	    .trilevel_sync = true,
 
 		.vsync_start_f1 = 4,              .vsync_start_f2     = 5,
 		.vsync_len      = 10,
 
-		.veq_ena	= true,	    .veq_start_f1	= 4,
+		.veq_ena	= true,		    .veq_start_f1    	= 4,
 		.veq_start_f2   = 4,	    .veq_len		= 10,
 
 
@@ -758,12 +854,12 @@ static const struct tv_mode tv_modes[] = {
 		.hsync_end      = 88,               .hblank_end         = 235,
 		.hblank_start   = 2155,             .htotal             = 2199,
 
-		.progressive	= false,	    .trilevel_sync = true,
+		.progressive    = false, 	    .trilevel_sync = true,
 
 		.vsync_start_f1 = 4,               .vsync_start_f2     = 5,
 		.vsync_len      = 10,
 
-		.veq_ena	= true,		    .veq_start_f1	= 4,
+		.veq_ena	= true,		    .veq_start_f1    	= 4,
 		.veq_start_f2	= 4,		    .veq_len		= 10,
 
 
@@ -774,6 +870,32 @@ static const struct tv_mode tv_modes[] = {
 
 		.filter_table = filter_table,
 	},
+	{
+		.name       = "1080i@59.94Hz",
+		.clock		= 148800,
+		.refresh	= 29970,
+		.oversample     = TV_OVERSAMPLE_2X,
+		.component_only = 1,
+
+		.hsync_end      = 88,               .hblank_end         = 235,
+		.hblank_start   = 2155,             .htotal             = 2201,
+
+		.progressive    = false, 	    .trilevel_sync = true,
+
+		.vsync_start_f1 = 4,            .vsync_start_f2    = 5,
+		.vsync_len      = 10,
+
+		.veq_ena	= true,		    .veq_start_f1	= 4,
+		.veq_start_f2 = 4,	    	    .veq_len = 10,
+
+
+		.vi_end_f1      = 21,           .vi_end_f2         	= 22,
+		.nbr_end        = 539,
+
+		.burst_ena      = false,
+
+		.filter_table = filter_table,
+	},
 };
 
 static struct intel_tv *enc_to_intel_tv(struct drm_encoder *encoder)
@@ -794,7 +916,7 @@ intel_tv_dpms(struct drm_encoder *encoder, int mode)
 	struct drm_device *dev = encoder->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
-	switch (mode) {
+	switch(mode) {
 	case DRM_MODE_DPMS_ON:
 		I915_WRITE(TV_CTL, I915_READ(TV_CTL) | TV_ENC_ENABLE);
 		break;
@@ -811,7 +933,7 @@ intel_tv_mode_lookup(const char *tv_format)
 {
 	int i;
 
-	for (i = 0; i < sizeof(tv_modes) / sizeof(tv_modes[0]); i++) {
+	for (i = 0; i < sizeof(tv_modes) / sizeof (tv_modes[0]); i++) {
 		const struct tv_mode *tv_mode = &tv_modes[i];
 
 		if (!strcmp(tv_format, tv_mode->name))
@@ -1006,7 +1128,7 @@ intel_tv_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
 	if (color_conversion) {
 		I915_WRITE(TV_CSC_Y, (color_conversion->ry << 16) |
 			   color_conversion->gy);
-		I915_WRITE(TV_CSC_Y2, (color_conversion->by << 16) |
+		I915_WRITE(TV_CSC_Y2,(color_conversion->by << 16) |
 			   color_conversion->ay);
 		I915_WRITE(TV_CSC_U, (color_conversion->ru << 16) |
 			   color_conversion->gu);
@@ -1110,12 +1232,10 @@ static const struct drm_display_mode reported_modes[] = {
  * \return false if TV is disconnected.
  */
 static int
-intel_tv_detect_type(struct intel_tv *intel_tv,
+intel_tv_detect_type (struct intel_tv *intel_tv,
 		      struct drm_connector *connector)
 {
 	struct drm_encoder *encoder = &intel_tv->base.base;
-	struct drm_crtc *crtc = encoder->crtc;
-	struct intel_crtc *intel_crtc = to_intel_crtc(crtc);
 	struct drm_device *dev = encoder->dev;
 	struct drm_i915_private *dev_priv = dev->dev_private;
 	unsigned long irqflags;
@@ -1138,10 +1258,6 @@ intel_tv_detect_type(struct intel_tv *intel_tv,
 	/* Poll for TV detection */
 	tv_ctl &= ~(TV_ENC_ENABLE | TV_TEST_MODE_MASK);
 	tv_ctl |= TV_TEST_MODE_MONITOR_DETECT;
-	if (intel_crtc->pipe == 1)
-		tv_ctl |= TV_ENC_PIPEB_SELECT;
-	else
-		tv_ctl &= ~TV_ENC_PIPEB_SELECT;
 
 	tv_dac &= ~(TVDAC_SENSE_MASK | DAC_A_MASK | DAC_B_MASK | DAC_C_MASK);
 	tv_dac |= (TVDAC_STATE_CHG_EN |
@@ -1161,26 +1277,26 @@ intel_tv_detect_type(struct intel_tv *intel_tv,
 			      to_intel_crtc(intel_tv->base.base.crtc)->pipe);
 
 	type = -1;
-	tv_dac = I915_READ(TV_DAC);
-	DRM_DEBUG_KMS("TV detected: %x, %x\n", tv_ctl, tv_dac);
-	/*
-	 *  A B C
-	 *  0 1 1 Composite
-	 *  1 0 X svideo
-	 *  0 0 0 Component
-	 */
-	if ((tv_dac & TVDAC_SENSE_MASK) == (TVDAC_B_SENSE | TVDAC_C_SENSE)) {
-		DRM_DEBUG_KMS("Detected Composite TV connection\n");
-		type = DRM_MODE_CONNECTOR_Composite;
-	} else if ((tv_dac & (TVDAC_A_SENSE|TVDAC_B_SENSE)) == TVDAC_A_SENSE) {
-		DRM_DEBUG_KMS("Detected S-Video TV connection\n");
-		type = DRM_MODE_CONNECTOR_SVIDEO;
-	} else if ((tv_dac & TVDAC_SENSE_MASK) == 0) {
-		DRM_DEBUG_KMS("Detected Component TV connection\n");
-		type = DRM_MODE_CONNECTOR_Component;
-	} else {
-		DRM_DEBUG_KMS("Unrecognised TV connection\n");
-		type = -1;
+	if (wait_for((tv_dac = I915_READ(TV_DAC)) & TVDAC_STATE_CHG, 20) == 0) {
+		DRM_DEBUG_KMS("TV detected: %x, %x\n", tv_ctl, tv_dac);
+		/*
+		 *  A B C
+		 *  0 1 1 Composite
+		 *  1 0 X svideo
+		 *  0 0 0 Component
+		 */
+		if ((tv_dac & TVDAC_SENSE_MASK) == (TVDAC_B_SENSE | TVDAC_C_SENSE)) {
+			DRM_DEBUG_KMS("Detected Composite TV connection\n");
+			type = DRM_MODE_CONNECTOR_Composite;
+		} else if ((tv_dac & (TVDAC_A_SENSE|TVDAC_B_SENSE)) == TVDAC_A_SENSE) {
+			DRM_DEBUG_KMS("Detected S-Video TV connection\n");
+			type = DRM_MODE_CONNECTOR_SVIDEO;
+		} else if ((tv_dac & TVDAC_SENSE_MASK) == 0) {
+			DRM_DEBUG_KMS("Detected Component TV connection\n");
+			type = DRM_MODE_CONNECTOR_Component;
+		} else {
+			DRM_DEBUG_KMS("Unrecognised TV connection\n");
+		}
 	}
 
 	I915_WRITE(TV_DAC, save_tv_dac & ~TVDAC_STATE_CHG_EN);
@@ -1245,7 +1361,7 @@ intel_tv_detect(struct drm_connector *connector, bool force)
 	int type;
 
 	mode = reported_modes[0];
-	drm_mode_set_crtcinfo(&mode, 0);
+	drm_mode_set_crtcinfo(&mode, CRTC_INTERLACE_HALVE_V);
 
 	if (intel_tv->base.base.crtc && intel_tv->base.base.crtc->enabled) {
 		type = intel_tv_detect_type(intel_tv, connector);
@@ -1369,7 +1485,7 @@ intel_tv_get_modes(struct drm_connector *connector)
 }
 
 static void
-intel_tv_destroy(struct drm_connector *connector)
+intel_tv_destroy (struct drm_connector *connector)
 {
 	drm_sysfs_connector_remove(connector);
 	drm_connector_cleanup(connector);
diff --git a/drivers/gpu/drm/mga/mga_dma.c b/drivers/gpu/drm/mga/mga_dma.c
index 507aa3d..5ccb65de 100644
--- a/drivers/gpu/drm/mga/mga_dma.c
+++ b/drivers/gpu/drm/mga/mga_dma.c
@@ -403,8 +403,6 @@ int mga_driver_load(struct drm_device *dev, unsigned long flags)
 	dev_priv->usec_timeout = MGA_DEFAULT_USEC_TIMEOUT;
 	dev_priv->chipset = flags;
 
-	pci_set_master(dev->pdev);
-
 	dev_priv->mmio_base = pci_resource_start(dev->pdev, 1);
 	dev_priv->mmio_size = pci_resource_len(dev->pdev, 1);
 
diff --git a/drivers/gpu/drm/mga/mga_drv.c b/drivers/gpu/drm/mga/mga_drv.c
index f9a925d..42d3187 100644
--- a/drivers/gpu/drm/mga/mga_drv.c
+++ b/drivers/gpu/drm/mga/mga_drv.c
@@ -29,8 +29,6 @@
  *    Gareth Hughes <gareth@valinux.com>
  */
 
-#include <linux/module.h>
-
 #include "drmP.h"
 #include "drm.h"
 #include "mga_drm.h"
@@ -44,20 +42,6 @@ static struct pci_device_id pciidlist[] = {
 	mga_PCI_IDS
 };
 
-static const struct file_operations mga_driver_fops = {
-	.owner = THIS_MODULE,
-	.open = drm_open,
-	.release = drm_release,
-	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_mmap,
-	.poll = drm_poll,
-	.fasync = drm_fasync,
-#ifdef CONFIG_COMPAT
-	.compat_ioctl = mga_compat_ioctl,
-#endif
-	.llseek = noop_llseek,
-};
-
 static struct drm_driver driver = {
 	.driver_features =
 	    DRIVER_USE_AGP | DRIVER_USE_MTRR | DRIVER_PCI_DMA |
@@ -78,7 +62,20 @@ static struct drm_driver driver = {
 	.reclaim_buffers = drm_core_reclaim_buffers,
 	.ioctls = mga_ioctls,
 	.dma_ioctl = mga_dma_buffers,
-	.fops = &mga_driver_fops,
+	.fops = {
+		.owner = THIS_MODULE,
+		.open = drm_open,
+		.release = drm_release,
+		.unlocked_ioctl = drm_ioctl,
+		.mmap = drm_mmap,
+		.poll = drm_poll,
+		.fasync = drm_fasync,
+#ifdef CONFIG_COMPAT
+		.compat_ioctl = mga_compat_ioctl,
+#endif
+		.llseek = noop_llseek,
+	},
+
 	.name = DRIVER_NAME,
 	.desc = DRIVER_DESC,
 	.date = DRIVER_DATE,
diff --git a/drivers/gpu/drm/mga/mga_warp.c b/drivers/gpu/drm/mga/mga_warp.c
index 722a91b..f172bd5 100644
--- a/drivers/gpu/drm/mga/mga_warp.c
+++ b/drivers/gpu/drm/mga/mga_warp.c
@@ -30,7 +30,6 @@
 #include <linux/firmware.h>
 #include <linux/ihex.h>
 #include <linux/platform_device.h>
-#include <linux/module.h>
 
 #include "drmP.h"
 #include "drm.h"
diff --git a/drivers/gpu/drm/nouveau/Makefile b/drivers/gpu/drm/nouveau/Makefile
index 1a2ad7e..0583677 100644
--- a/drivers/gpu/drm/nouveau/Makefile
+++ b/drivers/gpu/drm/nouveau/Makefile
@@ -9,33 +9,28 @@ nouveau-y := nouveau_drv.o nouveau_state.o nouveau_channel.o nouveau_mem.o \
              nouveau_bo.o nouveau_fence.o nouveau_gem.o nouveau_ttm.o \
              nouveau_hw.o nouveau_calc.o nouveau_bios.o nouveau_i2c.o \
              nouveau_display.o nouveau_connector.o nouveau_fbcon.o \
-             nouveau_hdmi.o nouveau_dp.o nouveau_ramht.o \
+             nouveau_dp.o nouveau_ramht.o \
 	     nouveau_pm.o nouveau_volt.o nouveau_perf.o nouveau_temp.o \
-	     nouveau_mm.o nouveau_vm.o nouveau_mxm.o nouveau_gpio.o \
+	     nouveau_mm.o nouveau_vm.o \
              nv04_timer.o \
              nv04_mc.o nv40_mc.o nv50_mc.o \
-             nv04_fb.o nv10_fb.o nv20_fb.o nv30_fb.o nv40_fb.o \
-             nv50_fb.o nvc0_fb.o \
+             nv04_fb.o nv10_fb.o nv30_fb.o nv40_fb.o nv50_fb.o nvc0_fb.o \
              nv04_fifo.o nv10_fifo.o nv40_fifo.o nv50_fifo.o nvc0_fifo.o \
              nv04_graph.o nv10_graph.o nv20_graph.o \
              nv40_graph.o nv50_graph.o nvc0_graph.o \
              nv40_grctx.o nv50_grctx.o nvc0_grctx.o \
-             nv84_crypt.o nv98_crypt.o \
+             nv84_crypt.o \
              nva3_copy.o nvc0_copy.o \
-             nv31_mpeg.o nv50_mpeg.o \
-             nv84_bsp.o \
-             nv84_vp.o \
-             nv98_ppp.o \
+             nv40_mpeg.o nv50_mpeg.o \
              nv04_instmem.o nv50_instmem.o nvc0_instmem.o \
-             nv04_dac.o nv04_dfp.o nv04_tv.o nv17_tv.o nv17_tv_modes.o \
-             nv04_crtc.o nv04_display.o nv04_cursor.o \
              nv50_evo.o nv50_crtc.o nv50_dac.o nv50_sor.o \
              nv50_cursor.o nv50_display.o \
-             nvd0_display.o \
+             nv04_dac.o nv04_dfp.o nv04_tv.o nv17_tv.o nv17_tv_modes.o \
+             nv04_crtc.o nv04_display.o nv04_cursor.o \
              nv04_fbcon.o nv50_fbcon.o nvc0_fbcon.o \
              nv10_gpio.o nv50_gpio.o \
 	     nv50_calc.o \
-	     nv04_pm.o nv40_pm.o nv50_pm.o nva3_pm.o nvc0_pm.o \
+	     nv04_pm.o nv50_pm.o nva3_pm.o \
 	     nv50_vram.o nvc0_vram.o \
 	     nv50_vm.o nvc0_vm.o
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_acpi.c b/drivers/gpu/drm/nouveau/nouveau_acpi.c
index 7814a76..525744d 100644
--- a/drivers/gpu/drm/nouveau/nouveau_acpi.c
+++ b/drivers/gpu/drm/nouveau/nouveau_acpi.c
@@ -18,6 +18,12 @@
 
 #include <linux/vga_switcheroo.h>
 
+#define NOUVEAU_DSM_SUPPORTED 0x00
+#define NOUVEAU_DSM_SUPPORTED_FUNCTIONS 0x00
+
+#define NOUVEAU_DSM_ACTIVE 0x01
+#define NOUVEAU_DSM_ACTIVE_QUERY 0x00
+
 #define NOUVEAU_DSM_LED 0x02
 #define NOUVEAU_DSM_LED_STATE 0x00
 #define NOUVEAU_DSM_LED_OFF 0x10
@@ -29,9 +35,6 @@
 #define NOUVEAU_DSM_POWER_SPEED 0x01
 #define NOUVEAU_DSM_POWER_STAMINA 0x02
 
-#define NOUVEAU_DSM_OPTIMUS_FN 0x1A
-#define NOUVEAU_DSM_OPTIMUS_ARGS 0x03000001
-
 static struct nouveau_dsm_priv {
 	bool dsm_detected;
 	bool optimus_detected;
@@ -58,8 +61,7 @@ static int nouveau_optimus_dsm(acpi_handle handle, int func, int arg, uint32_t *
 	struct acpi_object_list input;
 	union acpi_object params[4];
 	union acpi_object *obj;
-	int i, err;
-	char args_buff[4];
+	int err;
 
 	input.count = 4;
 	input.pointer = params;
@@ -71,11 +73,7 @@ static int nouveau_optimus_dsm(acpi_handle handle, int func, int arg, uint32_t *
 	params[2].type = ACPI_TYPE_INTEGER;
 	params[2].integer.value = func;
 	params[3].type = ACPI_TYPE_BUFFER;
-	params[3].buffer.length = 4;
-	/* ACPI is little endian, AABBCCDD becomes {DD,CC,BB,AA} */
-	for (i = 0; i < 4; i++)
-		args_buff[i] = (arg >> i * 8) & 0xFF;
-	params[3].buffer.pointer = args_buff;
+	params[3].buffer.length = 0;
 
 	err = acpi_evaluate_object(handle, "_DSM", &input, &output);
 	if (err) {
@@ -150,23 +148,6 @@ static int nouveau_dsm(acpi_handle handle, int func, int arg, uint32_t *result)
 	return 0;
 }
 
-/* Returns 1 if a DSM function is usable and 0 otherwise */
-static int nouveau_test_dsm(acpi_handle test_handle,
-	int (*dsm_func)(acpi_handle, int, int, uint32_t *),
-	int sfnc)
-{
-	u32 result = 0;
-
-	/* Function 0 returns a Buffer containing available functions. The args
-	 * parameter is ignored for function 0, so just put 0 in it */
-	if (dsm_func(test_handle, 0, 0, &result))
-		return 0;
-
-	/* ACPI Spec v4 9.14.1: if bit 0 is zero, no function is supported. If
-	 * the n-th bit is enabled, function n is supported */
-	return result & 1 && result & (1 << sfnc);
-}
-
 static int nouveau_dsm_switch_mux(acpi_handle handle, int mux_id)
 {
 	mxm_wmi_call_mxmx(mux_id == NOUVEAU_DSM_LED_STAMINA ? MXM_MXDS_ADAPTER_IGD : MXM_MXDS_ADAPTER_0);
@@ -187,10 +168,6 @@ static int nouveau_dsm_set_discrete_state(acpi_handle handle, enum vga_switchero
 
 static int nouveau_dsm_switchto(enum vga_switcheroo_client_id id)
 {
-	/* perhaps the _DSM functions are mutually exclusive, but prepare for
-	 * the future */
-	if (!nouveau_dsm_priv.dsm_detected && nouveau_dsm_priv.optimus_detected)
-		return 0;
 	if (id == VGA_SWITCHEROO_IGD)
 		return nouveau_dsm_switch_mux(nouveau_dsm_priv.dhandle, NOUVEAU_DSM_LED_STAMINA);
 	else
@@ -203,11 +180,6 @@ static int nouveau_dsm_power_state(enum vga_switcheroo_client_id id,
 	if (id == VGA_SWITCHEROO_IGD)
 		return 0;
 
-	/* Optimus laptops have the card already disabled in
-	 * nouveau_switcheroo_set_state */
-	if (!nouveau_dsm_priv.dsm_detected && nouveau_dsm_priv.optimus_detected)
-		return 0;
-
 	return nouveau_dsm_set_discrete_state(nouveau_dsm_priv.dhandle, state);
 }
 
@@ -240,7 +212,8 @@ static int nouveau_dsm_pci_probe(struct pci_dev *pdev)
 {
 	acpi_handle dhandle, nvidia_handle;
 	acpi_status status;
-	int retval = 0;
+	int ret, retval = 0;
+	uint32_t result;
 
 	dhandle = DEVICE_ACPI_HANDLE(&pdev->dev);
 	if (!dhandle)
@@ -251,11 +224,13 @@ static int nouveau_dsm_pci_probe(struct pci_dev *pdev)
 		return false;
 	}
 
-	if (nouveau_test_dsm(dhandle, nouveau_dsm, NOUVEAU_DSM_POWER))
+	ret = nouveau_dsm(dhandle, NOUVEAU_DSM_SUPPORTED,
+			  NOUVEAU_DSM_SUPPORTED_FUNCTIONS, &result);
+	if (ret == 0)
 		retval |= NOUVEAU_DSM_HAS_MUX;
 
-	if (nouveau_test_dsm(dhandle, nouveau_optimus_dsm,
-		NOUVEAU_DSM_OPTIMUS_FN))
+	ret = nouveau_optimus_dsm(dhandle, 0, 0, &result);
+	if (ret == 0)
 		retval |= NOUVEAU_DSM_HAS_OPT;
 
 	if (retval)
@@ -294,22 +269,15 @@ static bool nouveau_dsm_detect(void)
 	}
 
 	if (vga_count == 2 && has_dsm && guid_valid) {
-		acpi_get_name(nouveau_dsm_priv.dhandle, ACPI_FULL_PATHNAME,
-			&buffer);
+		acpi_get_name(nouveau_dsm_priv.dhandle, ACPI_FULL_PATHNAME, &buffer);
 		printk(KERN_INFO "VGA switcheroo: detected DSM switching method %s handle\n",
-			acpi_method_name);
+		       acpi_method_name);
 		nouveau_dsm_priv.dsm_detected = true;
 		ret = true;
 	}
 
-	if (has_optimus == 1) {
-		acpi_get_name(nouveau_dsm_priv.dhandle, ACPI_FULL_PATHNAME,
-			&buffer);
-		printk(KERN_INFO "VGA switcheroo: detected Optimus DSM method %s handle\n",
-			acpi_method_name);
+	if (has_optimus == 1)
 		nouveau_dsm_priv.optimus_detected = true;
-		ret = true;
-	}
 
 	return ret;
 }
@@ -325,17 +293,6 @@ void nouveau_register_dsm_handler(void)
 	vga_switcheroo_register_handler(&nouveau_dsm_handler);
 }
 
-/* Must be called for Optimus models before the card can be turned off */
-void nouveau_switcheroo_optimus_dsm(void)
-{
-	u32 result = 0;
-	if (!nouveau_dsm_priv.optimus_detected)
-		return;
-
-	nouveau_optimus_dsm(nouveau_dsm_priv.dhandle, NOUVEAU_DSM_OPTIMUS_FN,
-		NOUVEAU_DSM_OPTIMUS_ARGS, &result);
-}
-
 void nouveau_unregister_dsm_handler(void)
 {
 	vga_switcheroo_unregister_handler();
diff --git a/drivers/gpu/drm/nouveau/nouveau_backlight.c b/drivers/gpu/drm/nouveau/nouveau_backlight.c
index fa22b28..00a55df 100644
--- a/drivers/gpu/drm/nouveau/nouveau_backlight.c
+++ b/drivers/gpu/drm/nouveau/nouveau_backlight.c
@@ -37,10 +37,8 @@
 #include "nouveau_drv.h"
 #include "nouveau_drm.h"
 #include "nouveau_reg.h"
-#include "nouveau_encoder.h"
 
-static int
-nv40_get_intensity(struct backlight_device *bd)
+static int nv40_get_intensity(struct backlight_device *bd)
 {
 	struct drm_device *dev = bl_get_data(bd);
 	int val = (nv_rd32(dev, NV40_PMC_BACKLIGHT) & NV40_PMC_BACKLIGHT_MASK)
@@ -49,8 +47,7 @@ nv40_get_intensity(struct backlight_device *bd)
 	return val;
 }
 
-static int
-nv40_set_intensity(struct backlight_device *bd)
+static int nv40_set_intensity(struct backlight_device *bd)
 {
 	struct drm_device *dev = bl_get_data(bd);
 	int val = bd->props.brightness;
@@ -68,8 +65,30 @@ static const struct backlight_ops nv40_bl_ops = {
 	.update_status = nv40_set_intensity,
 };
 
-static int
-nv40_backlight_init(struct drm_connector *connector)
+static int nv50_get_intensity(struct backlight_device *bd)
+{
+	struct drm_device *dev = bl_get_data(bd);
+
+	return nv_rd32(dev, NV50_PDISPLAY_SOR_BACKLIGHT);
+}
+
+static int nv50_set_intensity(struct backlight_device *bd)
+{
+	struct drm_device *dev = bl_get_data(bd);
+	int val = bd->props.brightness;
+
+	nv_wr32(dev, NV50_PDISPLAY_SOR_BACKLIGHT,
+		val | NV50_PDISPLAY_SOR_BACKLIGHT_ENABLE);
+	return 0;
+}
+
+static const struct backlight_ops nv50_bl_ops = {
+	.options = BL_CORE_SUSPENDRESUME,
+	.get_brightness = nv50_get_intensity,
+	.update_status = nv50_set_intensity,
+};
+
+static int nouveau_nv40_backlight_init(struct drm_connector *connector)
 {
 	struct drm_device *dev = connector->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
@@ -94,129 +113,34 @@ nv40_backlight_init(struct drm_connector *connector)
 	return 0;
 }
 
-static int
-nv50_get_intensity(struct backlight_device *bd)
-{
-	struct nouveau_encoder *nv_encoder = bl_get_data(bd);
-	struct drm_device *dev = nv_encoder->base.base.dev;
-	int or = nv_encoder->or;
-	u32 div = 1025;
-	u32 val;
-
-	val  = nv_rd32(dev, NV50_PDISP_SOR_PWM_CTL(or));
-	val &= NV50_PDISP_SOR_PWM_CTL_VAL;
-	return ((val * 100) + (div / 2)) / div;
-}
-
-static int
-nv50_set_intensity(struct backlight_device *bd)
-{
-	struct nouveau_encoder *nv_encoder = bl_get_data(bd);
-	struct drm_device *dev = nv_encoder->base.base.dev;
-	int or = nv_encoder->or;
-	u32 div = 1025;
-	u32 val = (bd->props.brightness * div) / 100;
-
-	nv_wr32(dev, NV50_PDISP_SOR_PWM_CTL(or),
-		     NV50_PDISP_SOR_PWM_CTL_NEW | val);
-	return 0;
-}
-
-static const struct backlight_ops nv50_bl_ops = {
-	.options = BL_CORE_SUSPENDRESUME,
-	.get_brightness = nv50_get_intensity,
-	.update_status = nv50_set_intensity,
-};
-
-static int
-nva3_get_intensity(struct backlight_device *bd)
-{
-	struct nouveau_encoder *nv_encoder = bl_get_data(bd);
-	struct drm_device *dev = nv_encoder->base.base.dev;
-	int or = nv_encoder->or;
-	u32 div, val;
-
-	div  = nv_rd32(dev, NV50_PDISP_SOR_PWM_DIV(or));
-	val  = nv_rd32(dev, NV50_PDISP_SOR_PWM_CTL(or));
-	val &= NVA3_PDISP_SOR_PWM_CTL_VAL;
-	if (div && div >= val)
-		return ((val * 100) + (div / 2)) / div;
-
-	return 100;
-}
-
-static int
-nva3_set_intensity(struct backlight_device *bd)
-{
-	struct nouveau_encoder *nv_encoder = bl_get_data(bd);
-	struct drm_device *dev = nv_encoder->base.base.dev;
-	int or = nv_encoder->or;
-	u32 div, val;
-
-	div = nv_rd32(dev, NV50_PDISP_SOR_PWM_DIV(or));
-	val = (bd->props.brightness * div) / 100;
-	if (div) {
-		nv_wr32(dev, NV50_PDISP_SOR_PWM_CTL(or), val |
-			     NV50_PDISP_SOR_PWM_CTL_NEW |
-			     NVA3_PDISP_SOR_PWM_CTL_UNK);
-		return 0;
-	}
-
-	return -EINVAL;
-}
-
-static const struct backlight_ops nva3_bl_ops = {
-	.options = BL_CORE_SUSPENDRESUME,
-	.get_brightness = nva3_get_intensity,
-	.update_status = nva3_set_intensity,
-};
-
-static int
-nv50_backlight_init(struct drm_connector *connector)
+static int nouveau_nv50_backlight_init(struct drm_connector *connector)
 {
 	struct drm_device *dev = connector->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_encoder *nv_encoder;
 	struct backlight_properties props;
 	struct backlight_device *bd;
-	const struct backlight_ops *ops;
-
-	nv_encoder = find_encoder(connector, OUTPUT_LVDS);
-	if (!nv_encoder) {
-		nv_encoder = find_encoder(connector, OUTPUT_DP);
-		if (!nv_encoder)
-			return -ENODEV;
-	}
 
-	if (!nv_rd32(dev, NV50_PDISP_SOR_PWM_CTL(nv_encoder->or)))
+	if (!nv_rd32(dev, NV50_PDISPLAY_SOR_BACKLIGHT))
 		return 0;
 
-	if (dev_priv->chipset <= 0xa0 ||
-	    dev_priv->chipset == 0xaa ||
-	    dev_priv->chipset == 0xac)
-		ops = &nv50_bl_ops;
-	else
-		ops = &nva3_bl_ops;
-
 	memset(&props, 0, sizeof(struct backlight_properties));
 	props.type = BACKLIGHT_RAW;
-	props.max_brightness = 100;
-	bd = backlight_device_register("nv_backlight", &connector->kdev,
-				       nv_encoder, ops, &props);
+	props.max_brightness = 1025;
+	bd = backlight_device_register("nv_backlight", &connector->kdev, dev,
+				       &nv50_bl_ops, &props);
 	if (IS_ERR(bd))
 		return PTR_ERR(bd);
 
 	dev_priv->backlight = bd;
-	bd->props.brightness = bd->ops->get_brightness(bd);
+	bd->props.brightness = nv50_get_intensity(bd);
 	backlight_update_status(bd);
 	return 0;
 }
 
-int
-nouveau_backlight_init(struct drm_device *dev)
+int nouveau_backlight_init(struct drm_connector *connector)
 {
+	struct drm_device *dev = connector->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct drm_connector *connector;
 
 #ifdef CONFIG_ACPI
 	if (acpi_video_backlight_support()) {
@@ -226,28 +150,21 @@ nouveau_backlight_init(struct drm_device *dev)
 	}
 #endif
 
-	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
-		if (connector->connector_type != DRM_MODE_CONNECTOR_LVDS &&
-		    connector->connector_type != DRM_MODE_CONNECTOR_eDP)
-			continue;
-
-		switch (dev_priv->card_type) {
-		case NV_40:
-			return nv40_backlight_init(connector);
-		case NV_50:
-			return nv50_backlight_init(connector);
-		default:
-			break;
-		}
+	switch (dev_priv->card_type) {
+	case NV_40:
+		return nouveau_nv40_backlight_init(connector);
+	case NV_50:
+		return nouveau_nv50_backlight_init(connector);
+	default:
+		break;
 	}
 
-
 	return 0;
 }
 
-void
-nouveau_backlight_exit(struct drm_device *dev)
+void nouveau_backlight_exit(struct drm_connector *connector)
 {
+	struct drm_device *dev = connector->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 
 	if (dev_priv->backlight) {
diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.c b/drivers/gpu/drm/nouveau/nouveau_bios.c
index 8dbeeea..729d5fd 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bios.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bios.c
@@ -27,7 +27,6 @@
 #include "nouveau_drv.h"
 #include "nouveau_hw.h"
 #include "nouveau_encoder.h"
-#include "nouveau_gpio.h"
 
 #include <linux/io-mapping.h>
 
@@ -35,6 +34,9 @@
 #define NV_CIO_CRE_44_HEADA 0x0
 #define NV_CIO_CRE_44_HEADB 0x3
 #define FEATURE_MOBILE 0x10	/* also FEATURE_QUADRO for BMP */
+#define LEGACY_I2C_CRT 0x80
+#define LEGACY_I2C_PANEL 0x81
+#define LEGACY_I2C_TV 0x82
 
 #define EDID1_LEN 128
 
@@ -65,232 +67,194 @@ static bool nv_cksum(const uint8_t *data, unsigned int length)
 }
 
 static int
-score_vbios(struct nvbios *bios, const bool writeable)
+score_vbios(struct drm_device *dev, const uint8_t *data, const bool writeable)
 {
-	if (!bios->data || bios->data[0] != 0x55 || bios->data[1] != 0xAA) {
-		NV_TRACEWARN(bios->dev, "... BIOS signature not found\n");
+	if (!(data[0] == 0x55 && data[1] == 0xAA)) {
+		NV_TRACEWARN(dev, "... BIOS signature not found\n");
 		return 0;
 	}
 
-	if (nv_cksum(bios->data, bios->data[2] * 512)) {
-		NV_TRACEWARN(bios->dev, "... BIOS checksum invalid\n");
+	if (nv_cksum(data, data[2] * 512)) {
+		NV_TRACEWARN(dev, "... BIOS checksum invalid\n");
 		/* if a ro image is somewhat bad, it's probably all rubbish */
 		return writeable ? 2 : 1;
-	}
+	} else
+		NV_TRACE(dev, "... appears to be valid\n");
 
-	NV_TRACE(bios->dev, "... appears to be valid\n");
 	return 3;
 }
 
-static void
-bios_shadow_prom(struct nvbios *bios)
+static void load_vbios_prom(struct drm_device *dev, uint8_t *data)
 {
-	struct drm_device *dev = bios->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	u32 pcireg, access;
-	u16 pcir;
+	uint32_t pci_nv_20, save_pci_nv_20;
+	int pcir_ptr;
 	int i;
 
-	/* enable access to rom */
 	if (dev_priv->card_type >= NV_50)
-		pcireg = 0x088050;
+		pci_nv_20 = 0x88050;
 	else
-		pcireg = NV_PBUS_PCI_NV_20;
-	access = nv_mask(dev, pcireg, 0x00000001, 0x00000000);
+		pci_nv_20 = NV_PBUS_PCI_NV_20;
 
-	/* bail if no rom signature, with a workaround for a PROM reading
-	 * issue on some chipsets.  the first read after a period of
-	 * inactivity returns the wrong result, so retry the first header
-	 * byte a few times before giving up as a workaround
-	 */
-	i = 16;
-	do {
-		if (nv_rd08(dev, NV_PROM_OFFSET + 0) == 0x55)
-			break;
-	} while (i--);
+	/* enable ROM access */
+	save_pci_nv_20 = nvReadMC(dev, pci_nv_20);
+	nvWriteMC(dev, pci_nv_20,
+		  save_pci_nv_20 & ~NV_PBUS_PCI_NV_20_ROM_SHADOW_ENABLED);
 
-	if (!i || nv_rd08(dev, NV_PROM_OFFSET + 1) != 0xaa)
+	/* bail if no rom signature */
+	if (nv_rd08(dev, NV_PROM_OFFSET) != 0x55 ||
+	    nv_rd08(dev, NV_PROM_OFFSET + 1) != 0xaa)
 		goto out;
 
 	/* additional check (see note below) - read PCI record header */
-	pcir = nv_rd08(dev, NV_PROM_OFFSET + 0x18) |
-	       nv_rd08(dev, NV_PROM_OFFSET + 0x19) << 8;
-	if (nv_rd08(dev, NV_PROM_OFFSET + pcir + 0) != 'P' ||
-	    nv_rd08(dev, NV_PROM_OFFSET + pcir + 1) != 'C' ||
-	    nv_rd08(dev, NV_PROM_OFFSET + pcir + 2) != 'I' ||
-	    nv_rd08(dev, NV_PROM_OFFSET + pcir + 3) != 'R')
+	pcir_ptr = nv_rd08(dev, NV_PROM_OFFSET + 0x18) |
+		   nv_rd08(dev, NV_PROM_OFFSET + 0x19) << 8;
+	if (nv_rd08(dev, NV_PROM_OFFSET + pcir_ptr) != 'P' ||
+	    nv_rd08(dev, NV_PROM_OFFSET + pcir_ptr + 1) != 'C' ||
+	    nv_rd08(dev, NV_PROM_OFFSET + pcir_ptr + 2) != 'I' ||
+	    nv_rd08(dev, NV_PROM_OFFSET + pcir_ptr + 3) != 'R')
 		goto out;
 
-	/* read entire bios image to system memory */
-	bios->length = nv_rd08(dev, NV_PROM_OFFSET + 2) * 512;
-	bios->data = kmalloc(bios->length, GFP_KERNEL);
-	if (bios->data) {
-		for (i = 0; i < bios->length; i++)
-			bios->data[i] = nv_rd08(dev, NV_PROM_OFFSET + i);
-	}
+	/* on some 6600GT/6800LE prom reads are messed up.  nvclock alleges a
+	 * a good read may be obtained by waiting or re-reading (cargocult: 5x)
+	 * each byte.  we'll hope pramin has something usable instead
+	 */
+	for (i = 0; i < NV_PROM_SIZE; i++)
+		data[i] = nv_rd08(dev, NV_PROM_OFFSET + i);
 
 out:
-	/* disable access to rom */
-	nv_wr32(dev, pcireg, access);
+	/* disable ROM access */
+	nvWriteMC(dev, pci_nv_20,
+		  save_pci_nv_20 | NV_PBUS_PCI_NV_20_ROM_SHADOW_ENABLED);
 }
 
-static void
-bios_shadow_pramin(struct nvbios *bios)
+static void load_vbios_pramin(struct drm_device *dev, uint8_t *data)
 {
-	struct drm_device *dev = bios->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	u32 bar0 = 0;
+	uint32_t old_bar0_pramin = 0;
 	int i;
 
 	if (dev_priv->card_type >= NV_50) {
-		u64 addr = (u64)(nv_rd32(dev, 0x619f04) & 0xffffff00) << 8;
-		if (!addr) {
-			addr  = (u64)nv_rd32(dev, 0x001700) << 16;
-			addr += 0xf0000;
-		}
+		uint32_t vbios_vram = (nv_rd32(dev, 0x619f04) & ~0xff) << 8;
+
+		if (!vbios_vram)
+			vbios_vram = (nv_rd32(dev, 0x1700) << 16) + 0xf0000;
 
-		bar0 = nv_mask(dev, 0x001700, 0xffffffff, addr >> 16);
+		old_bar0_pramin = nv_rd32(dev, 0x1700);
+		nv_wr32(dev, 0x1700, vbios_vram >> 16);
 	}
 
 	/* bail if no rom signature */
-	if (nv_rd08(dev, NV_PRAMIN_OFFSET + 0) != 0x55 ||
+	if (nv_rd08(dev, NV_PRAMIN_OFFSET) != 0x55 ||
 	    nv_rd08(dev, NV_PRAMIN_OFFSET + 1) != 0xaa)
 		goto out;
 
-	bios->length = nv_rd08(dev, NV_PRAMIN_OFFSET + 2) * 512;
-	bios->data = kmalloc(bios->length, GFP_KERNEL);
-	if (bios->data) {
-		for (i = 0; i < bios->length; i++)
-			bios->data[i] = nv_rd08(dev, NV_PRAMIN_OFFSET + i);
-	}
+	for (i = 0; i < NV_PROM_SIZE; i++)
+		data[i] = nv_rd08(dev, NV_PRAMIN_OFFSET + i);
 
 out:
 	if (dev_priv->card_type >= NV_50)
-		nv_wr32(dev, 0x001700, bar0);
+		nv_wr32(dev, 0x1700, old_bar0_pramin);
 }
 
-static void
-bios_shadow_pci(struct nvbios *bios)
-{
-	struct pci_dev *pdev = bios->dev->pdev;
-	size_t length;
-
-	if (!pci_enable_rom(pdev)) {
-		void __iomem *rom = pci_map_rom(pdev, &length);
-		if (rom) {
-			bios->data = kmalloc(length, GFP_KERNEL);
-			if (bios->data) {
-				memcpy_fromio(bios->data, rom, length);
-				bios->length = length;
-			}
-			pci_unmap_rom(pdev, rom);
-		}
-
-		pci_disable_rom(pdev);
-	}
-}
-
-static void
-bios_shadow_acpi(struct nvbios *bios)
+static void load_vbios_pci(struct drm_device *dev, uint8_t *data)
 {
-	struct pci_dev *pdev = bios->dev->pdev;
-	int ptr, len, ret;
-	u8 data[3];
+	void __iomem *rom = NULL;
+	size_t rom_len;
+	int ret;
 
-	if (!nouveau_acpi_rom_supported(pdev))
+	ret = pci_enable_rom(dev->pdev);
+	if (ret)
 		return;
 
-	ret = nouveau_acpi_get_bios_chunk(data, 0, sizeof(data));
-	if (ret != sizeof(data))
-		return;
+	rom = pci_map_rom(dev->pdev, &rom_len);
+	if (!rom)
+		goto out;
+	memcpy_fromio(data, rom, rom_len);
+	pci_unmap_rom(dev->pdev, rom);
 
-	bios->length = min(data[2] * 512, 65536);
-	bios->data = kmalloc(bios->length, GFP_KERNEL);
-	if (!bios->data)
-		return;
+out:
+	pci_disable_rom(dev->pdev);
+}
 
-	len = bios->length;
-	ptr = 0;
-	while (len) {
-		int size = (len > ROM_BIOS_PAGE) ? ROM_BIOS_PAGE : len;
+static void load_vbios_acpi(struct drm_device *dev, uint8_t *data)
+{
+	int i;
+	int ret;
+	int size = 64 * 1024;
 
-		ret = nouveau_acpi_get_bios_chunk(bios->data, ptr, size);
-		if (ret != size) {
-			kfree(bios->data);
-			bios->data = NULL;
-			return;
-		}
+	if (!nouveau_acpi_rom_supported(dev->pdev))
+		return;
 
-		len -= size;
-		ptr += size;
+	for (i = 0; i < (size / ROM_BIOS_PAGE); i++) {
+		ret = nouveau_acpi_get_bios_chunk(data,
+						  (i * ROM_BIOS_PAGE),
+						  ROM_BIOS_PAGE);
+		if (ret <= 0)
+			break;
 	}
+	return;
 }
 
 struct methods {
 	const char desc[8];
-	void (*shadow)(struct nvbios *);
+	void (*loadbios)(struct drm_device *, uint8_t *);
 	const bool rw;
-	int score;
-	u32 size;
-	u8 *data;
 };
 
-static bool
-bios_shadow(struct drm_device *dev)
-{
-	struct methods shadow_methods[] = {
-		{ "PRAMIN", bios_shadow_pramin, true, 0, 0, NULL },
-		{ "PROM", bios_shadow_prom, false, 0, 0, NULL },
-		{ "ACPI", bios_shadow_acpi, true, 0, 0, NULL },
-		{ "PCIROM", bios_shadow_pci, true, 0, 0, NULL },
-		{}
-	};
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nvbios *bios = &dev_priv->vbios;
-	struct methods *mthd, *best;
+static struct methods shadow_methods[] = {
+	{ "PRAMIN", load_vbios_pramin, true },
+	{ "PROM", load_vbios_prom, false },
+	{ "PCIROM", load_vbios_pci, true },
+	{ "ACPI", load_vbios_acpi, true },
+};
+#define NUM_SHADOW_METHODS ARRAY_SIZE(shadow_methods)
+
+static bool NVShadowVBIOS(struct drm_device *dev, uint8_t *data)
+{
+	struct methods *methods = shadow_methods;
+	int testscore = 3;
+	int scores[NUM_SHADOW_METHODS], i;
 
 	if (nouveau_vbios) {
-		mthd = shadow_methods;
-		do {
-			if (strcasecmp(nouveau_vbios, mthd->desc))
-				continue;
-			NV_INFO(dev, "VBIOS source: %s\n", mthd->desc);
+		for (i = 0; i < NUM_SHADOW_METHODS; i++)
+			if (!strcasecmp(nouveau_vbios, methods[i].desc))
+				break;
 
-			mthd->shadow(bios);
-			mthd->score = score_vbios(bios, mthd->rw);
-			if (mthd->score)
+		if (i < NUM_SHADOW_METHODS) {
+			NV_INFO(dev, "Attempting to use BIOS image from %s\n",
+				methods[i].desc);
+
+			methods[i].loadbios(dev, data);
+			if (score_vbios(dev, data, methods[i].rw))
 				return true;
-		} while ((++mthd)->shadow);
+		}
 
 		NV_ERROR(dev, "VBIOS source \'%s\' invalid\n", nouveau_vbios);
 	}
 
-	mthd = shadow_methods;
-	do {
-		NV_TRACE(dev, "Checking %s for VBIOS\n", mthd->desc);
-		mthd->shadow(bios);
-		mthd->score = score_vbios(bios, mthd->rw);
-		mthd->size = bios->length;
-		mthd->data = bios->data;
-	} while (mthd->score != 3 && (++mthd)->shadow);
-
-	mthd = shadow_methods;
-	best = mthd;
-	do {
-		if (mthd->score > best->score) {
-			kfree(best->data);
-			best = mthd;
-		}
-	} while ((++mthd)->shadow);
+	for (i = 0; i < NUM_SHADOW_METHODS; i++) {
+		NV_TRACE(dev, "Attempting to load BIOS image from %s\n",
+			 methods[i].desc);
+		data[0] = data[1] = 0;	/* avoid reuse of previous image */
+		methods[i].loadbios(dev, data);
+		scores[i] = score_vbios(dev, data, methods[i].rw);
+		if (scores[i] == testscore)
+			return true;
+	}
 
-	if (best->score) {
-		NV_TRACE(dev, "Using VBIOS from %s\n", best->desc);
-		bios->length = best->size;
-		bios->data = best->data;
-		return true;
+	while (--testscore > 0) {
+		for (i = 0; i < NUM_SHADOW_METHODS; i++) {
+			if (scores[i] == testscore) {
+				NV_TRACE(dev, "Using BIOS image from %s\n",
+					 methods[i].desc);
+				methods[i].loadbios(dev, data);
+				return true;
+			}
+		}
 	}
 
-	NV_ERROR(dev, "No valid VBIOS image found\n");
+	NV_ERROR(dev, "No valid BIOS image found\n");
 	return false;
 }
 
@@ -331,11 +295,6 @@ munge_reg(struct nvbios *bios, uint32_t reg)
 	if (dev_priv->card_type < NV_50)
 		return reg;
 
-	if (reg & 0x80000000) {
-		BUG_ON(bios->display.crtc < 0);
-		reg += bios->display.crtc * 0x800;
-	}
-
 	if (reg & 0x40000000) {
 		BUG_ON(!dcbent);
 
@@ -344,7 +303,7 @@ munge_reg(struct nvbios *bios, uint32_t reg)
 			reg += 0x00000080;
 	}
 
-	reg &= ~0xe0000000;
+	reg &= ~0x60000000;
 	return reg;
 }
 
@@ -675,9 +634,10 @@ static int
 nv50_pll_set(struct drm_device *dev, uint32_t reg, uint32_t clk)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	uint32_t reg0 = nv_rd32(dev, reg + 0);
+	uint32_t reg1 = nv_rd32(dev, reg + 4);
 	struct nouveau_pll_vals pll;
 	struct pll_lims pll_limits;
-	u32 ctrl, mask, coef;
 	int ret;
 
 	ret = get_pll_limits(dev, reg, &pll_limits);
@@ -688,20 +648,15 @@ nv50_pll_set(struct drm_device *dev, uint32_t reg, uint32_t clk)
 	if (!clk)
 		return -ERANGE;
 
-	coef = pll.N1 << 8 | pll.M1;
-	ctrl = pll.log2P << 16;
-	mask = 0x00070000;
-	if (reg == 0x004008) {
-		mask |= 0x01f80000;
-		ctrl |= (pll_limits.log2p_bias << 19);
-		ctrl |= (pll.log2P << 22);
-	}
+	reg0 = (reg0 & 0xfff8ffff) | (pll.log2P << 16);
+	reg1 = (reg1 & 0xffff0000) | (pll.N1 << 8) | pll.M1;
 
-	if (!dev_priv->vbios.execute)
-		return 0;
+	if (dev_priv->vbios.execute) {
+		still_alive();
+		nv_wr32(dev, reg + 4, reg1);
+		nv_wr32(dev, reg + 0, reg0);
+	}
 
-	nv_mask(dev, reg + 0, mask, ctrl);
-	nv_wr32(dev, reg + 4, coef);
 	return 0;
 }
 
@@ -758,19 +713,115 @@ static int dcb_entry_idx_from_crtchead(struct drm_device *dev)
 	return dcb_entry;
 }
 
+static int
+read_dcb_i2c_entry(struct drm_device *dev, int dcb_version, uint8_t *i2ctable, int index, struct dcb_i2c_entry *i2c)
+{
+	uint8_t dcb_i2c_ver = dcb_version, headerlen = 0, entry_len = 4;
+	int i2c_entries = DCB_MAX_NUM_I2C_ENTRIES;
+	int recordoffset = 0, rdofs = 1, wrofs = 0;
+	uint8_t port_type = 0;
+
+	if (!i2ctable)
+		return -EINVAL;
+
+	if (dcb_version >= 0x30) {
+		if (i2ctable[0] != dcb_version) /* necessary? */
+			NV_WARN(dev,
+				"DCB I2C table version mismatch (%02X vs %02X)\n",
+				i2ctable[0], dcb_version);
+		dcb_i2c_ver = i2ctable[0];
+		headerlen = i2ctable[1];
+		if (i2ctable[2] <= DCB_MAX_NUM_I2C_ENTRIES)
+			i2c_entries = i2ctable[2];
+		else
+			NV_WARN(dev,
+				"DCB I2C table has more entries than indexable "
+				"(%d entries, max %d)\n", i2ctable[2],
+				DCB_MAX_NUM_I2C_ENTRIES);
+		entry_len = i2ctable[3];
+		/* [4] is i2c_default_indices, read in parse_dcb_table() */
+	}
+	/*
+	 * It's your own fault if you call this function on a DCB 1.1 BIOS --
+	 * the test below is for DCB 1.2
+	 */
+	if (dcb_version < 0x14) {
+		recordoffset = 2;
+		rdofs = 0;
+		wrofs = 1;
+	}
+
+	if (index == 0xf)
+		return 0;
+	if (index >= i2c_entries) {
+		NV_ERROR(dev, "DCB I2C index too big (%d >= %d)\n",
+			 index, i2ctable[2]);
+		return -ENOENT;
+	}
+	if (i2ctable[headerlen + entry_len * index + 3] == 0xff) {
+		NV_ERROR(dev, "DCB I2C entry invalid\n");
+		return -EINVAL;
+	}
+
+	if (dcb_i2c_ver >= 0x30) {
+		port_type = i2ctable[headerlen + recordoffset + 3 + entry_len * index];
+
+		/*
+		 * Fixup for chips using same address offset for read and
+		 * write.
+		 */
+		if (port_type == 4)	/* seen on C51 */
+			rdofs = wrofs = 1;
+		if (port_type >= 5)	/* G80+ */
+			rdofs = wrofs = 0;
+	}
+
+	if (dcb_i2c_ver >= 0x40) {
+		if (port_type != 5 && port_type != 6)
+			NV_WARN(dev, "DCB I2C table has port type %d\n", port_type);
+
+		i2c->entry = ROM32(i2ctable[headerlen + recordoffset + entry_len * index]);
+	}
+
+	i2c->port_type = port_type;
+	i2c->read = i2ctable[headerlen + recordoffset + rdofs + entry_len * index];
+	i2c->write = i2ctable[headerlen + recordoffset + wrofs + entry_len * index];
+
+	return 0;
+}
+
 static struct nouveau_i2c_chan *
 init_i2c_device_find(struct drm_device *dev, int i2c_index)
 {
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct dcb_table *dcb = &dev_priv->vbios.dcb;
+
 	if (i2c_index == 0xff) {
-		struct drm_nouveau_private *dev_priv = dev->dev_private;
-		struct dcb_table *dcb = &dev_priv->vbios.dcb;
 		/* note: dcb_entry_idx_from_crtchead needs pre-script set-up */
-		int idx = dcb_entry_idx_from_crtchead(dev);
+		int idx = dcb_entry_idx_from_crtchead(dev), shift = 0;
+		int default_indices = dcb->i2c_default_indices;
 
-		i2c_index = NV_I2C_DEFAULT(0);
 		if (idx != 0x7f && dcb->entry[idx].i2c_upper_default)
-			i2c_index = NV_I2C_DEFAULT(1);
+			shift = 4;
+
+		i2c_index = (default_indices >> shift) & 0xf;
 	}
+	if (i2c_index == 0x80)	/* g80+ */
+		i2c_index = dcb->i2c_default_indices & 0xf;
+	else
+	if (i2c_index == 0x81)
+		i2c_index = (dcb->i2c_default_indices & 0xf0) >> 4;
+
+	if (i2c_index >= DCB_MAX_NUM_I2C_ENTRIES) {
+		NV_ERROR(dev, "invalid i2c_index 0x%x\n", i2c_index);
+		return NULL;
+	}
+
+	/* Make sure i2c table entry has been parsed, it may not
+	 * have been if this is a bus not referenced by a DCB encoder
+	 */
+	read_dcb_i2c_entry(dev, dcb->version, dcb->i2c_table,
+			   i2c_index, &dcb->i2c[i2c_index]);
 
 	return nouveau_i2c_find(dev, i2c_index);
 }
@@ -1122,29 +1173,36 @@ init_dp_condition(struct nvbios *bios, uint16_t offset, struct init_exec *iexec)
 	 *
 	 */
 
+	struct bit_displayport_encoder_table *dpe = NULL;
 	struct dcb_entry *dcb = bios->display.output;
 	struct drm_device *dev = bios->dev;
 	uint8_t cond = bios->data[offset + 1];
-	uint8_t *table, *entry;
+	int dummy;
 
 	BIOSLOG(bios, "0x%04X: subop 0x%02X\n", offset, cond);
 
 	if (!iexec->execute)
 		return 3;
 
-	table = nouveau_dp_bios_data(dev, dcb, &entry);
-	if (!table)
+	dpe = nouveau_bios_dp_table(dev, dcb, &dummy);
+	if (!dpe) {
+		NV_ERROR(dev, "0x%04X: INIT_3A: no encoder table!!\n", offset);
 		return 3;
+	}
 
 	switch (cond) {
 	case 0:
-		entry = dcb_conn(dev, dcb->connector);
-		if (!entry || entry[0] != DCB_CONNECTOR_eDP)
+	{
+		struct dcb_connector_table_entry *ent =
+			&bios->dcb.connector.entry[dcb->connector];
+
+		if (ent->type != DCB_CONNECTOR_eDP)
 			iexec->execute = false;
+	}
 		break;
 	case 1:
 	case 2:
-		if (!(entry[5] & cond))
+		if (!(dpe->unknown & cond))
 			iexec->execute = false;
 		break;
 	case 5:
@@ -3174,8 +3232,49 @@ init_gpio(struct nvbios *bios, uint16_t offset, struct init_exec *iexec)
 	 * each GPIO according to various values listed in each entry
 	 */
 
-	if (iexec->execute && bios->execute)
-		nouveau_gpio_reset(bios->dev);
+	struct drm_nouveau_private *dev_priv = bios->dev->dev_private;
+	struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio;
+	const uint32_t nv50_gpio_ctl[2] = { 0xe100, 0xe28c };
+	int i;
+
+	if (dev_priv->card_type < NV_50) {
+		NV_ERROR(bios->dev, "INIT_GPIO on unsupported chipset\n");
+		return 1;
+	}
+
+	if (!iexec->execute)
+		return 1;
+
+	for (i = 0; i < bios->dcb.gpio.entries; i++) {
+		struct dcb_gpio_entry *gpio = &bios->dcb.gpio.entry[i];
+		uint32_t r, s, v;
+
+		BIOSLOG(bios, "0x%04X: Entry: 0x%08X\n", offset, gpio->entry);
+
+		BIOSLOG(bios, "0x%04X: set gpio 0x%02x, state %d\n",
+			offset, gpio->tag, gpio->state_default);
+		if (bios->execute)
+			pgpio->set(bios->dev, gpio->tag, gpio->state_default);
+
+		/* The NVIDIA binary driver doesn't appear to actually do
+		 * any of this, my VBIOS does however.
+		 */
+		/* Not a clue, needs de-magicing */
+		r = nv50_gpio_ctl[gpio->line >> 4];
+		s = (gpio->line & 0x0f);
+		v = bios_rd32(bios, r) & ~(0x00010001 << s);
+		switch ((gpio->entry & 0x06000000) >> 25) {
+		case 1:
+			v |= (0x00000001 << s);
+			break;
+		case 2:
+			v |= (0x00010000 << s);
+			break;
+		default:
+			break;
+		}
+		bios_wr32(bios, r, v);
+	}
 
 	return 1;
 }
@@ -3637,10 +3736,6 @@ parse_init_table(struct nvbios *bios, uint16_t offset, struct init_exec *iexec)
 	int count = 0, i, ret;
 	uint8_t id;
 
-	/* catch NULL script pointers */
-	if (offset == 0)
-		return 0;
-
 	/*
 	 * Loop until INIT_DONE causes us to break out of the loop
 	 * (or until offset > bios length just in case... )
@@ -4272,6 +4367,18 @@ int nouveau_bios_parse_lvds_table(struct drm_device *dev, int pxclk, bool *dl, b
 		break;
 	}
 
+	/* Dell Latitude D620 reports a too-high value for the dual-link
+	 * transition freq, causing us to program the panel incorrectly.
+	 *
+	 * It doesn't appear the VBIOS actually uses its transition freq
+	 * (90000kHz), instead it uses the "Number of LVDS channels" field
+	 * out of the panel ID structure (http://www.spwg.org/).
+	 *
+	 * For the moment, a quirk will do :)
+	 */
+	if (nv_match_device(dev, 0x01d7, 0x1028, 0x01c2))
+		bios->fp.duallink_transition_clk = 80000;
+
 	/* set dual_link flag for EDID case */
 	if (pxclk && (chip_version < 0x25 || chip_version > 0x28))
 		bios->fp.dual_link = (pxclk >= bios->fp.duallink_transition_clk);
@@ -4281,37 +4388,86 @@ int nouveau_bios_parse_lvds_table(struct drm_device *dev, int pxclk, bool *dl, b
 	return 0;
 }
 
-/* BIT 'U'/'d' table encoder subtables have hashes matching them to
- * a particular set of encoders.
- *
- * This function returns true if a particular DCB entry matches.
- */
-bool
-bios_encoder_match(struct dcb_entry *dcb, u32 hash)
+static uint8_t *
+bios_output_config_match(struct drm_device *dev, struct dcb_entry *dcbent,
+			 uint16_t record, int record_len, int record_nr,
+			 bool match_link)
 {
-	if ((hash & 0x000000f0) != (dcb->location << 4))
-		return false;
-	if ((hash & 0x0000000f) != dcb->type)
-		return false;
-	if (!(hash & (dcb->or << 16)))
-		return false;
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nvbios *bios = &dev_priv->vbios;
+	uint32_t entry;
+	uint16_t table;
+	int i, v;
 
-	switch (dcb->type) {
+	switch (dcbent->type) {
 	case OUTPUT_TMDS:
 	case OUTPUT_LVDS:
 	case OUTPUT_DP:
-		if (hash & 0x00c00000) {
-			if (!(hash & (dcb->sorconf.link << 22)))
-				return false;
-		}
+		break;
 	default:
-		return true;
+		match_link = false;
+		break;
 	}
+
+	for (i = 0; i < record_nr; i++, record += record_len) {
+		table = ROM16(bios->data[record]);
+		if (!table)
+			continue;
+		entry = ROM32(bios->data[table]);
+
+		if (match_link) {
+			v = (entry & 0x00c00000) >> 22;
+			if (!(v & dcbent->sorconf.link))
+				continue;
+		}
+
+		v = (entry & 0x000f0000) >> 16;
+		if (!(v & dcbent->or))
+			continue;
+
+		v = (entry & 0x000000f0) >> 4;
+		if (v != dcbent->location)
+			continue;
+
+		v = (entry & 0x0000000f);
+		if (v != dcbent->type)
+			continue;
+
+		return &bios->data[table];
+	}
+
+	return NULL;
+}
+
+void *
+nouveau_bios_dp_table(struct drm_device *dev, struct dcb_entry *dcbent,
+		      int *length)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nvbios *bios = &dev_priv->vbios;
+	uint8_t *table;
+
+	if (!bios->display.dp_table_ptr) {
+		NV_ERROR(dev, "No pointer to DisplayPort table\n");
+		return NULL;
+	}
+	table = &bios->data[bios->display.dp_table_ptr];
+
+	if (table[0] != 0x20 && table[0] != 0x21) {
+		NV_ERROR(dev, "DisplayPort table version 0x%02x unknown\n",
+			 table[0]);
+		return NULL;
+	}
+
+	*length = table[4];
+	return bios_output_config_match(dev, dcbent,
+					bios->display.dp_table_ptr + table[1],
+					table[2], table[3], table[0] >= 0x21);
 }
 
 int
-nouveau_bios_run_display_table(struct drm_device *dev, u16 type, int pclk,
-			       struct dcb_entry *dcbent, int crtc)
+nouveau_bios_run_display_table(struct drm_device *dev, struct dcb_entry *dcbent,
+			       uint32_t sub, int pxclk)
 {
 	/*
 	 * The display script table is located by the BIT 'U' table.
@@ -4341,7 +4497,7 @@ nouveau_bios_run_display_table(struct drm_device *dev, u16 type, int pclk,
 	uint8_t *table = &bios->data[bios->display.script_table_ptr];
 	uint8_t *otable = NULL;
 	uint16_t script;
-	int i;
+	int i = 0;
 
 	if (!bios->display.script_table_ptr) {
 		NV_ERROR(dev, "No pointer to output script table\n");
@@ -4393,33 +4549,30 @@ nouveau_bios_run_display_table(struct drm_device *dev, u16 type, int pclk,
 
 	NV_DEBUG_KMS(dev, "Searching for output entry for %d %d %d\n",
 			dcbent->type, dcbent->location, dcbent->or);
-	for (i = 0; i < table[3]; i++) {
-		otable = ROMPTR(dev, table[table[1] + (i * table[2])]);
-		if (otable && bios_encoder_match(dcbent, ROM32(otable[0])))
-			break;
-	}
-
+	otable = bios_output_config_match(dev, dcbent, table[1] +
+					  bios->display.script_table_ptr,
+					  table[2], table[3], table[0] >= 0x21);
 	if (!otable) {
 		NV_DEBUG_KMS(dev, "failed to match any output table\n");
 		return 1;
 	}
 
-	if (pclk < -2 || pclk > 0) {
+	if (pxclk < -2 || pxclk > 0) {
 		/* Try to find matching script table entry */
 		for (i = 0; i < otable[5]; i++) {
-			if (ROM16(otable[table[4] + i*6]) == type)
+			if (ROM16(otable[table[4] + i*6]) == sub)
 				break;
 		}
 
 		if (i == otable[5]) {
 			NV_ERROR(dev, "Table 0x%04x not found for %d/%d, "
 				      "using first\n",
-				 type, dcbent->type, dcbent->or);
+				 sub, dcbent->type, dcbent->or);
 			i = 0;
 		}
 	}
 
-	if (pclk == 0) {
+	if (pxclk == 0) {
 		script = ROM16(otable[6]);
 		if (!script) {
 			NV_DEBUG_KMS(dev, "output script 0 not found\n");
@@ -4427,9 +4580,9 @@ nouveau_bios_run_display_table(struct drm_device *dev, u16 type, int pclk,
 		}
 
 		NV_DEBUG_KMS(dev, "0x%04X: parsing output script 0\n", script);
-		nouveau_bios_run_init_table(dev, script, dcbent, crtc);
+		nouveau_bios_run_init_table(dev, script, dcbent);
 	} else
-	if (pclk == -1) {
+	if (pxclk == -1) {
 		script = ROM16(otable[8]);
 		if (!script) {
 			NV_DEBUG_KMS(dev, "output script 1 not found\n");
@@ -4437,9 +4590,9 @@ nouveau_bios_run_display_table(struct drm_device *dev, u16 type, int pclk,
 		}
 
 		NV_DEBUG_KMS(dev, "0x%04X: parsing output script 1\n", script);
-		nouveau_bios_run_init_table(dev, script, dcbent, crtc);
+		nouveau_bios_run_init_table(dev, script, dcbent);
 	} else
-	if (pclk == -2) {
+	if (pxclk == -2) {
 		if (table[4] >= 12)
 			script = ROM16(otable[10]);
 		else
@@ -4450,31 +4603,31 @@ nouveau_bios_run_display_table(struct drm_device *dev, u16 type, int pclk,
 		}
 
 		NV_DEBUG_KMS(dev, "0x%04X: parsing output script 2\n", script);
-		nouveau_bios_run_init_table(dev, script, dcbent, crtc);
+		nouveau_bios_run_init_table(dev, script, dcbent);
 	} else
-	if (pclk > 0) {
+	if (pxclk > 0) {
 		script = ROM16(otable[table[4] + i*6 + 2]);
 		if (script)
-			script = clkcmptable(bios, script, pclk);
+			script = clkcmptable(bios, script, pxclk);
 		if (!script) {
 			NV_DEBUG_KMS(dev, "clock script 0 not found\n");
 			return 1;
 		}
 
 		NV_DEBUG_KMS(dev, "0x%04X: parsing clock script 0\n", script);
-		nouveau_bios_run_init_table(dev, script, dcbent, crtc);
+		nouveau_bios_run_init_table(dev, script, dcbent);
 	} else
-	if (pclk < 0) {
+	if (pxclk < 0) {
 		script = ROM16(otable[table[4] + i*6 + 4]);
 		if (script)
-			script = clkcmptable(bios, script, -pclk);
+			script = clkcmptable(bios, script, -pxclk);
 		if (!script) {
 			NV_DEBUG_KMS(dev, "clock script 1 not found\n");
 			return 1;
 		}
 
 		NV_DEBUG_KMS(dev, "0x%04X: parsing clock script 1\n", script);
-		nouveau_bios_run_init_table(dev, script, dcbent, crtc);
+		nouveau_bios_run_init_table(dev, script, dcbent);
 	}
 
 	return 0;
@@ -4572,7 +4725,7 @@ static struct pll_mapping nv84_pll_mapping[] = {
 	{ PLL_CORE  , 0x004028 },
 	{ PLL_SHADER, 0x004020 },
 	{ PLL_MEMORY, 0x004008 },
-	{ PLL_VDEC  , 0x004030 },
+	{ PLL_UNK05 , 0x004030 },
 	{ PLL_UNK41 , 0x00e818 },
 	{ PLL_VPLL0 , 0x614100 },
 	{ PLL_VPLL1 , 0x614900 },
@@ -5033,7 +5186,7 @@ static int parse_bit_A_tbl_entry(struct drm_device *dev, struct nvbios *bios, st
 	load_table_ptr = ROM16(bios->data[bitentry->offset]);
 
 	if (load_table_ptr == 0x0) {
-		NV_DEBUG(dev, "Pointer to BIT loadval table invalid\n");
+		NV_ERROR(dev, "Pointer to BIT loadval table invalid\n");
 		return -EINVAL;
 	}
 
@@ -5324,6 +5477,14 @@ parse_bit_U_tbl_entry(struct drm_device *dev, struct nvbios *bios,
 	return 0;
 }
 
+static int
+parse_bit_displayport_tbl_entry(struct drm_device *dev, struct nvbios *bios,
+				struct bit_entry *bitentry)
+{
+	bios->display.dp_table_ptr = ROM16(bios->data[bitentry->offset]);
+	return 0;
+}
+
 struct bit_table {
 	const char id;
 	int (* const parse_fn)(struct drm_device *, struct nvbios *, struct bit_entry *);
@@ -5338,9 +5499,6 @@ bit_table(struct drm_device *dev, u8 id, struct bit_entry *bit)
 	struct nvbios *bios = &dev_priv->vbios;
 	u8 entries, *entry;
 
-	if (bios->type != NVBIOS_BIT)
-		return -ENODEV;
-
 	entries = bios->data[bios->offset + 10];
 	entry   = &bios->data[bios->offset + 12];
 	while (entries--) {
@@ -5349,7 +5507,7 @@ bit_table(struct drm_device *dev, u8 id, struct bit_entry *bit)
 			bit->version = entry[1];
 			bit->length = ROM16(entry[2]);
 			bit->offset = ROM16(entry[4]);
-			bit->data = ROMPTR(dev, entry[4]);
+			bit->data = ROMPTR(bios, entry[4]);
 			return 0;
 		}
 
@@ -5400,6 +5558,7 @@ parse_bit_structure(struct nvbios *bios, const uint16_t bitoffset)
 	parse_bit_table(bios, bitoffset, &BIT_TABLE('L', lvds));
 	parse_bit_table(bios, bitoffset, &BIT_TABLE('T', tmds));
 	parse_bit_table(bios, bitoffset, &BIT_TABLE('U', U));
+	parse_bit_table(bios, bitoffset, &BIT_TABLE('d', displayport));
 
 	return 0;
 }
@@ -5454,6 +5613,10 @@ static int parse_bmp_structure(struct drm_device *dev, struct nvbios *bios, unsi
 	uint16_t legacy_scripts_offset, legacy_i2c_offset;
 
 	/* load needed defaults in case we can't parse this info */
+	bios->dcb.i2c[0].write = NV_CIO_CRE_DDC_WR__INDEX;
+	bios->dcb.i2c[0].read = NV_CIO_CRE_DDC_STATUS__INDEX;
+	bios->dcb.i2c[1].write = NV_CIO_CRE_DDC0_WR__INDEX;
+	bios->dcb.i2c[1].read = NV_CIO_CRE_DDC0_STATUS__INDEX;
 	bios->digital_min_front_porch = 0x4b;
 	bios->fmaxvco = 256000;
 	bios->fminvco = 128000;
@@ -5561,6 +5724,14 @@ static int parse_bmp_structure(struct drm_device *dev, struct nvbios *bios, unsi
 	bios->legacy.i2c_indices.crt = bios->data[legacy_i2c_offset];
 	bios->legacy.i2c_indices.tv = bios->data[legacy_i2c_offset + 1];
 	bios->legacy.i2c_indices.panel = bios->data[legacy_i2c_offset + 2];
+	if (bios->data[legacy_i2c_offset + 4])
+		bios->dcb.i2c[0].write = bios->data[legacy_i2c_offset + 4];
+	if (bios->data[legacy_i2c_offset + 5])
+		bios->dcb.i2c[0].read = bios->data[legacy_i2c_offset + 5];
+	if (bios->data[legacy_i2c_offset + 6])
+		bios->dcb.i2c[1].write = bios->data[legacy_i2c_offset + 6];
+	if (bios->data[legacy_i2c_offset + 7])
+		bios->dcb.i2c[1].read = bios->data[legacy_i2c_offset + 7];
 
 	if (bmplength > 74) {
 		bios->fmaxvco = ROM32(bmp[67]);
@@ -5611,128 +5782,274 @@ static uint16_t findstr(uint8_t *data, int n, const uint8_t *str, int len)
 	return 0;
 }
 
-void *
-dcb_table(struct drm_device *dev)
+static struct dcb_gpio_entry *
+new_gpio_entry(struct nvbios *bios)
 {
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	u8 *dcb = NULL;
+	struct drm_device *dev = bios->dev;
+	struct dcb_gpio_table *gpio = &bios->dcb.gpio;
 
-	if (dev_priv->card_type > NV_04)
-		dcb = ROMPTR(dev, dev_priv->vbios.data[0x36]);
-	if (!dcb) {
-		NV_WARNONCE(dev, "No DCB data found in VBIOS\n");
+	if (gpio->entries >= DCB_MAX_NUM_GPIO_ENTRIES) {
+		NV_ERROR(dev, "exceeded maximum number of gpio entries!!\n");
 		return NULL;
 	}
 
-	if (dcb[0] >= 0x41) {
-		NV_WARNONCE(dev, "DCB version 0x%02x unknown\n", dcb[0]);
-		return NULL;
-	} else
-	if (dcb[0] >= 0x30) {
-		if (ROM32(dcb[6]) == 0x4edcbdcb)
-			return dcb;
-	} else
-	if (dcb[0] >= 0x20) {
-		if (ROM32(dcb[4]) == 0x4edcbdcb)
-			return dcb;
-	} else
-	if (dcb[0] >= 0x15) {
-		if (!memcmp(&dcb[-7], "DEV_REC", 7))
-			return dcb;
-	} else {
-		/*
-		 * v1.4 (some NV15/16, NV11+) seems the same as v1.5, but
-		 * always has the same single (crt) entry, even when tv-out
-		 * present, so the conclusion is this version cannot really
-		 * be used.
-		 *
-		 * v1.2 tables (some NV6/10, and NV15+) normally have the
-		 * same 5 entries, which are not specific to the card and so
-		 * no use.
-		 *
-		 * v1.2 does have an I2C table that read_dcb_i2c_table can
-		 * handle, but cards exist (nv11 in #14821) with a bad i2c
-		 * table pointer, so use the indices parsed in
-		 * parse_bmp_structure.
-		 *
-		 * v1.1 (NV5+, maybe some NV4) is entirely unhelpful
-		 */
-		NV_WARNONCE(dev, "No useful DCB data in VBIOS\n");
-		return NULL;
+	return &gpio->entry[gpio->entries++];
+}
+
+struct dcb_gpio_entry *
+nouveau_bios_gpio_entry(struct drm_device *dev, enum dcb_gpio_tag tag)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nvbios *bios = &dev_priv->vbios;
+	int i;
+
+	for (i = 0; i < bios->dcb.gpio.entries; i++) {
+		if (bios->dcb.gpio.entry[i].tag != tag)
+			continue;
+
+		return &bios->dcb.gpio.entry[i];
 	}
 
-	NV_WARNONCE(dev, "DCB header validation failed\n");
 	return NULL;
 }
 
-void *
-dcb_outp(struct drm_device *dev, u8 idx)
+static void
+parse_dcb_gpio_table(struct nvbios *bios)
 {
-	u8 *dcb = dcb_table(dev);
-	if (dcb && dcb[0] >= 0x30) {
-		if (idx < dcb[2])
-			return dcb + dcb[1] + (idx * dcb[3]);
+	struct drm_device *dev = bios->dev;
+	struct dcb_gpio_entry *e;
+	u8 headerlen, entries, recordlen;
+	u8 *dcb, *gpio = NULL, *entry;
+	int i;
+
+	dcb = ROMPTR(bios, bios->data[0x36]);
+	if (dcb[0] >= 0x30) {
+		gpio = ROMPTR(bios, dcb[10]);
+		if (!gpio)
+			goto no_table;
+
+		headerlen = gpio[1];
+		entries   = gpio[2];
+		recordlen = gpio[3];
 	} else
-	if (dcb && dcb[0] >= 0x20) {
-		u8 *i2c = ROMPTR(dev, dcb[2]);
-		u8 *ent = dcb + 8 + (idx * 8);
-		if (i2c && ent < i2c)
-			return ent;
+	if (dcb[0] >= 0x22 && dcb[-1] >= 0x13) {
+		gpio = ROMPTR(bios, dcb[-15]);
+		if (!gpio)
+			goto no_table;
+
+		headerlen = 3;
+		entries   = gpio[2];
+		recordlen = gpio[1];
 	} else
-	if (dcb && dcb[0] >= 0x15) {
-		u8 *i2c = ROMPTR(dev, dcb[2]);
-		u8 *ent = dcb + 4 + (idx * 10);
-		if (i2c && ent < i2c)
-			return ent;
+	if (dcb[0] >= 0x22) {
+		/* No GPIO table present, parse the TVDAC GPIO data. */
+		uint8_t *tvdac_gpio = &dcb[-5];
+
+		if (tvdac_gpio[0] & 1) {
+			e = new_gpio_entry(bios);
+			e->tag = DCB_GPIO_TVDAC0;
+			e->line = tvdac_gpio[1] >> 4;
+			e->invert = tvdac_gpio[0] & 2;
+		}
+
+		goto no_table;
+	} else {
+		NV_DEBUG(dev, "no/unknown gpio table on DCB 0x%02x\n", dcb[0]);
+		goto no_table;
 	}
 
-	return NULL;
+	entry = gpio + headerlen;
+	for (i = 0; i < entries; i++, entry += recordlen) {
+		e = new_gpio_entry(bios);
+		if (!e)
+			break;
+
+		if (gpio[0] < 0x40) {
+			e->entry = ROM16(entry[0]);
+			e->tag = (e->entry & 0x07e0) >> 5;
+			if (e->tag == 0x3f) {
+				bios->dcb.gpio.entries--;
+				continue;
+			}
+
+			e->line = (e->entry & 0x001f);
+			e->invert = ((e->entry & 0xf800) >> 11) != 4;
+		} else {
+			e->entry = ROM32(entry[0]);
+			e->tag = (e->entry & 0x0000ff00) >> 8;
+			if (e->tag == 0xff) {
+				bios->dcb.gpio.entries--;
+				continue;
+			}
+
+			e->line = (e->entry & 0x0000001f) >> 0;
+			e->state_default = (e->entry & 0x01000000) >> 24;
+			e->state[0] = (e->entry & 0x18000000) >> 27;
+			e->state[1] = (e->entry & 0x60000000) >> 29;
+		}
+	}
+
+no_table:
+	/* Apple iMac G4 NV18 */
+	if (nv_match_device(dev, 0x0189, 0x10de, 0x0010)) {
+		e = new_gpio_entry(bios);
+		if (e) {
+			e->tag = DCB_GPIO_TVDAC0;
+			e->line = 4;
+		}
+	}
 }
 
-int
-dcb_outp_foreach(struct drm_device *dev, void *data,
-		 int (*exec)(struct drm_device *, void *, int idx, u8 *outp))
-{
-	int ret, idx = -1;
-	u8 *outp = NULL;
-	while ((outp = dcb_outp(dev, ++idx))) {
-		if (ROM32(outp[0]) == 0x00000000)
-			break; /* seen on an NV11 with DCB v1.5 */
-		if (ROM32(outp[0]) == 0xffffffff)
-			break; /* seen on an NV17 with DCB v2.0 */
-
-		if ((outp[0] & 0x0f) == OUTPUT_UNUSED)
-			continue;
-		if ((outp[0] & 0x0f) == OUTPUT_EOL)
-			break;
+struct dcb_connector_table_entry *
+nouveau_bios_connector_entry(struct drm_device *dev, int index)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nvbios *bios = &dev_priv->vbios;
+	struct dcb_connector_table_entry *cte;
+
+	if (index >= bios->dcb.connector.entries)
+		return NULL;
 
-		ret = exec(dev, data, idx, outp);
-		if (ret)
-			return ret;
+	cte = &bios->dcb.connector.entry[index];
+	if (cte->type == 0xff)
+		return NULL;
+
+	return cte;
+}
+
+static enum dcb_connector_type
+divine_connector_type(struct nvbios *bios, int index)
+{
+	struct dcb_table *dcb = &bios->dcb;
+	unsigned encoders = 0, type = DCB_CONNECTOR_NONE;
+	int i;
+
+	for (i = 0; i < dcb->entries; i++) {
+		if (dcb->entry[i].connector == index)
+			encoders |= (1 << dcb->entry[i].type);
 	}
 
-	return 0;
+	if (encoders & (1 << OUTPUT_DP)) {
+		if (encoders & (1 << OUTPUT_TMDS))
+			type = DCB_CONNECTOR_DP;
+		else
+			type = DCB_CONNECTOR_eDP;
+	} else
+	if (encoders & (1 << OUTPUT_TMDS)) {
+		if (encoders & (1 << OUTPUT_ANALOG))
+			type = DCB_CONNECTOR_DVI_I;
+		else
+			type = DCB_CONNECTOR_DVI_D;
+	} else
+	if (encoders & (1 << OUTPUT_ANALOG)) {
+		type = DCB_CONNECTOR_VGA;
+	} else
+	if (encoders & (1 << OUTPUT_LVDS)) {
+		type = DCB_CONNECTOR_LVDS;
+	} else
+	if (encoders & (1 << OUTPUT_TV)) {
+		type = DCB_CONNECTOR_TV_0;
+	}
+
+	return type;
 }
 
-u8 *
-dcb_conntab(struct drm_device *dev)
+static void
+apply_dcb_connector_quirks(struct nvbios *bios, int idx)
 {
-	u8 *dcb = dcb_table(dev);
-	if (dcb && dcb[0] >= 0x30 && dcb[1] >= 0x16) {
-		u8 *conntab = ROMPTR(dev, dcb[0x14]);
-		if (conntab && conntab[0] >= 0x30 && conntab[0] <= 0x40)
-			return conntab;
+	struct dcb_connector_table_entry *cte = &bios->dcb.connector.entry[idx];
+	struct drm_device *dev = bios->dev;
+
+	/* Gigabyte NX85T */
+	if (nv_match_device(dev, 0x0421, 0x1458, 0x344c)) {
+		if (cte->type == DCB_CONNECTOR_HDMI_1)
+			cte->type = DCB_CONNECTOR_DVI_I;
 	}
-	return NULL;
 }
 
-u8 *
-dcb_conn(struct drm_device *dev, u8 idx)
+static const u8 hpd_gpio[16] = {
+	0xff, 0x07, 0x08, 0xff, 0xff, 0x51, 0x52, 0xff,
+	0xff, 0xff, 0xff, 0xff, 0xff, 0x5e, 0x5f, 0x60,
+};
+
+static void
+parse_dcb_connector_table(struct nvbios *bios)
 {
-	u8 *conntab = dcb_conntab(dev);
-	if (conntab && idx < conntab[2])
-		return conntab + conntab[1] + (idx * conntab[3]);
-	return NULL;
+	struct drm_device *dev = bios->dev;
+	struct dcb_connector_table *ct = &bios->dcb.connector;
+	struct dcb_connector_table_entry *cte;
+	uint8_t *conntab = &bios->data[bios->dcb.connector_table_ptr];
+	uint8_t *entry;
+	int i;
+
+	if (!bios->dcb.connector_table_ptr) {
+		NV_DEBUG_KMS(dev, "No DCB connector table present\n");
+		return;
+	}
+
+	NV_INFO(dev, "DCB connector table: VHER 0x%02x %d %d %d\n",
+		conntab[0], conntab[1], conntab[2], conntab[3]);
+	if ((conntab[0] != 0x30 && conntab[0] != 0x40) ||
+	    (conntab[3] != 2 && conntab[3] != 4)) {
+		NV_ERROR(dev, "  Unknown!  Please report.\n");
+		return;
+	}
+
+	ct->entries = conntab[2];
+
+	entry = conntab + conntab[1];
+	cte = &ct->entry[0];
+	for (i = 0; i < conntab[2]; i++, entry += conntab[3], cte++) {
+		cte->index = i;
+		if (conntab[3] == 2)
+			cte->entry = ROM16(entry[0]);
+		else
+			cte->entry = ROM32(entry[0]);
+
+		cte->type  = (cte->entry & 0x000000ff) >> 0;
+		cte->index2 = (cte->entry & 0x00000f00) >> 8;
+
+		cte->gpio_tag = ffs((cte->entry & 0x07033000) >> 12);
+		cte->gpio_tag = hpd_gpio[cte->gpio_tag];
+
+		if (cte->type == 0xff)
+			continue;
+
+		apply_dcb_connector_quirks(bios, i);
+
+		NV_INFO(dev, "  %d: 0x%08x: type 0x%02x idx %d tag 0x%02x\n",
+			i, cte->entry, cte->type, cte->index, cte->gpio_tag);
+
+		/* check for known types, fallback to guessing the type
+		 * from attached encoders if we hit an unknown.
+		 */
+		switch (cte->type) {
+		case DCB_CONNECTOR_VGA:
+		case DCB_CONNECTOR_TV_0:
+		case DCB_CONNECTOR_TV_1:
+		case DCB_CONNECTOR_TV_3:
+		case DCB_CONNECTOR_DVI_I:
+		case DCB_CONNECTOR_DVI_D:
+		case DCB_CONNECTOR_LVDS:
+		case DCB_CONNECTOR_LVDS_SPWG:
+		case DCB_CONNECTOR_DP:
+		case DCB_CONNECTOR_eDP:
+		case DCB_CONNECTOR_HDMI_0:
+		case DCB_CONNECTOR_HDMI_1:
+			break;
+		default:
+			cte->type = divine_connector_type(bios, cte->index);
+			NV_WARN(dev, "unknown type, using 0x%02x\n", cte->type);
+			break;
+		}
+
+		if (nouveau_override_conntype) {
+			int type = divine_connector_type(bios, cte->index);
+			if (type != cte->type)
+				NV_WARN(dev, " -> type 0x%02x\n", cte->type);
+		}
+
+	}
 }
 
 static struct dcb_entry *new_dcb_entry(struct dcb_table *dcb)
@@ -5765,7 +6082,8 @@ parse_dcb20_entry(struct drm_device *dev, struct dcb_table *dcb,
 	entry->type = conn & 0xf;
 	entry->i2c_index = (conn >> 4) & 0xf;
 	entry->heads = (conn >> 8) & 0xf;
-	entry->connector = (conn >> 12) & 0xf;
+	if (dcb->version >= 0x40)
+		entry->connector = (conn >> 12) & 0xf;
 	entry->bus = (conn >> 16) & 0xf;
 	entry->location = (conn >> 20) & 0x3;
 	entry->or = (conn >> 24) & 0xf;
@@ -5831,14 +6149,7 @@ parse_dcb20_entry(struct drm_device *dev, struct dcb_table *dcb,
 	}
 	case OUTPUT_DP:
 		entry->dpconf.sor.link = (conf & 0x00000030) >> 4;
-		switch ((conf & 0x00e00000) >> 21) {
-		case 0:
-			entry->dpconf.link_bw = 162000;
-			break;
-		default:
-			entry->dpconf.link_bw = 270000;
-			break;
-		}
+		entry->dpconf.link_bw = (conf & 0x00e00000) >> 21;
 		switch ((conf & 0x0f000000) >> 24) {
 		case 0xf:
 			entry->dpconf.link_nr = 4;
@@ -5937,6 +6248,25 @@ parse_dcb15_entry(struct drm_device *dev, struct dcb_table *dcb,
 	return true;
 }
 
+static bool parse_dcb_entry(struct drm_device *dev, struct dcb_table *dcb,
+			    uint32_t conn, uint32_t conf)
+{
+	struct dcb_entry *entry = new_dcb_entry(dcb);
+	bool ret;
+
+	if (dcb->version >= 0x20)
+		ret = parse_dcb20_entry(dev, dcb, conn, conf, entry);
+	else
+		ret = parse_dcb15_entry(dev, dcb, conn, conf, entry);
+	if (!ret)
+		return ret;
+
+	read_dcb_i2c_entry(dev, dcb->version, dcb->i2c_table,
+			   entry->i2c_index, &dcb->i2c[entry->i2c_index]);
+
+	return true;
+}
+
 static
 void merge_like_dcb_entries(struct drm_device *dev, struct dcb_table *dcb)
 {
@@ -6047,37 +6377,6 @@ apply_dcb_encoder_quirks(struct drm_device *dev, int idx, u32 *conn, u32 *conf)
 		}
 	}
 
-	/* Some other twisted XFX board (rhbz#694914)
-	 *
-	 * The DVI/VGA encoder combo that's supposed to represent the
-	 * DVI-I connector actually point at two different ones, and
-	 * the HDMI connector ends up paired with the VGA instead.
-	 *
-	 * Connector table is missing anything for VGA at all, pointing it
-	 * an invalid conntab entry 2 so we figure it out ourself.
-	 */
-	if (nv_match_device(dev, 0x0615, 0x1682, 0x2605)) {
-		if (idx == 0) {
-			*conn = 0x02002300; /* VGA, connector 2 */
-			*conf = 0x00000028;
-		} else
-		if (idx == 1) {
-			*conn = 0x01010312; /* DVI, connector 0 */
-			*conf = 0x00020030;
-		} else
-		if (idx == 2) {
-			*conn = 0x04020310; /* VGA, connector 0 */
-			*conf = 0x00000028;
-		} else
-		if (idx == 3) {
-			*conn = 0x02021322; /* HDMI, connector 1 */
-			*conf = 0x00020010;
-		} else {
-			*conn = 0x0000000e; /* EOL */
-			*conf = 0x00000000;
-		}
-	}
-
 	return true;
 }
 
@@ -6097,118 +6396,154 @@ fabricate_dcb_encoder_table(struct drm_device *dev, struct nvbios *bios)
 #endif
 
 	/* Make up some sane defaults */
-	fabricate_dcb_output(dcb, OUTPUT_ANALOG,
-			     bios->legacy.i2c_indices.crt, 1, 1);
+	fabricate_dcb_output(dcb, OUTPUT_ANALOG, LEGACY_I2C_CRT, 1, 1);
 
 	if (nv04_tv_identify(dev, bios->legacy.i2c_indices.tv) >= 0)
-		fabricate_dcb_output(dcb, OUTPUT_TV,
-				     bios->legacy.i2c_indices.tv,
+		fabricate_dcb_output(dcb, OUTPUT_TV, LEGACY_I2C_TV,
 				     all_heads, 0);
 
 	else if (bios->tmds.output0_script_ptr ||
 		 bios->tmds.output1_script_ptr)
-		fabricate_dcb_output(dcb, OUTPUT_TMDS,
-				     bios->legacy.i2c_indices.panel,
+		fabricate_dcb_output(dcb, OUTPUT_TMDS, LEGACY_I2C_PANEL,
 				     all_heads, 1);
 }
 
 static int
-parse_dcb_entry(struct drm_device *dev, void *data, int idx, u8 *outp)
+parse_dcb_table(struct drm_device *dev, struct nvbios *bios)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct dcb_table *dcb = &dev_priv->vbios.dcb;
-	u32 conf = (dcb->version >= 0x20) ? ROM32(outp[4]) : ROM32(outp[6]);
-	u32 conn = ROM32(outp[0]);
-	bool ret;
-
-	if (apply_dcb_encoder_quirks(dev, idx, &conn, &conf)) {
-		struct dcb_entry *entry = new_dcb_entry(dcb);
-
-		NV_TRACEWARN(dev, "DCB outp %02d: %08x %08x\n", idx, conn, conf);
+	struct dcb_table *dcb = &bios->dcb;
+	uint16_t dcbptr = 0, i2ctabptr = 0;
+	uint8_t *dcbtable;
+	uint8_t headerlen = 0x4, entries = DCB_MAX_NUM_ENTRIES;
+	bool configblock = true;
+	int recordlength = 8, confofs = 4;
+	int i;
 
-		if (dcb->version >= 0x20)
-			ret = parse_dcb20_entry(dev, dcb, conn, conf, entry);
-		else
-			ret = parse_dcb15_entry(dev, dcb, conn, conf, entry);
-		if (!ret)
-			return 1; /* stop parsing */
+	/* get the offset from 0x36 */
+	if (dev_priv->card_type > NV_04) {
+		dcbptr = ROM16(bios->data[0x36]);
+		if (dcbptr == 0x0000)
+			NV_WARN(dev, "No output data (DCB) found in BIOS\n");
+	}
 
-		/* Ignore the I2C index for on-chip TV-out, as there
-		 * are cards with bogus values (nv31m in bug 23212),
-		 * and it's otherwise useless.
-		 */
-		if (entry->type == OUTPUT_TV &&
-		    entry->location == DCB_LOC_ON_CHIP)
-			entry->i2c_index = 0x0f;
+	/* this situation likely means a really old card, pre DCB */
+	if (dcbptr == 0x0) {
+		fabricate_dcb_encoder_table(dev, bios);
+		return 0;
 	}
 
-	return 0;
-}
+	dcbtable = &bios->data[dcbptr];
 
-static void
-dcb_fake_connectors(struct nvbios *bios)
-{
-	struct dcb_table *dcbt = &bios->dcb;
-	u8 map[16] = { };
-	int i, idx = 0;
+	/* get DCB version */
+	dcb->version = dcbtable[0];
+	NV_TRACE(dev, "Found Display Configuration Block version %d.%d\n",
+		 dcb->version >> 4, dcb->version & 0xf);
 
-	/* heuristic: if we ever get a non-zero connector field, assume
-	 * that all the indices are valid and we don't need fake them.
-	 */
-	for (i = 0; i < dcbt->entries; i++) {
-		if (dcbt->entry[i].connector)
-			return;
-	}
+	if (dcb->version >= 0x20) { /* NV17+ */
+		uint32_t sig;
 
-	/* no useful connector info available, we need to make it up
-	 * ourselves.  the rule here is: anything on the same i2c bus
-	 * is considered to be on the same connector.  any output
-	 * without an associated i2c bus is assigned its own unique
-	 * connector index.
-	 */
-	for (i = 0; i < dcbt->entries; i++) {
-		u8 i2c = dcbt->entry[i].i2c_index;
-		if (i2c == 0x0f) {
-			dcbt->entry[i].connector = idx++;
+		if (dcb->version >= 0x30) { /* NV40+ */
+			headerlen = dcbtable[1];
+			entries = dcbtable[2];
+			recordlength = dcbtable[3];
+			i2ctabptr = ROM16(dcbtable[4]);
+			sig = ROM32(dcbtable[6]);
+			dcb->gpio_table_ptr = ROM16(dcbtable[10]);
+			dcb->connector_table_ptr = ROM16(dcbtable[20]);
 		} else {
-			if (!map[i2c])
-				map[i2c] = ++idx;
-			dcbt->entry[i].connector = map[i2c] - 1;
+			i2ctabptr = ROM16(dcbtable[2]);
+			sig = ROM32(dcbtable[4]);
+			headerlen = 8;
 		}
-	}
 
-	/* if we created more than one connector, destroy the connector
-	 * table - just in case it has random, rather than stub, entries.
-	 */
-	if (i > 1) {
-		u8 *conntab = dcb_conntab(bios->dev);
-		if (conntab)
-			conntab[0] = 0x00;
-	}
-}
+		if (sig != 0x4edcbdcb) {
+			NV_ERROR(dev, "Bad Display Configuration Block "
+					"signature (%08X)\n", sig);
+			return -EINVAL;
+		}
+	} else if (dcb->version >= 0x15) { /* some NV11 and NV20 */
+		char sig[8] = { 0 };
 
-static int
-parse_dcb_table(struct drm_device *dev, struct nvbios *bios)
-{
-	struct dcb_table *dcb = &bios->dcb;
-	u8 *dcbt, *conn;
-	int idx;
-
-	dcbt = dcb_table(dev);
-	if (!dcbt) {
-		/* handle pre-DCB boards */
-		if (bios->type == NVBIOS_BMP) {
-			fabricate_dcb_encoder_table(dev, bios);
-			return 0;
+		strncpy(sig, (char *)&dcbtable[-7], 7);
+		i2ctabptr = ROM16(dcbtable[2]);
+		recordlength = 10;
+		confofs = 6;
+
+		if (strcmp(sig, "DEV_REC")) {
+			NV_ERROR(dev, "Bad Display Configuration Block "
+					"signature (%s)\n", sig);
+			return -EINVAL;
 		}
+	} else {
+		/*
+		 * v1.4 (some NV15/16, NV11+) seems the same as v1.5, but always
+		 * has the same single (crt) entry, even when tv-out present, so
+		 * the conclusion is this version cannot really be used.
+		 * v1.2 tables (some NV6/10, and NV15+) normally have the same
+		 * 5 entries, which are not specific to the card and so no use.
+		 * v1.2 does have an I2C table that read_dcb_i2c_table can
+		 * handle, but cards exist (nv11 in #14821) with a bad i2c table
+		 * pointer, so use the indices parsed in parse_bmp_structure.
+		 * v1.1 (NV5+, maybe some NV4) is entirely unhelpful
+		 */
+		NV_TRACEWARN(dev, "No useful information in BIOS output table; "
+				  "adding all possible outputs\n");
+		fabricate_dcb_encoder_table(dev, bios);
+		return 0;
+	}
 
-		return -EINVAL;
+	if (!i2ctabptr)
+		NV_WARN(dev, "No pointer to DCB I2C port table\n");
+	else {
+		dcb->i2c_table = &bios->data[i2ctabptr];
+		if (dcb->version >= 0x30)
+			dcb->i2c_default_indices = dcb->i2c_table[4];
+
+		/*
+		 * Parse the "management" I2C bus, used for hardware
+		 * monitoring and some external TMDS transmitters.
+		 */
+		if (dcb->version >= 0x22) {
+			int idx = (dcb->version >= 0x40 ?
+				   dcb->i2c_default_indices & 0xf :
+				   2);
+
+			read_dcb_i2c_entry(dev, dcb->version, dcb->i2c_table,
+					   idx, &dcb->i2c[idx]);
+		}
 	}
 
-	NV_TRACE(dev, "DCB version %d.%d\n", dcbt[0] >> 4, dcbt[0] & 0xf);
+	if (entries > DCB_MAX_NUM_ENTRIES)
+		entries = DCB_MAX_NUM_ENTRIES;
+
+	for (i = 0; i < entries; i++) {
+		uint32_t connection, config = 0;
 
-	dcb->version = dcbt[0];
-	dcb_outp_foreach(dev, NULL, parse_dcb_entry);
+		connection = ROM32(dcbtable[headerlen + recordlength * i]);
+		if (configblock)
+			config = ROM32(dcbtable[headerlen + confofs + recordlength * i]);
+
+		/* seen on an NV11 with DCB v1.5 */
+		if (connection == 0x00000000)
+			break;
+
+		/* seen on an NV17 with DCB v2.0 */
+		if (connection == 0xffffffff)
+			break;
+
+		if ((connection & 0x0000000f) == 0x0000000f)
+			continue;
+
+		if (!apply_dcb_encoder_quirks(dev, i, &connection, &config))
+			continue;
+
+		NV_TRACEWARN(dev, "Raw DCB entry %d: %08x %08x\n",
+			     dcb->entries, connection, config);
+
+		if (!parse_dcb_entry(dev, dcb, connection, config))
+			break;
+	}
 
 	/*
 	 * apart for v2.1+ not being known for requiring merging, this
@@ -6220,19 +6555,77 @@ parse_dcb_table(struct drm_device *dev, struct nvbios *bios)
 	if (!dcb->entries)
 		return -ENXIO;
 
-	/* dump connector table entries to log, if any exist */
-	idx = -1;
-	while ((conn = dcb_conn(dev, ++idx))) {
-		if (conn[0] != 0xff) {
-			NV_TRACE(dev, "DCB conn %02d: ", idx);
-			if (dcb_conntab(dev)[3] < 4)
-				printk("%04x\n", ROM16(conn[0]));
-			else
-				printk("%08x\n", ROM32(conn[0]));
+	parse_dcb_gpio_table(bios);
+	parse_dcb_connector_table(bios);
+	return 0;
+}
+
+static void
+fixup_legacy_connector(struct nvbios *bios)
+{
+	struct dcb_table *dcb = &bios->dcb;
+	int i, i2c, i2c_conn[DCB_MAX_NUM_I2C_ENTRIES] = { };
+
+	/*
+	 * DCB 3.0 also has the table in most cases, but there are some cards
+	 * where the table is filled with stub entries, and the DCB entriy
+	 * indices are all 0.  We don't need the connector indices on pre-G80
+	 * chips (yet?) so limit the use to DCB 4.0 and above.
+	 */
+	if (dcb->version >= 0x40)
+		return;
+
+	dcb->connector.entries = 0;
+
+	/*
+	 * No known connector info before v3.0, so make it up.  the rule here
+	 * is: anything on the same i2c bus is considered to be on the same
+	 * connector.  any output without an associated i2c bus is assigned
+	 * its own unique connector index.
+	 */
+	for (i = 0; i < dcb->entries; i++) {
+		/*
+		 * Ignore the I2C index for on-chip TV-out, as there
+		 * are cards with bogus values (nv31m in bug 23212),
+		 * and it's otherwise useless.
+		 */
+		if (dcb->entry[i].type == OUTPUT_TV &&
+		    dcb->entry[i].location == DCB_LOC_ON_CHIP)
+			dcb->entry[i].i2c_index = 0xf;
+		i2c = dcb->entry[i].i2c_index;
+
+		if (i2c_conn[i2c]) {
+			dcb->entry[i].connector = i2c_conn[i2c] - 1;
+			continue;
 		}
+
+		dcb->entry[i].connector = dcb->connector.entries++;
+		if (i2c != 0xf)
+			i2c_conn[i2c] = dcb->connector.entries;
+	}
+
+	/* Fake the connector table as well as just connector indices */
+	for (i = 0; i < dcb->connector.entries; i++) {
+		dcb->connector.entry[i].index = i;
+		dcb->connector.entry[i].type = divine_connector_type(bios, i);
+		dcb->connector.entry[i].gpio_tag = 0xff;
+	}
+}
+
+static void
+fixup_legacy_i2c(struct nvbios *bios)
+{
+	struct dcb_table *dcb = &bios->dcb;
+	int i;
+
+	for (i = 0; i < dcb->entries; i++) {
+		if (dcb->entry[i].i2c_index == LEGACY_I2C_CRT)
+			dcb->entry[i].i2c_index = bios->legacy.i2c_indices.crt;
+		if (dcb->entry[i].i2c_index == LEGACY_I2C_PANEL)
+			dcb->entry[i].i2c_index = bios->legacy.i2c_indices.panel;
+		if (dcb->entry[i].i2c_index == LEGACY_I2C_TV)
+			dcb->entry[i].i2c_index = bios->legacy.i2c_indices.tv;
 	}
-	dcb_fake_connectors(bios);
-	return 0;
 }
 
 static int load_nv17_hwsq_ucode_entry(struct drm_device *dev, struct nvbios *bios, uint16_t hwsq_offset, int entry)
@@ -6338,7 +6731,7 @@ uint8_t *nouveau_bios_embedded_edid(struct drm_device *dev)
 
 void
 nouveau_bios_run_init_table(struct drm_device *dev, uint16_t table,
-			    struct dcb_entry *dcbent, int crtc)
+			    struct dcb_entry *dcbent)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nvbios *bios = &dev_priv->vbios;
@@ -6346,22 +6739,11 @@ nouveau_bios_run_init_table(struct drm_device *dev, uint16_t table,
 
 	spin_lock_bh(&bios->lock);
 	bios->display.output = dcbent;
-	bios->display.crtc = crtc;
 	parse_init_table(bios, table, &iexec);
 	bios->display.output = NULL;
 	spin_unlock_bh(&bios->lock);
 }
 
-void
-nouveau_bios_init_exec(struct drm_device *dev, uint16_t table)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nvbios *bios = &dev_priv->vbios;
-	struct init_exec iexec = { true, false };
-
-	parse_init_table(bios, table, &iexec);
-}
-
 static bool NVInitVBIOS(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
@@ -6371,7 +6753,11 @@ static bool NVInitVBIOS(struct drm_device *dev)
 	spin_lock_init(&bios->lock);
 	bios->dev = dev;
 
-	return bios_shadow(dev);
+	if (!NVShadowVBIOS(dev, bios->data))
+		return false;
+
+	bios->length = NV_PROM_SIZE;
+	return true;
 }
 
 static int nouveau_parse_vbios_struct(struct drm_device *dev)
@@ -6439,14 +6825,28 @@ nouveau_run_vbios_init(struct drm_device *dev)
 
 	if (dev_priv->card_type >= NV_50) {
 		for (i = 0; i < bios->dcb.entries; i++) {
-			nouveau_bios_run_display_table(dev, 0, 0,
-						       &bios->dcb.entry[i], -1);
+			nouveau_bios_run_display_table(dev,
+						       &bios->dcb.entry[i],
+						       0, 0);
 		}
 	}
 
 	return ret;
 }
 
+static void
+nouveau_bios_i2c_devices_takedown(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nvbios *bios = &dev_priv->vbios;
+	struct dcb_i2c_entry *entry;
+	int i;
+
+	entry = &bios->dcb.i2c[0];
+	for (i = 0; i < DCB_MAX_NUM_I2C_ENTRIES; i++, entry++)
+		nouveau_i2c_fini(dev, entry);
+}
+
 static bool
 nouveau_bios_posted(struct drm_device *dev)
 {
@@ -6483,18 +6883,13 @@ nouveau_bios_init(struct drm_device *dev)
 	if (ret)
 		return ret;
 
-	ret = nouveau_i2c_init(dev);
-	if (ret)
-		return ret;
-
-	ret = nouveau_mxm_init(dev);
-	if (ret)
-		return ret;
-
 	ret = parse_dcb_table(dev, bios);
 	if (ret)
 		return ret;
 
+	fixup_legacy_i2c(bios);
+	fixup_legacy_connector(bios);
+
 	if (!bios->major_version)	/* we don't run version 0 bios */
 		return 0;
 
@@ -6531,10 +6926,5 @@ nouveau_bios_init(struct drm_device *dev)
 void
 nouveau_bios_takedown(struct drm_device *dev)
 {
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-
-	nouveau_mxm_fini(dev);
-	nouveau_i2c_fini(dev);
-
-	kfree(dev_priv->vbios.data);
+	nouveau_bios_i2c_devices_takedown(dev);
 }
diff --git a/drivers/gpu/drm/nouveau/nouveau_bios.h b/drivers/gpu/drm/nouveau/nouveau_bios.h
index 1f3233d..050c314 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bios.h
+++ b/drivers/gpu/drm/nouveau/nouveau_bios.h
@@ -34,14 +34,9 @@
 
 #define DCB_LOC_ON_CHIP 0
 
-#define ROM16(x) le16_to_cpu(*(u16 *)&(x))
-#define ROM32(x) le32_to_cpu(*(u32 *)&(x))
-#define ROM48(x) ({ u8 *p = &(x); (u64)ROM16(p[4]) << 32 | ROM32(p[0]); })
-#define ROM64(x) le64_to_cpu(*(u64 *)&(x))
-#define ROMPTR(d,x) ({            \
-	struct drm_nouveau_private *dev_priv = (d)->dev_private; \
-	ROM16(x) ? &dev_priv->vbios.data[ROM16(x)] : NULL; \
-})
+#define ROM16(x) le16_to_cpu(*(uint16_t *)&(x))
+#define ROM32(x) le32_to_cpu(*(uint32_t *)&(x))
+#define ROMPTR(bios, x) (ROM16(x) ? &(bios)->data[ROM16(x)] : NULL)
 
 struct bit_entry {
 	uint8_t  id;
@@ -53,13 +48,30 @@ struct bit_entry {
 
 int bit_table(struct drm_device *, u8 id, struct bit_entry *);
 
+struct dcb_i2c_entry {
+	uint32_t entry;
+	uint8_t port_type;
+	uint8_t read, write;
+	struct nouveau_i2c_chan *chan;
+};
+
 enum dcb_gpio_tag {
-	DCB_GPIO_PANEL_POWER = 0x01,
-	DCB_GPIO_TVDAC0 = 0x0c,
+	DCB_GPIO_TVDAC0 = 0xc,
 	DCB_GPIO_TVDAC1 = 0x2d,
-	DCB_GPIO_PWM_FAN = 0x09,
-	DCB_GPIO_FAN_SENSE = 0x3d,
-	DCB_GPIO_UNUSED = 0xff
+};
+
+struct dcb_gpio_entry {
+	enum dcb_gpio_tag tag;
+	int line;
+	bool invert;
+	uint32_t entry;
+	uint8_t state_default;
+	uint8_t state[2];
+};
+
+struct dcb_gpio_table {
+	int entries;
+	struct dcb_gpio_entry entry[DCB_MAX_NUM_GPIO_ENTRIES];
 };
 
 enum dcb_connector_type {
@@ -75,11 +87,23 @@ enum dcb_connector_type {
 	DCB_CONNECTOR_eDP = 0x47,
 	DCB_CONNECTOR_HDMI_0 = 0x60,
 	DCB_CONNECTOR_HDMI_1 = 0x61,
-	DCB_CONNECTOR_DMS59_DP0 = 0x64,
-	DCB_CONNECTOR_DMS59_DP1 = 0x65,
 	DCB_CONNECTOR_NONE = 0xff
 };
 
+struct dcb_connector_table_entry {
+	uint8_t index;
+	uint32_t entry;
+	enum dcb_connector_type type;
+	uint8_t index2;
+	uint8_t gpio_tag;
+	void *drm;
+};
+
+struct dcb_connector_table {
+	int entries;
+	struct dcb_connector_table_entry entry[DCB_MAX_NUM_CONNECTOR_ENTRIES];
+};
+
 enum dcb_type {
 	OUTPUT_ANALOG = 0,
 	OUTPUT_TV = 1,
@@ -87,7 +111,6 @@ enum dcb_type {
 	OUTPUT_LVDS = 3,
 	OUTPUT_DP = 6,
 	OUTPUT_EOL = 14, /* DCB 4.0+, appears to be end-of-list */
-	OUTPUT_UNUSED = 15,
 	OUTPUT_ANY = -1
 };
 
@@ -132,8 +155,18 @@ struct dcb_entry {
 
 struct dcb_table {
 	uint8_t version;
+
 	int entries;
 	struct dcb_entry entry[DCB_MAX_NUM_ENTRIES];
+
+	uint8_t *i2c_table;
+	uint8_t i2c_default_indices;
+	struct dcb_i2c_entry i2c[DCB_MAX_NUM_I2C_ENTRIES];
+
+	uint16_t gpio_table_ptr;
+	struct dcb_gpio_table gpio;
+	uint16_t connector_table_ptr;
+	struct dcb_connector_table connector;
 };
 
 enum nouveau_or {
@@ -162,7 +195,7 @@ enum pll_types {
 	PLL_SHADER = 0x02,
 	PLL_UNK03  = 0x03,
 	PLL_MEMORY = 0x04,
-	PLL_VDEC   = 0x05,
+	PLL_UNK05  = 0x05,
 	PLL_UNK40  = 0x40,
 	PLL_UNK41  = 0x41,
 	PLL_UNK42  = 0x42,
@@ -211,8 +244,6 @@ struct nvbios {
 		NVBIOS_BIT
 	} type;
 	uint16_t offset;
-	uint32_t length;
-	uint8_t *data;
 
 	uint8_t chip_version;
 
@@ -223,6 +254,8 @@ struct nvbios {
 
 	spinlock_t lock;
 
+	uint8_t data[NV_PROM_SIZE];
+	unsigned int length;
 	bool execute;
 
 	uint8_t major_version;
@@ -256,8 +289,8 @@ struct nvbios {
 
 	struct {
 		struct dcb_entry *output;
-		int crtc;
 		uint16_t script_table_ptr;
+		uint16_t dp_table_ptr;
 	} display;
 
 	struct {
@@ -300,11 +333,4 @@ struct nvbios {
 	} legacy;
 };
 
-void *dcb_table(struct drm_device *);
-void *dcb_outp(struct drm_device *, u8 idx);
-int dcb_outp_foreach(struct drm_device *, void *data,
-		     int (*)(struct drm_device *, void *, int idx, u8 *outp));
-u8 *dcb_conntab(struct drm_device *);
-u8 *dcb_conn(struct drm_device *, u8 idx);
-
 #endif
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.c b/drivers/gpu/drm/nouveau/nouveau_bo.c
index b1d919f..5fb98de 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.c
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.c
@@ -28,7 +28,6 @@
  */
 
 #include "drmP.h"
-#include "ttm/ttm_page_alloc.h"
 
 #include "nouveau_drm.h"
 #include "nouveau_drv.h"
@@ -50,12 +49,16 @@ nouveau_bo_del_ttm(struct ttm_buffer_object *bo)
 		DRM_ERROR("bo %p still attached to GEM object\n", bo);
 
 	nv10_mem_put_tile_region(dev, nvbo->tile, NULL);
+	if (nvbo->vma.node) {
+		nouveau_vm_unmap(&nvbo->vma);
+		nouveau_vm_put(&nvbo->vma);
+	}
 	kfree(nvbo);
 }
 
 static void
 nouveau_bo_fixup_align(struct nouveau_bo *nvbo, u32 flags,
-		       int *align, int *size)
+		       int *align, int *size, int *page_shift)
 {
 	struct drm_nouveau_private *dev_priv = nouveau_bdev(nvbo->bo.bdev);
 
@@ -79,55 +82,67 @@ nouveau_bo_fixup_align(struct nouveau_bo *nvbo, u32 flags,
 			}
 		}
 	} else {
-		*size = roundup(*size, (1 << nvbo->page_shift));
-		*align = max((1 <<  nvbo->page_shift), *align);
+		if (likely(dev_priv->chan_vm)) {
+			if (!(flags & TTM_PL_FLAG_TT) &&  *size > 256 * 1024)
+				*page_shift = dev_priv->chan_vm->lpg_shift;
+			else
+				*page_shift = dev_priv->chan_vm->spg_shift;
+		} else {
+			*page_shift = 12;
+		}
+
+		*size = roundup(*size, (1 << *page_shift));
+		*align = max((1 << *page_shift), *align);
 	}
 
 	*size = roundup(*size, PAGE_SIZE);
 }
 
 int
-nouveau_bo_new(struct drm_device *dev, int size, int align,
-	       uint32_t flags, uint32_t tile_mode, uint32_t tile_flags,
-	       struct nouveau_bo **pnvbo)
+nouveau_bo_new(struct drm_device *dev, struct nouveau_channel *chan,
+	       int size, int align, uint32_t flags, uint32_t tile_mode,
+	       uint32_t tile_flags, struct nouveau_bo **pnvbo)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_bo *nvbo;
-	size_t acc_size;
-	int ret;
+	int ret = 0, page_shift = 0;
 
 	nvbo = kzalloc(sizeof(struct nouveau_bo), GFP_KERNEL);
 	if (!nvbo)
 		return -ENOMEM;
 	INIT_LIST_HEAD(&nvbo->head);
 	INIT_LIST_HEAD(&nvbo->entry);
-	INIT_LIST_HEAD(&nvbo->vma_list);
 	nvbo->tile_mode = tile_mode;
 	nvbo->tile_flags = tile_flags;
 	nvbo->bo.bdev = &dev_priv->ttm.bdev;
 
-	nvbo->page_shift = 12;
-	if (dev_priv->bar1_vm) {
-		if (!(flags & TTM_PL_FLAG_TT) && size > 256 * 1024)
-			nvbo->page_shift = dev_priv->bar1_vm->lpg_shift;
+	nouveau_bo_fixup_align(nvbo, flags, &align, &size, &page_shift);
+	align >>= PAGE_SHIFT;
+
+	if (dev_priv->chan_vm) {
+		ret = nouveau_vm_get(dev_priv->chan_vm, size, page_shift,
+				     NV_MEM_ACCESS_RW, &nvbo->vma);
+		if (ret) {
+			kfree(nvbo);
+			return ret;
+		}
 	}
 
-	nouveau_bo_fixup_align(nvbo, flags, &align, &size);
 	nvbo->bo.mem.num_pages = size >> PAGE_SHIFT;
 	nouveau_bo_placement_set(nvbo, flags, 0);
 
-	acc_size = ttm_bo_dma_acc_size(&dev_priv->ttm.bdev, size,
-				       sizeof(struct nouveau_bo));
-
+	nvbo->channel = chan;
 	ret = ttm_bo_init(&dev_priv->ttm.bdev, &nvbo->bo, size,
-			  ttm_bo_type_device, &nvbo->placement,
-			  align >> PAGE_SHIFT, 0, false, NULL, acc_size,
-			  nouveau_bo_del_ttm);
+			  ttm_bo_type_device, &nvbo->placement, align, 0,
+			  false, NULL, size, nouveau_bo_del_ttm);
 	if (ret) {
 		/* ttm will call nouveau_bo_del_ttm if it fails.. */
 		return ret;
 	}
+	nvbo->channel = NULL;
 
+	if (nvbo->vma.node)
+		nvbo->bo.offset = nvbo->vma.offset;
 	*pnvbo = nvbo;
 	return 0;
 }
@@ -153,7 +168,7 @@ set_placement_range(struct nouveau_bo *nvbo, uint32_t type)
 
 	if (dev_priv->card_type == NV_10 &&
 	    nvbo->tile_mode && (type & TTM_PL_FLAG_VRAM) &&
-	    nvbo->bo.mem.num_pages < vram_pages / 4) {
+	    nvbo->bo.mem.num_pages < vram_pages / 2) {
 		/*
 		 * Make sure that the color and depth buffers are handled
 		 * by independent memory controller units. Up to a 9x
@@ -297,6 +312,8 @@ nouveau_bo_validate(struct nouveau_bo *nvbo, bool interruptible,
 	if (ret)
 		return ret;
 
+	if (nvbo->vma.node)
+		nvbo->bo.offset = nvbo->vma.offset;
 	return 0;
 }
 
@@ -348,10 +365,8 @@ nouveau_bo_wr32(struct nouveau_bo *nvbo, unsigned index, u32 val)
 		*mem = val;
 }
 
-static struct ttm_tt *
-nouveau_ttm_tt_create(struct ttm_bo_device *bdev,
-		      unsigned long size, uint32_t page_flags,
-		      struct page *dummy_read_page)
+static struct ttm_backend *
+nouveau_bo_create_ttm_backend_entry(struct ttm_bo_device *bdev)
 {
 	struct drm_nouveau_private *dev_priv = nouveau_bdev(bdev);
 	struct drm_device *dev = dev_priv->dev;
@@ -359,13 +374,11 @@ nouveau_ttm_tt_create(struct ttm_bo_device *bdev,
 	switch (dev_priv->gart_info.type) {
 #if __OS_HAS_AGP
 	case NOUVEAU_GART_AGP:
-		return ttm_agp_tt_create(bdev, dev->agp->bridge,
-					 size, page_flags, dummy_read_page);
+		return ttm_agp_backend_init(bdev, dev->agp->bridge);
 #endif
 	case NOUVEAU_GART_PDMA:
 	case NOUVEAU_GART_HW:
-		return nouveau_sgdma_create_ttm(bdev, size, page_flags,
-						dummy_read_page);
+		return nouveau_sgdma_init_ttm(dev);
 	default:
 		NV_ERROR(dev, "Unknown GART type %d\n",
 			 dev_priv->gart_info.type);
@@ -427,6 +440,7 @@ nouveau_bo_init_mem_type(struct ttm_bo_device *bdev, uint32_t type,
 				     TTM_MEMTYPE_FLAG_CMA;
 			man->available_caching = TTM_PL_MASK_CACHING;
 			man->default_caching = TTM_PL_FLAG_CACHED;
+			man->gpu_offset = dev_priv->gart_info.aper_base;
 			break;
 		default:
 			NV_ERROR(dev, "Unknown GART type: %d\n",
@@ -487,12 +501,19 @@ static int
 nvc0_bo_move_m2mf(struct nouveau_channel *chan, struct ttm_buffer_object *bo,
 		  struct ttm_mem_reg *old_mem, struct ttm_mem_reg *new_mem)
 {
-	struct nouveau_mem *node = old_mem->mm_node;
-	u64 src_offset = node->vma[0].offset;
-	u64 dst_offset = node->vma[1].offset;
+	struct nouveau_mem *old_node = old_mem->mm_node;
+	struct nouveau_mem *new_node = new_mem->mm_node;
+	struct nouveau_bo *nvbo = nouveau_bo(bo);
 	u32 page_count = new_mem->num_pages;
+	u64 src_offset, dst_offset;
 	int ret;
 
+	src_offset = old_node->tmp_vma.offset;
+	if (new_node->tmp_vma.node)
+		dst_offset = new_node->tmp_vma.offset;
+	else
+		dst_offset = nvbo->vma.offset;
+
 	page_count = new_mem->num_pages;
 	while (page_count) {
 		int line_count = (page_count > 2047) ? 2047 : page_count;
@@ -526,13 +547,19 @@ static int
 nv50_bo_move_m2mf(struct nouveau_channel *chan, struct ttm_buffer_object *bo,
 		  struct ttm_mem_reg *old_mem, struct ttm_mem_reg *new_mem)
 {
-	struct nouveau_mem *node = old_mem->mm_node;
+	struct nouveau_mem *old_node = old_mem->mm_node;
+	struct nouveau_mem *new_node = new_mem->mm_node;
 	struct nouveau_bo *nvbo = nouveau_bo(bo);
 	u64 length = (new_mem->num_pages << PAGE_SHIFT);
-	u64 src_offset = node->vma[0].offset;
-	u64 dst_offset = node->vma[1].offset;
+	u64 src_offset, dst_offset;
 	int ret;
 
+	src_offset = old_node->tmp_vma.offset;
+	if (new_node->tmp_vma.node)
+		dst_offset = new_node->tmp_vma.offset;
+	else
+		dst_offset = nvbo->vma.offset;
+
 	while (length) {
 		u32 amount, stride, height;
 
@@ -668,26 +695,6 @@ nv04_bo_move_m2mf(struct nouveau_channel *chan, struct ttm_buffer_object *bo,
 }
 
 static int
-nouveau_vma_getmap(struct nouveau_channel *chan, struct nouveau_bo *nvbo,
-		   struct ttm_mem_reg *mem, struct nouveau_vma *vma)
-{
-	struct nouveau_mem *node = mem->mm_node;
-	int ret;
-
-	ret = nouveau_vm_get(chan->vm, mem->num_pages << PAGE_SHIFT,
-			     node->page_shift, NV_MEM_ACCESS_RO, vma);
-	if (ret)
-		return ret;
-
-	if (mem->mem_type == TTM_PL_VRAM)
-		nouveau_vm_map(vma, node);
-	else
-		nouveau_vm_map_sg(vma, 0, mem->num_pages << PAGE_SHIFT, node);
-
-	return 0;
-}
-
-static int
 nouveau_bo_move_m2mf(struct ttm_buffer_object *bo, int evict, bool intr,
 		     bool no_wait_reserve, bool no_wait_gpu,
 		     struct ttm_mem_reg *new_mem)
@@ -704,20 +711,31 @@ nouveau_bo_move_m2mf(struct ttm_buffer_object *bo, int evict, bool intr,
 		mutex_lock_nested(&chan->mutex, NOUVEAU_KCHANNEL_MUTEX);
 	}
 
-	/* create temporary vmas for the transfer and attach them to the
-	 * old nouveau_mem node, these will get cleaned up after ttm has
-	 * destroyed the ttm_mem_reg
+	/* create temporary vma for old memory, this will get cleaned
+	 * up after ttm destroys the ttm_mem_reg
 	 */
 	if (dev_priv->card_type >= NV_50) {
 		struct nouveau_mem *node = old_mem->mm_node;
+		if (!node->tmp_vma.node) {
+			u32 page_shift = nvbo->vma.node->type;
+			if (old_mem->mem_type == TTM_PL_TT)
+				page_shift = nvbo->vma.vm->spg_shift;
+
+			ret = nouveau_vm_get(chan->vm,
+					     old_mem->num_pages << PAGE_SHIFT,
+					     page_shift, NV_MEM_ACCESS_RO,
+					     &node->tmp_vma);
+			if (ret)
+				goto out;
+		}
 
-		ret = nouveau_vma_getmap(chan, nvbo, old_mem, &node->vma[0]);
-		if (ret)
-			goto out;
-
-		ret = nouveau_vma_getmap(chan, nvbo, new_mem, &node->vma[1]);
-		if (ret)
-			goto out;
+		if (old_mem->mem_type == TTM_PL_VRAM)
+			nouveau_vm_map(&node->tmp_vma, node);
+		else {
+			nouveau_vm_map_sg(&node->tmp_vma, 0,
+					  old_mem->num_pages << PAGE_SHIFT,
+					  node, node->pages);
+		}
 	}
 
 	if (dev_priv->card_type < NV_50)
@@ -744,6 +762,7 @@ nouveau_bo_move_flipd(struct ttm_buffer_object *bo, bool evict, bool intr,
 		      bool no_wait_reserve, bool no_wait_gpu,
 		      struct ttm_mem_reg *new_mem)
 {
+	struct drm_nouveau_private *dev_priv = nouveau_bdev(bo->bdev);
 	u32 placement_memtype = TTM_PL_FLAG_TT | TTM_PL_MASK_CACHING;
 	struct ttm_placement placement;
 	struct ttm_mem_reg tmp_mem;
@@ -763,7 +782,23 @@ nouveau_bo_move_flipd(struct ttm_buffer_object *bo, bool evict, bool intr,
 	if (ret)
 		goto out;
 
+	if (dev_priv->card_type >= NV_50) {
+		struct nouveau_bo *nvbo = nouveau_bo(bo);
+		struct nouveau_mem *node = tmp_mem.mm_node;
+		struct nouveau_vma *vma = &nvbo->vma;
+		if (vma->node->type != vma->vm->spg_shift)
+			vma = &node->tmp_vma;
+		nouveau_vm_map_sg(vma, 0, tmp_mem.num_pages << PAGE_SHIFT,
+				  node, node->pages);
+	}
+
 	ret = nouveau_bo_move_m2mf(bo, true, intr, no_wait_reserve, no_wait_gpu, &tmp_mem);
+
+	if (dev_priv->card_type >= NV_50) {
+		struct nouveau_bo *nvbo = nouveau_bo(bo);
+		nouveau_vm_unmap(&nvbo->vma);
+	}
+
 	if (ret)
 		goto out;
 
@@ -809,25 +844,30 @@ out:
 static void
 nouveau_bo_move_ntfy(struct ttm_buffer_object *bo, struct ttm_mem_reg *new_mem)
 {
+	struct drm_nouveau_private *dev_priv = nouveau_bdev(bo->bdev);
+	struct nouveau_mem *node = new_mem->mm_node;
 	struct nouveau_bo *nvbo = nouveau_bo(bo);
-	struct nouveau_vma *vma;
+	struct nouveau_vma *vma = &nvbo->vma;
+	struct nouveau_vm *vm = vma->vm;
 
-	/* ttm can now (stupidly) pass the driver bos it didn't create... */
-	if (bo->destroy != nouveau_bo_del_ttm)
+	if (dev_priv->card_type < NV_50)
 		return;
 
-	list_for_each_entry(vma, &nvbo->vma_list, head) {
-		if (new_mem && new_mem->mem_type == TTM_PL_VRAM) {
-			nouveau_vm_map(vma, new_mem->mm_node);
-		} else
-		if (new_mem && new_mem->mem_type == TTM_PL_TT &&
-		    nvbo->page_shift == vma->vm->spg_shift) {
-			nouveau_vm_map_sg(vma, 0, new_mem->
-					  num_pages << PAGE_SHIFT,
-					  new_mem->mm_node);
-		} else {
+	switch (new_mem->mem_type) {
+	case TTM_PL_VRAM:
+		nouveau_vm_map(vma, node);
+		break;
+	case TTM_PL_TT:
+		if (vma->node->type != vm->spg_shift) {
 			nouveau_vm_unmap(vma);
+			vma = &node->tmp_vma;
 		}
+		nouveau_vm_map_sg(vma, 0, new_mem->num_pages << PAGE_SHIFT,
+				  node, node->pages);
+		break;
+	default:
+		nouveau_vm_unmap(&nvbo->vma);
+		break;
 	}
 }
 
@@ -967,7 +1007,7 @@ nouveau_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_mem_reg *mem)
 			break;
 		}
 
-		if (dev_priv->card_type >= NV_C0)
+		if (dev_priv->card_type == NV_C0)
 			page_shift = node->page_shift;
 		else
 			page_shift = 12;
@@ -1055,94 +1095,8 @@ nouveau_bo_fence(struct nouveau_bo *nvbo, struct nouveau_fence *fence)
 	nouveau_fence_unref(&old_fence);
 }
 
-static int
-nouveau_ttm_tt_populate(struct ttm_tt *ttm)
-{
-	struct ttm_dma_tt *ttm_dma = (void *)ttm;
-	struct drm_nouveau_private *dev_priv;
-	struct drm_device *dev;
-	unsigned i;
-	int r;
-
-	if (ttm->state != tt_unpopulated)
-		return 0;
-
-	dev_priv = nouveau_bdev(ttm->bdev);
-	dev = dev_priv->dev;
-
-#if __OS_HAS_AGP
-	if (dev_priv->gart_info.type == NOUVEAU_GART_AGP) {
-		return ttm_agp_tt_populate(ttm);
-	}
-#endif
-
-#ifdef CONFIG_SWIOTLB
-	if (swiotlb_nr_tbl()) {
-		return ttm_dma_populate((void *)ttm, dev->dev);
-	}
-#endif
-
-	r = ttm_pool_populate(ttm);
-	if (r) {
-		return r;
-	}
-
-	for (i = 0; i < ttm->num_pages; i++) {
-		ttm_dma->dma_address[i] = pci_map_page(dev->pdev, ttm->pages[i],
-						   0, PAGE_SIZE,
-						   PCI_DMA_BIDIRECTIONAL);
-		if (pci_dma_mapping_error(dev->pdev, ttm_dma->dma_address[i])) {
-			while (--i) {
-				pci_unmap_page(dev->pdev, ttm_dma->dma_address[i],
-					       PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
-				ttm_dma->dma_address[i] = 0;
-			}
-			ttm_pool_unpopulate(ttm);
-			return -EFAULT;
-		}
-	}
-	return 0;
-}
-
-static void
-nouveau_ttm_tt_unpopulate(struct ttm_tt *ttm)
-{
-	struct ttm_dma_tt *ttm_dma = (void *)ttm;
-	struct drm_nouveau_private *dev_priv;
-	struct drm_device *dev;
-	unsigned i;
-
-	dev_priv = nouveau_bdev(ttm->bdev);
-	dev = dev_priv->dev;
-
-#if __OS_HAS_AGP
-	if (dev_priv->gart_info.type == NOUVEAU_GART_AGP) {
-		ttm_agp_tt_unpopulate(ttm);
-		return;
-	}
-#endif
-
-#ifdef CONFIG_SWIOTLB
-	if (swiotlb_nr_tbl()) {
-		ttm_dma_unpopulate((void *)ttm, dev->dev);
-		return;
-	}
-#endif
-
-	for (i = 0; i < ttm->num_pages; i++) {
-		if (ttm_dma->dma_address[i]) {
-			pci_unmap_page(dev->pdev, ttm_dma->dma_address[i],
-				       PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
-		}
-	}
-
-	ttm_pool_unpopulate(ttm);
-}
-
 struct ttm_bo_driver nouveau_bo_driver = {
-	.ttm_tt_create = &nouveau_ttm_tt_create,
-	.ttm_tt_populate = &nouveau_ttm_tt_populate,
-	.ttm_tt_unpopulate = &nouveau_ttm_tt_unpopulate,
+	.create_ttm_backend_entry = nouveau_bo_create_ttm_backend_entry,
 	.invalidate_caches = nouveau_bo_invalidate_caches,
 	.init_mem_type = nouveau_bo_init_mem_type,
 	.evict_flags = nouveau_bo_evict_flags,
@@ -1159,54 +1113,3 @@ struct ttm_bo_driver nouveau_bo_driver = {
 	.io_mem_free = &nouveau_ttm_io_mem_free,
 };
 
-struct nouveau_vma *
-nouveau_bo_vma_find(struct nouveau_bo *nvbo, struct nouveau_vm *vm)
-{
-	struct nouveau_vma *vma;
-	list_for_each_entry(vma, &nvbo->vma_list, head) {
-		if (vma->vm == vm)
-			return vma;
-	}
-
-	return NULL;
-}
-
-int
-nouveau_bo_vma_add(struct nouveau_bo *nvbo, struct nouveau_vm *vm,
-		   struct nouveau_vma *vma)
-{
-	const u32 size = nvbo->bo.mem.num_pages << PAGE_SHIFT;
-	struct nouveau_mem *node = nvbo->bo.mem.mm_node;
-	int ret;
-
-	ret = nouveau_vm_get(vm, size, nvbo->page_shift,
-			     NV_MEM_ACCESS_RW, vma);
-	if (ret)
-		return ret;
-
-	if (nvbo->bo.mem.mem_type == TTM_PL_VRAM)
-		nouveau_vm_map(vma, nvbo->bo.mem.mm_node);
-	else
-	if (nvbo->bo.mem.mem_type == TTM_PL_TT)
-		nouveau_vm_map_sg(vma, 0, size, node);
-
-	list_add_tail(&vma->head, &nvbo->vma_list);
-	vma->refcount = 1;
-	return 0;
-}
-
-void
-nouveau_bo_vma_del(struct nouveau_bo *nvbo, struct nouveau_vma *vma)
-{
-	if (vma->node) {
-		if (nvbo->bo.mem.mem_type != TTM_PL_SYSTEM) {
-			spin_lock(&nvbo->bo.bdev->fence_lock);
-			ttm_bo_wait(&nvbo->bo, false, false, false);
-			spin_unlock(&nvbo->bo.bdev->fence_lock);
-			nouveau_vm_unmap(vma);
-		}
-
-		nouveau_vm_put(vma);
-		list_del(&vma->head);
-	}
-}
diff --git a/drivers/gpu/drm/nouveau/nouveau_channel.c b/drivers/gpu/drm/nouveau/nouveau_channel.c
index a018def..d31d355 100644
--- a/drivers/gpu/drm/nouveau/nouveau_channel.c
+++ b/drivers/gpu/drm/nouveau/nouveau_channel.c
@@ -27,63 +27,40 @@
 #include "nouveau_drv.h"
 #include "nouveau_drm.h"
 #include "nouveau_dma.h"
-#include "nouveau_ramht.h"
 
 static int
-nouveau_channel_pushbuf_init(struct nouveau_channel *chan)
+nouveau_channel_pushbuf_ctxdma_init(struct nouveau_channel *chan)
 {
-	u32 mem = nouveau_vram_pushbuf ? TTM_PL_FLAG_VRAM : TTM_PL_FLAG_TT;
 	struct drm_device *dev = chan->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	int ret;
-
-	/* allocate buffer object */
-	ret = nouveau_bo_new(dev, 65536, 0, mem, 0, 0, &chan->pushbuf_bo);
-	if (ret)
-		goto out;
-
-	ret = nouveau_bo_pin(chan->pushbuf_bo, mem);
-	if (ret)
-		goto out;
-
-	ret = nouveau_bo_map(chan->pushbuf_bo);
-	if (ret)
-		goto out;
+	struct nouveau_bo *pb = chan->pushbuf_bo;
+	struct nouveau_gpuobj *pushbuf = NULL;
+	int ret = 0;
 
-	/* create DMA object covering the entire memtype where the push
-	 * buffer resides, userspace can submit its own push buffers from
-	 * anywhere within the same memtype.
-	 */
-	chan->pushbuf_base = chan->pushbuf_bo->bo.offset;
 	if (dev_priv->card_type >= NV_50) {
-		ret = nouveau_bo_vma_add(chan->pushbuf_bo, chan->vm,
-					 &chan->pushbuf_vma);
-		if (ret)
-			goto out;
-
 		if (dev_priv->card_type < NV_C0) {
 			ret = nouveau_gpuobj_dma_new(chan,
 						     NV_CLASS_DMA_IN_MEMORY, 0,
 						     (1ULL << 40),
 						     NV_MEM_ACCESS_RO,
 						     NV_MEM_TARGET_VM,
-						     &chan->pushbuf);
+						     &pushbuf);
 		}
-		chan->pushbuf_base = chan->pushbuf_vma.offset;
+		chan->pushbuf_base = pb->bo.offset;
 	} else
-	if (chan->pushbuf_bo->bo.mem.mem_type == TTM_PL_TT) {
+	if (pb->bo.mem.mem_type == TTM_PL_TT) {
 		ret = nouveau_gpuobj_dma_new(chan, NV_CLASS_DMA_IN_MEMORY, 0,
 					     dev_priv->gart_info.aper_size,
 					     NV_MEM_ACCESS_RO,
-					     NV_MEM_TARGET_GART,
-					     &chan->pushbuf);
+					     NV_MEM_TARGET_GART, &pushbuf);
+		chan->pushbuf_base = pb->bo.mem.start << PAGE_SHIFT;
 	} else
 	if (dev_priv->card_type != NV_04) {
 		ret = nouveau_gpuobj_dma_new(chan, NV_CLASS_DMA_IN_MEMORY, 0,
 					     dev_priv->fb_available_size,
 					     NV_MEM_ACCESS_RO,
-					     NV_MEM_TARGET_VRAM,
-					     &chan->pushbuf);
+					     NV_MEM_TARGET_VRAM, &pushbuf);
+		chan->pushbuf_base = pb->bo.mem.start << PAGE_SHIFT;
 	} else {
 		/* NV04 cmdbuf hack, from original ddx.. not sure of it's
 		 * exact reason for existing :)  PCI access to cmdbuf in
@@ -93,22 +70,47 @@ nouveau_channel_pushbuf_init(struct nouveau_channel *chan)
 					     pci_resource_start(dev->pdev, 1),
 					     dev_priv->fb_available_size,
 					     NV_MEM_ACCESS_RO,
-					     NV_MEM_TARGET_PCI,
-					     &chan->pushbuf);
+					     NV_MEM_TARGET_PCI, &pushbuf);
+		chan->pushbuf_base = pb->bo.mem.start << PAGE_SHIFT;
+	}
+
+	nouveau_gpuobj_ref(pushbuf, &chan->pushbuf);
+	nouveau_gpuobj_ref(NULL, &pushbuf);
+	return ret;
+}
+
+static struct nouveau_bo *
+nouveau_channel_user_pushbuf_alloc(struct drm_device *dev)
+{
+	struct nouveau_bo *pushbuf = NULL;
+	int location, ret;
+
+	if (nouveau_vram_pushbuf)
+		location = TTM_PL_FLAG_VRAM;
+	else
+		location = TTM_PL_FLAG_TT;
+
+	ret = nouveau_bo_new(dev, NULL, 65536, 0, location, 0, 0x0000, &pushbuf);
+	if (ret) {
+		NV_ERROR(dev, "error allocating DMA push buffer: %d\n", ret);
+		return NULL;
 	}
 
-out:
+	ret = nouveau_bo_pin(pushbuf, location);
 	if (ret) {
-		NV_ERROR(dev, "error initialising pushbuf: %d\n", ret);
-		nouveau_bo_vma_del(chan->pushbuf_bo, &chan->pushbuf_vma);
-		nouveau_gpuobj_ref(NULL, &chan->pushbuf);
-		if (chan->pushbuf_bo) {
-			nouveau_bo_unmap(chan->pushbuf_bo);
-			nouveau_bo_ref(NULL, &chan->pushbuf_bo);
-		}
+		NV_ERROR(dev, "error pinning DMA push buffer: %d\n", ret);
+		nouveau_bo_ref(NULL, &pushbuf);
+		return NULL;
 	}
 
-	return 0;
+	ret = nouveau_bo_map(pushbuf);
+	if (ret) {
+		nouveau_bo_unpin(pushbuf);
+		nouveau_bo_ref(NULL, &pushbuf);
+		return NULL;
+	}
+
+	return pushbuf;
 }
 
 /* allocates and initializes a fifo for user space consumption */
@@ -119,7 +121,6 @@ nouveau_channel_alloc(struct drm_device *dev, struct nouveau_channel **chan_ret,
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_fifo_engine *pfifo = &dev_priv->engine.fifo;
-	struct nouveau_fpriv *fpriv = nouveau_fpriv(file_priv);
 	struct nouveau_channel *chan;
 	unsigned long flags;
 	int ret;
@@ -160,14 +161,19 @@ nouveau_channel_alloc(struct drm_device *dev, struct nouveau_channel **chan_ret,
 	INIT_LIST_HEAD(&chan->fence.pending);
 	spin_lock_init(&chan->fence.lock);
 
-	/* setup channel's memory and vm */
-	ret = nouveau_gpuobj_channel_init(chan, vram_handle, gart_handle);
-	if (ret) {
-		NV_ERROR(dev, "gpuobj %d\n", ret);
+	/* Allocate DMA push buffer */
+	chan->pushbuf_bo = nouveau_channel_user_pushbuf_alloc(dev);
+	if (!chan->pushbuf_bo) {
+		ret = -ENOMEM;
+		NV_ERROR(dev, "pushbuf %d\n", ret);
 		nouveau_channel_put(&chan);
 		return ret;
 	}
 
+	nouveau_dma_pre_init(chan);
+	chan->user_put = 0x40;
+	chan->user_get = 0x44;
+
 	/* Allocate space for per-channel fixed notifier memory */
 	ret = nouveau_notifier_init_channel(chan);
 	if (ret) {
@@ -176,19 +182,21 @@ nouveau_channel_alloc(struct drm_device *dev, struct nouveau_channel **chan_ret,
 		return ret;
 	}
 
-	/* Allocate DMA push buffer */
-	ret = nouveau_channel_pushbuf_init(chan);
+	/* Setup channel's default objects */
+	ret = nouveau_gpuobj_channel_init(chan, vram_handle, gart_handle);
 	if (ret) {
-		NV_ERROR(dev, "pushbuf %d\n", ret);
+		NV_ERROR(dev, "gpuobj %d\n", ret);
 		nouveau_channel_put(&chan);
 		return ret;
 	}
 
-	nouveau_dma_pre_init(chan);
-	chan->user_put = 0x40;
-	chan->user_get = 0x44;
-	if (dev_priv->card_type >= NV_50)
-                chan->user_get_hi = 0x60;
+	/* Create a dma object for the push buffer */
+	ret = nouveau_channel_pushbuf_ctxdma_init(chan);
+	if (ret) {
+		NV_ERROR(dev, "pbctxdma %d\n", ret);
+		nouveau_channel_put(&chan);
+		return ret;
+	}
 
 	/* disable the fifo caches */
 	pfifo->reassign(dev, false);
@@ -213,11 +221,6 @@ nouveau_channel_alloc(struct drm_device *dev, struct nouveau_channel **chan_ret,
 	nouveau_debugfs_channel_init(chan);
 
 	NV_DEBUG(dev, "channel %d initialised\n", chan->id);
-	if (fpriv) {
-		spin_lock(&fpriv->lock);
-		list_add(&chan->list, &fpriv->channels);
-		spin_unlock(&fpriv->lock);
-	}
 	*chan_ret = chan;
 	return 0;
 }
@@ -234,23 +237,29 @@ nouveau_channel_get_unlocked(struct nouveau_channel *ref)
 }
 
 struct nouveau_channel *
-nouveau_channel_get(struct drm_file *file_priv, int id)
+nouveau_channel_get(struct drm_device *dev, struct drm_file *file_priv, int id)
 {
-	struct nouveau_fpriv *fpriv = nouveau_fpriv(file_priv);
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_channel *chan;
+	unsigned long flags;
 
-	spin_lock(&fpriv->lock);
-	list_for_each_entry(chan, &fpriv->channels, list) {
-		if (chan->id == id) {
-			chan = nouveau_channel_get_unlocked(chan);
-			spin_unlock(&fpriv->lock);
-			mutex_lock(&chan->mutex);
-			return chan;
-		}
+	if (unlikely(id < 0 || id >= NOUVEAU_MAX_CHANNEL_NR))
+		return ERR_PTR(-EINVAL);
+
+	spin_lock_irqsave(&dev_priv->channels.lock, flags);
+	chan = nouveau_channel_get_unlocked(dev_priv->channels.ptr[id]);
+	spin_unlock_irqrestore(&dev_priv->channels.lock, flags);
+
+	if (unlikely(!chan))
+		return ERR_PTR(-EINVAL);
+
+	if (unlikely(file_priv && chan->file_priv != file_priv)) {
+		nouveau_channel_put_unlocked(&chan);
+		return ERR_PTR(-EINVAL);
 	}
-	spin_unlock(&fpriv->lock);
 
-	return ERR_PTR(-EINVAL);
+	mutex_lock(&chan->mutex);
+	return chan;
 }
 
 void
@@ -304,14 +313,12 @@ nouveau_channel_put_unlocked(struct nouveau_channel **pchan)
 	/* destroy any resources the channel owned */
 	nouveau_gpuobj_ref(NULL, &chan->pushbuf);
 	if (chan->pushbuf_bo) {
-		nouveau_bo_vma_del(chan->pushbuf_bo, &chan->pushbuf_vma);
 		nouveau_bo_unmap(chan->pushbuf_bo);
 		nouveau_bo_unpin(chan->pushbuf_bo);
 		nouveau_bo_ref(NULL, &chan->pushbuf_bo);
 	}
-	nouveau_ramht_ref(NULL, &chan->ramht, chan);
-	nouveau_notifier_takedown_channel(chan);
 	nouveau_gpuobj_channel_takedown(chan);
+	nouveau_notifier_takedown_channel(chan);
 
 	nouveau_channel_ref(NULL, pchan);
 }
@@ -377,11 +384,10 @@ nouveau_channel_cleanup(struct drm_device *dev, struct drm_file *file_priv)
 
 	NV_DEBUG(dev, "clearing FIFO enables from file_priv\n");
 	for (i = 0; i < engine->fifo.channels; i++) {
-		chan = nouveau_channel_get(file_priv, i);
+		chan = nouveau_channel_get(dev, file_priv, i);
 		if (IS_ERR(chan))
 			continue;
 
-		list_del(&chan->list);
 		atomic_dec(&chan->users);
 		nouveau_channel_put(&chan);
 	}
@@ -414,17 +420,13 @@ nouveau_ioctl_fifo_alloc(struct drm_device *dev, void *data,
 		return ret;
 	init->channel  = chan->id;
 
-	if (nouveau_vram_pushbuf == 0) {
-		if (chan->dma.ib_max)
-			init->pushbuf_domains = NOUVEAU_GEM_DOMAIN_VRAM |
-						NOUVEAU_GEM_DOMAIN_GART;
-		else if (chan->pushbuf_bo->bo.mem.mem_type == TTM_PL_VRAM)
-			init->pushbuf_domains = NOUVEAU_GEM_DOMAIN_VRAM;
-		else
-			init->pushbuf_domains = NOUVEAU_GEM_DOMAIN_GART;
-	} else {
+	if (chan->dma.ib_max)
+		init->pushbuf_domains = NOUVEAU_GEM_DOMAIN_VRAM |
+					NOUVEAU_GEM_DOMAIN_GART;
+	else if (chan->pushbuf_bo->bo.mem.mem_type == TTM_PL_VRAM)
 		init->pushbuf_domains = NOUVEAU_GEM_DOMAIN_VRAM;
-	}
+	else
+		init->pushbuf_domains = NOUVEAU_GEM_DOMAIN_GART;
 
 	if (dev_priv->card_type < NV_C0) {
 		init->subchan[0].handle = NvM2MF;
@@ -458,11 +460,10 @@ nouveau_ioctl_fifo_free(struct drm_device *dev, void *data,
 	struct drm_nouveau_channel_free *req = data;
 	struct nouveau_channel *chan;
 
-	chan = nouveau_channel_get(file_priv, req->channel);
+	chan = nouveau_channel_get(dev, file_priv, req->channel);
 	if (IS_ERR(chan))
 		return PTR_ERR(chan);
 
-	list_del(&chan->list);
 	atomic_dec(&chan->users);
 	nouveau_channel_put(&chan);
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.c b/drivers/gpu/drm/nouveau/nouveau_connector.c
index 9f9d50d..1595d0b 100644
--- a/drivers/gpu/drm/nouveau/nouveau_connector.c
+++ b/drivers/gpu/drm/nouveau/nouveau_connector.c
@@ -35,13 +35,12 @@
 #include "nouveau_encoder.h"
 #include "nouveau_crtc.h"
 #include "nouveau_connector.h"
-#include "nouveau_gpio.h"
 #include "nouveau_hw.h"
 
 static void nouveau_connector_hotplug(void *, int);
 
-struct nouveau_encoder *
-find_encoder(struct drm_connector *connector, int type)
+static struct nouveau_encoder *
+find_encoder_by_type(struct drm_connector *connector, int type)
 {
 	struct drm_device *dev = connector->dev;
 	struct nouveau_encoder *nv_encoder;
@@ -79,11 +78,29 @@ nouveau_encoder_connector_get(struct nouveau_encoder *encoder)
 	return NULL;
 }
 
+/*TODO: This could use improvement, and learn to handle the fixed
+ *      BIOS tables etc.  It's fine currently, for its only user.
+ */
+int
+nouveau_connector_bpp(struct drm_connector *connector)
+{
+	struct nouveau_connector *nv_connector = nouveau_connector(connector);
+
+	if (nv_connector->edid && nv_connector->edid->revision >= 4) {
+		u8 bpc = ((nv_connector->edid->input & 0x70) >> 3) + 4;
+		if (bpc > 4)
+			return bpc;
+	}
+
+	return 18;
+}
+
 static void
 nouveau_connector_destroy(struct drm_connector *connector)
 {
 	struct nouveau_connector *nv_connector = nouveau_connector(connector);
 	struct drm_nouveau_private *dev_priv;
+	struct nouveau_gpio_engine *pgpio;
 	struct drm_device *dev;
 
 	if (!nv_connector)
@@ -93,11 +110,16 @@ nouveau_connector_destroy(struct drm_connector *connector)
 	dev_priv = dev->dev_private;
 	NV_DEBUG_KMS(dev, "\n");
 
-	if (nv_connector->hpd != DCB_GPIO_UNUSED) {
-		nouveau_gpio_isr_del(dev, 0, nv_connector->hpd, 0xff,
-				     nouveau_connector_hotplug, connector);
+	pgpio = &dev_priv->engine.gpio;
+	if (pgpio->irq_unregister) {
+		pgpio->irq_unregister(dev, nv_connector->dcb->gpio_tag,
+				      nouveau_connector_hotplug, connector);
 	}
 
+	if (connector->connector_type == DRM_MODE_CONNECTOR_LVDS ||
+	    connector->connector_type == DRM_MODE_CONNECTOR_eDP)
+		nouveau_backlight_exit(connector);
+
 	kfree(nv_connector->edid);
 	drm_sysfs_connector_remove(connector);
 	drm_connector_cleanup(connector);
@@ -148,8 +170,8 @@ nouveau_connector_of_detect(struct drm_connector *connector)
 	struct device_node *cn, *dn = pci_device_to_OF_node(dev->pdev);
 
 	if (!dn ||
-	    !((nv_encoder = find_encoder(connector, OUTPUT_TMDS)) ||
-	      (nv_encoder = find_encoder(connector, OUTPUT_ANALOG))))
+	    !((nv_encoder = find_encoder_by_type(connector, OUTPUT_TMDS)) ||
+	      (nv_encoder = find_encoder_by_type(connector, OUTPUT_ANALOG))))
 		return NULL;
 
 	for_each_child_of_node(dn, cn) {
@@ -180,10 +202,6 @@ nouveau_connector_set_encoder(struct drm_connector *connector,
 		return;
 	nv_connector->detected_encoder = nv_encoder;
 
-	if (dev_priv->card_type >= NV_50) {
-		connector->interlace_allowed = true;
-		connector->doublescan_allowed = true;
-	} else
 	if (nv_encoder->dcb->type == OUTPUT_LVDS ||
 	    nv_encoder->dcb->type == OUTPUT_TMDS) {
 		connector->doublescan_allowed = false;
@@ -200,7 +218,7 @@ nouveau_connector_set_encoder(struct drm_connector *connector,
 			connector->interlace_allowed = true;
 	}
 
-	if (nv_connector->type == DCB_CONNECTOR_DVI_I) {
+	if (nv_connector->dcb->type == DCB_CONNECTOR_DVI_I) {
 		drm_connector_property_set_value(connector,
 			dev->mode_config.dvi_i_subconnector_property,
 			nv_encoder->dcb->type == OUTPUT_TMDS ?
@@ -215,7 +233,6 @@ nouveau_connector_detect(struct drm_connector *connector, bool force)
 	struct drm_device *dev = connector->dev;
 	struct nouveau_connector *nv_connector = nouveau_connector(connector);
 	struct nouveau_encoder *nv_encoder = NULL;
-	struct nouveau_encoder *nv_partner;
 	struct nouveau_i2c_chan *i2c;
 	int type;
 
@@ -249,22 +266,19 @@ nouveau_connector_detect(struct drm_connector *connector, bool force)
 		 * same i2c channel so the value returned from ddc_detect
 		 * isn't necessarily correct.
 		 */
-		nv_partner = NULL;
-		if (nv_encoder->dcb->type == OUTPUT_TMDS)
-			nv_partner = find_encoder(connector, OUTPUT_ANALOG);
-		if (nv_encoder->dcb->type == OUTPUT_ANALOG)
-			nv_partner = find_encoder(connector, OUTPUT_TMDS);
-
-		if (nv_partner && ((nv_encoder->dcb->type == OUTPUT_ANALOG &&
-				    nv_partner->dcb->type == OUTPUT_TMDS) ||
-				   (nv_encoder->dcb->type == OUTPUT_TMDS &&
-				    nv_partner->dcb->type == OUTPUT_ANALOG))) {
+		if (nv_connector->dcb->type == DCB_CONNECTOR_DVI_I) {
 			if (nv_connector->edid->input & DRM_EDID_INPUT_DIGITAL)
 				type = OUTPUT_TMDS;
 			else
 				type = OUTPUT_ANALOG;
 
-			nv_encoder = find_encoder(connector, type);
+			nv_encoder = find_encoder_by_type(connector, type);
+			if (!nv_encoder) {
+				NV_ERROR(dev, "Detected %d encoder on %s, "
+					      "but no object!\n", type,
+					 drm_get_connector_name(connector));
+				return connector_status_disconnected;
+			}
 		}
 
 		nouveau_connector_set_encoder(connector, nv_encoder);
@@ -278,9 +292,9 @@ nouveau_connector_detect(struct drm_connector *connector, bool force)
 	}
 
 detect_analog:
-	nv_encoder = find_encoder(connector, OUTPUT_ANALOG);
+	nv_encoder = find_encoder_by_type(connector, OUTPUT_ANALOG);
 	if (!nv_encoder && !nouveau_tv_disable)
-		nv_encoder = find_encoder(connector, OUTPUT_TV);
+		nv_encoder = find_encoder_by_type(connector, OUTPUT_TV);
 	if (nv_encoder && force) {
 		struct drm_encoder *encoder = to_drm_encoder(nv_encoder);
 		struct drm_encoder_helper_funcs *helper =
@@ -313,7 +327,7 @@ nouveau_connector_detect_lvds(struct drm_connector *connector, bool force)
 		nv_connector->edid = NULL;
 	}
 
-	nv_encoder = find_encoder(connector, OUTPUT_LVDS);
+	nv_encoder = find_encoder_by_type(connector, OUTPUT_LVDS);
 	if (!nv_encoder)
 		return connector_status_disconnected;
 
@@ -383,7 +397,7 @@ nouveau_connector_force(struct drm_connector *connector)
 	struct nouveau_encoder *nv_encoder;
 	int type;
 
-	if (nv_connector->type == DCB_CONNECTOR_DVI_I) {
+	if (nv_connector->dcb->type == DCB_CONNECTOR_DVI_I) {
 		if (connector->force == DRM_FORCE_ON_DIGITAL)
 			type = OUTPUT_TMDS;
 		else
@@ -391,7 +405,7 @@ nouveau_connector_force(struct drm_connector *connector)
 	} else
 		type = OUTPUT_ANY;
 
-	nv_encoder = find_encoder(connector, type);
+	nv_encoder = find_encoder_by_type(connector, type);
 	if (!nv_encoder) {
 		NV_ERROR(connector->dev, "can't find encoder to force %s on!\n",
 			 drm_get_connector_name(connector));
@@ -406,21 +420,15 @@ static int
 nouveau_connector_set_property(struct drm_connector *connector,
 			       struct drm_property *property, uint64_t value)
 {
-	struct drm_nouveau_private *dev_priv = connector->dev->dev_private;
-	struct nouveau_display_engine *disp = &dev_priv->engine.display;
 	struct nouveau_connector *nv_connector = nouveau_connector(connector);
 	struct nouveau_encoder *nv_encoder = nv_connector->detected_encoder;
 	struct drm_encoder *encoder = to_drm_encoder(nv_encoder);
 	struct drm_device *dev = connector->dev;
-	struct nouveau_crtc *nv_crtc;
 	int ret;
 
-	nv_crtc = NULL;
-	if (connector->encoder && connector->encoder->crtc)
-		nv_crtc = nouveau_crtc(connector->encoder->crtc);
-
 	/* Scaling mode */
 	if (property == dev->mode_config.scaling_mode_property) {
+		struct nouveau_crtc *nv_crtc = NULL;
 		bool modeset = false;
 
 		switch (value) {
@@ -446,6 +454,8 @@ nouveau_connector_set_property(struct drm_connector *connector,
 			modeset = true;
 		nv_connector->scaling_mode = value;
 
+		if (connector->encoder && connector->encoder->crtc)
+			nv_crtc = nouveau_crtc(connector->encoder->crtc);
 		if (!nv_crtc)
 			return 0;
 
@@ -457,7 +467,7 @@ nouveau_connector_set_property(struct drm_connector *connector,
 			if (!ret)
 				return -EINVAL;
 		} else {
-			ret = nv_crtc->set_scale(nv_crtc, true);
+			ret = nv_crtc->set_scale(nv_crtc, value, true);
 			if (ret)
 				return ret;
 		}
@@ -465,71 +475,23 @@ nouveau_connector_set_property(struct drm_connector *connector,
 		return 0;
 	}
 
-	/* Underscan */
-	if (property == disp->underscan_property) {
-		if (nv_connector->underscan != value) {
-			nv_connector->underscan = value;
-			if (!nv_crtc || !nv_crtc->set_scale)
-				return 0;
-
-			return nv_crtc->set_scale(nv_crtc, true);
-		}
-
-		return 0;
-	}
-
-	if (property == disp->underscan_hborder_property) {
-		if (nv_connector->underscan_hborder != value) {
-			nv_connector->underscan_hborder = value;
-			if (!nv_crtc || !nv_crtc->set_scale)
-				return 0;
-
-			return nv_crtc->set_scale(nv_crtc, true);
-		}
-
-		return 0;
-	}
-
-	if (property == disp->underscan_vborder_property) {
-		if (nv_connector->underscan_vborder != value) {
-			nv_connector->underscan_vborder = value;
-			if (!nv_crtc || !nv_crtc->set_scale)
-				return 0;
-
-			return nv_crtc->set_scale(nv_crtc, true);
-		}
-
-		return 0;
-	}
-
 	/* Dithering */
-	if (property == disp->dithering_mode) {
-		nv_connector->dithering_mode = value;
-		if (!nv_crtc || !nv_crtc->set_dither)
-			return 0;
+	if (property == dev->mode_config.dithering_mode_property) {
+		struct nouveau_crtc *nv_crtc = NULL;
 
-		return nv_crtc->set_dither(nv_crtc, true);
-	}
+		if (value == DRM_MODE_DITHERING_ON)
+			nv_connector->use_dithering = true;
+		else
+			nv_connector->use_dithering = false;
+
+		if (connector->encoder && connector->encoder->crtc)
+			nv_crtc = nouveau_crtc(connector->encoder->crtc);
 
-	if (property == disp->dithering_depth) {
-		nv_connector->dithering_depth = value;
 		if (!nv_crtc || !nv_crtc->set_dither)
 			return 0;
 
-		return nv_crtc->set_dither(nv_crtc, true);
-	}
-
-	if (nv_crtc && nv_crtc->set_color_vibrance) {
-		/* Hue */
-		if (property == disp->vibrant_hue_property) {
-			nv_crtc->vibrant_hue = value - 90;
-			return nv_crtc->set_color_vibrance(nv_crtc, true);
-		}
-		/* Saturation */
-		if (property == disp->color_vibrance_property) {
-			nv_crtc->color_vibrance = value - 100;
-			return nv_crtc->set_color_vibrance(nv_crtc, true);
-		}
+		return nv_crtc->set_dither(nv_crtc, nv_connector->use_dithering,
+					   true);
 	}
 
 	if (nv_encoder && nv_encoder->dcb->type == OUTPUT_TV)
@@ -640,46 +602,6 @@ nouveau_connector_scaler_modes_add(struct drm_connector *connector)
 	return modes;
 }
 
-static void
-nouveau_connector_detect_depth(struct drm_connector *connector)
-{
-	struct drm_nouveau_private *dev_priv = connector->dev->dev_private;
-	struct nouveau_connector *nv_connector = nouveau_connector(connector);
-	struct nouveau_encoder *nv_encoder = nv_connector->detected_encoder;
-	struct nvbios *bios = &dev_priv->vbios;
-	struct drm_display_mode *mode = nv_connector->native_mode;
-	bool duallink;
-
-	/* if the edid is feeling nice enough to provide this info, use it */
-	if (nv_connector->edid && connector->display_info.bpc)
-		return;
-
-	/* if not, we're out of options unless we're LVDS, default to 6bpc */
-	connector->display_info.bpc = 6;
-	if (nv_encoder->dcb->type != OUTPUT_LVDS)
-		return;
-
-	/* LVDS: panel straps */
-	if (bios->fp_no_ddc) {
-		if (bios->fp.if_is_24bit)
-			connector->display_info.bpc = 8;
-		return;
-	}
-
-	/* LVDS: DDC panel, need to first determine the number of links to
-	 * know which if_is_24bit flag to check...
-	 */
-	if (nv_connector->edid &&
-	    nv_connector->type == DCB_CONNECTOR_LVDS_SPWG)
-		duallink = ((u8 *)nv_connector->edid)[121] == 2;
-	else
-		duallink = mode->clock >= bios->fp.duallink_transition_clk;
-
-	if ((!duallink && (bios->fp.strapless_is_24bit & 1)) ||
-	    ( duallink && (bios->fp.strapless_is_24bit & 2)))
-		connector->display_info.bpc = 8;
-}
-
 static int
 nouveau_connector_get_modes(struct drm_connector *connector)
 {
@@ -709,12 +631,6 @@ nouveau_connector_get_modes(struct drm_connector *connector)
 		nv_connector->native_mode = drm_mode_duplicate(dev, &mode);
 	}
 
-	/* Determine display colour depth for everything except LVDS now,
-	 * DP requires this before mode_valid() is called.
-	 */
-	if (connector->connector_type != DRM_MODE_CONNECTOR_LVDS)
-		nouveau_connector_detect_depth(connector);
-
 	/* Find the native mode if this is a digital panel, if we didn't
 	 * find any modes through DDC previously add the native mode to
 	 * the list of modes.
@@ -730,19 +646,12 @@ nouveau_connector_get_modes(struct drm_connector *connector)
 		ret = 1;
 	}
 
-	/* Determine LVDS colour depth, must happen after determining
-	 * "native" mode as some VBIOS tables require us to use the
-	 * pixel clock as part of the lookup...
-	 */
-	if (connector->connector_type == DRM_MODE_CONNECTOR_LVDS)
-		nouveau_connector_detect_depth(connector);
-
 	if (nv_encoder->dcb->type == OUTPUT_TV)
 		ret = get_slave_funcs(encoder)->get_modes(encoder, connector);
 
-	if (nv_connector->type == DCB_CONNECTOR_LVDS ||
-	    nv_connector->type == DCB_CONNECTOR_LVDS_SPWG ||
-	    nv_connector->type == DCB_CONNECTOR_eDP)
+	if (nv_connector->dcb->type == DCB_CONNECTOR_LVDS ||
+	    nv_connector->dcb->type == DCB_CONNECTOR_LVDS_SPWG ||
+	    nv_connector->dcb->type == DCB_CONNECTOR_eDP)
 		ret += nouveau_connector_scaler_modes_add(connector);
 
 	return ret;
@@ -799,9 +708,12 @@ nouveau_connector_mode_valid(struct drm_connector *connector,
 	case OUTPUT_TV:
 		return get_slave_funcs(encoder)->mode_valid(encoder, mode);
 	case OUTPUT_DP:
-		max_clock  = nv_encoder->dp.link_nr;
-		max_clock *= nv_encoder->dp.link_bw;
-		clock = clock * (connector->display_info.bpc * 3) / 10;
+		if (nv_encoder->dp.link_bw == DP_LINK_BW_2_7)
+			max_clock = nv_encoder->dp.link_nr * 270000;
+		else
+			max_clock = nv_encoder->dp.link_nr * 162000;
+
+		clock = clock * nouveau_connector_bpp(connector) / 8;
 		break;
 	default:
 		BUG_ON(1);
@@ -859,195 +771,103 @@ nouveau_connector_funcs_lvds = {
 	.force = nouveau_connector_force
 };
 
-static int
-drm_conntype_from_dcb(enum dcb_connector_type dcb)
-{
-	switch (dcb) {
-	case DCB_CONNECTOR_VGA      : return DRM_MODE_CONNECTOR_VGA;
-	case DCB_CONNECTOR_TV_0     :
-	case DCB_CONNECTOR_TV_1     :
-	case DCB_CONNECTOR_TV_3     : return DRM_MODE_CONNECTOR_TV;
-	case DCB_CONNECTOR_DVI_I    : return DRM_MODE_CONNECTOR_DVII;
-	case DCB_CONNECTOR_DVI_D    : return DRM_MODE_CONNECTOR_DVID;
-	case DCB_CONNECTOR_LVDS     :
-	case DCB_CONNECTOR_LVDS_SPWG: return DRM_MODE_CONNECTOR_LVDS;
-	case DCB_CONNECTOR_DMS59_DP0:
-	case DCB_CONNECTOR_DMS59_DP1:
-	case DCB_CONNECTOR_DP       : return DRM_MODE_CONNECTOR_DisplayPort;
-	case DCB_CONNECTOR_eDP      : return DRM_MODE_CONNECTOR_eDP;
-	case DCB_CONNECTOR_HDMI_0   :
-	case DCB_CONNECTOR_HDMI_1   : return DRM_MODE_CONNECTOR_HDMIA;
-	default:
-		break;
-	}
-
-	return DRM_MODE_CONNECTOR_Unknown;
-}
-
 struct drm_connector *
 nouveau_connector_create(struct drm_device *dev, int index)
 {
 	const struct drm_connector_funcs *funcs = &nouveau_connector_funcs;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_display_engine *disp = &dev_priv->engine.display;
+	struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio;
 	struct nouveau_connector *nv_connector = NULL;
+	struct dcb_connector_table_entry *dcb = NULL;
 	struct drm_connector *connector;
 	int type, ret = 0;
-	bool dummy;
 
 	NV_DEBUG_KMS(dev, "\n");
 
-	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
-		nv_connector = nouveau_connector(connector);
-		if (nv_connector->index == index)
-			return connector;
+	if (index >= dev_priv->vbios.dcb.connector.entries)
+		return ERR_PTR(-EINVAL);
+
+	dcb = &dev_priv->vbios.dcb.connector.entry[index];
+	if (dcb->drm)
+		return dcb->drm;
+
+	switch (dcb->type) {
+	case DCB_CONNECTOR_VGA:
+		type = DRM_MODE_CONNECTOR_VGA;
+		break;
+	case DCB_CONNECTOR_TV_0:
+	case DCB_CONNECTOR_TV_1:
+	case DCB_CONNECTOR_TV_3:
+		type = DRM_MODE_CONNECTOR_TV;
+		break;
+	case DCB_CONNECTOR_DVI_I:
+		type = DRM_MODE_CONNECTOR_DVII;
+		break;
+	case DCB_CONNECTOR_DVI_D:
+		type = DRM_MODE_CONNECTOR_DVID;
+		break;
+	case DCB_CONNECTOR_HDMI_0:
+	case DCB_CONNECTOR_HDMI_1:
+		type = DRM_MODE_CONNECTOR_HDMIA;
+		break;
+	case DCB_CONNECTOR_LVDS:
+	case DCB_CONNECTOR_LVDS_SPWG:
+		type = DRM_MODE_CONNECTOR_LVDS;
+		funcs = &nouveau_connector_funcs_lvds;
+		break;
+	case DCB_CONNECTOR_DP:
+		type = DRM_MODE_CONNECTOR_DisplayPort;
+		break;
+	case DCB_CONNECTOR_eDP:
+		type = DRM_MODE_CONNECTOR_eDP;
+		break;
+	default:
+		NV_ERROR(dev, "unknown connector type: 0x%02x!!\n", dcb->type);
+		return ERR_PTR(-EINVAL);
 	}
 
 	nv_connector = kzalloc(sizeof(*nv_connector), GFP_KERNEL);
 	if (!nv_connector)
 		return ERR_PTR(-ENOMEM);
-
+	nv_connector->dcb = dcb;
 	connector = &nv_connector->base;
-	nv_connector->index = index;
-
-	/* attempt to parse vbios connector type and hotplug gpio */
-	nv_connector->dcb = dcb_conn(dev, index);
-	if (nv_connector->dcb) {
-		static const u8 hpd[16] = {
-			0xff, 0x07, 0x08, 0xff, 0xff, 0x51, 0x52, 0xff,
-			0xff, 0xff, 0xff, 0xff, 0xff, 0x5e, 0x5f, 0x60,
-		};
-
-		u32 entry = ROM16(nv_connector->dcb[0]);
-		if (dcb_conntab(dev)[3] >= 4)
-			entry |= (u32)ROM16(nv_connector->dcb[2]) << 16;
-
-		nv_connector->hpd = ffs((entry & 0x07033000) >> 12);
-		nv_connector->hpd = hpd[nv_connector->hpd];
-
-		nv_connector->type = nv_connector->dcb[0];
-		if (drm_conntype_from_dcb(nv_connector->type) ==
-					  DRM_MODE_CONNECTOR_Unknown) {
-			NV_WARN(dev, "unknown connector type %02x\n",
-				nv_connector->type);
-			nv_connector->type = DCB_CONNECTOR_NONE;
-		}
 
-		/* Gigabyte NX85T */
-		if (nv_match_device(dev, 0x0421, 0x1458, 0x344c)) {
-			if (nv_connector->type == DCB_CONNECTOR_HDMI_1)
-				nv_connector->type = DCB_CONNECTOR_DVI_I;
-		}
-
-		/* Gigabyte GV-NX86T512H */
-		if (nv_match_device(dev, 0x0402, 0x1458, 0x3455)) {
-			if (nv_connector->type == DCB_CONNECTOR_HDMI_1)
-				nv_connector->type = DCB_CONNECTOR_DVI_I;
-		}
-	} else {
-		nv_connector->type = DCB_CONNECTOR_NONE;
-		nv_connector->hpd = DCB_GPIO_UNUSED;
-	}
+	/* defaults, will get overridden in detect() */
+	connector->interlace_allowed = false;
+	connector->doublescan_allowed = false;
 
-	/* no vbios data, or an unknown dcb connector type - attempt to
-	 * figure out something suitable ourselves
-	 */
-	if (nv_connector->type == DCB_CONNECTOR_NONE) {
-		struct drm_nouveau_private *dev_priv = dev->dev_private;
-		struct dcb_table *dcbt = &dev_priv->vbios.dcb;
-		u32 encoders = 0;
-		int i;
-
-		for (i = 0; i < dcbt->entries; i++) {
-			if (dcbt->entry[i].connector == nv_connector->index)
-				encoders |= (1 << dcbt->entry[i].type);
-		}
+	drm_connector_init(dev, connector, funcs, type);
+	drm_connector_helper_add(connector, &nouveau_connector_helper_funcs);
 
-		if (encoders & (1 << OUTPUT_DP)) {
-			if (encoders & (1 << OUTPUT_TMDS))
-				nv_connector->type = DCB_CONNECTOR_DP;
-			else
-				nv_connector->type = DCB_CONNECTOR_eDP;
-		} else
-		if (encoders & (1 << OUTPUT_TMDS)) {
-			if (encoders & (1 << OUTPUT_ANALOG))
-				nv_connector->type = DCB_CONNECTOR_DVI_I;
-			else
-				nv_connector->type = DCB_CONNECTOR_DVI_D;
-		} else
-		if (encoders & (1 << OUTPUT_ANALOG)) {
-			nv_connector->type = DCB_CONNECTOR_VGA;
-		} else
-		if (encoders & (1 << OUTPUT_LVDS)) {
-			nv_connector->type = DCB_CONNECTOR_LVDS;
-		} else
-		if (encoders & (1 << OUTPUT_TV)) {
-			nv_connector->type = DCB_CONNECTOR_TV_0;
-		}
-	}
+	/* Check if we need dithering enabled */
+	if (connector->connector_type == DRM_MODE_CONNECTOR_LVDS) {
+		bool dummy, is_24bit = false;
 
-	type = drm_conntype_from_dcb(nv_connector->type);
-	if (type == DRM_MODE_CONNECTOR_LVDS) {
-		ret = nouveau_bios_parse_lvds_table(dev, 0, &dummy, &dummy);
+		ret = nouveau_bios_parse_lvds_table(dev, 0, &dummy, &is_24bit);
 		if (ret) {
-			NV_ERROR(dev, "Error parsing LVDS table, disabling\n");
-			kfree(nv_connector);
-			return ERR_PTR(ret);
+			NV_ERROR(dev, "Error parsing LVDS table, disabling "
+				 "LVDS\n");
+			goto fail;
 		}
 
-		funcs = &nouveau_connector_funcs_lvds;
-	} else {
-		funcs = &nouveau_connector_funcs;
+		nv_connector->use_dithering = !is_24bit;
 	}
 
-	/* defaults, will get overridden in detect() */
-	connector->interlace_allowed = false;
-	connector->doublescan_allowed = false;
-
-	drm_connector_init(dev, connector, funcs, type);
-	drm_connector_helper_add(connector, &nouveau_connector_helper_funcs);
-
 	/* Init DVI-I specific properties */
-	if (nv_connector->type == DCB_CONNECTOR_DVI_I)
+	if (dcb->type == DCB_CONNECTOR_DVI_I) {
+		drm_mode_create_dvi_i_properties(dev);
 		drm_connector_attach_property(connector, dev->mode_config.dvi_i_subconnector_property, 0);
-
-	/* Add overscan compensation options to digital outputs */
-	if (disp->underscan_property &&
-	    (nv_connector->type == DCB_CONNECTOR_DVI_D ||
-	     nv_connector->type == DCB_CONNECTOR_DVI_I ||
-	     nv_connector->type == DCB_CONNECTOR_HDMI_0 ||
-	     nv_connector->type == DCB_CONNECTOR_HDMI_1 ||
-	     nv_connector->type == DCB_CONNECTOR_DP ||
-	     nv_connector->type == DCB_CONNECTOR_DMS59_DP0 ||
-	     nv_connector->type == DCB_CONNECTOR_DMS59_DP1)) {
-		drm_connector_attach_property(connector,
-					      disp->underscan_property,
-					      UNDERSCAN_OFF);
-		drm_connector_attach_property(connector,
-					      disp->underscan_hborder_property,
-					      0);
-		drm_connector_attach_property(connector,
-					      disp->underscan_vborder_property,
-					      0);
+		drm_connector_attach_property(connector, dev->mode_config.dvi_i_select_subconnector_property, 0);
 	}
 
-	/* Add hue and saturation options */
-	if (disp->vibrant_hue_property)
-		drm_connector_attach_property(connector,
-					      disp->vibrant_hue_property,
-					      90);
-	if (disp->color_vibrance_property)
-		drm_connector_attach_property(connector,
-					      disp->color_vibrance_property,
-					      150);
-
-	switch (nv_connector->type) {
+	switch (dcb->type) {
 	case DCB_CONNECTOR_VGA:
 		if (dev_priv->card_type >= NV_50) {
 			drm_connector_attach_property(connector,
 					dev->mode_config.scaling_mode_property,
 					nv_connector->scaling_mode);
 		}
+		connector->polled = DRM_CONNECTOR_POLL_CONNECT;
 		/* fall-through */
 	case DCB_CONNECTOR_TV_0:
 	case DCB_CONNECTOR_TV_1:
@@ -1060,32 +880,39 @@ nouveau_connector_create(struct drm_device *dev, int index)
 		drm_connector_attach_property(connector,
 				dev->mode_config.scaling_mode_property,
 				nv_connector->scaling_mode);
-		if (disp->dithering_mode) {
-			nv_connector->dithering_mode = DITHERING_MODE_AUTO;
-			drm_connector_attach_property(connector,
-						disp->dithering_mode,
-						nv_connector->dithering_mode);
-		}
-		if (disp->dithering_depth) {
-			nv_connector->dithering_depth = DITHERING_DEPTH_AUTO;
-			drm_connector_attach_property(connector,
-						disp->dithering_depth,
-						nv_connector->dithering_depth);
+		drm_connector_attach_property(connector,
+				dev->mode_config.dithering_mode_property,
+				nv_connector->use_dithering ?
+				DRM_MODE_DITHERING_ON : DRM_MODE_DITHERING_OFF);
+
+		if (connector->connector_type != DRM_MODE_CONNECTOR_LVDS) {
+			if (dev_priv->card_type >= NV_50)
+				connector->polled = DRM_CONNECTOR_POLL_HPD;
+			else
+				connector->polled = DRM_CONNECTOR_POLL_CONNECT;
 		}
 		break;
 	}
 
-	connector->polled = DRM_CONNECTOR_POLL_CONNECT;
-	if (nv_connector->hpd != DCB_GPIO_UNUSED) {
-		ret = nouveau_gpio_isr_add(dev, 0, nv_connector->hpd, 0xff,
-					   nouveau_connector_hotplug,
-					   connector);
-		if (ret == 0)
-			connector->polled = DRM_CONNECTOR_POLL_HPD;
+	if (pgpio->irq_register) {
+		pgpio->irq_register(dev, nv_connector->dcb->gpio_tag,
+				    nouveau_connector_hotplug, connector);
 	}
 
 	drm_sysfs_connector_add(connector);
-	return connector;
+
+	if (connector->connector_type == DRM_MODE_CONNECTOR_LVDS ||
+	    connector->connector_type == DRM_MODE_CONNECTOR_eDP)
+		nouveau_backlight_init(connector);
+
+	dcb->drm = connector;
+	return dcb->drm;
+
+fail:
+	drm_connector_cleanup(connector);
+	kfree(connector);
+	return ERR_PTR(ret);
+
 }
 
 static void
@@ -1094,13 +921,22 @@ nouveau_connector_hotplug(void *data, int plugged)
 	struct drm_connector *connector = data;
 	struct drm_device *dev = connector->dev;
 
-	NV_DEBUG(dev, "%splugged %s\n", plugged ? "" : "un",
-		 drm_get_connector_name(connector));
+	NV_INFO(dev, "%splugged %s\n", plugged ? "" : "un",
+		drm_get_connector_name(connector));
 
-	if (plugged)
-		drm_helper_connector_dpms(connector, DRM_MODE_DPMS_ON);
-	else
-		drm_helper_connector_dpms(connector, DRM_MODE_DPMS_OFF);
+	if (connector->encoder && connector->encoder->crtc &&
+	    connector->encoder->crtc->enabled) {
+		struct nouveau_encoder *nv_encoder = nouveau_encoder(connector->encoder);
+		struct drm_encoder_helper_funcs *helper =
+			connector->encoder->helper_private;
+
+		if (nv_encoder->dcb->type == OUTPUT_DP) {
+			if (plugged)
+				helper->dpms(connector->encoder, DRM_MODE_DPMS_ON);
+			else
+				helper->dpms(connector->encoder, DRM_MODE_DPMS_OFF);
+		}
+	}
 
 	drm_helper_hpd_irq_event(dev);
 }
diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.h b/drivers/gpu/drm/nouveau/nouveau_connector.h
index e485702..711b1e9 100644
--- a/drivers/gpu/drm/nouveau/nouveau_connector.h
+++ b/drivers/gpu/drm/nouveau/nouveau_connector.h
@@ -30,43 +30,13 @@
 #include "drm_edid.h"
 #include "nouveau_i2c.h"
 
-enum nouveau_underscan_type {
-	UNDERSCAN_OFF,
-	UNDERSCAN_ON,
-	UNDERSCAN_AUTO,
-};
-
-/* the enum values specifically defined here match nv50/nvd0 hw values, and
- * the code relies on this
- */
-enum nouveau_dithering_mode {
-	DITHERING_MODE_OFF = 0x00,
-	DITHERING_MODE_ON = 0x01,
-	DITHERING_MODE_DYNAMIC2X2 = 0x10 | DITHERING_MODE_ON,
-	DITHERING_MODE_STATIC2X2 = 0x18 | DITHERING_MODE_ON,
-	DITHERING_MODE_TEMPORAL = 0x20 | DITHERING_MODE_ON,
-	DITHERING_MODE_AUTO
-};
-
-enum nouveau_dithering_depth {
-	DITHERING_DEPTH_6BPC = 0x00,
-	DITHERING_DEPTH_8BPC = 0x02,
-	DITHERING_DEPTH_AUTO
-};
-
 struct nouveau_connector {
 	struct drm_connector base;
-	enum dcb_connector_type type;
-	u8 index;
-	u8 *dcb;
-	u8 hpd;
 
-	int dithering_mode;
-	int dithering_depth;
+	struct dcb_connector_table_entry *dcb;
+
 	int scaling_mode;
-	enum nouveau_underscan_type underscan;
-	u32 underscan_hborder;
-	u32 underscan_vborder;
+	bool use_dithering;
 
 	struct nouveau_encoder *detected_encoder;
 	struct edid *edid;
diff --git a/drivers/gpu/drm/nouveau/nouveau_crtc.h b/drivers/gpu/drm/nouveau/nouveau_crtc.h
index e6d0d1e..cb1ce2a 100644
--- a/drivers/gpu/drm/nouveau/nouveau_crtc.h
+++ b/drivers/gpu/drm/nouveau/nouveau_crtc.h
@@ -32,11 +32,11 @@ struct nouveau_crtc {
 
 	int index;
 
+	struct drm_display_mode *mode;
+
 	uint32_t dpms_saved_fp_control;
 	uint32_t fp_users;
 	int saturation;
-	int color_vibrance;
-	int vibrant_hue;
 	int sharpness;
 	int last_dpms;
 
@@ -67,9 +67,8 @@ struct nouveau_crtc {
 		int depth;
 	} lut;
 
-	int (*set_dither)(struct nouveau_crtc *crtc, bool update);
-	int (*set_scale)(struct nouveau_crtc *crtc, bool update);
-	int (*set_color_vibrance)(struct nouveau_crtc *crtc, bool update);
+	int (*set_dither)(struct nouveau_crtc *crtc, bool on, bool update);
+	int (*set_scale)(struct nouveau_crtc *crtc, int mode, bool update);
 };
 
 static inline struct nouveau_crtc *nouveau_crtc(struct drm_crtc *crtc)
@@ -83,13 +82,14 @@ static inline struct drm_crtc *to_drm_crtc(struct nouveau_crtc *crtc)
 }
 
 int nv50_crtc_create(struct drm_device *dev, int index);
+int nv50_cursor_init(struct nouveau_crtc *);
+void nv50_cursor_fini(struct nouveau_crtc *);
 int nv50_crtc_cursor_set(struct drm_crtc *drm_crtc, struct drm_file *file_priv,
 			 uint32_t buffer_handle, uint32_t width,
 			 uint32_t height);
 int nv50_crtc_cursor_move(struct drm_crtc *drm_crtc, int x, int y);
 
 int nv04_cursor_init(struct nouveau_crtc *);
-int nv50_cursor_init(struct nouveau_crtc *);
 
 struct nouveau_connector *
 nouveau_crtc_connector_get(struct nouveau_crtc *crtc);
diff --git a/drivers/gpu/drm/nouveau/nouveau_debugfs.c b/drivers/gpu/drm/nouveau/nouveau_debugfs.c
index fa2ec49..8e15923 100644
--- a/drivers/gpu/drm/nouveau/nouveau_debugfs.c
+++ b/drivers/gpu/drm/nouveau/nouveau_debugfs.c
@@ -44,7 +44,7 @@ nouveau_debugfs_channel_info(struct seq_file *m, void *data)
 	seq_printf(m, "channel id    : %d\n", chan->id);
 
 	seq_printf(m, "cpu fifo state:\n");
-	seq_printf(m, "          base: 0x%10llx\n", chan->pushbuf_base);
+	seq_printf(m, "          base: 0x%08x\n", chan->pushbuf_base);
 	seq_printf(m, "           max: 0x%08x\n", chan->dma.max << 2);
 	seq_printf(m, "           cur: 0x%08x\n", chan->dma.cur << 2);
 	seq_printf(m, "           put: 0x%08x\n", chan->dma.put << 2);
@@ -178,7 +178,6 @@ static struct drm_info_list nouveau_debugfs_list[] = {
 	{ "memory", nouveau_debugfs_memory_info, 0, NULL },
 	{ "vbios.rom", nouveau_debugfs_vbios_image, 0, NULL },
 	{ "ttm_page_pool", ttm_page_alloc_debugfs, 0, NULL },
-	{ "ttm_dma_page_pool", ttm_dma_page_alloc_debugfs, 0, NULL },
 };
 #define NOUVEAU_DEBUGFS_ENTRIES ARRAY_SIZE(nouveau_debugfs_list)
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_display.c b/drivers/gpu/drm/nouveau/nouveau_display.c
index c01ae78..eb514ea 100644
--- a/drivers/gpu/drm/nouveau/nouveau_display.c
+++ b/drivers/gpu/drm/nouveau/nouveau_display.c
@@ -32,8 +32,6 @@
 #include "nouveau_hw.h"
 #include "nouveau_crtc.h"
 #include "nouveau_dma.h"
-#include "nouveau_connector.h"
-#include "nouveau_gpio.h"
 #include "nv50_display.h"
 
 static void
@@ -66,7 +64,7 @@ static const struct drm_framebuffer_funcs nouveau_framebuffer_funcs = {
 int
 nouveau_framebuffer_init(struct drm_device *dev,
 			 struct nouveau_framebuffer *nv_fb,
-			 struct drm_mode_fb_cmd2 *mode_cmd,
+			 struct drm_mode_fb_cmd *mode_cmd,
 			 struct nouveau_bo *nvbo)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
@@ -107,16 +105,13 @@ nouveau_framebuffer_init(struct drm_device *dev,
 		if (dev_priv->chipset == 0x50)
 			nv_fb->r_format |= (tile_flags << 8);
 
-		if (!tile_flags) {
-			if (dev_priv->card_type < NV_D0)
-				nv_fb->r_pitch = 0x00100000 | fb->pitches[0];
-			else
-				nv_fb->r_pitch = 0x01000000 | fb->pitches[0];
-		} else {
+		if (!tile_flags)
+			nv_fb->r_pitch = 0x00100000 | fb->pitch;
+		else {
 			u32 mode = nvbo->tile_mode;
 			if (dev_priv->card_type >= NV_C0)
 				mode >>= 4;
-			nv_fb->r_pitch = ((fb->pitches[0] / 4) << 4) | mode;
+			nv_fb->r_pitch = ((fb->pitch / 4) << 4) | mode;
 		}
 	}
 
@@ -126,13 +121,13 @@ nouveau_framebuffer_init(struct drm_device *dev,
 static struct drm_framebuffer *
 nouveau_user_framebuffer_create(struct drm_device *dev,
 				struct drm_file *file_priv,
-				struct drm_mode_fb_cmd2 *mode_cmd)
+				struct drm_mode_fb_cmd *mode_cmd)
 {
 	struct nouveau_framebuffer *nouveau_fb;
 	struct drm_gem_object *gem;
 	int ret;
 
-	gem = drm_gem_object_lookup(dev, file_priv, mode_cmd->handles[0]);
+	gem = drm_gem_object_lookup(dev, file_priv, mode_cmd->handle);
 	if (!gem)
 		return ERR_PTR(-ENOENT);
 
@@ -149,207 +144,11 @@ nouveau_user_framebuffer_create(struct drm_device *dev,
 	return &nouveau_fb->base;
 }
 
-static const struct drm_mode_config_funcs nouveau_mode_config_funcs = {
+const struct drm_mode_config_funcs nouveau_mode_config_funcs = {
 	.fb_create = nouveau_user_framebuffer_create,
 	.output_poll_changed = nouveau_fbcon_output_poll_changed,
 };
 
-
-struct nouveau_drm_prop_enum_list {
-	u8 gen_mask;
-	int type;
-	char *name;
-};
-
-static struct nouveau_drm_prop_enum_list underscan[] = {
-	{ 6, UNDERSCAN_AUTO, "auto" },
-	{ 6, UNDERSCAN_OFF, "off" },
-	{ 6, UNDERSCAN_ON, "on" },
-	{}
-};
-
-static struct nouveau_drm_prop_enum_list dither_mode[] = {
-	{ 7, DITHERING_MODE_AUTO, "auto" },
-	{ 7, DITHERING_MODE_OFF, "off" },
-	{ 1, DITHERING_MODE_ON, "on" },
-	{ 6, DITHERING_MODE_STATIC2X2, "static 2x2" },
-	{ 6, DITHERING_MODE_DYNAMIC2X2, "dynamic 2x2" },
-	{ 4, DITHERING_MODE_TEMPORAL, "temporal" },
-	{}
-};
-
-static struct nouveau_drm_prop_enum_list dither_depth[] = {
-	{ 6, DITHERING_DEPTH_AUTO, "auto" },
-	{ 6, DITHERING_DEPTH_6BPC, "6 bpc" },
-	{ 6, DITHERING_DEPTH_8BPC, "8 bpc" },
-	{}
-};
-
-#define PROP_ENUM(p,gen,n,list) do {                                           \
-	struct nouveau_drm_prop_enum_list *l = (list);                         \
-	int c = 0;                                                             \
-	while (l->gen_mask) {                                                  \
-		if (l->gen_mask & (1 << (gen)))                                \
-			c++;                                                   \
-		l++;                                                           \
-	}                                                                      \
-	if (c) {                                                               \
-		p = drm_property_create(dev, DRM_MODE_PROP_ENUM, n, c);        \
-		l = (list);                                                    \
-		c = 0;                                                         \
-		while (p && l->gen_mask) {                                     \
-			if (l->gen_mask & (1 << (gen))) {                      \
-				drm_property_add_enum(p, c, l->type, l->name); \
-				c++;                                           \
-			}                                                      \
-			l++;                                                   \
-		}                                                              \
-	}                                                                      \
-} while(0)
-
-int
-nouveau_display_init(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_display_engine *disp = &dev_priv->engine.display;
-	struct drm_connector *connector;
-	int ret;
-
-	ret = disp->init(dev);
-	if (ret)
-		return ret;
-
-	/* power on internal panel if it's not already.  the init tables of
-	 * some vbios default this to off for some reason, causing the
-	 * panel to not work after resume
-	 */
-	if (nouveau_gpio_func_get(dev, DCB_GPIO_PANEL_POWER) == 0) {
-		nouveau_gpio_func_set(dev, DCB_GPIO_PANEL_POWER, true);
-		msleep(300);
-	}
-
-	/* enable polling for external displays */
-	drm_kms_helper_poll_enable(dev);
-
-	/* enable hotplug interrupts */
-	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
-		struct nouveau_connector *conn = nouveau_connector(connector);
-		nouveau_gpio_irq(dev, 0, conn->hpd, 0xff, true);
-	}
-
-	return ret;
-}
-
-void
-nouveau_display_fini(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_display_engine *disp = &dev_priv->engine.display;
-	struct drm_connector *connector;
-
-	/* disable hotplug interrupts */
-	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
-		struct nouveau_connector *conn = nouveau_connector(connector);
-		nouveau_gpio_irq(dev, 0, conn->hpd, 0xff, false);
-	}
-
-	drm_kms_helper_poll_disable(dev);
-	disp->fini(dev);
-}
-
-int
-nouveau_display_create(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_display_engine *disp = &dev_priv->engine.display;
-	int ret, gen;
-
-	drm_mode_config_init(dev);
-	drm_mode_create_scaling_mode_property(dev);
-	drm_mode_create_dvi_i_properties(dev);
-
-	if (dev_priv->card_type < NV_50)
-		gen = 0;
-	else
-	if (dev_priv->card_type < NV_D0)
-		gen = 1;
-	else
-		gen = 2;
-
-	PROP_ENUM(disp->dithering_mode, gen, "dithering mode", dither_mode);
-	PROP_ENUM(disp->dithering_depth, gen, "dithering depth", dither_depth);
-	PROP_ENUM(disp->underscan_property, gen, "underscan", underscan);
-
-	disp->underscan_hborder_property =
-		drm_property_create_range(dev, 0, "underscan hborder", 0, 128);
-
-	disp->underscan_vborder_property =
-		drm_property_create_range(dev, 0, "underscan vborder", 0, 128);
-
-	if (gen == 1) {
-		disp->vibrant_hue_property =
-			drm_property_create(dev, DRM_MODE_PROP_RANGE,
-					    "vibrant hue", 2);
-		disp->vibrant_hue_property->values[0] = 0;
-		disp->vibrant_hue_property->values[1] = 180; /* -90..+90 */
-
-		disp->color_vibrance_property =
-			drm_property_create(dev, DRM_MODE_PROP_RANGE,
-					    "color vibrance", 2);
-		disp->color_vibrance_property->values[0] = 0;
-		disp->color_vibrance_property->values[1] = 200; /* -100..+100 */
-	}
-
-	dev->mode_config.funcs = (void *)&nouveau_mode_config_funcs;
-	dev->mode_config.fb_base = pci_resource_start(dev->pdev, 1);
-
-	dev->mode_config.min_width = 0;
-	dev->mode_config.min_height = 0;
-	if (dev_priv->card_type < NV_10) {
-		dev->mode_config.max_width = 2048;
-		dev->mode_config.max_height = 2048;
-	} else
-	if (dev_priv->card_type < NV_50) {
-		dev->mode_config.max_width = 4096;
-		dev->mode_config.max_height = 4096;
-	} else {
-		dev->mode_config.max_width = 8192;
-		dev->mode_config.max_height = 8192;
-	}
-
-	dev->mode_config.preferred_depth = 24;
-	dev->mode_config.prefer_shadow = 1;
-
-	drm_kms_helper_poll_init(dev);
-	drm_kms_helper_poll_disable(dev);
-
-	ret = disp->create(dev);
-	if (ret)
-		return ret;
-
-	if (dev->mode_config.num_crtc) {
-		ret = drm_vblank_init(dev, dev->mode_config.num_crtc);
-		if (ret)
-			return ret;
-	}
-
-	return ret;
-}
-
-void
-nouveau_display_destroy(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_display_engine *disp = &dev_priv->engine.display;
-
-	drm_vblank_cleanup(dev);
-
-	disp->destroy(dev);
-
-	drm_kms_helper_poll_fini(dev);
-	drm_mode_config_cleanup(dev);
-}
-
 int
 nouveau_vblank_enable(struct drm_device *dev, int crtc)
 {
@@ -492,7 +291,7 @@ nouveau_crtc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 	/* Initialize a page flip struct */
 	*s = (struct nouveau_page_flip_state)
 		{ { }, event, nouveau_crtc(crtc)->index,
-		  fb->bits_per_pixel, fb->pitches[0], crtc->x, crtc->y,
+		  fb->bits_per_pixel, fb->pitch, crtc->x, crtc->y,
 		  new_bo->bo.offset };
 
 	/* Choose the channel the flip will be handled in */
@@ -503,10 +302,7 @@ nouveau_crtc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 
 	/* Emit a page flip */
 	if (dev_priv->card_type >= NV_50) {
-		if (dev_priv->card_type >= NV_D0)
-			ret = nvd0_display_flip_next(crtc, fb, chan, 0);
-		else
-			ret = nv50_display_flip_next(crtc, fb, chan);
+		ret = nv50_display_flip_next(crtc, fb, chan);
 		if (ret) {
 			nouveau_channel_put(&chan);
 			goto fail_unreserve;
@@ -570,48 +366,3 @@ nouveau_finish_page_flip(struct nouveau_channel *chan,
 	spin_unlock_irqrestore(&dev->event_lock, flags);
 	return 0;
 }
-
-int
-nouveau_display_dumb_create(struct drm_file *file_priv, struct drm_device *dev,
-			    struct drm_mode_create_dumb *args)
-{
-	struct nouveau_bo *bo;
-	int ret;
-
-	args->pitch = roundup(args->width * (args->bpp / 8), 256);
-	args->size = args->pitch * args->height;
-	args->size = roundup(args->size, PAGE_SIZE);
-
-	ret = nouveau_gem_new(dev, args->size, 0, TTM_PL_FLAG_VRAM, 0, 0, &bo);
-	if (ret)
-		return ret;
-
-	ret = drm_gem_handle_create(file_priv, bo->gem, &args->handle);
-	drm_gem_object_unreference_unlocked(bo->gem);
-	return ret;
-}
-
-int
-nouveau_display_dumb_destroy(struct drm_file *file_priv, struct drm_device *dev,
-			     uint32_t handle)
-{
-	return drm_gem_handle_delete(file_priv, handle);
-}
-
-int
-nouveau_display_dumb_map_offset(struct drm_file *file_priv,
-				struct drm_device *dev,
-				uint32_t handle, uint64_t *poffset)
-{
-	struct drm_gem_object *gem;
-
-	gem = drm_gem_object_lookup(dev, file_priv, handle);
-	if (gem) {
-		struct nouveau_bo *bo = gem->driver_private;
-		*poffset = bo->bo.addr_space_offset;
-		drm_gem_object_unreference_unlocked(gem);
-		return 0;
-	}
-
-	return -ENOENT;
-}
diff --git a/drivers/gpu/drm/nouveau/nouveau_dma.c b/drivers/gpu/drm/nouveau/nouveau_dma.c
index 4c2e4e5..568caed 100644
--- a/drivers/gpu/drm/nouveau/nouveau_dma.c
+++ b/drivers/gpu/drm/nouveau/nouveau_dma.c
@@ -134,13 +134,11 @@ OUT_RINGp(struct nouveau_channel *chan, const void *data, unsigned nr_dwords)
  *  -EBUSY if timeout exceeded
  */
 static inline int
-READ_GET(struct nouveau_channel *chan, uint64_t *prev_get, int *timeout)
+READ_GET(struct nouveau_channel *chan, uint32_t *prev_get, uint32_t *timeout)
 {
-	uint64_t val;
+	uint32_t val;
 
 	val = nvchan_rd32(chan, chan->user_get);
-        if (chan->user_get_hi)
-                val |= (uint64_t)nvchan_rd32(chan, chan->user_get_hi) << 32;
 
 	/* reset counter as long as GET is still advancing, this is
 	 * to avoid misdetecting a GPU lockup if the GPU happens to
@@ -169,13 +167,8 @@ nv50_dma_push(struct nouveau_channel *chan, struct nouveau_bo *bo,
 	      int delta, int length)
 {
 	struct nouveau_bo *pb = chan->pushbuf_bo;
-	struct nouveau_vma *vma;
+	uint64_t offset = bo->bo.offset + delta;
 	int ip = (chan->dma.ib_put * 2) + chan->dma.ib_base;
-	u64 offset;
-
-	vma = nouveau_bo_vma_find(bo, chan->vm);
-	BUG_ON(!vma);
-	offset = vma->offset + delta;
 
 	BUG_ON(chan->dma.ib_free < 1);
 	nouveau_bo_wr32(pb, ip++, lower_32_bits(offset));
@@ -220,8 +213,8 @@ nv50_dma_push_wait(struct nouveau_channel *chan, int count)
 static int
 nv50_dma_wait(struct nouveau_channel *chan, int slots, int count)
 {
-	uint64_t prev_get = 0;
-	int ret, cnt = 0;
+	uint32_t cnt = 0, prev_get = 0;
+	int ret;
 
 	ret = nv50_dma_push_wait(chan, slots + 1);
 	if (unlikely(ret))
@@ -263,8 +256,8 @@ nv50_dma_wait(struct nouveau_channel *chan, int slots, int count)
 int
 nouveau_dma_wait(struct nouveau_channel *chan, int slots, int size)
 {
-	uint64_t prev_get = 0;
-	int cnt = 0, get;
+	uint32_t prev_get = 0, cnt = 0;
+	int get;
 
 	if (chan->dma.ib_max)
 		return nv50_dma_wait(chan, slots, size);
diff --git a/drivers/gpu/drm/nouveau/nouveau_dp.c b/drivers/gpu/drm/nouveau/nouveau_dp.c
index 302b2f7..7beb82a 100644
--- a/drivers/gpu/drm/nouveau/nouveau_dp.c
+++ b/drivers/gpu/drm/nouveau/nouveau_dp.c
@@ -28,525 +28,557 @@
 #include "nouveau_i2c.h"
 #include "nouveau_connector.h"
 #include "nouveau_encoder.h"
-#include "nouveau_crtc.h"
-#include "nouveau_gpio.h"
-
-/******************************************************************************
- * aux channel util functions
- *****************************************************************************/
-#define AUX_DBG(fmt, args...) do {                                             \
-	if (nouveau_reg_debug & NOUVEAU_REG_DEBUG_AUXCH) {                     \
-		NV_PRINTK(KERN_DEBUG, dev, "AUXCH(%d): " fmt, ch, ##args);     \
-	}                                                                      \
-} while (0)
-#define AUX_ERR(fmt, args...) NV_ERROR(dev, "AUXCH(%d): " fmt, ch, ##args)
-
-static void
-auxch_fini(struct drm_device *dev, int ch)
+
+static int
+auxch_rd(struct drm_encoder *encoder, int address, uint8_t *buf, int size)
 {
-	nv_mask(dev, 0x00e4e4 + (ch * 0x50), 0x00310000, 0x00000000);
+	struct drm_device *dev = encoder->dev;
+	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
+	struct nouveau_i2c_chan *auxch;
+	int ret;
+
+	auxch = nouveau_i2c_find(dev, nv_encoder->dcb->i2c_index);
+	if (!auxch)
+		return -ENODEV;
+
+	ret = nouveau_dp_auxch(auxch, 9, address, buf, size);
+	if (ret)
+		return ret;
+
+	return 0;
 }
 
 static int
-auxch_init(struct drm_device *dev, int ch)
+auxch_wr(struct drm_encoder *encoder, int address, uint8_t *buf, int size)
 {
-	const u32 unksel = 1; /* nfi which to use, or if it matters.. */
-	const u32 ureq = unksel ? 0x00100000 : 0x00200000;
-	const u32 urep = unksel ? 0x01000000 : 0x02000000;
-	u32 ctrl, timeout;
-
-	/* wait up to 1ms for any previous transaction to be done... */
-	timeout = 1000;
-	do {
-		ctrl = nv_rd32(dev, 0x00e4e4 + (ch * 0x50));
-		udelay(1);
-		if (!timeout--) {
-			AUX_ERR("begin idle timeout 0x%08x", ctrl);
-			return -EBUSY;
-		}
-	} while (ctrl & 0x03010000);
-
-	/* set some magic, and wait up to 1ms for it to appear */
-	nv_mask(dev, 0x00e4e4 + (ch * 0x50), 0x00300000, ureq);
-	timeout = 1000;
-	do {
-		ctrl = nv_rd32(dev, 0x00e4e4 + (ch * 0x50));
-		udelay(1);
-		if (!timeout--) {
-			AUX_ERR("magic wait 0x%08x\n", ctrl);
-			auxch_fini(dev, ch);
-			return -EBUSY;
-		}
-	} while ((ctrl & 0x03000000) != urep);
+	struct drm_device *dev = encoder->dev;
+	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
+	struct nouveau_i2c_chan *auxch;
+	int ret;
 
-	return 0;
+	auxch = nouveau_i2c_find(dev, nv_encoder->dcb->i2c_index);
+	if (!auxch)
+		return -ENODEV;
+
+	ret = nouveau_dp_auxch(auxch, 8, address, buf, size);
+	return ret;
 }
 
 static int
-auxch_tx(struct drm_device *dev, int ch, u8 type, u32 addr, u8 *data, u8 size)
+nouveau_dp_lane_count_set(struct drm_encoder *encoder, uint8_t cmd)
 {
-	u32 ctrl, stat, timeout, retries;
-	u32 xbuf[4] = {};
-	int ret, i;
+	struct drm_device *dev = encoder->dev;
+	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
+	uint32_t tmp;
+	int or = nv_encoder->or, link = !(nv_encoder->dcb->sorconf.link & 1);
+
+	tmp  = nv_rd32(dev, NV50_SOR_DP_CTRL(or, link));
+	tmp &= ~(NV50_SOR_DP_CTRL_ENHANCED_FRAME_ENABLED |
+		 NV50_SOR_DP_CTRL_LANE_MASK);
+	tmp |= ((1 << (cmd & DP_LANE_COUNT_MASK)) - 1) << 16;
+	if (cmd & DP_LANE_COUNT_ENHANCED_FRAME_EN)
+		tmp |= NV50_SOR_DP_CTRL_ENHANCED_FRAME_ENABLED;
+	nv_wr32(dev, NV50_SOR_DP_CTRL(or, link), tmp);
+
+	return auxch_wr(encoder, DP_LANE_COUNT_SET, &cmd, 1);
+}
 
-	AUX_DBG("%d: 0x%08x %d\n", type, addr, size);
+static int
+nouveau_dp_link_bw_set(struct drm_encoder *encoder, uint8_t cmd)
+{
+	struct drm_device *dev = encoder->dev;
+	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
+	uint32_t tmp;
+	int reg = 0x614300 + (nv_encoder->or * 0x800);
+
+	tmp  = nv_rd32(dev, reg);
+	tmp &= 0xfff3ffff;
+	if (cmd == DP_LINK_BW_2_7)
+		tmp |= 0x00040000;
+	nv_wr32(dev, reg, tmp);
+
+	return auxch_wr(encoder, DP_LINK_BW_SET, &cmd, 1);
+}
+
+static int
+nouveau_dp_link_train_set(struct drm_encoder *encoder, int pattern)
+{
+	struct drm_device *dev = encoder->dev;
+	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
+	uint32_t tmp;
+	uint8_t cmd;
+	int or = nv_encoder->or, link = !(nv_encoder->dcb->sorconf.link & 1);
+	int ret;
+
+	tmp  = nv_rd32(dev, NV50_SOR_DP_CTRL(or, link));
+	tmp &= ~NV50_SOR_DP_CTRL_TRAINING_PATTERN;
+	tmp |= (pattern << 24);
+	nv_wr32(dev, NV50_SOR_DP_CTRL(or, link), tmp);
 
-	ret = auxch_init(dev, ch);
+	ret = auxch_rd(encoder, DP_TRAINING_PATTERN_SET, &cmd, 1);
 	if (ret)
-		goto out;
+		return ret;
+	cmd &= ~DP_TRAINING_PATTERN_MASK;
+	cmd |= (pattern & DP_TRAINING_PATTERN_MASK);
+	return auxch_wr(encoder, DP_TRAINING_PATTERN_SET, &cmd, 1);
+}
 
-	stat = nv_rd32(dev, 0x00e4e8 + (ch * 0x50));
-	if (!(stat & 0x10000000)) {
-		AUX_DBG("sink not detected\n");
-		ret = -ENXIO;
-		goto out;
-	}
+static int
+nouveau_dp_max_voltage_swing(struct drm_encoder *encoder)
+{
+	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
+	struct drm_device *dev = encoder->dev;
+	struct bit_displayport_encoder_table_entry *dpse;
+	struct bit_displayport_encoder_table *dpe;
+	int i, dpe_headerlen, max_vs = 0;
 
-	if (!(type & 1)) {
-		memcpy(xbuf, data, size);
-		for (i = 0; i < 16; i += 4) {
-			AUX_DBG("wr 0x%08x\n", xbuf[i / 4]);
-			nv_wr32(dev, 0x00e4c0 + (ch * 0x50) + i, xbuf[i / 4]);
-		}
+	dpe = nouveau_bios_dp_table(dev, nv_encoder->dcb, &dpe_headerlen);
+	if (!dpe)
+		return false;
+	dpse = (void *)((char *)dpe + dpe_headerlen);
+
+	for (i = 0; i < dpe_headerlen; i++, dpse++) {
+		if (dpse->vs_level > max_vs)
+			max_vs = dpse->vs_level;
 	}
 
-	ctrl  = nv_rd32(dev, 0x00e4e4 + (ch * 0x50));
-	ctrl &= ~0x0001f0ff;
-	ctrl |= type << 12;
-	ctrl |= size - 1;
-	nv_wr32(dev, 0x00e4e0 + (ch * 0x50), addr);
-
-	/* retry transaction a number of times on failure... */
-	ret = -EREMOTEIO;
-	for (retries = 0; retries < 32; retries++) {
-		/* reset, and delay a while if this is a retry */
-		nv_wr32(dev, 0x00e4e4 + (ch * 0x50), 0x80000000 | ctrl);
-		nv_wr32(dev, 0x00e4e4 + (ch * 0x50), 0x00000000 | ctrl);
-		if (retries)
-			udelay(400);
-
-		/* transaction request, wait up to 1ms for it to complete */
-		nv_wr32(dev, 0x00e4e4 + (ch * 0x50), 0x00010000 | ctrl);
-
-		timeout = 1000;
-		do {
-			ctrl = nv_rd32(dev, 0x00e4e4 + (ch * 0x50));
-			udelay(1);
-			if (!timeout--) {
-				AUX_ERR("tx req timeout 0x%08x\n", ctrl);
-				goto out;
-			}
-		} while (ctrl & 0x00010000);
+	return max_vs;
+}
 
-		/* read status, and check if transaction completed ok */
-		stat = nv_mask(dev, 0x00e4e8 + (ch * 0x50), 0, 0);
-		if (!(stat & 0x000f0f00)) {
-			ret = 0;
-			break;
-		}
+static int
+nouveau_dp_max_pre_emphasis(struct drm_encoder *encoder, int vs)
+{
+	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
+	struct drm_device *dev = encoder->dev;
+	struct bit_displayport_encoder_table_entry *dpse;
+	struct bit_displayport_encoder_table *dpe;
+	int i, dpe_headerlen, max_pre = 0;
 
-		AUX_DBG("%02d 0x%08x 0x%08x\n", retries, ctrl, stat);
-	}
+	dpe = nouveau_bios_dp_table(dev, nv_encoder->dcb, &dpe_headerlen);
+	if (!dpe)
+		return false;
+	dpse = (void *)((char *)dpe + dpe_headerlen);
 
-	if (type & 1) {
-		for (i = 0; i < 16; i += 4) {
-			xbuf[i / 4] = nv_rd32(dev, 0x00e4d0 + (ch * 0x50) + i);
-			AUX_DBG("rd 0x%08x\n", xbuf[i / 4]);
-		}
-		memcpy(data, xbuf, size);
+	for (i = 0; i < dpe_headerlen; i++, dpse++) {
+		if (dpse->vs_level != vs)
+			continue;
+
+		if (dpse->pre_level > max_pre)
+			max_pre = dpse->pre_level;
 	}
 
-out:
-	auxch_fini(dev, ch);
-	return ret;
+	return max_pre;
 }
 
-u8 *
-nouveau_dp_bios_data(struct drm_device *dev, struct dcb_entry *dcb, u8 **entry)
+static bool
+nouveau_dp_link_train_adjust(struct drm_encoder *encoder, uint8_t *config)
 {
-	struct bit_entry d;
-	u8 *table;
-	int i;
+	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
+	struct drm_device *dev = encoder->dev;
+	struct bit_displayport_encoder_table *dpe;
+	int ret, i, dpe_headerlen, vs = 0, pre = 0;
+	uint8_t request[2];
 
-	if (bit_table(dev, 'd', &d)) {
-		NV_ERROR(dev, "BIT 'd' table not found\n");
-		return NULL;
-	}
+	dpe = nouveau_bios_dp_table(dev, nv_encoder->dcb, &dpe_headerlen);
+	if (!dpe)
+		return false;
 
-	if (d.version != 1) {
-		NV_ERROR(dev, "BIT 'd' table version %d unknown\n", d.version);
-		return NULL;
-	}
+	ret = auxch_rd(encoder, DP_ADJUST_REQUEST_LANE0_1, request, 2);
+	if (ret)
+		return false;
+
+	NV_DEBUG_KMS(dev, "\t\tadjust 0x%02x 0x%02x\n", request[0], request[1]);
 
-	table = ROMPTR(dev, d.data[0]);
-	if (!table) {
-		NV_ERROR(dev, "displayport table pointer invalid\n");
-		return NULL;
+	/* Keep all lanes at the same level.. */
+	for (i = 0; i < nv_encoder->dp.link_nr; i++) {
+		int lane_req = (request[i >> 1] >> ((i & 1) << 2)) & 0xf;
+		int lane_vs = lane_req & 3;
+		int lane_pre = (lane_req >> 2) & 3;
+
+		if (lane_vs > vs)
+			vs = lane_vs;
+		if (lane_pre > pre)
+			pre = lane_pre;
 	}
 
-	switch (table[0]) {
-	case 0x20:
-	case 0x21:
-	case 0x30:
-		break;
-	default:
-		NV_ERROR(dev, "displayport table 0x%02x unknown\n", table[0]);
-		return NULL;
+	if (vs >= nouveau_dp_max_voltage_swing(encoder)) {
+		vs  = nouveau_dp_max_voltage_swing(encoder);
+		vs |= 4;
 	}
 
-	for (i = 0; i < table[3]; i++) {
-		*entry = ROMPTR(dev, table[table[1] + (i * table[2])]);
-		if (*entry && bios_encoder_match(dcb, ROM32((*entry)[0])))
-			return table;
+	if (pre >= nouveau_dp_max_pre_emphasis(encoder, vs & 3)) {
+		pre  = nouveau_dp_max_pre_emphasis(encoder, vs & 3);
+		pre |= 4;
 	}
 
-	NV_ERROR(dev, "displayport encoder table not found\n");
-	return NULL;
-}
+	/* Update the configuration for all lanes.. */
+	for (i = 0; i < nv_encoder->dp.link_nr; i++)
+		config[i] = (pre << 3) | vs;
 
-/******************************************************************************
- * link training
- *****************************************************************************/
-struct dp_state {
-	struct dp_train_func *func;
-	struct dcb_entry *dcb;
-	int auxch;
-	int crtc;
-	u8 *dpcd;
-	int link_nr;
-	u32 link_bw;
-	u8  stat[6];
-	u8  conf[4];
-};
+	return true;
+}
 
-static void
-dp_set_link_config(struct drm_device *dev, struct dp_state *dp)
+static bool
+nouveau_dp_link_train_commit(struct drm_encoder *encoder, uint8_t *config)
 {
-	u8 sink[2];
+	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
+	struct drm_device *dev = encoder->dev;
+	struct bit_displayport_encoder_table_entry *dpse;
+	struct bit_displayport_encoder_table *dpe;
+	int or = nv_encoder->or, link = !(nv_encoder->dcb->sorconf.link & 1);
+	int dpe_headerlen, ret, i;
 
-	NV_DEBUG_KMS(dev, "%d lanes at %d KB/s\n", dp->link_nr, dp->link_bw);
+	NV_DEBUG_KMS(dev, "\t\tconfig 0x%02x 0x%02x 0x%02x 0x%02x\n",
+		 config[0], config[1], config[2], config[3]);
 
-	/* set desired link configuration on the source */
-	dp->func->link_set(dev, dp->dcb, dp->crtc, dp->link_nr, dp->link_bw,
-			   dp->dpcd[2] & DP_ENHANCED_FRAME_CAP);
+	dpe = nouveau_bios_dp_table(dev, nv_encoder->dcb, &dpe_headerlen);
+	if (!dpe)
+		return false;
+	dpse = (void *)((char *)dpe + dpe_headerlen);
 
-	/* inform the sink of the new configuration */
-	sink[0] = dp->link_bw / 27000;
-	sink[1] = dp->link_nr;
-	if (dp->dpcd[2] & DP_ENHANCED_FRAME_CAP)
-		sink[1] |= DP_LANE_COUNT_ENHANCED_FRAME_EN;
+	for (i = 0; i < dpe->record_nr; i++, dpse++) {
+		if (dpse->vs_level == (config[0] & 3) &&
+		    dpse->pre_level == ((config[0] >> 3) & 3))
+			break;
+	}
+	BUG_ON(i == dpe->record_nr);
+
+	for (i = 0; i < nv_encoder->dp.link_nr; i++) {
+		const int shift[4] = { 16, 8, 0, 24 };
+		uint32_t mask = 0xff << shift[i];
+		uint32_t reg0, reg1, reg2;
+
+		reg0  = nv_rd32(dev, NV50_SOR_DP_UNK118(or, link)) & ~mask;
+		reg0 |= (dpse->reg0 << shift[i]);
+		reg1  = nv_rd32(dev, NV50_SOR_DP_UNK120(or, link)) & ~mask;
+		reg1 |= (dpse->reg1 << shift[i]);
+		reg2  = nv_rd32(dev, NV50_SOR_DP_UNK130(or, link)) & 0xffff00ff;
+		reg2 |= (dpse->reg2 << 8);
+		nv_wr32(dev, NV50_SOR_DP_UNK118(or, link), reg0);
+		nv_wr32(dev, NV50_SOR_DP_UNK120(or, link), reg1);
+		nv_wr32(dev, NV50_SOR_DP_UNK130(or, link), reg2);
+	}
+
+	ret = auxch_wr(encoder, DP_TRAINING_LANE0_SET, config, 4);
+	if (ret)
+		return false;
 
-	auxch_tx(dev, dp->auxch, 8, DP_LINK_BW_SET, sink, 2);
+	return true;
 }
 
-static void
-dp_set_training_pattern(struct drm_device *dev, struct dp_state *dp, u8 pattern)
+bool
+nouveau_dp_link_train(struct drm_encoder *encoder)
 {
-	u8 sink_tp;
-
-	NV_DEBUG_KMS(dev, "training pattern %d\n", pattern);
-
-	dp->func->train_set(dev, dp->dcb, pattern);
+	struct drm_device *dev = encoder->dev;
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio;
+	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
+	struct nouveau_connector *nv_connector;
+	struct bit_displayport_encoder_table *dpe;
+	int dpe_headerlen;
+	uint8_t config[4], status[3];
+	bool cr_done, cr_max_vs, eq_done, hpd_state;
+	int ret = 0, i, tries, voltage;
 
-	auxch_tx(dev, dp->auxch, 9, DP_TRAINING_PATTERN_SET, &sink_tp, 1);
-	sink_tp &= ~DP_TRAINING_PATTERN_MASK;
-	sink_tp |= pattern;
-	auxch_tx(dev, dp->auxch, 8, DP_TRAINING_PATTERN_SET, &sink_tp, 1);
-}
+	NV_DEBUG_KMS(dev, "link training!!\n");
 
-static int
-dp_link_train_commit(struct drm_device *dev, struct dp_state *dp)
-{
-	int i;
+	nv_connector = nouveau_encoder_connector_get(nv_encoder);
+	if (!nv_connector)
+		return false;
 
-	for (i = 0; i < dp->link_nr; i++) {
-		u8 lane = (dp->stat[4 + (i >> 1)] >> ((i & 1) * 4)) & 0xf;
-		u8 lpre = (lane & 0x0c) >> 2;
-		u8 lvsw = (lane & 0x03) >> 0;
+	dpe = nouveau_bios_dp_table(dev, nv_encoder->dcb, &dpe_headerlen);
+	if (!dpe) {
+		NV_ERROR(dev, "SOR-%d: no DP encoder table!\n", nv_encoder->or);
+		return false;
+	}
 
-		dp->conf[i] = (lpre << 3) | lvsw;
-		if (lvsw == DP_TRAIN_VOLTAGE_SWING_1200)
-			dp->conf[i] |= DP_TRAIN_MAX_SWING_REACHED;
-		if ((lpre << 3) == DP_TRAIN_PRE_EMPHASIS_9_5)
-			dp->conf[i] |= DP_TRAIN_MAX_PRE_EMPHASIS_REACHED;
+	/* disable hotplug detect, this flips around on some panels during
+	 * link training.
+	 */
+	hpd_state = pgpio->irq_enable(dev, nv_connector->dcb->gpio_tag, false);
 
-		NV_DEBUG_KMS(dev, "config lane %d %02x\n", i, dp->conf[i]);
-		dp->func->train_adj(dev, dp->dcb, i, lvsw, lpre);
+	if (dpe->script0) {
+		NV_DEBUG_KMS(dev, "SOR-%d: running DP script 0\n", nv_encoder->or);
+		nouveau_bios_run_init_table(dev, le16_to_cpu(dpe->script0),
+					    nv_encoder->dcb);
 	}
 
-	return auxch_tx(dev, dp->auxch, 8, DP_TRAINING_LANE0_SET, dp->conf, 4);
-}
+train:
+	cr_done = eq_done = false;
 
-static int
-dp_link_train_update(struct drm_device *dev, struct dp_state *dp, u32 delay)
-{
-	int ret;
+	/* set link configuration */
+	NV_DEBUG_KMS(dev, "\tbegin train: bw %d, lanes %d\n",
+		 nv_encoder->dp.link_bw, nv_encoder->dp.link_nr);
+
+	ret = nouveau_dp_link_bw_set(encoder, nv_encoder->dp.link_bw);
+	if (ret)
+		return false;
 
-	udelay(delay);
+	config[0] = nv_encoder->dp.link_nr;
+	if (nv_encoder->dp.dpcd_version >= 0x11 &&
+	    nv_encoder->dp.enhanced_frame)
+		config[0] |= DP_LANE_COUNT_ENHANCED_FRAME_EN;
 
-	ret = auxch_tx(dev, dp->auxch, 9, DP_LANE0_1_STATUS, dp->stat, 6);
+	ret = nouveau_dp_lane_count_set(encoder, config[0]);
 	if (ret)
-		return ret;
+		return false;
 
-	NV_DEBUG_KMS(dev, "status %02x %02x %02x %02x %02x %02x\n",
-		     dp->stat[0], dp->stat[1], dp->stat[2], dp->stat[3],
-		     dp->stat[4], dp->stat[5]);
-	return 0;
-}
+	/* clock recovery */
+	NV_DEBUG_KMS(dev, "\tbegin cr\n");
+	ret = nouveau_dp_link_train_set(encoder, DP_TRAINING_PATTERN_1);
+	if (ret)
+		goto stop;
 
-static int
-dp_link_train_cr(struct drm_device *dev, struct dp_state *dp)
-{
-	bool cr_done = false, abort = false;
-	int voltage = dp->conf[0] & DP_TRAIN_VOLTAGE_SWING_MASK;
-	int tries = 0, i;
+	tries = 0;
+	voltage = -1;
+	memset(config, 0x00, sizeof(config));
+	for (;;) {
+		if (!nouveau_dp_link_train_commit(encoder, config))
+			break;
 
-	dp_set_training_pattern(dev, dp, DP_TRAINING_PATTERN_1);
+		udelay(100);
 
-	do {
-		if (dp_link_train_commit(dev, dp) ||
-		    dp_link_train_update(dev, dp, 100))
+		ret = auxch_rd(encoder, DP_LANE0_1_STATUS, status, 2);
+		if (ret)
 			break;
+		NV_DEBUG_KMS(dev, "\t\tstatus: 0x%02x 0x%02x\n",
+			 status[0], status[1]);
 
 		cr_done = true;
-		for (i = 0; i < dp->link_nr; i++) {
-			u8 lane = (dp->stat[i >> 1] >> ((i & 1) * 4)) & 0xf;
+		cr_max_vs = false;
+		for (i = 0; i < nv_encoder->dp.link_nr; i++) {
+			int lane = (status[i >> 1] >> ((i & 1) * 4)) & 0xf;
+
 			if (!(lane & DP_LANE_CR_DONE)) {
 				cr_done = false;
-				if (dp->conf[i] & DP_TRAIN_MAX_SWING_REACHED)
-					abort = true;
+				if (config[i] & DP_TRAIN_MAX_PRE_EMPHASIS_REACHED)
+					cr_max_vs = true;
 				break;
 			}
 		}
 
-		if ((dp->conf[0] & DP_TRAIN_VOLTAGE_SWING_MASK) != voltage) {
-			voltage = dp->conf[0] & DP_TRAIN_VOLTAGE_SWING_MASK;
+		if ((config[0] & DP_TRAIN_VOLTAGE_SWING_MASK) != voltage) {
+			voltage = config[0] & DP_TRAIN_VOLTAGE_SWING_MASK;
 			tries = 0;
 		}
-	} while (!cr_done && !abort && ++tries < 5);
 
-	return cr_done ? 0 : -1;
-}
+		if (cr_done || cr_max_vs || (++tries == 5))
+			break;
 
-static int
-dp_link_train_eq(struct drm_device *dev, struct dp_state *dp)
-{
-	bool eq_done, cr_done = true;
-	int tries = 0, i;
+		if (!nouveau_dp_link_train_adjust(encoder, config))
+			break;
+	}
+
+	if (!cr_done)
+		goto stop;
+
+	/* channel equalisation */
+	NV_DEBUG_KMS(dev, "\tbegin eq\n");
+	ret = nouveau_dp_link_train_set(encoder, DP_TRAINING_PATTERN_2);
+	if (ret)
+		goto stop;
 
-	dp_set_training_pattern(dev, dp, DP_TRAINING_PATTERN_2);
+	for (tries = 0; tries <= 5; tries++) {
+		udelay(400);
 
-	do {
-		if (dp_link_train_update(dev, dp, 400))
+		ret = auxch_rd(encoder, DP_LANE0_1_STATUS, status, 3);
+		if (ret)
 			break;
+		NV_DEBUG_KMS(dev, "\t\tstatus: 0x%02x 0x%02x\n",
+			 status[0], status[1]);
 
-		eq_done = !!(dp->stat[2] & DP_INTERLANE_ALIGN_DONE);
-		for (i = 0; i < dp->link_nr && eq_done; i++) {
-			u8 lane = (dp->stat[i >> 1] >> ((i & 1) * 4)) & 0xf;
-			if (!(lane & DP_LANE_CR_DONE))
+		eq_done = true;
+		if (!(status[2] & DP_INTERLANE_ALIGN_DONE))
+			eq_done = false;
+
+		for (i = 0; eq_done && i < nv_encoder->dp.link_nr; i++) {
+			int lane = (status[i >> 1] >> ((i & 1) * 4)) & 0xf;
+
+			if (!(lane & DP_LANE_CR_DONE)) {
 				cr_done = false;
+				break;
+			}
+
 			if (!(lane & DP_LANE_CHANNEL_EQ_DONE) ||
-			    !(lane & DP_LANE_SYMBOL_LOCKED))
+			    !(lane & DP_LANE_SYMBOL_LOCKED)) {
 				eq_done = false;
+				break;
+			}
 		}
 
-		if (dp_link_train_commit(dev, dp))
+		if (eq_done || !cr_done)
 			break;
-	} while (!eq_done && cr_done && ++tries <= 5);
-
-	return eq_done ? 0 : -1;
-}
 
-static void
-dp_set_downspread(struct drm_device *dev, struct dp_state *dp, bool enable)
-{
-	u16 script = 0x0000;
-	u8 *entry, *table = nouveau_dp_bios_data(dev, dp->dcb, &entry);
-	if (table) {
-		if (table[0] >= 0x20 && table[0] <= 0x30) {
-			if (enable) script = ROM16(entry[12]);
-			else        script = ROM16(entry[14]);
-		}
+		if (!nouveau_dp_link_train_adjust(encoder, config) ||
+		    !nouveau_dp_link_train_commit(encoder, config))
+			break;
 	}
 
-	nouveau_bios_run_init_table(dev, script, dp->dcb, dp->crtc);
-}
+stop:
+	/* end link training */
+	ret = nouveau_dp_link_train_set(encoder, DP_TRAINING_PATTERN_DISABLE);
+	if (ret)
+		return false;
 
-static void
-dp_link_train_init(struct drm_device *dev, struct dp_state *dp)
-{
-	u16 script = 0x0000;
-	u8 *entry, *table = nouveau_dp_bios_data(dev, dp->dcb, &entry);
-	if (table) {
-		if (table[0] >= 0x20 && table[0] <= 0x30)
-			script = ROM16(entry[6]);
+	/* retry at a lower setting, if possible */
+	if (!ret && !(eq_done && cr_done)) {
+		NV_DEBUG_KMS(dev, "\twe failed\n");
+		if (nv_encoder->dp.link_bw != DP_LINK_BW_1_62) {
+			NV_DEBUG_KMS(dev, "retry link training at low rate\n");
+			nv_encoder->dp.link_bw = DP_LINK_BW_1_62;
+			goto train;
+		}
 	}
 
-	nouveau_bios_run_init_table(dev, script, dp->dcb, dp->crtc);
-}
-
-static void
-dp_link_train_fini(struct drm_device *dev, struct dp_state *dp)
-{
-	u16 script = 0x0000;
-	u8 *entry, *table = nouveau_dp_bios_data(dev, dp->dcb, &entry);
-	if (table) {
-		if (table[0] >= 0x20 && table[0] <= 0x30)
-			script = ROM16(entry[8]);
+	if (dpe->script1) {
+		NV_DEBUG_KMS(dev, "SOR-%d: running DP script 1\n", nv_encoder->or);
+		nouveau_bios_run_init_table(dev, le16_to_cpu(dpe->script1),
+					    nv_encoder->dcb);
 	}
 
-	nouveau_bios_run_init_table(dev, script, dp->dcb, dp->crtc);
+	/* re-enable hotplug detect */
+	pgpio->irq_enable(dev, nv_connector->dcb->gpio_tag, hpd_state);
+
+	return eq_done;
 }
 
 bool
-nouveau_dp_link_train(struct drm_encoder *encoder, u32 datarate,
-		      struct dp_train_func *func)
+nouveau_dp_detect(struct drm_encoder *encoder)
 {
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc);
-	struct nouveau_connector *nv_connector =
-		nouveau_encoder_connector_get(nv_encoder);
 	struct drm_device *dev = encoder->dev;
-	struct nouveau_i2c_chan *auxch;
-	const u32 bw_list[] = { 270000, 162000, 0 };
-	const u32 *link_bw = bw_list;
-	struct dp_state dp;
+	uint8_t dpcd[4];
+	int ret;
 
-	auxch = nouveau_i2c_find(dev, nv_encoder->dcb->i2c_index);
-	if (!auxch)
+	ret = auxch_rd(encoder, 0x0000, dpcd, 4);
+	if (ret)
 		return false;
 
-	dp.func = func;
-	dp.dcb = nv_encoder->dcb;
-	dp.crtc = nv_crtc->index;
-	dp.auxch = auxch->drive;
-	dp.dpcd = nv_encoder->dp.dpcd;
-
-	/* adjust required bandwidth for 8B/10B coding overhead */
-	datarate = (datarate / 8) * 10;
-
-	/* some sinks toggle hotplug in response to some of the actions
-	 * we take during link training (DP_SET_POWER is one), we need
-	 * to ignore them for the moment to avoid races.
-	 */
-	nouveau_gpio_irq(dev, 0, nv_connector->hpd, 0xff, false);
-
-	/* enable down-spreading, if possible */
-	dp_set_downspread(dev, &dp, nv_encoder->dp.dpcd[3] & 1);
-
-	/* execute pre-train script from vbios */
-	dp_link_train_init(dev, &dp);
-
-	/* start off at highest link rate supported by encoder and display */
-	while (*link_bw > nv_encoder->dp.link_bw)
-		link_bw++;
-
-	while (link_bw[0]) {
-		/* find minimum required lane count at this link rate */
-		dp.link_nr = nv_encoder->dp.link_nr;
-		while ((dp.link_nr >> 1) * link_bw[0] > datarate)
-			dp.link_nr >>= 1;
-
-		/* drop link rate to minimum with this lane count */
-		while ((link_bw[1] * dp.link_nr) > datarate)
-			link_bw++;
-		dp.link_bw = link_bw[0];
+	NV_DEBUG_KMS(dev, "encoder: link_bw %d, link_nr %d\n"
+		      "display: link_bw %d, link_nr %d version 0x%02x\n",
+		 nv_encoder->dcb->dpconf.link_bw,
+		 nv_encoder->dcb->dpconf.link_nr,
+		 dpcd[1], dpcd[2] & 0x0f, dpcd[0]);
 
-		/* program selected link configuration */
-		dp_set_link_config(dev, &dp);
+	nv_encoder->dp.dpcd_version = dpcd[0];
 
-		/* attempt to train the link at this configuration */
-		memset(dp.stat, 0x00, sizeof(dp.stat));
-		if (!dp_link_train_cr(dev, &dp) &&
-		    !dp_link_train_eq(dev, &dp))
-			break;
-
-		/* retry at lower rate */
-		link_bw++;
-	}
+	nv_encoder->dp.link_bw = dpcd[1];
+	if (nv_encoder->dp.link_bw != DP_LINK_BW_1_62 &&
+	    !nv_encoder->dcb->dpconf.link_bw)
+		nv_encoder->dp.link_bw = DP_LINK_BW_1_62;
 
-	/* finish link training */
-	dp_set_training_pattern(dev, &dp, DP_TRAINING_PATTERN_DISABLE);
+	nv_encoder->dp.link_nr = dpcd[2] & DP_MAX_LANE_COUNT_MASK;
+	if (nv_encoder->dp.link_nr > nv_encoder->dcb->dpconf.link_nr)
+		nv_encoder->dp.link_nr = nv_encoder->dcb->dpconf.link_nr;
 
-	/* execute post-train script from vbios */
-	dp_link_train_fini(dev, &dp);
+	nv_encoder->dp.enhanced_frame = (dpcd[2] & DP_ENHANCED_FRAME_CAP);
 
-	/* re-enable hotplug detect */
-	nouveau_gpio_irq(dev, 0, nv_connector->hpd, 0xff, true);
 	return true;
 }
 
-void
-nouveau_dp_dpms(struct drm_encoder *encoder, int mode, u32 datarate,
-		struct dp_train_func *func)
+int
+nouveau_dp_auxch(struct nouveau_i2c_chan *auxch, int cmd, int addr,
+		 uint8_t *data, int data_nr)
 {
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct nouveau_i2c_chan *auxch;
-	u8 status;
-
-	auxch = nouveau_i2c_find(encoder->dev, nv_encoder->dcb->i2c_index);
-	if (!auxch)
-		return;
+	struct drm_device *dev = auxch->dev;
+	uint32_t tmp, ctrl, stat = 0, data32[4] = {};
+	int ret = 0, i, index = auxch->rd;
+
+	NV_DEBUG_KMS(dev, "ch %d cmd %d addr 0x%x len %d\n", index, cmd, addr, data_nr);
+
+	tmp = nv_rd32(dev, NV50_AUXCH_CTRL(auxch->rd));
+	nv_wr32(dev, NV50_AUXCH_CTRL(auxch->rd), tmp | 0x00100000);
+	tmp = nv_rd32(dev, NV50_AUXCH_CTRL(auxch->rd));
+	if (!(tmp & 0x01000000)) {
+		NV_ERROR(dev, "expected bit 24 == 1, got 0x%08x\n", tmp);
+		ret = -EIO;
+		goto out;
+	}
 
-	if (mode == DRM_MODE_DPMS_ON)
-		status = DP_SET_POWER_D0;
-	else
-		status = DP_SET_POWER_D3;
+	for (i = 0; i < 3; i++) {
+		tmp = nv_rd32(dev, NV50_AUXCH_STAT(auxch->rd));
+		if (tmp & NV50_AUXCH_STAT_STATE_READY)
+			break;
+		udelay(100);
+	}
 
-	nouveau_dp_auxch(auxch, 8, DP_SET_POWER, &status, 1);
+	if (i == 3) {
+		ret = -EBUSY;
+		goto out;
+	}
 
-	if (mode == DRM_MODE_DPMS_ON)
-		nouveau_dp_link_train(encoder, datarate, func);
-}
+	if (!(cmd & 1)) {
+		memcpy(data32, data, data_nr);
+		for (i = 0; i < 4; i++) {
+			NV_DEBUG_KMS(dev, "wr %d: 0x%08x\n", i, data32[i]);
+			nv_wr32(dev, NV50_AUXCH_DATA_OUT(index, i), data32[i]);
+		}
+	}
 
-bool
-nouveau_dp_detect(struct drm_encoder *encoder)
-{
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct drm_device *dev = encoder->dev;
-	struct nouveau_i2c_chan *auxch;
-	u8 *dpcd = nv_encoder->dp.dpcd;
-	int ret;
+	nv_wr32(dev, NV50_AUXCH_ADDR(index), addr);
+	ctrl  = nv_rd32(dev, NV50_AUXCH_CTRL(index));
+	ctrl &= ~(NV50_AUXCH_CTRL_CMD | NV50_AUXCH_CTRL_LEN);
+	ctrl |= (cmd << NV50_AUXCH_CTRL_CMD_SHIFT);
+	ctrl |= ((data_nr - 1) << NV50_AUXCH_CTRL_LEN_SHIFT);
+
+	for (i = 0; i < 16; i++) {
+		nv_wr32(dev, NV50_AUXCH_CTRL(index), ctrl | 0x80000000);
+		nv_wr32(dev, NV50_AUXCH_CTRL(index), ctrl);
+		nv_wr32(dev, NV50_AUXCH_CTRL(index), ctrl | 0x00010000);
+		if (!nv_wait(dev, NV50_AUXCH_CTRL(index),
+			     0x00010000, 0x00000000)) {
+			NV_ERROR(dev, "expected bit 16 == 0, got 0x%08x\n",
+				 nv_rd32(dev, NV50_AUXCH_CTRL(index)));
+			ret = -EBUSY;
+			goto out;
+		}
 
-	auxch = nouveau_i2c_find(dev, nv_encoder->dcb->i2c_index);
-	if (!auxch)
-		return false;
+		udelay(400);
 
-	ret = auxch_tx(dev, auxch->drive, 9, DP_DPCD_REV, dpcd, 8);
-	if (ret)
-		return false;
+		stat = nv_rd32(dev, NV50_AUXCH_STAT(index));
+		if ((stat & NV50_AUXCH_STAT_REPLY_AUX) !=
+			    NV50_AUXCH_STAT_REPLY_AUX_DEFER)
+			break;
+	}
 
-	nv_encoder->dp.link_bw = 27000 * dpcd[1];
-	nv_encoder->dp.link_nr = dpcd[2] & DP_MAX_LANE_COUNT_MASK;
+	if (i == 16) {
+		NV_ERROR(dev, "auxch DEFER too many times, bailing\n");
+		ret = -EREMOTEIO;
+		goto out;
+	}
 
-	NV_DEBUG_KMS(dev, "display: %dx%d dpcd 0x%02x\n",
-		     nv_encoder->dp.link_nr, nv_encoder->dp.link_bw, dpcd[0]);
-	NV_DEBUG_KMS(dev, "encoder: %dx%d\n",
-		     nv_encoder->dcb->dpconf.link_nr,
-		     nv_encoder->dcb->dpconf.link_bw);
+	if (cmd & 1) {
+		if ((stat & NV50_AUXCH_STAT_COUNT) != data_nr) {
+			ret = -EREMOTEIO;
+			goto out;
+		}
 
-	if (nv_encoder->dcb->dpconf.link_nr < nv_encoder->dp.link_nr)
-		nv_encoder->dp.link_nr = nv_encoder->dcb->dpconf.link_nr;
-	if (nv_encoder->dcb->dpconf.link_bw < nv_encoder->dp.link_bw)
-		nv_encoder->dp.link_bw = nv_encoder->dcb->dpconf.link_bw;
+		for (i = 0; i < 4; i++) {
+			data32[i] = nv_rd32(dev, NV50_AUXCH_DATA_IN(index, i));
+			NV_DEBUG_KMS(dev, "rd %d: 0x%08x\n", i, data32[i]);
+		}
+		memcpy(data, data32, data_nr);
+	}
 
-	NV_DEBUG_KMS(dev, "maximum: %dx%d\n",
-		     nv_encoder->dp.link_nr, nv_encoder->dp.link_bw);
+out:
+	tmp = nv_rd32(dev, NV50_AUXCH_CTRL(auxch->rd));
+	nv_wr32(dev, NV50_AUXCH_CTRL(auxch->rd), tmp & ~0x00100000);
+	tmp = nv_rd32(dev, NV50_AUXCH_CTRL(auxch->rd));
+	if (tmp & 0x01000000) {
+		NV_ERROR(dev, "expected bit 24 == 0, got 0x%08x\n", tmp);
+		ret = -EIO;
+	}
 
-	return true;
-}
+	udelay(400);
 
-int
-nouveau_dp_auxch(struct nouveau_i2c_chan *auxch, int cmd, int addr,
-		 uint8_t *data, int data_nr)
-{
-	return auxch_tx(auxch->dev, auxch->drive, cmd, addr, data, data_nr);
+	return ret ? ret : (stat & NV50_AUXCH_STAT_REPLY);
 }
 
 static int
 nouveau_dp_i2c_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, int num)
 {
 	struct nouveau_i2c_chan *auxch = (struct nouveau_i2c_chan *)adap;
+	struct drm_device *dev = auxch->dev;
 	struct i2c_msg *msg = msgs;
 	int ret, mcnt = num;
 
@@ -570,6 +602,19 @@ nouveau_dp_i2c_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, int num)
 			if (ret < 0)
 				return ret;
 
+			switch (ret & NV50_AUXCH_STAT_REPLY_I2C) {
+			case NV50_AUXCH_STAT_REPLY_I2C_ACK:
+				break;
+			case NV50_AUXCH_STAT_REPLY_I2C_NACK:
+				return -EREMOTEIO;
+			case NV50_AUXCH_STAT_REPLY_I2C_DEFER:
+				udelay(100);
+				continue;
+			default:
+				NV_ERROR(dev, "bad auxch reply: 0x%08x\n", ret);
+				return -EREMOTEIO;
+			}
+
 			ptr += cnt;
 			remaining -= cnt;
 		}
diff --git a/drivers/gpu/drm/nouveau/nouveau_drv.c b/drivers/gpu/drm/nouveau/nouveau_drv.c
index 4f2030b..02c6f37 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drv.c
+++ b/drivers/gpu/drm/nouveau/nouveau_drv.c
@@ -23,7 +23,6 @@
  */
 
 #include <linux/console.h>
-#include <linux/module.h>
 
 #include "drmP.h"
 #include "drm.h"
@@ -42,7 +41,7 @@ int nouveau_agpmode = -1;
 module_param_named(agpmode, nouveau_agpmode, int, 0400);
 
 MODULE_PARM_DESC(modeset, "Enable kernel modesetting");
-int nouveau_modeset = -1;
+static int nouveau_modeset = -1; /* kms */
 module_param_named(modeset, nouveau_modeset, int, 0400);
 
 MODULE_PARM_DESC(vbios, "Override default VBIOS location");
@@ -57,10 +56,6 @@ MODULE_PARM_DESC(vram_notify, "Force DMA notifiers to be in VRAM");
 int nouveau_vram_notify = 0;
 module_param_named(vram_notify, nouveau_vram_notify, int, 0400);
 
-MODULE_PARM_DESC(vram_type, "Override detected VRAM type");
-char *nouveau_vram_type;
-module_param_named(vram_type, nouveau_vram_type, charp, 0400);
-
 MODULE_PARM_DESC(duallink, "Allow dual-link TMDS (>=GeForce 8)");
 int nouveau_duallink = 1;
 module_param_named(duallink, nouveau_duallink, int, 0400);
@@ -78,7 +73,7 @@ int nouveau_ignorelid = 0;
 module_param_named(ignorelid, nouveau_ignorelid, int, 0400);
 
 MODULE_PARM_DESC(noaccel, "Disable all acceleration");
-int nouveau_noaccel = -1;
+int nouveau_noaccel = 0;
 module_param_named(noaccel, nouveau_noaccel, int, 0400);
 
 MODULE_PARM_DESC(nofbaccel, "Disable fbcon acceleration");
@@ -93,7 +88,7 @@ MODULE_PARM_DESC(override_conntype, "Ignore DCB connector type");
 int nouveau_override_conntype = 0;
 module_param_named(override_conntype, nouveau_override_conntype, int, 0400);
 
-MODULE_PARM_DESC(tv_disable, "Disable TV-out detection");
+MODULE_PARM_DESC(tv_disable, "Disable TV-out detection\n");
 int nouveau_tv_disable = 0;
 module_param_named(tv_disable, nouveau_tv_disable, int, 0400);
 
@@ -108,30 +103,22 @@ module_param_named(tv_norm, nouveau_tv_norm, charp, 0400);
 MODULE_PARM_DESC(reg_debug, "Register access debug bitmask:\n"
 		"\t\t0x1 mc, 0x2 video, 0x4 fb, 0x8 extdev,\n"
 		"\t\t0x10 crtc, 0x20 ramdac, 0x40 vgacrtc, 0x80 rmvio,\n"
-		"\t\t0x100 vgaattr, 0x200 EVO (G80+)");
+		"\t\t0x100 vgaattr, 0x200 EVO (G80+). ");
 int nouveau_reg_debug;
 module_param_named(reg_debug, nouveau_reg_debug, int, 0600);
 
-MODULE_PARM_DESC(perflvl, "Performance level (default: boot)");
+MODULE_PARM_DESC(perflvl, "Performance level (default: boot)\n");
 char *nouveau_perflvl;
 module_param_named(perflvl, nouveau_perflvl, charp, 0400);
 
-MODULE_PARM_DESC(perflvl_wr, "Allow perflvl changes (warning: dangerous!)");
+MODULE_PARM_DESC(perflvl_wr, "Allow perflvl changes (warning: dangerous!)\n");
 int nouveau_perflvl_wr;
 module_param_named(perflvl_wr, nouveau_perflvl_wr, int, 0400);
 
-MODULE_PARM_DESC(msi, "Enable MSI (default: off)");
+MODULE_PARM_DESC(msi, "Enable MSI (default: off)\n");
 int nouveau_msi;
 module_param_named(msi, nouveau_msi, int, 0400);
 
-MODULE_PARM_DESC(ctxfw, "Use external HUB/GPC ucode (fermi)");
-int nouveau_ctxfw;
-module_param_named(ctxfw, nouveau_ctxfw, int, 0400);
-
-MODULE_PARM_DESC(mxmdcb, "Santise DCB table according to MXM-SIS");
-int nouveau_mxmdcb = 1;
-module_param_named(mxmdcb, nouveau_mxmdcb, int, 0400);
-
 int nouveau_fbpercrtc;
 #if 0
 module_param_named(fbpercrtc, nouveau_fbpercrtc, int, 0400);
@@ -186,11 +173,8 @@ nouveau_pci_suspend(struct pci_dev *pdev, pm_message_t pm_state)
 	if (dev->switch_power_state == DRM_SWITCH_POWER_OFF)
 		return 0;
 
-	NV_INFO(dev, "Disabling display...\n");
-	nouveau_display_fini(dev);
-
-	NV_INFO(dev, "Disabling fbcon...\n");
-	nouveau_fbcon_set_suspend(dev, 1);
+	NV_INFO(dev, "Disabling fbcon acceleration...\n");
+	nouveau_fbcon_save_disable_accel(dev);
 
 	NV_INFO(dev, "Unpinning framebuffer(s)...\n");
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
@@ -226,13 +210,10 @@ nouveau_pci_suspend(struct pci_dev *pdev, pm_message_t pm_state)
 	pfifo->unload_context(dev);
 
 	for (e = NVOBJ_ENGINE_NR - 1; e >= 0; e--) {
-		if (!dev_priv->eng[e])
-			continue;
-
-		ret = dev_priv->eng[e]->fini(dev, e, true);
-		if (ret) {
-			NV_ERROR(dev, "... engine %d failed: %d\n", e, ret);
-			goto out_abort;
+		if (dev_priv->eng[e]) {
+			ret = dev_priv->eng[e]->fini(dev, e);
+			if (ret)
+				goto out_abort;
 		}
 	}
 
@@ -257,6 +238,10 @@ nouveau_pci_suspend(struct pci_dev *pdev, pm_message_t pm_state)
 		pci_set_power_state(pdev, PCI_D3hot);
 	}
 
+	console_lock();
+	nouveau_fbcon_set_suspend(dev, 1);
+	console_unlock();
+	nouveau_fbcon_restore_accel(dev);
 	return 0;
 
 out_abort:
@@ -282,6 +267,8 @@ nouveau_pci_resume(struct pci_dev *pdev)
 	if (dev->switch_power_state == DRM_SWITCH_POWER_OFF)
 		return 0;
 
+	nouveau_fbcon_save_disable_accel(dev);
+
 	NV_INFO(dev, "We're back, enabling device...\n");
 	pci_set_power_state(pdev, PCI_D0);
 	pci_restore_state(pdev);
@@ -301,6 +288,8 @@ nouveau_pci_resume(struct pci_dev *pdev)
 	if (ret)
 		return ret;
 
+	nouveau_pm_resume(dev);
+
 	if (dev_priv->gart_info.type == NOUVEAU_GART_AGP) {
 		ret = nouveau_mem_init_agp(dev);
 		if (ret) {
@@ -340,8 +329,6 @@ nouveau_pci_resume(struct pci_dev *pdev)
 		}
 	}
 
-	nouveau_pm_resume(dev);
-
 	NV_INFO(dev, "Restoring mode...\n");
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
 		struct nouveau_framebuffer *nouveau_fb;
@@ -363,10 +350,16 @@ nouveau_pci_resume(struct pci_dev *pdev)
 			NV_ERROR(dev, "Could not pin/map cursor.\n");
 	}
 
-	nouveau_fbcon_set_suspend(dev, 0);
-	nouveau_fbcon_zfill_all(dev);
+	engine->display.init(dev);
 
-	nouveau_display_init(dev);
+	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
+		struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
+		u32 offset = nv_crtc->cursor.nvbo->bo.mem.start << PAGE_SHIFT;
+
+		nv_crtc->cursor.set_offset(nv_crtc, offset);
+		nv_crtc->cursor.set_pos(nv_crtc, nv_crtc->cursor_saved_x,
+						 nv_crtc->cursor_saved_y);
+	}
 
 	/* Force CLUT to get re-loaded during modeset */
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
@@ -375,35 +368,18 @@ nouveau_pci_resume(struct pci_dev *pdev)
 		nv_crtc->lut.depth = 0;
 	}
 
-	drm_helper_resume_force_mode(dev);
+	console_lock();
+	nouveau_fbcon_set_suspend(dev, 0);
+	console_unlock();
 
-	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) {
-		struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
-		u32 offset = nv_crtc->cursor.nvbo->bo.offset;
+	nouveau_fbcon_zfill_all(dev);
 
-		nv_crtc->cursor.set_offset(nv_crtc, offset);
-		nv_crtc->cursor.set_pos(nv_crtc, nv_crtc->cursor_saved_x,
-						 nv_crtc->cursor_saved_y);
-	}
+	drm_helper_resume_force_mode(dev);
 
+	nouveau_fbcon_restore_accel(dev);
 	return 0;
 }
 
-static const struct file_operations nouveau_driver_fops = {
-	.owner = THIS_MODULE,
-	.open = drm_open,
-	.release = drm_release,
-	.unlocked_ioctl = drm_ioctl,
-	.mmap = nouveau_ttm_mmap,
-	.poll = drm_poll,
-	.fasync = drm_fasync,
-	.read = drm_read,
-#if defined(CONFIG_COMPAT)
-	.compat_ioctl = nouveau_compat_ioctl,
-#endif
-	.llseek = noop_llseek,
-};
-
 static struct drm_driver driver = {
 	.driver_features =
 		DRIVER_USE_AGP | DRIVER_PCI_DMA | DRIVER_SG |
@@ -413,9 +389,7 @@ static struct drm_driver driver = {
 	.firstopen = nouveau_firstopen,
 	.lastclose = nouveau_lastclose,
 	.unload = nouveau_unload,
-	.open = nouveau_open,
 	.preclose = nouveau_preclose,
-	.postclose = nouveau_postclose,
 #if defined(CONFIG_DRM_NOUVEAU_DEBUG)
 	.debugfs_init = nouveau_debugfs_init,
 	.debugfs_cleanup = nouveau_debugfs_takedown,
@@ -429,15 +403,23 @@ static struct drm_driver driver = {
 	.disable_vblank = nouveau_vblank_disable,
 	.reclaim_buffers = drm_core_reclaim_buffers,
 	.ioctls = nouveau_ioctls,
-	.fops = &nouveau_driver_fops,
+	.fops = {
+		.owner = THIS_MODULE,
+		.open = drm_open,
+		.release = drm_release,
+		.unlocked_ioctl = drm_ioctl,
+		.mmap = nouveau_ttm_mmap,
+		.poll = drm_poll,
+		.fasync = drm_fasync,
+		.read = drm_read,
+#if defined(CONFIG_COMPAT)
+		.compat_ioctl = nouveau_compat_ioctl,
+#endif
+		.llseek = noop_llseek,
+	},
+
 	.gem_init_object = nouveau_gem_object_new,
 	.gem_free_object = nouveau_gem_object_del,
-	.gem_open_object = nouveau_gem_object_open,
-	.gem_close_object = nouveau_gem_object_close,
-
-	.dumb_create = nouveau_display_dumb_create,
-	.dumb_map_offset = nouveau_display_dumb_map_offset,
-	.dumb_destroy = nouveau_display_dumb_destroy,
 
 	.name = DRIVER_NAME,
 	.desc = DRIVER_DESC,
diff --git a/drivers/gpu/drm/nouveau/nouveau_drv.h b/drivers/gpu/drm/nouveau/nouveau_drv.h
index a184ba3..9c56331 100644
--- a/drivers/gpu/drm/nouveau/nouveau_drv.h
+++ b/drivers/gpu/drm/nouveau/nouveau_drv.h
@@ -46,17 +46,9 @@
 #include "ttm/ttm_module.h"
 
 struct nouveau_fpriv {
-	spinlock_t lock;
-	struct list_head channels;
-	struct nouveau_vm *vm;
+	struct ttm_object_file *tfile;
 };
 
-static inline struct nouveau_fpriv *
-nouveau_fpriv(struct drm_file *file_priv)
-{
-	return file_priv ? file_priv->driver_priv : NULL;
-}
-
 #define DRM_FILE_PAGE_OFFSET (0x100000000ULL >> PAGE_SHIFT)
 
 #include "nouveau_drm.h"
@@ -77,7 +69,7 @@ struct nouveau_mem {
 	struct drm_device *dev;
 
 	struct nouveau_vma bar_vma;
-	struct nouveau_vma vma[2];
+	struct nouveau_vma tmp_vma;
 	u8  page_shift;
 
 	struct drm_mm_node *tag;
@@ -115,8 +107,7 @@ struct nouveau_bo {
 
 	struct nouveau_channel *channel;
 
-	struct list_head vma_list;
-	unsigned page_shift;
+	struct nouveau_vma vma;
 
 	uint32_t tile_mode;
 	uint32_t tile_flags;
@@ -163,9 +154,6 @@ enum nouveau_flags {
 #define NVOBJ_ENGINE_COPY0	3
 #define NVOBJ_ENGINE_COPY1	4
 #define NVOBJ_ENGINE_MPEG	5
-#define NVOBJ_ENGINE_PPP	NVOBJ_ENGINE_MPEG
-#define NVOBJ_ENGINE_BSP	6
-#define NVOBJ_ENGINE_VP		7
 #define NVOBJ_ENGINE_DISPLAY	15
 #define NVOBJ_ENGINE_NR		16
 
@@ -188,10 +176,9 @@ struct nouveau_gpuobj {
 	uint32_t flags;
 
 	u32 size;
-	u32 pinst;	/* PRAMIN BAR offset */
-	u32 cinst;	/* Channel offset */
-	u64 vinst;	/* VRAM address */
-	u64 linst;	/* VM address */
+	u32 pinst;
+	u32 cinst;
+	u64 vinst;
 
 	uint32_t engine;
 	uint32_t class;
@@ -214,7 +201,6 @@ enum nouveau_channel_mutex_class {
 
 struct nouveau_channel {
 	struct drm_device *dev;
-	struct list_head list;
 	int id;
 
 	/* references to the channel data structure */
@@ -232,7 +218,6 @@ struct nouveau_channel {
 	/* mapping of the regs controlling the fifo */
 	void __iomem *user;
 	uint32_t user_get;
-	uint32_t user_get_hi;
 	uint32_t user_put;
 
 	/* Fencing */
@@ -243,18 +228,15 @@ struct nouveau_channel {
 		uint32_t sequence;
 		uint32_t sequence_ack;
 		atomic_t last_sequence_irq;
-		struct nouveau_vma vma;
 	} fence;
 
 	/* DMA push buffer */
 	struct nouveau_gpuobj *pushbuf;
 	struct nouveau_bo     *pushbuf_bo;
-	struct nouveau_vma     pushbuf_vma;
-	uint64_t               pushbuf_base;
+	uint32_t               pushbuf_base;
 
 	/* Notifier memory */
 	struct nouveau_bo *notifier_bo;
-	struct nouveau_vma notifier_vma;
 	struct drm_mm notifier_heap;
 
 	/* PFIFO context */
@@ -296,7 +278,6 @@ struct nouveau_channel {
 
 	uint32_t sw_subchannel[8];
 
-	struct nouveau_vma dispc_vma[2];
 	struct {
 		struct nouveau_gpuobj *vblsem;
 		uint32_t vblsem_head;
@@ -316,7 +297,7 @@ struct nouveau_channel {
 struct nouveau_exec_engine {
 	void (*destroy)(struct drm_device *, int engine);
 	int  (*init)(struct drm_device *, int engine);
-	int  (*fini)(struct drm_device *, int engine, bool suspend);
+	int  (*fini)(struct drm_device *, int engine);
 	int  (*context_new)(struct nouveau_channel *, int engine);
 	void (*context_del)(struct nouveau_channel *, int engine);
 	int  (*object_new)(struct nouveau_channel *, int engine,
@@ -333,8 +314,7 @@ struct nouveau_instmem_engine {
 	int	(*suspend)(struct drm_device *dev);
 	void	(*resume)(struct drm_device *dev);
 
-	int	(*get)(struct nouveau_gpuobj *, struct nouveau_channel *,
-		       u32 size, u32 align);
+	int	(*get)(struct nouveau_gpuobj *, u32 size, u32 align);
 	void	(*put)(struct nouveau_gpuobj *);
 	int	(*map)(struct nouveau_gpuobj *);
 	void	(*unmap)(struct nouveau_gpuobj *);
@@ -397,145 +377,77 @@ struct nouveau_display_engine {
 	int (*early_init)(struct drm_device *);
 	void (*late_takedown)(struct drm_device *);
 	int (*create)(struct drm_device *);
-	void (*destroy)(struct drm_device *);
 	int (*init)(struct drm_device *);
-	void (*fini)(struct drm_device *);
-
-	struct drm_property *dithering_mode;
-	struct drm_property *dithering_depth;
-	struct drm_property *underscan_property;
-	struct drm_property *underscan_hborder_property;
-	struct drm_property *underscan_vborder_property;
-	/* not really hue and saturation: */
-	struct drm_property *vibrant_hue_property;
-	struct drm_property *color_vibrance_property;
+	void (*destroy)(struct drm_device *);
 };
 
 struct nouveau_gpio_engine {
-	spinlock_t lock;
-	struct list_head isr;
-	int (*init)(struct drm_device *);
-	void (*fini)(struct drm_device *);
-	int (*drive)(struct drm_device *, int line, int dir, int out);
-	int (*sense)(struct drm_device *, int line);
-	void (*irq_enable)(struct drm_device *, int line, bool);
+	void *priv;
+
+	int  (*init)(struct drm_device *);
+	void (*takedown)(struct drm_device *);
+
+	int  (*get)(struct drm_device *, enum dcb_gpio_tag);
+	int  (*set)(struct drm_device *, enum dcb_gpio_tag, int state);
+
+	int  (*irq_register)(struct drm_device *, enum dcb_gpio_tag,
+			     void (*)(void *, int), void *);
+	void (*irq_unregister)(struct drm_device *, enum dcb_gpio_tag,
+			       void (*)(void *, int), void *);
+	bool (*irq_enable)(struct drm_device *, enum dcb_gpio_tag, bool on);
 };
 
 struct nouveau_pm_voltage_level {
-	u32 voltage; /* microvolts */
-	u8  vid;
+	u8 voltage;
+	u8 vid;
 };
 
 struct nouveau_pm_voltage {
 	bool supported;
-	u8 version;
 	u8 vid_mask;
 
 	struct nouveau_pm_voltage_level *level;
 	int nr_level;
 };
 
-/* Exclusive upper limits */
-#define NV_MEM_CL_DDR2_MAX 8
-#define NV_MEM_WR_DDR2_MAX 9
-#define NV_MEM_CL_DDR3_MAX 17
-#define NV_MEM_WR_DDR3_MAX 17
-#define NV_MEM_CL_GDDR3_MAX 16
-#define NV_MEM_WR_GDDR3_MAX 18
-#define NV_MEM_CL_GDDR5_MAX 21
-#define NV_MEM_WR_GDDR5_MAX 20
-
 struct nouveau_pm_memtiming {
 	int id;
-
-	u32 reg[9];
-	u32 mr[4];
-
-	u8 tCWL;
-
-	u8 odt;
-	u8 drive_strength;
-};
-
-struct nouveau_pm_tbl_header {
-	u8 version;
-	u8 header_len;
-	u8 entry_cnt;
-	u8 entry_len;
-};
-
-struct nouveau_pm_tbl_entry {
-	u8 tWR;
-	u8 tWTR;
-	u8 tCL;
-	u8 tRC;
-	u8 empty_4;
-	u8 tRFC;	/* Byte 5 */
-	u8 empty_6;
-	u8 tRAS;	/* Byte 7 */
-	u8 empty_8;
-	u8 tRP;		/* Byte 9 */
-	u8 tRCDRD;
-	u8 tRCDWR;
-	u8 tRRD;
-	u8 tUNK_13;
-	u8 RAM_FT1;		/* 14, a bitmask of random RAM features */
-	u8 empty_15;
-	u8 tUNK_16;
-	u8 empty_17;
-	u8 tUNK_18;
-	u8 tCWL;
-	u8 tUNK_20, tUNK_21;
-};
-
-struct nouveau_pm_profile;
-struct nouveau_pm_profile_func {
-	void (*destroy)(struct nouveau_pm_profile *);
-	void (*init)(struct nouveau_pm_profile *);
-	void (*fini)(struct nouveau_pm_profile *);
-	struct nouveau_pm_level *(*select)(struct nouveau_pm_profile *);
-};
-
-struct nouveau_pm_profile {
-	const struct nouveau_pm_profile_func *func;
-	struct list_head head;
-	char name[8];
+	u32 reg_100220;
+	u32 reg_100224;
+	u32 reg_100228;
+	u32 reg_10022c;
+	u32 reg_100230;
+	u32 reg_100234;
+	u32 reg_100238;
+	u32 reg_10023c;
+	u32 reg_100240;
 };
 
 #define NOUVEAU_PM_MAX_LEVEL 8
 struct nouveau_pm_level {
-	struct nouveau_pm_profile profile;
 	struct device_attribute dev_attr;
 	char name[32];
 	int id;
 
-	struct nouveau_pm_memtiming timing;
-	u32 memory;
-	u16 memscript;
-
 	u32 core;
+	u32 memory;
 	u32 shader;
-	u32 rop;
-	u32 copy;
-	u32 daemon;
-	u32 vdec;
-	u32 dom6;
-	u32 unka0;	/* nva3:nvc0 */
-	u32 hub01;	/* nvc0- */
-	u32 hub06;	/* nvc0- */
-	u32 hub07;	/* nvc0- */
-
-	u32 volt_min; /* microvolts */
-	u32 volt_max;
-	u8  fanspeed;
+	u32 unk05;
+	u32 unk0a;
+
+	u8 voltage;
+	u8 fanspeed;
+
+	u16 memscript;
+	struct nouveau_pm_memtiming *timing;
 };
 
 struct nouveau_pm_temp_sensor_constants {
 	u16 offset_constant;
 	s16 offset_mult;
-	s16 offset_div;
-	s16 slope_mult;
-	s16 slope_div;
+	u16 offset_div;
+	u16 slope_mult;
+	u16 slope_div;
 };
 
 struct nouveau_pm_threshold_temp {
@@ -544,26 +456,19 @@ struct nouveau_pm_threshold_temp {
 	s16 fan_boost;
 };
 
-struct nouveau_pm_fan {
-	u32 percent;
-	u32 min_duty;
-	u32 max_duty;
-	u32 pwm_freq;
-	u32 pwm_divisor;
+struct nouveau_pm_memtimings {
+	bool supported;
+	struct nouveau_pm_memtiming *timing;
+	int nr_timing;
 };
 
 struct nouveau_pm_engine {
 	struct nouveau_pm_voltage voltage;
 	struct nouveau_pm_level perflvl[NOUVEAU_PM_MAX_LEVEL];
 	int nr_perflvl;
+	struct nouveau_pm_memtimings memtimings;
 	struct nouveau_pm_temp_sensor_constants sensor_constants;
 	struct nouveau_pm_threshold_temp threshold_temp;
-	struct nouveau_pm_fan fan;
-
-	struct nouveau_pm_profile *profile_ac;
-	struct nouveau_pm_profile *profile_dc;
-	struct nouveau_pm_profile *profile;
-	struct list_head profiles;
 
 	struct nouveau_pm_level boot;
 	struct nouveau_pm_level *cur;
@@ -571,22 +476,19 @@ struct nouveau_pm_engine {
 	struct device *hwmon;
 	struct notifier_block acpi_nb;
 
-	int  (*clocks_get)(struct drm_device *, struct nouveau_pm_level *);
-	void *(*clocks_pre)(struct drm_device *, struct nouveau_pm_level *);
-	int (*clocks_set)(struct drm_device *, void *);
-
+	int (*clock_get)(struct drm_device *, u32 id);
+	void *(*clock_pre)(struct drm_device *, struct nouveau_pm_level *,
+			   u32 id, int khz);
+	void (*clock_set)(struct drm_device *, void *);
 	int (*voltage_get)(struct drm_device *);
 	int (*voltage_set)(struct drm_device *, int voltage);
-	int (*pwm_get)(struct drm_device *, int line, u32*, u32*);
-	int (*pwm_set)(struct drm_device *, int line, u32, u32);
+	int (*fanspeed_get)(struct drm_device *);
+	int (*fanspeed_set)(struct drm_device *, int fanspeed);
 	int (*temp_get)(struct drm_device *);
 };
 
 struct nouveau_vram_engine {
-	struct nouveau_mm mm;
-
 	int  (*init)(struct drm_device *);
-	void (*takedown)(struct drm_device *dev);
 	int  (*get)(struct drm_device *, u64, u32 align, u32 size_nc,
 		    u32 type, struct nouveau_mem **);
 	void (*put)(struct drm_device *, struct nouveau_mem **);
@@ -695,26 +597,24 @@ struct nv04_mode_state {
 };
 
 enum nouveau_card_type {
-	NV_04      = 0x04,
+	NV_04      = 0x00,
 	NV_10      = 0x10,
 	NV_20      = 0x20,
 	NV_30      = 0x30,
 	NV_40      = 0x40,
 	NV_50      = 0x50,
 	NV_C0      = 0xc0,
-	NV_D0      = 0xd0,
 };
 
 struct drm_nouveau_private {
 	struct drm_device *dev;
-	bool noaccel;
 
 	/* the card type, takes NV_* as values */
 	enum nouveau_card_type card_type;
 	/* exact chipset, derived from NV_PMC_BOOT_0 */
 	int chipset;
+	int stepping;
 	int flags;
-	u32 crystal;
 
 	void __iomem *mmio;
 
@@ -798,23 +698,11 @@ struct drm_nouveau_private {
 	} tile;
 
 	/* VRAM/fb configuration */
-	enum {
-		NV_MEM_TYPE_UNKNOWN = 0,
-		NV_MEM_TYPE_STOLEN,
-		NV_MEM_TYPE_SGRAM,
-		NV_MEM_TYPE_SDRAM,
-		NV_MEM_TYPE_DDR1,
-		NV_MEM_TYPE_DDR2,
-		NV_MEM_TYPE_DDR3,
-		NV_MEM_TYPE_GDDR2,
-		NV_MEM_TYPE_GDDR3,
-		NV_MEM_TYPE_GDDR4,
-		NV_MEM_TYPE_GDDR5
-	} vram_type;
 	uint64_t vram_size;
 	uint64_t vram_sys_base;
-	bool vram_rank_B;
+	u32 vram_rblock_size;
 
+	uint64_t fb_phys;
 	uint64_t fb_available_size;
 	uint64_t fb_mappable_pages;
 	uint64_t fb_aper_free;
@@ -828,8 +716,6 @@ struct drm_nouveau_private {
 	struct nouveau_vm *chan_vm;
 
 	struct nvbios vbios;
-	u8 *mxms;
-	struct list_head i2c_ports;
 
 	struct nv04_mode_state mode_reg;
 	struct nv04_mode_state saved_reg;
@@ -879,14 +765,12 @@ nouveau_bo_ref(struct nouveau_bo *ref, struct nouveau_bo **pnvbo)
 }
 
 /* nouveau_drv.c */
-extern int nouveau_modeset;
 extern int nouveau_agpmode;
 extern int nouveau_duallink;
 extern int nouveau_uscript_lvds;
 extern int nouveau_uscript_tmds;
 extern int nouveau_vram_pushbuf;
 extern int nouveau_vram_notify;
-extern char *nouveau_vram_type;
 extern int nouveau_fbpercrtc;
 extern int nouveau_tv_disable;
 extern char *nouveau_tv_norm;
@@ -900,16 +784,12 @@ extern int nouveau_override_conntype;
 extern char *nouveau_perflvl;
 extern int nouveau_perflvl_wr;
 extern int nouveau_msi;
-extern int nouveau_ctxfw;
-extern int nouveau_mxmdcb;
 
 extern int nouveau_pci_suspend(struct pci_dev *pdev, pm_message_t pm_state);
 extern int nouveau_pci_resume(struct pci_dev *pdev);
 
 /* nouveau_state.c */
-extern int  nouveau_open(struct drm_device *, struct drm_file *);
 extern void nouveau_preclose(struct drm_device *dev, struct drm_file *);
-extern void nouveau_postclose(struct drm_device *, struct drm_file *);
 extern int  nouveau_load(struct drm_device *, unsigned long flags);
 extern int  nouveau_firstopen(struct drm_device *);
 extern void nouveau_lastclose(struct drm_device *);
@@ -922,8 +802,6 @@ extern bool nouveau_wait_eq(struct drm_device *, uint64_t timeout,
 			    uint32_t reg, uint32_t mask, uint32_t val);
 extern bool nouveau_wait_ne(struct drm_device *, uint64_t timeout,
 			    uint32_t reg, uint32_t mask, uint32_t val);
-extern bool nouveau_wait_cb(struct drm_device *, u64 timeout,
-			    bool (*cond)(void *), void *);
 extern bool nouveau_wait_for_idle(struct drm_device *);
 extern int  nouveau_card_init(struct drm_device *);
 
@@ -935,12 +813,8 @@ extern void nouveau_mem_gart_fini(struct drm_device *);
 extern int  nouveau_mem_init_agp(struct drm_device *);
 extern int  nouveau_mem_reset_agp(struct drm_device *);
 extern void nouveau_mem_close(struct drm_device *);
+extern int  nouveau_mem_detect(struct drm_device *);
 extern bool nouveau_mem_flags_valid(struct drm_device *, u32 tile_flags);
-extern int  nouveau_mem_timing_calc(struct drm_device *, u32 freq,
-				    struct nouveau_pm_memtiming *);
-extern void nouveau_mem_timing_read(struct drm_device *,
-				    struct nouveau_pm_memtiming *);
-extern int nouveau_mem_vbios_type(struct drm_device *);
 extern struct nouveau_tile_reg *nv10_mem_set_tiling(
 	struct drm_device *dev, uint32_t addr, uint32_t size,
 	uint32_t pitch, uint32_t flags);
@@ -973,7 +847,7 @@ extern int  nouveau_channel_alloc(struct drm_device *dev,
 extern struct nouveau_channel *
 nouveau_channel_get_unlocked(struct nouveau_channel *);
 extern struct nouveau_channel *
-nouveau_channel_get(struct drm_file *, int id);
+nouveau_channel_get(struct drm_device *, struct drm_file *, int id);
 extern void nouveau_channel_put_unlocked(struct nouveau_channel **);
 extern void nouveau_channel_put(struct nouveau_channel **);
 extern void nouveau_channel_ref(struct nouveau_channel *chan,
@@ -1056,10 +930,7 @@ extern int nouveau_sgdma_init(struct drm_device *);
 extern void nouveau_sgdma_takedown(struct drm_device *);
 extern uint32_t nouveau_sgdma_get_physical(struct drm_device *,
 					   uint32_t offset);
-extern struct ttm_tt *nouveau_sgdma_create_ttm(struct ttm_bo_device *bdev,
-					       unsigned long size,
-					       uint32_t page_flags,
-					       struct page *dummy_read_page);
+extern struct ttm_backend *nouveau_sgdma_init_ttm(struct drm_device *);
 
 /* nouveau_debugfs.c */
 #if defined(CONFIG_DRM_NOUVEAU_DEBUG)
@@ -1100,14 +971,12 @@ extern int  nouveau_dma_wait(struct nouveau_channel *, int slots, int size);
 #if defined(CONFIG_ACPI)
 void nouveau_register_dsm_handler(void);
 void nouveau_unregister_dsm_handler(void);
-void nouveau_switcheroo_optimus_dsm(void);
 int nouveau_acpi_get_bios_chunk(uint8_t *bios, int offset, int len);
 bool nouveau_acpi_rom_supported(struct pci_dev *pdev);
 int nouveau_acpi_edid(struct drm_device *, struct drm_connector *);
 #else
 static inline void nouveau_register_dsm_handler(void) {}
 static inline void nouveau_unregister_dsm_handler(void) {}
-static inline void nouveau_switcheroo_optimus_dsm(void) {}
 static inline bool nouveau_acpi_rom_supported(struct pci_dev *pdev) { return false; }
 static inline int nouveau_acpi_get_bios_chunk(uint8_t *bios, int offset, int len) { return -EINVAL; }
 static inline int nouveau_acpi_edid(struct drm_device *dev, struct drm_connector *connector) { return -EINVAL; }
@@ -1115,15 +984,15 @@ static inline int nouveau_acpi_edid(struct drm_device *dev, struct drm_connector
 
 /* nouveau_backlight.c */
 #ifdef CONFIG_DRM_NOUVEAU_BACKLIGHT
-extern int nouveau_backlight_init(struct drm_device *);
-extern void nouveau_backlight_exit(struct drm_device *);
+extern int nouveau_backlight_init(struct drm_connector *);
+extern void nouveau_backlight_exit(struct drm_connector *);
 #else
-static inline int nouveau_backlight_init(struct drm_device *dev)
+static inline int nouveau_backlight_init(struct drm_connector *dev)
 {
 	return 0;
 }
 
-static inline void nouveau_backlight_exit(struct drm_device *dev) { }
+static inline void nouveau_backlight_exit(struct drm_connector *dev) { }
 #endif
 
 /* nouveau_bios.c */
@@ -1131,15 +1000,19 @@ extern int nouveau_bios_init(struct drm_device *);
 extern void nouveau_bios_takedown(struct drm_device *dev);
 extern int nouveau_run_vbios_init(struct drm_device *);
 extern void nouveau_bios_run_init_table(struct drm_device *, uint16_t table,
-					struct dcb_entry *, int crtc);
-extern void nouveau_bios_init_exec(struct drm_device *, uint16_t table);
+					struct dcb_entry *);
+extern struct dcb_gpio_entry *nouveau_bios_gpio_entry(struct drm_device *,
+						      enum dcb_gpio_tag);
 extern struct dcb_connector_table_entry *
 nouveau_bios_connector_entry(struct drm_device *, int index);
 extern u32 get_pll_register(struct drm_device *, enum pll_types);
 extern int get_pll_limits(struct drm_device *, uint32_t limit_match,
 			  struct pll_lims *);
-extern int nouveau_bios_run_display_table(struct drm_device *, u16 id, int clk,
-					  struct dcb_entry *, int crtc);
+extern int nouveau_bios_run_display_table(struct drm_device *,
+					  struct dcb_entry *,
+					  uint32_t script, int pxclk);
+extern void *nouveau_bios_dp_table(struct drm_device *, struct dcb_entry *,
+				   int *length);
 extern bool nouveau_bios_fp_mode(struct drm_device *, struct drm_display_mode *);
 extern uint8_t *nouveau_bios_embedded_edid(struct drm_device *);
 extern int nouveau_bios_parse_lvds_table(struct drm_device *, int pxclk,
@@ -1148,28 +1021,23 @@ extern int run_tmds_table(struct drm_device *, struct dcb_entry *,
 			  int head, int pxclk);
 extern int call_lvds_script(struct drm_device *, struct dcb_entry *, int head,
 			    enum LVDS_script, int pxclk);
-bool bios_encoder_match(struct dcb_entry *, u32 hash);
-
-/* nouveau_mxm.c */
-int  nouveau_mxm_init(struct drm_device *dev);
-void nouveau_mxm_fini(struct drm_device *dev);
 
 /* nouveau_ttm.c */
 int nouveau_ttm_global_init(struct drm_nouveau_private *);
 void nouveau_ttm_global_release(struct drm_nouveau_private *);
 int nouveau_ttm_mmap(struct file *, struct vm_area_struct *);
 
-/* nouveau_hdmi.c */
-void nouveau_hdmi_mode_set(struct drm_encoder *, struct drm_display_mode *);
+/* nouveau_dp.c */
+int nouveau_dp_auxch(struct nouveau_i2c_chan *auxch, int cmd, int addr,
+		     uint8_t *data, int data_nr);
+bool nouveau_dp_detect(struct drm_encoder *);
+bool nouveau_dp_link_train(struct drm_encoder *);
 
 /* nv04_fb.c */
-extern int  nv04_fb_vram_init(struct drm_device *);
 extern int  nv04_fb_init(struct drm_device *);
 extern void nv04_fb_takedown(struct drm_device *);
 
 /* nv10_fb.c */
-extern int  nv10_fb_vram_init(struct drm_device *dev);
-extern int  nv1a_fb_vram_init(struct drm_device *dev);
 extern int  nv10_fb_init(struct drm_device *);
 extern void nv10_fb_takedown(struct drm_device *);
 extern void nv10_fb_init_tile_region(struct drm_device *dev, int i,
@@ -1178,16 +1046,6 @@ extern void nv10_fb_init_tile_region(struct drm_device *dev, int i,
 extern void nv10_fb_set_tile_region(struct drm_device *dev, int i);
 extern void nv10_fb_free_tile_region(struct drm_device *dev, int i);
 
-/* nv20_fb.c */
-extern int  nv20_fb_vram_init(struct drm_device *dev);
-extern int  nv20_fb_init(struct drm_device *);
-extern void nv20_fb_takedown(struct drm_device *);
-extern void nv20_fb_init_tile_region(struct drm_device *dev, int i,
-				     uint32_t addr, uint32_t size,
-				     uint32_t pitch, uint32_t flags);
-extern void nv20_fb_set_tile_region(struct drm_device *dev, int i);
-extern void nv20_fb_free_tile_region(struct drm_device *dev, int i);
-
 /* nv30_fb.c */
 extern int  nv30_fb_init(struct drm_device *);
 extern void nv30_fb_takedown(struct drm_device *);
@@ -1197,7 +1055,6 @@ extern void nv30_fb_init_tile_region(struct drm_device *dev, int i,
 extern void nv30_fb_free_tile_region(struct drm_device *dev, int i);
 
 /* nv40_fb.c */
-extern int  nv40_fb_vram_init(struct drm_device *dev);
 extern int  nv40_fb_init(struct drm_device *);
 extern void nv40_fb_takedown(struct drm_device *);
 extern void nv40_fb_set_tile_region(struct drm_device *dev, int i);
@@ -1263,6 +1120,7 @@ extern int  nvc0_fifo_unload_context(struct drm_device *);
 
 /* nv04_graph.c */
 extern int  nv04_graph_create(struct drm_device *);
+extern void nv04_graph_fifo_access(struct drm_device *, bool);
 extern int  nv04_graph_object_new(struct nouveau_channel *, int, u32, u16);
 extern int  nv04_graph_mthd_page_flip(struct nouveau_channel *chan,
 				      u32 class, u32 mthd, u32 data);
@@ -1294,39 +1152,24 @@ extern int  nvc0_graph_isr_chid(struct drm_device *dev, u64 inst);
 /* nv84_crypt.c */
 extern int  nv84_crypt_create(struct drm_device *);
 
-/* nv98_crypt.c */
-extern int  nv98_crypt_create(struct drm_device *dev);
-
 /* nva3_copy.c */
 extern int  nva3_copy_create(struct drm_device *dev);
 
 /* nvc0_copy.c */
 extern int  nvc0_copy_create(struct drm_device *dev, int engine);
 
-/* nv31_mpeg.c */
-extern int  nv31_mpeg_create(struct drm_device *dev);
+/* nv40_mpeg.c */
+extern int  nv40_mpeg_create(struct drm_device *dev);
 
 /* nv50_mpeg.c */
 extern int  nv50_mpeg_create(struct drm_device *dev);
 
-/* nv84_bsp.c */
-/* nv98_bsp.c */
-extern int  nv84_bsp_create(struct drm_device *dev);
-
-/* nv84_vp.c */
-/* nv98_vp.c */
-extern int  nv84_vp_create(struct drm_device *dev);
-
-/* nv98_ppp.c */
-extern int  nv98_ppp_create(struct drm_device *dev);
-
 /* nv04_instmem.c */
 extern int  nv04_instmem_init(struct drm_device *);
 extern void nv04_instmem_takedown(struct drm_device *);
 extern int  nv04_instmem_suspend(struct drm_device *);
 extern void nv04_instmem_resume(struct drm_device *);
-extern int  nv04_instmem_get(struct nouveau_gpuobj *, struct nouveau_channel *,
-			     u32 size, u32 align);
+extern int  nv04_instmem_get(struct nouveau_gpuobj *, u32 size, u32 align);
 extern void nv04_instmem_put(struct nouveau_gpuobj *);
 extern int  nv04_instmem_map(struct nouveau_gpuobj *);
 extern void nv04_instmem_unmap(struct nouveau_gpuobj *);
@@ -1337,8 +1180,7 @@ extern int  nv50_instmem_init(struct drm_device *);
 extern void nv50_instmem_takedown(struct drm_device *);
 extern int  nv50_instmem_suspend(struct drm_device *);
 extern void nv50_instmem_resume(struct drm_device *);
-extern int  nv50_instmem_get(struct nouveau_gpuobj *, struct nouveau_channel *,
-			     u32 size, u32 align);
+extern int  nv50_instmem_get(struct nouveau_gpuobj *, u32 size, u32 align);
 extern void nv50_instmem_put(struct nouveau_gpuobj *);
 extern int  nv50_instmem_map(struct nouveau_gpuobj *);
 extern void nv50_instmem_unmap(struct nouveau_gpuobj *);
@@ -1397,28 +1239,18 @@ extern int nv17_tv_create(struct drm_connector *, struct dcb_entry *);
 extern int nv04_display_early_init(struct drm_device *);
 extern void nv04_display_late_takedown(struct drm_device *);
 extern int nv04_display_create(struct drm_device *);
-extern void nv04_display_destroy(struct drm_device *);
 extern int nv04_display_init(struct drm_device *);
-extern void nv04_display_fini(struct drm_device *);
-
-/* nvd0_display.c */
-extern int nvd0_display_create(struct drm_device *);
-extern void nvd0_display_destroy(struct drm_device *);
-extern int nvd0_display_init(struct drm_device *);
-extern void nvd0_display_fini(struct drm_device *);
-struct nouveau_bo *nvd0_display_crtc_sema(struct drm_device *, int crtc);
-void nvd0_display_flip_stop(struct drm_crtc *);
-int nvd0_display_flip_next(struct drm_crtc *, struct drm_framebuffer *,
-			   struct nouveau_channel *, u32 swap_interval);
+extern void nv04_display_destroy(struct drm_device *);
 
 /* nv04_crtc.c */
 extern int nv04_crtc_create(struct drm_device *, int index);
 
 /* nouveau_bo.c */
 extern struct ttm_bo_driver nouveau_bo_driver;
-extern int nouveau_bo_new(struct drm_device *, int size, int align,
-			  uint32_t flags, uint32_t tile_mode,
-			  uint32_t tile_flags, struct nouveau_bo **);
+extern int nouveau_bo_new(struct drm_device *, struct nouveau_channel *,
+			  int size, int align, uint32_t flags,
+			  uint32_t tile_mode, uint32_t tile_flags,
+			  struct nouveau_bo **);
 extern int nouveau_bo_pin(struct nouveau_bo *, uint32_t flags);
 extern int nouveau_bo_unpin(struct nouveau_bo *);
 extern int nouveau_bo_map(struct nouveau_bo *);
@@ -1433,12 +1265,6 @@ extern void nouveau_bo_fence(struct nouveau_bo *, struct nouveau_fence *);
 extern int nouveau_bo_validate(struct nouveau_bo *, bool interruptible,
 			       bool no_wait_reserve, bool no_wait_gpu);
 
-extern struct nouveau_vma *
-nouveau_bo_vma_find(struct nouveau_bo *, struct nouveau_vm *);
-extern int  nouveau_bo_vma_add(struct nouveau_bo *, struct nouveau_vm *,
-			       struct nouveau_vma *);
-extern void nouveau_bo_vma_del(struct nouveau_bo *, struct nouveau_vma *);
-
 /* nouveau_fence.c */
 struct nouveau_fence;
 extern int nouveau_fence_init(struct drm_device *);
@@ -1484,14 +1310,12 @@ static inline struct nouveau_fence *nouveau_fence_ref(struct nouveau_fence *obj)
 }
 
 /* nouveau_gem.c */
-extern int nouveau_gem_new(struct drm_device *, int size, int align,
-			   uint32_t domain, uint32_t tile_mode,
-			   uint32_t tile_flags, struct nouveau_bo **);
+extern int nouveau_gem_new(struct drm_device *, struct nouveau_channel *,
+			   int size, int align, uint32_t domain,
+			   uint32_t tile_mode, uint32_t tile_flags,
+			   struct nouveau_bo **);
 extern int nouveau_gem_object_new(struct drm_gem_object *);
 extern void nouveau_gem_object_del(struct drm_gem_object *);
-extern int nouveau_gem_object_open(struct drm_gem_object *, struct drm_file *);
-extern void nouveau_gem_object_close(struct drm_gem_object *,
-				     struct drm_file *);
 extern int nouveau_gem_ioctl_new(struct drm_device *, void *,
 				 struct drm_file *);
 extern int nouveau_gem_ioctl_pushbuf(struct drm_device *, void *,
@@ -1504,40 +1328,29 @@ extern int nouveau_gem_ioctl_info(struct drm_device *, void *,
 				  struct drm_file *);
 
 /* nouveau_display.c */
-int nouveau_display_create(struct drm_device *dev);
-void nouveau_display_destroy(struct drm_device *dev);
-int nouveau_display_init(struct drm_device *dev);
-void nouveau_display_fini(struct drm_device *dev);
 int nouveau_vblank_enable(struct drm_device *dev, int crtc);
 void nouveau_vblank_disable(struct drm_device *dev, int crtc);
 int nouveau_crtc_page_flip(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 			   struct drm_pending_vblank_event *event);
 int nouveau_finish_page_flip(struct nouveau_channel *,
 			     struct nouveau_page_flip_state *);
-int nouveau_display_dumb_create(struct drm_file *, struct drm_device *,
-				struct drm_mode_create_dumb *args);
-int nouveau_display_dumb_map_offset(struct drm_file *, struct drm_device *,
-				    uint32_t handle, uint64_t *offset);
-int nouveau_display_dumb_destroy(struct drm_file *, struct drm_device *,
-				 uint32_t handle);
 
 /* nv10_gpio.c */
-int nv10_gpio_init(struct drm_device *dev);
-void nv10_gpio_fini(struct drm_device *dev);
-int nv10_gpio_drive(struct drm_device *dev, int line, int dir, int out);
-int nv10_gpio_sense(struct drm_device *dev, int line);
-void nv10_gpio_irq_enable(struct drm_device *, int line, bool on);
+int nv10_gpio_get(struct drm_device *dev, enum dcb_gpio_tag tag);
+int nv10_gpio_set(struct drm_device *dev, enum dcb_gpio_tag tag, int state);
 
 /* nv50_gpio.c */
 int nv50_gpio_init(struct drm_device *dev);
 void nv50_gpio_fini(struct drm_device *dev);
-int nv50_gpio_drive(struct drm_device *dev, int line, int dir, int out);
-int nv50_gpio_sense(struct drm_device *dev, int line);
-void nv50_gpio_irq_enable(struct drm_device *, int line, bool on);
-int nvd0_gpio_drive(struct drm_device *dev, int line, int dir, int out);
-int nvd0_gpio_sense(struct drm_device *dev, int line);
-
-/* nv50_calc.c */
+int nv50_gpio_get(struct drm_device *dev, enum dcb_gpio_tag tag);
+int nv50_gpio_set(struct drm_device *dev, enum dcb_gpio_tag tag, int state);
+int  nv50_gpio_irq_register(struct drm_device *, enum dcb_gpio_tag,
+			    void (*)(void *, int), void *);
+void nv50_gpio_irq_unregister(struct drm_device *, enum dcb_gpio_tag,
+			      void (*)(void *, int), void *);
+bool nv50_gpio_irq_enable(struct drm_device *, enum dcb_gpio_tag, bool on);
+
+/* nv50_calc. */
 int nv50_calc_pll(struct drm_device *, struct pll_lims *, int clk,
 		  int *N1, int *M1, int *N2, int *M2, int *P);
 int nva3_calc_pll(struct drm_device *, struct pll_lims *,
@@ -1605,8 +1418,6 @@ static inline void nv_wr08(struct drm_device *dev, unsigned reg, u8 val)
 	nouveau_wait_eq(dev, 2000000000ULL, (reg), (mask), (val))
 #define nv_wait_ne(dev, reg, mask, val) \
 	nouveau_wait_ne(dev, 2000000000ULL, (reg), (mask), (val))
-#define nv_wait_cb(dev, func, data) \
-	nouveau_wait_cb(dev, 2000000000ULL, (func), (data))
 
 /* PRAMIN access */
 static inline u32 nv_ri32(struct drm_device *dev, unsigned offset)
@@ -1660,13 +1471,6 @@ extern void nv_wo32(struct nouveau_gpuobj *, u32 offset, u32 val);
 #define NV_TRACEWARN(d, fmt, arg...) NV_PRINTK(KERN_NOTICE, d, fmt, ##arg)
 #define NV_TRACE(d, fmt, arg...) NV_PRINTK(KERN_INFO, d, fmt, ##arg)
 #define NV_WARN(d, fmt, arg...) NV_PRINTK(KERN_WARNING, d, fmt, ##arg)
-#define NV_WARNONCE(d, fmt, arg...) do {                                       \
-	static int _warned = 0;                                                \
-	if (!_warned) {                                                        \
-		NV_WARN(d, fmt, ##arg);                                        \
-		_warned = 1;                                                   \
-	}                                                                      \
-} while(0)
 
 /* nouveau_reg_debug bitmask */
 enum {
@@ -1680,7 +1484,6 @@ enum {
 	NOUVEAU_REG_DEBUG_RMVIO          = 0x80,
 	NOUVEAU_REG_DEBUG_VGAATTR        = 0x100,
 	NOUVEAU_REG_DEBUG_EVO            = 0x200,
-	NOUVEAU_REG_DEBUG_AUXCH          = 0x400
 };
 
 #define NV_REG_DEBUG(type, dev, fmt, arg...) do { \
@@ -1754,7 +1557,6 @@ nv44_graph_class(struct drm_device *dev)
 #define NV_MEM_ACCESS_RW (NV_MEM_ACCESS_RO | NV_MEM_ACCESS_WO)
 #define NV_MEM_ACCESS_SYS 4
 #define NV_MEM_ACCESS_VM  8
-#define NV_MEM_ACCESS_NOSNOOP 16
 
 #define NV_MEM_TARGET_VRAM        0
 #define NV_MEM_TARGET_PCI         1
diff --git a/drivers/gpu/drm/nouveau/nouveau_encoder.h b/drivers/gpu/drm/nouveau/nouveau_encoder.h
index 3dc14a3..ae69b61 100644
--- a/drivers/gpu/drm/nouveau/nouveau_encoder.h
+++ b/drivers/gpu/drm/nouveau/nouveau_encoder.h
@@ -32,14 +32,6 @@
 
 #define NV_DPMS_CLEARED 0x80
 
-struct dp_train_func {
-	void (*link_set)(struct drm_device *, struct dcb_entry *, int crtc,
-			 int nr, u32 bw, bool enhframe);
-	void (*train_set)(struct drm_device *, struct dcb_entry *, u8 pattern);
-	void (*train_adj)(struct drm_device *, struct dcb_entry *,
-			  u8 lane, u8 swing, u8 preem);
-};
-
 struct nouveau_encoder {
 	struct drm_encoder_slave base;
 
@@ -57,17 +49,17 @@ struct nouveau_encoder {
 
 	union {
 		struct {
-			u8  dpcd[8];
+			int mc_unknown;
+			uint32_t unk0;
+			uint32_t unk1;
+			int dpcd_version;
 			int link_nr;
 			int link_bw;
-			u32 datarate;
+			bool enhanced_frame;
 		} dp;
 	};
 };
 
-struct nouveau_encoder *
-find_encoder(struct drm_connector *connector, int type);
-
 static inline struct nouveau_encoder *nouveau_encoder(struct drm_encoder *enc)
 {
 	struct drm_encoder_slave *slave = to_encoder_slave(enc);
@@ -86,19 +78,26 @@ get_slave_funcs(struct drm_encoder *enc)
 	return to_encoder_slave(enc)->slave_funcs;
 }
 
-/* nouveau_dp.c */
-int nouveau_dp_auxch(struct nouveau_i2c_chan *auxch, int cmd, int addr,
-		     uint8_t *data, int data_nr);
-bool nouveau_dp_detect(struct drm_encoder *);
-void nouveau_dp_dpms(struct drm_encoder *, int mode, u32 datarate,
-		     struct dp_train_func *);
-u8 *nouveau_dp_bios_data(struct drm_device *, struct dcb_entry *, u8 **);
-
 struct nouveau_connector *
 nouveau_encoder_connector_get(struct nouveau_encoder *encoder);
 int nv50_sor_create(struct drm_connector *, struct dcb_entry *);
-void nv50_sor_dp_calc_tu(struct drm_device *, int, int, u32, u32);
 int nv50_dac_create(struct drm_connector *, struct dcb_entry *);
 
+struct bit_displayport_encoder_table {
+	uint32_t match;
+	uint8_t  record_nr;
+	uint8_t  unknown;
+	uint16_t script0;
+	uint16_t script1;
+	uint16_t unknown_table;
+} __attribute__ ((packed));
+
+struct bit_displayport_encoder_table_entry {
+	uint8_t vs_level;
+	uint8_t pre_level;
+	uint8_t reg0;
+	uint8_t reg1;
+	uint8_t reg2;
+} __attribute__ ((packed));
 
 #endif /* __NOUVEAU_ENCODER_H__ */
diff --git a/drivers/gpu/drm/nouveau/nouveau_fb.h b/drivers/gpu/drm/nouveau/nouveau_fb.h
index f3fb649..a3a88ad 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fb.h
+++ b/drivers/gpu/drm/nouveau/nouveau_fb.h
@@ -30,7 +30,6 @@
 struct nouveau_framebuffer {
 	struct drm_framebuffer base;
 	struct nouveau_bo *nvbo;
-	struct nouveau_vma vma;
 	u32 r_dma;
 	u32 r_format;
 	u32 r_pitch;
@@ -42,6 +41,8 @@ nouveau_framebuffer(struct drm_framebuffer *fb)
 	return container_of(fb, struct nouveau_framebuffer, base);
 }
 
+extern const struct drm_mode_config_funcs nouveau_mode_config_funcs;
+
 int nouveau_framebuffer_init(struct drm_device *dev, struct nouveau_framebuffer *nouveau_fb,
-			     struct drm_mode_fb_cmd2 *mode_cmd, struct nouveau_bo *nvbo);
+			     struct drm_mode_fb_cmd *mode_cmd, struct nouveau_bo *nvbo);
 #endif /* __NOUVEAU_FB_H__ */
diff --git a/drivers/gpu/drm/nouveau/nouveau_fbcon.c b/drivers/gpu/drm/nouveau/nouveau_fbcon.c
index 6fd2211..ea71f78 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fbcon.c
+++ b/drivers/gpu/drm/nouveau/nouveau_fbcon.c
@@ -36,7 +36,6 @@
 #include <linux/init.h>
 #include <linux/screen_info.h>
 #include <linux/vga_switcheroo.h>
-#include <linux/console.h>
 
 #include "drmP.h"
 #include "drm.h"
@@ -280,9 +279,8 @@ nouveau_fbcon_create(struct nouveau_fbdev *nfbdev,
 	struct fb_info *info;
 	struct drm_framebuffer *fb;
 	struct nouveau_framebuffer *nouveau_fb;
-	struct nouveau_channel *chan;
 	struct nouveau_bo *nvbo;
-	struct drm_mode_fb_cmd2 mode_cmd;
+	struct drm_mode_fb_cmd mode_cmd;
 	struct pci_dev *pdev = dev->pdev;
 	struct device *device = &pdev->dev;
 	int size, ret;
@@ -290,17 +288,16 @@ nouveau_fbcon_create(struct nouveau_fbdev *nfbdev,
 	mode_cmd.width = sizes->surface_width;
 	mode_cmd.height = sizes->surface_height;
 
-	mode_cmd.pitches[0] = mode_cmd.width * (sizes->surface_bpp >> 3);
-	mode_cmd.pitches[0] = roundup(mode_cmd.pitches[0], 256);
+	mode_cmd.bpp = sizes->surface_bpp;
+	mode_cmd.pitch = mode_cmd.width * (mode_cmd.bpp >> 3);
+	mode_cmd.pitch = roundup(mode_cmd.pitch, 256);
+	mode_cmd.depth = sizes->surface_depth;
 
-	mode_cmd.pixel_format = drm_mode_legacy_fb_format(sizes->surface_bpp,
-							  sizes->surface_depth);
-
-	size = mode_cmd.pitches[0] * mode_cmd.height;
+	size = mode_cmd.pitch * mode_cmd.height;
 	size = roundup(size, PAGE_SIZE);
 
-	ret = nouveau_gem_new(dev, size, 0, NOUVEAU_GEM_DOMAIN_VRAM,
-			      0, 0x0000, &nvbo);
+	ret = nouveau_gem_new(dev, dev_priv->channel, size, 0,
+			      NOUVEAU_GEM_DOMAIN_VRAM, 0, 0x0000, &nvbo);
 	if (ret) {
 		NV_ERROR(dev, "failed to allocate framebuffer\n");
 		goto out;
@@ -321,15 +318,6 @@ nouveau_fbcon_create(struct nouveau_fbdev *nfbdev,
 		goto out;
 	}
 
-	chan = nouveau_nofbaccel ? NULL : dev_priv->channel;
-	if (chan && dev_priv->card_type >= NV_50) {
-		ret = nouveau_bo_vma_add(nvbo, chan->vm, &nfbdev->nouveau_fb.vma);
-		if (ret) {
-			NV_ERROR(dev, "failed to map fb into chan: %d\n", ret);
-			chan = NULL;
-		}
-	}
-
 	mutex_lock(&dev->struct_mutex);
 
 	info = framebuffer_alloc(0, device);
@@ -371,7 +359,7 @@ nouveau_fbcon_create(struct nouveau_fbdev *nfbdev,
 	info->screen_base = nvbo_kmap_obj_iovirtual(nouveau_fb->nvbo);
 	info->screen_size = size;
 
-	drm_fb_helper_fill_fix(info, fb->pitches[0], fb->depth);
+	drm_fb_helper_fill_fix(info, fb->pitch, fb->depth);
 	drm_fb_helper_fill_var(info, &nfbdev->helper, sizes->fb_width, sizes->fb_height);
 
 	/* Set aperture base/size for vesafb takeover */
@@ -381,7 +369,11 @@ nouveau_fbcon_create(struct nouveau_fbdev *nfbdev,
 		goto out_unref;
 	}
 
-	/* Use default scratch pixmap (info->pixmap.flags = FB_PIXMAP_SYSTEM) */
+	info->pixmap.size = 64*1024;
+	info->pixmap.buf_align = 8;
+	info->pixmap.access_align = 32;
+	info->pixmap.flags = FB_PIXMAP_SYSTEM;
+	info->pixmap.scan_align = 1;
 
 	mutex_unlock(&dev->struct_mutex);
 
@@ -456,7 +448,6 @@ nouveau_fbcon_destroy(struct drm_device *dev, struct nouveau_fbdev *nfbdev)
 
 	if (nouveau_fb->nvbo) {
 		nouveau_bo_unmap(nouveau_fb->nvbo);
-		nouveau_bo_vma_del(nouveau_fb->nvbo, &nouveau_fb->vma);
 		drm_gem_object_unreference_unlocked(nouveau_fb->nvbo->gem);
 		nouveau_fb->nvbo = NULL;
 	}
@@ -485,7 +476,6 @@ int nouveau_fbcon_init(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_fbdev *nfbdev;
-	int preferred_bpp;
 	int ret;
 
 	nfbdev = kzalloc(sizeof(struct nouveau_fbdev), GFP_KERNEL);
@@ -504,15 +494,7 @@ int nouveau_fbcon_init(struct drm_device *dev)
 	}
 
 	drm_fb_helper_single_add_all_connectors(&nfbdev->helper);
-
-	if (dev_priv->vram_size <= 32 * 1024 * 1024)
-		preferred_bpp = 8;
-	else if (dev_priv->vram_size <= 64 * 1024 * 1024)
-		preferred_bpp = 16;
-	else
-		preferred_bpp = 32;
-
-	drm_fb_helper_initial_config(&nfbdev->helper, preferred_bpp);
+	drm_fb_helper_initial_config(&nfbdev->helper, 32);
 	return 0;
 }
 
@@ -545,13 +527,7 @@ void nouveau_fbcon_restore_accel(struct drm_device *dev)
 void nouveau_fbcon_set_suspend(struct drm_device *dev, int state)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	console_lock();
-	if (state == 0)
-		nouveau_fbcon_save_disable_accel(dev);
 	fb_set_suspend(dev_priv->nfbdev->helper.fbdev, state);
-	if (state == 1)
-		nouveau_fbcon_restore_accel(dev);
-	console_unlock();
 }
 
 void nouveau_fbcon_zfill_all(struct drm_device *dev)
diff --git a/drivers/gpu/drm/nouveau/nouveau_fence.c b/drivers/gpu/drm/nouveau/nouveau_fence.c
index 2f6daae..56f06b0 100644
--- a/drivers/gpu/drm/nouveau/nouveau_fence.c
+++ b/drivers/gpu/drm/nouveau/nouveau_fence.c
@@ -336,7 +336,6 @@ semaphore_acquire(struct nouveau_channel *chan, struct nouveau_semaphore *sema)
 {
 	struct drm_nouveau_private *dev_priv = chan->dev->dev_private;
 	struct nouveau_fence *fence = NULL;
-	u64 offset = chan->fence.vma.offset + sema->mem->start;
 	int ret;
 
 	if (dev_priv->chipset < 0x84) {
@@ -346,10 +345,13 @@ semaphore_acquire(struct nouveau_channel *chan, struct nouveau_semaphore *sema)
 
 		BEGIN_RING(chan, NvSubSw, NV_SW_DMA_SEMAPHORE, 3);
 		OUT_RING  (chan, NvSema);
-		OUT_RING  (chan, offset);
+		OUT_RING  (chan, sema->mem->start);
 		OUT_RING  (chan, 1);
 	} else
 	if (dev_priv->chipset < 0xc0) {
+		struct nouveau_vma *vma = &dev_priv->fence.bo->vma;
+		u64 offset = vma->offset + sema->mem->start;
+
 		ret = RING_SPACE(chan, 7);
 		if (ret)
 			return ret;
@@ -362,6 +364,9 @@ semaphore_acquire(struct nouveau_channel *chan, struct nouveau_semaphore *sema)
 		OUT_RING  (chan, 1);
 		OUT_RING  (chan, 1); /* ACQUIRE_EQ */
 	} else {
+		struct nouveau_vma *vma = &dev_priv->fence.bo->vma;
+		u64 offset = vma->offset + sema->mem->start;
+
 		ret = RING_SPACE(chan, 5);
 		if (ret)
 			return ret;
@@ -389,7 +394,6 @@ semaphore_release(struct nouveau_channel *chan, struct nouveau_semaphore *sema)
 {
 	struct drm_nouveau_private *dev_priv = chan->dev->dev_private;
 	struct nouveau_fence *fence = NULL;
-	u64 offset = chan->fence.vma.offset + sema->mem->start;
 	int ret;
 
 	if (dev_priv->chipset < 0x84) {
@@ -399,11 +403,14 @@ semaphore_release(struct nouveau_channel *chan, struct nouveau_semaphore *sema)
 
 		BEGIN_RING(chan, NvSubSw, NV_SW_DMA_SEMAPHORE, 2);
 		OUT_RING  (chan, NvSema);
-		OUT_RING  (chan, offset);
+		OUT_RING  (chan, sema->mem->start);
 		BEGIN_RING(chan, NvSubSw, NV_SW_SEMAPHORE_RELEASE, 1);
 		OUT_RING  (chan, 1);
 	} else
 	if (dev_priv->chipset < 0xc0) {
+		struct nouveau_vma *vma = &dev_priv->fence.bo->vma;
+		u64 offset = vma->offset + sema->mem->start;
+
 		ret = RING_SPACE(chan, 7);
 		if (ret)
 			return ret;
@@ -416,6 +423,9 @@ semaphore_release(struct nouveau_channel *chan, struct nouveau_semaphore *sema)
 		OUT_RING  (chan, 1);
 		OUT_RING  (chan, 2); /* RELEASE */
 	} else {
+		struct nouveau_vma *vma = &dev_priv->fence.bo->vma;
+		u64 offset = vma->offset + sema->mem->start;
+
 		ret = RING_SPACE(chan, 5);
 		if (ret)
 			return ret;
@@ -519,7 +529,7 @@ nouveau_fence_channel_init(struct nouveau_channel *chan)
 	if (USE_SEMA(dev) && dev_priv->chipset < 0x84) {
 		struct ttm_mem_reg *mem = &dev_priv->fence.bo->bo.mem;
 
-		ret = nouveau_gpuobj_dma_new(chan, NV_CLASS_DMA_FROM_MEMORY,
+		ret = nouveau_gpuobj_dma_new(chan, NV_CLASS_DMA_IN_MEMORY,
 					     mem->start << PAGE_SHIFT,
 					     mem->size, NV_MEM_ACCESS_RW,
 					     NV_MEM_TARGET_VRAM, &obj);
@@ -530,13 +540,6 @@ nouveau_fence_channel_init(struct nouveau_channel *chan)
 		nouveau_gpuobj_ref(NULL, &obj);
 		if (ret)
 			return ret;
-	} else
-	if (USE_SEMA(dev)) {
-		/* map fence bo into channel's vm */
-		ret = nouveau_bo_vma_add(dev_priv->fence.bo, chan->vm,
-					 &chan->fence.vma);
-		if (ret)
-			return ret;
 	}
 
 	atomic_set(&chan->fence.last_sequence_irq, 0);
@@ -546,10 +549,10 @@ nouveau_fence_channel_init(struct nouveau_channel *chan)
 void
 nouveau_fence_channel_fini(struct nouveau_channel *chan)
 {
-	struct drm_nouveau_private *dev_priv = chan->dev->dev_private;
 	struct nouveau_fence *tmp, *fence;
 
 	spin_lock(&chan->fence.lock);
+
 	list_for_each_entry_safe(fence, tmp, &chan->fence.pending, entry) {
 		fence->signalled = true;
 		list_del(&fence->entry);
@@ -559,9 +562,8 @@ nouveau_fence_channel_fini(struct nouveau_channel *chan)
 
 		kref_put(&fence->refcount, nouveau_fence_del);
 	}
-	spin_unlock(&chan->fence.lock);
 
-	nouveau_bo_vma_del(dev_priv->fence.bo, &chan->fence.vma);
+	spin_unlock(&chan->fence.lock);
 }
 
 int
@@ -573,7 +575,7 @@ nouveau_fence_init(struct drm_device *dev)
 
 	/* Create a shared VRAM heap for cross-channel sync. */
 	if (USE_SEMA(dev)) {
-		ret = nouveau_bo_new(dev, size, 0, TTM_PL_FLAG_VRAM,
+		ret = nouveau_bo_new(dev, NULL, size, 0, TTM_PL_FLAG_VRAM,
 				     0, 0, &dev_priv->fence.bo);
 		if (ret)
 			return ret;
diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c
index 7ce3fde..cee78b2 100644
--- a/drivers/gpu/drm/nouveau/nouveau_gem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_gem.c
@@ -60,71 +60,9 @@ nouveau_gem_object_del(struct drm_gem_object *gem)
 }
 
 int
-nouveau_gem_object_open(struct drm_gem_object *gem, struct drm_file *file_priv)
-{
-	struct nouveau_fpriv *fpriv = nouveau_fpriv(file_priv);
-	struct nouveau_bo *nvbo = nouveau_gem_object(gem);
-	struct nouveau_vma *vma;
-	int ret;
-
-	if (!fpriv->vm)
-		return 0;
-
-	ret = ttm_bo_reserve(&nvbo->bo, false, false, false, 0);
-	if (ret)
-		return ret;
-
-	vma = nouveau_bo_vma_find(nvbo, fpriv->vm);
-	if (!vma) {
-		vma = kzalloc(sizeof(*vma), GFP_KERNEL);
-		if (!vma) {
-			ret = -ENOMEM;
-			goto out;
-		}
-
-		ret = nouveau_bo_vma_add(nvbo, fpriv->vm, vma);
-		if (ret) {
-			kfree(vma);
-			goto out;
-		}
-	} else {
-		vma->refcount++;
-	}
-
-out:
-	ttm_bo_unreserve(&nvbo->bo);
-	return ret;
-}
-
-void
-nouveau_gem_object_close(struct drm_gem_object *gem, struct drm_file *file_priv)
-{
-	struct nouveau_fpriv *fpriv = nouveau_fpriv(file_priv);
-	struct nouveau_bo *nvbo = nouveau_gem_object(gem);
-	struct nouveau_vma *vma;
-	int ret;
-
-	if (!fpriv->vm)
-		return;
-
-	ret = ttm_bo_reserve(&nvbo->bo, false, false, false, 0);
-	if (ret)
-		return;
-
-	vma = nouveau_bo_vma_find(nvbo, fpriv->vm);
-	if (vma) {
-		if (--vma->refcount == 0) {
-			nouveau_bo_vma_del(nvbo, vma);
-			kfree(vma);
-		}
-	}
-	ttm_bo_unreserve(&nvbo->bo);
-}
-
-int
-nouveau_gem_new(struct drm_device *dev, int size, int align, uint32_t domain,
-		uint32_t tile_mode, uint32_t tile_flags,
-		struct nouveau_bo **pnvbo)
+nouveau_gem_new(struct drm_device *dev, struct nouveau_channel *chan,
+		int size, int align, uint32_t domain, uint32_t tile_mode,
+		uint32_t tile_flags, struct nouveau_bo **pnvbo)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_bo *nvbo;
@@ -138,7 +76,7 @@ nouveau_gem_new(struct drm_device *dev, int size, int align, uint32_t domain,
 	if (!flags || domain & NOUVEAU_GEM_DOMAIN_CPU)
 		flags |= TTM_PL_FLAG_SYSTEM;
 
-	ret = nouveau_bo_new(dev, size, align, flags, tile_mode,
+	ret = nouveau_bo_new(dev, chan, size, align, flags, tile_mode,
 			     tile_flags, pnvbo);
 	if (ret)
 		return ret;
@@ -165,28 +103,17 @@ nouveau_gem_new(struct drm_device *dev, int size, int align, uint32_t domain,
 }
 
 static int
-nouveau_gem_info(struct drm_file *file_priv, struct drm_gem_object *gem,
-		 struct drm_nouveau_gem_info *rep)
+nouveau_gem_info(struct drm_gem_object *gem, struct drm_nouveau_gem_info *rep)
 {
-	struct nouveau_fpriv *fpriv = nouveau_fpriv(file_priv);
 	struct nouveau_bo *nvbo = nouveau_gem_object(gem);
-	struct nouveau_vma *vma;
 
 	if (nvbo->bo.mem.mem_type == TTM_PL_TT)
 		rep->domain = NOUVEAU_GEM_DOMAIN_GART;
 	else
 		rep->domain = NOUVEAU_GEM_DOMAIN_VRAM;
 
-	rep->offset = nvbo->bo.offset;
-	if (fpriv->vm) {
-		vma = nouveau_bo_vma_find(nvbo, fpriv->vm);
-		if (!vma)
-			return -EINVAL;
-
-		rep->offset = vma->offset;
-	}
-
 	rep->size = nvbo->bo.mem.num_pages << PAGE_SHIFT;
+	rep->offset = nvbo->bo.offset;
 	rep->map_handle = nvbo->bo.addr_space_offset;
 	rep->tile_mode = nvbo->tile_mode;
 	rep->tile_flags = nvbo->tile_flags;
@@ -200,6 +127,7 @@ nouveau_gem_ioctl_new(struct drm_device *dev, void *data,
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct drm_nouveau_gem_new *req = data;
 	struct nouveau_bo *nvbo = NULL;
+	struct nouveau_channel *chan = NULL;
 	int ret = 0;
 
 	if (unlikely(dev_priv->ttm.bdev.dev_mapping == NULL))
@@ -210,21 +138,28 @@ nouveau_gem_ioctl_new(struct drm_device *dev, void *data,
 		return -EINVAL;
 	}
 
-	ret = nouveau_gem_new(dev, req->info.size, req->align,
+	if (req->channel_hint) {
+		chan = nouveau_channel_get(dev, file_priv, req->channel_hint);
+		if (IS_ERR(chan))
+			return PTR_ERR(chan);
+	}
+
+	ret = nouveau_gem_new(dev, chan, req->info.size, req->align,
 			      req->info.domain, req->info.tile_mode,
 			      req->info.tile_flags, &nvbo);
+	if (chan)
+		nouveau_channel_put(&chan);
 	if (ret)
 		return ret;
 
-	ret = drm_gem_handle_create(file_priv, nvbo->gem, &req->info.handle);
-	if (ret == 0) {
-		ret = nouveau_gem_info(file_priv, nvbo->gem, &req->info);
-		if (ret)
-			drm_gem_handle_delete(file_priv, req->info.handle);
-	}
+	ret = nouveau_gem_info(nvbo->gem, &req->info);
+	if (ret)
+		goto out;
 
+	ret = drm_gem_handle_create(file_priv, nvbo->gem, &req->info.handle);
 	/* drop reference from allocate - handle holds it now */
 	drm_gem_object_unreference_unlocked(nvbo->gem);
+out:
 	return ret;
 }
 
@@ -402,7 +337,6 @@ static int
 validate_list(struct nouveau_channel *chan, struct list_head *list,
 	      struct drm_nouveau_gem_pushbuf_bo *pbbo, uint64_t user_pbbo_ptr)
 {
-	struct drm_nouveau_private *dev_priv = chan->dev->dev_private;
 	struct drm_nouveau_gem_pushbuf_bo __user *upbbo =
 				(void __force __user *)(uintptr_t)user_pbbo_ptr;
 	struct drm_device *dev = chan->dev;
@@ -441,26 +375,24 @@ validate_list(struct nouveau_channel *chan, struct list_head *list,
 			return ret;
 		}
 
-		if (dev_priv->card_type < NV_50) {
-			if (nvbo->bo.offset == b->presumed.offset &&
-			    ((nvbo->bo.mem.mem_type == TTM_PL_VRAM &&
-			      b->presumed.domain & NOUVEAU_GEM_DOMAIN_VRAM) ||
-			     (nvbo->bo.mem.mem_type == TTM_PL_TT &&
-			      b->presumed.domain & NOUVEAU_GEM_DOMAIN_GART)))
-				continue;
+		if (nvbo->bo.offset == b->presumed.offset &&
+		    ((nvbo->bo.mem.mem_type == TTM_PL_VRAM &&
+		      b->presumed.domain & NOUVEAU_GEM_DOMAIN_VRAM) ||
+		     (nvbo->bo.mem.mem_type == TTM_PL_TT &&
+		      b->presumed.domain & NOUVEAU_GEM_DOMAIN_GART)))
+			continue;
 
-			if (nvbo->bo.mem.mem_type == TTM_PL_TT)
-				b->presumed.domain = NOUVEAU_GEM_DOMAIN_GART;
-			else
-				b->presumed.domain = NOUVEAU_GEM_DOMAIN_VRAM;
-			b->presumed.offset = nvbo->bo.offset;
-			b->presumed.valid = 0;
-			relocs++;
-
-			if (DRM_COPY_TO_USER(&upbbo[nvbo->pbbo_index].presumed,
-					     &b->presumed, sizeof(b->presumed)))
-				return -EFAULT;
-		}
+		if (nvbo->bo.mem.mem_type == TTM_PL_TT)
+			b->presumed.domain = NOUVEAU_GEM_DOMAIN_GART;
+		else
+			b->presumed.domain = NOUVEAU_GEM_DOMAIN_VRAM;
+		b->presumed.offset = nvbo->bo.offset;
+		b->presumed.valid = 0;
+		relocs++;
+
+		if (DRM_COPY_TO_USER(&upbbo[nvbo->pbbo_index].presumed,
+				     &b->presumed, sizeof(b->presumed)))
+			return -EFAULT;
 	}
 
 	return relocs;
@@ -635,7 +567,7 @@ nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data,
 	struct nouveau_fence *fence = NULL;
 	int i, j, ret = 0, do_reloc = 0;
 
-	chan = nouveau_channel_get(file_priv, req->channel);
+	chan = nouveau_channel_get(dev, file_priv, req->channel);
 	if (IS_ERR(chan))
 		return PTR_ERR(chan);
 
@@ -869,7 +801,7 @@ nouveau_gem_ioctl_info(struct drm_device *dev, void *data,
 	if (!gem)
 		return -ENOENT;
 
-	ret = nouveau_gem_info(file_priv, gem, req);
+	ret = nouveau_gem_info(gem, req);
 	drm_gem_object_unreference_unlocked(gem);
 	return ret;
 }
diff --git a/drivers/gpu/drm/nouveau/nouveau_gpio.c b/drivers/gpu/drm/nouveau/nouveau_gpio.c
deleted file mode 100644
index a580cc6..0000000
--- a/drivers/gpu/drm/nouveau/nouveau_gpio.c
+++ /dev/null
@@ -1,400 +0,0 @@
-/*
- * Copyright 2011 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Ben Skeggs
- */
-
-#include "drmP.h"
-#include "nouveau_drv.h"
-#include "nouveau_i2c.h"
-#include "nouveau_gpio.h"
-
-static u8 *
-dcb_gpio_table(struct drm_device *dev)
-{
-	u8 *dcb = dcb_table(dev);
-	if (dcb) {
-		if (dcb[0] >= 0x30 && dcb[1] >= 0x0c)
-			return ROMPTR(dev, dcb[0x0a]);
-		if (dcb[0] >= 0x22 && dcb[-1] >= 0x13)
-			return ROMPTR(dev, dcb[-15]);
-	}
-	return NULL;
-}
-
-static u8 *
-dcb_gpio_entry(struct drm_device *dev, int idx, int ent, u8 *version)
-{
-	u8 *table = dcb_gpio_table(dev);
-	if (table) {
-		*version = table[0];
-		if (*version < 0x30 && ent < table[2])
-			return table + 3 + (ent * table[1]);
-		else if (ent < table[2])
-			return table + table[1] + (ent * table[3]);
-	}
-	return NULL;
-}
-
-int
-nouveau_gpio_drive(struct drm_device *dev, int idx, int line, int dir, int out)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio;
-
-	return pgpio->drive ? pgpio->drive(dev, line, dir, out) : -ENODEV;
-}
-
-int
-nouveau_gpio_sense(struct drm_device *dev, int idx, int line)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio;
-
-	return pgpio->sense ? pgpio->sense(dev, line) : -ENODEV;
-}
-
-int
-nouveau_gpio_find(struct drm_device *dev, int idx, u8 func, u8 line,
-		  struct gpio_func *gpio)
-{
-	u8 *table, *entry, version;
-	int i = -1;
-
-	if (line == 0xff && func == 0xff)
-		return -EINVAL;
-
-	while ((entry = dcb_gpio_entry(dev, idx, ++i, &version))) {
-		if (version < 0x40) {
-			u16 data = ROM16(entry[0]);
-			*gpio = (struct gpio_func) {
-				.line = (data & 0x001f) >> 0,
-				.func = (data & 0x07e0) >> 5,
-				.log[0] = (data & 0x1800) >> 11,
-				.log[1] = (data & 0x6000) >> 13,
-			};
-		} else
-		if (version < 0x41) {
-			*gpio = (struct gpio_func) {
-				.line = entry[0] & 0x1f,
-				.func = entry[1],
-				.log[0] = (entry[3] & 0x18) >> 3,
-				.log[1] = (entry[3] & 0x60) >> 5,
-			};
-		} else {
-			*gpio = (struct gpio_func) {
-				.line = entry[0] & 0x3f,
-				.func = entry[1],
-				.log[0] = (entry[4] & 0x30) >> 4,
-				.log[1] = (entry[4] & 0xc0) >> 6,
-			};
-		}
-
-		if ((line == 0xff || line == gpio->line) &&
-		    (func == 0xff || func == gpio->func))
-			return 0;
-	}
-
-	/* DCB 2.2, fixed TVDAC GPIO data */
-	if ((table = dcb_table(dev)) && table[0] >= 0x22) {
-		if (func == DCB_GPIO_TVDAC0) {
-			*gpio = (struct gpio_func) {
-				.func = DCB_GPIO_TVDAC0,
-				.line = table[-4] >> 4,
-				.log[0] = !!(table[-5] & 2),
-				.log[1] =  !(table[-5] & 2),
-			};
-			return 0;
-		}
-	}
-
-	/* Apple iMac G4 NV18 */
-	if (nv_match_device(dev, 0x0189, 0x10de, 0x0010)) {
-		if (func == DCB_GPIO_TVDAC0) {
-			*gpio = (struct gpio_func) {
-				.func = DCB_GPIO_TVDAC0,
-				.line = 4,
-				.log[0] = 0,
-				.log[1] = 1,
-			};
-			return 0;
-		}
-	}
-
-	return -EINVAL;
-}
-
-int
-nouveau_gpio_set(struct drm_device *dev, int idx, u8 tag, u8 line, int state)
-{
-	struct gpio_func gpio;
-	int ret;
-
-	ret = nouveau_gpio_find(dev, idx, tag, line, &gpio);
-	if (ret == 0) {
-		int dir = !!(gpio.log[state] & 0x02);
-		int out = !!(gpio.log[state] & 0x01);
-		ret = nouveau_gpio_drive(dev, idx, gpio.line, dir, out);
-	}
-
-	return ret;
-}
-
-int
-nouveau_gpio_get(struct drm_device *dev, int idx, u8 tag, u8 line)
-{
-	struct gpio_func gpio;
-	int ret;
-
-	ret = nouveau_gpio_find(dev, idx, tag, line, &gpio);
-	if (ret == 0) {
-		ret = nouveau_gpio_sense(dev, idx, gpio.line);
-		if (ret >= 0)
-			ret = (ret == (gpio.log[1] & 1));
-	}
-
-	return ret;
-}
-
-int
-nouveau_gpio_irq(struct drm_device *dev, int idx, u8 tag, u8 line, bool on)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio;
-	struct gpio_func gpio;
-	int ret;
-
-	ret = nouveau_gpio_find(dev, idx, tag, line, &gpio);
-	if (ret == 0) {
-		if (idx == 0 && pgpio->irq_enable)
-			pgpio->irq_enable(dev, gpio.line, on);
-		else
-			ret = -ENODEV;
-	}
-
-	return ret;
-}
-
-struct gpio_isr {
-	struct drm_device *dev;
-	struct list_head head;
-	struct work_struct work;
-	int idx;
-	struct gpio_func func;
-	void (*handler)(void *, int);
-	void *data;
-	bool inhibit;
-};
-
-static void
-nouveau_gpio_isr_bh(struct work_struct *work)
-{
-	struct gpio_isr *isr = container_of(work, struct gpio_isr, work);
-	struct drm_device *dev = isr->dev;
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio;
-	unsigned long flags;
-	int state;
-
-	state = nouveau_gpio_get(dev, isr->idx, isr->func.func, isr->func.line);
-	if (state >= 0)
-		isr->handler(isr->data, state);
-
-	spin_lock_irqsave(&pgpio->lock, flags);
-	isr->inhibit = false;
-	spin_unlock_irqrestore(&pgpio->lock, flags);
-}
-
-void
-nouveau_gpio_isr(struct drm_device *dev, int idx, u32 line_mask)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio;
-	struct gpio_isr *isr;
-
-	if (idx != 0)
-		return;
-
-	spin_lock(&pgpio->lock);
-	list_for_each_entry(isr, &pgpio->isr, head) {
-		if (line_mask & (1 << isr->func.line)) {
-			if (isr->inhibit)
-				continue;
-			isr->inhibit = true;
-			schedule_work(&isr->work);
-		}
-	}
-	spin_unlock(&pgpio->lock);
-}
-
-int
-nouveau_gpio_isr_add(struct drm_device *dev, int idx, u8 tag, u8 line,
-		     void (*handler)(void *, int), void *data)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio;
-	struct gpio_isr *isr;
-	unsigned long flags;
-	int ret;
-
-	isr = kzalloc(sizeof(*isr), GFP_KERNEL);
-	if (!isr)
-		return -ENOMEM;
-
-	ret = nouveau_gpio_find(dev, idx, tag, line, &isr->func);
-	if (ret) {
-		kfree(isr);
-		return ret;
-	}
-
-	INIT_WORK(&isr->work, nouveau_gpio_isr_bh);
-	isr->dev = dev;
-	isr->handler = handler;
-	isr->data = data;
-	isr->idx = idx;
-
-	spin_lock_irqsave(&pgpio->lock, flags);
-	list_add(&isr->head, &pgpio->isr);
-	spin_unlock_irqrestore(&pgpio->lock, flags);
-	return 0;
-}
-
-void
-nouveau_gpio_isr_del(struct drm_device *dev, int idx, u8 tag, u8 line,
-		     void (*handler)(void *, int), void *data)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio;
-	struct gpio_isr *isr, *tmp;
-	struct gpio_func func;
-	unsigned long flags;
-	LIST_HEAD(tofree);
-	int ret;
-
-	ret = nouveau_gpio_find(dev, idx, tag, line, &func);
-	if (ret == 0) {
-		spin_lock_irqsave(&pgpio->lock, flags);
-		list_for_each_entry_safe(isr, tmp, &pgpio->isr, head) {
-			if (memcmp(&isr->func, &func, sizeof(func)) ||
-			    isr->idx != idx ||
-			    isr->handler != handler || isr->data != data)
-				continue;
-			list_move(&isr->head, &tofree);
-		}
-		spin_unlock_irqrestore(&pgpio->lock, flags);
-
-		list_for_each_entry_safe(isr, tmp, &tofree, head) {
-			flush_work_sync(&isr->work);
-			kfree(isr);
-		}
-	}
-}
-
-int
-nouveau_gpio_create(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio;
-
-	INIT_LIST_HEAD(&pgpio->isr);
-	spin_lock_init(&pgpio->lock);
-
-	return nouveau_gpio_init(dev);
-}
-
-void
-nouveau_gpio_destroy(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio;
-
-	nouveau_gpio_fini(dev);
-	BUG_ON(!list_empty(&pgpio->isr));
-}
-
-int
-nouveau_gpio_init(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio;
-	int ret = 0;
-
-	if (pgpio->init)
-		ret = pgpio->init(dev);
-
-	return ret;
-}
-
-void
-nouveau_gpio_fini(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio;
-
-	if (pgpio->fini)
-		pgpio->fini(dev);
-}
-
-void
-nouveau_gpio_reset(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	u8 *entry, version;
-	int ent = -1;
-
-	while ((entry = dcb_gpio_entry(dev, 0, ++ent, &version))) {
-		u8 func = 0xff, line, defs, unk0, unk1;
-		if (version >= 0x41) {
-			defs = !!(entry[0] & 0x80);
-			line = entry[0] & 0x3f;
-			func = entry[1];
-			unk0 = entry[2];
-			unk1 = entry[3] & 0x1f;
-		} else
-		if (version >= 0x40) {
-			line = entry[0] & 0x1f;
-			func = entry[1];
-			defs = !!(entry[3] & 0x01);
-			unk0 = !!(entry[3] & 0x02);
-			unk1 = !!(entry[3] & 0x04);
-		} else {
-			break;
-		}
-
-		if (func == 0xff)
-			continue;
-
-		nouveau_gpio_func_set(dev, func, defs);
-
-		if (dev_priv->card_type >= NV_D0) {
-			nv_mask(dev, 0x00d610 + (line * 4), 0xff, unk0);
-			if (unk1--)
-				nv_mask(dev, 0x00d640 + (unk1 * 4), 0xff, line);
-		} else
-		if (dev_priv->card_type >= NV_50) {
-			static const u32 regs[] = { 0xe100, 0xe28c };
-			u32 val = (unk1 << 16) | unk0;
-			u32 reg = regs[line >> 4]; line &= 0x0f;
-
-			nv_mask(dev, reg, 0x00010001 << line, val << line);
-		}
-	}
-}
diff --git a/drivers/gpu/drm/nouveau/nouveau_gpio.h b/drivers/gpu/drm/nouveau/nouveau_gpio.h
deleted file mode 100644
index 64c5cb0..0000000
--- a/drivers/gpu/drm/nouveau/nouveau_gpio.h
+++ /dev/null
@@ -1,71 +0,0 @@
-/*
- * Copyright 2011 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- */
-
-#ifndef __NOUVEAU_GPIO_H__
-#define __NOUVEAU_GPIO_H__
-
-struct gpio_func {
-	u8 func;
-	u8 line;
-	u8 log[2];
-};
-
-/* nouveau_gpio.c */
-int  nouveau_gpio_create(struct drm_device *);
-void nouveau_gpio_destroy(struct drm_device *);
-int  nouveau_gpio_init(struct drm_device *);
-void nouveau_gpio_fini(struct drm_device *);
-void nouveau_gpio_reset(struct drm_device *);
-int  nouveau_gpio_drive(struct drm_device *, int idx, int line,
-			int dir, int out);
-int  nouveau_gpio_sense(struct drm_device *, int idx, int line);
-int  nouveau_gpio_find(struct drm_device *, int idx, u8 tag, u8 line,
-		       struct gpio_func *);
-int  nouveau_gpio_set(struct drm_device *, int idx, u8 tag, u8 line, int state);
-int  nouveau_gpio_get(struct drm_device *, int idx, u8 tag, u8 line);
-int  nouveau_gpio_irq(struct drm_device *, int idx, u8 tag, u8 line, bool on);
-void nouveau_gpio_isr(struct drm_device *, int idx, u32 mask);
-int  nouveau_gpio_isr_add(struct drm_device *, int idx, u8 tag, u8 line,
-			  void (*)(void *, int state), void *data);
-void nouveau_gpio_isr_del(struct drm_device *, int idx, u8 tag, u8 line,
-			  void (*)(void *, int state), void *data);
-
-static inline bool
-nouveau_gpio_func_valid(struct drm_device *dev, u8 tag)
-{
-	struct gpio_func func;
-	return (nouveau_gpio_find(dev, 0, tag, 0xff, &func)) == 0;
-}
-
-static inline int
-nouveau_gpio_func_set(struct drm_device *dev, u8 tag, int state)
-{
-	return nouveau_gpio_set(dev, 0, tag, 0xff, state);
-}
-
-static inline int
-nouveau_gpio_func_get(struct drm_device *dev, u8 tag)
-{
-	return nouveau_gpio_get(dev, 0, tag, 0xff);
-}
-
-#endif
diff --git a/drivers/gpu/drm/nouveau/nouveau_hdmi.c b/drivers/gpu/drm/nouveau/nouveau_hdmi.c
deleted file mode 100644
index 59ea1c1..0000000
--- a/drivers/gpu/drm/nouveau/nouveau_hdmi.c
+++ /dev/null
@@ -1,258 +0,0 @@
-/*
- * Copyright 2011 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Ben Skeggs
- */
-
-#include "drmP.h"
-#include "nouveau_drv.h"
-#include "nouveau_connector.h"
-#include "nouveau_encoder.h"
-#include "nouveau_crtc.h"
-
-static bool
-hdmi_sor(struct drm_encoder *encoder)
-{
-	struct drm_nouveau_private *dev_priv = encoder->dev->dev_private;
-	if (dev_priv->chipset < 0xa3)
-		return false;
-	return true;
-}
-
-static inline u32
-hdmi_base(struct drm_encoder *encoder)
-{
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(nv_encoder->crtc);
-	if (!hdmi_sor(encoder))
-		return 0x616500 + (nv_crtc->index * 0x800);
-	return 0x61c500 + (nv_encoder->or * 0x800);
-}
-
-static void
-hdmi_wr32(struct drm_encoder *encoder, u32 reg, u32 val)
-{
-	nv_wr32(encoder->dev, hdmi_base(encoder) + reg, val);
-}
-
-static u32
-hdmi_rd32(struct drm_encoder *encoder, u32 reg)
-{
-	return nv_rd32(encoder->dev, hdmi_base(encoder) + reg);
-}
-
-static u32
-hdmi_mask(struct drm_encoder *encoder, u32 reg, u32 mask, u32 val)
-{
-	u32 tmp = hdmi_rd32(encoder, reg);
-	hdmi_wr32(encoder, reg, (tmp & ~mask) | val);
-	return tmp;
-}
-
-static void
-nouveau_audio_disconnect(struct drm_encoder *encoder)
-{
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct drm_device *dev = encoder->dev;
-	u32 or = nv_encoder->or * 0x800;
-
-	if (hdmi_sor(encoder)) {
-		nv_mask(dev, 0x61c448 + or, 0x00000003, 0x00000000);
-	}
-}
-
-static void
-nouveau_audio_mode_set(struct drm_encoder *encoder,
-		       struct drm_display_mode *mode)
-{
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct nouveau_connector *nv_connector;
-	struct drm_device *dev = encoder->dev;
-	u32 or = nv_encoder->or * 0x800;
-	int i;
-
-	nv_connector = nouveau_encoder_connector_get(nv_encoder);
-	if (!drm_detect_monitor_audio(nv_connector->edid)) {
-		nouveau_audio_disconnect(encoder);
-		return;
-	}
-
-	if (hdmi_sor(encoder)) {
-		nv_mask(dev, 0x61c448 + or, 0x00000001, 0x00000001);
-
-		drm_edid_to_eld(&nv_connector->base, nv_connector->edid);
-		if (nv_connector->base.eld[0]) {
-			u8 *eld = nv_connector->base.eld;
-			for (i = 0; i < eld[2] * 4; i++)
-				nv_wr32(dev, 0x61c440 + or, (i << 8) | eld[i]);
-			for (i = eld[2] * 4; i < 0x60; i++)
-				nv_wr32(dev, 0x61c440 + or, (i << 8) | 0x00);
-			nv_mask(dev, 0x61c448 + or, 0x00000002, 0x00000002);
-		}
-	}
-}
-
-static void
-nouveau_hdmi_infoframe(struct drm_encoder *encoder, u32 ctrl, u8 *frame)
-{
-	/* calculate checksum for the infoframe */
-	u8 sum = 0, i;
-	for (i = 0; i < frame[2]; i++)
-		sum += frame[i];
-	frame[3] = 256 - sum;
-
-	/* disable infoframe, and write header */
-	hdmi_mask(encoder, ctrl + 0x00, 0x00000001, 0x00000000);
-	hdmi_wr32(encoder, ctrl + 0x08, *(u32 *)frame & 0xffffff);
-
-	/* register scans tell me the audio infoframe has only one set of
-	 * subpack regs, according to tegra (gee nvidia, it'd be nice if we
-	 * could get those docs too!), the hdmi block pads out the rest of
-	 * the packet on its own.
-	 */
-	if (ctrl == 0x020)
-		frame[2] = 6;
-
-	/* write out checksum and data, weird weird 7 byte register pairs */
-	for (i = 0; i < frame[2] + 1; i += 7) {
-		u32 rsubpack = ctrl + 0x0c + ((i / 7) * 8);
-		u32 *subpack = (u32 *)&frame[3 + i];
-		hdmi_wr32(encoder, rsubpack + 0, subpack[0]);
-		hdmi_wr32(encoder, rsubpack + 4, subpack[1] & 0xffffff);
-	}
-
-	/* enable the infoframe */
-	hdmi_mask(encoder, ctrl, 0x00000001, 0x00000001);
-}
-
-static void
-nouveau_hdmi_video_infoframe(struct drm_encoder *encoder,
-			     struct drm_display_mode *mode)
-{
-	const u8 Y = 0, A = 0, B = 0, S = 0, C = 0, M = 0, R = 0;
-	const u8 ITC = 0, EC = 0, Q = 0, SC = 0, VIC = 0, PR = 0;
-	const u8 bar_top = 0, bar_bottom = 0, bar_left = 0, bar_right = 0;
-	u8 frame[20];
-
-	frame[0x00] = 0x82; /* AVI infoframe */
-	frame[0x01] = 0x02; /* version */
-	frame[0x02] = 0x0d; /* length */
-	frame[0x03] = 0x00;
-	frame[0x04] = (Y << 5) | (A << 4) | (B << 2) | S;
-	frame[0x05] = (C << 6) | (M << 4) | R;
-	frame[0x06] = (ITC << 7) | (EC << 4) | (Q << 2) | SC;
-	frame[0x07] = VIC;
-	frame[0x08] = PR;
-	frame[0x09] = bar_top & 0xff;
-	frame[0x0a] = bar_top >> 8;
-	frame[0x0b] = bar_bottom & 0xff;
-	frame[0x0c] = bar_bottom >> 8;
-	frame[0x0d] = bar_left & 0xff;
-	frame[0x0e] = bar_left >> 8;
-	frame[0x0f] = bar_right & 0xff;
-	frame[0x10] = bar_right >> 8;
-	frame[0x11] = 0x00;
-	frame[0x12] = 0x00;
-	frame[0x13] = 0x00;
-
-	nouveau_hdmi_infoframe(encoder, 0x020, frame);
-}
-
-static void
-nouveau_hdmi_audio_infoframe(struct drm_encoder *encoder,
-			     struct drm_display_mode *mode)
-{
-	const u8 CT = 0x00, CC = 0x01, ceaSS = 0x00, SF = 0x00, FMT = 0x00;
-	const u8 CA = 0x00, DM_INH = 0, LSV = 0x00;
-	u8 frame[12];
-
-	frame[0x00] = 0x84;	/* Audio infoframe */
-	frame[0x01] = 0x01;	/* version */
-	frame[0x02] = 0x0a;	/* length */
-	frame[0x03] = 0x00;
-	frame[0x04] = (CT << 4) | CC;
-	frame[0x05] = (SF << 2) | ceaSS;
-	frame[0x06] = FMT;
-	frame[0x07] = CA;
-	frame[0x08] = (DM_INH << 7) | (LSV << 3);
-	frame[0x09] = 0x00;
-	frame[0x0a] = 0x00;
-	frame[0x0b] = 0x00;
-
-	nouveau_hdmi_infoframe(encoder, 0x000, frame);
-}
-
-static void
-nouveau_hdmi_disconnect(struct drm_encoder *encoder)
-{
-	nouveau_audio_disconnect(encoder);
-
-	/* disable audio and avi infoframes */
-	hdmi_mask(encoder, 0x000, 0x00000001, 0x00000000);
-	hdmi_mask(encoder, 0x020, 0x00000001, 0x00000000);
-
-	/* disable hdmi */
-	hdmi_mask(encoder, 0x0a4, 0x40000000, 0x00000000);
-}
-
-void
-nouveau_hdmi_mode_set(struct drm_encoder *encoder,
-		      struct drm_display_mode *mode)
-{
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct nouveau_connector *nv_connector;
-	struct drm_device *dev = encoder->dev;
-	u32 max_ac_packet, rekey;
-
-	nv_connector = nouveau_encoder_connector_get(nv_encoder);
-	if (!mode || !nv_connector || !nv_connector->edid ||
-	    !drm_detect_hdmi_monitor(nv_connector->edid)) {
-		nouveau_hdmi_disconnect(encoder);
-		return;
-	}
-
-	nouveau_hdmi_video_infoframe(encoder, mode);
-	nouveau_hdmi_audio_infoframe(encoder, mode);
-
-	hdmi_mask(encoder, 0x0d0, 0x00070001, 0x00010001); /* SPARE, HW_CTS */
-	hdmi_mask(encoder, 0x068, 0x00010101, 0x00000000); /* ACR_CTRL, ?? */
-	hdmi_mask(encoder, 0x078, 0x80000000, 0x80000000); /* ACR_0441_ENABLE */
-
-	nv_mask(dev, 0x61733c, 0x00100000, 0x00100000); /* RESETF */
-	nv_mask(dev, 0x61733c, 0x10000000, 0x10000000); /* LOOKUP_EN */
-	nv_mask(dev, 0x61733c, 0x00100000, 0x00000000); /* !RESETF */
-
-	/* value matches nvidia binary driver, and tegra constant */
-	rekey = 56;
-
-	max_ac_packet  = mode->htotal - mode->hdisplay;
-	max_ac_packet -= rekey;
-	max_ac_packet -= 18; /* constant from tegra */
-	max_ac_packet /= 32;
-
-	/* enable hdmi */
-	hdmi_mask(encoder, 0x0a4, 0x5f1f003f, 0x40000000 | /* enable */
-					      0x1f000000 | /* unknown */
-					      max_ac_packet << 16 |
-					      rekey);
-
-	nouveau_audio_mode_set(encoder, mode);
-}
diff --git a/drivers/gpu/drm/nouveau/nouveau_hwsq.h b/drivers/gpu/drm/nouveau/nouveau_hwsq.h
deleted file mode 100644
index 6976875..0000000
--- a/drivers/gpu/drm/nouveau/nouveau_hwsq.h
+++ /dev/null
@@ -1,115 +0,0 @@
-/*
- * Copyright 2010 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Ben Skeggs
- */
-
-#ifndef __NOUVEAU_HWSQ_H__
-#define __NOUVEAU_HWSQ_H__
-
-struct hwsq_ucode {
-	u8 data[0x200];
-	union {
-		u8  *u08;
-		u16 *u16;
-		u32 *u32;
-	} ptr;
-	u16 len;
-
-	u32 reg;
-	u32 val;
-};
-
-static inline void
-hwsq_init(struct hwsq_ucode *hwsq)
-{
-	hwsq->ptr.u08 = hwsq->data;
-	hwsq->reg = 0xffffffff;
-	hwsq->val = 0xffffffff;
-}
-
-static inline void
-hwsq_fini(struct hwsq_ucode *hwsq)
-{
-	do {
-		*hwsq->ptr.u08++ = 0x7f;
-		hwsq->len = hwsq->ptr.u08 - hwsq->data;
-	} while (hwsq->len & 3);
-	hwsq->ptr.u08 = hwsq->data;
-}
-
-static inline void
-hwsq_usec(struct hwsq_ucode *hwsq, u8 usec)
-{
-	u32 shift = 0;
-	while (usec & ~3) {
-		usec >>= 2;
-		shift++;
-	}
-
-	*hwsq->ptr.u08++ = (shift << 2) | usec;
-}
-
-static inline void
-hwsq_setf(struct hwsq_ucode *hwsq, u8 flag, int val)
-{
-	flag += 0x80;
-	if (val >= 0)
-		flag += 0x20;
-	if (val >= 1)
-		flag += 0x20;
-	*hwsq->ptr.u08++ = flag;
-}
-
-static inline void
-hwsq_op5f(struct hwsq_ucode *hwsq, u8 v0, u8 v1)
-{
-	*hwsq->ptr.u08++ = 0x5f;
-	*hwsq->ptr.u08++ = v0;
-	*hwsq->ptr.u08++ = v1;
-}
-
-static inline void
-hwsq_wr32(struct hwsq_ucode *hwsq, u32 reg, u32 val)
-{
-	if (val != hwsq->val) {
-		if ((val & 0xffff0000) == (hwsq->val & 0xffff0000)) {
-			*hwsq->ptr.u08++ = 0x42;
-			*hwsq->ptr.u16++ = (val & 0x0000ffff);
-		} else {
-			*hwsq->ptr.u08++ = 0xe2;
-			*hwsq->ptr.u32++ = val;
-		}
-
-		hwsq->val = val;
-	}
-
-	if ((reg & 0xffff0000) == (hwsq->reg & 0xffff0000)) {
-		*hwsq->ptr.u08++ = 0x40;
-		*hwsq->ptr.u16++ = (reg & 0x0000ffff);
-	} else {
-		*hwsq->ptr.u08++ = 0xe0;
-		*hwsq->ptr.u32++ = reg;
-	}
-	hwsq->reg = reg;
-}
-
-#endif
diff --git a/drivers/gpu/drm/nouveau/nouveau_i2c.c b/drivers/gpu/drm/nouveau/nouveau_i2c.c
index 8f4f914..cb389d0 100644
--- a/drivers/gpu/drm/nouveau/nouveau_i2c.c
+++ b/drivers/gpu/drm/nouveau/nouveau_i2c.c
@@ -22,482 +22,257 @@
  * Authors: Ben Skeggs
  */
 
-#include <linux/module.h>
-
 #include "drmP.h"
 #include "nouveau_drv.h"
 #include "nouveau_i2c.h"
 #include "nouveau_hw.h"
 
-#define T_TIMEOUT  2200000
-#define T_RISEFALL 1000
-#define T_HOLD     5000
-
 static void
-i2c_drive_scl(void *data, int state)
+nv04_i2c_setscl(void *data, int state)
 {
-	struct nouveau_i2c_chan *port = data;
-	if (port->type == 0) {
-		u8 val = NVReadVgaCrtc(port->dev, 0, port->drive);
-		if (state) val |= 0x20;
-		else	   val &= 0xdf;
-		NVWriteVgaCrtc(port->dev, 0, port->drive, val | 0x01);
-	} else
-	if (port->type == 4) {
-		nv_mask(port->dev, port->drive, 0x2f, state ? 0x21 : 0x01);
-	} else
-	if (port->type == 5) {
-		if (state) port->state |= 0x01;
-		else	   port->state &= 0xfe;
-		nv_wr32(port->dev, port->drive, 4 | port->state);
-	}
+	struct nouveau_i2c_chan *i2c = data;
+	struct drm_device *dev = i2c->dev;
+	uint8_t val;
+
+	val = (NVReadVgaCrtc(dev, 0, i2c->wr) & 0xd0) | (state ? 0x20 : 0);
+	NVWriteVgaCrtc(dev, 0, i2c->wr, val | 0x01);
 }
 
 static void
-i2c_drive_sda(void *data, int state)
+nv04_i2c_setsda(void *data, int state)
 {
-	struct nouveau_i2c_chan *port = data;
-	if (port->type == 0) {
-		u8 val = NVReadVgaCrtc(port->dev, 0, port->drive);
-		if (state) val |= 0x10;
-		else	   val &= 0xef;
-		NVWriteVgaCrtc(port->dev, 0, port->drive, val | 0x01);
-	} else
-	if (port->type == 4) {
-		nv_mask(port->dev, port->drive, 0x1f, state ? 0x11 : 0x01);
-	} else
-	if (port->type == 5) {
-		if (state) port->state |= 0x02;
-		else	   port->state &= 0xfd;
-		nv_wr32(port->dev, port->drive, 4 | port->state);
-	}
-}
+	struct nouveau_i2c_chan *i2c = data;
+	struct drm_device *dev = i2c->dev;
+	uint8_t val;
 
-static int
-i2c_sense_scl(void *data)
-{
-	struct nouveau_i2c_chan *port = data;
-	struct drm_nouveau_private *dev_priv = port->dev->dev_private;
-	if (port->type == 0) {
-		return !!(NVReadVgaCrtc(port->dev, 0, port->sense) & 0x04);
-	} else
-	if (port->type == 4) {
-		return !!(nv_rd32(port->dev, port->sense) & 0x00040000);
-	} else
-	if (port->type == 5) {
-		if (dev_priv->card_type < NV_D0)
-			return !!(nv_rd32(port->dev, port->sense) & 0x01);
-		else
-			return !!(nv_rd32(port->dev, port->sense) & 0x10);
-	}
-	return 0;
+	val = (NVReadVgaCrtc(dev, 0, i2c->wr) & 0xe0) | (state ? 0x10 : 0);
+	NVWriteVgaCrtc(dev, 0, i2c->wr, val | 0x01);
 }
 
 static int
-i2c_sense_sda(void *data)
+nv04_i2c_getscl(void *data)
 {
-	struct nouveau_i2c_chan *port = data;
-	struct drm_nouveau_private *dev_priv = port->dev->dev_private;
-	if (port->type == 0) {
-		return !!(NVReadVgaCrtc(port->dev, 0, port->sense) & 0x08);
-	} else
-	if (port->type == 4) {
-		return !!(nv_rd32(port->dev, port->sense) & 0x00080000);
-	} else
-	if (port->type == 5) {
-		if (dev_priv->card_type < NV_D0)
-			return !!(nv_rd32(port->dev, port->sense) & 0x02);
-		else
-			return !!(nv_rd32(port->dev, port->sense) & 0x20);
-	}
-	return 0;
-}
-
-static void
-i2c_delay(struct nouveau_i2c_chan *port, u32 nsec)
-{
-	udelay((nsec + 500) / 1000);
-}
-
-static bool
-i2c_raise_scl(struct nouveau_i2c_chan *port)
-{
-	u32 timeout = T_TIMEOUT / T_RISEFALL;
-
-	i2c_drive_scl(port, 1);
-	do {
-		i2c_delay(port, T_RISEFALL);
-	} while (!i2c_sense_scl(port) && --timeout);
+	struct nouveau_i2c_chan *i2c = data;
+	struct drm_device *dev = i2c->dev;
 
-	return timeout != 0;
+	return !!(NVReadVgaCrtc(dev, 0, i2c->rd) & 4);
 }
 
 static int
-i2c_start(struct nouveau_i2c_chan *port)
+nv04_i2c_getsda(void *data)
 {
-	int ret = 0;
-
-	port->state  = i2c_sense_scl(port);
-	port->state |= i2c_sense_sda(port) << 1;
-	if (port->state != 3) {
-		i2c_drive_scl(port, 0);
-		i2c_drive_sda(port, 1);
-		if (!i2c_raise_scl(port))
-			ret = -EBUSY;
-	}
+	struct nouveau_i2c_chan *i2c = data;
+	struct drm_device *dev = i2c->dev;
 
-	i2c_drive_sda(port, 0);
-	i2c_delay(port, T_HOLD);
-	i2c_drive_scl(port, 0);
-	i2c_delay(port, T_HOLD);
-	return ret;
+	return !!(NVReadVgaCrtc(dev, 0, i2c->rd) & 8);
 }
 
 static void
-i2c_stop(struct nouveau_i2c_chan *port)
+nv4e_i2c_setscl(void *data, int state)
 {
-	i2c_drive_scl(port, 0);
-	i2c_drive_sda(port, 0);
-	i2c_delay(port, T_RISEFALL);
-
-	i2c_drive_scl(port, 1);
-	i2c_delay(port, T_HOLD);
-	i2c_drive_sda(port, 1);
-	i2c_delay(port, T_HOLD);
+	struct nouveau_i2c_chan *i2c = data;
+	struct drm_device *dev = i2c->dev;
+	uint8_t val;
+
+	val = (nv_rd32(dev, i2c->wr) & 0xd0) | (state ? 0x20 : 0);
+	nv_wr32(dev, i2c->wr, val | 0x01);
 }
 
-static int
-i2c_bitw(struct nouveau_i2c_chan *port, int sda)
+static void
+nv4e_i2c_setsda(void *data, int state)
 {
-	i2c_drive_sda(port, sda);
-	i2c_delay(port, T_RISEFALL);
-
-	if (!i2c_raise_scl(port))
-		return -ETIMEDOUT;
-	i2c_delay(port, T_HOLD);
+	struct nouveau_i2c_chan *i2c = data;
+	struct drm_device *dev = i2c->dev;
+	uint8_t val;
 
-	i2c_drive_scl(port, 0);
-	i2c_delay(port, T_HOLD);
-	return 0;
+	val = (nv_rd32(dev, i2c->wr) & 0xe0) | (state ? 0x10 : 0);
+	nv_wr32(dev, i2c->wr, val | 0x01);
 }
 
 static int
-i2c_bitr(struct nouveau_i2c_chan *port)
+nv4e_i2c_getscl(void *data)
 {
-	int sda;
-
-	i2c_drive_sda(port, 1);
-	i2c_delay(port, T_RISEFALL);
-
-	if (!i2c_raise_scl(port))
-		return -ETIMEDOUT;
-	i2c_delay(port, T_HOLD);
+	struct nouveau_i2c_chan *i2c = data;
+	struct drm_device *dev = i2c->dev;
 
-	sda = i2c_sense_sda(port);
-
-	i2c_drive_scl(port, 0);
-	i2c_delay(port, T_HOLD);
-	return sda;
+	return !!((nv_rd32(dev, i2c->rd) >> 16) & 4);
 }
 
 static int
-i2c_get_byte(struct nouveau_i2c_chan *port, u8 *byte, bool last)
+nv4e_i2c_getsda(void *data)
 {
-	int i, bit;
-
-	*byte = 0;
-	for (i = 7; i >= 0; i--) {
-		bit = i2c_bitr(port);
-		if (bit < 0)
-			return bit;
-		*byte |= bit << i;
-	}
+	struct nouveau_i2c_chan *i2c = data;
+	struct drm_device *dev = i2c->dev;
 
-	return i2c_bitw(port, last ? 1 : 0);
+	return !!((nv_rd32(dev, i2c->rd) >> 16) & 8);
 }
 
 static int
-i2c_put_byte(struct nouveau_i2c_chan *port, u8 byte)
+nv50_i2c_getscl(void *data)
 {
-	int i, ret;
-	for (i = 7; i >= 0; i--) {
-		ret = i2c_bitw(port, !!(byte & (1 << i)));
-		if (ret < 0)
-			return ret;
-	}
+	struct nouveau_i2c_chan *i2c = data;
+	struct drm_device *dev = i2c->dev;
 
-	ret = i2c_bitr(port);
-	if (ret == 1) /* nack */
-		ret = -EIO;
-	return ret;
+	return !!(nv_rd32(dev, i2c->rd) & 1);
 }
 
-static int
-i2c_addr(struct nouveau_i2c_chan *port, struct i2c_msg *msg)
-{
-	u32 addr = msg->addr << 1;
-	if (msg->flags & I2C_M_RD)
-		addr |= 1;
-	return i2c_put_byte(port, addr);
-}
 
 static int
-i2c_bit_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, int num)
+nv50_i2c_getsda(void *data)
 {
-	struct nouveau_i2c_chan *port = (struct nouveau_i2c_chan *)adap;
-	struct i2c_msg *msg = msgs;
-	int ret = 0, mcnt = num;
-
-	while (!ret && mcnt--) {
-		u8 remaining = msg->len;
-		u8 *ptr = msg->buf;
+	struct nouveau_i2c_chan *i2c = data;
+	struct drm_device *dev = i2c->dev;
 
-		ret = i2c_start(port);
-		if (ret == 0)
-			ret = i2c_addr(port, msg);
-
-		if (msg->flags & I2C_M_RD) {
-			while (!ret && remaining--)
-				ret = i2c_get_byte(port, ptr++, !remaining);
-		} else {
-			while (!ret && remaining--)
-				ret = i2c_put_byte(port, *ptr++);
-		}
+	return !!(nv_rd32(dev, i2c->rd) & 2);
+}
 
-		msg++;
-	}
+static void
+nv50_i2c_setscl(void *data, int state)
+{
+	struct nouveau_i2c_chan *i2c = data;
+	struct drm_device *dev = i2c->dev;
 
-	i2c_stop(port);
-	return (ret < 0) ? ret : num;
+	nv_wr32(dev, i2c->wr, 4 | (i2c->data ? 2 : 0) | (state ? 1 : 0));
 }
 
-static u32
-i2c_bit_func(struct i2c_adapter *adap)
+static void
+nv50_i2c_setsda(void *data, int state)
 {
-	return I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL;
-}
+	struct nouveau_i2c_chan *i2c = data;
+	struct drm_device *dev = i2c->dev;
 
-const struct i2c_algorithm nouveau_i2c_bit_algo = {
-	.master_xfer = i2c_bit_xfer,
-	.functionality = i2c_bit_func
-};
+	nv_wr32(dev, i2c->wr,
+			(nv_rd32(dev, i2c->rd) & 1) | 4 | (state ? 2 : 0));
+	i2c->data = state;
+}
 
 static const uint32_t nv50_i2c_port[] = {
 	0x00e138, 0x00e150, 0x00e168, 0x00e180,
 	0x00e254, 0x00e274, 0x00e764, 0x00e780,
 	0x00e79c, 0x00e7b8
 };
+#define NV50_I2C_PORTS ARRAY_SIZE(nv50_i2c_port)
 
-static u8 *
-i2c_table(struct drm_device *dev, u8 *version)
+int
+nouveau_i2c_init(struct drm_device *dev, struct dcb_i2c_entry *entry, int index)
 {
-	u8 *dcb = dcb_table(dev), *i2c = NULL;
-	if (dcb) {
-		if (dcb[0] >= 0x15)
-			i2c = ROMPTR(dev, dcb[2]);
-		if (dcb[0] >= 0x30)
-			i2c = ROMPTR(dev, dcb[4]);
-	}
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_i2c_chan *i2c;
+	int ret;
 
-	/* early revisions had no version number, use dcb version */
-	if (i2c) {
-		*version = dcb[0];
-		if (*version >= 0x30)
-			*version = i2c[0];
-	}
+	if (entry->chan)
+		return -EEXIST;
 
-	return i2c;
-}
+	if (dev_priv->card_type >= NV_50 && entry->read >= NV50_I2C_PORTS) {
+		NV_ERROR(dev, "unknown i2c port %d\n", entry->read);
+		return -EINVAL;
+	}
 
-int
-nouveau_i2c_init(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nvbios *bios = &dev_priv->vbios;
-	struct nouveau_i2c_chan *port;
-	u8 *i2c, *entry, legacy[2][4] = {};
-	u8 version, entries, recordlen;
-	int ret, i;
-
-	INIT_LIST_HEAD(&dev_priv->i2c_ports);
-
-	i2c = i2c_table(dev, &version);
-	if (!i2c) {
-		u8 *bmp = &bios->data[bios->offset];
-		if (bios->type != NVBIOS_BMP)
-			return -ENODEV;
-
-		legacy[0][0] = NV_CIO_CRE_DDC_WR__INDEX;
-		legacy[0][1] = NV_CIO_CRE_DDC_STATUS__INDEX;
-		legacy[1][0] = NV_CIO_CRE_DDC0_WR__INDEX;
-		legacy[1][1] = NV_CIO_CRE_DDC0_STATUS__INDEX;
-
-		/* BMP (from v4.0) has i2c info in the structure, it's in a
-		 * fixed location on earlier VBIOS
-		 */
-		if (bmp[5] < 4)
-			i2c = &bios->data[0x48];
-		else
-			i2c = &bmp[0x36];
-
-		if (i2c[4]) legacy[0][0] = i2c[4];
-		if (i2c[5]) legacy[0][1] = i2c[5];
-		if (i2c[6]) legacy[1][0] = i2c[6];
-		if (i2c[7]) legacy[1][1] = i2c[7];
+	i2c = kzalloc(sizeof(*i2c), GFP_KERNEL);
+	if (i2c == NULL)
+		return -ENOMEM;
+
+	switch (entry->port_type) {
+	case 0:
+		i2c->bit.setsda = nv04_i2c_setsda;
+		i2c->bit.setscl = nv04_i2c_setscl;
+		i2c->bit.getsda = nv04_i2c_getsda;
+		i2c->bit.getscl = nv04_i2c_getscl;
+		i2c->rd = entry->read;
+		i2c->wr = entry->write;
+		break;
+	case 4:
+		i2c->bit.setsda = nv4e_i2c_setsda;
+		i2c->bit.setscl = nv4e_i2c_setscl;
+		i2c->bit.getsda = nv4e_i2c_getsda;
+		i2c->bit.getscl = nv4e_i2c_getscl;
+		i2c->rd = 0x600800 + entry->read;
+		i2c->wr = 0x600800 + entry->write;
+		break;
+	case 5:
+		i2c->bit.setsda = nv50_i2c_setsda;
+		i2c->bit.setscl = nv50_i2c_setscl;
+		i2c->bit.getsda = nv50_i2c_getsda;
+		i2c->bit.getscl = nv50_i2c_getscl;
+		i2c->rd = nv50_i2c_port[entry->read];
+		i2c->wr = i2c->rd;
+		break;
+	case 6:
+		i2c->rd = entry->read;
+		i2c->wr = entry->write;
+		break;
+	default:
+		NV_ERROR(dev, "DCB I2C port type %d unknown\n",
+			 entry->port_type);
+		kfree(i2c);
+		return -EINVAL;
 	}
 
-	if (i2c && version >= 0x30) {
-		entry     = i2c[1] + i2c;
-		entries   = i2c[2];
-		recordlen = i2c[3];
-	} else
-	if (i2c) {
-		entry     = i2c;
-		entries   = 16;
-		recordlen = 4;
+	snprintf(i2c->adapter.name, sizeof(i2c->adapter.name),
+		 "nouveau-%s-%d", pci_name(dev->pdev), index);
+	i2c->adapter.owner = THIS_MODULE;
+	i2c->adapter.dev.parent = &dev->pdev->dev;
+	i2c->dev = dev;
+	i2c_set_adapdata(&i2c->adapter, i2c);
+
+	if (entry->port_type < 6) {
+		i2c->adapter.algo_data = &i2c->bit;
+		i2c->bit.udelay = 40;
+		i2c->bit.timeout = usecs_to_jiffies(5000);
+		i2c->bit.data = i2c;
+		ret = i2c_bit_add_bus(&i2c->adapter);
 	} else {
-		entry     = legacy[0];
-		entries   = 2;
-		recordlen = 4;
+		i2c->adapter.algo = &nouveau_dp_i2c_algo;
+		ret = i2c_add_adapter(&i2c->adapter);
 	}
 
-	for (i = 0; i < entries; i++, entry += recordlen) {
-		port = kzalloc(sizeof(*port), GFP_KERNEL);
-		if (port == NULL) {
-			nouveau_i2c_fini(dev);
-			return -ENOMEM;
-		}
-
-		port->type = entry[3];
-		if (version < 0x30) {
-			port->type &= 0x07;
-			if (port->type == 0x07)
-				port->type = 0xff;
-		}
-
-		if (port->type == 0xff) {
-			kfree(port);
-			continue;
-		}
-
-		switch (port->type) {
-		case 0: /* NV04:NV50 */
-			port->drive = entry[0];
-			port->sense = entry[1];
-			port->adapter.algo = &nouveau_i2c_bit_algo;
-			break;
-		case 4: /* NV4E */
-			port->drive = 0x600800 + entry[1];
-			port->sense = port->drive;
-			port->adapter.algo = &nouveau_i2c_bit_algo;
-			break;
-		case 5: /* NV50- */
-			port->drive = entry[0] & 0x0f;
-			if (dev_priv->card_type < NV_D0) {
-				if (port->drive >= ARRAY_SIZE(nv50_i2c_port))
-					break;
-				port->drive = nv50_i2c_port[port->drive];
-				port->sense = port->drive;
-			} else {
-				port->drive = 0x00d014 + (port->drive * 0x20);
-				port->sense = port->drive;
-			}
-			port->adapter.algo = &nouveau_i2c_bit_algo;
-			break;
-		case 6: /* NV50- DP AUX */
-			port->drive = entry[0];
-			port->sense = port->drive;
-			port->adapter.algo = &nouveau_dp_i2c_algo;
-			break;
-		default:
-			break;
-		}
-
-		if (!port->adapter.algo) {
-			NV_ERROR(dev, "I2C%d: type %d index %x/%x unknown\n",
-				 i, port->type, port->drive, port->sense);
-			kfree(port);
-			continue;
-		}
-
-		snprintf(port->adapter.name, sizeof(port->adapter.name),
-			 "nouveau-%s-%d", pci_name(dev->pdev), i);
-		port->adapter.owner = THIS_MODULE;
-		port->adapter.dev.parent = &dev->pdev->dev;
-		port->dev = dev;
-		port->index = i;
-		port->dcb = ROM32(entry[0]);
-		i2c_set_adapdata(&port->adapter, i2c);
-
-		ret = i2c_add_adapter(&port->adapter);
-		if (ret) {
-			NV_ERROR(dev, "I2C%d: failed register: %d\n", i, ret);
-			kfree(port);
-			continue;
-		}
-
-		list_add_tail(&port->head, &dev_priv->i2c_ports);
+	if (ret) {
+		NV_ERROR(dev, "Failed to register i2c %d\n", index);
+		kfree(i2c);
+		return ret;
 	}
 
+	entry->chan = i2c;
 	return 0;
 }
 
 void
-nouveau_i2c_fini(struct drm_device *dev)
+nouveau_i2c_fini(struct drm_device *dev, struct dcb_i2c_entry *entry)
 {
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_i2c_chan *port, *tmp;
+	if (!entry->chan)
+		return;
 
-	list_for_each_entry_safe(port, tmp, &dev_priv->i2c_ports, head) {
-		i2c_del_adapter(&port->adapter);
-		kfree(port);
-	}
+	i2c_del_adapter(&entry->chan->adapter);
+	kfree(entry->chan);
+	entry->chan = NULL;
 }
 
 struct nouveau_i2c_chan *
-nouveau_i2c_find(struct drm_device *dev, u8 index)
+nouveau_i2c_find(struct drm_device *dev, int index)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_i2c_chan *port;
-
-	if (index == NV_I2C_DEFAULT(0) ||
-	    index == NV_I2C_DEFAULT(1)) {
-		u8 version, *i2c = i2c_table(dev, &version);
-		if (i2c && version >= 0x30) {
-			if (index == NV_I2C_DEFAULT(0))
-				index = (i2c[4] & 0x0f);
-			else
-				index = (i2c[4] & 0xf0) >> 4;
-		} else {
-			index = 2;
-		}
-	}
-
-	list_for_each_entry(port, &dev_priv->i2c_ports, head) {
-		if (port->index == index)
-			break;
-	}
+	struct dcb_i2c_entry *i2c = &dev_priv->vbios.dcb.i2c[index];
 
-	if (&port->head == &dev_priv->i2c_ports)
+	if (index >= DCB_MAX_NUM_I2C_ENTRIES)
 		return NULL;
 
-	if (dev_priv->card_type >= NV_50 && (port->dcb & 0x00000100)) {
-		u32 reg = 0x00e500, val;
-		if (port->type == 6) {
-			reg += port->drive * 0x50;
+	if (dev_priv->card_type >= NV_50 && (i2c->entry & 0x00000100)) {
+		uint32_t reg = 0xe500, val;
+
+		if (i2c->port_type == 6) {
+			reg += i2c->read * 0x50;
 			val  = 0x2002;
 		} else {
-			reg += ((port->dcb & 0x1e00) >> 9) * 0x50;
+			reg += ((i2c->entry & 0x1e00) >> 9) * 0x50;
 			val  = 0xe001;
 		}
 
-		/* nfi, but neither auxch or i2c work if it's 1 */
-		nv_mask(dev, reg + 0x0c, 0x00000001, 0x00000000);
-		/* nfi, but switches auxch vs normal i2c */
-		nv_mask(dev, reg + 0x00, 0x0000f003, val);
+		nv_wr32(dev, reg, (nv_rd32(dev, reg) & ~0xf003) | val);
 	}
 
-	return port;
+	if (!i2c->chan && nouveau_i2c_init(dev, i2c, index))
+		return NULL;
+	return i2c->chan;
 }
 
 bool
@@ -532,12 +307,8 @@ nouveau_i2c_identify(struct drm_device *dev, const char *what,
 	struct nouveau_i2c_chan *i2c = nouveau_i2c_find(dev, index);
 	int i;
 
-	if (!i2c) {
-		NV_DEBUG(dev, "No bus when probing %s on %d\n", what, index);
-		return -ENODEV;
-	}
+	NV_DEBUG(dev, "Probing %ss on I2C bus: %d\n", what, index);
 
-	NV_DEBUG(dev, "Probing %ss on I2C bus: %d\n", what, i2c->index);
 	for (i = 0; info[i].addr; i++) {
 		if (nouveau_probe_i2c_addr(i2c, info[i].addr) &&
 		    (!match || match(i2c, &info[i]))) {
@@ -547,5 +318,6 @@ nouveau_i2c_identify(struct drm_device *dev, const char *what,
 	}
 
 	NV_DEBUG(dev, "No devices found.\n");
+
 	return -ENODEV;
 }
diff --git a/drivers/gpu/drm/nouveau/nouveau_i2c.h b/drivers/gpu/drm/nouveau/nouveau_i2c.h
index 4d2e4e9..422b62f 100644
--- a/drivers/gpu/drm/nouveau/nouveau_i2c.h
+++ b/drivers/gpu/drm/nouveau/nouveau_i2c.h
@@ -27,25 +27,20 @@
 #include <linux/i2c-algo-bit.h>
 #include "drm_dp_helper.h"
 
-#define NV_I2C_PORT(n)    (0x00 + (n))
-#define NV_I2C_PORT_NUM    0x10
-#define NV_I2C_DEFAULT(n) (0x80 + (n))
+struct dcb_i2c_entry;
 
 struct nouveau_i2c_chan {
 	struct i2c_adapter adapter;
 	struct drm_device *dev;
-	struct list_head head;
-	u8  index;
-	u8  type;
-	u32 dcb;
-	u32 drive;
-	u32 sense;
-	u32 state;
+	struct i2c_algo_bit_data bit;
+	unsigned rd;
+	unsigned wr;
+	unsigned data;
 };
 
-int  nouveau_i2c_init(struct drm_device *);
-void nouveau_i2c_fini(struct drm_device *);
-struct nouveau_i2c_chan *nouveau_i2c_find(struct drm_device *, u8 index);
+int nouveau_i2c_init(struct drm_device *, struct dcb_i2c_entry *, int index);
+void nouveau_i2c_fini(struct drm_device *, struct dcb_i2c_entry *);
+struct nouveau_i2c_chan *nouveau_i2c_find(struct drm_device *, int index);
 bool nouveau_probe_i2c_addr(struct nouveau_i2c_chan *i2c, int addr);
 int nouveau_i2c_identify(struct drm_device *dev, const char *what,
 			 struct i2c_board_info *info,
diff --git a/drivers/gpu/drm/nouveau/nouveau_irq.c b/drivers/gpu/drm/nouveau/nouveau_irq.c
index 868c7fd..2ba7265 100644
--- a/drivers/gpu/drm/nouveau/nouveau_irq.c
+++ b/drivers/gpu/drm/nouveau/nouveau_irq.c
@@ -79,7 +79,7 @@ nouveau_irq_handler(DRM_IRQ_ARGS)
 	int i;
 
 	stat = nv_rd32(dev, NV03_PMC_INTR_0);
-	if (stat == 0 || stat == ~0)
+	if (!stat)
 		return IRQ_NONE;
 
 	spin_lock_irqsave(&dev_priv->context_switch_lock, flags);
diff --git a/drivers/gpu/drm/nouveau/nouveau_mem.c b/drivers/gpu/drm/nouveau/nouveau_mem.c
index b08065f..5ee14d2 100644
--- a/drivers/gpu/drm/nouveau/nouveau_mem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_mem.c
@@ -26,8 +26,7 @@
  * DEALINGS IN THE SOFTWARE.
  *
  * Authors:
- *    Ben Skeggs <bskeggs@redhat.com>
- *    Roy Spliet <r.spliet@student.tudelft.nl>
+ *    Keith Whitwell <keith@tungstengraphics.com>
  */
 
 
@@ -193,6 +192,75 @@ nouveau_mem_gart_fini(struct drm_device *dev)
 	}
 }
 
+static uint32_t
+nouveau_mem_detect_nv04(struct drm_device *dev)
+{
+	uint32_t boot0 = nv_rd32(dev, NV04_PFB_BOOT_0);
+
+	if (boot0 & 0x00000100)
+		return (((boot0 >> 12) & 0xf) * 2 + 2) * 1024 * 1024;
+
+	switch (boot0 & NV04_PFB_BOOT_0_RAM_AMOUNT) {
+	case NV04_PFB_BOOT_0_RAM_AMOUNT_32MB:
+		return 32 * 1024 * 1024;
+	case NV04_PFB_BOOT_0_RAM_AMOUNT_16MB:
+		return 16 * 1024 * 1024;
+	case NV04_PFB_BOOT_0_RAM_AMOUNT_8MB:
+		return 8 * 1024 * 1024;
+	case NV04_PFB_BOOT_0_RAM_AMOUNT_4MB:
+		return 4 * 1024 * 1024;
+	}
+
+	return 0;
+}
+
+static uint32_t
+nouveau_mem_detect_nforce(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct pci_dev *bridge;
+	uint32_t mem;
+
+	bridge = pci_get_bus_and_slot(0, PCI_DEVFN(0, 1));
+	if (!bridge) {
+		NV_ERROR(dev, "no bridge device\n");
+		return 0;
+	}
+
+	if (dev_priv->flags & NV_NFORCE) {
+		pci_read_config_dword(bridge, 0x7C, &mem);
+		return (uint64_t)(((mem >> 6) & 31) + 1)*1024*1024;
+	} else
+	if (dev_priv->flags & NV_NFORCE2) {
+		pci_read_config_dword(bridge, 0x84, &mem);
+		return (uint64_t)(((mem >> 4) & 127) + 1)*1024*1024;
+	}
+
+	NV_ERROR(dev, "impossible!\n");
+	return 0;
+}
+
+int
+nouveau_mem_detect(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+
+	if (dev_priv->card_type == NV_04) {
+		dev_priv->vram_size = nouveau_mem_detect_nv04(dev);
+	} else
+	if (dev_priv->flags & (NV_NFORCE | NV_NFORCE2)) {
+		dev_priv->vram_size = nouveau_mem_detect_nforce(dev);
+	} else
+	if (dev_priv->card_type < NV_50) {
+		dev_priv->vram_size  = nv_rd32(dev, NV04_PFB_FIFO_DATA);
+		dev_priv->vram_size &= NV10_PFB_FIFO_DATA_RAM_AMOUNT_MB_MASK;
+	}
+
+	if (dev_priv->vram_size)
+		return 0;
+	return -ENOMEM;
+}
+
 bool
 nouveau_mem_flags_valid(struct drm_device *dev, u32 tile_flags)
 {
@@ -317,29 +385,11 @@ nouveau_mem_init_agp(struct drm_device *dev)
 	return 0;
 }
 
-static const struct vram_types {
-	int value;
-	const char *name;
-} vram_type_map[] = {
-	{ NV_MEM_TYPE_STOLEN , "stolen system memory" },
-	{ NV_MEM_TYPE_SGRAM  , "SGRAM" },
-	{ NV_MEM_TYPE_SDRAM  , "SDRAM" },
-	{ NV_MEM_TYPE_DDR1   , "DDR1" },
-	{ NV_MEM_TYPE_DDR2   , "DDR2" },
-	{ NV_MEM_TYPE_DDR3   , "DDR3" },
-	{ NV_MEM_TYPE_GDDR2  , "GDDR2" },
-	{ NV_MEM_TYPE_GDDR3  , "GDDR3" },
-	{ NV_MEM_TYPE_GDDR4  , "GDDR4" },
-	{ NV_MEM_TYPE_GDDR5  , "GDDR5" },
-	{ NV_MEM_TYPE_UNKNOWN, "unknown type" }
-};
-
 int
 nouveau_mem_vram_init(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct ttm_bo_device *bdev = &dev_priv->ttm.bdev;
-	const struct vram_types *vram_type;
 	int ret, dma_bits;
 
 	dma_bits = 32;
@@ -347,7 +397,7 @@ nouveau_mem_vram_init(struct drm_device *dev)
 		if (pci_dma_supported(dev->pdev, DMA_BIT_MASK(40)))
 			dma_bits = 40;
 	} else
-	if (0 && pci_is_pcie(dev->pdev) &&
+	if (0 && drm_pci_device_is_pcie(dev) &&
 	    dev_priv->chipset  > 0x40 &&
 	    dev_priv->chipset != 0x45) {
 		if (pci_dma_supported(dev->pdev, DMA_BIT_MASK(39)))
@@ -357,12 +407,8 @@ nouveau_mem_vram_init(struct drm_device *dev)
 	ret = pci_set_dma_mask(dev->pdev, DMA_BIT_MASK(dma_bits));
 	if (ret)
 		return ret;
-	ret = pci_set_consistent_dma_mask(dev->pdev, DMA_BIT_MASK(dma_bits));
-	if (ret) {
-		/* Reset to default value. */
-		pci_set_consistent_dma_mask(dev->pdev, DMA_BIT_MASK(32));
-	}
 
+	dev_priv->fb_phys = pci_resource_start(dev->pdev, 1);
 
 	ret = nouveau_ttm_global_init(dev_priv);
 	if (ret)
@@ -377,21 +423,39 @@ nouveau_mem_vram_init(struct drm_device *dev)
 		return ret;
 	}
 
-	vram_type = vram_type_map;
-	while (vram_type->value != NV_MEM_TYPE_UNKNOWN) {
-		if (nouveau_vram_type) {
-			if (!strcasecmp(nouveau_vram_type, vram_type->name))
-				break;
-			dev_priv->vram_type = vram_type->value;
-		} else {
-			if (vram_type->value == dev_priv->vram_type)
-				break;
-		}
-		vram_type++;
+	/* reserve space at end of VRAM for PRAMIN */
+	if (dev_priv->card_type >= NV_50) {
+		dev_priv->ramin_rsvd_vram = 1 * 1024 * 1024;
+	} else
+	if (dev_priv->card_type >= NV_40) {
+		u32 vs = hweight8((nv_rd32(dev, 0x001540) & 0x0000ff00) >> 8);
+		u32 rsvd;
+
+		/* estimate grctx size, the magics come from nv40_grctx.c */
+		if      (dev_priv->chipset == 0x40) rsvd = 0x6aa0 * vs;
+		else if (dev_priv->chipset  < 0x43) rsvd = 0x4f00 * vs;
+		else if (nv44_graph_class(dev))	    rsvd = 0x4980 * vs;
+		else				    rsvd = 0x4a40 * vs;
+		rsvd += 16 * 1024;
+		rsvd *= dev_priv->engine.fifo.channels;
+
+		/* pciegart table */
+		if (drm_pci_device_is_pcie(dev))
+			rsvd += 512 * 1024;
+
+		/* object storage */
+		rsvd += 512 * 1024;
+
+		dev_priv->ramin_rsvd_vram = round_up(rsvd, 4096);
+	} else {
+		dev_priv->ramin_rsvd_vram = 512 * 1024;
 	}
 
-	NV_INFO(dev, "Detected %dMiB VRAM (%s)\n",
-		(int)(dev_priv->vram_size >> 20), vram_type->name);
+	ret = dev_priv->engine.vram.init(dev);
+	if (ret)
+		return ret;
+
+	NV_INFO(dev, "Detected %dMiB VRAM\n", (int)(dev_priv->vram_size >> 20));
 	if (dev_priv->vram_sys_base) {
 		NV_INFO(dev, "Stolen system memory at: 0x%010llx\n",
 			dev_priv->vram_sys_base);
@@ -415,7 +479,7 @@ nouveau_mem_vram_init(struct drm_device *dev)
 	}
 
 	if (dev_priv->card_type < NV_50) {
-		ret = nouveau_bo_new(dev, 256*1024, 0, TTM_PL_FLAG_VRAM,
+		ret = nouveau_bo_new(dev, NULL, 256*1024, 0, TTM_PL_FLAG_VRAM,
 				     0, 0, &dev_priv->vga_ram);
 		if (ret == 0)
 			ret = nouveau_bo_pin(dev_priv->vga_ram,
@@ -472,645 +536,230 @@ nouveau_mem_gart_init(struct drm_device *dev)
 	return 0;
 }
 
-static int
-nv40_mem_timing_calc(struct drm_device *dev, u32 freq,
-		     struct nouveau_pm_tbl_entry *e, u8 len,
-		     struct nouveau_pm_memtiming *boot,
-		     struct nouveau_pm_memtiming *t)
-{
-	t->reg[0] = (e->tRP << 24 | e->tRAS << 16 | e->tRFC << 8 | e->tRC);
-
-	/* XXX: I don't trust the -1's and +1's... they must come
-	 *      from somewhere! */
-	t->reg[1] = (e->tWR + 2 + (t->tCWL - 1)) << 24 |
-		    1 << 16 |
-		    (e->tWTR + 2 + (t->tCWL - 1)) << 8 |
-		    (e->tCL + 2 - (t->tCWL - 1));
-
-	t->reg[2] = 0x20200000 |
-		    ((t->tCWL - 1) << 24 |
-		     e->tRRD << 16 |
-		     e->tRCDWR << 8 |
-		     e->tRCDRD);
-
-	NV_DEBUG(dev, "Entry %d: 220: %08x %08x %08x\n", t->id,
-		 t->reg[0], t->reg[1], t->reg[2]);
-	return 0;
-}
-
-static int
-nv50_mem_timing_calc(struct drm_device *dev, u32 freq,
-		     struct nouveau_pm_tbl_entry *e, u8 len,
-		     struct nouveau_pm_memtiming *boot,
-		     struct nouveau_pm_memtiming *t)
+void
+nouveau_mem_timing_init(struct drm_device *dev)
 {
+	/* cards < NVC0 only */
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	struct nouveau_pm_memtimings *memtimings = &pm->memtimings;
+	struct nvbios *bios = &dev_priv->vbios;
 	struct bit_entry P;
-	uint8_t unk18 = 1, unk20 = 0, unk21 = 0, tmp7_3;
-
-	if (bit_table(dev, 'P', &P))
-		return -EINVAL;
-
-	switch (min(len, (u8) 22)) {
-	case 22:
-		unk21 = e->tUNK_21;
-	case 21:
-		unk20 = e->tUNK_20;
-	case 20:
-		if (e->tCWL > 0)
-			t->tCWL = e->tCWL;
-	case 19:
-		unk18 = e->tUNK_18;
-		break;
-	}
-
-	t->reg[0] = (e->tRP << 24 | e->tRAS << 16 | e->tRFC << 8 | e->tRC);
-
-	t->reg[1] = (e->tWR + 2 + (t->tCWL - 1)) << 24 |
-				max(unk18, (u8) 1) << 16 |
-				(e->tWTR + 2 + (t->tCWL - 1)) << 8;
-
-	t->reg[2] = ((t->tCWL - 1) << 24 |
-		    e->tRRD << 16 |
-		    e->tRCDWR << 8 |
-		    e->tRCDRD);
-
-	t->reg[4] = e->tUNK_13 << 8  | e->tUNK_13;
-
-	t->reg[5] = (e->tRFC << 24 | max(e->tRCDRD, e->tRCDWR) << 16 | e->tRP);
-
-	t->reg[8] = boot->reg[8] & 0xffffff00;
-
-	if (P.version == 1) {
-		t->reg[1] |= (e->tCL + 2 - (t->tCWL - 1));
-
-		t->reg[3] = (0x14 + e->tCL) << 24 |
-			    0x16 << 16 |
-			    (e->tCL - 1) << 8 |
-			    (e->tCL - 1);
-
-		t->reg[4] |= boot->reg[4] & 0xffff0000;
-
-		t->reg[6] = (0x33 - t->tCWL) << 16 |
-			    t->tCWL << 8 |
-			    (0x2e + e->tCL - t->tCWL);
-
-		t->reg[7] = 0x4000202 | (e->tCL - 1) << 16;
-
-		/* XXX: P.version == 1 only has DDR2 and GDDR3? */
-		if (dev_priv->vram_type == NV_MEM_TYPE_DDR2) {
-			t->reg[5] |= (e->tCL + 3) << 8;
-			t->reg[6] |= (t->tCWL - 2) << 8;
-			t->reg[8] |= (e->tCL - 4);
-		} else {
-			t->reg[5] |= (e->tCL + 2) << 8;
-			t->reg[6] |= t->tCWL << 8;
-			t->reg[8] |= (e->tCL - 2);
+	u8 tUNK_0, tUNK_1, tUNK_2;
+	u8 tRP;		/* Byte 3 */
+	u8 tRAS;	/* Byte 5 */
+	u8 tRFC;	/* Byte 7 */
+	u8 tRC;		/* Byte 9 */
+	u8 tUNK_10, tUNK_11, tUNK_12, tUNK_13, tUNK_14;
+	u8 tUNK_18, tUNK_19, tUNK_20, tUNK_21;
+	u8 magic_number = 0; /* Yeah... sorry*/
+	u8 *mem = NULL, *entry;
+	int i, recordlen, entries;
+
+	if (bios->type == NVBIOS_BIT) {
+		if (bit_table(dev, 'P', &P))
+			return;
+
+		if (P.version == 1)
+			mem = ROMPTR(bios, P.data[4]);
+		else
+		if (P.version == 2)
+			mem = ROMPTR(bios, P.data[8]);
+		else {
+			NV_WARN(dev, "unknown mem for BIT P %d\n", P.version);
 		}
 	} else {
-		t->reg[1] |= (5 + e->tCL - (t->tCWL));
-
-		/* XXX: 0xb? 0x30? */
-		t->reg[3] = (0x30 + e->tCL) << 24 |
-			    (boot->reg[3] & 0x00ff0000)|
-			    (0xb + e->tCL) << 8 |
-			    (e->tCL - 1);
-
-		t->reg[4] |= (unk20 << 24 | unk21 << 16);
-
-		/* XXX: +6? */
-		t->reg[5] |= (t->tCWL + 6) << 8;
-
-		t->reg[6] = (0x5a + e->tCL) << 16 |
-			    (6 - e->tCL + t->tCWL) << 8 |
-			    (0x50 + e->tCL - t->tCWL);
-
-		tmp7_3 = (boot->reg[7] & 0xff000000) >> 24;
-		t->reg[7] = (tmp7_3 << 24) |
-			    ((tmp7_3 - 6 + e->tCL) << 16) |
-			    0x202;
-	}
-
-	NV_DEBUG(dev, "Entry %d: 220: %08x %08x %08x %08x\n", t->id,
-		 t->reg[0], t->reg[1], t->reg[2], t->reg[3]);
-	NV_DEBUG(dev, "         230: %08x %08x %08x %08x\n",
-		 t->reg[4], t->reg[5], t->reg[6], t->reg[7]);
-	NV_DEBUG(dev, "         240: %08x\n", t->reg[8]);
-	return 0;
-}
-
-static int
-nvc0_mem_timing_calc(struct drm_device *dev, u32 freq,
-		     struct nouveau_pm_tbl_entry *e, u8 len,
-		     struct nouveau_pm_memtiming *boot,
-		     struct nouveau_pm_memtiming *t)
-{
-	if (e->tCWL > 0)
-		t->tCWL = e->tCWL;
-
-	t->reg[0] = (e->tRP << 24 | (e->tRAS & 0x7f) << 17 |
-		     e->tRFC << 8 | e->tRC);
-
-	t->reg[1] = (boot->reg[1] & 0xff000000) |
-		    (e->tRCDWR & 0x0f) << 20 |
-		    (e->tRCDRD & 0x0f) << 14 |
-		    (t->tCWL << 7) |
-		    (e->tCL & 0x0f);
-
-	t->reg[2] = (boot->reg[2] & 0xff0000ff) |
-		    e->tWR << 16 | e->tWTR << 8;
-
-	t->reg[3] = (e->tUNK_20 & 0x1f) << 9 |
-		    (e->tUNK_21 & 0xf) << 5 |
-		    (e->tUNK_13 & 0x1f);
-
-	t->reg[4] = (boot->reg[4] & 0xfff00fff) |
-		    (e->tRRD&0x1f) << 15;
-
-	NV_DEBUG(dev, "Entry %d: 290: %08x %08x %08x %08x\n", t->id,
-		 t->reg[0], t->reg[1], t->reg[2], t->reg[3]);
-	NV_DEBUG(dev, "         2a0: %08x\n", t->reg[4]);
-	return 0;
-}
-
-/**
- * MR generation methods
- */
-
-static int
-nouveau_mem_ddr2_mr(struct drm_device *dev, u32 freq,
-		    struct nouveau_pm_tbl_entry *e, u8 len,
-		    struct nouveau_pm_memtiming *boot,
-		    struct nouveau_pm_memtiming *t)
-{
-	t->drive_strength = 0;
-	if (len < 15) {
-		t->odt = boot->odt;
-	} else {
-		t->odt = e->RAM_FT1 & 0x07;
-	}
-
-	if (e->tCL >= NV_MEM_CL_DDR2_MAX) {
-		NV_WARN(dev, "(%u) Invalid tCL: %u", t->id, e->tCL);
-		return -ERANGE;
-	}
-
-	if (e->tWR >= NV_MEM_WR_DDR2_MAX) {
-		NV_WARN(dev, "(%u) Invalid tWR: %u", t->id, e->tWR);
-		return -ERANGE;
-	}
-
-	if (t->odt > 3) {
-		NV_WARN(dev, "(%u) Invalid odt value, assuming disabled: %x",
-			t->id, t->odt);
-		t->odt = 0;
-	}
-
-	t->mr[0] = (boot->mr[0] & 0x100f) |
-		   (e->tCL) << 4 |
-		   (e->tWR - 1) << 9;
-	t->mr[1] = (boot->mr[1] & 0x101fbb) |
-		   (t->odt & 0x1) << 2 |
-		   (t->odt & 0x2) << 5;
-
-	NV_DEBUG(dev, "(%u) MR: %08x", t->id, t->mr[0]);
-	return 0;
-}
-
-uint8_t nv_mem_wr_lut_ddr3[NV_MEM_WR_DDR3_MAX] = {
-	0, 0, 0, 0, 0, 1, 2, 3, 4, 5, 5, 6, 6, 7, 7, 0, 0};
-
-static int
-nouveau_mem_ddr3_mr(struct drm_device *dev, u32 freq,
-		    struct nouveau_pm_tbl_entry *e, u8 len,
-		    struct nouveau_pm_memtiming *boot,
-		    struct nouveau_pm_memtiming *t)
-{
-	u8 cl = e->tCL - 4;
-
-	t->drive_strength = 0;
-	if (len < 15) {
-		t->odt = boot->odt;
-	} else {
-		t->odt = e->RAM_FT1 & 0x07;
-	}
-
-	if (e->tCL >= NV_MEM_CL_DDR3_MAX || e->tCL < 4) {
-		NV_WARN(dev, "(%u) Invalid tCL: %u", t->id, e->tCL);
-		return -ERANGE;
-	}
-
-	if (e->tWR >= NV_MEM_WR_DDR3_MAX || e->tWR < 4) {
-		NV_WARN(dev, "(%u) Invalid tWR: %u", t->id, e->tWR);
-		return -ERANGE;
-	}
-
-	if (e->tCWL < 5) {
-		NV_WARN(dev, "(%u) Invalid tCWL: %u", t->id, e->tCWL);
-		return -ERANGE;
-	}
-
-	t->mr[0] = (boot->mr[0] & 0x180b) |
-		   /* CAS */
-		   (cl & 0x7) << 4 |
-		   (cl & 0x8) >> 1 |
-		   (nv_mem_wr_lut_ddr3[e->tWR]) << 9;
-	t->mr[1] = (boot->mr[1] & 0x101dbb) |
-		   (t->odt & 0x1) << 2 |
-		   (t->odt & 0x2) << 5 |
-		   (t->odt & 0x4) << 7;
-	t->mr[2] = (boot->mr[2] & 0x20ffb7) | (e->tCWL - 5) << 3;
-
-	NV_DEBUG(dev, "(%u) MR: %08x %08x", t->id, t->mr[0], t->mr[2]);
-	return 0;
-}
-
-uint8_t nv_mem_cl_lut_gddr3[NV_MEM_CL_GDDR3_MAX] = {
-	0, 0, 0, 0, 4, 5, 6, 7, 0, 1, 2, 3, 8, 9, 10, 11};
-uint8_t nv_mem_wr_lut_gddr3[NV_MEM_WR_GDDR3_MAX] = {
-	0, 0, 0, 0, 0, 2, 3, 8, 9, 10, 11, 0, 0, 1, 1, 0, 3};
-
-static int
-nouveau_mem_gddr3_mr(struct drm_device *dev, u32 freq,
-		     struct nouveau_pm_tbl_entry *e, u8 len,
-		     struct nouveau_pm_memtiming *boot,
-		     struct nouveau_pm_memtiming *t)
-{
-	if (len < 15) {
-		t->drive_strength = boot->drive_strength;
-		t->odt = boot->odt;
-	} else {
-		t->drive_strength = (e->RAM_FT1 & 0x30) >> 4;
-		t->odt = e->RAM_FT1 & 0x07;
+		NV_DEBUG(dev, "BMP version too old for memory\n");
+		return;
 	}
 
-	if (e->tCL >= NV_MEM_CL_GDDR3_MAX) {
-		NV_WARN(dev, "(%u) Invalid tCL: %u", t->id, e->tCL);
-		return -ERANGE;
+	if (!mem) {
+		NV_DEBUG(dev, "memory timing table pointer invalid\n");
+		return;
 	}
 
-	if (e->tWR >= NV_MEM_WR_GDDR3_MAX) {
-		NV_WARN(dev, "(%u) Invalid tWR: %u", t->id, e->tWR);
-		return -ERANGE;
+	if (mem[0] != 0x10) {
+		NV_WARN(dev, "memory timing table 0x%02x unknown\n", mem[0]);
+		return;
 	}
 
-	if (t->odt > 3) {
-		NV_WARN(dev, "(%u) Invalid odt value, assuming autocal: %x",
-			t->id, t->odt);
-		t->odt = 0;
+	/* validate record length */
+	entries   = mem[2];
+	recordlen = mem[3];
+	if (recordlen < 15) {
+		NV_ERROR(dev, "mem timing table length unknown: %d\n", mem[3]);
+		return;
 	}
 
-	t->mr[0] = (boot->mr[0] & 0xe0b) |
-		   /* CAS */
-		   ((nv_mem_cl_lut_gddr3[e->tCL] & 0x7) << 4) |
-		   ((nv_mem_cl_lut_gddr3[e->tCL] & 0x8) >> 2);
-	t->mr[1] = (boot->mr[1] & 0x100f40) | t->drive_strength |
-		   (t->odt << 2) |
-		   (nv_mem_wr_lut_gddr3[e->tWR] & 0xf) << 4;
-	t->mr[2] = boot->mr[2];
-
-	NV_DEBUG(dev, "(%u) MR: %08x %08x %08x", t->id,
-		      t->mr[0], t->mr[1], t->mr[2]);
-	return 0;
-}
-
-static int
-nouveau_mem_gddr5_mr(struct drm_device *dev, u32 freq,
-		     struct nouveau_pm_tbl_entry *e, u8 len,
-		     struct nouveau_pm_memtiming *boot,
-		     struct nouveau_pm_memtiming *t)
-{
-	if (len < 15) {
-		t->drive_strength = boot->drive_strength;
-		t->odt = boot->odt;
-	} else {
-		t->drive_strength = (e->RAM_FT1 & 0x30) >> 4;
-		t->odt = e->RAM_FT1 & 0x03;
-	}
+	/* parse vbios entries into common format */
+	memtimings->timing =
+		kcalloc(entries, sizeof(*memtimings->timing), GFP_KERNEL);
+	if (!memtimings->timing)
+		return;
 
-	if (e->tCL >= NV_MEM_CL_GDDR5_MAX) {
-		NV_WARN(dev, "(%u) Invalid tCL: %u", t->id, e->tCL);
-		return -ERANGE;
+	/* Get "some number" from the timing reg for NV_40 and NV_50
+	 * Used in calculations later */
+	if (dev_priv->card_type >= NV_40 && dev_priv->chipset < 0x98) {
+		magic_number = (nv_rd32(dev, 0x100228) & 0x0f000000) >> 24;
 	}
 
-	if (e->tWR >= NV_MEM_WR_GDDR5_MAX) {
-		NV_WARN(dev, "(%u) Invalid tWR: %u", t->id, e->tWR);
-		return -ERANGE;
-	}
+	entry = mem + mem[1];
+	for (i = 0; i < entries; i++, entry += recordlen) {
+		struct nouveau_pm_memtiming *timing = &pm->memtimings.timing[i];
+		if (entry[0] == 0)
+			continue;
 
-	if (t->odt > 3) {
-		NV_WARN(dev, "(%u) Invalid odt value, assuming autocal: %x",
-			t->id, t->odt);
-		t->odt = 0;
-	}
+		tUNK_18 = 1;
+		tUNK_19 = 1;
+		tUNK_20 = 0;
+		tUNK_21 = 0;
+		switch (min(recordlen, 22)) {
+		case 22:
+			tUNK_21 = entry[21];
+		case 21:
+			tUNK_20 = entry[20];
+		case 20:
+			tUNK_19 = entry[19];
+		case 19:
+			tUNK_18 = entry[18];
+		default:
+			tUNK_0  = entry[0];
+			tUNK_1  = entry[1];
+			tUNK_2  = entry[2];
+			tRP     = entry[3];
+			tRAS    = entry[5];
+			tRFC    = entry[7];
+			tRC     = entry[9];
+			tUNK_10 = entry[10];
+			tUNK_11 = entry[11];
+			tUNK_12 = entry[12];
+			tUNK_13 = entry[13];
+			tUNK_14 = entry[14];
+			break;
+		}
 
-	t->mr[0] = (boot->mr[0] & 0x007) |
-		   ((e->tCL - 5) << 3) |
-		   ((e->tWR - 4) << 8);
-	t->mr[1] = (boot->mr[1] & 0x1007f0) |
-		   t->drive_strength |
-		   (t->odt << 2);
+		timing->reg_100220 = (tRC << 24 | tRFC << 16 | tRAS << 8 | tRP);
 
-	NV_DEBUG(dev, "(%u) MR: %08x %08x", t->id, t->mr[0], t->mr[1]);
-	return 0;
-}
+		/* XXX: I don't trust the -1's and +1's... they must come
+		 *      from somewhere! */
+		timing->reg_100224 = (tUNK_0 + tUNK_19 + 1 + magic_number) << 24 |
+				      max(tUNK_18, (u8) 1) << 16 |
+				      (tUNK_1 + tUNK_19 + 1 + magic_number) << 8;
+		if (dev_priv->chipset == 0xa8) {
+			timing->reg_100224 |= (tUNK_2 - 1);
+		} else {
+			timing->reg_100224 |= (tUNK_2 + 2 - magic_number);
+		}
 
-int
-nouveau_mem_timing_calc(struct drm_device *dev, u32 freq,
-			struct nouveau_pm_memtiming *t)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
-	struct nouveau_pm_memtiming *boot = &pm->boot.timing;
-	struct nouveau_pm_tbl_entry *e;
-	u8 ver, len, *ptr, *ramcfg;
-	int ret;
+		timing->reg_100228 = (tUNK_12 << 16 | tUNK_11 << 8 | tUNK_10);
+		if (dev_priv->chipset >= 0xa3 && dev_priv->chipset < 0xaa)
+			timing->reg_100228 |= (tUNK_19 - 1) << 24;
+		else
+			timing->reg_100228 |= magic_number << 24;
+
+		if (dev_priv->card_type == NV_40) {
+			/* NV40: don't know what the rest of the regs are..
+			 * And don't need to know either */
+			timing->reg_100228 |= 0x20200000;
+		} else if (dev_priv->card_type >= NV_50) {
+			if (dev_priv->chipset < 0x98 ||
+			    (dev_priv->chipset == 0x98 &&
+			     dev_priv->stepping <= 0xa1)) {
+				timing->reg_10022c = (0x14 + tUNK_2) << 24 |
+						     0x16 << 16 |
+						     (tUNK_2 - 1) << 8 |
+						     (tUNK_2 - 1);
+			} else {
+				/* XXX: reg_10022c for recentish cards */
+				timing->reg_10022c = tUNK_2 - 1;
+			}
 
-	ptr = nouveau_perf_timing(dev, freq, &ver, &len);
-	if (!ptr || ptr[0] == 0x00) {
-		*t = *boot;
-		return 0;
-	}
-	e = (struct nouveau_pm_tbl_entry *)ptr;
-
-	t->tCWL = boot->tCWL;
-
-	switch (dev_priv->card_type) {
-	case NV_40:
-		ret = nv40_mem_timing_calc(dev, freq, e, len, boot, t);
-		break;
-	case NV_50:
-		ret = nv50_mem_timing_calc(dev, freq, e, len, boot, t);
-		break;
-	case NV_C0:
-		ret = nvc0_mem_timing_calc(dev, freq, e, len, boot, t);
-		break;
-	default:
-		ret = -ENODEV;
-		break;
-	}
+			timing->reg_100230 = (tUNK_20 << 24 | tUNK_21 << 16 |
+						  tUNK_13 << 8  | tUNK_13);
 
-	switch (dev_priv->vram_type * !ret) {
-	case NV_MEM_TYPE_GDDR3:
-		ret = nouveau_mem_gddr3_mr(dev, freq, e, len, boot, t);
-		break;
-	case NV_MEM_TYPE_GDDR5:
-		ret = nouveau_mem_gddr5_mr(dev, freq, e, len, boot, t);
-		break;
-	case NV_MEM_TYPE_DDR2:
-		ret = nouveau_mem_ddr2_mr(dev, freq, e, len, boot, t);
-		break;
-	case NV_MEM_TYPE_DDR3:
-		ret = nouveau_mem_ddr3_mr(dev, freq, e, len, boot, t);
-		break;
-	default:
-		ret = -EINVAL;
-		break;
-	}
+			timing->reg_100234 = (tRAS << 24 | tRC);
+			timing->reg_100234 += max(tUNK_10, tUNK_11) << 16;
 
-	ramcfg = nouveau_perf_ramcfg(dev, freq, &ver, &len);
-	if (ramcfg) {
-		int dll_off;
+			if (dev_priv->chipset < 0x98 ||
+			    (dev_priv->chipset == 0x98 &&
+			     dev_priv->stepping <= 0xa1)) {
+				timing->reg_100234 |= (tUNK_2 + 2) << 8;
+			} else {
+				/* XXX: +6? */
+				timing->reg_100234 |= (tUNK_19 + 6) << 8;
+			}
 
-		if (ver == 0x00)
-			dll_off = !!(ramcfg[3] & 0x04);
-		else
-			dll_off = !!(ramcfg[2] & 0x40);
+			/* XXX; reg_100238
+			 * reg_100238: 0x00?????? */
+			timing->reg_10023c = 0x202;
+			if (dev_priv->chipset < 0x98 ||
+			    (dev_priv->chipset == 0x98 &&
+			     dev_priv->stepping <= 0xa1)) {
+				timing->reg_10023c |= 0x4000000 | (tUNK_2 - 1) << 16;
+			} else {
+				/* XXX: reg_10023c
+				 * currently unknown
+				 * 10023c seen as 06xxxxxx, 0bxxxxxx or 0fxxxxxx */
+			}
 
-		switch (dev_priv->vram_type) {
-		case NV_MEM_TYPE_GDDR3:
-			t->mr[1] &= ~0x00000040;
-			t->mr[1] |=  0x00000040 * dll_off;
-			break;
-		default:
-			t->mr[1] &= ~0x00000001;
-			t->mr[1] |=  0x00000001 * dll_off;
-			break;
+			/* XXX: reg_100240? */
 		}
+		timing->id = i;
+
+		NV_DEBUG(dev, "Entry %d: 220: %08x %08x %08x %08x\n", i,
+			 timing->reg_100220, timing->reg_100224,
+			 timing->reg_100228, timing->reg_10022c);
+		NV_DEBUG(dev, "         230: %08x %08x %08x %08x\n",
+			 timing->reg_100230, timing->reg_100234,
+			 timing->reg_100238, timing->reg_10023c);
+		NV_DEBUG(dev, "         240: %08x\n", timing->reg_100240);
 	}
 
-	return ret;
+	memtimings->nr_timing = entries;
+	memtimings->supported = (dev_priv->chipset <= 0x98);
 }
 
 void
-nouveau_mem_timing_read(struct drm_device *dev, struct nouveau_pm_memtiming *t)
+nouveau_mem_timing_fini(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	u32 timing_base, timing_regs, mr_base;
-	int i;
-
-	if (dev_priv->card_type >= 0xC0) {
-		timing_base = 0x10f290;
-		mr_base = 0x10f300;
-	} else {
-		timing_base = 0x100220;
-		mr_base = 0x1002c0;
-	}
+	struct nouveau_pm_memtimings *mem = &dev_priv->engine.pm.memtimings;
 
-	t->id = -1;
-
-	switch (dev_priv->card_type) {
-	case NV_50:
-		timing_regs = 9;
-		break;
-	case NV_C0:
-	case NV_D0:
-		timing_regs = 5;
-		break;
-	case NV_30:
-	case NV_40:
-		timing_regs = 3;
-		break;
-	default:
-		timing_regs = 0;
-		return;
-	}
-	for(i = 0; i < timing_regs; i++)
-		t->reg[i] = nv_rd32(dev, timing_base + (0x04 * i));
-
-	t->tCWL = 0;
-	if (dev_priv->card_type < NV_C0) {
-		t->tCWL = ((nv_rd32(dev, 0x100228) & 0x0f000000) >> 24) + 1;
-	} else if (dev_priv->card_type <= NV_D0) {
-		t->tCWL = ((nv_rd32(dev, 0x10f294) & 0x00000f80) >> 7);
-	}
-
-	t->mr[0] = nv_rd32(dev, mr_base);
-	t->mr[1] = nv_rd32(dev, mr_base + 0x04);
-	t->mr[2] = nv_rd32(dev, mr_base + 0x20);
-	t->mr[3] = nv_rd32(dev, mr_base + 0x24);
-
-	t->odt = 0;
-	t->drive_strength = 0;
-
-	switch (dev_priv->vram_type) {
-	case NV_MEM_TYPE_DDR3:
-		t->odt |= (t->mr[1] & 0x200) >> 7;
-	case NV_MEM_TYPE_DDR2:
-		t->odt |= (t->mr[1] & 0x04) >> 2 |
-			  (t->mr[1] & 0x40) >> 5;
-		break;
-	case NV_MEM_TYPE_GDDR3:
-	case NV_MEM_TYPE_GDDR5:
-		t->drive_strength = t->mr[1] & 0x03;
-		t->odt = (t->mr[1] & 0x0c) >> 2;
-		break;
-	default:
-		break;
-	}
+	kfree(mem->timing);
 }
 
-int
-nouveau_mem_exec(struct nouveau_mem_exec_func *exec,
-		 struct nouveau_pm_level *perflvl)
+static int
+nouveau_vram_manager_init(struct ttm_mem_type_manager *man, unsigned long p_size)
 {
-	struct drm_nouveau_private *dev_priv = exec->dev->dev_private;
-	struct nouveau_pm_memtiming *info = &perflvl->timing;
-	u32 tMRD = 1000, tCKSRE = 0, tCKSRX = 0, tXS = 0, tDLLK = 0;
-	u32 mr[3] = { info->mr[0], info->mr[1], info->mr[2] };
-	u32 mr1_dlloff;
-
-	switch (dev_priv->vram_type) {
-	case NV_MEM_TYPE_DDR2:
-		tDLLK = 2000;
-		mr1_dlloff = 0x00000001;
-		break;
-	case NV_MEM_TYPE_DDR3:
-		tDLLK = 12000;
-		mr1_dlloff = 0x00000001;
-		break;
-	case NV_MEM_TYPE_GDDR3:
-		tDLLK = 40000;
-		mr1_dlloff = 0x00000040;
-		break;
-	default:
-		NV_ERROR(exec->dev, "cannot reclock unsupported memtype\n");
-		return -ENODEV;
-	}
-
-	/* fetch current MRs */
-	switch (dev_priv->vram_type) {
-	case NV_MEM_TYPE_GDDR3:
-	case NV_MEM_TYPE_DDR3:
-		mr[2] = exec->mrg(exec, 2);
-	default:
-		mr[1] = exec->mrg(exec, 1);
-		mr[0] = exec->mrg(exec, 0);
-		break;
-	}
-
-	/* DLL 'on' -> DLL 'off' mode, disable before entering self-refresh  */
-	if (!(mr[1] & mr1_dlloff) && (info->mr[1] & mr1_dlloff)) {
-		exec->precharge(exec);
-		exec->mrs (exec, 1, mr[1] | mr1_dlloff);
-		exec->wait(exec, tMRD);
-	}
-
-	/* enter self-refresh mode */
-	exec->precharge(exec);
-	exec->refresh(exec);
-	exec->refresh(exec);
-	exec->refresh_auto(exec, false);
-	exec->refresh_self(exec, true);
-	exec->wait(exec, tCKSRE);
-
-	/* modify input clock frequency */
-	exec->clock_set(exec);
-
-	/* exit self-refresh mode */
-	exec->wait(exec, tCKSRX);
-	exec->precharge(exec);
-	exec->refresh_self(exec, false);
-	exec->refresh_auto(exec, true);
-	exec->wait(exec, tXS);
-
-	/* update MRs */
-	if (mr[2] != info->mr[2]) {
-		exec->mrs (exec, 2, info->mr[2]);
-		exec->wait(exec, tMRD);
-	}
-
-	if (mr[1] != info->mr[1]) {
-		/* need to keep DLL off until later, at least on GDDR3 */
-		exec->mrs (exec, 1, info->mr[1] | (mr[1] & mr1_dlloff));
-		exec->wait(exec, tMRD);
-	}
-
-	if (mr[0] != info->mr[0]) {
-		exec->mrs (exec, 0, info->mr[0]);
-		exec->wait(exec, tMRD);
-	}
-
-	/* update PFB timing registers */
-	exec->timing_set(exec);
-
-	/* DLL (enable + ) reset */
-	if (!(info->mr[1] & mr1_dlloff)) {
-		if (mr[1] & mr1_dlloff) {
-			exec->mrs (exec, 1, info->mr[1]);
-			exec->wait(exec, tMRD);
-		}
-		exec->mrs (exec, 0, info->mr[0] | 0x00000100);
-		exec->wait(exec, tMRD);
-		exec->mrs (exec, 0, info->mr[0] | 0x00000000);
-		exec->wait(exec, tMRD);
-		exec->wait(exec, tDLLK);
-		if (dev_priv->vram_type == NV_MEM_TYPE_GDDR3)
-			exec->precharge(exec);
-	}
-
-	return 0;
-}
+	struct drm_nouveau_private *dev_priv = nouveau_bdev(man->bdev);
+	struct nouveau_mm *mm;
+	u64 size, block, rsvd;
+	int ret;
 
-int
-nouveau_mem_vbios_type(struct drm_device *dev)
-{
-	struct bit_entry M;
-	u8 ramcfg = (nv_rd32(dev, 0x101000) & 0x0000003c) >> 2;
-	if (!bit_table(dev, 'M', &M) || M.version != 2 || M.length < 5) {
-		u8 *table = ROMPTR(dev, M.data[3]);
-		if (table && table[0] == 0x10 && ramcfg < table[3]) {
-			u8 *entry = table + table[1] + (ramcfg * table[2]);
-			switch (entry[0] & 0x0f) {
-			case 0: return NV_MEM_TYPE_DDR2;
-			case 1: return NV_MEM_TYPE_DDR3;
-			case 2: return NV_MEM_TYPE_GDDR3;
-			case 3: return NV_MEM_TYPE_GDDR5;
-			default:
-				break;
-			}
+	rsvd  = (256 * 1024); /* vga memory */
+	size  = (p_size << PAGE_SHIFT) - rsvd;
+	block = dev_priv->vram_rblock_size;
 
-		}
-	}
-	return NV_MEM_TYPE_UNKNOWN;
-}
+	ret = nouveau_mm_init(&mm, rsvd >> 12, size >> 12, block >> 12);
+	if (ret)
+		return ret;
 
-static int
-nouveau_vram_manager_init(struct ttm_mem_type_manager *man, unsigned long psize)
-{
-	/* nothing to do */
+	man->priv = mm;
 	return 0;
 }
 
 static int
 nouveau_vram_manager_fini(struct ttm_mem_type_manager *man)
 {
-	/* nothing to do */
-	return 0;
-}
+	struct nouveau_mm *mm = man->priv;
+	int ret;
 
-static inline void
-nouveau_mem_node_cleanup(struct nouveau_mem *node)
-{
-	if (node->vma[0].node) {
-		nouveau_vm_unmap(&node->vma[0]);
-		nouveau_vm_put(&node->vma[0]);
-	}
+	ret = nouveau_mm_fini(&mm);
+	if (ret)
+		return ret;
 
-	if (node->vma[1].node) {
-		nouveau_vm_unmap(&node->vma[1]);
-		nouveau_vm_put(&node->vma[1]);
-	}
+	man->priv = NULL;
+	return 0;
 }
 
 static void
@@ -1119,9 +768,14 @@ nouveau_vram_manager_del(struct ttm_mem_type_manager *man,
 {
 	struct drm_nouveau_private *dev_priv = nouveau_bdev(man->bdev);
 	struct nouveau_vram_engine *vram = &dev_priv->engine.vram;
+	struct nouveau_mem *node = mem->mm_node;
 	struct drm_device *dev = dev_priv->dev;
 
-	nouveau_mem_node_cleanup(mem->mm_node);
+	if (node->tmp_vma.node) {
+		nouveau_vm_unmap(&node->tmp_vma);
+		nouveau_vm_put(&node->tmp_vma);
+	}
+
 	vram->put(dev, (struct nouveau_mem **)&mem->mm_node);
 }
 
@@ -1140,7 +794,7 @@ nouveau_vram_manager_new(struct ttm_mem_type_manager *man,
 	int ret;
 
 	if (nvbo->tile_flags & NOUVEAU_GEM_TILE_NONCONTIG)
-		size_nc = 1 << nvbo->page_shift;
+		size_nc = 1 << nvbo->vma.node->type;
 
 	ret = vram->get(dev, mem->num_pages << PAGE_SHIFT,
 			mem->page_alignment << PAGE_SHIFT, size_nc,
@@ -1150,7 +804,9 @@ nouveau_vram_manager_new(struct ttm_mem_type_manager *man,
 		return (ret == -ENOSPC) ? 0 : ret;
 	}
 
-	node->page_shift = nvbo->page_shift;
+	node->page_shift = 12;
+	if (nvbo->vma.node)
+		node->page_shift = nvbo->vma.node->type;
 
 	mem->mm_node = node;
 	mem->start   = node->offset >> PAGE_SHIFT;
@@ -1206,9 +862,15 @@ static void
 nouveau_gart_manager_del(struct ttm_mem_type_manager *man,
 			 struct ttm_mem_reg *mem)
 {
-	nouveau_mem_node_cleanup(mem->mm_node);
-	kfree(mem->mm_node);
+	struct nouveau_mem *node = mem->mm_node;
+
+	if (node->tmp_vma.node) {
+		nouveau_vm_unmap(&node->tmp_vma);
+		nouveau_vm_put(&node->tmp_vma);
+	}
+
 	mem->mm_node = NULL;
+	kfree(node);
 }
 
 static int
@@ -1218,7 +880,11 @@ nouveau_gart_manager_new(struct ttm_mem_type_manager *man,
 			 struct ttm_mem_reg *mem)
 {
 	struct drm_nouveau_private *dev_priv = nouveau_bdev(bo->bdev);
+	struct nouveau_bo *nvbo = nouveau_bo(bo);
+	struct nouveau_vma *vma = &nvbo->vma;
+	struct nouveau_vm *vm = vma->vm;
 	struct nouveau_mem *node;
+	int ret;
 
 	if (unlikely((mem->num_pages << PAGE_SHIFT) >=
 		     dev_priv->gart_info.aper_size))
@@ -1227,8 +893,24 @@ nouveau_gart_manager_new(struct ttm_mem_type_manager *man,
 	node = kzalloc(sizeof(*node), GFP_KERNEL);
 	if (!node)
 		return -ENOMEM;
-	node->page_shift = 12;
 
+	/* This node must be for evicting large-paged VRAM
+	 * to system memory.  Due to a nv50 limitation of
+	 * not being able to mix large/small pages within
+	 * the same PDE, we need to create a temporary
+	 * small-paged VMA for the eviction.
+	 */
+	if (vma->node->type != vm->spg_shift) {
+		ret = nouveau_vm_get(vm, (u64)vma->node->length << 12,
+				     vm->spg_shift, NV_MEM_ACCESS_RW,
+				     &node->tmp_vma);
+		if (ret) {
+			kfree(node);
+			return ret;
+		}
+	}
+
+	node->page_shift = nvbo->vma.node->type;
 	mem->mm_node = node;
 	mem->start   = 0;
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/nouveau_mm.c b/drivers/gpu/drm/nouveau/nouveau_mm.c
index b29ffb3..7609756 100644
--- a/drivers/gpu/drm/nouveau/nouveau_mm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_mm.c
@@ -27,7 +27,7 @@
 #include "nouveau_mm.h"
 
 static inline void
-region_put(struct nouveau_mm *mm, struct nouveau_mm_node *a)
+region_put(struct nouveau_mm *rmm, struct nouveau_mm_node *a)
 {
 	list_del(&a->nl_entry);
 	list_del(&a->fl_entry);
@@ -35,7 +35,7 @@ region_put(struct nouveau_mm *mm, struct nouveau_mm_node *a)
 }
 
 static struct nouveau_mm_node *
-region_split(struct nouveau_mm *mm, struct nouveau_mm_node *a, u32 size)
+region_split(struct nouveau_mm *rmm, struct nouveau_mm_node *a, u32 size)
 {
 	struct nouveau_mm_node *b;
 
@@ -57,33 +57,33 @@ region_split(struct nouveau_mm *mm, struct nouveau_mm_node *a, u32 size)
 	return b;
 }
 
-#define node(root, dir) ((root)->nl_entry.dir == &mm->nodes) ? NULL : \
+#define node(root, dir) ((root)->nl_entry.dir == &rmm->nodes) ? NULL : \
 	list_entry((root)->nl_entry.dir, struct nouveau_mm_node, nl_entry)
 
 void
-nouveau_mm_put(struct nouveau_mm *mm, struct nouveau_mm_node *this)
+nouveau_mm_put(struct nouveau_mm *rmm, struct nouveau_mm_node *this)
 {
 	struct nouveau_mm_node *prev = node(this, prev);
 	struct nouveau_mm_node *next = node(this, next);
 
-	list_add(&this->fl_entry, &mm->free);
+	list_add(&this->fl_entry, &rmm->free);
 	this->type = 0;
 
 	if (prev && prev->type == 0) {
 		prev->length += this->length;
-		region_put(mm, this);
+		region_put(rmm, this);
 		this = prev;
 	}
 
 	if (next && next->type == 0) {
 		next->offset  = this->offset;
 		next->length += this->length;
-		region_put(mm, this);
+		region_put(rmm, this);
 	}
 }
 
 int
-nouveau_mm_get(struct nouveau_mm *mm, int type, u32 size, u32 size_nc,
+nouveau_mm_get(struct nouveau_mm *rmm, int type, u32 size, u32 size_nc,
 	       u32 align, struct nouveau_mm_node **pnode)
 {
 	struct nouveau_mm_node *prev, *this, *next;
@@ -92,17 +92,17 @@ nouveau_mm_get(struct nouveau_mm *mm, int type, u32 size, u32 size_nc,
 	u32 splitoff;
 	u32 s, e;
 
-	list_for_each_entry(this, &mm->free, fl_entry) {
+	list_for_each_entry(this, &rmm->free, fl_entry) {
 		e = this->offset + this->length;
 		s = this->offset;
 
 		prev = node(this, prev);
 		if (prev && prev->type != type)
-			s = roundup(s, mm->block_size);
+			s = roundup(s, rmm->block_size);
 
 		next = node(this, next);
 		if (next && next->type != type)
-			e = rounddown(e, mm->block_size);
+			e = rounddown(e, rmm->block_size);
 
 		s  = (s + align_mask) & ~align_mask;
 		e &= ~align_mask;
@@ -110,10 +110,10 @@ nouveau_mm_get(struct nouveau_mm *mm, int type, u32 size, u32 size_nc,
 			continue;
 
 		splitoff = s - this->offset;
-		if (splitoff && !region_split(mm, this, splitoff))
+		if (splitoff && !region_split(rmm, this, splitoff))
 			return -ENOMEM;
 
-		this = region_split(mm, this, min(size, e - s));
+		this = region_split(rmm, this, min(size, e - s));
 		if (!this)
 			return -ENOMEM;
 
@@ -127,49 +127,45 @@ nouveau_mm_get(struct nouveau_mm *mm, int type, u32 size, u32 size_nc,
 }
 
 int
-nouveau_mm_init(struct nouveau_mm *mm, u32 offset, u32 length, u32 block)
+nouveau_mm_init(struct nouveau_mm **prmm, u32 offset, u32 length, u32 block)
 {
-	struct nouveau_mm_node *node;
-
-	if (block) {
-		mutex_init(&mm->mutex);
-		INIT_LIST_HEAD(&mm->nodes);
-		INIT_LIST_HEAD(&mm->free);
-		mm->block_size = block;
-		mm->heap_nodes = 0;
-	}
+	struct nouveau_mm *rmm;
+	struct nouveau_mm_node *heap;
 
-	node = kzalloc(sizeof(*node), GFP_KERNEL);
-	if (!node)
+	heap = kzalloc(sizeof(*heap), GFP_KERNEL);
+	if (!heap)
 		return -ENOMEM;
-	node->offset = roundup(offset, mm->block_size);
-	node->length = rounddown(offset + length, mm->block_size) - node->offset;
+	heap->offset = roundup(offset, block);
+	heap->length = rounddown(offset + length, block) - heap->offset;
 
-	list_add_tail(&node->nl_entry, &mm->nodes);
-	list_add_tail(&node->fl_entry, &mm->free);
-	mm->heap_nodes++;
+	rmm = kzalloc(sizeof(*rmm), GFP_KERNEL);
+	if (!rmm) {
+		kfree(heap);
+		return -ENOMEM;
+	}
+	rmm->block_size = block;
+	mutex_init(&rmm->mutex);
+	INIT_LIST_HEAD(&rmm->nodes);
+	INIT_LIST_HEAD(&rmm->free);
+	list_add(&heap->nl_entry, &rmm->nodes);
+	list_add(&heap->fl_entry, &rmm->free);
+
+	*prmm = rmm;
 	return 0;
 }
 
 int
-nouveau_mm_fini(struct nouveau_mm *mm)
+nouveau_mm_fini(struct nouveau_mm **prmm)
 {
-	struct nouveau_mm_node *node, *heap =
-		list_first_entry(&mm->nodes, struct nouveau_mm_node, nl_entry);
-	int nodes = 0;
-
-	list_for_each_entry(node, &mm->nodes, nl_entry) {
-		if (nodes++ == mm->heap_nodes) {
-			printk(KERN_ERR "nouveau_mm in use at destroy time!\n");
-			list_for_each_entry(node, &mm->nodes, nl_entry) {
-				printk(KERN_ERR "0x%02x: 0x%08x 0x%08x\n",
-				       node->type, node->offset, node->length);
-			}
-			WARN_ON(1);
-			return -EBUSY;
-		}
-	}
+	struct nouveau_mm *rmm = *prmm;
+	struct nouveau_mm_node *heap =
+		list_first_entry(&rmm->nodes, struct nouveau_mm_node, nl_entry);
+
+	if (!list_is_singular(&rmm->nodes))
+		return -EBUSY;
 
 	kfree(heap);
+	kfree(rmm);
+	*prmm = NULL;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nouveau_mm.h b/drivers/gpu/drm/nouveau/nouveau_mm.h
index 57a600c..1f7483a 100644
--- a/drivers/gpu/drm/nouveau/nouveau_mm.h
+++ b/drivers/gpu/drm/nouveau/nouveau_mm.h
@@ -42,18 +42,16 @@ struct nouveau_mm {
 	struct mutex mutex;
 
 	u32 block_size;
-	int heap_nodes;
 };
 
-int  nouveau_mm_init(struct nouveau_mm *, u32 offset, u32 length, u32 block);
-int  nouveau_mm_fini(struct nouveau_mm *);
+int  nouveau_mm_init(struct nouveau_mm **, u32 offset, u32 length, u32 block);
+int  nouveau_mm_fini(struct nouveau_mm **);
 int  nouveau_mm_pre(struct nouveau_mm *);
 int  nouveau_mm_get(struct nouveau_mm *, int type, u32 size, u32 size_nc,
 		    u32 align, struct nouveau_mm_node **);
 void nouveau_mm_put(struct nouveau_mm *, struct nouveau_mm_node *);
 
 int  nv50_vram_init(struct drm_device *);
-void nv50_vram_fini(struct drm_device *);
 int  nv50_vram_new(struct drm_device *, u64 size, u32 align, u32 size_nc,
 		    u32 memtype, struct nouveau_mem **);
 void nv50_vram_del(struct drm_device *, struct nouveau_mem **);
diff --git a/drivers/gpu/drm/nouveau/nouveau_mxm.c b/drivers/gpu/drm/nouveau/nouveau_mxm.c
deleted file mode 100644
index 07d0d1e..0000000
--- a/drivers/gpu/drm/nouveau/nouveau_mxm.c
+++ /dev/null
@@ -1,723 +0,0 @@
-/*
- * Copyright 2011 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Ben Skeggs
- */
-
-#include <linux/acpi.h>
-
-#include "drmP.h"
-#include "nouveau_drv.h"
-
-#define MXM_DBG(dev, fmt, args...) NV_DEBUG((dev), "MXM: " fmt, ##args)
-#define MXM_MSG(dev, fmt, args...) NV_INFO((dev), "MXM: " fmt, ##args)
-
-static u8 *
-mxms_data(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	return dev_priv->mxms;
-
-}
-
-static u16
-mxms_version(struct drm_device *dev)
-{
-	u8 *mxms = mxms_data(dev);
-	u16 version = (mxms[4] << 8) | mxms[5];
-	switch (version ) {
-	case 0x0200:
-	case 0x0201:
-	case 0x0300:
-		return version;
-	default:
-		break;
-	}
-
-	MXM_DBG(dev, "unknown version %d.%d\n", mxms[4], mxms[5]);
-	return 0x0000;
-}
-
-static u16
-mxms_headerlen(struct drm_device *dev)
-{
-	return 8;
-}
-
-static u16
-mxms_structlen(struct drm_device *dev)
-{
-	return *(u16 *)&mxms_data(dev)[6];
-}
-
-static bool
-mxms_checksum(struct drm_device *dev)
-{
-	u16 size = mxms_headerlen(dev) + mxms_structlen(dev);
-	u8 *mxms = mxms_data(dev), sum = 0;
-	while (size--)
-		sum += *mxms++;
-	if (sum) {
-		MXM_DBG(dev, "checksum invalid\n");
-		return false;
-	}
-	return true;
-}
-
-static bool
-mxms_valid(struct drm_device *dev)
-{
-	u8 *mxms = mxms_data(dev);
-	if (*(u32 *)mxms != 0x5f4d584d) {
-		MXM_DBG(dev, "signature invalid\n");
-		return false;
-	}
-
-	if (!mxms_version(dev) || !mxms_checksum(dev))
-		return false;
-
-	return true;
-}
-
-static bool
-mxms_foreach(struct drm_device *dev, u8 types,
-	     bool (*exec)(struct drm_device *, u8 *, void *), void *info)
-{
-	u8 *mxms = mxms_data(dev);
-	u8 *desc = mxms + mxms_headerlen(dev);
-	u8 *fini = desc + mxms_structlen(dev) - 1;
-	while (desc < fini) {
-		u8 type = desc[0] & 0x0f;
-		u8 headerlen = 0;
-		u8 recordlen = 0;
-		u8 entries = 0;
-
-		switch (type) {
-		case 0: /* Output Device Structure */
-			if (mxms_version(dev) >= 0x0300)
-				headerlen = 8;
-			else
-				headerlen = 6;
-			break;
-		case 1: /* System Cooling Capability Structure */
-		case 2: /* Thermal Structure */
-		case 3: /* Input Power Structure */
-			headerlen = 4;
-			break;
-		case 4: /* GPIO Device Structure */
-			headerlen = 4;
-			recordlen = 2;
-			entries   = (ROM32(desc[0]) & 0x01f00000) >> 20;
-			break;
-		case 5: /* Vendor Specific Structure */
-			headerlen = 8;
-			break;
-		case 6: /* Backlight Control Structure */
-			if (mxms_version(dev) >= 0x0300) {
-				headerlen = 4;
-				recordlen = 8;
-				entries   = (desc[1] & 0xf0) >> 4;
-			} else {
-				headerlen = 8;
-			}
-			break;
-		case 7: /* Fan Control Structure */
-			headerlen = 8;
-			recordlen = 4;
-			entries   = desc[1] & 0x07;
-			break;
-		default:
-			MXM_DBG(dev, "unknown descriptor type %d\n", type);
-			return false;
-		}
-
-		if ((drm_debug & DRM_UT_DRIVER) && (exec == NULL)) {
-			static const char * mxms_desc_name[] = {
-				"ODS", "SCCS", "TS", "IPS",
-				"GSD", "VSS", "BCS", "FCS",
-			};
-			u8 *dump = desc;
-			int i, j;
-
-			MXM_DBG(dev, "%4s: ", mxms_desc_name[type]);
-			for (j = headerlen - 1; j >= 0; j--)
-				printk("%02x", dump[j]);
-			printk("\n");
-			dump += headerlen;
-
-			for (i = 0; i < entries; i++, dump += recordlen) {
-				MXM_DBG(dev, "      ");
-				for (j = recordlen - 1; j >= 0; j--)
-					printk("%02x", dump[j]);
-				printk("\n");
-			}
-		}
-
-		if (types & (1 << type)) {
-			if (!exec(dev, desc, info))
-				return false;
-		}
-
-		desc += headerlen + (entries * recordlen);
-	}
-
-	return true;
-}
-
-static u8 *
-mxm_table(struct drm_device *dev, u8 *size)
-{
-	struct bit_entry x;
-
-	if (bit_table(dev, 'x', &x)) {
-		MXM_DBG(dev, "BIT 'x' table not present\n");
-		return NULL;
-	}
-
-	if (x.version != 1 || x.length < 3) {
-		MXM_MSG(dev, "BIT x table %d/%d unknown\n",
-			x.version, x.length);
-		return NULL;
-	}
-
-	*size = x.length;
-	return x.data;
-}
-
-/* These map MXM v2.x digital connection values to the appropriate SOR/link,
- * hopefully they're correct for all boards within the same chipset...
- *
- * MXM v3.x VBIOS are nicer and provide pointers to these tables.
- */
-static u8 nv84_sor_map[16] = {
-	0x00, 0x12, 0x22, 0x11, 0x32, 0x31, 0x11, 0x31,
-	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00
-};
-
-static u8 nv92_sor_map[16] = {
-	0x00, 0x12, 0x22, 0x11, 0x32, 0x31, 0x11, 0x31,
-	0x11, 0x31, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00
-};
-
-static u8 nv94_sor_map[16] = {
-	0x00, 0x14, 0x24, 0x11, 0x34, 0x31, 0x11, 0x31,
-	0x11, 0x31, 0x12, 0x00, 0x00, 0x00, 0x00, 0x00
-};
-
-static u8 nv96_sor_map[16] = {
-	0x00, 0x14, 0x24, 0x00, 0x34, 0x00, 0x11, 0x31,
-	0x11, 0x31, 0x12, 0x00, 0x00, 0x00, 0x00, 0x00
-};
-
-static u8 nv98_sor_map[16] = {
-	0x00, 0x14, 0x12, 0x11, 0x00, 0x31, 0x11, 0x31,
-	0x11, 0x31, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00
-};
-
-static u8
-mxm_sor_map(struct drm_device *dev, u8 conn)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	u8 len, *mxm = mxm_table(dev, &len);
-	if (mxm && len >= 6) {
-		u8 *map = ROMPTR(dev, mxm[4]);
-		if (map) {
-			if (map[0] == 0x10) {
-				if (conn < map[3])
-					return map[map[1] + conn];
-				return 0x00;
-			}
-
-			MXM_MSG(dev, "unknown sor map 0x%02x\n", map[0]);
-		}
-	}
-
-	if (dev_priv->chipset == 0x84 || dev_priv->chipset == 0x86)
-		return nv84_sor_map[conn];
-	if (dev_priv->chipset == 0x92)
-		return nv92_sor_map[conn];
-	if (dev_priv->chipset == 0x94)
-		return nv94_sor_map[conn];
-	if (dev_priv->chipset == 0x96)
-		return nv96_sor_map[conn];
-	if (dev_priv->chipset == 0x98)
-		return nv98_sor_map[conn];
-
-	MXM_MSG(dev, "missing sor map\n");
-	return 0x00;
-}
-
-static u8
-mxm_ddc_map(struct drm_device *dev, u8 port)
-{
-	u8 len, *mxm = mxm_table(dev, &len);
-	if (mxm && len >= 8) {
-		u8 *map = ROMPTR(dev, mxm[6]);
-		if (map) {
-			if (map[0] == 0x10) {
-				if (port < map[3])
-					return map[map[1] + port];
-				return 0x00;
-			}
-
-			MXM_MSG(dev, "unknown ddc map 0x%02x\n", map[0]);
-		}
-	}
-
-	/* v2.x: directly write port as dcb i2cidx */
-	return (port << 4) | port;
-}
-
-struct mxms_odev {
-	u8 outp_type;
-	u8 conn_type;
-	u8 ddc_port;
-	u8 dig_conn;
-};
-
-static void
-mxms_output_device(struct drm_device *dev, u8 *pdata, struct mxms_odev *desc)
-{
-	u64 data = ROM32(pdata[0]);
-	if (mxms_version(dev) >= 0x0300)
-		data |= (u64)ROM16(pdata[4]) << 32;
-
-	desc->outp_type = (data & 0x00000000000000f0ULL) >> 4;
-	desc->ddc_port  = (data & 0x0000000000000f00ULL) >> 8;
-	desc->conn_type = (data & 0x000000000001f000ULL) >> 12;
-	desc->dig_conn  = (data & 0x0000000000780000ULL) >> 19;
-}
-
-struct context {
-	u32 *outp;
-	struct mxms_odev desc;
-};
-
-static bool
-mxm_match_tmds_partner(struct drm_device *dev, u8 *data, void *info)
-{
-	struct context *ctx = info;
-	struct mxms_odev desc;
-
-	mxms_output_device(dev, data, &desc);
-	if (desc.outp_type == 2 &&
-	    desc.dig_conn == ctx->desc.dig_conn)
-		return false;
-	return true;
-}
-
-static bool
-mxm_match_dcb(struct drm_device *dev, u8 *data, void *info)
-{
-	struct context *ctx = info;
-	u64 desc = *(u64 *)data;
-
-	mxms_output_device(dev, data, &ctx->desc);
-
-	/* match dcb encoder type to mxm-ods device type */
-	if ((ctx->outp[0] & 0x0000000f) != ctx->desc.outp_type)
-		return true;
-
-	/* digital output, have some extra stuff to match here, there's a
-	 * table in the vbios that provides a mapping from the mxm digital
-	 * connection enum values to SOR/link
-	 */
-	if ((desc & 0x00000000000000f0) >= 0x20) {
-		/* check against sor index */
-		u8 link = mxm_sor_map(dev, ctx->desc.dig_conn);
-		if ((ctx->outp[0] & 0x0f000000) != (link & 0x0f) << 24)
-			return true;
-
-		/* check dcb entry has a compatible link field */
-		link = (link & 0x30) >> 4;
-		if ((link & ((ctx->outp[1] & 0x00000030) >> 4)) != link)
-			return true;
-	}
-
-	/* mark this descriptor accounted for by setting invalid device type,
-	 * except of course some manufactures don't follow specs properly and
-	 * we need to avoid killing off the TMDS function on DP connectors
-	 * if MXM-SIS is missing an entry for it.
-	 */
-	data[0] &= ~0xf0;
-	if (ctx->desc.outp_type == 6 && ctx->desc.conn_type == 6 &&
-	    mxms_foreach(dev, 0x01, mxm_match_tmds_partner, ctx)) {
-		data[0] |= 0x20; /* modify descriptor to match TMDS now */
-	} else {
-		data[0] |= 0xf0;
-	}
-
-	return false;
-}
-
-static int
-mxm_dcb_sanitise_entry(struct drm_device *dev, void *data, int idx, u8 *dcbe)
-{
-	struct context ctx = { .outp = (u32 *)dcbe };
-	u8 type, i2cidx, link;
-	u8 *conn;
-
-	/* look for an output device structure that matches this dcb entry.
-	 * if one isn't found, disable it.
-	 */
-	if (mxms_foreach(dev, 0x01, mxm_match_dcb, &ctx)) {
-		MXM_DBG(dev, "disable %d: 0x%08x 0x%08x\n",
-			idx, ctx.outp[0], ctx.outp[1]);
-		ctx.outp[0] |= 0x0000000f;
-		return 0;
-	}
-
-	/* modify the output's ddc/aux port, there's a pointer to a table
-	 * with the mapping from mxm ddc/aux port to dcb i2c_index in the
-	 * vbios mxm table
-	 */
-	i2cidx = mxm_ddc_map(dev, ctx.desc.ddc_port);
-	if ((ctx.outp[0] & 0x0000000f) != OUTPUT_DP)
-		i2cidx = (i2cidx & 0x0f) << 4;
-	else
-		i2cidx = (i2cidx & 0xf0);
-
-	if (i2cidx != 0xf0) {
-		ctx.outp[0] &= ~0x000000f0;
-		ctx.outp[0] |= i2cidx;
-	}
-
-	/* override dcb sorconf.link, based on what mxm data says */
-	switch (ctx.desc.outp_type) {
-	case 0x00: /* Analog CRT */
-	case 0x01: /* Analog TV/HDTV */
-		break;
-	default:
-		link = mxm_sor_map(dev, ctx.desc.dig_conn) & 0x30;
-		ctx.outp[1] &= ~0x00000030;
-		ctx.outp[1] |= link;
-		break;
-	}
-
-	/* we may need to fixup various other vbios tables based on what
-	 * the descriptor says the connector type should be.
-	 *
-	 * in a lot of cases, the vbios tables will claim DVI-I is possible,
-	 * and the mxm data says the connector is really HDMI.  another
-	 * common example is DP->eDP.
-	 */
-	conn = dcb_conn(dev, (ctx.outp[0] & 0x0000f000) >> 12);
-	type = conn[0];
-	switch (ctx.desc.conn_type) {
-	case 0x01: /* LVDS */
-		ctx.outp[1] |= 0x00000004; /* use_power_scripts */
-		/* XXX: modify default link width in LVDS table */
-		break;
-	case 0x02: /* HDMI */
-		type = DCB_CONNECTOR_HDMI_1;
-		break;
-	case 0x03: /* DVI-D */
-		type = DCB_CONNECTOR_DVI_D;
-		break;
-	case 0x0e: /* eDP, falls through to DPint */
-		ctx.outp[1] |= 0x00010000;
-	case 0x07: /* DP internal, wtf is this?? HP8670w */
-		ctx.outp[1] |= 0x00000004; /* use_power_scripts? */
-		type = DCB_CONNECTOR_eDP;
-		break;
-	default:
-		break;
-	}
-
-	if (mxms_version(dev) >= 0x0300)
-		conn[0] = type;
-
-	return 0;
-}
-
-static bool
-mxm_show_unmatched(struct drm_device *dev, u8 *data, void *info)
-{
-	u64 desc = *(u64 *)data;
-	if ((desc & 0xf0) != 0xf0)
-		MXM_MSG(dev, "unmatched output device 0x%016llx\n", desc);
-	return true;
-}
-
-static void
-mxm_dcb_sanitise(struct drm_device *dev)
-{
-	u8 *dcb = dcb_table(dev);
-	if (!dcb || dcb[0] != 0x40) {
-		MXM_DBG(dev, "unsupported DCB version\n");
-		return;
-	}
-
-	dcb_outp_foreach(dev, NULL, mxm_dcb_sanitise_entry);
-	mxms_foreach(dev, 0x01, mxm_show_unmatched, NULL);
-}
-
-static bool
-mxm_shadow_rom_fetch(struct nouveau_i2c_chan *i2c, u8 addr,
-		     u8 offset, u8 size, u8 *data)
-{
-	struct i2c_msg msgs[] = {
-		{ .addr = addr, .flags = 0, .len = 1, .buf = &offset },
-		{ .addr = addr, .flags = I2C_M_RD, .len = size, .buf = data, },
-	};
-
-	return i2c_transfer(&i2c->adapter, msgs, 2) == 2;
-}
-
-static bool
-mxm_shadow_rom(struct drm_device *dev, u8 version)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_i2c_chan *i2c = NULL;
-	u8 i2cidx, mxms[6], addr, size;
-
-	i2cidx = mxm_ddc_map(dev, 1 /* LVDS_DDC */) & 0x0f;
-	if (i2cidx < 0x0f)
-		i2c = nouveau_i2c_find(dev, i2cidx);
-	if (!i2c)
-		return false;
-
-	addr = 0x54;
-	if (!mxm_shadow_rom_fetch(i2c, addr, 0, 6, mxms)) {
-		addr = 0x56;
-		if (!mxm_shadow_rom_fetch(i2c, addr, 0, 6, mxms))
-			return false;
-	}
-
-	dev_priv->mxms = mxms;
-	size = mxms_headerlen(dev) + mxms_structlen(dev);
-	dev_priv->mxms = kmalloc(size, GFP_KERNEL);
-
-	if (dev_priv->mxms &&
-	    mxm_shadow_rom_fetch(i2c, addr, 0, size, dev_priv->mxms))
-		return true;
-
-	kfree(dev_priv->mxms);
-	dev_priv->mxms = NULL;
-	return false;
-}
-
-#if defined(CONFIG_ACPI)
-static bool
-mxm_shadow_dsm(struct drm_device *dev, u8 version)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	static char muid[] = {
-		0x00, 0xA4, 0x04, 0x40, 0x7D, 0x91, 0xF2, 0x4C,
-		0xB8, 0x9C, 0x79, 0xB6, 0x2F, 0xD5, 0x56, 0x65
-	};
-	u32 mxms_args[] = { 0x00000000 };
-	union acpi_object args[4] = {
-		/* _DSM MUID */
-		{ .buffer.type = 3,
-		  .buffer.length = sizeof(muid),
-		  .buffer.pointer = muid,
-		},
-		/* spec says this can be zero to mean "highest revision", but
-		 * of course there's at least one bios out there which fails
-		 * unless you pass in exactly the version it supports..
-		 */
-		{ .integer.type = ACPI_TYPE_INTEGER,
-		  .integer.value = (version & 0xf0) << 4 | (version & 0x0f),
-		},
-		/* MXMS function */
-		{ .integer.type = ACPI_TYPE_INTEGER,
-		  .integer.value = 0x00000010,
-		},
-		/* Pointer to MXMS arguments */
-		{ .buffer.type = ACPI_TYPE_BUFFER,
-		  .buffer.length = sizeof(mxms_args),
-		  .buffer.pointer = (char *)mxms_args,
-		},
-	};
-	struct acpi_object_list list = { ARRAY_SIZE(args), args };
-	struct acpi_buffer retn = { ACPI_ALLOCATE_BUFFER, NULL };
-	union acpi_object *obj;
-	acpi_handle handle;
-	int ret;
-
-	handle = DEVICE_ACPI_HANDLE(&dev->pdev->dev);
-	if (!handle)
-		return false;
-
-	ret = acpi_evaluate_object(handle, "_DSM", &list, &retn);
-	if (ret) {
-		MXM_DBG(dev, "DSM MXMS failed: %d\n", ret);
-		return false;
-	}
-
-	obj = retn.pointer;
-	if (obj->type == ACPI_TYPE_BUFFER) {
-		dev_priv->mxms = kmemdup(obj->buffer.pointer,
-					 obj->buffer.length, GFP_KERNEL);
-	} else
-	if (obj->type == ACPI_TYPE_INTEGER) {
-		MXM_DBG(dev, "DSM MXMS returned 0x%llx\n", obj->integer.value);
-	}
-
-	kfree(obj);
-	return dev_priv->mxms != NULL;
-}
-#endif
-
-#if defined(CONFIG_ACPI_WMI) || defined(CONFIG_ACPI_WMI_MODULE)
-
-#define WMI_WMMX_GUID "F6CB5C3C-9CAE-4EBD-B577-931EA32A2CC0"
-
-static u8
-wmi_wmmx_mxmi(struct drm_device *dev, u8 version)
-{
-	u32 mxmi_args[] = { 0x494D584D /* MXMI */, version, 0 };
-	struct acpi_buffer args = { sizeof(mxmi_args), mxmi_args };
-	struct acpi_buffer retn = { ACPI_ALLOCATE_BUFFER, NULL };
-	union acpi_object *obj;
-	acpi_status status;
-
-	status = wmi_evaluate_method(WMI_WMMX_GUID, 0, 0, &args, &retn);
-	if (ACPI_FAILURE(status)) {
-		MXM_DBG(dev, "WMMX MXMI returned %d\n", status);
-		return 0x00;
-	}
-
-	obj = retn.pointer;
-	if (obj->type == ACPI_TYPE_INTEGER) {
-		version = obj->integer.value;
-		MXM_DBG(dev, "WMMX MXMI version %d.%d\n",
-			     (version >> 4), version & 0x0f);
-	} else {
-		version = 0;
-		MXM_DBG(dev, "WMMX MXMI returned non-integer\n");
-	}
-
-	kfree(obj);
-	return version;
-}
-
-static bool
-mxm_shadow_wmi(struct drm_device *dev, u8 version)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	u32 mxms_args[] = { 0x534D584D /* MXMS */, version, 0 };
-	struct acpi_buffer args = { sizeof(mxms_args), mxms_args };
-	struct acpi_buffer retn = { ACPI_ALLOCATE_BUFFER, NULL };
-	union acpi_object *obj;
-	acpi_status status;
-
-	if (!wmi_has_guid(WMI_WMMX_GUID)) {
-		MXM_DBG(dev, "WMMX GUID not found\n");
-		return false;
-	}
-
-	mxms_args[1] = wmi_wmmx_mxmi(dev, 0x00);
-	if (!mxms_args[1])
-		mxms_args[1] = wmi_wmmx_mxmi(dev, version);
-	if (!mxms_args[1])
-		return false;
-
-	status = wmi_evaluate_method(WMI_WMMX_GUID, 0, 0, &args, &retn);
-	if (ACPI_FAILURE(status)) {
-		MXM_DBG(dev, "WMMX MXMS returned %d\n", status);
-		return false;
-	}
-
-	obj = retn.pointer;
-	if (obj->type == ACPI_TYPE_BUFFER) {
-		dev_priv->mxms = kmemdup(obj->buffer.pointer,
-					 obj->buffer.length, GFP_KERNEL);
-	}
-
-	kfree(obj);
-	return dev_priv->mxms != NULL;
-}
-#endif
-
-struct mxm_shadow_h {
-	const char *name;
-	bool (*exec)(struct drm_device *, u8 version);
-} _mxm_shadow[] = {
-	{ "ROM", mxm_shadow_rom },
-#if defined(CONFIG_ACPI)
-	{ "DSM", mxm_shadow_dsm },
-#endif
-#if defined(CONFIG_ACPI_WMI) || defined(CONFIG_ACPI_WMI_MODULE)
-	{ "WMI", mxm_shadow_wmi },
-#endif
-	{}
-};
-
-static int
-mxm_shadow(struct drm_device *dev, u8 version)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct mxm_shadow_h *shadow = _mxm_shadow;
-	do {
-		MXM_DBG(dev, "checking %s\n", shadow->name);
-		if (shadow->exec(dev, version)) {
-			if (mxms_valid(dev))
-				return 0;
-			kfree(dev_priv->mxms);
-			dev_priv->mxms = NULL;
-		}
-	} while ((++shadow)->name);
-	return -ENOENT;
-}
-
-int
-nouveau_mxm_init(struct drm_device *dev)
-{
-	u8 mxm_size, *mxm = mxm_table(dev, &mxm_size);
-	if (!mxm || !mxm[0]) {
-		MXM_MSG(dev, "no VBIOS data, nothing to do\n");
-		return 0;
-	}
-
-	MXM_MSG(dev, "BIOS version %d.%d\n", mxm[0] >> 4, mxm[0] & 0x0f);
-
-	if (mxm_shadow(dev, mxm[0])) {
-		MXM_MSG(dev, "failed to locate valid SIS\n");
-#if 0
-		/* we should, perhaps, fall back to some kind of limited
-		 * mode here if the x86 vbios hasn't already done the
-		 * work for us (so we prevent loading with completely
-		 * whacked vbios tables).
-		 */
-		return -EINVAL;
-#else
-		return 0;
-#endif
-	}
-
-	MXM_MSG(dev, "MXMS Version %d.%d\n",
-		mxms_version(dev) >> 8, mxms_version(dev) & 0xff);
-	mxms_foreach(dev, 0, NULL, NULL);
-
-	if (nouveau_mxmdcb)
-		mxm_dcb_sanitise(dev);
-	return 0;
-}
-
-void
-nouveau_mxm_fini(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	kfree(dev_priv->mxms);
-	dev_priv->mxms = NULL;
-}
diff --git a/drivers/gpu/drm/nouveau/nouveau_notifier.c b/drivers/gpu/drm/nouveau/nouveau_notifier.c
index 2ef883c..5b39718 100644
--- a/drivers/gpu/drm/nouveau/nouveau_notifier.c
+++ b/drivers/gpu/drm/nouveau/nouveau_notifier.c
@@ -34,7 +34,6 @@ int
 nouveau_notifier_init_channel(struct nouveau_channel *chan)
 {
 	struct drm_device *dev = chan->dev;
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_bo *ntfy = NULL;
 	uint32_t flags, ttmpl;
 	int ret;
@@ -47,7 +46,7 @@ nouveau_notifier_init_channel(struct nouveau_channel *chan)
 		ttmpl = TTM_PL_FLAG_TT;
 	}
 
-	ret = nouveau_gem_new(dev, PAGE_SIZE, 0, flags, 0, 0, &ntfy);
+	ret = nouveau_gem_new(dev, NULL, PAGE_SIZE, 0, flags, 0, 0, &ntfy);
 	if (ret)
 		return ret;
 
@@ -59,22 +58,14 @@ nouveau_notifier_init_channel(struct nouveau_channel *chan)
 	if (ret)
 		goto out_err;
 
-	if (dev_priv->card_type >= NV_50) {
-		ret = nouveau_bo_vma_add(ntfy, chan->vm, &chan->notifier_vma);
-		if (ret)
-			goto out_err;
-	}
-
 	ret = drm_mm_init(&chan->notifier_heap, 0, ntfy->bo.mem.size);
 	if (ret)
 		goto out_err;
 
 	chan->notifier_bo = ntfy;
 out_err:
-	if (ret) {
-		nouveau_bo_vma_del(ntfy, &chan->notifier_vma);
+	if (ret)
 		drm_gem_object_unreference_unlocked(ntfy->gem);
-	}
 
 	return ret;
 }
@@ -87,7 +78,6 @@ nouveau_notifier_takedown_channel(struct nouveau_channel *chan)
 	if (!chan->notifier_bo)
 		return;
 
-	nouveau_bo_vma_del(chan->notifier_bo, &chan->notifier_vma);
 	nouveau_bo_unmap(chan->notifier_bo);
 	mutex_lock(&dev->struct_mutex);
 	nouveau_bo_unpin(chan->notifier_bo);
@@ -115,7 +105,7 @@ nouveau_notifier_alloc(struct nouveau_channel *chan, uint32_t handle,
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_gpuobj *nobj = NULL;
 	struct drm_mm_node *mem;
-	uint64_t offset;
+	uint32_t offset;
 	int target, ret;
 
 	mem = drm_mm_search_free_in_range(&chan->notifier_heap, size, 0,
@@ -132,10 +122,10 @@ nouveau_notifier_alloc(struct nouveau_channel *chan, uint32_t handle,
 			target = NV_MEM_TARGET_VRAM;
 		else
 			target = NV_MEM_TARGET_GART;
-		offset  = chan->notifier_bo->bo.offset;
+		offset  = chan->notifier_bo->bo.mem.start << PAGE_SHIFT;
 	} else {
 		target = NV_MEM_TARGET_VM;
-		offset = chan->notifier_vma.offset;
+		offset = chan->notifier_bo->vma.offset;
 	}
 	offset += mem->start;
 
@@ -193,7 +183,7 @@ nouveau_ioctl_notifier_alloc(struct drm_device *dev, void *data,
 	if (unlikely(dev_priv->card_type >= NV_C0))
 		return -EINVAL;
 
-	chan = nouveau_channel_get(file_priv, na->channel);
+	chan = nouveau_channel_get(dev, file_priv, na->channel);
 	if (IS_ERR(chan))
 		return PTR_ERR(chan);
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_object.c b/drivers/gpu/drm/nouveau/nouveau_object.c
index cc419fa..8f97016 100644
--- a/drivers/gpu/drm/nouveau/nouveau_object.c
+++ b/drivers/gpu/drm/nouveau/nouveau_object.c
@@ -125,7 +125,7 @@ nouveau_gpuobj_mthd_call2(struct drm_device *dev, int chid,
 	int ret = -EINVAL;
 
 	spin_lock_irqsave(&dev_priv->channels.lock, flags);
-	if (chid >= 0 && chid < dev_priv->engine.fifo.channels)
+	if (chid > 0 && chid < dev_priv->engine.fifo.channels)
 		chan = dev_priv->channels.ptr[chid];
 	if (chan)
 		ret = nouveau_gpuobj_mthd_call(chan, class, mthd, data);
@@ -191,7 +191,7 @@ nouveau_gpuobj_new(struct drm_device *dev, struct nouveau_channel *chan,
 	list_add_tail(&gpuobj->list, &dev_priv->gpuobj_list);
 	spin_unlock(&dev_priv->ramin_lock);
 
-	if (!(flags & NVOBJ_FLAG_VM) && chan) {
+	if (chan) {
 		ramin = drm_mm_search_free(&chan->ramin_heap, size, align, 0);
 		if (ramin)
 			ramin = drm_mm_get_block(ramin, size, align);
@@ -208,7 +208,7 @@ nouveau_gpuobj_new(struct drm_device *dev, struct nouveau_channel *chan,
 		gpuobj->vinst = ramin->start + chan->ramin->vinst;
 		gpuobj->node  = ramin;
 	} else {
-		ret = instmem->get(gpuobj, chan, size, align);
+		ret = instmem->get(gpuobj, size, align);
 		if (ret) {
 			nouveau_gpuobj_ref(NULL, &gpuobj);
 			return ret;
@@ -680,7 +680,7 @@ nouveau_gpuobj_channel_init_pramin(struct nouveau_channel *chan)
 		return ret;
 	}
 
-	ret = drm_mm_init(&chan->ramin_heap, base, size - base);
+	ret = drm_mm_init(&chan->ramin_heap, base, size);
 	if (ret) {
 		NV_ERROR(dev, "Error creating PRAMIN heap: %d\n", ret);
 		nouveau_gpuobj_ref(NULL, &chan->ramin);
@@ -690,68 +690,35 @@ nouveau_gpuobj_channel_init_pramin(struct nouveau_channel *chan)
 	return 0;
 }
 
-static int
-nvc0_gpuobj_channel_init(struct nouveau_channel *chan, struct nouveau_vm *vm)
-{
-	struct drm_nouveau_private *dev_priv = chan->dev->dev_private;
-	struct drm_device *dev = chan->dev;
-	struct nouveau_gpuobj *pgd = NULL;
-	struct nouveau_vm_pgd *vpgd;
-	int ret, i;
-
-	ret = nouveau_gpuobj_new(dev, NULL, 4096, 0x1000, 0, &chan->ramin);
-	if (ret)
-		return ret;
-
-	/* create page directory for this vm if none currently exists,
-	 * will be destroyed automagically when last reference to the
-	 * vm is removed
-	 */
-	if (list_empty(&vm->pgd_list)) {
-		ret = nouveau_gpuobj_new(dev, NULL, 65536, 0x1000, 0, &pgd);
-		if (ret)
-			return ret;
-	}
-	nouveau_vm_ref(vm, &chan->vm, pgd);
-	nouveau_gpuobj_ref(NULL, &pgd);
-
-	/* point channel at vm's page directory */
-	vpgd = list_first_entry(&vm->pgd_list, struct nouveau_vm_pgd, head);
-	nv_wo32(chan->ramin, 0x0200, lower_32_bits(vpgd->obj->vinst));
-	nv_wo32(chan->ramin, 0x0204, upper_32_bits(vpgd->obj->vinst));
-	nv_wo32(chan->ramin, 0x0208, 0xffffffff);
-	nv_wo32(chan->ramin, 0x020c, 0x000000ff);
-
-	/* map display semaphore buffers into channel's vm */
-	for (i = 0; i < dev->mode_config.num_crtc; i++) {
-		struct nouveau_bo *bo;
-		if (dev_priv->card_type >= NV_D0)
-			bo = nvd0_display_crtc_sema(dev, i);
-		else
-			bo = nv50_display(dev)->crtc[i].sem.bo;
-
-		ret = nouveau_bo_vma_add(bo, chan->vm, &chan->dispc_vma[i]);
-		if (ret)
-			return ret;
-	}
-
-	return 0;
-}
-
 int
 nouveau_gpuobj_channel_init(struct nouveau_channel *chan,
 			    uint32_t vram_h, uint32_t tt_h)
 {
 	struct drm_device *dev = chan->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_fpriv *fpriv = nouveau_fpriv(chan->file_priv);
-	struct nouveau_vm *vm = fpriv ? fpriv->vm : dev_priv->chan_vm;
 	struct nouveau_gpuobj *vram = NULL, *tt = NULL;
 	int ret, i;
 
 	NV_DEBUG(dev, "ch%d vram=0x%08x tt=0x%08x\n", chan->id, vram_h, tt_h);
-	if (dev_priv->card_type >= NV_C0)
-		return nvc0_gpuobj_channel_init(chan, vm);
+
+	if (dev_priv->card_type == NV_C0) {
+		struct nouveau_vm *vm = dev_priv->chan_vm;
+		struct nouveau_vm_pgd *vpgd;
+
+		ret = nouveau_gpuobj_new(dev, NULL, 4096, 0x1000, 0,
+					 &chan->ramin);
+		if (ret)
+			return ret;
+
+		nouveau_vm_ref(vm, &chan->vm, NULL);
+
+		vpgd = list_first_entry(&vm->pgd_list, struct nouveau_vm_pgd, head);
+		nv_wo32(chan->ramin, 0x0200, lower_32_bits(vpgd->obj->vinst));
+		nv_wo32(chan->ramin, 0x0204, upper_32_bits(vpgd->obj->vinst));
+		nv_wo32(chan->ramin, 0x0208, 0xffffffff);
+		nv_wo32(chan->ramin, 0x020c, 0x000000ff);
+		return 0;
+	}
 
 	/* Allocate a chunk of memory for per-channel object storage */
 	ret = nouveau_gpuobj_channel_init_pramin(chan);
@@ -764,7 +731,7 @@ nouveau_gpuobj_channel_init(struct nouveau_channel *chan,
 	 *  - Allocate per-channel page-directory
 	 *  - Link with shared channel VM
 	 */
-	if (vm) {
+	if (dev_priv->chan_vm) {
 		u32 pgd_offs = (dev_priv->chipset == 0x50) ? 0x1400 : 0x0200;
 		u64 vm_vinst = chan->ramin->vinst + pgd_offs;
 		u32 vm_pinst = chan->ramin->pinst;
@@ -777,7 +744,7 @@ nouveau_gpuobj_channel_init(struct nouveau_channel *chan,
 		if (ret)
 			return ret;
 
-		nouveau_vm_ref(vm, &chan->vm, chan->vm_pd);
+		nouveau_vm_ref(dev_priv->chan_vm, &chan->vm, chan->vm_pd);
 	}
 
 	/* RAMHT */
@@ -797,11 +764,11 @@ nouveau_gpuobj_channel_init(struct nouveau_channel *chan,
 			return ret;
 
 		/* dma objects for display sync channel semaphore blocks */
-		for (i = 0; i < dev->mode_config.num_crtc; i++) {
+		for (i = 0; i < 2; i++) {
 			struct nouveau_gpuobj *sem = NULL;
 			struct nv50_display_crtc *dispc =
 				&nv50_display(dev)->crtc[i];
-			u64 offset = dispc->sem.bo->bo.offset;
+			u64 offset = dispc->sem.bo->bo.mem.start << PAGE_SHIFT;
 
 			ret = nouveau_gpuobj_dma_new(chan, 0x3d, offset, 0xfff,
 						     NV_MEM_ACCESS_RW,
@@ -874,24 +841,10 @@ void
 nouveau_gpuobj_channel_takedown(struct nouveau_channel *chan)
 {
 	struct drm_device *dev = chan->dev;
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	int i;
 
 	NV_DEBUG(dev, "ch%d\n", chan->id);
 
-	if (dev_priv->card_type >= NV_D0) {
-		for (i = 0; i < dev->mode_config.num_crtc; i++) {
-			struct nouveau_bo *bo = nvd0_display_crtc_sema(dev, i);
-			nouveau_bo_vma_del(bo, &chan->dispc_vma[i]);
-		}
-	} else
-	if (dev_priv->card_type >= NV_50) {
-		struct nv50_display *disp = nv50_display(dev);
-		for (i = 0; i < dev->mode_config.num_crtc; i++) {
-			struct nv50_display_crtc *dispc = &disp->crtc[i];
-			nouveau_bo_vma_del(dispc->sem.bo, &chan->dispc_vma[i]);
-		}
-	}
+	nouveau_ramht_ref(NULL, &chan->ramht, chan);
 
 	nouveau_vm_ref(NULL, &chan->vm, chan->vm_pd);
 	nouveau_gpuobj_ref(NULL, &chan->vm_pd);
@@ -956,7 +909,7 @@ int nouveau_ioctl_grobj_alloc(struct drm_device *dev, void *data,
 	if (init->handle == ~0)
 		return -EINVAL;
 
-	chan = nouveau_channel_get(file_priv, init->channel);
+	chan = nouveau_channel_get(dev, file_priv, init->channel);
 	if (IS_ERR(chan))
 		return PTR_ERR(chan);
 
@@ -983,7 +936,7 @@ int nouveau_ioctl_gpuobj_free(struct drm_device *dev, void *data,
 	struct nouveau_channel *chan;
 	int ret;
 
-	chan = nouveau_channel_get(file_priv, objfree->channel);
+	chan = nouveau_channel_get(dev, file_priv, objfree->channel);
 	if (IS_ERR(chan))
 		return PTR_ERR(chan);
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_perf.c b/drivers/gpu/drm/nouveau/nouveau_perf.c
index 69a528d..ef9dec0 100644
--- a/drivers/gpu/drm/nouveau/nouveau_perf.c
+++ b/drivers/gpu/drm/nouveau/nouveau_perf.c
@@ -27,178 +27,6 @@
 #include "nouveau_drv.h"
 #include "nouveau_pm.h"
 
-static u8 *
-nouveau_perf_table(struct drm_device *dev, u8 *ver)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nvbios *bios = &dev_priv->vbios;
-	struct bit_entry P;
-
-	if (!bit_table(dev, 'P', &P) && P.version && P.version <= 2) {
-		u8 *perf = ROMPTR(dev, P.data[0]);
-		if (perf) {
-			*ver = perf[0];
-			return perf;
-		}
-	}
-
-	if (bios->type == NVBIOS_BMP) {
-		if (bios->data[bios->offset + 6] >= 0x25) {
-			u8 *perf = ROMPTR(dev, bios->data[bios->offset + 0x94]);
-			if (perf) {
-				*ver = perf[1];
-				return perf;
-			}
-		}
-	}
-
-	return NULL;
-}
-
-static u8 *
-nouveau_perf_entry(struct drm_device *dev, int idx,
-		   u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
-{
-	u8 *perf = nouveau_perf_table(dev, ver);
-	if (perf) {
-		if (*ver >= 0x12 && *ver < 0x20 && idx < perf[2]) {
-			*hdr = perf[3];
-			*cnt = 0;
-			*len = 0;
-			return perf + perf[0] + idx * perf[3];
-		} else
-		if (*ver >= 0x20 && *ver < 0x40 && idx < perf[2]) {
-			*hdr = perf[3];
-			*cnt = perf[4];
-			*len = perf[5];
-			return perf + perf[1] + idx * (*hdr + (*cnt * *len));
-		} else
-		if (*ver >= 0x40 && *ver < 0x41 && idx < perf[5]) {
-			*hdr = perf[2];
-			*cnt = perf[4];
-			*len = perf[3];
-			return perf + perf[1] + idx * (*hdr + (*cnt * *len));
-		}
-	}
-	return NULL;
-}
-
-static u8 *
-nouveau_perf_rammap(struct drm_device *dev, u32 freq,
-		    u8 *ver, u8 *hdr, u8 *cnt, u8 *len)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct bit_entry P;
-	u8 *perf, i = 0;
-
-	if (!bit_table(dev, 'P', &P) && P.version == 2) {
-		u8 *rammap = ROMPTR(dev, P.data[4]);
-		if (rammap) {
-			u8 *ramcfg = rammap + rammap[1];
-
-			*ver = rammap[0];
-			*hdr = rammap[2];
-			*cnt = rammap[4];
-			*len = rammap[3];
-
-			freq /= 1000;
-			for (i = 0; i < rammap[5]; i++) {
-				if (freq >= ROM16(ramcfg[0]) &&
-				    freq <= ROM16(ramcfg[2]))
-					return ramcfg;
-
-				ramcfg += *hdr + (*cnt * *len);
-			}
-		}
-
-		return NULL;
-	}
-
-	if (dev_priv->chipset == 0x49 ||
-	    dev_priv->chipset == 0x4b)
-		freq /= 2;
-
-	while ((perf = nouveau_perf_entry(dev, i++, ver, hdr, cnt, len))) {
-		if (*ver >= 0x20 && *ver < 0x25) {
-			if (perf[0] != 0xff && freq <= ROM16(perf[11]) * 1000)
-				break;
-		} else
-		if (*ver >= 0x25 && *ver < 0x40) {
-			if (perf[0] != 0xff && freq <= ROM16(perf[12]) * 1000)
-				break;
-		}
-	}
-
-	if (perf) {
-		u8 *ramcfg = perf + *hdr;
-		*ver = 0x00;
-		*hdr = 0;
-		return ramcfg;
-	}
-
-	return NULL;
-}
-
-u8 *
-nouveau_perf_ramcfg(struct drm_device *dev, u32 freq, u8 *ver, u8 *len)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nvbios *bios = &dev_priv->vbios;
-	u8 strap, hdr, cnt;
-	u8 *rammap;
-
-	strap = (nv_rd32(dev, 0x101000) & 0x0000003c) >> 2;
-	if (bios->ram_restrict_tbl_ptr)
-		strap = bios->data[bios->ram_restrict_tbl_ptr + strap];
-
-	rammap = nouveau_perf_rammap(dev, freq, ver, &hdr, &cnt, len);
-	if (rammap && strap < cnt)
-		return rammap + hdr + (strap * *len);
-
-	return NULL;
-}
-
-u8 *
-nouveau_perf_timing(struct drm_device *dev, u32 freq, u8 *ver, u8 *len)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nvbios *bios = &dev_priv->vbios;
-	struct bit_entry P;
-	u8 *perf, *timing = NULL;
-	u8 i = 0, hdr, cnt;
-
-	if (bios->type == NVBIOS_BMP) {
-		while ((perf = nouveau_perf_entry(dev, i++, ver, &hdr, &cnt,
-						  len)) && *ver == 0x15) {
-			if (freq <= ROM32(perf[5]) * 20) {
-				*ver = 0x00;
-				*len = 14;
-				return perf + 41;
-			}
-		}
-		return NULL;
-	}
-
-	if (!bit_table(dev, 'P', &P)) {
-		if (P.version == 1)
-			timing = ROMPTR(dev, P.data[4]);
-		else
-		if (P.version == 2)
-			timing = ROMPTR(dev, P.data[8]);
-	}
-
-	if (timing && timing[0] == 0x10) {
-		u8 *ramcfg = nouveau_perf_ramcfg(dev, freq, ver, len);
-		if (ramcfg && ramcfg[1] < timing[2]) {
-			*ver = timing[0];
-			*len = timing[3];
-			return timing + timing[1] + (ramcfg[1] * timing[3]);
-		}
-	}
-
-	return NULL;
-}
-
 static void
 legacy_perf_init(struct drm_device *dev)
 {
@@ -213,7 +41,7 @@ legacy_perf_init(struct drm_device *dev)
 		return;
 	}
 
-	perf = ROMPTR(dev, bmp[0x73]);
+	perf = ROMPTR(bios, bmp[0x73]);
 	if (!perf) {
 		NV_DEBUG(dev, "No memclock table pointer found.\n");
 		return;
@@ -244,46 +72,66 @@ legacy_perf_init(struct drm_device *dev)
 	pm->nr_perflvl = 1;
 }
 
-static void
-nouveau_perf_voltage(struct drm_device *dev, struct nouveau_pm_level *perflvl)
+static struct nouveau_pm_memtiming *
+nouveau_perf_timing(struct drm_device *dev, struct bit_entry *P,
+		    u16 memclk, u8 *entry, u8 recordlen, u8 entries)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct bit_entry P;
-	u8 *vmap;
-	int id;
-
-	id = perflvl->volt_min;
-	perflvl->volt_min = 0;
+	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
+	struct nvbios *bios = &dev_priv->vbios;
+	u8 ramcfg;
+	int i;
 
-	/* boards using voltage table version <0x40 store the voltage
-	 * level directly in the perflvl entry as a multiple of 10mV
+	/* perf v2 has a separate "timing map" table, we have to match
+	 * the target memory clock to a specific entry, *then* use
+	 * ramcfg to select the correct subentry
 	 */
-	if (dev_priv->engine.pm.voltage.version < 0x40) {
-		perflvl->volt_min = id * 10000;
-		perflvl->volt_max = perflvl->volt_min;
-		return;
-	}
+	if (P->version == 2) {
+		u8 *tmap = ROMPTR(bios, P->data[4]);
+		if (!tmap) {
+			NV_DEBUG(dev, "no timing map pointer\n");
+			return NULL;
+		}
 
-	/* on newer ones, the perflvl stores an index into yet another
-	 * vbios table containing a min/max voltage value for the perflvl
-	 */
-	if (bit_table(dev, 'P', &P) || P.version != 2 || P.length < 34) {
-		NV_DEBUG(dev, "where's our volt map table ptr? %d %d\n",
-			 P.version, P.length);
-		return;
+		if (tmap[0] != 0x10) {
+			NV_WARN(dev, "timing map 0x%02x unknown\n", tmap[0]);
+			return NULL;
+		}
+
+		entry = tmap + tmap[1];
+		recordlen = tmap[2] + (tmap[4] * tmap[3]);
+		for (i = 0; i < tmap[5]; i++, entry += recordlen) {
+			if (memclk >= ROM16(entry[0]) &&
+			    memclk <= ROM16(entry[2]))
+				break;
+		}
+
+		if (i == tmap[5]) {
+			NV_WARN(dev, "no match in timing map table\n");
+			return NULL;
+		}
+
+		entry += tmap[2];
+		recordlen = tmap[3];
+		entries   = tmap[4];
 	}
 
-	vmap = ROMPTR(dev, P.data[32]);
-	if (!vmap) {
-		NV_DEBUG(dev, "volt map table pointer invalid\n");
-		return;
+	ramcfg = (nv_rd32(dev, NV_PEXTDEV_BOOT_0) & 0x0000003c) >> 2;
+	if (bios->ram_restrict_tbl_ptr)
+		ramcfg = bios->data[bios->ram_restrict_tbl_ptr + ramcfg];
+
+	if (ramcfg >= entries) {
+		NV_WARN(dev, "ramcfg strap out of bounds!\n");
+		return NULL;
 	}
 
-	if (id < vmap[3]) {
-		vmap += vmap[1] + (vmap[2] * id);
-		perflvl->volt_min = ROM32(vmap[0]);
-		perflvl->volt_max = ROM32(vmap[4]);
+	entry += ramcfg * recordlen;
+	if (entry[1] >= pm->memtimings.nr_timing) {
+		NV_WARN(dev, "timingset %d does not exist\n", entry[1]);
+		return NULL;
 	}
+
+	return &pm->memtimings.timing[entry[1]];
 }
 
 void
@@ -292,120 +140,153 @@ nouveau_perf_init(struct drm_device *dev)
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
 	struct nvbios *bios = &dev_priv->vbios;
-	u8 *perf, ver, hdr, cnt, len;
-	int ret, vid, i = -1;
+	struct bit_entry P;
+	u8 version, headerlen, recordlen, entries;
+	u8 *perf, *entry;
+	int vid, i;
 
-	if (bios->type == NVBIOS_BMP && bios->data[bios->offset + 6] < 0x25) {
-		legacy_perf_init(dev);
-		return;
+	if (bios->type == NVBIOS_BIT) {
+		if (bit_table(dev, 'P', &P))
+			return;
+
+		if (P.version != 1 && P.version != 2) {
+			NV_WARN(dev, "unknown perf for BIT P %d\n", P.version);
+			return;
+		}
+
+		perf = ROMPTR(bios, P.data[0]);
+		version   = perf[0];
+		headerlen = perf[1];
+		if (version < 0x40) {
+			recordlen = perf[3] + (perf[4] * perf[5]);
+			entries   = perf[2];
+		} else {
+			recordlen = perf[2] + (perf[3] * perf[4]);
+			entries   = perf[5];
+		}
+	} else {
+		if (bios->data[bios->offset + 6] < 0x25) {
+			legacy_perf_init(dev);
+			return;
+		}
+
+		perf = ROMPTR(bios, bios->data[bios->offset + 0x94]);
+		if (!perf) {
+			NV_DEBUG(dev, "perf table pointer invalid\n");
+			return;
+		}
+
+		version   = perf[1];
+		headerlen = perf[0];
+		recordlen = perf[3];
+		entries   = perf[2];
 	}
 
-	perf = nouveau_perf_table(dev, &ver);
-	if (ver >= 0x20 && ver < 0x40)
-		pm->fan.pwm_divisor = ROM16(perf[6]);
+	if (entries > NOUVEAU_PM_MAX_LEVEL) {
+		NV_DEBUG(dev, "perf table has too many entries - buggy vbios?\n");
+		entries = NOUVEAU_PM_MAX_LEVEL;
+	}
 
-	while ((perf = nouveau_perf_entry(dev, ++i, &ver, &hdr, &cnt, &len))) {
+	entry = perf + headerlen;
+	for (i = 0; i < entries; i++) {
 		struct nouveau_pm_level *perflvl = &pm->perflvl[pm->nr_perflvl];
 
-		if (perf[0] == 0xff)
+		perflvl->timing = NULL;
+
+		if (entry[0] == 0xff) {
+			entry += recordlen;
 			continue;
+		}
 
-		switch (ver) {
+		switch (version) {
 		case 0x12:
 		case 0x13:
 		case 0x15:
-			perflvl->fanspeed = perf[55];
-			if (hdr > 56)
-				perflvl->volt_min = perf[56];
-			perflvl->core = ROM32(perf[1]) * 10;
-			perflvl->memory = ROM32(perf[5]) * 20;
+			perflvl->fanspeed = entry[55];
+			perflvl->voltage = (recordlen > 56) ? entry[56] : 0;
+			perflvl->core = ROM32(entry[1]) * 10;
+			perflvl->memory = ROM32(entry[5]) * 20;
 			break;
 		case 0x21:
 		case 0x23:
 		case 0x24:
-			perflvl->fanspeed = perf[4];
-			perflvl->volt_min = perf[5];
-			perflvl->shader = ROM16(perf[6]) * 1000;
-			perflvl->core = perflvl->shader;
-			perflvl->core += (signed char)perf[8] * 1000;
+			perflvl->fanspeed = entry[4];
+			perflvl->voltage = entry[5];
+			perflvl->core = ROM16(entry[6]) * 1000;
+
 			if (dev_priv->chipset == 0x49 ||
 			    dev_priv->chipset == 0x4b)
-				perflvl->memory = ROM16(perf[11]) * 1000;
+				perflvl->memory = ROM16(entry[11]) * 1000;
 			else
-				perflvl->memory = ROM16(perf[11]) * 2000;
+				perflvl->memory = ROM16(entry[11]) * 2000;
+
 			break;
 		case 0x25:
-			perflvl->fanspeed = perf[4];
-			perflvl->volt_min = perf[5];
-			perflvl->core = ROM16(perf[6]) * 1000;
-			perflvl->shader = ROM16(perf[10]) * 1000;
-			perflvl->memory = ROM16(perf[12]) * 1000;
+			perflvl->fanspeed = entry[4];
+			perflvl->voltage = entry[5];
+			perflvl->core = ROM16(entry[6]) * 1000;
+			perflvl->shader = ROM16(entry[10]) * 1000;
+			perflvl->memory = ROM16(entry[12]) * 1000;
 			break;
 		case 0x30:
-			perflvl->memscript = ROM16(perf[2]);
+			perflvl->memscript = ROM16(entry[2]);
 		case 0x35:
-			perflvl->fanspeed = perf[6];
-			perflvl->volt_min = perf[7];
-			perflvl->core = ROM16(perf[8]) * 1000;
-			perflvl->shader = ROM16(perf[10]) * 1000;
-			perflvl->memory = ROM16(perf[12]) * 1000;
-			perflvl->vdec = ROM16(perf[16]) * 1000;
-			perflvl->dom6 = ROM16(perf[20]) * 1000;
+			perflvl->fanspeed = entry[6];
+			perflvl->voltage = entry[7];
+			perflvl->core = ROM16(entry[8]) * 1000;
+			perflvl->shader = ROM16(entry[10]) * 1000;
+			perflvl->memory = ROM16(entry[12]) * 1000;
+			/*XXX: confirm on 0x35 */
+			perflvl->unk05 = ROM16(entry[16]) * 1000;
 			break;
 		case 0x40:
-#define subent(n) ((ROM16(perf[hdr + (n) * len]) & 0xfff) * 1000)
+#define subent(n) entry[perf[2] + ((n) * perf[3])]
 			perflvl->fanspeed = 0; /*XXX*/
-			perflvl->volt_min = perf[2];
+			perflvl->voltage = entry[2];
 			if (dev_priv->card_type == NV_50) {
-				perflvl->core   = subent(0);
-				perflvl->shader = subent(1);
-				perflvl->memory = subent(2);
-				perflvl->vdec   = subent(3);
-				perflvl->unka0  = subent(4);
+				perflvl->core = ROM16(subent(0)) & 0xfff;
+				perflvl->shader = ROM16(subent(1)) & 0xfff;
+				perflvl->memory = ROM16(subent(2)) & 0xfff;
 			} else {
-				perflvl->hub06  = subent(0);
-				perflvl->hub01  = subent(1);
-				perflvl->copy   = subent(2);
-				perflvl->shader = subent(3);
-				perflvl->rop    = subent(4);
-				perflvl->memory = subent(5);
-				perflvl->vdec   = subent(6);
-				perflvl->daemon = subent(10);
-				perflvl->hub07  = subent(11);
+				perflvl->shader = ROM16(subent(3)) & 0xfff;
 				perflvl->core   = perflvl->shader / 2;
+				perflvl->unk0a  = ROM16(subent(4)) & 0xfff;
+				perflvl->memory = ROM16(subent(5)) & 0xfff;
 			}
+
+			perflvl->core *= 1000;
+			perflvl->shader *= 1000;
+			perflvl->memory *= 1000;
+			perflvl->unk0a *= 1000;
 			break;
 		}
 
 		/* make sure vid is valid */
-		nouveau_perf_voltage(dev, perflvl);
-		if (pm->voltage.supported && perflvl->volt_min) {
-			vid = nouveau_volt_vid_lookup(dev, perflvl->volt_min);
+		if (pm->voltage.supported && perflvl->voltage) {
+			vid = nouveau_volt_vid_lookup(dev, perflvl->voltage);
 			if (vid < 0) {
-				NV_DEBUG(dev, "perflvl %d, bad vid\n", i);
+				NV_DEBUG(dev, "drop perflvl %d, bad vid\n", i);
+				entry += recordlen;
 				continue;
 			}
 		}
 
 		/* get the corresponding memory timings */
-		ret = nouveau_mem_timing_calc(dev, perflvl->memory,
-					          &perflvl->timing);
-		if (ret) {
-			NV_DEBUG(dev, "perflvl %d, bad timing: %d\n", i, ret);
-			continue;
+		if (version > 0x15) {
+			/* last 3 args are for < 0x40, ignored for >= 0x40 */
+			perflvl->timing =
+				nouveau_perf_timing(dev, &P,
+						    perflvl->memory / 1000,
+						    entry + perf[3],
+						    perf[5], perf[4]);
 		}
 
 		snprintf(perflvl->name, sizeof(perflvl->name),
 			 "performance_level_%d", i);
 		perflvl->id = i;
-
-		snprintf(perflvl->profile.name, sizeof(perflvl->profile.name),
-			 "%d", perflvl->id);
-		perflvl->profile.func = &nouveau_pm_static_profile_func;
-		list_add_tail(&perflvl->profile.head, &pm->profiles);
-
-
 		pm->nr_perflvl++;
+
+		entry += recordlen;
 	}
 }
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_pm.c b/drivers/gpu/drm/nouveau/nouveau_pm.c
index 34d591b..da8d994 100644
--- a/drivers/gpu/drm/nouveau/nouveau_pm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_pm.c
@@ -26,7 +26,6 @@
 
 #include "nouveau_drv.h"
 #include "nouveau_pm.h"
-#include "nouveau_gpio.h"
 
 #ifdef CONFIG_ACPI
 #include <linux/acpi.h>
@@ -36,98 +35,22 @@
 #include <linux/hwmon-sysfs.h>
 
 static int
-nouveau_pwmfan_get(struct drm_device *dev)
+nouveau_pm_clock_set(struct drm_device *dev, struct nouveau_pm_level *perflvl,
+		     u8 id, u32 khz)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
-	struct gpio_func gpio;
-	u32 divs, duty;
-	int ret;
-
-	if (!pm->pwm_get)
-		return -ENODEV;
-
-	ret = nouveau_gpio_find(dev, 0, DCB_GPIO_PWM_FAN, 0xff, &gpio);
-	if (ret == 0) {
-		ret = pm->pwm_get(dev, gpio.line, &divs, &duty);
-		if (ret == 0 && divs) {
-			divs = max(divs, duty);
-			if (dev_priv->card_type <= NV_40 || (gpio.log[0] & 1))
-				duty = divs - duty;
-			return (duty * 100) / divs;
-		}
-
-		return nouveau_gpio_func_get(dev, gpio.func) * 100;
-	}
-
-	return -ENODEV;
-}
-
-static int
-nouveau_pwmfan_set(struct drm_device *dev, int percent)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
-	struct gpio_func gpio;
-	u32 divs, duty;
-	int ret;
-
-	if (!pm->pwm_set)
-		return -ENODEV;
-
-	ret = nouveau_gpio_find(dev, 0, DCB_GPIO_PWM_FAN, 0xff, &gpio);
-	if (ret == 0) {
-		divs = pm->fan.pwm_divisor;
-		if (pm->fan.pwm_freq) {
-			/*XXX: PNVIO clock more than likely... */
-			divs = 135000 / pm->fan.pwm_freq;
-			if (dev_priv->chipset < 0xa3)
-				divs /= 4;
-		}
-
-		duty = ((divs * percent) + 99) / 100;
-		if (dev_priv->card_type <= NV_40 || (gpio.log[0] & 1))
-			duty = divs - duty;
-
-		ret = pm->pwm_set(dev, gpio.line, divs, duty);
-		if (!ret)
-			pm->fan.percent = percent;
-		return ret;
-	}
-
-	return -ENODEV;
-}
-
-static int
-nouveau_pm_perflvl_aux(struct drm_device *dev, struct nouveau_pm_level *perflvl,
-		       struct nouveau_pm_level *a, struct nouveau_pm_level *b)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
-	int ret;
+	void *pre_state;
 
-	/*XXX: not on all boards, we should control based on temperature
-	 *     on recent boards..  or maybe on some other factor we don't
-	 *     know about?
-	 */
-	if (a->fanspeed && b->fanspeed && b->fanspeed > a->fanspeed) {
-		ret = nouveau_pwmfan_set(dev, perflvl->fanspeed);
-		if (ret && ret != -ENODEV) {
-			NV_ERROR(dev, "fanspeed set failed: %d\n", ret);
-			return ret;
-		}
-	}
+	if (khz == 0)
+		return 0;
 
-	if (pm->voltage.supported && pm->voltage_set) {
-		if (perflvl->volt_min && b->volt_min > a->volt_min) {
-			ret = pm->voltage_set(dev, perflvl->volt_min);
-			if (ret) {
-				NV_ERROR(dev, "voltage set failed: %d\n", ret);
-				return ret;
-			}
-		}
-	}
+	pre_state = pm->clock_pre(dev, perflvl, id, khz);
+	if (IS_ERR(pre_state))
+		return PTR_ERR(pre_state);
 
+	if (pre_state)
+		pm->clock_set(dev, pre_state);
 	return 0;
 }
 
@@ -136,90 +59,26 @@ nouveau_pm_perflvl_set(struct drm_device *dev, struct nouveau_pm_level *perflvl)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
-	void *state;
 	int ret;
 
 	if (perflvl == pm->cur)
 		return 0;
 
-	ret = nouveau_pm_perflvl_aux(dev, perflvl, pm->cur, perflvl);
-	if (ret)
-		return ret;
-
-	state = pm->clocks_pre(dev, perflvl);
-	if (IS_ERR(state)) {
-		ret = PTR_ERR(state);
-		goto error;
+	if (pm->voltage.supported && pm->voltage_set && perflvl->voltage) {
+		ret = pm->voltage_set(dev, perflvl->voltage);
+		if (ret) {
+			NV_ERROR(dev, "voltage_set %d failed: %d\n",
+				 perflvl->voltage, ret);
+		}
 	}
-	ret = pm->clocks_set(dev, state);
-	if (ret)
-		goto error;
 
-	ret = nouveau_pm_perflvl_aux(dev, perflvl, perflvl, pm->cur);
-	if (ret)
-		return ret;
+	nouveau_pm_clock_set(dev, perflvl, PLL_CORE, perflvl->core);
+	nouveau_pm_clock_set(dev, perflvl, PLL_SHADER, perflvl->shader);
+	nouveau_pm_clock_set(dev, perflvl, PLL_MEMORY, perflvl->memory);
+	nouveau_pm_clock_set(dev, perflvl, PLL_UNK05, perflvl->unk05);
 
 	pm->cur = perflvl;
 	return 0;
-
-error:
-	/* restore the fan speed and voltage before leaving */
-	nouveau_pm_perflvl_aux(dev, perflvl, perflvl, pm->cur);
-	return ret;
-}
-
-void
-nouveau_pm_trigger(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
-	struct nouveau_pm_profile *profile = NULL;
-	struct nouveau_pm_level *perflvl = NULL;
-	int ret;
-
-	/* select power profile based on current power source */
-	if (power_supply_is_system_supplied())
-		profile = pm->profile_ac;
-	else
-		profile = pm->profile_dc;
-
-	if (profile != pm->profile) {
-		pm->profile->func->fini(pm->profile);
-		pm->profile = profile;
-		pm->profile->func->init(pm->profile);
-	}
-
-	/* select performance level based on profile */
-	perflvl = profile->func->select(profile);
-
-	/* change perflvl, if necessary */
-	if (perflvl != pm->cur) {
-		struct nouveau_timer_engine *ptimer = &dev_priv->engine.timer;
-		u64 time0 = ptimer->read(dev);
-
-		NV_INFO(dev, "setting performance level: %d", perflvl->id);
-		ret = nouveau_pm_perflvl_set(dev, perflvl);
-		if (ret)
-			NV_INFO(dev, "> reclocking failed: %d\n\n", ret);
-
-		NV_INFO(dev, "> reclocking took %lluns\n\n",
-			     ptimer->read(dev) - time0);
-	}
-}
-
-static struct nouveau_pm_profile *
-profile_find(struct drm_device *dev, const char *string)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
-	struct nouveau_pm_profile *profile;
-
-	list_for_each_entry(profile, &pm->profiles, head) {
-		if (!strncmp(profile->name, string, sizeof(profile->name)))
-			return profile;
-	}
-
-	return NULL;
 }
 
 static int
@@ -227,53 +86,35 @@ nouveau_pm_profile_set(struct drm_device *dev, const char *profile)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
-	struct nouveau_pm_profile *ac = NULL, *dc = NULL;
-	char string[16], *cur = string, *ptr;
+	struct nouveau_pm_level *perflvl = NULL;
 
 	/* safety precaution, for now */
 	if (nouveau_perflvl_wr != 7777)
 		return -EPERM;
 
-	strncpy(string, profile, sizeof(string));
-	if ((ptr = strchr(string, '\n')))
-		*ptr = '\0';
-
-	ptr = strsep(&cur, ",");
-	if (ptr)
-		ac = profile_find(dev, ptr);
-
-	ptr = strsep(&cur, ",");
-	if (ptr)
-		dc = profile_find(dev, ptr);
-	else
-		dc = ac;
-
-	if (ac == NULL || dc == NULL)
+	if (!pm->clock_set)
 		return -EINVAL;
 
-	pm->profile_ac = ac;
-	pm->profile_dc = dc;
-	nouveau_pm_trigger(dev);
-	return 0;
-}
+	if (!strncmp(profile, "boot", 4))
+		perflvl = &pm->boot;
+	else {
+		int pl = simple_strtol(profile, NULL, 10);
+		int i;
 
-static void
-nouveau_pm_static_dummy(struct nouveau_pm_profile *profile)
-{
-}
+		for (i = 0; i < pm->nr_perflvl; i++) {
+			if (pm->perflvl[i].id == pl) {
+				perflvl = &pm->perflvl[i];
+				break;
+			}
+		}
 
-static struct nouveau_pm_level *
-nouveau_pm_static_select(struct nouveau_pm_profile *profile)
-{
-	return container_of(profile, struct nouveau_pm_level, profile);
-}
+		if (!perflvl)
+			return -EINVAL;
+	}
 
-const struct nouveau_pm_profile_func nouveau_pm_static_profile_func = {
-	.destroy = nouveau_pm_static_dummy,
-	.init = nouveau_pm_static_dummy,
-	.fini = nouveau_pm_static_dummy,
-	.select = nouveau_pm_static_select,
-};
+	NV_INFO(dev, "setting performance level: %s\n", profile);
+	return nouveau_pm_perflvl_set(dev, perflvl);
+}
 
 static int
 nouveau_pm_perflvl_get(struct drm_device *dev, struct nouveau_pm_level *perflvl)
@@ -282,34 +123,40 @@ nouveau_pm_perflvl_get(struct drm_device *dev, struct nouveau_pm_level *perflvl)
 	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
 	int ret;
 
+	if (!pm->clock_get)
+		return -EINVAL;
+
 	memset(perflvl, 0, sizeof(*perflvl));
 
-	if (pm->clocks_get) {
-		ret = pm->clocks_get(dev, perflvl);
-		if (ret)
-			return ret;
-	}
+	ret = pm->clock_get(dev, PLL_CORE);
+	if (ret > 0)
+		perflvl->core = ret;
+
+	ret = pm->clock_get(dev, PLL_MEMORY);
+	if (ret > 0)
+		perflvl->memory = ret;
+
+	ret = pm->clock_get(dev, PLL_SHADER);
+	if (ret > 0)
+		perflvl->shader = ret;
+
+	ret = pm->clock_get(dev, PLL_UNK05);
+	if (ret > 0)
+		perflvl->unk05 = ret;
 
 	if (pm->voltage.supported && pm->voltage_get) {
 		ret = pm->voltage_get(dev);
-		if (ret > 0) {
-			perflvl->volt_min = ret;
-			perflvl->volt_max = ret;
-		}
+		if (ret > 0)
+			perflvl->voltage = ret;
 	}
 
-	ret = nouveau_pwmfan_get(dev);
-	if (ret > 0)
-		perflvl->fanspeed = ret;
-
-	nouveau_mem_timing_read(dev, &perflvl->timing);
 	return 0;
 }
 
 static void
 nouveau_pm_perflvl_info(struct nouveau_pm_level *perflvl, char *ptr, int len)
 {
-	char c[16], s[16], v[32], f[16], m[16];
+	char c[16], s[16], v[16], f[16], t[16];
 
 	c[0] = '\0';
 	if (perflvl->core)
@@ -319,37 +166,31 @@ nouveau_pm_perflvl_info(struct nouveau_pm_level *perflvl, char *ptr, int len)
 	if (perflvl->shader)
 		snprintf(s, sizeof(s), " shader %dMHz", perflvl->shader / 1000);
 
-	m[0] = '\0';
-	if (perflvl->memory)
-		snprintf(m, sizeof(m), " memory %dMHz", perflvl->memory / 1000);
-
 	v[0] = '\0';
-	if (perflvl->volt_min && perflvl->volt_min != perflvl->volt_max) {
-		snprintf(v, sizeof(v), " voltage %dmV-%dmV",
-			 perflvl->volt_min / 1000, perflvl->volt_max / 1000);
-	} else
-	if (perflvl->volt_min) {
-		snprintf(v, sizeof(v), " voltage %dmV",
-			 perflvl->volt_min / 1000);
-	}
+	if (perflvl->voltage)
+		snprintf(v, sizeof(v), " voltage %dmV", perflvl->voltage * 10);
 
 	f[0] = '\0';
 	if (perflvl->fanspeed)
 		snprintf(f, sizeof(f), " fanspeed %d%%", perflvl->fanspeed);
 
-	snprintf(ptr, len, "%s%s%s%s%s\n", c, s, m, v, f);
+	t[0] = '\0';
+	if (perflvl->timing)
+		snprintf(t, sizeof(t), " timing %d", perflvl->timing->id);
+
+	snprintf(ptr, len, "memory %dMHz%s%s%s%s%s\n", perflvl->memory / 1000,
+		 c, s, v, f, t);
 }
 
 static ssize_t
 nouveau_pm_get_perflvl_info(struct device *d,
 			    struct device_attribute *a, char *buf)
 {
-	struct nouveau_pm_level *perflvl =
-		container_of(a, struct nouveau_pm_level, dev_attr);
+	struct nouveau_pm_level *perflvl = (struct nouveau_pm_level *)a;
 	char *ptr = buf;
 	int len = PAGE_SIZE;
 
-	snprintf(ptr, len, "%d:", perflvl->id);
+	snprintf(ptr, len, "%d: ", perflvl->id);
 	ptr += strlen(buf);
 	len -= strlen(buf);
 
@@ -367,8 +208,12 @@ nouveau_pm_get_perflvl(struct device *d, struct device_attribute *a, char *buf)
 	int len = PAGE_SIZE, ret;
 	char *ptr = buf;
 
-	snprintf(ptr, len, "profile: %s, %s\nc:",
-		 pm->profile_ac->name, pm->profile_dc->name);
+	if (!pm->cur)
+		snprintf(ptr, len, "setting: boot\n");
+	else if (pm->cur == &pm->boot)
+		snprintf(ptr, len, "setting: boot\nc: ");
+	else
+		snprintf(ptr, len, "setting: static %d\nc: ", pm->cur->id);
 	ptr += strlen(buf);
 	len -= strlen(buf);
 
@@ -447,7 +292,7 @@ nouveau_sysfs_fini(struct drm_device *dev)
 	}
 }
 
-#if defined(CONFIG_HWMON) || (defined(MODULE) && defined(CONFIG_HWMON_MODULE))
+#ifdef CONFIG_HWMON
 static ssize_t
 nouveau_hwmon_show_temp(struct device *d, struct device_attribute *a, char *buf)
 {
@@ -480,7 +325,7 @@ nouveau_hwmon_set_max_temp(struct device *d, struct device_attribute *a,
 	struct nouveau_pm_threshold_temp *temp = &pm->threshold_temp;
 	long value;
 
-	if (kstrtol(buf, 10, &value) == -EINVAL)
+	if (strict_strtol(buf, 10, &value) == -EINVAL)
 		return count;
 
 	temp->down_clock = value/1000;
@@ -515,7 +360,7 @@ nouveau_hwmon_set_critical_temp(struct device *d, struct device_attribute *a,
 	struct nouveau_pm_threshold_temp *temp = &pm->threshold_temp;
 	long value;
 
-	if (kstrtol(buf, 10, &value) == -EINVAL)
+	if (strict_strtol(buf, 10, &value) == -EINVAL)
 		return count;
 
 	temp->critical = value/1000;
@@ -547,172 +392,6 @@ static SENSOR_DEVICE_ATTR(update_rate, S_IRUGO,
 						nouveau_hwmon_show_update_rate,
 						NULL, 0);
 
-static ssize_t
-nouveau_hwmon_show_fan0_input(struct device *d, struct device_attribute *attr,
-			      char *buf)
-{
-	struct drm_device *dev = dev_get_drvdata(d);
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_timer_engine *ptimer = &dev_priv->engine.timer;
-	struct gpio_func gpio;
-	u32 cycles, cur, prev;
-	u64 start;
-	int ret;
-
-	ret = nouveau_gpio_find(dev, 0, DCB_GPIO_FAN_SENSE, 0xff, &gpio);
-	if (ret)
-		return ret;
-
-	/* Monitor the GPIO input 0x3b for 250ms.
-	 * When the fan spins, it changes the value of GPIO FAN_SENSE.
-	 * We get 4 changes (0 -> 1 -> 0 -> 1 -> [...]) per complete rotation.
-	 */
-	start = ptimer->read(dev);
-	prev = nouveau_gpio_sense(dev, 0, gpio.line);
-	cycles = 0;
-	do {
-		cur = nouveau_gpio_sense(dev, 0, gpio.line);
-		if (prev != cur) {
-			cycles++;
-			prev = cur;
-		}
-
-		usleep_range(500, 1000); /* supports 0 < rpm < 7500 */
-	} while (ptimer->read(dev) - start < 250000000);
-
-	/* interpolate to get rpm */
-	return sprintf(buf, "%i\n", cycles / 4 * 4 * 60);
-}
-static SENSOR_DEVICE_ATTR(fan0_input, S_IRUGO, nouveau_hwmon_show_fan0_input,
-			  NULL, 0);
-
-static ssize_t
-nouveau_hwmon_get_pwm0(struct device *d, struct device_attribute *a, char *buf)
-{
-	struct drm_device *dev = dev_get_drvdata(d);
-	int ret;
-
-	ret = nouveau_pwmfan_get(dev);
-	if (ret < 0)
-		return ret;
-
-	return sprintf(buf, "%i\n", ret);
-}
-
-static ssize_t
-nouveau_hwmon_set_pwm0(struct device *d, struct device_attribute *a,
-		       const char *buf, size_t count)
-{
-	struct drm_device *dev = dev_get_drvdata(d);
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
-	int ret = -ENODEV;
-	long value;
-
-	if (nouveau_perflvl_wr != 7777)
-		return -EPERM;
-
-	if (kstrtol(buf, 10, &value) == -EINVAL)
-		return -EINVAL;
-
-	if (value < pm->fan.min_duty)
-		value = pm->fan.min_duty;
-	if (value > pm->fan.max_duty)
-		value = pm->fan.max_duty;
-
-	ret = nouveau_pwmfan_set(dev, value);
-	if (ret)
-		return ret;
-
-	return count;
-}
-
-static SENSOR_DEVICE_ATTR(pwm0, S_IRUGO | S_IWUSR,
-			  nouveau_hwmon_get_pwm0,
-			  nouveau_hwmon_set_pwm0, 0);
-
-static ssize_t
-nouveau_hwmon_get_pwm0_min(struct device *d,
-			   struct device_attribute *a, char *buf)
-{
-	struct drm_device *dev = dev_get_drvdata(d);
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
-
-	return sprintf(buf, "%i\n", pm->fan.min_duty);
-}
-
-static ssize_t
-nouveau_hwmon_set_pwm0_min(struct device *d, struct device_attribute *a,
-			   const char *buf, size_t count)
-{
-	struct drm_device *dev = dev_get_drvdata(d);
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
-	long value;
-
-	if (kstrtol(buf, 10, &value) == -EINVAL)
-		return -EINVAL;
-
-	if (value < 0)
-		value = 0;
-
-	if (pm->fan.max_duty - value < 10)
-		value = pm->fan.max_duty - 10;
-
-	if (value < 10)
-		pm->fan.min_duty = 10;
-	else
-		pm->fan.min_duty = value;
-
-	return count;
-}
-
-static SENSOR_DEVICE_ATTR(pwm0_min, S_IRUGO | S_IWUSR,
-			  nouveau_hwmon_get_pwm0_min,
-			  nouveau_hwmon_set_pwm0_min, 0);
-
-static ssize_t
-nouveau_hwmon_get_pwm0_max(struct device *d,
-			   struct device_attribute *a, char *buf)
-{
-	struct drm_device *dev = dev_get_drvdata(d);
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
-
-	return sprintf(buf, "%i\n", pm->fan.max_duty);
-}
-
-static ssize_t
-nouveau_hwmon_set_pwm0_max(struct device *d, struct device_attribute *a,
-			   const char *buf, size_t count)
-{
-	struct drm_device *dev = dev_get_drvdata(d);
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
-	long value;
-
-	if (kstrtol(buf, 10, &value) == -EINVAL)
-		return -EINVAL;
-
-	if (value < 0)
-		value = 0;
-
-	if (value - pm->fan.min_duty < 10)
-		value = pm->fan.min_duty + 10;
-
-	if (value > 100)
-		pm->fan.max_duty = 100;
-	else
-		pm->fan.max_duty = value;
-
-	return count;
-}
-
-static SENSOR_DEVICE_ATTR(pwm0_max, S_IRUGO | S_IWUSR,
-			  nouveau_hwmon_get_pwm0_max,
-			  nouveau_hwmon_set_pwm0_max, 0);
-
 static struct attribute *hwmon_attributes[] = {
 	&sensor_dev_attr_temp1_input.dev_attr.attr,
 	&sensor_dev_attr_temp1_max.dev_attr.attr,
@@ -721,36 +400,20 @@ static struct attribute *hwmon_attributes[] = {
 	&sensor_dev_attr_update_rate.dev_attr.attr,
 	NULL
 };
-static struct attribute *hwmon_fan_rpm_attributes[] = {
-	&sensor_dev_attr_fan0_input.dev_attr.attr,
-	NULL
-};
-static struct attribute *hwmon_pwm_fan_attributes[] = {
-	&sensor_dev_attr_pwm0.dev_attr.attr,
-	&sensor_dev_attr_pwm0_min.dev_attr.attr,
-	&sensor_dev_attr_pwm0_max.dev_attr.attr,
-	NULL
-};
 
 static const struct attribute_group hwmon_attrgroup = {
 	.attrs = hwmon_attributes,
 };
-static const struct attribute_group hwmon_fan_rpm_attrgroup = {
-	.attrs = hwmon_fan_rpm_attributes,
-};
-static const struct attribute_group hwmon_pwm_fan_attrgroup = {
-	.attrs = hwmon_pwm_fan_attributes,
-};
 #endif
 
 static int
 nouveau_hwmon_init(struct drm_device *dev)
 {
+#ifdef CONFIG_HWMON
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
-#if defined(CONFIG_HWMON) || (defined(MODULE) && defined(CONFIG_HWMON_MODULE))
 	struct device *hwmon_dev;
-	int ret = 0;
+	int ret;
 
 	if (!pm->temp_get)
 		return -ENODEV;
@@ -763,62 +426,28 @@ nouveau_hwmon_init(struct drm_device *dev)
 		return ret;
 	}
 	dev_set_drvdata(hwmon_dev, dev);
-
-	/* default sysfs entries */
 	ret = sysfs_create_group(&dev->pdev->dev.kobj, &hwmon_attrgroup);
 	if (ret) {
-		if (ret)
-			goto error;
-	}
-
-	/* if the card has a pwm fan */
-	/*XXX: incorrect, need better detection for this, some boards have
-	 *     the gpio entries for pwm fan control even when there's no
-	 *     actual fan connected to it... therm table? */
-	if (nouveau_pwmfan_get(dev) >= 0) {
-		ret = sysfs_create_group(&dev->pdev->dev.kobj,
-					 &hwmon_pwm_fan_attrgroup);
-		if (ret)
-			goto error;
-	}
-
-	/* if the card can read the fan rpm */
-	if (nouveau_gpio_func_valid(dev, DCB_GPIO_FAN_SENSE)) {
-		ret = sysfs_create_group(&dev->pdev->dev.kobj,
-					 &hwmon_fan_rpm_attrgroup);
-		if (ret)
-			goto error;
+		NV_ERROR(dev,
+			"Unable to create hwmon sysfs file: %d\n", ret);
+		hwmon_device_unregister(hwmon_dev);
+		return ret;
 	}
 
 	pm->hwmon = hwmon_dev;
-
-	return 0;
-
-error:
-	NV_ERROR(dev, "Unable to create some hwmon sysfs files: %d\n", ret);
-	hwmon_device_unregister(hwmon_dev);
-	pm->hwmon = NULL;
-	return ret;
-#else
-	pm->hwmon = NULL;
-	return 0;
 #endif
+	return 0;
 }
 
 static void
 nouveau_hwmon_fini(struct drm_device *dev)
 {
-#if defined(CONFIG_HWMON) || (defined(MODULE) && defined(CONFIG_HWMON_MODULE))
+#ifdef CONFIG_HWMON
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
 
 	if (pm->hwmon) {
 		sysfs_remove_group(&dev->pdev->dev.kobj, &hwmon_attrgroup);
-		sysfs_remove_group(&dev->pdev->dev.kobj,
-				   &hwmon_pwm_fan_attrgroup);
-		sysfs_remove_group(&dev->pdev->dev.kobj,
-				   &hwmon_fan_rpm_attrgroup);
-
 		hwmon_device_unregister(pm->hwmon);
 	}
 #endif
@@ -837,7 +466,6 @@ nouveau_pm_acpi_event(struct notifier_block *nb, unsigned long val, void *data)
 		bool ac = power_supply_is_system_supplied();
 
 		NV_DEBUG(dev, "power supply changed: %s\n", ac ? "AC" : "DC");
-		nouveau_pm_trigger(dev);
 	}
 
 	return NOTIFY_OK;
@@ -852,48 +480,35 @@ nouveau_pm_init(struct drm_device *dev)
 	char info[256];
 	int ret, i;
 
-	/* parse aux tables from vbios */
+	nouveau_mem_timing_init(dev);
 	nouveau_volt_init(dev);
-	nouveau_temp_init(dev);
-
-	/* determine current ("boot") performance level */
-	ret = nouveau_pm_perflvl_get(dev, &pm->boot);
-	if (ret) {
-		NV_ERROR(dev, "failed to determine boot perflvl\n");
-		return ret;
-	}
-
-	strncpy(pm->boot.name, "boot", 4);
-	strncpy(pm->boot.profile.name, "boot", 4);
-	pm->boot.profile.func = &nouveau_pm_static_profile_func;
-
-	INIT_LIST_HEAD(&pm->profiles);
-	list_add(&pm->boot.profile.head, &pm->profiles);
-
-	pm->profile_ac = &pm->boot.profile;
-	pm->profile_dc = &pm->boot.profile;
-	pm->profile = &pm->boot.profile;
-	pm->cur = &pm->boot;
-
-	/* add performance levels from vbios */
 	nouveau_perf_init(dev);
+	nouveau_temp_init(dev);
 
-	/* display available performance levels */
 	NV_INFO(dev, "%d available performance level(s)\n", pm->nr_perflvl);
 	for (i = 0; i < pm->nr_perflvl; i++) {
 		nouveau_pm_perflvl_info(&pm->perflvl[i], info, sizeof(info));
-		NV_INFO(dev, "%d:%s", pm->perflvl[i].id, info);
+		NV_INFO(dev, "%d: %s", pm->perflvl[i].id, info);
 	}
 
-	nouveau_pm_perflvl_info(&pm->boot, info, sizeof(info));
-	NV_INFO(dev, "c:%s", info);
+	/* determine current ("boot") performance level */
+	ret = nouveau_pm_perflvl_get(dev, &pm->boot);
+	if (ret == 0) {
+		strncpy(pm->boot.name, "boot", 4);
+		pm->cur = &pm->boot;
 
-	/* switch performance levels now if requested */
-	if (nouveau_perflvl != NULL)
-		nouveau_pm_profile_set(dev, nouveau_perflvl);
+		nouveau_pm_perflvl_info(&pm->boot, info, sizeof(info));
+		NV_INFO(dev, "c: %s", info);
+	}
 
-	/* determine the current fan speed */
-	pm->fan.percent = nouveau_pwmfan_get(dev);
+	/* switch performance levels now if requested */
+	if (nouveau_perflvl != NULL) {
+		ret = nouveau_pm_profile_set(dev, nouveau_perflvl);
+		if (ret) {
+			NV_ERROR(dev, "error setting perflvl \"%s\": %d\n",
+				 nouveau_perflvl, ret);
+		}
+	}
 
 	nouveau_sysfs_init(dev);
 	nouveau_hwmon_init(dev);
@@ -910,12 +525,6 @@ nouveau_pm_fini(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
-	struct nouveau_pm_profile *profile, *tmp;
-
-	list_for_each_entry_safe(profile, tmp, &pm->profiles, head) {
-		list_del(&profile->head);
-		profile->func->destroy(profile);
-	}
 
 	if (pm->cur != &pm->boot)
 		nouveau_pm_perflvl_set(dev, &pm->boot);
@@ -923,6 +532,7 @@ nouveau_pm_fini(struct drm_device *dev)
 	nouveau_temp_fini(dev);
 	nouveau_perf_fini(dev);
 	nouveau_volt_fini(dev);
+	nouveau_mem_timing_fini(dev);
 
 #if defined(CONFIG_ACPI) && defined(CONFIG_POWER_SUPPLY)
 	unregister_acpi_notifier(&pm->acpi_nb);
@@ -944,5 +554,4 @@ nouveau_pm_resume(struct drm_device *dev)
 	perflvl = pm->cur;
 	pm->cur = &pm->boot;
 	nouveau_pm_perflvl_set(dev, perflvl);
-	nouveau_pwmfan_set(dev, pm->fan.percent);
 }
diff --git a/drivers/gpu/drm/nouveau/nouveau_pm.h b/drivers/gpu/drm/nouveau/nouveau_pm.h
index 3f82dfe..4a9838dd 100644
--- a/drivers/gpu/drm/nouveau/nouveau_pm.h
+++ b/drivers/gpu/drm/nouveau/nouveau_pm.h
@@ -25,30 +25,10 @@
 #ifndef __NOUVEAU_PM_H__
 #define __NOUVEAU_PM_H__
 
-struct nouveau_mem_exec_func {
-	struct drm_device *dev;
-	void (*precharge)(struct nouveau_mem_exec_func *);
-	void (*refresh)(struct nouveau_mem_exec_func *);
-	void (*refresh_auto)(struct nouveau_mem_exec_func *, bool);
-	void (*refresh_self)(struct nouveau_mem_exec_func *, bool);
-	void (*wait)(struct nouveau_mem_exec_func *, u32 nsec);
-	u32  (*mrg)(struct nouveau_mem_exec_func *, int mr);
-	void (*mrs)(struct nouveau_mem_exec_func *, int mr, u32 data);
-	void (*clock_set)(struct nouveau_mem_exec_func *);
-	void (*timing_set)(struct nouveau_mem_exec_func *);
-	void *priv;
-};
-
-/* nouveau_mem.c */
-int  nouveau_mem_exec(struct nouveau_mem_exec_func *,
-		      struct nouveau_pm_level *);
-
 /* nouveau_pm.c */
 int  nouveau_pm_init(struct drm_device *dev);
 void nouveau_pm_fini(struct drm_device *dev);
 void nouveau_pm_resume(struct drm_device *dev);
-extern const struct nouveau_pm_profile_func nouveau_pm_static_profile_func;
-void nouveau_pm_trigger(struct drm_device *dev);
 
 /* nouveau_volt.c */
 void nouveau_volt_init(struct drm_device *);
@@ -61,41 +41,28 @@ int  nouveau_voltage_gpio_set(struct drm_device *, int voltage);
 /* nouveau_perf.c */
 void nouveau_perf_init(struct drm_device *);
 void nouveau_perf_fini(struct drm_device *);
-u8 *nouveau_perf_timing(struct drm_device *, u32 freq, u8 *ver, u8 *len);
-u8 *nouveau_perf_ramcfg(struct drm_device *, u32 freq, u8 *ver, u8 *len);
 
 /* nouveau_mem.c */
 void nouveau_mem_timing_init(struct drm_device *);
 void nouveau_mem_timing_fini(struct drm_device *);
 
 /* nv04_pm.c */
-int nv04_pm_clocks_get(struct drm_device *, struct nouveau_pm_level *);
-void *nv04_pm_clocks_pre(struct drm_device *, struct nouveau_pm_level *);
-int nv04_pm_clocks_set(struct drm_device *, void *);
-
-/* nv40_pm.c */
-int nv40_pm_clocks_get(struct drm_device *, struct nouveau_pm_level *);
-void *nv40_pm_clocks_pre(struct drm_device *, struct nouveau_pm_level *);
-int nv40_pm_clocks_set(struct drm_device *, void *);
-int nv40_pm_pwm_get(struct drm_device *, int, u32 *, u32 *);
-int nv40_pm_pwm_set(struct drm_device *, int, u32, u32);
+int nv04_pm_clock_get(struct drm_device *, u32 id);
+void *nv04_pm_clock_pre(struct drm_device *, struct nouveau_pm_level *,
+			u32 id, int khz);
+void nv04_pm_clock_set(struct drm_device *, void *);
 
 /* nv50_pm.c */
-int nv50_pm_clocks_get(struct drm_device *, struct nouveau_pm_level *);
-void *nv50_pm_clocks_pre(struct drm_device *, struct nouveau_pm_level *);
-int nv50_pm_clocks_set(struct drm_device *, void *);
-int nv50_pm_pwm_get(struct drm_device *, int, u32 *, u32 *);
-int nv50_pm_pwm_set(struct drm_device *, int, u32, u32);
+int nv50_pm_clock_get(struct drm_device *, u32 id);
+void *nv50_pm_clock_pre(struct drm_device *, struct nouveau_pm_level *,
+			u32 id, int khz);
+void nv50_pm_clock_set(struct drm_device *, void *);
 
 /* nva3_pm.c */
-int nva3_pm_clocks_get(struct drm_device *, struct nouveau_pm_level *);
-void *nva3_pm_clocks_pre(struct drm_device *, struct nouveau_pm_level *);
-int nva3_pm_clocks_set(struct drm_device *, void *);
-
-/* nvc0_pm.c */
-int nvc0_pm_clocks_get(struct drm_device *, struct nouveau_pm_level *);
-void *nvc0_pm_clocks_pre(struct drm_device *, struct nouveau_pm_level *);
-int nvc0_pm_clocks_set(struct drm_device *, void *);
+int nva3_pm_clock_get(struct drm_device *, u32 id);
+void *nva3_pm_clock_pre(struct drm_device *, struct nouveau_pm_level *,
+			u32 id, int khz);
+void nva3_pm_clock_set(struct drm_device *, void *);
 
 /* nouveau_temp.c */
 void nouveau_temp_init(struct drm_device *dev);
diff --git a/drivers/gpu/drm/nouveau/nouveau_reg.h b/drivers/gpu/drm/nouveau/nouveau_reg.h
index 43a96b9..f18cdfc 100644
--- a/drivers/gpu/drm/nouveau/nouveau_reg.h
+++ b/drivers/gpu/drm/nouveau/nouveau_reg.h
@@ -826,12 +826,9 @@
 #define NV50_PDISPLAY_SOR_DPMS_STATE_ACTIVE                          0x00030000
 #define NV50_PDISPLAY_SOR_DPMS_STATE_BLANKED                         0x00080000
 #define NV50_PDISPLAY_SOR_DPMS_STATE_WAIT                            0x10000000
-#define NV50_PDISP_SOR_PWM_DIV(i)                     (0x0061c080 + (i) * 0x800)
-#define NV50_PDISP_SOR_PWM_CTL(i)                     (0x0061c084 + (i) * 0x800)
-#define NV50_PDISP_SOR_PWM_CTL_NEW                                   0x80000000
-#define NVA3_PDISP_SOR_PWM_CTL_UNK                                   0x40000000
-#define NV50_PDISP_SOR_PWM_CTL_VAL                                   0x000007ff
-#define NVA3_PDISP_SOR_PWM_CTL_VAL                                   0x00ffffff
+#define NV50_PDISPLAY_SOR_BACKLIGHT                                  0x0061c084
+#define NV50_PDISPLAY_SOR_BACKLIGHT_ENABLE                           0x80000000
+#define NV50_PDISPLAY_SOR_BACKLIGHT_LEVEL                            0x00000fff
 #define NV50_SOR_DP_CTRL(i, l)           (0x0061c10c + (i) * 0x800 + (l) * 0x80)
 #define NV50_SOR_DP_CTRL_ENABLED                                     0x00000001
 #define NV50_SOR_DP_CTRL_ENHANCED_FRAME_ENABLED                      0x00004000
@@ -846,7 +843,7 @@
 #define NV50_SOR_DP_CTRL_TRAINING_PATTERN_2                          0x02000000
 #define NV50_SOR_DP_UNK118(i, l)         (0x0061c118 + (i) * 0x800 + (l) * 0x80)
 #define NV50_SOR_DP_UNK120(i, l)         (0x0061c120 + (i) * 0x800 + (l) * 0x80)
-#define NV50_SOR_DP_SCFG(i, l)           (0x0061c128 + (i) * 0x800 + (l) * 0x80)
+#define NV50_SOR_DP_UNK128(i, l)         (0x0061c128 + (i) * 0x800 + (l) * 0x80)
 #define NV50_SOR_DP_UNK130(i, l)         (0x0061c130 + (i) * 0x800 + (l) * 0x80)
 
 #define NV50_PDISPLAY_USER(i)                        ((i) * 0x1000 + 0x00640000)
diff --git a/drivers/gpu/drm/nouveau/nouveau_sgdma.c b/drivers/gpu/drm/nouveau/nouveau_sgdma.c
index 47f245e..ca6028f 100644
--- a/drivers/gpu/drm/nouveau/nouveau_sgdma.c
+++ b/drivers/gpu/drm/nouveau/nouveau_sgdma.c
@@ -8,30 +8,112 @@
 #define NV_CTXDMA_PAGE_MASK  (NV_CTXDMA_PAGE_SIZE - 1)
 
 struct nouveau_sgdma_be {
-	/* this has to be the first field so populate/unpopulated in
-	 * nouve_bo.c works properly, otherwise have to move them here
-	 */
-	struct ttm_dma_tt ttm;
+	struct ttm_backend backend;
 	struct drm_device *dev;
+
+	dma_addr_t *pages;
+	bool *ttm_alloced;
+	unsigned nr_pages;
+
 	u64 offset;
+	bool bound;
 };
 
+static int
+nouveau_sgdma_populate(struct ttm_backend *be, unsigned long num_pages,
+		       struct page **pages, struct page *dummy_read_page,
+		       dma_addr_t *dma_addrs)
+{
+	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be;
+	struct drm_device *dev = nvbe->dev;
+
+	NV_DEBUG(nvbe->dev, "num_pages = %ld\n", num_pages);
+
+	if (nvbe->pages)
+		return -EINVAL;
+
+	nvbe->pages = kmalloc(sizeof(dma_addr_t) * num_pages, GFP_KERNEL);
+	if (!nvbe->pages)
+		return -ENOMEM;
+
+	nvbe->ttm_alloced = kmalloc(sizeof(bool) * num_pages, GFP_KERNEL);
+	if (!nvbe->ttm_alloced) {
+		kfree(nvbe->pages);
+		nvbe->pages = NULL;
+		return -ENOMEM;
+	}
+
+	nvbe->nr_pages = 0;
+	while (num_pages--) {
+		/* this code path isn't called and is incorrect anyways */
+		if (0) { /*dma_addrs[nvbe->nr_pages] != DMA_ERROR_CODE)*/
+			nvbe->pages[nvbe->nr_pages] =
+					dma_addrs[nvbe->nr_pages];
+		 	nvbe->ttm_alloced[nvbe->nr_pages] = true;
+		} else {
+			nvbe->pages[nvbe->nr_pages] =
+				pci_map_page(dev->pdev, pages[nvbe->nr_pages], 0,
+				     PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
+			if (pci_dma_mapping_error(dev->pdev,
+						  nvbe->pages[nvbe->nr_pages])) {
+				be->func->clear(be);
+				return -EFAULT;
+			}
+			nvbe->ttm_alloced[nvbe->nr_pages] = false;
+		}
+
+		nvbe->nr_pages++;
+	}
+
+	return 0;
+}
+
 static void
-nouveau_sgdma_destroy(struct ttm_tt *ttm)
+nouveau_sgdma_clear(struct ttm_backend *be)
 {
-	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)ttm;
+	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be;
+	struct drm_device *dev;
+
+	if (nvbe && nvbe->pages) {
+		dev = nvbe->dev;
+		NV_DEBUG(dev, "\n");
+
+		if (nvbe->bound)
+			be->func->unbind(be);
+
+		while (nvbe->nr_pages--) {
+			if (!nvbe->ttm_alloced[nvbe->nr_pages])
+				pci_unmap_page(dev->pdev, nvbe->pages[nvbe->nr_pages],
+				       PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
+		}
+		kfree(nvbe->pages);
+		kfree(nvbe->ttm_alloced);
+		nvbe->pages = NULL;
+		nvbe->ttm_alloced = NULL;
+		nvbe->nr_pages = 0;
+	}
+}
 
-	if (ttm) {
+static void
+nouveau_sgdma_destroy(struct ttm_backend *be)
+{
+	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be;
+
+	if (be) {
 		NV_DEBUG(nvbe->dev, "\n");
-		ttm_dma_tt_fini(&nvbe->ttm);
-		kfree(nvbe);
+
+		if (nvbe) {
+			if (nvbe->pages)
+				be->func->clear(be);
+			kfree(nvbe);
+		}
 	}
 }
 
 static int
-nv04_sgdma_bind(struct ttm_tt *ttm, struct ttm_mem_reg *mem)
+nv04_sgdma_bind(struct ttm_backend *be, struct ttm_mem_reg *mem)
 {
-	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)ttm;
+	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be;
 	struct drm_device *dev = nvbe->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_gpuobj *gpuobj = dev_priv->gart_info.sg_ctxdma;
@@ -41,23 +123,24 @@ nv04_sgdma_bind(struct ttm_tt *ttm, struct ttm_mem_reg *mem)
 
 	nvbe->offset = mem->start << PAGE_SHIFT;
 	pte = (nvbe->offset >> NV_CTXDMA_PAGE_SHIFT) + 2;
-	for (i = 0; i < ttm->num_pages; i++) {
-		dma_addr_t dma_offset = nvbe->ttm.dma_address[i];
+	for (i = 0; i < nvbe->nr_pages; i++) {
+		dma_addr_t dma_offset = nvbe->pages[i];
 		uint32_t offset_l = lower_32_bits(dma_offset);
 
 		for (j = 0; j < PAGE_SIZE / NV_CTXDMA_PAGE_SIZE; j++, pte++) {
 			nv_wo32(gpuobj, (pte * 4) + 0, offset_l | 3);
-			offset_l += NV_CTXDMA_PAGE_SIZE;
+			dma_offset += NV_CTXDMA_PAGE_SIZE;
 		}
 	}
 
+	nvbe->bound = true;
 	return 0;
 }
 
 static int
-nv04_sgdma_unbind(struct ttm_tt *ttm)
+nv04_sgdma_unbind(struct ttm_backend *be)
 {
-	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)ttm;
+	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be;
 	struct drm_device *dev = nvbe->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_gpuobj *gpuobj = dev_priv->gart_info.sg_ctxdma;
@@ -65,19 +148,22 @@ nv04_sgdma_unbind(struct ttm_tt *ttm)
 
 	NV_DEBUG(dev, "\n");
 
-	if (ttm->state != tt_bound)
+	if (!nvbe->bound)
 		return 0;
 
 	pte = (nvbe->offset >> NV_CTXDMA_PAGE_SHIFT) + 2;
-	for (i = 0; i < ttm->num_pages; i++) {
+	for (i = 0; i < nvbe->nr_pages; i++) {
 		for (j = 0; j < PAGE_SIZE / NV_CTXDMA_PAGE_SIZE; j++, pte++)
 			nv_wo32(gpuobj, (pte * 4) + 0, 0x00000000);
 	}
 
+	nvbe->bound = false;
 	return 0;
 }
 
 static struct ttm_backend_func nv04_sgdma_backend = {
+	.populate		= nouveau_sgdma_populate,
+	.clear			= nouveau_sgdma_clear,
 	.bind			= nv04_sgdma_bind,
 	.unbind			= nv04_sgdma_unbind,
 	.destroy		= nouveau_sgdma_destroy
@@ -96,14 +182,14 @@ nv41_sgdma_flush(struct nouveau_sgdma_be *nvbe)
 }
 
 static int
-nv41_sgdma_bind(struct ttm_tt *ttm, struct ttm_mem_reg *mem)
+nv41_sgdma_bind(struct ttm_backend *be, struct ttm_mem_reg *mem)
 {
-	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)ttm;
+	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be;
 	struct drm_nouveau_private *dev_priv = nvbe->dev->dev_private;
 	struct nouveau_gpuobj *pgt = dev_priv->gart_info.sg_ctxdma;
-	dma_addr_t *list = nvbe->ttm.dma_address;
+	dma_addr_t *list = nvbe->pages;
 	u32 pte = mem->start << 2;
-	u32 cnt = ttm->num_pages;
+	u32 cnt = nvbe->nr_pages;
 
 	nvbe->offset = mem->start << PAGE_SHIFT;
 
@@ -113,17 +199,18 @@ nv41_sgdma_bind(struct ttm_tt *ttm, struct ttm_mem_reg *mem)
 	}
 
 	nv41_sgdma_flush(nvbe);
+	nvbe->bound = true;
 	return 0;
 }
 
 static int
-nv41_sgdma_unbind(struct ttm_tt *ttm)
+nv41_sgdma_unbind(struct ttm_backend *be)
 {
-	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)ttm;
+	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be;
 	struct drm_nouveau_private *dev_priv = nvbe->dev->dev_private;
 	struct nouveau_gpuobj *pgt = dev_priv->gart_info.sg_ctxdma;
 	u32 pte = (nvbe->offset >> 12) << 2;
-	u32 cnt = ttm->num_pages;
+	u32 cnt = nvbe->nr_pages;
 
 	while (cnt--) {
 		nv_wo32(pgt, pte, 0x00000000);
@@ -131,22 +218,24 @@ nv41_sgdma_unbind(struct ttm_tt *ttm)
 	}
 
 	nv41_sgdma_flush(nvbe);
+	nvbe->bound = false;
 	return 0;
 }
 
 static struct ttm_backend_func nv41_sgdma_backend = {
+	.populate		= nouveau_sgdma_populate,
+	.clear			= nouveau_sgdma_clear,
 	.bind			= nv41_sgdma_bind,
 	.unbind			= nv41_sgdma_unbind,
 	.destroy		= nouveau_sgdma_destroy
 };
 
 static void
-nv44_sgdma_flush(struct ttm_tt *ttm)
+nv44_sgdma_flush(struct nouveau_sgdma_be *nvbe)
 {
-	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)ttm;
 	struct drm_device *dev = nvbe->dev;
 
-	nv_wr32(dev, 0x100814, (ttm->num_pages - 1) << 12);
+	nv_wr32(dev, 0x100814, (nvbe->nr_pages - 1) << 12);
 	nv_wr32(dev, 0x100808, nvbe->offset | 0x20);
 	if (!nv_wait(dev, 0x100808, 0x00000001, 0x00000001))
 		NV_ERROR(dev, "gart flush timeout: 0x%08x\n",
@@ -205,14 +294,14 @@ nv44_sgdma_fill(struct nouveau_gpuobj *pgt, dma_addr_t *list, u32 base, u32 cnt)
 }
 
 static int
-nv44_sgdma_bind(struct ttm_tt *ttm, struct ttm_mem_reg *mem)
+nv44_sgdma_bind(struct ttm_backend *be, struct ttm_mem_reg *mem)
 {
-	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)ttm;
+	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be;
 	struct drm_nouveau_private *dev_priv = nvbe->dev->dev_private;
 	struct nouveau_gpuobj *pgt = dev_priv->gart_info.sg_ctxdma;
-	dma_addr_t *list = nvbe->ttm.dma_address;
+	dma_addr_t *list = nvbe->pages;
 	u32 pte = mem->start << 2, tmp[4];
-	u32 cnt = ttm->num_pages;
+	u32 cnt = nvbe->nr_pages;
 	int i;
 
 	nvbe->offset = mem->start << PAGE_SHIFT;
@@ -240,18 +329,19 @@ nv44_sgdma_bind(struct ttm_tt *ttm, struct ttm_mem_reg *mem)
 	if (cnt)
 		nv44_sgdma_fill(pgt, list, pte, cnt);
 
-	nv44_sgdma_flush(ttm);
+	nv44_sgdma_flush(nvbe);
+	nvbe->bound = true;
 	return 0;
 }
 
 static int
-nv44_sgdma_unbind(struct ttm_tt *ttm)
+nv44_sgdma_unbind(struct ttm_backend *be)
 {
-	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)ttm;
+	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be;
 	struct drm_nouveau_private *dev_priv = nvbe->dev->dev_private;
 	struct nouveau_gpuobj *pgt = dev_priv->gart_info.sg_ctxdma;
 	u32 pte = (nvbe->offset >> 12) << 2;
-	u32 cnt = ttm->num_pages;
+	u32 cnt = nvbe->nr_pages;
 
 	if (pte & 0x0000000c) {
 		u32  max = 4 - ((pte >> 2) & 0x3);
@@ -273,47 +363,55 @@ nv44_sgdma_unbind(struct ttm_tt *ttm)
 	if (cnt)
 		nv44_sgdma_fill(pgt, NULL, pte, cnt);
 
-	nv44_sgdma_flush(ttm);
+	nv44_sgdma_flush(nvbe);
+	nvbe->bound = false;
 	return 0;
 }
 
 static struct ttm_backend_func nv44_sgdma_backend = {
+	.populate		= nouveau_sgdma_populate,
+	.clear			= nouveau_sgdma_clear,
 	.bind			= nv44_sgdma_bind,
 	.unbind			= nv44_sgdma_unbind,
 	.destroy		= nouveau_sgdma_destroy
 };
 
 static int
-nv50_sgdma_bind(struct ttm_tt *ttm, struct ttm_mem_reg *mem)
+nv50_sgdma_bind(struct ttm_backend *be, struct ttm_mem_reg *mem)
 {
-	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)ttm;
+	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be;
 	struct nouveau_mem *node = mem->mm_node;
-
 	/* noop: bound in move_notify() */
-	node->pages = nvbe->ttm.dma_address;
+	node->pages = nvbe->pages;
+	nvbe->pages = (dma_addr_t *)node;
+	nvbe->bound = true;
 	return 0;
 }
 
 static int
-nv50_sgdma_unbind(struct ttm_tt *ttm)
+nv50_sgdma_unbind(struct ttm_backend *be)
 {
+	struct nouveau_sgdma_be *nvbe = (struct nouveau_sgdma_be *)be;
+	struct nouveau_mem *node = (struct nouveau_mem *)nvbe->pages;
 	/* noop: unbound in move_notify() */
+	nvbe->pages = node->pages;
+	node->pages = NULL;
+	nvbe->bound = false;
 	return 0;
 }
 
 static struct ttm_backend_func nv50_sgdma_backend = {
+	.populate		= nouveau_sgdma_populate,
+	.clear			= nouveau_sgdma_clear,
 	.bind			= nv50_sgdma_bind,
 	.unbind			= nv50_sgdma_unbind,
 	.destroy		= nouveau_sgdma_destroy
 };
 
-struct ttm_tt *
-nouveau_sgdma_create_ttm(struct ttm_bo_device *bdev,
-			 unsigned long size, uint32_t page_flags,
-			 struct page *dummy_read_page)
+struct ttm_backend *
+nouveau_sgdma_init_ttm(struct drm_device *dev)
 {
-	struct drm_nouveau_private *dev_priv = nouveau_bdev(bdev);
-	struct drm_device *dev = dev_priv->dev;
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_sgdma_be *nvbe;
 
 	nvbe = kzalloc(sizeof(*nvbe), GFP_KERNEL);
@@ -321,13 +419,9 @@ nouveau_sgdma_create_ttm(struct ttm_bo_device *bdev,
 		return NULL;
 
 	nvbe->dev = dev;
-	nvbe->ttm.ttm.func = dev_priv->gart_info.func;
 
-	if (ttm_dma_tt_init(&nvbe->ttm, bdev, size, page_flags, dummy_read_page)) {
-		kfree(nvbe);
-		return NULL;
-	}
-	return &nvbe->ttm.ttm;
+	nvbe->backend.func = dev_priv->gart_info.func;
+	return &nvbe->backend;
 }
 
 int
@@ -338,7 +432,7 @@ nouveau_sgdma_init(struct drm_device *dev)
 	u32 aper_size, align;
 	int ret;
 
-	if (dev_priv->card_type >= NV_40 && pci_is_pcie(dev->pdev))
+	if (dev_priv->card_type >= NV_40 && drm_pci_device_is_pcie(dev))
 		aper_size = 512 * 1024 * 1024;
 	else
 		aper_size = 64 * 1024 * 1024;
@@ -367,7 +461,7 @@ nouveau_sgdma_init(struct drm_device *dev)
 		dev_priv->gart_info.type = NOUVEAU_GART_HW;
 		dev_priv->gart_info.func = &nv50_sgdma_backend;
 	} else
-	if (0 && pci_is_pcie(dev->pdev) &&
+	if (0 && drm_pci_device_is_pcie(dev) &&
 	    dev_priv->chipset > 0x40 && dev_priv->chipset != 0x45) {
 		if (nv44_graph_class(dev)) {
 			dev_priv->gart_info.func = &nv44_sgdma_backend;
diff --git a/drivers/gpu/drm/nouveau/nouveau_state.c b/drivers/gpu/drm/nouveau/nouveau_state.c
index 9c144fb..731acea 100644
--- a/drivers/gpu/drm/nouveau/nouveau_state.c
+++ b/drivers/gpu/drm/nouveau/nouveau_state.c
@@ -36,7 +36,6 @@
 #include "nouveau_drm.h"
 #include "nouveau_fbcon.h"
 #include "nouveau_ramht.h"
-#include "nouveau_gpio.h"
 #include "nouveau_pm.h"
 #include "nv50_display.h"
 
@@ -81,14 +80,17 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->display.early_init	= nv04_display_early_init;
 		engine->display.late_takedown	= nv04_display_late_takedown;
 		engine->display.create		= nv04_display_create;
-		engine->display.destroy		= nv04_display_destroy;
 		engine->display.init		= nv04_display_init;
-		engine->display.fini		= nv04_display_fini;
-		engine->pm.clocks_get		= nv04_pm_clocks_get;
-		engine->pm.clocks_pre		= nv04_pm_clocks_pre;
-		engine->pm.clocks_set		= nv04_pm_clocks_set;
-		engine->vram.init		= nv04_fb_vram_init;
-		engine->vram.takedown		= nouveau_stub_takedown;
+		engine->display.destroy		= nv04_display_destroy;
+		engine->gpio.init		= nouveau_stub_init;
+		engine->gpio.takedown		= nouveau_stub_takedown;
+		engine->gpio.get		= NULL;
+		engine->gpio.set		= NULL;
+		engine->gpio.irq_enable		= NULL;
+		engine->pm.clock_get		= nv04_pm_clock_get;
+		engine->pm.clock_pre		= nv04_pm_clock_pre;
+		engine->pm.clock_set		= nv04_pm_clock_set;
+		engine->vram.init		= nouveau_mem_detect;
 		engine->vram.flags_valid	= nouveau_mem_flags_valid;
 		break;
 	case 0x10:
@@ -126,20 +128,17 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->display.early_init	= nv04_display_early_init;
 		engine->display.late_takedown	= nv04_display_late_takedown;
 		engine->display.create		= nv04_display_create;
-		engine->display.destroy		= nv04_display_destroy;
 		engine->display.init		= nv04_display_init;
-		engine->display.fini		= nv04_display_fini;
-		engine->gpio.drive		= nv10_gpio_drive;
-		engine->gpio.sense		= nv10_gpio_sense;
-		engine->pm.clocks_get		= nv04_pm_clocks_get;
-		engine->pm.clocks_pre		= nv04_pm_clocks_pre;
-		engine->pm.clocks_set		= nv04_pm_clocks_set;
-		if (dev_priv->chipset == 0x1a ||
-		    dev_priv->chipset == 0x1f)
-			engine->vram.init	= nv1a_fb_vram_init;
-		else
-			engine->vram.init	= nv10_fb_vram_init;
-		engine->vram.takedown		= nouveau_stub_takedown;
+		engine->display.destroy		= nv04_display_destroy;
+		engine->gpio.init		= nouveau_stub_init;
+		engine->gpio.takedown		= nouveau_stub_takedown;
+		engine->gpio.get		= nv10_gpio_get;
+		engine->gpio.set		= nv10_gpio_set;
+		engine->gpio.irq_enable		= NULL;
+		engine->pm.clock_get		= nv04_pm_clock_get;
+		engine->pm.clock_pre		= nv04_pm_clock_pre;
+		engine->pm.clock_set		= nv04_pm_clock_set;
+		engine->vram.init		= nouveau_mem_detect;
 		engine->vram.flags_valid	= nouveau_mem_flags_valid;
 		break;
 	case 0x20:
@@ -157,11 +156,11 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->timer.init		= nv04_timer_init;
 		engine->timer.read		= nv04_timer_read;
 		engine->timer.takedown		= nv04_timer_takedown;
-		engine->fb.init			= nv20_fb_init;
-		engine->fb.takedown		= nv20_fb_takedown;
-		engine->fb.init_tile_region	= nv20_fb_init_tile_region;
-		engine->fb.set_tile_region	= nv20_fb_set_tile_region;
-		engine->fb.free_tile_region	= nv20_fb_free_tile_region;
+		engine->fb.init			= nv10_fb_init;
+		engine->fb.takedown		= nv10_fb_takedown;
+		engine->fb.init_tile_region	= nv10_fb_init_tile_region;
+		engine->fb.set_tile_region	= nv10_fb_set_tile_region;
+		engine->fb.free_tile_region	= nv10_fb_free_tile_region;
 		engine->fifo.channels		= 32;
 		engine->fifo.init		= nv10_fifo_init;
 		engine->fifo.takedown		= nv04_fifo_fini;
@@ -177,16 +176,17 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->display.early_init	= nv04_display_early_init;
 		engine->display.late_takedown	= nv04_display_late_takedown;
 		engine->display.create		= nv04_display_create;
-		engine->display.destroy		= nv04_display_destroy;
 		engine->display.init		= nv04_display_init;
-		engine->display.fini		= nv04_display_fini;
-		engine->gpio.drive		= nv10_gpio_drive;
-		engine->gpio.sense		= nv10_gpio_sense;
-		engine->pm.clocks_get		= nv04_pm_clocks_get;
-		engine->pm.clocks_pre		= nv04_pm_clocks_pre;
-		engine->pm.clocks_set		= nv04_pm_clocks_set;
-		engine->vram.init		= nv20_fb_vram_init;
-		engine->vram.takedown		= nouveau_stub_takedown;
+		engine->display.destroy		= nv04_display_destroy;
+		engine->gpio.init		= nouveau_stub_init;
+		engine->gpio.takedown		= nouveau_stub_takedown;
+		engine->gpio.get		= nv10_gpio_get;
+		engine->gpio.set		= nv10_gpio_set;
+		engine->gpio.irq_enable		= NULL;
+		engine->pm.clock_get		= nv04_pm_clock_get;
+		engine->pm.clock_pre		= nv04_pm_clock_pre;
+		engine->pm.clock_set		= nv04_pm_clock_set;
+		engine->vram.init		= nouveau_mem_detect;
 		engine->vram.flags_valid	= nouveau_mem_flags_valid;
 		break;
 	case 0x30:
@@ -224,18 +224,19 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->display.early_init	= nv04_display_early_init;
 		engine->display.late_takedown	= nv04_display_late_takedown;
 		engine->display.create		= nv04_display_create;
-		engine->display.destroy		= nv04_display_destroy;
 		engine->display.init		= nv04_display_init;
-		engine->display.fini		= nv04_display_fini;
-		engine->gpio.drive		= nv10_gpio_drive;
-		engine->gpio.sense		= nv10_gpio_sense;
-		engine->pm.clocks_get		= nv04_pm_clocks_get;
-		engine->pm.clocks_pre		= nv04_pm_clocks_pre;
-		engine->pm.clocks_set		= nv04_pm_clocks_set;
+		engine->display.destroy		= nv04_display_destroy;
+		engine->gpio.init		= nouveau_stub_init;
+		engine->gpio.takedown		= nouveau_stub_takedown;
+		engine->gpio.get		= nv10_gpio_get;
+		engine->gpio.set		= nv10_gpio_set;
+		engine->gpio.irq_enable		= NULL;
+		engine->pm.clock_get		= nv04_pm_clock_get;
+		engine->pm.clock_pre		= nv04_pm_clock_pre;
+		engine->pm.clock_set		= nv04_pm_clock_set;
 		engine->pm.voltage_get		= nouveau_voltage_gpio_get;
 		engine->pm.voltage_set		= nouveau_voltage_gpio_set;
-		engine->vram.init		= nv20_fb_vram_init;
-		engine->vram.takedown		= nouveau_stub_takedown;
+		engine->vram.init		= nouveau_mem_detect;
 		engine->vram.flags_valid	= nouveau_mem_flags_valid;
 		break;
 	case 0x40:
@@ -274,30 +275,26 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->display.early_init	= nv04_display_early_init;
 		engine->display.late_takedown	= nv04_display_late_takedown;
 		engine->display.create		= nv04_display_create;
-		engine->display.destroy		= nv04_display_destroy;
 		engine->display.init		= nv04_display_init;
-		engine->display.fini		= nv04_display_fini;
-		engine->gpio.init		= nv10_gpio_init;
-		engine->gpio.fini		= nv10_gpio_fini;
-		engine->gpio.drive		= nv10_gpio_drive;
-		engine->gpio.sense		= nv10_gpio_sense;
-		engine->gpio.irq_enable		= nv10_gpio_irq_enable;
-		engine->pm.clocks_get		= nv40_pm_clocks_get;
-		engine->pm.clocks_pre		= nv40_pm_clocks_pre;
-		engine->pm.clocks_set		= nv40_pm_clocks_set;
+		engine->display.destroy		= nv04_display_destroy;
+		engine->gpio.init		= nouveau_stub_init;
+		engine->gpio.takedown		= nouveau_stub_takedown;
+		engine->gpio.get		= nv10_gpio_get;
+		engine->gpio.set		= nv10_gpio_set;
+		engine->gpio.irq_enable		= NULL;
+		engine->pm.clock_get		= nv04_pm_clock_get;
+		engine->pm.clock_pre		= nv04_pm_clock_pre;
+		engine->pm.clock_set		= nv04_pm_clock_set;
 		engine->pm.voltage_get		= nouveau_voltage_gpio_get;
 		engine->pm.voltage_set		= nouveau_voltage_gpio_set;
 		engine->pm.temp_get		= nv40_temp_get;
-		engine->pm.pwm_get		= nv40_pm_pwm_get;
-		engine->pm.pwm_set		= nv40_pm_pwm_set;
-		engine->vram.init		= nv40_fb_vram_init;
-		engine->vram.takedown		= nouveau_stub_takedown;
+		engine->vram.init		= nouveau_mem_detect;
 		engine->vram.flags_valid	= nouveau_mem_flags_valid;
 		break;
 	case 0x50:
 	case 0x80: /* gotta love NVIDIA's consistency.. */
 	case 0x90:
-	case 0xa0:
+	case 0xA0:
 		engine->instmem.init		= nv50_instmem_init;
 		engine->instmem.takedown	= nv50_instmem_takedown;
 		engine->instmem.suspend		= nv50_instmem_suspend;
@@ -332,13 +329,14 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->display.early_init	= nv50_display_early_init;
 		engine->display.late_takedown	= nv50_display_late_takedown;
 		engine->display.create		= nv50_display_create;
-		engine->display.destroy		= nv50_display_destroy;
 		engine->display.init		= nv50_display_init;
-		engine->display.fini		= nv50_display_fini;
+		engine->display.destroy		= nv50_display_destroy;
 		engine->gpio.init		= nv50_gpio_init;
-		engine->gpio.fini		= nv50_gpio_fini;
-		engine->gpio.drive		= nv50_gpio_drive;
-		engine->gpio.sense		= nv50_gpio_sense;
+		engine->gpio.takedown		= nv50_gpio_fini;
+		engine->gpio.get		= nv50_gpio_get;
+		engine->gpio.set		= nv50_gpio_set;
+		engine->gpio.irq_register	= nv50_gpio_irq_register;
+		engine->gpio.irq_unregister	= nv50_gpio_irq_unregister;
 		engine->gpio.irq_enable		= nv50_gpio_irq_enable;
 		switch (dev_priv->chipset) {
 		case 0x84:
@@ -351,14 +349,14 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		case 0xaa:
 		case 0xac:
 		case 0x50:
-			engine->pm.clocks_get	= nv50_pm_clocks_get;
-			engine->pm.clocks_pre	= nv50_pm_clocks_pre;
-			engine->pm.clocks_set	= nv50_pm_clocks_set;
+			engine->pm.clock_get	= nv50_pm_clock_get;
+			engine->pm.clock_pre	= nv50_pm_clock_pre;
+			engine->pm.clock_set	= nv50_pm_clock_set;
 			break;
 		default:
-			engine->pm.clocks_get	= nva3_pm_clocks_get;
-			engine->pm.clocks_pre	= nva3_pm_clocks_pre;
-			engine->pm.clocks_set	= nva3_pm_clocks_set;
+			engine->pm.clock_get	= nva3_pm_clock_get;
+			engine->pm.clock_pre	= nva3_pm_clock_pre;
+			engine->pm.clock_set	= nva3_pm_clock_set;
 			break;
 		}
 		engine->pm.voltage_get		= nouveau_voltage_gpio_get;
@@ -367,15 +365,12 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 			engine->pm.temp_get	= nv84_temp_get;
 		else
 			engine->pm.temp_get	= nv40_temp_get;
-		engine->pm.pwm_get		= nv50_pm_pwm_get;
-		engine->pm.pwm_set		= nv50_pm_pwm_set;
 		engine->vram.init		= nv50_vram_init;
-		engine->vram.takedown		= nv50_vram_fini;
 		engine->vram.get		= nv50_vram_new;
 		engine->vram.put		= nv50_vram_del;
 		engine->vram.flags_valid	= nv50_vram_flags_valid;
 		break;
-	case 0xc0:
+	case 0xC0:
 		engine->instmem.init		= nvc0_instmem_init;
 		engine->instmem.takedown	= nvc0_instmem_takedown;
 		engine->instmem.suspend		= nvc0_instmem_suspend;
@@ -406,93 +401,25 @@ static int nouveau_init_engine_ptrs(struct drm_device *dev)
 		engine->display.early_init	= nv50_display_early_init;
 		engine->display.late_takedown	= nv50_display_late_takedown;
 		engine->display.create		= nv50_display_create;
-		engine->display.destroy		= nv50_display_destroy;
 		engine->display.init		= nv50_display_init;
-		engine->display.fini		= nv50_display_fini;
-		engine->gpio.init		= nv50_gpio_init;
-		engine->gpio.fini		= nv50_gpio_fini;
-		engine->gpio.drive		= nv50_gpio_drive;
-		engine->gpio.sense		= nv50_gpio_sense;
-		engine->gpio.irq_enable		= nv50_gpio_irq_enable;
-		engine->vram.init		= nvc0_vram_init;
-		engine->vram.takedown		= nv50_vram_fini;
-		engine->vram.get		= nvc0_vram_new;
-		engine->vram.put		= nv50_vram_del;
-		engine->vram.flags_valid	= nvc0_vram_flags_valid;
-		engine->pm.temp_get		= nv84_temp_get;
-		engine->pm.clocks_get		= nvc0_pm_clocks_get;
-		engine->pm.clocks_pre		= nvc0_pm_clocks_pre;
-		engine->pm.clocks_set		= nvc0_pm_clocks_set;
-		engine->pm.voltage_get		= nouveau_voltage_gpio_get;
-		engine->pm.voltage_set		= nouveau_voltage_gpio_set;
-		engine->pm.pwm_get		= nv50_pm_pwm_get;
-		engine->pm.pwm_set		= nv50_pm_pwm_set;
-		break;
-	case 0xd0:
-		engine->instmem.init		= nvc0_instmem_init;
-		engine->instmem.takedown	= nvc0_instmem_takedown;
-		engine->instmem.suspend		= nvc0_instmem_suspend;
-		engine->instmem.resume		= nvc0_instmem_resume;
-		engine->instmem.get		= nv50_instmem_get;
-		engine->instmem.put		= nv50_instmem_put;
-		engine->instmem.map		= nv50_instmem_map;
-		engine->instmem.unmap		= nv50_instmem_unmap;
-		engine->instmem.flush		= nv84_instmem_flush;
-		engine->mc.init			= nv50_mc_init;
-		engine->mc.takedown		= nv50_mc_takedown;
-		engine->timer.init		= nv04_timer_init;
-		engine->timer.read		= nv04_timer_read;
-		engine->timer.takedown		= nv04_timer_takedown;
-		engine->fb.init			= nvc0_fb_init;
-		engine->fb.takedown		= nvc0_fb_takedown;
-		engine->fifo.channels		= 128;
-		engine->fifo.init		= nvc0_fifo_init;
-		engine->fifo.takedown		= nvc0_fifo_takedown;
-		engine->fifo.disable		= nvc0_fifo_disable;
-		engine->fifo.enable		= nvc0_fifo_enable;
-		engine->fifo.reassign		= nvc0_fifo_reassign;
-		engine->fifo.channel_id		= nvc0_fifo_channel_id;
-		engine->fifo.create_context	= nvc0_fifo_create_context;
-		engine->fifo.destroy_context	= nvc0_fifo_destroy_context;
-		engine->fifo.load_context	= nvc0_fifo_load_context;
-		engine->fifo.unload_context	= nvc0_fifo_unload_context;
-		engine->display.early_init	= nouveau_stub_init;
-		engine->display.late_takedown	= nouveau_stub_takedown;
-		engine->display.create		= nvd0_display_create;
-		engine->display.destroy		= nvd0_display_destroy;
-		engine->display.init		= nvd0_display_init;
-		engine->display.fini		= nvd0_display_fini;
+		engine->display.destroy		= nv50_display_destroy;
 		engine->gpio.init		= nv50_gpio_init;
-		engine->gpio.fini		= nv50_gpio_fini;
-		engine->gpio.drive		= nvd0_gpio_drive;
-		engine->gpio.sense		= nvd0_gpio_sense;
+		engine->gpio.takedown		= nouveau_stub_takedown;
+		engine->gpio.get		= nv50_gpio_get;
+		engine->gpio.set		= nv50_gpio_set;
+		engine->gpio.irq_register	= nv50_gpio_irq_register;
+		engine->gpio.irq_unregister	= nv50_gpio_irq_unregister;
 		engine->gpio.irq_enable		= nv50_gpio_irq_enable;
 		engine->vram.init		= nvc0_vram_init;
-		engine->vram.takedown		= nv50_vram_fini;
 		engine->vram.get		= nvc0_vram_new;
 		engine->vram.put		= nv50_vram_del;
 		engine->vram.flags_valid	= nvc0_vram_flags_valid;
-		engine->pm.temp_get		= nv84_temp_get;
-		engine->pm.clocks_get		= nvc0_pm_clocks_get;
-		engine->pm.clocks_pre		= nvc0_pm_clocks_pre;
-		engine->pm.clocks_set		= nvc0_pm_clocks_set;
-		engine->pm.voltage_get		= nouveau_voltage_gpio_get;
-		engine->pm.voltage_set		= nouveau_voltage_gpio_set;
 		break;
 	default:
 		NV_ERROR(dev, "NV%02x unsupported\n", dev_priv->chipset);
 		return 1;
 	}
 
-	/* headless mode */
-	if (nouveau_modeset == 2) {
-		engine->display.early_init = nouveau_stub_init;
-		engine->display.late_takedown = nouveau_stub_takedown;
-		engine->display.create = nouveau_stub_init;
-		engine->display.init = nouveau_stub_init;
-		engine->display.destroy = nouveau_stub_takedown;
-	}
-
 	return 0;
 }
 
@@ -514,6 +441,21 @@ nouveau_vga_set_decode(void *priv, bool state)
 		return VGA_RSRC_NORMAL_IO | VGA_RSRC_NORMAL_MEM;
 }
 
+static int
+nouveau_card_init_channel(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	int ret;
+
+	ret = nouveau_channel_alloc(dev, &dev_priv->channel,
+				    (struct drm_file *)-2, NvDmaFB, NvDmaTT);
+	if (ret)
+		return ret;
+
+	mutex_unlock(&dev_priv->channel->mutex);
+	return 0;
+}
+
 static void nouveau_switcheroo_set_state(struct pci_dev *pdev,
 					 enum vga_switcheroo_state state)
 {
@@ -529,7 +471,6 @@ static void nouveau_switcheroo_set_state(struct pci_dev *pdev,
 		printk(KERN_ERR "VGA switcheroo: switched nouveau off\n");
 		dev->switch_power_state = DRM_SWITCH_POWER_CHANGING;
 		drm_kms_helper_poll_disable(dev);
-		nouveau_switcheroo_optimus_dsm();
 		nouveau_pci_suspend(pdev, pmm);
 		dev->switch_power_state = DRM_SWITCH_POWER_OFF;
 	}
@@ -584,55 +525,45 @@ nouveau_card_init(struct drm_device *dev)
 	if (ret)
 		goto out_display_early;
 
-	/* workaround an odd issue on nvc1 by disabling the device's
-	 * nosnoop capability.  hopefully won't cause issues until a
-	 * better fix is found - assuming there is one...
-	 */
-	if (dev_priv->chipset == 0xc1) {
-		nv_mask(dev, 0x00088080, 0x00000800, 0x00000000);
-	}
+	nouveau_pm_init(dev);
 
-	/* PMC */
-	ret = engine->mc.init(dev);
+	ret = nouveau_mem_vram_init(dev);
 	if (ret)
 		goto out_bios;
 
-	/* PTIMER */
-	ret = engine->timer.init(dev);
-	if (ret)
-		goto out_mc;
-
-	/* PFB */
-	ret = engine->fb.init(dev);
+	ret = nouveau_gpuobj_init(dev);
 	if (ret)
-		goto out_timer;
+		goto out_vram;
 
-	ret = engine->vram.init(dev);
+	ret = engine->instmem.init(dev);
 	if (ret)
-		goto out_fb;
+		goto out_gpuobj;
 
-	/* PGPIO */
-	ret = nouveau_gpio_create(dev);
+	ret = nouveau_mem_gart_init(dev);
 	if (ret)
-		goto out_vram;
+		goto out_instmem;
 
-	ret = nouveau_gpuobj_init(dev);
+	/* PMC */
+	ret = engine->mc.init(dev);
 	if (ret)
-		goto out_gpio;
+		goto out_gart;
 
-	ret = engine->instmem.init(dev);
+	/* PGPIO */
+	ret = engine->gpio.init(dev);
 	if (ret)
-		goto out_gpuobj;
+		goto out_mc;
 
-	ret = nouveau_mem_vram_init(dev);
+	/* PTIMER */
+	ret = engine->timer.init(dev);
 	if (ret)
-		goto out_instmem;
+		goto out_gpio;
 
-	ret = nouveau_mem_gart_init(dev);
+	/* PFB */
+	ret = engine->fb.init(dev);
 	if (ret)
-		goto out_ttmvram;
+		goto out_timer;
 
-	if (!dev_priv->noaccel) {
+	if (!nouveau_noaccel) {
 		switch (dev_priv->card_type) {
 		case NV_04:
 			nv04_graph_create(dev);
@@ -651,7 +582,6 @@ nouveau_card_init(struct drm_device *dev)
 			nv50_graph_create(dev);
 			break;
 		case NV_C0:
-		case NV_D0:
 			nvc0_graph_create(dev);
 			break;
 		default:
@@ -667,11 +597,6 @@ nouveau_card_init(struct drm_device *dev)
 		case 0xa0:
 			nv84_crypt_create(dev);
 			break;
-		case 0x98:
-		case 0xaa:
-		case 0xac:
-			nv98_crypt_create(dev);
-			break;
 		}
 
 		switch (dev_priv->card_type) {
@@ -693,25 +618,12 @@ nouveau_card_init(struct drm_device *dev)
 			break;
 		}
 
-		if (dev_priv->chipset >= 0xa3 || dev_priv->chipset == 0x98) {
-			nv84_bsp_create(dev);
-			nv84_vp_create(dev);
-			nv98_ppp_create(dev);
-		} else
-		if (dev_priv->chipset >= 0x84) {
-			nv50_mpeg_create(dev);
-			nv84_bsp_create(dev);
-			nv84_vp_create(dev);
-		} else
-		if (dev_priv->chipset >= 0x50) {
+		if (dev_priv->card_type == NV_40)
+			nv40_mpeg_create(dev);
+		else
+		if (dev_priv->card_type == NV_50 &&
+		    (dev_priv->chipset < 0x98 || dev_priv->chipset == 0xa0))
 			nv50_mpeg_create(dev);
-		} else
-		if (dev_priv->card_type == NV_40 ||
-		    dev_priv->chipset == 0x31 ||
-		    dev_priv->chipset == 0x34 ||
-		    dev_priv->chipset == 0x36) {
-			nv31_mpeg_create(dev);
-		}
 
 		for (e = 0; e < NVOBJ_ENGINE_NR; e++) {
 			if (dev_priv->eng[e]) {
@@ -727,80 +639,71 @@ nouveau_card_init(struct drm_device *dev)
 			goto out_engine;
 	}
 
-	ret = nouveau_irq_init(dev);
+	ret = engine->display.create(dev);
 	if (ret)
 		goto out_fifo;
 
-	ret = nouveau_display_create(dev);
+	ret = drm_vblank_init(dev, nv_two_heads(dev) ? 2 : 1);
 	if (ret)
-		goto out_irq;
+		goto out_vblank;
 
-	nouveau_backlight_init(dev);
-	nouveau_pm_init(dev);
+	ret = nouveau_irq_init(dev);
+	if (ret)
+		goto out_vblank;
+
+	/* what about PVIDEO/PCRTC/PRAMDAC etc? */
 
 	if (dev_priv->eng[NVOBJ_ENGINE_GR]) {
 		ret = nouveau_fence_init(dev);
 		if (ret)
-			goto out_pm;
+			goto out_irq;
 
-		ret = nouveau_channel_alloc(dev, &dev_priv->channel, NULL,
-					    NvDmaFB, NvDmaTT);
+		ret = nouveau_card_init_channel(dev);
 		if (ret)
 			goto out_fence;
-
-		mutex_unlock(&dev_priv->channel->mutex);
-	}
-
-	if (dev->mode_config.num_crtc) {
-		ret = nouveau_display_init(dev);
-		if (ret)
-			goto out_chan;
-
-		nouveau_fbcon_init(dev);
 	}
 
+	nouveau_fbcon_init(dev);
+	drm_kms_helper_poll_init(dev);
 	return 0;
 
-out_chan:
-	nouveau_channel_put_unlocked(&dev_priv->channel);
 out_fence:
 	nouveau_fence_fini(dev);
-out_pm:
-	nouveau_pm_fini(dev);
-	nouveau_backlight_exit(dev);
-	nouveau_display_destroy(dev);
 out_irq:
 	nouveau_irq_fini(dev);
+out_vblank:
+	drm_vblank_cleanup(dev);
+	engine->display.destroy(dev);
 out_fifo:
-	if (!dev_priv->noaccel)
+	if (!nouveau_noaccel)
 		engine->fifo.takedown(dev);
 out_engine:
-	if (!dev_priv->noaccel) {
+	if (!nouveau_noaccel) {
 		for (e = e - 1; e >= 0; e--) {
 			if (!dev_priv->eng[e])
 				continue;
-			dev_priv->eng[e]->fini(dev, e, false);
+			dev_priv->eng[e]->fini(dev, e);
 			dev_priv->eng[e]->destroy(dev,e );
 		}
 	}
+
+	engine->fb.takedown(dev);
+out_timer:
+	engine->timer.takedown(dev);
+out_gpio:
+	engine->gpio.takedown(dev);
+out_mc:
+	engine->mc.takedown(dev);
+out_gart:
 	nouveau_mem_gart_fini(dev);
-out_ttmvram:
-	nouveau_mem_vram_fini(dev);
 out_instmem:
 	engine->instmem.takedown(dev);
 out_gpuobj:
 	nouveau_gpuobj_takedown(dev);
-out_gpio:
-	nouveau_gpio_destroy(dev);
 out_vram:
-	engine->vram.takedown(dev);
-out_fb:
-	engine->fb.takedown(dev);
-out_timer:
-	engine->timer.takedown(dev);
-out_mc:
-	engine->mc.takedown(dev);
+	nouveau_mem_vram_fini(dev);
 out_bios:
+	nouveau_pm_fini(dev);
 	nouveau_bios_takedown(dev);
 out_display_early:
 	engine->display.late_takedown(dev);
@@ -815,29 +718,25 @@ static void nouveau_card_takedown(struct drm_device *dev)
 	struct nouveau_engine *engine = &dev_priv->engine;
 	int e;
 
-	if (dev->mode_config.num_crtc) {
-		nouveau_fbcon_fini(dev);
-		nouveau_display_fini(dev);
-	}
-
 	if (dev_priv->channel) {
-		nouveau_channel_put_unlocked(&dev_priv->channel);
 		nouveau_fence_fini(dev);
+		nouveau_channel_put_unlocked(&dev_priv->channel);
 	}
 
-	nouveau_pm_fini(dev);
-	nouveau_backlight_exit(dev);
-	nouveau_display_destroy(dev);
-
-	if (!dev_priv->noaccel) {
+	if (!nouveau_noaccel) {
 		engine->fifo.takedown(dev);
 		for (e = NVOBJ_ENGINE_NR - 1; e >= 0; e--) {
 			if (dev_priv->eng[e]) {
-				dev_priv->eng[e]->fini(dev, e, false);
+				dev_priv->eng[e]->fini(dev, e);
 				dev_priv->eng[e]->destroy(dev,e );
 			}
 		}
 	}
+	engine->fb.takedown(dev);
+	engine->timer.takedown(dev);
+	engine->gpio.takedown(dev);
+	engine->mc.takedown(dev);
+	engine->display.late_takedown(dev);
 
 	if (dev_priv->vga_ram) {
 		nouveau_bo_unpin(dev_priv->vga_ram);
@@ -849,60 +748,20 @@ static void nouveau_card_takedown(struct drm_device *dev)
 	ttm_bo_clean_mm(&dev_priv->ttm.bdev, TTM_PL_TT);
 	mutex_unlock(&dev->struct_mutex);
 	nouveau_mem_gart_fini(dev);
-	nouveau_mem_vram_fini(dev);
 
 	engine->instmem.takedown(dev);
 	nouveau_gpuobj_takedown(dev);
+	nouveau_mem_vram_fini(dev);
 
-	nouveau_gpio_destroy(dev);
-	engine->vram.takedown(dev);
-	engine->fb.takedown(dev);
-	engine->timer.takedown(dev);
-	engine->mc.takedown(dev);
+	nouveau_irq_fini(dev);
+	drm_vblank_cleanup(dev);
 
+	nouveau_pm_fini(dev);
 	nouveau_bios_takedown(dev);
-	engine->display.late_takedown(dev);
-
-	nouveau_irq_fini(dev);
 
 	vga_client_register(dev->pdev, NULL, NULL, NULL);
 }
 
-int
-nouveau_open(struct drm_device *dev, struct drm_file *file_priv)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_fpriv *fpriv;
-	int ret;
-
-	fpriv = kzalloc(sizeof(*fpriv), GFP_KERNEL);
-	if (unlikely(!fpriv))
-		return -ENOMEM;
-
-	spin_lock_init(&fpriv->lock);
-	INIT_LIST_HEAD(&fpriv->channels);
-
-	if (dev_priv->card_type == NV_50) {
-		ret = nouveau_vm_new(dev, 0, (1ULL << 40), 0x0020000000ULL,
-				     &fpriv->vm);
-		if (ret) {
-			kfree(fpriv);
-			return ret;
-		}
-	} else
-	if (dev_priv->card_type >= NV_C0) {
-		ret = nouveau_vm_new(dev, 0, (1ULL << 40), 0x0008000000ULL,
-				     &fpriv->vm);
-		if (ret) {
-			kfree(fpriv);
-			return ret;
-		}
-	}
-
-	file_priv->driver_priv = fpriv;
-	return 0;
-}
-
 /* here a client dies, release the stuff that was allocated for its
  * file_priv */
 void nouveau_preclose(struct drm_device *dev, struct drm_file *file_priv)
@@ -910,14 +769,6 @@ void nouveau_preclose(struct drm_device *dev, struct drm_file *file_priv)
 	nouveau_channel_cleanup(dev, file_priv);
 }
 
-void
-nouveau_postclose(struct drm_device *dev, struct drm_file *file_priv)
-{
-	struct nouveau_fpriv *fpriv = nouveau_fpriv(file_priv);
-	nouveau_vm_ref(NULL, &fpriv->vm, NULL);
-	kfree(fpriv);
-}
-
 /* first module load, setup the mmio/fb mapping */
 /* KMS: we need mmio at load time, not when the first drm client opens. */
 int nouveau_firstopen(struct drm_device *dev)
@@ -993,7 +844,7 @@ static int nouveau_remove_conflicting_drivers(struct drm_device *dev)
 int nouveau_load(struct drm_device *dev, unsigned long flags)
 {
 	struct drm_nouveau_private *dev_priv;
-	uint32_t reg0 = ~0, strap;
+	uint32_t reg0;
 	resource_size_t mmio_start_offs;
 	int ret;
 
@@ -1005,72 +856,15 @@ int nouveau_load(struct drm_device *dev, unsigned long flags)
 	dev->dev_private = dev_priv;
 	dev_priv->dev = dev;
 
-	pci_set_master(dev->pdev);
-
 	dev_priv->flags = flags & NOUVEAU_FLAGS;
 
 	NV_DEBUG(dev, "vendor: 0x%X device: 0x%X class: 0x%X\n",
 		 dev->pci_vendor, dev->pci_device, dev->pdev->class);
 
-	/* first up, map the start of mmio and determine the chipset */
-	dev_priv->mmio = ioremap(pci_resource_start(dev->pdev, 0), PAGE_SIZE);
-	if (dev_priv->mmio) {
-#ifdef __BIG_ENDIAN
-		/* put the card into big-endian mode if it's not */
-		if (nv_rd32(dev, NV03_PMC_BOOT_1) != 0x01000001)
-			nv_wr32(dev, NV03_PMC_BOOT_1, 0x01000001);
-		DRM_MEMORYBARRIER();
-#endif
-
-		/* determine chipset and derive architecture from it */
-		reg0 = nv_rd32(dev, NV03_PMC_BOOT_0);
-		if ((reg0 & 0x0f000000) > 0) {
-			dev_priv->chipset = (reg0 & 0xff00000) >> 20;
-			switch (dev_priv->chipset & 0xf0) {
-			case 0x10:
-			case 0x20:
-			case 0x30:
-				dev_priv->card_type = dev_priv->chipset & 0xf0;
-				break;
-			case 0x40:
-			case 0x60:
-				dev_priv->card_type = NV_40;
-				break;
-			case 0x50:
-			case 0x80:
-			case 0x90:
-			case 0xa0:
-				dev_priv->card_type = NV_50;
-				break;
-			case 0xc0:
-				dev_priv->card_type = NV_C0;
-				break;
-			case 0xd0:
-				dev_priv->card_type = NV_D0;
-				break;
-			default:
-				break;
-			}
-		} else
-		if ((reg0 & 0xff00fff0) == 0x20004000) {
-			if (reg0 & 0x00f00000)
-				dev_priv->chipset = 0x05;
-			else
-				dev_priv->chipset = 0x04;
-			dev_priv->card_type = NV_04;
-		}
-
-		iounmap(dev_priv->mmio);
-	}
-
-	if (!dev_priv->card_type) {
-		NV_ERROR(dev, "unsupported chipset 0x%08x\n", reg0);
-		ret = -EINVAL;
-		goto err_priv;
-	}
-
-	NV_INFO(dev, "Detected an NV%2x generation card (0x%08x)\n",
-		     dev_priv->card_type, reg0);
+	/* resource 0 is mmio regs */
+	/* resource 1 is linear FB */
+	/* resource 2 is RAMIN (mmio regs + 0x1000000) */
+	/* resource 6 is bios */
 
 	/* map the mmio regs */
 	mmio_start_offs = pci_resource_start(dev->pdev, 0);
@@ -1084,41 +878,61 @@ int nouveau_load(struct drm_device *dev, unsigned long flags)
 	NV_DEBUG(dev, "regs mapped ok at 0x%llx\n",
 					(unsigned long long)mmio_start_offs);
 
-	/* determine frequency of timing crystal */
-	strap = nv_rd32(dev, 0x101000);
-	if ( dev_priv->chipset < 0x17 ||
-	    (dev_priv->chipset >= 0x20 && dev_priv->chipset <= 0x25))
-		strap &= 0x00000040;
-	else
-		strap &= 0x00400040;
+#ifdef __BIG_ENDIAN
+	/* Put the card in BE mode if it's not */
+	if (nv_rd32(dev, NV03_PMC_BOOT_1) != 0x01000001)
+		nv_wr32(dev, NV03_PMC_BOOT_1, 0x01000001);
 
-	switch (strap) {
-	case 0x00000000: dev_priv->crystal = 13500; break;
-	case 0x00000040: dev_priv->crystal = 14318; break;
-	case 0x00400000: dev_priv->crystal = 27000; break;
-	case 0x00400040: dev_priv->crystal = 25000; break;
-	}
+	DRM_MEMORYBARRIER();
+#endif
 
-	NV_DEBUG(dev, "crystal freq: %dKHz\n", dev_priv->crystal);
+	/* Time to determine the card architecture */
+	reg0 = nv_rd32(dev, NV03_PMC_BOOT_0);
+	dev_priv->stepping = 0; /* XXX: add stepping for pre-NV10? */
+
+	/* We're dealing with >=NV10 */
+	if ((reg0 & 0x0f000000) > 0) {
+		/* Bit 27-20 contain the architecture in hex */
+		dev_priv->chipset = (reg0 & 0xff00000) >> 20;
+		dev_priv->stepping = (reg0 & 0xff);
+	/* NV04 or NV05 */
+	} else if ((reg0 & 0xff00fff0) == 0x20004000) {
+		if (reg0 & 0x00f00000)
+			dev_priv->chipset = 0x05;
+		else
+			dev_priv->chipset = 0x04;
+	} else
+		dev_priv->chipset = 0xff;
 
-	/* Determine whether we'll attempt acceleration or not, some
-	 * cards are disabled by default here due to them being known
-	 * non-functional, or never been tested due to lack of hw.
-	 */
-	dev_priv->noaccel = !!nouveau_noaccel;
-	if (nouveau_noaccel == -1) {
-		switch (dev_priv->chipset) {
-		case 0xd9: /* known broken */
-			NV_INFO(dev, "acceleration disabled by default, pass "
-				     "noaccel=0 to force enable\n");
-			dev_priv->noaccel = true;
-			break;
-		default:
-			dev_priv->noaccel = false;
-			break;
-		}
+	switch (dev_priv->chipset & 0xf0) {
+	case 0x00:
+	case 0x10:
+	case 0x20:
+	case 0x30:
+		dev_priv->card_type = dev_priv->chipset & 0xf0;
+		break;
+	case 0x40:
+	case 0x60:
+		dev_priv->card_type = NV_40;
+		break;
+	case 0x50:
+	case 0x80:
+	case 0x90:
+	case 0xa0:
+		dev_priv->card_type = NV_50;
+		break;
+	case 0xc0:
+		dev_priv->card_type = NV_C0;
+		break;
+	default:
+		NV_INFO(dev, "Unsupported chipset 0x%08x\n", reg0);
+		ret = -EINVAL;
+		goto err_mmio;
 	}
 
+	NV_INFO(dev, "Detected an NV%2x generation card (0x%08x)\n",
+		dev_priv->card_type, reg0);
+
 	ret = nouveau_remove_conflicting_drivers(dev);
 	if (ret)
 		goto err_mmio;
@@ -1134,7 +948,7 @@ int nouveau_load(struct drm_device *dev, unsigned long flags)
 			ioremap(pci_resource_start(dev->pdev, ramin_bar),
 				dev_priv->ramin_size);
 		if (!dev_priv->ramin) {
-			NV_ERROR(dev, "Failed to map PRAMIN BAR\n");
+			NV_ERROR(dev, "Failed to PRAMIN BAR");
 			ret = -ENOMEM;
 			goto err_mmio;
 		}
@@ -1183,7 +997,11 @@ void nouveau_lastclose(struct drm_device *dev)
 int nouveau_unload(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_engine *engine = &dev_priv->engine;
 
+	drm_kms_helper_poll_fini(dev);
+	nouveau_fbcon_fini(dev);
+	engine->display.destroy(dev);
 	nouveau_card_takedown(dev);
 
 	iounmap(dev_priv->mmio);
@@ -1213,7 +1031,7 @@ int nouveau_ioctl_getparam(struct drm_device *dev, void *data,
 	case NOUVEAU_GETPARAM_BUS_TYPE:
 		if (drm_pci_device_is_agp(dev))
 			getparam->value = NV_AGP;
-		else if (pci_is_pcie(dev->pdev))
+		else if (drm_pci_device_is_pcie(dev))
 			getparam->value = NV_PCIE;
 		else
 			getparam->value = NV_PCI;
@@ -1302,23 +1120,6 @@ nouveau_wait_ne(struct drm_device *dev, uint64_t timeout,
 	return false;
 }
 
-/* Wait until cond(data) == true, up until timeout has hit */
-bool
-nouveau_wait_cb(struct drm_device *dev, u64 timeout,
-		bool (*cond)(void *), void *data)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_timer_engine *ptimer = &dev_priv->engine.timer;
-	u64 start = ptimer->read(dev);
-
-	do {
-		if (cond(data) == true)
-			return true;
-	} while (ptimer->read(dev) - start < timeout);
-
-	return false;
-}
-
 /* Waits for PGRAPH to go completely idle */
 bool nouveau_wait_for_idle(struct drm_device *dev)
 {
diff --git a/drivers/gpu/drm/nouveau/nouveau_temp.c b/drivers/gpu/drm/nouveau/nouveau_temp.c
index 0f5a301..649b041 100644
--- a/drivers/gpu/drm/nouveau/nouveau_temp.c
+++ b/drivers/gpu/drm/nouveau/nouveau_temp.c
@@ -22,8 +22,6 @@
  * Authors: Martin Peres
  */
 
-#include <linux/module.h>
-
 #include "drmP.h"
 
 #include "nouveau_drv.h"
@@ -45,7 +43,7 @@ nouveau_temp_vbios_parse(struct drm_device *dev, u8 *temp)
 
 	/* Set the default sensor's contants */
 	sensor->offset_constant = 0;
-	sensor->offset_mult = 0;
+	sensor->offset_mult = 1;
 	sensor->offset_div = 1;
 	sensor->slope_mult = 1;
 	sensor->slope_div = 1;
@@ -55,10 +53,6 @@ nouveau_temp_vbios_parse(struct drm_device *dev, u8 *temp)
 	temps->down_clock = 100;
 	temps->fan_boost = 90;
 
-	/* Set the default range for the pwm fan */
-	pm->fan.min_duty = 30;
-	pm->fan.max_duty = 100;
-
 	/* Set the known default values to setup the temperature sensor */
 	if (dev_priv->card_type >= NV_40) {
 		switch (dev_priv->chipset) {
@@ -105,13 +99,6 @@ nouveau_temp_vbios_parse(struct drm_device *dev, u8 *temp)
 			sensor->slope_mult = 431;
 			sensor->slope_div = 10000;
 			break;
-
-		case 0x67:
-			sensor->offset_mult = -26149;
-			sensor->offset_div = 100;
-			sensor->slope_mult = 484;
-			sensor->slope_div = 10000;
-			break;
 		}
 	}
 
@@ -122,7 +109,7 @@ nouveau_temp_vbios_parse(struct drm_device *dev, u8 *temp)
 
 	/* Read the entries from the table */
 	for (i = 0; i < entries; i++) {
-		s16 value = ROM16(temp[1]);
+		u16 value = ROM16(temp[1]);
 
 		switch (temp[0]) {
 		case 0x01:
@@ -160,26 +147,11 @@ nouveau_temp_vbios_parse(struct drm_device *dev, u8 *temp)
 		case 0x13:
 			sensor->slope_div = value;
 			break;
-		case 0x22:
-			pm->fan.min_duty = value & 0xff;
-			pm->fan.max_duty = (value & 0xff00) >> 8;
-			break;
-		case 0x26:
-			pm->fan.pwm_freq = value;
-			break;
 		}
 		temp += recordlen;
 	}
 
 	nouveau_temp_safety_checks(dev);
-
-	/* check the fan min/max settings */
-	if (pm->fan.min_duty < 10)
-		pm->fan.min_duty = 10;
-	if (pm->fan.max_duty > 100)
-		pm->fan.max_duty = 100;
-	if (pm->fan.max_duty < pm->fan.min_duty)
-		pm->fan.max_duty = pm->fan.min_duty;
 }
 
 static int
@@ -188,8 +160,8 @@ nv40_sensor_setup(struct drm_device *dev)
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_pm_engine *pm = &dev_priv->engine.pm;
 	struct nouveau_pm_temp_sensor_constants *sensor = &pm->sensor_constants;
-	s32 offset = sensor->offset_mult / sensor->offset_div;
-	s32 sensor_calibration;
+	u32 offset = sensor->offset_mult / sensor->offset_div;
+	u32 sensor_calibration;
 
 	/* set up the sensors */
 	sensor_calibration = 120 - offset - sensor->offset_constant;
@@ -286,6 +258,8 @@ probe_monitoring_device(struct nouveau_i2c_chan *i2c,
 static void
 nouveau_temp_probe_i2c(struct drm_device *dev)
 {
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct dcb_table *dcb = &dev_priv->vbios.dcb;
 	struct i2c_board_info info[] = {
 		{ I2C_BOARD_INFO("w83l785ts", 0x2d) },
 		{ I2C_BOARD_INFO("w83781d", 0x2d) },
@@ -294,9 +268,11 @@ nouveau_temp_probe_i2c(struct drm_device *dev)
 		{ I2C_BOARD_INFO("lm99", 0x4c) },
 		{ }
 	};
+	int idx = (dcb->version >= 0x40 ?
+		   dcb->i2c_default_indices & 0xf : 2);
 
 	nouveau_i2c_identify(dev, "monitoring device", info,
-			     probe_monitoring_device, NV_I2C_DEFAULT(0));
+			     probe_monitoring_device, idx);
 }
 
 void
@@ -312,9 +288,9 @@ nouveau_temp_init(struct drm_device *dev)
 			return;
 
 		if (P.version == 1)
-			temp = ROMPTR(dev, P.data[12]);
+			temp = ROMPTR(bios, P.data[12]);
 		else if (P.version == 2)
-			temp = ROMPTR(dev, P.data[16]);
+			temp = ROMPTR(bios, P.data[16]);
 		else
 			NV_WARN(dev, "unknown temp for BIT P %d\n", P.version);
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_vm.c b/drivers/gpu/drm/nouveau/nouveau_vm.c
index 2bf6c03..519a6b4 100644
--- a/drivers/gpu/drm/nouveau/nouveau_vm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_vm.c
@@ -78,10 +78,9 @@ nouveau_vm_map(struct nouveau_vma *vma, struct nouveau_mem *node)
 
 void
 nouveau_vm_map_sg(struct nouveau_vma *vma, u64 delta, u64 length,
-		  struct nouveau_mem *mem)
+		  struct nouveau_mem *mem, dma_addr_t *list)
 {
 	struct nouveau_vm *vm = vma->vm;
-	dma_addr_t *list = mem->pages;
 	int big = vma->node->type != vm->spg_shift;
 	u32 offset = vma->node->offset + (delta >> 12);
 	u32 bits = vma->node->type - 12;
@@ -173,9 +172,9 @@ nouveau_vm_unmap_pgt(struct nouveau_vm *vm, int big, u32 fpde, u32 lpde)
 			vm->map_pgt(vpgd->obj, pde, vpgt->obj);
 		}
 
-		mutex_unlock(&vm->mm.mutex);
+		mutex_unlock(&vm->mm->mutex);
 		nouveau_gpuobj_ref(NULL, &pgt);
-		mutex_lock(&vm->mm.mutex);
+		mutex_lock(&vm->mm->mutex);
 	}
 }
 
@@ -192,18 +191,18 @@ nouveau_vm_map_pgt(struct nouveau_vm *vm, u32 pde, u32 type)
 	pgt_size  = (1 << (vm->pgt_bits + 12)) >> type;
 	pgt_size *= 8;
 
-	mutex_unlock(&vm->mm.mutex);
+	mutex_unlock(&vm->mm->mutex);
 	ret = nouveau_gpuobj_new(vm->dev, NULL, pgt_size, 0x1000,
 				 NVOBJ_FLAG_ZERO_ALLOC, &pgt);
-	mutex_lock(&vm->mm.mutex);
+	mutex_lock(&vm->mm->mutex);
 	if (unlikely(ret))
 		return ret;
 
 	/* someone beat us to filling the PDE while we didn't have the lock */
 	if (unlikely(vpgt->refcount[big]++)) {
-		mutex_unlock(&vm->mm.mutex);
+		mutex_unlock(&vm->mm->mutex);
 		nouveau_gpuobj_ref(NULL, &pgt);
-		mutex_lock(&vm->mm.mutex);
+		mutex_lock(&vm->mm->mutex);
 		return 0;
 	}
 
@@ -224,10 +223,10 @@ nouveau_vm_get(struct nouveau_vm *vm, u64 size, u32 page_shift,
 	u32 fpde, lpde, pde;
 	int ret;
 
-	mutex_lock(&vm->mm.mutex);
-	ret = nouveau_mm_get(&vm->mm, page_shift, msize, 0, align, &vma->node);
+	mutex_lock(&vm->mm->mutex);
+	ret = nouveau_mm_get(vm->mm, page_shift, msize, 0, align, &vma->node);
 	if (unlikely(ret != 0)) {
-		mutex_unlock(&vm->mm.mutex);
+		mutex_unlock(&vm->mm->mutex);
 		return ret;
 	}
 
@@ -246,13 +245,13 @@ nouveau_vm_get(struct nouveau_vm *vm, u64 size, u32 page_shift,
 		if (ret) {
 			if (pde != fpde)
 				nouveau_vm_unmap_pgt(vm, big, fpde, pde - 1);
-			nouveau_mm_put(&vm->mm, vma->node);
-			mutex_unlock(&vm->mm.mutex);
+			nouveau_mm_put(vm->mm, vma->node);
+			mutex_unlock(&vm->mm->mutex);
 			vma->node = NULL;
 			return ret;
 		}
 	}
-	mutex_unlock(&vm->mm.mutex);
+	mutex_unlock(&vm->mm->mutex);
 
 	vma->vm     = vm;
 	vma->offset = (u64)vma->node->offset << 12;
@@ -271,11 +270,11 @@ nouveau_vm_put(struct nouveau_vma *vma)
 	fpde = (vma->node->offset >> vm->pgt_bits);
 	lpde = (vma->node->offset + vma->node->length - 1) >> vm->pgt_bits;
 
-	mutex_lock(&vm->mm.mutex);
+	mutex_lock(&vm->mm->mutex);
 	nouveau_vm_unmap_pgt(vm, vma->node->type != vm->spg_shift, fpde, lpde);
-	nouveau_mm_put(&vm->mm, vma->node);
+	nouveau_mm_put(vm->mm, vma->node);
 	vma->node = NULL;
-	mutex_unlock(&vm->mm.mutex);
+	mutex_unlock(&vm->mm->mutex);
 }
 
 int
@@ -307,7 +306,7 @@ nouveau_vm_new(struct drm_device *dev, u64 offset, u64 length, u64 mm_offset,
 			block = length;
 
 	} else
-	if (dev_priv->card_type >= NV_C0) {
+	if (dev_priv->card_type == NV_C0) {
 		vm->map_pgt = nvc0_vm_map_pgt;
 		vm->map = nvc0_vm_map;
 		vm->map_sg = nvc0_vm_map_sg;
@@ -361,35 +360,32 @@ nouveau_vm_link(struct nouveau_vm *vm, struct nouveau_gpuobj *pgd)
 
 	nouveau_gpuobj_ref(pgd, &vpgd->obj);
 
-	mutex_lock(&vm->mm.mutex);
+	mutex_lock(&vm->mm->mutex);
 	for (i = vm->fpde; i <= vm->lpde; i++)
 		vm->map_pgt(pgd, i, vm->pgt[i - vm->fpde].obj);
 	list_add(&vpgd->head, &vm->pgd_list);
-	mutex_unlock(&vm->mm.mutex);
+	mutex_unlock(&vm->mm->mutex);
 	return 0;
 }
 
 static void
-nouveau_vm_unlink(struct nouveau_vm *vm, struct nouveau_gpuobj *mpgd)
+nouveau_vm_unlink(struct nouveau_vm *vm, struct nouveau_gpuobj *pgd)
 {
 	struct nouveau_vm_pgd *vpgd, *tmp;
-	struct nouveau_gpuobj *pgd = NULL;
 
-	if (!mpgd)
+	if (!pgd)
 		return;
 
-	mutex_lock(&vm->mm.mutex);
+	mutex_lock(&vm->mm->mutex);
 	list_for_each_entry_safe(vpgd, tmp, &vm->pgd_list, head) {
-		if (vpgd->obj == mpgd) {
-			pgd = vpgd->obj;
-			list_del(&vpgd->head);
-			kfree(vpgd);
-			break;
-		}
-	}
-	mutex_unlock(&vm->mm.mutex);
+		if (vpgd->obj != pgd)
+			continue;
 
-	nouveau_gpuobj_ref(NULL, &pgd);
+		list_del(&vpgd->head);
+		nouveau_gpuobj_ref(NULL, &vpgd->obj);
+		kfree(vpgd);
+	}
+	mutex_unlock(&vm->mm->mutex);
 }
 
 static void
@@ -400,8 +396,8 @@ nouveau_vm_del(struct nouveau_vm *vm)
 	list_for_each_entry_safe(vpgd, tmp, &vm->pgd_list, head) {
 		nouveau_vm_unlink(vm, vpgd->obj);
 	}
+	WARN_ON(nouveau_mm_fini(&vm->mm) != 0);
 
-	nouveau_mm_fini(&vm->mm);
 	kfree(vm->pgt);
 	kfree(vm);
 }
diff --git a/drivers/gpu/drm/nouveau/nouveau_vm.h b/drivers/gpu/drm/nouveau/nouveau_vm.h
index 4fb6e72..c48a9fc 100644
--- a/drivers/gpu/drm/nouveau/nouveau_vm.h
+++ b/drivers/gpu/drm/nouveau/nouveau_vm.h
@@ -41,8 +41,6 @@ struct nouveau_vm_pgd {
 };
 
 struct nouveau_vma {
-	struct list_head head;
-	int refcount;
 	struct nouveau_vm *vm;
 	struct nouveau_mm_node *node;
 	u64 offset;
@@ -51,7 +49,7 @@ struct nouveau_vma {
 
 struct nouveau_vm {
 	struct drm_device *dev;
-	struct nouveau_mm mm;
+	struct nouveau_mm *mm;
 	int refcount;
 
 	struct list_head pgd_list;
@@ -89,7 +87,7 @@ void nouveau_vm_map_at(struct nouveau_vma *, u64 offset, struct nouveau_mem *);
 void nouveau_vm_unmap(struct nouveau_vma *);
 void nouveau_vm_unmap_at(struct nouveau_vma *, u64 offset, u64 length);
 void nouveau_vm_map_sg(struct nouveau_vma *, u64 offset, u64 length,
-		       struct nouveau_mem *);
+		       struct nouveau_mem *, dma_addr_t *);
 
 /* nv50_vm.c */
 void nv50_vm_map_pgt(struct nouveau_gpuobj *pgd, u32 pde,
diff --git a/drivers/gpu/drm/nouveau/nouveau_volt.c b/drivers/gpu/drm/nouveau/nouveau_volt.c
index b010cb9..75e87274 100644
--- a/drivers/gpu/drm/nouveau/nouveau_volt.c
+++ b/drivers/gpu/drm/nouveau/nouveau_volt.c
@@ -26,15 +26,15 @@
 
 #include "nouveau_drv.h"
 #include "nouveau_pm.h"
-#include "nouveau_gpio.h"
 
-static const enum dcb_gpio_tag vidtag[] = { 0x04, 0x05, 0x06, 0x1a, 0x73 };
+static const enum dcb_gpio_tag vidtag[] = { 0x04, 0x05, 0x06, 0x1a };
 static int nr_vidtag = sizeof(vidtag) / sizeof(vidtag[0]);
 
 int
 nouveau_voltage_gpio_get(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_gpio_engine *gpio = &dev_priv->engine.gpio;
 	struct nouveau_pm_voltage *volt = &dev_priv->engine.pm.voltage;
 	u8 vid = 0;
 	int i;
@@ -43,7 +43,7 @@ nouveau_voltage_gpio_get(struct drm_device *dev)
 		if (!(volt->vid_mask & (1 << i)))
 			continue;
 
-		vid |= nouveau_gpio_func_get(dev, vidtag[i]) << i;
+		vid |= gpio->get(dev, vidtag[i]) << i;
 	}
 
 	return nouveau_volt_lvl_lookup(dev, vid);
@@ -53,6 +53,7 @@ int
 nouveau_voltage_gpio_set(struct drm_device *dev, int voltage)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_gpio_engine *gpio = &dev_priv->engine.gpio;
 	struct nouveau_pm_voltage *volt = &dev_priv->engine.pm.voltage;
 	int vid, i;
 
@@ -64,7 +65,7 @@ nouveau_voltage_gpio_set(struct drm_device *dev, int voltage)
 		if (!(volt->vid_mask & (1 << i)))
 			continue;
 
-		nouveau_gpio_func_set(dev, vidtag[i], !!(vid & (1 << i)));
+		gpio->set(dev, vidtag[i], !!(vid & (1 << i)));
 	}
 
 	return 0;
@@ -116,10 +117,10 @@ nouveau_volt_init(struct drm_device *dev)
 			return;
 
 		if (P.version == 1)
-			volt = ROMPTR(dev, P.data[16]);
+			volt = ROMPTR(bios, P.data[16]);
 		else
 		if (P.version == 2)
-			volt = ROMPTR(dev, P.data[12]);
+			volt = ROMPTR(bios, P.data[12]);
 		else {
 			NV_WARN(dev, "unknown volt for BIT P %d\n", P.version);
 		}
@@ -129,7 +130,7 @@ nouveau_volt_init(struct drm_device *dev)
 			return;
 		}
 
-		volt = ROMPTR(dev, bios->data[bios->offset + 0x98]);
+		volt = ROMPTR(bios, bios->data[bios->offset + 0x98]);
 	}
 
 	if (!volt) {
@@ -169,13 +170,6 @@ nouveau_volt_init(struct drm_device *dev)
 		 */
 		vidshift  = 2;
 		break;
-	case 0x40:
-		headerlen = volt[1];
-		recordlen = volt[2];
-		entries   = volt[3]; /* not a clue what the entries are for.. */
-		vidmask   = volt[11]; /* guess.. */
-		vidshift  = 0;
-		break;
 	default:
 		NV_WARN(dev, "voltage table 0x%02x unknown\n", volt[0]);
 		return;
@@ -193,7 +187,7 @@ nouveau_volt_init(struct drm_device *dev)
 			return;
 		}
 
-		if (!nouveau_gpio_func_valid(dev, vidtag[i])) {
+		if (!nouveau_bios_gpio_entry(dev, vidtag[i])) {
 			NV_DEBUG(dev, "vid bit %d has no gpio tag\n", i);
 			return;
 		}
@@ -203,37 +197,16 @@ nouveau_volt_init(struct drm_device *dev)
 	}
 
 	/* parse vbios entries into common format */
-	voltage->version = volt[0];
-	if (voltage->version < 0x40) {
-		voltage->nr_level = entries;
-		voltage->level =
-			kcalloc(entries, sizeof(*voltage->level), GFP_KERNEL);
-		if (!voltage->level)
-			return;
-
-		entry = volt + headerlen;
-		for (i = 0; i < entries; i++, entry += recordlen) {
-			voltage->level[i].voltage = entry[0] * 10000;
-			voltage->level[i].vid     = entry[1] >> vidshift;
-		}
-	} else {
-		u32 volt_uv = ROM32(volt[4]);
-		s16 step_uv = ROM16(volt[8]);
-		u8 vid;
-
-		voltage->nr_level = voltage->vid_mask + 1;
-		voltage->level = kcalloc(voltage->nr_level,
-					 sizeof(*voltage->level), GFP_KERNEL);
-		if (!voltage->level)
-			return;
+	voltage->level = kcalloc(entries, sizeof(*voltage->level), GFP_KERNEL);
+	if (!voltage->level)
+		return;
 
-		for (vid = 0; vid <= voltage->vid_mask; vid++) {
-			voltage->level[vid].voltage = volt_uv;
-			voltage->level[vid].vid = vid;
-			volt_uv += step_uv;
-		}
+	entry = volt + headerlen;
+	for (i = 0; i < entries; i++, entry += recordlen) {
+		voltage->level[i].voltage = entry[0];
+		voltage->level[i].vid     = entry[1] >> vidshift;
 	}
-
+	voltage->nr_level  = entries;
 	voltage->supported = true;
 }
 
diff --git a/drivers/gpu/drm/nouveau/nv04_crtc.c b/drivers/gpu/drm/nouveau/nv04_crtc.c
index 728d075..f1a3ae4 100644
--- a/drivers/gpu/drm/nouveau/nv04_crtc.c
+++ b/drivers/gpu/drm/nouveau/nv04_crtc.c
@@ -364,7 +364,7 @@ nv_crtc_mode_set_vga(struct drm_crtc *crtc, struct drm_display_mode *mode)
 	regp->CRTC[NV_CIO_CR_VRE_INDEX] = 1 << 5 | XLATE(vertEnd, 0, NV_CIO_CR_VRE_3_0);
 	regp->CRTC[NV_CIO_CR_VDE_INDEX] = vertDisplay;
 	/* framebuffer can be larger than crtc scanout area. */
-	regp->CRTC[NV_CIO_CR_OFFSET_INDEX] = fb->pitches[0] / 8;
+	regp->CRTC[NV_CIO_CR_OFFSET_INDEX] = fb->pitch / 8;
 	regp->CRTC[NV_CIO_CR_ULINE_INDEX] = 0x00;
 	regp->CRTC[NV_CIO_CR_VBS_INDEX] = vertBlankStart;
 	regp->CRTC[NV_CIO_CR_VBE_INDEX] = vertBlankEnd;
@@ -377,9 +377,9 @@ nv_crtc_mode_set_vga(struct drm_crtc *crtc, struct drm_display_mode *mode)
 
 	/* framebuffer can be larger than crtc scanout area. */
 	regp->CRTC[NV_CIO_CRE_RPC0_INDEX] =
-		XLATE(fb->pitches[0] / 8, 8, NV_CIO_CRE_RPC0_OFFSET_10_8);
+		XLATE(fb->pitch / 8, 8, NV_CIO_CRE_RPC0_OFFSET_10_8);
 	regp->CRTC[NV_CIO_CRE_42] =
-		XLATE(fb->pitches[0] / 8, 11, NV_CIO_CRE_42_OFFSET_11);
+		XLATE(fb->pitch / 8, 11, NV_CIO_CRE_42_OFFSET_11);
 	regp->CRTC[NV_CIO_CRE_RPC1_INDEX] = mode->crtc_hdisplay < 1280 ?
 					    MASK(NV_CIO_CRE_RPC1_LARGE) : 0x00;
 	regp->CRTC[NV_CIO_CRE_LSR_INDEX] = XLATE(horizBlankEnd, 6, NV_CIO_CRE_LSR_HBE_6) |
@@ -781,20 +781,11 @@ nv04_crtc_do_mode_set_base(struct drm_crtc *crtc,
 	struct drm_device *dev = crtc->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nv04_crtc_reg *regp = &dev_priv->mode_reg.crtc_reg[nv_crtc->index];
-	struct drm_framebuffer *drm_fb;
-	struct nouveau_framebuffer *fb;
+	struct drm_framebuffer *drm_fb = nv_crtc->base.fb;
+	struct nouveau_framebuffer *fb = nouveau_framebuffer(drm_fb);
 	int arb_burst, arb_lwm;
 	int ret;
 
-	NV_DEBUG_KMS(dev, "index %d\n", nv_crtc->index);
-
-	/* no fb bound */
-	if (!atomic && !crtc->fb) {
-		NV_DEBUG_KMS(dev, "No FB bound\n");
-		return 0;
-	}
-
-
 	/* If atomic, we want to switch to the fb we were passed, so
 	 * now we update pointers to do that.  (We don't pin; just
 	 * assume we're already pinned and update the base address.)
@@ -803,8 +794,6 @@ nv04_crtc_do_mode_set_base(struct drm_crtc *crtc,
 		drm_fb = passed_fb;
 		fb = nouveau_framebuffer(passed_fb);
 	} else {
-		drm_fb = crtc->fb;
-		fb = nouveau_framebuffer(crtc->fb);
 		/* If not atomic, we can go ahead and pin, and unpin the
 		 * old fb we were passed.
 		 */
@@ -835,18 +824,18 @@ nv04_crtc_do_mode_set_base(struct drm_crtc *crtc,
 	NVWriteRAMDAC(dev, nv_crtc->index, NV_PRAMDAC_GENERAL_CONTROL,
 		      regp->ramdac_gen_ctrl);
 
-	regp->CRTC[NV_CIO_CR_OFFSET_INDEX] = drm_fb->pitches[0] >> 3;
+	regp->CRTC[NV_CIO_CR_OFFSET_INDEX] = drm_fb->pitch >> 3;
 	regp->CRTC[NV_CIO_CRE_RPC0_INDEX] =
-		XLATE(drm_fb->pitches[0] >> 3, 8, NV_CIO_CRE_RPC0_OFFSET_10_8);
+		XLATE(drm_fb->pitch >> 3, 8, NV_CIO_CRE_RPC0_OFFSET_10_8);
 	regp->CRTC[NV_CIO_CRE_42] =
-		XLATE(drm_fb->pitches[0] / 8, 11, NV_CIO_CRE_42_OFFSET_11);
+		XLATE(drm_fb->pitch / 8, 11, NV_CIO_CRE_42_OFFSET_11);
 	crtc_wr_cio_state(crtc, regp, NV_CIO_CRE_RPC0_INDEX);
 	crtc_wr_cio_state(crtc, regp, NV_CIO_CR_OFFSET_INDEX);
 	crtc_wr_cio_state(crtc, regp, NV_CIO_CRE_42);
 
 	/* Update the framebuffer location. */
 	regp->fb_start = nv_crtc->fb.offset & ~3;
-	regp->fb_start += (y * drm_fb->pitches[0]) + (x * drm_fb->bits_per_pixel / 8);
+	regp->fb_start += (y * drm_fb->pitch) + (x * drm_fb->bits_per_pixel / 8);
 	nv_set_crtc_base(dev, nv_crtc->index, regp->fb_start);
 
 	/* Update the arbitration parameters. */
@@ -1046,7 +1035,7 @@ nv04_crtc_create(struct drm_device *dev, int crtc_num)
 	drm_crtc_helper_add(&nv_crtc->base, &nv04_crtc_helper_funcs);
 	drm_mode_crtc_set_gamma_size(&nv_crtc->base, 256);
 
-	ret = nouveau_bo_new(dev, 64*64*4, 0x100, TTM_PL_FLAG_VRAM,
+	ret = nouveau_bo_new(dev, NULL, 64*64*4, 0x100, TTM_PL_FLAG_VRAM,
 			     0, 0x0000, &nv_crtc->cursor.nvbo);
 	if (!ret) {
 		ret = nouveau_bo_pin(nv_crtc->cursor.nvbo, TTM_PL_FLAG_VRAM);
diff --git a/drivers/gpu/drm/nouveau/nv04_dac.c b/drivers/gpu/drm/nouveau/nv04_dac.c
index f180dcf..2d6bfd0 100644
--- a/drivers/gpu/drm/nouveau/nv04_dac.c
+++ b/drivers/gpu/drm/nouveau/nv04_dac.c
@@ -32,7 +32,6 @@
 #include "nouveau_connector.h"
 #include "nouveau_crtc.h"
 #include "nouveau_hw.h"
-#include "nouveau_gpio.h"
 #include "nvreg.h"
 
 int nv04_dac_output_offset(struct drm_encoder *encoder)
@@ -221,6 +220,7 @@ uint32_t nv17_dac_sample_load(struct drm_encoder *encoder)
 {
 	struct drm_device *dev = encoder->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_gpio_engine *gpio = &dev_priv->engine.gpio;
 	struct dcb_entry *dcb = nouveau_encoder(encoder)->dcb;
 	uint32_t sample, testval, regoffset = nv04_dac_output_offset(encoder);
 	uint32_t saved_powerctrl_2 = 0, saved_powerctrl_4 = 0, saved_routput,
@@ -252,11 +252,11 @@ uint32_t nv17_dac_sample_load(struct drm_encoder *encoder)
 		nvWriteMC(dev, NV_PBUS_POWERCTRL_4, saved_powerctrl_4 & 0xffffffcf);
 	}
 
-	saved_gpio1 = nouveau_gpio_func_get(dev, DCB_GPIO_TVDAC1);
-	saved_gpio0 = nouveau_gpio_func_get(dev, DCB_GPIO_TVDAC0);
+	saved_gpio1 = gpio->get(dev, DCB_GPIO_TVDAC1);
+	saved_gpio0 = gpio->get(dev, DCB_GPIO_TVDAC0);
 
-	nouveau_gpio_func_set(dev, DCB_GPIO_TVDAC1, dcb->type == OUTPUT_TV);
-	nouveau_gpio_func_set(dev, DCB_GPIO_TVDAC0, dcb->type == OUTPUT_TV);
+	gpio->set(dev, DCB_GPIO_TVDAC1, dcb->type == OUTPUT_TV);
+	gpio->set(dev, DCB_GPIO_TVDAC0, dcb->type == OUTPUT_TV);
 
 	msleep(4);
 
@@ -306,8 +306,8 @@ uint32_t nv17_dac_sample_load(struct drm_encoder *encoder)
 		nvWriteMC(dev, NV_PBUS_POWERCTRL_4, saved_powerctrl_4);
 	nvWriteMC(dev, NV_PBUS_POWERCTRL_2, saved_powerctrl_2);
 
-	nouveau_gpio_func_set(dev, DCB_GPIO_TVDAC1, saved_gpio1);
-	nouveau_gpio_func_set(dev, DCB_GPIO_TVDAC0, saved_gpio0);
+	gpio->set(dev, DCB_GPIO_TVDAC1, saved_gpio1);
+	gpio->set(dev, DCB_GPIO_TVDAC0, saved_gpio0);
 
 	return sample;
 }
diff --git a/drivers/gpu/drm/nouveau/nv04_dfp.c b/drivers/gpu/drm/nouveau/nv04_dfp.c
index 473f30a..752440c 100644
--- a/drivers/gpu/drm/nouveau/nv04_dfp.c
+++ b/drivers/gpu/drm/nouveau/nv04_dfp.c
@@ -289,7 +289,6 @@ static void nv04_dfp_mode_set(struct drm_encoder *encoder,
 	struct nouveau_connector *nv_connector = nouveau_crtc_connector_get(nv_crtc);
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
 	struct drm_display_mode *output_mode = &nv_encoder->mode;
-	struct drm_connector *connector = &nv_connector->base;
 	uint32_t mode_ratio, panel_ratio;
 
 	NV_DEBUG_KMS(dev, "Output mode on CRTC %d:\n", nv_crtc->index);
@@ -341,15 +340,10 @@ static void nv04_dfp_mode_set(struct drm_encoder *encoder,
 	    output_mode->clock > 165000)
 		regp->fp_control |= (2 << 24);
 	if (nv_encoder->dcb->type == OUTPUT_LVDS) {
-		bool duallink = false, dummy;
-		if (nv_connector->edid &&
-		    nv_connector->type == DCB_CONNECTOR_LVDS_SPWG) {
-			duallink = (((u8 *)nv_connector->edid)[121] == 2);
-		} else {
-			nouveau_bios_parse_lvds_table(dev, output_mode->clock,
-						      &duallink, &dummy);
-		}
+		bool duallink, dummy;
 
+		nouveau_bios_parse_lvds_table(dev, output_mode->clock,
+					      &duallink, &dummy);
 		if (duallink)
 			regp->fp_control |= (8 << 28);
 	} else
@@ -413,9 +407,7 @@ static void nv04_dfp_mode_set(struct drm_encoder *encoder,
 	}
 
 	/* Output property. */
-	if ((nv_connector->dithering_mode == DITHERING_MODE_ON) ||
-	    (nv_connector->dithering_mode == DITHERING_MODE_AUTO &&
-	     encoder->crtc->fb->depth > connector->display_info.bpc * 3)) {
+	if (nv_connector->use_dithering) {
 		if (dev_priv->chipset == 0x11)
 			regp->dither = savep->dither | 0x00010000;
 		else {
diff --git a/drivers/gpu/drm/nouveau/nv04_display.c b/drivers/gpu/drm/nouveau/nv04_display.c
index 7047d37..1715e14 100644
--- a/drivers/gpu/drm/nouveau/nv04_display.c
+++ b/drivers/gpu/drm/nouveau/nv04_display.c
@@ -126,6 +126,27 @@ nv04_display_create(struct drm_device *dev)
 
 	nouveau_hw_save_vga_fonts(dev, 1);
 
+	drm_mode_config_init(dev);
+	drm_mode_create_scaling_mode_property(dev);
+	drm_mode_create_dithering_property(dev);
+
+	dev->mode_config.funcs = (void *)&nouveau_mode_config_funcs;
+
+	dev->mode_config.min_width = 0;
+	dev->mode_config.min_height = 0;
+	switch (dev_priv->card_type) {
+	case NV_04:
+		dev->mode_config.max_width = 2048;
+		dev->mode_config.max_height = 2048;
+		break;
+	default:
+		dev->mode_config.max_width = 4096;
+		dev->mode_config.max_height = 4096;
+		break;
+	}
+
+	dev->mode_config.fb_base = dev_priv->fb_phys;
+
 	nv04_crtc_create(dev, 0);
 	if (nv_two_heads(dev))
 		nv04_crtc_create(dev, 1);
@@ -214,6 +235,8 @@ nv04_display_destroy(struct drm_device *dev)
 	list_for_each_entry(crtc, &dev->mode_config.crtc_list, head)
 		crtc->funcs->restore(crtc);
 
+	drm_mode_config_cleanup(dev);
+
 	nouveau_hw_save_vga_fonts(dev, 0);
 }
 
@@ -243,11 +266,6 @@ nv04_display_init(struct drm_device *dev)
 	return 0;
 }
 
-void
-nv04_display_fini(struct drm_device *dev)
-{
-}
-
 static void
 nv04_vblank_crtc0_isr(struct drm_device *dev)
 {
diff --git a/drivers/gpu/drm/nouveau/nv04_fb.c b/drivers/gpu/drm/nouveau/nv04_fb.c
index d5eedd6..638cf60 100644
--- a/drivers/gpu/drm/nouveau/nv04_fb.c
+++ b/drivers/gpu/drm/nouveau/nv04_fb.c
@@ -4,40 +4,6 @@
 #include "nouveau_drm.h"
 
 int
-nv04_fb_vram_init(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	u32 boot0 = nv_rd32(dev, NV04_PFB_BOOT_0);
-
-	if (boot0 & 0x00000100) {
-		dev_priv->vram_size  = ((boot0 >> 12) & 0xf) * 2 + 2;
-		dev_priv->vram_size *= 1024 * 1024;
-	} else {
-		switch (boot0 & NV04_PFB_BOOT_0_RAM_AMOUNT) {
-		case NV04_PFB_BOOT_0_RAM_AMOUNT_32MB:
-			dev_priv->vram_size = 32 * 1024 * 1024;
-			break;
-		case NV04_PFB_BOOT_0_RAM_AMOUNT_16MB:
-			dev_priv->vram_size = 16 * 1024 * 1024;
-			break;
-		case NV04_PFB_BOOT_0_RAM_AMOUNT_8MB:
-			dev_priv->vram_size = 8 * 1024 * 1024;
-			break;
-		case NV04_PFB_BOOT_0_RAM_AMOUNT_4MB:
-			dev_priv->vram_size = 4 * 1024 * 1024;
-			break;
-		}
-	}
-
-	if ((boot0 & 0x00000038) <= 0x10)
-		dev_priv->vram_type = NV_MEM_TYPE_SGRAM;
-	else
-		dev_priv->vram_type = NV_MEM_TYPE_SDRAM;
-
-	return 0;
-}
-
-int
 nv04_fb_init(struct drm_device *dev)
 {
 	/* This is what the DDX did for NV_ARCH_04, but a mmio-trace shows
diff --git a/drivers/gpu/drm/nouveau/nv04_graph.c b/drivers/gpu/drm/nouveau/nv04_graph.c
index dbdea8e..3626ee7 100644
--- a/drivers/gpu/drm/nouveau/nv04_graph.c
+++ b/drivers/gpu/drm/nouveau/nv04_graph.c
@@ -450,13 +450,13 @@ nv04_graph_context_del(struct nouveau_channel *chan, int engine)
 	unsigned long flags;
 
 	spin_lock_irqsave(&dev_priv->context_switch_lock, flags);
-	nv_mask(dev, NV04_PGRAPH_FIFO, 0x00000001, 0x00000000);
+	nv04_graph_fifo_access(dev, false);
 
 	/* Unload the context if it's the currently active one */
 	if (nv04_graph_channel(dev) == chan)
 		nv04_graph_unload_context(dev);
 
-	nv_mask(dev, NV04_PGRAPH_FIFO, 0x00000001, 0x00000001);
+	nv04_graph_fifo_access(dev, true);
 	spin_unlock_irqrestore(&dev_priv->context_switch_lock, flags);
 
 	/* Free the context resources */
@@ -538,18 +538,24 @@ nv04_graph_init(struct drm_device *dev, int engine)
 }
 
 static int
-nv04_graph_fini(struct drm_device *dev, int engine, bool suspend)
+nv04_graph_fini(struct drm_device *dev, int engine)
 {
-	nv_mask(dev, NV04_PGRAPH_FIFO, 0x00000001, 0x00000000);
-	if (!nv_wait(dev, NV04_PGRAPH_STATUS, ~0, 0) && suspend) {
-		nv_mask(dev, NV04_PGRAPH_FIFO, 0x00000001, 0x00000001);
-		return -EBUSY;
-	}
 	nv04_graph_unload_context(dev);
 	nv_wr32(dev, NV03_PGRAPH_INTR_EN, 0x00000000);
 	return 0;
 }
 
+void
+nv04_graph_fifo_access(struct drm_device *dev, bool enabled)
+{
+	if (enabled)
+		nv_wr32(dev, NV04_PGRAPH_FIFO,
+					nv_rd32(dev, NV04_PGRAPH_FIFO) | 1);
+	else
+		nv_wr32(dev, NV04_PGRAPH_FIFO,
+					nv_rd32(dev, NV04_PGRAPH_FIFO) & ~1);
+}
+
 static int
 nv04_graph_mthd_set_ref(struct nouveau_channel *chan,
 			u32 class, u32 mthd, u32 data)
diff --git a/drivers/gpu/drm/nouveau/nv04_instmem.c b/drivers/gpu/drm/nouveau/nv04_instmem.c
index c1248e0..b8611b9 100644
--- a/drivers/gpu/drm/nouveau/nv04_instmem.c
+++ b/drivers/gpu/drm/nouveau/nv04_instmem.c
@@ -28,31 +28,6 @@ int nv04_instmem_init(struct drm_device *dev)
 	/* RAMIN always available */
 	dev_priv->ramin_available = true;
 
-	/* Reserve space at end of VRAM for PRAMIN */
-	if (dev_priv->card_type >= NV_40) {
-		u32 vs = hweight8((nv_rd32(dev, 0x001540) & 0x0000ff00) >> 8);
-		u32 rsvd;
-
-		/* estimate grctx size, the magics come from nv40_grctx.c */
-		if      (dev_priv->chipset == 0x40) rsvd = 0x6aa0 * vs;
-		else if (dev_priv->chipset  < 0x43) rsvd = 0x4f00 * vs;
-		else if (nv44_graph_class(dev))	    rsvd = 0x4980 * vs;
-		else				    rsvd = 0x4a40 * vs;
-		rsvd += 16 * 1024;
-		rsvd *= dev_priv->engine.fifo.channels;
-
-		/* pciegart table */
-		if (pci_is_pcie(dev->pdev))
-			rsvd += 512 * 1024;
-
-		/* object storage */
-		rsvd += 512 * 1024;
-
-		dev_priv->ramin_rsvd_vram = round_up(rsvd, 4096);
-	} else {
-		dev_priv->ramin_rsvd_vram = 512 * 1024;
-	}
-
 	/* Setup shared RAMHT */
 	ret = nouveau_gpuobj_new_fake(dev, 0x10000, ~0, 4096,
 				      NVOBJ_FLAG_ZERO_ALLOC, &ramht);
@@ -137,8 +112,7 @@ nv04_instmem_resume(struct drm_device *dev)
 }
 
 int
-nv04_instmem_get(struct nouveau_gpuobj *gpuobj, struct nouveau_channel *chan,
-		 u32 size, u32 align)
+nv04_instmem_get(struct nouveau_gpuobj *gpuobj, u32 size, u32 align)
 {
 	struct drm_nouveau_private *dev_priv = gpuobj->dev->dev_private;
 	struct drm_mm_node *ramin = NULL;
diff --git a/drivers/gpu/drm/nouveau/nv04_pm.c b/drivers/gpu/drm/nouveau/nv04_pm.c
index 6e75899..eb1c70d 100644
--- a/drivers/gpu/drm/nouveau/nv04_pm.c
+++ b/drivers/gpu/drm/nouveau/nv04_pm.c
@@ -27,111 +27,64 @@
 #include "nouveau_hw.h"
 #include "nouveau_pm.h"
 
-int
-nv04_pm_clocks_get(struct drm_device *dev, struct nouveau_pm_level *perflvl)
-{
-	int ret;
-
-	ret = nouveau_hw_get_clock(dev, PLL_CORE);
-	if (ret < 0)
-		return ret;
-	perflvl->core = ret;
-
-	ret = nouveau_hw_get_clock(dev, PLL_MEMORY);
-	if (ret < 0)
-		return ret;
-	perflvl->memory = ret;
-
-	return 0;
-}
-
-struct nv04_pm_clock {
+struct nv04_pm_state {
 	struct pll_lims pll;
 	struct nouveau_pll_vals calc;
 };
 
-struct nv04_pm_state {
-	struct nv04_pm_clock core;
-	struct nv04_pm_clock memory;
-};
-
-static int
-calc_pll(struct drm_device *dev, u32 id, int khz, struct nv04_pm_clock *clk)
+int
+nv04_pm_clock_get(struct drm_device *dev, u32 id)
 {
-	int ret;
-
-	ret = get_pll_limits(dev, id, &clk->pll);
-	if (ret)
-		return ret;
-
-	ret = nouveau_calc_pll_mnp(dev, &clk->pll, khz, &clk->calc);
-	if (!ret)
-		return -EINVAL;
-
-	return 0;
+	return nouveau_hw_get_clock(dev, id);
 }
 
 void *
-nv04_pm_clocks_pre(struct drm_device *dev, struct nouveau_pm_level *perflvl)
+nv04_pm_clock_pre(struct drm_device *dev, struct nouveau_pm_level *perflvl,
+		  u32 id, int khz)
 {
-	struct nv04_pm_state *info;
+	struct nv04_pm_state *state;
 	int ret;
 
-	info = kzalloc(sizeof(*info), GFP_KERNEL);
-	if (!info)
+	state = kzalloc(sizeof(*state), GFP_KERNEL);
+	if (!state)
 		return ERR_PTR(-ENOMEM);
 
-	ret = calc_pll(dev, PLL_CORE, perflvl->core, &info->core);
-	if (ret)
-		goto error;
+	ret = get_pll_limits(dev, id, &state->pll);
+	if (ret) {
+		kfree(state);
+		return (ret == -ENOENT) ? NULL : ERR_PTR(ret);
+	}
 
-	if (perflvl->memory) {
-		ret = calc_pll(dev, PLL_MEMORY, perflvl->memory, &info->memory);
-		if (ret)
-			goto error;
+	ret = nouveau_calc_pll_mnp(dev, &state->pll, khz, &state->calc);
+	if (!ret) {
+		kfree(state);
+		return ERR_PTR(-EINVAL);
 	}
 
-	return info;
-error:
-	kfree(info);
-	return ERR_PTR(ret);
+	return state;
 }
 
-static void
-prog_pll(struct drm_device *dev, struct nv04_pm_clock *clk)
+void
+nv04_pm_clock_set(struct drm_device *dev, void *pre_state)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	u32 reg = clk->pll.reg;
+	struct nv04_pm_state *state = pre_state;
+	u32 reg = state->pll.reg;
 
 	/* thank the insane nouveau_hw_setpll() interface for this */
 	if (dev_priv->card_type >= NV_40)
 		reg += 4;
 
-	nouveau_hw_setpll(dev, reg, &clk->calc);
-}
-
-int
-nv04_pm_clocks_set(struct drm_device *dev, void *pre_state)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_timer_engine *ptimer = &dev_priv->engine.timer;
-	struct nv04_pm_state *state = pre_state;
+	nouveau_hw_setpll(dev, reg, &state->calc);
 
-	prog_pll(dev, &state->core);
+	if (dev_priv->card_type < NV_30 && reg == NV_PRAMDAC_MPLL_COEFF) {
+		if (dev_priv->card_type == NV_20)
+			nv_mask(dev, 0x1002c4, 0, 1 << 20);
 
-	if (state->memory.pll.reg) {
-		prog_pll(dev, &state->memory);
-		if (dev_priv->card_type < NV_30) {
-			if (dev_priv->card_type == NV_20)
-				nv_mask(dev, 0x1002c4, 0, 1 << 20);
-
-			/* Reset the DLLs */
-			nv_mask(dev, 0x1002c0, 0, 1 << 8);
-		}
+		/* Reset the DLLs */
+		nv_mask(dev, 0x1002c0, 0, 1 << 8);
 	}
 
-	ptimer->init(dev);
-
 	kfree(state);
-	return 0;
 }
+
diff --git a/drivers/gpu/drm/nouveau/nv04_timer.c b/drivers/gpu/drm/nouveau/nv04_timer.c
index 55c9452..1d09ddd 100644
--- a/drivers/gpu/drm/nouveau/nv04_timer.c
+++ b/drivers/gpu/drm/nouveau/nv04_timer.c
@@ -2,80 +2,47 @@
 #include "drm.h"
 #include "nouveau_drv.h"
 #include "nouveau_drm.h"
-#include "nouveau_hw.h"
 
 int
 nv04_timer_init(struct drm_device *dev)
 {
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	u32 m, n, d;
-
 	nv_wr32(dev, NV04_PTIMER_INTR_EN_0, 0x00000000);
 	nv_wr32(dev, NV04_PTIMER_INTR_0, 0xFFFFFFFF);
 
-	/* aim for 31.25MHz, which gives us nanosecond timestamps */
-	d = 1000000 / 32;
-
-	/* determine base clock for timer source */
-	if (dev_priv->chipset < 0x40) {
-		n = nouveau_hw_get_clock(dev, PLL_CORE);
-	} else
-	if (dev_priv->chipset == 0x40) {
-		/*XXX: figure this out */
-		n = 0;
-	} else {
-		n = dev_priv->crystal;
-		m = 1;
-		while (n < (d * 2)) {
-			n += (n / m);
-			m++;
-		}
-
-		nv_wr32(dev, 0x009220, m - 1);
-	}
-
-	if (!n) {
-		NV_WARN(dev, "PTIMER: unknown input clock freq\n");
-		if (!nv_rd32(dev, NV04_PTIMER_NUMERATOR) ||
-		    !nv_rd32(dev, NV04_PTIMER_DENOMINATOR)) {
-			nv_wr32(dev, NV04_PTIMER_NUMERATOR, 1);
-			nv_wr32(dev, NV04_PTIMER_DENOMINATOR, 1);
-		}
-		return 0;
-	}
-
-	/* reduce ratio to acceptable values */
-	while (((n % 5) == 0) && ((d % 5) == 0)) {
-		n /= 5;
-		d /= 5;
+	/* Just use the pre-existing values when possible for now; these regs
+	 * are not written in nv (driver writer missed a /4 on the address), and
+	 * writing 8 and 3 to the correct regs breaks the timings on the LVDS
+	 * hardware sequencing microcode.
+	 * A correct solution (involving calculations with the GPU PLL) can
+	 * be done when kernel modesetting lands
+	 */
+	if (!nv_rd32(dev, NV04_PTIMER_NUMERATOR) ||
+				!nv_rd32(dev, NV04_PTIMER_DENOMINATOR)) {
+		nv_wr32(dev, NV04_PTIMER_NUMERATOR, 0x00000008);
+		nv_wr32(dev, NV04_PTIMER_DENOMINATOR, 0x00000003);
 	}
 
-	while (((n % 2) == 0) && ((d % 2) == 0)) {
-		n /= 2;
-		d /= 2;
-	}
-
-	while (n > 0xffff || d > 0xffff) {
-		n >>= 1;
-		d >>= 1;
-	}
-
-	nv_wr32(dev, NV04_PTIMER_NUMERATOR, n);
-	nv_wr32(dev, NV04_PTIMER_DENOMINATOR, d);
 	return 0;
 }
 
-u64
+uint64_t
 nv04_timer_read(struct drm_device *dev)
 {
-	u32 hi, lo;
-
+	uint32_t low;
+	/* From kmmio dumps on nv28 this looks like how the blob does this.
+	 * It reads the high dword twice, before and after.
+	 * The only explanation seems to be that the 64-bit timer counter
+	 * advances between high and low dword reads and may corrupt the
+	 * result. Not confirmed.
+	 */
+	uint32_t high2 = nv_rd32(dev, NV04_PTIMER_TIME_1);
+	uint32_t high1;
 	do {
-		hi = nv_rd32(dev, NV04_PTIMER_TIME_1);
-		lo = nv_rd32(dev, NV04_PTIMER_TIME_0);
-	} while (hi != nv_rd32(dev, NV04_PTIMER_TIME_1));
-
-	return ((u64)hi << 32 | lo);
+		high1 = high2;
+		low = nv_rd32(dev, NV04_PTIMER_TIME_0);
+		high2 = nv_rd32(dev, NV04_PTIMER_TIME_1);
+	} while (high1 != high2);
+	return (((uint64_t)high2) << 32) | (uint64_t)low;
 }
 
 void
diff --git a/drivers/gpu/drm/nouveau/nv10_fb.c b/drivers/gpu/drm/nouveau/nv10_fb.c
index 420b1608..f78181a 100644
--- a/drivers/gpu/drm/nouveau/nv10_fb.c
+++ b/drivers/gpu/drm/nouveau/nv10_fb.c
@@ -3,16 +3,81 @@
 #include "nouveau_drv.h"
 #include "nouveau_drm.h"
 
+static struct drm_mm_node *
+nv20_fb_alloc_tag(struct drm_device *dev, uint32_t size)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_fb_engine *pfb = &dev_priv->engine.fb;
+	struct drm_mm_node *mem;
+	int ret;
+
+	ret = drm_mm_pre_get(&pfb->tag_heap);
+	if (ret)
+		return NULL;
+
+	spin_lock(&dev_priv->tile.lock);
+	mem = drm_mm_search_free(&pfb->tag_heap, size, 0, 0);
+	if (mem)
+		mem = drm_mm_get_block_atomic(mem, size, 0);
+	spin_unlock(&dev_priv->tile.lock);
+
+	return mem;
+}
+
+static void
+nv20_fb_free_tag(struct drm_device *dev, struct drm_mm_node *mem)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+
+	spin_lock(&dev_priv->tile.lock);
+	drm_mm_put_block(mem);
+	spin_unlock(&dev_priv->tile.lock);
+}
+
 void
 nv10_fb_init_tile_region(struct drm_device *dev, int i, uint32_t addr,
 			 uint32_t size, uint32_t pitch, uint32_t flags)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_tile_reg *tile = &dev_priv->tile.reg[i];
+	int bpp = (flags & NOUVEAU_GEM_TILE_32BPP ? 32 : 16);
 
-	tile->addr  = 0x80000000 | addr;
+	tile->addr = addr;
 	tile->limit = max(1u, addr + size) - 1;
 	tile->pitch = pitch;
+
+	if (dev_priv->card_type == NV_20) {
+		if (flags & NOUVEAU_GEM_TILE_ZETA) {
+			/*
+			 * Allocate some of the on-die tag memory,
+			 * used to store Z compression meta-data (most
+			 * likely just a bitmap determining if a given
+			 * tile is compressed or not).
+			 */
+			tile->tag_mem = nv20_fb_alloc_tag(dev, size / 256);
+
+			if (tile->tag_mem) {
+				/* Enable Z compression */
+				if (dev_priv->chipset >= 0x25)
+					tile->zcomp = tile->tag_mem->start |
+						(bpp == 16 ?
+						 NV25_PFB_ZCOMP_MODE_16 :
+						 NV25_PFB_ZCOMP_MODE_32);
+				else
+					tile->zcomp = tile->tag_mem->start |
+						NV20_PFB_ZCOMP_EN |
+						(bpp == 16 ? 0 :
+						 NV20_PFB_ZCOMP_MODE_32);
+			}
+
+			tile->addr |= 3;
+		} else {
+			tile->addr |= 1;
+		}
+
+	} else {
+		tile->addr |= 1 << 31;
+	}
 }
 
 void
@@ -21,6 +86,11 @@ nv10_fb_free_tile_region(struct drm_device *dev, int i)
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_tile_reg *tile = &dev_priv->tile.reg[i];
 
+	if (tile->tag_mem) {
+		nv20_fb_free_tag(dev, tile->tag_mem);
+		tile->tag_mem = NULL;
+	}
+
 	tile->addr = tile->limit = tile->pitch = tile->zcomp = 0;
 }
 
@@ -33,48 +103,9 @@ nv10_fb_set_tile_region(struct drm_device *dev, int i)
 	nv_wr32(dev, NV10_PFB_TLIMIT(i), tile->limit);
 	nv_wr32(dev, NV10_PFB_TSIZE(i), tile->pitch);
 	nv_wr32(dev, NV10_PFB_TILE(i), tile->addr);
-}
-
-int
-nv1a_fb_vram_init(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct pci_dev *bridge;
-	uint32_t mem, mib;
-
-	bridge = pci_get_bus_and_slot(0, PCI_DEVFN(0, 1));
-	if (!bridge) {
-		NV_ERROR(dev, "no bridge device\n");
-		return 0;
-	}
-
-	if (dev_priv->chipset == 0x1a) {
-		pci_read_config_dword(bridge, 0x7c, &mem);
-		mib = ((mem >> 6) & 31) + 1;
-	} else {
-		pci_read_config_dword(bridge, 0x84, &mem);
-		mib = ((mem >> 4) & 127) + 1;
-	}
-
-	dev_priv->vram_size = mib * 1024 * 1024;
-	return 0;
-}
-
-int
-nv10_fb_vram_init(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	u32 fifo_data = nv_rd32(dev, NV04_PFB_FIFO_DATA);
-	u32 cfg0 = nv_rd32(dev, 0x100200);
 
-	dev_priv->vram_size = fifo_data & NV10_PFB_FIFO_DATA_RAM_AMOUNT_MB_MASK;
-
-	if (cfg0 & 0x00000001)
-		dev_priv->vram_type = NV_MEM_TYPE_DDR1;
-	else
-		dev_priv->vram_type = NV_MEM_TYPE_SDRAM;
-
-	return 0;
+	if (dev_priv->card_type == NV_20)
+		nv_wr32(dev, NV20_PFB_ZCOMP(i), tile->zcomp);
 }
 
 int
@@ -84,8 +115,14 @@ nv10_fb_init(struct drm_device *dev)
 	struct nouveau_fb_engine *pfb = &dev_priv->engine.fb;
 	int i;
 
-	/* Turn all the tiling regions off. */
 	pfb->num_tiles = NV10_PFB_TILE__SIZE;
+
+	if (dev_priv->card_type == NV_20)
+		drm_mm_init(&pfb->tag_heap, 0,
+			    (dev_priv->chipset >= 0x25 ?
+			     64 * 1024 : 32 * 1024));
+
+	/* Turn all the tiling regions off. */
 	for (i = 0; i < pfb->num_tiles; i++)
 		pfb->set_tile_region(dev, i);
 
@@ -101,4 +138,7 @@ nv10_fb_takedown(struct drm_device *dev)
 
 	for (i = 0; i < pfb->num_tiles; i++)
 		pfb->free_tile_region(dev, i);
+
+	if (dev_priv->card_type == NV_20)
+		drm_mm_takedown(&pfb->tag_heap);
 }
diff --git a/drivers/gpu/drm/nouveau/nv10_gpio.c b/drivers/gpu/drm/nouveau/nv10_gpio.c
index 550ad3f..007fc29 100644
--- a/drivers/gpu/drm/nouveau/nv10_gpio.c
+++ b/drivers/gpu/drm/nouveau/nv10_gpio.c
@@ -27,97 +27,66 @@
 #include "drmP.h"
 #include "nouveau_drv.h"
 #include "nouveau_hw.h"
-#include "nouveau_gpio.h"
 
-int
-nv10_gpio_sense(struct drm_device *dev, int line)
+static bool
+get_gpio_location(struct dcb_gpio_entry *ent, uint32_t *reg, uint32_t *shift,
+		  uint32_t *mask)
 {
-	if (line < 2) {
-		line = line * 16;
-		line = NVReadCRTC(dev, 0, NV_PCRTC_GPIO) >> line;
-		return !!(line & 0x0100);
-	} else
-	if (line < 10) {
-		line = (line - 2) * 4;
-		line = NVReadCRTC(dev, 0, NV_PCRTC_GPIO_EXT) >> line;
-		return !!(line & 0x04);
-	} else
-	if (line < 14) {
-		line = (line - 10) * 4;
-		line = NVReadCRTC(dev, 0, NV_PCRTC_850) >> line;
-		return !!(line & 0x04);
-	}
+	if (ent->line < 2) {
+		*reg = NV_PCRTC_GPIO;
+		*shift = ent->line * 16;
+		*mask = 0x11;
 
-	return -EINVAL;
-}
+	} else if (ent->line < 10) {
+		*reg = NV_PCRTC_GPIO_EXT;
+		*shift = (ent->line - 2) * 4;
+		*mask = 0x3;
+
+	} else if (ent->line < 14) {
+		*reg = NV_PCRTC_850;
+		*shift = (ent->line - 10) * 4;
+		*mask = 0x3;
 
-int
-nv10_gpio_drive(struct drm_device *dev, int line, int dir, int out)
-{
-	u32 reg, mask, data;
-
-	if (line < 2) {
-		line = line * 16;
-		reg  = NV_PCRTC_GPIO;
-		mask = 0x00000011;
-		data = (dir << 4) | out;
-	} else
-	if (line < 10) {
-		line = (line - 2) * 4;
-		reg  = NV_PCRTC_GPIO_EXT;
-		mask = 0x00000003 << ((line - 2) * 4);
-		data = (dir << 1) | out;
-	} else
-	if (line < 14) {
-		line = (line - 10) * 4;
-		reg  = NV_PCRTC_850;
-		mask = 0x00000003;
-		data = (dir << 1) | out;
 	} else {
-		return -EINVAL;
+		return false;
 	}
 
-	mask = NVReadCRTC(dev, 0, reg) & ~(mask << line);
-	NVWriteCRTC(dev, 0, reg, mask | (data << line));
-	return 0;
+	return true;
 }
 
-void
-nv10_gpio_irq_enable(struct drm_device *dev, int line, bool on)
+int
+nv10_gpio_get(struct drm_device *dev, enum dcb_gpio_tag tag)
 {
-	u32 mask = 0x00010001 << line;
+	struct dcb_gpio_entry *ent = nouveau_bios_gpio_entry(dev, tag);
+	uint32_t reg, shift, mask, value;
 
-	nv_wr32(dev, 0x001104, mask);
-	nv_mask(dev, 0x001144, mask, on ? mask : 0);
-}
+	if (!ent)
+		return -ENODEV;
 
-static void
-nv10_gpio_isr(struct drm_device *dev)
-{
-	u32 intr = nv_rd32(dev, 0x1104);
-	u32 hi = (intr & 0x0000ffff) >> 0;
-	u32 lo = (intr & 0xffff0000) >> 16;
+	if (!get_gpio_location(ent, &reg, &shift, &mask))
+		return -ENODEV;
 
-	nouveau_gpio_isr(dev, 0, hi | lo);
+	value = NVReadCRTC(dev, 0, reg) >> shift;
 
-	nv_wr32(dev, 0x001104, intr);
+	return (ent->invert ? 1 : 0) ^ (value & 1);
 }
 
 int
-nv10_gpio_init(struct drm_device *dev)
+nv10_gpio_set(struct drm_device *dev, enum dcb_gpio_tag tag, int state)
 {
-	nv_wr32(dev, 0x001140, 0x00000000);
-	nv_wr32(dev, 0x001100, 0xffffffff);
-	nv_wr32(dev, 0x001144, 0x00000000);
-	nv_wr32(dev, 0x001104, 0xffffffff);
-	nouveau_irq_register(dev, 28, nv10_gpio_isr); /* PBUS */
-	return 0;
-}
+	struct dcb_gpio_entry *ent = nouveau_bios_gpio_entry(dev, tag);
+	uint32_t reg, shift, mask, value;
 
-void
-nv10_gpio_fini(struct drm_device *dev)
-{
-	nv_wr32(dev, 0x001140, 0x00000000);
-	nv_wr32(dev, 0x001144, 0x00000000);
-	nouveau_irq_unregister(dev, 28);
+	if (!ent)
+		return -ENODEV;
+
+	if (!get_gpio_location(ent, &reg, &shift, &mask))
+		return -ENODEV;
+
+	value = ((ent->invert ? 1 : 0) ^ (state ? 1 : 0)) << shift;
+	mask = ~(mask << shift);
+
+	NVWriteCRTC(dev, 0, reg, value | (NVReadCRTC(dev, 0, reg) & mask));
+
+	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nv10_graph.c b/drivers/gpu/drm/nouveau/nv10_graph.c
index 7255e4a..0930c6c 100644
--- a/drivers/gpu/drm/nouveau/nv10_graph.c
+++ b/drivers/gpu/drm/nouveau/nv10_graph.c
@@ -708,8 +708,8 @@ static void nv10_graph_load_dma_vtxbuf(struct nouveau_channel *chan,
 		0x2c000000 | chan->id << 20 | subchan << 16 | 0x18c);
 	nv_wr32(dev, NV10_PGRAPH_FFINTFC_ST2_DL, inst);
 	nv_mask(dev, NV10_PGRAPH_CTX_CONTROL, 0, 0x10000);
-	nv_mask(dev, NV04_PGRAPH_FIFO, 0x00000001, 0x00000001);
-	nv_mask(dev, NV04_PGRAPH_FIFO, 0x00000001, 0x00000000);
+	nv04_graph_fifo_access(dev, true);
+	nv04_graph_fifo_access(dev, false);
 
 	/* Restore the FIFO state */
 	for (i = 0; i < ARRAY_SIZE(fifo); i++)
@@ -879,13 +879,13 @@ nv10_graph_context_del(struct nouveau_channel *chan, int engine)
 	unsigned long flags;
 
 	spin_lock_irqsave(&dev_priv->context_switch_lock, flags);
-	nv_mask(dev, NV04_PGRAPH_FIFO, 0x00000001, 0x00000000);
+	nv04_graph_fifo_access(dev, false);
 
 	/* Unload the context if it's the currently active one */
 	if (nv10_graph_channel(dev) == chan)
 		nv10_graph_unload_context(dev);
 
-	nv_mask(dev, NV04_PGRAPH_FIFO, 0x00000001, 0x00000001);
+	nv04_graph_fifo_access(dev, true);
 	spin_unlock_irqrestore(&dev_priv->context_switch_lock, flags);
 
 	/* Free the context resources */
@@ -957,13 +957,8 @@ nv10_graph_init(struct drm_device *dev, int engine)
 }
 
 static int
-nv10_graph_fini(struct drm_device *dev, int engine, bool suspend)
+nv10_graph_fini(struct drm_device *dev, int engine)
 {
-	nv_mask(dev, NV04_PGRAPH_FIFO, 0x00000001, 0x00000000);
-	if (!nv_wait(dev, NV04_PGRAPH_STATUS, ~0, 0) && suspend) {
-		nv_mask(dev, NV04_PGRAPH_FIFO, 0x00000001, 0x00000001);
-		return -EBUSY;
-	}
 	nv10_graph_unload_context(dev);
 	nv_wr32(dev, NV03_PGRAPH_INTR_EN, 0x00000000);
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/nv17_tv.c b/drivers/gpu/drm/nouveau/nv17_tv.c
index 696d7e7..3900ceb 100644
--- a/drivers/gpu/drm/nouveau/nv17_tv.c
+++ b/drivers/gpu/drm/nouveau/nv17_tv.c
@@ -30,7 +30,6 @@
 #include "nouveau_encoder.h"
 #include "nouveau_connector.h"
 #include "nouveau_crtc.h"
-#include "nouveau_gpio.h"
 #include "nouveau_hw.h"
 #include "nv17_tv.h"
 
@@ -38,6 +37,7 @@ static uint32_t nv42_tv_sample_load(struct drm_encoder *encoder)
 {
 	struct drm_device *dev = encoder->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_gpio_engine *gpio = &dev_priv->engine.gpio;
 	uint32_t testval, regoffset = nv04_dac_output_offset(encoder);
 	uint32_t gpio0, gpio1, fp_htotal, fp_hsync_start, fp_hsync_end,
 		fp_control, test_ctrl, dacclk, ctv_14, ctv_1c, ctv_6c;
@@ -53,8 +53,8 @@ static uint32_t nv42_tv_sample_load(struct drm_encoder *encoder)
 	head = (dacclk & 0x100) >> 8;
 
 	/* Save the previous state. */
-	gpio1 = nouveau_gpio_func_get(dev, DCB_GPIO_TVDAC1);
-	gpio0 = nouveau_gpio_func_get(dev, DCB_GPIO_TVDAC0);
+	gpio1 = gpio->get(dev, DCB_GPIO_TVDAC1);
+	gpio0 = gpio->get(dev, DCB_GPIO_TVDAC0);
 	fp_htotal = NVReadRAMDAC(dev, head, NV_PRAMDAC_FP_HTOTAL);
 	fp_hsync_start = NVReadRAMDAC(dev, head, NV_PRAMDAC_FP_HSYNC_START);
 	fp_hsync_end = NVReadRAMDAC(dev, head, NV_PRAMDAC_FP_HSYNC_END);
@@ -65,8 +65,8 @@ static uint32_t nv42_tv_sample_load(struct drm_encoder *encoder)
 	ctv_6c = NVReadRAMDAC(dev, head, 0x680c6c);
 
 	/* Prepare the DAC for load detection.  */
-	nouveau_gpio_func_set(dev, DCB_GPIO_TVDAC1, true);
-	nouveau_gpio_func_set(dev, DCB_GPIO_TVDAC0, true);
+	gpio->set(dev, DCB_GPIO_TVDAC1, true);
+	gpio->set(dev, DCB_GPIO_TVDAC0, true);
 
 	NVWriteRAMDAC(dev, head, NV_PRAMDAC_FP_HTOTAL, 1343);
 	NVWriteRAMDAC(dev, head, NV_PRAMDAC_FP_HSYNC_START, 1047);
@@ -111,8 +111,8 @@ static uint32_t nv42_tv_sample_load(struct drm_encoder *encoder)
 	NVWriteRAMDAC(dev, head, NV_PRAMDAC_FP_HSYNC_END, fp_hsync_end);
 	NVWriteRAMDAC(dev, head, NV_PRAMDAC_FP_HSYNC_START, fp_hsync_start);
 	NVWriteRAMDAC(dev, head, NV_PRAMDAC_FP_HTOTAL, fp_htotal);
-	nouveau_gpio_func_set(dev, DCB_GPIO_TVDAC1, gpio1);
-	nouveau_gpio_func_set(dev, DCB_GPIO_TVDAC0, gpio0);
+	gpio->set(dev, DCB_GPIO_TVDAC1, gpio1);
+	gpio->set(dev, DCB_GPIO_TVDAC0, gpio0);
 
 	return sample;
 }
@@ -357,6 +357,8 @@ static bool nv17_tv_mode_fixup(struct drm_encoder *encoder,
 static void  nv17_tv_dpms(struct drm_encoder *encoder, int mode)
 {
 	struct drm_device *dev = encoder->dev;
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_gpio_engine *gpio = &dev_priv->engine.gpio;
 	struct nv17_tv_state *regs = &to_tv_enc(encoder)->state;
 	struct nv17_tv_norm_params *tv_norm = get_tv_norm(encoder);
 
@@ -381,8 +383,8 @@ static void  nv17_tv_dpms(struct drm_encoder *encoder, int mode)
 
 	nv_load_ptv(dev, regs, 200);
 
-	nouveau_gpio_func_set(dev, DCB_GPIO_TVDAC1, mode == DRM_MODE_DPMS_ON);
-	nouveau_gpio_func_set(dev, DCB_GPIO_TVDAC0, mode == DRM_MODE_DPMS_ON);
+	gpio->set(dev, DCB_GPIO_TVDAC1, mode == DRM_MODE_DPMS_ON);
+	gpio->set(dev, DCB_GPIO_TVDAC0, mode == DRM_MODE_DPMS_ON);
 
 	nv04_dac_update_dacclk(encoder, mode == DRM_MODE_DPMS_ON);
 }
diff --git a/drivers/gpu/drm/nouveau/nv20_fb.c b/drivers/gpu/drm/nouveau/nv20_fb.c
deleted file mode 100644
index 19bd640..0000000
--- a/drivers/gpu/drm/nouveau/nv20_fb.c
+++ /dev/null
@@ -1,148 +0,0 @@
-#include "drmP.h"
-#include "drm.h"
-#include "nouveau_drv.h"
-#include "nouveau_drm.h"
-
-static struct drm_mm_node *
-nv20_fb_alloc_tag(struct drm_device *dev, uint32_t size)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_fb_engine *pfb = &dev_priv->engine.fb;
-	struct drm_mm_node *mem;
-	int ret;
-
-	ret = drm_mm_pre_get(&pfb->tag_heap);
-	if (ret)
-		return NULL;
-
-	spin_lock(&dev_priv->tile.lock);
-	mem = drm_mm_search_free(&pfb->tag_heap, size, 0, 0);
-	if (mem)
-		mem = drm_mm_get_block_atomic(mem, size, 0);
-	spin_unlock(&dev_priv->tile.lock);
-
-	return mem;
-}
-
-static void
-nv20_fb_free_tag(struct drm_device *dev, struct drm_mm_node **pmem)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct drm_mm_node *mem = *pmem;
-	if (mem) {
-		spin_lock(&dev_priv->tile.lock);
-		drm_mm_put_block(mem);
-		spin_unlock(&dev_priv->tile.lock);
-		*pmem = NULL;
-	}
-}
-
-void
-nv20_fb_init_tile_region(struct drm_device *dev, int i, uint32_t addr,
-			 uint32_t size, uint32_t pitch, uint32_t flags)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_tile_reg *tile = &dev_priv->tile.reg[i];
-	int bpp = (flags & NOUVEAU_GEM_TILE_32BPP ? 32 : 16);
-
-	tile->addr  = 0x00000001 | addr;
-	tile->limit = max(1u, addr + size) - 1;
-	tile->pitch = pitch;
-
-	/* Allocate some of the on-die tag memory, used to store Z
-	 * compression meta-data (most likely just a bitmap determining
-	 * if a given tile is compressed or not).
-	 */
-	if (flags & NOUVEAU_GEM_TILE_ZETA) {
-		tile->tag_mem = nv20_fb_alloc_tag(dev, size / 256);
-		if (tile->tag_mem) {
-			/* Enable Z compression */
-			tile->zcomp = tile->tag_mem->start;
-			if (dev_priv->chipset >= 0x25) {
-				if (bpp == 16)
-					tile->zcomp |= NV25_PFB_ZCOMP_MODE_16;
-				else
-					tile->zcomp |= NV25_PFB_ZCOMP_MODE_32;
-			} else {
-				tile->zcomp |= NV20_PFB_ZCOMP_EN;
-				if (bpp != 16)
-					tile->zcomp |= NV20_PFB_ZCOMP_MODE_32;
-			}
-		}
-
-		tile->addr |= 2;
-	}
-}
-
-void
-nv20_fb_free_tile_region(struct drm_device *dev, int i)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_tile_reg *tile = &dev_priv->tile.reg[i];
-
-	tile->addr = tile->limit = tile->pitch = tile->zcomp = 0;
-	nv20_fb_free_tag(dev, &tile->tag_mem);
-}
-
-void
-nv20_fb_set_tile_region(struct drm_device *dev, int i)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_tile_reg *tile = &dev_priv->tile.reg[i];
-
-	nv_wr32(dev, NV10_PFB_TLIMIT(i), tile->limit);
-	nv_wr32(dev, NV10_PFB_TSIZE(i), tile->pitch);
-	nv_wr32(dev, NV10_PFB_TILE(i), tile->addr);
-	nv_wr32(dev, NV20_PFB_ZCOMP(i), tile->zcomp);
-}
-
-int
-nv20_fb_vram_init(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	u32 mem_size = nv_rd32(dev, 0x10020c);
-	u32 pbus1218 = nv_rd32(dev, 0x001218);
-
-	dev_priv->vram_size = mem_size & 0xff000000;
-	switch (pbus1218 & 0x00000300) {
-	case 0x00000000: dev_priv->vram_type = NV_MEM_TYPE_SDRAM; break;
-	case 0x00000100: dev_priv->vram_type = NV_MEM_TYPE_DDR1; break;
-	case 0x00000200: dev_priv->vram_type = NV_MEM_TYPE_GDDR3; break;
-	case 0x00000300: dev_priv->vram_type = NV_MEM_TYPE_GDDR2; break;
-	}
-
-	return 0;
-}
-
-int
-nv20_fb_init(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_fb_engine *pfb = &dev_priv->engine.fb;
-	int i;
-
-	if (dev_priv->chipset >= 0x25)
-		drm_mm_init(&pfb->tag_heap, 0, 64 * 1024);
-	else
-		drm_mm_init(&pfb->tag_heap, 0, 32 * 1024);
-
-	/* Turn all the tiling regions off. */
-	pfb->num_tiles = NV10_PFB_TILE__SIZE;
-	for (i = 0; i < pfb->num_tiles; i++)
-		pfb->set_tile_region(dev, i);
-
-	return 0;
-}
-
-void
-nv20_fb_takedown(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_fb_engine *pfb = &dev_priv->engine.fb;
-	int i;
-
-	for (i = 0; i < pfb->num_tiles; i++)
-		pfb->free_tile_region(dev, i);
-
-	drm_mm_takedown(&pfb->tag_heap);
-}
diff --git a/drivers/gpu/drm/nouveau/nv20_graph.c b/drivers/gpu/drm/nouveau/nv20_graph.c
index 183e375..affc7d7 100644
--- a/drivers/gpu/drm/nouveau/nv20_graph.c
+++ b/drivers/gpu/drm/nouveau/nv20_graph.c
@@ -454,13 +454,13 @@ nv20_graph_context_del(struct nouveau_channel *chan, int engine)
 	unsigned long flags;
 
 	spin_lock_irqsave(&dev_priv->context_switch_lock, flags);
-	nv_mask(dev, NV04_PGRAPH_FIFO, 0x00000001, 0x00000000);
+	nv04_graph_fifo_access(dev, false);
 
 	/* Unload the context if it's the currently active one */
 	if (nv10_graph_channel(dev) == chan)
 		nv20_graph_unload_context(dev);
 
-	nv_mask(dev, NV04_PGRAPH_FIFO, 0x00000001, 0x00000001);
+	nv04_graph_fifo_access(dev, true);
 	spin_unlock_irqrestore(&dev_priv->context_switch_lock, flags);
 
 	/* Free the context resources */
@@ -654,13 +654,8 @@ nv30_graph_init(struct drm_device *dev, int engine)
 }
 
 int
-nv20_graph_fini(struct drm_device *dev, int engine, bool suspend)
+nv20_graph_fini(struct drm_device *dev, int engine)
 {
-	nv_mask(dev, NV04_PGRAPH_FIFO, 0x00000001, 0x00000000);
-	if (!nv_wait(dev, NV04_PGRAPH_STATUS, ~0, 0) && suspend) {
-		nv_mask(dev, NV04_PGRAPH_FIFO, 0x00000001, 0x00000001);
-		return -EBUSY;
-	}
 	nv20_graph_unload_context(dev);
 	nv_wr32(dev, NV03_PGRAPH_INTR_EN, 0x00000000);
 	return 0;
@@ -758,7 +753,6 @@ nv20_graph_create(struct drm_device *dev)
 			break;
 		default:
 			NV_ERROR(dev, "PGRAPH: unknown chipset\n");
-			kfree(pgraph);
 			return 0;
 		}
 	} else {
@@ -780,7 +774,6 @@ nv20_graph_create(struct drm_device *dev)
 			break;
 		default:
 			NV_ERROR(dev, "PGRAPH: unknown chipset\n");
-			kfree(pgraph);
 			return 0;
 		}
 	}
diff --git a/drivers/gpu/drm/nouveau/nv31_mpeg.c b/drivers/gpu/drm/nouveau/nv31_mpeg.c
deleted file mode 100644
index 6f06a07..0000000
--- a/drivers/gpu/drm/nouveau/nv31_mpeg.c
+++ /dev/null
@@ -1,344 +0,0 @@
-/*
- * Copyright 2011 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Ben Skeggs
- */
-
-#include "drmP.h"
-#include "nouveau_drv.h"
-#include "nouveau_ramht.h"
-
-struct nv31_mpeg_engine {
-	struct nouveau_exec_engine base;
-	atomic_t refcount;
-};
-
-
-static int
-nv31_mpeg_context_new(struct nouveau_channel *chan, int engine)
-{
-	struct nv31_mpeg_engine *pmpeg = nv_engine(chan->dev, engine);
-
-	if (!atomic_add_unless(&pmpeg->refcount, 1, 1))
-		return -EBUSY;
-
-	chan->engctx[engine] = (void *)0xdeadcafe;
-	return 0;
-}
-
-static void
-nv31_mpeg_context_del(struct nouveau_channel *chan, int engine)
-{
-	struct nv31_mpeg_engine *pmpeg = nv_engine(chan->dev, engine);
-	atomic_dec(&pmpeg->refcount);
-	chan->engctx[engine] = NULL;
-}
-
-static int
-nv40_mpeg_context_new(struct nouveau_channel *chan, int engine)
-{
-	struct drm_device *dev = chan->dev;
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_gpuobj *ctx = NULL;
-	unsigned long flags;
-	int ret;
-
-	NV_DEBUG(dev, "ch%d\n", chan->id);
-
-	ret = nouveau_gpuobj_new(dev, NULL, 264 * 4, 16, NVOBJ_FLAG_ZERO_ALLOC |
-				 NVOBJ_FLAG_ZERO_FREE, &ctx);
-	if (ret)
-		return ret;
-
-	nv_wo32(ctx, 0x78, 0x02001ec1);
-
-	spin_lock_irqsave(&dev_priv->context_switch_lock, flags);
-	nv_mask(dev, 0x002500, 0x00000001, 0x00000000);
-	if ((nv_rd32(dev, 0x003204) & 0x1f) == chan->id)
-		nv_wr32(dev, 0x00330c, ctx->pinst >> 4);
-	nv_wo32(chan->ramfc, 0x54, ctx->pinst >> 4);
-	nv_mask(dev, 0x002500, 0x00000001, 0x00000001);
-	spin_unlock_irqrestore(&dev_priv->context_switch_lock, flags);
-
-	chan->engctx[engine] = ctx;
-	return 0;
-}
-
-static void
-nv40_mpeg_context_del(struct nouveau_channel *chan, int engine)
-{
-	struct drm_nouveau_private *dev_priv = chan->dev->dev_private;
-	struct nouveau_gpuobj *ctx = chan->engctx[engine];
-	struct drm_device *dev = chan->dev;
-	unsigned long flags;
-	u32 inst = 0x80000000 | (ctx->pinst >> 4);
-
-	spin_lock_irqsave(&dev_priv->context_switch_lock, flags);
-	nv_mask(dev, 0x00b32c, 0x00000001, 0x00000000);
-	if (nv_rd32(dev, 0x00b318) == inst)
-		nv_mask(dev, 0x00b318, 0x80000000, 0x00000000);
-	nv_mask(dev, 0x00b32c, 0x00000001, 0x00000001);
-	spin_unlock_irqrestore(&dev_priv->context_switch_lock, flags);
-
-	nouveau_gpuobj_ref(NULL, &ctx);
-	chan->engctx[engine] = NULL;
-}
-
-static int
-nv31_mpeg_object_new(struct nouveau_channel *chan, int engine,
-		      u32 handle, u16 class)
-{
-	struct drm_device *dev = chan->dev;
-	struct nouveau_gpuobj *obj = NULL;
-	int ret;
-
-	ret = nouveau_gpuobj_new(dev, chan, 20, 16, NVOBJ_FLAG_ZERO_ALLOC |
-				 NVOBJ_FLAG_ZERO_FREE, &obj);
-	if (ret)
-		return ret;
-	obj->engine = 2;
-	obj->class  = class;
-
-	nv_wo32(obj, 0x00, class);
-
-	ret = nouveau_ramht_insert(chan, handle, obj);
-	nouveau_gpuobj_ref(NULL, &obj);
-	return ret;
-}
-
-static int
-nv31_mpeg_init(struct drm_device *dev, int engine)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nv31_mpeg_engine *pmpeg = nv_engine(dev, engine);
-	int i;
-
-	/* VPE init */
-	nv_mask(dev, 0x000200, 0x00000002, 0x00000000);
-	nv_mask(dev, 0x000200, 0x00000002, 0x00000002);
-	nv_wr32(dev, 0x00b0e0, 0x00000020); /* nvidia: rd 0x01, wr 0x20 */
-	nv_wr32(dev, 0x00b0e8, 0x00000020); /* nvidia: rd 0x01, wr 0x20 */
-
-	for (i = 0; i < dev_priv->engine.fb.num_tiles; i++)
-		pmpeg->base.set_tile_region(dev, i);
-
-	/* PMPEG init */
-	nv_wr32(dev, 0x00b32c, 0x00000000);
-	nv_wr32(dev, 0x00b314, 0x00000100);
-	nv_wr32(dev, 0x00b220, nv44_graph_class(dev) ? 0x00000044 : 0x00000031);
-	nv_wr32(dev, 0x00b300, 0x02001ec1);
-	nv_mask(dev, 0x00b32c, 0x00000001, 0x00000001);
-
-	nv_wr32(dev, 0x00b100, 0xffffffff);
-	nv_wr32(dev, 0x00b140, 0xffffffff);
-
-	if (!nv_wait(dev, 0x00b200, 0x00000001, 0x00000000)) {
-		NV_ERROR(dev, "PMPEG init: 0x%08x\n", nv_rd32(dev, 0x00b200));
-		return -EBUSY;
-	}
-
-	return 0;
-}
-
-static int
-nv31_mpeg_fini(struct drm_device *dev, int engine, bool suspend)
-{
-	/*XXX: context save? */
-	nv_mask(dev, 0x00b32c, 0x00000001, 0x00000000);
-	nv_wr32(dev, 0x00b140, 0x00000000);
-	return 0;
-}
-
-static int
-nv31_mpeg_mthd_dma(struct nouveau_channel *chan, u32 class, u32 mthd, u32 data)
-{
-	struct drm_device *dev = chan->dev;
-	u32 inst = data << 4;
-	u32 dma0 = nv_ri32(dev, inst + 0);
-	u32 dma1 = nv_ri32(dev, inst + 4);
-	u32 dma2 = nv_ri32(dev, inst + 8);
-	u32 base = (dma2 & 0xfffff000) | (dma0 >> 20);
-	u32 size = dma1 + 1;
-
-	/* only allow linear DMA objects */
-	if (!(dma0 & 0x00002000))
-		return -EINVAL;
-
-	if (mthd == 0x0190) {
-		/* DMA_CMD */
-		nv_mask(dev, 0x00b300, 0x00030000, (dma0 & 0x00030000));
-		nv_wr32(dev, 0x00b334, base);
-		nv_wr32(dev, 0x00b324, size);
-	} else
-	if (mthd == 0x01a0) {
-		/* DMA_DATA */
-		nv_mask(dev, 0x00b300, 0x000c0000, (dma0 & 0x00030000) << 2);
-		nv_wr32(dev, 0x00b360, base);
-		nv_wr32(dev, 0x00b364, size);
-	} else {
-		/* DMA_IMAGE, VRAM only */
-		if (dma0 & 0x000c0000)
-			return -EINVAL;
-
-		nv_wr32(dev, 0x00b370, base);
-		nv_wr32(dev, 0x00b374, size);
-	}
-
-	return 0;
-}
-
-static int
-nv31_mpeg_isr_chid(struct drm_device *dev, u32 inst)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_gpuobj *ctx;
-	unsigned long flags;
-	int i;
-
-	/* hardcode drm channel id on nv3x, so swmthd lookup works */
-	if (dev_priv->card_type < NV_40)
-		return 0;
-
-	spin_lock_irqsave(&dev_priv->channels.lock, flags);
-	for (i = 0; i < dev_priv->engine.fifo.channels; i++) {
-		if (!dev_priv->channels.ptr[i])
-			continue;
-
-		ctx = dev_priv->channels.ptr[i]->engctx[NVOBJ_ENGINE_MPEG];
-		if (ctx && ctx->pinst == inst)
-			break;
-	}
-	spin_unlock_irqrestore(&dev_priv->channels.lock, flags);
-	return i;
-}
-
-static void
-nv31_vpe_set_tile_region(struct drm_device *dev, int i)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_tile_reg *tile = &dev_priv->tile.reg[i];
-
-	nv_wr32(dev, 0x00b008 + (i * 0x10), tile->pitch);
-	nv_wr32(dev, 0x00b004 + (i * 0x10), tile->limit);
-	nv_wr32(dev, 0x00b000 + (i * 0x10), tile->addr);
-}
-
-static void
-nv31_mpeg_isr(struct drm_device *dev)
-{
-	u32 inst = (nv_rd32(dev, 0x00b318) & 0x000fffff) << 4;
-	u32 chid = nv31_mpeg_isr_chid(dev, inst);
-	u32 stat = nv_rd32(dev, 0x00b100);
-	u32 type = nv_rd32(dev, 0x00b230);
-	u32 mthd = nv_rd32(dev, 0x00b234);
-	u32 data = nv_rd32(dev, 0x00b238);
-	u32 show = stat;
-
-	if (stat & 0x01000000) {
-		/* happens on initial binding of the object */
-		if (type == 0x00000020 && mthd == 0x0000) {
-			nv_mask(dev, 0x00b308, 0x00000000, 0x00000000);
-			show &= ~0x01000000;
-		}
-
-		if (type == 0x00000010) {
-			if (!nouveau_gpuobj_mthd_call2(dev, chid, 0x3174, mthd, data))
-				show &= ~0x01000000;
-		}
-	}
-
-	nv_wr32(dev, 0x00b100, stat);
-	nv_wr32(dev, 0x00b230, 0x00000001);
-
-	if (show && nouveau_ratelimit()) {
-		NV_INFO(dev, "PMPEG: Ch %d [0x%08x] 0x%08x 0x%08x 0x%08x 0x%08x\n",
-			chid, inst, stat, type, mthd, data);
-	}
-}
-
-static void
-nv31_vpe_isr(struct drm_device *dev)
-{
-	if (nv_rd32(dev, 0x00b100))
-		nv31_mpeg_isr(dev);
-
-	if (nv_rd32(dev, 0x00b800)) {
-		u32 stat = nv_rd32(dev, 0x00b800);
-		NV_INFO(dev, "PMSRCH: 0x%08x\n", stat);
-		nv_wr32(dev, 0xb800, stat);
-	}
-}
-
-static void
-nv31_mpeg_destroy(struct drm_device *dev, int engine)
-{
-	struct nv31_mpeg_engine *pmpeg = nv_engine(dev, engine);
-
-	nouveau_irq_unregister(dev, 0);
-
-	NVOBJ_ENGINE_DEL(dev, MPEG);
-	kfree(pmpeg);
-}
-
-int
-nv31_mpeg_create(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nv31_mpeg_engine *pmpeg;
-
-	pmpeg = kzalloc(sizeof(*pmpeg), GFP_KERNEL);
-	if (!pmpeg)
-		return -ENOMEM;
-	atomic_set(&pmpeg->refcount, 0);
-
-	pmpeg->base.destroy = nv31_mpeg_destroy;
-	pmpeg->base.init = nv31_mpeg_init;
-	pmpeg->base.fini = nv31_mpeg_fini;
-	if (dev_priv->card_type < NV_40) {
-		pmpeg->base.context_new = nv31_mpeg_context_new;
-		pmpeg->base.context_del = nv31_mpeg_context_del;
-	} else {
-		pmpeg->base.context_new = nv40_mpeg_context_new;
-		pmpeg->base.context_del = nv40_mpeg_context_del;
-	}
-	pmpeg->base.object_new = nv31_mpeg_object_new;
-
-	/* ISR vector, PMC_ENABLE bit,  and TILE regs are shared between
-	 * all VPE engines, for this driver's purposes the PMPEG engine
-	 * will be treated as the "master" and handle the global VPE
-	 * bits too
-	 */
-	pmpeg->base.set_tile_region = nv31_vpe_set_tile_region;
-	nouveau_irq_register(dev, 0, nv31_vpe_isr);
-
-	NVOBJ_ENGINE_ADD(dev, MPEG, &pmpeg->base);
-	NVOBJ_CLASS(dev, 0x3174, MPEG);
-	NVOBJ_MTHD (dev, 0x3174, 0x0190, nv31_mpeg_mthd_dma);
-	NVOBJ_MTHD (dev, 0x3174, 0x01a0, nv31_mpeg_mthd_dma);
-	NVOBJ_MTHD (dev, 0x3174, 0x01b0, nv31_mpeg_mthd_dma);
-
-#if 0
-	NVOBJ_ENGINE_ADD(dev, ME, &pme->base);
-	NVOBJ_CLASS(dev, 0x4075, ME);
-#endif
-	return 0;
-
-}
diff --git a/drivers/gpu/drm/nouveau/nv40_fb.c b/drivers/gpu/drm/nouveau/nv40_fb.c
index 7fbcb33..f0ac2a7 100644
--- a/drivers/gpu/drm/nouveau/nv40_fb.c
+++ b/drivers/gpu/drm/nouveau/nv40_fb.c
@@ -72,51 +72,6 @@ nv44_fb_init_gart(struct drm_device *dev)
 }
 
 int
-nv40_fb_vram_init(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-
-	/* 0x001218 is actually present on a few other NV4X I looked at,
-	 * and even contains sane values matching 0x100474.  From looking
-	 * at various vbios images however, this isn't the case everywhere.
-	 * So, I chose to use the same regs I've seen NVIDIA reading around
-	 * the memory detection, hopefully that'll get us the right numbers
-	 */
-	if (dev_priv->chipset == 0x40) {
-		u32 pbus1218 = nv_rd32(dev, 0x001218);
-		switch (pbus1218 & 0x00000300) {
-		case 0x00000000: dev_priv->vram_type = NV_MEM_TYPE_SDRAM; break;
-		case 0x00000100: dev_priv->vram_type = NV_MEM_TYPE_DDR1; break;
-		case 0x00000200: dev_priv->vram_type = NV_MEM_TYPE_GDDR3; break;
-		case 0x00000300: dev_priv->vram_type = NV_MEM_TYPE_DDR2; break;
-		}
-	} else
-	if (dev_priv->chipset == 0x49 || dev_priv->chipset == 0x4b) {
-		u32 pfb914 = nv_rd32(dev, 0x100914);
-		switch (pfb914 & 0x00000003) {
-		case 0x00000000: dev_priv->vram_type = NV_MEM_TYPE_DDR1; break;
-		case 0x00000001: dev_priv->vram_type = NV_MEM_TYPE_DDR2; break;
-		case 0x00000002: dev_priv->vram_type = NV_MEM_TYPE_GDDR3; break;
-		case 0x00000003: break;
-		}
-	} else
-	if (dev_priv->chipset != 0x4e) {
-		u32 pfb474 = nv_rd32(dev, 0x100474);
-		if (pfb474 & 0x00000004)
-			dev_priv->vram_type = NV_MEM_TYPE_GDDR3;
-		if (pfb474 & 0x00000002)
-			dev_priv->vram_type = NV_MEM_TYPE_DDR2;
-		if (pfb474 & 0x00000001)
-			dev_priv->vram_type = NV_MEM_TYPE_DDR1;
-	} else {
-		dev_priv->vram_type = NV_MEM_TYPE_STOLEN;
-	}
-
-	dev_priv->vram_size = nv_rd32(dev, 0x10020c) & 0xff000000;
-	return 0;
-}
-
-int
 nv40_fb_init(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
diff --git a/drivers/gpu/drm/nouveau/nv40_graph.c b/drivers/gpu/drm/nouveau/nv40_graph.c
index ba14a93..5beb01b 100644
--- a/drivers/gpu/drm/nouveau/nv40_graph.c
+++ b/drivers/gpu/drm/nouveau/nv40_graph.c
@@ -35,6 +35,89 @@ struct nv40_graph_engine {
 	u32 grctx_size;
 };
 
+static struct nouveau_channel *
+nv40_graph_channel(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_gpuobj *grctx;
+	uint32_t inst;
+	int i;
+
+	inst = nv_rd32(dev, NV40_PGRAPH_CTXCTL_CUR);
+	if (!(inst & NV40_PGRAPH_CTXCTL_CUR_LOADED))
+		return NULL;
+	inst = (inst & NV40_PGRAPH_CTXCTL_CUR_INSTANCE) << 4;
+
+	for (i = 0; i < dev_priv->engine.fifo.channels; i++) {
+		if (!dev_priv->channels.ptr[i])
+			continue;
+
+		grctx = dev_priv->channels.ptr[i]->engctx[NVOBJ_ENGINE_GR];
+		if (grctx && grctx->pinst == inst)
+			return dev_priv->channels.ptr[i];
+	}
+
+	return NULL;
+}
+
+static int
+nv40_graph_transfer_context(struct drm_device *dev, uint32_t inst, int save)
+{
+	uint32_t old_cp, tv = 1000, tmp;
+	int i;
+
+	old_cp = nv_rd32(dev, NV20_PGRAPH_CHANNEL_CTX_POINTER);
+	nv_wr32(dev, NV20_PGRAPH_CHANNEL_CTX_POINTER, inst);
+
+	tmp  = nv_rd32(dev, NV40_PGRAPH_CTXCTL_0310);
+	tmp |= save ? NV40_PGRAPH_CTXCTL_0310_XFER_SAVE :
+		      NV40_PGRAPH_CTXCTL_0310_XFER_LOAD;
+	nv_wr32(dev, NV40_PGRAPH_CTXCTL_0310, tmp);
+
+	tmp  = nv_rd32(dev, NV40_PGRAPH_CTXCTL_0304);
+	tmp |= NV40_PGRAPH_CTXCTL_0304_XFER_CTX;
+	nv_wr32(dev, NV40_PGRAPH_CTXCTL_0304, tmp);
+
+	nouveau_wait_for_idle(dev);
+
+	for (i = 0; i < tv; i++) {
+		if (nv_rd32(dev, NV40_PGRAPH_CTXCTL_030C) == 0)
+			break;
+	}
+
+	nv_wr32(dev, NV20_PGRAPH_CHANNEL_CTX_POINTER, old_cp);
+
+	if (i == tv) {
+		uint32_t ucstat = nv_rd32(dev, NV40_PGRAPH_CTXCTL_UCODE_STAT);
+		NV_ERROR(dev, "Failed: Instance=0x%08x Save=%d\n", inst, save);
+		NV_ERROR(dev, "IP: 0x%02x, Opcode: 0x%08x\n",
+			 ucstat >> NV40_PGRAPH_CTXCTL_UCODE_STAT_IP_SHIFT,
+			 ucstat  & NV40_PGRAPH_CTXCTL_UCODE_STAT_OP_MASK);
+		NV_ERROR(dev, "0x40030C = 0x%08x\n",
+			 nv_rd32(dev, NV40_PGRAPH_CTXCTL_030C));
+		return -EBUSY;
+	}
+
+	return 0;
+}
+
+static int
+nv40_graph_unload_context(struct drm_device *dev)
+{
+	uint32_t inst;
+	int ret;
+
+	inst = nv_rd32(dev, NV40_PGRAPH_CTXCTL_CUR);
+	if (!(inst & NV40_PGRAPH_CTXCTL_CUR_LOADED))
+		return 0;
+	inst &= NV40_PGRAPH_CTXCTL_CUR_INSTANCE;
+
+	ret = nv40_graph_transfer_context(dev, inst, 1);
+
+	nv_wr32(dev, NV40_PGRAPH_CTXCTL_CUR, inst);
+	return ret;
+}
+
 static int
 nv40_graph_context_new(struct nouveau_channel *chan, int engine)
 {
@@ -80,16 +163,16 @@ nv40_graph_context_del(struct nouveau_channel *chan, int engine)
 	struct nouveau_gpuobj *grctx = chan->engctx[engine];
 	struct drm_device *dev = chan->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	u32 inst = 0x01000000 | (grctx->pinst >> 4);
 	unsigned long flags;
 
 	spin_lock_irqsave(&dev_priv->context_switch_lock, flags);
-	nv_mask(dev, 0x400720, 0x00000000, 0x00000001);
-	if (nv_rd32(dev, 0x40032c) == inst)
-		nv_mask(dev, 0x40032c, 0x01000000, 0x00000000);
-	if (nv_rd32(dev, 0x400330) == inst)
-		nv_mask(dev, 0x400330, 0x01000000, 0x00000000);
-	nv_mask(dev, 0x400720, 0x00000001, 0x00000001);
+	nv04_graph_fifo_access(dev, false);
+
+	/* Unload the context if it's the currently active one */
+	if (nv40_graph_channel(dev) == chan)
+		nv40_graph_unload_context(dev);
+
+	nv04_graph_fifo_access(dev, true);
 	spin_unlock_irqrestore(&dev_priv->context_switch_lock, flags);
 
 	/* Free the context resources */
@@ -346,20 +429,9 @@ nv40_graph_init(struct drm_device *dev, int engine)
 }
 
 static int
-nv40_graph_fini(struct drm_device *dev, int engine, bool suspend)
+nv40_graph_fini(struct drm_device *dev, int engine)
 {
-	u32 inst = nv_rd32(dev, 0x40032c);
-	if (inst & 0x01000000) {
-		nv_wr32(dev, 0x400720, 0x00000000);
-		nv_wr32(dev, 0x400784, inst);
-		nv_mask(dev, 0x400310, 0x00000020, 0x00000020);
-		nv_mask(dev, 0x400304, 0x00000001, 0x00000001);
-		if (!nv_wait(dev, 0x400300, 0x00000001, 0x00000000)) {
-			u32 insn = nv_rd32(dev, 0x400308);
-			NV_ERROR(dev, "PGRAPH: ctxprog timeout 0x%08x\n", insn);
-		}
-		nv_mask(dev, 0x40032c, 0x01000000, 0x00000000);
-	}
+	nv40_graph_unload_context(dev);
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/nouveau/nv40_pm.c b/drivers/gpu/drm/nouveau/nv40_pm.c
deleted file mode 100644
index c761538..0000000
--- a/drivers/gpu/drm/nouveau/nv40_pm.c
+++ /dev/null
@@ -1,394 +0,0 @@
-/*
- * Copyright 2011 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Ben Skeggs
- */
-
-#include "drmP.h"
-#include "nouveau_drv.h"
-#include "nouveau_bios.h"
-#include "nouveau_pm.h"
-#include "nouveau_hw.h"
-
-#define min2(a,b) ((a) < (b) ? (a) : (b))
-
-static u32
-read_pll_1(struct drm_device *dev, u32 reg)
-{
-	u32 ctrl = nv_rd32(dev, reg + 0x00);
-	int P = (ctrl & 0x00070000) >> 16;
-	int N = (ctrl & 0x0000ff00) >> 8;
-	int M = (ctrl & 0x000000ff) >> 0;
-	u32 ref = 27000, clk = 0;
-
-	if (ctrl & 0x80000000)
-		clk = ref * N / M;
-
-	return clk >> P;
-}
-
-static u32
-read_pll_2(struct drm_device *dev, u32 reg)
-{
-	u32 ctrl = nv_rd32(dev, reg + 0x00);
-	u32 coef = nv_rd32(dev, reg + 0x04);
-	int N2 = (coef & 0xff000000) >> 24;
-	int M2 = (coef & 0x00ff0000) >> 16;
-	int N1 = (coef & 0x0000ff00) >> 8;
-	int M1 = (coef & 0x000000ff) >> 0;
-	int P = (ctrl & 0x00070000) >> 16;
-	u32 ref = 27000, clk = 0;
-
-	if ((ctrl & 0x80000000) && M1) {
-		clk = ref * N1 / M1;
-		if ((ctrl & 0x40000100) == 0x40000000) {
-			if (M2)
-				clk = clk * N2 / M2;
-			else
-				clk = 0;
-		}
-	}
-
-	return clk >> P;
-}
-
-static u32
-read_clk(struct drm_device *dev, u32 src)
-{
-	switch (src) {
-	case 3:
-		return read_pll_2(dev, 0x004000);
-	case 2:
-		return read_pll_1(dev, 0x004008);
-	default:
-		break;
-	}
-
-	return 0;
-}
-
-int
-nv40_pm_clocks_get(struct drm_device *dev, struct nouveau_pm_level *perflvl)
-{
-	u32 ctrl = nv_rd32(dev, 0x00c040);
-
-	perflvl->core   = read_clk(dev, (ctrl & 0x00000003) >> 0);
-	perflvl->shader = read_clk(dev, (ctrl & 0x00000030) >> 4);
-	perflvl->memory = read_pll_2(dev, 0x4020);
-	return 0;
-}
-
-struct nv40_pm_state {
-	u32 ctrl;
-	u32 npll_ctrl;
-	u32 npll_coef;
-	u32 spll;
-	u32 mpll_ctrl;
-	u32 mpll_coef;
-};
-
-static int
-nv40_calc_pll(struct drm_device *dev, u32 reg, struct pll_lims *pll,
-	      u32 clk, int *N1, int *M1, int *N2, int *M2, int *log2P)
-{
-	struct nouveau_pll_vals coef;
-	int ret;
-
-	ret = get_pll_limits(dev, reg, pll);
-	if (ret)
-		return ret;
-
-	if (clk < pll->vco1.maxfreq)
-		pll->vco2.maxfreq = 0;
-
-	ret = nouveau_calc_pll_mnp(dev, pll, clk, &coef);
-	if (ret == 0)
-		return -ERANGE;
-
-	*N1 = coef.N1;
-	*M1 = coef.M1;
-	if (N2 && M2) {
-		if (pll->vco2.maxfreq) {
-			*N2 = coef.N2;
-			*M2 = coef.M2;
-		} else {
-			*N2 = 1;
-			*M2 = 1;
-		}
-	}
-	*log2P = coef.log2P;
-	return 0;
-}
-
-void *
-nv40_pm_clocks_pre(struct drm_device *dev, struct nouveau_pm_level *perflvl)
-{
-	struct nv40_pm_state *info;
-	struct pll_lims pll;
-	int N1, N2, M1, M2, log2P;
-	int ret;
-
-	info = kmalloc(sizeof(*info), GFP_KERNEL);
-	if (!info)
-		return ERR_PTR(-ENOMEM);
-
-	/* core/geometric clock */
-	ret = nv40_calc_pll(dev, 0x004000, &pll, perflvl->core,
-			    &N1, &M1, &N2, &M2, &log2P);
-	if (ret < 0)
-		goto out;
-
-	if (N2 == M2) {
-		info->npll_ctrl = 0x80000100 | (log2P << 16);
-		info->npll_coef = (N1 << 8) | M1;
-	} else {
-		info->npll_ctrl = 0xc0000000 | (log2P << 16);
-		info->npll_coef = (N2 << 24) | (M2 << 16) | (N1 << 8) | M1;
-	}
-
-	/* use the second PLL for shader/rop clock, if it differs from core */
-	if (perflvl->shader && perflvl->shader != perflvl->core) {
-		ret = nv40_calc_pll(dev, 0x004008, &pll, perflvl->shader,
-				    &N1, &M1, NULL, NULL, &log2P);
-		if (ret < 0)
-			goto out;
-
-		info->spll = 0xc0000000 | (log2P << 16) | (N1 << 8) | M1;
-		info->ctrl = 0x00000223;
-	} else {
-		info->spll = 0x00000000;
-		info->ctrl = 0x00000333;
-	}
-
-	/* memory clock */
-	if (!perflvl->memory) {
-		info->mpll_ctrl = 0x00000000;
-		goto out;
-	}
-
-	ret = nv40_calc_pll(dev, 0x004020, &pll, perflvl->memory,
-			    &N1, &M1, &N2, &M2, &log2P);
-	if (ret < 0)
-		goto out;
-
-	info->mpll_ctrl  = 0x80000000 | (log2P << 16);
-	info->mpll_ctrl |= min2(pll.log2p_bias + log2P, pll.max_log2p) << 20;
-	if (N2 == M2) {
-		info->mpll_ctrl |= 0x00000100;
-		info->mpll_coef  = (N1 << 8) | M1;
-	} else {
-		info->mpll_ctrl |= 0x40000000;
-		info->mpll_coef  = (N2 << 24) | (M2 << 16) | (N1 << 8) | M1;
-	}
-
-out:
-	if (ret < 0) {
-		kfree(info);
-		info = ERR_PTR(ret);
-	}
-	return info;
-}
-
-static bool
-nv40_pm_gr_idle(void *data)
-{
-	struct drm_device *dev = data;
-
-	if ((nv_rd32(dev, 0x400760) & 0x000000f0) >> 4 !=
-	    (nv_rd32(dev, 0x400760) & 0x0000000f))
-		return false;
-
-	if (nv_rd32(dev, 0x400700))
-		return false;
-
-	return true;
-}
-
-int
-nv40_pm_clocks_set(struct drm_device *dev, void *pre_state)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nv40_pm_state *info = pre_state;
-	unsigned long flags;
-	struct bit_entry M;
-	u32 crtc_mask = 0;
-	u8 sr1[2];
-	int i, ret = -EAGAIN;
-
-	/* determine which CRTCs are active, fetch VGA_SR1 for each */
-	for (i = 0; i < 2; i++) {
-		u32 vbl = nv_rd32(dev, 0x600808 + (i * 0x2000));
-		u32 cnt = 0;
-		do {
-			if (vbl != nv_rd32(dev, 0x600808 + (i * 0x2000))) {
-				nv_wr08(dev, 0x0c03c4 + (i * 0x2000), 0x01);
-				sr1[i] = nv_rd08(dev, 0x0c03c5 + (i * 0x2000));
-				if (!(sr1[i] & 0x20))
-					crtc_mask |= (1 << i);
-				break;
-			}
-			udelay(1);
-		} while (cnt++ < 32);
-	}
-
-	/* halt and idle engines */
-	spin_lock_irqsave(&dev_priv->context_switch_lock, flags);
-	nv_mask(dev, 0x002500, 0x00000001, 0x00000000);
-	if (!nv_wait(dev, 0x002500, 0x00000010, 0x00000000))
-		goto resume;
-	nv_mask(dev, 0x003220, 0x00000001, 0x00000000);
-	if (!nv_wait(dev, 0x003220, 0x00000010, 0x00000000))
-		goto resume;
-	nv_mask(dev, 0x003200, 0x00000001, 0x00000000);
-	nv04_fifo_cache_pull(dev, false);
-
-	if (!nv_wait_cb(dev, nv40_pm_gr_idle, dev))
-		goto resume;
-
-	ret = 0;
-
-	/* set engine clocks */
-	nv_mask(dev, 0x00c040, 0x00000333, 0x00000000);
-	nv_wr32(dev, 0x004004, info->npll_coef);
-	nv_mask(dev, 0x004000, 0xc0070100, info->npll_ctrl);
-	nv_mask(dev, 0x004008, 0xc007ffff, info->spll);
-	mdelay(5);
-	nv_mask(dev, 0x00c040, 0x00000333, info->ctrl);
-
-	if (!info->mpll_ctrl)
-		goto resume;
-
-	/* wait for vblank start on active crtcs, disable memory access */
-	for (i = 0; i < 2; i++) {
-		if (!(crtc_mask & (1 << i)))
-			continue;
-		nv_wait(dev, 0x600808 + (i * 0x2000), 0x00010000, 0x00000000);
-		nv_wait(dev, 0x600808 + (i * 0x2000), 0x00010000, 0x00010000);
-		nv_wr08(dev, 0x0c03c4 + (i * 0x2000), 0x01);
-		nv_wr08(dev, 0x0c03c5 + (i * 0x2000), sr1[i] | 0x20);
-	}
-
-	/* prepare ram for reclocking */
-	nv_wr32(dev, 0x1002d4, 0x00000001); /* precharge */
-	nv_wr32(dev, 0x1002d0, 0x00000001); /* refresh */
-	nv_wr32(dev, 0x1002d0, 0x00000001); /* refresh */
-	nv_mask(dev, 0x100210, 0x80000000, 0x00000000); /* no auto refresh */
-	nv_wr32(dev, 0x1002dc, 0x00000001); /* enable self-refresh */
-
-	/* change the PLL of each memory partition */
-	nv_mask(dev, 0x00c040, 0x0000c000, 0x00000000);
-	switch (dev_priv->chipset) {
-	case 0x40:
-	case 0x45:
-	case 0x41:
-	case 0x42:
-	case 0x47:
-		nv_mask(dev, 0x004044, 0xc0771100, info->mpll_ctrl);
-		nv_mask(dev, 0x00402c, 0xc0771100, info->mpll_ctrl);
-		nv_wr32(dev, 0x004048, info->mpll_coef);
-		nv_wr32(dev, 0x004030, info->mpll_coef);
-	case 0x43:
-	case 0x49:
-	case 0x4b:
-		nv_mask(dev, 0x004038, 0xc0771100, info->mpll_ctrl);
-		nv_wr32(dev, 0x00403c, info->mpll_coef);
-	default:
-		nv_mask(dev, 0x004020, 0xc0771100, info->mpll_ctrl);
-		nv_wr32(dev, 0x004024, info->mpll_coef);
-		break;
-	}
-	udelay(100);
-	nv_mask(dev, 0x00c040, 0x0000c000, 0x0000c000);
-
-	/* re-enable normal operation of memory controller */
-	nv_wr32(dev, 0x1002dc, 0x00000000);
-	nv_mask(dev, 0x100210, 0x80000000, 0x80000000);
-	udelay(100);
-
-	/* execute memory reset script from vbios */
-	if (!bit_table(dev, 'M', &M))
-		nouveau_bios_init_exec(dev, ROM16(M.data[0]));
-
-	/* make sure we're in vblank (hopefully the same one as before), and
-	 * then re-enable crtc memory access
-	 */
-	for (i = 0; i < 2; i++) {
-		if (!(crtc_mask & (1 << i)))
-			continue;
-		nv_wait(dev, 0x600808 + (i * 0x2000), 0x00010000, 0x00010000);
-		nv_wr08(dev, 0x0c03c4 + (i * 0x2000), 0x01);
-		nv_wr08(dev, 0x0c03c5 + (i * 0x2000), sr1[i]);
-	}
-
-	/* resume engines */
-resume:
-	nv_wr32(dev, 0x003250, 0x00000001);
-	nv_mask(dev, 0x003220, 0x00000001, 0x00000001);
-	nv_wr32(dev, 0x003200, 0x00000001);
-	nv_wr32(dev, 0x002500, 0x00000001);
-	spin_unlock_irqrestore(&dev_priv->context_switch_lock, flags);
-
-	kfree(info);
-	return ret;
-}
-
-int
-nv40_pm_pwm_get(struct drm_device *dev, int line, u32 *divs, u32 *duty)
-{
-	if (line == 2) {
-		u32 reg = nv_rd32(dev, 0x0010f0);
-		if (reg & 0x80000000) {
-			*duty = (reg & 0x7fff0000) >> 16;
-			*divs = (reg & 0x00007fff);
-			return 0;
-		}
-	} else
-	if (line == 9) {
-		u32 reg = nv_rd32(dev, 0x0015f4);
-		if (reg & 0x80000000) {
-			*divs = nv_rd32(dev, 0x0015f8);
-			*duty = (reg & 0x7fffffff);
-			return 0;
-		}
-	} else {
-		NV_ERROR(dev, "unknown pwm ctrl for gpio %d\n", line);
-		return -ENODEV;
-	}
-
-	return -EINVAL;
-}
-
-int
-nv40_pm_pwm_set(struct drm_device *dev, int line, u32 divs, u32 duty)
-{
-	if (line == 2) {
-		nv_wr32(dev, 0x0010f0, 0x80000000 | (duty << 16) | divs);
-	} else
-	if (line == 9) {
-		nv_wr32(dev, 0x0015f8, divs);
-		nv_wr32(dev, 0x0015f4, duty | 0x80000000);
-	} else {
-		NV_ERROR(dev, "unknown pwm ctrl for gpio %d\n", line);
-		return -ENODEV;
-	}
-
-	return 0;
-}
diff --git a/drivers/gpu/drm/nouveau/nv50_crtc.c b/drivers/gpu/drm/nouveau/nv50_crtc.c
index 701b927..ebabacf 100644
--- a/drivers/gpu/drm/nouveau/nv50_crtc.c
+++ b/drivers/gpu/drm/nouveau/nv50_crtc.c
@@ -104,7 +104,7 @@ nv50_crtc_blank(struct nouveau_crtc *nv_crtc, bool blanked)
 		OUT_RING(evo, nv_crtc->lut.depth == 8 ?
 				NV50_EVO_CRTC_CLUT_MODE_OFF :
 				NV50_EVO_CRTC_CLUT_MODE_ON);
-		OUT_RING(evo, nv_crtc->lut.nvbo->bo.offset >> 8);
+		OUT_RING(evo, (nv_crtc->lut.nvbo->bo.mem.start << PAGE_SHIFT) >> 8);
 		if (dev_priv->chipset != 0x50) {
 			BEGIN_RING(evo, 0, NV84_EVO_CRTC(index, CLUT_DMA), 1);
 			OUT_RING(evo, NvEvoVRAM);
@@ -132,74 +132,30 @@ nv50_crtc_blank(struct nouveau_crtc *nv_crtc, bool blanked)
 }
 
 static int
-nv50_crtc_set_dither(struct nouveau_crtc *nv_crtc, bool update)
-{
-	struct nouveau_channel *evo = nv50_display(nv_crtc->base.dev)->master;
-	struct nouveau_connector *nv_connector;
-	struct drm_connector *connector;
-	int head = nv_crtc->index, ret;
-	u32 mode = 0x00;
-
-	nv_connector = nouveau_crtc_connector_get(nv_crtc);
-	connector = &nv_connector->base;
-	if (nv_connector->dithering_mode == DITHERING_MODE_AUTO) {
-		if (nv_crtc->base.fb->depth > connector->display_info.bpc * 3)
-			mode = DITHERING_MODE_DYNAMIC2X2;
-	} else {
-		mode = nv_connector->dithering_mode;
-	}
-
-	if (nv_connector->dithering_depth == DITHERING_DEPTH_AUTO) {
-		if (connector->display_info.bpc >= 8)
-			mode |= DITHERING_DEPTH_8BPC;
-	} else {
-		mode |= nv_connector->dithering_depth;
-	}
-
-	ret = RING_SPACE(evo, 2 + (update ? 2 : 0));
-	if (ret == 0) {
-		BEGIN_RING(evo, 0, NV50_EVO_CRTC(head, DITHER_CTRL), 1);
-		OUT_RING  (evo, mode);
-		if (update) {
-			BEGIN_RING(evo, 0, NV50_EVO_UPDATE, 1);
-			OUT_RING  (evo, 0);
-			FIRE_RING (evo);
-		}
-	}
-
-	return ret;
-}
-
-static int
-nv50_crtc_set_color_vibrance(struct nouveau_crtc *nv_crtc, bool update)
+nv50_crtc_set_dither(struct nouveau_crtc *nv_crtc, bool on, bool update)
 {
 	struct drm_device *dev = nv_crtc->base.dev;
 	struct nouveau_channel *evo = nv50_display(dev)->master;
 	int ret;
-	int adj;
-	u32 hue, vib;
 
-	NV_DEBUG_KMS(dev, "vibrance = %i, hue = %i\n",
-		     nv_crtc->color_vibrance, nv_crtc->vibrant_hue);
+	NV_DEBUG_KMS(dev, "\n");
 
 	ret = RING_SPACE(evo, 2 + (update ? 2 : 0));
 	if (ret) {
-		NV_ERROR(dev, "no space while setting color vibrance\n");
+		NV_ERROR(dev, "no space while setting dither\n");
 		return ret;
 	}
 
-	adj = (nv_crtc->color_vibrance > 0) ? 50 : 0;
-	vib = ((nv_crtc->color_vibrance * 2047 + adj) / 100) & 0xfff;
-
-	hue = ((nv_crtc->vibrant_hue * 2047) / 100) & 0xfff;
-
-	BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, COLOR_CTRL), 1);
-	OUT_RING  (evo, (hue << 20) | (vib << 8));
+	BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, DITHER_CTRL), 1);
+	if (on)
+		OUT_RING(evo, NV50_EVO_CRTC_DITHER_CTRL_ON);
+	else
+		OUT_RING(evo, NV50_EVO_CRTC_DITHER_CTRL_OFF);
 
 	if (update) {
 		BEGIN_RING(evo, 0, NV50_EVO_UPDATE, 1);
-		OUT_RING  (evo, 0);
-		FIRE_RING (evo);
+		OUT_RING(evo, 0);
+		FIRE_RING(evo);
 	}
 
 	return 0;
@@ -224,103 +180,80 @@ nouveau_crtc_connector_get(struct nouveau_crtc *nv_crtc)
 }
 
 static int
-nv50_crtc_set_scale(struct nouveau_crtc *nv_crtc, bool update)
+nv50_crtc_set_scale(struct nouveau_crtc *nv_crtc, int scaling_mode, bool update)
 {
-	struct nouveau_connector *nv_connector;
-	struct drm_crtc *crtc = &nv_crtc->base;
-	struct drm_device *dev = crtc->dev;
+	struct nouveau_connector *nv_connector =
+		nouveau_crtc_connector_get(nv_crtc);
+	struct drm_device *dev = nv_crtc->base.dev;
 	struct nouveau_channel *evo = nv50_display(dev)->master;
-	struct drm_display_mode *umode = &crtc->mode;
-	struct drm_display_mode *omode;
-	int scaling_mode, ret;
-	u32 ctrl = 0, oX, oY;
+	struct drm_display_mode *native_mode = NULL;
+	struct drm_display_mode *mode = &nv_crtc->base.mode;
+	uint32_t outX, outY, horiz, vert;
+	int ret;
 
 	NV_DEBUG_KMS(dev, "\n");
 
-	nv_connector = nouveau_crtc_connector_get(nv_crtc);
-	if (!nv_connector || !nv_connector->native_mode) {
-		NV_ERROR(dev, "no native mode, forcing panel scaling\n");
-		scaling_mode = DRM_MODE_SCALE_NONE;
-	} else {
-		scaling_mode = nv_connector->scaling_mode;
-	}
-
-	/* start off at the resolution we programmed the crtc for, this
-	 * effectively handles NONE/FULL scaling
-	 */
-	if (scaling_mode != DRM_MODE_SCALE_NONE)
-		omode = nv_connector->native_mode;
-	else
-		omode = umode;
-
-	oX = omode->hdisplay;
-	oY = omode->vdisplay;
-	if (omode->flags & DRM_MODE_FLAG_DBLSCAN)
-		oY *= 2;
-
-	/* add overscan compensation if necessary, will keep the aspect
-	 * ratio the same as the backend mode unless overridden by the
-	 * user setting both hborder and vborder properties.
-	 */
-	if (nv_connector && ( nv_connector->underscan == UNDERSCAN_ON ||
-			     (nv_connector->underscan == UNDERSCAN_AUTO &&
-			      nv_connector->edid &&
-			      drm_detect_hdmi_monitor(nv_connector->edid)))) {
-		u32 bX = nv_connector->underscan_hborder;
-		u32 bY = nv_connector->underscan_vborder;
-		u32 aspect = (oY << 19) / oX;
-
-		if (bX) {
-			oX -= (bX * 2);
-			if (bY) oY -= (bY * 2);
-			else    oY  = ((oX * aspect) + (aspect / 2)) >> 19;
+	switch (scaling_mode) {
+	case DRM_MODE_SCALE_NONE:
+		break;
+	default:
+		if (!nv_connector || !nv_connector->native_mode) {
+			NV_ERROR(dev, "No native mode, forcing panel scaling\n");
+			scaling_mode = DRM_MODE_SCALE_NONE;
 		} else {
-			oX -= (oX >> 4) + 32;
-			if (bY) oY -= (bY * 2);
-			else    oY  = ((oX * aspect) + (aspect / 2)) >> 19;
+			native_mode = nv_connector->native_mode;
 		}
+		break;
 	}
 
-	/* handle CENTER/ASPECT scaling, taking into account the areas
-	 * removed already for overscan compensation
-	 */
 	switch (scaling_mode) {
-	case DRM_MODE_SCALE_CENTER:
-		oX = min((u32)umode->hdisplay, oX);
-		oY = min((u32)umode->vdisplay, oY);
-		/* fall-through */
 	case DRM_MODE_SCALE_ASPECT:
-		if (oY < oX) {
-			u32 aspect = (umode->hdisplay << 19) / umode->vdisplay;
-			oX = ((oY * aspect) + (aspect / 2)) >> 19;
+		horiz = (native_mode->hdisplay << 19) / mode->hdisplay;
+		vert = (native_mode->vdisplay << 19) / mode->vdisplay;
+
+		if (vert > horiz) {
+			outX = (mode->hdisplay * horiz) >> 19;
+			outY = (mode->vdisplay * horiz) >> 19;
 		} else {
-			u32 aspect = (umode->vdisplay << 19) / umode->hdisplay;
-			oY = ((oX * aspect) + (aspect / 2)) >> 19;
+			outX = (mode->hdisplay * vert) >> 19;
+			outY = (mode->vdisplay * vert) >> 19;
 		}
 		break;
+	case DRM_MODE_SCALE_FULLSCREEN:
+		outX = native_mode->hdisplay;
+		outY = native_mode->vdisplay;
+		break;
+	case DRM_MODE_SCALE_CENTER:
+	case DRM_MODE_SCALE_NONE:
 	default:
+		outX = mode->hdisplay;
+		outY = mode->vdisplay;
 		break;
 	}
 
-	if (umode->hdisplay != oX || umode->vdisplay != oY ||
-	    umode->flags & DRM_MODE_FLAG_INTERLACE ||
-	    umode->flags & DRM_MODE_FLAG_DBLSCAN)
-		ctrl |= NV50_EVO_CRTC_SCALE_CTRL_ACTIVE;
-
-	ret = RING_SPACE(evo, 5);
+	ret = RING_SPACE(evo, update ? 7 : 5);
 	if (ret)
 		return ret;
 
+	/* Got a better name for SCALER_ACTIVE? */
+	/* One day i've got to really figure out why this is needed. */
 	BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, SCALE_CTRL), 1);
-	OUT_RING  (evo, ctrl);
+	if ((mode->flags & DRM_MODE_FLAG_DBLSCAN) ||
+	    (mode->flags & DRM_MODE_FLAG_INTERLACE) ||
+	    mode->hdisplay != outX || mode->vdisplay != outY) {
+		OUT_RING(evo, NV50_EVO_CRTC_SCALE_CTRL_ACTIVE);
+	} else {
+		OUT_RING(evo, NV50_EVO_CRTC_SCALE_CTRL_INACTIVE);
+	}
+
 	BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, SCALE_RES1), 2);
-	OUT_RING  (evo, oY << 16 | oX);
-	OUT_RING  (evo, oY << 16 | oX);
+	OUT_RING(evo, outY << 16 | outX);
+	OUT_RING(evo, outY << 16 | outX);
 
 	if (update) {
-		nv50_display_flip_stop(crtc);
-		nv50_display_sync(dev);
-		nv50_display_flip_next(crtc, crtc->fb, NULL);
+		BEGIN_RING(evo, 0, NV50_EVO_UPDATE, 1);
+		OUT_RING(evo, 0);
+		FIRE_RING(evo);
 	}
 
 	return 0;
@@ -396,10 +329,13 @@ nv50_crtc_destroy(struct drm_crtc *crtc)
 
 	drm_crtc_cleanup(&nv_crtc->base);
 
+	nv50_cursor_fini(nv_crtc);
+
 	nouveau_bo_unmap(nv_crtc->lut.nvbo);
 	nouveau_bo_ref(NULL, &nv_crtc->lut.nvbo);
 	nouveau_bo_unmap(nv_crtc->cursor.nvbo);
 	nouveau_bo_ref(NULL, &nv_crtc->cursor.nvbo);
+	kfree(nv_crtc->mode);
 	kfree(nv_crtc);
 }
 
@@ -436,7 +372,7 @@ nv50_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv,
 
 	nouveau_bo_unmap(cursor);
 
-	nv_crtc->cursor.set_offset(nv_crtc, nv_crtc->cursor.nvbo->bo.offset);
+	nv_crtc->cursor.set_offset(nv_crtc, nv_crtc->cursor.nvbo->bo.mem.start << PAGE_SHIFT);
 	nv_crtc->cursor.show(nv_crtc, true);
 
 out:
@@ -507,6 +443,39 @@ nv50_crtc_dpms(struct drm_crtc *crtc, int mode)
 {
 }
 
+static int
+nv50_crtc_wait_complete(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_timer_engine *ptimer = &dev_priv->engine.timer;
+	struct nv50_display *disp = nv50_display(dev);
+	struct nouveau_channel *evo = disp->master;
+	u64 start;
+	int ret;
+
+	ret = RING_SPACE(evo, 6);
+	if (ret)
+		return ret;
+	BEGIN_RING(evo, 0, 0x0084, 1);
+	OUT_RING  (evo, 0x80000000);
+	BEGIN_RING(evo, 0, 0x0080, 1);
+	OUT_RING  (evo, 0);
+	BEGIN_RING(evo, 0, 0x0084, 1);
+	OUT_RING  (evo, 0x00000000);
+
+	nv_wo32(disp->ntfy, 0x000, 0x00000000);
+	FIRE_RING (evo);
+
+	start = ptimer->read(dev);
+	do {
+		if (nv_ro32(disp->ntfy, 0x000))
+			return 0;
+	} while (ptimer->read(dev) - start < 2000000000ULL);
+
+	return -EBUSY;
+}
+
 static void
 nv50_crtc_prepare(struct drm_crtc *crtc)
 {
@@ -530,7 +499,7 @@ nv50_crtc_commit(struct drm_crtc *crtc)
 
 	nv50_crtc_blank(nv_crtc, false);
 	drm_vblank_post_modeset(dev, nv_crtc->index);
-	nv50_display_sync(dev);
+	nv50_crtc_wait_complete(crtc);
 	nv50_display_flip_next(crtc, crtc->fb, NULL);
 }
 
@@ -550,18 +519,12 @@ nv50_crtc_do_mode_set_base(struct drm_crtc *crtc,
 	struct drm_device *dev = nv_crtc->base.dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_channel *evo = nv50_display(dev)->master;
-	struct drm_framebuffer *drm_fb;
-	struct nouveau_framebuffer *fb;
+	struct drm_framebuffer *drm_fb = nv_crtc->base.fb;
+	struct nouveau_framebuffer *fb = nouveau_framebuffer(drm_fb);
 	int ret;
 
 	NV_DEBUG_KMS(dev, "index %d\n", nv_crtc->index);
 
-	/* no fb bound */
-	if (!atomic && !crtc->fb) {
-		NV_DEBUG_KMS(dev, "No FB bound\n");
-		return 0;
-	}
-
 	/* If atomic, we want to switch to the fb we were passed, so
 	 * now we update pointers to do that.  (We don't pin; just
 	 * assume we're already pinned and update the base address.)
@@ -570,8 +533,6 @@ nv50_crtc_do_mode_set_base(struct drm_crtc *crtc,
 		drm_fb = passed_fb;
 		fb = nouveau_framebuffer(passed_fb);
 	} else {
-		drm_fb = crtc->fb;
-		fb = nouveau_framebuffer(crtc->fb);
 		/* If not atomic, we can go ahead and pin, and unpin the
 		 * old fb we were passed.
 		 */
@@ -585,7 +546,7 @@ nv50_crtc_do_mode_set_base(struct drm_crtc *crtc,
 		}
 	}
 
-	nv_crtc->fb.offset = fb->nvbo->bo.offset;
+	nv_crtc->fb.offset = fb->nvbo->bo.mem.start << PAGE_SHIFT;
 	nv_crtc->fb.tile_flags = nouveau_bo_tile_layout(fb->nvbo);
 	nv_crtc->fb.cpp = drm_fb->bits_per_pixel / 8;
 	if (!nv_crtc->fb.blanked && dev_priv->chipset != 0x50) {
@@ -612,6 +573,8 @@ nv50_crtc_do_mode_set_base(struct drm_crtc *crtc,
 	OUT_RING  (evo, fb->base.depth == 8 ?
 		   NV50_EVO_CRTC_CLUT_MODE_OFF : NV50_EVO_CRTC_CLUT_MODE_ON);
 
+	BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, COLOR_CTRL), 1);
+	OUT_RING  (evo, NV50_EVO_CRTC_COLOR_CTRL_COLOR);
 	BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, FB_POS), 1);
 	OUT_RING  (evo, (y << 16) | x);
 
@@ -624,77 +587,90 @@ nv50_crtc_do_mode_set_base(struct drm_crtc *crtc,
 }
 
 static int
-nv50_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *umode,
-		   struct drm_display_mode *mode, int x, int y,
+nv50_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *mode,
+		   struct drm_display_mode *adjusted_mode, int x, int y,
 		   struct drm_framebuffer *old_fb)
 {
 	struct drm_device *dev = crtc->dev;
 	struct nouveau_channel *evo = nv50_display(dev)->master;
 	struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
-	u32 head = nv_crtc->index * 0x400;
-	u32 ilace = (mode->flags & DRM_MODE_FLAG_INTERLACE) ? 2 : 1;
-	u32 vscan = (mode->flags & DRM_MODE_FLAG_DBLSCAN) ? 2 : 1;
-	u32 hactive, hsynce, hbackp, hfrontp, hblanke, hblanks;
-	u32 vactive, vsynce, vbackp, vfrontp, vblanke, vblanks;
-	u32 vblan2e = 0, vblan2s = 1;
+	struct nouveau_connector *nv_connector = NULL;
+	uint32_t hsync_dur,  vsync_dur, hsync_start_to_end, vsync_start_to_end;
+	uint32_t hunk1, vunk1, vunk2a, vunk2b;
 	int ret;
 
-	/* hw timing description looks like this:
-	 *
-	 * <sync> <back porch> <---------display---------> <front porch>
-	 * ______
-	 *       |____________|---------------------------|____________|
-	 *
-	 *       ^ synce      ^ blanke                    ^ blanks     ^ active
-	 *
-	 * interlaced modes also have 2 additional values pointing at the end
-	 * and start of the next field's blanking period.
-	 */
+	/* Find the connector attached to this CRTC */
+	nv_connector = nouveau_crtc_connector_get(nv_crtc);
+
+	*nv_crtc->mode = *adjusted_mode;
 
-	hactive = mode->htotal;
-	hsynce  = mode->hsync_end - mode->hsync_start - 1;
-	hbackp  = mode->htotal - mode->hsync_end;
-	hblanke = hsynce + hbackp;
-	hfrontp = mode->hsync_start - mode->hdisplay;
-	hblanks = mode->htotal - hfrontp - 1;
-
-	vactive = mode->vtotal * vscan / ilace;
-	vsynce  = ((mode->vsync_end - mode->vsync_start) * vscan / ilace) - 1;
-	vbackp  = (mode->vtotal - mode->vsync_end) * vscan / ilace;
-	vblanke = vsynce + vbackp;
-	vfrontp = (mode->vsync_start - mode->vdisplay) * vscan / ilace;
-	vblanks = vactive - vfrontp - 1;
-	if (mode->flags & DRM_MODE_FLAG_INTERLACE) {
-		vblan2e = vactive + vsynce + vbackp;
-		vblan2s = vblan2e + (mode->vdisplay * vscan / ilace);
-		vactive = (vactive * 2) + 1;
+	NV_DEBUG_KMS(dev, "index %d\n", nv_crtc->index);
+
+	hsync_dur = adjusted_mode->hsync_end - adjusted_mode->hsync_start;
+	vsync_dur = adjusted_mode->vsync_end - adjusted_mode->vsync_start;
+	hsync_start_to_end = adjusted_mode->htotal - adjusted_mode->hsync_start;
+	vsync_start_to_end = adjusted_mode->vtotal - adjusted_mode->vsync_start;
+	/* I can't give this a proper name, anyone else can? */
+	hunk1 = adjusted_mode->htotal -
+		adjusted_mode->hsync_start + adjusted_mode->hdisplay;
+	vunk1 = adjusted_mode->vtotal -
+		adjusted_mode->vsync_start + adjusted_mode->vdisplay;
+	/* Another strange value, this time only for interlaced adjusted_modes. */
+	vunk2a = 2 * adjusted_mode->vtotal -
+		 adjusted_mode->vsync_start + adjusted_mode->vdisplay;
+	vunk2b = adjusted_mode->vtotal -
+		 adjusted_mode->vsync_start + adjusted_mode->vtotal;
+
+	if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) {
+		vsync_dur /= 2;
+		vsync_start_to_end  /= 2;
+		vunk1 /= 2;
+		vunk2a /= 2;
+		vunk2b /= 2;
+		/* magic */
+		if (adjusted_mode->flags & DRM_MODE_FLAG_DBLSCAN) {
+			vsync_start_to_end -= 1;
+			vunk1 -= 1;
+			vunk2a -= 1;
+			vunk2b -= 1;
+		}
 	}
 
-	ret = RING_SPACE(evo, 18);
-	if (ret == 0) {
-		BEGIN_RING(evo, 0, 0x0804 + head, 2);
-		OUT_RING  (evo, 0x00800000 | mode->clock);
-		OUT_RING  (evo, (ilace == 2) ? 2 : 0);
-		BEGIN_RING(evo, 0, 0x0810 + head, 6);
-		OUT_RING  (evo, 0x00000000); /* border colour */
-		OUT_RING  (evo, (vactive << 16) | hactive);
-		OUT_RING  (evo, ( vsynce << 16) | hsynce);
-		OUT_RING  (evo, (vblanke << 16) | hblanke);
-		OUT_RING  (evo, (vblanks << 16) | hblanks);
-		OUT_RING  (evo, (vblan2e << 16) | vblan2s);
-		BEGIN_RING(evo, 0, 0x082c + head, 1);
-		OUT_RING  (evo, 0x00000000);
-		BEGIN_RING(evo, 0, 0x0900 + head, 1);
-		OUT_RING  (evo, 0x00000311); /* makes sync channel work */
-		BEGIN_RING(evo, 0, 0x08c8 + head, 1);
-		OUT_RING  (evo, (umode->vdisplay << 16) | umode->hdisplay);
-		BEGIN_RING(evo, 0, 0x08d4 + head, 1);
-		OUT_RING  (evo, 0x00000000); /* screen position */
+	ret = RING_SPACE(evo, 17);
+	if (ret)
+		return ret;
+
+	BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, CLOCK), 2);
+	OUT_RING(evo, adjusted_mode->clock | 0x800000);
+	OUT_RING(evo, (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) ? 2 : 0);
+
+	BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, DISPLAY_START), 5);
+	OUT_RING(evo, 0);
+	OUT_RING(evo, (adjusted_mode->vtotal << 16) | adjusted_mode->htotal);
+	OUT_RING(evo, (vsync_dur - 1) << 16 | (hsync_dur - 1));
+	OUT_RING(evo, (vsync_start_to_end - 1) << 16 |
+			(hsync_start_to_end - 1));
+	OUT_RING(evo, (vunk1 - 1) << 16 | (hunk1 - 1));
+
+	if (adjusted_mode->flags & DRM_MODE_FLAG_INTERLACE) {
+		BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, UNK0824), 1);
+		OUT_RING(evo, (vunk2b - 1) << 16 | (vunk2a - 1));
+	} else {
+		OUT_RING(evo, 0);
+		OUT_RING(evo, 0);
 	}
 
-	nv_crtc->set_dither(nv_crtc, false);
-	nv_crtc->set_scale(nv_crtc, false);
-	nv_crtc->set_color_vibrance(nv_crtc, false);
+	BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, UNK082C), 1);
+	OUT_RING(evo, 0);
+
+	/* This is the actual resolution of the mode. */
+	BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, REAL_RES), 1);
+	OUT_RING(evo, (mode->vdisplay << 16) | mode->hdisplay);
+	BEGIN_RING(evo, 0, NV50_EVO_CRTC(nv_crtc->index, SCALE_CENTER_OFFSET), 1);
+	OUT_RING(evo, NV50_EVO_CRTC_SCALE_CENTER_OFFSET_VAL(0, 0));
+
+	nv_crtc->set_dither(nv_crtc, nv_connector->use_dithering, false);
+	nv_crtc->set_scale(nv_crtc, nv_connector->scaling_mode, false);
 
 	return nv50_crtc_do_mode_set_base(crtc, old_fb, x, y, false);
 }
@@ -710,7 +686,7 @@ nv50_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
 	if (ret)
 		return ret;
 
-	ret = nv50_display_sync(crtc->dev);
+	ret = nv50_crtc_wait_complete(crtc);
 	if (ret)
 		return ret;
 
@@ -729,7 +705,7 @@ nv50_crtc_mode_set_base_atomic(struct drm_crtc *crtc,
 	if (ret)
 		return ret;
 
-	return nv50_display_sync(crtc->dev);
+	return nv50_crtc_wait_complete(crtc);
 }
 
 static const struct drm_crtc_helper_funcs nv50_crtc_helper_funcs = {
@@ -755,8 +731,11 @@ nv50_crtc_create(struct drm_device *dev, int index)
 	if (!nv_crtc)
 		return -ENOMEM;
 
-	nv_crtc->color_vibrance = 50;
-	nv_crtc->vibrant_hue = 0;
+	nv_crtc->mode = kzalloc(sizeof(*nv_crtc->mode), GFP_KERNEL);
+	if (!nv_crtc->mode) {
+		kfree(nv_crtc);
+		return -ENOMEM;
+	}
 
 	/* Default CLUT parameters, will be activated on the hw upon
 	 * first mode set.
@@ -768,7 +747,7 @@ nv50_crtc_create(struct drm_device *dev, int index)
 	}
 	nv_crtc->lut.depth = 0;
 
-	ret = nouveau_bo_new(dev, 4096, 0x100, TTM_PL_FLAG_VRAM,
+	ret = nouveau_bo_new(dev, NULL, 4096, 0x100, TTM_PL_FLAG_VRAM,
 			     0, 0x0000, &nv_crtc->lut.nvbo);
 	if (!ret) {
 		ret = nouveau_bo_pin(nv_crtc->lut.nvbo, TTM_PL_FLAG_VRAM);
@@ -779,6 +758,7 @@ nv50_crtc_create(struct drm_device *dev, int index)
 	}
 
 	if (ret) {
+		kfree(nv_crtc->mode);
 		kfree(nv_crtc);
 		return ret;
 	}
@@ -788,13 +768,12 @@ nv50_crtc_create(struct drm_device *dev, int index)
 	/* set function pointers */
 	nv_crtc->set_dither = nv50_crtc_set_dither;
 	nv_crtc->set_scale = nv50_crtc_set_scale;
-	nv_crtc->set_color_vibrance = nv50_crtc_set_color_vibrance;
 
 	drm_crtc_init(dev, &nv_crtc->base, &nv50_crtc_funcs);
 	drm_crtc_helper_add(&nv_crtc->base, &nv50_crtc_helper_funcs);
 	drm_mode_crtc_set_gamma_size(&nv_crtc->base, 256);
 
-	ret = nouveau_bo_new(dev, 64*64*4, 0x100, TTM_PL_FLAG_VRAM,
+	ret = nouveau_bo_new(dev, NULL, 64*64*4, 0x100, TTM_PL_FLAG_VRAM,
 			     0, 0x0000, &nv_crtc->cursor.nvbo);
 	if (!ret) {
 		ret = nouveau_bo_pin(nv_crtc->cursor.nvbo, TTM_PL_FLAG_VRAM);
diff --git a/drivers/gpu/drm/nouveau/nv50_cursor.c b/drivers/gpu/drm/nouveau/nv50_cursor.c
index adfc9b6..9752c35 100644
--- a/drivers/gpu/drm/nouveau/nv50_cursor.c
+++ b/drivers/gpu/drm/nouveau/nv50_cursor.c
@@ -137,3 +137,21 @@ nv50_cursor_init(struct nouveau_crtc *nv_crtc)
 	nv_crtc->cursor.show = nv50_cursor_show;
 	return 0;
 }
+
+void
+nv50_cursor_fini(struct nouveau_crtc *nv_crtc)
+{
+	struct drm_device *dev = nv_crtc->base.dev;
+	int idx = nv_crtc->index;
+
+	NV_DEBUG_KMS(dev, "\n");
+
+	nv_wr32(dev, NV50_PDISPLAY_CURSOR_CURSOR_CTRL2(idx), 0);
+	if (!nv_wait(dev, NV50_PDISPLAY_CURSOR_CURSOR_CTRL2(idx),
+		     NV50_PDISPLAY_CURSOR_CURSOR_CTRL2_STATUS, 0)) {
+		NV_ERROR(dev, "timeout: CURSOR_CTRL2_STATUS == 0\n");
+		NV_ERROR(dev, "CURSOR_CTRL2 = 0x%08x\n",
+			 nv_rd32(dev, NV50_PDISPLAY_CURSOR_CURSOR_CTRL2(idx)));
+	}
+}
+
diff --git a/drivers/gpu/drm/nouveau/nv50_dac.c b/drivers/gpu/drm/nouveau/nv50_dac.c
index 55c5633..808f3ec 100644
--- a/drivers/gpu/drm/nouveau/nv50_dac.c
+++ b/drivers/gpu/drm/nouveau/nv50_dac.c
@@ -190,13 +190,21 @@ nv50_dac_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *mode,
 	}
 
 	if (connector->scaling_mode != DRM_MODE_SCALE_NONE &&
-	     connector->native_mode)
-		drm_mode_copy(adjusted_mode, connector->native_mode);
+	     connector->native_mode) {
+		int id = adjusted_mode->base.id;
+		*adjusted_mode = *connector->native_mode;
+		adjusted_mode->base.id = id;
+	}
 
 	return true;
 }
 
 static void
+nv50_dac_prepare(struct drm_encoder *encoder)
+{
+}
+
+static void
 nv50_dac_commit(struct drm_encoder *encoder)
 {
 }
@@ -258,7 +266,7 @@ static const struct drm_encoder_helper_funcs nv50_dac_helper_funcs = {
 	.save = nv50_dac_save,
 	.restore = nv50_dac_restore,
 	.mode_fixup = nv50_dac_mode_fixup,
-	.prepare = nv50_dac_disconnect,
+	.prepare = nv50_dac_prepare,
 	.commit = nv50_dac_commit,
 	.mode_set = nv50_dac_mode_set,
 	.get_crtc = nv50_dac_crtc_get,
diff --git a/drivers/gpu/drm/nouveau/nv50_display.c b/drivers/gpu/drm/nouveau/nv50_display.c
index 0e47a89..08da478 100644
--- a/drivers/gpu/drm/nouveau/nv50_display.c
+++ b/drivers/gpu/drm/nouveau/nv50_display.c
@@ -50,76 +50,9 @@ nv50_sor_nr(struct drm_device *dev)
 	return 4;
 }
 
-u32
-nv50_display_active_crtcs(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	u32 mask = 0;
-	int i;
-
-	if (dev_priv->chipset  < 0x90 ||
-	    dev_priv->chipset == 0x92 ||
-	    dev_priv->chipset == 0xa0) {
-		for (i = 0; i < 2; i++)
-			mask |= nv_rd32(dev, NV50_PDISPLAY_SOR_MODE_CTRL_C(i));
-	} else {
-		for (i = 0; i < 4; i++)
-			mask |= nv_rd32(dev, NV90_PDISPLAY_SOR_MODE_CTRL_C(i));
-	}
-
-	for (i = 0; i < 3; i++)
-		mask |= nv_rd32(dev, NV50_PDISPLAY_DAC_MODE_CTRL_C(i));
-
-	return mask & 3;
-}
-
-static int
-evo_icmd(struct drm_device *dev, int ch, u32 mthd, u32 data)
-{
-	int ret = 0;
-	nv_mask(dev, 0x610300 + (ch * 0x08), 0x00000001, 0x00000001);
-	nv_wr32(dev, 0x610304 + (ch * 0x08), data);
-	nv_wr32(dev, 0x610300 + (ch * 0x08), 0x80000001 | mthd);
-	if (!nv_wait(dev, 0x610300 + (ch * 0x08), 0x80000000, 0x00000000))
-		ret = -EBUSY;
-	if (ret || (nouveau_reg_debug & NOUVEAU_REG_DEBUG_EVO))
-		NV_INFO(dev, "EvoPIO: %d 0x%04x 0x%08x\n", ch, mthd, data);
-	nv_mask(dev, 0x610300 + (ch * 0x08), 0x00000001, 0x00000000);
-	return ret;
-}
-
 int
 nv50_display_early_init(struct drm_device *dev)
 {
-	u32 ctrl = nv_rd32(dev, 0x610200);
-	int i;
-
-	/* check if master evo channel is already active, a good a sign as any
-	 * that the display engine is in a weird state (hibernate/kexec), if
-	 * it is, do our best to reset the display engine...
-	 */
-	if ((ctrl & 0x00000003) == 0x00000003) {
-		NV_INFO(dev, "PDISP: EVO(0) 0x%08x, resetting...\n", ctrl);
-
-		/* deactivate both heads first, PDISP will disappear forever
-		 * (well, until you power cycle) on some boards as soon as
-		 * PMC_ENABLE is hit unless they are..
-		 */
-		for (i = 0; i < 2; i++) {
-			evo_icmd(dev, 0, 0x0880 + (i * 0x400), 0x05000000);
-			evo_icmd(dev, 0, 0x089c + (i * 0x400), 0);
-			evo_icmd(dev, 0, 0x0840 + (i * 0x400), 0);
-			evo_icmd(dev, 0, 0x0844 + (i * 0x400), 0);
-			evo_icmd(dev, 0, 0x085c + (i * 0x400), 0);
-			evo_icmd(dev, 0, 0x0874 + (i * 0x400), 0);
-		}
-		evo_icmd(dev, 0, 0x0080, 0);
-
-		/* reset PDISP */
-		nv_mask(dev, 0x000200, 0x40000000, 0x00000000);
-		nv_mask(dev, 0x000200, 0x40000000, 0x40000000);
-	}
-
 	return 0;
 }
 
@@ -129,40 +62,11 @@ nv50_display_late_takedown(struct drm_device *dev)
 }
 
 int
-nv50_display_sync(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_timer_engine *ptimer = &dev_priv->engine.timer;
-	struct nv50_display *disp = nv50_display(dev);
-	struct nouveau_channel *evo = disp->master;
-	u64 start;
-	int ret;
-
-	ret = RING_SPACE(evo, 6);
-	if (ret == 0) {
-		BEGIN_RING(evo, 0, 0x0084, 1);
-		OUT_RING  (evo, 0x80000000);
-		BEGIN_RING(evo, 0, 0x0080, 1);
-		OUT_RING  (evo, 0);
-		BEGIN_RING(evo, 0, 0x0084, 1);
-		OUT_RING  (evo, 0x00000000);
-
-		nv_wo32(disp->ntfy, 0x000, 0x00000000);
-		FIRE_RING (evo);
-
-		start = ptimer->read(dev);
-		do {
-			if (nv_ro32(disp->ntfy, 0x000))
-				return 0;
-		} while (ptimer->read(dev) - start < 2000000000ULL);
-	}
-
-	return -EBUSY;
-}
-
-int
 nv50_display_init(struct drm_device *dev)
 {
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio;
+	struct drm_connector *connector;
 	struct nouveau_channel *evo;
 	int ret, i;
 	u32 val;
@@ -257,6 +161,16 @@ nv50_display_init(struct drm_device *dev)
 		     NV50_PDISPLAY_INTR_EN_1_CLK_UNK20 |
 		     NV50_PDISPLAY_INTR_EN_1_CLK_UNK40);
 
+	/* enable hotplug interrupts */
+	list_for_each_entry(connector, &dev->mode_config.connector_list, head) {
+		struct nouveau_connector *conn = nouveau_connector(connector);
+
+		if (conn->dcb->gpio_tag == 0xff)
+			continue;
+
+		pgpio->irq_enable(dev, conn->dcb->gpio_tag, true);
+	}
+
 	ret = nv50_evo_init(dev);
 	if (ret)
 		return ret;
@@ -264,19 +178,36 @@ nv50_display_init(struct drm_device *dev)
 
 	nv_wr32(dev, NV50_PDISPLAY_OBJECTS, (evo->ramin->vinst >> 8) | 9);
 
-	ret = RING_SPACE(evo, 3);
+	ret = RING_SPACE(evo, 15);
 	if (ret)
 		return ret;
 	BEGIN_RING(evo, 0, NV50_EVO_UNK84, 2);
-	OUT_RING  (evo, NV50_EVO_UNK84_NOTIFY_DISABLED);
-	OUT_RING  (evo, NvEvoSync);
+	OUT_RING(evo, NV50_EVO_UNK84_NOTIFY_DISABLED);
+	OUT_RING(evo, NvEvoSync);
+	BEGIN_RING(evo, 0, NV50_EVO_CRTC(0, FB_DMA), 1);
+	OUT_RING(evo, NV50_EVO_CRTC_FB_DMA_HANDLE_NONE);
+	BEGIN_RING(evo, 0, NV50_EVO_CRTC(0, UNK0800), 1);
+	OUT_RING(evo, 0);
+	BEGIN_RING(evo, 0, NV50_EVO_CRTC(0, DISPLAY_START), 1);
+	OUT_RING(evo, 0);
+	BEGIN_RING(evo, 0, NV50_EVO_CRTC(0, UNK082C), 1);
+	OUT_RING(evo, 0);
+	/* required to make display sync channels not hate life */
+	BEGIN_RING(evo, 0, NV50_EVO_CRTC(0, UNK900), 1);
+	OUT_RING  (evo, 0x00000311);
+	BEGIN_RING(evo, 0, NV50_EVO_CRTC(1, UNK900), 1);
+	OUT_RING  (evo, 0x00000311);
+	FIRE_RING(evo);
+	if (!nv_wait(dev, 0x640004, 0xffffffff, evo->dma.put << 2))
+		NV_ERROR(dev, "evo pushbuf stalled\n");
 
-	return nv50_display_sync(dev);
+
+	return 0;
 }
 
-void
-nv50_display_fini(struct drm_device *dev)
+static int nv50_display_disable(struct drm_device *dev)
 {
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nv50_display *disp = nv50_display(dev);
 	struct nouveau_channel *evo = disp->master;
 	struct drm_crtc *drm_crtc;
@@ -316,16 +247,6 @@ nv50_display_fini(struct drm_device *dev)
 		}
 	}
 
-	for (i = 0; i < 2; i++) {
-		nv_wr32(dev, NV50_PDISPLAY_CURSOR_CURSOR_CTRL2(i), 0);
-		if (!nv_wait(dev, NV50_PDISPLAY_CURSOR_CURSOR_CTRL2(i),
-			     NV50_PDISPLAY_CURSOR_CURSOR_CTRL2_STATUS, 0)) {
-			NV_ERROR(dev, "timeout: CURSOR_CTRL2_STATUS == 0\n");
-			NV_ERROR(dev, "CURSOR_CTRL2 = 0x%08x\n",
-				 nv_rd32(dev, NV50_PDISPLAY_CURSOR_CURSOR_CTRL2(i)));
-		}
-	}
-
 	nv50_evo_fini(dev);
 
 	for (i = 0; i < 3; i++) {
@@ -339,10 +260,18 @@ nv50_display_fini(struct drm_device *dev)
 
 	/* disable interrupts. */
 	nv_wr32(dev, NV50_PDISPLAY_INTR_EN_1, 0x00000000);
+
+	/* disable hotplug interrupts */
+	nv_wr32(dev, 0xe054, 0xffffffff);
+	nv_wr32(dev, 0xe050, 0x00000000);
+	if (dev_priv->chipset >= 0x90) {
+		nv_wr32(dev, 0xe074, 0xffffffff);
+		nv_wr32(dev, 0xe070, 0x00000000);
+	}
+	return 0;
 }
 
-int
-nv50_display_create(struct drm_device *dev)
+int nv50_display_create(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct dcb_table *dcb = &dev_priv->vbios.dcb;
@@ -357,6 +286,23 @@ nv50_display_create(struct drm_device *dev)
 		return -ENOMEM;
 	dev_priv->engine.display.priv = priv;
 
+	/* init basic kernel modesetting */
+	drm_mode_config_init(dev);
+
+	/* Initialise some optional connector properties. */
+	drm_mode_create_scaling_mode_property(dev);
+	drm_mode_create_dithering_property(dev);
+
+	dev->mode_config.min_width = 0;
+	dev->mode_config.min_height = 0;
+
+	dev->mode_config.funcs = (void *)&nouveau_mode_config_funcs;
+
+	dev->mode_config.max_width = 8192;
+	dev->mode_config.max_height = 8192;
+
+	dev->mode_config.fb_base = dev_priv->fb_phys;
+
 	/* Create CRTC objects */
 	for (i = 0; i < 2; i++)
 		nv50_crtc_create(dev, i);
@@ -402,7 +348,7 @@ nv50_display_create(struct drm_device *dev)
 	tasklet_init(&priv->tasklet, nv50_display_bh, (unsigned long)dev);
 	nouveau_irq_register(dev, 26, nv50_display_isr);
 
-	ret = nv50_evo_create(dev);
+	ret = nv50_display_init(dev);
 	if (ret) {
 		nv50_display_destroy(dev);
 		return ret;
@@ -418,7 +364,9 @@ nv50_display_destroy(struct drm_device *dev)
 
 	NV_DEBUG_KMS(dev, "\n");
 
-	nv50_evo_destroy(dev);
+	drm_mode_config_cleanup(dev);
+
+	nv50_display_disable(dev);
 	nouveau_irq_unregister(dev, 26);
 	kfree(disp);
 }
@@ -467,6 +415,8 @@ nv50_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 
 	/* synchronise with the rendering channel, if necessary */
 	if (likely(chan)) {
+		u64 offset = dispc->sem.bo->vma.offset + dispc->sem.offset;
+
 		ret = RING_SPACE(chan, 10);
 		if (ret) {
 			WIND_RING(evo);
@@ -488,8 +438,6 @@ nv50_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 			else
 				OUT_RING  (chan, chan->vram_handle);
 		} else {
-			u64 offset = chan->dispc_vma[nv_crtc->index].offset;
-			offset += dispc->sem.offset;
 			BEGIN_NVC0(chan, 2, NvSubM2MF, 0x0010, 4);
 			OUT_RING  (chan, upper_32_bits(offset));
 			OUT_RING  (chan, lower_32_bits(offset));
@@ -536,7 +484,7 @@ nv50_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb,
 	OUT_RING  (evo, 0x00000000);
 	OUT_RING  (evo, 0x00000000);
 	BEGIN_RING(evo, 0, 0x0800, 5);
-	OUT_RING  (evo, nv_fb->nvbo->bo.offset >> 8);
+	OUT_RING  (evo, (nv_fb->nvbo->bo.mem.start << PAGE_SHIFT) >> 8);
 	OUT_RING  (evo, 0);
 	OUT_RING  (evo, (fb->height << 16) | fb->width);
 	OUT_RING  (evo, nv_fb->r_pitch);
@@ -582,7 +530,7 @@ nv50_display_script_select(struct drm_device *dev, struct dcb_entry *dcb,
 		} else {
 			/* determine number of lvds links */
 			if (nv_connector && nv_connector->edid &&
-			    nv_connector->type == DCB_CONNECTOR_LVDS_SPWG) {
+			    nv_connector->dcb->type == DCB_CONNECTOR_LVDS_SPWG) {
 				/* http://www.spwg.org */
 				if (((u8 *)nv_connector->edid)[121] == 2)
 					script |= 0x0100;
@@ -677,7 +625,7 @@ nv50_display_unk10_handler(struct drm_device *dev)
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nv50_display *disp = nv50_display(dev);
 	u32 unk30 = nv_rd32(dev, 0x610030), mc;
-	int i, crtc, or = 0, type = OUTPUT_ANY;
+	int i, crtc, or, type = OUTPUT_ANY;
 
 	NV_DEBUG_KMS(dev, "0x610030: 0x%08x\n", unk30);
 	disp->irq.dcb = NULL;
@@ -750,7 +698,7 @@ nv50_display_unk10_handler(struct drm_device *dev)
 		struct dcb_entry *dcb = &dev_priv->vbios.dcb.entry[i];
 
 		if (dcb->type == type && (dcb->or & (1 << or))) {
-			nouveau_bios_run_display_table(dev, 0, -1, dcb, -1);
+			nouveau_bios_run_display_table(dev, dcb, 0, -1);
 			disp->irq.dcb = dcb;
 			goto ack;
 		}
@@ -763,18 +711,49 @@ ack:
 }
 
 static void
+nv50_display_unk20_dp_hack(struct drm_device *dev, struct dcb_entry *dcb)
+{
+	int or = ffs(dcb->or) - 1, link = !(dcb->dpconf.sor.link & 1);
+	struct drm_encoder *encoder;
+	uint32_t tmp, unk0 = 0, unk1 = 0;
+
+	if (dcb->type != OUTPUT_DP)
+		return;
+
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
+		struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
+
+		if (nv_encoder->dcb == dcb) {
+			unk0 = nv_encoder->dp.unk0;
+			unk1 = nv_encoder->dp.unk1;
+			break;
+		}
+	}
+
+	if (unk0 || unk1) {
+		tmp  = nv_rd32(dev, NV50_SOR_DP_CTRL(or, link));
+		tmp &= 0xfffffe03;
+		nv_wr32(dev, NV50_SOR_DP_CTRL(or, link), tmp | unk0);
+
+		tmp  = nv_rd32(dev, NV50_SOR_DP_UNK128(or, link));
+		tmp &= 0xfef080c0;
+		nv_wr32(dev, NV50_SOR_DP_UNK128(or, link), tmp | unk1);
+	}
+}
+
+static void
 nv50_display_unk20_handler(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nv50_display *disp = nv50_display(dev);
 	u32 unk30 = nv_rd32(dev, 0x610030), tmp, pclk, script, mc = 0;
 	struct dcb_entry *dcb;
-	int i, crtc, or = 0, type = OUTPUT_ANY;
+	int i, crtc, or, type = OUTPUT_ANY;
 
 	NV_DEBUG_KMS(dev, "0x610030: 0x%08x\n", unk30);
 	dcb = disp->irq.dcb;
 	if (dcb) {
-		nouveau_bios_run_display_table(dev, 0, -2, dcb, -1);
+		nouveau_bios_run_display_table(dev, dcb, 0, -2);
 		disp->irq.dcb = NULL;
 	}
 
@@ -783,8 +762,8 @@ nv50_display_unk20_handler(struct drm_device *dev)
 	if (crtc >= 0) {
 		pclk  = nv_rd32(dev, NV50_PDISPLAY_CRTC_P(crtc, CLOCK));
 		pclk &= 0x003fffff;
-		if (pclk)
-			nv50_crtc_set_clock(dev, crtc, pclk);
+
+		nv50_crtc_set_clock(dev, crtc, pclk);
 
 		tmp = nv_rd32(dev, NV50_PDISPLAY_CRTC_CLK_CTRL2(crtc));
 		tmp &= ~0x000000f;
@@ -858,15 +837,9 @@ nv50_display_unk20_handler(struct drm_device *dev)
 	}
 
 	script = nv50_display_script_select(dev, dcb, mc, pclk);
-	nouveau_bios_run_display_table(dev, script, pclk, dcb, -1);
+	nouveau_bios_run_display_table(dev, dcb, script, pclk);
 
-	if (type == OUTPUT_DP) {
-		int link = !(dcb->dpconf.sor.link & 1);
-		if ((mc & 0x000f0000) == 0x00020000)
-			nv50_sor_dp_calc_tu(dev, or, link, pclk, 18);
-		else
-			nv50_sor_dp_calc_tu(dev, or, link, pclk, 24);
-	}
+	nv50_display_unk20_dp_hack(dev, dcb);
 
 	if (dcb->type != OUTPUT_ANALOG) {
 		tmp = nv_rd32(dev, NV50_PDISPLAY_SOR_CLK_CTRL2(or));
@@ -931,7 +904,7 @@ nv50_display_unk40_handler(struct drm_device *dev)
 	if (!dcb)
 		goto ack;
 
-	nouveau_bios_run_display_table(dev, script, -pclk, dcb, -1);
+	nouveau_bios_run_display_table(dev, dcb, script, -pclk);
 	nv50_display_unk40_dp_set_tmds(dev, dcb);
 
 ack:
diff --git a/drivers/gpu/drm/nouveau/nv50_display.h b/drivers/gpu/drm/nouveau/nv50_display.h
index 5d3dd14..c2da503 100644
--- a/drivers/gpu/drm/nouveau/nv50_display.h
+++ b/drivers/gpu/drm/nouveau/nv50_display.h
@@ -69,20 +69,14 @@ int nv50_display_early_init(struct drm_device *dev);
 void nv50_display_late_takedown(struct drm_device *dev);
 int nv50_display_create(struct drm_device *dev);
 int nv50_display_init(struct drm_device *dev);
-void nv50_display_fini(struct drm_device *dev);
 void nv50_display_destroy(struct drm_device *dev);
 int nv50_crtc_blank(struct nouveau_crtc *, bool blank);
 int nv50_crtc_set_clock(struct drm_device *, int head, int pclk);
 
-u32  nv50_display_active_crtcs(struct drm_device *);
-
-int  nv50_display_sync(struct drm_device *);
 int  nv50_display_flip_next(struct drm_crtc *, struct drm_framebuffer *,
 			    struct nouveau_channel *chan);
 void nv50_display_flip_stop(struct drm_crtc *);
 
-int  nv50_evo_create(struct drm_device *dev);
-void nv50_evo_destroy(struct drm_device *dev);
 int  nv50_evo_init(struct drm_device *dev);
 void nv50_evo_fini(struct drm_device *dev);
 void nv50_evo_dmaobj_init(struct nouveau_gpuobj *, u32 memtype, u64 base,
diff --git a/drivers/gpu/drm/nouveau/nv50_evo.c b/drivers/gpu/drm/nouveau/nv50_evo.c
index 9b962e9..c8e83c1 100644
--- a/drivers/gpu/drm/nouveau/nv50_evo.c
+++ b/drivers/gpu/drm/nouveau/nv50_evo.c
@@ -38,7 +38,6 @@ nv50_evo_channel_del(struct nouveau_channel **pevo)
 		return;
 	*pevo = NULL;
 
-	nouveau_ramht_ref(NULL, &evo->ramht, evo);
 	nouveau_gpuobj_channel_takedown(evo);
 	nouveau_bo_unmap(evo->pushbuf_bo);
 	nouveau_bo_ref(NULL, &evo->pushbuf_bo);
@@ -117,7 +116,7 @@ nv50_evo_channel_new(struct drm_device *dev, int chid,
 	evo->user_get = 4;
 	evo->user_put = 0;
 
-	ret = nouveau_bo_new(dev, 4096, 0, TTM_PL_FLAG_VRAM, 0, 0,
+	ret = nouveau_bo_new(dev, NULL, 4096, 0, TTM_PL_FLAG_VRAM, 0, 0,
 			     &evo->pushbuf_bo);
 	if (ret == 0)
 		ret = nouveau_bo_pin(evo->pushbuf_bo, TTM_PL_FLAG_VRAM);
@@ -154,7 +153,7 @@ nv50_evo_channel_init(struct nouveau_channel *evo)
 {
 	struct drm_device *dev = evo->dev;
 	int id = evo->id, ret, i;
-	u64 pushbuf = evo->pushbuf_bo->bo.offset;
+	u64 pushbuf = evo->pushbuf_bo->bo.mem.start << PAGE_SHIFT;
 	u32 tmp;
 
 	tmp = nv_rd32(dev, NV50_PDISPLAY_EVO_CTRL(id));
@@ -218,7 +217,7 @@ nv50_evo_channel_fini(struct nouveau_channel *evo)
 	}
 }
 
-void
+static void
 nv50_evo_destroy(struct drm_device *dev)
 {
 	struct nv50_display *disp = nv50_display(dev);
@@ -235,7 +234,7 @@ nv50_evo_destroy(struct drm_device *dev)
 	nv50_evo_channel_del(&disp->master);
 }
 
-int
+static int
 nv50_evo_create(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
@@ -332,15 +331,16 @@ nv50_evo_create(struct drm_device *dev)
 		if (ret)
 			goto err;
 
-		ret = nouveau_bo_new(dev, 4096, 0x1000, TTM_PL_FLAG_VRAM,
+		ret = nouveau_bo_new(dev, NULL, 4096, 0x1000, TTM_PL_FLAG_VRAM,
 				     0, 0x0000, &dispc->sem.bo);
 		if (!ret) {
+			offset = dispc->sem.bo->bo.mem.start << PAGE_SHIFT;
+
 			ret = nouveau_bo_pin(dispc->sem.bo, TTM_PL_FLAG_VRAM);
 			if (!ret)
 				ret = nouveau_bo_map(dispc->sem.bo);
 			if (ret)
 				nouveau_bo_ref(NULL, &dispc->sem.bo);
-			offset = dispc->sem.bo->bo.offset;
 		}
 
 		if (ret)
@@ -388,6 +388,12 @@ nv50_evo_init(struct drm_device *dev)
 	struct nv50_display *disp = nv50_display(dev);
 	int ret, i;
 
+	if (!disp->master) {
+		ret = nv50_evo_create(dev);
+		if (ret)
+			return ret;
+	}
+
 	ret = nv50_evo_channel_init(disp->master);
 	if (ret)
 		return ret;
@@ -414,4 +420,6 @@ nv50_evo_fini(struct drm_device *dev)
 
 	if (disp->master)
 		nv50_evo_channel_fini(disp->master);
+
+	nv50_evo_destroy(dev);
 }
diff --git a/drivers/gpu/drm/nouveau/nv50_evo.h b/drivers/gpu/drm/nouveau/nv50_evo.h
index 771d879..3860ca6 100644
--- a/drivers/gpu/drm/nouveau/nv50_evo.h
+++ b/drivers/gpu/drm/nouveau/nv50_evo.h
@@ -104,8 +104,7 @@
 #define NV50_EVO_CRTC_SCALE_CTRL_INACTIVE                            0x00000000
 #define NV50_EVO_CRTC_SCALE_CTRL_ACTIVE                              0x00000009
 #define NV50_EVO_CRTC_COLOR_CTRL                                     0x000008a8
-#define NV50_EVO_CRTC_COLOR_CTRL_VIBRANCE                            0x000fff00
-#define NV50_EVO_CRTC_COLOR_CTRL_HUE                                 0xfff00000
+#define NV50_EVO_CRTC_COLOR_CTRL_COLOR                               0x00040000
 #define NV50_EVO_CRTC_FB_POS                                         0x000008c0
 #define NV50_EVO_CRTC_REAL_RES                                       0x000008c8
 #define NV50_EVO_CRTC_SCALE_CENTER_OFFSET                            0x000008d4
diff --git a/drivers/gpu/drm/nouveau/nv50_fbcon.c b/drivers/gpu/drm/nouveau/nv50_fbcon.c
index dc75a72..791ded1 100644
--- a/drivers/gpu/drm/nouveau/nv50_fbcon.c
+++ b/drivers/gpu/drm/nouveau/nv50_fbcon.c
@@ -159,7 +159,7 @@ nv50_fbcon_accel_init(struct fb_info *info)
 	struct drm_device *dev = nfbdev->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_channel *chan = dev_priv->channel;
-	struct nouveau_framebuffer *fb = &nfbdev->nouveau_fb;
+	struct nouveau_bo *nvbo = nfbdev->nouveau_fb.nvbo;
 	int ret, format;
 
 	switch (info->var.bits_per_pixel) {
@@ -247,8 +247,8 @@ nv50_fbcon_accel_init(struct fb_info *info)
 	OUT_RING(chan, info->fix.line_length);
 	OUT_RING(chan, info->var.xres_virtual);
 	OUT_RING(chan, info->var.yres_virtual);
-	OUT_RING(chan, upper_32_bits(fb->vma.offset));
-	OUT_RING(chan, lower_32_bits(fb->vma.offset));
+	OUT_RING(chan, upper_32_bits(nvbo->vma.offset));
+	OUT_RING(chan, lower_32_bits(nvbo->vma.offset));
 	BEGIN_RING(chan, NvSub2D, 0x0230, 2);
 	OUT_RING(chan, format);
 	OUT_RING(chan, 1);
@@ -256,8 +256,8 @@ nv50_fbcon_accel_init(struct fb_info *info)
 	OUT_RING(chan, info->fix.line_length);
 	OUT_RING(chan, info->var.xres_virtual);
 	OUT_RING(chan, info->var.yres_virtual);
-	OUT_RING(chan, upper_32_bits(fb->vma.offset));
-	OUT_RING(chan, lower_32_bits(fb->vma.offset));
+	OUT_RING(chan, upper_32_bits(nvbo->vma.offset));
+	OUT_RING(chan, lower_32_bits(nvbo->vma.offset));
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nv50_fifo.c b/drivers/gpu/drm/nouveau/nv50_fifo.c
index 3bc2a56..c34a074 100644
--- a/drivers/gpu/drm/nouveau/nv50_fifo.c
+++ b/drivers/gpu/drm/nouveau/nv50_fifo.c
@@ -230,7 +230,6 @@ nv50_fifo_create_context(struct nouveau_channel *chan)
 	struct drm_device *dev = chan->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_gpuobj *ramfc = NULL;
-        uint64_t ib_offset = chan->pushbuf_base + chan->dma.ib_base * 4;
 	unsigned long flags;
 	int ret;
 
@@ -281,9 +280,8 @@ nv50_fifo_create_context(struct nouveau_channel *chan)
 	nv_wo32(ramfc, 0x7c, 0x30000001);
 	nv_wo32(ramfc, 0x78, 0x00000000);
 	nv_wo32(ramfc, 0x3c, 0x403f6078);
-	nv_wo32(ramfc, 0x50, lower_32_bits(ib_offset));
-	nv_wo32(ramfc, 0x54, upper_32_bits(ib_offset) |
-                drm_order(chan->dma.ib_max + 1) << 16);
+	nv_wo32(ramfc, 0x50, chan->pushbuf_base + chan->dma.ib_base * 4);
+	nv_wo32(ramfc, 0x54, drm_order(chan->dma.ib_max + 1) << 16);
 
 	if (dev_priv->chipset != 0x50) {
 		nv_wo32(chan->ramin, 0, chan->id);
diff --git a/drivers/gpu/drm/nouveau/nv50_gpio.c b/drivers/gpu/drm/nouveau/nv50_gpio.c
index f429e6a..d4f4206 100644
--- a/drivers/gpu/drm/nouveau/nv50_gpio.c
+++ b/drivers/gpu/drm/nouveau/nv50_gpio.c
@@ -25,95 +25,198 @@
 #include "drmP.h"
 #include "nouveau_drv.h"
 #include "nouveau_hw.h"
-#include "nouveau_gpio.h"
 
 #include "nv50_display.h"
 
+static void nv50_gpio_isr(struct drm_device *dev);
+static void nv50_gpio_isr_bh(struct work_struct *work);
+
+struct nv50_gpio_priv {
+	struct list_head handlers;
+	spinlock_t lock;
+};
+
+struct nv50_gpio_handler {
+	struct drm_device *dev;
+	struct list_head head;
+	struct work_struct work;
+	bool inhibit;
+
+	struct dcb_gpio_entry *gpio;
+
+	void (*handler)(void *data, int state);
+	void *data;
+};
+
 static int
-nv50_gpio_location(int line, u32 *reg, u32 *shift)
+nv50_gpio_location(struct dcb_gpio_entry *gpio, uint32_t *reg, uint32_t *shift)
 {
 	const uint32_t nv50_gpio_reg[4] = { 0xe104, 0xe108, 0xe280, 0xe284 };
 
-	if (line >= 32)
+	if (gpio->line >= 32)
 		return -EINVAL;
 
-	*reg = nv50_gpio_reg[line >> 3];
-	*shift = (line & 7) << 2;
+	*reg = nv50_gpio_reg[gpio->line >> 3];
+	*shift = (gpio->line & 7) << 2;
 	return 0;
 }
 
 int
-nv50_gpio_drive(struct drm_device *dev, int line, int dir, int out)
+nv50_gpio_get(struct drm_device *dev, enum dcb_gpio_tag tag)
 {
-	u32 reg, shift;
+	struct dcb_gpio_entry *gpio;
+	uint32_t r, s, v;
+
+	gpio = nouveau_bios_gpio_entry(dev, tag);
+	if (!gpio)
+		return -ENOENT;
 
-	if (nv50_gpio_location(line, &reg, &shift))
+	if (nv50_gpio_location(gpio, &r, &s))
 		return -EINVAL;
 
-	nv_mask(dev, reg, 7 << shift, (((dir ^ 1) << 1) | out) << shift);
-	return 0;
+	v = nv_rd32(dev, r) >> (s + 2);
+	return ((v & 1) == (gpio->state[1] & 1));
 }
 
 int
-nv50_gpio_sense(struct drm_device *dev, int line)
+nv50_gpio_set(struct drm_device *dev, enum dcb_gpio_tag tag, int state)
 {
-	u32 reg, shift;
+	struct dcb_gpio_entry *gpio;
+	uint32_t r, s, v;
+
+	gpio = nouveau_bios_gpio_entry(dev, tag);
+	if (!gpio)
+		return -ENOENT;
 
-	if (nv50_gpio_location(line, &reg, &shift))
+	if (nv50_gpio_location(gpio, &r, &s))
 		return -EINVAL;
 
-	return !!(nv_rd32(dev, reg) & (4 << shift));
+	v  = nv_rd32(dev, r) & ~(0x3 << s);
+	v |= (gpio->state[state] ^ 2) << s;
+	nv_wr32(dev, r, v);
+	return 0;
 }
 
-void
-nv50_gpio_irq_enable(struct drm_device *dev, int line, bool on)
+int
+nv50_gpio_irq_register(struct drm_device *dev, enum dcb_gpio_tag tag,
+		       void (*handler)(void *, int), void *data)
 {
-	u32 reg  = line < 16 ? 0xe050 : 0xe070;
-	u32 mask = 0x00010001 << (line & 0xf);
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio;
+	struct nv50_gpio_priv *priv = pgpio->priv;
+	struct nv50_gpio_handler *gpioh;
+	struct dcb_gpio_entry *gpio;
+	unsigned long flags;
 
-	nv_wr32(dev, reg + 4, mask);
-	nv_mask(dev, reg + 0, mask, on ? mask : 0);
+	gpio = nouveau_bios_gpio_entry(dev, tag);
+	if (!gpio)
+		return -ENOENT;
+
+	gpioh = kzalloc(sizeof(*gpioh), GFP_KERNEL);
+	if (!gpioh)
+		return -ENOMEM;
+
+	INIT_WORK(&gpioh->work, nv50_gpio_isr_bh);
+	gpioh->dev  = dev;
+	gpioh->gpio = gpio;
+	gpioh->handler = handler;
+	gpioh->data = data;
+
+	spin_lock_irqsave(&priv->lock, flags);
+	list_add(&gpioh->head, &priv->handlers);
+	spin_unlock_irqrestore(&priv->lock, flags);
+	return 0;
 }
 
-int
-nvd0_gpio_drive(struct drm_device *dev, int line, int dir, int out)
+void
+nv50_gpio_irq_unregister(struct drm_device *dev, enum dcb_gpio_tag tag,
+			 void (*handler)(void *, int), void *data)
 {
-	u32 data = ((dir ^ 1) << 13) | (out << 12);
-	nv_mask(dev, 0x00d610 + (line * 4), 0x00003000, data);
-	nv_mask(dev, 0x00d604, 0x00000001, 0x00000001); /* update? */
-	return 0;
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio;
+	struct nv50_gpio_priv *priv = pgpio->priv;
+	struct nv50_gpio_handler *gpioh, *tmp;
+	struct dcb_gpio_entry *gpio;
+	LIST_HEAD(tofree);
+	unsigned long flags;
+
+	gpio = nouveau_bios_gpio_entry(dev, tag);
+	if (!gpio)
+		return;
+
+	spin_lock_irqsave(&priv->lock, flags);
+	list_for_each_entry_safe(gpioh, tmp, &priv->handlers, head) {
+		if (gpioh->gpio != gpio ||
+		    gpioh->handler != handler ||
+		    gpioh->data != data)
+			continue;
+		list_move(&gpioh->head, &tofree);
+	}
+	spin_unlock_irqrestore(&priv->lock, flags);
+
+	list_for_each_entry_safe(gpioh, tmp, &tofree, head) {
+		flush_work_sync(&gpioh->work);
+		kfree(gpioh);
+	}
 }
 
-int
-nvd0_gpio_sense(struct drm_device *dev, int line)
+bool
+nv50_gpio_irq_enable(struct drm_device *dev, enum dcb_gpio_tag tag, bool on)
 {
-	return !!(nv_rd32(dev, 0x00d610 + (line * 4)) & 0x00004000);
+	struct dcb_gpio_entry *gpio;
+	u32 reg, mask;
+
+	gpio = nouveau_bios_gpio_entry(dev, tag);
+	if (!gpio)
+		return false;
+
+	reg  = gpio->line < 16 ? 0xe050 : 0xe070;
+	mask = 0x00010001 << (gpio->line & 0xf);
+
+	nv_wr32(dev, reg + 4, mask);
+	reg = nv_mask(dev, reg + 0, mask, on ? mask : 0);
+	return (reg & mask) == mask;
 }
 
-static void
-nv50_gpio_isr(struct drm_device *dev)
+static int
+nv50_gpio_create(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	u32 intr0, intr1 = 0;
-	u32 hi, lo;
+	struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio;
+	struct nv50_gpio_priv *priv;
 
-	intr0 = nv_rd32(dev, 0xe054) & nv_rd32(dev, 0xe050);
-	if (dev_priv->chipset >= 0x90)
-		intr1 = nv_rd32(dev, 0xe074) & nv_rd32(dev, 0xe070);
+	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
+	if (!priv)
+		return -ENOMEM;
 
-	hi = (intr0 & 0x0000ffff) | (intr1 << 16);
-	lo = (intr0 >> 16) | (intr1 & 0xffff0000);
-	nouveau_gpio_isr(dev, 0, hi | lo);
+	INIT_LIST_HEAD(&priv->handlers);
+	spin_lock_init(&priv->lock);
+	pgpio->priv = priv;
+	return 0;
+}
 
-	nv_wr32(dev, 0xe054, intr0);
-	if (dev_priv->chipset >= 0x90)
-		nv_wr32(dev, 0xe074, intr1);
+static void
+nv50_gpio_destroy(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio;
+
+	kfree(pgpio->priv);
+	pgpio->priv = NULL;
 }
 
 int
 nv50_gpio_init(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio;
+	int ret;
+
+	if (!pgpio->priv) {
+		ret = nv50_gpio_create(dev);
+		if (ret)
+			return ret;
+	}
 
 	/* disable, and ack any pending gpio interrupts */
 	nv_wr32(dev, 0xe050, 0x00000000);
@@ -136,4 +239,64 @@ nv50_gpio_fini(struct drm_device *dev)
 	if (dev_priv->chipset >= 0x90)
 		nv_wr32(dev, 0xe070, 0x00000000);
 	nouveau_irq_unregister(dev, 21);
+
+	nv50_gpio_destroy(dev);
+}
+
+static void
+nv50_gpio_isr_bh(struct work_struct *work)
+{
+	struct nv50_gpio_handler *gpioh =
+		container_of(work, struct nv50_gpio_handler, work);
+	struct drm_nouveau_private *dev_priv = gpioh->dev->dev_private;
+	struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio;
+	struct nv50_gpio_priv *priv = pgpio->priv;
+	unsigned long flags;
+	int state;
+
+	state = pgpio->get(gpioh->dev, gpioh->gpio->tag);
+	if (state < 0)
+		return;
+
+	gpioh->handler(gpioh->data, state);
+
+	spin_lock_irqsave(&priv->lock, flags);
+	gpioh->inhibit = false;
+	spin_unlock_irqrestore(&priv->lock, flags);
+}
+
+static void
+nv50_gpio_isr(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	struct nouveau_gpio_engine *pgpio = &dev_priv->engine.gpio;
+	struct nv50_gpio_priv *priv = pgpio->priv;
+	struct nv50_gpio_handler *gpioh;
+	u32 intr0, intr1 = 0;
+	u32 hi, lo, ch;
+
+	intr0 = nv_rd32(dev, 0xe054) & nv_rd32(dev, 0xe050);
+	if (dev_priv->chipset >= 0x90)
+		intr1 = nv_rd32(dev, 0xe074) & nv_rd32(dev, 0xe070);
+
+	hi = (intr0 & 0x0000ffff) | (intr1 << 16);
+	lo = (intr0 >> 16) | (intr1 & 0xffff0000);
+	ch = hi | lo;
+
+	nv_wr32(dev, 0xe054, intr0);
+	if (dev_priv->chipset >= 0x90)
+		nv_wr32(dev, 0xe074, intr1);
+
+	spin_lock(&priv->lock);
+	list_for_each_entry(gpioh, &priv->handlers, head) {
+		if (!(ch & (1 << gpioh->gpio->line)))
+			continue;
+
+		if (gpioh->inhibit)
+			continue;
+		gpioh->inhibit = true;
+
+		schedule_work(&gpioh->work);
+	}
+	spin_unlock(&priv->lock);
 }
diff --git a/drivers/gpu/drm/nouveau/nv50_graph.c b/drivers/gpu/drm/nouveau/nv50_graph.c
index 33d5711..e25cbb4 100644
--- a/drivers/gpu/drm/nouveau/nv50_graph.c
+++ b/drivers/gpu/drm/nouveau/nv50_graph.c
@@ -31,6 +31,7 @@
 #include "nouveau_grctx.h"
 #include "nouveau_dma.h"
 #include "nouveau_vm.h"
+#include "nouveau_ramht.h"
 #include "nv50_evo.h"
 
 struct nv50_graph_engine {
@@ -120,62 +121,71 @@ nv50_graph_unload_context(struct drm_device *dev)
 	return 0;
 }
 
-static int
-nv50_graph_init(struct drm_device *dev, int engine)
+static void
+nv50_graph_init_reset(struct drm_device *dev)
+{
+	uint32_t pmc_e = NV_PMC_ENABLE_PGRAPH | (1 << 21);
+
+	NV_DEBUG(dev, "\n");
+
+	nv_wr32(dev, NV03_PMC_ENABLE, nv_rd32(dev, NV03_PMC_ENABLE) & ~pmc_e);
+	nv_wr32(dev, NV03_PMC_ENABLE, nv_rd32(dev, NV03_PMC_ENABLE) |  pmc_e);
+}
+
+static void
+nv50_graph_init_intr(struct drm_device *dev)
+{
+	NV_DEBUG(dev, "\n");
+
+	nv_wr32(dev, NV03_PGRAPH_INTR, 0xffffffff);
+	nv_wr32(dev, 0x400138, 0xffffffff);
+	nv_wr32(dev, NV40_PGRAPH_INTR_EN, 0xffffffff);
+}
+
+static void
+nv50_graph_init_regs__nv(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nv50_graph_engine *pgraph = nv_engine(dev, engine);
-	u32 units = nv_rd32(dev, 0x001540);
+	uint32_t units = nv_rd32(dev, 0x1540);
 	int i;
 
 	NV_DEBUG(dev, "\n");
 
-	/* master reset */
-	nv_mask(dev, 0x000200, 0x00201000, 0x00000000);
-	nv_mask(dev, 0x000200, 0x00201000, 0x00201000);
-	nv_wr32(dev, 0x40008c, 0x00000004); /* HW_CTX_SWITCH_ENABLED */
-
-	/* reset/enable traps and interrupts */
 	nv_wr32(dev, 0x400804, 0xc0000000);
 	nv_wr32(dev, 0x406800, 0xc0000000);
 	nv_wr32(dev, 0x400c04, 0xc0000000);
 	nv_wr32(dev, 0x401800, 0xc0000000);
 	nv_wr32(dev, 0x405018, 0xc0000000);
 	nv_wr32(dev, 0x402000, 0xc0000000);
-	for (i = 0; i < 16; i++) {
-		if (!(units & (1 << i)))
-			continue;
 
-		if (dev_priv->chipset < 0xa0) {
-			nv_wr32(dev, 0x408900 + (i << 12), 0xc0000000);
-			nv_wr32(dev, 0x408e08 + (i << 12), 0xc0000000);
-			nv_wr32(dev, 0x408314 + (i << 12), 0xc0000000);
-		} else {
-			nv_wr32(dev, 0x408600 + (i << 11), 0xc0000000);
-			nv_wr32(dev, 0x408708 + (i << 11), 0xc0000000);
-			nv_wr32(dev, 0x40831c + (i << 11), 0xc0000000);
+	for (i = 0; i < 16; i++) {
+		if (units & 1 << i) {
+			if (dev_priv->chipset < 0xa0) {
+				nv_wr32(dev, 0x408900 + (i << 12), 0xc0000000);
+				nv_wr32(dev, 0x408e08 + (i << 12), 0xc0000000);
+				nv_wr32(dev, 0x408314 + (i << 12), 0xc0000000);
+			} else {
+				nv_wr32(dev, 0x408600 + (i << 11), 0xc0000000);
+				nv_wr32(dev, 0x408708 + (i << 11), 0xc0000000);
+				nv_wr32(dev, 0x40831c + (i << 11), 0xc0000000);
+			}
 		}
 	}
 
 	nv_wr32(dev, 0x400108, 0xffffffff);
-	nv_wr32(dev, 0x400138, 0xffffffff);
-	nv_wr32(dev, 0x400100, 0xffffffff);
-	nv_wr32(dev, 0x40013c, 0xffffffff);
+
+	nv_wr32(dev, 0x400824, 0x00004000);
 	nv_wr32(dev, 0x400500, 0x00010001);
+}
+
+static void
+nv50_graph_init_zcull(struct drm_device *dev)
+{
+	struct drm_nouveau_private *dev_priv = dev->dev_private;
+	int i;
+
+	NV_DEBUG(dev, "\n");
 
-	/* upload context program, initialise ctxctl defaults */
-	nv_wr32(dev, 0x400324, 0x00000000);
-	for (i = 0; i < pgraph->ctxprog_size; i++)
-		nv_wr32(dev, 0x400328, pgraph->ctxprog[i]);
-	nv_wr32(dev, 0x400824, 0x00000000);
-	nv_wr32(dev, 0x400828, 0x00000000);
-	nv_wr32(dev, 0x40082c, 0x00000000);
-	nv_wr32(dev, 0x400830, 0x00000000);
-	nv_wr32(dev, 0x400724, 0x00000000);
-	nv_wr32(dev, 0x40032c, 0x00000000);
-	nv_wr32(dev, 0x400320, 4);	/* CTXCTL_CMD = NEWCTXDMA */
-
-	/* some unknown zcull magic */
 	switch (dev_priv->chipset & 0xf0) {
 	case 0x50:
 	case 0x80:
@@ -204,18 +214,50 @@ nv50_graph_init(struct drm_device *dev, int engine)
 		nv_wr32(dev, 0x402c28 + (i * 8), 0x00000000);
 		nv_wr32(dev, 0x402c2c + (i * 8), 0x00000000);
 	}
+}
+
+static int
+nv50_graph_init_ctxctl(struct drm_device *dev)
+{
+	struct nv50_graph_engine *pgraph = nv_engine(dev, NVOBJ_ENGINE_GR);
+	int i;
+
+	NV_DEBUG(dev, "\n");
+
+	nv_wr32(dev, NV40_PGRAPH_CTXCTL_UCODE_INDEX, 0);
+	for (i = 0; i < pgraph->ctxprog_size; i++)
+		nv_wr32(dev, NV40_PGRAPH_CTXCTL_UCODE_DATA, pgraph->ctxprog[i]);
+
+	nv_wr32(dev, 0x40008c, 0x00000004); /* HW_CTX_SWITCH_ENABLED */
+	nv_wr32(dev, 0x400320, 4);
+	nv_wr32(dev, NV40_PGRAPH_CTXCTL_CUR, 0);
+	nv_wr32(dev, NV20_PGRAPH_CHANNEL_CTX_POINTER, 0);
+	return 0;
+}
+
+static int
+nv50_graph_init(struct drm_device *dev, int engine)
+{
+	int ret;
 
+	NV_DEBUG(dev, "\n");
+
+	nv50_graph_init_reset(dev);
+	nv50_graph_init_regs__nv(dev);
+	nv50_graph_init_zcull(dev);
+
+	ret = nv50_graph_init_ctxctl(dev);
+	if (ret)
+		return ret;
+
+	nv50_graph_init_intr(dev);
 	return 0;
 }
 
 static int
-nv50_graph_fini(struct drm_device *dev, int engine, bool suspend)
+nv50_graph_fini(struct drm_device *dev, int engine)
 {
-	nv_mask(dev, 0x400500, 0x00010001, 0x00000000);
-	if (!nv_wait(dev, 0x400700, ~0, 0) && suspend) {
-		nv_mask(dev, 0x400500, 0x00010001, 0x00010001);
-		return -EBUSY;
-	}
+	NV_DEBUG(dev, "\n");
 	nv50_graph_unload_context(dev);
 	nv_wr32(dev, 0x40013c, 0x00000000);
 	return 0;
@@ -616,9 +658,9 @@ nv50_pgraph_tp_trap(struct drm_device *dev, int type, uint32_t ustatus_old,
 			}
 			break;
 		case 7: /* MP error */
-			if (ustatus & 0x04030000) {
+			if (ustatus & 0x00010000) {
 				nv50_pgraph_mp_trap(dev, i, display);
-				ustatus &= ~0x04030000;
+				ustatus &= ~0x00010000;
 			}
 			break;
 		case 8: /* TPDMA error */
diff --git a/drivers/gpu/drm/nouveau/nv50_grctx.c b/drivers/gpu/drm/nouveau/nv50_grctx.c
index 4b46d69..de9abff 100644
--- a/drivers/gpu/drm/nouveau/nv50_grctx.c
+++ b/drivers/gpu/drm/nouveau/nv50_grctx.c
@@ -40,12 +40,6 @@
 #define CP_FLAG_UNK0B                 ((0 * 32) + 0xb)
 #define CP_FLAG_UNK0B_CLEAR           0
 #define CP_FLAG_UNK0B_SET             1
-#define CP_FLAG_XFER_SWITCH           ((0 * 32) + 0xe)
-#define CP_FLAG_XFER_SWITCH_DISABLE   0
-#define CP_FLAG_XFER_SWITCH_ENABLE    1
-#define CP_FLAG_STATE                 ((0 * 32) + 0x1c)
-#define CP_FLAG_STATE_STOPPED         0
-#define CP_FLAG_STATE_RUNNING         1
 #define CP_FLAG_UNK1D                 ((0 * 32) + 0x1d)
 #define CP_FLAG_UNK1D_CLEAR           0
 #define CP_FLAG_UNK1D_SET             1
@@ -200,9 +194,6 @@ nv50_grctx_init(struct nouveau_grctx *ctx)
 				   "the devs.\n");
 		return -ENOSYS;
 	}
-
-	cp_set (ctx, STATE, RUNNING);
-	cp_set (ctx, XFER_SWITCH, ENABLE);
 	/* decide whether we're loading/unloading the context */
 	cp_bra (ctx, AUTO_SAVE, PENDING, cp_setup_save);
 	cp_bra (ctx, USER_SAVE, PENDING, cp_setup_save);
@@ -269,8 +260,6 @@ nv50_grctx_init(struct nouveau_grctx *ctx)
 	cp_name(ctx, cp_exit);
 	cp_set (ctx, USER_SAVE, NOT_PENDING);
 	cp_set (ctx, USER_LOAD, NOT_PENDING);
-	cp_set (ctx, XFER_SWITCH, DISABLE);
-	cp_set (ctx, STATE, STOPPED);
 	cp_out (ctx, CP_END);
 	ctx->ctxvals_pos += 0x400; /* padding... no idea why you need it */
 
@@ -601,7 +590,7 @@ nv50_graph_construct_mmio(struct nouveau_grctx *ctx)
 					gr_def(ctx, offset + 0x1c, 0x00880000);
 					break;
 				case 0x86:
-					gr_def(ctx, offset + 0x1c, 0x018c0000);
+					gr_def(ctx, offset + 0x1c, 0x008c0000);
 					break;
 				case 0x92:
 				case 0x96:
diff --git a/drivers/gpu/drm/nouveau/nv50_instmem.c b/drivers/gpu/drm/nouveau/nv50_instmem.c
index a7c12c9..4f95a1e 100644
--- a/drivers/gpu/drm/nouveau/nv50_instmem.c
+++ b/drivers/gpu/drm/nouveau/nv50_instmem.c
@@ -305,9 +305,9 @@ struct nv50_gpuobj_node {
 	u32 align;
 };
 
+
 int
-nv50_instmem_get(struct nouveau_gpuobj *gpuobj, struct nouveau_channel *chan,
-		 u32 size, u32 align)
+nv50_instmem_get(struct nouveau_gpuobj *gpuobj, u32 size, u32 align)
 {
 	struct drm_device *dev = gpuobj->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
@@ -336,7 +336,7 @@ nv50_instmem_get(struct nouveau_gpuobj *gpuobj, struct nouveau_channel *chan,
 		if (!(gpuobj->flags & NVOBJ_FLAG_VM_USER))
 			flags |= NV_MEM_ACCESS_SYS;
 
-		ret = nouveau_vm_get(chan->vm, size, 12, flags,
+		ret = nouveau_vm_get(dev_priv->chan_vm, size, 12, flags,
 				     &node->chan_vma);
 		if (ret) {
 			vram->put(dev, &node->vram);
@@ -345,7 +345,7 @@ nv50_instmem_get(struct nouveau_gpuobj *gpuobj, struct nouveau_channel *chan,
 		}
 
 		nouveau_vm_map(&node->chan_vma, node->vram);
-		gpuobj->linst = node->chan_vma.offset;
+		gpuobj->vinst = node->chan_vma.offset;
 	}
 
 	gpuobj->size = size;
diff --git a/drivers/gpu/drm/nouveau/nv50_mpeg.c b/drivers/gpu/drm/nouveau/nv50_mpeg.c
index b57a2d1..1dc5913 100644
--- a/drivers/gpu/drm/nouveau/nv50_mpeg.c
+++ b/drivers/gpu/drm/nouveau/nv50_mpeg.c
@@ -160,7 +160,7 @@ nv50_mpeg_init(struct drm_device *dev, int engine)
 }
 
 static int
-nv50_mpeg_fini(struct drm_device *dev, int engine, bool suspend)
+nv50_mpeg_fini(struct drm_device *dev, int engine)
 {
 	/*XXX: context save for s/r */
 	nv_mask(dev, 0x00b32c, 0x00000001, 0x00000000);
diff --git a/drivers/gpu/drm/nouveau/nv50_pm.c b/drivers/gpu/drm/nouveau/nv50_pm.c
index d020ed4..8a28100 100644
--- a/drivers/gpu/drm/nouveau/nv50_pm.c
+++ b/drivers/gpu/drm/nouveau/nv50_pm.c
@@ -25,862 +25,122 @@
 #include "drmP.h"
 #include "nouveau_drv.h"
 #include "nouveau_bios.h"
-#include "nouveau_hw.h"
 #include "nouveau_pm.h"
-#include "nouveau_hwsq.h"
-#include "nv50_display.h"
-
-enum clk_src {
-	clk_src_crystal,
-	clk_src_href,
-	clk_src_hclk,
-	clk_src_hclkm3,
-	clk_src_hclkm3d2,
-	clk_src_host,
-	clk_src_nvclk,
-	clk_src_sclk,
-	clk_src_mclk,
-	clk_src_vdec,
-	clk_src_dom6
-};
-
-static u32 read_clk(struct drm_device *, enum clk_src);
-
-static u32
-read_div(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-
-	switch (dev_priv->chipset) {
-	case 0x50: /* it exists, but only has bit 31, not the dividers.. */
-	case 0x84:
-	case 0x86:
-	case 0x98:
-	case 0xa0:
-		return nv_rd32(dev, 0x004700);
-	case 0x92:
-	case 0x94:
-	case 0x96:
-		return nv_rd32(dev, 0x004800);
-	default:
-		return 0x00000000;
-	}
-}
-
-static u32
-read_pll_src(struct drm_device *dev, u32 base)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	u32 coef, ref = read_clk(dev, clk_src_crystal);
-	u32 rsel = nv_rd32(dev, 0x00e18c);
-	int P, N, M, id;
-
-	switch (dev_priv->chipset) {
-	case 0x50:
-	case 0xa0:
-		switch (base) {
-		case 0x4020:
-		case 0x4028: id = !!(rsel & 0x00000004); break;
-		case 0x4008: id = !!(rsel & 0x00000008); break;
-		case 0x4030: id = 0; break;
-		default:
-			NV_ERROR(dev, "ref: bad pll 0x%06x\n", base);
-			return 0;
-		}
-
-		coef = nv_rd32(dev, 0x00e81c + (id * 0x0c));
-		ref *=  (coef & 0x01000000) ? 2 : 4;
-		P    =  (coef & 0x00070000) >> 16;
-		N    = ((coef & 0x0000ff00) >> 8) + 1;
-		M    = ((coef & 0x000000ff) >> 0) + 1;
-		break;
-	case 0x84:
-	case 0x86:
-	case 0x92:
-		coef = nv_rd32(dev, 0x00e81c);
-		P    = (coef & 0x00070000) >> 16;
-		N    = (coef & 0x0000ff00) >> 8;
-		M    = (coef & 0x000000ff) >> 0;
-		break;
-	case 0x94:
-	case 0x96:
-	case 0x98:
-		rsel = nv_rd32(dev, 0x00c050);
-		switch (base) {
-		case 0x4020: rsel = (rsel & 0x00000003) >> 0; break;
-		case 0x4008: rsel = (rsel & 0x0000000c) >> 2; break;
-		case 0x4028: rsel = (rsel & 0x00001800) >> 11; break;
-		case 0x4030: rsel = 3; break;
-		default:
-			NV_ERROR(dev, "ref: bad pll 0x%06x\n", base);
-			return 0;
-		}
-
-		switch (rsel) {
-		case 0: id = 1; break;
-		case 1: return read_clk(dev, clk_src_crystal);
-		case 2: return read_clk(dev, clk_src_href);
-		case 3: id = 0; break;
-		}
-
-		coef =  nv_rd32(dev, 0x00e81c + (id * 0x28));
-		P    = (nv_rd32(dev, 0x00e824 + (id * 0x28)) >> 16) & 7;
-		P   += (coef & 0x00070000) >> 16;
-		N    = (coef & 0x0000ff00) >> 8;
-		M    = (coef & 0x000000ff) >> 0;
-		break;
-	default:
-		BUG_ON(1);
-	}
-
-	if (M)
-		return (ref * N / M) >> P;
-	return 0;
-}
-
-static u32
-read_pll_ref(struct drm_device *dev, u32 base)
-{
-	u32 src, mast = nv_rd32(dev, 0x00c040);
-
-	switch (base) {
-	case 0x004028:
-		src = !!(mast & 0x00200000);
-		break;
-	case 0x004020:
-		src = !!(mast & 0x00400000);
-		break;
-	case 0x004008:
-		src = !!(mast & 0x00010000);
-		break;
-	case 0x004030:
-		src = !!(mast & 0x02000000);
-		break;
-	case 0x00e810:
-		return read_clk(dev, clk_src_crystal);
-	default:
-		NV_ERROR(dev, "bad pll 0x%06x\n", base);
-		return 0;
-	}
-
-	if (src)
-		return read_clk(dev, clk_src_href);
-	return read_pll_src(dev, base);
-}
-
-static u32
-read_pll(struct drm_device *dev, u32 base)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	u32 mast = nv_rd32(dev, 0x00c040);
-	u32 ctrl = nv_rd32(dev, base + 0);
-	u32 coef = nv_rd32(dev, base + 4);
-	u32 ref = read_pll_ref(dev, base);
-	u32 clk = 0;
-	int N1, N2, M1, M2;
-
-	if (base == 0x004028 && (mast & 0x00100000)) {
-		/* wtf, appears to only disable post-divider on nva0 */
-		if (dev_priv->chipset != 0xa0)
-			return read_clk(dev, clk_src_dom6);
-	}
-
-	N2 = (coef & 0xff000000) >> 24;
-	M2 = (coef & 0x00ff0000) >> 16;
-	N1 = (coef & 0x0000ff00) >> 8;
-	M1 = (coef & 0x000000ff);
-	if ((ctrl & 0x80000000) && M1) {
-		clk = ref * N1 / M1;
-		if ((ctrl & 0x40000100) == 0x40000000) {
-			if (M2)
-				clk = clk * N2 / M2;
-			else
-				clk = 0;
-		}
-	}
-
-	return clk;
-}
-
-static u32
-read_clk(struct drm_device *dev, enum clk_src src)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	u32 mast = nv_rd32(dev, 0x00c040);
-	u32 P = 0;
-
-	switch (src) {
-	case clk_src_crystal:
-		return dev_priv->crystal;
-	case clk_src_href:
-		return 100000; /* PCIE reference clock */
-	case clk_src_hclk:
-		return read_clk(dev, clk_src_href) * 27778 / 10000;
-	case clk_src_hclkm3:
-		return read_clk(dev, clk_src_hclk) * 3;
-	case clk_src_hclkm3d2:
-		return read_clk(dev, clk_src_hclk) * 3 / 2;
-	case clk_src_host:
-		switch (mast & 0x30000000) {
-		case 0x00000000: return read_clk(dev, clk_src_href);
-		case 0x10000000: break;
-		case 0x20000000: /* !0x50 */
-		case 0x30000000: return read_clk(dev, clk_src_hclk);
-		}
-		break;
-	case clk_src_nvclk:
-		if (!(mast & 0x00100000))
-			P = (nv_rd32(dev, 0x004028) & 0x00070000) >> 16;
-		switch (mast & 0x00000003) {
-		case 0x00000000: return read_clk(dev, clk_src_crystal) >> P;
-		case 0x00000001: return read_clk(dev, clk_src_dom6);
-		case 0x00000002: return read_pll(dev, 0x004020) >> P;
-		case 0x00000003: return read_pll(dev, 0x004028) >> P;
-		}
-		break;
-	case clk_src_sclk:
-		P = (nv_rd32(dev, 0x004020) & 0x00070000) >> 16;
-		switch (mast & 0x00000030) {
-		case 0x00000000:
-			if (mast & 0x00000080)
-				return read_clk(dev, clk_src_host) >> P;
-			return read_clk(dev, clk_src_crystal) >> P;
-		case 0x00000010: break;
-		case 0x00000020: return read_pll(dev, 0x004028) >> P;
-		case 0x00000030: return read_pll(dev, 0x004020) >> P;
-		}
-		break;
-	case clk_src_mclk:
-		P = (nv_rd32(dev, 0x004008) & 0x00070000) >> 16;
-		if (nv_rd32(dev, 0x004008) & 0x00000200) {
-			switch (mast & 0x0000c000) {
-			case 0x00000000:
-				return read_clk(dev, clk_src_crystal) >> P;
-			case 0x00008000:
-			case 0x0000c000:
-				return read_clk(dev, clk_src_href) >> P;
-			}
-		} else {
-			return read_pll(dev, 0x004008) >> P;
-		}
-		break;
-	case clk_src_vdec:
-		P = (read_div(dev) & 0x00000700) >> 8;
-		switch (dev_priv->chipset) {
-		case 0x84:
-		case 0x86:
-		case 0x92:
-		case 0x94:
-		case 0x96:
-		case 0xa0:
-			switch (mast & 0x00000c00) {
-			case 0x00000000:
-				if (dev_priv->chipset == 0xa0) /* wtf?? */
-					return read_clk(dev, clk_src_nvclk) >> P;
-				return read_clk(dev, clk_src_crystal) >> P;
-			case 0x00000400:
-				return 0;
-			case 0x00000800:
-				if (mast & 0x01000000)
-					return read_pll(dev, 0x004028) >> P;
-				return read_pll(dev, 0x004030) >> P;
-			case 0x00000c00:
-				return read_clk(dev, clk_src_nvclk) >> P;
-			}
-			break;
-		case 0x98:
-			switch (mast & 0x00000c00) {
-			case 0x00000000:
-				return read_clk(dev, clk_src_nvclk) >> P;
-			case 0x00000400:
-				return 0;
-			case 0x00000800:
-				return read_clk(dev, clk_src_hclkm3d2) >> P;
-			case 0x00000c00:
-				return read_clk(dev, clk_src_mclk) >> P;
-			}
-			break;
-		}
-		break;
-	case clk_src_dom6:
-		switch (dev_priv->chipset) {
-		case 0x50:
-		case 0xa0:
-			return read_pll(dev, 0x00e810) >> 2;
-		case 0x84:
-		case 0x86:
-		case 0x92:
-		case 0x94:
-		case 0x96:
-		case 0x98:
-			P = (read_div(dev) & 0x00000007) >> 0;
-			switch (mast & 0x0c000000) {
-			case 0x00000000: return read_clk(dev, clk_src_href);
-			case 0x04000000: break;
-			case 0x08000000: return read_clk(dev, clk_src_hclk);
-			case 0x0c000000:
-				return read_clk(dev, clk_src_hclkm3) >> P;
-			}
-			break;
-		default:
-			break;
-		}
-	default:
-		break;
-	}
-
-	NV_DEBUG(dev, "unknown clock source %d 0x%08x\n", src, mast);
-	return 0;
-}
-
-int
-nv50_pm_clocks_get(struct drm_device *dev, struct nouveau_pm_level *perflvl)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	if (dev_priv->chipset == 0xaa ||
-	    dev_priv->chipset == 0xac)
-		return 0;
-
-	perflvl->core   = read_clk(dev, clk_src_nvclk);
-	perflvl->shader = read_clk(dev, clk_src_sclk);
-	perflvl->memory = read_clk(dev, clk_src_mclk);
-	if (dev_priv->chipset != 0x50) {
-		perflvl->vdec = read_clk(dev, clk_src_vdec);
-		perflvl->dom6 = read_clk(dev, clk_src_dom6);
-	}
-
-	return 0;
-}
 
 struct nv50_pm_state {
 	struct nouveau_pm_level *perflvl;
-	struct hwsq_ucode eclk_hwsq;
-	struct hwsq_ucode mclk_hwsq;
-	u32 mscript;
-	u32 mmast;
-	u32 mctrl;
-	u32 mcoef;
+	struct pll_lims pll;
+	enum pll_types type;
+	int N, M, P;
 };
 
-static u32
-calc_pll(struct drm_device *dev, u32 reg, struct pll_lims *pll,
-	 u32 clk, int *N1, int *M1, int *log2P)
+int
+nv50_pm_clock_get(struct drm_device *dev, u32 id)
 {
-	struct nouveau_pll_vals coef;
-	int ret;
+	struct pll_lims pll;
+	int P, N, M, ret;
+	u32 reg0, reg1;
 
-	ret = get_pll_limits(dev, reg, pll);
+	ret = get_pll_limits(dev, id, &pll);
 	if (ret)
-		return 0;
-
-	pll->vco2.maxfreq = 0;
-	pll->refclk = read_pll_ref(dev, reg);
-	if (!pll->refclk)
-		return 0;
-
-	ret = nouveau_calc_pll_mnp(dev, pll, clk, &coef);
-	if (ret == 0)
-		return 0;
-
-	*N1 = coef.N1;
-	*M1 = coef.M1;
-	*log2P = coef.log2P;
-	return ret;
-}
+		return ret;
 
-static inline u32
-calc_div(u32 src, u32 target, int *div)
-{
-	u32 clk0 = src, clk1 = src;
-	for (*div = 0; *div <= 7; (*div)++) {
-		if (clk0 <= target) {
-			clk1 = clk0 << (*div ? 1 : 0);
-			break;
+	reg0 = nv_rd32(dev, pll.reg + 0);
+	reg1 = nv_rd32(dev, pll.reg + 4);
+
+	if ((reg0 & 0x80000000) == 0) {
+		if (id == PLL_SHADER) {
+			NV_DEBUG(dev, "Shader PLL is disabled. "
+				"Shader clock is twice the core\n");
+			ret = nv50_pm_clock_get(dev, PLL_CORE);
+			if (ret > 0)
+				return ret << 1;
+		} else if (id == PLL_MEMORY) {
+			NV_DEBUG(dev, "Memory PLL is disabled. "
+				"Memory clock is equal to the ref_clk\n");
+			return pll.refclk;
 		}
-		clk0 >>= 1;
-	}
-
-	if (target - clk0 <= clk1 - target)
-		return clk0;
-	(*div)--;
-	return clk1;
-}
-
-static inline u32
-clk_same(u32 a, u32 b)
-{
-	return ((a / 1000) == (b / 1000));
-}
-
-static void
-mclk_precharge(struct nouveau_mem_exec_func *exec)
-{
-	struct nv50_pm_state *info = exec->priv;
-	struct hwsq_ucode *hwsq = &info->mclk_hwsq;
-
-	hwsq_wr32(hwsq, 0x1002d4, 0x00000001);
-}
-
-static void
-mclk_refresh(struct nouveau_mem_exec_func *exec)
-{
-	struct nv50_pm_state *info = exec->priv;
-	struct hwsq_ucode *hwsq = &info->mclk_hwsq;
-
-	hwsq_wr32(hwsq, 0x1002d0, 0x00000001);
-}
-
-static void
-mclk_refresh_auto(struct nouveau_mem_exec_func *exec, bool enable)
-{
-	struct nv50_pm_state *info = exec->priv;
-	struct hwsq_ucode *hwsq = &info->mclk_hwsq;
-
-	hwsq_wr32(hwsq, 0x100210, enable ? 0x80000000 : 0x00000000);
-}
-
-static void
-mclk_refresh_self(struct nouveau_mem_exec_func *exec, bool enable)
-{
-	struct nv50_pm_state *info = exec->priv;
-	struct hwsq_ucode *hwsq = &info->mclk_hwsq;
-
-	hwsq_wr32(hwsq, 0x1002dc, enable ? 0x00000001 : 0x00000000);
-}
-
-static void
-mclk_wait(struct nouveau_mem_exec_func *exec, u32 nsec)
-{
-	struct nv50_pm_state *info = exec->priv;
-	struct hwsq_ucode *hwsq = &info->mclk_hwsq;
-
-	if (nsec > 1000)
-		hwsq_usec(hwsq, (nsec + 500) / 1000);
-}
-
-static u32
-mclk_mrg(struct nouveau_mem_exec_func *exec, int mr)
-{
-	if (mr <= 1)
-		return nv_rd32(exec->dev, 0x1002c0 + ((mr - 0) * 4));
-	if (mr <= 3)
-		return nv_rd32(exec->dev, 0x1002e0 + ((mr - 2) * 4));
-	return 0;
-}
-
-static void
-mclk_mrs(struct nouveau_mem_exec_func *exec, int mr, u32 data)
-{
-	struct drm_nouveau_private *dev_priv = exec->dev->dev_private;
-	struct nv50_pm_state *info = exec->priv;
-	struct hwsq_ucode *hwsq = &info->mclk_hwsq;
-
-	if (mr <= 1) {
-		if (dev_priv->vram_rank_B)
-			hwsq_wr32(hwsq, 0x1002c8 + ((mr - 0) * 4), data);
-		hwsq_wr32(hwsq, 0x1002c0 + ((mr - 0) * 4), data);
-	} else
-	if (mr <= 3) {
-		if (dev_priv->vram_rank_B)
-			hwsq_wr32(hwsq, 0x1002e8 + ((mr - 2) * 4), data);
-		hwsq_wr32(hwsq, 0x1002e0 + ((mr - 2) * 4), data);
-	}
-}
-
-static void
-mclk_clock_set(struct nouveau_mem_exec_func *exec)
-{
-	struct nv50_pm_state *info = exec->priv;
-	struct hwsq_ucode *hwsq = &info->mclk_hwsq;
-	u32 ctrl = nv_rd32(exec->dev, 0x004008);
-
-	info->mmast = nv_rd32(exec->dev, 0x00c040);
-	info->mmast &= ~0xc0000000; /* get MCLK_2 from HREF */
-	info->mmast |=  0x0000c000; /* use MCLK_2 as MPLL_BYPASS clock */
-
-	hwsq_wr32(hwsq, 0xc040, info->mmast);
-	hwsq_wr32(hwsq, 0x4008, ctrl | 0x00000200); /* bypass MPLL */
-	if (info->mctrl & 0x80000000)
-		hwsq_wr32(hwsq, 0x400c, info->mcoef);
-	hwsq_wr32(hwsq, 0x4008, info->mctrl);
-}
-
-static void
-mclk_timing_set(struct nouveau_mem_exec_func *exec)
-{
-	struct drm_device *dev = exec->dev;
-	struct nv50_pm_state *info = exec->priv;
-	struct nouveau_pm_level *perflvl = info->perflvl;
-	struct hwsq_ucode *hwsq = &info->mclk_hwsq;
-	int i;
-
-	for (i = 0; i < 9; i++) {
-		u32 reg = 0x100220 + (i * 4);
-		u32 val = nv_rd32(dev, reg);
-		if (val != perflvl->timing.reg[i])
-			hwsq_wr32(hwsq, reg, perflvl->timing.reg[i]);
-	}
-}
-
-static int
-calc_mclk(struct drm_device *dev, struct nouveau_pm_level *perflvl,
-	  struct nv50_pm_state *info)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	u32 crtc_mask = nv50_display_active_crtcs(dev);
-	struct nouveau_mem_exec_func exec = {
-		.dev = dev,
-		.precharge = mclk_precharge,
-		.refresh = mclk_refresh,
-		.refresh_auto = mclk_refresh_auto,
-		.refresh_self = mclk_refresh_self,
-		.wait = mclk_wait,
-		.mrg = mclk_mrg,
-		.mrs = mclk_mrs,
-		.clock_set = mclk_clock_set,
-		.timing_set = mclk_timing_set,
-		.priv = info
-	};
-	struct hwsq_ucode *hwsq = &info->mclk_hwsq;
-	struct pll_lims pll;
-	int N, M, P;
-	int ret;
-
-	/* use pcie refclock if possible, otherwise use mpll */
-	info->mctrl  = nv_rd32(dev, 0x004008);
-	info->mctrl &= ~0x81ff0200;
-	if (clk_same(perflvl->memory, read_clk(dev, clk_src_href))) {
-		info->mctrl |= 0x00000200 | (pll.log2p_bias << 19);
-	} else {
-		ret = calc_pll(dev, 0x4008, &pll, perflvl->memory, &N, &M, &P);
-		if (ret == 0)
-			return -EINVAL;
-
-		info->mctrl |= 0x80000000 | (P << 22) | (P << 16);
-		info->mctrl |= pll.log2p_bias << 19;
-		info->mcoef  = (N << 8) | M;
 	}
 
-	/* build the ucode which will reclock the memory for us */
-	hwsq_init(hwsq);
-	if (crtc_mask) {
-		hwsq_op5f(hwsq, crtc_mask, 0x00); /* wait for scanout */
-		hwsq_op5f(hwsq, crtc_mask, 0x01); /* wait for vblank */
-	}
-	if (dev_priv->chipset >= 0x92)
-		hwsq_wr32(hwsq, 0x611200, 0x00003300); /* disable scanout */
-	hwsq_setf(hwsq, 0x10, 0); /* disable bus access */
-	hwsq_op5f(hwsq, 0x00, 0x01); /* no idea :s */
+	P = (reg0 & 0x00070000) >> 16;
+	N = (reg1 & 0x0000ff00) >> 8;
+	M = (reg1 & 0x000000ff);
 
-	ret = nouveau_mem_exec(&exec, perflvl);
-	if (ret)
-		return ret;
-
-	hwsq_setf(hwsq, 0x10, 1); /* enable bus access */
-	hwsq_op5f(hwsq, 0x00, 0x00); /* no idea, reverse of 0x00, 0x01? */
-	if (dev_priv->chipset >= 0x92)
-		hwsq_wr32(hwsq, 0x611200, 0x00003330); /* enable scanout */
-	hwsq_fini(hwsq);
-	return 0;
+	return ((pll.refclk * N / M) >> P);
 }
 
 void *
-nv50_pm_clocks_pre(struct drm_device *dev, struct nouveau_pm_level *perflvl)
+nv50_pm_clock_pre(struct drm_device *dev, struct nouveau_pm_level *perflvl,
+		  u32 id, int khz)
 {
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nv50_pm_state *info;
-	struct hwsq_ucode *hwsq;
-	struct pll_lims pll;
-	u32 out, mast, divs, ctrl;
-	int clk, ret = -EINVAL;
-	int N, M, P1, P2;
+	struct nv50_pm_state *state;
+	int dummy, ret;
 
-	if (dev_priv->chipset == 0xaa ||
-	    dev_priv->chipset == 0xac)
-		return ERR_PTR(-ENODEV);
-
-	info = kmalloc(sizeof(*info), GFP_KERNEL);
-	if (!info)
+	state = kzalloc(sizeof(*state), GFP_KERNEL);
+	if (!state)
 		return ERR_PTR(-ENOMEM);
-	info->perflvl = perflvl;
-
-	/* memory: build hwsq ucode which we'll use to reclock memory.
-	 *         use pcie refclock if possible, otherwise use mpll */
-	info->mclk_hwsq.len = 0;
-	if (perflvl->memory) {
-		ret = calc_mclk(dev, perflvl, info);
-		if (ret)
-			goto error;
-		info->mscript = perflvl->memscript;
-	}
-
-	divs = read_div(dev);
-	mast = info->mmast;
-
-	/* start building HWSQ script for engine reclocking */
-	hwsq = &info->eclk_hwsq;
-	hwsq_init(hwsq);
-	hwsq_setf(hwsq, 0x10, 0); /* disable bus access */
-	hwsq_op5f(hwsq, 0x00, 0x01); /* wait for access disabled? */
-
-	/* vdec/dom6: switch to "safe" clocks temporarily */
-	if (perflvl->vdec) {
-		mast &= ~0x00000c00;
-		divs &= ~0x00000700;
-	}
-
-	if (perflvl->dom6) {
-		mast &= ~0x0c000000;
-		divs &= ~0x00000007;
-	}
-
-	hwsq_wr32(hwsq, 0x00c040, mast);
-
-	/* vdec: avoid modifying xpll until we know exactly how the other
-	 * clock domains work, i suspect at least some of them can also be
-	 * tied to xpll...
-	 */
-	if (perflvl->vdec) {
-		/* see how close we can get using nvclk as a source */
-		clk = calc_div(perflvl->core, perflvl->vdec, &P1);
-
-		/* see how close we can get using xpll/hclk as a source */
-		if (dev_priv->chipset != 0x98)
-			out = read_pll(dev, 0x004030);
-		else
-			out = read_clk(dev, clk_src_hclkm3d2);
-		out = calc_div(out, perflvl->vdec, &P2);
-
-		/* select whichever gets us closest */
-		if (abs((int)perflvl->vdec - clk) <=
-		    abs((int)perflvl->vdec - out)) {
-			if (dev_priv->chipset != 0x98)
-				mast |= 0x00000c00;
-			divs |= P1 << 8;
-		} else {
-			mast |= 0x00000800;
-			divs |= P2 << 8;
-		}
-	}
-
-	/* dom6: nfi what this is, but we're limited to various combinations
-	 * of the host clock frequency
-	 */
-	if (perflvl->dom6) {
-		if (clk_same(perflvl->dom6, read_clk(dev, clk_src_href))) {
-			mast |= 0x00000000;
-		} else
-		if (clk_same(perflvl->dom6, read_clk(dev, clk_src_hclk))) {
-			mast |= 0x08000000;
-		} else {
-			clk = read_clk(dev, clk_src_hclk) * 3;
-			clk = calc_div(clk, perflvl->dom6, &P1);
-
-			mast |= 0x0c000000;
-			divs |= P1;
-		}
-	}
-
-	/* vdec/dom6: complete switch to new clocks */
-	switch (dev_priv->chipset) {
-	case 0x92:
-	case 0x94:
-	case 0x96:
-		hwsq_wr32(hwsq, 0x004800, divs);
-		break;
-	default:
-		hwsq_wr32(hwsq, 0x004700, divs);
-		break;
-	}
-
-	hwsq_wr32(hwsq, 0x00c040, mast);
-
-	/* core/shader: make sure sclk/nvclk are disconnected from their
-	 * PLLs (nvclk to dom6, sclk to hclk)
-	 */
-	if (dev_priv->chipset < 0x92)
-		mast = (mast & ~0x001000b0) | 0x00100080;
-	else
-		mast = (mast & ~0x000000b3) | 0x00000081;
-
-	hwsq_wr32(hwsq, 0x00c040, mast);
-
-	/* core: for the moment at least, always use nvpll */
-	clk = calc_pll(dev, 0x4028, &pll, perflvl->core, &N, &M, &P1);
-	if (clk == 0)
-		goto error;
-
-	ctrl  = nv_rd32(dev, 0x004028) & ~0xc03f0100;
-	mast &= ~0x00100000;
-	mast |= 3;
-
-	hwsq_wr32(hwsq, 0x004028, 0x80000000 | (P1 << 19) | (P1 << 16) | ctrl);
-	hwsq_wr32(hwsq, 0x00402c, (N << 8) | M);
-
-	/* shader: tie to nvclk if possible, otherwise use spll.  have to be
-	 * very careful that the shader clock is at least twice the core, or
-	 * some chipsets will be very unhappy.  i expect most or all of these
-	 * cases will be handled by tying to nvclk, but it's possible there's
-	 * corners
-	 */
-	ctrl = nv_rd32(dev, 0x004020) & ~0xc03f0100;
-
-	if (P1-- && perflvl->shader == (perflvl->core << 1)) {
-		hwsq_wr32(hwsq, 0x004020, (P1 << 19) | (P1 << 16) | ctrl);
-		hwsq_wr32(hwsq, 0x00c040, 0x00000020 | mast);
-	} else {
-		clk = calc_pll(dev, 0x4020, &pll, perflvl->shader, &N, &M, &P1);
-		if (clk == 0)
-			goto error;
-		ctrl |= 0x80000000;
-
-		hwsq_wr32(hwsq, 0x004020, (P1 << 19) | (P1 << 16) | ctrl);
-		hwsq_wr32(hwsq, 0x004024, (N << 8) | M);
-		hwsq_wr32(hwsq, 0x00c040, 0x00000030 | mast);
-	}
-
-	hwsq_setf(hwsq, 0x10, 1); /* enable bus access */
-	hwsq_op5f(hwsq, 0x00, 0x00); /* wait for access enabled? */
-	hwsq_fini(hwsq);
-
-	return info;
-error:
-	kfree(info);
-	return ERR_PTR(ret);
-}
-
-static int
-prog_hwsq(struct drm_device *dev, struct hwsq_ucode *hwsq)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	u32 hwsq_data, hwsq_kick;
-	int i;
+	state->type = id;
+	state->perflvl = perflvl;
 
-	if (dev_priv->chipset < 0x94) {
-		hwsq_data = 0x001400;
-		hwsq_kick = 0x00000003;
-	} else {
-		hwsq_data = 0x080000;
-		hwsq_kick = 0x00000001;
+	ret = get_pll_limits(dev, id, &state->pll);
+	if (ret < 0) {
+		kfree(state);
+		return (ret == -ENOENT) ? NULL : ERR_PTR(ret);
 	}
-	/* upload hwsq ucode */
-	nv_mask(dev, 0x001098, 0x00000008, 0x00000000);
-	nv_wr32(dev, 0x001304, 0x00000000);
-	if (dev_priv->chipset >= 0x92)
-		nv_wr32(dev, 0x001318, 0x00000000);
-	for (i = 0; i < hwsq->len / 4; i++)
-		nv_wr32(dev, hwsq_data + (i * 4), hwsq->ptr.u32[i]);
-	nv_mask(dev, 0x001098, 0x00000018, 0x00000018);
-
-	/* launch, and wait for completion */
-	nv_wr32(dev, 0x00130c, hwsq_kick);
-	if (!nv_wait(dev, 0x001308, 0x00000100, 0x00000000)) {
-		NV_ERROR(dev, "hwsq ucode exec timed out\n");
-		NV_ERROR(dev, "0x001308: 0x%08x\n", nv_rd32(dev, 0x001308));
-		for (i = 0; i < hwsq->len / 4; i++) {
-			NV_ERROR(dev, "0x%06x: 0x%08x\n", 0x1400 + (i * 4),
-				 nv_rd32(dev, 0x001400 + (i * 4)));
-		}
 
-		return -EIO;
+	ret = nv50_calc_pll(dev, &state->pll, khz, &state->N, &state->M,
+			    &dummy, &dummy, &state->P);
+	if (ret < 0) {
+		kfree(state);
+		return ERR_PTR(ret);
 	}
 
-	return 0;
+	return state;
 }
 
-int
-nv50_pm_clocks_set(struct drm_device *dev, void *data)
+void
+nv50_pm_clock_set(struct drm_device *dev, void *pre_state)
 {
-	struct nv50_pm_state *info = data;
-	struct bit_entry M;
-	int ret = -EBUSY;
-
-	/* halt and idle execution engines */
-	nv_mask(dev, 0x002504, 0x00000001, 0x00000001);
-	if (!nv_wait(dev, 0x002504, 0x00000010, 0x00000010))
-		goto resume;
-	if (!nv_wait(dev, 0x00251c, 0x0000003f, 0x0000003f))
-		goto resume;
+	struct nv50_pm_state *state = pre_state;
+	struct nouveau_pm_level *perflvl = state->perflvl;
+	u32 reg = state->pll.reg, tmp;
+	struct bit_entry BIT_M;
+	u16 script;
+	int N = state->N;
+	int M = state->M;
+	int P = state->P;
 
-	/* program memory clock, if necessary - must come before engine clock
-	 * reprogramming due to how we construct the hwsq scripts in pre()
-	 */
-	if (info->mclk_hwsq.len) {
-		/* execute some scripts that do ??? from the vbios.. */
-		if (!bit_table(dev, 'M', &M) && M.version == 1) {
-			if (M.length >= 6)
-				nouveau_bios_init_exec(dev, ROM16(M.data[5]));
-			if (M.length >= 8)
-				nouveau_bios_init_exec(dev, ROM16(M.data[7]));
-			if (M.length >= 10)
-				nouveau_bios_init_exec(dev, ROM16(M.data[9]));
-			nouveau_bios_init_exec(dev, info->mscript);
-		}
+	if (state->type == PLL_MEMORY && perflvl->memscript &&
+	    bit_table(dev, 'M', &BIT_M) == 0 &&
+	    BIT_M.version == 1 && BIT_M.length >= 0x0b) {
+		script = ROM16(BIT_M.data[0x05]);
+		if (script)
+			nouveau_bios_run_init_table(dev, script, NULL);
+		script = ROM16(BIT_M.data[0x07]);
+		if (script)
+			nouveau_bios_run_init_table(dev, script, NULL);
+		script = ROM16(BIT_M.data[0x09]);
+		if (script)
+			nouveau_bios_run_init_table(dev, script, NULL);
 
-		ret = prog_hwsq(dev, &info->mclk_hwsq);
-		if (ret)
-			goto resume;
+		nouveau_bios_run_init_table(dev, perflvl->memscript, NULL);
 	}
 
-	/* program engine clocks */
-	ret = prog_hwsq(dev, &info->eclk_hwsq);
-
-resume:
-	nv_mask(dev, 0x002504, 0x00000001, 0x00000000);
-	kfree(info);
-	return ret;
-}
-
-static int
-pwm_info(struct drm_device *dev, int *line, int *ctrl, int *indx)
-{
-	if (*line == 0x04) {
-		*ctrl = 0x00e100;
-		*line = 4;
-		*indx = 0;
-	} else
-	if (*line == 0x09) {
-		*ctrl = 0x00e100;
-		*line = 9;
-		*indx = 1;
-	} else
-	if (*line == 0x10) {
-		*ctrl = 0x00e28c;
-		*line = 0;
-		*indx = 0;
-	} else {
-		NV_ERROR(dev, "unknown pwm ctrl for gpio %d\n", *line);
-		return -ENODEV;
+	if (state->type == PLL_MEMORY) {
+		nv_wr32(dev, 0x100210, 0);
+		nv_wr32(dev, 0x1002dc, 1);
 	}
 
-	return 0;
-}
-
-int
-nv50_pm_pwm_get(struct drm_device *dev, int line, u32 *divs, u32 *duty)
-{
-	int ctrl, id, ret = pwm_info(dev, &line, &ctrl, &id);
-	if (ret)
-		return ret;
+	tmp  = nv_rd32(dev, reg + 0) & 0xfff8ffff;
+	tmp |= 0x80000000 | (P << 16);
+	nv_wr32(dev, reg + 0, tmp);
+	nv_wr32(dev, reg + 4, (N << 8) | M);
 
-	if (nv_rd32(dev, ctrl) & (1 << line)) {
-		*divs = nv_rd32(dev, 0x00e114 + (id * 8));
-		*duty = nv_rd32(dev, 0x00e118 + (id * 8));
-		return 0;
+	if (state->type == PLL_MEMORY) {
+		nv_wr32(dev, 0x1002dc, 0);
+		nv_wr32(dev, 0x100210, 0x80000000);
 	}
 
-	return -EINVAL;
+	kfree(state);
 }
 
-int
-nv50_pm_pwm_set(struct drm_device *dev, int line, u32 divs, u32 duty)
-{
-	int ctrl, id, ret = pwm_info(dev, &line, &ctrl, &id);
-	if (ret)
-		return ret;
-
-	nv_mask(dev, ctrl, 0x00010001 << line, 0x00000001 << line);
-	nv_wr32(dev, 0x00e114 + (id * 8), divs);
-	nv_wr32(dev, 0x00e118 + (id * 8), duty | 0x80000000);
-	return 0;
-}
diff --git a/drivers/gpu/drm/nouveau/nv50_sor.c b/drivers/gpu/drm/nouveau/nv50_sor.c
index a7844ab..c25c5938 100644
--- a/drivers/gpu/drm/nouveau/nv50_sor.c
+++ b/drivers/gpu/drm/nouveau/nv50_sor.c
@@ -36,193 +36,6 @@
 #include "nouveau_crtc.h"
 #include "nv50_display.h"
 
-static u32
-nv50_sor_dp_lane_map(struct drm_device *dev, struct dcb_entry *dcb, u8 lane)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	static const u8 nvaf[] = { 24, 16, 8, 0 }; /* thanks, apple.. */
-	static const u8 nv50[] = { 16, 8, 0, 24 };
-	if (dev_priv->card_type == 0xaf)
-		return nvaf[lane];
-	return nv50[lane];
-}
-
-static void
-nv50_sor_dp_train_set(struct drm_device *dev, struct dcb_entry *dcb, u8 pattern)
-{
-	u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1);
-	nv_mask(dev, NV50_SOR_DP_CTRL(or, link), 0x0f000000, pattern << 24);
-}
-
-static void
-nv50_sor_dp_train_adj(struct drm_device *dev, struct dcb_entry *dcb,
-		      u8 lane, u8 swing, u8 preem)
-{
-	u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1);
-	u32 shift = nv50_sor_dp_lane_map(dev, dcb, lane);
-	u32 mask = 0x000000ff << shift;
-	u8 *table, *entry, *config;
-
-	table = nouveau_dp_bios_data(dev, dcb, &entry);
-	if (!table || (table[0] != 0x20 && table[0] != 0x21)) {
-		NV_ERROR(dev, "PDISP: unsupported DP table for chipset\n");
-		return;
-	}
-
-	config = entry + table[4];
-	while (config[0] != swing || config[1] != preem) {
-		config += table[5];
-		if (config >= entry + table[4] + entry[4] * table[5])
-			return;
-	}
-
-	nv_mask(dev, NV50_SOR_DP_UNK118(or, link), mask, config[2] << shift);
-	nv_mask(dev, NV50_SOR_DP_UNK120(or, link), mask, config[3] << shift);
-	nv_mask(dev, NV50_SOR_DP_UNK130(or, link), 0x0000ff00, config[4] << 8);
-}
-
-static void
-nv50_sor_dp_link_set(struct drm_device *dev, struct dcb_entry *dcb, int crtc,
-		     int link_nr, u32 link_bw, bool enhframe)
-{
-	u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1);
-	u32 dpctrl = nv_rd32(dev, NV50_SOR_DP_CTRL(or, link)) & ~0x001f4000;
-	u32 clksor = nv_rd32(dev, 0x614300 + (or * 0x800)) & ~0x000c0000;
-	u8 *table, *entry, mask;
-	int i;
-
-	table = nouveau_dp_bios_data(dev, dcb, &entry);
-	if (!table || (table[0] != 0x20 && table[0] != 0x21)) {
-		NV_ERROR(dev, "PDISP: unsupported DP table for chipset\n");
-		return;
-	}
-
-	entry = ROMPTR(dev, entry[10]);
-	if (entry) {
-		while (link_bw < ROM16(entry[0]) * 10)
-			entry += 4;
-
-		nouveau_bios_run_init_table(dev, ROM16(entry[2]), dcb, crtc);
-	}
-
-	dpctrl |= ((1 << link_nr) - 1) << 16;
-	if (enhframe)
-		dpctrl |= 0x00004000;
-
-	if (link_bw > 162000)
-		clksor |= 0x00040000;
-
-	nv_wr32(dev, 0x614300 + (or * 0x800), clksor);
-	nv_wr32(dev, NV50_SOR_DP_CTRL(or, link), dpctrl);
-
-	mask = 0;
-	for (i = 0; i < link_nr; i++)
-		mask |= 1 << (nv50_sor_dp_lane_map(dev, dcb, i) >> 3);
-	nv_mask(dev, NV50_SOR_DP_UNK130(or, link), 0x0000000f, mask);
-}
-
-static void
-nv50_sor_dp_link_get(struct drm_device *dev, u32 or, u32 link, u32 *nr, u32 *bw)
-{
-	u32 dpctrl = nv_rd32(dev, NV50_SOR_DP_CTRL(or, link)) & 0x000f0000;
-	u32 clksor = nv_rd32(dev, 0x614300 + (or * 0x800));
-	if (clksor & 0x000c0000)
-		*bw = 270000;
-	else
-		*bw = 162000;
-
-	if      (dpctrl > 0x00030000) *nr = 4;
-	else if (dpctrl > 0x00010000) *nr = 2;
-	else			      *nr = 1;
-}
-
-void
-nv50_sor_dp_calc_tu(struct drm_device *dev, int or, int link, u32 clk, u32 bpp)
-{
-	const u32 symbol = 100000;
-	int bestTU = 0, bestVTUi = 0, bestVTUf = 0, bestVTUa = 0;
-	int TU, VTUi, VTUf, VTUa;
-	u64 link_data_rate, link_ratio, unk;
-	u32 best_diff = 64 * symbol;
-	u32 link_nr, link_bw, r;
-
-	/* calculate packed data rate for each lane */
-	nv50_sor_dp_link_get(dev, or, link, &link_nr, &link_bw);
-	link_data_rate = (clk * bpp / 8) / link_nr;
-
-	/* calculate ratio of packed data rate to link symbol rate */
-	link_ratio = link_data_rate * symbol;
-	r = do_div(link_ratio, link_bw);
-
-	for (TU = 64; TU >= 32; TU--) {
-		/* calculate average number of valid symbols in each TU */
-		u32 tu_valid = link_ratio * TU;
-		u32 calc, diff;
-
-		/* find a hw representation for the fraction.. */
-		VTUi = tu_valid / symbol;
-		calc = VTUi * symbol;
-		diff = tu_valid - calc;
-		if (diff) {
-			if (diff >= (symbol / 2)) {
-				VTUf = symbol / (symbol - diff);
-				if (symbol - (VTUf * diff))
-					VTUf++;
-
-				if (VTUf <= 15) {
-					VTUa  = 1;
-					calc += symbol - (symbol / VTUf);
-				} else {
-					VTUa  = 0;
-					VTUf  = 1;
-					calc += symbol;
-				}
-			} else {
-				VTUa  = 0;
-				VTUf  = min((int)(symbol / diff), 15);
-				calc += symbol / VTUf;
-			}
-
-			diff = calc - tu_valid;
-		} else {
-			/* no remainder, but the hw doesn't like the fractional
-			 * part to be zero.  decrement the integer part and
-			 * have the fraction add a whole symbol back
-			 */
-			VTUa = 0;
-			VTUf = 1;
-			VTUi--;
-		}
-
-		if (diff < best_diff) {
-			best_diff = diff;
-			bestTU = TU;
-			bestVTUa = VTUa;
-			bestVTUf = VTUf;
-			bestVTUi = VTUi;
-			if (diff == 0)
-				break;
-		}
-	}
-
-	if (!bestTU) {
-		NV_ERROR(dev, "DP: unable to find suitable config\n");
-		return;
-	}
-
-	/* XXX close to vbios numbers, but not right */
-	unk  = (symbol - link_ratio) * bestTU;
-	unk *= link_ratio;
-	r = do_div(unk, symbol);
-	r = do_div(unk, symbol);
-	unk += 6;
-
-	nv_mask(dev, NV50_SOR_DP_CTRL(or, link), 0x000001fc, bestTU << 2);
-	nv_mask(dev, NV50_SOR_DP_SCFG(or, link), 0x010f7f3f, bestVTUa << 24 |
-							     bestVTUf << 16 |
-							     bestVTUi << 8 |
-							     unk);
-}
 static void
 nv50_sor_disconnect(struct drm_encoder *encoder)
 {
@@ -247,8 +60,6 @@ nv50_sor_disconnect(struct drm_encoder *encoder)
 	BEGIN_RING(evo, 0, NV50_EVO_UPDATE, 1);
 	OUT_RING  (evo, 0);
 
-	nouveau_hdmi_mode_set(encoder, NULL);
-
 	nv_encoder->crtc = NULL;
 	nv_encoder->last_dpms = DRM_MODE_DPMS_OFF;
 }
@@ -304,13 +115,20 @@ nv50_sor_dpms(struct drm_encoder *encoder, int mode)
 	}
 
 	if (nv_encoder->dcb->type == OUTPUT_DP) {
-		struct dp_train_func func = {
-			.link_set = nv50_sor_dp_link_set,
-			.train_set = nv50_sor_dp_train_set,
-			.train_adj = nv50_sor_dp_train_adj
-		};
+		struct nouveau_i2c_chan *auxch;
+
+		auxch = nouveau_i2c_find(dev, nv_encoder->dcb->i2c_index);
+		if (!auxch)
+			return;
 
-		nouveau_dp_dpms(encoder, mode, nv_encoder->dp.datarate, &func);
+		if (mode == DRM_MODE_DPMS_ON) {
+			u8 status = DP_SET_POWER_D0;
+			nouveau_dp_auxch(auxch, 8, DP_SET_POWER, &status, 1);
+			nouveau_dp_link_train(encoder);
+		} else {
+			u8 status = DP_SET_POWER_D3;
+			nouveau_dp_auxch(auxch, 8, DP_SET_POWER, &status, 1);
+		}
 	}
 }
 
@@ -342,8 +160,11 @@ nv50_sor_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *mode,
 	}
 
 	if (connector->scaling_mode != DRM_MODE_SCALE_NONE &&
-	     connector->native_mode)
-		drm_mode_copy(adjusted_mode, connector->native_mode);
+	     connector->native_mode) {
+		int id = adjusted_mode->base.id;
+		*adjusted_mode = *connector->native_mode;
+		adjusted_mode->base.id = id;
+	}
 
 	return true;
 }
@@ -351,12 +172,6 @@ nv50_sor_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *mode,
 static void
 nv50_sor_prepare(struct drm_encoder *encoder)
 {
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	nv50_sor_disconnect(encoder);
-	if (nv_encoder->dcb->type == OUTPUT_DP) {
-		/* avoid race between link training and supervisor intr */
-		nv50_display_sync(encoder->dev);
-	}
 }
 
 static void
@@ -365,43 +180,33 @@ nv50_sor_commit(struct drm_encoder *encoder)
 }
 
 static void
-nv50_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *umode,
-		  struct drm_display_mode *mode)
+nv50_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
+		  struct drm_display_mode *adjusted_mode)
 {
 	struct nouveau_channel *evo = nv50_display(encoder->dev)->master;
 	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
 	struct drm_device *dev = encoder->dev;
 	struct nouveau_crtc *crtc = nouveau_crtc(encoder->crtc);
-	struct nouveau_connector *nv_connector;
 	uint32_t mode_ctl = 0;
 	int ret;
 
 	NV_DEBUG_KMS(dev, "or %d type %d -> crtc %d\n",
 		     nv_encoder->or, nv_encoder->dcb->type, crtc->index);
-	nv_encoder->crtc = encoder->crtc;
+
+	nv50_sor_dpms(encoder, DRM_MODE_DPMS_ON);
 
 	switch (nv_encoder->dcb->type) {
 	case OUTPUT_TMDS:
 		if (nv_encoder->dcb->sorconf.link & 1) {
-			if (mode->clock < 165000)
+			if (adjusted_mode->clock < 165000)
 				mode_ctl = 0x0100;
 			else
 				mode_ctl = 0x0500;
 		} else
 			mode_ctl = 0x0200;
-
-		nouveau_hdmi_mode_set(encoder, mode);
 		break;
 	case OUTPUT_DP:
-		nv_connector = nouveau_encoder_connector_get(nv_encoder);
-		if (nv_connector && nv_connector->base.display_info.bpc == 6) {
-			nv_encoder->dp.datarate = mode->clock * 18 / 8;
-			mode_ctl |= 0x00020000;
-		} else {
-			nv_encoder->dp.datarate = mode->clock * 24 / 8;
-			mode_ctl |= 0x00050000;
-		}
-
+		mode_ctl |= (nv_encoder->dp.mc_unknown << 16);
 		if (nv_encoder->dcb->sorconf.link & 1)
 			mode_ctl |= 0x00000800;
 		else
@@ -416,22 +221,21 @@ nv50_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *umode,
 	else
 		mode_ctl |= NV50_EVO_SOR_MODE_CTRL_CRTC0;
 
-	if (mode->flags & DRM_MODE_FLAG_NHSYNC)
+	if (adjusted_mode->flags & DRM_MODE_FLAG_NHSYNC)
 		mode_ctl |= NV50_EVO_SOR_MODE_CTRL_NHSYNC;
 
-	if (mode->flags & DRM_MODE_FLAG_NVSYNC)
+	if (adjusted_mode->flags & DRM_MODE_FLAG_NVSYNC)
 		mode_ctl |= NV50_EVO_SOR_MODE_CTRL_NVSYNC;
 
-	nv50_sor_dpms(encoder, DRM_MODE_DPMS_ON);
-
 	ret = RING_SPACE(evo, 2);
 	if (ret) {
 		NV_ERROR(dev, "no space while connecting SOR\n");
-		nv_encoder->crtc = NULL;
 		return;
 	}
 	BEGIN_RING(evo, 0, NV50_EVO_SOR(nv_encoder->or, MODE_CTRL), 1);
 	OUT_RING(evo, mode_ctl);
+
+	nv_encoder->crtc = encoder->crtc;
 }
 
 static struct drm_crtc *
@@ -509,6 +313,29 @@ nv50_sor_create(struct drm_connector *connector, struct dcb_entry *entry)
 	encoder->possible_crtcs = entry->heads;
 	encoder->possible_clones = 0;
 
+	if (nv_encoder->dcb->type == OUTPUT_DP) {
+		int or = nv_encoder->or, link = !(entry->dpconf.sor.link & 1);
+		uint32_t tmp;
+
+		tmp = nv_rd32(dev, 0x61c700 + (or * 0x800));
+
+		switch ((tmp & 0x00000f00) >> 8) {
+		case 8:
+		case 9:
+			nv_encoder->dp.mc_unknown = (tmp & 0x000f0000) >> 16;
+			tmp = nv_rd32(dev, NV50_SOR_DP_CTRL(or, link));
+			nv_encoder->dp.unk0 = tmp & 0x000001fc;
+			tmp = nv_rd32(dev, NV50_SOR_DP_UNK128(or, link));
+			nv_encoder->dp.unk1 = tmp & 0x010f7f3f;
+			break;
+		default:
+			break;
+		}
+
+		if (!nv_encoder->dp.mc_unknown)
+			nv_encoder->dp.mc_unknown = 5;
+	}
+
 	drm_mode_connector_attach_encoder(connector, encoder);
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nv50_vm.c b/drivers/gpu/drm/nouveau/nv50_vm.c
index 44fbac9..1a0dd49 100644
--- a/drivers/gpu/drm/nouveau/nv50_vm.c
+++ b/drivers/gpu/drm/nouveau/nv50_vm.c
@@ -48,7 +48,7 @@ nv50_vm_map_pgt(struct nouveau_gpuobj *pgd, u32 pde,
 			phys |= 0x60;
 		else if (coverage <= 64 * 1024 * 1024)
 			phys |= 0x40;
-		else if (coverage <= 128 * 1024 * 1024)
+		else if (coverage < 128 * 1024 * 1024)
 			phys |= 0x20;
 	}
 
@@ -57,15 +57,27 @@ nv50_vm_map_pgt(struct nouveau_gpuobj *pgd, u32 pde,
 }
 
 static inline u64
-vm_addr(struct nouveau_vma *vma, u64 phys, u32 memtype, u32 target)
+nv50_vm_addr(struct nouveau_vma *vma, u64 phys, u32 memtype, u32 target)
 {
+	struct drm_nouveau_private *dev_priv = vma->vm->dev->dev_private;
+
 	phys |= 1; /* present */
 	phys |= (u64)memtype << 40;
+
+	/* IGPs don't have real VRAM, re-target to stolen system memory */
+	if (target == 0 && dev_priv->vram_sys_base) {
+		phys  += dev_priv->vram_sys_base;
+		target = 3;
+	}
+
 	phys |= target << 4;
+
 	if (vma->access & NV_MEM_ACCESS_SYS)
 		phys |= (1 << 6);
+
 	if (!(vma->access & NV_MEM_ACCESS_WO))
 		phys |= (1 << 3);
+
 	return phys;
 }
 
@@ -73,19 +85,11 @@ void
 nv50_vm_map(struct nouveau_vma *vma, struct nouveau_gpuobj *pgt,
 	    struct nouveau_mem *mem, u32 pte, u32 cnt, u64 phys, u64 delta)
 {
-	struct drm_nouveau_private *dev_priv = vma->vm->dev->dev_private;
 	u32 comp = (mem->memtype & 0x180) >> 7;
-	u32 block, target;
+	u32 block;
 	int i;
 
-	/* IGPs don't have real VRAM, re-target to stolen system memory */
-	target = 0;
-	if (dev_priv->vram_sys_base) {
-		phys += dev_priv->vram_sys_base;
-		target = 3;
-	}
-
-	phys  = vm_addr(vma, phys, mem->memtype, target);
+	phys  = nv50_vm_addr(vma, phys, mem->memtype, 0);
 	pte <<= 3;
 	cnt <<= 3;
 
@@ -121,10 +125,9 @@ void
 nv50_vm_map_sg(struct nouveau_vma *vma, struct nouveau_gpuobj *pgt,
 	       struct nouveau_mem *mem, u32 pte, u32 cnt, dma_addr_t *list)
 {
-	u32 target = (vma->access & NV_MEM_ACCESS_NOSNOOP) ? 3 : 2;
 	pte <<= 3;
 	while (cnt--) {
-		u64 phys = vm_addr(vma, (u64)*list++, mem->memtype, target);
+		u64 phys = nv50_vm_addr(vma, (u64)*list++, mem->memtype, 2);
 		nv_wo32(pgt, pte + 0, lower_32_bits(phys));
 		nv_wo32(pgt, pte + 4, upper_32_bits(phys));
 		pte += 8;
@@ -153,7 +156,7 @@ nv50_vm_flush(struct nouveau_vm *vm)
 	pinstmem->flush(vm->dev);
 
 	/* BAR */
-	if (vm == dev_priv->bar1_vm || vm == dev_priv->bar3_vm) {
+	if (vm != dev_priv->chan_vm) {
 		nv50_vm_flush_engine(vm->dev, 6);
 		return;
 	}
diff --git a/drivers/gpu/drm/nouveau/nv50_vram.c b/drivers/gpu/drm/nouveau/nv50_vram.c
index 9ed9ae39..ffbc3d8 100644
--- a/drivers/gpu/drm/nouveau/nv50_vram.c
+++ b/drivers/gpu/drm/nouveau/nv50_vram.c
@@ -51,7 +51,9 @@ void
 nv50_vram_del(struct drm_device *dev, struct nouveau_mem **pmem)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_mm *mm = &dev_priv->engine.vram.mm;
+	struct ttm_bo_device *bdev = &dev_priv->ttm.bdev;
+	struct ttm_mem_type_manager *man = &bdev->man[TTM_PL_VRAM];
+	struct nouveau_mm *mm = man->priv;
 	struct nouveau_mm_node *this;
 	struct nouveau_mem *mem;
 
@@ -82,7 +84,9 @@ nv50_vram_new(struct drm_device *dev, u64 size, u32 align, u32 size_nc,
 	      u32 memtype, struct nouveau_mem **pmem)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_mm *mm = &dev_priv->engine.vram.mm;
+	struct ttm_bo_device *bdev = &dev_priv->ttm.bdev;
+	struct ttm_mem_type_manager *man = &bdev->man[TTM_PL_VRAM];
+	struct nouveau_mm *mm = man->priv;
 	struct nouveau_mm_node *r;
 	struct nouveau_mem *mem;
 	int comp = (memtype & 0x300) >> 8;
@@ -160,7 +164,7 @@ nv50_vram_rblock(struct drm_device *dev)
 	colbits  =  (r4 & 0x0000f000) >> 12;
 	rowbitsa = ((r4 & 0x000f0000) >> 16) + 8;
 	rowbitsb = ((r4 & 0x00f00000) >> 20) + 8;
-	banks    = 1 << (((r4 & 0x03000000) >> 24) + 2);
+	banks    = ((r4 & 0x01000000) ? 8 : 4);
 
 	rowsize = parts * banks * (1 << colbits) * 8;
 	predicted = rowsize << rowbitsa;
@@ -186,52 +190,22 @@ int
 nv50_vram_init(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_vram_engine *vram = &dev_priv->engine.vram;
-	const u32 rsvd_head = ( 256 * 1024) >> 12; /* vga memory */
-	const u32 rsvd_tail = (1024 * 1024) >> 12; /* vbios etc */
-	u32 pfb714 = nv_rd32(dev, 0x100714);
-	u32 rblock, length;
-
-	switch (pfb714 & 0x00000007) {
-	case 0: dev_priv->vram_type = NV_MEM_TYPE_DDR1; break;
-	case 1:
-		if (nouveau_mem_vbios_type(dev) == NV_MEM_TYPE_DDR3)
-			dev_priv->vram_type = NV_MEM_TYPE_DDR3;
-		else
-			dev_priv->vram_type = NV_MEM_TYPE_DDR2;
-		break;
-	case 2: dev_priv->vram_type = NV_MEM_TYPE_GDDR3; break;
-	case 3: dev_priv->vram_type = NV_MEM_TYPE_GDDR4; break;
-	case 4: dev_priv->vram_type = NV_MEM_TYPE_GDDR5; break;
-	default:
-		break;
-	}
 
-	dev_priv->vram_rank_B = !!(nv_rd32(dev, 0x100200) & 0x4);
 	dev_priv->vram_size  = nv_rd32(dev, 0x10020c);
 	dev_priv->vram_size |= (dev_priv->vram_size & 0xff) << 32;
 	dev_priv->vram_size &= 0xffffffff00ULL;
 
-	/* IGPs, no funky reordering happens here, they don't have VRAM */
-	if (dev_priv->chipset == 0xaa ||
-	    dev_priv->chipset == 0xac ||
-	    dev_priv->chipset == 0xaf) {
+	switch (dev_priv->chipset) {
+	case 0xaa:
+	case 0xac:
+	case 0xaf:
 		dev_priv->vram_sys_base = (u64)nv_rd32(dev, 0x100e10) << 12;
-		rblock = 4096 >> 12;
-	} else {
-		rblock = nv50_vram_rblock(dev) >> 12;
+		dev_priv->vram_rblock_size = 4096;
+		break;
+	default:
+		dev_priv->vram_rblock_size = nv50_vram_rblock(dev);
+		break;
 	}
 
-	length = (dev_priv->vram_size >> 12) - rsvd_head - rsvd_tail;
-
-	return nouveau_mm_init(&vram->mm, rsvd_head, length, rblock);
-}
-
-void
-nv50_vram_fini(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_vram_engine *vram = &dev_priv->engine.vram;
-
-	nouveau_mm_fini(&vram->mm);
+	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nv84_bsp.c b/drivers/gpu/drm/nouveau/nv84_bsp.c
deleted file mode 100644
index 7487573..0000000
--- a/drivers/gpu/drm/nouveau/nv84_bsp.c
+++ /dev/null
@@ -1,83 +0,0 @@
-/*
- * Copyright 2011 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Ben Skeggs
- */
-
-#include "drmP.h"
-#include "nouveau_drv.h"
-#include "nouveau_util.h"
-#include "nouveau_vm.h"
-#include "nouveau_ramht.h"
-
-/*XXX: This stub is currently used on NV98+ also, as soon as this becomes
- *     more than just an enable/disable stub this needs to be split out to
- *     nv98_bsp.c...
- */
-
-struct nv84_bsp_engine {
-	struct nouveau_exec_engine base;
-};
-
-static int
-nv84_bsp_fini(struct drm_device *dev, int engine, bool suspend)
-{
-	if (!(nv_rd32(dev, 0x000200) & 0x00008000))
-		return 0;
-
-	nv_mask(dev, 0x000200, 0x00008000, 0x00000000);
-	return 0;
-}
-
-static int
-nv84_bsp_init(struct drm_device *dev, int engine)
-{
-	nv_mask(dev, 0x000200, 0x00008000, 0x00000000);
-	nv_mask(dev, 0x000200, 0x00008000, 0x00008000);
-	return 0;
-}
-
-static void
-nv84_bsp_destroy(struct drm_device *dev, int engine)
-{
-	struct nv84_bsp_engine *pbsp = nv_engine(dev, engine);
-
-	NVOBJ_ENGINE_DEL(dev, BSP);
-
-	kfree(pbsp);
-}
-
-int
-nv84_bsp_create(struct drm_device *dev)
-{
-	struct nv84_bsp_engine *pbsp;
-
-	pbsp = kzalloc(sizeof(*pbsp), GFP_KERNEL);
-	if (!pbsp)
-		return -ENOMEM;
-
-	pbsp->base.destroy = nv84_bsp_destroy;
-	pbsp->base.init = nv84_bsp_init;
-	pbsp->base.fini = nv84_bsp_fini;
-
-	NVOBJ_ENGINE_ADD(dev, BSP, &pbsp->base);
-	return 0;
-}
diff --git a/drivers/gpu/drm/nouveau/nv84_crypt.c b/drivers/gpu/drm/nouveau/nv84_crypt.c
index edece9c..75b809a 100644
--- a/drivers/gpu/drm/nouveau/nv84_crypt.c
+++ b/drivers/gpu/drm/nouveau/nv84_crypt.c
@@ -138,7 +138,7 @@ nv84_crypt_isr(struct drm_device *dev)
 }
 
 static int
-nv84_crypt_fini(struct drm_device *dev, int engine, bool suspend)
+nv84_crypt_fini(struct drm_device *dev, int engine)
 {
 	nv_wr32(dev, 0x102140, 0x00000000);
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/nv84_vp.c b/drivers/gpu/drm/nouveau/nv84_vp.c
deleted file mode 100644
index 6570d30..0000000
--- a/drivers/gpu/drm/nouveau/nv84_vp.c
+++ /dev/null
@@ -1,83 +0,0 @@
-/*
- * Copyright 2011 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Ben Skeggs
- */
-
-#include "drmP.h"
-#include "nouveau_drv.h"
-#include "nouveau_util.h"
-#include "nouveau_vm.h"
-#include "nouveau_ramht.h"
-
-/*XXX: This stub is currently used on NV98+ also, as soon as this becomes
- *     more than just an enable/disable stub this needs to be split out to
- *     nv98_vp.c...
- */
-
-struct nv84_vp_engine {
-	struct nouveau_exec_engine base;
-};
-
-static int
-nv84_vp_fini(struct drm_device *dev, int engine, bool suspend)
-{
-	if (!(nv_rd32(dev, 0x000200) & 0x00020000))
-		return 0;
-
-	nv_mask(dev, 0x000200, 0x00020000, 0x00000000);
-	return 0;
-}
-
-static int
-nv84_vp_init(struct drm_device *dev, int engine)
-{
-	nv_mask(dev, 0x000200, 0x00020000, 0x00000000);
-	nv_mask(dev, 0x000200, 0x00020000, 0x00020000);
-	return 0;
-}
-
-static void
-nv84_vp_destroy(struct drm_device *dev, int engine)
-{
-	struct nv84_vp_engine *pvp = nv_engine(dev, engine);
-
-	NVOBJ_ENGINE_DEL(dev, VP);
-
-	kfree(pvp);
-}
-
-int
-nv84_vp_create(struct drm_device *dev)
-{
-	struct nv84_vp_engine *pvp;
-
-	pvp = kzalloc(sizeof(*pvp), GFP_KERNEL);
-	if (!pvp)
-		return -ENOMEM;
-
-	pvp->base.destroy = nv84_vp_destroy;
-	pvp->base.init = nv84_vp_init;
-	pvp->base.fini = nv84_vp_fini;
-
-	NVOBJ_ENGINE_ADD(dev, VP, &pvp->base);
-	return 0;
-}
diff --git a/drivers/gpu/drm/nouveau/nv98_crypt.c b/drivers/gpu/drm/nouveau/nv98_crypt.c
deleted file mode 100644
index db94ff0..0000000
--- a/drivers/gpu/drm/nouveau/nv98_crypt.c
+++ /dev/null
@@ -1,78 +0,0 @@
-/*
- * Copyright 2011 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Ben Skeggs
- */
-
-#include "drmP.h"
-#include "nouveau_drv.h"
-#include "nouveau_util.h"
-#include "nouveau_vm.h"
-#include "nouveau_ramht.h"
-
-struct nv98_crypt_engine {
-	struct nouveau_exec_engine base;
-};
-
-static int
-nv98_crypt_fini(struct drm_device *dev, int engine, bool suspend)
-{
-	if (!(nv_rd32(dev, 0x000200) & 0x00004000))
-		return 0;
-
-	nv_mask(dev, 0x000200, 0x00004000, 0x00000000);
-	return 0;
-}
-
-static int
-nv98_crypt_init(struct drm_device *dev, int engine)
-{
-	nv_mask(dev, 0x000200, 0x00004000, 0x00000000);
-	nv_mask(dev, 0x000200, 0x00004000, 0x00004000);
-	return 0;
-}
-
-static void
-nv98_crypt_destroy(struct drm_device *dev, int engine)
-{
-	struct nv98_crypt_engine *pcrypt = nv_engine(dev, engine);
-
-	NVOBJ_ENGINE_DEL(dev, CRYPT);
-
-	kfree(pcrypt);
-}
-
-int
-nv98_crypt_create(struct drm_device *dev)
-{
-	struct nv98_crypt_engine *pcrypt;
-
-	pcrypt = kzalloc(sizeof(*pcrypt), GFP_KERNEL);
-	if (!pcrypt)
-		return -ENOMEM;
-
-	pcrypt->base.destroy = nv98_crypt_destroy;
-	pcrypt->base.init = nv98_crypt_init;
-	pcrypt->base.fini = nv98_crypt_fini;
-
-	NVOBJ_ENGINE_ADD(dev, CRYPT, &pcrypt->base);
-	return 0;
-}
diff --git a/drivers/gpu/drm/nouveau/nv98_ppp.c b/drivers/gpu/drm/nouveau/nv98_ppp.c
deleted file mode 100644
index a987dd6..0000000
--- a/drivers/gpu/drm/nouveau/nv98_ppp.c
+++ /dev/null
@@ -1,78 +0,0 @@
-/*
- * Copyright 2011 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Ben Skeggs
- */
-
-#include "drmP.h"
-#include "nouveau_drv.h"
-#include "nouveau_util.h"
-#include "nouveau_vm.h"
-#include "nouveau_ramht.h"
-
-struct nv98_ppp_engine {
-	struct nouveau_exec_engine base;
-};
-
-static int
-nv98_ppp_fini(struct drm_device *dev, int engine, bool suspend)
-{
-	if (!(nv_rd32(dev, 0x000200) & 0x00000002))
-		return 0;
-
-	nv_mask(dev, 0x000200, 0x00000002, 0x00000000);
-	return 0;
-}
-
-static int
-nv98_ppp_init(struct drm_device *dev, int engine)
-{
-	nv_mask(dev, 0x000200, 0x00000002, 0x00000000);
-	nv_mask(dev, 0x000200, 0x00000002, 0x00000002);
-	return 0;
-}
-
-static void
-nv98_ppp_destroy(struct drm_device *dev, int engine)
-{
-	struct nv98_ppp_engine *pppp = nv_engine(dev, engine);
-
-	NVOBJ_ENGINE_DEL(dev, PPP);
-
-	kfree(pppp);
-}
-
-int
-nv98_ppp_create(struct drm_device *dev)
-{
-	struct nv98_ppp_engine *pppp;
-
-	pppp = kzalloc(sizeof(*pppp), GFP_KERNEL);
-	if (!pppp)
-		return -ENOMEM;
-
-	pppp->base.destroy = nv98_ppp_destroy;
-	pppp->base.init = nv98_ppp_init;
-	pppp->base.fini = nv98_ppp_fini;
-
-	NVOBJ_ENGINE_ADD(dev, PPP, &pppp->base);
-	return 0;
-}
diff --git a/drivers/gpu/drm/nouveau/nva3_copy.c b/drivers/gpu/drm/nouveau/nva3_copy.c
index 8f356d5..b86820a 100644
--- a/drivers/gpu/drm/nouveau/nva3_copy.c
+++ b/drivers/gpu/drm/nouveau/nva3_copy.c
@@ -140,7 +140,7 @@ nva3_copy_init(struct drm_device *dev, int engine)
 }
 
 static int
-nva3_copy_fini(struct drm_device *dev, int engine, bool suspend)
+nva3_copy_fini(struct drm_device *dev, int engine)
 {
 	nv_mask(dev, 0x104048, 0x00000003, 0x00000000);
 
diff --git a/drivers/gpu/drm/nouveau/nva3_copy.fuc b/drivers/gpu/drm/nouveau/nva3_copy.fuc
index abc3662..eaf35f8 100644
--- a/drivers/gpu/drm/nouveau/nva3_copy.fuc
+++ b/drivers/gpu/drm/nouveau/nva3_copy.fuc
@@ -31,9 +31,8 @@
  */
 
 ifdef(`NVA3',
-.section #nva3_pcopy_data
-,
-.section #nvc0_pcopy_data
+.section nva3_pcopy_data,
+.section nvc0_pcopy_data
 )
 
 ctx_object:                   .b32 0
@@ -43,7 +42,7 @@ ctx_dma_query:                .b32 0
 ctx_dma_src:                  .b32 0
 ctx_dma_dst:                  .b32 0
 ,)
-.equ #ctx_dma_count 3
+.equ ctx_dma_count 3
 ctx_query_address_high:       .b32 0
 ctx_query_address_low:        .b32 0
 ctx_query_counter:            .b32 0
@@ -79,65 +78,64 @@ ctx_ycnt:                     .b32 0
 dispatch_table:
 // mthd 0x0000, NAME
 .b16 0x000 1
-.b32 #ctx_object                     ~0xffffffff
+.b32 ctx_object                     ~0xffffffff
 // mthd 0x0100, NOP
 .b16 0x040 1
-.b32 0x00010000 + #cmd_nop           ~0xffffffff
+.b32 0x00010000 + cmd_nop           ~0xffffffff
 // mthd 0x0140, PM_TRIGGER
 .b16 0x050 1
-.b32 0x00010000 + #cmd_pm_trigger    ~0xffffffff
+.b32 0x00010000 + cmd_pm_trigger    ~0xffffffff
 ifdef(`NVA3', `
 // mthd 0x0180-0x018c, DMA_
-.b16 0x060 #ctx_dma_count
+.b16 0x060 ctx_dma_count
 dispatch_dma:
-.b32 0x00010000 + #cmd_dma           ~0xffffffff
-.b32 0x00010000 + #cmd_dma           ~0xffffffff
-.b32 0x00010000 + #cmd_dma           ~0xffffffff
+.b32 0x00010000 + cmd_dma           ~0xffffffff
+.b32 0x00010000 + cmd_dma           ~0xffffffff
+.b32 0x00010000 + cmd_dma           ~0xffffffff
 ',)
 // mthd 0x0200-0x0218, SRC_TILE
 .b16 0x80 7
-.b32 #ctx_src_tile_mode              ~0x00000fff
-.b32 #ctx_src_xsize                  ~0x0007ffff
-.b32 #ctx_src_ysize                  ~0x00001fff
-.b32 #ctx_src_zsize                  ~0x000007ff
-.b32 #ctx_src_zoff                   ~0x00000fff
-.b32 #ctx_src_xoff                   ~0x0007ffff
-.b32 #ctx_src_yoff                   ~0x00001fff
+.b32 ctx_src_tile_mode              ~0x00000fff
+.b32 ctx_src_xsize                  ~0x0007ffff
+.b32 ctx_src_ysize                  ~0x00001fff
+.b32 ctx_src_zsize                  ~0x000007ff
+.b32 ctx_src_zoff                   ~0x00000fff
+.b32 ctx_src_xoff                   ~0x0007ffff
+.b32 ctx_src_yoff                   ~0x00001fff
 // mthd 0x0220-0x0238, DST_TILE
 .b16 0x88 7
-.b32 #ctx_dst_tile_mode              ~0x00000fff
-.b32 #ctx_dst_xsize                  ~0x0007ffff
-.b32 #ctx_dst_ysize                  ~0x00001fff
-.b32 #ctx_dst_zsize                  ~0x000007ff
-.b32 #ctx_dst_zoff                   ~0x00000fff
-.b32 #ctx_dst_xoff                   ~0x0007ffff
-.b32 #ctx_dst_yoff                   ~0x00001fff
+.b32 ctx_dst_tile_mode              ~0x00000fff
+.b32 ctx_dst_xsize                  ~0x0007ffff
+.b32 ctx_dst_ysize                  ~0x00001fff
+.b32 ctx_dst_zsize                  ~0x000007ff
+.b32 ctx_dst_zoff                   ~0x00000fff
+.b32 ctx_dst_xoff                   ~0x0007ffff
+.b32 ctx_dst_yoff                   ~0x00001fff
 // mthd 0x0300-0x0304, EXEC, WRCACHE_FLUSH
 .b16 0xc0 2
-.b32 0x00010000 + #cmd_exec          ~0xffffffff
-.b32 0x00010000 + #cmd_wrcache_flush ~0xffffffff
+.b32 0x00010000 + cmd_exec          ~0xffffffff
+.b32 0x00010000 + cmd_wrcache_flush ~0xffffffff
 // mthd 0x030c-0x0340, various stuff
 .b16 0xc3 14
-.b32 #ctx_src_address_high           ~0x000000ff
-.b32 #ctx_src_address_low            ~0xfffffff0
-.b32 #ctx_dst_address_high           ~0x000000ff
-.b32 #ctx_dst_address_low            ~0xfffffff0
-.b32 #ctx_src_pitch                  ~0x0007ffff
-.b32 #ctx_dst_pitch                  ~0x0007ffff
-.b32 #ctx_xcnt                       ~0x0000ffff
-.b32 #ctx_ycnt                       ~0x00001fff
-.b32 #ctx_format                     ~0x0333ffff
-.b32 #ctx_swz_const0                 ~0xffffffff
-.b32 #ctx_swz_const1                 ~0xffffffff
-.b32 #ctx_query_address_high         ~0x000000ff
-.b32 #ctx_query_address_low          ~0xffffffff
-.b32 #ctx_query_counter              ~0xffffffff
+.b32 ctx_src_address_high           ~0x000000ff
+.b32 ctx_src_address_low            ~0xfffffff0
+.b32 ctx_dst_address_high           ~0x000000ff
+.b32 ctx_dst_address_low            ~0xfffffff0
+.b32 ctx_src_pitch                  ~0x0007ffff
+.b32 ctx_dst_pitch                  ~0x0007ffff
+.b32 ctx_xcnt                       ~0x0000ffff
+.b32 ctx_ycnt                       ~0x00001fff
+.b32 ctx_format                     ~0x0333ffff
+.b32 ctx_swz_const0                 ~0xffffffff
+.b32 ctx_swz_const1                 ~0xffffffff
+.b32 ctx_query_address_high         ~0x000000ff
+.b32 ctx_query_address_low          ~0xffffffff
+.b32 ctx_query_counter              ~0xffffffff
 .b16 0x800 0
 
 ifdef(`NVA3',
-.section #nva3_pcopy_code
-,
-.section #nvc0_pcopy_code
+.section nva3_pcopy_code,
+.section nvc0_pcopy_code
 )
 
 main:
@@ -145,12 +143,12 @@ main:
    mov $sp $r0
 
    // setup i0 handler and route fifo and ctxswitch to it
-   mov $r1 #ih
+   mov $r1 ih
    mov $iv0 $r1
    mov $r1 0x400
    movw $r2 0xfff3
    sethi $r2 0
-   iowr I[$r1 + 0x300] $r2
+   iowr I[$r2 + 0x300] $r2
 
    // enable interrupts
    or $r2 0xc
@@ -166,19 +164,19 @@ main:
    bset $flags $p0
    spin:
       sleep $p0
-      bra #spin
+      bra spin
 
 // i0 handler
 ih:
    iord $r1 I[$r0 + 0x200]
 
    and $r2 $r1 0x00000008
-   bra e #ih_no_chsw
-      call #chsw
+   bra e ih_no_chsw
+      call chsw
    ih_no_chsw:
    and $r2 $r1 0x00000004
-   bra e #ih_no_cmd
-      call #dispatch
+   bra e ih_no_cmd
+      call dispatch
 
    ih_no_cmd:
    and $r1 $r1 0x0000000c
@@ -237,9 +235,9 @@ ifdef(`NVA3', `
    sethi $r4 0x60000
 
    // swap!
-   bra $p1 #swctx_load
+   bra $p1 swctx_load
       xdst $r0 $r4
-      bra #swctx_done
+      bra swctx_done
    swctx_load:
       xdld $r0 $r4
    swctx_done:
@@ -253,9 +251,9 @@ chsw:
 
    // if it's active, unload it and return
    xbit $r15 $r3 0x1e
-   bra e #chsw_no_unload
+   bra e chsw_no_unload
       bclr $flags $p1
-      call #swctx
+      call swctx
       bclr $r3 0x1e
       iowr I[$r2] $r3
       mov $r4 1
@@ -268,20 +266,20 @@ chsw:
 
    // is there a channel waiting to be loaded?
    xbit $r13 $r3 0x1e
-   bra e #chsw_finish_load
+   bra e chsw_finish_load
       bset $flags $p1
-      call #swctx
+      call swctx
 ifdef(`NVA3',
       // load dma objects back into TARGET regs
-      mov $r5 #ctx_dma
-      mov $r6 #ctx_dma_count
+      mov $r5 ctx_dma
+      mov $r6 ctx_dma_count
       chsw_load_ctx_dma:
          ld b32 $r7 D[$r5 + $r6 * 4]
          add b32 $r8 $r6 0x180
          shl b32 $r8 8
          iowr I[$r8] $r7
          sub b32 $r6 1
-         bra nc #chsw_load_ctx_dma
+         bra nc chsw_load_ctx_dma
 ,)
 
    chsw_finish_load:
@@ -299,7 +297,7 @@ dispatch:
    shl b32 $r2 0x10
 
    // lookup method in the dispatch table, ILLEGAL_MTHD if not found
-   mov $r5 #dispatch_table
+   mov $r5 dispatch_table
    clear b32 $r6
    clear b32 $r7
    dispatch_loop:
@@ -307,14 +305,14 @@ dispatch:
       ld b16 $r7 D[$r5 + 2]
       add b32 $r5 4
       cmpu b32 $r4 $r6
-      bra c #dispatch_illegal_mthd
+      bra c dispatch_illegal_mthd
       add b32 $r7 $r6
       cmpu b32 $r4 $r7
-      bra c #dispatch_valid_mthd
+      bra c dispatch_valid_mthd
       sub b32 $r7 $r6
       shl b32 $r7 3
       add b32 $r5 $r7
-      bra #dispatch_loop
+      bra dispatch_loop
 
    // ensure no bits set in reserved fields, INVALID_BITFIELD
    dispatch_valid_mthd:
@@ -324,20 +322,20 @@ dispatch:
    ld b32 $r5 D[$r4 + 4]
    and $r5 $r3
    cmpu b32 $r5 0
-   bra ne #dispatch_invalid_bitfield
+   bra ne dispatch_invalid_bitfield
 
    // depending on dispatch flags: execute method, or save data as state
    ld b16 $r5 D[$r4 + 0]
    ld b16 $r6 D[$r4 + 2]
    cmpu b32 $r6 0
-   bra ne #dispatch_cmd
+   bra ne dispatch_cmd
       st b32 D[$r5] $r3
-      bra #dispatch_done
+      bra dispatch_done
    dispatch_cmd:
       bclr $flags $p1
       call $r5
-      bra $p1 #dispatch_error
-      bra #dispatch_done
+      bra $p1 dispatch_error
+      bra dispatch_done
 
    dispatch_invalid_bitfield:
    or $r2 2
@@ -355,7 +353,7 @@ dispatch:
       iord $r2 I[$r0 + 0x200]
       and $r2 0x40
       cmpu b32 $r2 0
-      bra ne #hostirq_wait
+      bra ne hostirq_wait
 
    dispatch_done:
    mov $r2 0x1d00
@@ -411,10 +409,10 @@ ifdef(`NVA3',
 //       $r2: hostirq state
 //       $r3: data
 cmd_dma:
-   sub b32 $r4 #dispatch_dma
+   sub b32 $r4 dispatch_dma
    shr b32 $r4 1
    bset $r3 0x1e
-   st b32 D[$r4 + #ctx_dma] $r3
+   st b32 D[$r4 + ctx_dma] $r3
    add b32 $r4 0x600
    shl b32 $r4 6
    iowr I[$r4] $r3
@@ -432,7 +430,7 @@ cmd_exec_set_format:
    st b32 D[$sp + 0x0c] $r0
 
    // extract cpp, src_ncomp and dst_ncomp from FORMAT
-   ld b32 $r4 D[$r0 + #ctx_format]
+   ld b32 $r4 D[$r0 + ctx_format]
    extr $r5 $r4 16:17
    add b32 $r5 1
    extr $r6 $r4 20:21
@@ -450,22 +448,22 @@ cmd_exec_set_format:
       clear b32 $r11
       bpc_loop:
          cmpu b8 $r10 4
-         bra nc #cmp_c0
+         bra nc cmp_c0
             mulu $r12 $r10 $r5
             add b32 $r12 $r11
             bset $flags $p2
-            bra #bpc_next
+            bra bpc_next
          cmp_c0:
-         bra ne #cmp_c1
+         bra ne cmp_c1
             mov $r12 0x10
             add b32 $r12 $r11
-            bra #bpc_next
+            bra bpc_next
          cmp_c1:
          cmpu b8 $r10 6
-         bra nc #cmp_zero
+         bra nc cmp_zero
             mov $r12 0x14
             add b32 $r12 $r11
-            bra #bpc_next
+            bra bpc_next
          cmp_zero:
             mov $r12 0x80
          bpc_next:
@@ -473,22 +471,22 @@ cmd_exec_set_format:
          add b32 $r8 1
          add b32 $r11 1
          cmpu b32 $r11 $r5
-         bra c #bpc_loop
+         bra c bpc_loop
       add b32 $r9 1
       cmpu b32 $r9 $r7
-      bra c #ncomp_loop
+      bra c ncomp_loop
 
    // SRC_XCNT = (xcnt * src_cpp), or 0 if no src ref in swz (hw will hang)
    mulu $r6 $r5
-   st b32 D[$r0 + #ctx_src_cpp] $r6
-   ld b32 $r8 D[$r0 + #ctx_xcnt]
+   st b32 D[$r0 + ctx_src_cpp] $r6
+   ld b32 $r8 D[$r0 + ctx_xcnt]
    mulu $r6 $r8
-   bra $p2 #dst_xcnt
+   bra $p2 dst_xcnt
    clear b32 $r6
 
    dst_xcnt:
    mulu $r7 $r5
-   st b32 D[$r0 + #ctx_dst_cpp] $r7
+   st b32 D[$r0 + ctx_dst_cpp] $r7
    mulu $r7 $r8
 
    mov $r5 0x810
@@ -496,10 +494,10 @@ cmd_exec_set_format:
    iowr I[$r5 + 0x000] $r6
    iowr I[$r5 + 0x100] $r7
    add b32 $r5 0x800
-   ld b32 $r6 D[$r0 + #ctx_dst_cpp]
+   ld b32 $r6 D[$r0 + ctx_dst_cpp]
    sub b32 $r6 1
    shl b32 $r6 8
-   ld b32 $r7 D[$r0 + #ctx_src_cpp]
+   ld b32 $r7 D[$r0 + ctx_src_cpp]
    sub b32 $r7 1
    or $r6 $r7
    iowr I[$r5 + 0x000] $r6
@@ -513,9 +511,9 @@ cmd_exec_set_format:
    ld b32 $r6 D[$sp + 0x0c]
    iowr I[$r5 + 0x300] $r6
    add b32 $r5 0x400
-   ld b32 $r6 D[$r0 + #ctx_swz_const0]
+   ld b32 $r6 D[$r0 + ctx_swz_const0]
    iowr I[$r5 + 0x000] $r6
-   ld b32 $r6 D[$r0 + #ctx_swz_const1]
+   ld b32 $r6 D[$r0 + ctx_swz_const1]
    iowr I[$r5 + 0x100] $r6
    add $sp 0x10
    ret
@@ -545,7 +543,7 @@ cmd_exec_set_format:
 //
 cmd_exec_set_surface_tiled:
    // translate TILE_MODE into Tp, Th, Td shift values
-   ld b32 $r7 D[$r5 + #ctx_src_tile_mode]
+   ld b32 $r7 D[$r5 + ctx_src_tile_mode]
    extr $r9 $r7 8:11
    extr $r8 $r7 4:7
 ifdef(`NVA3',
@@ -555,9 +553,9 @@ ifdef(`NVA3',
 )
    extr $r7 $r7 0:3
    cmp b32 $r7 0xe
-   bra ne #xtile64
+   bra ne xtile64
    mov $r7 4
-   bra #xtileok
+   bra xtileok
    xtile64:
    xbit $r7 $flags $p2
    add b32 $r7 17
@@ -567,8 +565,8 @@ ifdef(`NVA3',
 
    // Op = (x * cpp) & ((1 << Tp) - 1)
    // Tx = (x * cpp) >> Tp
-   ld b32 $r10 D[$r5 + #ctx_src_xoff]
-   ld b32 $r11 D[$r5 + #ctx_src_cpp]
+   ld b32 $r10 D[$r5 + ctx_src_xoff]
+   ld b32 $r11 D[$r5 + ctx_src_cpp]
    mulu $r10 $r11
    mov $r11 1
    shl b32 $r11 $r7
@@ -578,7 +576,7 @@ ifdef(`NVA3',
 
    // Tyo = y & ((1 << Th) - 1)
    // Ty  = y >> Th
-   ld b32 $r13 D[$r5 + #ctx_src_yoff]
+   ld b32 $r13 D[$r5 + ctx_src_yoff]
    mov $r14 1
    shl b32 $r14 $r8
    sub b32 $r14 1
@@ -600,8 +598,8 @@ ifdef(`NVA3',
    add b32 $r12 $r11
 
    // nTx = ((w * cpp) + ((1 << Tp) - 1) >> Tp)
-   ld b32 $r15 D[$r5 + #ctx_src_xsize]
-   ld b32 $r11 D[$r5 + #ctx_src_cpp]
+   ld b32 $r15 D[$r5 + ctx_src_xsize]
+   ld b32 $r11 D[$r5 + ctx_src_cpp]
    mulu $r15 $r11
    mov $r11 1
    shl b32 $r11 $r7
@@ -611,7 +609,7 @@ ifdef(`NVA3',
    push $r15
 
    // nTy = (h + ((1 << Th) - 1)) >> Th
-   ld b32 $r15 D[$r5 + #ctx_src_ysize]
+   ld b32 $r15 D[$r5 + ctx_src_ysize]
    mov $r11 1
    shl b32 $r11 $r8
    sub b32 $r11 1
@@ -631,7 +629,7 @@ ifdef(`NVA3',
    // Tz  = z >> Td
    // Op += Tzo << Tys
    // Ts  = Tys + Td
-   ld b32 $r8 D[$r5 + #ctx_src_zoff]
+   ld b32 $r8 D[$r5 + ctx_src_zoff]
    mov $r14 1
    shl b32 $r14 $r9
    sub b32 $r14 1
@@ -658,8 +656,8 @@ ifdef(`NVA3',
 
    // SRC_ADDRESS_LOW   = (Ot + Op) & 0xffffffff
    // CFG_ADDRESS_HIGH |= ((Ot + Op) >> 32) << 16
-   ld b32 $r7 D[$r5 + #ctx_src_address_low]
-   ld b32 $r8 D[$r5 + #ctx_src_address_high]
+   ld b32 $r7 D[$r5 + ctx_src_address_low]
+   ld b32 $r8 D[$r5 + ctx_src_address_high]
    add b32 $r10 $r12
    add b32 $r7 $r10
    adc b32 $r8 0
@@ -679,14 +677,14 @@ cmd_exec_set_surface_linear:
    xbit $r6 $flags $p2
    add b32 $r6 0x202
    shl b32 $r6 8
-   ld b32 $r7 D[$r5 + #ctx_src_address_low]
+   ld b32 $r7 D[$r5 + ctx_src_address_low]
    iowr I[$r6 + 0x000] $r7
    add b32 $r6 0x400
-   ld b32 $r7 D[$r5 + #ctx_src_address_high]
+   ld b32 $r7 D[$r5 + ctx_src_address_high]
    shl b32 $r7 16
    iowr I[$r6 + 0x000] $r7
    add b32 $r6 0x400
-   ld b32 $r7 D[$r5 + #ctx_src_pitch]
+   ld b32 $r7 D[$r5 + ctx_src_pitch]
    iowr I[$r6 + 0x000] $r7
    ret
 
@@ -699,7 +697,7 @@ cmd_exec_wait:
    loop:
       iord $r1 I[$r0]
       and $r1 1
-      bra ne #loop
+      bra ne loop
    pop $r1
    pop $r0
    ret
@@ -707,18 +705,18 @@ cmd_exec_wait:
 cmd_exec_query:
    // if QUERY_SHORT not set, write out { -, 0, TIME_LO, TIME_HI }
    xbit $r4 $r3 13
-   bra ne #query_counter
-      call #cmd_exec_wait
+   bra ne query_counter
+      call cmd_exec_wait
       mov $r4 0x80c
       shl b32 $r4 6
-      ld b32 $r5 D[$r0 + #ctx_query_address_low]
+      ld b32 $r5 D[$r0 + ctx_query_address_low]
       add b32 $r5 4
       iowr I[$r4 + 0x000] $r5
       iowr I[$r4 + 0x100] $r0
       mov $r5 0xc
       iowr I[$r4 + 0x200] $r5
       add b32 $r4 0x400
-      ld b32 $r5 D[$r0 + #ctx_query_address_high]
+      ld b32 $r5 D[$r0 + ctx_query_address_high]
       shl b32 $r5 16
       iowr I[$r4 + 0x000] $r5
       add b32 $r4 0x500
@@ -743,16 +741,16 @@ cmd_exec_query:
 
    // write COUNTER
    query_counter:
-   call #cmd_exec_wait
+   call cmd_exec_wait
    mov $r4 0x80c
    shl b32 $r4 6
-   ld b32 $r5 D[$r0 + #ctx_query_address_low]
+   ld b32 $r5 D[$r0 + ctx_query_address_low]
    iowr I[$r4 + 0x000] $r5
    iowr I[$r4 + 0x100] $r0
    mov $r5 0x4
    iowr I[$r4 + 0x200] $r5
    add b32 $r4 0x400
-   ld b32 $r5 D[$r0 + #ctx_query_address_high]
+   ld b32 $r5 D[$r0 + ctx_query_address_high]
    shl b32 $r5 16
    iowr I[$r4 + 0x000] $r5
    add b32 $r4 0x500
@@ -761,7 +759,7 @@ cmd_exec_query:
    mov $r5 0x00001110
    sethi $r5 0x13120000
    iowr I[$r4 + 0x100] $r5
-   ld b32 $r5 D[$r0 + #ctx_query_counter]
+   ld b32 $r5 D[$r0 + ctx_query_counter]
    add b32 $r4 0x500
    iowr I[$r4 + 0x000] $r5
    mov $r5 0x00002601
@@ -789,22 +787,22 @@ cmd_exec_query:
 //       $r2: hostirq state
 //       $r3: data
 cmd_exec:
-   call #cmd_exec_wait
+   call cmd_exec_wait
 
    // if format requested, call function to calculate it, otherwise
    // fill in cpp/xcnt for both surfaces as if (cpp == 1)
    xbit $r15 $r3 0
-   bra e #cmd_exec_no_format
-      call #cmd_exec_set_format
+   bra e cmd_exec_no_format
+      call cmd_exec_set_format
       mov $r4 0x200
-      bra #cmd_exec_init_src_surface
+      bra cmd_exec_init_src_surface
    cmd_exec_no_format:
       mov $r6 0x810
       shl b32 $r6 6
       mov $r7 1
-      st b32 D[$r0 + #ctx_src_cpp] $r7
-      st b32 D[$r0 + #ctx_dst_cpp] $r7
-      ld b32 $r7 D[$r0 + #ctx_xcnt]
+      st b32 D[$r0 + ctx_src_cpp] $r7
+      st b32 D[$r0 + ctx_dst_cpp] $r7
+      ld b32 $r7 D[$r0 + ctx_xcnt]
       iowr I[$r6 + 0x000] $r7
       iowr I[$r6 + 0x100] $r7
       clear b32 $r4
@@ -813,28 +811,28 @@ cmd_exec:
    bclr $flags $p2
    clear b32 $r5
    xbit $r15 $r3 4
-   bra e #src_tiled
-      call #cmd_exec_set_surface_linear
-      bra #cmd_exec_init_dst_surface
+   bra e src_tiled
+      call cmd_exec_set_surface_linear
+      bra cmd_exec_init_dst_surface
    src_tiled:
-      call #cmd_exec_set_surface_tiled
+      call cmd_exec_set_surface_tiled
       bset $r4 7
 
    cmd_exec_init_dst_surface:
    bset $flags $p2
-   mov $r5 #ctx_dst_address_high - #ctx_src_address_high
+   mov $r5 ctx_dst_address_high - ctx_src_address_high
    xbit $r15 $r3 8
-   bra e #dst_tiled
-      call #cmd_exec_set_surface_linear
-      bra #cmd_exec_kick
+   bra e dst_tiled
+      call cmd_exec_set_surface_linear
+      bra cmd_exec_kick
    dst_tiled:
-      call #cmd_exec_set_surface_tiled
+      call cmd_exec_set_surface_tiled
       bset $r4 8
 
    cmd_exec_kick:
    mov $r5 0x800
    shl b32 $r5 6
-   ld b32 $r6 D[$r0 + #ctx_ycnt]
+   ld b32 $r6 D[$r0 + ctx_ycnt]
    iowr I[$r5 + 0x100] $r6
    mov $r6 0x0041
    // SRC_TARGET = 1, DST_TARGET = 2
@@ -844,8 +842,8 @@ cmd_exec:
 
    // if requested, queue up a QUERY write after the copy has completed
    xbit $r15 $r3 12
-   bra e #cmd_exec_done
-      call #cmd_exec_query
+   bra e cmd_exec_done
+      call cmd_exec_query
 
    cmd_exec_done:
    ret
diff --git a/drivers/gpu/drm/nouveau/nva3_copy.fuc.h b/drivers/gpu/drm/nouveau/nva3_copy.fuc.h
index 1f33fbd..2731de2 100644
--- a/drivers/gpu/drm/nouveau/nva3_copy.fuc.h
+++ b/drivers/gpu/drm/nouveau/nva3_copy.fuc.h
@@ -152,7 +152,7 @@ uint32_t nva3_pcopy_code[] = {
 	0xf10010fe,
 	0xf1040017,
 	0xf0fff327,
-	0x12d00023,
+	0x22d00023,
 	0x0c25f0c0,
 	0xf40012d0,
 	0x17f11031,
diff --git a/drivers/gpu/drm/nouveau/nva3_pm.c b/drivers/gpu/drm/nouveau/nva3_pm.c
index 9e636e6..e4b2b9e 100644
--- a/drivers/gpu/drm/nouveau/nva3_pm.c
+++ b/drivers/gpu/drm/nouveau/nva3_pm.c
@@ -27,320 +27,178 @@
 #include "nouveau_bios.h"
 #include "nouveau_pm.h"
 
-static u32 read_clk(struct drm_device *, int, bool);
-static u32 read_pll(struct drm_device *, int, u32);
-
-static u32
-read_vco(struct drm_device *dev, int clk)
-{
-	u32 sctl = nv_rd32(dev, 0x4120 + (clk * 4));
-	if ((sctl & 0x00000030) != 0x00000030)
-		return read_pll(dev, 0x41, 0x00e820);
-	return read_pll(dev, 0x42, 0x00e8a0);
-}
-
-static u32
-read_clk(struct drm_device *dev, int clk, bool ignore_en)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	u32 sctl, sdiv, sclk;
-
-	/* refclk for the 0xe8xx plls is a fixed frequency */
-	if (clk >= 0x40) {
-		if (dev_priv->chipset == 0xaf) {
-			/* no joke.. seriously.. sigh.. */
-			return nv_rd32(dev, 0x00471c) * 1000;
-		}
-
-		return dev_priv->crystal;
-	}
-
-	sctl = nv_rd32(dev, 0x4120 + (clk * 4));
-	if (!ignore_en && !(sctl & 0x00000100))
-		return 0;
-
-	switch (sctl & 0x00003000) {
-	case 0x00000000:
-		return dev_priv->crystal;
-	case 0x00002000:
-		if (sctl & 0x00000040)
-			return 108000;
-		return 100000;
-	case 0x00003000:
-		sclk = read_vco(dev, clk);
-		sdiv = ((sctl & 0x003f0000) >> 16) + 2;
-		return (sclk * 2) / sdiv;
-	default:
-		return 0;
-	}
-}
-
-static u32
-read_pll(struct drm_device *dev, int clk, u32 pll)
-{
-	u32 ctrl = nv_rd32(dev, pll + 0);
-	u32 sclk = 0, P = 1, N = 1, M = 1;
-
-	if (!(ctrl & 0x00000008)) {
-		if (ctrl & 0x00000001) {
-			u32 coef = nv_rd32(dev, pll + 4);
-			M = (coef & 0x000000ff) >> 0;
-			N = (coef & 0x0000ff00) >> 8;
-			P = (coef & 0x003f0000) >> 16;
-
-			/* no post-divider on these.. */
-			if ((pll & 0x00ff00) == 0x00e800)
-				P = 1;
-
-			sclk = read_clk(dev, 0x00 + clk, false);
-		}
-	} else {
-		sclk = read_clk(dev, 0x10 + clk, false);
-	}
-
-	return sclk * N / (M * P);
-}
+/* This is actually a lot more complex than it appears here, but hopefully
+ * this should be able to deal with what the VBIOS leaves for us..
+ *
+ * If not, well, I'll jump off that bridge when I come to it.
+ */
 
-struct creg {
-	u32 clk;
-	u32 pll;
+struct nva3_pm_state {
+	enum pll_types type;
+	u32 src0;
+	u32 src1;
+	u32 ctrl;
+	u32 coef;
+	u32 old_pnm;
+	u32 new_pnm;
+	u32 new_div;
 };
 
 static int
-calc_clk(struct drm_device *dev, int clk, u32 pll, u32 khz, struct creg *reg)
+nva3_pm_pll_offset(u32 id)
 {
-	struct pll_lims limits;
-	u32 oclk, sclk, sdiv;
-	int P, N, M, diff;
-	int ret;
-
-	reg->pll = 0;
-	reg->clk = 0;
-	if (!khz) {
-		NV_DEBUG(dev, "no clock for 0x%04x/0x%02x\n", pll, clk);
-		return 0;
+	static const u32 pll_map[] = {
+		0x00, PLL_CORE,
+		0x01, PLL_SHADER,
+		0x02, PLL_MEMORY,
+		0x00, 0x00
+	};
+	const u32 *map = pll_map;
+
+	while (map[1]) {
+		if (id == map[1])
+			return map[0];
+		map += 2;
 	}
 
-	switch (khz) {
-	case 27000:
-		reg->clk = 0x00000100;
-		return khz;
-	case 100000:
-		reg->clk = 0x00002100;
-		return khz;
-	case 108000:
-		reg->clk = 0x00002140;
-		return khz;
-	default:
-		sclk = read_vco(dev, clk);
-		sdiv = min((sclk * 2) / (khz - 2999), (u32)65);
-		/* if the clock has a PLL attached, and we can get a within
-		 * [-2, 3) MHz of a divider, we'll disable the PLL and use
-		 * the divider instead.
-		 *
-		 * divider can go as low as 2, limited here because NVIDIA
-		 * and the VBIOS on my NVA8 seem to prefer using the PLL
-		 * for 810MHz - is there a good reason?
-		 */
-		if (sdiv > 4) {
-			oclk = (sclk * 2) / sdiv;
-			diff = khz - oclk;
-			if (!pll || (diff >= -2000 && diff < 3000)) {
-				reg->clk = (((sdiv - 2) << 16) | 0x00003100);
-				return oclk;
-			}
-		}
-
-		if (!pll) {
-			NV_ERROR(dev, "bad freq %02x: %d %d\n", clk, khz, sclk);
-			return -ERANGE;
-		}
+	return -ENOENT;
+}
 
-		break;
-	}
+int
+nva3_pm_clock_get(struct drm_device *dev, u32 id)
+{
+	u32 src0, src1, ctrl, coef;
+	struct pll_lims pll;
+	int ret, off;
+	int P, N, M;
 
-	ret = get_pll_limits(dev, pll, &limits);
+	ret = get_pll_limits(dev, id, &pll);
 	if (ret)
 		return ret;
 
-	limits.refclk = read_clk(dev, clk - 0x10, true);
-	if (!limits.refclk)
-		return -EINVAL;
-
-	ret = nva3_calc_pll(dev, &limits, khz, &N, NULL, &M, &P);
-	if (ret >= 0) {
-		reg->clk = nv_rd32(dev, 0x4120 + (clk * 4));
-		reg->pll = (P << 16) | (N << 8) | M;
-	}
-	return ret;
-}
-
-static void
-prog_pll(struct drm_device *dev, int clk, u32 pll, struct creg *reg)
-{
-	const u32 src0 = 0x004120 + (clk * 4);
-	const u32 src1 = 0x004160 + (clk * 4);
-	const u32 ctrl = pll + 0;
-	const u32 coef = pll + 4;
-	u32 cntl;
-
-	if (!reg->clk && !reg->pll) {
-		NV_DEBUG(dev, "no clock for %02x\n", clk);
-		return;
-	}
+	off = nva3_pm_pll_offset(id);
+	if (off < 0)
+		return off;
 
-	cntl = nv_rd32(dev, ctrl) & 0xfffffff2;
-	if (reg->pll) {
-		nv_mask(dev, src0, 0x00000101, 0x00000101);
-		nv_wr32(dev, coef, reg->pll);
-		nv_wr32(dev, ctrl, cntl | 0x00000015);
-		nv_mask(dev, src1, 0x00000100, 0x00000000);
-		nv_mask(dev, src1, 0x00000001, 0x00000000);
-	} else {
-		nv_mask(dev, src1, 0x003f3141, 0x00000101 | reg->clk);
-		nv_wr32(dev, ctrl, cntl | 0x0000001d);
-		nv_mask(dev, ctrl, 0x00000001, 0x00000000);
-		nv_mask(dev, src0, 0x00000100, 0x00000000);
-		nv_mask(dev, src0, 0x00000001, 0x00000000);
-	}
-}
+	src0 = nv_rd32(dev, 0x4120 + (off * 4));
+	src1 = nv_rd32(dev, 0x4160 + (off * 4));
+	ctrl = nv_rd32(dev, pll.reg + 0);
+	coef = nv_rd32(dev, pll.reg + 4);
+	NV_DEBUG(dev, "PLL %02x: 0x%08x 0x%08x 0x%08x 0x%08x\n",
+		      id, src0, src1, ctrl, coef);
 
-static void
-prog_clk(struct drm_device *dev, int clk, struct creg *reg)
-{
-	if (!reg->clk) {
-		NV_DEBUG(dev, "no clock for %02x\n", clk);
-		return;
+	if (ctrl & 0x00000008) {
+		u32 div = ((src1 & 0x003c0000) >> 18) + 1;
+		return (pll.refclk * 2) / div;
 	}
 
-	nv_mask(dev, 0x004120 + (clk * 4), 0x003f3141, 0x00000101 | reg->clk);
+	P = (coef & 0x003f0000) >> 16;
+	N = (coef & 0x0000ff00) >> 8;
+	M = (coef & 0x000000ff);
+	return pll.refclk * N / M / P;
 }
 
-int
-nva3_pm_clocks_get(struct drm_device *dev, struct nouveau_pm_level *perflvl)
-{
-	perflvl->core   = read_pll(dev, 0x00, 0x4200);
-	perflvl->shader = read_pll(dev, 0x01, 0x4220);
-	perflvl->memory = read_pll(dev, 0x02, 0x4000);
-	perflvl->unka0  = read_clk(dev, 0x20, false);
-	perflvl->vdec   = read_clk(dev, 0x21, false);
-	perflvl->daemon = read_clk(dev, 0x25, false);
-	perflvl->copy   = perflvl->core;
-	return 0;
-}
-
-struct nva3_pm_state {
-	struct creg nclk;
-	struct creg sclk;
-	struct creg mclk;
-	struct creg vdec;
-	struct creg unka0;
-};
-
 void *
-nva3_pm_clocks_pre(struct drm_device *dev, struct nouveau_pm_level *perflvl)
+nva3_pm_clock_pre(struct drm_device *dev, struct nouveau_pm_level *perflvl,
+		  u32 id, int khz)
 {
-	struct nva3_pm_state *info;
-	int ret;
-
-	info = kzalloc(sizeof(*info), GFP_KERNEL);
-	if (!info)
-		return ERR_PTR(-ENOMEM);
+	struct nva3_pm_state *pll;
+	struct pll_lims limits;
+	int N, M, P, diff;
+	int ret, off;
 
-	ret = calc_clk(dev, 0x10, 0x4200, perflvl->core, &info->nclk);
+	ret = get_pll_limits(dev, id, &limits);
 	if (ret < 0)
-		goto out;
+		return (ret == -ENOENT) ? NULL : ERR_PTR(ret);
 
-	ret = calc_clk(dev, 0x11, 0x4220, perflvl->shader, &info->sclk);
-	if (ret < 0)
-		goto out;
+	off = nva3_pm_pll_offset(id);
+	if (id < 0)
+		return ERR_PTR(-EINVAL);
 
-	ret = calc_clk(dev, 0x12, 0x4000, perflvl->memory, &info->mclk);
-	if (ret < 0)
-		goto out;
 
-	ret = calc_clk(dev, 0x20, 0x0000, perflvl->unka0, &info->unka0);
-	if (ret < 0)
-		goto out;
+	pll = kzalloc(sizeof(*pll), GFP_KERNEL);
+	if (!pll)
+		return ERR_PTR(-ENOMEM);
+	pll->type = id;
+	pll->src0 = 0x004120 + (off * 4);
+	pll->src1 = 0x004160 + (off * 4);
+	pll->ctrl = limits.reg + 0;
+	pll->coef = limits.reg + 4;
+
+	/* If target clock is within [-2, 3) MHz of a divisor, we'll
+	 * use that instead of calculating MNP values
+	 */
+	pll->new_div = min((limits.refclk * 2) / (khz - 2999), 16);
+	if (pll->new_div) {
+		diff = khz - ((limits.refclk * 2) / pll->new_div);
+		if (diff < -2000 || diff >= 3000)
+			pll->new_div = 0;
+	}
 
-	ret = calc_clk(dev, 0x21, 0x0000, perflvl->vdec, &info->vdec);
-	if (ret < 0)
-		goto out;
+	if (!pll->new_div) {
+		ret = nva3_calc_pll(dev, &limits, khz, &N, NULL, &M, &P);
+		if (ret < 0)
+			return ERR_PTR(ret);
 
-out:
-	if (ret < 0) {
-		kfree(info);
-		info = ERR_PTR(ret);
+		pll->new_pnm = (P << 16) | (N << 8) | M;
+		pll->new_div = 2 - 1;
+	} else {
+		pll->new_pnm = 0;
+		pll->new_div--;
 	}
-	return info;
-}
 
-static bool
-nva3_pm_grcp_idle(void *data)
-{
-	struct drm_device *dev = data;
-
-	if (!(nv_rd32(dev, 0x400304) & 0x00000001))
-		return true;
-	if (nv_rd32(dev, 0x400308) == 0x0050001c)
-		return true;
-	return false;
+	if ((nv_rd32(dev, pll->src1) & 0x00000101) != 0x00000101)
+		pll->old_pnm = nv_rd32(dev, pll->coef);
+	return pll;
 }
 
-int
-nva3_pm_clocks_set(struct drm_device *dev, void *pre_state)
+void
+nva3_pm_clock_set(struct drm_device *dev, void *pre_state)
 {
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nva3_pm_state *info = pre_state;
-	unsigned long flags;
-	int ret = -EAGAIN;
+	struct nva3_pm_state *pll = pre_state;
+	u32 ctrl = 0;
 
-	/* prevent any new grctx switches from starting */
-	spin_lock_irqsave(&dev_priv->context_switch_lock, flags);
-	nv_wr32(dev, 0x400324, 0x00000000);
-	nv_wr32(dev, 0x400328, 0x0050001c); /* wait flag 0x1c */
-	/* wait for any pending grctx switches to complete */
-	if (!nv_wait_cb(dev, nva3_pm_grcp_idle, dev)) {
-		NV_ERROR(dev, "pm: ctxprog didn't go idle\n");
-		goto cleanup;
+	/* For the memory clock, NVIDIA will build a "script" describing
+	 * the reclocking process and ask PDAEMON to execute it.
+	 */
+	if (pll->type == PLL_MEMORY) {
+		nv_wr32(dev, 0x100210, 0);
+		nv_wr32(dev, 0x1002dc, 1);
+		nv_wr32(dev, 0x004018, 0x00001000);
+		ctrl = 0x18000100;
 	}
-	/* freeze PFIFO */
-	nv_mask(dev, 0x002504, 0x00000001, 0x00000001);
-	if (!nv_wait(dev, 0x002504, 0x00000010, 0x00000010)) {
-		NV_ERROR(dev, "pm: fifo didn't go idle\n");
-		goto cleanup;
+
+	if (pll->old_pnm || !pll->new_pnm) {
+		nv_mask(dev, pll->src1, 0x003c0101, 0x00000101 |
+						    (pll->new_div << 18));
+		nv_wr32(dev, pll->ctrl, 0x0001001d | ctrl);
+		nv_mask(dev, pll->ctrl, 0x00000001, 0x00000000);
 	}
 
-	prog_pll(dev, 0x00, 0x004200, &info->nclk);
-	prog_pll(dev, 0x01, 0x004220, &info->sclk);
-	prog_clk(dev, 0x20, &info->unka0);
-	prog_clk(dev, 0x21, &info->vdec);
+	if (pll->new_pnm) {
+		nv_mask(dev, pll->src0, 0x00000101, 0x00000101);
+		nv_wr32(dev, pll->coef, pll->new_pnm);
+		nv_wr32(dev, pll->ctrl, 0x0001001d | ctrl);
+		nv_mask(dev, pll->ctrl, 0x00000010, 0x00000000);
+		nv_mask(dev, pll->ctrl, 0x00020010, 0x00020010);
+		nv_wr32(dev, pll->ctrl, 0x00010015 | ctrl);
+		nv_mask(dev, pll->src1, 0x00000100, 0x00000000);
+		nv_mask(dev, pll->src1, 0x00000001, 0x00000000);
+		if (pll->type == PLL_MEMORY)
+			nv_wr32(dev, 0x4018, 0x10005000);
+	} else {
+		nv_mask(dev, pll->ctrl, 0x00000001, 0x00000000);
+		nv_mask(dev, pll->src0, 0x00000100, 0x00000000);
+		nv_mask(dev, pll->src0, 0x00000001, 0x00000000);
+		if (pll->type == PLL_MEMORY)
+			nv_wr32(dev, 0x4018, 0x1000d000);
+	}
 
-	if (info->mclk.clk || info->mclk.pll) {
-		nv_wr32(dev, 0x100210, 0);
-		nv_wr32(dev, 0x1002dc, 1);
-		nv_wr32(dev, 0x004018, 0x00001000);
-		prog_pll(dev, 0x02, 0x004000, &info->mclk);
-		if (nv_rd32(dev, 0x4000) & 0x00000008)
-			nv_wr32(dev, 0x004018, 0x1000d000);
-		else
-			nv_wr32(dev, 0x004018, 0x10005000);
+	if (pll->type == PLL_MEMORY) {
 		nv_wr32(dev, 0x1002dc, 0);
 		nv_wr32(dev, 0x100210, 0x80000000);
 	}
 
-	ret = 0;
-
-cleanup:
-	/* unfreeze PFIFO */
-	nv_mask(dev, 0x002504, 0x00000001, 0x00000000);
-	/* restore ctxprog to normal */
-	nv_wr32(dev, 0x400324, 0x00000000);
-	nv_wr32(dev, 0x400328, 0x0070009c); /* set flag 0x1c */
-	/* unblock it if necessary */
-	if (nv_rd32(dev, 0x400308) == 0x0050001c)
-		nv_mask(dev, 0x400824, 0x10000000, 0x10000000);
-	spin_unlock_irqrestore(&dev_priv->context_switch_lock, flags);
-	kfree(info);
-	return ret;
+	kfree(pll);
 }
+
diff --git a/drivers/gpu/drm/nouveau/nvc0_copy.c b/drivers/gpu/drm/nouveau/nvc0_copy.c
index dddf006..208fa7a 100644
--- a/drivers/gpu/drm/nouveau/nvc0_copy.c
+++ b/drivers/gpu/drm/nouveau/nvc0_copy.c
@@ -48,14 +48,14 @@ nvc0_copy_context_new(struct nouveau_channel *chan, int engine)
 	struct nouveau_gpuobj *ctx = NULL;
 	int ret;
 
-	ret = nouveau_gpuobj_new(dev, chan, 256, 256,
+	ret = nouveau_gpuobj_new(dev, NULL, 256, 256,
 				 NVOBJ_FLAG_VM | NVOBJ_FLAG_VM_USER |
 				 NVOBJ_FLAG_ZERO_ALLOC, &ctx);
 	if (ret)
 		return ret;
 
-	nv_wo32(ramin, pcopy->ctx + 0, lower_32_bits(ctx->linst));
-	nv_wo32(ramin, pcopy->ctx + 4, upper_32_bits(ctx->linst));
+	nv_wo32(ramin, pcopy->ctx + 0, lower_32_bits(ctx->vinst));
+	nv_wo32(ramin, pcopy->ctx + 4, upper_32_bits(ctx->vinst));
 	dev_priv->engine.instmem.flush(dev);
 
 	chan->engctx[engine] = ctx;
@@ -127,7 +127,7 @@ nvc0_copy_init(struct drm_device *dev, int engine)
 }
 
 static int
-nvc0_copy_fini(struct drm_device *dev, int engine, bool suspend)
+nvc0_copy_fini(struct drm_device *dev, int engine)
 {
 	struct nvc0_copy_engine *pcopy = nv_engine(dev, engine);
 
diff --git a/drivers/gpu/drm/nouveau/nvc0_copy.fuc.h b/drivers/gpu/drm/nouveau/nvc0_copy.fuc.h
index a8d1745..4199038 100644
--- a/drivers/gpu/drm/nouveau/nvc0_copy.fuc.h
+++ b/drivers/gpu/drm/nouveau/nvc0_copy.fuc.h
@@ -145,7 +145,7 @@ uint32_t nvc0_pcopy_code[] = {
 	0xf10010fe,
 	0xf1040017,
 	0xf0fff327,
-	0x12d00023,
+	0x22d00023,
 	0x0c25f0c0,
 	0xf40012d0,
 	0x17f11031,
diff --git a/drivers/gpu/drm/nouveau/nvc0_fb.c b/drivers/gpu/drm/nouveau/nvc0_fb.c
index 5bf5503..26a9960 100644
--- a/drivers/gpu/drm/nouveau/nvc0_fb.c
+++ b/drivers/gpu/drm/nouveau/nvc0_fb.c
@@ -1,5 +1,5 @@
 /*
- * Copyright 2011 Red Hat Inc.
+ * Copyright 2010 Red Hat Inc.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -23,107 +23,16 @@
  */
 
 #include "drmP.h"
-#include "drm.h"
-#include "nouveau_drv.h"
-#include "nouveau_drm.h"
-
-struct nvc0_fb_priv {
-	struct page *r100c10_page;
-	dma_addr_t r100c10;
-};
-
-static inline void
-nvc0_mfb_subp_isr(struct drm_device *dev, int unit, int subp)
-{
-	u32 subp_base = 0x141000 + (unit * 0x2000) + (subp * 0x400);
-	u32 stat = nv_rd32(dev, subp_base + 0x020);
-
-	if (stat) {
-		NV_INFO(dev, "PMFB%d_SUBP%d: 0x%08x\n", unit, subp, stat);
-		nv_wr32(dev, subp_base + 0x020, stat);
-	}
-}
-
-static void
-nvc0_mfb_isr(struct drm_device *dev)
-{
-	u32 units = nv_rd32(dev, 0x00017c);
-	while (units) {
-		u32 subp, unit = ffs(units) - 1;
-		for (subp = 0; subp < 2; subp++)
-			nvc0_mfb_subp_isr(dev, unit, subp);
-		units &= ~(1 << unit);
-	}
-}
-
-static void
-nvc0_fb_destroy(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_fb_engine *pfb = &dev_priv->engine.fb;
-	struct nvc0_fb_priv *priv = pfb->priv;
-
-	nouveau_irq_unregister(dev, 25);
-
-	if (priv->r100c10_page) {
-		pci_unmap_page(dev->pdev, priv->r100c10, PAGE_SIZE,
-			       PCI_DMA_BIDIRECTIONAL);
-		__free_page(priv->r100c10_page);
-	}
 
-	kfree(priv);
-	pfb->priv = NULL;
-}
-
-static int
-nvc0_fb_create(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_fb_engine *pfb = &dev_priv->engine.fb;
-	struct nvc0_fb_priv *priv;
-
-	priv = kzalloc(sizeof(*priv), GFP_KERNEL);
-	if (!priv)
-		return -ENOMEM;
-	pfb->priv = priv;
-
-	priv->r100c10_page = alloc_page(GFP_KERNEL | __GFP_ZERO);
-	if (!priv->r100c10_page) {
-		nvc0_fb_destroy(dev);
-		return -ENOMEM;
-	}
-
-	priv->r100c10 = pci_map_page(dev->pdev, priv->r100c10_page, 0,
-				     PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
-	if (pci_dma_mapping_error(dev->pdev, priv->r100c10)) {
-		nvc0_fb_destroy(dev);
-		return -EFAULT;
-	}
-
-	nouveau_irq_register(dev, 25, nvc0_mfb_isr);
-	return 0;
-}
+#include "nouveau_drv.h"
 
 int
 nvc0_fb_init(struct drm_device *dev)
 {
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nvc0_fb_priv *priv;
-	int ret;
-
-	if (!dev_priv->engine.fb.priv) {
-		ret = nvc0_fb_create(dev);
-		if (ret)
-			return ret;
-	}
-	priv = dev_priv->engine.fb.priv;
-
-	nv_wr32(dev, 0x100c10, priv->r100c10 >> 8);
 	return 0;
 }
 
 void
 nvc0_fb_takedown(struct drm_device *dev)
 {
-	nvc0_fb_destroy(dev);
 }
diff --git a/drivers/gpu/drm/nouveau/nvc0_fbcon.c b/drivers/gpu/drm/nouveau/nvc0_fbcon.c
index a495e48..fa5d4c2 100644
--- a/drivers/gpu/drm/nouveau/nvc0_fbcon.c
+++ b/drivers/gpu/drm/nouveau/nvc0_fbcon.c
@@ -159,7 +159,7 @@ nvc0_fbcon_accel_init(struct fb_info *info)
 	struct drm_device *dev = nfbdev->dev;
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nouveau_channel *chan = dev_priv->channel;
-	struct nouveau_framebuffer *fb = &nfbdev->nouveau_fb;
+	struct nouveau_bo *nvbo = nfbdev->nouveau_fb.nvbo;
 	int ret, format;
 
 	ret = nouveau_gpuobj_gr_new(chan, 0x902d, 0x902d);
@@ -203,8 +203,8 @@ nvc0_fbcon_accel_init(struct fb_info *info)
 	BEGIN_NVC0(chan, 2, NvSub2D, 0x0000, 1);
 	OUT_RING  (chan, 0x0000902d);
 	BEGIN_NVC0(chan, 2, NvSub2D, 0x0104, 2);
-	OUT_RING  (chan, upper_32_bits(chan->notifier_vma.offset));
-	OUT_RING  (chan, lower_32_bits(chan->notifier_vma.offset));
+	OUT_RING  (chan, upper_32_bits(chan->notifier_bo->bo.offset));
+	OUT_RING  (chan, lower_32_bits(chan->notifier_bo->bo.offset));
 	BEGIN_NVC0(chan, 2, NvSub2D, 0x0290, 1);
 	OUT_RING  (chan, 0);
 	BEGIN_NVC0(chan, 2, NvSub2D, 0x0888, 1);
@@ -249,8 +249,8 @@ nvc0_fbcon_accel_init(struct fb_info *info)
 	OUT_RING  (chan, info->fix.line_length);
 	OUT_RING  (chan, info->var.xres_virtual);
 	OUT_RING  (chan, info->var.yres_virtual);
-	OUT_RING  (chan, upper_32_bits(fb->vma.offset));
-	OUT_RING  (chan, lower_32_bits(fb->vma.offset));
+	OUT_RING  (chan, upper_32_bits(nvbo->vma.offset));
+	OUT_RING  (chan, lower_32_bits(nvbo->vma.offset));
 	BEGIN_NVC0(chan, 2, NvSub2D, 0x0230, 10);
 	OUT_RING  (chan, format);
 	OUT_RING  (chan, 1);
@@ -260,8 +260,8 @@ nvc0_fbcon_accel_init(struct fb_info *info)
 	OUT_RING  (chan, info->fix.line_length);
 	OUT_RING  (chan, info->var.xres_virtual);
 	OUT_RING  (chan, info->var.yres_virtual);
-	OUT_RING  (chan, upper_32_bits(fb->vma.offset));
-	OUT_RING  (chan, lower_32_bits(fb->vma.offset));
+	OUT_RING  (chan, upper_32_bits(nvbo->vma.offset));
+	OUT_RING  (chan, lower_32_bits(nvbo->vma.offset));
 	FIRE_RING (chan);
 
 	return 0;
diff --git a/drivers/gpu/drm/nouveau/nvc0_fifo.c b/drivers/gpu/drm/nouveau/nvc0_fifo.c
index dcbe0d5..fb4f594 100644
--- a/drivers/gpu/drm/nouveau/nvc0_fifo.c
+++ b/drivers/gpu/drm/nouveau/nvc0_fifo.c
@@ -210,10 +210,10 @@ nvc0_fifo_unload_context(struct drm_device *dev)
 	int i;
 
 	for (i = 0; i < 128; i++) {
-		if (!(nv_rd32(dev, 0x003004 + (i * 8)) & 1))
+		if (!(nv_rd32(dev, 0x003004 + (i * 4)) & 1))
 			continue;
 
-		nv_mask(dev, 0x003004 + (i * 8), 0x00000001, 0x00000000);
+		nv_mask(dev, 0x003004 + (i * 4), 0x00000001, 0x00000000);
 		nv_wr32(dev, 0x002634, i);
 		if (!nv_wait(dev, 0x002634, 0xffffffff, i)) {
 			NV_INFO(dev, "PFIFO: kick ch %d failed: 0x%08x\n",
@@ -322,7 +322,7 @@ nvc0_fifo_init(struct drm_device *dev)
 	}
 
 	/* PSUBFIFO[n] */
-	for (i = 0; i < priv->spoon_nr; i++) {
+	for (i = 0; i < 3; i++) {
 		nv_mask(dev, 0x04013c + (i * 0x2000), 0x10000100, 0x00000000);
 		nv_wr32(dev, 0x040108 + (i * 0x2000), 0xffffffff); /* INTR */
 		nv_wr32(dev, 0x04010c + (i * 0x2000), 0xfffffeff); /* INTR_EN */
diff --git a/drivers/gpu/drm/nouveau/nvc0_graph.c b/drivers/gpu/drm/nouveau/nvc0_graph.c
index 8ee3963..ca6db20 100644
--- a/drivers/gpu/drm/nouveau/nvc0_graph.c
+++ b/drivers/gpu/drm/nouveau/nvc0_graph.c
@@ -23,40 +23,12 @@
  */
 
 #include <linux/firmware.h>
-#include <linux/module.h>
 
 #include "drmP.h"
 
 #include "nouveau_drv.h"
 #include "nouveau_mm.h"
-
 #include "nvc0_graph.h"
-#include "nvc0_grhub.fuc.h"
-#include "nvc0_grgpc.fuc.h"
-
-static void
-nvc0_graph_ctxctl_debug_unit(struct drm_device *dev, u32 base)
-{
-	NV_INFO(dev, "PGRAPH: %06x - done 0x%08x\n", base,
-		nv_rd32(dev, base + 0x400));
-	NV_INFO(dev, "PGRAPH: %06x - stat 0x%08x 0x%08x 0x%08x 0x%08x\n", base,
-		nv_rd32(dev, base + 0x800), nv_rd32(dev, base + 0x804),
-		nv_rd32(dev, base + 0x808), nv_rd32(dev, base + 0x80c));
-	NV_INFO(dev, "PGRAPH: %06x - stat 0x%08x 0x%08x 0x%08x 0x%08x\n", base,
-		nv_rd32(dev, base + 0x810), nv_rd32(dev, base + 0x814),
-		nv_rd32(dev, base + 0x818), nv_rd32(dev, base + 0x81c));
-}
-
-static void
-nvc0_graph_ctxctl_debug(struct drm_device *dev)
-{
-	u32 gpcnr = nv_rd32(dev, 0x409604) & 0xffff;
-	u32 gpc;
-
-	nvc0_graph_ctxctl_debug_unit(dev, 0x409000);
-	for (gpc = 0; gpc < gpcnr; gpc++)
-		nvc0_graph_ctxctl_debug_unit(dev, 0x502000 + (gpc * 0x8000));
-}
 
 static int
 nvc0_graph_load_context(struct nouveau_channel *chan)
@@ -100,44 +72,24 @@ nvc0_graph_construct_context(struct nouveau_channel *chan)
 	if (!ctx)
 		return -ENOMEM;
 
-	if (!nouveau_ctxfw) {
-		nv_wr32(dev, 0x409840, 0x80000000);
-		nv_wr32(dev, 0x409500, 0x80000000 | chan->ramin->vinst >> 12);
-		nv_wr32(dev, 0x409504, 0x00000001);
-		if (!nv_wait(dev, 0x409800, 0x80000000, 0x80000000)) {
-			NV_ERROR(dev, "PGRAPH: HUB_SET_CHAN timeout\n");
-			nvc0_graph_ctxctl_debug(dev);
-			ret = -EBUSY;
-			goto err;
-		}
-	} else {
-		nvc0_graph_load_context(chan);
-
-		nv_wo32(grch->grctx, 0x1c, 1);
-		nv_wo32(grch->grctx, 0x20, 0);
-		nv_wo32(grch->grctx, 0x28, 0);
-		nv_wo32(grch->grctx, 0x2c, 0);
-		dev_priv->engine.instmem.flush(dev);
-	}
+	nvc0_graph_load_context(chan);
+
+	nv_wo32(grch->grctx, 0x1c, 1);
+	nv_wo32(grch->grctx, 0x20, 0);
+	nv_wo32(grch->grctx, 0x28, 0);
+	nv_wo32(grch->grctx, 0x2c, 0);
+	dev_priv->engine.instmem.flush(dev);
 
 	ret = nvc0_grctx_generate(chan);
-	if (ret)
-		goto err;
-
-	if (!nouveau_ctxfw) {
-		nv_wr32(dev, 0x409840, 0x80000000);
-		nv_wr32(dev, 0x409500, 0x80000000 | chan->ramin->vinst >> 12);
-		nv_wr32(dev, 0x409504, 0x00000002);
-		if (!nv_wait(dev, 0x409800, 0x80000000, 0x80000000)) {
-			NV_ERROR(dev, "PGRAPH: HUB_CTX_SAVE timeout\n");
-			nvc0_graph_ctxctl_debug(dev);
-			ret = -EBUSY;
-			goto err;
-		}
-	} else {
-		ret = nvc0_graph_unload_context_to(dev, chan->ramin->vinst);
-		if (ret)
-			goto err;
+	if (ret) {
+		kfree(ctx);
+		return ret;
+	}
+
+	ret = nvc0_graph_unload_context_to(dev, chan->ramin->vinst);
+	if (ret) {
+		kfree(ctx);
+		return ret;
 	}
 
 	for (i = 0; i < priv->grctx_size; i += 4)
@@ -145,10 +97,6 @@ nvc0_graph_construct_context(struct nouveau_channel *chan)
 
 	priv->grctx_vals = ctx;
 	return 0;
-
-err:
-	kfree(ctx);
-	return ret;
 }
 
 static int
@@ -157,87 +105,64 @@ nvc0_graph_create_context_mmio_list(struct nouveau_channel *chan)
 	struct nvc0_graph_priv *priv = nv_engine(chan->dev, NVOBJ_ENGINE_GR);
 	struct nvc0_graph_chan *grch = chan->engctx[NVOBJ_ENGINE_GR];
 	struct drm_device *dev = chan->dev;
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	int i = 0, gpc, tp, ret;
+	u32 magic;
 
-	ret = nouveau_gpuobj_new(dev, chan, 0x2000, 256, NVOBJ_FLAG_VM,
+	ret = nouveau_gpuobj_new(dev, NULL, 0x2000, 256, NVOBJ_FLAG_VM,
 				 &grch->unk408004);
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(dev, chan, 0x8000, 256, NVOBJ_FLAG_VM,
+	ret = nouveau_gpuobj_new(dev, NULL, 0x8000, 256, NVOBJ_FLAG_VM,
 				 &grch->unk40800c);
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(dev, chan, 384 * 1024, 4096,
+	ret = nouveau_gpuobj_new(dev, NULL, 384 * 1024, 4096,
 				 NVOBJ_FLAG_VM | NVOBJ_FLAG_VM_USER,
 				 &grch->unk418810);
 	if (ret)
 		return ret;
 
-	ret = nouveau_gpuobj_new(dev, chan, 0x1000, 0, NVOBJ_FLAG_VM,
+	ret = nouveau_gpuobj_new(dev, NULL, 0x1000, 0, NVOBJ_FLAG_VM,
 				 &grch->mmio);
 	if (ret)
 		return ret;
 
 
 	nv_wo32(grch->mmio, i++ * 4, 0x00408004);
-	nv_wo32(grch->mmio, i++ * 4, grch->unk408004->linst >> 8);
+	nv_wo32(grch->mmio, i++ * 4, grch->unk408004->vinst >> 8);
 	nv_wo32(grch->mmio, i++ * 4, 0x00408008);
 	nv_wo32(grch->mmio, i++ * 4, 0x80000018);
 
 	nv_wo32(grch->mmio, i++ * 4, 0x0040800c);
-	nv_wo32(grch->mmio, i++ * 4, grch->unk40800c->linst >> 8);
+	nv_wo32(grch->mmio, i++ * 4, grch->unk40800c->vinst >> 8);
 	nv_wo32(grch->mmio, i++ * 4, 0x00408010);
 	nv_wo32(grch->mmio, i++ * 4, 0x80000000);
 
 	nv_wo32(grch->mmio, i++ * 4, 0x00418810);
-	nv_wo32(grch->mmio, i++ * 4, 0x80000000 | grch->unk418810->linst >> 12);
+	nv_wo32(grch->mmio, i++ * 4, 0x80000000 | grch->unk418810->vinst >> 12);
 	nv_wo32(grch->mmio, i++ * 4, 0x00419848);
-	nv_wo32(grch->mmio, i++ * 4, 0x10000000 | grch->unk418810->linst >> 12);
+	nv_wo32(grch->mmio, i++ * 4, 0x10000000 | grch->unk418810->vinst >> 12);
 
 	nv_wo32(grch->mmio, i++ * 4, 0x00419004);
-	nv_wo32(grch->mmio, i++ * 4, grch->unk40800c->linst >> 8);
+	nv_wo32(grch->mmio, i++ * 4, grch->unk40800c->vinst >> 8);
 	nv_wo32(grch->mmio, i++ * 4, 0x00419008);
 	nv_wo32(grch->mmio, i++ * 4, 0x00000000);
 
 	nv_wo32(grch->mmio, i++ * 4, 0x00418808);
-	nv_wo32(grch->mmio, i++ * 4, grch->unk408004->linst >> 8);
+	nv_wo32(grch->mmio, i++ * 4, grch->unk408004->vinst >> 8);
 	nv_wo32(grch->mmio, i++ * 4, 0x0041880c);
 	nv_wo32(grch->mmio, i++ * 4, 0x80000018);
 
-	if (dev_priv->chipset != 0xc1) {
-		u32 magic = 0x02180000;
-		nv_wo32(grch->mmio, i++ * 4, 0x00405830);
-		nv_wo32(grch->mmio, i++ * 4, magic);
-		for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-			for (tp = 0; tp < priv->tp_nr[gpc]; tp++) {
-				u32 reg = TP_UNIT(gpc, tp, 0x520);
-				nv_wo32(grch->mmio, i++ * 4, reg);
-				nv_wo32(grch->mmio, i++ * 4, magic);
-				magic += 0x0324;
-			}
-		}
-	} else {
-		u32 magic = 0x02180000;
-		nv_wo32(grch->mmio, i++ * 4, 0x00405830);
-		nv_wo32(grch->mmio, i++ * 4, magic | 0x0000218);
-		nv_wo32(grch->mmio, i++ * 4, 0x004064c4);
-		nv_wo32(grch->mmio, i++ * 4, 0x0086ffff);
-		for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
-			for (tp = 0; tp < priv->tp_nr[gpc]; tp++) {
-				u32 reg = TP_UNIT(gpc, tp, 0x520);
-				nv_wo32(grch->mmio, i++ * 4, reg);
-				nv_wo32(grch->mmio, i++ * 4, (1 << 28) | magic);
-				magic += 0x0324;
-			}
-			for (tp = 0; tp < priv->tp_nr[gpc]; tp++) {
-				u32 reg = TP_UNIT(gpc, tp, 0x544);
-				nv_wo32(grch->mmio, i++ * 4, reg);
-				nv_wo32(grch->mmio, i++ * 4, magic);
-				magic += 0x0324;
-			}
+	magic = 0x02180000;
+	nv_wo32(grch->mmio, i++ * 4, 0x00405830);
+	nv_wo32(grch->mmio, i++ * 4, magic);
+	for (gpc = 0; gpc < priv->gpc_nr; gpc++) {
+		for (tp = 0; tp < priv->tp_nr[gpc]; tp++, magic += 0x02fc) {
+			u32 reg = 0x504520 + (gpc * 0x8000) + (tp * 0x0800);
+			nv_wo32(grch->mmio, i++ * 4, reg);
+			nv_wo32(grch->mmio, i++ * 4, magic);
 		}
 	}
 
@@ -261,7 +186,7 @@ nvc0_graph_context_new(struct nouveau_channel *chan, int engine)
 		return -ENOMEM;
 	chan->engctx[NVOBJ_ENGINE_GR] = grch;
 
-	ret = nouveau_gpuobj_new(dev, chan, priv->grctx_size, 256,
+	ret = nouveau_gpuobj_new(dev, NULL, priv->grctx_size, 256,
 				 NVOBJ_FLAG_VM | NVOBJ_FLAG_ZERO_ALLOC,
 				 &grch->grctx);
 	if (ret)
@@ -272,8 +197,8 @@ nvc0_graph_context_new(struct nouveau_channel *chan, int engine)
 	if (ret)
 		goto error;
 
-	nv_wo32(chan->ramin, 0x0210, lower_32_bits(grctx->linst) | 4);
-	nv_wo32(chan->ramin, 0x0214, upper_32_bits(grctx->linst));
+	nv_wo32(chan->ramin, 0x0210, lower_32_bits(grctx->vinst) | 4);
+	nv_wo32(chan->ramin, 0x0214, upper_32_bits(grctx->vinst));
 	pinstmem->flush(dev);
 
 	if (!priv->grctx_vals) {
@@ -285,20 +210,15 @@ nvc0_graph_context_new(struct nouveau_channel *chan, int engine)
 	for (i = 0; i < priv->grctx_size; i += 4)
 		nv_wo32(grctx, i, priv->grctx_vals[i / 4]);
 
-	if (!nouveau_ctxfw) {
-		nv_wo32(grctx, 0x00, grch->mmio_nr);
-		nv_wo32(grctx, 0x04, grch->mmio->linst >> 8);
-	} else {
-		nv_wo32(grctx, 0xf4, 0);
-		nv_wo32(grctx, 0xf8, 0);
-		nv_wo32(grctx, 0x10, grch->mmio_nr);
-		nv_wo32(grctx, 0x14, lower_32_bits(grch->mmio->linst));
-		nv_wo32(grctx, 0x18, upper_32_bits(grch->mmio->linst));
-		nv_wo32(grctx, 0x1c, 1);
-		nv_wo32(grctx, 0x20, 0);
-		nv_wo32(grctx, 0x28, 0);
-		nv_wo32(grctx, 0x2c, 0);
-	}
+	nv_wo32(grctx, 0xf4, 0);
+	nv_wo32(grctx, 0xf8, 0);
+	nv_wo32(grctx, 0x10, grch->mmio_nr);
+	nv_wo32(grctx, 0x14, lower_32_bits(grch->mmio->vinst));
+	nv_wo32(grctx, 0x18, upper_32_bits(grch->mmio->vinst));
+	nv_wo32(grctx, 0x1c, 1);
+	nv_wo32(grctx, 0x20, 0);
+	nv_wo32(grctx, 0x28, 0);
+	nv_wo32(grctx, 0x2c, 0);
 	pinstmem->flush(dev);
 	return 0;
 
@@ -328,7 +248,7 @@ nvc0_graph_object_new(struct nouveau_channel *chan, int engine,
 }
 
 static int
-nvc0_graph_fini(struct drm_device *dev, int engine, bool suspend)
+nvc0_graph_fini(struct drm_device *dev, int engine)
 {
 	return 0;
 }
@@ -376,13 +296,10 @@ static void
 nvc0_graph_init_gpc_0(struct drm_device *dev)
 {
 	struct nvc0_graph_priv *priv = nv_engine(dev, NVOBJ_ENGINE_GR);
-	const u32 magicgpc918 = DIV_ROUND_UP(0x00800000, priv->tp_total);
 	u32 data[TP_MAX / 8];
 	u8  tpnr[GPC_MAX];
 	int i, gpc, tpc;
 
-	nv_wr32(dev, TP_UNIT(0, 0, 0x5c), 1); /* affects TFB offset queries */
-
 	/*
 	 *      TP      ROP UNKVAL(magic_not_rop_nr)
 	 * 450: 4/0/0/0 2        3
@@ -390,6 +307,13 @@ nvc0_graph_init_gpc_0(struct drm_device *dev)
 	 * 465: 3/4/4/0 4        7
 	 * 470: 3/3/4/4 5        5
 	 * 480: 3/4/4/4 6        6
+	 *
+	 * magicgpc918
+	 * 450: 00200000 00000000001000000000000000000000
+	 * 460: 00124925 00000000000100100100100100100101
+	 * 465: 000ba2e9 00000000000010111010001011101001
+	 * 470: 00092493 00000000000010010010010010010011
+	 * 480: 00088889 00000000000010001000100010001001
 	 */
 
 	memset(data, 0x00, sizeof(data));
@@ -412,11 +336,11 @@ nvc0_graph_init_gpc_0(struct drm_device *dev)
 		nv_wr32(dev, GPC_UNIT(gpc, 0x0914), priv->magic_not_rop_nr << 8 |
 						  priv->tp_nr[gpc]);
 		nv_wr32(dev, GPC_UNIT(gpc, 0x0910), 0x00040000 | priv->tp_total);
-		nv_wr32(dev, GPC_UNIT(gpc, 0x0918), magicgpc918);
+		nv_wr32(dev, GPC_UNIT(gpc, 0x0918), priv->magicgpc918);
 	}
 
-	nv_wr32(dev, GPC_BCAST(0x1bd4), magicgpc918);
-	nv_wr32(dev, GPC_BCAST(0x08ac), nv_rd32(dev, 0x100800));
+	nv_wr32(dev, GPC_BCAST(0x1bd4), priv->magicgpc918);
+	nv_wr32(dev, GPC_BCAST(0x08ac), priv->rop_nr);
 }
 
 static void
@@ -495,51 +419,8 @@ nvc0_graph_init_fuc(struct drm_device *dev, u32 fuc_base,
 static int
 nvc0_graph_init_ctxctl(struct drm_device *dev)
 {
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nvc0_graph_priv *priv = nv_engine(dev, NVOBJ_ENGINE_GR);
 	u32 r000260;
-	int i;
-
-	if (!nouveau_ctxfw) {
-		/* load HUB microcode */
-		r000260 = nv_mask(dev, 0x000260, 0x00000001, 0x00000000);
-		nv_wr32(dev, 0x4091c0, 0x01000000);
-		for (i = 0; i < sizeof(nvc0_grhub_data) / 4; i++)
-			nv_wr32(dev, 0x4091c4, nvc0_grhub_data[i]);
-
-		nv_wr32(dev, 0x409180, 0x01000000);
-		for (i = 0; i < sizeof(nvc0_grhub_code) / 4; i++) {
-			if ((i & 0x3f) == 0)
-				nv_wr32(dev, 0x409188, i >> 6);
-			nv_wr32(dev, 0x409184, nvc0_grhub_code[i]);
-		}
-
-		/* load GPC microcode */
-		nv_wr32(dev, 0x41a1c0, 0x01000000);
-		for (i = 0; i < sizeof(nvc0_grgpc_data) / 4; i++)
-			nv_wr32(dev, 0x41a1c4, nvc0_grgpc_data[i]);
-
-		nv_wr32(dev, 0x41a180, 0x01000000);
-		for (i = 0; i < sizeof(nvc0_grgpc_code) / 4; i++) {
-			if ((i & 0x3f) == 0)
-				nv_wr32(dev, 0x41a188, i >> 6);
-			nv_wr32(dev, 0x41a184, nvc0_grgpc_code[i]);
-		}
-		nv_wr32(dev, 0x000260, r000260);
-
-		/* start HUB ucode running, it'll init the GPCs */
-		nv_wr32(dev, 0x409800, dev_priv->chipset);
-		nv_wr32(dev, 0x40910c, 0x00000000);
-		nv_wr32(dev, 0x409100, 0x00000002);
-		if (!nv_wait(dev, 0x409800, 0x80000000, 0x80000000)) {
-			NV_ERROR(dev, "PGRAPH: HUB_INIT timed out\n");
-			nvc0_graph_ctxctl_debug(dev);
-			return -EBUSY;
-		}
-
-		priv->grctx_size = nv_rd32(dev, 0x409804);
-		return 0;
-	}
 
 	/* load fuc microcode */
 	r000260 = nv_mask(dev, 0x000260, 0x00000001, 0x00000000);
@@ -647,22 +528,6 @@ nvc0_graph_isr_chid(struct drm_device *dev, u64 inst)
 }
 
 static void
-nvc0_graph_ctxctl_isr(struct drm_device *dev)
-{
-	u32 ustat = nv_rd32(dev, 0x409c18);
-
-	if (ustat & 0x00000001)
-		NV_INFO(dev, "PGRAPH: CTXCTRL ucode error\n");
-	if (ustat & 0x00080000)
-		NV_INFO(dev, "PGRAPH: CTXCTRL watchdog timeout\n");
-	if (ustat & ~0x00080001)
-		NV_INFO(dev, "PGRAPH: CTXCTRL 0x%08x\n", ustat);
-
-	nvc0_graph_ctxctl_debug(dev);
-	nv_wr32(dev, 0x409c20, ustat);
-}
-
-static void
 nvc0_graph_isr(struct drm_device *dev)
 {
 	u64 inst = (u64)(nv_rd32(dev, 0x409b00) & 0x0fffffff) << 12;
@@ -713,7 +578,11 @@ nvc0_graph_isr(struct drm_device *dev)
 	}
 
 	if (stat & 0x00080000) {
-		nvc0_graph_ctxctl_isr(dev);
+		u32 ustat = nv_rd32(dev, 0x409c18);
+
+		NV_INFO(dev, "PGRAPH: CTXCTRL ustat 0x%08x\n", ustat);
+
+		nv_wr32(dev, 0x409c20, ustat);
 		nv_wr32(dev, 0x400100, 0x00080000);
 		stat &= ~0x00080000;
 	}
@@ -726,6 +595,22 @@ nvc0_graph_isr(struct drm_device *dev)
 	nv_wr32(dev, 0x400500, 0x00010001);
 }
 
+static void
+nvc0_runk140_isr(struct drm_device *dev)
+{
+	u32 units = nv_rd32(dev, 0x00017c) & 0x1f;
+
+	while (units) {
+		u32 unit = ffs(units) - 1;
+		u32 reg = 0x140000 + unit * 0x2000;
+		u32 st0 = nv_mask(dev, reg + 0x1020, 0, 0);
+		u32 st1 = nv_mask(dev, reg + 0x1420, 0, 0);
+
+		NV_INFO(dev, "PRUNK140: %d 0x%08x 0x%08x\n", unit, st0, st1);
+		units &= ~(1 << unit);
+	}
+}
+
 static int
 nvc0_graph_create_fw(struct drm_device *dev, const char *fwname,
 		     struct nvc0_graph_fuc *fuc)
@@ -766,14 +651,13 @@ nvc0_graph_destroy(struct drm_device *dev, int engine)
 {
 	struct nvc0_graph_priv *priv = nv_engine(dev, engine);
 
-	if (nouveau_ctxfw) {
-		nvc0_graph_destroy_fw(&priv->fuc409c);
-		nvc0_graph_destroy_fw(&priv->fuc409d);
-		nvc0_graph_destroy_fw(&priv->fuc41ac);
-		nvc0_graph_destroy_fw(&priv->fuc41ad);
-	}
+	nvc0_graph_destroy_fw(&priv->fuc409c);
+	nvc0_graph_destroy_fw(&priv->fuc409d);
+	nvc0_graph_destroy_fw(&priv->fuc41ac);
+	nvc0_graph_destroy_fw(&priv->fuc41ad);
 
 	nouveau_irq_unregister(dev, 12);
+	nouveau_irq_unregister(dev, 25);
 
 	nouveau_gpuobj_ref(NULL, &priv->unk4188b8);
 	nouveau_gpuobj_ref(NULL, &priv->unk4188b4);
@@ -791,10 +675,13 @@ nvc0_graph_create(struct drm_device *dev)
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
 	struct nvc0_graph_priv *priv;
 	int ret, gpc, i;
-	u32 fermi;
 
-	fermi = nvc0_graph_class(dev);
-	if (!fermi) {
+	switch (dev_priv->chipset) {
+	case 0xc0:
+	case 0xc3:
+	case 0xc4:
+		break;
+	default:
 		NV_ERROR(dev, "PGRAPH: unsupported chipset, please report!\n");
 		return 0;
 	}
@@ -812,18 +699,17 @@ nvc0_graph_create(struct drm_device *dev)
 
 	NVOBJ_ENGINE_ADD(dev, GR, &priv->base);
 	nouveau_irq_register(dev, 12, nvc0_graph_isr);
+	nouveau_irq_register(dev, 25, nvc0_runk140_isr);
 
-	if (nouveau_ctxfw) {
-		NV_INFO(dev, "PGRAPH: using external firmware\n");
-		if (nvc0_graph_create_fw(dev, "fuc409c", &priv->fuc409c) ||
-		    nvc0_graph_create_fw(dev, "fuc409d", &priv->fuc409d) ||
-		    nvc0_graph_create_fw(dev, "fuc41ac", &priv->fuc41ac) ||
-		    nvc0_graph_create_fw(dev, "fuc41ad", &priv->fuc41ad)) {
-			ret = 0;
-			goto error;
-		}
+	if (nvc0_graph_create_fw(dev, "fuc409c", &priv->fuc409c) ||
+	    nvc0_graph_create_fw(dev, "fuc409d", &priv->fuc409d) ||
+	    nvc0_graph_create_fw(dev, "fuc41ac", &priv->fuc41ac) ||
+	    nvc0_graph_create_fw(dev, "fuc41ad", &priv->fuc41ad)) {
+		ret = 0;
+		goto error;
 	}
 
+
 	ret = nouveau_gpuobj_new(dev, NULL, 0x1000, 256, 0, &priv->unk4188b4);
 	if (ret)
 		goto error;
@@ -849,34 +735,25 @@ nvc0_graph_create(struct drm_device *dev)
 	case 0xc0:
 		if (priv->tp_total == 11) { /* 465, 3/4/4/0, 4 */
 			priv->magic_not_rop_nr = 0x07;
+			/* filled values up to tp_total, the rest 0 */
+			priv->magicgpc918      = 0x000ba2e9;
 		} else
 		if (priv->tp_total == 14) { /* 470, 3/3/4/4, 5 */
 			priv->magic_not_rop_nr = 0x05;
+			priv->magicgpc918      = 0x00092493;
 		} else
 		if (priv->tp_total == 15) { /* 480, 3/4/4/4, 6 */
 			priv->magic_not_rop_nr = 0x06;
+			priv->magicgpc918      = 0x00088889;
 		}
 		break;
 	case 0xc3: /* 450, 4/0/0/0, 2 */
 		priv->magic_not_rop_nr = 0x03;
+		priv->magicgpc918      = 0x00200000;
 		break;
 	case 0xc4: /* 460, 3/4/0/0, 4 */
 		priv->magic_not_rop_nr = 0x01;
-		break;
-	case 0xc1: /* 2/0/0/0, 1 */
-		priv->magic_not_rop_nr = 0x01;
-		break;
-	case 0xc8: /* 4/4/3/4, 5 */
-		priv->magic_not_rop_nr = 0x06;
-		break;
-	case 0xce: /* 4/4/0/0, 4 */
-		priv->magic_not_rop_nr = 0x03;
-		break;
-	case 0xcf: /* 4/0/0/0, 3 */
-		priv->magic_not_rop_nr = 0x03;
-		break;
-	case 0xd9: /* 1/0/0/0, 1 */
-		priv->magic_not_rop_nr = 0x01;
+		priv->magicgpc918      = 0x00124925;
 		break;
 	}
 
@@ -884,17 +761,15 @@ nvc0_graph_create(struct drm_device *dev)
 		NV_ERROR(dev, "PGRAPH: unknown config: %d/%d/%d/%d, %d\n",
 			 priv->tp_nr[0], priv->tp_nr[1], priv->tp_nr[2],
 			 priv->tp_nr[3], priv->rop_nr);
-		priv->magic_not_rop_nr = 0x00;
+		/* use 0xc3's values... */
+		priv->magic_not_rop_nr = 0x03;
+		priv->magicgpc918      = 0x00200000;
 	}
 
 	NVOBJ_CLASS(dev, 0x902d, GR); /* 2D */
 	NVOBJ_CLASS(dev, 0x9039, GR); /* M2MF */
 	NVOBJ_MTHD (dev, 0x9039, 0x0500, nvc0_graph_mthd_page_flip);
 	NVOBJ_CLASS(dev, 0x9097, GR); /* 3D */
-	if (fermi >= 0x9197)
-		NVOBJ_CLASS(dev, 0x9197, GR); /* 3D (NVC1-) */
-	if (fermi >= 0x9297)
-		NVOBJ_CLASS(dev, 0x9297, GR); /* 3D (NVC8-) */
 	NVOBJ_CLASS(dev, 0x90c0, GR); /* COMPUTE */
 	return 0;
 
@@ -902,3 +777,20 @@ error:
 	nvc0_graph_destroy(dev, NVOBJ_ENGINE_GR);
 	return ret;
 }
+
+MODULE_FIRMWARE("nouveau/nvc0_fuc409c");
+MODULE_FIRMWARE("nouveau/nvc0_fuc409d");
+MODULE_FIRMWARE("nouveau/nvc0_fuc41ac");
+MODULE_FIRMWARE("nouveau/nvc0_fuc41ad");
+MODULE_FIRMWARE("nouveau/nvc3_fuc409c");
+MODULE_FIRMWARE("nouveau/nvc3_fuc409d");
+MODULE_FIRMWARE("nouveau/nvc3_fuc41ac");
+MODULE_FIRMWARE("nouveau/nvc3_fuc41ad");
+MODULE_FIRMWARE("nouveau/nvc4_fuc409c");
+MODULE_FIRMWARE("nouveau/nvc4_fuc409d");
+MODULE_FIRMWARE("nouveau/nvc4_fuc41ac");
+MODULE_FIRMWARE("nouveau/nvc4_fuc41ad");
+MODULE_FIRMWARE("nouveau/fuc409c");
+MODULE_FIRMWARE("nouveau/fuc409d");
+MODULE_FIRMWARE("nouveau/fuc41ac");
+MODULE_FIRMWARE("nouveau/fuc41ad");
diff --git a/drivers/gpu/drm/nouveau/nvc0_graph.fuc b/drivers/gpu/drm/nouveau/nvc0_graph.fuc
deleted file mode 100644
index e6b2288..0000000
--- a/drivers/gpu/drm/nouveau/nvc0_graph.fuc
+++ /dev/null
@@ -1,400 +0,0 @@
-/* fuc microcode util functions for nvc0 PGRAPH
- *
- * Copyright 2011 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Ben Skeggs
- */
-
-define(`mmctx_data', `.b32 eval((($2 - 1) << 26) | $1)')
-define(`queue_init', `.skip eval((2 * 4) + ((8 * 4) * 2))')
-
-ifdef(`include_code', `
-// Error codes
-define(`E_BAD_COMMAND', 0x01)
-define(`E_CMD_OVERFLOW', 0x02)
-
-// Util macros to help with debugging ucode hangs etc
-define(`T_WAIT', 0)
-define(`T_MMCTX', 1)
-define(`T_STRWAIT', 2)
-define(`T_STRINIT', 3)
-define(`T_AUTO', 4)
-define(`T_CHAN', 5)
-define(`T_LOAD', 6)
-define(`T_SAVE', 7)
-define(`T_LCHAN', 8)
-define(`T_LCTXH', 9)
-
-define(`trace_set', `
-	mov $r8 0x83c
-	shl b32 $r8 6
-	clear b32 $r9
-	bset $r9 $1
-	iowr I[$r8 + 0x000] $r9		// CC_SCRATCH[7]
-')
-
-define(`trace_clr', `
-	mov $r8 0x85c
-	shl b32 $r8 6
-	clear b32 $r9
-	bset $r9 $1
-	iowr I[$r8 + 0x000] $r9		// CC_SCRATCH[7]
-')
-
-// queue_put - add request to queue
-//
-// In : $r13 queue pointer
-//	$r14 command
-//	$r15 data
-//
-queue_put:
-	// make sure we have space..
-	ld b32 $r8 D[$r13 + 0x0]	// GET
-	ld b32 $r9 D[$r13 + 0x4]	// PUT
-	xor $r8 8
-	cmpu b32 $r8 $r9
-	bra ne #queue_put_next
-		mov $r15 E_CMD_OVERFLOW
-		call #error
-		ret
-
-	// store cmd/data on queue
-	queue_put_next:
-	and $r8 $r9 7
-	shl b32 $r8 3
-	add b32 $r8 $r13
-	add b32 $r8 8
-	st b32 D[$r8 + 0x0] $r14
-	st b32 D[$r8 + 0x4] $r15
-
-	// update PUT
-	add b32 $r9 1
-	and $r9 0xf
-	st b32 D[$r13 + 0x4] $r9
-	ret
-
-// queue_get - fetch request from queue
-//
-// In : $r13 queue pointer
-//
-// Out:	$p1  clear on success (data available)
-//	$r14 command
-// 	$r15 data
-//
-queue_get:
-	bset $flags $p1
-	ld b32 $r8 D[$r13 + 0x0]	// GET
-	ld b32 $r9 D[$r13 + 0x4]	// PUT
-	cmpu b32 $r8 $r9
-	bra e #queue_get_done
-		// fetch first cmd/data pair
-		and $r9 $r8 7
-		shl b32 $r9 3
-		add b32 $r9 $r13
-		add b32 $r9 8
-		ld b32 $r14 D[$r9 + 0x0]
-		ld b32 $r15 D[$r9 + 0x4]
-
-		// update GET
-		add b32 $r8 1
-		and $r8 0xf
-		st b32 D[$r13 + 0x0] $r8
-		bclr $flags $p1
-queue_get_done:
-	ret
-
-// nv_rd32 - read 32-bit value from nv register
-//
-// In : $r14 register
-// Out: $r15 value
-//
-nv_rd32:
-	mov $r11 0x728
-	shl b32 $r11 6
-	mov b32 $r12 $r14
-	bset $r12 31			// MMIO_CTRL_PENDING
-	iowr I[$r11 + 0x000] $r12	// MMIO_CTRL
-	nv_rd32_wait:
-		iord $r12 I[$r11 + 0x000]
-		xbit $r12 $r12 31
-		bra ne #nv_rd32_wait
-	mov $r10 6			// DONE_MMIO_RD
-	call #wait_doneo
-	iord $r15 I[$r11 + 0x100]	// MMIO_RDVAL
-	ret
-
-// nv_wr32 - write 32-bit value to nv register
-//
-// In : $r14 register
-//      $r15 value
-//
-nv_wr32:
-	mov $r11 0x728
-	shl b32 $r11 6
-	iowr I[$r11 + 0x200] $r15	// MMIO_WRVAL
-	mov b32 $r12 $r14
-	bset $r12 31			// MMIO_CTRL_PENDING
-	bset $r12 30			// MMIO_CTRL_WRITE
-	iowr I[$r11 + 0x000] $r12	// MMIO_CTRL
-	nv_wr32_wait:
-		iord $r12 I[$r11 + 0x000]
-		xbit $r12 $r12 31
-		bra ne #nv_wr32_wait
-	ret
-
-// (re)set watchdog timer
-//
-// In : $r15 timeout
-//
-watchdog_reset:
-	mov $r8 0x430
-	shl b32 $r8 6
-	bset $r15 31
-	iowr I[$r8 + 0x000] $r15
-	ret
-
-// clear watchdog timer
-watchdog_clear:
-	mov $r8 0x430
-	shl b32 $r8 6
-	iowr I[$r8 + 0x000] $r0
-	ret
-
-// wait_done{z,o} - wait on FUC_DONE bit to become clear/set
-//
-// In : $r10 bit to wait on
-//
-define(`wait_done', `
-$1:
-	trace_set(T_WAIT);
-	mov $r8 0x818
-	shl b32 $r8 6
-	iowr I[$r8 + 0x000] $r10	// CC_SCRATCH[6] = wait bit
-	wait_done_$1:
-		mov $r8 0x400
-		shl b32 $r8 6
-		iord $r8 I[$r8 + 0x000]	// DONE
-		xbit $r8 $r8 $r10
-		bra $2 #wait_done_$1
-	trace_clr(T_WAIT)
-	ret
-')
-wait_done(wait_donez, ne)
-wait_done(wait_doneo, e)
-
-// mmctx_size - determine size of a mmio list transfer
-//
-// In : $r14 mmio list head
-//      $r15 mmio list tail
-// Out: $r15 transfer size (in bytes)
-//
-mmctx_size:
-	clear b32 $r9
-	nv_mmctx_size_loop:
-		ld b32 $r8 D[$r14]
-		shr b32 $r8 26
-		add b32 $r8 1
-		shl b32 $r8 2
-		add b32 $r9 $r8
-		add b32 $r14 4
-		cmpu b32 $r14 $r15
-		bra ne #nv_mmctx_size_loop
-	mov b32 $r15 $r9
-	ret
-
-// mmctx_xfer - execute a list of mmio transfers
-//
-// In : $r10 flags
-//		bit 0: direction (0 = save, 1 = load)
-//		bit 1: set if first transfer
-//		bit 2: set if last transfer
-//	$r11 base
-//	$r12 mmio list head
-//	$r13 mmio list tail
-//	$r14 multi_stride
-//	$r15 multi_mask
-//
-mmctx_xfer:
-	trace_set(T_MMCTX)
-	mov $r8 0x710
-	shl b32 $r8 6
-	clear b32 $r9
-	or $r11 $r11
-	bra e #mmctx_base_disabled
-		iowr I[$r8 + 0x000] $r11	// MMCTX_BASE
-		bset $r9 0			// BASE_EN
-	mmctx_base_disabled:
-	or $r14 $r14
-	bra e #mmctx_multi_disabled
-		iowr I[$r8 + 0x200] $r14 	// MMCTX_MULTI_STRIDE
-		iowr I[$r8 + 0x300] $r15 	// MMCTX_MULTI_MASK
-		bset $r9 1			// MULTI_EN
-	mmctx_multi_disabled:
-	add b32 $r8 0x100
-
-	xbit $r11 $r10 0
-	shl b32 $r11 16			// DIR
-	bset $r11 12			// QLIMIT = 0x10
-	xbit $r14 $r10 1
-	shl b32 $r14 17
-	or $r11 $r14			// START_TRIGGER
-	iowr I[$r8 + 0x000] $r11	// MMCTX_CTRL
-
-	// loop over the mmio list, and send requests to the hw
-	mmctx_exec_loop:
-		// wait for space in mmctx queue
-		mmctx_wait_free:
-			iord $r14 I[$r8 + 0x000] // MMCTX_CTRL
-			and $r14 0x1f
-			bra e #mmctx_wait_free
-
-		// queue up an entry
-		ld b32 $r14 D[$r12]
-		or $r14 $r9
-		iowr I[$r8 + 0x300] $r14
-		add b32 $r12 4
-		cmpu b32 $r12 $r13
-		bra ne #mmctx_exec_loop
-
-	xbit $r11 $r10 2
-	bra ne #mmctx_stop
-		// wait for queue to empty
-		mmctx_fini_wait:
-			iord $r11 I[$r8 + 0x000]	// MMCTX_CTRL
-			and $r11 0x1f
-			cmpu b32 $r11 0x10
-			bra ne #mmctx_fini_wait
-		mov $r10 2				// DONE_MMCTX
-		call #wait_donez
-		bra #mmctx_done
-	mmctx_stop:
-		xbit $r11 $r10 0
-		shl b32 $r11 16			// DIR
-		bset $r11 12			// QLIMIT = 0x10
-		bset $r11 18			// STOP_TRIGGER
-		iowr I[$r8 + 0x000] $r11	// MMCTX_CTRL
-		mmctx_stop_wait:
-			// wait for STOP_TRIGGER to clear
-			iord $r11 I[$r8 + 0x000] // MMCTX_CTRL
-			xbit $r11 $r11 18
-			bra ne #mmctx_stop_wait
-	mmctx_done:
-	trace_clr(T_MMCTX)
-	ret
-
-// Wait for DONE_STRAND
-//
-strand_wait:
-	push $r10
-	mov $r10 2
-	call #wait_donez
-	pop $r10
-	ret
-
-// unknown - call before issuing strand commands
-//
-strand_pre:
-	mov $r8 0x4afc
-	sethi $r8 0x20000
-	mov $r9 0xc
-	iowr I[$r8] $r9
-	call #strand_wait
-	ret
-
-// unknown - call after issuing strand commands
-//
-strand_post:
-	mov $r8 0x4afc
-	sethi $r8 0x20000
-	mov $r9 0xd
-	iowr I[$r8] $r9
-	call #strand_wait
-	ret
-
-// Selects strand set?!
-//
-// In: $r14 id
-//
-strand_set:
-	mov $r10 0x4ffc
-	sethi $r10 0x20000
-	sub b32 $r11 $r10 0x500
-	mov $r12 0xf
-	iowr I[$r10 + 0x000] $r12		// 0x93c = 0xf
-	mov $r12 0xb
-	iowr I[$r11 + 0x000] $r12		// 0x928 = 0xb
-	call #strand_wait
-	iowr I[$r10 + 0x000] $r14		// 0x93c = <id>
-	mov $r12 0xa
-	iowr I[$r11 + 0x000] $r12		// 0x928 = 0xa
-	call #strand_wait
-	ret
-
-// Initialise strand context data
-//
-// In : $r15 context base
-// Out: $r15 context size (in bytes)
-//
-// Strandset(?) 3 hardcoded currently
-//
-strand_ctx_init:
-	trace_set(T_STRINIT)
-	call #strand_pre
-	mov $r14 3
-	call #strand_set
-	mov $r10 0x46fc
-	sethi $r10 0x20000
-	add b32 $r11 $r10 0x400
-	iowr I[$r10 + 0x100] $r0	// STRAND_FIRST_GENE = 0
-	mov $r12 1
-	iowr I[$r11 + 0x000] $r12	// STRAND_CMD = LATCH_FIRST_GENE
-	call #strand_wait
-	sub b32 $r12 $r0 1
-	iowr I[$r10 + 0x000] $r12	// STRAND_GENE_CNT = 0xffffffff
-	mov $r12 2
-	iowr I[$r11 + 0x000] $r12	// STRAND_CMD = LATCH_GENE_CNT
-	call #strand_wait
-	call #strand_post
-
-	// read the size of each strand, poke the context offset of
-	// each into STRAND_{SAVE,LOAD}_SWBASE now, no need to worry
-	// about it later then.
-	mov $r8 0x880
-	shl b32 $r8 6
-	iord $r9 I[$r8 + 0x000]		// STRANDS
-	add b32 $r8 0x2200
-	shr b32 $r14 $r15 8
-	ctx_init_strand_loop:
-		iowr I[$r8 + 0x000] $r14	// STRAND_SAVE_SWBASE
-		iowr I[$r8 + 0x100] $r14	// STRAND_LOAD_SWBASE
-		iord $r10 I[$r8 + 0x200]	// STRAND_SIZE
-		shr b32 $r10 6
-		add b32 $r10 1
-		add b32 $r14 $r10
-		add b32 $r8 4
-		sub b32 $r9 1
-		bra ne #ctx_init_strand_loop
-
-	shl b32 $r14 8
-	sub b32 $r15 $r14 $r15
-	trace_clr(T_STRINIT)
-	ret
-')
diff --git a/drivers/gpu/drm/nouveau/nvc0_graph.h b/drivers/gpu/drm/nouveau/nvc0_graph.h
index 91d44ea..f5d184e0 100644
--- a/drivers/gpu/drm/nouveau/nvc0_graph.h
+++ b/drivers/gpu/drm/nouveau/nvc0_graph.h
@@ -57,7 +57,8 @@ struct nvc0_graph_priv {
 	struct nouveau_gpuobj *unk4188b4;
 	struct nouveau_gpuobj *unk4188b8;
 
-	u8 magic_not_rop_nr;
+	u8  magic_not_rop_nr;
+	u32 magicgpc918;
 };
 
 struct nvc0_graph_chan {
@@ -71,27 +72,4 @@ struct nvc0_graph_chan {
 
 int nvc0_grctx_generate(struct nouveau_channel *);
 
-/* nvc0_graph.c uses this also to determine supported chipsets */
-static inline u32
-nvc0_graph_class(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-
-	switch (dev_priv->chipset) {
-	case 0xc0:
-	case 0xc3:
-	case 0xc4:
-	case 0xce: /* guess, mmio trace shows only 0x9097 state */
-	case 0xcf: /* guess, mmio trace shows only 0x9097 state */
-		return 0x9097;
-	case 0xc1:
-		return 0x9197;
-	case 0xc8:
-	case 0xd9:
-		return 0x9297;
-	default:
-		return 0;
-	}
-}
-
 #endif
diff --git a/drivers/gpu/drm/nouveau/nvc0_grctx.c b/drivers/gpu/drm/nouveau/nvc0_grctx.c
index de77842..6df0661 100644
--- a/drivers/gpu/drm/nouveau/nvc0_grctx.c
+++ b/drivers/gpu/drm/nouveau/nvc0_grctx.c
@@ -45,9 +45,6 @@ nv_mthd(struct drm_device *dev, u32 class, u32 mthd, u32 data)
 static void
 nvc0_grctx_generate_9097(struct drm_device *dev)
 {
-	u32 fermi = nvc0_graph_class(dev);
-	u32 mthd;
-
 	nv_mthd(dev, 0x9097, 0x0800, 0x00000000);
 	nv_mthd(dev, 0x9097, 0x0840, 0x00000000);
 	nv_mthd(dev, 0x9097, 0x0880, 0x00000000);
@@ -827,10 +824,134 @@ nvc0_grctx_generate_9097(struct drm_device *dev)
 	nv_mthd(dev, 0x9097, 0x1eb8, 0x00000001);
 	nv_mthd(dev, 0x9097, 0x1ed8, 0x00000001);
 	nv_mthd(dev, 0x9097, 0x1ef8, 0x00000001);
-	if (fermi == 0x9097) {
-		for (mthd = 0x3400; mthd <= 0x35fc; mthd += 4)
-			nv_mthd(dev, 0x9097, mthd, 0x00000000);
-	}
+	nv_mthd(dev, 0x9097, 0x3400, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3404, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3408, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x340c, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3410, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3414, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3418, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x341c, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3420, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3424, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3428, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x342c, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3430, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3434, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3438, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x343c, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3440, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3444, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3448, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x344c, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3450, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3454, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3458, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x345c, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3460, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3464, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3468, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x346c, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3470, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3474, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3478, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x347c, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3480, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3484, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3488, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x348c, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3490, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3494, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3498, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x349c, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34a0, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34a4, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34a8, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34ac, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34b0, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34b4, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34b8, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34bc, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34c0, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34c4, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34c8, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34cc, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34d0, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34d4, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34d8, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34dc, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34e0, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34e4, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34e8, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34ec, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34f0, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34f4, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34f8, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x34fc, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3500, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3504, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3508, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x350c, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3510, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3514, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3518, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x351c, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3520, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3524, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3528, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x352c, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3530, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3534, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3538, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x353c, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3540, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3544, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3548, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x354c, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3550, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3554, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3558, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x355c, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3560, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3564, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3568, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x356c, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3570, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3574, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3578, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x357c, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3580, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3584, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3588, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x358c, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3590, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3594, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x3598, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x359c, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35a0, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35a4, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35a8, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35ac, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35b0, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35b4, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35b8, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35bc, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35c0, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35c4, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35c8, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35cc, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35d0, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35d4, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35d8, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35dc, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35e0, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35e4, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35e8, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35ec, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35f0, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35f4, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35f8, 0x00000000);
+	nv_mthd(dev, 0x9097, 0x35fc, 0x00000000);
 	nv_mthd(dev, 0x9097, 0x030c, 0x00000001);
 	nv_mthd(dev, 0x9097, 0x1944, 0x00000000);
 	nv_mthd(dev, 0x9097, 0x1514, 0x00000000);
@@ -1200,37 +1321,6 @@ nvc0_grctx_generate_9097(struct drm_device *dev)
 }
 
 static void
-nvc0_grctx_generate_9197(struct drm_device *dev)
-{
-	u32 fermi = nvc0_graph_class(dev);
-	u32 mthd;
-
-	if (fermi == 0x9197) {
-		for (mthd = 0x3400; mthd <= 0x35fc; mthd += 4)
-			nv_mthd(dev, 0x9197, mthd, 0x00000000);
-	}
-	nv_mthd(dev, 0x9197, 0x02e4, 0x0000b001);
-}
-
-static void
-nvc0_grctx_generate_9297(struct drm_device *dev)
-{
-	u32 fermi = nvc0_graph_class(dev);
-	u32 mthd;
-
-	if (fermi == 0x9297) {
-		for (mthd = 0x3400; mthd <= 0x35fc; mthd += 4)
-			nv_mthd(dev, 0x9297, mthd, 0x00000000);
-	}
-	nv_mthd(dev, 0x9297, 0x036c, 0x00000000);
-	nv_mthd(dev, 0x9297, 0x0370, 0x00000000);
-	nv_mthd(dev, 0x9297, 0x07a4, 0x00000000);
-	nv_mthd(dev, 0x9297, 0x07a8, 0x00000000);
-	nv_mthd(dev, 0x9297, 0x0374, 0x00000000);
-	nv_mthd(dev, 0x9297, 0x0378, 0x00000020);
-}
-
-static void
 nvc0_grctx_generate_902d(struct drm_device *dev)
 {
 	nv_mthd(dev, 0x902d, 0x0200, 0x000000cf);
@@ -1268,17 +1358,6 @@ nvc0_grctx_generate_9039(struct drm_device *dev)
 static void
 nvc0_grctx_generate_90c0(struct drm_device *dev)
 {
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	int i;
-
-	for (i = 0; dev_priv->chipset == 0xd9 && i < 4; i++) {
-		nv_mthd(dev, 0x90c0, 0x2700 + (i * 0x40), 0x00000000);
-		nv_mthd(dev, 0x90c0, 0x2720 + (i * 0x40), 0x00000000);
-		nv_mthd(dev, 0x90c0, 0x2704 + (i * 0x40), 0x00000000);
-		nv_mthd(dev, 0x90c0, 0x2724 + (i * 0x40), 0x00000000);
-		nv_mthd(dev, 0x90c0, 0x2708 + (i * 0x40), 0x00000000);
-		nv_mthd(dev, 0x90c0, 0x2728 + (i * 0x40), 0x00000000);
-	}
 	nv_mthd(dev, 0x90c0, 0x270c, 0x00000000);
 	nv_mthd(dev, 0x90c0, 0x272c, 0x00000000);
 	nv_mthd(dev, 0x90c0, 0x274c, 0x00000000);
@@ -1287,12 +1366,6 @@ nvc0_grctx_generate_90c0(struct drm_device *dev)
 	nv_mthd(dev, 0x90c0, 0x27ac, 0x00000000);
 	nv_mthd(dev, 0x90c0, 0x27cc, 0x00000000);
 	nv_mthd(dev, 0x90c0, 0x27ec, 0x00000000);
-	for (i = 0; dev_priv->chipset == 0xd9 && i < 4; i++) {
-		nv_mthd(dev, 0x90c0, 0x2710 + (i * 0x40), 0x00014000);
-		nv_mthd(dev, 0x90c0, 0x2730 + (i * 0x40), 0x00014000);
-		nv_mthd(dev, 0x90c0, 0x2714 + (i * 0x40), 0x00000040);
-		nv_mthd(dev, 0x90c0, 0x2734 + (i * 0x40), 0x00000040);
-	}
 	nv_mthd(dev, 0x90c0, 0x030c, 0x00000001);
 	nv_mthd(dev, 0x90c0, 0x1944, 0x00000000);
 	nv_mthd(dev, 0x90c0, 0x0758, 0x00000100);
@@ -1486,22 +1559,9 @@ nvc0_grctx_generate_unk47xx(struct drm_device *dev)
 static void
 nvc0_grctx_generate_shaders(struct drm_device *dev)
 {
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-
-	if (dev_priv->chipset == 0xd9) {
-		nv_wr32(dev, 0x405800, 0x0f8000bf);
-		nv_wr32(dev, 0x405830, 0x02180218);
-		nv_wr32(dev, 0x405834, 0x08000000);
-	} else
-	if (dev_priv->chipset == 0xc1) {
-		nv_wr32(dev, 0x405800, 0x0f8000bf);
-		nv_wr32(dev, 0x405830, 0x02180218);
-		nv_wr32(dev, 0x405834, 0x00000000);
-	} else {
-		nv_wr32(dev, 0x405800, 0x078000bf);
-		nv_wr32(dev, 0x405830, 0x02180000);
-		nv_wr32(dev, 0x405834, 0x00000000);
-	}
+	nv_wr32(dev, 0x405800, 0x078000bf);
+	nv_wr32(dev, 0x405830, 0x02180000);
+	nv_wr32(dev, 0x405834, 0x00000000);
 	nv_wr32(dev, 0x405838, 0x00000000);
 	nv_wr32(dev, 0x405854, 0x00000000);
 	nv_wr32(dev, 0x405870, 0x00000001);
@@ -1526,19 +1586,10 @@ nvc0_grctx_generate_unk60xx(struct drm_device *dev)
 static void
 nvc0_grctx_generate_unk64xx(struct drm_device *dev)
 {
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-
 	nv_wr32(dev, 0x4064a8, 0x00000000);
 	nv_wr32(dev, 0x4064ac, 0x00003fff);
 	nv_wr32(dev, 0x4064b4, 0x00000000);
 	nv_wr32(dev, 0x4064b8, 0x00000000);
-	if (dev_priv->chipset == 0xd9)
-		nv_wr32(dev, 0x4064bc, 0x00000000);
-	if (dev_priv->chipset == 0xc1 ||
-	    dev_priv->chipset == 0xd9) {
-		nv_wr32(dev, 0x4064c0, 0x80140078);
-		nv_wr32(dev, 0x4064c4, 0x0086ffff);
-	}
 }
 
 static void
@@ -1571,28 +1622,22 @@ static void
 nvc0_grctx_generate_rop(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	int chipset = dev_priv->chipset;
 
 	/* ROPC_BROADCAST */
 	nv_wr32(dev, 0x408800, 0x02802a3c);
 	nv_wr32(dev, 0x408804, 0x00000040);
-	if (chipset == 0xd9) {
-		nv_wr32(dev, 0x408808, 0x1043e005);
+	nv_wr32(dev, 0x408808, 0x0003e00d);
+	switch (dev_priv->chipset) {
+	case 0xc0:
+		nv_wr32(dev, 0x408900, 0x0080b801);
+		break;
+	case 0xc3:
+	case 0xc4:
 		nv_wr32(dev, 0x408900, 0x3080b801);
-		nv_wr32(dev, 0x408904, 0x1043e005);
-		nv_wr32(dev, 0x408908, 0x00c8102f);
-	} else
-	if (chipset == 0xc1) {
-		nv_wr32(dev, 0x408808, 0x1003e005);
-		nv_wr32(dev, 0x408900, 0x3080b801);
-		nv_wr32(dev, 0x408904, 0x62000001);
-		nv_wr32(dev, 0x408908, 0x00c80929);
-	} else {
-		nv_wr32(dev, 0x408808, 0x0003e00d);
-		nv_wr32(dev, 0x408900, 0x3080b801);
-		nv_wr32(dev, 0x408904, 0x02000001);
-		nv_wr32(dev, 0x408908, 0x00c80929);
+		break;
 	}
+	nv_wr32(dev, 0x408904, 0x02000001);
+	nv_wr32(dev, 0x408908, 0x00c80929);
 	nv_wr32(dev, 0x40890c, 0x00000000);
 	nv_wr32(dev, 0x408980, 0x0000011d);
 }
@@ -1600,8 +1645,6 @@ nvc0_grctx_generate_rop(struct drm_device *dev)
 static void
 nvc0_grctx_generate_gpc(struct drm_device *dev)
 {
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	int chipset = dev_priv->chipset;
 	int i;
 
 	/* GPC_BROADCAST */
@@ -1611,7 +1654,7 @@ nvc0_grctx_generate_gpc(struct drm_device *dev)
 	nv_wr32(dev, 0x418408, 0x00000000);
 	nv_wr32(dev, 0x41840c, 0x00001008);
 	nv_wr32(dev, 0x418410, 0x0fff0fff);
-	nv_wr32(dev, 0x418414, chipset != 0xd9 ? 0x00200fff : 0x02200fff);
+	nv_wr32(dev, 0x418414, 0x00200fff);
 	nv_wr32(dev, 0x418450, 0x00000000);
 	nv_wr32(dev, 0x418454, 0x00000000);
 	nv_wr32(dev, 0x418458, 0x00000000);
@@ -1626,17 +1669,14 @@ nvc0_grctx_generate_gpc(struct drm_device *dev)
 	nv_wr32(dev, 0x418700, 0x00000002);
 	nv_wr32(dev, 0x418704, 0x00000080);
 	nv_wr32(dev, 0x418708, 0x00000000);
-	nv_wr32(dev, 0x41870c, chipset != 0xd9 ? 0x07c80000 : 0x00000000);
+	nv_wr32(dev, 0x41870c, 0x07c80000);
 	nv_wr32(dev, 0x418710, 0x00000000);
-	nv_wr32(dev, 0x418800, chipset != 0xd9 ? 0x0006860a : 0x7006860a);
+	nv_wr32(dev, 0x418800, 0x0006860a);
 	nv_wr32(dev, 0x418808, 0x00000000);
 	nv_wr32(dev, 0x41880c, 0x00000000);
 	nv_wr32(dev, 0x418810, 0x00000000);
 	nv_wr32(dev, 0x418828, 0x00008442);
-	if (chipset == 0xc1 || chipset == 0xd9)
-		nv_wr32(dev, 0x418830, 0x10000001);
-	else
-		nv_wr32(dev, 0x418830, 0x00000001);
+	nv_wr32(dev, 0x418830, 0x00000001);
 	nv_wr32(dev, 0x4188d8, 0x00000008);
 	nv_wr32(dev, 0x4188e0, 0x01000000);
 	nv_wr32(dev, 0x4188e8, 0x00000000);
@@ -1644,12 +1684,7 @@ nvc0_grctx_generate_gpc(struct drm_device *dev)
 	nv_wr32(dev, 0x4188f0, 0x00000000);
 	nv_wr32(dev, 0x4188f4, 0x00000000);
 	nv_wr32(dev, 0x4188f8, 0x00000000);
-	if (chipset == 0xd9)
-		nv_wr32(dev, 0x4188fc, 0x20100008);
-	else if (chipset == 0xc1)
-		nv_wr32(dev, 0x4188fc, 0x00100018);
-	else
-		nv_wr32(dev, 0x4188fc, 0x00100000);
+	nv_wr32(dev, 0x4188fc, 0x00100000);
 	nv_wr32(dev, 0x41891c, 0x00ff00ff);
 	nv_wr32(dev, 0x418924, 0x00000000);
 	nv_wr32(dev, 0x418928, 0x00ffff00);
@@ -1663,7 +1698,7 @@ nvc0_grctx_generate_gpc(struct drm_device *dev)
 		nv_wr32(dev, 0x418a14 + (i * 0x20), 0x00000000);
 		nv_wr32(dev, 0x418a18 + (i * 0x20), 0x00000000);
 	}
-	nv_wr32(dev, 0x418b00, chipset != 0xd9 ? 0x00000000 : 0x00000006);
+	nv_wr32(dev, 0x418b00, 0x00000000);
 	nv_wr32(dev, 0x418b08, 0x0a418820);
 	nv_wr32(dev, 0x418b0c, 0x062080e6);
 	nv_wr32(dev, 0x418b10, 0x020398a4);
@@ -1680,8 +1715,6 @@ nvc0_grctx_generate_gpc(struct drm_device *dev)
 	nv_wr32(dev, 0x418c24, 0x00000000);
 	nv_wr32(dev, 0x418c28, 0x00000000);
 	nv_wr32(dev, 0x418c2c, 0x00000000);
-	if (chipset == 0xc1 || chipset == 0xd9)
-		nv_wr32(dev, 0x418c6c, 0x00000001);
 	nv_wr32(dev, 0x418c80, 0x20200004);
 	nv_wr32(dev, 0x418c8c, 0x00000001);
 	nv_wr32(dev, 0x419000, 0x00000780);
@@ -1694,16 +1727,10 @@ static void
 nvc0_grctx_generate_tp(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	int chipset = dev_priv->chipset;
 
 	/* GPC_BROADCAST.TP_BROADCAST */
-	nv_wr32(dev, 0x419818, 0x00000000);
-	nv_wr32(dev, 0x41983c, 0x00038bc7);
 	nv_wr32(dev, 0x419848, 0x00000000);
-	if (chipset == 0xc1 || chipset == 0xd9)
-		nv_wr32(dev, 0x419864, 0x00000129);
-	else
-		nv_wr32(dev, 0x419864, 0x0000012a);
+	nv_wr32(dev, 0x419864, 0x0000012a);
 	nv_wr32(dev, 0x419888, 0x00000000);
 	nv_wr32(dev, 0x419a00, 0x000001f0);
 	nv_wr32(dev, 0x419a04, 0x00000001);
@@ -1713,10 +1740,8 @@ nvc0_grctx_generate_tp(struct drm_device *dev)
 	nv_wr32(dev, 0x419a14, 0x00000200);
 	nv_wr32(dev, 0x419a1c, 0x00000000);
 	nv_wr32(dev, 0x419a20, 0x00000800);
-	if (chipset == 0xd9)
-		nv_wr32(dev, 0x00419ac4, 0x0017f440);
-	else if (chipset != 0xc0 && chipset != 0xc8)
-		nv_wr32(dev, 0x00419ac4, 0x0007f440);
+	if (dev_priv->chipset != 0xc0)
+		nv_wr32(dev, 0x00419ac4, 0x0007f440); /* 0xc3 */
 	nv_wr32(dev, 0x419b00, 0x0a418820);
 	nv_wr32(dev, 0x419b04, 0x062080e6);
 	nv_wr32(dev, 0x419b08, 0x020398a4);
@@ -1724,34 +1749,17 @@ nvc0_grctx_generate_tp(struct drm_device *dev)
 	nv_wr32(dev, 0x419b10, 0x0a418820);
 	nv_wr32(dev, 0x419b14, 0x000000e6);
 	nv_wr32(dev, 0x419bd0, 0x00900103);
-	if (chipset == 0xc1 || chipset == 0xd9)
-		nv_wr32(dev, 0x419be0, 0x00400001);
-	else
-		nv_wr32(dev, 0x419be0, 0x00000001);
+	nv_wr32(dev, 0x419be0, 0x00000001);
 	nv_wr32(dev, 0x419be4, 0x00000000);
-	nv_wr32(dev, 0x419c00, chipset != 0xd9 ? 0x00000002 : 0x0000000a);
+	nv_wr32(dev, 0x419c00, 0x00000002);
 	nv_wr32(dev, 0x419c04, 0x00000006);
 	nv_wr32(dev, 0x419c08, 0x00000002);
 	nv_wr32(dev, 0x419c20, 0x00000000);
-	if (dev_priv->chipset == 0xd9) {
-		nv_wr32(dev, 0x419c24, 0x00084210);
-		nv_wr32(dev, 0x419c28, 0x3cf3cf3c);
-		nv_wr32(dev, 0x419cb0, 0x00020048);
-	} else
-	if (chipset == 0xce || chipset == 0xcf) {
-		nv_wr32(dev, 0x419cb0, 0x00020048);
-	} else {
-		nv_wr32(dev, 0x419cb0, 0x00060048);
-	}
+	nv_wr32(dev, 0x419cbc, 0x28137606);
 	nv_wr32(dev, 0x419ce8, 0x00000000);
 	nv_wr32(dev, 0x419cf4, 0x00000183);
-	if (chipset == 0xc1 || chipset == 0xd9)
-		nv_wr32(dev, 0x419d20, 0x12180000);
-	else
-		nv_wr32(dev, 0x419d20, 0x02180000);
+	nv_wr32(dev, 0x419d20, 0x02180000);
 	nv_wr32(dev, 0x419d24, 0x00001fff);
-	if (chipset == 0xc1 || chipset == 0xd9)
-		nv_wr32(dev, 0x419d44, 0x02180218);
 	nv_wr32(dev, 0x419e04, 0x00000000);
 	nv_wr32(dev, 0x419e08, 0x00000000);
 	nv_wr32(dev, 0x419e0c, 0x00000000);
@@ -1777,11 +1785,11 @@ nvc0_grctx_generate_tp(struct drm_device *dev)
 	nv_wr32(dev, 0x419e8c, 0x00000000);
 	nv_wr32(dev, 0x419e90, 0x00000000);
 	nv_wr32(dev, 0x419e98, 0x00000000);
-	if (chipset != 0xc0 && chipset != 0xc8)
+	if (dev_priv->chipset != 0xc0)
 		nv_wr32(dev, 0x419ee0, 0x00011110);
 	nv_wr32(dev, 0x419f50, 0x00000000);
 	nv_wr32(dev, 0x419f54, 0x00000000);
-	if (chipset != 0xc0 && chipset != 0xc8)
+	if (dev_priv->chipset != 0xc0)
 		nv_wr32(dev, 0x419f58, 0x00000000);
 }
 
@@ -1793,7 +1801,6 @@ nvc0_grctx_generate(struct nouveau_channel *chan)
 	struct nvc0_graph_chan *grch = chan->engctx[NVOBJ_ENGINE_GR];
 	struct drm_device *dev = chan->dev;
 	int i, gpc, tp, id;
-	u32 fermi = nvc0_graph_class(dev);
 	u32 r000260, tmp;
 
 	r000260 = nv_rd32(dev, 0x000260);
@@ -1850,7 +1857,10 @@ nvc0_grctx_generate(struct nouveau_channel *chan)
 	nv_wr32(dev, 0x40587c, 0x00000000);
 
 	if (1) {
-		u8 tpnr[GPC_MAX], data[TP_MAX];
+		const u8 chipset_tp_max[] = { 16, 0, 0, 4, 8 };
+		u8 max = chipset_tp_max[dev_priv->chipset & 0x0f];
+		u8 tpnr[GPC_MAX];
+		u8 data[32];
 
 		memcpy(tpnr, priv->tp_nr, sizeof(priv->tp_nr));
 		memset(data, 0x1f, sizeof(data));
@@ -1864,7 +1874,7 @@ nvc0_grctx_generate(struct nouveau_channel *chan)
 			data[tp] = gpc;
 		}
 
-		for (i = 0; i < 4; i++)
+		for (i = 0; i < max / 4; i++)
 			nv_wr32(dev, 0x4060a8 + (i * 4), ((u32 *)data)[i]);
 	}
 
@@ -1876,7 +1886,6 @@ nvc0_grctx_generate(struct nouveau_channel *chan)
 		/* calculate first set of magics */
 		memcpy(tpnr, priv->tp_nr, sizeof(priv->tp_nr));
 
-		gpc = -1;
 		for (tp = 0; tp < priv->tp_total; tp++) {
 			do {
 				gpc = (gpc + 1) % priv->gpc_nr;
@@ -1926,26 +1935,30 @@ nvc0_grctx_generate(struct nouveau_channel *chan)
 
 	if (1) {
 		u32 tp_mask = 0, tp_set = 0;
-		u8  tpnr[GPC_MAX], a, b;
+		u8  tpnr[GPC_MAX];
 
 		memcpy(tpnr, priv->tp_nr, sizeof(priv->tp_nr));
 		for (gpc = 0; gpc < priv->gpc_nr; gpc++)
 			tp_mask |= ((1 << priv->tp_nr[gpc]) - 1) << (gpc * 8);
 
-		for (i = 0, gpc = -1, b = -1; i < 32; i++) {
-			a = (i * (priv->tp_total - 1)) / 32;
-			if (a != b) {
-				b = a;
-				do {
-					gpc = (gpc + 1) % priv->gpc_nr;
-				} while (!tpnr[gpc]);
-				tp = priv->tp_nr[gpc] - tpnr[gpc]--;
+		gpc = -1;
+		for (i = 0, gpc = -1; i < 32; i++) {
+			int ltp = i * (priv->tp_total - 1) / 32;
 
-				tp_set |= 1 << ((gpc * 8) + tp);
-			}
+			do {
+				gpc = (gpc + 1) % priv->gpc_nr;
+			} while (!tpnr[gpc]);
+			tp = priv->tp_nr[gpc] - tpnr[gpc]--;
+
+			tp_set |= 1 << ((gpc * 8) + tp);
 
-			nv_wr32(dev, 0x406800 + (i * 0x20), tp_set);
-			nv_wr32(dev, 0x406c00 + (i * 0x20), tp_set ^ tp_mask);
+			do {
+				nv_wr32(dev, 0x406800 + (i * 0x20), tp_set);
+				tp_set ^= tp_mask;
+				nv_wr32(dev, 0x406c00 + (i * 0x20), tp_set);
+				tp_set ^= tp_mask;
+			} while (ltp == (++i * (priv->tp_total - 1) / 32));
+			i--;
 		}
 	}
 
@@ -2050,10 +2063,6 @@ nvc0_grctx_generate(struct nouveau_channel *chan)
 	nv_icmd(dev, 0x00000215, 0x00000040);
 	nv_icmd(dev, 0x00000216, 0x00000040);
 	nv_icmd(dev, 0x00000217, 0x00000040);
-	if (dev_priv->chipset == 0xd9) {
-		for (i = 0x0400; i <= 0x0417; i++)
-			nv_icmd(dev, i, 0x00000040);
-	}
 	nv_icmd(dev, 0x00000218, 0x0000c080);
 	nv_icmd(dev, 0x00000219, 0x0000c080);
 	nv_icmd(dev, 0x0000021a, 0x0000c080);
@@ -2062,10 +2071,6 @@ nvc0_grctx_generate(struct nouveau_channel *chan)
 	nv_icmd(dev, 0x0000021d, 0x0000c080);
 	nv_icmd(dev, 0x0000021e, 0x0000c080);
 	nv_icmd(dev, 0x0000021f, 0x0000c080);
-	if (dev_priv->chipset == 0xd9) {
-		for (i = 0x0440; i <= 0x0457; i++)
-			nv_icmd(dev, i, 0x0000c080);
-	}
 	nv_icmd(dev, 0x000000ad, 0x0000013e);
 	nv_icmd(dev, 0x000000e1, 0x00000010);
 	nv_icmd(dev, 0x00000290, 0x00000000);
@@ -2628,9 +2633,6 @@ nvc0_grctx_generate(struct nouveau_channel *chan)
 	nv_icmd(dev, 0x0000053f, 0xffff0000);
 	nv_icmd(dev, 0x00000585, 0x0000003f);
 	nv_icmd(dev, 0x00000576, 0x00000003);
-	if (dev_priv->chipset == 0xc1 ||
-	    dev_priv->chipset == 0xd9)
-		nv_icmd(dev, 0x0000057b, 0x00000059);
 	nv_icmd(dev, 0x00000586, 0x00000040);
 	nv_icmd(dev, 0x00000582, 0x00000080);
 	nv_icmd(dev, 0x00000583, 0x00000080);
@@ -2731,8 +2733,6 @@ nvc0_grctx_generate(struct nouveau_channel *chan)
 	nv_icmd(dev, 0x00000957, 0x00000003);
 	nv_icmd(dev, 0x0000095e, 0x20164010);
 	nv_icmd(dev, 0x0000095f, 0x00000020);
-	if (dev_priv->chipset == 0xd9)
-		nv_icmd(dev, 0x0000097d, 0x00000020);
 	nv_icmd(dev, 0x00000683, 0x00000006);
 	nv_icmd(dev, 0x00000685, 0x003fffff);
 	nv_icmd(dev, 0x00000687, 0x00000c48);
@@ -2865,10 +2865,6 @@ nvc0_grctx_generate(struct nouveau_channel *chan)
 	nv_wr32(dev, 0x404154, 0x00000400);
 
 	nvc0_grctx_generate_9097(dev);
-	if (fermi >= 0x9197)
-		nvc0_grctx_generate_9197(dev);
-	if (fermi >= 0x9297)
-		nvc0_grctx_generate_9297(dev);
 	nvc0_grctx_generate_902d(dev);
 	nvc0_grctx_generate_9039(dev);
 	nvc0_grctx_generate_90c0(dev);
diff --git a/drivers/gpu/drm/nouveau/nvc0_grgpc.fuc b/drivers/gpu/drm/nouveau/nvc0_grgpc.fuc
deleted file mode 100644
index 15272be..0000000
--- a/drivers/gpu/drm/nouveau/nvc0_grgpc.fuc
+++ /dev/null
@@ -1,539 +0,0 @@
-/* fuc microcode for nvc0 PGRAPH/GPC
- *
- * Copyright 2011 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Ben Skeggs
- */
-
-/* To build:
- *    m4 nvc0_grgpc.fuc | envyas -a -w -m fuc -V nva3 -o nvc0_grgpc.fuc.h
- */
-
-/* TODO
- * - bracket certain functions with scratch writes, useful for debugging
- * - watchdog timer around ctx operations
- */
-
-.section #nvc0_grgpc_data
-include(`nvc0_graph.fuc')
-gpc_id:			.b32 0
-gpc_mmio_list_head:	.b32 0
-gpc_mmio_list_tail:	.b32 0
-
-tpc_count:		.b32 0
-tpc_mask:		.b32 0
-tpc_mmio_list_head:	.b32 0
-tpc_mmio_list_tail:	.b32 0
-
-cmd_queue:		queue_init
-
-// chipset descriptions
-chipsets:
-.b8  0xc0 0 0 0
-.b16 #nvc0_gpc_mmio_head
-.b16 #nvc0_gpc_mmio_tail
-.b16 #nvc0_tpc_mmio_head
-.b16 #nvc0_tpc_mmio_tail
-.b8  0xc1 0 0 0
-.b16 #nvc0_gpc_mmio_head
-.b16 #nvc1_gpc_mmio_tail
-.b16 #nvc0_tpc_mmio_head
-.b16 #nvc1_tpc_mmio_tail
-.b8  0xc3 0 0 0
-.b16 #nvc0_gpc_mmio_head
-.b16 #nvc0_gpc_mmio_tail
-.b16 #nvc0_tpc_mmio_head
-.b16 #nvc3_tpc_mmio_tail
-.b8  0xc4 0 0 0
-.b16 #nvc0_gpc_mmio_head
-.b16 #nvc0_gpc_mmio_tail
-.b16 #nvc0_tpc_mmio_head
-.b16 #nvc3_tpc_mmio_tail
-.b8  0xc8 0 0 0
-.b16 #nvc0_gpc_mmio_head
-.b16 #nvc0_gpc_mmio_tail
-.b16 #nvc0_tpc_mmio_head
-.b16 #nvc0_tpc_mmio_tail
-.b8  0xce 0 0 0
-.b16 #nvc0_gpc_mmio_head
-.b16 #nvc0_gpc_mmio_tail
-.b16 #nvc0_tpc_mmio_head
-.b16 #nvc3_tpc_mmio_tail
-.b8  0xcf 0 0 0
-.b16 #nvc0_gpc_mmio_head
-.b16 #nvc0_gpc_mmio_tail
-.b16 #nvc0_tpc_mmio_head
-.b16 #nvcf_tpc_mmio_tail
-.b8  0xd9 0 0 0
-.b16 #nvd9_gpc_mmio_head
-.b16 #nvd9_gpc_mmio_tail
-.b16 #nvd9_tpc_mmio_head
-.b16 #nvd9_tpc_mmio_tail
-.b8  0 0 0 0
-
-// GPC mmio lists
-nvc0_gpc_mmio_head:
-mmctx_data(0x000380, 1)
-mmctx_data(0x000400, 6)
-mmctx_data(0x000450, 9)
-mmctx_data(0x000600, 1)
-mmctx_data(0x000684, 1)
-mmctx_data(0x000700, 5)
-mmctx_data(0x000800, 1)
-mmctx_data(0x000808, 3)
-mmctx_data(0x000828, 1)
-mmctx_data(0x000830, 1)
-mmctx_data(0x0008d8, 1)
-mmctx_data(0x0008e0, 1)
-mmctx_data(0x0008e8, 6)
-mmctx_data(0x00091c, 1)
-mmctx_data(0x000924, 3)
-mmctx_data(0x000b00, 1)
-mmctx_data(0x000b08, 6)
-mmctx_data(0x000bb8, 1)
-mmctx_data(0x000c08, 1)
-mmctx_data(0x000c10, 8)
-mmctx_data(0x000c80, 1)
-mmctx_data(0x000c8c, 1)
-mmctx_data(0x001000, 3)
-mmctx_data(0x001014, 1)
-nvc0_gpc_mmio_tail:
-mmctx_data(0x000c6c, 1);
-nvc1_gpc_mmio_tail:
-
-nvd9_gpc_mmio_head:
-mmctx_data(0x000380, 1)
-mmctx_data(0x000400, 2)
-mmctx_data(0x00040c, 3)
-mmctx_data(0x000450, 9)
-mmctx_data(0x000600, 1)
-mmctx_data(0x000684, 1)
-mmctx_data(0x000700, 5)
-mmctx_data(0x000800, 1)
-mmctx_data(0x000808, 3)
-mmctx_data(0x000828, 1)
-mmctx_data(0x000830, 1)
-mmctx_data(0x0008d8, 1)
-mmctx_data(0x0008e0, 1)
-mmctx_data(0x0008e8, 6)
-mmctx_data(0x00091c, 1)
-mmctx_data(0x000924, 3)
-mmctx_data(0x000b00, 1)
-mmctx_data(0x000b08, 6)
-mmctx_data(0x000bb8, 1)
-mmctx_data(0x000c08, 1)
-mmctx_data(0x000c10, 8)
-mmctx_data(0x000c6c, 1)
-mmctx_data(0x000c80, 1)
-mmctx_data(0x000c8c, 1)
-mmctx_data(0x001000, 3)
-mmctx_data(0x001014, 1)
-nvd9_gpc_mmio_tail:
-
-// TPC mmio lists
-nvc0_tpc_mmio_head:
-mmctx_data(0x000018, 1)
-mmctx_data(0x00003c, 1)
-mmctx_data(0x000048, 1)
-mmctx_data(0x000064, 1)
-mmctx_data(0x000088, 1)
-mmctx_data(0x000200, 6)
-mmctx_data(0x00021c, 2)
-mmctx_data(0x000300, 6)
-mmctx_data(0x0003d0, 1)
-mmctx_data(0x0003e0, 2)
-mmctx_data(0x000400, 3)
-mmctx_data(0x000420, 1)
-mmctx_data(0x0004b0, 1)
-mmctx_data(0x0004e8, 1)
-mmctx_data(0x0004f4, 1)
-mmctx_data(0x000520, 2)
-mmctx_data(0x000604, 4)
-mmctx_data(0x000644, 20)
-mmctx_data(0x000698, 1)
-mmctx_data(0x000750, 2)
-nvc0_tpc_mmio_tail:
-mmctx_data(0x000758, 1)
-mmctx_data(0x0002c4, 1)
-mmctx_data(0x0006e0, 1)
-nvcf_tpc_mmio_tail:
-mmctx_data(0x0004bc, 1)
-nvc3_tpc_mmio_tail:
-mmctx_data(0x000544, 1)
-nvc1_tpc_mmio_tail:
-
-nvd9_tpc_mmio_head:
-mmctx_data(0x000018, 1)
-mmctx_data(0x00003c, 1)
-mmctx_data(0x000048, 1)
-mmctx_data(0x000064, 1)
-mmctx_data(0x000088, 1)
-mmctx_data(0x000200, 6)
-mmctx_data(0x00021c, 2)
-mmctx_data(0x0002c4, 1)
-mmctx_data(0x000300, 6)
-mmctx_data(0x0003d0, 1)
-mmctx_data(0x0003e0, 2)
-mmctx_data(0x000400, 3)
-mmctx_data(0x000420, 3)
-mmctx_data(0x0004b0, 1)
-mmctx_data(0x0004e8, 1)
-mmctx_data(0x0004f4, 1)
-mmctx_data(0x000520, 2)
-mmctx_data(0x000544, 1)
-mmctx_data(0x000604, 4)
-mmctx_data(0x000644, 20)
-mmctx_data(0x000698, 1)
-mmctx_data(0x0006e0, 1)
-mmctx_data(0x000750, 3)
-nvd9_tpc_mmio_tail:
-
-.section #nvc0_grgpc_code
-bra #init
-define(`include_code')
-include(`nvc0_graph.fuc')
-
-// reports an exception to the host
-//
-// In: $r15 error code (see nvc0_graph.fuc)
-//
-error:
-	push $r14
-	mov $r14 -0x67ec 	// 0x9814
-	sethi $r14 0x400000
-	call #nv_wr32		// HUB_CTXCTL_CC_SCRATCH[5] = error code
-	add b32 $r14 0x41c
-	mov $r15 1
-	call #nv_wr32		// HUB_CTXCTL_INTR_UP_SET
-	pop $r14
-	ret
-
-// GPC fuc initialisation, executed by triggering ucode start, will
-// fall through to main loop after completion.
-//
-// Input:
-//   CC_SCRATCH[0]: chipset (PMC_BOOT_0 read returns 0x0bad0bad... sigh)
-//   CC_SCRATCH[1]: context base
-//
-// Output:
-//   CC_SCRATCH[0]:
-//	     31:31: set to signal completion
-//   CC_SCRATCH[1]:
-//	      31:0: GPC context size
-//
-init:
-	clear b32 $r0
-	mov $sp $r0
-
-	// enable fifo access
-	mov $r1 0x1200
-	mov $r2 2
-	iowr I[$r1 + 0x000] $r2		// FIFO_ENABLE
-
-	// setup i0 handler, and route all interrupts to it
-	mov $r1 #ih
-	mov $iv0 $r1
-	mov $r1 0x400
-	iowr I[$r1 + 0x300] $r0		// INTR_DISPATCH
-
-	// enable fifo interrupt
-	mov $r2 4
-	iowr I[$r1 + 0x000] $r2		// INTR_EN_SET
-
-	// enable interrupts
-	bset $flags ie0
-
-	// figure out which GPC we are, and how many TPCs we have
-	mov $r1 0x608
-	shl b32 $r1 6
-	iord $r2 I[$r1 + 0x000]		// UNITS
-	mov $r3 1
-	and $r2 0x1f
-	shl b32 $r3 $r2
-	sub b32 $r3 1
-	st b32 D[$r0 + #tpc_count] $r2
-	st b32 D[$r0 + #tpc_mask] $r3
-	add b32 $r1 0x400
-	iord $r2 I[$r1 + 0x000]		// MYINDEX
-	st b32 D[$r0 + #gpc_id] $r2
-
-	// find context data for this chipset
-	mov $r2 0x800
-	shl b32 $r2 6
-	iord $r2 I[$r2 + 0x000]		// CC_SCRATCH[0]
-	mov $r1 #chipsets - 12
-	init_find_chipset:
-		add b32 $r1 12
-		ld b32 $r3 D[$r1 + 0x00]
-		cmpu b32 $r3 $r2
-		bra e #init_context
-		cmpu b32 $r3 0
-		bra ne #init_find_chipset
-		// unknown chipset
-		ret
-
-	// initialise context base, and size tracking
-	init_context:
-	mov $r2 0x800
-	shl b32 $r2 6
-	iord $r2 I[$r2 + 0x100]	// CC_SCRATCH[1], initial base
-	clear b32 $r3		// track GPC context size here
-
-	// set mmctx base addresses now so we don't have to do it later,
-	// they don't currently ever change
-	mov $r4 0x700
-	shl b32 $r4 6
-	shr b32 $r5 $r2 8
-	iowr I[$r4 + 0x000] $r5		// MMCTX_SAVE_SWBASE
-	iowr I[$r4 + 0x100] $r5		// MMCTX_LOAD_SWBASE
-
-	// calculate GPC mmio context size, store the chipset-specific
-	// mmio list pointers somewhere we can get at them later without
-	// re-parsing the chipset list
-	clear b32 $r14
-	clear b32 $r15
-	ld b16 $r14 D[$r1 + 4]
-	ld b16 $r15 D[$r1 + 6]
-	st b16 D[$r0 + #gpc_mmio_list_head] $r14
-	st b16 D[$r0 + #gpc_mmio_list_tail] $r15
-	call #mmctx_size
-	add b32 $r2 $r15
-	add b32 $r3 $r15
-
-	// calculate per-TPC mmio context size, store the list pointers
-	ld b16 $r14 D[$r1 + 8]
-	ld b16 $r15 D[$r1 + 10]
-	st b16 D[$r0 + #tpc_mmio_list_head] $r14
-	st b16 D[$r0 + #tpc_mmio_list_tail] $r15
-	call #mmctx_size
-	ld b32 $r14 D[$r0 + #tpc_count]
-	mulu $r14 $r15
-	add b32 $r2 $r14
-	add b32 $r3 $r14
-
-	// round up base/size to 256 byte boundary (for strand SWBASE)
-	add b32 $r4 0x1300
-	shr b32 $r3 2
-	iowr I[$r4 + 0x000] $r3		// MMCTX_LOAD_COUNT, wtf for?!?
-	shr b32 $r2 8
-	shr b32 $r3 6
-	add b32 $r2 1
-	add b32 $r3 1
-	shl b32 $r2 8
-	shl b32 $r3 8
-
-	// calculate size of strand context data
-	mov b32 $r15 $r2
-	call #strand_ctx_init
-	add b32 $r3 $r15
-
-	// save context size, and tell HUB we're done
-	mov $r1 0x800
-	shl b32 $r1 6
-	iowr I[$r1 + 0x100] $r3		// CC_SCRATCH[1]  = context size
-	add b32 $r1 0x800
-	clear b32 $r2
-	bset $r2 31
-	iowr I[$r1 + 0x000] $r2		// CC_SCRATCH[0] |= 0x80000000
-
-// Main program loop, very simple, sleeps until woken up by the interrupt
-// handler, pulls a command from the queue and executes its handler
-//
-main:
-	bset $flags $p0
-	sleep $p0
-	mov $r13 #cmd_queue
-	call #queue_get
-	bra $p1 #main
-
-	// 0x0000-0x0003 are all context transfers
-	cmpu b32 $r14 0x04
-	bra nc #main_not_ctx_xfer
-		// fetch $flags and mask off $p1/$p2
-		mov $r1 $flags
-		mov $r2 0x0006
-		not b32 $r2
-		and $r1 $r2
-		// set $p1/$p2 according to transfer type
-		shl b32 $r14 1
-		or $r1 $r14
-		mov $flags $r1
-		// transfer context data
-		call #ctx_xfer
-		bra #main
-
-	main_not_ctx_xfer:
-	shl b32 $r15 $r14 16
-	or $r15 E_BAD_COMMAND
-	call #error
-	bra #main
-
-// interrupt handler
-ih:
-	push $r8
-	mov $r8 $flags
-	push $r8
-	push $r9
-	push $r10
-	push $r11
-	push $r13
-	push $r14
-	push $r15
-
-	// incoming fifo command?
-	iord $r10 I[$r0 + 0x200]	// INTR
-	and $r11 $r10 0x00000004
-	bra e #ih_no_fifo
-		// queue incoming fifo command for later processing
-		mov $r11 0x1900
-		mov $r13 #cmd_queue
-		iord $r14 I[$r11 + 0x100]	// FIFO_CMD
-		iord $r15 I[$r11 + 0x000]	// FIFO_DATA
-		call #queue_put
-		add b32 $r11 0x400
-		mov $r14 1
-		iowr I[$r11 + 0x000] $r14	// FIFO_ACK
-
-	// ack, and wake up main()
-	ih_no_fifo:
-	iowr I[$r0 + 0x100] $r10	// INTR_ACK
-
-	pop $r15
-	pop $r14
-	pop $r13
-	pop $r11
-	pop $r10
-	pop $r9
-	pop $r8
-	mov $flags $r8
-	pop $r8
-	bclr $flags $p0
-	iret
-
-// Set this GPC's bit in HUB_BAR, used to signal completion of various
-// activities to the HUB fuc
-//
-hub_barrier_done:
-	mov $r15 1
-	ld b32 $r14 D[$r0 + #gpc_id]
-	shl b32 $r15 $r14
-	mov $r14 -0x6be8 	// 0x409418 - HUB_BAR_SET
-	sethi $r14 0x400000
-	call #nv_wr32
-	ret
-
-// Disables various things, waits a bit, and re-enables them..
-//
-// Not sure how exactly this helps, perhaps "ENABLE" is not such a
-// good description for the bits we turn off?  Anyways, without this,
-// funny things happen.
-//
-ctx_redswitch:
-	mov $r14 0x614
-	shl b32 $r14 6
-	mov $r15 0x020
-	iowr I[$r14] $r15	// GPC_RED_SWITCH = POWER
-	mov $r15 8
-	ctx_redswitch_delay:
-		sub b32 $r15 1
-		bra ne #ctx_redswitch_delay
-	mov $r15 0xa20
-	iowr I[$r14] $r15	// GPC_RED_SWITCH = UNK11, ENABLE, POWER
-	ret
-
-// Transfer GPC context data between GPU and storage area
-//
-// In: $r15 context base address
-//     $p1 clear on save, set on load
-//     $p2 set if opposite direction done/will be done, so:
-//		on save it means: "a load will follow this save"
-//		on load it means: "a save preceeded this load"
-//
-ctx_xfer:
-	// set context base address
-	mov $r1 0xa04
-	shl b32 $r1 6
-	iowr I[$r1 + 0x000] $r15// MEM_BASE
-	bra not $p1 #ctx_xfer_not_load
-		call #ctx_redswitch
-	ctx_xfer_not_load:
-
-	// strands
-	mov $r1 0x4afc
-	sethi $r1 0x20000
-	mov $r2 0xc
-	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x0c
-	call #strand_wait
-	mov $r2 0x47fc
-	sethi $r2 0x20000
-	iowr I[$r2] $r0		// STRAND_FIRST_GENE(0x3f) = 0x00
-	xbit $r2 $flags $p1
-	add b32 $r2 3
-	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x03/0x04 (SAVE/LOAD)
-
-	// mmio context
-	xbit $r10 $flags $p1	// direction
-	or $r10 2		// first
-	mov $r11 0x0000
-	sethi $r11 0x500000
-	ld b32 $r12 D[$r0 + #gpc_id]
-	shl b32 $r12 15
-	add b32 $r11 $r12	// base = NV_PGRAPH_GPCn
-	ld b32 $r12 D[$r0 + #gpc_mmio_list_head]
-	ld b32 $r13 D[$r0 + #gpc_mmio_list_tail]
-	mov $r14 0		// not multi
-	call #mmctx_xfer
-
-	// per-TPC mmio context
-	xbit $r10 $flags $p1	// direction
-	or $r10 4		// last
-	mov $r11 0x4000
-	sethi $r11 0x500000	// base = NV_PGRAPH_GPC0_TPC0
-	ld b32 $r12 D[$r0 + #gpc_id]
-	shl b32 $r12 15
-	add b32 $r11 $r12	// base = NV_PGRAPH_GPCn_TPC0
-	ld b32 $r12 D[$r0 + #tpc_mmio_list_head]
-	ld b32 $r13 D[$r0 + #tpc_mmio_list_tail]
-	ld b32 $r15 D[$r0 + #tpc_mask]
-	mov $r14 0x800		// stride = 0x800
-	call #mmctx_xfer
-
-	// wait for strands to finish
-	call #strand_wait
-
-	// if load, or a save without a load following, do some
-	// unknown stuff that's done after finishing a block of
-	// strand commands
-	bra $p1 #ctx_xfer_post
-	bra not $p2 #ctx_xfer_done
-	ctx_xfer_post:
-		mov $r1 0x4afc
-		sethi $r1 0x20000
-		mov $r2 0xd
-		iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x0d
-		call #strand_wait
-
-	// mark completion in HUB's barrier
-	ctx_xfer_done:
-	call #hub_barrier_done
-	ret
-
-.align 256
diff --git a/drivers/gpu/drm/nouveau/nvc0_grgpc.fuc.h b/drivers/gpu/drm/nouveau/nvc0_grgpc.fuc.h
deleted file mode 100644
index a988b8a..0000000
--- a/drivers/gpu/drm/nouveau/nvc0_grgpc.fuc.h
+++ /dev/null
@@ -1,538 +0,0 @@
-uint32_t nvc0_grgpc_data[] = {
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x000000c0,
-	0x012800c8,
-	0x01e40194,
-	0x000000c1,
-	0x012c00c8,
-	0x01f80194,
-	0x000000c3,
-	0x012800c8,
-	0x01f40194,
-	0x000000c4,
-	0x012800c8,
-	0x01f40194,
-	0x000000c8,
-	0x012800c8,
-	0x01e40194,
-	0x000000ce,
-	0x012800c8,
-	0x01f40194,
-	0x000000cf,
-	0x012800c8,
-	0x01f00194,
-	0x000000d9,
-	0x0194012c,
-	0x025401f8,
-	0x00000000,
-	0x00000380,
-	0x14000400,
-	0x20000450,
-	0x00000600,
-	0x00000684,
-	0x10000700,
-	0x00000800,
-	0x08000808,
-	0x00000828,
-	0x00000830,
-	0x000008d8,
-	0x000008e0,
-	0x140008e8,
-	0x0000091c,
-	0x08000924,
-	0x00000b00,
-	0x14000b08,
-	0x00000bb8,
-	0x00000c08,
-	0x1c000c10,
-	0x00000c80,
-	0x00000c8c,
-	0x08001000,
-	0x00001014,
-	0x00000c6c,
-	0x00000380,
-	0x04000400,
-	0x0800040c,
-	0x20000450,
-	0x00000600,
-	0x00000684,
-	0x10000700,
-	0x00000800,
-	0x08000808,
-	0x00000828,
-	0x00000830,
-	0x000008d8,
-	0x000008e0,
-	0x140008e8,
-	0x0000091c,
-	0x08000924,
-	0x00000b00,
-	0x14000b08,
-	0x00000bb8,
-	0x00000c08,
-	0x1c000c10,
-	0x00000c6c,
-	0x00000c80,
-	0x00000c8c,
-	0x08001000,
-	0x00001014,
-	0x00000018,
-	0x0000003c,
-	0x00000048,
-	0x00000064,
-	0x00000088,
-	0x14000200,
-	0x0400021c,
-	0x14000300,
-	0x000003d0,
-	0x040003e0,
-	0x08000400,
-	0x00000420,
-	0x000004b0,
-	0x000004e8,
-	0x000004f4,
-	0x04000520,
-	0x0c000604,
-	0x4c000644,
-	0x00000698,
-	0x04000750,
-	0x00000758,
-	0x000002c4,
-	0x000006e0,
-	0x000004bc,
-	0x00000544,
-	0x00000018,
-	0x0000003c,
-	0x00000048,
-	0x00000064,
-	0x00000088,
-	0x14000200,
-	0x0400021c,
-	0x000002c4,
-	0x14000300,
-	0x000003d0,
-	0x040003e0,
-	0x08000400,
-	0x08000420,
-	0x000004b0,
-	0x000004e8,
-	0x000004f4,
-	0x04000520,
-	0x00000544,
-	0x0c000604,
-	0x4c000644,
-	0x00000698,
-	0x000006e0,
-	0x08000750,
-};
-
-uint32_t nvc0_grgpc_code[] = {
-	0x03060ef5,
-	0x9800d898,
-	0x86f001d9,
-	0x0489b808,
-	0xf00c1bf4,
-	0x21f502f7,
-	0x00f802ec,
-	0xb60798c4,
-	0x8dbb0384,
-	0x0880b600,
-	0x80008e80,
-	0x90b6018f,
-	0x0f94f001,
-	0xf801d980,
-	0x0131f400,
-	0x9800d898,
-	0x89b801d9,
-	0x210bf404,
-	0xb60789c4,
-	0x9dbb0394,
-	0x0890b600,
-	0x98009e98,
-	0x80b6019f,
-	0x0f84f001,
-	0xf400d880,
-	0x00f80132,
-	0x0728b7f1,
-	0xb906b4b6,
-	0xc9f002ec,
-	0x00bcd01f,
-	0xc800bccf,
-	0x1bf41fcc,
-	0x06a7f0fa,
-	0x010321f5,
-	0xf840bfcf,
-	0x28b7f100,
-	0x06b4b607,
-	0xb980bfd0,
-	0xc9f002ec,
-	0x1ec9f01f,
-	0xcf00bcd0,
-	0xccc800bc,
-	0xfa1bf41f,
-	0x87f100f8,
-	0x84b60430,
-	0x1ff9f006,
-	0xf8008fd0,
-	0x3087f100,
-	0x0684b604,
-	0xf80080d0,
-	0x3c87f100,
-	0x0684b608,
-	0x99f094bd,
-	0x0089d000,
-	0x081887f1,
-	0xd00684b6,
-	0x87f1008a,
-	0x84b60400,
-	0x0088cf06,
-	0xf4888aff,
-	0x87f1f31b,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00099,
-	0xf100f800,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00099f0,
-	0x87f10089,
-	0x84b60818,
-	0x008ad006,
-	0x040087f1,
-	0xcf0684b6,
-	0x8aff0088,
-	0xf30bf488,
-	0x085c87f1,
-	0xbd0684b6,
-	0x0099f094,
-	0xf80089d0,
-	0x9894bd00,
-	0x85b600e8,
-	0x0180b61a,
-	0xbb0284b6,
-	0xe0b60098,
-	0x04efb804,
-	0xb9eb1bf4,
-	0x00f8029f,
-	0x083c87f1,
-	0xbd0684b6,
-	0x0199f094,
-	0xf10089d0,
-	0xb6071087,
-	0x94bd0684,
-	0xf405bbfd,
-	0x8bd0090b,
-	0x0099f000,
-	0xf405eefd,
-	0x8ed00c0b,
-	0xc08fd080,
-	0xb70199f0,
-	0xc8010080,
-	0xb4b600ab,
-	0x0cb9f010,
-	0xb601aec8,
-	0xbefd11e4,
-	0x008bd005,
-	0xf0008ecf,
-	0x0bf41fe4,
-	0x00ce98fa,
-	0xd005e9fd,
-	0xc0b6c08e,
-	0x04cdb804,
-	0xc8e81bf4,
-	0x1bf402ab,
-	0x008bcf18,
-	0xb01fb4f0,
-	0x1bf410b4,
-	0x02a7f0f7,
-	0xf4c921f4,
-	0xabc81b0e,
-	0x10b4b600,
-	0xf00cb9f0,
-	0x8bd012b9,
-	0x008bcf00,
-	0xf412bbc8,
-	0x87f1fa1b,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00199,
-	0xf900f800,
-	0x02a7f0a0,
-	0xfcc921f4,
-	0xf100f8a0,
-	0xf04afc87,
-	0x97f00283,
-	0x0089d00c,
-	0x020721f5,
-	0x87f100f8,
-	0x83f04afc,
-	0x0d97f002,
-	0xf50089d0,
-	0xf8020721,
-	0xfca7f100,
-	0x02a3f04f,
-	0x0500aba2,
-	0xd00fc7f0,
-	0xc7f000ac,
-	0x00bcd00b,
-	0x020721f5,
-	0xf000aed0,
-	0xbcd00ac7,
-	0x0721f500,
-	0xf100f802,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00399f0,
-	0x21f50089,
-	0xe7f00213,
-	0x3921f503,
-	0xfca7f102,
-	0x02a3f046,
-	0x0400aba0,
-	0xf040a0d0,
-	0xbcd001c7,
-	0x0721f500,
-	0x010c9202,
-	0xf000acd0,
-	0xbcd002c7,
-	0x0721f500,
-	0x2621f502,
-	0x8087f102,
-	0x0684b608,
-	0xb70089cf,
-	0x95220080,
-	0x8ed008fe,
-	0x408ed000,
-	0xb6808acf,
-	0xa0b606a5,
-	0x00eabb01,
-	0xb60480b6,
-	0x1bf40192,
-	0x08e4b6e8,
-	0xf1f2efbc,
-	0xb6085c87,
-	0x94bd0684,
-	0xd00399f0,
-	0x00f80089,
-	0xe7f1e0f9,
-	0xe3f09814,
-	0x8d21f440,
-	0x041ce0b7,
-	0xf401f7f0,
-	0xe0fc8d21,
-	0x04bd00f8,
-	0xf10004fe,
-	0xf0120017,
-	0x12d00227,
-	0x3e17f100,
-	0x0010fe04,
-	0x040017f1,
-	0xf0c010d0,
-	0x12d00427,
-	0x1031f400,
-	0x060817f1,
-	0xcf0614b6,
-	0x37f00012,
-	0x1f24f001,
-	0xb60432bb,
-	0x02800132,
-	0x04038003,
-	0x040010b7,
-	0x800012cf,
-	0x27f10002,
-	0x24b60800,
-	0x0022cf06,
-	0xb65817f0,
-	0x13980c10,
-	0x0432b800,
-	0xb00b0bf4,
-	0x1bf40034,
-	0xf100f8f1,
-	0xb6080027,
-	0x22cf0624,
-	0xf134bd40,
-	0xb6070047,
-	0x25950644,
-	0x0045d008,
-	0xbd4045d0,
-	0x58f4bde4,
-	0x1f58021e,
-	0x020e4003,
-	0xf5040f40,
-	0xbb013d21,
-	0x3fbb002f,
-	0x041e5800,
-	0x40051f58,
-	0x0f400a0e,
-	0x3d21f50c,
-	0x030e9801,
-	0xbb00effd,
-	0x3ebb002e,
-	0x0040b700,
-	0x0235b613,
-	0xb60043d0,
-	0x35b60825,
-	0x0120b606,
-	0xb60130b6,
-	0x34b60824,
-	0x022fb908,
-	0x026321f5,
-	0xf1003fbb,
-	0xb6080017,
-	0x13d00614,
-	0x0010b740,
-	0xf024bd08,
-	0x12d01f29,
-	0x0031f400,
-	0xf00028f4,
-	0x21f41cd7,
-	0xf401f439,
-	0xf404e4b0,
-	0x81fe1e18,
-	0x0627f001,
-	0x12fd20bd,
-	0x01e4b604,
-	0xfe051efd,
-	0x21f50018,
-	0x0ef404c3,
-	0x10ef94d3,
-	0xf501f5f0,
-	0xf402ec21,
-	0x80f9c60e,
-	0xf90188fe,
-	0xf990f980,
-	0xf9b0f9a0,
-	0xf9e0f9d0,
-	0x800acff0,
-	0xf404abc4,
-	0xb7f11d0b,
-	0xd7f01900,
-	0x40becf1c,
-	0xf400bfcf,
-	0xb0b70421,
-	0xe7f00400,
-	0x00bed001,
-	0xfc400ad0,
-	0xfce0fcf0,
-	0xfcb0fcd0,
-	0xfc90fca0,
-	0x0088fe80,
-	0x32f480fc,
-	0xf001f800,
-	0x0e9801f7,
-	0x04febb00,
-	0x9418e7f1,
-	0xf440e3f0,
-	0x00f88d21,
-	0x0614e7f1,
-	0xf006e4b6,
-	0xefd020f7,
-	0x08f7f000,
-	0xf401f2b6,
-	0xf7f1fd1b,
-	0xefd00a20,
-	0xf100f800,
-	0xb60a0417,
-	0x1fd00614,
-	0x0711f400,
-	0x04a421f5,
-	0x4afc17f1,
-	0xf00213f0,
-	0x12d00c27,
-	0x0721f500,
-	0xfc27f102,
-	0x0223f047,
-	0xf00020d0,
-	0x20b6012c,
-	0x0012d003,
-	0xf001acf0,
-	0xb7f002a5,
-	0x50b3f000,
-	0xb6000c98,
-	0xbcbb0fc4,
-	0x010c9800,
-	0xf0020d98,
-	0x21f500e7,
-	0xacf0015c,
-	0x04a5f001,
-	0x4000b7f1,
-	0x9850b3f0,
-	0xc4b6000c,
-	0x00bcbb0f,
-	0x98050c98,
-	0x0f98060d,
-	0x00e7f104,
-	0x5c21f508,
-	0x0721f501,
-	0x0601f402,
-	0xf11412f4,
-	0xf04afc17,
-	0x27f00213,
-	0x0012d00d,
-	0x020721f5,
-	0x048f21f5,
-	0x000000f8,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-};
diff --git a/drivers/gpu/drm/nouveau/nvc0_grhub.fuc b/drivers/gpu/drm/nouveau/nvc0_grhub.fuc
deleted file mode 100644
index 98acddb..0000000
--- a/drivers/gpu/drm/nouveau/nvc0_grhub.fuc
+++ /dev/null
@@ -1,856 +0,0 @@
-/* fuc microcode for nvc0 PGRAPH/HUB
- *
- * Copyright 2011 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Ben Skeggs
- */
-
-/* To build:
- *    m4 nvc0_grhub.fuc | envyas -a -w -m fuc -V nva3 -o nvc0_grhub.fuc.h
- */
-
-.section #nvc0_grhub_data
-include(`nvc0_graph.fuc')
-gpc_count:		.b32 0
-rop_count:		.b32 0
-cmd_queue:		queue_init
-hub_mmio_list_head:	.b32 0
-hub_mmio_list_tail:	.b32 0
-
-ctx_current:		.b32 0
-
-chipsets:
-.b8  0xc0 0 0 0
-.b16 #nvc0_hub_mmio_head
-.b16 #nvc0_hub_mmio_tail
-.b8  0xc1 0 0 0
-.b16 #nvc0_hub_mmio_head
-.b16 #nvc1_hub_mmio_tail
-.b8  0xc3 0 0 0
-.b16 #nvc0_hub_mmio_head
-.b16 #nvc0_hub_mmio_tail
-.b8  0xc4 0 0 0
-.b16 #nvc0_hub_mmio_head
-.b16 #nvc0_hub_mmio_tail
-.b8  0xc8 0 0 0
-.b16 #nvc0_hub_mmio_head
-.b16 #nvc0_hub_mmio_tail
-.b8  0xce 0 0 0
-.b16 #nvc0_hub_mmio_head
-.b16 #nvc0_hub_mmio_tail
-.b8  0xcf 0 0 0
-.b16 #nvc0_hub_mmio_head
-.b16 #nvc0_hub_mmio_tail
-.b8  0xd9 0 0 0
-.b16 #nvd9_hub_mmio_head
-.b16 #nvd9_hub_mmio_tail
-.b8  0 0 0 0
-
-nvc0_hub_mmio_head:
-mmctx_data(0x17e91c, 2)
-mmctx_data(0x400204, 2)
-mmctx_data(0x404004, 11)
-mmctx_data(0x404044, 1)
-mmctx_data(0x404094, 14)
-mmctx_data(0x4040d0, 7)
-mmctx_data(0x4040f8, 1)
-mmctx_data(0x404130, 3)
-mmctx_data(0x404150, 3)
-mmctx_data(0x404164, 2)
-mmctx_data(0x404174, 3)
-mmctx_data(0x404200, 8)
-mmctx_data(0x404404, 14)
-mmctx_data(0x404460, 4)
-mmctx_data(0x404480, 1)
-mmctx_data(0x404498, 1)
-mmctx_data(0x404604, 4)
-mmctx_data(0x404618, 32)
-mmctx_data(0x404698, 21)
-mmctx_data(0x4046f0, 2)
-mmctx_data(0x404700, 22)
-mmctx_data(0x405800, 1)
-mmctx_data(0x405830, 3)
-mmctx_data(0x405854, 1)
-mmctx_data(0x405870, 4)
-mmctx_data(0x405a00, 2)
-mmctx_data(0x405a18, 1)
-mmctx_data(0x406020, 1)
-mmctx_data(0x406028, 4)
-mmctx_data(0x4064a8, 2)
-mmctx_data(0x4064b4, 2)
-mmctx_data(0x407804, 1)
-mmctx_data(0x40780c, 6)
-mmctx_data(0x4078bc, 1)
-mmctx_data(0x408000, 7)
-mmctx_data(0x408064, 1)
-mmctx_data(0x408800, 3)
-mmctx_data(0x408900, 4)
-mmctx_data(0x408980, 1)
-nvc0_hub_mmio_tail:
-mmctx_data(0x4064c0, 2)
-nvc1_hub_mmio_tail:
-
-nvd9_hub_mmio_head:
-mmctx_data(0x17e91c, 2)
-mmctx_data(0x400204, 2)
-mmctx_data(0x404004, 10)
-mmctx_data(0x404044, 1)
-mmctx_data(0x404094, 14)
-mmctx_data(0x4040d0, 7)
-mmctx_data(0x4040f8, 1)
-mmctx_data(0x404130, 3)
-mmctx_data(0x404150, 3)
-mmctx_data(0x404164, 2)
-mmctx_data(0x404178, 2)
-mmctx_data(0x404200, 8)
-mmctx_data(0x404404, 14)
-mmctx_data(0x404460, 4)
-mmctx_data(0x404480, 1)
-mmctx_data(0x404498, 1)
-mmctx_data(0x404604, 4)
-mmctx_data(0x404618, 32)
-mmctx_data(0x404698, 21)
-mmctx_data(0x4046f0, 2)
-mmctx_data(0x404700, 22)
-mmctx_data(0x405800, 1)
-mmctx_data(0x405830, 3)
-mmctx_data(0x405854, 1)
-mmctx_data(0x405870, 4)
-mmctx_data(0x405a00, 2)
-mmctx_data(0x405a18, 1)
-mmctx_data(0x406020, 1)
-mmctx_data(0x406028, 4)
-mmctx_data(0x4064a8, 2)
-mmctx_data(0x4064b4, 5)
-mmctx_data(0x407804, 1)
-mmctx_data(0x40780c, 6)
-mmctx_data(0x4078bc, 1)
-mmctx_data(0x408000, 7)
-mmctx_data(0x408064, 1)
-mmctx_data(0x408800, 3)
-mmctx_data(0x408900, 4)
-mmctx_data(0x408980, 1)
-nvd9_hub_mmio_tail:
-
-.align 256
-chan_data:
-chan_mmio_count:	.b32 0
-chan_mmio_address:	.b32 0
-
-.align 256
-xfer_data: 		.b32 0
-
-.section #nvc0_grhub_code
-bra #init
-define(`include_code')
-include(`nvc0_graph.fuc')
-
-// reports an exception to the host
-//
-// In: $r15 error code (see nvc0_graph.fuc)
-//
-error:
-	push $r14
-	mov $r14 0x814
-	shl b32 $r14 6
-	iowr I[$r14 + 0x000] $r15	// CC_SCRATCH[5] = error code
-	mov $r14 0xc1c
-	shl b32 $r14 6
-	mov $r15 1
-	iowr I[$r14 + 0x000] $r15	// INTR_UP_SET
-	pop $r14
-	ret
-
-// HUB fuc initialisation, executed by triggering ucode start, will
-// fall through to main loop after completion.
-//
-// Input:
-//   CC_SCRATCH[0]: chipset (PMC_BOOT_0 read returns 0x0bad0bad... sigh)
-//
-// Output:
-//   CC_SCRATCH[0]:
-//	     31:31: set to signal completion
-//   CC_SCRATCH[1]:
-//	      31:0: total PGRAPH context size
-//
-init:
-	clear b32 $r0
-	mov $sp $r0
-	mov $xdbase $r0
-
-	// enable fifo access
-	mov $r1 0x1200
-	mov $r2 2
-	iowr I[$r1 + 0x000] $r2	// FIFO_ENABLE
-
-	// setup i0 handler, and route all interrupts to it
-	mov $r1 #ih
-	mov $iv0 $r1
-	mov $r1 0x400
-	iowr I[$r1 + 0x300] $r0	// INTR_DISPATCH
-
-	// route HUB_CHANNEL_SWITCH to fuc interrupt 8
-	mov $r3 0x404
-	shl b32 $r3 6
-	mov $r2 0x2003		// { HUB_CHANNEL_SWITCH, ZERO } -> intr 8
-	iowr I[$r3 + 0x000] $r2
-
-	// not sure what these are, route them because NVIDIA does, and
-	// the IRQ handler will signal the host if we ever get one.. we
-	// may find out if/why we need to handle these if so..
-	//
-	mov $r2 0x2004
-	iowr I[$r3 + 0x004] $r2 // { 0x04, ZERO } -> intr 9
-	mov $r2 0x200b
-	iowr I[$r3 + 0x008] $r2 // { 0x0b, ZERO } -> intr 10
-	mov $r2 0x200c
-	iowr I[$r3 + 0x01c] $r2 // { 0x0c, ZERO } -> intr 15
-
-	// enable all INTR_UP interrupts
-	mov $r2 0xc24
-	shl b32 $r2 6
-	not b32 $r3 $r0
-	iowr I[$r2] $r3
-
-	// enable fifo, ctxsw, 9, 10, 15 interrupts
-	mov $r2 -0x78fc		// 0x8704
-	sethi $r2 0
-	iowr I[$r1 + 0x000] $r2	// INTR_EN_SET
-
-	// fifo level triggered, rest edge
-	sub b32 $r1 0x100
-	mov $r2 4
-	iowr I[$r1] $r2
-
-	// enable interrupts
-	bset $flags ie0
-
-	// fetch enabled GPC/ROP counts
-	mov $r14 -0x69fc	// 0x409604
-	sethi $r14 0x400000
-	call #nv_rd32
-	extr $r1 $r15 16:20
-	st b32 D[$r0 + #rop_count] $r1
-	and $r15 0x1f
-	st b32 D[$r0 + #gpc_count] $r15
-
-	// set BAR_REQMASK to GPC mask
-	mov $r1 1
-	shl b32 $r1 $r15
-	sub b32 $r1 1
-	mov $r2 0x40c
-	shl b32 $r2 6
-	iowr I[$r2 + 0x000] $r1
-	iowr I[$r2 + 0x100] $r1
-
-	// find context data for this chipset
-	mov $r2 0x800
-	shl b32 $r2 6
-	iord $r2 I[$r2 + 0x000]		// CC_SCRATCH[0]
-	mov $r15 #chipsets - 8
-	init_find_chipset:
-		add b32 $r15 8
-		ld b32 $r3 D[$r15 + 0x00]
-		cmpu b32 $r3 $r2
-		bra e #init_context
-		cmpu b32 $r3 0
-		bra ne #init_find_chipset
-		// unknown chipset
-		ret
-
-	// context size calculation, reserve first 256 bytes for use by fuc
-	init_context:
-	mov $r1 256
-
-	// calculate size of mmio context data
-	ld b16 $r14 D[$r15 + 4]
-	ld b16 $r15 D[$r15 + 6]
-	sethi $r14 0
-	st b32 D[$r0 + #hub_mmio_list_head] $r14
-	st b32 D[$r0 + #hub_mmio_list_tail] $r15
-	call #mmctx_size
-
-	// set mmctx base addresses now so we don't have to do it later,
-	// they don't (currently) ever change
-	mov $r3 0x700
-	shl b32 $r3 6
-	shr b32 $r4 $r1 8
-	iowr I[$r3 + 0x000] $r4		// MMCTX_SAVE_SWBASE
-	iowr I[$r3 + 0x100] $r4		// MMCTX_LOAD_SWBASE
-	add b32 $r3 0x1300
-	add b32 $r1 $r15
-	shr b32 $r15 2
-	iowr I[$r3 + 0x000] $r15	// MMCTX_LOAD_COUNT, wtf for?!?
-
-	// strands, base offset needs to be aligned to 256 bytes
-	shr b32 $r1 8
-	add b32 $r1 1
-	shl b32 $r1 8
-	mov b32 $r15 $r1
-	call #strand_ctx_init
-	add b32 $r1 $r15
-
-	// initialise each GPC in sequence by passing in the offset of its
-	// context data in GPCn_CC_SCRATCH[1], and starting its FUC (which
-	// has previously been uploaded by the host) running.
-	//
-	// the GPC fuc init sequence will set GPCn_CC_SCRATCH[0] bit 31
-	// when it has completed, and return the size of its context data
-	// in GPCn_CC_SCRATCH[1]
-	//
-	ld b32 $r3 D[$r0 + #gpc_count]
-	mov $r4 0x2000
-	sethi $r4 0x500000
-	init_gpc:
-		// setup, and start GPC ucode running
-		add b32 $r14 $r4 0x804
-		mov b32 $r15 $r1
-		call #nv_wr32			// CC_SCRATCH[1] = ctx offset
-		add b32 $r14 $r4 0x800
-		mov b32 $r15 $r2
-		call #nv_wr32			// CC_SCRATCH[0] = chipset
-		add b32 $r14 $r4 0x10c
-		clear b32 $r15
-		call #nv_wr32
-		add b32 $r14 $r4 0x104
-		call #nv_wr32			// ENTRY
-		add b32 $r14 $r4 0x100
-		mov $r15 2			// CTRL_START_TRIGGER
-		call #nv_wr32			// CTRL
-
-		// wait for it to complete, and adjust context size
-		add b32 $r14 $r4 0x800
-		init_gpc_wait:
-			call #nv_rd32
-			xbit $r15 $r15 31
-			bra e #init_gpc_wait
-		add b32 $r14 $r4 0x804
-		call #nv_rd32
-		add b32 $r1 $r15
-
-		// next!
-		add b32 $r4 0x8000
-		sub b32 $r3 1
-		bra ne #init_gpc
-
-	// save context size, and tell host we're ready
-	mov $r2 0x800
-	shl b32 $r2 6
-	iowr I[$r2 + 0x100] $r1		// CC_SCRATCH[1]  = context size
-	add b32 $r2 0x800
-	clear b32 $r1
-	bset $r1 31
-	iowr I[$r2 + 0x000] $r1		// CC_SCRATCH[0] |= 0x80000000
-
-// Main program loop, very simple, sleeps until woken up by the interrupt
-// handler, pulls a command from the queue and executes its handler
-//
-main:
-	// sleep until we have something to do
-	bset $flags $p0
-	sleep $p0
-	mov $r13 #cmd_queue
-	call #queue_get
-	bra $p1 #main
-
-	// context switch, requested by GPU?
-	cmpu b32 $r14 0x4001
-	bra ne #main_not_ctx_switch
-		trace_set(T_AUTO)
-		mov $r1 0xb00
-		shl b32 $r1 6
-		iord $r2 I[$r1 + 0x100]		// CHAN_NEXT
-		iord $r1 I[$r1 + 0x000]		// CHAN_CUR
-
-		xbit $r3 $r1 31
-		bra e #chsw_no_prev
-			xbit $r3 $r2 31
-			bra e #chsw_prev_no_next
-				push $r2
-				mov b32 $r2 $r1
-				trace_set(T_SAVE)
-				bclr $flags $p1
-				bset $flags $p2
-				call #ctx_xfer
-				trace_clr(T_SAVE);
-				pop $r2
-				trace_set(T_LOAD);
-				bset $flags $p1
-				call #ctx_xfer
-				trace_clr(T_LOAD);
-				bra #chsw_done
-			chsw_prev_no_next:
-				push $r2
-				mov b32 $r2 $r1
-				bclr $flags $p1
-				bclr $flags $p2
-				call #ctx_xfer
-				pop $r2
-				mov $r1 0xb00
-				shl b32 $r1 6
-				iowr I[$r1] $r2
-				bra #chsw_done
-		chsw_no_prev:
-			xbit $r3 $r2 31
-			bra e #chsw_done
-				bset $flags $p1
-				bclr $flags $p2
-				call #ctx_xfer
-
-		// ack the context switch request
-		chsw_done:
-		mov $r1 0xb0c
-		shl b32 $r1 6
-		mov $r2 1
-		iowr I[$r1 + 0x000] $r2		// 0x409b0c
-		trace_clr(T_AUTO)
-		bra #main
-
-	// request to set current channel? (*not* a context switch)
-	main_not_ctx_switch:
-	cmpu b32 $r14 0x0001
-	bra ne #main_not_ctx_chan
-		mov b32 $r2 $r15
-		call #ctx_chan
-		bra #main_done
-
-	// request to store current channel context?
-	main_not_ctx_chan:
-	cmpu b32 $r14 0x0002
-	bra ne #main_not_ctx_save
-		trace_set(T_SAVE)
-		bclr $flags $p1
-		bclr $flags $p2
-		call #ctx_xfer
-		trace_clr(T_SAVE)
-		bra #main_done
-
-	main_not_ctx_save:
-		shl b32 $r15 $r14 16
-		or $r15 E_BAD_COMMAND
-		call #error
-		bra #main
-
-	main_done:
-	mov $r1 0x820
-	shl b32 $r1 6
-	clear b32 $r2
-	bset $r2 31
-	iowr I[$r1 + 0x000] $r2		// CC_SCRATCH[0] |= 0x80000000
-	bra #main
-
-// interrupt handler
-ih:
-	push $r8
-	mov $r8 $flags
-	push $r8
-	push $r9
-	push $r10
-	push $r11
-	push $r13
-	push $r14
-	push $r15
-
-	// incoming fifo command?
-	iord $r10 I[$r0 + 0x200]	// INTR
-	and $r11 $r10 0x00000004
-	bra e #ih_no_fifo
-		// queue incoming fifo command for later processing
-		mov $r11 0x1900
-		mov $r13 #cmd_queue
-		iord $r14 I[$r11 + 0x100]	// FIFO_CMD
-		iord $r15 I[$r11 + 0x000]	// FIFO_DATA
-		call #queue_put
-		add b32 $r11 0x400
-		mov $r14 1
-		iowr I[$r11 + 0x000] $r14	// FIFO_ACK
-
-	// context switch request?
-	ih_no_fifo:
-	and $r11 $r10 0x00000100
-	bra e #ih_no_ctxsw
-		// enqueue a context switch for later processing
-		mov $r13 #cmd_queue
-		mov $r14 0x4001
-		call #queue_put
-
-	// anything we didn't handle, bring it to the host's attention
-	ih_no_ctxsw:
-	mov $r11 0x104
-	not b32 $r11
-	and $r11 $r10 $r11
-	bra e #ih_no_other
-		mov $r10 0xc1c
-		shl b32 $r10 6
-		iowr I[$r10] $r11	// INTR_UP_SET
-
-	// ack, and wake up main()
-	ih_no_other:
-	iowr I[$r0 + 0x100] $r10	// INTR_ACK
-
-	pop $r15
-	pop $r14
-	pop $r13
-	pop $r11
-	pop $r10
-	pop $r9
-	pop $r8
-	mov $flags $r8
-	pop $r8
-	bclr $flags $p0
-	iret
-
-// Not real sure, but, MEM_CMD 7 will hang forever if this isn't done
-ctx_4160s:
-	mov $r14 0x4160
-	sethi $r14 0x400000
-	mov $r15 1
-	call #nv_wr32
-	ctx_4160s_wait:
-		call #nv_rd32
-		xbit $r15 $r15 4
-		bra e #ctx_4160s_wait
-	ret
-
-// Without clearing again at end of xfer, some things cause PGRAPH
-// to hang with STATUS=0x00000007 until it's cleared.. fbcon can
-// still function with it set however...
-ctx_4160c:
-	mov $r14 0x4160
-	sethi $r14 0x400000
-	clear b32 $r15
-	call #nv_wr32
-	ret
-
-// Again, not real sure
-//
-// In: $r15 value to set 0x404170 to
-//
-ctx_4170s:
-	mov $r14 0x4170
-	sethi $r14 0x400000
-	or $r15 0x10
-	call #nv_wr32
-	ret
-
-// Waits for a ctx_4170s() call to complete
-//
-ctx_4170w:
-	mov $r14 0x4170
-	sethi $r14 0x400000
-	call #nv_rd32
-	and $r15 0x10
-	bra ne #ctx_4170w
-	ret
-
-// Disables various things, waits a bit, and re-enables them..
-//
-// Not sure how exactly this helps, perhaps "ENABLE" is not such a
-// good description for the bits we turn off?  Anyways, without this,
-// funny things happen.
-//
-ctx_redswitch:
-	mov $r14 0x614
-	shl b32 $r14 6
-	mov $r15 0x270
-	iowr I[$r14] $r15	// HUB_RED_SWITCH = ENABLE_GPC, POWER_ALL
-	mov $r15 8
-	ctx_redswitch_delay:
-		sub b32 $r15 1
-		bra ne #ctx_redswitch_delay
-	mov $r15 0x770
-	iowr I[$r14] $r15	// HUB_RED_SWITCH = ENABLE_ALL, POWER_ALL
-	ret
-
-// Not a clue what this is for, except that unless the value is 0x10, the
-// strand context is saved (and presumably restored) incorrectly..
-//
-// In: $r15 value to set to (0x00/0x10 are used)
-//
-ctx_86c:
-	mov $r14 0x86c
-	shl b32 $r14 6
-	iowr I[$r14] $r15	// HUB(0x86c) = val
-	mov $r14 -0x75ec
-	sethi $r14 0x400000
-	call #nv_wr32		// ROP(0xa14) = val
-	mov $r14 -0x5794
-	sethi $r14 0x410000
-	call #nv_wr32		// GPC(0x86c) = val
-	ret
-
-// ctx_load - load's a channel's ctxctl data, and selects its vm
-//
-// In: $r2 channel address
-//
-ctx_load:
-	trace_set(T_CHAN)
-
-	// switch to channel, somewhat magic in parts..
-	mov $r10 12		// DONE_UNK12
-	call #wait_donez
-	mov $r1 0xa24
-	shl b32 $r1 6
-	iowr I[$r1 + 0x000] $r0	// 0x409a24
-	mov $r3 0xb00
-	shl b32 $r3 6
-	iowr I[$r3 + 0x100] $r2	// CHAN_NEXT
-	mov $r1 0xa0c
-	shl b32 $r1 6
-	mov $r4 7
-	iowr I[$r1 + 0x000] $r2 // MEM_CHAN
-	iowr I[$r1 + 0x100] $r4	// MEM_CMD
-	ctx_chan_wait_0:
-		iord $r4 I[$r1 + 0x100]
-		and $r4 0x1f
-		bra ne #ctx_chan_wait_0
-	iowr I[$r3 + 0x000] $r2	// CHAN_CUR
-
-	// load channel header, fetch PGRAPH context pointer
-	mov $xtargets $r0
-	bclr $r2 31
-	shl b32 $r2 4
-	add b32 $r2 2
-
-	trace_set(T_LCHAN)
-	mov $r1 0xa04
-	shl b32 $r1 6
-	iowr I[$r1 + 0x000] $r2		// MEM_BASE
-	mov $r1 0xa20
-	shl b32 $r1 6
-	mov $r2 0x0002
-	sethi $r2 0x80000000
-	iowr I[$r1 + 0x000] $r2		// MEM_TARGET = vram
-	mov $r1 0x10			// chan + 0x0210
-	mov $r2 #xfer_data
-	sethi $r2 0x00020000		// 16 bytes
-	xdld $r1 $r2
-	xdwait
-	trace_clr(T_LCHAN)
-
-	// update current context
-	ld b32 $r1 D[$r0 + #xfer_data + 4]
-	shl b32 $r1 24
-	ld b32 $r2 D[$r0 + #xfer_data + 0]
-	shr b32 $r2 8
-	or $r1 $r2
-	st b32 D[$r0 + #ctx_current] $r1
-
-	// set transfer base to start of context, and fetch context header
-	trace_set(T_LCTXH)
-	mov $r2 0xa04
-	shl b32 $r2 6
-	iowr I[$r2 + 0x000] $r1		// MEM_BASE
-	mov $r2 1
-	mov $r1 0xa20
-	shl b32 $r1 6
-	iowr I[$r1 + 0x000] $r2		// MEM_TARGET = vm
-	mov $r1 #chan_data
-	sethi $r1 0x00060000		// 256 bytes
-	xdld $r0 $r1
-	xdwait
-	trace_clr(T_LCTXH)
-
-	trace_clr(T_CHAN)
-	ret
-
-// ctx_chan - handler for HUB_SET_CHAN command, will set a channel as
-//            the active channel for ctxctl, but not actually transfer
-//            any context data.  intended for use only during initial
-//            context construction.
-//
-// In: $r2 channel address
-//
-ctx_chan:
-	call #ctx_4160s
-	call #ctx_load
-	mov $r10 12			// DONE_UNK12
-	call #wait_donez
-	mov $r1 0xa10
-	shl b32 $r1 6
-	mov $r2 5
-	iowr I[$r1 + 0x000] $r2		// MEM_CMD = 5 (???)
-	ctx_chan_wait:
-		iord $r2 I[$r1 + 0x000]
-		or $r2 $r2
-		bra ne #ctx_chan_wait
-	call #ctx_4160c
-	ret
-
-// Execute per-context state overrides list
-//
-// Only executed on the first load of a channel.  Might want to look into
-// removing this and having the host directly modify the channel's context
-// to change this state...  The nouveau DRM already builds this list as
-// it's definitely needed for NVIDIA's, so we may as well use it for now
-//
-// Input: $r1 mmio list length
-//
-ctx_mmio_exec:
-	// set transfer base to be the mmio list
-	ld b32 $r3 D[$r0 + #chan_mmio_address]
-	mov $r2 0xa04
-	shl b32 $r2 6
-	iowr I[$r2 + 0x000] $r3		// MEM_BASE
-
-	clear b32 $r3
-	ctx_mmio_loop:
-		// fetch next 256 bytes of mmio list if necessary
-		and $r4 $r3 0xff
-		bra ne #ctx_mmio_pull
-			mov $r5 #xfer_data
-			sethi $r5 0x00060000	// 256 bytes
-			xdld $r3 $r5
-			xdwait
-
-		// execute a single list entry
-		ctx_mmio_pull:
-		ld b32 $r14 D[$r4 + #xfer_data + 0x00]
-		ld b32 $r15 D[$r4 + #xfer_data + 0x04]
-		call #nv_wr32
-
-		// next!
-		add b32 $r3 8
-		sub b32 $r1 1
-		bra ne #ctx_mmio_loop
-
-	// set transfer base back to the current context
-	ctx_mmio_done:
-	ld b32 $r3 D[$r0 + #ctx_current]
-	iowr I[$r2 + 0x000] $r3		// MEM_BASE
-
-	// disable the mmio list now, we don't need/want to execute it again
-	st b32 D[$r0 + #chan_mmio_count] $r0
-	mov $r1 #chan_data
-	sethi $r1 0x00060000		// 256 bytes
-	xdst $r0 $r1
-	xdwait
-	ret
-
-// Transfer HUB context data between GPU and storage area
-//
-// In: $r2 channel address
-//     $p1 clear on save, set on load
-//     $p2 set if opposite direction done/will be done, so:
-//		on save it means: "a load will follow this save"
-//		on load it means: "a save preceeded this load"
-//
-ctx_xfer:
-	bra not $p1 #ctx_xfer_pre
-	bra $p2 #ctx_xfer_pre_load
-	ctx_xfer_pre:
-		mov $r15 0x10
-		call #ctx_86c
-		call #ctx_4160s
-		bra not $p1 #ctx_xfer_exec
-
-	ctx_xfer_pre_load:
-		mov $r15 2
-		call #ctx_4170s
-		call #ctx_4170w
-		call #ctx_redswitch
-		clear b32 $r15
-		call #ctx_4170s
-		call #ctx_load
-
-	// fetch context pointer, and initiate xfer on all GPCs
-	ctx_xfer_exec:
-	ld b32 $r1 D[$r0 + #ctx_current]
-	mov $r2 0x414
-	shl b32 $r2 6
-	iowr I[$r2 + 0x000] $r0	// BAR_STATUS = reset
-	mov $r14 -0x5b00
-	sethi $r14 0x410000
-	mov b32 $r15 $r1
-	call #nv_wr32		// GPC_BCAST_WRCMD_DATA = ctx pointer
-	add b32 $r14 4
-	xbit $r15 $flags $p1
-	xbit $r2 $flags $p2
-	shl b32 $r2 1
-	or $r15 $r2
-	call #nv_wr32		// GPC_BCAST_WRCMD_CMD = GPC_XFER(type)
-
-	// strands
-	mov $r1 0x4afc
-	sethi $r1 0x20000
-	mov $r2 0xc
-	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x0c
-	call #strand_wait
-	mov $r2 0x47fc
-	sethi $r2 0x20000
-	iowr I[$r2] $r0		// STRAND_FIRST_GENE(0x3f) = 0x00
-	xbit $r2 $flags $p1
-	add b32 $r2 3
-	iowr I[$r1] $r2		// STRAND_CMD(0x3f) = 0x03/0x04 (SAVE/LOAD)
-
-	// mmio context
-	xbit $r10 $flags $p1	// direction
-	or $r10 6		// first, last
-	mov $r11 0		// base = 0
-	ld b32 $r12 D[$r0 + #hub_mmio_list_head]
-	ld b32 $r13 D[$r0 + #hub_mmio_list_tail]
-	mov $r14 0		// not multi
-	call #mmctx_xfer
-
-	// wait for GPCs to all complete
-	mov $r10 8		// DONE_BAR
-	call #wait_doneo
-
-	// wait for strand xfer to complete
-	call #strand_wait
-
-	// post-op
-	bra $p1 #ctx_xfer_post
-		mov $r10 12		// DONE_UNK12
-		call #wait_donez
-		mov $r1 0xa10
-		shl b32 $r1 6
-		mov $r2 5
-		iowr I[$r1] $r2		// MEM_CMD
-		ctx_xfer_post_save_wait:
-			iord $r2 I[$r1]
-			or $r2 $r2
-			bra ne #ctx_xfer_post_save_wait
-
-	bra $p2 #ctx_xfer_done
-	ctx_xfer_post:
-		mov $r15 2
-		call #ctx_4170s
-		clear b32 $r15
-		call #ctx_86c
-		call #strand_post
-		call #ctx_4170w
-		clear b32 $r15
-		call #ctx_4170s
-
-		bra not $p1 #ctx_xfer_no_post_mmio
-		ld b32 $r1 D[$r0 + #chan_mmio_count]
-		or $r1 $r1
-		bra e #ctx_xfer_no_post_mmio
-			call #ctx_mmio_exec
-
-		ctx_xfer_no_post_mmio:
-		call #ctx_4160c
-
-	ctx_xfer_done:
-	ret
-
-.align 256
diff --git a/drivers/gpu/drm/nouveau/nvc0_grhub.fuc.h b/drivers/gpu/drm/nouveau/nvc0_grhub.fuc.h
deleted file mode 100644
index c5ed307..0000000
--- a/drivers/gpu/drm/nouveau/nvc0_grhub.fuc.h
+++ /dev/null
@@ -1,838 +0,0 @@
-uint32_t nvc0_grhub_data[] = {
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x000000c0,
-	0x013c00a0,
-	0x000000c1,
-	0x014000a0,
-	0x000000c3,
-	0x013c00a0,
-	0x000000c4,
-	0x013c00a0,
-	0x000000c8,
-	0x013c00a0,
-	0x000000ce,
-	0x013c00a0,
-	0x000000cf,
-	0x013c00a0,
-	0x000000d9,
-	0x01dc0140,
-	0x00000000,
-	0x0417e91c,
-	0x04400204,
-	0x28404004,
-	0x00404044,
-	0x34404094,
-	0x184040d0,
-	0x004040f8,
-	0x08404130,
-	0x08404150,
-	0x04404164,
-	0x08404174,
-	0x1c404200,
-	0x34404404,
-	0x0c404460,
-	0x00404480,
-	0x00404498,
-	0x0c404604,
-	0x7c404618,
-	0x50404698,
-	0x044046f0,
-	0x54404700,
-	0x00405800,
-	0x08405830,
-	0x00405854,
-	0x0c405870,
-	0x04405a00,
-	0x00405a18,
-	0x00406020,
-	0x0c406028,
-	0x044064a8,
-	0x044064b4,
-	0x00407804,
-	0x1440780c,
-	0x004078bc,
-	0x18408000,
-	0x00408064,
-	0x08408800,
-	0x0c408900,
-	0x00408980,
-	0x044064c0,
-	0x0417e91c,
-	0x04400204,
-	0x24404004,
-	0x00404044,
-	0x34404094,
-	0x184040d0,
-	0x004040f8,
-	0x08404130,
-	0x08404150,
-	0x04404164,
-	0x04404178,
-	0x1c404200,
-	0x34404404,
-	0x0c404460,
-	0x00404480,
-	0x00404498,
-	0x0c404604,
-	0x7c404618,
-	0x50404698,
-	0x044046f0,
-	0x54404700,
-	0x00405800,
-	0x08405830,
-	0x00405854,
-	0x0c405870,
-	0x04405a00,
-	0x00405a18,
-	0x00406020,
-	0x0c406028,
-	0x044064a8,
-	0x104064b4,
-	0x00407804,
-	0x1440780c,
-	0x004078bc,
-	0x18408000,
-	0x00408064,
-	0x08408800,
-	0x0c408900,
-	0x00408980,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-};
-
-uint32_t nvc0_grhub_code[] = {
-	0x03090ef5,
-	0x9800d898,
-	0x86f001d9,
-	0x0489b808,
-	0xf00c1bf4,
-	0x21f502f7,
-	0x00f802ec,
-	0xb60798c4,
-	0x8dbb0384,
-	0x0880b600,
-	0x80008e80,
-	0x90b6018f,
-	0x0f94f001,
-	0xf801d980,
-	0x0131f400,
-	0x9800d898,
-	0x89b801d9,
-	0x210bf404,
-	0xb60789c4,
-	0x9dbb0394,
-	0x0890b600,
-	0x98009e98,
-	0x80b6019f,
-	0x0f84f001,
-	0xf400d880,
-	0x00f80132,
-	0x0728b7f1,
-	0xb906b4b6,
-	0xc9f002ec,
-	0x00bcd01f,
-	0xc800bccf,
-	0x1bf41fcc,
-	0x06a7f0fa,
-	0x010321f5,
-	0xf840bfcf,
-	0x28b7f100,
-	0x06b4b607,
-	0xb980bfd0,
-	0xc9f002ec,
-	0x1ec9f01f,
-	0xcf00bcd0,
-	0xccc800bc,
-	0xfa1bf41f,
-	0x87f100f8,
-	0x84b60430,
-	0x1ff9f006,
-	0xf8008fd0,
-	0x3087f100,
-	0x0684b604,
-	0xf80080d0,
-	0x3c87f100,
-	0x0684b608,
-	0x99f094bd,
-	0x0089d000,
-	0x081887f1,
-	0xd00684b6,
-	0x87f1008a,
-	0x84b60400,
-	0x0088cf06,
-	0xf4888aff,
-	0x87f1f31b,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00099,
-	0xf100f800,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00099f0,
-	0x87f10089,
-	0x84b60818,
-	0x008ad006,
-	0x040087f1,
-	0xcf0684b6,
-	0x8aff0088,
-	0xf30bf488,
-	0x085c87f1,
-	0xbd0684b6,
-	0x0099f094,
-	0xf80089d0,
-	0x9894bd00,
-	0x85b600e8,
-	0x0180b61a,
-	0xbb0284b6,
-	0xe0b60098,
-	0x04efb804,
-	0xb9eb1bf4,
-	0x00f8029f,
-	0x083c87f1,
-	0xbd0684b6,
-	0x0199f094,
-	0xf10089d0,
-	0xb6071087,
-	0x94bd0684,
-	0xf405bbfd,
-	0x8bd0090b,
-	0x0099f000,
-	0xf405eefd,
-	0x8ed00c0b,
-	0xc08fd080,
-	0xb70199f0,
-	0xc8010080,
-	0xb4b600ab,
-	0x0cb9f010,
-	0xb601aec8,
-	0xbefd11e4,
-	0x008bd005,
-	0xf0008ecf,
-	0x0bf41fe4,
-	0x00ce98fa,
-	0xd005e9fd,
-	0xc0b6c08e,
-	0x04cdb804,
-	0xc8e81bf4,
-	0x1bf402ab,
-	0x008bcf18,
-	0xb01fb4f0,
-	0x1bf410b4,
-	0x02a7f0f7,
-	0xf4c921f4,
-	0xabc81b0e,
-	0x10b4b600,
-	0xf00cb9f0,
-	0x8bd012b9,
-	0x008bcf00,
-	0xf412bbc8,
-	0x87f1fa1b,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00199,
-	0xf900f800,
-	0x02a7f0a0,
-	0xfcc921f4,
-	0xf100f8a0,
-	0xf04afc87,
-	0x97f00283,
-	0x0089d00c,
-	0x020721f5,
-	0x87f100f8,
-	0x83f04afc,
-	0x0d97f002,
-	0xf50089d0,
-	0xf8020721,
-	0xfca7f100,
-	0x02a3f04f,
-	0x0500aba2,
-	0xd00fc7f0,
-	0xc7f000ac,
-	0x00bcd00b,
-	0x020721f5,
-	0xf000aed0,
-	0xbcd00ac7,
-	0x0721f500,
-	0xf100f802,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00399f0,
-	0x21f50089,
-	0xe7f00213,
-	0x3921f503,
-	0xfca7f102,
-	0x02a3f046,
-	0x0400aba0,
-	0xf040a0d0,
-	0xbcd001c7,
-	0x0721f500,
-	0x010c9202,
-	0xf000acd0,
-	0xbcd002c7,
-	0x0721f500,
-	0x2621f502,
-	0x8087f102,
-	0x0684b608,
-	0xb70089cf,
-	0x95220080,
-	0x8ed008fe,
-	0x408ed000,
-	0xb6808acf,
-	0xa0b606a5,
-	0x00eabb01,
-	0xb60480b6,
-	0x1bf40192,
-	0x08e4b6e8,
-	0xf1f2efbc,
-	0xb6085c87,
-	0x94bd0684,
-	0xd00399f0,
-	0x00f80089,
-	0xe7f1e0f9,
-	0xe4b60814,
-	0x00efd006,
-	0x0c1ce7f1,
-	0xf006e4b6,
-	0xefd001f7,
-	0xf8e0fc00,
-	0xfe04bd00,
-	0x07fe0004,
-	0x0017f100,
-	0x0227f012,
-	0xf10012d0,
-	0xfe05b917,
-	0x17f10010,
-	0x10d00400,
-	0x0437f1c0,
-	0x0634b604,
-	0x200327f1,
-	0xf10032d0,
-	0xd0200427,
-	0x27f10132,
-	0x32d0200b,
-	0x0c27f102,
-	0x0732d020,
-	0x0c2427f1,
-	0xb90624b6,
-	0x23d00003,
-	0x0427f100,
-	0x0023f087,
-	0xb70012d0,
-	0xf0010012,
-	0x12d00427,
-	0x1031f400,
-	0x9604e7f1,
-	0xf440e3f0,
-	0xf1c76821,
-	0x01018090,
-	0x801ff4f0,
-	0x17f0000f,
-	0x041fbb01,
-	0xf10112b6,
-	0xb6040c27,
-	0x21d00624,
-	0x4021d000,
-	0x080027f1,
-	0xcf0624b6,
-	0xf7f00022,
-	0x08f0b654,
-	0xb800f398,
-	0x0bf40432,
-	0x0034b00b,
-	0xf8f11bf4,
-	0x0017f100,
-	0x02fe5801,
-	0xf003ff58,
-	0x0e8000e3,
-	0x150f8014,
-	0x013d21f5,
-	0x070037f1,
-	0x950634b6,
-	0x34d00814,
-	0x4034d000,
-	0x130030b7,
-	0xb6001fbb,
-	0x3fd002f5,
-	0x0815b600,
-	0xb60110b6,
-	0x1fb90814,
-	0x6321f502,
-	0x001fbb02,
-	0xf1000398,
-	0xf0200047,
-	0x4ea05043,
-	0x1fb90804,
-	0x8d21f402,
-	0x08004ea0,
-	0xf4022fb9,
-	0x4ea08d21,
-	0xf4bd010c,
-	0xa08d21f4,
-	0xf401044e,
-	0x4ea08d21,
-	0xf7f00100,
-	0x8d21f402,
-	0x08004ea0,
-	0xc86821f4,
-	0x0bf41fff,
-	0x044ea0fa,
-	0x6821f408,
-	0xb7001fbb,
-	0xb6800040,
-	0x1bf40132,
-	0x0027f1b4,
-	0x0624b608,
-	0xb74021d0,
-	0xbd080020,
-	0x1f19f014,
-	0xf40021d0,
-	0x28f40031,
-	0x08d7f000,
-	0xf43921f4,
-	0xe4b1f401,
-	0x1bf54001,
-	0x87f100d1,
-	0x84b6083c,
-	0xf094bd06,
-	0x89d00499,
-	0x0017f100,
-	0x0614b60b,
-	0xcf4012cf,
-	0x13c80011,
-	0x7e0bf41f,
-	0xf41f23c8,
-	0x20f95a0b,
-	0xf10212b9,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00799f0,
-	0x32f40089,
-	0x0231f401,
-	0x082921f5,
-	0x085c87f1,
-	0xbd0684b6,
-	0x0799f094,
-	0xfc0089d0,
-	0x3c87f120,
-	0x0684b608,
-	0x99f094bd,
-	0x0089d006,
-	0xf50131f4,
-	0xf1082921,
-	0xb6085c87,
-	0x94bd0684,
-	0xd00699f0,
-	0x0ef40089,
-	0xb920f931,
-	0x32f40212,
-	0x0232f401,
-	0x082921f5,
-	0x17f120fc,
-	0x14b60b00,
-	0x0012d006,
-	0xc8130ef4,
-	0x0bf41f23,
-	0x0131f40d,
-	0xf50232f4,
-	0xf1082921,
-	0xb60b0c17,
-	0x27f00614,
-	0x0012d001,
-	0x085c87f1,
-	0xbd0684b6,
-	0x0499f094,
-	0xf50089d0,
-	0xb0ff200e,
-	0x1bf401e4,
-	0x02f2b90d,
-	0x07b521f5,
-	0xb0420ef4,
-	0x1bf402e4,
-	0x3c87f12e,
-	0x0684b608,
-	0x99f094bd,
-	0x0089d007,
-	0xf40132f4,
-	0x21f50232,
-	0x87f10829,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00799,
-	0x110ef400,
-	0xf010ef94,
-	0x21f501f5,
-	0x0ef502ec,
-	0x17f1fed1,
-	0x14b60820,
-	0xf024bd06,
-	0x12d01f29,
-	0xbe0ef500,
-	0xfe80f9fe,
-	0x80f90188,
-	0xa0f990f9,
-	0xd0f9b0f9,
-	0xf0f9e0f9,
-	0xc4800acf,
-	0x0bf404ab,
-	0x00b7f11d,
-	0x08d7f019,
-	0xcf40becf,
-	0x21f400bf,
-	0x00b0b704,
-	0x01e7f004,
-	0xe400bed0,
-	0xf40100ab,
-	0xd7f00d0b,
-	0x01e7f108,
-	0x0421f440,
-	0x0104b7f1,
-	0xabffb0bd,
-	0x0d0bf4b4,
-	0x0c1ca7f1,
-	0xd006a4b6,
-	0x0ad000ab,
-	0xfcf0fc40,
-	0xfcd0fce0,
-	0xfca0fcb0,
-	0xfe80fc90,
-	0x80fc0088,
-	0xf80032f4,
-	0x60e7f101,
-	0x40e3f041,
-	0xf401f7f0,
-	0x21f48d21,
-	0x04ffc868,
-	0xf8fa0bf4,
-	0x60e7f100,
-	0x40e3f041,
-	0x21f4f4bd,
-	0xf100f88d,
-	0xf04170e7,
-	0xf5f040e3,
-	0x8d21f410,
-	0xe7f100f8,
-	0xe3f04170,
-	0x6821f440,
-	0xf410f4f0,
-	0x00f8f31b,
-	0x0614e7f1,
-	0xf106e4b6,
-	0xd00270f7,
-	0xf7f000ef,
-	0x01f2b608,
-	0xf1fd1bf4,
-	0xd00770f7,
-	0x00f800ef,
-	0x086ce7f1,
-	0xd006e4b6,
-	0xe7f100ef,
-	0xe3f08a14,
-	0x8d21f440,
-	0xa86ce7f1,
-	0xf441e3f0,
-	0x00f88d21,
-	0x083c87f1,
-	0xbd0684b6,
-	0x0599f094,
-	0xf00089d0,
-	0x21f40ca7,
-	0x2417f1c9,
-	0x0614b60a,
-	0xf10010d0,
-	0xb60b0037,
-	0x32d00634,
-	0x0c17f140,
-	0x0614b60a,
-	0xd00747f0,
-	0x14d00012,
-	0x4014cf40,
-	0xf41f44f0,
-	0x32d0fa1b,
-	0x000bfe00,
-	0xb61f2af0,
-	0x20b60424,
-	0x3c87f102,
-	0x0684b608,
-	0x99f094bd,
-	0x0089d008,
-	0x0a0417f1,
-	0xd00614b6,
-	0x17f10012,
-	0x14b60a20,
-	0x0227f006,
-	0x800023f1,
-	0xf00012d0,
-	0x27f11017,
-	0x23f00300,
-	0x0512fa02,
-	0x87f103f8,
-	0x84b6085c,
-	0xf094bd06,
-	0x89d00899,
-	0xc1019800,
-	0x981814b6,
-	0x25b6c002,
-	0x0512fd08,
-	0xf1160180,
-	0xb6083c87,
-	0x94bd0684,
-	0xd00999f0,
-	0x27f10089,
-	0x24b60a04,
-	0x0021d006,
-	0xf10127f0,
-	0xb60a2017,
-	0x12d00614,
-	0x0017f100,
-	0x0613f002,
-	0xf80501fa,
-	0x5c87f103,
-	0x0684b608,
-	0x99f094bd,
-	0x0089d009,
-	0x085c87f1,
-	0xbd0684b6,
-	0x0599f094,
-	0xf80089d0,
-	0x3121f500,
-	0xb821f506,
-	0x0ca7f006,
-	0xf1c921f4,
-	0xb60a1017,
-	0x27f00614,
-	0x0012d005,
-	0xfd0012cf,
-	0x1bf40522,
-	0x4921f5fa,
-	0x9800f806,
-	0x27f18103,
-	0x24b60a04,
-	0x0023d006,
-	0x34c434bd,
-	0x0f1bf4ff,
-	0x030057f1,
-	0xfa0653f0,
-	0x03f80535,
-	0x98c04e98,
-	0x21f4c14f,
-	0x0830b68d,
-	0xf40112b6,
-	0x0398df1b,
-	0x0023d016,
-	0xf1800080,
-	0xf0020017,
-	0x01fa0613,
-	0xf803f806,
-	0x0611f400,
-	0xf01102f4,
-	0x21f510f7,
-	0x21f50698,
-	0x11f40631,
-	0x02f7f01c,
-	0x065721f5,
-	0x066621f5,
-	0x067821f5,
-	0x21f5f4bd,
-	0x21f50657,
-	0x019806b8,
-	0x1427f116,
-	0x0624b604,
-	0xf10020d0,
-	0xf0a500e7,
-	0x1fb941e3,
-	0x8d21f402,
-	0xf004e0b6,
-	0x2cf001fc,
-	0x0124b602,
-	0xf405f2fd,
-	0x17f18d21,
-	0x13f04afc,
-	0x0c27f002,
-	0xf50012d0,
-	0xf1020721,
-	0xf047fc27,
-	0x20d00223,
-	0x012cf000,
-	0xd00320b6,
-	0xacf00012,
-	0x06a5f001,
-	0x9800b7f0,
-	0x0d98140c,
-	0x00e7f015,
-	0x015c21f5,
-	0xf508a7f0,
-	0xf5010321,
-	0xf4020721,
-	0xa7f02201,
-	0xc921f40c,
-	0x0a1017f1,
-	0xf00614b6,
-	0x12d00527,
-	0x0012cf00,
-	0xf40522fd,
-	0x02f4fa1b,
-	0x02f7f032,
-	0x065721f5,
-	0x21f5f4bd,
-	0x21f50698,
-	0x21f50226,
-	0xf4bd0666,
-	0x065721f5,
-	0x981011f4,
-	0x11fd8001,
-	0x070bf405,
-	0x07df21f5,
-	0x064921f5,
-	0x000000f8,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-	0x00000000,
-};
diff --git a/drivers/gpu/drm/nouveau/nvc0_instmem.c b/drivers/gpu/drm/nouveau/nvc0_instmem.c
index b701c43..82357d2 100644
--- a/drivers/gpu/drm/nouveau/nvc0_instmem.c
+++ b/drivers/gpu/drm/nouveau/nvc0_instmem.c
@@ -32,6 +32,7 @@ struct nvc0_instmem_priv {
 	struct nouveau_channel *bar1;
 	struct nouveau_gpuobj  *bar3_pgd;
 	struct nouveau_channel *bar3;
+	struct nouveau_gpuobj  *chan_pgd;
 };
 
 int
@@ -180,11 +181,17 @@ nvc0_instmem_init(struct drm_device *dev)
 		goto error;
 
 	/* channel vm */
-	ret = nouveau_vm_new(dev, 0, (1ULL << 40), 0x0008000000ULL,
-			     &dev_priv->chan_vm);
+	ret = nouveau_vm_new(dev, 0, (1ULL << 40), 0x0008000000ULL, &vm);
 	if (ret)
 		goto error;
 
+	ret = nouveau_gpuobj_new(dev, NULL, 0x8000, 4096, 0, &priv->chan_pgd);
+	if (ret)
+		goto error;
+
+	nouveau_vm_ref(vm, &dev_priv->chan_vm, priv->chan_pgd);
+	nouveau_vm_ref(NULL, &vm, NULL);
+
 	nvc0_instmem_resume(dev);
 	return 0;
 error:
@@ -204,7 +211,8 @@ nvc0_instmem_takedown(struct drm_device *dev)
 	nv_wr32(dev, 0x1704, 0x00000000);
 	nv_wr32(dev, 0x1714, 0x00000000);
 
-	nouveau_vm_ref(NULL, &dev_priv->chan_vm, NULL);
+	nouveau_vm_ref(NULL, &dev_priv->chan_vm, priv->chan_pgd);
+	nouveau_gpuobj_ref(NULL, &priv->chan_pgd);
 
 	nvc0_channel_del(&priv->bar1);
 	nouveau_vm_ref(NULL, &dev_priv->bar1_vm, priv->bar1_pgd);
diff --git a/drivers/gpu/drm/nouveau/nvc0_pm.c b/drivers/gpu/drm/nouveau/nvc0_pm.c
deleted file mode 100644
index ce65f81..0000000
--- a/drivers/gpu/drm/nouveau/nvc0_pm.c
+++ /dev/null
@@ -1,392 +0,0 @@
-/*
- * Copyright 2011 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Ben Skeggs
- */
-
-#include "drmP.h"
-#include "nouveau_drv.h"
-#include "nouveau_bios.h"
-#include "nouveau_pm.h"
-
-static u32 read_div(struct drm_device *, int, u32, u32);
-static u32 read_pll(struct drm_device *, u32);
-
-static u32
-read_vco(struct drm_device *dev, u32 dsrc)
-{
-	u32 ssrc = nv_rd32(dev, dsrc);
-	if (!(ssrc & 0x00000100))
-		return read_pll(dev, 0x00e800);
-	return read_pll(dev, 0x00e820);
-}
-
-static u32
-read_pll(struct drm_device *dev, u32 pll)
-{
-	u32 ctrl = nv_rd32(dev, pll + 0);
-	u32 coef = nv_rd32(dev, pll + 4);
-	u32 P = (coef & 0x003f0000) >> 16;
-	u32 N = (coef & 0x0000ff00) >> 8;
-	u32 M = (coef & 0x000000ff) >> 0;
-	u32 sclk, doff;
-
-	if (!(ctrl & 0x00000001))
-		return 0;
-
-	switch (pll & 0xfff000) {
-	case 0x00e000:
-		sclk = 27000;
-		P = 1;
-		break;
-	case 0x137000:
-		doff = (pll - 0x137000) / 0x20;
-		sclk = read_div(dev, doff, 0x137120, 0x137140);
-		break;
-	case 0x132000:
-		switch (pll) {
-		case 0x132000:
-			sclk = read_pll(dev, 0x132020);
-			break;
-		case 0x132020:
-			sclk = read_div(dev, 0, 0x137320, 0x137330);
-			break;
-		default:
-			return 0;
-		}
-		break;
-	default:
-		return 0;
-	}
-
-	return sclk * N / M / P;
-}
-
-static u32
-read_div(struct drm_device *dev, int doff, u32 dsrc, u32 dctl)
-{
-	u32 ssrc = nv_rd32(dev, dsrc + (doff * 4));
-	u32 sctl = nv_rd32(dev, dctl + (doff * 4));
-
-	switch (ssrc & 0x00000003) {
-	case 0:
-		if ((ssrc & 0x00030000) != 0x00030000)
-			return 27000;
-		return 108000;
-	case 2:
-		return 100000;
-	case 3:
-		if (sctl & 0x80000000) {
-			u32 sclk = read_vco(dev, dsrc + (doff * 4));
-			u32 sdiv = (sctl & 0x0000003f) + 2;
-			return (sclk * 2) / sdiv;
-		}
-
-		return read_vco(dev, dsrc + (doff * 4));
-	default:
-		return 0;
-	}
-}
-
-static u32
-read_mem(struct drm_device *dev)
-{
-	u32 ssel = nv_rd32(dev, 0x1373f0);
-	if (ssel & 0x00000001)
-		return read_div(dev, 0, 0x137300, 0x137310);
-	return read_pll(dev, 0x132000);
-}
-
-static u32
-read_clk(struct drm_device *dev, int clk)
-{
-	u32 sctl = nv_rd32(dev, 0x137250 + (clk * 4));
-	u32 ssel = nv_rd32(dev, 0x137100);
-	u32 sclk, sdiv;
-
-	if (ssel & (1 << clk)) {
-		if (clk < 7)
-			sclk = read_pll(dev, 0x137000 + (clk * 0x20));
-		else
-			sclk = read_pll(dev, 0x1370e0);
-		sdiv = ((sctl & 0x00003f00) >> 8) + 2;
-	} else {
-		sclk = read_div(dev, clk, 0x137160, 0x1371d0);
-		sdiv = ((sctl & 0x0000003f) >> 0) + 2;
-	}
-
-	if (sctl & 0x80000000)
-		return (sclk * 2) / sdiv;
-	return sclk;
-}
-
-int
-nvc0_pm_clocks_get(struct drm_device *dev, struct nouveau_pm_level *perflvl)
-{
-	perflvl->shader = read_clk(dev, 0x00);
-	perflvl->core   = perflvl->shader / 2;
-	perflvl->memory = read_mem(dev);
-	perflvl->rop    = read_clk(dev, 0x01);
-	perflvl->hub07  = read_clk(dev, 0x02);
-	perflvl->hub06  = read_clk(dev, 0x07);
-	perflvl->hub01  = read_clk(dev, 0x08);
-	perflvl->copy   = read_clk(dev, 0x09);
-	perflvl->daemon = read_clk(dev, 0x0c);
-	perflvl->vdec   = read_clk(dev, 0x0e);
-	return 0;
-}
-
-struct nvc0_pm_clock {
-	u32 freq;
-	u32 ssel;
-	u32 mdiv;
-	u32 dsrc;
-	u32 ddiv;
-	u32 coef;
-};
-
-struct nvc0_pm_state {
-	struct nvc0_pm_clock eng[16];
-};
-
-static u32
-calc_div(struct drm_device *dev, int clk, u32 ref, u32 freq, u32 *ddiv)
-{
-	u32 div = min((ref * 2) / freq, (u32)65);
-	if (div < 2)
-		div = 2;
-
-	*ddiv = div - 2;
-	return (ref * 2) / div;
-}
-
-static u32
-calc_src(struct drm_device *dev, int clk, u32 freq, u32 *dsrc, u32 *ddiv)
-{
-	u32 sclk;
-
-	/* use one of the fixed frequencies if possible */
-	*ddiv = 0x00000000;
-	switch (freq) {
-	case  27000:
-	case 108000:
-		*dsrc = 0x00000000;
-		if (freq == 108000)
-			*dsrc |= 0x00030000;
-		return freq;
-	case 100000:
-		*dsrc = 0x00000002;
-		return freq;
-	default:
-		*dsrc = 0x00000003;
-		break;
-	}
-
-	/* otherwise, calculate the closest divider */
-	sclk = read_vco(dev, clk);
-	if (clk < 7)
-		sclk = calc_div(dev, clk, sclk, freq, ddiv);
-	return sclk;
-}
-
-static u32
-calc_pll(struct drm_device *dev, int clk, u32 freq, u32 *coef)
-{
-	struct pll_lims limits;
-	int N, M, P, ret;
-
-	ret = get_pll_limits(dev, 0x137000 + (clk * 0x20), &limits);
-	if (ret)
-		return 0;
-
-	limits.refclk = read_div(dev, clk, 0x137120, 0x137140);
-	if (!limits.refclk)
-		return 0;
-
-	ret = nva3_calc_pll(dev, &limits, freq, &N, NULL, &M, &P);
-	if (ret <= 0)
-		return 0;
-
-	*coef = (P << 16) | (N << 8) | M;
-	return ret;
-}
-
-/* A (likely rather simplified and incomplete) view of the clock tree
- *
- * Key:
- *
- * S: source select
- * D: divider
- * P: pll
- * F: switch
- *
- * Engine clocks:
- *
- * 137250(D) ---- 137100(F0) ---- 137160(S)/1371d0(D) ------------------- ref
- *                      (F1) ---- 1370X0(P) ---- 137120(S)/137140(D) ---- ref
- *
- * Not all registers exist for all clocks.  For example: clocks >= 8 don't
- * have their own PLL (all tied to clock 7's PLL when in PLL mode), nor do
- * they have the divider at 1371d0, though the source selection at 137160
- * still exists.  You must use the divider at 137250 for these instead.
- *
- * Memory clock:
- *
- * TBD, read_mem() above is likely very wrong...
- *
- */
-
-static int
-calc_clk(struct drm_device *dev, int clk, struct nvc0_pm_clock *info, u32 freq)
-{
-	u32 src0, div0, div1D, div1P = 0;
-	u32 clk0, clk1 = 0;
-
-	/* invalid clock domain */
-	if (!freq)
-		return 0;
-
-	/* first possible path, using only dividers */
-	clk0 = calc_src(dev, clk, freq, &src0, &div0);
-	clk0 = calc_div(dev, clk, clk0, freq, &div1D);
-
-	/* see if we can get any closer using PLLs */
-	if (clk0 != freq && (0x00004387 & (1 << clk))) {
-		if (clk < 7)
-			clk1 = calc_pll(dev, clk, freq, &info->coef);
-		else
-			clk1 = read_pll(dev, 0x1370e0);
-		clk1 = calc_div(dev, clk, clk1, freq, &div1P);
-	}
-
-	/* select the method which gets closest to target freq */
-	if (abs((int)freq - clk0) <= abs((int)freq - clk1)) {
-		info->dsrc = src0;
-		if (div0) {
-			info->ddiv |= 0x80000000;
-			info->ddiv |= div0 << 8;
-			info->ddiv |= div0;
-		}
-		if (div1D) {
-			info->mdiv |= 0x80000000;
-			info->mdiv |= div1D;
-		}
-		info->ssel = 0;
-		info->freq = clk0;
-	} else {
-		if (div1P) {
-			info->mdiv |= 0x80000000;
-			info->mdiv |= div1P << 8;
-		}
-		info->ssel = (1 << clk);
-		info->freq = clk1;
-	}
-
-	return 0;
-}
-
-void *
-nvc0_pm_clocks_pre(struct drm_device *dev, struct nouveau_pm_level *perflvl)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nvc0_pm_state *info;
-	int ret;
-
-	info = kzalloc(sizeof(*info), GFP_KERNEL);
-	if (!info)
-		return ERR_PTR(-ENOMEM);
-
-	/* NFI why this is still in the performance table, the ROPCs appear
-	 * to get their clock from clock 2 ("hub07", actually hub05 on this
-	 * chip, but, anyway...) as well.  nvatiming confirms hub05 and ROP
-	 * are always the same freq with the binary driver even when the
-	 * performance table says they should differ.
-	 */
-	if (dev_priv->chipset == 0xd9)
-		perflvl->rop = 0;
-
-	if ((ret = calc_clk(dev, 0x00, &info->eng[0x00], perflvl->shader)) ||
-	    (ret = calc_clk(dev, 0x01, &info->eng[0x01], perflvl->rop)) ||
-	    (ret = calc_clk(dev, 0x02, &info->eng[0x02], perflvl->hub07)) ||
-	    (ret = calc_clk(dev, 0x07, &info->eng[0x07], perflvl->hub06)) ||
-	    (ret = calc_clk(dev, 0x08, &info->eng[0x08], perflvl->hub01)) ||
-	    (ret = calc_clk(dev, 0x09, &info->eng[0x09], perflvl->copy)) ||
-	    (ret = calc_clk(dev, 0x0c, &info->eng[0x0c], perflvl->daemon)) ||
-	    (ret = calc_clk(dev, 0x0e, &info->eng[0x0e], perflvl->vdec))) {
-		kfree(info);
-		return ERR_PTR(ret);
-	}
-
-	return info;
-}
-
-static void
-prog_clk(struct drm_device *dev, int clk, struct nvc0_pm_clock *info)
-{
-	/* program dividers at 137160/1371d0 first */
-	if (clk < 7 && !info->ssel) {
-		nv_mask(dev, 0x1371d0 + (clk * 0x04), 0x80003f3f, info->ddiv);
-		nv_wr32(dev, 0x137160 + (clk * 0x04), info->dsrc);
-	}
-
-	/* switch clock to non-pll mode */
-	nv_mask(dev, 0x137100, (1 << clk), 0x00000000);
-	nv_wait(dev, 0x137100, (1 << clk), 0x00000000);
-
-	/* reprogram pll */
-	if (clk < 7) {
-		/* make sure it's disabled first... */
-		u32 base = 0x137000 + (clk * 0x20);
-		u32 ctrl = nv_rd32(dev, base + 0x00);
-		if (ctrl & 0x00000001) {
-			nv_mask(dev, base + 0x00, 0x00000004, 0x00000000);
-			nv_mask(dev, base + 0x00, 0x00000001, 0x00000000);
-		}
-		/* program it to new values, if necessary */
-		if (info->ssel) {
-			nv_wr32(dev, base + 0x04, info->coef);
-			nv_mask(dev, base + 0x00, 0x00000001, 0x00000001);
-			nv_wait(dev, base + 0x00, 0x00020000, 0x00020000);
-			nv_mask(dev, base + 0x00, 0x00020004, 0x00000004);
-		}
-	}
-
-	/* select pll/non-pll mode, and program final clock divider */
-	nv_mask(dev, 0x137100, (1 << clk), info->ssel);
-	nv_wait(dev, 0x137100, (1 << clk), info->ssel);
-	nv_mask(dev, 0x137250 + (clk * 0x04), 0x00003f3f, info->mdiv);
-}
-
-int
-nvc0_pm_clocks_set(struct drm_device *dev, void *data)
-{
-	struct nvc0_pm_state *info = data;
-	int i;
-
-	for (i = 0; i < 16; i++) {
-		if (!info->eng[i].freq)
-			continue;
-		prog_clk(dev, i, &info->eng[i]);
-	}
-
-	kfree(info);
-	return 0;
-}
diff --git a/drivers/gpu/drm/nouveau/nvc0_vm.c b/drivers/gpu/drm/nouveau/nvc0_vm.c
index 30d2bd5..a179e6c 100644
--- a/drivers/gpu/drm/nouveau/nvc0_vm.c
+++ b/drivers/gpu/drm/nouveau/nvc0_vm.c
@@ -77,11 +77,9 @@ void
 nvc0_vm_map_sg(struct nouveau_vma *vma, struct nouveau_gpuobj *pgt,
 	       struct nouveau_mem *mem, u32 pte, u32 cnt, dma_addr_t *list)
 {
-	u32 target = (vma->access & NV_MEM_ACCESS_NOSNOOP) ? 7 : 5;
-
 	pte <<= 3;
 	while (cnt--) {
-		u64 phys = nvc0_vm_addr(vma, *list++, mem->memtype, target);
+		u64 phys = nvc0_vm_addr(vma, *list++, mem->memtype, 5);
 		nv_wo32(pgt, pte + 0, lower_32_bits(phys));
 		nv_wo32(pgt, pte + 4, upper_32_bits(phys));
 		pte += 8;
@@ -107,11 +105,7 @@ nvc0_vm_flush(struct nouveau_vm *vm)
 	struct drm_device *dev = vm->dev;
 	struct nouveau_vm_pgd *vpgd;
 	unsigned long flags;
-	u32 engine;
-
-	engine = 1;
-	if (vm == dev_priv->bar1_vm || vm == dev_priv->bar3_vm)
-		engine |= 4;
+	u32 engine = (dev_priv->chan_vm == vm) ? 1 : 5;
 
 	pinstmem->flush(vm->dev);
 
diff --git a/drivers/gpu/drm/nouveau/nvc0_vram.c b/drivers/gpu/drm/nouveau/nvc0_vram.c
index a7eef89..67c6ec6 100644
--- a/drivers/gpu/drm/nouveau/nvc0_vram.c
+++ b/drivers/gpu/drm/nouveau/nvc0_vram.c
@@ -43,7 +43,7 @@ static const u8 types[256] = {
 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
 	0, 0, 0, 3, 3, 3, 3, 1, 1, 1, 1, 0, 0, 0, 0, 0,
 	0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3,
-	3, 3, 3, 1, 0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3,
+	3, 3, 3, 3, 0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3,
 	3, 3, 0, 0, 0, 0, 0, 0, 3, 0, 0, 3, 0, 3, 0, 3,
 	3, 0, 3, 3, 3, 3, 3, 0, 0, 3, 0, 3, 0, 3, 3, 0,
 	3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 0, 1, 1, 0
@@ -61,7 +61,9 @@ nvc0_vram_new(struct drm_device *dev, u64 size, u32 align, u32 ncmin,
 	      u32 type, struct nouveau_mem **pmem)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_mm *mm = &dev_priv->engine.vram.mm;
+	struct ttm_bo_device *bdev = &dev_priv->ttm.bdev;
+	struct ttm_mem_type_manager *man = &bdev->man[TTM_PL_VRAM];
+	struct nouveau_mm *mm = man->priv;
 	struct nouveau_mm_node *r;
 	struct nouveau_mem *mem;
 	int ret;
@@ -103,58 +105,9 @@ int
 nvc0_vram_init(struct drm_device *dev)
 {
 	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_vram_engine *vram = &dev_priv->engine.vram;
-	const u32 rsvd_head = ( 256 * 1024) >> 12; /* vga memory */
-	const u32 rsvd_tail = (1024 * 1024) >> 12; /* vbios etc */
-	u32 parts = nv_rd32(dev, 0x022438);
-	u32 pmask = nv_rd32(dev, 0x022554);
-	u32 bsize = nv_rd32(dev, 0x10f20c);
-	u32 offset, length;
-	bool uniform = true;
-	int ret, part;
-
-	NV_DEBUG(dev, "0x100800: 0x%08x\n", nv_rd32(dev, 0x100800));
-	NV_DEBUG(dev, "parts 0x%08x mask 0x%08x\n", parts, pmask);
-
-	dev_priv->vram_type = nouveau_mem_vbios_type(dev);
-	dev_priv->vram_rank_B = !!(nv_rd32(dev, 0x10f200) & 0x00000004);
-
-	/* read amount of vram attached to each memory controller */
-	for (part = 0; part < parts; part++) {
-		if (!(pmask & (1 << part))) {
-			u32 psize = nv_rd32(dev, 0x11020c + (part * 0x1000));
-			if (psize != bsize) {
-				if (psize < bsize)
-					bsize = psize;
-				uniform = false;
-			}
-
-			NV_DEBUG(dev, "%d: mem_amount 0x%08x\n", part, psize);
-			dev_priv->vram_size += (u64)psize << 20;
-		}
-	}
-
-	/* if all controllers have the same amount attached, there's no holes */
-	if (uniform) {
-		offset = rsvd_head;
-		length = (dev_priv->vram_size >> 12) - rsvd_head - rsvd_tail;
-		return nouveau_mm_init(&vram->mm, offset, length, 1);
-	}
-
-	/* otherwise, address lowest common amount from 0GiB */
-	ret = nouveau_mm_init(&vram->mm, rsvd_head, (bsize << 8) * parts, 1);
-	if (ret)
-		return ret;
-
-	/* and the rest starting from (8GiB + common_size) */
-	offset = (0x0200000000ULL >> 12) + (bsize << 8);
-	length = (dev_priv->vram_size >> 12) - (bsize << 8) - rsvd_tail;
-
-	ret = nouveau_mm_init(&vram->mm, offset, length, 0);
-	if (ret) {
-		nouveau_mm_fini(&vram->mm);
-		return ret;
-	}
 
+	dev_priv->vram_size  = nv_rd32(dev, 0x10f20c) << 20;
+	dev_priv->vram_size *= nv_rd32(dev, 0x121c74);
+	dev_priv->vram_rblock_size = 4096;
 	return 0;
 }
diff --git a/drivers/gpu/drm/nouveau/nvd0_display.c b/drivers/gpu/drm/nouveau/nvd0_display.c
deleted file mode 100644
index dfb8a95..0000000
--- a/drivers/gpu/drm/nouveau/nvd0_display.c
+++ /dev/null
@@ -1,2098 +0,0 @@
-/*
- * Copyright 2011 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Ben Skeggs
- */
-
-#include <linux/dma-mapping.h>
-
-#include "drmP.h"
-#include "drm_crtc_helper.h"
-
-#include "nouveau_drv.h"
-#include "nouveau_connector.h"
-#include "nouveau_encoder.h"
-#include "nouveau_crtc.h"
-#include "nouveau_dma.h"
-#include "nouveau_fb.h"
-#include "nv50_display.h"
-
-#define EVO_DMA_NR 9
-
-#define EVO_MASTER  (0x00)
-#define EVO_FLIP(c) (0x01 + (c))
-#define EVO_OVLY(c) (0x05 + (c))
-#define EVO_OIMM(c) (0x09 + (c))
-#define EVO_CURS(c) (0x0d + (c))
-
-/* offsets in shared sync bo of various structures */
-#define EVO_SYNC(c, o) ((c) * 0x0100 + (o))
-#define EVO_MAST_NTFY     EVO_SYNC(  0, 0x00)
-#define EVO_FLIP_SEM0(c)  EVO_SYNC((c), 0x00)
-#define EVO_FLIP_SEM1(c)  EVO_SYNC((c), 0x10)
-
-struct evo {
-	int idx;
-	dma_addr_t handle;
-	u32 *ptr;
-	struct {
-		u32 offset;
-		u16 value;
-	} sem;
-};
-
-struct nvd0_display {
-	struct nouveau_gpuobj *mem;
-	struct nouveau_bo *sync;
-	struct evo evo[9];
-
-	struct tasklet_struct tasklet;
-	u32 modeset;
-};
-
-static struct nvd0_display *
-nvd0_display(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	return dev_priv->engine.display.priv;
-}
-
-static struct drm_crtc *
-nvd0_display_crtc_get(struct drm_encoder *encoder)
-{
-	return nouveau_encoder(encoder)->crtc;
-}
-
-/******************************************************************************
- * EVO channel helpers
- *****************************************************************************/
-static inline int
-evo_icmd(struct drm_device *dev, int id, u32 mthd, u32 data)
-{
-	int ret = 0;
-	nv_mask(dev, 0x610700 + (id * 0x10), 0x00000001, 0x00000001);
-	nv_wr32(dev, 0x610704 + (id * 0x10), data);
-	nv_mask(dev, 0x610704 + (id * 0x10), 0x80000ffc, 0x80000000 | mthd);
-	if (!nv_wait(dev, 0x610704 + (id * 0x10), 0x80000000, 0x00000000))
-		ret = -EBUSY;
-	nv_mask(dev, 0x610700 + (id * 0x10), 0x00000001, 0x00000000);
-	return ret;
-}
-
-static u32 *
-evo_wait(struct drm_device *dev, int id, int nr)
-{
-	struct nvd0_display *disp = nvd0_display(dev);
-	u32 put = nv_rd32(dev, 0x640000 + (id * 0x1000)) / 4;
-
-	if (put + nr >= (PAGE_SIZE / 4)) {
-		disp->evo[id].ptr[put] = 0x20000000;
-
-		nv_wr32(dev, 0x640000 + (id * 0x1000), 0x00000000);
-		if (!nv_wait(dev, 0x640004 + (id * 0x1000), ~0, 0x00000000)) {
-			NV_ERROR(dev, "evo %d dma stalled\n", id);
-			return NULL;
-		}
-
-		put = 0;
-	}
-
-	if (nouveau_reg_debug & NOUVEAU_REG_DEBUG_EVO)
-		NV_INFO(dev, "Evo%d: %p START\n", id, disp->evo[id].ptr + put);
-
-	return disp->evo[id].ptr + put;
-}
-
-static void
-evo_kick(u32 *push, struct drm_device *dev, int id)
-{
-	struct nvd0_display *disp = nvd0_display(dev);
-
-	if (nouveau_reg_debug & NOUVEAU_REG_DEBUG_EVO) {
-		u32 curp = nv_rd32(dev, 0x640000 + (id * 0x1000)) >> 2;
-		u32 *cur = disp->evo[id].ptr + curp;
-
-		while (cur < push)
-			NV_INFO(dev, "Evo%d: 0x%08x\n", id, *cur++);
-		NV_INFO(dev, "Evo%d: %p KICK!\n", id, push);
-	}
-
-	nv_wr32(dev, 0x640000 + (id * 0x1000), (push - disp->evo[id].ptr) << 2);
-}
-
-#define evo_mthd(p,m,s) *((p)++) = (((s) << 18) | (m))
-#define evo_data(p,d)   *((p)++) = (d)
-
-static int
-evo_init_dma(struct drm_device *dev, int ch)
-{
-	struct nvd0_display *disp = nvd0_display(dev);
-	u32 flags;
-
-	flags = 0x00000000;
-	if (ch == EVO_MASTER)
-		flags |= 0x01000000;
-
-	nv_wr32(dev, 0x610494 + (ch * 0x0010), (disp->evo[ch].handle >> 8) | 3);
-	nv_wr32(dev, 0x610498 + (ch * 0x0010), 0x00010000);
-	nv_wr32(dev, 0x61049c + (ch * 0x0010), 0x00000001);
-	nv_mask(dev, 0x610490 + (ch * 0x0010), 0x00000010, 0x00000010);
-	nv_wr32(dev, 0x640000 + (ch * 0x1000), 0x00000000);
-	nv_wr32(dev, 0x610490 + (ch * 0x0010), 0x00000013 | flags);
-	if (!nv_wait(dev, 0x610490 + (ch * 0x0010), 0x80000000, 0x00000000)) {
-		NV_ERROR(dev, "PDISP: ch%d 0x%08x\n", ch,
-			      nv_rd32(dev, 0x610490 + (ch * 0x0010)));
-		return -EBUSY;
-	}
-
-	nv_mask(dev, 0x610090, (1 << ch), (1 << ch));
-	nv_mask(dev, 0x6100a0, (1 << ch), (1 << ch));
-	return 0;
-}
-
-static void
-evo_fini_dma(struct drm_device *dev, int ch)
-{
-	if (!(nv_rd32(dev, 0x610490 + (ch * 0x0010)) & 0x00000010))
-		return;
-
-	nv_mask(dev, 0x610490 + (ch * 0x0010), 0x00000010, 0x00000000);
-	nv_mask(dev, 0x610490 + (ch * 0x0010), 0x00000003, 0x00000000);
-	nv_wait(dev, 0x610490 + (ch * 0x0010), 0x80000000, 0x00000000);
-	nv_mask(dev, 0x610090, (1 << ch), 0x00000000);
-	nv_mask(dev, 0x6100a0, (1 << ch), 0x00000000);
-}
-
-static inline void
-evo_piow(struct drm_device *dev, int ch, u16 mthd, u32 data)
-{
-	nv_wr32(dev, 0x640000 + (ch * 0x1000) + mthd, data);
-}
-
-static int
-evo_init_pio(struct drm_device *dev, int ch)
-{
-	nv_wr32(dev, 0x610490 + (ch * 0x0010), 0x00000001);
-	if (!nv_wait(dev, 0x610490 + (ch * 0x0010), 0x00010000, 0x00010000)) {
-		NV_ERROR(dev, "PDISP: ch%d 0x%08x\n", ch,
-			      nv_rd32(dev, 0x610490 + (ch * 0x0010)));
-		return -EBUSY;
-	}
-
-	nv_mask(dev, 0x610090, (1 << ch), (1 << ch));
-	nv_mask(dev, 0x6100a0, (1 << ch), (1 << ch));
-	return 0;
-}
-
-static void
-evo_fini_pio(struct drm_device *dev, int ch)
-{
-	if (!(nv_rd32(dev, 0x610490 + (ch * 0x0010)) & 0x00000001))
-		return;
-
-	nv_mask(dev, 0x610490 + (ch * 0x0010), 0x00000010, 0x00000010);
-	nv_mask(dev, 0x610490 + (ch * 0x0010), 0x00000001, 0x00000000);
-	nv_wait(dev, 0x610490 + (ch * 0x0010), 0x00010000, 0x00000000);
-	nv_mask(dev, 0x610090, (1 << ch), 0x00000000);
-	nv_mask(dev, 0x6100a0, (1 << ch), 0x00000000);
-}
-
-static bool
-evo_sync_wait(void *data)
-{
-	return nouveau_bo_rd32(data, EVO_MAST_NTFY) != 0x00000000;
-}
-
-static int
-evo_sync(struct drm_device *dev, int ch)
-{
-	struct nvd0_display *disp = nvd0_display(dev);
-	u32 *push = evo_wait(dev, ch, 8);
-	if (push) {
-		nouveau_bo_wr32(disp->sync, EVO_MAST_NTFY, 0x00000000);
-		evo_mthd(push, 0x0084, 1);
-		evo_data(push, 0x80000000 | EVO_MAST_NTFY);
-		evo_mthd(push, 0x0080, 2);
-		evo_data(push, 0x00000000);
-		evo_data(push, 0x00000000);
-		evo_kick(push, dev, ch);
-		if (nv_wait_cb(dev, evo_sync_wait, disp->sync))
-			return 0;
-	}
-
-	return -EBUSY;
-}
-
-/******************************************************************************
- * Page flipping channel
- *****************************************************************************/
-struct nouveau_bo *
-nvd0_display_crtc_sema(struct drm_device *dev, int crtc)
-{
-	return nvd0_display(dev)->sync;
-}
-
-void
-nvd0_display_flip_stop(struct drm_crtc *crtc)
-{
-	struct nvd0_display *disp = nvd0_display(crtc->dev);
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
-	struct evo *evo = &disp->evo[EVO_FLIP(nv_crtc->index)];
-	u32 *push;
-
-	push = evo_wait(crtc->dev, evo->idx, 8);
-	if (push) {
-		evo_mthd(push, 0x0084, 1);
-		evo_data(push, 0x00000000);
-		evo_mthd(push, 0x0094, 1);
-		evo_data(push, 0x00000000);
-		evo_mthd(push, 0x00c0, 1);
-		evo_data(push, 0x00000000);
-		evo_mthd(push, 0x0080, 1);
-		evo_data(push, 0x00000000);
-		evo_kick(push, crtc->dev, evo->idx);
-	}
-}
-
-int
-nvd0_display_flip_next(struct drm_crtc *crtc, struct drm_framebuffer *fb,
-		       struct nouveau_channel *chan, u32 swap_interval)
-{
-	struct nouveau_framebuffer *nv_fb = nouveau_framebuffer(fb);
-	struct nvd0_display *disp = nvd0_display(crtc->dev);
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
-	struct evo *evo = &disp->evo[EVO_FLIP(nv_crtc->index)];
-	u64 offset;
-	u32 *push;
-	int ret;
-
-	evo_sync(crtc->dev, EVO_MASTER);
-
-	swap_interval <<= 4;
-	if (swap_interval == 0)
-		swap_interval |= 0x100;
-
-	push = evo_wait(crtc->dev, evo->idx, 128);
-	if (unlikely(push == NULL))
-		return -EBUSY;
-
-	/* synchronise with the rendering channel, if necessary */
-	if (likely(chan)) {
-		ret = RING_SPACE(chan, 10);
-		if (ret)
-			return ret;
-
-		offset  = chan->dispc_vma[nv_crtc->index].offset;
-		offset += evo->sem.offset;
-
-		BEGIN_NVC0(chan, 2, NvSubM2MF, 0x0010, 4);
-		OUT_RING  (chan, upper_32_bits(offset));
-		OUT_RING  (chan, lower_32_bits(offset));
-		OUT_RING  (chan, 0xf00d0000 | evo->sem.value);
-		OUT_RING  (chan, 0x1002);
-		BEGIN_NVC0(chan, 2, NvSubM2MF, 0x0010, 4);
-		OUT_RING  (chan, upper_32_bits(offset));
-		OUT_RING  (chan, lower_32_bits(offset ^ 0x10));
-		OUT_RING  (chan, 0x74b1e000);
-		OUT_RING  (chan, 0x1001);
-		FIRE_RING (chan);
-	} else {
-		nouveau_bo_wr32(disp->sync, evo->sem.offset / 4,
-				0xf00d0000 | evo->sem.value);
-		evo_sync(crtc->dev, EVO_MASTER);
-	}
-
-	/* queue the flip */
-	evo_mthd(push, 0x0100, 1);
-	evo_data(push, 0xfffe0000);
-	evo_mthd(push, 0x0084, 1);
-	evo_data(push, swap_interval);
-	if (!(swap_interval & 0x00000100)) {
-		evo_mthd(push, 0x00e0, 1);
-		evo_data(push, 0x40000000);
-	}
-	evo_mthd(push, 0x0088, 4);
-	evo_data(push, evo->sem.offset);
-	evo_data(push, 0xf00d0000 | evo->sem.value);
-	evo_data(push, 0x74b1e000);
-	evo_data(push, NvEvoSync);
-	evo_mthd(push, 0x00a0, 2);
-	evo_data(push, 0x00000000);
-	evo_data(push, 0x00000000);
-	evo_mthd(push, 0x00c0, 1);
-	evo_data(push, nv_fb->r_dma);
-	evo_mthd(push, 0x0110, 2);
-	evo_data(push, 0x00000000);
-	evo_data(push, 0x00000000);
-	evo_mthd(push, 0x0400, 5);
-	evo_data(push, nv_fb->nvbo->bo.offset >> 8);
-	evo_data(push, 0);
-	evo_data(push, (fb->height << 16) | fb->width);
-	evo_data(push, nv_fb->r_pitch);
-	evo_data(push, nv_fb->r_format);
-	evo_mthd(push, 0x0080, 1);
-	evo_data(push, 0x00000000);
-	evo_kick(push, crtc->dev, evo->idx);
-
-	evo->sem.offset ^= 0x10;
-	evo->sem.value++;
-	return 0;
-}
-
-/******************************************************************************
- * CRTC
- *****************************************************************************/
-static int
-nvd0_crtc_set_dither(struct nouveau_crtc *nv_crtc, bool update)
-{
-	struct drm_device *dev = nv_crtc->base.dev;
-	struct nouveau_connector *nv_connector;
-	struct drm_connector *connector;
-	u32 *push, mode = 0x00;
-
-	nv_connector = nouveau_crtc_connector_get(nv_crtc);
-	connector = &nv_connector->base;
-	if (nv_connector->dithering_mode == DITHERING_MODE_AUTO) {
-		if (nv_crtc->base.fb->depth > connector->display_info.bpc * 3)
-			mode = DITHERING_MODE_DYNAMIC2X2;
-	} else {
-		mode = nv_connector->dithering_mode;
-	}
-
-	if (nv_connector->dithering_depth == DITHERING_DEPTH_AUTO) {
-		if (connector->display_info.bpc >= 8)
-			mode |= DITHERING_DEPTH_8BPC;
-	} else {
-		mode |= nv_connector->dithering_depth;
-	}
-
-	push = evo_wait(dev, EVO_MASTER, 4);
-	if (push) {
-		evo_mthd(push, 0x0490 + (nv_crtc->index * 0x300), 1);
-		evo_data(push, mode);
-		if (update) {
-			evo_mthd(push, 0x0080, 1);
-			evo_data(push, 0x00000000);
-		}
-		evo_kick(push, dev, EVO_MASTER);
-	}
-
-	return 0;
-}
-
-static int
-nvd0_crtc_set_scale(struct nouveau_crtc *nv_crtc, bool update)
-{
-	struct drm_display_mode *omode, *umode = &nv_crtc->base.mode;
-	struct drm_device *dev = nv_crtc->base.dev;
-	struct drm_crtc *crtc = &nv_crtc->base;
-	struct nouveau_connector *nv_connector;
-	int mode = DRM_MODE_SCALE_NONE;
-	u32 oX, oY, *push;
-
-	/* start off at the resolution we programmed the crtc for, this
-	 * effectively handles NONE/FULL scaling
-	 */
-	nv_connector = nouveau_crtc_connector_get(nv_crtc);
-	if (nv_connector && nv_connector->native_mode)
-		mode = nv_connector->scaling_mode;
-
-	if (mode != DRM_MODE_SCALE_NONE)
-		omode = nv_connector->native_mode;
-	else
-		omode = umode;
-
-	oX = omode->hdisplay;
-	oY = omode->vdisplay;
-	if (omode->flags & DRM_MODE_FLAG_DBLSCAN)
-		oY *= 2;
-
-	/* add overscan compensation if necessary, will keep the aspect
-	 * ratio the same as the backend mode unless overridden by the
-	 * user setting both hborder and vborder properties.
-	 */
-	if (nv_connector && ( nv_connector->underscan == UNDERSCAN_ON ||
-			     (nv_connector->underscan == UNDERSCAN_AUTO &&
-			      nv_connector->edid &&
-			      drm_detect_hdmi_monitor(nv_connector->edid)))) {
-		u32 bX = nv_connector->underscan_hborder;
-		u32 bY = nv_connector->underscan_vborder;
-		u32 aspect = (oY << 19) / oX;
-
-		if (bX) {
-			oX -= (bX * 2);
-			if (bY) oY -= (bY * 2);
-			else    oY  = ((oX * aspect) + (aspect / 2)) >> 19;
-		} else {
-			oX -= (oX >> 4) + 32;
-			if (bY) oY -= (bY * 2);
-			else    oY  = ((oX * aspect) + (aspect / 2)) >> 19;
-		}
-	}
-
-	/* handle CENTER/ASPECT scaling, taking into account the areas
-	 * removed already for overscan compensation
-	 */
-	switch (mode) {
-	case DRM_MODE_SCALE_CENTER:
-		oX = min((u32)umode->hdisplay, oX);
-		oY = min((u32)umode->vdisplay, oY);
-		/* fall-through */
-	case DRM_MODE_SCALE_ASPECT:
-		if (oY < oX) {
-			u32 aspect = (umode->hdisplay << 19) / umode->vdisplay;
-			oX = ((oY * aspect) + (aspect / 2)) >> 19;
-		} else {
-			u32 aspect = (umode->vdisplay << 19) / umode->hdisplay;
-			oY = ((oX * aspect) + (aspect / 2)) >> 19;
-		}
-		break;
-	default:
-		break;
-	}
-
-	push = evo_wait(dev, EVO_MASTER, 8);
-	if (push) {
-		evo_mthd(push, 0x04c0 + (nv_crtc->index * 0x300), 3);
-		evo_data(push, (oY << 16) | oX);
-		evo_data(push, (oY << 16) | oX);
-		evo_data(push, (oY << 16) | oX);
-		evo_mthd(push, 0x0494 + (nv_crtc->index * 0x300), 1);
-		evo_data(push, 0x00000000);
-		evo_mthd(push, 0x04b8 + (nv_crtc->index * 0x300), 1);
-		evo_data(push, (umode->vdisplay << 16) | umode->hdisplay);
-		evo_kick(push, dev, EVO_MASTER);
-		if (update) {
-			nvd0_display_flip_stop(crtc);
-			nvd0_display_flip_next(crtc, crtc->fb, NULL, 1);
-		}
-	}
-
-	return 0;
-}
-
-static int
-nvd0_crtc_set_image(struct nouveau_crtc *nv_crtc, struct drm_framebuffer *fb,
-		    int x, int y, bool update)
-{
-	struct nouveau_framebuffer *nvfb = nouveau_framebuffer(fb);
-	u32 *push;
-
-	push = evo_wait(fb->dev, EVO_MASTER, 16);
-	if (push) {
-		evo_mthd(push, 0x0460 + (nv_crtc->index * 0x300), 1);
-		evo_data(push, nvfb->nvbo->bo.offset >> 8);
-		evo_mthd(push, 0x0468 + (nv_crtc->index * 0x300), 4);
-		evo_data(push, (fb->height << 16) | fb->width);
-		evo_data(push, nvfb->r_pitch);
-		evo_data(push, nvfb->r_format);
-		evo_data(push, nvfb->r_dma);
-		evo_mthd(push, 0x04b0 + (nv_crtc->index * 0x300), 1);
-		evo_data(push, (y << 16) | x);
-		if (update) {
-			evo_mthd(push, 0x0080, 1);
-			evo_data(push, 0x00000000);
-		}
-		evo_kick(push, fb->dev, EVO_MASTER);
-	}
-
-	nv_crtc->fb.tile_flags = nvfb->r_dma;
-	return 0;
-}
-
-static void
-nvd0_crtc_cursor_show(struct nouveau_crtc *nv_crtc, bool show, bool update)
-{
-	struct drm_device *dev = nv_crtc->base.dev;
-	u32 *push = evo_wait(dev, EVO_MASTER, 16);
-	if (push) {
-		if (show) {
-			evo_mthd(push, 0x0480 + (nv_crtc->index * 0x300), 2);
-			evo_data(push, 0x85000000);
-			evo_data(push, nv_crtc->cursor.nvbo->bo.offset >> 8);
-			evo_mthd(push, 0x048c + (nv_crtc->index * 0x300), 1);
-			evo_data(push, NvEvoVRAM);
-		} else {
-			evo_mthd(push, 0x0480 + (nv_crtc->index * 0x300), 1);
-			evo_data(push, 0x05000000);
-			evo_mthd(push, 0x048c + (nv_crtc->index * 0x300), 1);
-			evo_data(push, 0x00000000);
-		}
-
-		if (update) {
-			evo_mthd(push, 0x0080, 1);
-			evo_data(push, 0x00000000);
-		}
-
-		evo_kick(push, dev, EVO_MASTER);
-	}
-}
-
-static void
-nvd0_crtc_dpms(struct drm_crtc *crtc, int mode)
-{
-}
-
-static void
-nvd0_crtc_prepare(struct drm_crtc *crtc)
-{
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
-	u32 *push;
-
-	nvd0_display_flip_stop(crtc);
-
-	push = evo_wait(crtc->dev, EVO_MASTER, 2);
-	if (push) {
-		evo_mthd(push, 0x0474 + (nv_crtc->index * 0x300), 1);
-		evo_data(push, 0x00000000);
-		evo_mthd(push, 0x0440 + (nv_crtc->index * 0x300), 1);
-		evo_data(push, 0x03000000);
-		evo_mthd(push, 0x045c + (nv_crtc->index * 0x300), 1);
-		evo_data(push, 0x00000000);
-		evo_kick(push, crtc->dev, EVO_MASTER);
-	}
-
-	nvd0_crtc_cursor_show(nv_crtc, false, false);
-}
-
-static void
-nvd0_crtc_commit(struct drm_crtc *crtc)
-{
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
-	u32 *push;
-
-	push = evo_wait(crtc->dev, EVO_MASTER, 32);
-	if (push) {
-		evo_mthd(push, 0x0474 + (nv_crtc->index * 0x300), 1);
-		evo_data(push, nv_crtc->fb.tile_flags);
-		evo_mthd(push, 0x0440 + (nv_crtc->index * 0x300), 4);
-		evo_data(push, 0x83000000);
-		evo_data(push, nv_crtc->lut.nvbo->bo.offset >> 8);
-		evo_data(push, 0x00000000);
-		evo_data(push, 0x00000000);
-		evo_mthd(push, 0x045c + (nv_crtc->index * 0x300), 1);
-		evo_data(push, NvEvoVRAM);
-		evo_mthd(push, 0x0430 + (nv_crtc->index * 0x300), 1);
-		evo_data(push, 0xffffff00);
-		evo_kick(push, crtc->dev, EVO_MASTER);
-	}
-
-	nvd0_crtc_cursor_show(nv_crtc, nv_crtc->cursor.visible, true);
-	nvd0_display_flip_next(crtc, crtc->fb, NULL, 1);
-}
-
-static bool
-nvd0_crtc_mode_fixup(struct drm_crtc *crtc, struct drm_display_mode *mode,
-		     struct drm_display_mode *adjusted_mode)
-{
-	return true;
-}
-
-static int
-nvd0_crtc_swap_fbs(struct drm_crtc *crtc, struct drm_framebuffer *old_fb)
-{
-	struct nouveau_framebuffer *nvfb = nouveau_framebuffer(crtc->fb);
-	int ret;
-
-	ret = nouveau_bo_pin(nvfb->nvbo, TTM_PL_FLAG_VRAM);
-	if (ret)
-		return ret;
-
-	if (old_fb) {
-		nvfb = nouveau_framebuffer(old_fb);
-		nouveau_bo_unpin(nvfb->nvbo);
-	}
-
-	return 0;
-}
-
-static int
-nvd0_crtc_mode_set(struct drm_crtc *crtc, struct drm_display_mode *umode,
-		   struct drm_display_mode *mode, int x, int y,
-		   struct drm_framebuffer *old_fb)
-{
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
-	struct nouveau_connector *nv_connector;
-	u32 ilace = (mode->flags & DRM_MODE_FLAG_INTERLACE) ? 2 : 1;
-	u32 vscan = (mode->flags & DRM_MODE_FLAG_DBLSCAN) ? 2 : 1;
-	u32 hactive, hsynce, hbackp, hfrontp, hblanke, hblanks;
-	u32 vactive, vsynce, vbackp, vfrontp, vblanke, vblanks;
-	u32 vblan2e = 0, vblan2s = 1;
-	u32 *push;
-	int ret;
-
-	hactive = mode->htotal;
-	hsynce  = mode->hsync_end - mode->hsync_start - 1;
-	hbackp  = mode->htotal - mode->hsync_end;
-	hblanke = hsynce + hbackp;
-	hfrontp = mode->hsync_start - mode->hdisplay;
-	hblanks = mode->htotal - hfrontp - 1;
-
-	vactive = mode->vtotal * vscan / ilace;
-	vsynce  = ((mode->vsync_end - mode->vsync_start) * vscan / ilace) - 1;
-	vbackp  = (mode->vtotal - mode->vsync_end) * vscan / ilace;
-	vblanke = vsynce + vbackp;
-	vfrontp = (mode->vsync_start - mode->vdisplay) * vscan / ilace;
-	vblanks = vactive - vfrontp - 1;
-	if (mode->flags & DRM_MODE_FLAG_INTERLACE) {
-		vblan2e = vactive + vsynce + vbackp;
-		vblan2s = vblan2e + (mode->vdisplay * vscan / ilace);
-		vactive = (vactive * 2) + 1;
-	}
-
-	ret = nvd0_crtc_swap_fbs(crtc, old_fb);
-	if (ret)
-		return ret;
-
-	push = evo_wait(crtc->dev, EVO_MASTER, 64);
-	if (push) {
-		evo_mthd(push, 0x0410 + (nv_crtc->index * 0x300), 6);
-		evo_data(push, 0x00000000);
-		evo_data(push, (vactive << 16) | hactive);
-		evo_data(push, ( vsynce << 16) | hsynce);
-		evo_data(push, (vblanke << 16) | hblanke);
-		evo_data(push, (vblanks << 16) | hblanks);
-		evo_data(push, (vblan2e << 16) | vblan2s);
-		evo_mthd(push, 0x042c + (nv_crtc->index * 0x300), 1);
-		evo_data(push, 0x00000000); /* ??? */
-		evo_mthd(push, 0x0450 + (nv_crtc->index * 0x300), 3);
-		evo_data(push, mode->clock * 1000);
-		evo_data(push, 0x00200000); /* ??? */
-		evo_data(push, mode->clock * 1000);
-		evo_mthd(push, 0x04d0 + (nv_crtc->index * 0x300), 2);
-		evo_data(push, 0x00000311);
-		evo_data(push, 0x00000100);
-		evo_kick(push, crtc->dev, EVO_MASTER);
-	}
-
-	nv_connector = nouveau_crtc_connector_get(nv_crtc);
-	nvd0_crtc_set_dither(nv_crtc, false);
-	nvd0_crtc_set_scale(nv_crtc, false);
-	nvd0_crtc_set_image(nv_crtc, crtc->fb, x, y, false);
-	return 0;
-}
-
-static int
-nvd0_crtc_mode_set_base(struct drm_crtc *crtc, int x, int y,
-			struct drm_framebuffer *old_fb)
-{
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
-	int ret;
-
-	if (!crtc->fb) {
-		NV_DEBUG_KMS(crtc->dev, "No FB bound\n");
-		return 0;
-	}
-
-	ret = nvd0_crtc_swap_fbs(crtc, old_fb);
-	if (ret)
-		return ret;
-
-	nvd0_display_flip_stop(crtc);
-	nvd0_crtc_set_image(nv_crtc, crtc->fb, x, y, true);
-	nvd0_display_flip_next(crtc, crtc->fb, NULL, 1);
-	return 0;
-}
-
-static int
-nvd0_crtc_mode_set_base_atomic(struct drm_crtc *crtc,
-			       struct drm_framebuffer *fb, int x, int y,
-			       enum mode_set_atomic state)
-{
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
-	nvd0_display_flip_stop(crtc);
-	nvd0_crtc_set_image(nv_crtc, fb, x, y, true);
-	return 0;
-}
-
-static void
-nvd0_crtc_lut_load(struct drm_crtc *crtc)
-{
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
-	void __iomem *lut = nvbo_kmap_obj_iovirtual(nv_crtc->lut.nvbo);
-	int i;
-
-	for (i = 0; i < 256; i++) {
-		writew(0x6000 + (nv_crtc->lut.r[i] >> 2), lut + (i * 0x20) + 0);
-		writew(0x6000 + (nv_crtc->lut.g[i] >> 2), lut + (i * 0x20) + 2);
-		writew(0x6000 + (nv_crtc->lut.b[i] >> 2), lut + (i * 0x20) + 4);
-	}
-}
-
-static int
-nvd0_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv,
-		     uint32_t handle, uint32_t width, uint32_t height)
-{
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
-	struct drm_device *dev = crtc->dev;
-	struct drm_gem_object *gem;
-	struct nouveau_bo *nvbo;
-	bool visible = (handle != 0);
-	int i, ret = 0;
-
-	if (visible) {
-		if (width != 64 || height != 64)
-			return -EINVAL;
-
-		gem = drm_gem_object_lookup(dev, file_priv, handle);
-		if (unlikely(!gem))
-			return -ENOENT;
-		nvbo = nouveau_gem_object(gem);
-
-		ret = nouveau_bo_map(nvbo);
-		if (ret == 0) {
-			for (i = 0; i < 64 * 64; i++) {
-				u32 v = nouveau_bo_rd32(nvbo, i);
-				nouveau_bo_wr32(nv_crtc->cursor.nvbo, i, v);
-			}
-			nouveau_bo_unmap(nvbo);
-		}
-
-		drm_gem_object_unreference_unlocked(gem);
-	}
-
-	if (visible != nv_crtc->cursor.visible) {
-		nvd0_crtc_cursor_show(nv_crtc, visible, true);
-		nv_crtc->cursor.visible = visible;
-	}
-
-	return ret;
-}
-
-static int
-nvd0_crtc_cursor_move(struct drm_crtc *crtc, int x, int y)
-{
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
-	int ch = EVO_CURS(nv_crtc->index);
-
-	evo_piow(crtc->dev, ch, 0x0084, (y << 16) | x);
-	evo_piow(crtc->dev, ch, 0x0080, 0x00000000);
-	return 0;
-}
-
-static void
-nvd0_crtc_gamma_set(struct drm_crtc *crtc, u16 *r, u16 *g, u16 *b,
-		    uint32_t start, uint32_t size)
-{
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
-	u32 end = max(start + size, (u32)256);
-	u32 i;
-
-	for (i = start; i < end; i++) {
-		nv_crtc->lut.r[i] = r[i];
-		nv_crtc->lut.g[i] = g[i];
-		nv_crtc->lut.b[i] = b[i];
-	}
-
-	nvd0_crtc_lut_load(crtc);
-}
-
-static void
-nvd0_crtc_destroy(struct drm_crtc *crtc)
-{
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(crtc);
-	nouveau_bo_unmap(nv_crtc->cursor.nvbo);
-	nouveau_bo_ref(NULL, &nv_crtc->cursor.nvbo);
-	nouveau_bo_unmap(nv_crtc->lut.nvbo);
-	nouveau_bo_ref(NULL, &nv_crtc->lut.nvbo);
-	drm_crtc_cleanup(crtc);
-	kfree(crtc);
-}
-
-static const struct drm_crtc_helper_funcs nvd0_crtc_hfunc = {
-	.dpms = nvd0_crtc_dpms,
-	.prepare = nvd0_crtc_prepare,
-	.commit = nvd0_crtc_commit,
-	.mode_fixup = nvd0_crtc_mode_fixup,
-	.mode_set = nvd0_crtc_mode_set,
-	.mode_set_base = nvd0_crtc_mode_set_base,
-	.mode_set_base_atomic = nvd0_crtc_mode_set_base_atomic,
-	.load_lut = nvd0_crtc_lut_load,
-};
-
-static const struct drm_crtc_funcs nvd0_crtc_func = {
-	.cursor_set = nvd0_crtc_cursor_set,
-	.cursor_move = nvd0_crtc_cursor_move,
-	.gamma_set = nvd0_crtc_gamma_set,
-	.set_config = drm_crtc_helper_set_config,
-	.destroy = nvd0_crtc_destroy,
-	.page_flip = nouveau_crtc_page_flip,
-};
-
-static void
-nvd0_cursor_set_pos(struct nouveau_crtc *nv_crtc, int x, int y)
-{
-}
-
-static void
-nvd0_cursor_set_offset(struct nouveau_crtc *nv_crtc, uint32_t offset)
-{
-}
-
-static int
-nvd0_crtc_create(struct drm_device *dev, int index)
-{
-	struct nouveau_crtc *nv_crtc;
-	struct drm_crtc *crtc;
-	int ret, i;
-
-	nv_crtc = kzalloc(sizeof(*nv_crtc), GFP_KERNEL);
-	if (!nv_crtc)
-		return -ENOMEM;
-
-	nv_crtc->index = index;
-	nv_crtc->set_dither = nvd0_crtc_set_dither;
-	nv_crtc->set_scale = nvd0_crtc_set_scale;
-	nv_crtc->cursor.set_offset = nvd0_cursor_set_offset;
-	nv_crtc->cursor.set_pos = nvd0_cursor_set_pos;
-	for (i = 0; i < 256; i++) {
-		nv_crtc->lut.r[i] = i << 8;
-		nv_crtc->lut.g[i] = i << 8;
-		nv_crtc->lut.b[i] = i << 8;
-	}
-
-	crtc = &nv_crtc->base;
-	drm_crtc_init(dev, crtc, &nvd0_crtc_func);
-	drm_crtc_helper_add(crtc, &nvd0_crtc_hfunc);
-	drm_mode_crtc_set_gamma_size(crtc, 256);
-
-	ret = nouveau_bo_new(dev, 64 * 64 * 4, 0x100, TTM_PL_FLAG_VRAM,
-			     0, 0x0000, &nv_crtc->cursor.nvbo);
-	if (!ret) {
-		ret = nouveau_bo_pin(nv_crtc->cursor.nvbo, TTM_PL_FLAG_VRAM);
-		if (!ret)
-			ret = nouveau_bo_map(nv_crtc->cursor.nvbo);
-		if (ret)
-			nouveau_bo_ref(NULL, &nv_crtc->cursor.nvbo);
-	}
-
-	if (ret)
-		goto out;
-
-	ret = nouveau_bo_new(dev, 8192, 0x100, TTM_PL_FLAG_VRAM,
-			     0, 0x0000, &nv_crtc->lut.nvbo);
-	if (!ret) {
-		ret = nouveau_bo_pin(nv_crtc->lut.nvbo, TTM_PL_FLAG_VRAM);
-		if (!ret)
-			ret = nouveau_bo_map(nv_crtc->lut.nvbo);
-		if (ret)
-			nouveau_bo_ref(NULL, &nv_crtc->lut.nvbo);
-	}
-
-	if (ret)
-		goto out;
-
-	nvd0_crtc_lut_load(crtc);
-
-out:
-	if (ret)
-		nvd0_crtc_destroy(crtc);
-	return ret;
-}
-
-/******************************************************************************
- * DAC
- *****************************************************************************/
-static void
-nvd0_dac_dpms(struct drm_encoder *encoder, int mode)
-{
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct drm_device *dev = encoder->dev;
-	int or = nv_encoder->or;
-	u32 dpms_ctrl;
-
-	dpms_ctrl = 0x80000000;
-	if (mode == DRM_MODE_DPMS_STANDBY || mode == DRM_MODE_DPMS_OFF)
-		dpms_ctrl |= 0x00000001;
-	if (mode == DRM_MODE_DPMS_SUSPEND || mode == DRM_MODE_DPMS_OFF)
-		dpms_ctrl |= 0x00000004;
-
-	nv_wait(dev, 0x61a004 + (or * 0x0800), 0x80000000, 0x00000000);
-	nv_mask(dev, 0x61a004 + (or * 0x0800), 0xc000007f, dpms_ctrl);
-	nv_wait(dev, 0x61a004 + (or * 0x0800), 0x80000000, 0x00000000);
-}
-
-static bool
-nvd0_dac_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *mode,
-		    struct drm_display_mode *adjusted_mode)
-{
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct nouveau_connector *nv_connector;
-
-	nv_connector = nouveau_encoder_connector_get(nv_encoder);
-	if (nv_connector && nv_connector->native_mode) {
-		if (nv_connector->scaling_mode != DRM_MODE_SCALE_NONE) {
-			int id = adjusted_mode->base.id;
-			*adjusted_mode = *nv_connector->native_mode;
-			adjusted_mode->base.id = id;
-		}
-	}
-
-	return true;
-}
-
-static void
-nvd0_dac_commit(struct drm_encoder *encoder)
-{
-}
-
-static void
-nvd0_dac_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode,
-		  struct drm_display_mode *adjusted_mode)
-{
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc);
-	u32 syncs, magic, *push;
-
-	syncs = 0x00000001;
-	if (mode->flags & DRM_MODE_FLAG_NHSYNC)
-		syncs |= 0x00000008;
-	if (mode->flags & DRM_MODE_FLAG_NVSYNC)
-		syncs |= 0x00000010;
-
-	magic = 0x31ec6000 | (nv_crtc->index << 25);
-	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
-		magic |= 0x00000001;
-
-	nvd0_dac_dpms(encoder, DRM_MODE_DPMS_ON);
-
-	push = evo_wait(encoder->dev, EVO_MASTER, 8);
-	if (push) {
-		evo_mthd(push, 0x0404 + (nv_crtc->index * 0x300), 2);
-		evo_data(push, syncs);
-		evo_data(push, magic);
-		evo_mthd(push, 0x0180 + (nv_encoder->or * 0x020), 2);
-		evo_data(push, 1 << nv_crtc->index);
-		evo_data(push, 0x00ff);
-		evo_kick(push, encoder->dev, EVO_MASTER);
-	}
-
-	nv_encoder->crtc = encoder->crtc;
-}
-
-static void
-nvd0_dac_disconnect(struct drm_encoder *encoder)
-{
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct drm_device *dev = encoder->dev;
-	u32 *push;
-
-	if (nv_encoder->crtc) {
-		nvd0_crtc_prepare(nv_encoder->crtc);
-
-		push = evo_wait(dev, EVO_MASTER, 4);
-		if (push) {
-			evo_mthd(push, 0x0180 + (nv_encoder->or * 0x20), 1);
-			evo_data(push, 0x00000000);
-			evo_mthd(push, 0x0080, 1);
-			evo_data(push, 0x00000000);
-			evo_kick(push, dev, EVO_MASTER);
-		}
-
-		nv_encoder->crtc = NULL;
-	}
-}
-
-static enum drm_connector_status
-nvd0_dac_detect(struct drm_encoder *encoder, struct drm_connector *connector)
-{
-	enum drm_connector_status status = connector_status_disconnected;
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct drm_device *dev = encoder->dev;
-	int or = nv_encoder->or;
-	u32 load;
-
-	nv_wr32(dev, 0x61a00c + (or * 0x800), 0x00100000);
-	udelay(9500);
-	nv_wr32(dev, 0x61a00c + (or * 0x800), 0x80000000);
-
-	load = nv_rd32(dev, 0x61a00c + (or * 0x800));
-	if ((load & 0x38000000) == 0x38000000)
-		status = connector_status_connected;
-
-	nv_wr32(dev, 0x61a00c + (or * 0x800), 0x00000000);
-	return status;
-}
-
-static void
-nvd0_dac_destroy(struct drm_encoder *encoder)
-{
-	drm_encoder_cleanup(encoder);
-	kfree(encoder);
-}
-
-static const struct drm_encoder_helper_funcs nvd0_dac_hfunc = {
-	.dpms = nvd0_dac_dpms,
-	.mode_fixup = nvd0_dac_mode_fixup,
-	.prepare = nvd0_dac_disconnect,
-	.commit = nvd0_dac_commit,
-	.mode_set = nvd0_dac_mode_set,
-	.disable = nvd0_dac_disconnect,
-	.get_crtc = nvd0_display_crtc_get,
-	.detect = nvd0_dac_detect
-};
-
-static const struct drm_encoder_funcs nvd0_dac_func = {
-	.destroy = nvd0_dac_destroy,
-};
-
-static int
-nvd0_dac_create(struct drm_connector *connector, struct dcb_entry *dcbe)
-{
-	struct drm_device *dev = connector->dev;
-	struct nouveau_encoder *nv_encoder;
-	struct drm_encoder *encoder;
-
-	nv_encoder = kzalloc(sizeof(*nv_encoder), GFP_KERNEL);
-	if (!nv_encoder)
-		return -ENOMEM;
-	nv_encoder->dcb = dcbe;
-	nv_encoder->or = ffs(dcbe->or) - 1;
-
-	encoder = to_drm_encoder(nv_encoder);
-	encoder->possible_crtcs = dcbe->heads;
-	encoder->possible_clones = 0;
-	drm_encoder_init(dev, encoder, &nvd0_dac_func, DRM_MODE_ENCODER_DAC);
-	drm_encoder_helper_add(encoder, &nvd0_dac_hfunc);
-
-	drm_mode_connector_attach_encoder(connector, encoder);
-	return 0;
-}
-
-/******************************************************************************
- * Audio
- *****************************************************************************/
-static void
-nvd0_audio_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode)
-{
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct nouveau_connector *nv_connector;
-	struct drm_device *dev = encoder->dev;
-	int i, or = nv_encoder->or * 0x30;
-
-	nv_connector = nouveau_encoder_connector_get(nv_encoder);
-	if (!drm_detect_monitor_audio(nv_connector->edid))
-		return;
-
-	nv_mask(dev, 0x10ec10 + or, 0x80000003, 0x80000001);
-
-	drm_edid_to_eld(&nv_connector->base, nv_connector->edid);
-	if (nv_connector->base.eld[0]) {
-		u8 *eld = nv_connector->base.eld;
-
-		for (i = 0; i < eld[2] * 4; i++)
-			nv_wr32(dev, 0x10ec00 + or, (i << 8) | eld[i]);
-		for (i = eld[2] * 4; i < 0x60; i++)
-			nv_wr32(dev, 0x10ec00 + or, (i << 8) | 0x00);
-
-		nv_mask(dev, 0x10ec10 + or, 0x80000002, 0x80000002);
-	}
-}
-
-static void
-nvd0_audio_disconnect(struct drm_encoder *encoder)
-{
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct drm_device *dev = encoder->dev;
-	int or = nv_encoder->or * 0x30;
-
-	nv_mask(dev, 0x10ec10 + or, 0x80000003, 0x80000000);
-}
-
-/******************************************************************************
- * HDMI
- *****************************************************************************/
-static void
-nvd0_hdmi_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode)
-{
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc);
-	struct nouveau_connector *nv_connector;
-	struct drm_device *dev = encoder->dev;
-	int head = nv_crtc->index * 0x800;
-	u32 rekey = 56; /* binary driver, and tegra constant */
-	u32 max_ac_packet;
-
-	nv_connector = nouveau_encoder_connector_get(nv_encoder);
-	if (!drm_detect_hdmi_monitor(nv_connector->edid))
-		return;
-
-	max_ac_packet  = mode->htotal - mode->hdisplay;
-	max_ac_packet -= rekey;
-	max_ac_packet -= 18; /* constant from tegra */
-	max_ac_packet /= 32;
-
-	/* AVI InfoFrame */
-	nv_mask(dev, 0x616714 + head, 0x00000001, 0x00000000);
-	nv_wr32(dev, 0x61671c + head, 0x000d0282);
-	nv_wr32(dev, 0x616720 + head, 0x0000006f);
-	nv_wr32(dev, 0x616724 + head, 0x00000000);
-	nv_wr32(dev, 0x616728 + head, 0x00000000);
-	nv_wr32(dev, 0x61672c + head, 0x00000000);
-	nv_mask(dev, 0x616714 + head, 0x00000001, 0x00000001);
-
-	/* ??? InfoFrame? */
-	nv_mask(dev, 0x6167a4 + head, 0x00000001, 0x00000000);
-	nv_wr32(dev, 0x6167ac + head, 0x00000010);
-	nv_mask(dev, 0x6167a4 + head, 0x00000001, 0x00000001);
-
-	/* HDMI_CTRL */
-	nv_mask(dev, 0x616798 + head, 0x401f007f, 0x40000000 | rekey |
-						  max_ac_packet << 16);
-
-	/* NFI, audio doesn't work without it though.. */
-	nv_mask(dev, 0x616548 + head, 0x00000070, 0x00000000);
-
-	nvd0_audio_mode_set(encoder, mode);
-}
-
-static void
-nvd0_hdmi_disconnect(struct drm_encoder *encoder)
-{
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(nv_encoder->crtc);
-	struct drm_device *dev = encoder->dev;
-	int head = nv_crtc->index * 0x800;
-
-	nvd0_audio_disconnect(encoder);
-
-	nv_mask(dev, 0x616798 + head, 0x40000000, 0x00000000);
-	nv_mask(dev, 0x6167a4 + head, 0x00000001, 0x00000000);
-	nv_mask(dev, 0x616714 + head, 0x00000001, 0x00000000);
-}
-
-/******************************************************************************
- * SOR
- *****************************************************************************/
-static inline u32
-nvd0_sor_dp_lane_map(struct drm_device *dev, struct dcb_entry *dcb, u8 lane)
-{
-	static const u8 nvd0[] = { 16, 8, 0, 24 };
-	return nvd0[lane];
-}
-
-static void
-nvd0_sor_dp_train_set(struct drm_device *dev, struct dcb_entry *dcb, u8 pattern)
-{
-	const u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1);
-	const u32 loff = (or * 0x800) + (link * 0x80);
-	nv_mask(dev, 0x61c110 + loff, 0x0f0f0f0f, 0x01010101 * pattern);
-}
-
-static void
-nvd0_sor_dp_train_adj(struct drm_device *dev, struct dcb_entry *dcb,
-		      u8 lane, u8 swing, u8 preem)
-{
-	const u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1);
-	const u32 loff = (or * 0x800) + (link * 0x80);
-	u32 shift = nvd0_sor_dp_lane_map(dev, dcb, lane);
-	u32 mask = 0x000000ff << shift;
-	u8 *table, *entry, *config = NULL;
-
-	switch (swing) {
-	case 0: preem += 0; break;
-	case 1: preem += 4; break;
-	case 2: preem += 7; break;
-	case 3: preem += 9; break;
-	}
-
-	table = nouveau_dp_bios_data(dev, dcb, &entry);
-	if (table) {
-		if (table[0] == 0x30) {
-			config  = entry + table[4];
-			config += table[5] * preem;
-		}
-	}
-
-	if (!config) {
-		NV_ERROR(dev, "PDISP: unsupported DP table for chipset\n");
-		return;
-	}
-
-	nv_mask(dev, 0x61c118 + loff, mask, config[1] << shift);
-	nv_mask(dev, 0x61c120 + loff, mask, config[2] << shift);
-	nv_mask(dev, 0x61c130 + loff, 0x0000ff00, config[3] << 8);
-	nv_mask(dev, 0x61c13c + loff, 0x00000000, 0x00000000);
-}
-
-static void
-nvd0_sor_dp_link_set(struct drm_device *dev, struct dcb_entry *dcb, int crtc,
-		     int link_nr, u32 link_bw, bool enhframe)
-{
-	const u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1);
-	const u32 loff = (or * 0x800) + (link * 0x80);
-	const u32 soff = (or * 0x800);
-	u32 dpctrl = nv_rd32(dev, 0x61c10c + loff) & ~0x001f4000;
-	u32 clksor = nv_rd32(dev, 0x612300 + soff) & ~0x007c0000;
-	u32 script = 0x0000, lane_mask = 0;
-	u8 *table, *entry;
-	int i;
-
-	link_bw /= 27000;
-
-	table = nouveau_dp_bios_data(dev, dcb, &entry);
-	if (table) {
-		if      (table[0] == 0x30) entry = ROMPTR(dev, entry[10]);
-		else                       entry = NULL;
-
-		while (entry) {
-			if (entry[0] >= link_bw)
-				break;
-			entry += 3;
-		}
-
-		nouveau_bios_run_init_table(dev, script, dcb, crtc);
-	}
-
-	clksor |= link_bw << 18;
-	dpctrl |= ((1 << link_nr) - 1) << 16;
-	if (enhframe)
-		dpctrl |= 0x00004000;
-
-	for (i = 0; i < link_nr; i++)
-		lane_mask |= 1 << (nvd0_sor_dp_lane_map(dev, dcb, i) >> 3);
-
-	nv_wr32(dev, 0x612300 + soff, clksor);
-	nv_wr32(dev, 0x61c10c + loff, dpctrl);
-	nv_mask(dev, 0x61c130 + loff, 0x0000000f, lane_mask);
-}
-
-static void
-nvd0_sor_dp_link_get(struct drm_device *dev, struct dcb_entry *dcb,
-		     u32 *link_nr, u32 *link_bw)
-{
-	const u32 or = ffs(dcb->or) - 1, link = !(dcb->sorconf.link & 1);
-	const u32 loff = (or * 0x800) + (link * 0x80);
-	const u32 soff = (or * 0x800);
-	u32 dpctrl = nv_rd32(dev, 0x61c10c + loff) & 0x000f0000;
-	u32 clksor = nv_rd32(dev, 0x612300 + soff);
-
-	if      (dpctrl > 0x00030000) *link_nr = 4;
-	else if (dpctrl > 0x00010000) *link_nr = 2;
-	else			      *link_nr = 1;
-
-	*link_bw  = (clksor & 0x007c0000) >> 18;
-	*link_bw *= 27000;
-}
-
-static void
-nvd0_sor_dp_calc_tu(struct drm_device *dev, struct dcb_entry *dcb,
-		    u32 crtc, u32 datarate)
-{
-	const u32 symbol = 100000;
-	const u32 TU = 64;
-	u32 link_nr, link_bw;
-	u64 ratio, value;
-
-	nvd0_sor_dp_link_get(dev, dcb, &link_nr, &link_bw);
-
-	ratio  = datarate;
-	ratio *= symbol;
-	do_div(ratio, link_nr * link_bw);
-
-	value  = (symbol - ratio) * TU;
-	value *= ratio;
-	do_div(value, symbol);
-	do_div(value, symbol);
-
-	value += 5;
-	value |= 0x08000000;
-
-	nv_wr32(dev, 0x616610 + (crtc * 0x800), value);
-}
-
-static void
-nvd0_sor_dpms(struct drm_encoder *encoder, int mode)
-{
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct drm_device *dev = encoder->dev;
-	struct drm_encoder *partner;
-	int or = nv_encoder->or;
-	u32 dpms_ctrl;
-
-	nv_encoder->last_dpms = mode;
-
-	list_for_each_entry(partner, &dev->mode_config.encoder_list, head) {
-		struct nouveau_encoder *nv_partner = nouveau_encoder(partner);
-
-		if (partner->encoder_type != DRM_MODE_ENCODER_TMDS)
-			continue;
-
-		if (nv_partner != nv_encoder &&
-		    nv_partner->dcb->or == nv_encoder->dcb->or) {
-			if (nv_partner->last_dpms == DRM_MODE_DPMS_ON)
-				return;
-			break;
-		}
-	}
-
-	dpms_ctrl  = (mode == DRM_MODE_DPMS_ON);
-	dpms_ctrl |= 0x80000000;
-
-	nv_wait(dev, 0x61c004 + (or * 0x0800), 0x80000000, 0x00000000);
-	nv_mask(dev, 0x61c004 + (or * 0x0800), 0x80000001, dpms_ctrl);
-	nv_wait(dev, 0x61c004 + (or * 0x0800), 0x80000000, 0x00000000);
-	nv_wait(dev, 0x61c030 + (or * 0x0800), 0x10000000, 0x00000000);
-
-	if (nv_encoder->dcb->type == OUTPUT_DP) {
-		struct dp_train_func func = {
-			.link_set = nvd0_sor_dp_link_set,
-			.train_set = nvd0_sor_dp_train_set,
-			.train_adj = nvd0_sor_dp_train_adj
-		};
-
-		nouveau_dp_dpms(encoder, mode, nv_encoder->dp.datarate, &func);
-	}
-}
-
-static bool
-nvd0_sor_mode_fixup(struct drm_encoder *encoder, struct drm_display_mode *mode,
-		    struct drm_display_mode *adjusted_mode)
-{
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct nouveau_connector *nv_connector;
-
-	nv_connector = nouveau_encoder_connector_get(nv_encoder);
-	if (nv_connector && nv_connector->native_mode) {
-		if (nv_connector->scaling_mode != DRM_MODE_SCALE_NONE) {
-			int id = adjusted_mode->base.id;
-			*adjusted_mode = *nv_connector->native_mode;
-			adjusted_mode->base.id = id;
-		}
-	}
-
-	return true;
-}
-
-static void
-nvd0_sor_disconnect(struct drm_encoder *encoder)
-{
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct drm_device *dev = encoder->dev;
-	u32 *push;
-
-	if (nv_encoder->crtc) {
-		nvd0_crtc_prepare(nv_encoder->crtc);
-
-		push = evo_wait(dev, EVO_MASTER, 4);
-		if (push) {
-			evo_mthd(push, 0x0200 + (nv_encoder->or * 0x20), 1);
-			evo_data(push, 0x00000000);
-			evo_mthd(push, 0x0080, 1);
-			evo_data(push, 0x00000000);
-			evo_kick(push, dev, EVO_MASTER);
-		}
-
-		nvd0_hdmi_disconnect(encoder);
-
-		nv_encoder->crtc = NULL;
-		nv_encoder->last_dpms = DRM_MODE_DPMS_OFF;
-	}
-}
-
-static void
-nvd0_sor_prepare(struct drm_encoder *encoder)
-{
-	nvd0_sor_disconnect(encoder);
-	if (nouveau_encoder(encoder)->dcb->type == OUTPUT_DP)
-		evo_sync(encoder->dev, EVO_MASTER);
-}
-
-static void
-nvd0_sor_commit(struct drm_encoder *encoder)
-{
-}
-
-static void
-nvd0_sor_mode_set(struct drm_encoder *encoder, struct drm_display_mode *umode,
-		  struct drm_display_mode *mode)
-{
-	struct drm_device *dev = encoder->dev;
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_encoder *nv_encoder = nouveau_encoder(encoder);
-	struct nouveau_crtc *nv_crtc = nouveau_crtc(encoder->crtc);
-	struct nouveau_connector *nv_connector;
-	struct nvbios *bios = &dev_priv->vbios;
-	u32 mode_ctrl = (1 << nv_crtc->index);
-	u32 syncs, magic, *push;
-	u32 or_config;
-
-	syncs = 0x00000001;
-	if (mode->flags & DRM_MODE_FLAG_NHSYNC)
-		syncs |= 0x00000008;
-	if (mode->flags & DRM_MODE_FLAG_NVSYNC)
-		syncs |= 0x00000010;
-
-	magic = 0x31ec6000 | (nv_crtc->index << 25);
-	if (mode->flags & DRM_MODE_FLAG_INTERLACE)
-		magic |= 0x00000001;
-
-	nv_connector = nouveau_encoder_connector_get(nv_encoder);
-	switch (nv_encoder->dcb->type) {
-	case OUTPUT_TMDS:
-		if (nv_encoder->dcb->sorconf.link & 1) {
-			if (mode->clock < 165000)
-				mode_ctrl |= 0x00000100;
-			else
-				mode_ctrl |= 0x00000500;
-		} else {
-			mode_ctrl |= 0x00000200;
-		}
-
-		or_config = (mode_ctrl & 0x00000f00) >> 8;
-		if (mode->clock >= 165000)
-			or_config |= 0x0100;
-
-		nvd0_hdmi_mode_set(encoder, mode);
-		break;
-	case OUTPUT_LVDS:
-		or_config = (mode_ctrl & 0x00000f00) >> 8;
-		if (bios->fp_no_ddc) {
-			if (bios->fp.dual_link)
-				or_config |= 0x0100;
-			if (bios->fp.if_is_24bit)
-				or_config |= 0x0200;
-		} else {
-			if (nv_connector->type == DCB_CONNECTOR_LVDS_SPWG) {
-				if (((u8 *)nv_connector->edid)[121] == 2)
-					or_config |= 0x0100;
-			} else
-			if (mode->clock >= bios->fp.duallink_transition_clk) {
-				or_config |= 0x0100;
-			}
-
-			if (or_config & 0x0100) {
-				if (bios->fp.strapless_is_24bit & 2)
-					or_config |= 0x0200;
-			} else {
-				if (bios->fp.strapless_is_24bit & 1)
-					or_config |= 0x0200;
-			}
-
-			if (nv_connector->base.display_info.bpc == 8)
-				or_config |= 0x0200;
-
-		}
-		break;
-	case OUTPUT_DP:
-		if (nv_connector->base.display_info.bpc == 6) {
-			nv_encoder->dp.datarate = mode->clock * 18 / 8;
-			syncs |= 0x00000140;
-		} else {
-			nv_encoder->dp.datarate = mode->clock * 24 / 8;
-			syncs |= 0x00000180;
-		}
-
-		if (nv_encoder->dcb->sorconf.link & 1)
-			mode_ctrl |= 0x00000800;
-		else
-			mode_ctrl |= 0x00000900;
-
-		or_config = (mode_ctrl & 0x00000f00) >> 8;
-		break;
-	default:
-		BUG_ON(1);
-		break;
-	}
-
-	nvd0_sor_dpms(encoder, DRM_MODE_DPMS_ON);
-
-	if (nv_encoder->dcb->type == OUTPUT_DP) {
-		nvd0_sor_dp_calc_tu(dev, nv_encoder->dcb, nv_crtc->index,
-					 nv_encoder->dp.datarate);
-	}
-
-	push = evo_wait(dev, EVO_MASTER, 8);
-	if (push) {
-		evo_mthd(push, 0x0404 + (nv_crtc->index * 0x300), 2);
-		evo_data(push, syncs);
-		evo_data(push, magic);
-		evo_mthd(push, 0x0200 + (nv_encoder->or * 0x020), 2);
-		evo_data(push, mode_ctrl);
-		evo_data(push, or_config);
-		evo_kick(push, dev, EVO_MASTER);
-	}
-
-	nv_encoder->crtc = encoder->crtc;
-}
-
-static void
-nvd0_sor_destroy(struct drm_encoder *encoder)
-{
-	drm_encoder_cleanup(encoder);
-	kfree(encoder);
-}
-
-static const struct drm_encoder_helper_funcs nvd0_sor_hfunc = {
-	.dpms = nvd0_sor_dpms,
-	.mode_fixup = nvd0_sor_mode_fixup,
-	.prepare = nvd0_sor_prepare,
-	.commit = nvd0_sor_commit,
-	.mode_set = nvd0_sor_mode_set,
-	.disable = nvd0_sor_disconnect,
-	.get_crtc = nvd0_display_crtc_get,
-};
-
-static const struct drm_encoder_funcs nvd0_sor_func = {
-	.destroy = nvd0_sor_destroy,
-};
-
-static int
-nvd0_sor_create(struct drm_connector *connector, struct dcb_entry *dcbe)
-{
-	struct drm_device *dev = connector->dev;
-	struct nouveau_encoder *nv_encoder;
-	struct drm_encoder *encoder;
-
-	nv_encoder = kzalloc(sizeof(*nv_encoder), GFP_KERNEL);
-	if (!nv_encoder)
-		return -ENOMEM;
-	nv_encoder->dcb = dcbe;
-	nv_encoder->or = ffs(dcbe->or) - 1;
-	nv_encoder->last_dpms = DRM_MODE_DPMS_OFF;
-
-	encoder = to_drm_encoder(nv_encoder);
-	encoder->possible_crtcs = dcbe->heads;
-	encoder->possible_clones = 0;
-	drm_encoder_init(dev, encoder, &nvd0_sor_func, DRM_MODE_ENCODER_TMDS);
-	drm_encoder_helper_add(encoder, &nvd0_sor_hfunc);
-
-	drm_mode_connector_attach_encoder(connector, encoder);
-	return 0;
-}
-
-/******************************************************************************
- * IRQ
- *****************************************************************************/
-static struct dcb_entry *
-lookup_dcb(struct drm_device *dev, int id, u32 mc)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	int type, or, i, link = -1;
-
-	if (id < 4) {
-		type = OUTPUT_ANALOG;
-		or   = id;
-	} else {
-		switch (mc & 0x00000f00) {
-		case 0x00000000: link = 0; type = OUTPUT_LVDS; break;
-		case 0x00000100: link = 0; type = OUTPUT_TMDS; break;
-		case 0x00000200: link = 1; type = OUTPUT_TMDS; break;
-		case 0x00000500: link = 0; type = OUTPUT_TMDS; break;
-		case 0x00000800: link = 0; type = OUTPUT_DP; break;
-		case 0x00000900: link = 1; type = OUTPUT_DP; break;
-		default:
-			NV_ERROR(dev, "PDISP: unknown SOR mc 0x%08x\n", mc);
-			return NULL;
-		}
-
-		or = id - 4;
-	}
-
-	for (i = 0; i < dev_priv->vbios.dcb.entries; i++) {
-		struct dcb_entry *dcb = &dev_priv->vbios.dcb.entry[i];
-		if (dcb->type == type && (dcb->or & (1 << or)) &&
-		    (link < 0 || link == !(dcb->sorconf.link & 1)))
-			return dcb;
-	}
-
-	NV_ERROR(dev, "PDISP: DCB for %d/0x%08x not found\n", id, mc);
-	return NULL;
-}
-
-static void
-nvd0_display_unk1_handler(struct drm_device *dev, u32 crtc, u32 mask)
-{
-	struct dcb_entry *dcb;
-	int i;
-
-	for (i = 0; mask && i < 8; i++) {
-		u32 mcc = nv_rd32(dev, 0x640180 + (i * 0x20));
-		if (!(mcc & (1 << crtc)))
-			continue;
-
-		dcb = lookup_dcb(dev, i, mcc);
-		if (!dcb)
-			continue;
-
-		nouveau_bios_run_display_table(dev, 0x0000, -1, dcb, crtc);
-	}
-
-	nv_wr32(dev, 0x6101d4, 0x00000000);
-	nv_wr32(dev, 0x6109d4, 0x00000000);
-	nv_wr32(dev, 0x6101d0, 0x80000000);
-}
-
-static void
-nvd0_display_unk2_handler(struct drm_device *dev, u32 crtc, u32 mask)
-{
-	struct dcb_entry *dcb;
-	u32 or, tmp, pclk;
-	int i;
-
-	for (i = 0; mask && i < 8; i++) {
-		u32 mcc = nv_rd32(dev, 0x640180 + (i * 0x20));
-		if (!(mcc & (1 << crtc)))
-			continue;
-
-		dcb = lookup_dcb(dev, i, mcc);
-		if (!dcb)
-			continue;
-
-		nouveau_bios_run_display_table(dev, 0x0000, -2, dcb, crtc);
-	}
-
-	pclk = nv_rd32(dev, 0x660450 + (crtc * 0x300)) / 1000;
-	if (mask & 0x00010000) {
-		nv50_crtc_set_clock(dev, crtc, pclk);
-	}
-
-	for (i = 0; mask && i < 8; i++) {
-		u32 mcp = nv_rd32(dev, 0x660180 + (i * 0x20));
-		u32 cfg = nv_rd32(dev, 0x660184 + (i * 0x20));
-		if (!(mcp & (1 << crtc)))
-			continue;
-
-		dcb = lookup_dcb(dev, i, mcp);
-		if (!dcb)
-			continue;
-		or = ffs(dcb->or) - 1;
-
-		nouveau_bios_run_display_table(dev, cfg, pclk, dcb, crtc);
-
-		nv_wr32(dev, 0x612200 + (crtc * 0x800), 0x00000000);
-		switch (dcb->type) {
-		case OUTPUT_ANALOG:
-			nv_wr32(dev, 0x612280 + (or * 0x800), 0x00000000);
-			break;
-		case OUTPUT_TMDS:
-		case OUTPUT_LVDS:
-		case OUTPUT_DP:
-			if (cfg & 0x00000100)
-				tmp = 0x00000101;
-			else
-				tmp = 0x00000000;
-
-			nv_mask(dev, 0x612300 + (or * 0x800), 0x00000707, tmp);
-			break;
-		default:
-			break;
-		}
-
-		break;
-	}
-
-	nv_wr32(dev, 0x6101d4, 0x00000000);
-	nv_wr32(dev, 0x6109d4, 0x00000000);
-	nv_wr32(dev, 0x6101d0, 0x80000000);
-}
-
-static void
-nvd0_display_unk4_handler(struct drm_device *dev, u32 crtc, u32 mask)
-{
-	struct dcb_entry *dcb;
-	int pclk, i;
-
-	pclk = nv_rd32(dev, 0x660450 + (crtc * 0x300)) / 1000;
-
-	for (i = 0; mask && i < 8; i++) {
-		u32 mcp = nv_rd32(dev, 0x660180 + (i * 0x20));
-		u32 cfg = nv_rd32(dev, 0x660184 + (i * 0x20));
-		if (!(mcp & (1 << crtc)))
-			continue;
-
-		dcb = lookup_dcb(dev, i, mcp);
-		if (!dcb)
-			continue;
-
-		nouveau_bios_run_display_table(dev, cfg, -pclk, dcb, crtc);
-	}
-
-	nv_wr32(dev, 0x6101d4, 0x00000000);
-	nv_wr32(dev, 0x6109d4, 0x00000000);
-	nv_wr32(dev, 0x6101d0, 0x80000000);
-}
-
-static void
-nvd0_display_bh(unsigned long data)
-{
-	struct drm_device *dev = (struct drm_device *)data;
-	struct nvd0_display *disp = nvd0_display(dev);
-	u32 mask = 0, crtc = ~0;
-	int i;
-
-	if (drm_debug & (DRM_UT_DRIVER | DRM_UT_KMS)) {
-		NV_INFO(dev, "PDISP: modeset req %d\n", disp->modeset);
-		NV_INFO(dev, " STAT: 0x%08x 0x%08x 0x%08x\n",
-			 nv_rd32(dev, 0x6101d0),
-			 nv_rd32(dev, 0x6101d4), nv_rd32(dev, 0x6109d4));
-		for (i = 0; i < 8; i++) {
-			NV_INFO(dev, " %s%d: 0x%08x 0x%08x\n",
-				i < 4 ? "DAC" : "SOR", i,
-				nv_rd32(dev, 0x640180 + (i * 0x20)),
-				nv_rd32(dev, 0x660180 + (i * 0x20)));
-		}
-	}
-
-	while (!mask && ++crtc < dev->mode_config.num_crtc)
-		mask = nv_rd32(dev, 0x6101d4 + (crtc * 0x800));
-
-	if (disp->modeset & 0x00000001)
-		nvd0_display_unk1_handler(dev, crtc, mask);
-	if (disp->modeset & 0x00000002)
-		nvd0_display_unk2_handler(dev, crtc, mask);
-	if (disp->modeset & 0x00000004)
-		nvd0_display_unk4_handler(dev, crtc, mask);
-}
-
-static void
-nvd0_display_intr(struct drm_device *dev)
-{
-	struct nvd0_display *disp = nvd0_display(dev);
-	u32 intr = nv_rd32(dev, 0x610088);
-	int i;
-
-	if (intr & 0x00000001) {
-		u32 stat = nv_rd32(dev, 0x61008c);
-		nv_wr32(dev, 0x61008c, stat);
-		intr &= ~0x00000001;
-	}
-
-	if (intr & 0x00000002) {
-		u32 stat = nv_rd32(dev, 0x61009c);
-		int chid = ffs(stat) - 1;
-		if (chid >= 0) {
-			u32 mthd = nv_rd32(dev, 0x6101f0 + (chid * 12));
-			u32 data = nv_rd32(dev, 0x6101f4 + (chid * 12));
-			u32 unkn = nv_rd32(dev, 0x6101f8 + (chid * 12));
-
-			NV_INFO(dev, "EvoCh: chid %d mthd 0x%04x data 0x%08x "
-				     "0x%08x 0x%08x\n",
-				chid, (mthd & 0x0000ffc), data, mthd, unkn);
-			nv_wr32(dev, 0x61009c, (1 << chid));
-			nv_wr32(dev, 0x6101f0 + (chid * 12), 0x90000000);
-		}
-
-		intr &= ~0x00000002;
-	}
-
-	if (intr & 0x00100000) {
-		u32 stat = nv_rd32(dev, 0x6100ac);
-
-		if (stat & 0x00000007) {
-			disp->modeset = stat;
-			tasklet_schedule(&disp->tasklet);
-
-			nv_wr32(dev, 0x6100ac, (stat & 0x00000007));
-			stat &= ~0x00000007;
-		}
-
-		if (stat) {
-			NV_INFO(dev, "PDISP: unknown intr24 0x%08x\n", stat);
-			nv_wr32(dev, 0x6100ac, stat);
-		}
-
-		intr &= ~0x00100000;
-	}
-
-	for (i = 0; i < dev->mode_config.num_crtc; i++) {
-		u32 mask = 0x01000000 << i;
-		if (intr & mask) {
-			u32 stat = nv_rd32(dev, 0x6100bc + (i * 0x800));
-			nv_wr32(dev, 0x6100bc + (i * 0x800), stat);
-			intr &= ~mask;
-		}
-	}
-
-	if (intr)
-		NV_INFO(dev, "PDISP: unknown intr 0x%08x\n", intr);
-}
-
-/******************************************************************************
- * Init
- *****************************************************************************/
-void
-nvd0_display_fini(struct drm_device *dev)
-{
-	int i;
-
-	/* fini cursors + overlays + flips */
-	for (i = 1; i >= 0; i--) {
-		evo_fini_pio(dev, EVO_CURS(i));
-		evo_fini_pio(dev, EVO_OIMM(i));
-		evo_fini_dma(dev, EVO_OVLY(i));
-		evo_fini_dma(dev, EVO_FLIP(i));
-	}
-
-	/* fini master */
-	evo_fini_dma(dev, EVO_MASTER);
-}
-
-int
-nvd0_display_init(struct drm_device *dev)
-{
-	struct nvd0_display *disp = nvd0_display(dev);
-	int ret, i;
-	u32 *push;
-
-	if (nv_rd32(dev, 0x6100ac) & 0x00000100) {
-		nv_wr32(dev, 0x6100ac, 0x00000100);
-		nv_mask(dev, 0x6194e8, 0x00000001, 0x00000000);
-		if (!nv_wait(dev, 0x6194e8, 0x00000002, 0x00000000)) {
-			NV_ERROR(dev, "PDISP: 0x6194e8 0x%08x\n",
-				 nv_rd32(dev, 0x6194e8));
-			return -EBUSY;
-		}
-	}
-
-	/* nfi what these are exactly, i do know that SOR_MODE_CTRL won't
-	 * work at all unless you do the SOR part below.
-	 */
-	for (i = 0; i < 3; i++) {
-		u32 dac = nv_rd32(dev, 0x61a000 + (i * 0x800));
-		nv_wr32(dev, 0x6101c0 + (i * 0x800), dac);
-	}
-
-	for (i = 0; i < 4; i++) {
-		u32 sor = nv_rd32(dev, 0x61c000 + (i * 0x800));
-		nv_wr32(dev, 0x6301c4 + (i * 0x800), sor);
-	}
-
-	for (i = 0; i < dev->mode_config.num_crtc; i++) {
-		u32 crtc0 = nv_rd32(dev, 0x616104 + (i * 0x800));
-		u32 crtc1 = nv_rd32(dev, 0x616108 + (i * 0x800));
-		u32 crtc2 = nv_rd32(dev, 0x61610c + (i * 0x800));
-		nv_wr32(dev, 0x6101b4 + (i * 0x800), crtc0);
-		nv_wr32(dev, 0x6101b8 + (i * 0x800), crtc1);
-		nv_wr32(dev, 0x6101bc + (i * 0x800), crtc2);
-	}
-
-	/* point at our hash table / objects, enable interrupts */
-	nv_wr32(dev, 0x610010, (disp->mem->vinst >> 8) | 9);
-	nv_mask(dev, 0x6100b0, 0x00000307, 0x00000307);
-
-	/* init master */
-	ret = evo_init_dma(dev, EVO_MASTER);
-	if (ret)
-		goto error;
-
-	/* init flips + overlays + cursors */
-	for (i = 0; i < dev->mode_config.num_crtc; i++) {
-		if ((ret = evo_init_dma(dev, EVO_FLIP(i))) ||
-		    (ret = evo_init_dma(dev, EVO_OVLY(i))) ||
-		    (ret = evo_init_pio(dev, EVO_OIMM(i))) ||
-		    (ret = evo_init_pio(dev, EVO_CURS(i))))
-			goto error;
-	}
-
-	push = evo_wait(dev, EVO_MASTER, 32);
-	if (!push) {
-		ret = -EBUSY;
-		goto error;
-	}
-	evo_mthd(push, 0x0088, 1);
-	evo_data(push, NvEvoSync);
-	evo_mthd(push, 0x0084, 1);
-	evo_data(push, 0x00000000);
-	evo_mthd(push, 0x0084, 1);
-	evo_data(push, 0x80000000);
-	evo_mthd(push, 0x008c, 1);
-	evo_data(push, 0x00000000);
-	evo_kick(push, dev, EVO_MASTER);
-
-error:
-	if (ret)
-		nvd0_display_fini(dev);
-	return ret;
-}
-
-void
-nvd0_display_destroy(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nvd0_display *disp = nvd0_display(dev);
-	struct pci_dev *pdev = dev->pdev;
-	int i;
-
-	for (i = 0; i < EVO_DMA_NR; i++) {
-		struct evo *evo = &disp->evo[i];
-		pci_free_consistent(pdev, PAGE_SIZE, evo->ptr, evo->handle);
-	}
-
-	nouveau_gpuobj_ref(NULL, &disp->mem);
-	nouveau_bo_unmap(disp->sync);
-	nouveau_bo_ref(NULL, &disp->sync);
-	nouveau_irq_unregister(dev, 26);
-
-	dev_priv->engine.display.priv = NULL;
-	kfree(disp);
-}
-
-int
-nvd0_display_create(struct drm_device *dev)
-{
-	struct drm_nouveau_private *dev_priv = dev->dev_private;
-	struct nouveau_instmem_engine *pinstmem = &dev_priv->engine.instmem;
-	struct dcb_table *dcb = &dev_priv->vbios.dcb;
-	struct drm_connector *connector, *tmp;
-	struct pci_dev *pdev = dev->pdev;
-	struct nvd0_display *disp;
-	struct dcb_entry *dcbe;
-	int crtcs, ret, i;
-
-	disp = kzalloc(sizeof(*disp), GFP_KERNEL);
-	if (!disp)
-		return -ENOMEM;
-	dev_priv->engine.display.priv = disp;
-
-	/* create crtc objects to represent the hw heads */
-	crtcs = nv_rd32(dev, 0x022448);
-	for (i = 0; i < crtcs; i++) {
-		ret = nvd0_crtc_create(dev, i);
-		if (ret)
-			goto out;
-	}
-
-	/* create encoder/connector objects based on VBIOS DCB table */
-	for (i = 0, dcbe = &dcb->entry[0]; i < dcb->entries; i++, dcbe++) {
-		connector = nouveau_connector_create(dev, dcbe->connector);
-		if (IS_ERR(connector))
-			continue;
-
-		if (dcbe->location != DCB_LOC_ON_CHIP) {
-			NV_WARN(dev, "skipping off-chip encoder %d/%d\n",
-				dcbe->type, ffs(dcbe->or) - 1);
-			continue;
-		}
-
-		switch (dcbe->type) {
-		case OUTPUT_TMDS:
-		case OUTPUT_LVDS:
-		case OUTPUT_DP:
-			nvd0_sor_create(connector, dcbe);
-			break;
-		case OUTPUT_ANALOG:
-			nvd0_dac_create(connector, dcbe);
-			break;
-		default:
-			NV_WARN(dev, "skipping unsupported encoder %d/%d\n",
-				dcbe->type, ffs(dcbe->or) - 1);
-			continue;
-		}
-	}
-
-	/* cull any connectors we created that don't have an encoder */
-	list_for_each_entry_safe(connector, tmp, &dev->mode_config.connector_list, head) {
-		if (connector->encoder_ids[0])
-			continue;
-
-		NV_WARN(dev, "%s has no encoders, removing\n",
-			drm_get_connector_name(connector));
-		connector->funcs->destroy(connector);
-	}
-
-	/* setup interrupt handling */
-	tasklet_init(&disp->tasklet, nvd0_display_bh, (unsigned long)dev);
-	nouveau_irq_register(dev, 26, nvd0_display_intr);
-
-	/* small shared memory area we use for notifiers and semaphores */
-	ret = nouveau_bo_new(dev, 4096, 0x1000, TTM_PL_FLAG_VRAM,
-			     0, 0x0000, &disp->sync);
-	if (!ret) {
-		ret = nouveau_bo_pin(disp->sync, TTM_PL_FLAG_VRAM);
-		if (!ret)
-			ret = nouveau_bo_map(disp->sync);
-		if (ret)
-			nouveau_bo_ref(NULL, &disp->sync);
-	}
-
-	if (ret)
-		goto out;
-
-	/* hash table and dma objects for the memory areas we care about */
-	ret = nouveau_gpuobj_new(dev, NULL, 0x4000, 0x10000,
-				 NVOBJ_FLAG_ZERO_ALLOC, &disp->mem);
-	if (ret)
-		goto out;
-
-	/* create evo dma channels */
-	for (i = 0; i < EVO_DMA_NR; i++) {
-		struct evo *evo = &disp->evo[i];
-		u64 offset = disp->sync->bo.offset;
-		u32 dmao = 0x1000 + (i * 0x100);
-		u32 hash = 0x0000 + (i * 0x040);
-
-		evo->idx = i;
-		evo->sem.offset = EVO_SYNC(evo->idx, 0x00);
-		evo->ptr = pci_alloc_consistent(pdev, PAGE_SIZE, &evo->handle);
-		if (!evo->ptr) {
-			ret = -ENOMEM;
-			goto out;
-		}
-
-		nv_wo32(disp->mem, dmao + 0x00, 0x00000049);
-		nv_wo32(disp->mem, dmao + 0x04, (offset + 0x0000) >> 8);
-		nv_wo32(disp->mem, dmao + 0x08, (offset + 0x0fff) >> 8);
-		nv_wo32(disp->mem, dmao + 0x0c, 0x00000000);
-		nv_wo32(disp->mem, dmao + 0x10, 0x00000000);
-		nv_wo32(disp->mem, dmao + 0x14, 0x00000000);
-		nv_wo32(disp->mem, hash + 0x00, NvEvoSync);
-		nv_wo32(disp->mem, hash + 0x04, 0x00000001 | (i << 27) |
-						((dmao + 0x00) << 9));
-
-		nv_wo32(disp->mem, dmao + 0x20, 0x00000049);
-		nv_wo32(disp->mem, dmao + 0x24, 0x00000000);
-		nv_wo32(disp->mem, dmao + 0x28, (dev_priv->vram_size - 1) >> 8);
-		nv_wo32(disp->mem, dmao + 0x2c, 0x00000000);
-		nv_wo32(disp->mem, dmao + 0x30, 0x00000000);
-		nv_wo32(disp->mem, dmao + 0x34, 0x00000000);
-		nv_wo32(disp->mem, hash + 0x08, NvEvoVRAM);
-		nv_wo32(disp->mem, hash + 0x0c, 0x00000001 | (i << 27) |
-						((dmao + 0x20) << 9));
-
-		nv_wo32(disp->mem, dmao + 0x40, 0x00000009);
-		nv_wo32(disp->mem, dmao + 0x44, 0x00000000);
-		nv_wo32(disp->mem, dmao + 0x48, (dev_priv->vram_size - 1) >> 8);
-		nv_wo32(disp->mem, dmao + 0x4c, 0x00000000);
-		nv_wo32(disp->mem, dmao + 0x50, 0x00000000);
-		nv_wo32(disp->mem, dmao + 0x54, 0x00000000);
-		nv_wo32(disp->mem, hash + 0x10, NvEvoVRAM_LP);
-		nv_wo32(disp->mem, hash + 0x14, 0x00000001 | (i << 27) |
-						((dmao + 0x40) << 9));
-
-		nv_wo32(disp->mem, dmao + 0x60, 0x0fe00009);
-		nv_wo32(disp->mem, dmao + 0x64, 0x00000000);
-		nv_wo32(disp->mem, dmao + 0x68, (dev_priv->vram_size - 1) >> 8);
-		nv_wo32(disp->mem, dmao + 0x6c, 0x00000000);
-		nv_wo32(disp->mem, dmao + 0x70, 0x00000000);
-		nv_wo32(disp->mem, dmao + 0x74, 0x00000000);
-		nv_wo32(disp->mem, hash + 0x18, NvEvoFB32);
-		nv_wo32(disp->mem, hash + 0x1c, 0x00000001 | (i << 27) |
-						((dmao + 0x60) << 9));
-	}
-
-	pinstmem->flush(dev);
-
-out:
-	if (ret)
-		nvd0_display_destroy(dev);
-	return ret;
-}
diff --git a/drivers/gpu/drm/r128/r128_cce.c b/drivers/gpu/drm/r128/r128_cce.c
index bcac90b..570e190 100644
--- a/drivers/gpu/drm/r128/r128_cce.c
+++ b/drivers/gpu/drm/r128/r128_cce.c
@@ -32,7 +32,6 @@
 #include <linux/firmware.h>
 #include <linux/platform_device.h>
 #include <linux/slab.h>
-#include <linux/module.h>
 
 #include "drmP.h"
 #include "drm.h"
diff --git a/drivers/gpu/drm/r128/r128_drv.c b/drivers/gpu/drm/r128/r128_drv.c
index 88718fa..b9e8efd 100644
--- a/drivers/gpu/drm/r128/r128_drv.c
+++ b/drivers/gpu/drm/r128/r128_drv.c
@@ -29,8 +29,6 @@
  *    Gareth Hughes <gareth@valinux.com>
  */
 
-#include <linux/module.h>
-
 #include "drmP.h"
 #include "drm.h"
 #include "r128_drm.h"
@@ -42,20 +40,6 @@ static struct pci_device_id pciidlist[] = {
 	r128_PCI_IDS
 };
 
-static const struct file_operations r128_driver_fops = {
-	.owner = THIS_MODULE,
-	.open = drm_open,
-	.release = drm_release,
-	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_mmap,
-	.poll = drm_poll,
-	.fasync = drm_fasync,
-#ifdef CONFIG_COMPAT
-	.compat_ioctl = r128_compat_ioctl,
-#endif
-	.llseek = noop_llseek,
-};
-
 static struct drm_driver driver = {
 	.driver_features =
 	    DRIVER_USE_AGP | DRIVER_USE_MTRR | DRIVER_PCI_DMA | DRIVER_SG |
@@ -74,7 +58,21 @@ static struct drm_driver driver = {
 	.reclaim_buffers = drm_core_reclaim_buffers,
 	.ioctls = r128_ioctls,
 	.dma_ioctl = r128_cce_buffers,
-	.fops = &r128_driver_fops,
+	.fops = {
+		.owner = THIS_MODULE,
+		.open = drm_open,
+		.release = drm_release,
+		.unlocked_ioctl = drm_ioctl,
+		.mmap = drm_mmap,
+		.poll = drm_poll,
+		.fasync = drm_fasync,
+#ifdef CONFIG_COMPAT
+		.compat_ioctl = r128_compat_ioctl,
+#endif
+		.llseek = noop_llseek,
+	},
+
+
 	.name = DRIVER_NAME,
 	.desc = DRIVER_DESC,
 	.date = DRIVER_DATE,
@@ -85,7 +83,6 @@ static struct drm_driver driver = {
 
 int r128_driver_load(struct drm_device *dev, unsigned long flags)
 {
-	pci_set_master(dev->pdev);
 	return drm_vblank_init(dev, 1);
 }
 
diff --git a/drivers/gpu/drm/radeon/.gitignore b/drivers/gpu/drm/radeon/.gitignore
new file mode 100644
index 0000000..403eb3a
--- /dev/null
+++ b/drivers/gpu/drm/radeon/.gitignore
@@ -0,0 +1,3 @@
+mkregtable
+*_reg_safe.h
+
diff --git a/drivers/gpu/drm/radeon/Makefile b/drivers/gpu/drm/radeon/Makefile
index 8410415..3896ef8 100644
--- a/drivers/gpu/drm/radeon/Makefile
+++ b/drivers/gpu/drm/radeon/Makefile
@@ -5,7 +5,6 @@
 ccflags-y := -Iinclude/drm
 
 hostprogs-y := mkregtable
-clean-files := rn50_reg_safe.h r100_reg_safe.h r200_reg_safe.h rv515_reg_safe.h r300_reg_safe.h r420_reg_safe.h rs600_reg_safe.h r600_reg_safe.h evergreen_reg_safe.h cayman_reg_safe.h
 
 quiet_cmd_mkregtable = MKREGTABLE $@
       cmd_mkregtable = $(obj)/mkregtable $< > $@
@@ -70,8 +69,7 @@ radeon-y += radeon_device.o radeon_asic.o radeon_kms.o \
 	r200.o radeon_legacy_tv.o r600_cs.o r600_blit.o r600_blit_shaders.o \
 	r600_blit_kms.o radeon_pm.o atombios_dp.o r600_audio.o r600_hdmi.o \
 	evergreen.o evergreen_cs.o evergreen_blit_shaders.o evergreen_blit_kms.o \
-	radeon_trace_points.o ni.o cayman_blit_shaders.o atombios_encoders.o \
-	radeon_semaphore.o radeon_sa.o atombios_i2c.o
+	radeon_trace_points.o ni.o cayman_blit_shaders.o
 
 radeon-$(CONFIG_COMPAT) += radeon_ioc32.o
 radeon-$(CONFIG_VGA_SWITCHEROO) += radeon_atpx_handler.o
@@ -79,4 +77,4 @@ radeon-$(CONFIG_ACPI) += radeon_acpi.o
 
 obj-$(CONFIG_DRM_RADEON)+= radeon.o
 
-CFLAGS_radeon_trace_points.o := -I$(src)
+CFLAGS_radeon_trace_points.o := -I$(src)
\ No newline at end of file
diff --git a/drivers/gpu/drm/radeon/atom.c b/drivers/gpu/drm/radeon/atom.c
index daa1e34..a1df7d7 100644
--- a/drivers/gpu/drm/radeon/atom.c
+++ b/drivers/gpu/drm/radeon/atom.c
@@ -665,8 +665,6 @@ static void atom_op_delay(atom_exec_context *ctx, int *ptr, int arg)
 	SDEBUG("   count: %d\n", count);
 	if (arg == ATOM_UNIT_MICROSEC)
 		udelay(count);
-	else if (!drm_can_sleep())
-		mdelay(count);
 	else
 		msleep(count);
 }
@@ -1261,9 +1259,6 @@ struct atom_context *atom_parse(struct card_info *card, void *bios)
 	char name[512];
 	int i;
 
-	if (!ctx)
-		return NULL;
-
 	ctx->card = card;
 	ctx->bios = bios;
 
diff --git a/drivers/gpu/drm/radeon/atombios_crtc.c b/drivers/gpu/drm/radeon/atombios_crtc.c
index 72672ea..071ded1 100644
--- a/drivers/gpu/drm/radeon/atombios_crtc.c
+++ b/drivers/gpu/drm/radeon/atombios_crtc.c
@@ -355,12 +355,15 @@ static void atombios_crtc_set_timing(struct drm_crtc *crtc,
 	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
 }
 
-static void atombios_disable_ss(struct radeon_device *rdev, int pll_id)
+static void atombios_disable_ss(struct drm_crtc *crtc)
 {
+	struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc);
+	struct drm_device *dev = crtc->dev;
+	struct radeon_device *rdev = dev->dev_private;
 	u32 ss_cntl;
 
 	if (ASIC_IS_DCE4(rdev)) {
-		switch (pll_id) {
+		switch (radeon_crtc->pll_id) {
 		case ATOM_PPLL1:
 			ss_cntl = RREG32(EVERGREEN_P1PLL_SS_CNTL);
 			ss_cntl &= ~EVERGREEN_PxPLL_SS_EN;
@@ -376,7 +379,7 @@ static void atombios_disable_ss(struct radeon_device *rdev, int pll_id)
 			return;
 		}
 	} else if (ASIC_IS_AVIVO(rdev)) {
-		switch (pll_id) {
+		switch (radeon_crtc->pll_id) {
 		case ATOM_PPLL1:
 			ss_cntl = RREG32(AVIVO_P1PLL_INT_SS_CNTL);
 			ss_cntl &= ~1;
@@ -403,11 +406,13 @@ union atom_enable_ss {
 	ENABLE_SPREAD_SPECTRUM_ON_PPLL_V3 v3;
 };
 
-static void atombios_crtc_program_ss(struct radeon_device *rdev,
+static void atombios_crtc_program_ss(struct drm_crtc *crtc,
 				     int enable,
 				     int pll_id,
 				     struct radeon_atom_ss *ss)
 {
+	struct drm_device *dev = crtc->dev;
+	struct radeon_device *rdev = dev->dev_private;
 	int index = GetIndexIntoMasterTable(COMMAND, EnableSpreadSpectrumOnPPLL);
 	union atom_enable_ss args;
 
@@ -461,7 +466,7 @@ static void atombios_crtc_program_ss(struct radeon_device *rdev,
 			return;
 		}
 		args.v2.ucEnable = enable;
-		if ((ss->percentage == 0) || (ss->type & ATOM_EXTERNAL_SS_MASK) || ASIC_IS_DCE41(rdev))
+		if ((ss->percentage == 0) || (ss->type & ATOM_EXTERNAL_SS_MASK))
 			args.v2.ucEnable = ATOM_DISABLE;
 	} else if (ASIC_IS_DCE3(rdev)) {
 		args.v1.usSpreadSpectrumPercentage = cpu_to_le16(ss->percentage);
@@ -474,7 +479,7 @@ static void atombios_crtc_program_ss(struct radeon_device *rdev,
 	} else if (ASIC_IS_AVIVO(rdev)) {
 		if ((enable == ATOM_DISABLE) || (ss->percentage == 0) ||
 		    (ss->type & ATOM_EXTERNAL_SS_MASK)) {
-			atombios_disable_ss(rdev, pll_id);
+			atombios_disable_ss(crtc);
 			return;
 		}
 		args.lvds_ss_2.usSpreadSpectrumPercentage = cpu_to_le16(ss->percentage);
@@ -486,7 +491,7 @@ static void atombios_crtc_program_ss(struct radeon_device *rdev,
 	} else {
 		if ((enable == ATOM_DISABLE) || (ss->percentage == 0) ||
 		    (ss->type & ATOM_EXTERNAL_SS_MASK)) {
-			atombios_disable_ss(rdev, pll_id);
+			atombios_disable_ss(crtc);
 			return;
 		}
 		args.lvds_ss.usSpreadSpectrumPercentage = cpu_to_le16(ss->percentage);
@@ -518,7 +523,6 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc,
 	int encoder_mode = 0;
 	u32 dp_clock = mode->clock;
 	int bpc = 8;
-	bool is_duallink = false;
 
 	/* reset the pll flags */
 	pll->flags = 0;
@@ -550,12 +554,11 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc,
 		if (encoder->crtc == crtc) {
 			radeon_encoder = to_radeon_encoder(encoder);
 			connector = radeon_get_connector_for_encoder(encoder);
-			if (connector && connector->display_info.bpc)
+			if (connector)
 				bpc = connector->display_info.bpc;
 			encoder_mode = atombios_get_encoder_mode(encoder);
-			is_duallink = radeon_dig_monitor_is_duallink(encoder, mode->clock);
 			if ((radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT | ATOM_DEVICE_DFP_SUPPORT)) ||
-			    (radeon_encoder_get_dp_bridge_encoder_id(encoder) != ENCODER_OBJECT_ID_NONE)) {
+			    radeon_encoder_is_dp_bridge(encoder)) {
 				if (connector) {
 					struct radeon_connector *radeon_connector = to_radeon_connector(connector);
 					struct radeon_connector_atom_dig *dig_connector =
@@ -635,29 +638,44 @@ static u32 atombios_adjust_pll(struct drm_crtc *crtc,
 				if (ss_enabled && ss->percentage)
 					args.v3.sInput.ucDispPllConfig |=
 						DISPPLL_CONFIG_SS_ENABLE;
-				if (ENCODER_MODE_IS_DP(encoder_mode)) {
-					args.v3.sInput.ucDispPllConfig |=
-						DISPPLL_CONFIG_COHERENT_MODE;
-					/* 16200 or 27000 */
-					args.v3.sInput.usPixelClock = cpu_to_le16(dp_clock / 10);
-				} else if (radeon_encoder->devices & (ATOM_DEVICE_DFP_SUPPORT)) {
+				if (radeon_encoder->devices & (ATOM_DEVICE_DFP_SUPPORT) ||
+				    radeon_encoder_is_dp_bridge(encoder)) {
 					struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
-					if (encoder_mode == ATOM_ENCODER_MODE_HDMI)
-						/* deep color support */
-						args.v3.sInput.usPixelClock =
-							cpu_to_le16((mode->clock * bpc / 8) / 10);
-					if (dig->coherent_mode)
+					if (encoder_mode == ATOM_ENCODER_MODE_DP) {
 						args.v3.sInput.ucDispPllConfig |=
 							DISPPLL_CONFIG_COHERENT_MODE;
-					if (is_duallink)
+						/* 16200 or 27000 */
+						args.v3.sInput.usPixelClock = cpu_to_le16(dp_clock / 10);
+					} else {
+						if (encoder_mode == ATOM_ENCODER_MODE_HDMI) {
+							/* deep color support */
+							args.v3.sInput.usPixelClock =
+								cpu_to_le16((mode->clock * bpc / 8) / 10);
+						}
+						if (dig->coherent_mode)
+							args.v3.sInput.ucDispPllConfig |=
+								DISPPLL_CONFIG_COHERENT_MODE;
+						if (mode->clock > 165000)
+							args.v3.sInput.ucDispPllConfig |=
+								DISPPLL_CONFIG_DUAL_LINK;
+					}
+				} else if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) {
+					if (encoder_mode == ATOM_ENCODER_MODE_DP) {
 						args.v3.sInput.ucDispPllConfig |=
-							DISPPLL_CONFIG_DUAL_LINK;
+							DISPPLL_CONFIG_COHERENT_MODE;
+						/* 16200 or 27000 */
+						args.v3.sInput.usPixelClock = cpu_to_le16(dp_clock / 10);
+					} else if (encoder_mode != ATOM_ENCODER_MODE_LVDS) {
+						if (mode->clock > 165000)
+							args.v3.sInput.ucDispPllConfig |=
+								DISPPLL_CONFIG_DUAL_LINK;
+					}
 				}
-				if (radeon_encoder_get_dp_bridge_encoder_id(encoder) !=
-				    ENCODER_OBJECT_ID_NONE)
-					args.v3.sInput.ucExtTransmitterID =
-						radeon_encoder_get_dp_bridge_encoder_id(encoder);
-				else
+				if (radeon_encoder_is_dp_bridge(encoder)) {
+					struct drm_encoder *ext_encoder = radeon_atom_get_external_encoder(encoder);
+					struct radeon_encoder *ext_radeon_encoder = to_radeon_encoder(ext_encoder);
+					args.v3.sInput.ucExtTransmitterID = ext_radeon_encoder->encoder_id;
+				} else
 					args.v3.sInput.ucExtTransmitterID = 0;
 
 				atom_execute_table(rdev->mode_info.atom_context,
@@ -699,9 +717,11 @@ union set_pixel_clock {
 /* on DCE5, make sure the voltage is high enough to support the
  * required disp clk.
  */
-static void atombios_crtc_set_dcpll(struct radeon_device *rdev,
+static void atombios_crtc_set_dcpll(struct drm_crtc *crtc,
 				    u32 dispclk)
 {
+	struct drm_device *dev = crtc->dev;
+	struct radeon_device *rdev = dev->dev_private;
 	u8 frev, crev;
 	int index;
 	union set_pixel_clock args;
@@ -744,7 +764,7 @@ static void atombios_crtc_set_dcpll(struct radeon_device *rdev,
 }
 
 static void atombios_crtc_program_pll(struct drm_crtc *crtc,
-				      u32 crtc_id,
+				      int crtc_id,
 				      int pll_id,
 				      u32 encoder_mode,
 				      u32 encoder_id,
@@ -831,7 +851,8 @@ static void atombios_crtc_program_pll(struct drm_crtc *crtc,
 			args.v5.ucPpll = pll_id;
 			break;
 		case 6:
-			args.v6.ulDispEngClkFreq = cpu_to_le32(crtc_id << 24 | clock / 10);
+			args.v6.ulCrtcPclkFreq.ucCRTC = crtc_id;
+			args.v6.ulCrtcPclkFreq.ulPixelClock = cpu_to_le32(clock / 10);
 			args.v6.ucRefDiv = ref_div;
 			args.v6.usFbDiv = cpu_to_le16(fb_div);
 			args.v6.ulFbDivDecFrac = cpu_to_le32(frac_fb_div * 100000);
@@ -925,7 +946,6 @@ static void atombios_crtc_set_pll(struct drm_crtc *crtc, struct drm_display_mode
 		bpc = connector->display_info.bpc;
 
 		switch (encoder_mode) {
-		case ATOM_ENCODER_MODE_DP_MST:
 		case ATOM_ENCODER_MODE_DP:
 			/* DP/eDP */
 			dp_clock = dig_connector->dp_clock / 10;
@@ -991,7 +1011,7 @@ static void atombios_crtc_set_pll(struct drm_crtc *crtc, struct drm_display_mode
 		radeon_compute_pll_legacy(pll, adjusted_clock, &pll_clock, &fb_div, &frac_fb_div,
 					  &ref_div, &post_div);
 
-	atombios_crtc_program_ss(rdev, ATOM_DISABLE, radeon_crtc->pll_id, &ss);
+	atombios_crtc_program_ss(crtc, ATOM_DISABLE, radeon_crtc->pll_id, &ss);
 
 	atombios_crtc_program_pll(crtc, radeon_crtc->crtc_id, radeon_crtc->pll_id,
 				  encoder_mode, radeon_encoder->encoder_id, mode->clock,
@@ -1014,7 +1034,7 @@ static void atombios_crtc_set_pll(struct drm_crtc *crtc, struct drm_display_mode
 			ss.step = step_size;
 		}
 
-		atombios_crtc_program_ss(rdev, ATOM_ENABLE, radeon_crtc->pll_id, &ss);
+		atombios_crtc_program_ss(crtc, ATOM_ENABLE, radeon_crtc->pll_id, &ss);
 	}
 }
 
@@ -1031,7 +1051,6 @@ static int dce4_crtc_do_set_base(struct drm_crtc *crtc,
 	struct radeon_bo *rbo;
 	uint64_t fb_location;
 	uint32_t fb_format, fb_pitch_pixels, tiling_flags;
-	unsigned bankw, bankh, mtaspect, tile_split;
 	u32 fb_swap = EVERGREEN_GRPH_ENDIAN_SWAP(EVERGREEN_GRPH_ENDIAN_NONE);
 	u32 tmp, viewport_w, viewport_h;
 	int r;
@@ -1103,33 +1122,9 @@ static int dce4_crtc_do_set_base(struct drm_crtc *crtc,
 		return -EINVAL;
 	}
 
-	if (tiling_flags & RADEON_TILING_MACRO) {
-		if (rdev->family >= CHIP_CAYMAN)
-			tmp = rdev->config.cayman.tile_config;
-		else
-			tmp = rdev->config.evergreen.tile_config;
-
-		switch ((tmp & 0xf0) >> 4) {
-		case 0: /* 4 banks */
-			fb_format |= EVERGREEN_GRPH_NUM_BANKS(EVERGREEN_ADDR_SURF_4_BANK);
-			break;
-		case 1: /* 8 banks */
-		default:
-			fb_format |= EVERGREEN_GRPH_NUM_BANKS(EVERGREEN_ADDR_SURF_8_BANK);
-			break;
-		case 2: /* 16 banks */
-			fb_format |= EVERGREEN_GRPH_NUM_BANKS(EVERGREEN_ADDR_SURF_16_BANK);
-			break;
-		}
-
+	if (tiling_flags & RADEON_TILING_MACRO)
 		fb_format |= EVERGREEN_GRPH_ARRAY_MODE(EVERGREEN_GRPH_ARRAY_2D_TILED_THIN1);
-
-		evergreen_tiling_fields(tiling_flags, &bankw, &bankh, &mtaspect, &tile_split);
-		fb_format |= EVERGREEN_GRPH_TILE_SPLIT(tile_split);
-		fb_format |= EVERGREEN_GRPH_BANK_WIDTH(bankw);
-		fb_format |= EVERGREEN_GRPH_BANK_HEIGHT(bankh);
-		fb_format |= EVERGREEN_GRPH_MACRO_TILE_ASPECT(mtaspect);
-	} else if (tiling_flags & RADEON_TILING_MICRO)
+	else if (tiling_flags & RADEON_TILING_MICRO)
 		fb_format |= EVERGREEN_GRPH_ARRAY_MODE(EVERGREEN_GRPH_ARRAY_1D_TILED_THIN1);
 
 	switch (radeon_crtc->crtc_id) {
@@ -1173,7 +1168,7 @@ static int dce4_crtc_do_set_base(struct drm_crtc *crtc,
 	WREG32(EVERGREEN_GRPH_X_END + radeon_crtc->crtc_offset, target_fb->width);
 	WREG32(EVERGREEN_GRPH_Y_END + radeon_crtc->crtc_offset, target_fb->height);
 
-	fb_pitch_pixels = target_fb->pitches[0] / (target_fb->bits_per_pixel / 8);
+	fb_pitch_pixels = target_fb->pitch / (target_fb->bits_per_pixel / 8);
 	WREG32(EVERGREEN_GRPH_PITCH + radeon_crtc->crtc_offset, fb_pitch_pixels);
 	WREG32(EVERGREEN_GRPH_ENABLE + radeon_crtc->crtc_offset, 1);
 
@@ -1342,7 +1337,7 @@ static int avivo_crtc_do_set_base(struct drm_crtc *crtc,
 	WREG32(AVIVO_D1GRPH_X_END + radeon_crtc->crtc_offset, target_fb->width);
 	WREG32(AVIVO_D1GRPH_Y_END + radeon_crtc->crtc_offset, target_fb->height);
 
-	fb_pitch_pixels = target_fb->pitches[0] / (target_fb->bits_per_pixel / 8);
+	fb_pitch_pixels = target_fb->pitch / (target_fb->bits_per_pixel / 8);
 	WREG32(AVIVO_D1GRPH_PITCH + radeon_crtc->crtc_offset, fb_pitch_pixels);
 	WREG32(AVIVO_D1GRPH_ENABLE + radeon_crtc->crtc_offset, 1);
 
@@ -1456,7 +1451,7 @@ static int radeon_atom_pick_pll(struct drm_crtc *crtc)
 				 * PPLL/DCPLL programming and only program the DP DTO for the
 				 * crtc virtual pixel clock.
 				 */
-				if (ENCODER_MODE_IS_DP(atombios_get_encoder_mode(test_encoder))) {
+				if (atombios_get_encoder_mode(test_encoder) == ATOM_ENCODER_MODE_DP) {
 					if (ASIC_IS_DCE5(rdev) || rdev->clock.dp_extclk)
 						return ATOM_PPLL_INVALID;
 				}
@@ -1483,24 +1478,6 @@ static int radeon_atom_pick_pll(struct drm_crtc *crtc)
 
 }
 
-void radeon_atom_dcpll_init(struct radeon_device *rdev)
-{
-	/* always set DCPLL */
-	if (ASIC_IS_DCE4(rdev)) {
-		struct radeon_atom_ss ss;
-		bool ss_enabled = radeon_atombios_get_asic_ss_info(rdev, &ss,
-								   ASIC_INTERNAL_SS_ON_DCPLL,
-								   rdev->clock.default_dispclk);
-		if (ss_enabled)
-			atombios_crtc_program_ss(rdev, ATOM_DISABLE, ATOM_DCPLL, &ss);
-		/* XXX: DCE5, make sure voltage, dispclk is high enough */
-		atombios_crtc_set_dcpll(rdev, rdev->clock.default_dispclk);
-		if (ss_enabled)
-			atombios_crtc_program_ss(rdev, ATOM_ENABLE, ATOM_DCPLL, &ss);
-	}
-
-}
-
 int atombios_crtc_mode_set(struct drm_crtc *crtc,
 			   struct drm_display_mode *mode,
 			   struct drm_display_mode *adjusted_mode,
@@ -1522,6 +1499,19 @@ int atombios_crtc_mode_set(struct drm_crtc *crtc,
 		}
 	}
 
+	/* always set DCPLL */
+	if (ASIC_IS_DCE4(rdev)) {
+		struct radeon_atom_ss ss;
+		bool ss_enabled = radeon_atombios_get_asic_ss_info(rdev, &ss,
+								   ASIC_INTERNAL_SS_ON_DCPLL,
+								   rdev->clock.default_dispclk);
+		if (ss_enabled)
+			atombios_crtc_program_ss(crtc, ATOM_DISABLE, ATOM_DCPLL, &ss);
+		/* XXX: DCE5, make sure voltage, dispclk is high enough */
+		atombios_crtc_set_dcpll(crtc, rdev->clock.default_dispclk);
+		if (ss_enabled)
+			atombios_crtc_program_ss(crtc, ATOM_ENABLE, ATOM_DCPLL, &ss);
+	}
 	atombios_crtc_set_pll(crtc, adjusted_mode);
 
 	if (ASIC_IS_DCE4(rdev))
@@ -1547,6 +1537,12 @@ static bool atombios_crtc_mode_fixup(struct drm_crtc *crtc,
 				     struct drm_display_mode *mode,
 				     struct drm_display_mode *adjusted_mode)
 {
+	struct drm_device *dev = crtc->dev;
+	struct radeon_device *rdev = dev->dev_private;
+
+	/* adjust pm to upcoming mode change */
+	radeon_pm_compute_clocks(rdev);
+
 	if (!radeon_crtc_scaling_mode_fixup(crtc, mode, adjusted_mode))
 		return false;
 	return true;
diff --git a/drivers/gpu/drm/radeon/atombios_dp.c b/drivers/gpu/drm/radeon/atombios_dp.c
index aee9a39..efc2b21 100644
--- a/drivers/gpu/drm/radeon/atombios_dp.c
+++ b/drivers/gpu/drm/radeon/atombios_dp.c
@@ -483,8 +483,7 @@ static int radeon_dp_get_dp_link_clock(struct drm_connector *connector,
 	int bpp = convert_bpc_to_bpp(connector->display_info.bpc);
 	int lane_num, max_pix_clock;
 
-	if (radeon_connector_encoder_get_dp_bridge_encoder_id(connector) ==
-	    ENCODER_OBJECT_ID_NUTMEG)
+	if (radeon_connector_encoder_is_dp_bridge(connector))
 		return 270000;
 
 	lane_num = radeon_dp_get_dp_lane_number(connector, dpcd, pix_clock);
@@ -550,8 +549,8 @@ bool radeon_dp_getdpcd(struct radeon_connector *radeon_connector)
 	return false;
 }
 
-int radeon_dp_get_panel_mode(struct drm_encoder *encoder,
-			     struct drm_connector *connector)
+static void radeon_dp_set_panel_mode(struct drm_encoder *encoder,
+				     struct drm_connector *connector)
 {
 	struct drm_device *dev = encoder->dev;
 	struct radeon_device *rdev = dev->dev_private;
@@ -559,33 +558,24 @@ int radeon_dp_get_panel_mode(struct drm_encoder *encoder,
 	int panel_mode = DP_PANEL_MODE_EXTERNAL_DP_MODE;
 
 	if (!ASIC_IS_DCE4(rdev))
-		return panel_mode;
+		return;
 
-	if (radeon_connector_encoder_get_dp_bridge_encoder_id(connector) ==
-	    ENCODER_OBJECT_ID_NUTMEG)
+	if (radeon_connector_encoder_is_dp_bridge(connector))
 		panel_mode = DP_PANEL_MODE_INTERNAL_DP1_MODE;
-	else if (radeon_connector_encoder_get_dp_bridge_encoder_id(connector) ==
-		 ENCODER_OBJECT_ID_TRAVIS) {
-		u8 id[6];
-		int i;
-		for (i = 0; i < 6; i++)
-			id[i] = radeon_read_dpcd_reg(radeon_connector, 0x503 + i);
-		if (id[0] == 0x73 &&
-		    id[1] == 0x69 &&
-		    id[2] == 0x76 &&
-		    id[3] == 0x61 &&
-		    id[4] == 0x72 &&
-		    id[5] == 0x54)
-			panel_mode = DP_PANEL_MODE_INTERNAL_DP1_MODE;
-		else
-			panel_mode = DP_PANEL_MODE_INTERNAL_DP2_MODE;
-	} else if (connector->connector_type == DRM_MODE_CONNECTOR_eDP) {
+	else if (connector->connector_type == DRM_MODE_CONNECTOR_eDP) {
 		u8 tmp = radeon_read_dpcd_reg(radeon_connector, DP_EDP_CONFIGURATION_CAP);
 		if (tmp & 1)
 			panel_mode = DP_PANEL_MODE_INTERNAL_DP2_MODE;
 	}
 
-	return panel_mode;
+	atombios_dig_encoder_setup(encoder,
+				   ATOM_ENCODER_CMD_SETUP_PANEL_MODE,
+				   panel_mode);
+
+	if ((connector->connector_type == DRM_MODE_CONNECTOR_eDP) &&
+	    (panel_mode == DP_PANEL_MODE_INTERNAL_DP2_MODE)) {
+		radeon_write_dpcd_reg(radeon_connector, DP_EDP_CONFIGURATION_SET, 1);
+	}
 }
 
 void radeon_dp_set_link_config(struct drm_connector *connector,
@@ -722,8 +712,6 @@ static void radeon_dp_set_tp(struct radeon_dp_link_train_info *dp_info, int tp)
 
 static int radeon_dp_link_train_init(struct radeon_dp_link_train_info *dp_info)
 {
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(dp_info->encoder);
-	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
 	u8 tmp;
 
 	/* power up the sink */
@@ -739,15 +727,11 @@ static int radeon_dp_link_train_init(struct radeon_dp_link_train_info *dp_info)
 		radeon_write_dpcd_reg(dp_info->radeon_connector,
 				      DP_DOWNSPREAD_CTRL, 0);
 
-	if ((dp_info->connector->connector_type == DRM_MODE_CONNECTOR_eDP) &&
-	    (dig->panel_mode == DP_PANEL_MODE_INTERNAL_DP2_MODE)) {
-		radeon_write_dpcd_reg(dp_info->radeon_connector, DP_EDP_CONFIGURATION_SET, 1);
-	}
+	radeon_dp_set_panel_mode(dp_info->encoder, dp_info->connector);
 
 	/* set the lane count on the sink */
 	tmp = dp_info->dp_lane_count;
-	if (dp_info->dpcd[DP_DPCD_REV] >= 0x11 &&
-	    dp_info->dpcd[DP_MAX_LANE_COUNT] & DP_ENHANCED_FRAME_CAP)
+	if (dp_info->dpcd[0] >= 0x11)
 		tmp |= DP_LANE_COUNT_ENHANCED_FRAME_EN;
 	radeon_write_dpcd_reg(dp_info->radeon_connector, DP_LANE_COUNT_SET, tmp);
 
diff --git a/drivers/gpu/drm/radeon/atombios_encoders.c b/drivers/gpu/drm/radeon/atombios_encoders.c
deleted file mode 100644
index b88c460..0000000
--- a/drivers/gpu/drm/radeon/atombios_encoders.c
+++ /dev/null
@@ -1,2359 +0,0 @@
-/*
- * Copyright 2007-11 Advanced Micro Devices, Inc.
- * Copyright 2008 Red Hat Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Dave Airlie
- *          Alex Deucher
- */
-#include "drmP.h"
-#include "drm_crtc_helper.h"
-#include "radeon_drm.h"
-#include "radeon.h"
-#include "atom.h"
-
-extern int atom_debug;
-
-/* evil but including atombios.h is much worse */
-bool radeon_atom_get_tv_timings(struct radeon_device *rdev, int index,
-				struct drm_display_mode *mode);
-
-
-static inline bool radeon_encoder_is_digital(struct drm_encoder *encoder)
-{
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	switch (radeon_encoder->encoder_id) {
-	case ENCODER_OBJECT_ID_INTERNAL_LVDS:
-	case ENCODER_OBJECT_ID_INTERNAL_TMDS1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
-	case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
-	case ENCODER_OBJECT_ID_INTERNAL_DVO1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
-	case ENCODER_OBJECT_ID_INTERNAL_DDI:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
-		return true;
-	default:
-		return false;
-	}
-}
-
-static bool radeon_atom_mode_fixup(struct drm_encoder *encoder,
-				   struct drm_display_mode *mode,
-				   struct drm_display_mode *adjusted_mode)
-{
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	struct drm_device *dev = encoder->dev;
-	struct radeon_device *rdev = dev->dev_private;
-
-	/* set the active encoder to connector routing */
-	radeon_encoder_set_active_device(encoder);
-	drm_mode_set_crtcinfo(adjusted_mode, 0);
-
-	/* hw bug */
-	if ((mode->flags & DRM_MODE_FLAG_INTERLACE)
-	    && (mode->crtc_vsync_start < (mode->crtc_vdisplay + 2)))
-		adjusted_mode->crtc_vsync_start = adjusted_mode->crtc_vdisplay + 2;
-
-	/* get the native mode for LVDS */
-	if (radeon_encoder->active_device & (ATOM_DEVICE_LCD_SUPPORT))
-		radeon_panel_mode_fixup(encoder, adjusted_mode);
-
-	/* get the native mode for TV */
-	if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT)) {
-		struct radeon_encoder_atom_dac *tv_dac = radeon_encoder->enc_priv;
-		if (tv_dac) {
-			if (tv_dac->tv_std == TV_STD_NTSC ||
-			    tv_dac->tv_std == TV_STD_NTSC_J ||
-			    tv_dac->tv_std == TV_STD_PAL_M)
-				radeon_atom_get_tv_timings(rdev, 0, adjusted_mode);
-			else
-				radeon_atom_get_tv_timings(rdev, 1, adjusted_mode);
-		}
-	}
-
-	if (ASIC_IS_DCE3(rdev) &&
-	    ((radeon_encoder->active_device & (ATOM_DEVICE_DFP_SUPPORT | ATOM_DEVICE_LCD_SUPPORT)) ||
-	     (radeon_encoder_get_dp_bridge_encoder_id(encoder) != ENCODER_OBJECT_ID_NONE))) {
-		struct drm_connector *connector = radeon_get_connector_for_encoder(encoder);
-		radeon_dp_set_link_config(connector, mode);
-	}
-
-	return true;
-}
-
-static void
-atombios_dac_setup(struct drm_encoder *encoder, int action)
-{
-	struct drm_device *dev = encoder->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	DAC_ENCODER_CONTROL_PS_ALLOCATION args;
-	int index = 0;
-	struct radeon_encoder_atom_dac *dac_info = radeon_encoder->enc_priv;
-
-	memset(&args, 0, sizeof(args));
-
-	switch (radeon_encoder->encoder_id) {
-	case ENCODER_OBJECT_ID_INTERNAL_DAC1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1:
-		index = GetIndexIntoMasterTable(COMMAND, DAC1EncoderControl);
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_DAC2:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC2:
-		index = GetIndexIntoMasterTable(COMMAND, DAC2EncoderControl);
-		break;
-	}
-
-	args.ucAction = action;
-
-	if (radeon_encoder->active_device & (ATOM_DEVICE_CRT_SUPPORT))
-		args.ucDacStandard = ATOM_DAC1_PS2;
-	else if (radeon_encoder->active_device & (ATOM_DEVICE_CV_SUPPORT))
-		args.ucDacStandard = ATOM_DAC1_CV;
-	else {
-		switch (dac_info->tv_std) {
-		case TV_STD_PAL:
-		case TV_STD_PAL_M:
-		case TV_STD_SCART_PAL:
-		case TV_STD_SECAM:
-		case TV_STD_PAL_CN:
-			args.ucDacStandard = ATOM_DAC1_PAL;
-			break;
-		case TV_STD_NTSC:
-		case TV_STD_NTSC_J:
-		case TV_STD_PAL_60:
-		default:
-			args.ucDacStandard = ATOM_DAC1_NTSC;
-			break;
-		}
-	}
-	args.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
-
-	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
-
-}
-
-static void
-atombios_tv_setup(struct drm_encoder *encoder, int action)
-{
-	struct drm_device *dev = encoder->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	TV_ENCODER_CONTROL_PS_ALLOCATION args;
-	int index = 0;
-	struct radeon_encoder_atom_dac *dac_info = radeon_encoder->enc_priv;
-
-	memset(&args, 0, sizeof(args));
-
-	index = GetIndexIntoMasterTable(COMMAND, TVEncoderControl);
-
-	args.sTVEncoder.ucAction = action;
-
-	if (radeon_encoder->active_device & (ATOM_DEVICE_CV_SUPPORT))
-		args.sTVEncoder.ucTvStandard = ATOM_TV_CV;
-	else {
-		switch (dac_info->tv_std) {
-		case TV_STD_NTSC:
-			args.sTVEncoder.ucTvStandard = ATOM_TV_NTSC;
-			break;
-		case TV_STD_PAL:
-			args.sTVEncoder.ucTvStandard = ATOM_TV_PAL;
-			break;
-		case TV_STD_PAL_M:
-			args.sTVEncoder.ucTvStandard = ATOM_TV_PALM;
-			break;
-		case TV_STD_PAL_60:
-			args.sTVEncoder.ucTvStandard = ATOM_TV_PAL60;
-			break;
-		case TV_STD_NTSC_J:
-			args.sTVEncoder.ucTvStandard = ATOM_TV_NTSCJ;
-			break;
-		case TV_STD_SCART_PAL:
-			args.sTVEncoder.ucTvStandard = ATOM_TV_PAL; /* ??? */
-			break;
-		case TV_STD_SECAM:
-			args.sTVEncoder.ucTvStandard = ATOM_TV_SECAM;
-			break;
-		case TV_STD_PAL_CN:
-			args.sTVEncoder.ucTvStandard = ATOM_TV_PALCN;
-			break;
-		default:
-			args.sTVEncoder.ucTvStandard = ATOM_TV_NTSC;
-			break;
-		}
-	}
-
-	args.sTVEncoder.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
-
-	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
-
-}
-
-union dvo_encoder_control {
-	ENABLE_EXTERNAL_TMDS_ENCODER_PS_ALLOCATION ext_tmds;
-	DVO_ENCODER_CONTROL_PS_ALLOCATION dvo;
-	DVO_ENCODER_CONTROL_PS_ALLOCATION_V3 dvo_v3;
-};
-
-void
-atombios_dvo_setup(struct drm_encoder *encoder, int action)
-{
-	struct drm_device *dev = encoder->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	union dvo_encoder_control args;
-	int index = GetIndexIntoMasterTable(COMMAND, DVOEncoderControl);
-	uint8_t frev, crev;
-
-	memset(&args, 0, sizeof(args));
-
-	if (!atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev))
-		return;
-
-	switch (frev) {
-	case 1:
-		switch (crev) {
-		case 1:
-			/* R4xx, R5xx */
-			args.ext_tmds.sXTmdsEncoder.ucEnable = action;
-
-			if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock))
-				args.ext_tmds.sXTmdsEncoder.ucMisc |= PANEL_ENCODER_MISC_DUAL;
-
-			args.ext_tmds.sXTmdsEncoder.ucMisc |= ATOM_PANEL_MISC_888RGB;
-			break;
-		case 2:
-			/* RS600/690/740 */
-			args.dvo.sDVOEncoder.ucAction = action;
-			args.dvo.sDVOEncoder.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
-			/* DFP1, CRT1, TV1 depending on the type of port */
-			args.dvo.sDVOEncoder.ucDeviceType = ATOM_DEVICE_DFP1_INDEX;
-
-			if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock))
-				args.dvo.sDVOEncoder.usDevAttr.sDigAttrib.ucAttribute |= PANEL_ENCODER_MISC_DUAL;
-			break;
-		case 3:
-			/* R6xx */
-			args.dvo_v3.ucAction = action;
-			args.dvo_v3.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
-			args.dvo_v3.ucDVOConfig = 0; /* XXX */
-			break;
-		default:
-			DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
-			break;
-		}
-		break;
-	default:
-		DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
-		break;
-	}
-
-	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
-}
-
-union lvds_encoder_control {
-	LVDS_ENCODER_CONTROL_PS_ALLOCATION    v1;
-	LVDS_ENCODER_CONTROL_PS_ALLOCATION_V2 v2;
-};
-
-void
-atombios_digital_setup(struct drm_encoder *encoder, int action)
-{
-	struct drm_device *dev = encoder->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
-	union lvds_encoder_control args;
-	int index = 0;
-	int hdmi_detected = 0;
-	uint8_t frev, crev;
-
-	if (!dig)
-		return;
-
-	if (atombios_get_encoder_mode(encoder) == ATOM_ENCODER_MODE_HDMI)
-		hdmi_detected = 1;
-
-	memset(&args, 0, sizeof(args));
-
-	switch (radeon_encoder->encoder_id) {
-	case ENCODER_OBJECT_ID_INTERNAL_LVDS:
-		index = GetIndexIntoMasterTable(COMMAND, LVDSEncoderControl);
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_TMDS1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
-		index = GetIndexIntoMasterTable(COMMAND, TMDS1EncoderControl);
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
-		if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT))
-			index = GetIndexIntoMasterTable(COMMAND, LVDSEncoderControl);
-		else
-			index = GetIndexIntoMasterTable(COMMAND, TMDS2EncoderControl);
-		break;
-	}
-
-	if (!atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev))
-		return;
-
-	switch (frev) {
-	case 1:
-	case 2:
-		switch (crev) {
-		case 1:
-			args.v1.ucMisc = 0;
-			args.v1.ucAction = action;
-			if (hdmi_detected)
-				args.v1.ucMisc |= PANEL_ENCODER_MISC_HDMI_TYPE;
-			args.v1.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
-			if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) {
-				if (dig->lcd_misc & ATOM_PANEL_MISC_DUAL)
-					args.v1.ucMisc |= PANEL_ENCODER_MISC_DUAL;
-				if (dig->lcd_misc & ATOM_PANEL_MISC_888RGB)
-					args.v1.ucMisc |= ATOM_PANEL_MISC_888RGB;
-			} else {
-				if (dig->linkb)
-					args.v1.ucMisc |= PANEL_ENCODER_MISC_TMDS_LINKB;
-				if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock))
-					args.v1.ucMisc |= PANEL_ENCODER_MISC_DUAL;
-				/*if (pScrn->rgbBits == 8) */
-				args.v1.ucMisc |= ATOM_PANEL_MISC_888RGB;
-			}
-			break;
-		case 2:
-		case 3:
-			args.v2.ucMisc = 0;
-			args.v2.ucAction = action;
-			if (crev == 3) {
-				if (dig->coherent_mode)
-					args.v2.ucMisc |= PANEL_ENCODER_MISC_COHERENT;
-			}
-			if (hdmi_detected)
-				args.v2.ucMisc |= PANEL_ENCODER_MISC_HDMI_TYPE;
-			args.v2.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
-			args.v2.ucTruncate = 0;
-			args.v2.ucSpatial = 0;
-			args.v2.ucTemporal = 0;
-			args.v2.ucFRC = 0;
-			if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) {
-				if (dig->lcd_misc & ATOM_PANEL_MISC_DUAL)
-					args.v2.ucMisc |= PANEL_ENCODER_MISC_DUAL;
-				if (dig->lcd_misc & ATOM_PANEL_MISC_SPATIAL) {
-					args.v2.ucSpatial = PANEL_ENCODER_SPATIAL_DITHER_EN;
-					if (dig->lcd_misc & ATOM_PANEL_MISC_888RGB)
-						args.v2.ucSpatial |= PANEL_ENCODER_SPATIAL_DITHER_DEPTH;
-				}
-				if (dig->lcd_misc & ATOM_PANEL_MISC_TEMPORAL) {
-					args.v2.ucTemporal = PANEL_ENCODER_TEMPORAL_DITHER_EN;
-					if (dig->lcd_misc & ATOM_PANEL_MISC_888RGB)
-						args.v2.ucTemporal |= PANEL_ENCODER_TEMPORAL_DITHER_DEPTH;
-					if (((dig->lcd_misc >> ATOM_PANEL_MISC_GREY_LEVEL_SHIFT) & 0x3) == 2)
-						args.v2.ucTemporal |= PANEL_ENCODER_TEMPORAL_LEVEL_4;
-				}
-			} else {
-				if (dig->linkb)
-					args.v2.ucMisc |= PANEL_ENCODER_MISC_TMDS_LINKB;
-				if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock))
-					args.v2.ucMisc |= PANEL_ENCODER_MISC_DUAL;
-			}
-			break;
-		default:
-			DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
-			break;
-		}
-		break;
-	default:
-		DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
-		break;
-	}
-
-	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
-}
-
-int
-atombios_get_encoder_mode(struct drm_encoder *encoder)
-{
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	struct drm_connector *connector;
-	struct radeon_connector *radeon_connector;
-	struct radeon_connector_atom_dig *dig_connector;
-
-	/* dp bridges are always DP */
-	if (radeon_encoder_get_dp_bridge_encoder_id(encoder) != ENCODER_OBJECT_ID_NONE)
-		return ATOM_ENCODER_MODE_DP;
-
-	/* DVO is always DVO */
-	if (radeon_encoder->encoder_id == ATOM_ENCODER_MODE_DVO)
-		return ATOM_ENCODER_MODE_DVO;
-
-	connector = radeon_get_connector_for_encoder(encoder);
-	/* if we don't have an active device yet, just use one of
-	 * the connectors tied to the encoder.
-	 */
-	if (!connector)
-		connector = radeon_get_connector_for_encoder_init(encoder);
-	radeon_connector = to_radeon_connector(connector);
-
-	switch (connector->connector_type) {
-	case DRM_MODE_CONNECTOR_DVII:
-	case DRM_MODE_CONNECTOR_HDMIB: /* HDMI-B is basically DL-DVI; analog works fine */
-		if (drm_detect_hdmi_monitor(radeon_connector->edid) &&
-		    radeon_audio)
-			return ATOM_ENCODER_MODE_HDMI;
-		else if (radeon_connector->use_digital)
-			return ATOM_ENCODER_MODE_DVI;
-		else
-			return ATOM_ENCODER_MODE_CRT;
-		break;
-	case DRM_MODE_CONNECTOR_DVID:
-	case DRM_MODE_CONNECTOR_HDMIA:
-	default:
-		if (drm_detect_hdmi_monitor(radeon_connector->edid) &&
-		    radeon_audio)
-			return ATOM_ENCODER_MODE_HDMI;
-		else
-			return ATOM_ENCODER_MODE_DVI;
-		break;
-	case DRM_MODE_CONNECTOR_LVDS:
-		return ATOM_ENCODER_MODE_LVDS;
-		break;
-	case DRM_MODE_CONNECTOR_DisplayPort:
-		dig_connector = radeon_connector->con_priv;
-		if ((dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_DISPLAYPORT) ||
-		    (dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_eDP))
-			return ATOM_ENCODER_MODE_DP;
-		else if (drm_detect_hdmi_monitor(radeon_connector->edid) &&
-			 radeon_audio)
-			return ATOM_ENCODER_MODE_HDMI;
-		else
-			return ATOM_ENCODER_MODE_DVI;
-		break;
-	case DRM_MODE_CONNECTOR_eDP:
-		return ATOM_ENCODER_MODE_DP;
-	case DRM_MODE_CONNECTOR_DVIA:
-	case DRM_MODE_CONNECTOR_VGA:
-		return ATOM_ENCODER_MODE_CRT;
-		break;
-	case DRM_MODE_CONNECTOR_Composite:
-	case DRM_MODE_CONNECTOR_SVIDEO:
-	case DRM_MODE_CONNECTOR_9PinDIN:
-		/* fix me */
-		return ATOM_ENCODER_MODE_TV;
-		/*return ATOM_ENCODER_MODE_CV;*/
-		break;
-	}
-}
-
-/*
- * DIG Encoder/Transmitter Setup
- *
- * DCE 3.0/3.1
- * - 2 DIG transmitter blocks. UNIPHY (links A and B) and LVTMA.
- * Supports up to 3 digital outputs
- * - 2 DIG encoder blocks.
- * DIG1 can drive UNIPHY link A or link B
- * DIG2 can drive UNIPHY link B or LVTMA
- *
- * DCE 3.2
- * - 3 DIG transmitter blocks. UNIPHY0/1/2 (links A and B).
- * Supports up to 5 digital outputs
- * - 2 DIG encoder blocks.
- * DIG1/2 can drive UNIPHY0/1/2 link A or link B
- *
- * DCE 4.0/5.0
- * - 3 DIG transmitter blocks UNIPHY0/1/2 (links A and B).
- * Supports up to 6 digital outputs
- * - 6 DIG encoder blocks.
- * - DIG to PHY mapping is hardcoded
- * DIG1 drives UNIPHY0 link A, A+B
- * DIG2 drives UNIPHY0 link B
- * DIG3 drives UNIPHY1 link A, A+B
- * DIG4 drives UNIPHY1 link B
- * DIG5 drives UNIPHY2 link A, A+B
- * DIG6 drives UNIPHY2 link B
- *
- * DCE 4.1
- * - 3 DIG transmitter blocks UNIPHY0/1/2 (links A and B).
- * Supports up to 6 digital outputs
- * - 2 DIG encoder blocks.
- * DIG1/2 can drive UNIPHY0/1/2 link A or link B
- *
- * Routing
- * crtc -> dig encoder -> UNIPHY/LVTMA (1 or 2 links)
- * Examples:
- * crtc0 -> dig2 -> LVTMA   links A+B -> TMDS/HDMI
- * crtc1 -> dig1 -> UNIPHY0 link  B   -> DP
- * crtc0 -> dig1 -> UNIPHY2 link  A   -> LVDS
- * crtc1 -> dig2 -> UNIPHY1 link  B+A -> TMDS/HDMI
- */
-
-union dig_encoder_control {
-	DIG_ENCODER_CONTROL_PS_ALLOCATION v1;
-	DIG_ENCODER_CONTROL_PARAMETERS_V2 v2;
-	DIG_ENCODER_CONTROL_PARAMETERS_V3 v3;
-	DIG_ENCODER_CONTROL_PARAMETERS_V4 v4;
-};
-
-void
-atombios_dig_encoder_setup(struct drm_encoder *encoder, int action, int panel_mode)
-{
-	struct drm_device *dev = encoder->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
-	struct drm_connector *connector = radeon_get_connector_for_encoder(encoder);
-	union dig_encoder_control args;
-	int index = 0;
-	uint8_t frev, crev;
-	int dp_clock = 0;
-	int dp_lane_count = 0;
-	int hpd_id = RADEON_HPD_NONE;
-	int bpc = 8;
-
-	if (connector) {
-		struct radeon_connector *radeon_connector = to_radeon_connector(connector);
-		struct radeon_connector_atom_dig *dig_connector =
-			radeon_connector->con_priv;
-
-		dp_clock = dig_connector->dp_clock;
-		dp_lane_count = dig_connector->dp_lane_count;
-		hpd_id = radeon_connector->hpd.hpd;
-		bpc = connector->display_info.bpc;
-	}
-
-	/* no dig encoder assigned */
-	if (dig->dig_encoder == -1)
-		return;
-
-	memset(&args, 0, sizeof(args));
-
-	if (ASIC_IS_DCE4(rdev))
-		index = GetIndexIntoMasterTable(COMMAND, DIGxEncoderControl);
-	else {
-		if (dig->dig_encoder)
-			index = GetIndexIntoMasterTable(COMMAND, DIG2EncoderControl);
-		else
-			index = GetIndexIntoMasterTable(COMMAND, DIG1EncoderControl);
-	}
-
-	if (!atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev))
-		return;
-
-	switch (frev) {
-	case 1:
-		switch (crev) {
-		case 1:
-			args.v1.ucAction = action;
-			args.v1.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
-			if (action == ATOM_ENCODER_CMD_SETUP_PANEL_MODE)
-				args.v3.ucPanelMode = panel_mode;
-			else
-				args.v1.ucEncoderMode = atombios_get_encoder_mode(encoder);
-
-			if (ENCODER_MODE_IS_DP(args.v1.ucEncoderMode))
-				args.v1.ucLaneNum = dp_lane_count;
-			else if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock))
-				args.v1.ucLaneNum = 8;
-			else
-				args.v1.ucLaneNum = 4;
-
-			if (ENCODER_MODE_IS_DP(args.v1.ucEncoderMode) && (dp_clock == 270000))
-				args.v1.ucConfig |= ATOM_ENCODER_CONFIG_DPLINKRATE_2_70GHZ;
-			switch (radeon_encoder->encoder_id) {
-			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
-				args.v1.ucConfig = ATOM_ENCODER_CONFIG_V2_TRANSMITTER1;
-				break;
-			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
-			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
-				args.v1.ucConfig = ATOM_ENCODER_CONFIG_V2_TRANSMITTER2;
-				break;
-			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
-				args.v1.ucConfig = ATOM_ENCODER_CONFIG_V2_TRANSMITTER3;
-				break;
-			}
-			if (dig->linkb)
-				args.v1.ucConfig |= ATOM_ENCODER_CONFIG_LINKB;
-			else
-				args.v1.ucConfig |= ATOM_ENCODER_CONFIG_LINKA;
-			break;
-		case 2:
-		case 3:
-			args.v3.ucAction = action;
-			args.v3.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
-			if (action == ATOM_ENCODER_CMD_SETUP_PANEL_MODE)
-				args.v3.ucPanelMode = panel_mode;
-			else
-				args.v3.ucEncoderMode = atombios_get_encoder_mode(encoder);
-
-			if (ENCODER_MODE_IS_DP(args.v1.ucEncoderMode))
-				args.v3.ucLaneNum = dp_lane_count;
-			else if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock))
-				args.v3.ucLaneNum = 8;
-			else
-				args.v3.ucLaneNum = 4;
-
-			if (ENCODER_MODE_IS_DP(args.v1.ucEncoderMode) && (dp_clock == 270000))
-				args.v1.ucConfig |= ATOM_ENCODER_CONFIG_V3_DPLINKRATE_2_70GHZ;
-			args.v3.acConfig.ucDigSel = dig->dig_encoder;
-			switch (bpc) {
-			case 0:
-				args.v3.ucBitPerColor = PANEL_BPC_UNDEFINE;
-				break;
-			case 6:
-				args.v3.ucBitPerColor = PANEL_6BIT_PER_COLOR;
-				break;
-			case 8:
-			default:
-				args.v3.ucBitPerColor = PANEL_8BIT_PER_COLOR;
-				break;
-			case 10:
-				args.v3.ucBitPerColor = PANEL_10BIT_PER_COLOR;
-				break;
-			case 12:
-				args.v3.ucBitPerColor = PANEL_12BIT_PER_COLOR;
-				break;
-			case 16:
-				args.v3.ucBitPerColor = PANEL_16BIT_PER_COLOR;
-				break;
-			}
-			break;
-		case 4:
-			args.v4.ucAction = action;
-			args.v4.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
-			if (action == ATOM_ENCODER_CMD_SETUP_PANEL_MODE)
-				args.v4.ucPanelMode = panel_mode;
-			else
-				args.v4.ucEncoderMode = atombios_get_encoder_mode(encoder);
-
-			if (ENCODER_MODE_IS_DP(args.v1.ucEncoderMode))
-				args.v4.ucLaneNum = dp_lane_count;
-			else if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock))
-				args.v4.ucLaneNum = 8;
-			else
-				args.v4.ucLaneNum = 4;
-
-			if (ENCODER_MODE_IS_DP(args.v1.ucEncoderMode)) {
-				if (dp_clock == 270000)
-					args.v1.ucConfig |= ATOM_ENCODER_CONFIG_V4_DPLINKRATE_2_70GHZ;
-				else if (dp_clock == 540000)
-					args.v1.ucConfig |= ATOM_ENCODER_CONFIG_V4_DPLINKRATE_5_40GHZ;
-			}
-			args.v4.acConfig.ucDigSel = dig->dig_encoder;
-			switch (bpc) {
-			case 0:
-				args.v4.ucBitPerColor = PANEL_BPC_UNDEFINE;
-				break;
-			case 6:
-				args.v4.ucBitPerColor = PANEL_6BIT_PER_COLOR;
-				break;
-			case 8:
-			default:
-				args.v4.ucBitPerColor = PANEL_8BIT_PER_COLOR;
-				break;
-			case 10:
-				args.v4.ucBitPerColor = PANEL_10BIT_PER_COLOR;
-				break;
-			case 12:
-				args.v4.ucBitPerColor = PANEL_12BIT_PER_COLOR;
-				break;
-			case 16:
-				args.v4.ucBitPerColor = PANEL_16BIT_PER_COLOR;
-				break;
-			}
-			if (hpd_id == RADEON_HPD_NONE)
-				args.v4.ucHPD_ID = 0;
-			else
-				args.v4.ucHPD_ID = hpd_id + 1;
-			break;
-		default:
-			DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
-			break;
-		}
-		break;
-	default:
-		DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
-		break;
-	}
-
-	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
-
-}
-
-union dig_transmitter_control {
-	DIG_TRANSMITTER_CONTROL_PS_ALLOCATION v1;
-	DIG_TRANSMITTER_CONTROL_PARAMETERS_V2 v2;
-	DIG_TRANSMITTER_CONTROL_PARAMETERS_V3 v3;
-	DIG_TRANSMITTER_CONTROL_PARAMETERS_V4 v4;
-};
-
-void
-atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t lane_num, uint8_t lane_set)
-{
-	struct drm_device *dev = encoder->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
-	struct drm_connector *connector;
-	union dig_transmitter_control args;
-	int index = 0;
-	uint8_t frev, crev;
-	bool is_dp = false;
-	int pll_id = 0;
-	int dp_clock = 0;
-	int dp_lane_count = 0;
-	int connector_object_id = 0;
-	int igp_lane_info = 0;
-	int dig_encoder = dig->dig_encoder;
-
-	if (action == ATOM_TRANSMITTER_ACTION_INIT) {
-		connector = radeon_get_connector_for_encoder_init(encoder);
-		/* just needed to avoid bailing in the encoder check.  the encoder
-		 * isn't used for init
-		 */
-		dig_encoder = 0;
-	} else
-		connector = radeon_get_connector_for_encoder(encoder);
-
-	if (connector) {
-		struct radeon_connector *radeon_connector = to_radeon_connector(connector);
-		struct radeon_connector_atom_dig *dig_connector =
-			radeon_connector->con_priv;
-
-		dp_clock = dig_connector->dp_clock;
-		dp_lane_count = dig_connector->dp_lane_count;
-		connector_object_id =
-			(radeon_connector->connector_object_id & OBJECT_ID_MASK) >> OBJECT_ID_SHIFT;
-		igp_lane_info = dig_connector->igp_lane_info;
-	}
-
-	if (encoder->crtc) {
-		struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc);
-		pll_id = radeon_crtc->pll_id;
-	}
-
-	/* no dig encoder assigned */
-	if (dig_encoder == -1)
-		return;
-
-	if (ENCODER_MODE_IS_DP(atombios_get_encoder_mode(encoder)))
-		is_dp = true;
-
-	memset(&args, 0, sizeof(args));
-
-	switch (radeon_encoder->encoder_id) {
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
-		index = GetIndexIntoMasterTable(COMMAND, DVOOutputControl);
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
-		index = GetIndexIntoMasterTable(COMMAND, UNIPHYTransmitterControl);
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
-		index = GetIndexIntoMasterTable(COMMAND, LVTMATransmitterControl);
-		break;
-	}
-
-	if (!atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev))
-		return;
-
-	switch (frev) {
-	case 1:
-		switch (crev) {
-		case 1:
-			args.v1.ucAction = action;
-			if (action == ATOM_TRANSMITTER_ACTION_INIT) {
-				args.v1.usInitInfo = cpu_to_le16(connector_object_id);
-			} else if (action == ATOM_TRANSMITTER_ACTION_SETUP_VSEMPH) {
-				args.v1.asMode.ucLaneSel = lane_num;
-				args.v1.asMode.ucLaneSet = lane_set;
-			} else {
-				if (is_dp)
-					args.v1.usPixelClock =
-						cpu_to_le16(dp_clock / 10);
-				else if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock))
-					args.v1.usPixelClock = cpu_to_le16((radeon_encoder->pixel_clock / 2) / 10);
-				else
-					args.v1.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
-			}
-
-			args.v1.ucConfig = ATOM_TRANSMITTER_CONFIG_CLKSRC_PPLL;
-
-			if (dig_encoder)
-				args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_DIG2_ENCODER;
-			else
-				args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_DIG1_ENCODER;
-
-			if ((rdev->flags & RADEON_IS_IGP) &&
-			    (radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_UNIPHY)) {
-				if (is_dp ||
-				    !radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock)) {
-					if (igp_lane_info & 0x1)
-						args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LANE_0_3;
-					else if (igp_lane_info & 0x2)
-						args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LANE_4_7;
-					else if (igp_lane_info & 0x4)
-						args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LANE_8_11;
-					else if (igp_lane_info & 0x8)
-						args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LANE_12_15;
-				} else {
-					if (igp_lane_info & 0x3)
-						args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LANE_0_7;
-					else if (igp_lane_info & 0xc)
-						args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LANE_8_15;
-				}
-			}
-
-			if (dig->linkb)
-				args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LINKB;
-			else
-				args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LINKA;
-
-			if (is_dp)
-				args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_COHERENT;
-			else if (radeon_encoder->devices & (ATOM_DEVICE_DFP_SUPPORT)) {
-				if (dig->coherent_mode)
-					args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_COHERENT;
-				if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock))
-					args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_8LANE_LINK;
-			}
-			break;
-		case 2:
-			args.v2.ucAction = action;
-			if (action == ATOM_TRANSMITTER_ACTION_INIT) {
-				args.v2.usInitInfo = cpu_to_le16(connector_object_id);
-			} else if (action == ATOM_TRANSMITTER_ACTION_SETUP_VSEMPH) {
-				args.v2.asMode.ucLaneSel = lane_num;
-				args.v2.asMode.ucLaneSet = lane_set;
-			} else {
-				if (is_dp)
-					args.v2.usPixelClock =
-						cpu_to_le16(dp_clock / 10);
-				else if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock))
-					args.v2.usPixelClock = cpu_to_le16((radeon_encoder->pixel_clock / 2) / 10);
-				else
-					args.v2.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
-			}
-
-			args.v2.acConfig.ucEncoderSel = dig_encoder;
-			if (dig->linkb)
-				args.v2.acConfig.ucLinkSel = 1;
-
-			switch (radeon_encoder->encoder_id) {
-			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
-				args.v2.acConfig.ucTransmitterSel = 0;
-				break;
-			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
-				args.v2.acConfig.ucTransmitterSel = 1;
-				break;
-			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
-				args.v2.acConfig.ucTransmitterSel = 2;
-				break;
-			}
-
-			if (is_dp) {
-				args.v2.acConfig.fCoherentMode = 1;
-				args.v2.acConfig.fDPConnector = 1;
-			} else if (radeon_encoder->devices & (ATOM_DEVICE_DFP_SUPPORT)) {
-				if (dig->coherent_mode)
-					args.v2.acConfig.fCoherentMode = 1;
-				if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock))
-					args.v2.acConfig.fDualLinkConnector = 1;
-			}
-			break;
-		case 3:
-			args.v3.ucAction = action;
-			if (action == ATOM_TRANSMITTER_ACTION_INIT) {
-				args.v3.usInitInfo = cpu_to_le16(connector_object_id);
-			} else if (action == ATOM_TRANSMITTER_ACTION_SETUP_VSEMPH) {
-				args.v3.asMode.ucLaneSel = lane_num;
-				args.v3.asMode.ucLaneSet = lane_set;
-			} else {
-				if (is_dp)
-					args.v3.usPixelClock =
-						cpu_to_le16(dp_clock / 10);
-				else if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock))
-					args.v3.usPixelClock = cpu_to_le16((radeon_encoder->pixel_clock / 2) / 10);
-				else
-					args.v3.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
-			}
-
-			if (is_dp)
-				args.v3.ucLaneNum = dp_lane_count;
-			else if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock))
-				args.v3.ucLaneNum = 8;
-			else
-				args.v3.ucLaneNum = 4;
-
-			if (dig->linkb)
-				args.v3.acConfig.ucLinkSel = 1;
-			if (dig_encoder & 1)
-				args.v3.acConfig.ucEncoderSel = 1;
-
-			/* Select the PLL for the PHY
-			 * DP PHY should be clocked from external src if there is
-			 * one.
-			 */
-			/* On DCE4, if there is an external clock, it generates the DP ref clock */
-			if (is_dp && rdev->clock.dp_extclk)
-				args.v3.acConfig.ucRefClkSource = 2; /* external src */
-			else
-				args.v3.acConfig.ucRefClkSource = pll_id;
-
-			switch (radeon_encoder->encoder_id) {
-			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
-				args.v3.acConfig.ucTransmitterSel = 0;
-				break;
-			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
-				args.v3.acConfig.ucTransmitterSel = 1;
-				break;
-			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
-				args.v3.acConfig.ucTransmitterSel = 2;
-				break;
-			}
-
-			if (is_dp)
-				args.v3.acConfig.fCoherentMode = 1; /* DP requires coherent */
-			else if (radeon_encoder->devices & (ATOM_DEVICE_DFP_SUPPORT)) {
-				if (dig->coherent_mode)
-					args.v3.acConfig.fCoherentMode = 1;
-				if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock))
-					args.v3.acConfig.fDualLinkConnector = 1;
-			}
-			break;
-		case 4:
-			args.v4.ucAction = action;
-			if (action == ATOM_TRANSMITTER_ACTION_INIT) {
-				args.v4.usInitInfo = cpu_to_le16(connector_object_id);
-			} else if (action == ATOM_TRANSMITTER_ACTION_SETUP_VSEMPH) {
-				args.v4.asMode.ucLaneSel = lane_num;
-				args.v4.asMode.ucLaneSet = lane_set;
-			} else {
-				if (is_dp)
-					args.v4.usPixelClock =
-						cpu_to_le16(dp_clock / 10);
-				else if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock))
-					args.v4.usPixelClock = cpu_to_le16((radeon_encoder->pixel_clock / 2) / 10);
-				else
-					args.v4.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
-			}
-
-			if (is_dp)
-				args.v4.ucLaneNum = dp_lane_count;
-			else if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock))
-				args.v4.ucLaneNum = 8;
-			else
-				args.v4.ucLaneNum = 4;
-
-			if (dig->linkb)
-				args.v4.acConfig.ucLinkSel = 1;
-			if (dig_encoder & 1)
-				args.v4.acConfig.ucEncoderSel = 1;
-
-			/* Select the PLL for the PHY
-			 * DP PHY should be clocked from external src if there is
-			 * one.
-			 */
-			/* On DCE5 DCPLL usually generates the DP ref clock */
-			if (is_dp) {
-				if (rdev->clock.dp_extclk)
-					args.v4.acConfig.ucRefClkSource = ENCODER_REFCLK_SRC_EXTCLK;
-				else
-					args.v4.acConfig.ucRefClkSource = ENCODER_REFCLK_SRC_DCPLL;
-			} else
-				args.v4.acConfig.ucRefClkSource = pll_id;
-
-			switch (radeon_encoder->encoder_id) {
-			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
-				args.v4.acConfig.ucTransmitterSel = 0;
-				break;
-			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
-				args.v4.acConfig.ucTransmitterSel = 1;
-				break;
-			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
-				args.v4.acConfig.ucTransmitterSel = 2;
-				break;
-			}
-
-			if (is_dp)
-				args.v4.acConfig.fCoherentMode = 1; /* DP requires coherent */
-			else if (radeon_encoder->devices & (ATOM_DEVICE_DFP_SUPPORT)) {
-				if (dig->coherent_mode)
-					args.v4.acConfig.fCoherentMode = 1;
-				if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock))
-					args.v4.acConfig.fDualLinkConnector = 1;
-			}
-			break;
-		default:
-			DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
-			break;
-		}
-		break;
-	default:
-		DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
-		break;
-	}
-
-	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
-}
-
-bool
-atombios_set_edp_panel_power(struct drm_connector *connector, int action)
-{
-	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
-	struct drm_device *dev = radeon_connector->base.dev;
-	struct radeon_device *rdev = dev->dev_private;
-	union dig_transmitter_control args;
-	int index = GetIndexIntoMasterTable(COMMAND, UNIPHYTransmitterControl);
-	uint8_t frev, crev;
-
-	if (connector->connector_type != DRM_MODE_CONNECTOR_eDP)
-		goto done;
-
-	if (!ASIC_IS_DCE4(rdev))
-		goto done;
-
-	if ((action != ATOM_TRANSMITTER_ACTION_POWER_ON) &&
-	    (action != ATOM_TRANSMITTER_ACTION_POWER_OFF))
-		goto done;
-
-	if (!atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev))
-		goto done;
-
-	memset(&args, 0, sizeof(args));
-
-	args.v1.ucAction = action;
-
-	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
-
-	/* wait for the panel to power up */
-	if (action == ATOM_TRANSMITTER_ACTION_POWER_ON) {
-		int i;
-
-		for (i = 0; i < 300; i++) {
-			if (radeon_hpd_sense(rdev, radeon_connector->hpd.hpd))
-				return true;
-			mdelay(1);
-		}
-		return false;
-	}
-done:
-	return true;
-}
-
-union external_encoder_control {
-	EXTERNAL_ENCODER_CONTROL_PS_ALLOCATION v1;
-	EXTERNAL_ENCODER_CONTROL_PS_ALLOCATION_V3 v3;
-};
-
-static void
-atombios_external_encoder_setup(struct drm_encoder *encoder,
-				struct drm_encoder *ext_encoder,
-				int action)
-{
-	struct drm_device *dev = encoder->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	struct radeon_encoder *ext_radeon_encoder = to_radeon_encoder(ext_encoder);
-	union external_encoder_control args;
-	struct drm_connector *connector;
-	int index = GetIndexIntoMasterTable(COMMAND, ExternalEncoderControl);
-	u8 frev, crev;
-	int dp_clock = 0;
-	int dp_lane_count = 0;
-	int connector_object_id = 0;
-	u32 ext_enum = (ext_radeon_encoder->encoder_enum & ENUM_ID_MASK) >> ENUM_ID_SHIFT;
-	int bpc = 8;
-
-	if (action == EXTERNAL_ENCODER_ACTION_V3_ENCODER_INIT)
-		connector = radeon_get_connector_for_encoder_init(encoder);
-	else
-		connector = radeon_get_connector_for_encoder(encoder);
-
-	if (connector) {
-		struct radeon_connector *radeon_connector = to_radeon_connector(connector);
-		struct radeon_connector_atom_dig *dig_connector =
-			radeon_connector->con_priv;
-
-		dp_clock = dig_connector->dp_clock;
-		dp_lane_count = dig_connector->dp_lane_count;
-		connector_object_id =
-			(radeon_connector->connector_object_id & OBJECT_ID_MASK) >> OBJECT_ID_SHIFT;
-		bpc = connector->display_info.bpc;
-	}
-
-	memset(&args, 0, sizeof(args));
-
-	if (!atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev))
-		return;
-
-	switch (frev) {
-	case 1:
-		/* no params on frev 1 */
-		break;
-	case 2:
-		switch (crev) {
-		case 1:
-		case 2:
-			args.v1.sDigEncoder.ucAction = action;
-			args.v1.sDigEncoder.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
-			args.v1.sDigEncoder.ucEncoderMode = atombios_get_encoder_mode(encoder);
-
-			if (ENCODER_MODE_IS_DP(args.v1.sDigEncoder.ucEncoderMode)) {
-				if (dp_clock == 270000)
-					args.v1.sDigEncoder.ucConfig |= ATOM_ENCODER_CONFIG_DPLINKRATE_2_70GHZ;
-				args.v1.sDigEncoder.ucLaneNum = dp_lane_count;
-			} else if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock))
-				args.v1.sDigEncoder.ucLaneNum = 8;
-			else
-				args.v1.sDigEncoder.ucLaneNum = 4;
-			break;
-		case 3:
-			args.v3.sExtEncoder.ucAction = action;
-			if (action == EXTERNAL_ENCODER_ACTION_V3_ENCODER_INIT)
-				args.v3.sExtEncoder.usConnectorId = cpu_to_le16(connector_object_id);
-			else
-				args.v3.sExtEncoder.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
-			args.v3.sExtEncoder.ucEncoderMode = atombios_get_encoder_mode(encoder);
-
-			if (ENCODER_MODE_IS_DP(args.v3.sExtEncoder.ucEncoderMode)) {
-				if (dp_clock == 270000)
-					args.v3.sExtEncoder.ucConfig |= EXTERNAL_ENCODER_CONFIG_V3_DPLINKRATE_2_70GHZ;
-				else if (dp_clock == 540000)
-					args.v3.sExtEncoder.ucConfig |= EXTERNAL_ENCODER_CONFIG_V3_DPLINKRATE_5_40GHZ;
-				args.v3.sExtEncoder.ucLaneNum = dp_lane_count;
-			} else if (radeon_dig_monitor_is_duallink(encoder, radeon_encoder->pixel_clock))
-				args.v3.sExtEncoder.ucLaneNum = 8;
-			else
-				args.v3.sExtEncoder.ucLaneNum = 4;
-			switch (ext_enum) {
-			case GRAPH_OBJECT_ENUM_ID1:
-				args.v3.sExtEncoder.ucConfig |= EXTERNAL_ENCODER_CONFIG_V3_ENCODER1;
-				break;
-			case GRAPH_OBJECT_ENUM_ID2:
-				args.v3.sExtEncoder.ucConfig |= EXTERNAL_ENCODER_CONFIG_V3_ENCODER2;
-				break;
-			case GRAPH_OBJECT_ENUM_ID3:
-				args.v3.sExtEncoder.ucConfig |= EXTERNAL_ENCODER_CONFIG_V3_ENCODER3;
-				break;
-			}
-			switch (bpc) {
-			case 0:
-				args.v3.sExtEncoder.ucBitPerColor = PANEL_BPC_UNDEFINE;
-				break;
-			case 6:
-				args.v3.sExtEncoder.ucBitPerColor = PANEL_6BIT_PER_COLOR;
-				break;
-			case 8:
-			default:
-				args.v3.sExtEncoder.ucBitPerColor = PANEL_8BIT_PER_COLOR;
-				break;
-			case 10:
-				args.v3.sExtEncoder.ucBitPerColor = PANEL_10BIT_PER_COLOR;
-				break;
-			case 12:
-				args.v3.sExtEncoder.ucBitPerColor = PANEL_12BIT_PER_COLOR;
-				break;
-			case 16:
-				args.v3.sExtEncoder.ucBitPerColor = PANEL_16BIT_PER_COLOR;
-				break;
-			}
-			break;
-		default:
-			DRM_ERROR("Unknown table version: %d, %d\n", frev, crev);
-			return;
-		}
-		break;
-	default:
-		DRM_ERROR("Unknown table version: %d, %d\n", frev, crev);
-		return;
-	}
-	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
-}
-
-static void
-atombios_yuv_setup(struct drm_encoder *encoder, bool enable)
-{
-	struct drm_device *dev = encoder->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc);
-	ENABLE_YUV_PS_ALLOCATION args;
-	int index = GetIndexIntoMasterTable(COMMAND, EnableYUV);
-	uint32_t temp, reg;
-
-	memset(&args, 0, sizeof(args));
-
-	if (rdev->family >= CHIP_R600)
-		reg = R600_BIOS_3_SCRATCH;
-	else
-		reg = RADEON_BIOS_3_SCRATCH;
-
-	/* XXX: fix up scratch reg handling */
-	temp = RREG32(reg);
-	if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT))
-		WREG32(reg, (ATOM_S3_TV1_ACTIVE |
-			     (radeon_crtc->crtc_id << 18)));
-	else if (radeon_encoder->active_device & (ATOM_DEVICE_CV_SUPPORT))
-		WREG32(reg, (ATOM_S3_CV_ACTIVE | (radeon_crtc->crtc_id << 24)));
-	else
-		WREG32(reg, 0);
-
-	if (enable)
-		args.ucEnable = ATOM_ENABLE;
-	args.ucCRTC = radeon_crtc->crtc_id;
-
-	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
-
-	WREG32(reg, temp);
-}
-
-static void
-radeon_atom_encoder_dpms_avivo(struct drm_encoder *encoder, int mode)
-{
-	struct drm_device *dev = encoder->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	DISPLAY_DEVICE_OUTPUT_CONTROL_PS_ALLOCATION args;
-	int index = 0;
-
-	memset(&args, 0, sizeof(args));
-
-	switch (radeon_encoder->encoder_id) {
-	case ENCODER_OBJECT_ID_INTERNAL_TMDS1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
-		index = GetIndexIntoMasterTable(COMMAND, TMDSAOutputControl);
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_DVO1:
-	case ENCODER_OBJECT_ID_INTERNAL_DDI:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
-		index = GetIndexIntoMasterTable(COMMAND, DVOOutputControl);
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_LVDS:
-		index = GetIndexIntoMasterTable(COMMAND, LCD1OutputControl);
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
-		if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT))
-			index = GetIndexIntoMasterTable(COMMAND, LCD1OutputControl);
-		else
-			index = GetIndexIntoMasterTable(COMMAND, LVTMAOutputControl);
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_DAC1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1:
-		if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT))
-			index = GetIndexIntoMasterTable(COMMAND, TV1OutputControl);
-		else if (radeon_encoder->active_device & (ATOM_DEVICE_CV_SUPPORT))
-			index = GetIndexIntoMasterTable(COMMAND, CV1OutputControl);
-		else
-			index = GetIndexIntoMasterTable(COMMAND, DAC1OutputControl);
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_DAC2:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC2:
-		if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT))
-			index = GetIndexIntoMasterTable(COMMAND, TV1OutputControl);
-		else if (radeon_encoder->active_device & (ATOM_DEVICE_CV_SUPPORT))
-			index = GetIndexIntoMasterTable(COMMAND, CV1OutputControl);
-		else
-			index = GetIndexIntoMasterTable(COMMAND, DAC2OutputControl);
-		break;
-	default:
-		return;
-	}
-
-	switch (mode) {
-	case DRM_MODE_DPMS_ON:
-		args.ucAction = ATOM_ENABLE;
-		/* workaround for DVOOutputControl on some RS690 systems */
-		if (radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_DDI) {
-			u32 reg = RREG32(RADEON_BIOS_3_SCRATCH);
-			WREG32(RADEON_BIOS_3_SCRATCH, reg & ~ATOM_S3_DFP2I_ACTIVE);
-			atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
-			WREG32(RADEON_BIOS_3_SCRATCH, reg);
-		} else
-			atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
-		if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) {
-			args.ucAction = ATOM_LCD_BLON;
-			atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
-		}
-		break;
-	case DRM_MODE_DPMS_STANDBY:
-	case DRM_MODE_DPMS_SUSPEND:
-	case DRM_MODE_DPMS_OFF:
-		args.ucAction = ATOM_DISABLE;
-		atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
-		if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) {
-			args.ucAction = ATOM_LCD_BLOFF;
-			atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
-		}
-		break;
-	}
-}
-
-static void
-radeon_atom_encoder_dpms_dig(struct drm_encoder *encoder, int mode)
-{
-	struct drm_device *dev = encoder->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	struct drm_connector *connector = radeon_get_connector_for_encoder(encoder);
-	struct radeon_connector *radeon_connector = NULL;
-	struct radeon_connector_atom_dig *radeon_dig_connector = NULL;
-
-	if (connector) {
-		radeon_connector = to_radeon_connector(connector);
-		radeon_dig_connector = radeon_connector->con_priv;
-	}
-
-	switch (mode) {
-	case DRM_MODE_DPMS_ON:
-		/* some early dce3.2 boards have a bug in their transmitter control table */
-		if ((rdev->family == CHIP_RV710) || (rdev->family == CHIP_RV730) ||
-		    ASIC_IS_DCE41(rdev) || ASIC_IS_DCE5(rdev))
-			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_ENABLE, 0, 0);
-		else
-			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_ENABLE_OUTPUT, 0, 0);
-		if (ENCODER_MODE_IS_DP(atombios_get_encoder_mode(encoder)) && connector) {
-			if (connector->connector_type == DRM_MODE_CONNECTOR_eDP) {
-				atombios_set_edp_panel_power(connector,
-							     ATOM_TRANSMITTER_ACTION_POWER_ON);
-				radeon_dig_connector->edp_on = true;
-			}
-			radeon_dp_link_train(encoder, connector);
-			if (ASIC_IS_DCE4(rdev))
-				atombios_dig_encoder_setup(encoder, ATOM_ENCODER_CMD_DP_VIDEO_ON, 0);
-		}
-		if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT))
-			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_LCD_BLON, 0, 0);
-		break;
-	case DRM_MODE_DPMS_STANDBY:
-	case DRM_MODE_DPMS_SUSPEND:
-	case DRM_MODE_DPMS_OFF:
-		if (ASIC_IS_DCE41(rdev) || ASIC_IS_DCE5(rdev))
-			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_DISABLE, 0, 0);
-		else
-			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_DISABLE_OUTPUT, 0, 0);
-		if (ENCODER_MODE_IS_DP(atombios_get_encoder_mode(encoder)) && connector) {
-			if (ASIC_IS_DCE4(rdev))
-				atombios_dig_encoder_setup(encoder, ATOM_ENCODER_CMD_DP_VIDEO_OFF, 0);
-			if (connector->connector_type == DRM_MODE_CONNECTOR_eDP) {
-				atombios_set_edp_panel_power(connector,
-							     ATOM_TRANSMITTER_ACTION_POWER_OFF);
-				radeon_dig_connector->edp_on = false;
-			}
-		}
-		if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT))
-			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_LCD_BLOFF, 0, 0);
-		break;
-	}
-}
-
-static void
-radeon_atom_encoder_dpms_ext(struct drm_encoder *encoder,
-			     struct drm_encoder *ext_encoder,
-			     int mode)
-{
-	struct drm_device *dev = encoder->dev;
-	struct radeon_device *rdev = dev->dev_private;
-
-	switch (mode) {
-	case DRM_MODE_DPMS_ON:
-	default:
-		if (ASIC_IS_DCE41(rdev)) {
-			atombios_external_encoder_setup(encoder, ext_encoder,
-							EXTERNAL_ENCODER_ACTION_V3_ENABLE_OUTPUT);
-			atombios_external_encoder_setup(encoder, ext_encoder,
-							EXTERNAL_ENCODER_ACTION_V3_ENCODER_BLANKING_OFF);
-		} else
-			atombios_external_encoder_setup(encoder, ext_encoder, ATOM_ENABLE);
-		break;
-	case DRM_MODE_DPMS_STANDBY:
-	case DRM_MODE_DPMS_SUSPEND:
-	case DRM_MODE_DPMS_OFF:
-		if (ASIC_IS_DCE41(rdev)) {
-			atombios_external_encoder_setup(encoder, ext_encoder,
-							EXTERNAL_ENCODER_ACTION_V3_ENCODER_BLANKING);
-			atombios_external_encoder_setup(encoder, ext_encoder,
-							EXTERNAL_ENCODER_ACTION_V3_DISABLE_OUTPUT);
-		} else
-			atombios_external_encoder_setup(encoder, ext_encoder, ATOM_DISABLE);
-		break;
-	}
-}
-
-static void
-radeon_atom_encoder_dpms(struct drm_encoder *encoder, int mode)
-{
-	struct drm_device *dev = encoder->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	struct drm_encoder *ext_encoder = radeon_get_external_encoder(encoder);
-
-	DRM_DEBUG_KMS("encoder dpms %d to mode %d, devices %08x, active_devices %08x\n",
-		  radeon_encoder->encoder_id, mode, radeon_encoder->devices,
-		  radeon_encoder->active_device);
-	switch (radeon_encoder->encoder_id) {
-	case ENCODER_OBJECT_ID_INTERNAL_TMDS1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
-	case ENCODER_OBJECT_ID_INTERNAL_LVDS:
-	case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
-	case ENCODER_OBJECT_ID_INTERNAL_DVO1:
-	case ENCODER_OBJECT_ID_INTERNAL_DDI:
-	case ENCODER_OBJECT_ID_INTERNAL_DAC2:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC2:
-		radeon_atom_encoder_dpms_avivo(encoder, mode);
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
-		radeon_atom_encoder_dpms_dig(encoder, mode);
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
-		if (ASIC_IS_DCE5(rdev)) {
-			switch (mode) {
-			case DRM_MODE_DPMS_ON:
-				atombios_dvo_setup(encoder, ATOM_ENABLE);
-				break;
-			case DRM_MODE_DPMS_STANDBY:
-			case DRM_MODE_DPMS_SUSPEND:
-			case DRM_MODE_DPMS_OFF:
-				atombios_dvo_setup(encoder, ATOM_DISABLE);
-				break;
-			}
-		} else if (ASIC_IS_DCE3(rdev))
-			radeon_atom_encoder_dpms_dig(encoder, mode);
-		else
-			radeon_atom_encoder_dpms_avivo(encoder, mode);
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_DAC1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1:
-		if (ASIC_IS_DCE5(rdev)) {
-			switch (mode) {
-			case DRM_MODE_DPMS_ON:
-				atombios_dac_setup(encoder, ATOM_ENABLE);
-				break;
-			case DRM_MODE_DPMS_STANDBY:
-			case DRM_MODE_DPMS_SUSPEND:
-			case DRM_MODE_DPMS_OFF:
-				atombios_dac_setup(encoder, ATOM_DISABLE);
-				break;
-			}
-		} else
-			radeon_atom_encoder_dpms_avivo(encoder, mode);
-		break;
-	default:
-		return;
-	}
-
-	if (ext_encoder)
-		radeon_atom_encoder_dpms_ext(encoder, ext_encoder, mode);
-
-	radeon_atombios_encoder_dpms_scratch_regs(encoder, (mode == DRM_MODE_DPMS_ON) ? true : false);
-
-}
-
-union crtc_source_param {
-	SELECT_CRTC_SOURCE_PS_ALLOCATION v1;
-	SELECT_CRTC_SOURCE_PARAMETERS_V2 v2;
-};
-
-static void
-atombios_set_encoder_crtc_source(struct drm_encoder *encoder)
-{
-	struct drm_device *dev = encoder->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc);
-	union crtc_source_param args;
-	int index = GetIndexIntoMasterTable(COMMAND, SelectCRTC_Source);
-	uint8_t frev, crev;
-	struct radeon_encoder_atom_dig *dig;
-
-	memset(&args, 0, sizeof(args));
-
-	if (!atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev))
-		return;
-
-	switch (frev) {
-	case 1:
-		switch (crev) {
-		case 1:
-		default:
-			if (ASIC_IS_AVIVO(rdev))
-				args.v1.ucCRTC = radeon_crtc->crtc_id;
-			else {
-				if (radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_DAC1) {
-					args.v1.ucCRTC = radeon_crtc->crtc_id;
-				} else {
-					args.v1.ucCRTC = radeon_crtc->crtc_id << 2;
-				}
-			}
-			switch (radeon_encoder->encoder_id) {
-			case ENCODER_OBJECT_ID_INTERNAL_TMDS1:
-			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
-				args.v1.ucDevice = ATOM_DEVICE_DFP1_INDEX;
-				break;
-			case ENCODER_OBJECT_ID_INTERNAL_LVDS:
-			case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
-				if (radeon_encoder->devices & ATOM_DEVICE_LCD1_SUPPORT)
-					args.v1.ucDevice = ATOM_DEVICE_LCD1_INDEX;
-				else
-					args.v1.ucDevice = ATOM_DEVICE_DFP3_INDEX;
-				break;
-			case ENCODER_OBJECT_ID_INTERNAL_DVO1:
-			case ENCODER_OBJECT_ID_INTERNAL_DDI:
-			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
-				args.v1.ucDevice = ATOM_DEVICE_DFP2_INDEX;
-				break;
-			case ENCODER_OBJECT_ID_INTERNAL_DAC1:
-			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1:
-				if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT))
-					args.v1.ucDevice = ATOM_DEVICE_TV1_INDEX;
-				else if (radeon_encoder->active_device & (ATOM_DEVICE_CV_SUPPORT))
-					args.v1.ucDevice = ATOM_DEVICE_CV_INDEX;
-				else
-					args.v1.ucDevice = ATOM_DEVICE_CRT1_INDEX;
-				break;
-			case ENCODER_OBJECT_ID_INTERNAL_DAC2:
-			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC2:
-				if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT))
-					args.v1.ucDevice = ATOM_DEVICE_TV1_INDEX;
-				else if (radeon_encoder->active_device & (ATOM_DEVICE_CV_SUPPORT))
-					args.v1.ucDevice = ATOM_DEVICE_CV_INDEX;
-				else
-					args.v1.ucDevice = ATOM_DEVICE_CRT2_INDEX;
-				break;
-			}
-			break;
-		case 2:
-			args.v2.ucCRTC = radeon_crtc->crtc_id;
-			if (radeon_encoder_get_dp_bridge_encoder_id(encoder) != ENCODER_OBJECT_ID_NONE) {
-				struct drm_connector *connector = radeon_get_connector_for_encoder(encoder);
-
-				if (connector->connector_type == DRM_MODE_CONNECTOR_LVDS)
-					args.v2.ucEncodeMode = ATOM_ENCODER_MODE_LVDS;
-				else if (connector->connector_type == DRM_MODE_CONNECTOR_VGA)
-					args.v2.ucEncodeMode = ATOM_ENCODER_MODE_CRT;
-				else
-					args.v2.ucEncodeMode = atombios_get_encoder_mode(encoder);
-			} else
-				args.v2.ucEncodeMode = atombios_get_encoder_mode(encoder);
-			switch (radeon_encoder->encoder_id) {
-			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
-			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
-			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
-			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
-				dig = radeon_encoder->enc_priv;
-				switch (dig->dig_encoder) {
-				case 0:
-					args.v2.ucEncoderID = ASIC_INT_DIG1_ENCODER_ID;
-					break;
-				case 1:
-					args.v2.ucEncoderID = ASIC_INT_DIG2_ENCODER_ID;
-					break;
-				case 2:
-					args.v2.ucEncoderID = ASIC_INT_DIG3_ENCODER_ID;
-					break;
-				case 3:
-					args.v2.ucEncoderID = ASIC_INT_DIG4_ENCODER_ID;
-					break;
-				case 4:
-					args.v2.ucEncoderID = ASIC_INT_DIG5_ENCODER_ID;
-					break;
-				case 5:
-					args.v2.ucEncoderID = ASIC_INT_DIG6_ENCODER_ID;
-					break;
-				}
-				break;
-			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
-				args.v2.ucEncoderID = ASIC_INT_DVO_ENCODER_ID;
-				break;
-			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1:
-				if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT))
-					args.v2.ucEncoderID = ASIC_INT_TV_ENCODER_ID;
-				else if (radeon_encoder->active_device & (ATOM_DEVICE_CV_SUPPORT))
-					args.v2.ucEncoderID = ASIC_INT_TV_ENCODER_ID;
-				else
-					args.v2.ucEncoderID = ASIC_INT_DAC1_ENCODER_ID;
-				break;
-			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC2:
-				if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT))
-					args.v2.ucEncoderID = ASIC_INT_TV_ENCODER_ID;
-				else if (radeon_encoder->active_device & (ATOM_DEVICE_CV_SUPPORT))
-					args.v2.ucEncoderID = ASIC_INT_TV_ENCODER_ID;
-				else
-					args.v2.ucEncoderID = ASIC_INT_DAC2_ENCODER_ID;
-				break;
-			}
-			break;
-		}
-		break;
-	default:
-		DRM_ERROR("Unknown table version: %d, %d\n", frev, crev);
-		return;
-	}
-
-	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
-
-	/* update scratch regs with new routing */
-	radeon_atombios_encoder_crtc_scratch_regs(encoder, radeon_crtc->crtc_id);
-}
-
-static void
-atombios_apply_encoder_quirks(struct drm_encoder *encoder,
-			      struct drm_display_mode *mode)
-{
-	struct drm_device *dev = encoder->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc);
-
-	/* Funky macbooks */
-	if ((dev->pdev->device == 0x71C5) &&
-	    (dev->pdev->subsystem_vendor == 0x106b) &&
-	    (dev->pdev->subsystem_device == 0x0080)) {
-		if (radeon_encoder->devices & ATOM_DEVICE_LCD1_SUPPORT) {
-			uint32_t lvtma_bit_depth_control = RREG32(AVIVO_LVTMA_BIT_DEPTH_CONTROL);
-
-			lvtma_bit_depth_control &= ~AVIVO_LVTMA_BIT_DEPTH_CONTROL_TRUNCATE_EN;
-			lvtma_bit_depth_control &= ~AVIVO_LVTMA_BIT_DEPTH_CONTROL_SPATIAL_DITHER_EN;
-
-			WREG32(AVIVO_LVTMA_BIT_DEPTH_CONTROL, lvtma_bit_depth_control);
-		}
-	}
-
-	/* set scaler clears this on some chips */
-	if (ASIC_IS_AVIVO(rdev) &&
-	    (!(radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT)))) {
-		if (ASIC_IS_DCE4(rdev)) {
-			if (mode->flags & DRM_MODE_FLAG_INTERLACE)
-				WREG32(EVERGREEN_DATA_FORMAT + radeon_crtc->crtc_offset,
-				       EVERGREEN_INTERLEAVE_EN);
-			else
-				WREG32(EVERGREEN_DATA_FORMAT + radeon_crtc->crtc_offset, 0);
-		} else {
-			if (mode->flags & DRM_MODE_FLAG_INTERLACE)
-				WREG32(AVIVO_D1MODE_DATA_FORMAT + radeon_crtc->crtc_offset,
-				       AVIVO_D1MODE_INTERLEAVE_EN);
-			else
-				WREG32(AVIVO_D1MODE_DATA_FORMAT + radeon_crtc->crtc_offset, 0);
-		}
-	}
-}
-
-static int radeon_atom_pick_dig_encoder(struct drm_encoder *encoder)
-{
-	struct drm_device *dev = encoder->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc);
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	struct drm_encoder *test_encoder;
-	struct radeon_encoder_atom_dig *dig;
-	uint32_t dig_enc_in_use = 0;
-
-	/* DCE4/5 */
-	if (ASIC_IS_DCE4(rdev)) {
-		dig = radeon_encoder->enc_priv;
-		if (ASIC_IS_DCE41(rdev)) {
-			/* ontario follows DCE4 */
-			if (rdev->family == CHIP_PALM) {
-				if (dig->linkb)
-					return 1;
-				else
-					return 0;
-			} else
-				/* llano follows DCE3.2 */
-				return radeon_crtc->crtc_id;
-		} else {
-			switch (radeon_encoder->encoder_id) {
-			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
-				if (dig->linkb)
-					return 1;
-				else
-					return 0;
-				break;
-			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
-				if (dig->linkb)
-					return 3;
-				else
-					return 2;
-				break;
-			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
-				if (dig->linkb)
-					return 5;
-				else
-					return 4;
-				break;
-			}
-		}
-	}
-
-	/* on DCE32 and encoder can driver any block so just crtc id */
-	if (ASIC_IS_DCE32(rdev)) {
-		return radeon_crtc->crtc_id;
-	}
-
-	/* on DCE3 - LVTMA can only be driven by DIGB */
-	list_for_each_entry(test_encoder, &dev->mode_config.encoder_list, head) {
-		struct radeon_encoder *radeon_test_encoder;
-
-		if (encoder == test_encoder)
-			continue;
-
-		if (!radeon_encoder_is_digital(test_encoder))
-			continue;
-
-		radeon_test_encoder = to_radeon_encoder(test_encoder);
-		dig = radeon_test_encoder->enc_priv;
-
-		if (dig->dig_encoder >= 0)
-			dig_enc_in_use |= (1 << dig->dig_encoder);
-	}
-
-	if (radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA) {
-		if (dig_enc_in_use & 0x2)
-			DRM_ERROR("LVDS required digital encoder 2 but it was in use - stealing\n");
-		return 1;
-	}
-	if (!(dig_enc_in_use & 1))
-		return 0;
-	return 1;
-}
-
-/* This only needs to be called once at startup */
-void
-radeon_atom_encoder_init(struct radeon_device *rdev)
-{
-	struct drm_device *dev = rdev->ddev;
-	struct drm_encoder *encoder;
-
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
-		struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-		struct drm_encoder *ext_encoder = radeon_get_external_encoder(encoder);
-
-		switch (radeon_encoder->encoder_id) {
-		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
-		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
-		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
-		case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
-			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_INIT, 0, 0);
-			break;
-		default:
-			break;
-		}
-
-		if (ext_encoder && ASIC_IS_DCE41(rdev))
-			atombios_external_encoder_setup(encoder, ext_encoder,
-							EXTERNAL_ENCODER_ACTION_V3_ENCODER_INIT);
-	}
-}
-
-static void
-radeon_atom_encoder_mode_set(struct drm_encoder *encoder,
-			     struct drm_display_mode *mode,
-			     struct drm_display_mode *adjusted_mode)
-{
-	struct drm_device *dev = encoder->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	struct drm_encoder *ext_encoder = radeon_get_external_encoder(encoder);
-
-	radeon_encoder->pixel_clock = adjusted_mode->clock;
-
-	if (ASIC_IS_AVIVO(rdev) && !ASIC_IS_DCE4(rdev)) {
-		if (radeon_encoder->active_device & (ATOM_DEVICE_CV_SUPPORT | ATOM_DEVICE_TV_SUPPORT))
-			atombios_yuv_setup(encoder, true);
-		else
-			atombios_yuv_setup(encoder, false);
-	}
-
-	switch (radeon_encoder->encoder_id) {
-	case ENCODER_OBJECT_ID_INTERNAL_TMDS1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
-	case ENCODER_OBJECT_ID_INTERNAL_LVDS:
-	case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
-		atombios_digital_setup(encoder, PANEL_ENCODER_ACTION_ENABLE);
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
-		if (ASIC_IS_DCE41(rdev) || ASIC_IS_DCE5(rdev)) {
-			struct drm_connector *connector = radeon_get_connector_for_encoder(encoder);
-			struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
-
-			if (!connector)
-				dig->panel_mode = DP_PANEL_MODE_EXTERNAL_DP_MODE;
-			else
-				dig->panel_mode = radeon_dp_get_panel_mode(encoder, connector);
-
-			/* setup and enable the encoder */
-			atombios_dig_encoder_setup(encoder, ATOM_ENCODER_CMD_SETUP, 0);
-			atombios_dig_encoder_setup(encoder,
-						   ATOM_ENCODER_CMD_SETUP_PANEL_MODE,
-						   dig->panel_mode);
-		} else if (ASIC_IS_DCE4(rdev)) {
-			/* disable the transmitter */
-			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_DISABLE, 0, 0);
-			/* setup and enable the encoder */
-			atombios_dig_encoder_setup(encoder, ATOM_ENCODER_CMD_SETUP, 0);
-
-			/* enable the transmitter */
-			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_ENABLE, 0, 0);
-		} else {
-			/* disable the encoder and transmitter */
-			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_DISABLE, 0, 0);
-			atombios_dig_encoder_setup(encoder, ATOM_DISABLE, 0);
-
-			/* setup and enable the encoder and transmitter */
-			atombios_dig_encoder_setup(encoder, ATOM_ENABLE, 0);
-			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_SETUP, 0, 0);
-			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_ENABLE, 0, 0);
-		}
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_DDI:
-	case ENCODER_OBJECT_ID_INTERNAL_DVO1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
-		atombios_dvo_setup(encoder, ATOM_ENABLE);
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_DAC1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1:
-	case ENCODER_OBJECT_ID_INTERNAL_DAC2:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC2:
-		atombios_dac_setup(encoder, ATOM_ENABLE);
-		if (radeon_encoder->devices & (ATOM_DEVICE_TV_SUPPORT | ATOM_DEVICE_CV_SUPPORT)) {
-			if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT | ATOM_DEVICE_CV_SUPPORT))
-				atombios_tv_setup(encoder, ATOM_ENABLE);
-			else
-				atombios_tv_setup(encoder, ATOM_DISABLE);
-		}
-		break;
-	}
-
-	if (ext_encoder) {
-		if (ASIC_IS_DCE41(rdev))
-			atombios_external_encoder_setup(encoder, ext_encoder,
-							EXTERNAL_ENCODER_ACTION_V3_ENCODER_SETUP);
-		else
-			atombios_external_encoder_setup(encoder, ext_encoder, ATOM_ENABLE);
-	}
-
-	atombios_apply_encoder_quirks(encoder, adjusted_mode);
-
-	if (atombios_get_encoder_mode(encoder) == ATOM_ENCODER_MODE_HDMI) {
-		r600_hdmi_enable(encoder);
-		r600_hdmi_setmode(encoder, adjusted_mode);
-	}
-}
-
-static bool
-atombios_dac_load_detect(struct drm_encoder *encoder, struct drm_connector *connector)
-{
-	struct drm_device *dev = encoder->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
-
-	if (radeon_encoder->devices & (ATOM_DEVICE_TV_SUPPORT |
-				       ATOM_DEVICE_CV_SUPPORT |
-				       ATOM_DEVICE_CRT_SUPPORT)) {
-		DAC_LOAD_DETECTION_PS_ALLOCATION args;
-		int index = GetIndexIntoMasterTable(COMMAND, DAC_LoadDetection);
-		uint8_t frev, crev;
-
-		memset(&args, 0, sizeof(args));
-
-		if (!atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev))
-			return false;
-
-		args.sDacload.ucMisc = 0;
-
-		if ((radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_DAC1) ||
-		    (radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1))
-			args.sDacload.ucDacType = ATOM_DAC_A;
-		else
-			args.sDacload.ucDacType = ATOM_DAC_B;
-
-		if (radeon_connector->devices & ATOM_DEVICE_CRT1_SUPPORT)
-			args.sDacload.usDeviceID = cpu_to_le16(ATOM_DEVICE_CRT1_SUPPORT);
-		else if (radeon_connector->devices & ATOM_DEVICE_CRT2_SUPPORT)
-			args.sDacload.usDeviceID = cpu_to_le16(ATOM_DEVICE_CRT2_SUPPORT);
-		else if (radeon_connector->devices & ATOM_DEVICE_CV_SUPPORT) {
-			args.sDacload.usDeviceID = cpu_to_le16(ATOM_DEVICE_CV_SUPPORT);
-			if (crev >= 3)
-				args.sDacload.ucMisc = DAC_LOAD_MISC_YPrPb;
-		} else if (radeon_connector->devices & ATOM_DEVICE_TV1_SUPPORT) {
-			args.sDacload.usDeviceID = cpu_to_le16(ATOM_DEVICE_TV1_SUPPORT);
-			if (crev >= 3)
-				args.sDacload.ucMisc = DAC_LOAD_MISC_YPrPb;
-		}
-
-		atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
-
-		return true;
-	} else
-		return false;
-}
-
-static enum drm_connector_status
-radeon_atom_dac_detect(struct drm_encoder *encoder, struct drm_connector *connector)
-{
-	struct drm_device *dev = encoder->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
-	uint32_t bios_0_scratch;
-
-	if (!atombios_dac_load_detect(encoder, connector)) {
-		DRM_DEBUG_KMS("detect returned false \n");
-		return connector_status_unknown;
-	}
-
-	if (rdev->family >= CHIP_R600)
-		bios_0_scratch = RREG32(R600_BIOS_0_SCRATCH);
-	else
-		bios_0_scratch = RREG32(RADEON_BIOS_0_SCRATCH);
-
-	DRM_DEBUG_KMS("Bios 0 scratch %x %08x\n", bios_0_scratch, radeon_encoder->devices);
-	if (radeon_connector->devices & ATOM_DEVICE_CRT1_SUPPORT) {
-		if (bios_0_scratch & ATOM_S0_CRT1_MASK)
-			return connector_status_connected;
-	}
-	if (radeon_connector->devices & ATOM_DEVICE_CRT2_SUPPORT) {
-		if (bios_0_scratch & ATOM_S0_CRT2_MASK)
-			return connector_status_connected;
-	}
-	if (radeon_connector->devices & ATOM_DEVICE_CV_SUPPORT) {
-		if (bios_0_scratch & (ATOM_S0_CV_MASK|ATOM_S0_CV_MASK_A))
-			return connector_status_connected;
-	}
-	if (radeon_connector->devices & ATOM_DEVICE_TV1_SUPPORT) {
-		if (bios_0_scratch & (ATOM_S0_TV1_COMPOSITE | ATOM_S0_TV1_COMPOSITE_A))
-			return connector_status_connected; /* CTV */
-		else if (bios_0_scratch & (ATOM_S0_TV1_SVIDEO | ATOM_S0_TV1_SVIDEO_A))
-			return connector_status_connected; /* STV */
-	}
-	return connector_status_disconnected;
-}
-
-static enum drm_connector_status
-radeon_atom_dig_detect(struct drm_encoder *encoder, struct drm_connector *connector)
-{
-	struct drm_device *dev = encoder->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
-	struct drm_encoder *ext_encoder = radeon_get_external_encoder(encoder);
-	u32 bios_0_scratch;
-
-	if (!ASIC_IS_DCE4(rdev))
-		return connector_status_unknown;
-
-	if (!ext_encoder)
-		return connector_status_unknown;
-
-	if ((radeon_connector->devices & ATOM_DEVICE_CRT_SUPPORT) == 0)
-		return connector_status_unknown;
-
-	/* load detect on the dp bridge */
-	atombios_external_encoder_setup(encoder, ext_encoder,
-					EXTERNAL_ENCODER_ACTION_V3_DACLOAD_DETECTION);
-
-	bios_0_scratch = RREG32(R600_BIOS_0_SCRATCH);
-
-	DRM_DEBUG_KMS("Bios 0 scratch %x %08x\n", bios_0_scratch, radeon_encoder->devices);
-	if (radeon_connector->devices & ATOM_DEVICE_CRT1_SUPPORT) {
-		if (bios_0_scratch & ATOM_S0_CRT1_MASK)
-			return connector_status_connected;
-	}
-	if (radeon_connector->devices & ATOM_DEVICE_CRT2_SUPPORT) {
-		if (bios_0_scratch & ATOM_S0_CRT2_MASK)
-			return connector_status_connected;
-	}
-	if (radeon_connector->devices & ATOM_DEVICE_CV_SUPPORT) {
-		if (bios_0_scratch & (ATOM_S0_CV_MASK|ATOM_S0_CV_MASK_A))
-			return connector_status_connected;
-	}
-	if (radeon_connector->devices & ATOM_DEVICE_TV1_SUPPORT) {
-		if (bios_0_scratch & (ATOM_S0_TV1_COMPOSITE | ATOM_S0_TV1_COMPOSITE_A))
-			return connector_status_connected; /* CTV */
-		else if (bios_0_scratch & (ATOM_S0_TV1_SVIDEO | ATOM_S0_TV1_SVIDEO_A))
-			return connector_status_connected; /* STV */
-	}
-	return connector_status_disconnected;
-}
-
-void
-radeon_atom_ext_encoder_setup_ddc(struct drm_encoder *encoder)
-{
-	struct drm_encoder *ext_encoder = radeon_get_external_encoder(encoder);
-
-	if (ext_encoder)
-		/* ddc_setup on the dp bridge */
-		atombios_external_encoder_setup(encoder, ext_encoder,
-						EXTERNAL_ENCODER_ACTION_V3_DDC_SETUP);
-
-}
-
-static void radeon_atom_encoder_prepare(struct drm_encoder *encoder)
-{
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	struct drm_connector *connector = radeon_get_connector_for_encoder(encoder);
-
-	if ((radeon_encoder->active_device &
-	     (ATOM_DEVICE_DFP_SUPPORT | ATOM_DEVICE_LCD_SUPPORT)) ||
-	    (radeon_encoder_get_dp_bridge_encoder_id(encoder) !=
-	     ENCODER_OBJECT_ID_NONE)) {
-		struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
-		if (dig)
-			dig->dig_encoder = radeon_atom_pick_dig_encoder(encoder);
-	}
-
-	radeon_atom_output_lock(encoder, true);
-	radeon_atom_encoder_dpms(encoder, DRM_MODE_DPMS_OFF);
-
-	if (connector) {
-		struct radeon_connector *radeon_connector = to_radeon_connector(connector);
-
-		/* select the clock/data port if it uses a router */
-		if (radeon_connector->router.cd_valid)
-			radeon_router_select_cd_port(radeon_connector);
-
-		/* turn eDP panel on for mode set */
-		if (connector->connector_type == DRM_MODE_CONNECTOR_eDP)
-			atombios_set_edp_panel_power(connector,
-						     ATOM_TRANSMITTER_ACTION_POWER_ON);
-	}
-
-	/* this is needed for the pll/ss setup to work correctly in some cases */
-	atombios_set_encoder_crtc_source(encoder);
-}
-
-static void radeon_atom_encoder_commit(struct drm_encoder *encoder)
-{
-	radeon_atom_encoder_dpms(encoder, DRM_MODE_DPMS_ON);
-	radeon_atom_output_lock(encoder, false);
-}
-
-static void radeon_atom_encoder_disable(struct drm_encoder *encoder)
-{
-	struct drm_device *dev = encoder->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	struct radeon_encoder_atom_dig *dig;
-
-	/* check for pre-DCE3 cards with shared encoders;
-	 * can't really use the links individually, so don't disable
-	 * the encoder if it's in use by another connector
-	 */
-	if (!ASIC_IS_DCE3(rdev)) {
-		struct drm_encoder *other_encoder;
-		struct radeon_encoder *other_radeon_encoder;
-
-		list_for_each_entry(other_encoder, &dev->mode_config.encoder_list, head) {
-			other_radeon_encoder = to_radeon_encoder(other_encoder);
-			if ((radeon_encoder->encoder_id == other_radeon_encoder->encoder_id) &&
-			    drm_helper_encoder_in_use(other_encoder))
-				goto disable_done;
-		}
-	}
-
-	radeon_atom_encoder_dpms(encoder, DRM_MODE_DPMS_OFF);
-
-	switch (radeon_encoder->encoder_id) {
-	case ENCODER_OBJECT_ID_INTERNAL_TMDS1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
-	case ENCODER_OBJECT_ID_INTERNAL_LVDS:
-	case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
-		atombios_digital_setup(encoder, PANEL_ENCODER_ACTION_DISABLE);
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
-		if (ASIC_IS_DCE4(rdev))
-			/* disable the transmitter */
-			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_DISABLE, 0, 0);
-		else {
-			/* disable the encoder and transmitter */
-			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_DISABLE, 0, 0);
-			atombios_dig_encoder_setup(encoder, ATOM_DISABLE, 0);
-		}
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_DDI:
-	case ENCODER_OBJECT_ID_INTERNAL_DVO1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
-		atombios_dvo_setup(encoder, ATOM_DISABLE);
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_DAC1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1:
-	case ENCODER_OBJECT_ID_INTERNAL_DAC2:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC2:
-		atombios_dac_setup(encoder, ATOM_DISABLE);
-		if (radeon_encoder->devices & (ATOM_DEVICE_TV_SUPPORT | ATOM_DEVICE_CV_SUPPORT))
-			atombios_tv_setup(encoder, ATOM_DISABLE);
-		break;
-	}
-
-disable_done:
-	if (radeon_encoder_is_digital(encoder)) {
-		if (atombios_get_encoder_mode(encoder) == ATOM_ENCODER_MODE_HDMI)
-			r600_hdmi_disable(encoder);
-		dig = radeon_encoder->enc_priv;
-		dig->dig_encoder = -1;
-	}
-	radeon_encoder->active_device = 0;
-}
-
-/* these are handled by the primary encoders */
-static void radeon_atom_ext_prepare(struct drm_encoder *encoder)
-{
-
-}
-
-static void radeon_atom_ext_commit(struct drm_encoder *encoder)
-{
-
-}
-
-static void
-radeon_atom_ext_mode_set(struct drm_encoder *encoder,
-			 struct drm_display_mode *mode,
-			 struct drm_display_mode *adjusted_mode)
-{
-
-}
-
-static void radeon_atom_ext_disable(struct drm_encoder *encoder)
-{
-
-}
-
-static void
-radeon_atom_ext_dpms(struct drm_encoder *encoder, int mode)
-{
-
-}
-
-static bool radeon_atom_ext_mode_fixup(struct drm_encoder *encoder,
-				       struct drm_display_mode *mode,
-				       struct drm_display_mode *adjusted_mode)
-{
-	return true;
-}
-
-static const struct drm_encoder_helper_funcs radeon_atom_ext_helper_funcs = {
-	.dpms = radeon_atom_ext_dpms,
-	.mode_fixup = radeon_atom_ext_mode_fixup,
-	.prepare = radeon_atom_ext_prepare,
-	.mode_set = radeon_atom_ext_mode_set,
-	.commit = radeon_atom_ext_commit,
-	.disable = radeon_atom_ext_disable,
-	/* no detect for TMDS/LVDS yet */
-};
-
-static const struct drm_encoder_helper_funcs radeon_atom_dig_helper_funcs = {
-	.dpms = radeon_atom_encoder_dpms,
-	.mode_fixup = radeon_atom_mode_fixup,
-	.prepare = radeon_atom_encoder_prepare,
-	.mode_set = radeon_atom_encoder_mode_set,
-	.commit = radeon_atom_encoder_commit,
-	.disable = radeon_atom_encoder_disable,
-	.detect = radeon_atom_dig_detect,
-};
-
-static const struct drm_encoder_helper_funcs radeon_atom_dac_helper_funcs = {
-	.dpms = radeon_atom_encoder_dpms,
-	.mode_fixup = radeon_atom_mode_fixup,
-	.prepare = radeon_atom_encoder_prepare,
-	.mode_set = radeon_atom_encoder_mode_set,
-	.commit = radeon_atom_encoder_commit,
-	.detect = radeon_atom_dac_detect,
-};
-
-void radeon_enc_destroy(struct drm_encoder *encoder)
-{
-	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
-	kfree(radeon_encoder->enc_priv);
-	drm_encoder_cleanup(encoder);
-	kfree(radeon_encoder);
-}
-
-static const struct drm_encoder_funcs radeon_atom_enc_funcs = {
-	.destroy = radeon_enc_destroy,
-};
-
-struct radeon_encoder_atom_dac *
-radeon_atombios_set_dac_info(struct radeon_encoder *radeon_encoder)
-{
-	struct drm_device *dev = radeon_encoder->base.dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_encoder_atom_dac *dac = kzalloc(sizeof(struct radeon_encoder_atom_dac), GFP_KERNEL);
-
-	if (!dac)
-		return NULL;
-
-	dac->tv_std = radeon_atombios_get_tv_info(rdev);
-	return dac;
-}
-
-struct radeon_encoder_atom_dig *
-radeon_atombios_set_dig_info(struct radeon_encoder *radeon_encoder)
-{
-	int encoder_enum = (radeon_encoder->encoder_enum & ENUM_ID_MASK) >> ENUM_ID_SHIFT;
-	struct radeon_encoder_atom_dig *dig = kzalloc(sizeof(struct radeon_encoder_atom_dig), GFP_KERNEL);
-
-	if (!dig)
-		return NULL;
-
-	/* coherent mode by default */
-	dig->coherent_mode = true;
-	dig->dig_encoder = -1;
-
-	if (encoder_enum == 2)
-		dig->linkb = true;
-	else
-		dig->linkb = false;
-
-	return dig;
-}
-
-void
-radeon_add_atom_encoder(struct drm_device *dev,
-			uint32_t encoder_enum,
-			uint32_t supported_device,
-			u16 caps)
-{
-	struct radeon_device *rdev = dev->dev_private;
-	struct drm_encoder *encoder;
-	struct radeon_encoder *radeon_encoder;
-
-	/* see if we already added it */
-	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
-		radeon_encoder = to_radeon_encoder(encoder);
-		if (radeon_encoder->encoder_enum == encoder_enum) {
-			radeon_encoder->devices |= supported_device;
-			return;
-		}
-
-	}
-
-	/* add a new one */
-	radeon_encoder = kzalloc(sizeof(struct radeon_encoder), GFP_KERNEL);
-	if (!radeon_encoder)
-		return;
-
-	encoder = &radeon_encoder->base;
-	switch (rdev->num_crtc) {
-	case 1:
-		encoder->possible_crtcs = 0x1;
-		break;
-	case 2:
-	default:
-		encoder->possible_crtcs = 0x3;
-		break;
-	case 4:
-		encoder->possible_crtcs = 0xf;
-		break;
-	case 6:
-		encoder->possible_crtcs = 0x3f;
-		break;
-	}
-
-	radeon_encoder->enc_priv = NULL;
-
-	radeon_encoder->encoder_enum = encoder_enum;
-	radeon_encoder->encoder_id = (encoder_enum & OBJECT_ID_MASK) >> OBJECT_ID_SHIFT;
-	radeon_encoder->devices = supported_device;
-	radeon_encoder->rmx_type = RMX_OFF;
-	radeon_encoder->underscan_type = UNDERSCAN_OFF;
-	radeon_encoder->is_ext_encoder = false;
-	radeon_encoder->caps = caps;
-
-	switch (radeon_encoder->encoder_id) {
-	case ENCODER_OBJECT_ID_INTERNAL_LVDS:
-	case ENCODER_OBJECT_ID_INTERNAL_TMDS1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
-	case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
-		if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) {
-			radeon_encoder->rmx_type = RMX_FULL;
-			drm_encoder_init(dev, encoder, &radeon_atom_enc_funcs, DRM_MODE_ENCODER_LVDS);
-			radeon_encoder->enc_priv = radeon_atombios_get_lvds_info(radeon_encoder);
-		} else {
-			drm_encoder_init(dev, encoder, &radeon_atom_enc_funcs, DRM_MODE_ENCODER_TMDS);
-			radeon_encoder->enc_priv = radeon_atombios_set_dig_info(radeon_encoder);
-		}
-		drm_encoder_helper_add(encoder, &radeon_atom_dig_helper_funcs);
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_DAC1:
-		drm_encoder_init(dev, encoder, &radeon_atom_enc_funcs, DRM_MODE_ENCODER_DAC);
-		radeon_encoder->enc_priv = radeon_atombios_set_dac_info(radeon_encoder);
-		drm_encoder_helper_add(encoder, &radeon_atom_dac_helper_funcs);
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_DAC2:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC2:
-		drm_encoder_init(dev, encoder, &radeon_atom_enc_funcs, DRM_MODE_ENCODER_TVDAC);
-		radeon_encoder->enc_priv = radeon_atombios_set_dac_info(radeon_encoder);
-		drm_encoder_helper_add(encoder, &radeon_atom_dac_helper_funcs);
-		break;
-	case ENCODER_OBJECT_ID_INTERNAL_DVO1:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
-	case ENCODER_OBJECT_ID_INTERNAL_DDI:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
-	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
-	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
-		if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) {
-			radeon_encoder->rmx_type = RMX_FULL;
-			drm_encoder_init(dev, encoder, &radeon_atom_enc_funcs, DRM_MODE_ENCODER_LVDS);
-			radeon_encoder->enc_priv = radeon_atombios_get_lvds_info(radeon_encoder);
-		} else if (radeon_encoder->devices & (ATOM_DEVICE_CRT_SUPPORT)) {
-			drm_encoder_init(dev, encoder, &radeon_atom_enc_funcs, DRM_MODE_ENCODER_DAC);
-			radeon_encoder->enc_priv = radeon_atombios_set_dig_info(radeon_encoder);
-		} else {
-			drm_encoder_init(dev, encoder, &radeon_atom_enc_funcs, DRM_MODE_ENCODER_TMDS);
-			radeon_encoder->enc_priv = radeon_atombios_set_dig_info(radeon_encoder);
-		}
-		drm_encoder_helper_add(encoder, &radeon_atom_dig_helper_funcs);
-		break;
-	case ENCODER_OBJECT_ID_SI170B:
-	case ENCODER_OBJECT_ID_CH7303:
-	case ENCODER_OBJECT_ID_EXTERNAL_SDVOA:
-	case ENCODER_OBJECT_ID_EXTERNAL_SDVOB:
-	case ENCODER_OBJECT_ID_TITFP513:
-	case ENCODER_OBJECT_ID_VT1623:
-	case ENCODER_OBJECT_ID_HDMI_SI1930:
-	case ENCODER_OBJECT_ID_TRAVIS:
-	case ENCODER_OBJECT_ID_NUTMEG:
-		/* these are handled by the primary encoders */
-		radeon_encoder->is_ext_encoder = true;
-		if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT))
-			drm_encoder_init(dev, encoder, &radeon_atom_enc_funcs, DRM_MODE_ENCODER_LVDS);
-		else if (radeon_encoder->devices & (ATOM_DEVICE_CRT_SUPPORT))
-			drm_encoder_init(dev, encoder, &radeon_atom_enc_funcs, DRM_MODE_ENCODER_DAC);
-		else
-			drm_encoder_init(dev, encoder, &radeon_atom_enc_funcs, DRM_MODE_ENCODER_TMDS);
-		drm_encoder_helper_add(encoder, &radeon_atom_ext_helper_funcs);
-		break;
-	}
-}
diff --git a/drivers/gpu/drm/radeon/atombios_i2c.c b/drivers/gpu/drm/radeon/atombios_i2c.c
deleted file mode 100644
index 44d87b6..0000000
--- a/drivers/gpu/drm/radeon/atombios_i2c.c
+++ /dev/null
@@ -1,139 +0,0 @@
-/*
- * Copyright 2011 Advanced Micro Devices, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included in
- * all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
- * OTHER DEALINGS IN THE SOFTWARE.
- *
- * Authors: Alex Deucher
- *
- */
-#include "drmP.h"
-#include "radeon_drm.h"
-#include "radeon.h"
-#include "atom.h"
-
-#define TARGET_HW_I2C_CLOCK 50
-
-/* these are a limitation of ProcessI2cChannelTransaction not the hw */
-#define ATOM_MAX_HW_I2C_WRITE 2
-#define ATOM_MAX_HW_I2C_READ  255
-
-static int radeon_process_i2c_ch(struct radeon_i2c_chan *chan,
-				 u8 slave_addr, u8 flags,
-				 u8 *buf, u8 num)
-{
-	struct drm_device *dev = chan->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	PROCESS_I2C_CHANNEL_TRANSACTION_PS_ALLOCATION args;
-	int index = GetIndexIntoMasterTable(COMMAND, ProcessI2cChannelTransaction);
-	unsigned char *base;
-	u16 out;
-
-	memset(&args, 0, sizeof(args));
-
-	base = (unsigned char *)rdev->mode_info.atom_context->scratch;
-
-	if (flags & HW_I2C_WRITE) {
-		if (num > ATOM_MAX_HW_I2C_WRITE) {
-			DRM_ERROR("hw i2c: tried to write too many bytes (%d vs 2)\n", num);
-			return -EINVAL;
-		}
-		memcpy(&out, buf, num);
-		args.lpI2CDataOut = cpu_to_le16(out);
-	} else {
-		if (num > ATOM_MAX_HW_I2C_READ) {
-			DRM_ERROR("hw i2c: tried to read too many bytes (%d vs 255)\n", num);
-			return -EINVAL;
-		}
-	}
-
-	args.ucI2CSpeed = TARGET_HW_I2C_CLOCK;
-	args.ucRegIndex = 0;
-	args.ucTransBytes = num;
-	args.ucSlaveAddr = slave_addr << 1;
-	args.ucLineNumber = chan->rec.i2c_id;
-
-	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
-
-	/* error */
-	if (args.ucStatus != HW_ASSISTED_I2C_STATUS_SUCCESS) {
-		DRM_DEBUG_KMS("hw_i2c error\n");
-		return -EIO;
-	}
-
-	if (!(flags & HW_I2C_WRITE))
-		memcpy(buf, base, num);
-
-	return 0;
-}
-
-int radeon_atom_hw_i2c_xfer(struct i2c_adapter *i2c_adap,
-			    struct i2c_msg *msgs, int num)
-{
-	struct radeon_i2c_chan *i2c = i2c_get_adapdata(i2c_adap);
-	struct i2c_msg *p;
-	int i, remaining, current_count, buffer_offset, max_bytes, ret;
-	u8 buf = 0, flags;
-
-	/* check for bus probe */
-	p = &msgs[0];
-	if ((num == 1) && (p->len == 0)) {
-		ret = radeon_process_i2c_ch(i2c,
-					    p->addr, HW_I2C_WRITE,
-					    &buf, 1);
-		if (ret)
-			return ret;
-		else
-			return num;
-	}
-
-	for (i = 0; i < num; i++) {
-		p = &msgs[i];
-		remaining = p->len;
-		buffer_offset = 0;
-		/* max_bytes are a limitation of ProcessI2cChannelTransaction not the hw */
-		if (p->flags & I2C_M_RD) {
-			max_bytes = ATOM_MAX_HW_I2C_READ;
-			flags = HW_I2C_READ;
-		} else {
-			max_bytes = ATOM_MAX_HW_I2C_WRITE;
-			flags = HW_I2C_WRITE;
-		}
-		while (remaining) {
-			if (remaining > max_bytes)
-				current_count = max_bytes;
-			else
-				current_count = remaining;
-			ret = radeon_process_i2c_ch(i2c,
-						    p->addr, flags,
-						    &p->buf[buffer_offset], current_count);
-			if (ret)
-				return ret;
-			remaining -= current_count;
-			buffer_offset += current_count;
-		}
-	}
-
-	return num;
-}
-
-u32 radeon_atom_hw_i2c_func(struct i2c_adapter *adap)
-{
-	return I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL;
-}
-
diff --git a/drivers/gpu/drm/radeon/evergreen.c b/drivers/gpu/drm/radeon/evergreen.c
index 04a5d83..87ff586 100644
--- a/drivers/gpu/drm/radeon/evergreen.c
+++ b/drivers/gpu/drm/radeon/evergreen.c
@@ -39,40 +39,7 @@
 
 static void evergreen_gpu_init(struct radeon_device *rdev);
 void evergreen_fini(struct radeon_device *rdev);
-void evergreen_pcie_gen2_enable(struct radeon_device *rdev);
-extern void cayman_cp_int_cntl_setup(struct radeon_device *rdev,
-				     int ring, u32 cp_int_cntl);
-
-void evergreen_tiling_fields(unsigned tiling_flags, unsigned *bankw,
-			     unsigned *bankh, unsigned *mtaspect,
-			     unsigned *tile_split)
-{
-	*bankw = (tiling_flags >> RADEON_TILING_EG_BANKW_SHIFT) & RADEON_TILING_EG_BANKW_MASK;
-	*bankh = (tiling_flags >> RADEON_TILING_EG_BANKH_SHIFT) & RADEON_TILING_EG_BANKH_MASK;
-	*mtaspect = (tiling_flags >> RADEON_TILING_EG_MACRO_TILE_ASPECT_SHIFT) & RADEON_TILING_EG_MACRO_TILE_ASPECT_MASK;
-	*tile_split = (tiling_flags >> RADEON_TILING_EG_TILE_SPLIT_SHIFT) & RADEON_TILING_EG_TILE_SPLIT_MASK;
-	switch (*bankw) {
-	default:
-	case 1: *bankw = EVERGREEN_ADDR_SURF_BANK_WIDTH_1; break;
-	case 2: *bankw = EVERGREEN_ADDR_SURF_BANK_WIDTH_2; break;
-	case 4: *bankw = EVERGREEN_ADDR_SURF_BANK_WIDTH_4; break;
-	case 8: *bankw = EVERGREEN_ADDR_SURF_BANK_WIDTH_8; break;
-	}
-	switch (*bankh) {
-	default:
-	case 1: *bankh = EVERGREEN_ADDR_SURF_BANK_HEIGHT_1; break;
-	case 2: *bankh = EVERGREEN_ADDR_SURF_BANK_HEIGHT_2; break;
-	case 4: *bankh = EVERGREEN_ADDR_SURF_BANK_HEIGHT_4; break;
-	case 8: *bankh = EVERGREEN_ADDR_SURF_BANK_HEIGHT_8; break;
-	}
-	switch (*mtaspect) {
-	default:
-	case 1: *mtaspect = EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_1; break;
-	case 2: *mtaspect = EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_2; break;
-	case 4: *mtaspect = EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_4; break;
-	case 8: *mtaspect = EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_8; break;
-	}
-}
+static void evergreen_pcie_gen2_enable(struct radeon_device *rdev);
 
 void evergreen_fix_pci_max_read_req_size(struct radeon_device *rdev)
 {
@@ -99,25 +66,6 @@ void evergreen_fix_pci_max_read_req_size(struct radeon_device *rdev)
 	}
 }
 
-void dce4_wait_for_vblank(struct radeon_device *rdev, int crtc)
-{
-	struct radeon_crtc *radeon_crtc = rdev->mode_info.crtcs[crtc];
-	int i;
-
-	if (RREG32(EVERGREEN_CRTC_CONTROL + radeon_crtc->crtc_offset) & EVERGREEN_CRTC_MASTER_EN) {
-		for (i = 0; i < rdev->usec_timeout; i++) {
-			if (!(RREG32(EVERGREEN_CRTC_STATUS + radeon_crtc->crtc_offset) & EVERGREEN_CRTC_V_BLANK))
-				break;
-			udelay(1);
-		}
-		for (i = 0; i < rdev->usec_timeout; i++) {
-			if (RREG32(EVERGREEN_CRTC_STATUS + radeon_crtc->crtc_offset) & EVERGREEN_CRTC_V_BLANK)
-				break;
-			udelay(1);
-		}
-	}
-}
-
 void evergreen_pre_page_flip(struct radeon_device *rdev, int crtc)
 {
 	/* enable the pflip int */
@@ -214,57 +162,6 @@ int sumo_get_temp(struct radeon_device *rdev)
 	return actual_temp * 1000;
 }
 
-void sumo_pm_init_profile(struct radeon_device *rdev)
-{
-	int idx;
-
-	/* default */
-	rdev->pm.profiles[PM_PROFILE_DEFAULT_IDX].dpms_off_ps_idx = rdev->pm.default_power_state_index;
-	rdev->pm.profiles[PM_PROFILE_DEFAULT_IDX].dpms_on_ps_idx = rdev->pm.default_power_state_index;
-	rdev->pm.profiles[PM_PROFILE_DEFAULT_IDX].dpms_off_cm_idx = 0;
-	rdev->pm.profiles[PM_PROFILE_DEFAULT_IDX].dpms_on_cm_idx = 0;
-
-	/* low,mid sh/mh */
-	if (rdev->flags & RADEON_IS_MOBILITY)
-		idx = radeon_pm_get_type_index(rdev, POWER_STATE_TYPE_BATTERY, 0);
-	else
-		idx = radeon_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 0);
-
-	rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_off_ps_idx = idx;
-	rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_ps_idx = idx;
-	rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_off_cm_idx = 0;
-	rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_cm_idx = 0;
-
-	rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_off_ps_idx = idx;
-	rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_ps_idx = idx;
-	rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_off_cm_idx = 0;
-	rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_cm_idx = 0;
-
-	rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_ps_idx = idx;
-	rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_ps_idx = idx;
-	rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_cm_idx = 0;
-	rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_cm_idx = 0;
-
-	rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_ps_idx = idx;
-	rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_ps_idx = idx;
-	rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_cm_idx = 0;
-	rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_cm_idx = 0;
-
-	/* high sh/mh */
-	idx = radeon_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 0);
-	rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_off_ps_idx = idx;
-	rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_on_ps_idx = idx;
-	rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_off_cm_idx = 0;
-	rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_on_cm_idx =
-		rdev->pm.power_state[idx].num_clock_modes - 1;
-
-	rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_off_ps_idx = idx;
-	rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_on_ps_idx = idx;
-	rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_off_cm_idx = 0;
-	rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_on_cm_idx =
-		rdev->pm.power_state[idx].num_clock_modes - 1;
-}
-
 void evergreen_pm_misc(struct radeon_device *rdev)
 {
 	int req_ps_idx = rdev->pm.requested_power_state_index;
@@ -904,7 +801,7 @@ static void evergreen_program_watermarks(struct radeon_device *rdev,
 		    !evergreen_average_bandwidth_vs_available_bandwidth(&wm) ||
 		    !evergreen_check_latency_hiding(&wm) ||
 		    (rdev->disp_priority == 2)) {
-			DRM_DEBUG_KMS("force priority to high\n");
+			DRM_INFO("force priority to high\n");
 			priority_a_cnt |= PRIORITY_ALWAYS_ON;
 			priority_b_cnt |= PRIORITY_ALWAYS_ON;
 		}
@@ -1029,7 +926,7 @@ int evergreen_pcie_gart_enable(struct radeon_device *rdev)
 	u32 tmp;
 	int r;
 
-	if (rdev->gart.robj == NULL) {
+	if (rdev->gart.table.vram.robj == NULL) {
 		dev_err(rdev->dev, "No VRAM object for PCIE GART.\n");
 		return -EINVAL;
 	}
@@ -1076,9 +973,6 @@ int evergreen_pcie_gart_enable(struct radeon_device *rdev)
 	WREG32(VM_CONTEXT1_CNTL, 0);
 
 	evergreen_pcie_gart_tlb_flush(rdev);
-	DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n",
-		 (unsigned)(rdev->mc.gtt_size >> 20),
-		 (unsigned long long)rdev->gart.table_addr);
 	rdev->gart.ready = true;
 	return 0;
 }
@@ -1086,6 +980,7 @@ int evergreen_pcie_gart_enable(struct radeon_device *rdev)
 void evergreen_pcie_gart_disable(struct radeon_device *rdev)
 {
 	u32 tmp;
+	int r;
 
 	/* Disable all tables */
 	WREG32(VM_CONTEXT0_CNTL, 0);
@@ -1105,7 +1000,14 @@ void evergreen_pcie_gart_disable(struct radeon_device *rdev)
 	WREG32(MC_VM_MB_L1_TLB1_CNTL, tmp);
 	WREG32(MC_VM_MB_L1_TLB2_CNTL, tmp);
 	WREG32(MC_VM_MB_L1_TLB3_CNTL, tmp);
-	radeon_gart_table_vram_unpin(rdev);
+	if (rdev->gart.table.vram.robj) {
+		r = radeon_bo_reserve(rdev->gart.table.vram.robj, false);
+		if (likely(r == 0)) {
+			radeon_bo_kunmap(rdev->gart.table.vram.robj);
+			radeon_bo_unpin(rdev->gart.table.vram.robj);
+			radeon_bo_unreserve(rdev->gart.table.vram.robj);
+		}
+	}
 }
 
 void evergreen_pcie_gart_fini(struct radeon_device *rdev)
@@ -1304,7 +1206,7 @@ void evergreen_mc_program(struct radeon_device *rdev)
 		WREG32(MC_VM_SYSTEM_APERTURE_HIGH_ADDR,
 			rdev->mc.vram_end >> 12);
 	}
-	WREG32(MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR, rdev->vram_scratch.gpu_addr >> 12);
+	WREG32(MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR, 0);
 	if (rdev->flags & RADEON_IS_IGP) {
 		tmp = RREG32(MC_FUS_VM_FB_OFFSET) & 0x000FFFFF;
 		tmp |= ((rdev->mc.vram_end >> 20) & 0xF) << 24;
@@ -1340,20 +1242,18 @@ void evergreen_mc_program(struct radeon_device *rdev)
  */
 void evergreen_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib)
 {
-	struct radeon_ring *ring = &rdev->ring[ib->fence->ring];
-
 	/* set to DX10/11 mode */
-	radeon_ring_write(ring, PACKET3(PACKET3_MODE_CONTROL, 0));
-	radeon_ring_write(ring, 1);
+	radeon_ring_write(rdev, PACKET3(PACKET3_MODE_CONTROL, 0));
+	radeon_ring_write(rdev, 1);
 	/* FIXME: implement */
-	radeon_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2));
-	radeon_ring_write(ring,
+	radeon_ring_write(rdev, PACKET3(PACKET3_INDIRECT_BUFFER, 2));
+	radeon_ring_write(rdev,
 #ifdef __BIG_ENDIAN
 			  (2 << 0) |
 #endif
 			  (ib->gpu_addr & 0xFFFFFFFC));
-	radeon_ring_write(ring, upper_32_bits(ib->gpu_addr) & 0xFF);
-	radeon_ring_write(ring, ib->length_dw);
+	radeon_ring_write(rdev, upper_32_bits(ib->gpu_addr) & 0xFF);
+	radeon_ring_write(rdev, ib->length_dw);
 }
 
 
@@ -1391,73 +1291,71 @@ static int evergreen_cp_load_microcode(struct radeon_device *rdev)
 
 static int evergreen_cp_start(struct radeon_device *rdev)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	int r, i;
 	uint32_t cp_me;
 
-	r = radeon_ring_lock(rdev, ring, 7);
+	r = radeon_ring_lock(rdev, 7);
 	if (r) {
 		DRM_ERROR("radeon: cp failed to lock ring (%d).\n", r);
 		return r;
 	}
-	radeon_ring_write(ring, PACKET3(PACKET3_ME_INITIALIZE, 5));
-	radeon_ring_write(ring, 0x1);
-	radeon_ring_write(ring, 0x0);
-	radeon_ring_write(ring, rdev->config.evergreen.max_hw_contexts - 1);
-	radeon_ring_write(ring, PACKET3_ME_INITIALIZE_DEVICE_ID(1));
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, 0);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_write(rdev, PACKET3(PACKET3_ME_INITIALIZE, 5));
+	radeon_ring_write(rdev, 0x1);
+	radeon_ring_write(rdev, 0x0);
+	radeon_ring_write(rdev, rdev->config.evergreen.max_hw_contexts - 1);
+	radeon_ring_write(rdev, PACKET3_ME_INITIALIZE_DEVICE_ID(1));
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_unlock_commit(rdev);
 
 	cp_me = 0xff;
 	WREG32(CP_ME_CNTL, cp_me);
 
-	r = radeon_ring_lock(rdev, ring, evergreen_default_size + 19);
+	r = radeon_ring_lock(rdev, evergreen_default_size + 19);
 	if (r) {
 		DRM_ERROR("radeon: cp failed to lock ring (%d).\n", r);
 		return r;
 	}
 
 	/* setup clear context state */
-	radeon_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0));
-	radeon_ring_write(ring, PACKET3_PREAMBLE_BEGIN_CLEAR_STATE);
+	radeon_ring_write(rdev, PACKET3(PACKET3_PREAMBLE_CNTL, 0));
+	radeon_ring_write(rdev, PACKET3_PREAMBLE_BEGIN_CLEAR_STATE);
 
 	for (i = 0; i < evergreen_default_size; i++)
-		radeon_ring_write(ring, evergreen_default_state[i]);
+		radeon_ring_write(rdev, evergreen_default_state[i]);
 
-	radeon_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0));
-	radeon_ring_write(ring, PACKET3_PREAMBLE_END_CLEAR_STATE);
+	radeon_ring_write(rdev, PACKET3(PACKET3_PREAMBLE_CNTL, 0));
+	radeon_ring_write(rdev, PACKET3_PREAMBLE_END_CLEAR_STATE);
 
 	/* set clear context state */
-	radeon_ring_write(ring, PACKET3(PACKET3_CLEAR_STATE, 0));
-	radeon_ring_write(ring, 0);
+	radeon_ring_write(rdev, PACKET3(PACKET3_CLEAR_STATE, 0));
+	radeon_ring_write(rdev, 0);
 
 	/* SQ_VTX_BASE_VTX_LOC */
-	radeon_ring_write(ring, 0xc0026f00);
-	radeon_ring_write(ring, 0x00000000);
-	radeon_ring_write(ring, 0x00000000);
-	radeon_ring_write(ring, 0x00000000);
+	radeon_ring_write(rdev, 0xc0026f00);
+	radeon_ring_write(rdev, 0x00000000);
+	radeon_ring_write(rdev, 0x00000000);
+	radeon_ring_write(rdev, 0x00000000);
 
 	/* Clear consts */
-	radeon_ring_write(ring, 0xc0036f00);
-	radeon_ring_write(ring, 0x00000bc4);
-	radeon_ring_write(ring, 0xffffffff);
-	radeon_ring_write(ring, 0xffffffff);
-	radeon_ring_write(ring, 0xffffffff);
+	radeon_ring_write(rdev, 0xc0036f00);
+	radeon_ring_write(rdev, 0x00000bc4);
+	radeon_ring_write(rdev, 0xffffffff);
+	radeon_ring_write(rdev, 0xffffffff);
+	radeon_ring_write(rdev, 0xffffffff);
 
-	radeon_ring_write(ring, 0xc0026900);
-	radeon_ring_write(ring, 0x00000316);
-	radeon_ring_write(ring, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */
-	radeon_ring_write(ring, 0x00000010); /*  */
+	radeon_ring_write(rdev, 0xc0026900);
+	radeon_ring_write(rdev, 0x00000316);
+	radeon_ring_write(rdev, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */
+	radeon_ring_write(rdev, 0x00000010); /*  */
 
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev);
 
 	return 0;
 }
 
 int evergreen_cp_resume(struct radeon_device *rdev)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	u32 tmp;
 	u32 rb_bufsz;
 	int r;
@@ -1475,14 +1373,13 @@ int evergreen_cp_resume(struct radeon_device *rdev)
 	RREG32(GRBM_SOFT_RESET);
 
 	/* Set ring buffer size */
-	rb_bufsz = drm_order(ring->ring_size / 8);
+	rb_bufsz = drm_order(rdev->cp.ring_size / 8);
 	tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz;
 #ifdef __BIG_ENDIAN
 	tmp |= BUF_SWAP_32BIT;
 #endif
 	WREG32(CP_RB_CNTL, tmp);
-	WREG32(CP_SEM_WAIT_TIMER, 0x0);
-	WREG32(CP_SEM_INCOMPLETE_TIMER_CNTL, 0x0);
+	WREG32(CP_SEM_WAIT_TIMER, 0x4);
 
 	/* Set the write pointer delay */
 	WREG32(CP_RB_WPTR_DELAY, 0);
@@ -1490,11 +1387,14 @@ int evergreen_cp_resume(struct radeon_device *rdev)
 	/* Initialize the ring buffer's read and write pointers */
 	WREG32(CP_RB_CNTL, tmp | RB_RPTR_WR_ENA);
 	WREG32(CP_RB_RPTR_WR, 0);
-	ring->wptr = 0;
-	WREG32(CP_RB_WPTR, ring->wptr);
+	rdev->cp.wptr = 0;
+	WREG32(CP_RB_WPTR, rdev->cp.wptr);
 
 	/* set the wb address wether it's enabled or not */
 	WREG32(CP_RB_RPTR_ADDR,
+#ifdef __BIG_ENDIAN
+	       RB_RPTR_SWAP(2) |
+#endif
 	       ((rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) & 0xFFFFFFFC));
 	WREG32(CP_RB_RPTR_ADDR_HI, upper_32_bits(rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) & 0xFF);
 	WREG32(SCRATCH_ADDR, ((rdev->wb.gpu_addr + RADEON_WB_SCRATCH_OFFSET) >> 8) & 0xFFFFFFFF);
@@ -1509,16 +1409,16 @@ int evergreen_cp_resume(struct radeon_device *rdev)
 	mdelay(1);
 	WREG32(CP_RB_CNTL, tmp);
 
-	WREG32(CP_RB_BASE, ring->gpu_addr >> 8);
+	WREG32(CP_RB_BASE, rdev->cp.gpu_addr >> 8);
 	WREG32(CP_DEBUG, (1 << 27) | (1 << 28));
 
-	ring->rptr = RREG32(CP_RB_RPTR);
+	rdev->cp.rptr = RREG32(CP_RB_RPTR);
 
 	evergreen_cp_start(rdev);
-	ring->ready = true;
-	r = radeon_ring_test(rdev, RADEON_RING_TYPE_GFX_INDEX, ring);
+	rdev->cp.ready = true;
+	r = radeon_ring_test(rdev);
 	if (r) {
-		ring->ready = false;
+		rdev->cp.ready = false;
 		return r;
 	}
 	return 0;
@@ -2116,7 +2016,6 @@ static void evergreen_gpu_init(struct radeon_device *rdev)
 	rdev->config.evergreen.tile_config |=
 		((gb_addr_config & 0x30000000) >> 28) << 12;
 
-	rdev->config.evergreen.backend_map = gb_backend_map;
 	WREG32(GB_BACKEND_MAP, gb_backend_map);
 	WREG32(GB_ADDR_CONFIG, gb_addr_config);
 	WREG32(DMIF_ADDR_CONFIG, gb_addr_config);
@@ -2387,7 +2286,7 @@ int evergreen_mc_init(struct radeon_device *rdev)
 	return 0;
 }
 
-bool evergreen_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
+bool evergreen_gpu_is_lockup(struct radeon_device *rdev)
 {
 	u32 srbm_status;
 	u32 grbm_status;
@@ -2400,19 +2299,19 @@ bool evergreen_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *rin
 	grbm_status_se0 = RREG32(GRBM_STATUS_SE0);
 	grbm_status_se1 = RREG32(GRBM_STATUS_SE1);
 	if (!(grbm_status & GUI_ACTIVE)) {
-		r100_gpu_lockup_update(lockup, ring);
+		r100_gpu_lockup_update(lockup, &rdev->cp);
 		return false;
 	}
 	/* force CP activities */
-	r = radeon_ring_lock(rdev, ring, 2);
+	r = radeon_ring_lock(rdev, 2);
 	if (!r) {
 		/* PACKET2 NOP */
-		radeon_ring_write(ring, 0x80000000);
-		radeon_ring_write(ring, 0x80000000);
-		radeon_ring_unlock_commit(rdev, ring);
+		radeon_ring_write(rdev, 0x80000000);
+		radeon_ring_write(rdev, 0x80000000);
+		radeon_ring_unlock_commit(rdev);
 	}
-	ring->rptr = RREG32(CP_RB_RPTR);
-	return r100_gpu_cp_is_lockup(rdev, lockup, ring);
+	rdev->cp.rptr = RREG32(CP_RB_RPTR);
+	return r100_gpu_cp_is_lockup(rdev, lockup, &rdev->cp);
 }
 
 static int evergreen_gpu_soft_reset(struct radeon_device *rdev)
@@ -2504,13 +2403,7 @@ void evergreen_disable_interrupt_state(struct radeon_device *rdev)
 {
 	u32 tmp;
 
-	if (rdev->family >= CHIP_CAYMAN) {
-		cayman_cp_int_cntl_setup(rdev, 0,
-					 CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE);
-		cayman_cp_int_cntl_setup(rdev, 1, 0);
-		cayman_cp_int_cntl_setup(rdev, 2, 0);
-	} else
-		WREG32(CP_INT_CNTL, CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE);
+	WREG32(CP_INT_CNTL, CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE);
 	WREG32(GRBM_INT_CNTL, 0);
 	WREG32(INT_MASK + EVERGREEN_CRTC0_REGISTER_OFFSET, 0);
 	WREG32(INT_MASK + EVERGREEN_CRTC1_REGISTER_OFFSET, 0);
@@ -2555,7 +2448,6 @@ void evergreen_disable_interrupt_state(struct radeon_device *rdev)
 int evergreen_irq_set(struct radeon_device *rdev)
 {
 	u32 cp_int_cntl = CNTX_BUSY_INT_ENABLE | CNTX_EMPTY_INT_ENABLE;
-	u32 cp_int_cntl1 = 0, cp_int_cntl2 = 0;
 	u32 crtc1 = 0, crtc2 = 0, crtc3 = 0, crtc4 = 0, crtc5 = 0, crtc6 = 0;
 	u32 hpd1, hpd2, hpd3, hpd4, hpd5, hpd6;
 	u32 grbm_int_cntl = 0;
@@ -2580,28 +2472,11 @@ int evergreen_irq_set(struct radeon_device *rdev)
 	hpd5 = RREG32(DC_HPD5_INT_CONTROL) & ~DC_HPDx_INT_EN;
 	hpd6 = RREG32(DC_HPD6_INT_CONTROL) & ~DC_HPDx_INT_EN;
 
-	if (rdev->family >= CHIP_CAYMAN) {
-		/* enable CP interrupts on all rings */
-		if (rdev->irq.sw_int[RADEON_RING_TYPE_GFX_INDEX]) {
-			DRM_DEBUG("evergreen_irq_set: sw int gfx\n");
-			cp_int_cntl |= TIME_STAMP_INT_ENABLE;
-		}
-		if (rdev->irq.sw_int[CAYMAN_RING_TYPE_CP1_INDEX]) {
-			DRM_DEBUG("evergreen_irq_set: sw int cp1\n");
-			cp_int_cntl1 |= TIME_STAMP_INT_ENABLE;
-		}
-		if (rdev->irq.sw_int[CAYMAN_RING_TYPE_CP2_INDEX]) {
-			DRM_DEBUG("evergreen_irq_set: sw int cp2\n");
-			cp_int_cntl2 |= TIME_STAMP_INT_ENABLE;
-		}
-	} else {
-		if (rdev->irq.sw_int[RADEON_RING_TYPE_GFX_INDEX]) {
-			DRM_DEBUG("evergreen_irq_set: sw int gfx\n");
-			cp_int_cntl |= RB_INT_ENABLE;
-			cp_int_cntl |= TIME_STAMP_INT_ENABLE;
-		}
+	if (rdev->irq.sw_int) {
+		DRM_DEBUG("evergreen_irq_set: sw int\n");
+		cp_int_cntl |= RB_INT_ENABLE;
+		cp_int_cntl |= TIME_STAMP_INT_ENABLE;
 	}
-
 	if (rdev->irq.crtc_vblank_int[0] ||
 	    rdev->irq.pflip[0]) {
 		DRM_DEBUG("evergreen_irq_set: vblank 0\n");
@@ -2661,12 +2536,7 @@ int evergreen_irq_set(struct radeon_device *rdev)
 		grbm_int_cntl |= GUI_IDLE_INT_ENABLE;
 	}
 
-	if (rdev->family >= CHIP_CAYMAN) {
-		cayman_cp_int_cntl_setup(rdev, 0, cp_int_cntl);
-		cayman_cp_int_cntl_setup(rdev, 1, cp_int_cntl1);
-		cayman_cp_int_cntl_setup(rdev, 2, cp_int_cntl2);
-	} else
-		WREG32(CP_INT_CNTL, cp_int_cntl);
+	WREG32(CP_INT_CNTL, cp_int_cntl);
 	WREG32(GRBM_INT_CNTL, grbm_int_cntl);
 
 	WREG32(INT_MASK + EVERGREEN_CRTC0_REGISTER_OFFSET, crtc1);
@@ -2701,7 +2571,7 @@ int evergreen_irq_set(struct radeon_device *rdev)
 	return 0;
 }
 
-static void evergreen_irq_ack(struct radeon_device *rdev)
+static inline void evergreen_irq_ack(struct radeon_device *rdev)
 {
 	u32 tmp;
 
@@ -2812,7 +2682,7 @@ void evergreen_irq_suspend(struct radeon_device *rdev)
 	r600_rlc_stop(rdev);
 }
 
-static u32 evergreen_get_ih_wptr(struct radeon_device *rdev)
+static inline u32 evergreen_get_ih_wptr(struct radeon_device *rdev)
 {
 	u32 wptr, tmp;
 
@@ -2858,9 +2728,6 @@ int evergreen_irq_process(struct radeon_device *rdev)
 		return IRQ_NONE;
 	}
 restart_ih:
-	/* Order reading of wptr vs. reading of IH ring data */
-	rmb();
-
 	/* display interrupts */
 	evergreen_irq_ack(rdev);
 
@@ -3081,24 +2948,11 @@ restart_ih:
 		case 177: /* CP_INT in IB1 */
 		case 178: /* CP_INT in IB2 */
 			DRM_DEBUG("IH: CP int: 0x%08x\n", src_data);
-			radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX);
+			radeon_fence_process(rdev);
 			break;
 		case 181: /* CP EOP event */
 			DRM_DEBUG("IH: CP EOP\n");
-			if (rdev->family >= CHIP_CAYMAN) {
-				switch (src_data) {
-				case 0:
-					radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX);
-					break;
-				case 1:
-					radeon_fence_process(rdev, CAYMAN_RING_TYPE_CP1_INDEX);
-					break;
-				case 2:
-					radeon_fence_process(rdev, CAYMAN_RING_TYPE_CP2_INDEX);
-					break;
-				}
-			} else
-				radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX);
+			radeon_fence_process(rdev);
 			break;
 		case 233: /* GUI IDLE */
 			DRM_DEBUG("IH: GUI idle\n");
@@ -3128,11 +2982,11 @@ restart_ih:
 
 static int evergreen_startup(struct radeon_device *rdev)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	int r;
 
 	/* enable pcie gen2 link */
-	evergreen_pcie_gen2_enable(rdev);
+	if (!ASIC_IS_DCE5(rdev))
+		evergreen_pcie_gen2_enable(rdev);
 
 	if (ASIC_IS_DCE5(rdev)) {
 		if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw || !rdev->mc_fw) {
@@ -3157,10 +3011,6 @@ static int evergreen_startup(struct radeon_device *rdev)
 		}
 	}
 
-	r = r600_vram_scratch_init(rdev);
-	if (r)
-		return r;
-
 	evergreen_mc_program(rdev);
 	if (rdev->flags & RADEON_IS_AGP) {
 		evergreen_agp_enable(rdev);
@@ -3173,8 +3023,8 @@ static int evergreen_startup(struct radeon_device *rdev)
 
 	r = evergreen_blit_init(rdev);
 	if (r) {
-		r600_blit_fini(rdev);
-		rdev->asic->copy.copy = NULL;
+		evergreen_blit_fini(rdev);
+		rdev->asic->copy = NULL;
 		dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r);
 	}
 
@@ -3183,12 +3033,6 @@ static int evergreen_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX);
-	if (r) {
-		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
-		return r;
-	}
-
 	/* Enable IRQ */
 	r = r600_irq_init(rdev);
 	if (r) {
@@ -3198,9 +3042,7 @@ static int evergreen_startup(struct radeon_device *rdev)
 	}
 	evergreen_irq_set(rdev);
 
-	r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP_RPTR_OFFSET,
-			     R600_CP_RB_RPTR, R600_CP_RB_WPTR,
-			     0, 0xfffff, RADEON_CP_PACKET2);
+	r = radeon_ring_init(rdev, rdev->cp.ring_size);
 	if (r)
 		return r;
 	r = evergreen_cp_load_microcode(rdev);
@@ -3210,23 +3052,6 @@ static int evergreen_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = radeon_ib_pool_start(rdev);
-	if (r)
-		return r;
-
-	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
-	if (r) {
-		DRM_ERROR("radeon: failed testing IB (%d).\n", r);
-		rdev->accel_working = false;
-		return r;
-	}
-
-	r = r600_audio_init(rdev);
-	if (r) {
-		DRM_ERROR("radeon: audio init failed\n");
-		return r;
-	}
-
 	return 0;
 }
 
@@ -3246,11 +3071,15 @@ int evergreen_resume(struct radeon_device *rdev)
 	/* post card */
 	atom_asic_init(rdev->mode_info.atom_context);
 
-	rdev->accel_working = true;
 	r = evergreen_startup(rdev);
 	if (r) {
 		DRM_ERROR("evergreen startup failed on resume\n");
-		rdev->accel_working = false;
+		return r;
+	}
+
+	r = r600_ib_test(rdev);
+	if (r) {
+		DRM_ERROR("radeon: failed testing IB (%d).\n", r);
 		return r;
 	}
 
@@ -3260,18 +3089,45 @@ int evergreen_resume(struct radeon_device *rdev)
 
 int evergreen_suspend(struct radeon_device *rdev)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
+	int r;
 
-	r600_audio_fini(rdev);
 	/* FIXME: we should wait for ring to be empty */
-	radeon_ib_pool_suspend(rdev);
-	r600_blit_suspend(rdev);
 	r700_cp_stop(rdev);
-	ring->ready = false;
+	rdev->cp.ready = false;
 	evergreen_irq_suspend(rdev);
 	radeon_wb_disable(rdev);
 	evergreen_pcie_gart_disable(rdev);
 
+	/* unpin shaders bo */
+	r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false);
+	if (likely(r == 0)) {
+		radeon_bo_unpin(rdev->r600_blit.shader_obj);
+		radeon_bo_unreserve(rdev->r600_blit.shader_obj);
+	}
+
+	return 0;
+}
+
+int evergreen_copy_blit(struct radeon_device *rdev,
+			uint64_t src_offset,
+			uint64_t dst_offset,
+			unsigned num_gpu_pages,
+			struct radeon_fence *fence)
+{
+	int r;
+
+	mutex_lock(&rdev->r600_blit.mutex);
+	rdev->r600_blit.vb_ib = NULL;
+	r = evergreen_blit_prepare_copy(rdev, num_gpu_pages * RADEON_GPU_PAGE_SIZE);
+	if (r) {
+		if (rdev->r600_blit.vb_ib)
+			radeon_ib_free(rdev, &rdev->r600_blit.vb_ib);
+		mutex_unlock(&rdev->r600_blit.mutex);
+		return r;
+	}
+	evergreen_kms_blit_copy(rdev, src_offset, dst_offset, num_gpu_pages * RADEON_GPU_PAGE_SIZE);
+	evergreen_blit_done_copy(rdev, fence);
+	mutex_unlock(&rdev->r600_blit.mutex);
 	return 0;
 }
 
@@ -3345,8 +3201,8 @@ int evergreen_init(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ring_obj = NULL;
-	r600_ring_init(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX], 1024 * 1024);
+	rdev->cp.ring_obj = NULL;
+	r600_ring_init(rdev, 1024 * 1024);
 
 	rdev->ih.ring_obj = NULL;
 	r600_ih_ring_init(rdev, 64 * 1024);
@@ -3355,24 +3211,29 @@ int evergreen_init(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = radeon_ib_pool_init(rdev);
 	rdev->accel_working = true;
-	if (r) {
-		dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
-		rdev->accel_working = false;
-	}
-
 	r = evergreen_startup(rdev);
 	if (r) {
 		dev_err(rdev->dev, "disabling GPU acceleration\n");
 		r700_cp_fini(rdev);
 		r600_irq_fini(rdev);
 		radeon_wb_fini(rdev);
-		r100_ib_fini(rdev);
 		radeon_irq_kms_fini(rdev);
 		evergreen_pcie_gart_fini(rdev);
 		rdev->accel_working = false;
 	}
+	if (rdev->accel_working) {
+		r = radeon_ib_pool_init(rdev);
+		if (r) {
+			DRM_ERROR("radeon: failed initializing IB pool (%d).\n", r);
+			rdev->accel_working = false;
+		}
+		r = r600_ib_test(rdev);
+		if (r) {
+			DRM_ERROR("radeon: failed testing IB (%d).\n", r);
+			rdev->accel_working = false;
+		}
+	}
 
 	/* Don't start up if the MC ucode is missing on BTC parts.
 	 * The default clocks and voltages before the MC ucode
@@ -3390,17 +3251,14 @@ int evergreen_init(struct radeon_device *rdev)
 
 void evergreen_fini(struct radeon_device *rdev)
 {
-	r600_audio_fini(rdev);
-	r600_blit_fini(rdev);
+	evergreen_blit_fini(rdev);
 	r700_cp_fini(rdev);
 	r600_irq_fini(rdev);
 	radeon_wb_fini(rdev);
-	r100_ib_fini(rdev);
+	radeon_ib_pool_fini(rdev);
 	radeon_irq_kms_fini(rdev);
 	evergreen_pcie_gart_fini(rdev);
-	r600_vram_scratch_fini(rdev);
 	radeon_gem_fini(rdev);
-	radeon_semaphore_driver_fini(rdev);
 	radeon_fence_driver_fini(rdev);
 	radeon_agp_fini(rdev);
 	radeon_bo_fini(rdev);
@@ -3409,7 +3267,7 @@ void evergreen_fini(struct radeon_device *rdev)
 	rdev->bios = NULL;
 }
 
-void evergreen_pcie_gen2_enable(struct radeon_device *rdev)
+static void evergreen_pcie_gen2_enable(struct radeon_device *rdev)
 {
 	u32 link_width_cntl, speed_cntl;
 
diff --git a/drivers/gpu/drm/radeon/evergreen_blit_kms.c b/drivers/gpu/drm/radeon/evergreen_blit_kms.c
index 4e83fdc..2eb2518 100644
--- a/drivers/gpu/drm/radeon/evergreen_blit_kms.c
+++ b/drivers/gpu/drm/radeon/evergreen_blit_kms.c
@@ -32,14 +32,23 @@
 #include "evergreend.h"
 #include "evergreen_blit_shaders.h"
 #include "cayman_blit_shaders.h"
-#include "radeon_blit_common.h"
+
+#define DI_PT_RECTLIST        0x11
+#define DI_INDEX_SIZE_16_BIT  0x0
+#define DI_SRC_SEL_AUTO_INDEX 0x2
+
+#define FMT_8                 0x1
+#define FMT_5_6_5             0x8
+#define FMT_8_8_8_8           0x1a
+#define COLOR_8               0x1
+#define COLOR_5_6_5           0x8
+#define COLOR_8_8_8_8         0x1a
 
 /* emits 17 */
 static void
 set_render_target(struct radeon_device *rdev, int format,
 		  int w, int h, u64 gpu_addr)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	u32 cb_color_info;
 	int pitch, slice;
 
@@ -47,29 +56,27 @@ set_render_target(struct radeon_device *rdev, int format,
 	if (h < 8)
 		h = 8;
 
-	cb_color_info = CB_FORMAT(format) |
-		CB_SOURCE_FORMAT(CB_SF_EXPORT_NORM) |
-		CB_ARRAY_MODE(ARRAY_1D_TILED_THIN1);
+	cb_color_info = ((format << 2) | (1 << 24) | (1 << 8));
 	pitch = (w / 8) - 1;
 	slice = ((w * h) / 64) - 1;
 
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 15));
-	radeon_ring_write(ring, (CB_COLOR0_BASE - PACKET3_SET_CONTEXT_REG_START) >> 2);
-	radeon_ring_write(ring, gpu_addr >> 8);
-	radeon_ring_write(ring, pitch);
-	radeon_ring_write(ring, slice);
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, cb_color_info);
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, (w - 1) | ((h - 1) << 16));
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, 0);
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 15));
+	radeon_ring_write(rdev, (CB_COLOR0_BASE - PACKET3_SET_CONTEXT_REG_START) >> 2);
+	radeon_ring_write(rdev, gpu_addr >> 8);
+	radeon_ring_write(rdev, pitch);
+	radeon_ring_write(rdev, slice);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, cb_color_info);
+	radeon_ring_write(rdev, (1 << 4));
+	radeon_ring_write(rdev, (w - 1) | ((h - 1) << 16));
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
 }
 
 /* emits 5dw */
@@ -78,7 +85,6 @@ cp_set_surface_sync(struct radeon_device *rdev,
 		    u32 sync_type, u32 size,
 		    u64 mc_addr)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	u32 cp_coher_size;
 
 	if (size == 0xffffffff)
@@ -86,45 +92,35 @@ cp_set_surface_sync(struct radeon_device *rdev,
 	else
 		cp_coher_size = ((size + 255) >> 8);
 
-	if (rdev->family >= CHIP_CAYMAN) {
-		/* CP_COHER_CNTL2 has to be set manually when submitting a surface_sync
-		 * to the RB directly. For IBs, the CP programs this as part of the
-		 * surface_sync packet.
-		 */
-		radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
-		radeon_ring_write(ring, (0x85e8 - PACKET3_SET_CONFIG_REG_START) >> 2);
-		radeon_ring_write(ring, 0); /* CP_COHER_CNTL2 */
-	}
-	radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3));
-	radeon_ring_write(ring, sync_type);
-	radeon_ring_write(ring, cp_coher_size);
-	radeon_ring_write(ring, mc_addr >> 8);
-	radeon_ring_write(ring, 10); /* poll interval */
+	radeon_ring_write(rdev, PACKET3(PACKET3_SURFACE_SYNC, 3));
+	radeon_ring_write(rdev, sync_type);
+	radeon_ring_write(rdev, cp_coher_size);
+	radeon_ring_write(rdev, mc_addr >> 8);
+	radeon_ring_write(rdev, 10); /* poll interval */
 }
 
 /* emits 11dw + 1 surface sync = 16dw */
 static void
 set_shaders(struct radeon_device *rdev)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	u64 gpu_addr;
 
 	/* VS */
 	gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.vs_offset;
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 3));
-	radeon_ring_write(ring, (SQ_PGM_START_VS - PACKET3_SET_CONTEXT_REG_START) >> 2);
-	radeon_ring_write(ring, gpu_addr >> 8);
-	radeon_ring_write(ring, 2);
-	radeon_ring_write(ring, 0);
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 3));
+	radeon_ring_write(rdev, (SQ_PGM_START_VS - PACKET3_SET_CONTEXT_REG_START) >> 2);
+	radeon_ring_write(rdev, gpu_addr >> 8);
+	radeon_ring_write(rdev, 2);
+	radeon_ring_write(rdev, 0);
 
 	/* PS */
 	gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.ps_offset;
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 4));
-	radeon_ring_write(ring, (SQ_PGM_START_PS - PACKET3_SET_CONTEXT_REG_START) >> 2);
-	radeon_ring_write(ring, gpu_addr >> 8);
-	radeon_ring_write(ring, 1);
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, 2);
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 4));
+	radeon_ring_write(rdev, (SQ_PGM_START_PS - PACKET3_SET_CONTEXT_REG_START) >> 2);
+	radeon_ring_write(rdev, gpu_addr >> 8);
+	radeon_ring_write(rdev, 1);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 2);
 
 	gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.vs_offset;
 	cp_set_surface_sync(rdev, PACKET3_SH_ACTION_ENA, 512, gpu_addr);
@@ -134,31 +130,26 @@ set_shaders(struct radeon_device *rdev)
 static void
 set_vtx_resource(struct radeon_device *rdev, u64 gpu_addr)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	u32 sq_vtx_constant_word2, sq_vtx_constant_word3;
 
 	/* high addr, stride */
-	sq_vtx_constant_word2 = SQ_VTXC_BASE_ADDR_HI(upper_32_bits(gpu_addr) & 0xff) |
-		SQ_VTXC_STRIDE(16);
+	sq_vtx_constant_word2 = ((upper_32_bits(gpu_addr) & 0xff) | (16 << 8));
 #ifdef __BIG_ENDIAN
-	sq_vtx_constant_word2 |= SQ_VTXC_ENDIAN_SWAP(SQ_ENDIAN_8IN32);
+	sq_vtx_constant_word2 |= (2 << 30);
 #endif
 	/* xyzw swizzles */
-	sq_vtx_constant_word3 = SQ_VTCX_SEL_X(SQ_SEL_X) |
-		SQ_VTCX_SEL_Y(SQ_SEL_Y) |
-		SQ_VTCX_SEL_Z(SQ_SEL_Z) |
-		SQ_VTCX_SEL_W(SQ_SEL_W);
-
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_RESOURCE, 8));
-	radeon_ring_write(ring, 0x580);
-	radeon_ring_write(ring, gpu_addr & 0xffffffff);
-	radeon_ring_write(ring, 48 - 1); /* size */
-	radeon_ring_write(ring, sq_vtx_constant_word2);
-	radeon_ring_write(ring, sq_vtx_constant_word3);
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, S__SQ_CONSTANT_TYPE(SQ_TEX_VTX_VALID_BUFFER));
+	sq_vtx_constant_word3 = (0 << 3) | (1 << 6) | (2 << 9) | (3 << 12);
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_RESOURCE, 8));
+	radeon_ring_write(rdev, 0x580);
+	radeon_ring_write(rdev, gpu_addr & 0xffffffff);
+	radeon_ring_write(rdev, 48 - 1); /* size */
+	radeon_ring_write(rdev, sq_vtx_constant_word2);
+	radeon_ring_write(rdev, sq_vtx_constant_word3);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, SQ_TEX_VTX_VALID_BUFFER << 30);
 
 	if ((rdev->family == CHIP_CEDAR) ||
 	    (rdev->family == CHIP_PALM) ||
@@ -177,42 +168,33 @@ set_vtx_resource(struct radeon_device *rdev, u64 gpu_addr)
 static void
 set_tex_resource(struct radeon_device *rdev,
 		 int format, int w, int h, int pitch,
-		 u64 gpu_addr, u32 size)
+		 u64 gpu_addr)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	u32 sq_tex_resource_word0, sq_tex_resource_word1;
 	u32 sq_tex_resource_word4, sq_tex_resource_word7;
 
 	if (h < 1)
 		h = 1;
 
-	sq_tex_resource_word0 = TEX_DIM(SQ_TEX_DIM_2D);
+	sq_tex_resource_word0 = (1 << 0); /* 2D */
 	sq_tex_resource_word0 |= ((((pitch >> 3) - 1) << 6) |
 				  ((w - 1) << 18));
-	sq_tex_resource_word1 = ((h - 1) << 0) |
-				TEX_ARRAY_MODE(ARRAY_1D_TILED_THIN1);
+	sq_tex_resource_word1 = ((h - 1) << 0) | (1 << 28);
 	/* xyzw swizzles */
-	sq_tex_resource_word4 = TEX_DST_SEL_X(SQ_SEL_X) |
-				TEX_DST_SEL_Y(SQ_SEL_Y) |
-				TEX_DST_SEL_Z(SQ_SEL_Z) |
-				TEX_DST_SEL_W(SQ_SEL_W);
-
-	sq_tex_resource_word7 = format |
-		S__SQ_CONSTANT_TYPE(SQ_TEX_VTX_VALID_TEXTURE);
-
-	cp_set_surface_sync(rdev,
-			    PACKET3_TC_ACTION_ENA, size, gpu_addr);
-
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_RESOURCE, 8));
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, sq_tex_resource_word0);
-	radeon_ring_write(ring, sq_tex_resource_word1);
-	radeon_ring_write(ring, gpu_addr >> 8);
-	radeon_ring_write(ring, gpu_addr >> 8);
-	radeon_ring_write(ring, sq_tex_resource_word4);
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, sq_tex_resource_word7);
+	sq_tex_resource_word4 = (0 << 16) | (1 << 19) | (2 << 22) | (3 << 25);
+
+	sq_tex_resource_word7 = format | (SQ_TEX_VTX_VALID_TEXTURE << 30);
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_RESOURCE, 8));
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, sq_tex_resource_word0);
+	radeon_ring_write(rdev, sq_tex_resource_word1);
+	radeon_ring_write(rdev, gpu_addr >> 8);
+	radeon_ring_write(rdev, gpu_addr >> 8);
+	radeon_ring_write(rdev, sq_tex_resource_word4);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, sq_tex_resource_word7);
 }
 
 /* emits 12 */
@@ -220,7 +202,6 @@ static void
 set_scissors(struct radeon_device *rdev, int x1, int y1,
 	     int x2, int y2)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	/* workaround some hw bugs */
 	if (x2 == 0)
 		x1 = 1;
@@ -231,44 +212,43 @@ set_scissors(struct radeon_device *rdev, int x1, int y1,
 			x2 = 2;
 	}
 
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2));
-	radeon_ring_write(ring, (PA_SC_SCREEN_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_START) >> 2);
-	radeon_ring_write(ring, (x1 << 0) | (y1 << 16));
-	radeon_ring_write(ring, (x2 << 0) | (y2 << 16));
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 2));
+	radeon_ring_write(rdev, (PA_SC_SCREEN_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_START) >> 2);
+	radeon_ring_write(rdev, (x1 << 0) | (y1 << 16));
+	radeon_ring_write(rdev, (x2 << 0) | (y2 << 16));
 
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2));
-	radeon_ring_write(ring, (PA_SC_GENERIC_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_START) >> 2);
-	radeon_ring_write(ring, (x1 << 0) | (y1 << 16) | (1 << 31));
-	radeon_ring_write(ring, (x2 << 0) | (y2 << 16));
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 2));
+	radeon_ring_write(rdev, (PA_SC_GENERIC_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_START) >> 2);
+	radeon_ring_write(rdev, (x1 << 0) | (y1 << 16) | (1 << 31));
+	radeon_ring_write(rdev, (x2 << 0) | (y2 << 16));
 
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2));
-	radeon_ring_write(ring, (PA_SC_WINDOW_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_START) >> 2);
-	radeon_ring_write(ring, (x1 << 0) | (y1 << 16) | (1 << 31));
-	radeon_ring_write(ring, (x2 << 0) | (y2 << 16));
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 2));
+	radeon_ring_write(rdev, (PA_SC_WINDOW_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_START) >> 2);
+	radeon_ring_write(rdev, (x1 << 0) | (y1 << 16) | (1 << 31));
+	radeon_ring_write(rdev, (x2 << 0) | (y2 << 16));
 }
 
 /* emits 10 */
 static void
 draw_auto(struct radeon_device *rdev)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
-	radeon_ring_write(ring, (VGT_PRIMITIVE_TYPE - PACKET3_SET_CONFIG_REG_START) >> 2);
-	radeon_ring_write(ring, DI_PT_RECTLIST);
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 1));
+	radeon_ring_write(rdev, (VGT_PRIMITIVE_TYPE - PACKET3_SET_CONFIG_REG_START) >> 2);
+	radeon_ring_write(rdev, DI_PT_RECTLIST);
 
-	radeon_ring_write(ring, PACKET3(PACKET3_INDEX_TYPE, 0));
-	radeon_ring_write(ring,
+	radeon_ring_write(rdev, PACKET3(PACKET3_INDEX_TYPE, 0));
+	radeon_ring_write(rdev,
 #ifdef __BIG_ENDIAN
 			  (2 << 2) |
 #endif
 			  DI_INDEX_SIZE_16_BIT);
 
-	radeon_ring_write(ring, PACKET3(PACKET3_NUM_INSTANCES, 0));
-	radeon_ring_write(ring, 1);
+	radeon_ring_write(rdev, PACKET3(PACKET3_NUM_INSTANCES, 0));
+	radeon_ring_write(rdev, 1);
 
-	radeon_ring_write(ring, PACKET3(PACKET3_DRAW_INDEX_AUTO, 1));
-	radeon_ring_write(ring, 3);
-	radeon_ring_write(ring, DI_SRC_SEL_AUTO_INDEX);
+	radeon_ring_write(rdev, PACKET3(PACKET3_DRAW_INDEX_AUTO, 1));
+	radeon_ring_write(rdev, 3);
+	radeon_ring_write(rdev, DI_SRC_SEL_AUTO_INDEX);
 
 }
 
@@ -276,7 +256,6 @@ draw_auto(struct radeon_device *rdev)
 static void
 set_default_state(struct radeon_device *rdev)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	u32 sq_config, sq_gpr_resource_mgmt_1, sq_gpr_resource_mgmt_2, sq_gpr_resource_mgmt_3;
 	u32 sq_thread_resource_mgmt, sq_thread_resource_mgmt_2;
 	u32 sq_stack_resource_mgmt_1, sq_stack_resource_mgmt_2, sq_stack_resource_mgmt_3;
@@ -290,8 +269,8 @@ set_default_state(struct radeon_device *rdev)
 	int dwords;
 
 	/* set clear context state */
-	radeon_ring_write(ring, PACKET3(PACKET3_CLEAR_STATE, 0));
-	radeon_ring_write(ring, 0);
+	radeon_ring_write(rdev, PACKET3(PACKET3_CLEAR_STATE, 0));
+	radeon_ring_write(rdev, 0);
 
 	if (rdev->family < CHIP_CAYMAN) {
 		switch (rdev->family) {
@@ -548,63 +527,88 @@ set_default_state(struct radeon_device *rdev)
 					    NUM_LS_STACK_ENTRIES(num_ls_stack_entries));
 
 		/* disable dyn gprs */
-		radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
-		radeon_ring_write(ring, (SQ_DYN_GPR_CNTL_PS_FLUSH_REQ - PACKET3_SET_CONFIG_REG_START) >> 2);
-		radeon_ring_write(ring, 0);
+		radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 1));
+		radeon_ring_write(rdev, (SQ_DYN_GPR_CNTL_PS_FLUSH_REQ - PACKET3_SET_CONFIG_REG_START) >> 2);
+		radeon_ring_write(rdev, 0);
 
 		/* setup LDS */
-		radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
-		radeon_ring_write(ring, (SQ_LDS_RESOURCE_MGMT - PACKET3_SET_CONFIG_REG_START) >> 2);
-		radeon_ring_write(ring, 0x10001000);
+		radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 1));
+		radeon_ring_write(rdev, (SQ_LDS_RESOURCE_MGMT - PACKET3_SET_CONFIG_REG_START) >> 2);
+		radeon_ring_write(rdev, 0x10001000);
 
 		/* SQ config */
-		radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 11));
-		radeon_ring_write(ring, (SQ_CONFIG - PACKET3_SET_CONFIG_REG_START) >> 2);
-		radeon_ring_write(ring, sq_config);
-		radeon_ring_write(ring, sq_gpr_resource_mgmt_1);
-		radeon_ring_write(ring, sq_gpr_resource_mgmt_2);
-		radeon_ring_write(ring, sq_gpr_resource_mgmt_3);
-		radeon_ring_write(ring, 0);
-		radeon_ring_write(ring, 0);
-		radeon_ring_write(ring, sq_thread_resource_mgmt);
-		radeon_ring_write(ring, sq_thread_resource_mgmt_2);
-		radeon_ring_write(ring, sq_stack_resource_mgmt_1);
-		radeon_ring_write(ring, sq_stack_resource_mgmt_2);
-		radeon_ring_write(ring, sq_stack_resource_mgmt_3);
+		radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 11));
+		radeon_ring_write(rdev, (SQ_CONFIG - PACKET3_SET_CONFIG_REG_START) >> 2);
+		radeon_ring_write(rdev, sq_config);
+		radeon_ring_write(rdev, sq_gpr_resource_mgmt_1);
+		radeon_ring_write(rdev, sq_gpr_resource_mgmt_2);
+		radeon_ring_write(rdev, sq_gpr_resource_mgmt_3);
+		radeon_ring_write(rdev, 0);
+		radeon_ring_write(rdev, 0);
+		radeon_ring_write(rdev, sq_thread_resource_mgmt);
+		radeon_ring_write(rdev, sq_thread_resource_mgmt_2);
+		radeon_ring_write(rdev, sq_stack_resource_mgmt_1);
+		radeon_ring_write(rdev, sq_stack_resource_mgmt_2);
+		radeon_ring_write(rdev, sq_stack_resource_mgmt_3);
 	}
 
 	/* CONTEXT_CONTROL */
-	radeon_ring_write(ring, 0xc0012800);
-	radeon_ring_write(ring, 0x80000000);
-	radeon_ring_write(ring, 0x80000000);
+	radeon_ring_write(rdev, 0xc0012800);
+	radeon_ring_write(rdev, 0x80000000);
+	radeon_ring_write(rdev, 0x80000000);
 
 	/* SQ_VTX_BASE_VTX_LOC */
-	radeon_ring_write(ring, 0xc0026f00);
-	radeon_ring_write(ring, 0x00000000);
-	radeon_ring_write(ring, 0x00000000);
-	radeon_ring_write(ring, 0x00000000);
+	radeon_ring_write(rdev, 0xc0026f00);
+	radeon_ring_write(rdev, 0x00000000);
+	radeon_ring_write(rdev, 0x00000000);
+	radeon_ring_write(rdev, 0x00000000);
 
 	/* SET_SAMPLER */
-	radeon_ring_write(ring, 0xc0036e00);
-	radeon_ring_write(ring, 0x00000000);
-	radeon_ring_write(ring, 0x00000012);
-	radeon_ring_write(ring, 0x00000000);
-	radeon_ring_write(ring, 0x00000000);
+	radeon_ring_write(rdev, 0xc0036e00);
+	radeon_ring_write(rdev, 0x00000000);
+	radeon_ring_write(rdev, 0x00000012);
+	radeon_ring_write(rdev, 0x00000000);
+	radeon_ring_write(rdev, 0x00000000);
 
 	/* set to DX10/11 mode */
-	radeon_ring_write(ring, PACKET3(PACKET3_MODE_CONTROL, 0));
-	radeon_ring_write(ring, 1);
+	radeon_ring_write(rdev, PACKET3(PACKET3_MODE_CONTROL, 0));
+	radeon_ring_write(rdev, 1);
 
 	/* emit an IB pointing at default state */
 	dwords = ALIGN(rdev->r600_blit.state_len, 0x10);
 	gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.state_offset;
-	radeon_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2));
-	radeon_ring_write(ring, gpu_addr & 0xFFFFFFFC);
-	radeon_ring_write(ring, upper_32_bits(gpu_addr) & 0xFF);
-	radeon_ring_write(ring, dwords);
+	radeon_ring_write(rdev, PACKET3(PACKET3_INDIRECT_BUFFER, 2));
+	radeon_ring_write(rdev, gpu_addr & 0xFFFFFFFC);
+	radeon_ring_write(rdev, upper_32_bits(gpu_addr) & 0xFF);
+	radeon_ring_write(rdev, dwords);
 
 }
 
+static inline uint32_t i2f(uint32_t input)
+{
+	u32 result, i, exponent, fraction;
+
+	if ((input & 0x3fff) == 0)
+		result = 0; /* 0 is a special case */
+	else {
+		exponent = 140; /* exponent biased by 127; */
+		fraction = (input & 0x3fff) << 10; /* cheat and only
+						      handle numbers below 2^^15 */
+		for (i = 0; i < 14; i++) {
+			if (fraction & 0x800000)
+				break;
+			else {
+				fraction = fraction << 1; /* keep
+							     shifting left until top bit = 1 */
+				exponent = exponent - 1;
+			}
+		}
+		result = exponent << 23 | (fraction & 0x7fffff); /* mask
+								    off top bit; assumed 1 */
+	}
+	return result;
+}
+
 int evergreen_blit_init(struct radeon_device *rdev)
 {
 	u32 obj_size;
@@ -613,26 +617,6 @@ int evergreen_blit_init(struct radeon_device *rdev)
 	u32 packet2s[16];
 	int num_packet2s = 0;
 
-	rdev->r600_blit.primitives.set_render_target = set_render_target;
-	rdev->r600_blit.primitives.cp_set_surface_sync = cp_set_surface_sync;
-	rdev->r600_blit.primitives.set_shaders = set_shaders;
-	rdev->r600_blit.primitives.set_vtx_resource = set_vtx_resource;
-	rdev->r600_blit.primitives.set_tex_resource = set_tex_resource;
-	rdev->r600_blit.primitives.set_scissors = set_scissors;
-	rdev->r600_blit.primitives.draw_auto = draw_auto;
-	rdev->r600_blit.primitives.set_default_state = set_default_state;
-
-	rdev->r600_blit.ring_size_common = 55; /* shaders + def state */
-	rdev->r600_blit.ring_size_common += 16; /* fence emit for VB IB */
-	rdev->r600_blit.ring_size_common += 5; /* done copy */
-	rdev->r600_blit.ring_size_common += 16; /* fence emit for done copy */
-
-	rdev->r600_blit.ring_size_per_loop = 74;
-	if (rdev->family >= CHIP_CAYMAN)
-		rdev->r600_blit.ring_size_per_loop += 9; /* additional DWs for surface sync */
-
-	rdev->r600_blit.max_dim = 16384;
-
 	/* pin copy shader into vram if already initialized */
 	if (rdev->r600_blit.shader_obj)
 		goto done;
@@ -728,3 +712,277 @@ done:
 	radeon_ttm_set_active_vram_size(rdev, rdev->mc.real_vram_size);
 	return 0;
 }
+
+void evergreen_blit_fini(struct radeon_device *rdev)
+{
+	int r;
+
+	radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size);
+	if (rdev->r600_blit.shader_obj == NULL)
+		return;
+	/* If we can't reserve the bo, unref should be enough to destroy
+	 * it when it becomes idle.
+	 */
+	r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false);
+	if (!r) {
+		radeon_bo_unpin(rdev->r600_blit.shader_obj);
+		radeon_bo_unreserve(rdev->r600_blit.shader_obj);
+	}
+	radeon_bo_unref(&rdev->r600_blit.shader_obj);
+}
+
+static int evergreen_vb_ib_get(struct radeon_device *rdev)
+{
+	int r;
+	r = radeon_ib_get(rdev, &rdev->r600_blit.vb_ib);
+	if (r) {
+		DRM_ERROR("failed to get IB for vertex buffer\n");
+		return r;
+	}
+
+	rdev->r600_blit.vb_total = 64*1024;
+	rdev->r600_blit.vb_used = 0;
+	return 0;
+}
+
+static void evergreen_vb_ib_put(struct radeon_device *rdev)
+{
+	radeon_fence_emit(rdev, rdev->r600_blit.vb_ib->fence);
+	radeon_ib_free(rdev, &rdev->r600_blit.vb_ib);
+}
+
+int evergreen_blit_prepare_copy(struct radeon_device *rdev, int size_bytes)
+{
+	int r;
+	int ring_size, line_size;
+	int max_size;
+	/* loops of emits + fence emit possible */
+	int dwords_per_loop = 74, num_loops;
+
+	r = evergreen_vb_ib_get(rdev);
+	if (r)
+		return r;
+
+	/* 8 bpp vs 32 bpp for xfer unit */
+	if (size_bytes & 3)
+		line_size = 8192;
+	else
+		line_size = 8192 * 4;
+
+	max_size = 8192 * line_size;
+
+	/* major loops cover the max size transfer */
+	num_loops = ((size_bytes + max_size) / max_size);
+	/* minor loops cover the extra non aligned bits */
+	num_loops += ((size_bytes % line_size) ? 1 : 0);
+	/* calculate number of loops correctly */
+	ring_size = num_loops * dwords_per_loop;
+	/* set default  + shaders */
+	ring_size += 55; /* shaders + def state */
+	ring_size += 10; /* fence emit for VB IB */
+	ring_size += 5; /* done copy */
+	ring_size += 10; /* fence emit for done copy */
+	r = radeon_ring_lock(rdev, ring_size);
+	if (r)
+		return r;
+
+	set_default_state(rdev); /* 36 */
+	set_shaders(rdev); /* 16 */
+	return 0;
+}
+
+void evergreen_blit_done_copy(struct radeon_device *rdev, struct radeon_fence *fence)
+{
+	int r;
+
+	if (rdev->r600_blit.vb_ib)
+		evergreen_vb_ib_put(rdev);
+
+	if (fence)
+		r = radeon_fence_emit(rdev, fence);
+
+	radeon_ring_unlock_commit(rdev);
+}
+
+void evergreen_kms_blit_copy(struct radeon_device *rdev,
+			     u64 src_gpu_addr, u64 dst_gpu_addr,
+			     int size_bytes)
+{
+	int max_bytes;
+	u64 vb_gpu_addr;
+	u32 *vb;
+
+	DRM_DEBUG("emitting copy %16llx %16llx %d %d\n", src_gpu_addr, dst_gpu_addr,
+		  size_bytes, rdev->r600_blit.vb_used);
+	vb = (u32 *)(rdev->r600_blit.vb_ib->ptr + rdev->r600_blit.vb_used);
+	if ((size_bytes & 3) || (src_gpu_addr & 3) || (dst_gpu_addr & 3)) {
+		max_bytes = 8192;
+
+		while (size_bytes) {
+			int cur_size = size_bytes;
+			int src_x = src_gpu_addr & 255;
+			int dst_x = dst_gpu_addr & 255;
+			int h = 1;
+			src_gpu_addr = src_gpu_addr & ~255ULL;
+			dst_gpu_addr = dst_gpu_addr & ~255ULL;
+
+			if (!src_x && !dst_x) {
+				h = (cur_size / max_bytes);
+				if (h > 8192)
+					h = 8192;
+				if (h == 0)
+					h = 1;
+				else
+					cur_size = max_bytes;
+			} else {
+				if (cur_size > max_bytes)
+					cur_size = max_bytes;
+				if (cur_size > (max_bytes - dst_x))
+					cur_size = (max_bytes - dst_x);
+				if (cur_size > (max_bytes - src_x))
+					cur_size = (max_bytes - src_x);
+			}
+
+			if ((rdev->r600_blit.vb_used + 48) > rdev->r600_blit.vb_total) {
+				WARN_ON(1);
+			}
+
+			vb[0] = i2f(dst_x);
+			vb[1] = 0;
+			vb[2] = i2f(src_x);
+			vb[3] = 0;
+
+			vb[4] = i2f(dst_x);
+			vb[5] = i2f(h);
+			vb[6] = i2f(src_x);
+			vb[7] = i2f(h);
+
+			vb[8] = i2f(dst_x + cur_size);
+			vb[9] = i2f(h);
+			vb[10] = i2f(src_x + cur_size);
+			vb[11] = i2f(h);
+
+			/* src 10 */
+			set_tex_resource(rdev, FMT_8,
+					 src_x + cur_size, h, src_x + cur_size,
+					 src_gpu_addr);
+
+			/* 5 */
+			cp_set_surface_sync(rdev,
+					    PACKET3_TC_ACTION_ENA, (src_x + cur_size * h), src_gpu_addr);
+
+
+			/* dst 17 */
+			set_render_target(rdev, COLOR_8,
+					  dst_x + cur_size, h,
+					  dst_gpu_addr);
+
+			/* scissors 12 */
+			set_scissors(rdev, dst_x, 0, dst_x + cur_size, h);
+
+			/* 15 */
+			vb_gpu_addr = rdev->r600_blit.vb_ib->gpu_addr + rdev->r600_blit.vb_used;
+			set_vtx_resource(rdev, vb_gpu_addr);
+
+			/* draw 10 */
+			draw_auto(rdev);
+
+			/* 5 */
+			cp_set_surface_sync(rdev,
+					    PACKET3_CB_ACTION_ENA | PACKET3_CB0_DEST_BASE_ENA,
+					    cur_size * h, dst_gpu_addr);
+
+			vb += 12;
+			rdev->r600_blit.vb_used += 12 * 4;
+
+			src_gpu_addr += cur_size * h;
+			dst_gpu_addr += cur_size * h;
+			size_bytes -= cur_size * h;
+		}
+	} else {
+		max_bytes = 8192 * 4;
+
+		while (size_bytes) {
+			int cur_size = size_bytes;
+			int src_x = (src_gpu_addr & 255);
+			int dst_x = (dst_gpu_addr & 255);
+			int h = 1;
+			src_gpu_addr = src_gpu_addr & ~255ULL;
+			dst_gpu_addr = dst_gpu_addr & ~255ULL;
+
+			if (!src_x && !dst_x) {
+				h = (cur_size / max_bytes);
+				if (h > 8192)
+					h = 8192;
+				if (h == 0)
+					h = 1;
+				else
+					cur_size = max_bytes;
+			} else {
+				if (cur_size > max_bytes)
+					cur_size = max_bytes;
+				if (cur_size > (max_bytes - dst_x))
+					cur_size = (max_bytes - dst_x);
+				if (cur_size > (max_bytes - src_x))
+					cur_size = (max_bytes - src_x);
+			}
+
+			if ((rdev->r600_blit.vb_used + 48) > rdev->r600_blit.vb_total) {
+				WARN_ON(1);
+			}
+
+			vb[0] = i2f(dst_x / 4);
+			vb[1] = 0;
+			vb[2] = i2f(src_x / 4);
+			vb[3] = 0;
+
+			vb[4] = i2f(dst_x / 4);
+			vb[5] = i2f(h);
+			vb[6] = i2f(src_x / 4);
+			vb[7] = i2f(h);
+
+			vb[8] = i2f((dst_x + cur_size) / 4);
+			vb[9] = i2f(h);
+			vb[10] = i2f((src_x + cur_size) / 4);
+			vb[11] = i2f(h);
+
+			/* src 10 */
+			set_tex_resource(rdev, FMT_8_8_8_8,
+					 (src_x + cur_size) / 4,
+					 h, (src_x + cur_size) / 4,
+					 src_gpu_addr);
+			/* 5 */
+			cp_set_surface_sync(rdev,
+					    PACKET3_TC_ACTION_ENA, (src_x + cur_size * h), src_gpu_addr);
+
+			/* dst 17 */
+			set_render_target(rdev, COLOR_8_8_8_8,
+					  (dst_x + cur_size) / 4, h,
+					  dst_gpu_addr);
+
+			/* scissors 12  */
+			set_scissors(rdev, (dst_x / 4), 0, (dst_x + cur_size / 4), h);
+
+			/* Vertex buffer setup 15 */
+			vb_gpu_addr = rdev->r600_blit.vb_ib->gpu_addr + rdev->r600_blit.vb_used;
+			set_vtx_resource(rdev, vb_gpu_addr);
+
+			/* draw 10 */
+			draw_auto(rdev);
+
+			/* 5 */
+			cp_set_surface_sync(rdev,
+					    PACKET3_CB_ACTION_ENA | PACKET3_CB0_DEST_BASE_ENA,
+					    cur_size * h, dst_gpu_addr);
+
+			/* 74 ring dwords per loop */
+			vb += 12;
+			rdev->r600_blit.vb_used += 12 * 4;
+
+			src_gpu_addr += cur_size * h;
+			dst_gpu_addr += cur_size * h;
+			size_bytes -= cur_size * h;
+		}
+	}
+}
+
diff --git a/drivers/gpu/drm/radeon/evergreen_cs.c b/drivers/gpu/drm/radeon/evergreen_cs.c
index a58b37a..23d3641 100644
--- a/drivers/gpu/drm/radeon/evergreen_cs.c
+++ b/drivers/gpu/drm/radeon/evergreen_cs.c
@@ -31,9 +31,6 @@
 #include "evergreen_reg_safe.h"
 #include "cayman_reg_safe.h"
 
-#define MAX(a,b)                   (((a)>(b))?(a):(b))
-#define MIN(a,b)                   (((a)<(b))?(a):(b))
-
 static int evergreen_cs_packet_next_reloc(struct radeon_cs_parser *p,
 					  struct radeon_cs_reloc **cs_reloc);
 
@@ -41,72 +38,45 @@ struct evergreen_cs_track {
 	u32			group_size;
 	u32			nbanks;
 	u32			npipes;
-	u32			row_size;
 	/* value we track */
-	u32			nsamples;		/* unused */
+	u32			nsamples;
+	u32			cb_color_base_last[12];
 	struct radeon_bo	*cb_color_bo[12];
 	u32			cb_color_bo_offset[12];
-	struct radeon_bo	*cb_color_fmask_bo[8];	/* unused */
-	struct radeon_bo	*cb_color_cmask_bo[8];	/* unused */
+	struct radeon_bo	*cb_color_fmask_bo[8];
+	struct radeon_bo	*cb_color_cmask_bo[8];
 	u32			cb_color_info[12];
 	u32			cb_color_view[12];
+	u32			cb_color_pitch_idx[12];
+	u32			cb_color_slice_idx[12];
+	u32			cb_color_dim_idx[12];
+	u32			cb_color_dim[12];
 	u32			cb_color_pitch[12];
 	u32			cb_color_slice[12];
-	u32			cb_color_attrib[12];
-	u32			cb_color_cmask_slice[8];/* unused */
-	u32			cb_color_fmask_slice[8];/* unused */
+	u32			cb_color_cmask_slice[8];
+	u32			cb_color_fmask_slice[8];
 	u32			cb_target_mask;
-	u32			cb_shader_mask; /* unused */
+	u32			cb_shader_mask;
 	u32			vgt_strmout_config;
 	u32			vgt_strmout_buffer_config;
-	struct radeon_bo	*vgt_strmout_bo[4];
-	u32			vgt_strmout_bo_offset[4];
-	u32			vgt_strmout_size[4];
 	u32			db_depth_control;
 	u32			db_depth_view;
-	u32			db_depth_slice;
 	u32			db_depth_size;
+	u32			db_depth_size_idx;
 	u32			db_z_info;
+	u32			db_z_idx;
 	u32			db_z_read_offset;
 	u32			db_z_write_offset;
 	struct radeon_bo	*db_z_read_bo;
 	struct radeon_bo	*db_z_write_bo;
 	u32			db_s_info;
+	u32			db_s_idx;
 	u32			db_s_read_offset;
 	u32			db_s_write_offset;
 	struct radeon_bo	*db_s_read_bo;
 	struct radeon_bo	*db_s_write_bo;
-	bool			sx_misc_kill_all_prims;
-	bool			cb_dirty;
-	bool			db_dirty;
-	bool			streamout_dirty;
 };
 
-static u32 evergreen_cs_get_aray_mode(u32 tiling_flags)
-{
-	if (tiling_flags & RADEON_TILING_MACRO)
-		return ARRAY_2D_TILED_THIN1;
-	else if (tiling_flags & RADEON_TILING_MICRO)
-		return ARRAY_1D_TILED_THIN1;
-	else
-		return ARRAY_LINEAR_GENERAL;
-}
-
-static u32 evergreen_cs_get_num_banks(u32 nbanks)
-{
-	switch (nbanks) {
-	case 2:
-		return ADDR_SURF_2_BANK;
-	case 4:
-		return ADDR_SURF_4_BANK;
-	case 8:
-	default:
-		return ADDR_SURF_8_BANK;
-	case 16:
-		return ADDR_SURF_16_BANK;
-	}
-}
-
 static void evergreen_cs_track_init(struct evergreen_cs_track *track)
 {
 	int i;
@@ -119,745 +89,56 @@ static void evergreen_cs_track_init(struct evergreen_cs_track *track)
 	}
 
 	for (i = 0; i < 12; i++) {
+		track->cb_color_base_last[i] = 0;
 		track->cb_color_bo[i] = NULL;
 		track->cb_color_bo_offset[i] = 0xFFFFFFFF;
 		track->cb_color_info[i] = 0;
-		track->cb_color_view[i] = 0xFFFFFFFF;
+		track->cb_color_view[i] = 0;
+		track->cb_color_pitch_idx[i] = 0;
+		track->cb_color_slice_idx[i] = 0;
+		track->cb_color_dim[i] = 0;
 		track->cb_color_pitch[i] = 0;
 		track->cb_color_slice[i] = 0;
+		track->cb_color_dim[i] = 0;
 	}
 	track->cb_target_mask = 0xFFFFFFFF;
 	track->cb_shader_mask = 0xFFFFFFFF;
-	track->cb_dirty = true;
 
 	track->db_depth_view = 0xFFFFC000;
 	track->db_depth_size = 0xFFFFFFFF;
+	track->db_depth_size_idx = 0;
 	track->db_depth_control = 0xFFFFFFFF;
 	track->db_z_info = 0xFFFFFFFF;
+	track->db_z_idx = 0xFFFFFFFF;
 	track->db_z_read_offset = 0xFFFFFFFF;
 	track->db_z_write_offset = 0xFFFFFFFF;
 	track->db_z_read_bo = NULL;
 	track->db_z_write_bo = NULL;
 	track->db_s_info = 0xFFFFFFFF;
+	track->db_s_idx = 0xFFFFFFFF;
 	track->db_s_read_offset = 0xFFFFFFFF;
 	track->db_s_write_offset = 0xFFFFFFFF;
 	track->db_s_read_bo = NULL;
 	track->db_s_write_bo = NULL;
-	track->db_dirty = true;
-
-	for (i = 0; i < 4; i++) {
-		track->vgt_strmout_size[i] = 0;
-		track->vgt_strmout_bo[i] = NULL;
-		track->vgt_strmout_bo_offset[i] = 0xFFFFFFFF;
-	}
-	track->streamout_dirty = true;
-	track->sx_misc_kill_all_prims = false;
-}
-
-struct eg_surface {
-	/* value gathered from cs */
-	unsigned	nbx;
-	unsigned	nby;
-	unsigned	format;
-	unsigned	mode;
-	unsigned	nbanks;
-	unsigned	bankw;
-	unsigned	bankh;
-	unsigned	tsplit;
-	unsigned	mtilea;
-	unsigned	nsamples;
-	/* output value */
-	unsigned	bpe;
-	unsigned	layer_size;
-	unsigned	palign;
-	unsigned	halign;
-	unsigned long	base_align;
-};
-
-static int evergreen_surface_check_linear(struct radeon_cs_parser *p,
-					  struct eg_surface *surf,
-					  const char *prefix)
-{
-	surf->layer_size = surf->nbx * surf->nby * surf->bpe * surf->nsamples;
-	surf->base_align = surf->bpe;
-	surf->palign = 1;
-	surf->halign = 1;
-	return 0;
-}
-
-static int evergreen_surface_check_linear_aligned(struct radeon_cs_parser *p,
-						  struct eg_surface *surf,
-						  const char *prefix)
-{
-	struct evergreen_cs_track *track = p->track;
-	unsigned palign;
-
-	palign = MAX(64, track->group_size / surf->bpe);
-	surf->layer_size = surf->nbx * surf->nby * surf->bpe * surf->nsamples;
-	surf->base_align = track->group_size;
-	surf->palign = palign;
-	surf->halign = 1;
-	if (surf->nbx & (palign - 1)) {
-		if (prefix) {
-			dev_warn(p->dev, "%s:%d %s pitch %d invalid must be aligned with %d\n",
-				 __func__, __LINE__, prefix, surf->nbx, palign);
-		}
-		return -EINVAL;
-	}
-	return 0;
-}
-
-static int evergreen_surface_check_1d(struct radeon_cs_parser *p,
-				      struct eg_surface *surf,
-				      const char *prefix)
-{
-	struct evergreen_cs_track *track = p->track;
-	unsigned palign;
-
-	palign = track->group_size / (8 * surf->bpe * surf->nsamples);
-	palign = MAX(8, palign);
-	surf->layer_size = surf->nbx * surf->nby * surf->bpe;
-	surf->base_align = track->group_size;
-	surf->palign = palign;
-	surf->halign = 8;
-	if ((surf->nbx & (palign - 1))) {
-		if (prefix) {
-			dev_warn(p->dev, "%s:%d %s pitch %d invalid must be aligned with %d (%d %d %d)\n",
-				 __func__, __LINE__, prefix, surf->nbx, palign,
-				 track->group_size, surf->bpe, surf->nsamples);
-		}
-		return -EINVAL;
-	}
-	if ((surf->nby & (8 - 1))) {
-		if (prefix) {
-			dev_warn(p->dev, "%s:%d %s height %d invalid must be aligned with 8\n",
-				 __func__, __LINE__, prefix, surf->nby);
-		}
-		return -EINVAL;
-	}
-	return 0;
-}
-
-static int evergreen_surface_check_2d(struct radeon_cs_parser *p,
-				      struct eg_surface *surf,
-				      const char *prefix)
-{
-	struct evergreen_cs_track *track = p->track;
-	unsigned palign, halign, tileb, slice_pt;
-
-	tileb = 64 * surf->bpe * surf->nsamples;
-	palign = track->group_size / (8 * surf->bpe * surf->nsamples);
-	palign = MAX(8, palign);
-	slice_pt = 1;
-	if (tileb > surf->tsplit) {
-		slice_pt = tileb / surf->tsplit;
-	}
-	tileb = tileb / slice_pt;
-	/* macro tile width & height */
-	palign = (8 * surf->bankw * track->npipes) * surf->mtilea;
-	halign = (8 * surf->bankh * surf->nbanks) / surf->mtilea;
-	surf->layer_size = surf->nbx * surf->nby * surf->bpe * slice_pt;
-	surf->base_align = (palign / 8) * (halign / 8) * tileb;
-	surf->palign = palign;
-	surf->halign = halign;
-
-	if ((surf->nbx & (palign - 1))) {
-		if (prefix) {
-			dev_warn(p->dev, "%s:%d %s pitch %d invalid must be aligned with %d\n",
-				 __func__, __LINE__, prefix, surf->nbx, palign);
-		}
-		return -EINVAL;
-	}
-	if ((surf->nby & (halign - 1))) {
-		if (prefix) {
-			dev_warn(p->dev, "%s:%d %s height %d invalid must be aligned with %d\n",
-				 __func__, __LINE__, prefix, surf->nby, halign);
-		}
-		return -EINVAL;
-	}
-
-	return 0;
-}
-
-static int evergreen_surface_check(struct radeon_cs_parser *p,
-				   struct eg_surface *surf,
-				   const char *prefix)
-{
-	/* some common value computed here */
-	surf->bpe = r600_fmt_get_blocksize(surf->format);
-
-	switch (surf->mode) {
-	case ARRAY_LINEAR_GENERAL:
-		return evergreen_surface_check_linear(p, surf, prefix);
-	case ARRAY_LINEAR_ALIGNED:
-		return evergreen_surface_check_linear_aligned(p, surf, prefix);
-	case ARRAY_1D_TILED_THIN1:
-		return evergreen_surface_check_1d(p, surf, prefix);
-	case ARRAY_2D_TILED_THIN1:
-		return evergreen_surface_check_2d(p, surf, prefix);
-	default:
-		dev_warn(p->dev, "%s:%d %s invalid array mode %d\n",
-				__func__, __LINE__, prefix, surf->mode);
-		return -EINVAL;
-	}
-	return -EINVAL;
-}
-
-static int evergreen_surface_value_conv_check(struct radeon_cs_parser *p,
-					      struct eg_surface *surf,
-					      const char *prefix)
-{
-	switch (surf->mode) {
-	case ARRAY_2D_TILED_THIN1:
-		break;
-	case ARRAY_LINEAR_GENERAL:
-	case ARRAY_LINEAR_ALIGNED:
-	case ARRAY_1D_TILED_THIN1:
-		return 0;
-	default:
-		dev_warn(p->dev, "%s:%d %s invalid array mode %d\n",
-				__func__, __LINE__, prefix, surf->mode);
-		return -EINVAL;
-	}
-
-	switch (surf->nbanks) {
-	case 0: surf->nbanks = 2; break;
-	case 1: surf->nbanks = 4; break;
-	case 2: surf->nbanks = 8; break;
-	case 3: surf->nbanks = 16; break;
-	default:
-		dev_warn(p->dev, "%s:%d %s invalid number of banks %d\n",
-			 __func__, __LINE__, prefix, surf->nbanks);
-		return -EINVAL;
-	}
-	switch (surf->bankw) {
-	case 0: surf->bankw = 1; break;
-	case 1: surf->bankw = 2; break;
-	case 2: surf->bankw = 4; break;
-	case 3: surf->bankw = 8; break;
-	default:
-		dev_warn(p->dev, "%s:%d %s invalid bankw %d\n",
-			 __func__, __LINE__, prefix, surf->bankw);
-		return -EINVAL;
-	}
-	switch (surf->bankh) {
-	case 0: surf->bankh = 1; break;
-	case 1: surf->bankh = 2; break;
-	case 2: surf->bankh = 4; break;
-	case 3: surf->bankh = 8; break;
-	default:
-		dev_warn(p->dev, "%s:%d %s invalid bankh %d\n",
-			 __func__, __LINE__, prefix, surf->bankh);
-		return -EINVAL;
-	}
-	switch (surf->mtilea) {
-	case 0: surf->mtilea = 1; break;
-	case 1: surf->mtilea = 2; break;
-	case 2: surf->mtilea = 4; break;
-	case 3: surf->mtilea = 8; break;
-	default:
-		dev_warn(p->dev, "%s:%d %s invalid macro tile aspect %d\n",
-			 __func__, __LINE__, prefix, surf->mtilea);
-		return -EINVAL;
-	}
-	switch (surf->tsplit) {
-	case 0: surf->tsplit = 64; break;
-	case 1: surf->tsplit = 128; break;
-	case 2: surf->tsplit = 256; break;
-	case 3: surf->tsplit = 512; break;
-	case 4: surf->tsplit = 1024; break;
-	case 5: surf->tsplit = 2048; break;
-	case 6: surf->tsplit = 4096; break;
-	default:
-		dev_warn(p->dev, "%s:%d %s invalid tile split %d\n",
-			 __func__, __LINE__, prefix, surf->tsplit);
-		return -EINVAL;
-	}
-	return 0;
 }
 
-static int evergreen_cs_track_validate_cb(struct radeon_cs_parser *p, unsigned id)
+static inline int evergreen_cs_track_validate_cb(struct radeon_cs_parser *p, int i)
 {
-	struct evergreen_cs_track *track = p->track;
-	struct eg_surface surf;
-	unsigned pitch, slice, mslice;
-	unsigned long offset;
-	int r;
-
-	mslice = G_028C6C_SLICE_MAX(track->cb_color_view[id]) + 1;
-	pitch = track->cb_color_pitch[id];
-	slice = track->cb_color_slice[id];
-	surf.nbx = (pitch + 1) * 8;
-	surf.nby = ((slice + 1) * 64) / surf.nbx;
-	surf.mode = G_028C70_ARRAY_MODE(track->cb_color_info[id]);
-	surf.format = G_028C70_FORMAT(track->cb_color_info[id]);
-	surf.tsplit = G_028C74_TILE_SPLIT(track->cb_color_attrib[id]);
-	surf.nbanks = G_028C74_NUM_BANKS(track->cb_color_attrib[id]);
-	surf.bankw = G_028C74_BANK_WIDTH(track->cb_color_attrib[id]);
-	surf.bankh = G_028C74_BANK_HEIGHT(track->cb_color_attrib[id]);
-	surf.mtilea = G_028C74_MACRO_TILE_ASPECT(track->cb_color_attrib[id]);
-	surf.nsamples = 1;
-
-	if (!r600_fmt_is_valid_color(surf.format)) {
-		dev_warn(p->dev, "%s:%d cb invalid format %d for %d (0x%08x)\n",
-			 __func__, __LINE__, surf.format,
-			id, track->cb_color_info[id]);
-		return -EINVAL;
-	}
-
-	r = evergreen_surface_value_conv_check(p, &surf, "cb");
-	if (r) {
-		return r;
-	}
-
-	r = evergreen_surface_check(p, &surf, "cb");
-	if (r) {
-		dev_warn(p->dev, "%s:%d cb[%d] invalid (0x%08x 0x%08x 0x%08x 0x%08x)\n",
-			 __func__, __LINE__, id, track->cb_color_pitch[id],
-			 track->cb_color_slice[id], track->cb_color_attrib[id],
-			 track->cb_color_info[id]);
-		return r;
-	}
-
-	offset = track->cb_color_bo_offset[id] << 8;
-	if (offset & (surf.base_align - 1)) {
-		dev_warn(p->dev, "%s:%d cb[%d] bo base %ld not aligned with %ld\n",
-			 __func__, __LINE__, id, offset, surf.base_align);
-		return -EINVAL;
-	}
-
-	offset += surf.layer_size * mslice;
-	if (offset > radeon_bo_size(track->cb_color_bo[id])) {
-		dev_warn(p->dev, "%s:%d cb[%d] bo too small (layer size %d, "
-			 "offset %d, max layer %d, bo size %ld, slice %d)\n",
-			 __func__, __LINE__, id, surf.layer_size,
-			track->cb_color_bo_offset[id] << 8, mslice,
-			radeon_bo_size(track->cb_color_bo[id]), slice);
-		dev_warn(p->dev, "%s:%d problematic surf: (%d %d) (%d %d %d %d %d %d %d)\n",
-			 __func__, __LINE__, surf.nbx, surf.nby,
-			surf.mode, surf.bpe, surf.nsamples,
-			surf.bankw, surf.bankh,
-			surf.tsplit, surf.mtilea);
-		return -EINVAL;
-	}
-
-	return 0;
-}
-
-static int evergreen_cs_track_validate_stencil(struct radeon_cs_parser *p)
-{
-	struct evergreen_cs_track *track = p->track;
-	struct eg_surface surf;
-	unsigned pitch, slice, mslice;
-	unsigned long offset;
-	int r;
-
-	mslice = G_028008_SLICE_MAX(track->db_depth_view) + 1;
-	pitch = G_028058_PITCH_TILE_MAX(track->db_depth_size);
-	slice = track->db_depth_slice;
-	surf.nbx = (pitch + 1) * 8;
-	surf.nby = ((slice + 1) * 64) / surf.nbx;
-	surf.mode = G_028040_ARRAY_MODE(track->db_z_info);
-	surf.format = G_028044_FORMAT(track->db_s_info);
-	surf.tsplit = G_028044_TILE_SPLIT(track->db_s_info);
-	surf.nbanks = G_028040_NUM_BANKS(track->db_z_info);
-	surf.bankw = G_028040_BANK_WIDTH(track->db_z_info);
-	surf.bankh = G_028040_BANK_HEIGHT(track->db_z_info);
-	surf.mtilea = G_028040_MACRO_TILE_ASPECT(track->db_z_info);
-	surf.nsamples = 1;
-
-	if (surf.format != 1) {
-		dev_warn(p->dev, "%s:%d stencil invalid format %d\n",
-			 __func__, __LINE__, surf.format);
-		return -EINVAL;
-	}
-	/* replace by color format so we can use same code */
-	surf.format = V_028C70_COLOR_8;
-
-	r = evergreen_surface_value_conv_check(p, &surf, "stencil");
-	if (r) {
-		return r;
-	}
-
-	r = evergreen_surface_check(p, &surf, NULL);
-	if (r) {
-		/* old userspace doesn't compute proper depth/stencil alignment
-		 * check that alignment against a bigger byte per elements and
-		 * only report if that alignment is wrong too.
-		 */
-		surf.format = V_028C70_COLOR_8_8_8_8;
-		r = evergreen_surface_check(p, &surf, "stencil");
-		if (r) {
-			dev_warn(p->dev, "%s:%d stencil invalid (0x%08x 0x%08x 0x%08x 0x%08x)\n",
-				 __func__, __LINE__, track->db_depth_size,
-				 track->db_depth_slice, track->db_s_info, track->db_z_info);
-		}
-		return r;
-	}
-
-	offset = track->db_s_read_offset << 8;
-	if (offset & (surf.base_align - 1)) {
-		dev_warn(p->dev, "%s:%d stencil read bo base %ld not aligned with %ld\n",
-			 __func__, __LINE__, offset, surf.base_align);
-		return -EINVAL;
-	}
-	offset += surf.layer_size * mslice;
-	if (offset > radeon_bo_size(track->db_s_read_bo)) {
-		dev_warn(p->dev, "%s:%d stencil read bo too small (layer size %d, "
-			 "offset %ld, max layer %d, bo size %ld)\n",
-			 __func__, __LINE__, surf.layer_size,
-			(unsigned long)track->db_s_read_offset << 8, mslice,
-			radeon_bo_size(track->db_s_read_bo));
-		dev_warn(p->dev, "%s:%d stencil invalid (0x%08x 0x%08x 0x%08x 0x%08x)\n",
-			 __func__, __LINE__, track->db_depth_size,
-			 track->db_depth_slice, track->db_s_info, track->db_z_info);
-		return -EINVAL;
-	}
-
-	offset = track->db_s_write_offset << 8;
-	if (offset & (surf.base_align - 1)) {
-		dev_warn(p->dev, "%s:%d stencil write bo base %ld not aligned with %ld\n",
-			 __func__, __LINE__, offset, surf.base_align);
-		return -EINVAL;
-	}
-	offset += surf.layer_size * mslice;
-	if (offset > radeon_bo_size(track->db_s_write_bo)) {
-		dev_warn(p->dev, "%s:%d stencil write bo too small (layer size %d, "
-			 "offset %ld, max layer %d, bo size %ld)\n",
-			 __func__, __LINE__, surf.layer_size,
-			(unsigned long)track->db_s_write_offset << 8, mslice,
-			radeon_bo_size(track->db_s_write_bo));
-		return -EINVAL;
-	}
-
-	return 0;
-}
-
-static int evergreen_cs_track_validate_depth(struct radeon_cs_parser *p)
-{
-	struct evergreen_cs_track *track = p->track;
-	struct eg_surface surf;
-	unsigned pitch, slice, mslice;
-	unsigned long offset;
-	int r;
-
-	mslice = G_028008_SLICE_MAX(track->db_depth_view) + 1;
-	pitch = G_028058_PITCH_TILE_MAX(track->db_depth_size);
-	slice = track->db_depth_slice;
-	surf.nbx = (pitch + 1) * 8;
-	surf.nby = ((slice + 1) * 64) / surf.nbx;
-	surf.mode = G_028040_ARRAY_MODE(track->db_z_info);
-	surf.format = G_028040_FORMAT(track->db_z_info);
-	surf.tsplit = G_028040_TILE_SPLIT(track->db_z_info);
-	surf.nbanks = G_028040_NUM_BANKS(track->db_z_info);
-	surf.bankw = G_028040_BANK_WIDTH(track->db_z_info);
-	surf.bankh = G_028040_BANK_HEIGHT(track->db_z_info);
-	surf.mtilea = G_028040_MACRO_TILE_ASPECT(track->db_z_info);
-	surf.nsamples = 1;
-
-	switch (surf.format) {
-	case V_028040_Z_16:
-		surf.format = V_028C70_COLOR_16;
-		break;
-	case V_028040_Z_24:
-	case V_028040_Z_32_FLOAT:
-		surf.format = V_028C70_COLOR_8_8_8_8;
-		break;
-	default:
-		dev_warn(p->dev, "%s:%d depth invalid format %d\n",
-			 __func__, __LINE__, surf.format);
-		return -EINVAL;
-	}
-
-	r = evergreen_surface_value_conv_check(p, &surf, "depth");
-	if (r) {
-		dev_warn(p->dev, "%s:%d depth invalid (0x%08x 0x%08x 0x%08x)\n",
-			 __func__, __LINE__, track->db_depth_size,
-			 track->db_depth_slice, track->db_z_info);
-		return r;
-	}
-
-	r = evergreen_surface_check(p, &surf, "depth");
-	if (r) {
-		dev_warn(p->dev, "%s:%d depth invalid (0x%08x 0x%08x 0x%08x)\n",
-			 __func__, __LINE__, track->db_depth_size,
-			 track->db_depth_slice, track->db_z_info);
-		return r;
-	}
-
-	offset = track->db_z_read_offset << 8;
-	if (offset & (surf.base_align - 1)) {
-		dev_warn(p->dev, "%s:%d stencil read bo base %ld not aligned with %ld\n",
-			 __func__, __LINE__, offset, surf.base_align);
-		return -EINVAL;
-	}
-	offset += surf.layer_size * mslice;
-	if (offset > radeon_bo_size(track->db_z_read_bo)) {
-		dev_warn(p->dev, "%s:%d depth read bo too small (layer size %d, "
-			 "offset %ld, max layer %d, bo size %ld)\n",
-			 __func__, __LINE__, surf.layer_size,
-			(unsigned long)track->db_z_read_offset << 8, mslice,
-			radeon_bo_size(track->db_z_read_bo));
-		return -EINVAL;
-	}
-
-	offset = track->db_z_write_offset << 8;
-	if (offset & (surf.base_align - 1)) {
-		dev_warn(p->dev, "%s:%d stencil write bo base %ld not aligned with %ld\n",
-			 __func__, __LINE__, offset, surf.base_align);
-		return -EINVAL;
-	}
-	offset += surf.layer_size * mslice;
-	if (offset > radeon_bo_size(track->db_z_write_bo)) {
-		dev_warn(p->dev, "%s:%d depth write bo too small (layer size %d, "
-			 "offset %ld, max layer %d, bo size %ld)\n",
-			 __func__, __LINE__, surf.layer_size,
-			(unsigned long)track->db_z_write_offset << 8, mslice,
-			radeon_bo_size(track->db_z_write_bo));
-		return -EINVAL;
-	}
-
-	return 0;
-}
-
-static int evergreen_cs_track_validate_texture(struct radeon_cs_parser *p,
-					       struct radeon_bo *texture,
-					       struct radeon_bo *mipmap,
-					       unsigned idx)
-{
-	struct eg_surface surf;
-	unsigned long toffset, moffset;
-	unsigned dim, llevel, mslice, width, height, depth, i;
-	u32 texdw[8];
-	int r;
-
-	texdw[0] = radeon_get_ib_value(p, idx + 0);
-	texdw[1] = radeon_get_ib_value(p, idx + 1);
-	texdw[2] = radeon_get_ib_value(p, idx + 2);
-	texdw[3] = radeon_get_ib_value(p, idx + 3);
-	texdw[4] = radeon_get_ib_value(p, idx + 4);
-	texdw[5] = radeon_get_ib_value(p, idx + 5);
-	texdw[6] = radeon_get_ib_value(p, idx + 6);
-	texdw[7] = radeon_get_ib_value(p, idx + 7);
-	dim = G_030000_DIM(texdw[0]);
-	llevel = G_030014_LAST_LEVEL(texdw[5]);
-	mslice = G_030014_LAST_ARRAY(texdw[5]) + 1;
-	width = G_030000_TEX_WIDTH(texdw[0]) + 1;
-	height =  G_030004_TEX_HEIGHT(texdw[1]) + 1;
-	depth = G_030004_TEX_DEPTH(texdw[1]) + 1;
-	surf.format = G_03001C_DATA_FORMAT(texdw[7]);
-	surf.nbx = (G_030000_PITCH(texdw[0]) + 1) * 8;
-	surf.nbx = r600_fmt_get_nblocksx(surf.format, surf.nbx);
-	surf.nby = r600_fmt_get_nblocksy(surf.format, height);
-	surf.mode = G_030004_ARRAY_MODE(texdw[1]);
-	surf.tsplit = G_030018_TILE_SPLIT(texdw[6]);
-	surf.nbanks = G_03001C_NUM_BANKS(texdw[7]);
-	surf.bankw = G_03001C_BANK_WIDTH(texdw[7]);
-	surf.bankh = G_03001C_BANK_HEIGHT(texdw[7]);
-	surf.mtilea = G_03001C_MACRO_TILE_ASPECT(texdw[7]);
-	surf.nsamples = 1;
-	toffset = texdw[2] << 8;
-	moffset = texdw[3] << 8;
-
-	if (!r600_fmt_is_valid_texture(surf.format, p->family)) {
-		dev_warn(p->dev, "%s:%d texture invalid format %d\n",
-			 __func__, __LINE__, surf.format);
-		return -EINVAL;
-	}
-	switch (dim) {
-	case V_030000_SQ_TEX_DIM_1D:
-	case V_030000_SQ_TEX_DIM_2D:
-	case V_030000_SQ_TEX_DIM_CUBEMAP:
-	case V_030000_SQ_TEX_DIM_1D_ARRAY:
-	case V_030000_SQ_TEX_DIM_2D_ARRAY:
-		depth = 1;
-	case V_030000_SQ_TEX_DIM_3D:
-		break;
-	default:
-		dev_warn(p->dev, "%s:%d texture invalid dimension %d\n",
-			 __func__, __LINE__, dim);
-		return -EINVAL;
-	}
-
-	r = evergreen_surface_value_conv_check(p, &surf, "texture");
-	if (r) {
-		return r;
-	}
-
-	/* align height */
-	evergreen_surface_check(p, &surf, NULL);
-	surf.nby = ALIGN(surf.nby, surf.halign);
-
-	r = evergreen_surface_check(p, &surf, "texture");
-	if (r) {
-		dev_warn(p->dev, "%s:%d texture invalid 0x%08x 0x%08x 0x%08x 0x%08x 0x%08x 0x%08x\n",
-			 __func__, __LINE__, texdw[0], texdw[1], texdw[4],
-			 texdw[5], texdw[6], texdw[7]);
-		return r;
-	}
-
-	/* check texture size */
-	if (toffset & (surf.base_align - 1)) {
-		dev_warn(p->dev, "%s:%d texture bo base %ld not aligned with %ld\n",
-			 __func__, __LINE__, toffset, surf.base_align);
-		return -EINVAL;
-	}
-	if (moffset & (surf.base_align - 1)) {
-		dev_warn(p->dev, "%s:%d mipmap bo base %ld not aligned with %ld\n",
-			 __func__, __LINE__, moffset, surf.base_align);
-		return -EINVAL;
-	}
-	if (dim == SQ_TEX_DIM_3D) {
-		toffset += surf.layer_size * depth;
-	} else {
-		toffset += surf.layer_size * mslice;
-	}
-	if (toffset > radeon_bo_size(texture)) {
-		dev_warn(p->dev, "%s:%d texture bo too small (layer size %d, "
-			 "offset %ld, max layer %d, depth %d, bo size %ld) (%d %d)\n",
-			 __func__, __LINE__, surf.layer_size,
-			(unsigned long)texdw[2] << 8, mslice,
-			depth, radeon_bo_size(texture),
-			surf.nbx, surf.nby);
-		return -EINVAL;
-	}
-
-	/* check mipmap size */
-	for (i = 1; i <= llevel; i++) {
-		unsigned w, h, d;
-
-		w = r600_mip_minify(width, i);
-		h = r600_mip_minify(height, i);
-		d = r600_mip_minify(depth, i);
-		surf.nbx = r600_fmt_get_nblocksx(surf.format, w);
-		surf.nby = r600_fmt_get_nblocksy(surf.format, h);
-
-		switch (surf.mode) {
-		case ARRAY_2D_TILED_THIN1:
-			if (surf.nbx < surf.palign || surf.nby < surf.halign) {
-				surf.mode = ARRAY_1D_TILED_THIN1;
-			}
-			/* recompute alignment */
-			evergreen_surface_check(p, &surf, NULL);
-			break;
-		case ARRAY_LINEAR_GENERAL:
-		case ARRAY_LINEAR_ALIGNED:
-		case ARRAY_1D_TILED_THIN1:
-			break;
-		default:
-			dev_warn(p->dev, "%s:%d invalid array mode %d\n",
-				 __func__, __LINE__, surf.mode);
-			return -EINVAL;
-		}
-		surf.nbx = ALIGN(surf.nbx, surf.palign);
-		surf.nby = ALIGN(surf.nby, surf.halign);
-
-		r = evergreen_surface_check(p, &surf, "mipmap");
-		if (r) {
-			return r;
-		}
-
-		if (dim == SQ_TEX_DIM_3D) {
-			moffset += surf.layer_size * d;
-		} else {
-			moffset += surf.layer_size * mslice;
-		}
-		if (moffset > radeon_bo_size(mipmap)) {
-			dev_warn(p->dev, "%s:%d mipmap [%d] bo too small (layer size %d, "
-					"offset %ld, coffset %ld, max layer %d, depth %d, "
-					"bo size %ld) level0 (%d %d %d)\n",
-					__func__, __LINE__, i, surf.layer_size,
-					(unsigned long)texdw[3] << 8, moffset, mslice,
-					d, radeon_bo_size(mipmap),
-					width, height, depth);
-			dev_warn(p->dev, "%s:%d problematic surf: (%d %d) (%d %d %d %d %d %d %d)\n",
-				 __func__, __LINE__, surf.nbx, surf.nby,
-				surf.mode, surf.bpe, surf.nsamples,
-				surf.bankw, surf.bankh,
-				surf.tsplit, surf.mtilea);
-			return -EINVAL;
-		}
-	}
-
+	/* XXX fill in */
 	return 0;
 }
 
 static int evergreen_cs_track_check(struct radeon_cs_parser *p)
 {
 	struct evergreen_cs_track *track = p->track;
-	unsigned tmp, i;
-	int r;
-	unsigned buffer_mask = 0;
 
-	/* check streamout */
-	if (track->streamout_dirty && track->vgt_strmout_config) {
-		for (i = 0; i < 4; i++) {
-			if (track->vgt_strmout_config & (1 << i)) {
-				buffer_mask |= (track->vgt_strmout_buffer_config >> (i * 4)) & 0xf;
-			}
-		}
-
-		for (i = 0; i < 4; i++) {
-			if (buffer_mask & (1 << i)) {
-				if (track->vgt_strmout_bo[i]) {
-					u64 offset = (u64)track->vgt_strmout_bo_offset[i] +
-							(u64)track->vgt_strmout_size[i];
-					if (offset > radeon_bo_size(track->vgt_strmout_bo[i])) {
-						DRM_ERROR("streamout %d bo too small: 0x%llx, 0x%lx\n",
-							  i, offset,
-							  radeon_bo_size(track->vgt_strmout_bo[i]));
-						return -EINVAL;
-					}
-				} else {
-					dev_warn(p->dev, "No buffer for streamout %d\n", i);
-					return -EINVAL;
-				}
-			}
-		}
-		track->streamout_dirty = false;
-	}
-
-	if (track->sx_misc_kill_all_prims)
-		return 0;
-
-	/* check that we have a cb for each enabled target
-	 */
-	if (track->cb_dirty) {
-		tmp = track->cb_target_mask;
-		for (i = 0; i < 8; i++) {
-			if ((tmp >> (i * 4)) & 0xF) {
-				/* at least one component is enabled */
-				if (track->cb_color_bo[i] == NULL) {
-					dev_warn(p->dev, "%s:%d mask 0x%08X | 0x%08X no cb for %d\n",
-						__func__, __LINE__, track->cb_target_mask, track->cb_shader_mask, i);
-					return -EINVAL;
-				}
-				/* check cb */
-				r = evergreen_cs_track_validate_cb(p, i);
-				if (r) {
-					return r;
-				}
-			}
-		}
-		track->cb_dirty = false;
-	}
-
-	if (track->db_dirty) {
-		/* Check stencil buffer */
-		if (G_028800_STENCIL_ENABLE(track->db_depth_control)) {
-			r = evergreen_cs_track_validate_stencil(p);
-			if (r)
-				return r;
-		}
-		/* Check depth buffer */
-		if (G_028800_Z_WRITE_ENABLE(track->db_depth_control)) {
-			r = evergreen_cs_track_validate_depth(p);
-			if (r)
-				return r;
-		}
-		track->db_dirty = false;
+	/* we don't support stream out buffer yet */
+	if (track->vgt_strmout_config || track->vgt_strmout_buffer_config) {
+		dev_warn(p->dev, "this kernel doesn't support SMX output buffer\n");
+		return -EINVAL;
 	}
 
+	/* XXX fill in */
 	return 0;
 }
 
@@ -955,6 +236,28 @@ static int evergreen_cs_packet_next_reloc(struct radeon_cs_parser *p,
 }
 
 /**
+ * evergreen_cs_packet_next_is_pkt3_nop() - test if next packet is packet3 nop for reloc
+ * @parser:		parser structure holding parsing context.
+ *
+ * Check next packet is relocation packet3, do bo validation and compute
+ * GPU offset using the provided start.
+ **/
+static inline int evergreen_cs_packet_next_is_pkt3_nop(struct radeon_cs_parser *p)
+{
+	struct radeon_cs_packet p3reloc;
+	int r;
+
+	r = evergreen_cs_packet_parse(p, &p3reloc, p->idx);
+	if (r) {
+		return 0;
+	}
+	if (p3reloc.type != PACKET_TYPE3 || p3reloc.opcode != PACKET3_NOP) {
+		return 0;
+	}
+	return 1;
+}
+
+/**
  * evergreen_cs_packet_next_vline() - parse userspace VLINE packet
  * @parser:		parser structure holding parsing context.
  *
@@ -1111,7 +414,7 @@ static int evergreen_cs_parse_packet0(struct radeon_cs_parser *p,
  * if register is safe. If register is not flag as safe this function
  * will test it against a list of register needind special handling.
  */
-static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
+static inline int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 {
 	struct evergreen_cs_track *track = (struct evergreen_cs_track *)p->track;
 	struct radeon_cs_reloc *reloc;
@@ -1125,7 +428,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		last_reg = ARRAY_SIZE(evergreen_reg_safe_bm);
 
 	i = (reg >> 7);
-	if (i >= last_reg) {
+	if (i > last_reg) {
 		dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx);
 		return -EINVAL;
 	}
@@ -1189,7 +492,6 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		break;
 	case DB_DEPTH_CONTROL:
 		track->db_depth_control = radeon_get_ib_value(p, idx);
-		track->db_dirty = true;
 		break;
 	case CAYMAN_DB_EQAA:
 		if (p->rdev->family < CHIP_CAYMAN) {
@@ -1206,48 +508,32 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		}
 		break;
 	case DB_Z_INFO:
+		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		if (r) {
+			dev_warn(p->dev, "bad SET_CONTEXT_REG "
+					"0x%04X\n", reg);
+			return -EINVAL;
+		}
 		track->db_z_info = radeon_get_ib_value(p, idx);
-		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
-			r = evergreen_cs_packet_next_reloc(p, &reloc);
-			if (r) {
-				dev_warn(p->dev, "bad SET_CONTEXT_REG "
-						"0x%04X\n", reg);
-				return -EINVAL;
-			}
-			ib[idx] &= ~Z_ARRAY_MODE(0xf);
-			track->db_z_info &= ~Z_ARRAY_MODE(0xf);
-			ib[idx] |= Z_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags));
-			track->db_z_info |= Z_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags));
-			if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) {
-				unsigned bankw, bankh, mtaspect, tile_split;
-
-				evergreen_tiling_fields(reloc->lobj.tiling_flags,
-							&bankw, &bankh, &mtaspect,
-							&tile_split);
-				ib[idx] |= DB_NUM_BANKS(evergreen_cs_get_num_banks(track->nbanks));
-				ib[idx] |= DB_TILE_SPLIT(tile_split) |
-						DB_BANK_WIDTH(bankw) |
-						DB_BANK_HEIGHT(bankh) |
-						DB_MACRO_TILE_ASPECT(mtaspect);
-			}
+		ib[idx] &= ~Z_ARRAY_MODE(0xf);
+		track->db_z_info &= ~Z_ARRAY_MODE(0xf);
+		if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) {
+			ib[idx] |= Z_ARRAY_MODE(ARRAY_2D_TILED_THIN1);
+			track->db_z_info |= Z_ARRAY_MODE(ARRAY_2D_TILED_THIN1);
+		} else {
+			ib[idx] |= Z_ARRAY_MODE(ARRAY_1D_TILED_THIN1);
+			track->db_z_info |= Z_ARRAY_MODE(ARRAY_1D_TILED_THIN1);
 		}
-		track->db_dirty = true;
 		break;
 	case DB_STENCIL_INFO:
 		track->db_s_info = radeon_get_ib_value(p, idx);
-		track->db_dirty = true;
 		break;
 	case DB_DEPTH_VIEW:
 		track->db_depth_view = radeon_get_ib_value(p, idx);
-		track->db_dirty = true;
 		break;
 	case DB_DEPTH_SIZE:
 		track->db_depth_size = radeon_get_ib_value(p, idx);
-		track->db_dirty = true;
-		break;
-	case R_02805C_DB_DEPTH_SLICE:
-		track->db_depth_slice = radeon_get_ib_value(p, idx);
-		track->db_dirty = true;
+		track->db_depth_size_idx = idx;
 		break;
 	case DB_Z_READ_BASE:
 		r = evergreen_cs_packet_next_reloc(p, &reloc);
@@ -1259,7 +545,6 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		track->db_z_read_offset = radeon_get_ib_value(p, idx);
 		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 		track->db_z_read_bo = reloc->robj;
-		track->db_dirty = true;
 		break;
 	case DB_Z_WRITE_BASE:
 		r = evergreen_cs_packet_next_reloc(p, &reloc);
@@ -1271,7 +556,6 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		track->db_z_write_offset = radeon_get_ib_value(p, idx);
 		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 		track->db_z_write_bo = reloc->robj;
-		track->db_dirty = true;
 		break;
 	case DB_STENCIL_READ_BASE:
 		r = evergreen_cs_packet_next_reloc(p, &reloc);
@@ -1283,7 +567,6 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		track->db_s_read_offset = radeon_get_ib_value(p, idx);
 		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 		track->db_s_read_bo = reloc->robj;
-		track->db_dirty = true;
 		break;
 	case DB_STENCIL_WRITE_BASE:
 		r = evergreen_cs_packet_next_reloc(p, &reloc);
@@ -1295,56 +578,18 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		track->db_s_write_offset = radeon_get_ib_value(p, idx);
 		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 		track->db_s_write_bo = reloc->robj;
-		track->db_dirty = true;
 		break;
 	case VGT_STRMOUT_CONFIG:
 		track->vgt_strmout_config = radeon_get_ib_value(p, idx);
-		track->streamout_dirty = true;
 		break;
 	case VGT_STRMOUT_BUFFER_CONFIG:
 		track->vgt_strmout_buffer_config = radeon_get_ib_value(p, idx);
-		track->streamout_dirty = true;
 		break;
-	case VGT_STRMOUT_BUFFER_BASE_0:
-	case VGT_STRMOUT_BUFFER_BASE_1:
-	case VGT_STRMOUT_BUFFER_BASE_2:
-	case VGT_STRMOUT_BUFFER_BASE_3:
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
-		if (r) {
-			dev_warn(p->dev, "bad SET_CONTEXT_REG "
-					"0x%04X\n", reg);
-			return -EINVAL;
-		}
-		tmp = (reg - VGT_STRMOUT_BUFFER_BASE_0) / 16;
-		track->vgt_strmout_bo_offset[tmp] = radeon_get_ib_value(p, idx) << 8;
-		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
-		track->vgt_strmout_bo[tmp] = reloc->robj;
-		track->streamout_dirty = true;
-		break;
-	case VGT_STRMOUT_BUFFER_SIZE_0:
-	case VGT_STRMOUT_BUFFER_SIZE_1:
-	case VGT_STRMOUT_BUFFER_SIZE_2:
-	case VGT_STRMOUT_BUFFER_SIZE_3:
-		tmp = (reg - VGT_STRMOUT_BUFFER_SIZE_0) / 16;
-		/* size in register is DWs, convert to bytes */
-		track->vgt_strmout_size[tmp] = radeon_get_ib_value(p, idx) * 4;
-		track->streamout_dirty = true;
-		break;
-	case CP_COHER_BASE:
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
-		if (r) {
-			dev_warn(p->dev, "missing reloc for CP_COHER_BASE "
-					"0x%04X\n", reg);
-			return -EINVAL;
-		}
-		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 	case CB_TARGET_MASK:
 		track->cb_target_mask = radeon_get_ib_value(p, idx);
-		track->cb_dirty = true;
 		break;
 	case CB_SHADER_MASK:
 		track->cb_shader_mask = radeon_get_ib_value(p, idx);
-		track->cb_dirty = true;
 		break;
 	case PA_SC_AA_CONFIG:
 		if (p->rdev->family >= CHIP_CAYMAN) {
@@ -1374,7 +619,6 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case CB_COLOR7_VIEW:
 		tmp = (reg - CB_COLOR0_VIEW) / 0x3c;
 		track->cb_color_view[tmp] = radeon_get_ib_value(p, idx);
-		track->cb_dirty = true;
 		break;
 	case CB_COLOR8_VIEW:
 	case CB_COLOR9_VIEW:
@@ -1382,7 +626,6 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case CB_COLOR11_VIEW:
 		tmp = ((reg - CB_COLOR8_VIEW) / 0x1c) + 8;
 		track->cb_color_view[tmp] = radeon_get_ib_value(p, idx);
-		track->cb_dirty = true;
 		break;
 	case CB_COLOR0_INFO:
 	case CB_COLOR1_INFO:
@@ -1392,37 +635,41 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case CB_COLOR5_INFO:
 	case CB_COLOR6_INFO:
 	case CB_COLOR7_INFO:
+		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		if (r) {
+			dev_warn(p->dev, "bad SET_CONTEXT_REG "
+					"0x%04X\n", reg);
+			return -EINVAL;
+		}
 		tmp = (reg - CB_COLOR0_INFO) / 0x3c;
 		track->cb_color_info[tmp] = radeon_get_ib_value(p, idx);
-		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
-			r = evergreen_cs_packet_next_reloc(p, &reloc);
-			if (r) {
-				dev_warn(p->dev, "bad SET_CONTEXT_REG "
-						"0x%04X\n", reg);
-				return -EINVAL;
-			}
-			ib[idx] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags));
-			track->cb_color_info[tmp] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags));
+		if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) {
+			ib[idx] |= CB_ARRAY_MODE(ARRAY_2D_TILED_THIN1);
+			track->cb_color_info[tmp] |= CB_ARRAY_MODE(ARRAY_2D_TILED_THIN1);
+		} else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) {
+			ib[idx] |= CB_ARRAY_MODE(ARRAY_1D_TILED_THIN1);
+			track->cb_color_info[tmp] |= CB_ARRAY_MODE(ARRAY_1D_TILED_THIN1);
 		}
-		track->cb_dirty = true;
 		break;
 	case CB_COLOR8_INFO:
 	case CB_COLOR9_INFO:
 	case CB_COLOR10_INFO:
 	case CB_COLOR11_INFO:
+		r = evergreen_cs_packet_next_reloc(p, &reloc);
+		if (r) {
+			dev_warn(p->dev, "bad SET_CONTEXT_REG "
+					"0x%04X\n", reg);
+			return -EINVAL;
+		}
 		tmp = ((reg - CB_COLOR8_INFO) / 0x1c) + 8;
 		track->cb_color_info[tmp] = radeon_get_ib_value(p, idx);
-		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
-			r = evergreen_cs_packet_next_reloc(p, &reloc);
-			if (r) {
-				dev_warn(p->dev, "bad SET_CONTEXT_REG "
-						"0x%04X\n", reg);
-				return -EINVAL;
-			}
-			ib[idx] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags));
-			track->cb_color_info[tmp] |= CB_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags));
+		if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) {
+			ib[idx] |= CB_ARRAY_MODE(ARRAY_2D_TILED_THIN1);
+			track->cb_color_info[tmp] |= CB_ARRAY_MODE(ARRAY_2D_TILED_THIN1);
+		} else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) {
+			ib[idx] |= CB_ARRAY_MODE(ARRAY_1D_TILED_THIN1);
+			track->cb_color_info[tmp] |= CB_ARRAY_MODE(ARRAY_1D_TILED_THIN1);
 		}
-		track->cb_dirty = true;
 		break;
 	case CB_COLOR0_PITCH:
 	case CB_COLOR1_PITCH:
@@ -1434,7 +681,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case CB_COLOR7_PITCH:
 		tmp = (reg - CB_COLOR0_PITCH) / 0x3c;
 		track->cb_color_pitch[tmp] = radeon_get_ib_value(p, idx);
-		track->cb_dirty = true;
+		track->cb_color_pitch_idx[tmp] = idx;
 		break;
 	case CB_COLOR8_PITCH:
 	case CB_COLOR9_PITCH:
@@ -1442,7 +689,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case CB_COLOR11_PITCH:
 		tmp = ((reg - CB_COLOR8_PITCH) / 0x1c) + 8;
 		track->cb_color_pitch[tmp] = radeon_get_ib_value(p, idx);
-		track->cb_dirty = true;
+		track->cb_color_pitch_idx[tmp] = idx;
 		break;
 	case CB_COLOR0_SLICE:
 	case CB_COLOR1_SLICE:
@@ -1454,7 +701,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case CB_COLOR7_SLICE:
 		tmp = (reg - CB_COLOR0_SLICE) / 0x3c;
 		track->cb_color_slice[tmp] = radeon_get_ib_value(p, idx);
-		track->cb_dirty = true;
+		track->cb_color_slice_idx[tmp] = idx;
 		break;
 	case CB_COLOR8_SLICE:
 	case CB_COLOR9_SLICE:
@@ -1462,7 +709,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case CB_COLOR11_SLICE:
 		tmp = ((reg - CB_COLOR8_SLICE) / 0x1c) + 8;
 		track->cb_color_slice[tmp] = radeon_get_ib_value(p, idx);
-		track->cb_dirty = true;
+		track->cb_color_slice_idx[tmp] = idx;
 		break;
 	case CB_COLOR0_ATTRIB:
 	case CB_COLOR1_ATTRIB:
@@ -1472,57 +719,30 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case CB_COLOR5_ATTRIB:
 	case CB_COLOR6_ATTRIB:
 	case CB_COLOR7_ATTRIB:
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
-		if (r) {
-			dev_warn(p->dev, "bad SET_CONTEXT_REG "
-					"0x%04X\n", reg);
-			return -EINVAL;
-		}
-		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
-			if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) {
-				unsigned bankw, bankh, mtaspect, tile_split;
-
-				evergreen_tiling_fields(reloc->lobj.tiling_flags,
-							&bankw, &bankh, &mtaspect,
-							&tile_split);
-				ib[idx] |= CB_NUM_BANKS(evergreen_cs_get_num_banks(track->nbanks));
-				ib[idx] |= CB_TILE_SPLIT(tile_split) |
-					   CB_BANK_WIDTH(bankw) |
-					   CB_BANK_HEIGHT(bankh) |
-					   CB_MACRO_TILE_ASPECT(mtaspect);
-			}
-		}
-		tmp = ((reg - CB_COLOR0_ATTRIB) / 0x3c);
-		track->cb_color_attrib[tmp] = ib[idx];
-		track->cb_dirty = true;
-		break;
 	case CB_COLOR8_ATTRIB:
 	case CB_COLOR9_ATTRIB:
 	case CB_COLOR10_ATTRIB:
 	case CB_COLOR11_ATTRIB:
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
-		if (r) {
-			dev_warn(p->dev, "bad SET_CONTEXT_REG "
-					"0x%04X\n", reg);
-			return -EINVAL;
-		}
-		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
-			if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) {
-				unsigned bankw, bankh, mtaspect, tile_split;
-
-				evergreen_tiling_fields(reloc->lobj.tiling_flags,
-							&bankw, &bankh, &mtaspect,
-							&tile_split);
-				ib[idx] |= CB_NUM_BANKS(evergreen_cs_get_num_banks(track->nbanks));
-				ib[idx] |= CB_TILE_SPLIT(tile_split) |
-					   CB_BANK_WIDTH(bankw) |
-					   CB_BANK_HEIGHT(bankh) |
-					   CB_MACRO_TILE_ASPECT(mtaspect);
-			}
-		}
-		tmp = ((reg - CB_COLOR8_ATTRIB) / 0x1c) + 8;
-		track->cb_color_attrib[tmp] = ib[idx];
-		track->cb_dirty = true;
+		break;
+	case CB_COLOR0_DIM:
+	case CB_COLOR1_DIM:
+	case CB_COLOR2_DIM:
+	case CB_COLOR3_DIM:
+	case CB_COLOR4_DIM:
+	case CB_COLOR5_DIM:
+	case CB_COLOR6_DIM:
+	case CB_COLOR7_DIM:
+		tmp = (reg - CB_COLOR0_DIM) / 0x3c;
+		track->cb_color_dim[tmp] = radeon_get_ib_value(p, idx);
+		track->cb_color_dim_idx[tmp] = idx;
+		break;
+	case CB_COLOR8_DIM:
+	case CB_COLOR9_DIM:
+	case CB_COLOR10_DIM:
+	case CB_COLOR11_DIM:
+		tmp = ((reg - CB_COLOR8_DIM) / 0x1c) + 8;
+		track->cb_color_dim[tmp] = radeon_get_ib_value(p, idx);
+		track->cb_color_dim_idx[tmp] = idx;
 		break;
 	case CB_COLOR0_FMASK:
 	case CB_COLOR1_FMASK:
@@ -1597,8 +817,8 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		tmp = (reg - CB_COLOR0_BASE) / 0x3c;
 		track->cb_color_bo_offset[tmp] = radeon_get_ib_value(p, idx);
 		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
+		track->cb_color_base_last[tmp] = ib[idx];
 		track->cb_color_bo[tmp] = reloc->robj;
-		track->cb_dirty = true;
 		break;
 	case CB_COLOR8_BASE:
 	case CB_COLOR9_BASE:
@@ -1613,8 +833,8 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		tmp = ((reg - CB_COLOR8_BASE) / 0x1c) + 8;
 		track->cb_color_bo_offset[tmp] = radeon_get_ib_value(p, idx);
 		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
+		track->cb_color_base_last[tmp] = ib[idx];
 		track->cb_color_bo[tmp] = reloc->robj;
-		track->cb_dirty = true;
 		break;
 	case CB_IMMED0_BASE:
 	case CB_IMMED1_BASE:
@@ -1636,6 +856,7 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case SQ_PGM_START_PS:
 	case SQ_PGM_START_HS:
 	case SQ_PGM_START_LS:
+	case GDS_ADDR_BASE:
 	case SQ_CONST_MEM_BASE:
 	case SQ_ALU_CONST_CACHE_GS_0:
 	case SQ_ALU_CONST_CACHE_GS_1:
@@ -1725,37 +946,6 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		}
 		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 		break;
-	case SX_MEMORY_EXPORT_BASE:
-		if (p->rdev->family >= CHIP_CAYMAN) {
-			dev_warn(p->dev, "bad SET_CONFIG_REG "
-				 "0x%04X\n", reg);
-			return -EINVAL;
-		}
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
-		if (r) {
-			dev_warn(p->dev, "bad SET_CONFIG_REG "
-					"0x%04X\n", reg);
-			return -EINVAL;
-		}
-		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
-		break;
-	case CAYMAN_SX_SCATTER_EXPORT_BASE:
-		if (p->rdev->family < CHIP_CAYMAN) {
-			dev_warn(p->dev, "bad SET_CONTEXT_REG "
-				 "0x%04X\n", reg);
-			return -EINVAL;
-		}
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
-		if (r) {
-			dev_warn(p->dev, "bad SET_CONTEXT_REG "
-					"0x%04X\n", reg);
-			return -EINVAL;
-		}
-		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
-		break;
-	case SX_MISC:
-		track->sx_misc_kill_all_prims = (radeon_get_ib_value(p, idx) & 0x1) != 0;
-		break;
 	default:
 		dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx);
 		return -EINVAL;
@@ -1763,30 +953,22 @@ static int evergreen_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	return 0;
 }
 
-static bool evergreen_is_safe_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
+/**
+ * evergreen_check_texture_resource() - check if register is authorized or not
+ * @p: parser structure holding parsing context
+ * @idx: index into the cs buffer
+ * @texture: texture's bo structure
+ * @mipmap: mipmap's bo structure
+ *
+ * This function will check that the resource has valid field and that
+ * the texture and mipmap bo object are big enough to cover this resource.
+ */
+static inline int evergreen_check_texture_resource(struct radeon_cs_parser *p,  u32 idx,
+						   struct radeon_bo *texture,
+						   struct radeon_bo *mipmap)
 {
-	u32 last_reg, m, i;
-
-	if (p->rdev->family >= CHIP_CAYMAN)
-		last_reg = ARRAY_SIZE(cayman_reg_safe_bm);
-	else
-		last_reg = ARRAY_SIZE(evergreen_reg_safe_bm);
-
-	i = (reg >> 7);
-	if (i >= last_reg) {
-		dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx);
-		return false;
-	}
-	m = 1 << ((reg >> 2) & 31);
-	if (p->rdev->family >= CHIP_CAYMAN) {
-		if (!(cayman_reg_safe_bm[i] & m))
-			return true;
-	} else {
-		if (!(evergreen_reg_safe_bm[i] & m))
-			return true;
-	}
-	dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx);
-	return false;
+	/* XXX fill in */
+	return 0;
 }
 
 static int evergreen_packet3_check(struct radeon_cs_parser *p,
@@ -1811,8 +993,6 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 	{
 		int pred_op;
 		int tmp;
-		uint64_t offset;
-
 		if (pkt->count != 1) {
 			DRM_ERROR("bad SET PREDICATION\n");
 			return -EINVAL;
@@ -1836,12 +1016,8 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			return -EINVAL;
 		}
 
-		offset = reloc->lobj.gpu_offset +
-		         (idx_value & 0xfffffff0) +
-		         ((u64)(tmp & 0xff) << 32);
-
-		ib[idx + 0] = offset;
-		ib[idx + 1] = (tmp & 0xffffff00) | (upper_32_bits(offset) & 0xff);
+		ib[idx + 0] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff);
+		ib[idx + 1] = tmp + (upper_32_bits(reloc->lobj.gpu_offset) & 0xff);
 	}
 	break;
 	case PACKET3_CONTEXT_CONTROL:
@@ -1869,9 +1045,6 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 		}
 		break;
 	case PACKET3_INDEX_BASE:
-	{
-		uint64_t offset;
-
 		if (pkt->count != 1) {
 			DRM_ERROR("bad INDEX_BASE\n");
 			return -EINVAL;
@@ -1881,24 +1054,15 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad INDEX_BASE\n");
 			return -EINVAL;
 		}
-
-		offset = reloc->lobj.gpu_offset +
-		         idx_value +
-		         ((u64)(radeon_get_ib_value(p, idx+1) & 0xff) << 32);
-
-		ib[idx+0] = offset;
-		ib[idx+1] = upper_32_bits(offset) & 0xff;
-
+		ib[idx+0] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff);
+		ib[idx+1] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
 		r = evergreen_cs_track_check(p);
 		if (r) {
 			dev_warn(p->dev, "%s:%d invalid cmd stream\n", __func__, __LINE__);
 			return r;
 		}
 		break;
-	}
 	case PACKET3_DRAW_INDEX:
-	{
-		uint64_t offset;
 		if (pkt->count != 3) {
 			DRM_ERROR("bad DRAW_INDEX\n");
 			return -EINVAL;
@@ -1908,25 +1072,15 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad DRAW_INDEX\n");
 			return -EINVAL;
 		}
-
-		offset = reloc->lobj.gpu_offset +
-		         idx_value +
-		         ((u64)(radeon_get_ib_value(p, idx+1) & 0xff) << 32);
-
-		ib[idx+0] = offset;
-		ib[idx+1] = upper_32_bits(offset) & 0xff;
-
+		ib[idx+0] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff);
+		ib[idx+1] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
 		r = evergreen_cs_track_check(p);
 		if (r) {
 			dev_warn(p->dev, "%s:%d invalid cmd stream\n", __func__, __LINE__);
 			return r;
 		}
 		break;
-	}
 	case PACKET3_DRAW_INDEX_2:
-	{
-		uint64_t offset;
-
 		if (pkt->count != 4) {
 			DRM_ERROR("bad DRAW_INDEX_2\n");
 			return -EINVAL;
@@ -1936,21 +1090,14 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad DRAW_INDEX_2\n");
 			return -EINVAL;
 		}
-
-		offset = reloc->lobj.gpu_offset +
-		         radeon_get_ib_value(p, idx+1) +
-		         ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32);
-
-		ib[idx+1] = offset;
-		ib[idx+2] = upper_32_bits(offset) & 0xff;
-
+		ib[idx+1] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff);
+		ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
 		r = evergreen_cs_track_check(p);
 		if (r) {
 			dev_warn(p->dev, "%s:%d invalid cmd stream\n", __func__, __LINE__);
 			return r;
 		}
 		break;
-	}
 	case PACKET3_DRAW_INDEX_AUTO:
 		if (pkt->count != 1) {
 			DRM_ERROR("bad DRAW_INDEX_AUTO\n");
@@ -2006,34 +1153,6 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			return r;
 		}
 		break;
-	case PACKET3_DISPATCH_DIRECT:
-		if (pkt->count != 3) {
-			DRM_ERROR("bad DISPATCH_DIRECT\n");
-			return -EINVAL;
-		}
-		r = evergreen_cs_track_check(p);
-		if (r) {
-			dev_warn(p->dev, "%s:%d invalid cmd stream %d\n", __func__, __LINE__, idx);
-			return r;
-		}
-		break;
-	case PACKET3_DISPATCH_INDIRECT:
-		if (pkt->count != 1) {
-			DRM_ERROR("bad DISPATCH_INDIRECT\n");
-			return -EINVAL;
-		}
-		r = evergreen_cs_packet_next_reloc(p, &reloc);
-		if (r) {
-			DRM_ERROR("bad DISPATCH_INDIRECT\n");
-			return -EINVAL;
-		}
-		ib[idx+0] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff);
-		r = evergreen_cs_track_check(p);
-		if (r) {
-			dev_warn(p->dev, "%s:%d invalid cmd stream\n", __func__, __LINE__);
-			return r;
-		}
-		break;
 	case PACKET3_WAIT_REG_MEM:
 		if (pkt->count != 5) {
 			DRM_ERROR("bad WAIT_REG_MEM\n");
@@ -2041,20 +1160,13 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 		}
 		/* bit 4 is reg (0) or mem (1) */
 		if (idx_value & 0x10) {
-			uint64_t offset;
-
 			r = evergreen_cs_packet_next_reloc(p, &reloc);
 			if (r) {
 				DRM_ERROR("bad WAIT_REG_MEM\n");
 				return -EINVAL;
 			}
-
-			offset = reloc->lobj.gpu_offset +
-			         (radeon_get_ib_value(p, idx+1) & 0xfffffffc) +
-			         ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32);
-
-			ib[idx+1] = (ib[idx+1] & 0x3) | (offset & 0xfffffffc);
-			ib[idx+2] = upper_32_bits(offset) & 0xff;
+			ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff);
+			ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
 		}
 		break;
 	case PACKET3_SURFACE_SYNC:
@@ -2079,25 +1191,16 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			return -EINVAL;
 		}
 		if (pkt->count) {
-			uint64_t offset;
-
 			r = evergreen_cs_packet_next_reloc(p, &reloc);
 			if (r) {
 				DRM_ERROR("bad EVENT_WRITE\n");
 				return -EINVAL;
 			}
-			offset = reloc->lobj.gpu_offset +
-			         (radeon_get_ib_value(p, idx+1) & 0xfffffff8) +
-			         ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32);
-
-			ib[idx+1] = offset & 0xfffffff8;
-			ib[idx+2] = upper_32_bits(offset) & 0xff;
+			ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff);
+			ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
 		}
 		break;
 	case PACKET3_EVENT_WRITE_EOP:
-	{
-		uint64_t offset;
-
 		if (pkt->count != 4) {
 			DRM_ERROR("bad EVENT_WRITE_EOP\n");
 			return -EINVAL;
@@ -2107,19 +1210,10 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad EVENT_WRITE_EOP\n");
 			return -EINVAL;
 		}
-
-		offset = reloc->lobj.gpu_offset +
-		         (radeon_get_ib_value(p, idx+1) & 0xfffffffc) +
-		         ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32);
-
-		ib[idx+1] = offset & 0xfffffffc;
-		ib[idx+2] = (ib[idx+2] & 0xffffff00) | (upper_32_bits(offset) & 0xff);
+		ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff);
+		ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
 		break;
-	}
 	case PACKET3_EVENT_WRITE_EOS:
-	{
-		uint64_t offset;
-
 		if (pkt->count != 3) {
 			DRM_ERROR("bad EVENT_WRITE_EOS\n");
 			return -EINVAL;
@@ -2129,15 +1223,9 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad EVENT_WRITE_EOS\n");
 			return -EINVAL;
 		}
-
-		offset = reloc->lobj.gpu_offset +
-		         (radeon_get_ib_value(p, idx+1) & 0xfffffffc) +
-		         ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32);
-
-		ib[idx+1] = offset & 0xfffffffc;
-		ib[idx+2] = (ib[idx+2] & 0xffffff00) | (upper_32_bits(offset) & 0xff);
+		ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff);
+		ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
 		break;
-	}
 	case PACKET3_SET_CONFIG_REG:
 		start_reg = (idx_value << 2) + PACKET3_SET_CONFIG_REG_START;
 		end_reg = 4 * pkt->count + start_reg - 4;
@@ -2185,7 +1273,6 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 		}
 		for (i = 0; i < (pkt->count / 8); i++) {
 			struct radeon_bo *texture, *mipmap;
-			u32 toffset, moffset;
 			u32 size, offset;
 
 			switch (G__SQ_CONSTANT_TYPE(radeon_get_ib_value(p, idx+1+(i*8)+7))) {
@@ -2196,42 +1283,26 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 					DRM_ERROR("bad SET_RESOURCE (tex)\n");
 					return -EINVAL;
 				}
-				if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
-					ib[idx+1+(i*8)+1] |=
-						TEX_ARRAY_MODE(evergreen_cs_get_aray_mode(reloc->lobj.tiling_flags));
-					if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO) {
-						unsigned bankw, bankh, mtaspect, tile_split;
-
-						evergreen_tiling_fields(reloc->lobj.tiling_flags,
-									&bankw, &bankh, &mtaspect,
-									&tile_split);
-						ib[idx+1+(i*8)+6] |= TEX_TILE_SPLIT(tile_split);
-						ib[idx+1+(i*8)+7] |=
-							TEX_BANK_WIDTH(bankw) |
-							TEX_BANK_HEIGHT(bankh) |
-							MACRO_TILE_ASPECT(mtaspect) |
-							TEX_NUM_BANKS(evergreen_cs_get_num_banks(track->nbanks));
-					}
-				}
+				ib[idx+1+(i*8)+2] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
+				if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
+					ib[idx+1+(i*8)+1] |= TEX_ARRAY_MODE(ARRAY_2D_TILED_THIN1);
+				else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
+					ib[idx+1+(i*8)+1] |= TEX_ARRAY_MODE(ARRAY_1D_TILED_THIN1);
 				texture = reloc->robj;
-				toffset = (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 				/* tex mip base */
 				r = evergreen_cs_packet_next_reloc(p, &reloc);
 				if (r) {
 					DRM_ERROR("bad SET_RESOURCE (tex)\n");
 					return -EINVAL;
 				}
-				moffset = (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
+				ib[idx+1+(i*8)+3] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 				mipmap = reloc->robj;
-				r = evergreen_cs_track_validate_texture(p, texture, mipmap, idx+1+(i*8));
+				r = evergreen_check_texture_resource(p,  idx+1+(i*8),
+						texture, mipmap);
 				if (r)
 					return r;
-				ib[idx+1+(i*8)+2] += toffset;
-				ib[idx+1+(i*8)+3] += moffset;
 				break;
 			case SQ_TEX_VTX_VALID_BUFFER:
-			{
-				uint64_t offset64;
 				/* vtx base */
 				r = evergreen_cs_packet_next_reloc(p, &reloc);
 				if (r) {
@@ -2243,15 +1314,11 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 				if (p->rdev && (size + offset) > radeon_bo_size(reloc->robj)) {
 					/* force size to size of the buffer */
 					dev_warn(p->dev, "vbo resource seems too big for the bo\n");
-					ib[idx+1+(i*8)+1] = radeon_bo_size(reloc->robj) - offset;
+					ib[idx+1+(i*8)+1] = radeon_bo_size(reloc->robj);
 				}
-
-				offset64 = reloc->lobj.gpu_offset + offset;
-				ib[idx+1+(i*8)+0] = offset64;
-				ib[idx+1+(i*8)+2] = (ib[idx+1+(i*8)+2] & 0xffffff00) |
-						    (upper_32_bits(offset64) & 0xff);
+				ib[idx+1+(i*8)+0] += (u32)((reloc->lobj.gpu_offset) & 0xffffffff);
+				ib[idx+1+(i*8)+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
 				break;
-			}
 			case SQ_TEX_VTX_INVALID_TEXTURE:
 			case SQ_TEX_VTX_INVALID_BUFFER:
 			default:
@@ -2307,104 +1374,6 @@ static int evergreen_packet3_check(struct radeon_cs_parser *p,
 			return -EINVAL;
 		}
 		break;
-	case PACKET3_STRMOUT_BUFFER_UPDATE:
-		if (pkt->count != 4) {
-			DRM_ERROR("bad STRMOUT_BUFFER_UPDATE (invalid count)\n");
-			return -EINVAL;
-		}
-		/* Updating memory at DST_ADDRESS. */
-		if (idx_value & 0x1) {
-			u64 offset;
-			r = evergreen_cs_packet_next_reloc(p, &reloc);
-			if (r) {
-				DRM_ERROR("bad STRMOUT_BUFFER_UPDATE (missing dst reloc)\n");
-				return -EINVAL;
-			}
-			offset = radeon_get_ib_value(p, idx+1);
-			offset += ((u64)(radeon_get_ib_value(p, idx+2) & 0xff)) << 32;
-			if ((offset + 4) > radeon_bo_size(reloc->robj)) {
-				DRM_ERROR("bad STRMOUT_BUFFER_UPDATE dst bo too small: 0x%llx, 0x%lx\n",
-					  offset + 4, radeon_bo_size(reloc->robj));
-				return -EINVAL;
-			}
-			offset += reloc->lobj.gpu_offset;
-			ib[idx+1] = offset;
-			ib[idx+2] = upper_32_bits(offset) & 0xff;
-		}
-		/* Reading data from SRC_ADDRESS. */
-		if (((idx_value >> 1) & 0x3) == 2) {
-			u64 offset;
-			r = evergreen_cs_packet_next_reloc(p, &reloc);
-			if (r) {
-				DRM_ERROR("bad STRMOUT_BUFFER_UPDATE (missing src reloc)\n");
-				return -EINVAL;
-			}
-			offset = radeon_get_ib_value(p, idx+3);
-			offset += ((u64)(radeon_get_ib_value(p, idx+4) & 0xff)) << 32;
-			if ((offset + 4) > radeon_bo_size(reloc->robj)) {
-				DRM_ERROR("bad STRMOUT_BUFFER_UPDATE src bo too small: 0x%llx, 0x%lx\n",
-					  offset + 4, radeon_bo_size(reloc->robj));
-				return -EINVAL;
-			}
-			offset += reloc->lobj.gpu_offset;
-			ib[idx+3] = offset;
-			ib[idx+4] = upper_32_bits(offset) & 0xff;
-		}
-		break;
-	case PACKET3_COPY_DW:
-		if (pkt->count != 4) {
-			DRM_ERROR("bad COPY_DW (invalid count)\n");
-			return -EINVAL;
-		}
-		if (idx_value & 0x1) {
-			u64 offset;
-			/* SRC is memory. */
-			r = evergreen_cs_packet_next_reloc(p, &reloc);
-			if (r) {
-				DRM_ERROR("bad COPY_DW (missing src reloc)\n");
-				return -EINVAL;
-			}
-			offset = radeon_get_ib_value(p, idx+1);
-			offset += ((u64)(radeon_get_ib_value(p, idx+2) & 0xff)) << 32;
-			if ((offset + 4) > radeon_bo_size(reloc->robj)) {
-				DRM_ERROR("bad COPY_DW src bo too small: 0x%llx, 0x%lx\n",
-					  offset + 4, radeon_bo_size(reloc->robj));
-				return -EINVAL;
-			}
-			offset += reloc->lobj.gpu_offset;
-			ib[idx+1] = offset;
-			ib[idx+2] = upper_32_bits(offset) & 0xff;
-		} else {
-			/* SRC is a reg. */
-			reg = radeon_get_ib_value(p, idx+1) << 2;
-			if (!evergreen_is_safe_reg(p, reg, idx+1))
-				return -EINVAL;
-		}
-		if (idx_value & 0x2) {
-			u64 offset;
-			/* DST is memory. */
-			r = evergreen_cs_packet_next_reloc(p, &reloc);
-			if (r) {
-				DRM_ERROR("bad COPY_DW (missing dst reloc)\n");
-				return -EINVAL;
-			}
-			offset = radeon_get_ib_value(p, idx+3);
-			offset += ((u64)(radeon_get_ib_value(p, idx+4) & 0xff)) << 32;
-			if ((offset + 4) > radeon_bo_size(reloc->robj)) {
-				DRM_ERROR("bad COPY_DW dst bo too small: 0x%llx, 0x%lx\n",
-					  offset + 4, radeon_bo_size(reloc->robj));
-				return -EINVAL;
-			}
-			offset += reloc->lobj.gpu_offset;
-			ib[idx+3] = offset;
-			ib[idx+4] = upper_32_bits(offset) & 0xff;
-		} else {
-			/* DST is a reg. */
-			reg = radeon_get_ib_value(p, idx+3) << 2;
-			if (!evergreen_is_safe_reg(p, reg, idx+3))
-				return -EINVAL;
-		}
-		break;
 	case PACKET3_NOP:
 		break;
 	default:
@@ -2418,7 +1387,6 @@ int evergreen_cs_parse(struct radeon_cs_parser *p)
 {
 	struct radeon_cs_packet pkt;
 	struct evergreen_cs_track *track;
-	u32 tmp;
 	int r;
 
 	if (p->track == NULL) {
@@ -2427,63 +1395,9 @@ int evergreen_cs_parse(struct radeon_cs_parser *p)
 		if (track == NULL)
 			return -ENOMEM;
 		evergreen_cs_track_init(track);
-		if (p->rdev->family >= CHIP_CAYMAN)
-			tmp = p->rdev->config.cayman.tile_config;
-		else
-			tmp = p->rdev->config.evergreen.tile_config;
-
-		switch (tmp & 0xf) {
-		case 0:
-			track->npipes = 1;
-			break;
-		case 1:
-		default:
-			track->npipes = 2;
-			break;
-		case 2:
-			track->npipes = 4;
-			break;
-		case 3:
-			track->npipes = 8;
-			break;
-		}
-
-		switch ((tmp & 0xf0) >> 4) {
-		case 0:
-			track->nbanks = 4;
-			break;
-		case 1:
-		default:
-			track->nbanks = 8;
-			break;
-		case 2:
-			track->nbanks = 16;
-			break;
-		}
-
-		switch ((tmp & 0xf00) >> 8) {
-		case 0:
-			track->group_size = 256;
-			break;
-		case 1:
-		default:
-			track->group_size = 512;
-			break;
-		}
-
-		switch ((tmp & 0xf000) >> 12) {
-		case 0:
-			track->row_size = 1;
-			break;
-		case 1:
-		default:
-			track->row_size = 2;
-			break;
-		case 2:
-			track->row_size = 4;
-			break;
-		}
-
+		track->npipes = p->rdev->config.evergreen.tiling_npipes;
+		track->nbanks = p->rdev->config.evergreen.tiling_nbanks;
+		track->group_size = p->rdev->config.evergreen.tiling_group_size;
 		p->track = track;
 	}
 	do {
@@ -2526,242 +1440,3 @@ int evergreen_cs_parse(struct radeon_cs_parser *p)
 	return 0;
 }
 
-/* vm parser */
-static bool evergreen_vm_reg_valid(u32 reg)
-{
-	/* context regs are fine */
-	if (reg >= 0x28000)
-		return true;
-
-	/* check config regs */
-	switch (reg) {
-	case GRBM_GFX_INDEX:
-	case VGT_VTX_VECT_EJECT_REG:
-	case VGT_CACHE_INVALIDATION:
-	case VGT_GS_VERTEX_REUSE:
-	case VGT_PRIMITIVE_TYPE:
-	case VGT_INDEX_TYPE:
-	case VGT_NUM_INDICES:
-	case VGT_NUM_INSTANCES:
-	case VGT_COMPUTE_DIM_X:
-	case VGT_COMPUTE_DIM_Y:
-	case VGT_COMPUTE_DIM_Z:
-	case VGT_COMPUTE_START_X:
-	case VGT_COMPUTE_START_Y:
-	case VGT_COMPUTE_START_Z:
-	case VGT_COMPUTE_INDEX:
-	case VGT_COMPUTE_THREAD_GROUP_SIZE:
-	case VGT_HS_OFFCHIP_PARAM:
-	case PA_CL_ENHANCE:
-	case PA_SU_LINE_STIPPLE_VALUE:
-	case PA_SC_LINE_STIPPLE_STATE:
-	case PA_SC_ENHANCE:
-	case SQ_DYN_GPR_CNTL_PS_FLUSH_REQ:
-	case SQ_DYN_GPR_SIMD_LOCK_EN:
-	case SQ_CONFIG:
-	case SQ_GPR_RESOURCE_MGMT_1:
-	case SQ_GLOBAL_GPR_RESOURCE_MGMT_1:
-	case SQ_GLOBAL_GPR_RESOURCE_MGMT_2:
-	case SQ_CONST_MEM_BASE:
-	case SQ_STATIC_THREAD_MGMT_1:
-	case SQ_STATIC_THREAD_MGMT_2:
-	case SQ_STATIC_THREAD_MGMT_3:
-	case SPI_CONFIG_CNTL:
-	case SPI_CONFIG_CNTL_1:
-	case TA_CNTL_AUX:
-	case DB_DEBUG:
-	case DB_DEBUG2:
-	case DB_DEBUG3:
-	case DB_DEBUG4:
-	case DB_WATERMARKS:
-	case TD_PS_BORDER_COLOR_INDEX:
-	case TD_PS_BORDER_COLOR_RED:
-	case TD_PS_BORDER_COLOR_GREEN:
-	case TD_PS_BORDER_COLOR_BLUE:
-	case TD_PS_BORDER_COLOR_ALPHA:
-	case TD_VS_BORDER_COLOR_INDEX:
-	case TD_VS_BORDER_COLOR_RED:
-	case TD_VS_BORDER_COLOR_GREEN:
-	case TD_VS_BORDER_COLOR_BLUE:
-	case TD_VS_BORDER_COLOR_ALPHA:
-	case TD_GS_BORDER_COLOR_INDEX:
-	case TD_GS_BORDER_COLOR_RED:
-	case TD_GS_BORDER_COLOR_GREEN:
-	case TD_GS_BORDER_COLOR_BLUE:
-	case TD_GS_BORDER_COLOR_ALPHA:
-	case TD_HS_BORDER_COLOR_INDEX:
-	case TD_HS_BORDER_COLOR_RED:
-	case TD_HS_BORDER_COLOR_GREEN:
-	case TD_HS_BORDER_COLOR_BLUE:
-	case TD_HS_BORDER_COLOR_ALPHA:
-	case TD_LS_BORDER_COLOR_INDEX:
-	case TD_LS_BORDER_COLOR_RED:
-	case TD_LS_BORDER_COLOR_GREEN:
-	case TD_LS_BORDER_COLOR_BLUE:
-	case TD_LS_BORDER_COLOR_ALPHA:
-	case TD_CS_BORDER_COLOR_INDEX:
-	case TD_CS_BORDER_COLOR_RED:
-	case TD_CS_BORDER_COLOR_GREEN:
-	case TD_CS_BORDER_COLOR_BLUE:
-	case TD_CS_BORDER_COLOR_ALPHA:
-	case SQ_ESGS_RING_SIZE:
-	case SQ_GSVS_RING_SIZE:
-	case SQ_ESTMP_RING_SIZE:
-	case SQ_GSTMP_RING_SIZE:
-	case SQ_HSTMP_RING_SIZE:
-	case SQ_LSTMP_RING_SIZE:
-	case SQ_PSTMP_RING_SIZE:
-	case SQ_VSTMP_RING_SIZE:
-	case SQ_ESGS_RING_ITEMSIZE:
-	case SQ_ESTMP_RING_ITEMSIZE:
-	case SQ_GSTMP_RING_ITEMSIZE:
-	case SQ_GSVS_RING_ITEMSIZE:
-	case SQ_GS_VERT_ITEMSIZE:
-	case SQ_GS_VERT_ITEMSIZE_1:
-	case SQ_GS_VERT_ITEMSIZE_2:
-	case SQ_GS_VERT_ITEMSIZE_3:
-	case SQ_GSVS_RING_OFFSET_1:
-	case SQ_GSVS_RING_OFFSET_2:
-	case SQ_GSVS_RING_OFFSET_3:
-	case SQ_HSTMP_RING_ITEMSIZE:
-	case SQ_LSTMP_RING_ITEMSIZE:
-	case SQ_PSTMP_RING_ITEMSIZE:
-	case SQ_VSTMP_RING_ITEMSIZE:
-	case VGT_TF_RING_SIZE:
-	case SQ_ESGS_RING_BASE:
-	case SQ_GSVS_RING_BASE:
-	case SQ_ESTMP_RING_BASE:
-	case SQ_GSTMP_RING_BASE:
-	case SQ_HSTMP_RING_BASE:
-	case SQ_LSTMP_RING_BASE:
-	case SQ_PSTMP_RING_BASE:
-	case SQ_VSTMP_RING_BASE:
-	case CAYMAN_VGT_OFFCHIP_LDS_BASE:
-	case CAYMAN_SQ_EX_ALLOC_TABLE_SLOTS:
-		return true;
-	default:
-		return false;
-	}
-}
-
-static int evergreen_vm_packet3_check(struct radeon_device *rdev,
-				      u32 *ib, struct radeon_cs_packet *pkt)
-{
-	u32 idx = pkt->idx + 1;
-	u32 idx_value = ib[idx];
-	u32 start_reg, end_reg, reg, i;
-
-	switch (pkt->opcode) {
-	case PACKET3_NOP:
-	case PACKET3_SET_BASE:
-	case PACKET3_CLEAR_STATE:
-	case PACKET3_INDEX_BUFFER_SIZE:
-	case PACKET3_DISPATCH_DIRECT:
-	case PACKET3_DISPATCH_INDIRECT:
-	case PACKET3_MODE_CONTROL:
-	case PACKET3_SET_PREDICATION:
-	case PACKET3_COND_EXEC:
-	case PACKET3_PRED_EXEC:
-	case PACKET3_DRAW_INDIRECT:
-	case PACKET3_DRAW_INDEX_INDIRECT:
-	case PACKET3_INDEX_BASE:
-	case PACKET3_DRAW_INDEX_2:
-	case PACKET3_CONTEXT_CONTROL:
-	case PACKET3_DRAW_INDEX_OFFSET:
-	case PACKET3_INDEX_TYPE:
-	case PACKET3_DRAW_INDEX:
-	case PACKET3_DRAW_INDEX_AUTO:
-	case PACKET3_DRAW_INDEX_IMMD:
-	case PACKET3_NUM_INSTANCES:
-	case PACKET3_DRAW_INDEX_MULTI_AUTO:
-	case PACKET3_STRMOUT_BUFFER_UPDATE:
-	case PACKET3_DRAW_INDEX_OFFSET_2:
-	case PACKET3_DRAW_INDEX_MULTI_ELEMENT:
-	case PACKET3_MPEG_INDEX:
-	case PACKET3_WAIT_REG_MEM:
-	case PACKET3_MEM_WRITE:
-	case PACKET3_SURFACE_SYNC:
-	case PACKET3_EVENT_WRITE:
-	case PACKET3_EVENT_WRITE_EOP:
-	case PACKET3_EVENT_WRITE_EOS:
-	case PACKET3_SET_CONTEXT_REG:
-	case PACKET3_SET_BOOL_CONST:
-	case PACKET3_SET_LOOP_CONST:
-	case PACKET3_SET_RESOURCE:
-	case PACKET3_SET_SAMPLER:
-	case PACKET3_SET_CTL_CONST:
-	case PACKET3_SET_RESOURCE_OFFSET:
-	case PACKET3_SET_CONTEXT_REG_INDIRECT:
-	case PACKET3_SET_RESOURCE_INDIRECT:
-	case CAYMAN_PACKET3_DEALLOC_STATE:
-		break;
-	case PACKET3_COND_WRITE:
-		if (idx_value & 0x100) {
-			reg = ib[idx + 5] * 4;
-			if (!evergreen_vm_reg_valid(reg))
-				return -EINVAL;
-		}
-		break;
-	case PACKET3_COPY_DW:
-		if (idx_value & 0x2) {
-			reg = ib[idx + 3] * 4;
-			if (!evergreen_vm_reg_valid(reg))
-				return -EINVAL;
-		}
-		break;
-	case PACKET3_SET_CONFIG_REG:
-		start_reg = (idx_value << 2) + PACKET3_SET_CONFIG_REG_START;
-		end_reg = 4 * pkt->count + start_reg - 4;
-		if ((start_reg < PACKET3_SET_CONFIG_REG_START) ||
-		    (start_reg >= PACKET3_SET_CONFIG_REG_END) ||
-		    (end_reg >= PACKET3_SET_CONFIG_REG_END)) {
-			DRM_ERROR("bad PACKET3_SET_CONFIG_REG\n");
-			return -EINVAL;
-		}
-		for (i = 0; i < pkt->count; i++) {
-			reg = start_reg + (4 * i);
-			if (!evergreen_vm_reg_valid(reg))
-				return -EINVAL;
-		}
-		break;
-	default:
-		return -EINVAL;
-	}
-	return 0;
-}
-
-int evergreen_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib)
-{
-	int ret = 0;
-	u32 idx = 0;
-	struct radeon_cs_packet pkt;
-
-	do {
-		pkt.idx = idx;
-		pkt.type = CP_PACKET_GET_TYPE(ib->ptr[idx]);
-		pkt.count = CP_PACKET_GET_COUNT(ib->ptr[idx]);
-		pkt.one_reg_wr = 0;
-		switch (pkt.type) {
-		case PACKET_TYPE0:
-			dev_err(rdev->dev, "Packet0 not allowed!\n");
-			ret = -EINVAL;
-			break;
-		case PACKET_TYPE2:
-			idx += 1;
-			break;
-		case PACKET_TYPE3:
-			pkt.opcode = CP_PACKET3_GET_OPCODE(ib->ptr[idx]);
-			ret = evergreen_vm_packet3_check(rdev, ib->ptr, &pkt);
-			idx += pkt.count + 2;
-			break;
-		default:
-			dev_err(rdev->dev, "Unknown packet type %d !\n", pkt.type);
-			ret = -EINVAL;
-			break;
-		}
-		if (ret)
-			break;
-	} while (idx < ib->length_dw);
-
-	return ret;
-}
diff --git a/drivers/gpu/drm/radeon/evergreen_reg.h b/drivers/gpu/drm/radeon/evergreen_reg.h
index 96c10b3..c781c92 100644
--- a/drivers/gpu/drm/radeon/evergreen_reg.h
+++ b/drivers/gpu/drm/radeon/evergreen_reg.h
@@ -35,14 +35,6 @@
 #define EVERGREEN_P1PLL_SS_CNTL                         0x414
 #define EVERGREEN_P2PLL_SS_CNTL                         0x454
 #       define EVERGREEN_PxPLL_SS_EN                    (1 << 12)
-
-#define EVERGREEN_AUDIO_PLL1_MUL			0x5b0
-#define EVERGREEN_AUDIO_PLL1_DIV			0x5b4
-#define EVERGREEN_AUDIO_PLL1_UNK			0x5bc
-
-#define EVERGREEN_AUDIO_ENABLE				0x5e78
-#define EVERGREEN_AUDIO_VENDOR_ID			0x5ec0
-
 /* GRPH blocks at 0x6800, 0x7400, 0x10000, 0x10c00, 0x11800, 0x12400 */
 #define EVERGREEN_GRPH_ENABLE                           0x6800
 #define EVERGREEN_GRPH_CONTROL                          0x6804
@@ -50,17 +42,6 @@
 #       define EVERGREEN_GRPH_DEPTH_8BPP                0
 #       define EVERGREEN_GRPH_DEPTH_16BPP               1
 #       define EVERGREEN_GRPH_DEPTH_32BPP               2
-#       define EVERGREEN_GRPH_NUM_BANKS(x)              (((x) & 0x3) << 2)
-#       define EVERGREEN_ADDR_SURF_2_BANK               0
-#       define EVERGREEN_ADDR_SURF_4_BANK               1
-#       define EVERGREEN_ADDR_SURF_8_BANK               2
-#       define EVERGREEN_ADDR_SURF_16_BANK              3
-#       define EVERGREEN_GRPH_Z(x)                      (((x) & 0x3) << 4)
-#       define EVERGREEN_GRPH_BANK_WIDTH(x)             (((x) & 0x3) << 6)
-#       define EVERGREEN_ADDR_SURF_BANK_WIDTH_1         0
-#       define EVERGREEN_ADDR_SURF_BANK_WIDTH_2         1
-#       define EVERGREEN_ADDR_SURF_BANK_WIDTH_4         2
-#       define EVERGREEN_ADDR_SURF_BANK_WIDTH_8         3
 #       define EVERGREEN_GRPH_FORMAT(x)                 (((x) & 0x7) << 8)
 /* 8 BPP */
 #       define EVERGREEN_GRPH_FORMAT_INDEXED            0
@@ -80,24 +61,6 @@
 #       define EVERGREEN_GRPH_FORMAT_8B_BGRA1010102     5
 #       define EVERGREEN_GRPH_FORMAT_RGB111110          6
 #       define EVERGREEN_GRPH_FORMAT_BGR101111          7
-#       define EVERGREEN_GRPH_BANK_HEIGHT(x)            (((x) & 0x3) << 11)
-#       define EVERGREEN_ADDR_SURF_BANK_HEIGHT_1        0
-#       define EVERGREEN_ADDR_SURF_BANK_HEIGHT_2        1
-#       define EVERGREEN_ADDR_SURF_BANK_HEIGHT_4        2
-#       define EVERGREEN_ADDR_SURF_BANK_HEIGHT_8        3
-#       define EVERGREEN_GRPH_TILE_SPLIT(x)             (((x) & 0x7) << 13)
-#       define EVERGREEN_ADDR_SURF_TILE_SPLIT_64B       0
-#       define EVERGREEN_ADDR_SURF_TILE_SPLIT_128B      1
-#       define EVERGREEN_ADDR_SURF_TILE_SPLIT_256B      2
-#       define EVERGREEN_ADDR_SURF_TILE_SPLIT_512B      3
-#       define EVERGREEN_ADDR_SURF_TILE_SPLIT_1KB       4
-#       define EVERGREEN_ADDR_SURF_TILE_SPLIT_2KB       5
-#       define EVERGREEN_ADDR_SURF_TILE_SPLIT_4KB       6
-#       define EVERGREEN_GRPH_MACRO_TILE_ASPECT(x)      (((x) & 0x3) << 18)
-#       define EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_1  0
-#       define EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_2  1
-#       define EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_4  2
-#       define EVERGREEN_ADDR_SURF_MACRO_TILE_ASPECT_8  3
 #       define EVERGREEN_GRPH_ARRAY_MODE(x)             (((x) & 0x7) << 20)
 #       define EVERGREEN_GRPH_ARRAY_LINEAR_GENERAL      0
 #       define EVERGREEN_GRPH_ARRAY_LINEAR_ALIGNED      1
@@ -219,7 +182,6 @@
 #       define EVERGREEN_CRTC_MASTER_EN                 (1 << 0)
 #       define EVERGREEN_CRTC_DISP_READ_REQUEST_DISABLE (1 << 24)
 #define EVERGREEN_CRTC_STATUS                           0x6e8c
-#       define EVERGREEN_CRTC_V_BLANK                   (1 << 0)
 #define EVERGREEN_CRTC_STATUS_POSITION                  0x6e90
 #define EVERGREEN_MASTER_UPDATE_MODE                    0x6ef8
 #define EVERGREEN_CRTC_UPDATE_LOCK                      0x6ed4
@@ -229,9 +191,4 @@
 #define EVERGREEN_DC_GPIO_HPD_EN                        0x64b8
 #define EVERGREEN_DC_GPIO_HPD_Y                         0x64bc
 
-/* HDMI blocks at 0x7030, 0x7c30, 0x10830, 0x11430, 0x12030, 0x12c30 */
-#define EVERGREEN_HDMI_BASE				0x7030
-
-#define EVERGREEN_HDMI_CONFIG_OFFSET			0xf0
-
 #endif
diff --git a/drivers/gpu/drm/radeon/evergreend.h b/drivers/gpu/drm/radeon/evergreend.h
index d83fd9c..ab670c3 100644
--- a/drivers/gpu/drm/radeon/evergreend.h
+++ b/drivers/gpu/drm/radeon/evergreend.h
@@ -77,7 +77,6 @@
 
 #define	CONFIG_MEMSIZE					0x5428
 
-#define	CP_COHER_BASE					0x85F8
 #define CP_ME_CNTL					0x86D8
 #define		CP_ME_HALT					(1 << 28)
 #define		CP_PFP_HALT					(1 << 26)
@@ -109,7 +108,6 @@
 #define	CP_RB_WPTR_ADDR_HI				0xC11C
 #define	CP_RB_WPTR_DELAY				0x8704
 #define	CP_SEM_WAIT_TIMER				0x85BC
-#define	CP_SEM_INCOMPLETE_TIMER_CNTL			0x85C8
 #define	CP_DEBUG					0xC1FC
 
 
@@ -245,7 +243,6 @@
 #define	PA_CL_ENHANCE					0x8A14
 #define		CLIP_VTX_REORDER_ENA				(1 << 0)
 #define		NUM_CLIP_SEQ(x)					((x) << 1)
-#define	PA_SC_ENHANCE					0x8BF0
 #define PA_SC_AA_CONFIG					0x28C04
 #define         MSAA_NUM_SAMPLES_SHIFT                  0
 #define         MSAA_NUM_SAMPLES_MASK                   0x3
@@ -323,8 +320,6 @@
 #define	SQ_GPR_RESOURCE_MGMT_3				0x8C0C
 #define		NUM_HS_GPRS(x)					((x) << 0)
 #define		NUM_LS_GPRS(x)					((x) << 16)
-#define	SQ_GLOBAL_GPR_RESOURCE_MGMT_1			0x8C10
-#define	SQ_GLOBAL_GPR_RESOURCE_MGMT_2			0x8C14
 #define	SQ_THREAD_RESOURCE_MGMT				0x8C18
 #define		NUM_PS_THREADS(x)				((x) << 0)
 #define		NUM_VS_THREADS(x)				((x) << 8)
@@ -343,10 +338,6 @@
 #define		NUM_HS_STACK_ENTRIES(x)				((x) << 0)
 #define		NUM_LS_STACK_ENTRIES(x)				((x) << 16)
 #define	SQ_DYN_GPR_CNTL_PS_FLUSH_REQ    		0x8D8C
-#define	SQ_DYN_GPR_SIMD_LOCK_EN    			0x8D94
-#define	SQ_STATIC_THREAD_MGMT_1    			0x8E20
-#define	SQ_STATIC_THREAD_MGMT_2    			0x8E24
-#define	SQ_STATIC_THREAD_MGMT_3    			0x8E28
 #define	SQ_LDS_RESOURCE_MGMT    			0x8E2C
 
 #define	SQ_MS_FIFO_SIZES				0x8CF0
@@ -361,7 +352,6 @@
 #define		COLOR_BUFFER_SIZE(x)				((x) << 0)
 #define		POSITION_BUFFER_SIZE(x)				((x) << 8)
 #define		SMX_BUFFER_SIZE(x)				((x) << 16)
-#define	SX_MEMORY_EXPORT_BASE				0x9010
 #define	SX_MISC						0x28350
 
 #define CB_PERF_CTR0_SEL_0				0x9A20
@@ -705,7 +695,6 @@
 #define	PACKET3_DRAW_INDEX_MULTI_ELEMENT		0x36
 #define	PACKET3_MEM_SEMAPHORE				0x39
 #define	PACKET3_MPEG_INDEX				0x3A
-#define	PACKET3_COPY_DW					0x3B
 #define	PACKET3_WAIT_REG_MEM				0x3C
 #define	PACKET3_MEM_WRITE				0x3D
 #define	PACKET3_INDIRECT_BUFFER				0x32
@@ -783,8 +772,6 @@
 #define			SQ_TEX_VTX_VALID_TEXTURE			0x2
 #define			SQ_TEX_VTX_VALID_BUFFER				0x3
 
-#define VGT_VTX_VECT_EJECT_REG				0x88b0
-
 #define SQ_CONST_MEM_BASE				0x8df8
 
 #define SQ_ESGS_RING_BASE				0x8c40
@@ -909,152 +896,19 @@
 #define PA_SC_SCREEN_SCISSOR_TL                         0x28030
 #define PA_SC_GENERIC_SCISSOR_TL                        0x28240
 #define PA_SC_WINDOW_SCISSOR_TL                         0x28204
-
 #define VGT_PRIMITIVE_TYPE                              0x8958
-#define VGT_INDEX_TYPE                                  0x895C
-
-#define VGT_NUM_INDICES                                 0x8970
-
-#define VGT_COMPUTE_DIM_X                               0x8990
-#define VGT_COMPUTE_DIM_Y                               0x8994
-#define VGT_COMPUTE_DIM_Z                               0x8998
-#define VGT_COMPUTE_START_X                             0x899C
-#define VGT_COMPUTE_START_Y                             0x89A0
-#define VGT_COMPUTE_START_Z                             0x89A4
-#define VGT_COMPUTE_INDEX                               0x89A8
-#define VGT_COMPUTE_THREAD_GROUP_SIZE                   0x89AC
-#define VGT_HS_OFFCHIP_PARAM                            0x89B0
-
-#define DB_DEBUG					0x9830
-#define DB_DEBUG2					0x9834
-#define DB_DEBUG3					0x9838
-#define DB_DEBUG4					0x983C
-#define DB_WATERMARKS					0x9854
+
 #define DB_DEPTH_CONTROL				0x28800
-#define R_028800_DB_DEPTH_CONTROL                    0x028800
-#define   S_028800_STENCIL_ENABLE(x)                   (((x) & 0x1) << 0)
-#define   G_028800_STENCIL_ENABLE(x)                   (((x) >> 0) & 0x1)
-#define   C_028800_STENCIL_ENABLE                      0xFFFFFFFE
-#define   S_028800_Z_ENABLE(x)                         (((x) & 0x1) << 1)
-#define   G_028800_Z_ENABLE(x)                         (((x) >> 1) & 0x1)
-#define   C_028800_Z_ENABLE                            0xFFFFFFFD
-#define   S_028800_Z_WRITE_ENABLE(x)                   (((x) & 0x1) << 2)
-#define   G_028800_Z_WRITE_ENABLE(x)                   (((x) >> 2) & 0x1)
-#define   C_028800_Z_WRITE_ENABLE                      0xFFFFFFFB
-#define   S_028800_ZFUNC(x)                            (((x) & 0x7) << 4)
-#define   G_028800_ZFUNC(x)                            (((x) >> 4) & 0x7)
-#define   C_028800_ZFUNC                               0xFFFFFF8F
-#define   S_028800_BACKFACE_ENABLE(x)                  (((x) & 0x1) << 7)
-#define   G_028800_BACKFACE_ENABLE(x)                  (((x) >> 7) & 0x1)
-#define   C_028800_BACKFACE_ENABLE                     0xFFFFFF7F
-#define   S_028800_STENCILFUNC(x)                      (((x) & 0x7) << 8)
-#define   G_028800_STENCILFUNC(x)                      (((x) >> 8) & 0x7)
-#define   C_028800_STENCILFUNC                         0xFFFFF8FF
-#define     V_028800_STENCILFUNC_NEVER                 0x00000000
-#define     V_028800_STENCILFUNC_LESS                  0x00000001
-#define     V_028800_STENCILFUNC_EQUAL                 0x00000002
-#define     V_028800_STENCILFUNC_LEQUAL                0x00000003
-#define     V_028800_STENCILFUNC_GREATER               0x00000004
-#define     V_028800_STENCILFUNC_NOTEQUAL              0x00000005
-#define     V_028800_STENCILFUNC_GEQUAL                0x00000006
-#define     V_028800_STENCILFUNC_ALWAYS                0x00000007
-#define   S_028800_STENCILFAIL(x)                      (((x) & 0x7) << 11)
-#define   G_028800_STENCILFAIL(x)                      (((x) >> 11) & 0x7)
-#define   C_028800_STENCILFAIL                         0xFFFFC7FF
-#define     V_028800_STENCIL_KEEP                      0x00000000
-#define     V_028800_STENCIL_ZERO                      0x00000001
-#define     V_028800_STENCIL_REPLACE                   0x00000002
-#define     V_028800_STENCIL_INCR                      0x00000003
-#define     V_028800_STENCIL_DECR                      0x00000004
-#define     V_028800_STENCIL_INVERT                    0x00000005
-#define     V_028800_STENCIL_INCR_WRAP                 0x00000006
-#define     V_028800_STENCIL_DECR_WRAP                 0x00000007
-#define   S_028800_STENCILZPASS(x)                     (((x) & 0x7) << 14)
-#define   G_028800_STENCILZPASS(x)                     (((x) >> 14) & 0x7)
-#define   C_028800_STENCILZPASS                        0xFFFE3FFF
-#define   S_028800_STENCILZFAIL(x)                     (((x) & 0x7) << 17)
-#define   G_028800_STENCILZFAIL(x)                     (((x) >> 17) & 0x7)
-#define   C_028800_STENCILZFAIL                        0xFFF1FFFF
-#define   S_028800_STENCILFUNC_BF(x)                   (((x) & 0x7) << 20)
-#define   G_028800_STENCILFUNC_BF(x)                   (((x) >> 20) & 0x7)
-#define   C_028800_STENCILFUNC_BF                      0xFF8FFFFF
-#define   S_028800_STENCILFAIL_BF(x)                   (((x) & 0x7) << 23)
-#define   G_028800_STENCILFAIL_BF(x)                   (((x) >> 23) & 0x7)
-#define   C_028800_STENCILFAIL_BF                      0xFC7FFFFF
-#define   S_028800_STENCILZPASS_BF(x)                  (((x) & 0x7) << 26)
-#define   G_028800_STENCILZPASS_BF(x)                  (((x) >> 26) & 0x7)
-#define   C_028800_STENCILZPASS_BF                     0xE3FFFFFF
-#define   S_028800_STENCILZFAIL_BF(x)                  (((x) & 0x7) << 29)
-#define   G_028800_STENCILZFAIL_BF(x)                  (((x) >> 29) & 0x7)
-#define   C_028800_STENCILZFAIL_BF                     0x1FFFFFFF
 #define DB_DEPTH_VIEW					0x28008
-#define R_028008_DB_DEPTH_VIEW                       0x00028008
-#define   S_028008_SLICE_START(x)                      (((x) & 0x7FF) << 0)
-#define   G_028008_SLICE_START(x)                      (((x) >> 0) & 0x7FF)
-#define   C_028008_SLICE_START                         0xFFFFF800
-#define   S_028008_SLICE_MAX(x)                        (((x) & 0x7FF) << 13)
-#define   G_028008_SLICE_MAX(x)                        (((x) >> 13) & 0x7FF)
-#define   C_028008_SLICE_MAX                           0xFF001FFF
 #define DB_HTILE_DATA_BASE				0x28014
 #define DB_Z_INFO					0x28040
 #       define Z_ARRAY_MODE(x)                          ((x) << 4)
-#       define DB_TILE_SPLIT(x)                         (((x) & 0x7) << 8)
-#       define DB_NUM_BANKS(x)                          (((x) & 0x3) << 12)
-#       define DB_BANK_WIDTH(x)                         (((x) & 0x3) << 16)
-#       define DB_BANK_HEIGHT(x)                        (((x) & 0x3) << 20)
-#       define DB_MACRO_TILE_ASPECT(x)                  (((x) & 0x3) << 24)
-#define R_028040_DB_Z_INFO                       0x028040
-#define   S_028040_FORMAT(x)                           (((x) & 0x3) << 0)
-#define   G_028040_FORMAT(x)                           (((x) >> 0) & 0x3)
-#define   C_028040_FORMAT                              0xFFFFFFFC
-#define     V_028040_Z_INVALID                     0x00000000
-#define     V_028040_Z_16                          0x00000001
-#define     V_028040_Z_24                          0x00000002
-#define     V_028040_Z_32_FLOAT                    0x00000003
-#define   S_028040_ARRAY_MODE(x)                       (((x) & 0xF) << 4)
-#define   G_028040_ARRAY_MODE(x)                       (((x) >> 4) & 0xF)
-#define   C_028040_ARRAY_MODE                          0xFFFFFF0F
-#define   S_028040_READ_SIZE(x)                        (((x) & 0x1) << 28)
-#define   G_028040_READ_SIZE(x)                        (((x) >> 28) & 0x1)
-#define   C_028040_READ_SIZE                           0xEFFFFFFF
-#define   S_028040_TILE_SURFACE_ENABLE(x)              (((x) & 0x1) << 29)
-#define   G_028040_TILE_SURFACE_ENABLE(x)              (((x) >> 29) & 0x1)
-#define   C_028040_TILE_SURFACE_ENABLE                 0xDFFFFFFF
-#define   S_028040_ZRANGE_PRECISION(x)                 (((x) & 0x1) << 31)
-#define   G_028040_ZRANGE_PRECISION(x)                 (((x) >> 31) & 0x1)
-#define   C_028040_ZRANGE_PRECISION                    0x7FFFFFFF
-#define   S_028040_TILE_SPLIT(x)                       (((x) & 0x7) << 8)
-#define   G_028040_TILE_SPLIT(x)                       (((x) >> 8) & 0x7)
-#define   S_028040_NUM_BANKS(x)                        (((x) & 0x3) << 12)
-#define   G_028040_NUM_BANKS(x)                        (((x) >> 12) & 0x3)
-#define   S_028040_BANK_WIDTH(x)                       (((x) & 0x3) << 16)
-#define   G_028040_BANK_WIDTH(x)                       (((x) >> 16) & 0x3)
-#define   S_028040_BANK_HEIGHT(x)                      (((x) & 0x3) << 20)
-#define   G_028040_BANK_HEIGHT(x)                      (((x) >> 20) & 0x3)
-#define   S_028040_MACRO_TILE_ASPECT(x)                (((x) & 0x3) << 24)
-#define   G_028040_MACRO_TILE_ASPECT(x)                (((x) >> 24) & 0x3)
 #define DB_STENCIL_INFO					0x28044
-#define R_028044_DB_STENCIL_INFO                     0x028044
-#define   S_028044_FORMAT(x)                           (((x) & 0x1) << 0)
-#define   G_028044_FORMAT(x)                           (((x) >> 0) & 0x1)
-#define   C_028044_FORMAT                              0xFFFFFFFE
-#define   G_028044_TILE_SPLIT(x)                       (((x) >> 8) & 0x7)
 #define DB_Z_READ_BASE					0x28048
 #define DB_STENCIL_READ_BASE				0x2804c
 #define DB_Z_WRITE_BASE					0x28050
 #define DB_STENCIL_WRITE_BASE				0x28054
 #define DB_DEPTH_SIZE					0x28058
-#define R_028058_DB_DEPTH_SIZE                       0x028058
-#define   S_028058_PITCH_TILE_MAX(x)                   (((x) & 0x7FF) << 0)
-#define   G_028058_PITCH_TILE_MAX(x)                   (((x) >> 0) & 0x7FF)
-#define   C_028058_PITCH_TILE_MAX                      0xFFFFF800
-#define   S_028058_HEIGHT_TILE_MAX(x)                   (((x) & 0x7FF) << 11)
-#define   G_028058_HEIGHT_TILE_MAX(x)                   (((x) >> 11) & 0x7FF)
-#define   C_028058_HEIGHT_TILE_MAX                      0xFFC007FF
-#define R_02805C_DB_DEPTH_SLICE                      0x02805C
-#define   S_02805C_SLICE_TILE_MAX(x)                   (((x) & 0x3FFFFF) << 0)
-#define   G_02805C_SLICE_TILE_MAX(x)                   (((x) >> 0) & 0x3FFFFF)
-#define   C_02805C_SLICE_TILE_MAX                      0xFFC00000
 
 #define SQ_PGM_START_PS					0x28840
 #define SQ_PGM_START_VS					0x2885c
@@ -1064,14 +918,6 @@
 #define SQ_PGM_START_HS					0x288b8
 #define SQ_PGM_START_LS					0x288d0
 
-#define	VGT_STRMOUT_BUFFER_BASE_0			0x28AD8
-#define	VGT_STRMOUT_BUFFER_BASE_1			0x28AE8
-#define	VGT_STRMOUT_BUFFER_BASE_2			0x28AF8
-#define	VGT_STRMOUT_BUFFER_BASE_3			0x28B08
-#define VGT_STRMOUT_BUFFER_SIZE_0			0x28AD0
-#define VGT_STRMOUT_BUFFER_SIZE_1			0x28AE0
-#define VGT_STRMOUT_BUFFER_SIZE_2			0x28AF0
-#define VGT_STRMOUT_BUFFER_SIZE_3			0x28B00
 #define VGT_STRMOUT_CONFIG				0x28b94
 #define VGT_STRMOUT_BUFFER_CONFIG			0x28b98
 
@@ -1098,163 +944,13 @@
 #define	CB_COLOR0_PITCH					0x28c64
 #define	CB_COLOR0_SLICE					0x28c68
 #define	CB_COLOR0_VIEW					0x28c6c
-#define R_028C6C_CB_COLOR0_VIEW                      0x00028C6C
-#define   S_028C6C_SLICE_START(x)                      (((x) & 0x7FF) << 0)
-#define   G_028C6C_SLICE_START(x)                      (((x) >> 0) & 0x7FF)
-#define   C_028C6C_SLICE_START                         0xFFFFF800
-#define   S_028C6C_SLICE_MAX(x)                        (((x) & 0x7FF) << 13)
-#define   G_028C6C_SLICE_MAX(x)                        (((x) >> 13) & 0x7FF)
-#define   C_028C6C_SLICE_MAX                           0xFF001FFF
-#define R_028C70_CB_COLOR0_INFO                      0x028C70
-#define   S_028C70_ENDIAN(x)                           (((x) & 0x3) << 0)
-#define   G_028C70_ENDIAN(x)                           (((x) >> 0) & 0x3)
-#define   C_028C70_ENDIAN                              0xFFFFFFFC
-#define   S_028C70_FORMAT(x)                           (((x) & 0x3F) << 2)
-#define   G_028C70_FORMAT(x)                           (((x) >> 2) & 0x3F)
-#define   C_028C70_FORMAT                              0xFFFFFF03
-#define     V_028C70_COLOR_INVALID                     0x00000000
-#define     V_028C70_COLOR_8                           0x00000001
-#define     V_028C70_COLOR_4_4                         0x00000002
-#define     V_028C70_COLOR_3_3_2                       0x00000003
-#define     V_028C70_COLOR_16                          0x00000005
-#define     V_028C70_COLOR_16_FLOAT                    0x00000006
-#define     V_028C70_COLOR_8_8                         0x00000007
-#define     V_028C70_COLOR_5_6_5                       0x00000008
-#define     V_028C70_COLOR_6_5_5                       0x00000009
-#define     V_028C70_COLOR_1_5_5_5                     0x0000000A
-#define     V_028C70_COLOR_4_4_4_4                     0x0000000B
-#define     V_028C70_COLOR_5_5_5_1                     0x0000000C
-#define     V_028C70_COLOR_32                          0x0000000D
-#define     V_028C70_COLOR_32_FLOAT                    0x0000000E
-#define     V_028C70_COLOR_16_16                       0x0000000F
-#define     V_028C70_COLOR_16_16_FLOAT                 0x00000010
-#define     V_028C70_COLOR_8_24                        0x00000011
-#define     V_028C70_COLOR_8_24_FLOAT                  0x00000012
-#define     V_028C70_COLOR_24_8                        0x00000013
-#define     V_028C70_COLOR_24_8_FLOAT                  0x00000014
-#define     V_028C70_COLOR_10_11_11                    0x00000015
-#define     V_028C70_COLOR_10_11_11_FLOAT              0x00000016
-#define     V_028C70_COLOR_11_11_10                    0x00000017
-#define     V_028C70_COLOR_11_11_10_FLOAT              0x00000018
-#define     V_028C70_COLOR_2_10_10_10                  0x00000019
-#define     V_028C70_COLOR_8_8_8_8                     0x0000001A
-#define     V_028C70_COLOR_10_10_10_2                  0x0000001B
-#define     V_028C70_COLOR_X24_8_32_FLOAT              0x0000001C
-#define     V_028C70_COLOR_32_32                       0x0000001D
-#define     V_028C70_COLOR_32_32_FLOAT                 0x0000001E
-#define     V_028C70_COLOR_16_16_16_16                 0x0000001F
-#define     V_028C70_COLOR_16_16_16_16_FLOAT           0x00000020
-#define     V_028C70_COLOR_32_32_32_32                 0x00000022
-#define     V_028C70_COLOR_32_32_32_32_FLOAT           0x00000023
-#define     V_028C70_COLOR_32_32_32_FLOAT              0x00000030
-#define   S_028C70_ARRAY_MODE(x)                       (((x) & 0xF) << 8)
-#define   G_028C70_ARRAY_MODE(x)                       (((x) >> 8) & 0xF)
-#define   C_028C70_ARRAY_MODE                          0xFFFFF0FF
-#define     V_028C70_ARRAY_LINEAR_GENERAL              0x00000000
-#define     V_028C70_ARRAY_LINEAR_ALIGNED              0x00000001
-#define     V_028C70_ARRAY_1D_TILED_THIN1              0x00000002
-#define     V_028C70_ARRAY_2D_TILED_THIN1              0x00000004
-#define   S_028C70_NUMBER_TYPE(x)                      (((x) & 0x7) << 12)
-#define   G_028C70_NUMBER_TYPE(x)                      (((x) >> 12) & 0x7)
-#define   C_028C70_NUMBER_TYPE                         0xFFFF8FFF
-#define     V_028C70_NUMBER_UNORM                      0x00000000
-#define     V_028C70_NUMBER_SNORM                      0x00000001
-#define     V_028C70_NUMBER_USCALED                    0x00000002
-#define     V_028C70_NUMBER_SSCALED                    0x00000003
-#define     V_028C70_NUMBER_UINT                       0x00000004
-#define     V_028C70_NUMBER_SINT                       0x00000005
-#define     V_028C70_NUMBER_SRGB                       0x00000006
-#define     V_028C70_NUMBER_FLOAT                      0x00000007
-#define   S_028C70_COMP_SWAP(x)                        (((x) & 0x3) << 15)
-#define   G_028C70_COMP_SWAP(x)                        (((x) >> 15) & 0x3)
-#define   C_028C70_COMP_SWAP                           0xFFFE7FFF
-#define     V_028C70_SWAP_STD                          0x00000000
-#define     V_028C70_SWAP_ALT                          0x00000001
-#define     V_028C70_SWAP_STD_REV                      0x00000002
-#define     V_028C70_SWAP_ALT_REV                      0x00000003
-#define   S_028C70_FAST_CLEAR(x)                       (((x) & 0x1) << 17)
-#define   G_028C70_FAST_CLEAR(x)                       (((x) >> 17) & 0x1)
-#define   C_028C70_FAST_CLEAR                          0xFFFDFFFF
-#define   S_028C70_COMPRESSION(x)                      (((x) & 0x3) << 18)
-#define   G_028C70_COMPRESSION(x)                      (((x) >> 18) & 0x3)
-#define   C_028C70_COMPRESSION                         0xFFF3FFFF
-#define   S_028C70_BLEND_CLAMP(x)                      (((x) & 0x1) << 19)
-#define   G_028C70_BLEND_CLAMP(x)                      (((x) >> 19) & 0x1)
-#define   C_028C70_BLEND_CLAMP                         0xFFF7FFFF
-#define   S_028C70_BLEND_BYPASS(x)                     (((x) & 0x1) << 20)
-#define   G_028C70_BLEND_BYPASS(x)                     (((x) >> 20) & 0x1)
-#define   C_028C70_BLEND_BYPASS                        0xFFEFFFFF
-#define   S_028C70_SIMPLE_FLOAT(x)                     (((x) & 0x1) << 21)
-#define   G_028C70_SIMPLE_FLOAT(x)                     (((x) >> 21) & 0x1)
-#define   C_028C70_SIMPLE_FLOAT                        0xFFDFFFFF
-#define   S_028C70_ROUND_MODE(x)                       (((x) & 0x1) << 22)
-#define   G_028C70_ROUND_MODE(x)                       (((x) >> 22) & 0x1)
-#define   C_028C70_ROUND_MODE                          0xFFBFFFFF
-#define   S_028C70_TILE_COMPACT(x)                     (((x) & 0x1) << 23)
-#define   G_028C70_TILE_COMPACT(x)                     (((x) >> 23) & 0x1)
-#define   C_028C70_TILE_COMPACT                        0xFF7FFFFF
-#define   S_028C70_SOURCE_FORMAT(x)                    (((x) & 0x3) << 24)
-#define   G_028C70_SOURCE_FORMAT(x)                    (((x) >> 24) & 0x3)
-#define   C_028C70_SOURCE_FORMAT                       0xFCFFFFFF
-#define     V_028C70_EXPORT_4C_32BPC                   0x0
-#define     V_028C70_EXPORT_4C_16BPC                   0x1
-#define     V_028C70_EXPORT_2C_32BPC                   0x2 /* Do not use */
-#define   S_028C70_RAT(x)                              (((x) & 0x1) << 26)
-#define   G_028C70_RAT(x)                              (((x) >> 26) & 0x1)
-#define   C_028C70_RAT                                 0xFBFFFFFF
-#define   S_028C70_RESOURCE_TYPE(x)                    (((x) & 0x7) << 27)
-#define   G_028C70_RESOURCE_TYPE(x)                    (((x) >> 27) & 0x7)
-#define   C_028C70_RESOURCE_TYPE                       0xC7FFFFFF
-
 #define	CB_COLOR0_INFO					0x28c70
-#	define CB_FORMAT(x)				((x) << 2)
 #       define CB_ARRAY_MODE(x)                         ((x) << 8)
 #       define ARRAY_LINEAR_GENERAL                     0
 #       define ARRAY_LINEAR_ALIGNED                     1
 #       define ARRAY_1D_TILED_THIN1                     2
 #       define ARRAY_2D_TILED_THIN1                     4
-#	define CB_SOURCE_FORMAT(x)			((x) << 24)
-#	define CB_SF_EXPORT_FULL			0
-#	define CB_SF_EXPORT_NORM			1
-#define R_028C74_CB_COLOR0_ATTRIB                      0x028C74
-#define   S_028C74_NON_DISP_TILING_ORDER(x)            (((x) & 0x1) << 4)
-#define   G_028C74_NON_DISP_TILING_ORDER(x)            (((x) >> 4) & 0x1)
-#define   C_028C74_NON_DISP_TILING_ORDER               0xFFFFFFEF
-#define   S_028C74_TILE_SPLIT(x)                       (((x) & 0xf) << 5)
-#define   G_028C74_TILE_SPLIT(x)                       (((x) >> 5) & 0xf)
-#define   S_028C74_NUM_BANKS(x)                        (((x) & 0x3) << 10)
-#define   G_028C74_NUM_BANKS(x)                        (((x) >> 10) & 0x3)
-#define   S_028C74_BANK_WIDTH(x)                       (((x) & 0x3) << 13)
-#define   G_028C74_BANK_WIDTH(x)                       (((x) >> 13) & 0x3)
-#define   S_028C74_BANK_HEIGHT(x)                      (((x) & 0x3) << 16)
-#define   G_028C74_BANK_HEIGHT(x)                      (((x) >> 16) & 0x3)
-#define   S_028C74_MACRO_TILE_ASPECT(x)                (((x) & 0x3) << 19)
-#define   G_028C74_MACRO_TILE_ASPECT(x)                (((x) >> 19) & 0x3)
 #define	CB_COLOR0_ATTRIB				0x28c74
-#       define CB_TILE_SPLIT(x)                         (((x) & 0x7) << 5)
-#       define ADDR_SURF_TILE_SPLIT_64B                 0
-#       define ADDR_SURF_TILE_SPLIT_128B                1
-#       define ADDR_SURF_TILE_SPLIT_256B                2
-#       define ADDR_SURF_TILE_SPLIT_512B                3
-#       define ADDR_SURF_TILE_SPLIT_1KB                 4
-#       define ADDR_SURF_TILE_SPLIT_2KB                 5
-#       define ADDR_SURF_TILE_SPLIT_4KB                 6
-#       define CB_NUM_BANKS(x)                          (((x) & 0x3) << 10)
-#       define ADDR_SURF_2_BANK                         0
-#       define ADDR_SURF_4_BANK                         1
-#       define ADDR_SURF_8_BANK                         2
-#       define ADDR_SURF_16_BANK                        3
-#       define CB_BANK_WIDTH(x)                         (((x) & 0x3) << 13)
-#       define ADDR_SURF_BANK_WIDTH_1                   0
-#       define ADDR_SURF_BANK_WIDTH_2                   1
-#       define ADDR_SURF_BANK_WIDTH_4                   2
-#       define ADDR_SURF_BANK_WIDTH_8                   3
-#       define CB_BANK_HEIGHT(x)                        (((x) & 0x3) << 16)
-#       define ADDR_SURF_BANK_HEIGHT_1                  0
-#       define ADDR_SURF_BANK_HEIGHT_2                  1
-#       define ADDR_SURF_BANK_HEIGHT_4                  2
-#       define ADDR_SURF_BANK_HEIGHT_8                  3
-#       define CB_MACRO_TILE_ASPECT(x)                  (((x) & 0x3) << 19)
 #define	CB_COLOR0_DIM					0x28c78
 /* only CB0-7 blocks have these regs */
 #define	CB_COLOR0_CMASK					0x28c7c
@@ -1415,232 +1111,22 @@
 #define	CB_COLOR7_CLEAR_WORD3				0x28e3c
 
 #define SQ_TEX_RESOURCE_WORD0_0                         0x30000
-#	define TEX_DIM(x)				((x) << 0)
-#	define SQ_TEX_DIM_1D				0
-#	define SQ_TEX_DIM_2D				1
-#	define SQ_TEX_DIM_3D				2
-#	define SQ_TEX_DIM_CUBEMAP			3
-#	define SQ_TEX_DIM_1D_ARRAY			4
-#	define SQ_TEX_DIM_2D_ARRAY			5
-#	define SQ_TEX_DIM_2D_MSAA			6
-#	define SQ_TEX_DIM_2D_ARRAY_MSAA			7
 #define SQ_TEX_RESOURCE_WORD1_0                         0x30004
 #       define TEX_ARRAY_MODE(x)                        ((x) << 28)
 #define SQ_TEX_RESOURCE_WORD2_0                         0x30008
 #define SQ_TEX_RESOURCE_WORD3_0                         0x3000C
 #define SQ_TEX_RESOURCE_WORD4_0                         0x30010
-#	define TEX_DST_SEL_X(x)				((x) << 16)
-#	define TEX_DST_SEL_Y(x)				((x) << 19)
-#	define TEX_DST_SEL_Z(x)				((x) << 22)
-#	define TEX_DST_SEL_W(x)				((x) << 25)
-#	define SQ_SEL_X					0
-#	define SQ_SEL_Y					1
-#	define SQ_SEL_Z					2
-#	define SQ_SEL_W					3
-#	define SQ_SEL_0					4
-#	define SQ_SEL_1					5
 #define SQ_TEX_RESOURCE_WORD5_0                         0x30014
 #define SQ_TEX_RESOURCE_WORD6_0                         0x30018
-#       define TEX_TILE_SPLIT(x)                        (((x) & 0x7) << 29)
 #define SQ_TEX_RESOURCE_WORD7_0                         0x3001c
-#       define MACRO_TILE_ASPECT(x)                     (((x) & 0x3) << 6)
-#       define TEX_BANK_WIDTH(x)                        (((x) & 0x3) << 8)
-#       define TEX_BANK_HEIGHT(x)                       (((x) & 0x3) << 10)
-#       define TEX_NUM_BANKS(x)                         (((x) & 0x3) << 16)
-#define R_030000_SQ_TEX_RESOURCE_WORD0_0             0x030000
-#define   S_030000_DIM(x)                              (((x) & 0x7) << 0)
-#define   G_030000_DIM(x)                              (((x) >> 0) & 0x7)
-#define   C_030000_DIM                                 0xFFFFFFF8
-#define     V_030000_SQ_TEX_DIM_1D                     0x00000000
-#define     V_030000_SQ_TEX_DIM_2D                     0x00000001
-#define     V_030000_SQ_TEX_DIM_3D                     0x00000002
-#define     V_030000_SQ_TEX_DIM_CUBEMAP                0x00000003
-#define     V_030000_SQ_TEX_DIM_1D_ARRAY               0x00000004
-#define     V_030000_SQ_TEX_DIM_2D_ARRAY               0x00000005
-#define     V_030000_SQ_TEX_DIM_2D_MSAA                0x00000006
-#define     V_030000_SQ_TEX_DIM_2D_ARRAY_MSAA          0x00000007
-#define   S_030000_NON_DISP_TILING_ORDER(x)            (((x) & 0x1) << 5)
-#define   G_030000_NON_DISP_TILING_ORDER(x)            (((x) >> 5) & 0x1)
-#define   C_030000_NON_DISP_TILING_ORDER               0xFFFFFFDF
-#define   S_030000_PITCH(x)                            (((x) & 0xFFF) << 6)
-#define   G_030000_PITCH(x)                            (((x) >> 6) & 0xFFF)
-#define   C_030000_PITCH                               0xFFFC003F
-#define   S_030000_TEX_WIDTH(x)                        (((x) & 0x3FFF) << 18)
-#define   G_030000_TEX_WIDTH(x)                        (((x) >> 18) & 0x3FFF)
-#define   C_030000_TEX_WIDTH                           0x0003FFFF
-#define R_030004_SQ_TEX_RESOURCE_WORD1_0             0x030004
-#define   S_030004_TEX_HEIGHT(x)                       (((x) & 0x3FFF) << 0)
-#define   G_030004_TEX_HEIGHT(x)                       (((x) >> 0) & 0x3FFF)
-#define   C_030004_TEX_HEIGHT                          0xFFFFC000
-#define   S_030004_TEX_DEPTH(x)                        (((x) & 0x1FFF) << 14)
-#define   G_030004_TEX_DEPTH(x)                        (((x) >> 14) & 0x1FFF)
-#define   C_030004_TEX_DEPTH                           0xF8003FFF
-#define   S_030004_ARRAY_MODE(x)                       (((x) & 0xF) << 28)
-#define   G_030004_ARRAY_MODE(x)                       (((x) >> 28) & 0xF)
-#define   C_030004_ARRAY_MODE                          0x0FFFFFFF
-#define R_030008_SQ_TEX_RESOURCE_WORD2_0             0x030008
-#define   S_030008_BASE_ADDRESS(x)                     (((x) & 0xFFFFFFFF) << 0)
-#define   G_030008_BASE_ADDRESS(x)                     (((x) >> 0) & 0xFFFFFFFF)
-#define   C_030008_BASE_ADDRESS                        0x00000000
-#define R_03000C_SQ_TEX_RESOURCE_WORD3_0             0x03000C
-#define   S_03000C_MIP_ADDRESS(x)                      (((x) & 0xFFFFFFFF) << 0)
-#define   G_03000C_MIP_ADDRESS(x)                      (((x) >> 0) & 0xFFFFFFFF)
-#define   C_03000C_MIP_ADDRESS                         0x00000000
-#define R_030010_SQ_TEX_RESOURCE_WORD4_0             0x030010
-#define   S_030010_FORMAT_COMP_X(x)                    (((x) & 0x3) << 0)
-#define   G_030010_FORMAT_COMP_X(x)                    (((x) >> 0) & 0x3)
-#define   C_030010_FORMAT_COMP_X                       0xFFFFFFFC
-#define     V_030010_SQ_FORMAT_COMP_UNSIGNED           0x00000000
-#define     V_030010_SQ_FORMAT_COMP_SIGNED             0x00000001
-#define     V_030010_SQ_FORMAT_COMP_UNSIGNED_BIASED    0x00000002
-#define   S_030010_FORMAT_COMP_Y(x)                    (((x) & 0x3) << 2)
-#define   G_030010_FORMAT_COMP_Y(x)                    (((x) >> 2) & 0x3)
-#define   C_030010_FORMAT_COMP_Y                       0xFFFFFFF3
-#define   S_030010_FORMAT_COMP_Z(x)                    (((x) & 0x3) << 4)
-#define   G_030010_FORMAT_COMP_Z(x)                    (((x) >> 4) & 0x3)
-#define   C_030010_FORMAT_COMP_Z                       0xFFFFFFCF
-#define   S_030010_FORMAT_COMP_W(x)                    (((x) & 0x3) << 6)
-#define   G_030010_FORMAT_COMP_W(x)                    (((x) >> 6) & 0x3)
-#define   C_030010_FORMAT_COMP_W                       0xFFFFFF3F
-#define   S_030010_NUM_FORMAT_ALL(x)                   (((x) & 0x3) << 8)
-#define   G_030010_NUM_FORMAT_ALL(x)                   (((x) >> 8) & 0x3)
-#define   C_030010_NUM_FORMAT_ALL                      0xFFFFFCFF
-#define     V_030010_SQ_NUM_FORMAT_NORM                0x00000000
-#define     V_030010_SQ_NUM_FORMAT_INT                 0x00000001
-#define     V_030010_SQ_NUM_FORMAT_SCALED              0x00000002
-#define   S_030010_SRF_MODE_ALL(x)                     (((x) & 0x1) << 10)
-#define   G_030010_SRF_MODE_ALL(x)                     (((x) >> 10) & 0x1)
-#define   C_030010_SRF_MODE_ALL                        0xFFFFFBFF
-#define     V_030010_SRF_MODE_ZERO_CLAMP_MINUS_ONE     0x00000000
-#define     V_030010_SRF_MODE_NO_ZERO                  0x00000001
-#define   S_030010_FORCE_DEGAMMA(x)                    (((x) & 0x1) << 11)
-#define   G_030010_FORCE_DEGAMMA(x)                    (((x) >> 11) & 0x1)
-#define   C_030010_FORCE_DEGAMMA                       0xFFFFF7FF
-#define   S_030010_ENDIAN_SWAP(x)                      (((x) & 0x3) << 12)
-#define   G_030010_ENDIAN_SWAP(x)                      (((x) >> 12) & 0x3)
-#define   C_030010_ENDIAN_SWAP                         0xFFFFCFFF
-#define   S_030010_DST_SEL_X(x)                        (((x) & 0x7) << 16)
-#define   G_030010_DST_SEL_X(x)                        (((x) >> 16) & 0x7)
-#define   C_030010_DST_SEL_X                           0xFFF8FFFF
-#define     V_030010_SQ_SEL_X                          0x00000000
-#define     V_030010_SQ_SEL_Y                          0x00000001
-#define     V_030010_SQ_SEL_Z                          0x00000002
-#define     V_030010_SQ_SEL_W                          0x00000003
-#define     V_030010_SQ_SEL_0                          0x00000004
-#define     V_030010_SQ_SEL_1                          0x00000005
-#define   S_030010_DST_SEL_Y(x)                        (((x) & 0x7) << 19)
-#define   G_030010_DST_SEL_Y(x)                        (((x) >> 19) & 0x7)
-#define   C_030010_DST_SEL_Y                           0xFFC7FFFF
-#define   S_030010_DST_SEL_Z(x)                        (((x) & 0x7) << 22)
-#define   G_030010_DST_SEL_Z(x)                        (((x) >> 22) & 0x7)
-#define   C_030010_DST_SEL_Z                           0xFE3FFFFF
-#define   S_030010_DST_SEL_W(x)                        (((x) & 0x7) << 25)
-#define   G_030010_DST_SEL_W(x)                        (((x) >> 25) & 0x7)
-#define   C_030010_DST_SEL_W                           0xF1FFFFFF
-#define   S_030010_BASE_LEVEL(x)                       (((x) & 0xF) << 28)
-#define   G_030010_BASE_LEVEL(x)                       (((x) >> 28) & 0xF)
-#define   C_030010_BASE_LEVEL                          0x0FFFFFFF
-#define R_030014_SQ_TEX_RESOURCE_WORD5_0             0x030014
-#define   S_030014_LAST_LEVEL(x)                       (((x) & 0xF) << 0)
-#define   G_030014_LAST_LEVEL(x)                       (((x) >> 0) & 0xF)
-#define   C_030014_LAST_LEVEL                          0xFFFFFFF0
-#define   S_030014_BASE_ARRAY(x)                       (((x) & 0x1FFF) << 4)
-#define   G_030014_BASE_ARRAY(x)                       (((x) >> 4) & 0x1FFF)
-#define   C_030014_BASE_ARRAY                          0xFFFE000F
-#define   S_030014_LAST_ARRAY(x)                       (((x) & 0x1FFF) << 17)
-#define   G_030014_LAST_ARRAY(x)                       (((x) >> 17) & 0x1FFF)
-#define   C_030014_LAST_ARRAY                          0xC001FFFF
-#define R_030018_SQ_TEX_RESOURCE_WORD6_0             0x030018
-#define   S_030018_MAX_ANISO(x)                        (((x) & 0x7) << 0)
-#define   G_030018_MAX_ANISO(x)                        (((x) >> 0) & 0x7)
-#define   C_030018_MAX_ANISO                           0xFFFFFFF8
-#define   S_030018_PERF_MODULATION(x)                  (((x) & 0x7) << 3)
-#define   G_030018_PERF_MODULATION(x)                  (((x) >> 3) & 0x7)
-#define   C_030018_PERF_MODULATION                     0xFFFFFFC7
-#define   S_030018_INTERLACED(x)                       (((x) & 0x1) << 6)
-#define   G_030018_INTERLACED(x)                       (((x) >> 6) & 0x1)
-#define   C_030018_INTERLACED                          0xFFFFFFBF
-#define   S_030018_TILE_SPLIT(x)                       (((x) & 0x7) << 29)
-#define   G_030018_TILE_SPLIT(x)                       (((x) >> 29) & 0x7)
-#define R_03001C_SQ_TEX_RESOURCE_WORD7_0             0x03001C
-#define   S_03001C_MACRO_TILE_ASPECT(x)                (((x) & 0x3) << 6)
-#define   G_03001C_MACRO_TILE_ASPECT(x)                (((x) >> 6) & 0x3)
-#define   S_03001C_BANK_WIDTH(x)                       (((x) & 0x3) << 8)
-#define   G_03001C_BANK_WIDTH(x)                       (((x) >> 8) & 0x3)
-#define   S_03001C_BANK_HEIGHT(x)                      (((x) & 0x3) << 10)
-#define   G_03001C_BANK_HEIGHT(x)                      (((x) >> 10) & 0x3)
-#define   S_03001C_NUM_BANKS(x)                        (((x) & 0x3) << 16)
-#define   G_03001C_NUM_BANKS(x)                        (((x) >> 16) & 0x3)
-#define   S_03001C_TYPE(x)                             (((x) & 0x3) << 30)
-#define   G_03001C_TYPE(x)                             (((x) >> 30) & 0x3)
-#define   C_03001C_TYPE                                0x3FFFFFFF
-#define     V_03001C_SQ_TEX_VTX_INVALID_TEXTURE        0x00000000
-#define     V_03001C_SQ_TEX_VTX_INVALID_BUFFER         0x00000001
-#define     V_03001C_SQ_TEX_VTX_VALID_TEXTURE          0x00000002
-#define     V_03001C_SQ_TEX_VTX_VALID_BUFFER           0x00000003
-#define   S_03001C_DATA_FORMAT(x)                      (((x) & 0x3F) << 0)
-#define   G_03001C_DATA_FORMAT(x)                      (((x) >> 0) & 0x3F)
-#define   C_03001C_DATA_FORMAT                         0xFFFFFFC0
-
-#define SQ_VTX_CONSTANT_WORD0_0				0x30000
-#define SQ_VTX_CONSTANT_WORD1_0				0x30004
-#define SQ_VTX_CONSTANT_WORD2_0				0x30008
-#	define SQ_VTXC_BASE_ADDR_HI(x)			((x) << 0)
-#	define SQ_VTXC_STRIDE(x)			((x) << 8)
-#	define SQ_VTXC_ENDIAN_SWAP(x)			((x) << 30)
-#	define SQ_ENDIAN_NONE				0
-#	define SQ_ENDIAN_8IN16				1
-#	define SQ_ENDIAN_8IN32				2
-#define SQ_VTX_CONSTANT_WORD3_0				0x3000C
-#	define SQ_VTCX_SEL_X(x)				((x) << 3)
-#	define SQ_VTCX_SEL_Y(x)				((x) << 6)
-#	define SQ_VTCX_SEL_Z(x)				((x) << 9)
-#	define SQ_VTCX_SEL_W(x)				((x) << 12)
-#define SQ_VTX_CONSTANT_WORD4_0				0x30010
-#define SQ_VTX_CONSTANT_WORD5_0                         0x30014
-#define SQ_VTX_CONSTANT_WORD6_0                         0x30018
-#define SQ_VTX_CONSTANT_WORD7_0                         0x3001c
-
-#define TD_PS_BORDER_COLOR_INDEX                        0xA400
-#define TD_PS_BORDER_COLOR_RED                          0xA404
-#define TD_PS_BORDER_COLOR_GREEN                        0xA408
-#define TD_PS_BORDER_COLOR_BLUE                         0xA40C
-#define TD_PS_BORDER_COLOR_ALPHA                        0xA410
-#define TD_VS_BORDER_COLOR_INDEX                        0xA414
-#define TD_VS_BORDER_COLOR_RED                          0xA418
-#define TD_VS_BORDER_COLOR_GREEN                        0xA41C
-#define TD_VS_BORDER_COLOR_BLUE                         0xA420
-#define TD_VS_BORDER_COLOR_ALPHA                        0xA424
-#define TD_GS_BORDER_COLOR_INDEX                        0xA428
-#define TD_GS_BORDER_COLOR_RED                          0xA42C
-#define TD_GS_BORDER_COLOR_GREEN                        0xA430
-#define TD_GS_BORDER_COLOR_BLUE                         0xA434
-#define TD_GS_BORDER_COLOR_ALPHA                        0xA438
-#define TD_HS_BORDER_COLOR_INDEX                        0xA43C
-#define TD_HS_BORDER_COLOR_RED                          0xA440
-#define TD_HS_BORDER_COLOR_GREEN                        0xA444
-#define TD_HS_BORDER_COLOR_BLUE                         0xA448
-#define TD_HS_BORDER_COLOR_ALPHA                        0xA44C
-#define TD_LS_BORDER_COLOR_INDEX                        0xA450
-#define TD_LS_BORDER_COLOR_RED                          0xA454
-#define TD_LS_BORDER_COLOR_GREEN                        0xA458
-#define TD_LS_BORDER_COLOR_BLUE                         0xA45C
-#define TD_LS_BORDER_COLOR_ALPHA                        0xA460
-#define TD_CS_BORDER_COLOR_INDEX                        0xA464
-#define TD_CS_BORDER_COLOR_RED                          0xA468
-#define TD_CS_BORDER_COLOR_GREEN                        0xA46C
-#define TD_CS_BORDER_COLOR_BLUE                         0xA470
-#define TD_CS_BORDER_COLOR_ALPHA                        0xA474
 
 /* cayman 3D regs */
-#define CAYMAN_VGT_OFFCHIP_LDS_BASE			0x89B4
-#define CAYMAN_SQ_EX_ALLOC_TABLE_SLOTS			0x8E48
+#define CAYMAN_VGT_OFFCHIP_LDS_BASE			0x89B0
 #define CAYMAN_DB_EQAA					0x28804
 #define CAYMAN_DB_DEPTH_INFO				0x2803C
 #define CAYMAN_PA_SC_AA_CONFIG				0x28BE0
 #define         CAYMAN_MSAA_NUM_SAMPLES_SHIFT           0
 #define         CAYMAN_MSAA_NUM_SAMPLES_MASK            0x7
-#define CAYMAN_SX_SCATTER_EXPORT_BASE			0x28358
 /* cayman packet3 addition */
 #define	CAYMAN_PACKET3_DEALLOC_STATE			0x14
 
diff --git a/drivers/gpu/drm/radeon/ni.c b/drivers/gpu/drm/radeon/ni.c
index 160799c..0c460c4 100644
--- a/drivers/gpu/drm/radeon/ni.c
+++ b/drivers/gpu/drm/radeon/ni.c
@@ -24,7 +24,6 @@
 #include <linux/firmware.h>
 #include <linux/platform_device.h>
 #include <linux/slab.h>
-#include <linux/module.h>
 #include "drmP.h"
 #include "radeon.h"
 #include "radeon_asic.h"
@@ -41,7 +40,6 @@ extern void evergreen_mc_program(struct radeon_device *rdev);
 extern void evergreen_irq_suspend(struct radeon_device *rdev);
 extern int evergreen_mc_init(struct radeon_device *rdev);
 extern void evergreen_fix_pci_max_read_req_size(struct radeon_device *rdev);
-extern void evergreen_pcie_gen2_enable(struct radeon_device *rdev);
 
 #define EVERGREEN_PFP_UCODE_SIZE 1120
 #define EVERGREEN_PM4_UCODE_SIZE 1376
@@ -262,11 +260,8 @@ int ni_mc_load_microcode(struct radeon_device *rdev)
 		WREG32(MC_SEQ_SUP_CNTL, 0x00000001);
 
 		/* wait for training to complete */
-		for (i = 0; i < rdev->usec_timeout; i++) {
-			if (RREG32(MC_IO_PAD_CNTL_D0) & MEM_FALL_OUT_CMD)
-				break;
-			udelay(1);
-		}
+		while (!(RREG32(MC_IO_PAD_CNTL_D0) & MEM_FALL_OUT_CMD))
+			udelay(10);
 
 		if (running)
 			WREG32(MC_SHARED_BLACKOUT_CNTL, blackout);
@@ -811,7 +806,6 @@ static void cayman_gpu_init(struct radeon_device *rdev)
 	rdev->config.cayman.tile_config |=
 		((gb_addr_config & ROW_SIZE_MASK) >> ROW_SIZE_SHIFT) << 12;
 
-	rdev->config.cayman.backend_map = gb_backend_map;
 	WREG32(GB_BACKEND_MAP, gb_backend_map);
 	WREG32(GB_ADDR_CONFIG, gb_addr_config);
 	WREG32(DMIF_ADDR_CONFIG, gb_addr_config);
@@ -934,9 +928,9 @@ void cayman_pcie_gart_tlb_flush(struct radeon_device *rdev)
 
 int cayman_pcie_gart_enable(struct radeon_device *rdev)
 {
-	int i, r;
+	int r;
 
-	if (rdev->gart.robj == NULL) {
+	if (rdev->gart.table.vram.robj == NULL) {
 		dev_err(rdev->dev, "No VRAM object for PCIE GART.\n");
 		return -EINVAL;
 	}
@@ -945,12 +939,9 @@ int cayman_pcie_gart_enable(struct radeon_device *rdev)
 		return r;
 	radeon_gart_restore(rdev);
 	/* Setup TLB control */
-	WREG32(MC_VM_MX_L1_TLB_CNTL,
-	       (0xA << 7) |
-	       ENABLE_L1_TLB |
+	WREG32(MC_VM_MX_L1_TLB_CNTL, ENABLE_L1_TLB |
 	       ENABLE_L1_FRAGMENT_PROCESSING |
 	       SYSTEM_ACCESS_MODE_NOT_IN_SYS |
-	       ENABLE_ADVANCED_DRIVER_MODEL |
 	       SYSTEM_APERTURE_UNMAPPED_ACCESS_PASS_THRU);
 	/* Setup L2 cache */
 	WREG32(VM_L2_CNTL, ENABLE_L2_CACHE |
@@ -970,37 +961,19 @@ int cayman_pcie_gart_enable(struct radeon_device *rdev)
 	WREG32(VM_CONTEXT0_CNTL2, 0);
 	WREG32(VM_CONTEXT0_CNTL, ENABLE_CONTEXT | PAGE_TABLE_DEPTH(0) |
 				RANGE_PROTECTION_FAULT_ENABLE_DEFAULT);
-
-	WREG32(0x15D4, 0);
-	WREG32(0x15D8, 0);
-	WREG32(0x15DC, 0);
-
-	/* empty context1-7 */
-	for (i = 1; i < 8; i++) {
-		WREG32(VM_CONTEXT0_PAGE_TABLE_START_ADDR + (i << 2), 0);
-		WREG32(VM_CONTEXT0_PAGE_TABLE_END_ADDR + (i << 2), 0);
-		WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (i << 2),
-			rdev->gart.table_addr >> 12);
-	}
-
-	/* enable context1-7 */
-	WREG32(VM_CONTEXT1_PROTECTION_FAULT_DEFAULT_ADDR,
-	       (u32)(rdev->dummy_page.addr >> 12));
+	/* disable context1-7 */
 	WREG32(VM_CONTEXT1_CNTL2, 0);
 	WREG32(VM_CONTEXT1_CNTL, 0);
-	WREG32(VM_CONTEXT1_CNTL, ENABLE_CONTEXT | PAGE_TABLE_DEPTH(0) |
-				RANGE_PROTECTION_FAULT_ENABLE_DEFAULT);
 
 	cayman_pcie_gart_tlb_flush(rdev);
-	DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n",
-		 (unsigned)(rdev->mc.gtt_size >> 20),
-		 (unsigned long long)rdev->gart.table_addr);
 	rdev->gart.ready = true;
 	return 0;
 }
 
 void cayman_pcie_gart_disable(struct radeon_device *rdev)
 {
+	int r;
+
 	/* Disable all tables */
 	WREG32(VM_CONTEXT0_CNTL, 0);
 	WREG32(VM_CONTEXT1_CNTL, 0);
@@ -1016,7 +989,14 @@ void cayman_pcie_gart_disable(struct radeon_device *rdev)
 	WREG32(VM_L2_CNTL2, 0);
 	WREG32(VM_L2_CNTL3, L2_CACHE_BIGK_ASSOCIATIVITY |
 	       L2_CACHE_BIGK_FRAGMENT_SIZE(6));
-	radeon_gart_table_vram_unpin(rdev);
+	if (rdev->gart.table.vram.robj) {
+		r = radeon_bo_reserve(rdev->gart.table.vram.robj, false);
+		if (likely(r == 0)) {
+			radeon_bo_kunmap(rdev->gart.table.vram.robj);
+			radeon_bo_unpin(rdev->gart.table.vram.robj);
+			radeon_bo_unreserve(rdev->gart.table.vram.robj);
+		}
+	}
 }
 
 void cayman_pcie_gart_fini(struct radeon_device *rdev)
@@ -1026,69 +1006,9 @@ void cayman_pcie_gart_fini(struct radeon_device *rdev)
 	radeon_gart_fini(rdev);
 }
 
-void cayman_cp_int_cntl_setup(struct radeon_device *rdev,
-			      int ring, u32 cp_int_cntl)
-{
-	u32 srbm_gfx_cntl = RREG32(SRBM_GFX_CNTL) & ~3;
-
-	WREG32(SRBM_GFX_CNTL, srbm_gfx_cntl | (ring & 3));
-	WREG32(CP_INT_CNTL, cp_int_cntl);
-}
-
 /*
  * CP.
  */
-void cayman_fence_ring_emit(struct radeon_device *rdev,
-			    struct radeon_fence *fence)
-{
-	struct radeon_ring *ring = &rdev->ring[fence->ring];
-	u64 addr = rdev->fence_drv[fence->ring].gpu_addr;
-
-	/* flush read cache over gart for this vmid */
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
-	radeon_ring_write(ring, (CP_COHER_CNTL2 - PACKET3_SET_CONFIG_REG_START) >> 2);
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3));
-	radeon_ring_write(ring, PACKET3_TC_ACTION_ENA | PACKET3_SH_ACTION_ENA);
-	radeon_ring_write(ring, 0xFFFFFFFF);
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, 10); /* poll interval */
-	/* EVENT_WRITE_EOP - flush caches, send int */
-	radeon_ring_write(ring, PACKET3(PACKET3_EVENT_WRITE_EOP, 4));
-	radeon_ring_write(ring, EVENT_TYPE(CACHE_FLUSH_AND_INV_EVENT_TS) | EVENT_INDEX(5));
-	radeon_ring_write(ring, addr & 0xffffffff);
-	radeon_ring_write(ring, (upper_32_bits(addr) & 0xff) | DATA_SEL(1) | INT_SEL(2));
-	radeon_ring_write(ring, fence->seq);
-	radeon_ring_write(ring, 0);
-}
-
-void cayman_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib)
-{
-	struct radeon_ring *ring = &rdev->ring[ib->fence->ring];
-
-	/* set to DX10/11 mode */
-	radeon_ring_write(ring, PACKET3(PACKET3_MODE_CONTROL, 0));
-	radeon_ring_write(ring, 1);
-	radeon_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2));
-	radeon_ring_write(ring,
-#ifdef __BIG_ENDIAN
-			  (2 << 0) |
-#endif
-			  (ib->gpu_addr & 0xFFFFFFFC));
-	radeon_ring_write(ring, upper_32_bits(ib->gpu_addr) & 0xFF);
-	radeon_ring_write(ring, ib->length_dw | (ib->vm_id << 24));
-
-	/* flush read cache over gart for this vmid */
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
-	radeon_ring_write(ring, (CP_COHER_CNTL2 - PACKET3_SET_CONFIG_REG_START) >> 2);
-	radeon_ring_write(ring, ib->vm_id);
-	radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3));
-	radeon_ring_write(ring, PACKET3_TC_ACTION_ENA | PACKET3_SH_ACTION_ENA);
-	radeon_ring_write(ring, 0xFFFFFFFF);
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, 10); /* poll interval */
-}
-
 static void cayman_cp_enable(struct radeon_device *rdev, bool enable)
 {
 	if (enable)
@@ -1129,64 +1049,63 @@ static int cayman_cp_load_microcode(struct radeon_device *rdev)
 
 static int cayman_cp_start(struct radeon_device *rdev)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	int r, i;
 
-	r = radeon_ring_lock(rdev, ring, 7);
+	r = radeon_ring_lock(rdev, 7);
 	if (r) {
 		DRM_ERROR("radeon: cp failed to lock ring (%d).\n", r);
 		return r;
 	}
-	radeon_ring_write(ring, PACKET3(PACKET3_ME_INITIALIZE, 5));
-	radeon_ring_write(ring, 0x1);
-	radeon_ring_write(ring, 0x0);
-	radeon_ring_write(ring, rdev->config.cayman.max_hw_contexts - 1);
-	radeon_ring_write(ring, PACKET3_ME_INITIALIZE_DEVICE_ID(1));
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, 0);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_write(rdev, PACKET3(PACKET3_ME_INITIALIZE, 5));
+	radeon_ring_write(rdev, 0x1);
+	radeon_ring_write(rdev, 0x0);
+	radeon_ring_write(rdev, rdev->config.cayman.max_hw_contexts - 1);
+	radeon_ring_write(rdev, PACKET3_ME_INITIALIZE_DEVICE_ID(1));
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_unlock_commit(rdev);
 
 	cayman_cp_enable(rdev, true);
 
-	r = radeon_ring_lock(rdev, ring, cayman_default_size + 19);
+	r = radeon_ring_lock(rdev, cayman_default_size + 19);
 	if (r) {
 		DRM_ERROR("radeon: cp failed to lock ring (%d).\n", r);
 		return r;
 	}
 
 	/* setup clear context state */
-	radeon_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0));
-	radeon_ring_write(ring, PACKET3_PREAMBLE_BEGIN_CLEAR_STATE);
+	radeon_ring_write(rdev, PACKET3(PACKET3_PREAMBLE_CNTL, 0));
+	radeon_ring_write(rdev, PACKET3_PREAMBLE_BEGIN_CLEAR_STATE);
 
 	for (i = 0; i < cayman_default_size; i++)
-		radeon_ring_write(ring, cayman_default_state[i]);
+		radeon_ring_write(rdev, cayman_default_state[i]);
 
-	radeon_ring_write(ring, PACKET3(PACKET3_PREAMBLE_CNTL, 0));
-	radeon_ring_write(ring, PACKET3_PREAMBLE_END_CLEAR_STATE);
+	radeon_ring_write(rdev, PACKET3(PACKET3_PREAMBLE_CNTL, 0));
+	radeon_ring_write(rdev, PACKET3_PREAMBLE_END_CLEAR_STATE);
 
 	/* set clear context state */
-	radeon_ring_write(ring, PACKET3(PACKET3_CLEAR_STATE, 0));
-	radeon_ring_write(ring, 0);
+	radeon_ring_write(rdev, PACKET3(PACKET3_CLEAR_STATE, 0));
+	radeon_ring_write(rdev, 0);
 
 	/* SQ_VTX_BASE_VTX_LOC */
-	radeon_ring_write(ring, 0xc0026f00);
-	radeon_ring_write(ring, 0x00000000);
-	radeon_ring_write(ring, 0x00000000);
-	radeon_ring_write(ring, 0x00000000);
+	radeon_ring_write(rdev, 0xc0026f00);
+	radeon_ring_write(rdev, 0x00000000);
+	radeon_ring_write(rdev, 0x00000000);
+	radeon_ring_write(rdev, 0x00000000);
 
 	/* Clear consts */
-	radeon_ring_write(ring, 0xc0036f00);
-	radeon_ring_write(ring, 0x00000bc4);
-	radeon_ring_write(ring, 0xffffffff);
-	radeon_ring_write(ring, 0xffffffff);
-	radeon_ring_write(ring, 0xffffffff);
+	radeon_ring_write(rdev, 0xc0036f00);
+	radeon_ring_write(rdev, 0x00000bc4);
+	radeon_ring_write(rdev, 0xffffffff);
+	radeon_ring_write(rdev, 0xffffffff);
+	radeon_ring_write(rdev, 0xffffffff);
 
-	radeon_ring_write(ring, 0xc0026900);
-	radeon_ring_write(ring, 0x00000316);
-	radeon_ring_write(ring, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */
-	radeon_ring_write(ring, 0x00000010); /*  */
+	radeon_ring_write(rdev, 0xc0026900);
+	radeon_ring_write(rdev, 0x00000316);
+	radeon_ring_write(rdev, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */
+	radeon_ring_write(rdev, 0x00000010); /*  */
 
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev);
 
 	/* XXX init other rings */
 
@@ -1196,12 +1115,11 @@ static int cayman_cp_start(struct radeon_device *rdev)
 static void cayman_cp_fini(struct radeon_device *rdev)
 {
 	cayman_cp_enable(rdev, false);
-	radeon_ring_fini(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
+	radeon_ring_fini(rdev);
 }
 
 int cayman_cp_resume(struct radeon_device *rdev)
 {
-	struct radeon_ring *ring;
 	u32 tmp;
 	u32 rb_bufsz;
 	int r;
@@ -1218,8 +1136,7 @@ int cayman_cp_resume(struct radeon_device *rdev)
 	WREG32(GRBM_SOFT_RESET, 0);
 	RREG32(GRBM_SOFT_RESET);
 
-	WREG32(CP_SEM_WAIT_TIMER, 0x0);
-	WREG32(CP_SEM_INCOMPLETE_TIMER_CNTL, 0x0);
+	WREG32(CP_SEM_WAIT_TIMER, 0x4);
 
 	/* Set the write pointer delay */
 	WREG32(CP_RB_WPTR_DELAY, 0);
@@ -1228,8 +1145,7 @@ int cayman_cp_resume(struct radeon_device *rdev)
 
 	/* ring 0 - compute and gfx */
 	/* Set ring buffer size */
-	ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
-	rb_bufsz = drm_order(ring->ring_size / 8);
+	rb_bufsz = drm_order(rdev->cp.ring_size / 8);
 	tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz;
 #ifdef __BIG_ENDIAN
 	tmp |= BUF_SWAP_32BIT;
@@ -1238,8 +1154,8 @@ int cayman_cp_resume(struct radeon_device *rdev)
 
 	/* Initialize the ring buffer's read and write pointers */
 	WREG32(CP_RB0_CNTL, tmp | RB_RPTR_WR_ENA);
-	ring->wptr = 0;
-	WREG32(CP_RB0_WPTR, ring->wptr);
+	rdev->cp.wptr = 0;
+	WREG32(CP_RB0_WPTR, rdev->cp.wptr);
 
 	/* set the wb address wether it's enabled or not */
 	WREG32(CP_RB0_RPTR_ADDR, (rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) & 0xFFFFFFFC);
@@ -1256,14 +1172,13 @@ int cayman_cp_resume(struct radeon_device *rdev)
 	mdelay(1);
 	WREG32(CP_RB0_CNTL, tmp);
 
-	WREG32(CP_RB0_BASE, ring->gpu_addr >> 8);
+	WREG32(CP_RB0_BASE, rdev->cp.gpu_addr >> 8);
 
-	ring->rptr = RREG32(CP_RB0_RPTR);
+	rdev->cp.rptr = RREG32(CP_RB0_RPTR);
 
 	/* ring1  - compute only */
 	/* Set ring buffer size */
-	ring = &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX];
-	rb_bufsz = drm_order(ring->ring_size / 8);
+	rb_bufsz = drm_order(rdev->cp1.ring_size / 8);
 	tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz;
 #ifdef __BIG_ENDIAN
 	tmp |= BUF_SWAP_32BIT;
@@ -1272,8 +1187,8 @@ int cayman_cp_resume(struct radeon_device *rdev)
 
 	/* Initialize the ring buffer's read and write pointers */
 	WREG32(CP_RB1_CNTL, tmp | RB_RPTR_WR_ENA);
-	ring->wptr = 0;
-	WREG32(CP_RB1_WPTR, ring->wptr);
+	rdev->cp1.wptr = 0;
+	WREG32(CP_RB1_WPTR, rdev->cp1.wptr);
 
 	/* set the wb address wether it's enabled or not */
 	WREG32(CP_RB1_RPTR_ADDR, (rdev->wb.gpu_addr + RADEON_WB_CP1_RPTR_OFFSET) & 0xFFFFFFFC);
@@ -1282,14 +1197,13 @@ int cayman_cp_resume(struct radeon_device *rdev)
 	mdelay(1);
 	WREG32(CP_RB1_CNTL, tmp);
 
-	WREG32(CP_RB1_BASE, ring->gpu_addr >> 8);
+	WREG32(CP_RB1_BASE, rdev->cp1.gpu_addr >> 8);
 
-	ring->rptr = RREG32(CP_RB1_RPTR);
+	rdev->cp1.rptr = RREG32(CP_RB1_RPTR);
 
 	/* ring2 - compute only */
 	/* Set ring buffer size */
-	ring = &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX];
-	rb_bufsz = drm_order(ring->ring_size / 8);
+	rb_bufsz = drm_order(rdev->cp2.ring_size / 8);
 	tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz;
 #ifdef __BIG_ENDIAN
 	tmp |= BUF_SWAP_32BIT;
@@ -1298,8 +1212,8 @@ int cayman_cp_resume(struct radeon_device *rdev)
 
 	/* Initialize the ring buffer's read and write pointers */
 	WREG32(CP_RB2_CNTL, tmp | RB_RPTR_WR_ENA);
-	ring->wptr = 0;
-	WREG32(CP_RB2_WPTR, ring->wptr);
+	rdev->cp2.wptr = 0;
+	WREG32(CP_RB2_WPTR, rdev->cp2.wptr);
 
 	/* set the wb address wether it's enabled or not */
 	WREG32(CP_RB2_RPTR_ADDR, (rdev->wb.gpu_addr + RADEON_WB_CP2_RPTR_OFFSET) & 0xFFFFFFFC);
@@ -1308,28 +1222,28 @@ int cayman_cp_resume(struct radeon_device *rdev)
 	mdelay(1);
 	WREG32(CP_RB2_CNTL, tmp);
 
-	WREG32(CP_RB2_BASE, ring->gpu_addr >> 8);
+	WREG32(CP_RB2_BASE, rdev->cp2.gpu_addr >> 8);
 
-	ring->rptr = RREG32(CP_RB2_RPTR);
+	rdev->cp2.rptr = RREG32(CP_RB2_RPTR);
 
 	/* start the rings */
 	cayman_cp_start(rdev);
-	rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = true;
-	rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX].ready = false;
-	rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX].ready = false;
+	rdev->cp.ready = true;
+	rdev->cp1.ready = true;
+	rdev->cp2.ready = true;
 	/* this only test cp0 */
-	r = radeon_ring_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
+	r = radeon_ring_test(rdev);
 	if (r) {
-		rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false;
-		rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX].ready = false;
-		rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX].ready = false;
+		rdev->cp.ready = false;
+		rdev->cp1.ready = false;
+		rdev->cp2.ready = false;
 		return r;
 	}
 
 	return 0;
 }
 
-bool cayman_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
+bool cayman_gpu_is_lockup(struct radeon_device *rdev)
 {
 	u32 srbm_status;
 	u32 grbm_status;
@@ -1342,20 +1256,20 @@ bool cayman_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
 	grbm_status_se0 = RREG32(GRBM_STATUS_SE0);
 	grbm_status_se1 = RREG32(GRBM_STATUS_SE1);
 	if (!(grbm_status & GUI_ACTIVE)) {
-		r100_gpu_lockup_update(lockup, ring);
+		r100_gpu_lockup_update(lockup, &rdev->cp);
 		return false;
 	}
 	/* force CP activities */
-	r = radeon_ring_lock(rdev, ring, 2);
+	r = radeon_ring_lock(rdev, 2);
 	if (!r) {
 		/* PACKET2 NOP */
-		radeon_ring_write(ring, 0x80000000);
-		radeon_ring_write(ring, 0x80000000);
-		radeon_ring_unlock_commit(rdev, ring);
+		radeon_ring_write(rdev, 0x80000000);
+		radeon_ring_write(rdev, 0x80000000);
+		radeon_ring_unlock_commit(rdev);
 	}
 	/* XXX deal with CP0,1,2 */
-	ring->rptr = RREG32(ring->rptr_reg);
-	return r100_gpu_cp_is_lockup(rdev, lockup, ring);
+	rdev->cp.rptr = RREG32(CP_RB0_RPTR);
+	return r100_gpu_cp_is_lockup(rdev, lockup, &rdev->cp);
 }
 
 static int cayman_gpu_soft_reset(struct radeon_device *rdev)
@@ -1375,15 +1289,6 @@ static int cayman_gpu_soft_reset(struct radeon_device *rdev)
 		RREG32(GRBM_STATUS_SE1));
 	dev_info(rdev->dev, "  SRBM_STATUS=0x%08X\n",
 		RREG32(SRBM_STATUS));
-	dev_info(rdev->dev, "  VM_CONTEXT0_PROTECTION_FAULT_ADDR   0x%08X\n",
-		 RREG32(0x14F8));
-	dev_info(rdev->dev, "  VM_CONTEXT0_PROTECTION_FAULT_STATUS 0x%08X\n",
-		 RREG32(0x14D8));
-	dev_info(rdev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x%08X\n",
-		 RREG32(0x14FC));
-	dev_info(rdev->dev, "  VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n",
-		 RREG32(0x14DC));
-
 	evergreen_mc_stop(rdev, &save);
 	if (evergreen_mc_wait_for_idle(rdev)) {
 		dev_warn(rdev->dev, "Wait for MC idle timedout !\n");
@@ -1414,7 +1319,6 @@ static int cayman_gpu_soft_reset(struct radeon_device *rdev)
 	(void)RREG32(GRBM_SOFT_RESET);
 	/* Wait a little for things to settle down */
 	udelay(50);
-
 	dev_info(rdev->dev, "  GRBM_STATUS=0x%08X\n",
 		RREG32(GRBM_STATUS));
 	dev_info(rdev->dev, "  GRBM_STATUS_SE0=0x%08X\n",
@@ -1434,12 +1338,8 @@ int cayman_asic_reset(struct radeon_device *rdev)
 
 static int cayman_startup(struct radeon_device *rdev)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	int r;
 
-	/* enable pcie gen2 link */
-	evergreen_pcie_gen2_enable(rdev);
-
 	if (!rdev->me_fw || !rdev->pfp_fw || !rdev->rlc_fw || !rdev->mc_fw) {
 		r = ni_init_microcode(rdev);
 		if (r) {
@@ -1453,10 +1353,6 @@ static int cayman_startup(struct radeon_device *rdev)
 		return r;
 	}
 
-	r = r600_vram_scratch_init(rdev);
-	if (r)
-		return r;
-
 	evergreen_mc_program(rdev);
 	r = cayman_pcie_gart_enable(rdev);
 	if (r)
@@ -1465,8 +1361,8 @@ static int cayman_startup(struct radeon_device *rdev)
 
 	r = evergreen_blit_init(rdev);
 	if (r) {
-		r600_blit_fini(rdev);
-		rdev->asic->copy.copy = NULL;
+		evergreen_blit_fini(rdev);
+		rdev->asic->copy = NULL;
 		dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r);
 	}
 
@@ -1475,24 +1371,6 @@ static int cayman_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX);
-	if (r) {
-		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
-		return r;
-	}
-
-	r = radeon_fence_driver_start_ring(rdev, CAYMAN_RING_TYPE_CP1_INDEX);
-	if (r) {
-		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
-		return r;
-	}
-
-	r = radeon_fence_driver_start_ring(rdev, CAYMAN_RING_TYPE_CP2_INDEX);
-	if (r) {
-		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
-		return r;
-	}
-
 	/* Enable IRQ */
 	r = r600_irq_init(rdev);
 	if (r) {
@@ -1502,9 +1380,7 @@ static int cayman_startup(struct radeon_device *rdev)
 	}
 	evergreen_irq_set(rdev);
 
-	r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP_RPTR_OFFSET,
-			     CP_RB0_RPTR, CP_RB0_WPTR,
-			     0, 0xfffff, RADEON_CP_PACKET2);
+	r = radeon_ring_init(rdev, rdev->cp.ring_size);
 	if (r)
 		return r;
 	r = cayman_cp_load_microcode(rdev);
@@ -1514,21 +1390,6 @@ static int cayman_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = radeon_ib_pool_start(rdev);
-	if (r)
-		return r;
-
-	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
-	if (r) {
-		DRM_ERROR("radeon: failed testing IB (%d).\n", r);
-		rdev->accel_working = false;
-		return r;
-	}
-
-	r = radeon_vm_manager_start(rdev);
-	if (r)
-		return r;
-
 	return 0;
 }
 
@@ -1543,27 +1404,40 @@ int cayman_resume(struct radeon_device *rdev)
 	/* post card */
 	atom_asic_init(rdev->mode_info.atom_context);
 
-	rdev->accel_working = true;
 	r = cayman_startup(rdev);
 	if (r) {
 		DRM_ERROR("cayman startup failed on resume\n");
-		rdev->accel_working = false;
 		return r;
 	}
+
+	r = r600_ib_test(rdev);
+	if (r) {
+		DRM_ERROR("radeon: failled testing IB (%d).\n", r);
+		return r;
+	}
+
 	return r;
+
 }
 
 int cayman_suspend(struct radeon_device *rdev)
 {
+	int r;
+
 	/* FIXME: we should wait for ring to be empty */
-	radeon_ib_pool_suspend(rdev);
-	radeon_vm_manager_suspend(rdev);
-	r600_blit_suspend(rdev);
 	cayman_cp_enable(rdev, false);
-	rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false;
+	rdev->cp.ready = false;
 	evergreen_irq_suspend(rdev);
 	radeon_wb_disable(rdev);
 	cayman_pcie_gart_disable(rdev);
+
+	/* unpin shaders bo */
+	r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false);
+	if (likely(r == 0)) {
+		radeon_bo_unpin(rdev->r600_blit.shader_obj);
+		radeon_bo_unreserve(rdev->r600_blit.shader_obj);
+	}
+
 	return 0;
 }
 
@@ -1575,7 +1449,6 @@ int cayman_suspend(struct radeon_device *rdev)
  */
 int cayman_init(struct radeon_device *rdev)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	int r;
 
 	/* This don't do much */
@@ -1628,8 +1501,8 @@ int cayman_init(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	ring->ring_obj = NULL;
-	r600_ring_init(rdev, ring, 1024 * 1024);
+	rdev->cp.ring_obj = NULL;
+	r600_ring_init(rdev, 1024 * 1024);
 
 	rdev->ih.ring_obj = NULL;
 	r600_ih_ring_init(rdev, 64 * 1024);
@@ -1638,29 +1511,29 @@ int cayman_init(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = radeon_ib_pool_init(rdev);
 	rdev->accel_working = true;
-	if (r) {
-		dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
-		rdev->accel_working = false;
-	}
-	r = radeon_vm_manager_init(rdev);
-	if (r) {
-		dev_err(rdev->dev, "vm manager initialization failed (%d).\n", r);
-	}
-
 	r = cayman_startup(rdev);
 	if (r) {
 		dev_err(rdev->dev, "disabling GPU acceleration\n");
 		cayman_cp_fini(rdev);
 		r600_irq_fini(rdev);
 		radeon_wb_fini(rdev);
-		r100_ib_fini(rdev);
-		radeon_vm_manager_fini(rdev);
 		radeon_irq_kms_fini(rdev);
 		cayman_pcie_gart_fini(rdev);
 		rdev->accel_working = false;
 	}
+	if (rdev->accel_working) {
+		r = radeon_ib_pool_init(rdev);
+		if (r) {
+			DRM_ERROR("radeon: failed initializing IB pool (%d).\n", r);
+			rdev->accel_working = false;
+		}
+		r = r600_ib_test(rdev);
+		if (r) {
+			DRM_ERROR("radeon: failed testing IB (%d).\n", r);
+			rdev->accel_working = false;
+		}
+	}
 
 	/* Don't start up if the MC ucode is missing.
 	 * The default clocks and voltages before the MC ucode
@@ -1676,17 +1549,14 @@ int cayman_init(struct radeon_device *rdev)
 
 void cayman_fini(struct radeon_device *rdev)
 {
-	r600_blit_fini(rdev);
+	evergreen_blit_fini(rdev);
 	cayman_cp_fini(rdev);
 	r600_irq_fini(rdev);
 	radeon_wb_fini(rdev);
-	radeon_vm_manager_fini(rdev);
-	r100_ib_fini(rdev);
+	radeon_ib_pool_fini(rdev);
 	radeon_irq_kms_fini(rdev);
 	cayman_pcie_gart_fini(rdev);
-	r600_vram_scratch_fini(rdev);
 	radeon_gem_fini(rdev);
-	radeon_semaphore_driver_fini(rdev);
 	radeon_fence_driver_fini(rdev);
 	radeon_bo_fini(rdev);
 	radeon_atombios_fini(rdev);
@@ -1694,84 +1564,3 @@ void cayman_fini(struct radeon_device *rdev)
 	rdev->bios = NULL;
 }
 
-/*
- * vm
- */
-int cayman_vm_init(struct radeon_device *rdev)
-{
-	/* number of VMs */
-	rdev->vm_manager.nvm = 8;
-	/* base offset of vram pages */
-	rdev->vm_manager.vram_base_offset = 0;
-	return 0;
-}
-
-void cayman_vm_fini(struct radeon_device *rdev)
-{
-}
-
-int cayman_vm_bind(struct radeon_device *rdev, struct radeon_vm *vm, int id)
-{
-	WREG32(VM_CONTEXT0_PAGE_TABLE_START_ADDR + (id << 2), 0);
-	WREG32(VM_CONTEXT0_PAGE_TABLE_END_ADDR + (id << 2), vm->last_pfn);
-	WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (id << 2), vm->pt_gpu_addr >> 12);
-	/* flush hdp cache */
-	WREG32(HDP_MEM_COHERENCY_FLUSH_CNTL, 0x1);
-	/* bits 0-7 are the VM contexts0-7 */
-	WREG32(VM_INVALIDATE_REQUEST, 1 << id);
-	return 0;
-}
-
-void cayman_vm_unbind(struct radeon_device *rdev, struct radeon_vm *vm)
-{
-	WREG32(VM_CONTEXT0_PAGE_TABLE_START_ADDR + (vm->id << 2), 0);
-	WREG32(VM_CONTEXT0_PAGE_TABLE_END_ADDR + (vm->id << 2), 0);
-	WREG32(VM_CONTEXT0_PAGE_TABLE_BASE_ADDR + (vm->id << 2), 0);
-	/* flush hdp cache */
-	WREG32(HDP_MEM_COHERENCY_FLUSH_CNTL, 0x1);
-	/* bits 0-7 are the VM contexts0-7 */
-	WREG32(VM_INVALIDATE_REQUEST, 1 << vm->id);
-}
-
-void cayman_vm_tlb_flush(struct radeon_device *rdev, struct radeon_vm *vm)
-{
-	if (vm->id == -1)
-		return;
-
-	/* flush hdp cache */
-	WREG32(HDP_MEM_COHERENCY_FLUSH_CNTL, 0x1);
-	/* bits 0-7 are the VM contexts0-7 */
-	WREG32(VM_INVALIDATE_REQUEST, 1 << vm->id);
-}
-
-#define R600_PTE_VALID     (1 << 0)
-#define R600_PTE_SYSTEM    (1 << 1)
-#define R600_PTE_SNOOPED   (1 << 2)
-#define R600_PTE_READABLE  (1 << 5)
-#define R600_PTE_WRITEABLE (1 << 6)
-
-uint32_t cayman_vm_page_flags(struct radeon_device *rdev,
-			      struct radeon_vm *vm,
-			      uint32_t flags)
-{
-	uint32_t r600_flags = 0;
-
-	r600_flags |= (flags & RADEON_VM_PAGE_VALID) ? R600_PTE_VALID : 0;
-	r600_flags |= (flags & RADEON_VM_PAGE_READABLE) ? R600_PTE_READABLE : 0;
-	r600_flags |= (flags & RADEON_VM_PAGE_WRITEABLE) ? R600_PTE_WRITEABLE : 0;
-	if (flags & RADEON_VM_PAGE_SYSTEM) {
-		r600_flags |= R600_PTE_SYSTEM;
-		r600_flags |= (flags & RADEON_VM_PAGE_SNOOPED) ? R600_PTE_SNOOPED : 0;
-	}
-	return r600_flags;
-}
-
-void cayman_vm_set_page(struct radeon_device *rdev, struct radeon_vm *vm,
-			unsigned pfn, uint64_t addr, uint32_t flags)
-{
-	void __iomem *ptr = (void *)vm->pt;
-
-	addr = addr & 0xFFFFFFFFFFFFF000ULL;
-	addr |= flags;
-	writeq(addr, ptr + (pfn * 8));
-}
diff --git a/drivers/gpu/drm/radeon/nid.h b/drivers/gpu/drm/radeon/nid.h
index 9a7f3b6..4672869 100644
--- a/drivers/gpu/drm/radeon/nid.h
+++ b/drivers/gpu/drm/radeon/nid.h
@@ -42,9 +42,6 @@
 #define CAYMAN_MAX_TCC_MASK          0xFF
 
 #define DMIF_ADDR_CONFIG  				0xBD4
-#define	SRBM_GFX_CNTL				        0x0E44
-#define		RINGID(x)					(((x) & 0x3) << 0)
-#define		VMID(x)						(((x) & 0x7) << 0)
 #define	SRBM_STATUS				        0x0E50
 
 #define VM_CONTEXT0_REQUEST_RESPONSE			0x1470
@@ -222,8 +219,6 @@
 #define	SCRATCH_UMSK					0x8540
 #define	SCRATCH_ADDR					0x8544
 #define	CP_SEM_WAIT_TIMER				0x85BC
-#define	CP_SEM_INCOMPLETE_TIMER_CNTL			0x85C8
-#define	CP_COHER_CNTL2					0x85E8
 #define CP_ME_CNTL					0x86D8
 #define		CP_ME_HALT					(1 << 28)
 #define		CP_PFP_HALT					(1 << 26)
@@ -399,12 +394,6 @@
 #define	CP_RB0_RPTR_ADDR				0xC10C
 #define	CP_RB0_RPTR_ADDR_HI				0xC110
 #define	CP_RB0_WPTR					0xC114
-
-#define CP_INT_CNTL                                     0xC124
-#       define CNTX_BUSY_INT_ENABLE                     (1 << 19)
-#       define CNTX_EMPTY_INT_ENABLE                    (1 << 20)
-#       define TIME_STAMP_INT_ENABLE                    (1 << 26)
-
 #define	CP_RB1_BASE					0xC180
 #define	CP_RB1_CNTL					0xC184
 #define	CP_RB1_RPTR_ADDR				0xC188
@@ -422,10 +411,6 @@
 #define	CP_ME_RAM_DATA					0xC160
 #define	CP_DEBUG					0xC1FC
 
-#define VGT_EVENT_INITIATOR                             0x28a90
-#       define CACHE_FLUSH_AND_INV_EVENT_TS                     (0x14 << 0)
-#       define CACHE_FLUSH_AND_INV_EVENT                        (0x16 << 0)
-
 /*
  * PM4
  */
@@ -460,7 +445,6 @@
 #define	PACKET3_DISPATCH_DIRECT				0x15
 #define	PACKET3_DISPATCH_INDIRECT			0x16
 #define	PACKET3_INDIRECT_BUFFER_END			0x17
-#define	PACKET3_MODE_CONTROL				0x18
 #define	PACKET3_SET_PREDICATION				0x20
 #define	PACKET3_REG_RMW					0x21
 #define	PACKET3_COND_EXEC				0x22
@@ -510,27 +494,7 @@
 #define		PACKET3_ME_INITIALIZE_DEVICE_ID(x) ((x) << 16)
 #define	PACKET3_COND_WRITE				0x45
 #define	PACKET3_EVENT_WRITE				0x46
-#define		EVENT_TYPE(x)                           ((x) << 0)
-#define		EVENT_INDEX(x)                          ((x) << 8)
-                /* 0 - any non-TS event
-		 * 1 - ZPASS_DONE
-		 * 2 - SAMPLE_PIPELINESTAT
-		 * 3 - SAMPLE_STREAMOUTSTAT*
-		 * 4 - *S_PARTIAL_FLUSH
-		 * 5 - TS events
-		 */
 #define	PACKET3_EVENT_WRITE_EOP				0x47
-#define		DATA_SEL(x)                             ((x) << 29)
-                /* 0 - discard
-		 * 1 - send low 32bit data
-		 * 2 - send 64bit data
-		 * 3 - send 64bit counter value
-		 */
-#define		INT_SEL(x)                              ((x) << 24)
-                /* 0 - none
-		 * 1 - interrupt only (DATA_SEL = 0)
-		 * 2 - interrupt when data write is confirmed
-		 */
 #define	PACKET3_EVENT_WRITE_EOS				0x48
 #define	PACKET3_PREAMBLE_CNTL				0x4A
 #              define PACKET3_PREAMBLE_BEGIN_CLEAR_STATE     (2 << 28)
diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c
index ae356cf..d94f440 100644
--- a/drivers/gpu/drm/radeon/r100.c
+++ b/drivers/gpu/drm/radeon/r100.c
@@ -41,7 +41,6 @@
 
 #include <linux/firmware.h>
 #include <linux/platform_device.h>
-#include <linux/module.h>
 
 #include "r100_reg_safe.h"
 #include "rn50_reg_safe.h"
@@ -65,150 +64,10 @@ MODULE_FIRMWARE(FIRMWARE_R520);
 
 #include "r100_track.h"
 
-void r100_wait_for_vblank(struct radeon_device *rdev, int crtc)
-{
-	struct radeon_crtc *radeon_crtc = rdev->mode_info.crtcs[crtc];
-	int i;
-
-	if (radeon_crtc->crtc_id == 0) {
-		if (RREG32(RADEON_CRTC_GEN_CNTL) & RADEON_CRTC_EN) {
-			for (i = 0; i < rdev->usec_timeout; i++) {
-				if (!(RREG32(RADEON_CRTC_STATUS) & RADEON_CRTC_VBLANK_CUR))
-					break;
-				udelay(1);
-			}
-			for (i = 0; i < rdev->usec_timeout; i++) {
-				if (RREG32(RADEON_CRTC_STATUS) & RADEON_CRTC_VBLANK_CUR)
-					break;
-				udelay(1);
-			}
-		}
-	} else {
-		if (RREG32(RADEON_CRTC2_GEN_CNTL) & RADEON_CRTC2_EN) {
-			for (i = 0; i < rdev->usec_timeout; i++) {
-				if (!(RREG32(RADEON_CRTC2_STATUS) & RADEON_CRTC2_VBLANK_CUR))
-					break;
-				udelay(1);
-			}
-			for (i = 0; i < rdev->usec_timeout; i++) {
-				if (RREG32(RADEON_CRTC2_STATUS) & RADEON_CRTC2_VBLANK_CUR)
-					break;
-				udelay(1);
-			}
-		}
-	}
-}
-
 /* This files gather functions specifics to:
  * r100,rv100,rs100,rv200,rs200,r200,rv250,rs300,rv280
  */
 
-int r100_reloc_pitch_offset(struct radeon_cs_parser *p,
-			    struct radeon_cs_packet *pkt,
-			    unsigned idx,
-			    unsigned reg)
-{
-	int r;
-	u32 tile_flags = 0;
-	u32 tmp;
-	struct radeon_cs_reloc *reloc;
-	u32 value;
-
-	r = r100_cs_packet_next_reloc(p, &reloc);
-	if (r) {
-		DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
-			  idx, reg);
-		r100_cs_dump_packet(p, pkt);
-		return r;
-	}
-
-	value = radeon_get_ib_value(p, idx);
-	tmp = value & 0x003fffff;
-	tmp += (((u32)reloc->lobj.gpu_offset) >> 10);
-
-	if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
-		if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
-			tile_flags |= RADEON_DST_TILE_MACRO;
-		if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) {
-			if (reg == RADEON_SRC_PITCH_OFFSET) {
-				DRM_ERROR("Cannot src blit from microtiled surface\n");
-				r100_cs_dump_packet(p, pkt);
-				return -EINVAL;
-			}
-			tile_flags |= RADEON_DST_TILE_MICRO;
-		}
-
-		tmp |= tile_flags;
-		p->ib->ptr[idx] = (value & 0x3fc00000) | tmp;
-	} else
-		p->ib->ptr[idx] = (value & 0xffc00000) | tmp;
-	return 0;
-}
-
-int r100_packet3_load_vbpntr(struct radeon_cs_parser *p,
-			     struct radeon_cs_packet *pkt,
-			     int idx)
-{
-	unsigned c, i;
-	struct radeon_cs_reloc *reloc;
-	struct r100_cs_track *track;
-	int r = 0;
-	volatile uint32_t *ib;
-	u32 idx_value;
-
-	ib = p->ib->ptr;
-	track = (struct r100_cs_track *)p->track;
-	c = radeon_get_ib_value(p, idx++) & 0x1F;
-	if (c > 16) {
-	    DRM_ERROR("Only 16 vertex buffers are allowed %d\n",
-		      pkt->opcode);
-	    r100_cs_dump_packet(p, pkt);
-	    return -EINVAL;
-	}
-	track->num_arrays = c;
-	for (i = 0; i < (c - 1); i+=2, idx+=3) {
-		r = r100_cs_packet_next_reloc(p, &reloc);
-		if (r) {
-			DRM_ERROR("No reloc for packet3 %d\n",
-				  pkt->opcode);
-			r100_cs_dump_packet(p, pkt);
-			return r;
-		}
-		idx_value = radeon_get_ib_value(p, idx);
-		ib[idx+1] = radeon_get_ib_value(p, idx + 1) + ((u32)reloc->lobj.gpu_offset);
-
-		track->arrays[i + 0].esize = idx_value >> 8;
-		track->arrays[i + 0].robj = reloc->robj;
-		track->arrays[i + 0].esize &= 0x7F;
-		r = r100_cs_packet_next_reloc(p, &reloc);
-		if (r) {
-			DRM_ERROR("No reloc for packet3 %d\n",
-				  pkt->opcode);
-			r100_cs_dump_packet(p, pkt);
-			return r;
-		}
-		ib[idx+2] = radeon_get_ib_value(p, idx + 2) + ((u32)reloc->lobj.gpu_offset);
-		track->arrays[i + 1].robj = reloc->robj;
-		track->arrays[i + 1].esize = idx_value >> 24;
-		track->arrays[i + 1].esize &= 0x7F;
-	}
-	if (c & 1) {
-		r = r100_cs_packet_next_reloc(p, &reloc);
-		if (r) {
-			DRM_ERROR("No reloc for packet3 %d\n",
-					  pkt->opcode);
-			r100_cs_dump_packet(p, pkt);
-			return r;
-		}
-		idx_value = radeon_get_ib_value(p, idx);
-		ib[idx+1] = radeon_get_ib_value(p, idx + 1) + ((u32)reloc->lobj.gpu_offset);
-		track->arrays[i + 0].robj = reloc->robj;
-		track->arrays[i + 0].esize = idx_value >> 8;
-		track->arrays[i + 0].esize &= 0x7F;
-	}
-	return r;
-}
-
 void r100_pre_page_flip(struct radeon_device *rdev, int crtc)
 {
 	/* enable the pflip int */
@@ -450,7 +309,7 @@ void r100_pm_misc(struct radeon_device *rdev)
 	/* set pcie lanes */
 	if ((rdev->flags & RADEON_IS_PCIE) &&
 	    !(rdev->flags & RADEON_IS_IGP) &&
-	    rdev->asic->pm.set_pcie_lanes &&
+	    rdev->asic->set_pcie_lanes &&
 	    (ps->pcie_lanes !=
 	     rdev->pm.power_state[rdev->pm.current_power_state_index].pcie_lanes)) {
 		radeon_set_pcie_lanes(rdev,
@@ -621,7 +480,7 @@ int r100_pci_gart_init(struct radeon_device *rdev)
 {
 	int r;
 
-	if (rdev->gart.ptr) {
+	if (rdev->gart.table.ram.ptr) {
 		WARN(1, "R100 PCI GART already initialized\n");
 		return 0;
 	}
@@ -630,8 +489,8 @@ int r100_pci_gart_init(struct radeon_device *rdev)
 	if (r)
 		return r;
 	rdev->gart.table_size = rdev->gart.num_gpu_pages * 4;
-	rdev->asic->gart.tlb_flush = &r100_pci_gart_tlb_flush;
-	rdev->asic->gart.set_page = &r100_pci_gart_set_page;
+	rdev->asic->gart_tlb_flush = &r100_pci_gart_tlb_flush;
+	rdev->asic->gart_set_page = &r100_pci_gart_set_page;
 	return radeon_gart_table_ram_alloc(rdev);
 }
 
@@ -660,9 +519,6 @@ int r100_pci_gart_enable(struct radeon_device *rdev)
 	tmp = RREG32(RADEON_AIC_CNTL) | RADEON_PCIGART_TRANSLATE_EN;
 	WREG32(RADEON_AIC_CNTL, tmp);
 	r100_pci_gart_tlb_flush(rdev);
-	DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n",
-		 (unsigned)(rdev->mc.gtt_size >> 20),
-		 (unsigned long long)rdev->gart.table_addr);
 	rdev->gart.ready = true;
 	return 0;
 }
@@ -680,12 +536,10 @@ void r100_pci_gart_disable(struct radeon_device *rdev)
 
 int r100_pci_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr)
 {
-	u32 *gtt = rdev->gart.ptr;
-
 	if (i < 0 || i > rdev->gart.num_gpu_pages) {
 		return -EINVAL;
 	}
-	gtt[i] = cpu_to_le32(lower_32_bits(addr));
+	rdev->gart.table.ram.ptr[i] = cpu_to_le32(lower_32_bits(addr));
 	return 0;
 }
 
@@ -705,7 +559,7 @@ int r100_irq_set(struct radeon_device *rdev)
 		WREG32(R_000040_GEN_INT_CNTL, 0);
 		return -EINVAL;
 	}
-	if (rdev->irq.sw_int[RADEON_RING_TYPE_GFX_INDEX]) {
+	if (rdev->irq.sw_int) {
 		tmp |= RADEON_SW_INT_ENABLE;
 	}
 	if (rdev->irq.gui_idle) {
@@ -740,7 +594,7 @@ void r100_irq_disable(struct radeon_device *rdev)
 	WREG32(R_000044_GEN_INT_STATUS, tmp);
 }
 
-static uint32_t r100_irq_ack(struct radeon_device *rdev)
+static inline uint32_t r100_irq_ack(struct radeon_device *rdev)
 {
 	uint32_t irqs = RREG32(RADEON_GEN_INT_STATUS);
 	uint32_t irq_mask = RADEON_SW_INT_TEST |
@@ -777,7 +631,7 @@ int r100_irq_process(struct radeon_device *rdev)
 	while (status) {
 		/* SW interrupt */
 		if (status & RADEON_SW_INT_TEST) {
-			radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX);
+			radeon_fence_process(rdev);
 		}
 		/* gui idle interrupt */
 		if (status & RADEON_GUI_IDLE_STAT) {
@@ -847,36 +701,25 @@ u32 r100_get_vblank_counter(struct radeon_device *rdev, int crtc)
 void r100_fence_ring_emit(struct radeon_device *rdev,
 			  struct radeon_fence *fence)
 {
-	struct radeon_ring *ring = &rdev->ring[fence->ring];
-
 	/* We have to make sure that caches are flushed before
 	 * CPU might read something from VRAM. */
-	radeon_ring_write(ring, PACKET0(RADEON_RB3D_DSTCACHE_CTLSTAT, 0));
-	radeon_ring_write(ring, RADEON_RB3D_DC_FLUSH_ALL);
-	radeon_ring_write(ring, PACKET0(RADEON_RB3D_ZCACHE_CTLSTAT, 0));
-	radeon_ring_write(ring, RADEON_RB3D_ZC_FLUSH_ALL);
+	radeon_ring_write(rdev, PACKET0(RADEON_RB3D_DSTCACHE_CTLSTAT, 0));
+	radeon_ring_write(rdev, RADEON_RB3D_DC_FLUSH_ALL);
+	radeon_ring_write(rdev, PACKET0(RADEON_RB3D_ZCACHE_CTLSTAT, 0));
+	radeon_ring_write(rdev, RADEON_RB3D_ZC_FLUSH_ALL);
 	/* Wait until IDLE & CLEAN */
-	radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0));
-	radeon_ring_write(ring, RADEON_WAIT_2D_IDLECLEAN | RADEON_WAIT_3D_IDLECLEAN);
-	radeon_ring_write(ring, PACKET0(RADEON_HOST_PATH_CNTL, 0));
-	radeon_ring_write(ring, rdev->config.r100.hdp_cntl |
+	radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0));
+	radeon_ring_write(rdev, RADEON_WAIT_2D_IDLECLEAN | RADEON_WAIT_3D_IDLECLEAN);
+	radeon_ring_write(rdev, PACKET0(RADEON_HOST_PATH_CNTL, 0));
+	radeon_ring_write(rdev, rdev->config.r100.hdp_cntl |
 				RADEON_HDP_READ_BUFFER_INVALIDATE);
-	radeon_ring_write(ring, PACKET0(RADEON_HOST_PATH_CNTL, 0));
-	radeon_ring_write(ring, rdev->config.r100.hdp_cntl);
+	radeon_ring_write(rdev, PACKET0(RADEON_HOST_PATH_CNTL, 0));
+	radeon_ring_write(rdev, rdev->config.r100.hdp_cntl);
 	/* Emit fence sequence & fire IRQ */
-	radeon_ring_write(ring, PACKET0(rdev->fence_drv[fence->ring].scratch_reg, 0));
-	radeon_ring_write(ring, fence->seq);
-	radeon_ring_write(ring, PACKET0(RADEON_GEN_INT_STATUS, 0));
-	radeon_ring_write(ring, RADEON_SW_INT_FIRE);
-}
-
-void r100_semaphore_ring_emit(struct radeon_device *rdev,
-			      struct radeon_ring *ring,
-			      struct radeon_semaphore *semaphore,
-			      bool emit_wait)
-{
-	/* Unused on older asics, since we don't have semaphores or multiple rings */
-	BUG();
+	radeon_ring_write(rdev, PACKET0(rdev->fence_drv.scratch_reg, 0));
+	radeon_ring_write(rdev, fence->seq);
+	radeon_ring_write(rdev, PACKET0(RADEON_GEN_INT_STATUS, 0));
+	radeon_ring_write(rdev, RADEON_SW_INT_FIRE);
 }
 
 int r100_copy_blit(struct radeon_device *rdev,
@@ -885,7 +728,6 @@ int r100_copy_blit(struct radeon_device *rdev,
 		   unsigned num_gpu_pages,
 		   struct radeon_fence *fence)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	uint32_t cur_pages;
 	uint32_t stride_bytes = RADEON_GPU_PAGE_SIZE;
 	uint32_t pitch;
@@ -903,7 +745,7 @@ int r100_copy_blit(struct radeon_device *rdev,
 
 	/* Ask for enough room for blit + flush + fence */
 	ndw = 64 + (10 * num_loops);
-	r = radeon_ring_lock(rdev, ring, ndw);
+	r = radeon_ring_lock(rdev, ndw);
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d) asking for %u dw.\n", r, ndw);
 		return -EINVAL;
@@ -917,8 +759,8 @@ int r100_copy_blit(struct radeon_device *rdev,
 
 		/* pages are in Y direction - height
 		   page width in X direction - width */
-		radeon_ring_write(ring, PACKET3(PACKET3_BITBLT_MULTI, 8));
-		radeon_ring_write(ring,
+		radeon_ring_write(rdev, PACKET3(PACKET3_BITBLT_MULTI, 8));
+		radeon_ring_write(rdev,
 				  RADEON_GMC_SRC_PITCH_OFFSET_CNTL |
 				  RADEON_GMC_DST_PITCH_OFFSET_CNTL |
 				  RADEON_GMC_SRC_CLIPPING |
@@ -930,26 +772,26 @@ int r100_copy_blit(struct radeon_device *rdev,
 				  RADEON_DP_SRC_SOURCE_MEMORY |
 				  RADEON_GMC_CLR_CMP_CNTL_DIS |
 				  RADEON_GMC_WR_MSK_DIS);
-		radeon_ring_write(ring, (pitch << 22) | (src_offset >> 10));
-		radeon_ring_write(ring, (pitch << 22) | (dst_offset >> 10));
-		radeon_ring_write(ring, (0x1fff) | (0x1fff << 16));
-		radeon_ring_write(ring, 0);
-		radeon_ring_write(ring, (0x1fff) | (0x1fff << 16));
-		radeon_ring_write(ring, num_gpu_pages);
-		radeon_ring_write(ring, num_gpu_pages);
-		radeon_ring_write(ring, cur_pages | (stride_pixels << 16));
-	}
-	radeon_ring_write(ring, PACKET0(RADEON_DSTCACHE_CTLSTAT, 0));
-	radeon_ring_write(ring, RADEON_RB2D_DC_FLUSH_ALL);
-	radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0));
-	radeon_ring_write(ring,
+		radeon_ring_write(rdev, (pitch << 22) | (src_offset >> 10));
+		radeon_ring_write(rdev, (pitch << 22) | (dst_offset >> 10));
+		radeon_ring_write(rdev, (0x1fff) | (0x1fff << 16));
+		radeon_ring_write(rdev, 0);
+		radeon_ring_write(rdev, (0x1fff) | (0x1fff << 16));
+		radeon_ring_write(rdev, num_gpu_pages);
+		radeon_ring_write(rdev, num_gpu_pages);
+		radeon_ring_write(rdev, cur_pages | (stride_pixels << 16));
+	}
+	radeon_ring_write(rdev, PACKET0(RADEON_DSTCACHE_CTLSTAT, 0));
+	radeon_ring_write(rdev, RADEON_RB2D_DC_FLUSH_ALL);
+	radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0));
+	radeon_ring_write(rdev,
 			  RADEON_WAIT_2D_IDLECLEAN |
 			  RADEON_WAIT_HOST_IDLECLEAN |
 			  RADEON_WAIT_DMA_GUI_IDLE);
 	if (fence) {
 		r = radeon_fence_emit(rdev, fence);
 	}
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev);
 	return r;
 }
 
@@ -968,21 +810,21 @@ static int r100_cp_wait_for_idle(struct radeon_device *rdev)
 	return -1;
 }
 
-void r100_ring_start(struct radeon_device *rdev, struct radeon_ring *ring)
+void r100_ring_start(struct radeon_device *rdev)
 {
 	int r;
 
-	r = radeon_ring_lock(rdev, ring, 2);
+	r = radeon_ring_lock(rdev, 2);
 	if (r) {
 		return;
 	}
-	radeon_ring_write(ring, PACKET0(RADEON_ISYNC_CNTL, 0));
-	radeon_ring_write(ring,
+	radeon_ring_write(rdev, PACKET0(RADEON_ISYNC_CNTL, 0));
+	radeon_ring_write(rdev,
 			  RADEON_ISYNC_ANY2D_IDLE3D |
 			  RADEON_ISYNC_ANY3D_IDLE2D |
 			  RADEON_ISYNC_WAIT_IDLEGUI |
 			  RADEON_ISYNC_CPSCRATCH_IDLEGUI);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev);
 }
 
 
@@ -1083,7 +925,6 @@ static void r100_cp_load_microcode(struct radeon_device *rdev)
 
 int r100_cp_init(struct radeon_device *rdev, unsigned ring_size)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	unsigned rb_bufsz;
 	unsigned rb_blksz;
 	unsigned max_fetch;
@@ -1109,9 +950,7 @@ int r100_cp_init(struct radeon_device *rdev, unsigned ring_size)
 	rb_bufsz = drm_order(ring_size / 8);
 	ring_size = (1 << (rb_bufsz + 1)) * 4;
 	r100_cp_load_microcode(rdev);
-	r = radeon_ring_init(rdev, ring, ring_size, RADEON_WB_CP_RPTR_OFFSET,
-			     RADEON_CP_RB_RPTR, RADEON_CP_RB_WPTR,
-			     0, 0x7fffff, RADEON_CP_PACKET2);
+	r = radeon_ring_init(rdev, ring_size);
 	if (r) {
 		return r;
 	}
@@ -1120,7 +959,7 @@ int r100_cp_init(struct radeon_device *rdev, unsigned ring_size)
 	rb_blksz = 9;
 	/* cp will read 128bytes at a time (4 dwords) */
 	max_fetch = 1;
-	ring->align_mask = 16 - 1;
+	rdev->cp.align_mask = 16 - 1;
 	/* Write to CP_RB_WPTR will be delayed for pre_write_timer clocks */
 	pre_write_timer = 64;
 	/* Force CP_RB_WPTR write if written more than one time before the
@@ -1150,13 +989,13 @@ int r100_cp_init(struct radeon_device *rdev, unsigned ring_size)
 	WREG32(RADEON_CP_RB_CNTL, tmp | RADEON_RB_NO_UPDATE);
 
 	/* Set ring address */
-	DRM_INFO("radeon: ring at 0x%016lX\n", (unsigned long)ring->gpu_addr);
-	WREG32(RADEON_CP_RB_BASE, ring->gpu_addr);
+	DRM_INFO("radeon: ring at 0x%016lX\n", (unsigned long)rdev->cp.gpu_addr);
+	WREG32(RADEON_CP_RB_BASE, rdev->cp.gpu_addr);
 	/* Force read & write ptr to 0 */
 	WREG32(RADEON_CP_RB_CNTL, tmp | RADEON_RB_RPTR_WR_ENA | RADEON_RB_NO_UPDATE);
 	WREG32(RADEON_CP_RB_RPTR_WR, 0);
-	ring->wptr = 0;
-	WREG32(RADEON_CP_RB_WPTR, ring->wptr);
+	rdev->cp.wptr = 0;
+	WREG32(RADEON_CP_RB_WPTR, rdev->cp.wptr);
 
 	/* set the wb address whether it's enabled or not */
 	WREG32(R_00070C_CP_RB_RPTR_ADDR,
@@ -1172,7 +1011,7 @@ int r100_cp_init(struct radeon_device *rdev, unsigned ring_size)
 
 	WREG32(RADEON_CP_RB_CNTL, tmp);
 	udelay(10);
-	ring->rptr = RREG32(RADEON_CP_RB_RPTR);
+	rdev->cp.rptr = RREG32(RADEON_CP_RB_RPTR);
 	/* Set cp mode to bus mastering & enable cp*/
 	WREG32(RADEON_CP_CSQ_MODE,
 	       REG_SET(RADEON_INDIRECT2_START, indirect2_start) |
@@ -1180,13 +1019,13 @@ int r100_cp_init(struct radeon_device *rdev, unsigned ring_size)
 	WREG32(RADEON_CP_RB_WPTR_DELAY, 0);
 	WREG32(RADEON_CP_CSQ_MODE, 0x00004D4D);
 	WREG32(RADEON_CP_CSQ_CNTL, RADEON_CSQ_PRIBM_INDBM);
-	radeon_ring_start(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
-	r = radeon_ring_test(rdev, RADEON_RING_TYPE_GFX_INDEX, ring);
+	radeon_ring_start(rdev);
+	r = radeon_ring_test(rdev);
 	if (r) {
 		DRM_ERROR("radeon: cp isn't working (%d).\n", r);
 		return r;
 	}
-	ring->ready = true;
+	rdev->cp.ready = true;
 	radeon_ttm_set_active_vram_size(rdev, rdev->mc.real_vram_size);
 	return 0;
 }
@@ -1198,7 +1037,7 @@ void r100_cp_fini(struct radeon_device *rdev)
 	}
 	/* Disable ring */
 	r100_cp_disable(rdev);
-	radeon_ring_fini(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
+	radeon_ring_fini(rdev);
 	DRM_INFO("radeon: cp finalized\n");
 }
 
@@ -1206,7 +1045,7 @@ void r100_cp_disable(struct radeon_device *rdev)
 {
 	/* Disable ring */
 	radeon_ttm_set_active_vram_size(rdev, rdev->mc.visible_vram_size);
-	rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false;
+	rdev->cp.ready = false;
 	WREG32(RADEON_CP_CSQ_MODE, 0);
 	WREG32(RADEON_CP_CSQ_CNTL, 0);
 	WREG32(R_000770_SCRATCH_UMSK, 0);
@@ -1216,6 +1055,13 @@ void r100_cp_disable(struct radeon_device *rdev)
 	}
 }
 
+void r100_cp_commit(struct radeon_device *rdev)
+{
+	WREG32(RADEON_CP_RB_WPTR, rdev->cp.wptr);
+	(void)RREG32(RADEON_CP_RB_WPTR);
+}
+
+
 /*
  * CS functions
  */
@@ -1589,17 +1435,7 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 			r100_cs_dump_packet(p, pkt);
 			return r;
 		}
-		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
-			if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
-				tile_flags |= RADEON_TXO_MACRO_TILE;
-			if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
-				tile_flags |= RADEON_TXO_MICRO_TILE_X2;
-
-			tmp = idx_value & ~(0x7 << 2);
-			tmp |= tile_flags;
-			ib[idx] = tmp + ((u32)reloc->lobj.gpu_offset);
-		} else
-			ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset);
+		ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset);
 		track->textures[i].robj = reloc->robj;
 		track->tex_dirty = true;
 		break;
@@ -1670,17 +1506,15 @@ static int r100_packet0_check(struct radeon_cs_parser *p,
 			r100_cs_dump_packet(p, pkt);
 			return r;
 		}
-		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
-			if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
-				tile_flags |= RADEON_COLOR_TILE_ENABLE;
-			if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
-				tile_flags |= RADEON_COLOR_MICROTILE_ENABLE;
-
-			tmp = idx_value & ~(0x7 << 16);
-			tmp |= tile_flags;
-			ib[idx] = tmp;
-		} else
-			ib[idx] = idx_value;
+
+		if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
+			tile_flags |= RADEON_COLOR_TILE_ENABLE;
+		if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
+			tile_flags |= RADEON_COLOR_MICROTILE_ENABLE;
+
+		tmp = idx_value & ~(0x7 << 16);
+		tmp |= tile_flags;
+		ib[idx] = tmp;
 
 		track->cb[0].pitch = idx_value & RADEON_COLORPITCH_MASK;
 		track->cb_dirty = true;
@@ -2155,9 +1989,9 @@ int r100_mc_wait_for_idle(struct radeon_device *rdev)
 	return -1;
 }
 
-void r100_gpu_lockup_update(struct r100_gpu_lockup *lockup, struct radeon_ring *ring)
+void r100_gpu_lockup_update(struct r100_gpu_lockup *lockup, struct radeon_cp *cp)
 {
-	lockup->last_cp_rptr = ring->rptr;
+	lockup->last_cp_rptr = cp->rptr;
 	lockup->last_jiffies = jiffies;
 }
 
@@ -2182,20 +2016,20 @@ void r100_gpu_lockup_update(struct r100_gpu_lockup *lockup, struct radeon_ring *
  * false positive when CP is just gived nothing to do.
  *
  **/
-bool r100_gpu_cp_is_lockup(struct radeon_device *rdev, struct r100_gpu_lockup *lockup, struct radeon_ring *ring)
+bool r100_gpu_cp_is_lockup(struct radeon_device *rdev, struct r100_gpu_lockup *lockup, struct radeon_cp *cp)
 {
 	unsigned long cjiffies, elapsed;
 
 	cjiffies = jiffies;
 	if (!time_after(cjiffies, lockup->last_jiffies)) {
 		/* likely a wrap around */
-		lockup->last_cp_rptr = ring->rptr;
+		lockup->last_cp_rptr = cp->rptr;
 		lockup->last_jiffies = jiffies;
 		return false;
 	}
-	if (ring->rptr != lockup->last_cp_rptr) {
+	if (cp->rptr != lockup->last_cp_rptr) {
 		/* CP is still working no lockup */
-		lockup->last_cp_rptr = ring->rptr;
+		lockup->last_cp_rptr = cp->rptr;
 		lockup->last_jiffies = jiffies;
 		return false;
 	}
@@ -2208,26 +2042,26 @@ bool r100_gpu_cp_is_lockup(struct radeon_device *rdev, struct r100_gpu_lockup *l
 	return false;
 }
 
-bool r100_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
+bool r100_gpu_is_lockup(struct radeon_device *rdev)
 {
 	u32 rbbm_status;
 	int r;
 
 	rbbm_status = RREG32(R_000E40_RBBM_STATUS);
 	if (!G_000E40_GUI_ACTIVE(rbbm_status)) {
-		r100_gpu_lockup_update(&rdev->config.r100.lockup, ring);
+		r100_gpu_lockup_update(&rdev->config.r100.lockup, &rdev->cp);
 		return false;
 	}
 	/* force CP activities */
-	r = radeon_ring_lock(rdev, ring, 2);
+	r = radeon_ring_lock(rdev, 2);
 	if (!r) {
 		/* PACKET2 NOP */
-		radeon_ring_write(ring, 0x80000000);
-		radeon_ring_write(ring, 0x80000000);
-		radeon_ring_unlock_commit(rdev, ring);
+		radeon_ring_write(rdev, 0x80000000);
+		radeon_ring_write(rdev, 0x80000000);
+		radeon_ring_unlock_commit(rdev);
 	}
-	ring->rptr = RREG32(ring->rptr_reg);
-	return r100_gpu_cp_is_lockup(rdev, &rdev->config.r100.lockup, ring);
+	rdev->cp.rptr = RREG32(RADEON_CP_RB_RPTR);
+	return r100_gpu_cp_is_lockup(rdev, &rdev->config.r100.lockup, &rdev->cp);
 }
 
 void r100_bm_disable(struct radeon_device *rdev)
@@ -2244,7 +2078,8 @@ void r100_bm_disable(struct radeon_device *rdev)
 	WREG32(R_000030_BUS_CNTL, (tmp & 0xFFFFFFFF) | 0x00000040);
 	tmp = RREG32(RADEON_BUS_CNTL);
 	mdelay(1);
-	pci_clear_master(rdev->pdev);
+	pci_read_config_word(rdev->pdev, 0x4, &tmp16);
+	pci_write_config_word(rdev->pdev, 0x4, tmp16 & 0xFFFB);
 	mdelay(1);
 }
 
@@ -2635,22 +2470,21 @@ static int r100_debugfs_cp_ring_info(struct seq_file *m, void *data)
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
 	struct drm_device *dev = node->minor->dev;
 	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	uint32_t rdp, wdp;
 	unsigned count, i, j;
 
-	radeon_ring_free_size(rdev, ring);
+	radeon_ring_free_size(rdev);
 	rdp = RREG32(RADEON_CP_RB_RPTR);
 	wdp = RREG32(RADEON_CP_RB_WPTR);
-	count = (rdp + ring->ring_size - wdp) & ring->ptr_mask;
+	count = (rdp + rdev->cp.ring_size - wdp) & rdev->cp.ptr_mask;
 	seq_printf(m, "CP_STAT 0x%08x\n", RREG32(RADEON_CP_STAT));
 	seq_printf(m, "CP_RB_WPTR 0x%08x\n", wdp);
 	seq_printf(m, "CP_RB_RPTR 0x%08x\n", rdp);
-	seq_printf(m, "%u free dwords in ring\n", ring->ring_free_dw);
+	seq_printf(m, "%u free dwords in ring\n", rdev->cp.ring_free_dw);
 	seq_printf(m, "%u dwords in ring\n", count);
 	for (j = 0; j <= count; j++) {
-		i = (rdp + j) & ring->ptr_mask;
-		seq_printf(m, "r[%04d]=0x%08x\n", i, ring->ring[i]);
+		i = (rdp + j) & rdev->cp.ptr_mask;
+		seq_printf(m, "r[%04d]=0x%08x\n", i, rdev->cp.ring[i]);
 	}
 	return 0;
 }
@@ -3318,7 +3152,7 @@ void r100_bandwidth_update(struct radeon_device *rdev)
 	}
 }
 
-static void r100_cs_track_texture_print(struct r100_cs_track_texture *t)
+static inline void r100_cs_track_texture_print(struct r100_cs_track_texture *t)
 {
 	DRM_ERROR("pitch                      %d\n", t->pitch);
 	DRM_ERROR("use_pitch                  %d\n", t->use_pitch);
@@ -3692,7 +3526,7 @@ void r100_cs_track_clear(struct radeon_device *rdev, struct r100_cs_track *track
 	}
 }
 
-int r100_ring_test(struct radeon_device *rdev, struct radeon_ring *ring)
+int r100_ring_test(struct radeon_device *rdev)
 {
 	uint32_t scratch;
 	uint32_t tmp = 0;
@@ -3705,15 +3539,15 @@ int r100_ring_test(struct radeon_device *rdev, struct radeon_ring *ring)
 		return r;
 	}
 	WREG32(scratch, 0xCAFEDEAD);
-	r = radeon_ring_lock(rdev, ring, 2);
+	r = radeon_ring_lock(rdev, 2);
 	if (r) {
 		DRM_ERROR("radeon: cp failed to lock ring (%d).\n", r);
 		radeon_scratch_free(rdev, scratch);
 		return r;
 	}
-	radeon_ring_write(ring, PACKET0(scratch, 0));
-	radeon_ring_write(ring, 0xDEADBEEF);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_write(rdev, PACKET0(scratch, 0));
+	radeon_ring_write(rdev, 0xDEADBEEF);
+	radeon_ring_unlock_commit(rdev);
 	for (i = 0; i < rdev->usec_timeout; i++) {
 		tmp = RREG32(scratch);
 		if (tmp == 0xDEADBEEF) {
@@ -3734,14 +3568,12 @@ int r100_ring_test(struct radeon_device *rdev, struct radeon_ring *ring)
 
 void r100_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
-
-	radeon_ring_write(ring, PACKET0(RADEON_CP_IB_BASE, 1));
-	radeon_ring_write(ring, ib->gpu_addr);
-	radeon_ring_write(ring, ib->length_dw);
+	radeon_ring_write(rdev, PACKET0(RADEON_CP_IB_BASE, 1));
+	radeon_ring_write(rdev, ib->gpu_addr);
+	radeon_ring_write(rdev, ib->length_dw);
 }
 
-int r100_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
+int r100_ib_test(struct radeon_device *rdev)
 {
 	struct radeon_ib *ib;
 	uint32_t scratch;
@@ -3755,7 +3587,7 @@ int r100_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 		return r;
 	}
 	WREG32(scratch, 0xCAFEDEAD);
-	r = radeon_ib_get(rdev, RADEON_RING_TYPE_GFX_INDEX, &ib, 256);
+	r = radeon_ib_get(rdev, &ib);
 	if (r) {
 		return r;
 	}
@@ -3799,16 +3631,34 @@ int r100_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 
 void r100_ib_fini(struct radeon_device *rdev)
 {
-	radeon_ib_pool_suspend(rdev);
 	radeon_ib_pool_fini(rdev);
 }
 
+int r100_ib_init(struct radeon_device *rdev)
+{
+	int r;
+
+	r = radeon_ib_pool_init(rdev);
+	if (r) {
+		dev_err(rdev->dev, "failed initializing IB pool (%d).\n", r);
+		r100_ib_fini(rdev);
+		return r;
+	}
+	r = r100_ib_test(rdev);
+	if (r) {
+		dev_err(rdev->dev, "failed testing IB (%d).\n", r);
+		r100_ib_fini(rdev);
+		return r;
+	}
+	return 0;
+}
+
 void r100_mc_stop(struct radeon_device *rdev, struct r100_mc_save *save)
 {
 	/* Shutdown CP we shouldn't need to do that but better be safe than
 	 * sorry
 	 */
-	rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false;
+	rdev->cp.ready = false;
 	WREG32(R_000740_CP_CSQ_CNTL, 0);
 
 	/* Save few CRTC registers */
@@ -3946,12 +3796,6 @@ static int r100_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX);
-	if (r) {
-		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
-		return r;
-	}
-
 	/* Enable IRQ */
 	r100_irq_set(rdev);
 	rdev->config.r100.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL);
@@ -3961,25 +3805,16 @@ static int r100_startup(struct radeon_device *rdev)
 		dev_err(rdev->dev, "failed initializing CP (%d).\n", r);
 		return r;
 	}
-
-	r = radeon_ib_pool_start(rdev);
-	if (r)
-		return r;
-
-	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
+	r = r100_ib_init(rdev);
 	if (r) {
-		dev_err(rdev->dev, "failed testing IB (%d).\n", r);
-		rdev->accel_working = false;
+		dev_err(rdev->dev, "failed initializing IB (%d).\n", r);
 		return r;
 	}
-
 	return 0;
 }
 
 int r100_resume(struct radeon_device *rdev)
 {
-	int r;
-
 	/* Make sur GART are not working */
 	if (rdev->flags & RADEON_IS_PCI)
 		r100_pci_gart_disable(rdev);
@@ -3997,18 +3832,11 @@ int r100_resume(struct radeon_device *rdev)
 	r100_clock_startup(rdev);
 	/* Initialize surface registers */
 	radeon_surface_init(rdev);
-
-	rdev->accel_working = true;
-	r = r100_startup(rdev);
-	if (r) {
-		rdev->accel_working = false;
-	}
-	return r;
+	return r100_startup(rdev);
 }
 
 int r100_suspend(struct radeon_device *rdev)
 {
-	radeon_ib_pool_suspend(rdev);
 	r100_cp_disable(rdev);
 	radeon_wb_disable(rdev);
 	r100_irq_disable(rdev);
@@ -4127,14 +3955,7 @@ int r100_init(struct radeon_device *rdev)
 			return r;
 	}
 	r100_set_safe_registers(rdev);
-
-	r = radeon_ib_pool_init(rdev);
 	rdev->accel_working = true;
-	if (r) {
-		dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
-		rdev->accel_working = false;
-	}
-
 	r = r100_startup(rdev);
 	if (r) {
 		/* Somethings want wront with the accel init stop accel */
@@ -4149,43 +3970,3 @@ int r100_init(struct radeon_device *rdev)
 	}
 	return 0;
 }
-
-uint32_t r100_mm_rreg(struct radeon_device *rdev, uint32_t reg)
-{
-	if (reg < rdev->rmmio_size)
-		return readl(((void __iomem *)rdev->rmmio) + reg);
-	else {
-		writel(reg, ((void __iomem *)rdev->rmmio) + RADEON_MM_INDEX);
-		return readl(((void __iomem *)rdev->rmmio) + RADEON_MM_DATA);
-	}
-}
-
-void r100_mm_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v)
-{
-	if (reg < rdev->rmmio_size)
-		writel(v, ((void __iomem *)rdev->rmmio) + reg);
-	else {
-		writel(reg, ((void __iomem *)rdev->rmmio) + RADEON_MM_INDEX);
-		writel(v, ((void __iomem *)rdev->rmmio) + RADEON_MM_DATA);
-	}
-}
-
-u32 r100_io_rreg(struct radeon_device *rdev, u32 reg)
-{
-	if (reg < rdev->rio_mem_size)
-		return ioread32(rdev->rio_mem + reg);
-	else {
-		iowrite32(reg, rdev->rio_mem + RADEON_MM_INDEX);
-		return ioread32(rdev->rio_mem + RADEON_MM_DATA);
-	}
-}
-
-void r100_io_wreg(struct radeon_device *rdev, u32 reg, u32 v)
-{
-	if (reg < rdev->rio_mem_size)
-		iowrite32(v, rdev->rio_mem + reg);
-	else {
-		iowrite32(reg, rdev->rio_mem + RADEON_MM_INDEX);
-		iowrite32(v, rdev->rio_mem + RADEON_MM_DATA);
-	}
-}
diff --git a/drivers/gpu/drm/radeon/r100_track.h b/drivers/gpu/drm/radeon/r100_track.h
index 6a603b3..686f9dc 100644
--- a/drivers/gpu/drm/radeon/r100_track.h
+++ b/drivers/gpu/drm/radeon/r100_track.h
@@ -92,10 +92,106 @@ int r200_packet0_check(struct radeon_cs_parser *p,
 		       struct radeon_cs_packet *pkt,
 		       unsigned idx, unsigned reg);
 
-int r100_reloc_pitch_offset(struct radeon_cs_parser *p,
-			    struct radeon_cs_packet *pkt,
-			    unsigned idx,
-			    unsigned reg);
-int r100_packet3_load_vbpntr(struct radeon_cs_parser *p,
-			     struct radeon_cs_packet *pkt,
-			     int idx);
+
+
+static inline int r100_reloc_pitch_offset(struct radeon_cs_parser *p,
+					  struct radeon_cs_packet *pkt,
+					  unsigned idx,
+					  unsigned reg)
+{
+	int r;
+	u32 tile_flags = 0;
+	u32 tmp;
+	struct radeon_cs_reloc *reloc;
+	u32 value;
+
+	r = r100_cs_packet_next_reloc(p, &reloc);
+	if (r) {
+		DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
+			  idx, reg);
+		r100_cs_dump_packet(p, pkt);
+		return r;
+	}
+	value = radeon_get_ib_value(p, idx);
+	tmp = value & 0x003fffff;
+	tmp += (((u32)reloc->lobj.gpu_offset) >> 10);
+
+	if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
+		tile_flags |= RADEON_DST_TILE_MACRO;
+	if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO) {
+		if (reg == RADEON_SRC_PITCH_OFFSET) {
+			DRM_ERROR("Cannot src blit from microtiled surface\n");
+			r100_cs_dump_packet(p, pkt);
+			return -EINVAL;
+		}
+		tile_flags |= RADEON_DST_TILE_MICRO;
+	}
+
+	tmp |= tile_flags;
+	p->ib->ptr[idx] = (value & 0x3fc00000) | tmp;
+	return 0;
+}
+
+static inline int r100_packet3_load_vbpntr(struct radeon_cs_parser *p,
+					   struct radeon_cs_packet *pkt,
+					   int idx)
+{
+	unsigned c, i;
+	struct radeon_cs_reloc *reloc;
+	struct r100_cs_track *track;
+	int r = 0;
+	volatile uint32_t *ib;
+	u32 idx_value;
+
+	ib = p->ib->ptr;
+	track = (struct r100_cs_track *)p->track;
+	c = radeon_get_ib_value(p, idx++) & 0x1F;
+	if (c > 16) {
+	    DRM_ERROR("Only 16 vertex buffers are allowed %d\n",
+		      pkt->opcode);
+	    r100_cs_dump_packet(p, pkt);
+	    return -EINVAL;
+	}
+	track->num_arrays = c;
+	for (i = 0; i < (c - 1); i+=2, idx+=3) {
+		r = r100_cs_packet_next_reloc(p, &reloc);
+		if (r) {
+			DRM_ERROR("No reloc for packet3 %d\n",
+				  pkt->opcode);
+			r100_cs_dump_packet(p, pkt);
+			return r;
+		}
+		idx_value = radeon_get_ib_value(p, idx);
+		ib[idx+1] = radeon_get_ib_value(p, idx + 1) + ((u32)reloc->lobj.gpu_offset);
+
+		track->arrays[i + 0].esize = idx_value >> 8;
+		track->arrays[i + 0].robj = reloc->robj;
+		track->arrays[i + 0].esize &= 0x7F;
+		r = r100_cs_packet_next_reloc(p, &reloc);
+		if (r) {
+			DRM_ERROR("No reloc for packet3 %d\n",
+				  pkt->opcode);
+			r100_cs_dump_packet(p, pkt);
+			return r;
+		}
+		ib[idx+2] = radeon_get_ib_value(p, idx + 2) + ((u32)reloc->lobj.gpu_offset);
+		track->arrays[i + 1].robj = reloc->robj;
+		track->arrays[i + 1].esize = idx_value >> 24;
+		track->arrays[i + 1].esize &= 0x7F;
+	}
+	if (c & 1) {
+		r = r100_cs_packet_next_reloc(p, &reloc);
+		if (r) {
+			DRM_ERROR("No reloc for packet3 %d\n",
+					  pkt->opcode);
+			r100_cs_dump_packet(p, pkt);
+			return r;
+		}
+		idx_value = radeon_get_ib_value(p, idx);
+		ib[idx+1] = radeon_get_ib_value(p, idx + 1) + ((u32)reloc->lobj.gpu_offset);
+		track->arrays[i + 0].robj = reloc->robj;
+		track->arrays[i + 0].esize = idx_value >> 8;
+		track->arrays[i + 0].esize &= 0x7F;
+	}
+	return r;
+}
diff --git a/drivers/gpu/drm/radeon/r200.c b/drivers/gpu/drm/radeon/r200.c
index a59cc47..a1f3ba0 100644
--- a/drivers/gpu/drm/radeon/r200.c
+++ b/drivers/gpu/drm/radeon/r200.c
@@ -87,7 +87,6 @@ int r200_copy_dma(struct radeon_device *rdev,
 		  unsigned num_gpu_pages,
 		  struct radeon_fence *fence)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	uint32_t size;
 	uint32_t cur_size;
 	int i, num_loops;
@@ -96,33 +95,33 @@ int r200_copy_dma(struct radeon_device *rdev,
 	/* radeon pitch is /64 */
 	size = num_gpu_pages << RADEON_GPU_PAGE_SHIFT;
 	num_loops = DIV_ROUND_UP(size, 0x1FFFFF);
-	r = radeon_ring_lock(rdev, ring, num_loops * 4 + 64);
+	r = radeon_ring_lock(rdev, num_loops * 4 + 64);
 	if (r) {
 		DRM_ERROR("radeon: moving bo (%d).\n", r);
 		return r;
 	}
 	/* Must wait for 2D idle & clean before DMA or hangs might happen */
-	radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0));
-	radeon_ring_write(ring, (1 << 16));
+	radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0));
+	radeon_ring_write(rdev, (1 << 16));
 	for (i = 0; i < num_loops; i++) {
 		cur_size = size;
 		if (cur_size > 0x1FFFFF) {
 			cur_size = 0x1FFFFF;
 		}
 		size -= cur_size;
-		radeon_ring_write(ring, PACKET0(0x720, 2));
-		radeon_ring_write(ring, src_offset);
-		radeon_ring_write(ring, dst_offset);
-		radeon_ring_write(ring, cur_size | (1 << 31) | (1 << 30));
+		radeon_ring_write(rdev, PACKET0(0x720, 2));
+		radeon_ring_write(rdev, src_offset);
+		radeon_ring_write(rdev, dst_offset);
+		radeon_ring_write(rdev, cur_size | (1 << 31) | (1 << 30));
 		src_offset += cur_size;
 		dst_offset += cur_size;
 	}
-	radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0));
-	radeon_ring_write(ring, RADEON_WAIT_DMA_GUI_IDLE);
+	radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0));
+	radeon_ring_write(rdev, RADEON_WAIT_DMA_GUI_IDLE);
 	if (fence) {
 		r = radeon_fence_emit(rdev, fence);
 	}
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev);
 	return r;
 }
 
@@ -215,17 +214,7 @@ int r200_packet0_check(struct radeon_cs_parser *p,
 			r100_cs_dump_packet(p, pkt);
 			return r;
 		}
-		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
-			if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
-				tile_flags |= R200_TXO_MACRO_TILE;
-			if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
-				tile_flags |= R200_TXO_MICRO_TILE;
-
-			tmp = idx_value & ~(0x7 << 2);
-			tmp |= tile_flags;
-			ib[idx] = tmp + ((u32)reloc->lobj.gpu_offset);
-		} else
-			ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset);
+		ib[idx] = idx_value + ((u32)reloc->lobj.gpu_offset);
 		track->textures[i].robj = reloc->robj;
 		track->tex_dirty = true;
 		break;
@@ -287,17 +276,14 @@ int r200_packet0_check(struct radeon_cs_parser *p,
 			return r;
 		}
 
-		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
-			if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
-				tile_flags |= RADEON_COLOR_TILE_ENABLE;
-			if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
-				tile_flags |= RADEON_COLOR_MICROTILE_ENABLE;
+		if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
+			tile_flags |= RADEON_COLOR_TILE_ENABLE;
+		if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
+			tile_flags |= RADEON_COLOR_MICROTILE_ENABLE;
 
-			tmp = idx_value & ~(0x7 << 16);
-			tmp |= tile_flags;
-			ib[idx] = tmp;
-		} else
-			ib[idx] = idx_value;
+		tmp = idx_value & ~(0x7 << 16);
+		tmp |= tile_flags;
+		ib[idx] = tmp;
 
 		track->cb[0].pitch = idx_value & RADEON_COLORPITCH_MASK;
 		track->cb_dirty = true;
diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c
index fa14383..55a7f19 100644
--- a/drivers/gpu/drm/radeon/r300.c
+++ b/drivers/gpu/drm/radeon/r300.c
@@ -74,7 +74,7 @@ void rv370_pcie_gart_tlb_flush(struct radeon_device *rdev)
 
 int rv370_pcie_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr)
 {
-	void __iomem *ptr = rdev->gart.ptr;
+	void __iomem *ptr = (void *)rdev->gart.table.vram.ptr;
 
 	if (i < 0 || i > rdev->gart.num_gpu_pages) {
 		return -EINVAL;
@@ -93,7 +93,7 @@ int rv370_pcie_gart_init(struct radeon_device *rdev)
 {
 	int r;
 
-	if (rdev->gart.robj) {
+	if (rdev->gart.table.vram.robj) {
 		WARN(1, "RV370 PCIE GART already initialized\n");
 		return 0;
 	}
@@ -105,8 +105,8 @@ int rv370_pcie_gart_init(struct radeon_device *rdev)
 	if (r)
 		DRM_ERROR("Failed to register debugfs file for PCIE gart !\n");
 	rdev->gart.table_size = rdev->gart.num_gpu_pages * 4;
-	rdev->asic->gart.tlb_flush = &rv370_pcie_gart_tlb_flush;
-	rdev->asic->gart.set_page = &rv370_pcie_gart_set_page;
+	rdev->asic->gart_tlb_flush = &rv370_pcie_gart_tlb_flush;
+	rdev->asic->gart_set_page = &rv370_pcie_gart_set_page;
 	return radeon_gart_table_vram_alloc(rdev);
 }
 
@@ -116,7 +116,7 @@ int rv370_pcie_gart_enable(struct radeon_device *rdev)
 	uint32_t tmp;
 	int r;
 
-	if (rdev->gart.robj == NULL) {
+	if (rdev->gart.table.vram.robj == NULL) {
 		dev_err(rdev->dev, "No VRAM object for PCIE GART.\n");
 		return -EINVAL;
 	}
@@ -144,9 +144,8 @@ int rv370_pcie_gart_enable(struct radeon_device *rdev)
 	tmp |= RADEON_PCIE_TX_GART_UNMAPPED_ACCESS_DISCARD;
 	WREG32_PCIE(RADEON_PCIE_TX_GART_CNTL, tmp);
 	rv370_pcie_gart_tlb_flush(rdev);
-	DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n",
-		 (unsigned)(rdev->mc.gtt_size >> 20),
-		 (unsigned long long)table_addr);
+	DRM_INFO("PCIE GART of %uM enabled (table at 0x%08X).\n",
+		 (unsigned)(rdev->mc.gtt_size >> 20), table_addr);
 	rdev->gart.ready = true;
 	return 0;
 }
@@ -154,6 +153,7 @@ int rv370_pcie_gart_enable(struct radeon_device *rdev)
 void rv370_pcie_gart_disable(struct radeon_device *rdev)
 {
 	u32 tmp;
+	int r;
 
 	WREG32_PCIE(RADEON_PCIE_TX_GART_START_LO, 0);
 	WREG32_PCIE(RADEON_PCIE_TX_GART_END_LO, 0);
@@ -162,7 +162,14 @@ void rv370_pcie_gart_disable(struct radeon_device *rdev)
 	tmp = RREG32_PCIE(RADEON_PCIE_TX_GART_CNTL);
 	tmp |= RADEON_PCIE_TX_GART_UNMAPPED_ACCESS_DISCARD;
 	WREG32_PCIE(RADEON_PCIE_TX_GART_CNTL, tmp & ~RADEON_PCIE_TX_GART_EN);
-	radeon_gart_table_vram_unpin(rdev);
+	if (rdev->gart.table.vram.robj) {
+		r = radeon_bo_reserve(rdev->gart.table.vram.robj, false);
+		if (likely(r == 0)) {
+			radeon_bo_kunmap(rdev->gart.table.vram.robj);
+			radeon_bo_unpin(rdev->gart.table.vram.robj);
+			radeon_bo_unreserve(rdev->gart.table.vram.robj);
+		}
+	}
 }
 
 void rv370_pcie_gart_fini(struct radeon_device *rdev)
@@ -175,38 +182,36 @@ void rv370_pcie_gart_fini(struct radeon_device *rdev)
 void r300_fence_ring_emit(struct radeon_device *rdev,
 			  struct radeon_fence *fence)
 {
-	struct radeon_ring *ring = &rdev->ring[fence->ring];
-
 	/* Who ever call radeon_fence_emit should call ring_lock and ask
 	 * for enough space (today caller are ib schedule and buffer move) */
 	/* Write SC register so SC & US assert idle */
-	radeon_ring_write(ring, PACKET0(R300_RE_SCISSORS_TL, 0));
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, PACKET0(R300_RE_SCISSORS_BR, 0));
-	radeon_ring_write(ring, 0);
+	radeon_ring_write(rdev, PACKET0(R300_RE_SCISSORS_TL, 0));
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, PACKET0(R300_RE_SCISSORS_BR, 0));
+	radeon_ring_write(rdev, 0);
 	/* Flush 3D cache */
-	radeon_ring_write(ring, PACKET0(R300_RB3D_DSTCACHE_CTLSTAT, 0));
-	radeon_ring_write(ring, R300_RB3D_DC_FLUSH);
-	radeon_ring_write(ring, PACKET0(R300_RB3D_ZCACHE_CTLSTAT, 0));
-	radeon_ring_write(ring, R300_ZC_FLUSH);
+	radeon_ring_write(rdev, PACKET0(R300_RB3D_DSTCACHE_CTLSTAT, 0));
+	radeon_ring_write(rdev, R300_RB3D_DC_FLUSH);
+	radeon_ring_write(rdev, PACKET0(R300_RB3D_ZCACHE_CTLSTAT, 0));
+	radeon_ring_write(rdev, R300_ZC_FLUSH);
 	/* Wait until IDLE & CLEAN */
-	radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0));
-	radeon_ring_write(ring, (RADEON_WAIT_3D_IDLECLEAN |
+	radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0));
+	radeon_ring_write(rdev, (RADEON_WAIT_3D_IDLECLEAN |
 				 RADEON_WAIT_2D_IDLECLEAN |
 				 RADEON_WAIT_DMA_GUI_IDLE));
-	radeon_ring_write(ring, PACKET0(RADEON_HOST_PATH_CNTL, 0));
-	radeon_ring_write(ring, rdev->config.r300.hdp_cntl |
+	radeon_ring_write(rdev, PACKET0(RADEON_HOST_PATH_CNTL, 0));
+	radeon_ring_write(rdev, rdev->config.r300.hdp_cntl |
 				RADEON_HDP_READ_BUFFER_INVALIDATE);
-	radeon_ring_write(ring, PACKET0(RADEON_HOST_PATH_CNTL, 0));
-	radeon_ring_write(ring, rdev->config.r300.hdp_cntl);
+	radeon_ring_write(rdev, PACKET0(RADEON_HOST_PATH_CNTL, 0));
+	radeon_ring_write(rdev, rdev->config.r300.hdp_cntl);
 	/* Emit fence sequence & fire IRQ */
-	radeon_ring_write(ring, PACKET0(rdev->fence_drv[fence->ring].scratch_reg, 0));
-	radeon_ring_write(ring, fence->seq);
-	radeon_ring_write(ring, PACKET0(RADEON_GEN_INT_STATUS, 0));
-	radeon_ring_write(ring, RADEON_SW_INT_FIRE);
+	radeon_ring_write(rdev, PACKET0(rdev->fence_drv.scratch_reg, 0));
+	radeon_ring_write(rdev, fence->seq);
+	radeon_ring_write(rdev, PACKET0(RADEON_GEN_INT_STATUS, 0));
+	radeon_ring_write(rdev, RADEON_SW_INT_FIRE);
 }
 
-void r300_ring_start(struct radeon_device *rdev, struct radeon_ring *ring)
+void r300_ring_start(struct radeon_device *rdev)
 {
 	unsigned gb_tile_config;
 	int r;
@@ -229,44 +234,44 @@ void r300_ring_start(struct radeon_device *rdev, struct radeon_ring *ring)
 		break;
 	}
 
-	r = radeon_ring_lock(rdev, ring, 64);
+	r = radeon_ring_lock(rdev, 64);
 	if (r) {
 		return;
 	}
-	radeon_ring_write(ring, PACKET0(RADEON_ISYNC_CNTL, 0));
-	radeon_ring_write(ring,
+	radeon_ring_write(rdev, PACKET0(RADEON_ISYNC_CNTL, 0));
+	radeon_ring_write(rdev,
 			  RADEON_ISYNC_ANY2D_IDLE3D |
 			  RADEON_ISYNC_ANY3D_IDLE2D |
 			  RADEON_ISYNC_WAIT_IDLEGUI |
 			  RADEON_ISYNC_CPSCRATCH_IDLEGUI);
-	radeon_ring_write(ring, PACKET0(R300_GB_TILE_CONFIG, 0));
-	radeon_ring_write(ring, gb_tile_config);
-	radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0));
-	radeon_ring_write(ring,
+	radeon_ring_write(rdev, PACKET0(R300_GB_TILE_CONFIG, 0));
+	radeon_ring_write(rdev, gb_tile_config);
+	radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0));
+	radeon_ring_write(rdev,
 			  RADEON_WAIT_2D_IDLECLEAN |
 			  RADEON_WAIT_3D_IDLECLEAN);
-	radeon_ring_write(ring, PACKET0(R300_DST_PIPE_CONFIG, 0));
-	radeon_ring_write(ring, R300_PIPE_AUTO_CONFIG);
-	radeon_ring_write(ring, PACKET0(R300_GB_SELECT, 0));
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, PACKET0(R300_GB_ENABLE, 0));
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, PACKET0(R300_RB3D_DSTCACHE_CTLSTAT, 0));
-	radeon_ring_write(ring, R300_RB3D_DC_FLUSH | R300_RB3D_DC_FREE);
-	radeon_ring_write(ring, PACKET0(R300_RB3D_ZCACHE_CTLSTAT, 0));
-	radeon_ring_write(ring, R300_ZC_FLUSH | R300_ZC_FREE);
-	radeon_ring_write(ring, PACKET0(RADEON_WAIT_UNTIL, 0));
-	radeon_ring_write(ring,
+	radeon_ring_write(rdev, PACKET0(R300_DST_PIPE_CONFIG, 0));
+	radeon_ring_write(rdev, R300_PIPE_AUTO_CONFIG);
+	radeon_ring_write(rdev, PACKET0(R300_GB_SELECT, 0));
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, PACKET0(R300_GB_ENABLE, 0));
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, PACKET0(R300_RB3D_DSTCACHE_CTLSTAT, 0));
+	radeon_ring_write(rdev, R300_RB3D_DC_FLUSH | R300_RB3D_DC_FREE);
+	radeon_ring_write(rdev, PACKET0(R300_RB3D_ZCACHE_CTLSTAT, 0));
+	radeon_ring_write(rdev, R300_ZC_FLUSH | R300_ZC_FREE);
+	radeon_ring_write(rdev, PACKET0(RADEON_WAIT_UNTIL, 0));
+	radeon_ring_write(rdev,
 			  RADEON_WAIT_2D_IDLECLEAN |
 			  RADEON_WAIT_3D_IDLECLEAN);
-	radeon_ring_write(ring, PACKET0(R300_GB_AA_CONFIG, 0));
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, PACKET0(R300_RB3D_DSTCACHE_CTLSTAT, 0));
-	radeon_ring_write(ring, R300_RB3D_DC_FLUSH | R300_RB3D_DC_FREE);
-	radeon_ring_write(ring, PACKET0(R300_RB3D_ZCACHE_CTLSTAT, 0));
-	radeon_ring_write(ring, R300_ZC_FLUSH | R300_ZC_FREE);
-	radeon_ring_write(ring, PACKET0(R300_GB_MSPOS0, 0));
-	radeon_ring_write(ring,
+	radeon_ring_write(rdev, PACKET0(R300_GB_AA_CONFIG, 0));
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, PACKET0(R300_RB3D_DSTCACHE_CTLSTAT, 0));
+	radeon_ring_write(rdev, R300_RB3D_DC_FLUSH | R300_RB3D_DC_FREE);
+	radeon_ring_write(rdev, PACKET0(R300_RB3D_ZCACHE_CTLSTAT, 0));
+	radeon_ring_write(rdev, R300_ZC_FLUSH | R300_ZC_FREE);
+	radeon_ring_write(rdev, PACKET0(R300_GB_MSPOS0, 0));
+	radeon_ring_write(rdev,
 			  ((6 << R300_MS_X0_SHIFT) |
 			   (6 << R300_MS_Y0_SHIFT) |
 			   (6 << R300_MS_X1_SHIFT) |
@@ -275,8 +280,8 @@ void r300_ring_start(struct radeon_device *rdev, struct radeon_ring *ring)
 			   (6 << R300_MS_Y2_SHIFT) |
 			   (6 << R300_MSBD0_Y_SHIFT) |
 			   (6 << R300_MSBD0_X_SHIFT)));
-	radeon_ring_write(ring, PACKET0(R300_GB_MSPOS1, 0));
-	radeon_ring_write(ring,
+	radeon_ring_write(rdev, PACKET0(R300_GB_MSPOS1, 0));
+	radeon_ring_write(rdev,
 			  ((6 << R300_MS_X3_SHIFT) |
 			   (6 << R300_MS_Y3_SHIFT) |
 			   (6 << R300_MS_X4_SHIFT) |
@@ -284,16 +289,16 @@ void r300_ring_start(struct radeon_device *rdev, struct radeon_ring *ring)
 			   (6 << R300_MS_X5_SHIFT) |
 			   (6 << R300_MS_Y5_SHIFT) |
 			   (6 << R300_MSBD1_SHIFT)));
-	radeon_ring_write(ring, PACKET0(R300_GA_ENHANCE, 0));
-	radeon_ring_write(ring, R300_GA_DEADLOCK_CNTL | R300_GA_FASTSYNC_CNTL);
-	radeon_ring_write(ring, PACKET0(R300_GA_POLY_MODE, 0));
-	radeon_ring_write(ring,
+	radeon_ring_write(rdev, PACKET0(R300_GA_ENHANCE, 0));
+	radeon_ring_write(rdev, R300_GA_DEADLOCK_CNTL | R300_GA_FASTSYNC_CNTL);
+	radeon_ring_write(rdev, PACKET0(R300_GA_POLY_MODE, 0));
+	radeon_ring_write(rdev,
 			  R300_FRONT_PTYPE_TRIANGE | R300_BACK_PTYPE_TRIANGE);
-	radeon_ring_write(ring, PACKET0(R300_GA_ROUND_MODE, 0));
-	radeon_ring_write(ring,
+	radeon_ring_write(rdev, PACKET0(R300_GA_ROUND_MODE, 0));
+	radeon_ring_write(rdev,
 			  R300_GEOMETRY_ROUND_NEAREST |
 			  R300_COLOR_ROUND_NEAREST);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_unlock_commit(rdev);
 }
 
 void r300_errata(struct radeon_device *rdev)
@@ -377,26 +382,26 @@ void r300_gpu_init(struct radeon_device *rdev)
 		 rdev->num_gb_pipes, rdev->num_z_pipes);
 }
 
-bool r300_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
+bool r300_gpu_is_lockup(struct radeon_device *rdev)
 {
 	u32 rbbm_status;
 	int r;
 
 	rbbm_status = RREG32(R_000E40_RBBM_STATUS);
 	if (!G_000E40_GUI_ACTIVE(rbbm_status)) {
-		r100_gpu_lockup_update(&rdev->config.r300.lockup, ring);
+		r100_gpu_lockup_update(&rdev->config.r300.lockup, &rdev->cp);
 		return false;
 	}
 	/* force CP activities */
-	r = radeon_ring_lock(rdev, ring, 2);
+	r = radeon_ring_lock(rdev, 2);
 	if (!r) {
 		/* PACKET2 NOP */
-		radeon_ring_write(ring, 0x80000000);
-		radeon_ring_write(ring, 0x80000000);
-		radeon_ring_unlock_commit(rdev, ring);
+		radeon_ring_write(rdev, 0x80000000);
+		radeon_ring_write(rdev, 0x80000000);
+		radeon_ring_unlock_commit(rdev);
 	}
-	ring->rptr = RREG32(RADEON_CP_RB_RPTR);
-	return r100_gpu_cp_is_lockup(rdev, &rdev->config.r300.lockup, ring);
+	rdev->cp.rptr = RREG32(RADEON_CP_RB_RPTR);
+	return r100_gpu_cp_is_lockup(rdev, &rdev->config.r300.lockup, &rdev->cp);
 }
 
 int r300_asic_reset(struct radeon_device *rdev)
@@ -703,21 +708,16 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 			return r;
 		}
 
-		if (p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS) {
-			ib[idx] = (idx_value & 31) | /* keep the 1st 5 bits */
-				  ((idx_value & ~31) + (u32)reloc->lobj.gpu_offset);
-		} else {
-			if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
-				tile_flags |= R300_TXO_MACRO_TILE;
-			if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
-				tile_flags |= R300_TXO_MICRO_TILE;
-			else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO_SQUARE)
-				tile_flags |= R300_TXO_MICRO_TILE_SQUARE;
-
-			tmp = idx_value + ((u32)reloc->lobj.gpu_offset);
-			tmp |= tile_flags;
-			ib[idx] = tmp;
-		}
+		if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
+			tile_flags |= R300_TXO_MACRO_TILE;
+		if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
+			tile_flags |= R300_TXO_MICRO_TILE;
+		else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO_SQUARE)
+			tile_flags |= R300_TXO_MICRO_TILE_SQUARE;
+
+		tmp = idx_value + ((u32)reloc->lobj.gpu_offset);
+		tmp |= tile_flags;
+		ib[idx] = tmp;
 		track->textures[i].robj = reloc->robj;
 		track->tex_dirty = true;
 		break;
@@ -767,26 +767,24 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 		/* RB3D_COLORPITCH1 */
 		/* RB3D_COLORPITCH2 */
 		/* RB3D_COLORPITCH3 */
-		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
-			r = r100_cs_packet_next_reloc(p, &reloc);
-			if (r) {
-				DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
-					  idx, reg);
-				r100_cs_dump_packet(p, pkt);
-				return r;
-			}
+		r = r100_cs_packet_next_reloc(p, &reloc);
+		if (r) {
+			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
+				  idx, reg);
+			r100_cs_dump_packet(p, pkt);
+			return r;
+		}
 
-			if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
-				tile_flags |= R300_COLOR_TILE_ENABLE;
-			if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
-				tile_flags |= R300_COLOR_MICROTILE_ENABLE;
-			else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO_SQUARE)
-				tile_flags |= R300_COLOR_MICROTILE_SQUARE_ENABLE;
+		if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
+			tile_flags |= R300_COLOR_TILE_ENABLE;
+		if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
+			tile_flags |= R300_COLOR_MICROTILE_ENABLE;
+		else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO_SQUARE)
+			tile_flags |= R300_COLOR_MICROTILE_SQUARE_ENABLE;
 
-			tmp = idx_value & ~(0x7 << 16);
-			tmp |= tile_flags;
-			ib[idx] = tmp;
-		}
+		tmp = idx_value & ~(0x7 << 16);
+		tmp |= tile_flags;
+		ib[idx] = tmp;
 		i = (reg - 0x4E38) >> 2;
 		track->cb[i].pitch = idx_value & 0x3FFE;
 		switch (((idx_value >> 21) & 0xF)) {
@@ -852,26 +850,25 @@ static int r300_packet0_check(struct radeon_cs_parser *p,
 		break;
 	case 0x4F24:
 		/* ZB_DEPTHPITCH */
-		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
-			r = r100_cs_packet_next_reloc(p, &reloc);
-			if (r) {
-				DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
-					  idx, reg);
-				r100_cs_dump_packet(p, pkt);
-				return r;
-			}
+		r = r100_cs_packet_next_reloc(p, &reloc);
+		if (r) {
+			DRM_ERROR("No reloc for ib[%d]=0x%04X\n",
+				  idx, reg);
+			r100_cs_dump_packet(p, pkt);
+			return r;
+		}
 
-			if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
-				tile_flags |= R300_DEPTHMACROTILE_ENABLE;
-			if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
-				tile_flags |= R300_DEPTHMICROTILE_TILED;
-			else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO_SQUARE)
-				tile_flags |= R300_DEPTHMICROTILE_TILED_SQUARE;
+		if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
+			tile_flags |= R300_DEPTHMACROTILE_ENABLE;
+		if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
+			tile_flags |= R300_DEPTHMICROTILE_TILED;
+		else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO_SQUARE)
+			tile_flags |= R300_DEPTHMICROTILE_TILED_SQUARE;
+
+		tmp = idx_value & ~(0x7 << 16);
+		tmp |= tile_flags;
+		ib[idx] = tmp;
 
-			tmp = idx_value & ~(0x7 << 16);
-			tmp |= tile_flags;
-			ib[idx] = tmp;
-		}
 		track->zb.pitch = idx_value & 0x3FFC;
 		track->zb_dirty = true;
 		break;
@@ -1398,12 +1395,6 @@ static int r300_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX);
-	if (r) {
-		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
-		return r;
-	}
-
 	/* Enable IRQ */
 	r100_irq_set(rdev);
 	rdev->config.r300.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL);
@@ -1413,25 +1404,16 @@ static int r300_startup(struct radeon_device *rdev)
 		dev_err(rdev->dev, "failed initializing CP (%d).\n", r);
 		return r;
 	}
-
-	r = radeon_ib_pool_start(rdev);
-	if (r)
-		return r;
-
-	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
+	r = r100_ib_init(rdev);
 	if (r) {
-		dev_err(rdev->dev, "failed testing IB (%d).\n", r);
-		rdev->accel_working = false;
+		dev_err(rdev->dev, "failed initializing IB (%d).\n", r);
 		return r;
 	}
-
 	return 0;
 }
 
 int r300_resume(struct radeon_device *rdev)
 {
-	int r;
-
 	/* Make sur GART are not working */
 	if (rdev->flags & RADEON_IS_PCIE)
 		rv370_pcie_gart_disable(rdev);
@@ -1451,18 +1433,11 @@ int r300_resume(struct radeon_device *rdev)
 	r300_clock_startup(rdev);
 	/* Initialize surface registers */
 	radeon_surface_init(rdev);
-
-	rdev->accel_working = true;
-	r = r300_startup(rdev);
-	if (r) {
-		rdev->accel_working = false;
-	}
-	return r;
+	return r300_startup(rdev);
 }
 
 int r300_suspend(struct radeon_device *rdev)
 {
-	radeon_ib_pool_suspend(rdev);
 	r100_cp_disable(rdev);
 	radeon_wb_disable(rdev);
 	r100_irq_disable(rdev);
@@ -1563,14 +1538,7 @@ int r300_init(struct radeon_device *rdev)
 			return r;
 	}
 	r300_set_reg_safe(rdev);
-
-	r = radeon_ib_pool_init(rdev);
 	rdev->accel_working = true;
-	if (r) {
-		dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
-		rdev->accel_working = false;
-	}
-
 	r = r300_startup(rdev);
 	if (r) {
 		/* Somethings want wront with the accel init stop accel */
diff --git a/drivers/gpu/drm/radeon/r300_cmdbuf.c b/drivers/gpu/drm/radeon/r300_cmdbuf.c
index 9aa02be..a12f373 100644
--- a/drivers/gpu/drm/radeon/r300_cmdbuf.c
+++ b/drivers/gpu/drm/radeon/r300_cmdbuf.c
@@ -791,7 +791,7 @@ static __inline__ int r300_emit_packet3(drm_radeon_private_t *dev_priv,
 /**
  * Emit the sequence to pacify R300.
  */
-static void r300_pacify(drm_radeon_private_t *dev_priv)
+static __inline__ void r300_pacify(drm_radeon_private_t *dev_priv)
 {
 	uint32_t cache_z, cache_3d, cache_2d;
 	RING_LOCALS;
diff --git a/drivers/gpu/drm/radeon/r420.c b/drivers/gpu/drm/radeon/r420.c
index f3fcaac..417fab8 100644
--- a/drivers/gpu/drm/radeon/r420.c
+++ b/drivers/gpu/drm/radeon/r420.c
@@ -199,8 +199,6 @@ static void r420_clock_resume(struct radeon_device *rdev)
 
 static void r420_cp_errata_init(struct radeon_device *rdev)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
-
 	/* RV410 and R420 can lock up if CP DMA to host memory happens
 	 * while the 2D engine is busy.
 	 *
@@ -208,24 +206,22 @@ static void r420_cp_errata_init(struct radeon_device *rdev)
 	 * of the CP init, apparently.
 	 */
 	radeon_scratch_get(rdev, &rdev->config.r300.resync_scratch);
-	radeon_ring_lock(rdev, ring, 8);
-	radeon_ring_write(ring, PACKET0(R300_CP_RESYNC_ADDR, 1));
-	radeon_ring_write(ring, rdev->config.r300.resync_scratch);
-	radeon_ring_write(ring, 0xDEADBEEF);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_lock(rdev, 8);
+	radeon_ring_write(rdev, PACKET0(R300_CP_RESYNC_ADDR, 1));
+	radeon_ring_write(rdev, rdev->config.r300.resync_scratch);
+	radeon_ring_write(rdev, 0xDEADBEEF);
+	radeon_ring_unlock_commit(rdev);
 }
 
 static void r420_cp_errata_fini(struct radeon_device *rdev)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
-
 	/* Catch the RESYNC we dispatched all the way back,
 	 * at the very beginning of the CP init.
 	 */
-	radeon_ring_lock(rdev, ring, 8);
-	radeon_ring_write(ring, PACKET0(R300_RB3D_DSTCACHE_CTLSTAT, 0));
-	radeon_ring_write(ring, R300_RB3D_DC_FINISH);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_lock(rdev, 8);
+	radeon_ring_write(rdev, PACKET0(R300_RB3D_DSTCACHE_CTLSTAT, 0));
+	radeon_ring_write(rdev, R300_RB3D_DC_FINISH);
+	radeon_ring_unlock_commit(rdev);
 	radeon_scratch_free(rdev, rdev->config.r300.resync_scratch);
 }
 
@@ -258,12 +254,6 @@ static int r420_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX);
-	if (r) {
-		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
-		return r;
-	}
-
 	/* Enable IRQ */
 	r100_irq_set(rdev);
 	rdev->config.r300.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL);
@@ -274,25 +264,16 @@ static int r420_startup(struct radeon_device *rdev)
 		return r;
 	}
 	r420_cp_errata_init(rdev);
-
-	r = radeon_ib_pool_start(rdev);
-	if (r)
-		return r;
-
-	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
+	r = r100_ib_init(rdev);
 	if (r) {
-		dev_err(rdev->dev, "failed testing IB (%d).\n", r);
-		rdev->accel_working = false;
+		dev_err(rdev->dev, "failed initializing IB (%d).\n", r);
 		return r;
 	}
-
 	return 0;
 }
 
 int r420_resume(struct radeon_device *rdev)
 {
-	int r;
-
 	/* Make sur GART are not working */
 	if (rdev->flags & RADEON_IS_PCIE)
 		rv370_pcie_gart_disable(rdev);
@@ -316,18 +297,11 @@ int r420_resume(struct radeon_device *rdev)
 	r420_clock_resume(rdev);
 	/* Initialize surface registers */
 	radeon_surface_init(rdev);
-
-	rdev->accel_working = true;
-	r = r420_startup(rdev);
-	if (r) {
-		rdev->accel_working = false;
-	}
-	return r;
+	return r420_startup(rdev);
 }
 
 int r420_suspend(struct radeon_device *rdev)
 {
-	radeon_ib_pool_suspend(rdev);
 	r420_cp_errata_fini(rdev);
 	r100_cp_disable(rdev);
 	radeon_wb_disable(rdev);
@@ -440,14 +414,7 @@ int r420_init(struct radeon_device *rdev)
 			return r;
 	}
 	r420_set_reg_safe(rdev);
-
-	r = radeon_ib_pool_init(rdev);
 	rdev->accel_working = true;
-	if (r) {
-		dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
-		rdev->accel_working = false;
-	}
-
 	r = r420_startup(rdev);
 	if (r) {
 		/* Somethings want wront with the accel init stop accel */
diff --git a/drivers/gpu/drm/radeon/r500_reg.h b/drivers/gpu/drm/radeon/r500_reg.h
index ec576aa..fc43705 100644
--- a/drivers/gpu/drm/radeon/r500_reg.h
+++ b/drivers/gpu/drm/radeon/r500_reg.h
@@ -351,8 +351,6 @@
 #define AVIVO_D1CRTC_BLANK_CONTROL                              0x6084
 #define AVIVO_D1CRTC_INTERLACE_CONTROL                          0x6088
 #define AVIVO_D1CRTC_INTERLACE_STATUS                           0x608c
-#define AVIVO_D1CRTC_STATUS                                     0x609c
-#       define AVIVO_D1CRTC_V_BLANK                             (1 << 0)
 #define AVIVO_D1CRTC_STATUS_POSITION                            0x60a0
 #define AVIVO_D1CRTC_FRAME_COUNT                                0x60a4
 #define AVIVO_D1CRTC_STEREO_CONTROL                             0x60c4
@@ -575,7 +573,6 @@
 
 #define AVIVO_TMDSA_CNTL                    0x7880
 #   define AVIVO_TMDSA_CNTL_ENABLE               (1 << 0)
-#   define AVIVO_TMDSA_CNTL_HDMI_EN              (1 << 2)
 #   define AVIVO_TMDSA_CNTL_HPD_MASK             (1 << 4)
 #   define AVIVO_TMDSA_CNTL_HPD_SELECT           (1 << 8)
 #   define AVIVO_TMDSA_CNTL_SYNC_PHASE           (1 << 12)
@@ -636,7 +633,6 @@
 
 #define AVIVO_LVTMA_CNTL					0x7a80
 #   define AVIVO_LVTMA_CNTL_ENABLE               (1 << 0)
-#   define AVIVO_LVTMA_CNTL_HDMI_EN              (1 << 2)
 #   define AVIVO_LVTMA_CNTL_HPD_MASK             (1 << 4)
 #   define AVIVO_LVTMA_CNTL_HPD_SELECT           (1 << 8)
 #   define AVIVO_LVTMA_CNTL_SYNC_PHASE           (1 << 12)
diff --git a/drivers/gpu/drm/radeon/r520.c b/drivers/gpu/drm/radeon/r520.c
index ebcc15b..3081d07 100644
--- a/drivers/gpu/drm/radeon/r520.c
+++ b/drivers/gpu/drm/radeon/r520.c
@@ -33,7 +33,7 @@
 
 /* This files gather functions specifics to: r520,rv530,rv560,rv570,r580 */
 
-int r520_mc_wait_for_idle(struct radeon_device *rdev)
+static int r520_mc_wait_for_idle(struct radeon_device *rdev)
 {
 	unsigned i;
 	uint32_t tmp;
@@ -187,12 +187,6 @@ static int r520_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX);
-	if (r) {
-		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
-		return r;
-	}
-
 	/* Enable IRQ */
 	rs600_irq_set(rdev);
 	rdev->config.r300.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL);
@@ -202,15 +196,9 @@ static int r520_startup(struct radeon_device *rdev)
 		dev_err(rdev->dev, "failed initializing CP (%d).\n", r);
 		return r;
 	}
-
-	r = radeon_ib_pool_start(rdev);
-	if (r)
-		return r;
-
-	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
+	r = r100_ib_init(rdev);
 	if (r) {
-		dev_err(rdev->dev, "failed testing IB (%d).\n", r);
-		rdev->accel_working = false;
+		dev_err(rdev->dev, "failed initializing IB (%d).\n", r);
 		return r;
 	}
 	return 0;
@@ -218,8 +206,6 @@ static int r520_startup(struct radeon_device *rdev)
 
 int r520_resume(struct radeon_device *rdev)
 {
-	int r;
-
 	/* Make sur GART are not working */
 	if (rdev->flags & RADEON_IS_PCIE)
 		rv370_pcie_gart_disable(rdev);
@@ -237,13 +223,7 @@ int r520_resume(struct radeon_device *rdev)
 	rv515_clock_startup(rdev);
 	/* Initialize surface registers */
 	radeon_surface_init(rdev);
-
-	rdev->accel_working = true;
-	r = r520_startup(rdev);
-	if (r) {
-		rdev->accel_working = false;
-	}
-	return r;
+	return r520_startup(rdev);
 }
 
 int r520_init(struct radeon_device *rdev)
@@ -312,14 +292,7 @@ int r520_init(struct radeon_device *rdev)
 	if (r)
 		return r;
 	rv515_set_safe_registers(rdev);
-
-	r = radeon_ib_pool_init(rdev);
 	rdev->accel_working = true;
-	if (r) {
-		dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
-		rdev->accel_working = false;
-	}
-
 	r = r520_startup(rdev);
 	if (r) {
 		/* Somethings want wront with the accel init stop accel */
diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c
index 5eb2382..1a4ed43 100644
--- a/drivers/gpu/drm/radeon/r600.c
+++ b/drivers/gpu/drm/radeon/r600.c
@@ -29,7 +29,6 @@
 #include <linux/seq_file.h>
 #include <linux/firmware.h>
 #include <linux/platform_device.h>
-#include <linux/module.h>
 #include "drmP.h"
 #include "radeon_drm.h"
 #include "radeon.h"
@@ -288,6 +287,24 @@ void r600_pm_get_dynpm_state(struct radeon_device *rdev)
 		  pcie_lanes);
 }
 
+static int r600_pm_get_type_index(struct radeon_device *rdev,
+				  enum radeon_pm_state_type ps_type,
+				  int instance)
+{
+	int i;
+	int found_instance = -1;
+
+	for (i = 0; i < rdev->pm.num_power_states; i++) {
+		if (rdev->pm.power_state[i].type == ps_type) {
+			found_instance++;
+			if (found_instance == instance)
+				return i;
+		}
+	}
+	/* return default if no match */
+	return rdev->pm.default_power_state_index;
+}
+
 void rs780_pm_init_profile(struct radeon_device *rdev)
 {
 	if (rdev->pm.num_power_states == 2) {
@@ -403,8 +420,6 @@ void rs780_pm_init_profile(struct radeon_device *rdev)
 
 void r600_pm_init_profile(struct radeon_device *rdev)
 {
-	int idx;
-
 	if (rdev->family == CHIP_R600) {
 		/* XXX */
 		/* default */
@@ -486,43 +501,81 @@ void r600_pm_init_profile(struct radeon_device *rdev)
 			rdev->pm.profiles[PM_PROFILE_DEFAULT_IDX].dpms_off_cm_idx = 0;
 			rdev->pm.profiles[PM_PROFILE_DEFAULT_IDX].dpms_on_cm_idx = 2;
 			/* low sh */
-			if (rdev->flags & RADEON_IS_MOBILITY)
-				idx = radeon_pm_get_type_index(rdev, POWER_STATE_TYPE_BATTERY, 0);
-			else
-				idx = radeon_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 0);
-			rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_off_ps_idx = idx;
-			rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_ps_idx = idx;
-			rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_off_cm_idx = 0;
-			rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_cm_idx = 0;
+			if (rdev->flags & RADEON_IS_MOBILITY) {
+				rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_off_ps_idx =
+					r600_pm_get_type_index(rdev, POWER_STATE_TYPE_BATTERY, 0);
+				rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_ps_idx =
+					r600_pm_get_type_index(rdev, POWER_STATE_TYPE_BATTERY, 0);
+				rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_off_cm_idx = 0;
+				rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_cm_idx = 0;
+			} else {
+				rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_off_ps_idx =
+					r600_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 0);
+				rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_ps_idx =
+					r600_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 0);
+				rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_off_cm_idx = 0;
+				rdev->pm.profiles[PM_PROFILE_LOW_SH_IDX].dpms_on_cm_idx = 0;
+			}
 			/* mid sh */
-			rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_ps_idx = idx;
-			rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_ps_idx = idx;
-			rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_cm_idx = 0;
-			rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_cm_idx = 1;
+			if (rdev->flags & RADEON_IS_MOBILITY) {
+				rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_ps_idx =
+					r600_pm_get_type_index(rdev, POWER_STATE_TYPE_BATTERY, 0);
+				rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_ps_idx =
+					r600_pm_get_type_index(rdev, POWER_STATE_TYPE_BATTERY, 0);
+				rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_cm_idx = 0;
+				rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_cm_idx = 1;
+			} else {
+				rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_ps_idx =
+					r600_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 0);
+				rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_ps_idx =
+					r600_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 0);
+				rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_off_cm_idx = 0;
+				rdev->pm.profiles[PM_PROFILE_MID_SH_IDX].dpms_on_cm_idx = 1;
+			}
 			/* high sh */
-			idx = radeon_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 0);
-			rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_off_ps_idx = idx;
-			rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_on_ps_idx = idx;
+			rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_off_ps_idx =
+				r600_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 0);
+			rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_on_ps_idx =
+				r600_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 0);
 			rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_off_cm_idx = 0;
 			rdev->pm.profiles[PM_PROFILE_HIGH_SH_IDX].dpms_on_cm_idx = 2;
 			/* low mh */
-			if (rdev->flags & RADEON_IS_MOBILITY)
-				idx = radeon_pm_get_type_index(rdev, POWER_STATE_TYPE_BATTERY, 1);
-			else
-				idx = radeon_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 1);
-			rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_off_ps_idx = idx;
-			rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_ps_idx = idx;
-			rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_off_cm_idx = 0;
-			rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_cm_idx = 0;
+			if (rdev->flags & RADEON_IS_MOBILITY) {
+				rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_off_ps_idx =
+					r600_pm_get_type_index(rdev, POWER_STATE_TYPE_BATTERY, 1);
+				rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_ps_idx =
+					r600_pm_get_type_index(rdev, POWER_STATE_TYPE_BATTERY, 1);
+				rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_off_cm_idx = 0;
+				rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_cm_idx = 0;
+			} else {
+				rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_off_ps_idx =
+					r600_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 1);
+				rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_ps_idx =
+					r600_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 1);
+				rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_off_cm_idx = 0;
+				rdev->pm.profiles[PM_PROFILE_LOW_MH_IDX].dpms_on_cm_idx = 0;
+			}
 			/* mid mh */
-			rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_ps_idx = idx;
-			rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_ps_idx = idx;
-			rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_cm_idx = 0;
-			rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_cm_idx = 1;
+			if (rdev->flags & RADEON_IS_MOBILITY) {
+				rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_ps_idx =
+					r600_pm_get_type_index(rdev, POWER_STATE_TYPE_BATTERY, 1);
+				rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_ps_idx =
+					r600_pm_get_type_index(rdev, POWER_STATE_TYPE_BATTERY, 1);
+				rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_cm_idx = 0;
+				rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_cm_idx = 1;
+			} else {
+				rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_ps_idx =
+					r600_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 1);
+				rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_ps_idx =
+					r600_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 1);
+				rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_off_cm_idx = 0;
+				rdev->pm.profiles[PM_PROFILE_MID_MH_IDX].dpms_on_cm_idx = 1;
+			}
 			/* high mh */
-			idx = radeon_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 1);
-			rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_off_ps_idx = idx;
-			rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_on_ps_idx = idx;
+			rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_off_ps_idx =
+				r600_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 1);
+			rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_on_ps_idx =
+				r600_pm_get_type_index(rdev, POWER_STATE_TYPE_PERFORMANCE, 1);
 			rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_off_cm_idx = 0;
 			rdev->pm.profiles[PM_PROFILE_HIGH_MH_IDX].dpms_on_cm_idx = 2;
 		}
@@ -842,7 +895,7 @@ void r600_pcie_gart_tlb_flush(struct radeon_device *rdev)
 	/* flush hdp cache so updates hit vram */
 	if ((rdev->family >= CHIP_RV770) && (rdev->family <= CHIP_RV740) &&
 	    !(rdev->flags & RADEON_IS_AGP)) {
-		void __iomem *ptr = (void *)rdev->gart.ptr;
+		void __iomem *ptr = (void *)rdev->gart.table.vram.ptr;
 		u32 tmp;
 
 		/* r7xx hw bug.  write to HDP_DEBUG1 followed by fb read
@@ -877,7 +930,7 @@ int r600_pcie_gart_init(struct radeon_device *rdev)
 {
 	int r;
 
-	if (rdev->gart.robj) {
+	if (rdev->gart.table.vram.robj) {
 		WARN(1, "R600 PCIE GART already initialized\n");
 		return 0;
 	}
@@ -894,7 +947,7 @@ int r600_pcie_gart_enable(struct radeon_device *rdev)
 	u32 tmp;
 	int r, i;
 
-	if (rdev->gart.robj == NULL) {
+	if (rdev->gart.table.vram.robj == NULL) {
 		dev_err(rdev->dev, "No VRAM object for PCIE GART.\n");
 		return -EINVAL;
 	}
@@ -939,9 +992,6 @@ int r600_pcie_gart_enable(struct radeon_device *rdev)
 		WREG32(VM_CONTEXT0_CNTL + (i * 4), 0);
 
 	r600_pcie_gart_tlb_flush(rdev);
-	DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n",
-		 (unsigned)(rdev->mc.gtt_size >> 20),
-		 (unsigned long long)rdev->gart.table_addr);
 	rdev->gart.ready = true;
 	return 0;
 }
@@ -949,7 +999,7 @@ int r600_pcie_gart_enable(struct radeon_device *rdev)
 void r600_pcie_gart_disable(struct radeon_device *rdev)
 {
 	u32 tmp;
-	int i;
+	int i, r;
 
 	/* Disable all tables */
 	for (i = 0; i < 7; i++)
@@ -976,7 +1026,14 @@ void r600_pcie_gart_disable(struct radeon_device *rdev)
 	WREG32(MC_VM_L1_TLB_MCB_WR_SYS_CNTL, tmp);
 	WREG32(MC_VM_L1_TLB_MCB_RD_HDP_CNTL, tmp);
 	WREG32(MC_VM_L1_TLB_MCB_WR_HDP_CNTL, tmp);
-	radeon_gart_table_vram_unpin(rdev);
+	if (rdev->gart.table.vram.robj) {
+		r = radeon_bo_reserve(rdev->gart.table.vram.robj, false);
+		if (likely(r == 0)) {
+			radeon_bo_kunmap(rdev->gart.table.vram.robj);
+			radeon_bo_unpin(rdev->gart.table.vram.robj);
+			radeon_bo_unreserve(rdev->gart.table.vram.robj);
+		}
+	}
 }
 
 void r600_pcie_gart_fini(struct radeon_device *rdev)
@@ -1076,7 +1133,7 @@ static void r600_mc_program(struct radeon_device *rdev)
 		WREG32(MC_VM_SYSTEM_APERTURE_LOW_ADDR, rdev->mc.vram_start >> 12);
 		WREG32(MC_VM_SYSTEM_APERTURE_HIGH_ADDR, rdev->mc.vram_end >> 12);
 	}
-	WREG32(MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR, rdev->vram_scratch.gpu_addr >> 12);
+	WREG32(MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR, 0);
 	tmp = ((rdev->mc.vram_end >> 24) & 0xFFFF) << 16;
 	tmp |= ((rdev->mc.vram_start >> 24) & 0xFFFF);
 	WREG32(MC_VM_FB_LOCATION, tmp);
@@ -1215,53 +1272,6 @@ int r600_mc_init(struct radeon_device *rdev)
 	return 0;
 }
 
-int r600_vram_scratch_init(struct radeon_device *rdev)
-{
-	int r;
-
-	if (rdev->vram_scratch.robj == NULL) {
-		r = radeon_bo_create(rdev, RADEON_GPU_PAGE_SIZE,
-				     PAGE_SIZE, true, RADEON_GEM_DOMAIN_VRAM,
-				     &rdev->vram_scratch.robj);
-		if (r) {
-			return r;
-		}
-	}
-
-	r = radeon_bo_reserve(rdev->vram_scratch.robj, false);
-	if (unlikely(r != 0))
-		return r;
-	r = radeon_bo_pin(rdev->vram_scratch.robj,
-			  RADEON_GEM_DOMAIN_VRAM, &rdev->vram_scratch.gpu_addr);
-	if (r) {
-		radeon_bo_unreserve(rdev->vram_scratch.robj);
-		return r;
-	}
-	r = radeon_bo_kmap(rdev->vram_scratch.robj,
-				(void **)&rdev->vram_scratch.ptr);
-	if (r)
-		radeon_bo_unpin(rdev->vram_scratch.robj);
-	radeon_bo_unreserve(rdev->vram_scratch.robj);
-
-	return r;
-}
-
-void r600_vram_scratch_fini(struct radeon_device *rdev)
-{
-	int r;
-
-	if (rdev->vram_scratch.robj == NULL) {
-		return;
-	}
-	r = radeon_bo_reserve(rdev->vram_scratch.robj, false);
-	if (likely(r == 0)) {
-		radeon_bo_kunmap(rdev->vram_scratch.robj);
-		radeon_bo_unpin(rdev->vram_scratch.robj);
-		radeon_bo_unreserve(rdev->vram_scratch.robj);
-	}
-	radeon_bo_unref(&rdev->vram_scratch.robj);
-}
-
 /* We doesn't check that the GPU really needs a reset we simply do the
  * reset, it's up to the caller to determine if the GPU needs one. We
  * might add an helper function to check that.
@@ -1344,7 +1354,7 @@ int r600_gpu_soft_reset(struct radeon_device *rdev)
 	return 0;
 }
 
-bool r600_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
+bool r600_gpu_is_lockup(struct radeon_device *rdev)
 {
 	u32 srbm_status;
 	u32 grbm_status;
@@ -1361,19 +1371,19 @@ bool r600_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *ring)
 	grbm_status = RREG32(R_008010_GRBM_STATUS);
 	grbm_status2 = RREG32(R_008014_GRBM_STATUS2);
 	if (!G_008010_GUI_ACTIVE(grbm_status)) {
-		r100_gpu_lockup_update(lockup, ring);
+		r100_gpu_lockup_update(lockup, &rdev->cp);
 		return false;
 	}
 	/* force CP activities */
-	r = radeon_ring_lock(rdev, ring, 2);
+	r = radeon_ring_lock(rdev, 2);
 	if (!r) {
 		/* PACKET2 NOP */
-		radeon_ring_write(ring, 0x80000000);
-		radeon_ring_write(ring, 0x80000000);
-		radeon_ring_unlock_commit(rdev, ring);
+		radeon_ring_write(rdev, 0x80000000);
+		radeon_ring_write(rdev, 0x80000000);
+		radeon_ring_unlock_commit(rdev);
 	}
-	ring->rptr = RREG32(ring->rptr_reg);
-	return r100_gpu_cp_is_lockup(rdev, lockup, ring);
+	rdev->cp.rptr = RREG32(R600_CP_RB_RPTR);
+	return r100_gpu_cp_is_lockup(rdev, lockup, &rdev->cp);
 }
 
 int r600_asic_reset(struct radeon_device *rdev)
@@ -1651,7 +1661,6 @@ void r600_gpu_init(struct radeon_device *rdev)
 									       R6XX_MAX_BACKENDS_MASK) >> 16)),
 							(cc_rb_backend_disable >> 16));
 	rdev->config.r600.tile_config = tiling_config;
-	rdev->config.r600.backend_map = backend_map;
 	tiling_config |= BACKEND_MAP(backend_map);
 	WREG32(GB_TILING_CONFIG, tiling_config);
 	WREG32(DCP_TILING_CONFIG, tiling_config & 0xffff);
@@ -2144,28 +2153,27 @@ static int r600_cp_load_microcode(struct radeon_device *rdev)
 
 int r600_cp_start(struct radeon_device *rdev)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	int r;
 	uint32_t cp_me;
 
-	r = radeon_ring_lock(rdev, ring, 7);
+	r = radeon_ring_lock(rdev, 7);
 	if (r) {
 		DRM_ERROR("radeon: cp failed to lock ring (%d).\n", r);
 		return r;
 	}
-	radeon_ring_write(ring, PACKET3(PACKET3_ME_INITIALIZE, 5));
-	radeon_ring_write(ring, 0x1);
+	radeon_ring_write(rdev, PACKET3(PACKET3_ME_INITIALIZE, 5));
+	radeon_ring_write(rdev, 0x1);
 	if (rdev->family >= CHIP_RV770) {
-		radeon_ring_write(ring, 0x0);
-		radeon_ring_write(ring, rdev->config.rv770.max_hw_contexts - 1);
+		radeon_ring_write(rdev, 0x0);
+		radeon_ring_write(rdev, rdev->config.rv770.max_hw_contexts - 1);
 	} else {
-		radeon_ring_write(ring, 0x3);
-		radeon_ring_write(ring, rdev->config.r600.max_hw_contexts - 1);
+		radeon_ring_write(rdev, 0x3);
+		radeon_ring_write(rdev, rdev->config.r600.max_hw_contexts - 1);
 	}
-	radeon_ring_write(ring, PACKET3_ME_INITIALIZE_DEVICE_ID(1));
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, 0);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_write(rdev, PACKET3_ME_INITIALIZE_DEVICE_ID(1));
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_unlock_commit(rdev);
 
 	cp_me = 0xff;
 	WREG32(R_0086D8_CP_ME_CNTL, cp_me);
@@ -2174,7 +2182,6 @@ int r600_cp_start(struct radeon_device *rdev)
 
 int r600_cp_resume(struct radeon_device *rdev)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	u32 tmp;
 	u32 rb_bufsz;
 	int r;
@@ -2186,13 +2193,13 @@ int r600_cp_resume(struct radeon_device *rdev)
 	WREG32(GRBM_SOFT_RESET, 0);
 
 	/* Set ring buffer size */
-	rb_bufsz = drm_order(ring->ring_size / 8);
+	rb_bufsz = drm_order(rdev->cp.ring_size / 8);
 	tmp = (drm_order(RADEON_GPU_PAGE_SIZE/8) << 8) | rb_bufsz;
 #ifdef __BIG_ENDIAN
 	tmp |= BUF_SWAP_32BIT;
 #endif
 	WREG32(CP_RB_CNTL, tmp);
-	WREG32(CP_SEM_WAIT_TIMER, 0x0);
+	WREG32(CP_SEM_WAIT_TIMER, 0x4);
 
 	/* Set the write pointer delay */
 	WREG32(CP_RB_WPTR_DELAY, 0);
@@ -2200,11 +2207,14 @@ int r600_cp_resume(struct radeon_device *rdev)
 	/* Initialize the ring buffer's read and write pointers */
 	WREG32(CP_RB_CNTL, tmp | RB_RPTR_WR_ENA);
 	WREG32(CP_RB_RPTR_WR, 0);
-	ring->wptr = 0;
-	WREG32(CP_RB_WPTR, ring->wptr);
+	rdev->cp.wptr = 0;
+	WREG32(CP_RB_WPTR, rdev->cp.wptr);
 
 	/* set the wb address whether it's enabled or not */
 	WREG32(CP_RB_RPTR_ADDR,
+#ifdef __BIG_ENDIAN
+	       RB_RPTR_SWAP(2) |
+#endif
 	       ((rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) & 0xFFFFFFFC));
 	WREG32(CP_RB_RPTR_ADDR_HI, upper_32_bits(rdev->wb.gpu_addr + RADEON_WB_CP_RPTR_OFFSET) & 0xFF);
 	WREG32(SCRATCH_ADDR, ((rdev->wb.gpu_addr + RADEON_WB_SCRATCH_OFFSET) >> 8) & 0xFFFFFFFF);
@@ -2219,36 +2229,42 @@ int r600_cp_resume(struct radeon_device *rdev)
 	mdelay(1);
 	WREG32(CP_RB_CNTL, tmp);
 
-	WREG32(CP_RB_BASE, ring->gpu_addr >> 8);
+	WREG32(CP_RB_BASE, rdev->cp.gpu_addr >> 8);
 	WREG32(CP_DEBUG, (1 << 27) | (1 << 28));
 
-	ring->rptr = RREG32(CP_RB_RPTR);
+	rdev->cp.rptr = RREG32(CP_RB_RPTR);
 
 	r600_cp_start(rdev);
-	ring->ready = true;
-	r = radeon_ring_test(rdev, RADEON_RING_TYPE_GFX_INDEX, ring);
+	rdev->cp.ready = true;
+	r = radeon_ring_test(rdev);
 	if (r) {
-		ring->ready = false;
+		rdev->cp.ready = false;
 		return r;
 	}
 	return 0;
 }
 
-void r600_ring_init(struct radeon_device *rdev, struct radeon_ring *ring, unsigned ring_size)
+void r600_cp_commit(struct radeon_device *rdev)
+{
+	WREG32(CP_RB_WPTR, rdev->cp.wptr);
+	(void)RREG32(CP_RB_WPTR);
+}
+
+void r600_ring_init(struct radeon_device *rdev, unsigned ring_size)
 {
 	u32 rb_bufsz;
 
 	/* Align ring size */
 	rb_bufsz = drm_order(ring_size / 8);
 	ring_size = (1 << (rb_bufsz + 1)) * 4;
-	ring->ring_size = ring_size;
-	ring->align_mask = 16 - 1;
+	rdev->cp.ring_size = ring_size;
+	rdev->cp.align_mask = 16 - 1;
 }
 
 void r600_cp_fini(struct radeon_device *rdev)
 {
 	r600_cp_stop(rdev);
-	radeon_ring_fini(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
+	radeon_ring_fini(rdev);
 }
 
 
@@ -2267,11 +2283,11 @@ void r600_scratch_init(struct radeon_device *rdev)
 	}
 }
 
-int r600_ring_test(struct radeon_device *rdev, struct radeon_ring *ring)
+int r600_ring_test(struct radeon_device *rdev)
 {
 	uint32_t scratch;
 	uint32_t tmp = 0;
-	unsigned i, ridx = radeon_ring_index(rdev, ring);
+	unsigned i;
 	int r;
 
 	r = radeon_scratch_get(rdev, &scratch);
@@ -2280,16 +2296,16 @@ int r600_ring_test(struct radeon_device *rdev, struct radeon_ring *ring)
 		return r;
 	}
 	WREG32(scratch, 0xCAFEDEAD);
-	r = radeon_ring_lock(rdev, ring, 3);
+	r = radeon_ring_lock(rdev, 3);
 	if (r) {
-		DRM_ERROR("radeon: cp failed to lock ring %d (%d).\n", ridx, r);
+		DRM_ERROR("radeon: cp failed to lock ring (%d).\n", r);
 		radeon_scratch_free(rdev, scratch);
 		return r;
 	}
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
-	radeon_ring_write(ring, ((scratch - PACKET3_SET_CONFIG_REG_OFFSET) >> 2));
-	radeon_ring_write(ring, 0xDEADBEEF);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 1));
+	radeon_ring_write(rdev, ((scratch - PACKET3_SET_CONFIG_REG_OFFSET) >> 2));
+	radeon_ring_write(rdev, 0xDEADBEEF);
+	radeon_ring_unlock_commit(rdev);
 	for (i = 0; i < rdev->usec_timeout; i++) {
 		tmp = RREG32(scratch);
 		if (tmp == 0xDEADBEEF)
@@ -2297,10 +2313,10 @@ int r600_ring_test(struct radeon_device *rdev, struct radeon_ring *ring)
 		DRM_UDELAY(1);
 	}
 	if (i < rdev->usec_timeout) {
-		DRM_INFO("ring test on %d succeeded in %d usecs\n", ridx, i);
+		DRM_INFO("ring test succeeded in %d usecs\n", i);
 	} else {
-		DRM_ERROR("radeon: ring %d test failed (scratch(0x%04X)=0x%08X)\n",
-			  ridx, scratch, tmp);
+		DRM_ERROR("radeon: ring test failed (scratch(0x%04X)=0x%08X)\n",
+			  scratch, tmp);
 		r = -EINVAL;
 	}
 	radeon_scratch_free(rdev, scratch);
@@ -2310,66 +2326,33 @@ int r600_ring_test(struct radeon_device *rdev, struct radeon_ring *ring)
 void r600_fence_ring_emit(struct radeon_device *rdev,
 			  struct radeon_fence *fence)
 {
-	struct radeon_ring *ring = &rdev->ring[fence->ring];
-
 	if (rdev->wb.use_event) {
-		u64 addr = rdev->fence_drv[fence->ring].gpu_addr;
-		/* flush read cache over gart */
-		radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3));
-		radeon_ring_write(ring, PACKET3_TC_ACTION_ENA |
-					PACKET3_VC_ACTION_ENA |
-					PACKET3_SH_ACTION_ENA);
-		radeon_ring_write(ring, 0xFFFFFFFF);
-		radeon_ring_write(ring, 0);
-		radeon_ring_write(ring, 10); /* poll interval */
+		u64 addr = rdev->wb.gpu_addr + R600_WB_EVENT_OFFSET +
+			(u64)(rdev->fence_drv.scratch_reg - rdev->scratch.reg_base);
 		/* EVENT_WRITE_EOP - flush caches, send int */
-		radeon_ring_write(ring, PACKET3(PACKET3_EVENT_WRITE_EOP, 4));
-		radeon_ring_write(ring, EVENT_TYPE(CACHE_FLUSH_AND_INV_EVENT_TS) | EVENT_INDEX(5));
-		radeon_ring_write(ring, addr & 0xffffffff);
-		radeon_ring_write(ring, (upper_32_bits(addr) & 0xff) | DATA_SEL(1) | INT_SEL(2));
-		radeon_ring_write(ring, fence->seq);
-		radeon_ring_write(ring, 0);
+		radeon_ring_write(rdev, PACKET3(PACKET3_EVENT_WRITE_EOP, 4));
+		radeon_ring_write(rdev, EVENT_TYPE(CACHE_FLUSH_AND_INV_EVENT_TS) | EVENT_INDEX(5));
+		radeon_ring_write(rdev, addr & 0xffffffff);
+		radeon_ring_write(rdev, (upper_32_bits(addr) & 0xff) | DATA_SEL(1) | INT_SEL(2));
+		radeon_ring_write(rdev, fence->seq);
+		radeon_ring_write(rdev, 0);
 	} else {
-		/* flush read cache over gart */
-		radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3));
-		radeon_ring_write(ring, PACKET3_TC_ACTION_ENA |
-					PACKET3_VC_ACTION_ENA |
-					PACKET3_SH_ACTION_ENA);
-		radeon_ring_write(ring, 0xFFFFFFFF);
-		radeon_ring_write(ring, 0);
-		radeon_ring_write(ring, 10); /* poll interval */
-		radeon_ring_write(ring, PACKET3(PACKET3_EVENT_WRITE, 0));
-		radeon_ring_write(ring, EVENT_TYPE(CACHE_FLUSH_AND_INV_EVENT) | EVENT_INDEX(0));
+		radeon_ring_write(rdev, PACKET3(PACKET3_EVENT_WRITE, 0));
+		radeon_ring_write(rdev, EVENT_TYPE(CACHE_FLUSH_AND_INV_EVENT) | EVENT_INDEX(0));
 		/* wait for 3D idle clean */
-		radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
-		radeon_ring_write(ring, (WAIT_UNTIL - PACKET3_SET_CONFIG_REG_OFFSET) >> 2);
-		radeon_ring_write(ring, WAIT_3D_IDLE_bit | WAIT_3D_IDLECLEAN_bit);
+		radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 1));
+		radeon_ring_write(rdev, (WAIT_UNTIL - PACKET3_SET_CONFIG_REG_OFFSET) >> 2);
+		radeon_ring_write(rdev, WAIT_3D_IDLE_bit | WAIT_3D_IDLECLEAN_bit);
 		/* Emit fence sequence & fire IRQ */
-		radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
-		radeon_ring_write(ring, ((rdev->fence_drv[fence->ring].scratch_reg - PACKET3_SET_CONFIG_REG_OFFSET) >> 2));
-		radeon_ring_write(ring, fence->seq);
+		radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 1));
+		radeon_ring_write(rdev, ((rdev->fence_drv.scratch_reg - PACKET3_SET_CONFIG_REG_OFFSET) >> 2));
+		radeon_ring_write(rdev, fence->seq);
 		/* CP_INTERRUPT packet 3 no longer exists, use packet 0 */
-		radeon_ring_write(ring, PACKET0(CP_INT_STATUS, 0));
-		radeon_ring_write(ring, RB_INT_STAT);
+		radeon_ring_write(rdev, PACKET0(CP_INT_STATUS, 0));
+		radeon_ring_write(rdev, RB_INT_STAT);
 	}
 }
 
-void r600_semaphore_ring_emit(struct radeon_device *rdev,
-			      struct radeon_ring *ring,
-			      struct radeon_semaphore *semaphore,
-			      bool emit_wait)
-{
-	uint64_t addr = semaphore->gpu_addr;
-	unsigned sel = emit_wait ? PACKET3_SEM_SEL_WAIT : PACKET3_SEM_SEL_SIGNAL;
-
-	if (rdev->family < CHIP_CAYMAN)
-		sel |= PACKET3_SEM_WAIT_ON_SIGNAL;
-
-	radeon_ring_write(ring, PACKET3(PACKET3_MEM_SEMAPHORE, 1));
-	radeon_ring_write(ring, addr & 0xffffffff);
-	radeon_ring_write(ring, (upper_32_bits(addr) & 0xff) | sel);
-}
-
 int r600_copy_blit(struct radeon_device *rdev,
 		   uint64_t src_offset,
 		   uint64_t dst_offset,
@@ -2380,33 +2363,19 @@ int r600_copy_blit(struct radeon_device *rdev,
 
 	mutex_lock(&rdev->r600_blit.mutex);
 	rdev->r600_blit.vb_ib = NULL;
-	r = r600_blit_prepare_copy(rdev, num_gpu_pages);
+	r = r600_blit_prepare_copy(rdev, num_gpu_pages * RADEON_GPU_PAGE_SIZE);
 	if (r) {
 		if (rdev->r600_blit.vb_ib)
 			radeon_ib_free(rdev, &rdev->r600_blit.vb_ib);
 		mutex_unlock(&rdev->r600_blit.mutex);
 		return r;
 	}
-	r600_kms_blit_copy(rdev, src_offset, dst_offset, num_gpu_pages);
+	r600_kms_blit_copy(rdev, src_offset, dst_offset, num_gpu_pages * RADEON_GPU_PAGE_SIZE);
 	r600_blit_done_copy(rdev, fence);
 	mutex_unlock(&rdev->r600_blit.mutex);
 	return 0;
 }
 
-void r600_blit_suspend(struct radeon_device *rdev)
-{
-	int r;
-
-	/* unpin shaders bo */
-	if (rdev->r600_blit.shader_obj) {
-		r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false);
-		if (!r) {
-			radeon_bo_unpin(rdev->r600_blit.shader_obj);
-			radeon_bo_unreserve(rdev->r600_blit.shader_obj);
-		}
-	}
-}
-
 int r600_set_surface_reg(struct radeon_device *rdev, int reg,
 			 uint32_t tiling_flags, uint32_t pitch,
 			 uint32_t offset, uint32_t obj_size)
@@ -2422,7 +2391,6 @@ void r600_clear_surface_reg(struct radeon_device *rdev, int reg)
 
 int r600_startup(struct radeon_device *rdev)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	int r;
 
 	/* enable pcie gen2 link */
@@ -2436,10 +2404,6 @@ int r600_startup(struct radeon_device *rdev)
 		}
 	}
 
-	r = r600_vram_scratch_init(rdev);
-	if (r)
-		return r;
-
 	r600_mc_program(rdev);
 	if (rdev->flags & RADEON_IS_AGP) {
 		r600_agp_enable(rdev);
@@ -2452,7 +2416,7 @@ int r600_startup(struct radeon_device *rdev)
 	r = r600_blit_init(rdev);
 	if (r) {
 		r600_blit_fini(rdev);
-		rdev->asic->copy.copy = NULL;
+		rdev->asic->copy = NULL;
 		dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r);
 	}
 
@@ -2461,12 +2425,6 @@ int r600_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX);
-	if (r) {
-		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
-		return r;
-	}
-
 	/* Enable IRQ */
 	r = r600_irq_init(rdev);
 	if (r) {
@@ -2476,10 +2434,7 @@ int r600_startup(struct radeon_device *rdev)
 	}
 	r600_irq_set(rdev);
 
-	r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP_RPTR_OFFSET,
-			     R600_CP_RB_RPTR, R600_CP_RB_WPTR,
-			     0, 0xfffff, RADEON_CP_PACKET2);
-
+	r = radeon_ring_init(rdev, rdev->cp.ring_size);
 	if (r)
 		return r;
 	r = r600_cp_load_microcode(rdev);
@@ -2489,17 +2444,6 @@ int r600_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = radeon_ib_pool_start(rdev);
-	if (r)
-		return r;
-
-	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
-	if (r) {
-		DRM_ERROR("radeon: failed testing IB (%d).\n", r);
-		rdev->accel_working = false;
-		return r;
-	}
-
 	return 0;
 }
 
@@ -2528,11 +2472,15 @@ int r600_resume(struct radeon_device *rdev)
 	/* post card */
 	atom_asic_init(rdev->mode_info.atom_context);
 
-	rdev->accel_working = true;
 	r = r600_startup(rdev);
 	if (r) {
 		DRM_ERROR("r600 startup failed on resume\n");
-		rdev->accel_working = false;
+		return r;
+	}
+
+	r = r600_ib_test(rdev);
+	if (r) {
+		DRM_ERROR("radeon: failed testing IB (%d).\n", r);
 		return r;
 	}
 
@@ -2547,16 +2495,23 @@ int r600_resume(struct radeon_device *rdev)
 
 int r600_suspend(struct radeon_device *rdev)
 {
+	int r;
+
 	r600_audio_fini(rdev);
-	radeon_ib_pool_suspend(rdev);
-	r600_blit_suspend(rdev);
 	/* FIXME: we should wait for ring to be empty */
 	r600_cp_stop(rdev);
-	rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false;
+	rdev->cp.ready = false;
 	r600_irq_suspend(rdev);
 	radeon_wb_disable(rdev);
 	r600_pcie_gart_disable(rdev);
-
+	/* unpin shaders bo */
+	if (rdev->r600_blit.shader_obj) {
+		r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false);
+		if (!r) {
+			radeon_bo_unpin(rdev->r600_blit.shader_obj);
+			radeon_bo_unreserve(rdev->r600_blit.shader_obj);
+		}
+	}
 	return 0;
 }
 
@@ -2626,8 +2581,8 @@ int r600_init(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ring_obj = NULL;
-	r600_ring_init(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX], 1024 * 1024);
+	rdev->cp.ring_obj = NULL;
+	r600_ring_init(rdev, 1024 * 1024);
 
 	rdev->ih.ring_obj = NULL;
 	r600_ih_ring_init(rdev, 64 * 1024);
@@ -2636,24 +2591,30 @@ int r600_init(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = radeon_ib_pool_init(rdev);
 	rdev->accel_working = true;
-	if (r) {
-		dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
-		rdev->accel_working = false;
-	}
-
 	r = r600_startup(rdev);
 	if (r) {
 		dev_err(rdev->dev, "disabling GPU acceleration\n");
 		r600_cp_fini(rdev);
 		r600_irq_fini(rdev);
 		radeon_wb_fini(rdev);
-		r100_ib_fini(rdev);
 		radeon_irq_kms_fini(rdev);
 		r600_pcie_gart_fini(rdev);
 		rdev->accel_working = false;
 	}
+	if (rdev->accel_working) {
+		r = radeon_ib_pool_init(rdev);
+		if (r) {
+			dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
+			rdev->accel_working = false;
+		} else {
+			r = r600_ib_test(rdev);
+			if (r) {
+				dev_err(rdev->dev, "IB test failed (%d).\n", r);
+				rdev->accel_working = false;
+			}
+		}
+	}
 
 	r = r600_audio_init(rdev);
 	if (r)
@@ -2668,13 +2629,11 @@ void r600_fini(struct radeon_device *rdev)
 	r600_cp_fini(rdev);
 	r600_irq_fini(rdev);
 	radeon_wb_fini(rdev);
-	r100_ib_fini(rdev);
+	radeon_ib_pool_fini(rdev);
 	radeon_irq_kms_fini(rdev);
 	r600_pcie_gart_fini(rdev);
-	r600_vram_scratch_fini(rdev);
 	radeon_agp_fini(rdev);
 	radeon_gem_fini(rdev);
-	radeon_semaphore_driver_fini(rdev);
 	radeon_fence_driver_fini(rdev);
 	radeon_bo_fini(rdev);
 	radeon_atombios_fini(rdev);
@@ -2688,27 +2647,24 @@ void r600_fini(struct radeon_device *rdev)
  */
 void r600_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib)
 {
-	struct radeon_ring *ring = &rdev->ring[ib->fence->ring];
-
 	/* FIXME: implement */
-	radeon_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2));
-	radeon_ring_write(ring,
+	radeon_ring_write(rdev, PACKET3(PACKET3_INDIRECT_BUFFER, 2));
+	radeon_ring_write(rdev,
 #ifdef __BIG_ENDIAN
 			  (2 << 0) |
 #endif
 			  (ib->gpu_addr & 0xFFFFFFFC));
-	radeon_ring_write(ring, upper_32_bits(ib->gpu_addr) & 0xFF);
-	radeon_ring_write(ring, ib->length_dw);
+	radeon_ring_write(rdev, upper_32_bits(ib->gpu_addr) & 0xFF);
+	radeon_ring_write(rdev, ib->length_dw);
 }
 
-int r600_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
+int r600_ib_test(struct radeon_device *rdev)
 {
 	struct radeon_ib *ib;
 	uint32_t scratch;
 	uint32_t tmp = 0;
 	unsigned i;
 	int r;
-	int ring_index = radeon_ring_index(rdev, ring);
 
 	r = radeon_scratch_get(rdev, &scratch);
 	if (r) {
@@ -2716,7 +2672,7 @@ int r600_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 		return r;
 	}
 	WREG32(scratch, 0xCAFEDEAD);
-	r = radeon_ib_get(rdev, ring_index, &ib, 256);
+	r = radeon_ib_get(rdev, &ib);
 	if (r) {
 		DRM_ERROR("radeon: failed to get ib (%d).\n", r);
 		return r;
@@ -2724,7 +2680,20 @@ int r600_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 	ib->ptr[0] = PACKET3(PACKET3_SET_CONFIG_REG, 1);
 	ib->ptr[1] = ((scratch - PACKET3_SET_CONFIG_REG_OFFSET) >> 2);
 	ib->ptr[2] = 0xDEADBEEF;
-	ib->length_dw = 3;
+	ib->ptr[3] = PACKET2(0);
+	ib->ptr[4] = PACKET2(0);
+	ib->ptr[5] = PACKET2(0);
+	ib->ptr[6] = PACKET2(0);
+	ib->ptr[7] = PACKET2(0);
+	ib->ptr[8] = PACKET2(0);
+	ib->ptr[9] = PACKET2(0);
+	ib->ptr[10] = PACKET2(0);
+	ib->ptr[11] = PACKET2(0);
+	ib->ptr[12] = PACKET2(0);
+	ib->ptr[13] = PACKET2(0);
+	ib->ptr[14] = PACKET2(0);
+	ib->ptr[15] = PACKET2(0);
+	ib->length_dw = 16;
 	r = radeon_ib_schedule(rdev, ib);
 	if (r) {
 		radeon_scratch_free(rdev, scratch);
@@ -2744,7 +2713,7 @@ int r600_ib_test(struct radeon_device *rdev, struct radeon_ring *ring)
 		DRM_UDELAY(1);
 	}
 	if (i < rdev->usec_timeout) {
-		DRM_INFO("ib test on ring %d succeeded in %u usecs\n", ib->fence->ring, i);
+		DRM_INFO("ib test succeeded in %u usecs\n", i);
 	} else {
 		DRM_ERROR("radeon: ib test failed (scratch(0x%04X)=0x%08X)\n",
 			  scratch, tmp);
@@ -3026,6 +2995,10 @@ int r600_irq_init(struct radeon_device *rdev)
 	/* RPTR_REARM only works if msi's are enabled */
 	if (rdev->msi_enabled)
 		ih_cntl |= RPTR_REARM;
+
+#ifdef __BIG_ENDIAN
+	ih_cntl |= IH_MC_SWAP(IH_MC_SWAP_32BIT);
+#endif
 	WREG32(IH_CNTL, ih_cntl);
 
 	/* force the active interrupt state to all disabled */
@@ -3091,7 +3064,7 @@ int r600_irq_set(struct radeon_device *rdev)
 		hpd3 = RREG32(DC_HOT_PLUG_DETECT3_INT_CONTROL) & ~DC_HPDx_INT_EN;
 	}
 
-	if (rdev->irq.sw_int[RADEON_RING_TYPE_GFX_INDEX]) {
+	if (rdev->irq.sw_int) {
 		DRM_DEBUG("r600_irq_set: sw int\n");
 		cp_int_cntl |= RB_INT_ENABLE;
 		cp_int_cntl |= TIME_STAMP_INT_ENABLE;
@@ -3169,7 +3142,7 @@ int r600_irq_set(struct radeon_device *rdev)
 	return 0;
 }
 
-static void r600_irq_ack(struct radeon_device *rdev)
+static inline void r600_irq_ack(struct radeon_device *rdev)
 {
 	u32 tmp;
 
@@ -3270,7 +3243,7 @@ void r600_irq_disable(struct radeon_device *rdev)
 	r600_disable_interrupt_state(rdev);
 }
 
-static u32 r600_get_ih_wptr(struct radeon_device *rdev)
+static inline u32 r600_get_ih_wptr(struct radeon_device *rdev)
 {
 	u32 wptr, tmp;
 
@@ -3336,10 +3309,6 @@ int r600_irq_process(struct radeon_device *rdev)
 	if (!rdev->ih.enabled || rdev->shutdown)
 		return IRQ_NONE;
 
-	/* No MSIs, need a dummy read to flush PCI DMAs */
-	if (!rdev->msi_enabled)
-		RREG32(IH_RB_WPTR);
-
 	wptr = r600_get_ih_wptr(rdev);
 	rptr = rdev->ih.rptr;
 	DRM_DEBUG("r600_irq_process start: rptr %d, wptr %d\n", rptr, wptr);
@@ -3352,9 +3321,6 @@ int r600_irq_process(struct radeon_device *rdev)
 	}
 
 restart_ih:
-	/* Order reading of wptr vs. reading of IH ring data */
-	rmb();
-
 	/* display interrupts */
 	r600_irq_ack(rdev);
 
@@ -3475,11 +3441,11 @@ restart_ih:
 		case 177: /* CP_INT in IB1 */
 		case 178: /* CP_INT in IB2 */
 			DRM_DEBUG("IH: CP int: 0x%08x\n", src_data);
-			radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX);
+			radeon_fence_process(rdev);
 			break;
 		case 181: /* CP EOP event */
 			DRM_DEBUG("IH: CP EOP\n");
-			radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX);
+			radeon_fence_process(rdev);
 			break;
 		case 233: /* GUI IDLE */
 			DRM_DEBUG("IH: GUI idle\n");
@@ -3512,6 +3478,30 @@ restart_ih:
  */
 #if defined(CONFIG_DEBUG_FS)
 
+static int r600_debugfs_cp_ring_info(struct seq_file *m, void *data)
+{
+	struct drm_info_node *node = (struct drm_info_node *) m->private;
+	struct drm_device *dev = node->minor->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	unsigned count, i, j;
+
+	radeon_ring_free_size(rdev);
+	count = (rdev->cp.ring_size / 4) - rdev->cp.ring_free_dw;
+	seq_printf(m, "CP_STAT 0x%08x\n", RREG32(CP_STAT));
+	seq_printf(m, "CP_RB_WPTR 0x%08x\n", RREG32(CP_RB_WPTR));
+	seq_printf(m, "CP_RB_RPTR 0x%08x\n", RREG32(CP_RB_RPTR));
+	seq_printf(m, "driver's copy of the CP_RB_WPTR 0x%08x\n", rdev->cp.wptr);
+	seq_printf(m, "driver's copy of the CP_RB_RPTR 0x%08x\n", rdev->cp.rptr);
+	seq_printf(m, "%u free dwords in ring\n", rdev->cp.ring_free_dw);
+	seq_printf(m, "%u dwords in ring\n", count);
+	i = rdev->cp.rptr;
+	for (j = 0; j <= count; j++) {
+		seq_printf(m, "r[%04d]=0x%08x\n", i, rdev->cp.ring[i]);
+		i = (i + 1) & rdev->cp.ptr_mask;
+	}
+	return 0;
+}
+
 static int r600_debugfs_mc_info(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
@@ -3525,6 +3515,7 @@ static int r600_debugfs_mc_info(struct seq_file *m, void *data)
 
 static struct drm_info_list r600_mc_info_list[] = {
 	{"r600_mc_info", r600_debugfs_mc_info, 0, NULL},
+	{"r600_ring_info", r600_debugfs_cp_ring_info, 0, NULL},
 };
 #endif
 
diff --git a/drivers/gpu/drm/radeon/r600_audio.c b/drivers/gpu/drm/radeon/r600_audio.c
index ba66f30..846fae5 100644
--- a/drivers/gpu/drm/radeon/r600_audio.c
+++ b/drivers/gpu/drm/radeon/r600_audio.c
@@ -36,7 +36,7 @@
  */
 static int r600_audio_chipset_supported(struct radeon_device *rdev)
 {
-	return (rdev->family >= CHIP_R600 && !ASIC_IS_DCE5(rdev))
+	return (rdev->family >= CHIP_R600 && rdev->family < CHIP_CEDAR)
 		|| rdev->family == CHIP_RS600
 		|| rdev->family == CHIP_RS690
 		|| rdev->family == CHIP_RS740;
@@ -161,18 +161,8 @@ static void r600_audio_update_hdmi(unsigned long param)
  */
 static void r600_audio_engine_enable(struct radeon_device *rdev, bool enable)
 {
-	u32 value = 0;
 	DRM_INFO("%s audio support\n", enable ? "Enabling" : "Disabling");
-	if (ASIC_IS_DCE4(rdev)) {
-		if (enable) {
-			value |= 0x81000000; /* Required to enable audio */
-			value |= 0x0e1000f0; /* fglrx sets that too */
-		}
-		WREG32(EVERGREEN_AUDIO_ENABLE, value);
-	} else {
-		WREG32_P(R600_AUDIO_ENABLE,
-			 enable ? 0x81000000 : 0x0, ~0x81000000);
-	}
+	WREG32_P(R600_AUDIO_ENABLE, enable ? 0x81000000 : 0x0, ~0x81000000);
 	rdev->audio_enabled = enable;
 }
 
@@ -258,33 +248,22 @@ void r600_audio_set_clock(struct drm_encoder *encoder, int clock)
 		return;
 	}
 
-	if (ASIC_IS_DCE4(rdev)) {
-		/* TODO: other PLLs? */
-		WREG32(EVERGREEN_AUDIO_PLL1_MUL, base_rate * 10);
-		WREG32(EVERGREEN_AUDIO_PLL1_DIV, clock * 10);
-		WREG32(EVERGREEN_AUDIO_PLL1_UNK, 0x00000071);
-
-		/* Some magic trigger or src sel? */
-		WREG32_P(0x5ac, 0x01, ~0x77);
-	} else {
-		switch (dig->dig_encoder) {
-		case 0:
-			WREG32(R600_AUDIO_PLL1_MUL, base_rate * 50);
-			WREG32(R600_AUDIO_PLL1_DIV, clock * 100);
-			WREG32(R600_AUDIO_CLK_SRCSEL, 0);
-			break;
-
-		case 1:
-			WREG32(R600_AUDIO_PLL2_MUL, base_rate * 50);
-			WREG32(R600_AUDIO_PLL2_DIV, clock * 100);
-			WREG32(R600_AUDIO_CLK_SRCSEL, 1);
-			break;
-		default:
-			dev_err(rdev->dev,
-				"Unsupported DIG on encoder 0x%02X\n",
-				radeon_encoder->encoder_id);
-			return;
-		}
+	switch (dig->dig_encoder) {
+	case 0:
+		WREG32(R600_AUDIO_PLL1_MUL, base_rate * 50);
+		WREG32(R600_AUDIO_PLL1_DIV, clock * 100);
+		WREG32(R600_AUDIO_CLK_SRCSEL, 0);
+		break;
+
+	case 1:
+		WREG32(R600_AUDIO_PLL2_MUL, base_rate * 50);
+		WREG32(R600_AUDIO_PLL2_DIV, clock * 100);
+		WREG32(R600_AUDIO_CLK_SRCSEL, 1);
+		break;
+	default:
+		dev_err(rdev->dev, "Unsupported DIG on encoder 0x%02X\n",
+			  radeon_encoder->encoder_id);
+		return;
 	}
 }
 
diff --git a/drivers/gpu/drm/radeon/r600_blit.c b/drivers/gpu/drm/radeon/r600_blit.c
index 3c031a4..7f10434 100644
--- a/drivers/gpu/drm/radeon/r600_blit.c
+++ b/drivers/gpu/drm/radeon/r600_blit.c
@@ -41,7 +41,7 @@
 #define COLOR_5_6_5           0x8
 #define COLOR_8_8_8_8         0x1a
 
-static void
+static inline void
 set_render_target(drm_radeon_private_t *dev_priv, int format, int w, int h, u64 gpu_addr)
 {
 	u32 cb_color_info;
@@ -99,7 +99,7 @@ set_render_target(drm_radeon_private_t *dev_priv, int format, int w, int h, u64
 	ADVANCE_RING();
 }
 
-static void
+static inline void
 cp_set_surface_sync(drm_radeon_private_t *dev_priv,
 		    u32 sync_type, u32 size, u64 mc_addr)
 {
@@ -121,7 +121,7 @@ cp_set_surface_sync(drm_radeon_private_t *dev_priv,
 	ADVANCE_RING();
 }
 
-static void
+static inline void
 set_shaders(struct drm_device *dev)
 {
 	drm_radeon_private_t *dev_priv = dev->dev_private;
@@ -184,7 +184,7 @@ set_shaders(struct drm_device *dev)
 			    R600_SH_ACTION_ENA, 512, gpu_addr);
 }
 
-static void
+static inline void
 set_vtx_resource(drm_radeon_private_t *dev_priv, u64 gpu_addr)
 {
 	uint32_t sq_vtx_constant_word2;
@@ -220,7 +220,7 @@ set_vtx_resource(drm_radeon_private_t *dev_priv, u64 gpu_addr)
 				    R600_VC_ACTION_ENA, 48, gpu_addr);
 }
 
-static void
+static inline void
 set_tex_resource(drm_radeon_private_t *dev_priv,
 		 int format, int w, int h, int pitch, u64 gpu_addr)
 {
@@ -258,7 +258,7 @@ set_tex_resource(drm_radeon_private_t *dev_priv,
 
 }
 
-static void
+static inline void
 set_scissors(drm_radeon_private_t *dev_priv, int x1, int y1, int x2, int y2)
 {
 	RING_LOCALS;
@@ -282,7 +282,7 @@ set_scissors(drm_radeon_private_t *dev_priv, int x1, int y1, int x2, int y2)
 	ADVANCE_RING();
 }
 
-static void
+static inline void
 draw_auto(drm_radeon_private_t *dev_priv)
 {
 	RING_LOCALS;
@@ -311,7 +311,7 @@ draw_auto(drm_radeon_private_t *dev_priv)
 	COMMIT_RING();
 }
 
-static void
+static inline void
 set_default_state(drm_radeon_private_t *dev_priv)
 {
 	int i;
@@ -489,7 +489,7 @@ set_default_state(drm_radeon_private_t *dev_priv)
 	ADVANCE_RING();
 }
 
-static uint32_t i2f(uint32_t input)
+static inline uint32_t i2f(uint32_t input)
 {
 	u32 result, i, exponent, fraction;
 
@@ -515,7 +515,7 @@ static uint32_t i2f(uint32_t input)
 }
 
 
-static int r600_nomm_get_vb(struct drm_device *dev)
+static inline int r600_nomm_get_vb(struct drm_device *dev)
 {
 	drm_radeon_private_t *dev_priv = dev->dev_private;
 	dev_priv->blit_vb = radeon_freelist_get(dev);
@@ -526,7 +526,7 @@ static int r600_nomm_get_vb(struct drm_device *dev)
 	return 0;
 }
 
-static void r600_nomm_put_vb(struct drm_device *dev)
+static inline void r600_nomm_put_vb(struct drm_device *dev)
 {
 	drm_radeon_private_t *dev_priv = dev->dev_private;
 
@@ -534,7 +534,7 @@ static void r600_nomm_put_vb(struct drm_device *dev)
 	radeon_cp_discard_buffer(dev, dev_priv->blit_vb->file_priv->master, dev_priv->blit_vb);
 }
 
-static void *r600_nomm_get_vb_ptr(struct drm_device *dev)
+static inline void *r600_nomm_get_vb_ptr(struct drm_device *dev)
 {
 	drm_radeon_private_t *dev_priv = dev->dev_private;
 	return (((char *)dev->agp_buffer_map->handle +
diff --git a/drivers/gpu/drm/radeon/r600_blit_kms.c b/drivers/gpu/drm/radeon/r600_blit_kms.c
index db38f58..9aa74c3 100644
--- a/drivers/gpu/drm/radeon/r600_blit_kms.c
+++ b/drivers/gpu/drm/radeon/r600_blit_kms.c
@@ -30,14 +30,23 @@
 
 #include "r600d.h"
 #include "r600_blit_shaders.h"
-#include "radeon_blit_common.h"
+
+#define DI_PT_RECTLIST        0x11
+#define DI_INDEX_SIZE_16_BIT  0x0
+#define DI_SRC_SEL_AUTO_INDEX 0x2
+
+#define FMT_8                 0x1
+#define FMT_5_6_5             0x8
+#define FMT_8_8_8_8           0x1a
+#define COLOR_8               0x1
+#define COLOR_5_6_5           0x8
+#define COLOR_8_8_8_8         0x1a
 
 /* emits 21 on rv770+, 23 on r600 */
 static void
 set_render_target(struct radeon_device *rdev, int format,
 		  int w, int h, u64 gpu_addr)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	u32 cb_color_info;
 	int pitch, slice;
 
@@ -45,44 +54,42 @@ set_render_target(struct radeon_device *rdev, int format,
 	if (h < 8)
 		h = 8;
 
-	cb_color_info = CB_FORMAT(format) |
-		CB_SOURCE_FORMAT(CB_SF_EXPORT_NORM) |
-		CB_ARRAY_MODE(ARRAY_1D_TILED_THIN1);
+	cb_color_info = ((format << 2) | (1 << 27) | (1 << 8));
 	pitch = (w / 8) - 1;
 	slice = ((w * h) / 64) - 1;
 
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
-	radeon_ring_write(ring, (CB_COLOR0_BASE - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
-	radeon_ring_write(ring, gpu_addr >> 8);
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (CB_COLOR0_BASE - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, gpu_addr >> 8);
 
 	if (rdev->family > CHIP_R600 && rdev->family < CHIP_RV770) {
-		radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_BASE_UPDATE, 0));
-		radeon_ring_write(ring, 2 << 0);
+		radeon_ring_write(rdev, PACKET3(PACKET3_SURFACE_BASE_UPDATE, 0));
+		radeon_ring_write(rdev, 2 << 0);
 	}
 
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
-	radeon_ring_write(ring, (CB_COLOR0_SIZE - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
-	radeon_ring_write(ring, (pitch << 0) | (slice << 10));
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (CB_COLOR0_SIZE - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, (pitch << 0) | (slice << 10));
 
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
-	radeon_ring_write(ring, (CB_COLOR0_VIEW - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
-	radeon_ring_write(ring, 0);
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (CB_COLOR0_VIEW - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, 0);
 
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
-	radeon_ring_write(ring, (CB_COLOR0_INFO - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
-	radeon_ring_write(ring, cb_color_info);
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (CB_COLOR0_INFO - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, cb_color_info);
 
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
-	radeon_ring_write(ring, (CB_COLOR0_TILE - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
-	radeon_ring_write(ring, 0);
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (CB_COLOR0_TILE - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, 0);
 
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
-	radeon_ring_write(ring, (CB_COLOR0_FRAG - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
-	radeon_ring_write(ring, 0);
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (CB_COLOR0_FRAG - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, 0);
 
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
-	radeon_ring_write(ring, (CB_COLOR0_MASK - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
-	radeon_ring_write(ring, 0);
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (CB_COLOR0_MASK - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, 0);
 }
 
 /* emits 5dw */
@@ -91,7 +98,6 @@ cp_set_surface_sync(struct radeon_device *rdev,
 		    u32 sync_type, u32 size,
 		    u64 mc_addr)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	u32 cp_coher_size;
 
 	if (size == 0xffffffff)
@@ -99,18 +105,17 @@ cp_set_surface_sync(struct radeon_device *rdev,
 	else
 		cp_coher_size = ((size + 255) >> 8);
 
-	radeon_ring_write(ring, PACKET3(PACKET3_SURFACE_SYNC, 3));
-	radeon_ring_write(ring, sync_type);
-	radeon_ring_write(ring, cp_coher_size);
-	radeon_ring_write(ring, mc_addr >> 8);
-	radeon_ring_write(ring, 10); /* poll interval */
+	radeon_ring_write(rdev, PACKET3(PACKET3_SURFACE_SYNC, 3));
+	radeon_ring_write(rdev, sync_type);
+	radeon_ring_write(rdev, cp_coher_size);
+	radeon_ring_write(rdev, mc_addr >> 8);
+	radeon_ring_write(rdev, 10); /* poll interval */
 }
 
 /* emits 21dw + 1 surface sync = 26dw */
 static void
 set_shaders(struct radeon_device *rdev)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	u64 gpu_addr;
 	u32 sq_pgm_resources;
 
@@ -119,35 +124,35 @@ set_shaders(struct radeon_device *rdev)
 
 	/* VS */
 	gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.vs_offset;
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
-	radeon_ring_write(ring, (SQ_PGM_START_VS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
-	radeon_ring_write(ring, gpu_addr >> 8);
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (SQ_PGM_START_VS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, gpu_addr >> 8);
 
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
-	radeon_ring_write(ring, (SQ_PGM_RESOURCES_VS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
-	radeon_ring_write(ring, sq_pgm_resources);
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (SQ_PGM_RESOURCES_VS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, sq_pgm_resources);
 
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
-	radeon_ring_write(ring, (SQ_PGM_CF_OFFSET_VS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
-	radeon_ring_write(ring, 0);
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (SQ_PGM_CF_OFFSET_VS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, 0);
 
 	/* PS */
 	gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.ps_offset;
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
-	radeon_ring_write(ring, (SQ_PGM_START_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
-	radeon_ring_write(ring, gpu_addr >> 8);
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (SQ_PGM_START_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, gpu_addr >> 8);
 
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
-	radeon_ring_write(ring, (SQ_PGM_RESOURCES_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
-	radeon_ring_write(ring, sq_pgm_resources | (1 << 28));
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (SQ_PGM_RESOURCES_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, sq_pgm_resources | (1 << 28));
 
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
-	radeon_ring_write(ring, (SQ_PGM_EXPORTS_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
-	radeon_ring_write(ring, 2);
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (SQ_PGM_EXPORTS_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, 2);
 
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
-	radeon_ring_write(ring, (SQ_PGM_CF_OFFSET_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
-	radeon_ring_write(ring, 0);
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 1));
+	radeon_ring_write(rdev, (SQ_PGM_CF_OFFSET_PS - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, 0);
 
 	gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.vs_offset;
 	cp_set_surface_sync(rdev, PACKET3_SH_ACTION_ENA, 512, gpu_addr);
@@ -157,24 +162,22 @@ set_shaders(struct radeon_device *rdev)
 static void
 set_vtx_resource(struct radeon_device *rdev, u64 gpu_addr)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	u32 sq_vtx_constant_word2;
 
-	sq_vtx_constant_word2 = SQ_VTXC_BASE_ADDR_HI(upper_32_bits(gpu_addr) & 0xff) |
-		SQ_VTXC_STRIDE(16);
+	sq_vtx_constant_word2 = ((upper_32_bits(gpu_addr) & 0xff) | (16 << 8));
 #ifdef __BIG_ENDIAN
-	sq_vtx_constant_word2 |=  SQ_VTXC_ENDIAN_SWAP(SQ_ENDIAN_8IN32);
+	sq_vtx_constant_word2 |= (2 << 30);
 #endif
 
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_RESOURCE, 7));
-	radeon_ring_write(ring, 0x460);
-	radeon_ring_write(ring, gpu_addr & 0xffffffff);
-	radeon_ring_write(ring, 48 - 1);
-	radeon_ring_write(ring, sq_vtx_constant_word2);
-	radeon_ring_write(ring, 1 << 0);
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, SQ_TEX_VTX_VALID_BUFFER << 30);
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_RESOURCE, 7));
+	radeon_ring_write(rdev, 0x460);
+	radeon_ring_write(rdev, gpu_addr & 0xffffffff);
+	radeon_ring_write(rdev, 48 - 1);
+	radeon_ring_write(rdev, sq_vtx_constant_word2);
+	radeon_ring_write(rdev, 1 << 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, SQ_TEX_VTX_VALID_BUFFER << 30);
 
 	if ((rdev->family == CHIP_RV610) ||
 	    (rdev->family == CHIP_RV620) ||
@@ -192,40 +195,35 @@ set_vtx_resource(struct radeon_device *rdev, u64 gpu_addr)
 static void
 set_tex_resource(struct radeon_device *rdev,
 		 int format, int w, int h, int pitch,
-		 u64 gpu_addr, u32 size)
+		 u64 gpu_addr)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	uint32_t sq_tex_resource_word0, sq_tex_resource_word1, sq_tex_resource_word4;
 
 	if (h < 1)
 		h = 1;
 
-	sq_tex_resource_word0 = S_038000_DIM(V_038000_SQ_TEX_DIM_2D) |
-		S_038000_TILE_MODE(V_038000_ARRAY_1D_TILED_THIN1);
-	sq_tex_resource_word0 |= S_038000_PITCH((pitch >> 3) - 1) |
-		S_038000_TEX_WIDTH(w - 1);
-
-	sq_tex_resource_word1 = S_038004_DATA_FORMAT(format);
-	sq_tex_resource_word1 |= S_038004_TEX_HEIGHT(h - 1);
-
-	sq_tex_resource_word4 = S_038010_REQUEST_SIZE(1) |
-		S_038010_DST_SEL_X(SQ_SEL_X) |
-		S_038010_DST_SEL_Y(SQ_SEL_Y) |
-		S_038010_DST_SEL_Z(SQ_SEL_Z) |
-		S_038010_DST_SEL_W(SQ_SEL_W);
-
-	cp_set_surface_sync(rdev,
-			    PACKET3_TC_ACTION_ENA, size, gpu_addr);
-
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_RESOURCE, 7));
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, sq_tex_resource_word0);
-	radeon_ring_write(ring, sq_tex_resource_word1);
-	radeon_ring_write(ring, gpu_addr >> 8);
-	radeon_ring_write(ring, gpu_addr >> 8);
-	radeon_ring_write(ring, sq_tex_resource_word4);
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, SQ_TEX_VTX_VALID_TEXTURE << 30);
+	sq_tex_resource_word0 = (1 << 0) | (1 << 3);
+	sq_tex_resource_word0 |= ((((pitch >> 3) - 1) << 8) |
+				  ((w - 1) << 19));
+
+	sq_tex_resource_word1 = (format << 26);
+	sq_tex_resource_word1 |= ((h - 1) << 0);
+
+	sq_tex_resource_word4 = ((1 << 14) |
+				 (0 << 16) |
+				 (1 << 19) |
+				 (2 << 22) |
+				 (3 << 25));
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_RESOURCE, 7));
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, sq_tex_resource_word0);
+	radeon_ring_write(rdev, sq_tex_resource_word1);
+	radeon_ring_write(rdev, gpu_addr >> 8);
+	radeon_ring_write(rdev, gpu_addr >> 8);
+	radeon_ring_write(rdev, sq_tex_resource_word4);
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, SQ_TEX_VTX_VALID_TEXTURE << 30);
 }
 
 /* emits 12 */
@@ -233,45 +231,43 @@ static void
 set_scissors(struct radeon_device *rdev, int x1, int y1,
 	     int x2, int y2)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2));
-	radeon_ring_write(ring, (PA_SC_SCREEN_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
-	radeon_ring_write(ring, (x1 << 0) | (y1 << 16));
-	radeon_ring_write(ring, (x2 << 0) | (y2 << 16));
-
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2));
-	radeon_ring_write(ring, (PA_SC_GENERIC_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
-	radeon_ring_write(ring, (x1 << 0) | (y1 << 16) | (1 << 31));
-	radeon_ring_write(ring, (x2 << 0) | (y2 << 16));
-
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONTEXT_REG, 2));
-	radeon_ring_write(ring, (PA_SC_WINDOW_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
-	radeon_ring_write(ring, (x1 << 0) | (y1 << 16) | (1 << 31));
-	radeon_ring_write(ring, (x2 << 0) | (y2 << 16));
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 2));
+	radeon_ring_write(rdev, (PA_SC_SCREEN_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, (x1 << 0) | (y1 << 16));
+	radeon_ring_write(rdev, (x2 << 0) | (y2 << 16));
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 2));
+	radeon_ring_write(rdev, (PA_SC_GENERIC_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, (x1 << 0) | (y1 << 16) | (1 << 31));
+	radeon_ring_write(rdev, (x2 << 0) | (y2 << 16));
+
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONTEXT_REG, 2));
+	radeon_ring_write(rdev, (PA_SC_WINDOW_SCISSOR_TL - PACKET3_SET_CONTEXT_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, (x1 << 0) | (y1 << 16) | (1 << 31));
+	radeon_ring_write(rdev, (x2 << 0) | (y2 << 16));
 }
 
 /* emits 10 */
 static void
 draw_auto(struct radeon_device *rdev)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 1));
-	radeon_ring_write(ring, (VGT_PRIMITIVE_TYPE - PACKET3_SET_CONFIG_REG_OFFSET) >> 2);
-	radeon_ring_write(ring, DI_PT_RECTLIST);
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 1));
+	radeon_ring_write(rdev, (VGT_PRIMITIVE_TYPE - PACKET3_SET_CONFIG_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, DI_PT_RECTLIST);
 
-	radeon_ring_write(ring, PACKET3(PACKET3_INDEX_TYPE, 0));
-	radeon_ring_write(ring,
+	radeon_ring_write(rdev, PACKET3(PACKET3_INDEX_TYPE, 0));
+	radeon_ring_write(rdev,
 #ifdef __BIG_ENDIAN
 			  (2 << 2) |
 #endif
 			  DI_INDEX_SIZE_16_BIT);
 
-	radeon_ring_write(ring, PACKET3(PACKET3_NUM_INSTANCES, 0));
-	radeon_ring_write(ring, 1);
+	radeon_ring_write(rdev, PACKET3(PACKET3_NUM_INSTANCES, 0));
+	radeon_ring_write(rdev, 1);
 
-	radeon_ring_write(ring, PACKET3(PACKET3_DRAW_INDEX_AUTO, 1));
-	radeon_ring_write(ring, 3);
-	radeon_ring_write(ring, DI_SRC_SEL_AUTO_INDEX);
+	radeon_ring_write(rdev, PACKET3(PACKET3_DRAW_INDEX_AUTO, 1));
+	radeon_ring_write(rdev, 3);
+	radeon_ring_write(rdev, DI_SRC_SEL_AUTO_INDEX);
 
 }
 
@@ -279,7 +275,6 @@ draw_auto(struct radeon_device *rdev)
 static void
 set_default_state(struct radeon_device *rdev)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	u32 sq_config, sq_gpr_resource_mgmt_1, sq_gpr_resource_mgmt_2;
 	u32 sq_thread_resource_mgmt, sq_stack_resource_mgmt_1, sq_stack_resource_mgmt_2;
 	int num_ps_gprs, num_vs_gprs, num_temp_gprs, num_gs_gprs, num_es_gprs;
@@ -435,62 +430,47 @@ set_default_state(struct radeon_device *rdev)
 	/* emit an IB pointing at default state */
 	dwords = ALIGN(rdev->r600_blit.state_len, 0x10);
 	gpu_addr = rdev->r600_blit.shader_gpu_addr + rdev->r600_blit.state_offset;
-	radeon_ring_write(ring, PACKET3(PACKET3_INDIRECT_BUFFER, 2));
-	radeon_ring_write(ring,
+	radeon_ring_write(rdev, PACKET3(PACKET3_INDIRECT_BUFFER, 2));
+	radeon_ring_write(rdev,
 #ifdef __BIG_ENDIAN
 			  (2 << 0) |
 #endif
 			  (gpu_addr & 0xFFFFFFFC));
-	radeon_ring_write(ring, upper_32_bits(gpu_addr) & 0xFF);
-	radeon_ring_write(ring, dwords);
+	radeon_ring_write(rdev, upper_32_bits(gpu_addr) & 0xFF);
+	radeon_ring_write(rdev, dwords);
 
 	/* SQ config */
-	radeon_ring_write(ring, PACKET3(PACKET3_SET_CONFIG_REG, 6));
-	radeon_ring_write(ring, (SQ_CONFIG - PACKET3_SET_CONFIG_REG_OFFSET) >> 2);
-	radeon_ring_write(ring, sq_config);
-	radeon_ring_write(ring, sq_gpr_resource_mgmt_1);
-	radeon_ring_write(ring, sq_gpr_resource_mgmt_2);
-	radeon_ring_write(ring, sq_thread_resource_mgmt);
-	radeon_ring_write(ring, sq_stack_resource_mgmt_1);
-	radeon_ring_write(ring, sq_stack_resource_mgmt_2);
+	radeon_ring_write(rdev, PACKET3(PACKET3_SET_CONFIG_REG, 6));
+	radeon_ring_write(rdev, (SQ_CONFIG - PACKET3_SET_CONFIG_REG_OFFSET) >> 2);
+	radeon_ring_write(rdev, sq_config);
+	radeon_ring_write(rdev, sq_gpr_resource_mgmt_1);
+	radeon_ring_write(rdev, sq_gpr_resource_mgmt_2);
+	radeon_ring_write(rdev, sq_thread_resource_mgmt);
+	radeon_ring_write(rdev, sq_stack_resource_mgmt_1);
+	radeon_ring_write(rdev, sq_stack_resource_mgmt_2);
 }
 
-#define I2F_MAX_BITS 15
-#define I2F_MAX_INPUT  ((1 << I2F_MAX_BITS) - 1)
-#define I2F_SHIFT (24 - I2F_MAX_BITS)
-
-/*
- * Converts unsigned integer into 32-bit IEEE floating point representation.
- * Conversion is not universal and only works for the range from 0
- * to 2^I2F_MAX_BITS-1. Currently we only use it with inputs between
- * 0 and 16384 (inclusive), so I2F_MAX_BITS=15 is enough. If necessary,
- * I2F_MAX_BITS can be increased, but that will add to the loop iterations
- * and slow us down. Conversion is done by shifting the input and counting
- * down until the first 1 reaches bit position 23. The resulting counter
- * and the shifted input are, respectively, the exponent and the fraction.
- * The sign is always zero.
- */
-static uint32_t i2f(uint32_t input)
+static inline uint32_t i2f(uint32_t input)
 {
 	u32 result, i, exponent, fraction;
 
-	WARN_ON_ONCE(input > I2F_MAX_INPUT);
-
-	if ((input & I2F_MAX_INPUT) == 0)
-		result = 0;
+	if ((input & 0x3fff) == 0)
+		result = 0; /* 0 is a special case */
 	else {
-		exponent = 126 + I2F_MAX_BITS;
-		fraction = (input & I2F_MAX_INPUT) << I2F_SHIFT;
-
-		for (i = 0; i < I2F_MAX_BITS; i++) {
+		exponent = 140; /* exponent biased by 127; */
+		fraction = (input & 0x3fff) << 10; /* cheat and only
+						      handle numbers below 2^^15 */
+		for (i = 0; i < 14; i++) {
 			if (fraction & 0x800000)
 				break;
 			else {
-				fraction = fraction << 1;
+				fraction = fraction << 1; /* keep
+							     shifting left until top bit = 1 */
 				exponent = exponent - 1;
 			}
 		}
-		result = exponent << 23 | (fraction & 0x7fffff);
+		result = exponent << 23 | (fraction & 0x7fffff); /* mask
+								    off top bit; assumed 1 */
 	}
 	return result;
 }
@@ -503,27 +483,6 @@ int r600_blit_init(struct radeon_device *rdev)
 	u32 packet2s[16];
 	int num_packet2s = 0;
 
-	rdev->r600_blit.primitives.set_render_target = set_render_target;
-	rdev->r600_blit.primitives.cp_set_surface_sync = cp_set_surface_sync;
-	rdev->r600_blit.primitives.set_shaders = set_shaders;
-	rdev->r600_blit.primitives.set_vtx_resource = set_vtx_resource;
-	rdev->r600_blit.primitives.set_tex_resource = set_tex_resource;
-	rdev->r600_blit.primitives.set_scissors = set_scissors;
-	rdev->r600_blit.primitives.draw_auto = draw_auto;
-	rdev->r600_blit.primitives.set_default_state = set_default_state;
-
-	rdev->r600_blit.ring_size_common = 40; /* shaders + def state */
-	rdev->r600_blit.ring_size_common += 16; /* fence emit for VB IB */
-	rdev->r600_blit.ring_size_common += 5; /* done copy */
-	rdev->r600_blit.ring_size_common += 16; /* fence emit for done copy */
-
-	rdev->r600_blit.ring_size_per_loop = 76;
-	/* set_render_target emits 2 extra dwords on rv6xx */
-	if (rdev->family > CHIP_R600 && rdev->family < CHIP_RV770)
-		rdev->r600_blit.ring_size_per_loop += 2;
-
-	rdev->r600_blit.max_dim = 8192;
-
 	/* pin copy shader into vram if already initialized */
 	if (rdev->r600_blit.shader_obj)
 		goto done;
@@ -621,17 +580,16 @@ void r600_blit_fini(struct radeon_device *rdev)
 	radeon_bo_unref(&rdev->r600_blit.shader_obj);
 }
 
-static int r600_vb_ib_get(struct radeon_device *rdev, unsigned size)
+static int r600_vb_ib_get(struct radeon_device *rdev)
 {
 	int r;
-	r = radeon_ib_get(rdev, RADEON_RING_TYPE_GFX_INDEX,
-			  &rdev->r600_blit.vb_ib, size);
+	r = radeon_ib_get(rdev, &rdev->r600_blit.vb_ib);
 	if (r) {
 		DRM_ERROR("failed to get IB for vertex buffer\n");
 		return r;
 	}
 
-	rdev->r600_blit.vb_total = size;
+	rdev->r600_blit.vb_total = 64*1024;
 	rdev->r600_blit.vb_used = 0;
 	return 0;
 }
@@ -642,82 +600,47 @@ static void r600_vb_ib_put(struct radeon_device *rdev)
 	radeon_ib_free(rdev, &rdev->r600_blit.vb_ib);
 }
 
-static unsigned r600_blit_create_rect(unsigned num_gpu_pages,
-				      int *width, int *height, int max_dim)
+int r600_blit_prepare_copy(struct radeon_device *rdev, int size_bytes)
 {
-	unsigned max_pages;
-	unsigned pages = num_gpu_pages;
-	int w, h;
-
-	if (num_gpu_pages == 0) {
-		/* not supposed to be called with no pages, but just in case */
-		h = 0;
-		w = 0;
-		pages = 0;
-		WARN_ON(1);
-	} else {
-		int rect_order = 2;
-		h = RECT_UNIT_H;
-		while (num_gpu_pages / rect_order) {
-			h *= 2;
-			rect_order *= 4;
-			if (h >= max_dim) {
-				h = max_dim;
-				break;
-			}
-		}
-		max_pages = (max_dim * h) / (RECT_UNIT_W * RECT_UNIT_H);
-		if (pages > max_pages)
-			pages = max_pages;
-		w = (pages * RECT_UNIT_W * RECT_UNIT_H) / h;
-		w = (w / RECT_UNIT_W) * RECT_UNIT_W;
-		pages = (w * h) / (RECT_UNIT_W * RECT_UNIT_H);
-		BUG_ON(pages == 0);
-	}
-
-
-	DRM_DEBUG("blit_rectangle: h=%d, w=%d, pages=%d\n", h, w, pages);
-
-	/* return width and height only of the caller wants it */
-	if (height)
-		*height = h;
-	if (width)
-		*width = w;
-
-	return pages;
-}
-
-
-int r600_blit_prepare_copy(struct radeon_device *rdev, unsigned num_gpu_pages)
-{
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	int r;
-	int ring_size;
-	int num_loops = 0;
-	int dwords_per_loop = rdev->r600_blit.ring_size_per_loop;
-
-	/* num loops */
-	while (num_gpu_pages) {
-		num_gpu_pages -=
-			r600_blit_create_rect(num_gpu_pages, NULL, NULL,
-					      rdev->r600_blit.max_dim);
-		num_loops++;
-	}
+	int ring_size, line_size;
+	int max_size;
+	/* loops of emits 64 + fence emit possible */
+	int dwords_per_loop = 76, num_loops;
 
-	/* 48 bytes for vertex per loop */
-	r = r600_vb_ib_get(rdev, (num_loops*48)+256);
+	r = r600_vb_ib_get(rdev);
 	if (r)
 		return r;
 
+	/* set_render_target emits 2 extra dwords on rv6xx */
+	if (rdev->family > CHIP_R600 && rdev->family < CHIP_RV770)
+		dwords_per_loop += 2;
+
+	/* 8 bpp vs 32 bpp for xfer unit */
+	if (size_bytes & 3)
+		line_size = 8192;
+	else
+		line_size = 8192*4;
+
+	max_size = 8192 * line_size;
+
+	/* major loops cover the max size transfer */
+	num_loops = ((size_bytes + max_size) / max_size);
+	/* minor loops cover the extra non aligned bits */
+	num_loops += ((size_bytes % line_size) ? 1 : 0);
 	/* calculate number of loops correctly */
 	ring_size = num_loops * dwords_per_loop;
-	ring_size += rdev->r600_blit.ring_size_common;
-	r = radeon_ring_lock(rdev, ring, ring_size);
+	/* set default  + shaders */
+	ring_size += 40; /* shaders + def state */
+	ring_size += 10; /* fence emit for VB IB */
+	ring_size += 5; /* done copy */
+	ring_size += 10; /* fence emit for done copy */
+	r = radeon_ring_lock(rdev, ring_size);
 	if (r)
 		return r;
 
-	rdev->r600_blit.primitives.set_default_state(rdev);
-	rdev->r600_blit.primitives.set_shaders(rdev);
+	set_default_state(rdev); /* 14 */
+	set_shaders(rdev); /* 26 */
 	return 0;
 }
 
@@ -731,66 +654,187 @@ void r600_blit_done_copy(struct radeon_device *rdev, struct radeon_fence *fence)
 	if (fence)
 		r = radeon_fence_emit(rdev, fence);
 
-	radeon_ring_unlock_commit(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
+	radeon_ring_unlock_commit(rdev);
 }
 
 void r600_kms_blit_copy(struct radeon_device *rdev,
 			u64 src_gpu_addr, u64 dst_gpu_addr,
-			unsigned num_gpu_pages)
+			int size_bytes)
 {
+	int max_bytes;
 	u64 vb_gpu_addr;
 	u32 *vb;
 
-	DRM_DEBUG("emitting copy %16llx %16llx %d %d\n",
-		  src_gpu_addr, dst_gpu_addr,
-		  num_gpu_pages, rdev->r600_blit.vb_used);
+	DRM_DEBUG("emitting copy %16llx %16llx %d %d\n", src_gpu_addr, dst_gpu_addr,
+		  size_bytes, rdev->r600_blit.vb_used);
 	vb = (u32 *)(rdev->r600_blit.vb_ib->ptr + rdev->r600_blit.vb_used);
+	if ((size_bytes & 3) || (src_gpu_addr & 3) || (dst_gpu_addr & 3)) {
+		max_bytes = 8192;
+
+		while (size_bytes) {
+			int cur_size = size_bytes;
+			int src_x = src_gpu_addr & 255;
+			int dst_x = dst_gpu_addr & 255;
+			int h = 1;
+			src_gpu_addr = src_gpu_addr & ~255ULL;
+			dst_gpu_addr = dst_gpu_addr & ~255ULL;
+
+			if (!src_x && !dst_x) {
+				h = (cur_size / max_bytes);
+				if (h > 8192)
+					h = 8192;
+				if (h == 0)
+					h = 1;
+				else
+					cur_size = max_bytes;
+			} else {
+				if (cur_size > max_bytes)
+					cur_size = max_bytes;
+				if (cur_size > (max_bytes - dst_x))
+					cur_size = (max_bytes - dst_x);
+				if (cur_size > (max_bytes - src_x))
+					cur_size = (max_bytes - src_x);
+			}
+
+			if ((rdev->r600_blit.vb_used + 48) > rdev->r600_blit.vb_total) {
+				WARN_ON(1);
+			}
+
+			vb[0] = i2f(dst_x);
+			vb[1] = 0;
+			vb[2] = i2f(src_x);
+			vb[3] = 0;
+
+			vb[4] = i2f(dst_x);
+			vb[5] = i2f(h);
+			vb[6] = i2f(src_x);
+			vb[7] = i2f(h);
+
+			vb[8] = i2f(dst_x + cur_size);
+			vb[9] = i2f(h);
+			vb[10] = i2f(src_x + cur_size);
+			vb[11] = i2f(h);
 
-	while (num_gpu_pages) {
-		int w, h;
-		unsigned size_in_bytes;
-		unsigned pages_per_loop =
-			r600_blit_create_rect(num_gpu_pages, &w, &h,
-					      rdev->r600_blit.max_dim);
+			/* src 9 */
+			set_tex_resource(rdev, FMT_8,
+					 src_x + cur_size, h, src_x + cur_size,
+					 src_gpu_addr);
 
-		size_in_bytes = pages_per_loop * RADEON_GPU_PAGE_SIZE;
-		DRM_DEBUG("rectangle w=%d h=%d\n", w, h);
+			/* 5 */
+			cp_set_surface_sync(rdev,
+					    PACKET3_TC_ACTION_ENA, (src_x + cur_size * h), src_gpu_addr);
 
-		if ((rdev->r600_blit.vb_used + 48) > rdev->r600_blit.vb_total) {
-			WARN_ON(1);
+			/* dst 23 */
+			set_render_target(rdev, COLOR_8,
+					  dst_x + cur_size, h,
+					  dst_gpu_addr);
+
+			/* scissors 12 */
+			set_scissors(rdev, dst_x, 0, dst_x + cur_size, h);
+
+			/* 14 */
+			vb_gpu_addr = rdev->r600_blit.vb_ib->gpu_addr + rdev->r600_blit.vb_used;
+			set_vtx_resource(rdev, vb_gpu_addr);
+
+			/* draw 10 */
+			draw_auto(rdev);
+
+			/* 5 */
+			cp_set_surface_sync(rdev,
+					    PACKET3_CB_ACTION_ENA | PACKET3_CB0_DEST_BASE_ENA,
+					    cur_size * h, dst_gpu_addr);
+
+			vb += 12;
+			rdev->r600_blit.vb_used += 12 * 4;
+
+			src_gpu_addr += cur_size * h;
+			dst_gpu_addr += cur_size * h;
+			size_bytes -= cur_size * h;
 		}
+	} else {
+		max_bytes = 8192 * 4;
+
+		while (size_bytes) {
+			int cur_size = size_bytes;
+			int src_x = (src_gpu_addr & 255);
+			int dst_x = (dst_gpu_addr & 255);
+			int h = 1;
+			src_gpu_addr = src_gpu_addr & ~255ULL;
+			dst_gpu_addr = dst_gpu_addr & ~255ULL;
+
+			if (!src_x && !dst_x) {
+				h = (cur_size / max_bytes);
+				if (h > 8192)
+					h = 8192;
+				if (h == 0)
+					h = 1;
+				else
+					cur_size = max_bytes;
+			} else {
+				if (cur_size > max_bytes)
+					cur_size = max_bytes;
+				if (cur_size > (max_bytes - dst_x))
+					cur_size = (max_bytes - dst_x);
+				if (cur_size > (max_bytes - src_x))
+					cur_size = (max_bytes - src_x);
+			}
+
+			if ((rdev->r600_blit.vb_used + 48) > rdev->r600_blit.vb_total) {
+				WARN_ON(1);
+			}
 
-		vb[0] = 0;
-		vb[1] = 0;
-		vb[2] = 0;
-		vb[3] = 0;
-
-		vb[4] = 0;
-		vb[5] = i2f(h);
-		vb[6] = 0;
-		vb[7] = i2f(h);
-
-		vb[8] = i2f(w);
-		vb[9] = i2f(h);
-		vb[10] = i2f(w);
-		vb[11] = i2f(h);
-
-		rdev->r600_blit.primitives.set_tex_resource(rdev, FMT_8_8_8_8,
-							    w, h, w, src_gpu_addr, size_in_bytes);
-		rdev->r600_blit.primitives.set_render_target(rdev, COLOR_8_8_8_8,
-							     w, h, dst_gpu_addr);
-		rdev->r600_blit.primitives.set_scissors(rdev, 0, 0, w, h);
-		vb_gpu_addr = rdev->r600_blit.vb_ib->gpu_addr + rdev->r600_blit.vb_used;
-		rdev->r600_blit.primitives.set_vtx_resource(rdev, vb_gpu_addr);
-		rdev->r600_blit.primitives.draw_auto(rdev);
-		rdev->r600_blit.primitives.cp_set_surface_sync(rdev,
-				    PACKET3_CB_ACTION_ENA | PACKET3_CB0_DEST_BASE_ENA,
-				    size_in_bytes, dst_gpu_addr);
-
-		vb += 12;
-		rdev->r600_blit.vb_used += 4*12;
-		src_gpu_addr += size_in_bytes;
-		dst_gpu_addr += size_in_bytes;
-		num_gpu_pages -= pages_per_loop;
+			vb[0] = i2f(dst_x / 4);
+			vb[1] = 0;
+			vb[2] = i2f(src_x / 4);
+			vb[3] = 0;
+
+			vb[4] = i2f(dst_x / 4);
+			vb[5] = i2f(h);
+			vb[6] = i2f(src_x / 4);
+			vb[7] = i2f(h);
+
+			vb[8] = i2f((dst_x + cur_size) / 4);
+			vb[9] = i2f(h);
+			vb[10] = i2f((src_x + cur_size) / 4);
+			vb[11] = i2f(h);
+
+			/* src 9 */
+			set_tex_resource(rdev, FMT_8_8_8_8,
+					 (src_x + cur_size) / 4,
+					 h, (src_x + cur_size) / 4,
+					 src_gpu_addr);
+			/* 5 */
+			cp_set_surface_sync(rdev,
+					    PACKET3_TC_ACTION_ENA, (src_x + cur_size * h), src_gpu_addr);
+
+			/* dst 23 */
+			set_render_target(rdev, COLOR_8_8_8_8,
+					  (dst_x + cur_size) / 4, h,
+					  dst_gpu_addr);
+
+			/* scissors 12  */
+			set_scissors(rdev, (dst_x / 4), 0, (dst_x + cur_size / 4), h);
+
+			/* Vertex buffer setup 14 */
+			vb_gpu_addr = rdev->r600_blit.vb_ib->gpu_addr + rdev->r600_blit.vb_used;
+			set_vtx_resource(rdev, vb_gpu_addr);
+
+			/* draw 10 */
+			draw_auto(rdev);
+
+			/* 5 */
+			cp_set_surface_sync(rdev,
+					    PACKET3_CB_ACTION_ENA | PACKET3_CB0_DEST_BASE_ENA,
+					    cur_size * h, dst_gpu_addr);
+
+			/* 78 ring dwords per loop */
+			vb += 12;
+			rdev->r600_blit.vb_used += 12 * 4;
+
+			src_gpu_addr += cur_size * h;
+			dst_gpu_addr += cur_size * h;
+			size_bytes -= cur_size * h;
+		}
 	}
 }
+
diff --git a/drivers/gpu/drm/radeon/r600_cp.c b/drivers/gpu/drm/radeon/r600_cp.c
index 84c5462..c3ab959 100644
--- a/drivers/gpu/drm/radeon/r600_cp.c
+++ b/drivers/gpu/drm/radeon/r600_cp.c
@@ -26,8 +26,6 @@
  *     Alex Deucher <alexander.deucher@amd.com>
  */
 
-#include <linux/module.h>
-
 #include "drmP.h"
 #include "drm.h"
 #include "radeon_drm.h"
@@ -1804,8 +1802,8 @@ static void r600_cp_init_ring_buffer(struct drm_device *dev,
 	/* Set ring buffer size */
 #ifdef __BIG_ENDIAN
 	RADEON_WRITE(R600_CP_RB_CNTL,
-		     R600_BUF_SWAP_32BIT |
-		     R600_RB_NO_UPDATE |
+		     RADEON_BUF_SWAP_32BIT |
+		     RADEON_RB_NO_UPDATE |
 		     (dev_priv->ring.rptr_update_l2qw << 8) |
 		     dev_priv->ring.size_l2qw);
 #else
@@ -1815,22 +1813,22 @@ static void r600_cp_init_ring_buffer(struct drm_device *dev,
 		     dev_priv->ring.size_l2qw);
 #endif
 
-	RADEON_WRITE(R600_CP_SEM_WAIT_TIMER, 0x0);
+	RADEON_WRITE(R600_CP_SEM_WAIT_TIMER, 0x4);
 
 	/* Set the write pointer delay */
 	RADEON_WRITE(R600_CP_RB_WPTR_DELAY, 0);
 
 #ifdef __BIG_ENDIAN
 	RADEON_WRITE(R600_CP_RB_CNTL,
-		     R600_BUF_SWAP_32BIT |
-		     R600_RB_NO_UPDATE |
-		     R600_RB_RPTR_WR_ENA |
+		     RADEON_BUF_SWAP_32BIT |
+		     RADEON_RB_NO_UPDATE |
+		     RADEON_RB_RPTR_WR_ENA |
 		     (dev_priv->ring.rptr_update_l2qw << 8) |
 		     dev_priv->ring.size_l2qw);
 #else
 	RADEON_WRITE(R600_CP_RB_CNTL,
-		     R600_RB_NO_UPDATE |
-		     R600_RB_RPTR_WR_ENA |
+		     RADEON_RB_NO_UPDATE |
+		     RADEON_RB_RPTR_WR_ENA |
 		     (dev_priv->ring.rptr_update_l2qw << 8) |
 		     dev_priv->ring.size_l2qw);
 #endif
@@ -1853,8 +1851,13 @@ static void r600_cp_init_ring_buffer(struct drm_device *dev,
 			- ((unsigned long) dev->sg->virtual)
 			+ dev_priv->gart_vm_start;
 	}
-	RADEON_WRITE(R600_CP_RB_RPTR_ADDR, (rptr_addr & 0xfffffffc));
-	RADEON_WRITE(R600_CP_RB_RPTR_ADDR_HI, upper_32_bits(rptr_addr));
+	RADEON_WRITE(R600_CP_RB_RPTR_ADDR,
+#ifdef __BIG_ENDIAN
+		     (2 << 0) |
+#endif
+		     (rptr_addr & 0xfffffffc));
+	RADEON_WRITE(R600_CP_RB_RPTR_ADDR_HI,
+		     upper_32_bits(rptr_addr));
 
 #ifdef __BIG_ENDIAN
 	RADEON_WRITE(R600_CP_RB_CNTL,
diff --git a/drivers/gpu/drm/radeon/r600_cs.c b/drivers/gpu/drm/radeon/r600_cs.c
index 0ec3f20..909bda8 100644
--- a/drivers/gpu/drm/radeon/r600_cs.c
+++ b/drivers/gpu/drm/radeon/r600_cs.c
@@ -52,20 +52,15 @@ struct r600_cs_track {
 	struct radeon_bo	*cb_color_bo[8];
 	u64			cb_color_bo_mc[8];
 	u32			cb_color_bo_offset[8];
-	struct radeon_bo	*cb_color_frag_bo[8]; /* unused */
-	struct radeon_bo	*cb_color_tile_bo[8]; /* unused */
+	struct radeon_bo	*cb_color_frag_bo[8];
+	struct radeon_bo	*cb_color_tile_bo[8];
 	u32			cb_color_info[8];
-	u32			cb_color_view[8];
-	u32			cb_color_size_idx[8]; /* unused */
+	u32			cb_color_size_idx[8];
 	u32			cb_target_mask;
-	u32			cb_shader_mask;  /* unused */
+	u32			cb_shader_mask;
 	u32			cb_color_size[8];
 	u32			vgt_strmout_en;
 	u32			vgt_strmout_buffer_en;
-	struct radeon_bo	*vgt_strmout_bo[4];
-	u64			vgt_strmout_bo_mc[4]; /* unused */
-	u32			vgt_strmout_bo_offset[4];
-	u32			vgt_strmout_size[4];
 	u32			db_depth_control;
 	u32			db_depth_info;
 	u32			db_depth_size_idx;
@@ -74,17 +69,13 @@ struct r600_cs_track {
 	u32			db_offset;
 	struct radeon_bo	*db_bo;
 	u64			db_bo_mc;
-	bool			sx_misc_kill_all_prims;
-	bool			cb_dirty;
-	bool			db_dirty;
-	bool			streamout_dirty;
 };
 
 #define FMT_8_BIT(fmt, vc)   [fmt] = { 1, 1, 1, vc, CHIP_R600 }
 #define FMT_16_BIT(fmt, vc)  [fmt] = { 1, 1, 2, vc, CHIP_R600 }
-#define FMT_24_BIT(fmt)      [fmt] = { 1, 1, 4,  0, CHIP_R600 }
+#define FMT_24_BIT(fmt)      [fmt] = { 1, 1, 3,  0, CHIP_R600 }
 #define FMT_32_BIT(fmt, vc)  [fmt] = { 1, 1, 4, vc, CHIP_R600 }
-#define FMT_48_BIT(fmt)      [fmt] = { 1, 1, 8,  0, CHIP_R600 }
+#define FMT_48_BIT(fmt)      [fmt] = { 1, 1, 6,  0, CHIP_R600 }
 #define FMT_64_BIT(fmt, vc)  [fmt] = { 1, 1, 8, vc, CHIP_R600 }
 #define FMT_96_BIT(fmt)      [fmt] = { 1, 1, 12, 0, CHIP_R600 }
 #define FMT_128_BIT(fmt, vc) [fmt] = { 1, 1, 16,vc, CHIP_R600 }
@@ -116,7 +107,7 @@ static const struct gpu_formats color_formats_table[] = {
 
 	/* 24-bit */
 	FMT_24_BIT(V_038004_FMT_8_8_8),
-
+					       
 	/* 32-bit */
 	FMT_32_BIT(V_038004_COLOR_32, 1),
 	FMT_32_BIT(V_038004_COLOR_32_FLOAT, 1),
@@ -171,22 +162,22 @@ static const struct gpu_formats color_formats_table[] = {
 	[V_038004_FMT_32_AS_32_32_32_32] = { 1, 1, 4, 0, CHIP_CEDAR},
 };
 
-bool r600_fmt_is_valid_color(u32 format)
+static inline bool fmt_is_valid_color(u32 format)
 {
 	if (format >= ARRAY_SIZE(color_formats_table))
 		return false;
-
+	
 	if (color_formats_table[format].valid_color)
 		return true;
 
 	return false;
 }
 
-bool r600_fmt_is_valid_texture(u32 format, enum radeon_family family)
+static inline bool fmt_is_valid_texture(u32 format, enum radeon_family family)
 {
 	if (format >= ARRAY_SIZE(color_formats_table))
 		return false;
-
+	
 	if (family < color_formats_table[format].min_family)
 		return false;
 
@@ -196,7 +187,7 @@ bool r600_fmt_is_valid_texture(u32 format, enum radeon_family family)
 	return false;
 }
 
-int r600_fmt_get_blocksize(u32 format)
+static inline int fmt_get_blocksize(u32 format)
 {
 	if (format >= ARRAY_SIZE(color_formats_table))
 		return 0;
@@ -204,7 +195,7 @@ int r600_fmt_get_blocksize(u32 format)
 	return color_formats_table[format].blocksize;
 }
 
-int r600_fmt_get_nblocksx(u32 format, u32 w)
+static inline int fmt_get_nblocksx(u32 format, u32 w)
 {
 	unsigned bw;
 
@@ -218,7 +209,7 @@ int r600_fmt_get_nblocksx(u32 format, u32 w)
 	return (w + bw - 1) / bw;
 }
 
-int r600_fmt_get_nblocksy(u32 format, u32 h)
+static inline int fmt_get_nblocksy(u32 format, u32 h)
 {
 	unsigned bh;
 
@@ -232,6 +223,25 @@ int r600_fmt_get_nblocksy(u32 format, u32 h)
 	return (h + bh - 1) / bh;
 }
 
+static inline int r600_bpe_from_format(u32 *bpe, u32 format)
+{
+ 	unsigned res;
+
+	if (format >= ARRAY_SIZE(color_formats_table))
+		goto fail;
+
+	res = color_formats_table[format].blocksize;
+	if (res == 0)
+		goto fail;
+
+	*bpe = res;
+	return 0;
+
+fail:
+	*bpe = 16;
+	return -EINVAL;
+}
+
 struct array_mode_checker {
 	int array_mode;
 	u32 group_size;
@@ -242,7 +252,7 @@ struct array_mode_checker {
 };
 
 /* returns alignment in pixels for pitch/height/depth and bytes for base */
-static int r600_get_array_mode_alignment(struct array_mode_checker *values,
+static inline int r600_get_array_mode_alignment(struct array_mode_checker *values,
 						u32 *pitch_align,
 						u32 *height_align,
 						u32 *depth_align,
@@ -265,7 +275,7 @@ static int r600_get_array_mode_alignment(struct array_mode_checker *values,
 		break;
 	case ARRAY_LINEAR_ALIGNED:
 		*pitch_align = max((u32)64, (u32)(values->group_size / values->blocksize));
-		*height_align = 1;
+		*height_align = tile_height;
 		*depth_align = 1;
 		*base_align = values->group_size;
 		break;
@@ -278,9 +288,10 @@ static int r600_get_array_mode_alignment(struct array_mode_checker *values,
 		*base_align = values->group_size;
 		break;
 	case ARRAY_2D_TILED_THIN1:
-		*pitch_align = max((u32)macro_tile_width * tile_width,
-				(u32)((values->group_size * values->nbanks) /
-				(values->blocksize * values->nsamples * tile_width)));
+		*pitch_align = max((u32)macro_tile_width,
+				  (u32)(((values->group_size / tile_height) /
+					 (values->blocksize * values->nsamples)) *
+					values->nbanks)) * tile_width;
 		*height_align = macro_tile_height * tile_height;
 		*depth_align = 1;
 		*base_align = max(macro_tile_bytes,
@@ -304,14 +315,12 @@ static void r600_cs_track_init(struct r600_cs_track *track)
 		track->cb_color_size[i] = 0;
 		track->cb_color_size_idx[i] = 0;
 		track->cb_color_info[i] = 0;
-		track->cb_color_view[i] = 0xFFFFFFFF;
 		track->cb_color_bo[i] = NULL;
 		track->cb_color_bo_offset[i] = 0xFFFFFFFF;
 		track->cb_color_bo_mc[i] = 0xFFFFFFFF;
 	}
 	track->cb_target_mask = 0xFFFFFFFF;
 	track->cb_shader_mask = 0xFFFFFFFF;
-	track->cb_dirty = true;
 	track->db_bo = NULL;
 	track->db_bo_mc = 0xFFFFFFFF;
 	/* assume the biggest format and that htile is enabled */
@@ -320,19 +329,9 @@ static void r600_cs_track_init(struct r600_cs_track *track)
 	track->db_depth_size = 0xFFFFFFFF;
 	track->db_depth_size_idx = 0;
 	track->db_depth_control = 0xFFFFFFFF;
-	track->db_dirty = true;
-
-	for (i = 0; i < 4; i++) {
-		track->vgt_strmout_size[i] = 0;
-		track->vgt_strmout_bo[i] = NULL;
-		track->vgt_strmout_bo_offset[i] = 0xFFFFFFFF;
-		track->vgt_strmout_bo_mc[i] = 0xFFFFFFFF;
-	}
-	track->streamout_dirty = true;
-	track->sx_misc_kill_all_prims = false;
 }
 
-static int r600_cs_track_validate_cb(struct radeon_cs_parser *p, int i)
+static inline int r600_cs_track_validate_cb(struct radeon_cs_parser *p, int i)
 {
 	struct r600_cs_track *track = p->track;
 	u32 slice_tile_max, size, tmp;
@@ -342,14 +341,13 @@ static int r600_cs_track_validate_cb(struct radeon_cs_parser *p, int i)
 	volatile u32 *ib = p->ib->ptr;
 	unsigned array_mode;
 	u32 format;
-
 	if (G_0280A0_TILE_MODE(track->cb_color_info[i])) {
 		dev_warn(p->dev, "FMASK or CMASK buffer are not supported by this kernel\n");
 		return -EINVAL;
 	}
 	size = radeon_bo_size(track->cb_color_bo[i]) - track->cb_color_bo_offset[i];
 	format = G_0280A0_FORMAT(track->cb_color_info[i]);
-	if (!r600_fmt_is_valid_color(format)) {
+	if (!fmt_is_valid_color(format)) {
 		dev_warn(p->dev, "%s:%d cb invalid format %d for %d (0x%08X)\n",
 			 __func__, __LINE__, format,
 			i, track->cb_color_info[i]);
@@ -370,7 +368,7 @@ static int r600_cs_track_validate_cb(struct radeon_cs_parser *p, int i)
 	array_check.nbanks = track->nbanks;
 	array_check.npipes = track->npipes;
 	array_check.nsamples = track->nsamples;
-	array_check.blocksize = r600_fmt_get_blocksize(format);
+	array_check.blocksize = fmt_get_blocksize(format);
 	if (r600_get_array_mode_alignment(&array_check,
 					  &pitch_align, &height_align, &depth_align, &base_align)) {
 		dev_warn(p->dev, "%s invalid tiling %d for %d (0x%08X)\n", __func__,
@@ -414,18 +412,7 @@ static int r600_cs_track_validate_cb(struct radeon_cs_parser *p, int i)
 	}
 
 	/* check offset */
-	tmp = r600_fmt_get_nblocksy(format, height) * r600_fmt_get_nblocksx(format, pitch) * r600_fmt_get_blocksize(format);
-	switch (array_mode) {
-	default:
-	case V_0280A0_ARRAY_LINEAR_GENERAL:
-	case V_0280A0_ARRAY_LINEAR_ALIGNED:
-		tmp += track->cb_color_view[i] & 0xFF;
-		break;
-	case V_0280A0_ARRAY_1D_TILED_THIN1:
-	case V_0280A0_ARRAY_2D_TILED_THIN1:
-		tmp += G_028080_SLICE_MAX(track->cb_color_view[i]) * tmp;
-		break;
-	}
+	tmp = fmt_get_nblocksy(format, height) * fmt_get_nblocksx(format, pitch) * fmt_get_blocksize(format);
 	if ((tmp + track->cb_color_bo_offset[i]) > radeon_bo_size(track->cb_color_bo[i])) {
 		if (array_mode == V_0280A0_ARRAY_LINEAR_GENERAL) {
 			/* the initial DDX does bad things with the CB size occasionally */
@@ -435,13 +422,10 @@ static int r600_cs_track_validate_cb(struct radeon_cs_parser *p, int i)
 			 * broken userspace.
 			 */
 		} else {
-			dev_warn(p->dev, "%s offset[%d] %d %d %d %lu too big (%d %d) (%d %d %d)\n",
-				 __func__, i, array_mode,
+			dev_warn(p->dev, "%s offset[%d] %d %d %d %lu too big\n", __func__, i,
+				 array_mode,
 				 track->cb_color_bo_offset[i], tmp,
-				 radeon_bo_size(track->cb_color_bo[i]),
-				 pitch, height, r600_fmt_get_nblocksx(format, pitch),
-				 r600_fmt_get_nblocksy(format, height),
-				 r600_fmt_get_blocksize(format));
+				 radeon_bo_size(track->cb_color_bo[i]));
 			return -EINVAL;
 		}
 	}
@@ -465,171 +449,143 @@ static int r600_cs_track_check(struct radeon_cs_parser *p)
 	/* on legacy kernel we don't perform advanced check */
 	if (p->rdev == NULL)
 		return 0;
-
-	/* check streamout */
-	if (track->streamout_dirty && track->vgt_strmout_en) {
-		for (i = 0; i < 4; i++) {
-			if (track->vgt_strmout_buffer_en & (1 << i)) {
-				if (track->vgt_strmout_bo[i]) {
-					u64 offset = (u64)track->vgt_strmout_bo_offset[i] +
-						(u64)track->vgt_strmout_size[i];
-					if (offset > radeon_bo_size(track->vgt_strmout_bo[i])) {
-						DRM_ERROR("streamout %d bo too small: 0x%llx, 0x%lx\n",
-							  i, offset,
-							  radeon_bo_size(track->vgt_strmout_bo[i]));
-						return -EINVAL;
-					}
-				} else {
-					dev_warn(p->dev, "No buffer for streamout %d\n", i);
-					return -EINVAL;
-				}
-			}
-		}
-		track->streamout_dirty = false;
+	/* we don't support out buffer yet */
+	if (track->vgt_strmout_en || track->vgt_strmout_buffer_en) {
+		dev_warn(p->dev, "this kernel doesn't support SMX output buffer\n");
+		return -EINVAL;
 	}
-
-	if (track->sx_misc_kill_all_prims)
-		return 0;
-
 	/* check that we have a cb for each enabled target, we don't check
 	 * shader_mask because it seems mesa isn't always setting it :(
 	 */
-	if (track->cb_dirty) {
-		tmp = track->cb_target_mask;
-		for (i = 0; i < 8; i++) {
-			if ((tmp >> (i * 4)) & 0xF) {
-				/* at least one component is enabled */
-				if (track->cb_color_bo[i] == NULL) {
-					dev_warn(p->dev, "%s:%d mask 0x%08X | 0x%08X no cb for %d\n",
-						__func__, __LINE__, track->cb_target_mask, track->cb_shader_mask, i);
-					return -EINVAL;
-				}
-				/* perform rewrite of CB_COLOR[0-7]_SIZE */
-				r = r600_cs_track_validate_cb(p, i);
-				if (r)
-					return r;
+	tmp = track->cb_target_mask;
+	for (i = 0; i < 8; i++) {
+		if ((tmp >> (i * 4)) & 0xF) {
+			/* at least one component is enabled */
+			if (track->cb_color_bo[i] == NULL) {
+				dev_warn(p->dev, "%s:%d mask 0x%08X | 0x%08X no cb for %d\n",
+					__func__, __LINE__, track->cb_target_mask, track->cb_shader_mask, i);
+				return -EINVAL;
 			}
+			/* perform rewrite of CB_COLOR[0-7]_SIZE */
+			r = r600_cs_track_validate_cb(p, i);
+			if (r)
+				return r;
 		}
-		track->cb_dirty = false;
 	}
-
-	if (track->db_dirty) {
-		/* Check depth buffer */
-		if (G_028800_STENCIL_ENABLE(track->db_depth_control) ||
-			G_028800_Z_ENABLE(track->db_depth_control)) {
-			u32 nviews, bpe, ntiles, size, slice_tile_max;
-			u32 height, height_align, pitch, pitch_align, depth_align;
-			u64 base_offset, base_align;
-			struct array_mode_checker array_check;
-			int array_mode;
-
-			if (track->db_bo == NULL) {
-				dev_warn(p->dev, "z/stencil with no depth buffer\n");
+	/* Check depth buffer */
+	if (G_028800_STENCIL_ENABLE(track->db_depth_control) ||
+		G_028800_Z_ENABLE(track->db_depth_control)) {
+		u32 nviews, bpe, ntiles, size, slice_tile_max;
+		u32 height, height_align, pitch, pitch_align, depth_align;
+		u64 base_offset, base_align;
+		struct array_mode_checker array_check;
+		int array_mode;
+
+		if (track->db_bo == NULL) {
+			dev_warn(p->dev, "z/stencil with no depth buffer\n");
+			return -EINVAL;
+		}
+		if (G_028010_TILE_SURFACE_ENABLE(track->db_depth_info)) {
+			dev_warn(p->dev, "this kernel doesn't support z/stencil htile\n");
+			return -EINVAL;
+		}
+		switch (G_028010_FORMAT(track->db_depth_info)) {
+		case V_028010_DEPTH_16:
+			bpe = 2;
+			break;
+		case V_028010_DEPTH_X8_24:
+		case V_028010_DEPTH_8_24:
+		case V_028010_DEPTH_X8_24_FLOAT:
+		case V_028010_DEPTH_8_24_FLOAT:
+		case V_028010_DEPTH_32_FLOAT:
+			bpe = 4;
+			break;
+		case V_028010_DEPTH_X24_8_32_FLOAT:
+			bpe = 8;
+			break;
+		default:
+			dev_warn(p->dev, "z/stencil with invalid format %d\n", G_028010_FORMAT(track->db_depth_info));
+			return -EINVAL;
+		}
+		if ((track->db_depth_size & 0xFFFFFC00) == 0xFFFFFC00) {
+			if (!track->db_depth_size_idx) {
+				dev_warn(p->dev, "z/stencil buffer size not set\n");
 				return -EINVAL;
 			}
-			if (G_028010_TILE_SURFACE_ENABLE(track->db_depth_info)) {
-				dev_warn(p->dev, "this kernel doesn't support z/stencil htile\n");
+			tmp = radeon_bo_size(track->db_bo) - track->db_offset;
+			tmp = (tmp / bpe) >> 6;
+			if (!tmp) {
+				dev_warn(p->dev, "z/stencil buffer too small (0x%08X %d %d %ld)\n",
+						track->db_depth_size, bpe, track->db_offset,
+						radeon_bo_size(track->db_bo));
 				return -EINVAL;
 			}
-			switch (G_028010_FORMAT(track->db_depth_info)) {
-			case V_028010_DEPTH_16:
-				bpe = 2;
-				break;
-			case V_028010_DEPTH_X8_24:
-			case V_028010_DEPTH_8_24:
-			case V_028010_DEPTH_X8_24_FLOAT:
-			case V_028010_DEPTH_8_24_FLOAT:
-			case V_028010_DEPTH_32_FLOAT:
-				bpe = 4;
+			ib[track->db_depth_size_idx] = S_028000_SLICE_TILE_MAX(tmp - 1) | (track->db_depth_size & 0x3FF);
+		} else {
+			size = radeon_bo_size(track->db_bo);
+			/* pitch in pixels */
+			pitch = (G_028000_PITCH_TILE_MAX(track->db_depth_size) + 1) * 8;
+			slice_tile_max = G_028000_SLICE_TILE_MAX(track->db_depth_size) + 1;
+			slice_tile_max *= 64;
+			height = slice_tile_max / pitch;
+			if (height > 8192)
+				height = 8192;
+			base_offset = track->db_bo_mc + track->db_offset;
+			array_mode = G_028010_ARRAY_MODE(track->db_depth_info);
+			array_check.array_mode = array_mode;
+			array_check.group_size = track->group_size;
+			array_check.nbanks = track->nbanks;
+			array_check.npipes = track->npipes;
+			array_check.nsamples = track->nsamples;
+			array_check.blocksize = bpe;
+			if (r600_get_array_mode_alignment(&array_check,
+							  &pitch_align, &height_align, &depth_align, &base_align)) {
+				dev_warn(p->dev, "%s invalid tiling %d (0x%08X)\n", __func__,
+					 G_028010_ARRAY_MODE(track->db_depth_info),
+					 track->db_depth_info);
+				return -EINVAL;
+			}
+			switch (array_mode) {
+			case V_028010_ARRAY_1D_TILED_THIN1:
+				/* don't break userspace */
+				height &= ~0x7;
 				break;
-			case V_028010_DEPTH_X24_8_32_FLOAT:
-				bpe = 8;
+			case V_028010_ARRAY_2D_TILED_THIN1:
 				break;
 			default:
-				dev_warn(p->dev, "z/stencil with invalid format %d\n", G_028010_FORMAT(track->db_depth_info));
+				dev_warn(p->dev, "%s invalid tiling %d (0x%08X)\n", __func__,
+					 G_028010_ARRAY_MODE(track->db_depth_info),
+					 track->db_depth_info);
 				return -EINVAL;
 			}
-			if ((track->db_depth_size & 0xFFFFFC00) == 0xFFFFFC00) {
-				if (!track->db_depth_size_idx) {
-					dev_warn(p->dev, "z/stencil buffer size not set\n");
-					return -EINVAL;
-				}
-				tmp = radeon_bo_size(track->db_bo) - track->db_offset;
-				tmp = (tmp / bpe) >> 6;
-				if (!tmp) {
-					dev_warn(p->dev, "z/stencil buffer too small (0x%08X %d %d %ld)\n",
-							track->db_depth_size, bpe, track->db_offset,
-							radeon_bo_size(track->db_bo));
-					return -EINVAL;
-				}
-				ib[track->db_depth_size_idx] = S_028000_SLICE_TILE_MAX(tmp - 1) | (track->db_depth_size & 0x3FF);
-			} else {
-				size = radeon_bo_size(track->db_bo);
-				/* pitch in pixels */
-				pitch = (G_028000_PITCH_TILE_MAX(track->db_depth_size) + 1) * 8;
-				slice_tile_max = G_028000_SLICE_TILE_MAX(track->db_depth_size) + 1;
-				slice_tile_max *= 64;
-				height = slice_tile_max / pitch;
-				if (height > 8192)
-					height = 8192;
-				base_offset = track->db_bo_mc + track->db_offset;
-				array_mode = G_028010_ARRAY_MODE(track->db_depth_info);
-				array_check.array_mode = array_mode;
-				array_check.group_size = track->group_size;
-				array_check.nbanks = track->nbanks;
-				array_check.npipes = track->npipes;
-				array_check.nsamples = track->nsamples;
-				array_check.blocksize = bpe;
-				if (r600_get_array_mode_alignment(&array_check,
-								  &pitch_align, &height_align, &depth_align, &base_align)) {
-					dev_warn(p->dev, "%s invalid tiling %d (0x%08X)\n", __func__,
-						 G_028010_ARRAY_MODE(track->db_depth_info),
-						 track->db_depth_info);
-					return -EINVAL;
-				}
-				switch (array_mode) {
-				case V_028010_ARRAY_1D_TILED_THIN1:
-					/* don't break userspace */
-					height &= ~0x7;
-					break;
-				case V_028010_ARRAY_2D_TILED_THIN1:
-					break;
-				default:
-					dev_warn(p->dev, "%s invalid tiling %d (0x%08X)\n", __func__,
-						 G_028010_ARRAY_MODE(track->db_depth_info),
-						 track->db_depth_info);
-					return -EINVAL;
-				}
 
-				if (!IS_ALIGNED(pitch, pitch_align)) {
-					dev_warn(p->dev, "%s:%d db pitch (%d, 0x%x, %d) invalid\n",
-						 __func__, __LINE__, pitch, pitch_align, array_mode);
-					return -EINVAL;
-				}
-				if (!IS_ALIGNED(height, height_align)) {
-					dev_warn(p->dev, "%s:%d db height (%d, 0x%x, %d) invalid\n",
-						 __func__, __LINE__, height, height_align, array_mode);
-					return -EINVAL;
-				}
-				if (!IS_ALIGNED(base_offset, base_align)) {
-					dev_warn(p->dev, "%s offset[%d] 0x%llx, 0x%llx, %d not aligned\n", __func__, i,
-						 base_offset, base_align, array_mode);
-					return -EINVAL;
-				}
+			if (!IS_ALIGNED(pitch, pitch_align)) {
+				dev_warn(p->dev, "%s:%d db pitch (%d, 0x%x, %d) invalid\n",
+					 __func__, __LINE__, pitch, pitch_align, array_mode);
+				return -EINVAL;
+			}
+			if (!IS_ALIGNED(height, height_align)) {
+				dev_warn(p->dev, "%s:%d db height (%d, 0x%x, %d) invalid\n",
+					 __func__, __LINE__, height, height_align, array_mode);
+				return -EINVAL;
+			}
+			if (!IS_ALIGNED(base_offset, base_align)) {
+				dev_warn(p->dev, "%s offset[%d] 0x%llx, 0x%llx, %d not aligned\n", __func__, i,
+					 base_offset, base_align, array_mode);
+				return -EINVAL;
+			}
 
-				ntiles = G_028000_SLICE_TILE_MAX(track->db_depth_size) + 1;
-				nviews = G_028004_SLICE_MAX(track->db_depth_view) + 1;
-				tmp = ntiles * bpe * 64 * nviews;
-				if ((tmp + track->db_offset) > radeon_bo_size(track->db_bo)) {
-					dev_warn(p->dev, "z/stencil buffer (%d) too small (0x%08X %d %d %d -> %u have %lu)\n",
-						 array_mode,
-						 track->db_depth_size, ntiles, nviews, bpe, tmp + track->db_offset,
-						 radeon_bo_size(track->db_bo));
-					return -EINVAL;
-				}
+			ntiles = G_028000_SLICE_TILE_MAX(track->db_depth_size) + 1;
+			nviews = G_028004_SLICE_MAX(track->db_depth_view) + 1;
+			tmp = ntiles * bpe * 64 * nviews;
+			if ((tmp + track->db_offset) > radeon_bo_size(track->db_bo)) {
+				dev_warn(p->dev, "z/stencil buffer (%d) too small (0x%08X %d %d %d -> %u have %lu)\n",
+					 array_mode,
+					 track->db_depth_size, ntiles, nviews, bpe, tmp + track->db_offset,
+					 radeon_bo_size(track->db_bo));
+				return -EINVAL;
 			}
 		}
-		track->db_dirty = false;
 	}
 	return 0;
 }
@@ -781,7 +737,7 @@ static int r600_cs_packet_next_reloc_nomm(struct radeon_cs_parser *p,
  * Check next packet is relocation packet3, do bo validation and compute
  * GPU offset using the provided start.
  **/
-static int r600_cs_packet_next_is_pkt3_nop(struct radeon_cs_parser *p)
+static inline int r600_cs_packet_next_is_pkt3_nop(struct radeon_cs_parser *p)
 {
 	struct radeon_cs_packet p3reloc;
 	int r;
@@ -955,15 +911,16 @@ static int r600_cs_parse_packet0(struct radeon_cs_parser *p,
  * if register is safe. If register is not flag as safe this function
  * will test it against a list of register needind special handling.
  */
-static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
+static inline int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 {
 	struct r600_cs_track *track = (struct r600_cs_track *)p->track;
 	struct radeon_cs_reloc *reloc;
+	u32 last_reg = ARRAY_SIZE(r600_reg_safe_bm);
 	u32 m, i, tmp, *ib;
 	int r;
 
 	i = (reg >> 7);
-	if (i >= ARRAY_SIZE(r600_reg_safe_bm)) {
+	if (i > last_reg) {
 		dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx);
 		return -EINVAL;
 	}
@@ -1002,11 +959,9 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		break;
 	case R_028800_DB_DEPTH_CONTROL:
 		track->db_depth_control = radeon_get_ib_value(p, idx);
-		track->db_dirty = true;
 		break;
 	case R_028010_DB_DEPTH_INFO:
-		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS) &&
-		    r600_cs_packet_next_is_pkt3_nop(p)) {
+		if (r600_cs_packet_next_is_pkt3_nop(p)) {
 			r = r600_cs_packet_next_reloc(p, &reloc);
 			if (r) {
 				dev_warn(p->dev, "bad SET_CONTEXT_REG "
@@ -1023,66 +978,24 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 				ib[idx] |= S_028010_ARRAY_MODE(V_028010_ARRAY_1D_TILED_THIN1);
 				track->db_depth_info |= S_028010_ARRAY_MODE(V_028010_ARRAY_1D_TILED_THIN1);
 			}
-		} else {
+		} else
 			track->db_depth_info = radeon_get_ib_value(p, idx);
-		}
-		track->db_dirty = true;
 		break;
 	case R_028004_DB_DEPTH_VIEW:
 		track->db_depth_view = radeon_get_ib_value(p, idx);
-		track->db_dirty = true;
 		break;
 	case R_028000_DB_DEPTH_SIZE:
 		track->db_depth_size = radeon_get_ib_value(p, idx);
 		track->db_depth_size_idx = idx;
-		track->db_dirty = true;
 		break;
 	case R_028AB0_VGT_STRMOUT_EN:
 		track->vgt_strmout_en = radeon_get_ib_value(p, idx);
-		track->streamout_dirty = true;
 		break;
 	case R_028B20_VGT_STRMOUT_BUFFER_EN:
 		track->vgt_strmout_buffer_en = radeon_get_ib_value(p, idx);
-		track->streamout_dirty = true;
-		break;
-	case VGT_STRMOUT_BUFFER_BASE_0:
-	case VGT_STRMOUT_BUFFER_BASE_1:
-	case VGT_STRMOUT_BUFFER_BASE_2:
-	case VGT_STRMOUT_BUFFER_BASE_3:
-		r = r600_cs_packet_next_reloc(p, &reloc);
-		if (r) {
-			dev_warn(p->dev, "bad SET_CONTEXT_REG "
-					"0x%04X\n", reg);
-			return -EINVAL;
-		}
-		tmp = (reg - VGT_STRMOUT_BUFFER_BASE_0) / 16;
-		track->vgt_strmout_bo_offset[tmp] = radeon_get_ib_value(p, idx) << 8;
-		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
-		track->vgt_strmout_bo[tmp] = reloc->robj;
-		track->vgt_strmout_bo_mc[tmp] = reloc->lobj.gpu_offset;
-		track->streamout_dirty = true;
-		break;
-	case VGT_STRMOUT_BUFFER_SIZE_0:
-	case VGT_STRMOUT_BUFFER_SIZE_1:
-	case VGT_STRMOUT_BUFFER_SIZE_2:
-	case VGT_STRMOUT_BUFFER_SIZE_3:
-		tmp = (reg - VGT_STRMOUT_BUFFER_SIZE_0) / 16;
-		/* size in register is DWs, convert to bytes */
-		track->vgt_strmout_size[tmp] = radeon_get_ib_value(p, idx) * 4;
-		track->streamout_dirty = true;
-		break;
-	case CP_COHER_BASE:
-		r = r600_cs_packet_next_reloc(p, &reloc);
-		if (r) {
-			dev_warn(p->dev, "missing reloc for CP_COHER_BASE "
-					"0x%04X\n", reg);
-			return -EINVAL;
-		}
-		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 		break;
 	case R_028238_CB_TARGET_MASK:
 		track->cb_target_mask = radeon_get_ib_value(p, idx);
-		track->cb_dirty = true;
 		break;
 	case R_02823C_CB_SHADER_MASK:
 		track->cb_shader_mask = radeon_get_ib_value(p, idx);
@@ -1090,7 +1003,6 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case R_028C04_PA_SC_AA_CONFIG:
 		tmp = G_028C04_MSAA_NUM_SAMPLES(radeon_get_ib_value(p, idx));
 		track->nsamples = 1 << tmp;
-		track->cb_dirty = true;
 		break;
 	case R_0280A0_CB_COLOR0_INFO:
 	case R_0280A4_CB_COLOR1_INFO:
@@ -1100,8 +1012,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	case R_0280B4_CB_COLOR5_INFO:
 	case R_0280B8_CB_COLOR6_INFO:
 	case R_0280BC_CB_COLOR7_INFO:
-		if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS) &&
-		     r600_cs_packet_next_is_pkt3_nop(p)) {
+		if (r600_cs_packet_next_is_pkt3_nop(p)) {
 			r = r600_cs_packet_next_reloc(p, &reloc);
 			if (r) {
 				dev_err(p->dev, "bad SET_CONTEXT_REG 0x%04X\n", reg);
@@ -1120,19 +1031,6 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 			tmp = (reg - R_0280A0_CB_COLOR0_INFO) / 4;
 			track->cb_color_info[tmp] = radeon_get_ib_value(p, idx);
 		}
-		track->cb_dirty = true;
-		break;
-	case R_028080_CB_COLOR0_VIEW:
-	case R_028084_CB_COLOR1_VIEW:
-	case R_028088_CB_COLOR2_VIEW:
-	case R_02808C_CB_COLOR3_VIEW:
-	case R_028090_CB_COLOR4_VIEW:
-	case R_028094_CB_COLOR5_VIEW:
-	case R_028098_CB_COLOR6_VIEW:
-	case R_02809C_CB_COLOR7_VIEW:
-		tmp = (reg - R_028080_CB_COLOR0_VIEW) / 4;
-		track->cb_color_view[tmp] = radeon_get_ib_value(p, idx);
-		track->cb_dirty = true;
 		break;
 	case R_028060_CB_COLOR0_SIZE:
 	case R_028064_CB_COLOR1_SIZE:
@@ -1145,7 +1043,6 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		tmp = (reg - R_028060_CB_COLOR0_SIZE) / 4;
 		track->cb_color_size[tmp] = radeon_get_ib_value(p, idx);
 		track->cb_color_size_idx[tmp] = idx;
-		track->cb_dirty = true;
 		break;
 		/* This register were added late, there is userspace
 		 * which does provide relocation for those but set
@@ -1228,7 +1125,6 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		track->cb_color_base_last[tmp] = ib[idx];
 		track->cb_color_bo[tmp] = reloc->robj;
 		track->cb_color_bo_mc[tmp] = reloc->lobj.gpu_offset;
-		track->cb_dirty = true;
 		break;
 	case DB_DEPTH_BASE:
 		r = r600_cs_packet_next_reloc(p, &reloc);
@@ -1241,7 +1137,6 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 		track->db_bo = reloc->robj;
 		track->db_bo_mc = reloc->lobj.gpu_offset;
-		track->db_dirty = true;
 		break;
 	case DB_HTILE_DATA_BASE:
 	case SQ_PGM_START_FS:
@@ -1305,18 +1200,6 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 		}
 		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
 		break;
-	case SX_MEMORY_EXPORT_BASE:
-		r = r600_cs_packet_next_reloc(p, &reloc);
-		if (r) {
-			dev_warn(p->dev, "bad SET_CONFIG_REG "
-					"0x%04X\n", reg);
-			return -EINVAL;
-		}
-		ib[idx] += (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
-		break;
-	case SX_MISC:
-		track->sx_misc_kill_all_prims = (radeon_get_ib_value(p, idx) & 0x1) != 0;
-		break;
 	default:
 		dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx);
 		return -EINVAL;
@@ -1324,7 +1207,7 @@ static int r600_cs_check_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
 	return 0;
 }
 
-unsigned r600_mip_minify(unsigned size, unsigned level)
+static inline unsigned mip_minify(unsigned size, unsigned level)
 {
 	unsigned val;
 
@@ -1346,22 +1229,22 @@ static void r600_texture_size(unsigned nfaces, unsigned blevel, unsigned llevel,
 	unsigned nlevels = llevel - blevel + 1;
 
 	*l0_size = -1;
-	blocksize = r600_fmt_get_blocksize(format);
+	blocksize = fmt_get_blocksize(format);
 
-	w0 = r600_mip_minify(w0, 0);
-	h0 = r600_mip_minify(h0, 0);
-	d0 = r600_mip_minify(d0, 0);
+	w0 = mip_minify(w0, 0);
+	h0 = mip_minify(h0, 0);
+	d0 = mip_minify(d0, 0);
 	for(i = 0, offset = 0, level = blevel; i < nlevels; i++, level++) {
-		width = r600_mip_minify(w0, i);
-		nbx = r600_fmt_get_nblocksx(format, width);
+		width = mip_minify(w0, i);
+		nbx = fmt_get_nblocksx(format, width);
 
 		nbx = round_up(nbx, block_align);
 
-		height = r600_mip_minify(h0, i);
-		nby = r600_fmt_get_nblocksy(format, height);
+		height = mip_minify(h0, i);
+		nby = fmt_get_nblocksy(format, height);
 		nby = round_up(nby, height_align);
 
-		depth = r600_mip_minify(d0, i);
+		depth = mip_minify(d0, i);
 
 		size = nbx * nby * blocksize;
 		if (nfaces)
@@ -1394,7 +1277,7 @@ static void r600_texture_size(unsigned nfaces, unsigned blevel, unsigned llevel,
  * This function will check that the resource has valid field and that
  * the texture and mipmap bo object are big enough to cover this resource.
  */
-static int r600_check_texture_resource(struct radeon_cs_parser *p,  u32 idx,
+static inline int r600_check_texture_resource(struct radeon_cs_parser *p,  u32 idx,
 					      struct radeon_bo *texture,
 					      struct radeon_bo *mipmap,
 					      u64 base_offset,
@@ -1419,18 +1302,15 @@ static int r600_check_texture_resource(struct radeon_cs_parser *p,  u32 idx,
 	mip_offset <<= 8;
 
 	word0 = radeon_get_ib_value(p, idx + 0);
-	if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
-		if (tiling_flags & RADEON_TILING_MACRO)
-			word0 |= S_038000_TILE_MODE(V_038000_ARRAY_2D_TILED_THIN1);
-		else if (tiling_flags & RADEON_TILING_MICRO)
-			word0 |= S_038000_TILE_MODE(V_038000_ARRAY_1D_TILED_THIN1);
-	}
+	if (tiling_flags & RADEON_TILING_MACRO)
+		word0 |= S_038000_TILE_MODE(V_038000_ARRAY_2D_TILED_THIN1);
+	else if (tiling_flags & RADEON_TILING_MICRO)
+		word0 |= S_038000_TILE_MODE(V_038000_ARRAY_1D_TILED_THIN1);
 	word1 = radeon_get_ib_value(p, idx + 1);
 	w0 = G_038000_TEX_WIDTH(word0) + 1;
 	h0 = G_038004_TEX_HEIGHT(word1) + 1;
 	d0 = G_038004_TEX_DEPTH(word1);
 	nfaces = 1;
-	array = 0;
 	switch (G_038000_DIM(word0)) {
 	case V_038000_SQ_TEX_DIM_1D:
 	case V_038000_SQ_TEX_DIM_2D:
@@ -1453,7 +1333,7 @@ static int r600_check_texture_resource(struct radeon_cs_parser *p,  u32 idx,
 		return -EINVAL;
 	}
 	format = G_038004_DATA_FORMAT(word1);
-	if (!r600_fmt_is_valid_texture(format, p->family)) {
+	if (!fmt_is_valid_texture(format, p->family)) {
 		dev_warn(p->dev, "%s:%d texture invalid format %d\n",
 			 __func__, __LINE__, format);
 		return -EINVAL;
@@ -1466,7 +1346,7 @@ static int r600_check_texture_resource(struct radeon_cs_parser *p,  u32 idx,
 	array_check.nbanks = track->nbanks;
 	array_check.npipes = track->npipes;
 	array_check.nsamples = 1;
-	array_check.blocksize = r600_fmt_get_blocksize(format);
+	array_check.blocksize = fmt_get_blocksize(format);
 	if (r600_get_array_mode_alignment(&array_check,
 					  &pitch_align, &height_align, &depth_align, &base_align)) {
 		dev_warn(p->dev, "%s:%d tex array mode (%d) invalid\n",
@@ -1499,10 +1379,6 @@ static int r600_check_texture_resource(struct radeon_cs_parser *p,  u32 idx,
 	word1 = radeon_get_ib_value(p, idx + 5);
 	blevel = G_038010_BASE_LEVEL(word0);
 	llevel = G_038014_LAST_LEVEL(word1);
-	if (blevel > llevel) {
-		dev_warn(p->dev, "texture blevel %d > llevel %d\n",
-			 blevel, llevel);
-	}
 	if (array == 1) {
 		barray = G_038014_BASE_ARRAY(word1);
 		larray = G_038014_LAST_ARRAY(word1);
@@ -1514,10 +1390,8 @@ static int r600_check_texture_resource(struct radeon_cs_parser *p,  u32 idx,
 			  &l0_size, &mipmap_size);
 	/* using get ib will give us the offset into the texture bo */
 	if ((l0_size + word2) > radeon_bo_size(texture)) {
-		dev_warn(p->dev, "texture bo too small ((%d %d) (%d %d) %d %d %d -> %d have %ld)\n",
-			 w0, h0, pitch_align, height_align,
-			 array_check.array_mode, format, word2,
-			 l0_size, radeon_bo_size(texture));
+		dev_warn(p->dev, "texture bo too small (%d %d %d %d -> %d have %ld)\n",
+			w0, h0, format, word2, l0_size, radeon_bo_size(texture));
 		dev_warn(p->dev, "alignments %d %d %d %lld\n", pitch, pitch_align, height_align, base_align);
 		return -EINVAL;
 	}
@@ -1530,22 +1404,6 @@ static int r600_check_texture_resource(struct radeon_cs_parser *p,  u32 idx,
 	return 0;
 }
 
-static bool r600_is_safe_reg(struct radeon_cs_parser *p, u32 reg, u32 idx)
-{
-	u32 m, i;
-
-	i = (reg >> 7);
-	if (i >= ARRAY_SIZE(r600_reg_safe_bm)) {
-		dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx);
-		return false;
-	}
-	m = 1 << ((reg >> 2) & 31);
-	if (!(r600_reg_safe_bm[i] & m))
-		return true;
-	dev_warn(p->dev, "forbidden register 0x%08x at %d\n", reg, idx);
-	return false;
-}
-
 static int r600_packet3_check(struct radeon_cs_parser *p,
 				struct radeon_cs_packet *pkt)
 {
@@ -1568,8 +1426,6 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 	{
 		int pred_op;
 		int tmp;
-		uint64_t offset;
-
 		if (pkt->count != 1) {
 			DRM_ERROR("bad SET PREDICATION\n");
 			return -EINVAL;
@@ -1593,12 +1449,8 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 			return -EINVAL;
 		}
 
-		offset = reloc->lobj.gpu_offset +
-		         (idx_value & 0xfffffff0) +
-		         ((u64)(tmp & 0xff) << 32);
-
-		ib[idx + 0] = offset;
-		ib[idx + 1] = (tmp & 0xffffff00) | (upper_32_bits(offset) & 0xff);
+		ib[idx + 0] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff);
+		ib[idx + 1] = tmp + (upper_32_bits(reloc->lobj.gpu_offset) & 0xff);
 	}
 	break;
 
@@ -1622,8 +1474,6 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 		}
 		break;
 	case PACKET3_DRAW_INDEX:
-	{
-		uint64_t offset;
 		if (pkt->count != 3) {
 			DRM_ERROR("bad DRAW_INDEX\n");
 			return -EINVAL;
@@ -1633,21 +1483,14 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad DRAW_INDEX\n");
 			return -EINVAL;
 		}
-
-		offset = reloc->lobj.gpu_offset +
-		         idx_value +
-		         ((u64)(radeon_get_ib_value(p, idx+1) & 0xff) << 32);
-
-		ib[idx+0] = offset;
-		ib[idx+1] = upper_32_bits(offset) & 0xff;
-
+		ib[idx+0] = idx_value + (u32)(reloc->lobj.gpu_offset & 0xffffffff);
+		ib[idx+1] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
 		r = r600_cs_track_check(p);
 		if (r) {
 			dev_warn(p->dev, "%s:%d invalid cmd stream\n", __func__, __LINE__);
 			return r;
 		}
 		break;
-	}
 	case PACKET3_DRAW_INDEX_AUTO:
 		if (pkt->count != 1) {
 			DRM_ERROR("bad DRAW_INDEX_AUTO\n");
@@ -1678,20 +1521,13 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 		}
 		/* bit 4 is reg (0) or mem (1) */
 		if (idx_value & 0x10) {
-			uint64_t offset;
-
 			r = r600_cs_packet_next_reloc(p, &reloc);
 			if (r) {
 				DRM_ERROR("bad WAIT_REG_MEM\n");
 				return -EINVAL;
 			}
-
-			offset = reloc->lobj.gpu_offset +
-			         (radeon_get_ib_value(p, idx+1) & 0xfffffff0) +
-			         ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32);
-
-			ib[idx+1] = (ib[idx+1] & 0x3) | (offset & 0xfffffff0);
-			ib[idx+2] = upper_32_bits(offset) & 0xff;
+			ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff);
+			ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
 		}
 		break;
 	case PACKET3_SURFACE_SYNC:
@@ -1716,25 +1552,16 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 			return -EINVAL;
 		}
 		if (pkt->count) {
-			uint64_t offset;
-
 			r = r600_cs_packet_next_reloc(p, &reloc);
 			if (r) {
 				DRM_ERROR("bad EVENT_WRITE\n");
 				return -EINVAL;
 			}
-			offset = reloc->lobj.gpu_offset +
-			         (radeon_get_ib_value(p, idx+1) & 0xfffffff8) +
-			         ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32);
-
-			ib[idx+1] = offset & 0xfffffff8;
-			ib[idx+2] = upper_32_bits(offset) & 0xff;
+			ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff);
+			ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
 		}
 		break;
 	case PACKET3_EVENT_WRITE_EOP:
-	{
-		uint64_t offset;
-
 		if (pkt->count != 4) {
 			DRM_ERROR("bad EVENT_WRITE_EOP\n");
 			return -EINVAL;
@@ -1744,15 +1571,9 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 			DRM_ERROR("bad EVENT_WRITE\n");
 			return -EINVAL;
 		}
-
-		offset = reloc->lobj.gpu_offset +
-		         (radeon_get_ib_value(p, idx+1) & 0xfffffffc) +
-		         ((u64)(radeon_get_ib_value(p, idx+2) & 0xff) << 32);
-
-		ib[idx+1] = offset & 0xfffffffc;
-		ib[idx+2] = (ib[idx+2] & 0xffffff00) | (upper_32_bits(offset) & 0xff);
+		ib[idx+1] += (u32)(reloc->lobj.gpu_offset & 0xffffffff);
+		ib[idx+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
 		break;
-	}
 	case PACKET3_SET_CONFIG_REG:
 		start_reg = (idx_value << 2) + PACKET3_SET_CONFIG_REG_OFFSET;
 		end_reg = 4 * pkt->count + start_reg - 4;
@@ -1811,12 +1632,10 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 					return -EINVAL;
 				}
 				base_offset = (u32)((reloc->lobj.gpu_offset >> 8) & 0xffffffff);
-				if (!(p->cs_flags & RADEON_CS_KEEP_TILING_FLAGS)) {
-					if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
-						ib[idx+1+(i*7)+0] |= S_038000_TILE_MODE(V_038000_ARRAY_2D_TILED_THIN1);
-					else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
-						ib[idx+1+(i*7)+0] |= S_038000_TILE_MODE(V_038000_ARRAY_1D_TILED_THIN1);
-				}
+				if (reloc->lobj.tiling_flags & RADEON_TILING_MACRO)
+					ib[idx+1+(i*7)+0] |= S_038000_TILE_MODE(V_038000_ARRAY_2D_TILED_THIN1);
+				else if (reloc->lobj.tiling_flags & RADEON_TILING_MICRO)
+					ib[idx+1+(i*7)+0] |= S_038000_TILE_MODE(V_038000_ARRAY_1D_TILED_THIN1);
 				texture = reloc->robj;
 				/* tex mip base */
 				r = r600_cs_packet_next_reloc(p, &reloc);
@@ -1837,8 +1656,6 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 				ib[idx+1+(i*7)+3] += mip_offset;
 				break;
 			case SQ_TEX_VTX_VALID_BUFFER:
-			{
-				uint64_t offset64;
 				/* vtx base */
 				r = r600_cs_packet_next_reloc(p, &reloc);
 				if (r) {
@@ -1851,15 +1668,11 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 					/* force size to size of the buffer */
 					dev_warn(p->dev, "vbo resource seems too big (%d) for the bo (%ld)\n",
 						 size + offset, radeon_bo_size(reloc->robj));
-					ib[idx+1+(i*7)+1] = radeon_bo_size(reloc->robj) - offset;
+					ib[idx+1+(i*7)+1] = radeon_bo_size(reloc->robj);
 				}
-
-				offset64 = reloc->lobj.gpu_offset + offset;
-				ib[idx+1+(i*8)+0] = offset64;
-				ib[idx+1+(i*8)+2] = (ib[idx+1+(i*8)+2] & 0xffffff00) |
-						    (upper_32_bits(offset64) & 0xff);
+				ib[idx+1+(i*7)+0] += (u32)((reloc->lobj.gpu_offset) & 0xffffffff);
+				ib[idx+1+(i*7)+2] += upper_32_bits(reloc->lobj.gpu_offset) & 0xff;
 				break;
-			}
 			case SQ_TEX_VTX_INVALID_TEXTURE:
 			case SQ_TEX_VTX_INVALID_BUFFER:
 			default:
@@ -1934,104 +1747,6 @@ static int r600_packet3_check(struct radeon_cs_parser *p,
 			return -EINVAL;
 		}
 		break;
-	case PACKET3_STRMOUT_BUFFER_UPDATE:
-		if (pkt->count != 4) {
-			DRM_ERROR("bad STRMOUT_BUFFER_UPDATE (invalid count)\n");
-			return -EINVAL;
-		}
-		/* Updating memory at DST_ADDRESS. */
-		if (idx_value & 0x1) {
-			u64 offset;
-			r = r600_cs_packet_next_reloc(p, &reloc);
-			if (r) {
-				DRM_ERROR("bad STRMOUT_BUFFER_UPDATE (missing dst reloc)\n");
-				return -EINVAL;
-			}
-			offset = radeon_get_ib_value(p, idx+1);
-			offset += ((u64)(radeon_get_ib_value(p, idx+2) & 0xff)) << 32;
-			if ((offset + 4) > radeon_bo_size(reloc->robj)) {
-				DRM_ERROR("bad STRMOUT_BUFFER_UPDATE dst bo too small: 0x%llx, 0x%lx\n",
-					  offset + 4, radeon_bo_size(reloc->robj));
-				return -EINVAL;
-			}
-			offset += reloc->lobj.gpu_offset;
-			ib[idx+1] = offset;
-			ib[idx+2] = upper_32_bits(offset) & 0xff;
-		}
-		/* Reading data from SRC_ADDRESS. */
-		if (((idx_value >> 1) & 0x3) == 2) {
-			u64 offset;
-			r = r600_cs_packet_next_reloc(p, &reloc);
-			if (r) {
-				DRM_ERROR("bad STRMOUT_BUFFER_UPDATE (missing src reloc)\n");
-				return -EINVAL;
-			}
-			offset = radeon_get_ib_value(p, idx+3);
-			offset += ((u64)(radeon_get_ib_value(p, idx+4) & 0xff)) << 32;
-			if ((offset + 4) > radeon_bo_size(reloc->robj)) {
-				DRM_ERROR("bad STRMOUT_BUFFER_UPDATE src bo too small: 0x%llx, 0x%lx\n",
-					  offset + 4, radeon_bo_size(reloc->robj));
-				return -EINVAL;
-			}
-			offset += reloc->lobj.gpu_offset;
-			ib[idx+3] = offset;
-			ib[idx+4] = upper_32_bits(offset) & 0xff;
-		}
-		break;
-	case PACKET3_COPY_DW:
-		if (pkt->count != 4) {
-			DRM_ERROR("bad COPY_DW (invalid count)\n");
-			return -EINVAL;
-		}
-		if (idx_value & 0x1) {
-			u64 offset;
-			/* SRC is memory. */
-			r = r600_cs_packet_next_reloc(p, &reloc);
-			if (r) {
-				DRM_ERROR("bad COPY_DW (missing src reloc)\n");
-				return -EINVAL;
-			}
-			offset = radeon_get_ib_value(p, idx+1);
-			offset += ((u64)(radeon_get_ib_value(p, idx+2) & 0xff)) << 32;
-			if ((offset + 4) > radeon_bo_size(reloc->robj)) {
-				DRM_ERROR("bad COPY_DW src bo too small: 0x%llx, 0x%lx\n",
-					  offset + 4, radeon_bo_size(reloc->robj));
-				return -EINVAL;
-			}
-			offset += reloc->lobj.gpu_offset;
-			ib[idx+1] = offset;
-			ib[idx+2] = upper_32_bits(offset) & 0xff;
-		} else {
-			/* SRC is a reg. */
-			reg = radeon_get_ib_value(p, idx+1) << 2;
-			if (!r600_is_safe_reg(p, reg, idx+1))
-				return -EINVAL;
-		}
-		if (idx_value & 0x2) {
-			u64 offset;
-			/* DST is memory. */
-			r = r600_cs_packet_next_reloc(p, &reloc);
-			if (r) {
-				DRM_ERROR("bad COPY_DW (missing dst reloc)\n");
-				return -EINVAL;
-			}
-			offset = radeon_get_ib_value(p, idx+3);
-			offset += ((u64)(radeon_get_ib_value(p, idx+4) & 0xff)) << 32;
-			if ((offset + 4) > radeon_bo_size(reloc->robj)) {
-				DRM_ERROR("bad COPY_DW dst bo too small: 0x%llx, 0x%lx\n",
-					  offset + 4, radeon_bo_size(reloc->robj));
-				return -EINVAL;
-			}
-			offset += reloc->lobj.gpu_offset;
-			ib[idx+3] = offset;
-			ib[idx+4] = upper_32_bits(offset) & 0xff;
-		} else {
-			/* DST is a reg. */
-			reg = radeon_get_ib_value(p, idx+3) << 2;
-			if (!r600_is_safe_reg(p, reg, idx+3))
-				return -EINVAL;
-		}
-		break;
 	case PACKET3_NOP:
 		break;
 	default:
diff --git a/drivers/gpu/drm/radeon/r600_hdmi.c b/drivers/gpu/drm/radeon/r600_hdmi.c
index 0b59206..c45d921 100644
--- a/drivers/gpu/drm/radeon/r600_hdmi.c
+++ b/drivers/gpu/drm/radeon/r600_hdmi.c
@@ -320,7 +320,7 @@ void r600_hdmi_setmode(struct drm_encoder *encoder, struct drm_display_mode *mod
 	struct radeon_device *rdev = dev->dev_private;
 	uint32_t offset = to_radeon_encoder(encoder)->hdmi_offset;
 
-	if (ASIC_IS_DCE5(rdev))
+	if (ASIC_IS_DCE4(rdev))
 		return;
 
 	if (!offset)
@@ -462,31 +462,13 @@ static void r600_hdmi_assign_block(struct drm_encoder *encoder)
 	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
 	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
 
-	u16 eg_offsets[] = {
-		EVERGREEN_CRTC0_REGISTER_OFFSET,
-		EVERGREEN_CRTC1_REGISTER_OFFSET,
-		EVERGREEN_CRTC2_REGISTER_OFFSET,
-		EVERGREEN_CRTC3_REGISTER_OFFSET,
-		EVERGREEN_CRTC4_REGISTER_OFFSET,
-		EVERGREEN_CRTC5_REGISTER_OFFSET,
-	};
-
 	if (!dig) {
 		dev_err(rdev->dev, "Enabling HDMI on non-dig encoder\n");
 		return;
 	}
 
-	if (ASIC_IS_DCE5(rdev)) {
+	if (ASIC_IS_DCE4(rdev)) {
 		/* TODO */
-	} else if (ASIC_IS_DCE4(rdev)) {
-		if (dig->dig_encoder >= ARRAY_SIZE(eg_offsets)) {
-			dev_err(rdev->dev, "Enabling HDMI on unknown dig\n");
-			return;
-		}
-		radeon_encoder->hdmi_offset = EVERGREEN_HDMI_BASE +
-						eg_offsets[dig->dig_encoder];
-		radeon_encoder->hdmi_config_offset = radeon_encoder->hdmi_offset
-						+ EVERGREEN_HDMI_CONFIG_OFFSET;
 	} else if (ASIC_IS_DCE3(rdev)) {
 		radeon_encoder->hdmi_offset = dig->dig_encoder ?
 			R600_HDMI_BLOCK3 : R600_HDMI_BLOCK1;
@@ -509,7 +491,7 @@ void r600_hdmi_enable(struct drm_encoder *encoder)
 	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
 	uint32_t offset;
 
-	if (ASIC_IS_DCE5(rdev))
+	if (ASIC_IS_DCE4(rdev))
 		return;
 
 	if (!radeon_encoder->hdmi_offset) {
@@ -522,24 +504,16 @@ void r600_hdmi_enable(struct drm_encoder *encoder)
 	}
 
 	offset = radeon_encoder->hdmi_offset;
-	if (ASIC_IS_DCE5(rdev)) {
-		/* TODO */
-	} else if (ASIC_IS_DCE4(rdev)) {
-		WREG32_P(radeon_encoder->hdmi_config_offset + 0xc, 0x1, ~0x1);
-	} else if (ASIC_IS_DCE32(rdev)) {
+	if (ASIC_IS_DCE32(rdev) && !ASIC_IS_DCE4(rdev)) {
 		WREG32_P(radeon_encoder->hdmi_config_offset + 0x4, 0x1, ~0x1);
-	} else if (ASIC_IS_DCE3(rdev)) {
-		/* TODO */
-	} else if (rdev->family >= CHIP_R600) {
+	} else if (rdev->family >= CHIP_R600 && !ASIC_IS_DCE3(rdev)) {
 		switch (radeon_encoder->encoder_id) {
 		case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
-			WREG32_P(AVIVO_TMDSA_CNTL, AVIVO_TMDSA_CNTL_HDMI_EN,
-				 ~AVIVO_TMDSA_CNTL_HDMI_EN);
+			WREG32_P(AVIVO_TMDSA_CNTL, 0x4, ~0x4);
 			WREG32(offset + R600_HDMI_ENABLE, 0x101);
 			break;
 		case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
-			WREG32_P(AVIVO_LVTMA_CNTL, AVIVO_LVTMA_CNTL_HDMI_EN,
-				 ~AVIVO_LVTMA_CNTL_HDMI_EN);
+			WREG32_P(AVIVO_LVTMA_CNTL, 0x4, ~0x4);
 			WREG32(offset + R600_HDMI_ENABLE, 0x105);
 			break;
 		default:
@@ -551,8 +525,8 @@ void r600_hdmi_enable(struct drm_encoder *encoder)
 	if (rdev->irq.installed
 	    && rdev->family != CHIP_RS600
 	    && rdev->family != CHIP_RS690
-	    && rdev->family != CHIP_RS740
-	    && !ASIC_IS_DCE4(rdev)) {
+	    && rdev->family != CHIP_RS740) {
+
 		/* if irq is available use it */
 		rdev->irq.hdmi[offset == R600_HDMI_BLOCK1 ? 0 : 1] = true;
 		radeon_irq_set(rdev);
@@ -577,7 +551,7 @@ void r600_hdmi_disable(struct drm_encoder *encoder)
 	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
 	uint32_t offset;
 
-	if (ASIC_IS_DCE5(rdev))
+	if (ASIC_IS_DCE4(rdev))
 		return;
 
 	offset = radeon_encoder->hdmi_offset;
@@ -596,22 +570,16 @@ void r600_hdmi_disable(struct drm_encoder *encoder)
 	/* disable polling */
 	r600_audio_disable_polling(encoder);
 
-	if (ASIC_IS_DCE5(rdev)) {
-		/* TODO */
-	} else if (ASIC_IS_DCE4(rdev)) {
-		WREG32_P(radeon_encoder->hdmi_config_offset + 0xc, 0, ~0x1);
-	} else if (ASIC_IS_DCE32(rdev)) {
+	if (ASIC_IS_DCE32(rdev) && !ASIC_IS_DCE4(rdev)) {
 		WREG32_P(radeon_encoder->hdmi_config_offset + 0x4, 0, ~0x1);
 	} else if (rdev->family >= CHIP_R600 && !ASIC_IS_DCE3(rdev)) {
 		switch (radeon_encoder->encoder_id) {
 		case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
-			WREG32_P(AVIVO_TMDSA_CNTL, 0,
-				 ~AVIVO_TMDSA_CNTL_HDMI_EN);
+			WREG32_P(AVIVO_TMDSA_CNTL, 0, ~0x4);
 			WREG32(offset + R600_HDMI_ENABLE, 0);
 			break;
 		case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
-			WREG32_P(AVIVO_LVTMA_CNTL, 0,
-				 ~AVIVO_LVTMA_CNTL_HDMI_EN);
+			WREG32_P(AVIVO_LVTMA_CNTL, 0, ~0x4);
 			WREG32(offset + R600_HDMI_ENABLE, 0);
 			break;
 		default:
diff --git a/drivers/gpu/drm/radeon/r600d.h b/drivers/gpu/drm/radeon/r600d.h
index 8ae328f..0245ae6 100644
--- a/drivers/gpu/drm/radeon/r600d.h
+++ b/drivers/gpu/drm/radeon/r600d.h
@@ -78,26 +78,7 @@
 
 #define CB_COLOR0_SIZE                                  0x28060
 #define CB_COLOR0_VIEW                                  0x28080
-#define R_028080_CB_COLOR0_VIEW                      0x028080
-#define   S_028080_SLICE_START(x)                      (((x) & 0x7FF) << 0)
-#define   G_028080_SLICE_START(x)                      (((x) >> 0) & 0x7FF)
-#define   C_028080_SLICE_START                         0xFFFFF800
-#define   S_028080_SLICE_MAX(x)                        (((x) & 0x7FF) << 13)
-#define   G_028080_SLICE_MAX(x)                        (((x) >> 13) & 0x7FF)
-#define   C_028080_SLICE_MAX                           0xFF001FFF
-#define R_028084_CB_COLOR1_VIEW                      0x028084
-#define R_028088_CB_COLOR2_VIEW                      0x028088
-#define R_02808C_CB_COLOR3_VIEW                      0x02808C
-#define R_028090_CB_COLOR4_VIEW                      0x028090
-#define R_028094_CB_COLOR5_VIEW                      0x028094
-#define R_028098_CB_COLOR6_VIEW                      0x028098
-#define R_02809C_CB_COLOR7_VIEW                      0x02809C
 #define CB_COLOR0_INFO                                  0x280a0
-#	define CB_FORMAT(x)				((x) << 2)
-#       define CB_ARRAY_MODE(x)                         ((x) << 8)
-#	define CB_SOURCE_FORMAT(x)			((x) << 27)
-#	define CB_SF_EXPORT_FULL			0
-#	define CB_SF_EXPORT_NORM			1
 #define CB_COLOR0_TILE                                  0x280c0
 #define CB_COLOR0_FRAG                                  0x280e0
 #define CB_COLOR0_MASK                                  0x28100
@@ -436,17 +417,6 @@
 #define	SQ_PGM_START_VS					0x28858
 #define SQ_PGM_RESOURCES_VS                             0x28868
 #define SQ_PGM_CF_OFFSET_VS                             0x288d0
-
-#define SQ_VTX_CONSTANT_WORD0_0				0x30000
-#define SQ_VTX_CONSTANT_WORD1_0				0x30004
-#define SQ_VTX_CONSTANT_WORD2_0				0x30008
-#	define SQ_VTXC_BASE_ADDR_HI(x)			((x) << 0)
-#	define SQ_VTXC_STRIDE(x)			((x) << 8)
-#	define SQ_VTXC_ENDIAN_SWAP(x)			((x) << 30)
-#	define SQ_ENDIAN_NONE				0
-#	define SQ_ENDIAN_8IN16				1
-#	define SQ_ENDIAN_8IN32				2
-#define SQ_VTX_CONSTANT_WORD3_0				0x3000c
 #define	SQ_VTX_CONSTANT_WORD6_0				0x38018
 #define		S__SQ_VTX_CONSTANT_TYPE(x)			(((x) & 3) << 30)
 #define		G__SQ_VTX_CONSTANT_TYPE(x)			(((x) >> 30) & 3)
@@ -507,11 +477,6 @@
 #define	VGT_STRMOUT_BUFFER_OFFSET_1			0x28AEC
 #define	VGT_STRMOUT_BUFFER_OFFSET_2			0x28AFC
 #define	VGT_STRMOUT_BUFFER_OFFSET_3			0x28B0C
-#define VGT_STRMOUT_BUFFER_SIZE_0			0x28AD0
-#define VGT_STRMOUT_BUFFER_SIZE_1			0x28AE0
-#define VGT_STRMOUT_BUFFER_SIZE_2			0x28AF0
-#define VGT_STRMOUT_BUFFER_SIZE_3			0x28B00
-
 #define	VGT_STRMOUT_EN					0x28AB0
 #define	VGT_VERTEX_REUSE_BLOCK_CNTL			0x28C58
 #define		VTX_REUSE_DEPTH_MASK				0x000000FF
@@ -850,11 +815,7 @@
 #define	PACKET3_STRMOUT_BUFFER_UPDATE			0x34
 #define	PACKET3_INDIRECT_BUFFER_MP			0x38
 #define	PACKET3_MEM_SEMAPHORE				0x39
-#              define PACKET3_SEM_WAIT_ON_SIGNAL    (0x1 << 12)
-#              define PACKET3_SEM_SEL_SIGNAL	    (0x6 << 29)
-#              define PACKET3_SEM_SEL_WAIT	    (0x7 << 29)
 #define	PACKET3_MPEG_INDEX				0x3A
-#define	PACKET3_COPY_DW					0x3B
 #define	PACKET3_WAIT_REG_MEM				0x3C
 #define	PACKET3_MEM_WRITE				0x3D
 #define	PACKET3_INDIRECT_BUFFER				0x32
@@ -1391,12 +1352,6 @@
 #define   S_038010_DST_SEL_W(x)                        (((x) & 0x7) << 25)
 #define   G_038010_DST_SEL_W(x)                        (((x) >> 25) & 0x7)
 #define   C_038010_DST_SEL_W                           0xF1FFFFFF
-#	define SQ_SEL_X					0
-#	define SQ_SEL_Y					1
-#	define SQ_SEL_Z					2
-#	define SQ_SEL_W					3
-#	define SQ_SEL_0					4
-#	define SQ_SEL_1					5
 #define   S_038010_BASE_LEVEL(x)                       (((x) & 0xF) << 28)
 #define   G_038010_BASE_LEVEL(x)                       (((x) >> 28) & 0xF)
 #define   C_038010_BASE_LEVEL                          0x0FFFFFFF
diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index d2870a0..59d72d0 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -60,7 +60,7 @@
  *                          are considered as fatal)
  */
 
-#include <linux/atomic.h>
+#include <asm/atomic.h>
 #include <linux/wait.h>
 #include <linux/list.h>
 #include <linux/kref.h>
@@ -103,25 +103,10 @@ extern int radeon_msi;
 #define RADEON_FENCE_JIFFIES_TIMEOUT	(HZ / 2)
 /* RADEON_IB_POOL_SIZE must be a power of 2 */
 #define RADEON_IB_POOL_SIZE		16
-#define RADEON_DEBUGFS_MAX_COMPONENTS	32
+#define RADEON_DEBUGFS_MAX_NUM_FILES	32
 #define RADEONFB_CONN_LIMIT		4
 #define RADEON_BIOS_NUM_SCRATCH		8
 
-/* max number of rings */
-#define RADEON_NUM_RINGS 3
-
-/* internal ring indices */
-/* r1xx+ has gfx CP ring */
-#define RADEON_RING_TYPE_GFX_INDEX  0
-
-/* cayman has 2 compute CP rings */
-#define CAYMAN_RING_TYPE_CP1_INDEX 1
-#define CAYMAN_RING_TYPE_CP2_INDEX 2
-
-/* hardcode those limit for now */
-#define RADEON_VA_RESERVED_SIZE		(8 << 20)
-#define RADEON_IB_VM_MAX_SIZE		(64 << 10)
-
 /*
  * Errata workarounds.
  */
@@ -157,47 +142,6 @@ bool radeon_get_bios(struct radeon_device *rdev);
 
 
 /*
- * Mutex which allows recursive locking from the same process.
- */
-struct radeon_mutex {
-	struct mutex		mutex;
-	struct task_struct	*owner;
-	int			level;
-};
-
-static inline void radeon_mutex_init(struct radeon_mutex *mutex)
-{
-	mutex_init(&mutex->mutex);
-	mutex->owner = NULL;
-	mutex->level = 0;
-}
-
-static inline void radeon_mutex_lock(struct radeon_mutex *mutex)
-{
-	if (mutex_trylock(&mutex->mutex)) {
-		/* The mutex was unlocked before, so it's ours now */
-		mutex->owner = current;
-	} else if (mutex->owner != current) {
-		/* Another process locked the mutex, take it */
-		mutex_lock(&mutex->mutex);
-		mutex->owner = current;
-	}
-	/* Otherwise the mutex was already locked by this process */
-
-	mutex->level++;
-}
-
-static inline void radeon_mutex_unlock(struct radeon_mutex *mutex)
-{
-	if (--mutex->level > 0)
-		return;
-
-	mutex->owner = NULL;
-	mutex_unlock(&mutex->mutex);
-}
-
-
-/*
  * Dummy page
  */
 struct radeon_dummy_page {
@@ -242,24 +186,20 @@ extern int rv6xx_get_temp(struct radeon_device *rdev);
 extern int rv770_get_temp(struct radeon_device *rdev);
 extern int evergreen_get_temp(struct radeon_device *rdev);
 extern int sumo_get_temp(struct radeon_device *rdev);
-extern void evergreen_tiling_fields(unsigned tiling_flags, unsigned *bankw,
-				    unsigned *bankh, unsigned *mtaspect,
-				    unsigned *tile_split);
 
 /*
  * Fences.
  */
 struct radeon_fence_driver {
 	uint32_t			scratch_reg;
-	uint64_t			gpu_addr;
-	volatile uint32_t		*cpu_addr;
 	atomic_t			seq;
 	uint32_t			last_seq;
 	unsigned long			last_jiffies;
 	unsigned long			last_timeout;
 	wait_queue_head_t		queue;
+	rwlock_t			lock;
 	struct list_head		created;
-	struct list_head		emitted;
+	struct list_head		emited;
 	struct list_head		signaled;
 	bool				initialized;
 };
@@ -270,26 +210,21 @@ struct radeon_fence {
 	struct list_head		list;
 	/* protected by radeon_fence.lock */
 	uint32_t			seq;
-	bool				emitted;
+	bool				emited;
 	bool				signaled;
-	/* RB, DMA, etc. */
-	int				ring;
-	struct radeon_semaphore		*semaphore;
 };
 
-int radeon_fence_driver_start_ring(struct radeon_device *rdev, int ring);
 int radeon_fence_driver_init(struct radeon_device *rdev);
 void radeon_fence_driver_fini(struct radeon_device *rdev);
-int radeon_fence_create(struct radeon_device *rdev, struct radeon_fence **fence, int ring);
+int radeon_fence_create(struct radeon_device *rdev, struct radeon_fence **fence);
 int radeon_fence_emit(struct radeon_device *rdev, struct radeon_fence *fence);
-void radeon_fence_process(struct radeon_device *rdev, int ring);
+void radeon_fence_process(struct radeon_device *rdev);
 bool radeon_fence_signaled(struct radeon_fence *fence);
 int radeon_fence_wait(struct radeon_fence *fence, bool interruptible);
-int radeon_fence_wait_next(struct radeon_device *rdev, int ring);
-int radeon_fence_wait_last(struct radeon_device *rdev, int ring);
+int radeon_fence_wait_next(struct radeon_device *rdev);
+int radeon_fence_wait_last(struct radeon_device *rdev);
 struct radeon_fence *radeon_fence_ref(struct radeon_fence *fence);
 void radeon_fence_unref(struct radeon_fence **fence);
-int radeon_fence_count_emitted(struct radeon_device *rdev, int ring);
 
 /*
  * Tiling registers
@@ -311,21 +246,6 @@ struct radeon_mman {
 	bool				initialized;
 };
 
-/* bo virtual address in a specific vm */
-struct radeon_bo_va {
-	/* bo list is protected by bo being reserved */
-	struct list_head		bo_list;
-	/* vm list is protected by vm mutex */
-	struct list_head		vm_list;
-	/* constant after initialization */
-	struct radeon_vm		*vm;
-	struct radeon_bo		*bo;
-	uint64_t			soffset;
-	uint64_t			eoffset;
-	uint32_t			flags;
-	bool				valid;
-};
-
 struct radeon_bo {
 	/* Protected by gem.mutex */
 	struct list_head		list;
@@ -339,10 +259,6 @@ struct radeon_bo {
 	u32				tiling_flags;
 	u32				pitch;
 	int				surface_reg;
-	/* list of all virtual address to which this bo
-	 * is associated to
-	 */
-	struct list_head		va;
 	/* Constant after initialization */
 	struct radeon_device		*rdev;
 	struct drm_gem_object		gem_base;
@@ -358,48 +274,6 @@ struct radeon_bo_list {
 	u32			tiling_flags;
 };
 
-/* sub-allocation manager, it has to be protected by another lock.
- * By conception this is an helper for other part of the driver
- * like the indirect buffer or semaphore, which both have their
- * locking.
- *
- * Principe is simple, we keep a list of sub allocation in offset
- * order (first entry has offset == 0, last entry has the highest
- * offset).
- *
- * When allocating new object we first check if there is room at
- * the end total_size - (last_object_offset + last_object_size) >=
- * alloc_size. If so we allocate new object there.
- *
- * When there is not enough room at the end, we start waiting for
- * each sub object until we reach object_offset+object_size >=
- * alloc_size, this object then become the sub object we return.
- *
- * Alignment can't be bigger than page size.
- *
- * Hole are not considered for allocation to keep things simple.
- * Assumption is that there won't be hole (all object on same
- * alignment).
- */
-struct radeon_sa_manager {
-	struct radeon_bo	*bo;
-	struct list_head	sa_bo;
-	unsigned		size;
-	uint64_t		gpu_addr;
-	void			*cpu_ptr;
-	uint32_t		domain;
-};
-
-struct radeon_sa_bo;
-
-/* sub-allocation buffer */
-struct radeon_sa_bo {
-	struct list_head		list;
-	struct radeon_sa_manager	*manager;
-	unsigned			offset;
-	unsigned			size;
-};
-
 /*
  * GEM objects.
  */
@@ -414,6 +288,9 @@ int radeon_gem_object_create(struct radeon_device *rdev, int size,
 				int alignment, int initial_domain,
 				bool discardable, bool kernel,
 				struct drm_gem_object **obj);
+int radeon_gem_object_pin(struct drm_gem_object *obj, uint32_t pin_domain,
+			  uint64_t *gpu_addr);
+void radeon_gem_object_unpin(struct drm_gem_object *obj);
 
 int radeon_mode_dumb_create(struct drm_file *file_priv,
 			    struct drm_device *dev,
@@ -426,64 +303,37 @@ int radeon_mode_dumb_destroy(struct drm_file *file_priv,
 			     uint32_t handle);
 
 /*
- * Semaphores.
+ * GART structures, functions & helpers
  */
-struct radeon_ring;
-
-#define	RADEON_SEMAPHORE_BO_SIZE	256
+struct radeon_mc;
 
-struct radeon_semaphore_driver {
-	rwlock_t			lock;
-	struct list_head		bo;
+struct radeon_gart_table_ram {
+	volatile uint32_t		*ptr;
 };
 
-struct radeon_semaphore_bo;
-
-/* everything here is constant */
-struct radeon_semaphore {
-	struct list_head		list;
-	uint64_t			gpu_addr;
-	uint32_t			*cpu_ptr;
-	struct radeon_semaphore_bo	*bo;
+struct radeon_gart_table_vram {
+	struct radeon_bo		*robj;
+	volatile uint32_t		*ptr;
 };
 
-struct radeon_semaphore_bo {
-	struct list_head		list;
-	struct radeon_ib		*ib;
-	struct list_head		free;
-	struct radeon_semaphore		semaphores[RADEON_SEMAPHORE_BO_SIZE/8];
-	unsigned			nused;
+union radeon_gart_table {
+	struct radeon_gart_table_ram	ram;
+	struct radeon_gart_table_vram	vram;
 };
 
-void radeon_semaphore_driver_fini(struct radeon_device *rdev);
-int radeon_semaphore_create(struct radeon_device *rdev,
-			    struct radeon_semaphore **semaphore);
-void radeon_semaphore_emit_signal(struct radeon_device *rdev, int ring,
-				  struct radeon_semaphore *semaphore);
-void radeon_semaphore_emit_wait(struct radeon_device *rdev, int ring,
-				struct radeon_semaphore *semaphore);
-void radeon_semaphore_free(struct radeon_device *rdev,
-			   struct radeon_semaphore *semaphore);
-
-/*
- * GART structures, functions & helpers
- */
-struct radeon_mc;
-
 #define RADEON_GPU_PAGE_SIZE 4096
 #define RADEON_GPU_PAGE_MASK (RADEON_GPU_PAGE_SIZE - 1)
 #define RADEON_GPU_PAGE_SHIFT 12
-#define RADEON_GPU_PAGE_ALIGN(a) (((a) + RADEON_GPU_PAGE_MASK) & ~RADEON_GPU_PAGE_MASK)
 
 struct radeon_gart {
 	dma_addr_t			table_addr;
-	struct radeon_bo		*robj;
-	void				*ptr;
 	unsigned			num_gpu_pages;
 	unsigned			num_cpu_pages;
 	unsigned			table_size;
+	union radeon_gart_table		table;
 	struct page			**pages;
 	dma_addr_t			*pages_addr;
+	bool				*ttm_alloced;
 	bool				ready;
 };
 
@@ -491,8 +341,6 @@ int radeon_gart_table_ram_alloc(struct radeon_device *rdev);
 void radeon_gart_table_ram_free(struct radeon_device *rdev);
 int radeon_gart_table_vram_alloc(struct radeon_device *rdev);
 void radeon_gart_table_vram_free(struct radeon_device *rdev);
-int radeon_gart_table_vram_pin(struct radeon_device *rdev);
-void radeon_gart_table_vram_unpin(struct radeon_device *rdev);
 int radeon_gart_init(struct radeon_device *rdev);
 void radeon_gart_fini(struct radeon_device *rdev);
 void radeon_gart_unbind(struct radeon_device *rdev, unsigned offset,
@@ -500,7 +348,6 @@ void radeon_gart_unbind(struct radeon_device *rdev, unsigned offset,
 int radeon_gart_bind(struct radeon_device *rdev, unsigned offset,
 		     int pages, struct page **pagelist,
 		     dma_addr_t *dma_addr);
-void radeon_gart_restore(struct radeon_device *rdev);
 
 
 /*
@@ -591,47 +438,45 @@ union radeon_irq_stat_regs {
 	struct evergreen_irq_stat_regs evergreen;
 };
 
-#define RADEON_MAX_HPD_PINS 6
-#define RADEON_MAX_CRTCS 6
-#define RADEON_MAX_HDMI_BLOCKS 2
-
 struct radeon_irq {
 	bool		installed;
-	bool		sw_int[RADEON_NUM_RINGS];
-	bool		crtc_vblank_int[RADEON_MAX_CRTCS];
-	bool		pflip[RADEON_MAX_CRTCS];
+	bool		sw_int;
+	/* FIXME: use a define max crtc rather than hardcode it */
+	bool		crtc_vblank_int[6];
+	bool		pflip[6];
 	wait_queue_head_t	vblank_queue;
-	bool            hpd[RADEON_MAX_HPD_PINS];
+	/* FIXME: use defines for max hpd/dacs */
+	bool            hpd[6];
 	bool            gui_idle;
 	bool            gui_idle_acked;
 	wait_queue_head_t	idle_queue;
-	bool		hdmi[RADEON_MAX_HDMI_BLOCKS];
+	/* FIXME: use defines for max HDMI blocks */
+	bool		hdmi[2];
 	spinlock_t sw_lock;
-	int sw_refcount[RADEON_NUM_RINGS];
+	int sw_refcount;
 	union radeon_irq_stat_regs stat_regs;
-	spinlock_t pflip_lock[RADEON_MAX_CRTCS];
-	int pflip_refcount[RADEON_MAX_CRTCS];
+	spinlock_t pflip_lock[6];
+	int pflip_refcount[6];
 };
 
 int radeon_irq_kms_init(struct radeon_device *rdev);
 void radeon_irq_kms_fini(struct radeon_device *rdev);
-void radeon_irq_kms_sw_irq_get(struct radeon_device *rdev, int ring);
-void radeon_irq_kms_sw_irq_put(struct radeon_device *rdev, int ring);
+void radeon_irq_kms_sw_irq_get(struct radeon_device *rdev);
+void radeon_irq_kms_sw_irq_put(struct radeon_device *rdev);
 void radeon_irq_kms_pflip_irq_get(struct radeon_device *rdev, int crtc);
 void radeon_irq_kms_pflip_irq_put(struct radeon_device *rdev, int crtc);
 
 /*
- * CP & rings.
+ * CP & ring.
  */
-
 struct radeon_ib {
-	struct radeon_sa_bo	sa_bo;
+	struct list_head	list;
 	unsigned		idx;
-	uint32_t		length_dw;
 	uint64_t		gpu_addr;
-	uint32_t		*ptr;
 	struct radeon_fence	*fence;
-	unsigned		vm_id;
+	uint32_t		*ptr;
+	uint32_t		length_dw;
+	bool			free;
 };
 
 /*
@@ -639,22 +484,20 @@ struct radeon_ib {
  * mutex protects scheduled_ibs, ready, alloc_bm
  */
 struct radeon_ib_pool {
-	struct radeon_mutex		mutex;
-	struct radeon_sa_manager	sa_manager;
-	struct radeon_ib		ibs[RADEON_IB_POOL_SIZE];
-	bool				ready;
-	unsigned			head_id;
+	struct mutex		mutex;
+	struct radeon_bo	*robj;
+	struct list_head	bogus_ib;
+	struct radeon_ib	ibs[RADEON_IB_POOL_SIZE];
+	bool			ready;
+	unsigned		head_id;
 };
 
-struct radeon_ring {
+struct radeon_cp {
 	struct radeon_bo	*ring_obj;
 	volatile uint32_t	*ring;
 	unsigned		rptr;
-	unsigned		rptr_offs;
-	unsigned		rptr_reg;
 	unsigned		wptr;
 	unsigned		wptr_old;
-	unsigned		wptr_reg;
 	unsigned		ring_size;
 	unsigned		ring_free_dw;
 	int			count_dw;
@@ -663,61 +506,6 @@ struct radeon_ring {
 	uint32_t		ptr_mask;
 	struct mutex		mutex;
 	bool			ready;
-	u32			ptr_reg_shift;
-	u32			ptr_reg_mask;
-	u32			nop;
-};
-
-/*
- * VM
- */
-struct radeon_vm {
-	struct list_head		list;
-	struct list_head		va;
-	int				id;
-	unsigned			last_pfn;
-	u64				pt_gpu_addr;
-	u64				*pt;
-	struct radeon_sa_bo		sa_bo;
-	struct mutex			mutex;
-	/* last fence for cs using this vm */
-	struct radeon_fence		*fence;
-};
-
-struct radeon_vm_funcs {
-	int (*init)(struct radeon_device *rdev);
-	void (*fini)(struct radeon_device *rdev);
-	/* cs mutex must be lock for schedule_ib */
-	int (*bind)(struct radeon_device *rdev, struct radeon_vm *vm, int id);
-	void (*unbind)(struct radeon_device *rdev, struct radeon_vm *vm);
-	void (*tlb_flush)(struct radeon_device *rdev, struct radeon_vm *vm);
-	uint32_t (*page_flags)(struct radeon_device *rdev,
-			       struct radeon_vm *vm,
-			       uint32_t flags);
-	void (*set_page)(struct radeon_device *rdev, struct radeon_vm *vm,
-			unsigned pfn, uint64_t addr, uint32_t flags);
-};
-
-struct radeon_vm_manager {
-	struct list_head		lru_vm;
-	uint32_t			use_bitmap;
-	struct radeon_sa_manager	sa_manager;
-	uint32_t			max_pfn;
-	/* fields constant after init */
-	const struct radeon_vm_funcs	*funcs;
-	/* number of VMIDs */
-	unsigned			nvm;
-	/* vram base address for page table entry  */
-	u64				vram_base_offset;
-	/* is vm enabled? */
-	bool				enabled;
-};
-
-/*
- * file private structure
- */
-struct radeon_fpriv {
-	struct radeon_vm		vm;
 };
 
 /*
@@ -727,7 +515,6 @@ struct r600_ih {
 	struct radeon_bo	*ring_obj;
 	volatile uint32_t	*ring;
 	unsigned		rptr;
-	unsigned		rptr_offs;
 	unsigned		wptr;
 	unsigned		wptr_old;
 	unsigned		ring_size;
@@ -737,30 +524,9 @@ struct r600_ih {
 	bool                    enabled;
 };
 
-struct r600_blit_cp_primitives {
-	void (*set_render_target)(struct radeon_device *rdev, int format,
-				  int w, int h, u64 gpu_addr);
-	void (*cp_set_surface_sync)(struct radeon_device *rdev,
-				    u32 sync_type, u32 size,
-				    u64 mc_addr);
-	void (*set_shaders)(struct radeon_device *rdev);
-	void (*set_vtx_resource)(struct radeon_device *rdev, u64 gpu_addr);
-	void (*set_tex_resource)(struct radeon_device *rdev,
-				 int format, int w, int h, int pitch,
-				 u64 gpu_addr, u32 size);
-	void (*set_scissors)(struct radeon_device *rdev, int x1, int y1,
-			     int x2, int y2);
-	void (*draw_auto)(struct radeon_device *rdev);
-	void (*set_default_state)(struct radeon_device *rdev);
-};
-
 struct r600_blit {
 	struct mutex		mutex;
 	struct radeon_bo	*shader_obj;
-	struct r600_blit_cp_primitives primitives;
-	int max_dim;
-	int ring_size_common;
-	int ring_size_per_loop;
 	u64 shader_gpu_addr;
 	u32 vs_offset, ps_offset;
 	u32 state_offset;
@@ -769,30 +535,23 @@ struct r600_blit {
 	struct radeon_ib *vb_ib;
 };
 
-void r600_blit_suspend(struct radeon_device *rdev);
-
-int radeon_ib_get(struct radeon_device *rdev, int ring,
-		  struct radeon_ib **ib, unsigned size);
+int radeon_ib_get(struct radeon_device *rdev, struct radeon_ib **ib);
 void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib **ib);
-bool radeon_ib_try_free(struct radeon_device *rdev, struct radeon_ib *ib);
 int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib);
 int radeon_ib_pool_init(struct radeon_device *rdev);
 void radeon_ib_pool_fini(struct radeon_device *rdev);
-int radeon_ib_pool_start(struct radeon_device *rdev);
-int radeon_ib_pool_suspend(struct radeon_device *rdev);
+int radeon_ib_test(struct radeon_device *rdev);
+extern void radeon_ib_bogus_add(struct radeon_device *rdev, struct radeon_ib *ib);
 /* Ring access between begin & end cannot sleep */
-int radeon_ring_index(struct radeon_device *rdev, struct radeon_ring *cp);
-void radeon_ring_free_size(struct radeon_device *rdev, struct radeon_ring *cp);
-int radeon_ring_alloc(struct radeon_device *rdev, struct radeon_ring *cp, unsigned ndw);
-int radeon_ring_lock(struct radeon_device *rdev, struct radeon_ring *cp, unsigned ndw);
-void radeon_ring_commit(struct radeon_device *rdev, struct radeon_ring *cp);
-void radeon_ring_unlock_commit(struct radeon_device *rdev, struct radeon_ring *cp);
-void radeon_ring_unlock_undo(struct radeon_device *rdev, struct radeon_ring *cp);
-int radeon_ring_test(struct radeon_device *rdev, struct radeon_ring *cp);
-int radeon_ring_init(struct radeon_device *rdev, struct radeon_ring *cp, unsigned ring_size,
-		     unsigned rptr_offs, unsigned rptr_reg, unsigned wptr_reg,
-		     u32 ptr_reg_shift, u32 ptr_reg_mask, u32 nop);
-void radeon_ring_fini(struct radeon_device *rdev, struct radeon_ring *cp);
+void radeon_ring_free_size(struct radeon_device *rdev);
+int radeon_ring_alloc(struct radeon_device *rdev, unsigned ndw);
+int radeon_ring_lock(struct radeon_device *rdev, unsigned ndw);
+void radeon_ring_commit(struct radeon_device *rdev);
+void radeon_ring_unlock_commit(struct radeon_device *rdev);
+void radeon_ring_unlock_undo(struct radeon_device *rdev);
+int radeon_ring_test(struct radeon_device *rdev);
+int radeon_ring_init(struct radeon_device *rdev, unsigned ring_size);
+void radeon_ring_fini(struct radeon_device *rdev);
 
 
 /*
@@ -809,12 +568,12 @@ struct radeon_cs_reloc {
 struct radeon_cs_chunk {
 	uint32_t		chunk_id;
 	uint32_t		length_dw;
-	int			kpage_idx[2];
-	uint32_t		*kpage[2];
+	int kpage_idx[2];
+	uint32_t                *kpage[2];
 	uint32_t		*kdata;
-	void __user		*user_ptr;
-	int			last_copied_page;
-	int			last_page_index;
+	void __user *user_ptr;
+	int last_copied_page;
+	int last_page_index;
 };
 
 struct radeon_cs_parser {
@@ -835,19 +594,40 @@ struct radeon_cs_parser {
 	/* indices of various chunks */
 	int			chunk_ib_idx;
 	int			chunk_relocs_idx;
-	int			chunk_flags_idx;
 	struct radeon_ib	*ib;
 	void			*track;
 	unsigned		family;
-	int			parser_error;
-	u32			cs_flags;
-	u32			ring;
-	s32			priority;
+	int parser_error;
 };
 
 extern int radeon_cs_update_pages(struct radeon_cs_parser *p, int pg_idx);
 extern int radeon_cs_finish_pages(struct radeon_cs_parser *p);
-extern u32 radeon_get_ib_value(struct radeon_cs_parser *p, int idx);
+
+
+static inline u32 radeon_get_ib_value(struct radeon_cs_parser *p, int idx)
+{
+	struct radeon_cs_chunk *ibc = &p->chunks[p->chunk_ib_idx];
+	u32 pg_idx, pg_offset;
+	u32 idx_value = 0;
+	int new_page;
+
+	pg_idx = (idx * 4) / PAGE_SIZE;
+	pg_offset = (idx * 4) % PAGE_SIZE;
+
+	if (ibc->kpage_idx[0] == pg_idx)
+		return ibc->kpage[0][pg_offset/4];
+	if (ibc->kpage_idx[1] == pg_idx)
+		return ibc->kpage[1][pg_offset/4];
+
+	new_page = radeon_cs_update_pages(p, pg_idx);
+	if (new_page < 0) {
+		p->parser_error = new_page;
+		return 0;
+	}
+
+	idx_value = ibc->kpage[new_page][pg_offset/4];
+	return idx_value;
+}
 
 struct radeon_cs_packet {
 	unsigned	idx;
@@ -1015,7 +795,8 @@ struct radeon_pm_clock_info {
 
 struct radeon_power_state {
 	enum radeon_pm_state_type type;
-	struct radeon_pm_clock_info *clock_info;
+	/* XXX: use a define for num clock modes */
+	struct radeon_pm_clock_info clock_info[8];
 	/* number of valid clock modes in this power state */
 	int num_clock_modes;
 	struct radeon_pm_clock_info *default_clock_mode;
@@ -1085,34 +866,22 @@ struct radeon_pm {
 	struct device	        *int_hwmon_dev;
 };
 
-int radeon_pm_get_type_index(struct radeon_device *rdev,
-			     enum radeon_pm_state_type ps_type,
-			     int instance);
 
 /*
  * Benchmarking
  */
-void radeon_benchmark(struct radeon_device *rdev, int test_number);
+void radeon_benchmark(struct radeon_device *rdev);
 
 
 /*
  * Testing
  */
 void radeon_test_moves(struct radeon_device *rdev);
-void radeon_test_ring_sync(struct radeon_device *rdev,
-			   struct radeon_ring *cpA,
-			   struct radeon_ring *cpB);
-void radeon_test_syncing(struct radeon_device *rdev);
 
 
 /*
  * Debugfs
  */
-struct radeon_debugfs {
-	struct drm_info_list	*files;
-	unsigned		num_files;
-};
-
 int radeon_debugfs_add_files(struct radeon_device *rdev,
 			     struct drm_info_list *files,
 			     unsigned nfiles);
@@ -1128,8 +897,53 @@ struct radeon_asic {
 	int (*resume)(struct radeon_device *rdev);
 	int (*suspend)(struct radeon_device *rdev);
 	void (*vga_set_state)(struct radeon_device *rdev, bool state);
-	bool (*gpu_is_lockup)(struct radeon_device *rdev, struct radeon_ring *cp);
+	bool (*gpu_is_lockup)(struct radeon_device *rdev);
 	int (*asic_reset)(struct radeon_device *rdev);
+	void (*gart_tlb_flush)(struct radeon_device *rdev);
+	int (*gart_set_page)(struct radeon_device *rdev, int i, uint64_t addr);
+	int (*cp_init)(struct radeon_device *rdev, unsigned ring_size);
+	void (*cp_fini)(struct radeon_device *rdev);
+	void (*cp_disable)(struct radeon_device *rdev);
+	void (*cp_commit)(struct radeon_device *rdev);
+	void (*ring_start)(struct radeon_device *rdev);
+	int (*ring_test)(struct radeon_device *rdev);
+	void (*ring_ib_execute)(struct radeon_device *rdev, struct radeon_ib *ib);
+	int (*irq_set)(struct radeon_device *rdev);
+	int (*irq_process)(struct radeon_device *rdev);
+	u32 (*get_vblank_counter)(struct radeon_device *rdev, int crtc);
+	void (*fence_ring_emit)(struct radeon_device *rdev, struct radeon_fence *fence);
+	int (*cs_parse)(struct radeon_cs_parser *p);
+	int (*copy_blit)(struct radeon_device *rdev,
+			 uint64_t src_offset,
+			 uint64_t dst_offset,
+			 unsigned num_gpu_pages,
+			 struct radeon_fence *fence);
+	int (*copy_dma)(struct radeon_device *rdev,
+			uint64_t src_offset,
+			uint64_t dst_offset,
+			unsigned num_gpu_pages,
+			struct radeon_fence *fence);
+	int (*copy)(struct radeon_device *rdev,
+		    uint64_t src_offset,
+		    uint64_t dst_offset,
+		    unsigned num_gpu_pages,
+		    struct radeon_fence *fence);
+	uint32_t (*get_engine_clock)(struct radeon_device *rdev);
+	void (*set_engine_clock)(struct radeon_device *rdev, uint32_t eng_clock);
+	uint32_t (*get_memory_clock)(struct radeon_device *rdev);
+	void (*set_memory_clock)(struct radeon_device *rdev, uint32_t mem_clock);
+	int (*get_pcie_lanes)(struct radeon_device *rdev);
+	void (*set_pcie_lanes)(struct radeon_device *rdev, int lanes);
+	void (*set_clock_gating)(struct radeon_device *rdev, int enable);
+	int (*set_surface_reg)(struct radeon_device *rdev, int reg,
+			       uint32_t tiling_flags, uint32_t pitch,
+			       uint32_t offset, uint32_t obj_size);
+	void (*clear_surface_reg)(struct radeon_device *rdev, int reg);
+	void (*bandwidth_update)(struct radeon_device *rdev);
+	void (*hpd_init)(struct radeon_device *rdev);
+	void (*hpd_fini)(struct radeon_device *rdev);
+	bool (*hpd_sense)(struct radeon_device *rdev, enum radeon_hpd_id hpd);
+	void (*hpd_set_polarity)(struct radeon_device *rdev, enum radeon_hpd_id hpd);
 	/* ioctl hw specific callback. Some hw might want to perform special
 	 * operation on specific ioctl. For instance on wait idle some hw
 	 * might want to perform and HDP flush through MMIO as it seems that
@@ -1137,99 +951,17 @@ struct radeon_asic {
 	 * through ring.
 	 */
 	void (*ioctl_wait_idle)(struct radeon_device *rdev, struct radeon_bo *bo);
-	/* check if 3D engine is idle */
 	bool (*gui_idle)(struct radeon_device *rdev);
-	/* wait for mc_idle */
-	int (*mc_wait_for_idle)(struct radeon_device *rdev);
-	/* gart */
-	struct {
-		void (*tlb_flush)(struct radeon_device *rdev);
-		int (*set_page)(struct radeon_device *rdev, int i, uint64_t addr);
-	} gart;
-	/* ring specific callbacks */
-	struct {
-		void (*ib_execute)(struct radeon_device *rdev, struct radeon_ib *ib);
-		int (*ib_parse)(struct radeon_device *rdev, struct radeon_ib *ib);
-		void (*emit_fence)(struct radeon_device *rdev, struct radeon_fence *fence);
-		void (*emit_semaphore)(struct radeon_device *rdev, struct radeon_ring *cp,
-				       struct radeon_semaphore *semaphore, bool emit_wait);
-		int (*cs_parse)(struct radeon_cs_parser *p);
-		void (*ring_start)(struct radeon_device *rdev, struct radeon_ring *cp);
-		int (*ring_test)(struct radeon_device *rdev, struct radeon_ring *cp);
-		int (*ib_test)(struct radeon_device *rdev, struct radeon_ring *cp);
-	} ring[RADEON_NUM_RINGS];
-	/* irqs */
-	struct {
-		int (*set)(struct radeon_device *rdev);
-		int (*process)(struct radeon_device *rdev);
-	} irq;
-	/* displays */
-	struct {
-		/* display watermarks */
-		void (*bandwidth_update)(struct radeon_device *rdev);
-		/* get frame count */
-		u32 (*get_vblank_counter)(struct radeon_device *rdev, int crtc);
-		/* wait for vblank */
-		void (*wait_for_vblank)(struct radeon_device *rdev, int crtc);
-	} display;
-	/* copy functions for bo handling */
-	struct {
-		int (*blit)(struct radeon_device *rdev,
-			    uint64_t src_offset,
-			    uint64_t dst_offset,
-			    unsigned num_gpu_pages,
-			    struct radeon_fence *fence);
-		u32 blit_ring_index;
-		int (*dma)(struct radeon_device *rdev,
-			   uint64_t src_offset,
-			   uint64_t dst_offset,
-			   unsigned num_gpu_pages,
-			   struct radeon_fence *fence);
-		u32 dma_ring_index;
-		/* method used for bo copy */
-		int (*copy)(struct radeon_device *rdev,
-			    uint64_t src_offset,
-			    uint64_t dst_offset,
-			    unsigned num_gpu_pages,
-			    struct radeon_fence *fence);
-		/* ring used for bo copies */
-		u32 copy_ring_index;
-	} copy;
-	/* surfaces */
-	struct {
-		int (*set_reg)(struct radeon_device *rdev, int reg,
-				       uint32_t tiling_flags, uint32_t pitch,
-				       uint32_t offset, uint32_t obj_size);
-		void (*clear_reg)(struct radeon_device *rdev, int reg);
-	} surface;
-	/* hotplug detect */
-	struct {
-		void (*init)(struct radeon_device *rdev);
-		void (*fini)(struct radeon_device *rdev);
-		bool (*sense)(struct radeon_device *rdev, enum radeon_hpd_id hpd);
-		void (*set_polarity)(struct radeon_device *rdev, enum radeon_hpd_id hpd);
-	} hpd;
 	/* power management */
-	struct {
-		void (*misc)(struct radeon_device *rdev);
-		void (*prepare)(struct radeon_device *rdev);
-		void (*finish)(struct radeon_device *rdev);
-		void (*init_profile)(struct radeon_device *rdev);
-		void (*get_dynpm_state)(struct radeon_device *rdev);
-		uint32_t (*get_engine_clock)(struct radeon_device *rdev);
-		void (*set_engine_clock)(struct radeon_device *rdev, uint32_t eng_clock);
-		uint32_t (*get_memory_clock)(struct radeon_device *rdev);
-		void (*set_memory_clock)(struct radeon_device *rdev, uint32_t mem_clock);
-		int (*get_pcie_lanes)(struct radeon_device *rdev);
-		void (*set_pcie_lanes)(struct radeon_device *rdev, int lanes);
-		void (*set_clock_gating)(struct radeon_device *rdev, int enable);
-	} pm;
+	void (*pm_misc)(struct radeon_device *rdev);
+	void (*pm_prepare)(struct radeon_device *rdev);
+	void (*pm_finish)(struct radeon_device *rdev);
+	void (*pm_init_profile)(struct radeon_device *rdev);
+	void (*pm_get_dynpm_state)(struct radeon_device *rdev);
 	/* pageflipping */
-	struct {
-		void (*pre_page_flip)(struct radeon_device *rdev, int crtc);
-		u32 (*page_flip)(struct radeon_device *rdev, int crtc, u64 crtc_base);
-		void (*post_page_flip)(struct radeon_device *rdev, int crtc);
-	} pflip;
+	void (*pre_page_flip)(struct radeon_device *rdev, int crtc);
+	u32 (*page_flip)(struct radeon_device *rdev, int crtc, u64 crtc_base);
+	void (*post_page_flip)(struct radeon_device *rdev, int crtc);
 };
 
 /*
@@ -1273,7 +1005,6 @@ struct r600_asic {
 	unsigned		tiling_npipes;
 	unsigned		tiling_group_size;
 	unsigned		tile_config;
-	unsigned		backend_map;
 	struct r100_gpu_lockup	lockup;
 };
 
@@ -1299,7 +1030,6 @@ struct rv770_asic {
 	unsigned		tiling_npipes;
 	unsigned		tiling_group_size;
 	unsigned		tile_config;
-	unsigned		backend_map;
 	struct r100_gpu_lockup	lockup;
 };
 
@@ -1326,7 +1056,6 @@ struct evergreen_asic {
 	unsigned tiling_npipes;
 	unsigned tiling_group_size;
 	unsigned tile_config;
-	unsigned backend_map;
 	struct r100_gpu_lockup	lockup;
 };
 
@@ -1408,22 +1137,18 @@ int radeon_gem_busy_ioctl(struct drm_device *dev, void *data,
 			  struct drm_file *filp);
 int radeon_gem_wait_idle_ioctl(struct drm_device *dev, void *data,
 			      struct drm_file *filp);
-int radeon_gem_va_ioctl(struct drm_device *dev, void *data,
-			  struct drm_file *filp);
 int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp);
 int radeon_gem_set_tiling_ioctl(struct drm_device *dev, void *data,
 				struct drm_file *filp);
 int radeon_gem_get_tiling_ioctl(struct drm_device *dev, void *data,
 				struct drm_file *filp);
 
-/* VRAM scratch page for HDP bug, default vram page */
-struct r600_vram_scratch {
+/* VRAM scratch page for HDP bug */
+struct r700_vram_scratch {
 	struct radeon_bo		*robj;
 	volatile uint32_t		*ptr;
-	u64				gpu_addr;
 };
 
-
 /*
  * Core structure, functions and helpers.
  */
@@ -1451,7 +1176,7 @@ struct radeon_device {
 	/* Register mmio */
 	resource_size_t			rmmio_base;
 	resource_size_t			rmmio_size;
-	void __iomem			*rmmio;
+	void				*rmmio;
 	radeon_rreg_t			mc_rreg;
 	radeon_wreg_t			mc_wreg;
 	radeon_rreg_t			pll_rreg;
@@ -1468,17 +1193,18 @@ struct radeon_device {
 	struct radeon_mode_info		mode_info;
 	struct radeon_scratch		scratch;
 	struct radeon_mman		mman;
-	rwlock_t			fence_lock;
-	struct radeon_fence_driver	fence_drv[RADEON_NUM_RINGS];
-	struct radeon_semaphore_driver	semaphore_drv;
-	struct radeon_ring		ring[RADEON_NUM_RINGS];
+	struct radeon_fence_driver	fence_drv;
+	struct radeon_cp		cp;
+	/* cayman compute rings */
+	struct radeon_cp		cp1;
+	struct radeon_cp		cp2;
 	struct radeon_ib_pool		ib_pool;
 	struct radeon_irq		irq;
 	struct radeon_asic		*asic;
 	struct radeon_gem		gem;
 	struct radeon_pm		pm;
 	uint32_t			bios_scratch[RADEON_BIOS_NUM_SCRATCH];
-	struct radeon_mutex		cs_mutex;
+	struct mutex			cs_mutex;
 	struct radeon_wb		wb;
 	struct radeon_dummy_page	dummy_page;
 	bool				gpu_lockup;
@@ -1492,7 +1218,7 @@ struct radeon_device {
 	const struct firmware *rlc_fw;	/* r6/700 RLC firmware */
 	const struct firmware *mc_fw;	/* NI MC firmware */
 	struct r600_blit r600_blit;
-	struct r600_vram_scratch vram_scratch;
+	struct r700_vram_scratch vram_scratch;
 	int msi_enabled; /* msi enabled */
 	struct r600_ih ih; /* r6/700 interrupt ring */
 	struct work_struct hotplug_work;
@@ -1515,11 +1241,6 @@ struct radeon_device {
 	struct drm_file *cmask_filp;
 	/* i2c buses */
 	struct radeon_i2c_chan *i2c_bus[RADEON_MAX_I2C_BUS];
-	/* debugfs */
-	struct radeon_debugfs	debugfs[RADEON_DEBUGFS_MAX_COMPONENTS];
-	unsigned 		debugfs_count;
-	/* virtual memory */
-	struct radeon_vm_manager	vm_manager;
 };
 
 int radeon_device_init(struct radeon_device *rdev,
@@ -1529,10 +1250,45 @@ int radeon_device_init(struct radeon_device *rdev,
 void radeon_device_fini(struct radeon_device *rdev);
 int radeon_gpu_wait_for_idle(struct radeon_device *rdev);
 
-uint32_t r100_mm_rreg(struct radeon_device *rdev, uint32_t reg);
-void r100_mm_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
-u32 r100_io_rreg(struct radeon_device *rdev, u32 reg);
-void r100_io_wreg(struct radeon_device *rdev, u32 reg, u32 v);
+static inline uint32_t r100_mm_rreg(struct radeon_device *rdev, uint32_t reg)
+{
+	if (reg < rdev->rmmio_size)
+		return readl(((void __iomem *)rdev->rmmio) + reg);
+	else {
+		writel(reg, ((void __iomem *)rdev->rmmio) + RADEON_MM_INDEX);
+		return readl(((void __iomem *)rdev->rmmio) + RADEON_MM_DATA);
+	}
+}
+
+static inline void r100_mm_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v)
+{
+	if (reg < rdev->rmmio_size)
+		writel(v, ((void __iomem *)rdev->rmmio) + reg);
+	else {
+		writel(reg, ((void __iomem *)rdev->rmmio) + RADEON_MM_INDEX);
+		writel(v, ((void __iomem *)rdev->rmmio) + RADEON_MM_DATA);
+	}
+}
+
+static inline u32 r100_io_rreg(struct radeon_device *rdev, u32 reg)
+{
+	if (reg < rdev->rio_mem_size)
+		return ioread32(rdev->rio_mem + reg);
+	else {
+		iowrite32(reg, rdev->rio_mem + RADEON_MM_INDEX);
+		return ioread32(rdev->rio_mem + RADEON_MM_DATA);
+	}
+}
+
+static inline void r100_io_wreg(struct radeon_device *rdev, u32 reg, u32 v)
+{
+	if (reg < rdev->rio_mem_size)
+		iowrite32(v, rdev->rio_mem + reg);
+	else {
+		iowrite32(reg, rdev->rio_mem + RADEON_MM_INDEX);
+		iowrite32(v, rdev->rio_mem + RADEON_MM_DATA);
+	}
+}
 
 /*
  * Cast helper
@@ -1542,10 +1298,10 @@ void r100_io_wreg(struct radeon_device *rdev, u32 reg, u32 v);
 /*
  * Registers read & write functions.
  */
-#define RREG8(reg) readb((rdev->rmmio) + (reg))
-#define WREG8(reg, v) writeb(v, (rdev->rmmio) + (reg))
-#define RREG16(reg) readw((rdev->rmmio) + (reg))
-#define WREG16(reg, v) writew(v, (rdev->rmmio) + (reg))
+#define RREG8(reg) readb(((void __iomem *)rdev->rmmio) + (reg))
+#define WREG8(reg, v) writeb(v, ((void __iomem *)rdev->rmmio) + (reg))
+#define RREG16(reg) readw(((void __iomem *)rdev->rmmio) + (reg))
+#define WREG16(reg, v) writew(v, ((void __iomem *)rdev->rmmio) + (reg))
 #define RREG32(reg) r100_mm_rreg(rdev, (reg))
 #define DREG32(reg) printk(KERN_INFO "REGISTER: " #reg " : 0x%08X\n", r100_mm_rreg(rdev, (reg)))
 #define WREG32(reg, v) r100_mm_wreg(rdev, (reg), (v))
@@ -1655,18 +1411,19 @@ void radeon_atombios_fini(struct radeon_device *rdev);
 /*
  * RING helpers.
  */
-#if DRM_DEBUG_CODE == 0
-static inline void radeon_ring_write(struct radeon_ring *ring, uint32_t v)
+static inline void radeon_ring_write(struct radeon_device *rdev, uint32_t v)
 {
-	ring->ring[ring->wptr++] = v;
-	ring->wptr &= ring->ptr_mask;
-	ring->count_dw--;
-	ring->ring_free_dw--;
-}
-#else
-/* With debugging this is just too big to inline */
-void radeon_ring_write(struct radeon_ring *ring, uint32_t v);
+#if DRM_DEBUG_CODE
+	if (rdev->cp.count_dw <= 0) {
+		DRM_ERROR("radeon: writting more dword to ring than expected !\n");
+	}
 #endif
+	rdev->cp.ring[rdev->cp.wptr++] = v;
+	rdev->cp.wptr &= rdev->cp.ptr_mask;
+	rdev->cp.count_dw--;
+	rdev->cp.ring_free_dw--;
+}
+
 
 /*
  * ASICs macro.
@@ -1675,58 +1432,53 @@ void radeon_ring_write(struct radeon_ring *ring, uint32_t v);
 #define radeon_fini(rdev) (rdev)->asic->fini((rdev))
 #define radeon_resume(rdev) (rdev)->asic->resume((rdev))
 #define radeon_suspend(rdev) (rdev)->asic->suspend((rdev))
-#define radeon_cs_parse(rdev, r, p) (rdev)->asic->ring[(r)].cs_parse((p))
+#define radeon_cs_parse(p) rdev->asic->cs_parse((p))
 #define radeon_vga_set_state(rdev, state) (rdev)->asic->vga_set_state((rdev), (state))
-#define radeon_gpu_is_lockup(rdev, cp) (rdev)->asic->gpu_is_lockup((rdev), (cp))
+#define radeon_gpu_is_lockup(rdev) (rdev)->asic->gpu_is_lockup((rdev))
 #define radeon_asic_reset(rdev) (rdev)->asic->asic_reset((rdev))
-#define radeon_gart_tlb_flush(rdev) (rdev)->asic->gart.tlb_flush((rdev))
-#define radeon_gart_set_page(rdev, i, p) (rdev)->asic->gart.set_page((rdev), (i), (p))
-#define radeon_ring_start(rdev, r, cp) (rdev)->asic->ring[(r)].ring_start((rdev), (cp))
-#define radeon_ring_test(rdev, r, cp) (rdev)->asic->ring[(r)].ring_test((rdev), (cp))
-#define radeon_ib_test(rdev, r, cp) (rdev)->asic->ring[(r)].ib_test((rdev), (cp))
-#define radeon_ring_ib_execute(rdev, r, ib) (rdev)->asic->ring[(r)].ib_execute((rdev), (ib))
-#define radeon_ring_ib_parse(rdev, r, ib) (rdev)->asic->ring[(r)].ib_parse((rdev), (ib))
-#define radeon_irq_set(rdev) (rdev)->asic->irq.set((rdev))
-#define radeon_irq_process(rdev) (rdev)->asic->irq.process((rdev))
-#define radeon_get_vblank_counter(rdev, crtc) (rdev)->asic->display.get_vblank_counter((rdev), (crtc))
-#define radeon_fence_ring_emit(rdev, r, fence) (rdev)->asic->ring[(r)].emit_fence((rdev), (fence))
-#define radeon_semaphore_ring_emit(rdev, r, cp, semaphore, emit_wait) (rdev)->asic->ring[(r)].emit_semaphore((rdev), (cp), (semaphore), (emit_wait))
-#define radeon_copy_blit(rdev, s, d, np, f) (rdev)->asic->copy.blit((rdev), (s), (d), (np), (f))
-#define radeon_copy_dma(rdev, s, d, np, f) (rdev)->asic->copy.dma((rdev), (s), (d), (np), (f))
-#define radeon_copy(rdev, s, d, np, f) (rdev)->asic->copy.copy((rdev), (s), (d), (np), (f))
-#define radeon_copy_blit_ring_index(rdev) (rdev)->asic->copy.blit_ring_index
-#define radeon_copy_dma_ring_index(rdev) (rdev)->asic->copy.dma_ring_index
-#define radeon_copy_ring_index(rdev) (rdev)->asic->copy.copy_ring_index
-#define radeon_get_engine_clock(rdev) (rdev)->asic->pm.get_engine_clock((rdev))
-#define radeon_set_engine_clock(rdev, e) (rdev)->asic->pm.set_engine_clock((rdev), (e))
-#define radeon_get_memory_clock(rdev) (rdev)->asic->pm.get_memory_clock((rdev))
-#define radeon_set_memory_clock(rdev, e) (rdev)->asic->pm.set_memory_clock((rdev), (e))
-#define radeon_get_pcie_lanes(rdev) (rdev)->asic->pm.get_pcie_lanes((rdev))
-#define radeon_set_pcie_lanes(rdev, l) (rdev)->asic->pm.set_pcie_lanes((rdev), (l))
-#define radeon_set_clock_gating(rdev, e) (rdev)->asic->pm.set_clock_gating((rdev), (e))
-#define radeon_set_surface_reg(rdev, r, f, p, o, s) ((rdev)->asic->surface.set_reg((rdev), (r), (f), (p), (o), (s)))
-#define radeon_clear_surface_reg(rdev, r) ((rdev)->asic->surface.clear_reg((rdev), (r)))
-#define radeon_bandwidth_update(rdev) (rdev)->asic->display.bandwidth_update((rdev))
-#define radeon_hpd_init(rdev) (rdev)->asic->hpd.init((rdev))
-#define radeon_hpd_fini(rdev) (rdev)->asic->hpd.fini((rdev))
-#define radeon_hpd_sense(rdev, h) (rdev)->asic->hpd.sense((rdev), (h))
-#define radeon_hpd_set_polarity(rdev, h) (rdev)->asic->hpd.set_polarity((rdev), (h))
+#define radeon_gart_tlb_flush(rdev) (rdev)->asic->gart_tlb_flush((rdev))
+#define radeon_gart_set_page(rdev, i, p) (rdev)->asic->gart_set_page((rdev), (i), (p))
+#define radeon_cp_commit(rdev) (rdev)->asic->cp_commit((rdev))
+#define radeon_ring_start(rdev) (rdev)->asic->ring_start((rdev))
+#define radeon_ring_test(rdev) (rdev)->asic->ring_test((rdev))
+#define radeon_ring_ib_execute(rdev, ib) (rdev)->asic->ring_ib_execute((rdev), (ib))
+#define radeon_irq_set(rdev) (rdev)->asic->irq_set((rdev))
+#define radeon_irq_process(rdev) (rdev)->asic->irq_process((rdev))
+#define radeon_get_vblank_counter(rdev, crtc) (rdev)->asic->get_vblank_counter((rdev), (crtc))
+#define radeon_fence_ring_emit(rdev, fence) (rdev)->asic->fence_ring_emit((rdev), (fence))
+#define radeon_copy_blit(rdev, s, d, np, f) (rdev)->asic->copy_blit((rdev), (s), (d), (np), (f))
+#define radeon_copy_dma(rdev, s, d, np, f) (rdev)->asic->copy_dma((rdev), (s), (d), (np), (f))
+#define radeon_copy(rdev, s, d, np, f) (rdev)->asic->copy((rdev), (s), (d), (np), (f))
+#define radeon_get_engine_clock(rdev) (rdev)->asic->get_engine_clock((rdev))
+#define radeon_set_engine_clock(rdev, e) (rdev)->asic->set_engine_clock((rdev), (e))
+#define radeon_get_memory_clock(rdev) (rdev)->asic->get_memory_clock((rdev))
+#define radeon_set_memory_clock(rdev, e) (rdev)->asic->set_memory_clock((rdev), (e))
+#define radeon_get_pcie_lanes(rdev) (rdev)->asic->get_pcie_lanes((rdev))
+#define radeon_set_pcie_lanes(rdev, l) (rdev)->asic->set_pcie_lanes((rdev), (l))
+#define radeon_set_clock_gating(rdev, e) (rdev)->asic->set_clock_gating((rdev), (e))
+#define radeon_set_surface_reg(rdev, r, f, p, o, s) ((rdev)->asic->set_surface_reg((rdev), (r), (f), (p), (o), (s)))
+#define radeon_clear_surface_reg(rdev, r) ((rdev)->asic->clear_surface_reg((rdev), (r)))
+#define radeon_bandwidth_update(rdev) (rdev)->asic->bandwidth_update((rdev))
+#define radeon_hpd_init(rdev) (rdev)->asic->hpd_init((rdev))
+#define radeon_hpd_fini(rdev) (rdev)->asic->hpd_fini((rdev))
+#define radeon_hpd_sense(rdev, hpd) (rdev)->asic->hpd_sense((rdev), (hpd))
+#define radeon_hpd_set_polarity(rdev, hpd) (rdev)->asic->hpd_set_polarity((rdev), (hpd))
 #define radeon_gui_idle(rdev) (rdev)->asic->gui_idle((rdev))
-#define radeon_pm_misc(rdev) (rdev)->asic->pm.misc((rdev))
-#define radeon_pm_prepare(rdev) (rdev)->asic->pm.prepare((rdev))
-#define radeon_pm_finish(rdev) (rdev)->asic->pm.finish((rdev))
-#define radeon_pm_init_profile(rdev) (rdev)->asic->pm.init_profile((rdev))
-#define radeon_pm_get_dynpm_state(rdev) (rdev)->asic->pm.get_dynpm_state((rdev))
-#define radeon_pre_page_flip(rdev, crtc) rdev->asic->pflip.pre_page_flip((rdev), (crtc))
-#define radeon_page_flip(rdev, crtc, base) rdev->asic->pflip.page_flip((rdev), (crtc), (base))
-#define radeon_post_page_flip(rdev, crtc) rdev->asic->pflip.post_page_flip((rdev), (crtc))
-#define radeon_wait_for_vblank(rdev, crtc) rdev->asic->display.wait_for_vblank((rdev), (crtc))
-#define radeon_mc_wait_for_idle(rdev) rdev->asic->mc_wait_for_idle((rdev))
+#define radeon_pm_misc(rdev) (rdev)->asic->pm_misc((rdev))
+#define radeon_pm_prepare(rdev) (rdev)->asic->pm_prepare((rdev))
+#define radeon_pm_finish(rdev) (rdev)->asic->pm_finish((rdev))
+#define radeon_pm_init_profile(rdev) (rdev)->asic->pm_init_profile((rdev))
+#define radeon_pm_get_dynpm_state(rdev) (rdev)->asic->pm_get_dynpm_state((rdev))
+#define radeon_pre_page_flip(rdev, crtc) rdev->asic->pre_page_flip((rdev), (crtc))
+#define radeon_page_flip(rdev, crtc, base) rdev->asic->page_flip((rdev), (crtc), (base))
+#define radeon_post_page_flip(rdev, crtc) rdev->asic->post_page_flip((rdev), (crtc))
 
 /* Common functions */
 /* AGP */
 extern int radeon_gpu_reset(struct radeon_device *rdev);
 extern void radeon_agp_disable(struct radeon_device *rdev);
+extern int radeon_gart_table_vram_pin(struct radeon_device *rdev);
+extern void radeon_gart_restore(struct radeon_device *rdev);
 extern int radeon_modeset_init(struct radeon_device *rdev);
 extern void radeon_modeset_fini(struct radeon_device *rdev);
 extern bool radeon_card_posted(struct radeon_device *rdev);
@@ -1750,49 +1502,6 @@ extern int radeon_suspend_kms(struct drm_device *dev, pm_message_t state);
 extern void radeon_ttm_set_active_vram_size(struct radeon_device *rdev, u64 size);
 
 /*
- * vm
- */
-int radeon_vm_manager_init(struct radeon_device *rdev);
-void radeon_vm_manager_fini(struct radeon_device *rdev);
-int radeon_vm_manager_start(struct radeon_device *rdev);
-int radeon_vm_manager_suspend(struct radeon_device *rdev);
-int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm);
-void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm);
-int radeon_vm_bind(struct radeon_device *rdev, struct radeon_vm *vm);
-void radeon_vm_unbind(struct radeon_device *rdev, struct radeon_vm *vm);
-int radeon_vm_bo_update_pte(struct radeon_device *rdev,
-			    struct radeon_vm *vm,
-			    struct radeon_bo *bo,
-			    struct ttm_mem_reg *mem);
-void radeon_vm_bo_invalidate(struct radeon_device *rdev,
-			     struct radeon_bo *bo);
-int radeon_vm_bo_add(struct radeon_device *rdev,
-		     struct radeon_vm *vm,
-		     struct radeon_bo *bo,
-		     uint64_t offset,
-		     uint32_t flags);
-int radeon_vm_bo_rmv(struct radeon_device *rdev,
-		     struct radeon_vm *vm,
-		     struct radeon_bo *bo);
-
-
-/*
- * R600 vram scratch functions
- */
-int r600_vram_scratch_init(struct radeon_device *rdev);
-void r600_vram_scratch_fini(struct radeon_device *rdev);
-
-/*
- * r600 cs checking helper
- */
-unsigned r600_mip_minify(unsigned size, unsigned level);
-bool r600_fmt_is_valid_color(u32 format);
-bool r600_fmt_is_valid_texture(u32 format, enum radeon_family family);
-int r600_fmt_get_blocksize(u32 format);
-int r600_fmt_get_nblocksx(u32 format, u32 w);
-int r600_fmt_get_nblocksy(u32 format, u32 h);
-
-/*
  * r600 functions used by radeon_encoder.c
  */
 extern void r600_hdmi_enable(struct drm_encoder *encoder);
diff --git a/drivers/gpu/drm/radeon/radeon_acpi.c b/drivers/gpu/drm/radeon/radeon_acpi.c
index 3516a60..3f6636b 100644
--- a/drivers/gpu/drm/radeon/radeon_acpi.c
+++ b/drivers/gpu/drm/radeon/radeon_acpi.c
@@ -35,8 +35,7 @@ static int radeon_atif_call(acpi_handle handle)
 
 	/* Fail only if calling the method fails and ATIF is supported */
 	if (ACPI_FAILURE(status) && status != AE_NOT_FOUND) {
-		DRM_DEBUG_DRIVER("failed to evaluate ATIF got %s\n",
-				 acpi_format_exception(status));
+		printk(KERN_DEBUG "failed to evaluate ATIF got %s\n", acpi_format_exception(status));
 		kfree(buffer.pointer);
 		return 1;
 	}
@@ -51,13 +50,13 @@ int radeon_acpi_init(struct radeon_device *rdev)
 	acpi_handle handle;
 	int ret;
 
-	/* Get the device handle */
-	handle = DEVICE_ACPI_HANDLE(&rdev->pdev->dev);
-
 	/* No need to proceed if we're sure that ATIF is not supported */
-	if (!ASIC_IS_AVIVO(rdev) || !rdev->bios || !handle)
+	if (!ASIC_IS_AVIVO(rdev) || !rdev->bios)
 		return 0;
 
+	/* Get the device handle */
+	handle = DEVICE_ACPI_HANDLE(&rdev->pdev->dev);
+
 	/* Call the ATIF method */
 	ret = radeon_atif_call(handle);
 	if (ret)
diff --git a/drivers/gpu/drm/radeon/radeon_asic.c b/drivers/gpu/drm/radeon/radeon_asic.c
index 479c89e..b244962 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.c
+++ b/drivers/gpu/drm/radeon/radeon_asic.c
@@ -114,13 +114,13 @@ void radeon_agp_disable(struct radeon_device *rdev)
 			rdev->family == CHIP_R423) {
 		DRM_INFO("Forcing AGP to PCIE mode\n");
 		rdev->flags |= RADEON_IS_PCIE;
-		rdev->asic->gart.tlb_flush = &rv370_pcie_gart_tlb_flush;
-		rdev->asic->gart.set_page = &rv370_pcie_gart_set_page;
+		rdev->asic->gart_tlb_flush = &rv370_pcie_gart_tlb_flush;
+		rdev->asic->gart_set_page = &rv370_pcie_gart_set_page;
 	} else {
 		DRM_INFO("Forcing AGP to PCI mode\n");
 		rdev->flags |= RADEON_IS_PCI;
-		rdev->asic->gart.tlb_flush = &r100_pci_gart_tlb_flush;
-		rdev->asic->gart.set_page = &r100_pci_gart_set_page;
+		rdev->asic->gart_tlb_flush = &r100_pci_gart_tlb_flush;
+		rdev->asic->gart_set_page = &r100_pci_gart_set_page;
 	}
 	rdev->mc.gtt_size = radeon_gart_size * 1024 * 1024;
 }
@@ -136,70 +136,44 @@ static struct radeon_asic r100_asic = {
 	.vga_set_state = &r100_vga_set_state,
 	.gpu_is_lockup = &r100_gpu_is_lockup,
 	.asic_reset = &r100_asic_reset,
+	.gart_tlb_flush = &r100_pci_gart_tlb_flush,
+	.gart_set_page = &r100_pci_gart_set_page,
+	.cp_commit = &r100_cp_commit,
+	.ring_start = &r100_ring_start,
+	.ring_test = &r100_ring_test,
+	.ring_ib_execute = &r100_ring_ib_execute,
+	.irq_set = &r100_irq_set,
+	.irq_process = &r100_irq_process,
+	.get_vblank_counter = &r100_get_vblank_counter,
+	.fence_ring_emit = &r100_fence_ring_emit,
+	.cs_parse = &r100_cs_parse,
+	.copy_blit = &r100_copy_blit,
+	.copy_dma = NULL,
+	.copy = &r100_copy_blit,
+	.get_engine_clock = &radeon_legacy_get_engine_clock,
+	.set_engine_clock = &radeon_legacy_set_engine_clock,
+	.get_memory_clock = &radeon_legacy_get_memory_clock,
+	.set_memory_clock = NULL,
+	.get_pcie_lanes = NULL,
+	.set_pcie_lanes = NULL,
+	.set_clock_gating = &radeon_legacy_set_clock_gating,
+	.set_surface_reg = r100_set_surface_reg,
+	.clear_surface_reg = r100_clear_surface_reg,
+	.bandwidth_update = &r100_bandwidth_update,
+	.hpd_init = &r100_hpd_init,
+	.hpd_fini = &r100_hpd_fini,
+	.hpd_sense = &r100_hpd_sense,
+	.hpd_set_polarity = &r100_hpd_set_polarity,
 	.ioctl_wait_idle = NULL,
 	.gui_idle = &r100_gui_idle,
-	.mc_wait_for_idle = &r100_mc_wait_for_idle,
-	.gart = {
-		.tlb_flush = &r100_pci_gart_tlb_flush,
-		.set_page = &r100_pci_gart_set_page,
-	},
-	.ring = {
-		[RADEON_RING_TYPE_GFX_INDEX] = {
-			.ib_execute = &r100_ring_ib_execute,
-			.emit_fence = &r100_fence_ring_emit,
-			.emit_semaphore = &r100_semaphore_ring_emit,
-			.cs_parse = &r100_cs_parse,
-			.ring_start = &r100_ring_start,
-			.ring_test = &r100_ring_test,
-			.ib_test = &r100_ib_test,
-		}
-	},
-	.irq = {
-		.set = &r100_irq_set,
-		.process = &r100_irq_process,
-	},
-	.display = {
-		.bandwidth_update = &r100_bandwidth_update,
-		.get_vblank_counter = &r100_get_vblank_counter,
-		.wait_for_vblank = &r100_wait_for_vblank,
-	},
-	.copy = {
-		.blit = &r100_copy_blit,
-		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.dma = NULL,
-		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.copy = &r100_copy_blit,
-		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-	},
-	.surface = {
-		.set_reg = r100_set_surface_reg,
-		.clear_reg = r100_clear_surface_reg,
-	},
-	.hpd = {
-		.init = &r100_hpd_init,
-		.fini = &r100_hpd_fini,
-		.sense = &r100_hpd_sense,
-		.set_polarity = &r100_hpd_set_polarity,
-	},
-	.pm = {
-		.misc = &r100_pm_misc,
-		.prepare = &r100_pm_prepare,
-		.finish = &r100_pm_finish,
-		.init_profile = &r100_pm_init_profile,
-		.get_dynpm_state = &r100_pm_get_dynpm_state,
-		.get_engine_clock = &radeon_legacy_get_engine_clock,
-		.set_engine_clock = &radeon_legacy_set_engine_clock,
-		.get_memory_clock = &radeon_legacy_get_memory_clock,
-		.set_memory_clock = NULL,
-		.get_pcie_lanes = NULL,
-		.set_pcie_lanes = NULL,
-		.set_clock_gating = &radeon_legacy_set_clock_gating,
-	},
-	.pflip = {
-		.pre_page_flip = &r100_pre_page_flip,
-		.page_flip = &r100_page_flip,
-		.post_page_flip = &r100_post_page_flip,
-	},
+	.pm_misc = &r100_pm_misc,
+	.pm_prepare = &r100_pm_prepare,
+	.pm_finish = &r100_pm_finish,
+	.pm_init_profile = &r100_pm_init_profile,
+	.pm_get_dynpm_state = &r100_pm_get_dynpm_state,
+	.pre_page_flip = &r100_pre_page_flip,
+	.page_flip = &r100_page_flip,
+	.post_page_flip = &r100_post_page_flip,
 };
 
 static struct radeon_asic r200_asic = {
@@ -210,70 +184,43 @@ static struct radeon_asic r200_asic = {
 	.vga_set_state = &r100_vga_set_state,
 	.gpu_is_lockup = &r100_gpu_is_lockup,
 	.asic_reset = &r100_asic_reset,
+	.gart_tlb_flush = &r100_pci_gart_tlb_flush,
+	.gart_set_page = &r100_pci_gart_set_page,
+	.cp_commit = &r100_cp_commit,
+	.ring_start = &r100_ring_start,
+	.ring_test = &r100_ring_test,
+	.ring_ib_execute = &r100_ring_ib_execute,
+	.irq_set = &r100_irq_set,
+	.irq_process = &r100_irq_process,
+	.get_vblank_counter = &r100_get_vblank_counter,
+	.fence_ring_emit = &r100_fence_ring_emit,
+	.cs_parse = &r100_cs_parse,
+	.copy_blit = &r100_copy_blit,
+	.copy_dma = &r200_copy_dma,
+	.copy = &r100_copy_blit,
+	.get_engine_clock = &radeon_legacy_get_engine_clock,
+	.set_engine_clock = &radeon_legacy_set_engine_clock,
+	.get_memory_clock = &radeon_legacy_get_memory_clock,
+	.set_memory_clock = NULL,
+	.set_pcie_lanes = NULL,
+	.set_clock_gating = &radeon_legacy_set_clock_gating,
+	.set_surface_reg = r100_set_surface_reg,
+	.clear_surface_reg = r100_clear_surface_reg,
+	.bandwidth_update = &r100_bandwidth_update,
+	.hpd_init = &r100_hpd_init,
+	.hpd_fini = &r100_hpd_fini,
+	.hpd_sense = &r100_hpd_sense,
+	.hpd_set_polarity = &r100_hpd_set_polarity,
 	.ioctl_wait_idle = NULL,
 	.gui_idle = &r100_gui_idle,
-	.mc_wait_for_idle = &r100_mc_wait_for_idle,
-	.gart = {
-		.tlb_flush = &r100_pci_gart_tlb_flush,
-		.set_page = &r100_pci_gart_set_page,
-	},
-	.ring = {
-		[RADEON_RING_TYPE_GFX_INDEX] = {
-			.ib_execute = &r100_ring_ib_execute,
-			.emit_fence = &r100_fence_ring_emit,
-			.emit_semaphore = &r100_semaphore_ring_emit,
-			.cs_parse = &r100_cs_parse,
-			.ring_start = &r100_ring_start,
-			.ring_test = &r100_ring_test,
-			.ib_test = &r100_ib_test,
-		}
-	},
-	.irq = {
-		.set = &r100_irq_set,
-		.process = &r100_irq_process,
-	},
-	.display = {
-		.bandwidth_update = &r100_bandwidth_update,
-		.get_vblank_counter = &r100_get_vblank_counter,
-		.wait_for_vblank = &r100_wait_for_vblank,
-	},
-	.copy = {
-		.blit = &r100_copy_blit,
-		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.dma = &r200_copy_dma,
-		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.copy = &r100_copy_blit,
-		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-	},
-	.surface = {
-		.set_reg = r100_set_surface_reg,
-		.clear_reg = r100_clear_surface_reg,
-	},
-	.hpd = {
-		.init = &r100_hpd_init,
-		.fini = &r100_hpd_fini,
-		.sense = &r100_hpd_sense,
-		.set_polarity = &r100_hpd_set_polarity,
-	},
-	.pm = {
-		.misc = &r100_pm_misc,
-		.prepare = &r100_pm_prepare,
-		.finish = &r100_pm_finish,
-		.init_profile = &r100_pm_init_profile,
-		.get_dynpm_state = &r100_pm_get_dynpm_state,
-		.get_engine_clock = &radeon_legacy_get_engine_clock,
-		.set_engine_clock = &radeon_legacy_set_engine_clock,
-		.get_memory_clock = &radeon_legacy_get_memory_clock,
-		.set_memory_clock = NULL,
-		.get_pcie_lanes = NULL,
-		.set_pcie_lanes = NULL,
-		.set_clock_gating = &radeon_legacy_set_clock_gating,
-	},
-	.pflip = {
-		.pre_page_flip = &r100_pre_page_flip,
-		.page_flip = &r100_page_flip,
-		.post_page_flip = &r100_post_page_flip,
-	},
+	.pm_misc = &r100_pm_misc,
+	.pm_prepare = &r100_pm_prepare,
+	.pm_finish = &r100_pm_finish,
+	.pm_init_profile = &r100_pm_init_profile,
+	.pm_get_dynpm_state = &r100_pm_get_dynpm_state,
+	.pre_page_flip = &r100_pre_page_flip,
+	.page_flip = &r100_page_flip,
+	.post_page_flip = &r100_post_page_flip,
 };
 
 static struct radeon_asic r300_asic = {
@@ -284,70 +231,44 @@ static struct radeon_asic r300_asic = {
 	.vga_set_state = &r100_vga_set_state,
 	.gpu_is_lockup = &r300_gpu_is_lockup,
 	.asic_reset = &r300_asic_reset,
+	.gart_tlb_flush = &r100_pci_gart_tlb_flush,
+	.gart_set_page = &r100_pci_gart_set_page,
+	.cp_commit = &r100_cp_commit,
+	.ring_start = &r300_ring_start,
+	.ring_test = &r100_ring_test,
+	.ring_ib_execute = &r100_ring_ib_execute,
+	.irq_set = &r100_irq_set,
+	.irq_process = &r100_irq_process,
+	.get_vblank_counter = &r100_get_vblank_counter,
+	.fence_ring_emit = &r300_fence_ring_emit,
+	.cs_parse = &r300_cs_parse,
+	.copy_blit = &r100_copy_blit,
+	.copy_dma = &r200_copy_dma,
+	.copy = &r100_copy_blit,
+	.get_engine_clock = &radeon_legacy_get_engine_clock,
+	.set_engine_clock = &radeon_legacy_set_engine_clock,
+	.get_memory_clock = &radeon_legacy_get_memory_clock,
+	.set_memory_clock = NULL,
+	.get_pcie_lanes = &rv370_get_pcie_lanes,
+	.set_pcie_lanes = &rv370_set_pcie_lanes,
+	.set_clock_gating = &radeon_legacy_set_clock_gating,
+	.set_surface_reg = r100_set_surface_reg,
+	.clear_surface_reg = r100_clear_surface_reg,
+	.bandwidth_update = &r100_bandwidth_update,
+	.hpd_init = &r100_hpd_init,
+	.hpd_fini = &r100_hpd_fini,
+	.hpd_sense = &r100_hpd_sense,
+	.hpd_set_polarity = &r100_hpd_set_polarity,
 	.ioctl_wait_idle = NULL,
 	.gui_idle = &r100_gui_idle,
-	.mc_wait_for_idle = &r300_mc_wait_for_idle,
-	.gart = {
-		.tlb_flush = &r100_pci_gart_tlb_flush,
-		.set_page = &r100_pci_gart_set_page,
-	},
-	.ring = {
-		[RADEON_RING_TYPE_GFX_INDEX] = {
-			.ib_execute = &r100_ring_ib_execute,
-			.emit_fence = &r300_fence_ring_emit,
-			.emit_semaphore = &r100_semaphore_ring_emit,
-			.cs_parse = &r300_cs_parse,
-			.ring_start = &r300_ring_start,
-			.ring_test = &r100_ring_test,
-			.ib_test = &r100_ib_test,
-		}
-	},
-	.irq = {
-		.set = &r100_irq_set,
-		.process = &r100_irq_process,
-	},
-	.display = {
-		.bandwidth_update = &r100_bandwidth_update,
-		.get_vblank_counter = &r100_get_vblank_counter,
-		.wait_for_vblank = &r100_wait_for_vblank,
-	},
-	.copy = {
-		.blit = &r100_copy_blit,
-		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.dma = &r200_copy_dma,
-		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.copy = &r100_copy_blit,
-		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-	},
-	.surface = {
-		.set_reg = r100_set_surface_reg,
-		.clear_reg = r100_clear_surface_reg,
-	},
-	.hpd = {
-		.init = &r100_hpd_init,
-		.fini = &r100_hpd_fini,
-		.sense = &r100_hpd_sense,
-		.set_polarity = &r100_hpd_set_polarity,
-	},
-	.pm = {
-		.misc = &r100_pm_misc,
-		.prepare = &r100_pm_prepare,
-		.finish = &r100_pm_finish,
-		.init_profile = &r100_pm_init_profile,
-		.get_dynpm_state = &r100_pm_get_dynpm_state,
-		.get_engine_clock = &radeon_legacy_get_engine_clock,
-		.set_engine_clock = &radeon_legacy_set_engine_clock,
-		.get_memory_clock = &radeon_legacy_get_memory_clock,
-		.set_memory_clock = NULL,
-		.get_pcie_lanes = &rv370_get_pcie_lanes,
-		.set_pcie_lanes = &rv370_set_pcie_lanes,
-		.set_clock_gating = &radeon_legacy_set_clock_gating,
-	},
-	.pflip = {
-		.pre_page_flip = &r100_pre_page_flip,
-		.page_flip = &r100_page_flip,
-		.post_page_flip = &r100_post_page_flip,
-	},
+	.pm_misc = &r100_pm_misc,
+	.pm_prepare = &r100_pm_prepare,
+	.pm_finish = &r100_pm_finish,
+	.pm_init_profile = &r100_pm_init_profile,
+	.pm_get_dynpm_state = &r100_pm_get_dynpm_state,
+	.pre_page_flip = &r100_pre_page_flip,
+	.page_flip = &r100_page_flip,
+	.post_page_flip = &r100_post_page_flip,
 };
 
 static struct radeon_asic r300_asic_pcie = {
@@ -358,70 +279,43 @@ static struct radeon_asic r300_asic_pcie = {
 	.vga_set_state = &r100_vga_set_state,
 	.gpu_is_lockup = &r300_gpu_is_lockup,
 	.asic_reset = &r300_asic_reset,
+	.gart_tlb_flush = &rv370_pcie_gart_tlb_flush,
+	.gart_set_page = &rv370_pcie_gart_set_page,
+	.cp_commit = &r100_cp_commit,
+	.ring_start = &r300_ring_start,
+	.ring_test = &r100_ring_test,
+	.ring_ib_execute = &r100_ring_ib_execute,
+	.irq_set = &r100_irq_set,
+	.irq_process = &r100_irq_process,
+	.get_vblank_counter = &r100_get_vblank_counter,
+	.fence_ring_emit = &r300_fence_ring_emit,
+	.cs_parse = &r300_cs_parse,
+	.copy_blit = &r100_copy_blit,
+	.copy_dma = &r200_copy_dma,
+	.copy = &r100_copy_blit,
+	.get_engine_clock = &radeon_legacy_get_engine_clock,
+	.set_engine_clock = &radeon_legacy_set_engine_clock,
+	.get_memory_clock = &radeon_legacy_get_memory_clock,
+	.set_memory_clock = NULL,
+	.set_pcie_lanes = &rv370_set_pcie_lanes,
+	.set_clock_gating = &radeon_legacy_set_clock_gating,
+	.set_surface_reg = r100_set_surface_reg,
+	.clear_surface_reg = r100_clear_surface_reg,
+	.bandwidth_update = &r100_bandwidth_update,
+	.hpd_init = &r100_hpd_init,
+	.hpd_fini = &r100_hpd_fini,
+	.hpd_sense = &r100_hpd_sense,
+	.hpd_set_polarity = &r100_hpd_set_polarity,
 	.ioctl_wait_idle = NULL,
 	.gui_idle = &r100_gui_idle,
-	.mc_wait_for_idle = &r300_mc_wait_for_idle,
-	.gart = {
-		.tlb_flush = &rv370_pcie_gart_tlb_flush,
-		.set_page = &rv370_pcie_gart_set_page,
-	},
-	.ring = {
-		[RADEON_RING_TYPE_GFX_INDEX] = {
-			.ib_execute = &r100_ring_ib_execute,
-			.emit_fence = &r300_fence_ring_emit,
-			.emit_semaphore = &r100_semaphore_ring_emit,
-			.cs_parse = &r300_cs_parse,
-			.ring_start = &r300_ring_start,
-			.ring_test = &r100_ring_test,
-			.ib_test = &r100_ib_test,
-		}
-	},
-	.irq = {
-		.set = &r100_irq_set,
-		.process = &r100_irq_process,
-	},
-	.display = {
-		.bandwidth_update = &r100_bandwidth_update,
-		.get_vblank_counter = &r100_get_vblank_counter,
-		.wait_for_vblank = &r100_wait_for_vblank,
-	},
-	.copy = {
-		.blit = &r100_copy_blit,
-		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.dma = &r200_copy_dma,
-		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.copy = &r100_copy_blit,
-		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-	},
-	.surface = {
-		.set_reg = r100_set_surface_reg,
-		.clear_reg = r100_clear_surface_reg,
-	},
-	.hpd = {
-		.init = &r100_hpd_init,
-		.fini = &r100_hpd_fini,
-		.sense = &r100_hpd_sense,
-		.set_polarity = &r100_hpd_set_polarity,
-	},
-	.pm = {
-		.misc = &r100_pm_misc,
-		.prepare = &r100_pm_prepare,
-		.finish = &r100_pm_finish,
-		.init_profile = &r100_pm_init_profile,
-		.get_dynpm_state = &r100_pm_get_dynpm_state,
-		.get_engine_clock = &radeon_legacy_get_engine_clock,
-		.set_engine_clock = &radeon_legacy_set_engine_clock,
-		.get_memory_clock = &radeon_legacy_get_memory_clock,
-		.set_memory_clock = NULL,
-		.get_pcie_lanes = &rv370_get_pcie_lanes,
-		.set_pcie_lanes = &rv370_set_pcie_lanes,
-		.set_clock_gating = &radeon_legacy_set_clock_gating,
-	},
-	.pflip = {
-		.pre_page_flip = &r100_pre_page_flip,
-		.page_flip = &r100_page_flip,
-		.post_page_flip = &r100_post_page_flip,
-	},
+	.pm_misc = &r100_pm_misc,
+	.pm_prepare = &r100_pm_prepare,
+	.pm_finish = &r100_pm_finish,
+	.pm_init_profile = &r100_pm_init_profile,
+	.pm_get_dynpm_state = &r100_pm_get_dynpm_state,
+	.pre_page_flip = &r100_pre_page_flip,
+	.page_flip = &r100_page_flip,
+	.post_page_flip = &r100_post_page_flip,
 };
 
 static struct radeon_asic r420_asic = {
@@ -432,70 +326,44 @@ static struct radeon_asic r420_asic = {
 	.vga_set_state = &r100_vga_set_state,
 	.gpu_is_lockup = &r300_gpu_is_lockup,
 	.asic_reset = &r300_asic_reset,
+	.gart_tlb_flush = &rv370_pcie_gart_tlb_flush,
+	.gart_set_page = &rv370_pcie_gart_set_page,
+	.cp_commit = &r100_cp_commit,
+	.ring_start = &r300_ring_start,
+	.ring_test = &r100_ring_test,
+	.ring_ib_execute = &r100_ring_ib_execute,
+	.irq_set = &r100_irq_set,
+	.irq_process = &r100_irq_process,
+	.get_vblank_counter = &r100_get_vblank_counter,
+	.fence_ring_emit = &r300_fence_ring_emit,
+	.cs_parse = &r300_cs_parse,
+	.copy_blit = &r100_copy_blit,
+	.copy_dma = &r200_copy_dma,
+	.copy = &r100_copy_blit,
+	.get_engine_clock = &radeon_atom_get_engine_clock,
+	.set_engine_clock = &radeon_atom_set_engine_clock,
+	.get_memory_clock = &radeon_atom_get_memory_clock,
+	.set_memory_clock = &radeon_atom_set_memory_clock,
+	.get_pcie_lanes = &rv370_get_pcie_lanes,
+	.set_pcie_lanes = &rv370_set_pcie_lanes,
+	.set_clock_gating = &radeon_atom_set_clock_gating,
+	.set_surface_reg = r100_set_surface_reg,
+	.clear_surface_reg = r100_clear_surface_reg,
+	.bandwidth_update = &r100_bandwidth_update,
+	.hpd_init = &r100_hpd_init,
+	.hpd_fini = &r100_hpd_fini,
+	.hpd_sense = &r100_hpd_sense,
+	.hpd_set_polarity = &r100_hpd_set_polarity,
 	.ioctl_wait_idle = NULL,
 	.gui_idle = &r100_gui_idle,
-	.mc_wait_for_idle = &r300_mc_wait_for_idle,
-	.gart = {
-		.tlb_flush = &rv370_pcie_gart_tlb_flush,
-		.set_page = &rv370_pcie_gart_set_page,
-	},
-	.ring = {
-		[RADEON_RING_TYPE_GFX_INDEX] = {
-			.ib_execute = &r100_ring_ib_execute,
-			.emit_fence = &r300_fence_ring_emit,
-			.emit_semaphore = &r100_semaphore_ring_emit,
-			.cs_parse = &r300_cs_parse,
-			.ring_start = &r300_ring_start,
-			.ring_test = &r100_ring_test,
-			.ib_test = &r100_ib_test,
-		}
-	},
-	.irq = {
-		.set = &r100_irq_set,
-		.process = &r100_irq_process,
-	},
-	.display = {
-		.bandwidth_update = &r100_bandwidth_update,
-		.get_vblank_counter = &r100_get_vblank_counter,
-		.wait_for_vblank = &r100_wait_for_vblank,
-	},
-	.copy = {
-		.blit = &r100_copy_blit,
-		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.dma = &r200_copy_dma,
-		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.copy = &r100_copy_blit,
-		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-	},
-	.surface = {
-		.set_reg = r100_set_surface_reg,
-		.clear_reg = r100_clear_surface_reg,
-	},
-	.hpd = {
-		.init = &r100_hpd_init,
-		.fini = &r100_hpd_fini,
-		.sense = &r100_hpd_sense,
-		.set_polarity = &r100_hpd_set_polarity,
-	},
-	.pm = {
-		.misc = &r100_pm_misc,
-		.prepare = &r100_pm_prepare,
-		.finish = &r100_pm_finish,
-		.init_profile = &r420_pm_init_profile,
-		.get_dynpm_state = &r100_pm_get_dynpm_state,
-		.get_engine_clock = &radeon_atom_get_engine_clock,
-		.set_engine_clock = &radeon_atom_set_engine_clock,
-		.get_memory_clock = &radeon_atom_get_memory_clock,
-		.set_memory_clock = &radeon_atom_set_memory_clock,
-		.get_pcie_lanes = &rv370_get_pcie_lanes,
-		.set_pcie_lanes = &rv370_set_pcie_lanes,
-		.set_clock_gating = &radeon_atom_set_clock_gating,
-	},
-	.pflip = {
-		.pre_page_flip = &r100_pre_page_flip,
-		.page_flip = &r100_page_flip,
-		.post_page_flip = &r100_post_page_flip,
-	},
+	.pm_misc = &r100_pm_misc,
+	.pm_prepare = &r100_pm_prepare,
+	.pm_finish = &r100_pm_finish,
+	.pm_init_profile = &r420_pm_init_profile,
+	.pm_get_dynpm_state = &r100_pm_get_dynpm_state,
+	.pre_page_flip = &r100_pre_page_flip,
+	.page_flip = &r100_page_flip,
+	.post_page_flip = &r100_post_page_flip,
 };
 
 static struct radeon_asic rs400_asic = {
@@ -506,70 +374,44 @@ static struct radeon_asic rs400_asic = {
 	.vga_set_state = &r100_vga_set_state,
 	.gpu_is_lockup = &r300_gpu_is_lockup,
 	.asic_reset = &r300_asic_reset,
+	.gart_tlb_flush = &rs400_gart_tlb_flush,
+	.gart_set_page = &rs400_gart_set_page,
+	.cp_commit = &r100_cp_commit,
+	.ring_start = &r300_ring_start,
+	.ring_test = &r100_ring_test,
+	.ring_ib_execute = &r100_ring_ib_execute,
+	.irq_set = &r100_irq_set,
+	.irq_process = &r100_irq_process,
+	.get_vblank_counter = &r100_get_vblank_counter,
+	.fence_ring_emit = &r300_fence_ring_emit,
+	.cs_parse = &r300_cs_parse,
+	.copy_blit = &r100_copy_blit,
+	.copy_dma = &r200_copy_dma,
+	.copy = &r100_copy_blit,
+	.get_engine_clock = &radeon_legacy_get_engine_clock,
+	.set_engine_clock = &radeon_legacy_set_engine_clock,
+	.get_memory_clock = &radeon_legacy_get_memory_clock,
+	.set_memory_clock = NULL,
+	.get_pcie_lanes = NULL,
+	.set_pcie_lanes = NULL,
+	.set_clock_gating = &radeon_legacy_set_clock_gating,
+	.set_surface_reg = r100_set_surface_reg,
+	.clear_surface_reg = r100_clear_surface_reg,
+	.bandwidth_update = &r100_bandwidth_update,
+	.hpd_init = &r100_hpd_init,
+	.hpd_fini = &r100_hpd_fini,
+	.hpd_sense = &r100_hpd_sense,
+	.hpd_set_polarity = &r100_hpd_set_polarity,
 	.ioctl_wait_idle = NULL,
 	.gui_idle = &r100_gui_idle,
-	.mc_wait_for_idle = &rs400_mc_wait_for_idle,
-	.gart = {
-		.tlb_flush = &rs400_gart_tlb_flush,
-		.set_page = &rs400_gart_set_page,
-	},
-	.ring = {
-		[RADEON_RING_TYPE_GFX_INDEX] = {
-			.ib_execute = &r100_ring_ib_execute,
-			.emit_fence = &r300_fence_ring_emit,
-			.emit_semaphore = &r100_semaphore_ring_emit,
-			.cs_parse = &r300_cs_parse,
-			.ring_start = &r300_ring_start,
-			.ring_test = &r100_ring_test,
-			.ib_test = &r100_ib_test,
-		}
-	},
-	.irq = {
-		.set = &r100_irq_set,
-		.process = &r100_irq_process,
-	},
-	.display = {
-		.bandwidth_update = &r100_bandwidth_update,
-		.get_vblank_counter = &r100_get_vblank_counter,
-		.wait_for_vblank = &r100_wait_for_vblank,
-	},
-	.copy = {
-		.blit = &r100_copy_blit,
-		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.dma = &r200_copy_dma,
-		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.copy = &r100_copy_blit,
-		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-	},
-	.surface = {
-		.set_reg = r100_set_surface_reg,
-		.clear_reg = r100_clear_surface_reg,
-	},
-	.hpd = {
-		.init = &r100_hpd_init,
-		.fini = &r100_hpd_fini,
-		.sense = &r100_hpd_sense,
-		.set_polarity = &r100_hpd_set_polarity,
-	},
-	.pm = {
-		.misc = &r100_pm_misc,
-		.prepare = &r100_pm_prepare,
-		.finish = &r100_pm_finish,
-		.init_profile = &r100_pm_init_profile,
-		.get_dynpm_state = &r100_pm_get_dynpm_state,
-		.get_engine_clock = &radeon_legacy_get_engine_clock,
-		.set_engine_clock = &radeon_legacy_set_engine_clock,
-		.get_memory_clock = &radeon_legacy_get_memory_clock,
-		.set_memory_clock = NULL,
-		.get_pcie_lanes = NULL,
-		.set_pcie_lanes = NULL,
-		.set_clock_gating = &radeon_legacy_set_clock_gating,
-	},
-	.pflip = {
-		.pre_page_flip = &r100_pre_page_flip,
-		.page_flip = &r100_page_flip,
-		.post_page_flip = &r100_post_page_flip,
-	},
+	.pm_misc = &r100_pm_misc,
+	.pm_prepare = &r100_pm_prepare,
+	.pm_finish = &r100_pm_finish,
+	.pm_init_profile = &r100_pm_init_profile,
+	.pm_get_dynpm_state = &r100_pm_get_dynpm_state,
+	.pre_page_flip = &r100_pre_page_flip,
+	.page_flip = &r100_page_flip,
+	.post_page_flip = &r100_post_page_flip,
 };
 
 static struct radeon_asic rs600_asic = {
@@ -580,70 +422,44 @@ static struct radeon_asic rs600_asic = {
 	.vga_set_state = &r100_vga_set_state,
 	.gpu_is_lockup = &r300_gpu_is_lockup,
 	.asic_reset = &rs600_asic_reset,
+	.gart_tlb_flush = &rs600_gart_tlb_flush,
+	.gart_set_page = &rs600_gart_set_page,
+	.cp_commit = &r100_cp_commit,
+	.ring_start = &r300_ring_start,
+	.ring_test = &r100_ring_test,
+	.ring_ib_execute = &r100_ring_ib_execute,
+	.irq_set = &rs600_irq_set,
+	.irq_process = &rs600_irq_process,
+	.get_vblank_counter = &rs600_get_vblank_counter,
+	.fence_ring_emit = &r300_fence_ring_emit,
+	.cs_parse = &r300_cs_parse,
+	.copy_blit = &r100_copy_blit,
+	.copy_dma = &r200_copy_dma,
+	.copy = &r100_copy_blit,
+	.get_engine_clock = &radeon_atom_get_engine_clock,
+	.set_engine_clock = &radeon_atom_set_engine_clock,
+	.get_memory_clock = &radeon_atom_get_memory_clock,
+	.set_memory_clock = &radeon_atom_set_memory_clock,
+	.get_pcie_lanes = NULL,
+	.set_pcie_lanes = NULL,
+	.set_clock_gating = &radeon_atom_set_clock_gating,
+	.set_surface_reg = r100_set_surface_reg,
+	.clear_surface_reg = r100_clear_surface_reg,
+	.bandwidth_update = &rs600_bandwidth_update,
+	.hpd_init = &rs600_hpd_init,
+	.hpd_fini = &rs600_hpd_fini,
+	.hpd_sense = &rs600_hpd_sense,
+	.hpd_set_polarity = &rs600_hpd_set_polarity,
 	.ioctl_wait_idle = NULL,
 	.gui_idle = &r100_gui_idle,
-	.mc_wait_for_idle = &rs600_mc_wait_for_idle,
-	.gart = {
-		.tlb_flush = &rs600_gart_tlb_flush,
-		.set_page = &rs600_gart_set_page,
-	},
-	.ring = {
-		[RADEON_RING_TYPE_GFX_INDEX] = {
-			.ib_execute = &r100_ring_ib_execute,
-			.emit_fence = &r300_fence_ring_emit,
-			.emit_semaphore = &r100_semaphore_ring_emit,
-			.cs_parse = &r300_cs_parse,
-			.ring_start = &r300_ring_start,
-			.ring_test = &r100_ring_test,
-			.ib_test = &r100_ib_test,
-		}
-	},
-	.irq = {
-		.set = &rs600_irq_set,
-		.process = &rs600_irq_process,
-	},
-	.display = {
-		.bandwidth_update = &rs600_bandwidth_update,
-		.get_vblank_counter = &rs600_get_vblank_counter,
-		.wait_for_vblank = &avivo_wait_for_vblank,
-	},
-	.copy = {
-		.blit = &r100_copy_blit,
-		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.dma = &r200_copy_dma,
-		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.copy = &r100_copy_blit,
-		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-	},
-	.surface = {
-		.set_reg = r100_set_surface_reg,
-		.clear_reg = r100_clear_surface_reg,
-	},
-	.hpd = {
-		.init = &rs600_hpd_init,
-		.fini = &rs600_hpd_fini,
-		.sense = &rs600_hpd_sense,
-		.set_polarity = &rs600_hpd_set_polarity,
-	},
-	.pm = {
-		.misc = &rs600_pm_misc,
-		.prepare = &rs600_pm_prepare,
-		.finish = &rs600_pm_finish,
-		.init_profile = &r420_pm_init_profile,
-		.get_dynpm_state = &r100_pm_get_dynpm_state,
-		.get_engine_clock = &radeon_atom_get_engine_clock,
-		.set_engine_clock = &radeon_atom_set_engine_clock,
-		.get_memory_clock = &radeon_atom_get_memory_clock,
-		.set_memory_clock = &radeon_atom_set_memory_clock,
-		.get_pcie_lanes = NULL,
-		.set_pcie_lanes = NULL,
-		.set_clock_gating = &radeon_atom_set_clock_gating,
-	},
-	.pflip = {
-		.pre_page_flip = &rs600_pre_page_flip,
-		.page_flip = &rs600_page_flip,
-		.post_page_flip = &rs600_post_page_flip,
-	},
+	.pm_misc = &rs600_pm_misc,
+	.pm_prepare = &rs600_pm_prepare,
+	.pm_finish = &rs600_pm_finish,
+	.pm_init_profile = &r420_pm_init_profile,
+	.pm_get_dynpm_state = &r100_pm_get_dynpm_state,
+	.pre_page_flip = &rs600_pre_page_flip,
+	.page_flip = &rs600_page_flip,
+	.post_page_flip = &rs600_post_page_flip,
 };
 
 static struct radeon_asic rs690_asic = {
@@ -654,70 +470,44 @@ static struct radeon_asic rs690_asic = {
 	.vga_set_state = &r100_vga_set_state,
 	.gpu_is_lockup = &r300_gpu_is_lockup,
 	.asic_reset = &rs600_asic_reset,
+	.gart_tlb_flush = &rs400_gart_tlb_flush,
+	.gart_set_page = &rs400_gart_set_page,
+	.cp_commit = &r100_cp_commit,
+	.ring_start = &r300_ring_start,
+	.ring_test = &r100_ring_test,
+	.ring_ib_execute = &r100_ring_ib_execute,
+	.irq_set = &rs600_irq_set,
+	.irq_process = &rs600_irq_process,
+	.get_vblank_counter = &rs600_get_vblank_counter,
+	.fence_ring_emit = &r300_fence_ring_emit,
+	.cs_parse = &r300_cs_parse,
+	.copy_blit = &r100_copy_blit,
+	.copy_dma = &r200_copy_dma,
+	.copy = &r200_copy_dma,
+	.get_engine_clock = &radeon_atom_get_engine_clock,
+	.set_engine_clock = &radeon_atom_set_engine_clock,
+	.get_memory_clock = &radeon_atom_get_memory_clock,
+	.set_memory_clock = &radeon_atom_set_memory_clock,
+	.get_pcie_lanes = NULL,
+	.set_pcie_lanes = NULL,
+	.set_clock_gating = &radeon_atom_set_clock_gating,
+	.set_surface_reg = r100_set_surface_reg,
+	.clear_surface_reg = r100_clear_surface_reg,
+	.bandwidth_update = &rs690_bandwidth_update,
+	.hpd_init = &rs600_hpd_init,
+	.hpd_fini = &rs600_hpd_fini,
+	.hpd_sense = &rs600_hpd_sense,
+	.hpd_set_polarity = &rs600_hpd_set_polarity,
 	.ioctl_wait_idle = NULL,
 	.gui_idle = &r100_gui_idle,
-	.mc_wait_for_idle = &rs690_mc_wait_for_idle,
-	.gart = {
-		.tlb_flush = &rs400_gart_tlb_flush,
-		.set_page = &rs400_gart_set_page,
-	},
-	.ring = {
-		[RADEON_RING_TYPE_GFX_INDEX] = {
-			.ib_execute = &r100_ring_ib_execute,
-			.emit_fence = &r300_fence_ring_emit,
-			.emit_semaphore = &r100_semaphore_ring_emit,
-			.cs_parse = &r300_cs_parse,
-			.ring_start = &r300_ring_start,
-			.ring_test = &r100_ring_test,
-			.ib_test = &r100_ib_test,
-		}
-	},
-	.irq = {
-		.set = &rs600_irq_set,
-		.process = &rs600_irq_process,
-	},
-	.display = {
-		.get_vblank_counter = &rs600_get_vblank_counter,
-		.bandwidth_update = &rs690_bandwidth_update,
-		.wait_for_vblank = &avivo_wait_for_vblank,
-	},
-	.copy = {
-		.blit = &r100_copy_blit,
-		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.dma = &r200_copy_dma,
-		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.copy = &r200_copy_dma,
-		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-	},
-	.surface = {
-		.set_reg = r100_set_surface_reg,
-		.clear_reg = r100_clear_surface_reg,
-	},
-	.hpd = {
-		.init = &rs600_hpd_init,
-		.fini = &rs600_hpd_fini,
-		.sense = &rs600_hpd_sense,
-		.set_polarity = &rs600_hpd_set_polarity,
-	},
-	.pm = {
-		.misc = &rs600_pm_misc,
-		.prepare = &rs600_pm_prepare,
-		.finish = &rs600_pm_finish,
-		.init_profile = &r420_pm_init_profile,
-		.get_dynpm_state = &r100_pm_get_dynpm_state,
-		.get_engine_clock = &radeon_atom_get_engine_clock,
-		.set_engine_clock = &radeon_atom_set_engine_clock,
-		.get_memory_clock = &radeon_atom_get_memory_clock,
-		.set_memory_clock = &radeon_atom_set_memory_clock,
-		.get_pcie_lanes = NULL,
-		.set_pcie_lanes = NULL,
-		.set_clock_gating = &radeon_atom_set_clock_gating,
-	},
-	.pflip = {
-		.pre_page_flip = &rs600_pre_page_flip,
-		.page_flip = &rs600_page_flip,
-		.post_page_flip = &rs600_post_page_flip,
-	},
+	.pm_misc = &rs600_pm_misc,
+	.pm_prepare = &rs600_pm_prepare,
+	.pm_finish = &rs600_pm_finish,
+	.pm_init_profile = &r420_pm_init_profile,
+	.pm_get_dynpm_state = &r100_pm_get_dynpm_state,
+	.pre_page_flip = &rs600_pre_page_flip,
+	.page_flip = &rs600_page_flip,
+	.post_page_flip = &rs600_post_page_flip,
 };
 
 static struct radeon_asic rv515_asic = {
@@ -728,70 +518,44 @@ static struct radeon_asic rv515_asic = {
 	.vga_set_state = &r100_vga_set_state,
 	.gpu_is_lockup = &r300_gpu_is_lockup,
 	.asic_reset = &rs600_asic_reset,
+	.gart_tlb_flush = &rv370_pcie_gart_tlb_flush,
+	.gart_set_page = &rv370_pcie_gart_set_page,
+	.cp_commit = &r100_cp_commit,
+	.ring_start = &rv515_ring_start,
+	.ring_test = &r100_ring_test,
+	.ring_ib_execute = &r100_ring_ib_execute,
+	.irq_set = &rs600_irq_set,
+	.irq_process = &rs600_irq_process,
+	.get_vblank_counter = &rs600_get_vblank_counter,
+	.fence_ring_emit = &r300_fence_ring_emit,
+	.cs_parse = &r300_cs_parse,
+	.copy_blit = &r100_copy_blit,
+	.copy_dma = &r200_copy_dma,
+	.copy = &r100_copy_blit,
+	.get_engine_clock = &radeon_atom_get_engine_clock,
+	.set_engine_clock = &radeon_atom_set_engine_clock,
+	.get_memory_clock = &radeon_atom_get_memory_clock,
+	.set_memory_clock = &radeon_atom_set_memory_clock,
+	.get_pcie_lanes = &rv370_get_pcie_lanes,
+	.set_pcie_lanes = &rv370_set_pcie_lanes,
+	.set_clock_gating = &radeon_atom_set_clock_gating,
+	.set_surface_reg = r100_set_surface_reg,
+	.clear_surface_reg = r100_clear_surface_reg,
+	.bandwidth_update = &rv515_bandwidth_update,
+	.hpd_init = &rs600_hpd_init,
+	.hpd_fini = &rs600_hpd_fini,
+	.hpd_sense = &rs600_hpd_sense,
+	.hpd_set_polarity = &rs600_hpd_set_polarity,
 	.ioctl_wait_idle = NULL,
 	.gui_idle = &r100_gui_idle,
-	.mc_wait_for_idle = &rv515_mc_wait_for_idle,
-	.gart = {
-		.tlb_flush = &rv370_pcie_gart_tlb_flush,
-		.set_page = &rv370_pcie_gart_set_page,
-	},
-	.ring = {
-		[RADEON_RING_TYPE_GFX_INDEX] = {
-			.ib_execute = &r100_ring_ib_execute,
-			.emit_fence = &r300_fence_ring_emit,
-			.emit_semaphore = &r100_semaphore_ring_emit,
-			.cs_parse = &r300_cs_parse,
-			.ring_start = &rv515_ring_start,
-			.ring_test = &r100_ring_test,
-			.ib_test = &r100_ib_test,
-		}
-	},
-	.irq = {
-		.set = &rs600_irq_set,
-		.process = &rs600_irq_process,
-	},
-	.display = {
-		.get_vblank_counter = &rs600_get_vblank_counter,
-		.bandwidth_update = &rv515_bandwidth_update,
-		.wait_for_vblank = &avivo_wait_for_vblank,
-	},
-	.copy = {
-		.blit = &r100_copy_blit,
-		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.dma = &r200_copy_dma,
-		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.copy = &r100_copy_blit,
-		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-	},
-	.surface = {
-		.set_reg = r100_set_surface_reg,
-		.clear_reg = r100_clear_surface_reg,
-	},
-	.hpd = {
-		.init = &rs600_hpd_init,
-		.fini = &rs600_hpd_fini,
-		.sense = &rs600_hpd_sense,
-		.set_polarity = &rs600_hpd_set_polarity,
-	},
-	.pm = {
-		.misc = &rs600_pm_misc,
-		.prepare = &rs600_pm_prepare,
-		.finish = &rs600_pm_finish,
-		.init_profile = &r420_pm_init_profile,
-		.get_dynpm_state = &r100_pm_get_dynpm_state,
-		.get_engine_clock = &radeon_atom_get_engine_clock,
-		.set_engine_clock = &radeon_atom_set_engine_clock,
-		.get_memory_clock = &radeon_atom_get_memory_clock,
-		.set_memory_clock = &radeon_atom_set_memory_clock,
-		.get_pcie_lanes = &rv370_get_pcie_lanes,
-		.set_pcie_lanes = &rv370_set_pcie_lanes,
-		.set_clock_gating = &radeon_atom_set_clock_gating,
-	},
-	.pflip = {
-		.pre_page_flip = &rs600_pre_page_flip,
-		.page_flip = &rs600_page_flip,
-		.post_page_flip = &rs600_post_page_flip,
-	},
+	.pm_misc = &rs600_pm_misc,
+	.pm_prepare = &rs600_pm_prepare,
+	.pm_finish = &rs600_pm_finish,
+	.pm_init_profile = &r420_pm_init_profile,
+	.pm_get_dynpm_state = &r100_pm_get_dynpm_state,
+	.pre_page_flip = &rs600_pre_page_flip,
+	.page_flip = &rs600_page_flip,
+	.post_page_flip = &rs600_post_page_flip,
 };
 
 static struct radeon_asic r520_asic = {
@@ -802,70 +566,44 @@ static struct radeon_asic r520_asic = {
 	.vga_set_state = &r100_vga_set_state,
 	.gpu_is_lockup = &r300_gpu_is_lockup,
 	.asic_reset = &rs600_asic_reset,
+	.gart_tlb_flush = &rv370_pcie_gart_tlb_flush,
+	.gart_set_page = &rv370_pcie_gart_set_page,
+	.cp_commit = &r100_cp_commit,
+	.ring_start = &rv515_ring_start,
+	.ring_test = &r100_ring_test,
+	.ring_ib_execute = &r100_ring_ib_execute,
+	.irq_set = &rs600_irq_set,
+	.irq_process = &rs600_irq_process,
+	.get_vblank_counter = &rs600_get_vblank_counter,
+	.fence_ring_emit = &r300_fence_ring_emit,
+	.cs_parse = &r300_cs_parse,
+	.copy_blit = &r100_copy_blit,
+	.copy_dma = &r200_copy_dma,
+	.copy = &r100_copy_blit,
+	.get_engine_clock = &radeon_atom_get_engine_clock,
+	.set_engine_clock = &radeon_atom_set_engine_clock,
+	.get_memory_clock = &radeon_atom_get_memory_clock,
+	.set_memory_clock = &radeon_atom_set_memory_clock,
+	.get_pcie_lanes = &rv370_get_pcie_lanes,
+	.set_pcie_lanes = &rv370_set_pcie_lanes,
+	.set_clock_gating = &radeon_atom_set_clock_gating,
+	.set_surface_reg = r100_set_surface_reg,
+	.clear_surface_reg = r100_clear_surface_reg,
+	.bandwidth_update = &rv515_bandwidth_update,
+	.hpd_init = &rs600_hpd_init,
+	.hpd_fini = &rs600_hpd_fini,
+	.hpd_sense = &rs600_hpd_sense,
+	.hpd_set_polarity = &rs600_hpd_set_polarity,
 	.ioctl_wait_idle = NULL,
 	.gui_idle = &r100_gui_idle,
-	.mc_wait_for_idle = &r520_mc_wait_for_idle,
-	.gart = {
-		.tlb_flush = &rv370_pcie_gart_tlb_flush,
-		.set_page = &rv370_pcie_gart_set_page,
-	},
-	.ring = {
-		[RADEON_RING_TYPE_GFX_INDEX] = {
-			.ib_execute = &r100_ring_ib_execute,
-			.emit_fence = &r300_fence_ring_emit,
-			.emit_semaphore = &r100_semaphore_ring_emit,
-			.cs_parse = &r300_cs_parse,
-			.ring_start = &rv515_ring_start,
-			.ring_test = &r100_ring_test,
-			.ib_test = &r100_ib_test,
-		}
-	},
-	.irq = {
-		.set = &rs600_irq_set,
-		.process = &rs600_irq_process,
-	},
-	.display = {
-		.bandwidth_update = &rv515_bandwidth_update,
-		.get_vblank_counter = &rs600_get_vblank_counter,
-		.wait_for_vblank = &avivo_wait_for_vblank,
-	},
-	.copy = {
-		.blit = &r100_copy_blit,
-		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.dma = &r200_copy_dma,
-		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.copy = &r100_copy_blit,
-		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-	},
-	.surface = {
-		.set_reg = r100_set_surface_reg,
-		.clear_reg = r100_clear_surface_reg,
-	},
-	.hpd = {
-		.init = &rs600_hpd_init,
-		.fini = &rs600_hpd_fini,
-		.sense = &rs600_hpd_sense,
-		.set_polarity = &rs600_hpd_set_polarity,
-	},
-	.pm = {
-		.misc = &rs600_pm_misc,
-		.prepare = &rs600_pm_prepare,
-		.finish = &rs600_pm_finish,
-		.init_profile = &r420_pm_init_profile,
-		.get_dynpm_state = &r100_pm_get_dynpm_state,
-		.get_engine_clock = &radeon_atom_get_engine_clock,
-		.set_engine_clock = &radeon_atom_set_engine_clock,
-		.get_memory_clock = &radeon_atom_get_memory_clock,
-		.set_memory_clock = &radeon_atom_set_memory_clock,
-		.get_pcie_lanes = &rv370_get_pcie_lanes,
-		.set_pcie_lanes = &rv370_set_pcie_lanes,
-		.set_clock_gating = &radeon_atom_set_clock_gating,
-	},
-	.pflip = {
-		.pre_page_flip = &rs600_pre_page_flip,
-		.page_flip = &rs600_page_flip,
-		.post_page_flip = &rs600_post_page_flip,
-	},
+	.pm_misc = &rs600_pm_misc,
+	.pm_prepare = &rs600_pm_prepare,
+	.pm_finish = &rs600_pm_finish,
+	.pm_init_profile = &r420_pm_init_profile,
+	.pm_get_dynpm_state = &r100_pm_get_dynpm_state,
+	.pre_page_flip = &rs600_pre_page_flip,
+	.page_flip = &rs600_page_flip,
+	.post_page_flip = &rs600_post_page_flip,
 };
 
 static struct radeon_asic r600_asic = {
@@ -873,72 +611,46 @@ static struct radeon_asic r600_asic = {
 	.fini = &r600_fini,
 	.suspend = &r600_suspend,
 	.resume = &r600_resume,
+	.cp_commit = &r600_cp_commit,
 	.vga_set_state = &r600_vga_set_state,
 	.gpu_is_lockup = &r600_gpu_is_lockup,
 	.asic_reset = &r600_asic_reset,
+	.gart_tlb_flush = &r600_pcie_gart_tlb_flush,
+	.gart_set_page = &rs600_gart_set_page,
+	.ring_test = &r600_ring_test,
+	.ring_ib_execute = &r600_ring_ib_execute,
+	.irq_set = &r600_irq_set,
+	.irq_process = &r600_irq_process,
+	.get_vblank_counter = &rs600_get_vblank_counter,
+	.fence_ring_emit = &r600_fence_ring_emit,
+	.cs_parse = &r600_cs_parse,
+	.copy_blit = &r600_copy_blit,
+	.copy_dma = &r600_copy_blit,
+	.copy = &r600_copy_blit,
+	.get_engine_clock = &radeon_atom_get_engine_clock,
+	.set_engine_clock = &radeon_atom_set_engine_clock,
+	.get_memory_clock = &radeon_atom_get_memory_clock,
+	.set_memory_clock = &radeon_atom_set_memory_clock,
+	.get_pcie_lanes = &r600_get_pcie_lanes,
+	.set_pcie_lanes = &r600_set_pcie_lanes,
+	.set_clock_gating = NULL,
+	.set_surface_reg = r600_set_surface_reg,
+	.clear_surface_reg = r600_clear_surface_reg,
+	.bandwidth_update = &rv515_bandwidth_update,
+	.hpd_init = &r600_hpd_init,
+	.hpd_fini = &r600_hpd_fini,
+	.hpd_sense = &r600_hpd_sense,
+	.hpd_set_polarity = &r600_hpd_set_polarity,
 	.ioctl_wait_idle = r600_ioctl_wait_idle,
 	.gui_idle = &r600_gui_idle,
-	.mc_wait_for_idle = &r600_mc_wait_for_idle,
-	.gart = {
-		.tlb_flush = &r600_pcie_gart_tlb_flush,
-		.set_page = &rs600_gart_set_page,
-	},
-	.ring = {
-		[RADEON_RING_TYPE_GFX_INDEX] = {
-			.ib_execute = &r600_ring_ib_execute,
-			.emit_fence = &r600_fence_ring_emit,
-			.emit_semaphore = &r600_semaphore_ring_emit,
-			.cs_parse = &r600_cs_parse,
-			.ring_test = &r600_ring_test,
-			.ib_test = &r600_ib_test,
-		}
-	},
-	.irq = {
-		.set = &r600_irq_set,
-		.process = &r600_irq_process,
-	},
-	.display = {
-		.bandwidth_update = &rv515_bandwidth_update,
-		.get_vblank_counter = &rs600_get_vblank_counter,
-		.wait_for_vblank = &avivo_wait_for_vblank,
-	},
-	.copy = {
-		.blit = &r600_copy_blit,
-		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.dma = NULL,
-		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.copy = &r600_copy_blit,
-		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-	},
-	.surface = {
-		.set_reg = r600_set_surface_reg,
-		.clear_reg = r600_clear_surface_reg,
-	},
-	.hpd = {
-		.init = &r600_hpd_init,
-		.fini = &r600_hpd_fini,
-		.sense = &r600_hpd_sense,
-		.set_polarity = &r600_hpd_set_polarity,
-	},
-	.pm = {
-		.misc = &r600_pm_misc,
-		.prepare = &rs600_pm_prepare,
-		.finish = &rs600_pm_finish,
-		.init_profile = &r600_pm_init_profile,
-		.get_dynpm_state = &r600_pm_get_dynpm_state,
-		.get_engine_clock = &radeon_atom_get_engine_clock,
-		.set_engine_clock = &radeon_atom_set_engine_clock,
-		.get_memory_clock = &radeon_atom_get_memory_clock,
-		.set_memory_clock = &radeon_atom_set_memory_clock,
-		.get_pcie_lanes = &r600_get_pcie_lanes,
-		.set_pcie_lanes = &r600_set_pcie_lanes,
-		.set_clock_gating = NULL,
-	},
-	.pflip = {
-		.pre_page_flip = &rs600_pre_page_flip,
-		.page_flip = &rs600_page_flip,
-		.post_page_flip = &rs600_post_page_flip,
-	},
+	.pm_misc = &r600_pm_misc,
+	.pm_prepare = &rs600_pm_prepare,
+	.pm_finish = &rs600_pm_finish,
+	.pm_init_profile = &r600_pm_init_profile,
+	.pm_get_dynpm_state = &r600_pm_get_dynpm_state,
+	.pre_page_flip = &rs600_pre_page_flip,
+	.page_flip = &rs600_page_flip,
+	.post_page_flip = &rs600_post_page_flip,
 };
 
 static struct radeon_asic rs780_asic = {
@@ -946,72 +658,46 @@ static struct radeon_asic rs780_asic = {
 	.fini = &r600_fini,
 	.suspend = &r600_suspend,
 	.resume = &r600_resume,
+	.cp_commit = &r600_cp_commit,
 	.gpu_is_lockup = &r600_gpu_is_lockup,
 	.vga_set_state = &r600_vga_set_state,
 	.asic_reset = &r600_asic_reset,
+	.gart_tlb_flush = &r600_pcie_gart_tlb_flush,
+	.gart_set_page = &rs600_gart_set_page,
+	.ring_test = &r600_ring_test,
+	.ring_ib_execute = &r600_ring_ib_execute,
+	.irq_set = &r600_irq_set,
+	.irq_process = &r600_irq_process,
+	.get_vblank_counter = &rs600_get_vblank_counter,
+	.fence_ring_emit = &r600_fence_ring_emit,
+	.cs_parse = &r600_cs_parse,
+	.copy_blit = &r600_copy_blit,
+	.copy_dma = &r600_copy_blit,
+	.copy = &r600_copy_blit,
+	.get_engine_clock = &radeon_atom_get_engine_clock,
+	.set_engine_clock = &radeon_atom_set_engine_clock,
+	.get_memory_clock = NULL,
+	.set_memory_clock = NULL,
+	.get_pcie_lanes = NULL,
+	.set_pcie_lanes = NULL,
+	.set_clock_gating = NULL,
+	.set_surface_reg = r600_set_surface_reg,
+	.clear_surface_reg = r600_clear_surface_reg,
+	.bandwidth_update = &rs690_bandwidth_update,
+	.hpd_init = &r600_hpd_init,
+	.hpd_fini = &r600_hpd_fini,
+	.hpd_sense = &r600_hpd_sense,
+	.hpd_set_polarity = &r600_hpd_set_polarity,
 	.ioctl_wait_idle = r600_ioctl_wait_idle,
 	.gui_idle = &r600_gui_idle,
-	.mc_wait_for_idle = &r600_mc_wait_for_idle,
-	.gart = {
-		.tlb_flush = &r600_pcie_gart_tlb_flush,
-		.set_page = &rs600_gart_set_page,
-	},
-	.ring = {
-		[RADEON_RING_TYPE_GFX_INDEX] = {
-			.ib_execute = &r600_ring_ib_execute,
-			.emit_fence = &r600_fence_ring_emit,
-			.emit_semaphore = &r600_semaphore_ring_emit,
-			.cs_parse = &r600_cs_parse,
-			.ring_test = &r600_ring_test,
-			.ib_test = &r600_ib_test,
-		}
-	},
-	.irq = {
-		.set = &r600_irq_set,
-		.process = &r600_irq_process,
-	},
-	.display = {
-		.bandwidth_update = &rs690_bandwidth_update,
-		.get_vblank_counter = &rs600_get_vblank_counter,
-		.wait_for_vblank = &avivo_wait_for_vblank,
-	},
-	.copy = {
-		.blit = &r600_copy_blit,
-		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.dma = NULL,
-		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.copy = &r600_copy_blit,
-		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-	},
-	.surface = {
-		.set_reg = r600_set_surface_reg,
-		.clear_reg = r600_clear_surface_reg,
-	},
-	.hpd = {
-		.init = &r600_hpd_init,
-		.fini = &r600_hpd_fini,
-		.sense = &r600_hpd_sense,
-		.set_polarity = &r600_hpd_set_polarity,
-	},
-	.pm = {
-		.misc = &r600_pm_misc,
-		.prepare = &rs600_pm_prepare,
-		.finish = &rs600_pm_finish,
-		.init_profile = &rs780_pm_init_profile,
-		.get_dynpm_state = &r600_pm_get_dynpm_state,
-		.get_engine_clock = &radeon_atom_get_engine_clock,
-		.set_engine_clock = &radeon_atom_set_engine_clock,
-		.get_memory_clock = NULL,
-		.set_memory_clock = NULL,
-		.get_pcie_lanes = NULL,
-		.set_pcie_lanes = NULL,
-		.set_clock_gating = NULL,
-	},
-	.pflip = {
-		.pre_page_flip = &rs600_pre_page_flip,
-		.page_flip = &rs600_page_flip,
-		.post_page_flip = &rs600_post_page_flip,
-	},
+	.pm_misc = &r600_pm_misc,
+	.pm_prepare = &rs600_pm_prepare,
+	.pm_finish = &rs600_pm_finish,
+	.pm_init_profile = &rs780_pm_init_profile,
+	.pm_get_dynpm_state = &r600_pm_get_dynpm_state,
+	.pre_page_flip = &rs600_pre_page_flip,
+	.page_flip = &rs600_page_flip,
+	.post_page_flip = &rs600_post_page_flip,
 };
 
 static struct radeon_asic rv770_asic = {
@@ -1019,72 +705,46 @@ static struct radeon_asic rv770_asic = {
 	.fini = &rv770_fini,
 	.suspend = &rv770_suspend,
 	.resume = &rv770_resume,
+	.cp_commit = &r600_cp_commit,
 	.asic_reset = &r600_asic_reset,
 	.gpu_is_lockup = &r600_gpu_is_lockup,
 	.vga_set_state = &r600_vga_set_state,
+	.gart_tlb_flush = &r600_pcie_gart_tlb_flush,
+	.gart_set_page = &rs600_gart_set_page,
+	.ring_test = &r600_ring_test,
+	.ring_ib_execute = &r600_ring_ib_execute,
+	.irq_set = &r600_irq_set,
+	.irq_process = &r600_irq_process,
+	.get_vblank_counter = &rs600_get_vblank_counter,
+	.fence_ring_emit = &r600_fence_ring_emit,
+	.cs_parse = &r600_cs_parse,
+	.copy_blit = &r600_copy_blit,
+	.copy_dma = &r600_copy_blit,
+	.copy = &r600_copy_blit,
+	.get_engine_clock = &radeon_atom_get_engine_clock,
+	.set_engine_clock = &radeon_atom_set_engine_clock,
+	.get_memory_clock = &radeon_atom_get_memory_clock,
+	.set_memory_clock = &radeon_atom_set_memory_clock,
+	.get_pcie_lanes = &r600_get_pcie_lanes,
+	.set_pcie_lanes = &r600_set_pcie_lanes,
+	.set_clock_gating = &radeon_atom_set_clock_gating,
+	.set_surface_reg = r600_set_surface_reg,
+	.clear_surface_reg = r600_clear_surface_reg,
+	.bandwidth_update = &rv515_bandwidth_update,
+	.hpd_init = &r600_hpd_init,
+	.hpd_fini = &r600_hpd_fini,
+	.hpd_sense = &r600_hpd_sense,
+	.hpd_set_polarity = &r600_hpd_set_polarity,
 	.ioctl_wait_idle = r600_ioctl_wait_idle,
 	.gui_idle = &r600_gui_idle,
-	.mc_wait_for_idle = &r600_mc_wait_for_idle,
-	.gart = {
-		.tlb_flush = &r600_pcie_gart_tlb_flush,
-		.set_page = &rs600_gart_set_page,
-	},
-	.ring = {
-		[RADEON_RING_TYPE_GFX_INDEX] = {
-			.ib_execute = &r600_ring_ib_execute,
-			.emit_fence = &r600_fence_ring_emit,
-			.emit_semaphore = &r600_semaphore_ring_emit,
-			.cs_parse = &r600_cs_parse,
-			.ring_test = &r600_ring_test,
-			.ib_test = &r600_ib_test,
-		}
-	},
-	.irq = {
-		.set = &r600_irq_set,
-		.process = &r600_irq_process,
-	},
-	.display = {
-		.bandwidth_update = &rv515_bandwidth_update,
-		.get_vblank_counter = &rs600_get_vblank_counter,
-		.wait_for_vblank = &avivo_wait_for_vblank,
-	},
-	.copy = {
-		.blit = &r600_copy_blit,
-		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.dma = NULL,
-		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.copy = &r600_copy_blit,
-		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-	},
-	.surface = {
-		.set_reg = r600_set_surface_reg,
-		.clear_reg = r600_clear_surface_reg,
-	},
-	.hpd = {
-		.init = &r600_hpd_init,
-		.fini = &r600_hpd_fini,
-		.sense = &r600_hpd_sense,
-		.set_polarity = &r600_hpd_set_polarity,
-	},
-	.pm = {
-		.misc = &rv770_pm_misc,
-		.prepare = &rs600_pm_prepare,
-		.finish = &rs600_pm_finish,
-		.init_profile = &r600_pm_init_profile,
-		.get_dynpm_state = &r600_pm_get_dynpm_state,
-		.get_engine_clock = &radeon_atom_get_engine_clock,
-		.set_engine_clock = &radeon_atom_set_engine_clock,
-		.get_memory_clock = &radeon_atom_get_memory_clock,
-		.set_memory_clock = &radeon_atom_set_memory_clock,
-		.get_pcie_lanes = &r600_get_pcie_lanes,
-		.set_pcie_lanes = &r600_set_pcie_lanes,
-		.set_clock_gating = &radeon_atom_set_clock_gating,
-	},
-	.pflip = {
-		.pre_page_flip = &rs600_pre_page_flip,
-		.page_flip = &rv770_page_flip,
-		.post_page_flip = &rs600_post_page_flip,
-	},
+	.pm_misc = &rv770_pm_misc,
+	.pm_prepare = &rs600_pm_prepare,
+	.pm_finish = &rs600_pm_finish,
+	.pm_init_profile = &r600_pm_init_profile,
+	.pm_get_dynpm_state = &r600_pm_get_dynpm_state,
+	.pre_page_flip = &rs600_pre_page_flip,
+	.page_flip = &rv770_page_flip,
+	.post_page_flip = &rs600_post_page_flip,
 };
 
 static struct radeon_asic evergreen_asic = {
@@ -1092,72 +752,46 @@ static struct radeon_asic evergreen_asic = {
 	.fini = &evergreen_fini,
 	.suspend = &evergreen_suspend,
 	.resume = &evergreen_resume,
+	.cp_commit = &r600_cp_commit,
 	.gpu_is_lockup = &evergreen_gpu_is_lockup,
 	.asic_reset = &evergreen_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
+	.gart_tlb_flush = &evergreen_pcie_gart_tlb_flush,
+	.gart_set_page = &rs600_gart_set_page,
+	.ring_test = &r600_ring_test,
+	.ring_ib_execute = &evergreen_ring_ib_execute,
+	.irq_set = &evergreen_irq_set,
+	.irq_process = &evergreen_irq_process,
+	.get_vblank_counter = &evergreen_get_vblank_counter,
+	.fence_ring_emit = &r600_fence_ring_emit,
+	.cs_parse = &evergreen_cs_parse,
+	.copy_blit = &evergreen_copy_blit,
+	.copy_dma = &evergreen_copy_blit,
+	.copy = &evergreen_copy_blit,
+	.get_engine_clock = &radeon_atom_get_engine_clock,
+	.set_engine_clock = &radeon_atom_set_engine_clock,
+	.get_memory_clock = &radeon_atom_get_memory_clock,
+	.set_memory_clock = &radeon_atom_set_memory_clock,
+	.get_pcie_lanes = &r600_get_pcie_lanes,
+	.set_pcie_lanes = &r600_set_pcie_lanes,
+	.set_clock_gating = NULL,
+	.set_surface_reg = r600_set_surface_reg,
+	.clear_surface_reg = r600_clear_surface_reg,
+	.bandwidth_update = &evergreen_bandwidth_update,
+	.hpd_init = &evergreen_hpd_init,
+	.hpd_fini = &evergreen_hpd_fini,
+	.hpd_sense = &evergreen_hpd_sense,
+	.hpd_set_polarity = &evergreen_hpd_set_polarity,
 	.ioctl_wait_idle = r600_ioctl_wait_idle,
 	.gui_idle = &r600_gui_idle,
-	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
-	.gart = {
-		.tlb_flush = &evergreen_pcie_gart_tlb_flush,
-		.set_page = &rs600_gart_set_page,
-	},
-	.ring = {
-		[RADEON_RING_TYPE_GFX_INDEX] = {
-			.ib_execute = &evergreen_ring_ib_execute,
-			.emit_fence = &r600_fence_ring_emit,
-			.emit_semaphore = &r600_semaphore_ring_emit,
-			.cs_parse = &evergreen_cs_parse,
-			.ring_test = &r600_ring_test,
-			.ib_test = &r600_ib_test,
-		}
-	},
-	.irq = {
-		.set = &evergreen_irq_set,
-		.process = &evergreen_irq_process,
-	},
-	.display = {
-		.bandwidth_update = &evergreen_bandwidth_update,
-		.get_vblank_counter = &evergreen_get_vblank_counter,
-		.wait_for_vblank = &dce4_wait_for_vblank,
-	},
-	.copy = {
-		.blit = &r600_copy_blit,
-		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.dma = NULL,
-		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.copy = &r600_copy_blit,
-		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-	},
-	.surface = {
-		.set_reg = r600_set_surface_reg,
-		.clear_reg = r600_clear_surface_reg,
-	},
-	.hpd = {
-		.init = &evergreen_hpd_init,
-		.fini = &evergreen_hpd_fini,
-		.sense = &evergreen_hpd_sense,
-		.set_polarity = &evergreen_hpd_set_polarity,
-	},
-	.pm = {
-		.misc = &evergreen_pm_misc,
-		.prepare = &evergreen_pm_prepare,
-		.finish = &evergreen_pm_finish,
-		.init_profile = &r600_pm_init_profile,
-		.get_dynpm_state = &r600_pm_get_dynpm_state,
-		.get_engine_clock = &radeon_atom_get_engine_clock,
-		.set_engine_clock = &radeon_atom_set_engine_clock,
-		.get_memory_clock = &radeon_atom_get_memory_clock,
-		.set_memory_clock = &radeon_atom_set_memory_clock,
-		.get_pcie_lanes = &r600_get_pcie_lanes,
-		.set_pcie_lanes = &r600_set_pcie_lanes,
-		.set_clock_gating = NULL,
-	},
-	.pflip = {
-		.pre_page_flip = &evergreen_pre_page_flip,
-		.page_flip = &evergreen_page_flip,
-		.post_page_flip = &evergreen_post_page_flip,
-	},
+	.pm_misc = &evergreen_pm_misc,
+	.pm_prepare = &evergreen_pm_prepare,
+	.pm_finish = &evergreen_pm_finish,
+	.pm_init_profile = &r600_pm_init_profile,
+	.pm_get_dynpm_state = &r600_pm_get_dynpm_state,
+	.pre_page_flip = &evergreen_pre_page_flip,
+	.page_flip = &evergreen_page_flip,
+	.post_page_flip = &evergreen_post_page_flip,
 };
 
 static struct radeon_asic sumo_asic = {
@@ -1165,72 +799,46 @@ static struct radeon_asic sumo_asic = {
 	.fini = &evergreen_fini,
 	.suspend = &evergreen_suspend,
 	.resume = &evergreen_resume,
+	.cp_commit = &r600_cp_commit,
 	.gpu_is_lockup = &evergreen_gpu_is_lockup,
 	.asic_reset = &evergreen_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
+	.gart_tlb_flush = &evergreen_pcie_gart_tlb_flush,
+	.gart_set_page = &rs600_gart_set_page,
+	.ring_test = &r600_ring_test,
+	.ring_ib_execute = &evergreen_ring_ib_execute,
+	.irq_set = &evergreen_irq_set,
+	.irq_process = &evergreen_irq_process,
+	.get_vblank_counter = &evergreen_get_vblank_counter,
+	.fence_ring_emit = &r600_fence_ring_emit,
+	.cs_parse = &evergreen_cs_parse,
+	.copy_blit = &evergreen_copy_blit,
+	.copy_dma = &evergreen_copy_blit,
+	.copy = &evergreen_copy_blit,
+	.get_engine_clock = &radeon_atom_get_engine_clock,
+	.set_engine_clock = &radeon_atom_set_engine_clock,
+	.get_memory_clock = NULL,
+	.set_memory_clock = NULL,
+	.get_pcie_lanes = NULL,
+	.set_pcie_lanes = NULL,
+	.set_clock_gating = NULL,
+	.set_surface_reg = r600_set_surface_reg,
+	.clear_surface_reg = r600_clear_surface_reg,
+	.bandwidth_update = &evergreen_bandwidth_update,
+	.hpd_init = &evergreen_hpd_init,
+	.hpd_fini = &evergreen_hpd_fini,
+	.hpd_sense = &evergreen_hpd_sense,
+	.hpd_set_polarity = &evergreen_hpd_set_polarity,
 	.ioctl_wait_idle = r600_ioctl_wait_idle,
 	.gui_idle = &r600_gui_idle,
-	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
-	.gart = {
-		.tlb_flush = &evergreen_pcie_gart_tlb_flush,
-		.set_page = &rs600_gart_set_page,
-	},
-	.ring = {
-		[RADEON_RING_TYPE_GFX_INDEX] = {
-			.ib_execute = &evergreen_ring_ib_execute,
-			.emit_fence = &r600_fence_ring_emit,
-			.emit_semaphore = &r600_semaphore_ring_emit,
-			.cs_parse = &evergreen_cs_parse,
-			.ring_test = &r600_ring_test,
-			.ib_test = &r600_ib_test,
-		},
-	},
-	.irq = {
-		.set = &evergreen_irq_set,
-		.process = &evergreen_irq_process,
-	},
-	.display = {
-		.bandwidth_update = &evergreen_bandwidth_update,
-		.get_vblank_counter = &evergreen_get_vblank_counter,
-		.wait_for_vblank = &dce4_wait_for_vblank,
-	},
-	.copy = {
-		.blit = &r600_copy_blit,
-		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.dma = NULL,
-		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.copy = &r600_copy_blit,
-		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-	},
-	.surface = {
-		.set_reg = r600_set_surface_reg,
-		.clear_reg = r600_clear_surface_reg,
-	},
-	.hpd = {
-		.init = &evergreen_hpd_init,
-		.fini = &evergreen_hpd_fini,
-		.sense = &evergreen_hpd_sense,
-		.set_polarity = &evergreen_hpd_set_polarity,
-	},
-	.pm = {
-		.misc = &evergreen_pm_misc,
-		.prepare = &evergreen_pm_prepare,
-		.finish = &evergreen_pm_finish,
-		.init_profile = &sumo_pm_init_profile,
-		.get_dynpm_state = &r600_pm_get_dynpm_state,
-		.get_engine_clock = &radeon_atom_get_engine_clock,
-		.set_engine_clock = &radeon_atom_set_engine_clock,
-		.get_memory_clock = NULL,
-		.set_memory_clock = NULL,
-		.get_pcie_lanes = NULL,
-		.set_pcie_lanes = NULL,
-		.set_clock_gating = NULL,
-	},
-	.pflip = {
-		.pre_page_flip = &evergreen_pre_page_flip,
-		.page_flip = &evergreen_page_flip,
-		.post_page_flip = &evergreen_post_page_flip,
-	},
+	.pm_misc = &evergreen_pm_misc,
+	.pm_prepare = &evergreen_pm_prepare,
+	.pm_finish = &evergreen_pm_finish,
+	.pm_init_profile = &rs780_pm_init_profile,
+	.pm_get_dynpm_state = &r600_pm_get_dynpm_state,
+	.pre_page_flip = &evergreen_pre_page_flip,
+	.page_flip = &evergreen_page_flip,
+	.post_page_flip = &evergreen_post_page_flip,
 };
 
 static struct radeon_asic btc_asic = {
@@ -1238,82 +846,46 @@ static struct radeon_asic btc_asic = {
 	.fini = &evergreen_fini,
 	.suspend = &evergreen_suspend,
 	.resume = &evergreen_resume,
+	.cp_commit = &r600_cp_commit,
 	.gpu_is_lockup = &evergreen_gpu_is_lockup,
 	.asic_reset = &evergreen_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
+	.gart_tlb_flush = &evergreen_pcie_gart_tlb_flush,
+	.gart_set_page = &rs600_gart_set_page,
+	.ring_test = &r600_ring_test,
+	.ring_ib_execute = &evergreen_ring_ib_execute,
+	.irq_set = &evergreen_irq_set,
+	.irq_process = &evergreen_irq_process,
+	.get_vblank_counter = &evergreen_get_vblank_counter,
+	.fence_ring_emit = &r600_fence_ring_emit,
+	.cs_parse = &evergreen_cs_parse,
+	.copy_blit = &evergreen_copy_blit,
+	.copy_dma = &evergreen_copy_blit,
+	.copy = &evergreen_copy_blit,
+	.get_engine_clock = &radeon_atom_get_engine_clock,
+	.set_engine_clock = &radeon_atom_set_engine_clock,
+	.get_memory_clock = &radeon_atom_get_memory_clock,
+	.set_memory_clock = &radeon_atom_set_memory_clock,
+	.get_pcie_lanes = NULL,
+	.set_pcie_lanes = NULL,
+	.set_clock_gating = NULL,
+	.set_surface_reg = r600_set_surface_reg,
+	.clear_surface_reg = r600_clear_surface_reg,
+	.bandwidth_update = &evergreen_bandwidth_update,
+	.hpd_init = &evergreen_hpd_init,
+	.hpd_fini = &evergreen_hpd_fini,
+	.hpd_sense = &evergreen_hpd_sense,
+	.hpd_set_polarity = &evergreen_hpd_set_polarity,
 	.ioctl_wait_idle = r600_ioctl_wait_idle,
 	.gui_idle = &r600_gui_idle,
-	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
-	.gart = {
-		.tlb_flush = &evergreen_pcie_gart_tlb_flush,
-		.set_page = &rs600_gart_set_page,
-	},
-	.ring = {
-		[RADEON_RING_TYPE_GFX_INDEX] = {
-			.ib_execute = &evergreen_ring_ib_execute,
-			.emit_fence = &r600_fence_ring_emit,
-			.emit_semaphore = &r600_semaphore_ring_emit,
-			.cs_parse = &evergreen_cs_parse,
-			.ring_test = &r600_ring_test,
-			.ib_test = &r600_ib_test,
-		}
-	},
-	.irq = {
-		.set = &evergreen_irq_set,
-		.process = &evergreen_irq_process,
-	},
-	.display = {
-		.bandwidth_update = &evergreen_bandwidth_update,
-		.get_vblank_counter = &evergreen_get_vblank_counter,
-		.wait_for_vblank = &dce4_wait_for_vblank,
-	},
-	.copy = {
-		.blit = &r600_copy_blit,
-		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.dma = NULL,
-		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.copy = &r600_copy_blit,
-		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-	},
-	.surface = {
-		.set_reg = r600_set_surface_reg,
-		.clear_reg = r600_clear_surface_reg,
-	},
-	.hpd = {
-		.init = &evergreen_hpd_init,
-		.fini = &evergreen_hpd_fini,
-		.sense = &evergreen_hpd_sense,
-		.set_polarity = &evergreen_hpd_set_polarity,
-	},
-	.pm = {
-		.misc = &evergreen_pm_misc,
-		.prepare = &evergreen_pm_prepare,
-		.finish = &evergreen_pm_finish,
-		.init_profile = &r600_pm_init_profile,
-		.get_dynpm_state = &r600_pm_get_dynpm_state,
-		.get_engine_clock = &radeon_atom_get_engine_clock,
-		.set_engine_clock = &radeon_atom_set_engine_clock,
-		.get_memory_clock = &radeon_atom_get_memory_clock,
-		.set_memory_clock = &radeon_atom_set_memory_clock,
-		.get_pcie_lanes = NULL,
-		.set_pcie_lanes = NULL,
-		.set_clock_gating = NULL,
-	},
-	.pflip = {
-		.pre_page_flip = &evergreen_pre_page_flip,
-		.page_flip = &evergreen_page_flip,
-		.post_page_flip = &evergreen_post_page_flip,
-	},
-};
-
-static const struct radeon_vm_funcs cayman_vm_funcs = {
-	.init = &cayman_vm_init,
-	.fini = &cayman_vm_fini,
-	.bind = &cayman_vm_bind,
-	.unbind = &cayman_vm_unbind,
-	.tlb_flush = &cayman_vm_tlb_flush,
-	.page_flags = &cayman_vm_page_flags,
-	.set_page = &cayman_vm_set_page,
+	.pm_misc = &evergreen_pm_misc,
+	.pm_prepare = &evergreen_pm_prepare,
+	.pm_finish = &evergreen_pm_finish,
+	.pm_init_profile = &r600_pm_init_profile,
+	.pm_get_dynpm_state = &r600_pm_get_dynpm_state,
+	.pre_page_flip = &evergreen_pre_page_flip,
+	.page_flip = &evergreen_page_flip,
+	.post_page_flip = &evergreen_post_page_flip,
 };
 
 static struct radeon_asic cayman_asic = {
@@ -1321,91 +893,46 @@ static struct radeon_asic cayman_asic = {
 	.fini = &cayman_fini,
 	.suspend = &cayman_suspend,
 	.resume = &cayman_resume,
+	.cp_commit = &r600_cp_commit,
 	.gpu_is_lockup = &cayman_gpu_is_lockup,
 	.asic_reset = &cayman_asic_reset,
 	.vga_set_state = &r600_vga_set_state,
+	.gart_tlb_flush = &cayman_pcie_gart_tlb_flush,
+	.gart_set_page = &rs600_gart_set_page,
+	.ring_test = &r600_ring_test,
+	.ring_ib_execute = &evergreen_ring_ib_execute,
+	.irq_set = &evergreen_irq_set,
+	.irq_process = &evergreen_irq_process,
+	.get_vblank_counter = &evergreen_get_vblank_counter,
+	.fence_ring_emit = &r600_fence_ring_emit,
+	.cs_parse = &evergreen_cs_parse,
+	.copy_blit = &evergreen_copy_blit,
+	.copy_dma = &evergreen_copy_blit,
+	.copy = &evergreen_copy_blit,
+	.get_engine_clock = &radeon_atom_get_engine_clock,
+	.set_engine_clock = &radeon_atom_set_engine_clock,
+	.get_memory_clock = &radeon_atom_get_memory_clock,
+	.set_memory_clock = &radeon_atom_set_memory_clock,
+	.get_pcie_lanes = NULL,
+	.set_pcie_lanes = NULL,
+	.set_clock_gating = NULL,
+	.set_surface_reg = r600_set_surface_reg,
+	.clear_surface_reg = r600_clear_surface_reg,
+	.bandwidth_update = &evergreen_bandwidth_update,
+	.hpd_init = &evergreen_hpd_init,
+	.hpd_fini = &evergreen_hpd_fini,
+	.hpd_sense = &evergreen_hpd_sense,
+	.hpd_set_polarity = &evergreen_hpd_set_polarity,
 	.ioctl_wait_idle = r600_ioctl_wait_idle,
 	.gui_idle = &r600_gui_idle,
-	.mc_wait_for_idle = &evergreen_mc_wait_for_idle,
-	.gart = {
-		.tlb_flush = &cayman_pcie_gart_tlb_flush,
-		.set_page = &rs600_gart_set_page,
-	},
-	.ring = {
-		[RADEON_RING_TYPE_GFX_INDEX] = {
-			.ib_execute = &cayman_ring_ib_execute,
-			.ib_parse = &evergreen_ib_parse,
-			.emit_fence = &cayman_fence_ring_emit,
-			.emit_semaphore = &r600_semaphore_ring_emit,
-			.cs_parse = &evergreen_cs_parse,
-			.ring_test = &r600_ring_test,
-			.ib_test = &r600_ib_test,
-		},
-		[CAYMAN_RING_TYPE_CP1_INDEX] = {
-			.ib_execute = &cayman_ring_ib_execute,
-			.ib_parse = &evergreen_ib_parse,
-			.emit_fence = &cayman_fence_ring_emit,
-			.emit_semaphore = &r600_semaphore_ring_emit,
-			.cs_parse = &evergreen_cs_parse,
-			.ring_test = &r600_ring_test,
-			.ib_test = &r600_ib_test,
-		},
-		[CAYMAN_RING_TYPE_CP2_INDEX] = {
-			.ib_execute = &cayman_ring_ib_execute,
-			.ib_parse = &evergreen_ib_parse,
-			.emit_fence = &cayman_fence_ring_emit,
-			.emit_semaphore = &r600_semaphore_ring_emit,
-			.cs_parse = &evergreen_cs_parse,
-			.ring_test = &r600_ring_test,
-			.ib_test = &r600_ib_test,
-		}
-	},
-	.irq = {
-		.set = &evergreen_irq_set,
-		.process = &evergreen_irq_process,
-	},
-	.display = {
-		.bandwidth_update = &evergreen_bandwidth_update,
-		.get_vblank_counter = &evergreen_get_vblank_counter,
-		.wait_for_vblank = &dce4_wait_for_vblank,
-	},
-	.copy = {
-		.blit = &r600_copy_blit,
-		.blit_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.dma = NULL,
-		.dma_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-		.copy = &r600_copy_blit,
-		.copy_ring_index = RADEON_RING_TYPE_GFX_INDEX,
-	},
-	.surface = {
-		.set_reg = r600_set_surface_reg,
-		.clear_reg = r600_clear_surface_reg,
-	},
-	.hpd = {
-		.init = &evergreen_hpd_init,
-		.fini = &evergreen_hpd_fini,
-		.sense = &evergreen_hpd_sense,
-		.set_polarity = &evergreen_hpd_set_polarity,
-	},
-	.pm = {
-		.misc = &evergreen_pm_misc,
-		.prepare = &evergreen_pm_prepare,
-		.finish = &evergreen_pm_finish,
-		.init_profile = &r600_pm_init_profile,
-		.get_dynpm_state = &r600_pm_get_dynpm_state,
-		.get_engine_clock = &radeon_atom_get_engine_clock,
-		.set_engine_clock = &radeon_atom_set_engine_clock,
-		.get_memory_clock = &radeon_atom_get_memory_clock,
-		.set_memory_clock = &radeon_atom_set_memory_clock,
-		.get_pcie_lanes = NULL,
-		.set_pcie_lanes = NULL,
-		.set_clock_gating = NULL,
-	},
-	.pflip = {
-		.pre_page_flip = &evergreen_pre_page_flip,
-		.page_flip = &evergreen_page_flip,
-		.post_page_flip = &evergreen_post_page_flip,
-	},
+	.pm_misc = &evergreen_pm_misc,
+	.pm_prepare = &evergreen_pm_prepare,
+	.pm_finish = &evergreen_pm_finish,
+	.pm_init_profile = &r600_pm_init_profile,
+	.pm_get_dynpm_state = &r600_pm_get_dynpm_state,
+	.pre_page_flip = &evergreen_pre_page_flip,
+	.page_flip = &evergreen_page_flip,
+	.post_page_flip = &evergreen_post_page_flip,
 };
 
 int radeon_asic_init(struct radeon_device *rdev)
@@ -1447,10 +974,10 @@ int radeon_asic_init(struct radeon_device *rdev)
 		rdev->asic = &r420_asic;
 		/* handle macs */
 		if (rdev->bios == NULL) {
-			rdev->asic->pm.get_engine_clock = &radeon_legacy_get_engine_clock;
-			rdev->asic->pm.set_engine_clock = &radeon_legacy_set_engine_clock;
-			rdev->asic->pm.get_memory_clock = &radeon_legacy_get_memory_clock;
-			rdev->asic->pm.set_memory_clock = NULL;
+			rdev->asic->get_engine_clock = &radeon_legacy_get_engine_clock;
+			rdev->asic->set_engine_clock = &radeon_legacy_set_engine_clock;
+			rdev->asic->get_memory_clock = &radeon_legacy_get_memory_clock;
+			rdev->asic->set_memory_clock = NULL;
 		}
 		break;
 	case CHIP_RS400:
@@ -1523,7 +1050,6 @@ int radeon_asic_init(struct radeon_device *rdev)
 		rdev->asic = &cayman_asic;
 		/* set num crtcs */
 		rdev->num_crtc = 6;
-		rdev->vm_manager.funcs = &cayman_vm_funcs;
 		break;
 	default:
 		/* FIXME: not supported yet */
@@ -1531,8 +1057,8 @@ int radeon_asic_init(struct radeon_device *rdev)
 	}
 
 	if (rdev->flags & RADEON_IS_IGP) {
-		rdev->asic->pm.get_memory_clock = NULL;
-		rdev->asic->pm.set_memory_clock = NULL;
+		rdev->asic->get_memory_clock = NULL;
+		rdev->asic->set_memory_clock = NULL;
 	}
 
 	return 0;
diff --git a/drivers/gpu/drm/radeon/radeon_asic.h b/drivers/gpu/drm/radeon/radeon_asic.h
index dec17b3..4d81e96 100644
--- a/drivers/gpu/drm/radeon/radeon_asic.h
+++ b/drivers/gpu/drm/radeon/radeon_asic.h
@@ -58,20 +58,17 @@ void r100_fini(struct radeon_device *rdev);
 int r100_suspend(struct radeon_device *rdev);
 int r100_resume(struct radeon_device *rdev);
 void r100_vga_set_state(struct radeon_device *rdev, bool state);
-bool r100_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp);
+bool r100_gpu_is_lockup(struct radeon_device *rdev);
 int r100_asic_reset(struct radeon_device *rdev);
 u32 r100_get_vblank_counter(struct radeon_device *rdev, int crtc);
 void r100_pci_gart_tlb_flush(struct radeon_device *rdev);
 int r100_pci_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr);
-void r100_ring_start(struct radeon_device *rdev, struct radeon_ring *ring);
+void r100_cp_commit(struct radeon_device *rdev);
+void r100_ring_start(struct radeon_device *rdev);
 int r100_irq_set(struct radeon_device *rdev);
 int r100_irq_process(struct radeon_device *rdev);
 void r100_fence_ring_emit(struct radeon_device *rdev,
 			  struct radeon_fence *fence);
-void r100_semaphore_ring_emit(struct radeon_device *rdev,
-			      struct radeon_ring *cp,
-			      struct radeon_semaphore *semaphore,
-			      bool emit_wait);
 int r100_cs_parse(struct radeon_cs_parser *p);
 void r100_pll_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
 uint32_t r100_pll_rreg(struct radeon_device *rdev, uint32_t reg);
@@ -86,7 +83,7 @@ int r100_set_surface_reg(struct radeon_device *rdev, int reg,
 void r100_clear_surface_reg(struct radeon_device *rdev, int reg);
 void r100_bandwidth_update(struct radeon_device *rdev);
 void r100_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib);
-int r100_ring_test(struct radeon_device *rdev, struct radeon_ring *cp);
+int r100_ring_test(struct radeon_device *rdev);
 void r100_hpd_init(struct radeon_device *rdev);
 void r100_hpd_fini(struct radeon_device *rdev);
 bool r100_hpd_sense(struct radeon_device *rdev, enum radeon_hpd_id hpd);
@@ -104,12 +101,12 @@ void r100_pci_gart_disable(struct radeon_device *rdev);
 int r100_debugfs_mc_info_init(struct radeon_device *rdev);
 int r100_gui_wait_for_idle(struct radeon_device *rdev);
 void r100_gpu_lockup_update(struct r100_gpu_lockup *lockup,
-			    struct radeon_ring *cp);
+			    struct radeon_cp *cp);
 bool r100_gpu_cp_is_lockup(struct radeon_device *rdev,
 			   struct r100_gpu_lockup *lockup,
-			   struct radeon_ring *cp);
+			   struct radeon_cp *cp);
 void r100_ib_fini(struct radeon_device *rdev);
-int r100_ib_test(struct radeon_device *rdev, struct radeon_ring *ring);
+int r100_ib_init(struct radeon_device *rdev);
 void r100_irq_disable(struct radeon_device *rdev);
 void r100_mc_stop(struct radeon_device *rdev, struct r100_mc_save *save);
 void r100_mc_resume(struct radeon_device *rdev, struct r100_mc_save *save);
@@ -139,8 +136,6 @@ extern void r100_pm_get_dynpm_state(struct radeon_device *rdev);
 extern void r100_pre_page_flip(struct radeon_device *rdev, int crtc);
 extern u32 r100_page_flip(struct radeon_device *rdev, int crtc, u64 crtc_base);
 extern void r100_post_page_flip(struct radeon_device *rdev, int crtc);
-extern void r100_wait_for_vblank(struct radeon_device *rdev, int crtc);
-extern int r100_mc_wait_for_idle(struct radeon_device *rdev);
 
 /*
  * r200,rv250,rs300,rv280
@@ -159,9 +154,9 @@ extern int r300_init(struct radeon_device *rdev);
 extern void r300_fini(struct radeon_device *rdev);
 extern int r300_suspend(struct radeon_device *rdev);
 extern int r300_resume(struct radeon_device *rdev);
-extern bool r300_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp);
+extern bool r300_gpu_is_lockup(struct radeon_device *rdev);
 extern int r300_asic_reset(struct radeon_device *rdev);
-extern void r300_ring_start(struct radeon_device *rdev, struct radeon_ring *ring);
+extern void r300_ring_start(struct radeon_device *rdev);
 extern void r300_fence_ring_emit(struct radeon_device *rdev,
 				struct radeon_fence *fence);
 extern int r300_cs_parse(struct radeon_cs_parser *p);
@@ -178,7 +173,6 @@ extern int rv370_pcie_gart_init(struct radeon_device *rdev);
 extern void rv370_pcie_gart_fini(struct radeon_device *rdev);
 extern int rv370_pcie_gart_enable(struct radeon_device *rdev);
 extern void rv370_pcie_gart_disable(struct radeon_device *rdev);
-extern int r300_mc_wait_for_idle(struct radeon_device *rdev);
 
 /*
  * r420,r423,rv410
@@ -209,7 +203,6 @@ int rs400_gart_enable(struct radeon_device *rdev);
 void rs400_gart_adjust_size(struct radeon_device *rdev);
 void rs400_gart_disable(struct radeon_device *rdev);
 void rs400_gart_fini(struct radeon_device *rdev);
-extern int rs400_mc_wait_for_idle(struct radeon_device *rdev);
 
 /*
  * rs600.
@@ -240,8 +233,7 @@ extern void rs600_pre_page_flip(struct radeon_device *rdev, int crtc);
 extern u32 rs600_page_flip(struct radeon_device *rdev, int crtc, u64 crtc_base);
 extern void rs600_post_page_flip(struct radeon_device *rdev, int crtc);
 void rs600_set_safe_registers(struct radeon_device *rdev);
-extern void avivo_wait_for_vblank(struct radeon_device *rdev, int crtc);
-extern int rs600_mc_wait_for_idle(struct radeon_device *rdev);
+
 
 /*
  * rs690,rs740
@@ -256,7 +248,6 @@ void rs690_bandwidth_update(struct radeon_device *rdev);
 void rs690_line_buffer_adjust(struct radeon_device *rdev,
 					struct drm_display_mode *mode1,
 					struct drm_display_mode *mode2);
-extern int rs690_mc_wait_for_idle(struct radeon_device *rdev);
 
 /*
  * rv515
@@ -270,7 +261,7 @@ int rv515_init(struct radeon_device *rdev);
 void rv515_fini(struct radeon_device *rdev);
 uint32_t rv515_mc_rreg(struct radeon_device *rdev, uint32_t reg);
 void rv515_mc_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
-void rv515_ring_start(struct radeon_device *rdev, struct radeon_ring *ring);
+void rv515_ring_start(struct radeon_device *rdev);
 void rv515_bandwidth_update(struct radeon_device *rdev);
 int rv515_resume(struct radeon_device *rdev);
 int rv515_suspend(struct radeon_device *rdev);
@@ -281,14 +272,13 @@ void rv515_mc_stop(struct radeon_device *rdev, struct rv515_mc_save *save);
 void rv515_mc_resume(struct radeon_device *rdev, struct rv515_mc_save *save);
 void rv515_clock_startup(struct radeon_device *rdev);
 void rv515_debugfs(struct radeon_device *rdev);
-int rv515_mc_wait_for_idle(struct radeon_device *rdev);
+
 
 /*
  * r520,rv530,rv560,rv570,r580
  */
 int r520_init(struct radeon_device *rdev);
 int r520_resume(struct radeon_device *rdev);
-int r520_mc_wait_for_idle(struct radeon_device *rdev);
 
 /*
  * r600,rv610,rv630,rv620,rv635,rv670,rs780,rs880
@@ -300,25 +290,22 @@ int r600_resume(struct radeon_device *rdev);
 void r600_vga_set_state(struct radeon_device *rdev, bool state);
 int r600_wb_init(struct radeon_device *rdev);
 void r600_wb_fini(struct radeon_device *rdev);
+void r600_cp_commit(struct radeon_device *rdev);
 void r600_pcie_gart_tlb_flush(struct radeon_device *rdev);
 uint32_t r600_pciep_rreg(struct radeon_device *rdev, uint32_t reg);
 void r600_pciep_wreg(struct radeon_device *rdev, uint32_t reg, uint32_t v);
 int r600_cs_parse(struct radeon_cs_parser *p);
 void r600_fence_ring_emit(struct radeon_device *rdev,
 			  struct radeon_fence *fence);
-void r600_semaphore_ring_emit(struct radeon_device *rdev,
-			      struct radeon_ring *cp,
-			      struct radeon_semaphore *semaphore,
-			      bool emit_wait);
-bool r600_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp);
+bool r600_gpu_is_lockup(struct radeon_device *rdev);
 int r600_asic_reset(struct radeon_device *rdev);
 int r600_set_surface_reg(struct radeon_device *rdev, int reg,
 			 uint32_t tiling_flags, uint32_t pitch,
 			 uint32_t offset, uint32_t obj_size);
 void r600_clear_surface_reg(struct radeon_device *rdev, int reg);
-int r600_ib_test(struct radeon_device *rdev, struct radeon_ring *ring);
+int r600_ib_test(struct radeon_device *rdev);
 void r600_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib);
-int r600_ring_test(struct radeon_device *rdev, struct radeon_ring *cp);
+int r600_ring_test(struct radeon_device *rdev);
 int r600_copy_blit(struct radeon_device *rdev,
 		   uint64_t src_offset, uint64_t dst_offset,
 		   unsigned num_gpu_pages, struct radeon_fence *fence);
@@ -338,7 +325,7 @@ extern int r600_get_pcie_lanes(struct radeon_device *rdev);
 bool r600_card_posted(struct radeon_device *rdev);
 void r600_cp_stop(struct radeon_device *rdev);
 int r600_cp_start(struct radeon_device *rdev);
-void r600_ring_init(struct radeon_device *rdev, struct radeon_ring *cp, unsigned ring_size);
+void r600_ring_init(struct radeon_device *rdev, unsigned ring_size);
 int r600_cp_resume(struct radeon_device *rdev);
 void r600_cp_fini(struct radeon_device *rdev);
 int r600_count_pipe_bits(uint32_t val);
@@ -374,12 +361,11 @@ void r600_hdmi_init(struct drm_encoder *encoder);
 int r600_hdmi_buffer_status_changed(struct drm_encoder *encoder);
 void r600_hdmi_update_audio_settings(struct drm_encoder *encoder);
 /* r600 blit */
-int r600_blit_prepare_copy(struct radeon_device *rdev, unsigned num_gpu_pages);
+int r600_blit_prepare_copy(struct radeon_device *rdev, int size_bytes);
 void r600_blit_done_copy(struct radeon_device *rdev, struct radeon_fence *fence);
 void r600_kms_blit_copy(struct radeon_device *rdev,
 			u64 src_gpu_addr, u64 dst_gpu_addr,
-			unsigned num_gpu_pages);
-int r600_mc_wait_for_idle(struct radeon_device *rdev);
+			int size_bytes);
 
 /*
  * rv770,rv730,rv710,rv740
@@ -407,10 +393,13 @@ int evergreen_init(struct radeon_device *rdev);
 void evergreen_fini(struct radeon_device *rdev);
 int evergreen_suspend(struct radeon_device *rdev);
 int evergreen_resume(struct radeon_device *rdev);
-bool evergreen_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp);
+bool evergreen_gpu_is_lockup(struct radeon_device *rdev);
 int evergreen_asic_reset(struct radeon_device *rdev);
 void evergreen_bandwidth_update(struct radeon_device *rdev);
 void evergreen_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib);
+int evergreen_copy_blit(struct radeon_device *rdev,
+			uint64_t src_offset, uint64_t dst_offset,
+			unsigned num_gpu_pages, struct radeon_fence *fence);
 void evergreen_hpd_init(struct radeon_device *rdev);
 void evergreen_hpd_fini(struct radeon_device *rdev);
 bool evergreen_hpd_sense(struct radeon_device *rdev, enum radeon_hpd_id hpd);
@@ -423,38 +412,28 @@ extern int evergreen_cs_parse(struct radeon_cs_parser *p);
 extern void evergreen_pm_misc(struct radeon_device *rdev);
 extern void evergreen_pm_prepare(struct radeon_device *rdev);
 extern void evergreen_pm_finish(struct radeon_device *rdev);
-extern void sumo_pm_init_profile(struct radeon_device *rdev);
 extern void evergreen_pre_page_flip(struct radeon_device *rdev, int crtc);
 extern u32 evergreen_page_flip(struct radeon_device *rdev, int crtc, u64 crtc_base);
 extern void evergreen_post_page_flip(struct radeon_device *rdev, int crtc);
-extern void dce4_wait_for_vblank(struct radeon_device *rdev, int crtc);
 void evergreen_disable_interrupt_state(struct radeon_device *rdev);
 int evergreen_blit_init(struct radeon_device *rdev);
-int evergreen_mc_wait_for_idle(struct radeon_device *rdev);
+void evergreen_blit_fini(struct radeon_device *rdev);
+/* evergreen blit */
+int evergreen_blit_prepare_copy(struct radeon_device *rdev, int size_bytes);
+void evergreen_blit_done_copy(struct radeon_device *rdev, struct radeon_fence *fence);
+void evergreen_kms_blit_copy(struct radeon_device *rdev,
+			     u64 src_gpu_addr, u64 dst_gpu_addr,
+			     int size_bytes);
 
 /*
  * cayman
  */
-void cayman_fence_ring_emit(struct radeon_device *rdev,
-			    struct radeon_fence *fence);
 void cayman_pcie_gart_tlb_flush(struct radeon_device *rdev);
 int cayman_init(struct radeon_device *rdev);
 void cayman_fini(struct radeon_device *rdev);
 int cayman_suspend(struct radeon_device *rdev);
 int cayman_resume(struct radeon_device *rdev);
-bool cayman_gpu_is_lockup(struct radeon_device *rdev, struct radeon_ring *cp);
+bool cayman_gpu_is_lockup(struct radeon_device *rdev);
 int cayman_asic_reset(struct radeon_device *rdev);
-void cayman_ring_ib_execute(struct radeon_device *rdev, struct radeon_ib *ib);
-int cayman_vm_init(struct radeon_device *rdev);
-void cayman_vm_fini(struct radeon_device *rdev);
-int cayman_vm_bind(struct radeon_device *rdev, struct radeon_vm *vm, int id);
-void cayman_vm_unbind(struct radeon_device *rdev, struct radeon_vm *vm);
-void cayman_vm_tlb_flush(struct radeon_device *rdev, struct radeon_vm *vm);
-uint32_t cayman_vm_page_flags(struct radeon_device *rdev,
-			      struct radeon_vm *vm,
-			      uint32_t flags);
-void cayman_vm_set_page(struct radeon_device *rdev, struct radeon_vm *vm,
-			unsigned pfn, uint64_t addr, uint32_t flags);
-int evergreen_ib_parse(struct radeon_device *rdev, struct radeon_ib *ib);
 
 #endif
diff --git a/drivers/gpu/drm/radeon/radeon_atombios.c b/drivers/gpu/drm/radeon/radeon_atombios.c
index 97ab2ec..f9d49e3 100644
--- a/drivers/gpu/drm/radeon/radeon_atombios.c
+++ b/drivers/gpu/drm/radeon/radeon_atombios.c
@@ -62,88 +62,7 @@ union atom_supported_devices {
 	struct _ATOM_SUPPORTED_DEVICES_INFO_2d1 info_2d1;
 };
 
-static void radeon_lookup_i2c_gpio_quirks(struct radeon_device *rdev,
-					  ATOM_GPIO_I2C_ASSIGMENT *gpio,
-					  u8 index)
-{
-	/* r4xx mask is technically not used by the hw, so patch in the legacy mask bits */
-	if ((rdev->family == CHIP_R420) ||
-	    (rdev->family == CHIP_R423) ||
-	    (rdev->family == CHIP_RV410)) {
-		if ((le16_to_cpu(gpio->usClkMaskRegisterIndex) == 0x0018) ||
-		    (le16_to_cpu(gpio->usClkMaskRegisterIndex) == 0x0019) ||
-		    (le16_to_cpu(gpio->usClkMaskRegisterIndex) == 0x001a)) {
-			gpio->ucClkMaskShift = 0x19;
-			gpio->ucDataMaskShift = 0x18;
-		}
-	}
-
-	/* some evergreen boards have bad data for this entry */
-	if (ASIC_IS_DCE4(rdev)) {
-		if ((index == 7) &&
-		    (le16_to_cpu(gpio->usClkMaskRegisterIndex) == 0x1936) &&
-		    (gpio->sucI2cId.ucAccess == 0)) {
-			gpio->sucI2cId.ucAccess = 0x97;
-			gpio->ucDataMaskShift = 8;
-			gpio->ucDataEnShift = 8;
-			gpio->ucDataY_Shift = 8;
-			gpio->ucDataA_Shift = 8;
-		}
-	}
-
-	/* some DCE3 boards have bad data for this entry */
-	if (ASIC_IS_DCE3(rdev)) {
-		if ((index == 4) &&
-		    (le16_to_cpu(gpio->usClkMaskRegisterIndex) == 0x1fda) &&
-		    (gpio->sucI2cId.ucAccess == 0x94))
-			gpio->sucI2cId.ucAccess = 0x14;
-	}
-}
-
-static struct radeon_i2c_bus_rec radeon_get_bus_rec_for_i2c_gpio(ATOM_GPIO_I2C_ASSIGMENT *gpio)
-{
-	struct radeon_i2c_bus_rec i2c;
-
-	memset(&i2c, 0, sizeof(struct radeon_i2c_bus_rec));
-
-	i2c.mask_clk_reg = le16_to_cpu(gpio->usClkMaskRegisterIndex) * 4;
-	i2c.mask_data_reg = le16_to_cpu(gpio->usDataMaskRegisterIndex) * 4;
-	i2c.en_clk_reg = le16_to_cpu(gpio->usClkEnRegisterIndex) * 4;
-	i2c.en_data_reg = le16_to_cpu(gpio->usDataEnRegisterIndex) * 4;
-	i2c.y_clk_reg = le16_to_cpu(gpio->usClkY_RegisterIndex) * 4;
-	i2c.y_data_reg = le16_to_cpu(gpio->usDataY_RegisterIndex) * 4;
-	i2c.a_clk_reg = le16_to_cpu(gpio->usClkA_RegisterIndex) * 4;
-	i2c.a_data_reg = le16_to_cpu(gpio->usDataA_RegisterIndex) * 4;
-	i2c.mask_clk_mask = (1 << gpio->ucClkMaskShift);
-	i2c.mask_data_mask = (1 << gpio->ucDataMaskShift);
-	i2c.en_clk_mask = (1 << gpio->ucClkEnShift);
-	i2c.en_data_mask = (1 << gpio->ucDataEnShift);
-	i2c.y_clk_mask = (1 << gpio->ucClkY_Shift);
-	i2c.y_data_mask = (1 << gpio->ucDataY_Shift);
-	i2c.a_clk_mask = (1 << gpio->ucClkA_Shift);
-	i2c.a_data_mask = (1 << gpio->ucDataA_Shift);
-
-	if (gpio->sucI2cId.sbfAccess.bfHW_Capable)
-		i2c.hw_capable = true;
-	else
-		i2c.hw_capable = false;
-
-	if (gpio->sucI2cId.ucAccess == 0xa0)
-		i2c.mm_i2c = true;
-	else
-		i2c.mm_i2c = false;
-
-	i2c.i2c_id = gpio->sucI2cId.ucAccess;
-
-	if (i2c.mask_clk_reg)
-		i2c.valid = true;
-	else
-		i2c.valid = false;
-
-	return i2c;
-}
-
-static struct radeon_i2c_bus_rec radeon_lookup_i2c_gpio(struct radeon_device *rdev,
+static inline struct radeon_i2c_bus_rec radeon_lookup_i2c_gpio(struct radeon_device *rdev,
 							       uint8_t id)
 {
 	struct atom_context *ctx = rdev->mode_info.atom_context;
@@ -166,10 +85,71 @@ static struct radeon_i2c_bus_rec radeon_lookup_i2c_gpio(struct radeon_device *rd
 		for (i = 0; i < num_indices; i++) {
 			gpio = &i2c_info->asGPIO_Info[i];
 
-			radeon_lookup_i2c_gpio_quirks(rdev, gpio, i);
+			/* r4xx mask is technically not used by the hw, so patch in the legacy mask bits */
+			if ((rdev->family == CHIP_R420) ||
+			    (rdev->family == CHIP_R423) ||
+			    (rdev->family == CHIP_RV410)) {
+				if ((le16_to_cpu(gpio->usClkMaskRegisterIndex) == 0x0018) ||
+				    (le16_to_cpu(gpio->usClkMaskRegisterIndex) == 0x0019) ||
+				    (le16_to_cpu(gpio->usClkMaskRegisterIndex) == 0x001a)) {
+					gpio->ucClkMaskShift = 0x19;
+					gpio->ucDataMaskShift = 0x18;
+				}
+			}
+
+			/* some evergreen boards have bad data for this entry */
+			if (ASIC_IS_DCE4(rdev)) {
+				if ((i == 7) &&
+				    (le16_to_cpu(gpio->usClkMaskRegisterIndex) == 0x1936) &&
+				    (gpio->sucI2cId.ucAccess == 0)) {
+					gpio->sucI2cId.ucAccess = 0x97;
+					gpio->ucDataMaskShift = 8;
+					gpio->ucDataEnShift = 8;
+					gpio->ucDataY_Shift = 8;
+					gpio->ucDataA_Shift = 8;
+				}
+			}
+
+			/* some DCE3 boards have bad data for this entry */
+			if (ASIC_IS_DCE3(rdev)) {
+				if ((i == 4) &&
+				    (le16_to_cpu(gpio->usClkMaskRegisterIndex) == 0x1fda) &&
+				    (gpio->sucI2cId.ucAccess == 0x94))
+					gpio->sucI2cId.ucAccess = 0x14;
+			}
 
 			if (gpio->sucI2cId.ucAccess == id) {
-				i2c = radeon_get_bus_rec_for_i2c_gpio(gpio);
+				i2c.mask_clk_reg = le16_to_cpu(gpio->usClkMaskRegisterIndex) * 4;
+				i2c.mask_data_reg = le16_to_cpu(gpio->usDataMaskRegisterIndex) * 4;
+				i2c.en_clk_reg = le16_to_cpu(gpio->usClkEnRegisterIndex) * 4;
+				i2c.en_data_reg = le16_to_cpu(gpio->usDataEnRegisterIndex) * 4;
+				i2c.y_clk_reg = le16_to_cpu(gpio->usClkY_RegisterIndex) * 4;
+				i2c.y_data_reg = le16_to_cpu(gpio->usDataY_RegisterIndex) * 4;
+				i2c.a_clk_reg = le16_to_cpu(gpio->usClkA_RegisterIndex) * 4;
+				i2c.a_data_reg = le16_to_cpu(gpio->usDataA_RegisterIndex) * 4;
+				i2c.mask_clk_mask = (1 << gpio->ucClkMaskShift);
+				i2c.mask_data_mask = (1 << gpio->ucDataMaskShift);
+				i2c.en_clk_mask = (1 << gpio->ucClkEnShift);
+				i2c.en_data_mask = (1 << gpio->ucDataEnShift);
+				i2c.y_clk_mask = (1 << gpio->ucClkY_Shift);
+				i2c.y_data_mask = (1 << gpio->ucDataY_Shift);
+				i2c.a_clk_mask = (1 << gpio->ucClkA_Shift);
+				i2c.a_data_mask = (1 << gpio->ucDataA_Shift);
+
+				if (gpio->sucI2cId.sbfAccess.bfHW_Capable)
+					i2c.hw_capable = true;
+				else
+					i2c.hw_capable = false;
+
+				if (gpio->sucI2cId.ucAccess == 0xa0)
+					i2c.mm_i2c = true;
+				else
+					i2c.mm_i2c = false;
+
+				i2c.i2c_id = gpio->sucI2cId.ucAccess;
+
+				if (i2c.mask_clk_reg)
+					i2c.valid = true;
 				break;
 			}
 		}
@@ -189,6 +169,8 @@ void radeon_atombios_i2c_init(struct radeon_device *rdev)
 	int i, num_indices;
 	char stmp[32];
 
+	memset(&i2c, 0, sizeof(struct radeon_i2c_bus_rec));
+
 	if (atom_parse_data_header(ctx, index, &size, NULL, NULL, &data_offset)) {
 		i2c_info = (struct _ATOM_GPIO_I2C_INFO *)(ctx->bios + data_offset);
 
@@ -197,12 +179,72 @@ void radeon_atombios_i2c_init(struct radeon_device *rdev)
 
 		for (i = 0; i < num_indices; i++) {
 			gpio = &i2c_info->asGPIO_Info[i];
+			i2c.valid = false;
+
+			/* r4xx mask is technically not used by the hw, so patch in the legacy mask bits */
+			if ((rdev->family == CHIP_R420) ||
+			    (rdev->family == CHIP_R423) ||
+			    (rdev->family == CHIP_RV410)) {
+				if ((le16_to_cpu(gpio->usClkMaskRegisterIndex) == 0x0018) ||
+				    (le16_to_cpu(gpio->usClkMaskRegisterIndex) == 0x0019) ||
+				    (le16_to_cpu(gpio->usClkMaskRegisterIndex) == 0x001a)) {
+					gpio->ucClkMaskShift = 0x19;
+					gpio->ucDataMaskShift = 0x18;
+				}
+			}
+
+			/* some evergreen boards have bad data for this entry */
+			if (ASIC_IS_DCE4(rdev)) {
+				if ((i == 7) &&
+				    (le16_to_cpu(gpio->usClkMaskRegisterIndex) == 0x1936) &&
+				    (gpio->sucI2cId.ucAccess == 0)) {
+					gpio->sucI2cId.ucAccess = 0x97;
+					gpio->ucDataMaskShift = 8;
+					gpio->ucDataEnShift = 8;
+					gpio->ucDataY_Shift = 8;
+					gpio->ucDataA_Shift = 8;
+				}
+			}
+
+			/* some DCE3 boards have bad data for this entry */
+			if (ASIC_IS_DCE3(rdev)) {
+				if ((i == 4) &&
+				    (le16_to_cpu(gpio->usClkMaskRegisterIndex) == 0x1fda) &&
+				    (gpio->sucI2cId.ucAccess == 0x94))
+					gpio->sucI2cId.ucAccess = 0x14;
+			}
 
-			radeon_lookup_i2c_gpio_quirks(rdev, gpio, i);
+			i2c.mask_clk_reg = le16_to_cpu(gpio->usClkMaskRegisterIndex) * 4;
+			i2c.mask_data_reg = le16_to_cpu(gpio->usDataMaskRegisterIndex) * 4;
+			i2c.en_clk_reg = le16_to_cpu(gpio->usClkEnRegisterIndex) * 4;
+			i2c.en_data_reg = le16_to_cpu(gpio->usDataEnRegisterIndex) * 4;
+			i2c.y_clk_reg = le16_to_cpu(gpio->usClkY_RegisterIndex) * 4;
+			i2c.y_data_reg = le16_to_cpu(gpio->usDataY_RegisterIndex) * 4;
+			i2c.a_clk_reg = le16_to_cpu(gpio->usClkA_RegisterIndex) * 4;
+			i2c.a_data_reg = le16_to_cpu(gpio->usDataA_RegisterIndex) * 4;
+			i2c.mask_clk_mask = (1 << gpio->ucClkMaskShift);
+			i2c.mask_data_mask = (1 << gpio->ucDataMaskShift);
+			i2c.en_clk_mask = (1 << gpio->ucClkEnShift);
+			i2c.en_data_mask = (1 << gpio->ucDataEnShift);
+			i2c.y_clk_mask = (1 << gpio->ucClkY_Shift);
+			i2c.y_data_mask = (1 << gpio->ucDataY_Shift);
+			i2c.a_clk_mask = (1 << gpio->ucClkA_Shift);
+			i2c.a_data_mask = (1 << gpio->ucDataA_Shift);
+
+			if (gpio->sucI2cId.sbfAccess.bfHW_Capable)
+				i2c.hw_capable = true;
+			else
+				i2c.hw_capable = false;
 
-			i2c = radeon_get_bus_rec_for_i2c_gpio(gpio);
+			if (gpio->sucI2cId.ucAccess == 0xa0)
+				i2c.mm_i2c = true;
+			else
+				i2c.mm_i2c = false;
 
-			if (i2c.valid) {
+			i2c.i2c_id = gpio->sucI2cId.ucAccess;
+
+			if (i2c.mask_clk_reg) {
+				i2c.valid = true;
 				sprintf(stmp, "0x%x", i2c.i2c_id);
 				rdev->i2c_bus[i] = radeon_i2c_create(rdev->ddev, &i2c, stmp);
 			}
@@ -210,7 +252,7 @@ void radeon_atombios_i2c_init(struct radeon_device *rdev)
 	}
 }
 
-static struct radeon_gpio_rec radeon_lookup_gpio(struct radeon_device *rdev,
+static inline struct radeon_gpio_rec radeon_lookup_gpio(struct radeon_device *rdev,
 							u8 id)
 {
 	struct atom_context *ctx = rdev->mode_info.atom_context;
@@ -2001,14 +2043,10 @@ static int radeon_atombios_parse_power_table_1_3(struct radeon_device *rdev)
 		return state_index;
 	/* last mode is usually default, array is low to high */
 	for (i = 0; i < num_modes; i++) {
-		rdev->pm.power_state[state_index].clock_info =
-			kzalloc(sizeof(struct radeon_pm_clock_info) * 1, GFP_KERNEL);
-		if (!rdev->pm.power_state[state_index].clock_info)
-			return state_index;
-		rdev->pm.power_state[state_index].num_clock_modes = 1;
 		rdev->pm.power_state[state_index].clock_info[0].voltage.type = VOLTAGE_NONE;
 		switch (frev) {
 		case 1:
+			rdev->pm.power_state[state_index].num_clock_modes = 1;
 			rdev->pm.power_state[state_index].clock_info[0].mclk =
 				le16_to_cpu(power_info->info.asPowerPlayInfo[i].usMemoryClock);
 			rdev->pm.power_state[state_index].clock_info[0].sclk =
@@ -2044,6 +2082,7 @@ static int radeon_atombios_parse_power_table_1_3(struct radeon_device *rdev)
 			state_index++;
 			break;
 		case 2:
+			rdev->pm.power_state[state_index].num_clock_modes = 1;
 			rdev->pm.power_state[state_index].clock_info[0].mclk =
 				le32_to_cpu(power_info->info_2.asPowerPlayInfo[i].ulMemoryClock);
 			rdev->pm.power_state[state_index].clock_info[0].sclk =
@@ -2080,6 +2119,7 @@ static int radeon_atombios_parse_power_table_1_3(struct radeon_device *rdev)
 			state_index++;
 			break;
 		case 3:
+			rdev->pm.power_state[state_index].num_clock_modes = 1;
 			rdev->pm.power_state[state_index].clock_info[0].mclk =
 				le32_to_cpu(power_info->info_3.asPowerPlayInfo[i].ulMemoryClock);
 			rdev->pm.power_state[state_index].clock_info[0].sclk =
@@ -2264,7 +2304,7 @@ static void radeon_atombios_parse_pplib_non_clock_info(struct radeon_device *rde
 		rdev->pm.default_power_state_index = state_index;
 		rdev->pm.power_state[state_index].default_clock_mode =
 			&rdev->pm.power_state[state_index].clock_info[mode_index - 1];
-		if (ASIC_IS_DCE5(rdev) && !(rdev->flags & RADEON_IS_IGP)) {
+		if (ASIC_IS_DCE5(rdev)) {
 			/* NI chips post without MC ucode, so default clocks are strobe mode only */
 			rdev->pm.default_sclk = rdev->pm.power_state[state_index].clock_info[0].sclk;
 			rdev->pm.default_mclk = rdev->pm.power_state[state_index].clock_info[0].mclk;
@@ -2386,31 +2426,17 @@ static int radeon_atombios_parse_power_table_4_5(struct radeon_device *rdev)
 			 le16_to_cpu(power_info->pplib.usNonClockInfoArrayOffset) +
 			 (power_state->v1.ucNonClockStateIndex *
 			  power_info->pplib.ucNonClockSize));
-		rdev->pm.power_state[i].clock_info = kzalloc(sizeof(struct radeon_pm_clock_info) *
-							     ((power_info->pplib.ucStateEntrySize - 1) ?
-							      (power_info->pplib.ucStateEntrySize - 1) : 1),
-							     GFP_KERNEL);
-		if (!rdev->pm.power_state[i].clock_info)
-			return state_index;
-		if (power_info->pplib.ucStateEntrySize - 1) {
-			for (j = 0; j < (power_info->pplib.ucStateEntrySize - 1); j++) {
-				clock_info = (union pplib_clock_info *)
-					(mode_info->atom_context->bios + data_offset +
-					 le16_to_cpu(power_info->pplib.usClockInfoArrayOffset) +
-					 (power_state->v1.ucClockStateIndices[j] *
-					  power_info->pplib.ucClockInfoSize));
-				valid = radeon_atombios_parse_pplib_clock_info(rdev,
-									       state_index, mode_index,
-									       clock_info);
-				if (valid)
-					mode_index++;
-			}
-		} else {
-			rdev->pm.power_state[state_index].clock_info[0].mclk =
-				rdev->clock.default_mclk;
-			rdev->pm.power_state[state_index].clock_info[0].sclk =
-				rdev->clock.default_sclk;
-			mode_index++;
+		for (j = 0; j < (power_info->pplib.ucStateEntrySize - 1); j++) {
+			clock_info = (union pplib_clock_info *)
+				(mode_info->atom_context->bios + data_offset +
+				 le16_to_cpu(power_info->pplib.usClockInfoArrayOffset) +
+				 (power_state->v1.ucClockStateIndices[j] *
+				  power_info->pplib.ucClockInfoSize));
+			valid = radeon_atombios_parse_pplib_clock_info(rdev,
+								       state_index, mode_index,
+								       clock_info);
+			if (valid)
+				mode_index++;
 		}
 		rdev->pm.power_state[state_index].num_clock_modes = mode_index;
 		if (mode_index) {
@@ -2481,32 +2507,18 @@ static int radeon_atombios_parse_power_table_6(struct radeon_device *rdev)
 		non_clock_array_index = i; /* power_state->v2.nonClockInfoIndex */
 		non_clock_info = (struct _ATOM_PPLIB_NONCLOCK_INFO *)
 			&non_clock_info_array->nonClockInfo[non_clock_array_index];
-		rdev->pm.power_state[i].clock_info = kzalloc(sizeof(struct radeon_pm_clock_info) *
-							     (power_state->v2.ucNumDPMLevels ?
-							      power_state->v2.ucNumDPMLevels : 1),
-							     GFP_KERNEL);
-		if (!rdev->pm.power_state[i].clock_info)
-			return state_index;
-		if (power_state->v2.ucNumDPMLevels) {
-			for (j = 0; j < power_state->v2.ucNumDPMLevels; j++) {
-				clock_array_index = power_state->v2.clockInfoIndex[j];
-				/* XXX this might be an inagua bug... */
-				if (clock_array_index >= clock_info_array->ucNumEntries)
-					continue;
-				clock_info = (union pplib_clock_info *)
-					&clock_info_array->clockInfo[clock_array_index];
-				valid = radeon_atombios_parse_pplib_clock_info(rdev,
-									       state_index, mode_index,
-									       clock_info);
-				if (valid)
-					mode_index++;
-			}
-		} else {
-			rdev->pm.power_state[state_index].clock_info[0].mclk =
-				rdev->clock.default_mclk;
-			rdev->pm.power_state[state_index].clock_info[0].sclk =
-				rdev->clock.default_sclk;
-			mode_index++;
+		for (j = 0; j < power_state->v2.ucNumDPMLevels; j++) {
+			clock_array_index = power_state->v2.clockInfoIndex[j];
+			/* XXX this might be an inagua bug... */
+			if (clock_array_index >= clock_info_array->ucNumEntries)
+				continue;
+			clock_info = (union pplib_clock_info *)
+				&clock_info_array->clockInfo[clock_array_index];
+			valid = radeon_atombios_parse_pplib_clock_info(rdev,
+								       state_index, mode_index,
+								       clock_info);
+			if (valid)
+				mode_index++;
 		}
 		rdev->pm.power_state[state_index].num_clock_modes = mode_index;
 		if (mode_index) {
@@ -2565,23 +2577,19 @@ void radeon_atombios_get_power_modes(struct radeon_device *rdev)
 	if (state_index == 0) {
 		rdev->pm.power_state = kzalloc(sizeof(struct radeon_power_state), GFP_KERNEL);
 		if (rdev->pm.power_state) {
-			rdev->pm.power_state[0].clock_info =
-				kzalloc(sizeof(struct radeon_pm_clock_info) * 1, GFP_KERNEL);
-			if (rdev->pm.power_state[0].clock_info) {
-				/* add the default mode */
-				rdev->pm.power_state[state_index].type =
-					POWER_STATE_TYPE_DEFAULT;
-				rdev->pm.power_state[state_index].num_clock_modes = 1;
-				rdev->pm.power_state[state_index].clock_info[0].mclk = rdev->clock.default_mclk;
-				rdev->pm.power_state[state_index].clock_info[0].sclk = rdev->clock.default_sclk;
-				rdev->pm.power_state[state_index].default_clock_mode =
-					&rdev->pm.power_state[state_index].clock_info[0];
-				rdev->pm.power_state[state_index].clock_info[0].voltage.type = VOLTAGE_NONE;
-				rdev->pm.power_state[state_index].pcie_lanes = 16;
-				rdev->pm.default_power_state_index = state_index;
-				rdev->pm.power_state[state_index].flags = 0;
-				state_index++;
-			}
+			/* add the default mode */
+			rdev->pm.power_state[state_index].type =
+				POWER_STATE_TYPE_DEFAULT;
+			rdev->pm.power_state[state_index].num_clock_modes = 1;
+			rdev->pm.power_state[state_index].clock_info[0].mclk = rdev->clock.default_mclk;
+			rdev->pm.power_state[state_index].clock_info[0].sclk = rdev->clock.default_sclk;
+			rdev->pm.power_state[state_index].default_clock_mode =
+				&rdev->pm.power_state[state_index].clock_info[0];
+			rdev->pm.power_state[state_index].clock_info[0].voltage.type = VOLTAGE_NONE;
+			rdev->pm.power_state[state_index].pcie_lanes = 16;
+			rdev->pm.default_power_state_index = state_index;
+			rdev->pm.power_state[state_index].flags = 0;
+			state_index++;
 		}
 	}
 
@@ -2960,20 +2968,6 @@ radeon_atombios_connected_scratch_regs(struct drm_connector *connector,
 			bios_6_scratch &= ~ATOM_S6_ACC_REQ_DFP5;
 		}
 	}
-	if ((radeon_encoder->devices & ATOM_DEVICE_DFP6_SUPPORT) &&
-	    (radeon_connector->devices & ATOM_DEVICE_DFP6_SUPPORT)) {
-		if (connected) {
-			DRM_DEBUG_KMS("DFP6 connected\n");
-			bios_0_scratch |= ATOM_S0_DFP6;
-			bios_3_scratch |= ATOM_S3_DFP6_ACTIVE;
-			bios_6_scratch |= ATOM_S6_ACC_REQ_DFP6;
-		} else {
-			DRM_DEBUG_KMS("DFP6 disconnected\n");
-			bios_0_scratch &= ~ATOM_S0_DFP6;
-			bios_3_scratch &= ~ATOM_S3_DFP6_ACTIVE;
-			bios_6_scratch &= ~ATOM_S6_ACC_REQ_DFP6;
-		}
-	}
 
 	if (rdev->family >= CHIP_R600) {
 		WREG32(R600_BIOS_0_SCRATCH, bios_0_scratch);
@@ -2994,9 +2988,6 @@ radeon_atombios_encoder_crtc_scratch_regs(struct drm_encoder *encoder, int crtc)
 	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
 	uint32_t bios_3_scratch;
 
-	if (ASIC_IS_DCE4(rdev))
-		return;
-
 	if (rdev->family >= CHIP_R600)
 		bios_3_scratch = RREG32(R600_BIOS_3_SCRATCH);
 	else
@@ -3049,9 +3040,6 @@ radeon_atombios_encoder_dpms_scratch_regs(struct drm_encoder *encoder, bool on)
 	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
 	uint32_t bios_2_scratch;
 
-	if (ASIC_IS_DCE4(rdev))
-		return;
-
 	if (rdev->family >= CHIP_R600)
 		bios_2_scratch = RREG32(R600_BIOS_2_SCRATCH);
 	else
diff --git a/drivers/gpu/drm/radeon/radeon_atpx_handler.c b/drivers/gpu/drm/radeon/radeon_atpx_handler.c
index 98724fc..9d95792 100644
--- a/drivers/gpu/drm/radeon/radeon_atpx_handler.c
+++ b/drivers/gpu/drm/radeon/radeon_atpx_handler.c
@@ -58,8 +58,7 @@ static int radeon_atrm_call(acpi_handle atrm_handle, uint8_t *bios,
 	}
 
 	obj = (union acpi_object *)buffer.pointer;
-	memcpy(bios+offset, obj->buffer.pointer, obj->buffer.length);
-	len = obj->buffer.length;
+	memcpy(bios+offset, obj->buffer.pointer, len);
 	kfree(buffer.pointer);
 	return len;
 }
diff --git a/drivers/gpu/drm/radeon/radeon_benchmark.c b/drivers/gpu/drm/radeon/radeon_benchmark.c
index fef7b72..10191d9 100644
--- a/drivers/gpu/drm/radeon/radeon_benchmark.c
+++ b/drivers/gpu/drm/radeon/radeon_benchmark.c
@@ -26,83 +26,21 @@
 #include "radeon_reg.h"
 #include "radeon.h"
 
-#define RADEON_BENCHMARK_COPY_BLIT 1
-#define RADEON_BENCHMARK_COPY_DMA  0
-
-#define RADEON_BENCHMARK_ITERATIONS 1024
-#define RADEON_BENCHMARK_COMMON_MODES_N 17
-
-static int radeon_benchmark_do_move(struct radeon_device *rdev, unsigned size,
-				    uint64_t saddr, uint64_t daddr,
-				    int flag, int n)
-{
-	unsigned long start_jiffies;
-	unsigned long end_jiffies;
-	struct radeon_fence *fence = NULL;
-	int i, r;
-
-	start_jiffies = jiffies;
-	for (i = 0; i < n; i++) {
-		switch (flag) {
-		case RADEON_BENCHMARK_COPY_DMA:
-			r = radeon_fence_create(rdev, &fence, radeon_copy_dma_ring_index(rdev));
-			if (r)
-				return r;
-			r = radeon_copy_dma(rdev, saddr, daddr,
-					    size / RADEON_GPU_PAGE_SIZE,
-					    fence);
-			break;
-		case RADEON_BENCHMARK_COPY_BLIT:
-			r = radeon_fence_create(rdev, &fence, radeon_copy_blit_ring_index(rdev));
-			if (r)
-				return r;
-			r = radeon_copy_blit(rdev, saddr, daddr,
-					     size / RADEON_GPU_PAGE_SIZE,
-					     fence);
-			break;
-		default:
-			DRM_ERROR("Unknown copy method\n");
-			r = -EINVAL;
-		}
-		if (r)
-			goto exit_do_move;
-		r = radeon_fence_wait(fence, false);
-		if (r)
-			goto exit_do_move;
-		radeon_fence_unref(&fence);
-	}
-	end_jiffies = jiffies;
-	r = jiffies_to_msecs(end_jiffies - start_jiffies);
-
-exit_do_move:
-	if (fence)
-		radeon_fence_unref(&fence);
-	return r;
-}
-
-
-static void radeon_benchmark_log_results(int n, unsigned size,
-					 unsigned int time,
-					 unsigned sdomain, unsigned ddomain,
-					 char *kind)
-{
-	unsigned int throughput = (n * (size >> 10)) / time;
-	DRM_INFO("radeon: %s %u bo moves of %u kB from"
-		 " %d to %d in %u ms, throughput: %u Mb/s or %u MB/s\n",
-		 kind, n, size >> 10, sdomain, ddomain, time,
-		 throughput * 8, throughput);
-}
-
-static void radeon_benchmark_move(struct radeon_device *rdev, unsigned size,
-				  unsigned sdomain, unsigned ddomain)
+void radeon_benchmark_move(struct radeon_device *rdev, unsigned bsize,
+			   unsigned sdomain, unsigned ddomain)
 {
 	struct radeon_bo *dobj = NULL;
 	struct radeon_bo *sobj = NULL;
+	struct radeon_fence *fence = NULL;
 	uint64_t saddr, daddr;
-	int r, n;
-	int time;
+	unsigned long start_jiffies;
+	unsigned long end_jiffies;
+	unsigned long time;
+	unsigned i, n, size;
+	int r;
 
-	n = RADEON_BENCHMARK_ITERATIONS;
+	size = bsize;
+	n = 1024;
 	r = radeon_bo_create(rdev, size, PAGE_SIZE, true, sdomain, &sobj);
 	if (r) {
 		goto out_cleanup;
@@ -129,26 +67,65 @@ static void radeon_benchmark_move(struct radeon_device *rdev, unsigned size,
 	}
 
 	/* r100 doesn't have dma engine so skip the test */
-	/* also, VRAM-to-VRAM test doesn't make much sense for DMA */
-	/* skip it as well if domains are the same */
-	if ((rdev->asic->copy.dma) && (sdomain != ddomain)) {
-		time = radeon_benchmark_do_move(rdev, size, saddr, daddr,
-						RADEON_BENCHMARK_COPY_DMA, n);
-		if (time < 0)
-			goto out_cleanup;
-		if (time > 0)
-			radeon_benchmark_log_results(n, size, time,
-						     sdomain, ddomain, "dma");
-	}
+	if (rdev->asic->copy_dma) {
 
-	time = radeon_benchmark_do_move(rdev, size, saddr, daddr,
-					RADEON_BENCHMARK_COPY_BLIT, n);
-	if (time < 0)
-		goto out_cleanup;
-	if (time > 0)
-		radeon_benchmark_log_results(n, size, time,
-					     sdomain, ddomain, "blit");
+		start_jiffies = jiffies;
+		for (i = 0; i < n; i++) {
+			r = radeon_fence_create(rdev, &fence);
+			if (r) {
+				goto out_cleanup;
+			}
+
+			r = radeon_copy_dma(rdev, saddr, daddr,
+					size / RADEON_GPU_PAGE_SIZE, fence);
+
+			if (r) {
+				goto out_cleanup;
+			}
+			r = radeon_fence_wait(fence, false);
+			if (r) {
+				goto out_cleanup;
+			}
+			radeon_fence_unref(&fence);
+		}
+		end_jiffies = jiffies;
+		time = end_jiffies - start_jiffies;
+		time = jiffies_to_msecs(time);
+		if (time > 0) {
+			i = ((n * size) >> 10) / time;
+			printk(KERN_INFO "radeon: dma %u bo moves of %ukb from"
+					" %d to %d in %lums (%ukb/ms %ukb/s %uM/s)\n",
+					n, size >> 10,
+					sdomain, ddomain, time,
+					i, i * 1000, (i * 1000) / 1024);
+		}
+	}
 
+	start_jiffies = jiffies;
+	for (i = 0; i < n; i++) {
+		r = radeon_fence_create(rdev, &fence);
+		if (r) {
+			goto out_cleanup;
+		}
+		r = radeon_copy_blit(rdev, saddr, daddr, size / RADEON_GPU_PAGE_SIZE, fence);
+		if (r) {
+			goto out_cleanup;
+		}
+		r = radeon_fence_wait(fence, false);
+		if (r) {
+			goto out_cleanup;
+		}
+		radeon_fence_unref(&fence);
+	}
+	end_jiffies = jiffies;
+	time = end_jiffies - start_jiffies;
+	time = jiffies_to_msecs(time);
+	if (time > 0) {
+		i = ((n * size) >> 10) / time;
+		printk(KERN_INFO "radeon: blit %u bo moves of %ukb from %d to %d"
+		       " in %lums (%ukb/ms %ukb/s %uM/s)\n", n, size >> 10,
+		       sdomain, ddomain, time, i, i * 1000, (i * 1000) / 1024);
+	}
 out_cleanup:
 	if (sobj) {
 		r = radeon_bo_reserve(sobj, false);
@@ -166,92 +143,18 @@ out_cleanup:
 		}
 		radeon_bo_unref(&dobj);
 	}
-
+	if (fence) {
+		radeon_fence_unref(&fence);
+	}
 	if (r) {
-		DRM_ERROR("Error while benchmarking BO move.\n");
+		printk(KERN_WARNING "Error while benchmarking BO move.\n");
 	}
 }
 
-void radeon_benchmark(struct radeon_device *rdev, int test_number)
+void radeon_benchmark(struct radeon_device *rdev)
 {
-	int i;
-	int common_modes[RADEON_BENCHMARK_COMMON_MODES_N] = {
-		640 * 480 * 4,
-		720 * 480 * 4,
-		800 * 600 * 4,
-		848 * 480 * 4,
-		1024 * 768 * 4,
-		1152 * 768 * 4,
-		1280 * 720 * 4,
-		1280 * 800 * 4,
-		1280 * 854 * 4,
-		1280 * 960 * 4,
-		1280 * 1024 * 4,
-		1440 * 900 * 4,
-		1400 * 1050 * 4,
-		1680 * 1050 * 4,
-		1600 * 1200 * 4,
-		1920 * 1080 * 4,
-		1920 * 1200 * 4
-	};
-
-	switch (test_number) {
-	case 1:
-		/* simple test, VRAM to GTT and GTT to VRAM */
-		radeon_benchmark_move(rdev, 1024*1024, RADEON_GEM_DOMAIN_GTT,
-				      RADEON_GEM_DOMAIN_VRAM);
-		radeon_benchmark_move(rdev, 1024*1024, RADEON_GEM_DOMAIN_VRAM,
-				      RADEON_GEM_DOMAIN_GTT);
-		break;
-	case 2:
-		/* simple test, VRAM to VRAM */
-		radeon_benchmark_move(rdev, 1024*1024, RADEON_GEM_DOMAIN_VRAM,
-				      RADEON_GEM_DOMAIN_VRAM);
-		break;
-	case 3:
-		/* GTT to VRAM, buffer size sweep, powers of 2 */
-		for (i = 1; i <= 16384; i <<= 1)
-			radeon_benchmark_move(rdev, i * RADEON_GPU_PAGE_SIZE,
-					      RADEON_GEM_DOMAIN_GTT,
-					      RADEON_GEM_DOMAIN_VRAM);
-		break;
-	case 4:
-		/* VRAM to GTT, buffer size sweep, powers of 2 */
-		for (i = 1; i <= 16384; i <<= 1)
-			radeon_benchmark_move(rdev, i * RADEON_GPU_PAGE_SIZE,
-					      RADEON_GEM_DOMAIN_VRAM,
-					      RADEON_GEM_DOMAIN_GTT);
-		break;
-	case 5:
-		/* VRAM to VRAM, buffer size sweep, powers of 2 */
-		for (i = 1; i <= 16384; i <<= 1)
-			radeon_benchmark_move(rdev, i * RADEON_GPU_PAGE_SIZE,
-					      RADEON_GEM_DOMAIN_VRAM,
-					      RADEON_GEM_DOMAIN_VRAM);
-		break;
-	case 6:
-		/* GTT to VRAM, buffer size sweep, common modes */
-		for (i = 0; i < RADEON_BENCHMARK_COMMON_MODES_N; i++)
-			radeon_benchmark_move(rdev, common_modes[i],
-					      RADEON_GEM_DOMAIN_GTT,
-					      RADEON_GEM_DOMAIN_VRAM);
-		break;
-	case 7:
-		/* VRAM to GTT, buffer size sweep, common modes */
-		for (i = 0; i < RADEON_BENCHMARK_COMMON_MODES_N; i++)
-			radeon_benchmark_move(rdev, common_modes[i],
-					      RADEON_GEM_DOMAIN_VRAM,
-					      RADEON_GEM_DOMAIN_GTT);
-		break;
-	case 8:
-		/* VRAM to VRAM, buffer size sweep, common modes */
-		for (i = 0; i < RADEON_BENCHMARK_COMMON_MODES_N; i++)
-			radeon_benchmark_move(rdev, common_modes[i],
-					      RADEON_GEM_DOMAIN_VRAM,
-					      RADEON_GEM_DOMAIN_VRAM);
-		break;
-
-	default:
-		DRM_ERROR("Unknown benchmark\n");
-	}
+	radeon_benchmark_move(rdev, 1024*1024, RADEON_GEM_DOMAIN_GTT,
+			      RADEON_GEM_DOMAIN_VRAM);
+	radeon_benchmark_move(rdev, 1024*1024, RADEON_GEM_DOMAIN_VRAM,
+			      RADEON_GEM_DOMAIN_GTT);
 }
diff --git a/drivers/gpu/drm/radeon/radeon_bios.c b/drivers/gpu/drm/radeon/radeon_bios.c
index 501f488..229a20f 100644
--- a/drivers/gpu/drm/radeon/radeon_bios.c
+++ b/drivers/gpu/drm/radeon/radeon_bios.c
@@ -120,7 +120,7 @@ static bool radeon_atrm_get_bios(struct radeon_device *rdev)
 		ret = radeon_atrm_get_bios_chunk(rdev->bios,
 						 (i * ATRM_BIOS_PAGE),
 						 ATRM_BIOS_PAGE);
-		if (ret < ATRM_BIOS_PAGE)
+		if (ret <= 0)
 			break;
 	}
 
diff --git a/drivers/gpu/drm/radeon/radeon_blit_common.h b/drivers/gpu/drm/radeon/radeon_blit_common.h
deleted file mode 100644
index 4ecbe72..0000000
--- a/drivers/gpu/drm/radeon/radeon_blit_common.h
+++ /dev/null
@@ -1,44 +0,0 @@
-/*
- * Copyright 2009 Advanced Micro Devices, Inc.
- * Copyright 2009 Red Hat Inc.
- * Copyright 2012 Alcatel-Lucent, Inc.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice (including the next
- * paragraph) shall be included in all copies or substantial portions of the
- * Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * THE COPYRIGHT HOLDER(S) AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR
- * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
- * DEALINGS IN THE SOFTWARE.
- *
- */
-
-#ifndef __RADEON_BLIT_COMMON_H__
-
-#define DI_PT_RECTLIST        0x11
-#define DI_INDEX_SIZE_16_BIT  0x0
-#define DI_SRC_SEL_AUTO_INDEX 0x2
-
-#define FMT_8                 0x1
-#define FMT_5_6_5             0x8
-#define FMT_8_8_8_8           0x1a
-#define COLOR_8               0x1
-#define COLOR_5_6_5           0x8
-#define COLOR_8_8_8_8         0x1a
-
-#define RECT_UNIT_H           32
-#define RECT_UNIT_W           (RADEON_GPU_PAGE_SIZE / 4 / RECT_UNIT_H)
-
-#define __RADEON_BLIT_COMMON_H__
-#endif
diff --git a/drivers/gpu/drm/radeon/radeon_clocks.c b/drivers/gpu/drm/radeon/radeon_clocks.c
index 6ae0c75..b956cf1 100644
--- a/drivers/gpu/drm/radeon/radeon_clocks.c
+++ b/drivers/gpu/drm/radeon/radeon_clocks.c
@@ -96,7 +96,7 @@ uint32_t radeon_legacy_get_memory_clock(struct radeon_device *rdev)
  * Read XTAL (ref clock), SCLK and MCLK from Open Firmware device
  * tree. Hopefully, ATI OF driver is kind enough to fill these
  */
-static bool radeon_read_clocks_OF(struct drm_device *dev)
+static bool __devinit radeon_read_clocks_OF(struct drm_device *dev)
 {
 	struct radeon_device *rdev = dev->dev_private;
 	struct device_node *dp = rdev->pdev->dev.of_node;
@@ -166,7 +166,7 @@ static bool radeon_read_clocks_OF(struct drm_device *dev)
 	return true;
 }
 #else
-static bool radeon_read_clocks_OF(struct drm_device *dev)
+static bool __devinit radeon_read_clocks_OF(struct drm_device *dev)
 {
 	return false;
 }
@@ -334,7 +334,7 @@ void radeon_get_clock_info(struct drm_device *dev)
 
 	if (!rdev->clock.default_sclk)
 		rdev->clock.default_sclk = radeon_get_engine_clock(rdev);
-	if ((!rdev->clock.default_mclk) && rdev->asic->pm.get_memory_clock)
+	if ((!rdev->clock.default_mclk) && rdev->asic->get_memory_clock)
 		rdev->clock.default_mclk = radeon_get_memory_clock(rdev);
 
 	rdev->pm.current_sclk = rdev->clock.default_sclk;
diff --git a/drivers/gpu/drm/radeon/radeon_combios.c b/drivers/gpu/drm/radeon/radeon_combios.c
index e158e68..f616e40 100644
--- a/drivers/gpu/drm/radeon/radeon_combios.c
+++ b/drivers/gpu/drm/radeon/radeon_combios.c
@@ -2513,28 +2513,25 @@ void radeon_combios_get_power_modes(struct radeon_device *rdev)
 	u16 offset, misc, misc2 = 0;
 	u8 rev, blocks, tmp;
 	int state_index = 0;
-	struct radeon_i2c_bus_rec i2c_bus;
 
 	rdev->pm.default_power_state_index = -1;
 
 	/* allocate 2 power states */
 	rdev->pm.power_state = kzalloc(sizeof(struct radeon_power_state) * 2, GFP_KERNEL);
-	if (rdev->pm.power_state) {
-		/* allocate 1 clock mode per state */
-		rdev->pm.power_state[0].clock_info =
-			kzalloc(sizeof(struct radeon_pm_clock_info) * 1, GFP_KERNEL);
-		rdev->pm.power_state[1].clock_info =
-			kzalloc(sizeof(struct radeon_pm_clock_info) * 1, GFP_KERNEL);
-		if (!rdev->pm.power_state[0].clock_info ||
-		    !rdev->pm.power_state[1].clock_info)
-			goto pm_failed;
-	} else
-		goto pm_failed;
+	if (!rdev->pm.power_state) {
+		rdev->pm.default_power_state_index = state_index;
+		rdev->pm.num_power_states = 0;
+
+		rdev->pm.current_power_state_index = rdev->pm.default_power_state_index;
+		rdev->pm.current_clock_mode_index = 0;
+		return;
+	}
 
 	/* check for a thermal chip */
 	offset = combios_get_table_offset(dev, COMBIOS_OVERDRIVE_INFO_TABLE);
 	if (offset) {
 		u8 thermal_controller = 0, gpio = 0, i2c_addr = 0, clk_bit = 0, data_bit = 0;
+		struct radeon_i2c_bus_rec i2c_bus;
 
 		rev = RBIOS8(offset);
 
@@ -2576,25 +2573,6 @@ void radeon_combios_get_power_modes(struct radeon_device *rdev)
 				i2c_new_device(&rdev->pm.i2c_bus->adapter, &info);
 			}
 		}
-	} else {
-		/* boards with a thermal chip, but no overdrive table */
-
-		/* Asus 9600xt has an f75375 on the monid bus */
-		if ((dev->pdev->device == 0x4152) &&
-		    (dev->pdev->subsystem_vendor == 0x1043) &&
-		    (dev->pdev->subsystem_device == 0xc002)) {
-			i2c_bus = combios_setup_i2c_bus(rdev, DDC_MONID, 0, 0);
-			rdev->pm.i2c_bus = radeon_i2c_lookup(rdev, &i2c_bus);
-			if (rdev->pm.i2c_bus) {
-				struct i2c_board_info info = { };
-				const char *name = "f75375";
-				info.addr = 0x28;
-				strlcpy(info.type, name, sizeof(info.type));
-				i2c_new_device(&rdev->pm.i2c_bus->adapter, &info);
-				DRM_INFO("Possible %s thermal controller at 0x%02x\n",
-					 name, info.addr);
-			}
-		}
 	}
 
 	if (rdev->flags & RADEON_IS_MOBILITY) {
@@ -2694,14 +2672,6 @@ default_mode:
 
 	rdev->pm.current_power_state_index = rdev->pm.default_power_state_index;
 	rdev->pm.current_clock_mode_index = 0;
-	return;
-
-pm_failed:
-	rdev->pm.default_power_state_index = state_index;
-	rdev->pm.num_power_states = 0;
-
-	rdev->pm.current_power_state_index = rdev->pm.default_power_state_index;
-	rdev->pm.current_clock_mode_index = 0;
 }
 
 void radeon_external_tmds_setup(struct drm_encoder *encoder)
diff --git a/drivers/gpu/drm/radeon/radeon_connectors.c b/drivers/gpu/drm/radeon/radeon_connectors.c
index 4c4c924..f1a1e8a 100644
--- a/drivers/gpu/drm/radeon/radeon_connectors.c
+++ b/drivers/gpu/drm/radeon/radeon_connectors.c
@@ -44,6 +44,8 @@ extern void
 radeon_legacy_backlight_init(struct radeon_encoder *radeon_encoder,
 			     struct drm_connector *drm_connector);
 
+bool radeon_connector_encoder_is_dp_bridge(struct drm_connector *connector);
+
 void radeon_connector_hotplug(struct drm_connector *connector)
 {
 	struct drm_device *dev = connector->dev;
@@ -449,24 +451,65 @@ int radeon_connector_set_property(struct drm_connector *connector, struct drm_pr
 	return 0;
 }
 
+/*
+ * Some integrated ATI Radeon chipset implementations (e. g.
+ * Asus M2A-VM HDMI) may indicate the availability of a DDC,
+ * even when there's no monitor connected. For these connectors
+ * following DDC probe extension will be applied: check also for the
+ * availability of EDID with at least a correct EDID header. Only then,
+ * DDC is assumed to be available. This prevents drm_get_edid() and
+ * drm_edid_block_valid() from periodically dumping data and kernel
+ * errors into the logs and onto the terminal.
+ */
+static bool radeon_connector_needs_extended_probe(struct radeon_device *dev,
+				     uint32_t supported_device,
+				     int connector_type)
+{
+	/* Asus M2A-VM HDMI board sends data to i2c bus even,
+	 * if HDMI add-on card is not plugged in or HDMI is disabled in
+	 * BIOS. Valid DDC can only be assumed, if also a valid EDID header
+	 * can be retrieved via i2c bus during DDC probe */
+	if ((dev->pdev->device == 0x791e) &&
+	    (dev->pdev->subsystem_vendor == 0x1043) &&
+	    (dev->pdev->subsystem_device == 0x826d)) {
+		if ((connector_type == DRM_MODE_CONNECTOR_HDMIA) &&
+		    (supported_device == ATOM_DEVICE_DFP2_SUPPORT))
+			return true;
+	}
+	/* ECS A740GM-M with ATI RADEON 2100 sends data to i2c bus
+	 * for a DVI connector that is not implemented */
+	if ((dev->pdev->device == 0x796e) &&
+	    (dev->pdev->subsystem_vendor == 0x1019) &&
+	    (dev->pdev->subsystem_device == 0x2615)) {
+		if ((connector_type == DRM_MODE_CONNECTOR_DVID) &&
+		    (supported_device == ATOM_DEVICE_DFP2_SUPPORT))
+			return true;
+	}
+	/* TOSHIBA Satellite L300D with ATI Mobility Radeon x1100
+	 * (RS690M) sends data to i2c bus for a HDMI connector that
+	 * is not implemented */
+	if ((dev->pdev->device == 0x791f) &&
+	    (dev->pdev->subsystem_vendor == 0x1179) &&
+	    (dev->pdev->subsystem_device == 0xff68)) {
+		if ((connector_type == DRM_MODE_CONNECTOR_HDMIA) &&
+		    (supported_device == ATOM_DEVICE_DFP2_SUPPORT))
+			return true;
+	}
+
+	/* Default: no EDID header probe required for DDC probing */
+	return false;
+}
+
 static void radeon_fixup_lvds_native_mode(struct drm_encoder *encoder,
 					  struct drm_connector *connector)
 {
 	struct radeon_encoder *radeon_encoder =	to_radeon_encoder(encoder);
 	struct drm_display_mode *native_mode = &radeon_encoder->native_mode;
-	struct drm_display_mode *t, *mode;
-
-	/* If the EDID preferred mode doesn't match the native mode, use it */
-	list_for_each_entry_safe(mode, t, &connector->probed_modes, head) {
-		if (mode->type & DRM_MODE_TYPE_PREFERRED) {
-			if (mode->hdisplay != native_mode->hdisplay ||
-			    mode->vdisplay != native_mode->vdisplay)
-				memcpy(native_mode, mode, sizeof(*mode));
-		}
-	}
 
 	/* Try to get native mode details from EDID if necessary */
 	if (!native_mode->clock) {
+		struct drm_display_mode *t, *mode;
+
 		list_for_each_entry_safe(mode, t, &connector->probed_modes, head) {
 			if (mode->hdisplay == native_mode->hdisplay &&
 			    mode->vdisplay == native_mode->vdisplay) {
@@ -477,7 +520,6 @@ static void radeon_fixup_lvds_native_mode(struct drm_encoder *encoder,
 			}
 		}
 	}
-
 	if (!native_mode->clock) {
 		DRM_DEBUG_KMS("No LVDS native mode details, disabling RMX\n");
 		radeon_encoder->rmx_type = RMX_OFF;
@@ -689,7 +731,8 @@ radeon_vga_detect(struct drm_connector *connector, bool force)
 		ret = connector_status_disconnected;
 
 	if (radeon_connector->ddc_bus)
-		dret = radeon_ddc_probe(radeon_connector);
+		dret = radeon_ddc_probe(radeon_connector,
+					radeon_connector->requires_extended_probe);
 	if (dret) {
 		radeon_connector->detected_by_load = false;
 		if (radeon_connector->edid) {
@@ -846,27 +889,6 @@ static int radeon_dvi_get_modes(struct drm_connector *connector)
 	return ret;
 }
 
-static bool radeon_check_hpd_status_unchanged(struct drm_connector *connector)
-{
-	struct drm_device *dev = connector->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
-	enum drm_connector_status status;
-
-	/* We only trust HPD on R600 and newer ASICS. */
-	if (rdev->family >= CHIP_R600
-	  && radeon_connector->hpd.hpd != RADEON_HPD_NONE) {
-		if (radeon_hpd_sense(rdev, radeon_connector->hpd.hpd))
-			status = connector_status_connected;
-		else
-			status = connector_status_disconnected;
-		if (connector->status == status)
-			return true;
-	}
-
-	return false;
-}
-
 /*
  * DVI is complicated
  * Do a DDC probe, if DDC probe passes, get the full EDID so
@@ -891,11 +913,9 @@ radeon_dvi_detect(struct drm_connector *connector, bool force)
 	enum drm_connector_status ret = connector_status_disconnected;
 	bool dret = false;
 
-	if (!force && radeon_check_hpd_status_unchanged(connector))
-		return connector->status;
-
 	if (radeon_connector->ddc_bus)
-		dret = radeon_ddc_probe(radeon_connector);
+		dret = radeon_ddc_probe(radeon_connector,
+					radeon_connector->requires_extended_probe);
 	if (dret) {
 		radeon_connector->detected_by_load = false;
 		if (radeon_connector->edid) {
@@ -1104,7 +1124,7 @@ static int radeon_dvi_mode_valid(struct drm_connector *connector,
 		    (radeon_connector->connector_object_id == CONNECTOR_OBJECT_ID_HDMI_TYPE_B))
 			return MODE_OK;
 		else if (radeon_connector->connector_object_id == CONNECTOR_OBJECT_ID_HDMI_TYPE_A) {
-			if (0) {
+			if (ASIC_IS_DCE3(rdev)) {
 				/* HDMI 1.3+ supports max clock of 340 Mhz */
 				if (mode->clock > 340000)
 					return MODE_CLOCK_HIGH;
@@ -1164,23 +1184,13 @@ static int radeon_dp_get_modes(struct drm_connector *connector)
 	    (connector->connector_type == DRM_MODE_CONNECTOR_LVDS)) {
 		struct drm_display_mode *mode;
 
-		if (connector->connector_type == DRM_MODE_CONNECTOR_eDP) {
-			if (!radeon_dig_connector->edp_on)
-				atombios_set_edp_panel_power(connector,
-							     ATOM_TRANSMITTER_ACTION_POWER_ON);
-			ret = radeon_ddc_get_modes(radeon_connector);
-			if (!radeon_dig_connector->edp_on)
-				atombios_set_edp_panel_power(connector,
-							     ATOM_TRANSMITTER_ACTION_POWER_OFF);
-		} else {
-			/* need to setup ddc on the bridge */
-			if (radeon_connector_encoder_get_dp_bridge_encoder_id(connector) !=
-			    ENCODER_OBJECT_ID_NONE) {
-				if (encoder)
-					radeon_atom_ext_encoder_setup_ddc(encoder);
-			}
-			ret = radeon_ddc_get_modes(radeon_connector);
-		}
+		if (!radeon_dig_connector->edp_on)
+			atombios_set_edp_panel_power(connector,
+						     ATOM_TRANSMITTER_ACTION_POWER_ON);
+		ret = radeon_ddc_get_modes(radeon_connector);
+		if (!radeon_dig_connector->edp_on)
+			atombios_set_edp_panel_power(connector,
+						     ATOM_TRANSMITTER_ACTION_POWER_OFF);
 
 		if (ret > 0) {
 			if (encoder) {
@@ -1191,6 +1201,7 @@ static int radeon_dp_get_modes(struct drm_connector *connector)
 			return ret;
 		}
 
+		encoder = radeon_best_single_encoder(connector);
 		if (!encoder)
 			return 0;
 
@@ -1207,8 +1218,7 @@ static int radeon_dp_get_modes(struct drm_connector *connector)
 		}
 	} else {
 		/* need to setup ddc on the bridge */
-		if (radeon_connector_encoder_get_dp_bridge_encoder_id(connector) !=
-			ENCODER_OBJECT_ID_NONE) {
+		if (radeon_connector_encoder_is_dp_bridge(connector)) {
 			if (encoder)
 				radeon_atom_ext_encoder_setup_ddc(encoder);
 		}
@@ -1218,12 +1228,13 @@ static int radeon_dp_get_modes(struct drm_connector *connector)
 	return ret;
 }
 
-u16 radeon_connector_encoder_get_dp_bridge_encoder_id(struct drm_connector *connector)
+bool radeon_connector_encoder_is_dp_bridge(struct drm_connector *connector)
 {
 	struct drm_mode_object *obj;
 	struct drm_encoder *encoder;
 	struct radeon_encoder *radeon_encoder;
 	int i;
+	bool found = false;
 
 	for (i = 0; i < DRM_CONNECTOR_MAX_ENCODER; i++) {
 		if (connector->encoder_ids[i] == 0)
@@ -1239,13 +1250,14 @@ u16 radeon_connector_encoder_get_dp_bridge_encoder_id(struct drm_connector *conn
 		switch (radeon_encoder->encoder_id) {
 		case ENCODER_OBJECT_ID_TRAVIS:
 		case ENCODER_OBJECT_ID_NUTMEG:
-			return radeon_encoder->encoder_id;
+			found = true;
+			break;
 		default:
 			break;
 		}
 	}
 
-	return ENCODER_OBJECT_ID_NONE;
+	return found;
 }
 
 bool radeon_connector_encoder_is_hbr2(struct drm_connector *connector)
@@ -1297,9 +1309,6 @@ radeon_dp_detect(struct drm_connector *connector, bool force)
 	struct radeon_connector_atom_dig *radeon_dig_connector = radeon_connector->con_priv;
 	struct drm_encoder *encoder = radeon_best_single_encoder(connector);
 
-	if (!force && radeon_check_hpd_status_unchanged(connector))
-		return connector->status;
-
 	if (radeon_connector->edid) {
 		kfree(radeon_connector->edid);
 		radeon_connector->edid = NULL;
@@ -1325,24 +1334,12 @@ radeon_dp_detect(struct drm_connector *connector, bool force)
 		if (!radeon_dig_connector->edp_on)
 			atombios_set_edp_panel_power(connector,
 						     ATOM_TRANSMITTER_ACTION_POWER_OFF);
-	} else if (radeon_connector_encoder_get_dp_bridge_encoder_id(connector) !=
-		   ENCODER_OBJECT_ID_NONE) {
-		/* DP bridges are always DP */
-		radeon_dig_connector->dp_sink_type = CONNECTOR_OBJECT_ID_DISPLAYPORT;
-		/* get the DPCD from the bridge */
-		radeon_dp_getdpcd(radeon_connector);
-
-		if (encoder) {
-			/* setup ddc on the bridge */
-			radeon_atom_ext_encoder_setup_ddc(encoder);
-			if (radeon_ddc_probe(radeon_connector)) /* try DDC */
-				ret = connector_status_connected;
-			else if (radeon_connector->dac_load_detect) { /* try load detection */
-				struct drm_encoder_helper_funcs *encoder_funcs = encoder->helper_private;
-				ret = encoder_funcs->detect(encoder, connector);
-			}
-		}
 	} else {
+		/* need to setup ddc on the bridge */
+		if (radeon_connector_encoder_is_dp_bridge(connector)) {
+			if (encoder)
+				radeon_atom_ext_encoder_setup_ddc(encoder);
+		}
 		radeon_dig_connector->dp_sink_type = radeon_dp_getsinktype(radeon_connector);
 		if (radeon_hpd_sense(rdev, radeon_connector->hpd.hpd)) {
 			ret = connector_status_connected;
@@ -1353,10 +1350,21 @@ radeon_dp_detect(struct drm_connector *connector, bool force)
 				if (radeon_dp_getdpcd(radeon_connector))
 					ret = connector_status_connected;
 			} else {
-				if (radeon_ddc_probe(radeon_connector))
+				if (radeon_ddc_probe(radeon_connector,
+						     radeon_connector->requires_extended_probe))
 					ret = connector_status_connected;
 			}
 		}
+
+		if ((ret == connector_status_disconnected) &&
+		    radeon_connector->dac_load_detect) {
+			struct drm_encoder *encoder = radeon_best_single_encoder(connector);
+			struct drm_encoder_helper_funcs *encoder_funcs;
+			if (encoder) {
+				encoder_funcs = encoder->helper_private;
+				ret = encoder_funcs->detect(encoder, connector);
+			}
+		}
 	}
 
 	radeon_connector_update_scratch_regs(connector, ret);
@@ -1498,7 +1506,9 @@ radeon_add_atom_connector(struct drm_device *dev,
 	radeon_connector->shared_ddc = shared_ddc;
 	radeon_connector->connector_object_id = connector_object_id;
 	radeon_connector->hpd = *hpd;
-
+	radeon_connector->requires_extended_probe =
+		radeon_connector_needs_extended_probe(rdev, supported_device,
+							connector_type);
 	radeon_connector->router = *router;
 	if (router->ddc_valid || router->cd_valid) {
 		radeon_connector->router_bus = radeon_i2c_lookup(rdev, &router->i2c_info);
@@ -1845,7 +1855,9 @@ radeon_add_legacy_connector(struct drm_device *dev,
 	radeon_connector->devices = supported_device;
 	radeon_connector->connector_object_id = connector_object_id;
 	radeon_connector->hpd = *hpd;
-
+	radeon_connector->requires_extended_probe =
+		radeon_connector_needs_extended_probe(rdev, supported_device,
+							connector_type);
 	switch (connector_type) {
 	case DRM_MODE_CONNECTOR_VGA:
 		drm_connector_init(dev, &radeon_connector->base, &radeon_vga_connector_funcs, connector_type);
diff --git a/drivers/gpu/drm/radeon/radeon_cp.c b/drivers/gpu/drm/radeon/radeon_cp.c
index 0ebb7d4..7586779 100644
--- a/drivers/gpu/drm/radeon/radeon_cp.c
+++ b/drivers/gpu/drm/radeon/radeon_cp.c
@@ -29,8 +29,6 @@
  *    Gareth Hughes <gareth@valinux.com>
  */
 
-#include <linux/module.h>
-
 #include "drmP.h"
 #include "drm.h"
 #include "drm_sarea.h"
@@ -2115,11 +2113,9 @@ int radeon_driver_load(struct drm_device *dev, unsigned long flags)
 		break;
 	}
 
-	pci_set_master(dev->pdev);
-
 	if (drm_pci_device_is_agp(dev))
 		dev_priv->flags |= RADEON_IS_AGP;
-	else if (pci_is_pcie(dev->pdev))
+	else if (drm_pci_device_is_pcie(dev))
 		dev_priv->flags |= RADEON_IS_PCIE;
 	else
 		dev_priv->flags |= RADEON_IS_PCI;
diff --git a/drivers/gpu/drm/radeon/radeon_cs.c b/drivers/gpu/drm/radeon/radeon_cs.c
index d9d9f5a..fae00c0 100644
--- a/drivers/gpu/drm/radeon/radeon_cs.c
+++ b/drivers/gpu/drm/radeon/radeon_cs.c
@@ -58,7 +58,7 @@ int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
 
 		duplicate = false;
 		r = (struct drm_radeon_cs_reloc *)&chunk->kdata[i*4];
-		for (j = 0; j < i; j++) {
+		for (j = 0; j < p->nrelocs; j++) {
 			if (r->handle == p->relocs[j].handle) {
 				p->relocs_ptr[i] = &p->relocs[j];
 				duplicate = true;
@@ -84,73 +84,9 @@ int radeon_cs_parser_relocs(struct radeon_cs_parser *p)
 			p->relocs[i].flags = r->flags;
 			radeon_bo_list_add_object(&p->relocs[i].lobj,
 						  &p->validated);
-
-		} else
-			p->relocs[i].handle = 0;
-	}
-	return radeon_bo_list_validate(&p->validated);
-}
-
-static int radeon_cs_get_ring(struct radeon_cs_parser *p, u32 ring, s32 priority)
-{
-	p->priority = priority;
-
-	switch (ring) {
-	default:
-		DRM_ERROR("unknown ring id: %d\n", ring);
-		return -EINVAL;
-	case RADEON_CS_RING_GFX:
-		p->ring = RADEON_RING_TYPE_GFX_INDEX;
-		break;
-	case RADEON_CS_RING_COMPUTE:
-		/* for now */
-		p->ring = RADEON_RING_TYPE_GFX_INDEX;
-		break;
-	}
-	return 0;
-}
-
-static int radeon_cs_sync_rings(struct radeon_cs_parser *p)
-{
-	bool sync_to_ring[RADEON_NUM_RINGS] = { };
-	int i, r;
-
-	for (i = 0; i < p->nrelocs; i++) {
-		if (!p->relocs[i].robj || !p->relocs[i].robj->tbo.sync_obj)
-			continue;
-
-		if (!(p->relocs[i].flags & RADEON_RELOC_DONT_SYNC)) {
-			struct radeon_fence *fence = p->relocs[i].robj->tbo.sync_obj;
-			if (!radeon_fence_signaled(fence)) {
-				sync_to_ring[fence->ring] = true;
-			}
-		}
-	}
-
-	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
-		/* no need to sync to our own or unused rings */
-		if (i == p->ring || !sync_to_ring[i] || !p->rdev->ring[i].ready)
-			continue;
-
-		if (!p->ib->fence->semaphore) {
-			r = radeon_semaphore_create(p->rdev, &p->ib->fence->semaphore);
-			if (r)
-				return r;
 		}
-
-		r = radeon_ring_lock(p->rdev, &p->rdev->ring[i], 3);
-		if (r)
-			return r;
-		radeon_semaphore_emit_signal(p->rdev, i, p->ib->fence->semaphore);
-		radeon_ring_unlock_commit(p->rdev, &p->rdev->ring[i]);
-
-		r = radeon_ring_lock(p->rdev, &p->rdev->ring[p->ring], 3);
-		if (r)
-			return r;
-		radeon_semaphore_emit_wait(p->rdev, p->ring, p->ib->fence->semaphore);
-		radeon_ring_unlock_commit(p->rdev, &p->rdev->ring[p->ring]);
 	}
-	return 0;
+	return radeon_bo_list_validate(&p->validated);
 }
 
 int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
@@ -158,8 +94,6 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 	struct drm_radeon_cs *cs = data;
 	uint64_t *chunk_array_ptr;
 	unsigned size, i;
-	u32 ring = RADEON_CS_RING_GFX;
-	s32 priority = 0;
 
 	if (!cs->num_chunks) {
 		return 0;
@@ -169,7 +103,6 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 	p->idx = 0;
 	p->chunk_ib_idx = -1;
 	p->chunk_relocs_idx = -1;
-	p->chunk_flags_idx = -1;
 	p->chunks_array = kcalloc(cs->num_chunks, sizeof(uint64_t), GFP_KERNEL);
 	if (p->chunks_array == NULL) {
 		return -ENOMEM;
@@ -179,7 +112,6 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 			       sizeof(uint64_t)*cs->num_chunks)) {
 		return -EFAULT;
 	}
-	p->cs_flags = 0;
 	p->nchunks = cs->num_chunks;
 	p->chunks = kcalloc(p->nchunks, sizeof(struct radeon_cs_chunk), GFP_KERNEL);
 	if (p->chunks == NULL) {
@@ -208,19 +140,12 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 			if (p->chunks[i].length_dw == 0)
 				return -EINVAL;
 		}
-		if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_FLAGS) {
-			p->chunk_flags_idx = i;
-			/* zero length flags aren't useful */
-			if (p->chunks[i].length_dw == 0)
-				return -EINVAL;
-		}
 
 		p->chunks[i].length_dw = user_chunk.length_dw;
 		p->chunks[i].user_ptr = (void __user *)(unsigned long)user_chunk.chunk_data;
 
 		cdata = (uint32_t *)(unsigned long)user_chunk.chunk_data;
-		if ((p->chunks[i].chunk_id == RADEON_CHUNK_ID_RELOCS) ||
-		    (p->chunks[i].chunk_id == RADEON_CHUNK_ID_FLAGS)) {
+		if (p->chunks[i].chunk_id != RADEON_CHUNK_ID_IB) {
 			size = p->chunks[i].length_dw * sizeof(uint32_t);
 			p->chunks[i].kdata = kmalloc(size, GFP_KERNEL);
 			if (p->chunks[i].kdata == NULL) {
@@ -230,47 +155,25 @@ int radeon_cs_parser_init(struct radeon_cs_parser *p, void *data)
 					       p->chunks[i].user_ptr, size)) {
 				return -EFAULT;
 			}
-			if (p->chunks[i].chunk_id == RADEON_CHUNK_ID_FLAGS) {
-				p->cs_flags = p->chunks[i].kdata[0];
-				if (p->chunks[i].length_dw > 1)
-					ring = p->chunks[i].kdata[1];
-				if (p->chunks[i].length_dw > 2)
-					priority = (s32)p->chunks[i].kdata[2];
+		} else {
+			p->chunks[i].kpage[0] = kmalloc(PAGE_SIZE, GFP_KERNEL);
+			p->chunks[i].kpage[1] = kmalloc(PAGE_SIZE, GFP_KERNEL);
+			if (p->chunks[i].kpage[0] == NULL || p->chunks[i].kpage[1] == NULL) {
+				kfree(p->chunks[i].kpage[0]);
+				kfree(p->chunks[i].kpage[1]);
+				return -ENOMEM;
 			}
+			p->chunks[i].kpage_idx[0] = -1;
+			p->chunks[i].kpage_idx[1] = -1;
+			p->chunks[i].last_copied_page = -1;
+			p->chunks[i].last_page_index = ((p->chunks[i].length_dw * 4) - 1) / PAGE_SIZE;
 		}
 	}
-
-	if ((p->cs_flags & RADEON_CS_USE_VM) &&
-	    !p->rdev->vm_manager.enabled) {
-		DRM_ERROR("VM not active on asic!\n");
+	if (p->chunks[p->chunk_ib_idx].length_dw > (16 * 1024)) {
+		DRM_ERROR("cs IB too big: %d\n",
+			  p->chunks[p->chunk_ib_idx].length_dw);
 		return -EINVAL;
 	}
-
-	if (radeon_cs_get_ring(p, ring, priority))
-		return -EINVAL;
-
-
-	/* deal with non-vm */
-	if ((p->chunk_ib_idx != -1) &&
-	    ((p->cs_flags & RADEON_CS_USE_VM) == 0) &&
-	    (p->chunks[p->chunk_ib_idx].chunk_id == RADEON_CHUNK_ID_IB)) {
-		if (p->chunks[p->chunk_ib_idx].length_dw > (16 * 1024)) {
-			DRM_ERROR("cs IB too big: %d\n",
-				  p->chunks[p->chunk_ib_idx].length_dw);
-			return -EINVAL;
-		}
-		p->chunks[p->chunk_ib_idx].kpage[0] = kmalloc(PAGE_SIZE, GFP_KERNEL);
-		p->chunks[p->chunk_ib_idx].kpage[1] = kmalloc(PAGE_SIZE, GFP_KERNEL);
-		if (p->chunks[p->chunk_ib_idx].kpage[0] == NULL ||
-		    p->chunks[p->chunk_ib_idx].kpage[1] == NULL)
-			return -ENOMEM;
-		p->chunks[p->chunk_ib_idx].kpage_idx[0] = -1;
-		p->chunks[p->chunk_ib_idx].kpage_idx[1] = -1;
-		p->chunks[p->chunk_ib_idx].last_copied_page = -1;
-		p->chunks[p->chunk_ib_idx].last_page_index =
-			((p->chunks[p->chunk_ib_idx].length_dw * 4) - 1) / PAGE_SIZE;
-	}
-
 	return 0;
 }
 
@@ -312,146 +215,14 @@ static void radeon_cs_parser_fini(struct radeon_cs_parser *parser, int error)
 	radeon_ib_free(parser->rdev, &parser->ib);
 }
 
-static int radeon_cs_ib_chunk(struct radeon_device *rdev,
-			      struct radeon_cs_parser *parser)
-{
-	struct radeon_cs_chunk *ib_chunk;
-	int r;
-
-	if (parser->chunk_ib_idx == -1)
-		return 0;
-
-	if (parser->cs_flags & RADEON_CS_USE_VM)
-		return 0;
-
-	ib_chunk = &parser->chunks[parser->chunk_ib_idx];
-	/* Copy the packet into the IB, the parser will read from the
-	 * input memory (cached) and write to the IB (which can be
-	 * uncached).
-	 */
-	r =  radeon_ib_get(rdev, parser->ring, &parser->ib,
-			   ib_chunk->length_dw * 4);
-	if (r) {
-		DRM_ERROR("Failed to get ib !\n");
-		return r;
-	}
-	parser->ib->length_dw = ib_chunk->length_dw;
-	r = radeon_cs_parse(rdev, parser->ring, parser);
-	if (r || parser->parser_error) {
-		DRM_ERROR("Invalid command stream !\n");
-		return r;
-	}
-	r = radeon_cs_finish_pages(parser);
-	if (r) {
-		DRM_ERROR("Invalid command stream !\n");
-		return r;
-	}
-	r = radeon_cs_sync_rings(parser);
-	if (r) {
-		DRM_ERROR("Failed to synchronize rings !\n");
-	}
-	parser->ib->vm_id = 0;
-	r = radeon_ib_schedule(rdev, parser->ib);
-	if (r) {
-		DRM_ERROR("Failed to schedule IB !\n");
-	}
-	return 0;
-}
-
-static int radeon_bo_vm_update_pte(struct radeon_cs_parser *parser,
-				   struct radeon_vm *vm)
-{
-	struct radeon_bo_list *lobj;
-	struct radeon_bo *bo;
-	int r;
-
-	list_for_each_entry(lobj, &parser->validated, tv.head) {
-		bo = lobj->bo;
-		r = radeon_vm_bo_update_pte(parser->rdev, vm, bo, &bo->tbo.mem);
-		if (r) {
-			return r;
-		}
-	}
-	return 0;
-}
-
-static int radeon_cs_ib_vm_chunk(struct radeon_device *rdev,
-				 struct radeon_cs_parser *parser)
-{
-	struct radeon_cs_chunk *ib_chunk;
-	struct radeon_fpriv *fpriv = parser->filp->driver_priv;
-	struct radeon_vm *vm = &fpriv->vm;
-	int r;
-
-	if (parser->chunk_ib_idx == -1)
-		return 0;
-
-	if ((parser->cs_flags & RADEON_CS_USE_VM) == 0)
-		return 0;
-
-	ib_chunk = &parser->chunks[parser->chunk_ib_idx];
-	if (ib_chunk->length_dw > RADEON_IB_VM_MAX_SIZE) {
-		DRM_ERROR("cs IB too big: %d\n", ib_chunk->length_dw);
-		return -EINVAL;
-	}
-	r =  radeon_ib_get(rdev, parser->ring, &parser->ib,
-			   ib_chunk->length_dw * 4);
-	if (r) {
-		DRM_ERROR("Failed to get ib !\n");
-		return r;
-	}
-	parser->ib->length_dw = ib_chunk->length_dw;
-	/* Copy the packet into the IB */
-	if (DRM_COPY_FROM_USER(parser->ib->ptr, ib_chunk->user_ptr,
-			       ib_chunk->length_dw * 4)) {
-		return -EFAULT;
-	}
-	r = radeon_ring_ib_parse(rdev, parser->ring, parser->ib);
-	if (r) {
-		return r;
-	}
-
-	mutex_lock(&vm->mutex);
-	r = radeon_vm_bind(rdev, vm);
-	if (r) {
-		goto out;
-	}
-	r = radeon_bo_vm_update_pte(parser, vm);
-	if (r) {
-		goto out;
-	}
-	r = radeon_cs_sync_rings(parser);
-	if (r) {
-		DRM_ERROR("Failed to synchronize rings !\n");
-	}
-	parser->ib->vm_id = vm->id;
-	/* ib pool is bind at 0 in virtual address space to gpu_addr is the
-	 * offset inside the pool bo
-	 */
-	parser->ib->gpu_addr = parser->ib->sa_bo.offset;
-	r = radeon_ib_schedule(rdev, parser->ib);
-out:
-	if (!r) {
-		if (vm->fence) {
-			radeon_fence_unref(&vm->fence);
-		}
-		vm->fence = radeon_fence_ref(parser->ib->fence);
-	}
-	mutex_unlock(&fpriv->vm.mutex);
-	return r;
-}
-
 int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 {
 	struct radeon_device *rdev = dev->dev_private;
 	struct radeon_cs_parser parser;
+	struct radeon_cs_chunk *ib_chunk;
 	int r;
 
-	radeon_mutex_lock(&rdev->cs_mutex);
-	if (!rdev->accel_working) {
-		radeon_mutex_unlock(&rdev->cs_mutex);
-		return -EBUSY;
-	}
+	mutex_lock(&rdev->cs_mutex);
 	/* initialize parser */
 	memset(&parser, 0, sizeof(struct radeon_cs_parser));
 	parser.filp = filp;
@@ -462,7 +233,14 @@ int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 	if (r) {
 		DRM_ERROR("Failed to initialize parser !\n");
 		radeon_cs_parser_fini(&parser, r);
-		radeon_mutex_unlock(&rdev->cs_mutex);
+		mutex_unlock(&rdev->cs_mutex);
+		return r;
+	}
+	r =  radeon_ib_get(rdev, &parser.ib);
+	if (r) {
+		DRM_ERROR("Failed to get ib !\n");
+		radeon_cs_parser_fini(&parser, r);
+		mutex_unlock(&rdev->cs_mutex);
 		return r;
 	}
 	r = radeon_cs_parser_relocs(&parser);
@@ -470,20 +248,34 @@ int radeon_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		if (r != -ERESTARTSYS)
 			DRM_ERROR("Failed to parse relocation %d!\n", r);
 		radeon_cs_parser_fini(&parser, r);
-		radeon_mutex_unlock(&rdev->cs_mutex);
+		mutex_unlock(&rdev->cs_mutex);
 		return r;
 	}
-	r = radeon_cs_ib_chunk(rdev, &parser);
+	/* Copy the packet into the IB, the parser will read from the
+	 * input memory (cached) and write to the IB (which can be
+	 * uncached). */
+	ib_chunk = &parser.chunks[parser.chunk_ib_idx];
+	parser.ib->length_dw = ib_chunk->length_dw;
+	r = radeon_cs_parse(&parser);
+	if (r || parser.parser_error) {
+		DRM_ERROR("Invalid command stream !\n");
+		radeon_cs_parser_fini(&parser, r);
+		mutex_unlock(&rdev->cs_mutex);
+		return r;
+	}
+	r = radeon_cs_finish_pages(&parser);
 	if (r) {
-		goto out;
+		DRM_ERROR("Invalid command stream !\n");
+		radeon_cs_parser_fini(&parser, r);
+		mutex_unlock(&rdev->cs_mutex);
+		return r;
 	}
-	r = radeon_cs_ib_vm_chunk(rdev, &parser);
+	r = radeon_ib_schedule(rdev, parser.ib);
 	if (r) {
-		goto out;
+		DRM_ERROR("Failed to schedule IB !\n");
 	}
-out:
 	radeon_cs_parser_fini(&parser, r);
-	radeon_mutex_unlock(&rdev->cs_mutex);
+	mutex_unlock(&rdev->cs_mutex);
 	return r;
 }
 
diff --git a/drivers/gpu/drm/radeon/radeon_cursor.c b/drivers/gpu/drm/radeon/radeon_cursor.c
index 711e95a..72f749d 100644
--- a/drivers/gpu/drm/radeon/radeon_cursor.c
+++ b/drivers/gpu/drm/radeon/radeon_cursor.c
@@ -197,12 +197,7 @@ int radeon_crtc_cursor_set(struct drm_crtc *crtc,
 
 unpin:
 	if (radeon_crtc->cursor_bo) {
-		robj = gem_to_radeon_bo(radeon_crtc->cursor_bo);
-		ret = radeon_bo_reserve(robj, false);
-		if (likely(ret == 0)) {
-			radeon_bo_unpin(robj);
-			radeon_bo_unreserve(robj);
-		}
+		radeon_gem_object_unpin(radeon_crtc->cursor_bo);
 		drm_gem_object_unreference_unlocked(radeon_crtc->cursor_bo);
 	}
 
@@ -229,14 +224,14 @@ int radeon_crtc_cursor_move(struct drm_crtc *crtc,
 	}
 	DRM_DEBUG("x %d y %d c->x %d c->y %d\n", x, y, crtc->x, crtc->y);
 
-	if (x < 0) {
-		xorigin = min(-x, CURSOR_WIDTH - 1);
-		x = 0;
-	}
-	if (y < 0) {
-		yorigin = min(-y, CURSOR_HEIGHT - 1);
-		y = 0;
-	}
+	if (x < 0)
+		xorigin = -x + 1;
+	if (y < 0)
+		yorigin = -y + 1;
+	if (xorigin >= CURSOR_WIDTH)
+		xorigin = CURSOR_WIDTH - 1;
+	if (yorigin >= CURSOR_HEIGHT)
+		yorigin = CURSOR_HEIGHT - 1;
 
 	if (ASIC_IS_AVIVO(rdev)) {
 		int i = 0;
@@ -275,12 +270,16 @@ int radeon_crtc_cursor_move(struct drm_crtc *crtc,
 
 	radeon_lock_cursor(crtc, true);
 	if (ASIC_IS_DCE4(rdev)) {
-		WREG32(EVERGREEN_CUR_POSITION + radeon_crtc->crtc_offset, (x << 16) | y);
+		WREG32(EVERGREEN_CUR_POSITION + radeon_crtc->crtc_offset,
+		       ((xorigin ? 0 : x) << 16) |
+		       (yorigin ? 0 : y));
 		WREG32(EVERGREEN_CUR_HOT_SPOT + radeon_crtc->crtc_offset, (xorigin << 16) | yorigin);
 		WREG32(EVERGREEN_CUR_SIZE + radeon_crtc->crtc_offset,
 		       ((w - 1) << 16) | (radeon_crtc->cursor_height - 1));
 	} else if (ASIC_IS_AVIVO(rdev)) {
-		WREG32(AVIVO_D1CUR_POSITION + radeon_crtc->crtc_offset, (x << 16) | y);
+		WREG32(AVIVO_D1CUR_POSITION + radeon_crtc->crtc_offset,
+			     ((xorigin ? 0 : x) << 16) |
+			     (yorigin ? 0 : y));
 		WREG32(AVIVO_D1CUR_HOT_SPOT + radeon_crtc->crtc_offset, (xorigin << 16) | yorigin);
 		WREG32(AVIVO_D1CUR_SIZE + radeon_crtc->crtc_offset,
 		       ((w - 1) << 16) | (radeon_crtc->cursor_height - 1));
@@ -294,8 +293,8 @@ int radeon_crtc_cursor_move(struct drm_crtc *crtc,
 			| yorigin));
 		WREG32(RADEON_CUR_HORZ_VERT_POSN + radeon_crtc->crtc_offset,
 		       (RADEON_CUR_LOCK
-			| (x << 16)
-			| y));
+			| ((xorigin ? 0 : x) << 16)
+			| (yorigin ? 0 : y)));
 		/* offset is from DISP(2)_BASE_ADDRESS */
 		WREG32(RADEON_CUR_OFFSET + radeon_crtc->crtc_offset, (radeon_crtc->legacy_cursor_offset +
 								      (yorigin * 256)));
diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
index 405c827..a275cf6 100644
--- a/drivers/gpu/drm/radeon/radeon_device.c
+++ b/drivers/gpu/drm/radeon/radeon_device.c
@@ -32,7 +32,6 @@
 #include <drm/radeon_drm.h>
 #include <linux/vgaarb.h>
 #include <linux/vga_switcheroo.h>
-#include <linux/efi.h>
 #include "radeon_reg.h"
 #include "radeon.h"
 #include "atom.h"
@@ -304,8 +303,6 @@ void radeon_vram_location(struct radeon_device *rdev, struct radeon_mc *mc, u64
 		mc->mc_vram_size = mc->aper_size;
 	}
 	mc->vram_end = mc->vram_start + mc->mc_vram_size - 1;
-	if (radeon_vram_limit && radeon_vram_limit < mc->real_vram_size)
-		mc->real_vram_size = radeon_vram_limit;
 	dev_info(rdev->dev, "VRAM: %lluM 0x%016llX - 0x%016llX (%lluM used)\n",
 			mc->mc_vram_size >> 20, mc->vram_start,
 			mc->vram_end, mc->real_vram_size >> 20);
@@ -354,9 +351,6 @@ bool radeon_card_posted(struct radeon_device *rdev)
 {
 	uint32_t reg;
 
-	if (efi_enabled && rdev->pdev->subsystem_vendor == PCI_VENDOR_ID_APPLE)
-		return false;
-
 	/* first check CRTCs */
 	if (ASIC_IS_DCE4(rdev)) {
 		reg = RREG32(EVERGREEN_CRTC_CONTROL + EVERGREEN_CRTC0_REGISTER_OFFSET) |
@@ -718,26 +712,19 @@ int radeon_device_init(struct radeon_device *rdev,
 
 	/* mutex initialization are all done here so we
 	 * can recall function without having locking issues */
-	radeon_mutex_init(&rdev->cs_mutex);
-	radeon_mutex_init(&rdev->ib_pool.mutex);
-	for (i = 0; i < RADEON_NUM_RINGS; ++i)
-		mutex_init(&rdev->ring[i].mutex);
+	mutex_init(&rdev->cs_mutex);
+	mutex_init(&rdev->ib_pool.mutex);
+	mutex_init(&rdev->cp.mutex);
 	mutex_init(&rdev->dc_hw_i2c_mutex);
 	if (rdev->family >= CHIP_R600)
 		spin_lock_init(&rdev->ih.lock);
 	mutex_init(&rdev->gem.mutex);
 	mutex_init(&rdev->pm.mutex);
 	mutex_init(&rdev->vram_mutex);
-	rwlock_init(&rdev->fence_lock);
-	rwlock_init(&rdev->semaphore_drv.lock);
+	rwlock_init(&rdev->fence_drv.lock);
 	INIT_LIST_HEAD(&rdev->gem.objects);
 	init_waitqueue_head(&rdev->irq.vblank_queue);
 	init_waitqueue_head(&rdev->irq.idle_queue);
-	INIT_LIST_HEAD(&rdev->semaphore_drv.bo);
-	/* initialize vm here */
-	rdev->vm_manager.use_bitmap = 1;
-	rdev->vm_manager.max_pfn = 1 << 20;
-	INIT_LIST_HEAD(&rdev->vm_manager.lru_vm);
 
 	/* Set asic functions */
 	r = radeon_asic_init(rdev);
@@ -759,29 +746,22 @@ int radeon_device_init(struct radeon_device *rdev,
 
 	/* set DMA mask + need_dma32 flags.
 	 * PCIE - can handle 40-bits.
-	 * IGP - can handle 40-bits
+	 * IGP - can handle 40-bits (in theory)
 	 * AGP - generally dma32 is safest
-	 * PCI - dma32 for legacy pci gart, 40 bits on newer asics
+	 * PCI - only dma32
 	 */
 	rdev->need_dma32 = false;
 	if (rdev->flags & RADEON_IS_AGP)
 		rdev->need_dma32 = true;
-	if ((rdev->flags & RADEON_IS_PCI) &&
-	    (rdev->family < CHIP_RS400))
+	if (rdev->flags & RADEON_IS_PCI)
 		rdev->need_dma32 = true;
 
 	dma_bits = rdev->need_dma32 ? 32 : 40;
 	r = pci_set_dma_mask(rdev->pdev, DMA_BIT_MASK(dma_bits));
 	if (r) {
 		rdev->need_dma32 = true;
-		dma_bits = 32;
 		printk(KERN_WARNING "radeon: No suitable DMA available.\n");
 	}
-	r = pci_set_consistent_dma_mask(rdev->pdev, DMA_BIT_MASK(dma_bits));
-	if (r) {
-		pci_set_consistent_dma_mask(rdev->pdev, DMA_BIT_MASK(32));
-		printk(KERN_WARNING "radeon: No coherent DMA available.\n");
-	}
 
 	/* Registers mapping */
 	/* TODO: block userspace mapping of io register */
@@ -829,20 +809,15 @@ int radeon_device_init(struct radeon_device *rdev,
 		if (r)
 			return r;
 	}
-	if ((radeon_testing & 1)) {
+	if (radeon_testing) {
 		radeon_test_moves(rdev);
 	}
-	if ((radeon_testing & 2)) {
-		radeon_test_syncing(rdev);
-	}
 	if (radeon_benchmarking) {
-		radeon_benchmark(rdev, radeon_benchmarking);
+		radeon_benchmark(rdev);
 	}
 	return 0;
 }
 
-static void radeon_debugfs_remove_files(struct radeon_device *rdev);
-
 void radeon_device_fini(struct radeon_device *rdev)
 {
 	DRM_INFO("radeon: finishing device.\n");
@@ -857,7 +832,6 @@ void radeon_device_fini(struct radeon_device *rdev)
 	rdev->rio_mem = NULL;
 	iounmap(rdev->rmmio);
 	rdev->rmmio = NULL;
-	radeon_debugfs_remove_files(rdev);
 }
 
 
@@ -869,7 +843,7 @@ int radeon_suspend_kms(struct drm_device *dev, pm_message_t state)
 	struct radeon_device *rdev;
 	struct drm_crtc *crtc;
 	struct drm_connector *connector;
-	int i, r;
+	int r;
 
 	if (dev == NULL || dev->dev_private == NULL) {
 		return -ENODEV;
@@ -910,8 +884,7 @@ int radeon_suspend_kms(struct drm_device *dev, pm_message_t state)
 	/* evict vram memory */
 	radeon_bo_evict_vram(rdev);
 	/* wait for gpu to finish processing current batch */
-	for (i = 0; i < RADEON_NUM_RINGS; i++)
-		radeon_fence_wait_last(rdev, i);
+	radeon_fence_wait_last(rdev);
 
 	radeon_save_bios_scratch_regs(rdev);
 
@@ -960,11 +933,9 @@ int radeon_resume_kms(struct drm_device *dev)
 	radeon_fbdev_set_suspend(rdev, 0);
 	console_unlock();
 
-	/* init dig PHYs, disp eng pll */
-	if (rdev->is_atom_bios) {
+	/* init dig PHYs */
+	if (rdev->is_atom_bios)
 		radeon_atom_encoder_init(rdev);
-		radeon_atom_dcpll_init(rdev);
-	}
 	/* reset hpd state */
 	radeon_hpd_init(rdev);
 	/* blat the mode back in */
@@ -983,9 +954,6 @@ int radeon_gpu_reset(struct radeon_device *rdev)
 	int r;
 	int resched;
 
-	/* Prevent CS ioctl from interfering */
-	radeon_mutex_lock(&rdev->cs_mutex);
-
 	radeon_save_bios_scratch_regs(rdev);
 	/* block TTM */
 	resched = ttm_bo_lock_delayed_workqueue(&rdev->mman.bdev);
@@ -998,15 +966,10 @@ int radeon_gpu_reset(struct radeon_device *rdev)
 		radeon_restore_bios_scratch_regs(rdev);
 		drm_helper_resume_force_mode(rdev->ddev);
 		ttm_bo_unlock_delayed_workqueue(&rdev->mman.bdev, resched);
+		return 0;
 	}
-
-	radeon_mutex_unlock(&rdev->cs_mutex);
-
-	if (r) {
-		/* bad news, how to tell it to userspace ? */
-		dev_info(rdev->dev, "GPU reset failed\n");
-	}
-
+	/* bad news, how to tell it to userspace ? */
+	dev_info(rdev->dev, "GPU reset failed\n");
 	return r;
 }
 
@@ -1014,29 +977,33 @@ int radeon_gpu_reset(struct radeon_device *rdev)
 /*
  * Debugfs
  */
+struct radeon_debugfs {
+	struct drm_info_list	*files;
+	unsigned		num_files;
+};
+static struct radeon_debugfs _radeon_debugfs[RADEON_DEBUGFS_MAX_NUM_FILES];
+static unsigned _radeon_debugfs_count = 0;
+
 int radeon_debugfs_add_files(struct radeon_device *rdev,
 			     struct drm_info_list *files,
 			     unsigned nfiles)
 {
 	unsigned i;
 
-	for (i = 0; i < rdev->debugfs_count; i++) {
-		if (rdev->debugfs[i].files == files) {
+	for (i = 0; i < _radeon_debugfs_count; i++) {
+		if (_radeon_debugfs[i].files == files) {
 			/* Already registered */
 			return 0;
 		}
 	}
-
-	i = rdev->debugfs_count + 1;
-	if (i > RADEON_DEBUGFS_MAX_COMPONENTS) {
-		DRM_ERROR("Reached maximum number of debugfs components.\n");
-		DRM_ERROR("Report so we increase "
-		          "RADEON_DEBUGFS_MAX_COMPONENTS.\n");
+	if ((_radeon_debugfs_count + nfiles) > RADEON_DEBUGFS_MAX_NUM_FILES) {
+		DRM_ERROR("Reached maximum number of debugfs files.\n");
+		DRM_ERROR("Report so we increase RADEON_DEBUGFS_MAX_NUM_FILES.\n");
 		return -EINVAL;
 	}
-	rdev->debugfs[rdev->debugfs_count].files = files;
-	rdev->debugfs[rdev->debugfs_count].num_files = nfiles;
-	rdev->debugfs_count = i;
+	_radeon_debugfs[_radeon_debugfs_count].files = files;
+	_radeon_debugfs[_radeon_debugfs_count].num_files = nfiles;
+	_radeon_debugfs_count++;
 #if defined(CONFIG_DEBUG_FS)
 	drm_debugfs_create_files(files, nfiles,
 				 rdev->ddev->control->debugfs_root,
@@ -1048,22 +1015,6 @@ int radeon_debugfs_add_files(struct radeon_device *rdev,
 	return 0;
 }
 
-static void radeon_debugfs_remove_files(struct radeon_device *rdev)
-{
-#if defined(CONFIG_DEBUG_FS)
-	unsigned i;
-
-	for (i = 0; i < rdev->debugfs_count; i++) {
-		drm_debugfs_remove_files(rdev->debugfs[i].files,
-					 rdev->debugfs[i].num_files,
-					 rdev->ddev->control);
-		drm_debugfs_remove_files(rdev->debugfs[i].files,
-					 rdev->debugfs[i].num_files,
-					 rdev->ddev->primary);
-	}
-#endif
-}
-
 #if defined(CONFIG_DEBUG_FS)
 int radeon_debugfs_init(struct drm_minor *minor)
 {
@@ -1072,5 +1023,11 @@ int radeon_debugfs_init(struct drm_minor *minor)
 
 void radeon_debugfs_cleanup(struct drm_minor *minor)
 {
+	unsigned i;
+
+	for (i = 0; i < _radeon_debugfs_count; i++) {
+		drm_debugfs_remove_files(_radeon_debugfs[i].files,
+					 _radeon_debugfs[i].num_files, minor);
+	}
 }
 #endif
diff --git a/drivers/gpu/drm/radeon/radeon_display.c b/drivers/gpu/drm/radeon/radeon_display.c
index a3e07c2..0896fae 100644
--- a/drivers/gpu/drm/radeon/radeon_display.c
+++ b/drivers/gpu/drm/radeon/radeon_display.c
@@ -33,6 +33,8 @@
 #include "drm_crtc_helper.h"
 #include "drm_edid.h"
 
+static int radeon_ddc_dump(struct drm_connector *connector);
+
 static void avivo_crtc_load_lut(struct drm_crtc *crtc)
 {
 	struct radeon_crtc *radeon_crtc = to_radeon_crtc(crtc);
@@ -280,7 +282,7 @@ void radeon_crtc_handle_flip(struct radeon_device *rdev, int crtc_id)
 	spin_lock_irqsave(&rdev->ddev->event_lock, flags);
 	work = radeon_crtc->unpin_work;
 	if (work == NULL ||
-	    (work->fence && !radeon_fence_signaled(work->fence))) {
+	    !radeon_fence_signaled(work->fence)) {
 		spin_unlock_irqrestore(&rdev->ddev->event_lock, flags);
 		return;
 	}
@@ -303,17 +305,8 @@ void radeon_crtc_handle_flip(struct radeon_device *rdev, int crtc_id)
 	if (update_pending &&
 	    (DRM_SCANOUTPOS_VALID & radeon_get_crtc_scanoutpos(rdev->ddev, crtc_id,
 							       &vpos, &hpos)) &&
-	    ((vpos >= (99 * rdev->mode_info.crtcs[crtc_id]->base.hwmode.crtc_vdisplay)/100) ||
-	     (vpos < 0 && !ASIC_IS_AVIVO(rdev)))) {
-		/* crtc didn't flip in this target vblank interval,
-		 * but flip is pending in crtc. Based on the current
-		 * scanout position we know that the current frame is
-		 * (nearly) complete and the flip will (likely)
-		 * complete before the start of the next frame.
-		 */
-		update_pending = 0;
-	}
-	if (update_pending) {
+	    (vpos >=0) &&
+	    (vpos < (99 * rdev->mode_info.crtcs[crtc_id]->base.hwmode.crtc_vdisplay)/100)) {
 		/* crtc didn't flip in this target vblank interval,
 		 * but flip is pending in crtc. It will complete it
 		 * in next vblank interval, so complete the flip at
@@ -355,6 +348,7 @@ static int radeon_crtc_page_flip(struct drm_crtc *crtc,
 	struct radeon_framebuffer *new_radeon_fb;
 	struct drm_gem_object *obj;
 	struct radeon_bo *rbo;
+	struct radeon_fence *fence;
 	struct radeon_unpin_work *work;
 	unsigned long flags;
 	u32 tiling_flags, pitch_pixels;
@@ -365,9 +359,16 @@ static int radeon_crtc_page_flip(struct drm_crtc *crtc,
 	if (work == NULL)
 		return -ENOMEM;
 
+	r = radeon_fence_create(rdev, &fence);
+	if (unlikely(r != 0)) {
+		kfree(work);
+		DRM_ERROR("flip queue: failed to create fence.\n");
+		return -ENOMEM;
+	}
 	work->event = event;
 	work->rdev = rdev;
 	work->crtc_id = radeon_crtc->crtc_id;
+	work->fence = radeon_fence_ref(fence);
 	old_radeon_fb = to_radeon_framebuffer(crtc->fb);
 	new_radeon_fb = to_radeon_framebuffer(fb);
 	/* schedule unpin of the old buffer */
@@ -376,10 +377,6 @@ static int radeon_crtc_page_flip(struct drm_crtc *crtc,
 	drm_gem_object_reference(obj);
 	rbo = gem_to_radeon_bo(obj);
 	work->old_rbo = rbo;
-	obj = new_radeon_fb->obj;
-	rbo = gem_to_radeon_bo(obj);
-	if (rbo->tbo.sync_obj)
-		work->fence = radeon_fence_ref(rbo->tbo.sync_obj);
 	INIT_WORK(&work->work, radeon_unpin_work_func);
 
 	/* We borrow the event spin lock for protecting unpin_work */
@@ -394,6 +391,9 @@ static int radeon_crtc_page_flip(struct drm_crtc *crtc,
 	spin_unlock_irqrestore(&dev->event_lock, flags);
 
 	/* pin the new buffer */
+	obj = new_radeon_fb->obj;
+	rbo = gem_to_radeon_bo(obj);
+
 	DRM_DEBUG_DRIVER("flip-ioctl() cur_fbo = %p, cur_bbo = %p\n",
 			 work->old_rbo, rbo);
 
@@ -402,9 +402,7 @@ static int radeon_crtc_page_flip(struct drm_crtc *crtc,
 		DRM_ERROR("failed to reserve new rbo buffer before flip\n");
 		goto pflip_cleanup;
 	}
-	/* Only 27 bit offset for legacy CRTC */
-	r = radeon_bo_pin_restricted(rbo, RADEON_GEM_DOMAIN_VRAM,
-				     ASIC_IS_AVIVO(rdev) ? 0 : 1 << 27, &base);
+	r = radeon_bo_pin(rbo, RADEON_GEM_DOMAIN_VRAM, &base);
 	if (unlikely(r != 0)) {
 		radeon_bo_unreserve(rbo);
 		r = -EINVAL;
@@ -417,7 +415,7 @@ static int radeon_crtc_page_flip(struct drm_crtc *crtc,
 	if (!ASIC_IS_AVIVO(rdev)) {
 		/* crtc offset is from display base addr not FB location */
 		base -= radeon_crtc->legacy_display_base_addr;
-		pitch_pixels = fb->pitches[0] / (fb->bits_per_pixel / 8);
+		pitch_pixels = fb->pitch / (fb->bits_per_pixel / 8);
 
 		if (tiling_flags & RADEON_TILING_MACRO) {
 			if (ASIC_IS_R300(rdev)) {
@@ -463,18 +461,37 @@ static int radeon_crtc_page_flip(struct drm_crtc *crtc,
 		goto pflip_cleanup1;
 	}
 
+	/* 32 ought to cover us */
+	r = radeon_ring_lock(rdev, 32);
+	if (r) {
+		DRM_ERROR("failed to lock the ring before flip\n");
+		goto pflip_cleanup2;
+	}
+
+	/* emit the fence */
+	radeon_fence_emit(rdev, fence);
 	/* set the proper interrupt */
 	radeon_pre_page_flip(rdev, radeon_crtc->crtc_id);
+	/* fire the ring */
+	radeon_ring_unlock_commit(rdev);
 
 	return 0;
 
+pflip_cleanup2:
+	drm_vblank_put(dev, radeon_crtc->crtc_id);
+
 pflip_cleanup1:
-	if (unlikely(radeon_bo_reserve(rbo, false) != 0)) {
+	r = radeon_bo_reserve(rbo, false);
+	if (unlikely(r != 0)) {
 		DRM_ERROR("failed to reserve new rbo in error path\n");
 		goto pflip_cleanup;
 	}
-	if (unlikely(radeon_bo_unpin(rbo) != 0)) {
+	r = radeon_bo_unpin(rbo);
+	if (unlikely(r != 0)) {
+		radeon_bo_unreserve(rbo);
+		r = -EINVAL;
 		DRM_ERROR("failed to unpin new rbo in error path\n");
+		goto pflip_cleanup;
 	}
 	radeon_bo_unreserve(rbo);
 
@@ -482,9 +499,9 @@ pflip_cleanup:
 	spin_lock_irqsave(&dev->event_lock, flags);
 	radeon_crtc->unpin_work = NULL;
 unlock_free:
-	spin_unlock_irqrestore(&dev->event_lock, flags);
 	drm_gem_object_unreference_unlocked(old_radeon_fb->obj);
-	radeon_fence_unref(&work->fence);
+	spin_unlock_irqrestore(&dev->event_lock, flags);
+	radeon_fence_unref(&fence);
 	kfree(work);
 
 	return r;
@@ -678,6 +695,7 @@ static void radeon_print_display_setup(struct drm_device *dev)
 static bool radeon_setup_enc_conn(struct drm_device *dev)
 {
 	struct radeon_device *rdev = dev->dev_private;
+	struct drm_connector *drm_connector;
 	bool ret = false;
 
 	if (rdev->bios) {
@@ -697,6 +715,8 @@ static bool radeon_setup_enc_conn(struct drm_device *dev)
 	if (ret) {
 		radeon_setup_encoder_clones(dev);
 		radeon_print_display_setup(dev);
+		list_for_each_entry(drm_connector, &dev->mode_config.connector_list, head)
+			radeon_ddc_dump(drm_connector);
 	}
 
 	return ret;
@@ -713,22 +733,16 @@ int radeon_ddc_get_modes(struct radeon_connector *radeon_connector)
 		radeon_router_select_ddc_port(radeon_connector);
 
 	if ((radeon_connector->base.connector_type == DRM_MODE_CONNECTOR_DisplayPort) ||
-	    (radeon_connector->base.connector_type == DRM_MODE_CONNECTOR_eDP) ||
-	    (radeon_connector_encoder_get_dp_bridge_encoder_id(&radeon_connector->base) !=
-	     ENCODER_OBJECT_ID_NONE)) {
+	    (radeon_connector->base.connector_type == DRM_MODE_CONNECTOR_eDP)) {
 		struct radeon_connector_atom_dig *dig = radeon_connector->con_priv;
-
 		if ((dig->dp_sink_type == CONNECTOR_OBJECT_ID_DISPLAYPORT ||
 		     dig->dp_sink_type == CONNECTOR_OBJECT_ID_eDP) && dig->dp_i2c_bus)
-			radeon_connector->edid = drm_get_edid(&radeon_connector->base,
-							      &dig->dp_i2c_bus->adapter);
-		else if (radeon_connector->ddc_bus && !radeon_connector->edid)
-			radeon_connector->edid = drm_get_edid(&radeon_connector->base,
-							      &radeon_connector->ddc_bus->adapter);
-	} else {
-		if (radeon_connector->ddc_bus && !radeon_connector->edid)
-			radeon_connector->edid = drm_get_edid(&radeon_connector->base,
-							      &radeon_connector->ddc_bus->adapter);
+			radeon_connector->edid = drm_get_edid(&radeon_connector->base, &dig->dp_i2c_bus->adapter);
+	}
+	if (!radeon_connector->ddc_bus)
+		return -1;
+	if (!radeon_connector->edid) {
+		radeon_connector->edid = drm_get_edid(&radeon_connector->base, &radeon_connector->ddc_bus->adapter);
 	}
 
 	if (!radeon_connector->edid) {
@@ -750,6 +764,34 @@ int radeon_ddc_get_modes(struct radeon_connector *radeon_connector)
 	return 0;
 }
 
+static int radeon_ddc_dump(struct drm_connector *connector)
+{
+	struct edid *edid;
+	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
+	int ret = 0;
+
+	/* on hw with routers, select right port */
+	if (radeon_connector->router.ddc_valid)
+		radeon_router_select_ddc_port(radeon_connector);
+
+	if (!radeon_connector->ddc_bus)
+		return -1;
+	edid = drm_get_edid(connector, &radeon_connector->ddc_bus->adapter);
+	/* Log EDID retrieval status here. In particular with regard to
+	 * connectors with requires_extended_probe flag set, that will prevent
+	 * function radeon_dvi_detect() to fetch EDID on this connector,
+	 * as long as there is no valid EDID header found */
+	if (edid) {
+		DRM_INFO("Radeon display connector %s: Found valid EDID",
+				drm_get_connector_name(connector));
+		kfree(edid);
+	} else {
+		DRM_INFO("Radeon display connector %s: No monitor connected or invalid EDID",
+				drm_get_connector_name(connector));
+	}
+	return ret;
+}
+
 /* avivo */
 static void avivo_get_fb_div(struct radeon_pll *pll,
 			     u32 target_clock,
@@ -1089,36 +1131,29 @@ static const struct drm_framebuffer_funcs radeon_fb_funcs = {
 	.create_handle = radeon_user_framebuffer_create_handle,
 };
 
-int
+void
 radeon_framebuffer_init(struct drm_device *dev,
 			struct radeon_framebuffer *rfb,
-			struct drm_mode_fb_cmd2 *mode_cmd,
+			struct drm_mode_fb_cmd *mode_cmd,
 			struct drm_gem_object *obj)
 {
-	int ret;
 	rfb->obj = obj;
-	ret = drm_framebuffer_init(dev, &rfb->base, &radeon_fb_funcs);
-	if (ret) {
-		rfb->obj = NULL;
-		return ret;
-	}
+	drm_framebuffer_init(dev, &rfb->base, &radeon_fb_funcs);
 	drm_helper_mode_fill_fb_struct(&rfb->base, mode_cmd);
-	return 0;
 }
 
 static struct drm_framebuffer *
 radeon_user_framebuffer_create(struct drm_device *dev,
 			       struct drm_file *file_priv,
-			       struct drm_mode_fb_cmd2 *mode_cmd)
+			       struct drm_mode_fb_cmd *mode_cmd)
 {
 	struct drm_gem_object *obj;
 	struct radeon_framebuffer *radeon_fb;
-	int ret;
 
-	obj = drm_gem_object_lookup(dev, file_priv, mode_cmd->handles[0]);
+	obj = drm_gem_object_lookup(dev, file_priv, mode_cmd->handle);
 	if (obj ==  NULL) {
 		dev_err(&dev->pdev->dev, "No GEM object associated to handle 0x%08X, "
-			"can't create framebuffer\n", mode_cmd->handles[0]);
+			"can't create framebuffer\n", mode_cmd->handle);
 		return ERR_PTR(-ENOENT);
 	}
 
@@ -1128,12 +1163,7 @@ radeon_user_framebuffer_create(struct drm_device *dev,
 		return ERR_PTR(-ENOMEM);
 	}
 
-	ret = radeon_framebuffer_init(dev, radeon_fb, mode_cmd, obj);
-	if (ret) {
-		kfree(radeon_fb);
-		drm_gem_object_unreference_unlocked(obj);
-		return NULL;
-	}
+	radeon_framebuffer_init(dev, radeon_fb, mode_cmd, obj);
 
 	return &radeon_fb->base;
 }
@@ -1149,6 +1179,11 @@ static const struct drm_mode_config_funcs radeon_mode_funcs = {
 	.output_poll_changed = radeon_output_poll_changed
 };
 
+struct drm_prop_enum_list {
+	int type;
+	char *name;
+};
+
 static struct drm_prop_enum_list radeon_tmds_pll_enum_list[] =
 {	{ 0, "driver" },
 	{ 1, "bios" },
@@ -1173,53 +1208,86 @@ static struct drm_prop_enum_list radeon_underscan_enum_list[] =
 
 static int radeon_modeset_create_props(struct radeon_device *rdev)
 {
-	int sz;
+	int i, sz;
 
 	if (rdev->is_atom_bios) {
 		rdev->mode_info.coherent_mode_property =
-			drm_property_create_range(rdev->ddev, 0 , "coherent", 0, 1);
+			drm_property_create(rdev->ddev,
+					    DRM_MODE_PROP_RANGE,
+					    "coherent", 2);
 		if (!rdev->mode_info.coherent_mode_property)
 			return -ENOMEM;
+
+		rdev->mode_info.coherent_mode_property->values[0] = 0;
+		rdev->mode_info.coherent_mode_property->values[1] = 1;
 	}
 
 	if (!ASIC_IS_AVIVO(rdev)) {
 		sz = ARRAY_SIZE(radeon_tmds_pll_enum_list);
 		rdev->mode_info.tmds_pll_property =
-			drm_property_create_enum(rdev->ddev, 0,
-					    "tmds_pll",
-					    radeon_tmds_pll_enum_list, sz);
+			drm_property_create(rdev->ddev,
+					    DRM_MODE_PROP_ENUM,
+					    "tmds_pll", sz);
+		for (i = 0; i < sz; i++) {
+			drm_property_add_enum(rdev->mode_info.tmds_pll_property,
+					      i,
+					      radeon_tmds_pll_enum_list[i].type,
+					      radeon_tmds_pll_enum_list[i].name);
+		}
 	}
 
 	rdev->mode_info.load_detect_property =
-		drm_property_create_range(rdev->ddev, 0, "load detection", 0, 1);
+		drm_property_create(rdev->ddev,
+				    DRM_MODE_PROP_RANGE,
+				    "load detection", 2);
 	if (!rdev->mode_info.load_detect_property)
 		return -ENOMEM;
+	rdev->mode_info.load_detect_property->values[0] = 0;
+	rdev->mode_info.load_detect_property->values[1] = 1;
 
 	drm_mode_create_scaling_mode_property(rdev->ddev);
 
 	sz = ARRAY_SIZE(radeon_tv_std_enum_list);
 	rdev->mode_info.tv_std_property =
-		drm_property_create_enum(rdev->ddev, 0,
-				    "tv standard",
-				    radeon_tv_std_enum_list, sz);
+		drm_property_create(rdev->ddev,
+				    DRM_MODE_PROP_ENUM,
+				    "tv standard", sz);
+	for (i = 0; i < sz; i++) {
+		drm_property_add_enum(rdev->mode_info.tv_std_property,
+				      i,
+				      radeon_tv_std_enum_list[i].type,
+				      radeon_tv_std_enum_list[i].name);
+	}
 
 	sz = ARRAY_SIZE(radeon_underscan_enum_list);
 	rdev->mode_info.underscan_property =
-		drm_property_create_enum(rdev->ddev, 0,
-				    "underscan",
-				    radeon_underscan_enum_list, sz);
+		drm_property_create(rdev->ddev,
+				    DRM_MODE_PROP_ENUM,
+				    "underscan", sz);
+	for (i = 0; i < sz; i++) {
+		drm_property_add_enum(rdev->mode_info.underscan_property,
+				      i,
+				      radeon_underscan_enum_list[i].type,
+				      radeon_underscan_enum_list[i].name);
+	}
 
 	rdev->mode_info.underscan_hborder_property =
-		drm_property_create_range(rdev->ddev, 0,
-					"underscan hborder", 0, 128);
+		drm_property_create(rdev->ddev,
+					DRM_MODE_PROP_RANGE,
+					"underscan hborder", 2);
 	if (!rdev->mode_info.underscan_hborder_property)
 		return -ENOMEM;
+	rdev->mode_info.underscan_hborder_property->values[0] = 0;
+	rdev->mode_info.underscan_hborder_property->values[1] = 128;
 
 	rdev->mode_info.underscan_vborder_property =
-		drm_property_create_range(rdev->ddev, 0,
-					"underscan vborder", 0, 128);
+		drm_property_create(rdev->ddev,
+					DRM_MODE_PROP_RANGE,
+					"underscan vborder", 2);
 	if (!rdev->mode_info.underscan_vborder_property)
 		return -ENOMEM;
+	rdev->mode_info.underscan_vborder_property->values[0] = 0;
+	rdev->mode_info.underscan_vborder_property->values[1] = 128;
 
 	return 0;
 }
@@ -1265,9 +1333,6 @@ int radeon_modeset_init(struct radeon_device *rdev)
 		rdev->ddev->mode_config.max_height = 4096;
 	}
 
-	rdev->ddev->mode_config.preferred_depth = 24;
-	rdev->ddev->mode_config.prefer_shadow = 1;
-
 	rdev->ddev->mode_config.fb_base = rdev->mc.aper_base;
 
 	ret = radeon_modeset_create_props(rdev);
@@ -1295,11 +1360,9 @@ int radeon_modeset_init(struct radeon_device *rdev)
 		return ret;
 	}
 
-	/* init dig PHYs, disp eng pll */
-	if (rdev->is_atom_bios) {
+	/* init dig PHYs */
+	if (rdev->is_atom_bios)
 		radeon_atom_encoder_init(rdev);
-		radeon_atom_dcpll_init(rdev);
-	}
 
 	/* initialize hpd */
 	radeon_hpd_init(rdev);
diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c
index 498d21d..60e1605 100644
--- a/drivers/gpu/drm/radeon/radeon_drv.c
+++ b/drivers/gpu/drm/radeon/radeon_drv.c
@@ -36,7 +36,6 @@
 
 #include "drm_pciids.h"
 #include <linux/console.h>
-#include <linux/module.h>
 
 
 /*
@@ -52,13 +51,9 @@
  *   2.8.0 - pageflip support, r500 US_FORMAT regs. r500 ARGB2101010 colorbuf, r300->r500 CMASK, clock crystal query
  *   2.9.0 - r600 tiling (s3tc,rgtc) working, SET_PREDICATION packet 3 on r600 + eg, backend query
  *   2.10.0 - fusion 2D tiling
- *   2.11.0 - backend map, initial compute support for the CS checker
- *   2.12.0 - RADEON_CS_KEEP_TILING_FLAGS
- *   2.13.0 - virtual memory support, streamout
- *   2.14.0 - add evergreen tiling informations
  */
 #define KMS_DRIVER_MAJOR	2
-#define KMS_DRIVER_MINOR	14
+#define KMS_DRIVER_MINOR	10
 #define KMS_DRIVER_PATCHLEVEL	0
 int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags);
 int radeon_driver_unload_kms(struct drm_device *dev);
@@ -86,10 +81,6 @@ int radeon_dma_ioctl_kms(struct drm_device *dev, void *data,
 			 struct drm_file *file_priv);
 int radeon_gem_object_init(struct drm_gem_object *obj);
 void radeon_gem_object_free(struct drm_gem_object *obj);
-int radeon_gem_object_open(struct drm_gem_object *obj,
-				struct drm_file *file_priv);
-void radeon_gem_object_close(struct drm_gem_object *obj,
-				struct drm_file *file_priv);
 extern int radeon_get_crtc_scanoutpos(struct drm_device *dev, int crtc,
 				      int *vpos, int *hpos);
 extern struct drm_ioctl_desc radeon_ioctls_kms[];
@@ -146,7 +137,7 @@ module_param_named(vramlimit, radeon_vram_limit, int, 0600);
 MODULE_PARM_DESC(agpmode, "AGP Mode (-1 == PCI)");
 module_param_named(agpmode, radeon_agpmode, int, 0444);
 
-MODULE_PARM_DESC(gartsize, "Size of PCIE/IGP gart to setup in megabytes (32, 64, etc)");
+MODULE_PARM_DESC(gartsize, "Size of PCIE/IGP gart to setup in megabytes (32,64, etc)\n");
 module_param_named(gartsize, radeon_gart_size, int, 0600);
 
 MODULE_PARM_DESC(benchmark, "Run benchmark");
@@ -212,21 +203,6 @@ static struct pci_device_id pciidlist[] = {
 MODULE_DEVICE_TABLE(pci, pciidlist);
 #endif
 
-static const struct file_operations radeon_driver_old_fops = {
-	.owner = THIS_MODULE,
-	.open = drm_open,
-	.release = drm_release,
-	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_mmap,
-	.poll = drm_poll,
-	.fasync = drm_fasync,
-	.read = drm_read,
-#ifdef CONFIG_COMPAT
-	.compat_ioctl = radeon_compat_ioctl,
-#endif
-	.llseek = noop_llseek,
-};
-
 static struct drm_driver driver_old = {
 	.driver_features =
 	    DRIVER_USE_AGP | DRIVER_USE_MTRR | DRIVER_PCI_DMA | DRIVER_SG |
@@ -253,7 +229,21 @@ static struct drm_driver driver_old = {
 	.reclaim_buffers = drm_core_reclaim_buffers,
 	.ioctls = radeon_ioctls,
 	.dma_ioctl = radeon_cp_buffers,
-	.fops = &radeon_driver_old_fops,
+	.fops = {
+		 .owner = THIS_MODULE,
+		 .open = drm_open,
+		 .release = drm_release,
+		 .unlocked_ioctl = drm_ioctl,
+		 .mmap = drm_mmap,
+		 .poll = drm_poll,
+		 .fasync = drm_fasync,
+		 .read = drm_read,
+#ifdef CONFIG_COMPAT
+		 .compat_ioctl = radeon_compat_ioctl,
+#endif
+		 .llseek = noop_llseek,
+	},
+
 	.name = DRIVER_NAME,
 	.desc = DRIVER_DESC,
 	.date = DRIVER_DATE,
@@ -311,20 +301,6 @@ radeon_pci_resume(struct pci_dev *pdev)
 	return radeon_resume_kms(dev);
 }
 
-static const struct file_operations radeon_driver_kms_fops = {
-	.owner = THIS_MODULE,
-	.open = drm_open,
-	.release = drm_release,
-	.unlocked_ioctl = drm_ioctl,
-	.mmap = radeon_mmap,
-	.poll = drm_poll,
-	.fasync = drm_fasync,
-	.read = drm_read,
-#ifdef CONFIG_COMPAT
-	.compat_ioctl = radeon_kms_compat_ioctl,
-#endif
-};
-
 static struct drm_driver kms_driver = {
 	.driver_features =
 	    DRIVER_USE_AGP | DRIVER_USE_MTRR | DRIVER_PCI_DMA | DRIVER_SG |
@@ -356,13 +332,24 @@ static struct drm_driver kms_driver = {
 	.ioctls = radeon_ioctls_kms,
 	.gem_init_object = radeon_gem_object_init,
 	.gem_free_object = radeon_gem_object_free,
-	.gem_open_object = radeon_gem_object_open,
-	.gem_close_object = radeon_gem_object_close,
 	.dma_ioctl = radeon_dma_ioctl_kms,
 	.dumb_create = radeon_mode_dumb_create,
 	.dumb_map_offset = radeon_mode_dumb_mmap,
 	.dumb_destroy = radeon_mode_dumb_destroy,
-	.fops = &radeon_driver_kms_fops,
+	.fops = {
+		 .owner = THIS_MODULE,
+		 .open = drm_open,
+		 .release = drm_release,
+		 .unlocked_ioctl = drm_ioctl,
+		 .mmap = radeon_mmap,
+		 .poll = drm_poll,
+		 .fasync = drm_fasync,
+		 .read = drm_read,
+#ifdef CONFIG_COMPAT
+		 .compat_ioctl = radeon_kms_compat_ioctl,
+#endif
+	},
+
 	.name = DRIVER_NAME,
 	.desc = DRIVER_DESC,
 	.date = DRIVER_DATE,
diff --git a/drivers/gpu/drm/radeon/radeon_encoders.c b/drivers/gpu/drm/radeon/radeon_encoders.c
index 26e9270..8a171b2 100644
--- a/drivers/gpu/drm/radeon/radeon_encoders.c
+++ b/drivers/gpu/drm/radeon/radeon_encoders.c
@@ -29,6 +29,12 @@
 #include "radeon.h"
 #include "atom.h"
 
+extern int atom_debug;
+
+/* evil but including atombios.h is much worse */
+bool radeon_atom_get_tv_timings(struct radeon_device *rdev, int index,
+				struct drm_display_mode *mode);
+
 static uint32_t radeon_encoder_clones(struct drm_encoder *encoder)
 {
 	struct drm_device *dev = encoder->dev;
@@ -150,6 +156,27 @@ radeon_get_encoder_enum(struct drm_device *dev, uint32_t supported_device, uint8
 	return ret;
 }
 
+static inline bool radeon_encoder_is_digital(struct drm_encoder *encoder)
+{
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	switch (radeon_encoder->encoder_id) {
+	case ENCODER_OBJECT_ID_INTERNAL_LVDS:
+	case ENCODER_OBJECT_ID_INTERNAL_TMDS1:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
+	case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
+	case ENCODER_OBJECT_ID_INTERNAL_DVO1:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
+	case ENCODER_OBJECT_ID_INTERNAL_DDI:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+		return true;
+	default:
+		return false;
+	}
+}
+
 void
 radeon_link_encoder_connector(struct drm_device *dev)
 {
@@ -202,7 +229,7 @@ radeon_get_connector_for_encoder(struct drm_encoder *encoder)
 	return NULL;
 }
 
-struct drm_connector *
+static struct drm_connector *
 radeon_get_connector_for_encoder_init(struct drm_encoder *encoder)
 {
 	struct drm_device *dev = encoder->dev;
@@ -218,7 +245,7 @@ radeon_get_connector_for_encoder_init(struct drm_encoder *encoder)
 	return NULL;
 }
 
-struct drm_encoder *radeon_get_external_encoder(struct drm_encoder *encoder)
+struct drm_encoder *radeon_atom_get_external_encoder(struct drm_encoder *encoder)
 {
 	struct drm_device *dev = encoder->dev;
 	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
@@ -239,9 +266,9 @@ struct drm_encoder *radeon_get_external_encoder(struct drm_encoder *encoder)
 	return NULL;
 }
 
-u16 radeon_encoder_get_dp_bridge_encoder_id(struct drm_encoder *encoder)
+bool radeon_encoder_is_dp_bridge(struct drm_encoder *encoder)
 {
-	struct drm_encoder *other_encoder = radeon_get_external_encoder(encoder);
+	struct drm_encoder *other_encoder = radeon_atom_get_external_encoder(encoder);
 
 	if (other_encoder) {
 		struct radeon_encoder *radeon_encoder = to_radeon_encoder(other_encoder);
@@ -249,12 +276,13 @@ u16 radeon_encoder_get_dp_bridge_encoder_id(struct drm_encoder *encoder)
 		switch (radeon_encoder->encoder_id) {
 		case ENCODER_OBJECT_ID_TRAVIS:
 		case ENCODER_OBJECT_ID_NUTMEG:
-			return radeon_encoder->encoder_id;
+			return true;
 		default:
-			return ENCODER_OBJECT_ID_NONE;
+			return false;
 		}
 	}
-	return ENCODER_OBJECT_ID_NONE;
+
+	return false;
 }
 
 void radeon_panel_mode_fixup(struct drm_encoder *encoder,
@@ -304,13 +332,339 @@ void radeon_panel_mode_fixup(struct drm_encoder *encoder,
 
 }
 
-bool radeon_dig_monitor_is_duallink(struct drm_encoder *encoder,
-				    u32 pixel_clock)
+static bool radeon_atom_mode_fixup(struct drm_encoder *encoder,
+				   struct drm_display_mode *mode,
+				   struct drm_display_mode *adjusted_mode)
+{
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
+
+	/* set the active encoder to connector routing */
+	radeon_encoder_set_active_device(encoder);
+	drm_mode_set_crtcinfo(adjusted_mode, 0);
+
+	/* hw bug */
+	if ((mode->flags & DRM_MODE_FLAG_INTERLACE)
+	    && (mode->crtc_vsync_start < (mode->crtc_vdisplay + 2)))
+		adjusted_mode->crtc_vsync_start = adjusted_mode->crtc_vdisplay + 2;
+
+	/* get the native mode for LVDS */
+	if (radeon_encoder->active_device & (ATOM_DEVICE_LCD_SUPPORT))
+		radeon_panel_mode_fixup(encoder, adjusted_mode);
+
+	/* get the native mode for TV */
+	if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT)) {
+		struct radeon_encoder_atom_dac *tv_dac = radeon_encoder->enc_priv;
+		if (tv_dac) {
+			if (tv_dac->tv_std == TV_STD_NTSC ||
+			    tv_dac->tv_std == TV_STD_NTSC_J ||
+			    tv_dac->tv_std == TV_STD_PAL_M)
+				radeon_atom_get_tv_timings(rdev, 0, adjusted_mode);
+			else
+				radeon_atom_get_tv_timings(rdev, 1, adjusted_mode);
+		}
+	}
+
+	if (ASIC_IS_DCE3(rdev) &&
+	    ((radeon_encoder->active_device & (ATOM_DEVICE_DFP_SUPPORT | ATOM_DEVICE_LCD_SUPPORT)) ||
+	     radeon_encoder_is_dp_bridge(encoder))) {
+		struct drm_connector *connector = radeon_get_connector_for_encoder(encoder);
+		radeon_dp_set_link_config(connector, mode);
+	}
+
+	return true;
+}
+
+static void
+atombios_dac_setup(struct drm_encoder *encoder, int action)
+{
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	DAC_ENCODER_CONTROL_PS_ALLOCATION args;
+	int index = 0;
+	struct radeon_encoder_atom_dac *dac_info = radeon_encoder->enc_priv;
+
+	memset(&args, 0, sizeof(args));
+
+	switch (radeon_encoder->encoder_id) {
+	case ENCODER_OBJECT_ID_INTERNAL_DAC1:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1:
+		index = GetIndexIntoMasterTable(COMMAND, DAC1EncoderControl);
+		break;
+	case ENCODER_OBJECT_ID_INTERNAL_DAC2:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC2:
+		index = GetIndexIntoMasterTable(COMMAND, DAC2EncoderControl);
+		break;
+	}
+
+	args.ucAction = action;
+
+	if (radeon_encoder->active_device & (ATOM_DEVICE_CRT_SUPPORT))
+		args.ucDacStandard = ATOM_DAC1_PS2;
+	else if (radeon_encoder->active_device & (ATOM_DEVICE_CV_SUPPORT))
+		args.ucDacStandard = ATOM_DAC1_CV;
+	else {
+		switch (dac_info->tv_std) {
+		case TV_STD_PAL:
+		case TV_STD_PAL_M:
+		case TV_STD_SCART_PAL:
+		case TV_STD_SECAM:
+		case TV_STD_PAL_CN:
+			args.ucDacStandard = ATOM_DAC1_PAL;
+			break;
+		case TV_STD_NTSC:
+		case TV_STD_NTSC_J:
+		case TV_STD_PAL_60:
+		default:
+			args.ucDacStandard = ATOM_DAC1_NTSC;
+			break;
+		}
+	}
+	args.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
+
+	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+}
+
+static void
+atombios_tv_setup(struct drm_encoder *encoder, int action)
+{
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	TV_ENCODER_CONTROL_PS_ALLOCATION args;
+	int index = 0;
+	struct radeon_encoder_atom_dac *dac_info = radeon_encoder->enc_priv;
+
+	memset(&args, 0, sizeof(args));
+
+	index = GetIndexIntoMasterTable(COMMAND, TVEncoderControl);
+
+	args.sTVEncoder.ucAction = action;
+
+	if (radeon_encoder->active_device & (ATOM_DEVICE_CV_SUPPORT))
+		args.sTVEncoder.ucTvStandard = ATOM_TV_CV;
+	else {
+		switch (dac_info->tv_std) {
+		case TV_STD_NTSC:
+			args.sTVEncoder.ucTvStandard = ATOM_TV_NTSC;
+			break;
+		case TV_STD_PAL:
+			args.sTVEncoder.ucTvStandard = ATOM_TV_PAL;
+			break;
+		case TV_STD_PAL_M:
+			args.sTVEncoder.ucTvStandard = ATOM_TV_PALM;
+			break;
+		case TV_STD_PAL_60:
+			args.sTVEncoder.ucTvStandard = ATOM_TV_PAL60;
+			break;
+		case TV_STD_NTSC_J:
+			args.sTVEncoder.ucTvStandard = ATOM_TV_NTSCJ;
+			break;
+		case TV_STD_SCART_PAL:
+			args.sTVEncoder.ucTvStandard = ATOM_TV_PAL; /* ??? */
+			break;
+		case TV_STD_SECAM:
+			args.sTVEncoder.ucTvStandard = ATOM_TV_SECAM;
+			break;
+		case TV_STD_PAL_CN:
+			args.sTVEncoder.ucTvStandard = ATOM_TV_PALCN;
+			break;
+		default:
+			args.sTVEncoder.ucTvStandard = ATOM_TV_NTSC;
+			break;
+		}
+	}
+
+	args.sTVEncoder.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
+
+	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+}
+
+union dvo_encoder_control {
+	ENABLE_EXTERNAL_TMDS_ENCODER_PS_ALLOCATION ext_tmds;
+	DVO_ENCODER_CONTROL_PS_ALLOCATION dvo;
+	DVO_ENCODER_CONTROL_PS_ALLOCATION_V3 dvo_v3;
+};
+
+void
+atombios_dvo_setup(struct drm_encoder *encoder, int action)
+{
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	union dvo_encoder_control args;
+	int index = GetIndexIntoMasterTable(COMMAND, DVOEncoderControl);
+
+	memset(&args, 0, sizeof(args));
+
+	if (ASIC_IS_DCE3(rdev)) {
+		/* DCE3+ */
+		args.dvo_v3.ucAction = action;
+		args.dvo_v3.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
+		args.dvo_v3.ucDVOConfig = 0; /* XXX */
+	} else if (ASIC_IS_DCE2(rdev)) {
+		/* DCE2 (pre-DCE3 R6xx, RS600/690/740 */
+		args.dvo.sDVOEncoder.ucAction = action;
+		args.dvo.sDVOEncoder.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
+		/* DFP1, CRT1, TV1 depending on the type of port */
+		args.dvo.sDVOEncoder.ucDeviceType = ATOM_DEVICE_DFP1_INDEX;
+
+		if (radeon_encoder->pixel_clock > 165000)
+			args.dvo.sDVOEncoder.usDevAttr.sDigAttrib.ucAttribute |= PANEL_ENCODER_MISC_DUAL;
+	} else {
+		/* R4xx, R5xx */
+		args.ext_tmds.sXTmdsEncoder.ucEnable = action;
+
+		if (radeon_encoder->pixel_clock > 165000)
+			args.ext_tmds.sXTmdsEncoder.ucMisc |= PANEL_ENCODER_MISC_DUAL;
+
+		/*if (pScrn->rgbBits == 8)*/
+		args.ext_tmds.sXTmdsEncoder.ucMisc |= ATOM_PANEL_MISC_888RGB;
+	}
+
+	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+}
+
+union lvds_encoder_control {
+	LVDS_ENCODER_CONTROL_PS_ALLOCATION    v1;
+	LVDS_ENCODER_CONTROL_PS_ALLOCATION_V2 v2;
+};
+
+void
+atombios_digital_setup(struct drm_encoder *encoder, int action)
 {
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
+	union lvds_encoder_control args;
+	int index = 0;
+	int hdmi_detected = 0;
+	uint8_t frev, crev;
+
+	if (!dig)
+		return;
+
+	if (atombios_get_encoder_mode(encoder) == ATOM_ENCODER_MODE_HDMI)
+		hdmi_detected = 1;
+
+	memset(&args, 0, sizeof(args));
+
+	switch (radeon_encoder->encoder_id) {
+	case ENCODER_OBJECT_ID_INTERNAL_LVDS:
+		index = GetIndexIntoMasterTable(COMMAND, LVDSEncoderControl);
+		break;
+	case ENCODER_OBJECT_ID_INTERNAL_TMDS1:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
+		index = GetIndexIntoMasterTable(COMMAND, TMDS1EncoderControl);
+		break;
+	case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
+		if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT))
+			index = GetIndexIntoMasterTable(COMMAND, LVDSEncoderControl);
+		else
+			index = GetIndexIntoMasterTable(COMMAND, TMDS2EncoderControl);
+		break;
+	}
+
+	if (!atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev))
+		return;
+
+	switch (frev) {
+	case 1:
+	case 2:
+		switch (crev) {
+		case 1:
+			args.v1.ucMisc = 0;
+			args.v1.ucAction = action;
+			if (hdmi_detected)
+				args.v1.ucMisc |= PANEL_ENCODER_MISC_HDMI_TYPE;
+			args.v1.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
+			if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) {
+				if (dig->lcd_misc & ATOM_PANEL_MISC_DUAL)
+					args.v1.ucMisc |= PANEL_ENCODER_MISC_DUAL;
+				if (dig->lcd_misc & ATOM_PANEL_MISC_888RGB)
+					args.v1.ucMisc |= ATOM_PANEL_MISC_888RGB;
+			} else {
+				if (dig->linkb)
+					args.v1.ucMisc |= PANEL_ENCODER_MISC_TMDS_LINKB;
+				if (radeon_encoder->pixel_clock > 165000)
+					args.v1.ucMisc |= PANEL_ENCODER_MISC_DUAL;
+				/*if (pScrn->rgbBits == 8) */
+				args.v1.ucMisc |= ATOM_PANEL_MISC_888RGB;
+			}
+			break;
+		case 2:
+		case 3:
+			args.v2.ucMisc = 0;
+			args.v2.ucAction = action;
+			if (crev == 3) {
+				if (dig->coherent_mode)
+					args.v2.ucMisc |= PANEL_ENCODER_MISC_COHERENT;
+			}
+			if (hdmi_detected)
+				args.v2.ucMisc |= PANEL_ENCODER_MISC_HDMI_TYPE;
+			args.v2.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
+			args.v2.ucTruncate = 0;
+			args.v2.ucSpatial = 0;
+			args.v2.ucTemporal = 0;
+			args.v2.ucFRC = 0;
+			if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) {
+				if (dig->lcd_misc & ATOM_PANEL_MISC_DUAL)
+					args.v2.ucMisc |= PANEL_ENCODER_MISC_DUAL;
+				if (dig->lcd_misc & ATOM_PANEL_MISC_SPATIAL) {
+					args.v2.ucSpatial = PANEL_ENCODER_SPATIAL_DITHER_EN;
+					if (dig->lcd_misc & ATOM_PANEL_MISC_888RGB)
+						args.v2.ucSpatial |= PANEL_ENCODER_SPATIAL_DITHER_DEPTH;
+				}
+				if (dig->lcd_misc & ATOM_PANEL_MISC_TEMPORAL) {
+					args.v2.ucTemporal = PANEL_ENCODER_TEMPORAL_DITHER_EN;
+					if (dig->lcd_misc & ATOM_PANEL_MISC_888RGB)
+						args.v2.ucTemporal |= PANEL_ENCODER_TEMPORAL_DITHER_DEPTH;
+					if (((dig->lcd_misc >> ATOM_PANEL_MISC_GREY_LEVEL_SHIFT) & 0x3) == 2)
+						args.v2.ucTemporal |= PANEL_ENCODER_TEMPORAL_LEVEL_4;
+				}
+			} else {
+				if (dig->linkb)
+					args.v2.ucMisc |= PANEL_ENCODER_MISC_TMDS_LINKB;
+				if (radeon_encoder->pixel_clock > 165000)
+					args.v2.ucMisc |= PANEL_ENCODER_MISC_DUAL;
+			}
+			break;
+		default:
+			DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
+			break;
+		}
+		break;
+	default:
+		DRM_ERROR("Unknown table version %d, %d\n", frev, crev);
+		break;
+	}
+
+	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+}
+
+int
+atombios_get_encoder_mode(struct drm_encoder *encoder)
+{
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
 	struct drm_connector *connector;
 	struct radeon_connector *radeon_connector;
 	struct radeon_connector_atom_dig *dig_connector;
 
+	/* dp bridges are always DP */
+	if (radeon_encoder_is_dp_bridge(encoder))
+		return ATOM_ENCODER_MODE_DP;
+
+	/* DVO is always DVO */
+	if (radeon_encoder->encoder_id == ATOM_ENCODER_MODE_DVO)
+		return ATOM_ENCODER_MODE_DVO;
+
 	connector = radeon_get_connector_for_encoder(encoder);
 	/* if we don't have an active device yet, just use one of
 	 * the connectors tied to the encoder.
@@ -321,45 +675,1747 @@ bool radeon_dig_monitor_is_duallink(struct drm_encoder *encoder,
 
 	switch (connector->connector_type) {
 	case DRM_MODE_CONNECTOR_DVII:
-	case DRM_MODE_CONNECTOR_HDMIB:
-		if (radeon_connector->use_digital) {
-			/* HDMI 1.3 supports up to 340 Mhz over single link */
-			if (0 && drm_detect_hdmi_monitor(radeon_connector->edid)) {
-				if (pixel_clock > 340000)
-					return true;
-				else
-					return false;
-			} else {
-				if (pixel_clock > 165000)
-					return true;
-				else
-					return false;
-			}
-		} else
-			return false;
+	case DRM_MODE_CONNECTOR_HDMIB: /* HDMI-B is basically DL-DVI; analog works fine */
+		if (drm_detect_monitor_audio(radeon_connector->edid) && radeon_audio) {
+			/* fix me */
+			if (ASIC_IS_DCE4(rdev))
+				return ATOM_ENCODER_MODE_DVI;
+			else
+				return ATOM_ENCODER_MODE_HDMI;
+		} else if (radeon_connector->use_digital)
+			return ATOM_ENCODER_MODE_DVI;
+		else
+			return ATOM_ENCODER_MODE_CRT;
+		break;
 	case DRM_MODE_CONNECTOR_DVID:
 	case DRM_MODE_CONNECTOR_HDMIA:
+	default:
+		if (drm_detect_monitor_audio(radeon_connector->edid) && radeon_audio) {
+			/* fix me */
+			if (ASIC_IS_DCE4(rdev))
+				return ATOM_ENCODER_MODE_DVI;
+			else
+				return ATOM_ENCODER_MODE_HDMI;
+		} else
+			return ATOM_ENCODER_MODE_DVI;
+		break;
+	case DRM_MODE_CONNECTOR_LVDS:
+		return ATOM_ENCODER_MODE_LVDS;
+		break;
 	case DRM_MODE_CONNECTOR_DisplayPort:
 		dig_connector = radeon_connector->con_priv;
 		if ((dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_DISPLAYPORT) ||
 		    (dig_connector->dp_sink_type == CONNECTOR_OBJECT_ID_eDP))
-			return false;
-		else {
-			/* HDMI 1.3 supports up to 340 Mhz over single link */
-			if (0 && drm_detect_hdmi_monitor(radeon_connector->edid)) {
-				if (pixel_clock > 340000)
-					return true;
+			return ATOM_ENCODER_MODE_DP;
+		else if (drm_detect_monitor_audio(radeon_connector->edid) && radeon_audio) {
+			/* fix me */
+			if (ASIC_IS_DCE4(rdev))
+				return ATOM_ENCODER_MODE_DVI;
+			else
+				return ATOM_ENCODER_MODE_HDMI;
+		} else
+			return ATOM_ENCODER_MODE_DVI;
+		break;
+	case DRM_MODE_CONNECTOR_eDP:
+		return ATOM_ENCODER_MODE_DP;
+	case DRM_MODE_CONNECTOR_DVIA:
+	case DRM_MODE_CONNECTOR_VGA:
+		return ATOM_ENCODER_MODE_CRT;
+		break;
+	case DRM_MODE_CONNECTOR_Composite:
+	case DRM_MODE_CONNECTOR_SVIDEO:
+	case DRM_MODE_CONNECTOR_9PinDIN:
+		/* fix me */
+		return ATOM_ENCODER_MODE_TV;
+		/*return ATOM_ENCODER_MODE_CV;*/
+		break;
+	}
+}
+
+/*
+ * DIG Encoder/Transmitter Setup
+ *
+ * DCE 3.0/3.1
+ * - 2 DIG transmitter blocks. UNIPHY (links A and B) and LVTMA.
+ * Supports up to 3 digital outputs
+ * - 2 DIG encoder blocks.
+ * DIG1 can drive UNIPHY link A or link B
+ * DIG2 can drive UNIPHY link B or LVTMA
+ *
+ * DCE 3.2
+ * - 3 DIG transmitter blocks. UNIPHY0/1/2 (links A and B).
+ * Supports up to 5 digital outputs
+ * - 2 DIG encoder blocks.
+ * DIG1/2 can drive UNIPHY0/1/2 link A or link B
+ *
+ * DCE 4.0/5.0
+ * - 3 DIG transmitter blocks UNIPHY0/1/2 (links A and B).
+ * Supports up to 6 digital outputs
+ * - 6 DIG encoder blocks.
+ * - DIG to PHY mapping is hardcoded
+ * DIG1 drives UNIPHY0 link A, A+B
+ * DIG2 drives UNIPHY0 link B
+ * DIG3 drives UNIPHY1 link A, A+B
+ * DIG4 drives UNIPHY1 link B
+ * DIG5 drives UNIPHY2 link A, A+B
+ * DIG6 drives UNIPHY2 link B
+ *
+ * DCE 4.1
+ * - 3 DIG transmitter blocks UNIPHY0/1/2 (links A and B).
+ * Supports up to 6 digital outputs
+ * - 2 DIG encoder blocks.
+ * DIG1/2 can drive UNIPHY0/1/2 link A or link B
+ *
+ * Routing
+ * crtc -> dig encoder -> UNIPHY/LVTMA (1 or 2 links)
+ * Examples:
+ * crtc0 -> dig2 -> LVTMA   links A+B -> TMDS/HDMI
+ * crtc1 -> dig1 -> UNIPHY0 link  B   -> DP
+ * crtc0 -> dig1 -> UNIPHY2 link  A   -> LVDS
+ * crtc1 -> dig2 -> UNIPHY1 link  B+A -> TMDS/HDMI
+ */
+
+union dig_encoder_control {
+	DIG_ENCODER_CONTROL_PS_ALLOCATION v1;
+	DIG_ENCODER_CONTROL_PARAMETERS_V2 v2;
+	DIG_ENCODER_CONTROL_PARAMETERS_V3 v3;
+	DIG_ENCODER_CONTROL_PARAMETERS_V4 v4;
+};
+
+void
+atombios_dig_encoder_setup(struct drm_encoder *encoder, int action, int panel_mode)
+{
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
+	struct drm_connector *connector = radeon_get_connector_for_encoder(encoder);
+	union dig_encoder_control args;
+	int index = 0;
+	uint8_t frev, crev;
+	int dp_clock = 0;
+	int dp_lane_count = 0;
+	int hpd_id = RADEON_HPD_NONE;
+	int bpc = 8;
+
+	if (connector) {
+		struct radeon_connector *radeon_connector = to_radeon_connector(connector);
+		struct radeon_connector_atom_dig *dig_connector =
+			radeon_connector->con_priv;
+
+		dp_clock = dig_connector->dp_clock;
+		dp_lane_count = dig_connector->dp_lane_count;
+		hpd_id = radeon_connector->hpd.hpd;
+		bpc = connector->display_info.bpc;
+	}
+
+	/* no dig encoder assigned */
+	if (dig->dig_encoder == -1)
+		return;
+
+	memset(&args, 0, sizeof(args));
+
+	if (ASIC_IS_DCE4(rdev))
+		index = GetIndexIntoMasterTable(COMMAND, DIGxEncoderControl);
+	else {
+		if (dig->dig_encoder)
+			index = GetIndexIntoMasterTable(COMMAND, DIG2EncoderControl);
+		else
+			index = GetIndexIntoMasterTable(COMMAND, DIG1EncoderControl);
+	}
+
+	if (!atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev))
+		return;
+
+	args.v1.ucAction = action;
+	args.v1.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
+	if (action == ATOM_ENCODER_CMD_SETUP_PANEL_MODE)
+		args.v3.ucPanelMode = panel_mode;
+	else
+		args.v1.ucEncoderMode = atombios_get_encoder_mode(encoder);
+
+	if ((args.v1.ucEncoderMode == ATOM_ENCODER_MODE_DP) ||
+	    (args.v1.ucEncoderMode == ATOM_ENCODER_MODE_DP_MST))
+		args.v1.ucLaneNum = dp_lane_count;
+	else if (radeon_encoder->pixel_clock > 165000)
+		args.v1.ucLaneNum = 8;
+	else
+		args.v1.ucLaneNum = 4;
+
+	if (ASIC_IS_DCE5(rdev)) {
+		if ((args.v1.ucEncoderMode == ATOM_ENCODER_MODE_DP) ||
+		    (args.v1.ucEncoderMode == ATOM_ENCODER_MODE_DP_MST)) {
+			if (dp_clock == 270000)
+				args.v1.ucConfig |= ATOM_ENCODER_CONFIG_V4_DPLINKRATE_2_70GHZ;
+			else if (dp_clock == 540000)
+				args.v1.ucConfig |= ATOM_ENCODER_CONFIG_V4_DPLINKRATE_5_40GHZ;
+		}
+		args.v4.acConfig.ucDigSel = dig->dig_encoder;
+		switch (bpc) {
+		case 0:
+			args.v4.ucBitPerColor = PANEL_BPC_UNDEFINE;
+			break;
+		case 6:
+			args.v4.ucBitPerColor = PANEL_6BIT_PER_COLOR;
+			break;
+		case 8:
+		default:
+			args.v4.ucBitPerColor = PANEL_8BIT_PER_COLOR;
+			break;
+		case 10:
+			args.v4.ucBitPerColor = PANEL_10BIT_PER_COLOR;
+			break;
+		case 12:
+			args.v4.ucBitPerColor = PANEL_12BIT_PER_COLOR;
+			break;
+		case 16:
+			args.v4.ucBitPerColor = PANEL_16BIT_PER_COLOR;
+			break;
+		}
+		if (hpd_id == RADEON_HPD_NONE)
+			args.v4.ucHPD_ID = 0;
+		else
+			args.v4.ucHPD_ID = hpd_id + 1;
+	} else if (ASIC_IS_DCE4(rdev)) {
+		if ((args.v1.ucEncoderMode == ATOM_ENCODER_MODE_DP) && (dp_clock == 270000))
+			args.v1.ucConfig |= ATOM_ENCODER_CONFIG_V3_DPLINKRATE_2_70GHZ;
+		args.v3.acConfig.ucDigSel = dig->dig_encoder;
+		switch (bpc) {
+		case 0:
+			args.v3.ucBitPerColor = PANEL_BPC_UNDEFINE;
+			break;
+		case 6:
+			args.v3.ucBitPerColor = PANEL_6BIT_PER_COLOR;
+			break;
+		case 8:
+		default:
+			args.v3.ucBitPerColor = PANEL_8BIT_PER_COLOR;
+			break;
+		case 10:
+			args.v3.ucBitPerColor = PANEL_10BIT_PER_COLOR;
+			break;
+		case 12:
+			args.v3.ucBitPerColor = PANEL_12BIT_PER_COLOR;
+			break;
+		case 16:
+			args.v3.ucBitPerColor = PANEL_16BIT_PER_COLOR;
+			break;
+		}
+	} else {
+		if ((args.v1.ucEncoderMode == ATOM_ENCODER_MODE_DP) && (dp_clock == 270000))
+			args.v1.ucConfig |= ATOM_ENCODER_CONFIG_DPLINKRATE_2_70GHZ;
+		switch (radeon_encoder->encoder_id) {
+		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
+			args.v1.ucConfig = ATOM_ENCODER_CONFIG_V2_TRANSMITTER1;
+			break;
+		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
+		case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
+			args.v1.ucConfig = ATOM_ENCODER_CONFIG_V2_TRANSMITTER2;
+			break;
+		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+			args.v1.ucConfig = ATOM_ENCODER_CONFIG_V2_TRANSMITTER3;
+			break;
+		}
+		if (dig->linkb)
+			args.v1.ucConfig |= ATOM_ENCODER_CONFIG_LINKB;
+		else
+			args.v1.ucConfig |= ATOM_ENCODER_CONFIG_LINKA;
+	}
+
+	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+}
+
+union dig_transmitter_control {
+	DIG_TRANSMITTER_CONTROL_PS_ALLOCATION v1;
+	DIG_TRANSMITTER_CONTROL_PARAMETERS_V2 v2;
+	DIG_TRANSMITTER_CONTROL_PARAMETERS_V3 v3;
+	DIG_TRANSMITTER_CONTROL_PARAMETERS_V4 v4;
+};
+
+void
+atombios_dig_transmitter_setup(struct drm_encoder *encoder, int action, uint8_t lane_num, uint8_t lane_set)
+{
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
+	struct drm_connector *connector;
+	union dig_transmitter_control args;
+	int index = 0;
+	uint8_t frev, crev;
+	bool is_dp = false;
+	int pll_id = 0;
+	int dp_clock = 0;
+	int dp_lane_count = 0;
+	int connector_object_id = 0;
+	int igp_lane_info = 0;
+	int dig_encoder = dig->dig_encoder;
+
+	if (action == ATOM_TRANSMITTER_ACTION_INIT) {
+		connector = radeon_get_connector_for_encoder_init(encoder);
+		/* just needed to avoid bailing in the encoder check.  the encoder
+		 * isn't used for init
+		 */
+		dig_encoder = 0;
+	} else
+		connector = radeon_get_connector_for_encoder(encoder);
+
+	if (connector) {
+		struct radeon_connector *radeon_connector = to_radeon_connector(connector);
+		struct radeon_connector_atom_dig *dig_connector =
+			radeon_connector->con_priv;
+
+		dp_clock = dig_connector->dp_clock;
+		dp_lane_count = dig_connector->dp_lane_count;
+		connector_object_id =
+			(radeon_connector->connector_object_id & OBJECT_ID_MASK) >> OBJECT_ID_SHIFT;
+		igp_lane_info = dig_connector->igp_lane_info;
+	}
+
+	/* no dig encoder assigned */
+	if (dig_encoder == -1)
+		return;
+
+	if (atombios_get_encoder_mode(encoder) == ATOM_ENCODER_MODE_DP)
+		is_dp = true;
+
+	memset(&args, 0, sizeof(args));
+
+	switch (radeon_encoder->encoder_id) {
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
+		index = GetIndexIntoMasterTable(COMMAND, DVOOutputControl);
+		break;
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+		index = GetIndexIntoMasterTable(COMMAND, UNIPHYTransmitterControl);
+		break;
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
+		index = GetIndexIntoMasterTable(COMMAND, LVTMATransmitterControl);
+		break;
+	}
+
+	if (!atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev))
+		return;
+
+	args.v1.ucAction = action;
+	if (action == ATOM_TRANSMITTER_ACTION_INIT) {
+		args.v1.usInitInfo = cpu_to_le16(connector_object_id);
+	} else if (action == ATOM_TRANSMITTER_ACTION_SETUP_VSEMPH) {
+		args.v1.asMode.ucLaneSel = lane_num;
+		args.v1.asMode.ucLaneSet = lane_set;
+	} else {
+		if (is_dp)
+			args.v1.usPixelClock =
+				cpu_to_le16(dp_clock / 10);
+		else if (radeon_encoder->pixel_clock > 165000)
+			args.v1.usPixelClock = cpu_to_le16((radeon_encoder->pixel_clock / 2) / 10);
+		else
+			args.v1.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
+	}
+	if (ASIC_IS_DCE4(rdev)) {
+		if (is_dp)
+			args.v3.ucLaneNum = dp_lane_count;
+		else if (radeon_encoder->pixel_clock > 165000)
+			args.v3.ucLaneNum = 8;
+		else
+			args.v3.ucLaneNum = 4;
+
+		if (dig->linkb)
+			args.v3.acConfig.ucLinkSel = 1;
+		if (dig_encoder & 1)
+			args.v3.acConfig.ucEncoderSel = 1;
+
+		/* Select the PLL for the PHY
+		 * DP PHY should be clocked from external src if there is
+		 * one.
+		 */
+		if (encoder->crtc) {
+			struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc);
+			pll_id = radeon_crtc->pll_id;
+		}
+
+		if (ASIC_IS_DCE5(rdev)) {
+			/* On DCE5 DCPLL usually generates the DP ref clock */
+			if (is_dp) {
+				if (rdev->clock.dp_extclk)
+					args.v4.acConfig.ucRefClkSource = ENCODER_REFCLK_SRC_EXTCLK;
 				else
-					return false;
+					args.v4.acConfig.ucRefClkSource = ENCODER_REFCLK_SRC_DCPLL;
+			} else
+				args.v4.acConfig.ucRefClkSource = pll_id;
+		} else {
+			/* On DCE4, if there is an external clock, it generates the DP ref clock */
+			if (is_dp && rdev->clock.dp_extclk)
+				args.v3.acConfig.ucRefClkSource = 2; /* external src */
+			else
+				args.v3.acConfig.ucRefClkSource = pll_id;
+		}
+
+		switch (radeon_encoder->encoder_id) {
+		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
+			args.v3.acConfig.ucTransmitterSel = 0;
+			break;
+		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
+			args.v3.acConfig.ucTransmitterSel = 1;
+			break;
+		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+			args.v3.acConfig.ucTransmitterSel = 2;
+			break;
+		}
+
+		if (is_dp)
+			args.v3.acConfig.fCoherentMode = 1; /* DP requires coherent */
+		else if (radeon_encoder->devices & (ATOM_DEVICE_DFP_SUPPORT)) {
+			if (dig->coherent_mode)
+				args.v3.acConfig.fCoherentMode = 1;
+			if (radeon_encoder->pixel_clock > 165000)
+				args.v3.acConfig.fDualLinkConnector = 1;
+		}
+	} else if (ASIC_IS_DCE32(rdev)) {
+		args.v2.acConfig.ucEncoderSel = dig_encoder;
+		if (dig->linkb)
+			args.v2.acConfig.ucLinkSel = 1;
+
+		switch (radeon_encoder->encoder_id) {
+		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
+			args.v2.acConfig.ucTransmitterSel = 0;
+			break;
+		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
+			args.v2.acConfig.ucTransmitterSel = 1;
+			break;
+		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+			args.v2.acConfig.ucTransmitterSel = 2;
+			break;
+		}
+
+		if (is_dp) {
+			args.v2.acConfig.fCoherentMode = 1;
+			args.v2.acConfig.fDPConnector = 1;
+		} else if (radeon_encoder->devices & (ATOM_DEVICE_DFP_SUPPORT)) {
+			if (dig->coherent_mode)
+				args.v2.acConfig.fCoherentMode = 1;
+			if (radeon_encoder->pixel_clock > 165000)
+				args.v2.acConfig.fDualLinkConnector = 1;
+		}
+	} else {
+		args.v1.ucConfig = ATOM_TRANSMITTER_CONFIG_CLKSRC_PPLL;
+
+		if (dig_encoder)
+			args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_DIG2_ENCODER;
+		else
+			args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_DIG1_ENCODER;
+
+		if ((rdev->flags & RADEON_IS_IGP) &&
+		    (radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_UNIPHY)) {
+			if (is_dp || (radeon_encoder->pixel_clock <= 165000)) {
+				if (igp_lane_info & 0x1)
+					args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LANE_0_3;
+				else if (igp_lane_info & 0x2)
+					args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LANE_4_7;
+				else if (igp_lane_info & 0x4)
+					args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LANE_8_11;
+				else if (igp_lane_info & 0x8)
+					args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LANE_12_15;
 			} else {
-				if (pixel_clock > 165000)
-					return true;
+				if (igp_lane_info & 0x3)
+					args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LANE_0_7;
+				else if (igp_lane_info & 0xc)
+					args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LANE_8_15;
+			}
+		}
+
+		if (dig->linkb)
+			args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LINKB;
+		else
+			args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_LINKA;
+
+		if (is_dp)
+			args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_COHERENT;
+		else if (radeon_encoder->devices & (ATOM_DEVICE_DFP_SUPPORT)) {
+			if (dig->coherent_mode)
+				args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_COHERENT;
+			if (radeon_encoder->pixel_clock > 165000)
+				args.v1.ucConfig |= ATOM_TRANSMITTER_CONFIG_8LANE_LINK;
+		}
+	}
+
+	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+}
+
+bool
+atombios_set_edp_panel_power(struct drm_connector *connector, int action)
+{
+	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
+	struct drm_device *dev = radeon_connector->base.dev;
+	struct radeon_device *rdev = dev->dev_private;
+	union dig_transmitter_control args;
+	int index = GetIndexIntoMasterTable(COMMAND, UNIPHYTransmitterControl);
+	uint8_t frev, crev;
+
+	if (connector->connector_type != DRM_MODE_CONNECTOR_eDP)
+		goto done;
+
+	if (!ASIC_IS_DCE4(rdev))
+		goto done;
+
+	if ((action != ATOM_TRANSMITTER_ACTION_POWER_ON) &&
+	    (action != ATOM_TRANSMITTER_ACTION_POWER_OFF))
+		goto done;
+
+	if (!atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev))
+		goto done;
+
+	memset(&args, 0, sizeof(args));
+
+	args.v1.ucAction = action;
+
+	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+	/* wait for the panel to power up */
+	if (action == ATOM_TRANSMITTER_ACTION_POWER_ON) {
+		int i;
+
+		for (i = 0; i < 300; i++) {
+			if (radeon_hpd_sense(rdev, radeon_connector->hpd.hpd))
+				return true;
+			mdelay(1);
+		}
+		return false;
+	}
+done:
+	return true;
+}
+
+union external_encoder_control {
+	EXTERNAL_ENCODER_CONTROL_PS_ALLOCATION v1;
+	EXTERNAL_ENCODER_CONTROL_PS_ALLOCATION_V3 v3;
+};
+
+static void
+atombios_external_encoder_setup(struct drm_encoder *encoder,
+				struct drm_encoder *ext_encoder,
+				int action)
+{
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	struct radeon_encoder *ext_radeon_encoder = to_radeon_encoder(ext_encoder);
+	union external_encoder_control args;
+	struct drm_connector *connector;
+	int index = GetIndexIntoMasterTable(COMMAND, ExternalEncoderControl);
+	u8 frev, crev;
+	int dp_clock = 0;
+	int dp_lane_count = 0;
+	int connector_object_id = 0;
+	u32 ext_enum = (ext_radeon_encoder->encoder_enum & ENUM_ID_MASK) >> ENUM_ID_SHIFT;
+	int bpc = 8;
+
+	if (action == EXTERNAL_ENCODER_ACTION_V3_ENCODER_INIT)
+		connector = radeon_get_connector_for_encoder_init(encoder);
+	else
+		connector = radeon_get_connector_for_encoder(encoder);
+
+	if (connector) {
+		struct radeon_connector *radeon_connector = to_radeon_connector(connector);
+		struct radeon_connector_atom_dig *dig_connector =
+			radeon_connector->con_priv;
+
+		dp_clock = dig_connector->dp_clock;
+		dp_lane_count = dig_connector->dp_lane_count;
+		connector_object_id =
+			(radeon_connector->connector_object_id & OBJECT_ID_MASK) >> OBJECT_ID_SHIFT;
+		bpc = connector->display_info.bpc;
+	}
+
+	memset(&args, 0, sizeof(args));
+
+	if (!atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev))
+		return;
+
+	switch (frev) {
+	case 1:
+		/* no params on frev 1 */
+		break;
+	case 2:
+		switch (crev) {
+		case 1:
+		case 2:
+			args.v1.sDigEncoder.ucAction = action;
+			args.v1.sDigEncoder.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
+			args.v1.sDigEncoder.ucEncoderMode = atombios_get_encoder_mode(encoder);
+
+			if (args.v1.sDigEncoder.ucEncoderMode == ATOM_ENCODER_MODE_DP) {
+				if (dp_clock == 270000)
+					args.v1.sDigEncoder.ucConfig |= ATOM_ENCODER_CONFIG_DPLINKRATE_2_70GHZ;
+				args.v1.sDigEncoder.ucLaneNum = dp_lane_count;
+			} else if (radeon_encoder->pixel_clock > 165000)
+				args.v1.sDigEncoder.ucLaneNum = 8;
+			else
+				args.v1.sDigEncoder.ucLaneNum = 4;
+			break;
+		case 3:
+			args.v3.sExtEncoder.ucAction = action;
+			if (action == EXTERNAL_ENCODER_ACTION_V3_ENCODER_INIT)
+				args.v3.sExtEncoder.usConnectorId = cpu_to_le16(connector_object_id);
+			else
+				args.v3.sExtEncoder.usPixelClock = cpu_to_le16(radeon_encoder->pixel_clock / 10);
+			args.v3.sExtEncoder.ucEncoderMode = atombios_get_encoder_mode(encoder);
+
+			if (args.v3.sExtEncoder.ucEncoderMode == ATOM_ENCODER_MODE_DP) {
+				if (dp_clock == 270000)
+					args.v3.sExtEncoder.ucConfig |= EXTERNAL_ENCODER_CONFIG_V3_DPLINKRATE_2_70GHZ;
+				else if (dp_clock == 540000)
+					args.v3.sExtEncoder.ucConfig |= EXTERNAL_ENCODER_CONFIG_V3_DPLINKRATE_5_40GHZ;
+				args.v3.sExtEncoder.ucLaneNum = dp_lane_count;
+			} else if (radeon_encoder->pixel_clock > 165000)
+				args.v3.sExtEncoder.ucLaneNum = 8;
+			else
+				args.v3.sExtEncoder.ucLaneNum = 4;
+			switch (ext_enum) {
+			case GRAPH_OBJECT_ENUM_ID1:
+				args.v3.sExtEncoder.ucConfig |= EXTERNAL_ENCODER_CONFIG_V3_ENCODER1;
+				break;
+			case GRAPH_OBJECT_ENUM_ID2:
+				args.v3.sExtEncoder.ucConfig |= EXTERNAL_ENCODER_CONFIG_V3_ENCODER2;
+				break;
+			case GRAPH_OBJECT_ENUM_ID3:
+				args.v3.sExtEncoder.ucConfig |= EXTERNAL_ENCODER_CONFIG_V3_ENCODER3;
+				break;
+			}
+			switch (bpc) {
+			case 0:
+				args.v3.sExtEncoder.ucBitPerColor = PANEL_BPC_UNDEFINE;
+				break;
+			case 6:
+				args.v3.sExtEncoder.ucBitPerColor = PANEL_6BIT_PER_COLOR;
+				break;
+			case 8:
+			default:
+				args.v3.sExtEncoder.ucBitPerColor = PANEL_8BIT_PER_COLOR;
+				break;
+			case 10:
+				args.v3.sExtEncoder.ucBitPerColor = PANEL_10BIT_PER_COLOR;
+				break;
+			case 12:
+				args.v3.sExtEncoder.ucBitPerColor = PANEL_12BIT_PER_COLOR;
+				break;
+			case 16:
+				args.v3.sExtEncoder.ucBitPerColor = PANEL_16BIT_PER_COLOR;
+				break;
+			}
+			break;
+		default:
+			DRM_ERROR("Unknown table version: %d, %d\n", frev, crev);
+			return;
+		}
+		break;
+	default:
+		DRM_ERROR("Unknown table version: %d, %d\n", frev, crev);
+		return;
+	}
+	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+}
+
+static void
+atombios_yuv_setup(struct drm_encoder *encoder, bool enable)
+{
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc);
+	ENABLE_YUV_PS_ALLOCATION args;
+	int index = GetIndexIntoMasterTable(COMMAND, EnableYUV);
+	uint32_t temp, reg;
+
+	memset(&args, 0, sizeof(args));
+
+	if (rdev->family >= CHIP_R600)
+		reg = R600_BIOS_3_SCRATCH;
+	else
+		reg = RADEON_BIOS_3_SCRATCH;
+
+	/* XXX: fix up scratch reg handling */
+	temp = RREG32(reg);
+	if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT))
+		WREG32(reg, (ATOM_S3_TV1_ACTIVE |
+			     (radeon_crtc->crtc_id << 18)));
+	else if (radeon_encoder->active_device & (ATOM_DEVICE_CV_SUPPORT))
+		WREG32(reg, (ATOM_S3_CV_ACTIVE | (radeon_crtc->crtc_id << 24)));
+	else
+		WREG32(reg, 0);
+
+	if (enable)
+		args.ucEnable = ATOM_ENABLE;
+	args.ucCRTC = radeon_crtc->crtc_id;
+
+	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+	WREG32(reg, temp);
+}
+
+static void
+radeon_atom_encoder_dpms(struct drm_encoder *encoder, int mode)
+{
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	struct drm_encoder *ext_encoder = radeon_atom_get_external_encoder(encoder);
+	DISPLAY_DEVICE_OUTPUT_CONTROL_PS_ALLOCATION args;
+	int index = 0;
+	bool is_dig = false;
+	bool is_dce5_dac = false;
+	bool is_dce5_dvo = false;
+
+	memset(&args, 0, sizeof(args));
+
+	DRM_DEBUG_KMS("encoder dpms %d to mode %d, devices %08x, active_devices %08x\n",
+		  radeon_encoder->encoder_id, mode, radeon_encoder->devices,
+		  radeon_encoder->active_device);
+	switch (radeon_encoder->encoder_id) {
+	case ENCODER_OBJECT_ID_INTERNAL_TMDS1:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
+		index = GetIndexIntoMasterTable(COMMAND, TMDSAOutputControl);
+		break;
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
+		is_dig = true;
+		break;
+	case ENCODER_OBJECT_ID_INTERNAL_DVO1:
+	case ENCODER_OBJECT_ID_INTERNAL_DDI:
+		index = GetIndexIntoMasterTable(COMMAND, DVOOutputControl);
+		break;
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
+		if (ASIC_IS_DCE5(rdev))
+			is_dce5_dvo = true;
+		else if (ASIC_IS_DCE3(rdev))
+			is_dig = true;
+		else
+			index = GetIndexIntoMasterTable(COMMAND, DVOOutputControl);
+		break;
+	case ENCODER_OBJECT_ID_INTERNAL_LVDS:
+		index = GetIndexIntoMasterTable(COMMAND, LCD1OutputControl);
+		break;
+	case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
+		if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT))
+			index = GetIndexIntoMasterTable(COMMAND, LCD1OutputControl);
+		else
+			index = GetIndexIntoMasterTable(COMMAND, LVTMAOutputControl);
+		break;
+	case ENCODER_OBJECT_ID_INTERNAL_DAC1:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1:
+		if (ASIC_IS_DCE5(rdev))
+			is_dce5_dac = true;
+		else {
+			if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT))
+				index = GetIndexIntoMasterTable(COMMAND, TV1OutputControl);
+			else if (radeon_encoder->active_device & (ATOM_DEVICE_CV_SUPPORT))
+				index = GetIndexIntoMasterTable(COMMAND, CV1OutputControl);
+			else
+				index = GetIndexIntoMasterTable(COMMAND, DAC1OutputControl);
+		}
+		break;
+	case ENCODER_OBJECT_ID_INTERNAL_DAC2:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC2:
+		if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT))
+			index = GetIndexIntoMasterTable(COMMAND, TV1OutputControl);
+		else if (radeon_encoder->active_device & (ATOM_DEVICE_CV_SUPPORT))
+			index = GetIndexIntoMasterTable(COMMAND, CV1OutputControl);
+		else
+			index = GetIndexIntoMasterTable(COMMAND, DAC2OutputControl);
+		break;
+	}
+
+	if (is_dig) {
+		switch (mode) {
+		case DRM_MODE_DPMS_ON:
+			/* some early dce3.2 boards have a bug in their transmitter control table */
+			if ((rdev->family == CHIP_RV710) || (rdev->family == CHIP_RV730))
+				atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_ENABLE, 0, 0);
+			else
+				atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_ENABLE_OUTPUT, 0, 0);
+			if (atombios_get_encoder_mode(encoder) == ATOM_ENCODER_MODE_DP) {
+				struct drm_connector *connector = radeon_get_connector_for_encoder(encoder);
+
+				if (connector &&
+				    (connector->connector_type == DRM_MODE_CONNECTOR_eDP)) {
+					struct radeon_connector *radeon_connector = to_radeon_connector(connector);
+					struct radeon_connector_atom_dig *radeon_dig_connector =
+						radeon_connector->con_priv;
+					atombios_set_edp_panel_power(connector,
+								     ATOM_TRANSMITTER_ACTION_POWER_ON);
+					radeon_dig_connector->edp_on = true;
+				}
+				if (ASIC_IS_DCE4(rdev))
+					atombios_dig_encoder_setup(encoder, ATOM_ENCODER_CMD_DP_VIDEO_OFF, 0);
+				radeon_dp_link_train(encoder, connector);
+				if (ASIC_IS_DCE4(rdev))
+					atombios_dig_encoder_setup(encoder, ATOM_ENCODER_CMD_DP_VIDEO_ON, 0);
+			}
+			if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT))
+				atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_LCD_BLON, 0, 0);
+			break;
+		case DRM_MODE_DPMS_STANDBY:
+		case DRM_MODE_DPMS_SUSPEND:
+		case DRM_MODE_DPMS_OFF:
+			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_DISABLE_OUTPUT, 0, 0);
+			if (atombios_get_encoder_mode(encoder) == ATOM_ENCODER_MODE_DP) {
+				struct drm_connector *connector = radeon_get_connector_for_encoder(encoder);
+
+				if (ASIC_IS_DCE4(rdev))
+					atombios_dig_encoder_setup(encoder, ATOM_ENCODER_CMD_DP_VIDEO_OFF, 0);
+				if (connector &&
+				    (connector->connector_type == DRM_MODE_CONNECTOR_eDP)) {
+					struct radeon_connector *radeon_connector = to_radeon_connector(connector);
+					struct radeon_connector_atom_dig *radeon_dig_connector =
+						radeon_connector->con_priv;
+					atombios_set_edp_panel_power(connector,
+								     ATOM_TRANSMITTER_ACTION_POWER_OFF);
+					radeon_dig_connector->edp_on = false;
+				}
+			}
+			if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT))
+				atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_LCD_BLOFF, 0, 0);
+			break;
+		}
+	} else if (is_dce5_dac) {
+		switch (mode) {
+		case DRM_MODE_DPMS_ON:
+			atombios_dac_setup(encoder, ATOM_ENABLE);
+			break;
+		case DRM_MODE_DPMS_STANDBY:
+		case DRM_MODE_DPMS_SUSPEND:
+		case DRM_MODE_DPMS_OFF:
+			atombios_dac_setup(encoder, ATOM_DISABLE);
+			break;
+		}
+	} else if (is_dce5_dvo) {
+		switch (mode) {
+		case DRM_MODE_DPMS_ON:
+			atombios_dvo_setup(encoder, ATOM_ENABLE);
+			break;
+		case DRM_MODE_DPMS_STANDBY:
+		case DRM_MODE_DPMS_SUSPEND:
+		case DRM_MODE_DPMS_OFF:
+			atombios_dvo_setup(encoder, ATOM_DISABLE);
+			break;
+		}
+	} else {
+		switch (mode) {
+		case DRM_MODE_DPMS_ON:
+			args.ucAction = ATOM_ENABLE;
+			/* workaround for DVOOutputControl on some RS690 systems */
+			if (radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_DDI) {
+				u32 reg = RREG32(RADEON_BIOS_3_SCRATCH);
+				WREG32(RADEON_BIOS_3_SCRATCH, reg & ~ATOM_S3_DFP2I_ACTIVE);
+				atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+				WREG32(RADEON_BIOS_3_SCRATCH, reg);
+			} else
+				atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+			if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) {
+				args.ucAction = ATOM_LCD_BLON;
+				atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+			}
+			break;
+		case DRM_MODE_DPMS_STANDBY:
+		case DRM_MODE_DPMS_SUSPEND:
+		case DRM_MODE_DPMS_OFF:
+			args.ucAction = ATOM_DISABLE;
+			atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+			if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) {
+				args.ucAction = ATOM_LCD_BLOFF;
+				atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+			}
+			break;
+		}
+	}
+
+	if (ext_encoder) {
+		switch (mode) {
+		case DRM_MODE_DPMS_ON:
+		default:
+			if (ASIC_IS_DCE41(rdev)) {
+				atombios_external_encoder_setup(encoder, ext_encoder,
+								EXTERNAL_ENCODER_ACTION_V3_ENABLE_OUTPUT);
+				atombios_external_encoder_setup(encoder, ext_encoder,
+								EXTERNAL_ENCODER_ACTION_V3_ENCODER_BLANKING_OFF);
+			} else
+				atombios_external_encoder_setup(encoder, ext_encoder, ATOM_ENABLE);
+			break;
+		case DRM_MODE_DPMS_STANDBY:
+		case DRM_MODE_DPMS_SUSPEND:
+		case DRM_MODE_DPMS_OFF:
+			if (ASIC_IS_DCE41(rdev)) {
+				atombios_external_encoder_setup(encoder, ext_encoder,
+								EXTERNAL_ENCODER_ACTION_V3_ENCODER_BLANKING);
+				atombios_external_encoder_setup(encoder, ext_encoder,
+								EXTERNAL_ENCODER_ACTION_V3_DISABLE_OUTPUT);
+			} else
+				atombios_external_encoder_setup(encoder, ext_encoder, ATOM_DISABLE);
+			break;
+		}
+	}
+
+	radeon_atombios_encoder_dpms_scratch_regs(encoder, (mode == DRM_MODE_DPMS_ON) ? true : false);
+
+}
+
+union crtc_source_param {
+	SELECT_CRTC_SOURCE_PS_ALLOCATION v1;
+	SELECT_CRTC_SOURCE_PARAMETERS_V2 v2;
+};
+
+static void
+atombios_set_encoder_crtc_source(struct drm_encoder *encoder)
+{
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc);
+	union crtc_source_param args;
+	int index = GetIndexIntoMasterTable(COMMAND, SelectCRTC_Source);
+	uint8_t frev, crev;
+	struct radeon_encoder_atom_dig *dig;
+
+	memset(&args, 0, sizeof(args));
+
+	if (!atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev))
+		return;
+
+	switch (frev) {
+	case 1:
+		switch (crev) {
+		case 1:
+		default:
+			if (ASIC_IS_AVIVO(rdev))
+				args.v1.ucCRTC = radeon_crtc->crtc_id;
+			else {
+				if (radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_DAC1) {
+					args.v1.ucCRTC = radeon_crtc->crtc_id;
+				} else {
+					args.v1.ucCRTC = radeon_crtc->crtc_id << 2;
+				}
+			}
+			switch (radeon_encoder->encoder_id) {
+			case ENCODER_OBJECT_ID_INTERNAL_TMDS1:
+			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
+				args.v1.ucDevice = ATOM_DEVICE_DFP1_INDEX;
+				break;
+			case ENCODER_OBJECT_ID_INTERNAL_LVDS:
+			case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
+				if (radeon_encoder->devices & ATOM_DEVICE_LCD1_SUPPORT)
+					args.v1.ucDevice = ATOM_DEVICE_LCD1_INDEX;
+				else
+					args.v1.ucDevice = ATOM_DEVICE_DFP3_INDEX;
+				break;
+			case ENCODER_OBJECT_ID_INTERNAL_DVO1:
+			case ENCODER_OBJECT_ID_INTERNAL_DDI:
+			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
+				args.v1.ucDevice = ATOM_DEVICE_DFP2_INDEX;
+				break;
+			case ENCODER_OBJECT_ID_INTERNAL_DAC1:
+			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1:
+				if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT))
+					args.v1.ucDevice = ATOM_DEVICE_TV1_INDEX;
+				else if (radeon_encoder->active_device & (ATOM_DEVICE_CV_SUPPORT))
+					args.v1.ucDevice = ATOM_DEVICE_CV_INDEX;
+				else
+					args.v1.ucDevice = ATOM_DEVICE_CRT1_INDEX;
+				break;
+			case ENCODER_OBJECT_ID_INTERNAL_DAC2:
+			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC2:
+				if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT))
+					args.v1.ucDevice = ATOM_DEVICE_TV1_INDEX;
+				else if (radeon_encoder->active_device & (ATOM_DEVICE_CV_SUPPORT))
+					args.v1.ucDevice = ATOM_DEVICE_CV_INDEX;
+				else
+					args.v1.ucDevice = ATOM_DEVICE_CRT2_INDEX;
+				break;
+			}
+			break;
+		case 2:
+			args.v2.ucCRTC = radeon_crtc->crtc_id;
+			args.v2.ucEncodeMode = atombios_get_encoder_mode(encoder);
+			switch (radeon_encoder->encoder_id) {
+			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
+			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
+			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
+				dig = radeon_encoder->enc_priv;
+				switch (dig->dig_encoder) {
+				case 0:
+					args.v2.ucEncoderID = ASIC_INT_DIG1_ENCODER_ID;
+					break;
+				case 1:
+					args.v2.ucEncoderID = ASIC_INT_DIG2_ENCODER_ID;
+					break;
+				case 2:
+					args.v2.ucEncoderID = ASIC_INT_DIG3_ENCODER_ID;
+					break;
+				case 3:
+					args.v2.ucEncoderID = ASIC_INT_DIG4_ENCODER_ID;
+					break;
+				case 4:
+					args.v2.ucEncoderID = ASIC_INT_DIG5_ENCODER_ID;
+					break;
+				case 5:
+					args.v2.ucEncoderID = ASIC_INT_DIG6_ENCODER_ID;
+					break;
+				}
+				break;
+			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
+				args.v2.ucEncoderID = ASIC_INT_DVO_ENCODER_ID;
+				break;
+			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1:
+				if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT))
+					args.v2.ucEncoderID = ASIC_INT_TV_ENCODER_ID;
+				else if (radeon_encoder->active_device & (ATOM_DEVICE_CV_SUPPORT))
+					args.v2.ucEncoderID = ASIC_INT_TV_ENCODER_ID;
+				else
+					args.v2.ucEncoderID = ASIC_INT_DAC1_ENCODER_ID;
+				break;
+			case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC2:
+				if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT))
+					args.v2.ucEncoderID = ASIC_INT_TV_ENCODER_ID;
+				else if (radeon_encoder->active_device & (ATOM_DEVICE_CV_SUPPORT))
+					args.v2.ucEncoderID = ASIC_INT_TV_ENCODER_ID;
 				else
-					return false;
+					args.v2.ucEncoderID = ASIC_INT_DAC2_ENCODER_ID;
+				break;
 			}
+			break;
 		}
+		break;
 	default:
+		DRM_ERROR("Unknown table version: %d, %d\n", frev, crev);
+		return;
+	}
+
+	atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+	/* update scratch regs with new routing */
+	radeon_atombios_encoder_crtc_scratch_regs(encoder, radeon_crtc->crtc_id);
+}
+
+static void
+atombios_apply_encoder_quirks(struct drm_encoder *encoder,
+			      struct drm_display_mode *mode)
+{
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc);
+
+	/* Funky macbooks */
+	if ((dev->pdev->device == 0x71C5) &&
+	    (dev->pdev->subsystem_vendor == 0x106b) &&
+	    (dev->pdev->subsystem_device == 0x0080)) {
+		if (radeon_encoder->devices & ATOM_DEVICE_LCD1_SUPPORT) {
+			uint32_t lvtma_bit_depth_control = RREG32(AVIVO_LVTMA_BIT_DEPTH_CONTROL);
+
+			lvtma_bit_depth_control &= ~AVIVO_LVTMA_BIT_DEPTH_CONTROL_TRUNCATE_EN;
+			lvtma_bit_depth_control &= ~AVIVO_LVTMA_BIT_DEPTH_CONTROL_SPATIAL_DITHER_EN;
+
+			WREG32(AVIVO_LVTMA_BIT_DEPTH_CONTROL, lvtma_bit_depth_control);
+		}
+	}
+
+	/* set scaler clears this on some chips */
+	if (ASIC_IS_AVIVO(rdev) &&
+	    (!(radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT)))) {
+		if (ASIC_IS_DCE4(rdev)) {
+			if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+				WREG32(EVERGREEN_DATA_FORMAT + radeon_crtc->crtc_offset,
+				       EVERGREEN_INTERLEAVE_EN);
+			else
+				WREG32(EVERGREEN_DATA_FORMAT + radeon_crtc->crtc_offset, 0);
+		} else {
+			if (mode->flags & DRM_MODE_FLAG_INTERLACE)
+				WREG32(AVIVO_D1MODE_DATA_FORMAT + radeon_crtc->crtc_offset,
+				       AVIVO_D1MODE_INTERLEAVE_EN);
+			else
+				WREG32(AVIVO_D1MODE_DATA_FORMAT + radeon_crtc->crtc_offset, 0);
+		}
+	}
+}
+
+static int radeon_atom_pick_dig_encoder(struct drm_encoder *encoder)
+{
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_crtc *radeon_crtc = to_radeon_crtc(encoder->crtc);
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	struct drm_encoder *test_encoder;
+	struct radeon_encoder_atom_dig *dig;
+	uint32_t dig_enc_in_use = 0;
+
+	/* DCE4/5 */
+	if (ASIC_IS_DCE4(rdev)) {
+		dig = radeon_encoder->enc_priv;
+		if (ASIC_IS_DCE41(rdev)) {
+			if (dig->linkb)
+				return 1;
+			else
+				return 0;
+		} else {
+			switch (radeon_encoder->encoder_id) {
+			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
+				if (dig->linkb)
+					return 1;
+				else
+					return 0;
+				break;
+			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
+				if (dig->linkb)
+					return 3;
+				else
+					return 2;
+				break;
+			case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+				if (dig->linkb)
+					return 5;
+				else
+					return 4;
+				break;
+			}
+		}
+	}
+
+	/* on DCE32 and encoder can driver any block so just crtc id */
+	if (ASIC_IS_DCE32(rdev)) {
+		return radeon_crtc->crtc_id;
+	}
+
+	/* on DCE3 - LVTMA can only be driven by DIGB */
+	list_for_each_entry(test_encoder, &dev->mode_config.encoder_list, head) {
+		struct radeon_encoder *radeon_test_encoder;
+
+		if (encoder == test_encoder)
+			continue;
+
+		if (!radeon_encoder_is_digital(test_encoder))
+			continue;
+
+		radeon_test_encoder = to_radeon_encoder(test_encoder);
+		dig = radeon_test_encoder->enc_priv;
+
+		if (dig->dig_encoder >= 0)
+			dig_enc_in_use |= (1 << dig->dig_encoder);
+	}
+
+	if (radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA) {
+		if (dig_enc_in_use & 0x2)
+			DRM_ERROR("LVDS required digital encoder 2 but it was in use - stealing\n");
+		return 1;
+	}
+	if (!(dig_enc_in_use & 1))
+		return 0;
+	return 1;
+}
+
+/* This only needs to be called once at startup */
+void
+radeon_atom_encoder_init(struct radeon_device *rdev)
+{
+	struct drm_device *dev = rdev->ddev;
+	struct drm_encoder *encoder;
+
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
+		struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+		struct drm_encoder *ext_encoder = radeon_atom_get_external_encoder(encoder);
+
+		switch (radeon_encoder->encoder_id) {
+		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
+		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
+		case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+		case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
+			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_INIT, 0, 0);
+			break;
+		default:
+			break;
+		}
+
+		if (ext_encoder && ASIC_IS_DCE41(rdev))
+			atombios_external_encoder_setup(encoder, ext_encoder,
+							EXTERNAL_ENCODER_ACTION_V3_ENCODER_INIT);
+	}
+}
+
+static void
+radeon_atom_encoder_mode_set(struct drm_encoder *encoder,
+			     struct drm_display_mode *mode,
+			     struct drm_display_mode *adjusted_mode)
+{
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	struct drm_encoder *ext_encoder = radeon_atom_get_external_encoder(encoder);
+
+	radeon_encoder->pixel_clock = adjusted_mode->clock;
+
+	if (ASIC_IS_AVIVO(rdev) && !ASIC_IS_DCE4(rdev)) {
+		if (radeon_encoder->active_device & (ATOM_DEVICE_CV_SUPPORT | ATOM_DEVICE_TV_SUPPORT))
+			atombios_yuv_setup(encoder, true);
+		else
+			atombios_yuv_setup(encoder, false);
+	}
+
+	switch (radeon_encoder->encoder_id) {
+	case ENCODER_OBJECT_ID_INTERNAL_TMDS1:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
+	case ENCODER_OBJECT_ID_INTERNAL_LVDS:
+	case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
+		atombios_digital_setup(encoder, PANEL_ENCODER_ACTION_ENABLE);
+		break;
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
+		if (ASIC_IS_DCE4(rdev)) {
+			/* disable the transmitter */
+			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_DISABLE, 0, 0);
+			/* setup and enable the encoder */
+			atombios_dig_encoder_setup(encoder, ATOM_ENCODER_CMD_SETUP, 0);
+
+			/* enable the transmitter */
+			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_ENABLE, 0, 0);
+		} else {
+			/* disable the encoder and transmitter */
+			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_DISABLE, 0, 0);
+			atombios_dig_encoder_setup(encoder, ATOM_DISABLE, 0);
+
+			/* setup and enable the encoder and transmitter */
+			atombios_dig_encoder_setup(encoder, ATOM_ENABLE, 0);
+			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_SETUP, 0, 0);
+			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_ENABLE, 0, 0);
+		}
+		break;
+	case ENCODER_OBJECT_ID_INTERNAL_DDI:
+	case ENCODER_OBJECT_ID_INTERNAL_DVO1:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
+		atombios_dvo_setup(encoder, ATOM_ENABLE);
+		break;
+	case ENCODER_OBJECT_ID_INTERNAL_DAC1:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1:
+	case ENCODER_OBJECT_ID_INTERNAL_DAC2:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC2:
+		atombios_dac_setup(encoder, ATOM_ENABLE);
+		if (radeon_encoder->devices & (ATOM_DEVICE_TV_SUPPORT | ATOM_DEVICE_CV_SUPPORT)) {
+			if (radeon_encoder->active_device & (ATOM_DEVICE_TV_SUPPORT | ATOM_DEVICE_CV_SUPPORT))
+				atombios_tv_setup(encoder, ATOM_ENABLE);
+			else
+				atombios_tv_setup(encoder, ATOM_DISABLE);
+		}
+		break;
+	}
+
+	if (ext_encoder) {
+		if (ASIC_IS_DCE41(rdev))
+			atombios_external_encoder_setup(encoder, ext_encoder,
+							EXTERNAL_ENCODER_ACTION_V3_ENCODER_SETUP);
+		else
+			atombios_external_encoder_setup(encoder, ext_encoder, ATOM_ENABLE);
+	}
+
+	atombios_apply_encoder_quirks(encoder, adjusted_mode);
+
+	if (atombios_get_encoder_mode(encoder) == ATOM_ENCODER_MODE_HDMI) {
+		r600_hdmi_enable(encoder);
+		r600_hdmi_setmode(encoder, adjusted_mode);
+	}
+}
+
+static bool
+atombios_dac_load_detect(struct drm_encoder *encoder, struct drm_connector *connector)
+{
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
+
+	if (radeon_encoder->devices & (ATOM_DEVICE_TV_SUPPORT |
+				       ATOM_DEVICE_CV_SUPPORT |
+				       ATOM_DEVICE_CRT_SUPPORT)) {
+		DAC_LOAD_DETECTION_PS_ALLOCATION args;
+		int index = GetIndexIntoMasterTable(COMMAND, DAC_LoadDetection);
+		uint8_t frev, crev;
+
+		memset(&args, 0, sizeof(args));
+
+		if (!atom_parse_cmd_header(rdev->mode_info.atom_context, index, &frev, &crev))
+			return false;
+
+		args.sDacload.ucMisc = 0;
+
+		if ((radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_DAC1) ||
+		    (radeon_encoder->encoder_id == ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1))
+			args.sDacload.ucDacType = ATOM_DAC_A;
+		else
+			args.sDacload.ucDacType = ATOM_DAC_B;
+
+		if (radeon_connector->devices & ATOM_DEVICE_CRT1_SUPPORT)
+			args.sDacload.usDeviceID = cpu_to_le16(ATOM_DEVICE_CRT1_SUPPORT);
+		else if (radeon_connector->devices & ATOM_DEVICE_CRT2_SUPPORT)
+			args.sDacload.usDeviceID = cpu_to_le16(ATOM_DEVICE_CRT2_SUPPORT);
+		else if (radeon_connector->devices & ATOM_DEVICE_CV_SUPPORT) {
+			args.sDacload.usDeviceID = cpu_to_le16(ATOM_DEVICE_CV_SUPPORT);
+			if (crev >= 3)
+				args.sDacload.ucMisc = DAC_LOAD_MISC_YPrPb;
+		} else if (radeon_connector->devices & ATOM_DEVICE_TV1_SUPPORT) {
+			args.sDacload.usDeviceID = cpu_to_le16(ATOM_DEVICE_TV1_SUPPORT);
+			if (crev >= 3)
+				args.sDacload.ucMisc = DAC_LOAD_MISC_YPrPb;
+		}
+
+		atom_execute_table(rdev->mode_info.atom_context, index, (uint32_t *)&args);
+
+		return true;
+	} else
 		return false;
+}
+
+static enum drm_connector_status
+radeon_atom_dac_detect(struct drm_encoder *encoder, struct drm_connector *connector)
+{
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
+	uint32_t bios_0_scratch;
+
+	if (!atombios_dac_load_detect(encoder, connector)) {
+		DRM_DEBUG_KMS("detect returned false \n");
+		return connector_status_unknown;
+	}
+
+	if (rdev->family >= CHIP_R600)
+		bios_0_scratch = RREG32(R600_BIOS_0_SCRATCH);
+	else
+		bios_0_scratch = RREG32(RADEON_BIOS_0_SCRATCH);
+
+	DRM_DEBUG_KMS("Bios 0 scratch %x %08x\n", bios_0_scratch, radeon_encoder->devices);
+	if (radeon_connector->devices & ATOM_DEVICE_CRT1_SUPPORT) {
+		if (bios_0_scratch & ATOM_S0_CRT1_MASK)
+			return connector_status_connected;
+	}
+	if (radeon_connector->devices & ATOM_DEVICE_CRT2_SUPPORT) {
+		if (bios_0_scratch & ATOM_S0_CRT2_MASK)
+			return connector_status_connected;
+	}
+	if (radeon_connector->devices & ATOM_DEVICE_CV_SUPPORT) {
+		if (bios_0_scratch & (ATOM_S0_CV_MASK|ATOM_S0_CV_MASK_A))
+			return connector_status_connected;
+	}
+	if (radeon_connector->devices & ATOM_DEVICE_TV1_SUPPORT) {
+		if (bios_0_scratch & (ATOM_S0_TV1_COMPOSITE | ATOM_S0_TV1_COMPOSITE_A))
+			return connector_status_connected; /* CTV */
+		else if (bios_0_scratch & (ATOM_S0_TV1_SVIDEO | ATOM_S0_TV1_SVIDEO_A))
+			return connector_status_connected; /* STV */
+	}
+	return connector_status_disconnected;
+}
+
+static enum drm_connector_status
+radeon_atom_dig_detect(struct drm_encoder *encoder, struct drm_connector *connector)
+{
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	struct radeon_connector *radeon_connector = to_radeon_connector(connector);
+	struct drm_encoder *ext_encoder = radeon_atom_get_external_encoder(encoder);
+	u32 bios_0_scratch;
+
+	if (!ASIC_IS_DCE4(rdev))
+		return connector_status_unknown;
+
+	if (!ext_encoder)
+		return connector_status_unknown;
+
+	if ((radeon_connector->devices & ATOM_DEVICE_CRT_SUPPORT) == 0)
+		return connector_status_unknown;
+
+	/* load detect on the dp bridge */
+	atombios_external_encoder_setup(encoder, ext_encoder,
+					EXTERNAL_ENCODER_ACTION_V3_DACLOAD_DETECTION);
+
+	bios_0_scratch = RREG32(R600_BIOS_0_SCRATCH);
+
+	DRM_DEBUG_KMS("Bios 0 scratch %x %08x\n", bios_0_scratch, radeon_encoder->devices);
+	if (radeon_connector->devices & ATOM_DEVICE_CRT1_SUPPORT) {
+		if (bios_0_scratch & ATOM_S0_CRT1_MASK)
+			return connector_status_connected;
+	}
+	if (radeon_connector->devices & ATOM_DEVICE_CRT2_SUPPORT) {
+		if (bios_0_scratch & ATOM_S0_CRT2_MASK)
+			return connector_status_connected;
 	}
+	if (radeon_connector->devices & ATOM_DEVICE_CV_SUPPORT) {
+		if (bios_0_scratch & (ATOM_S0_CV_MASK|ATOM_S0_CV_MASK_A))
+			return connector_status_connected;
+	}
+	if (radeon_connector->devices & ATOM_DEVICE_TV1_SUPPORT) {
+		if (bios_0_scratch & (ATOM_S0_TV1_COMPOSITE | ATOM_S0_TV1_COMPOSITE_A))
+			return connector_status_connected; /* CTV */
+		else if (bios_0_scratch & (ATOM_S0_TV1_SVIDEO | ATOM_S0_TV1_SVIDEO_A))
+			return connector_status_connected; /* STV */
+	}
+	return connector_status_disconnected;
 }
 
+void
+radeon_atom_ext_encoder_setup_ddc(struct drm_encoder *encoder)
+{
+	struct drm_encoder *ext_encoder = radeon_atom_get_external_encoder(encoder);
+
+	if (ext_encoder)
+		/* ddc_setup on the dp bridge */
+		atombios_external_encoder_setup(encoder, ext_encoder,
+						EXTERNAL_ENCODER_ACTION_V3_DDC_SETUP);
+
+}
+
+static void radeon_atom_encoder_prepare(struct drm_encoder *encoder)
+{
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	struct drm_connector *connector = radeon_get_connector_for_encoder(encoder);
+
+	if ((radeon_encoder->active_device &
+	     (ATOM_DEVICE_DFP_SUPPORT | ATOM_DEVICE_LCD_SUPPORT)) ||
+	    radeon_encoder_is_dp_bridge(encoder)) {
+		struct radeon_encoder_atom_dig *dig = radeon_encoder->enc_priv;
+		if (dig)
+			dig->dig_encoder = radeon_atom_pick_dig_encoder(encoder);
+	}
+
+	radeon_atom_output_lock(encoder, true);
+	radeon_atom_encoder_dpms(encoder, DRM_MODE_DPMS_OFF);
+
+	if (connector) {
+		struct radeon_connector *radeon_connector = to_radeon_connector(connector);
+
+		/* select the clock/data port if it uses a router */
+		if (radeon_connector->router.cd_valid)
+			radeon_router_select_cd_port(radeon_connector);
+
+		/* turn eDP panel on for mode set */
+		if (connector->connector_type == DRM_MODE_CONNECTOR_eDP)
+			atombios_set_edp_panel_power(connector,
+						     ATOM_TRANSMITTER_ACTION_POWER_ON);
+	}
+
+	/* this is needed for the pll/ss setup to work correctly in some cases */
+	atombios_set_encoder_crtc_source(encoder);
+}
+
+static void radeon_atom_encoder_commit(struct drm_encoder *encoder)
+{
+	radeon_atom_encoder_dpms(encoder, DRM_MODE_DPMS_ON);
+	radeon_atom_output_lock(encoder, false);
+}
+
+static void radeon_atom_encoder_disable(struct drm_encoder *encoder)
+{
+	struct drm_device *dev = encoder->dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	struct radeon_encoder_atom_dig *dig;
+
+	/* check for pre-DCE3 cards with shared encoders;
+	 * can't really use the links individually, so don't disable
+	 * the encoder if it's in use by another connector
+	 */
+	if (!ASIC_IS_DCE3(rdev)) {
+		struct drm_encoder *other_encoder;
+		struct radeon_encoder *other_radeon_encoder;
+
+		list_for_each_entry(other_encoder, &dev->mode_config.encoder_list, head) {
+			other_radeon_encoder = to_radeon_encoder(other_encoder);
+			if ((radeon_encoder->encoder_id == other_radeon_encoder->encoder_id) &&
+			    drm_helper_encoder_in_use(other_encoder))
+				goto disable_done;
+		}
+	}
+
+	radeon_atom_encoder_dpms(encoder, DRM_MODE_DPMS_OFF);
+
+	switch (radeon_encoder->encoder_id) {
+	case ENCODER_OBJECT_ID_INTERNAL_TMDS1:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
+	case ENCODER_OBJECT_ID_INTERNAL_LVDS:
+	case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
+		atombios_digital_setup(encoder, PANEL_ENCODER_ACTION_DISABLE);
+		break;
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
+		if (ASIC_IS_DCE4(rdev))
+			/* disable the transmitter */
+			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_DISABLE, 0, 0);
+		else {
+			/* disable the encoder and transmitter */
+			atombios_dig_transmitter_setup(encoder, ATOM_TRANSMITTER_ACTION_DISABLE, 0, 0);
+			atombios_dig_encoder_setup(encoder, ATOM_DISABLE, 0);
+		}
+		break;
+	case ENCODER_OBJECT_ID_INTERNAL_DDI:
+	case ENCODER_OBJECT_ID_INTERNAL_DVO1:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
+		atombios_dvo_setup(encoder, ATOM_DISABLE);
+		break;
+	case ENCODER_OBJECT_ID_INTERNAL_DAC1:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1:
+	case ENCODER_OBJECT_ID_INTERNAL_DAC2:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC2:
+		atombios_dac_setup(encoder, ATOM_DISABLE);
+		if (radeon_encoder->devices & (ATOM_DEVICE_TV_SUPPORT | ATOM_DEVICE_CV_SUPPORT))
+			atombios_tv_setup(encoder, ATOM_DISABLE);
+		break;
+	}
+
+disable_done:
+	if (radeon_encoder_is_digital(encoder)) {
+		if (atombios_get_encoder_mode(encoder) == ATOM_ENCODER_MODE_HDMI)
+			r600_hdmi_disable(encoder);
+		dig = radeon_encoder->enc_priv;
+		dig->dig_encoder = -1;
+	}
+	radeon_encoder->active_device = 0;
+}
+
+/* these are handled by the primary encoders */
+static void radeon_atom_ext_prepare(struct drm_encoder *encoder)
+{
+
+}
+
+static void radeon_atom_ext_commit(struct drm_encoder *encoder)
+{
+
+}
+
+static void
+radeon_atom_ext_mode_set(struct drm_encoder *encoder,
+			 struct drm_display_mode *mode,
+			 struct drm_display_mode *adjusted_mode)
+{
+
+}
+
+static void radeon_atom_ext_disable(struct drm_encoder *encoder)
+{
+
+}
+
+static void
+radeon_atom_ext_dpms(struct drm_encoder *encoder, int mode)
+{
+
+}
+
+static bool radeon_atom_ext_mode_fixup(struct drm_encoder *encoder,
+				       struct drm_display_mode *mode,
+				       struct drm_display_mode *adjusted_mode)
+{
+	return true;
+}
+
+static const struct drm_encoder_helper_funcs radeon_atom_ext_helper_funcs = {
+	.dpms = radeon_atom_ext_dpms,
+	.mode_fixup = radeon_atom_ext_mode_fixup,
+	.prepare = radeon_atom_ext_prepare,
+	.mode_set = radeon_atom_ext_mode_set,
+	.commit = radeon_atom_ext_commit,
+	.disable = radeon_atom_ext_disable,
+	/* no detect for TMDS/LVDS yet */
+};
+
+static const struct drm_encoder_helper_funcs radeon_atom_dig_helper_funcs = {
+	.dpms = radeon_atom_encoder_dpms,
+	.mode_fixup = radeon_atom_mode_fixup,
+	.prepare = radeon_atom_encoder_prepare,
+	.mode_set = radeon_atom_encoder_mode_set,
+	.commit = radeon_atom_encoder_commit,
+	.disable = radeon_atom_encoder_disable,
+	.detect = radeon_atom_dig_detect,
+};
+
+static const struct drm_encoder_helper_funcs radeon_atom_dac_helper_funcs = {
+	.dpms = radeon_atom_encoder_dpms,
+	.mode_fixup = radeon_atom_mode_fixup,
+	.prepare = radeon_atom_encoder_prepare,
+	.mode_set = radeon_atom_encoder_mode_set,
+	.commit = radeon_atom_encoder_commit,
+	.detect = radeon_atom_dac_detect,
+};
+
+void radeon_enc_destroy(struct drm_encoder *encoder)
+{
+	struct radeon_encoder *radeon_encoder = to_radeon_encoder(encoder);
+	kfree(radeon_encoder->enc_priv);
+	drm_encoder_cleanup(encoder);
+	kfree(radeon_encoder);
+}
+
+static const struct drm_encoder_funcs radeon_atom_enc_funcs = {
+	.destroy = radeon_enc_destroy,
+};
+
+struct radeon_encoder_atom_dac *
+radeon_atombios_set_dac_info(struct radeon_encoder *radeon_encoder)
+{
+	struct drm_device *dev = radeon_encoder->base.dev;
+	struct radeon_device *rdev = dev->dev_private;
+	struct radeon_encoder_atom_dac *dac = kzalloc(sizeof(struct radeon_encoder_atom_dac), GFP_KERNEL);
+
+	if (!dac)
+		return NULL;
+
+	dac->tv_std = radeon_atombios_get_tv_info(rdev);
+	return dac;
+}
+
+struct radeon_encoder_atom_dig *
+radeon_atombios_set_dig_info(struct radeon_encoder *radeon_encoder)
+{
+	int encoder_enum = (radeon_encoder->encoder_enum & ENUM_ID_MASK) >> ENUM_ID_SHIFT;
+	struct radeon_encoder_atom_dig *dig = kzalloc(sizeof(struct radeon_encoder_atom_dig), GFP_KERNEL);
+
+	if (!dig)
+		return NULL;
+
+	/* coherent mode by default */
+	dig->coherent_mode = true;
+	dig->dig_encoder = -1;
+
+	if (encoder_enum == 2)
+		dig->linkb = true;
+	else
+		dig->linkb = false;
+
+	return dig;
+}
+
+void
+radeon_add_atom_encoder(struct drm_device *dev,
+			uint32_t encoder_enum,
+			uint32_t supported_device,
+			u16 caps)
+{
+	struct radeon_device *rdev = dev->dev_private;
+	struct drm_encoder *encoder;
+	struct radeon_encoder *radeon_encoder;
+
+	/* see if we already added it */
+	list_for_each_entry(encoder, &dev->mode_config.encoder_list, head) {
+		radeon_encoder = to_radeon_encoder(encoder);
+		if (radeon_encoder->encoder_enum == encoder_enum) {
+			radeon_encoder->devices |= supported_device;
+			return;
+		}
+
+	}
+
+	/* add a new one */
+	radeon_encoder = kzalloc(sizeof(struct radeon_encoder), GFP_KERNEL);
+	if (!radeon_encoder)
+		return;
+
+	encoder = &radeon_encoder->base;
+	switch (rdev->num_crtc) {
+	case 1:
+		encoder->possible_crtcs = 0x1;
+		break;
+	case 2:
+	default:
+		encoder->possible_crtcs = 0x3;
+		break;
+	case 4:
+		encoder->possible_crtcs = 0xf;
+		break;
+	case 6:
+		encoder->possible_crtcs = 0x3f;
+		break;
+	}
+
+	radeon_encoder->enc_priv = NULL;
+
+	radeon_encoder->encoder_enum = encoder_enum;
+	radeon_encoder->encoder_id = (encoder_enum & OBJECT_ID_MASK) >> OBJECT_ID_SHIFT;
+	radeon_encoder->devices = supported_device;
+	radeon_encoder->rmx_type = RMX_OFF;
+	radeon_encoder->underscan_type = UNDERSCAN_OFF;
+	radeon_encoder->is_ext_encoder = false;
+	radeon_encoder->caps = caps;
+
+	switch (radeon_encoder->encoder_id) {
+	case ENCODER_OBJECT_ID_INTERNAL_LVDS:
+	case ENCODER_OBJECT_ID_INTERNAL_TMDS1:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_TMDS1:
+	case ENCODER_OBJECT_ID_INTERNAL_LVTM1:
+		if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) {
+			radeon_encoder->rmx_type = RMX_FULL;
+			drm_encoder_init(dev, encoder, &radeon_atom_enc_funcs, DRM_MODE_ENCODER_LVDS);
+			radeon_encoder->enc_priv = radeon_atombios_get_lvds_info(radeon_encoder);
+		} else {
+			drm_encoder_init(dev, encoder, &radeon_atom_enc_funcs, DRM_MODE_ENCODER_TMDS);
+			radeon_encoder->enc_priv = radeon_atombios_set_dig_info(radeon_encoder);
+		}
+		drm_encoder_helper_add(encoder, &radeon_atom_dig_helper_funcs);
+		break;
+	case ENCODER_OBJECT_ID_INTERNAL_DAC1:
+		drm_encoder_init(dev, encoder, &radeon_atom_enc_funcs, DRM_MODE_ENCODER_DAC);
+		radeon_encoder->enc_priv = radeon_atombios_set_dac_info(radeon_encoder);
+		drm_encoder_helper_add(encoder, &radeon_atom_dac_helper_funcs);
+		break;
+	case ENCODER_OBJECT_ID_INTERNAL_DAC2:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC1:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DAC2:
+		drm_encoder_init(dev, encoder, &radeon_atom_enc_funcs, DRM_MODE_ENCODER_TVDAC);
+		radeon_encoder->enc_priv = radeon_atombios_set_dac_info(radeon_encoder);
+		drm_encoder_helper_add(encoder, &radeon_atom_dac_helper_funcs);
+		break;
+	case ENCODER_OBJECT_ID_INTERNAL_DVO1:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_DVO1:
+	case ENCODER_OBJECT_ID_INTERNAL_DDI:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY:
+	case ENCODER_OBJECT_ID_INTERNAL_KLDSCP_LVTMA:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY1:
+	case ENCODER_OBJECT_ID_INTERNAL_UNIPHY2:
+		if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT)) {
+			radeon_encoder->rmx_type = RMX_FULL;
+			drm_encoder_init(dev, encoder, &radeon_atom_enc_funcs, DRM_MODE_ENCODER_LVDS);
+			radeon_encoder->enc_priv = radeon_atombios_get_lvds_info(radeon_encoder);
+		} else if (radeon_encoder->devices & (ATOM_DEVICE_CRT_SUPPORT)) {
+			drm_encoder_init(dev, encoder, &radeon_atom_enc_funcs, DRM_MODE_ENCODER_DAC);
+			radeon_encoder->enc_priv = radeon_atombios_set_dig_info(radeon_encoder);
+		} else {
+			drm_encoder_init(dev, encoder, &radeon_atom_enc_funcs, DRM_MODE_ENCODER_TMDS);
+			radeon_encoder->enc_priv = radeon_atombios_set_dig_info(radeon_encoder);
+		}
+		drm_encoder_helper_add(encoder, &radeon_atom_dig_helper_funcs);
+		break;
+	case ENCODER_OBJECT_ID_SI170B:
+	case ENCODER_OBJECT_ID_CH7303:
+	case ENCODER_OBJECT_ID_EXTERNAL_SDVOA:
+	case ENCODER_OBJECT_ID_EXTERNAL_SDVOB:
+	case ENCODER_OBJECT_ID_TITFP513:
+	case ENCODER_OBJECT_ID_VT1623:
+	case ENCODER_OBJECT_ID_HDMI_SI1930:
+	case ENCODER_OBJECT_ID_TRAVIS:
+	case ENCODER_OBJECT_ID_NUTMEG:
+		/* these are handled by the primary encoders */
+		radeon_encoder->is_ext_encoder = true;
+		if (radeon_encoder->devices & (ATOM_DEVICE_LCD_SUPPORT))
+			drm_encoder_init(dev, encoder, &radeon_atom_enc_funcs, DRM_MODE_ENCODER_LVDS);
+		else if (radeon_encoder->devices & (ATOM_DEVICE_CRT_SUPPORT))
+			drm_encoder_init(dev, encoder, &radeon_atom_enc_funcs, DRM_MODE_ENCODER_DAC);
+		else
+			drm_encoder_init(dev, encoder, &radeon_atom_enc_funcs, DRM_MODE_ENCODER_TMDS);
+		drm_encoder_helper_add(encoder, &radeon_atom_ext_helper_funcs);
+		break;
+	}
+}
diff --git a/drivers/gpu/drm/radeon/radeon_fb.c b/drivers/gpu/drm/radeon/radeon_fb.c
index 5906914..0b7b486 100644
--- a/drivers/gpu/drm/radeon/radeon_fb.c
+++ b/drivers/gpu/drm/radeon/radeon_fb.c
@@ -103,7 +103,7 @@ static void radeonfb_destroy_pinned_object(struct drm_gem_object *gobj)
 }
 
 static int radeonfb_create_pinned_object(struct radeon_fbdev *rfbdev,
-					 struct drm_mode_fb_cmd2 *mode_cmd,
+					 struct drm_mode_fb_cmd *mode_cmd,
 					 struct drm_gem_object **gobj_p)
 {
 	struct radeon_device *rdev = rfbdev->rdev;
@@ -114,17 +114,13 @@ static int radeonfb_create_pinned_object(struct radeon_fbdev *rfbdev,
 	int ret;
 	int aligned_size, size;
 	int height = mode_cmd->height;
-	u32 bpp, depth;
-
-	drm_fb_get_bpp_depth(mode_cmd->pixel_format, &depth, &bpp);
 
 	/* need to align pitch with crtc limits */
-	mode_cmd->pitches[0] = radeon_align_pitch(rdev, mode_cmd->width, bpp,
-						  fb_tiled) * ((bpp + 1) / 8);
+	mode_cmd->pitch = radeon_align_pitch(rdev, mode_cmd->width, mode_cmd->bpp, fb_tiled) * ((mode_cmd->bpp + 1) / 8);
 
 	if (rdev->family >= CHIP_R600)
 		height = ALIGN(mode_cmd->height, 8);
-	size = mode_cmd->pitches[0] * height;
+	size = mode_cmd->pitch * height;
 	aligned_size = ALIGN(size, PAGE_SIZE);
 	ret = radeon_gem_object_create(rdev, aligned_size, 0,
 				       RADEON_GEM_DOMAIN_VRAM,
@@ -141,7 +137,7 @@ static int radeonfb_create_pinned_object(struct radeon_fbdev *rfbdev,
 		tiling_flags = RADEON_TILING_MACRO;
 
 #ifdef __BIG_ENDIAN
-	switch (bpp) {
+	switch (mode_cmd->bpp) {
 	case 32:
 		tiling_flags |= RADEON_TILING_SWAP_32BIT;
 		break;
@@ -155,7 +151,7 @@ static int radeonfb_create_pinned_object(struct radeon_fbdev *rfbdev,
 	if (tiling_flags) {
 		ret = radeon_bo_set_tiling_flags(rbo,
 						 tiling_flags | RADEON_TILING_SURFACE,
-						 mode_cmd->pitches[0]);
+						 mode_cmd->pitch);
 		if (ret)
 			dev_err(rdev->dev, "FB failed to set tiling flags\n");
 	}
@@ -164,10 +160,7 @@ static int radeonfb_create_pinned_object(struct radeon_fbdev *rfbdev,
 	ret = radeon_bo_reserve(rbo, false);
 	if (unlikely(ret != 0))
 		goto out_unref;
-	/* Only 27 bit offset for legacy CRTC */
-	ret = radeon_bo_pin_restricted(rbo, RADEON_GEM_DOMAIN_VRAM,
-				       ASIC_IS_AVIVO(rdev) ? 0 : 1 << 27,
-				       NULL);
+	ret = radeon_bo_pin(rbo, RADEON_GEM_DOMAIN_VRAM, NULL);
 	if (ret) {
 		radeon_bo_unreserve(rbo);
 		goto out_unref;
@@ -194,7 +187,7 @@ static int radeonfb_create(struct radeon_fbdev *rfbdev,
 	struct radeon_device *rdev = rfbdev->rdev;
 	struct fb_info *info;
 	struct drm_framebuffer *fb = NULL;
-	struct drm_mode_fb_cmd2 mode_cmd;
+	struct drm_mode_fb_cmd mode_cmd;
 	struct drm_gem_object *gobj = NULL;
 	struct radeon_bo *rbo = NULL;
 	struct device *device = &rdev->pdev->dev;
@@ -208,15 +201,10 @@ static int radeonfb_create(struct radeon_fbdev *rfbdev,
 	if ((sizes->surface_bpp == 24) && ASIC_IS_AVIVO(rdev))
 		sizes->surface_bpp = 32;
 
-	mode_cmd.pixel_format = drm_mode_legacy_fb_format(sizes->surface_bpp,
-							  sizes->surface_depth);
+	mode_cmd.bpp = sizes->surface_bpp;
+	mode_cmd.depth = sizes->surface_depth;
 
 	ret = radeonfb_create_pinned_object(rfbdev, &mode_cmd, &gobj);
-	if (ret) {
-		DRM_ERROR("failed to create fbcon object %d\n", ret);
-		return ret;
-	}
-
 	rbo = gem_to_radeon_bo(gobj);
 
 	/* okay we have an object now allocate the framebuffer */
@@ -228,11 +216,7 @@ static int radeonfb_create(struct radeon_fbdev *rfbdev,
 
 	info->par = rfbdev;
 
-	ret = radeon_framebuffer_init(rdev->ddev, &rfbdev->rfb, &mode_cmd, gobj);
-	if (ret) {
-		DRM_ERROR("failed to initalise framebuffer %d\n", ret);
-		goto out_unref;
-	}
+	radeon_framebuffer_init(rdev->ddev, &rfbdev->rfb, &mode_cmd, gobj);
 
 	fb = &rfbdev->rfb.base;
 
@@ -244,7 +228,7 @@ static int radeonfb_create(struct radeon_fbdev *rfbdev,
 
 	strcpy(info->fix.id, "radeondrmfb");
 
-	drm_fb_helper_fill_fix(info, fb->pitches[0], fb->depth);
+	drm_fb_helper_fill_fix(info, fb->pitch, fb->depth);
 
 	info->flags = FBINFO_DEFAULT | FBINFO_CAN_FORCE_OUTPUT;
 	info->fbops = &radeonfb_ops;
@@ -266,7 +250,11 @@ static int radeonfb_create(struct radeon_fbdev *rfbdev,
 	info->apertures->ranges[0].base = rdev->ddev->mode_config.fb_base;
 	info->apertures->ranges[0].size = rdev->mc.aper_size;
 
-	/* Use default scratch pixmap (info->pixmap.flags = FB_PIXMAP_SYSTEM) */
+	info->pixmap.size = 64*1024;
+	info->pixmap.buf_align = 8;
+	info->pixmap.access_align = 32;
+	info->pixmap.flags = FB_PIXMAP_SYSTEM;
+	info->pixmap.scan_align = 1;
 
 	if (info->screen_base == NULL) {
 		ret = -ENOSPC;
@@ -283,7 +271,7 @@ static int radeonfb_create(struct radeon_fbdev *rfbdev,
 	DRM_INFO("vram apper at 0x%lX\n",  (unsigned long)rdev->mc.aper_base);
 	DRM_INFO("size %lu\n", (unsigned long)radeon_bo_size(rbo));
 	DRM_INFO("fb depth is %d\n", fb->depth);
-	DRM_INFO("   pitch is %d\n", fb->pitches[0]);
+	DRM_INFO("   pitch is %d\n", fb->pitch);
 
 	vga_switcheroo_client_fb_set(rdev->ddev->pdev, info);
 	return 0;
diff --git a/drivers/gpu/drm/radeon/radeon_fence.c b/drivers/gpu/drm/radeon/radeon_fence.c
index 4bd36a3..021d2b6 100644
--- a/drivers/gpu/drm/radeon/radeon_fence.c
+++ b/drivers/gpu/drm/radeon/radeon_fence.c
@@ -29,7 +29,7 @@
  *    Dave Airlie
  */
 #include <linux/seq_file.h>
-#include <linux/atomic.h>
+#include <asm/atomic.h>
 #include <linux/wait.h>
 #include <linux/list.h>
 #include <linux/kref.h>
@@ -40,24 +40,32 @@
 #include "radeon.h"
 #include "radeon_trace.h"
 
-static void radeon_fence_write(struct radeon_device *rdev, u32 seq, int ring)
+static void radeon_fence_write(struct radeon_device *rdev, u32 seq)
 {
 	if (rdev->wb.enabled) {
-		*rdev->fence_drv[ring].cpu_addr = cpu_to_le32(seq);
-	} else {
-		WREG32(rdev->fence_drv[ring].scratch_reg, seq);
-	}
+		u32 scratch_index;
+		if (rdev->wb.use_event)
+			scratch_index = R600_WB_EVENT_OFFSET + rdev->fence_drv.scratch_reg - rdev->scratch.reg_base;
+		else
+			scratch_index = RADEON_WB_SCRATCH_OFFSET + rdev->fence_drv.scratch_reg - rdev->scratch.reg_base;
+		rdev->wb.wb[scratch_index/4] = cpu_to_le32(seq);;
+	} else
+		WREG32(rdev->fence_drv.scratch_reg, seq);
 }
 
-static u32 radeon_fence_read(struct radeon_device *rdev, int ring)
+static u32 radeon_fence_read(struct radeon_device *rdev)
 {
-	u32 seq = 0;
+	u32 seq;
 
 	if (rdev->wb.enabled) {
-		seq = le32_to_cpu(*rdev->fence_drv[ring].cpu_addr);
-	} else {
-		seq = RREG32(rdev->fence_drv[ring].scratch_reg);
-	}
+		u32 scratch_index;
+		if (rdev->wb.use_event)
+			scratch_index = R600_WB_EVENT_OFFSET + rdev->fence_drv.scratch_reg - rdev->scratch.reg_base;
+		else
+			scratch_index = RADEON_WB_SCRATCH_OFFSET + rdev->fence_drv.scratch_reg - rdev->scratch.reg_base;
+		seq = le32_to_cpu(rdev->wb.wb[scratch_index/4]);
+	} else
+		seq = RREG32(rdev->fence_drv.scratch_reg);
 	return seq;
 }
 
@@ -65,28 +73,28 @@ int radeon_fence_emit(struct radeon_device *rdev, struct radeon_fence *fence)
 {
 	unsigned long irq_flags;
 
-	write_lock_irqsave(&rdev->fence_lock, irq_flags);
-	if (fence->emitted) {
-		write_unlock_irqrestore(&rdev->fence_lock, irq_flags);
+	write_lock_irqsave(&rdev->fence_drv.lock, irq_flags);
+	if (fence->emited) {
+		write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags);
 		return 0;
 	}
-	fence->seq = atomic_add_return(1, &rdev->fence_drv[fence->ring].seq);
-	if (!rdev->ring[fence->ring].ready)
+	fence->seq = atomic_add_return(1, &rdev->fence_drv.seq);
+	if (!rdev->cp.ready)
 		/* FIXME: cp is not running assume everythings is done right
 		 * away
 		 */
-		radeon_fence_write(rdev, fence->seq, fence->ring);
+		radeon_fence_write(rdev, fence->seq);
 	else
-		radeon_fence_ring_emit(rdev, fence->ring, fence);
+		radeon_fence_ring_emit(rdev, fence);
 
 	trace_radeon_fence_emit(rdev->ddev, fence->seq);
-	fence->emitted = true;
-	list_move_tail(&fence->list, &rdev->fence_drv[fence->ring].emitted);
-	write_unlock_irqrestore(&rdev->fence_lock, irq_flags);
+	fence->emited = true;
+	list_move_tail(&fence->list, &rdev->fence_drv.emited);
+	write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags);
 	return 0;
 }
 
-static bool radeon_fence_poll_locked(struct radeon_device *rdev, int ring)
+static bool radeon_fence_poll_locked(struct radeon_device *rdev)
 {
 	struct radeon_fence *fence;
 	struct list_head *i, *n;
@@ -94,34 +102,34 @@ static bool radeon_fence_poll_locked(struct radeon_device *rdev, int ring)
 	bool wake = false;
 	unsigned long cjiffies;
 
-	seq = radeon_fence_read(rdev, ring);
-	if (seq != rdev->fence_drv[ring].last_seq) {
-		rdev->fence_drv[ring].last_seq = seq;
-		rdev->fence_drv[ring].last_jiffies = jiffies;
-		rdev->fence_drv[ring].last_timeout = RADEON_FENCE_JIFFIES_TIMEOUT;
+	seq = radeon_fence_read(rdev);
+	if (seq != rdev->fence_drv.last_seq) {
+		rdev->fence_drv.last_seq = seq;
+		rdev->fence_drv.last_jiffies = jiffies;
+		rdev->fence_drv.last_timeout = RADEON_FENCE_JIFFIES_TIMEOUT;
 	} else {
 		cjiffies = jiffies;
-		if (time_after(cjiffies, rdev->fence_drv[ring].last_jiffies)) {
-			cjiffies -= rdev->fence_drv[ring].last_jiffies;
-			if (time_after(rdev->fence_drv[ring].last_timeout, cjiffies)) {
+		if (time_after(cjiffies, rdev->fence_drv.last_jiffies)) {
+			cjiffies -= rdev->fence_drv.last_jiffies;
+			if (time_after(rdev->fence_drv.last_timeout, cjiffies)) {
 				/* update the timeout */
-				rdev->fence_drv[ring].last_timeout -= cjiffies;
+				rdev->fence_drv.last_timeout -= cjiffies;
 			} else {
 				/* the 500ms timeout is elapsed we should test
 				 * for GPU lockup
 				 */
-				rdev->fence_drv[ring].last_timeout = 1;
+				rdev->fence_drv.last_timeout = 1;
 			}
 		} else {
 			/* wrap around update last jiffies, we will just wait
 			 * a little longer
 			 */
-			rdev->fence_drv[ring].last_jiffies = cjiffies;
+			rdev->fence_drv.last_jiffies = cjiffies;
 		}
 		return false;
 	}
 	n = NULL;
-	list_for_each(i, &rdev->fence_drv[ring].emitted) {
+	list_for_each(i, &rdev->fence_drv.emited) {
 		fence = list_entry(i, struct radeon_fence, list);
 		if (fence->seq == seq) {
 			n = i;
@@ -133,11 +141,11 @@ static bool radeon_fence_poll_locked(struct radeon_device *rdev, int ring)
 		i = n;
 		do {
 			n = i->prev;
-			list_move_tail(i, &rdev->fence_drv[ring].signaled);
+			list_move_tail(i, &rdev->fence_drv.signaled);
 			fence = list_entry(i, struct radeon_fence, list);
 			fence->signaled = true;
 			i = n;
-		} while (i != &rdev->fence_drv[ring].emitted);
+		} while (i != &rdev->fence_drv.emited);
 		wake = true;
 	}
 	return wake;
@@ -149,18 +157,14 @@ static void radeon_fence_destroy(struct kref *kref)
         struct radeon_fence *fence;
 
 	fence = container_of(kref, struct radeon_fence, kref);
-	write_lock_irqsave(&fence->rdev->fence_lock, irq_flags);
+	write_lock_irqsave(&fence->rdev->fence_drv.lock, irq_flags);
 	list_del(&fence->list);
-	fence->emitted = false;
-	write_unlock_irqrestore(&fence->rdev->fence_lock, irq_flags);
-	if (fence->semaphore)
-		radeon_semaphore_free(fence->rdev, fence->semaphore);
+	fence->emited = false;
+	write_unlock_irqrestore(&fence->rdev->fence_drv.lock, irq_flags);
 	kfree(fence);
 }
 
-int radeon_fence_create(struct radeon_device *rdev,
-			struct radeon_fence **fence,
-			int ring)
+int radeon_fence_create(struct radeon_device *rdev, struct radeon_fence **fence)
 {
 	unsigned long irq_flags;
 
@@ -170,19 +174,18 @@ int radeon_fence_create(struct radeon_device *rdev,
 	}
 	kref_init(&((*fence)->kref));
 	(*fence)->rdev = rdev;
-	(*fence)->emitted = false;
+	(*fence)->emited = false;
 	(*fence)->signaled = false;
 	(*fence)->seq = 0;
-	(*fence)->ring = ring;
-	(*fence)->semaphore = NULL;
 	INIT_LIST_HEAD(&(*fence)->list);
 
-	write_lock_irqsave(&rdev->fence_lock, irq_flags);
-	list_add_tail(&(*fence)->list, &rdev->fence_drv[ring].created);
-	write_unlock_irqrestore(&rdev->fence_lock, irq_flags);
+	write_lock_irqsave(&rdev->fence_drv.lock, irq_flags);
+	list_add_tail(&(*fence)->list, &rdev->fence_drv.created);
+	write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags);
 	return 0;
 }
 
+
 bool radeon_fence_signaled(struct radeon_fence *fence)
 {
 	unsigned long irq_flags;
@@ -194,21 +197,21 @@ bool radeon_fence_signaled(struct radeon_fence *fence)
 	if (fence->rdev->gpu_lockup)
 		return true;
 
-	write_lock_irqsave(&fence->rdev->fence_lock, irq_flags);
+	write_lock_irqsave(&fence->rdev->fence_drv.lock, irq_flags);
 	signaled = fence->signaled;
 	/* if we are shuting down report all fence as signaled */
 	if (fence->rdev->shutdown) {
 		signaled = true;
 	}
-	if (!fence->emitted) {
-		WARN(1, "Querying an unemitted fence : %p !\n", fence);
+	if (!fence->emited) {
+		WARN(1, "Querying an unemited fence : %p !\n", fence);
 		signaled = true;
 	}
 	if (!signaled) {
-		radeon_fence_poll_locked(fence->rdev, fence->ring);
+		radeon_fence_poll_locked(fence->rdev);
 		signaled = fence->signaled;
 	}
-	write_unlock_irqrestore(&fence->rdev->fence_lock, irq_flags);
+	write_unlock_irqrestore(&fence->rdev->fence_drv.lock, irq_flags);
 	return signaled;
 }
 
@@ -227,24 +230,24 @@ int radeon_fence_wait(struct radeon_fence *fence, bool intr)
 	if (radeon_fence_signaled(fence)) {
 		return 0;
 	}
-	timeout = rdev->fence_drv[fence->ring].last_timeout;
+	timeout = rdev->fence_drv.last_timeout;
 retry:
 	/* save current sequence used to check for GPU lockup */
-	seq = rdev->fence_drv[fence->ring].last_seq;
+	seq = rdev->fence_drv.last_seq;
 	trace_radeon_fence_wait_begin(rdev->ddev, seq);
 	if (intr) {
-		radeon_irq_kms_sw_irq_get(rdev, fence->ring);
-		r = wait_event_interruptible_timeout(rdev->fence_drv[fence->ring].queue,
+		radeon_irq_kms_sw_irq_get(rdev);
+		r = wait_event_interruptible_timeout(rdev->fence_drv.queue,
 				radeon_fence_signaled(fence), timeout);
-		radeon_irq_kms_sw_irq_put(rdev, fence->ring);
+		radeon_irq_kms_sw_irq_put(rdev);
 		if (unlikely(r < 0)) {
 			return r;
 		}
 	} else {
-		radeon_irq_kms_sw_irq_get(rdev, fence->ring);
-		r = wait_event_timeout(rdev->fence_drv[fence->ring].queue,
+		radeon_irq_kms_sw_irq_get(rdev);
+		r = wait_event_timeout(rdev->fence_drv.queue,
 			 radeon_fence_signaled(fence), timeout);
-		radeon_irq_kms_sw_irq_put(rdev, fence->ring);
+		radeon_irq_kms_sw_irq_put(rdev);
 	}
 	trace_radeon_fence_wait_end(rdev->ddev, seq);
 	if (unlikely(!radeon_fence_signaled(fence))) {
@@ -255,13 +258,12 @@ retry:
 			timeout = r;
 			goto retry;
 		}
-		/* don't protect read access to rdev->fence_drv[t].last_seq
+		/* don't protect read access to rdev->fence_drv.last_seq
 		 * if we experiencing a lockup the value doesn't change
 		 */
-		if (seq == rdev->fence_drv[fence->ring].last_seq &&
-		    radeon_gpu_is_lockup(rdev, &rdev->ring[fence->ring])) {
+		if (seq == rdev->fence_drv.last_seq && radeon_gpu_is_lockup(rdev)) {
 			/* good news we believe it's a lockup */
-			printk(KERN_WARNING "GPU lockup (waiting for 0x%08X last fence id 0x%08X)\n",
+			WARN(1, "GPU lockup (waiting for 0x%08X last fence id 0x%08X)\n",
 			     fence->seq, seq);
 			/* FIXME: what should we do ? marking everyone
 			 * as signaled for now
@@ -270,20 +272,20 @@ retry:
 			r = radeon_gpu_reset(rdev);
 			if (r)
 				return r;
-			radeon_fence_write(rdev, fence->seq, fence->ring);
+			radeon_fence_write(rdev, fence->seq);
 			rdev->gpu_lockup = false;
 		}
 		timeout = RADEON_FENCE_JIFFIES_TIMEOUT;
-		write_lock_irqsave(&rdev->fence_lock, irq_flags);
-		rdev->fence_drv[fence->ring].last_timeout = RADEON_FENCE_JIFFIES_TIMEOUT;
-		rdev->fence_drv[fence->ring].last_jiffies = jiffies;
-		write_unlock_irqrestore(&rdev->fence_lock, irq_flags);
+		write_lock_irqsave(&rdev->fence_drv.lock, irq_flags);
+		rdev->fence_drv.last_timeout = RADEON_FENCE_JIFFIES_TIMEOUT;
+		rdev->fence_drv.last_jiffies = jiffies;
+		write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags);
 		goto retry;
 	}
 	return 0;
 }
 
-int radeon_fence_wait_next(struct radeon_device *rdev, int ring)
+int radeon_fence_wait_next(struct radeon_device *rdev)
 {
 	unsigned long irq_flags;
 	struct radeon_fence *fence;
@@ -292,21 +294,21 @@ int radeon_fence_wait_next(struct radeon_device *rdev, int ring)
 	if (rdev->gpu_lockup) {
 		return 0;
 	}
-	write_lock_irqsave(&rdev->fence_lock, irq_flags);
-	if (list_empty(&rdev->fence_drv[ring].emitted)) {
-		write_unlock_irqrestore(&rdev->fence_lock, irq_flags);
+	write_lock_irqsave(&rdev->fence_drv.lock, irq_flags);
+	if (list_empty(&rdev->fence_drv.emited)) {
+		write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags);
 		return 0;
 	}
-	fence = list_entry(rdev->fence_drv[ring].emitted.next,
+	fence = list_entry(rdev->fence_drv.emited.next,
 			   struct radeon_fence, list);
 	radeon_fence_ref(fence);
-	write_unlock_irqrestore(&rdev->fence_lock, irq_flags);
+	write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags);
 	r = radeon_fence_wait(fence, false);
 	radeon_fence_unref(&fence);
 	return r;
 }
 
-int radeon_fence_wait_last(struct radeon_device *rdev, int ring)
+int radeon_fence_wait_last(struct radeon_device *rdev)
 {
 	unsigned long irq_flags;
 	struct radeon_fence *fence;
@@ -315,15 +317,15 @@ int radeon_fence_wait_last(struct radeon_device *rdev, int ring)
 	if (rdev->gpu_lockup) {
 		return 0;
 	}
-	write_lock_irqsave(&rdev->fence_lock, irq_flags);
-	if (list_empty(&rdev->fence_drv[ring].emitted)) {
-		write_unlock_irqrestore(&rdev->fence_lock, irq_flags);
+	write_lock_irqsave(&rdev->fence_drv.lock, irq_flags);
+	if (list_empty(&rdev->fence_drv.emited)) {
+		write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags);
 		return 0;
 	}
-	fence = list_entry(rdev->fence_drv[ring].emitted.prev,
+	fence = list_entry(rdev->fence_drv.emited.prev,
 			   struct radeon_fence, list);
 	radeon_fence_ref(fence);
-	write_unlock_irqrestore(&rdev->fence_lock, irq_flags);
+	write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags);
 	r = radeon_fence_wait(fence, false);
 	radeon_fence_unref(&fence);
 	return r;
@@ -345,97 +347,39 @@ void radeon_fence_unref(struct radeon_fence **fence)
 	}
 }
 
-void radeon_fence_process(struct radeon_device *rdev, int ring)
+void radeon_fence_process(struct radeon_device *rdev)
 {
 	unsigned long irq_flags;
 	bool wake;
 
-	write_lock_irqsave(&rdev->fence_lock, irq_flags);
-	wake = radeon_fence_poll_locked(rdev, ring);
-	write_unlock_irqrestore(&rdev->fence_lock, irq_flags);
+	write_lock_irqsave(&rdev->fence_drv.lock, irq_flags);
+	wake = radeon_fence_poll_locked(rdev);
+	write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags);
 	if (wake) {
-		wake_up_all(&rdev->fence_drv[ring].queue);
-	}
-}
-
-int radeon_fence_count_emitted(struct radeon_device *rdev, int ring)
-{
-	unsigned long irq_flags;
-	int not_processed = 0;
-
-	read_lock_irqsave(&rdev->fence_lock, irq_flags);
-	if (!rdev->fence_drv[ring].initialized) {
-		read_unlock_irqrestore(&rdev->fence_lock, irq_flags);
-		return 0;
-	}
-
-	if (!list_empty(&rdev->fence_drv[ring].emitted)) {
-		struct list_head *ptr;
-		list_for_each(ptr, &rdev->fence_drv[ring].emitted) {
-			/* count up to 3, that's enought info */
-			if (++not_processed >= 3)
-				break;
-		}
-	}
-	read_unlock_irqrestore(&rdev->fence_lock, irq_flags);
-	return not_processed;
-}
-
-int radeon_fence_driver_start_ring(struct radeon_device *rdev, int ring)
-{
-	unsigned long irq_flags;
-	uint64_t index;
-	int r;
-
-	write_lock_irqsave(&rdev->fence_lock, irq_flags);
-	radeon_scratch_free(rdev, rdev->fence_drv[ring].scratch_reg);
-	if (rdev->wb.use_event) {
-		rdev->fence_drv[ring].scratch_reg = 0;
-		index = R600_WB_EVENT_OFFSET + ring * 4;
-	} else {
-		r = radeon_scratch_get(rdev, &rdev->fence_drv[ring].scratch_reg);
-		if (r) {
-			dev_err(rdev->dev, "fence failed to get scratch register\n");
-			write_unlock_irqrestore(&rdev->fence_lock, irq_flags);
-			return r;
-		}
-		index = RADEON_WB_SCRATCH_OFFSET +
-			rdev->fence_drv[ring].scratch_reg -
-			rdev->scratch.reg_base;
+		wake_up_all(&rdev->fence_drv.queue);
 	}
-	rdev->fence_drv[ring].cpu_addr = &rdev->wb.wb[index/4];
-	rdev->fence_drv[ring].gpu_addr = rdev->wb.gpu_addr + index;
-	radeon_fence_write(rdev, atomic_read(&rdev->fence_drv[ring].seq), ring);
-	rdev->fence_drv[ring].initialized = true;
-	DRM_INFO("fence driver on ring %d use gpu addr 0x%08Lx and cpu addr 0x%p\n",
-		 ring, rdev->fence_drv[ring].gpu_addr, rdev->fence_drv[ring].cpu_addr);
-	write_unlock_irqrestore(&rdev->fence_lock, irq_flags);
-	return 0;
-}
-
-static void radeon_fence_driver_init_ring(struct radeon_device *rdev, int ring)
-{
-	rdev->fence_drv[ring].scratch_reg = -1;
-	rdev->fence_drv[ring].cpu_addr = NULL;
-	rdev->fence_drv[ring].gpu_addr = 0;
-	atomic_set(&rdev->fence_drv[ring].seq, 0);
-	INIT_LIST_HEAD(&rdev->fence_drv[ring].created);
-	INIT_LIST_HEAD(&rdev->fence_drv[ring].emitted);
-	INIT_LIST_HEAD(&rdev->fence_drv[ring].signaled);
-	init_waitqueue_head(&rdev->fence_drv[ring].queue);
-	rdev->fence_drv[ring].initialized = false;
 }
 
 int radeon_fence_driver_init(struct radeon_device *rdev)
 {
 	unsigned long irq_flags;
-	int ring;
+	int r;
 
-	write_lock_irqsave(&rdev->fence_lock, irq_flags);
-	for (ring = 0; ring < RADEON_NUM_RINGS; ring++) {
-		radeon_fence_driver_init_ring(rdev, ring);
+	write_lock_irqsave(&rdev->fence_drv.lock, irq_flags);
+	r = radeon_scratch_get(rdev, &rdev->fence_drv.scratch_reg);
+	if (r) {
+		dev_err(rdev->dev, "fence failed to get scratch register\n");
+		write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags);
+		return r;
 	}
-	write_unlock_irqrestore(&rdev->fence_lock, irq_flags);
+	radeon_fence_write(rdev, 0);
+	atomic_set(&rdev->fence_drv.seq, 0);
+	INIT_LIST_HEAD(&rdev->fence_drv.created);
+	INIT_LIST_HEAD(&rdev->fence_drv.emited);
+	INIT_LIST_HEAD(&rdev->fence_drv.signaled);
+	init_waitqueue_head(&rdev->fence_drv.queue);
+	rdev->fence_drv.initialized = true;
+	write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags);
 	if (radeon_debugfs_fence_init(rdev)) {
 		dev_err(rdev->dev, "fence debugfs file creation failed\n");
 	}
@@ -445,18 +389,14 @@ int radeon_fence_driver_init(struct radeon_device *rdev)
 void radeon_fence_driver_fini(struct radeon_device *rdev)
 {
 	unsigned long irq_flags;
-	int ring;
-
-	for (ring = 0; ring < RADEON_NUM_RINGS; ring++) {
-		if (!rdev->fence_drv[ring].initialized)
-			continue;
-		radeon_fence_wait_last(rdev, ring);
-		wake_up_all(&rdev->fence_drv[ring].queue);
-		write_lock_irqsave(&rdev->fence_lock, irq_flags);
-		radeon_scratch_free(rdev, rdev->fence_drv[ring].scratch_reg);
-		write_unlock_irqrestore(&rdev->fence_lock, irq_flags);
-		rdev->fence_drv[ring].initialized = false;
-	}
+
+	if (!rdev->fence_drv.initialized)
+		return;
+	wake_up_all(&rdev->fence_drv.queue);
+	write_lock_irqsave(&rdev->fence_drv.lock, irq_flags);
+	radeon_scratch_free(rdev, rdev->fence_drv.scratch_reg);
+	write_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags);
+	rdev->fence_drv.initialized = false;
 }
 
 
@@ -470,21 +410,14 @@ static int radeon_debugfs_fence_info(struct seq_file *m, void *data)
 	struct drm_device *dev = node->minor->dev;
 	struct radeon_device *rdev = dev->dev_private;
 	struct radeon_fence *fence;
-	int i;
-
-	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
-		if (!rdev->fence_drv[i].initialized)
-			continue;
-
-		seq_printf(m, "--- ring %d ---\n", i);
-		seq_printf(m, "Last signaled fence 0x%08X\n",
-			   radeon_fence_read(rdev, i));
-		if (!list_empty(&rdev->fence_drv[i].emitted)) {
-			fence = list_entry(rdev->fence_drv[i].emitted.prev,
-					   struct radeon_fence, list);
-			seq_printf(m, "Last emitted fence %p with 0x%08X\n",
-				   fence,  fence->seq);
-		}
+
+	seq_printf(m, "Last signaled fence 0x%08X\n",
+		   radeon_fence_read(rdev));
+	if (!list_empty(&rdev->fence_drv.emited)) {
+		   fence = list_entry(rdev->fence_drv.emited.prev,
+				      struct radeon_fence, list);
+		   seq_printf(m, "Last emited fence %p with 0x%08X\n",
+			      fence,  fence->seq);
 	}
 	return 0;
 }
diff --git a/drivers/gpu/drm/radeon/radeon_gart.c b/drivers/gpu/drm/radeon/radeon_gart.c
index c58a036..a533f52 100644
--- a/drivers/gpu/drm/radeon/radeon_gart.c
+++ b/drivers/gpu/drm/radeon/radeon_gart.c
@@ -49,27 +49,27 @@ int radeon_gart_table_ram_alloc(struct radeon_device *rdev)
 			      rdev->gart.table_size >> PAGE_SHIFT);
 	}
 #endif
-	rdev->gart.ptr = ptr;
-	memset((void *)rdev->gart.ptr, 0, rdev->gart.table_size);
+	rdev->gart.table.ram.ptr = ptr;
+	memset((void *)rdev->gart.table.ram.ptr, 0, rdev->gart.table_size);
 	return 0;
 }
 
 void radeon_gart_table_ram_free(struct radeon_device *rdev)
 {
-	if (rdev->gart.ptr == NULL) {
+	if (rdev->gart.table.ram.ptr == NULL) {
 		return;
 	}
 #ifdef CONFIG_X86
 	if (rdev->family == CHIP_RS400 || rdev->family == CHIP_RS480 ||
 	    rdev->family == CHIP_RS690 || rdev->family == CHIP_RS740) {
-		set_memory_wb((unsigned long)rdev->gart.ptr,
+		set_memory_wb((unsigned long)rdev->gart.table.ram.ptr,
 			      rdev->gart.table_size >> PAGE_SHIFT);
 	}
 #endif
 	pci_free_consistent(rdev->pdev, rdev->gart.table_size,
-			    (void *)rdev->gart.ptr,
+			    (void *)rdev->gart.table.ram.ptr,
 			    rdev->gart.table_addr);
-	rdev->gart.ptr = NULL;
+	rdev->gart.table.ram.ptr = NULL;
 	rdev->gart.table_addr = 0;
 }
 
@@ -77,10 +77,10 @@ int radeon_gart_table_vram_alloc(struct radeon_device *rdev)
 {
 	int r;
 
-	if (rdev->gart.robj == NULL) {
+	if (rdev->gart.table.vram.robj == NULL) {
 		r = radeon_bo_create(rdev, rdev->gart.table_size,
 				     PAGE_SIZE, true, RADEON_GEM_DOMAIN_VRAM,
-				     &rdev->gart.robj);
+				     &rdev->gart.table.vram.robj);
 		if (r) {
 			return r;
 		}
@@ -93,46 +93,38 @@ int radeon_gart_table_vram_pin(struct radeon_device *rdev)
 	uint64_t gpu_addr;
 	int r;
 
-	r = radeon_bo_reserve(rdev->gart.robj, false);
+	r = radeon_bo_reserve(rdev->gart.table.vram.robj, false);
 	if (unlikely(r != 0))
 		return r;
-	r = radeon_bo_pin(rdev->gart.robj,
+	r = radeon_bo_pin(rdev->gart.table.vram.robj,
 				RADEON_GEM_DOMAIN_VRAM, &gpu_addr);
 	if (r) {
-		radeon_bo_unreserve(rdev->gart.robj);
+		radeon_bo_unreserve(rdev->gart.table.vram.robj);
 		return r;
 	}
-	r = radeon_bo_kmap(rdev->gart.robj, &rdev->gart.ptr);
+	r = radeon_bo_kmap(rdev->gart.table.vram.robj,
+				(void **)&rdev->gart.table.vram.ptr);
 	if (r)
-		radeon_bo_unpin(rdev->gart.robj);
-	radeon_bo_unreserve(rdev->gart.robj);
+		radeon_bo_unpin(rdev->gart.table.vram.robj);
+	radeon_bo_unreserve(rdev->gart.table.vram.robj);
 	rdev->gart.table_addr = gpu_addr;
 	return r;
 }
 
-void radeon_gart_table_vram_unpin(struct radeon_device *rdev)
+void radeon_gart_table_vram_free(struct radeon_device *rdev)
 {
 	int r;
 
-	if (rdev->gart.robj == NULL) {
+	if (rdev->gart.table.vram.robj == NULL) {
 		return;
 	}
-	r = radeon_bo_reserve(rdev->gart.robj, false);
+	r = radeon_bo_reserve(rdev->gart.table.vram.robj, false);
 	if (likely(r == 0)) {
-		radeon_bo_kunmap(rdev->gart.robj);
-		radeon_bo_unpin(rdev->gart.robj);
-		radeon_bo_unreserve(rdev->gart.robj);
-		rdev->gart.ptr = NULL;
-	}
-}
-
-void radeon_gart_table_vram_free(struct radeon_device *rdev)
-{
-	if (rdev->gart.robj == NULL) {
-		return;
+		radeon_bo_kunmap(rdev->gart.table.vram.robj);
+		radeon_bo_unpin(rdev->gart.table.vram.robj);
+		radeon_bo_unreserve(rdev->gart.table.vram.robj);
 	}
-	radeon_gart_table_vram_unpin(rdev);
-	radeon_bo_unref(&rdev->gart.robj);
+	radeon_bo_unref(&rdev->gart.table.vram.robj);
 }
 
 
@@ -150,20 +142,21 @@ void radeon_gart_unbind(struct radeon_device *rdev, unsigned offset,
 	u64 page_base;
 
 	if (!rdev->gart.ready) {
-		WARN(1, "trying to unbind memory from uninitialized GART !\n");
+		WARN(1, "trying to unbind memory to unitialized GART !\n");
 		return;
 	}
 	t = offset / RADEON_GPU_PAGE_SIZE;
 	p = t / (PAGE_SIZE / RADEON_GPU_PAGE_SIZE);
 	for (i = 0; i < pages; i++, p++) {
 		if (rdev->gart.pages[p]) {
+			if (!rdev->gart.ttm_alloced[p])
+				pci_unmap_page(rdev->pdev, rdev->gart.pages_addr[p],
+				       		PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
 			rdev->gart.pages[p] = NULL;
 			rdev->gart.pages_addr[p] = rdev->dummy_page.addr;
 			page_base = rdev->gart.pages_addr[p];
 			for (j = 0; j < (PAGE_SIZE / RADEON_GPU_PAGE_SIZE); j++, t++) {
-				if (rdev->gart.ptr) {
-					radeon_gart_set_page(rdev, t, page_base);
-				}
+				radeon_gart_set_page(rdev, t, page_base);
 				page_base += RADEON_GPU_PAGE_SIZE;
 			}
 		}
@@ -181,22 +174,36 @@ int radeon_gart_bind(struct radeon_device *rdev, unsigned offset,
 	int i, j;
 
 	if (!rdev->gart.ready) {
-		WARN(1, "trying to bind memory to uninitialized GART !\n");
+		WARN(1, "trying to bind memory to unitialized GART !\n");
 		return -EINVAL;
 	}
 	t = offset / RADEON_GPU_PAGE_SIZE;
 	p = t / (PAGE_SIZE / RADEON_GPU_PAGE_SIZE);
 
 	for (i = 0; i < pages; i++, p++) {
-		rdev->gart.pages_addr[p] = dma_addr[i];
-		rdev->gart.pages[p] = pagelist[i];
-		if (rdev->gart.ptr) {
-			page_base = rdev->gart.pages_addr[p];
-			for (j = 0; j < (PAGE_SIZE / RADEON_GPU_PAGE_SIZE); j++, t++) {
-				radeon_gart_set_page(rdev, t, page_base);
-				page_base += RADEON_GPU_PAGE_SIZE;
+		/* we reverted the patch using dma_addr in TTM for now but this
+		 * code stops building on alpha so just comment it out for now */
+		if (0) { /*dma_addr[i] != DMA_ERROR_CODE) */
+			rdev->gart.ttm_alloced[p] = true;
+			rdev->gart.pages_addr[p] = dma_addr[i];
+		} else {
+			/* we need to support large memory configurations */
+			/* assume that unbind have already been call on the range */
+			rdev->gart.pages_addr[p] = pci_map_page(rdev->pdev, pagelist[i],
+							0, PAGE_SIZE,
+							PCI_DMA_BIDIRECTIONAL);
+			if (pci_dma_mapping_error(rdev->pdev, rdev->gart.pages_addr[p])) {
+				/* FIXME: failed to map page (return -ENOMEM?) */
+				radeon_gart_unbind(rdev, offset, pages);
+				return -ENOMEM;
 			}
 		}
+		rdev->gart.pages[p] = pagelist[i];
+		page_base = rdev->gart.pages_addr[p];
+		for (j = 0; j < (PAGE_SIZE / RADEON_GPU_PAGE_SIZE); j++, t++) {
+			radeon_gart_set_page(rdev, t, page_base);
+			page_base += RADEON_GPU_PAGE_SIZE;
+		}
 	}
 	mb();
 	radeon_gart_tlb_flush(rdev);
@@ -208,9 +215,6 @@ void radeon_gart_restore(struct radeon_device *rdev)
 	int i, j, t;
 	u64 page_base;
 
-	if (!rdev->gart.ptr) {
-		return;
-	}
 	for (i = 0, t = 0; i < rdev->gart.num_cpu_pages; i++) {
 		page_base = rdev->gart.pages_addr[i];
 		for (j = 0; j < (PAGE_SIZE / RADEON_GPU_PAGE_SIZE); j++, t++) {
@@ -255,6 +259,12 @@ int radeon_gart_init(struct radeon_device *rdev)
 		radeon_gart_fini(rdev);
 		return -ENOMEM;
 	}
+	rdev->gart.ttm_alloced = kzalloc(sizeof(bool) *
+					 rdev->gart.num_cpu_pages, GFP_KERNEL);
+	if (rdev->gart.ttm_alloced == NULL) {
+		radeon_gart_fini(rdev);
+		return -ENOMEM;
+	}
 	/* set GART entry to point to the dummy page by default */
 	for (i = 0; i < rdev->gart.num_cpu_pages; i++) {
 		rdev->gart.pages_addr[i] = rdev->dummy_page.addr;
@@ -271,404 +281,10 @@ void radeon_gart_fini(struct radeon_device *rdev)
 	rdev->gart.ready = false;
 	kfree(rdev->gart.pages);
 	kfree(rdev->gart.pages_addr);
+	kfree(rdev->gart.ttm_alloced);
 	rdev->gart.pages = NULL;
 	rdev->gart.pages_addr = NULL;
+	rdev->gart.ttm_alloced = NULL;
 
 	radeon_dummy_page_fini(rdev);
 }
-
-/*
- * vm helpers
- *
- * TODO bind a default page at vm initialization for default address
- */
-int radeon_vm_manager_init(struct radeon_device *rdev)
-{
-	int r;
-
-	rdev->vm_manager.enabled = false;
-
-	/* mark first vm as always in use, it's the system one */
-	r = radeon_sa_bo_manager_init(rdev, &rdev->vm_manager.sa_manager,
-				      rdev->vm_manager.max_pfn * 8,
-				      RADEON_GEM_DOMAIN_VRAM);
-	if (r) {
-		dev_err(rdev->dev, "failed to allocate vm bo (%dKB)\n",
-			(rdev->vm_manager.max_pfn * 8) >> 10);
-		return r;
-	}
-
-	r = rdev->vm_manager.funcs->init(rdev);
-	if (r == 0)
-		rdev->vm_manager.enabled = true;
-
-	return r;
-}
-
-/* cs mutex must be lock */
-static void radeon_vm_unbind_locked(struct radeon_device *rdev,
-				    struct radeon_vm *vm)
-{
-	struct radeon_bo_va *bo_va;
-
-	if (vm->id == -1) {
-		return;
-	}
-
-	/* wait for vm use to end */
-	if (vm->fence) {
-		radeon_fence_wait(vm->fence, false);
-		radeon_fence_unref(&vm->fence);
-	}
-
-	/* hw unbind */
-	rdev->vm_manager.funcs->unbind(rdev, vm);
-	rdev->vm_manager.use_bitmap &= ~(1 << vm->id);
-	list_del_init(&vm->list);
-	vm->id = -1;
-	radeon_sa_bo_free(rdev, &vm->sa_bo);
-	vm->pt = NULL;
-
-	list_for_each_entry(bo_va, &vm->va, vm_list) {
-		bo_va->valid = false;
-	}
-}
-
-void radeon_vm_manager_fini(struct radeon_device *rdev)
-{
-	if (rdev->vm_manager.sa_manager.bo == NULL)
-		return;
-	radeon_vm_manager_suspend(rdev);
-	rdev->vm_manager.funcs->fini(rdev);
-	radeon_sa_bo_manager_fini(rdev, &rdev->vm_manager.sa_manager);
-	rdev->vm_manager.enabled = false;
-}
-
-int radeon_vm_manager_start(struct radeon_device *rdev)
-{
-	if (rdev->vm_manager.sa_manager.bo == NULL) {
-		return -EINVAL;
-	}
-	return radeon_sa_bo_manager_start(rdev, &rdev->vm_manager.sa_manager);
-}
-
-int radeon_vm_manager_suspend(struct radeon_device *rdev)
-{
-	struct radeon_vm *vm, *tmp;
-
-	radeon_mutex_lock(&rdev->cs_mutex);
-	/* unbind all active vm */
-	list_for_each_entry_safe(vm, tmp, &rdev->vm_manager.lru_vm, list) {
-		radeon_vm_unbind_locked(rdev, vm);
-	}
-	rdev->vm_manager.funcs->fini(rdev);
-	radeon_mutex_unlock(&rdev->cs_mutex);
-	return radeon_sa_bo_manager_suspend(rdev, &rdev->vm_manager.sa_manager);
-}
-
-/* cs mutex must be lock */
-void radeon_vm_unbind(struct radeon_device *rdev, struct radeon_vm *vm)
-{
-	mutex_lock(&vm->mutex);
-	radeon_vm_unbind_locked(rdev, vm);
-	mutex_unlock(&vm->mutex);
-}
-
-/* cs mutex must be lock & vm mutex must be lock */
-int radeon_vm_bind(struct radeon_device *rdev, struct radeon_vm *vm)
-{
-	struct radeon_vm *vm_evict;
-	unsigned i;
-	int id = -1, r;
-
-	if (vm == NULL) {
-		return -EINVAL;
-	}
-
-	if (vm->id != -1) {
-		/* update lru */
-		list_del_init(&vm->list);
-		list_add_tail(&vm->list, &rdev->vm_manager.lru_vm);
-		return 0;
-	}
-
-retry:
-	r = radeon_sa_bo_new(rdev, &rdev->vm_manager.sa_manager, &vm->sa_bo,
-			     RADEON_GPU_PAGE_ALIGN(vm->last_pfn * 8),
-			     RADEON_GPU_PAGE_SIZE);
-	if (r) {
-		if (list_empty(&rdev->vm_manager.lru_vm)) {
-			return r;
-		}
-		vm_evict = list_first_entry(&rdev->vm_manager.lru_vm, struct radeon_vm, list);
-		radeon_vm_unbind(rdev, vm_evict);
-		goto retry;
-	}
-	vm->pt = rdev->vm_manager.sa_manager.cpu_ptr;
-	vm->pt += (vm->sa_bo.offset >> 3);
-	vm->pt_gpu_addr = rdev->vm_manager.sa_manager.gpu_addr;
-	vm->pt_gpu_addr += vm->sa_bo.offset;
-	memset(vm->pt, 0, RADEON_GPU_PAGE_ALIGN(vm->last_pfn * 8));
-
-retry_id:
-	/* search for free vm */
-	for (i = 0; i < rdev->vm_manager.nvm; i++) {
-		if (!(rdev->vm_manager.use_bitmap & (1 << i))) {
-			id = i;
-			break;
-		}
-	}
-	/* evict vm if necessary */
-	if (id == -1) {
-		vm_evict = list_first_entry(&rdev->vm_manager.lru_vm, struct radeon_vm, list);
-		radeon_vm_unbind(rdev, vm_evict);
-		goto retry_id;
-	}
-
-	/* do hw bind */
-	r = rdev->vm_manager.funcs->bind(rdev, vm, id);
-	if (r) {
-		radeon_sa_bo_free(rdev, &vm->sa_bo);
-		return r;
-	}
-	rdev->vm_manager.use_bitmap |= 1 << id;
-	vm->id = id;
-	list_add_tail(&vm->list, &rdev->vm_manager.lru_vm);
-	return radeon_vm_bo_update_pte(rdev, vm, rdev->ib_pool.sa_manager.bo,
-				       &rdev->ib_pool.sa_manager.bo->tbo.mem);
-}
-
-/* object have to be reserved */
-int radeon_vm_bo_add(struct radeon_device *rdev,
-		     struct radeon_vm *vm,
-		     struct radeon_bo *bo,
-		     uint64_t offset,
-		     uint32_t flags)
-{
-	struct radeon_bo_va *bo_va, *tmp;
-	struct list_head *head;
-	uint64_t size = radeon_bo_size(bo), last_offset = 0;
-	unsigned last_pfn;
-
-	bo_va = kzalloc(sizeof(struct radeon_bo_va), GFP_KERNEL);
-	if (bo_va == NULL) {
-		return -ENOMEM;
-	}
-	bo_va->vm = vm;
-	bo_va->bo = bo;
-	bo_va->soffset = offset;
-	bo_va->eoffset = offset + size;
-	bo_va->flags = flags;
-	bo_va->valid = false;
-	INIT_LIST_HEAD(&bo_va->bo_list);
-	INIT_LIST_HEAD(&bo_va->vm_list);
-	/* make sure object fit at this offset */
-	if (bo_va->soffset >= bo_va->eoffset) {
-		kfree(bo_va);
-		return -EINVAL;
-	}
-
-	last_pfn = bo_va->eoffset / RADEON_GPU_PAGE_SIZE;
-	if (last_pfn > rdev->vm_manager.max_pfn) {
-		kfree(bo_va);
-		dev_err(rdev->dev, "va above limit (0x%08X > 0x%08X)\n",
-			last_pfn, rdev->vm_manager.max_pfn);
-		return -EINVAL;
-	}
-
-	mutex_lock(&vm->mutex);
-	if (last_pfn > vm->last_pfn) {
-		/* grow va space 32M by 32M */
-		unsigned align = ((32 << 20) >> 12) - 1;
-		radeon_mutex_lock(&rdev->cs_mutex);
-		radeon_vm_unbind_locked(rdev, vm);
-		radeon_mutex_unlock(&rdev->cs_mutex);
-		vm->last_pfn = (last_pfn + align) & ~align;
-	}
-	head = &vm->va;
-	last_offset = 0;
-	list_for_each_entry(tmp, &vm->va, vm_list) {
-		if (bo_va->soffset >= last_offset && bo_va->eoffset < tmp->soffset) {
-			/* bo can be added before this one */
-			break;
-		}
-		if (bo_va->soffset >= tmp->soffset && bo_va->soffset < tmp->eoffset) {
-			/* bo and tmp overlap, invalid offset */
-			dev_err(rdev->dev, "bo %p va 0x%08X conflict with (bo %p 0x%08X 0x%08X)\n",
-				bo, (unsigned)bo_va->soffset, tmp->bo,
-				(unsigned)tmp->soffset, (unsigned)tmp->eoffset);
-			kfree(bo_va);
-			mutex_unlock(&vm->mutex);
-			return -EINVAL;
-		}
-		last_offset = tmp->eoffset;
-		head = &tmp->vm_list;
-	}
-	list_add(&bo_va->vm_list, head);
-	list_add_tail(&bo_va->bo_list, &bo->va);
-	mutex_unlock(&vm->mutex);
-	return 0;
-}
-
-static u64 radeon_vm_get_addr(struct radeon_device *rdev,
-			      struct ttm_mem_reg *mem,
-			      unsigned pfn)
-{
-	u64 addr = 0;
-
-	switch (mem->mem_type) {
-	case TTM_PL_VRAM:
-		addr = (mem->start << PAGE_SHIFT);
-		addr += pfn * RADEON_GPU_PAGE_SIZE;
-		addr += rdev->vm_manager.vram_base_offset;
-		break;
-	case TTM_PL_TT:
-		/* offset inside page table */
-		addr = mem->start << PAGE_SHIFT;
-		addr += pfn * RADEON_GPU_PAGE_SIZE;
-		addr = addr >> PAGE_SHIFT;
-		/* page table offset */
-		addr = rdev->gart.pages_addr[addr];
-		/* in case cpu page size != gpu page size*/
-		addr += (pfn * RADEON_GPU_PAGE_SIZE) & (~PAGE_MASK);
-		break;
-	default:
-		break;
-	}
-	return addr;
-}
-
-/* object have to be reserved & cs mutex took & vm mutex took */
-int radeon_vm_bo_update_pte(struct radeon_device *rdev,
-			    struct radeon_vm *vm,
-			    struct radeon_bo *bo,
-			    struct ttm_mem_reg *mem)
-{
-	struct radeon_bo_va *bo_va;
-	unsigned ngpu_pages, i;
-	uint64_t addr = 0, pfn;
-	uint32_t flags;
-
-	/* nothing to do if vm isn't bound */
-	if (vm->id == -1)
-		return 0;;
-
-	bo_va = radeon_bo_va(bo, vm);
-	if (bo_va == NULL) {
-		dev_err(rdev->dev, "bo %p not in vm %p\n", bo, vm);
-		return -EINVAL;
-	}
-
-	if (bo_va->valid)
-		return 0;
-
-	ngpu_pages = radeon_bo_ngpu_pages(bo);
-	bo_va->flags &= ~RADEON_VM_PAGE_VALID;
-	bo_va->flags &= ~RADEON_VM_PAGE_SYSTEM;
-	if (mem) {
-		if (mem->mem_type != TTM_PL_SYSTEM) {
-			bo_va->flags |= RADEON_VM_PAGE_VALID;
-			bo_va->valid = true;
-		}
-		if (mem->mem_type == TTM_PL_TT) {
-			bo_va->flags |= RADEON_VM_PAGE_SYSTEM;
-		}
-	}
-	pfn = bo_va->soffset / RADEON_GPU_PAGE_SIZE;
-	flags = rdev->vm_manager.funcs->page_flags(rdev, bo_va->vm, bo_va->flags);
-	for (i = 0, addr = 0; i < ngpu_pages; i++) {
-		if (mem && bo_va->valid) {
-			addr = radeon_vm_get_addr(rdev, mem, i);
-		}
-		rdev->vm_manager.funcs->set_page(rdev, bo_va->vm, i + pfn, addr, flags);
-	}
-	rdev->vm_manager.funcs->tlb_flush(rdev, bo_va->vm);
-	return 0;
-}
-
-/* object have to be reserved */
-int radeon_vm_bo_rmv(struct radeon_device *rdev,
-		     struct radeon_vm *vm,
-		     struct radeon_bo *bo)
-{
-	struct radeon_bo_va *bo_va;
-
-	bo_va = radeon_bo_va(bo, vm);
-	if (bo_va == NULL)
-		return 0;
-
-	mutex_lock(&vm->mutex);
-	radeon_mutex_lock(&rdev->cs_mutex);
-	radeon_vm_bo_update_pte(rdev, vm, bo, NULL);
-	radeon_mutex_unlock(&rdev->cs_mutex);
-	list_del(&bo_va->vm_list);
-	mutex_unlock(&vm->mutex);
-	list_del(&bo_va->bo_list);
-
-	kfree(bo_va);
-	return 0;
-}
-
-void radeon_vm_bo_invalidate(struct radeon_device *rdev,
-			     struct radeon_bo *bo)
-{
-	struct radeon_bo_va *bo_va;
-
-	BUG_ON(!atomic_read(&bo->tbo.reserved));
-	list_for_each_entry(bo_va, &bo->va, bo_list) {
-		bo_va->valid = false;
-	}
-}
-
-int radeon_vm_init(struct radeon_device *rdev, struct radeon_vm *vm)
-{
-	int r;
-
-	vm->id = -1;
-	vm->fence = NULL;
-	mutex_init(&vm->mutex);
-	INIT_LIST_HEAD(&vm->list);
-	INIT_LIST_HEAD(&vm->va);
-	vm->last_pfn = 0;
-	/* map the ib pool buffer at 0 in virtual address space, set
-	 * read only
-	 */
-	r = radeon_vm_bo_add(rdev, vm, rdev->ib_pool.sa_manager.bo, 0,
-			     RADEON_VM_PAGE_READABLE | RADEON_VM_PAGE_SNOOPED);
-	return r;
-}
-
-void radeon_vm_fini(struct radeon_device *rdev, struct radeon_vm *vm)
-{
-	struct radeon_bo_va *bo_va, *tmp;
-	int r;
-
-	mutex_lock(&vm->mutex);
-
-	radeon_mutex_lock(&rdev->cs_mutex);
-	radeon_vm_unbind_locked(rdev, vm);
-	radeon_mutex_unlock(&rdev->cs_mutex);
-
-	/* remove all bo */
-	r = radeon_bo_reserve(rdev->ib_pool.sa_manager.bo, false);
-	if (!r) {
-		bo_va = radeon_bo_va(rdev->ib_pool.sa_manager.bo, vm);
-		list_del_init(&bo_va->bo_list);
-		list_del_init(&bo_va->vm_list);
-		radeon_bo_unreserve(rdev->ib_pool.sa_manager.bo);
-		kfree(bo_va);
-	}
-	if (!list_empty(&vm->va)) {
-		dev_err(rdev->dev, "still active bo inside vm\n");
-	}
-	list_for_each_entry_safe(bo_va, tmp, &vm->va, vm_list) {
-		list_del_init(&bo_va->vm_list);
-		r = radeon_bo_reserve(bo_va->bo, false);
-		if (!r) {
-			list_del_init(&bo_va->bo_list);
-			radeon_bo_unreserve(bo_va->bo);
-			kfree(bo_va);
-		}
-	}
-	mutex_unlock(&vm->mutex);
-}
diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c
index c7008b5..aa1ca2d 100644
--- a/drivers/gpu/drm/radeon/radeon_gem.c
+++ b/drivers/gpu/drm/radeon/radeon_gem.c
@@ -75,6 +75,32 @@ int radeon_gem_object_create(struct radeon_device *rdev, int size,
 	return 0;
 }
 
+int radeon_gem_object_pin(struct drm_gem_object *obj, uint32_t pin_domain,
+			  uint64_t *gpu_addr)
+{
+	struct radeon_bo *robj = gem_to_radeon_bo(obj);
+	int r;
+
+	r = radeon_bo_reserve(robj, false);
+	if (unlikely(r != 0))
+		return r;
+	r = radeon_bo_pin(robj, pin_domain, gpu_addr);
+	radeon_bo_unreserve(robj);
+	return r;
+}
+
+void radeon_gem_object_unpin(struct drm_gem_object *obj)
+{
+	struct radeon_bo *robj = gem_to_radeon_bo(obj);
+	int r;
+
+	r = radeon_bo_reserve(robj, false);
+	if (likely(r == 0)) {
+		radeon_bo_unpin(robj);
+		radeon_bo_unreserve(robj);
+	}
+}
+
 int radeon_gem_set_domain(struct drm_gem_object *gobj,
 			  uint32_t rdomain, uint32_t wdomain)
 {
@@ -116,44 +142,6 @@ void radeon_gem_fini(struct radeon_device *rdev)
 	radeon_bo_force_delete(rdev);
 }
 
-/*
- * Call from drm_gem_handle_create which appear in both new and open ioctl
- * case.
- */
-int radeon_gem_object_open(struct drm_gem_object *obj, struct drm_file *file_priv)
-{
-	return 0;
-}
-
-void radeon_gem_object_close(struct drm_gem_object *obj,
-			     struct drm_file *file_priv)
-{
-	struct radeon_bo *rbo = gem_to_radeon_bo(obj);
-	struct radeon_device *rdev = rbo->rdev;
-	struct radeon_fpriv *fpriv = file_priv->driver_priv;
-	struct radeon_vm *vm = &fpriv->vm;
-	struct radeon_bo_va *bo_va, *tmp;
-
-	if (rdev->family < CHIP_CAYMAN) {
-		return;
-	}
-
-	if (radeon_bo_reserve(rbo, false)) {
-		return;
-	}
-	list_for_each_entry_safe(bo_va, tmp, &rbo->va, bo_list) {
-		if (bo_va->vm == vm) {
-			/* remove from this vm address space */
-			mutex_lock(&vm->mutex);
-			list_del(&bo_va->vm_list);
-			mutex_unlock(&vm->mutex);
-			list_del(&bo_va->bo_list);
-			kfree(bo_va);
-		}
-	}
-	radeon_bo_unreserve(rbo);
-}
-
 
 /*
  * GEM ioctls.
@@ -164,7 +152,6 @@ int radeon_gem_info_ioctl(struct drm_device *dev, void *data,
 	struct radeon_device *rdev = dev->dev_private;
 	struct drm_radeon_gem_info *args = data;
 	struct ttm_mem_type_manager *man;
-	unsigned i;
 
 	man = &rdev->mman.bdev.man[TTM_PL_VRAM];
 
@@ -173,9 +160,8 @@ int radeon_gem_info_ioctl(struct drm_device *dev, void *data,
 	if (rdev->stollen_vga_memory)
 		args->vram_visible -= radeon_bo_size(rdev->stollen_vga_memory);
 	args->vram_visible -= radeon_fbdev_total_size(rdev);
-	args->gart_size = rdev->mc.gtt_size - 4096 - RADEON_IB_POOL_SIZE*64*1024;
-	for(i = 0; i < RADEON_NUM_RINGS; ++i)
-		args->gart_size -= rdev->ring[i].ring_size;
+	args->gart_size = rdev->mc.gtt_size - rdev->cp.ring_size - 4096 -
+		RADEON_IB_POOL_SIZE*64*1024;
 	return 0;
 }
 
@@ -366,109 +352,6 @@ out:
 	return r;
 }
 
-int radeon_gem_va_ioctl(struct drm_device *dev, void *data,
-			  struct drm_file *filp)
-{
-	struct drm_radeon_gem_va *args = data;
-	struct drm_gem_object *gobj;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_fpriv *fpriv = filp->driver_priv;
-	struct radeon_bo *rbo;
-	struct radeon_bo_va *bo_va;
-	u32 invalid_flags;
-	int r = 0;
-
-	if (!rdev->vm_manager.enabled) {
-		args->operation = RADEON_VA_RESULT_ERROR;
-		return -ENOTTY;
-	}
-
-	/* !! DONT REMOVE !!
-	 * We don't support vm_id yet, to be sure we don't have have broken
-	 * userspace, reject anyone trying to use non 0 value thus moving
-	 * forward we can use those fields without breaking existant userspace
-	 */
-	if (args->vm_id) {
-		args->operation = RADEON_VA_RESULT_ERROR;
-		return -EINVAL;
-	}
-
-	if (args->offset < RADEON_VA_RESERVED_SIZE) {
-		dev_err(&dev->pdev->dev,
-			"offset 0x%lX is in reserved area 0x%X\n",
-			(unsigned long)args->offset,
-			RADEON_VA_RESERVED_SIZE);
-		args->operation = RADEON_VA_RESULT_ERROR;
-		return -EINVAL;
-	}
-
-	/* don't remove, we need to enforce userspace to set the snooped flag
-	 * otherwise we will endup with broken userspace and we won't be able
-	 * to enable this feature without adding new interface
-	 */
-	invalid_flags = RADEON_VM_PAGE_VALID | RADEON_VM_PAGE_SYSTEM;
-	if ((args->flags & invalid_flags)) {
-		dev_err(&dev->pdev->dev, "invalid flags 0x%08X vs 0x%08X\n",
-			args->flags, invalid_flags);
-		args->operation = RADEON_VA_RESULT_ERROR;
-		return -EINVAL;
-	}
-	if (!(args->flags & RADEON_VM_PAGE_SNOOPED)) {
-		dev_err(&dev->pdev->dev, "only supported snooped mapping for now\n");
-		args->operation = RADEON_VA_RESULT_ERROR;
-		return -EINVAL;
-	}
-
-	switch (args->operation) {
-	case RADEON_VA_MAP:
-	case RADEON_VA_UNMAP:
-		break;
-	default:
-		dev_err(&dev->pdev->dev, "unsupported operation %d\n",
-			args->operation);
-		args->operation = RADEON_VA_RESULT_ERROR;
-		return -EINVAL;
-	}
-
-	gobj = drm_gem_object_lookup(dev, filp, args->handle);
-	if (gobj == NULL) {
-		args->operation = RADEON_VA_RESULT_ERROR;
-		return -ENOENT;
-	}
-	rbo = gem_to_radeon_bo(gobj);
-	r = radeon_bo_reserve(rbo, false);
-	if (r) {
-		args->operation = RADEON_VA_RESULT_ERROR;
-		drm_gem_object_unreference_unlocked(gobj);
-		return r;
-	}
-	switch (args->operation) {
-	case RADEON_VA_MAP:
-		bo_va = radeon_bo_va(rbo, &fpriv->vm);
-		if (bo_va) {
-			args->operation = RADEON_VA_RESULT_VA_EXIST;
-			args->offset = bo_va->soffset;
-			goto out;
-		}
-		r = radeon_vm_bo_add(rdev, &fpriv->vm, rbo,
-				     args->offset, args->flags);
-		break;
-	case RADEON_VA_UNMAP:
-		r = radeon_vm_bo_rmv(rdev, &fpriv->vm, rbo);
-		break;
-	default:
-		break;
-	}
-	args->operation = RADEON_VA_RESULT_OK;
-	if (r) {
-		args->operation = RADEON_VA_RESULT_ERROR;
-	}
-out:
-	radeon_bo_unreserve(rbo);
-	drm_gem_object_unreference_unlocked(gobj);
-	return r;
-}
-
 int radeon_mode_dumb_create(struct drm_file *file_priv,
 			    struct drm_device *dev,
 			    struct drm_mode_create_dumb *args)
diff --git a/drivers/gpu/drm/radeon/radeon_i2c.c b/drivers/gpu/drm/radeon/radeon_i2c.c
index 91ef8f6..c90425c 100644
--- a/drivers/gpu/drm/radeon/radeon_i2c.c
+++ b/drivers/gpu/drm/radeon/radeon_i2c.c
@@ -23,41 +23,39 @@
  * Authors: Dave Airlie
  *          Alex Deucher
  */
-
 #include "drmP.h"
-#include "drm_edid.h"
 #include "radeon_drm.h"
 #include "radeon.h"
 #include "atom.h"
 
-extern int radeon_atom_hw_i2c_xfer(struct i2c_adapter *i2c_adap,
-				   struct i2c_msg *msgs, int num);
-extern u32 radeon_atom_hw_i2c_func(struct i2c_adapter *adap);
-
 /**
  * radeon_ddc_probe
  *
  */
-bool radeon_ddc_probe(struct radeon_connector *radeon_connector)
+bool radeon_ddc_probe(struct radeon_connector *radeon_connector, bool requires_extended_probe)
 {
 	u8 out = 0x0;
 	u8 buf[8];
 	int ret;
 	struct i2c_msg msgs[] = {
 		{
-			.addr = DDC_ADDR,
+			.addr = 0x50,
 			.flags = 0,
 			.len = 1,
 			.buf = &out,
 		},
 		{
-			.addr = DDC_ADDR,
+			.addr = 0x50,
 			.flags = I2C_M_RD,
-			.len = 8,
+			.len = 1,
 			.buf = buf,
 		}
 	};
 
+	/* Read 8 bytes from i2c for extended probe of EDID header */
+	if (requires_extended_probe)
+		msgs[1].len = 8;
+
 	/* on hw with routers, select right port */
 	if (radeon_connector->router.ddc_valid)
 		radeon_router_select_ddc_port(radeon_connector);
@@ -66,24 +64,25 @@ bool radeon_ddc_probe(struct radeon_connector *radeon_connector)
 	if (ret != 2)
 		/* Couldn't find an accessible DDC on this connector */
 		return false;
-	/* Probe also for valid EDID header
-	 * EDID header starts with:
-	 * 0x00,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0x00.
-	 * Only the first 6 bytes must be valid as
-	 * drm_edid_block_valid() can fix the last 2 bytes */
-	if (drm_edid_header_is_valid(buf) < 6) {
-		/* Couldn't find an accessible EDID on this
-		 * connector */
-		return false;
+	if (requires_extended_probe) {
+		/* Probe also for valid EDID header
+		 * EDID header starts with:
+		 * 0x00,0xFF,0xFF,0xFF,0xFF,0xFF,0xFF,0x00.
+		 * Only the first 6 bytes must be valid as
+		 * drm_edid_block_valid() can fix the last 2 bytes */
+		if (drm_edid_header_is_valid(buf) < 6) {
+			/* Couldn't find an accessible EDID on this
+			 * connector */
+			return false;
+		}
 	}
 	return true;
 }
 
 /* bit banging i2c */
 
-static int pre_xfer(struct i2c_adapter *i2c_adap)
+static void radeon_i2c_do_lock(struct radeon_i2c_chan *i2c, int lock_state)
 {
-	struct radeon_i2c_chan *i2c = i2c_get_adapdata(i2c_adap);
 	struct radeon_device *rdev = i2c->dev->dev_private;
 	struct radeon_i2c_bus_rec *rec = &i2c->rec;
 	uint32_t temp;
@@ -138,30 +137,19 @@ static int pre_xfer(struct i2c_adapter *i2c_adap)
 	WREG32(rec->en_data_reg, temp);
 
 	/* mask the gpio pins for software use */
-	temp = RREG32(rec->mask_clk_reg) | rec->mask_clk_mask;
-	WREG32(rec->mask_clk_reg, temp);
 	temp = RREG32(rec->mask_clk_reg);
-
-	temp = RREG32(rec->mask_data_reg) | rec->mask_data_mask;
-	WREG32(rec->mask_data_reg, temp);
-	temp = RREG32(rec->mask_data_reg);
-
-	return 0;
-}
-
-static void post_xfer(struct i2c_adapter *i2c_adap)
-{
-	struct radeon_i2c_chan *i2c = i2c_get_adapdata(i2c_adap);
-	struct radeon_device *rdev = i2c->dev->dev_private;
-	struct radeon_i2c_bus_rec *rec = &i2c->rec;
-	uint32_t temp;
-
-	/* unmask the gpio pins for software use */
-	temp = RREG32(rec->mask_clk_reg) & ~rec->mask_clk_mask;
+	if (lock_state)
+		temp |= rec->mask_clk_mask;
+	else
+		temp &= ~rec->mask_clk_mask;
 	WREG32(rec->mask_clk_reg, temp);
 	temp = RREG32(rec->mask_clk_reg);
 
-	temp = RREG32(rec->mask_data_reg) & ~rec->mask_data_mask;
+	temp = RREG32(rec->mask_data_reg);
+	if (lock_state)
+		temp |= rec->mask_data_mask;
+	else
+		temp &= ~rec->mask_data_mask;
 	WREG32(rec->mask_data_reg, temp);
 	temp = RREG32(rec->mask_data_reg);
 }
@@ -221,6 +209,22 @@ static void set_data(void *i2c_priv, int data)
 	WREG32(rec->en_data_reg, val);
 }
 
+static int pre_xfer(struct i2c_adapter *i2c_adap)
+{
+	struct radeon_i2c_chan *i2c = i2c_get_adapdata(i2c_adap);
+
+	radeon_i2c_do_lock(i2c, 1);
+
+	return 0;
+}
+
+static void post_xfer(struct i2c_adapter *i2c_adap)
+{
+	struct radeon_i2c_chan *i2c = i2c_get_adapdata(i2c_adap);
+
+	radeon_i2c_do_lock(i2c, 0);
+}
+
 /* hw i2c */
 
 static u32 radeon_get_i2c_prescale(struct radeon_device *rdev)
@@ -886,11 +890,6 @@ static const struct i2c_algorithm radeon_i2c_algo = {
 	.functionality = radeon_hw_i2c_func,
 };
 
-static const struct i2c_algorithm radeon_atom_i2c_algo = {
-	.master_xfer = radeon_atom_hw_i2c_xfer,
-	.functionality = radeon_atom_hw_i2c_func,
-};
-
 struct radeon_i2c_chan *radeon_i2c_create(struct drm_device *dev,
 					  struct radeon_i2c_bus_rec *rec,
 					  const char *name)
@@ -910,7 +909,6 @@ struct radeon_i2c_chan *radeon_i2c_create(struct drm_device *dev,
 	i2c->rec = *rec;
 	i2c->adapter.owner = THIS_MODULE;
 	i2c->adapter.class = I2C_CLASS_DDC;
-	i2c->adapter.dev.parent = &dev->pdev->dev;
 	i2c->dev = dev;
 	i2c_set_adapdata(&i2c->adapter, i2c);
 	if (rec->mm_i2c ||
@@ -927,18 +925,6 @@ struct radeon_i2c_chan *radeon_i2c_create(struct drm_device *dev,
 			DRM_ERROR("Failed to register hw i2c %s\n", name);
 			goto out_free;
 		}
-	} else if (rec->hw_capable &&
-		   radeon_hw_i2c &&
-		   ASIC_IS_DCE3(rdev)) {
-		/* hw i2c using atom */
-		snprintf(i2c->adapter.name, sizeof(i2c->adapter.name),
-			 "Radeon i2c hw bus %s", name);
-		i2c->adapter.algo = &radeon_atom_i2c_algo;
-		ret = i2c_add_adapter(&i2c->adapter);
-		if (ret) {
-			DRM_ERROR("Failed to register hw i2c %s\n", name);
-			goto out_free;
-		}
 	} else {
 		/* set the radeon bit adapter */
 		snprintf(i2c->adapter.name, sizeof(i2c->adapter.name),
@@ -950,8 +936,10 @@ struct radeon_i2c_chan *radeon_i2c_create(struct drm_device *dev,
 		i2c->algo.bit.setscl = set_clock;
 		i2c->algo.bit.getsda = get_data;
 		i2c->algo.bit.getscl = get_clock;
-		i2c->algo.bit.udelay = 10;
-		i2c->algo.bit.timeout = usecs_to_jiffies(2200);	/* from VESA */
+		i2c->algo.bit.udelay = 20;
+		/* vesa says 2.2 ms is enough, 1 jiffy doesn't seem to always
+		 * make this, 2 jiffies is a lot more reliable */
+		i2c->algo.bit.timeout = 2;
 		i2c->algo.bit.data = i2c;
 		ret = i2c_bit_add_bus(&i2c->adapter);
 		if (ret) {
@@ -981,7 +969,6 @@ struct radeon_i2c_chan *radeon_i2c_create_dp(struct drm_device *dev,
 	i2c->rec = *rec;
 	i2c->adapter.owner = THIS_MODULE;
 	i2c->adapter.class = I2C_CLASS_DDC;
-	i2c->adapter.dev.parent = &dev->pdev->dev;
 	i2c->dev = dev;
 	snprintf(i2c->adapter.name, sizeof(i2c->adapter.name),
 		 "Radeon aux bus %s", name);
diff --git a/drivers/gpu/drm/radeon/radeon_irq.c b/drivers/gpu/drm/radeon/radeon_irq.c
index 00da384..465746b 100644
--- a/drivers/gpu/drm/radeon/radeon_irq.c
+++ b/drivers/gpu/drm/radeon/radeon_irq.c
@@ -129,7 +129,7 @@ void radeon_disable_vblank(struct drm_device *dev, int crtc)
 	}
 }
 
-static u32 radeon_acknowledge_irqs(drm_radeon_private_t *dev_priv, u32 *r500_disp_int)
+static inline u32 radeon_acknowledge_irqs(drm_radeon_private_t *dev_priv, u32 *r500_disp_int)
 {
 	u32 irqs = RADEON_READ(RADEON_GEN_INT_STATUS);
 	u32 irq_mask = RADEON_SW_INT_TEST;
diff --git a/drivers/gpu/drm/radeon/radeon_irq_kms.c b/drivers/gpu/drm/radeon/radeon_irq_kms.c
index 645dcbf..1cfe753 100644
--- a/drivers/gpu/drm/radeon/radeon_irq_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_irq_kms.c
@@ -65,13 +65,12 @@ void radeon_driver_irq_preinstall_kms(struct drm_device *dev)
 	unsigned i;
 
 	/* Disable *all* interrupts */
-	for (i = 0; i < RADEON_NUM_RINGS; i++)
-		rdev->irq.sw_int[i] = false;
+	rdev->irq.sw_int = false;
 	rdev->irq.gui_idle = false;
-	for (i = 0; i < RADEON_MAX_HPD_PINS; i++)
-		rdev->irq.hpd[i] = false;
-	for (i = 0; i < RADEON_MAX_CRTCS; i++) {
+	for (i = 0; i < rdev->num_crtc; i++)
 		rdev->irq.crtc_vblank_int[i] = false;
+	for (i = 0; i < 6; i++) {
+		rdev->irq.hpd[i] = false;
 		rdev->irq.pflip[i] = false;
 	}
 	radeon_irq_set(rdev);
@@ -82,11 +81,9 @@ void radeon_driver_irq_preinstall_kms(struct drm_device *dev)
 int radeon_driver_irq_postinstall_kms(struct drm_device *dev)
 {
 	struct radeon_device *rdev = dev->dev_private;
-	unsigned i;
 
 	dev->max_vblank_count = 0x001fffff;
-	for (i = 0; i < RADEON_NUM_RINGS; i++)
-		rdev->irq.sw_int[i] = true;
+	rdev->irq.sw_int = true;
 	radeon_irq_set(rdev);
 	return 0;
 }
@@ -100,13 +97,12 @@ void radeon_driver_irq_uninstall_kms(struct drm_device *dev)
 		return;
 	}
 	/* Disable *all* interrupts */
-	for (i = 0; i < RADEON_NUM_RINGS; i++)
-		rdev->irq.sw_int[i] = false;
+	rdev->irq.sw_int = false;
 	rdev->irq.gui_idle = false;
-	for (i = 0; i < RADEON_MAX_HPD_PINS; i++)
-		rdev->irq.hpd[i] = false;
-	for (i = 0; i < RADEON_MAX_CRTCS; i++) {
+	for (i = 0; i < rdev->num_crtc; i++)
 		rdev->irq.crtc_vblank_int[i] = false;
+	for (i = 0; i < 6; i++) {
+		rdev->irq.hpd[i] = false;
 		rdev->irq.pflip[i] = false;
 	}
 	radeon_irq_set(rdev);
@@ -220,26 +216,26 @@ void radeon_irq_kms_fini(struct radeon_device *rdev)
 	flush_work_sync(&rdev->hotplug_work);
 }
 
-void radeon_irq_kms_sw_irq_get(struct radeon_device *rdev, int ring)
+void radeon_irq_kms_sw_irq_get(struct radeon_device *rdev)
 {
 	unsigned long irqflags;
 
 	spin_lock_irqsave(&rdev->irq.sw_lock, irqflags);
-	if (rdev->ddev->irq_enabled && (++rdev->irq.sw_refcount[ring] == 1)) {
-		rdev->irq.sw_int[ring] = true;
+	if (rdev->ddev->irq_enabled && (++rdev->irq.sw_refcount == 1)) {
+		rdev->irq.sw_int = true;
 		radeon_irq_set(rdev);
 	}
 	spin_unlock_irqrestore(&rdev->irq.sw_lock, irqflags);
 }
 
-void radeon_irq_kms_sw_irq_put(struct radeon_device *rdev, int ring)
+void radeon_irq_kms_sw_irq_put(struct radeon_device *rdev)
 {
 	unsigned long irqflags;
 
 	spin_lock_irqsave(&rdev->irq.sw_lock, irqflags);
-	BUG_ON(rdev->ddev->irq_enabled && rdev->irq.sw_refcount[ring] <= 0);
-	if (rdev->ddev->irq_enabled && (--rdev->irq.sw_refcount[ring] == 0)) {
-		rdev->irq.sw_int[ring] = false;
+	BUG_ON(rdev->ddev->irq_enabled && rdev->irq.sw_refcount <= 0);
+	if (rdev->ddev->irq_enabled && (--rdev->irq.sw_refcount == 0)) {
+		rdev->irq.sw_int = false;
 		radeon_irq_set(rdev);
 	}
 	spin_unlock_irqrestore(&rdev->irq.sw_lock, irqflags);
diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index 1986eba..bd58af6 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -57,12 +57,10 @@ int radeon_driver_load_kms(struct drm_device *dev, unsigned long flags)
 	}
 	dev->dev_private = (void *)rdev;
 
-	pci_set_master(dev->pdev);
-
 	/* update BUS flag */
 	if (drm_pci_device_is_agp(dev)) {
 		flags |= RADEON_IS_AGP;
-	} else if (pci_is_pcie(dev->pdev)) {
+	} else if (drm_pci_device_is_pcie(dev)) {
 		flags |= RADEON_IS_PCIE;
 	} else {
 		flags |= RADEON_IS_PCI;
@@ -239,31 +237,6 @@ int radeon_info_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 	case RADEON_INFO_FUSION_GART_WORKING:
 		value = 1;
 		break;
-	case RADEON_INFO_BACKEND_MAP:
-		if (rdev->family >= CHIP_CAYMAN)
-			value = rdev->config.cayman.backend_map;
-		else if (rdev->family >= CHIP_CEDAR)
-			value = rdev->config.evergreen.backend_map;
-		else if (rdev->family >= CHIP_RV770)
-			value = rdev->config.rv770.backend_map;
-		else if (rdev->family >= CHIP_R600)
-			value = rdev->config.r600.backend_map;
-		else {
-			return -EINVAL;
-		}
-		break;
-	case RADEON_INFO_VA_START:
-		/* this is where we report if vm is supported or not */
-		if (rdev->family < CHIP_CAYMAN)
-			return -EINVAL;
-		value = RADEON_VA_RESERVED_SIZE;
-		break;
-	case RADEON_INFO_IB_VM_MAX_SIZE:
-		/* this is where we report if vm is supported or not */
-		if (rdev->family < CHIP_CAYMAN)
-			return -EINVAL;
-		value = RADEON_IB_VM_MAX_SIZE;
-		break;
 	default:
 		DRM_DEBUG_KMS("Invalid request %d\n", info->request);
 		return -EINVAL;
@@ -284,6 +257,7 @@ int radeon_driver_firstopen_kms(struct drm_device *dev)
 	return 0;
 }
 
+
 void radeon_driver_lastclose_kms(struct drm_device *dev)
 {
 	vga_switcheroo_process_delayed_switch();
@@ -291,45 +265,12 @@ void radeon_driver_lastclose_kms(struct drm_device *dev)
 
 int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
 {
-	struct radeon_device *rdev = dev->dev_private;
-
-	file_priv->driver_priv = NULL;
-
-	/* new gpu have virtual address space support */
-	if (rdev->family >= CHIP_CAYMAN) {
-		struct radeon_fpriv *fpriv;
-		int r;
-
-		fpriv = kzalloc(sizeof(*fpriv), GFP_KERNEL);
-		if (unlikely(!fpriv)) {
-			return -ENOMEM;
-		}
-
-		r = radeon_vm_init(rdev, &fpriv->vm);
-		if (r) {
-			radeon_vm_fini(rdev, &fpriv->vm);
-			kfree(fpriv);
-			return r;
-		}
-
-		file_priv->driver_priv = fpriv;
-	}
 	return 0;
 }
 
 void radeon_driver_postclose_kms(struct drm_device *dev,
 				 struct drm_file *file_priv)
 {
-	struct radeon_device *rdev = dev->dev_private;
-
-	/* new gpu have virtual address space support */
-	if (rdev->family >= CHIP_CAYMAN && file_priv->driver_priv) {
-		struct radeon_fpriv *fpriv = file_priv->driver_priv;
-
-		radeon_vm_fini(rdev, &fpriv->vm);
-		kfree(fpriv);
-		file_priv->driver_priv = NULL;
-	}
 }
 
 void radeon_driver_preclose_kms(struct drm_device *dev,
@@ -497,6 +438,5 @@ struct drm_ioctl_desc radeon_ioctls_kms[] = {
 	DRM_IOCTL_DEF_DRV(RADEON_GEM_SET_TILING, radeon_gem_set_tiling_ioctl, DRM_AUTH|DRM_UNLOCKED),
 	DRM_IOCTL_DEF_DRV(RADEON_GEM_GET_TILING, radeon_gem_get_tiling_ioctl, DRM_AUTH|DRM_UNLOCKED),
 	DRM_IOCTL_DEF_DRV(RADEON_GEM_BUSY, radeon_gem_busy_ioctl, DRM_AUTH|DRM_UNLOCKED),
-	DRM_IOCTL_DEF_DRV(RADEON_GEM_VA, radeon_gem_va_ioctl, DRM_AUTH|DRM_UNLOCKED),
 };
 int radeon_max_kms_ioctl = DRM_ARRAY_SIZE(radeon_ioctls_kms);
diff --git a/drivers/gpu/drm/radeon/radeon_legacy_crtc.c b/drivers/gpu/drm/radeon/radeon_legacy_crtc.c
index 210317c..41a5d48 100644
--- a/drivers/gpu/drm/radeon/radeon_legacy_crtc.c
+++ b/drivers/gpu/drm/radeon/radeon_legacy_crtc.c
@@ -419,9 +419,7 @@ int radeon_crtc_do_set_base(struct drm_crtc *crtc,
 	r = radeon_bo_reserve(rbo, false);
 	if (unlikely(r != 0))
 		return r;
-	/* Only 27 bit offset for legacy CRTC */
-	r = radeon_bo_pin_restricted(rbo, RADEON_GEM_DOMAIN_VRAM, 1 << 27,
-				     &base);
+	r = radeon_bo_pin(rbo, RADEON_GEM_DOMAIN_VRAM, &base);
 	if (unlikely(r != 0)) {
 		radeon_bo_unreserve(rbo);
 		return -EINVAL;
@@ -439,7 +437,7 @@ int radeon_crtc_do_set_base(struct drm_crtc *crtc,
 
 	crtc_offset_cntl = 0;
 
-	pitch_pixels = target_fb->pitches[0] / (target_fb->bits_per_pixel / 8);
+	pitch_pixels = target_fb->pitch / (target_fb->bits_per_pixel / 8);
 	crtc_pitch  = (((pitch_pixels * target_fb->bits_per_pixel) +
 			((target_fb->bits_per_pixel * 8) - 1)) /
 		       (target_fb->bits_per_pixel * 8));
@@ -993,6 +991,12 @@ static bool radeon_crtc_mode_fixup(struct drm_crtc *crtc,
 				   struct drm_display_mode *mode,
 				   struct drm_display_mode *adjusted_mode)
 {
+	struct drm_device *dev = crtc->dev;
+	struct radeon_device *rdev = dev->dev_private;
+
+	/* adjust pm to upcoming mode change */
+	radeon_pm_compute_clocks(rdev);
+
 	if (!radeon_crtc_scaling_mode_fixup(crtc, mode, adjusted_mode))
 		return false;
 	return true;
diff --git a/drivers/gpu/drm/radeon/radeon_legacy_tv.c b/drivers/gpu/drm/radeon/radeon_legacy_tv.c
index b37ec0f..c7b6cb4 100644
--- a/drivers/gpu/drm/radeon/radeon_legacy_tv.c
+++ b/drivers/gpu/drm/radeon/radeon_legacy_tv.c
@@ -864,7 +864,7 @@ void radeon_legacy_tv_adjust_crtc_reg(struct drm_encoder *encoder,
 	*v_sync_strt_wid = tmp;
 }
 
-static int get_post_div(int value)
+static inline int get_post_div(int value)
 {
 	int post_div;
 	switch (value) {
diff --git a/drivers/gpu/drm/radeon/radeon_mem.c b/drivers/gpu/drm/radeon/radeon_mem.c
index 988548e..ed95155 100644
--- a/drivers/gpu/drm/radeon/radeon_mem.c
+++ b/drivers/gpu/drm/radeon/radeon_mem.c
@@ -139,7 +139,7 @@ static int init_heap(struct mem_block **heap, int start, int size)
 	if (!blocks)
 		return -ENOMEM;
 
-	*heap = kzalloc(sizeof(**heap), GFP_KERNEL);
+	*heap = kmalloc(sizeof(**heap), GFP_KERNEL);
 	if (!*heap) {
 		kfree(blocks);
 		return -ENOMEM;
@@ -150,6 +150,7 @@ static int init_heap(struct mem_block **heap, int start, int size)
 	blocks->file_priv = NULL;
 	blocks->next = blocks->prev = *heap;
 
+	memset(*heap, 0, sizeof(**heap));
 	(*heap)->file_priv = (struct drm_file *) - 1;
 	(*heap)->next = (*heap)->prev = blocks;
 	return 0;
diff --git a/drivers/gpu/drm/radeon/radeon_mode.h b/drivers/gpu/drm/radeon/radeon_mode.h
index 8a85598..ed0178f 100644
--- a/drivers/gpu/drm/radeon/radeon_mode.h
+++ b/drivers/gpu/drm/radeon/radeon_mode.h
@@ -362,7 +362,6 @@ struct radeon_encoder_atom_dig {
 	struct backlight_device *bl_dev;
 	int dpms_mode;
 	uint8_t backlight_level;
-	int panel_mode;
 };
 
 struct radeon_encoder_atom_dac {
@@ -439,6 +438,9 @@ struct radeon_connector {
 	struct radeon_i2c_chan *ddc_bus;
 	/* some systems have an hdmi and vga port with a shared ddc line */
 	bool shared_ddc;
+	/* for some Radeon chip families we apply an additional EDID header
+	   check as part of the DDC probe */
+	bool requires_extended_probe;
 	bool use_digital;
 	/* we need to mind the EDID between detect
 	   and get modes due to analog/digital/tvencoder */
@@ -457,8 +459,6 @@ struct radeon_framebuffer {
 	struct drm_gem_object *obj;
 };
 
-#define ENCODER_MODE_IS_DP(em) (((em) == ATOM_ENCODER_MODE_DP) || \
-				((em) == ATOM_ENCODER_MODE_DP_MST))
 
 extern enum radeon_tv_std
 radeon_combios_get_tv_info(struct radeon_device *rdev);
@@ -467,13 +467,9 @@ radeon_atombios_get_tv_info(struct radeon_device *rdev);
 
 extern struct drm_connector *
 radeon_get_connector_for_encoder(struct drm_encoder *encoder);
-extern struct drm_connector *
-radeon_get_connector_for_encoder_init(struct drm_encoder *encoder);
-extern bool radeon_dig_monitor_is_duallink(struct drm_encoder *encoder,
-				    u32 pixel_clock);
 
-extern u16 radeon_encoder_get_dp_bridge_encoder_id(struct drm_encoder *encoder);
-extern u16 radeon_connector_encoder_get_dp_bridge_encoder_id(struct drm_connector *connector);
+extern bool radeon_encoder_is_dp_bridge(struct drm_encoder *encoder);
+extern bool radeon_connector_encoder_is_dp_bridge(struct drm_connector *connector);
 extern bool radeon_connector_encoder_is_hbr2(struct drm_connector *connector);
 extern bool radeon_connector_is_dp12_capable(struct drm_connector *connector);
 
@@ -487,16 +483,13 @@ extern void radeon_dp_link_train(struct drm_encoder *encoder,
 extern bool radeon_dp_needs_link_train(struct radeon_connector *radeon_connector);
 extern u8 radeon_dp_getsinktype(struct radeon_connector *radeon_connector);
 extern bool radeon_dp_getdpcd(struct radeon_connector *radeon_connector);
-extern int radeon_dp_get_panel_mode(struct drm_encoder *encoder,
-				    struct drm_connector *connector);
 extern void atombios_dig_encoder_setup(struct drm_encoder *encoder, int action, int panel_mode);
 extern void radeon_atom_encoder_init(struct radeon_device *rdev);
-extern void radeon_atom_dcpll_init(struct radeon_device *rdev);
 extern void atombios_dig_transmitter_setup(struct drm_encoder *encoder,
 					   int action, uint8_t lane_num,
 					   uint8_t lane_set);
 extern void radeon_atom_ext_encoder_setup_ddc(struct drm_encoder *encoder);
-extern struct drm_encoder *radeon_get_external_encoder(struct drm_encoder *encoder);
+extern struct drm_encoder *radeon_atom_get_external_encoder(struct drm_encoder *encoder);
 extern int radeon_dp_i2c_aux_ch(struct i2c_adapter *adapter, int mode,
 				u8 write_byte, u8 *read_byte);
 
@@ -526,7 +519,8 @@ extern void radeon_i2c_put_byte(struct radeon_i2c_chan *i2c,
 				u8 val);
 extern void radeon_router_select_ddc_port(struct radeon_connector *radeon_connector);
 extern void radeon_router_select_cd_port(struct radeon_connector *radeon_connector);
-extern bool radeon_ddc_probe(struct radeon_connector *radeon_connector);
+extern bool radeon_ddc_probe(struct radeon_connector *radeon_connector,
+			bool requires_extended_probe);
 extern int radeon_ddc_get_modes(struct radeon_connector *radeon_connector);
 
 extern struct drm_encoder *radeon_best_encoder(struct drm_connector *connector);
@@ -649,9 +643,9 @@ extern void radeon_crtc_fb_gamma_set(struct drm_crtc *crtc, u16 red, u16 green,
 				     u16 blue, int regno);
 extern void radeon_crtc_fb_gamma_get(struct drm_crtc *crtc, u16 *red, u16 *green,
 				     u16 *blue, int regno);
-int radeon_framebuffer_init(struct drm_device *dev,
+void radeon_framebuffer_init(struct drm_device *dev,
 			     struct radeon_framebuffer *rfb,
-			     struct drm_mode_fb_cmd2 *mode_cmd,
+			     struct drm_mode_fb_cmd *mode_cmd,
 			     struct drm_gem_object *obj);
 
 int radeonfb_remove(struct drm_device *dev, struct drm_framebuffer *fb);
diff --git a/drivers/gpu/drm/radeon/radeon_object.c b/drivers/gpu/drm/radeon/radeon_object.c
index 91541e6..35da1b4 100644
--- a/drivers/gpu/drm/radeon/radeon_object.c
+++ b/drivers/gpu/drm/radeon/radeon_object.c
@@ -46,20 +46,6 @@ static void radeon_bo_clear_surface_reg(struct radeon_bo *bo);
  * function are calling it.
  */
 
-void radeon_bo_clear_va(struct radeon_bo *bo)
-{
-	struct radeon_bo_va *bo_va, *tmp;
-
-	list_for_each_entry_safe(bo_va, tmp, &bo->va, bo_list) {
-		/* remove from all vm address space */
-		mutex_lock(&bo_va->vm->mutex);
-		list_del(&bo_va->vm_list);
-		mutex_unlock(&bo_va->vm->mutex);
-		list_del(&bo_va->bo_list);
-		kfree(bo_va);
-	}
-}
-
 static void radeon_ttm_bo_destroy(struct ttm_buffer_object *tbo)
 {
 	struct radeon_bo *bo;
@@ -69,7 +55,6 @@ static void radeon_ttm_bo_destroy(struct ttm_buffer_object *tbo)
 	list_del_init(&bo->list);
 	mutex_unlock(&bo->rdev->gem.mutex);
 	radeon_bo_clear_surface_reg(bo);
-	radeon_bo_clear_va(bo);
 	drm_gem_object_release(&bo->gem_base);
 	kfree(bo);
 }
@@ -110,7 +95,6 @@ int radeon_bo_create(struct radeon_device *rdev,
 	enum ttm_bo_type type;
 	unsigned long page_align = roundup(byte_align, PAGE_SIZE) >> PAGE_SHIFT;
 	unsigned long max_size = 0;
-	size_t acc_size;
 	int r;
 
 	size = ALIGN(size, PAGE_SIZE);
@@ -133,9 +117,6 @@ int radeon_bo_create(struct radeon_device *rdev,
 		return -ENOMEM;
 	}
 
-	acc_size = ttm_bo_dma_acc_size(&rdev->mman.bdev, size,
-				       sizeof(struct radeon_bo));
-
 retry:
 	bo = kzalloc(sizeof(struct radeon_bo), GFP_KERNEL);
 	if (bo == NULL)
@@ -149,13 +130,12 @@ retry:
 	bo->gem_base.driver_private = NULL;
 	bo->surface_reg = -1;
 	INIT_LIST_HEAD(&bo->list);
-	INIT_LIST_HEAD(&bo->va);
 	radeon_ttm_placement_from_domain(bo, domain);
 	/* Kernel allocation are uninterruptible */
 	mutex_lock(&rdev->vram_mutex);
 	r = ttm_bo_init(&rdev->mman.bdev, &bo->tbo, size, type,
-			&bo->placement, page_align, 0, !kernel, NULL,
-			acc_size, &radeon_ttm_bo_destroy);
+			&bo->placement, page_align, 0, !kernel, NULL, size,
+			&radeon_ttm_bo_destroy);
 	mutex_unlock(&rdev->vram_mutex);
 	if (unlikely(r != 0)) {
 		if (r != -ERESTARTSYS) {
@@ -461,54 +441,8 @@ static void radeon_bo_clear_surface_reg(struct radeon_bo *bo)
 int radeon_bo_set_tiling_flags(struct radeon_bo *bo,
 				uint32_t tiling_flags, uint32_t pitch)
 {
-	struct radeon_device *rdev = bo->rdev;
 	int r;
 
-	if (rdev->family >= CHIP_CEDAR) {
-		unsigned bankw, bankh, mtaspect, tilesplit, stilesplit;
-
-		bankw = (tiling_flags >> RADEON_TILING_EG_BANKW_SHIFT) & RADEON_TILING_EG_BANKW_MASK;
-		bankh = (tiling_flags >> RADEON_TILING_EG_BANKH_SHIFT) & RADEON_TILING_EG_BANKH_MASK;
-		mtaspect = (tiling_flags >> RADEON_TILING_EG_MACRO_TILE_ASPECT_SHIFT) & RADEON_TILING_EG_MACRO_TILE_ASPECT_MASK;
-		tilesplit = (tiling_flags >> RADEON_TILING_EG_TILE_SPLIT_SHIFT) & RADEON_TILING_EG_TILE_SPLIT_MASK;
-		stilesplit = (tiling_flags >> RADEON_TILING_EG_STENCIL_TILE_SPLIT_SHIFT) & RADEON_TILING_EG_STENCIL_TILE_SPLIT_MASK;
-		switch (bankw) {
-		case 0:
-		case 1:
-		case 2:
-		case 4:
-		case 8:
-			break;
-		default:
-			return -EINVAL;
-		}
-		switch (bankh) {
-		case 0:
-		case 1:
-		case 2:
-		case 4:
-		case 8:
-			break;
-		default:
-			return -EINVAL;
-		}
-		switch (mtaspect) {
-		case 0:
-		case 1:
-		case 2:
-		case 4:
-		case 8:
-			break;
-		default:
-			return -EINVAL;
-		}
-		if (tilesplit > 6) {
-			return -EINVAL;
-		}
-		if (stilesplit > 6) {
-			return -EINVAL;
-		}
-	}
 	r = radeon_bo_reserve(bo, false);
 	if (unlikely(r != 0))
 		return r;
@@ -565,7 +499,6 @@ void radeon_bo_move_notify(struct ttm_buffer_object *bo,
 		return;
 	rbo = container_of(bo, struct radeon_bo, tbo);
 	radeon_bo_check_tiling(rbo, 0, 1);
-	radeon_vm_bo_invalidate(rbo->rdev, rbo);
 }
 
 int radeon_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
@@ -598,57 +531,3 @@ int radeon_bo_fault_reserve_notify(struct ttm_buffer_object *bo)
 	}
 	return 0;
 }
-
-int radeon_bo_wait(struct radeon_bo *bo, u32 *mem_type, bool no_wait)
-{
-	int r;
-
-	r = ttm_bo_reserve(&bo->tbo, true, no_wait, false, 0);
-	if (unlikely(r != 0))
-		return r;
-	spin_lock(&bo->tbo.bdev->fence_lock);
-	if (mem_type)
-		*mem_type = bo->tbo.mem.mem_type;
-	if (bo->tbo.sync_obj)
-		r = ttm_bo_wait(&bo->tbo, true, true, no_wait);
-	spin_unlock(&bo->tbo.bdev->fence_lock);
-	ttm_bo_unreserve(&bo->tbo);
-	return r;
-}
-
-
-/**
- * radeon_bo_reserve - reserve bo
- * @bo:		bo structure
- * @no_wait:		don't sleep while trying to reserve (return -EBUSY)
- *
- * Returns:
- * -EBUSY: buffer is busy and @no_wait is true
- * -ERESTARTSYS: A wait for the buffer to become unreserved was interrupted by
- * a signal. Release all buffer reservations and return to user-space.
- */
-int radeon_bo_reserve(struct radeon_bo *bo, bool no_wait)
-{
-	int r;
-
-	r = ttm_bo_reserve(&bo->tbo, true, no_wait, false, 0);
-	if (unlikely(r != 0)) {
-		if (r != -ERESTARTSYS)
-			dev_err(bo->rdev->dev, "%p reserve failed\n", bo);
-		return r;
-	}
-	return 0;
-}
-
-/* object have to be reserved */
-struct radeon_bo_va *radeon_bo_va(struct radeon_bo *rbo, struct radeon_vm *vm)
-{
-	struct radeon_bo_va *bo_va;
-
-	list_for_each_entry(bo_va, &rbo->va, bo_list) {
-		if (bo_va->vm == vm) {
-			return bo_va;
-		}
-	}
-	return NULL;
-}
diff --git a/drivers/gpu/drm/radeon/radeon_object.h b/drivers/gpu/drm/radeon/radeon_object.h
index f9104be..7199c6a 100644
--- a/drivers/gpu/drm/radeon/radeon_object.h
+++ b/drivers/gpu/drm/radeon/radeon_object.h
@@ -52,7 +52,28 @@ static inline unsigned radeon_mem_type_to_domain(u32 mem_type)
 	return 0;
 }
 
-int radeon_bo_reserve(struct radeon_bo *bo, bool no_wait);
+/**
+ * radeon_bo_reserve - reserve bo
+ * @bo:		bo structure
+ * @no_wait:		don't sleep while trying to reserve (return -EBUSY)
+ *
+ * Returns:
+ * -EBUSY: buffer is busy and @no_wait is true
+ * -ERESTARTSYS: A wait for the buffer to become unreserved was interrupted by
+ * a signal. Release all buffer reservations and return to user-space.
+ */
+static inline int radeon_bo_reserve(struct radeon_bo *bo, bool no_wait)
+{
+	int r;
+
+	r = ttm_bo_reserve(&bo->tbo, true, no_wait, false, 0);
+	if (unlikely(r != 0)) {
+		if (r != -ERESTARTSYS)
+			dev_err(bo->rdev->dev, "%p reserve failed\n", bo);
+		return r;
+	}
+	return 0;
+}
 
 static inline void radeon_bo_unreserve(struct radeon_bo *bo)
 {
@@ -83,16 +104,6 @@ static inline bool radeon_bo_is_reserved(struct radeon_bo *bo)
 	return !!atomic_read(&bo->tbo.reserved);
 }
 
-static inline unsigned radeon_bo_ngpu_pages(struct radeon_bo *bo)
-{
-	return (bo->tbo.num_pages << PAGE_SHIFT) / RADEON_GPU_PAGE_SIZE;
-}
-
-static inline unsigned radeon_bo_gpu_page_alignment(struct radeon_bo *bo)
-{
-	return (bo->tbo.mem.page_alignment << PAGE_SHIFT) / RADEON_GPU_PAGE_SIZE;
-}
-
 /**
  * radeon_bo_mmap_offset - return mmap offset of bo
  * @bo:	radeon object for which we query the offset
@@ -107,8 +118,23 @@ static inline u64 radeon_bo_mmap_offset(struct radeon_bo *bo)
 	return bo->tbo.addr_space_offset;
 }
 
-extern int radeon_bo_wait(struct radeon_bo *bo, u32 *mem_type,
-			  bool no_wait);
+static inline int radeon_bo_wait(struct radeon_bo *bo, u32 *mem_type,
+					bool no_wait)
+{
+	int r;
+
+	r = ttm_bo_reserve(&bo->tbo, true, no_wait, false, 0);
+	if (unlikely(r != 0))
+		return r;
+	spin_lock(&bo->tbo.bdev->fence_lock);
+	if (mem_type)
+		*mem_type = bo->tbo.mem.mem_type;
+	if (bo->tbo.sync_obj)
+		r = ttm_bo_wait(&bo->tbo, true, true, no_wait);
+	spin_unlock(&bo->tbo.bdev->fence_lock);
+	ttm_bo_unreserve(&bo->tbo);
+	return r;
+}
 
 extern int radeon_bo_create(struct radeon_device *rdev,
 				unsigned long size, int byte_align,
@@ -140,26 +166,4 @@ extern void radeon_bo_move_notify(struct ttm_buffer_object *bo,
 					struct ttm_mem_reg *mem);
 extern int radeon_bo_fault_reserve_notify(struct ttm_buffer_object *bo);
 extern int radeon_bo_get_surface_reg(struct radeon_bo *bo);
-extern struct radeon_bo_va *radeon_bo_va(struct radeon_bo *rbo,
-					 struct radeon_vm *vm);
-
-/*
- * sub allocation
- */
-extern int radeon_sa_bo_manager_init(struct radeon_device *rdev,
-				     struct radeon_sa_manager *sa_manager,
-				     unsigned size, u32 domain);
-extern void radeon_sa_bo_manager_fini(struct radeon_device *rdev,
-				      struct radeon_sa_manager *sa_manager);
-extern int radeon_sa_bo_manager_start(struct radeon_device *rdev,
-				      struct radeon_sa_manager *sa_manager);
-extern int radeon_sa_bo_manager_suspend(struct radeon_device *rdev,
-					struct radeon_sa_manager *sa_manager);
-extern int radeon_sa_bo_new(struct radeon_device *rdev,
-			    struct radeon_sa_manager *sa_manager,
-			    struct radeon_sa_bo *sa_bo,
-			    unsigned size, unsigned align);
-extern void radeon_sa_bo_free(struct radeon_device *rdev,
-			      struct radeon_sa_bo *sa_bo);
-
 #endif
diff --git a/drivers/gpu/drm/radeon/radeon_pm.c b/drivers/gpu/drm/radeon/radeon_pm.c
index 44d0e5b..8270a85 100644
--- a/drivers/gpu/drm/radeon/radeon_pm.c
+++ b/drivers/gpu/drm/radeon/radeon_pm.c
@@ -53,24 +53,6 @@ static void radeon_pm_set_clocks(struct radeon_device *rdev);
 
 #define ACPI_AC_CLASS           "ac_adapter"
 
-int radeon_pm_get_type_index(struct radeon_device *rdev,
-			     enum radeon_pm_state_type ps_type,
-			     int instance)
-{
-	int i;
-	int found_instance = -1;
-
-	for (i = 0; i < rdev->pm.num_power_states; i++) {
-		if (rdev->pm.power_state[i].type == ps_type) {
-			found_instance++;
-			if (found_instance == instance)
-				return i;
-		}
-	}
-	/* return default if no match */
-	return rdev->pm.default_power_state_index;
-}
-
 #ifdef CONFIG_ACPI
 static int radeon_acpi_event(struct notifier_block *nb,
 			     unsigned long val,
@@ -221,7 +203,7 @@ static void radeon_set_power_state(struct radeon_device *rdev)
 		}
 
 		/* set memory clock */
-		if (rdev->asic->pm.set_memory_clock && (mclk != rdev->pm.current_mclk)) {
+		if (rdev->asic->set_memory_clock && (mclk != rdev->pm.current_mclk)) {
 			radeon_pm_debug_check_in_vbl(rdev, false);
 			radeon_set_memory_clock(rdev, mclk);
 			radeon_pm_debug_check_in_vbl(rdev, true);
@@ -252,10 +234,7 @@ static void radeon_pm_set_clocks(struct radeon_device *rdev)
 
 	mutex_lock(&rdev->ddev->struct_mutex);
 	mutex_lock(&rdev->vram_mutex);
-	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
-		if (rdev->ring[i].ring_obj)
-			mutex_lock(&rdev->ring[i].mutex);
-	}
+	mutex_lock(&rdev->cp.mutex);
 
 	/* gui idle int has issues on older chips it seems */
 	if (rdev->family >= CHIP_R600) {
@@ -271,13 +250,12 @@ static void radeon_pm_set_clocks(struct radeon_device *rdev)
 			radeon_irq_set(rdev);
 		}
 	} else {
-		struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
-		if (ring->ready) {
+		if (rdev->cp.ready) {
 			struct radeon_fence *fence;
-			radeon_ring_alloc(rdev, ring, 64);
-			radeon_fence_create(rdev, &fence, radeon_ring_index(rdev, ring));
+			radeon_ring_alloc(rdev, 64);
+			radeon_fence_create(rdev, &fence);
 			radeon_fence_emit(rdev, fence);
-			radeon_ring_commit(rdev, ring);
+			radeon_ring_commit(rdev);
 			radeon_fence_wait(fence, false);
 			radeon_fence_unref(&fence);
 		}
@@ -311,10 +289,7 @@ static void radeon_pm_set_clocks(struct radeon_device *rdev)
 
 	rdev->pm.dynpm_planned_action = DYNPM_ACTION_NONE;
 
-	for (i = 0; i < RADEON_NUM_RINGS; ++i) {
-		if (rdev->ring[i].ring_obj)
-			mutex_unlock(&rdev->ring[i].mutex);
-	}
+	mutex_unlock(&rdev->cp.mutex);
 	mutex_unlock(&rdev->vram_mutex);
 	mutex_unlock(&rdev->ddev->struct_mutex);
 }
@@ -806,14 +781,19 @@ static void radeon_dynpm_idle_work_handler(struct work_struct *work)
 	resched = ttm_bo_lock_delayed_workqueue(&rdev->mman.bdev);
 	mutex_lock(&rdev->pm.mutex);
 	if (rdev->pm.dynpm_state == DYNPM_STATE_ACTIVE) {
+		unsigned long irq_flags;
 		int not_processed = 0;
-		int i;
 
-		for (i = 0; i < RADEON_NUM_RINGS; ++i) {
-			not_processed += radeon_fence_count_emitted(rdev, i);
-			if (not_processed >= 3)
-				break;
+		read_lock_irqsave(&rdev->fence_drv.lock, irq_flags);
+		if (!list_empty(&rdev->fence_drv.emited)) {
+			struct list_head *ptr;
+			list_for_each(ptr, &rdev->fence_drv.emited) {
+				/* count up to 3, that's enought info */
+				if (++not_processed >= 3)
+					break;
+			}
 		}
+		read_unlock_irqrestore(&rdev->fence_drv.lock, irq_flags);
 
 		if (not_processed >= 3) { /* should upclock */
 			if (rdev->pm.dynpm_planned_action == DYNPM_ACTION_DOWNCLOCK) {
@@ -871,11 +851,11 @@ static int radeon_debugfs_pm_info(struct seq_file *m, void *data)
 	else
 		seq_printf(m, "current engine clock: %u0 kHz\n", radeon_get_engine_clock(rdev));
 	seq_printf(m, "default memory clock: %u0 kHz\n", rdev->pm.default_mclk);
-	if (rdev->asic->pm.get_memory_clock)
+	if (rdev->asic->get_memory_clock)
 		seq_printf(m, "current memory clock: %u0 kHz\n", radeon_get_memory_clock(rdev));
 	if (rdev->pm.current_vddc)
 		seq_printf(m, "voltage: %u mV\n", rdev->pm.current_vddc);
-	if (rdev->asic->pm.get_pcie_lanes)
+	if (rdev->asic->get_pcie_lanes)
 		seq_printf(m, "PCIE lanes: %d\n", radeon_get_pcie_lanes(rdev));
 
 	return 0;
diff --git a/drivers/gpu/drm/radeon/radeon_reg.h b/drivers/gpu/drm/radeon/radeon_reg.h
index 5098634..bc44a3d 100644
--- a/drivers/gpu/drm/radeon/radeon_reg.h
+++ b/drivers/gpu/drm/radeon/radeon_reg.h
@@ -539,11 +539,9 @@
 
 #define RADEON_CRTC2_PITCH                  0x032c
 #define RADEON_CRTC_STATUS                  0x005c
-#       define RADEON_CRTC_VBLANK_CUR       (1 <<  0)
 #       define RADEON_CRTC_VBLANK_SAVE      (1 <<  1)
 #       define RADEON_CRTC_VBLANK_SAVE_CLEAR  (1 <<  1)
 #define RADEON_CRTC2_STATUS                  0x03fc
-#       define RADEON_CRTC2_VBLANK_CUR       (1 <<  0)
 #       define RADEON_CRTC2_VBLANK_SAVE      (1 <<  1)
 #       define RADEON_CRTC2_VBLANK_SAVE_CLEAR  (1 <<  1)
 #define RADEON_CRTC_V_SYNC_STRT_WID         0x020c
@@ -3297,7 +3295,7 @@
 #	define RADEON_RB_BUFSZ_MASK		(0x3f << 0)
 #	define RADEON_RB_BLKSZ_SHIFT		8
 #	define RADEON_RB_BLKSZ_MASK		(0x3f << 8)
-#	define RADEON_BUF_SWAP_32BIT		(2 << 16)
+#	define RADEON_BUF_SWAP_32BIT		(1 << 17)
 #	define RADEON_MAX_FETCH_SHIFT		18
 #	define RADEON_MAX_FETCH_MASK		(0x3 << 18)
 #	define RADEON_RB_NO_UPDATE		(1 << 27)
diff --git a/drivers/gpu/drm/radeon/radeon_ring.c b/drivers/gpu/drm/radeon/radeon_ring.c
index 3056620..08c0233 100644
--- a/drivers/gpu/drm/radeon/radeon_ring.c
+++ b/drivers/gpu/drm/radeon/radeon_ring.c
@@ -34,133 +34,93 @@
 #include "atom.h"
 
 int radeon_debugfs_ib_init(struct radeon_device *rdev);
-int radeon_debugfs_ring_init(struct radeon_device *rdev);
 
-u32 radeon_get_ib_value(struct radeon_cs_parser *p, int idx)
+void radeon_ib_bogus_cleanup(struct radeon_device *rdev)
 {
-	struct radeon_cs_chunk *ibc = &p->chunks[p->chunk_ib_idx];
-	u32 pg_idx, pg_offset;
-	u32 idx_value = 0;
-	int new_page;
-
-	pg_idx = (idx * 4) / PAGE_SIZE;
-	pg_offset = (idx * 4) % PAGE_SIZE;
-
-	if (ibc->kpage_idx[0] == pg_idx)
-		return ibc->kpage[0][pg_offset/4];
-	if (ibc->kpage_idx[1] == pg_idx)
-		return ibc->kpage[1][pg_offset/4];
-
-	new_page = radeon_cs_update_pages(p, pg_idx);
-	if (new_page < 0) {
-		p->parser_error = new_page;
-		return 0;
-	}
+	struct radeon_ib *ib, *n;
 
-	idx_value = ibc->kpage[new_page][pg_offset/4];
-	return idx_value;
+	list_for_each_entry_safe(ib, n, &rdev->ib_pool.bogus_ib, list) {
+		list_del(&ib->list);
+		vfree(ib->ptr);
+		kfree(ib);
+	}
 }
 
-void radeon_ring_write(struct radeon_ring *ring, uint32_t v)
+void radeon_ib_bogus_add(struct radeon_device *rdev, struct radeon_ib *ib)
 {
-#if DRM_DEBUG_CODE
-	if (ring->count_dw <= 0) {
-		DRM_ERROR("radeon: writting more dword to ring than expected !\n");
+	struct radeon_ib *bib;
+
+	bib = kmalloc(sizeof(*bib), GFP_KERNEL);
+	if (bib == NULL)
+		return;
+	bib->ptr = vmalloc(ib->length_dw * 4);
+	if (bib->ptr == NULL) {
+		kfree(bib);
+		return;
 	}
-#endif
-	ring->ring[ring->wptr++] = v;
-	ring->wptr &= ring->ptr_mask;
-	ring->count_dw--;
-	ring->ring_free_dw--;
+	memcpy(bib->ptr, ib->ptr, ib->length_dw * 4);
+	bib->length_dw = ib->length_dw;
+	mutex_lock(&rdev->ib_pool.mutex);
+	list_add_tail(&bib->list, &rdev->ib_pool.bogus_ib);
+	mutex_unlock(&rdev->ib_pool.mutex);
 }
 
 /*
  * IB.
  */
-bool radeon_ib_try_free(struct radeon_device *rdev, struct radeon_ib *ib)
-{
-	bool done = false;
-
-	/* only free ib which have been emited */
-	if (ib->fence && ib->fence->emitted) {
-		if (radeon_fence_signaled(ib->fence)) {
-			radeon_fence_unref(&ib->fence);
-			radeon_sa_bo_free(rdev, &ib->sa_bo);
-			done = true;
-		}
-	}
-	return done;
-}
-
-int radeon_ib_get(struct radeon_device *rdev, int ring,
-		  struct radeon_ib **ib, unsigned size)
+int radeon_ib_get(struct radeon_device *rdev, struct radeon_ib **ib)
 {
 	struct radeon_fence *fence;
-	unsigned cretry = 0;
-	int r = 0, i, idx;
+	struct radeon_ib *nib;
+	int r = 0, i, c;
 
 	*ib = NULL;
-	/* align size on 256 bytes */
-	size = ALIGN(size, 256);
-
-	r = radeon_fence_create(rdev, &fence, ring);
+	r = radeon_fence_create(rdev, &fence);
 	if (r) {
 		dev_err(rdev->dev, "failed to create fence for new IB\n");
 		return r;
 	}
-
-	radeon_mutex_lock(&rdev->ib_pool.mutex);
-	idx = rdev->ib_pool.head_id;
-retry:
-	if (cretry > 5) {
-		dev_err(rdev->dev, "failed to get an ib after 5 retry\n");
-		radeon_mutex_unlock(&rdev->ib_pool.mutex);
-		radeon_fence_unref(&fence);
-		return -ENOMEM;
-	}
-	cretry++;
-	for (i = 0; i < RADEON_IB_POOL_SIZE; i++) {
-		radeon_ib_try_free(rdev, &rdev->ib_pool.ibs[idx]);
-		if (rdev->ib_pool.ibs[idx].fence == NULL) {
-			r = radeon_sa_bo_new(rdev, &rdev->ib_pool.sa_manager,
-					     &rdev->ib_pool.ibs[idx].sa_bo,
-					     size, 256);
-			if (!r) {
-				*ib = &rdev->ib_pool.ibs[idx];
-				(*ib)->ptr = rdev->ib_pool.sa_manager.cpu_ptr;
-				(*ib)->ptr += ((*ib)->sa_bo.offset >> 2);
-				(*ib)->gpu_addr = rdev->ib_pool.sa_manager.gpu_addr;
-				(*ib)->gpu_addr += (*ib)->sa_bo.offset;
-				(*ib)->fence = fence;
-				(*ib)->vm_id = 0;
-				/* ib are most likely to be allocated in a ring fashion
-				 * thus rdev->ib_pool.head_id should be the id of the
-				 * oldest ib
-				 */
-				rdev->ib_pool.head_id = (1 + idx);
-				rdev->ib_pool.head_id &= (RADEON_IB_POOL_SIZE - 1);
-				radeon_mutex_unlock(&rdev->ib_pool.mutex);
-				return 0;
-			}
+	mutex_lock(&rdev->ib_pool.mutex);
+	for (i = rdev->ib_pool.head_id, c = 0, nib = NULL; c < RADEON_IB_POOL_SIZE; c++, i++) {
+		i &= (RADEON_IB_POOL_SIZE - 1);
+		if (rdev->ib_pool.ibs[i].free) {
+			nib = &rdev->ib_pool.ibs[i];
+			break;
 		}
-		idx = (idx + 1) & (RADEON_IB_POOL_SIZE - 1);
 	}
-	/* this should be rare event, ie all ib scheduled none signaled yet.
-	 */
-	for (i = 0; i < RADEON_IB_POOL_SIZE; i++) {
-		if (rdev->ib_pool.ibs[idx].fence && rdev->ib_pool.ibs[idx].fence->emitted) {
-			r = radeon_fence_wait(rdev->ib_pool.ibs[idx].fence, false);
-			if (!r) {
-				goto retry;
-			}
-			/* an error happened */
-			break;
+	if (nib == NULL) {
+		/* This should never happen, it means we allocated all
+		 * IB and haven't scheduled one yet, return EBUSY to
+		 * userspace hoping that on ioctl recall we get better
+		 * luck
+		 */
+		dev_err(rdev->dev, "no free indirect buffer !\n");
+		mutex_unlock(&rdev->ib_pool.mutex);
+		radeon_fence_unref(&fence);
+		return -EBUSY;
+	}
+	rdev->ib_pool.head_id = (nib->idx + 1) & (RADEON_IB_POOL_SIZE - 1);
+	nib->free = false;
+	if (nib->fence) {
+		mutex_unlock(&rdev->ib_pool.mutex);
+		r = radeon_fence_wait(nib->fence, false);
+		if (r) {
+			dev_err(rdev->dev, "error waiting fence of IB(%u:0x%016lX:%u)\n",
+				nib->idx, (unsigned long)nib->gpu_addr, nib->length_dw);
+			mutex_lock(&rdev->ib_pool.mutex);
+			nib->free = true;
+			mutex_unlock(&rdev->ib_pool.mutex);
+			radeon_fence_unref(&fence);
+			return r;
 		}
-		idx = (idx + 1) & (RADEON_IB_POOL_SIZE - 1);
+		mutex_lock(&rdev->ib_pool.mutex);
 	}
-	radeon_mutex_unlock(&rdev->ib_pool.mutex);
-	radeon_fence_unref(&fence);
-	return r;
+	radeon_fence_unref(&nib->fence);
+	nib->fence = fence;
+	nib->length_dw = 0;
+	mutex_unlock(&rdev->ib_pool.mutex);
+	*ib = nib;
+	return 0;
 }
 
 void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib **ib)
@@ -171,258 +131,247 @@ void radeon_ib_free(struct radeon_device *rdev, struct radeon_ib **ib)
 	if (tmp == NULL) {
 		return;
 	}
-	radeon_mutex_lock(&rdev->ib_pool.mutex);
-	if (tmp->fence && !tmp->fence->emitted) {
-		radeon_sa_bo_free(rdev, &tmp->sa_bo);
+	if (!tmp->fence->emited)
 		radeon_fence_unref(&tmp->fence);
-	}
-	radeon_mutex_unlock(&rdev->ib_pool.mutex);
+	mutex_lock(&rdev->ib_pool.mutex);
+	tmp->free = true;
+	mutex_unlock(&rdev->ib_pool.mutex);
 }
 
 int radeon_ib_schedule(struct radeon_device *rdev, struct radeon_ib *ib)
 {
-	struct radeon_ring *ring = &rdev->ring[ib->fence->ring];
 	int r = 0;
 
-	if (!ib->length_dw || !ring->ready) {
+	if (!ib->length_dw || !rdev->cp.ready) {
 		/* TODO: Nothings in the ib we should report. */
 		DRM_ERROR("radeon: couldn't schedule IB(%u).\n", ib->idx);
 		return -EINVAL;
 	}
 
 	/* 64 dwords should be enough for fence too */
-	r = radeon_ring_lock(rdev, ring, 64);
+	r = radeon_ring_lock(rdev, 64);
 	if (r) {
 		DRM_ERROR("radeon: scheduling IB failed (%d).\n", r);
 		return r;
 	}
-	radeon_ring_ib_execute(rdev, ib->fence->ring, ib);
+	radeon_ring_ib_execute(rdev, ib);
 	radeon_fence_emit(rdev, ib->fence);
-	radeon_ring_unlock_commit(rdev, ring);
+	mutex_lock(&rdev->ib_pool.mutex);
+	/* once scheduled IB is considered free and protected by the fence */
+	ib->free = true;
+	mutex_unlock(&rdev->ib_pool.mutex);
+	radeon_ring_unlock_commit(rdev);
 	return 0;
 }
 
 int radeon_ib_pool_init(struct radeon_device *rdev)
 {
-	struct radeon_sa_manager tmp;
-	int i, r;
+	void *ptr;
+	uint64_t gpu_addr;
+	int i;
+	int r = 0;
 
-	r = radeon_sa_bo_manager_init(rdev, &tmp,
-				      RADEON_IB_POOL_SIZE*64*1024,
-				      RADEON_GEM_DOMAIN_GTT);
+	if (rdev->ib_pool.robj)
+		return 0;
+	INIT_LIST_HEAD(&rdev->ib_pool.bogus_ib);
+	/* Allocate 1M object buffer */
+	r = radeon_bo_create(rdev, RADEON_IB_POOL_SIZE*64*1024,
+			     PAGE_SIZE, true, RADEON_GEM_DOMAIN_GTT,
+			     &rdev->ib_pool.robj);
 	if (r) {
+		DRM_ERROR("radeon: failed to ib pool (%d).\n", r);
 		return r;
 	}
-
-	radeon_mutex_lock(&rdev->ib_pool.mutex);
-	if (rdev->ib_pool.ready) {
-		radeon_mutex_unlock(&rdev->ib_pool.mutex);
-		radeon_sa_bo_manager_fini(rdev, &tmp);
-		return 0;
+	r = radeon_bo_reserve(rdev->ib_pool.robj, false);
+	if (unlikely(r != 0))
+		return r;
+	r = radeon_bo_pin(rdev->ib_pool.robj, RADEON_GEM_DOMAIN_GTT, &gpu_addr);
+	if (r) {
+		radeon_bo_unreserve(rdev->ib_pool.robj);
+		DRM_ERROR("radeon: failed to pin ib pool (%d).\n", r);
+		return r;
+	}
+	r = radeon_bo_kmap(rdev->ib_pool.robj, &ptr);
+	radeon_bo_unreserve(rdev->ib_pool.robj);
+	if (r) {
+		DRM_ERROR("radeon: failed to map ib pool (%d).\n", r);
+		return r;
 	}
-
-	rdev->ib_pool.sa_manager = tmp;
-	INIT_LIST_HEAD(&rdev->ib_pool.sa_manager.sa_bo);
 	for (i = 0; i < RADEON_IB_POOL_SIZE; i++) {
-		rdev->ib_pool.ibs[i].fence = NULL;
+		unsigned offset;
+
+		offset = i * 64 * 1024;
+		rdev->ib_pool.ibs[i].gpu_addr = gpu_addr + offset;
+		rdev->ib_pool.ibs[i].ptr = ptr + offset;
 		rdev->ib_pool.ibs[i].idx = i;
 		rdev->ib_pool.ibs[i].length_dw = 0;
-		INIT_LIST_HEAD(&rdev->ib_pool.ibs[i].sa_bo.list);
+		rdev->ib_pool.ibs[i].free = true;
 	}
 	rdev->ib_pool.head_id = 0;
 	rdev->ib_pool.ready = true;
 	DRM_INFO("radeon: ib pool ready.\n");
-
 	if (radeon_debugfs_ib_init(rdev)) {
 		DRM_ERROR("Failed to register debugfs file for IB !\n");
 	}
-	if (radeon_debugfs_ring_init(rdev)) {
-		DRM_ERROR("Failed to register debugfs file for rings !\n");
-	}
-	radeon_mutex_unlock(&rdev->ib_pool.mutex);
-	return 0;
+	return r;
 }
 
 void radeon_ib_pool_fini(struct radeon_device *rdev)
 {
-	unsigned i;
+	int r;
+	struct radeon_bo *robj;
 
-	radeon_mutex_lock(&rdev->ib_pool.mutex);
-	if (rdev->ib_pool.ready) {
-		for (i = 0; i < RADEON_IB_POOL_SIZE; i++) {
-			radeon_sa_bo_free(rdev, &rdev->ib_pool.ibs[i].sa_bo);
-			radeon_fence_unref(&rdev->ib_pool.ibs[i].fence);
+	if (!rdev->ib_pool.ready) {
+		return;
+	}
+	mutex_lock(&rdev->ib_pool.mutex);
+	radeon_ib_bogus_cleanup(rdev);
+	robj = rdev->ib_pool.robj;
+	rdev->ib_pool.robj = NULL;
+	mutex_unlock(&rdev->ib_pool.mutex);
+
+	if (robj) {
+		r = radeon_bo_reserve(robj, false);
+		if (likely(r == 0)) {
+			radeon_bo_kunmap(robj);
+			radeon_bo_unpin(robj);
+			radeon_bo_unreserve(robj);
 		}
-		radeon_sa_bo_manager_fini(rdev, &rdev->ib_pool.sa_manager);
-		rdev->ib_pool.ready = false;
+		radeon_bo_unref(&robj);
 	}
-	radeon_mutex_unlock(&rdev->ib_pool.mutex);
 }
 
-int radeon_ib_pool_start(struct radeon_device *rdev)
-{
-	return radeon_sa_bo_manager_start(rdev, &rdev->ib_pool.sa_manager);
-}
-
-int radeon_ib_pool_suspend(struct radeon_device *rdev)
-{
-	return radeon_sa_bo_manager_suspend(rdev, &rdev->ib_pool.sa_manager);
-}
 
 /*
  * Ring.
  */
-int radeon_ring_index(struct radeon_device *rdev, struct radeon_ring *ring)
-{
-	/* r1xx-r5xx only has CP ring */
-	if (rdev->family < CHIP_R600)
-		return RADEON_RING_TYPE_GFX_INDEX;
-
-	if (rdev->family >= CHIP_CAYMAN) {
-		if (ring == &rdev->ring[CAYMAN_RING_TYPE_CP1_INDEX])
-			return CAYMAN_RING_TYPE_CP1_INDEX;
-		else if (ring == &rdev->ring[CAYMAN_RING_TYPE_CP2_INDEX])
-			return CAYMAN_RING_TYPE_CP2_INDEX;
-	}
-	return RADEON_RING_TYPE_GFX_INDEX;
-}
-
-void radeon_ring_free_size(struct radeon_device *rdev, struct radeon_ring *ring)
+void radeon_ring_free_size(struct radeon_device *rdev)
 {
-	u32 rptr;
-
 	if (rdev->wb.enabled)
-		rptr = le32_to_cpu(rdev->wb.wb[ring->rptr_offs/4]);
-	else
-		rptr = RREG32(ring->rptr_reg);
-	ring->rptr = (rptr & ring->ptr_reg_mask) >> ring->ptr_reg_shift;
+		rdev->cp.rptr = le32_to_cpu(rdev->wb.wb[RADEON_WB_CP_RPTR_OFFSET/4]);
+	else {
+		if (rdev->family >= CHIP_R600)
+			rdev->cp.rptr = RREG32(R600_CP_RB_RPTR);
+		else
+			rdev->cp.rptr = RREG32(RADEON_CP_RB_RPTR);
+	}
 	/* This works because ring_size is a power of 2 */
-	ring->ring_free_dw = (ring->rptr + (ring->ring_size / 4));
-	ring->ring_free_dw -= ring->wptr;
-	ring->ring_free_dw &= ring->ptr_mask;
-	if (!ring->ring_free_dw) {
-		ring->ring_free_dw = ring->ring_size / 4;
+	rdev->cp.ring_free_dw = (rdev->cp.rptr + (rdev->cp.ring_size / 4));
+	rdev->cp.ring_free_dw -= rdev->cp.wptr;
+	rdev->cp.ring_free_dw &= rdev->cp.ptr_mask;
+	if (!rdev->cp.ring_free_dw) {
+		rdev->cp.ring_free_dw = rdev->cp.ring_size / 4;
 	}
 }
 
-
-int radeon_ring_alloc(struct radeon_device *rdev, struct radeon_ring *ring, unsigned ndw)
+int radeon_ring_alloc(struct radeon_device *rdev, unsigned ndw)
 {
 	int r;
 
 	/* Align requested size with padding so unlock_commit can
 	 * pad safely */
-	ndw = (ndw + ring->align_mask) & ~ring->align_mask;
-	while (ndw > (ring->ring_free_dw - 1)) {
-		radeon_ring_free_size(rdev, ring);
-		if (ndw < ring->ring_free_dw) {
+	ndw = (ndw + rdev->cp.align_mask) & ~rdev->cp.align_mask;
+	while (ndw > (rdev->cp.ring_free_dw - 1)) {
+		radeon_ring_free_size(rdev);
+		if (ndw < rdev->cp.ring_free_dw) {
 			break;
 		}
-		r = radeon_fence_wait_next(rdev, radeon_ring_index(rdev, ring));
+		r = radeon_fence_wait_next(rdev);
 		if (r)
 			return r;
 	}
-	ring->count_dw = ndw;
-	ring->wptr_old = ring->wptr;
+	rdev->cp.count_dw = ndw;
+	rdev->cp.wptr_old = rdev->cp.wptr;
 	return 0;
 }
 
-int radeon_ring_lock(struct radeon_device *rdev, struct radeon_ring *ring, unsigned ndw)
+int radeon_ring_lock(struct radeon_device *rdev, unsigned ndw)
 {
 	int r;
 
-	mutex_lock(&ring->mutex);
-	r = radeon_ring_alloc(rdev, ring, ndw);
+	mutex_lock(&rdev->cp.mutex);
+	r = radeon_ring_alloc(rdev, ndw);
 	if (r) {
-		mutex_unlock(&ring->mutex);
+		mutex_unlock(&rdev->cp.mutex);
 		return r;
 	}
 	return 0;
 }
 
-void radeon_ring_commit(struct radeon_device *rdev, struct radeon_ring *ring)
+void radeon_ring_commit(struct radeon_device *rdev)
 {
 	unsigned count_dw_pad;
 	unsigned i;
 
 	/* We pad to match fetch size */
-	count_dw_pad = (ring->align_mask + 1) -
-		       (ring->wptr & ring->align_mask);
+	count_dw_pad = (rdev->cp.align_mask + 1) -
+		       (rdev->cp.wptr & rdev->cp.align_mask);
 	for (i = 0; i < count_dw_pad; i++) {
-		radeon_ring_write(ring, ring->nop);
+		radeon_ring_write(rdev, 2 << 30);
 	}
 	DRM_MEMORYBARRIER();
-	WREG32(ring->wptr_reg, (ring->wptr << ring->ptr_reg_shift) & ring->ptr_reg_mask);
-	(void)RREG32(ring->wptr_reg);
+	radeon_cp_commit(rdev);
 }
 
-void radeon_ring_unlock_commit(struct radeon_device *rdev, struct radeon_ring *ring)
+void radeon_ring_unlock_commit(struct radeon_device *rdev)
 {
-	radeon_ring_commit(rdev, ring);
-	mutex_unlock(&ring->mutex);
+	radeon_ring_commit(rdev);
+	mutex_unlock(&rdev->cp.mutex);
 }
 
-void radeon_ring_unlock_undo(struct radeon_device *rdev, struct radeon_ring *ring)
+void radeon_ring_unlock_undo(struct radeon_device *rdev)
 {
-	ring->wptr = ring->wptr_old;
-	mutex_unlock(&ring->mutex);
+	rdev->cp.wptr = rdev->cp.wptr_old;
+	mutex_unlock(&rdev->cp.mutex);
 }
 
-int radeon_ring_init(struct radeon_device *rdev, struct radeon_ring *ring, unsigned ring_size,
-		     unsigned rptr_offs, unsigned rptr_reg, unsigned wptr_reg,
-		     u32 ptr_reg_shift, u32 ptr_reg_mask, u32 nop)
+int radeon_ring_init(struct radeon_device *rdev, unsigned ring_size)
 {
 	int r;
 
-	ring->ring_size = ring_size;
-	ring->rptr_offs = rptr_offs;
-	ring->rptr_reg = rptr_reg;
-	ring->wptr_reg = wptr_reg;
-	ring->ptr_reg_shift = ptr_reg_shift;
-	ring->ptr_reg_mask = ptr_reg_mask;
-	ring->nop = nop;
+	rdev->cp.ring_size = ring_size;
 	/* Allocate ring buffer */
-	if (ring->ring_obj == NULL) {
-		r = radeon_bo_create(rdev, ring->ring_size, PAGE_SIZE, true,
+	if (rdev->cp.ring_obj == NULL) {
+		r = radeon_bo_create(rdev, rdev->cp.ring_size, PAGE_SIZE, true,
 					RADEON_GEM_DOMAIN_GTT,
-					&ring->ring_obj);
+					&rdev->cp.ring_obj);
 		if (r) {
 			dev_err(rdev->dev, "(%d) ring create failed\n", r);
 			return r;
 		}
-		r = radeon_bo_reserve(ring->ring_obj, false);
+		r = radeon_bo_reserve(rdev->cp.ring_obj, false);
 		if (unlikely(r != 0))
 			return r;
-		r = radeon_bo_pin(ring->ring_obj, RADEON_GEM_DOMAIN_GTT,
-					&ring->gpu_addr);
+		r = radeon_bo_pin(rdev->cp.ring_obj, RADEON_GEM_DOMAIN_GTT,
+					&rdev->cp.gpu_addr);
 		if (r) {
-			radeon_bo_unreserve(ring->ring_obj);
+			radeon_bo_unreserve(rdev->cp.ring_obj);
 			dev_err(rdev->dev, "(%d) ring pin failed\n", r);
 			return r;
 		}
-		r = radeon_bo_kmap(ring->ring_obj,
-				       (void **)&ring->ring);
-		radeon_bo_unreserve(ring->ring_obj);
+		r = radeon_bo_kmap(rdev->cp.ring_obj,
+				       (void **)&rdev->cp.ring);
+		radeon_bo_unreserve(rdev->cp.ring_obj);
 		if (r) {
 			dev_err(rdev->dev, "(%d) ring map failed\n", r);
 			return r;
 		}
 	}
-	ring->ptr_mask = (ring->ring_size / 4) - 1;
-	ring->ring_free_dw = ring->ring_size / 4;
+	rdev->cp.ptr_mask = (rdev->cp.ring_size / 4) - 1;
+	rdev->cp.ring_free_dw = rdev->cp.ring_size / 4;
 	return 0;
 }
 
-void radeon_ring_fini(struct radeon_device *rdev, struct radeon_ring *ring)
+void radeon_ring_fini(struct radeon_device *rdev)
 {
 	int r;
 	struct radeon_bo *ring_obj;
 
-	mutex_lock(&ring->mutex);
-	ring_obj = ring->ring_obj;
-	ring->ring = NULL;
-	ring->ring_obj = NULL;
-	mutex_unlock(&ring->mutex);
+	mutex_lock(&rdev->cp.mutex);
+	ring_obj = rdev->cp.ring_obj;
+	rdev->cp.ring = NULL;
+	rdev->cp.ring_obj = NULL;
+	mutex_unlock(&rdev->cp.mutex);
 
 	if (ring_obj) {
 		r = radeon_bo_reserve(ring_obj, false);
@@ -435,96 +384,78 @@ void radeon_ring_fini(struct radeon_device *rdev, struct radeon_ring *ring)
 	}
 }
 
+
 /*
  * Debugfs info
  */
 #if defined(CONFIG_DEBUG_FS)
-
-static int radeon_debugfs_ring_info(struct seq_file *m, void *data)
+static int radeon_debugfs_ib_info(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
-	struct drm_device *dev = node->minor->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	int ridx = *(int*)node->info_ent->data;
-	struct radeon_ring *ring = &rdev->ring[ridx];
-	unsigned count, i, j;
-
-	radeon_ring_free_size(rdev, ring);
-	count = (ring->ring_size / 4) - ring->ring_free_dw;
-	seq_printf(m, "wptr(0x%04x): 0x%08x\n", ring->wptr_reg, RREG32(ring->wptr_reg));
-	seq_printf(m, "rptr(0x%04x): 0x%08x\n", ring->rptr_reg, RREG32(ring->rptr_reg));
-	seq_printf(m, "driver's copy of the wptr: 0x%08x\n", ring->wptr);
-	seq_printf(m, "driver's copy of the rptr: 0x%08x\n", ring->rptr);
-	seq_printf(m, "%u free dwords in ring\n", ring->ring_free_dw);
-	seq_printf(m, "%u dwords in ring\n", count);
-	i = ring->rptr;
-	for (j = 0; j <= count; j++) {
-		seq_printf(m, "r[%04d]=0x%08x\n", i, ring->ring[i]);
-		i = (i + 1) & ring->ptr_mask;
+	struct radeon_ib *ib = node->info_ent->data;
+	unsigned i;
+
+	if (ib == NULL) {
+		return 0;
+	}
+	seq_printf(m, "IB %04u\n", ib->idx);
+	seq_printf(m, "IB fence %p\n", ib->fence);
+	seq_printf(m, "IB size %05u dwords\n", ib->length_dw);
+	for (i = 0; i < ib->length_dw; i++) {
+		seq_printf(m, "[%05u]=0x%08X\n", i, ib->ptr[i]);
 	}
 	return 0;
 }
 
-static int radeon_ring_type_gfx_index = RADEON_RING_TYPE_GFX_INDEX;
-static int cayman_ring_type_cp1_index = CAYMAN_RING_TYPE_CP1_INDEX;
-static int cayman_ring_type_cp2_index = CAYMAN_RING_TYPE_CP2_INDEX;
-
-static struct drm_info_list radeon_debugfs_ring_info_list[] = {
-	{"radeon_ring_gfx", radeon_debugfs_ring_info, 0, &radeon_ring_type_gfx_index},
-	{"radeon_ring_cp1", radeon_debugfs_ring_info, 0, &cayman_ring_type_cp1_index},
-	{"radeon_ring_cp2", radeon_debugfs_ring_info, 0, &cayman_ring_type_cp2_index},
-};
-
-static int radeon_debugfs_ib_info(struct seq_file *m, void *data)
+static int radeon_debugfs_ib_bogus_info(struct seq_file *m, void *data)
 {
 	struct drm_info_node *node = (struct drm_info_node *) m->private;
-	struct drm_device *dev = node->minor->dev;
-	struct radeon_device *rdev = dev->dev_private;
-	struct radeon_ib *ib = &rdev->ib_pool.ibs[*((unsigned*)node->info_ent->data)];
+	struct radeon_device *rdev = node->info_ent->data;
+	struct radeon_ib *ib;
 	unsigned i;
 
-	if (ib == NULL) {
+	mutex_lock(&rdev->ib_pool.mutex);
+	if (list_empty(&rdev->ib_pool.bogus_ib)) {
+		mutex_unlock(&rdev->ib_pool.mutex);
+		seq_printf(m, "no bogus IB recorded\n");
 		return 0;
 	}
-	seq_printf(m, "IB %04u\n", ib->idx);
-	seq_printf(m, "IB fence %p\n", ib->fence);
+	ib = list_first_entry(&rdev->ib_pool.bogus_ib, struct radeon_ib, list);
+	list_del_init(&ib->list);
+	mutex_unlock(&rdev->ib_pool.mutex);
 	seq_printf(m, "IB size %05u dwords\n", ib->length_dw);
 	for (i = 0; i < ib->length_dw; i++) {
 		seq_printf(m, "[%05u]=0x%08X\n", i, ib->ptr[i]);
 	}
+	vfree(ib->ptr);
+	kfree(ib);
 	return 0;
 }
 
 static struct drm_info_list radeon_debugfs_ib_list[RADEON_IB_POOL_SIZE];
 static char radeon_debugfs_ib_names[RADEON_IB_POOL_SIZE][32];
-static unsigned radeon_debugfs_ib_idx[RADEON_IB_POOL_SIZE];
-#endif
 
-int radeon_debugfs_ring_init(struct radeon_device *rdev)
-{
-#if defined(CONFIG_DEBUG_FS)
-	if (rdev->family >= CHIP_CAYMAN)
-		return radeon_debugfs_add_files(rdev, radeon_debugfs_ring_info_list,
-						ARRAY_SIZE(radeon_debugfs_ring_info_list));
-	else
-		return radeon_debugfs_add_files(rdev, radeon_debugfs_ring_info_list, 1);
-#else
-	return 0;
+static struct drm_info_list radeon_debugfs_ib_bogus_info_list[] = {
+	{"radeon_ib_bogus", radeon_debugfs_ib_bogus_info, 0, NULL},
+};
 #endif
-}
 
 int radeon_debugfs_ib_init(struct radeon_device *rdev)
 {
 #if defined(CONFIG_DEBUG_FS)
 	unsigned i;
+	int r;
 
+	radeon_debugfs_ib_bogus_info_list[0].data = rdev;
+	r = radeon_debugfs_add_files(rdev, radeon_debugfs_ib_bogus_info_list, 1);
+	if (r)
+		return r;
 	for (i = 0; i < RADEON_IB_POOL_SIZE; i++) {
 		sprintf(radeon_debugfs_ib_names[i], "radeon_ib_%04u", i);
-		radeon_debugfs_ib_idx[i] = i;
 		radeon_debugfs_ib_list[i].name = radeon_debugfs_ib_names[i];
 		radeon_debugfs_ib_list[i].show = &radeon_debugfs_ib_info;
 		radeon_debugfs_ib_list[i].driver_features = 0;
-		radeon_debugfs_ib_list[i].data = &radeon_debugfs_ib_idx[i];
+		radeon_debugfs_ib_list[i].data = &rdev->ib_pool.ibs[i];
 	}
 	return radeon_debugfs_add_files(rdev, radeon_debugfs_ib_list,
 					RADEON_IB_POOL_SIZE);
diff --git a/drivers/gpu/drm/radeon/radeon_sa.c b/drivers/gpu/drm/radeon/radeon_sa.c
deleted file mode 100644
index 4cce47e..0000000
--- a/drivers/gpu/drm/radeon/radeon_sa.c
+++ /dev/null
@@ -1,189 +0,0 @@
-/*
- * Copyright 2011 Red Hat Inc.
- * All Rights Reserved.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the
- * "Software"), to deal in the Software without restriction, including
- * without limitation the rights to use, copy, modify, merge, publish,
- * distribute, sub license, and/or sell copies of the Software, and to
- * permit persons to whom the Software is furnished to do so, subject to
- * the following conditions:
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
- * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
- * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
- * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
- * USE OR OTHER DEALINGS IN THE SOFTWARE.
- *
- * The above copyright notice and this permission notice (including the
- * next paragraph) shall be included in all copies or substantial portions
- * of the Software.
- *
- */
-/*
- * Authors:
- *    Jerome Glisse <glisse@freedesktop.org>
- */
-#include "drmP.h"
-#include "drm.h"
-#include "radeon.h"
-
-int radeon_sa_bo_manager_init(struct radeon_device *rdev,
-			      struct radeon_sa_manager *sa_manager,
-			      unsigned size, u32 domain)
-{
-	int r;
-
-	sa_manager->bo = NULL;
-	sa_manager->size = size;
-	sa_manager->domain = domain;
-	INIT_LIST_HEAD(&sa_manager->sa_bo);
-
-	r = radeon_bo_create(rdev, size, RADEON_GPU_PAGE_SIZE, true,
-			     RADEON_GEM_DOMAIN_CPU, &sa_manager->bo);
-	if (r) {
-		dev_err(rdev->dev, "(%d) failed to allocate bo for manager\n", r);
-		return r;
-	}
-
-	return r;
-}
-
-void radeon_sa_bo_manager_fini(struct radeon_device *rdev,
-			       struct radeon_sa_manager *sa_manager)
-{
-	struct radeon_sa_bo *sa_bo, *tmp;
-
-	if (!list_empty(&sa_manager->sa_bo)) {
-		dev_err(rdev->dev, "sa_manager is not empty, clearing anyway\n");
-	}
-	list_for_each_entry_safe(sa_bo, tmp, &sa_manager->sa_bo, list) {
-		list_del_init(&sa_bo->list);
-	}
-	radeon_bo_unref(&sa_manager->bo);
-	sa_manager->size = 0;
-}
-
-int radeon_sa_bo_manager_start(struct radeon_device *rdev,
-			       struct radeon_sa_manager *sa_manager)
-{
-	int r;
-
-	if (sa_manager->bo == NULL) {
-		dev_err(rdev->dev, "no bo for sa manager\n");
-		return -EINVAL;
-	}
-
-	/* map the buffer */
-	r = radeon_bo_reserve(sa_manager->bo, false);
-	if (r) {
-		dev_err(rdev->dev, "(%d) failed to reserve manager bo\n", r);
-		return r;
-	}
-	r = radeon_bo_pin(sa_manager->bo, sa_manager->domain, &sa_manager->gpu_addr);
-	if (r) {
-		radeon_bo_unreserve(sa_manager->bo);
-		dev_err(rdev->dev, "(%d) failed to pin manager bo\n", r);
-		return r;
-	}
-	r = radeon_bo_kmap(sa_manager->bo, &sa_manager->cpu_ptr);
-	radeon_bo_unreserve(sa_manager->bo);
-	return r;
-}
-
-int radeon_sa_bo_manager_suspend(struct radeon_device *rdev,
-				 struct radeon_sa_manager *sa_manager)
-{
-	int r;
-
-	if (sa_manager->bo == NULL) {
-		dev_err(rdev->dev, "no bo for sa manager\n");
-		return -EINVAL;
-	}
-
-	r = radeon_bo_reserve(sa_manager->bo, false);
-	if (!r) {
-		radeon_bo_kunmap(sa_manager->bo);
-		radeon_bo_unpin(sa_manager->bo);
-		radeon_bo_unreserve(sa_manager->bo);
-	}
-	return r;
-}
-
-/*
- * Principe is simple, we keep a list of sub allocation in offset
- * order (first entry has offset == 0, last entry has the highest
- * offset).
- *
- * When allocating new object we first check if there is room at
- * the end total_size - (last_object_offset + last_object_size) >=
- * alloc_size. If so we allocate new object there.
- *
- * When there is not enough room at the end, we start waiting for
- * each sub object until we reach object_offset+object_size >=
- * alloc_size, this object then become the sub object we return.
- *
- * Alignment can't be bigger than page size
- */
-int radeon_sa_bo_new(struct radeon_device *rdev,
-		     struct radeon_sa_manager *sa_manager,
-		     struct radeon_sa_bo *sa_bo,
-		     unsigned size, unsigned align)
-{
-	struct radeon_sa_bo *tmp;
-	struct list_head *head;
-	unsigned offset = 0, wasted = 0;
-
-	BUG_ON(align > RADEON_GPU_PAGE_SIZE);
-	BUG_ON(size > sa_manager->size);
-
-	/* no one ? */
-	head = sa_manager->sa_bo.prev;
-	if (list_empty(&sa_manager->sa_bo)) {
-		goto out;
-	}
-
-	/* look for a hole big enough */
-	offset = 0;
-	list_for_each_entry(tmp, &sa_manager->sa_bo, list) {
-		/* room before this object ? */
-		if ((tmp->offset - offset) >= size) {
-			head = tmp->list.prev;
-			goto out;
-		}
-		offset = tmp->offset + tmp->size;
-		wasted = offset % align;
-		if (wasted) {
-			wasted = align - wasted;
-		}
-		offset += wasted;
-	}
-	/* room at the end ? */
-	head = sa_manager->sa_bo.prev;
-	tmp = list_entry(head, struct radeon_sa_bo, list);
-	offset = tmp->offset + tmp->size;
-	wasted = offset % align;
-	if (wasted) {
-		wasted = align - wasted;
-	}
-	offset += wasted;
-	if ((sa_manager->size - offset) < size) {
-		/* failed to find somethings big enough */
-		return -ENOMEM;
-	}
-
-out:
-	sa_bo->manager = sa_manager;
-	sa_bo->offset = offset;
-	sa_bo->size = size;
-	list_add(&sa_bo->list, head);
-	return 0;
-}
-
-void radeon_sa_bo_free(struct radeon_device *rdev, struct radeon_sa_bo *sa_bo)
-{
-	list_del_init(&sa_bo->list);
-}
diff --git a/drivers/gpu/drm/radeon/radeon_semaphore.c b/drivers/gpu/drm/radeon/radeon_semaphore.c
deleted file mode 100644
index 61dd4e3..0000000
--- a/drivers/gpu/drm/radeon/radeon_semaphore.c
+++ /dev/null
@@ -1,178 +0,0 @@
-/*
- * Copyright 2011 Christian König.
- * All Rights Reserved.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the
- * "Software"), to deal in the Software without restriction, including
- * without limitation the rights to use, copy, modify, merge, publish,
- * distribute, sub license, and/or sell copies of the Software, and to
- * permit persons to whom the Software is furnished to do so, subject to
- * the following conditions:
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
- * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
- * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
- * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
- * USE OR OTHER DEALINGS IN THE SOFTWARE.
- *
- * The above copyright notice and this permission notice (including the
- * next paragraph) shall be included in all copies or substantial portions
- * of the Software.
- *
- */
-/*
- * Authors:
- *    Christian König <deathsimple@vodafone.de>
- */
-#include "drmP.h"
-#include "drm.h"
-#include "radeon.h"
-
-static int radeon_semaphore_add_bo(struct radeon_device *rdev)
-{
-	struct radeon_semaphore_bo *bo;
-	unsigned long irq_flags;
-	uint64_t gpu_addr;
-	uint32_t *cpu_ptr;
-	int r, i;
-
-
-	bo = kmalloc(sizeof(struct radeon_semaphore_bo), GFP_KERNEL);
-	if (bo == NULL) {
-		return -ENOMEM;
-	}
-	INIT_LIST_HEAD(&bo->free);
-	INIT_LIST_HEAD(&bo->list);
-	bo->nused = 0;
-
-	r = radeon_ib_get(rdev, 0, &bo->ib, RADEON_SEMAPHORE_BO_SIZE);
-	if (r) {
-		dev_err(rdev->dev, "failed to get a bo after 5 retry\n");
-		kfree(bo);
-		return r;
-	}
-	gpu_addr = rdev->ib_pool.sa_manager.gpu_addr;
-	gpu_addr += bo->ib->sa_bo.offset;
-	cpu_ptr = rdev->ib_pool.sa_manager.cpu_ptr;
-	cpu_ptr += (bo->ib->sa_bo.offset >> 2);
-	for (i = 0; i < (RADEON_SEMAPHORE_BO_SIZE/8); i++) {
-		bo->semaphores[i].gpu_addr = gpu_addr;
-		bo->semaphores[i].cpu_ptr = cpu_ptr;
-		bo->semaphores[i].bo = bo;
-		list_add_tail(&bo->semaphores[i].list, &bo->free);
-		gpu_addr += 8;
-		cpu_ptr += 2;
-	}
-	write_lock_irqsave(&rdev->semaphore_drv.lock, irq_flags);
-	list_add_tail(&bo->list, &rdev->semaphore_drv.bo);
-	write_unlock_irqrestore(&rdev->semaphore_drv.lock, irq_flags);
-	return 0;
-}
-
-static void radeon_semaphore_del_bo_locked(struct radeon_device *rdev,
-					   struct radeon_semaphore_bo *bo)
-{
-	radeon_sa_bo_free(rdev, &bo->ib->sa_bo);
-	radeon_fence_unref(&bo->ib->fence);
-	list_del(&bo->list);
-	kfree(bo);
-}
-
-void radeon_semaphore_shrink_locked(struct radeon_device *rdev)
-{
-	struct radeon_semaphore_bo *bo, *n;
-
-	if (list_empty(&rdev->semaphore_drv.bo)) {
-		return;
-	}
-	/* only shrink if first bo has free semaphore */
-	bo = list_first_entry(&rdev->semaphore_drv.bo, struct radeon_semaphore_bo, list);
-	if (list_empty(&bo->free)) {
-		return;
-	}
-	list_for_each_entry_safe_continue(bo, n, &rdev->semaphore_drv.bo, list) {
-		if (bo->nused)
-			continue;
-		radeon_semaphore_del_bo_locked(rdev, bo);
-	}
-}
-
-int radeon_semaphore_create(struct radeon_device *rdev,
-			    struct radeon_semaphore **semaphore)
-{
-	struct radeon_semaphore_bo *bo;
-	unsigned long irq_flags;
-	bool do_retry = true;
-	int r;
-
-retry:
-	*semaphore = NULL;
-	write_lock_irqsave(&rdev->semaphore_drv.lock, irq_flags);
-	list_for_each_entry(bo, &rdev->semaphore_drv.bo, list) {
-		if (list_empty(&bo->free))
-			continue;
-		*semaphore = list_first_entry(&bo->free, struct radeon_semaphore, list);
-		(*semaphore)->cpu_ptr[0] = 0;
-		(*semaphore)->cpu_ptr[1] = 0;
-		list_del(&(*semaphore)->list);
-		bo->nused++;
-		break;
-	}
-	write_unlock_irqrestore(&rdev->semaphore_drv.lock, irq_flags);
-
-	if (*semaphore == NULL) {
-		if (do_retry) {
-			do_retry = false;
-			r = radeon_semaphore_add_bo(rdev);
-			if (r)
-				return r;
-			goto retry;
-		}
-		return -ENOMEM;
-	}
-
-	return 0;
-}
-
-void radeon_semaphore_emit_signal(struct radeon_device *rdev, int ring,
-			          struct radeon_semaphore *semaphore)
-{
-	radeon_semaphore_ring_emit(rdev, ring, &rdev->ring[ring], semaphore, false);
-}
-
-void radeon_semaphore_emit_wait(struct radeon_device *rdev, int ring,
-			        struct radeon_semaphore *semaphore)
-{
-	radeon_semaphore_ring_emit(rdev, ring, &rdev->ring[ring], semaphore, true);
-}
-
-void radeon_semaphore_free(struct radeon_device *rdev,
-			   struct radeon_semaphore *semaphore)
-{
-	unsigned long irq_flags;
-
-	write_lock_irqsave(&rdev->semaphore_drv.lock, irq_flags);
-	semaphore->bo->nused--;
-	list_add_tail(&semaphore->list, &semaphore->bo->free);
-	radeon_semaphore_shrink_locked(rdev);
-	write_unlock_irqrestore(&rdev->semaphore_drv.lock, irq_flags);
-}
-
-void radeon_semaphore_driver_fini(struct radeon_device *rdev)
-{
-	struct radeon_semaphore_bo *bo, *n;
-	unsigned long irq_flags;
-
-	write_lock_irqsave(&rdev->semaphore_drv.lock, irq_flags);
-	/* we force to free everything */
-	list_for_each_entry_safe(bo, n, &rdev->semaphore_drv.bo, list) {
-		if (!list_empty(&bo->free)) {
-			dev_err(rdev->dev, "still in use semaphore\n");
-		}
-		radeon_semaphore_del_bo_locked(rdev, bo);
-	}
-	write_unlock_irqrestore(&rdev->semaphore_drv.lock, irq_flags);
-}
diff --git a/drivers/gpu/drm/radeon/radeon_state.c b/drivers/gpu/drm/radeon/radeon_state.c
index e8422ae..92e7ea7 100644
--- a/drivers/gpu/drm/radeon/radeon_state.c
+++ b/drivers/gpu/drm/radeon/radeon_state.c
@@ -272,12 +272,12 @@ static __inline__ int radeon_check_and_fixup_packets(drm_radeon_private_t *
 	return 0;
 }
 
-static int radeon_check_and_fixup_packet3(drm_radeon_private_t *
-					  dev_priv,
-					  struct drm_file *file_priv,
-					  drm_radeon_kcmd_buffer_t *
-					  cmdbuf,
-					  unsigned int *cmdsz)
+static __inline__ int radeon_check_and_fixup_packet3(drm_radeon_private_t *
+						     dev_priv,
+						     struct drm_file *file_priv,
+						     drm_radeon_kcmd_buffer_t *
+						     cmdbuf,
+						     unsigned int *cmdsz)
 {
 	u32 *cmd = drm_buffer_pointer_to_dword(cmdbuf->buffer, 0);
 	u32 offset, narrays;
@@ -446,8 +446,8 @@ static int radeon_check_and_fixup_packet3(drm_radeon_private_t *
  * CP hardware state programming functions
  */
 
-static void radeon_emit_clip_rect(drm_radeon_private_t * dev_priv,
-				  struct drm_clip_rect * box)
+static __inline__ void radeon_emit_clip_rect(drm_radeon_private_t * dev_priv,
+					     struct drm_clip_rect * box)
 {
 	RING_LOCALS;
 
diff --git a/drivers/gpu/drm/radeon/radeon_test.c b/drivers/gpu/drm/radeon/radeon_test.c
index dc5dcf4..dee4a0c 100644
--- a/drivers/gpu/drm/radeon/radeon_test.c
+++ b/drivers/gpu/drm/radeon/radeon_test.c
@@ -40,16 +40,10 @@ void radeon_test_moves(struct radeon_device *rdev)
 	size = 1024 * 1024;
 
 	/* Number of tests =
-	 * (Total GTT - IB pool - writeback page - ring buffers) / test size
+	 * (Total GTT - IB pool - writeback page - ring buffer) / test size
 	 */
-	n = rdev->mc.gtt_size - RADEON_IB_POOL_SIZE*64*1024;
-	for (i = 0; i < RADEON_NUM_RINGS; ++i)
-		n -= rdev->ring[i].ring_size;
-	if (rdev->wb.wb_obj)
-		n -= RADEON_GPU_PAGE_SIZE;
-	if (rdev->ih.ring_obj)
-		n -= rdev->ih.ring_size;
-	n /= size;
+	n = ((u32)(rdev->mc.gtt_size - RADEON_IB_POOL_SIZE*64*1024 - RADEON_GPU_PAGE_SIZE -
+	     rdev->cp.ring_size)) / size;
 
 	gtt_obj = kzalloc(n * sizeof(*gtt_obj), GFP_KERNEL);
 	if (!gtt_obj) {
@@ -106,7 +100,7 @@ void radeon_test_moves(struct radeon_device *rdev)
 
 		radeon_bo_kunmap(gtt_obj[i]);
 
-		r = radeon_fence_create(rdev, &fence, RADEON_RING_TYPE_GFX_INDEX);
+		r = radeon_fence_create(rdev, &fence);
 		if (r) {
 			DRM_ERROR("Failed to create GTT->VRAM fence %d\n", i);
 			goto out_cleanup;
@@ -138,15 +132,9 @@ void radeon_test_moves(struct radeon_device *rdev)
 		     gtt_start++, vram_start++) {
 			if (*vram_start != gtt_start) {
 				DRM_ERROR("Incorrect GTT->VRAM copy %d: Got 0x%p, "
-					  "expected 0x%p (GTT/VRAM offset "
-					  "0x%16llx/0x%16llx)\n",
-					  i, *vram_start, gtt_start,
-					  (unsigned long long)
-					  (gtt_addr - rdev->mc.gtt_start +
-					   (void*)gtt_start - gtt_map),
-					  (unsigned long long)
-					  (vram_addr - rdev->mc.vram_start +
-					   (void*)gtt_start - gtt_map));
+					  "expected 0x%p (GTT map 0x%p-0x%p)\n",
+					  i, *vram_start, gtt_start, gtt_map,
+					  gtt_end);
 				radeon_bo_kunmap(vram_obj);
 				goto out_cleanup;
 			}
@@ -155,7 +143,7 @@ void radeon_test_moves(struct radeon_device *rdev)
 
 		radeon_bo_kunmap(vram_obj);
 
-		r = radeon_fence_create(rdev, &fence, RADEON_RING_TYPE_GFX_INDEX);
+		r = radeon_fence_create(rdev, &fence);
 		if (r) {
 			DRM_ERROR("Failed to create VRAM->GTT fence %d\n", i);
 			goto out_cleanup;
@@ -187,15 +175,9 @@ void radeon_test_moves(struct radeon_device *rdev)
 		     gtt_start++, vram_start++) {
 			if (*gtt_start != vram_start) {
 				DRM_ERROR("Incorrect VRAM->GTT copy %d: Got 0x%p, "
-					  "expected 0x%p (VRAM/GTT offset "
-					  "0x%16llx/0x%16llx)\n",
-					  i, *gtt_start, vram_start,
-					  (unsigned long long)
-					  (vram_addr - rdev->mc.vram_start +
-					   (void*)vram_start - vram_map),
-					  (unsigned long long)
-					  (gtt_addr - rdev->mc.gtt_start +
-					   (void*)vram_start - vram_map));
+					  "expected 0x%p (VRAM map 0x%p-0x%p)\n",
+					  i, *gtt_start, vram_start, vram_map,
+					  vram_end);
 				radeon_bo_kunmap(gtt_obj[i]);
 				goto out_cleanup;
 			}
@@ -234,264 +216,3 @@ out_cleanup:
 		printk(KERN_WARNING "Error while testing BO move.\n");
 	}
 }
-
-void radeon_test_ring_sync(struct radeon_device *rdev,
-			   struct radeon_ring *ringA,
-			   struct radeon_ring *ringB)
-{
-	struct radeon_fence *fence1 = NULL, *fence2 = NULL;
-	struct radeon_semaphore *semaphore = NULL;
-	int ridxA = radeon_ring_index(rdev, ringA);
-	int ridxB = radeon_ring_index(rdev, ringB);
-	int r;
-
-	r = radeon_fence_create(rdev, &fence1, ridxA);
-	if (r) {
-		DRM_ERROR("Failed to create sync fence 1\n");
-		goto out_cleanup;
-	}
-	r = radeon_fence_create(rdev, &fence2, ridxA);
-	if (r) {
-		DRM_ERROR("Failed to create sync fence 2\n");
-		goto out_cleanup;
-	}
-
-	r = radeon_semaphore_create(rdev, &semaphore);
-	if (r) {
-		DRM_ERROR("Failed to create semaphore\n");
-		goto out_cleanup;
-	}
-
-	r = radeon_ring_lock(rdev, ringA, 64);
-	if (r) {
-		DRM_ERROR("Failed to lock ring A %d\n", ridxA);
-		goto out_cleanup;
-	}
-	radeon_semaphore_emit_wait(rdev, ridxA, semaphore);
-	radeon_fence_emit(rdev, fence1);
-	radeon_semaphore_emit_wait(rdev, ridxA, semaphore);
-	radeon_fence_emit(rdev, fence2);
-	radeon_ring_unlock_commit(rdev, ringA);
-
-	mdelay(1000);
-
-	if (radeon_fence_signaled(fence1)) {
-		DRM_ERROR("Fence 1 signaled without waiting for semaphore.\n");
-		goto out_cleanup;
-	}
-
-	r = radeon_ring_lock(rdev, ringB, 64);
-	if (r) {
-		DRM_ERROR("Failed to lock ring B %p\n", ringB);
-		goto out_cleanup;
-	}
-	radeon_semaphore_emit_signal(rdev, ridxB, semaphore);
-	radeon_ring_unlock_commit(rdev, ringB);
-
-	r = radeon_fence_wait(fence1, false);
-	if (r) {
-		DRM_ERROR("Failed to wait for sync fence 1\n");
-		goto out_cleanup;
-	}
-
-	mdelay(1000);
-
-	if (radeon_fence_signaled(fence2)) {
-		DRM_ERROR("Fence 2 signaled without waiting for semaphore.\n");
-		goto out_cleanup;
-	}
-
-	r = radeon_ring_lock(rdev, ringB, 64);
-	if (r) {
-		DRM_ERROR("Failed to lock ring B %p\n", ringB);
-		goto out_cleanup;
-	}
-	radeon_semaphore_emit_signal(rdev, ridxB, semaphore);
-	radeon_ring_unlock_commit(rdev, ringB);
-
-	r = radeon_fence_wait(fence2, false);
-	if (r) {
-		DRM_ERROR("Failed to wait for sync fence 1\n");
-		goto out_cleanup;
-	}
-
-out_cleanup:
-	if (semaphore)
-		radeon_semaphore_free(rdev, semaphore);
-
-	if (fence1)
-		radeon_fence_unref(&fence1);
-
-	if (fence2)
-		radeon_fence_unref(&fence2);
-
-	if (r)
-		printk(KERN_WARNING "Error while testing ring sync (%d).\n", r);
-}
-
-void radeon_test_ring_sync2(struct radeon_device *rdev,
-			    struct radeon_ring *ringA,
-			    struct radeon_ring *ringB,
-			    struct radeon_ring *ringC)
-{
-	struct radeon_fence *fenceA = NULL, *fenceB = NULL;
-	struct radeon_semaphore *semaphore = NULL;
-	int ridxA = radeon_ring_index(rdev, ringA);
-	int ridxB = radeon_ring_index(rdev, ringB);
-	int ridxC = radeon_ring_index(rdev, ringC);
-	bool sigA, sigB;
-	int i, r;
-
-	r = radeon_fence_create(rdev, &fenceA, ridxA);
-	if (r) {
-		DRM_ERROR("Failed to create sync fence 1\n");
-		goto out_cleanup;
-	}
-	r = radeon_fence_create(rdev, &fenceB, ridxB);
-	if (r) {
-		DRM_ERROR("Failed to create sync fence 2\n");
-		goto out_cleanup;
-	}
-
-	r = radeon_semaphore_create(rdev, &semaphore);
-	if (r) {
-		DRM_ERROR("Failed to create semaphore\n");
-		goto out_cleanup;
-	}
-
-	r = radeon_ring_lock(rdev, ringA, 64);
-	if (r) {
-		DRM_ERROR("Failed to lock ring A %d\n", ridxA);
-		goto out_cleanup;
-	}
-	radeon_semaphore_emit_wait(rdev, ridxA, semaphore);
-	radeon_fence_emit(rdev, fenceA);
-	radeon_ring_unlock_commit(rdev, ringA);
-
-	r = radeon_ring_lock(rdev, ringB, 64);
-	if (r) {
-		DRM_ERROR("Failed to lock ring B %d\n", ridxB);
-		goto out_cleanup;
-	}
-	radeon_semaphore_emit_wait(rdev, ridxB, semaphore);
-	radeon_fence_emit(rdev, fenceB);
-	radeon_ring_unlock_commit(rdev, ringB);
-
-	mdelay(1000);
-
-	if (radeon_fence_signaled(fenceA)) {
-		DRM_ERROR("Fence A signaled without waiting for semaphore.\n");
-		goto out_cleanup;
-	}
-	if (radeon_fence_signaled(fenceB)) {
-		DRM_ERROR("Fence A signaled without waiting for semaphore.\n");
-		goto out_cleanup;
-	}
-
-	r = radeon_ring_lock(rdev, ringC, 64);
-	if (r) {
-		DRM_ERROR("Failed to lock ring B %p\n", ringC);
-		goto out_cleanup;
-	}
-	radeon_semaphore_emit_signal(rdev, ridxC, semaphore);
-	radeon_ring_unlock_commit(rdev, ringC);
-
-	for (i = 0; i < 30; ++i) {
-		mdelay(100);
-		sigA = radeon_fence_signaled(fenceA);
-		sigB = radeon_fence_signaled(fenceB);
-		if (sigA || sigB)
-			break;
-	}
-
-	if (!sigA && !sigB) {
-		DRM_ERROR("Neither fence A nor B has been signaled\n");
-		goto out_cleanup;
-	} else if (sigA && sigB) {
-		DRM_ERROR("Both fence A and B has been signaled\n");
-		goto out_cleanup;
-	}
-
-	DRM_INFO("Fence %c was first signaled\n", sigA ? 'A' : 'B');
-
-	r = radeon_ring_lock(rdev, ringC, 64);
-	if (r) {
-		DRM_ERROR("Failed to lock ring B %p\n", ringC);
-		goto out_cleanup;
-	}
-	radeon_semaphore_emit_signal(rdev, ridxC, semaphore);
-	radeon_ring_unlock_commit(rdev, ringC);
-
-	mdelay(1000);
-
-	r = radeon_fence_wait(fenceA, false);
-	if (r) {
-		DRM_ERROR("Failed to wait for sync fence A\n");
-		goto out_cleanup;
-	}
-	r = radeon_fence_wait(fenceB, false);
-	if (r) {
-		DRM_ERROR("Failed to wait for sync fence B\n");
-		goto out_cleanup;
-	}
-
-out_cleanup:
-	if (semaphore)
-		radeon_semaphore_free(rdev, semaphore);
-
-	if (fenceA)
-		radeon_fence_unref(&fenceA);
-
-	if (fenceB)
-		radeon_fence_unref(&fenceB);
-
-	if (r)
-		printk(KERN_WARNING "Error while testing ring sync (%d).\n", r);
-}
-
-void radeon_test_syncing(struct radeon_device *rdev)
-{
-	int i, j, k;
-
-	for (i = 1; i < RADEON_NUM_RINGS; ++i) {
-		struct radeon_ring *ringA = &rdev->ring[i];
-		if (!ringA->ready)
-			continue;
-
-		for (j = 0; j < i; ++j) {
-			struct radeon_ring *ringB = &rdev->ring[j];
-			if (!ringB->ready)
-				continue;
-
-			DRM_INFO("Testing syncing between rings %d and %d...\n", i, j);
-			radeon_test_ring_sync(rdev, ringA, ringB);
-
-			DRM_INFO("Testing syncing between rings %d and %d...\n", j, i);
-			radeon_test_ring_sync(rdev, ringB, ringA);
-
-			for (k = 0; k < j; ++k) {
-				struct radeon_ring *ringC = &rdev->ring[k];
-				if (!ringC->ready)
-					continue;
-
-				DRM_INFO("Testing syncing between rings %d, %d and %d...\n", i, j, k);
-				radeon_test_ring_sync2(rdev, ringA, ringB, ringC);
-
-				DRM_INFO("Testing syncing between rings %d, %d and %d...\n", i, k, j);
-				radeon_test_ring_sync2(rdev, ringA, ringC, ringB);
-
-				DRM_INFO("Testing syncing between rings %d, %d and %d...\n", j, i, k);
-				radeon_test_ring_sync2(rdev, ringB, ringA, ringC);
-
-				DRM_INFO("Testing syncing between rings %d, %d and %d...\n", j, k, i);
-				radeon_test_ring_sync2(rdev, ringB, ringC, ringA);
-
-				DRM_INFO("Testing syncing between rings %d, %d and %d...\n", k, i, j);
-				radeon_test_ring_sync2(rdev, ringC, ringA, ringB);
-
-				DRM_INFO("Testing syncing between rings %d, %d and %d...\n", k, j, i);
-				radeon_test_ring_sync2(rdev, ringC, ringB, ringA);
-			}
-		}
-	}
-}
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index f493c64..3e9b41b 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -114,6 +114,24 @@ static void radeon_ttm_global_fini(struct radeon_device *rdev)
 	}
 }
 
+struct ttm_backend *radeon_ttm_backend_create(struct radeon_device *rdev);
+
+static struct ttm_backend*
+radeon_create_ttm_backend_entry(struct ttm_bo_device *bdev)
+{
+	struct radeon_device *rdev;
+
+	rdev = radeon_get_rdev(bdev);
+#if __OS_HAS_AGP
+	if (rdev->flags & RADEON_IS_AGP) {
+		return ttm_agp_backend_init(bdev, rdev->ddev->agp->bridge);
+	} else
+#endif
+	{
+		return radeon_ttm_backend_create(rdev);
+	}
+}
+
 static int radeon_invalidate_caches(struct ttm_bo_device *bdev, uint32_t flags)
 {
 	return 0;
@@ -188,7 +206,7 @@ static void radeon_evict_flags(struct ttm_buffer_object *bo,
 	rbo = container_of(bo, struct radeon_bo, tbo);
 	switch (bo->mem.mem_type) {
 	case TTM_PL_VRAM:
-		if (rbo->rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready == false)
+		if (rbo->rdev->cp.ready == false)
 			radeon_ttm_placement_from_domain(rbo, RADEON_GEM_DOMAIN_CPU);
 		else
 			radeon_ttm_placement_from_domain(rbo, RADEON_GEM_DOMAIN_GTT);
@@ -223,10 +241,10 @@ static int radeon_move_blit(struct ttm_buffer_object *bo,
 	struct radeon_device *rdev;
 	uint64_t old_start, new_start;
 	struct radeon_fence *fence;
-	int r, i;
+	int r;
 
 	rdev = radeon_get_rdev(bo->bdev);
-	r = radeon_fence_create(rdev, &fence, radeon_copy_ring_index(rdev));
+	r = radeon_fence_create(rdev, &fence);
 	if (unlikely(r)) {
 		return r;
 	}
@@ -255,43 +273,13 @@ static int radeon_move_blit(struct ttm_buffer_object *bo,
 		DRM_ERROR("Unknown placement %d\n", old_mem->mem_type);
 		return -EINVAL;
 	}
-	if (!rdev->ring[radeon_copy_ring_index(rdev)].ready) {
-		DRM_ERROR("Trying to move memory with ring turned off.\n");
+	if (!rdev->cp.ready) {
+		DRM_ERROR("Trying to move memory with CP turned off.\n");
 		return -EINVAL;
 	}
 
 	BUILD_BUG_ON((PAGE_SIZE % RADEON_GPU_PAGE_SIZE) != 0);
 
-	/* sync other rings */
-	if (rdev->family >= CHIP_R600) {
-		for (i = 0; i < RADEON_NUM_RINGS; ++i) {
-			/* no need to sync to our own or unused rings */
-			if (i == radeon_copy_ring_index(rdev) || !rdev->ring[i].ready)
-				continue;
-
-			if (!fence->semaphore) {
-				r = radeon_semaphore_create(rdev, &fence->semaphore);
-				/* FIXME: handle semaphore error */
-				if (r)
-					continue;
-			}
-
-			r = radeon_ring_lock(rdev, &rdev->ring[i], 3);
-			/* FIXME: handle ring lock error */
-			if (r)
-				continue;
-			radeon_semaphore_emit_signal(rdev, i, fence->semaphore);
-			radeon_ring_unlock_commit(rdev, &rdev->ring[i]);
-
-			r = radeon_ring_lock(rdev, &rdev->ring[radeon_copy_ring_index(rdev)], 3);
-			/* FIXME: handle ring lock error */
-			if (r)
-				continue;
-			radeon_semaphore_emit_wait(rdev, radeon_copy_ring_index(rdev), fence->semaphore);
-			radeon_ring_unlock_commit(rdev, &rdev->ring[radeon_copy_ring_index(rdev)]);
-		}
-	}
-
 	r = radeon_copy(rdev, old_start, new_start,
 			new_mem->num_pages * (PAGE_SIZE / RADEON_GPU_PAGE_SIZE), /* GPU pages */
 			fence);
@@ -410,8 +398,7 @@ static int radeon_bo_move(struct ttm_buffer_object *bo,
 		radeon_move_null(bo, new_mem);
 		return 0;
 	}
-	if (!rdev->ring[radeon_copy_ring_index(rdev)].ready ||
-	    rdev->asic->copy.copy == NULL) {
+	if (!rdev->cp.ready || rdev->asic->copy == NULL) {
 		/* use memcpy */
 		goto memcpy;
 	}
@@ -468,29 +455,6 @@ static int radeon_ttm_io_mem_reserve(struct ttm_bo_device *bdev, struct ttm_mem_
 			return -EINVAL;
 		mem->bus.base = rdev->mc.aper_base;
 		mem->bus.is_iomem = true;
-#ifdef __alpha__
-		/*
-		 * Alpha: use bus.addr to hold the ioremap() return,
-		 * so we can modify bus.base below.
-		 */
-		if (mem->placement & TTM_PL_FLAG_WC)
-			mem->bus.addr =
-				ioremap_wc(mem->bus.base + mem->bus.offset,
-					   mem->bus.size);
-		else
-			mem->bus.addr =
-				ioremap_nocache(mem->bus.base + mem->bus.offset,
-						mem->bus.size);
-
-		/*
-		 * Alpha: Use just the bus offset plus
-		 * the hose/domain memory base for bus.base.
-		 * It then can be used to build PTEs for VRAM
-		 * access, as done in ttm_bo_vm_fault().
-		 */
-		mem->bus.base = (mem->bus.base & 0x0ffffffffUL) +
-			rdev->ddev->hose->dense_mem_base;
-#endif
 		break;
 	default:
 		return -EINVAL;
@@ -528,166 +492,8 @@ static bool radeon_sync_obj_signaled(void *sync_obj, void *sync_arg)
 	return radeon_fence_signaled((struct radeon_fence *)sync_obj);
 }
 
-/*
- * TTM backend functions.
- */
-struct radeon_ttm_tt {
-	struct ttm_dma_tt		ttm;
-	struct radeon_device		*rdev;
-	u64				offset;
-};
-
-static int radeon_ttm_backend_bind(struct ttm_tt *ttm,
-				   struct ttm_mem_reg *bo_mem)
-{
-	struct radeon_ttm_tt *gtt = (void*)ttm;
-	int r;
-
-	gtt->offset = (unsigned long)(bo_mem->start << PAGE_SHIFT);
-	if (!ttm->num_pages) {
-		WARN(1, "nothing to bind %lu pages for mreg %p back %p!\n",
-		     ttm->num_pages, bo_mem, ttm);
-	}
-	r = radeon_gart_bind(gtt->rdev, gtt->offset,
-			     ttm->num_pages, ttm->pages, gtt->ttm.dma_address);
-	if (r) {
-		DRM_ERROR("failed to bind %lu pages at 0x%08X\n",
-			  ttm->num_pages, (unsigned)gtt->offset);
-		return r;
-	}
-	return 0;
-}
-
-static int radeon_ttm_backend_unbind(struct ttm_tt *ttm)
-{
-	struct radeon_ttm_tt *gtt = (void *)ttm;
-
-	radeon_gart_unbind(gtt->rdev, gtt->offset, ttm->num_pages);
-	return 0;
-}
-
-static void radeon_ttm_backend_destroy(struct ttm_tt *ttm)
-{
-	struct radeon_ttm_tt *gtt = (void *)ttm;
-
-	ttm_dma_tt_fini(&gtt->ttm);
-	kfree(gtt);
-}
-
-static struct ttm_backend_func radeon_backend_func = {
-	.bind = &radeon_ttm_backend_bind,
-	.unbind = &radeon_ttm_backend_unbind,
-	.destroy = &radeon_ttm_backend_destroy,
-};
-
-struct ttm_tt *radeon_ttm_tt_create(struct ttm_bo_device *bdev,
-				    unsigned long size, uint32_t page_flags,
-				    struct page *dummy_read_page)
-{
-	struct radeon_device *rdev;
-	struct radeon_ttm_tt *gtt;
-
-	rdev = radeon_get_rdev(bdev);
-#if __OS_HAS_AGP
-	if (rdev->flags & RADEON_IS_AGP) {
-		return ttm_agp_tt_create(bdev, rdev->ddev->agp->bridge,
-					 size, page_flags, dummy_read_page);
-	}
-#endif
-
-	gtt = kzalloc(sizeof(struct radeon_ttm_tt), GFP_KERNEL);
-	if (gtt == NULL) {
-		return NULL;
-	}
-	gtt->ttm.ttm.func = &radeon_backend_func;
-	gtt->rdev = rdev;
-	if (ttm_dma_tt_init(&gtt->ttm, bdev, size, page_flags, dummy_read_page)) {
-		kfree(gtt);
-		return NULL;
-	}
-	return &gtt->ttm.ttm;
-}
-
-static int radeon_ttm_tt_populate(struct ttm_tt *ttm)
-{
-	struct radeon_device *rdev;
-	struct radeon_ttm_tt *gtt = (void *)ttm;
-	unsigned i;
-	int r;
-
-	if (ttm->state != tt_unpopulated)
-		return 0;
-
-	rdev = radeon_get_rdev(ttm->bdev);
-#if __OS_HAS_AGP
-	if (rdev->flags & RADEON_IS_AGP) {
-		return ttm_agp_tt_populate(ttm);
-	}
-#endif
-
-#ifdef CONFIG_SWIOTLB
-	if (swiotlb_nr_tbl()) {
-		return ttm_dma_populate(&gtt->ttm, rdev->dev);
-	}
-#endif
-
-	r = ttm_pool_populate(ttm);
-	if (r) {
-		return r;
-	}
-
-	for (i = 0; i < ttm->num_pages; i++) {
-		gtt->ttm.dma_address[i] = pci_map_page(rdev->pdev, ttm->pages[i],
-						       0, PAGE_SIZE,
-						       PCI_DMA_BIDIRECTIONAL);
-		if (pci_dma_mapping_error(rdev->pdev, gtt->ttm.dma_address[i])) {
-			while (--i) {
-				pci_unmap_page(rdev->pdev, gtt->ttm.dma_address[i],
-					       PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
-				gtt->ttm.dma_address[i] = 0;
-			}
-			ttm_pool_unpopulate(ttm);
-			return -EFAULT;
-		}
-	}
-	return 0;
-}
-
-static void radeon_ttm_tt_unpopulate(struct ttm_tt *ttm)
-{
-	struct radeon_device *rdev;
-	struct radeon_ttm_tt *gtt = (void *)ttm;
-	unsigned i;
-
-	rdev = radeon_get_rdev(ttm->bdev);
-#if __OS_HAS_AGP
-	if (rdev->flags & RADEON_IS_AGP) {
-		ttm_agp_tt_unpopulate(ttm);
-		return;
-	}
-#endif
-
-#ifdef CONFIG_SWIOTLB
-	if (swiotlb_nr_tbl()) {
-		ttm_dma_unpopulate(&gtt->ttm, rdev->dev);
-		return;
-	}
-#endif
-
-	for (i = 0; i < ttm->num_pages; i++) {
-		if (gtt->ttm.dma_address[i]) {
-			pci_unmap_page(rdev->pdev, gtt->ttm.dma_address[i],
-				       PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
-		}
-	}
-
-	ttm_pool_unpopulate(ttm);
-}
-
 static struct ttm_bo_driver radeon_bo_driver = {
-	.ttm_tt_create = &radeon_ttm_tt_create,
-	.ttm_tt_populate = &radeon_ttm_tt_populate,
-	.ttm_tt_unpopulate = &radeon_ttm_tt_unpopulate,
+	.create_ttm_backend_entry = &radeon_create_ttm_backend_entry,
 	.invalidate_caches = &radeon_invalidate_caches,
 	.init_mem_type = &radeon_init_mem_type,
 	.evict_flags = &radeon_evict_flags,
@@ -851,6 +657,124 @@ int radeon_mmap(struct file *filp, struct vm_area_struct *vma)
 }
 
 
+/*
+ * TTM backend functions.
+ */
+struct radeon_ttm_backend {
+	struct ttm_backend		backend;
+	struct radeon_device		*rdev;
+	unsigned long			num_pages;
+	struct page			**pages;
+	struct page			*dummy_read_page;
+	dma_addr_t			*dma_addrs;
+	bool				populated;
+	bool				bound;
+	unsigned			offset;
+};
+
+static int radeon_ttm_backend_populate(struct ttm_backend *backend,
+				       unsigned long num_pages,
+				       struct page **pages,
+				       struct page *dummy_read_page,
+				       dma_addr_t *dma_addrs)
+{
+	struct radeon_ttm_backend *gtt;
+
+	gtt = container_of(backend, struct radeon_ttm_backend, backend);
+	gtt->pages = pages;
+	gtt->dma_addrs = dma_addrs;
+	gtt->num_pages = num_pages;
+	gtt->dummy_read_page = dummy_read_page;
+	gtt->populated = true;
+	return 0;
+}
+
+static void radeon_ttm_backend_clear(struct ttm_backend *backend)
+{
+	struct radeon_ttm_backend *gtt;
+
+	gtt = container_of(backend, struct radeon_ttm_backend, backend);
+	gtt->pages = NULL;
+	gtt->dma_addrs = NULL;
+	gtt->num_pages = 0;
+	gtt->dummy_read_page = NULL;
+	gtt->populated = false;
+	gtt->bound = false;
+}
+
+
+static int radeon_ttm_backend_bind(struct ttm_backend *backend,
+				   struct ttm_mem_reg *bo_mem)
+{
+	struct radeon_ttm_backend *gtt;
+	int r;
+
+	gtt = container_of(backend, struct radeon_ttm_backend, backend);
+	gtt->offset = bo_mem->start << PAGE_SHIFT;
+	if (!gtt->num_pages) {
+		WARN(1, "nothing to bind %lu pages for mreg %p back %p!\n",
+		     gtt->num_pages, bo_mem, backend);
+	}
+	r = radeon_gart_bind(gtt->rdev, gtt->offset,
+			     gtt->num_pages, gtt->pages, gtt->dma_addrs);
+	if (r) {
+		DRM_ERROR("failed to bind %lu pages at 0x%08X\n",
+			  gtt->num_pages, gtt->offset);
+		return r;
+	}
+	gtt->bound = true;
+	return 0;
+}
+
+static int radeon_ttm_backend_unbind(struct ttm_backend *backend)
+{
+	struct radeon_ttm_backend *gtt;
+
+	gtt = container_of(backend, struct radeon_ttm_backend, backend);
+	radeon_gart_unbind(gtt->rdev, gtt->offset, gtt->num_pages);
+	gtt->bound = false;
+	return 0;
+}
+
+static void radeon_ttm_backend_destroy(struct ttm_backend *backend)
+{
+	struct radeon_ttm_backend *gtt;
+
+	gtt = container_of(backend, struct radeon_ttm_backend, backend);
+	if (gtt->bound) {
+		radeon_ttm_backend_unbind(backend);
+	}
+	kfree(gtt);
+}
+
+static struct ttm_backend_func radeon_backend_func = {
+	.populate = &radeon_ttm_backend_populate,
+	.clear = &radeon_ttm_backend_clear,
+	.bind = &radeon_ttm_backend_bind,
+	.unbind = &radeon_ttm_backend_unbind,
+	.destroy = &radeon_ttm_backend_destroy,
+};
+
+struct ttm_backend *radeon_ttm_backend_create(struct radeon_device *rdev)
+{
+	struct radeon_ttm_backend *gtt;
+
+	gtt = kzalloc(sizeof(struct radeon_ttm_backend), GFP_KERNEL);
+	if (gtt == NULL) {
+		return NULL;
+	}
+	gtt->backend.bdev = &rdev->mman.bdev;
+	gtt->backend.flags = 0;
+	gtt->backend.func = &radeon_backend_func;
+	gtt->rdev = rdev;
+	gtt->pages = NULL;
+	gtt->num_pages = 0;
+	gtt->dummy_read_page = NULL;
+	gtt->populated = false;
+	gtt->bound = false;
+	return &gtt->backend;
+}
+
 #define RADEON_DEBUGFS_MEM_TYPES 2
 
 #if defined(CONFIG_DEBUG_FS)
@@ -873,8 +797,8 @@ static int radeon_mm_dump_table(struct seq_file *m, void *data)
 static int radeon_ttm_debugfs_init(struct radeon_device *rdev)
 {
 #if defined(CONFIG_DEBUG_FS)
-	static struct drm_info_list radeon_mem_types_list[RADEON_DEBUGFS_MEM_TYPES+2];
-	static char radeon_mem_types_names[RADEON_DEBUGFS_MEM_TYPES+2][32];
+	static struct drm_info_list radeon_mem_types_list[RADEON_DEBUGFS_MEM_TYPES+1];
+	static char radeon_mem_types_names[RADEON_DEBUGFS_MEM_TYPES+1][32];
 	unsigned i;
 
 	for (i = 0; i < RADEON_DEBUGFS_MEM_TYPES; i++) {
@@ -896,17 +820,8 @@ static int radeon_ttm_debugfs_init(struct radeon_device *rdev)
 	radeon_mem_types_list[i].name = radeon_mem_types_names[i];
 	radeon_mem_types_list[i].show = &ttm_page_alloc_debugfs;
 	radeon_mem_types_list[i].driver_features = 0;
-	radeon_mem_types_list[i++].data = NULL;
-#ifdef CONFIG_SWIOTLB
-	if (swiotlb_nr_tbl()) {
-		sprintf(radeon_mem_types_names[i], "ttm_dma_page_pool");
-		radeon_mem_types_list[i].name = radeon_mem_types_names[i];
-		radeon_mem_types_list[i].show = &ttm_dma_page_alloc_debugfs;
-		radeon_mem_types_list[i].driver_features = 0;
-		radeon_mem_types_list[i++].data = NULL;
-	}
-#endif
-	return radeon_debugfs_add_files(rdev, radeon_mem_types_list, i);
+	radeon_mem_types_list[i].data = NULL;
+	return radeon_debugfs_add_files(rdev, radeon_mem_types_list, RADEON_DEBUGFS_MEM_TYPES+1);
 
 #endif
 	return 0;
diff --git a/drivers/gpu/drm/radeon/reg_srcs/cayman b/drivers/gpu/drm/radeon/reg_srcs/cayman
index aea63c4..0aa8e85 100644
--- a/drivers/gpu/drm/radeon/reg_srcs/cayman
+++ b/drivers/gpu/drm/radeon/reg_srcs/cayman
@@ -1,8 +1,5 @@
 cayman 0x9400
 0x0000802C GRBM_GFX_INDEX
-0x000084FC CP_STRMOUT_CNTL
-0x000085F0 CP_COHER_CNTL
-0x000085F4 CP_COHER_SIZE
 0x000088B0 VGT_VTX_VECT_EJECT_REG
 0x000088C4 VGT_CACHE_INVALIDATION
 0x000088D4 VGT_GS_VERTEX_REUSE
@@ -80,6 +77,7 @@ cayman 0x9400
 0x0002802C DB_DEPTH_CLEAR
 0x00028030 PA_SC_SCREEN_SCISSOR_TL
 0x00028034 PA_SC_SCREEN_SCISSOR_BR
+0x0002805C DB_DEPTH_SLICE
 0x00028140 SQ_ALU_CONST_BUFFER_SIZE_PS_0
 0x00028144 SQ_ALU_CONST_BUFFER_SIZE_PS_1
 0x00028148 SQ_ALU_CONST_BUFFER_SIZE_PS_2
@@ -208,8 +206,8 @@ cayman 0x9400
 0x00028344 PA_SC_VPORT_ZMAX_14
 0x00028348 PA_SC_VPORT_ZMIN_15
 0x0002834C PA_SC_VPORT_ZMAX_15
+0x00028350 SX_MISC
 0x00028354 SX_SURFACE_SYNC
-0x0002835C SX_SCATTER_EXPORT_SIZE
 0x00028380 SQ_VTX_SEMANTIC_0
 0x00028384 SQ_VTX_SEMANTIC_1
 0x00028388 SQ_VTX_SEMANTIC_2
@@ -434,7 +432,6 @@ cayman 0x9400
 0x00028700 SPI_STACK_MGMT
 0x00028704 SPI_WAVE_MGMT_1
 0x00028708 SPI_WAVE_MGMT_2
-0x00028720 GDS_ADDR_BASE
 0x00028724 GDS_ADDR_SIZE
 0x00028780 CB_BLEND0_CONTROL
 0x00028784 CB_BLEND1_CONTROL
@@ -513,13 +510,6 @@ cayman 0x9400
 0x00028AC0 DB_SRESULTS_COMPARE_STATE0
 0x00028AC4 DB_SRESULTS_COMPARE_STATE1
 0x00028AC8 DB_PRELOAD_CONTROL
-0x00028AD4 VGT_STRMOUT_VTX_STRIDE_0
-0x00028AE4 VGT_STRMOUT_VTX_STRIDE_1
-0x00028AF4 VGT_STRMOUT_VTX_STRIDE_2
-0x00028B04 VGT_STRMOUT_VTX_STRIDE_3
-0x00028B28 VGT_STRMOUT_DRAW_OPAQUE_OFFSET
-0x00028B2C VGT_STRMOUT_DRAW_OPAQUE_BUFFER_FILLED_SIZE
-0x00028B30 VGT_STRMOUT_DRAW_OPAQUE_VERTEX_STRIDE
 0x00028B38 VGT_GS_MAX_VERT_OUT
 0x00028B54 VGT_SHADER_STAGES_EN
 0x00028B58 VGT_LS_HS_CONFIG
@@ -559,18 +549,6 @@ cayman 0x9400
 0x00028C34 PA_SC_AA_SAMPLE_LOCS_PIXEL_X1_Y1_3
 0x00028C38 PA_SC_AA_MASK_X0_Y0_X1_Y0
 0x00028C3C PA_SC_AA_MASK_X0_Y1_X1_Y1
-0x00028C78 CB_COLOR0_DIM
-0x00028CB4 CB_COLOR1_DIM
-0x00028CF0 CB_COLOR2_DIM
-0x00028D2C CB_COLOR3_DIM
-0x00028D68 CB_COLOR4_DIM
-0x00028DA4 CB_COLOR5_DIM
-0x00028DE0 CB_COLOR6_DIM
-0x00028E1C CB_COLOR7_DIM
-0x00028E58 CB_COLOR8_DIM
-0x00028E74 CB_COLOR9_DIM
-0x00028E90 CB_COLOR10_DIM
-0x00028EAC CB_COLOR11_DIM
 0x00028C8C CB_COLOR0_CLEAR_WORD0
 0x00028C90 CB_COLOR0_CLEAR_WORD1
 0x00028C94 CB_COLOR0_CLEAR_WORD2
diff --git a/drivers/gpu/drm/radeon/reg_srcs/evergreen b/drivers/gpu/drm/radeon/reg_srcs/evergreen
index 77c3720..0e28cae 100644
--- a/drivers/gpu/drm/radeon/reg_srcs/evergreen
+++ b/drivers/gpu/drm/radeon/reg_srcs/evergreen
@@ -4,9 +4,6 @@ evergreen 0x9400
 0x00008044 WAIT_UNTIL_POLL_CNTL
 0x00008048 WAIT_UNTIL_POLL_MASK
 0x0000804c WAIT_UNTIL_POLL_REFDATA
-0x000084FC CP_STRMOUT_CNTL
-0x000085F0 CP_COHER_CNTL
-0x000085F4 CP_COHER_SIZE
 0x000088B0 VGT_VTX_VECT_EJECT_REG
 0x000088C4 VGT_CACHE_INVALIDATION
 0x000088D4 VGT_GS_VERTEX_REUSE
@@ -47,7 +44,6 @@ evergreen 0x9400
 0x00008E28 SQ_STATIC_THREAD_MGMT_3
 0x00008E2C SQ_LDS_RESOURCE_MGMT
 0x00008E48 SQ_EX_ALLOC_TABLE_SLOTS
-0x00009014 SX_MEMORY_EXPORT_SIZE
 0x00009100 SPI_CONFIG_CNTL
 0x0000913C SPI_CONFIG_CNTL_1
 0x00009508 TA_CNTL_AUX
@@ -96,6 +92,7 @@ evergreen 0x9400
 0x0002802C DB_DEPTH_CLEAR
 0x00028030 PA_SC_SCREEN_SCISSOR_TL
 0x00028034 PA_SC_SCREEN_SCISSOR_BR
+0x0002805C DB_DEPTH_SLICE
 0x00028140 SQ_ALU_CONST_BUFFER_SIZE_PS_0
 0x00028144 SQ_ALU_CONST_BUFFER_SIZE_PS_1
 0x00028148 SQ_ALU_CONST_BUFFER_SIZE_PS_2
@@ -224,6 +221,7 @@ evergreen 0x9400
 0x00028344 PA_SC_VPORT_ZMAX_14
 0x00028348 PA_SC_VPORT_ZMIN_15
 0x0002834C PA_SC_VPORT_ZMAX_15
+0x00028350 SX_MISC
 0x00028354 SX_SURFACE_SYNC
 0x00028380 SQ_VTX_SEMANTIC_0
 0x00028384 SQ_VTX_SEMANTIC_1
@@ -444,9 +442,7 @@ evergreen 0x9400
 0x000286EC SPI_COMPUTE_NUM_THREAD_X
 0x000286F0 SPI_COMPUTE_NUM_THREAD_Y
 0x000286F4 SPI_COMPUTE_NUM_THREAD_Z
-0x00028720 GDS_ADDR_BASE
 0x00028724 GDS_ADDR_SIZE
-0x00028728 GDS_ORDERED_WAVE_PER_SE
 0x00028780 CB_BLEND0_CONTROL
 0x00028784 CB_BLEND1_CONTROL
 0x00028788 CB_BLEND2_CONTROL
@@ -523,13 +519,6 @@ evergreen 0x9400
 0x00028AC0 DB_SRESULTS_COMPARE_STATE0
 0x00028AC4 DB_SRESULTS_COMPARE_STATE1
 0x00028AC8 DB_PRELOAD_CONTROL
-0x00028AD4 VGT_STRMOUT_VTX_STRIDE_0
-0x00028AE4 VGT_STRMOUT_VTX_STRIDE_1
-0x00028AF4 VGT_STRMOUT_VTX_STRIDE_2
-0x00028B04 VGT_STRMOUT_VTX_STRIDE_3
-0x00028B28 VGT_STRMOUT_DRAW_OPAQUE_OFFSET
-0x00028B2C VGT_STRMOUT_DRAW_OPAQUE_BUFFER_FILLED_SIZE
-0x00028B30 VGT_STRMOUT_DRAW_OPAQUE_VERTEX_STRIDE
 0x00028B38 VGT_GS_MAX_VERT_OUT
 0x00028B54 VGT_SHADER_STAGES_EN
 0x00028B58 VGT_LS_HS_CONFIG
@@ -562,18 +551,6 @@ evergreen 0x9400
 0x00028C34 PA_SC_AA_SAMPLE_LOCS_6
 0x00028C38 PA_SC_AA_SAMPLE_LOCS_7
 0x00028C3C PA_SC_AA_MASK
-0x00028C78 CB_COLOR0_DIM
-0x00028CB4 CB_COLOR1_DIM
-0x00028CF0 CB_COLOR2_DIM
-0x00028D2C CB_COLOR3_DIM
-0x00028D68 CB_COLOR4_DIM
-0x00028DA4 CB_COLOR5_DIM
-0x00028DE0 CB_COLOR6_DIM
-0x00028E1C CB_COLOR7_DIM
-0x00028E58 CB_COLOR8_DIM
-0x00028E74 CB_COLOR9_DIM
-0x00028E90 CB_COLOR10_DIM
-0x00028EAC CB_COLOR11_DIM
 0x00028C8C CB_COLOR0_CLEAR_WORD0
 0x00028C90 CB_COLOR0_CLEAR_WORD1
 0x00028C94 CB_COLOR0_CLEAR_WORD2
diff --git a/drivers/gpu/drm/radeon/reg_srcs/r600 b/drivers/gpu/drm/radeon/reg_srcs/r600
index 626c24e..ea49752 100644
--- a/drivers/gpu/drm/radeon/reg_srcs/r600
+++ b/drivers/gpu/drm/radeon/reg_srcs/r600
@@ -3,9 +3,6 @@ r600 0x9400
 0x00028230 R7xx_PA_SC_EDGERULE
 0x000286C8 R7xx_SPI_THREAD_GROUPING
 0x00008D8C R7xx_SQ_DYN_GPR_CNTL_PS_FLUSH_REQ
-0x00008490 CP_STRMOUT_CNTL
-0x000085F0 CP_COHER_CNTL
-0x000085F4 CP_COHER_SIZE
 0x000088C4 VGT_CACHE_INVALIDATION
 0x00028A50 VGT_ENHANCE
 0x000088CC VGT_ES_PER_GS
@@ -41,13 +38,6 @@ r600 0x9400
 0x00028AB4 VGT_REUSE_OFF
 0x00028AB8 VGT_VTX_CNT_EN
 0x000088B0 VGT_VTX_VECT_EJECT_REG
-0x00028AD4 VGT_STRMOUT_VTX_STRIDE_0
-0x00028AE4 VGT_STRMOUT_VTX_STRIDE_1
-0x00028AF4 VGT_STRMOUT_VTX_STRIDE_2
-0x00028B04 VGT_STRMOUT_VTX_STRIDE_3
-0x00028B28 VGT_STRMOUT_DRAW_OPAQUE_OFFSET
-0x00028B2C VGT_STRMOUT_DRAW_OPAQUE_BUFFER_FILLED_SIZE
-0x00028B30 VGT_STRMOUT_DRAW_OPAQUE_VERTEX_STRIDE
 0x00028810 PA_CL_CLIP_CNTL
 0x00008A14 PA_CL_ENHANCE
 0x00028C14 PA_CL_GB_HORZ_CLIP_ADJ
@@ -438,8 +428,7 @@ r600 0x9400
 0x00028638 SPI_VS_OUT_ID_9
 0x00028438 SX_ALPHA_REF
 0x00028410 SX_ALPHA_TEST_CONTROL
-0x00028354 SX_SURFACE_SYNC
-0x00009014 SX_MEMORY_EXPORT_SIZE
+0x00028350 SX_MISC
 0x00009604 TC_INVALIDATE
 0x00009400 TD_FILTER4
 0x00009404 TD_FILTER4_1
@@ -753,6 +742,14 @@ r600 0x9400
 0x00028114 CB_COLOR5_MASK
 0x00028118 CB_COLOR6_MASK
 0x0002811C CB_COLOR7_MASK
+0x00028080 CB_COLOR0_VIEW
+0x00028084 CB_COLOR1_VIEW
+0x00028088 CB_COLOR2_VIEW
+0x0002808C CB_COLOR3_VIEW
+0x00028090 CB_COLOR4_VIEW
+0x00028094 CB_COLOR5_VIEW
+0x00028098 CB_COLOR6_VIEW
+0x0002809C CB_COLOR7_VIEW
 0x00028808 CB_COLOR_CONTROL
 0x0002842C CB_FOG_BLUE
 0x00028428 CB_FOG_GREEN
diff --git a/drivers/gpu/drm/radeon/rs400.c b/drivers/gpu/drm/radeon/rs400.c
index aba2093..317eac1 100644
--- a/drivers/gpu/drm/radeon/rs400.c
+++ b/drivers/gpu/drm/radeon/rs400.c
@@ -77,7 +77,7 @@ int rs400_gart_init(struct radeon_device *rdev)
 {
 	int r;
 
-	if (rdev->gart.ptr) {
+	if (rdev->gart.table.ram.ptr) {
 		WARN(1, "RS400 GART already initialized\n");
 		return 0;
 	}
@@ -185,9 +185,6 @@ int rs400_gart_enable(struct radeon_device *rdev)
 	/* Enable gart */
 	WREG32_MC(RS480_AGP_ADDRESS_SPACE_SIZE, (RS480_GART_EN | size_reg));
 	rs400_gart_tlb_flush(rdev);
-	DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n",
-		 (unsigned)(rdev->mc.gtt_size >> 20),
-		 (unsigned long long)rdev->gart.table_addr);
 	rdev->gart.ready = true;
 	return 0;
 }
@@ -215,7 +212,6 @@ void rs400_gart_fini(struct radeon_device *rdev)
 int rs400_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr)
 {
 	uint32_t entry;
-	u32 *gtt = rdev->gart.ptr;
 
 	if (i < 0 || i > rdev->gart.num_gpu_pages) {
 		return -EINVAL;
@@ -225,7 +221,7 @@ int rs400_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr)
 		((upper_32_bits(addr) & 0xff) << 4) |
 		RS400_PTE_WRITEABLE | RS400_PTE_READABLE;
 	entry = cpu_to_le32(entry);
-	gtt[i] = entry;
+	rdev->gart.table.ram.ptr[i] = entry;
 	return 0;
 }
 
@@ -413,12 +409,6 @@ static int rs400_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX);
-	if (r) {
-		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
-		return r;
-	}
-
 	/* Enable IRQ */
 	r100_irq_set(rdev);
 	rdev->config.r300.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL);
@@ -428,25 +418,16 @@ static int rs400_startup(struct radeon_device *rdev)
 		dev_err(rdev->dev, "failed initializing CP (%d).\n", r);
 		return r;
 	}
-
-	r = radeon_ib_pool_start(rdev);
-	if (r)
-		return r;
-
-	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
+	r = r100_ib_init(rdev);
 	if (r) {
-		dev_err(rdev->dev, "failed testing IB (%d).\n", r);
-		rdev->accel_working = false;
+		dev_err(rdev->dev, "failed initializing IB (%d).\n", r);
 		return r;
 	}
-
 	return 0;
 }
 
 int rs400_resume(struct radeon_device *rdev)
 {
-	int r;
-
 	/* Make sur GART are not working */
 	rs400_gart_disable(rdev);
 	/* Resume clock before doing reset */
@@ -465,18 +446,11 @@ int rs400_resume(struct radeon_device *rdev)
 	r300_clock_startup(rdev);
 	/* Initialize surface registers */
 	radeon_surface_init(rdev);
-
-	rdev->accel_working = true;
-	r = rs400_startup(rdev);
-	if (r) {
-		rdev->accel_working = false;
-	}
-	return r;
+	return rs400_startup(rdev);
 }
 
 int rs400_suspend(struct radeon_device *rdev)
 {
-	radeon_ib_pool_suspend(rdev);
 	r100_cp_disable(rdev);
 	radeon_wb_disable(rdev);
 	r100_irq_disable(rdev);
@@ -555,14 +529,7 @@ int rs400_init(struct radeon_device *rdev)
 	if (r)
 		return r;
 	r300_set_reg_safe(rdev);
-
-	r = radeon_ib_pool_init(rdev);
 	rdev->accel_working = true;
-	if (r) {
-		dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
-		rdev->accel_working = false;
-	}
-
 	r = rs400_startup(rdev);
 	if (r) {
 		/* Somethings want wront with the accel init stop accel */
diff --git a/drivers/gpu/drm/radeon/rs600.c b/drivers/gpu/drm/radeon/rs600.c
index d25cf86..2026c2d 100644
--- a/drivers/gpu/drm/radeon/rs600.c
+++ b/drivers/gpu/drm/radeon/rs600.c
@@ -46,25 +46,6 @@
 void rs600_gpu_init(struct radeon_device *rdev);
 int rs600_mc_wait_for_idle(struct radeon_device *rdev);
 
-void avivo_wait_for_vblank(struct radeon_device *rdev, int crtc)
-{
-	struct radeon_crtc *radeon_crtc = rdev->mode_info.crtcs[crtc];
-	int i;
-
-	if (RREG32(AVIVO_D1CRTC_CONTROL + radeon_crtc->crtc_offset) & AVIVO_CRTC_EN) {
-		for (i = 0; i < rdev->usec_timeout; i++) {
-			if (!(RREG32(AVIVO_D1CRTC_STATUS + radeon_crtc->crtc_offset) & AVIVO_D1CRTC_V_BLANK))
-				break;
-			udelay(1);
-		}
-		for (i = 0; i < rdev->usec_timeout; i++) {
-			if (RREG32(AVIVO_D1CRTC_STATUS + radeon_crtc->crtc_offset) & AVIVO_D1CRTC_V_BLANK)
-				break;
-			udelay(1);
-		}
-	}
-}
-
 void rs600_pre_page_flip(struct radeon_device *rdev, int crtc)
 {
 	/* enable the pflip int */
@@ -194,7 +175,7 @@ void rs600_pm_misc(struct radeon_device *rdev)
 	/* set pcie lanes */
 	if ((rdev->flags & RADEON_IS_PCIE) &&
 	    !(rdev->flags & RADEON_IS_IGP) &&
-	    rdev->asic->pm.set_pcie_lanes &&
+	    rdev->asic->set_pcie_lanes &&
 	    (ps->pcie_lanes !=
 	     rdev->pm.power_state[rdev->pm.current_power_state_index].pcie_lanes)) {
 		radeon_set_pcie_lanes(rdev,
@@ -341,6 +322,16 @@ void rs600_hpd_fini(struct radeon_device *rdev)
 	}
 }
 
+void rs600_bm_disable(struct radeon_device *rdev)
+{
+	u16 tmp;
+
+	/* disable bus mastering */
+	pci_read_config_word(rdev->pdev, 0x4, &tmp);
+	pci_write_config_word(rdev->pdev, 0x4, tmp & 0xFFFB);
+	mdelay(1);
+}
+
 int rs600_asic_reset(struct radeon_device *rdev)
 {
 	struct rv515_mc_save save;
@@ -364,8 +355,7 @@ int rs600_asic_reset(struct radeon_device *rdev)
 	WREG32(RADEON_CP_RB_CNTL, tmp);
 	pci_save_state(rdev->pdev);
 	/* disable bus mastering */
-	pci_clear_master(rdev->pdev);
-	mdelay(1);
+	rs600_bm_disable(rdev);
 	/* reset GA+VAP */
 	WREG32(R_0000F0_RBBM_SOFT_RESET, S_0000F0_SOFT_RESET_VAP(1) |
 					S_0000F0_SOFT_RESET_GA(1));
@@ -429,7 +419,7 @@ int rs600_gart_init(struct radeon_device *rdev)
 {
 	int r;
 
-	if (rdev->gart.robj) {
+	if (rdev->gart.table.vram.robj) {
 		WARN(1, "RS600 GART already initialized\n");
 		return 0;
 	}
@@ -447,7 +437,7 @@ static int rs600_gart_enable(struct radeon_device *rdev)
 	u32 tmp;
 	int r, i;
 
-	if (rdev->gart.robj == NULL) {
+	if (rdev->gart.table.vram.robj == NULL) {
 		dev_err(rdev->dev, "No VRAM object for PCIE GART.\n");
 		return -EINVAL;
 	}
@@ -500,9 +490,6 @@ static int rs600_gart_enable(struct radeon_device *rdev)
 	tmp = RREG32_MC(R_000009_MC_CNTL1);
 	WREG32_MC(R_000009_MC_CNTL1, (tmp | S_000009_ENABLE_PAGE_TABLES(1)));
 	rs600_gart_tlb_flush(rdev);
-	DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n",
-		 (unsigned)(rdev->mc.gtt_size >> 20),
-		 (unsigned long long)rdev->gart.table_addr);
 	rdev->gart.ready = true;
 	return 0;
 }
@@ -510,12 +497,20 @@ static int rs600_gart_enable(struct radeon_device *rdev)
 void rs600_gart_disable(struct radeon_device *rdev)
 {
 	u32 tmp;
+	int r;
 
 	/* FIXME: disable out of gart access */
 	WREG32_MC(R_000100_MC_PT0_CNTL, 0);
 	tmp = RREG32_MC(R_000009_MC_CNTL1);
 	WREG32_MC(R_000009_MC_CNTL1, tmp & C_000009_ENABLE_PAGE_TABLES);
-	radeon_gart_table_vram_unpin(rdev);
+	if (rdev->gart.table.vram.robj) {
+		r = radeon_bo_reserve(rdev->gart.table.vram.robj, false);
+		if (r == 0) {
+			radeon_bo_kunmap(rdev->gart.table.vram.robj);
+			radeon_bo_unpin(rdev->gart.table.vram.robj);
+			radeon_bo_unreserve(rdev->gart.table.vram.robj);
+		}
+	}
 }
 
 void rs600_gart_fini(struct radeon_device *rdev)
@@ -533,7 +528,7 @@ void rs600_gart_fini(struct radeon_device *rdev)
 
 int rs600_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr)
 {
-	void __iomem *ptr = (void *)rdev->gart.ptr;
+	void __iomem *ptr = (void *)rdev->gart.table.vram.ptr;
 
 	if (i < 0 || i > rdev->gart.num_gpu_pages) {
 		return -EINVAL;
@@ -541,7 +536,7 @@ int rs600_gart_set_page(struct radeon_device *rdev, int i, uint64_t addr)
 	addr = addr & 0xFFFFFFFFFFFFF000ULL;
 	addr |= R600_PTE_VALID | R600_PTE_SYSTEM | R600_PTE_SNOOPED;
 	addr |= R600_PTE_READABLE | R600_PTE_WRITEABLE;
-	writeq(addr, ptr + (i * 8));
+	writeq(addr, ((void __iomem *)ptr) + (i * 8));
 	return 0;
 }
 
@@ -559,7 +554,7 @@ int rs600_irq_set(struct radeon_device *rdev)
 		WREG32(R_000040_GEN_INT_CNTL, 0);
 		return -EINVAL;
 	}
-	if (rdev->irq.sw_int[RADEON_RING_TYPE_GFX_INDEX]) {
+	if (rdev->irq.sw_int) {
 		tmp |= S_000040_SW_INT_EN(1);
 	}
 	if (rdev->irq.gui_idle) {
@@ -652,7 +647,7 @@ int rs600_irq_process(struct radeon_device *rdev)
 	while (status || rdev->irq.stat_regs.r500.disp_int) {
 		/* SW interrupt */
 		if (G_000044_SW_INT(status)) {
-			radeon_fence_process(rdev, RADEON_RING_TYPE_GFX_INDEX);
+			radeon_fence_process(rdev);
 		}
 		/* GUI idle */
 		if (G_000040_GUI_IDLE(status)) {
@@ -857,12 +852,6 @@ static int rs600_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX);
-	if (r) {
-		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
-		return r;
-	}
-
 	/* Enable IRQ */
 	rs600_irq_set(rdev);
 	rdev->config.r300.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL);
@@ -872,21 +861,15 @@ static int rs600_startup(struct radeon_device *rdev)
 		dev_err(rdev->dev, "failed initializing CP (%d).\n", r);
 		return r;
 	}
-
-	r = r600_audio_init(rdev);
+	r = r100_ib_init(rdev);
 	if (r) {
-		dev_err(rdev->dev, "failed initializing audio\n");
+		dev_err(rdev->dev, "failed initializing IB (%d).\n", r);
 		return r;
 	}
 
-	r = radeon_ib_pool_start(rdev);
-	if (r)
-		return r;
-
-	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
+	r = r600_audio_init(rdev);
 	if (r) {
-		dev_err(rdev->dev, "failed testing IB (%d).\n", r);
-		rdev->accel_working = false;
+		dev_err(rdev->dev, "failed initializing audio\n");
 		return r;
 	}
 
@@ -895,8 +878,6 @@ static int rs600_startup(struct radeon_device *rdev)
 
 int rs600_resume(struct radeon_device *rdev)
 {
-	int r;
-
 	/* Make sur GART are not working */
 	rs600_gart_disable(rdev);
 	/* Resume clock before doing reset */
@@ -913,18 +894,11 @@ int rs600_resume(struct radeon_device *rdev)
 	rv515_clock_startup(rdev);
 	/* Initialize surface registers */
 	radeon_surface_init(rdev);
-
-	rdev->accel_working = true;
-	r = rs600_startup(rdev);
-	if (r) {
-		rdev->accel_working = false;
-	}
-	return r;
+	return rs600_startup(rdev);
 }
 
 int rs600_suspend(struct radeon_device *rdev)
 {
-	radeon_ib_pool_suspend(rdev);
 	r600_audio_fini(rdev);
 	r100_cp_disable(rdev);
 	radeon_wb_disable(rdev);
@@ -1005,14 +979,7 @@ int rs600_init(struct radeon_device *rdev)
 	if (r)
 		return r;
 	rs600_set_safe_registers(rdev);
-
-	r = radeon_ib_pool_init(rdev);
 	rdev->accel_working = true;
-	if (r) {
-		dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
-		rdev->accel_working = false;
-	}
-
 	r = rs600_startup(rdev);
 	if (r) {
 		/* Somethings want wront with the accel init stop accel */
diff --git a/drivers/gpu/drm/radeon/rs690.c b/drivers/gpu/drm/radeon/rs690.c
index f2c3b9d..a9049ed 100644
--- a/drivers/gpu/drm/radeon/rs690.c
+++ b/drivers/gpu/drm/radeon/rs690.c
@@ -31,7 +31,7 @@
 #include "atom.h"
 #include "rs690d.h"
 
-int rs690_mc_wait_for_idle(struct radeon_device *rdev)
+static int rs690_mc_wait_for_idle(struct radeon_device *rdev)
 {
 	unsigned i;
 	uint32_t tmp;
@@ -621,12 +621,6 @@ static int rs690_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX);
-	if (r) {
-		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
-		return r;
-	}
-
 	/* Enable IRQ */
 	rs600_irq_set(rdev);
 	rdev->config.r300.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL);
@@ -636,21 +630,15 @@ static int rs690_startup(struct radeon_device *rdev)
 		dev_err(rdev->dev, "failed initializing CP (%d).\n", r);
 		return r;
 	}
-
-	r = r600_audio_init(rdev);
+	r = r100_ib_init(rdev);
 	if (r) {
-		dev_err(rdev->dev, "failed initializing audio\n");
+		dev_err(rdev->dev, "failed initializing IB (%d).\n", r);
 		return r;
 	}
 
-	r = radeon_ib_pool_start(rdev);
-	if (r)
-		return r;
-
-	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
+	r = r600_audio_init(rdev);
 	if (r) {
-		dev_err(rdev->dev, "failed testing IB (%d).\n", r);
-		rdev->accel_working = false;
+		dev_err(rdev->dev, "failed initializing audio\n");
 		return r;
 	}
 
@@ -659,8 +647,6 @@ static int rs690_startup(struct radeon_device *rdev)
 
 int rs690_resume(struct radeon_device *rdev)
 {
-	int r;
-
 	/* Make sur GART are not working */
 	rs400_gart_disable(rdev);
 	/* Resume clock before doing reset */
@@ -677,18 +663,11 @@ int rs690_resume(struct radeon_device *rdev)
 	rv515_clock_startup(rdev);
 	/* Initialize surface registers */
 	radeon_surface_init(rdev);
-
-	rdev->accel_working = true;
-	r = rs690_startup(rdev);
-	if (r) {
-		rdev->accel_working = false;
-	}
-	return r;
+	return rs690_startup(rdev);
 }
 
 int rs690_suspend(struct radeon_device *rdev)
 {
-	radeon_ib_pool_suspend(rdev);
 	r600_audio_fini(rdev);
 	r100_cp_disable(rdev);
 	radeon_wb_disable(rdev);
@@ -770,14 +749,7 @@ int rs690_init(struct radeon_device *rdev)
 	if (r)
 		return r;
 	rs600_set_safe_registers(rdev);
-
-	r = radeon_ib_pool_init(rdev);
 	rdev->accel_working = true;
-	if (r) {
-		dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
-		rdev->accel_working = false;
-	}
-
 	r = rs690_startup(rdev);
 	if (r) {
 		/* Somethings want wront with the accel init stop accel */
diff --git a/drivers/gpu/drm/radeon/rv515.c b/drivers/gpu/drm/radeon/rv515.c
index ea18e58..d5f45b4 100644
--- a/drivers/gpu/drm/radeon/rv515.c
+++ b/drivers/gpu/drm/radeon/rv515.c
@@ -53,46 +53,46 @@ void rv515_debugfs(struct radeon_device *rdev)
 	}
 }
 
-void rv515_ring_start(struct radeon_device *rdev, struct radeon_ring *ring)
+void rv515_ring_start(struct radeon_device *rdev)
 {
 	int r;
 
-	r = radeon_ring_lock(rdev, ring, 64);
+	r = radeon_ring_lock(rdev, 64);
 	if (r) {
 		return;
 	}
-	radeon_ring_write(ring, PACKET0(ISYNC_CNTL, 0));
-	radeon_ring_write(ring,
+	radeon_ring_write(rdev, PACKET0(ISYNC_CNTL, 0));
+	radeon_ring_write(rdev,
 			  ISYNC_ANY2D_IDLE3D |
 			  ISYNC_ANY3D_IDLE2D |
 			  ISYNC_WAIT_IDLEGUI |
 			  ISYNC_CPSCRATCH_IDLEGUI);
-	radeon_ring_write(ring, PACKET0(WAIT_UNTIL, 0));
-	radeon_ring_write(ring, WAIT_2D_IDLECLEAN | WAIT_3D_IDLECLEAN);
-	radeon_ring_write(ring, PACKET0(R300_DST_PIPE_CONFIG, 0));
-	radeon_ring_write(ring, R300_PIPE_AUTO_CONFIG);
-	radeon_ring_write(ring, PACKET0(GB_SELECT, 0));
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, PACKET0(GB_ENABLE, 0));
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, PACKET0(R500_SU_REG_DEST, 0));
-	radeon_ring_write(ring, (1 << rdev->num_gb_pipes) - 1);
-	radeon_ring_write(ring, PACKET0(VAP_INDEX_OFFSET, 0));
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, PACKET0(RB3D_DSTCACHE_CTLSTAT, 0));
-	radeon_ring_write(ring, RB3D_DC_FLUSH | RB3D_DC_FREE);
-	radeon_ring_write(ring, PACKET0(ZB_ZCACHE_CTLSTAT, 0));
-	radeon_ring_write(ring, ZC_FLUSH | ZC_FREE);
-	radeon_ring_write(ring, PACKET0(WAIT_UNTIL, 0));
-	radeon_ring_write(ring, WAIT_2D_IDLECLEAN | WAIT_3D_IDLECLEAN);
-	radeon_ring_write(ring, PACKET0(GB_AA_CONFIG, 0));
-	radeon_ring_write(ring, 0);
-	radeon_ring_write(ring, PACKET0(RB3D_DSTCACHE_CTLSTAT, 0));
-	radeon_ring_write(ring, RB3D_DC_FLUSH | RB3D_DC_FREE);
-	radeon_ring_write(ring, PACKET0(ZB_ZCACHE_CTLSTAT, 0));
-	radeon_ring_write(ring, ZC_FLUSH | ZC_FREE);
-	radeon_ring_write(ring, PACKET0(GB_MSPOS0, 0));
-	radeon_ring_write(ring,
+	radeon_ring_write(rdev, PACKET0(WAIT_UNTIL, 0));
+	radeon_ring_write(rdev, WAIT_2D_IDLECLEAN | WAIT_3D_IDLECLEAN);
+	radeon_ring_write(rdev, PACKET0(R300_DST_PIPE_CONFIG, 0));
+	radeon_ring_write(rdev, R300_PIPE_AUTO_CONFIG);
+	radeon_ring_write(rdev, PACKET0(GB_SELECT, 0));
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, PACKET0(GB_ENABLE, 0));
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, PACKET0(R500_SU_REG_DEST, 0));
+	radeon_ring_write(rdev, (1 << rdev->num_gb_pipes) - 1);
+	radeon_ring_write(rdev, PACKET0(VAP_INDEX_OFFSET, 0));
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, PACKET0(RB3D_DSTCACHE_CTLSTAT, 0));
+	radeon_ring_write(rdev, RB3D_DC_FLUSH | RB3D_DC_FREE);
+	radeon_ring_write(rdev, PACKET0(ZB_ZCACHE_CTLSTAT, 0));
+	radeon_ring_write(rdev, ZC_FLUSH | ZC_FREE);
+	radeon_ring_write(rdev, PACKET0(WAIT_UNTIL, 0));
+	radeon_ring_write(rdev, WAIT_2D_IDLECLEAN | WAIT_3D_IDLECLEAN);
+	radeon_ring_write(rdev, PACKET0(GB_AA_CONFIG, 0));
+	radeon_ring_write(rdev, 0);
+	radeon_ring_write(rdev, PACKET0(RB3D_DSTCACHE_CTLSTAT, 0));
+	radeon_ring_write(rdev, RB3D_DC_FLUSH | RB3D_DC_FREE);
+	radeon_ring_write(rdev, PACKET0(ZB_ZCACHE_CTLSTAT, 0));
+	radeon_ring_write(rdev, ZC_FLUSH | ZC_FREE);
+	radeon_ring_write(rdev, PACKET0(GB_MSPOS0, 0));
+	radeon_ring_write(rdev,
 			  ((6 << MS_X0_SHIFT) |
 			   (6 << MS_Y0_SHIFT) |
 			   (6 << MS_X1_SHIFT) |
@@ -101,8 +101,8 @@ void rv515_ring_start(struct radeon_device *rdev, struct radeon_ring *ring)
 			   (6 << MS_Y2_SHIFT) |
 			   (6 << MSBD0_Y_SHIFT) |
 			   (6 << MSBD0_X_SHIFT)));
-	radeon_ring_write(ring, PACKET0(GB_MSPOS1, 0));
-	radeon_ring_write(ring,
+	radeon_ring_write(rdev, PACKET0(GB_MSPOS1, 0));
+	radeon_ring_write(rdev,
 			  ((6 << MS_X3_SHIFT) |
 			   (6 << MS_Y3_SHIFT) |
 			   (6 << MS_X4_SHIFT) |
@@ -110,15 +110,15 @@ void rv515_ring_start(struct radeon_device *rdev, struct radeon_ring *ring)
 			   (6 << MS_X5_SHIFT) |
 			   (6 << MS_Y5_SHIFT) |
 			   (6 << MSBD1_SHIFT)));
-	radeon_ring_write(ring, PACKET0(GA_ENHANCE, 0));
-	radeon_ring_write(ring, GA_DEADLOCK_CNTL | GA_FASTSYNC_CNTL);
-	radeon_ring_write(ring, PACKET0(GA_POLY_MODE, 0));
-	radeon_ring_write(ring, FRONT_PTYPE_TRIANGE | BACK_PTYPE_TRIANGE);
-	radeon_ring_write(ring, PACKET0(GA_ROUND_MODE, 0));
-	radeon_ring_write(ring, GEOMETRY_ROUND_NEAREST | COLOR_ROUND_NEAREST);
-	radeon_ring_write(ring, PACKET0(0x20C8, 0));
-	radeon_ring_write(ring, 0);
-	radeon_ring_unlock_commit(rdev, ring);
+	radeon_ring_write(rdev, PACKET0(GA_ENHANCE, 0));
+	radeon_ring_write(rdev, GA_DEADLOCK_CNTL | GA_FASTSYNC_CNTL);
+	radeon_ring_write(rdev, PACKET0(GA_POLY_MODE, 0));
+	radeon_ring_write(rdev, FRONT_PTYPE_TRIANGE | BACK_PTYPE_TRIANGE);
+	radeon_ring_write(rdev, PACKET0(GA_ROUND_MODE, 0));
+	radeon_ring_write(rdev, GEOMETRY_ROUND_NEAREST | COLOR_ROUND_NEAREST);
+	radeon_ring_write(rdev, PACKET0(0x20C8, 0));
+	radeon_ring_write(rdev, 0);
+	radeon_ring_unlock_commit(rdev);
 }
 
 int rv515_mc_wait_for_idle(struct radeon_device *rdev)
@@ -379,12 +379,6 @@ static int rv515_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX);
-	if (r) {
-		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
-		return r;
-	}
-
 	/* Enable IRQ */
 	rs600_irq_set(rdev);
 	rdev->config.r300.hdp_cntl = RREG32(RADEON_HOST_PATH_CNTL);
@@ -394,15 +388,9 @@ static int rv515_startup(struct radeon_device *rdev)
 		dev_err(rdev->dev, "failed initializing CP (%d).\n", r);
 		return r;
 	}
-
-	r = radeon_ib_pool_start(rdev);
-	if (r)
-		return r;
-
-	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
+	r = r100_ib_init(rdev);
 	if (r) {
-		dev_err(rdev->dev, "failed testing IB (%d).\n", r);
-		rdev->accel_working = false;
+		dev_err(rdev->dev, "failed initializing IB (%d).\n", r);
 		return r;
 	}
 	return 0;
@@ -410,8 +398,6 @@ static int rv515_startup(struct radeon_device *rdev)
 
 int rv515_resume(struct radeon_device *rdev)
 {
-	int r;
-
 	/* Make sur GART are not working */
 	if (rdev->flags & RADEON_IS_PCIE)
 		rv370_pcie_gart_disable(rdev);
@@ -429,13 +415,7 @@ int rv515_resume(struct radeon_device *rdev)
 	rv515_clock_startup(rdev);
 	/* Initialize surface registers */
 	radeon_surface_init(rdev);
-
-	rdev->accel_working = true;
-	r =  rv515_startup(rdev);
-	if (r) {
-		rdev->accel_working = false;
-	}
-	return r;
+	return rv515_startup(rdev);
 }
 
 int rv515_suspend(struct radeon_device *rdev)
@@ -531,14 +511,7 @@ int rv515_init(struct radeon_device *rdev)
 	if (r)
 		return r;
 	rv515_set_safe_registers(rdev);
-
-	r = radeon_ib_pool_init(rdev);
 	rdev->accel_working = true;
-	if (r) {
-		dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
-		rdev->accel_working = false;
-	}
-
 	r = rv515_startup(rdev);
 	if (r) {
 		/* Somethings want wront with the accel init stop accel */
diff --git a/drivers/gpu/drm/radeon/rv770.c b/drivers/gpu/drm/radeon/rv770.c
index c73ea54..51d20aa 100644
--- a/drivers/gpu/drm/radeon/rv770.c
+++ b/drivers/gpu/drm/radeon/rv770.c
@@ -129,7 +129,7 @@ int rv770_pcie_gart_enable(struct radeon_device *rdev)
 	u32 tmp;
 	int r, i;
 
-	if (rdev->gart.robj == NULL) {
+	if (rdev->gart.table.vram.robj == NULL) {
 		dev_err(rdev->dev, "No VRAM object for PCIE GART.\n");
 		return -EINVAL;
 	}
@@ -168,9 +168,6 @@ int rv770_pcie_gart_enable(struct radeon_device *rdev)
 		WREG32(VM_CONTEXT0_CNTL + (i * 4), 0);
 
 	r600_pcie_gart_tlb_flush(rdev);
-	DRM_INFO("PCIE GART of %uM enabled (table at 0x%016llX).\n",
-		 (unsigned)(rdev->mc.gtt_size >> 20),
-		 (unsigned long long)rdev->gart.table_addr);
 	rdev->gart.ready = true;
 	return 0;
 }
@@ -178,7 +175,7 @@ int rv770_pcie_gart_enable(struct radeon_device *rdev)
 void rv770_pcie_gart_disable(struct radeon_device *rdev)
 {
 	u32 tmp;
-	int i;
+	int i, r;
 
 	/* Disable all tables */
 	for (i = 0; i < 7; i++)
@@ -198,7 +195,14 @@ void rv770_pcie_gart_disable(struct radeon_device *rdev)
 	WREG32(MC_VM_MB_L1_TLB1_CNTL, tmp);
 	WREG32(MC_VM_MB_L1_TLB2_CNTL, tmp);
 	WREG32(MC_VM_MB_L1_TLB3_CNTL, tmp);
-	radeon_gart_table_vram_unpin(rdev);
+	if (rdev->gart.table.vram.robj) {
+		r = radeon_bo_reserve(rdev->gart.table.vram.robj, false);
+		if (likely(r == 0)) {
+			radeon_bo_kunmap(rdev->gart.table.vram.robj);
+			radeon_bo_unpin(rdev->gart.table.vram.robj);
+			radeon_bo_unreserve(rdev->gart.table.vram.robj);
+		}
+	}
 }
 
 void rv770_pcie_gart_fini(struct radeon_device *rdev)
@@ -282,7 +286,7 @@ static void rv770_mc_program(struct radeon_device *rdev)
 		WREG32(MC_VM_SYSTEM_APERTURE_HIGH_ADDR,
 			rdev->mc.vram_end >> 12);
 	}
-	WREG32(MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR, rdev->vram_scratch.gpu_addr >> 12);
+	WREG32(MC_VM_SYSTEM_APERTURE_DEFAULT_ADDR, 0);
 	tmp = ((rdev->mc.vram_end >> 24) & 0xFFFF) << 16;
 	tmp |= ((rdev->mc.vram_start >> 24) & 0xFFFF);
 	WREG32(MC_VM_FB_LOCATION, tmp);
@@ -359,7 +363,7 @@ static int rv770_cp_load_microcode(struct radeon_device *rdev)
 void r700_cp_fini(struct radeon_device *rdev)
 {
 	r700_cp_stop(rdev);
-	radeon_ring_fini(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
+	radeon_ring_fini(rdev);
 }
 
 /*
@@ -732,7 +736,6 @@ static void rv770_gpu_init(struct radeon_device *rdev)
 								(cc_rb_backend_disable >> 16));
 
 	rdev->config.rv770.tile_config = gb_tiling_config;
-	rdev->config.rv770.backend_map = backend_map;
 	gb_tiling_config |= BACKEND_MAP(backend_map);
 
 	WREG32(GB_TILING_CONFIG, gb_tiling_config);
@@ -959,6 +962,54 @@ static void rv770_gpu_init(struct radeon_device *rdev)
 
 }
 
+static int rv770_vram_scratch_init(struct radeon_device *rdev)
+{
+	int r;
+	u64 gpu_addr;
+
+	if (rdev->vram_scratch.robj == NULL) {
+		r = radeon_bo_create(rdev, RADEON_GPU_PAGE_SIZE,
+				     PAGE_SIZE, true, RADEON_GEM_DOMAIN_VRAM,
+				     &rdev->vram_scratch.robj);
+		if (r) {
+			return r;
+		}
+	}
+
+	r = radeon_bo_reserve(rdev->vram_scratch.robj, false);
+	if (unlikely(r != 0))
+		return r;
+	r = radeon_bo_pin(rdev->vram_scratch.robj,
+			  RADEON_GEM_DOMAIN_VRAM, &gpu_addr);
+	if (r) {
+		radeon_bo_unreserve(rdev->vram_scratch.robj);
+		return r;
+	}
+	r = radeon_bo_kmap(rdev->vram_scratch.robj,
+				(void **)&rdev->vram_scratch.ptr);
+	if (r)
+		radeon_bo_unpin(rdev->vram_scratch.robj);
+	radeon_bo_unreserve(rdev->vram_scratch.robj);
+
+	return r;
+}
+
+static void rv770_vram_scratch_fini(struct radeon_device *rdev)
+{
+	int r;
+
+	if (rdev->vram_scratch.robj == NULL) {
+		return;
+	}
+	r = radeon_bo_reserve(rdev->vram_scratch.robj, false);
+	if (likely(r == 0)) {
+		radeon_bo_kunmap(rdev->vram_scratch.robj);
+		radeon_bo_unpin(rdev->vram_scratch.robj);
+		radeon_bo_unreserve(rdev->vram_scratch.robj);
+	}
+	radeon_bo_unref(&rdev->vram_scratch.robj);
+}
+
 void r700_vram_gtt_location(struct radeon_device *rdev, struct radeon_mc *mc)
 {
 	u64 size_bf, size_af;
@@ -1045,7 +1096,6 @@ int rv770_mc_init(struct radeon_device *rdev)
 
 static int rv770_startup(struct radeon_device *rdev)
 {
-	struct radeon_ring *ring = &rdev->ring[RADEON_RING_TYPE_GFX_INDEX];
 	int r;
 
 	/* enable pcie gen2 link */
@@ -1059,10 +1109,6 @@ static int rv770_startup(struct radeon_device *rdev)
 		}
 	}
 
-	r = r600_vram_scratch_init(rdev);
-	if (r)
-		return r;
-
 	rv770_mc_program(rdev);
 	if (rdev->flags & RADEON_IS_AGP) {
 		rv770_agp_enable(rdev);
@@ -1071,12 +1117,14 @@ static int rv770_startup(struct radeon_device *rdev)
 		if (r)
 			return r;
 	}
-
+	r = rv770_vram_scratch_init(rdev);
+	if (r)
+		return r;
 	rv770_gpu_init(rdev);
 	r = r600_blit_init(rdev);
 	if (r) {
 		r600_blit_fini(rdev);
-		rdev->asic->copy.copy = NULL;
+		rdev->asic->copy = NULL;
 		dev_warn(rdev->dev, "failed blitter (%d) falling back to memcpy\n", r);
 	}
 
@@ -1085,12 +1133,6 @@ static int rv770_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = radeon_fence_driver_start_ring(rdev, RADEON_RING_TYPE_GFX_INDEX);
-	if (r) {
-		dev_err(rdev->dev, "failed initializing CP fences (%d).\n", r);
-		return r;
-	}
-
 	/* Enable IRQ */
 	r = r600_irq_init(rdev);
 	if (r) {
@@ -1100,9 +1142,7 @@ static int rv770_startup(struct radeon_device *rdev)
 	}
 	r600_irq_set(rdev);
 
-	r = radeon_ring_init(rdev, ring, ring->ring_size, RADEON_WB_CP_RPTR_OFFSET,
-			     R600_CP_RB_RPTR, R600_CP_RB_WPTR,
-			     0, 0xfffff, RADEON_CP_PACKET2);
+	r = radeon_ring_init(rdev, rdev->cp.ring_size);
 	if (r)
 		return r;
 	r = rv770_cp_load_microcode(rdev);
@@ -1112,17 +1152,6 @@ static int rv770_startup(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = radeon_ib_pool_start(rdev);
-	if (r)
-		return r;
-
-	r = radeon_ib_test(rdev, RADEON_RING_TYPE_GFX_INDEX, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX]);
-	if (r) {
-		dev_err(rdev->dev, "IB test failed (%d).\n", r);
-		rdev->accel_working = false;
-		return r;
-	}
-
 	return 0;
 }
 
@@ -1137,11 +1166,15 @@ int rv770_resume(struct radeon_device *rdev)
 	/* post card */
 	atom_asic_init(rdev->mode_info.atom_context);
 
-	rdev->accel_working = true;
 	r = rv770_startup(rdev);
 	if (r) {
 		DRM_ERROR("r600 startup failed on resume\n");
-		rdev->accel_working = false;
+		return r;
+	}
+
+	r = r600_ib_test(rdev);
+	if (r) {
+		DRM_ERROR("radeon: failed testing IB (%d).\n", r);
 		return r;
 	}
 
@@ -1157,16 +1190,23 @@ int rv770_resume(struct radeon_device *rdev)
 
 int rv770_suspend(struct radeon_device *rdev)
 {
+	int r;
+
 	r600_audio_fini(rdev);
-	radeon_ib_pool_suspend(rdev);
-	r600_blit_suspend(rdev);
 	/* FIXME: we should wait for ring to be empty */
 	r700_cp_stop(rdev);
-	rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ready = false;
+	rdev->cp.ready = false;
 	r600_irq_suspend(rdev);
 	radeon_wb_disable(rdev);
 	rv770_pcie_gart_disable(rdev);
-
+	/* unpin shaders bo */
+	if (rdev->r600_blit.shader_obj) {
+		r = radeon_bo_reserve(rdev->r600_blit.shader_obj, false);
+		if (likely(r == 0)) {
+			radeon_bo_unpin(rdev->r600_blit.shader_obj);
+			radeon_bo_unreserve(rdev->r600_blit.shader_obj);
+		}
+	}
 	return 0;
 }
 
@@ -1234,8 +1274,8 @@ int rv770_init(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	rdev->ring[RADEON_RING_TYPE_GFX_INDEX].ring_obj = NULL;
-	r600_ring_init(rdev, &rdev->ring[RADEON_RING_TYPE_GFX_INDEX], 1024 * 1024);
+	rdev->cp.ring_obj = NULL;
+	r600_ring_init(rdev, 1024 * 1024);
 
 	rdev->ih.ring_obj = NULL;
 	r600_ih_ring_init(rdev, 64 * 1024);
@@ -1244,24 +1284,30 @@ int rv770_init(struct radeon_device *rdev)
 	if (r)
 		return r;
 
-	r = radeon_ib_pool_init(rdev);
 	rdev->accel_working = true;
-	if (r) {
-		dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
-		rdev->accel_working = false;
-	}
-
 	r = rv770_startup(rdev);
 	if (r) {
 		dev_err(rdev->dev, "disabling GPU acceleration\n");
 		r700_cp_fini(rdev);
 		r600_irq_fini(rdev);
 		radeon_wb_fini(rdev);
-		r100_ib_fini(rdev);
 		radeon_irq_kms_fini(rdev);
 		rv770_pcie_gart_fini(rdev);
 		rdev->accel_working = false;
 	}
+	if (rdev->accel_working) {
+		r = radeon_ib_pool_init(rdev);
+		if (r) {
+			dev_err(rdev->dev, "IB initialization failed (%d).\n", r);
+			rdev->accel_working = false;
+		} else {
+			r = r600_ib_test(rdev);
+			if (r) {
+				dev_err(rdev->dev, "IB test failed (%d).\n", r);
+				rdev->accel_working = false;
+			}
+		}
+	}
 
 	r = r600_audio_init(rdev);
 	if (r) {
@@ -1278,12 +1324,11 @@ void rv770_fini(struct radeon_device *rdev)
 	r700_cp_fini(rdev);
 	r600_irq_fini(rdev);
 	radeon_wb_fini(rdev);
-	r100_ib_fini(rdev);
+	radeon_ib_pool_fini(rdev);
 	radeon_irq_kms_fini(rdev);
 	rv770_pcie_gart_fini(rdev);
-	r600_vram_scratch_fini(rdev);
+	rv770_vram_scratch_fini(rdev);
 	radeon_gem_fini(rdev);
-	radeon_semaphore_driver_fini(rdev);
 	radeon_fence_driver_fini(rdev);
 	radeon_agp_fini(rdev);
 	radeon_bo_fini(rdev);
diff --git a/drivers/gpu/drm/savage/savage_drv.c b/drivers/gpu/drm/savage/savage_drv.c
index 89afe0b..6464490 100644
--- a/drivers/gpu/drm/savage/savage_drv.c
+++ b/drivers/gpu/drm/savage/savage_drv.c
@@ -23,8 +23,6 @@
  * WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  */
 
-#include <linux/module.h>
-
 #include "drmP.h"
 #include "savage_drm.h"
 #include "savage_drv.h"
@@ -35,17 +33,6 @@ static struct pci_device_id pciidlist[] = {
 	savage_PCI_IDS
 };
 
-static const struct file_operations savage_driver_fops = {
-	.owner = THIS_MODULE,
-	.open = drm_open,
-	.release = drm_release,
-	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_mmap,
-	.poll = drm_poll,
-	.fasync = drm_fasync,
-	.llseek = noop_llseek,
-};
-
 static struct drm_driver driver = {
 	.driver_features =
 	    DRIVER_USE_AGP | DRIVER_USE_MTRR | DRIVER_HAVE_DMA | DRIVER_PCI_DMA,
@@ -57,7 +44,17 @@ static struct drm_driver driver = {
 	.reclaim_buffers = savage_reclaim_buffers,
 	.ioctls = savage_ioctls,
 	.dma_ioctl = savage_bci_buffers,
-	.fops = &savage_driver_fops,
+	.fops = {
+		 .owner = THIS_MODULE,
+		 .open = drm_open,
+		 .release = drm_release,
+		 .unlocked_ioctl = drm_ioctl,
+		 .mmap = drm_mmap,
+		 .poll = drm_poll,
+		 .fasync = drm_fasync,
+		 .llseek = noop_llseek,
+	},
+
 	.name = DRIVER_NAME,
 	.desc = DRIVER_DESC,
 	.date = DRIVER_DATE,
diff --git a/drivers/gpu/drm/savage/savage_state.c b/drivers/gpu/drm/savage/savage_state.c
index 031aaaf..8a3e315 100644
--- a/drivers/gpu/drm/savage/savage_state.c
+++ b/drivers/gpu/drm/savage/savage_state.c
@@ -1057,8 +1057,7 @@ int savage_bci_cmdbuf(struct drm_device *dev, void *data, struct drm_file *file_
 				DRM_ERROR("indexed drawing command extends "
 					  "beyond end of command buffer\n");
 				DMA_FLUSH();
-				ret = -EINVAL;
-				goto done;
+				return -EINVAL;
 			}
 			/* fall through */
 		case SAVAGE_CMD_DMA_PRIM:
@@ -1077,7 +1076,7 @@ int savage_bci_cmdbuf(struct drm_device *dev, void *data, struct drm_file *file_
 				      cmdbuf->vb_stride,
 				      cmdbuf->nbox, cmdbuf->box_addr);
 				if (ret != 0)
-					goto done;
+					return ret;
 				first_draw_cmd = NULL;
 			}
 		}
diff --git a/drivers/gpu/drm/sis/sis_drv.c b/drivers/gpu/drm/sis/sis_drv.c
index 30d98d1..46d5be6 100644
--- a/drivers/gpu/drm/sis/sis_drv.c
+++ b/drivers/gpu/drm/sis/sis_drv.c
@@ -25,8 +25,6 @@
  *
  */
 
-#include <linux/module.h>
-
 #include "drmP.h"
 #include "sis_drm.h"
 #include "sis_drv.h"
@@ -40,8 +38,7 @@ static struct pci_device_id pciidlist[] = {
 static int sis_driver_load(struct drm_device *dev, unsigned long chipset)
 {
 	drm_sis_private_t *dev_priv;
-
-	pci_set_master(dev->pdev);
+	int ret;
 
 	dev_priv = kzalloc(sizeof(drm_sis_private_t), GFP_KERNEL);
 	if (dev_priv == NULL)
@@ -49,69 +46,43 @@ static int sis_driver_load(struct drm_device *dev, unsigned long chipset)
 
 	dev->dev_private = (void *)dev_priv;
 	dev_priv->chipset = chipset;
-	idr_init(&dev->object_name_idr);
+	ret = drm_sman_init(&dev_priv->sman, 2, 12, 8);
+	if (ret)
+		kfree(dev_priv);
 
-	return 0;
+	return ret;
 }
 
 static int sis_driver_unload(struct drm_device *dev)
 {
 	drm_sis_private_t *dev_priv = dev->dev_private;
 
-	idr_remove_all(&dev_priv->object_idr);
-	idr_destroy(&dev_priv->object_idr);
-
+	drm_sman_takedown(&dev_priv->sman);
 	kfree(dev_priv);
 
 	return 0;
 }
 
-static const struct file_operations sis_driver_fops = {
-	.owner = THIS_MODULE,
-	.open = drm_open,
-	.release = drm_release,
-	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_mmap,
-	.poll = drm_poll,
-	.fasync = drm_fasync,
-	.llseek = noop_llseek,
-};
-
-static int sis_driver_open(struct drm_device *dev, struct drm_file *file)
-{
-	struct sis_file_private *file_priv;
-
-	DRM_DEBUG_DRIVER("\n");
-	file_priv = kmalloc(sizeof(*file_priv), GFP_KERNEL);
-	if (!file_priv)
-		return -ENOMEM;
-
-	file->driver_priv = file_priv;
-
-	INIT_LIST_HEAD(&file_priv->obj_list);
-
-	return 0;
-}
-
-void sis_driver_postclose(struct drm_device *dev, struct drm_file *file)
-{
-	struct sis_file_private *file_priv = file->driver_priv;
-
-	kfree(file_priv);
-}
-
 static struct drm_driver driver = {
 	.driver_features = DRIVER_USE_AGP | DRIVER_USE_MTRR,
 	.load = sis_driver_load,
 	.unload = sis_driver_unload,
-	.open = sis_driver_open,
-	.postclose = sis_driver_postclose,
 	.dma_quiescent = sis_idle,
 	.reclaim_buffers = NULL,
 	.reclaim_buffers_idlelocked = sis_reclaim_buffers_locked,
 	.lastclose = sis_lastclose,
 	.ioctls = sis_ioctls,
-	.fops = &sis_driver_fops,
+	.fops = {
+		 .owner = THIS_MODULE,
+		 .open = drm_open,
+		 .release = drm_release,
+		 .unlocked_ioctl = drm_ioctl,
+		 .mmap = drm_mmap,
+		 .poll = drm_poll,
+		 .fasync = drm_fasync,
+		 .llseek = noop_llseek,
+	},
+
 	.name = DRIVER_NAME,
 	.desc = DRIVER_DESC,
 	.date = DRIVER_DATE,
diff --git a/drivers/gpu/drm/sis/sis_drv.h b/drivers/gpu/drm/sis/sis_drv.h
index 573758b..ef940ba 100644
--- a/drivers/gpu/drm/sis/sis_drv.h
+++ b/drivers/gpu/drm/sis/sis_drv.h
@@ -44,25 +44,22 @@ enum sis_family {
 	SIS_CHIP_315 = 1,
 };
 
-#include "drm_mm.h"
+#include "drm_sman.h"
 
 
 #define SIS_BASE (dev_priv->mmio)
-#define SIS_READ(reg)         DRM_READ32(SIS_BASE, reg)
-#define SIS_WRITE(reg, val)   DRM_WRITE32(SIS_BASE, reg, val)
+#define SIS_READ(reg)         DRM_READ32(SIS_BASE, reg);
+#define SIS_WRITE(reg, val)   DRM_WRITE32(SIS_BASE, reg, val);
 
 typedef struct drm_sis_private {
 	drm_local_map_t *mmio;
 	unsigned int idle_fault;
+	struct drm_sman sman;
 	unsigned int chipset;
 	int vram_initialized;
 	int agp_initialized;
 	unsigned long vram_offset;
 	unsigned long agp_offset;
-	struct drm_mm vram_mm;
-	struct drm_mm agp_mm;
-	/** Mapping of userspace keys to mm objects */
-	struct idr object_idr;
 } drm_sis_private_t;
 
 extern int sis_idle(struct drm_device *dev);
diff --git a/drivers/gpu/drm/sis/sis_mm.c b/drivers/gpu/drm/sis/sis_mm.c
index dd4a316..7fe2b63 100644
--- a/drivers/gpu/drm/sis/sis_mm.c
+++ b/drivers/gpu/drm/sis/sis_mm.c
@@ -41,18 +41,40 @@
 #define AGP_TYPE 1
 
 
-struct sis_memblock {
-	struct drm_mm_node mm_node;
-	struct sis_memreq req;
-	struct list_head owner_list;
-};
-
 #if defined(CONFIG_FB_SIS) || defined(CONFIG_FB_SIS_MODULE)
 /* fb management via fb device */
 
 #define SIS_MM_ALIGN_SHIFT 0
 #define SIS_MM_ALIGN_MASK 0
 
+static void *sis_sman_mm_allocate(void *private, unsigned long size,
+				  unsigned alignment)
+{
+	struct sis_memreq req;
+
+	req.size = size;
+	sis_malloc(&req);
+	if (req.size == 0)
+		return NULL;
+	else
+		return (void *)(unsigned long)~req.offset;
+}
+
+static void sis_sman_mm_free(void *private, void *ref)
+{
+	sis_free(~((unsigned long)ref));
+}
+
+static void sis_sman_mm_destroy(void *private)
+{
+	;
+}
+
+static unsigned long sis_sman_mm_offset(void *private, void *ref)
+{
+	return ~((unsigned long)ref);
+}
+
 #else /* CONFIG_FB_SIS[_MODULE] */
 
 #define SIS_MM_ALIGN_SHIFT 4
@@ -64,11 +86,30 @@ static int sis_fb_init(struct drm_device *dev, void *data, struct drm_file *file
 {
 	drm_sis_private_t *dev_priv = dev->dev_private;
 	drm_sis_fb_t *fb = data;
+	int ret;
 
 	mutex_lock(&dev->struct_mutex);
-	/* Unconditionally init the drm_mm, even though we don't use it when the
-	 * fb sis driver is available - make cleanup easier. */
-	drm_mm_init(&dev_priv->vram_mm, 0, fb->size >> SIS_MM_ALIGN_SHIFT);
+#if defined(CONFIG_FB_SIS) || defined(CONFIG_FB_SIS_MODULE)
+	{
+		struct drm_sman_mm sman_mm;
+		sman_mm.private = (void *)0xFFFFFFFF;
+		sman_mm.allocate = sis_sman_mm_allocate;
+		sman_mm.free = sis_sman_mm_free;
+		sman_mm.destroy = sis_sman_mm_destroy;
+		sman_mm.offset = sis_sman_mm_offset;
+		ret =
+		    drm_sman_set_manager(&dev_priv->sman, VIDEO_TYPE, &sman_mm);
+	}
+#else
+	ret = drm_sman_set_range(&dev_priv->sman, VIDEO_TYPE, 0,
+				 fb->size >> SIS_MM_ALIGN_SHIFT);
+#endif
+
+	if (ret) {
+		DRM_ERROR("VRAM memory manager initialisation error\n");
+		mutex_unlock(&dev->struct_mutex);
+		return ret;
+	}
 
 	dev_priv->vram_initialized = 1;
 	dev_priv->vram_offset = fb->offset;
@@ -79,15 +120,13 @@ static int sis_fb_init(struct drm_device *dev, void *data, struct drm_file *file
 	return 0;
 }
 
-static int sis_drm_alloc(struct drm_device *dev, struct drm_file *file,
+static int sis_drm_alloc(struct drm_device *dev, struct drm_file *file_priv,
 			 void *data, int pool)
 {
 	drm_sis_private_t *dev_priv = dev->dev_private;
 	drm_sis_mem_t *mem = data;
-	int retval = 0, user_key;
-	struct sis_memblock *item;
-	struct sis_file_private *file_priv = file->driver_priv;
-	unsigned long offset;
+	int retval = 0;
+	struct drm_memblock_item *item;
 
 	mutex_lock(&dev->struct_mutex);
 
@@ -99,68 +138,25 @@ static int sis_drm_alloc(struct drm_device *dev, struct drm_file *file,
 		return -EINVAL;
 	}
 
-	item = kzalloc(sizeof(*item), GFP_KERNEL);
-	if (!item) {
-		retval = -ENOMEM;
-		goto fail_alloc;
-	}
-
 	mem->size = (mem->size + SIS_MM_ALIGN_MASK) >> SIS_MM_ALIGN_SHIFT;
-	if (pool == AGP_TYPE) {
-		retval = drm_mm_insert_node(&dev_priv->agp_mm,
-					    &item->mm_node,
-					    mem->size, 0);
-		offset = item->mm_node.start;
-	} else {
-#if defined(CONFIG_FB_SIS) || defined(CONFIG_FB_SIS_MODULE)
-		item->req.size = mem->size;
-		sis_malloc(&item->req);
-		if (item->req.size == 0)
-			retval = -ENOMEM;
-		offset = item->req.offset;
-#else
-		retval = drm_mm_insert_node(&dev_priv->vram_mm,
-					    &item->mm_node,
-					    mem->size, 0);
-		offset = item->mm_node.start;
-#endif
-	}
-	if (retval)
-		goto fail_alloc;
+	item = drm_sman_alloc(&dev_priv->sman, pool, mem->size, 0,
+			      (unsigned long)file_priv);
 
-again:
-	if (idr_pre_get(&dev_priv->object_idr, GFP_KERNEL) == 0) {
+	mutex_unlock(&dev->struct_mutex);
+	if (item) {
+		mem->offset = ((pool == 0) ?
+			      dev_priv->vram_offset : dev_priv->agp_offset) +
+		    (item->mm->
+		     offset(item->mm, item->mm_info) << SIS_MM_ALIGN_SHIFT);
+		mem->free = item->user_hash.key;
+		mem->size = mem->size << SIS_MM_ALIGN_SHIFT;
+	} else {
+		mem->offset = 0;
+		mem->size = 0;
+		mem->free = 0;
 		retval = -ENOMEM;
-		goto fail_idr;
 	}
 
-	retval = idr_get_new_above(&dev_priv->object_idr, item, 1, &user_key);
-	if (retval == -EAGAIN)
-		goto again;
-	if (retval)
-		goto fail_idr;
-
-	list_add(&item->owner_list, &file_priv->obj_list);
-	mutex_unlock(&dev->struct_mutex);
-
-	mem->offset = ((pool == 0) ?
-		      dev_priv->vram_offset : dev_priv->agp_offset) +
-	    (offset << SIS_MM_ALIGN_SHIFT);
-	mem->free = user_key;
-	mem->size = mem->size << SIS_MM_ALIGN_SHIFT;
-
-	return 0;
-
-fail_idr:
-	drm_mm_remove_node(&item->mm_node);
-fail_alloc:
-	kfree(item);
-	mutex_unlock(&dev->struct_mutex);
-
-	mem->offset = 0;
-	mem->size = 0;
-	mem->free = 0;
-
 	DRM_DEBUG("alloc %d, size = %d, offset = %d\n", pool, mem->size,
 		  mem->offset);
 
@@ -171,28 +167,14 @@ static int sis_drm_free(struct drm_device *dev, void *data, struct drm_file *fil
 {
 	drm_sis_private_t *dev_priv = dev->dev_private;
 	drm_sis_mem_t *mem = data;
-	struct sis_memblock *obj;
+	int ret;
 
 	mutex_lock(&dev->struct_mutex);
-	obj = idr_find(&dev_priv->object_idr, mem->free);
-	if (obj == NULL) {
-		mutex_unlock(&dev->struct_mutex);
-		return -EINVAL;
-	}
-
-	idr_remove(&dev_priv->object_idr, mem->free);
-	list_del(&obj->owner_list);
-	if (drm_mm_node_allocated(&obj->mm_node))
-		drm_mm_remove_node(&obj->mm_node);
-#if defined(CONFIG_FB_SIS) || defined(CONFIG_FB_SIS_MODULE)
-	else
-		sis_free(obj->req.offset);
-#endif
-	kfree(obj);
+	ret = drm_sman_free_key(&dev_priv->sman, mem->free);
 	mutex_unlock(&dev->struct_mutex);
 	DRM_DEBUG("free = 0x%lx\n", mem->free);
 
-	return 0;
+	return ret;
 }
 
 static int sis_fb_alloc(struct drm_device *dev, void *data,
@@ -206,10 +188,18 @@ static int sis_ioctl_agp_init(struct drm_device *dev, void *data,
 {
 	drm_sis_private_t *dev_priv = dev->dev_private;
 	drm_sis_agp_t *agp = data;
+	int ret;
 	dev_priv = dev->dev_private;
 
 	mutex_lock(&dev->struct_mutex);
-	drm_mm_init(&dev_priv->agp_mm, 0, agp->size >> SIS_MM_ALIGN_SHIFT);
+	ret = drm_sman_set_range(&dev_priv->sman, AGP_TYPE, 0,
+				 agp->size >> SIS_MM_ALIGN_SHIFT);
+
+	if (ret) {
+		DRM_ERROR("AGP memory manager initialisation error\n");
+		mutex_unlock(&dev->struct_mutex);
+		return ret;
+	}
 
 	dev_priv->agp_initialized = 1;
 	dev_priv->agp_offset = agp->offset;
@@ -303,26 +293,20 @@ void sis_lastclose(struct drm_device *dev)
 		return;
 
 	mutex_lock(&dev->struct_mutex);
-	if (dev_priv->vram_initialized) {
-		drm_mm_takedown(&dev_priv->vram_mm);
-		dev_priv->vram_initialized = 0;
-	}
-	if (dev_priv->agp_initialized) {
-		drm_mm_takedown(&dev_priv->agp_mm);
-		dev_priv->agp_initialized = 0;
-	}
+	drm_sman_cleanup(&dev_priv->sman);
+	dev_priv->vram_initialized = 0;
+	dev_priv->agp_initialized = 0;
 	dev_priv->mmio = NULL;
 	mutex_unlock(&dev->struct_mutex);
 }
 
 void sis_reclaim_buffers_locked(struct drm_device *dev,
-				struct drm_file *file)
+				struct drm_file *file_priv)
 {
-	struct sis_file_private *file_priv = file->driver_priv;
-	struct sis_memblock *entry, *next;
+	drm_sis_private_t *dev_priv = dev->dev_private;
 
 	mutex_lock(&dev->struct_mutex);
-	if (list_empty(&file_priv->obj_list)) {
+	if (drm_sman_owner_clean(&dev_priv->sman, (unsigned long)file_priv)) {
 		mutex_unlock(&dev->struct_mutex);
 		return;
 	}
@@ -330,18 +314,7 @@ void sis_reclaim_buffers_locked(struct drm_device *dev,
 	if (dev->driver->dma_quiescent)
 		dev->driver->dma_quiescent(dev);
 
-
-	list_for_each_entry_safe(entry, next, &file_priv->obj_list,
-				 owner_list) {
-		list_del(&entry->owner_list);
-		if (drm_mm_node_allocated(&entry->mm_node))
-			drm_mm_remove_node(&entry->mm_node);
-#if defined(CONFIG_FB_SIS) || defined(CONFIG_FB_SIS_MODULE)
-		else
-			sis_free(entry->req.offset);
-#endif
-		kfree(entry);
-	}
+	drm_sman_owner_cleanup(&dev_priv->sman, (unsigned long)file_priv);
 	mutex_unlock(&dev->struct_mutex);
 	return;
 }
diff --git a/drivers/gpu/drm/tdfx/tdfx_drv.c b/drivers/gpu/drm/tdfx/tdfx_drv.c
index 1613c78..8bf9881 100644
--- a/drivers/gpu/drm/tdfx/tdfx_drv.c
+++ b/drivers/gpu/drm/tdfx/tdfx_drv.c
@@ -30,8 +30,6 @@
  *    Gareth Hughes <gareth@valinux.com>
  */
 
-#include <linux/module.h>
-
 #include "drmP.h"
 #include "tdfx_drv.h"
 
@@ -41,21 +39,20 @@ static struct pci_device_id pciidlist[] = {
 	tdfx_PCI_IDS
 };
 
-static const struct file_operations tdfx_driver_fops = {
-	.owner = THIS_MODULE,
-	.open = drm_open,
-	.release = drm_release,
-	.unlocked_ioctl = drm_ioctl,
-	.mmap = drm_mmap,
-	.poll = drm_poll,
-	.fasync = drm_fasync,
-	.llseek = noop_llseek,
-};
-
 static struct drm_driver driver = {
 	.driver_features = DRIVER_USE_MTRR,
 	.reclaim_buffers = drm_core_reclaim_buffers,
-	.fops = &tdfx_driver_fops,
+	.fops = {
+		 .owner = THIS_MODULE,
+		 .open = drm_open,
+		 .release = drm_release,
+		 .unlocked_ioctl = drm_ioctl,
+		 .mmap = drm_mmap,
+		 .poll = drm_poll,
+		 .fasync = drm_fasync,
+		 .llseek = noop_llseek,
+	},
+
 	.name = DRIVER_NAME,
 	.desc = DRIVER_DESC,
 	.date = DRIVER_DATE,
diff --git a/drivers/gpu/drm/vmwgfx/Kconfig b/drivers/gpu/drm/vmwgfx/Kconfig
index 794ff67..30ad133 100644
--- a/drivers/gpu/drm/vmwgfx/Kconfig
+++ b/drivers/gpu/drm/vmwgfx/Kconfig
@@ -7,8 +7,7 @@ config DRM_VMWGFX
 	select FB_CFB_IMAGEBLIT
 	select DRM_TTM
 	help
-	  Choose this option if you would like to run 3D acceleration
-	  in a VMware virtual machine.
-	  This is a KMS enabled DRM driver for the VMware SVGA2
-	  virtual hardware.
-	  The compiled module will be called "vmwgfx.ko".
+	  KMS enabled DRM driver for SVGA2 virtual hardware.
+
+	  If unsure say n. The compiled module will be
+	  called vmwgfx.ko
diff --git a/drivers/gpu/drm/vmwgfx/Makefile b/drivers/gpu/drm/vmwgfx/Makefile
index 586869c..c9281a1 100644
--- a/drivers/gpu/drm/vmwgfx/Makefile
+++ b/drivers/gpu/drm/vmwgfx/Makefile
@@ -4,7 +4,6 @@ ccflags-y := -Iinclude/drm
 vmwgfx-y := vmwgfx_execbuf.o vmwgfx_gmr.o vmwgfx_kms.o vmwgfx_drv.o \
 	    vmwgfx_fb.o vmwgfx_ioctl.o vmwgfx_resource.o vmwgfx_buffer.o \
 	    vmwgfx_fifo.o vmwgfx_irq.o vmwgfx_ldu.o vmwgfx_ttm_glue.o \
-	    vmwgfx_overlay.o vmwgfx_marker.o vmwgfx_gmrid_manager.o \
-	    vmwgfx_fence.o vmwgfx_dmabuf.o vmwgfx_scrn.o
+	    vmwgfx_overlay.o vmwgfx_fence.o vmwgfx_gmrid_manager.o
 
 obj-$(CONFIG_DRM_VMWGFX) := vmwgfx.o
diff --git a/drivers/gpu/drm/vmwgfx/svga3d_reg.h b/drivers/gpu/drm/vmwgfx/svga3d_reg.h
index d0e085e..77cb453 100644
--- a/drivers/gpu/drm/vmwgfx/svga3d_reg.h
+++ b/drivers/gpu/drm/vmwgfx/svga3d_reg.h
@@ -57,8 +57,7 @@ typedef enum {
    SVGA3D_HWVERSION_WS6_B1    = SVGA3D_MAKE_HWVERSION(1, 1),
    SVGA3D_HWVERSION_FUSION_11 = SVGA3D_MAKE_HWVERSION(1, 4),
    SVGA3D_HWVERSION_WS65_B1   = SVGA3D_MAKE_HWVERSION(2, 0),
-   SVGA3D_HWVERSION_WS8_B1    = SVGA3D_MAKE_HWVERSION(2, 1),
-   SVGA3D_HWVERSION_CURRENT   = SVGA3D_HWVERSION_WS8_B1,
+   SVGA3D_HWVERSION_CURRENT   = SVGA3D_HWVERSION_WS65_B1,
 } SVGA3dHardwareVersion;
 
 /*
@@ -68,8 +67,7 @@ typedef enum {
 typedef uint32 SVGA3dBool; /* 32-bit Bool definition */
 #define SVGA3D_NUM_CLIPPLANES                   6
 #define SVGA3D_MAX_SIMULTANEOUS_RENDER_TARGETS  8
-#define SVGA3D_MAX_CONTEXT_IDS                  256
-#define SVGA3D_MAX_SURFACE_IDS                  (32 * 1024)
+
 
 /*
  * Surface formats.
@@ -81,91 +79,76 @@ typedef uint32 SVGA3dBool; /* 32-bit Bool definition */
  */
 
 typedef enum SVGA3dSurfaceFormat {
-   SVGA3D_FORMAT_INVALID               = 0,
+   SVGA3D_FORMAT_INVALID = 0,
 
-   SVGA3D_X8R8G8B8                     = 1,
-   SVGA3D_A8R8G8B8                     = 2,
+   SVGA3D_X8R8G8B8       = 1,
+   SVGA3D_A8R8G8B8       = 2,
 
-   SVGA3D_R5G6B5                       = 3,
-   SVGA3D_X1R5G5B5                     = 4,
-   SVGA3D_A1R5G5B5                     = 5,
-   SVGA3D_A4R4G4B4                     = 6,
+   SVGA3D_R5G6B5         = 3,
+   SVGA3D_X1R5G5B5       = 4,
+   SVGA3D_A1R5G5B5       = 5,
+   SVGA3D_A4R4G4B4       = 6,
 
-   SVGA3D_Z_D32                        = 7,
-   SVGA3D_Z_D16                        = 8,
-   SVGA3D_Z_D24S8                      = 9,
-   SVGA3D_Z_D15S1                      = 10,
+   SVGA3D_Z_D32          = 7,
+   SVGA3D_Z_D16          = 8,
+   SVGA3D_Z_D24S8        = 9,
+   SVGA3D_Z_D15S1        = 10,
 
-   SVGA3D_LUMINANCE8                   = 11,
-   SVGA3D_LUMINANCE4_ALPHA4            = 12,
-   SVGA3D_LUMINANCE16                  = 13,
-   SVGA3D_LUMINANCE8_ALPHA8            = 14,
+   SVGA3D_LUMINANCE8            = 11,
+   SVGA3D_LUMINANCE4_ALPHA4     = 12,
+   SVGA3D_LUMINANCE16           = 13,
+   SVGA3D_LUMINANCE8_ALPHA8     = 14,
 
-   SVGA3D_DXT1                         = 15,
-   SVGA3D_DXT2                         = 16,
-   SVGA3D_DXT3                         = 17,
-   SVGA3D_DXT4                         = 18,
-   SVGA3D_DXT5                         = 19,
+   SVGA3D_DXT1           = 15,
+   SVGA3D_DXT2           = 16,
+   SVGA3D_DXT3           = 17,
+   SVGA3D_DXT4           = 18,
+   SVGA3D_DXT5           = 19,
 
-   SVGA3D_BUMPU8V8                     = 20,
-   SVGA3D_BUMPL6V5U5                   = 21,
-   SVGA3D_BUMPX8L8V8U8                 = 22,
-   SVGA3D_BUMPL8V8U8                   = 23,
+   SVGA3D_BUMPU8V8       = 20,
+   SVGA3D_BUMPL6V5U5     = 21,
+   SVGA3D_BUMPX8L8V8U8   = 22,
+   SVGA3D_BUMPL8V8U8     = 23,
 
-   SVGA3D_ARGB_S10E5                   = 24,   /* 16-bit floating-point ARGB */
-   SVGA3D_ARGB_S23E8                   = 25,   /* 32-bit floating-point ARGB */
+   SVGA3D_ARGB_S10E5     = 24,   /* 16-bit floating-point ARGB */
+   SVGA3D_ARGB_S23E8     = 25,   /* 32-bit floating-point ARGB */
 
-   SVGA3D_A2R10G10B10                  = 26,
+   SVGA3D_A2R10G10B10    = 26,
 
    /* signed formats */
-   SVGA3D_V8U8                         = 27,
-   SVGA3D_Q8W8V8U8                     = 28,
-   SVGA3D_CxV8U8                       = 29,
+   SVGA3D_V8U8           = 27,
+   SVGA3D_Q8W8V8U8       = 28,
+   SVGA3D_CxV8U8         = 29,
 
    /* mixed formats */
-   SVGA3D_X8L8V8U8                     = 30,
-   SVGA3D_A2W10V10U10                  = 31,
+   SVGA3D_X8L8V8U8       = 30,
+   SVGA3D_A2W10V10U10    = 31,
 
-   SVGA3D_ALPHA8                       = 32,
+   SVGA3D_ALPHA8         = 32,
 
    /* Single- and dual-component floating point formats */
-   SVGA3D_R_S10E5                      = 33,
-   SVGA3D_R_S23E8                      = 34,
-   SVGA3D_RG_S10E5                     = 35,
-   SVGA3D_RG_S23E8                     = 36,
+   SVGA3D_R_S10E5        = 33,
+   SVGA3D_R_S23E8        = 34,
+   SVGA3D_RG_S10E5       = 35,
+   SVGA3D_RG_S23E8       = 36,
 
    /*
     * Any surface can be used as a buffer object, but SVGA3D_BUFFER is
     * the most efficient format to use when creating new surfaces
     * expressly for index or vertex data.
     */
+   SVGA3D_BUFFER         = 37,
 
-   SVGA3D_BUFFER                       = 37,
-
-   SVGA3D_Z_D24X8                      = 38,
+   SVGA3D_Z_D24X8        = 38,
 
-   SVGA3D_V16U16                       = 39,
+   SVGA3D_V16U16         = 39,
 
-   SVGA3D_G16R16                       = 40,
-   SVGA3D_A16B16G16R16                 = 41,
+   SVGA3D_G16R16         = 40,
+   SVGA3D_A16B16G16R16   = 41,
 
    /* Packed Video formats */
-   SVGA3D_UYVY                         = 42,
-   SVGA3D_YUY2                         = 43,
-
-   /* Planar video formats */
-   SVGA3D_NV12                         = 44,
-
-   /* Video format with alpha */
-   SVGA3D_AYUV                         = 45,
-
-   SVGA3D_BC4_UNORM                    = 108,
-   SVGA3D_BC5_UNORM                    = 111,
-
-   /* Advanced D3D9 depth formats. */
-   SVGA3D_Z_DF16                       = 118,
-   SVGA3D_Z_DF24                       = 119,
-   SVGA3D_Z_D24S8_INT                  = 120,
+   SVGA3D_UYVY           = 42,
+   SVGA3D_YUY2           = 43,
 
    SVGA3D_FORMAT_MAX
 } SVGA3dSurfaceFormat;
@@ -431,20 +414,10 @@ typedef enum {
    SVGA3D_RS_SRCBLENDALPHA             = 94,    /* SVGA3dBlendOp */
    SVGA3D_RS_DSTBLENDALPHA             = 95,    /* SVGA3dBlendOp */
    SVGA3D_RS_BLENDEQUATIONALPHA        = 96,    /* SVGA3dBlendEquation */
-   SVGA3D_RS_TRANSPARENCYANTIALIAS     = 97,    /* SVGA3dTransparencyAntialiasType */
-   SVGA3D_RS_LINEAA                    = 98,    /* SVGA3dBool */
-   SVGA3D_RS_LINEWIDTH                 = 99,    /* float */
    SVGA3D_RS_MAX
 } SVGA3dRenderStateName;
 
 typedef enum {
-   SVGA3D_TRANSPARENCYANTIALIAS_NORMAL            = 0,
-   SVGA3D_TRANSPARENCYANTIALIAS_ALPHATOCOVERAGE   = 1,
-   SVGA3D_TRANSPARENCYANTIALIAS_SUPERSAMPLE       = 2,
-   SVGA3D_TRANSPARENCYANTIALIAS_MAX
-} SVGA3dTransparencyAntialiasType;
-
-typedef enum {
    SVGA3D_VERTEXMATERIAL_NONE     = 0,    /* Use the value in the current material */
    SVGA3D_VERTEXMATERIAL_DIFFUSE  = 1,    /* Use the value in the diffuse component */
    SVGA3D_VERTEXMATERIAL_SPECULAR = 2,    /* Use the value in the specular component */
@@ -755,10 +728,10 @@ typedef enum {
    SVGA3D_TEX_FILTER_NEAREST        = 1,
    SVGA3D_TEX_FILTER_LINEAR         = 2,
    SVGA3D_TEX_FILTER_ANISOTROPIC    = 3,
-   SVGA3D_TEX_FILTER_FLATCUBIC      = 4, /* Deprecated, not implemented */
-   SVGA3D_TEX_FILTER_GAUSSIANCUBIC  = 5, /* Deprecated, not implemented */
-   SVGA3D_TEX_FILTER_PYRAMIDALQUAD  = 6, /* Not currently implemented */
-   SVGA3D_TEX_FILTER_GAUSSIANQUAD   = 7, /* Not currently implemented */
+   SVGA3D_TEX_FILTER_FLATCUBIC      = 4, // Deprecated, not implemented
+   SVGA3D_TEX_FILTER_GAUSSIANCUBIC  = 5, // Deprecated, not implemented
+   SVGA3D_TEX_FILTER_PYRAMIDALQUAD  = 6, // Not currently implemented
+   SVGA3D_TEX_FILTER_GAUSSIANQUAD   = 7, // Not currently implemented
    SVGA3D_TEX_FILTER_MAX
 } SVGA3dTextureFilter;
 
@@ -826,19 +799,19 @@ typedef enum {
 
 typedef enum {
    SVGA3D_DECLUSAGE_POSITION     = 0,
-   SVGA3D_DECLUSAGE_BLENDWEIGHT,       /*  1 */
-   SVGA3D_DECLUSAGE_BLENDINDICES,      /*  2 */
-   SVGA3D_DECLUSAGE_NORMAL,            /*  3 */
-   SVGA3D_DECLUSAGE_PSIZE,             /*  4 */
-   SVGA3D_DECLUSAGE_TEXCOORD,          /*  5 */
-   SVGA3D_DECLUSAGE_TANGENT,           /*  6 */
-   SVGA3D_DECLUSAGE_BINORMAL,          /*  7 */
-   SVGA3D_DECLUSAGE_TESSFACTOR,        /*  8 */
-   SVGA3D_DECLUSAGE_POSITIONT,         /*  9 */
-   SVGA3D_DECLUSAGE_COLOR,             /* 10 */
-   SVGA3D_DECLUSAGE_FOG,               /* 11 */
-   SVGA3D_DECLUSAGE_DEPTH,             /* 12 */
-   SVGA3D_DECLUSAGE_SAMPLE,            /* 13 */
+   SVGA3D_DECLUSAGE_BLENDWEIGHT,       //  1
+   SVGA3D_DECLUSAGE_BLENDINDICES,      //  2
+   SVGA3D_DECLUSAGE_NORMAL,            //  3
+   SVGA3D_DECLUSAGE_PSIZE,             //  4
+   SVGA3D_DECLUSAGE_TEXCOORD,          //  5
+   SVGA3D_DECLUSAGE_TANGENT,           //  6
+   SVGA3D_DECLUSAGE_BINORMAL,          //  7
+   SVGA3D_DECLUSAGE_TESSFACTOR,        //  8
+   SVGA3D_DECLUSAGE_POSITIONT,         //  9
+   SVGA3D_DECLUSAGE_COLOR,             // 10
+   SVGA3D_DECLUSAGE_FOG,               // 11
+   SVGA3D_DECLUSAGE_DEPTH,             // 12
+   SVGA3D_DECLUSAGE_SAMPLE,            // 13
    SVGA3D_DECLUSAGE_MAX
 } SVGA3dDeclUsage;
 
@@ -846,10 +819,10 @@ typedef enum {
    SVGA3D_DECLMETHOD_DEFAULT     = 0,
    SVGA3D_DECLMETHOD_PARTIALU,
    SVGA3D_DECLMETHOD_PARTIALV,
-   SVGA3D_DECLMETHOD_CROSSUV,          /* Normal */
+   SVGA3D_DECLMETHOD_CROSSUV,          // Normal
    SVGA3D_DECLMETHOD_UV,
-   SVGA3D_DECLMETHOD_LOOKUP,           /* Lookup a displacement map */
-   SVGA3D_DECLMETHOD_LOOKUPPRESAMPLED, /* Lookup a pre-sampled displacement map */
+   SVGA3D_DECLMETHOD_LOOKUP,           // Lookup a displacement map
+   SVGA3D_DECLMETHOD_LOOKUPPRESAMPLED, // Lookup a pre-sampled displacement map
 } SVGA3dDeclMethod;
 
 typedef enum {
@@ -957,6 +930,7 @@ typedef enum {
 } SVGA3dCubeFace;
 
 typedef enum {
+   SVGA3D_SHADERTYPE_COMPILED_DX8               = 0,
    SVGA3D_SHADERTYPE_VS                         = 1,
    SVGA3D_SHADERTYPE_PS                         = 2,
    SVGA3D_SHADERTYPE_MAX
@@ -994,18 +968,12 @@ typedef enum {
 } SVGA3dTransferType;
 
 /*
- * The maximum number of vertex arrays we're guaranteed to support in
+ * The maximum number vertex arrays we're guaranteed to support in
  * SVGA_3D_CMD_DRAWPRIMITIVES.
  */
 #define SVGA3D_MAX_VERTEX_ARRAYS   32
 
 /*
- * The maximum number of primitive ranges we're guaranteed to support
- * in SVGA_3D_CMD_DRAWPRIMITIVES.
- */
-#define SVGA3D_MAX_DRAW_PRIMITIVE_RANGES 32
-
-/*
  * Identifiers for commands in the command FIFO.
  *
  * IDs between 1000 and 1039 (inclusive) were used by obsolete versions of
@@ -1022,7 +990,7 @@ typedef enum {
 #define SVGA_3D_CMD_LEGACY_BASE            1000
 #define SVGA_3D_CMD_BASE                   1040
 
-#define SVGA_3D_CMD_SURFACE_DEFINE         SVGA_3D_CMD_BASE + 0     /* Deprecated */
+#define SVGA_3D_CMD_SURFACE_DEFINE         SVGA_3D_CMD_BASE + 0
 #define SVGA_3D_CMD_SURFACE_DESTROY        SVGA_3D_CMD_BASE + 1
 #define SVGA_3D_CMD_SURFACE_COPY           SVGA_3D_CMD_BASE + 2
 #define SVGA_3D_CMD_SURFACE_STRETCHBLT     SVGA_3D_CMD_BASE + 3
@@ -1040,7 +1008,7 @@ typedef enum {
 #define SVGA_3D_CMD_SETVIEWPORT            SVGA_3D_CMD_BASE + 15
 #define SVGA_3D_CMD_SETCLIPPLANE           SVGA_3D_CMD_BASE + 16
 #define SVGA_3D_CMD_CLEAR                  SVGA_3D_CMD_BASE + 17
-#define SVGA_3D_CMD_PRESENT                SVGA_3D_CMD_BASE + 18    /* Deprecated */
+#define SVGA_3D_CMD_PRESENT                SVGA_3D_CMD_BASE + 18    // Deprecated
 #define SVGA_3D_CMD_SHADER_DEFINE          SVGA_3D_CMD_BASE + 19
 #define SVGA_3D_CMD_SHADER_DESTROY         SVGA_3D_CMD_BASE + 20
 #define SVGA_3D_CMD_SET_SHADER             SVGA_3D_CMD_BASE + 21
@@ -1050,13 +1018,9 @@ typedef enum {
 #define SVGA_3D_CMD_BEGIN_QUERY            SVGA_3D_CMD_BASE + 25
 #define SVGA_3D_CMD_END_QUERY              SVGA_3D_CMD_BASE + 26
 #define SVGA_3D_CMD_WAIT_FOR_QUERY         SVGA_3D_CMD_BASE + 27
-#define SVGA_3D_CMD_PRESENT_READBACK       SVGA_3D_CMD_BASE + 28    /* Deprecated */
+#define SVGA_3D_CMD_PRESENT_READBACK       SVGA_3D_CMD_BASE + 28    // Deprecated
 #define SVGA_3D_CMD_BLIT_SURFACE_TO_SCREEN SVGA_3D_CMD_BASE + 29
-#define SVGA_3D_CMD_SURFACE_DEFINE_V2      SVGA_3D_CMD_BASE + 30
-#define SVGA_3D_CMD_GENERATE_MIPMAPS       SVGA_3D_CMD_BASE + 31
-#define SVGA_3D_CMD_ACTIVATE_SURFACE       SVGA_3D_CMD_BASE + 40
-#define SVGA_3D_CMD_DEACTIVATE_SURFACE     SVGA_3D_CMD_BASE + 41
-#define SVGA_3D_CMD_MAX                    SVGA_3D_CMD_BASE + 42
+#define SVGA_3D_CMD_MAX                    SVGA_3D_CMD_BASE + 30
 
 #define SVGA_3D_CMD_FUTURE_MAX             2000
 
@@ -1067,9 +1031,9 @@ typedef enum {
 typedef struct {
    union {
       struct {
-         uint16  function;       /* SVGA3dFogFunction */
-         uint8   type;           /* SVGA3dFogType */
-         uint8   base;           /* SVGA3dFogBase */
+         uint16  function;       // SVGA3dFogFunction
+         uint8   type;           // SVGA3dFogType
+         uint8   base;           // SVGA3dFogBase
       };
       uint32     uintValue;
    };
@@ -1145,8 +1109,6 @@ typedef enum {
    SVGA3D_SURFACE_HINT_RENDERTARGET    = (1 << 6),
    SVGA3D_SURFACE_HINT_DEPTHSTENCIL    = (1 << 7),
    SVGA3D_SURFACE_HINT_WRITEONLY       = (1 << 8),
-   SVGA3D_SURFACE_MASKABLE_ANTIALIAS   = (1 << 9),
-   SVGA3D_SURFACE_AUTOGENMIPMAPS       = (1 << 10),
 } SVGA3dSurfaceFlags;
 
 typedef
@@ -1159,12 +1121,6 @@ struct {
    uint32                      sid;
    SVGA3dSurfaceFlags          surfaceFlags;
    SVGA3dSurfaceFormat         format;
-   /*
-    * If surfaceFlags has SVGA3D_SURFACE_CUBEMAP bit set, all SVGA3dSurfaceFace
-    * structures must have the same value of numMipLevels field.
-    * Otherwise, all but the first SVGA3dSurfaceFace structures must have the
-    * numMipLevels set to 0.
-    */
    SVGA3dSurfaceFace           face[SVGA3D_MAX_SURFACE_FACES];
    /*
     * Followed by an SVGA3dSize structure for each mip level in each face.
@@ -1179,31 +1135,6 @@ struct {
 
 typedef
 struct {
-   uint32                      sid;
-   SVGA3dSurfaceFlags          surfaceFlags;
-   SVGA3dSurfaceFormat         format;
-   /*
-    * If surfaceFlags has SVGA3D_SURFACE_CUBEMAP bit set, all SVGA3dSurfaceFace
-    * structures must have the same value of numMipLevels field.
-    * Otherwise, all but the first SVGA3dSurfaceFace structures must have the
-    * numMipLevels set to 0.
-    */
-   SVGA3dSurfaceFace           face[SVGA3D_MAX_SURFACE_FACES];
-   uint32                      multisampleCount;
-   SVGA3dTextureFilter         autogenFilter;
-   /*
-    * Followed by an SVGA3dSize structure for each mip level in each face.
-    *
-    * A note on surface sizes: Sizes are always specified in pixels,
-    * even if the true surface size is not a multiple of the minimum
-    * block size of the surface's format. For example, a 3x3x1 DXT1
-    * compressed texture would actually be stored as a 4x4x1 image in
-    * memory.
-    */
-} SVGA3dCmdDefineSurface_v2;     /* SVGA_3D_CMD_SURFACE_DEFINE_V2 */
-
-typedef
-struct {
    uint32               sid;
 } SVGA3dCmdDestroySurface;      /* SVGA_3D_CMD_SURFACE_DESTROY */
 
@@ -1543,12 +1474,10 @@ struct {
     * SVGA3dCmdDrawPrimitives structure. In order,
     * they are:
     *
-    * 1. SVGA3dVertexDecl, quantity 'numVertexDecls', but no more than
-    *    SVGA3D_MAX_VERTEX_ARRAYS;
-    * 2. SVGA3dPrimitiveRange, quantity 'numRanges', but no more than
-    *    SVGA3D_MAX_DRAW_PRIMITIVE_RANGES;
+    * 1. SVGA3dVertexDecl, quantity 'numVertexDecls'
+    * 2. SVGA3dPrimitiveRange, quantity 'numRanges'
     * 3. Optionally, SVGA3dVertexDivisor, quantity 'numVertexDecls' (contains
-    *    the frequency divisor for the corresponding vertex decl).
+    *    the frequency divisor for this the corresponding vertex decl)
     */
 } SVGA3dCmdDrawPrimitives;      /* SVGA_3D_CMD_DRAWPRIMITIVES */
 
@@ -1742,12 +1671,6 @@ struct {
    /* Clipping: zero or more SVGASignedRects follow */
 } SVGA3dCmdBlitSurfaceToScreen;         /* SVGA_3D_CMD_BLIT_SURFACE_TO_SCREEN */
 
-typedef
-struct {
-   uint32               sid;
-   SVGA3dTextureFilter  filter;
-} SVGA3dCmdGenerateMipmaps;             /* SVGA_3D_CMD_GENERATE_MIPMAPS */
-
 
 /*
  * Capability query index.
@@ -1851,32 +1774,6 @@ typedef enum {
    SVGA3D_DEVCAP_SURFACEFMT_A16B16G16R16           = 67,
    SVGA3D_DEVCAP_SURFACEFMT_UYVY                   = 68,
    SVGA3D_DEVCAP_SURFACEFMT_YUY2                   = 69,
-   SVGA3D_DEVCAP_MULTISAMPLE_NONMASKABLESAMPLES    = 70,
-   SVGA3D_DEVCAP_MULTISAMPLE_MASKABLESAMPLES       = 71,
-   SVGA3D_DEVCAP_ALPHATOCOVERAGE                   = 72,
-   SVGA3D_DEVCAP_SUPERSAMPLE                       = 73,
-   SVGA3D_DEVCAP_AUTOGENMIPMAPS                    = 74,
-   SVGA3D_DEVCAP_SURFACEFMT_NV12                   = 75,
-   SVGA3D_DEVCAP_SURFACEFMT_AYUV                   = 76,
-
-   /*
-    * This is the maximum number of SVGA context IDs that the guest
-    * can define using SVGA_3D_CMD_CONTEXT_DEFINE.
-    */
-   SVGA3D_DEVCAP_MAX_CONTEXT_IDS                   = 77,
-
-   /*
-    * This is the maximum number of SVGA surface IDs that the guest
-    * can define using SVGA_3D_CMD_SURFACE_DEFINE*.
-    */
-   SVGA3D_DEVCAP_MAX_SURFACE_IDS                   = 78,
-
-   SVGA3D_DEVCAP_SURFACEFMT_Z_DF16                 = 79,
-   SVGA3D_DEVCAP_SURFACEFMT_Z_DF24                 = 80,
-   SVGA3D_DEVCAP_SURFACEFMT_Z_D24S8_INT            = 81,
-
-   SVGA3D_DEVCAP_SURFACEFMT_BC4_UNORM              = 82,
-   SVGA3D_DEVCAP_SURFACEFMT_BC5_UNORM              = 83,
 
    /*
     * Don't add new caps into the previous section; the values in this
diff --git a/drivers/gpu/drm/vmwgfx/svga_escape.h b/drivers/gpu/drm/vmwgfx/svga_escape.h
index 8e8d968..7b85e9b 100644
--- a/drivers/gpu/drm/vmwgfx/svga_escape.h
+++ b/drivers/gpu/drm/vmwgfx/svga_escape.h
@@ -75,7 +75,7 @@
  */
 
 #define SVGA_ESCAPE_VMWARE_HINT               0x00030000
-#define SVGA_ESCAPE_VMWARE_HINT_FULLSCREEN    0x00030001  /* Deprecated */
+#define SVGA_ESCAPE_VMWARE_HINT_FULLSCREEN    0x00030001  // Deprecated
 
 typedef
 struct {
diff --git a/drivers/gpu/drm/vmwgfx/svga_overlay.h b/drivers/gpu/drm/vmwgfx/svga_overlay.h
index f38416f..f753d73 100644
--- a/drivers/gpu/drm/vmwgfx/svga_overlay.h
+++ b/drivers/gpu/drm/vmwgfx/svga_overlay.h
@@ -38,9 +38,9 @@
  * Video formats we support
  */
 
-#define VMWARE_FOURCC_YV12 0x32315659 /* 'Y' 'V' '1' '2' */
-#define VMWARE_FOURCC_YUY2 0x32595559 /* 'Y' 'U' 'Y' '2' */
-#define VMWARE_FOURCC_UYVY 0x59565955 /* 'U' 'Y' 'V' 'Y' */
+#define VMWARE_FOURCC_YV12 0x32315659 // 'Y' 'V' '1' '2'
+#define VMWARE_FOURCC_YUY2 0x32595559 // 'Y' 'U' 'Y' '2'
+#define VMWARE_FOURCC_UYVY 0x59565955 // 'U' 'Y' 'V' 'Y'
 
 typedef enum {
    SVGA_OVERLAY_FORMAT_INVALID = 0,
@@ -68,7 +68,7 @@ struct SVGAEscapeVideoSetRegs {
       uint32 streamId;
    } header;
 
-   /* May include zero or more items. */
+   // May include zero or more items.
    struct {
       uint32 registerId;
       uint32 value;
@@ -134,12 +134,12 @@ struct {
  */
 
 static inline bool
-VMwareVideoGetAttributes(const SVGAOverlayFormat format,    /* IN */
-                         uint32 *width,                     /* IN / OUT */
-                         uint32 *height,                    /* IN / OUT */
-                         uint32 *size,                      /* OUT */
-                         uint32 *pitches,                   /* OUT (optional) */
-                         uint32 *offsets)                   /* OUT (optional) */
+VMwareVideoGetAttributes(const SVGAOverlayFormat format,    // IN
+                         uint32 *width,                     // IN / OUT
+                         uint32 *height,                    // IN / OUT
+                         uint32 *size,                      // OUT
+                         uint32 *pitches,                   // OUT (optional)
+                         uint32 *offsets)                   // OUT (optional)
 {
     int tmp;
 
@@ -198,4 +198,4 @@ VMwareVideoGetAttributes(const SVGAOverlayFormat format,    /* IN */
     return true;
 }
 
-#endif /* _SVGA_OVERLAY_H_ */
+#endif // _SVGA_OVERLAY_H_
diff --git a/drivers/gpu/drm/vmwgfx/svga_reg.h b/drivers/gpu/drm/vmwgfx/svga_reg.h
index 01f63cb..1b96c2e 100644
--- a/drivers/gpu/drm/vmwgfx/svga_reg.h
+++ b/drivers/gpu/drm/vmwgfx/svga_reg.h
@@ -39,15 +39,6 @@
 #define PCI_DEVICE_ID_VMWARE_SVGA2      0x0405
 
 /*
- * SVGA_REG_ENABLE bit definitions.
- */
-#define SVGA_REG_ENABLE_DISABLE     0
-#define SVGA_REG_ENABLE_ENABLE      1
-#define SVGA_REG_ENABLE_HIDE        2
-#define SVGA_REG_ENABLE_ENABLE_HIDE (SVGA_REG_ENABLE_ENABLE |\
-				     SVGA_REG_ENABLE_HIDE)
-
-/*
  * Legal values for the SVGA_REG_CURSOR_ON register in old-fashioned
  * cursor bypass mode. This is still supported, but no new guest
  * drivers should use it.
@@ -167,9 +158,7 @@ enum {
    SVGA_REG_GMR_MAX_DESCRIPTOR_LENGTH = 44,
 
    SVGA_REG_TRACES = 45,            /* Enable trace-based updates even when FIFO is on */
-   SVGA_REG_GMRS_MAX_PAGES = 46,    /* Maximum number of 4KB pages for all GMRs */
-   SVGA_REG_MEMORY_SIZE = 47,       /* Total dedicated device memory excluding FIFO */
-   SVGA_REG_TOP = 48,               /* Must be 1 more than the last register */
+   SVGA_REG_TOP = 46,               /* Must be 1 more than the last register */
 
    SVGA_PALETTE_BASE = 1024,        /* Base of SVGA color map */
    /* Next 768 (== 256*3) registers exist for colormap */
@@ -276,7 +265,7 @@ enum {
  * possible.
  */
 #define SVGA_GMR_NULL         ((uint32) -1)
-#define SVGA_GMR_FRAMEBUFFER  ((uint32) -2)  /* Guest Framebuffer (GFB) */
+#define SVGA_GMR_FRAMEBUFFER  ((uint32) -2)  // Guest Framebuffer (GFB)
 
 typedef
 struct SVGAGuestMemDescriptor {
@@ -317,35 +306,13 @@ struct SVGAGMRImageFormat {
       struct {
          uint32 bitsPerPixel : 8;
          uint32 colorDepth   : 8;
-         uint32 reserved     : 16;  /* Must be zero */
+         uint32 reserved     : 16;  // Must be zero
       };
 
       uint32 value;
    };
 } SVGAGMRImageFormat;
 
-typedef
-struct SVGAGuestImage {
-   SVGAGuestPtr         ptr;
-
-   /*
-    * A note on interpretation of pitch: This value of pitch is the
-    * number of bytes between vertically adjacent image
-    * blocks. Normally this is the number of bytes between the first
-    * pixel of two adjacent scanlines. With compressed textures,
-    * however, this may represent the number of bytes between
-    * compression blocks rather than between rows of pixels.
-    *
-    * XXX: Compressed textures currently must be tightly packed in guest memory.
-    *
-    * If the image is 1-dimensional, pitch is ignored.
-    *
-    * If 'pitch' is zero, the SVGA3D device calculates a pitch value
-    * assuming each row of blocks is tightly packed.
-    */
-   uint32 pitch;
-} SVGAGuestImage;
-
 /*
  * SVGAColorBGRX --
  *
@@ -361,7 +328,7 @@ struct SVGAColorBGRX {
          uint32 b : 8;
          uint32 g : 8;
          uint32 r : 8;
-         uint32 x : 8;  /* Unused */
+         uint32 x : 8;  // Unused
       };
 
       uint32 value;
@@ -403,34 +370,23 @@ struct SVGASignedPoint {
  *  Note the holes in the bitfield. Missing bits have been deprecated,
  *  and must not be reused. Those capabilities will never be reported
  *  by new versions of the SVGA device.
- *
- * SVGA_CAP_GMR2 --
- *    Provides asynchronous commands to define and remap guest memory
- *    regions.  Adds device registers SVGA_REG_GMRS_MAX_PAGES and
- *    SVGA_REG_MEMORY_SIZE.
- *
- * SVGA_CAP_SCREEN_OBJECT_2 --
- *    Allow screen object support, and require backing stores from the
- *    guest for each screen object.
  */
 
 #define SVGA_CAP_NONE               0x00000000
 #define SVGA_CAP_RECT_COPY          0x00000002
 #define SVGA_CAP_CURSOR             0x00000020
-#define SVGA_CAP_CURSOR_BYPASS      0x00000040   /* Legacy (Use Cursor Bypass 3 instead) */
-#define SVGA_CAP_CURSOR_BYPASS_2    0x00000080   /* Legacy (Use Cursor Bypass 3 instead) */
+#define SVGA_CAP_CURSOR_BYPASS      0x00000040   // Legacy (Use Cursor Bypass 3 instead)
+#define SVGA_CAP_CURSOR_BYPASS_2    0x00000080   // Legacy (Use Cursor Bypass 3 instead)
 #define SVGA_CAP_8BIT_EMULATION     0x00000100
 #define SVGA_CAP_ALPHA_CURSOR       0x00000200
 #define SVGA_CAP_3D                 0x00004000
 #define SVGA_CAP_EXTENDED_FIFO      0x00008000
-#define SVGA_CAP_MULTIMON           0x00010000   /* Legacy multi-monitor support */
+#define SVGA_CAP_MULTIMON           0x00010000   // Legacy multi-monitor support
 #define SVGA_CAP_PITCHLOCK          0x00020000
 #define SVGA_CAP_IRQMASK            0x00040000
-#define SVGA_CAP_DISPLAY_TOPOLOGY   0x00080000   /* Legacy multi-monitor support */
+#define SVGA_CAP_DISPLAY_TOPOLOGY   0x00080000   // Legacy multi-monitor support
 #define SVGA_CAP_GMR                0x00100000
 #define SVGA_CAP_TRACES             0x00200000
-#define SVGA_CAP_GMR2               0x00400000
-#define SVGA_CAP_SCREEN_OBJECT_2    0x00800000
 
 
 /*
@@ -475,7 +431,7 @@ enum {
 
    SVGA_FIFO_CAPABILITIES = 4,
    SVGA_FIFO_FLAGS,
-   /* Valid with SVGA_FIFO_CAP_FENCE: */
+   // Valid with SVGA_FIFO_CAP_FENCE:
    SVGA_FIFO_FENCE,
 
    /*
@@ -488,47 +444,33 @@ enum {
     * extended FIFO.
     */
 
-   /* Valid if exists (i.e. if extended FIFO enabled): */
+   // Valid if exists (i.e. if extended FIFO enabled):
    SVGA_FIFO_3D_HWVERSION,       /* See SVGA3dHardwareVersion in svga3d_reg.h */
-   /* Valid with SVGA_FIFO_CAP_PITCHLOCK: */
+   // Valid with SVGA_FIFO_CAP_PITCHLOCK:
    SVGA_FIFO_PITCHLOCK,
 
-   /* Valid with SVGA_FIFO_CAP_CURSOR_BYPASS_3: */
+   // Valid with SVGA_FIFO_CAP_CURSOR_BYPASS_3:
    SVGA_FIFO_CURSOR_ON,          /* Cursor bypass 3 show/hide register */
    SVGA_FIFO_CURSOR_X,           /* Cursor bypass 3 x register */
    SVGA_FIFO_CURSOR_Y,           /* Cursor bypass 3 y register */
    SVGA_FIFO_CURSOR_COUNT,       /* Incremented when any of the other 3 change */
    SVGA_FIFO_CURSOR_LAST_UPDATED,/* Last time the host updated the cursor */
 
-   /* Valid with SVGA_FIFO_CAP_RESERVE: */
+   // Valid with SVGA_FIFO_CAP_RESERVE:
    SVGA_FIFO_RESERVED,           /* Bytes past NEXT_CMD with real contents */
 
    /*
-    * Valid with SVGA_FIFO_CAP_SCREEN_OBJECT or SVGA_FIFO_CAP_SCREEN_OBJECT_2:
+    * Valid with SVGA_FIFO_CAP_SCREEN_OBJECT:
     *
     * By default this is SVGA_ID_INVALID, to indicate that the cursor
     * coordinates are specified relative to the virtual root. If this
     * is set to a specific screen ID, cursor position is reinterpreted
-    * as a signed offset relative to that screen's origin.
+    * as a signed offset relative to that screen's origin. This is the
+    * only way to place the cursor on a non-rooted screen.
     */
    SVGA_FIFO_CURSOR_SCREEN_ID,
 
    /*
-    * Valid with SVGA_FIFO_CAP_DEAD
-    *
-    * An arbitrary value written by the host, drivers should not use it.
-    */
-   SVGA_FIFO_DEAD,
-
-   /*
-    * Valid with SVGA_FIFO_CAP_3D_HWVERSION_REVISED:
-    *
-    * Contains 3D HWVERSION (see SVGA3dHardwareVersion in svga3d_reg.h)
-    * on platforms that can enforce graphics resource limits.
-    */
-   SVGA_FIFO_3D_HWVERSION_REVISED,
-
-   /*
     * XXX: The gap here, up until SVGA_FIFO_3D_CAPS, can be used for new
     * registers, but this must be done carefully and with judicious use of
     * capability bits, since comparisons based on SVGA_FIFO_MIN aren't
@@ -566,7 +508,7 @@ enum {
     * sets SVGA_FIFO_MIN high enough to leave room for them.
     */
 
-   /* Valid if register exists: */
+   // Valid if register exists:
    SVGA_FIFO_GUEST_3D_HWVERSION, /* Guest driver's 3D version */
    SVGA_FIFO_FENCE_GOAL,         /* Matching target for SVGA_IRQFLAG_FENCE_GOAL */
    SVGA_FIFO_BUSY,               /* See "FIFO Synchronization Registers" */
@@ -767,37 +709,6 @@ enum {
  *
  *       - When a screen is resized, either using Screen Object commands or
  *         legacy multimon registers, its contents are preserved.
- *
- * SVGA_FIFO_CAP_GMR2 --
- *
- *    Provides new commands to define and remap guest memory regions (GMR).
- *
- *    New 2D commands:
- *       DEFINE_GMR2, REMAP_GMR2.
- *
- * SVGA_FIFO_CAP_3D_HWVERSION_REVISED --
- *
- *    Indicates new register SVGA_FIFO_3D_HWVERSION_REVISED exists.
- *    This register may replace SVGA_FIFO_3D_HWVERSION on platforms
- *    that enforce graphics resource limits.  This allows the platform
- *    to clear SVGA_FIFO_3D_HWVERSION and disable 3D in legacy guest
- *    drivers that do not limit their resources.
- *
- *    Note this is an alias to SVGA_FIFO_CAP_GMR2 because these indicators
- *    are codependent (and thus we use a single capability bit).
- *
- * SVGA_FIFO_CAP_SCREEN_OBJECT_2 --
- *
- *    Modifies the DEFINE_SCREEN command to include a guest provided
- *    backing store in GMR memory and the bytesPerLine for the backing
- *    store.  This capability requires the use of a backing store when
- *    creating screen objects.  However if SVGA_FIFO_CAP_SCREEN_OBJECT
- *    is present then backing stores are optional.
- *
- * SVGA_FIFO_CAP_DEAD --
- *
- *    Drivers should not use this cap bit.  This cap bit can not be
- *    reused since some hosts already expose it.
  */
 
 #define SVGA_FIFO_CAP_NONE                  0
@@ -809,10 +720,6 @@ enum {
 #define SVGA_FIFO_CAP_ESCAPE            (1<<5)
 #define SVGA_FIFO_CAP_RESERVE           (1<<6)
 #define SVGA_FIFO_CAP_SCREEN_OBJECT     (1<<7)
-#define SVGA_FIFO_CAP_GMR2              (1<<8)
-#define SVGA_FIFO_CAP_3D_HWVERSION_REVISED  SVGA_FIFO_CAP_GMR2
-#define SVGA_FIFO_CAP_SCREEN_OBJECT_2   (1<<9)
-#define SVGA_FIFO_CAP_DEAD              (1<<10)
 
 
 /*
@@ -823,7 +730,7 @@ enum {
 
 #define SVGA_FIFO_FLAG_NONE                 0
 #define SVGA_FIFO_FLAG_ACCELFRONT       (1<<0)
-#define SVGA_FIFO_FLAG_RESERVED        (1<<31) /* Internal use only */
+#define SVGA_FIFO_FLAG_RESERVED        (1<<31) // Internal use only
 
 /*
  * FIFO reservation sentinel value
@@ -856,22 +763,22 @@ enum {
    SVGA_VIDEO_DATA_OFFSET,
    SVGA_VIDEO_FORMAT,
    SVGA_VIDEO_COLORKEY,
-   SVGA_VIDEO_SIZE,          /* Deprecated */
+   SVGA_VIDEO_SIZE,          // Deprecated
    SVGA_VIDEO_WIDTH,
    SVGA_VIDEO_HEIGHT,
    SVGA_VIDEO_SRC_X,
    SVGA_VIDEO_SRC_Y,
    SVGA_VIDEO_SRC_WIDTH,
    SVGA_VIDEO_SRC_HEIGHT,
-   SVGA_VIDEO_DST_X,         /* Signed int32 */
-   SVGA_VIDEO_DST_Y,         /* Signed int32 */
+   SVGA_VIDEO_DST_X,         // Signed int32
+   SVGA_VIDEO_DST_Y,         // Signed int32
    SVGA_VIDEO_DST_WIDTH,
    SVGA_VIDEO_DST_HEIGHT,
    SVGA_VIDEO_PITCH_1,
    SVGA_VIDEO_PITCH_2,
    SVGA_VIDEO_PITCH_3,
-   SVGA_VIDEO_DATA_GMRID,    /* Optional, defaults to SVGA_GMR_FRAMEBUFFER */
-   SVGA_VIDEO_DST_SCREEN_ID, /* Optional, defaults to virtual coords (SVGA_ID_INVALID) */
+   SVGA_VIDEO_DATA_GMRID,    // Optional, defaults to SVGA_GMR_FRAMEBUFFER
+   SVGA_VIDEO_DST_SCREEN_ID, // Optional, defaults to virtual coords (SVGA_ID_INVALID)
    SVGA_VIDEO_NUM_REGS
 };
 
@@ -922,51 +829,15 @@ typedef struct SVGAOverlayUnit {
  *    compatibility. New flags can be added, and the struct may grow,
  *    but existing fields must retain their meaning.
  *
- *    Added with SVGA_FIFO_CAP_SCREEN_OBJECT_2 are required fields of
- *    a SVGAGuestPtr that is used to back the screen contents.  This
- *    memory must come from the GFB.  The guest is not allowed to
- *    access the memory and doing so will have undefined results.  The
- *    backing store is required to be page aligned and the size is
- *    padded to the next page boundry.  The number of pages is:
- *       (bytesPerLine * size.width * 4 + PAGE_SIZE - 1) / PAGE_SIZE
- *
- *    The pitch in the backingStore is required to be at least large
- *    enough to hold a 32bbp scanline.  It is recommended that the
- *    driver pad bytesPerLine for a potential performance win.
- *
- *    The cloneCount field is treated as a hint from the guest that
- *    the user wants this display to be cloned, countCount times.  A
- *    value of zero means no cloning should happen.
- */
-
-#define SVGA_SCREEN_MUST_BE_SET     (1 << 0) /* Must be set or results undefined */
-#define SVGA_SCREEN_HAS_ROOT SVGA_SCREEN_MUST_BE_SET /* Deprecated */
-#define SVGA_SCREEN_IS_PRIMARY      (1 << 1) /* Guest considers this screen to be 'primary' */
-#define SVGA_SCREEN_FULLSCREEN_HINT (1 << 2) /* Guest is running a fullscreen app here */
-
-/*
- * Added with SVGA_FIFO_CAP_SCREEN_OBJECT_2.  When the screen is
- * deactivated the base layer is defined to lose all contents and
- * become black.  When a screen is deactivated the backing store is
- * optional.  When set backingPtr and bytesPerLine will be ignored.
  */
-#define SVGA_SCREEN_DEACTIVATE  (1 << 3)
 
-/*
- * Added with SVGA_FIFO_CAP_SCREEN_OBJECT_2.  When this flag is set
- * the screen contents will be outputted as all black to the user
- * though the base layer contents is preserved.  The screen base layer
- * can still be read and written to like normal though the no visible
- * effect will be seen by the user.  When the flag is changed the
- * screen will be blanked or redrawn to the current contents as needed
- * without any extra commands from the driver.  This flag only has an
- * effect when the screen is not deactivated.
- */
-#define SVGA_SCREEN_BLANKING (1 << 4)
+#define SVGA_SCREEN_HAS_ROOT    (1 << 0)  // Screen is present in the virtual coord space
+#define SVGA_SCREEN_IS_PRIMARY  (1 << 1)  // Guest considers this screen to be 'primary'
+#define SVGA_SCREEN_FULLSCREEN_HINT (1 << 2)   // Guest is running a fullscreen app here
 
 typedef
 struct SVGAScreenObject {
-   uint32 structSize;   /* sizeof(SVGAScreenObject) */
+   uint32 structSize;   // sizeof(SVGAScreenObject)
    uint32 id;
    uint32 flags;
    struct {
@@ -976,14 +847,7 @@ struct SVGAScreenObject {
    struct {
       int32 x;
       int32 y;
-   } root;
-
-   /*
-    * Added and required by SVGA_FIFO_CAP_SCREEN_OBJECT_2, optional
-    * with SVGA_FIFO_CAP_SCREEN_OBJECT.
-    */
-   SVGAGuestImage backingStore;
-   uint32 cloneCount;
+   } root;              // Only used if SVGA_SCREEN_HAS_ROOT is set.
 } SVGAScreenObject;
 
 
@@ -1021,8 +885,6 @@ typedef enum {
    SVGA_CMD_BLIT_SCREEN_TO_GMRFB  = 38,
    SVGA_CMD_ANNOTATION_FILL       = 39,
    SVGA_CMD_ANNOTATION_COPY       = 40,
-   SVGA_CMD_DEFINE_GMR2           = 41,
-   SVGA_CMD_REMAP_GMR2            = 42,
    SVGA_CMD_MAX
 } SVGAFifoCmdId;
 
@@ -1058,7 +920,7 @@ typedef enum {
  */
 
 typedef
-struct SVGAFifoCmdUpdate {
+struct {
    uint32 x;
    uint32 y;
    uint32 width;
@@ -1077,7 +939,7 @@ struct SVGAFifoCmdUpdate {
  */
 
 typedef
-struct SVGAFifoCmdRectCopy {
+struct {
    uint32 srcX;
    uint32 srcY;
    uint32 destX;
@@ -1101,14 +963,14 @@ struct SVGAFifoCmdRectCopy {
  */
 
 typedef
-struct SVGAFifoCmdDefineCursor {
-   uint32 id;             /* Reserved, must be zero. */
+struct {
+   uint32 id;             // Reserved, must be zero.
    uint32 hotspotX;
    uint32 hotspotY;
    uint32 width;
    uint32 height;
-   uint32 andMaskDepth;   /* Value must be 1 or equal to BITS_PER_PIXEL */
-   uint32 xorMaskDepth;   /* Value must be 1 or equal to BITS_PER_PIXEL */
+   uint32 andMaskDepth;   // Value must be 1 or equal to BITS_PER_PIXEL
+   uint32 xorMaskDepth;   // Value must be 1 or equal to BITS_PER_PIXEL
    /*
     * Followed by scanline data for AND mask, then XOR mask.
     * Each scanline is padded to a 32-bit boundary.
@@ -1130,8 +992,8 @@ struct SVGAFifoCmdDefineCursor {
  */
 
 typedef
-struct SVGAFifoCmdDefineAlphaCursor {
-   uint32 id;             /* Reserved, must be zero. */
+struct {
+   uint32 id;             // Reserved, must be zero.
    uint32 hotspotX;
    uint32 hotspotY;
    uint32 width;
@@ -1153,7 +1015,7 @@ struct SVGAFifoCmdDefineAlphaCursor {
  */
 
 typedef
-struct SVGAFifoCmdUpdateVerbose {
+struct {
    uint32 x;
    uint32 y;
    uint32 width;
@@ -1178,13 +1040,13 @@ struct SVGAFifoCmdUpdateVerbose {
 #define  SVGA_ROP_COPY                    0x03
 
 typedef
-struct SVGAFifoCmdFrontRopFill {
-   uint32 color;     /* In the same format as the GFB */
+struct {
+   uint32 color;     // In the same format as the GFB
    uint32 x;
    uint32 y;
    uint32 width;
    uint32 height;
-   uint32 rop;       /* Must be SVGA_ROP_COPY */
+   uint32 rop;       // Must be SVGA_ROP_COPY
 } SVGAFifoCmdFrontRopFill;
 
 
@@ -1221,7 +1083,7 @@ struct {
  */
 
 typedef
-struct SVGAFifoCmdEscape {
+struct {
    uint32 nsid;
    uint32 size;
    /* followed by 'size' bytes of data */
@@ -1251,12 +1113,12 @@ struct SVGAFifoCmdEscape {
  *    registers (SVGA_REG_NUM_GUEST_DISPLAYS, SVGA_REG_DISPLAY_*).
  *
  * Availability:
- *    SVGA_FIFO_CAP_SCREEN_OBJECT or SVGA_FIFO_CAP_SCREEN_OBJECT_2
+ *    SVGA_FIFO_CAP_SCREEN_OBJECT
  */
 
 typedef
 struct {
-   SVGAScreenObject screen;   /* Variable-length according to version */
+   SVGAScreenObject screen;   // Variable-length according to version
 } SVGAFifoCmdDefineScreen;
 
 
@@ -1267,7 +1129,7 @@ struct {
  *    re-use.
  *
  * Availability:
- *    SVGA_FIFO_CAP_SCREEN_OBJECT or SVGA_FIFO_CAP_SCREEN_OBJECT_2
+ *    SVGA_FIFO_CAP_SCREEN_OBJECT
  */
 
 typedef
@@ -1320,7 +1182,7 @@ struct {
  *    GMRFB.
  *
  * Availability:
- *    SVGA_FIFO_CAP_SCREEN_OBJECT or SVGA_FIFO_CAP_SCREEN_OBJECT_2
+ *    SVGA_FIFO_CAP_SCREEN_OBJECT
  */
 
 typedef
@@ -1357,7 +1219,7 @@ struct {
  *    SVGA_CMD_ANNOTATION_* commands for details.
  *
  * Availability:
- *    SVGA_FIFO_CAP_SCREEN_OBJECT or SVGA_FIFO_CAP_SCREEN_OBJECT_2
+ *    SVGA_FIFO_CAP_SCREEN_OBJECT
  */
 
 typedef
@@ -1405,7 +1267,7 @@ struct {
  *    the time any subsequent FENCE commands are reached.
  *
  * Availability:
- *    SVGA_FIFO_CAP_SCREEN_OBJECT or SVGA_FIFO_CAP_SCREEN_OBJECT_2
+ *    SVGA_FIFO_CAP_SCREEN_OBJECT
  */
 
 typedef
@@ -1440,7 +1302,7 @@ struct {
  *    user's display is being remoted over a network connection.
  *
  * Availability:
- *    SVGA_FIFO_CAP_SCREEN_OBJECT or SVGA_FIFO_CAP_SCREEN_OBJECT_2
+ *    SVGA_FIFO_CAP_SCREEN_OBJECT
  */
 
 typedef
@@ -1472,7 +1334,7 @@ struct {
  *    undefined.
  *
  * Availability:
- *    SVGA_FIFO_CAP_SCREEN_OBJECT or SVGA_FIFO_CAP_SCREEN_OBJECT_2
+ *    SVGA_FIFO_CAP_SCREEN_OBJECT
  */
 
 typedef
@@ -1481,72 +1343,4 @@ struct {
    uint32           srcScreenId;
 } SVGAFifoCmdAnnotationCopy;
 
-
-/*
- * SVGA_CMD_DEFINE_GMR2 --
- *
- *    Define guest memory region v2.  See the description of GMRs above.
- *
- * Availability:
- *    SVGA_CAP_GMR2
- */
-
-typedef
-struct {
-   uint32 gmrId;
-   uint32 numPages;
-} SVGAFifoCmdDefineGMR2;
-
-
-/*
- * SVGA_CMD_REMAP_GMR2 --
- *
- *    Remap guest memory region v2.  See the description of GMRs above.
- *
- *    This command allows guest to modify a portion of an existing GMR by
- *    invalidating it or reassigning it to different guest physical pages.
- *    The pages are identified by physical page number (PPN).  The pages
- *    are assumed to be pinned and valid for DMA operations.
- *
- *    Description of command flags:
- *
- *    SVGA_REMAP_GMR2_VIA_GMR: If enabled, references a PPN list in a GMR.
- *       The PPN list must not overlap with the remap region (this can be
- *       handled trivially by referencing a separate GMR).  If flag is
- *       disabled, PPN list is appended to SVGARemapGMR command.
- *
- *    SVGA_REMAP_GMR2_PPN64: If set, PPN list is in PPN64 format, otherwise
- *       it is in PPN32 format.
- *
- *    SVGA_REMAP_GMR2_SINGLE_PPN: If set, PPN list contains a single entry.
- *       A single PPN can be used to invalidate a portion of a GMR or
- *       map it to to a single guest scratch page.
- *
- * Availability:
- *    SVGA_CAP_GMR2
- */
-
-typedef enum {
-   SVGA_REMAP_GMR2_PPN32         = 0,
-   SVGA_REMAP_GMR2_VIA_GMR       = (1 << 0),
-   SVGA_REMAP_GMR2_PPN64         = (1 << 1),
-   SVGA_REMAP_GMR2_SINGLE_PPN    = (1 << 2),
-} SVGARemapGMR2Flags;
-
-typedef
-struct {
-   uint32 gmrId;
-   SVGARemapGMR2Flags flags;
-   uint32 offsetPages; /* offset in pages to begin remap */
-   uint32 numPages; /* number of pages to remap */
-   /*
-    * Followed by additional data depending on SVGARemapGMR2Flags.
-    *
-    * If flag SVGA_REMAP_GMR2_VIA_GMR is set, single SVGAGuestPtr follows.
-    * Otherwise an array of page descriptors in PPN32 or PPN64 format
-    * (according to flag SVGA_REMAP_GMR2_PPN64) follows.  If flag
-    * SVGA_REMAP_GMR2_SINGLE_PPN is set, array contains a single entry.
-    */
-} SVGAFifoCmdRemapGMR2;
-
 #endif
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c b/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c
index 1e2c0fb..87e43e0 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_buffer.c
@@ -28,7 +28,6 @@
 #include "vmwgfx_drv.h"
 #include "ttm/ttm_bo_driver.h"
 #include "ttm/ttm_placement.h"
-#include "ttm/ttm_page_alloc.h"
 
 static uint32_t vram_placement_flags = TTM_PL_FLAG_VRAM |
 	TTM_PL_FLAG_CACHED;
@@ -43,10 +42,6 @@ static uint32_t sys_placement_flags = TTM_PL_FLAG_SYSTEM |
 static uint32_t gmr_placement_flags = VMW_PL_FLAG_GMR |
 	TTM_PL_FLAG_CACHED;
 
-static uint32_t gmr_ne_placement_flags = VMW_PL_FLAG_GMR |
-	TTM_PL_FLAG_CACHED |
-	TTM_PL_FLAG_NO_EVICT;
-
 struct ttm_placement vmw_vram_placement = {
 	.fpfn = 0,
 	.lpfn = 0,
@@ -61,11 +56,6 @@ static uint32_t vram_gmr_placement_flags[] = {
 	VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED
 };
 
-static uint32_t gmr_vram_placement_flags[] = {
-	VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED,
-	TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED
-};
-
 struct ttm_placement vmw_vram_gmr_placement = {
 	.fpfn = 0,
 	.lpfn = 0,
@@ -75,20 +65,6 @@ struct ttm_placement vmw_vram_gmr_placement = {
 	.busy_placement = &gmr_placement_flags
 };
 
-static uint32_t vram_gmr_ne_placement_flags[] = {
-	TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT,
-	VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED | TTM_PL_FLAG_NO_EVICT
-};
-
-struct ttm_placement vmw_vram_gmr_ne_placement = {
-	.fpfn = 0,
-	.lpfn = 0,
-	.num_placement = 2,
-	.placement = vram_gmr_ne_placement_flags,
-	.num_busy_placement = 1,
-	.busy_placement = &gmr_ne_placement_flags
-};
-
 struct ttm_placement vmw_vram_sys_placement = {
 	.fpfn = 0,
 	.lpfn = 0,
@@ -116,87 +92,85 @@ struct ttm_placement vmw_sys_placement = {
 	.busy_placement = &sys_placement_flags
 };
 
-static uint32_t evictable_placement_flags[] = {
-	TTM_PL_FLAG_SYSTEM | TTM_PL_FLAG_CACHED,
-	TTM_PL_FLAG_VRAM | TTM_PL_FLAG_CACHED,
-	VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED
+struct vmw_ttm_backend {
+	struct ttm_backend backend;
+	struct page **pages;
+	unsigned long num_pages;
+	struct vmw_private *dev_priv;
+	int gmr_id;
 };
 
-struct ttm_placement vmw_evictable_placement = {
-	.fpfn = 0,
-	.lpfn = 0,
-	.num_placement = 3,
-	.placement = evictable_placement_flags,
-	.num_busy_placement = 1,
-	.busy_placement = &sys_placement_flags
-};
+static int vmw_ttm_populate(struct ttm_backend *backend,
+			    unsigned long num_pages, struct page **pages,
+			    struct page *dummy_read_page,
+			    dma_addr_t *dma_addrs)
+{
+	struct vmw_ttm_backend *vmw_be =
+	    container_of(backend, struct vmw_ttm_backend, backend);
 
-struct ttm_placement vmw_srf_placement = {
-	.fpfn = 0,
-	.lpfn = 0,
-	.num_placement = 1,
-	.num_busy_placement = 2,
-	.placement = &gmr_placement_flags,
-	.busy_placement = gmr_vram_placement_flags
-};
+	vmw_be->pages = pages;
+	vmw_be->num_pages = num_pages;
 
-struct vmw_ttm_tt {
-	struct ttm_tt ttm;
-	struct vmw_private *dev_priv;
-	int gmr_id;
-};
+	return 0;
+}
 
-static int vmw_ttm_bind(struct ttm_tt *ttm, struct ttm_mem_reg *bo_mem)
+static int vmw_ttm_bind(struct ttm_backend *backend, struct ttm_mem_reg *bo_mem)
 {
-	struct vmw_ttm_tt *vmw_be = container_of(ttm, struct vmw_ttm_tt, ttm);
+	struct vmw_ttm_backend *vmw_be =
+	    container_of(backend, struct vmw_ttm_backend, backend);
 
 	vmw_be->gmr_id = bo_mem->start;
 
-	return vmw_gmr_bind(vmw_be->dev_priv, ttm->pages,
-			    ttm->num_pages, vmw_be->gmr_id);
+	return vmw_gmr_bind(vmw_be->dev_priv, vmw_be->pages,
+			    vmw_be->num_pages, vmw_be->gmr_id);
 }
 
-static int vmw_ttm_unbind(struct ttm_tt *ttm)
+static int vmw_ttm_unbind(struct ttm_backend *backend)
 {
-	struct vmw_ttm_tt *vmw_be = container_of(ttm, struct vmw_ttm_tt, ttm);
+	struct vmw_ttm_backend *vmw_be =
+	    container_of(backend, struct vmw_ttm_backend, backend);
 
 	vmw_gmr_unbind(vmw_be->dev_priv, vmw_be->gmr_id);
 	return 0;
 }
 
-static void vmw_ttm_destroy(struct ttm_tt *ttm)
+static void vmw_ttm_clear(struct ttm_backend *backend)
+{
+	struct vmw_ttm_backend *vmw_be =
+		container_of(backend, struct vmw_ttm_backend, backend);
+
+	vmw_be->pages = NULL;
+	vmw_be->num_pages = 0;
+}
+
+static void vmw_ttm_destroy(struct ttm_backend *backend)
 {
-	struct vmw_ttm_tt *vmw_be = container_of(ttm, struct vmw_ttm_tt, ttm);
+	struct vmw_ttm_backend *vmw_be =
+	    container_of(backend, struct vmw_ttm_backend, backend);
 
-	ttm_tt_fini(ttm);
 	kfree(vmw_be);
 }
 
 static struct ttm_backend_func vmw_ttm_func = {
+	.populate = vmw_ttm_populate,
+	.clear = vmw_ttm_clear,
 	.bind = vmw_ttm_bind,
 	.unbind = vmw_ttm_unbind,
 	.destroy = vmw_ttm_destroy,
 };
 
-struct ttm_tt *vmw_ttm_tt_create(struct ttm_bo_device *bdev,
-				 unsigned long size, uint32_t page_flags,
-				 struct page *dummy_read_page)
+struct ttm_backend *vmw_ttm_backend_init(struct ttm_bo_device *bdev)
 {
-	struct vmw_ttm_tt *vmw_be;
+	struct vmw_ttm_backend *vmw_be;
 
 	vmw_be = kmalloc(sizeof(*vmw_be), GFP_KERNEL);
 	if (!vmw_be)
 		return NULL;
 
-	vmw_be->ttm.func = &vmw_ttm_func;
+	vmw_be->backend.func = &vmw_ttm_func;
 	vmw_be->dev_priv = container_of(bdev, struct vmw_private, bdev);
 
-	if (ttm_tt_init(&vmw_be->ttm, bdev, size, page_flags, dummy_read_page)) {
-		kfree(vmw_be);
-		return NULL;
-	}
-
-	return &vmw_be->ttm;
+	return &vmw_be->backend;
 }
 
 int vmw_invalidate_caches(struct ttm_bo_device *bdev, uint32_t flags)
@@ -300,45 +274,43 @@ static int vmw_ttm_fault_reserve_notify(struct ttm_buffer_object *bo)
 
 static void *vmw_sync_obj_ref(void *sync_obj)
 {
-
-	return (void *)
-		vmw_fence_obj_reference((struct vmw_fence_obj *) sync_obj);
+	return sync_obj;
 }
 
 static void vmw_sync_obj_unref(void **sync_obj)
 {
-	vmw_fence_obj_unreference((struct vmw_fence_obj **) sync_obj);
+	*sync_obj = NULL;
 }
 
 static int vmw_sync_obj_flush(void *sync_obj, void *sync_arg)
 {
-	vmw_fence_obj_flush((struct vmw_fence_obj *) sync_obj);
+	struct vmw_private *dev_priv = (struct vmw_private *)sync_arg;
+
+	mutex_lock(&dev_priv->hw_mutex);
+	vmw_write(dev_priv, SVGA_REG_SYNC, SVGA_SYNC_GENERIC);
+	mutex_unlock(&dev_priv->hw_mutex);
 	return 0;
 }
 
 static bool vmw_sync_obj_signaled(void *sync_obj, void *sync_arg)
 {
-	unsigned long flags = (unsigned long) sync_arg;
-	return	vmw_fence_obj_signaled((struct vmw_fence_obj *) sync_obj,
-				       (uint32_t) flags);
+	struct vmw_private *dev_priv = (struct vmw_private *)sync_arg;
+	uint32_t sequence = (unsigned long) sync_obj;
 
+	return vmw_fence_signaled(dev_priv, sequence);
 }
 
 static int vmw_sync_obj_wait(void *sync_obj, void *sync_arg,
 			     bool lazy, bool interruptible)
 {
-	unsigned long flags = (unsigned long) sync_arg;
+	struct vmw_private *dev_priv = (struct vmw_private *)sync_arg;
+	uint32_t sequence = (unsigned long) sync_obj;
 
-	return vmw_fence_obj_wait((struct vmw_fence_obj *) sync_obj,
-				  (uint32_t) flags,
-				  lazy, interruptible,
-				  VMW_FENCE_WAIT_TIMEOUT);
+	return vmw_wait_fence(dev_priv, false, sequence, false, 3*HZ);
 }
 
 struct ttm_bo_driver vmw_bo_driver = {
-	.ttm_tt_create = &vmw_ttm_tt_create,
-	.ttm_tt_populate = &ttm_pool_populate,
-	.ttm_tt_unpopulate = &ttm_pool_unpopulate,
+	.create_ttm_backend_entry = vmw_ttm_backend_init,
 	.invalidate_caches = vmw_invalidate_caches,
 	.init_mem_type = vmw_init_mem_type,
 	.evict_flags = vmw_evict_flags,
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c
deleted file mode 100644
index 3fa884d..0000000
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_dmabuf.c
+++ /dev/null
@@ -1,322 +0,0 @@
-/**************************************************************************
- *
- * Copyright © 2011 VMware, Inc., Palo Alto, CA., USA
- * All Rights Reserved.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the
- * "Software"), to deal in the Software without restriction, including
- * without limitation the rights to use, copy, modify, merge, publish,
- * distribute, sub license, and/or sell copies of the Software, and to
- * permit persons to whom the Software is furnished to do so, subject to
- * the following conditions:
- *
- * The above copyright notice and this permission notice (including the
- * next paragraph) shall be included in all copies or substantial portions
- * of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
- * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
- * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
- * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
- * USE OR OTHER DEALINGS IN THE SOFTWARE.
- *
- **************************************************************************/
-
-#include "ttm/ttm_placement.h"
-
-#include "drmP.h"
-#include "vmwgfx_drv.h"
-
-
-/**
- * vmw_dmabuf_to_placement - Validate a buffer to placement.
- *
- * @dev_priv:  Driver private.
- * @buf:  DMA buffer to move.
- * @pin:  Pin buffer if true.
- * @interruptible:  Use interruptible wait.
- *
- * May only be called by the current master since it assumes that the
- * master lock is the current master's lock.
- * This function takes the master's lock in write mode.
- * Flushes and unpins the query bo to avoid failures.
- *
- * Returns
- *  -ERESTARTSYS if interrupted by a signal.
- */
-int vmw_dmabuf_to_placement(struct vmw_private *dev_priv,
-			    struct vmw_dma_buffer *buf,
-			    struct ttm_placement *placement,
-			    bool interruptible)
-{
-	struct vmw_master *vmaster = dev_priv->active_master;
-	struct ttm_buffer_object *bo = &buf->base;
-	int ret;
-
-	ret = ttm_write_lock(&vmaster->lock, interruptible);
-	if (unlikely(ret != 0))
-		return ret;
-
-	vmw_execbuf_release_pinned_bo(dev_priv, false, 0);
-
-	ret = ttm_bo_reserve(bo, interruptible, false, false, 0);
-	if (unlikely(ret != 0))
-		goto err;
-
-	ret = ttm_bo_validate(bo, placement, interruptible, false, false);
-
-	ttm_bo_unreserve(bo);
-
-err:
-	ttm_write_unlock(&vmaster->lock);
-	return ret;
-}
-
-/**
- * vmw_dmabuf_to_vram_or_gmr - Move a buffer to vram or gmr.
- *
- * May only be called by the current master since it assumes that the
- * master lock is the current master's lock.
- * This function takes the master's lock in write mode.
- * Flushes and unpins the query bo if @pin == true to avoid failures.
- *
- * @dev_priv:  Driver private.
- * @buf:  DMA buffer to move.
- * @pin:  Pin buffer if true.
- * @interruptible:  Use interruptible wait.
- *
- * Returns
- * -ERESTARTSYS if interrupted by a signal.
- */
-int vmw_dmabuf_to_vram_or_gmr(struct vmw_private *dev_priv,
-			      struct vmw_dma_buffer *buf,
-			      bool pin, bool interruptible)
-{
-	struct vmw_master *vmaster = dev_priv->active_master;
-	struct ttm_buffer_object *bo = &buf->base;
-	struct ttm_placement *placement;
-	int ret;
-
-	ret = ttm_write_lock(&vmaster->lock, interruptible);
-	if (unlikely(ret != 0))
-		return ret;
-
-	if (pin)
-		vmw_execbuf_release_pinned_bo(dev_priv, false, 0);
-
-	ret = ttm_bo_reserve(bo, interruptible, false, false, 0);
-	if (unlikely(ret != 0))
-		goto err;
-
-	/**
-	 * Put BO in VRAM if there is space, otherwise as a GMR.
-	 * If there is no space in VRAM and GMR ids are all used up,
-	 * start evicting GMRs to make room. If the DMA buffer can't be
-	 * used as a GMR, this will return -ENOMEM.
-	 */
-
-	if (pin)
-		placement = &vmw_vram_gmr_ne_placement;
-	else
-		placement = &vmw_vram_gmr_placement;
-
-	ret = ttm_bo_validate(bo, placement, interruptible, false, false);
-	if (likely(ret == 0) || ret == -ERESTARTSYS)
-		goto err_unreserve;
-
-
-	/**
-	 * If that failed, try VRAM again, this time evicting
-	 * previous contents.
-	 */
-
-	if (pin)
-		placement = &vmw_vram_ne_placement;
-	else
-		placement = &vmw_vram_placement;
-
-	ret = ttm_bo_validate(bo, placement, interruptible, false, false);
-
-err_unreserve:
-	ttm_bo_unreserve(bo);
-err:
-	ttm_write_unlock(&vmaster->lock);
-	return ret;
-}
-
-/**
- * vmw_dmabuf_to_vram - Move a buffer to vram.
- *
- * May only be called by the current master since it assumes that the
- * master lock is the current master's lock.
- * This function takes the master's lock in write mode.
- *
- * @dev_priv:  Driver private.
- * @buf:  DMA buffer to move.
- * @pin:  Pin buffer in vram if true.
- * @interruptible:  Use interruptible wait.
- *
- * Returns
- * -ERESTARTSYS if interrupted by a signal.
- */
-int vmw_dmabuf_to_vram(struct vmw_private *dev_priv,
-		       struct vmw_dma_buffer *buf,
-		       bool pin, bool interruptible)
-{
-	struct ttm_placement *placement;
-
-	if (pin)
-		placement = &vmw_vram_ne_placement;
-	else
-		placement = &vmw_vram_placement;
-
-	return vmw_dmabuf_to_placement(dev_priv, buf,
-				       placement,
-				       interruptible);
-}
-
-/**
- * vmw_dmabuf_to_start_of_vram - Move a buffer to start of vram.
- *
- * May only be called by the current master since it assumes that the
- * master lock is the current master's lock.
- * This function takes the master's lock in write mode.
- * Flushes and unpins the query bo if @pin == true to avoid failures.
- *
- * @dev_priv:  Driver private.
- * @buf:  DMA buffer to move.
- * @pin:  Pin buffer in vram if true.
- * @interruptible:  Use interruptible wait.
- *
- * Returns
- * -ERESTARTSYS if interrupted by a signal.
- */
-int vmw_dmabuf_to_start_of_vram(struct vmw_private *dev_priv,
-				struct vmw_dma_buffer *buf,
-				bool pin, bool interruptible)
-{
-	struct vmw_master *vmaster = dev_priv->active_master;
-	struct ttm_buffer_object *bo = &buf->base;
-	struct ttm_placement placement;
-	int ret = 0;
-
-	if (pin)
-		placement = vmw_vram_ne_placement;
-	else
-		placement = vmw_vram_placement;
-	placement.lpfn = bo->num_pages;
-
-	ret = ttm_write_lock(&vmaster->lock, interruptible);
-	if (unlikely(ret != 0))
-		return ret;
-
-	if (pin)
-		vmw_execbuf_release_pinned_bo(dev_priv, false, 0);
-
-	ret = ttm_bo_reserve(bo, interruptible, false, false, 0);
-	if (unlikely(ret != 0))
-		goto err_unlock;
-
-	/* Is this buffer already in vram but not at the start of it? */
-	if (bo->mem.mem_type == TTM_PL_VRAM &&
-	    bo->mem.start < bo->num_pages &&
-	    bo->mem.start > 0)
-		(void) ttm_bo_validate(bo, &vmw_sys_placement, false,
-				       false, false);
-
-	ret = ttm_bo_validate(bo, &placement, interruptible, false, false);
-
-	/* For some reason we didn't up at the start of vram */
-	WARN_ON(ret == 0 && bo->offset != 0);
-
-	ttm_bo_unreserve(bo);
-err_unlock:
-	ttm_write_unlock(&vmaster->lock);
-
-	return ret;
-}
-
-
-/**
- * vmw_dmabuf_upin - Unpin the buffer given buffer, does not move the buffer.
- *
- * May only be called by the current master since it assumes that the
- * master lock is the current master's lock.
- * This function takes the master's lock in write mode.
- *
- * @dev_priv:  Driver private.
- * @buf:  DMA buffer to unpin.
- * @interruptible:  Use interruptible wait.
- *
- * Returns
- * -ERESTARTSYS if interrupted by a signal.
- */
-int vmw_dmabuf_unpin(struct vmw_private *dev_priv,
-		     struct vmw_dma_buffer *buf,
-		     bool interruptible)
-{
-	/*
-	 * We could in theory early out if the buffer is
-	 * unpinned but we need to lock and reserve the buffer
-	 * anyways so we don't gain much by that.
-	 */
-	return vmw_dmabuf_to_placement(dev_priv, buf,
-				       &vmw_evictable_placement,
-				       interruptible);
-}
-
-
-/**
- * vmw_bo_get_guest_ptr - Get the guest ptr representing the current placement
- * of a buffer.
- *
- * @bo: Pointer to a struct ttm_buffer_object. Must be pinned or reserved.
- * @ptr: SVGAGuestPtr returning the result.
- */
-void vmw_bo_get_guest_ptr(const struct ttm_buffer_object *bo,
-			  SVGAGuestPtr *ptr)
-{
-	if (bo->mem.mem_type == TTM_PL_VRAM) {
-		ptr->gmrId = SVGA_GMR_FRAMEBUFFER;
-		ptr->offset = bo->offset;
-	} else {
-		ptr->gmrId = bo->mem.start;
-		ptr->offset = 0;
-	}
-}
-
-
-/**
- * vmw_bo_pin - Pin or unpin a buffer object without moving it.
- *
- * @bo: The buffer object. Must be reserved, and present either in VRAM
- * or GMR memory.
- * @pin: Whether to pin or unpin.
- *
- */
-void vmw_bo_pin(struct ttm_buffer_object *bo, bool pin)
-{
-	uint32_t pl_flags;
-	struct ttm_placement placement;
-	uint32_t old_mem_type = bo->mem.mem_type;
-	int ret;
-
-	BUG_ON(!atomic_read(&bo->reserved));
-	BUG_ON(old_mem_type != TTM_PL_VRAM &&
-	       old_mem_type != VMW_PL_FLAG_GMR);
-
-	pl_flags = TTM_PL_FLAG_VRAM | VMW_PL_FLAG_GMR | TTM_PL_FLAG_CACHED;
-	if (pin)
-		pl_flags |= TTM_PL_FLAG_NO_EVICT;
-
-	memset(&placement, 0, sizeof(placement));
-	placement.num_placement = 1;
-	placement.placement = &pl_flags;
-
-	ret = ttm_bo_validate(bo, &placement, false, true, true);
-
-	BUG_ON(ret != 0 || bo->mem.mem_type != old_mem_type);
-}
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
index eb78beb..55d4b29 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
@@ -24,7 +24,6 @@
  * USE OR OTHER DEALINGS IN THE SOFTWARE.
  *
  **************************************************************************/
-#include <linux/module.h>
 
 #include "drmP.h"
 #include "vmwgfx_drv.h"
@@ -38,10 +37,6 @@
 #define VMWGFX_CHIP_SVGAII 0
 #define VMW_FB_RESERVATION 0
 
-#define VMW_MIN_INITIAL_WIDTH 800
-#define VMW_MIN_INITIAL_HEIGHT 600
-
-
 /**
  * Fully encoded drm commands. Might move to vmw_drm.h
  */
@@ -87,31 +82,17 @@
 #define DRM_IOCTL_VMW_EXECBUF					\
 	DRM_IOW(DRM_COMMAND_BASE + DRM_VMW_EXECBUF,		\
 		struct drm_vmw_execbuf_arg)
-#define DRM_IOCTL_VMW_GET_3D_CAP				\
-	DRM_IOW(DRM_COMMAND_BASE + DRM_VMW_GET_3D_CAP,		\
-		 struct drm_vmw_get_3d_cap_arg)
+#define DRM_IOCTL_VMW_FIFO_DEBUG				\
+	DRM_IOWR(DRM_COMMAND_BASE + DRM_VMW_FIFO_DEBUG,		\
+		 struct drm_vmw_fifo_debug_arg)
 #define DRM_IOCTL_VMW_FENCE_WAIT				\
 	DRM_IOWR(DRM_COMMAND_BASE + DRM_VMW_FENCE_WAIT,		\
 		 struct drm_vmw_fence_wait_arg)
-#define DRM_IOCTL_VMW_FENCE_SIGNALED				\
-	DRM_IOWR(DRM_COMMAND_BASE + DRM_VMW_FENCE_SIGNALED,	\
-		 struct drm_vmw_fence_signaled_arg)
-#define DRM_IOCTL_VMW_FENCE_UNREF				\
-	DRM_IOW(DRM_COMMAND_BASE + DRM_VMW_FENCE_UNREF,		\
-		 struct drm_vmw_fence_arg)
-#define DRM_IOCTL_VMW_FENCE_EVENT				\
-	DRM_IOW(DRM_COMMAND_BASE + DRM_VMW_FENCE_EVENT,		\
-		 struct drm_vmw_fence_event_arg)
-#define DRM_IOCTL_VMW_PRESENT					\
-	DRM_IOW(DRM_COMMAND_BASE + DRM_VMW_PRESENT,		\
-		 struct drm_vmw_present_arg)
-#define DRM_IOCTL_VMW_PRESENT_READBACK				\
-	DRM_IOW(DRM_COMMAND_BASE + DRM_VMW_PRESENT_READBACK,	\
-		 struct drm_vmw_present_readback_arg)
 #define DRM_IOCTL_VMW_UPDATE_LAYOUT				\
-	DRM_IOW(DRM_COMMAND_BASE + DRM_VMW_UPDATE_LAYOUT,	\
+	DRM_IOWR(DRM_COMMAND_BASE + DRM_VMW_UPDATE_LAYOUT,	\
 		 struct drm_vmw_update_layout_arg)
 
+
 /**
  * The core DRM version of this macro doesn't account for
  * DRM_COMMAND_BASE.
@@ -154,28 +135,12 @@ static struct drm_ioctl_desc vmw_ioctls[] = {
 		      DRM_AUTH | DRM_UNLOCKED),
 	VMW_IOCTL_DEF(VMW_EXECBUF, vmw_execbuf_ioctl,
 		      DRM_AUTH | DRM_UNLOCKED),
-	VMW_IOCTL_DEF(VMW_FENCE_WAIT, vmw_fence_obj_wait_ioctl,
-		      DRM_AUTH | DRM_UNLOCKED),
-	VMW_IOCTL_DEF(VMW_FENCE_SIGNALED,
-		      vmw_fence_obj_signaled_ioctl,
-		      DRM_AUTH | DRM_UNLOCKED),
-	VMW_IOCTL_DEF(VMW_FENCE_UNREF, vmw_fence_obj_unref_ioctl,
-		      DRM_AUTH | DRM_UNLOCKED),
-	VMW_IOCTL_DEF(VMW_FENCE_EVENT,
-		      vmw_fence_event_ioctl,
-		      DRM_AUTH | DRM_UNLOCKED),
-	VMW_IOCTL_DEF(VMW_GET_3D_CAP, vmw_get_cap_3d_ioctl,
+	VMW_IOCTL_DEF(VMW_FIFO_DEBUG, vmw_fifo_debug_ioctl,
+		      DRM_AUTH | DRM_ROOT_ONLY | DRM_MASTER | DRM_UNLOCKED),
+	VMW_IOCTL_DEF(VMW_FENCE_WAIT, vmw_fence_wait_ioctl,
 		      DRM_AUTH | DRM_UNLOCKED),
-
-	/* these allow direct access to the framebuffers mark as master only */
-	VMW_IOCTL_DEF(VMW_PRESENT, vmw_present_ioctl,
-		      DRM_MASTER | DRM_AUTH | DRM_UNLOCKED),
-	VMW_IOCTL_DEF(VMW_PRESENT_READBACK,
-		      vmw_present_readback_ioctl,
-		      DRM_MASTER | DRM_AUTH | DRM_UNLOCKED),
-	VMW_IOCTL_DEF(VMW_UPDATE_LAYOUT,
-		      vmw_kms_update_layout_ioctl,
-		      DRM_MASTER | DRM_UNLOCKED),
+	VMW_IOCTL_DEF(VMW_UPDATE_LAYOUT, vmw_kms_update_layout_ioctl,
+		      DRM_MASTER | DRM_CONTROL_ALLOW | DRM_UNLOCKED)
 };
 
 static struct pci_device_id vmw_pci_id_list[] = {
@@ -225,78 +190,8 @@ static void vmw_print_capabilities(uint32_t capabilities)
 		DRM_INFO("  GMR.\n");
 	if (capabilities & SVGA_CAP_TRACES)
 		DRM_INFO("  Traces.\n");
-	if (capabilities & SVGA_CAP_GMR2)
-		DRM_INFO("  GMR2.\n");
-	if (capabilities & SVGA_CAP_SCREEN_OBJECT_2)
-		DRM_INFO("  Screen Object 2.\n");
-}
-
-
-/**
- * vmw_execbuf_prepare_dummy_query - Initialize a query result structure at
- * the start of a buffer object.
- *
- * @dev_priv: The device private structure.
- *
- * This function will idle the buffer using an uninterruptible wait, then
- * map the first page and initialize a pending occlusion query result structure,
- * Finally it will unmap the buffer.
- *
- * TODO: Since we're only mapping a single page, we should optimize the map
- * to use kmap_atomic / iomap_atomic.
- */
-static void vmw_dummy_query_bo_prepare(struct vmw_private *dev_priv)
-{
-	struct ttm_bo_kmap_obj map;
-	volatile SVGA3dQueryResult *result;
-	bool dummy;
-	int ret;
-	struct ttm_bo_device *bdev = &dev_priv->bdev;
-	struct ttm_buffer_object *bo = dev_priv->dummy_query_bo;
-
-	ttm_bo_reserve(bo, false, false, false, 0);
-	spin_lock(&bdev->fence_lock);
-	ret = ttm_bo_wait(bo, false, false, false);
-	spin_unlock(&bdev->fence_lock);
-	if (unlikely(ret != 0))
-		(void) vmw_fallback_wait(dev_priv, false, true, 0, false,
-					 10*HZ);
-
-	ret = ttm_bo_kmap(bo, 0, 1, &map);
-	if (likely(ret == 0)) {
-		result = ttm_kmap_obj_virtual(&map, &dummy);
-		result->totalSize = sizeof(*result);
-		result->state = SVGA3D_QUERYSTATE_PENDING;
-		result->result32 = 0xff;
-		ttm_bo_kunmap(&map);
-	} else
-		DRM_ERROR("Dummy query buffer map failed.\n");
-	ttm_bo_unreserve(bo);
-}
-
-
-/**
- * vmw_dummy_query_bo_create - create a bo to hold a dummy query result
- *
- * @dev_priv: A device private structure.
- *
- * This function creates a small buffer object that holds the query
- * result for dummy queries emitted as query barriers.
- * No interruptible waits are done within this function.
- *
- * Returns an error if bo creation fails.
- */
-static int vmw_dummy_query_bo_create(struct vmw_private *dev_priv)
-{
-	return ttm_bo_create(&dev_priv->bdev,
-			     PAGE_SIZE,
-			     ttm_bo_type_device,
-			     &vmw_vram_sys_placement,
-			     0, 0, false, NULL,
-			     &dev_priv->dummy_query_bo);
 }
 
-
 static int vmw_request_device(struct vmw_private *dev_priv)
 {
 	int ret;
@@ -306,42 +201,16 @@ static int vmw_request_device(struct vmw_private *dev_priv)
 		DRM_ERROR("Unable to initialize FIFO.\n");
 		return ret;
 	}
-	vmw_fence_fifo_up(dev_priv->fman);
-	ret = vmw_dummy_query_bo_create(dev_priv);
-	if (unlikely(ret != 0))
-		goto out_no_query_bo;
-	vmw_dummy_query_bo_prepare(dev_priv);
 
 	return 0;
-
-out_no_query_bo:
-	vmw_fence_fifo_down(dev_priv->fman);
-	vmw_fifo_release(dev_priv, &dev_priv->fifo);
-	return ret;
 }
 
 static void vmw_release_device(struct vmw_private *dev_priv)
 {
-	/*
-	 * Previous destructions should've released
-	 * the pinned bo.
-	 */
-
-	BUG_ON(dev_priv->pinned_bo != NULL);
-
-	ttm_bo_unref(&dev_priv->dummy_query_bo);
-	vmw_fence_fifo_down(dev_priv->fman);
 	vmw_fifo_release(dev_priv, &dev_priv->fifo);
 }
 
-/**
- * Increase the 3d resource refcount.
- * If the count was prevously zero, initialize the fifo, switching to svga
- * mode. Note that the master holds a ref as well, and may request an
- * explicit switch to svga mode if fb is not running, using @unhide_svga.
- */
-int vmw_3d_resource_inc(struct vmw_private *dev_priv,
-			bool unhide_svga)
+int vmw_3d_resource_inc(struct vmw_private *dev_priv)
 {
 	int ret = 0;
 
@@ -350,83 +219,25 @@ int vmw_3d_resource_inc(struct vmw_private *dev_priv,
 		ret = vmw_request_device(dev_priv);
 		if (unlikely(ret != 0))
 			--dev_priv->num_3d_resources;
-	} else if (unhide_svga) {
-		mutex_lock(&dev_priv->hw_mutex);
-		vmw_write(dev_priv, SVGA_REG_ENABLE,
-			  vmw_read(dev_priv, SVGA_REG_ENABLE) &
-			  ~SVGA_REG_ENABLE_HIDE);
-		mutex_unlock(&dev_priv->hw_mutex);
 	}
-
 	mutex_unlock(&dev_priv->release_mutex);
 	return ret;
 }
 
-/**
- * Decrease the 3d resource refcount.
- * If the count reaches zero, disable the fifo, switching to vga mode.
- * Note that the master holds a refcount as well, and may request an
- * explicit switch to vga mode when it releases its refcount to account
- * for the situation of an X server vt switch to VGA with 3d resources
- * active.
- */
-void vmw_3d_resource_dec(struct vmw_private *dev_priv,
-			 bool hide_svga)
+
+void vmw_3d_resource_dec(struct vmw_private *dev_priv)
 {
 	int32_t n3d;
 
 	mutex_lock(&dev_priv->release_mutex);
 	if (unlikely(--dev_priv->num_3d_resources == 0))
 		vmw_release_device(dev_priv);
-	else if (hide_svga) {
-		mutex_lock(&dev_priv->hw_mutex);
-		vmw_write(dev_priv, SVGA_REG_ENABLE,
-			  vmw_read(dev_priv, SVGA_REG_ENABLE) |
-			  SVGA_REG_ENABLE_HIDE);
-		mutex_unlock(&dev_priv->hw_mutex);
-	}
-
 	n3d = (int32_t) dev_priv->num_3d_resources;
 	mutex_unlock(&dev_priv->release_mutex);
 
 	BUG_ON(n3d < 0);
 }
 
-/**
- * Sets the initial_[width|height] fields on the given vmw_private.
- *
- * It does so by reading SVGA_REG_[WIDTH|HEIGHT] regs and then
- * clamping the value to fb_max_[width|height] fields and the
- * VMW_MIN_INITIAL_[WIDTH|HEIGHT].
- * If the values appear to be invalid, set them to
- * VMW_MIN_INITIAL_[WIDTH|HEIGHT].
- */
-static void vmw_get_initial_size(struct vmw_private *dev_priv)
-{
-	uint32_t width;
-	uint32_t height;
-
-	width = vmw_read(dev_priv, SVGA_REG_WIDTH);
-	height = vmw_read(dev_priv, SVGA_REG_HEIGHT);
-
-	width = max_t(uint32_t, width, VMW_MIN_INITIAL_WIDTH);
-	height = max_t(uint32_t, height, VMW_MIN_INITIAL_HEIGHT);
-
-	if (width > dev_priv->fb_max_width ||
-	    height > dev_priv->fb_max_height) {
-
-		/*
-		 * This is a host error and shouldn't occur.
-		 */
-
-		width = VMW_MIN_INITIAL_WIDTH;
-		height = VMW_MIN_INITIAL_HEIGHT;
-	}
-
-	dev_priv->initial_width = width;
-	dev_priv->initial_height = height;
-}
-
 static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
 {
 	struct vmw_private *dev_priv;
@@ -440,11 +251,9 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
 	}
 	memset(dev_priv, 0, sizeof(*dev_priv));
 
-	pci_set_master(dev->pdev);
-
 	dev_priv->dev = dev;
 	dev_priv->vmw_chipset = chipset;
-	dev_priv->last_read_seqno = (uint32_t) -100;
+	dev_priv->last_read_sequence = (uint32_t) -100;
 	mutex_init(&dev_priv->hw_mutex);
 	mutex_init(&dev_priv->cmdbuf_mutex);
 	mutex_init(&dev_priv->release_mutex);
@@ -455,10 +264,8 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
 	mutex_init(&dev_priv->init_mutex);
 	init_waitqueue_head(&dev_priv->fence_queue);
 	init_waitqueue_head(&dev_priv->fifo_queue);
-	dev_priv->fence_queue_waiters = 0;
+	atomic_set(&dev_priv->fence_queue_waiters, 0);
 	atomic_set(&dev_priv->fifo_queue_waiters, 0);
-	INIT_LIST_HEAD(&dev_priv->surface_lru);
-	dev_priv->used_memory_size = 0;
 
 	dev_priv->io_start = pci_resource_start(dev->pdev, 0);
 	dev_priv->vram_start = pci_resource_start(dev->pdev, 1);
@@ -479,13 +286,6 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
 
 	dev_priv->capabilities = vmw_read(dev_priv, SVGA_REG_CAPABILITIES);
 
-	dev_priv->vram_size = vmw_read(dev_priv, SVGA_REG_VRAM_SIZE);
-	dev_priv->mmio_size = vmw_read(dev_priv, SVGA_REG_MEM_SIZE);
-	dev_priv->fb_max_width = vmw_read(dev_priv, SVGA_REG_MAX_WIDTH);
-	dev_priv->fb_max_height = vmw_read(dev_priv, SVGA_REG_MAX_HEIGHT);
-
-	vmw_get_initial_size(dev_priv);
-
 	if (dev_priv->capabilities & SVGA_CAP_GMR) {
 		dev_priv->max_gmr_descriptors =
 			vmw_read(dev_priv,
@@ -493,19 +293,11 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
 		dev_priv->max_gmr_ids =
 			vmw_read(dev_priv, SVGA_REG_GMR_MAX_IDS);
 	}
-	if (dev_priv->capabilities & SVGA_CAP_GMR2) {
-		dev_priv->max_gmr_pages =
-			vmw_read(dev_priv, SVGA_REG_GMRS_MAX_PAGES);
-		dev_priv->memory_size =
-			vmw_read(dev_priv, SVGA_REG_MEMORY_SIZE);
-		dev_priv->memory_size -= dev_priv->vram_size;
-	} else {
-		/*
-		 * An arbitrary limit of 512MiB on surface
-		 * memory. But all HWV8 hardware supports GMR2.
-		 */
-		dev_priv->memory_size = 512*1024*1024;
-	}
+
+	dev_priv->vram_size = vmw_read(dev_priv, SVGA_REG_VRAM_SIZE);
+	dev_priv->mmio_size = vmw_read(dev_priv, SVGA_REG_MEM_SIZE);
+	dev_priv->fb_max_width = vmw_read(dev_priv, SVGA_REG_MAX_WIDTH);
+	dev_priv->fb_max_height = vmw_read(dev_priv, SVGA_REG_MAX_HEIGHT);
 
 	mutex_unlock(&dev_priv->hw_mutex);
 
@@ -517,12 +309,6 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
 		DRM_INFO("Max GMR descriptors is %u\n",
 			 (unsigned)dev_priv->max_gmr_descriptors);
 	}
-	if (dev_priv->capabilities & SVGA_CAP_GMR2) {
-		DRM_INFO("Max number of GMR pages is %u\n",
-			 (unsigned)dev_priv->max_gmr_pages);
-		DRM_INFO("Max dedicated hypervisor surface memory is %u kiB\n",
-			 (unsigned)dev_priv->memory_size / 1024);
-	}
 	DRM_INFO("VRAM at 0x%08x size is %u kiB\n",
 		 dev_priv->vram_start, dev_priv->vram_size / 1024);
 	DRM_INFO("MMIO at 0x%08x size is %u kiB\n",
@@ -609,34 +395,22 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
 			goto out_no_device;
 		}
 	}
-
-	dev_priv->fman = vmw_fence_manager_init(dev_priv);
-	if (unlikely(dev_priv->fman == NULL))
-		goto out_no_fman;
-
-	/* Need to start the fifo to check if we can do screen objects */
-	ret = vmw_3d_resource_inc(dev_priv, true);
-	if (unlikely(ret != 0))
-		goto out_no_fifo;
-	vmw_kms_save_vga(dev_priv);
-
-	/* Start kms and overlay systems, needs fifo. */
 	ret = vmw_kms_init(dev_priv);
 	if (unlikely(ret != 0))
 		goto out_no_kms;
 	vmw_overlay_init(dev_priv);
-
-	/* 3D Depends on Screen Objects being used. */
-	DRM_INFO("Detected %sdevice 3D availability.\n",
-		 vmw_fifo_have_3d(dev_priv) ?
-		 "" : "no ");
-
-	/* We might be done with the fifo now */
 	if (dev_priv->enable_fb) {
+		ret = vmw_3d_resource_inc(dev_priv);
+		if (unlikely(ret != 0))
+			goto out_no_fifo;
+		vmw_kms_save_vga(dev_priv);
 		vmw_fb_init(dev_priv);
+		DRM_INFO("%s", vmw_fifo_have_3d(dev_priv) ?
+			 "Detected device 3D availability.\n" :
+			 "Detected no device 3D availability.\n");
 	} else {
-		vmw_kms_restore_vga(dev_priv);
-		vmw_3d_resource_dec(dev_priv, true);
+		DRM_INFO("Delayed 3D detection since we're not "
+			 "running the device in SVGA mode yet.\n");
 	}
 
 	if (dev_priv->capabilities & SVGA_CAP_IRQMASK) {
@@ -653,19 +427,15 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
 	return 0;
 
 out_no_irq:
-	if (dev_priv->enable_fb)
-		vmw_fb_close(dev_priv);
-	vmw_overlay_close(dev_priv);
-	vmw_kms_close(dev_priv);
-out_no_kms:
-	/* We still have a 3D resource reference held */
 	if (dev_priv->enable_fb) {
+		vmw_fb_close(dev_priv);
 		vmw_kms_restore_vga(dev_priv);
-		vmw_3d_resource_dec(dev_priv, false);
+		vmw_3d_resource_dec(dev_priv);
 	}
 out_no_fifo:
-	vmw_fence_manager_takedown(dev_priv->fman);
-out_no_fman:
+	vmw_overlay_close(dev_priv);
+	vmw_kms_close(dev_priv);
+out_no_kms:
 	if (dev_priv->stealth)
 		pci_release_region(dev->pdev, 2);
 	else
@@ -698,18 +468,15 @@ static int vmw_driver_unload(struct drm_device *dev)
 
 	unregister_pm_notifier(&dev_priv->pm_nb);
 
-	if (dev_priv->ctx.cmd_bounce)
-		vfree(dev_priv->ctx.cmd_bounce);
 	if (dev_priv->capabilities & SVGA_CAP_IRQMASK)
 		drm_irq_uninstall(dev_priv->dev);
 	if (dev_priv->enable_fb) {
 		vmw_fb_close(dev_priv);
 		vmw_kms_restore_vga(dev_priv);
-		vmw_3d_resource_dec(dev_priv, false);
+		vmw_3d_resource_dec(dev_priv);
 	}
 	vmw_kms_close(dev_priv);
 	vmw_overlay_close(dev_priv);
-	vmw_fence_manager_takedown(dev_priv->fman);
 	if (dev_priv->stealth)
 		pci_release_region(dev->pdev, 2);
 	else
@@ -733,15 +500,6 @@ static int vmw_driver_unload(struct drm_device *dev)
 	return 0;
 }
 
-static void vmw_preclose(struct drm_device *dev,
-			 struct drm_file *file_priv)
-{
-	struct vmw_fpriv *vmw_fp = vmw_fpriv(file_priv);
-	struct vmw_private *dev_priv = vmw_priv(dev);
-
-	vmw_event_fence_fpriv_gone(dev_priv->fman, &vmw_fp->fence_events);
-}
-
 static void vmw_postclose(struct drm_device *dev,
 			 struct drm_file *file_priv)
 {
@@ -764,7 +522,6 @@ static int vmw_driver_open(struct drm_device *dev, struct drm_file *file_priv)
 	if (unlikely(vmw_fp == NULL))
 		return ret;
 
-	INIT_LIST_HEAD(&vmw_fp->fence_events);
 	vmw_fp->tfile = ttm_object_file_init(dev_priv->tdev, 10);
 	if (unlikely(vmw_fp->tfile == NULL))
 		goto out_no_tfile;
@@ -890,7 +647,7 @@ static int vmw_master_set(struct drm_device *dev,
 	int ret = 0;
 
 	if (!dev_priv->enable_fb) {
-		ret = vmw_3d_resource_inc(dev_priv, true);
+		ret = vmw_3d_resource_inc(dev_priv);
 		if (unlikely(ret != 0))
 			return ret;
 		vmw_kms_save_vga(dev_priv);
@@ -932,7 +689,7 @@ out_no_active_lock:
 		vmw_write(dev_priv, SVGA_REG_TRACES, 1);
 		mutex_unlock(&dev_priv->hw_mutex);
 		vmw_kms_restore_vga(dev_priv);
-		vmw_3d_resource_dec(dev_priv, true);
+		vmw_3d_resource_dec(dev_priv);
 	}
 	return ret;
 }
@@ -953,7 +710,7 @@ static void vmw_master_drop(struct drm_device *dev,
 
 	vmw_fp->locked_master = drm_master_get(file_priv->master);
 	ret = ttm_vt_lock(&vmaster->lock, false, vmw_fp->tfile);
-	vmw_execbuf_release_pinned_bo(dev_priv, false, 0);
+	vmw_kms_idle_workqueues(vmaster);
 
 	if (unlikely((ret != 0))) {
 		DRM_ERROR("Unable to lock TTM at VT switch.\n");
@@ -970,7 +727,7 @@ static void vmw_master_drop(struct drm_device *dev,
 		vmw_write(dev_priv, SVGA_REG_TRACES, 1);
 		mutex_unlock(&dev_priv->hw_mutex);
 		vmw_kms_restore_vga(dev_priv);
-		vmw_3d_resource_dec(dev_priv, true);
+		vmw_3d_resource_dec(dev_priv);
 	}
 
 	dev_priv->active_master = &dev_priv->fbdev_master;
@@ -1005,7 +762,6 @@ static int vmwgfx_pm_notifier(struct notifier_block *nb, unsigned long val,
 		 * This empties VRAM and unbinds all GMR bindings.
 		 * Buffer contents is moved to swappable memory.
 		 */
-		vmw_execbuf_release_pinned_bo(dev_priv, false, 0);
 		ttm_bo_swapout_all(&dev_priv->bdev);
 
 		break;
@@ -1080,7 +836,7 @@ static int vmw_pm_prepare(struct device *kdev)
 	 */
 	dev_priv->suspended = true;
 	if (dev_priv->enable_fb)
-			vmw_3d_resource_dec(dev_priv, true);
+		vmw_3d_resource_dec(dev_priv);
 
 	if (dev_priv->num_3d_resources != 0) {
 
@@ -1088,7 +844,7 @@ static int vmw_pm_prepare(struct device *kdev)
 			 "while 3D resources are active.\n");
 
 		if (dev_priv->enable_fb)
-			vmw_3d_resource_inc(dev_priv, true);
+			vmw_3d_resource_inc(dev_priv);
 		dev_priv->suspended = false;
 		return -EBUSY;
 	}
@@ -1112,7 +868,7 @@ static void vmw_pm_complete(struct device *kdev)
 	 * start fifo.
 	 */
 	if (dev_priv->enable_fb)
-			vmw_3d_resource_inc(dev_priv, false);
+		vmw_3d_resource_inc(dev_priv);
 
 	dev_priv->suspended = false;
 }
@@ -1124,21 +880,6 @@ static const struct dev_pm_ops vmw_pm_ops = {
 	.resume = vmw_pm_resume,
 };
 
-static const struct file_operations vmwgfx_driver_fops = {
-	.owner = THIS_MODULE,
-	.open = drm_open,
-	.release = drm_release,
-	.unlocked_ioctl = vmw_unlocked_ioctl,
-	.mmap = vmw_mmap,
-	.poll = vmw_fops_poll,
-	.read = vmw_fops_read,
-	.fasync = drm_fasync,
-#if defined(CONFIG_COMPAT)
-	.compat_ioctl = drm_compat_ioctl,
-#endif
-	.llseek = noop_llseek,
-};
-
 static struct drm_driver driver = {
 	.driver_features = DRIVER_HAVE_IRQ | DRIVER_IRQ_SHARED |
 	DRIVER_MODESET,
@@ -1151,8 +892,6 @@ static struct drm_driver driver = {
 	.irq_uninstall = vmw_irq_uninstall,
 	.irq_handler = vmw_irq_handler,
 	.get_vblank_counter = vmw_get_vblank_counter,
-	.enable_vblank = vmw_enable_vblank,
-	.disable_vblank = vmw_disable_vblank,
 	.reclaim_buffers_locked = NULL,
 	.ioctls = vmw_ioctls,
 	.num_ioctls = DRM_ARRAY_SIZE(vmw_ioctls),
@@ -1162,9 +901,20 @@ static struct drm_driver driver = {
 	.master_set = vmw_master_set,
 	.master_drop = vmw_master_drop,
 	.open = vmw_driver_open,
-	.preclose = vmw_preclose,
 	.postclose = vmw_postclose,
-	.fops = &vmwgfx_driver_fops,
+	.fops = {
+		 .owner = THIS_MODULE,
+		 .open = drm_open,
+		 .release = drm_release,
+		 .unlocked_ioctl = vmw_unlocked_ioctl,
+		 .mmap = vmw_mmap,
+		 .poll = drm_poll,
+		 .fasync = drm_fasync,
+#if defined(CONFIG_COMPAT)
+		 .compat_ioctl = drm_compat_ioctl,
+#endif
+		 .llseek = noop_llseek,
+	},
 	.name = VMWGFX_DRIVER_NAME,
 	.desc = VMWGFX_DRIVER_DESC,
 	.date = VMWGFX_DRIVER_DATE,
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
index d0f2c07..10fc01f 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h
@@ -38,31 +38,23 @@
 #include "ttm/ttm_lock.h"
 #include "ttm/ttm_execbuf_util.h"
 #include "ttm/ttm_module.h"
-#include "vmwgfx_fence.h"
 
-#define VMWGFX_DRIVER_DATE "20120209"
-#define VMWGFX_DRIVER_MAJOR 2
+#define VMWGFX_DRIVER_DATE "20100927"
+#define VMWGFX_DRIVER_MAJOR 1
 #define VMWGFX_DRIVER_MINOR 4
 #define VMWGFX_DRIVER_PATCHLEVEL 0
 #define VMWGFX_FILE_PAGE_OFFSET 0x00100000
 #define VMWGFX_FIFO_STATIC_SIZE (1024*1024)
 #define VMWGFX_MAX_RELOCATIONS 2048
-#define VMWGFX_MAX_VALIDATIONS 2048
+#define VMWGFX_MAX_GMRS 2048
 #define VMWGFX_MAX_DISPLAYS 16
-#define VMWGFX_CMD_BOUNCE_INIT_SIZE 32768
 
 #define VMW_PL_GMR TTM_PL_PRIV0
 #define VMW_PL_FLAG_GMR TTM_PL_FLAG_PRIV0
 
-#define VMW_RES_CONTEXT ttm_driver_type0
-#define VMW_RES_SURFACE ttm_driver_type1
-#define VMW_RES_STREAM ttm_driver_type2
-#define VMW_RES_FENCE ttm_driver_type3
-
 struct vmw_fpriv {
 	struct drm_master *locked_master;
 	struct ttm_object_file *tfile;
-	struct list_head fence_events;
 };
 
 struct vmw_dma_buffer {
@@ -80,11 +72,9 @@ struct vmw_resource {
 	int id;
 	enum ttm_object_type res_type;
 	bool avail;
-	void (*remove_from_lists) (struct vmw_resource *res);
 	void (*hw_destroy) (struct vmw_resource *res);
 	void (*res_free) (struct vmw_resource *res);
-	struct list_head validate_head;
-	struct list_head query_head; /* Protected by the cmdbuf mutex */
+
 	/* TODO is a generic snooper needed? */
 #if 0
 	void (*snoop)(struct vmw_resource *res,
@@ -100,12 +90,8 @@ struct vmw_cursor_snooper {
 	uint32_t *image;
 };
 
-struct vmw_framebuffer;
-struct vmw_surface_offset;
-
 struct vmw_surface {
 	struct vmw_resource res;
-	struct list_head lru_head; /* Protected by the resource lock */
 	uint32_t flags;
 	uint32_t format;
 	uint32_t mip_levels[DRM_VMW_MAX_SURFACE_FACES];
@@ -116,12 +102,9 @@ struct vmw_surface {
 
 	/* TODO so far just a extra pointer */
 	struct vmw_cursor_snooper snooper;
-	struct ttm_buffer_object *backup;
-	struct vmw_surface_offset *offsets;
-	uint32_t backup_size;
 };
 
-struct vmw_marker_queue {
+struct vmw_fence_queue {
 	struct list_head head;
 	struct timespec lag;
 	struct timespec lag_time;
@@ -132,12 +115,16 @@ struct vmw_fifo_state {
 	unsigned long reserved_size;
 	__le32 *dynamic_buffer;
 	__le32 *static_buffer;
+	__le32 *last_buffer;
+	uint32_t last_data_size;
+	uint32_t last_buffer_size;
+	bool last_buffer_add;
 	unsigned long static_buffer_size;
 	bool using_bounce_buffer;
 	uint32_t capabilities;
 	struct mutex fifo_mutex;
 	struct rw_semaphore rwsem;
-	struct vmw_marker_queue marker_queue;
+	struct vmw_fence_queue fence_queue;
 };
 
 struct vmw_relocation {
@@ -149,8 +136,6 @@ struct vmw_sw_context{
 	struct ida bo_list;
 	uint32_t last_cid;
 	bool cid_valid;
-	bool kernel; /**< is the called made from the kernel */
-	struct vmw_resource *cur_ctx;
 	uint32_t last_sid;
 	uint32_t sid_translation;
 	bool sid_valid;
@@ -158,16 +143,8 @@ struct vmw_sw_context{
 	struct list_head validate_nodes;
 	struct vmw_relocation relocs[VMWGFX_MAX_RELOCATIONS];
 	uint32_t cur_reloc;
-	struct ttm_validate_buffer val_bufs[VMWGFX_MAX_VALIDATIONS];
+	struct ttm_validate_buffer val_bufs[VMWGFX_MAX_GMRS];
 	uint32_t cur_val_buf;
-	uint32_t *cmd_bounce;
-	uint32_t cmd_bounce_size;
-	struct list_head resource_list;
-	uint32_t fence_flags;
-	struct list_head query_list;
-	struct ttm_buffer_object *cur_query_bo;
-	uint32_t cur_query_cid;
-	bool query_cid_valid;
 };
 
 struct vmw_legacy_display;
@@ -203,15 +180,11 @@ struct vmw_private {
 	uint32_t mmio_size;
 	uint32_t fb_max_width;
 	uint32_t fb_max_height;
-	uint32_t initial_width;
-	uint32_t initial_height;
 	__le32 __iomem *mmio_virt;
 	int mmio_mtrr;
 	uint32_t capabilities;
 	uint32_t max_gmr_descriptors;
 	uint32_t max_gmr_ids;
-	uint32_t max_gmr_pages;
-	uint32_t memory_size;
 	bool has_gmr;
 	struct mutex hw_mutex;
 
@@ -222,7 +195,12 @@ struct vmw_private {
 	struct vmw_vga_topology_state vga_save[VMWGFX_MAX_DISPLAYS];
 	uint32_t vga_width;
 	uint32_t vga_height;
+	uint32_t vga_depth;
 	uint32_t vga_bpp;
+	uint32_t vga_pseudo;
+	uint32_t vga_red_mask;
+	uint32_t vga_green_mask;
+	uint32_t vga_blue_mask;
 	uint32_t vga_bpl;
 	uint32_t vga_pitchlock;
 
@@ -234,7 +212,6 @@ struct vmw_private {
 
 	void *fb_info;
 	struct vmw_legacy_display *ldu_priv;
-	struct vmw_screen_object_display *sou_priv;
 	struct vmw_overlay *overlay_priv;
 
 	/*
@@ -263,16 +240,13 @@ struct vmw_private {
 	 * Fencing and IRQs.
 	 */
 
-	atomic_t marker_seq;
+	atomic_t fence_seq;
 	wait_queue_head_t fence_queue;
 	wait_queue_head_t fifo_queue;
-	int fence_queue_waiters; /* Protected by hw_mutex */
-	int goal_queue_waiters; /* Protected by hw_mutex */
+	atomic_t fence_queue_waiters;
 	atomic_t fifo_queue_waiters;
-	uint32_t last_read_seqno;
+	uint32_t last_read_sequence;
 	spinlock_t irq_lock;
-	struct vmw_fence_manager *fman;
-	uint32_t irq_mask;
 
 	/*
 	 * Device state
@@ -311,26 +285,6 @@ struct vmw_private {
 
 	struct mutex release_mutex;
 	uint32_t num_3d_resources;
-
-	/*
-	 * Query processing. These members
-	 * are protected by the cmdbuf mutex.
-	 */
-
-	struct ttm_buffer_object *dummy_query_bo;
-	struct ttm_buffer_object *pinned_bo;
-	uint32_t query_cid;
-	bool dummy_query_bo_pinned;
-
-	/*
-	 * Surface swapping. The "surface_lru" list is protected by the
-	 * resource lock in order to be able to destroy a surface and take
-	 * it off the lru atomically. "used_memory_size" is currently
-	 * protected by the cmdbuf mutex for simplicity.
-	 */
-
-	struct list_head surface_lru;
-	uint32_t used_memory_size;
 };
 
 static inline struct vmw_private *vmw_priv(struct drm_device *dev)
@@ -365,8 +319,8 @@ static inline uint32_t vmw_read(struct vmw_private *dev_priv,
 	return val;
 }
 
-int vmw_3d_resource_inc(struct vmw_private *dev_priv, bool unhide_svga);
-void vmw_3d_resource_dec(struct vmw_private *dev_priv, bool hide_svga);
+int vmw_3d_resource_inc(struct vmw_private *dev_priv);
+void vmw_3d_resource_dec(struct vmw_private *dev_priv);
 
 /**
  * GMR utilities - vmwgfx_gmr.c
@@ -391,13 +345,7 @@ extern int vmw_context_define_ioctl(struct drm_device *dev, void *data,
 				    struct drm_file *file_priv);
 extern int vmw_context_check(struct vmw_private *dev_priv,
 			     struct ttm_object_file *tfile,
-			     int id,
-			     struct vmw_resource **p_res);
-extern int vmw_user_lookup_handle(struct vmw_private *dev_priv,
-				  struct ttm_object_file *tfile,
-				  uint32_t handle,
-				  struct vmw_surface **out_surf,
-				  struct vmw_dma_buffer **out_buf);
+			     int id);
 extern void vmw_surface_res_free(struct vmw_resource *res);
 extern int vmw_surface_init(struct vmw_private *dev_priv,
 			    struct vmw_surface *srf,
@@ -415,8 +363,6 @@ extern int vmw_surface_reference_ioctl(struct drm_device *dev, void *data,
 extern int vmw_surface_check(struct vmw_private *dev_priv,
 			     struct ttm_object_file *tfile,
 			     uint32_t handle, int *id);
-extern int vmw_surface_validate(struct vmw_private *dev_priv,
-				struct vmw_surface *srf);
 extern void vmw_dmabuf_bo_free(struct ttm_buffer_object *bo);
 extern int vmw_dmabuf_init(struct vmw_private *dev_priv,
 			   struct vmw_dma_buffer *vmw_bo,
@@ -432,6 +378,10 @@ extern uint32_t vmw_dmabuf_validate_node(struct ttm_buffer_object *bo,
 extern void vmw_dmabuf_validate_clear(struct ttm_buffer_object *bo);
 extern int vmw_user_dmabuf_lookup(struct ttm_object_file *tfile,
 				  uint32_t id, struct vmw_dma_buffer **out);
+extern int vmw_dmabuf_to_start_of_vram(struct vmw_private *vmw_priv,
+				       struct vmw_dma_buffer *bo);
+extern int vmw_dmabuf_from_vram(struct vmw_private *vmw_priv,
+				struct vmw_dma_buffer *bo);
 extern int vmw_stream_claim_ioctl(struct drm_device *dev, void *data,
 				  struct drm_file *file_priv);
 extern int vmw_stream_unref_ioctl(struct drm_device *dev, void *data,
@@ -440,30 +390,7 @@ extern int vmw_user_stream_lookup(struct vmw_private *dev_priv,
 				  struct ttm_object_file *tfile,
 				  uint32_t *inout_id,
 				  struct vmw_resource **out);
-extern void vmw_resource_unreserve(struct list_head *list);
 
-/**
- * DMA buffer helper routines - vmwgfx_dmabuf.c
- */
-extern int vmw_dmabuf_to_placement(struct vmw_private *vmw_priv,
-				   struct vmw_dma_buffer *bo,
-				   struct ttm_placement *placement,
-				   bool interruptible);
-extern int vmw_dmabuf_to_vram(struct vmw_private *dev_priv,
-			      struct vmw_dma_buffer *buf,
-			      bool pin, bool interruptible);
-extern int vmw_dmabuf_to_vram_or_gmr(struct vmw_private *dev_priv,
-				     struct vmw_dma_buffer *buf,
-				     bool pin, bool interruptible);
-extern int vmw_dmabuf_to_start_of_vram(struct vmw_private *vmw_priv,
-				       struct vmw_dma_buffer *bo,
-				       bool pin, bool interruptible);
-extern int vmw_dmabuf_unpin(struct vmw_private *vmw_priv,
-			    struct vmw_dma_buffer *bo,
-			    bool interruptible);
-extern void vmw_bo_get_guest_ptr(const struct ttm_buffer_object *buf,
-				 SVGAGuestPtr *ptr);
-extern void vmw_bo_pin(struct ttm_buffer_object *bo, bool pin);
 
 /**
  * Misc Ioctl functionality - vmwgfx_ioctl.c
@@ -471,16 +398,8 @@ extern void vmw_bo_pin(struct ttm_buffer_object *bo, bool pin);
 
 extern int vmw_getparam_ioctl(struct drm_device *dev, void *data,
 			      struct drm_file *file_priv);
-extern int vmw_get_cap_3d_ioctl(struct drm_device *dev, void *data,
+extern int vmw_fifo_debug_ioctl(struct drm_device *dev, void *data,
 				struct drm_file *file_priv);
-extern int vmw_present_ioctl(struct drm_device *dev, void *data,
-			     struct drm_file *file_priv);
-extern int vmw_present_readback_ioctl(struct drm_device *dev, void *data,
-				      struct drm_file *file_priv);
-extern unsigned int vmw_fops_poll(struct file *filp,
-				  struct poll_table_struct *wait);
-extern ssize_t vmw_fops_read(struct file *filp, char __user *buffer,
-			     size_t count, loff_t *offset);
 
 /**
  * Fifo utilities - vmwgfx_fifo.c
@@ -493,12 +412,11 @@ extern void vmw_fifo_release(struct vmw_private *dev_priv,
 extern void *vmw_fifo_reserve(struct vmw_private *dev_priv, uint32_t bytes);
 extern void vmw_fifo_commit(struct vmw_private *dev_priv, uint32_t bytes);
 extern int vmw_fifo_send_fence(struct vmw_private *dev_priv,
-			       uint32_t *seqno);
+			       uint32_t *sequence);
 extern void vmw_fifo_ping_host(struct vmw_private *dev_priv, uint32_t reason);
+extern int vmw_fifo_mmap(struct file *filp, struct vm_area_struct *vma);
 extern bool vmw_fifo_have_3d(struct vmw_private *dev_priv);
 extern bool vmw_fifo_have_pitchlock(struct vmw_private *dev_priv);
-extern int vmw_fifo_emit_dummy_query(struct vmw_private *dev_priv,
-				     uint32_t cid);
 
 /**
  * TTM glue - vmwgfx_ttm_glue.c
@@ -516,10 +434,7 @@ extern struct ttm_placement vmw_vram_placement;
 extern struct ttm_placement vmw_vram_ne_placement;
 extern struct ttm_placement vmw_vram_sys_placement;
 extern struct ttm_placement vmw_vram_gmr_placement;
-extern struct ttm_placement vmw_vram_gmr_ne_placement;
 extern struct ttm_placement vmw_sys_placement;
-extern struct ttm_placement vmw_evictable_placement;
-extern struct ttm_placement vmw_srf_placement;
 extern struct ttm_bo_driver vmw_bo_driver;
 extern int vmw_dma_quiescent(struct drm_device *dev);
 
@@ -529,71 +444,45 @@ extern int vmw_dma_quiescent(struct drm_device *dev);
 
 extern int vmw_execbuf_ioctl(struct drm_device *dev, void *data,
 			     struct drm_file *file_priv);
-extern int vmw_execbuf_process(struct drm_file *file_priv,
-			       struct vmw_private *dev_priv,
-			       void __user *user_commands,
-			       void *kernel_commands,
-			       uint32_t command_size,
-			       uint64_t throttle_us,
-			       struct drm_vmw_fence_rep __user
-			       *user_fence_rep,
-			       struct vmw_fence_obj **out_fence);
-
-extern void
-vmw_execbuf_release_pinned_bo(struct vmw_private *dev_priv,
-			      bool only_on_cid_match, uint32_t cid);
-
-extern int vmw_execbuf_fence_commands(struct drm_file *file_priv,
-				      struct vmw_private *dev_priv,
-				      struct vmw_fence_obj **p_fence,
-				      uint32_t *p_handle);
-extern void vmw_execbuf_copy_fence_user(struct vmw_private *dev_priv,
-					struct vmw_fpriv *vmw_fp,
-					int ret,
-					struct drm_vmw_fence_rep __user
-					*user_fence_rep,
-					struct vmw_fence_obj *fence,
-					uint32_t fence_handle);
 
 /**
  * IRQs and wating - vmwgfx_irq.c
  */
 
 extern irqreturn_t vmw_irq_handler(DRM_IRQ_ARGS);
-extern int vmw_wait_seqno(struct vmw_private *dev_priv, bool lazy,
-			     uint32_t seqno, bool interruptible,
-			     unsigned long timeout);
+extern int vmw_wait_fence(struct vmw_private *dev_priv, bool lazy,
+			  uint32_t sequence, bool interruptible,
+			  unsigned long timeout);
 extern void vmw_irq_preinstall(struct drm_device *dev);
 extern int vmw_irq_postinstall(struct drm_device *dev);
 extern void vmw_irq_uninstall(struct drm_device *dev);
-extern bool vmw_seqno_passed(struct vmw_private *dev_priv,
-				uint32_t seqno);
+extern bool vmw_fence_signaled(struct vmw_private *dev_priv,
+			       uint32_t sequence);
+extern int vmw_fence_wait_ioctl(struct drm_device *dev, void *data,
+				struct drm_file *file_priv);
 extern int vmw_fallback_wait(struct vmw_private *dev_priv,
 			     bool lazy,
 			     bool fifo_idle,
-			     uint32_t seqno,
+			     uint32_t sequence,
 			     bool interruptible,
 			     unsigned long timeout);
-extern void vmw_update_seqno(struct vmw_private *dev_priv,
+extern void vmw_update_sequence(struct vmw_private *dev_priv,
 				struct vmw_fifo_state *fifo_state);
-extern void vmw_seqno_waiter_add(struct vmw_private *dev_priv);
-extern void vmw_seqno_waiter_remove(struct vmw_private *dev_priv);
-extern void vmw_goal_waiter_add(struct vmw_private *dev_priv);
-extern void vmw_goal_waiter_remove(struct vmw_private *dev_priv);
+
 
 /**
- * Rudimentary fence-like objects currently used only for throttling -
- * vmwgfx_marker.c
+ * Rudimentary fence objects currently used only for throttling -
+ * vmwgfx_fence.c
  */
 
-extern void vmw_marker_queue_init(struct vmw_marker_queue *queue);
-extern void vmw_marker_queue_takedown(struct vmw_marker_queue *queue);
-extern int vmw_marker_push(struct vmw_marker_queue *queue,
-			  uint32_t seqno);
-extern int vmw_marker_pull(struct vmw_marker_queue *queue,
-			  uint32_t signaled_seqno);
+extern void vmw_fence_queue_init(struct vmw_fence_queue *queue);
+extern void vmw_fence_queue_takedown(struct vmw_fence_queue *queue);
+extern int vmw_fence_push(struct vmw_fence_queue *queue,
+			  uint32_t sequence);
+extern int vmw_fence_pull(struct vmw_fence_queue *queue,
+			  uint32_t signaled_sequence);
 extern int vmw_wait_lag(struct vmw_private *dev_priv,
-			struct vmw_marker_queue *queue, uint32_t us);
+			struct vmw_fence_queue *queue, uint32_t us);
 
 /**
  * Kernel framebuffer - vmwgfx_fb.c
@@ -619,31 +508,16 @@ void vmw_kms_cursor_snoop(struct vmw_surface *srf,
 			  struct ttm_object_file *tfile,
 			  struct ttm_buffer_object *bo,
 			  SVGA3dCmdHeader *header);
-int vmw_kms_write_svga(struct vmw_private *vmw_priv,
-		       unsigned width, unsigned height, unsigned pitch,
-		       unsigned bpp, unsigned depth);
+void vmw_kms_write_svga(struct vmw_private *vmw_priv,
+			unsigned width, unsigned height, unsigned pitch,
+			unsigned bbp, unsigned depth);
+int vmw_kms_update_layout_ioctl(struct drm_device *dev, void *data,
+				struct drm_file *file_priv);
 void vmw_kms_idle_workqueues(struct vmw_master *vmaster);
 bool vmw_kms_validate_mode_vram(struct vmw_private *dev_priv,
 				uint32_t pitch,
 				uint32_t height);
 u32 vmw_get_vblank_counter(struct drm_device *dev, int crtc);
-int vmw_enable_vblank(struct drm_device *dev, int crtc);
-void vmw_disable_vblank(struct drm_device *dev, int crtc);
-int vmw_kms_present(struct vmw_private *dev_priv,
-		    struct drm_file *file_priv,
-		    struct vmw_framebuffer *vfb,
-		    struct vmw_surface *surface,
-		    uint32_t sid, int32_t destX, int32_t destY,
-		    struct drm_vmw_rect *clips,
-		    uint32_t num_clips);
-int vmw_kms_readback(struct vmw_private *dev_priv,
-		     struct drm_file *file_priv,
-		     struct vmw_framebuffer *vfb,
-		     struct drm_vmw_fence_rep __user *user_fence_rep,
-		     struct drm_vmw_rect *clips,
-		     uint32_t num_clips);
-int vmw_kms_update_layout_ioctl(struct drm_device *dev, void *data,
-				struct drm_file *file_priv);
 
 /**
  * Overlay control - vmwgfx_overlay.c
@@ -702,8 +576,4 @@ static inline struct vmw_dma_buffer *vmw_dmabuf_reference(struct vmw_dma_buffer
 	return NULL;
 }
 
-static inline struct ttm_mem_global *vmw_mem_glob(struct vmw_private *dev_priv)
-{
-	return (struct ttm_mem_global *) dev_priv->mem_global_ref.object;
-}
 #endif
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
index 4acced4..41b95ed 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_execbuf.c
@@ -44,71 +44,10 @@ static int vmw_cmd_ok(struct vmw_private *dev_priv,
 	return 0;
 }
 
-static void vmw_resource_to_validate_list(struct vmw_sw_context *sw_context,
-					  struct vmw_resource **p_res)
-{
-	struct vmw_resource *res = *p_res;
-
-	if (list_empty(&res->validate_head)) {
-		list_add_tail(&res->validate_head, &sw_context->resource_list);
-		*p_res = NULL;
-	} else
-		vmw_resource_unreference(p_res);
-}
-
-/**
- * vmw_bo_to_validate_list - add a bo to a validate list
- *
- * @sw_context: The software context used for this command submission batch.
- * @bo: The buffer object to add.
- * @fence_flags: Fence flags to be or'ed with any other fence flags for
- * this buffer on this submission batch.
- * @p_val_node: If non-NULL Will be updated with the validate node number
- * on return.
- *
- * Returns -EINVAL if the limit of number of buffer objects per command
- * submission is reached.
- */
-static int vmw_bo_to_validate_list(struct vmw_sw_context *sw_context,
-				   struct ttm_buffer_object *bo,
-				   uint32_t fence_flags,
-				   uint32_t *p_val_node)
-{
-	uint32_t val_node;
-	struct ttm_validate_buffer *val_buf;
-
-	val_node = vmw_dmabuf_validate_node(bo, sw_context->cur_val_buf);
-
-	if (unlikely(val_node >= VMWGFX_MAX_VALIDATIONS)) {
-		DRM_ERROR("Max number of DMA buffers per submission"
-			  " exceeded.\n");
-		return -EINVAL;
-	}
-
-	val_buf = &sw_context->val_bufs[val_node];
-	if (unlikely(val_node == sw_context->cur_val_buf)) {
-		val_buf->new_sync_obj_arg = NULL;
-		val_buf->bo = ttm_bo_reference(bo);
-		list_add_tail(&val_buf->head, &sw_context->validate_nodes);
-		++sw_context->cur_val_buf;
-	}
-
-	val_buf->new_sync_obj_arg = (void *)
-		((unsigned long) val_buf->new_sync_obj_arg | fence_flags);
-	sw_context->fence_flags |= fence_flags;
-
-	if (p_val_node)
-		*p_val_node = val_node;
-
-	return 0;
-}
-
 static int vmw_cmd_cid_check(struct vmw_private *dev_priv,
 			     struct vmw_sw_context *sw_context,
 			     SVGA3dCmdHeader *header)
 {
-	struct vmw_resource *ctx;
-
 	struct vmw_cid_cmd {
 		SVGA3dCmdHeader header;
 		__le32 cid;
@@ -119,8 +58,7 @@ static int vmw_cmd_cid_check(struct vmw_private *dev_priv,
 	if (likely(sw_context->cid_valid && cmd->cid == sw_context->last_cid))
 		return 0;
 
-	ret = vmw_context_check(dev_priv, sw_context->tfile, cmd->cid,
-				&ctx);
+	ret = vmw_context_check(dev_priv, sw_context->tfile, cmd->cid);
 	if (unlikely(ret != 0)) {
 		DRM_ERROR("Could not find or use context %u\n",
 			  (unsigned) cmd->cid);
@@ -129,8 +67,6 @@ static int vmw_cmd_cid_check(struct vmw_private *dev_priv,
 
 	sw_context->last_cid = cmd->cid;
 	sw_context->cid_valid = true;
-	sw_context->cur_ctx = ctx;
-	vmw_resource_to_validate_list(sw_context, &ctx);
 
 	return 0;
 }
@@ -139,45 +75,29 @@ static int vmw_cmd_sid_check(struct vmw_private *dev_priv,
 			     struct vmw_sw_context *sw_context,
 			     uint32_t *sid)
 {
-	struct vmw_surface *srf;
-	int ret;
-	struct vmw_resource *res;
-
 	if (*sid == SVGA3D_INVALID_ID)
 		return 0;
 
-	if (likely((sw_context->sid_valid  &&
-		      *sid == sw_context->last_sid))) {
-		*sid = sw_context->sid_translation;
-		return 0;
-	}
-
-	ret = vmw_user_surface_lookup_handle(dev_priv,
-					     sw_context->tfile,
-					     *sid, &srf);
-	if (unlikely(ret != 0)) {
-		DRM_ERROR("Could ot find or use surface 0x%08x "
-			  "address 0x%08lx\n",
-			  (unsigned int) *sid,
-			  (unsigned long) sid);
-		return ret;
-	}
+	if (unlikely((!sw_context->sid_valid  ||
+		      *sid != sw_context->last_sid))) {
+		int real_id;
+		int ret = vmw_surface_check(dev_priv, sw_context->tfile,
+					    *sid, &real_id);
 
-	ret = vmw_surface_validate(dev_priv, srf);
-	if (unlikely(ret != 0)) {
-		if (ret != -ERESTARTSYS)
-			DRM_ERROR("Could not validate surface.\n");
-		vmw_surface_unreference(&srf);
-		return ret;
-	}
-
-	sw_context->last_sid = *sid;
-	sw_context->sid_valid = true;
-	sw_context->sid_translation = srf->res.id;
-	*sid = sw_context->sid_translation;
+		if (unlikely(ret != 0)) {
+			DRM_ERROR("Could ot find or use surface 0x%08x "
+				  "address 0x%08lx\n",
+				  (unsigned int) *sid,
+				  (unsigned long) sid);
+			return ret;
+		}
 
-	res = &srf->res;
-	vmw_resource_to_validate_list(sw_context, &res);
+		sw_context->last_sid = *sid;
+		sw_context->sid_valid = true;
+		*sid = real_id;
+		sw_context->sid_translation = real_id;
+	} else
+		*sid = sw_context->sid_translation;
 
 	return 0;
 }
@@ -246,12 +166,6 @@ static int vmw_cmd_blt_surf_screen_check(struct vmw_private *dev_priv,
 	} *cmd;
 
 	cmd = container_of(header, struct vmw_sid_cmd, header);
-
-	if (unlikely(!sw_context->kernel)) {
-		DRM_ERROR("Kernel only SVGA3d command: %u.\n", cmd->header.id);
-		return -EPERM;
-	}
-
 	return vmw_cmd_sid_check(dev_priv, sw_context, &cmd->body.srcImage.sid);
 }
 
@@ -264,179 +178,10 @@ static int vmw_cmd_present_check(struct vmw_private *dev_priv,
 		SVGA3dCmdPresent body;
 	} *cmd;
 
-
 	cmd = container_of(header, struct vmw_sid_cmd, header);
-
-	if (unlikely(!sw_context->kernel)) {
-		DRM_ERROR("Kernel only SVGA3d command: %u.\n", cmd->header.id);
-		return -EPERM;
-	}
-
 	return vmw_cmd_sid_check(dev_priv, sw_context, &cmd->body.sid);
 }
 
-/**
- * vmw_query_bo_switch_prepare - Prepare to switch pinned buffer for queries.
- *
- * @dev_priv: The device private structure.
- * @cid: The hardware context for the next query.
- * @new_query_bo: The new buffer holding query results.
- * @sw_context: The software context used for this command submission.
- *
- * This function checks whether @new_query_bo is suitable for holding
- * query results, and if another buffer currently is pinned for query
- * results. If so, the function prepares the state of @sw_context for
- * switching pinned buffers after successful submission of the current
- * command batch. It also checks whether we're using a new query context.
- * In that case, it makes sure we emit a query barrier for the old
- * context before the current query buffer is fenced.
- */
-static int vmw_query_bo_switch_prepare(struct vmw_private *dev_priv,
-				       uint32_t cid,
-				       struct ttm_buffer_object *new_query_bo,
-				       struct vmw_sw_context *sw_context)
-{
-	int ret;
-	bool add_cid = false;
-	uint32_t cid_to_add;
-
-	if (unlikely(new_query_bo != sw_context->cur_query_bo)) {
-
-		if (unlikely(new_query_bo->num_pages > 4)) {
-			DRM_ERROR("Query buffer too large.\n");
-			return -EINVAL;
-		}
-
-		if (unlikely(sw_context->cur_query_bo != NULL)) {
-			BUG_ON(!sw_context->query_cid_valid);
-			add_cid = true;
-			cid_to_add = sw_context->cur_query_cid;
-			ret = vmw_bo_to_validate_list(sw_context,
-						      sw_context->cur_query_bo,
-						      DRM_VMW_FENCE_FLAG_EXEC,
-						      NULL);
-			if (unlikely(ret != 0))
-				return ret;
-		}
-		sw_context->cur_query_bo = new_query_bo;
-
-		ret = vmw_bo_to_validate_list(sw_context,
-					      dev_priv->dummy_query_bo,
-					      DRM_VMW_FENCE_FLAG_EXEC,
-					      NULL);
-		if (unlikely(ret != 0))
-			return ret;
-
-	}
-
-	if (unlikely(cid != sw_context->cur_query_cid &&
-		     sw_context->query_cid_valid)) {
-		add_cid = true;
-		cid_to_add = sw_context->cur_query_cid;
-	}
-
-	sw_context->cur_query_cid = cid;
-	sw_context->query_cid_valid = true;
-
-	if (add_cid) {
-		struct vmw_resource *ctx = sw_context->cur_ctx;
-
-		if (list_empty(&ctx->query_head))
-			list_add_tail(&ctx->query_head,
-				      &sw_context->query_list);
-		ret = vmw_bo_to_validate_list(sw_context,
-					      dev_priv->dummy_query_bo,
-					      DRM_VMW_FENCE_FLAG_EXEC,
-					      NULL);
-		if (unlikely(ret != 0))
-			return ret;
-	}
-	return 0;
-}
-
-
-/**
- * vmw_query_bo_switch_commit - Finalize switching pinned query buffer
- *
- * @dev_priv: The device private structure.
- * @sw_context: The software context used for this command submission batch.
- *
- * This function will check if we're switching query buffers, and will then,
- * if no other query waits are issued this command submission batch,
- * issue a dummy occlusion query wait used as a query barrier. When the fence
- * object following that query wait has signaled, we are sure that all
- * preseding queries have finished, and the old query buffer can be unpinned.
- * However, since both the new query buffer and the old one are fenced with
- * that fence, we can do an asynchronus unpin now, and be sure that the
- * old query buffer won't be moved until the fence has signaled.
- *
- * As mentioned above, both the new - and old query buffers need to be fenced
- * using a sequence emitted *after* calling this function.
- */
-static void vmw_query_bo_switch_commit(struct vmw_private *dev_priv,
-				     struct vmw_sw_context *sw_context)
-{
-
-	struct vmw_resource *ctx, *next_ctx;
-	int ret;
-
-	/*
-	 * The validate list should still hold references to all
-	 * contexts here.
-	 */
-
-	list_for_each_entry_safe(ctx, next_ctx, &sw_context->query_list,
-				 query_head) {
-		list_del_init(&ctx->query_head);
-
-		BUG_ON(list_empty(&ctx->validate_head));
-
-		ret = vmw_fifo_emit_dummy_query(dev_priv, ctx->id);
-
-		if (unlikely(ret != 0))
-			DRM_ERROR("Out of fifo space for dummy query.\n");
-	}
-
-	if (dev_priv->pinned_bo != sw_context->cur_query_bo) {
-		if (dev_priv->pinned_bo) {
-			vmw_bo_pin(dev_priv->pinned_bo, false);
-			ttm_bo_unref(&dev_priv->pinned_bo);
-		}
-
-		vmw_bo_pin(sw_context->cur_query_bo, true);
-
-		/*
-		 * We pin also the dummy_query_bo buffer so that we
-		 * don't need to validate it when emitting
-		 * dummy queries in context destroy paths.
-		 */
-
-		vmw_bo_pin(dev_priv->dummy_query_bo, true);
-		dev_priv->dummy_query_bo_pinned = true;
-
-		dev_priv->query_cid = sw_context->cur_query_cid;
-		dev_priv->pinned_bo =
-			ttm_bo_reference(sw_context->cur_query_bo);
-	}
-}
-
-/**
- * vmw_query_switch_backoff - clear query barrier list
- * @sw_context: The sw context used for this submission batch.
- *
- * This function is used as part of an error path, where a previously
- * set up list of query barriers needs to be cleared.
- *
- */
-static void vmw_query_switch_backoff(struct vmw_sw_context *sw_context)
-{
-	struct list_head *list, *next;
-
-	list_for_each_safe(list, next, &sw_context->query_list) {
-		list_del_init(list);
-	}
-}
-
 static int vmw_translate_guest_ptr(struct vmw_private *dev_priv,
 				   struct vmw_sw_context *sw_context,
 				   SVGAGuestPtr *ptr,
@@ -446,6 +191,8 @@ static int vmw_translate_guest_ptr(struct vmw_private *dev_priv,
 	struct ttm_buffer_object *bo;
 	uint32_t handle = ptr->gmrId;
 	struct vmw_relocation *reloc;
+	uint32_t cur_validate_node;
+	struct ttm_validate_buffer *val_buf;
 	int ret;
 
 	ret = vmw_user_dmabuf_lookup(sw_context->tfile, handle, &vmw_bo);
@@ -465,11 +212,22 @@ static int vmw_translate_guest_ptr(struct vmw_private *dev_priv,
 	reloc = &sw_context->relocs[sw_context->cur_reloc++];
 	reloc->location = ptr;
 
-	ret = vmw_bo_to_validate_list(sw_context, bo, DRM_VMW_FENCE_FLAG_EXEC,
-				      &reloc->index);
-	if (unlikely(ret != 0))
+	cur_validate_node = vmw_dmabuf_validate_node(bo, sw_context->cur_val_buf);
+	if (unlikely(cur_validate_node >= VMWGFX_MAX_GMRS)) {
+		DRM_ERROR("Max number of DMA buffers per submission"
+			  " exceeded.\n");
+		ret = -EINVAL;
 		goto out_no_reloc;
+	}
 
+	reloc->index = cur_validate_node;
+	if (unlikely(cur_validate_node == sw_context->cur_val_buf)) {
+		val_buf = &sw_context->val_bufs[cur_validate_node];
+		val_buf->bo = ttm_bo_reference(bo);
+		val_buf->new_sync_obj_arg = (void *) dev_priv;
+		list_add_tail(&val_buf->head, &sw_context->validate_nodes);
+		++sw_context->cur_val_buf;
+	}
 	*vmw_bo_p = vmw_bo;
 	return 0;
 
@@ -501,11 +259,8 @@ static int vmw_cmd_end_query(struct vmw_private *dev_priv,
 	if (unlikely(ret != 0))
 		return ret;
 
-	ret = vmw_query_bo_switch_prepare(dev_priv, cmd->q.cid,
-					  &vmw_bo->base, sw_context);
-
 	vmw_dmabuf_unreference(&vmw_bo);
-	return ret;
+	return 0;
 }
 
 static int vmw_cmd_wait_query(struct vmw_private *dev_priv,
@@ -518,7 +273,6 @@ static int vmw_cmd_wait_query(struct vmw_private *dev_priv,
 		SVGA3dCmdWaitForQuery q;
 	} *cmd;
 	int ret;
-	struct vmw_resource *ctx;
 
 	cmd = container_of(header, struct vmw_query_cmd, header);
 	ret = vmw_cmd_cid_check(dev_priv, sw_context, header);
@@ -532,19 +286,10 @@ static int vmw_cmd_wait_query(struct vmw_private *dev_priv,
 		return ret;
 
 	vmw_dmabuf_unreference(&vmw_bo);
-
-	/*
-	 * This wait will act as a barrier for previous waits for this
-	 * context.
-	 */
-
-	ctx = sw_context->cur_ctx;
-	if (!list_empty(&ctx->query_head))
-		list_del_init(&ctx->query_head);
-
 	return 0;
 }
 
+
 static int vmw_cmd_dma(struct vmw_private *dev_priv,
 		       struct vmw_sw_context *sw_context,
 		       SVGA3dCmdHeader *header)
@@ -557,7 +302,6 @@ static int vmw_cmd_dma(struct vmw_private *dev_priv,
 		SVGA3dCmdSurfaceDMA dma;
 	} *cmd;
 	int ret;
-	struct vmw_resource *res;
 
 	cmd = container_of(header, struct vmw_dma_cmd, header);
 	ret = vmw_translate_guest_ptr(dev_priv, sw_context,
@@ -574,28 +318,18 @@ static int vmw_cmd_dma(struct vmw_private *dev_priv,
 		goto out_no_reloc;
 	}
 
-	ret = vmw_surface_validate(dev_priv, srf);
-	if (unlikely(ret != 0)) {
-		if (ret != -ERESTARTSYS)
-			DRM_ERROR("Culd not validate surface.\n");
-		goto out_no_validate;
-	}
-
-	/*
+	/**
 	 * Patch command stream with device SID.
 	 */
+
 	cmd->dma.host.sid = srf->res.id;
 	vmw_kms_cursor_snoop(srf, sw_context->tfile, bo, header);
-
-	vmw_dmabuf_unreference(&vmw_bo);
-
-	res = &srf->res;
-	vmw_resource_to_validate_list(sw_context, &res);
-
-	return 0;
-
-out_no_validate:
+	/**
+	 * FIXME: May deadlock here when called from the
+	 * command parsing code.
+	 */
 	vmw_surface_unreference(&srf);
+
 out_no_reloc:
 	vmw_dmabuf_unreference(&vmw_bo);
 	return ret;
@@ -685,71 +419,6 @@ static int vmw_cmd_tex_state(struct vmw_private *dev_priv,
 	return 0;
 }
 
-static int vmw_cmd_check_define_gmrfb(struct vmw_private *dev_priv,
-				      struct vmw_sw_context *sw_context,
-				      void *buf)
-{
-	struct vmw_dma_buffer *vmw_bo;
-	int ret;
-
-	struct {
-		uint32_t header;
-		SVGAFifoCmdDefineGMRFB body;
-	} *cmd = buf;
-
-	ret = vmw_translate_guest_ptr(dev_priv, sw_context,
-				      &cmd->body.ptr,
-				      &vmw_bo);
-	if (unlikely(ret != 0))
-		return ret;
-
-	vmw_dmabuf_unreference(&vmw_bo);
-
-	return ret;
-}
-
-static int vmw_cmd_check_not_3d(struct vmw_private *dev_priv,
-				struct vmw_sw_context *sw_context,
-				void *buf, uint32_t *size)
-{
-	uint32_t size_remaining = *size;
-	uint32_t cmd_id;
-
-	cmd_id = le32_to_cpu(((uint32_t *)buf)[0]);
-	switch (cmd_id) {
-	case SVGA_CMD_UPDATE:
-		*size = sizeof(uint32_t) + sizeof(SVGAFifoCmdUpdate);
-		break;
-	case SVGA_CMD_DEFINE_GMRFB:
-		*size = sizeof(uint32_t) + sizeof(SVGAFifoCmdDefineGMRFB);
-		break;
-	case SVGA_CMD_BLIT_GMRFB_TO_SCREEN:
-		*size = sizeof(uint32_t) + sizeof(SVGAFifoCmdBlitGMRFBToScreen);
-		break;
-	case SVGA_CMD_BLIT_SCREEN_TO_GMRFB:
-		*size = sizeof(uint32_t) + sizeof(SVGAFifoCmdBlitGMRFBToScreen);
-		break;
-	default:
-		DRM_ERROR("Unsupported SVGA command: %u.\n", cmd_id);
-		return -EINVAL;
-	}
-
-	if (*size > size_remaining) {
-		DRM_ERROR("Invalid SVGA command (size mismatch):"
-			  " %u.\n", cmd_id);
-		return -EINVAL;
-	}
-
-	if (unlikely(!sw_context->kernel)) {
-		DRM_ERROR("Kernel only SVGA command: %u.\n", cmd_id);
-		return -EPERM;
-	}
-
-	if (cmd_id == SVGA_CMD_DEFINE_GMRFB)
-		return vmw_cmd_check_define_gmrfb(dev_priv, sw_context, buf);
-
-	return 0;
-}
 
 typedef int (*vmw_cmd_func) (struct vmw_private *,
 			     struct vmw_sw_context *,
@@ -802,11 +471,11 @@ static int vmw_cmd_check(struct vmw_private *dev_priv,
 	SVGA3dCmdHeader *header = (SVGA3dCmdHeader *) buf;
 	int ret;
 
-	cmd_id = le32_to_cpu(((uint32_t *)buf)[0]);
-	/* Handle any none 3D commands */
-	if (unlikely(cmd_id < SVGA_CMD_MAX))
-		return vmw_cmd_check_not_3d(dev_priv, sw_context, buf, size);
-
+	cmd_id = ((uint32_t *)buf)[0];
+	if (cmd_id == SVGA_CMD_UPDATE) {
+		*size = 5 << 2;
+		return 0;
+	}
 
 	cmd_id = le32_to_cpu(header->id);
 	*size = le32_to_cpu(header->size) + sizeof(SVGA3dCmdHeader);
@@ -831,8 +500,7 @@ out_err:
 
 static int vmw_cmd_check_all(struct vmw_private *dev_priv,
 			     struct vmw_sw_context *sw_context,
-			     void *buf,
-			     uint32_t size)
+			     void *buf, uint32_t size)
 {
 	int32_t cur_size = size;
 	int ret;
@@ -882,11 +550,7 @@ static void vmw_apply_relocations(struct vmw_sw_context *sw_context)
 static void vmw_clear_validations(struct vmw_sw_context *sw_context)
 {
 	struct ttm_validate_buffer *entry, *next;
-	struct vmw_resource *res, *res_next;
 
-	/*
-	 * Drop references to DMA buffers held during command submission.
-	 */
 	list_for_each_entry_safe(entry, next, &sw_context->validate_nodes,
 				 head) {
 		list_del(&entry->head);
@@ -895,16 +559,6 @@ static void vmw_clear_validations(struct vmw_sw_context *sw_context)
 		sw_context->cur_val_buf--;
 	}
 	BUG_ON(sw_context->cur_val_buf != 0);
-
-	/*
-	 * Drop references to resources held during command submission.
-	 */
-	vmw_resource_unreserve(&sw_context->resource_list);
-	list_for_each_entry_safe(res, res_next, &sw_context->resource_list,
-				 validate_head) {
-		list_del_init(&res->validate_head);
-		vmw_resource_unreference(&res);
-	}
 }
 
 static int vmw_validate_single_buffer(struct vmw_private *dev_priv,
@@ -912,16 +566,6 @@ static int vmw_validate_single_buffer(struct vmw_private *dev_priv,
 {
 	int ret;
 
-
-	/*
-	 * Don't validate pinned buffers.
-	 */
-
-	if (bo == dev_priv->pinned_bo ||
-	    (bo == dev_priv->dummy_query_bo &&
-	     dev_priv->dummy_query_bo_pinned))
-		return 0;
-
 	/**
 	 * Put BO in VRAM if there is space, otherwise as a GMR.
 	 * If there is no space in VRAM and GMR ids are all used up,
@@ -958,209 +602,57 @@ static int vmw_validate_buffers(struct vmw_private *dev_priv,
 	return 0;
 }
 
-static int vmw_resize_cmd_bounce(struct vmw_sw_context *sw_context,
-				 uint32_t size)
-{
-	if (likely(sw_context->cmd_bounce_size >= size))
-		return 0;
-
-	if (sw_context->cmd_bounce_size == 0)
-		sw_context->cmd_bounce_size = VMWGFX_CMD_BOUNCE_INIT_SIZE;
-
-	while (sw_context->cmd_bounce_size < size) {
-		sw_context->cmd_bounce_size =
-			PAGE_ALIGN(sw_context->cmd_bounce_size +
-				   (sw_context->cmd_bounce_size >> 1));
-	}
-
-	if (sw_context->cmd_bounce != NULL)
-		vfree(sw_context->cmd_bounce);
-
-	sw_context->cmd_bounce = vmalloc(sw_context->cmd_bounce_size);
-
-	if (sw_context->cmd_bounce == NULL) {
-		DRM_ERROR("Failed to allocate command bounce buffer.\n");
-		sw_context->cmd_bounce_size = 0;
-		return -ENOMEM;
-	}
-
-	return 0;
-}
-
-/**
- * vmw_execbuf_fence_commands - create and submit a command stream fence
- *
- * Creates a fence object and submits a command stream marker.
- * If this fails for some reason, We sync the fifo and return NULL.
- * It is then safe to fence buffers with a NULL pointer.
- *
- * If @p_handle is not NULL @file_priv must also not be NULL. Creates
- * a userspace handle if @p_handle is not NULL, otherwise not.
- */
-
-int vmw_execbuf_fence_commands(struct drm_file *file_priv,
-			       struct vmw_private *dev_priv,
-			       struct vmw_fence_obj **p_fence,
-			       uint32_t *p_handle)
+int vmw_execbuf_ioctl(struct drm_device *dev, void *data,
+		      struct drm_file *file_priv)
 {
-	uint32_t sequence;
+	struct vmw_private *dev_priv = vmw_priv(dev);
+	struct drm_vmw_execbuf_arg *arg = (struct drm_vmw_execbuf_arg *)data;
+	struct drm_vmw_fence_rep fence_rep;
+	struct drm_vmw_fence_rep __user *user_fence_rep;
 	int ret;
-	bool synced = false;
+	void *user_cmd;
+	void *cmd;
+	uint32_t sequence;
+	struct vmw_sw_context *sw_context = &dev_priv->ctx;
+	struct vmw_master *vmaster = vmw_master(file_priv->master);
 
-	/* p_handle implies file_priv. */
-	BUG_ON(p_handle != NULL && file_priv == NULL);
+	ret = ttm_read_lock(&vmaster->lock, true);
+	if (unlikely(ret != 0))
+		return ret;
 
-	ret = vmw_fifo_send_fence(dev_priv, &sequence);
+	ret = mutex_lock_interruptible(&dev_priv->cmdbuf_mutex);
 	if (unlikely(ret != 0)) {
-		DRM_ERROR("Fence submission error. Syncing.\n");
-		synced = true;
-	}
-
-	if (p_handle != NULL)
-		ret = vmw_user_fence_create(file_priv, dev_priv->fman,
-					    sequence,
-					    DRM_VMW_FENCE_FLAG_EXEC,
-					    p_fence, p_handle);
-	else
-		ret = vmw_fence_create(dev_priv->fman, sequence,
-				       DRM_VMW_FENCE_FLAG_EXEC,
-				       p_fence);
-
-	if (unlikely(ret != 0 && !synced)) {
-		(void) vmw_fallback_wait(dev_priv, false, false,
-					 sequence, false,
-					 VMW_FENCE_WAIT_TIMEOUT);
-		*p_fence = NULL;
+		ret = -ERESTARTSYS;
+		goto out_no_cmd_mutex;
 	}
 
-	return 0;
-}
-
-/**
- * vmw_execbuf_copy_fence_user - copy fence object information to
- * user-space.
- *
- * @dev_priv: Pointer to a vmw_private struct.
- * @vmw_fp: Pointer to the struct vmw_fpriv representing the calling file.
- * @ret: Return value from fence object creation.
- * @user_fence_rep: User space address of a struct drm_vmw_fence_rep to
- * which the information should be copied.
- * @fence: Pointer to the fenc object.
- * @fence_handle: User-space fence handle.
- *
- * This function copies fence information to user-space. If copying fails,
- * The user-space struct drm_vmw_fence_rep::error member is hopefully
- * left untouched, and if it's preloaded with an -EFAULT by user-space,
- * the error will hopefully be detected.
- * Also if copying fails, user-space will be unable to signal the fence
- * object so we wait for it immediately, and then unreference the
- * user-space reference.
- */
-void
-vmw_execbuf_copy_fence_user(struct vmw_private *dev_priv,
-			    struct vmw_fpriv *vmw_fp,
-			    int ret,
-			    struct drm_vmw_fence_rep __user *user_fence_rep,
-			    struct vmw_fence_obj *fence,
-			    uint32_t fence_handle)
-{
-	struct drm_vmw_fence_rep fence_rep;
-
-	if (user_fence_rep == NULL)
-		return;
-
-	memset(&fence_rep, 0, sizeof(fence_rep));
-
-	fence_rep.error = ret;
-	if (ret == 0) {
-		BUG_ON(fence == NULL);
-
-		fence_rep.handle = fence_handle;
-		fence_rep.seqno = fence->seqno;
-		vmw_update_seqno(dev_priv, &dev_priv->fifo);
-		fence_rep.passed_seqno = dev_priv->last_read_seqno;
+	cmd = vmw_fifo_reserve(dev_priv, arg->command_size);
+	if (unlikely(cmd == NULL)) {
+		DRM_ERROR("Failed reserving fifo space for commands.\n");
+		ret = -ENOMEM;
+		goto out_unlock;
 	}
 
-	/*
-	 * copy_to_user errors will be detected by user space not
-	 * seeing fence_rep::error filled in. Typically
-	 * user-space would have pre-set that member to -EFAULT.
-	 */
-	ret = copy_to_user(user_fence_rep, &fence_rep,
-			   sizeof(fence_rep));
+	user_cmd = (void __user *)(unsigned long)arg->commands;
+	ret = copy_from_user(cmd, user_cmd, arg->command_size);
 
-	/*
-	 * User-space lost the fence object. We need to sync
-	 * and unreference the handle.
-	 */
-	if (unlikely(ret != 0) && (fence_rep.error == 0)) {
-		ttm_ref_object_base_unref(vmw_fp->tfile,
-					  fence_handle, TTM_REF_USAGE);
-		DRM_ERROR("Fence copy error. Syncing.\n");
-		(void) vmw_fence_obj_wait(fence, fence->signal_mask,
-					  false, false,
-					  VMW_FENCE_WAIT_TIMEOUT);
+	if (unlikely(ret != 0)) {
+		ret = -EFAULT;
+		DRM_ERROR("Failed copying commands.\n");
+		goto out_commit;
 	}
-}
-
-int vmw_execbuf_process(struct drm_file *file_priv,
-			struct vmw_private *dev_priv,
-			void __user *user_commands,
-			void *kernel_commands,
-			uint32_t command_size,
-			uint64_t throttle_us,
-			struct drm_vmw_fence_rep __user *user_fence_rep,
-			struct vmw_fence_obj **out_fence)
-{
-	struct vmw_sw_context *sw_context = &dev_priv->ctx;
-	struct vmw_fence_obj *fence = NULL;
-	uint32_t handle;
-	void *cmd;
-	int ret;
-
-	ret = mutex_lock_interruptible(&dev_priv->cmdbuf_mutex);
-	if (unlikely(ret != 0))
-		return -ERESTARTSYS;
-
-	if (kernel_commands == NULL) {
-		sw_context->kernel = false;
-
-		ret = vmw_resize_cmd_bounce(sw_context, command_size);
-		if (unlikely(ret != 0))
-			goto out_unlock;
-
-
-		ret = copy_from_user(sw_context->cmd_bounce,
-				     user_commands, command_size);
-
-		if (unlikely(ret != 0)) {
-			ret = -EFAULT;
-			DRM_ERROR("Failed copying commands.\n");
-			goto out_unlock;
-		}
-		kernel_commands = sw_context->cmd_bounce;
-	} else
-		sw_context->kernel = true;
 
 	sw_context->tfile = vmw_fpriv(file_priv)->tfile;
 	sw_context->cid_valid = false;
 	sw_context->sid_valid = false;
 	sw_context->cur_reloc = 0;
 	sw_context->cur_val_buf = 0;
-	sw_context->fence_flags = 0;
-	INIT_LIST_HEAD(&sw_context->query_list);
-	INIT_LIST_HEAD(&sw_context->resource_list);
-	sw_context->cur_query_bo = dev_priv->pinned_bo;
-	sw_context->cur_query_cid = dev_priv->query_cid;
-	sw_context->query_cid_valid = (dev_priv->pinned_bo != NULL);
 
 	INIT_LIST_HEAD(&sw_context->validate_nodes);
 
-	ret = vmw_cmd_check_all(dev_priv, sw_context, kernel_commands,
-				command_size);
+	ret = vmw_cmd_check_all(dev_priv, sw_context, cmd, arg->command_size);
 	if (unlikely(ret != 0))
 		goto out_err;
-
 	ret = ttm_eu_reserve_buffers(&sw_context->validate_nodes);
 	if (unlikely(ret != 0))
 		goto out_err;
@@ -1171,212 +663,57 @@ int vmw_execbuf_process(struct drm_file *file_priv,
 
 	vmw_apply_relocations(sw_context);
 
-	if (throttle_us) {
-		ret = vmw_wait_lag(dev_priv, &dev_priv->fifo.marker_queue,
-				   throttle_us);
+	if (arg->throttle_us) {
+		ret = vmw_wait_lag(dev_priv, &dev_priv->fifo.fence_queue,
+				   arg->throttle_us);
 
 		if (unlikely(ret != 0))
-			goto out_throttle;
+			goto out_err;
 	}
 
-	cmd = vmw_fifo_reserve(dev_priv, command_size);
-	if (unlikely(cmd == NULL)) {
-		DRM_ERROR("Failed reserving fifo space for commands.\n");
-		ret = -ENOMEM;
-		goto out_throttle;
-	}
+	vmw_fifo_commit(dev_priv, arg->command_size);
 
-	memcpy(cmd, kernel_commands, command_size);
-	vmw_fifo_commit(dev_priv, command_size);
+	ret = vmw_fifo_send_fence(dev_priv, &sequence);
+
+	ttm_eu_fence_buffer_objects(&sw_context->validate_nodes,
+				    (void *)(unsigned long) sequence);
+	vmw_clear_validations(sw_context);
+	mutex_unlock(&dev_priv->cmdbuf_mutex);
 
-	vmw_query_bo_switch_commit(dev_priv, sw_context);
-	ret = vmw_execbuf_fence_commands(file_priv, dev_priv,
-					 &fence,
-					 (user_fence_rep) ? &handle : NULL);
 	/*
 	 * This error is harmless, because if fence submission fails,
-	 * vmw_fifo_send_fence will sync. The error will be propagated to
-	 * user-space in @fence_rep
+	 * vmw_fifo_send_fence will sync.
 	 */
 
 	if (ret != 0)
 		DRM_ERROR("Fence submission error. Syncing.\n");
 
-	ttm_eu_fence_buffer_objects(&sw_context->validate_nodes,
-				    (void *) fence);
+	fence_rep.error = ret;
+	fence_rep.fence_seq = (uint64_t) sequence;
+	fence_rep.pad64 = 0;
 
-	vmw_clear_validations(sw_context);
-	vmw_execbuf_copy_fence_user(dev_priv, vmw_fpriv(file_priv), ret,
-				    user_fence_rep, fence, handle);
-
-	/* Don't unreference when handing fence out */
-	if (unlikely(out_fence != NULL)) {
-		*out_fence = fence;
-		fence = NULL;
-	} else if (likely(fence != NULL)) {
-		vmw_fence_obj_unreference(&fence);
-	}
+	user_fence_rep = (struct drm_vmw_fence_rep __user *)
+	    (unsigned long)arg->fence_rep;
 
-	mutex_unlock(&dev_priv->cmdbuf_mutex);
-	return 0;
+	/*
+	 * copy_to_user errors will be detected by user space not
+	 * seeing fence_rep::error filled in.
+	 */
+
+	ret = copy_to_user(user_fence_rep, &fence_rep, sizeof(fence_rep));
 
+	vmw_kms_cursor_post_execbuf(dev_priv);
+	ttm_read_unlock(&vmaster->lock);
+	return 0;
 out_err:
 	vmw_free_relocations(sw_context);
-out_throttle:
-	vmw_query_switch_backoff(sw_context);
 	ttm_eu_backoff_reservation(&sw_context->validate_nodes);
 	vmw_clear_validations(sw_context);
+out_commit:
+	vmw_fifo_commit(dev_priv, 0);
 out_unlock:
 	mutex_unlock(&dev_priv->cmdbuf_mutex);
-	return ret;
-}
-
-/**
- * vmw_execbuf_unpin_panic - Idle the fifo and unpin the query buffer.
- *
- * @dev_priv: The device private structure.
- *
- * This function is called to idle the fifo and unpin the query buffer
- * if the normal way to do this hits an error, which should typically be
- * extremely rare.
- */
-static void vmw_execbuf_unpin_panic(struct vmw_private *dev_priv)
-{
-	DRM_ERROR("Can't unpin query buffer. Trying to recover.\n");
-
-	(void) vmw_fallback_wait(dev_priv, false, true, 0, false, 10*HZ);
-	vmw_bo_pin(dev_priv->pinned_bo, false);
-	vmw_bo_pin(dev_priv->dummy_query_bo, false);
-	dev_priv->dummy_query_bo_pinned = false;
-}
-
-
-/**
- * vmw_execbuf_release_pinned_bo - Flush queries and unpin the pinned
- * query bo.
- *
- * @dev_priv: The device private structure.
- * @only_on_cid_match: Only flush and unpin if the current active query cid
- * matches @cid.
- * @cid: Optional context id to match.
- *
- * This function should be used to unpin the pinned query bo, or
- * as a query barrier when we need to make sure that all queries have
- * finished before the next fifo command. (For example on hardware
- * context destructions where the hardware may otherwise leak unfinished
- * queries).
- *
- * This function does not return any failure codes, but make attempts
- * to do safe unpinning in case of errors.
- *
- * The function will synchronize on the previous query barrier, and will
- * thus not finish until that barrier has executed.
- */
-void vmw_execbuf_release_pinned_bo(struct vmw_private *dev_priv,
-				   bool only_on_cid_match, uint32_t cid)
-{
-	int ret = 0;
-	struct list_head validate_list;
-	struct ttm_validate_buffer pinned_val, query_val;
-	struct vmw_fence_obj *fence;
-
-	mutex_lock(&dev_priv->cmdbuf_mutex);
-
-	if (dev_priv->pinned_bo == NULL)
-		goto out_unlock;
-
-	if (only_on_cid_match && cid != dev_priv->query_cid)
-		goto out_unlock;
-
-	INIT_LIST_HEAD(&validate_list);
-
-	pinned_val.new_sync_obj_arg = (void *)(unsigned long)
-		DRM_VMW_FENCE_FLAG_EXEC;
-	pinned_val.bo = ttm_bo_reference(dev_priv->pinned_bo);
-	list_add_tail(&pinned_val.head, &validate_list);
-
-	query_val.new_sync_obj_arg = pinned_val.new_sync_obj_arg;
-	query_val.bo = ttm_bo_reference(dev_priv->dummy_query_bo);
-	list_add_tail(&query_val.head, &validate_list);
-
-	do {
-		ret = ttm_eu_reserve_buffers(&validate_list);
-	} while (ret == -ERESTARTSYS);
-
-	if (unlikely(ret != 0)) {
-		vmw_execbuf_unpin_panic(dev_priv);
-		goto out_no_reserve;
-	}
-
-	ret = vmw_fifo_emit_dummy_query(dev_priv, dev_priv->query_cid);
-	if (unlikely(ret != 0)) {
-		vmw_execbuf_unpin_panic(dev_priv);
-		goto out_no_emit;
-	}
-
-	vmw_bo_pin(dev_priv->pinned_bo, false);
-	vmw_bo_pin(dev_priv->dummy_query_bo, false);
-	dev_priv->dummy_query_bo_pinned = false;
-
-	(void) vmw_execbuf_fence_commands(NULL, dev_priv, &fence, NULL);
-	ttm_eu_fence_buffer_objects(&validate_list, (void *) fence);
-
-	ttm_bo_unref(&query_val.bo);
-	ttm_bo_unref(&pinned_val.bo);
-	ttm_bo_unref(&dev_priv->pinned_bo);
-
-out_unlock:
-	mutex_unlock(&dev_priv->cmdbuf_mutex);
-	return;
-
-out_no_emit:
-	ttm_eu_backoff_reservation(&validate_list);
-out_no_reserve:
-	ttm_bo_unref(&query_val.bo);
-	ttm_bo_unref(&pinned_val.bo);
-	ttm_bo_unref(&dev_priv->pinned_bo);
-	mutex_unlock(&dev_priv->cmdbuf_mutex);
-}
-
-
-int vmw_execbuf_ioctl(struct drm_device *dev, void *data,
-		      struct drm_file *file_priv)
-{
-	struct vmw_private *dev_priv = vmw_priv(dev);
-	struct drm_vmw_execbuf_arg *arg = (struct drm_vmw_execbuf_arg *)data;
-	struct vmw_master *vmaster = vmw_master(file_priv->master);
-	int ret;
-
-	/*
-	 * This will allow us to extend the ioctl argument while
-	 * maintaining backwards compatibility:
-	 * We take different code paths depending on the value of
-	 * arg->version.
-	 */
-
-	if (unlikely(arg->version != DRM_VMW_EXECBUF_VERSION)) {
-		DRM_ERROR("Incorrect execbuf version.\n");
-		DRM_ERROR("You're running outdated experimental "
-			  "vmwgfx user-space drivers.");
-		return -EINVAL;
-	}
-
-	ret = ttm_read_lock(&vmaster->lock, true);
-	if (unlikely(ret != 0))
-		return ret;
-
-	ret = vmw_execbuf_process(file_priv, dev_priv,
-				  (void __user *)(unsigned long)arg->commands,
-				  NULL, arg->command_size, arg->throttle_us,
-				  (void __user *)(unsigned long)arg->fence_rep,
-				  NULL);
-
-	if (unlikely(ret != 0))
-		goto out_unlock;
-
-	vmw_kms_cursor_post_execbuf(dev_priv);
-
-out_unlock:
+out_no_cmd_mutex:
 	ttm_read_unlock(&vmaster->lock);
 	return ret;
 }
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c
index b19a7ba..bfab60c 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fb.c
@@ -158,14 +158,10 @@ static int vmw_fb_set_par(struct fb_info *info)
 {
 	struct vmw_fb_par *par = info->par;
 	struct vmw_private *vmw_priv = par->vmw_priv;
-	int ret;
-
-	ret = vmw_kms_write_svga(vmw_priv, info->var.xres, info->var.yres,
-				 info->fix.line_length,
-				 par->bpp, par->depth);
-	if (ret)
-		return ret;
 
+	vmw_kms_write_svga(vmw_priv, info->var.xres, info->var.yres,
+			   info->fix.line_length,
+			   par->bpp, par->depth);
 	if (vmw_priv->capabilities & SVGA_CAP_DISPLAY_TOPOLOGY) {
 		/* TODO check if pitch and offset changes */
 		vmw_write(vmw_priv, SVGA_REG_NUM_GUEST_DISPLAYS, 1);
@@ -409,20 +405,24 @@ int vmw_fb_init(struct vmw_private *vmw_priv)
 	struct fb_info *info;
 	unsigned initial_width, initial_height;
 	unsigned fb_width, fb_height;
-	unsigned fb_bpp, fb_depth, fb_offset, fb_pitch, fb_size;
+	unsigned fb_bbp, fb_depth, fb_offset, fb_pitch, fb_size;
 	int ret;
 
-	fb_bpp = 32;
+	/* XXX These shouldn't be hardcoded. */
+	initial_width = 800;
+	initial_height = 600;
+
+	fb_bbp = 32;
 	fb_depth = 24;
 
 	/* XXX As shouldn't these be as well. */
 	fb_width = min(vmw_priv->fb_max_width, (unsigned)2048);
 	fb_height = min(vmw_priv->fb_max_height, (unsigned)2048);
 
-	initial_width = min(vmw_priv->initial_width, fb_width);
-	initial_height = min(vmw_priv->initial_height, fb_height);
+	initial_width = min(fb_width, initial_width);
+	initial_height = min(fb_height, initial_height);
 
-	fb_pitch = fb_width * fb_bpp / 8;
+	fb_pitch = fb_width * fb_bbp / 8;
 	fb_size = fb_pitch * fb_height;
 	fb_offset = vmw_read(vmw_priv, SVGA_REG_FB_OFFSET);
 
@@ -437,7 +437,7 @@ int vmw_fb_init(struct vmw_private *vmw_priv)
 	par = info->par;
 	par->vmw_priv = vmw_priv;
 	par->depth = fb_depth;
-	par->bpp = fb_bpp;
+	par->bpp = fb_bbp;
 	par->vmalloc = NULL;
 	par->max_width = fb_width;
 	par->max_height = fb_height;
@@ -509,7 +509,19 @@ int vmw_fb_init(struct vmw_private *vmw_priv)
 	info->var.xres = initial_width;
 	info->var.yres = initial_height;
 
-	/* Use default scratch pixmap (info->pixmap.flags = FB_PIXMAP_SYSTEM) */
+#if 0
+	info->pixmap.size = 64*1024;
+	info->pixmap.buf_align = 8;
+	info->pixmap.access_align = 32;
+	info->pixmap.flags = FB_PIXMAP_SYSTEM;
+	info->pixmap.scan_align = 1;
+#else
+	info->pixmap.size = 0;
+	info->pixmap.buf_align = 8;
+	info->pixmap.access_align = 32;
+	info->pixmap.flags = FB_PIXMAP_SYSTEM;
+	info->pixmap.scan_align = 1;
+#endif
 
 	info->apertures = alloc_apertures(1);
 	if (!info->apertures) {
@@ -576,6 +588,58 @@ int vmw_fb_close(struct vmw_private *vmw_priv)
 	return 0;
 }
 
+int vmw_dmabuf_from_vram(struct vmw_private *vmw_priv,
+			 struct vmw_dma_buffer *vmw_bo)
+{
+	struct ttm_buffer_object *bo = &vmw_bo->base;
+	int ret = 0;
+
+	ret = ttm_bo_reserve(bo, false, false, false, 0);
+	if (unlikely(ret != 0))
+		return ret;
+
+	ret = ttm_bo_validate(bo, &vmw_sys_placement, false, false, false);
+	ttm_bo_unreserve(bo);
+
+	return ret;
+}
+
+int vmw_dmabuf_to_start_of_vram(struct vmw_private *vmw_priv,
+				struct vmw_dma_buffer *vmw_bo)
+{
+	struct ttm_buffer_object *bo = &vmw_bo->base;
+	struct ttm_placement ne_placement = vmw_vram_ne_placement;
+	int ret = 0;
+
+	ne_placement.lpfn = bo->num_pages;
+
+	/* interuptable? */
+	ret = ttm_write_lock(&vmw_priv->active_master->lock, false);
+	if (unlikely(ret != 0))
+		return ret;
+
+	ret = ttm_bo_reserve(bo, false, false, false, 0);
+	if (unlikely(ret != 0))
+		goto err_unlock;
+
+	if (bo->mem.mem_type == TTM_PL_VRAM &&
+	    bo->mem.start < bo->num_pages &&
+	    bo->mem.start > 0)
+		(void) ttm_bo_validate(bo, &vmw_sys_placement, false,
+				       false, false);
+
+	ret = ttm_bo_validate(bo, &ne_placement, false, false, false);
+
+	/* Could probably bug on */
+	WARN_ON(bo->offset != 0);
+
+	ttm_bo_unreserve(bo);
+err_unlock:
+	ttm_write_unlock(&vmw_priv->active_master->lock);
+
+	return ret;
+}
+
 int vmw_fb_off(struct vmw_private *vmw_priv)
 {
 	struct fb_info *info;
@@ -597,7 +661,7 @@ int vmw_fb_off(struct vmw_private *vmw_priv)
 	par->bo_ptr = NULL;
 	ttm_bo_kunmap(&par->map);
 
-	vmw_dmabuf_unpin(vmw_priv, par->vmw_bo, false);
+	vmw_dmabuf_from_vram(vmw_priv, par->vmw_bo);
 
 	return 0;
 }
@@ -623,7 +687,7 @@ int vmw_fb_on(struct vmw_private *vmw_priv)
 	/* Make sure that all overlays are stoped when we take over */
 	vmw_overlay_stop_all(vmw_priv);
 
-	ret = vmw_dmabuf_to_start_of_vram(vmw_priv, par->vmw_bo, true, false);
+	ret = vmw_dmabuf_to_start_of_vram(vmw_priv, par->vmw_bo);
 	if (unlikely(ret != 0)) {
 		DRM_ERROR("could not move buffer to start of VRAM\n");
 		goto err_no_buffer;
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c
index f2fb8f1..61eacc1 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c
@@ -1,6 +1,6 @@
 /**************************************************************************
  *
- * Copyright © 2011 VMware, Inc., Palo Alto, CA., USA
+ * Copyright (C) 2010 VMware, Inc., Palo Alto, CA., USA
  * All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
@@ -25,1130 +25,149 @@
  *
  **************************************************************************/
 
-#include "drmP.h"
-#include "vmwgfx_drv.h"
-
-#define VMW_FENCE_WRAP (1 << 31)
 
-struct vmw_fence_manager {
-	int num_fence_objects;
-	struct vmw_private *dev_priv;
-	spinlock_t lock;
-	struct list_head fence_list;
-	struct work_struct work;
-	u32 user_fence_size;
-	u32 fence_size;
-	u32 event_fence_action_size;
-	bool fifo_down;
-	struct list_head cleanup_list;
-	uint32_t pending_actions[VMW_ACTION_MAX];
-	struct mutex goal_irq_mutex;
-	bool goal_irq_on; /* Protected by @goal_irq_mutex */
-	bool seqno_valid; /* Protected by @lock, and may not be set to true
-			     without the @goal_irq_mutex held. */
-};
-
-struct vmw_user_fence {
-	struct ttm_base_object base;
-	struct vmw_fence_obj fence;
-};
-
-/**
- * struct vmw_event_fence_action - fence action that delivers a drm event.
- *
- * @e: A struct drm_pending_event that controls the event delivery.
- * @action: A struct vmw_fence_action to hook up to a fence.
- * @fence: A referenced pointer to the fence to keep it alive while @action
- * hangs on it.
- * @dev: Pointer to a struct drm_device so we can access the event stuff.
- * @kref: Both @e and @action has destructors, so we need to refcount.
- * @size: Size accounted for this object.
- * @tv_sec: If non-null, the variable pointed to will be assigned
- * current time tv_sec val when the fence signals.
- * @tv_usec: Must be set if @tv_sec is set, and the variable pointed to will
- * be assigned the current time tv_usec val when the fence signals.
- */
-struct vmw_event_fence_action {
-	struct vmw_fence_action action;
-	struct list_head fpriv_head;
-
-	struct drm_pending_event *event;
-	struct vmw_fence_obj *fence;
-	struct drm_device *dev;
+#include "vmwgfx_drv.h"
 
-	uint32_t *tv_sec;
-	uint32_t *tv_usec;
+struct vmw_fence {
+	struct list_head head;
+	uint32_t sequence;
+	struct timespec submitted;
 };
 
-/**
- * Note on fencing subsystem usage of irqs:
- * Typically the vmw_fences_update function is called
- *
- * a) When a new fence seqno has been submitted by the fifo code.
- * b) On-demand when we have waiters. Sleeping waiters will switch on the
- * ANY_FENCE irq and call vmw_fences_update function each time an ANY_FENCE
- * irq is received. When the last fence waiter is gone, that IRQ is masked
- * away.
- *
- * In situations where there are no waiters and we don't submit any new fences,
- * fence objects may not be signaled. This is perfectly OK, since there are
- * no consumers of the signaled data, but that is NOT ok when there are fence
- * actions attached to a fence. The fencing subsystem then makes use of the
- * FENCE_GOAL irq and sets the fence goal seqno to that of the next fence
- * which has an action attached, and each time vmw_fences_update is called,
- * the subsystem makes sure the fence goal seqno is updated.
- *
- * The fence goal seqno irq is on as long as there are unsignaled fence
- * objects with actions attached to them.
- */
-
-static void vmw_fence_obj_destroy_locked(struct kref *kref)
+void vmw_fence_queue_init(struct vmw_fence_queue *queue)
 {
-	struct vmw_fence_obj *fence =
-		container_of(kref, struct vmw_fence_obj, kref);
-
-	struct vmw_fence_manager *fman = fence->fman;
-	unsigned int num_fences;
-
-	list_del_init(&fence->head);
-	num_fences = --fman->num_fence_objects;
-	spin_unlock_irq(&fman->lock);
-	if (fence->destroy)
-		fence->destroy(fence);
-	else
-		kfree(fence);
-
-	spin_lock_irq(&fman->lock);
+	INIT_LIST_HEAD(&queue->head);
+	queue->lag = ns_to_timespec(0);
+	getrawmonotonic(&queue->lag_time);
+	spin_lock_init(&queue->lock);
 }
 
-
-/**
- * Execute signal actions on fences recently signaled.
- * This is done from a workqueue so we don't have to execute
- * signal actions from atomic context.
- */
-
-static void vmw_fence_work_func(struct work_struct *work)
+void vmw_fence_queue_takedown(struct vmw_fence_queue *queue)
 {
-	struct vmw_fence_manager *fman =
-		container_of(work, struct vmw_fence_manager, work);
-	struct list_head list;
-	struct vmw_fence_action *action, *next_action;
-	bool seqno_valid;
-
-	do {
-		INIT_LIST_HEAD(&list);
-		mutex_lock(&fman->goal_irq_mutex);
-
-		spin_lock_irq(&fman->lock);
-		list_splice_init(&fman->cleanup_list, &list);
-		seqno_valid = fman->seqno_valid;
-		spin_unlock_irq(&fman->lock);
-
-		if (!seqno_valid && fman->goal_irq_on) {
-			fman->goal_irq_on = false;
-			vmw_goal_waiter_remove(fman->dev_priv);
-		}
-		mutex_unlock(&fman->goal_irq_mutex);
+	struct vmw_fence *fence, *next;
 
-		if (list_empty(&list))
-			return;
-
-		/*
-		 * At this point, only we should be able to manipulate the
-		 * list heads of the actions we have on the private list.
-		 * hence fman::lock not held.
-		 */
-
-		list_for_each_entry_safe(action, next_action, &list, head) {
-			list_del_init(&action->head);
-			if (action->cleanup)
-				action->cleanup(action);
-		}
-	} while (1);
-}
-
-struct vmw_fence_manager *vmw_fence_manager_init(struct vmw_private *dev_priv)
-{
-	struct vmw_fence_manager *fman = kzalloc(sizeof(*fman), GFP_KERNEL);
-
-	if (unlikely(fman == NULL))
-		return NULL;
-
-	fman->dev_priv = dev_priv;
-	spin_lock_init(&fman->lock);
-	INIT_LIST_HEAD(&fman->fence_list);
-	INIT_LIST_HEAD(&fman->cleanup_list);
-	INIT_WORK(&fman->work, &vmw_fence_work_func);
-	fman->fifo_down = true;
-	fman->user_fence_size = ttm_round_pot(sizeof(struct vmw_user_fence));
-	fman->fence_size = ttm_round_pot(sizeof(struct vmw_fence_obj));
-	fman->event_fence_action_size =
-		ttm_round_pot(sizeof(struct vmw_event_fence_action));
-	mutex_init(&fman->goal_irq_mutex);
-
-	return fman;
-}
-
-void vmw_fence_manager_takedown(struct vmw_fence_manager *fman)
-{
-	unsigned long irq_flags;
-	bool lists_empty;
-
-	(void) cancel_work_sync(&fman->work);
-
-	spin_lock_irqsave(&fman->lock, irq_flags);
-	lists_empty = list_empty(&fman->fence_list) &&
-		list_empty(&fman->cleanup_list);
-	spin_unlock_irqrestore(&fman->lock, irq_flags);
-
-	BUG_ON(!lists_empty);
-	kfree(fman);
-}
-
-static int vmw_fence_obj_init(struct vmw_fence_manager *fman,
-			      struct vmw_fence_obj *fence,
-			      u32 seqno,
-			      uint32_t mask,
-			      void (*destroy) (struct vmw_fence_obj *fence))
-{
-	unsigned long irq_flags;
-	unsigned int num_fences;
-	int ret = 0;
-
-	fence->seqno = seqno;
-	INIT_LIST_HEAD(&fence->seq_passed_actions);
-	fence->fman = fman;
-	fence->signaled = 0;
-	fence->signal_mask = mask;
-	kref_init(&fence->kref);
-	fence->destroy = destroy;
-	init_waitqueue_head(&fence->queue);
-
-	spin_lock_irqsave(&fman->lock, irq_flags);
-	if (unlikely(fman->fifo_down)) {
-		ret = -EBUSY;
-		goto out_unlock;
-	}
-	list_add_tail(&fence->head, &fman->fence_list);
-	num_fences = ++fman->num_fence_objects;
-
-out_unlock:
-	spin_unlock_irqrestore(&fman->lock, irq_flags);
-	return ret;
-
-}
-
-struct vmw_fence_obj *vmw_fence_obj_reference(struct vmw_fence_obj *fence)
-{
-	if (unlikely(fence == NULL))
-		return NULL;
-
-	kref_get(&fence->kref);
-	return fence;
-}
-
-/**
- * vmw_fence_obj_unreference
- *
- * Note that this function may not be entered with disabled irqs since
- * it may re-enable them in the destroy function.
- *
- */
-void vmw_fence_obj_unreference(struct vmw_fence_obj **fence_p)
-{
-	struct vmw_fence_obj *fence = *fence_p;
-	struct vmw_fence_manager *fman;
-
-	if (unlikely(fence == NULL))
-		return;
-
-	fman = fence->fman;
-	*fence_p = NULL;
-	spin_lock_irq(&fman->lock);
-	BUG_ON(atomic_read(&fence->kref.refcount) == 0);
-	kref_put(&fence->kref, vmw_fence_obj_destroy_locked);
-	spin_unlock_irq(&fman->lock);
-}
-
-void vmw_fences_perform_actions(struct vmw_fence_manager *fman,
-				struct list_head *list)
-{
-	struct vmw_fence_action *action, *next_action;
-
-	list_for_each_entry_safe(action, next_action, list, head) {
-		list_del_init(&action->head);
-		fman->pending_actions[action->type]--;
-		if (action->seq_passed != NULL)
-			action->seq_passed(action);
-
-		/*
-		 * Add the cleanup action to the cleanup list so that
-		 * it will be performed by a worker task.
-		 */
-
-		list_add_tail(&action->head, &fman->cleanup_list);
-	}
-}
-
-/**
- * vmw_fence_goal_new_locked - Figure out a new device fence goal
- * seqno if needed.
- *
- * @fman: Pointer to a fence manager.
- * @passed_seqno: The seqno the device currently signals as passed.
- *
- * This function should be called with the fence manager lock held.
- * It is typically called when we have a new passed_seqno, and
- * we might need to update the fence goal. It checks to see whether
- * the current fence goal has already passed, and, in that case,
- * scans through all unsignaled fences to get the next fence object with an
- * action attached, and sets the seqno of that fence as a new fence goal.
- *
- * returns true if the device goal seqno was updated. False otherwise.
- */
-static bool vmw_fence_goal_new_locked(struct vmw_fence_manager *fman,
-				      u32 passed_seqno)
-{
-	u32 goal_seqno;
-	__le32 __iomem *fifo_mem;
-	struct vmw_fence_obj *fence;
-
-	if (likely(!fman->seqno_valid))
-		return false;
-
-	fifo_mem = fman->dev_priv->mmio_virt;
-	goal_seqno = ioread32(fifo_mem + SVGA_FIFO_FENCE_GOAL);
-	if (likely(passed_seqno - goal_seqno >= VMW_FENCE_WRAP))
-		return false;
-
-	fman->seqno_valid = false;
-	list_for_each_entry(fence, &fman->fence_list, head) {
-		if (!list_empty(&fence->seq_passed_actions)) {
-			fman->seqno_valid = true;
-			iowrite32(fence->seqno,
-				  fifo_mem + SVGA_FIFO_FENCE_GOAL);
-			break;
-		}
-	}
-
-	return true;
-}
-
-
-/**
- * vmw_fence_goal_check_locked - Replace the device fence goal seqno if
- * needed.
- *
- * @fence: Pointer to a struct vmw_fence_obj the seqno of which should be
- * considered as a device fence goal.
- *
- * This function should be called with the fence manager lock held.
- * It is typically called when an action has been attached to a fence to
- * check whether the seqno of that fence should be used for a fence
- * goal interrupt. This is typically needed if the current fence goal is
- * invalid, or has a higher seqno than that of the current fence object.
- *
- * returns true if the device goal seqno was updated. False otherwise.
- */
-static bool vmw_fence_goal_check_locked(struct vmw_fence_obj *fence)
-{
-	u32 goal_seqno;
-	__le32 __iomem *fifo_mem;
-
-	if (fence->signaled & DRM_VMW_FENCE_FLAG_EXEC)
-		return false;
-
-	fifo_mem = fence->fman->dev_priv->mmio_virt;
-	goal_seqno = ioread32(fifo_mem + SVGA_FIFO_FENCE_GOAL);
-	if (likely(fence->fman->seqno_valid &&
-		   goal_seqno - fence->seqno < VMW_FENCE_WRAP))
-		return false;
-
-	iowrite32(fence->seqno, fifo_mem + SVGA_FIFO_FENCE_GOAL);
-	fence->fman->seqno_valid = true;
-
-	return true;
-}
-
-void vmw_fences_update(struct vmw_fence_manager *fman)
-{
-	unsigned long flags;
-	struct vmw_fence_obj *fence, *next_fence;
-	struct list_head action_list;
-	bool needs_rerun;
-	uint32_t seqno, new_seqno;
-	__le32 __iomem *fifo_mem = fman->dev_priv->mmio_virt;
-
-	seqno = ioread32(fifo_mem + SVGA_FIFO_FENCE);
-rerun:
-	spin_lock_irqsave(&fman->lock, flags);
-	list_for_each_entry_safe(fence, next_fence, &fman->fence_list, head) {
-		if (seqno - fence->seqno < VMW_FENCE_WRAP) {
-			list_del_init(&fence->head);
-			fence->signaled |= DRM_VMW_FENCE_FLAG_EXEC;
-			INIT_LIST_HEAD(&action_list);
-			list_splice_init(&fence->seq_passed_actions,
-					 &action_list);
-			vmw_fences_perform_actions(fman, &action_list);
-			wake_up_all(&fence->queue);
-		} else
-			break;
-	}
-
-	needs_rerun = vmw_fence_goal_new_locked(fman, seqno);
-
-	if (!list_empty(&fman->cleanup_list))
-		(void) schedule_work(&fman->work);
-	spin_unlock_irqrestore(&fman->lock, flags);
-
-	/*
-	 * Rerun if the fence goal seqno was updated, and the
-	 * hardware might have raced with that update, so that
-	 * we missed a fence_goal irq.
-	 */
-
-	if (unlikely(needs_rerun)) {
-		new_seqno = ioread32(fifo_mem + SVGA_FIFO_FENCE);
-		if (new_seqno != seqno) {
-			seqno = new_seqno;
-			goto rerun;
-		}
-	}
-}
-
-bool vmw_fence_obj_signaled(struct vmw_fence_obj *fence,
-			    uint32_t flags)
-{
-	struct vmw_fence_manager *fman = fence->fman;
-	unsigned long irq_flags;
-	uint32_t signaled;
-
-	spin_lock_irqsave(&fman->lock, irq_flags);
-	signaled = fence->signaled;
-	spin_unlock_irqrestore(&fman->lock, irq_flags);
-
-	flags &= fence->signal_mask;
-	if ((signaled & flags) == flags)
-		return 1;
-
-	if ((signaled & DRM_VMW_FENCE_FLAG_EXEC) == 0)
-		vmw_fences_update(fman);
-
-	spin_lock_irqsave(&fman->lock, irq_flags);
-	signaled = fence->signaled;
-	spin_unlock_irqrestore(&fman->lock, irq_flags);
-
-	return ((signaled & flags) == flags);
-}
-
-int vmw_fence_obj_wait(struct vmw_fence_obj *fence,
-		       uint32_t flags, bool lazy,
-		       bool interruptible, unsigned long timeout)
-{
-	struct vmw_private *dev_priv = fence->fman->dev_priv;
-	long ret;
-
-	if (likely(vmw_fence_obj_signaled(fence, flags)))
-		return 0;
-
-	vmw_fifo_ping_host(dev_priv, SVGA_SYNC_GENERIC);
-	vmw_seqno_waiter_add(dev_priv);
-
-	if (interruptible)
-		ret = wait_event_interruptible_timeout
-			(fence->queue,
-			 vmw_fence_obj_signaled(fence, flags),
-			 timeout);
-	else
-		ret = wait_event_timeout
-			(fence->queue,
-			 vmw_fence_obj_signaled(fence, flags),
-			 timeout);
-
-	vmw_seqno_waiter_remove(dev_priv);
-
-	if (unlikely(ret == 0))
-		ret = -EBUSY;
-	else if (likely(ret > 0))
-		ret = 0;
-
-	return ret;
-}
-
-void vmw_fence_obj_flush(struct vmw_fence_obj *fence)
-{
-	struct vmw_private *dev_priv = fence->fman->dev_priv;
-
-	vmw_fifo_ping_host(dev_priv, SVGA_SYNC_GENERIC);
-}
-
-static void vmw_fence_destroy(struct vmw_fence_obj *fence)
-{
-	struct vmw_fence_manager *fman = fence->fman;
-
-	kfree(fence);
-	/*
-	 * Free kernel space accounting.
-	 */
-	ttm_mem_global_free(vmw_mem_glob(fman->dev_priv),
-			    fman->fence_size);
-}
-
-int vmw_fence_create(struct vmw_fence_manager *fman,
-		     uint32_t seqno,
-		     uint32_t mask,
-		     struct vmw_fence_obj **p_fence)
-{
-	struct ttm_mem_global *mem_glob = vmw_mem_glob(fman->dev_priv);
-	struct vmw_fence_obj *fence;
-	int ret;
-
-	ret = ttm_mem_global_alloc(mem_glob, fman->fence_size,
-				   false, false);
-	if (unlikely(ret != 0))
-		return ret;
-
-	fence = kzalloc(sizeof(*fence), GFP_KERNEL);
-	if (unlikely(fence == NULL)) {
-		ret = -ENOMEM;
-		goto out_no_object;
+	spin_lock(&queue->lock);
+	list_for_each_entry_safe(fence, next, &queue->head, head) {
+		kfree(fence);
 	}
-
-	ret = vmw_fence_obj_init(fman, fence, seqno, mask,
-				 vmw_fence_destroy);
-	if (unlikely(ret != 0))
-		goto out_err_init;
-
-	*p_fence = fence;
-	return 0;
-
-out_err_init:
-	kfree(fence);
-out_no_object:
-	ttm_mem_global_free(mem_glob, fman->fence_size);
-	return ret;
+	spin_unlock(&queue->lock);
 }
 
-
-static void vmw_user_fence_destroy(struct vmw_fence_obj *fence)
+int vmw_fence_push(struct vmw_fence_queue *queue,
+		   uint32_t sequence)
 {
-	struct vmw_user_fence *ufence =
-		container_of(fence, struct vmw_user_fence, fence);
-	struct vmw_fence_manager *fman = fence->fman;
+	struct vmw_fence *fence = kmalloc(sizeof(*fence), GFP_KERNEL);
 
-	kfree(ufence);
-	/*
-	 * Free kernel space accounting.
-	 */
-	ttm_mem_global_free(vmw_mem_glob(fman->dev_priv),
-			    fman->user_fence_size);
-}
-
-static void vmw_user_fence_base_release(struct ttm_base_object **p_base)
-{
-	struct ttm_base_object *base = *p_base;
-	struct vmw_user_fence *ufence =
-		container_of(base, struct vmw_user_fence, base);
-	struct vmw_fence_obj *fence = &ufence->fence;
-
-	*p_base = NULL;
-	vmw_fence_obj_unreference(&fence);
-}
-
-int vmw_user_fence_create(struct drm_file *file_priv,
-			  struct vmw_fence_manager *fman,
-			  uint32_t seqno,
-			  uint32_t mask,
-			  struct vmw_fence_obj **p_fence,
-			  uint32_t *p_handle)
-{
-	struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile;
-	struct vmw_user_fence *ufence;
-	struct vmw_fence_obj *tmp;
-	struct ttm_mem_global *mem_glob = vmw_mem_glob(fman->dev_priv);
-	int ret;
-
-	/*
-	 * Kernel memory space accounting, since this object may
-	 * be created by a user-space request.
-	 */
-
-	ret = ttm_mem_global_alloc(mem_glob, fman->user_fence_size,
-				   false, false);
-	if (unlikely(ret != 0))
-		return ret;
-
-	ufence = kzalloc(sizeof(*ufence), GFP_KERNEL);
-	if (unlikely(ufence == NULL)) {
-		ret = -ENOMEM;
-		goto out_no_object;
-	}
-
-	ret = vmw_fence_obj_init(fman, &ufence->fence, seqno,
-				 mask, vmw_user_fence_destroy);
-	if (unlikely(ret != 0)) {
-		kfree(ufence);
-		goto out_no_object;
-	}
-
-	/*
-	 * The base object holds a reference which is freed in
-	 * vmw_user_fence_base_release.
-	 */
-	tmp = vmw_fence_obj_reference(&ufence->fence);
-	ret = ttm_base_object_init(tfile, &ufence->base, false,
-				   VMW_RES_FENCE,
-				   &vmw_user_fence_base_release, NULL);
-
-
-	if (unlikely(ret != 0)) {
-		/*
-		 * Free the base object's reference
-		 */
-		vmw_fence_obj_unreference(&tmp);
-		goto out_err;
-	}
+	if (unlikely(!fence))
+		return -ENOMEM;
 
-	*p_fence = &ufence->fence;
-	*p_handle = ufence->base.hash.key;
+	fence->sequence = sequence;
+	getrawmonotonic(&fence->submitted);
+	spin_lock(&queue->lock);
+	list_add_tail(&fence->head, &queue->head);
+	spin_unlock(&queue->lock);
 
 	return 0;
-out_err:
-	tmp = &ufence->fence;
-	vmw_fence_obj_unreference(&tmp);
-out_no_object:
-	ttm_mem_global_free(mem_glob, fman->user_fence_size);
-	return ret;
-}
-
-
-/**
- * vmw_fence_fifo_down - signal all unsignaled fence objects.
- */
-
-void vmw_fence_fifo_down(struct vmw_fence_manager *fman)
-{
-	unsigned long irq_flags;
-	struct list_head action_list;
-	int ret;
-
-	/*
-	 * The list may be altered while we traverse it, so always
-	 * restart when we've released the fman->lock.
-	 */
-
-	spin_lock_irqsave(&fman->lock, irq_flags);
-	fman->fifo_down = true;
-	while (!list_empty(&fman->fence_list)) {
-		struct vmw_fence_obj *fence =
-			list_entry(fman->fence_list.prev, struct vmw_fence_obj,
-				   head);
-		kref_get(&fence->kref);
-		spin_unlock_irq(&fman->lock);
-
-		ret = vmw_fence_obj_wait(fence, fence->signal_mask,
-					 false, false,
-					 VMW_FENCE_WAIT_TIMEOUT);
-
-		if (unlikely(ret != 0)) {
-			list_del_init(&fence->head);
-			fence->signaled |= DRM_VMW_FENCE_FLAG_EXEC;
-			INIT_LIST_HEAD(&action_list);
-			list_splice_init(&fence->seq_passed_actions,
-					 &action_list);
-			vmw_fences_perform_actions(fman, &action_list);
-			wake_up_all(&fence->queue);
-		}
-
-		spin_lock_irq(&fman->lock);
-
-		BUG_ON(!list_empty(&fence->head));
-		kref_put(&fence->kref, vmw_fence_obj_destroy_locked);
-	}
-	spin_unlock_irqrestore(&fman->lock, irq_flags);
-}
-
-void vmw_fence_fifo_up(struct vmw_fence_manager *fman)
-{
-	unsigned long irq_flags;
-
-	spin_lock_irqsave(&fman->lock, irq_flags);
-	fman->fifo_down = false;
-	spin_unlock_irqrestore(&fman->lock, irq_flags);
 }
 
-
-int vmw_fence_obj_wait_ioctl(struct drm_device *dev, void *data,
-			     struct drm_file *file_priv)
+int vmw_fence_pull(struct vmw_fence_queue *queue,
+		   uint32_t signaled_sequence)
 {
-	struct drm_vmw_fence_wait_arg *arg =
-	    (struct drm_vmw_fence_wait_arg *)data;
-	unsigned long timeout;
-	struct ttm_base_object *base;
-	struct vmw_fence_obj *fence;
-	struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile;
-	int ret;
-	uint64_t wait_timeout = ((uint64_t)arg->timeout_us * HZ);
+	struct vmw_fence *fence, *next;
+	struct timespec now;
+	bool updated = false;
 
-	/*
-	 * 64-bit division not present on 32-bit systems, so do an
-	 * approximation. (Divide by 1000000).
-	 */
+	spin_lock(&queue->lock);
+	getrawmonotonic(&now);
 
-	wait_timeout = (wait_timeout >> 20) + (wait_timeout >> 24) -
-	  (wait_timeout >> 26);
-
-	if (!arg->cookie_valid) {
-		arg->cookie_valid = 1;
-		arg->kernel_cookie = jiffies + wait_timeout;
-	}
-
-	base = ttm_base_object_lookup(tfile, arg->handle);
-	if (unlikely(base == NULL)) {
-		printk(KERN_ERR "Wait invalid fence object handle "
-		       "0x%08lx.\n",
-		       (unsigned long)arg->handle);
-		return -EINVAL;
-	}
-
-	fence = &(container_of(base, struct vmw_user_fence, base)->fence);
-
-	timeout = jiffies;
-	if (time_after_eq(timeout, (unsigned long)arg->kernel_cookie)) {
-		ret = ((vmw_fence_obj_signaled(fence, arg->flags)) ?
-		       0 : -EBUSY);
-		goto out;
+	if (list_empty(&queue->head)) {
+		queue->lag = ns_to_timespec(0);
+		queue->lag_time = now;
+		updated = true;
+		goto out_unlock;
 	}
 
-	timeout = (unsigned long)arg->kernel_cookie - timeout;
-
-	ret = vmw_fence_obj_wait(fence, arg->flags, arg->lazy, true, timeout);
-
-out:
-	ttm_base_object_unref(&base);
-
-	/*
-	 * Optionally unref the fence object.
-	 */
-
-	if (ret == 0 && (arg->wait_options & DRM_VMW_WAIT_OPTION_UNREF))
-		return ttm_ref_object_base_unref(tfile, arg->handle,
-						 TTM_REF_USAGE);
-	return ret;
-}
-
-int vmw_fence_obj_signaled_ioctl(struct drm_device *dev, void *data,
-				 struct drm_file *file_priv)
-{
-	struct drm_vmw_fence_signaled_arg *arg =
-		(struct drm_vmw_fence_signaled_arg *) data;
-	struct ttm_base_object *base;
-	struct vmw_fence_obj *fence;
-	struct vmw_fence_manager *fman;
-	struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile;
-	struct vmw_private *dev_priv = vmw_priv(dev);
+	list_for_each_entry_safe(fence, next, &queue->head, head) {
+		if (signaled_sequence - fence->sequence > (1 << 30))
+			continue;
 
-	base = ttm_base_object_lookup(tfile, arg->handle);
-	if (unlikely(base == NULL)) {
-		printk(KERN_ERR "Fence signaled invalid fence object handle "
-		       "0x%08lx.\n",
-		       (unsigned long)arg->handle);
-		return -EINVAL;
+		queue->lag = timespec_sub(now, fence->submitted);
+		queue->lag_time = now;
+		updated = true;
+		list_del(&fence->head);
+		kfree(fence);
 	}
 
-	fence = &(container_of(base, struct vmw_user_fence, base)->fence);
-	fman = fence->fman;
-
-	arg->signaled = vmw_fence_obj_signaled(fence, arg->flags);
-	spin_lock_irq(&fman->lock);
-
-	arg->signaled_flags = fence->signaled;
-	arg->passed_seqno = dev_priv->last_read_seqno;
-	spin_unlock_irq(&fman->lock);
-
-	ttm_base_object_unref(&base);
-
-	return 0;
-}
-
-
-int vmw_fence_obj_unref_ioctl(struct drm_device *dev, void *data,
-			      struct drm_file *file_priv)
-{
-	struct drm_vmw_fence_arg *arg =
-		(struct drm_vmw_fence_arg *) data;
-
-	return ttm_ref_object_base_unref(vmw_fpriv(file_priv)->tfile,
-					 arg->handle,
-					 TTM_REF_USAGE);
-}
-
-/**
- * vmw_event_fence_fpriv_gone - Remove references to struct drm_file objects
- *
- * @fman: Pointer to a struct vmw_fence_manager
- * @event_list: Pointer to linked list of struct vmw_event_fence_action objects
- * with pointers to a struct drm_file object about to be closed.
- *
- * This function removes all pending fence events with references to a
- * specific struct drm_file object about to be closed. The caller is required
- * to pass a list of all struct vmw_event_fence_action objects with such
- * events attached. This function is typically called before the
- * struct drm_file object's event management is taken down.
- */
-void vmw_event_fence_fpriv_gone(struct vmw_fence_manager *fman,
-				struct list_head *event_list)
-{
-	struct vmw_event_fence_action *eaction;
-	struct drm_pending_event *event;
-	unsigned long irq_flags;
-
-	while (1) {
-		spin_lock_irqsave(&fman->lock, irq_flags);
-		if (list_empty(event_list))
-			goto out_unlock;
-		eaction = list_first_entry(event_list,
-					   struct vmw_event_fence_action,
-					   fpriv_head);
-		list_del_init(&eaction->fpriv_head);
-		event = eaction->event;
-		eaction->event = NULL;
-		spin_unlock_irqrestore(&fman->lock, irq_flags);
-		event->destroy(event);
-	}
 out_unlock:
-	spin_unlock_irqrestore(&fman->lock, irq_flags);
-}
-
-
-/**
- * vmw_event_fence_action_seq_passed
- *
- * @action: The struct vmw_fence_action embedded in a struct
- * vmw_event_fence_action.
- *
- * This function is called when the seqno of the fence where @action is
- * attached has passed. It queues the event on the submitter's event list.
- * This function is always called from atomic context, and may be called
- * from irq context.
- */
-static void vmw_event_fence_action_seq_passed(struct vmw_fence_action *action)
-{
-	struct vmw_event_fence_action *eaction =
-		container_of(action, struct vmw_event_fence_action, action);
-	struct drm_device *dev = eaction->dev;
-	struct drm_pending_event *event = eaction->event;
-	struct drm_file *file_priv;
-	unsigned long irq_flags;
-
-	if (unlikely(event == NULL))
-		return;
-
-	file_priv = event->file_priv;
-	spin_lock_irqsave(&dev->event_lock, irq_flags);
-
-	if (likely(eaction->tv_sec != NULL)) {
-		struct timeval tv;
-
-		do_gettimeofday(&tv);
-		*eaction->tv_sec = tv.tv_sec;
-		*eaction->tv_usec = tv.tv_usec;
-	}
-
-	list_del_init(&eaction->fpriv_head);
-	list_add_tail(&eaction->event->link, &file_priv->event_list);
-	eaction->event = NULL;
-	wake_up_all(&file_priv->event_wait);
-	spin_unlock_irqrestore(&dev->event_lock, irq_flags);
-}
-
-/**
- * vmw_event_fence_action_cleanup
- *
- * @action: The struct vmw_fence_action embedded in a struct
- * vmw_event_fence_action.
- *
- * This function is the struct vmw_fence_action destructor. It's typically
- * called from a workqueue.
- */
-static void vmw_event_fence_action_cleanup(struct vmw_fence_action *action)
-{
-	struct vmw_event_fence_action *eaction =
-		container_of(action, struct vmw_event_fence_action, action);
-	struct vmw_fence_manager *fman = eaction->fence->fman;
-	unsigned long irq_flags;
+	spin_unlock(&queue->lock);
 
-	spin_lock_irqsave(&fman->lock, irq_flags);
-	list_del(&eaction->fpriv_head);
-	spin_unlock_irqrestore(&fman->lock, irq_flags);
-
-	vmw_fence_obj_unreference(&eaction->fence);
-	kfree(eaction);
+	return (updated) ? 0 : -EBUSY;
 }
 
-
-/**
- * vmw_fence_obj_add_action - Add an action to a fence object.
- *
- * @fence - The fence object.
- * @action - The action to add.
- *
- * Note that the action callbacks may be executed before this function
- * returns.
- */
-void vmw_fence_obj_add_action(struct vmw_fence_obj *fence,
-			      struct vmw_fence_action *action)
+static struct timespec vmw_timespec_add(struct timespec t1,
+					struct timespec t2)
 {
-	struct vmw_fence_manager *fman = fence->fman;
-	unsigned long irq_flags;
-	bool run_update = false;
-
-	mutex_lock(&fman->goal_irq_mutex);
-	spin_lock_irqsave(&fman->lock, irq_flags);
-
-	fman->pending_actions[action->type]++;
-	if (fence->signaled & DRM_VMW_FENCE_FLAG_EXEC) {
-		struct list_head action_list;
-
-		INIT_LIST_HEAD(&action_list);
-		list_add_tail(&action->head, &action_list);
-		vmw_fences_perform_actions(fman, &action_list);
-	} else {
-		list_add_tail(&action->head, &fence->seq_passed_actions);
-
-		/*
-		 * This function may set fman::seqno_valid, so it must
-		 * be run with the goal_irq_mutex held.
-		 */
-		run_update = vmw_fence_goal_check_locked(fence);
-	}
-
-	spin_unlock_irqrestore(&fman->lock, irq_flags);
-
-	if (run_update) {
-		if (!fman->goal_irq_on) {
-			fman->goal_irq_on = true;
-			vmw_goal_waiter_add(fman->dev_priv);
-		}
-		vmw_fences_update(fman);
+	t1.tv_sec += t2.tv_sec;
+	t1.tv_nsec += t2.tv_nsec;
+	if (t1.tv_nsec >= 1000000000L) {
+		t1.tv_sec += 1;
+		t1.tv_nsec -= 1000000000L;
 	}
-	mutex_unlock(&fman->goal_irq_mutex);
 
+	return t1;
 }
 
-/**
- * vmw_event_fence_action_create - Post an event for sending when a fence
- * object seqno has passed.
- *
- * @file_priv: The file connection on which the event should be posted.
- * @fence: The fence object on which to post the event.
- * @event: Event to be posted. This event should've been alloced
- * using k[mz]alloc, and should've been completely initialized.
- * @interruptible: Interruptible waits if possible.
- *
- * As a side effect, the object pointed to by @event may have been
- * freed when this function returns. If this function returns with
- * an error code, the caller needs to free that object.
- */
-
-int vmw_event_fence_action_queue(struct drm_file *file_priv,
-				 struct vmw_fence_obj *fence,
-				 struct drm_pending_event *event,
-				 uint32_t *tv_sec,
-				 uint32_t *tv_usec,
-				 bool interruptible)
+static struct timespec vmw_fifo_lag(struct vmw_fence_queue *queue)
 {
-	struct vmw_event_fence_action *eaction;
-	struct vmw_fence_manager *fman = fence->fman;
-	struct vmw_fpriv *vmw_fp = vmw_fpriv(file_priv);
-	unsigned long irq_flags;
-
-	eaction = kzalloc(sizeof(*eaction), GFP_KERNEL);
-	if (unlikely(eaction == NULL))
-		return -ENOMEM;
-
-	eaction->event = event;
-
-	eaction->action.seq_passed = vmw_event_fence_action_seq_passed;
-	eaction->action.cleanup = vmw_event_fence_action_cleanup;
-	eaction->action.type = VMW_ACTION_EVENT;
+	struct timespec now;
 
-	eaction->fence = vmw_fence_obj_reference(fence);
-	eaction->dev = fman->dev_priv->dev;
-	eaction->tv_sec = tv_sec;
-	eaction->tv_usec = tv_usec;
-
-	spin_lock_irqsave(&fman->lock, irq_flags);
-	list_add_tail(&eaction->fpriv_head, &vmw_fp->fence_events);
-	spin_unlock_irqrestore(&fman->lock, irq_flags);
-
-	vmw_fence_obj_add_action(fence, &eaction->action);
-
-	return 0;
+	spin_lock(&queue->lock);
+	getrawmonotonic(&now);
+	queue->lag = vmw_timespec_add(queue->lag,
+				      timespec_sub(now, queue->lag_time));
+	queue->lag_time = now;
+	spin_unlock(&queue->lock);
+	return queue->lag;
 }
 
-struct vmw_event_fence_pending {
-	struct drm_pending_event base;
-	struct drm_vmw_event_fence event;
-};
 
-int vmw_event_fence_action_create(struct drm_file *file_priv,
-				  struct vmw_fence_obj *fence,
-				  uint32_t flags,
-				  uint64_t user_data,
-				  bool interruptible)
+static bool vmw_lag_lt(struct vmw_fence_queue *queue,
+		       uint32_t us)
 {
-	struct vmw_event_fence_pending *event;
-	struct drm_device *dev = fence->fman->dev_priv->dev;
-	unsigned long irq_flags;
-	int ret;
-
-	spin_lock_irqsave(&dev->event_lock, irq_flags);
-
-	ret = (file_priv->event_space < sizeof(event->event)) ? -EBUSY : 0;
-	if (likely(ret == 0))
-		file_priv->event_space -= sizeof(event->event);
-
-	spin_unlock_irqrestore(&dev->event_lock, irq_flags);
-
-	if (unlikely(ret != 0)) {
-		DRM_ERROR("Failed to allocate event space for this file.\n");
-		goto out_no_space;
-	}
-
+	struct timespec lag, cond;
 
-	event = kzalloc(sizeof(event->event), GFP_KERNEL);
-	if (unlikely(event == NULL)) {
-		DRM_ERROR("Failed to allocate an event.\n");
-		ret = -ENOMEM;
-		goto out_no_event;
-	}
-
-	event->event.base.type = DRM_VMW_EVENT_FENCE_SIGNALED;
-	event->event.base.length = sizeof(*event);
-	event->event.user_data = user_data;
-
-	event->base.event = &event->event.base;
-	event->base.file_priv = file_priv;
-	event->base.destroy = (void (*) (struct drm_pending_event *)) kfree;
-
-
-	if (flags & DRM_VMW_FE_FLAG_REQ_TIME)
-		ret = vmw_event_fence_action_queue(file_priv, fence,
-						   &event->base,
-						   &event->event.tv_sec,
-						   &event->event.tv_usec,
-						   interruptible);
-	else
-		ret = vmw_event_fence_action_queue(file_priv, fence,
-						   &event->base,
-						   NULL,
-						   NULL,
-						   interruptible);
-	if (ret != 0)
-		goto out_no_queue;
-
-out_no_queue:
-	event->base.destroy(&event->base);
-out_no_event:
-	spin_lock_irqsave(&dev->event_lock, irq_flags);
-	file_priv->event_space += sizeof(*event);
-	spin_unlock_irqrestore(&dev->event_lock, irq_flags);
-out_no_space:
-	return ret;
+	cond = ns_to_timespec((s64) us * 1000);
+	lag = vmw_fifo_lag(queue);
+	return (timespec_compare(&lag, &cond) < 1);
 }
 
-int vmw_fence_event_ioctl(struct drm_device *dev, void *data,
-			  struct drm_file *file_priv)
+int vmw_wait_lag(struct vmw_private *dev_priv,
+		 struct vmw_fence_queue *queue, uint32_t us)
 {
-	struct vmw_private *dev_priv = vmw_priv(dev);
-	struct drm_vmw_fence_event_arg *arg =
-		(struct drm_vmw_fence_event_arg *) data;
-	struct vmw_fence_obj *fence = NULL;
-	struct vmw_fpriv *vmw_fp = vmw_fpriv(file_priv);
-	struct drm_vmw_fence_rep __user *user_fence_rep =
-		(struct drm_vmw_fence_rep __user *)(unsigned long)
-		arg->fence_rep;
-	uint32_t handle;
+	struct vmw_fence *fence;
+	uint32_t sequence;
 	int ret;
 
-	/*
-	 * Look up an existing fence object,
-	 * and if user-space wants a new reference,
-	 * add one.
-	 */
-	if (arg->handle) {
-		struct ttm_base_object *base =
-			ttm_base_object_lookup(vmw_fp->tfile, arg->handle);
-
-		if (unlikely(base == NULL)) {
-			DRM_ERROR("Fence event invalid fence object handle "
-				  "0x%08lx.\n",
-				  (unsigned long)arg->handle);
-			return -EINVAL;
+	while (!vmw_lag_lt(queue, us)) {
+		spin_lock(&queue->lock);
+		if (list_empty(&queue->head))
+			sequence = atomic_read(&dev_priv->fence_seq);
+		else {
+			fence = list_first_entry(&queue->head,
+						 struct vmw_fence, head);
+			sequence = fence->sequence;
 		}
-		fence = &(container_of(base, struct vmw_user_fence,
-				       base)->fence);
-		(void) vmw_fence_obj_reference(fence);
+		spin_unlock(&queue->lock);
 
-		if (user_fence_rep != NULL) {
-			bool existed;
+		ret = vmw_wait_fence(dev_priv, false, sequence, true,
+				     3*HZ);
 
-			ret = ttm_ref_object_add(vmw_fp->tfile, base,
-						 TTM_REF_USAGE, &existed);
-			if (unlikely(ret != 0)) {
-				DRM_ERROR("Failed to reference a fence "
-					  "object.\n");
-				goto out_no_ref_obj;
-			}
-			handle = base->hash.key;
-		}
-		ttm_base_object_unref(&base);
-	}
-
-	/*
-	 * Create a new fence object.
-	 */
-	if (!fence) {
-		ret = vmw_execbuf_fence_commands(file_priv, dev_priv,
-						 &fence,
-						 (user_fence_rep) ?
-						 &handle : NULL);
-		if (unlikely(ret != 0)) {
-			DRM_ERROR("Fence event failed to create fence.\n");
+		if (unlikely(ret != 0))
 			return ret;
-		}
-	}
-
-	BUG_ON(fence == NULL);
 
-	if (arg->flags & DRM_VMW_FE_FLAG_REQ_TIME)
-		ret = vmw_event_fence_action_create(file_priv, fence,
-						    arg->flags,
-						    arg->user_data,
-						    true);
-	else
-		ret = vmw_event_fence_action_create(file_priv, fence,
-						    arg->flags,
-						    arg->user_data,
-						    true);
-
-	if (unlikely(ret != 0)) {
-		if (ret != -ERESTARTSYS)
-			DRM_ERROR("Failed to attach event to fence.\n");
-		goto out_no_create;
+		(void) vmw_fence_pull(queue, sequence);
 	}
-
-	vmw_execbuf_copy_fence_user(dev_priv, vmw_fp, 0, user_fence_rep, fence,
-				    handle);
-	vmw_fence_obj_unreference(&fence);
 	return 0;
-out_no_create:
-	if (user_fence_rep != NULL)
-		ttm_ref_object_base_unref(vmw_fpriv(file_priv)->tfile,
-					  handle, TTM_REF_USAGE);
-out_no_ref_obj:
-	vmw_fence_obj_unreference(&fence);
-	return ret;
 }
+
+
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.h b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.h
deleted file mode 100644
index faf2e78..0000000
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.h
+++ /dev/null
@@ -1,120 +0,0 @@
-/**************************************************************************
- *
- * Copyright © 2011 VMware, Inc., Palo Alto, CA., USA
- * All Rights Reserved.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the
- * "Software"), to deal in the Software without restriction, including
- * without limitation the rights to use, copy, modify, merge, publish,
- * distribute, sub license, and/or sell copies of the Software, and to
- * permit persons to whom the Software is furnished to do so, subject to
- * the following conditions:
- *
- * The above copyright notice and this permission notice (including the
- * next paragraph) shall be included in all copies or substantial portions
- * of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
- * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
- * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
- * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
- * USE OR OTHER DEALINGS IN THE SOFTWARE.
- *
- **************************************************************************/
-
-#ifndef _VMWGFX_FENCE_H_
-
-#define VMW_FENCE_WAIT_TIMEOUT (5*HZ)
-
-struct vmw_private;
-
-struct vmw_fence_manager;
-
-/**
- *
- *
- */
-enum vmw_action_type {
-	VMW_ACTION_EVENT = 0,
-	VMW_ACTION_MAX
-};
-
-struct vmw_fence_action {
-	struct list_head head;
-	enum vmw_action_type type;
-	void (*seq_passed) (struct vmw_fence_action *action);
-	void (*cleanup) (struct vmw_fence_action *action);
-};
-
-struct vmw_fence_obj {
-	struct kref kref;
-	u32 seqno;
-
-	struct vmw_fence_manager *fman;
-	struct list_head head;
-	uint32_t signaled;
-	uint32_t signal_mask;
-	struct list_head seq_passed_actions;
-	void (*destroy)(struct vmw_fence_obj *fence);
-	wait_queue_head_t queue;
-};
-
-extern struct vmw_fence_manager *
-vmw_fence_manager_init(struct vmw_private *dev_priv);
-
-extern void vmw_fence_manager_takedown(struct vmw_fence_manager *fman);
-
-extern void vmw_fence_obj_unreference(struct vmw_fence_obj **fence_p);
-
-extern struct vmw_fence_obj *
-vmw_fence_obj_reference(struct vmw_fence_obj *fence);
-
-extern void vmw_fences_update(struct vmw_fence_manager *fman);
-
-extern bool vmw_fence_obj_signaled(struct vmw_fence_obj *fence,
-				   uint32_t flags);
-
-extern int vmw_fence_obj_wait(struct vmw_fence_obj *fence, uint32_t flags,
-			      bool lazy,
-			      bool interruptible, unsigned long timeout);
-
-extern void vmw_fence_obj_flush(struct vmw_fence_obj *fence);
-
-extern int vmw_fence_create(struct vmw_fence_manager *fman,
-			    uint32_t seqno,
-			    uint32_t mask,
-			    struct vmw_fence_obj **p_fence);
-
-extern int vmw_user_fence_create(struct drm_file *file_priv,
-				 struct vmw_fence_manager *fman,
-				 uint32_t sequence,
-				 uint32_t mask,
-				 struct vmw_fence_obj **p_fence,
-				 uint32_t *p_handle);
-
-extern void vmw_fence_fifo_up(struct vmw_fence_manager *fman);
-
-extern void vmw_fence_fifo_down(struct vmw_fence_manager *fman);
-
-extern int vmw_fence_obj_wait_ioctl(struct drm_device *dev, void *data,
-				    struct drm_file *file_priv);
-
-extern int vmw_fence_obj_signaled_ioctl(struct drm_device *dev, void *data,
-					struct drm_file *file_priv);
-
-extern int vmw_fence_obj_unref_ioctl(struct drm_device *dev, void *data,
-				     struct drm_file *file_priv);
-extern int vmw_fence_event_ioctl(struct drm_device *dev, void *data,
-				 struct drm_file *file_priv);
-extern void vmw_event_fence_fpriv_gone(struct vmw_fence_manager *fman,
-				       struct list_head *event_list);
-extern int vmw_event_fence_action_queue(struct drm_file *filee_priv,
-					struct vmw_fence_obj *fence,
-					struct drm_pending_event *event,
-					uint32_t *tv_sec,
-					uint32_t *tv_usec,
-					bool interruptible);
-#endif /* _VMWGFX_FENCE_H_ */
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c
index a0c2f12..635c0ff 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c
@@ -33,7 +33,6 @@ bool vmw_fifo_have_3d(struct vmw_private *dev_priv)
 {
 	__le32 __iomem *fifo_mem = dev_priv->mmio_virt;
 	uint32_t fifo_min, hwversion;
-	const struct vmw_fifo_state *fifo = &dev_priv->fifo;
 
 	if (!(dev_priv->capabilities & SVGA_CAP_EXTENDED_FIFO))
 		return false;
@@ -42,20 +41,11 @@ bool vmw_fifo_have_3d(struct vmw_private *dev_priv)
 	if (fifo_min <= SVGA_FIFO_3D_HWVERSION * sizeof(unsigned int))
 		return false;
 
-	hwversion = ioread32(fifo_mem +
-			     ((fifo->capabilities &
-			       SVGA_FIFO_CAP_3D_HWVERSION_REVISED) ?
-			      SVGA_FIFO_3D_HWVERSION_REVISED :
-			      SVGA_FIFO_3D_HWVERSION));
-
+	hwversion = ioread32(fifo_mem + SVGA_FIFO_3D_HWVERSION);
 	if (hwversion == 0)
 		return false;
 
-	if (hwversion < SVGA3D_HWVERSION_WS8_B1)
-		return false;
-
-	/* Non-Screen Object path does not support surfaces */
-	if (!dev_priv->sou_priv)
+	if (hwversion < SVGA3D_HWVERSION_WS65_B1)
 		return false;
 
 	return true;
@@ -82,12 +72,22 @@ int vmw_fifo_init(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo)
 	uint32_t max;
 	uint32_t min;
 	uint32_t dummy;
+	int ret;
 
 	fifo->static_buffer_size = VMWGFX_FIFO_STATIC_SIZE;
 	fifo->static_buffer = vmalloc(fifo->static_buffer_size);
 	if (unlikely(fifo->static_buffer == NULL))
 		return -ENOMEM;
 
+	fifo->last_buffer_size = VMWGFX_FIFO_STATIC_SIZE;
+	fifo->last_data_size = 0;
+	fifo->last_buffer_add = false;
+	fifo->last_buffer = vmalloc(fifo->last_buffer_size);
+	if (unlikely(fifo->last_buffer == NULL)) {
+		ret = -ENOMEM;
+		goto out_err;
+	}
+
 	fifo->dynamic_buffer = NULL;
 	fifo->reserved_size = 0;
 	fifo->using_bounce_buffer = false;
@@ -137,10 +137,14 @@ int vmw_fifo_init(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo)
 		 (unsigned int) min,
 		 (unsigned int) fifo->capabilities);
 
-	atomic_set(&dev_priv->marker_seq, dev_priv->last_read_seqno);
-	iowrite32(dev_priv->last_read_seqno, fifo_mem + SVGA_FIFO_FENCE);
-	vmw_marker_queue_init(&fifo->marker_queue);
+	atomic_set(&dev_priv->fence_seq, dev_priv->last_read_sequence);
+	iowrite32(dev_priv->last_read_sequence, fifo_mem + SVGA_FIFO_FENCE);
+	vmw_fence_queue_init(&fifo->fence_queue);
 	return vmw_fifo_send_fence(dev_priv, &dummy);
+out_err:
+	vfree(fifo->static_buffer);
+	fifo->static_buffer = NULL;
+	return ret;
 }
 
 void vmw_fifo_ping_host(struct vmw_private *dev_priv, uint32_t reason)
@@ -166,7 +170,7 @@ void vmw_fifo_release(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo)
 	while (vmw_read(dev_priv, SVGA_REG_BUSY) != 0)
 		vmw_write(dev_priv, SVGA_REG_SYNC, SVGA_SYNC_GENERIC);
 
-	dev_priv->last_read_seqno = ioread32(fifo_mem + SVGA_FIFO_FENCE);
+	dev_priv->last_read_sequence = ioread32(fifo_mem + SVGA_FIFO_FENCE);
 
 	vmw_write(dev_priv, SVGA_REG_CONFIG_DONE,
 		  dev_priv->config_done_state);
@@ -176,7 +180,12 @@ void vmw_fifo_release(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo)
 		  dev_priv->traces_state);
 
 	mutex_unlock(&dev_priv->hw_mutex);
-	vmw_marker_queue_takedown(&fifo->marker_queue);
+	vmw_fence_queue_takedown(&fifo->fence_queue);
+
+	if (likely(fifo->last_buffer != NULL)) {
+		vfree(fifo->last_buffer);
+		fifo->last_buffer = NULL;
+	}
 
 	if (likely(fifo->static_buffer != NULL)) {
 		vfree(fifo->static_buffer);
@@ -253,8 +262,9 @@ static int vmw_fifo_wait(struct vmw_private *dev_priv,
 		spin_lock_irqsave(&dev_priv->irq_lock, irq_flags);
 		outl(SVGA_IRQFLAG_FIFO_PROGRESS,
 		     dev_priv->io_start + VMWGFX_IRQSTATUS_PORT);
-		dev_priv->irq_mask |= SVGA_IRQFLAG_FIFO_PROGRESS;
-		vmw_write(dev_priv, SVGA_REG_IRQMASK, dev_priv->irq_mask);
+		vmw_write(dev_priv, SVGA_REG_IRQMASK,
+			  vmw_read(dev_priv, SVGA_REG_IRQMASK) |
+			  SVGA_IRQFLAG_FIFO_PROGRESS);
 		spin_unlock_irqrestore(&dev_priv->irq_lock, irq_flags);
 	}
 	mutex_unlock(&dev_priv->hw_mutex);
@@ -276,8 +286,9 @@ static int vmw_fifo_wait(struct vmw_private *dev_priv,
 	mutex_lock(&dev_priv->hw_mutex);
 	if (atomic_dec_and_test(&dev_priv->fifo_queue_waiters)) {
 		spin_lock_irqsave(&dev_priv->irq_lock, irq_flags);
-		dev_priv->irq_mask &= ~SVGA_IRQFLAG_FIFO_PROGRESS;
-		vmw_write(dev_priv, SVGA_REG_IRQMASK, dev_priv->irq_mask);
+		vmw_write(dev_priv, SVGA_REG_IRQMASK,
+			  vmw_read(dev_priv, SVGA_REG_IRQMASK) &
+			  ~SVGA_IRQFLAG_FIFO_PROGRESS);
 		spin_unlock_irqrestore(&dev_priv->irq_lock, irq_flags);
 	}
 	mutex_unlock(&dev_priv->hw_mutex);
@@ -285,16 +296,6 @@ static int vmw_fifo_wait(struct vmw_private *dev_priv,
 	return ret;
 }
 
-/**
- * Reserve @bytes number of bytes in the fifo.
- *
- * This function will return NULL (error) on two conditions:
- *  If it timeouts waiting for fifo space, or if @bytes is larger than the
- *   available fifo space.
- *
- * Returns:
- *   Pointer to the fifo, or null on error (possible hardware hang).
- */
 void *vmw_fifo_reserve(struct vmw_private *dev_priv, uint32_t bytes)
 {
 	struct vmw_fifo_state *fifo_state = &dev_priv->fifo;
@@ -465,7 +466,7 @@ void vmw_fifo_commit(struct vmw_private *dev_priv, uint32_t bytes)
 	mutex_unlock(&fifo_state->fifo_mutex);
 }
 
-int vmw_fifo_send_fence(struct vmw_private *dev_priv, uint32_t *seqno)
+int vmw_fifo_send_fence(struct vmw_private *dev_priv, uint32_t *sequence)
 {
 	struct vmw_fifo_state *fifo_state = &dev_priv->fifo;
 	struct svga_fifo_cmd_fence *cmd_fence;
@@ -475,16 +476,16 @@ int vmw_fifo_send_fence(struct vmw_private *dev_priv, uint32_t *seqno)
 
 	fm = vmw_fifo_reserve(dev_priv, bytes);
 	if (unlikely(fm == NULL)) {
-		*seqno = atomic_read(&dev_priv->marker_seq);
+		*sequence = atomic_read(&dev_priv->fence_seq);
 		ret = -ENOMEM;
-		(void)vmw_fallback_wait(dev_priv, false, true, *seqno,
+		(void)vmw_fallback_wait(dev_priv, false, true, *sequence,
 					false, 3*HZ);
 		goto out_err;
 	}
 
 	do {
-		*seqno = atomic_add_return(1, &dev_priv->marker_seq);
-	} while (*seqno == 0);
+		*sequence = atomic_add_return(1, &dev_priv->fence_seq);
+	} while (*sequence == 0);
 
 	if (!(fifo_state->capabilities & SVGA_FIFO_CAP_FENCE)) {
 
@@ -501,68 +502,61 @@ int vmw_fifo_send_fence(struct vmw_private *dev_priv, uint32_t *seqno)
 	cmd_fence = (struct svga_fifo_cmd_fence *)
 	    ((unsigned long)fm + sizeof(__le32));
 
-	iowrite32(*seqno, &cmd_fence->fence);
+	iowrite32(*sequence, &cmd_fence->fence);
+	fifo_state->last_buffer_add = true;
 	vmw_fifo_commit(dev_priv, bytes);
-	(void) vmw_marker_push(&fifo_state->marker_queue, *seqno);
-	vmw_update_seqno(dev_priv, fifo_state);
+	fifo_state->last_buffer_add = false;
+	(void) vmw_fence_push(&fifo_state->fence_queue, *sequence);
+	vmw_update_sequence(dev_priv, fifo_state);
 
 out_err:
 	return ret;
 }
 
 /**
- * vmw_fifo_emit_dummy_query - emits a dummy query to the fifo.
- *
- * @dev_priv: The device private structure.
- * @cid: The hardware context id used for the query.
- *
- * This function is used to emit a dummy occlusion query with
- * no primitives rendered between query begin and query end.
- * It's used to provide a query barrier, in order to know that when
- * this query is finished, all preceding queries are also finished.
- *
- * A Query results structure should have been initialized at the start
- * of the dev_priv->dummy_query_bo buffer object. And that buffer object
- * must also be either reserved or pinned when this function is called.
- *
- * Returns -ENOMEM on failure to reserve fifo space.
+ * Map the first page of the FIFO read-only to user-space.
  */
-int vmw_fifo_emit_dummy_query(struct vmw_private *dev_priv,
-			      uint32_t cid)
-{
-	/*
-	 * A query wait without a preceding query end will
-	 * actually finish all queries for this cid
-	 * without writing to the query result structure.
-	 */
 
-	struct ttm_buffer_object *bo = dev_priv->dummy_query_bo;
-	struct {
-		SVGA3dCmdHeader header;
-		SVGA3dCmdWaitForQuery body;
-	} *cmd;
+static int vmw_fifo_vm_fault(struct vm_area_struct *vma, struct vm_fault *vmf)
+{
+	int ret;
+	unsigned long address = (unsigned long)vmf->virtual_address;
 
-	cmd = vmw_fifo_reserve(dev_priv, sizeof(*cmd));
+	if (address != vma->vm_start)
+		return VM_FAULT_SIGBUS;
 
-	if (unlikely(cmd == NULL)) {
-		DRM_ERROR("Out of fifo space for dummy query.\n");
-		return -ENOMEM;
-	}
+	ret = vm_insert_pfn(vma, address, vma->vm_pgoff);
+	if (likely(ret == -EBUSY || ret == 0))
+		return VM_FAULT_NOPAGE;
+	else if (ret == -ENOMEM)
+		return VM_FAULT_OOM;
 
-	cmd->header.id = SVGA_3D_CMD_WAIT_FOR_QUERY;
-	cmd->header.size = sizeof(cmd->body);
-	cmd->body.cid = cid;
-	cmd->body.type = SVGA3D_QUERYTYPE_OCCLUSION;
-
-	if (bo->mem.mem_type == TTM_PL_VRAM) {
-		cmd->body.guestResult.gmrId = SVGA_GMR_FRAMEBUFFER;
-		cmd->body.guestResult.offset = bo->offset;
-	} else {
-		cmd->body.guestResult.gmrId = bo->mem.start;
-		cmd->body.guestResult.offset = 0;
-	}
+	return VM_FAULT_SIGBUS;
+}
 
-	vmw_fifo_commit(dev_priv, sizeof(*cmd));
+static struct vm_operations_struct vmw_fifo_vm_ops = {
+	.fault = vmw_fifo_vm_fault,
+	.open = NULL,
+	.close = NULL
+};
 
+int vmw_fifo_mmap(struct file *filp, struct vm_area_struct *vma)
+{
+	struct drm_file *file_priv;
+	struct vmw_private *dev_priv;
+
+	file_priv = filp->private_data;
+	dev_priv = vmw_priv(file_priv->minor->dev);
+
+	if (vma->vm_pgoff != (dev_priv->mmio_start >> PAGE_SHIFT) ||
+	    (vma->vm_end - vma->vm_start) != PAGE_SIZE)
+		return -EINVAL;
+
+	vma->vm_flags &= ~(VM_WRITE | VM_MAYWRITE);
+	vma->vm_flags |= VM_IO | VM_PFNMAP | VM_DONTEXPAND | VM_SHARED;
+	vma->vm_page_prot = vm_get_page_prot(vma->vm_flags);
+	vma->vm_page_prot = ttm_io_prot(TTM_PL_FLAG_UNCACHED,
+					vma->vm_page_prot);
+	vma->vm_ops = &vmw_fifo_vm_ops;
 	return 0;
 }
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_gmr.c b/drivers/gpu/drm/vmwgfx/vmwgfx_gmr.c
index f4e7763..de0c594 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_gmr.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_gmr.c
@@ -1,6 +1,6 @@
 /**************************************************************************
  *
- * Copyright © 2009-2011 VMware, Inc., Palo Alto, CA., USA
+ * Copyright © 2009 VMware, Inc., Palo Alto, CA., USA
  * All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
@@ -29,77 +29,6 @@
 #include "drmP.h"
 #include "ttm/ttm_bo_driver.h"
 
-#define VMW_PPN_SIZE sizeof(unsigned long)
-
-static int vmw_gmr2_bind(struct vmw_private *dev_priv,
-			 struct page *pages[],
-			 unsigned long num_pages,
-			 int gmr_id)
-{
-	SVGAFifoCmdDefineGMR2 define_cmd;
-	SVGAFifoCmdRemapGMR2 remap_cmd;
-	uint32_t define_size = sizeof(define_cmd) + 4;
-	uint32_t remap_size = VMW_PPN_SIZE * num_pages + sizeof(remap_cmd) + 4;
-	uint32_t *cmd;
-	uint32_t *cmd_orig;
-	uint32_t i;
-
-	cmd_orig = cmd = vmw_fifo_reserve(dev_priv, define_size + remap_size);
-	if (unlikely(cmd == NULL))
-		return -ENOMEM;
-
-	define_cmd.gmrId = gmr_id;
-	define_cmd.numPages = num_pages;
-
-	remap_cmd.gmrId = gmr_id;
-	remap_cmd.flags = (VMW_PPN_SIZE > sizeof(*cmd)) ?
-		SVGA_REMAP_GMR2_PPN64 : SVGA_REMAP_GMR2_PPN32;
-	remap_cmd.offsetPages = 0;
-	remap_cmd.numPages = num_pages;
-
-	*cmd++ = SVGA_CMD_DEFINE_GMR2;
-	memcpy(cmd, &define_cmd, sizeof(define_cmd));
-	cmd += sizeof(define_cmd) / sizeof(uint32);
-
-	*cmd++ = SVGA_CMD_REMAP_GMR2;
-	memcpy(cmd, &remap_cmd, sizeof(remap_cmd));
-	cmd += sizeof(remap_cmd) / sizeof(uint32);
-
-	for (i = 0; i < num_pages; ++i) {
-		if (VMW_PPN_SIZE > 4)
-			*cmd = page_to_pfn(*pages++);
-		else
-			*((uint64_t *)cmd) = page_to_pfn(*pages++);
-
-		cmd += VMW_PPN_SIZE / sizeof(*cmd);
-	}
-
-	vmw_fifo_commit(dev_priv, define_size + remap_size);
-
-	return 0;
-}
-
-static void vmw_gmr2_unbind(struct vmw_private *dev_priv,
-			    int gmr_id)
-{
-	SVGAFifoCmdDefineGMR2 define_cmd;
-	uint32_t define_size = sizeof(define_cmd) + 4;
-	uint32_t *cmd;
-
-	cmd = vmw_fifo_reserve(dev_priv, define_size);
-	if (unlikely(cmd == NULL)) {
-		DRM_ERROR("GMR2 unbind failed.\n");
-		return;
-	}
-	define_cmd.gmrId = gmr_id;
-	define_cmd.numPages = 0;
-
-	*cmd++ = SVGA_CMD_DEFINE_GMR2;
-	memcpy(cmd, &define_cmd, sizeof(define_cmd));
-
-	vmw_fifo_commit(dev_priv, define_size);
-}
-
 /**
  * FIXME: Adjust to the ttm lowmem / highmem storage to minimize
  * the number of used descriptors.
@@ -241,9 +170,6 @@ int vmw_gmr_bind(struct vmw_private *dev_priv,
 	struct list_head desc_pages;
 	int ret;
 
-	if (likely(dev_priv->capabilities & SVGA_CAP_GMR2))
-		return vmw_gmr2_bind(dev_priv, pages, num_pages, gmr_id);
-
 	if (unlikely(!(dev_priv->capabilities & SVGA_CAP_GMR)))
 		return -EINVAL;
 
@@ -266,11 +192,6 @@ int vmw_gmr_bind(struct vmw_private *dev_priv,
 
 void vmw_gmr_unbind(struct vmw_private *dev_priv, int gmr_id)
 {
-	if (likely(dev_priv->capabilities & SVGA_CAP_GMR2)) {
-		vmw_gmr2_unbind(dev_priv, gmr_id);
-		return;
-	}
-
 	mutex_lock(&dev_priv->hw_mutex);
 	vmw_write(dev_priv, SVGA_REG_GMR_ID, gmr_id);
 	wmb();
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_gmrid_manager.c b/drivers/gpu/drm/vmwgfx/vmwgfx_gmrid_manager.c
index 5f71715..ac6e0d1 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_gmrid_manager.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_gmrid_manager.c
@@ -40,8 +40,6 @@ struct vmwgfx_gmrid_man {
 	spinlock_t lock;
 	struct ida gmr_ida;
 	uint32_t max_gmr_ids;
-	uint32_t max_gmr_pages;
-	uint32_t used_gmr_pages;
 };
 
 static int vmw_gmrid_man_get_node(struct ttm_mem_type_manager *man,
@@ -51,50 +49,33 @@ static int vmw_gmrid_man_get_node(struct ttm_mem_type_manager *man,
 {
 	struct vmwgfx_gmrid_man *gman =
 		(struct vmwgfx_gmrid_man *)man->priv;
-	int ret = 0;
+	int ret;
 	int id;
 
 	mem->mm_node = NULL;
 
-	spin_lock(&gman->lock);
-
-	if (gman->max_gmr_pages > 0) {
-		gman->used_gmr_pages += bo->num_pages;
-		if (unlikely(gman->used_gmr_pages > gman->max_gmr_pages))
-			goto out_err_locked;
-	}
-
 	do {
-		spin_unlock(&gman->lock);
-		if (unlikely(ida_pre_get(&gman->gmr_ida, GFP_KERNEL) == 0)) {
-			ret = -ENOMEM;
-			goto out_err;
-		}
-		spin_lock(&gman->lock);
+		if (unlikely(ida_pre_get(&gman->gmr_ida, GFP_KERNEL) == 0))
+			return -ENOMEM;
 
+		spin_lock(&gman->lock);
 		ret = ida_get_new(&gman->gmr_ida, &id);
+
 		if (unlikely(ret == 0 && id >= gman->max_gmr_ids)) {
 			ida_remove(&gman->gmr_ida, id);
-			ret = 0;
-			goto out_err_locked;
+			spin_unlock(&gman->lock);
+			return 0;
 		}
+
+		spin_unlock(&gman->lock);
+
 	} while (ret == -EAGAIN);
 
 	if (likely(ret == 0)) {
 		mem->mm_node = gman;
 		mem->start = id;
-		mem->num_pages = bo->num_pages;
-	} else
-		goto out_err_locked;
-
-	spin_unlock(&gman->lock);
-	return 0;
+	}
 
-out_err:
-	spin_lock(&gman->lock);
-out_err_locked:
-	gman->used_gmr_pages -= bo->num_pages;
-	spin_unlock(&gman->lock);
 	return ret;
 }
 
@@ -107,7 +88,6 @@ static void vmw_gmrid_man_put_node(struct ttm_mem_type_manager *man,
 	if (mem->mm_node) {
 		spin_lock(&gman->lock);
 		ida_remove(&gman->gmr_ida, mem->start);
-		gman->used_gmr_pages -= mem->num_pages;
 		spin_unlock(&gman->lock);
 		mem->mm_node = NULL;
 	}
@@ -116,8 +96,6 @@ static void vmw_gmrid_man_put_node(struct ttm_mem_type_manager *man,
 static int vmw_gmrid_man_init(struct ttm_mem_type_manager *man,
 			      unsigned long p_size)
 {
-	struct vmw_private *dev_priv =
-		container_of(man->bdev, struct vmw_private, bdev);
 	struct vmwgfx_gmrid_man *gman =
 		kzalloc(sizeof(*gman), GFP_KERNEL);
 
@@ -125,8 +103,6 @@ static int vmw_gmrid_man_init(struct ttm_mem_type_manager *man,
 		return -ENOMEM;
 
 	spin_lock_init(&gman->lock);
-	gman->max_gmr_pages = dev_priv->max_gmr_pages;
-	gman->used_gmr_pages = 0;
 	ida_init(&gman->gmr_ida);
 	gman->max_gmr_ids = p_size;
 	man->priv = (void *) gman;
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c
index 66917c6..570d577 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c
@@ -27,7 +27,6 @@
 
 #include "vmwgfx_drv.h"
 #include "vmwgfx_drm.h"
-#include "vmwgfx_kms.h"
 
 int vmw_getparam_ioctl(struct drm_device *dev, void *data,
 		       struct drm_file *file_priv)
@@ -46,6 +45,9 @@ int vmw_getparam_ioctl(struct drm_device *dev, void *data,
 	case DRM_VMW_PARAM_3D:
 		param->value = vmw_fifo_have_3d(dev_priv) ? 1 : 0;
 		break;
+	case DRM_VMW_PARAM_FIFO_OFFSET:
+		param->value = dev_priv->mmio_start;
+		break;
 	case DRM_VMW_PARAM_HW_CAPS:
 		param->value = dev_priv->capabilities;
 		break;
@@ -55,19 +57,6 @@ int vmw_getparam_ioctl(struct drm_device *dev, void *data,
 	case DRM_VMW_PARAM_MAX_FB_SIZE:
 		param->value = dev_priv->vram_size;
 		break;
-	case DRM_VMW_PARAM_FIFO_HW_VERSION:
-	{
-		__le32 __iomem *fifo_mem = dev_priv->mmio_virt;
-		const struct vmw_fifo_state *fifo = &dev_priv->fifo;
-
-		param->value =
-			ioread32(fifo_mem +
-				 ((fifo->capabilities &
-				   SVGA_FIFO_CAP_3D_HWVERSION_REVISED) ?
-				  SVGA_FIFO_3D_HWVERSION_REVISED :
-				  SVGA_FIFO_3D_HWVERSION));
-		break;
-	}
 	default:
 		DRM_ERROR("Illegal vmwgfx get param request: %d\n",
 			  param->param);
@@ -77,253 +66,25 @@ int vmw_getparam_ioctl(struct drm_device *dev, void *data,
 	return 0;
 }
 
-
-int vmw_get_cap_3d_ioctl(struct drm_device *dev, void *data,
+int vmw_fifo_debug_ioctl(struct drm_device *dev, void *data,
 			 struct drm_file *file_priv)
 {
-	struct drm_vmw_get_3d_cap_arg *arg =
-		(struct drm_vmw_get_3d_cap_arg *) data;
 	struct vmw_private *dev_priv = vmw_priv(dev);
-	uint32_t size;
-	__le32 __iomem *fifo_mem;
-	void __user *buffer = (void __user *)((unsigned long)(arg->buffer));
-	void *bounce;
-	int ret;
+	struct vmw_fifo_state *fifo_state = &dev_priv->fifo;
+	struct drm_vmw_fifo_debug_arg *arg =
+	    (struct drm_vmw_fifo_debug_arg *)data;
+	__le32 __user *buffer = (__le32 __user *)
+	    (unsigned long)arg->debug_buffer;
 
-	if (unlikely(arg->pad64 != 0)) {
-		DRM_ERROR("Illegal GET_3D_CAP argument.\n");
+	if (unlikely(fifo_state->last_buffer == NULL))
 		return -EINVAL;
-	}
-
-	size = (SVGA_FIFO_3D_CAPS_LAST - SVGA_FIFO_3D_CAPS + 1) << 2;
-
-	if (arg->max_size < size)
-		size = arg->max_size;
-
-	bounce = vmalloc(size);
-	if (unlikely(bounce == NULL)) {
-		DRM_ERROR("Failed to allocate bounce buffer for 3D caps.\n");
-		return -ENOMEM;
-	}
-
-	fifo_mem = dev_priv->mmio_virt;
-	memcpy_fromio(bounce, &fifo_mem[SVGA_FIFO_3D_CAPS], size);
-
-	ret = copy_to_user(buffer, bounce, size);
-	vfree(bounce);
-
-	if (unlikely(ret != 0))
-		DRM_ERROR("Failed to report 3D caps info.\n");
-
-	return ret;
-}
-
-int vmw_present_ioctl(struct drm_device *dev, void *data,
-		      struct drm_file *file_priv)
-{
-	struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile;
-	struct vmw_private *dev_priv = vmw_priv(dev);
-	struct drm_vmw_present_arg *arg =
-		(struct drm_vmw_present_arg *)data;
-	struct vmw_surface *surface;
-	struct vmw_master *vmaster = vmw_master(file_priv->master);
-	struct drm_vmw_rect __user *clips_ptr;
-	struct drm_vmw_rect *clips = NULL;
-	struct drm_mode_object *obj;
-	struct vmw_framebuffer *vfb;
-	uint32_t num_clips;
-	int ret;
-
-	num_clips = arg->num_clips;
-	clips_ptr = (struct drm_vmw_rect *)(unsigned long)arg->clips_ptr;
-
-	if (unlikely(num_clips == 0))
-		return 0;
-
-	if (clips_ptr == NULL) {
-		DRM_ERROR("Variable clips_ptr must be specified.\n");
-		ret = -EINVAL;
-		goto out_clips;
-	}
-
-	clips = kcalloc(num_clips, sizeof(*clips), GFP_KERNEL);
-	if (clips == NULL) {
-		DRM_ERROR("Failed to allocate clip rect list.\n");
-		ret = -ENOMEM;
-		goto out_clips;
-	}
-
-	ret = copy_from_user(clips, clips_ptr, num_clips * sizeof(*clips));
-	if (ret) {
-		DRM_ERROR("Failed to copy clip rects from userspace.\n");
-		ret = -EFAULT;
-		goto out_no_copy;
-	}
-
-	ret = mutex_lock_interruptible(&dev->mode_config.mutex);
-	if (unlikely(ret != 0)) {
-		ret = -ERESTARTSYS;
-		goto out_no_mode_mutex;
-	}
-
-	obj = drm_mode_object_find(dev, arg->fb_id, DRM_MODE_OBJECT_FB);
-	if (!obj) {
-		DRM_ERROR("Invalid framebuffer id.\n");
-		ret = -EINVAL;
-		goto out_no_fb;
-	}
-	vfb = vmw_framebuffer_to_vfb(obj_to_fb(obj));
-
-	ret = ttm_read_lock(&vmaster->lock, true);
-	if (unlikely(ret != 0))
-		goto out_no_ttm_lock;
-
-	ret = vmw_user_surface_lookup_handle(dev_priv, tfile, arg->sid,
-					     &surface);
-	if (ret)
-		goto out_no_surface;
-
-	ret = vmw_kms_present(dev_priv, file_priv,
-			      vfb, surface, arg->sid,
-			      arg->dest_x, arg->dest_y,
-			      clips, num_clips);
-
-	/* vmw_user_surface_lookup takes one ref so does new_fb */
-	vmw_surface_unreference(&surface);
 
-out_no_surface:
-	ttm_read_unlock(&vmaster->lock);
-out_no_ttm_lock:
-out_no_fb:
-	mutex_unlock(&dev->mode_config.mutex);
-out_no_mode_mutex:
-out_no_copy:
-	kfree(clips);
-out_clips:
-	return ret;
-}
-
-int vmw_present_readback_ioctl(struct drm_device *dev, void *data,
-			       struct drm_file *file_priv)
-{
-	struct vmw_private *dev_priv = vmw_priv(dev);
-	struct drm_vmw_present_readback_arg *arg =
-		(struct drm_vmw_present_readback_arg *)data;
-	struct drm_vmw_fence_rep __user *user_fence_rep =
-		(struct drm_vmw_fence_rep __user *)
-		(unsigned long)arg->fence_rep;
-	struct vmw_master *vmaster = vmw_master(file_priv->master);
-	struct drm_vmw_rect __user *clips_ptr;
-	struct drm_vmw_rect *clips = NULL;
-	struct drm_mode_object *obj;
-	struct vmw_framebuffer *vfb;
-	uint32_t num_clips;
-	int ret;
-
-	num_clips = arg->num_clips;
-	clips_ptr = (struct drm_vmw_rect *)(unsigned long)arg->clips_ptr;
-
-	if (unlikely(num_clips == 0))
-		return 0;
-
-	if (clips_ptr == NULL) {
-		DRM_ERROR("Argument clips_ptr must be specified.\n");
-		ret = -EINVAL;
-		goto out_clips;
-	}
-
-	clips = kcalloc(num_clips, sizeof(*clips), GFP_KERNEL);
-	if (clips == NULL) {
-		DRM_ERROR("Failed to allocate clip rect list.\n");
-		ret = -ENOMEM;
-		goto out_clips;
-	}
-
-	ret = copy_from_user(clips, clips_ptr, num_clips * sizeof(*clips));
-	if (ret) {
-		DRM_ERROR("Failed to copy clip rects from userspace.\n");
-		ret = -EFAULT;
-		goto out_no_copy;
-	}
-
-	ret = mutex_lock_interruptible(&dev->mode_config.mutex);
-	if (unlikely(ret != 0)) {
-		ret = -ERESTARTSYS;
-		goto out_no_mode_mutex;
-	}
-
-	obj = drm_mode_object_find(dev, arg->fb_id, DRM_MODE_OBJECT_FB);
-	if (!obj) {
-		DRM_ERROR("Invalid framebuffer id.\n");
-		ret = -EINVAL;
-		goto out_no_fb;
-	}
-
-	vfb = vmw_framebuffer_to_vfb(obj_to_fb(obj));
-	if (!vfb->dmabuf) {
-		DRM_ERROR("Framebuffer not dmabuf backed.\n");
-		ret = -EINVAL;
-		goto out_no_fb;
+	if (arg->debug_buffer_size < fifo_state->last_data_size) {
+		arg->used_size = arg->debug_buffer_size;
+		arg->did_not_fit = 1;
+	} else {
+		arg->used_size = fifo_state->last_data_size;
+		arg->did_not_fit = 0;
 	}
-
-	ret = ttm_read_lock(&vmaster->lock, true);
-	if (unlikely(ret != 0))
-		goto out_no_ttm_lock;
-
-	ret = vmw_kms_readback(dev_priv, file_priv,
-			       vfb, user_fence_rep,
-			       clips, num_clips);
-
-	ttm_read_unlock(&vmaster->lock);
-out_no_ttm_lock:
-out_no_fb:
-	mutex_unlock(&dev->mode_config.mutex);
-out_no_mode_mutex:
-out_no_copy:
-	kfree(clips);
-out_clips:
-	return ret;
-}
-
-
-/**
- * vmw_fops_poll - wrapper around the drm_poll function
- *
- * @filp: See the linux fops poll documentation.
- * @wait: See the linux fops poll documentation.
- *
- * Wrapper around the drm_poll function that makes sure the device is
- * processing the fifo if drm_poll decides to wait.
- */
-unsigned int vmw_fops_poll(struct file *filp, struct poll_table_struct *wait)
-{
-	struct drm_file *file_priv = filp->private_data;
-	struct vmw_private *dev_priv =
-		vmw_priv(file_priv->minor->dev);
-
-	vmw_fifo_ping_host(dev_priv, SVGA_SYNC_GENERIC);
-	return drm_poll(filp, wait);
-}
-
-
-/**
- * vmw_fops_read - wrapper around the drm_read function
- *
- * @filp: See the linux fops read documentation.
- * @buffer: See the linux fops read documentation.
- * @count: See the linux fops read documentation.
- * offset: See the linux fops read documentation.
- *
- * Wrapper around the drm_read function that makes sure the device is
- * processing the fifo if drm_read decides to wait.
- */
-ssize_t vmw_fops_read(struct file *filp, char __user *buffer,
-		      size_t count, loff_t *offset)
-{
-	struct drm_file *file_priv = filp->private_data;
-	struct vmw_private *dev_priv =
-		vmw_priv(file_priv->minor->dev);
-
-	vmw_fifo_ping_host(dev_priv, SVGA_SYNC_GENERIC);
-	return drm_read(filp, buffer, count, offset);
+	return copy_to_user(buffer, fifo_state->last_buffer, arg->used_size);
 }
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c b/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c
index cabc95f..e92298a 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c
@@ -34,33 +34,26 @@ irqreturn_t vmw_irq_handler(DRM_IRQ_ARGS)
 {
 	struct drm_device *dev = (struct drm_device *)arg;
 	struct vmw_private *dev_priv = vmw_priv(dev);
-	uint32_t status, masked_status;
+	uint32_t status;
 
 	spin_lock(&dev_priv->irq_lock);
 	status = inl(dev_priv->io_start + VMWGFX_IRQSTATUS_PORT);
-	masked_status = status & dev_priv->irq_mask;
 	spin_unlock(&dev_priv->irq_lock);
 
-	if (likely(status))
-		outl(status, dev_priv->io_start + VMWGFX_IRQSTATUS_PORT);
-
-	if (!masked_status)
-		return IRQ_NONE;
-
-	if (masked_status & (SVGA_IRQFLAG_ANY_FENCE |
-			     SVGA_IRQFLAG_FENCE_GOAL)) {
-		vmw_fences_update(dev_priv->fman);
+	if (status & SVGA_IRQFLAG_ANY_FENCE)
 		wake_up_all(&dev_priv->fence_queue);
-	}
-
-	if (masked_status & SVGA_IRQFLAG_FIFO_PROGRESS)
+	if (status & SVGA_IRQFLAG_FIFO_PROGRESS)
 		wake_up_all(&dev_priv->fifo_queue);
 
+	if (likely(status)) {
+		outl(status, dev_priv->io_start + VMWGFX_IRQSTATUS_PORT);
+		return IRQ_HANDLED;
+	}
 
-	return IRQ_HANDLED;
+	return IRQ_NONE;
 }
 
-static bool vmw_fifo_idle(struct vmw_private *dev_priv, uint32_t seqno)
+static bool vmw_fifo_idle(struct vmw_private *dev_priv, uint32_t sequence)
 {
 	uint32_t busy;
 
@@ -71,43 +64,43 @@ static bool vmw_fifo_idle(struct vmw_private *dev_priv, uint32_t seqno)
 	return (busy == 0);
 }
 
-void vmw_update_seqno(struct vmw_private *dev_priv,
+void vmw_update_sequence(struct vmw_private *dev_priv,
 			 struct vmw_fifo_state *fifo_state)
 {
 	__le32 __iomem *fifo_mem = dev_priv->mmio_virt;
-	uint32_t seqno = ioread32(fifo_mem + SVGA_FIFO_FENCE);
 
-	if (dev_priv->last_read_seqno != seqno) {
-		dev_priv->last_read_seqno = seqno;
-		vmw_marker_pull(&fifo_state->marker_queue, seqno);
-		vmw_fences_update(dev_priv->fman);
+	uint32_t sequence = ioread32(fifo_mem + SVGA_FIFO_FENCE);
+
+	if (dev_priv->last_read_sequence != sequence) {
+		dev_priv->last_read_sequence = sequence;
+		vmw_fence_pull(&fifo_state->fence_queue, sequence);
 	}
 }
 
-bool vmw_seqno_passed(struct vmw_private *dev_priv,
-			 uint32_t seqno)
+bool vmw_fence_signaled(struct vmw_private *dev_priv,
+			uint32_t sequence)
 {
 	struct vmw_fifo_state *fifo_state;
 	bool ret;
 
-	if (likely(dev_priv->last_read_seqno - seqno < VMW_FENCE_WRAP))
+	if (likely(dev_priv->last_read_sequence - sequence < VMW_FENCE_WRAP))
 		return true;
 
 	fifo_state = &dev_priv->fifo;
-	vmw_update_seqno(dev_priv, fifo_state);
-	if (likely(dev_priv->last_read_seqno - seqno < VMW_FENCE_WRAP))
+	vmw_update_sequence(dev_priv, fifo_state);
+	if (likely(dev_priv->last_read_sequence - sequence < VMW_FENCE_WRAP))
 		return true;
 
 	if (!(fifo_state->capabilities & SVGA_FIFO_CAP_FENCE) &&
-	    vmw_fifo_idle(dev_priv, seqno))
+	    vmw_fifo_idle(dev_priv, sequence))
 		return true;
 
 	/**
-	 * Then check if the seqno is higher than what we've actually
+	 * Then check if the sequence is higher than what we've actually
 	 * emitted. Then the fence is stale and signaled.
 	 */
 
-	ret = ((atomic_read(&dev_priv->marker_seq) - seqno)
+	ret = ((atomic_read(&dev_priv->fence_seq) - sequence)
 	       > VMW_FENCE_WRAP);
 
 	return ret;
@@ -116,7 +109,7 @@ bool vmw_seqno_passed(struct vmw_private *dev_priv,
 int vmw_fallback_wait(struct vmw_private *dev_priv,
 		      bool lazy,
 		      bool fifo_idle,
-		      uint32_t seqno,
+		      uint32_t sequence,
 		      bool interruptible,
 		      unsigned long timeout)
 {
@@ -130,7 +123,7 @@ int vmw_fallback_wait(struct vmw_private *dev_priv,
 	DEFINE_WAIT(__wait);
 
 	wait_condition = (fifo_idle) ? &vmw_fifo_idle :
-		&vmw_seqno_passed;
+		&vmw_fence_signaled;
 
 	/**
 	 * Block command submission while waiting for idle.
@@ -138,14 +131,14 @@ int vmw_fallback_wait(struct vmw_private *dev_priv,
 
 	if (fifo_idle)
 		down_read(&fifo_state->rwsem);
-	signal_seq = atomic_read(&dev_priv->marker_seq);
+	signal_seq = atomic_read(&dev_priv->fence_seq);
 	ret = 0;
 
 	for (;;) {
 		prepare_to_wait(&dev_priv->fence_queue, &__wait,
 				(interruptible) ?
 				TASK_INTERRUPTIBLE : TASK_UNINTERRUPTIBLE);
-		if (wait_condition(dev_priv, seqno))
+		if (wait_condition(dev_priv, sequence))
 			break;
 		if (time_after_eq(jiffies, end_jiffies)) {
 			DRM_ERROR("SVGA device lockup.\n");
@@ -182,110 +175,68 @@ int vmw_fallback_wait(struct vmw_private *dev_priv,
 	return ret;
 }
 
-void vmw_seqno_waiter_add(struct vmw_private *dev_priv)
-{
-	mutex_lock(&dev_priv->hw_mutex);
-	if (dev_priv->fence_queue_waiters++ == 0) {
-		unsigned long irq_flags;
-
-		spin_lock_irqsave(&dev_priv->irq_lock, irq_flags);
-		outl(SVGA_IRQFLAG_ANY_FENCE,
-		     dev_priv->io_start + VMWGFX_IRQSTATUS_PORT);
-		dev_priv->irq_mask |= SVGA_IRQFLAG_ANY_FENCE;
-		vmw_write(dev_priv, SVGA_REG_IRQMASK, dev_priv->irq_mask);
-		spin_unlock_irqrestore(&dev_priv->irq_lock, irq_flags);
-	}
-	mutex_unlock(&dev_priv->hw_mutex);
-}
-
-void vmw_seqno_waiter_remove(struct vmw_private *dev_priv)
-{
-	mutex_lock(&dev_priv->hw_mutex);
-	if (--dev_priv->fence_queue_waiters == 0) {
-		unsigned long irq_flags;
-
-		spin_lock_irqsave(&dev_priv->irq_lock, irq_flags);
-		dev_priv->irq_mask &= ~SVGA_IRQFLAG_ANY_FENCE;
-		vmw_write(dev_priv, SVGA_REG_IRQMASK, dev_priv->irq_mask);
-		spin_unlock_irqrestore(&dev_priv->irq_lock, irq_flags);
-	}
-	mutex_unlock(&dev_priv->hw_mutex);
-}
-
-
-void vmw_goal_waiter_add(struct vmw_private *dev_priv)
-{
-	mutex_lock(&dev_priv->hw_mutex);
-	if (dev_priv->goal_queue_waiters++ == 0) {
-		unsigned long irq_flags;
-
-		spin_lock_irqsave(&dev_priv->irq_lock, irq_flags);
-		outl(SVGA_IRQFLAG_FENCE_GOAL,
-		     dev_priv->io_start + VMWGFX_IRQSTATUS_PORT);
-		dev_priv->irq_mask |= SVGA_IRQFLAG_FENCE_GOAL;
-		vmw_write(dev_priv, SVGA_REG_IRQMASK, dev_priv->irq_mask);
-		spin_unlock_irqrestore(&dev_priv->irq_lock, irq_flags);
-	}
-	mutex_unlock(&dev_priv->hw_mutex);
-}
-
-void vmw_goal_waiter_remove(struct vmw_private *dev_priv)
-{
-	mutex_lock(&dev_priv->hw_mutex);
-	if (--dev_priv->goal_queue_waiters == 0) {
-		unsigned long irq_flags;
-
-		spin_lock_irqsave(&dev_priv->irq_lock, irq_flags);
-		dev_priv->irq_mask &= ~SVGA_IRQFLAG_FENCE_GOAL;
-		vmw_write(dev_priv, SVGA_REG_IRQMASK, dev_priv->irq_mask);
-		spin_unlock_irqrestore(&dev_priv->irq_lock, irq_flags);
-	}
-	mutex_unlock(&dev_priv->hw_mutex);
-}
-
-int vmw_wait_seqno(struct vmw_private *dev_priv,
-		      bool lazy, uint32_t seqno,
-		      bool interruptible, unsigned long timeout)
+int vmw_wait_fence(struct vmw_private *dev_priv,
+		   bool lazy, uint32_t sequence,
+		   bool interruptible, unsigned long timeout)
 {
 	long ret;
+	unsigned long irq_flags;
 	struct vmw_fifo_state *fifo = &dev_priv->fifo;
 
-	if (likely(dev_priv->last_read_seqno - seqno < VMW_FENCE_WRAP))
+	if (likely(dev_priv->last_read_sequence - sequence < VMW_FENCE_WRAP))
 		return 0;
 
-	if (likely(vmw_seqno_passed(dev_priv, seqno)))
+	if (likely(vmw_fence_signaled(dev_priv, sequence)))
 		return 0;
 
 	vmw_fifo_ping_host(dev_priv, SVGA_SYNC_GENERIC);
 
 	if (!(fifo->capabilities & SVGA_FIFO_CAP_FENCE))
-		return vmw_fallback_wait(dev_priv, lazy, true, seqno,
+		return vmw_fallback_wait(dev_priv, lazy, true, sequence,
 					 interruptible, timeout);
 
 	if (!(dev_priv->capabilities & SVGA_CAP_IRQMASK))
-		return vmw_fallback_wait(dev_priv, lazy, false, seqno,
+		return vmw_fallback_wait(dev_priv, lazy, false, sequence,
 					 interruptible, timeout);
 
-	vmw_seqno_waiter_add(dev_priv);
+	mutex_lock(&dev_priv->hw_mutex);
+	if (atomic_add_return(1, &dev_priv->fence_queue_waiters) > 0) {
+		spin_lock_irqsave(&dev_priv->irq_lock, irq_flags);
+		outl(SVGA_IRQFLAG_ANY_FENCE,
+		     dev_priv->io_start + VMWGFX_IRQSTATUS_PORT);
+		vmw_write(dev_priv, SVGA_REG_IRQMASK,
+			  vmw_read(dev_priv, SVGA_REG_IRQMASK) |
+			  SVGA_IRQFLAG_ANY_FENCE);
+		spin_unlock_irqrestore(&dev_priv->irq_lock, irq_flags);
+	}
+	mutex_unlock(&dev_priv->hw_mutex);
 
 	if (interruptible)
 		ret = wait_event_interruptible_timeout
 		    (dev_priv->fence_queue,
-		     vmw_seqno_passed(dev_priv, seqno),
+		     vmw_fence_signaled(dev_priv, sequence),
 		     timeout);
 	else
 		ret = wait_event_timeout
 		    (dev_priv->fence_queue,
-		     vmw_seqno_passed(dev_priv, seqno),
+		     vmw_fence_signaled(dev_priv, sequence),
 		     timeout);
 
-	vmw_seqno_waiter_remove(dev_priv);
-
 	if (unlikely(ret == 0))
 		ret = -EBUSY;
 	else if (likely(ret > 0))
 		ret = 0;
 
+	mutex_lock(&dev_priv->hw_mutex);
+	if (atomic_dec_and_test(&dev_priv->fence_queue_waiters)) {
+		spin_lock_irqsave(&dev_priv->irq_lock, irq_flags);
+		vmw_write(dev_priv, SVGA_REG_IRQMASK,
+			  vmw_read(dev_priv, SVGA_REG_IRQMASK) &
+			  ~SVGA_IRQFLAG_ANY_FENCE);
+		spin_unlock_irqrestore(&dev_priv->irq_lock, irq_flags);
+	}
+	mutex_unlock(&dev_priv->hw_mutex);
+
 	return ret;
 }
 
@@ -322,3 +273,25 @@ void vmw_irq_uninstall(struct drm_device *dev)
 	status = inl(dev_priv->io_start + VMWGFX_IRQSTATUS_PORT);
 	outl(status, dev_priv->io_start + VMWGFX_IRQSTATUS_PORT);
 }
+
+#define VMW_FENCE_WAIT_TIMEOUT 3*HZ;
+
+int vmw_fence_wait_ioctl(struct drm_device *dev, void *data,
+			 struct drm_file *file_priv)
+{
+	struct drm_vmw_fence_wait_arg *arg =
+	    (struct drm_vmw_fence_wait_arg *)data;
+	unsigned long timeout;
+
+	if (!arg->cookie_valid) {
+		arg->cookie_valid = 1;
+		arg->kernel_cookie = jiffies + VMW_FENCE_WAIT_TIMEOUT;
+	}
+
+	timeout = jiffies;
+	if (time_after_eq(timeout, (unsigned long)arg->kernel_cookie))
+		return -EBUSY;
+
+	timeout = (unsigned long)arg->kernel_cookie - timeout;
+	return vmw_wait_fence(vmw_priv(dev), true, arg->sequence, true, timeout);
+}
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
index 2286d47..8a38c91 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.c
@@ -27,47 +27,11 @@
 
 #include "vmwgfx_kms.h"
 
-
 /* Might need a hrtimer here? */
 #define VMWGFX_PRESENT_RATE ((HZ / 60 > 0) ? HZ / 60 : 1)
 
-
-struct vmw_clip_rect {
-	int x1, x2, y1, y2;
-};
-
-/**
- * Clip @num_rects number of @rects against @clip storing the
- * results in @out_rects and the number of passed rects in @out_num.
- */
-void vmw_clip_cliprects(struct drm_clip_rect *rects,
-			int num_rects,
-			struct vmw_clip_rect clip,
-			SVGASignedRect *out_rects,
-			int *out_num)
-{
-	int i, k;
-
-	for (i = 0, k = 0; i < num_rects; i++) {
-		int x1 = max_t(int, clip.x1, rects[i].x1);
-		int y1 = max_t(int, clip.y1, rects[i].y1);
-		int x2 = min_t(int, clip.x2, rects[i].x2);
-		int y2 = min_t(int, clip.y2, rects[i].y2);
-
-		if (x1 >= x2)
-			continue;
-		if (y1 >= y2)
-			continue;
-
-		out_rects[k].left   = x1;
-		out_rects[k].top    = y1;
-		out_rects[k].right  = x2;
-		out_rects[k].bottom = y2;
-		k++;
-	}
-
-	*out_num = k;
-}
+static int vmw_surface_dmabuf_pin(struct vmw_framebuffer *vfb);
+static int vmw_surface_dmabuf_unpin(struct vmw_framebuffer *vfb);
 
 void vmw_display_unit_cleanup(struct vmw_display_unit *du)
 {
@@ -120,43 +84,6 @@ int vmw_cursor_update_image(struct vmw_private *dev_priv,
 	return 0;
 }
 
-int vmw_cursor_update_dmabuf(struct vmw_private *dev_priv,
-			     struct vmw_dma_buffer *dmabuf,
-			     u32 width, u32 height,
-			     u32 hotspotX, u32 hotspotY)
-{
-	struct ttm_bo_kmap_obj map;
-	unsigned long kmap_offset;
-	unsigned long kmap_num;
-	void *virtual;
-	bool dummy;
-	int ret;
-
-	kmap_offset = 0;
-	kmap_num = (width*height*4 + PAGE_SIZE - 1) >> PAGE_SHIFT;
-
-	ret = ttm_bo_reserve(&dmabuf->base, true, false, false, 0);
-	if (unlikely(ret != 0)) {
-		DRM_ERROR("reserve failed\n");
-		return -EINVAL;
-	}
-
-	ret = ttm_bo_kmap(&dmabuf->base, kmap_offset, kmap_num, &map);
-	if (unlikely(ret != 0))
-		goto err_unreserve;
-
-	virtual = ttm_kmap_obj_virtual(&map, &dummy);
-	ret = vmw_cursor_update_image(dev_priv, virtual, width, height,
-				      hotspotX, hotspotY);
-
-	ttm_bo_kunmap(&map);
-err_unreserve:
-	ttm_bo_unreserve(&dmabuf->base);
-
-	return ret;
-}
-
-
 void vmw_cursor_update_position(struct vmw_private *dev_priv,
 				bool show, int x, int y)
 {
@@ -180,26 +107,24 @@ int vmw_du_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv,
 	struct vmw_dma_buffer *dmabuf = NULL;
 	int ret;
 
-	/* A lot of the code assumes this */
-	if (handle && (width != 64 || height != 64))
-		return -EINVAL;
-
 	if (handle) {
-		ret = vmw_user_lookup_handle(dev_priv, tfile,
-					     handle, &surface, &dmabuf);
-		if (ret) {
-			DRM_ERROR("failed to find surface or dmabuf: %i\n", ret);
-			return -EINVAL;
+		ret = vmw_user_surface_lookup_handle(dev_priv, tfile,
+						     handle, &surface);
+		if (!ret) {
+			if (!surface->snooper.image) {
+				DRM_ERROR("surface not suitable for cursor\n");
+				return -EINVAL;
+			}
+		} else {
+			ret = vmw_user_dmabuf_lookup(tfile,
+						     handle, &dmabuf);
+			if (ret) {
+				DRM_ERROR("failed to find surface or dmabuf: %i\n", ret);
+				return -EINVAL;
+			}
 		}
 	}
 
-	/* need to do this before taking down old image */
-	if (surface && !surface->snooper.image) {
-		DRM_ERROR("surface not suitable for cursor\n");
-		vmw_surface_unreference(&surface);
-		return -EINVAL;
-	}
-
 	/* takedown old cursor */
 	if (du->cursor_surface) {
 		du->cursor_surface->snooper.crtc = NULL;
@@ -218,19 +143,42 @@ int vmw_du_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv,
 		vmw_cursor_update_image(dev_priv, surface->snooper.image,
 					64, 64, du->hotspot_x, du->hotspot_y);
 	} else if (dmabuf) {
+		struct ttm_bo_kmap_obj map;
+		unsigned long kmap_offset;
+		unsigned long kmap_num;
+		void *virtual;
+		bool dummy;
+
 		/* vmw_user_surface_lookup takes one reference */
 		du->cursor_dmabuf = dmabuf;
 
-		ret = vmw_cursor_update_dmabuf(dev_priv, dmabuf, width, height,
-					       du->hotspot_x, du->hotspot_y);
+		kmap_offset = 0;
+		kmap_num = (64*64*4) >> PAGE_SHIFT;
+
+		ret = ttm_bo_reserve(&dmabuf->base, true, false, false, 0);
+		if (unlikely(ret != 0)) {
+			DRM_ERROR("reserve failed\n");
+			return -EINVAL;
+		}
+
+		ret = ttm_bo_kmap(&dmabuf->base, kmap_offset, kmap_num, &map);
+		if (unlikely(ret != 0))
+			goto err_unreserve;
+
+		virtual = ttm_kmap_obj_virtual(&map, &dummy);
+		vmw_cursor_update_image(dev_priv, virtual, 64, 64,
+					du->hotspot_x, du->hotspot_y);
+
+		ttm_bo_kunmap(&map);
+err_unreserve:
+		ttm_bo_unreserve(&dmabuf->base);
+
 	} else {
 		vmw_cursor_update_position(dev_priv, false, 0, 0);
 		return 0;
 	}
 
-	vmw_cursor_update_position(dev_priv, true,
-				   du->cursor_x + du->hotspot_x,
-				   du->cursor_y + du->hotspot_y);
+	vmw_cursor_update_position(dev_priv, true, du->cursor_x, du->cursor_y);
 
 	return 0;
 }
@@ -245,8 +193,7 @@ int vmw_du_crtc_cursor_move(struct drm_crtc *crtc, int x, int y)
 	du->cursor_y = y + crtc->y;
 
 	vmw_cursor_update_position(dev_priv, shown,
-				   du->cursor_x + du->hotspot_x,
-				   du->cursor_y + du->hotspot_y);
+				   du->cursor_x, du->cursor_y);
 
 	return 0;
 }
@@ -267,7 +214,7 @@ void vmw_kms_cursor_snoop(struct vmw_surface *srf,
 		SVGA3dCmdHeader header;
 		SVGA3dCmdSurfaceDMA dma;
 	} *cmd;
-	int i, ret;
+	int ret;
 
 	cmd = container_of(header, struct vmw_dma_cmd, header);
 
@@ -289,19 +236,16 @@ void vmw_kms_cursor_snoop(struct vmw_surface *srf,
 	box_count = (cmd->header.size - sizeof(SVGA3dCmdSurfaceDMA)) /
 			sizeof(SVGA3dCopyBox);
 
-	if (cmd->dma.guest.ptr.offset % PAGE_SIZE ||
+	if (cmd->dma.guest.pitch != (64 * 4) ||
+	    cmd->dma.guest.ptr.offset % PAGE_SIZE ||
 	    box->x != 0    || box->y != 0    || box->z != 0    ||
 	    box->srcx != 0 || box->srcy != 0 || box->srcz != 0 ||
-	    box->d != 1    || box_count != 1) {
+	    box->w != 64   || box->h != 64   || box->d != 1    ||
+	    box_count != 1) {
 		/* TODO handle none page aligned offsets */
-		/* TODO handle more dst & src != 0 */
-		/* TODO handle more then one copy */
-		DRM_ERROR("Cant snoop dma request for cursor!\n");
-		DRM_ERROR("(%u, %u, %u) (%u, %u, %u) (%ux%ux%u) %u %u\n",
-			  box->srcx, box->srcy, box->srcz,
-			  box->x, box->y, box->z,
-			  box->w, box->h, box->d, box_count,
-			  cmd->dma.guest.ptr.offset);
+		/* TODO handle partial uploads and pitch != 256 */
+		/* TODO handle more then one copy (size != 64) */
+		DRM_ERROR("lazy programmer, can't handle weird stuff\n");
 		return;
 	}
 
@@ -320,16 +264,7 @@ void vmw_kms_cursor_snoop(struct vmw_surface *srf,
 
 	virtual = ttm_kmap_obj_virtual(&map, &dummy);
 
-	if (box->w == 64 && cmd->dma.guest.pitch == 64*4) {
-		memcpy(srf->snooper.image, virtual, 64*64*4);
-	} else {
-		/* Image is unsigned pointer. */
-		for (i = 0; i < box->h; i++)
-			memcpy(srf->snooper.image + i * 64,
-			       virtual + i * cmd->dma.guest.pitch,
-			       box->w * 4);
-	}
-
+	memcpy(srf->snooper.image, virtual, 64*64*4);
 	srf->snooper.age++;
 
 	/* we can't call this function from this function since execbuf has
@@ -394,10 +329,41 @@ struct vmw_framebuffer_surface {
 	struct vmw_framebuffer base;
 	struct vmw_surface *surface;
 	struct vmw_dma_buffer *buffer;
+	struct delayed_work d_work;
+	struct mutex work_lock;
+	bool present_fs;
 	struct list_head head;
 	struct drm_master *master;
 };
 
+/**
+ * vmw_kms_idle_workqueues - Flush workqueues on this master
+ *
+ * @vmaster - Pointer identifying the master, for the surfaces of which
+ * we idle the dirty work queues.
+ *
+ * This function should be called with the ttm lock held in exclusive mode
+ * to idle all dirty work queues before the fifo is taken down.
+ *
+ * The work task may actually requeue itself, but after the flush returns we're
+ * sure that there's nothing to present, since the ttm lock is held in
+ * exclusive mode, so the fifo will never get used.
+ */
+
+void vmw_kms_idle_workqueues(struct vmw_master *vmaster)
+{
+	struct vmw_framebuffer_surface *entry;
+
+	mutex_lock(&vmaster->fb_surf_mutex);
+	list_for_each_entry(entry, &vmaster->fb_surf, head) {
+		if (cancel_delayed_work_sync(&entry->d_work))
+			(void) entry->d_work.work.func(&entry->d_work.work);
+
+		(void) cancel_delayed_work_sync(&entry->d_work);
+	}
+	mutex_unlock(&vmaster->fb_surf_mutex);
+}
+
 void vmw_framebuffer_surface_destroy(struct drm_framebuffer *framebuffer)
 {
 	struct vmw_framebuffer_surface *vfbs =
@@ -409,162 +375,65 @@ void vmw_framebuffer_surface_destroy(struct drm_framebuffer *framebuffer)
 	list_del(&vfbs->head);
 	mutex_unlock(&vmaster->fb_surf_mutex);
 
+	cancel_delayed_work_sync(&vfbs->d_work);
 	drm_master_put(&vfbs->master);
 	drm_framebuffer_cleanup(framebuffer);
 	vmw_surface_unreference(&vfbs->surface);
-	ttm_base_object_unref(&vfbs->base.user_obj);
 
 	kfree(vfbs);
 }
 
-static int do_surface_dirty_sou(struct vmw_private *dev_priv,
-				struct drm_file *file_priv,
-				struct vmw_framebuffer *framebuffer,
-				unsigned flags, unsigned color,
-				struct drm_clip_rect *clips,
-				unsigned num_clips, int inc,
-				struct vmw_fence_obj **out_fence)
+static void vmw_framebuffer_present_fs_callback(struct work_struct *work)
 {
-	struct vmw_display_unit *units[VMWGFX_NUM_DISPLAY_UNITS];
-	struct drm_clip_rect *clips_ptr;
-	struct drm_clip_rect *tmp;
-	struct drm_crtc *crtc;
-	size_t fifo_size;
-	int i, num_units;
-	int ret = 0; /* silence warning */
-	int left, right, top, bottom;
+	struct delayed_work *d_work =
+		container_of(work, struct delayed_work, work);
+	struct vmw_framebuffer_surface *vfbs =
+		container_of(d_work, struct vmw_framebuffer_surface, d_work);
+	struct vmw_surface *surf = vfbs->surface;
+	struct drm_framebuffer *framebuffer = &vfbs->base.base;
+	struct vmw_private *dev_priv = vmw_priv(framebuffer->dev);
 
 	struct {
 		SVGA3dCmdHeader header;
-		SVGA3dCmdBlitSurfaceToScreen body;
+		SVGA3dCmdPresent body;
+		SVGA3dCopyRect cr;
 	} *cmd;
-	SVGASignedRect *blits;
-
-	num_units = 0;
-	list_for_each_entry(crtc, &dev_priv->dev->mode_config.crtc_list,
-			    head) {
-		if (crtc->fb != &framebuffer->base)
-			continue;
-		units[num_units++] = vmw_crtc_to_du(crtc);
-	}
-
-	BUG_ON(!clips || !num_clips);
-
-	tmp = kzalloc(sizeof(*tmp) * num_clips, GFP_KERNEL);
-	if (unlikely(tmp == NULL)) {
-		DRM_ERROR("Temporary cliprect memory alloc failed.\n");
-		return -ENOMEM;
-	}
-
-	fifo_size = sizeof(*cmd) + sizeof(SVGASignedRect) * num_clips;
-	cmd = kzalloc(fifo_size, GFP_KERNEL);
-	if (unlikely(cmd == NULL)) {
-		DRM_ERROR("Temporary fifo memory alloc failed.\n");
-		ret = -ENOMEM;
-		goto out_free_tmp;
-	}
-
-	/* setup blits pointer */
-	blits = (SVGASignedRect *)&cmd[1];
-
-	/* initial clip region */
-	left = clips->x1;
-	right = clips->x2;
-	top = clips->y1;
-	bottom = clips->y2;
-
-	/* skip the first clip rect */
-	for (i = 1, clips_ptr = clips + inc;
-	     i < num_clips; i++, clips_ptr += inc) {
-		left = min_t(int, left, (int)clips_ptr->x1);
-		right = max_t(int, right, (int)clips_ptr->x2);
-		top = min_t(int, top, (int)clips_ptr->y1);
-		bottom = max_t(int, bottom, (int)clips_ptr->y2);
-	}
 
-	/* only need to do this once */
-	memset(cmd, 0, fifo_size);
-	cmd->header.id = cpu_to_le32(SVGA_3D_CMD_BLIT_SURFACE_TO_SCREEN);
-	cmd->header.size = cpu_to_le32(fifo_size - sizeof(cmd->header));
-
-	cmd->body.srcRect.left = left;
-	cmd->body.srcRect.right = right;
-	cmd->body.srcRect.top = top;
-	cmd->body.srcRect.bottom = bottom;
-
-	clips_ptr = clips;
-	for (i = 0; i < num_clips; i++, clips_ptr += inc) {
-		tmp[i].x1 = clips_ptr->x1 - left;
-		tmp[i].x2 = clips_ptr->x2 - left;
-		tmp[i].y1 = clips_ptr->y1 - top;
-		tmp[i].y2 = clips_ptr->y2 - top;
-	}
-
-	/* do per unit writing, reuse fifo for each */
-	for (i = 0; i < num_units; i++) {
-		struct vmw_display_unit *unit = units[i];
-		struct vmw_clip_rect clip;
-		int num;
-
-		clip.x1 = left - unit->crtc.x;
-		clip.y1 = top - unit->crtc.y;
-		clip.x2 = right - unit->crtc.x;
-		clip.y2 = bottom - unit->crtc.y;
-
-		/* skip any crtcs that misses the clip region */
-		if (clip.x1 >= unit->crtc.mode.hdisplay ||
-		    clip.y1 >= unit->crtc.mode.vdisplay ||
-		    clip.x2 <= 0 || clip.y2 <= 0)
-			continue;
-
-		/*
-		 * In order for the clip rects to be correctly scaled
-		 * the src and dest rects needs to be the same size.
-		 */
-		cmd->body.destRect.left = clip.x1;
-		cmd->body.destRect.right = clip.x2;
-		cmd->body.destRect.top = clip.y1;
-		cmd->body.destRect.bottom = clip.y2;
-
-		/* create a clip rect of the crtc in dest coords */
-		clip.x2 = unit->crtc.mode.hdisplay - clip.x1;
-		clip.y2 = unit->crtc.mode.vdisplay - clip.y1;
-		clip.x1 = 0 - clip.x1;
-		clip.y1 = 0 - clip.y1;
-
-		/* need to reset sid as it is changed by execbuf */
-		cmd->body.srcImage.sid = cpu_to_le32(framebuffer->user_handle);
-		cmd->body.destScreenId = unit->unit;
-
-		/* clip and write blits to cmd stream */
-		vmw_clip_cliprects(tmp, num_clips, clip, blits, &num);
-
-		/* if no cliprects hit skip this */
-		if (num == 0)
-			continue;
-
-		/* only return the last fence */
-		if (out_fence && *out_fence)
-			vmw_fence_obj_unreference(out_fence);
-
-		/* recalculate package length */
-		fifo_size = sizeof(*cmd) + sizeof(SVGASignedRect) * num;
-		cmd->header.size = cpu_to_le32(fifo_size - sizeof(cmd->header));
-		ret = vmw_execbuf_process(file_priv, dev_priv, NULL, cmd,
-					  fifo_size, 0, NULL, out_fence);
-
-		if (unlikely(ret != 0))
-			break;
-	}
-
-
-	kfree(cmd);
-out_free_tmp:
-	kfree(tmp);
+	/**
+	 * Strictly we should take the ttm_lock in read mode before accessing
+	 * the fifo, to make sure the fifo is present and up. However,
+	 * instead we flush all workqueues under the ttm lock in exclusive mode
+	 * before taking down the fifo.
+	 */
+	mutex_lock(&vfbs->work_lock);
+	if (!vfbs->present_fs)
+		goto out_unlock;
 
-	return ret;
+	cmd = vmw_fifo_reserve(dev_priv, sizeof(*cmd));
+	if (unlikely(cmd == NULL))
+		goto out_resched;
+
+	cmd->header.id = cpu_to_le32(SVGA_3D_CMD_PRESENT);
+	cmd->header.size = cpu_to_le32(sizeof(cmd->body) + sizeof(cmd->cr));
+	cmd->body.sid = cpu_to_le32(surf->res.id);
+	cmd->cr.x = cpu_to_le32(0);
+	cmd->cr.y = cpu_to_le32(0);
+	cmd->cr.srcx = cmd->cr.x;
+	cmd->cr.srcy = cmd->cr.y;
+	cmd->cr.w = cpu_to_le32(framebuffer->width);
+	cmd->cr.h = cpu_to_le32(framebuffer->height);
+	vfbs->present_fs = false;
+	vmw_fifo_commit(dev_priv, sizeof(*cmd));
+out_resched:
+	/**
+	 * Will not re-add if already pending.
+	 */
+	schedule_delayed_work(&vfbs->d_work, VMWGFX_PRESENT_RATE);
+out_unlock:
+	mutex_unlock(&vfbs->work_lock);
 }
 
+
 int vmw_framebuffer_surface_dirty(struct drm_framebuffer *framebuffer,
 				  struct drm_file *file_priv,
 				  unsigned flags, unsigned color,
@@ -575,20 +444,44 @@ int vmw_framebuffer_surface_dirty(struct drm_framebuffer *framebuffer,
 	struct vmw_master *vmaster = vmw_master(file_priv->master);
 	struct vmw_framebuffer_surface *vfbs =
 		vmw_framebuffer_to_vfbs(framebuffer);
+	struct vmw_surface *surf = vfbs->surface;
 	struct drm_clip_rect norect;
-	int ret, inc = 1;
+	SVGA3dCopyRect *cr;
+	int i, inc = 1;
+	int ret;
 
-	if (unlikely(vfbs->master != file_priv->master))
-		return -EINVAL;
+	struct {
+		SVGA3dCmdHeader header;
+		SVGA3dCmdPresent body;
+		SVGA3dCopyRect cr;
+	} *cmd;
 
-	/* Require ScreenObject support for 3D */
-	if (!dev_priv->sou_priv)
+	if (unlikely(vfbs->master != file_priv->master))
 		return -EINVAL;
 
 	ret = ttm_read_lock(&vmaster->lock, true);
 	if (unlikely(ret != 0))
 		return ret;
 
+	if (!num_clips ||
+	    !(dev_priv->fifo.capabilities &
+	      SVGA_FIFO_CAP_SCREEN_OBJECT)) {
+		int ret;
+
+		mutex_lock(&vfbs->work_lock);
+		vfbs->present_fs = true;
+		ret = schedule_delayed_work(&vfbs->d_work, VMWGFX_PRESENT_RATE);
+		mutex_unlock(&vfbs->work_lock);
+		if (ret) {
+			/**
+			 * No work pending, Force immediate present.
+			 */
+			vmw_framebuffer_present_fs_callback(&vfbs->d_work.work);
+		}
+		ttm_read_unlock(&vmaster->lock);
+		return 0;
+	}
+
 	if (!num_clips) {
 		num_clips = 1;
 		clips = &norect;
@@ -600,10 +493,29 @@ int vmw_framebuffer_surface_dirty(struct drm_framebuffer *framebuffer,
 		inc = 2; /* skip source rects */
 	}
 
-	ret = do_surface_dirty_sou(dev_priv, file_priv, &vfbs->base,
-				   flags, color,
-				   clips, num_clips, inc, NULL);
+	cmd = vmw_fifo_reserve(dev_priv, sizeof(*cmd) + (num_clips - 1) * sizeof(cmd->cr));
+	if (unlikely(cmd == NULL)) {
+		DRM_ERROR("Fifo reserve failed.\n");
+		ttm_read_unlock(&vmaster->lock);
+		return -ENOMEM;
+	}
+
+	memset(cmd, 0, sizeof(*cmd));
+
+	cmd->header.id = cpu_to_le32(SVGA_3D_CMD_PRESENT);
+	cmd->header.size = cpu_to_le32(sizeof(cmd->body) + num_clips * sizeof(cmd->cr));
+	cmd->body.sid = cpu_to_le32(surf->res.id);
+
+	for (i = 0, cr = &cmd->cr; i < num_clips; i++, cr++, clips += inc) {
+		cr->x = cpu_to_le16(clips->x1);
+		cr->y = cpu_to_le16(clips->y1);
+		cr->srcx = cr->x;
+		cr->srcy = cr->y;
+		cr->w = cpu_to_le16(clips->x2 - clips->x1);
+		cr->h = cpu_to_le16(clips->y2 - clips->y1);
+	}
 
+	vmw_fifo_commit(dev_priv, sizeof(*cmd) + (num_clips - 1) * sizeof(cmd->cr));
 	ttm_read_unlock(&vmaster->lock);
 	return 0;
 }
@@ -628,18 +540,10 @@ static int vmw_kms_new_framebuffer_surface(struct vmw_private *dev_priv,
 	struct vmw_master *vmaster = vmw_master(file_priv->master);
 	int ret;
 
-	/* 3D is only supported on HWv8 hosts which supports screen objects */
-	if (!dev_priv->sou_priv)
-		return -ENOSYS;
-
 	/*
 	 * Sanity checks.
 	 */
 
-	/* Surface must be marked as a scanout. */
-	if (unlikely(!surface->scanout))
-		return -EINVAL;
-
 	if (unlikely(surface->mip_levels[0] != 1 ||
 		     surface->num_sizes != 1 ||
 		     surface->sizes[0].width < mode_cmd->width ||
@@ -663,9 +567,6 @@ static int vmw_kms_new_framebuffer_surface(struct vmw_private *dev_priv,
 	case 15:
 		format = SVGA3D_A1R5G5B5;
 		break;
-	case 8:
-		format = SVGA3D_LUMINANCE8;
-		break;
 	default:
 		DRM_ERROR("Invalid color depth: %d\n", mode_cmd->depth);
 		return -EINVAL;
@@ -694,15 +595,18 @@ static int vmw_kms_new_framebuffer_surface(struct vmw_private *dev_priv,
 
 	/* XXX get the first 3 from the surface info */
 	vfbs->base.base.bits_per_pixel = mode_cmd->bpp;
-	vfbs->base.base.pitches[0] = mode_cmd->pitch;
+	vfbs->base.base.pitch = mode_cmd->pitch;
 	vfbs->base.base.depth = mode_cmd->depth;
 	vfbs->base.base.width = mode_cmd->width;
 	vfbs->base.base.height = mode_cmd->height;
+	vfbs->base.pin = &vmw_surface_dmabuf_pin;
+	vfbs->base.unpin = &vmw_surface_dmabuf_unpin;
 	vfbs->surface = surface;
-	vfbs->base.user_handle = mode_cmd->handle;
 	vfbs->master = drm_master_get(file_priv->master);
+	mutex_init(&vfbs->work_lock);
 
 	mutex_lock(&vmaster->fb_surf_mutex);
+	INIT_DELAYED_WORK(&vfbs->d_work, &vmw_framebuffer_present_fs_callback);
 	list_add_tail(&vfbs->head, &vmaster->fb_surf);
 	mutex_unlock(&vmaster->fb_surf_mutex);
 
@@ -737,33 +641,48 @@ void vmw_framebuffer_dmabuf_destroy(struct drm_framebuffer *framebuffer)
 
 	drm_framebuffer_cleanup(framebuffer);
 	vmw_dmabuf_unreference(&vfbd->buffer);
-	ttm_base_object_unref(&vfbd->base.user_obj);
 
 	kfree(vfbd);
 }
 
-static int do_dmabuf_dirty_ldu(struct vmw_private *dev_priv,
-			       struct vmw_framebuffer *framebuffer,
-			       unsigned flags, unsigned color,
-			       struct drm_clip_rect *clips,
-			       unsigned num_clips, int increment)
+int vmw_framebuffer_dmabuf_dirty(struct drm_framebuffer *framebuffer,
+				 struct drm_file *file_priv,
+				 unsigned flags, unsigned color,
+				 struct drm_clip_rect *clips,
+				 unsigned num_clips)
 {
-	size_t fifo_size;
-	int i;
-
+	struct vmw_private *dev_priv = vmw_priv(framebuffer->dev);
+	struct vmw_master *vmaster = vmw_master(file_priv->master);
+	struct drm_clip_rect norect;
+	int ret;
 	struct {
 		uint32_t header;
 		SVGAFifoCmdUpdate body;
 	} *cmd;
+	int i, increment = 1;
+
+	ret = ttm_read_lock(&vmaster->lock, true);
+	if (unlikely(ret != 0))
+		return ret;
+
+	if (!num_clips) {
+		num_clips = 1;
+		clips = &norect;
+		norect.x1 = norect.y1 = 0;
+		norect.x2 = framebuffer->width;
+		norect.y2 = framebuffer->height;
+	} else if (flags & DRM_MODE_FB_DIRTY_ANNOTATE_COPY) {
+		num_clips /= 2;
+		increment = 2;
+	}
 
-	fifo_size = sizeof(*cmd) * num_clips;
-	cmd = vmw_fifo_reserve(dev_priv, fifo_size);
+	cmd = vmw_fifo_reserve(dev_priv, sizeof(*cmd) * num_clips);
 	if (unlikely(cmd == NULL)) {
 		DRM_ERROR("Fifo reserve failed.\n");
+		ttm_read_unlock(&vmaster->lock);
 		return -ENOMEM;
 	}
 
-	memset(cmd, 0, fifo_size);
 	for (i = 0; i < num_clips; i++, clips += increment) {
 		cmd[i].header = cpu_to_le32(SVGA_CMD_UPDATE);
 		cmd[i].body.x = cpu_to_le32(clips->x1);
@@ -772,201 +691,57 @@ static int do_dmabuf_dirty_ldu(struct vmw_private *dev_priv,
 		cmd[i].body.height = cpu_to_le32(clips->y2 - clips->y1);
 	}
 
-	vmw_fifo_commit(dev_priv, fifo_size);
+	vmw_fifo_commit(dev_priv, sizeof(*cmd) * num_clips);
+	ttm_read_unlock(&vmaster->lock);
+
 	return 0;
 }
 
-static int do_dmabuf_define_gmrfb(struct drm_file *file_priv,
-				  struct vmw_private *dev_priv,
-				  struct vmw_framebuffer *framebuffer)
+static struct drm_framebuffer_funcs vmw_framebuffer_dmabuf_funcs = {
+	.destroy = vmw_framebuffer_dmabuf_destroy,
+	.dirty = vmw_framebuffer_dmabuf_dirty,
+	.create_handle = vmw_framebuffer_create_handle,
+};
+
+static int vmw_surface_dmabuf_pin(struct vmw_framebuffer *vfb)
 {
-	int depth = framebuffer->base.depth;
-	size_t fifo_size;
+	struct vmw_private *dev_priv = vmw_priv(vfb->base.dev);
+	struct vmw_framebuffer_surface *vfbs =
+		vmw_framebuffer_to_vfbs(&vfb->base);
+	unsigned long size = vfbs->base.base.pitch * vfbs->base.base.height;
 	int ret;
 
-	struct {
-		uint32_t header;
-		SVGAFifoCmdDefineGMRFB body;
-	} *cmd;
-
-	/* Emulate RGBA support, contrary to svga_reg.h this is not
-	 * supported by hosts. This is only a problem if we are reading
-	 * this value later and expecting what we uploaded back.
-	 */
-	if (depth == 32)
-		depth = 24;
-
-	fifo_size = sizeof(*cmd);
-	cmd = kmalloc(fifo_size, GFP_KERNEL);
-	if (unlikely(cmd == NULL)) {
-		DRM_ERROR("Failed to allocate temporary cmd buffer.\n");
+	vfbs->buffer = kzalloc(sizeof(*vfbs->buffer), GFP_KERNEL);
+	if (unlikely(vfbs->buffer == NULL))
 		return -ENOMEM;
-	}
-
-	memset(cmd, 0, fifo_size);
-	cmd->header = SVGA_CMD_DEFINE_GMRFB;
-	cmd->body.format.bitsPerPixel = framebuffer->base.bits_per_pixel;
-	cmd->body.format.colorDepth = depth;
-	cmd->body.format.reserved = 0;
-	cmd->body.bytesPerLine = framebuffer->base.pitches[0];
-	cmd->body.ptr.gmrId = framebuffer->user_handle;
-	cmd->body.ptr.offset = 0;
-
-	ret = vmw_execbuf_process(file_priv, dev_priv, NULL, cmd,
-				  fifo_size, 0, NULL, NULL);
-
-	kfree(cmd);
 
-	return ret;
-}
-
-static int do_dmabuf_dirty_sou(struct drm_file *file_priv,
-			       struct vmw_private *dev_priv,
-			       struct vmw_framebuffer *framebuffer,
-			       unsigned flags, unsigned color,
-			       struct drm_clip_rect *clips,
-			       unsigned num_clips, int increment,
-			       struct vmw_fence_obj **out_fence)
-{
-	struct vmw_display_unit *units[VMWGFX_NUM_DISPLAY_UNITS];
-	struct drm_clip_rect *clips_ptr;
-	int i, k, num_units, ret;
-	struct drm_crtc *crtc;
-	size_t fifo_size;
-
-	struct {
-		uint32_t header;
-		SVGAFifoCmdBlitGMRFBToScreen body;
-	} *blits;
-
-	ret = do_dmabuf_define_gmrfb(file_priv, dev_priv, framebuffer);
+	vmw_overlay_pause_all(dev_priv);
+	ret = vmw_dmabuf_init(dev_priv, vfbs->buffer, size,
+			       &vmw_vram_ne_placement,
+			       false, &vmw_dmabuf_bo_free);
+	vmw_overlay_resume_all(dev_priv);
 	if (unlikely(ret != 0))
-		return ret; /* define_gmrfb prints warnings */
-
-	fifo_size = sizeof(*blits) * num_clips;
-	blits = kmalloc(fifo_size, GFP_KERNEL);
-	if (unlikely(blits == NULL)) {
-		DRM_ERROR("Failed to allocate temporary cmd buffer.\n");
-		return -ENOMEM;
-	}
-
-	num_units = 0;
-	list_for_each_entry(crtc, &dev_priv->dev->mode_config.crtc_list, head) {
-		if (crtc->fb != &framebuffer->base)
-			continue;
-		units[num_units++] = vmw_crtc_to_du(crtc);
-	}
-
-	for (k = 0; k < num_units; k++) {
-		struct vmw_display_unit *unit = units[k];
-		int hit_num = 0;
-
-		clips_ptr = clips;
-		for (i = 0; i < num_clips; i++, clips_ptr += increment) {
-			int clip_x1 = clips_ptr->x1 - unit->crtc.x;
-			int clip_y1 = clips_ptr->y1 - unit->crtc.y;
-			int clip_x2 = clips_ptr->x2 - unit->crtc.x;
-			int clip_y2 = clips_ptr->y2 - unit->crtc.y;
-			int move_x, move_y;
-
-			/* skip any crtcs that misses the clip region */
-			if (clip_x1 >= unit->crtc.mode.hdisplay ||
-			    clip_y1 >= unit->crtc.mode.vdisplay ||
-			    clip_x2 <= 0 || clip_y2 <= 0)
-				continue;
-
-			/* clip size to crtc size */
-			clip_x2 = min_t(int, clip_x2, unit->crtc.mode.hdisplay);
-			clip_y2 = min_t(int, clip_y2, unit->crtc.mode.vdisplay);
-
-			/* translate both src and dest to bring clip into screen */
-			move_x = min_t(int, clip_x1, 0);
-			move_y = min_t(int, clip_y1, 0);
-
-			/* actual translate done here */
-			blits[hit_num].header = SVGA_CMD_BLIT_GMRFB_TO_SCREEN;
-			blits[hit_num].body.destScreenId = unit->unit;
-			blits[hit_num].body.srcOrigin.x = clips_ptr->x1 - move_x;
-			blits[hit_num].body.srcOrigin.y = clips_ptr->y1 - move_y;
-			blits[hit_num].body.destRect.left = clip_x1 - move_x;
-			blits[hit_num].body.destRect.top = clip_y1 - move_y;
-			blits[hit_num].body.destRect.right = clip_x2;
-			blits[hit_num].body.destRect.bottom = clip_y2;
-			hit_num++;
-		}
-
-		/* no clips hit the crtc */
-		if (hit_num == 0)
-			continue;
-
-		/* only return the last fence */
-		if (out_fence && *out_fence)
-			vmw_fence_obj_unreference(out_fence);
-
-		fifo_size = sizeof(*blits) * hit_num;
-		ret = vmw_execbuf_process(file_priv, dev_priv, NULL, blits,
-					  fifo_size, 0, NULL, out_fence);
-
-		if (unlikely(ret != 0))
-			break;
-	}
-
-	kfree(blits);
+		vfbs->buffer = NULL;
 
 	return ret;
 }
 
-int vmw_framebuffer_dmabuf_dirty(struct drm_framebuffer *framebuffer,
-				 struct drm_file *file_priv,
-				 unsigned flags, unsigned color,
-				 struct drm_clip_rect *clips,
-				 unsigned num_clips)
+static int vmw_surface_dmabuf_unpin(struct vmw_framebuffer *vfb)
 {
-	struct vmw_private *dev_priv = vmw_priv(framebuffer->dev);
-	struct vmw_master *vmaster = vmw_master(file_priv->master);
-	struct vmw_framebuffer_dmabuf *vfbd =
-		vmw_framebuffer_to_vfbd(framebuffer);
-	struct drm_clip_rect norect;
-	int ret, increment = 1;
-
-	ret = ttm_read_lock(&vmaster->lock, true);
-	if (unlikely(ret != 0))
-		return ret;
+	struct ttm_buffer_object *bo;
+	struct vmw_framebuffer_surface *vfbs =
+		vmw_framebuffer_to_vfbs(&vfb->base);
 
-	if (!num_clips) {
-		num_clips = 1;
-		clips = &norect;
-		norect.x1 = norect.y1 = 0;
-		norect.x2 = framebuffer->width;
-		norect.y2 = framebuffer->height;
-	} else if (flags & DRM_MODE_FB_DIRTY_ANNOTATE_COPY) {
-		num_clips /= 2;
-		increment = 2;
-	}
+	if (unlikely(vfbs->buffer == NULL))
+		return 0;
 
-	if (dev_priv->ldu_priv) {
-		ret = do_dmabuf_dirty_ldu(dev_priv, &vfbd->base,
-					  flags, color,
-					  clips, num_clips, increment);
-	} else {
-		ret = do_dmabuf_dirty_sou(file_priv, dev_priv, &vfbd->base,
-					  flags, color,
-					  clips, num_clips, increment, NULL);
-	}
+	bo = &vfbs->buffer->base;
+	ttm_bo_unref(&bo);
+	vfbs->buffer = NULL;
 
-	ttm_read_unlock(&vmaster->lock);
-	return ret;
+	return 0;
 }
 
-static struct drm_framebuffer_funcs vmw_framebuffer_dmabuf_funcs = {
-	.destroy = vmw_framebuffer_dmabuf_destroy,
-	.dirty = vmw_framebuffer_dmabuf_dirty,
-	.create_handle = vmw_framebuffer_create_handle,
-};
-
-/**
- * Pin the dmabuffer to the start of vram.
- */
 static int vmw_framebuffer_dmabuf_pin(struct vmw_framebuffer *vfb)
 {
 	struct vmw_private *dev_priv = vmw_priv(vfb->base.dev);
@@ -974,12 +749,10 @@ static int vmw_framebuffer_dmabuf_pin(struct vmw_framebuffer *vfb)
 		vmw_framebuffer_to_vfbd(&vfb->base);
 	int ret;
 
-	/* This code should not be used with screen objects */
-	BUG_ON(dev_priv->sou_priv);
 
 	vmw_overlay_pause_all(dev_priv);
 
-	ret = vmw_dmabuf_to_start_of_vram(dev_priv, vfbd->buffer, true, false);
+	ret = vmw_dmabuf_to_start_of_vram(dev_priv, vfbd->buffer);
 
 	vmw_overlay_resume_all(dev_priv);
 
@@ -999,7 +772,7 @@ static int vmw_framebuffer_dmabuf_unpin(struct vmw_framebuffer *vfb)
 		return 0;
 	}
 
-	return vmw_dmabuf_unpin(dev_priv, vfbd->buffer, false);
+	return vmw_dmabuf_from_vram(dev_priv, vfbd->buffer);
 }
 
 static int vmw_kms_new_framebuffer_dmabuf(struct vmw_private *dev_priv,
@@ -1021,33 +794,6 @@ static int vmw_kms_new_framebuffer_dmabuf(struct vmw_private *dev_priv,
 		return -EINVAL;
 	}
 
-	/* Limited framebuffer color depth support for screen objects */
-	if (dev_priv->sou_priv) {
-		switch (mode_cmd->depth) {
-		case 32:
-		case 24:
-			/* Only support 32 bpp for 32 and 24 depth fbs */
-			if (mode_cmd->bpp == 32)
-				break;
-
-			DRM_ERROR("Invalid color depth/bbp: %d %d\n",
-				  mode_cmd->depth, mode_cmd->bpp);
-			return -EINVAL;
-		case 16:
-		case 15:
-			/* Only support 16 bpp for 16 and 15 depth fbs */
-			if (mode_cmd->bpp == 16)
-				break;
-
-			DRM_ERROR("Invalid color depth/bbp: %d %d\n",
-				  mode_cmd->depth, mode_cmd->bpp);
-			return -EINVAL;
-		default:
-			DRM_ERROR("Invalid color depth: %d\n", mode_cmd->depth);
-			return -EINVAL;
-		}
-	}
-
 	vfbd = kzalloc(sizeof(*vfbd), GFP_KERNEL);
 	if (!vfbd) {
 		ret = -ENOMEM;
@@ -1065,17 +811,13 @@ static int vmw_kms_new_framebuffer_dmabuf(struct vmw_private *dev_priv,
 	}
 
 	vfbd->base.base.bits_per_pixel = mode_cmd->bpp;
-	vfbd->base.base.pitches[0] = mode_cmd->pitch;
+	vfbd->base.base.pitch = mode_cmd->pitch;
 	vfbd->base.base.depth = mode_cmd->depth;
 	vfbd->base.base.width = mode_cmd->width;
 	vfbd->base.base.height = mode_cmd->height;
-	if (!dev_priv->sou_priv) {
-		vfbd->base.pin = vmw_framebuffer_dmabuf_pin;
-		vfbd->base.unpin = vmw_framebuffer_dmabuf_unpin;
-	}
-	vfbd->base.dmabuf = true;
+	vfbd->base.pin = vmw_framebuffer_dmabuf_pin;
+	vfbd->base.unpin = vmw_framebuffer_dmabuf_unpin;
 	vfbd->buffer = dmabuf;
-	vfbd->base.user_handle = mode_cmd->handle;
 	*out = &vfbd->base;
 
 	return 0;
@@ -1094,336 +836,85 @@ out_err1:
 
 static struct drm_framebuffer *vmw_kms_fb_create(struct drm_device *dev,
 						 struct drm_file *file_priv,
-						 struct drm_mode_fb_cmd2 *mode_cmd2)
+						 struct drm_mode_fb_cmd *mode_cmd)
 {
 	struct vmw_private *dev_priv = vmw_priv(dev);
 	struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile;
 	struct vmw_framebuffer *vfb = NULL;
 	struct vmw_surface *surface = NULL;
 	struct vmw_dma_buffer *bo = NULL;
-	struct ttm_base_object *user_obj;
-	struct drm_mode_fb_cmd mode_cmd;
+	u64 required_size;
 	int ret;
 
-	mode_cmd.width = mode_cmd2->width;
-	mode_cmd.height = mode_cmd2->height;
-	mode_cmd.pitch = mode_cmd2->pitches[0];
-	mode_cmd.handle = mode_cmd2->handles[0];
-	drm_fb_get_bpp_depth(mode_cmd2->pixel_format, &mode_cmd.depth,
-				    &mode_cmd.bpp);
-
 	/**
 	 * This code should be conditioned on Screen Objects not being used.
 	 * If screen objects are used, we can allocate a GMR to hold the
 	 * requested framebuffer.
 	 */
 
-	if (!vmw_kms_validate_mode_vram(dev_priv,
-					mode_cmd.pitch,
-					mode_cmd.height)) {
+	required_size = mode_cmd->pitch * mode_cmd->height;
+	if (unlikely(required_size > (u64) dev_priv->vram_size)) {
 		DRM_ERROR("VRAM size is too small for requested mode.\n");
-		return ERR_PTR(-ENOMEM);
-	}
-
-	/*
-	 * Take a reference on the user object of the resource
-	 * backing the kms fb. This ensures that user-space handle
-	 * lookups on that resource will always work as long as
-	 * it's registered with a kms framebuffer. This is important,
-	 * since vmw_execbuf_process identifies resources in the
-	 * command stream using user-space handles.
-	 */
-
-	user_obj = ttm_base_object_lookup(tfile, mode_cmd.handle);
-	if (unlikely(user_obj == NULL)) {
-		DRM_ERROR("Could not locate requested kms frame buffer.\n");
-		return ERR_PTR(-ENOENT);
+		return NULL;
 	}
 
 	/**
 	 * End conditioned code.
 	 */
 
-	/* returns either a dmabuf or surface */
-	ret = vmw_user_lookup_handle(dev_priv, tfile,
-				     mode_cmd.handle,
-				     &surface, &bo);
+	ret = vmw_user_surface_lookup_handle(dev_priv, tfile,
+					     mode_cmd->handle, &surface);
 	if (ret)
-		goto err_out;
-
-	/* Create the new framebuffer depending one what we got back */
-	if (bo)
-		ret = vmw_kms_new_framebuffer_dmabuf(dev_priv, bo, &vfb,
-						     &mode_cmd);
-	else if (surface)
-		ret = vmw_kms_new_framebuffer_surface(dev_priv, file_priv,
-						      surface, &vfb, &mode_cmd);
-	else
-		BUG();
-
-err_out:
-	/* vmw_user_lookup_handle takes one ref so does new_fb */
-	if (bo)
-		vmw_dmabuf_unreference(&bo);
-	if (surface)
-		vmw_surface_unreference(&surface);
-
-	if (ret) {
-		DRM_ERROR("failed to create vmw_framebuffer: %i\n", ret);
-		ttm_base_object_unref(&user_obj);
-		return ERR_PTR(ret);
-	} else
-		vfb->user_obj = user_obj;
-
-	return &vfb->base;
-}
-
-static struct drm_mode_config_funcs vmw_kms_funcs = {
-	.fb_create = vmw_kms_fb_create,
-};
-
-int vmw_kms_present(struct vmw_private *dev_priv,
-		    struct drm_file *file_priv,
-		    struct vmw_framebuffer *vfb,
-		    struct vmw_surface *surface,
-		    uint32_t sid,
-		    int32_t destX, int32_t destY,
-		    struct drm_vmw_rect *clips,
-		    uint32_t num_clips)
-{
-	struct vmw_display_unit *units[VMWGFX_NUM_DISPLAY_UNITS];
-	struct drm_clip_rect *tmp;
-	struct drm_crtc *crtc;
-	size_t fifo_size;
-	int i, k, num_units;
-	int ret = 0; /* silence warning */
-	int left, right, top, bottom;
-
-	struct {
-		SVGA3dCmdHeader header;
-		SVGA3dCmdBlitSurfaceToScreen body;
-	} *cmd;
-	SVGASignedRect *blits;
+		goto try_dmabuf;
 
-	num_units = 0;
-	list_for_each_entry(crtc, &dev_priv->dev->mode_config.crtc_list, head) {
-		if (crtc->fb != &vfb->base)
-			continue;
-		units[num_units++] = vmw_crtc_to_du(crtc);
-	}
+	if (!surface->scanout)
+		goto err_not_scanout;
 
-	BUG_ON(surface == NULL);
-	BUG_ON(!clips || !num_clips);
+	ret = vmw_kms_new_framebuffer_surface(dev_priv, file_priv, surface,
+					      &vfb, mode_cmd);
 
-	tmp = kzalloc(sizeof(*tmp) * num_clips, GFP_KERNEL);
-	if (unlikely(tmp == NULL)) {
-		DRM_ERROR("Temporary cliprect memory alloc failed.\n");
-		return -ENOMEM;
-	}
+	/* vmw_user_surface_lookup takes one ref so does new_fb */
+	vmw_surface_unreference(&surface);
 
-	fifo_size = sizeof(*cmd) + sizeof(SVGASignedRect) * num_clips;
-	cmd = kmalloc(fifo_size, GFP_KERNEL);
-	if (unlikely(cmd == NULL)) {
-		DRM_ERROR("Failed to allocate temporary fifo memory.\n");
-		ret = -ENOMEM;
-		goto out_free_tmp;
+	if (ret) {
+		DRM_ERROR("failed to create vmw_framebuffer: %i\n", ret);
+		return ERR_PTR(ret);
 	}
+	return &vfb->base;
 
-	left = clips->x;
-	right = clips->x + clips->w;
-	top = clips->y;
-	bottom = clips->y + clips->h;
+try_dmabuf:
+	DRM_INFO("%s: trying buffer\n", __func__);
 
-	for (i = 1; i < num_clips; i++) {
-		left = min_t(int, left, (int)clips[i].x);
-		right = max_t(int, right, (int)clips[i].x + clips[i].w);
-		top = min_t(int, top, (int)clips[i].y);
-		bottom = max_t(int, bottom, (int)clips[i].y + clips[i].h);
+	ret = vmw_user_dmabuf_lookup(tfile, mode_cmd->handle, &bo);
+	if (ret) {
+		DRM_ERROR("failed to find buffer: %i\n", ret);
+		return ERR_PTR(-ENOENT);
 	}
 
-	/* only need to do this once */
-	memset(cmd, 0, fifo_size);
-	cmd->header.id = cpu_to_le32(SVGA_3D_CMD_BLIT_SURFACE_TO_SCREEN);
+	ret = vmw_kms_new_framebuffer_dmabuf(dev_priv, bo, &vfb,
+					     mode_cmd);
 
-	blits = (SVGASignedRect *)&cmd[1];
+	/* vmw_user_dmabuf_lookup takes one ref so does new_fb */
+	vmw_dmabuf_unreference(&bo);
 
-	cmd->body.srcRect.left = left;
-	cmd->body.srcRect.right = right;
-	cmd->body.srcRect.top = top;
-	cmd->body.srcRect.bottom = bottom;
-
-	for (i = 0; i < num_clips; i++) {
-		tmp[i].x1 = clips[i].x - left;
-		tmp[i].x2 = clips[i].x + clips[i].w - left;
-		tmp[i].y1 = clips[i].y - top;
-		tmp[i].y2 = clips[i].y + clips[i].h - top;
+	if (ret) {
+		DRM_ERROR("failed to create vmw_framebuffer: %i\n", ret);
+		return ERR_PTR(ret);
 	}
 
-	for (k = 0; k < num_units; k++) {
-		struct vmw_display_unit *unit = units[k];
-		struct vmw_clip_rect clip;
-		int num;
-
-		clip.x1 = left + destX - unit->crtc.x;
-		clip.y1 = top + destY - unit->crtc.y;
-		clip.x2 = right + destX - unit->crtc.x;
-		clip.y2 = bottom + destY - unit->crtc.y;
-
-		/* skip any crtcs that misses the clip region */
-		if (clip.x1 >= unit->crtc.mode.hdisplay ||
-		    clip.y1 >= unit->crtc.mode.vdisplay ||
-		    clip.x2 <= 0 || clip.y2 <= 0)
-			continue;
-
-		/*
-		 * In order for the clip rects to be correctly scaled
-		 * the src and dest rects needs to be the same size.
-		 */
-		cmd->body.destRect.left = clip.x1;
-		cmd->body.destRect.right = clip.x2;
-		cmd->body.destRect.top = clip.y1;
-		cmd->body.destRect.bottom = clip.y2;
-
-		/* create a clip rect of the crtc in dest coords */
-		clip.x2 = unit->crtc.mode.hdisplay - clip.x1;
-		clip.y2 = unit->crtc.mode.vdisplay - clip.y1;
-		clip.x1 = 0 - clip.x1;
-		clip.y1 = 0 - clip.y1;
-
-		/* need to reset sid as it is changed by execbuf */
-		cmd->body.srcImage.sid = sid;
-		cmd->body.destScreenId = unit->unit;
-
-		/* clip and write blits to cmd stream */
-		vmw_clip_cliprects(tmp, num_clips, clip, blits, &num);
-
-		/* if no cliprects hit skip this */
-		if (num == 0)
-			continue;
-
-		/* recalculate package length */
-		fifo_size = sizeof(*cmd) + sizeof(SVGASignedRect) * num;
-		cmd->header.size = cpu_to_le32(fifo_size - sizeof(cmd->header));
-		ret = vmw_execbuf_process(file_priv, dev_priv, NULL, cmd,
-					  fifo_size, 0, NULL, NULL);
-
-		if (unlikely(ret != 0))
-			break;
-	}
+	return &vfb->base;
 
-	kfree(cmd);
-out_free_tmp:
-	kfree(tmp);
+err_not_scanout:
+	DRM_ERROR("surface not marked as scanout\n");
+	/* vmw_user_surface_lookup takes one ref */
+	vmw_surface_unreference(&surface);
 
-	return ret;
+	return ERR_PTR(-EINVAL);
 }
 
-int vmw_kms_readback(struct vmw_private *dev_priv,
-		     struct drm_file *file_priv,
-		     struct vmw_framebuffer *vfb,
-		     struct drm_vmw_fence_rep __user *user_fence_rep,
-		     struct drm_vmw_rect *clips,
-		     uint32_t num_clips)
-{
-	struct vmw_framebuffer_dmabuf *vfbd =
-		vmw_framebuffer_to_vfbd(&vfb->base);
-	struct vmw_dma_buffer *dmabuf = vfbd->buffer;
-	struct vmw_display_unit *units[VMWGFX_NUM_DISPLAY_UNITS];
-	struct drm_crtc *crtc;
-	size_t fifo_size;
-	int i, k, ret, num_units, blits_pos;
-
-	struct {
-		uint32_t header;
-		SVGAFifoCmdDefineGMRFB body;
-	} *cmd;
-	struct {
-		uint32_t header;
-		SVGAFifoCmdBlitScreenToGMRFB body;
-	} *blits;
-
-	num_units = 0;
-	list_for_each_entry(crtc, &dev_priv->dev->mode_config.crtc_list, head) {
-		if (crtc->fb != &vfb->base)
-			continue;
-		units[num_units++] = vmw_crtc_to_du(crtc);
-	}
-
-	BUG_ON(dmabuf == NULL);
-	BUG_ON(!clips || !num_clips);
-
-	/* take a safe guess at fifo size */
-	fifo_size = sizeof(*cmd) + sizeof(*blits) * num_clips * num_units;
-	cmd = kmalloc(fifo_size, GFP_KERNEL);
-	if (unlikely(cmd == NULL)) {
-		DRM_ERROR("Failed to allocate temporary fifo memory.\n");
-		return -ENOMEM;
-	}
-
-	memset(cmd, 0, fifo_size);
-	cmd->header = SVGA_CMD_DEFINE_GMRFB;
-	cmd->body.format.bitsPerPixel = vfb->base.bits_per_pixel;
-	cmd->body.format.colorDepth = vfb->base.depth;
-	cmd->body.format.reserved = 0;
-	cmd->body.bytesPerLine = vfb->base.pitches[0];
-	cmd->body.ptr.gmrId = vfb->user_handle;
-	cmd->body.ptr.offset = 0;
-
-	blits = (void *)&cmd[1];
-	blits_pos = 0;
-	for (i = 0; i < num_units; i++) {
-		struct drm_vmw_rect *c = clips;
-		for (k = 0; k < num_clips; k++, c++) {
-			/* transform clip coords to crtc origin based coords */
-			int clip_x1 = c->x - units[i]->crtc.x;
-			int clip_x2 = c->x - units[i]->crtc.x + c->w;
-			int clip_y1 = c->y - units[i]->crtc.y;
-			int clip_y2 = c->y - units[i]->crtc.y + c->h;
-			int dest_x = c->x;
-			int dest_y = c->y;
-
-			/* compensate for clipping, we negate
-			 * a negative number and add that.
-			 */
-			if (clip_x1 < 0)
-				dest_x += -clip_x1;
-			if (clip_y1 < 0)
-				dest_y += -clip_y1;
-
-			/* clip */
-			clip_x1 = max(clip_x1, 0);
-			clip_y1 = max(clip_y1, 0);
-			clip_x2 = min(clip_x2, units[i]->crtc.mode.hdisplay);
-			clip_y2 = min(clip_y2, units[i]->crtc.mode.vdisplay);
-
-			/* and cull any rects that misses the crtc */
-			if (clip_x1 >= units[i]->crtc.mode.hdisplay ||
-			    clip_y1 >= units[i]->crtc.mode.vdisplay ||
-			    clip_x2 <= 0 || clip_y2 <= 0)
-				continue;
-
-			blits[blits_pos].header = SVGA_CMD_BLIT_SCREEN_TO_GMRFB;
-			blits[blits_pos].body.srcScreenId = units[i]->unit;
-			blits[blits_pos].body.destOrigin.x = dest_x;
-			blits[blits_pos].body.destOrigin.y = dest_y;
-
-			blits[blits_pos].body.srcRect.left = clip_x1;
-			blits[blits_pos].body.srcRect.top = clip_y1;
-			blits[blits_pos].body.srcRect.right = clip_x2;
-			blits[blits_pos].body.srcRect.bottom = clip_y2;
-			blits_pos++;
-		}
-	}
-	/* reset size here and use calculated exact size from loops */
-	fifo_size = sizeof(*cmd) + sizeof(*blits) * blits_pos;
-
-	ret = vmw_execbuf_process(file_priv, dev_priv, NULL, cmd, fifo_size,
-				  0, user_fence_rep, NULL);
-
-	kfree(cmd);
-
-	return ret;
-}
+static struct drm_mode_config_funcs vmw_kms_funcs = {
+	.fb_create = vmw_kms_fb_create,
+};
 
 int vmw_kms_init(struct vmw_private *dev_priv)
 {
@@ -1438,9 +929,7 @@ int vmw_kms_init(struct vmw_private *dev_priv)
 	dev->mode_config.max_width = 8192;
 	dev->mode_config.max_height = 8192;
 
-	ret = vmw_kms_init_screen_object_display(dev_priv);
-	if (ret) /* Fallback */
-		(void)vmw_kms_init_legacy_display_system(dev_priv);
+	ret = vmw_kms_init_legacy_display_system(dev_priv);
 
 	return 0;
 }
@@ -1453,10 +942,7 @@ int vmw_kms_close(struct vmw_private *dev_priv)
 	 * drm_encoder_cleanup which takes the lock we deadlock.
 	 */
 	drm_mode_config_cleanup(dev_priv->dev);
-	if (dev_priv->sou_priv)
-		vmw_kms_close_screen_object_display(dev_priv);
-	else
-		vmw_kms_close_legacy_display_system(dev_priv);
+	vmw_kms_close_legacy_display_system(dev_priv);
 	return 0;
 }
 
@@ -1501,9 +987,9 @@ out:
 	return ret;
 }
 
-int vmw_kms_write_svga(struct vmw_private *vmw_priv,
+void vmw_kms_write_svga(struct vmw_private *vmw_priv,
 			unsigned width, unsigned height, unsigned pitch,
-			unsigned bpp, unsigned depth)
+			unsigned bbp, unsigned depth)
 {
 	if (vmw_priv->capabilities & SVGA_CAP_PITCHLOCK)
 		vmw_write(vmw_priv, SVGA_REG_PITCHLOCK, pitch);
@@ -1511,15 +997,11 @@ int vmw_kms_write_svga(struct vmw_private *vmw_priv,
 		iowrite32(pitch, vmw_priv->mmio_virt + SVGA_FIFO_PITCHLOCK);
 	vmw_write(vmw_priv, SVGA_REG_WIDTH, width);
 	vmw_write(vmw_priv, SVGA_REG_HEIGHT, height);
-	vmw_write(vmw_priv, SVGA_REG_BITS_PER_PIXEL, bpp);
-
-	if (vmw_read(vmw_priv, SVGA_REG_DEPTH) != depth) {
-		DRM_ERROR("Invalid depth %u for %u bpp, host expects %u\n",
-			  depth, bpp, vmw_read(vmw_priv, SVGA_REG_DEPTH));
-		return -EINVAL;
-	}
-
-	return 0;
+	vmw_write(vmw_priv, SVGA_REG_BITS_PER_PIXEL, bbp);
+	vmw_write(vmw_priv, SVGA_REG_DEPTH, depth);
+	vmw_write(vmw_priv, SVGA_REG_RED_MASK, 0x00ff0000);
+	vmw_write(vmw_priv, SVGA_REG_GREEN_MASK, 0x0000ff00);
+	vmw_write(vmw_priv, SVGA_REG_BLUE_MASK, 0x000000ff);
 }
 
 int vmw_kms_save_vga(struct vmw_private *vmw_priv)
@@ -1529,7 +1011,12 @@ int vmw_kms_save_vga(struct vmw_private *vmw_priv)
 
 	vmw_priv->vga_width = vmw_read(vmw_priv, SVGA_REG_WIDTH);
 	vmw_priv->vga_height = vmw_read(vmw_priv, SVGA_REG_HEIGHT);
+	vmw_priv->vga_depth = vmw_read(vmw_priv, SVGA_REG_DEPTH);
 	vmw_priv->vga_bpp = vmw_read(vmw_priv, SVGA_REG_BITS_PER_PIXEL);
+	vmw_priv->vga_pseudo = vmw_read(vmw_priv, SVGA_REG_PSEUDOCOLOR);
+	vmw_priv->vga_red_mask = vmw_read(vmw_priv, SVGA_REG_RED_MASK);
+	vmw_priv->vga_blue_mask = vmw_read(vmw_priv, SVGA_REG_BLUE_MASK);
+	vmw_priv->vga_green_mask = vmw_read(vmw_priv, SVGA_REG_GREEN_MASK);
 	if (vmw_priv->capabilities & SVGA_CAP_PITCHLOCK)
 		vmw_priv->vga_pitchlock =
 		  vmw_read(vmw_priv, SVGA_REG_PITCHLOCK);
@@ -1578,7 +1065,12 @@ int vmw_kms_restore_vga(struct vmw_private *vmw_priv)
 
 	vmw_write(vmw_priv, SVGA_REG_WIDTH, vmw_priv->vga_width);
 	vmw_write(vmw_priv, SVGA_REG_HEIGHT, vmw_priv->vga_height);
+	vmw_write(vmw_priv, SVGA_REG_DEPTH, vmw_priv->vga_depth);
 	vmw_write(vmw_priv, SVGA_REG_BITS_PER_PIXEL, vmw_priv->vga_bpp);
+	vmw_write(vmw_priv, SVGA_REG_PSEUDOCOLOR, vmw_priv->vga_pseudo);
+	vmw_write(vmw_priv, SVGA_REG_RED_MASK, vmw_priv->vga_red_mask);
+	vmw_write(vmw_priv, SVGA_REG_GREEN_MASK, vmw_priv->vga_green_mask);
+	vmw_write(vmw_priv, SVGA_REG_BLUE_MASK, vmw_priv->vga_blue_mask);
 	if (vmw_priv->capabilities & SVGA_CAP_PITCHLOCK)
 		vmw_write(vmw_priv, SVGA_REG_PITCHLOCK,
 			  vmw_priv->vga_pitchlock);
@@ -1603,376 +1095,6 @@ int vmw_kms_restore_vga(struct vmw_private *vmw_priv)
 	return 0;
 }
 
-bool vmw_kms_validate_mode_vram(struct vmw_private *dev_priv,
-				uint32_t pitch,
-				uint32_t height)
-{
-	return ((u64) pitch * (u64) height) < (u64) dev_priv->vram_size;
-}
-
-
-/**
- * Function called by DRM code called with vbl_lock held.
- */
-u32 vmw_get_vblank_counter(struct drm_device *dev, int crtc)
-{
-	return 0;
-}
-
-/**
- * Function called by DRM code called with vbl_lock held.
- */
-int vmw_enable_vblank(struct drm_device *dev, int crtc)
-{
-	return -ENOSYS;
-}
-
-/**
- * Function called by DRM code called with vbl_lock held.
- */
-void vmw_disable_vblank(struct drm_device *dev, int crtc)
-{
-}
-
-
-/*
- * Small shared kms functions.
- */
-
-int vmw_du_update_layout(struct vmw_private *dev_priv, unsigned num,
-			 struct drm_vmw_rect *rects)
-{
-	struct drm_device *dev = dev_priv->dev;
-	struct vmw_display_unit *du;
-	struct drm_connector *con;
-
-	mutex_lock(&dev->mode_config.mutex);
-
-#if 0
-	{
-		unsigned int i;
-
-		DRM_INFO("%s: new layout ", __func__);
-		for (i = 0; i < num; i++)
-			DRM_INFO("(%i, %i %ux%u) ", rects[i].x, rects[i].y,
-				 rects[i].w, rects[i].h);
-		DRM_INFO("\n");
-	}
-#endif
-
-	list_for_each_entry(con, &dev->mode_config.connector_list, head) {
-		du = vmw_connector_to_du(con);
-		if (num > du->unit) {
-			du->pref_width = rects[du->unit].w;
-			du->pref_height = rects[du->unit].h;
-			du->pref_active = true;
-			du->gui_x = rects[du->unit].x;
-			du->gui_y = rects[du->unit].y;
-		} else {
-			du->pref_width = 800;
-			du->pref_height = 600;
-			du->pref_active = false;
-		}
-		con->status = vmw_du_connector_detect(con, true);
-	}
-
-	mutex_unlock(&dev->mode_config.mutex);
-
-	return 0;
-}
-
-int vmw_du_page_flip(struct drm_crtc *crtc,
-		     struct drm_framebuffer *fb,
-		     struct drm_pending_vblank_event *event)
-{
-	struct vmw_private *dev_priv = vmw_priv(crtc->dev);
-	struct drm_framebuffer *old_fb = crtc->fb;
-	struct vmw_framebuffer *vfb = vmw_framebuffer_to_vfb(fb);
-	struct drm_file *file_priv = event->base.file_priv;
-	struct vmw_fence_obj *fence = NULL;
-	struct drm_clip_rect clips;
-	int ret;
-
-	/* require ScreenObject support for page flipping */
-	if (!dev_priv->sou_priv)
-		return -ENOSYS;
-
-	if (!vmw_kms_screen_object_flippable(dev_priv, crtc))
-		return -EINVAL;
-
-	crtc->fb = fb;
-
-	/* do a full screen dirty update */
-	clips.x1 = clips.y1 = 0;
-	clips.x2 = fb->width;
-	clips.y2 = fb->height;
-
-	if (vfb->dmabuf)
-		ret = do_dmabuf_dirty_sou(file_priv, dev_priv, vfb,
-					  0, 0, &clips, 1, 1, &fence);
-	else
-		ret = do_surface_dirty_sou(dev_priv, file_priv, vfb,
-					   0, 0, &clips, 1, 1, &fence);
-
-
-	if (ret != 0)
-		goto out_no_fence;
-	if (!fence) {
-		ret = -EINVAL;
-		goto out_no_fence;
-	}
-
-	ret = vmw_event_fence_action_queue(file_priv, fence,
-					   &event->base,
-					   &event->event.tv_sec,
-					   &event->event.tv_usec,
-					   true);
-
-	/*
-	 * No need to hold on to this now. The only cleanup
-	 * we need to do if we fail is unref the fence.
-	 */
-	vmw_fence_obj_unreference(&fence);
-
-	if (vmw_crtc_to_du(crtc)->is_implicit)
-		vmw_kms_screen_object_update_implicit_fb(dev_priv, crtc);
-
-	return ret;
-
-out_no_fence:
-	crtc->fb = old_fb;
-	return ret;
-}
-
-
-void vmw_du_crtc_save(struct drm_crtc *crtc)
-{
-}
-
-void vmw_du_crtc_restore(struct drm_crtc *crtc)
-{
-}
-
-void vmw_du_crtc_gamma_set(struct drm_crtc *crtc,
-			   u16 *r, u16 *g, u16 *b,
-			   uint32_t start, uint32_t size)
-{
-	struct vmw_private *dev_priv = vmw_priv(crtc->dev);
-	int i;
-
-	for (i = 0; i < size; i++) {
-		DRM_DEBUG("%d r/g/b = 0x%04x / 0x%04x / 0x%04x\n", i,
-			  r[i], g[i], b[i]);
-		vmw_write(dev_priv, SVGA_PALETTE_BASE + i * 3 + 0, r[i] >> 8);
-		vmw_write(dev_priv, SVGA_PALETTE_BASE + i * 3 + 1, g[i] >> 8);
-		vmw_write(dev_priv, SVGA_PALETTE_BASE + i * 3 + 2, b[i] >> 8);
-	}
-}
-
-void vmw_du_connector_dpms(struct drm_connector *connector, int mode)
-{
-}
-
-void vmw_du_connector_save(struct drm_connector *connector)
-{
-}
-
-void vmw_du_connector_restore(struct drm_connector *connector)
-{
-}
-
-enum drm_connector_status
-vmw_du_connector_detect(struct drm_connector *connector, bool force)
-{
-	uint32_t num_displays;
-	struct drm_device *dev = connector->dev;
-	struct vmw_private *dev_priv = vmw_priv(dev);
-	struct vmw_display_unit *du = vmw_connector_to_du(connector);
-
-	mutex_lock(&dev_priv->hw_mutex);
-	num_displays = vmw_read(dev_priv, SVGA_REG_NUM_DISPLAYS);
-	mutex_unlock(&dev_priv->hw_mutex);
-
-	return ((vmw_connector_to_du(connector)->unit < num_displays &&
-		 du->pref_active) ?
-		connector_status_connected : connector_status_disconnected);
-}
-
-static struct drm_display_mode vmw_kms_connector_builtin[] = {
-	/* 640x480@60Hz */
-	{ DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 25175, 640, 656,
-		   752, 800, 0, 480, 489, 492, 525, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 800x600@60Hz */
-	{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 40000, 800, 840,
-		   968, 1056, 0, 600, 601, 605, 628, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1024x768@60Hz */
-	{ DRM_MODE("1024x768", DRM_MODE_TYPE_DRIVER, 65000, 1024, 1048,
-		   1184, 1344, 0, 768, 771, 777, 806, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1152x864@75Hz */
-	{ DRM_MODE("1152x864", DRM_MODE_TYPE_DRIVER, 108000, 1152, 1216,
-		   1344, 1600, 0, 864, 865, 868, 900, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1280x768@60Hz */
-	{ DRM_MODE("1280x768", DRM_MODE_TYPE_DRIVER, 79500, 1280, 1344,
-		   1472, 1664, 0, 768, 771, 778, 798, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1280x800@60Hz */
-	{ DRM_MODE("1280x800", DRM_MODE_TYPE_DRIVER, 83500, 1280, 1352,
-		   1480, 1680, 0, 800, 803, 809, 831, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
-	/* 1280x960@60Hz */
-	{ DRM_MODE("1280x960", DRM_MODE_TYPE_DRIVER, 108000, 1280, 1376,
-		   1488, 1800, 0, 960, 961, 964, 1000, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1280x1024@60Hz */
-	{ DRM_MODE("1280x1024", DRM_MODE_TYPE_DRIVER, 108000, 1280, 1328,
-		   1440, 1688, 0, 1024, 1025, 1028, 1066, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1360x768@60Hz */
-	{ DRM_MODE("1360x768", DRM_MODE_TYPE_DRIVER, 85500, 1360, 1424,
-		   1536, 1792, 0, 768, 771, 777, 795, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1440x1050@60Hz */
-	{ DRM_MODE("1400x1050", DRM_MODE_TYPE_DRIVER, 121750, 1400, 1488,
-		   1632, 1864, 0, 1050, 1053, 1057, 1089, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1440x900@60Hz */
-	{ DRM_MODE("1440x900", DRM_MODE_TYPE_DRIVER, 106500, 1440, 1520,
-		   1672, 1904, 0, 900, 903, 909, 934, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1600x1200@60Hz */
-	{ DRM_MODE("1600x1200", DRM_MODE_TYPE_DRIVER, 162000, 1600, 1664,
-		   1856, 2160, 0, 1200, 1201, 1204, 1250, 0,
-		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1680x1050@60Hz */
-	{ DRM_MODE("1680x1050", DRM_MODE_TYPE_DRIVER, 146250, 1680, 1784,
-		   1960, 2240, 0, 1050, 1053, 1059, 1089, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1792x1344@60Hz */
-	{ DRM_MODE("1792x1344", DRM_MODE_TYPE_DRIVER, 204750, 1792, 1920,
-		   2120, 2448, 0, 1344, 1345, 1348, 1394, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1853x1392@60Hz */
-	{ DRM_MODE("1856x1392", DRM_MODE_TYPE_DRIVER, 218250, 1856, 1952,
-		   2176, 2528, 0, 1392, 1393, 1396, 1439, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1920x1200@60Hz */
-	{ DRM_MODE("1920x1200", DRM_MODE_TYPE_DRIVER, 193250, 1920, 2056,
-		   2256, 2592, 0, 1200, 1203, 1209, 1245, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 1920x1440@60Hz */
-	{ DRM_MODE("1920x1440", DRM_MODE_TYPE_DRIVER, 234000, 1920, 2048,
-		   2256, 2600, 0, 1440, 1441, 1444, 1500, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* 2560x1600@60Hz */
-	{ DRM_MODE("2560x1600", DRM_MODE_TYPE_DRIVER, 348500, 2560, 2752,
-		   3032, 3504, 0, 1600, 1603, 1609, 1658, 0,
-		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
-	/* Terminate */
-	{ DRM_MODE("", 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0) },
-};
-
-/**
- * vmw_guess_mode_timing - Provide fake timings for a
- * 60Hz vrefresh mode.
- *
- * @mode - Pointer to a struct drm_display_mode with hdisplay and vdisplay
- * members filled in.
- */
-static void vmw_guess_mode_timing(struct drm_display_mode *mode)
-{
-	mode->hsync_start = mode->hdisplay + 50;
-	mode->hsync_end = mode->hsync_start + 50;
-	mode->htotal = mode->hsync_end + 50;
-
-	mode->vsync_start = mode->vdisplay + 50;
-	mode->vsync_end = mode->vsync_start + 50;
-	mode->vtotal = mode->vsync_end + 50;
-
-	mode->clock = (u32)mode->htotal * (u32)mode->vtotal / 100 * 6;
-	mode->vrefresh = drm_mode_vrefresh(mode);
-}
-
-
-int vmw_du_connector_fill_modes(struct drm_connector *connector,
-				uint32_t max_width, uint32_t max_height)
-{
-	struct vmw_display_unit *du = vmw_connector_to_du(connector);
-	struct drm_device *dev = connector->dev;
-	struct vmw_private *dev_priv = vmw_priv(dev);
-	struct drm_display_mode *mode = NULL;
-	struct drm_display_mode *bmode;
-	struct drm_display_mode prefmode = { DRM_MODE("preferred",
-		DRM_MODE_TYPE_DRIVER | DRM_MODE_TYPE_PREFERRED,
-		0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
-		DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC)
-	};
-	int i;
-
-	/* Add preferred mode */
-	{
-		mode = drm_mode_duplicate(dev, &prefmode);
-		if (!mode)
-			return 0;
-		mode->hdisplay = du->pref_width;
-		mode->vdisplay = du->pref_height;
-		vmw_guess_mode_timing(mode);
-
-		if (vmw_kms_validate_mode_vram(dev_priv, mode->hdisplay * 2,
-					       mode->vdisplay)) {
-			drm_mode_probed_add(connector, mode);
-		} else {
-			drm_mode_destroy(dev, mode);
-			mode = NULL;
-		}
-
-		if (du->pref_mode) {
-			list_del_init(&du->pref_mode->head);
-			drm_mode_destroy(dev, du->pref_mode);
-		}
-
-		/* mode might be null here, this is intended */
-		du->pref_mode = mode;
-	}
-
-	for (i = 0; vmw_kms_connector_builtin[i].type != 0; i++) {
-		bmode = &vmw_kms_connector_builtin[i];
-		if (bmode->hdisplay > max_width ||
-		    bmode->vdisplay > max_height)
-			continue;
-
-		if (!vmw_kms_validate_mode_vram(dev_priv, bmode->hdisplay * 2,
-						bmode->vdisplay))
-			continue;
-
-		mode = drm_mode_duplicate(dev, bmode);
-		if (!mode)
-			return 0;
-		mode->vrefresh = drm_mode_vrefresh(mode);
-
-		drm_mode_probed_add(connector, mode);
-	}
-
-	/* Move the prefered mode first, help apps pick the right mode. */
-	if (du->pref_mode)
-		list_move(&du->pref_mode->head, &connector->probed_modes);
-
-	drm_mode_connector_list_update(connector);
-
-	return 1;
-}
-
-int vmw_du_connector_set_property(struct drm_connector *connector,
-				  struct drm_property *property,
-				  uint64_t val)
-{
-	return 0;
-}
-
-
 int vmw_kms_update_layout_ioctl(struct drm_device *dev, void *data,
 				struct drm_file *file_priv)
 {
@@ -1984,8 +1106,6 @@ int vmw_kms_update_layout_ioctl(struct drm_device *dev, void *data,
 	struct drm_vmw_rect *rects;
 	unsigned rects_size;
 	int ret;
-	int i;
-	struct drm_mode_config *mode_config = &dev->mode_config;
 
 	ret = ttm_read_lock(&vmaster->lock, true);
 	if (unlikely(ret != 0))
@@ -1993,13 +1113,12 @@ int vmw_kms_update_layout_ioctl(struct drm_device *dev, void *data,
 
 	if (!arg->num_outputs) {
 		struct drm_vmw_rect def_rect = {0, 0, 800, 600};
-		vmw_du_update_layout(dev_priv, 1, &def_rect);
+		vmw_kms_ldu_update_layout(dev_priv, 1, &def_rect);
 		goto out_unlock;
 	}
 
 	rects_size = arg->num_outputs * sizeof(struct drm_vmw_rect);
-	rects = kcalloc(arg->num_outputs, sizeof(struct drm_vmw_rect),
-			GFP_KERNEL);
+	rects = kzalloc(rects_size, GFP_KERNEL);
 	if (unlikely(!rects)) {
 		ret = -ENOMEM;
 		goto out_unlock;
@@ -2013,18 +1132,7 @@ int vmw_kms_update_layout_ioctl(struct drm_device *dev, void *data,
 		goto out_free;
 	}
 
-	for (i = 0; i < arg->num_outputs; ++i) {
-		if (rects[i].x < 0 ||
-		    rects[i].y < 0 ||
-		    rects[i].x + rects[i].w > mode_config->max_width ||
-		    rects[i].y + rects[i].h > mode_config->max_height) {
-			DRM_ERROR("Invalid GUI layout.\n");
-			ret = -EINVAL;
-			goto out_free;
-		}
-	}
-
-	vmw_du_update_layout(dev_priv, arg->num_outputs, rects);
+	vmw_kms_ldu_update_layout(dev_priv, arg->num_outputs, rects);
 
 out_free:
 	kfree(rects);
@@ -2032,3 +1140,15 @@ out_unlock:
 	ttm_read_unlock(&vmaster->lock);
 	return ret;
 }
+
+bool vmw_kms_validate_mode_vram(struct vmw_private *dev_priv,
+				uint32_t pitch,
+				uint32_t height)
+{
+	return ((u64) pitch * (u64) height) < (u64) dev_priv->vram_size;
+}
+
+u32 vmw_get_vblank_counter(struct drm_device *dev, int crtc)
+{
+	return 0;
+}
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h
index 8184bc5..8a398a0 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_kms.h
@@ -29,11 +29,8 @@
 #define VMWGFX_KMS_H_
 
 #include "drmP.h"
-#include "drm_crtc_helper.h"
 #include "vmwgfx_drv.h"
 
-#define VMWGFX_NUM_DISPLAY_UNITS 8
-
 
 #define vmw_framebuffer_to_vfb(x) \
 	container_of(x, struct vmw_framebuffer, base)
@@ -48,9 +45,6 @@ struct vmw_framebuffer {
 	struct drm_framebuffer base;
 	int (*pin)(struct vmw_framebuffer *fb);
 	int (*unpin)(struct vmw_framebuffer *fb);
-	bool dmabuf;
-	struct ttm_base_object *user_obj;
-	uint32_t user_handle;
 };
 
 
@@ -63,14 +57,9 @@ struct vmw_framebuffer {
 int vmw_cursor_update_image(struct vmw_private *dev_priv,
 			    u32 *image, u32 width, u32 height,
 			    u32 hotspotX, u32 hotspotY);
-int vmw_cursor_update_dmabuf(struct vmw_private *dev_priv,
-			     struct vmw_dma_buffer *dmabuf,
-			     u32 width, u32 height,
-			     u32 hotspotX, u32 hotspotY);
 void vmw_cursor_update_position(struct vmw_private *dev_priv,
 				bool show, int x, int y);
 
-
 /**
  * Base class display unit.
  *
@@ -94,73 +83,22 @@ struct vmw_display_unit {
 	int hotspot_y;
 
 	unsigned unit;
-
-	/*
-	 * Prefered mode tracking.
-	 */
-	unsigned pref_width;
-	unsigned pref_height;
-	bool pref_active;
-	struct drm_display_mode *pref_mode;
-
-	/*
-	 * Gui positioning
-	 */
-	int gui_x;
-	int gui_y;
-	bool is_implicit;
 };
 
-#define vmw_crtc_to_du(x) \
-	container_of(x, struct vmw_display_unit, crtc)
-#define vmw_connector_to_du(x) \
-	container_of(x, struct vmw_display_unit, connector)
-
-
 /*
  * Shared display unit functions - vmwgfx_kms.c
  */
 void vmw_display_unit_cleanup(struct vmw_display_unit *du);
-int vmw_du_page_flip(struct drm_crtc *crtc,
-		     struct drm_framebuffer *fb,
-		     struct drm_pending_vblank_event *event);
-void vmw_du_crtc_save(struct drm_crtc *crtc);
-void vmw_du_crtc_restore(struct drm_crtc *crtc);
-void vmw_du_crtc_gamma_set(struct drm_crtc *crtc,
-			   u16 *r, u16 *g, u16 *b,
-			   uint32_t start, uint32_t size);
 int vmw_du_crtc_cursor_set(struct drm_crtc *crtc, struct drm_file *file_priv,
 			   uint32_t handle, uint32_t width, uint32_t height);
 int vmw_du_crtc_cursor_move(struct drm_crtc *crtc, int x, int y);
-void vmw_du_connector_dpms(struct drm_connector *connector, int mode);
-void vmw_du_connector_save(struct drm_connector *connector);
-void vmw_du_connector_restore(struct drm_connector *connector);
-enum drm_connector_status
-vmw_du_connector_detect(struct drm_connector *connector, bool force);
-int vmw_du_connector_fill_modes(struct drm_connector *connector,
-				uint32_t max_width, uint32_t max_height);
-int vmw_du_connector_set_property(struct drm_connector *connector,
-				  struct drm_property *property,
-				  uint64_t val);
-
 
 /*
  * Legacy display unit functions - vmwgfx_ldu.c
  */
 int vmw_kms_init_legacy_display_system(struct vmw_private *dev_priv);
 int vmw_kms_close_legacy_display_system(struct vmw_private *dev_priv);
-
-/*
- * Screen Objects display functions - vmwgfx_scrn.c
- */
-int vmw_kms_init_screen_object_display(struct vmw_private *dev_priv);
-int vmw_kms_close_screen_object_display(struct vmw_private *dev_priv);
-int vmw_kms_sou_update_layout(struct vmw_private *dev_priv, unsigned num,
+int vmw_kms_ldu_update_layout(struct vmw_private *dev_priv, unsigned num,
 			      struct drm_vmw_rect *rects);
-bool vmw_kms_screen_object_flippable(struct vmw_private *dev_priv,
-				     struct drm_crtc *crtc);
-void vmw_kms_screen_object_update_implicit_fb(struct vmw_private *dev_priv,
-					      struct drm_crtc *crtc);
-
 
 #endif
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c
index 070fb23..b3a2cd5 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ldu.c
@@ -27,6 +27,7 @@
 
 #include "vmwgfx_kms.h"
 
+#define VMWGFX_LDU_NUM_DU 8
 
 #define vmw_crtc_to_ldu(x) \
 	container_of(x, struct vmw_legacy_display_unit, base.crtc)
@@ -50,6 +51,11 @@ struct vmw_legacy_display {
 struct vmw_legacy_display_unit {
 	struct vmw_display_unit base;
 
+	unsigned pref_width;
+	unsigned pref_height;
+	bool pref_active;
+	struct drm_display_mode *pref_mode;
+
 	struct list_head active;
 };
 
@@ -65,6 +71,20 @@ static void vmw_ldu_destroy(struct vmw_legacy_display_unit *ldu)
  * Legacy Display Unit CRTC functions
  */
 
+static void vmw_ldu_crtc_save(struct drm_crtc *crtc)
+{
+}
+
+static void vmw_ldu_crtc_restore(struct drm_crtc *crtc)
+{
+}
+
+static void vmw_ldu_crtc_gamma_set(struct drm_crtc *crtc,
+				   u16 *r, u16 *g, u16 *b,
+				   uint32_t start, uint32_t size)
+{
+}
+
 static void vmw_ldu_crtc_destroy(struct drm_crtc *crtc)
 {
 	vmw_ldu_destroy(vmw_crtc_to_ldu(crtc));
@@ -74,10 +94,9 @@ static int vmw_ldu_commit_list(struct vmw_private *dev_priv)
 {
 	struct vmw_legacy_display *lds = dev_priv->ldu_priv;
 	struct vmw_legacy_display_unit *entry;
-	struct vmw_display_unit *du = NULL;
 	struct drm_framebuffer *fb = NULL;
 	struct drm_crtc *crtc = NULL;
-	int i = 0, ret;
+	int i = 0;
 
 	/* If there is no display topology the host just assumes
 	 * that the guest will set the same layout as the host.
@@ -95,15 +114,17 @@ static int vmw_ldu_commit_list(struct vmw_private *dev_priv)
 			return 0;
 		fb = entry->base.crtc.fb;
 
-		return vmw_kms_write_svga(dev_priv, w, h, fb->pitches[0],
-					  fb->bits_per_pixel, fb->depth);
+		vmw_kms_write_svga(dev_priv, w, h, fb->pitch,
+				   fb->bits_per_pixel, fb->depth);
+
+		return 0;
 	}
 
 	if (!list_empty(&lds->active)) {
 		entry = list_entry(lds->active.next, typeof(*entry), active);
 		fb = entry->base.crtc.fb;
 
-		vmw_kms_write_svga(dev_priv, fb->width, fb->height, fb->pitches[0],
+		vmw_kms_write_svga(dev_priv, fb->width, fb->height, fb->pitch,
 				   fb->bits_per_pixel, fb->depth);
 	}
 
@@ -130,25 +151,6 @@ static int vmw_ldu_commit_list(struct vmw_private *dev_priv)
 
 	lds->last_num_active = lds->num_active;
 
-
-	/* Find the first du with a cursor. */
-	list_for_each_entry(entry, &lds->active, active) {
-		du = &entry->base;
-
-		if (!du->cursor_dmabuf)
-			continue;
-
-		ret = vmw_cursor_update_dmabuf(dev_priv,
-					       du->cursor_dmabuf,
-					       64, 64,
-					       du->hotspot_x,
-					       du->hotspot_y);
-		if (ret == 0)
-			break;
-
-		DRM_ERROR("Could not update cursor image\n");
-	}
-
 	return 0;
 }
 
@@ -263,7 +265,9 @@ static int vmw_ldu_crtc_set_config(struct drm_mode_set *set)
 
 		vmw_ldu_del_active(dev_priv, ldu);
 
-		return vmw_ldu_commit_list(dev_priv);
+		vmw_ldu_commit_list(dev_priv);
+
+		return 0;
 	}
 
 
@@ -288,20 +292,21 @@ static int vmw_ldu_crtc_set_config(struct drm_mode_set *set)
 
 	vmw_ldu_add_active(dev_priv, ldu, vfb);
 
-	return vmw_ldu_commit_list(dev_priv);
+	vmw_ldu_commit_list(dev_priv);
+
+	return 0;
 }
 
 static struct drm_crtc_funcs vmw_legacy_crtc_funcs = {
-	.save = vmw_du_crtc_save,
-	.restore = vmw_du_crtc_restore,
+	.save = vmw_ldu_crtc_save,
+	.restore = vmw_ldu_crtc_restore,
 	.cursor_set = vmw_du_crtc_cursor_set,
 	.cursor_move = vmw_du_crtc_cursor_move,
-	.gamma_set = vmw_du_crtc_gamma_set,
+	.gamma_set = vmw_ldu_crtc_gamma_set,
 	.destroy = vmw_ldu_crtc_destroy,
 	.set_config = vmw_ldu_crtc_set_config,
 };
 
-
 /*
  * Legacy Display Unit encoder functions
  */
@@ -319,18 +324,183 @@ static struct drm_encoder_funcs vmw_legacy_encoder_funcs = {
  * Legacy Display Unit connector functions
  */
 
+static void vmw_ldu_connector_dpms(struct drm_connector *connector, int mode)
+{
+}
+
+static void vmw_ldu_connector_save(struct drm_connector *connector)
+{
+}
+
+static void vmw_ldu_connector_restore(struct drm_connector *connector)
+{
+}
+
+static enum drm_connector_status
+	vmw_ldu_connector_detect(struct drm_connector *connector,
+				 bool force)
+{
+	if (vmw_connector_to_ldu(connector)->pref_active)
+		return connector_status_connected;
+	return connector_status_disconnected;
+}
+
+static const struct drm_display_mode vmw_ldu_connector_builtin[] = {
+	/* 640x480@60Hz */
+	{ DRM_MODE("640x480", DRM_MODE_TYPE_DRIVER, 25175, 640, 656,
+		   752, 800, 0, 480, 489, 492, 525, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 800x600@60Hz */
+	{ DRM_MODE("800x600", DRM_MODE_TYPE_DRIVER, 40000, 800, 840,
+		   968, 1056, 0, 600, 601, 605, 628, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1024x768@60Hz */
+	{ DRM_MODE("1024x768", DRM_MODE_TYPE_DRIVER, 65000, 1024, 1048,
+		   1184, 1344, 0, 768, 771, 777, 806, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1152x864@75Hz */
+	{ DRM_MODE("1152x864", DRM_MODE_TYPE_DRIVER, 108000, 1152, 1216,
+		   1344, 1600, 0, 864, 865, 868, 900, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1280x768@60Hz */
+	{ DRM_MODE("1280x768", DRM_MODE_TYPE_DRIVER, 79500, 1280, 1344,
+		   1472, 1664, 0, 768, 771, 778, 798, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1280x800@60Hz */
+	{ DRM_MODE("1280x800", DRM_MODE_TYPE_DRIVER, 83500, 1280, 1352,
+		   1480, 1680, 0, 800, 803, 809, 831, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_NVSYNC) },
+	/* 1280x960@60Hz */
+	{ DRM_MODE("1280x960", DRM_MODE_TYPE_DRIVER, 108000, 1280, 1376,
+		   1488, 1800, 0, 960, 961, 964, 1000, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1280x1024@60Hz */
+	{ DRM_MODE("1280x1024", DRM_MODE_TYPE_DRIVER, 108000, 1280, 1328,
+		   1440, 1688, 0, 1024, 1025, 1028, 1066, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1360x768@60Hz */
+	{ DRM_MODE("1360x768", DRM_MODE_TYPE_DRIVER, 85500, 1360, 1424,
+		   1536, 1792, 0, 768, 771, 777, 795, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1440x1050@60Hz */
+	{ DRM_MODE("1400x1050", DRM_MODE_TYPE_DRIVER, 121750, 1400, 1488,
+		   1632, 1864, 0, 1050, 1053, 1057, 1089, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1440x900@60Hz */
+	{ DRM_MODE("1440x900", DRM_MODE_TYPE_DRIVER, 106500, 1440, 1520,
+		   1672, 1904, 0, 900, 903, 909, 934, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1600x1200@60Hz */
+	{ DRM_MODE("1600x1200", DRM_MODE_TYPE_DRIVER, 162000, 1600, 1664,
+		   1856, 2160, 0, 1200, 1201, 1204, 1250, 0,
+		   DRM_MODE_FLAG_PHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1680x1050@60Hz */
+	{ DRM_MODE("1680x1050", DRM_MODE_TYPE_DRIVER, 146250, 1680, 1784,
+		   1960, 2240, 0, 1050, 1053, 1059, 1089, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1792x1344@60Hz */
+	{ DRM_MODE("1792x1344", DRM_MODE_TYPE_DRIVER, 204750, 1792, 1920,
+		   2120, 2448, 0, 1344, 1345, 1348, 1394, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1853x1392@60Hz */
+	{ DRM_MODE("1856x1392", DRM_MODE_TYPE_DRIVER, 218250, 1856, 1952,
+		   2176, 2528, 0, 1392, 1393, 1396, 1439, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1920x1200@60Hz */
+	{ DRM_MODE("1920x1200", DRM_MODE_TYPE_DRIVER, 193250, 1920, 2056,
+		   2256, 2592, 0, 1200, 1203, 1209, 1245, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 1920x1440@60Hz */
+	{ DRM_MODE("1920x1440", DRM_MODE_TYPE_DRIVER, 234000, 1920, 2048,
+		   2256, 2600, 0, 1440, 1441, 1444, 1500, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* 2560x1600@60Hz */
+	{ DRM_MODE("2560x1600", DRM_MODE_TYPE_DRIVER, 348500, 2560, 2752,
+		   3032, 3504, 0, 1600, 1603, 1609, 1658, 0,
+		   DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC) },
+	/* Terminate */
+	{ DRM_MODE("", 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0) },
+};
+
+static int vmw_ldu_connector_fill_modes(struct drm_connector *connector,
+					uint32_t max_width, uint32_t max_height)
+{
+	struct vmw_legacy_display_unit *ldu = vmw_connector_to_ldu(connector);
+	struct drm_device *dev = connector->dev;
+	struct vmw_private *dev_priv = vmw_priv(dev);
+	struct drm_display_mode *mode = NULL;
+	struct drm_display_mode prefmode = { DRM_MODE("preferred",
+		DRM_MODE_TYPE_DRIVER | DRM_MODE_TYPE_PREFERRED,
+		0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
+		DRM_MODE_FLAG_NHSYNC | DRM_MODE_FLAG_PVSYNC)
+	};
+	int i;
+
+	/* Add preferred mode */
+	{
+		mode = drm_mode_duplicate(dev, &prefmode);
+		if (!mode)
+			return 0;
+		mode->hdisplay = ldu->pref_width;
+		mode->vdisplay = ldu->pref_height;
+		mode->vrefresh = drm_mode_vrefresh(mode);
+		if (vmw_kms_validate_mode_vram(dev_priv, mode->hdisplay * 2,
+					       mode->vdisplay)) {
+			drm_mode_probed_add(connector, mode);
+
+			if (ldu->pref_mode) {
+				list_del_init(&ldu->pref_mode->head);
+				drm_mode_destroy(dev, ldu->pref_mode);
+			}
+
+			ldu->pref_mode = mode;
+		}
+	}
+
+	for (i = 0; vmw_ldu_connector_builtin[i].type != 0; i++) {
+		const struct drm_display_mode *bmode;
+
+		bmode = &vmw_ldu_connector_builtin[i];
+		if (bmode->hdisplay > max_width ||
+		    bmode->vdisplay > max_height)
+			continue;
+
+		if (!vmw_kms_validate_mode_vram(dev_priv, bmode->hdisplay * 2,
+						bmode->vdisplay))
+			continue;
+
+		mode = drm_mode_duplicate(dev, bmode);
+		if (!mode)
+			return 0;
+		mode->vrefresh = drm_mode_vrefresh(mode);
+
+		drm_mode_probed_add(connector, mode);
+	}
+
+	drm_mode_connector_list_update(connector);
+
+	return 1;
+}
+
+static int vmw_ldu_connector_set_property(struct drm_connector *connector,
+					  struct drm_property *property,
+					  uint64_t val)
+{
+	return 0;
+}
+
 static void vmw_ldu_connector_destroy(struct drm_connector *connector)
 {
 	vmw_ldu_destroy(vmw_connector_to_ldu(connector));
 }
 
 static struct drm_connector_funcs vmw_legacy_connector_funcs = {
-	.dpms = vmw_du_connector_dpms,
-	.save = vmw_du_connector_save,
-	.restore = vmw_du_connector_restore,
-	.detect = vmw_du_connector_detect,
-	.fill_modes = vmw_du_connector_fill_modes,
-	.set_property = vmw_du_connector_set_property,
+	.dpms = vmw_ldu_connector_dpms,
+	.save = vmw_ldu_connector_save,
+	.restore = vmw_ldu_connector_restore,
+	.detect = vmw_ldu_connector_detect,
+	.fill_modes = vmw_ldu_connector_fill_modes,
+	.set_property = vmw_ldu_connector_set_property,
 	.destroy = vmw_ldu_connector_destroy,
 };
 
@@ -353,26 +523,23 @@ static int vmw_ldu_init(struct vmw_private *dev_priv, unsigned unit)
 
 	INIT_LIST_HEAD(&ldu->active);
 
-	ldu->base.pref_active = (unit == 0);
-	ldu->base.pref_width = dev_priv->initial_width;
-	ldu->base.pref_height = dev_priv->initial_height;
-	ldu->base.pref_mode = NULL;
-	ldu->base.is_implicit = true;
+	ldu->pref_active = (unit == 0);
+	ldu->pref_width = 800;
+	ldu->pref_height = 600;
+	ldu->pref_mode = NULL;
 
 	drm_connector_init(dev, connector, &vmw_legacy_connector_funcs,
-			   DRM_MODE_CONNECTOR_VIRTUAL);
-	connector->status = vmw_du_connector_detect(connector, true);
+			   DRM_MODE_CONNECTOR_LVDS);
+	connector->status = vmw_ldu_connector_detect(connector, true);
 
 	drm_encoder_init(dev, encoder, &vmw_legacy_encoder_funcs,
-			 DRM_MODE_ENCODER_VIRTUAL);
+			 DRM_MODE_ENCODER_LVDS);
 	drm_mode_connector_attach_encoder(connector, encoder);
 	encoder->possible_crtcs = (1 << unit);
 	encoder->possible_clones = 0;
 
 	drm_crtc_init(dev, crtc, &vmw_legacy_crtc_funcs);
 
-	drm_mode_crtc_set_gamma_size(crtc, 256);
-
 	drm_connector_attach_property(connector,
 				      dev->mode_config.dirty_info_property,
 				      1);
@@ -383,7 +550,8 @@ static int vmw_ldu_init(struct vmw_private *dev_priv, unsigned unit)
 int vmw_kms_init_legacy_display_system(struct vmw_private *dev_priv)
 {
 	struct drm_device *dev = dev_priv->dev;
-	int i, ret;
+	int i;
+	int ret;
 
 	if (dev_priv->ldu_priv) {
 		DRM_INFO("ldu system already on\n");
@@ -391,6 +559,7 @@ int vmw_kms_init_legacy_display_system(struct vmw_private *dev_priv)
 	}
 
 	dev_priv->ldu_priv = kmalloc(sizeof(*dev_priv->ldu_priv), GFP_KERNEL);
+
 	if (!dev_priv->ldu_priv)
 		return -ENOMEM;
 
@@ -399,31 +568,18 @@ int vmw_kms_init_legacy_display_system(struct vmw_private *dev_priv)
 	dev_priv->ldu_priv->last_num_active = 0;
 	dev_priv->ldu_priv->fb = NULL;
 
-	/* for old hardware without multimon only enable one display */
-	if (dev_priv->capabilities & SVGA_CAP_MULTIMON)
-		ret = drm_vblank_init(dev, VMWGFX_NUM_DISPLAY_UNITS);
-	else
-		ret = drm_vblank_init(dev, 1);
-	if (ret != 0)
-		goto err_free;
-
-	ret = drm_mode_create_dirty_info_property(dev);
-	if (ret != 0)
-		goto err_vblank_cleanup;
+	drm_mode_create_dirty_info_property(dev_priv->dev);
 
-	if (dev_priv->capabilities & SVGA_CAP_MULTIMON)
-		for (i = 0; i < VMWGFX_NUM_DISPLAY_UNITS; ++i)
+	if (dev_priv->capabilities & SVGA_CAP_MULTIMON) {
+		for (i = 0; i < VMWGFX_LDU_NUM_DU; ++i)
 			vmw_ldu_init(dev_priv, i);
-	else
+		ret = drm_vblank_init(dev, VMWGFX_LDU_NUM_DU);
+	} else {
+		/* for old hardware without multimon only enable one display */
 		vmw_ldu_init(dev_priv, 0);
+		ret = drm_vblank_init(dev, 1);
+	}
 
-	return 0;
-
-err_vblank_cleanup:
-	drm_vblank_cleanup(dev);
-err_free:
-	kfree(dev_priv->ldu_priv);
-	dev_priv->ldu_priv = NULL;
 	return ret;
 }
 
@@ -431,14 +587,52 @@ int vmw_kms_close_legacy_display_system(struct vmw_private *dev_priv)
 {
 	struct drm_device *dev = dev_priv->dev;
 
+	drm_vblank_cleanup(dev);
 	if (!dev_priv->ldu_priv)
 		return -ENOSYS;
 
-	drm_vblank_cleanup(dev);
-
 	BUG_ON(!list_empty(&dev_priv->ldu_priv->active));
 
 	kfree(dev_priv->ldu_priv);
 
 	return 0;
 }
+
+int vmw_kms_ldu_update_layout(struct vmw_private *dev_priv, unsigned num,
+			      struct drm_vmw_rect *rects)
+{
+	struct drm_device *dev = dev_priv->dev;
+	struct vmw_legacy_display_unit *ldu;
+	struct drm_connector *con;
+	int i;
+
+	mutex_lock(&dev->mode_config.mutex);
+
+#if 0
+	DRM_INFO("%s: new layout ", __func__);
+	for (i = 0; i < (int)num; i++)
+		DRM_INFO("(%i, %i %ux%u) ", rects[i].x, rects[i].y,
+			 rects[i].w, rects[i].h);
+	DRM_INFO("\n");
+#else
+	(void)i;
+#endif
+
+	list_for_each_entry(con, &dev->mode_config.connector_list, head) {
+		ldu = vmw_connector_to_ldu(con);
+		if (num > ldu->base.unit) {
+			ldu->pref_width = rects[ldu->base.unit].w;
+			ldu->pref_height = rects[ldu->base.unit].h;
+			ldu->pref_active = true;
+		} else {
+			ldu->pref_width = 800;
+			ldu->pref_height = 600;
+			ldu->pref_active = false;
+		}
+		con->status = vmw_ldu_connector_detect(con, true);
+	}
+
+	mutex_unlock(&dev->mode_config.mutex);
+
+	return 0;
+}
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_marker.c b/drivers/gpu/drm/vmwgfx/vmwgfx_marker.c
deleted file mode 100644
index 8a8725c..0000000
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_marker.c
+++ /dev/null
@@ -1,171 +0,0 @@
-/**************************************************************************
- *
- * Copyright (C) 2010 VMware, Inc., Palo Alto, CA., USA
- * All Rights Reserved.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the
- * "Software"), to deal in the Software without restriction, including
- * without limitation the rights to use, copy, modify, merge, publish,
- * distribute, sub license, and/or sell copies of the Software, and to
- * permit persons to whom the Software is furnished to do so, subject to
- * the following conditions:
- *
- * The above copyright notice and this permission notice (including the
- * next paragraph) shall be included in all copies or substantial portions
- * of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
- * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
- * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
- * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
- * USE OR OTHER DEALINGS IN THE SOFTWARE.
- *
- **************************************************************************/
-
-
-#include "vmwgfx_drv.h"
-
-struct vmw_marker {
-	struct list_head head;
-	uint32_t seqno;
-	struct timespec submitted;
-};
-
-void vmw_marker_queue_init(struct vmw_marker_queue *queue)
-{
-	INIT_LIST_HEAD(&queue->head);
-	queue->lag = ns_to_timespec(0);
-	getrawmonotonic(&queue->lag_time);
-	spin_lock_init(&queue->lock);
-}
-
-void vmw_marker_queue_takedown(struct vmw_marker_queue *queue)
-{
-	struct vmw_marker *marker, *next;
-
-	spin_lock(&queue->lock);
-	list_for_each_entry_safe(marker, next, &queue->head, head) {
-		kfree(marker);
-	}
-	spin_unlock(&queue->lock);
-}
-
-int vmw_marker_push(struct vmw_marker_queue *queue,
-		   uint32_t seqno)
-{
-	struct vmw_marker *marker = kmalloc(sizeof(*marker), GFP_KERNEL);
-
-	if (unlikely(!marker))
-		return -ENOMEM;
-
-	marker->seqno = seqno;
-	getrawmonotonic(&marker->submitted);
-	spin_lock(&queue->lock);
-	list_add_tail(&marker->head, &queue->head);
-	spin_unlock(&queue->lock);
-
-	return 0;
-}
-
-int vmw_marker_pull(struct vmw_marker_queue *queue,
-		   uint32_t signaled_seqno)
-{
-	struct vmw_marker *marker, *next;
-	struct timespec now;
-	bool updated = false;
-
-	spin_lock(&queue->lock);
-	getrawmonotonic(&now);
-
-	if (list_empty(&queue->head)) {
-		queue->lag = ns_to_timespec(0);
-		queue->lag_time = now;
-		updated = true;
-		goto out_unlock;
-	}
-
-	list_for_each_entry_safe(marker, next, &queue->head, head) {
-		if (signaled_seqno - marker->seqno > (1 << 30))
-			continue;
-
-		queue->lag = timespec_sub(now, marker->submitted);
-		queue->lag_time = now;
-		updated = true;
-		list_del(&marker->head);
-		kfree(marker);
-	}
-
-out_unlock:
-	spin_unlock(&queue->lock);
-
-	return (updated) ? 0 : -EBUSY;
-}
-
-static struct timespec vmw_timespec_add(struct timespec t1,
-					struct timespec t2)
-{
-	t1.tv_sec += t2.tv_sec;
-	t1.tv_nsec += t2.tv_nsec;
-	if (t1.tv_nsec >= 1000000000L) {
-		t1.tv_sec += 1;
-		t1.tv_nsec -= 1000000000L;
-	}
-
-	return t1;
-}
-
-static struct timespec vmw_fifo_lag(struct vmw_marker_queue *queue)
-{
-	struct timespec now;
-
-	spin_lock(&queue->lock);
-	getrawmonotonic(&now);
-	queue->lag = vmw_timespec_add(queue->lag,
-				      timespec_sub(now, queue->lag_time));
-	queue->lag_time = now;
-	spin_unlock(&queue->lock);
-	return queue->lag;
-}
-
-
-static bool vmw_lag_lt(struct vmw_marker_queue *queue,
-		       uint32_t us)
-{
-	struct timespec lag, cond;
-
-	cond = ns_to_timespec((s64) us * 1000);
-	lag = vmw_fifo_lag(queue);
-	return (timespec_compare(&lag, &cond) < 1);
-}
-
-int vmw_wait_lag(struct vmw_private *dev_priv,
-		 struct vmw_marker_queue *queue, uint32_t us)
-{
-	struct vmw_marker *marker;
-	uint32_t seqno;
-	int ret;
-
-	while (!vmw_lag_lt(queue, us)) {
-		spin_lock(&queue->lock);
-		if (list_empty(&queue->head))
-			seqno = atomic_read(&dev_priv->marker_seq);
-		else {
-			marker = list_first_entry(&queue->head,
-						 struct vmw_marker, head);
-			seqno = marker->seqno;
-		}
-		spin_unlock(&queue->lock);
-
-		ret = vmw_wait_seqno(dev_priv, false, seqno, true,
-					3*HZ);
-
-		if (unlikely(ret != 0))
-			return ret;
-
-		(void) vmw_marker_pull(queue, seqno);
-	}
-	return 0;
-}
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_overlay.c b/drivers/gpu/drm/vmwgfx/vmwgfx_overlay.c
index 14399ee..f1a52f9 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_overlay.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_overlay.c
@@ -87,6 +87,48 @@ static inline void fill_flush(struct vmw_escape_video_flush *cmd,
 }
 
 /**
+ * Pin or unpin a buffer in vram.
+ *
+ * @dev_priv:  Driver private.
+ * @buf:  DMA buffer to pin or unpin.
+ * @pin:  Pin buffer in vram if true.
+ * @interruptible:  Use interruptible wait.
+ *
+ * Takes the current masters ttm lock in read.
+ *
+ * Returns
+ * -ERESTARTSYS if interrupted by a signal.
+ */
+static int vmw_dmabuf_pin_in_vram(struct vmw_private *dev_priv,
+				  struct vmw_dma_buffer *buf,
+				  bool pin, bool interruptible)
+{
+	struct ttm_buffer_object *bo = &buf->base;
+	struct ttm_placement *overlay_placement = &vmw_vram_placement;
+	int ret;
+
+	ret = ttm_read_lock(&dev_priv->active_master->lock, interruptible);
+	if (unlikely(ret != 0))
+		return ret;
+
+	ret = ttm_bo_reserve(bo, interruptible, false, false, 0);
+	if (unlikely(ret != 0))
+		goto err;
+
+	if (pin)
+		overlay_placement = &vmw_vram_ne_placement;
+
+	ret = ttm_bo_validate(bo, overlay_placement, interruptible, false, false);
+
+	ttm_bo_unreserve(bo);
+
+err:
+	ttm_read_unlock(&dev_priv->active_master->lock);
+
+	return ret;
+}
+
+/**
  * Send put command to hw.
  *
  * Returns
@@ -97,80 +139,68 @@ static int vmw_overlay_send_put(struct vmw_private *dev_priv,
 				struct drm_vmw_control_stream_arg *arg,
 				bool interruptible)
 {
-	struct vmw_escape_video_flush *flush;
-	size_t fifo_size;
-	bool have_so = dev_priv->sou_priv ? true : false;
-	int i, num_items;
-	SVGAGuestPtr ptr;
-
 	struct {
 		struct vmw_escape_header escape;
 		struct {
-			uint32_t cmdType;
-			uint32_t streamId;
-		} header;
+			struct {
+				uint32_t cmdType;
+				uint32_t streamId;
+			} header;
+			struct {
+				uint32_t registerId;
+				uint32_t value;
+			} items[SVGA_VIDEO_PITCH_3 + 1];
+		} body;
+		struct vmw_escape_video_flush flush;
 	} *cmds;
-	struct {
-		uint32_t registerId;
-		uint32_t value;
-	} *items;
-
-	/* defines are a index needs + 1 */
-	if (have_so)
-		num_items = SVGA_VIDEO_DST_SCREEN_ID + 1;
-	else
-		num_items = SVGA_VIDEO_PITCH_3 + 1;
-
-	fifo_size = sizeof(*cmds) + sizeof(*flush) + sizeof(*items) * num_items;
+	uint32_t offset;
+	int i, ret;
 
-	cmds = vmw_fifo_reserve(dev_priv, fifo_size);
-	/* hardware has hung, can't do anything here */
-	if (!cmds)
-		return -ENOMEM;
+	for (;;) {
+		cmds = vmw_fifo_reserve(dev_priv, sizeof(*cmds));
+		if (cmds)
+			break;
 
-	items = (typeof(items))&cmds[1];
-	flush = (struct vmw_escape_video_flush *)&items[num_items];
-
-	/* the size is header + number of items */
-	fill_escape(&cmds->escape, sizeof(*items) * (num_items + 1));
-
-	cmds->header.cmdType = SVGA_ESCAPE_VMWARE_VIDEO_SET_REGS;
-	cmds->header.streamId = arg->stream_id;
-
-	/* the IDs are neatly numbered */
-	for (i = 0; i < num_items; i++)
-		items[i].registerId = i;
-
-	vmw_bo_get_guest_ptr(&buf->base, &ptr);
-	ptr.offset += arg->offset;
-
-	items[SVGA_VIDEO_ENABLED].value     = true;
-	items[SVGA_VIDEO_FLAGS].value       = arg->flags;
-	items[SVGA_VIDEO_DATA_OFFSET].value = ptr.offset;
-	items[SVGA_VIDEO_FORMAT].value      = arg->format;
-	items[SVGA_VIDEO_COLORKEY].value    = arg->color_key;
-	items[SVGA_VIDEO_SIZE].value        = arg->size;
-	items[SVGA_VIDEO_WIDTH].value       = arg->width;
-	items[SVGA_VIDEO_HEIGHT].value      = arg->height;
-	items[SVGA_VIDEO_SRC_X].value       = arg->src.x;
-	items[SVGA_VIDEO_SRC_Y].value       = arg->src.y;
-	items[SVGA_VIDEO_SRC_WIDTH].value   = arg->src.w;
-	items[SVGA_VIDEO_SRC_HEIGHT].value  = arg->src.h;
-	items[SVGA_VIDEO_DST_X].value       = arg->dst.x;
-	items[SVGA_VIDEO_DST_Y].value       = arg->dst.y;
-	items[SVGA_VIDEO_DST_WIDTH].value   = arg->dst.w;
-	items[SVGA_VIDEO_DST_HEIGHT].value  = arg->dst.h;
-	items[SVGA_VIDEO_PITCH_1].value     = arg->pitch[0];
-	items[SVGA_VIDEO_PITCH_2].value     = arg->pitch[1];
-	items[SVGA_VIDEO_PITCH_3].value     = arg->pitch[2];
-	if (have_so) {
-		items[SVGA_VIDEO_DATA_GMRID].value    = ptr.gmrId;
-		items[SVGA_VIDEO_DST_SCREEN_ID].value = SVGA_ID_INVALID;
+		ret = vmw_fallback_wait(dev_priv, false, true, 0,
+					interruptible, 3*HZ);
+		if (interruptible && ret == -ERESTARTSYS)
+			return ret;
+		else
+			BUG_ON(ret != 0);
 	}
 
-	fill_flush(flush, arg->stream_id);
+	fill_escape(&cmds->escape, sizeof(cmds->body));
+	cmds->body.header.cmdType = SVGA_ESCAPE_VMWARE_VIDEO_SET_REGS;
+	cmds->body.header.streamId = arg->stream_id;
+
+	for (i = 0; i <= SVGA_VIDEO_PITCH_3; i++)
+		cmds->body.items[i].registerId = i;
+
+	offset = buf->base.offset + arg->offset;
+
+	cmds->body.items[SVGA_VIDEO_ENABLED].value     = true;
+	cmds->body.items[SVGA_VIDEO_FLAGS].value       = arg->flags;
+	cmds->body.items[SVGA_VIDEO_DATA_OFFSET].value = offset;
+	cmds->body.items[SVGA_VIDEO_FORMAT].value      = arg->format;
+	cmds->body.items[SVGA_VIDEO_COLORKEY].value    = arg->color_key;
+	cmds->body.items[SVGA_VIDEO_SIZE].value        = arg->size;
+	cmds->body.items[SVGA_VIDEO_WIDTH].value       = arg->width;
+	cmds->body.items[SVGA_VIDEO_HEIGHT].value      = arg->height;
+	cmds->body.items[SVGA_VIDEO_SRC_X].value       = arg->src.x;
+	cmds->body.items[SVGA_VIDEO_SRC_Y].value       = arg->src.y;
+	cmds->body.items[SVGA_VIDEO_SRC_WIDTH].value   = arg->src.w;
+	cmds->body.items[SVGA_VIDEO_SRC_HEIGHT].value  = arg->src.h;
+	cmds->body.items[SVGA_VIDEO_DST_X].value       = arg->dst.x;
+	cmds->body.items[SVGA_VIDEO_DST_Y].value       = arg->dst.y;
+	cmds->body.items[SVGA_VIDEO_DST_WIDTH].value   = arg->dst.w;
+	cmds->body.items[SVGA_VIDEO_DST_HEIGHT].value  = arg->dst.h;
+	cmds->body.items[SVGA_VIDEO_PITCH_1].value     = arg->pitch[0];
+	cmds->body.items[SVGA_VIDEO_PITCH_2].value     = arg->pitch[1];
+	cmds->body.items[SVGA_VIDEO_PITCH_3].value     = arg->pitch[2];
+
+	fill_flush(&cmds->flush, arg->stream_id);
 
-	vmw_fifo_commit(dev_priv, fifo_size);
+	vmw_fifo_commit(dev_priv, sizeof(*cmds));
 
 	return 0;
 }
@@ -218,25 +248,6 @@ static int vmw_overlay_send_stop(struct vmw_private *dev_priv,
 }
 
 /**
- * Move a buffer to vram or gmr if @pin is set, else unpin the buffer.
- *
- * With the introduction of screen objects buffers could now be
- * used with GMRs instead of being locked to vram.
- */
-static int vmw_overlay_move_buffer(struct vmw_private *dev_priv,
-				   struct vmw_dma_buffer *buf,
-				   bool pin, bool inter)
-{
-	if (!pin)
-		return vmw_dmabuf_unpin(dev_priv, buf, inter);
-
-	if (!dev_priv->sou_priv)
-		return vmw_dmabuf_to_vram(dev_priv, buf, true, inter);
-
-	return vmw_dmabuf_to_vram_or_gmr(dev_priv, buf, true, inter);
-}
-
-/**
  * Stop or pause a stream.
  *
  * If the stream is paused the no evict flag is removed from the buffer
@@ -268,8 +279,8 @@ static int vmw_overlay_stop(struct vmw_private *dev_priv,
 			return ret;
 
 		/* We just remove the NO_EVICT flag so no -ENOMEM */
-		ret = vmw_overlay_move_buffer(dev_priv, stream->buf, false,
-					      interruptible);
+		ret = vmw_dmabuf_pin_in_vram(dev_priv, stream->buf, false,
+					     interruptible);
 		if (interruptible && ret == -ERESTARTSYS)
 			return ret;
 		else
@@ -331,7 +342,7 @@ static int vmw_overlay_update_stream(struct vmw_private *dev_priv,
 	/* We don't start the old stream if we are interrupted.
 	 * Might return -ENOMEM if it can't fit the buffer in vram.
 	 */
-	ret = vmw_overlay_move_buffer(dev_priv, buf, true, interruptible);
+	ret = vmw_dmabuf_pin_in_vram(dev_priv, buf, true, interruptible);
 	if (ret)
 		return ret;
 
@@ -340,8 +351,7 @@ static int vmw_overlay_update_stream(struct vmw_private *dev_priv,
 		/* This one needs to happen no matter what. We only remove
 		 * the NO_EVICT flag so this is safe from -ENOMEM.
 		 */
-		BUG_ON(vmw_overlay_move_buffer(dev_priv, buf, false, false)
-		       != 0);
+		BUG_ON(vmw_dmabuf_pin_in_vram(dev_priv, buf, false, false) != 0);
 		return ret;
 	}
 
@@ -575,10 +585,11 @@ int vmw_overlay_init(struct vmw_private *dev_priv)
 		return -ENOSYS;
 	}
 
-	overlay = kzalloc(sizeof(*overlay), GFP_KERNEL);
+	overlay = kmalloc(sizeof(*overlay), GFP_KERNEL);
 	if (!overlay)
 		return -ENOMEM;
 
+	memset(overlay, 0, sizeof(*overlay));
 	mutex_init(&overlay->mutex);
 	for (i = 0; i < VMW_MAX_NUM_STREAMS; i++) {
 		overlay->stream[i].buf = NULL;
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
index a37abb5..5408b1b 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_resource.c
@@ -31,6 +31,10 @@
 #include "ttm/ttm_placement.h"
 #include "drmP.h"
 
+#define VMW_RES_CONTEXT ttm_driver_type0
+#define VMW_RES_SURFACE ttm_driver_type1
+#define VMW_RES_STREAM ttm_driver_type2
+
 struct vmw_user_context {
 	struct ttm_base_object base;
 	struct vmw_resource res;
@@ -39,7 +43,6 @@ struct vmw_user_context {
 struct vmw_user_surface {
 	struct ttm_base_object base;
 	struct vmw_surface srf;
-	uint32_t size;
 };
 
 struct vmw_user_dma_buffer {
@@ -62,17 +65,6 @@ struct vmw_user_stream {
 	struct vmw_stream stream;
 };
 
-struct vmw_surface_offset {
-	uint32_t face;
-	uint32_t mip;
-	uint32_t bo_offset;
-};
-
-
-static uint64_t vmw_user_context_size;
-static uint64_t vmw_user_surface_size;
-static uint64_t vmw_user_stream_size;
-
 static inline struct vmw_dma_buffer *
 vmw_dma_buffer(struct ttm_buffer_object *bo)
 {
@@ -92,36 +84,13 @@ struct vmw_resource *vmw_resource_reference(struct vmw_resource *res)
 	return res;
 }
 
-
-/**
- * vmw_resource_release_id - release a resource id to the id manager.
- *
- * @res: Pointer to the resource.
- *
- * Release the resource id to the resource id manager and set it to -1
- */
-static void vmw_resource_release_id(struct vmw_resource *res)
-{
-	struct vmw_private *dev_priv = res->dev_priv;
-
-	write_lock(&dev_priv->resource_lock);
-	if (res->id != -1)
-		idr_remove(res->idr, res->id);
-	res->id = -1;
-	write_unlock(&dev_priv->resource_lock);
-}
-
 static void vmw_resource_release(struct kref *kref)
 {
 	struct vmw_resource *res =
 	    container_of(kref, struct vmw_resource, kref);
 	struct vmw_private *dev_priv = res->dev_priv;
-	int id = res->id;
-	struct idr *idr = res->idr;
 
-	res->avail = false;
-	if (res->remove_from_lists != NULL)
-		res->remove_from_lists(res);
+	idr_remove(res->idr, res->id);
 	write_unlock(&dev_priv->resource_lock);
 
 	if (likely(res->hw_destroy != NULL))
@@ -133,9 +102,6 @@ static void vmw_resource_release(struct kref *kref)
 		kfree(res);
 
 	write_lock(&dev_priv->resource_lock);
-
-	if (id != -1)
-		idr_remove(idr, id);
 }
 
 void vmw_resource_unreference(struct vmw_resource **p_res)
@@ -149,29 +115,28 @@ void vmw_resource_unreference(struct vmw_resource **p_res)
 	write_unlock(&dev_priv->resource_lock);
 }
 
-
-/**
- * vmw_resource_alloc_id - release a resource id to the id manager.
- *
- * @dev_priv: Pointer to the device private structure.
- * @res: Pointer to the resource.
- *
- * Allocate the lowest free resource from the resource manager, and set
- * @res->id to that id. Returns 0 on success and -ENOMEM on failure.
- */
-static int vmw_resource_alloc_id(struct vmw_private *dev_priv,
-				 struct vmw_resource *res)
+static int vmw_resource_init(struct vmw_private *dev_priv,
+			     struct vmw_resource *res,
+			     struct idr *idr,
+			     enum ttm_object_type obj_type,
+			     void (*res_free) (struct vmw_resource *res))
 {
 	int ret;
 
-	BUG_ON(res->id != -1);
+	kref_init(&res->kref);
+	res->hw_destroy = NULL;
+	res->res_free = res_free;
+	res->res_type = obj_type;
+	res->idr = idr;
+	res->avail = false;
+	res->dev_priv = dev_priv;
 
 	do {
-		if (unlikely(idr_pre_get(res->idr, GFP_KERNEL) == 0))
+		if (unlikely(idr_pre_get(idr, GFP_KERNEL) == 0))
 			return -ENOMEM;
 
 		write_lock(&dev_priv->resource_lock);
-		ret = idr_get_new_above(res->idr, res, 1, &res->id);
+		ret = idr_get_new_above(idr, res, 1, &res->id);
 		write_unlock(&dev_priv->resource_lock);
 
 	} while (ret == -EAGAIN);
@@ -179,33 +144,6 @@ static int vmw_resource_alloc_id(struct vmw_private *dev_priv,
 	return ret;
 }
 
-
-static int vmw_resource_init(struct vmw_private *dev_priv,
-			     struct vmw_resource *res,
-			     struct idr *idr,
-			     enum ttm_object_type obj_type,
-			     bool delay_id,
-			     void (*res_free) (struct vmw_resource *res),
-			     void (*remove_from_lists)
-			     (struct vmw_resource *res))
-{
-	kref_init(&res->kref);
-	res->hw_destroy = NULL;
-	res->res_free = res_free;
-	res->remove_from_lists = remove_from_lists;
-	res->res_type = obj_type;
-	res->idr = idr;
-	res->avail = false;
-	res->dev_priv = dev_priv;
-	INIT_LIST_HEAD(&res->query_head);
-	INIT_LIST_HEAD(&res->validate_head);
-	res->id = -1;
-	if (delay_id)
-		return 0;
-	else
-		return vmw_resource_alloc_id(dev_priv, res);
-}
-
 /**
  * vmw_resource_activate
  *
@@ -260,12 +198,8 @@ static void vmw_hw_context_destroy(struct vmw_resource *res)
 	struct {
 		SVGA3dCmdHeader header;
 		SVGA3dCmdDestroyContext body;
-	} *cmd;
+	} *cmd = vmw_fifo_reserve(dev_priv, sizeof(*cmd));
 
-
-	vmw_execbuf_release_pinned_bo(dev_priv, true, res->id);
-
-	cmd = vmw_fifo_reserve(dev_priv, sizeof(*cmd));
 	if (unlikely(cmd == NULL)) {
 		DRM_ERROR("Failed reserving FIFO space for surface "
 			  "destruction.\n");
@@ -277,7 +211,7 @@ static void vmw_hw_context_destroy(struct vmw_resource *res)
 	cmd->body.cid = cpu_to_le32(res->id);
 
 	vmw_fifo_commit(dev_priv, sizeof(*cmd));
-	vmw_3d_resource_dec(dev_priv, false);
+	vmw_3d_resource_dec(dev_priv);
 }
 
 static int vmw_context_init(struct vmw_private *dev_priv,
@@ -292,17 +226,14 @@ static int vmw_context_init(struct vmw_private *dev_priv,
 	} *cmd;
 
 	ret = vmw_resource_init(dev_priv, res, &dev_priv->context_idr,
-				VMW_RES_CONTEXT, false, res_free, NULL);
+				VMW_RES_CONTEXT, res_free);
 
 	if (unlikely(ret != 0)) {
-		DRM_ERROR("Failed to allocate a resource id.\n");
-		goto out_early;
-	}
-
-	if (unlikely(res->id >= SVGA3D_MAX_CONTEXT_IDS)) {
-		DRM_ERROR("Out of hw context ids.\n");
-		vmw_resource_unreference(&res);
-		return -ENOMEM;
+		if (res_free == NULL)
+			kfree(res);
+		else
+			res_free(res);
+		return ret;
 	}
 
 	cmd = vmw_fifo_reserve(dev_priv, sizeof(*cmd));
@@ -317,16 +248,9 @@ static int vmw_context_init(struct vmw_private *dev_priv,
 	cmd->body.cid = cpu_to_le32(res->id);
 
 	vmw_fifo_commit(dev_priv, sizeof(*cmd));
-	(void) vmw_3d_resource_inc(dev_priv, false);
+	(void) vmw_3d_resource_inc(dev_priv);
 	vmw_resource_activate(res, vmw_hw_context_destroy);
 	return 0;
-
-out_early:
-	if (res_free == NULL)
-		kfree(res);
-	else
-		res_free(res);
-	return ret;
 }
 
 struct vmw_resource *vmw_context_alloc(struct vmw_private *dev_priv)
@@ -349,11 +273,8 @@ static void vmw_user_context_free(struct vmw_resource *res)
 {
 	struct vmw_user_context *ctx =
 	    container_of(res, struct vmw_user_context, res);
-	struct vmw_private *dev_priv = res->dev_priv;
 
 	kfree(ctx);
-	ttm_mem_global_free(vmw_mem_glob(dev_priv),
-			    vmw_user_context_size);
 }
 
 /**
@@ -407,56 +328,23 @@ int vmw_context_define_ioctl(struct drm_device *dev, void *data,
 			     struct drm_file *file_priv)
 {
 	struct vmw_private *dev_priv = vmw_priv(dev);
-	struct vmw_user_context *ctx;
+	struct vmw_user_context *ctx = kmalloc(sizeof(*ctx), GFP_KERNEL);
 	struct vmw_resource *res;
 	struct vmw_resource *tmp;
 	struct drm_vmw_context_arg *arg = (struct drm_vmw_context_arg *)data;
 	struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile;
-	struct vmw_master *vmaster = vmw_master(file_priv->master);
 	int ret;
 
-
-	/*
-	 * Approximate idr memory usage with 128 bytes. It will be limited
-	 * by maximum number_of contexts anyway.
-	 */
-
-	if (unlikely(vmw_user_context_size == 0))
-		vmw_user_context_size = ttm_round_pot(sizeof(*ctx)) + 128;
-
-	ret = ttm_read_lock(&vmaster->lock, true);
-	if (unlikely(ret != 0))
-		return ret;
-
-	ret = ttm_mem_global_alloc(vmw_mem_glob(dev_priv),
-				   vmw_user_context_size,
-				   false, true);
-	if (unlikely(ret != 0)) {
-		if (ret != -ERESTARTSYS)
-			DRM_ERROR("Out of graphics memory for context"
-				  " creation.\n");
-		goto out_unlock;
-	}
-
-	ctx = kmalloc(sizeof(*ctx), GFP_KERNEL);
-	if (unlikely(ctx == NULL)) {
-		ttm_mem_global_free(vmw_mem_glob(dev_priv),
-				    vmw_user_context_size);
-		ret = -ENOMEM;
-		goto out_unlock;
-	}
+	if (unlikely(ctx == NULL))
+		return -ENOMEM;
 
 	res = &ctx->res;
 	ctx->base.shareable = false;
 	ctx->base.tfile = NULL;
 
-	/*
-	 * From here on, the destructor takes over resource freeing.
-	 */
-
 	ret = vmw_context_init(dev_priv, res, vmw_user_context_free);
 	if (unlikely(ret != 0))
-		goto out_unlock;
+		return ret;
 
 	tmp = vmw_resource_reference(&ctx->res);
 	ret = ttm_base_object_init(tfile, &ctx->base, false, VMW_RES_CONTEXT,
@@ -470,16 +358,13 @@ int vmw_context_define_ioctl(struct drm_device *dev, void *data,
 	arg->cid = res->id;
 out_err:
 	vmw_resource_unreference(&res);
-out_unlock:
-	ttm_read_unlock(&vmaster->lock);
 	return ret;
 
 }
 
 int vmw_context_check(struct vmw_private *dev_priv,
 		      struct ttm_object_file *tfile,
-		      int id,
-		      struct vmw_resource **p_res)
+		      int id)
 {
 	struct vmw_resource *res;
 	int ret = 0;
@@ -491,8 +376,6 @@ int vmw_context_check(struct vmw_private *dev_priv,
 			container_of(res, struct vmw_user_context, res);
 		if (ctx->base.tfile != tfile && !ctx->base.shareable)
 			ret = -EPERM;
-		if (p_res)
-			*p_res = vmw_resource_reference(res);
 	} else
 		ret = -EINVAL;
 	read_unlock(&dev_priv->resource_lock);
@@ -500,638 +383,102 @@ int vmw_context_check(struct vmw_private *dev_priv,
 	return ret;
 }
 
-struct vmw_bpp {
-	uint8_t bpp;
-	uint8_t s_bpp;
-};
-
-/*
- * Size table for the supported SVGA3D surface formats. It consists of
- * two values. The bpp value and the s_bpp value which is short for
- * "stride bits per pixel" The values are given in such a way that the
- * minimum stride for the image is calculated using
- *
- * min_stride = w*s_bpp
- *
- * and the total memory requirement for the image is
- *
- * h*min_stride*bpp/s_bpp
- *
- */
-static const struct vmw_bpp vmw_sf_bpp[] = {
-	[SVGA3D_FORMAT_INVALID] = {0, 0},
-	[SVGA3D_X8R8G8B8] = {32, 32},
-	[SVGA3D_A8R8G8B8] = {32, 32},
-	[SVGA3D_R5G6B5] = {16, 16},
-	[SVGA3D_X1R5G5B5] = {16, 16},
-	[SVGA3D_A1R5G5B5] = {16, 16},
-	[SVGA3D_A4R4G4B4] = {16, 16},
-	[SVGA3D_Z_D32] = {32, 32},
-	[SVGA3D_Z_D16] = {16, 16},
-	[SVGA3D_Z_D24S8] = {32, 32},
-	[SVGA3D_Z_D15S1] = {16, 16},
-	[SVGA3D_LUMINANCE8] = {8, 8},
-	[SVGA3D_LUMINANCE4_ALPHA4] = {8, 8},
-	[SVGA3D_LUMINANCE16] = {16, 16},
-	[SVGA3D_LUMINANCE8_ALPHA8] = {16, 16},
-	[SVGA3D_DXT1] = {4, 16},
-	[SVGA3D_DXT2] = {8, 32},
-	[SVGA3D_DXT3] = {8, 32},
-	[SVGA3D_DXT4] = {8, 32},
-	[SVGA3D_DXT5] = {8, 32},
-	[SVGA3D_BUMPU8V8] = {16, 16},
-	[SVGA3D_BUMPL6V5U5] = {16, 16},
-	[SVGA3D_BUMPX8L8V8U8] = {32, 32},
-	[SVGA3D_ARGB_S10E5] = {16, 16},
-	[SVGA3D_ARGB_S23E8] = {32, 32},
-	[SVGA3D_A2R10G10B10] = {32, 32},
-	[SVGA3D_V8U8] = {16, 16},
-	[SVGA3D_Q8W8V8U8] = {32, 32},
-	[SVGA3D_CxV8U8] = {16, 16},
-	[SVGA3D_X8L8V8U8] = {32, 32},
-	[SVGA3D_A2W10V10U10] = {32, 32},
-	[SVGA3D_ALPHA8] = {8, 8},
-	[SVGA3D_R_S10E5] = {16, 16},
-	[SVGA3D_R_S23E8] = {32, 32},
-	[SVGA3D_RG_S10E5] = {16, 16},
-	[SVGA3D_RG_S23E8] = {32, 32},
-	[SVGA3D_BUFFER] = {8, 8},
-	[SVGA3D_Z_D24X8] = {32, 32},
-	[SVGA3D_V16U16] = {32, 32},
-	[SVGA3D_G16R16] = {32, 32},
-	[SVGA3D_A16B16G16R16] = {64,  64},
-	[SVGA3D_UYVY] = {12, 12},
-	[SVGA3D_YUY2] = {12, 12},
-	[SVGA3D_NV12] = {12, 8},
-	[SVGA3D_AYUV] = {32, 32},
-	[SVGA3D_BC4_UNORM] = {4,  16},
-	[SVGA3D_BC5_UNORM] = {8,  32},
-	[SVGA3D_Z_DF16] = {16,  16},
-	[SVGA3D_Z_DF24] = {24,  24},
-	[SVGA3D_Z_D24S8_INT] = {32,  32}
-};
-
 
 /**
  * Surface management.
  */
 
-struct vmw_surface_dma {
-	SVGA3dCmdHeader header;
-	SVGA3dCmdSurfaceDMA body;
-	SVGA3dCopyBox cb;
-	SVGA3dCmdSurfaceDMASuffix suffix;
-};
-
-struct vmw_surface_define {
-	SVGA3dCmdHeader header;
-	SVGA3dCmdDefineSurface body;
-};
-
-struct vmw_surface_destroy {
-	SVGA3dCmdHeader header;
-	SVGA3dCmdDestroySurface body;
-};
-
-
-/**
- * vmw_surface_dma_size - Compute fifo size for a dma command.
- *
- * @srf: Pointer to a struct vmw_surface
- *
- * Computes the required size for a surface dma command for backup or
- * restoration of the surface represented by @srf.
- */
-static inline uint32_t vmw_surface_dma_size(const struct vmw_surface *srf)
-{
-	return srf->num_sizes * sizeof(struct vmw_surface_dma);
-}
-
-
-/**
- * vmw_surface_define_size - Compute fifo size for a surface define command.
- *
- * @srf: Pointer to a struct vmw_surface
- *
- * Computes the required size for a surface define command for the definition
- * of the surface represented by @srf.
- */
-static inline uint32_t vmw_surface_define_size(const struct vmw_surface *srf)
-{
-	return sizeof(struct vmw_surface_define) + srf->num_sizes *
-		sizeof(SVGA3dSize);
-}
-
-
-/**
- * vmw_surface_destroy_size - Compute fifo size for a surface destroy command.
- *
- * Computes the required size for a surface destroy command for the destruction
- * of a hw surface.
- */
-static inline uint32_t vmw_surface_destroy_size(void)
-{
-	return sizeof(struct vmw_surface_destroy);
-}
-
-/**
- * vmw_surface_destroy_encode - Encode a surface_destroy command.
- *
- * @id: The surface id
- * @cmd_space: Pointer to memory area in which the commands should be encoded.
- */
-static void vmw_surface_destroy_encode(uint32_t id,
-				       void *cmd_space)
-{
-	struct vmw_surface_destroy *cmd = (struct vmw_surface_destroy *)
-		cmd_space;
-
-	cmd->header.id = SVGA_3D_CMD_SURFACE_DESTROY;
-	cmd->header.size = sizeof(cmd->body);
-	cmd->body.sid = id;
-}
-
-/**
- * vmw_surface_define_encode - Encode a surface_define command.
- *
- * @srf: Pointer to a struct vmw_surface object.
- * @cmd_space: Pointer to memory area in which the commands should be encoded.
- */
-static void vmw_surface_define_encode(const struct vmw_surface *srf,
-				      void *cmd_space)
-{
-	struct vmw_surface_define *cmd = (struct vmw_surface_define *)
-		cmd_space;
-	struct drm_vmw_size *src_size;
-	SVGA3dSize *cmd_size;
-	uint32_t cmd_len;
-	int i;
-
-	cmd_len = sizeof(cmd->body) + srf->num_sizes * sizeof(SVGA3dSize);
-
-	cmd->header.id = SVGA_3D_CMD_SURFACE_DEFINE;
-	cmd->header.size = cmd_len;
-	cmd->body.sid = srf->res.id;
-	cmd->body.surfaceFlags = srf->flags;
-	cmd->body.format = cpu_to_le32(srf->format);
-	for (i = 0; i < DRM_VMW_MAX_SURFACE_FACES; ++i)
-		cmd->body.face[i].numMipLevels = srf->mip_levels[i];
-
-	cmd += 1;
-	cmd_size = (SVGA3dSize *) cmd;
-	src_size = srf->sizes;
-
-	for (i = 0; i < srf->num_sizes; ++i, cmd_size++, src_size++) {
-		cmd_size->width = src_size->width;
-		cmd_size->height = src_size->height;
-		cmd_size->depth = src_size->depth;
-	}
-}
-
-
-/**
- * vmw_surface_dma_encode - Encode a surface_dma command.
- *
- * @srf: Pointer to a struct vmw_surface object.
- * @cmd_space: Pointer to memory area in which the commands should be encoded.
- * @ptr: Pointer to an SVGAGuestPtr indicating where the surface contents
- * should be placed or read from.
- * @to_surface: Boolean whether to DMA to the surface or from the surface.
- */
-static void vmw_surface_dma_encode(struct vmw_surface *srf,
-				   void *cmd_space,
-				   const SVGAGuestPtr *ptr,
-				   bool to_surface)
-{
-	uint32_t i;
-	uint32_t bpp = vmw_sf_bpp[srf->format].bpp;
-	uint32_t stride_bpp = vmw_sf_bpp[srf->format].s_bpp;
-	struct vmw_surface_dma *cmd = (struct vmw_surface_dma *)cmd_space;
-
-	for (i = 0; i < srf->num_sizes; ++i) {
-		SVGA3dCmdHeader *header = &cmd->header;
-		SVGA3dCmdSurfaceDMA *body = &cmd->body;
-		SVGA3dCopyBox *cb = &cmd->cb;
-		SVGA3dCmdSurfaceDMASuffix *suffix = &cmd->suffix;
-		const struct vmw_surface_offset *cur_offset = &srf->offsets[i];
-		const struct drm_vmw_size *cur_size = &srf->sizes[i];
-
-		header->id = SVGA_3D_CMD_SURFACE_DMA;
-		header->size = sizeof(*body) + sizeof(*cb) + sizeof(*suffix);
-
-		body->guest.ptr = *ptr;
-		body->guest.ptr.offset += cur_offset->bo_offset;
-		body->guest.pitch = (cur_size->width * stride_bpp + 7) >> 3;
-		body->host.sid = srf->res.id;
-		body->host.face = cur_offset->face;
-		body->host.mipmap = cur_offset->mip;
-		body->transfer = ((to_surface) ?  SVGA3D_WRITE_HOST_VRAM :
-				  SVGA3D_READ_HOST_VRAM);
-		cb->x = 0;
-		cb->y = 0;
-		cb->z = 0;
-		cb->srcx = 0;
-		cb->srcy = 0;
-		cb->srcz = 0;
-		cb->w = cur_size->width;
-		cb->h = cur_size->height;
-		cb->d = cur_size->depth;
-
-		suffix->suffixSize = sizeof(*suffix);
-		suffix->maximumOffset = body->guest.pitch*cur_size->height*
-			cur_size->depth*bpp / stride_bpp;
-		suffix->flags.discard = 0;
-		suffix->flags.unsynchronized = 0;
-		suffix->flags.reserved = 0;
-		++cmd;
-	}
-};
-
-
 static void vmw_hw_surface_destroy(struct vmw_resource *res)
 {
 
 	struct vmw_private *dev_priv = res->dev_priv;
-	struct vmw_surface *srf;
-	void *cmd;
-
-	if (res->id != -1) {
-
-		cmd = vmw_fifo_reserve(dev_priv, vmw_surface_destroy_size());
-		if (unlikely(cmd == NULL)) {
-			DRM_ERROR("Failed reserving FIFO space for surface "
-				  "destruction.\n");
-			return;
-		}
-
-		vmw_surface_destroy_encode(res->id, cmd);
-		vmw_fifo_commit(dev_priv, vmw_surface_destroy_size());
+	struct {
+		SVGA3dCmdHeader header;
+		SVGA3dCmdDestroySurface body;
+	} *cmd = vmw_fifo_reserve(dev_priv, sizeof(*cmd));
 
-		/*
-		 * used_memory_size_atomic, or separate lock
-		 * to avoid taking dev_priv::cmdbuf_mutex in
-		 * the destroy path.
-		 */
+	if (unlikely(cmd == NULL)) {
+		DRM_ERROR("Failed reserving FIFO space for surface "
+			  "destruction.\n");
+		return;
+	}
 
-		mutex_lock(&dev_priv->cmdbuf_mutex);
-		srf = container_of(res, struct vmw_surface, res);
-		dev_priv->used_memory_size -= srf->backup_size;
-		mutex_unlock(&dev_priv->cmdbuf_mutex);
+	cmd->header.id = cpu_to_le32(SVGA_3D_CMD_SURFACE_DESTROY);
+	cmd->header.size = cpu_to_le32(sizeof(cmd->body));
+	cmd->body.sid = cpu_to_le32(res->id);
 
-	}
-	vmw_3d_resource_dec(dev_priv, false);
+	vmw_fifo_commit(dev_priv, sizeof(*cmd));
+	vmw_3d_resource_dec(dev_priv);
 }
 
 void vmw_surface_res_free(struct vmw_resource *res)
 {
 	struct vmw_surface *srf = container_of(res, struct vmw_surface, res);
 
-	if (srf->backup)
-		ttm_bo_unref(&srf->backup);
-	kfree(srf->offsets);
 	kfree(srf->sizes);
 	kfree(srf->snooper.image);
 	kfree(srf);
 }
 
-
-/**
- * vmw_surface_do_validate - make a surface available to the device.
- *
- * @dev_priv: Pointer to a device private struct.
- * @srf: Pointer to a struct vmw_surface.
- *
- * If the surface doesn't have a hw id, allocate one, and optionally
- * DMA the backed up surface contents to the device.
- *
- * Returns -EBUSY if there wasn't sufficient device resources to
- * complete the validation. Retry after freeing up resources.
- *
- * May return other errors if the kernel is out of guest resources.
- */
-int vmw_surface_do_validate(struct vmw_private *dev_priv,
-			    struct vmw_surface *srf)
+int vmw_surface_init(struct vmw_private *dev_priv,
+		     struct vmw_surface *srf,
+		     void (*res_free) (struct vmw_resource *res))
 {
-	struct vmw_resource *res = &srf->res;
-	struct list_head val_list;
-	struct ttm_validate_buffer val_buf;
-	uint32_t submit_size;
-	uint8_t *cmd;
 	int ret;
+	struct {
+		SVGA3dCmdHeader header;
+		SVGA3dCmdDefineSurface body;
+	} *cmd;
+	SVGA3dSize *cmd_size;
+	struct vmw_resource *res = &srf->res;
+	struct drm_vmw_size *src_size;
+	size_t submit_size;
+	uint32_t cmd_len;
+	int i;
 
-	if (likely(res->id != -1))
-		return 0;
-
-	if (unlikely(dev_priv->used_memory_size + srf->backup_size >=
-		     dev_priv->memory_size))
-		return -EBUSY;
-
-	/*
-	 * Reserve- and validate the backup DMA bo.
-	 */
-
-	if (srf->backup) {
-		INIT_LIST_HEAD(&val_list);
-		val_buf.bo = ttm_bo_reference(srf->backup);
-		val_buf.new_sync_obj_arg = (void *)((unsigned long)
-						    DRM_VMW_FENCE_FLAG_EXEC);
-		list_add_tail(&val_buf.head, &val_list);
-		ret = ttm_eu_reserve_buffers(&val_list);
-		if (unlikely(ret != 0))
-			goto out_no_reserve;
-
-		ret = ttm_bo_validate(srf->backup, &vmw_srf_placement,
-				      true, false, false);
-		if (unlikely(ret != 0))
-			goto out_no_validate;
-	}
-
-	/*
-	 * Alloc id for the resource.
-	 */
+	BUG_ON(res_free == NULL);
+	ret = vmw_resource_init(dev_priv, res, &dev_priv->surface_idr,
+				VMW_RES_SURFACE, res_free);
 
-	ret = vmw_resource_alloc_id(dev_priv, res);
 	if (unlikely(ret != 0)) {
-		DRM_ERROR("Failed to allocate a surface id.\n");
-		goto out_no_id;
-	}
-	if (unlikely(res->id >= SVGA3D_MAX_SURFACE_IDS)) {
-		ret = -EBUSY;
-		goto out_no_fifo;
+		res_free(res);
+		return ret;
 	}
 
-
-	/*
-	 * Encode surface define- and dma commands.
-	 */
-
-	submit_size = vmw_surface_define_size(srf);
-	if (srf->backup)
-		submit_size += vmw_surface_dma_size(srf);
+	submit_size = sizeof(*cmd) + srf->num_sizes * sizeof(SVGA3dSize);
+	cmd_len = sizeof(cmd->body) + srf->num_sizes * sizeof(SVGA3dSize);
 
 	cmd = vmw_fifo_reserve(dev_priv, submit_size);
 	if (unlikely(cmd == NULL)) {
-		DRM_ERROR("Failed reserving FIFO space for surface "
-			  "validation.\n");
-		ret = -ENOMEM;
-		goto out_no_fifo;
-	}
-
-	vmw_surface_define_encode(srf, cmd);
-	if (srf->backup) {
-		SVGAGuestPtr ptr;
-
-		cmd += vmw_surface_define_size(srf);
-		vmw_bo_get_guest_ptr(srf->backup, &ptr);
-		vmw_surface_dma_encode(srf, cmd, &ptr, true);
-	}
-
-	vmw_fifo_commit(dev_priv, submit_size);
-
-	/*
-	 * Create a fence object and fence the backup buffer.
-	 */
-
-	if (srf->backup) {
-		struct vmw_fence_obj *fence;
-
-		(void) vmw_execbuf_fence_commands(NULL, dev_priv,
-						  &fence, NULL);
-		ttm_eu_fence_buffer_objects(&val_list, fence);
-		if (likely(fence != NULL))
-			vmw_fence_obj_unreference(&fence);
-		ttm_bo_unref(&val_buf.bo);
-		ttm_bo_unref(&srf->backup);
+		DRM_ERROR("Fifo reserve failed for create surface.\n");
+		vmw_resource_unreference(&res);
+		return -ENOMEM;
 	}
 
-	/*
-	 * Surface memory usage accounting.
-	 */
-
-	dev_priv->used_memory_size += srf->backup_size;
-
-	return 0;
-
-out_no_fifo:
-	vmw_resource_release_id(res);
-out_no_id:
-out_no_validate:
-	if (srf->backup)
-		ttm_eu_backoff_reservation(&val_list);
-out_no_reserve:
-	if (srf->backup)
-		ttm_bo_unref(&val_buf.bo);
-	return ret;
-}
-
-/**
- * vmw_surface_evict - Evict a hw surface.
- *
- * @dev_priv: Pointer to a device private struct.
- * @srf: Pointer to a struct vmw_surface
- *
- * DMA the contents of a hw surface to a backup guest buffer object,
- * and destroy the hw surface, releasing its id.
- */
-int vmw_surface_evict(struct vmw_private *dev_priv,
-		      struct vmw_surface *srf)
-{
-	struct vmw_resource *res = &srf->res;
-	struct list_head val_list;
-	struct ttm_validate_buffer val_buf;
-	uint32_t submit_size;
-	uint8_t *cmd;
-	int ret;
-	struct vmw_fence_obj *fence;
-	SVGAGuestPtr ptr;
-
-	BUG_ON(res->id == -1);
-
-	/*
-	 * Create a surface backup buffer object.
-	 */
-
-	if (!srf->backup) {
-		ret = ttm_bo_create(&dev_priv->bdev, srf->backup_size,
-				    ttm_bo_type_device,
-				    &vmw_srf_placement, 0, 0, true,
-				    NULL, &srf->backup);
-		if (unlikely(ret != 0))
-			return ret;
+	cmd->header.id = cpu_to_le32(SVGA_3D_CMD_SURFACE_DEFINE);
+	cmd->header.size = cpu_to_le32(cmd_len);
+	cmd->body.sid = cpu_to_le32(res->id);
+	cmd->body.surfaceFlags = cpu_to_le32(srf->flags);
+	cmd->body.format = cpu_to_le32(srf->format);
+	for (i = 0; i < DRM_VMW_MAX_SURFACE_FACES; ++i) {
+		cmd->body.face[i].numMipLevels =
+		    cpu_to_le32(srf->mip_levels[i]);
 	}
 
-	/*
-	 * Reserve- and validate the backup DMA bo.
-	 */
-
-	INIT_LIST_HEAD(&val_list);
-	val_buf.bo = ttm_bo_reference(srf->backup);
-	val_buf.new_sync_obj_arg = (void *)(unsigned long)
-		DRM_VMW_FENCE_FLAG_EXEC;
-	list_add_tail(&val_buf.head, &val_list);
-	ret = ttm_eu_reserve_buffers(&val_list);
-	if (unlikely(ret != 0))
-		goto out_no_reserve;
-
-	ret = ttm_bo_validate(srf->backup, &vmw_srf_placement,
-			      true, false, false);
-	if (unlikely(ret != 0))
-		goto out_no_validate;
-
-
-	/*
-	 * Encode the dma- and surface destroy commands.
-	 */
+	cmd += 1;
+	cmd_size = (SVGA3dSize *) cmd;
+	src_size = srf->sizes;
 
-	submit_size = vmw_surface_dma_size(srf) + vmw_surface_destroy_size();
-	cmd = vmw_fifo_reserve(dev_priv, submit_size);
-	if (unlikely(cmd == NULL)) {
-		DRM_ERROR("Failed reserving FIFO space for surface "
-			  "eviction.\n");
-		ret = -ENOMEM;
-		goto out_no_fifo;
+	for (i = 0; i < srf->num_sizes; ++i, cmd_size++, src_size++) {
+		cmd_size->width = cpu_to_le32(src_size->width);
+		cmd_size->height = cpu_to_le32(src_size->height);
+		cmd_size->depth = cpu_to_le32(src_size->depth);
 	}
 
-	vmw_bo_get_guest_ptr(srf->backup, &ptr);
-	vmw_surface_dma_encode(srf, cmd, &ptr, false);
-	cmd += vmw_surface_dma_size(srf);
-	vmw_surface_destroy_encode(res->id, cmd);
 	vmw_fifo_commit(dev_priv, submit_size);
-
-	/*
-	 * Surface memory usage accounting.
-	 */
-
-	dev_priv->used_memory_size -= srf->backup_size;
-
-	/*
-	 * Create a fence object and fence the DMA buffer.
-	 */
-
-	(void) vmw_execbuf_fence_commands(NULL, dev_priv,
-					  &fence, NULL);
-	ttm_eu_fence_buffer_objects(&val_list, fence);
-	if (likely(fence != NULL))
-		vmw_fence_obj_unreference(&fence);
-	ttm_bo_unref(&val_buf.bo);
-
-	/*
-	 * Release the surface ID.
-	 */
-
-	vmw_resource_release_id(res);
-
-	return 0;
-
-out_no_fifo:
-out_no_validate:
-	if (srf->backup)
-		ttm_eu_backoff_reservation(&val_list);
-out_no_reserve:
-	ttm_bo_unref(&val_buf.bo);
-	ttm_bo_unref(&srf->backup);
-	return ret;
-}
-
-
-/**
- * vmw_surface_validate - make a surface available to the device, evicting
- * other surfaces if needed.
- *
- * @dev_priv: Pointer to a device private struct.
- * @srf: Pointer to a struct vmw_surface.
- *
- * Try to validate a surface and if it fails due to limited device resources,
- * repeatedly try to evict other surfaces until the request can be
- * acommodated.
- *
- * May return errors if out of resources.
- */
-int vmw_surface_validate(struct vmw_private *dev_priv,
-			 struct vmw_surface *srf)
-{
-	int ret;
-	struct vmw_surface *evict_srf;
-
-	do {
-		write_lock(&dev_priv->resource_lock);
-		list_del_init(&srf->lru_head);
-		write_unlock(&dev_priv->resource_lock);
-
-		ret = vmw_surface_do_validate(dev_priv, srf);
-		if (likely(ret != -EBUSY))
-			break;
-
-		write_lock(&dev_priv->resource_lock);
-		if (list_empty(&dev_priv->surface_lru)) {
-			DRM_ERROR("Out of device memory for surfaces.\n");
-			ret = -EBUSY;
-			write_unlock(&dev_priv->resource_lock);
-			break;
-		}
-
-		evict_srf = vmw_surface_reference
-			(list_first_entry(&dev_priv->surface_lru,
-					  struct vmw_surface,
-					  lru_head));
-		list_del_init(&evict_srf->lru_head);
-
-		write_unlock(&dev_priv->resource_lock);
-		(void) vmw_surface_evict(dev_priv, evict_srf);
-
-		vmw_surface_unreference(&evict_srf);
-
-	} while (1);
-
-	if (unlikely(ret != 0 && srf->res.id != -1)) {
-		write_lock(&dev_priv->resource_lock);
-		list_add_tail(&srf->lru_head, &dev_priv->surface_lru);
-		write_unlock(&dev_priv->resource_lock);
-	}
-
-	return ret;
-}
-
-
-/**
- * vmw_surface_remove_from_lists - Remove surface resources from lookup lists
- *
- * @res: Pointer to a struct vmw_resource embedded in a struct vmw_surface
- *
- * As part of the resource destruction, remove the surface from any
- * lookup lists.
- */
-static void vmw_surface_remove_from_lists(struct vmw_resource *res)
-{
-	struct vmw_surface *srf = container_of(res, struct vmw_surface, res);
-
-	list_del_init(&srf->lru_head);
-}
-
-int vmw_surface_init(struct vmw_private *dev_priv,
-		     struct vmw_surface *srf,
-		     void (*res_free) (struct vmw_resource *res))
-{
-	int ret;
-	struct vmw_resource *res = &srf->res;
-
-	BUG_ON(res_free == NULL);
-	INIT_LIST_HEAD(&srf->lru_head);
-	ret = vmw_resource_init(dev_priv, res, &dev_priv->surface_idr,
-				VMW_RES_SURFACE, true, res_free,
-				vmw_surface_remove_from_lists);
-
-	if (unlikely(ret != 0))
-		res_free(res);
-
-	/*
-	 * The surface won't be visible to hardware until a
-	 * surface validate.
-	 */
-
-	(void) vmw_3d_resource_inc(dev_priv, false);
+	(void) vmw_3d_resource_inc(dev_priv);
 	vmw_resource_activate(res, vmw_hw_surface_destroy);
-	return ret;
+	return 0;
 }
 
 static void vmw_user_surface_free(struct vmw_resource *res)
@@ -1139,81 +486,12 @@ static void vmw_user_surface_free(struct vmw_resource *res)
 	struct vmw_surface *srf = container_of(res, struct vmw_surface, res);
 	struct vmw_user_surface *user_srf =
 	    container_of(srf, struct vmw_user_surface, srf);
-	struct vmw_private *dev_priv = srf->res.dev_priv;
-	uint32_t size = user_srf->size;
 
-	if (srf->backup)
-		ttm_bo_unref(&srf->backup);
-	kfree(srf->offsets);
 	kfree(srf->sizes);
 	kfree(srf->snooper.image);
 	kfree(user_srf);
-	ttm_mem_global_free(vmw_mem_glob(dev_priv), size);
-}
-
-/**
- * vmw_resource_unreserve - unreserve resources previously reserved for
- * command submission.
- *
- * @list_head: list of resources to unreserve.
- *
- * Currently only surfaces are considered, and unreserving a surface
- * means putting it back on the device's surface lru list,
- * so that it can be evicted if necessary.
- * This function traverses the resource list and
- * checks whether resources are surfaces, and in that case puts them back
- * on the device's surface LRU list.
- */
-void vmw_resource_unreserve(struct list_head *list)
-{
-	struct vmw_resource *res;
-	struct vmw_surface *srf;
-	rwlock_t *lock = NULL;
-
-	list_for_each_entry(res, list, validate_head) {
-
-		if (res->res_free != &vmw_surface_res_free &&
-		    res->res_free != &vmw_user_surface_free)
-			continue;
-
-		if (unlikely(lock == NULL)) {
-			lock = &res->dev_priv->resource_lock;
-			write_lock(lock);
-		}
-
-		srf = container_of(res, struct vmw_surface, res);
-		list_del_init(&srf->lru_head);
-		list_add_tail(&srf->lru_head, &res->dev_priv->surface_lru);
-	}
-
-	if (lock != NULL)
-		write_unlock(lock);
 }
 
-/**
- * Helper function that looks either a surface or dmabuf.
- *
- * The pointer this pointed at by out_surf and out_buf needs to be null.
- */
-int vmw_user_lookup_handle(struct vmw_private *dev_priv,
-			   struct ttm_object_file *tfile,
-			   uint32_t handle,
-			   struct vmw_surface **out_surf,
-			   struct vmw_dma_buffer **out_buf)
-{
-	int ret;
-
-	BUG_ON(*out_surf || *out_buf);
-
-	ret = vmw_user_surface_lookup_handle(dev_priv, tfile, handle, out_surf);
-	if (!ret)
-		return 0;
-
-	ret = vmw_user_dmabuf_lookup(tfile, handle, out_buf);
-	return ret;
-}
-
-
 int vmw_user_surface_lookup_handle(struct vmw_private *dev_priv,
 				   struct ttm_object_file *tfile,
 				   uint32_t handle, struct vmw_surface **out)
@@ -1278,7 +556,8 @@ int vmw_surface_define_ioctl(struct drm_device *dev, void *data,
 			     struct drm_file *file_priv)
 {
 	struct vmw_private *dev_priv = vmw_priv(dev);
-	struct vmw_user_surface *user_srf;
+	struct vmw_user_surface *user_srf =
+	    kmalloc(sizeof(*user_srf), GFP_KERNEL);
 	struct vmw_surface *srf;
 	struct vmw_resource *res;
 	struct vmw_resource *tmp;
@@ -1289,51 +568,10 @@ int vmw_surface_define_ioctl(struct drm_device *dev, void *data,
 	struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile;
 	struct drm_vmw_size __user *user_sizes;
 	int ret;
-	int i, j;
-	uint32_t cur_bo_offset;
-	struct drm_vmw_size *cur_size;
-	struct vmw_surface_offset *cur_offset;
-	uint32_t stride_bpp;
-	uint32_t bpp;
-	uint32_t num_sizes;
-	uint32_t size;
-	struct vmw_master *vmaster = vmw_master(file_priv->master);
-
-	if (unlikely(vmw_user_surface_size == 0))
-		vmw_user_surface_size = ttm_round_pot(sizeof(*user_srf)) +
-			128;
-
-	num_sizes = 0;
-	for (i = 0; i < DRM_VMW_MAX_SURFACE_FACES; ++i)
-		num_sizes += req->mip_levels[i];
-
-	if (num_sizes > DRM_VMW_MAX_SURFACE_FACES *
-	    DRM_VMW_MAX_MIP_LEVELS)
-		return -EINVAL;
-
-	size = vmw_user_surface_size + 128 +
-		ttm_round_pot(num_sizes * sizeof(struct drm_vmw_size)) +
-		ttm_round_pot(num_sizes * sizeof(struct vmw_surface_offset));
-
-
-	ret = ttm_read_lock(&vmaster->lock, true);
-	if (unlikely(ret != 0))
-		return ret;
-
-	ret = ttm_mem_global_alloc(vmw_mem_glob(dev_priv),
-				   size, false, true);
-	if (unlikely(ret != 0)) {
-		if (ret != -ERESTARTSYS)
-			DRM_ERROR("Out of graphics memory for surface"
-				  " creation.\n");
-		goto out_unlock;
-	}
+	int i;
 
-	user_srf = kmalloc(sizeof(*user_srf), GFP_KERNEL);
-	if (unlikely(user_srf == NULL)) {
-		ret = -ENOMEM;
-		goto out_no_user_srf;
-	}
+	if (unlikely(user_srf == NULL))
+		return -ENOMEM;
 
 	srf = &user_srf->srf;
 	res = &srf->res;
@@ -1341,22 +579,21 @@ int vmw_surface_define_ioctl(struct drm_device *dev, void *data,
 	srf->flags = req->flags;
 	srf->format = req->format;
 	srf->scanout = req->scanout;
-	srf->backup = NULL;
-
 	memcpy(srf->mip_levels, req->mip_levels, sizeof(srf->mip_levels));
-	srf->num_sizes = num_sizes;
-	user_srf->size = size;
+	srf->num_sizes = 0;
+	for (i = 0; i < DRM_VMW_MAX_SURFACE_FACES; ++i)
+		srf->num_sizes += srf->mip_levels[i];
 
-	srf->sizes = kmalloc(srf->num_sizes * sizeof(*srf->sizes), GFP_KERNEL);
-	if (unlikely(srf->sizes == NULL)) {
-		ret = -ENOMEM;
-		goto out_no_sizes;
+	if (srf->num_sizes > DRM_VMW_MAX_SURFACE_FACES *
+	    DRM_VMW_MAX_MIP_LEVELS) {
+		ret = -EINVAL;
+		goto out_err0;
 	}
-	srf->offsets = kmalloc(srf->num_sizes * sizeof(*srf->offsets),
-			       GFP_KERNEL);
+
+	srf->sizes = kmalloc(srf->num_sizes * sizeof(*srf->sizes), GFP_KERNEL);
 	if (unlikely(srf->sizes == NULL)) {
 		ret = -ENOMEM;
-		goto out_no_offsets;
+		goto out_err0;
 	}
 
 	user_sizes = (struct drm_vmw_size __user *)(unsigned long)
@@ -1366,31 +603,8 @@ int vmw_surface_define_ioctl(struct drm_device *dev, void *data,
 			     srf->num_sizes * sizeof(*srf->sizes));
 	if (unlikely(ret != 0)) {
 		ret = -EFAULT;
-		goto out_no_copy;
-	}
-
-	cur_bo_offset = 0;
-	cur_offset = srf->offsets;
-	cur_size = srf->sizes;
-
-	bpp = vmw_sf_bpp[srf->format].bpp;
-	stride_bpp = vmw_sf_bpp[srf->format].s_bpp;
-
-	for (i = 0; i < DRM_VMW_MAX_SURFACE_FACES; ++i) {
-		for (j = 0; j < srf->mip_levels[i]; ++j) {
-			uint32_t stride =
-				(cur_size->width * stride_bpp + 7) >> 3;
-
-			cur_offset->face = i;
-			cur_offset->mip = j;
-			cur_offset->bo_offset = cur_bo_offset;
-			cur_bo_offset += stride * cur_size->height *
-				cur_size->depth * bpp / stride_bpp;
-			++cur_offset;
-			++cur_size;
-		}
+		goto out_err1;
 	}
-	srf->backup_size = cur_bo_offset;
 
 	if (srf->scanout &&
 	    srf->num_sizes == 1 &&
@@ -1398,12 +612,14 @@ int vmw_surface_define_ioctl(struct drm_device *dev, void *data,
 	    srf->sizes[0].height == 64 &&
 	    srf->format == SVGA3D_A8R8G8B8) {
 
-		/* allocate image area and clear it */
-		srf->snooper.image = kzalloc(64 * 64 * 4, GFP_KERNEL);
-		if (!srf->snooper.image) {
+		srf->snooper.image = kmalloc(64 * 64 * 4, GFP_KERNEL);
+		/* clear the image */
+		if (srf->snooper.image) {
+			memset(srf->snooper.image, 0x00, 64 * 64 * 4);
+		} else {
 			DRM_ERROR("Failed to allocate cursor_image\n");
 			ret = -ENOMEM;
-			goto out_no_copy;
+			goto out_err1;
 		}
 	} else {
 		srf->snooper.image = NULL;
@@ -1420,7 +636,7 @@ int vmw_surface_define_ioctl(struct drm_device *dev, void *data,
 
 	ret = vmw_surface_init(dev_priv, srf, vmw_user_surface_free);
 	if (unlikely(ret != 0))
-		goto out_unlock;
+		return ret;
 
 	tmp = vmw_resource_reference(&srf->res);
 	ret = ttm_base_object_init(tfile, &user_srf->base,
@@ -1430,7 +646,7 @@ int vmw_surface_define_ioctl(struct drm_device *dev, void *data,
 	if (unlikely(ret != 0)) {
 		vmw_resource_unreference(&tmp);
 		vmw_resource_unreference(&res);
-		goto out_unlock;
+		return ret;
 	}
 
 	rep->sid = user_srf->base.hash.key;
@@ -1438,19 +654,11 @@ int vmw_surface_define_ioctl(struct drm_device *dev, void *data,
 		DRM_ERROR("Created bad Surface ID.\n");
 
 	vmw_resource_unreference(&res);
-
-	ttm_read_unlock(&vmaster->lock);
 	return 0;
-out_no_copy:
-	kfree(srf->offsets);
-out_no_offsets:
+out_err1:
 	kfree(srf->sizes);
-out_no_sizes:
+out_err0:
 	kfree(user_srf);
-out_no_user_srf:
-	ttm_mem_global_free(vmw_mem_glob(dev_priv), size);
-out_unlock:
-	ttm_read_unlock(&vmaster->lock);
 	return ret;
 }
 
@@ -1540,10 +748,29 @@ out_bad_surface:
 /**
  * Buffer management.
  */
+
+static size_t vmw_dmabuf_acc_size(struct ttm_bo_global *glob,
+				  unsigned long num_pages)
+{
+	static size_t bo_user_size = ~0;
+
+	size_t page_array_size =
+	    (num_pages * sizeof(void *) + PAGE_SIZE - 1) & PAGE_MASK;
+
+	if (unlikely(bo_user_size == ~0)) {
+		bo_user_size = glob->ttm_bo_extra_size +
+		    ttm_round_pot(sizeof(struct vmw_dma_buffer));
+	}
+
+	return bo_user_size + page_array_size;
+}
+
 void vmw_dmabuf_bo_free(struct ttm_buffer_object *bo)
 {
 	struct vmw_dma_buffer *vmw_bo = vmw_dma_buffer(bo);
+	struct ttm_bo_global *glob = bo->glob;
 
+	ttm_mem_global_free(glob->mem_glob, bo->acc_size);
 	kfree(vmw_bo);
 }
 
@@ -1554,12 +781,24 @@ int vmw_dmabuf_init(struct vmw_private *dev_priv,
 		    void (*bo_free) (struct ttm_buffer_object *bo))
 {
 	struct ttm_bo_device *bdev = &dev_priv->bdev;
+	struct ttm_mem_global *mem_glob = bdev->glob->mem_glob;
 	size_t acc_size;
 	int ret;
 
 	BUG_ON(!bo_free);
 
-	acc_size = ttm_bo_acc_size(bdev, size, sizeof(struct vmw_dma_buffer));
+	acc_size =
+	    vmw_dmabuf_acc_size(bdev->glob,
+				(size + PAGE_SIZE - 1) >> PAGE_SHIFT);
+
+	ret = ttm_mem_global_alloc(mem_glob, acc_size, false, false);
+	if (unlikely(ret != 0)) {
+		/* we must free the bo here as
+		 * ttm_buffer_object_init does so as well */
+		bo_free(&vmw_bo->base);
+		return ret;
+	}
+
 	memset(vmw_bo, 0, sizeof(*vmw_bo));
 
 	INIT_LIST_HEAD(&vmw_bo->validate_list);
@@ -1574,7 +813,9 @@ int vmw_dmabuf_init(struct vmw_private *dev_priv,
 static void vmw_user_dmabuf_destroy(struct ttm_buffer_object *bo)
 {
 	struct vmw_user_dma_buffer *vmw_user_bo = vmw_user_dma_buffer(bo);
+	struct ttm_bo_global *glob = bo->glob;
 
+	ttm_mem_global_free(glob->mem_glob, bo->acc_size);
 	kfree(vmw_user_bo);
 }
 
@@ -1731,7 +972,7 @@ static int vmw_stream_init(struct vmw_private *dev_priv,
 	int ret;
 
 	ret = vmw_resource_init(dev_priv, res, &dev_priv->stream_idr,
-				VMW_RES_STREAM, false, res_free, NULL);
+				VMW_RES_STREAM, res_free);
 
 	if (unlikely(ret != 0)) {
 		if (res_free == NULL)
@@ -1761,11 +1002,8 @@ static void vmw_user_stream_free(struct vmw_resource *res)
 {
 	struct vmw_user_stream *stream =
 	    container_of(res, struct vmw_user_stream, stream.res);
-	struct vmw_private *dev_priv = res->dev_priv;
 
 	kfree(stream);
-	ttm_mem_global_free(vmw_mem_glob(dev_priv),
-			    vmw_user_stream_size);
 }
 
 /**
@@ -1819,56 +1057,23 @@ int vmw_stream_claim_ioctl(struct drm_device *dev, void *data,
 			   struct drm_file *file_priv)
 {
 	struct vmw_private *dev_priv = vmw_priv(dev);
-	struct vmw_user_stream *stream;
+	struct vmw_user_stream *stream = kmalloc(sizeof(*stream), GFP_KERNEL);
 	struct vmw_resource *res;
 	struct vmw_resource *tmp;
 	struct drm_vmw_stream_arg *arg = (struct drm_vmw_stream_arg *)data;
 	struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile;
-	struct vmw_master *vmaster = vmw_master(file_priv->master);
 	int ret;
 
-	/*
-	 * Approximate idr memory usage with 128 bytes. It will be limited
-	 * by maximum number_of streams anyway?
-	 */
-
-	if (unlikely(vmw_user_stream_size == 0))
-		vmw_user_stream_size = ttm_round_pot(sizeof(*stream)) + 128;
-
-	ret = ttm_read_lock(&vmaster->lock, true);
-	if (unlikely(ret != 0))
-		return ret;
-
-	ret = ttm_mem_global_alloc(vmw_mem_glob(dev_priv),
-				   vmw_user_stream_size,
-				   false, true);
-	if (unlikely(ret != 0)) {
-		if (ret != -ERESTARTSYS)
-			DRM_ERROR("Out of graphics memory for stream"
-				  " creation.\n");
-		goto out_unlock;
-	}
-
-
-	stream = kmalloc(sizeof(*stream), GFP_KERNEL);
-	if (unlikely(stream == NULL)) {
-		ttm_mem_global_free(vmw_mem_glob(dev_priv),
-				    vmw_user_stream_size);
-		ret = -ENOMEM;
-		goto out_unlock;
-	}
+	if (unlikely(stream == NULL))
+		return -ENOMEM;
 
 	res = &stream->stream.res;
 	stream->base.shareable = false;
 	stream->base.tfile = NULL;
 
-	/*
-	 * From here on, the destructor takes over resource freeing.
-	 */
-
 	ret = vmw_stream_init(dev_priv, &stream->stream, vmw_user_stream_free);
 	if (unlikely(ret != 0))
-		goto out_unlock;
+		return ret;
 
 	tmp = vmw_resource_reference(res);
 	ret = ttm_base_object_init(tfile, &stream->base, false, VMW_RES_STREAM,
@@ -1882,8 +1087,6 @@ int vmw_stream_claim_ioctl(struct drm_device *dev, void *data,
 	arg->stream_id = res->id;
 out_err:
 	vmw_resource_unreference(&res);
-out_unlock:
-	ttm_read_unlock(&vmaster->lock);
 	return ret;
 }
 
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_scrn.c b/drivers/gpu/drm/vmwgfx/vmwgfx_scrn.c
deleted file mode 100644
index 6deaf2f..0000000
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_scrn.c
+++ /dev/null
@@ -1,571 +0,0 @@
-/**************************************************************************
- *
- * Copyright © 2011 VMware, Inc., Palo Alto, CA., USA
- * All Rights Reserved.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the
- * "Software"), to deal in the Software without restriction, including
- * without limitation the rights to use, copy, modify, merge, publish,
- * distribute, sub license, and/or sell copies of the Software, and to
- * permit persons to whom the Software is furnished to do so, subject to
- * the following conditions:
- *
- * The above copyright notice and this permission notice (including the
- * next paragraph) shall be included in all copies or substantial portions
- * of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
- * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
- * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
- * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
- * USE OR OTHER DEALINGS IN THE SOFTWARE.
- *
- **************************************************************************/
-
-#include "vmwgfx_kms.h"
-
-
-#define vmw_crtc_to_sou(x) \
-	container_of(x, struct vmw_screen_object_unit, base.crtc)
-#define vmw_encoder_to_sou(x) \
-	container_of(x, struct vmw_screen_object_unit, base.encoder)
-#define vmw_connector_to_sou(x) \
-	container_of(x, struct vmw_screen_object_unit, base.connector)
-
-struct vmw_screen_object_display {
-	unsigned num_implicit;
-
-	struct vmw_framebuffer *implicit_fb;
-};
-
-/**
- * Display unit using screen objects.
- */
-struct vmw_screen_object_unit {
-	struct vmw_display_unit base;
-
-	unsigned long buffer_size; /**< Size of allocated buffer */
-	struct vmw_dma_buffer *buffer; /**< Backing store buffer */
-
-	bool defined;
-	bool active_implicit;
-};
-
-static void vmw_sou_destroy(struct vmw_screen_object_unit *sou)
-{
-	vmw_display_unit_cleanup(&sou->base);
-	kfree(sou);
-}
-
-
-/*
- * Screen Object Display Unit CRTC functions
- */
-
-static void vmw_sou_crtc_destroy(struct drm_crtc *crtc)
-{
-	vmw_sou_destroy(vmw_crtc_to_sou(crtc));
-}
-
-static void vmw_sou_del_active(struct vmw_private *vmw_priv,
-			      struct vmw_screen_object_unit *sou)
-{
-	struct vmw_screen_object_display *ld = vmw_priv->sou_priv;
-
-	if (sou->active_implicit) {
-		if (--(ld->num_implicit) == 0)
-			ld->implicit_fb = NULL;
-		sou->active_implicit = false;
-	}
-}
-
-static void vmw_sou_add_active(struct vmw_private *vmw_priv,
-			      struct vmw_screen_object_unit *sou,
-			      struct vmw_framebuffer *vfb)
-{
-	struct vmw_screen_object_display *ld = vmw_priv->sou_priv;
-
-	BUG_ON(!ld->num_implicit && ld->implicit_fb);
-
-	if (!sou->active_implicit && sou->base.is_implicit) {
-		ld->implicit_fb = vfb;
-		sou->active_implicit = true;
-		ld->num_implicit++;
-	}
-}
-
-/**
- * Send the fifo command to create a screen.
- */
-static int vmw_sou_fifo_create(struct vmw_private *dev_priv,
-			       struct vmw_screen_object_unit *sou,
-			       uint32_t x, uint32_t y,
-			       struct drm_display_mode *mode)
-{
-	size_t fifo_size;
-
-	struct {
-		struct {
-			uint32_t cmdType;
-		} header;
-		SVGAScreenObject obj;
-	} *cmd;
-
-	BUG_ON(!sou->buffer);
-
-	fifo_size = sizeof(*cmd);
-	cmd = vmw_fifo_reserve(dev_priv, fifo_size);
-	/* The hardware has hung, nothing we can do about it here. */
-	if (unlikely(cmd == NULL)) {
-		DRM_ERROR("Fifo reserve failed.\n");
-		return -ENOMEM;
-	}
-
-	memset(cmd, 0, fifo_size);
-	cmd->header.cmdType = SVGA_CMD_DEFINE_SCREEN;
-	cmd->obj.structSize = sizeof(SVGAScreenObject);
-	cmd->obj.id = sou->base.unit;
-	cmd->obj.flags = SVGA_SCREEN_HAS_ROOT |
-		(sou->base.unit == 0 ? SVGA_SCREEN_IS_PRIMARY : 0);
-	cmd->obj.size.width = mode->hdisplay;
-	cmd->obj.size.height = mode->vdisplay;
-	if (sou->base.is_implicit) {
-		cmd->obj.root.x = x;
-		cmd->obj.root.y = y;
-	} else {
-		cmd->obj.root.x = sou->base.gui_x;
-		cmd->obj.root.y = sou->base.gui_y;
-	}
-
-	/* Ok to assume that buffer is pinned in vram */
-	vmw_bo_get_guest_ptr(&sou->buffer->base, &cmd->obj.backingStore.ptr);
-	cmd->obj.backingStore.pitch = mode->hdisplay * 4;
-
-	vmw_fifo_commit(dev_priv, fifo_size);
-
-	sou->defined = true;
-
-	return 0;
-}
-
-/**
- * Send the fifo command to destroy a screen.
- */
-static int vmw_sou_fifo_destroy(struct vmw_private *dev_priv,
-				struct vmw_screen_object_unit *sou)
-{
-	size_t fifo_size;
-	int ret;
-
-	struct {
-		struct {
-			uint32_t cmdType;
-		} header;
-		SVGAFifoCmdDestroyScreen body;
-	} *cmd;
-
-	/* no need to do anything */
-	if (unlikely(!sou->defined))
-		return 0;
-
-	fifo_size = sizeof(*cmd);
-	cmd = vmw_fifo_reserve(dev_priv, fifo_size);
-	/* the hardware has hung, nothing we can do about it here */
-	if (unlikely(cmd == NULL)) {
-		DRM_ERROR("Fifo reserve failed.\n");
-		return -ENOMEM;
-	}
-
-	memset(cmd, 0, fifo_size);
-	cmd->header.cmdType = SVGA_CMD_DESTROY_SCREEN;
-	cmd->body.screenId = sou->base.unit;
-
-	vmw_fifo_commit(dev_priv, fifo_size);
-
-	/* Force sync */
-	ret = vmw_fallback_wait(dev_priv, false, true, 0, false, 3*HZ);
-	if (unlikely(ret != 0))
-		DRM_ERROR("Failed to sync with HW");
-	else
-		sou->defined = false;
-
-	return ret;
-}
-
-/**
- * Free the backing store.
- */
-static void vmw_sou_backing_free(struct vmw_private *dev_priv,
-				 struct vmw_screen_object_unit *sou)
-{
-	struct ttm_buffer_object *bo;
-
-	if (unlikely(sou->buffer == NULL))
-		return;
-
-	bo = &sou->buffer->base;
-	ttm_bo_unref(&bo);
-	sou->buffer = NULL;
-	sou->buffer_size = 0;
-}
-
-/**
- * Allocate the backing store for the buffer.
- */
-static int vmw_sou_backing_alloc(struct vmw_private *dev_priv,
-				 struct vmw_screen_object_unit *sou,
-				 unsigned long size)
-{
-	int ret;
-
-	if (sou->buffer_size == size)
-		return 0;
-
-	if (sou->buffer)
-		vmw_sou_backing_free(dev_priv, sou);
-
-	sou->buffer = kzalloc(sizeof(*sou->buffer), GFP_KERNEL);
-	if (unlikely(sou->buffer == NULL))
-		return -ENOMEM;
-
-	/* After we have alloced the backing store might not be able to
-	 * resume the overlays, this is preferred to failing to alloc.
-	 */
-	vmw_overlay_pause_all(dev_priv);
-	ret = vmw_dmabuf_init(dev_priv, sou->buffer, size,
-			      &vmw_vram_ne_placement,
-			      false, &vmw_dmabuf_bo_free);
-	vmw_overlay_resume_all(dev_priv);
-
-	if (unlikely(ret != 0))
-		sou->buffer = NULL; /* vmw_dmabuf_init frees on error */
-	else
-		sou->buffer_size = size;
-
-	return ret;
-}
-
-static int vmw_sou_crtc_set_config(struct drm_mode_set *set)
-{
-	struct vmw_private *dev_priv;
-	struct vmw_screen_object_unit *sou;
-	struct drm_connector *connector;
-	struct drm_display_mode *mode;
-	struct drm_encoder *encoder;
-	struct vmw_framebuffer *vfb;
-	struct drm_framebuffer *fb;
-	struct drm_crtc *crtc;
-	int ret = 0;
-
-	if (!set)
-		return -EINVAL;
-
-	if (!set->crtc)
-		return -EINVAL;
-
-	/* get the sou */
-	crtc = set->crtc;
-	sou = vmw_crtc_to_sou(crtc);
-	vfb = set->fb ? vmw_framebuffer_to_vfb(set->fb) : NULL;
-	dev_priv = vmw_priv(crtc->dev);
-
-	if (set->num_connectors > 1) {
-		DRM_ERROR("to many connectors\n");
-		return -EINVAL;
-	}
-
-	if (set->num_connectors == 1 &&
-	    set->connectors[0] != &sou->base.connector) {
-		DRM_ERROR("connector doesn't match %p %p\n",
-			set->connectors[0], &sou->base.connector);
-		return -EINVAL;
-	}
-
-	/* sou only supports one fb active at the time */
-	if (sou->base.is_implicit &&
-	    dev_priv->sou_priv->implicit_fb && vfb &&
-	    !(dev_priv->sou_priv->num_implicit == 1 &&
-	      sou->active_implicit) &&
-	    dev_priv->sou_priv->implicit_fb != vfb) {
-		DRM_ERROR("Multiple framebuffers not supported\n");
-		return -EINVAL;
-	}
-
-	/* since they always map one to one these are safe */
-	connector = &sou->base.connector;
-	encoder = &sou->base.encoder;
-
-	/* should we turn the crtc off */
-	if (set->num_connectors == 0 || !set->mode || !set->fb) {
-		ret = vmw_sou_fifo_destroy(dev_priv, sou);
-		/* the hardware has hung don't do anything more */
-		if (unlikely(ret != 0))
-			return ret;
-
-		connector->encoder = NULL;
-		encoder->crtc = NULL;
-		crtc->fb = NULL;
-		crtc->x = 0;
-		crtc->y = 0;
-
-		vmw_sou_del_active(dev_priv, sou);
-
-		vmw_sou_backing_free(dev_priv, sou);
-
-		return 0;
-	}
-
-
-	/* we now know we want to set a mode */
-	mode = set->mode;
-	fb = set->fb;
-
-	if (set->x + mode->hdisplay > fb->width ||
-	    set->y + mode->vdisplay > fb->height) {
-		DRM_ERROR("set outside of framebuffer\n");
-		return -EINVAL;
-	}
-
-	vmw_fb_off(dev_priv);
-
-	if (mode->hdisplay != crtc->mode.hdisplay ||
-	    mode->vdisplay != crtc->mode.vdisplay) {
-		/* no need to check if depth is different, because backing
-		 * store depth is forced to 4 by the device.
-		 */
-
-		ret = vmw_sou_fifo_destroy(dev_priv, sou);
-		/* the hardware has hung don't do anything more */
-		if (unlikely(ret != 0))
-			return ret;
-
-		vmw_sou_backing_free(dev_priv, sou);
-	}
-
-	if (!sou->buffer) {
-		/* forced to depth 4 by the device */
-		size_t size = mode->hdisplay * mode->vdisplay * 4;
-		ret = vmw_sou_backing_alloc(dev_priv, sou, size);
-		if (unlikely(ret != 0))
-			return ret;
-	}
-
-	ret = vmw_sou_fifo_create(dev_priv, sou, set->x, set->y, mode);
-	if (unlikely(ret != 0)) {
-		/*
-		 * We are in a bit of a situation here, the hardware has
-		 * hung and we may or may not have a buffer hanging of
-		 * the screen object, best thing to do is not do anything
-		 * if we where defined, if not just turn the crtc of.
-		 * Not what userspace wants but it needs to htfu.
-		 */
-		if (sou->defined)
-			return ret;
-
-		connector->encoder = NULL;
-		encoder->crtc = NULL;
-		crtc->fb = NULL;
-		crtc->x = 0;
-		crtc->y = 0;
-
-		return ret;
-	}
-
-	vmw_sou_add_active(dev_priv, sou, vfb);
-
-	connector->encoder = encoder;
-	encoder->crtc = crtc;
-	crtc->mode = *mode;
-	crtc->fb = fb;
-	crtc->x = set->x;
-	crtc->y = set->y;
-
-	return 0;
-}
-
-static struct drm_crtc_funcs vmw_screen_object_crtc_funcs = {
-	.save = vmw_du_crtc_save,
-	.restore = vmw_du_crtc_restore,
-	.cursor_set = vmw_du_crtc_cursor_set,
-	.cursor_move = vmw_du_crtc_cursor_move,
-	.gamma_set = vmw_du_crtc_gamma_set,
-	.destroy = vmw_sou_crtc_destroy,
-	.set_config = vmw_sou_crtc_set_config,
-	.page_flip = vmw_du_page_flip,
-};
-
-/*
- * Screen Object Display Unit encoder functions
- */
-
-static void vmw_sou_encoder_destroy(struct drm_encoder *encoder)
-{
-	vmw_sou_destroy(vmw_encoder_to_sou(encoder));
-}
-
-static struct drm_encoder_funcs vmw_screen_object_encoder_funcs = {
-	.destroy = vmw_sou_encoder_destroy,
-};
-
-/*
- * Screen Object Display Unit connector functions
- */
-
-static void vmw_sou_connector_destroy(struct drm_connector *connector)
-{
-	vmw_sou_destroy(vmw_connector_to_sou(connector));
-}
-
-static struct drm_connector_funcs vmw_legacy_connector_funcs = {
-	.dpms = vmw_du_connector_dpms,
-	.save = vmw_du_connector_save,
-	.restore = vmw_du_connector_restore,
-	.detect = vmw_du_connector_detect,
-	.fill_modes = vmw_du_connector_fill_modes,
-	.set_property = vmw_du_connector_set_property,
-	.destroy = vmw_sou_connector_destroy,
-};
-
-static int vmw_sou_init(struct vmw_private *dev_priv, unsigned unit)
-{
-	struct vmw_screen_object_unit *sou;
-	struct drm_device *dev = dev_priv->dev;
-	struct drm_connector *connector;
-	struct drm_encoder *encoder;
-	struct drm_crtc *crtc;
-
-	sou = kzalloc(sizeof(*sou), GFP_KERNEL);
-	if (!sou)
-		return -ENOMEM;
-
-	sou->base.unit = unit;
-	crtc = &sou->base.crtc;
-	encoder = &sou->base.encoder;
-	connector = &sou->base.connector;
-
-	sou->active_implicit = false;
-
-	sou->base.pref_active = (unit == 0);
-	sou->base.pref_width = dev_priv->initial_width;
-	sou->base.pref_height = dev_priv->initial_height;
-	sou->base.pref_mode = NULL;
-	sou->base.is_implicit = true;
-
-	drm_connector_init(dev, connector, &vmw_legacy_connector_funcs,
-			   DRM_MODE_CONNECTOR_VIRTUAL);
-	connector->status = vmw_du_connector_detect(connector, true);
-
-	drm_encoder_init(dev, encoder, &vmw_screen_object_encoder_funcs,
-			 DRM_MODE_ENCODER_VIRTUAL);
-	drm_mode_connector_attach_encoder(connector, encoder);
-	encoder->possible_crtcs = (1 << unit);
-	encoder->possible_clones = 0;
-
-	drm_crtc_init(dev, crtc, &vmw_screen_object_crtc_funcs);
-
-	drm_mode_crtc_set_gamma_size(crtc, 256);
-
-	drm_connector_attach_property(connector,
-				      dev->mode_config.dirty_info_property,
-				      1);
-
-	return 0;
-}
-
-int vmw_kms_init_screen_object_display(struct vmw_private *dev_priv)
-{
-	struct drm_device *dev = dev_priv->dev;
-	int i, ret;
-
-	if (dev_priv->sou_priv) {
-		DRM_INFO("sou system already on\n");
-		return -EINVAL;
-	}
-
-	if (!(dev_priv->fifo.capabilities & SVGA_FIFO_CAP_SCREEN_OBJECT_2)) {
-		DRM_INFO("Not using screen objects,"
-			 " missing cap SCREEN_OBJECT_2\n");
-		return -ENOSYS;
-	}
-
-	ret = -ENOMEM;
-	dev_priv->sou_priv = kmalloc(sizeof(*dev_priv->sou_priv), GFP_KERNEL);
-	if (unlikely(!dev_priv->sou_priv))
-		goto err_no_mem;
-
-	dev_priv->sou_priv->num_implicit = 0;
-	dev_priv->sou_priv->implicit_fb = NULL;
-
-	ret = drm_vblank_init(dev, VMWGFX_NUM_DISPLAY_UNITS);
-	if (unlikely(ret != 0))
-		goto err_free;
-
-	ret = drm_mode_create_dirty_info_property(dev);
-	if (unlikely(ret != 0))
-		goto err_vblank_cleanup;
-
-	for (i = 0; i < VMWGFX_NUM_DISPLAY_UNITS; ++i)
-		vmw_sou_init(dev_priv, i);
-
-	DRM_INFO("Screen objects system initialized\n");
-
-	return 0;
-
-err_vblank_cleanup:
-	drm_vblank_cleanup(dev);
-err_free:
-	kfree(dev_priv->sou_priv);
-	dev_priv->sou_priv = NULL;
-err_no_mem:
-	return ret;
-}
-
-int vmw_kms_close_screen_object_display(struct vmw_private *dev_priv)
-{
-	struct drm_device *dev = dev_priv->dev;
-
-	if (!dev_priv->sou_priv)
-		return -ENOSYS;
-
-	drm_vblank_cleanup(dev);
-
-	kfree(dev_priv->sou_priv);
-
-	return 0;
-}
-
-/**
- * Returns if this unit can be page flipped.
- * Must be called with the mode_config mutex held.
- */
-bool vmw_kms_screen_object_flippable(struct vmw_private *dev_priv,
-				     struct drm_crtc *crtc)
-{
-	struct vmw_screen_object_unit *sou = vmw_crtc_to_sou(crtc);
-
-	if (!sou->base.is_implicit)
-		return true;
-
-	if (dev_priv->sou_priv->num_implicit != 1)
-		return false;
-
-	return true;
-}
-
-/**
- * Update the implicit fb to the current fb of this crtc.
- * Must be called with the mode_config mutex held.
- */
-void vmw_kms_screen_object_update_implicit_fb(struct vmw_private *dev_priv,
-					      struct drm_crtc *crtc)
-{
-	struct vmw_screen_object_unit *sou = vmw_crtc_to_sou(crtc);
-
-	BUG_ON(!sou->base.is_implicit);
-
-	dev_priv->sou_priv->implicit_fb =
-		vmw_framebuffer_to_vfb(sou->base.crtc.fb);
-}
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_glue.c b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_glue.c
index d3c11f5..1e8eedd 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_glue.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_ttm_glue.c
@@ -34,8 +34,9 @@ int vmw_mmap(struct file *filp, struct vm_area_struct *vma)
 	struct vmw_private *dev_priv;
 
 	if (unlikely(vma->vm_pgoff < VMWGFX_FILE_PAGE_OFFSET)) {
-		DRM_ERROR("Illegal attempt to mmap old fifo space.\n");
-		return -EINVAL;
+		if (vmw_fifo_mmap(filp, vma) == 0)
+			return 0;
+		return drm_mmap(filp, vma);
 	}
 
 	file_priv = filp->private_data;