1 files changed, 165 insertions, 1490 deletions

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;
-}