Merge remote-tracking branch 'mesa/13.0' into nougat-x86

8 files changed, 86 insertions, 70 deletions

diff --git a/src/gallium/auxiliary/hud/hud_cpufreq.c b/src/gallium/auxiliary/hud/hud_cpufreq.c
index 4501bbb..19a6f08 100644
--- a/src/gallium/auxiliary/hud/hud_cpufreq.c
+++ b/src/gallium/auxiliary/hud/hud_cpufreq.c
@@ -36,6 +36,7 @@
 #include "hud/hud_private.h"
 #include "util/list.h"
 #include "os/os_time.h"
+#include "os/os_thread.h"
 #include "util/u_memory.h"
 #include <stdio.h>
 #include <unistd.h>
@@ -61,6 +62,7 @@ struct cpufreq_info
 
 static int gcpufreq_count = 0;
 static struct list_head gcpufreq_list;
+pipe_static_mutex(gcpufreq_mutex);
 
 static struct cpufreq_info *
 find_cfi_by_index(int cpu_index, int mode)
@@ -112,14 +114,6 @@ query_cfi_load(struct hud_graph *gr)
    }
 }
 
-static void
-free_query_data(void *p)
-{
-   struct cpufreq_info *cfi = (struct cpufreq_info *)p;
-   list_del(&cfi->list);
-   FREE(cfi);
-}
-
 /**
   * Create and initialize a new object for a specific CPU.
   * \param  pane  parent context.
@@ -162,11 +156,6 @@ hud_cpufreq_graph_install(struct hud_pane *pane, int cpu_index,
    gr->query_data = cfi;
    gr->query_new_value = query_cfi_load;
 
-   /* Don't use free() as our callback as that messes up Gallium's
-    * memory debugger.  Use simple free_query_data() wrapper.
-    */
-   gr->free_query_data = free_query_data;
-
    hud_pane_add_graph(pane, gr);
    hud_pane_set_max_value(pane, 3000000 /* 3 GHz */);
 }
@@ -199,16 +188,21 @@ hud_get_num_cpufreq(bool displayhelp)
    int cpu_index;
 
    /* Return the number of CPU metrics we support. */
-   if (gcpufreq_count)
+   pipe_mutex_lock(gcpufreq_mutex);
+   if (gcpufreq_count) {
+      pipe_mutex_unlock(gcpufreq_mutex);
       return gcpufreq_count;
+   }
 
    /* Scan /sys/devices.../cpu, for every object type we support, create
     * and persist an object to represent its different metrics.
     */
    list_inithead(&gcpufreq_list);
    DIR *dir = opendir("/sys/devices/system/cpu");
-   if (!dir)
+   if (!dir) {
+      pipe_mutex_unlock(gcpufreq_mutex);
       return 0;
+   }
 
    while ((dp = readdir(dir)) != NULL) {
 
@@ -238,6 +232,7 @@ hud_get_num_cpufreq(bool displayhelp)
       snprintf(fn, sizeof(fn), "%s/cpufreq/scaling_max_freq", basename);
       add_object(dp->d_name, fn, CPUFREQ_MAXIMUM, cpu_index);
    }
+   closedir(dir);
 
    if (displayhelp) {
       list_for_each_entry(struct cpufreq_info, cfi, &gcpufreq_list, list) {
@@ -251,6 +246,7 @@ hud_get_num_cpufreq(bool displayhelp)
       }
    }
 
+   pipe_mutex_unlock(gcpufreq_mutex);
    return gcpufreq_count;
 }
 
diff --git a/src/gallium/auxiliary/hud/hud_diskstat.c b/src/gallium/auxiliary/hud/hud_diskstat.c
index b248baf..af6e62d 100644
--- a/src/gallium/auxiliary/hud/hud_diskstat.c
+++ b/src/gallium/auxiliary/hud/hud_diskstat.c
@@ -35,6 +35,7 @@
 #include "hud/hud_private.h"
 #include "util/list.h"
 #include "os/os_time.h"
+#include "os/os_thread.h"
 #include "util/u_memory.h"
 #include <stdio.h>
 #include <unistd.h>
@@ -81,6 +82,7 @@ struct diskstat_info
  */
 static int gdiskstat_count = 0;
 static struct list_head gdiskstat_list;
+pipe_static_mutex(gdiskstat_mutex);
 
 static struct diskstat_info *
 find_dsi_by_name(const char *n, int mode)
@@ -162,14 +164,6 @@ query_dsi_load(struct hud_graph *gr)
    }
 }
 
-static void
-free_query_data(void *p)
-{
-   struct diskstat_info *nic = (struct diskstat_info *) p;
-   list_del(&nic->list);
-   FREE(nic);
-}
-
 /**
   * Create and initialize a new object for a specific block I/O device.
   * \param  pane  parent context.
@@ -208,11 +202,6 @@ hud_diskstat_graph_install(struct hud_pane *pane, const char *dev_name,
    gr->query_data = dsi;
    gr->query_new_value = query_dsi_load;
 
-   /* Don't use free() as our callback as that messes up Gallium's
-    * memory debugger.  Use simple free_query_data() wrapper.
-    */
-   gr->free_query_data = free_query_data;
-
    hud_pane_add_graph(pane, gr);
    hud_pane_set_max_value(pane, 100);
 }
@@ -257,16 +246,21 @@ hud_get_num_disks(bool displayhelp)
    char name[64];
 
    /* Return the number of block devices and partitions. */
-   if (gdiskstat_count)
+   pipe_mutex_lock(gdiskstat_mutex);
+   if (gdiskstat_count) {
+      pipe_mutex_unlock(gdiskstat_mutex);
       return gdiskstat_count;
+   }
 
    /* Scan /sys/block, for every object type we support, create and
     * persist an object to represent its different statistics.
     */
    list_inithead(&gdiskstat_list);
    DIR *dir = opendir("/sys/block/");
-   if (!dir)
+   if (!dir) {
+      pipe_mutex_unlock(gdiskstat_mutex);
       return 0;
+   }
 
    while ((dp = readdir(dir)) != NULL) {
 
@@ -290,8 +284,11 @@ hud_get_num_disks(bool displayhelp)
       /* Add any partitions */
       struct dirent *dpart;
       DIR *pdir = opendir(basename);
-      if (!pdir)
+      if (!pdir) {
+         pipe_mutex_unlock(gdiskstat_mutex);
+         closedir(dir);
          return 0;
+      }
 
       while ((dpart = readdir(pdir)) != NULL) {
          /* Avoid 'lo' and '..' and '.' */
@@ -311,6 +308,7 @@ hud_get_num_disks(bool displayhelp)
          add_object_part(basename, dpart->d_name, DISKSTAT_WR);
       }
    }
+   closedir(dir);
 
    if (displayhelp) {
       list_for_each_entry(struct diskstat_info, dsi, &gdiskstat_list, list) {
@@ -322,6 +320,7 @@ hud_get_num_disks(bool displayhelp)
          puts(line);
       }
    }
+   pipe_mutex_unlock(gdiskstat_mutex);
 
    return gdiskstat_count;
 }
diff --git a/src/gallium/auxiliary/hud/hud_nic.c b/src/gallium/auxiliary/hud/hud_nic.c
index fb6b8c0..f9935de 100644
--- a/src/gallium/auxiliary/hud/hud_nic.c
+++ b/src/gallium/auxiliary/hud/hud_nic.c
@@ -35,6 +35,7 @@
 #include "hud/hud_private.h"
 #include "util/list.h"
 #include "os/os_time.h"
+#include "os/os_thread.h"
 #include "util/u_memory.h"
 #include <stdio.h>
 #include <unistd.h>
@@ -66,6 +67,7 @@ struct nic_info
  */
 static int gnic_count = 0;
 static struct list_head gnic_list;
+pipe_static_mutex(gnic_mutex);
 
 static struct nic_info *
 find_nic_by_name(const char *n, int mode)
@@ -234,14 +236,6 @@ query_nic_load(struct hud_graph *gr)
    }
 }
 
-static void
-free_query_data(void *p)
-{
-   struct nic_info *nic = (struct nic_info *) p;
-   list_del(&nic->list);
-   FREE(nic);
-}
-
 /**
   * Create and initialize a new object for a specific network interface dev.
   * \param  pane  parent context.
@@ -284,11 +278,6 @@ hud_nic_graph_install(struct hud_pane *pane, const char *nic_name,
    gr->query_data = nic;
    gr->query_new_value = query_nic_load;
 
-   /* Don't use free() as our callback as that messes up Gallium's
-    * memory debugger.  Use simple free_query_data() wrapper.
-    */
-   gr->free_query_data = free_query_data;
-
    hud_pane_add_graph(pane, gr);
    hud_pane_set_max_value(pane, 100);
 }
@@ -342,16 +331,21 @@ hud_get_num_nics(bool displayhelp)
    char name[64];
 
    /* Return the number if network interfaces. */
-   if (gnic_count)
+   pipe_mutex_lock(gnic_mutex);
+   if (gnic_count) {
+      pipe_mutex_unlock(gnic_mutex);
       return gnic_count;
+   }
 
    /* Scan /sys/block, for every object type we support, create and
     * persist an object to represent its different statistics.
     */
    list_inithead(&gnic_list);
    DIR *dir = opendir("/sys/class/net/");
-   if (!dir)
+   if (!dir) {
+      pipe_mutex_unlock(gnic_mutex);
       return 0;
+   }
 
    while ((dp = readdir(dir)) != NULL) {
 
@@ -412,6 +406,7 @@ hud_get_num_nics(bool displayhelp)
       }
 
    }
+   closedir(dir);
 
    list_for_each_entry(struct nic_info, nic, &gnic_list, list) {
       char line[64];
@@ -424,6 +419,7 @@ hud_get_num_nics(bool displayhelp)
 
    }
 
+   pipe_mutex_unlock(gnic_mutex);
    return gnic_count;
 }
 
diff --git a/src/gallium/auxiliary/hud/hud_sensors_temp.c b/src/gallium/auxiliary/hud/hud_sensors_temp.c
index e41b847..11b8a4c 100644
--- a/src/gallium/auxiliary/hud/hud_sensors_temp.c
+++ b/src/gallium/auxiliary/hud/hud_sensors_temp.c
@@ -32,6 +32,7 @@
 #include "hud/hud_private.h"
 #include "util/list.h"
 #include "os/os_time.h"
+#include "os/os_thread.h"
 #include "util/u_memory.h"
 #include <stdio.h>
 #include <unistd.h>
@@ -49,6 +50,7 @@
  */
 static int gsensors_temp_count = 0;
 static struct list_head gsensors_temp_list;
+pipe_static_mutex(gsensor_temp_mutex);
 
 struct sensors_temp_info
 {
@@ -189,17 +191,6 @@ query_sti_load(struct hud_graph *gr)
    }
 }
 
-static void
-free_query_data(void *p)
-{
-   struct sensors_temp_info *sti = (struct sensors_temp_info *) p;
-   list_del(&sti->list);
-   if (sti->chip)
-      sensors_free_chip_name(sti->chip);
-   FREE(sti);
-   sensors_cleanup();
-}
-
 /**
   * Create and initialize a new object for a specific sensor interface dev.
   * \param  pane  parent context.
@@ -237,11 +228,6 @@ hud_sensors_temp_graph_install(struct hud_pane *pane, const char *dev_name,
    gr->query_data = sti;
    gr->query_new_value = query_sti_load;
 
-   /* Don't use free() as our callback as that messes up Gallium's
-    * memory debugger.  Use simple free_query_data() wrapper.
-    */
-   gr->free_query_data = free_query_data;
-
    hud_pane_add_graph(pane, gr);
    switch (sti->mode) {
    case SENSORS_TEMP_CURRENT:
@@ -338,12 +324,17 @@ int
 hud_get_num_sensors(bool displayhelp)
 {
    /* Return the number of sensors detected. */
-   if (gsensors_temp_count)
+   pipe_mutex_lock(gsensor_temp_mutex);
+   if (gsensors_temp_count) {
+      pipe_mutex_unlock(gsensor_temp_mutex);
       return gsensors_temp_count;
+   }
 
    int ret = sensors_init(NULL);
-   if (ret)
+   if (ret) {
+      pipe_mutex_unlock(gsensor_temp_mutex);
       return 0;
+   }
 
    list_inithead(&gsensors_temp_list);
 
@@ -377,6 +368,7 @@ hud_get_num_sensors(bool displayhelp)
       }
    }
 
+   pipe_mutex_unlock(gsensor_temp_mutex);
    return gsensors_temp_count;
 }
 
diff --git a/src/gallium/drivers/radeonsi/si_blit.c b/src/gallium/drivers/radeonsi/si_blit.c
index db41f56..dd8f83b 100644
--- a/src/gallium/drivers/radeonsi/si_blit.c
+++ b/src/gallium/drivers/radeonsi/si_blit.c
@@ -487,7 +487,9 @@ si_decompress_sampler_color_textures(struct si_context *sctx,
 		assert(view);
 
 		tex = (struct r600_texture *)view->texture;
-		assert(tex->cmask.size || tex->fmask.size || tex->dcc_offset);
+		/* CMASK or DCC can be discarded and we can still end up here. */
+		if (!tex->cmask.size && !tex->fmask.size && !tex->dcc_offset)
+			continue;
 
 		si_blit_decompress_color(&sctx->b.b, tex,
 					 view->u.tex.first_level, view->u.tex.last_level,
diff --git a/src/gallium/drivers/radeonsi/si_shader_tgsi_alu.c b/src/gallium/drivers/radeonsi/si_shader_tgsi_alu.c
index 123ff5d..18e905b 100644
--- a/src/gallium/drivers/radeonsi/si_shader_tgsi_alu.c
+++ b/src/gallium/drivers/radeonsi/si_shader_tgsi_alu.c
@@ -459,6 +459,8 @@ static void emit_bfi(const struct lp_build_tgsi_action *action,
 	struct gallivm_state *gallivm = bld_base->base.gallivm;
 	LLVMBuilderRef builder = gallivm->builder;
 	LLVMValueRef bfi_args[3];
+	LLVMValueRef bfi_sm5;
+	LLVMValueRef cond;
 
 	// Calculate the bitmask: (((1 << src3) - 1) << src2
 	bfi_args[0] = LLVMBuildShl(builder,
@@ -478,11 +480,40 @@ static void emit_bfi(const struct lp_build_tgsi_action *action,
 	 *   (arg0 & arg1) | (~arg0 & arg2) = arg2 ^ (arg0 & (arg1 ^ arg2)
 	 * Use the right-hand side, which the LLVM backend can convert to V_BFI.
 	 */
-	emit_data->output[emit_data->chan] =
+	bfi_sm5 =
 		LLVMBuildXor(builder, bfi_args[2],
 			LLVMBuildAnd(builder, bfi_args[0],
 				LLVMBuildXor(builder, bfi_args[1], bfi_args[2],
 					     ""), ""), "");
+
+	/* Since shifts of >= 32 bits are undefined in LLVM IR, the backend
+	 * uses the convenient V_BFI lowering for the above, which follows SM5
+	 * and disagrees with GLSL semantics when bits (src3) is 32.
+	 */
+	cond = LLVMBuildICmp(builder, LLVMIntUGE, emit_data->args[3],
+			     lp_build_const_int32(gallivm, 32), "");
+	emit_data->output[emit_data->chan] =
+		LLVMBuildSelect(builder, cond, emit_data->args[1], bfi_sm5, "");
+}
+
+static void emit_bfe(const struct lp_build_tgsi_action *action,
+		     struct lp_build_tgsi_context *bld_base,
+		     struct lp_build_emit_data *emit_data)
+{
+	struct gallivm_state *gallivm = bld_base->base.gallivm;
+	LLVMBuilderRef builder = gallivm->builder;
+	LLVMValueRef bfe_sm5;
+	LLVMValueRef cond;
+
+	bfe_sm5 = lp_build_intrinsic(builder, action->intr_name,
+				     emit_data->dst_type, emit_data->args,
+				     emit_data->arg_count, LLVMReadNoneAttribute);
+
+	/* Correct for GLSL semantics. */
+	cond = LLVMBuildICmp(builder, LLVMIntUGE, emit_data->args[2],
+			     lp_build_const_int32(gallivm, 32), "");
+	emit_data->output[emit_data->chan] =
+		LLVMBuildSelect(builder, cond, emit_data->args[0], bfe_sm5, "");
 }
 
 /* this is ffs in C */
@@ -783,7 +814,7 @@ void si_shader_context_init_alu(struct lp_build_tgsi_context *bld_base)
 	bld_base->op_actions[TGSI_OPCODE_FSLT].emit = emit_fcmp;
 	bld_base->op_actions[TGSI_OPCODE_FSNE].emit = emit_fcmp;
 	bld_base->op_actions[TGSI_OPCODE_IABS].emit = emit_iabs;
-	bld_base->op_actions[TGSI_OPCODE_IBFE].emit = build_tgsi_intrinsic_nomem;
+	bld_base->op_actions[TGSI_OPCODE_IBFE].emit = emit_bfe;
 	bld_base->op_actions[TGSI_OPCODE_IBFE].intr_name = "llvm.AMDGPU.bfe.i32";
 	bld_base->op_actions[TGSI_OPCODE_IDIV].emit = emit_idiv;
 	bld_base->op_actions[TGSI_OPCODE_IMAX].emit = emit_minmax_int;
@@ -835,7 +866,7 @@ void si_shader_context_init_alu(struct lp_build_tgsi_context *bld_base)
 	bld_base->op_actions[TGSI_OPCODE_TRUNC].emit = build_tgsi_intrinsic_nomem;
 	bld_base->op_actions[TGSI_OPCODE_TRUNC].intr_name = "llvm.trunc.f32";
 	bld_base->op_actions[TGSI_OPCODE_UADD].emit = emit_uadd;
-	bld_base->op_actions[TGSI_OPCODE_UBFE].emit = build_tgsi_intrinsic_nomem;
+	bld_base->op_actions[TGSI_OPCODE_UBFE].emit = emit_bfe;
 	bld_base->op_actions[TGSI_OPCODE_UBFE].intr_name = "llvm.AMDGPU.bfe.u32";
 	bld_base->op_actions[TGSI_OPCODE_UDIV].emit = emit_udiv;
 	bld_base->op_actions[TGSI_OPCODE_UMAX].emit = emit_minmax_int;
diff --git a/src/gallium/drivers/vc4/vc4_program.c b/src/gallium/drivers/vc4/vc4_program.c
index 81ac070..0145488 100644
--- a/src/gallium/drivers/vc4/vc4_program.c
+++ b/src/gallium/drivers/vc4/vc4_program.c
@@ -1370,7 +1370,7 @@ emit_vert_end(struct vc4_compile *c,
               struct vc4_varying_slot *fs_inputs,
               uint32_t num_fs_inputs)
 {
-        struct qreg rcp_w = qir_RCP(c, c->outputs[c->output_position_index + 3]);
+        struct qreg rcp_w = ntq_rcp(c, c->outputs[c->output_position_index + 3]);
 
         emit_stub_vpm_read(c);
 
diff --git a/src/gallium/state_trackers/vdpau/output.c b/src/gallium/state_trackers/vdpau/output.c
index f4d62a3..8c29a3f 100644
--- a/src/gallium/state_trackers/vdpau/output.c
+++ b/src/gallium/state_trackers/vdpau/output.c
@@ -82,7 +82,7 @@ vlVdpOutputSurfaceCreate(VdpDevice device,
    res_tmpl.depth0 = 1;
    res_tmpl.array_size = 1;
    res_tmpl.bind = PIPE_BIND_SAMPLER_VIEW | PIPE_BIND_RENDER_TARGET |
-                   PIPE_BIND_LINEAR | PIPE_BIND_SHARED;
+                   PIPE_BIND_SHARED;
    res_tmpl.usage = PIPE_USAGE_DEFAULT;
 
    pipe_mutex_lock(dev->mutex);