radv: Add RADV_PROFILE_PSTATE envvar

Enable selecting the specific pstate to enter when using thread tracing
and when acquiring the profiling lock for performance queries.

Signed-off-by: Josh Simmons <josh@nega.tv>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30139>

docs/envvars.rst

@@ -1465,6 +1465,20 @@ RADV driver environment variables
    decrease the resolution used for dumping the ray history resolution when capturing
    RRA traces. This allows for dumping every Nth invocation along each dispatch dimension.
 
+.. envvar:: RADV_PROFILE_PSTATE
+
+   choose the specific pstate to enter when using thread tracing or when acquiring the
+   profiling lock for performance queries.
+
+   ``standard``
+      force GPU clocks to an arbitrary fixed level
+   ``min_sclk``
+      force the shader clock to its minimum level
+   ``min_mclk``
+      force the memory clock to its minimum level
+   ``peak``
+      force GPU clocks to their maximum level, this is the default value
+
 .. envvar:: ACO_DEBUG
 
    a comma-separated list of named flags, which do various things:

src/amd/vulkan/radv_device.c

@@ -1863,8 +1863,9 @@ bool
 radv_device_set_pstate(struct radv_device *device, bool enable)
 {
    const struct radv_physical_device *pdev = radv_device_physical(device);
+   const struct radv_instance *instance = radv_physical_device_instance(pdev);
    struct radeon_winsys *ws = device->ws;
-   enum radeon_ctx_pstate pstate = enable ? RADEON_CTX_PSTATE_PEAK : RADEON_CTX_PSTATE_NONE;
+   enum radeon_ctx_pstate pstate = enable ? instance->profile_pstate : RADEON_CTX_PSTATE_NONE;
 
    if (pdev->info.has_stable_pstate) {
       /* pstate is per-device; setting it for one ctx is sufficient.

src/amd/vulkan/radv_instance.c

@@ -308,6 +308,22 @@ radv_handle_legacy_sqtt_trigger(struct vk_instance *instance)
    }
 }
 
+static enum radeon_ctx_pstate
+radv_parse_pstate(const char* str)
+{
+   if (!strcmp(str, "peak")) {
+      return RADEON_CTX_PSTATE_PEAK;
+   } else if (!strcmp(str, "standard")) {
+      return RADEON_CTX_PSTATE_STANDARD;
+   } else if (!strcmp(str, "min_sclk")) {
+      return RADEON_CTX_PSTATE_MIN_SCLK;
+   } else if (!strcmp(str, "min_mclk")) {
+      return RADEON_CTX_PSTATE_MIN_MCLK;
+   } else {
+      return RADEON_CTX_PSTATE_NONE;
+   }
+}
+
 VKAPI_ATTR VkResult VKAPI_CALL
 radv_CreateInstance(const VkInstanceCreateInfo *pCreateInfo, const VkAllocationCallbacks *pAllocator,
                     VkInstance *pInstance)
@@ -338,6 +354,7 @@ radv_CreateInstance(const VkInstanceCreateInfo *pCreateInfo, const VkAllocationC
 
    instance->debug_flags = parse_debug_string(getenv("RADV_DEBUG"), radv_debug_options);
    instance->perftest_flags = parse_debug_string(getenv("RADV_PERFTEST"), radv_perftest_options);
+   instance->profile_pstate = radv_parse_pstate(debug_get_option("RADV_PROFILE_PSTATE", "peak"));
 
    /* When RADV_FORCE_FAMILY is set, the driver creates a null
     * device that allows to test the compiler without having an

src/amd/vulkan/radv_instance.h

@@ -12,7 +12,7 @@
 #define RADV_INSTANCE_H
 
 #include "util/xmlconfig.h"
-
+#include "radv_radeon_winsys.h"
 #include "vk_instance.h"
 
 #ifdef ANDROID_STRICT
@@ -39,6 +39,7 @@ struct radv_instance {
 
    uint64_t debug_flags;
    uint64_t perftest_flags;
+   enum radeon_ctx_pstate profile_pstate;
 
    struct {
       struct driOptionCache options;