OpenHarmony/third_party_mesa3d
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

radeonsi: lock a mutex when updating scratch_va for compute shaders

Browse Source 
Fixes: 3b0bfd254f - radeonsi/gfx11: make flat_scratch changes for compute

Reviewed-by: Qiang Yu <yuq825@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/30071>
This commit is contained in:
Marek Olšák
2024-07-08 08:39:04 -04:00
parent a5b4ae67ae
commit 2afd233145
1 changed files with 18 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

18
src/gallium/drivers/radeonsi/si_compute.c
View File

@@ -218,6 +218,7 @@ static void *si_create_compute_state(struct pipe_context *ctx, const struct pipe
pipe_reference_init(&sel->base.reference, 1); pipe_reference_init(&sel->base.reference, 1);
sel->stage = MESA_SHADER_COMPUTE; sel->stage = MESA_SHADER_COMPUTE;
sel->screen = sscreen; sel->screen = sscreen;
simple_mtx_init(&sel->mutex, mtx_plain);
sel->const_and_shader_buf_descriptors_index = sel->const_and_shader_buf_descriptors_index =
si_const_and_shader_buffer_descriptors_idx(PIPE_SHADER_COMPUTE); si_const_and_shader_buffer_descriptors_idx(PIPE_SHADER_COMPUTE);
sel->sampler_and_images_descriptors_index = sel->sampler_and_images_descriptors_index =
@@ -480,6 +481,16 @@ static bool si_switch_compute_shader(struct si_context *sctx, struct si_compute
} }
if (config->scratch_bytes_per_wave) { if (config->scratch_bytes_per_wave) {
/* Prevent race conditions for accesses to shader->scratch_va and shader->bo, which
* can change when scratch_va is updated. Any accesses to shader->bo must also be inside
* the lock.
*
* TODO: This lock could be removed if the scratch address was passed via user SGPRs instead
* of the shader binary.
*/
if (!sctx->screen->info.has_scratch_base_registers)
simple_mtx_lock(&shader->selector->mutex);
/* Update max_seen_compute_scratch_bytes_per_wave and compute_tmpring_size. */ /* Update max_seen_compute_scratch_bytes_per_wave and compute_tmpring_size. */
ac_get_scratch_tmpring_size(&sctx->screen->info, ac_get_scratch_tmpring_size(&sctx->screen->info,
config->scratch_bytes_per_wave, config->scratch_bytes_per_wave,
@@ -503,6 +514,12 @@ static bool si_switch_compute_shader(struct si_context *sctx, struct si_compute
radeon_add_to_buffer_list(sctx, &sctx->gfx_cs, shader->bo, radeon_add_to_buffer_list(sctx, &sctx->gfx_cs, shader->bo,
RADEON_USAGE_READ | RADEON_PRIO_SHADER_BINARY); RADEON_USAGE_READ | RADEON_PRIO_SHADER_BINARY);
/* shader->bo can't be used after this if the scratch address is inserted into the shader
* binary.
*/
if (config->scratch_bytes_per_wave && !sctx->screen->info.has_scratch_base_registers)
simple_mtx_unlock(&shader->selector->mutex);
if (sctx->gfx_level >= GFX12) { if (sctx->gfx_level >= GFX12) {
unsigned rsrc3 = S_00B8A0_INST_PREF_SIZE_GFX12(si_get_shader_prefetch_size(shader)); unsigned rsrc3 = S_00B8A0_INST_PREF_SIZE_GFX12(si_get_shader_prefetch_size(shader));
@@ -1307,6 +1324,7 @@ void si_destroy_compute(struct si_compute *program)
si_shader_destroy(&program->shader); si_shader_destroy(&program->shader);
ralloc_free(program->sel.nir); ralloc_free(program->sel.nir);
simple_mtx_destroy(&sel->mutex);
FREE(program); FREE(program);
} }