Mesa (staging/21.1): ir3: memory_barrier also controls shared memory access order
GitLab Mirror
gitlab-mirror at kemper.freedesktop.org
Mon May 10 21:01:49 UTC 2021
Module: Mesa
Branch: staging/21.1
Commit: 1ea435427de3b165f6e69774ab665f757ecef36b
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=1ea435427de3b165f6e69774ab665f757ecef36b
Author: Danylo Piliaiev <dpiliaiev at igalia.com>
Date: Mon Feb 15 15:53:46 2021 +0200
ir3: memory_barrier also controls shared memory access order
nir_intrinsic_memory_barrier has the same semantic as memoryBarrier()
in GLSL, which is:
GLSL 4.60, 4.10. "Memory Qualifiers":
"The built-in function memoryBarrier() can be used if needed to
guarantee the completion and relative ordering of memory accesses
performed by a single shader invocation."
GLSL 4.60, 8.17. "Shader Memory Control Functions":
"The built-in functions memoryBarrier() and groupMemoryBarrier() wait
for the completion of accesses to all of the above variable types."
Fixes tests:
dEQP-VK.memory_model.message_passing.core11.u32.coherent.fence_fence.atomicwrite.device.payload_local.image.guard_nonlocal.workgroup.comp
dEQP-VK.memory_model.message_passing.core11.u32.coherent.fence_fence.atomicwrite.device.payload_nonlocal.workgroup.guard_local.image.comp
Fixes: 819a613a ("freedreno/ir3: moar better scheduler")
Signed-off-by: Danylo Piliaiev <dpiliaiev at igalia.com>
Reviewed-by: Samuel Iglesias Gonsálvez <siglesias at igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/9054>
(cherry picked from commit cb8a00791cc9618f5be903c31abf737b42e4cf46)
---
.pick_status.json | 2 +-
src/freedreno/ci/deqp-freedreno-a630-fails.txt | 2 --
src/freedreno/ir3/ir3.h | 2 +-
src/freedreno/ir3/ir3_compiler_nir.c | 13 +------------
4 files changed, 3 insertions(+), 16 deletions(-)
diff --git a/.pick_status.json b/.pick_status.json
index 1a4a95ffd11..e534829c841 100644
--- a/.pick_status.json
+++ b/.pick_status.json
@@ -211,7 +211,7 @@
"description": "ir3: memory_barrier also controls shared memory access order",
"nominated": true,
"nomination_type": 1,
- "resolution": 0,
+ "resolution": 1,
"master_sha": null,
"because_sha": "819a613ae33410584e13e78e78af82c71716f67d"
},
diff --git a/src/freedreno/ci/deqp-freedreno-a630-fails.txt b/src/freedreno/ci/deqp-freedreno-a630-fails.txt
index f3639f35712..fb130509949 100644
--- a/src/freedreno/ci/deqp-freedreno-a630-fails.txt
+++ b/src/freedreno/ci/deqp-freedreno-a630-fails.txt
@@ -129,8 +129,6 @@ dEQP-VK.image.subresource_layout.3d.all_levels.r16g16b16a16_snorm,Fail
dEQP-VK.image.subresource_layout.3d.all_levels.r8_snorm,Fail
dEQP-VK.image.subresource_layout.3d.all_levels.r8g8b8a8_snorm,Fail
dEQP-VK.info.device_mandatory_features,Fail
-dEQP-VK.memory_model.message_passing.core11.u32.coherent.fence_fence.atomicwrite.device.payload_local.image.guard_nonlocal.workgroup.comp,Fail
-dEQP-VK.memory_model.message_passing.core11.u32.coherent.fence_fence.atomicwrite.device.payload_nonlocal.workgroup.guard_local.image.comp,Fail
dEQP-VK.pipeline.extended_dynamic_state.after_pipelines.depth_compare_always_greater,Fail
dEQP-VK.pipeline.extended_dynamic_state.after_pipelines.depth_compare_greater,Fail
dEQP-VK.pipeline.extended_dynamic_state.after_pipelines.depth_compare_greater_equal_greater,Fail
diff --git a/src/freedreno/ir3/ir3.h b/src/freedreno/ir3/ir3.h
index bc11f89b94b..eaabef6c5bd 100644
--- a/src/freedreno/ir3/ir3.h
+++ b/src/freedreno/ir3/ir3.h
@@ -392,7 +392,7 @@ struct ir3_instruction {
* shared image atomic SSBO everything
* barrier()/ - R/W R/W R/W R/W X
* groupMemoryBarrier()
- * memoryBarrier() - R/W R/W
+ * memoryBarrier()
* (but only images declared coherent?)
* memoryBarrierAtomic() - R/W
* memoryBarrierBuffer() - R/W
diff --git a/src/freedreno/ir3/ir3_compiler_nir.c b/src/freedreno/ir3/ir3_compiler_nir.c
index 4db7e87cb14..14754036a24 100644
--- a/src/freedreno/ir3/ir3_compiler_nir.c
+++ b/src/freedreno/ir3/ir3_compiler_nir.c
@@ -1312,18 +1312,6 @@ emit_intrinsic_barrier(struct ir3_context *ctx, nir_intrinsic_instr *intr)
barrier->flags = IR3_INSTR_SS | IR3_INSTR_SY;
barrier->barrier_class = IR3_BARRIER_EVERYTHING;
break;
- case nir_intrinsic_memory_barrier:
- barrier = ir3_FENCE(b);
- barrier->cat7.g = true;
- barrier->cat7.r = true;
- barrier->cat7.w = true;
- barrier->cat7.l = true;
- barrier->barrier_class = IR3_BARRIER_IMAGE_W |
- IR3_BARRIER_BUFFER_W;
- barrier->barrier_conflict =
- IR3_BARRIER_IMAGE_R | IR3_BARRIER_IMAGE_W |
- IR3_BARRIER_BUFFER_R | IR3_BARRIER_BUFFER_W;
- break;
case nir_intrinsic_memory_barrier_buffer:
barrier = ir3_FENCE(b);
barrier->cat7.g = true;
@@ -1353,6 +1341,7 @@ emit_intrinsic_barrier(struct ir3_context *ctx, nir_intrinsic_instr *intr)
barrier->barrier_conflict = IR3_BARRIER_SHARED_R |
IR3_BARRIER_SHARED_W;
break;
+ case nir_intrinsic_memory_barrier:
case nir_intrinsic_group_memory_barrier:
barrier = ir3_FENCE(b);
barrier->cat7.g = true;
More information about the mesa-commit
mailing list