fdo-mirrors/mesa
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/mesa/mesa.git synced 2026-05-22 02:18:10 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions 2
217010 commits 66 branches 936 tags 1 GiB
Commit graph

19620 commits

Author SHA1 Message Date
Konstantin Seurer
58a35647e1 radv: Fix crash if proceed comes before initialize
Some checks are pending
macOS-CI / macOS-CI (dri) (push) Waiting to run
Details
macOS-CI / macOS-CI (xlib) (push) Waiting to run
Details 
"initialize" can be NULL if the rq_proceed was visited before
rq_initialize.

Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/14626
cc: mesa-stable

Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39251>
 2026-01-12 22:34:32 +00:00
Natalie Vock
473cf6046a aco/spill_preserved: Preserve linear VGPRs even if they aren't p_spill operands 
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39157>
 2026-01-12 21:46:50 +00:00
Natalie Vock
1ef2691221 aco/spill: Fix preserved reload operand update 
p_logical_end is actually after p_reload_preserved, so this didn't do
anything.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39157>
 2026-01-12 21:46:50 +00:00
Natalie Vock
548062f10e aco/insert_waitcnt: Don't determine linearity by reg number 
VGPRs can be linear too, and RT function calls will add VMEM
instructions acting on linear VGPRs. Using the linear VGPR in a block
with only linear preds will cause the pass to incorrectly skip inserting
s_waitcnt.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39157>
 2026-01-12 21:46:50 +00:00
Natalie Vock
7c12603933 aco/lower_to_hw_instr: Preserve linearity of lowered linear VGPRs 
So subsequent passes like waitcnt insertion can know these are linear.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39157>
 2026-01-12 21:46:50 +00:00
Natalie Vock
0d93e8ce54 aco: Don't insert p_reload_preserved in loops 
This can't work.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39157>
 2026-01-12 21:46:50 +00:00
Natalie Vock
c816f699b2 aco/spill_preserved: Only reload linear VGPRs at end 
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39157>
 2026-01-12 21:46:50 +00:00
Natalie Vock
897c95c37e aco: Include arbitrarily fixed registers in max_reg_demand 
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39157>
 2026-01-12 21:46:50 +00:00
Georg Lehmann
daf235c607 aco/tests: don't destroy vk_device if it was never created 
Happens if you only run one test that doesn't need a vk_device.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39268>
 2026-01-12 16:16:54 +00:00
Georg Lehmann
fad95030a7 aco/tests: test VALUMaskWriteHazard with v_cmpx 
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39252>
 2026-01-12 15:48:39 +00:00
Georg Lehmann
1d85552745 aco/tests: test VALUReadSGPRHazard with v_cmpx 
To avoid regressing this in a future rework.

Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39252>
 2026-01-12 15:48:39 +00:00
Georg Lehmann
3e10ab34e1 aco/insert_NOPs: explicitly wait for sa_sdst to resolve SALU -> VALU hazards 
The assumption that these waits are not required has been proven incorrect
in at least some cases.

Totals from 190 (0.24% of 79825) affected shaders: (Navi31)
Instrs: 499718 -> 500491 (+0.15%)
CodeSize: 2658228 -> 2661916 (+0.14%)
Latency: 5964632 -> 5965453 (+0.01%); split: -0.00%, +0.01%
InvThroughput: 794221 -> 794289 (+0.01%)

Totals from 17093 (21.41% of 79839) affected shaders: (Navi48)
Instrs: 22805214 -> 22854313 (+0.22%)
CodeSize: 121240428 -> 121432904 (+0.16%); split: -0.00%, +0.16%
Latency: 166500300 -> 166530529 (+0.02%); split: -0.00%, +0.02%
InvThroughput: 28770053 -> 28772870 (+0.01%); split: -0.00%, +0.01%

Fixes: 018f45f981 ("aco/insert_NOPs: remove redundant VALUReadSGPRHazard waits")
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/14516

Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Reviewed-by: Daniel Schürmann <daniel@schuermann.dev>
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39252>
 2026-01-12 15:48:38 +00:00
Samuel Pitoiset
b65cc9d587 ac,radv: sample and set correct shader/memory clocks for RGP 
These clocks need to be the clocks at trace time. This shouldn't fix
anything given that RADV sets profile_peak when SQTT is enabled but
better to report it correctly anyways.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39208>
 2026-01-12 11:58:43 +00:00
David Rosca
0518784b62 radv/amdgpu: Only wait on queue syncobj when needed 
This would always wait on the queue syncobj if there was any other
wait syncobj, but it should only wait after zero submit.

Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39193>
 2026-01-12 10:59:03 +00:00
Dave Airlie
ab9e904f24 radv/coopmat: fix deref stride
Some checks are pending
macOS-CI / macOS-CI (dri) (push) Waiting to run
Details
macOS-CI / macOS-CI (xlib) (push) Waiting to run
Details 
This at least fixes the nir debug output to have correct values.

Fixes: 48fc8c8d1c ("radv/nir/lower_cmat: set optimal load/store alignment")
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39256>
 2026-01-12 10:39:05 +00:00
David Rosca
df4220d500 radv/video: Use different dpb swizzle mode for 10 bit encode 
Acked-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Benjamin Cheng <benjamin.cheng@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39189>
 2026-01-12 10:18:18 +00:00
David Rosca
587a7aa510 radv: Enable DCC modifiers for multi plane formats on GFX12 
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39190>
 2026-01-12 09:57:56 +00:00
Samuel Pitoiset
5bcca4a832 radv/spm: use a staging buffer for faster reads on dGPUS 
This allows us to move the SPM buffer to VRAM because I think it must
be in VRAM too.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39195>
 2026-01-12 09:35:37 +00:00
Samuel Pitoiset
6863a90486 radv/spm: rework allocating the SPM buffer 
For using a staging buffer.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39195>
 2026-01-12 09:35:37 +00:00
Samuel Pitoiset
c7d0aa6671 radv/sqtt: use a staging buffer for faster reads on dGPUS 
This is way faster.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39195>
 2026-01-12 09:35:36 +00:00
Samuel Pitoiset
5d430940d2 radv/sqtt: rework allocating the SQTT buffer 
For using a staging buffer.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39195>
 2026-01-12 09:35:36 +00:00
Samuel Pitoiset
1c611c2dac radv/sqtt: use VkCommandBuffer objects for SQTT start/stop sequences 
For using a staging buffer.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39195>
 2026-01-12 09:35:35 +00:00
Samuel Pitoiset
6f4b3c7c9b ac/perfcounter: re-order GPU perf blocks on GFX11 
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39199>
 2026-01-12 08:10:33 +00:00
Samuel Pitoiset
653b39989a ac/perfcounter: define more GPU blocks on GFX11 
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39199>
 2026-01-12 08:10:33 +00:00
Samuel Pitoiset
d391cd0c4d ac/perfcounter: fix number of scoped instances for RMI block 
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39199>
 2026-01-12 08:10:33 +00:00
Samuel Pitoiset
ea63aa3e8e ac/perfcounter: add missing configuration for GCEA on GFX11 
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39199>
 2026-01-12 08:10:33 +00:00
Samuel Pitoiset
6722a6332a ac,radv,radeonsi: rename num_spm_counters to num_spm_modules 
A module can have different number of counters.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39199>
 2026-01-12 08:10:32 +00:00
Samuel Pitoiset
fb43d7bff2 ac/perfcounter: re-order GPU perf blocks on GFX12 
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39199>
 2026-01-12 08:10:31 +00:00
Samuel Pitoiset
3b6ff80d48 ac/perfcounter: define more GPU blocks on GFX12 
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39199>
 2026-01-12 08:10:31 +00:00
Samuel Pitoiset
eb37d6ceb7 ac/perfcounter: fix computing number of 16-bit/32-bit SPM counters 
Determine them only when both are explicitly 0.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39199>
 2026-01-12 08:10:30 +00:00
Samuel Pitoiset
d1efdc7e76 ac/perfcounter: fix number of 32-bit SPM counters 
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39199>
 2026-01-12 08:10:29 +00:00
Samuel Pitoiset
9fe57d3882 ac/spm: define new per-shader engine blocks 
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39199>
 2026-01-12 08:10:29 +00:00
Samuel Pitoiset
60fac38491 ac/spm: fix typo in one GPU perf block name 
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39199>
 2026-01-12 08:10:29 +00:00
Samuel Pitoiset
db02077c8a radv: remove extra instructions after UNREACHABLE 
Minor cleanups.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39237>
 2026-01-12 07:41:08 +00:00
Samuel Pitoiset
e1e2517664 radv: use UNREACHABLE for illegal texture filter 
Found this with a broken CTS test, way easier to crash for isolating
the test case.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39237>
 2026-01-12 07:41:08 +00:00
Samuel Pitoiset
91e0f8f1e5 radv/rt: fix a compilation warning about uninitialized fields 
Just zero-initialize the layout struct to fix the following warning
because radv_use_bvh8() might return FALSE.

../src/amd/vulkan/radv_acceleration_structure.c: In function ‘radv_update_as_gfx12’:
../src/amd/vulkan/radv_acceleration_structure.c:873:70: warning: ‘layout.bounds_offsets’ may be used uninitialized [-Wmaybe-uninitialized]
  873 |       .bounds = state->build_info->scratchData.deviceAddress + layout.bounds_offsets,
      |                                                                ~~~~~~^~~~~~~~~~~~~~~
../src/amd/vulkan/radv_acceleration_structure.c:866:33: note: ‘layout.bounds_offsets’ was declared here
  866 |    struct update_scratch_layout layout;
      |                                 ^~~~~~

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39228>
 2026-01-12 07:18:50 +00:00
Konstantin Seurer
077292f65b radv/bvh: Use box16 nodes when bvh8 is not used
Some checks are pending
macOS-CI / macOS-CI (dri) (push) Waiting to run
Details
macOS-CI / macOS-CI (xlib) (push) Waiting to run
Details 
Using box16 nodes trades bvh quality for memory bandwidth which seems to
be roughly equal in performance.

Stats assuming box16 nodes are as expensive as box32 nodes:
Totals from 7668 (79.68% of 9624) affected BVHs:
compacted_size: 951666944 -> 742347648 (-22.00%)
max_depth: 57606 -> 57615 (+0.02%)
sah: 129114796242 -> 129998517775 (+0.68%); split: -0.00%, +0.68%
scene_sah: 188564162 -> 192063633 (+1.86%); split: -0.02%, +1.88%
box16_node_count: 0 -> 3270600 (+inf%)
box32_node_count: 3365707 -> 95100 (-97.17%)

Reviewed-by: Natalie Vock <natalie.vock@gmx.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37883>
 2026-01-10 11:36:28 +01:00
Konstantin Seurer
543a88af99 radv/bvh: Add radv_aabb16 and use it for box16 nodes 
Reviewed-by: Natalie Vock <natalie.vock@gmx.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37883>
 2026-01-10 11:36:19 +01:00
Konstantin Seurer
fefdad9249 radv/rra: Count box16 nodes properly 
Otherwise rra won't allocate memory when loading the capture.

Reviewed-by: Natalie Vock <natalie.vock@gmx.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37883>
 2026-01-10 11:34:18 +01:00
Konstantin Seurer
39d58a55a7 aco: Add support to f2f16 with rtpi/rtni 
Those rounding modes are needed when computing 16-bit bounding boxes
since the bounding box must not get smaller.

Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/37883>
 2026-01-10 11:34:12 +01:00
Alyssa Rosenzweig
235e868ef7 ac/nir: use nir_is_shared_access 
Signed-off-by: Alyssa Rosenzweig <alyssa.rosenzweig@intel.com>
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39219>
 2026-01-09 20:51:13 +00:00
Benjamin Cheng
499d9e2e98 radv/video: Allow aliasing of video images 
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39109>
 2026-01-09 13:52:56 +00:00
Georg Lehmann
6d07a56c6a ac/nir/lower_ps_late: preserve signed zero, inf, nan for exports 
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39187>
 2026-01-09 11:58:52 +00:00
Georg Lehmann
84ecac58a6 ac/nir/opt_pack_half: preserve fp_math_ctrl 
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39187>
 2026-01-09 11:58:52 +00:00
Georg Lehmann
5241343ccb ac/nir/lower_sin_cos: preserve fp_math_ctrl 
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39187>
 2026-01-09 11:58:52 +00:00
Georg Lehmann
9331726157 ac/nir/lower_sin_cos: use nir_shader_alu_pass 
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39187>
 2026-01-09 11:58:52 +00:00
Samuel Pitoiset
4fa20bacac radv/ci: document a regression with transfer queue on RENOIR
Some checks are pending
macOS-CI / macOS-CI (dri) (push) Waiting to run
Details
macOS-CI / macOS-CI (xlib) (push) Waiting to run
Details 
Weird that only RENOIR fails given that ASTC/ETC2 aren't natively
supported too.

Needs to be investigated but SDMA supports these formats to some
extent it seems.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39230>
 2026-01-09 10:47:31 +00:00
Samuel Pitoiset
edb730f647 radv: fix flushing gang semaphore with SDMA/ACE
Some checks are pending
macOS-CI / macOS-CI (dri) (push) Waiting to run
Details
macOS-CI / macOS-CI (xlib) (push) Waiting to run
Details 
If the main CS is SDMA and the gang CS is ACE, this would emit a
SDMA_FENCE packet on ACE which just hangs.

Fixes: b1938901d0 ("radv: Use SDMA fence packet when flushing gang semaphores")
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39211>
 2026-01-09 09:07:45 +00:00
Natalie Vock
60dd9d797e aco: Swizzle ray launch IDs in the RT prolog 
This converts from 1D workgroups to 2D ray launch IDs entirely via
shader ALU, including handling partial/cut-off workgroups optimally.

Doing this entirely in-shader means it Just Works(TM) with indirect
dispatches as well. Previous approaches manipulating various things on
CPU depending on the dispatch size couldn't handle indirect dispatches.

The swizzle implemented here also swizzles with a recursive Z-order
pattern, which should be a little more optimal than arranging
invocations linearly within the wave.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39142>
 2026-01-08 19:49:55 +01:00
Natalie Vock
1f6ac3fa93 radv/rt,aco: Always dispatch 1D workgroups for RT 
We will swizzle the workgroups ourselves in the next commit.
Removes the need for 1D dispatch workarounds.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/39142>
 2026-01-08 19:49:54 +01:00