Alyssa Rosenzweig [Mon, 28 Aug 2023 16:32:47 +0000 (12:32 -0400)]
agx/nir_lower_texture: Allow disabling layer clamping
For background program with layered.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Alyssa Rosenzweig [Mon, 28 Aug 2023 13:40:46 +0000 (09:40 -0400)]
nir: Support arrays in block_image_store_agx
For layered rendering, runs once per layer.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Alyssa Rosenzweig [Wed, 16 Aug 2023 21:58:27 +0000 (17:58 -0400)]
nir: Add layer_id_written_agx sysval
We'll implement layer ID reads in the frag shader with a varying read, but if
the VS doesn't write the varying we need to return 0 per the spec. Add a sysval
to detect that case so we can handle it at runtime without keys.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Alyssa Rosenzweig [Wed, 30 Aug 2023 20:03:46 +0000 (16:03 -0400)]
agx: Insert jmp_exec_none instructions
With the exception of the backwards branch for loops, all the control flow we
insert during instruction selection just predicates instructions rather than
actually jumping around. That means, for example, we execute both sides of the
if even for a uniform condition! That's inefficient. The solution is insert
jmp_exec_none instructions after control flow in order to skip unexecuted
regions, which is much faster than predicating them out. However, jmp_exec_none
is costly in itself, so we need to use a heuristic to determine when it's
actually beneficial.
This uses a very simple heuristic for this purpose. However, it is a massive
performance speed-up for Dolphin uber shaders: 39fps -> 67fps at 2x resolution.
Nearly a doubling of performance!
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Alyssa Rosenzweig [Wed, 30 Aug 2023 20:47:31 +0000 (16:47 -0400)]
agx: Add agx_prev_block helper
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Alyssa Rosenzweig [Wed, 30 Aug 2023 20:25:02 +0000 (16:25 -0400)]
agx: Add jumps to block ends
jmp_exec_none variant that jumps to the last instruction of the target block,
rather than the beginning. This is convenient for skipping over elses, while
still executing the block-final pop_exec instruction. Similarly for skipping
over loop bodies while still executing the block-final pop_exec, after break
instructions.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Alyssa Rosenzweig [Wed, 30 Aug 2023 19:43:51 +0000 (15:43 -0400)]
agx: Augment if/else/while_cmp with a target
Add an optional pointer to a target block for these instructions. This does NOT
act like a logical branch, and does NOT get added to the logical control flow.
It is ignored wholesale until after RA, when physical edges may be inserted by a
pass we add later in this series.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Alyssa Rosenzweig [Sun, 3 Sep 2023 23:03:04 +0000 (19:03 -0400)]
agx: Set PIPE_SHADER_CAP_CONT_SUPPORTED
So we get adequate testing of continues, rather than lowering them in GLSL. We
don't really /want/ to see continues but lowering them away will just make them
harder to test... and besides, we should be optimizing them in NIR (not GLSL) so
we can get the win on Vulkan too.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Alyssa Rosenzweig [Wed, 30 Aug 2023 21:56:37 +0000 (17:56 -0400)]
agx: Hoist sample_mask/zs_emit
Although this is well-motivated, perf effect seems to be neglible for Dolphin.
It does prevent the scheduler from making things worse by sinking these
instructions though, so as a way to prevent future problems this seems sensible.
The kind of problem this affects (late discard) isn't modelled in shader-db.
Nevertheless, nothing concerning there:
total instructions in shared programs: 1756699 -> 1756722 (<.01%)
instructions in affected programs: 10106 -> 10129 (0.23%)
helped: 21
HURT: 41
Inconclusive result (value mean confidence interval includes 0).
total bytes in shared programs:
11525404 ->
11525452 (<.01%)
bytes in affected programs: 72900 -> 72948 (0.07%)
helped: 27
HURT: 41
Inconclusive result (value mean confidence interval includes 0).
total halfregs in shared programs: 483394 -> 483286 (-0.02%)
halfregs in affected programs: 4945 -> 4837 (-2.18%)
helped: 88
HURT: 78
Inconclusive result (value mean confidence interval includes 0).
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Alyssa Rosenzweig [Mon, 4 Sep 2023 14:03:57 +0000 (10:03 -0400)]
agx: Align the reg file for 256-bit vectors
This fixes live range splitting with 3D textureGrad(), which involves vectors
larger than the natural 128-bit maximum and hence requires special handling.
Fixes this assert with a combination of debug flags and new patches:
unsigned int find_best_region_to_evict(struct ra_ctx *, unsigned int,
unsigned int *, unsigned int *):
Assertion `(rctx->bound % size) == 0 && "register file size must be aligned
to the maximum vector size"' failed
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Alyssa Rosenzweig [Thu, 31 Aug 2023 00:29:43 +0000 (20:29 -0400)]
asahi: Clamp index buffer extent to what's read
This makes for cleaner agxdecodes, I think this matches what I've seen on the
macOS side but I might be misremembering. Certainly shouldn't hurt.
This only applies for direct draws.
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Friedrich Vock [Fri, 18 Aug 2023 21:01:26 +0000 (23:01 +0200)]
radv: Fix check in insert_block
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25496>
Friedrich Vock [Fri, 18 Aug 2023 20:59:09 +0000 (22:59 +0200)]
radv: Initialize shader freelist on allocation
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25496>
Vitaliy Triang3l Kuzmin [Sat, 5 Aug 2023 13:53:36 +0000 (16:53 +0300)]
r600: Replace R600_BIG_ENDIAN with UTIL_ARCH_BIG_ENDIAN
In particular, removes the dependency of r600_formats.h on r600_pipe.h so
it can be shared between Gallium and Vulkan.
Reviewed-by: Gert Wollny <gert.wollny@collabora.com>
Signed-off-by: Vitaliy Triang3l Kuzmin <triang3l@yandex.ru>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24513>
Marek Olšák [Mon, 7 Aug 2023 04:59:05 +0000 (00:59 -0400)]
winsys/amdgpu: pad gfx and compute IBs with a single NOP packet
to minimize CP overhead
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25043>
Marek Olšák [Mon, 7 Aug 2023 03:59:52 +0000 (23:59 -0400)]
ac/gpu_info: pad IBs according to ib_size_alignment
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25043>
Marek Olšák [Mon, 7 Aug 2023 03:17:48 +0000 (23:17 -0400)]
ac/gpu_info: replace ib_alignment with per-IP IB base and size alignments
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25043>
Eric Engestrom [Wed, 16 Aug 2023 21:16:17 +0000 (22:16 +0100)]
ci/deqp-runner: restore exit-on-error after getting deqp-runner's exit code
Signed-off-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24738>
Eric Engestrom [Wed, 16 Aug 2023 18:44:37 +0000 (19:44 +0100)]
ci/deqp-runner: fix indentation
Signed-off-by: Eric Engestrom <eric@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24738>
Marek Olšák [Mon, 4 Sep 2023 17:23:53 +0000 (13:23 -0400)]
amd: rename GFX110x to NAVI31-33
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Acked-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25492>
Marek Olšák [Sat, 2 Sep 2023 02:45:05 +0000 (22:45 -0400)]
ac/llvm: fix flat PS input corruption
Fixes:
0a54fbb5b40fc4e4a - radeonsi/gfx11: interp changes for 32bit
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25492>
Marek Olšák [Mon, 7 Aug 2023 05:04:04 +0000 (01:04 -0400)]
ac/gpu_info: don't align IBs to the GL2 cache line size
PAL doesn't do it. If drivers want IBs not to share cache lines with other buffers,
they should align the size manually.
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/25492>
Marek Olšák [Thu, 3 Aug 2023 00:24:04 +0000 (20:24 -0400)]
ac/llvm: set !fpmath 3.0 for llvm.sqrt
LLVM will be changed to increase precision with additional instructions,
which we don't want.
Reviewed-by: Timur Kristóf <timur.kristof@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25492>
Marek Olšák [Sat, 30 Sep 2023 07:36:58 +0000 (03:36 -0400)]
ac/perfcounter: remove a bogus assert to fix an assertion failure on gfx11
Reviewed-by: Konstantin Seurer <konstantin.seurer@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25485>
Eric Engestrom [Fri, 4 Aug 2023 17:22:55 +0000 (18:22 +0100)]
ci: move shader-db clone/build into its own script
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24504>
Eric Engestrom [Fri, 8 Sep 2023 17:34:52 +0000 (18:34 +0100)]
ci: drop unused shader-db clone + build from fedora image
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24504>
Eric Engestrom [Fri, 8 Sep 2023 17:34:52 +0000 (18:34 +0100)]
ci: drop unused shader-db clone + build from alpine image
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24504>
Konstantin Seurer [Wed, 23 Aug 2023 12:30:51 +0000 (14:30 +0200)]
radv: Do not sync after radv_update_buffer_cp
WR_CONFIRM is set.
Reviewed-by: Friedrich Vock <friedrich.vock@gmx.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24846>
Konstantin Seurer [Wed, 23 Aug 2023 12:17:18 +0000 (14:17 +0200)]
radv: Merge the sync_data and header initialization
Reviewed-by: Friedrich Vock <friedrich.vock@gmx.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24846>
Konstantin Seurer [Wed, 23 Aug 2023 10:13:27 +0000 (12:13 +0200)]
radv/bvh/ploc: Load child bounds from LDS
The bounds are already in LDS so there is no need to load them from
VRAM.
Reviewed-by: Friedrich Vock <friedrich.vock@gmx.de>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24846>
Eric Engestrom [Fri, 8 Sep 2023 09:28:59 +0000 (10:28 +0100)]
docs: document the merging process and what is allowed or not
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25117>
Cong Liu [Thu, 28 Sep 2023 10:36:06 +0000 (18:36 +0800)]
virgl:Fix ITEM_CPY macro pointer copy bug
The ITEM_CPY macro uses the memcpy function to copy the item variable.
When item is a pointer, the memcpy function will copy the value of the
pointer, not the address that the pointer points to.
Signed-off-by: Cong Liu <liucong2@kylinos.cn>
Reviewerd-by: Feng Jiang <jiangfeng@kylinos.cn>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25453>
Guilherme Gallo [Fri, 29 Sep 2023 06:35:10 +0000 (03:35 -0300)]
ci/anv: Catch some flakes
Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25469>
Guilherme Gallo [Fri, 29 Sep 2023 06:25:26 +0000 (03:25 -0300)]
ci/zink: Found some flakes
Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25469>
Guilherme Gallo [Fri, 29 Sep 2023 06:22:27 +0000 (03:22 -0300)]
ci/freedreno: Add a new flake
KHR-GLES31.core.compute_shader.simple-compute-shared_context resulted in
an UnexpectedPass during execution of the job
https://gitlab.freedesktop.org/mesa/mesa/-/jobs/
49445812
Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25469>
Guilherme Gallo [Fri, 29 Sep 2023 06:18:02 +0000 (03:18 -0300)]
ci/radeonsi: Update flake list
Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25469>
Mike Blumenkrantz [Fri, 29 Sep 2023 13:15:31 +0000 (09:15 -0400)]
radv: fix external handle type queries for dmabuf/fd
the intent of the VkExternalMemoryProperties API is that all compatible
handle types are returned, not just the type being queried. these two
types are compatible, so return both when both are supported
cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25474>
Rhys Perry [Fri, 29 Sep 2023 10:36:43 +0000 (11:36 +0100)]
aco: don't combine DPP into v_cmpx
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25471>
Rhys Perry [Fri, 29 Sep 2023 10:01:45 +0000 (11:01 +0100)]
aco/optimizer_postRA: don't combine DPP across exec on GFX8/9
GFX8/9 seem to use FI=0 behaviour.
fossil-db (vega10):
Totals from 1 (0.00% of 63053) affected shaders:
Instrs: 542 -> 570 (+5.17%)
CodeSize: 2928 -> 3040 (+3.83%)
Latency: 2087 -> 2118 (+1.49%)
InvThroughput: 1103 -> 1143 (+3.63%)
Affected shader is from Cyberpunk 2077 fossil.
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Cc: 23.2 <mesa-stable>
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/9784
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25471>
Helen Koike [Fri, 29 Sep 2023 12:40:32 +0000 (09:40 -0300)]
ci/ci_run_n_monitor: add docs for multiple targets
clarify how multiple targets can be used with the pipe character.
Signed-off-by: Helen Koike <helen.koike@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25472>
Lionel Landwerlin [Mon, 7 Aug 2023 11:03:57 +0000 (14:03 +0300)]
intel/fs: switch from SIMD 1 to 8 instructions surface/sampler rematerialization
SIMD1 instructions are problematic because they are considered partial
writes. This increases the liveness of the destination register
written by those instructions. To workaround this we use UNDEF
instructions to bound the liveness of the register. But this causing
other issues like in this case :
undef(1) vgrf2
mov(1) vgrf2, u4.0
add(1) vgrf3, vgrf2.0, 64UD
In this case the copy propagation pass in unable to see that vgrf2 in
the add() instruction can be replaced with the uniform u4.0.
To fix this problem, we switch NoMask SIMD8 instructions that cover
the entire register. We can drop the UNDEF instructions and now copy
propagation can do its job.
Good results on 2 apps :
Cyberpunk 2077 :
Totals from 7258 (68.80% of 10549) affected shaders:
Instrs: 6332210 -> 6073833 (-4.08%); split: -4.11%, +0.03%
Cycles:
130667501 ->
127351268 (-2.54%); split: -3.12%, +0.58%
Subgroup size: 90320 -> 90400 (+0.09%)
Spill count: 90 -> 68 (-24.44%)
Fill count: 82 -> 64 (-21.95%)
Scratch Memory Size: 8192 -> 6144 (-25.00%)
Max live registers: 385464 -> 375152 (-2.68%)
Max dispatch width: 64336 -> 64424 (+0.14%); split: +0.96%, -0.82%
Gaining 60 SIMD16/SIMD32 shaders, loosing 33
Strange Brigade :
Totals from 2137 (53.12% of 4023) affected shaders:
Instrs: 1544031 -> 1457544 (-5.60%); split: -5.60%, +0.00%
Cycles:
22292564 ->
21868978 (-1.90%); split: -2.43%, +0.53%
Subgroup size: 25328 -> 25344 (+0.06%)
Max live registers: 113716 -> 111214 (-2.20%)
Max dispatch width: 17232 -> 18608 (+7.99%); split: +8.36%, -0.37%
Gaining 138 SIMD16/SIMD32 shaders, loosing 4
On app slightly negatively affected :
Dota2 :
Totals from 232 (14.73% of 1575) affected shaders:
Instrs: 30029 -> 28194 (-6.11%)
Cycles: 385155 -> 371422 (-3.57%); split: -3.59%, +0.02%
Max live registers: 6792 -> 6780 (-0.18%)
Max dispatch width: 2256 -> 2160 (-4.26%)
Loosing 6 SIMD32 shaders
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24554>
Lionel Landwerlin [Mon, 7 Aug 2023 11:09:21 +0000 (14:09 +0300)]
intel/fs: handle add3 in surface/sampler rematerialization
Some recent NIR changes started generated those instructions. We need
to handle them to be able to rematerialize.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24554>
Lionel Landwerlin [Fri, 8 Sep 2023 06:52:14 +0000 (09:52 +0300)]
intel/fs: handle ishl in surface/sampler rematerialization
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24554>
Samuel Pitoiset [Wed, 27 Sep 2023 13:15:51 +0000 (15:15 +0200)]
radv: emit missing PA_{SC,SU}_LINE_STIPPLE_xxx regs in gfx preamble
Ported from RadeonSI and PAL.
This might fix the line stipple flakes in CI.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25428>
Samuel Pitoiset [Tue, 26 Sep 2023 16:19:59 +0000 (18:19 +0200)]
radv: fix COMPUTE_SHADER_INVOCATIONS query on compute queue
The VA needs to be adjusted, otherwise the hw always writes at offset 0.
This fixes dEQP-VK.query_pool.statistics_query.*_cq.
Cc: mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25406>
Samuel Pitoiset [Wed, 27 Sep 2023 15:20:01 +0000 (17:20 +0200)]
radv: re-enable DCC with mipmaps on GFX11
This is now fixed.
Hi-Fi Rush, Sonic Frontiers and Hogwarts Legacy were known broken games.
I personally reproduced the issue with Hi-Fi Rush which has been fixed
since
e6735409eef ("radv: disable DCC with signedness reinterpretation
on GFX11"). I also tested Sonic Frontiers which has been fixed since
52b68869922 ("amd: update addrlib"). I didn't check Hogwarts Legacy but
I think it was also fixed by
e6735409eef.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25435>
Caio Oliveira [Fri, 8 Sep 2023 18:57:48 +0000 (11:57 -0700)]
compiler/types: Add void parameter to ensure these are valid C prototypes
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25445>
Caio Oliveira [Fri, 8 Sep 2023 18:17:16 +0000 (11:17 -0700)]
compiler/types: Spell struct and enum in type names
This is a preparation for moving compiler/types from C++ to C.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25445>
Caio Oliveira [Fri, 8 Sep 2023 18:36:19 +0000 (11:36 -0700)]
compiler/types: Use C compatible cast syntax
This is a preparation for moving compiler/types from C++ to C.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25445>
Caio Oliveira [Wed, 27 Sep 2023 23:02:20 +0000 (16:02 -0700)]
compiler/types: Remove use of auto
This is a preparation for moving compiler/types from C++ to C.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25445>
Caio Oliveira [Wed, 27 Sep 2023 22:54:42 +0000 (15:54 -0700)]
compiler/types: Remove use of references
This is a preparation for moving compiler/types from C++ to C.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25445>
Caio Oliveira [Wed, 27 Sep 2023 22:49:26 +0000 (15:49 -0700)]
compiler/types: Remove use of new/delete
This is a preparation for moving compiler/types from C++ to C.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25445>
Caio Oliveira [Fri, 1 Sep 2023 21:07:41 +0000 (14:07 -0700)]
compiler/types: Remove private related declarations
Now there's no private data in glsl_type anymore, so we can remove
the friend declaration and move the private functions into file local
statics.
Reviewed-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25445>
Karol Herbst [Thu, 28 Sep 2023 21:02:20 +0000 (23:02 +0200)]
rusticl/mesa: create COMPUTE_ONLY contexts
This allows drivers to skip initializing unnecessary bits.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25466>
Sagar Ghuge [Fri, 8 Sep 2023 21:15:22 +0000 (14:15 -0700)]
anv: Enable barrier handling on video engines
v1: (Lionel)
- Don't check for the layout transition
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/9776
Signed-off-by: Sagar Ghuge <sagar.ghuge@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25131>
Marek Olšák [Tue, 26 Sep 2023 04:28:36 +0000 (00:28 -0400)]
ac/llvm: replace removed amdgcn.ldexp for LLVM 18
Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25393>
Emma Anholt [Wed, 27 Sep 2023 20:15:48 +0000 (13:15 -0700)]
ci/etnaviv: Drop some gc2k flakes that I think are resolved.
All these basic GLSL tests should be stable now that other processes
aren't hanging the GPU on us.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25442>
Emma Anholt [Wed, 27 Sep 2023 19:34:07 +0000 (12:34 -0700)]
ci/etnaviv: Skip some tests that hang the GPU and knock out other tests.
Just a few tests were causing GPU hangs that seem to cause spurious
failures in other tests. With those skipped, the results mostly
stabilize.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25442>
Marek Olšák [Tue, 26 Sep 2023 00:09:17 +0000 (20:09 -0400)]
radeonsi: fix compute-only contexts
We can't skip redundant register updates on compute queues, at least
not all of them.
Fixes:
315231b5a519bdc14 - radeonsi: eliminate redundant compute SH register changes
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25390>
Faith Ekstrand [Thu, 28 Sep 2023 15:13:31 +0000 (10:13 -0500)]
nvk: Init pipelineCacheUUID
Fixes:
968cefbff113 ("nvk: Cache NIR shaders")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25458>
Martin Roukala (né Peres) [Sun, 17 Sep 2023 10:55:34 +0000 (13:55 +0300)]
radv/ci: add a manual job for vkd3d-proton on navi31
Signed-off-by: Martin Roukala (né Peres) <martin.roukala@mupuf.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25266>
Martin Roukala (né Peres) [Sun, 17 Sep 2023 10:50:52 +0000 (13:50 +0300)]
radv/ci: add a manual job to run vkcts on navi31
We currently only have a single navi31, but we'll get more in the
next month or so.
Signed-off-by: Martin Roukala (né Peres) <martin.roukala@mupuf.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25266>
Martin Roukala (né Peres) [Wed, 27 Sep 2023 07:01:02 +0000 (10:01 +0300)]
radv/ci: update the vkcts gfx1100 flake/fail lists
Signed-off-by: Martin Roukala (né Peres) <martin.roukala@mupuf.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25266>
Timothy Arceri [Mon, 25 Sep 2023 08:18:44 +0000 (18:18 +1000)]
glsl: remove now unused varying linker code
All of this varying linking code has now been ported to the NIR linker.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25371>
Timothy Arceri [Mon, 25 Sep 2023 00:57:15 +0000 (10:57 +1000)]
glsl: switch to nir linkers cross_validate_outputs_to_inputs()
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25371>
Timothy Arceri [Mon, 25 Sep 2023 00:45:24 +0000 (10:45 +1000)]
glsl: implement cross_validate_outputs_to_inputs() in nir linker
This is mostly a direct port of the GLSL IR code there are only
2 real functional changes.
1. The direct use of mesa symbol_table instead of glsl_symbol_table.
However since none of the extra functionality offered by
glsl_symbol_table was ever used here this can be seen as an
improvement.
2. Because interface blocks are lowered before this new nir linker
sees them we must explicitly skip them (they are validated
elsewhere) to avoid errors.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25371>
Timothy Arceri [Mon, 25 Sep 2023 00:37:05 +0000 (10:37 +1000)]
nir: add used field to nir variables
Will be use in a following path by the glsl nir based linker.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25371>
Timothy Arceri [Mon, 25 Sep 2023 00:16:15 +0000 (10:16 +1000)]
glsl: move is_gl_identifier() to linker_util
This had started to multiply around the NIR linker. Here we move it to a
common helper shared by GLSL IR and NIR.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25371>
Timothy Arceri [Mon, 25 Sep 2023 00:04:24 +0000 (10:04 +1000)]
glsl: move interpolation_string() to linker_util
Moving it here allows us to use in in the NIR linker too.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25371>
Timothy Arceri [Fri, 22 Sep 2023 06:07:25 +0000 (16:07 +1000)]
nir: copy explicit_invariant flag to nir vars
This will be used in the following patch.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25371>
Samuel Pitoiset [Tue, 26 Sep 2023 14:37:28 +0000 (16:37 +0200)]
radv: stop skip emitting CB states when there is no color attachment
This is actually wrong. For example, if there is a DCC decompress
draw followed by a draw without any color attachments,
CB_COLOR_CONTROL.MODE is still CB_DCC_DECOMPRESS but it should be
CB_DISABLED. For some reasons, this hangs on RDNA3 (VM faults are also
reported through dmesg).
This fixes GPU hangs with Resident Evil 6, Star Wars The Old Republic
and probably more games on RDNA3.
Strictly speaking, I don't think this dynamic state optimization is
worth a try, even for other states, and I think it would be safer to
remove it completely.
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/9335
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/8327
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/9878
Fixes:
c08082e8615 ("radv: ignore all CB dynamic states when there is no color attachments")
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25402>
Mike Blumenkrantz [Wed, 27 Sep 2023 18:43:09 +0000 (14:43 -0400)]
glx: XFree visual info
cc: mesa-stable
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25441>
Cong Liu [Wed, 27 Sep 2023 07:30:14 +0000 (15:30 +0800)]
r300: Fix out-of-bounds access in ntr_emit_store_output()
This patch resolves the problem by modifying the for loop condition
to ensure that it stays within the bounds of the array (i.e., i < 4)
Signed-off-by: Cong Liu <liucong2@kylinos.cn>
Reviewed-by: Pavel Ondračka <pavel.ondracka@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25419>
Eric Engestrom [Wed, 27 Sep 2023 13:14:14 +0000 (14:14 +0100)]
ci/zink+radv: document flake
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25427>
Christian Gmeiner [Thu, 31 Aug 2023 13:38:04 +0000 (15:38 +0200)]
isaspec: python does not need ';'
Signed-off-by: Christian Gmeiner <cgmeiner@igalia.com>
Reviewed-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25437>
Ian Romanick [Wed, 26 Jul 2023 21:42:59 +0000 (14:42 -0700)]
spirv: Track when a shader has a cooperative matrix
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23825>
Caio Oliveira [Sat, 17 Jun 2023 00:02:39 +0000 (17:02 -0700)]
spirv: Implement SPV_KHR_cooperative_matrix
Includes a modified version of using extract/insert for OpLoad/OpStore
from Ian.
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com> (earlier version)
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl> (earlier version)
Acked-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23825>
Caio Oliveira [Sat, 23 Sep 2023 08:19:26 +0000 (01:19 -0700)]
spirv: Let vtn_ssa_value hold references to variables
In certain cases, we have complex opaque objects that are loaded
into (SPIR-V) SSA values. To represent these, we now can store a
reference to a variable in vtn_ssa_value.
Also implements a few operations we know will have to be supported,
like Select and Copy.
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23825>
Caio Oliveira [Sat, 17 Jun 2023 00:22:04 +0000 (17:22 -0700)]
spirv: Expose some memory related functions in vtn_private.h
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23825>
Caio Oliveira [Tue, 8 Aug 2023 18:02:33 +0000 (11:02 -0700)]
nir: Handle cooperative matrix in various passes
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23825>
Caio Oliveira [Tue, 8 Aug 2023 18:02:14 +0000 (11:02 -0700)]
nir: Add new intrinsics for Cooperative Matrix
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23825>
Caio Oliveira [Wed, 31 May 2023 06:26:14 +0000 (23:26 -0700)]
compiler/types: Add support for Cooperative Matrix types
Reviewed-by: Jesse Natalie <jenatali@microsoft.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23825>
Paulo Zanoni [Mon, 14 Aug 2023 16:41:11 +0000 (09:41 -0700)]
anv: enable sparse resources by default
This of course only applies to xe.ko. There is no reason to keep it
disabled by default.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23045>
Paulo Zanoni [Thu, 20 Jul 2023 21:30:16 +0000 (14:30 -0700)]
anv/sparse: add INTEL_DEBUG=sparse
This pollutes stderr a lot, but I've used it countless times while
developing this code.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23045>
Paulo Zanoni [Mon, 7 Aug 2023 22:07:37 +0000 (15:07 -0700)]
anv/sparse: get ready to issue a single vm_bind ioctl per non-opaque bind
Game testing shows it's common for this operation to result in
multiple bind regions, so try to use a single ioctl when we can.
Actual testing reveals 136 shader-related tests fail when we actually
do this, so for now keep doing a single bind per ioctl while leaving a
very easy way to the desired behavior when we figure this out.
It should also be possible to go even higher-level and do this at the
anv_queue_submit_sparse_bind_locked() layer, but that should happen in
future commits.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23045>
Paulo Zanoni [Wed, 19 Apr 2023 00:26:05 +0000 (17:26 -0700)]
anv/sparse: add the initial code for Sparse Resources
This giant patch implements a huge chunk of the Vulkan Sparse
Resources API. I previously had this as a nice series of many smaller
patches that evolved as the xe.ko added more features, but once I was
asked to squash some of the major reworks I realized I wouldn't be
able easily rewrite history, so I just squased basically the whole
series into a giant patch. I may end up splitting this again later if
I find a way to properly do it.
If we want to support the DX12 API through vkd3d we need to support
part of the the Sparse Resources API. If we don't, a bunch of Steam
games won't work.
For now we only support the xe.ko backend, but the vast majority of
the code is KMD-independent and so an i915.ko implementation would use
most of what's here, just extending the part that binds and unbinds
memory.
v2+: There's no way to sanely track the version history of this patch
in this commit message. Please refer to Gitlab.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23045>
Paulo Zanoni [Mon, 25 Sep 2023 20:45:06 +0000 (13:45 -0700)]
intel/isl: simplify the check for maximum surface size
The only thing that changes between these 3 checks is the size.
This entire patch was suggested by Kenneth Graunke, I just converted
his gitlab comment to a git commit.
Credits-to: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23045>
Paulo Zanoni [Fri, 21 Apr 2023 23:33:00 +0000 (16:33 -0700)]
intel/isl: add ISL_SURF_USAGE_SPARSE_BIT
Vulkan Sparse resources have their own set of rules, so here we try to
make ISL aware of them through ISL_SURF_USAGE_SPARSE_BIT.
The big deal here is when some image ends up not using Tile64 nor
TileYs. Previously Ys was not supported on TGL at all, and Tile64 did
not have support for 3D. Now we still have some formats that end up
not being used with either Tile64 and Ys, but need to support Sparse
on them (e.g., YUV on Tile64). In the future we may have new tiling
formats or hardware restrictions that would force this case to happen
again.
So here we do some adjustments so we can make sparse work with other
tiling formats, although with limited functionality (e.g., those
formats may be restricted to opaque binds, and certainly don't support
the standard block shapes).
v2: before we had Ys support, we had defined TGL's block size as 4k.
v3: move the size_B chunk to before nte notify_failure() checks (Ken).
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/23045>
Faith Ekstrand [Wed, 27 Sep 2023 20:10:22 +0000 (15:10 -0500)]
nvk: Cache NIR shaders
We can't cache shader binaries just yet but this at least lets us cache
the output of spirv_to_nir and the initial optimize.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25443>
Faith Ekstrand [Wed, 27 Sep 2023 20:09:53 +0000 (15:09 -0500)]
nvk: Add a default pipeline cache
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25443>
Faith Ekstrand [Wed, 27 Sep 2023 19:48:29 +0000 (14:48 -0500)]
nvk: Re-structure early shader compilation a bit
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25443>
Faith Ekstrand [Wed, 27 Sep 2023 19:28:05 +0000 (14:28 -0500)]
nvk: Hook up the disk cache
This won't actually do much yet because we don't have pipeline caches
yet but it turns on the infrastructure.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25443>
Faith Ekstrand [Wed, 27 Sep 2023 17:45:39 +0000 (12:45 -0500)]
nvk: Store a 20-bit driver_build_sha in nvk_instance
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25443>
Timothy Arceri [Tue, 26 Sep 2023 04:47:50 +0000 (14:47 +1000)]
nir: fix typo in comment
The variable is unused or dead, not used.
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com>
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25414>
antonino [Thu, 17 Aug 2023 13:09:15 +0000 (15:09 +0200)]
vulkan: Handle vkSetDebugUtilsObjectNameEXT on WSI objects
Some WSI objects don't extend `vk_object_base` therefore they need
special handling.
Fixes:
3c87618d357 ("vulkan: Handle vkGet/SetPrivateDataEXT on Android swapchains")
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24752>
antonino [Thu, 17 Aug 2023 13:07:04 +0000 (15:07 +0200)]
vulkan: Extend vkGet/SetPrivateDataEXT handling to VkSurface
VkSurface is handled by WSI and it doesn't extend `vk_object_base` so it
needs special handling.
Fixes:
3c87618d357 ("vulkan: Handle vkGet/SetPrivateDataEXT on Android swapchains")
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24752>
antonino [Thu, 17 Aug 2023 13:06:18 +0000 (15:06 +0200)]
vulkan: Extend vkGet/SetPrivateDataEXT handling to all platforms
Non-android platforms use mesa WSI, however some WSI object still don't
extend `vk_object_base` so they still need special handling.
Fixes:
3c87618d357 ("vulkan: Handle vkGet/SetPrivateDataEXT on Android swapchains")
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24752>
Dave Airlie [Wed, 27 Sep 2023 05:30:05 +0000 (15:30 +1000)]
llvmpipe/cs: migrate cs image handle to common jit code.
This moves some of the code over, and uses the generic paths.
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25416>
Dave Airlie [Wed, 27 Sep 2023 05:26:47 +0000 (15:26 +1000)]
llvmpipe/cs: migrate to generic jit texture from pipe code.
This moves some cs specific bits to the generic code and uses it.
Reviewed-by: Emma Anholt <emma@anholt.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25416>
Marcin Ślusarz [Mon, 20 Mar 2023 10:32:07 +0000 (11:32 +0100)]
intel/compiler: mask GS URB handles at thread payload construction
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25195>