193897 commits

Author	SHA1	Message	Date
Christian Gmeiner	faa3d901ab	etnaviv: Only emit VIVS_PS_MSAA_CONFIG if GPU support it ... Context buffer initialization in the nxp-imx kernel driver (lf-6.6.3-1.0.0) indicates that this state is only defined when the GPU supports MSAA_FRAGMENT_OPERATION. Signed-off-by: Christian Gmeiner <cgmeiner@igalia.com> Reviewed-by: Lucas Stach <l.stach@pengutronix.de> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36358>	2025-07-31 16:31:37 +00:00
Christian Gmeiner	71ac1647e7	etnaviv: hwdb: Add MSAA_FRAGMENT_OPERATION feature ... Signed-off-by: Christian Gmeiner <cgmeiner@igalia.com> Reviewed-by: Lucas Stach <l.stach@pengutronix.de> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36358>	2025-07-31 16:31:37 +00:00
Christian Gmeiner	4ad333103f	etnaviv: Emulate rasterizer_discard ... When rasterizer discard is enabled, we need to prevent all rasterization without affecting vertex processing or transform feedback. Emulate this by setting a zero-sized scissor rectangle (0,0,0,0) which effectively clips all fragments while allowing the vertex shader to continue running normally. This matches the behavior implemented by the binary blob. Passes all dEQP-GLES3.functional.rasterizer_discard.* on GC7000. Signed-off-by: Christian Gmeiner <cgmeiner@igalia.com> Reviewed-by: Lucas Stach <l.stach@pengutronix.de> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36373>	2025-07-31 16:12:43 +00:00
Gurchetan Singh	7390276d9b	util: rust: make stubs simpler ... Apparently, this works with a newer rustc. Suggested-by: <nkgold@google.com> Reviewed-by: Marcin Radomski <dextero@google.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36456>	2025-07-31 15:11:07 +00:00
Georg Lehmann	a6a6c2f691	aco/ra: convert bitwise instruction to gfx11+ 16bit on demand ... The 32bit versions are smaller, allow more optimizations and VOPD, so only use the 16bit opcodes if nessecary. Foz-DB Navi31: Totals from 84 (0.10% of 80237) affected shaders: Instrs: 176673 -> 176347 (-0.18%); split: -0.20%, +0.01% CodeSize: 970148 -> 969716 (-0.04%); split: -0.08%, +0.03% VGPRs: 5876 -> 5864 (-0.20%) Latency: 2805974 -> 2805674 (-0.01%); split: -0.02%, +0.01% InvThroughput: 769007 -> 768738 (-0.03%); split: -0.04%, +0.01% VClause: 2593 -> 2597 (+0.15%) Copies: 23749 -> 23487 (-1.10%); split: -1.11%, +0.00% VALU: 107124 -> 106862 (-0.24%); split: -0.25%, +0.00% Reviewed-by: Daniel Schürmann <daniel@schuermann.dev> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35919>	2025-07-31 12:07:07 +00:00
Georg Lehmann	404e1f13e8	aco/print_asm: use real true16 instr on gfx11+ ... Fake16 doesn't print opsel on v_cndmask_b16, so it looks really broken. Restrict to LLVM20+ because older versions have incomplete tru16 support. Reviewed-by: Daniel Schürmann <daniel@schuermann.dev> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35919>	2025-07-31 12:07:07 +00:00
David Rosca	2c3bb204a5	vulkan/video: Fix h265 level values ... From spec: "general_level_idc shall be set equal to a value of 30 times the level number." Fixes: 8243145f02 ("vulkan/video: add a h265 level translator.") Acked-by: Dave Airlie <airlied@redhat.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36285>	2025-07-31 10:08:05 +00:00
Juan A. Suarez Romero	77aa8d44c4	broadcom/ci: disable baremetal rusticl jobs for ci-tron ... After a while testing in parallel baremetal and ci-tron rusticl jobs, disable baremetal ones and keep ci-tron jobs. Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36486>	2025-07-31 09:49:37 +00:00
David Rosca	4011344bc9	radeonsi/vcn: Always enable decode tier2 when supported ... This enables tier2 for VDPAU H264/5. Reviewed-by: Ruijing Dong <ruijing.dong@amd.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36462>	2025-07-31 09:15:33 +00:00
Rhys Perry	8d93684694	nir/load_store_vectorize: always set num_components correctly ... 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/36442>	2025-07-31 08:39:10 +00:00
Rhys Perry	d3476650f1	nir/load_store_vectorize: set is_store for shared append/consume ... So that check_for_aliasing() doesn't ignore them. Signed-off-by: Rhys Perry <pendingchaos02@gmail.com> Backport-to: 25.1 Reviewed-by: Georg Lehmann <dadschoorse@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36442>	2025-07-31 08:39:10 +00:00
Rhys Perry	eeddf6b9e2	nir/load_store_vectorize: check for interfering shared2 before vectorizing ... Only affected shaders in radv_fossils are f1_23. fossil-db (navi21): Totals from 3 (0.00% of 79825) affected shaders: Instrs: 2700 -> 2730 (+1.11%) CodeSize: 17096 -> 17228 (+0.77%) Latency: 8424 -> 8726 (+3.58%) InvThroughput: 3768 -> 3778 (+0.27%); split: -0.05%, +0.32% Copies: 224 -> 234 (+4.46%) PreVGPRs: 291 -> 287 (-1.37%) VALU: 1989 -> 1996 (+0.35%); split: -0.05%, +0.40% fossil-db (gfx1201): Totals from 3 (0.00% of 79839) affected shaders: Instrs: 2862 -> 2908 (+1.61%) CodeSize: 17868 -> 18064 (+1.10%) Latency: 7567 -> 7854 (+3.79%) InvThroughput: 2810 -> 2802 (-0.28%) Copies: 122 -> 120 (-1.64%) PreVGPRs: 291 -> 287 (-1.37%) VALU: 1890 -> 1885 (-0.26%) Signed-off-by: Rhys Perry <pendingchaos02@gmail.com> Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/13616 Backport-to: 25.1 Reviewed-by: Georg Lehmann <dadschoorse@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36442>	2025-07-31 08:39:09 +00:00
Konstantin Seurer	656acb96b0	vulkan/cmd_queue: Reorder memcpy in get_struct_copy ... Using the temporary variable for the memcpy makes sure they are always used so the "(void)tmp_src123" can be removed. Acked-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36331>	2025-07-31 08:13:59 +00:00
Konstantin Seurer	d29f446aa3	vulkan/cmd_queue: Clean up generating copies ... Using the builder makes it much easier to see what is happening and fixes indentation in the process. Acked-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36331>	2025-07-31 08:13:58 +00:00
Konstantin Seurer	d824525a01	vulkan/cmd_queue: Recursively free struct members ... Acked-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36331>	2025-07-31 08:13:58 +00:00
Konstantin Seurer	ea7a2f9834	vulkan/cmd_queue: Improve struct free code indentation ... It also cleans up the way the string is generated because using large templates was a mess. Acked-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36331>	2025-07-31 08:13:58 +00:00
Konstantin Seurer	c29db0965c	vulkan/cmd_queue: Do not free if driver_free_cb is provided ... Avoids crashes when the custom implementation allocates differently. Acked-by: Mike Blumenkrantz <michael.blumenkrantz@gmail.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36331>	2025-07-31 08:13:58 +00:00
Lionel Landwerlin	9efb3ee511	anv: fix wsi image aliasing ... The private BO can get removed due to WSI aliasing and that breaks the submission code (expecting one). Delay the registration on the device to when the image actually gets bound and there is a private BO. Fixes: b21e62b71a ("anv: avoid leaking private binding for aliased wsi image") Reviewed-by: Yiwei Zhang <zzyiwei@chromium.org> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36474>	2025-07-31 07:48:23 +00:00
Job Noorman	19d2f5e086	freedreno/a7xx: disable consts_ubo upload ... consts_ubo is treated as any UBO on a7xx and loaded to the const file via the preamble. Signed-off-by: Job Noorman <jnoorman@igalia.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36470>	2025-07-31 06:03:41 +00:00
Job Noorman	0abaae5a39	tu: remove consts_ubo upload code ... consts_ubo is now treated as any UBO and loaded to the const file via the preamble. Signed-off-by: Job Noorman <jnoorman@igalia.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36470>	2025-07-31 06:03:41 +00:00
Job Noorman	bd58347769	ir3: treat consts_ubo as normal UBO ... Lowering UBO ranges to the const file happens via the preamble on turnip (all gens) and freedreno (a7xx+). The only exception is the consts_ubo, which is uploaded to the const file via CP. Make this more consistent, and remove some special-casing code, by treating consts_ubo as any other UBO. Signed-off-by: Job Noorman <jnoorman@igalia.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36470>	2025-07-31 06:03:41 +00:00
Guilherme Gallo	5ad4e6d3d1	ci/panfrost: Disable DUTs under maintenance ... Collabora's farm 2025-07-31 maintenance will make the following DUTs offline for a few hours: - rk3288-veyron-jaq - sun50i-h6-pine-h64 So let's disable them before it blocks the pipelines. ---- Planned downtime for those DUTs in the farm: Start: 2025-07-31 08:00 BST End: 2025-07-31 10:00 BST Signed-off-by: Guilherme Gallo <guilherme.gallo@collabora.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36480>	2025-07-30 22:19:52 -03:00
Eric Engestrom	fdcad6e547	meson: include VkLayer_MESA_anti_lag in the devenv ... See https://github.com/KhronosGroup/Vulkan-Loader/blob/main/docs/LoaderLayerInterface.md#linux-layer-discovery Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36346>	2025-07-30 23:53:51 +00:00
Eric Engestrom	55d5c3dd8f	meson: include VkLayer_MESA_vram_report_limit in the devenv ... See https://github.com/KhronosGroup/Vulkan-Loader/blob/main/docs/LoaderLayerInterface.md#linux-layer-discovery Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36346>	2025-07-30 23:53:51 +00:00
Eric Engestrom	56c910aafc	meson: include VkLayer_MESA_screenshot in the devenv ... See https://github.com/KhronosGroup/Vulkan-Loader/blob/main/docs/LoaderLayerInterface.md#linux-layer-discovery Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36346>	2025-07-30 23:53:51 +00:00
Eric Engestrom	45789ac396	meson: fix VkLayer_MESA_device_select in the devenv ... It's an implicit layer, not an explicit one like the others in that MR, which I missed in the previous commit. Fixes: 294d8ce80a ("meson: include VkLayer_MESA_device_select in the devenv") Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36346>	2025-07-30 23:53:51 +00:00
Eric Engestrom	6d5f17371c	llvmpipe/ci: set DRIVER_NAME to not have to manually add llvmpipe-skips.txt in asan job variant ... Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36354>	2025-07-30 23:18:09 +00:00
Eric Engestrom	4b4e1a81b9	nouveau/ci: drop already included *-skips.tx ... `all-skips.txt` is always included, and `nouveau-gm20b-skips.txt` matches `$GPU_VERSION-skips.txt`. Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36354>	2025-07-30 23:18:08 +00:00
Eric Engestrom	77f2b9bcc3	iris/ci: drop already included skip list ... It matches `$DRIVER_NAME-skips.txt`. Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36354>	2025-07-30 23:18:07 +00:00
Eric Engestrom	e43ab3fab7	anv/ci: drop already included skip list ... It matches `$GPU_VERSION-skips.txt`. Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36354>	2025-07-30 23:18:07 +00:00
Eric Engestrom	eee467f78a	broadcom/ci: drop redundant HWCI_TEST_SCRIPT already set by .broadcom-test ... Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36354>	2025-07-30 23:18:06 +00:00
Eric Engestrom	fb32b79550	broadcom/ci: drop redundant script: already set by .broadcom-test ... Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36354>	2025-07-30 23:18:05 +00:00
Eric Engestrom	a4dc14dedc	turnip/ci: drop redundant GPU_VERSION ... It's already set by `.a750-mupuf`. Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36354>	2025-07-30 23:18:05 +00:00
Eric Engestrom	d40438031c	radv/ci: deduplicate GPU_VERSION in ci-tron jobs ... Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36354>	2025-07-30 23:18:04 +00:00
Eric Engestrom	84ca8c54f7	radv/ci: deduplicate DEQP_SUITE: radv-valve in ci-tron jobs ... Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36354>	2025-07-30 23:18:03 +00:00
Eric Engestrom	3752723a78	ci-tron: drop unnecessary HWCI_TEST_SCRIPT: deqp-runner.sh re-defines ... This is already the default for ci-tron jobs. Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36354>	2025-07-30 23:18:01 +00:00
Karol Herbst	c2c0e13b78	nak: optimize load_subgroup_id ... If a dimension in workgroup_size is 1 we can simply emit a 0 instead of loading a system value. Helps a couple of compute shaders. Totals: CodeSize: 934788544 -> 934788048 (-0.00%) Static cycle count: 222856984 -> 222856908 (-0.00%) Totals from 8 (0.01% of 87622) affected shaders: CodeSize: 34992 -> 34496 (-1.42%) Static cycle count: 8593 -> 8517 (-0.88%) Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36428>	2025-07-30 23:03:19 +00:00
Alejandro Piñeiro	fa8731b859	broadcom/compiler: update compact arrays comment ... PIPE_CAP_NIR_COMPACT_ARRAYS is gone since commit 2e5d49b3dd v3d properly uses the compact_arrays option from nir_shader_compiler_options since commit d694c1b094 Reviewed-by: Juan A. Suarez <jasuarez@igalia.com> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36446>	2025-07-30 22:45:33 +00:00
Alyssa Rosenzweig	4f1bafa6d5	nir: drop load_sample_id_no_per_sample ... unused now. Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36429>	2025-07-30 22:13:23 +00:00
Alyssa Rosenzweig	04be5a8910	nir/lower_system_values: simplify load_helper_invocation lowering ... The two backends (ir3, dxil) using the lowering have info->fs.uses_sample_shading matching set when sample shading is used -- the VK drivers pass the rasterization state flag into the compiler, while freedreno and d3d12 have caps->force_persample_interp so the frontend creates a shader variant with info->fs.uses_sample_shading set. This means that we can drop the sample-id SHL/AND in the pixel-rate shading case, which in turn means that drivers don't need to have a load_sample_id() that doesn't trigger sample-rate shading (which Adreno doesn't have). Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io> Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36429>	2025-07-30 22:13:23 +00:00
Emma Anholt	ceaa8de981	tu: Implement sampleShadingEnable by flagging uses_sample_shading. ... This will unblock getting correct helper invocation lowering without forcing sample shading, by correctly setting NIR's uses_sample_shading whenever it should be active. Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36429>	2025-07-30 22:13:23 +00:00
Emma Anholt	cfeedd2025	freedreno: Drop min_samples handling code. ... The frontend already does it for us based on !caps->force_persample_interp, and the caller in util_blitter also handles things through per-sample interpolation. Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36429>	2025-07-30 22:13:23 +00:00
Emma Anholt	c1392168e5	tu: Rename per_samp to sample_shading to match ir3. ... Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36429>	2025-07-30 22:13:23 +00:00
Emma Anholt	6e0e6b0caf	ir3: Rename per_samp to sample_shading. ... per_samp ("do sample shading") was confusing compared to has_per_samp ("shader key has some flags per texture sampler to inspect"). Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36429>	2025-07-30 22:13:23 +00:00
Mike Blumenkrantz	dedfb5ef78	zink: trigger multi-context buffer invalidate on internal buffer invalidate ... this is the same as zink_context_replace_buffer_storage() cc: mesa-stable Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36473>	2025-07-30 20:20:15 +00:00
Mike Blumenkrantz	f1ba8f44d9	zink: fix tc buffer replacement rebind condition ... this check is to detect multi-context rebinds, not to determine whether all rebinds on the same context were successful cc: mesa-stable Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36473>	2025-07-30 20:20:15 +00:00
Mike Blumenkrantz	b4280887ed	zink: check for multi-context image/buffer rebinds during dispatch ... Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36473>	2025-07-30 20:20:14 +00:00
Mike Blumenkrantz	1c5e5aa4eb	zink: also add access stage sync when rebinding buffers ... cc: mesa-stable Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36473>	2025-07-30 20:20:14 +00:00
Jason Macnak	72d07c16a3	gfxstream: Add gfxstream TLS connection manager reset ... ... which is still needed to keep the end2end tests happy when one thread creates and destroys instances multiple times (which happens in the MultiThreadedShutdown test). Test: GfxstreamEnd2EndTests Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36471>	2025-07-30 19:41:14 +00:00
Thomas H.P. Andersen	b08a23550f	zink: do not overwrite existing error for miptail on uncommit ... a7f86e38ca added the call to sparse_backing_free. It reuses the ok variable and overwrites the existing value Fixes: a7f86e38ca Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36467>	2025-07-30 19:19:05 +00:00