Commit graph

11552 commits

Author SHA1 Message Date
Samuel Pitoiset
dba9809e0c radv: remove redundant radv_sdma_surf::is_3d
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40186>
2026-03-09 08:40:44 +00:00
Samuel Pitoiset
1a00587c44 radv: fix a GPU hang with PS epilogs and secondary command buffers
If the secondary changes the fragment output state and if the same
PS epilog used before ExecuteCommands() is re-bind immediately after
that call, the PS epilog state wouldn't be re-emitted.

Apply the same change for VS prologs, although the logic is slightly
different and the bug shouldn't occur. The whole logic of secondaries
should be completely rewritten because it's definitely not robust.

This fixes a GPU hang in Where Winds Meet, see
https://github.com/doitsujin/dxvk/issues/5436.

Cc: mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40221>
2026-03-09 08:16:49 +00:00
Samuel Pitoiset
ac3fd06987 radv: always enable DISABLE_CONSERVATIVE_ZPASS_COUNTS on GFX11
Some checks are pending
macOS-CI / macOS-CI (dri) (push) Waiting to run
macOS-CI / macOS-CI (xlib) (push) Waiting to run
This might cause incorrect occlusion queries count.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40235>
2026-03-09 07:26:25 +00:00
Georg Lehmann
406935c6fe radv: use nir_opt_fp_math_ctrl
Some checks are pending
macOS-CI / macOS-CI (dri) (push) Waiting to run
macOS-CI / macOS-CI (xlib) (push) Waiting to run
Foz-DB Navi21:
Totals from 10833 (9.63% of 112497) affected shaders:
Instrs: 10090308 -> 10043030 (-0.47%); split: -0.48%, +0.01%
CodeSize: 53681564 -> 53556756 (-0.23%); split: -0.25%, +0.01%
VGPRs: 511568 -> 511296 (-0.05%); split: -0.08%, +0.03%
SpillSGPRs: 2442 -> 2438 (-0.16%); split: -0.20%, +0.04%
Latency: 58989785 -> 58935280 (-0.09%); split: -0.18%, +0.09%
InvThroughput: 15142587 -> 15067217 (-0.50%); split: -0.52%, +0.02%
VClause: 200588 -> 200410 (-0.09%); split: -0.20%, +0.11%
SClause: 257273 -> 257262 (-0.00%); split: -0.20%, +0.19%
Copies: 741430 -> 741397 (-0.00%); split: -0.22%, +0.22%
Branches: 211023 -> 211020 (-0.00%); split: -0.00%, +0.00%
PreSGPRs: 491752 -> 491663 (-0.02%); split: -0.02%, +0.00%
PreVGPRs: 418558 -> 418089 (-0.11%); split: -0.12%, +0.01%
VALU: 7064149 -> 7017847 (-0.66%); split: -0.66%, +0.01%
SALU: 1227287 -> 1226639 (-0.05%); split: -0.13%, +0.07%
SMEM: 449268 -> 449343 (+0.02%)

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40098>
2026-03-07 08:16:29 +01:00
Samuel Pitoiset
395132432b radv: initialize HiZ also for depth-only images
This is actually needed to initialize it to full expanded HiZ range.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40213>
2026-03-05 13:58:18 +00:00
Samuel Pitoiset
eaa51f4737 radv: remove unused radv_gfx12_get_hiz_clear_value()
This isn't necessary because depth/stencil clears use graphics.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40213>
2026-03-05 13:58:18 +00:00
Samuel Pitoiset
ab3ae53906 radv: add radv_image_has_hiz_metadata()
And rename few things.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40213>
2026-03-05 13:58:18 +00:00
Samuel Pitoiset
6358e48820 radv: add radv_image_has_hiz()
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40213>
2026-03-05 13:58:17 +00:00
Rhys Perry
7de3552f3f radv: don't mask PS epilog spi_shader_col_format with dual source blending
In case the FS only writes one output.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40005>
2026-03-05 09:38:24 +00:00
Rhys Perry
b4dc39f632 ac/nir: fix when both dual source outputs are unwritten
When this is the case. we shouldn't hang or crash.

Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40005>
2026-03-05 09:38:23 +00:00
Samuel Pitoiset
783c065f3b radv: always fast-clear non-MSAA color image with comp-to-single on GFX10-10.3
Some checks are pending
macOS-CI / macOS-CI (dri) (push) Waiting to run
macOS-CI / macOS-CI (xlib) (push) Waiting to run
This should enable fast-clear for color images with the GENERAL layout
on GFX10-10.3. This seems important because DXVK tends to use that
layout more often now.

There are still issues with MSAA images, so it's only enabled for
single-sampled.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40145>
2026-03-05 08:34:43 +00:00
Samuel Pitoiset
9de465da81 radv: remove radv_disable_depth_storage drirc
This was only for TW3.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40214>
2026-03-05 07:55:27 +00:00
Samuel Pitoiset
74defc5f24 radv: disable STORAGE for depth-only formats
This is actually not needed because nobody is using storage with
depth-only formats and compression doesn't work at all anyways.

PAL and native don't allow this either.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40214>
2026-03-05 07:55:26 +00:00
Marek Olšák
f421eecebe radv,radeonsi: PA_SC_BINNER changes for gfx12
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/40022>
2026-03-04 21:14:56 +00:00
Marek Olšák
22f3761579 radv: make radv_postprocess_binary_config more correct and more readable
- don't set fields that don't exist on some generations
- add gfx_level checks for MEM_ORDERED even when it's technically not needed

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/40022>
2026-03-04 21:14:56 +00:00
Marek Olšák
fd6941650c ac: add ac_cu_info::has_fmask, adjust FMASK checks
Some of these should check has_fmask, others should check < GFX11.

v2: move to ac_cu_info

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/40022>
2026-03-04 21:14:56 +00:00
Samuel Pitoiset
37ba573877 radv/meta: set the depth/stencil resolve region depth to 1
Some checks are pending
macOS-CI / macOS-CI (dri) (push) Waiting to run
macOS-CI / macOS-CI (xlib) (push) Waiting to run
No 3D.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40135>
2026-03-04 15:12:54 +00:00
Samuel Pitoiset
4d3549d372 radv/meta: rework the barrier for depth/stencil resolves
Yes, COLOR_ATTACHMENT_WRITE_BIT can be used to synchronize ds resolves.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40135>
2026-03-04 15:12:54 +00:00
Samuel Pitoiset
e7d135ae51 radv/meta: optimize synchronization with compute resolves
It's only required to synchronize at the end, so that all compute
resolves can run in parallel.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40135>
2026-03-04 15:12:53 +00:00
Samuel Pitoiset
fc0c3f6d84 radv/meta: remove an useless barrier before color resolves with graphics
No reason to require a barrier either because there is already one
before doing resolves and decompressions should already be correctly
synchronized.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40135>
2026-03-04 15:12:53 +00:00
Samuel Pitoiset
c9c16d62dd radv/meta: remove an useless barrier after color resolves with graphics
The app must synchronize with
VK_PIPELINE_STAGE_2_COLOR_ATTACHMENT_OUTPUT_BIT anyways.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40135>
2026-03-04 15:12:52 +00:00
Samuel Pitoiset
421b15cd3a radv/meta: remove redundant barriers in vkCmdResolveImage2()
The source image layout must be either TRANSFER_SRC or GENERAL and the
application must emit the image layout transition. There is no reason
the source image wouldn't be readable by shaders.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40135>
2026-03-04 15:12:52 +00:00
Rhys Perry
5c3b5688a1 amd: rename ac_cu_info to ac_compiler_info
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40042>
2026-03-03 08:50:12 +00:00
Rhys Perry
a65089dfce ac/nir: pass ac_cu_info to ac_nir_compute_tess_wg_info
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40042>
2026-03-03 08:50:11 +00:00
Rhys Perry
8801ca188d ac/nir: don't pass radeon_info to ac_nir_set_options
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40042>
2026-03-03 08:50:10 +00:00
Rhys Perry
5a8a7dbb22 ac/nir: don't pass radeon_info to NGG lowering
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40042>
2026-03-03 08:50:09 +00:00
Rhys Perry
36feec61c8 ac/nir: use ac_nir_lower_ngg_options for ac_nir_lower_ngg_mesh
Signed-off-by: Rhys Perry <pendingchaos02@gmail.com>
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40042>
2026-03-03 08:50:09 +00:00
Samuel Pitoiset
27b6ad994a radv: fix computing pitch/slice_pitch for compressed block formats
Also there is no need to multiply for dividing later.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40121>
2026-03-02 14:44:14 +00:00
Samuel Pitoiset
67c985d754 radv: mark linear images for SDMA as potentially compressed on GFX12
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40121>
2026-03-02 14:44:14 +00:00
Samuel Pitoiset
e6e305988c ac,radv,radeonsi: merge tiled/linear surfaces into one struct
This will be used to rework/fix metadata config on SDMA7.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40121>
2026-03-02 14:44:14 +00:00
Samuel Pitoiset
1a695d50b2 radv: tidy up determining 3D alignment for SDMA
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40121>
2026-03-02 14:44:14 +00:00
Samuel Pitoiset
afc261d43a radv: remove redundant radv_sdma_surf::micro_tile_mode
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40121>
2026-03-02 14:44:14 +00:00
Samuel Pitoiset
cc21e61e43 radv: fix copying images with different swizzle modes on SDMA7
Swizzle modes must match on SDMA7 (GFX12), and the micro tile mode
doesn't exist.

Cc: mesa-stable
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40121>
2026-03-02 14:44:14 +00:00
Marek Olšák
f22f117d1a amd: add meson variable idep_amd_generated_headers for all generated headers
group all generated header under the same variable

Acked-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40084>
2026-02-28 05:23:59 +00:00
Mike Blumenkrantz
b8ee6f3d30 radv: fix multiview fast clears
this was only clearing layer0 because it was ignoring the viewmask

cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40123>
2026-02-27 15:07:01 +00:00
Samuel Pitoiset
5fa94dba34 radv: only enable DCC/HTILE if it's compressed with SDMA
Some checks are pending
macOS-CI / macOS-CI (dri) (push) Waiting to run
macOS-CI / macOS-CI (xlib) (push) Waiting to run
It's wrong to enable compression if it's not compressed.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40101>
2026-02-27 11:23:48 +00:00
Samuel Pitoiset
d583339e79 radv/meta: add a single function for color/depth&stencil fragment resolves
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40064>
2026-02-27 08:21:39 +00:00
Samuel Pitoiset
8d0206c95f radv/meta: simplify creating pipelines for color/depth&stencil fragment resolves
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40064>
2026-02-27 08:21:39 +00:00
Samuel Pitoiset
009cf9bfa7 radv/meta: pass a resolve mode for fragment color resolves
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40064>
2026-02-27 08:21:38 +00:00
Samuel Pitoiset
2c5685ed89 radv/meta: inline one function in the fragment resolve path
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40064>
2026-02-27 08:21:38 +00:00
Samuel Pitoiset
876b204703 radv/meta: use the same shader for color/depth&stencil fragment resolves
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40064>
2026-02-27 08:21:37 +00:00
Samuel Pitoiset
fd9cbff41a radv/meta: move the barrier for color fragment resolves outside
This barrier is only needed for rendering resolves (ie. not for
vkCmdResolveImage()). Though, these barriers are likely unnecessary
but let's keep them for now.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40064>
2026-02-27 08:21:37 +00:00
Samuel Pitoiset
255439fd9c radv/meta: fix the shader stage for push constants in the fragment resolve path
RADV doesn't use it but better to set it correctly.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40064>
2026-02-27 08:21:37 +00:00
Samuel Pitoiset
ed546ce809 radv/meta: add a single function for color/depth&stencil compute resolves
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40064>
2026-02-27 08:21:36 +00:00
Samuel Pitoiset
94234f35fd radv/meta: simplify creating pipelines for color/depth&stencil compute resolves
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40064>
2026-02-27 08:21:36 +00:00
Samuel Pitoiset
1c7b416daa radv/meta: pass a resolve mode for compute color resolves
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40064>
2026-02-27 08:21:36 +00:00
Samuel Pitoiset
2ae32d709c radv/meta: inline one function in the compute resolve path
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40064>
2026-02-27 08:21:35 +00:00
Samuel Pitoiset
e3e52c348c radv/meta: use the same shader for color/depth&stencil compute resolves
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40064>
2026-02-27 08:21:35 +00:00
Samuel Pitoiset
08f147d92c radv/meta: add depth/stencil support to the core resolve shader
This removes lot of duplicated code.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40064>
2026-02-27 08:21:34 +00:00
Samuel Pitoiset
1e561eb1d7 radv/meta: replace radv_meta_resolve_type by VkImageAspectFlags
It's either depth or stencil.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/40064>
2026-02-27 08:21:34 +00:00