fe38fb858c
brw: workaround broken indirect RT messages on Gfx11
...
Unfortunately we cannot use the indirect descriptor on Gfx11, it
appears to just drop writes. Other platforms appear to be fine.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Cc: mesa-stable
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36883 >
2025-08-20 15:01:50 +00:00
a0844458b8
brw: enable opt_register_coalesce to work with multiple EOT blocks
...
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com >
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36883 >
2025-08-20 15:01:50 +00:00
c4c7ff3f8f
brw: enable register allocation to deal with multiple EOTs
...
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Reviewed-by: Caio Oliveira <caio.oliveira@intel.com >
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36883 >
2025-08-20 15:01:50 +00:00
ed471927e5
vulkan/runtime: use a pipeline flag for unaligned dispatches
...
The problem with the current flag is that it seems to belong to
VkShaderCreateFlagsEXT, not VkPipelineShaderStageCreateFlagBits.
Also it is completely skipped by the vk_pipeline.c code.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Fixes: 7b634ebb63sagar.ghuge@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36828 >
2025-08-20 11:17:52 +00:00
6b88e2bd38
anv/ci: Update expectations from nightly jobs
...
Document current failures and flakes from the nightly jobs, and add a
skip for tests that are timing out.
Signed-off-by: Valentine Burley <valentine.burley@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36608 >
2025-08-20 08:53:36 +00:00
e4fc3e4ee6
anv/ci: Lower concurrency for nightly jobs
...
The nightly jobs can hit OOMs on JSL and ADL, so reduce the number of
threads used by deqp-runner to avoid that.
Signed-off-by: Valentine Burley <valentine.burley@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36608 >
2025-08-20 08:53:36 +00:00
4fda724fd4
brw: Avoid invalid access when compacting out-of-bounds JIP/UIP
...
Usually JIP will be valid, but as part of other changes, it will be
possible to have a shader that have multiple EOT messages and end with
and ENDIF instruction. Its JIP will point after the program ends.
This is fine but was tripping up the compaction code.
Change compaction to not read its internal structures beyond the last
instruction.
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36822 >
2025-08-20 00:54:41 +00:00
148063670d
brw: If the instruction is already a SEND, no need to resize sources
...
Kept an assert as a placeholder in case we had something odd going on
that this code was protecting.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36817 >
2025-08-19 13:54:43 +00:00
cebac156c4
brw: Only access valid sources in lower_btd_logical_send()
...
Only the SHADER_OPCODE_BTD_SPAWN_LOGICAL has sources, so
only reach for them when handling that instruction.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36817 >
2025-08-19 13:54:43 +00:00
dc960936fc
brw: Move resize_sources() earlier when lowering FIND_LIVE_CHANNELS
...
Move it before the new source is used. This currently works because all
instructions have a minimum amount of sources allocated, but a later commit
will change that.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36817 >
2025-08-19 13:54:43 +00:00
fe2e2fabcd
brw: Make sure copied instruction don't copy the list pointers
...
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36817 >
2025-08-19 13:54:43 +00:00
5a34f676a5
brw: Define order for fixes in 3-src operand fix
...
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36817 >
2025-08-19 13:54:43 +00:00
78d3b9cd0a
perfetto: allow specifying clock domain for cpu timestamps
...
Everything is currently using CLOCK_BOOTTIME, which is perfetto's
default, and matches the previous behavior. On some hardware, different
clocks may be better synchronized with the gpu clock.
Signed-off-by: Olivia Lee <olivia.lee@collabora.com >
Reviewed-by: Eric R. Smith <eric.smith@collabora.com >
Reviewed-by: Mary Guillemard <mary.guillemard@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/34390 >
2025-08-19 09:50:36 +00:00
49b917baaf
intel/compiler: Fix ray geometry index
...
We have only 24-bit wide geometry index, not the 28-bit wide.
Signed-off-by: Sagar Ghuge <sagar.ghuge@intel.com >
Reviewed-by: Iván Briano <ivan.briano@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36796 >
2025-08-19 09:32:55 +00:00
7ca356d5db
intel/genxml: Drop all unused struct/fields
...
Signed-off-by: Sagar Ghuge <sagar.ghuge@intel.com >
Reviewed-by: Iván Briano <ivan.briano@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36796 >
2025-08-19 09:32:55 +00:00
6fd4dc353c
elk/algebraic: Protect SHUFFLE from OOB indices
...
Akin to b67230de63sagar.ghuge@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36779 >
2025-08-19 09:15:19 +00:00
b4b692c486
brw/algebraic: Protect SHUFFLE from OOB indices
...
Akin to b67230de63https://gitlab.freedesktop.org/mesa/mesa/-/issues/13351
Reviewed-by: Sagar Ghuge <sagar.ghuge@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36779 >
2025-08-19 09:15:19 +00:00
7b945df668
hasvk: Dead code anv_bo_sync
...
Acked-by: Daniel Stone <daniels@collabora.com >
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36783 >
2025-08-16 00:04:47 -04:00
9944be0be9
hasvk/wsi: Stop requesting signal_*_with_memory
...
Now that we require the dma-buf sync file import/export path, these
legacy paths should never be invoked so we can stop requesting them.
Acked-by: Daniel Stone <daniels@collabora.com >
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36783 >
2025-08-16 00:04:47 -04:00
9cf4872475
hasvk: Require Linux 6.0 for dma-buf sync file import/export
...
This also implies all the other syncobj features we care about so those
become dead code. We'll delete them in following commits.
Acked-by: Daniel Stone <daniels@collabora.com >
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36783 >
2025-08-16 00:04:47 -04:00
5f7c6b2810
hasvk: Require HAS_EXEC_TIMELINE_FENCES
...
i915 has had support for timeline syncobjs for a long time. We might as
well require it at this point.
Acked-by: Daniel Stone <daniels@collabora.com >
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36783 >
2025-08-16 00:04:46 -04:00
3ec62d3a09
hasvk: Require HAS_EXEC_CAPTURE
...
This feature is almost as old as the Vulkan driver itself. We've
required newer kernels for a long time. There's no point in having this
feature bit kicking around.
Acked-by: Daniel Stone <daniels@collabora.com >
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36783 >
2025-08-16 00:04:46 -04:00
5802d2c090
hasvk: Require HAS_EXEC_ASYNC
...
This feature is as old as the Vulkan driver itself. We've required
newer kernels for a long time. There's no point in having this feature
bit kicking around.
Acked-by: Daniel Stone <daniels@collabora.com >
Acked-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36783 >
2025-08-16 00:04:46 -04:00
57aceb96aa
anv: Dead code anv_bo_sync
...
Acked-by: Daniel Stone <daniels@collabora.com >
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36783 >
2025-08-16 00:04:46 -04:00
7ebe93aa9f
anv/wsi: Stop requesting signal_*_with_memory
...
Now that we require the dma-buf sync file import/export path, these
legacy paths should never be invoked so we can stop requesting them.
Acked-by: Daniel Stone <daniels@collabora.com >
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36783 >
2025-08-16 00:04:46 -04:00
affee04bd9
anv: Require Linux 6.0 for dma-buf sync file import/export
...
This also implies all the other syncobj features we care about so those
become dead code. We'll delete them in following commits.
We don't need a check for Xe because the Xe driver was merged into Linux
6.8 while dma-buf sync file import/export landed in 6.0.
Acked-by: Daniel Stone <daniels@collabora.com >
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36783 >
2025-08-16 00:04:46 -04:00
d7416ebc19
intel/gem: Add an intel_gem_supports_dma_buf_sync_file() helper
...
Acked-by: Daniel Stone <daniels@collabora.com >
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36783 >
2025-08-16 00:04:46 -04:00
8044f16bd6
anv/i915: Require HAS_EXEC_TIMELINE_FENCES
...
i915 has had support for timeline syncobjs for a long time. We might as
well require it at this point.
Acked-by: Daniel Stone <daniels@collabora.com >
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36783 >
2025-08-16 00:04:46 -04:00
cb5a2eafd5
anv/i915: Require HAS_EXEC_CAPTURE
...
This feature is almost as old as the Vulkan driver itself. We've
required newer kernels for a long time. There's no point in having this
feature bit kicking around.
Acked-by: Daniel Stone <daniels@collabora.com >
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36783 >
2025-08-16 00:04:46 -04:00
f28eb1bae6
anv/i915: Require HAS_EXEC_ASYNC
...
This feature is as old as the Vulkan driver itself. We've required
newer kernels for a long time. There's no point in having this feature
bit kicking around.
Acked-by: Daniel Stone <daniels@collabora.com >
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36783 >
2025-08-16 00:04:46 -04:00
94931fd4f4
anv: Set the Shader capability when compiling the FP64 shader
...
Otherwise the SPIR-V parser prints a warning the first time the driver
is loaded after a fresh compile.
Fixes: 91b62e9868lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36783 >
2025-08-16 00:04:46 -04:00
381cb2cabd
anv: fix R64* vertex buffer format support
...
Disabled by mistake :(
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Fixes: 5c7397c751ivan.briano@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36771 >
2025-08-14 08:34:00 +00:00
35190aac91
anv: dirty descriptor state on CmdSetDescriptorBufferOffets
...
Calling vkCmdBindDescriptorBuffersEXT() does not invalidate previously
set descriptor sets. Move the state dirtying to
vkCmdSetDescriptorBufferOffets.
Fixes: ab7641b8dclionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36768 >
2025-08-14 07:05:31 +00:00
382b8e3aa2
anv: check for pending_db_mode when dirtying descriptor mode
...
If an application switches back and forth between descriptor sets and
descriptor buffers before executing a draw/dispatch, we could end up in
a wrong state due to pending_db_mode not getting updated.
Fixes: ab7641b8dclionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36768 >
2025-08-14 07:05:31 +00:00
42088cd602
isl/blorp: handle failing 96bpp linear blit case
...
Fix the aux usage assert in blorp for 96bpp linear blit and provide CMF
values for RGB formats supported by isl_format_rgb_to_rgba.
CC: mesa-stable
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/13670
Signed-off-by: Tapani Pälli <tapani.palli@intel.com >
Reviewed-by: Sagar Ghuge <sagar.ghuge@intel.com >
Reviewed-by: Nanley Chery <nanley.g.chery@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36709 >
2025-08-13 16:09:12 +00:00
8a2bf596aa
iris/ci: consistently use x11- prefix for deqp-egl-x11
...
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36759 >
2025-08-13 15:12:31 +00:00
c871a62a75
brw: move URB channel mask shifting to the lowering pass
...
For example Xe2 uses the LSC and doesn´t need the shifting, so let's
just apply it where it's needed.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Reviewed-by: Ivan Briano <ivan.briano@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36757 >
2025-08-13 12:01:49 +00:00
99cf8273f6
anv: stop using descriptor layouts for descriptor buffers push sizes
...
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Reviewed-by: Ivan Briano <ivan.briano@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36757 >
2025-08-13 12:01:49 +00:00
68838d7001
brw: reorder reloc enums to leave embedded samplers at the end
...
So that the driver can allocate an array of relocations using
BRW_SHADER_RELOC_EMBEDDED_SAMPLER_HANDLE + number_of_embedded_samplers
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Reviewed-by: Ivan Briano <ivan.briano@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36757 >
2025-08-13 12:01:49 +00:00
beadc1f93a
anv/i915: print bo->map when dumping exec buffers bos
...
This makes it easier to verify if the host allocation a user-ptr bo is
assigned to still exists. The kernel rejects command submissions with
user-ptr bos pointing to non-mapped host memory, so this makes it easier
to debug those.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36701 >
2025-08-13 08:30:43 +00:00
8d8f5558eb
anv: do not map from_host_ptr bos in image_bind_address
...
Rusticl running on zink might end up creating an 1D image from a host_ptr
and zink might bind it with VK_IMAGE_USAGE_HOST_TRANSFER_BIT_EXT.
That ended up hitting an assert inside anv_device_map_bo.
Cc: mesa-stable
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36701 >
2025-08-13 08:30:43 +00:00
a973fb1bb0
anv: fix missing meson dep
...
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Cc: mesa-stable
Closes: https://gitlab.freedesktop.org/mesa/mesa/-/issues/13627
Reviewed-by: Tapani Pälli <tapani.palli@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36737 >
2025-08-13 08:10:39 +00:00
9a6065852d
anv: Do not consider task as prerasterization
...
What drivers state programming in the 3D pipeline is the mesh shader.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Reviewed-by: Ivan Briano <ivan.briano@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36734 >
2025-08-13 06:24:44 +00:00
46c16f854e
brw: compute consistent clip/cull distance masks with VUE
...
We can optimize the VUE layout in cases where all shaders are compiled
together and some outputs are unused. So we need to have consistent
clip/cull_distance_mask with the VUE.
Previously we could have a VUE without ClipDistance present in the
header and yet have a non zero clip_distance_mask. This would trip the
HW into taking into account a VUE field that doesn't exist.
Here we set the clip/cull_distance_mask to 0 if the associated output
is not written by the shader. The written outputs are always
consistent with what's in the VUE.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Fixes: 2d396f6085https://gitlab.freedesktop.org/mesa/mesa/-/issues/13685
Reviewed-by: Ivan Briano <ivan.briano@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36734 >
2025-08-13 06:24:44 +00:00
cac3b4f404
anv: Mask off excessive invocations
...
For unaligned invocations, don't launch two COMPUTE_WALKER, instead we
can mask off excessive invocations in the shader itself at nir level and
launch one additional workgroup.
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/36245 >
2025-08-12 23:17:02 +00:00
7b634ebb63
vulkan/runtime: Add VK_SHADER_CREATE_UNALIGNED_DISPATCH_BIT_MESA flag
...
Drivers that doesn't support direct unaligned dispatches, they can use
the shader creation flag to lower unaligned dispatches.
Suggested-by: Alyssa Rosenzweig <alyssa@rosenzweig.io >
Signed-off-by: Sagar Ghuge <sagar.ghuge@intel.com >
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io >
Reviewed-by: Konstantin Seurer <konstantin.seurer@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36245 >
2025-08-12 23:17:02 +00:00
349de5b0be
anv: Use vk_get_bvh_build_pipeline_spv helper
...
Took inspiration from RADV driver changes. This allow us to get rid of
our local helper get_pipeline_spv().
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/36245 >
2025-08-12 23:17:02 +00:00
3f995bf817
anv: Enable G8_B8_R8_3PLANE_422 and G8_B8_R8_3PLANE_444 formats
...
They are well supported and useful for efficient playback of software
decoded videos - e.g. when allocating buffers with udmabuf or dma heaps
and then importing them into Vulkan.
Signed-off-by: Robert Mader <robert.mader@posteo.de >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/35273 >
2025-08-12 19:16:29 +00:00
c7045e3e63
perfetto: unify init
...
util_perfetto_init() was called in some places, util_cpu_trace_init()
was called in other places, and some places used tracing without ever
calling either of them
util_cpu_trace_init() is now guaranteed to be called:
* on gallium screen create
* on VK instance create
thus no driver/frontend/etc should ever need to call this manually
cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36628 >
2025-08-12 13:17:43 +00:00
6ece4f3fa0
vulkan: Add a vk_video_session_finish() helper
...
It's always better if init/finish come in pairs.
Reviewed-by: Dave Airlie <airlied@redhat.com >
Reviewed-by: Hyunjun Ko <zzoon@igalia.com >
Reviewed-by: David Rosca <david.rosca@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/36646 >
2025-08-11 20:34:56 +00:00
Lionel Landwerlin