787ddadf73
util: add mesa_pin_threads environment variable that sets a static affinity mask
...
This is required for deterministic benchmark results.
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27247 >
2024-01-29 19:40:46 +00:00
ea08690d1c
st/mesa: rename ST_L3_PINNING_DISABLED -> ST_THREAD_SCHEDULER_DISABLED
...
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27247 >
2024-01-29 19:40:46 +00:00
386822c1b0
gallium: rename PIPE_.._PIN_THREADS_TO_L3_CACHE -> .._UPDATE_THREAD_SCHEDULING
...
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27247 >
2024-01-29 19:40:46 +00:00
d230434ae3
mesa,gallium: move the thread scheduler to src/util
...
The only change in behavior is that setting the affinity mask is skipped
when it has no effect, which happens when the app thread hasn't been moved
under a different L3 cache by the kernel (the sched_state variable tracks
that). This improves performance because setting the affinity mask incurs
measurable CPU overhead even if it has no effect.
Reviewed-By: Mike Blumenkrantz <michael.blumenkrantz@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27247 >
2024-01-29 19:40:46 +00:00
1c29869c20
Uprev Piglit to e9316bcd12544aaf7e753ce37fe50d64165d9598
...
https://gitlab.freedesktop.org/mesa/piglit/-/compare/f7db20b03de6896d013826c0a731bc4417c1a5a0...e9316bcd12544aaf7e753ce37fe50d64165d9598
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/25449 >
2024-01-29 18:27:07 +00:00
24f14bfacc
dzn: Add missing blit source barriers for enhanced barriers
...
I was under the impression these would never be needed, since the source
should already be in the right GENERIC_READ layout, but when resolving
rendering attachments with a blit, the source is still in the layout that
it was in at the end of rendering.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27313 >
2024-01-29 17:46:11 +00:00
f91ac2ff41
dzn: Add missing condition to immutable sampler init loop
...
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27313 >
2024-01-29 17:46:11 +00:00
fd89ca82a6
dzn: Add a stencil blit fallback
...
Not all D3D hardware supports specifying the stencil ref value from the
pixel shader. For such hardware, stencil blits need to do the awful
8-pass one-bit-per-pass discard-to-not-write blit approach.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27313 >
2024-01-29 17:46:11 +00:00
5cdcb7134a
microsoft/compiler: Use double pack/unpack instead of int for reduce ops on doubles
...
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27313 >
2024-01-29 17:46:11 +00:00
b75ee1a067
frontends/va: make vlVaSyncSurface blocking
...
According to VA documentation, vaSyncSurface should block,
so this mapping function vlVaSyncSurface is changed to be blocking.
add vlVaSyncSurface2 which can be used as non-blocking version.
Signed-off-by: David (Ming Qiang) Wu <David.Wu3@amd.com >
Reviewed-by: Boyuan Zhang <Boyuan.Zhang@amd.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27309 >
2024-01-29 14:02:47 +00:00
dc1b16c010
radv: export alpha-to-coverage via MRTZ for ESO on GFX11
...
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27236 >
2024-01-29 09:47:01 +00:00
6a2b264526
radv: emit more default states for ESO on GFX10.3+
...
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27236 >
2024-01-29 09:47:01 +00:00
b7fdfdc901
radv: do not ignore RADV_DYNAMIC_FRAGMENT_SHADING_RATE for ESO on GFX10.3+
...
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27236 >
2024-01-29 09:47:01 +00:00
9a7c551f95
radv: set NGG fields in vgt_shader_key for ESO on GFX10+
...
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27236 >
2024-01-29 09:47:01 +00:00
70c288558a
radv: fix the late scissor emission workaround with ESO on GFX9
...
graphics_pipeline is NULL when ESO is used.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27236 >
2024-01-29 09:47:01 +00:00
411e23e193
radv: use the non-emitted graphics pipeline for the needed dynamic states
...
This is similar but it will allow us to re-use the helper for the
late scissor bug on GFX9.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27236 >
2024-01-29 09:47:01 +00:00
bbcdd1dc09
radv: refactor gfx103_pipeline_emit_vrs_state()
...
For using it with ESO.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27236 >
2024-01-29 09:47:01 +00:00
e7b4dcd522
radv: refactor gfx103_pipeline_emit_vgt_draw_payload_cntl()
...
For using it with ESO.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27236 >
2024-01-29 09:47:01 +00:00
b9fea5ea6b
virgl: Use better reporting for mirror_clamp features
...
Fixes: 9efe50c83bgert.wollny@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27106 >
2024-01-29 09:01:26 +00:00
fe19405c46
vulkan/runtime: handle new dynamic states for attachment remapping
...
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27307 >
2024-01-29 08:24:14 +00:00
0105cfd93c
radv/ci: enable RADV_PERFTEST=shader_object for vkcts-polaris10-valve
...
This now runs +571k tests for ESO in ~34 minutes compared to ~30 minutes
without shader_object support. Definitely acceptable for a manual job.
Note that the next VKCTS uprev will reduce the number of ESO tests by
a lot because most of them weren't useful for coverage.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27139 >
2024-01-29 07:50:39 +00:00
c68f96878c
radv: fix segfault when getting device vm fault info
...
pFaultInfo can be NULL.
Fixes: 8097becc7fsamuel.pitoiset@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27292 >
2024-01-29 07:26:28 +00:00
2f4d6df372
wsi/common: Allow KHR_present_wait on WL.
...
The fallback in place ensures a good enough solution.
Signed-off-by: Hans-Kristian Arntzen <post@arntzen-software.no >
Reviewed-by: Joshua Ashton <joshua@froggi.es >
Reviewed-by: Sebastian Wick <sebastian.wick@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27275 >
2024-01-28 23:14:36 +00:00
0d51cd4808
wsi/wl: Improve fallback for present_wait.
...
When presentation feedback protocol is not supported,
fallback to using frame callbacks.
In some sense, frame callbacks functions like
present complete + latch delay, so it's a reasonable approach, given the
alternative.
Xwl uses frame callback for COMPLETE events, so it's not a new approach.
To guard against lack of forward progress guarantee,
add a timeout for present complete to avoid deadlocking applications.
Signed-off-by: Hans-Kristian Arntzen <post@arntzen-software.no >
Reviewed-by: Joshua Ashton <joshua@froggi.es >
Reviewed-by: Sebastian Wick <sebastian.wick@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27275 >
2024-01-28 23:14:36 +00:00
1af7f41174
wsi/wl: Refactor out code to update current present ID.
...
Signed-off-by: Hans-Kristian Arntzen <post@arntzen-software.no >
Reviewed-by: Joshua Ashton <joshua@froggi.es >
Reviewed-by: Sebastian Wick <sebastian.wick@redhat.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27275 >
2024-01-28 23:14:36 +00:00
820859a6ab
r600: lower dround_even also on hardware that supports fp64
...
Fixes: aed6a39c10gert.wollny@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27329 >
2024-01-28 22:33:30 +00:00
4a23b28623
ci/VK-GL-CTS: add patches to fix dEQP-VK.glsl.derivate crashes
...
Acked-by: Rob Clark <robdclark@chromium.org >
Signed-off-by: David Heidelberg <david.heidelberg@collabora.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27324 >
2024-01-28 20:41:37 +01:00
68b3b1573b
d3d10umd: Fix compilation
...
Commit eb20ef92eb20ef927https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27238 >
2024-01-27 18:18:35 +00:00
55413e33dc
nvk: Disable all cbufs in nvk_queue_init_context_draw_state()
...
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27320 >
2024-01-27 17:54:19 +00:00
35bd36c1b0
nvk: Whitespace fixes
...
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27320 >
2024-01-27 17:54:19 +00:00
c4254f374f
nvk: Do a second submit to check for errors in the sync case
...
With NVK_DEBUG=push_sync, we signal and immediately wait on a syncobj
to better detect whitch push failed and then print it. However, the
syncobj wait itself isn't enough to check for errors. We have to try
pushing again. Adding an empty push lets us get errors immediately and
print the actually failing pushbuf to stderr.
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27320 >
2024-01-27 17:54:19 +00:00
614e4269cf
nvk: promote load_store_op_none to KHR
...
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27297 >
2024-01-27 17:38:58 +00:00
132b3a79f4
nvk: promote VK_EXT_line_rasterization to KHR
...
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27297 >
2024-01-27 17:38:58 +00:00
da910fc2fd
nvk: promote VK_EXT_index_type_uint8 to KHR
...
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27297 >
2024-01-27 17:38:58 +00:00
32babbc416
nvk: Set ICD version to 1.3
...
This should complete the Vulkan 1.3 puzzle
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27319 >
2024-01-27 16:19:10 +00:00
c467a87e06
radeonsi: Destroy queues before the aux contexts
...
Otherwise the queue might access the already destroyed aux contexts
Signed-off-by: Sebastian Wick <sebastian.wick@redhat.com >
Reviewed-by: Marek Olšák <marek.olsak@amd.com >
Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com >
Cc: mesa-stable
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27300 >
2024-01-27 14:59:11 +00:00
a0a453e43f
nvk: Advertise Vulkan 1.3
...
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27317 >
2024-01-27 06:30:08 +00:00
25d40e8e2c
nvk: Stop requiring dedicated allocations
...
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27317 >
2024-01-27 06:30:08 +00:00
305256e344
venus: handle empty resolved query feedback list
...
On the off chance the combined list resolves to empty due to resets,
skip adding query feedback by not increasing the total cmd buffer
count for query feedback.
Signed-off-by: Juston Li <justonli@google.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27163 >
2024-01-26 23:40:52 +00:00
df1cd20096
venus: recycle linked query feedback immediately during submission
...
Instead of just recyling 1 linked query feedback cmd for use and
defering the actualy recycle, recycle all linked cmds found when
setting up submission immediately.
Signed-off-by: Juston Li <justonli@google.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27163 >
2024-01-26 23:40:52 +00:00
decc5d4539
venus: add comments for query feedback batch free list
...
Signed-off-by: Juston Li <justonli@google.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27163 >
2024-01-26 23:40:52 +00:00
e2c4bafccc
venus: free query batches for VK_COMMAND_POOL_RESET_RELEASE_RESOURCES_BIT
...
When a pool is reset with VK_COMMAND_POOL_RESET_RELEASE_RESOURCES_BIT,
free up all query batches in the pool.
Signed-off-by: Juston Li <justonli@google.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27163 >
2024-01-26 23:40:52 +00:00
aceda1f5f6
venus: acquire mutex when recycling query feedback cmds
...
Fixes: 5b24ab91e4justonli@google.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27163 >
2024-01-26 23:40:52 +00:00
c3417c5bb8
venus: refactor query feedback cmds
...
The list free_query_feedback_cmds for recycling query feedback cmds was
only used in vn_command_pool when it was a vn_feedback_cmd_pool.
For clarity, refactor and store this list in vn_feedback_cmd_pool
instead and introduce a new struct vn_query_feedback_cmd that references
the feedback cmd and the feedback cmd pool for tracking.
Refactor out the allocation portion of query feedback cmds into its own
function for allocating the new vn_query_feedback_cmd struct.
Fixes: 5b24ab91e4justonli@google.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27163 >
2024-01-26 23:40:52 +00:00
3b9466dd51
anv: fix transfer barriers flushes with compute queue
...
Transfer operation are implemented differently on the compute engine
and require a different kind of cache flush.
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Cc: mesa-stable
Reviewed-by: Sagar Ghuge <sagar.ghuge@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27233 >
2024-01-26 20:47:18 +00:00
9f99685c1b
virgl: use PIPE_MAX_SAMPLERS in bind_samplers_states
...
Fix incorrect use of PIPE_MAX_SHADER_SAMPLER_VIEWS in
virgl_bind_samplers_states.
Signed-off-by: Ryan Neph <ryanneph@google.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27251 >
2024-01-26 18:20:13 +00:00
5275df0cfe
gallium: Add ssd130x to the list of kmsro drivers
...
Add an entry point for the ssd130x modesetting driver, to expose a _dri.so
and allow Solomon panels to display the contents rendered by a GPU driver.
Suggested-by: Erico Nunes <nunes.erico@gmail.com >
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com >
Reviewed-by: Erico Nunes <nunes.erico@gmail.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27135 >
2024-01-26 17:52:45 +00:00
9d31680e79
anv: retain ccs image binding address
...
Memory can be free before images it is bound to. When unmapping the
CCS range in the AUX-TT, we cannot rely on the anv_bo::offset field
because the anv_bo might have been freed.
Just save the mapping address/size and use those values at unmapping
time.
Fixes an assert on CI with :
dEQP-VK.synchronization.internally_synchronized_objects.pipeline_cache_graphics
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Fixes: e519e06f4btapani.palli@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27304 >
2024-01-26 17:29:31 +00:00
eead86ad8e
anv: rename aux_tt image field
...
We'll add more to the sub-struct
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Reviewed-by: Jianxun Zhang <jianxun.zhang@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27304 >
2024-01-26 17:29:31 +00:00
fdc2f0a52e
anv: factor out aux-tt binding logic for future reuse
...
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com >
Reviewed-by: Jianxun Zhang <jianxun.zhang@intel.com >
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/27304 >
2024-01-26 17:29:30 +00:00
Marek Olšák