AlexIndustrial/mesa
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
202,423 Commits 1 Branch 0 Tags
dcb5cfbfcc4431b6f72002fd7743cbd274dc43d2
Commit Graph

202423 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Lionel Landwerlin dcb5cfbfcc vulkan/runtime: add a multialloc vk_shader allocator 
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33792>
 2025-02-27 10:01:17 +02:00
Lionel Landwerlin 009ef67c8d vulkan/runtime: pass robustness state to preprocess vfunc 
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33792>
 2025-02-27 10:01:16 +02:00
Lionel Landwerlin 4dba1ad93f vulkan/runtime: ensure robustness state is fully initialized 
This is part of the hashing key :

==25753== Uninitialised byte(s) found during client check request
==25753==    at 0x93D29AE: blob_write_bytes (blob.c:164)
==25753==    by 0x93A62C6: vk_pipeline_precomp_shader_serialize (vk_pipeline.c:722)
==25753==    by 0x93AC55E: vk_pipeline_cache_add_object (vk_pipeline_cache.c:433)
==25753==    by 0x93A691B: vk_pipeline_precompile_shader (vk_pipeline.c:875)
==25753==    by 0x93A8FB9: vk_create_graphics_pipeline (vk_pipeline.c:1715)
==25753==    by 0x93A9799: vk_common_CreateGraphicsPipelines (vk_pipeline.c:1860)
==25753==  Address 0xf1adf82 is 82 bytes inside a block of size 152 alloc'd
==25753==    at 0x64FA858: malloc (in /usr/libexec/valgrind/vgpreload_memcheck-amd64-linux.so)
==25753==    by 0x99AAC38: vk_default_alloc (vk_alloc.c:26)
==25753==    by 0x93A403B: vk_alloc (vk_alloc.h:48)
==25753==    by 0x93A406B: vk_zalloc (vk_alloc.h:56)
==25753==    by 0x93A60A0: vk_pipeline_precomp_shader_create (vk_pipeline.c:680)
==25753==    by 0x93A689D: vk_pipeline_precompile_shader (vk_pipeline.c:866)
==25753==    by 0x93A8FB9: vk_create_graphics_pipeline (vk_pipeline.c:1715)
==25753==    by 0x93A9799: vk_common_CreateGraphicsPipelines (vk_pipeline.c:1860)

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Fixes: 9308e8d90d ("vulkan: Add generic graphics and compute VkPipeline implementations")
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33792>
 2025-02-27 10:01:02 +02:00
Tapani Pälli 78e5157a9c intel/compiler: add a spec note about L1WT types being uncached 
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33755>
 2025-02-27 05:38:35 +00:00
Peyton Lee 7c8d58c26c radeonsi/vpe: vpe support hdr input 
when an application asks for supported formats
will return HDR formats(2020, explicit) is supported.

Signed-off-by: Peyton Lee <peytolee@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33731>
 2025-02-27 03:15:17 +00:00
Peyton Lee 43ce5b1138 radeonsi/vpe: vpe support tonemapping 
if input source is HDR stream, vpe can use gmlib generating tonemapping
table to convert HDR image to SDR image.

Signed-off-by: Peyton Lee <peytolee@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33731>
 2025-02-27 03:15:17 +00:00
Peyton Lee 2e46c41448 amd/gmlib: add gmlib for radeonsi 
radeonsi drivers can use gmlib to generate 3dlut used to do tonemapping.

Signed-off-by: Peyton Lee <peytolee@amd.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33731>
 2025-02-27 03:15:16 +00:00
Marek Olšák 2e124dd389 util: remove glthread enablement from app profiles 
It's mature, so if you want it, just enable it for your driver by default.

Reviewed-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33480>
 2025-02-27 02:28:58 +00:00
Faith Ekstrand 8fffcdb18b nak/nir: Re-materialize load_const instructions in use blocks 
This is useful both for correctness (to ensure that things we think are
constant stay constant) and it improves performance a bit by reducing
register pressure and avoiding spilling.

Pipeline-db stats:

    CodeSize: 29665072 -> 29437344 (-0.77%); split: -0.92%, +0.16%
    Number of GPRs: 157124 -> 156082 (-0.66%)
    SLM Size: 148900 -> 146436 (-1.65%)
    Static cycle count: 6840286 -> 6805711 (-0.51%); split: -0.98%, +0.47%
    Spills to memory: 177779 -> 173337 (-2.50%)
    Fills from memory: 177779 -> 173337 (-2.50%)
    Spills to reg: 17692 -> 16731 (-5.43%)
    Fills from reg: 12013 -> 11897 (-0.97%)
    Max warps/SM: 309128 -> 309456 (+0.11%)

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33771>
 2025-02-27 00:26:54 +00:00
Faith Ekstrand 8de37b142e nvk: Only support compute shader derivatives on Turing+ 
Fixes: e0e7d8d910 ("nvk: Advertise VK_NV/KHR_compute_shader_derivatives")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33771>
 2025-02-27 00:26:54 +00:00
Faith Ekstrand bd04fdcb2b nvk: Only support deviceGeneratedCommandsMultiDrawIndirectCount on Turing+ 
Indirect draws on Maxwell involve patching pushbufs together and doing
that isn't possible with device generated commands.

Fixes: 83b220f833 ("nvk: Advertise VK_EXT_device_generated_commands")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33771>
 2025-02-27 00:26:54 +00:00
Faith Ekstrand 7e12ba8709 nvk: Handle pre-Turing dispatch indirect commands 
The QMD layout is a bit different.

Fixes: 976f22a5da ("nvk: Implement CmdProcess/ExecuteGeneratedCommandsEXT")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33771>
 2025-02-27 00:26:54 +00:00
Faith Ekstrand c540e5e2cc nak/qmd: Add a nak_get_qmd_cbuf_desc_layout() helper 
Fixes: 976f22a5da ("nvk: Implement CmdProcess/ExecuteGeneratedCommandsEXT")
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33771>
 2025-02-27 00:26:54 +00:00
Faith Ekstrand 755a6884d3 nak/qmd: Drop some unnecessary .try_into().unwrap() 
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33771>
 2025-02-27 00:26:54 +00:00
Faith Ekstrand 59f7f76166 nouveau/class_parser: Make strided element functions const 
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33771>
 2025-02-27 00:26:54 +00:00
Faith Ekstrand 0c4be9e0ff nvk: Fix indentation in begin_end_query() 
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33771>
 2025-02-27 00:26:54 +00:00
Mel Henning 2a0302967f nak: Add spill/fill statistics 
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33773>
 2025-02-26 23:52:31 +00:00
Mel Henning 0480d8294c nak/spill_values: Make Spill take &mut self 
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33773>
 2025-02-26 23:52:31 +00:00
Mel Henning ae0576a7f8 nak: Add an occupancy statistic 
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33773>
 2025-02-26 23:52:31 +00:00
Mel Henning 890bab86b4 nak: Add static cycle count statistic 
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33773>
 2025-02-26 23:52:31 +00:00
Lionel Landwerlin b72c772f83 spirv: fix racy build 
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reported-by: Ki'sak
Acked-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33769>
 2025-02-26 23:09:51 +00:00
Dave Airlie 8a64eee4d7 nvk: enable float16 on turing. 
Reviewed-by: Mel Henning <drawoc@darkrefraction.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33521>
 2025-02-26 22:48:49 +00:00
Dave Airlie b45feed3b2 nak: adjust latencies on fp16/64 instructions on Turing 
These instructions on Turing require longer latencies,

this fixes the float16 tests on Turing.

Reviewed-by: Mel Henning <drawoc@darkrefraction.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33521>
 2025-02-26 22:48:49 +00:00
Paulo Zanoni fd10764cff brw: extend the NOP+WHILE workaround 
It turns out that we need to add a NOP not only in between two
consecutive WHILE instructions, but also after every control flow
instruction that immediately precedes a WHILE.

v2: Rebase after the renames.

Fixes: 5ca883505e ("brw: add a NOP in between WHILE instructions on LNL")
Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33021>
 2025-02-26 22:23:16 +00:00
Paulo Zanoni 3596b4e325 brw: add instructions missing from is_control_flow() 
I'm not aware of any workloads that will be impacted by this change,
but let's keep our list of control flow instructions complete. A
shader-db run on MTL tells me nothing changes.

v2: "The scheduler relies on HALT not being considered control flow to
be able to move code past HALT instructions. Doing this would prevent
such optimization from happening and would reduce performance
dramatically in some cases." - Francisco.

Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Paulo Zanoni <paulo.r.zanoni@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33021>
 2025-02-26 22:23:16 +00:00
Mike Blumenkrantz 0e87acb2f3 dri: delete INVALIDATE extension 
this shouldn't be needed anymore since it doesn't do anything

Reviewed-by: Adam Jackson <ajax@redhat.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33774>
 2025-02-26 21:46:06 +00:00
Erik Faye-Lund 6ab4a0b5c9 panfrost: respect pipe_rasterizer_state::line_rectangular 
The state-tracker already tells us if we should use rectangular ends or
not on our lines, so we don't need to manually infer this from
combination of states.

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33382>
 2025-02-26 21:11:06 +00:00
Erik Faye-Lund 6bb4971497 pan/genxml: rename field 
This field is really about the line-shape, not multisampling or not.

Yeah, in OpenGL, these two concepts are kinda intertwined. But this is
what the state actually does, so let's name it based on that.

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33382>
 2025-02-26 21:11:06 +00:00
Karol Herbst dad5ee1039 intel/brw, lp: enable lower_pack_64_4x16 
The compiler won't be able to emit pack_64_4x16, so we should prevent
nir_opt_algebraic to optimize to it. This fixes an infinite optimization
loop inside brw_nir_optimize:

nir_copy_prop
    16x4     %77 = @load_global (%80)
    32    %61995 = pack_32_2x16_split %77.x, %77.y
    32    %61998 = pack_32_2x16_split %77.z, %77.w
    64    %61999 = pack_64_2x32_split %61995, %61998
    64       %76 = iadd %100, %79
                   @store_global (%61999, %76)

nir_opt_algebraic
    16x4     %77 = @load_global (%80)
    32    %61995 = pack_32_2x16_split %77.x, %77.y
    32    %61998 = pack_32_2x16_split %77.z, %77.w
    16x4  %62000 = vec4 %77.x, %77.y, %77.z, %77.w
    64    %62001 = pack_64_4x16 %62000
    64       %76 = iadd %100, %79
                   @store_global (%62001, %76)

nir_lower_pack
    16x4     %77 = @load_global (%80)
    16x4  %62000 = vec4 %77.x, %77.y, %77.z, %77.w
    16    %62002 = mov %62000.y
    16    %62003 = mov %62000.x
    32    %62004 = pack_32_2x16_split %62003, %62002
    16    %62005 = mov %62000.w
    16    %62006 = mov %62000.z
    32    %62007 = pack_32_2x16_split %62006, %62005
    64    %62008 = pack_64_2x32_split %62004, %62007
    64       %76 = iadd %100, %79
                   @store_global (%62008, %76)

// brw_nir_optimize loops here

nir_copy_prop
    16x4     %77 = @load_global (%80)
    32    %62004 = pack_32_2x16_split %77.x, %77.y
    32    %62007 = pack_32_2x16_split %77.z, %77.w
    64    %62008 = pack_64_2x32_split %62004, %62007
    64       %76 = iadd %100, %79
                   @store_global (%62008, %76)

llvmpipe has a similar issue inside lp_build_opt_nir

Fixes: b1bc691b0f ("nir/algebraic: add and improve pack/unpack patterns")
Acked-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33347>
 2025-02-26 20:43:39 +00:00
Mike Blumenkrantz 480d8bea2b zink: put zink_program::reference on separate cacheline 
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33738>
 2025-02-26 20:19:50 +00:00
Mike Blumenkrantz 08dc6aa354 zink: use a separate ralloc ctx for zink_program objects 
I considered removing ralloc at all here but it was more deeply embedded
than I realized. maybe a project for another time

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33738>
 2025-02-26 20:19:50 +00:00
Ian Romanick 495812d8e0 brw/print: Don't let SHADER_OPCODE_FLOW affect indentation 
In `fossilize-replay --pipeline-hash 375a63e14afa96c4
fossils/fossil-db/steam-dxvk/f1_22_abu_dhabi.dx12vk-ultra.foz`,
`cf_count` would get decremented below zero. This would lead trying to
print `UINT_MAX` levels of indentation just a few lines below. I ran
out of disk space and patience before that finished. 🤣

Reviewed-by: Caio Oliveira <caio.oliveira@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33748>
 2025-02-26 19:50:30 +00:00
Yiwei Zhang bef1c23a23 venus: re-enable 1.4 support 
Have made core venus protocol update to be able to support host image
copy.

Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33757>
 2025-02-26 19:16:48 +00:00
Yiwei Zhang 577fa1793d venus: enable VK_EXT_host_image_copy support 
Fill the core features and properties properly, and conditionally pass
through support of the extension based on the renderer venus protocol
spec version.

Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33757>
 2025-02-26 19:16:48 +00:00
Yiwei Zhang f3da8167d6 vulkan/util: clean up copy_property and fix for setter 
For setters, e.g. vk_set_physical_device_properties_struct used by venus
to fill all props, the out array storage comes from the driver, so we'd
assign directly. This change also fixes the template indent and drops an
unused arg.

Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33757>
 2025-02-26 19:16:48 +00:00
Yiwei Zhang c13da1c744 venus: implement host image copy commands 
The img-2-img and layout transition are trivial passthrough. For
img-2-mem and mem-2-img copies, host pointer has to be sized for proper
protocol encoding and decoding, and we have to either query or calculate
on our own based on VK_HOST_IMAGE_COPY_MEMCPY_EXT flag being used or
not.

Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33757>
 2025-02-26 19:16:48 +00:00
Yiwei Zhang 036493f43d venus: extend image format cache for host copy props 
Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33757>
 2025-02-26 19:16:48 +00:00
Yiwei Zhang 2eb5a75d51 venus: sync latest protocol v3 support for host copy 
Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33757>
 2025-02-26 19:16:48 +00:00
Yiwei Zhang fde5cebec5 venus: fix image format cache miss with AHB usage query 
should skip updating cache key instead of marking as a miss

Fixes: e48645250c ("venus: image format properties cache")
Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33757>
 2025-02-26 19:16:48 +00:00
Yiwei Zhang 15777727c9 venus: use sharing_mode from common vk_image 
Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33757>
 2025-02-26 19:16:48 +00:00
Yiwei Zhang f9afd65695 venus: drop unnecessary struct 
Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33757>
 2025-02-26 19:16:48 +00:00
Yiwei Zhang bcb0b8c7f6 venus: suppress a few -Wmaybe-uninitialized 
Signed-off-by: Yiwei Zhang <zzyiwei@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33757>
 2025-02-26 19:16:48 +00:00
Mike Blumenkrantz f9fe08740a zink: always fully unwrap contexts 
threaded_context_unwrap_sync() can be called safely on non-threaded
contexts

cc: mesa-stable

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33742>
 2025-02-26 18:54:25 +00:00
Lionel Landwerlin d0c980caa7 brw: avoid setting up the sampler header bits when unused 
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33704>
 2025-02-26 17:19:04 +00:00
Lionel Landwerlin 8b4f997168 brw: optimize load payload with immediate headers 
Currently the condition to use a single MOV is failing on immediate
values, so we emit 2 MOVs in SIMD8 instead of a single SIMD16.

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33704>
 2025-02-26 17:19:04 +00:00
Rob Clark 513184fa44 tu: Suballoc VkEvent BOs 
No need to burn an entire PAGE_SIZE BO for an event.  And in particular
the pattern of allocate + immediate mmap is expensive in a VM.
Suballocating cuts down the # of times we do this in
dEQP-VK.api.command_buffers.execute_large_primary from 10000 to 157,
avoiding problems with the test running up against watchdog timeout.

Signed-off-by: Rob Clark <robdclark@chromium.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33721>
 2025-02-26 16:51:08 +00:00
Alyssa Rosenzweig c3cc756cf9 ir3: clean up progress manually 
Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33722>
 2025-02-26 15:19:53 +00:00
Alyssa Rosenzweig 266638d10a nir: clean up progress 
semantic patch made a few bad choices.

Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33722>
 2025-02-26 15:19:53 +00:00
Alyssa Rosenzweig 593308a685 nir: eliminate nir_metadata_preserve 
Everybody uses the wrapper now.

Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33722>
 2025-02-26 15:19:53 +00:00
Alyssa Rosenzweig 24d088104f nir,nak: update comments referencing nir_metadata_preserve 
in prep for removing this method.

Signed-off-by: Alyssa Rosenzweig <alyssa@rosenzweig.io>
Reviewed-by: Georg Lehmann <dadschoorse@gmail.com>
Reviewed-by: Faith Ekstrand <faith.ekstrand@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33722>
 2025-02-26 15:19:53 +00:00