Commit graph

187891 commits

Author SHA1 Message Date
Rhys Perry
35f9318cee ac/llvm: implement mqsad_4x8 and shfr
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/26251>
2024-04-05 11:01:39 +00:00
Rhys Perry
6b301eae36 aco: implement mqsad_4x8 and shfr
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/26251>
2024-04-05 11:01:39 +00:00
Rhys Perry
08903bbe89 nir: add mqsad_4x8, shfr and nir_opt_mqsad
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/26251>
2024-04-05 11:01:39 +00:00
Georg Lehmann
be93e38152 aco/vn: remove instruction hash templates
This reuses the instruction data size helper used for creating instruction.
It also removes the hashing of the aco::spans used for operands and definitions,
which was redundant with hash finalization.

Reviewed-by: Rhys Perry <pendingchaos02@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28527>
2024-04-05 10:23:16 +00:00
Lionel Landwerlin
ea84b36592 anv: fix incorrect blorp dynamic state heap usage
Found with valgrind :

==253563== Invalid free() / delete / delete[] / realloc()
==253563==    at 0x6EEBB88: anv_state_pool_free (anv_allocator.c:962)
==253563==    by 0x6EFB563: anv_device_finish_blorp (anv_blorp.c:143)
==253563==    by 0x6F204F8: anv_DestroyDevice (anv_device.c:4063)
==253563==    by 0x6DE1CD7: loader_layer_destroy_device (loader.c:4387)
==253563==    by 0x6DF1D5E: vkDestroyDevice (trampoline.c:1025)
==253563==    by 0x407C54: vk::refdetails::Deleter<vk::VkDevice_s*>::operator()(vk::VkDevice_s*) const (vkRef.hpp:131)
==253563==    by 0x42C016: vk::refdetails::RefBase<vk::VkDevice_s*>::reset() (vkRef.hpp:303)
==253563==    by 0x40B385: vk::refdetails::RefBase<vk::VkDevice_s*>::~RefBase() (vkRef.hpp:296)
==253563==    by 0x40A95D: vk::refdetails::Unique<vk::VkDevice_s*>::~Unique() (vkRef.hpp:376)
==253563==    by 0x402501: vkt::DefaultDevice::~DefaultDevice() (vktTestCase.cpp:658)
==253563==    by 0x444807: de::DefaultDeleter<vkt::DefaultDevice>::operator()(vkt::DefaultDevice*) const (deDefs.hpp:112)
==253563==    by 0x43922D: de::details::UniqueBase<vkt::DefaultDevice, de::DefaultDeleter<vkt::DefaultDevice> >::reset() (deUniquePtr.hpp:90)
==253563==  Address 0xd3df000 is 0 bytes inside a block of size 272 client-defined
==253563==    at 0x6EEBA0B: anv_state_pool_alloc (anv_allocator.c:940)
==253563==    by 0x6EFA610: anv_state_pool_emit_data (anv_private.h:852)
==253563==    by 0x6EFB206: upload_dynamic_state (anv_blorp.c:106)
==253563==    by 0x6FC8C31: blorp_init_dynamic_states (blorp_genX_exec_brw.h:2211)
==253563==    by 0x6FCD02D: gfx9_blorp_init_dynamic_states (genX_blorp_exec.c:507)
==253563==    by 0x6EFB47E: anv_device_init_blorp (anv_blorp.c:129)
==253563==    by 0x6F1FBBA: anv_CreateDevice (anv_device.c:3908)
==253563==    by 0x840B669: vk_tramp_CreateDevice (vk_dispatch_trampolines.c:78)
==253563==    by 0x6DE663A: terminator_CreateDevice (loader.c:5836)
==253563==    by 0x6DE3E39: loader_create_device_chain (loader.c:4940)
==253563==    by 0x6DE1AC5: loader_layer_create_device (loader.c:4320)
==253563==    by 0x6DF1CBF: vkCreateDevice (trampoline.c:1005)

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Fixes: fe1baa6481 ("anv: reduce blorp dynamic state emissions")
Reviewed-by: Ivan Briano <ivan.briano@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28578>
2024-04-05 09:50:41 +00:00
Corentin Noël
0624130933 zink: Remove ctx from zink_gfx_program
It is already available in the base structure.

Signed-off-by: Corentin Noël <corentin.noel@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28567>
2024-04-05 08:01:01 +00:00
Corentin Noël
b0d680f54c zink: Removed unused push_valid member
This member is only set but never used.

Signed-off-by: Corentin Noël <corentin.noel@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28567>
2024-04-05 08:01:01 +00:00
Gert Wollny
6a0f3b1aa5 zink: another fence that is better off as a batch state
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28560>
2024-04-05 07:29:04 +00:00
Gert Wollny
d862cb0024 zink: ctx->last_fence really wants to be a batch_state, so accomodate it
Signed-off-by: Gert Wollny <gert.wollny@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28560>
2024-04-05 07:29:03 +00:00
Jordan Justen
98f3d072b4 intel/dev: Add 0x56be and 0x56bf DG2 PCI IDs
Ref: bspec 44477
Backport-to: 24.0
Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28584>
2024-04-05 06:36:44 +00:00
Lionel Landwerlin
9b0f028c7e anv: update protection fault property
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Fixes: 794b0496e9 ("anv: enable protected memory")
Reviewed-by: Rohan Garg <rohan.garg@intel.com>
Tested-by: Mark Janes <markjanes@swizzler.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/26540>
2024-04-05 09:07:21 +03:00
Lionel Landwerlin
d2e490dc4d anv: disable generated draws in protected command buffers
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Fixes: 794b0496e9 ("anv: enable protected memory")
Reviewed-by: Rohan Garg <rohan.garg@intel.com>
Tested-by: Mark Janes <markjanes@swizzler.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/26540>
2024-04-05 09:07:21 +03:00
Lionel Landwerlin
034a1cdb58 anv: disable protected content around surface state copies
Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Fixes: 794b0496e9 ("anv: enable protected memory")
Reviewed-by: Rohan Garg <rohan.garg@intel.com>
Tested-by: Mark Janes <markjanes@swizzler.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/26540>
2024-04-05 09:07:21 +03:00
Lionel Landwerlin
27a3771227 anv: pull surface state copies for secondary in one loop
It'll be easier for the next commit.

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Tested-by: Mark Janes <markjanes@swizzler.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/26540>
2024-04-05 09:07:21 +03:00
Lionel Landwerlin
07bf480856 anv: fix protected memory allocations
Using the wrong flag field...

Signed-off-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
Fixes: 5f2c77a10a ("anv: handle protected memory allocation")
Reviewed-by: Rohan Garg <rohan.garg@intel.com>
Tested-by: Mark Janes <markjanes@swizzler.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/26540>
2024-04-05 09:07:21 +03:00
Ian Romanick
0e817ba548 intel/brw/xe2+: Implement Wa 22016140776
HF sources to math instructions cannot be scalar. This is very similar
to an old Gfx6 restriction on POW, so let's fix it in a similar way.

As an extra bit of saftey, lower any occurances that might slip through
in brw_fs_lower_regioning.

The primary change is to prevent copy propagation from violating the
restriction. With that change, nothing should be able to generate these
invalid source strides. The modification to fs_visitor::validate should
detect potential problems sooner rather than later.

Previous attempts to implement this Wa when emitting the math
instruction (in brw_eu_emit.c gfx6_math) didn't work for several
reasons. The lowering happens after the SWSB pass, so the scoreboarding
was incorrect (thanks to Curro for finding that). In addition, the
lowering happens after register allocation, so it's impossible to
allocate a non-scalar register to expand the scalar value.

Fixes 113 tests in the dEQP-VK.spirv_assembly.* group on LNL.

v2: Add changes to brw_fs_lower_regioning. Suggested by Curro.

Reviewed-by: Francisco Jerez <currojerez@riseup.net>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28480>
2024-04-04 21:04:09 -07:00
Jordan Justen
50c7d25a9e intel/dev/mesa_defs.json: Add LNL WA entries
Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Acked-by: Ian Romanick <ian.d.romanick@intel.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28480>
2024-04-04 21:03:51 -07:00
Jesse Natalie
c891a384e7 dzn: Don't copy app indirect args if we don't need to
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28383>
2024-04-05 00:25:41 +00:00
Jesse Natalie
db16dcf0f9 dzn: Add a hash table of command signatures with non-default strides
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28383>
2024-04-05 00:25:41 +00:00
Jesse Natalie
d7bd87aa27 dzn: Rework indirect drawing keys for shaders and command signatures
When we don't need emulation of first vertex, base instance, draw ID
sysvals, or triangle fans, we can have very simple command signatures
and indirect arg buffer generation shaders. The next step is to handle
the case where everything can be supported straight from the app's
buffer.

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28383>
2024-04-05 00:25:41 +00:00
Jesse Natalie
ccf439629f dzn: Update pipeline cache params to take all options into account
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28383>
2024-04-05 00:25:41 +00:00
Jesse Natalie
4de88025cc dzn: Understand whether first-vertex and base-instance are needed for a pipeline
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28383>
2024-04-05 00:25:41 +00:00
Jesse Natalie
fbd4e5d8aa dzn: Query options21
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28383>
2024-04-05 00:25:41 +00:00
Jesse Natalie
ed465bf738 dzn: Delete dzn structs for indirect draw args and use D3D ones
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28383>
2024-04-05 00:25:41 +00:00
Jesse Natalie
f6c3d73700 spirv2dxil: Output more specific metadata for whether draw sysvals are needed
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28383>
2024-04-05 00:25:41 +00:00
Jesse Natalie
30b1a6750c spirv2dxil: Support passing first vertex / base instance to DXIL backend
These values now have 3 modes of operation so switch from a bool to an enum:
1. Zero
2. Native DXIL sysvals
3. "Runtime data" via a constant buffer or root constants

Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28383>
2024-04-05 00:25:41 +00:00
Jesse Natalie
607720151f microsoft/compiler: Handle base vertex/instance sysvals as DXIL intrinsics
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28383>
2024-04-05 00:25:41 +00:00
Jesse Natalie
349c4d84c0 ci/windows: Bump Agility SDK to 1.613.2 for ExecuteIndirect validation fix
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28383>
2024-04-05 00:25:41 +00:00
Ian Romanick
0b67d3d909 intel/elk: Delete stray nir_opt_dce
No shader-db changes on any Intel platform.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28136>
2024-04-04 23:42:28 +00:00
Ian Romanick
24cdbbdaa2 intel/brw: Delete stray nir_opt_dce
No shader-db or fossil-db changes on any Intel platform.

Fixes: f76f4be301 ("intel/compiler: move gen5 final pass to actually be final pass")
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28136>
2024-04-04 23:42:27 +00:00
Ian Romanick
44fb57b827 intel/elk: Don't call nir_opt_remove_phis before nir_convert_from_ssa
shader-db:

All platforms had similar results. (Ivy Bridge shown)
total instructions in shared programs: 15831424 -> 15831637 (<.01%)
instructions in affected programs: 38880 -> 39093 (0.55%)
helped: 0 / HURT: 179

total cycles in shared programs: 432140353 -> 432170199 (<.01%)
cycles in affected programs: 11798080 -> 11827926 (0.25%)
helped: 77 / HURT: 123

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28136>
2024-04-04 23:42:27 +00:00
Ian Romanick
6377e8fd29 intel/brw: Don't call nir_opt_remove_phis before nir_convert_from_ssa
Per discussion in #10727, removing phis breaks LCSSA form which in turn
invalidates divergence analysis.

shader-db:

All Skylake and newer platforms had similar results. (Ice Lake shown)
total instructions in shared programs: 20299612 -> 20299695 (<.01%)
instructions in affected programs: 20829 -> 20912 (0.40%)
helped: 6 / HURT: 13

total cycles in shared programs: 842149085 -> 842148399 (<.01%)
cycles in affected programs: 15146222 -> 15145536 (<.01%)
helped: 40 / HURT: 45

fossil-db:

All Intel platforms had similar results. (Ice Lake shown)
Totals:
Instrs: 165505077 -> 165505603 (+0.00%); split: -0.00%, +0.00%
Cycles: 15144183575 -> 15144235695 (+0.00%); split: -0.00%, +0.00%
Spill count: 45213 -> 45220 (+0.02%)
Fill count: 74166 -> 74184 (+0.02%)

Totals from 94 (0.01% of 656116) affected shaders:
Instrs: 263079 -> 263605 (+0.20%); split: -0.00%, +0.20%
Cycles: 28411487 -> 28463607 (+0.18%); split: -0.18%, +0.37%
Spill count: 3474 -> 3481 (+0.20%)
Fill count: 6713 -> 6731 (+0.27%)

Fixes: 6dbb5f1e07 ("intel/fs: rerun divergence analysis prior to convert_from_ssa")
Closes: #10727
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28136>
2024-04-04 23:42:27 +00:00
Ian Romanick
87101e7d83 intel/compiler: Ensure load_barycentric_at_sample and load_interpolated_input remain together
This previously worked by luck because we were incorrectly calling
nir_opt_remove_phis before calling nir_convert_from_ssa. See also #10727.

No shader-db or fossil-db changes on any Intel platform.

v2: Handle the load_interpolated_input and load_barycentric_at_sample as
separate passes. Based on discussion with Ken starting at
https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28136#note_2330424.

Fixes: 74a40cc4b6 ("intel/fs: move lower of non-uniform at_sample barycentric to NIR")
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28136>
2024-04-04 23:42:27 +00:00
Erico Nunes
49217c2547 ci: enable shader-db on lima
Run shader-db on CI with the lima drm_shim as done for other drivers.

Signed-off-by: Erico Nunes <nunes.erico@gmail.com>
Acked-by: Vasily Khoruzhick <anarsoul@gmail.com>
Reviewed-by: David Heidelberg <david.heidelberg@collabora.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/24357>
2024-04-04 22:48:10 +00:00
Samuel Pitoiset
69b911bc53 radv: remove radv_private.h
This file was a giant mess and I'm very happy to remove it entirely.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28568>
2024-04-04 21:57:46 +00:00
Samuel Pitoiset
367cf70a29 radv: move RADV_USE_WSI_PLATFORM define to radv_wsi.h
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28568>
2024-04-04 21:57:46 +00:00
Samuel Pitoiset
28eefbadeb radv: move CLOCK_MONOTONIC_RAW define to radv_physical_device.h
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28568>
2024-04-04 21:57:46 +00:00
Samuel Pitoiset
e8269c01cb radv: move RADV_API_VERSION to radv_instance.h
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28568>
2024-04-04 21:57:46 +00:00
Samuel Pitoiset
e25882352b radv: move RADV_SUPPORT_CALIBRATED_TIMESTAMPS to radv_physical_device.c
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28568>
2024-04-04 21:57:46 +00:00
Samuel Pitoiset
cc1526eeac radv: replace radv_is_aligned() by util_is_aligned()
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28568>
2024-04-04 21:57:46 +00:00
Samuel Pitoiset
81e3c46d06 util: add util_is_aligned()
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28568>
2024-04-04 21:57:46 +00:00
Samuel Pitoiset
5cc3258533 radv: replace radv_float_to_{u,s}fixed() by util_{un}signed_fixed()
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28568>
2024-04-04 21:57:46 +00:00
Samuel Pitoiset
6a27c5e9a8 radv: replace align_u32_npot() by ALIGN_NPOT
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28568>
2024-04-04 21:57:45 +00:00
Samuel Pitoiset
e75fdac533 radv: replace align_{u32,u64}() by align{64}()
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28568>
2024-04-04 21:57:45 +00:00
Samuel Pitoiset
ba153fc06b radv: replace radv_minify() by u_minify()
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28568>
2024-04-04 21:57:45 +00:00
Samuel Pitoiset
799e99fe21 radv: rename remaining phys_dev occurrences to pdev
Missed those.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28568>
2024-04-04 21:57:45 +00:00
Samuel Pitoiset
b409936e97 radv: add missing endif comment for some headers
For consistency.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28568>
2024-04-04 21:57:45 +00:00
Samuel Pitoiset
0388df3d08 radv: replace RADV_FROM_HANDLE by VK_FROM_HANDLE
It was exactly the same thing.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28568>
2024-04-04 21:57:45 +00:00
Samuel Pitoiset
67ac6e75c6 radv: remove remaining forward declarations and comments in radv_private.h
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28568>
2024-04-04 21:57:45 +00:00
Samuel Pitoiset
8ade8f28a8 radv: remove pre-declarations needed for WSI entrypoints
Probably a very old thing.

Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Part-of: <https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/28568>
2024-04-04 21:57:45 +00:00