[dawn] Validate total invocations per workgroup against explicit subg…

…roup size

The total invocations per workgroup must be a multiple of subgroup
size when `chromium_experimental_subgroup_size_control` is enabled.

Bug: 463721943
Change-Id: I88b9bb34a7165f9839dd5911ddd6310bbbc362b0
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/284575
Reviewed-by: James Price <jrprice@google.com>
Reviewed-by: Loko Kung <lokokung@google.com>
Commit-Queue: Shao, Jiawei <jiawei.shao@intel.com>

[dawn] Many more pixel 10 cts expectations

Bug: 40238674,475503907
Change-Id: I84cd9ff6b9a02ff39e5959f041cf9b0fd0a2970f
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/285435
Bot-Commit: rubber-stamper@appspot.gserviceaccount.com <rubber-stamper@appspot.gserviceaccount.com>
Commit-Queue: Peter McNeeley <petermcneeley@google.com>

[D3D12] Implement `chromium_experimental_subgroup_size_control`

This patch implements `chromium_experimental_subgroup_size_control`
on D3D12 backend. This feature requires at least Shader Model 6.6
for the HLSL attribute `[WaveSize]`.

As D3D12 runtime only accept the values between
`D3D12_FEATURE_DATA_D3D12_OPTIONS1::waveLaneCountMin` and
`D3D12_FEATURE_DATA_D3D12_OPTIONS1::waveLaneCountMax` as `[WaveSize]`,
while currently on D3D12 backend we don't always use these two values
as `AdapterInfo::subgroupMinSize` or `AdapterInfo::subgroupMaxSize`,
we introduce two new limitations (`minExplicitComputeSubgroupSize` and
`maxExplicitComputeSubgroupSize`) to provide valid values for
`[WaveSize]`.

Bug: 463721943
Test: dawn_end2end_tests
Change-Id: Ib86a6e2034957b5e1ab5091203d862d981247363
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/284855
Reviewed-by: James Price <jrprice@google.com>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Shao, Jiawei <jiawei.shao@intel.com>

Remove stale WebGPU CTS expectations

Autogenerated CL from running:

//content/test/gpu/unexpected_pass_finder.py --project chrome-unexpected-pass-data --include-internal-builders --remove-stale-expectations --narrow-semi-stale-expectation-scope --num-samples 100 --suite webgpu_cts

Affected bugs for CL description:

R=rubber-stamper@appspot.gserviceaccount.com

Bug: 0000, 373485785, 407159836, 417770938, 418226002
Bug: 469120957, 469328920, 470115243
Fixed: 470127837
Change-Id: Ic876ee8751ad173cfc52de9872f7c608b9c82e36
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/284995
Commit-Queue: rubber-stamper@appspot.gserviceaccount.com <rubber-stamper@appspot.gserviceaccount.com>
Auto-Submit: chrome-automated-expectation@chops-service-accounts.iam.gserviceaccount.com <chrome-automated-expectation@chops-service-accounts.iam.gserviceaccount.com>
Bot-Commit: rubber-stamper@appspot.gserviceaccount.com <rubber-stamper@appspot.gserviceaccount.com>

Skip ResolveMultipleAttachmentInSeparatePasses if transient attachment

Bug: 475358683
Change-Id: I0fa82f19cdc3bdbfc77b45f3880442b9aa6fbe49
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/284635
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Fr <beaufort.francois@gmail.com>
Reviewed-by: Brandon Jones <bajones@chromium.org>

dawn/native: Add feature `chromium_experimental_subgroup_size_control`

This patch adds feature `chromium_experimental_subgroup_size_control`
which allows the use of WGSL attribute `@subgroup_size` and enables
the extension `subgroups` on the API side.

Bug: 463721943
Test: dawn_unittests
Change-Id: I1a0565e998dddece6781ebaf25c4803d60f10285
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/283795
Commit-Queue: Shao, Jiawei <jiawei.shao@intel.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>

Roll third_party/webgpu-cts/ 18183dc33..acb7d7902 (14 commits)

Manually rebased to remove duplicate expectations

Regenerated:
 - expectations.txt
 - compat-expectations.txt
 - ts_sources.txt
 - test_list.txt
 - resource_files.txt
 - webtest .html files

https://chromium.googlesource.com/external/github.com/gpuweb/cts/+log/18183dc33dcc..acb7d7902a1d
 - acb7d7 Compat: getCurrentTexture and textureBindingViewDimension (#4553)
 - f1aca4 Tweaks to case count reports on PRs (#4551)
 - 8f6f73 PR-comment automation works now. Fix nits
 - 279abe Attempt to fix PR-comment automation (3)
 - deab3a Attempt to fix PR-comment automation (2)
 - 8c90ec Attempt to fix PR-comment automation
 - 4ed1a4 Generate a case-count report on pull requests (#4548)
 - 77591d Cleanup some texture checks (#4544)
 - 2ded3f Test reflection of GPUTexture.textureBindingViewDimension. (#4537)
 - 19f59f Compat: Add copy T2T, T2B, B2T tests for textureBindingViewDimesion (#4542)
 - 41c23d `maxInterStageShaderVariables`: decrement max fragment input outputs per fragment input built-in (#4539)
 - d5a131 Minor fix for operator desc (#4536)
 - e0c2ca Bump @webgpu/types to 0.1.68 (#4534)
 - 5c24db Require depthOrArrayLayers=1 and mipLevelCount=1 for transient textures (#4533)

Created with './tools/run cts roll'

Cq-Include-Trybots: luci.chromium.try:android-dawn-arm-rel,android-dawn-arm64-rel,dawn-try-linux-x64-intel-uhd770-rel,dawn-try-mac-amd-exp,dawn-try-mac-arm64-m2-exp,dawn-try-mac-intel-exp,dawn-try-win-x64-intel-uhd770-rel,dawn-try-win10-x86-rel,linux-dawn-rel,mac-arm64-dawn-rel,mac-dawn-rel,win-dawn-rel
Include-Ci-Only-Tests: true
Change-Id: Ic0cef0eb0c740d124ed5a2c26c55827e9c3e98f6
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/283957
Commit-Queue: David Neto <dneto@google.com>
Reviewed-by: David Neto <dneto@google.com>

[dawn] Make the Instance simply RefCounted.

- Before, the Instance needed to be externally refcounted because
  there wasn't a way to get the Instance once the last external
  reference was dropped, hence it could cause any WaitAnyOnly or
  AllowProcessEvent events to be leaked. Since then, we have
  added a `GetInstance()` API on the Adapter which means that as
  long as we have a Device (which has `GetAdapter()`) on it, the
  instance can be reached. As a result, we only need to clear all
  events when the last device reference is dropped. This means
  that the Instance can just be a simple refcounted object now.
- This is also necessary because the `WaitAny()` and
  `ProcessEvents()` APIs implicitly assume that if they are called,
  that the Instance is externally alive. However, because users
  can drop the last external reference, then get a new one when
  they need it via the Device/Adapter, some teardown code in
  Chromium could cause a crash by calling those APIs on an
  Instance retrieved via the getter APIs after the original
  external reference was dropped.

Bug: 467732049, 474662276
Change-Id: Ib04c6243538983c2ba4c3060beb23487da66a499
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/283515
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
Reviewed-by: Sunny Sachanandani <sunnyps@chromium.org>
Auto-Submit: Loko Kung <lokokung@google.com>
Commit-Queue: Kai Ninomiya <kainino@chromium.org>

Add ConstExprIf test

Add test that the result is the result of the executed block, not the
original conditional result.

Change-Id: Ia20f5c524053c82588fdad2900a3a9e08beee816
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/284035
Commit-Queue: dan sinclair <dsinclair@chromium.org>
Reviewed-by: James Price <jrprice@google.com>

Native: Add SetOwnershipForMemoryDump for textures

This CL add a Dawn Native API SetOwnershipForMemoryDump for textures to
allow setting and resetting owner GUID of TextureBase and get reported
during memory dumping. This can help us better understand the texture
memory usage for shared images in Chromium.

Issue: 397989391
Change-Id: I5cfca178b9474ac62ec6d783dd879530eec63d38
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/280095
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Loko Kung <lokokung@google.com>
Commit-Queue: Zhaoming Jiang <zhaoming.jiang@microsoft.com>