This workaround replaces the toggle that forbids creating mipmapped
R8Unorm and RG8Unorm textures on Metal. Instead it adds a Metal render
pass workaround that renders to a temporary texture under certain
conditions, then copies back to the correct mip level.
The texture lazy-clearing code is also changed to use the helper that
does workarounds for Metal render passes.
A test is added that triggers the issue with a small amount of code, but
more extensive coverage is left to the CTS. For example texture_zero
tests caught multiple issues during the creation of this CL.
Bug: dawn:1071
Change-Id: I7ef1151524e71e5a9a8e8f5205d9b554bee438b3
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/87864
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
It is failing for a single format when trying to copy to a buffer
because the depth-specific format for depth23unorm-stencil8 buffer is
depth24plus which is not copyable.
Bug: dawn:1388
Change-Id: I034145bf55d42403ded23b1b9acf974d85770d42
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/87701
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
intel-0x3e92 was not added, but intel-0x5912 was - leading
to flaky CQ rejection depending on what bot happened to run
the task.
Generalize this expectation since it fails on all platforms
anyway.
Bug: dawn:1387
Change-Id: Iccd4806de456a817edf538041445ee2ac93099e7
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/87675
Bot-Commit: rubber-stamper@appspot.gserviceaccount.com <rubber-stamper@appspot.gserviceaccount.com>
Commit-Queue: Austin Eng <enga@chromium.org>
Reviewed-by: Ben Clayton <bclayton@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Update expectations and ts_sources
a06666e000..45d2ccd250
- 45d2cc Fixed TS errors
- 5eacfb Take new createView() dimension default logic into account
- cbdcbd Explicitly destroy devices created by requestDevice_limits
Created with './tools/run cts roll'
Change-Id: I701bc9e6fa6b1996b1ed40b08766cc9dedaf20d8
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/87621
Reviewed-by: Brandon Jones <bajones@chromium.org>
Commit-Queue: Ben Clayton <bclayton@google.com>
• Remove a number of expectations that are no longer needed.
• Change a number of expectation tags to be more specific.
• Sort the expectations per chunk.
• Add a few more wide-wildcard 'Failure' rules for tests that commonly flake.
Bug: dawn:1342
Change-Id: I0e23cd546f7ab54a5720e24c52e3782b5df6b99f
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/87520
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Ben Clayton <bclayton@google.com>
Adds support for non-zero first/baseVertex on Draw[Indexed]Indirect by
duplicating the first/baseVertex indirect parameter onto a root
constant in the indirect buffer.
Change-Id: I280149065179806d3e57b07f1a396f9e2e4e8fcb
Bug: dawn:548
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/84240
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Enrico Galli <enrico.galli@intel.com>
Auto-Submit: Enrico Galli <enrico.galli@intel.com>
Fixed: dawn:1370
Change-Id: I395b3e6e2506608ca4aefdbcef18443b90ca75ea
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/87302
Reviewed-by: Brandon Jones <bajones@chromium.org>
Reviewed-by: Loko Kung <lokokung@google.com>
Commit-Queue: Austin Eng <enga@chromium.org>
These are known crashes from the old WebGPUExpectations. Running
them is a waste of resources.
Bug: chromium:1309194, chromium:1234041, chromium:1236130, chromium:1315987
Change-Id: I7bbf70d0bc5abd09ebf47e2e65521d4a4fa5363e
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/86844
Bot-Commit: rubber-stamper@appspot.gserviceaccount.com <rubber-stamper@appspot.gserviceaccount.com>
Commit-Queue: Austin Eng <enga@chromium.org>
dcba3ac6b8..5e83d8349b
$ git log dcba3ac6b..5e83d8349 --date=short --no-merges --format='%ad %ae %s'
2022-04-13 cwallez Reduce the length of subresource test names (#1164)
2022-04-13 cwallez zero_init: add constant zero to prevent driver unrolling loops. (#1163)
2022-04-13 jiawei.shao Add validation,resource_usages,texture,in_render_misc:* - Part II (#1161)
2022-04-13 jiawei.shao Add validation,resource_usages,texture,in_render_misc:* - Part I (#1157)
2022-04-11 enga Test view format reinterpretation (#1126)
2022-04-11 litherum Only 2D textures can be marked as render targets (#1155)
2022-04-08 cwallez operation,image_copy: Add support for depth16unorm (#1154)
2022-04-08 cwallez image_copy,mip_levels: reduce the mip level count. (#1151)
2022-04-08 cwallez depth_compare_func: Don't use 0.5 as depth16unorm rounds either way (#1152)
2022-04-08 cwallez Fix tests trying to copy aspects of depth-stencil textures. (#1146)
2022-04-08 cwallez depth_clip_clamp: only set stencilLoadOp and friends when needed. (#1153)
2022-04-08 cwallez Fix validation errors with depth/StencilLoad/StoreOp set when not allowed. (#1145)
2022-04-08 enga Shorten test and param lengths in resource_usages,texture,in_render_common:* (#1144)
2022-04-08 jiawei.shao Add validation,resource_usages,texture,in_render_common:* - Part V (#1150)
2022-04-06 cwallez val,buffer,create: Buffer sizes don't have to be multiple of 4. (#1140)
2022-04-06 cwallez image_copy: Ensure bufferOffset % 4 == 0 for depth-stencil. (#1139)
2022-04-06 jiawei.shao Add validation,resource_usages,texture,in_render_common:* - Part IV (#1132)
2022-04-05 enga Fix infinite recursion in CopyExternalImageToTexture.spec.ts (#1143)
2022-04-05 cwallez Use globalThis instead of self. (#1141)
2022-04-05 cwallez val/copyT2T,format_compat: Don't assume compressed is always 4x4 (#1134)
2022-04-05 cwallez Fix DSAttachment load/store ops being specified when disallowed. (#1136)
2022-04-05 cwallez Fix test of BGL binding limits to correctly check limits (#1137)
2022-04-04 cwallez Fix a validation error caused by WriteTexture bytesPerRow: t.width (#1133)
2022-04-04 cwallez Add unicode tests for entryPoint matching. (#1131)
2022-04-04 cwallez writeBuffer: an unaligned bufferOffset is a validation error. (#1129)
Created with:
roll-dep third_party/webgpu-cts
Change-Id: I048afcaf3d33bcd67401412bf235181013ebf6d2
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/86533
Auto-Submit: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Austin Eng <enga@chromium.org>
So we can begin running the CTS on SwiftShader and add targeted
suppressions when necessary.
Bug: chromium:1315703
Change-Id: Ic68e43ad0d79830b3ef5559ea1de0133d48ea43c
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/86626
Commit-Queue: Austin Eng <enga@chromium.org>
Auto-Submit: Austin Eng <enga@chromium.org>
Reviewed-by: Ben Clayton <bclayton@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Roll CTS to dcba3ac6b819abb0920e589ca1809ae5b5f87119
Update ts_sources.txt
Regenerate expectations
Change-Id: Ifa0cae3fa1abf9da9e4fbd63acf9cefe2683f50d
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/85560
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Ben Clayton <bclayton@google.com>
Corentin Wallez
Ben Clayton
Austin Eng
Yunchao He
Enrico Galli
Jiawei Shao
Shrek Shao