This patch adds an option to declare the workgroup variables with zero
initializer in Build() instead of transform::ZeroInitWorkgroupMemory
in Sanitize(). This option will be enabled when the Vulkan extension
VK_KHR_zero_initialize_workgroup_memory is enabled on the API side.
BUG=dawn:1302
Change-Id: Ia580df98ec161ec6f2d3099a01dbedb8bf848bf2
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/82580
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@chromium.org>
Commit-Queue: Ben Clayton <bclayton@chromium.org>
Fixes vulkan validation layer assertions on Linux in CTS reftests:
canvas_complex_bgra8unorm_copy.https.html
canvas_size_different_with_back_buffer_size.https.html
Bug: chromium:1083478
Change-Id: I36580184d15fe36423e5f759eaf959191c6b571c
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/82480
Reviewed-by: Loko Kung <lokokung@google.com>
Commit-Queue: Austin Eng <enga@chromium.org>
In GLSL, runtime-sized arrays are only valid in interface blocks, not
in structs. The existing code was attempting to avoid emitting structs
containing runtime-sized arrays but was confused by type aliases in
the AST resulting in arrays being missed.
The fix is to do the work on the semantic types instead, where type
aliases have been resolved.
Bug: tint:1339
Change-Id: I8c305ee9bddd75f975dd13f1d19d623d71410693
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/82360
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Commit-Queue: Stephen White <senorblanco@chromium.org>
Unlike other texture functions in GLSL, textureGather() and
textureGatherOffset() do not expect the refZ value to be appended to
the texture coordinates. It is passed as a regular argument. So append
refZ to coordinates by default, and pass as a regular parameter only
for the gather functions.
Bug: tint:1459
Change-Id: Iad1255be3de5915aeff4adb9054479b9e92c45cb
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/82340
Reviewed-by: Ben Clayton <bclayton@google.com>
Commit-Queue: Stephen White <senorblanco@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
Multiple bugs here:
1) Like texture(), GLSL's textureLod() on depth textures returns a
scalar, and does not need to be swizzled. So set glsl_ret_width to
1 in that case.
2) Like texture(), GLSL's textureLod() always requires a Dref
parameter, so append a zero if not present.
3) GLSL's "lod" parameter to textureLod() is always a float, unlike
WGSL's textureSampleLevel() which is an i32 for depth textures,
so cast it.
Along the way, I discovered that textureLod() is not supported on
samplerCubeShadow or sampler2DArrayShadow (even on Desktop GL). So some
tests will never pass. Logged as https://crbug.com/dawn/1313
Bug: tint:1456
Change-Id: If67d8d288704142278d7a4e52b46e8010776f381
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/82300
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Commit-Queue: Stephen White <senorblanco@chromium.org>
The Mesa Gallium llvmpipe driver is a software rasterizer that uses LLVM
to do runtime code generation. It seems to put incorrect value for
maxFragmentCombinedOutputResources, like we see on desktop GPUs.
Some relative values are as below.
vkLimits.maxFragmentCombinedOutputResources: 8
kMaxColorAttachments: 8
baseLimits.v1.maxStorageTexturesPerShaderStage: 4
baseLimits.v1.maxStorageBuffersPerShaderStage: 8
This CL will skip the check of this limit on llvmpipe, like we did for
desktop GPUs.
BUG=dawn:1311
Change-Id: I1698f6e26dd62b56f3819980fa2dcc986c9c5ba3
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/81941
Reviewed-by: Jiawei Shao <jiawei.shao@intel.com>
Reviewed-by: Hao Li <hao.x.li@intel.com>
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Austin Eng <enga@chromium.org>
This patch fixes a use-after-free issue in Create*PipelineAsyncTasks
that when pipeline->Initialize() returns error, the pipeline object
will be deleted, while we still attempt to call its member function
after it is deleted.
BUG=dawn:1310
TEST=dawn_unittests
Change-Id: I57d5ca98d6c97b14df1d7c3bf2941c9cc87adeff
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/81800
Reviewed-by: Loko Kung <lokokung@google.com>
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Austin Eng <enga@chromium.org>
CopyTextureForBrowser() will support all uploading paths for WebGPU
CopyExternalImageToTexture(). The validation checks for texture states
should report accurate errors to meet CopyExternalImageToTexture() validate
rules.
Bug: dawn:1306
Change-Id: Ie3b25ec82246d53e6c82968b5dc2f8a253c560c1
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/81240
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Shaobo Yan <shaobo.yan@intel.com>
Although the error rate of timestamp tick->ns conversion is very small
(3e-5), some profiling scenarios, such as CPU/GPU timestamp calibration
on Windows, require absolutely accurate timestamps.
Add new toggle to resolve timestamps to ticks for those cases where zero
error is required.
Add an end2end test for GPU timestamp calibration on D3D12 backend.
Disable timestamp period calculation on Device when the
DisableTimestampQueryConversion is enabled.
Bug: dawn:1305
Change-Id: I31ee6b4c1686d5dd2ac29ccb0bd398e650481c26
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/81023
Reviewed-by: Shrek Shao <shrekshao@google.com>
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Austin Eng <enga@chromium.org>
This patch adds the validations on the format of a resolve target
according to the latest updates in WebGPU SPEC. WebGPU SPEC only
supports the below texture formats being used as the resolve target
of a render pass encoder:
- R8Unorm
- RG8Unorm
- RGBA8Unorm
- RGBA8UnormSrgb
- BGRA8Unorm
- BGRA8UnormSrgb
- R16Float
- RG16Float
- RGBA16Float
- RGB10A2Unorm
BUG=dawn:1244
TEST=dawn_unittests
Change-Id: I16f9c3984f4ffb4641f4f43ecdb3cc76be6a562a
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/81080
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Brandon Jones <bajones@chromium.org>
Commit-Queue: Jiawei Shao <jiawei.shao@intel.com>
Renames multiple attachment clear values to match their new names in the
spec, deprecating the old ones.
- GPURenderPassColorAttachment.clearColor -> clearValue
- GPURenderPassDepthStencilAttachment.clearDepth -> depthClearValue
- GPURenderPassDepthStencilAttachment.clearStencil -> stencilClearValue
Additionally, the old names are marked as deprecated with appropriate
warnings if they are used during the deprecation period.
Bug: dawn:1269
Change-Id: I6649184d65578118942c1f51a41f350719665272
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/80941
Reviewed-by: Shrek Shao <shrekshao@google.com>
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Brandon Jones <bajones@chromium.org>
Use `static constexpr bool` instead of `std::integral_constant` for `IsTypeOrDerived`. This is often cleaner to use.
Add `IsCastable`, a helper for determining if all the template types derive from `CastableBase`.
Add `CastableCommonBase`, some template magic for determinine the most derived, common base type for all castable types.
Change-Id: Ia3d33548424750f8260f518ecd63d39949e4a826
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/81105
Kokoro: Kokoro <noreply+kokoro@google.com>
Auto-Submit: Ben Clayton <bclayton@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Antonio Maiorano <amaiorano@google.com>
The fix involves unifying the dependency logic of DependencyScanner, which also cleans up the code quite a bit.
Also:
* Correctly resolve parameter type before declaring the parameter variable.
* Fix bonkers typos in utils::Lookup.
Fixed: tint:1442
Change-Id: I77b548e148b461f87ec10e8272c398f6fee297bd
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/81102
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: James Price <jrprice@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
