Async functions should fail by rejecting the callback instead of
generating a device error. This fixes a leak in the wire where
the allocation for the callback was never delete since it was
never called.
Fixed: chromium:1181627
Change-Id: I840073c1d1b5f1401aa8ed29d3c8f0e1e4fefd35
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/42540
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Austin Eng <enga@chromium.org>
Currently on D3D12 backend a device lost will occur when we do a T2T
copy under the following situations:
1. with Depth32Float
2. only copy one row
3. bufferCopyOffset == 256
This is because in current implementation it is possible for us to do
a copy with D3D12_SUBRESOURCE_FOOTPRINT.Depth > 1, which is not allowed
with DXGI_FORMAT_D32_FLOAT because this format is not supported to be
used as 3D textures.
This patch adds a regression test for this bug and we will fix it
later.
BUG=dawn:693
TEST=dawn_end2end_tests
Change-Id: Ib6fe70988b5b217d5f14d3f32999b3841e5d23b9
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/42600
Commit-Queue: Austin Eng <enga@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Brandon Jones <brandon1.jones@intel.com>
Reviewed-by: Austin Eng <enga@chromium.org>
The are replaced by Queue::OnSubmittedWorkDone. Only Queue::CreateFence
is deprecated since all other fence-related calls require a fence to
work.
Also ports a number of uses of fences in test harness to use
Queue::OnSubmittedWorkDone instead.
Bug: chromium:1177476
Change-Id: I479415f72b08158a3484013e00db8facd11e6f33
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/42660
Auto-Submit: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Brandon Jones <bajones@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Brandon Jones <bajones@chromium.org>
It was being taken from the pre-transformed program, not the post-transformed program.
If we did get a transform error, the string would be empty.
Also change the signature of RunTransforms() to take a Transform* instead of a Manager*. There's nothing special about Manager anymore - it is just a transform that acts as a group of other transforms.
Change-Id: I4ea6cb022b5967b3c6b8c628517727dc3da3be8e
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/42640
Auto-Submit: Ben Clayton <bclayton@google.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Ben Clayton <bclayton@google.com>
Adds dawn_native::ResetDeviceInternal, which allows us to destroy and
create a new ID3D12Device. The device should be reset after every test
when GPU-based validation is enabled in order to prevent GBV objects
from using a significant amount of memory over time.
Bug: dawn:623
Change-Id: I654d093d993ab0198c6c240bd0f3f843d2762680
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/41340
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Rafael Cintron <rafael.cintron@microsoft.com>
Commit-Queue: Brandon Jones <brandon1.jones@intel.com>
This reverts commit 9fdbb74072.
Reason for revert: The SPIRV-Cross GN files have a bug which is fixed upstream, but Skia is on an old version of SPIRV-Cross so they don't have the fix. So, the build fails when rolling Dawn into Skia complaining about unknown `-fno-exceptions` on the windows bots. Need Skia to either move to vulkan-deps, or roll their spirv-cross forward before this can re-land.
Original change's description:
> Use upstream SPIRV-Cross GN files.
>
> Previously spirv-cross didn't have its own sources files but this
> changed recently so we should use them. This will ultimately allow
> sharing targets between multiple components in Chromium.
>
> Bug: chromium:1179277
> Change-Id: Ib4bb1884b9ba9c4c3804e96d8adbb8905c60c9a3
> Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/41725
> Auto-Submit: Corentin Wallez <cwallez@chromium.org>
> Reviewed-by: Peng Huang <penghuang@chromium.org>
> Reviewed-by: Austin Eng <enga@chromium.org>
> Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Bug: chromium:1179277
Change-Id: If1003bafa7b35f502c08b2dab91dd3d416aab077
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/42420
Commit-Queue: dan sinclair <dsinclair@chromium.org>
Commit-Queue: Austin Eng <enga@chromium.org>
Auto-Submit: dan sinclair <dsinclair@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
The new tint::transform::[Spirv,Hlsl,Msl] transforms sanitize the tint::Program for the given backend.
The tint::transform::Spirv transform handles edge cases for sample masks (crbug.com/tint/372). We can now enable these tests.
Rework dawn_native::[opengl,vulkan]::ShaderModule::Initialize() so that transforms are applied *before* calling ShaderModuleBase::InitializeBase(). This is done as InitializeBase() wants to validate the SPIR-V, which requires the pre-processing of tint::transform::Spirv.
InitializeBase() also performs shader reflection which needs to be performed on the post-transformed program for the information to be correct.
Bug: tint:372
Change-Id: I4c96ce89b6ae286972549d8c7efe59e77c469063
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/42223
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Ben Clayton <bclayton@google.com>
SwANGLE does not support GL_EXT_texture_compression_s3tc.
However, it does support GL_EXT_texture_compression_dxt1,
GL_ANGLE_texture_compression_dxt3 and GL_ANGLE_texture_compression_dxt5,
which should be sufficient for Dawn's purposes.
Note that sRGB formats may be problematic for OpenGL ES in general,
but ANGLE does support them, so we can satisfy the requirements for
Dawn's texture_compression_bc and exercise the codepaths on ES.
glPixelStorei is not only unsupported for CompressedTexSubImage*D()
but causes asserts in ANGLE and validation errors elsewhere. The fix
is to scope the glPixelStorei() calls to the non-compressed codepath.
Bug: dawn:580
Change-Id: I68fa019eda2aee37a097e697cfc87bcdc28c1f12
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/42120
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Stephen White <senorblanco@chromium.org>
In upstream WebGPU, error scopes do not wait for queue operations or
async operations like create*PipelineAsync or mapAsync. This simplifies
the implementation so we don't need to track error scopes by parent
pointers but can instead have a simple stack.
Bug: dawn:22, chromium:1177107
Change-Id: Ic7344cbd96e257cbabc0f414934a5e42a4020a13
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/41980
Commit-Queue: Austin Eng <enga@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Previously spirv-cross didn't have its own sources files but this
changed recently so we should use them. This will ultimately allow
sharing targets between multiple components in Chromium.
Bug: chromium:1179277
Change-Id: Ib4bb1884b9ba9c4c3804e96d8adbb8905c60c9a3
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/41725
Auto-Submit: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Peng Huang <penghuang@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
This makes the primitives in the serialized wire protocol
the same across platforms and architectures which is better
for both fuzzing and remoting Dawn.
Commands that used size_t are updated to use uint64_t, and
the server-side implementation checks if conversion to
size_t would narrow.
Bug: dawn:680
Change-Id: Icef9dc11a72699685ed7191c34d6a922b652c887
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/41582
Commit-Queue: Austin Eng <enga@chromium.org>
Reviewed-by: Stephen White <senorblanco@chromium.org>
BufferConsumer wraps a buffer pointer and size and exposes a
limited number of operations to get data while decrementing
the remaining available size. This makes it so that code
reading or writing into a buffer cannot easily consume more
bytes than available.
This CL guards against serialization overflows using
BufferConsumer, and it implements GetPtrFromBuffer
(for deserialization) on top of BufferConsumer. A future patch
will make the rest of the deserialization code use BufferConsumer.
Bug: dawn:680
Change-Id: Ic2bd6e7039e83ce70307c2ff47aaca9891c16d91
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/41780
Commit-Queue: Austin Eng <enga@chromium.org>
Reviewed-by: Stephen White <senorblanco@chromium.org>
This is a reland of fb0bf70459
Reland after making libx11-xcb dynamically loaded since it isn't present
on all Linux deployment targets of Chromium. Also includes a couple of
additional cosmetic changes to d3d12/PlatformFunctions noticed while
looking at it for inspiration.
Original change's description:
> Vulkan: Fallback to XCB for Xlib surfaces
>
> Chromium builds the Vulkan loader without support Xlib (because it
> prefers XCB) which caused Xlib wgpu::SwapChain creation to fail on the
> Vulkan backend.
>
> This CL adds a fallback to use VK_KHR_xcb_surface if VK_KHR_xlib_surface
> isn't present.
>
> Bug: dawn:662
> Change-Id: I0e0128ee6b5c75da03998dbae231d17e48bacc81
> Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/41180
> Reviewed-by: Austin Eng <enga@chromium.org>
> Commit-Queue: Austin Eng <enga@chromium.org>
> Auto-Submit: Corentin Wallez <cwallez@chromium.org>
Bug: dawn:662
Change-Id: I617fcd1059dddfa05c29ac20d77f891ca6962342
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/41380
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Jiawei Shao <jiawei.shao@intel.com>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Auto-Submit: Corentin Wallez <cwallez@chromium.org>
Tint automatically infers a storage class of `handle` for these types
of variable, as per the WGSL spec. Explicitly specifying a storage
class for them will soon become an error.
Bug: tint:332
Change-Id: Ib67e44a5afcdd364470488fd4b456a2b42304daa
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/41402
Commit-Queue: Austin Eng <enga@chromium.org>
Auto-Submit: James Price <jrprice@google.com>
Reviewed-by: Austin Eng <enga@chromium.org>
Currently the queries availability info is only stored on the encoders
where they are written, and we need the info in the compute shader. If
resolving them from a different encoder, 0s are returned because the
queries are not written on resolving encoder.
Besides the encoders, we also need to add availability info to query
set, and use it in compute shader instead.
When resolving query set without any written, we need to reset queries
before resolveQuerySet based on the query set availability on Vulkan.
Added more end2end tests for these cases.
Bug: dawn:645
Change-Id: I09bf1230934aa885587fa4f671925940c1795cd9
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/39740
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Hao Li <hao.x.li@intel.com>
Sample mask variables and all functionality introduced by GL_OES_sample_variables was not made core until ES 3.2.
Implement a toggle to disable the functionality if not supported by the backend.
Bug: dawn:673
Change-Id: I7a5ec61fb57da343f0f72ffd3b0c69031eaaff8f
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/41142
Commit-Queue: Stephen White <senorblanco@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Readback of stencil is not supported on vanilla ES, so we introduce a toggle to disable it. NVidia GLES drivers support NV_stencil_read and NV_depth_stencil_read extensions, so we use those where available.
It turns out ANGLE supports NV_stencil_read but not NV_depth_stencil_read (for reading from the packed depth/stencil buffers which Dawn uses), so that's the extension we check for.
Bug: dawn:667 dawn:634
Change-Id: I136674d3d47fecee2b8b390d5d219bab07e3bb64
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/41141
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Stephen White <senorblanco@chromium.org>
Chromium builds the Vulkan loader without support Xlib (because it
prefers XCB) which caused Xlib wgpu::SwapChain creation to fail on the
Vulkan backend.
This CL adds a fallback to use VK_KHR_xcb_surface if VK_KHR_xlib_surface
isn't present.
Bug: dawn:662
Change-Id: I0e0128ee6b5c75da03998dbae231d17e48bacc81
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/41180
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Austin Eng <enga@chromium.org>
Auto-Submit: Corentin Wallez <cwallez@chromium.org>
SNORM textures are non-renderable on vanilla ES, which means they're also
non-readable. Use the EXT_render_snorm extension where available, otherwise
skip the test for now.
Bug: dawn:624 dawn:636 dawn:647 dawn:667
Change-Id: Ic50368032d6168060b6b52889b4ba952ce662f02
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/40420
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Stephen White <senorblanco@chromium.org>
