Deprecates the computeStage member of the descriptor in favor of compute
as described by the spec. In order to support both variants without
breaking backwards compatibility some code had to be manually added to
the wire client to copy from the deprecated member to the new one and
visa versa.
Change-Id: I9d5c2fc9c446c927c5792c9af9ed56c90060b65b
Bug: dawn:800
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/53884
Commit-Queue: Brandon Jones <bajones@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
This patch implements the asynchronous path of CreateComputePipelineAsync
on D3D12 backend with the basic framework of the dawn_unittest
AsyncTaskTest.Basic.
1. Call the constructor of dawn_native::d3d12::ComputePipeline in the main
thread.
2. Execute dawn_native::ComputePipelineBase::Initialize() (a virtual function)
asynchronously.
3. Ensure every operation in dawn_native::d3d12::ComputePipeline::Initialize()
is thread-safe (PersistentCache).
4. Save all the return values (pipeline object or error message, userdata, etc)
in a CreateComputePipelineAsyncWaitableCallbackTask object and insert this
callback task into CallbackTaskManager.
5. In Callback.Finish():
- Insert the pipeline object into the pipeline cache if necessary
- Call WGPUCreateComputePipelineAsyncCallback
Note that as we always handle the front-end pipeline cache in the main thread,
we don't need to make it thread-safe right now.
BUG=dawn:529
TEST=dawn_end2end_tests
Change-Id: I7eba2ce550b32439a94b2a4d1aa7f1b3383aa514
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/47900
Commit-Queue: Jiawei Shao <jiawei.shao@intel.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
In the service of keeping these changes incremental and easier to
review this change only validates the RenderPipelineDescriptor2
before converting it back to a RenderPipelineDescriptor.
BUG: dawn:642
Change-Id: Ie24c999756fa74f7d3cd10662a6bc95649602b14
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/46120
Auto-Submit: Brandon Jones <bajones@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Dawn's object-based cache creates keys from memory
addresses. These keys cannot be used in a persistent
cache. This change modifies the keys to only
use hashes so they can be re-used for caching
pipelines.
BUG=dawn:549
Change-Id: Ica64d58ae6a3c6266435cfc3f776c820190f7895
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/30740
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Bryan Bernhart <bryan.bernhart@intel.com>
Previously both Render and Compute pipelines handled extracting data
from the ProgrammableStageDescriptors. Unify them in PipelineBase in
preparation for gathering EntryPointMetadata in the PipelineBase.
Bug: dawn:216
Change-Id: I633dd2d8c9fdd0c08bb34cbf18955445951e312f
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/27263
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Ryan Harrison <rharrison@chromium.org>
This patch makes the |layout| member of the Render|ComputePipelineDescriptor
optional. If it is not provided, a default layout is created from the
ShaderModules provided and used to replace the layout in the descriptor.
Then, pipeline.GetBindGroupLayout may be called to get the existing, or
the computed bind group layout. If no bind group layout exists at the
provided index, an empty bind group layout is returned.
Bug: dawn:276
Change-Id: I276ed0296a2f1f2d8131fa906a4aefe85d75b3a7
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/13741
Commit-Queue: Austin Eng <enga@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jiawei Shao <jiawei.shao@intel.com>
This fixes bugs where we try to uncache objects that fail creation.
Bug: dawn:249
Change-Id: Ic60b3ce702dfdda18baa6d263911885a43d3cda7
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/12820
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
Commit-Queue: Austin Eng <enga@chromium.org>
This is to match the work in progress webgpu.h header.
BUG=dawn:22
Change-Id: I0904297bb4411b12f9d99e8457d32613058ef9b2
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/9380
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
This is to follow changes to WebGPU's IDL and header file where the Bit
suffix disappeared. This caused a conflict with
utils/dawn_native::ShaderStage that were both renamed to
SingleShaderStage.
BUG-dawn: 22
Change-Id: I7b085686775fa19e4b4a3b54979903d72f3ef660
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/10462
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Instead it is replaced by one enum in dawn_native and another in utils.
BUG=dawn:22
Change-Id: I094a40c8d4e22b704e59aea60cbefd1f05c5352a
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/8800
Commit-Queue: Kai Ninomiya <kainino@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
Error objects that are cached still have mIsBlueprint to false and would
try to uncache themselves, which isn't valid.
Also adds a regression test.
BUG=dawn:143
Change-Id: Icd8bab52982b8520519d52296fffa6ed85088b8e
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/6900
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
The contents of PipelineStageDescriptor were inlined inside of
ComputePipelineDescriptor. This changes updates
ComputePipelineDescriptor to contain PipelineStageDescriptor to match
WebGPU.
Bug: chromium:877147
Change-Id: Ic030b7bd7a237945cbbaf4c567cc361940e1ad00
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/6400
Commit-Queue: Kai Ninomiya <kainino@chromium.org>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Dawn used to return "nullptr" when an error happened while creating an
object. To match WebGPU we want to return valid pointers but to "error"
objects.
This commit implements WebGPU error handling for all "descriptorized"
objects and changes the nullptr error checks into "ValidateObject"
checks. This method is used both to check that the object isn't an
error, but also that all objects in a function call are from the same
device.
New validation is added to objects with methods (apart from Device) so
they check they aren't error objects.
A large number of ASSERTs were added to check that frontend objects
aren't accessed when they are errors, so that missing validation would
hit the asserts instead of crashing randomly.
The bind group validation tests were modified to test the behavior with
both nullptrs and error objects.
Future commits will change the CommandBufferBuilder to not be a builder
anymore but an "Encoder" instead with special-cased error handling. Then
once all objects are descriptorized, the notion of builders will be
removed from the code.
BUG=dawn:8
Change-Id: I8647712d5de3deb0e99e3bc58f34496f67710edd
Reviewed-on: https://dawn-review.googlesource.com/c/4360
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
The fuzzer is able to trigger nullptr reads by failing to create objects
and then using the resulting nullptr in other operations. The proper fix
is to implement WebGPU error handling where creation failure returns a
valid but "error" object.
However implementing this error handling is a lot of work, so in the
meantime we use nullptr checks in relevant places to fix the fuzzer
issue. These checks will be removed once the error handling is changed.
BUG=dawn:8
Change-Id: I6777a7fa40383b3d2235e071c3f0109de7605a22
Reviewed-on: https://dawn-review.googlesource.com/c/2565
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
The error type will help distinguish between validation errors, context
losts and others which should be handled differently.
Take advantage of advantage of this to change DAWN_RETURN_ERROR to
"return DAWN_FOO_ERROR" to have the return be more explicit. Also
removes usage of DAWN_TRY_ASSERT for more explicit checks.
Change-Id: Icbce16b0c8d8eb084b0af2fc132acee776909a36
Brandon Jones
Jiawei Shao
Bryan Bernhart
Corentin Wallez
Idan Raiter
Austin Eng
Jinho Bang