Bug: dawn:628
Change-Id: Iaaa62507592e12a724673a0226783c0425b04f35
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/67601
Commit-Queue: Loko Kung <lokokung@google.com>
Reviewed-by: Austin Eng <enga@chromium.org>
Instead of using BufferLocation as another layer of indirection,
the indirectBuffer can be set directly on the indirect command.
This makes the indirect validation a bit simpler, but introduces
additional lifetime dependencies in that the indirect draw validation
MUST be encoded while the DrawIndexedIndirectCmds it references
are still valid.
Bug: dawn:809
Change-Id: I1ef084622d8737ad5ec1b0247bf9062712e35008
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/67241
Commit-Queue: Austin Eng <enga@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
This will help with review load and finding an owner once mandatory
owner review is enabled.
Change-Id: I1a90b83fed68f062c1406552aec218bf9c2092ce
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/67821
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
The changes should pass through the destroy changes such that when the device is destroyed, the respective destroy functionality currently existing in the backends should be called.
For buffers, destroy no longer causes validation errors since even error buffers may need to be destroyed in the case of mappedAtCreation.
Bug: dawn:628, dawn:1002
Change-Id: I42a475af5d67cc60f86d95ac53c2b377a9fd2e82
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/65863
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Loko Kung <lokokung@google.com>
We now always use Tint.
Bug: dawn:571
Change-Id: Ic65669a9e00493292ed60cde00af6fa3f6e55ec2
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/65665
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Ben Clayton <bclayton@google.com>
This commit also unifies the initialization process for Adapters.
InitializeImpl() initializes the actual backend adapter.
InitializeSupportedFeaturesImpl() checks base WebGPU features and
discovers additional supported features.
InitializeSupportedLimitsImpl() checks base WebGPU limits and
queries the adapter's maximum supported limits.
Some of these limits from the backend are still overriden in the
frontend because they are limited by internal Dawn constants.
Bug: dawn:685
Change-Id: I43efb0b678dd45f8f89cd62d13104dd00b197da1
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/64980
Commit-Queue: Austin Eng <enga@chromium.org>
Reviewed-by: Brandon Jones <bajones@chromium.org>
- Moving the call into DeleteThis should make it so that derived classes don't need to explicitly implement a destructor that calls DestroyApiObject.
Bug: dawn:628
Change-Id: I145f42e7e4c144cc0d2d7c7f609744399d514fe1
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/66840
Commit-Queue: Loko Kung <lokokung@google.com>
Reviewed-by: Austin Eng <enga@chromium.org>
This patch removes a redundant creation of ComputePipelineBase object
in GetCachedComputePipeline(). Instead, we directly compute the blueprint
hash from the uninitialized backend compute pipeline object.
BUG=dawn:529
TEST=dawn_end2end_tests
Change-Id: I9b982664aa140ab385418a202270b9988cfcb9f3
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/66221
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Jiawei Shao <jiawei.shao@intel.com>
- Start tracking BindGroupLayout objects at construction
- Utilizes untrack tag for blueprint layouts for caching purposes
- Adds dawn native test file for testing utilities that require static dawn native lib
- Adds testing macros and mocks for simple sanity unit testing
Bug: dawn:628
Change-Id: Ic85b60e9574e67cc5fc1804cc5300cd1f3a0f6fd
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/65862
Commit-Queue: Loko Kung <lokokung@google.com>
Reviewed-by: Austin Eng <enga@chromium.org>
- Renames some of the Device functions to be consistent with documentation
- Reverts change in https://dawn-review.googlesource.com/c/dawn/+/64820 for overloading mDevice == nullptr to determine if objects are alive because device is needed for error propagation. Instead, use list existence to determine if objects are alive
- Updates destroy api to return bool upwards in case we need to further process the extending objects
- Adds tracking functions in ObjectBase
- Adds final tag to all backend Device implementations
- Adds MoveInto LinkedList support to move list elements in O(1)
Bug: dawn:628
Change-Id: Iff70f4f7d55f46ee52d1bd0e02e3671819f2eed4
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/65861
Commit-Queue: Loko Kung <lokokung@google.com>
Reviewed-by: Austin Eng <enga@chromium.org>
This is a reland of a57c1db878
in current Dawn implementation RenderPipelineGL and ComputePipelineGL
will be destroyed before calling their initialize() function when we
can find a proper pipeline object in the front-end cache, so mProgram
may not be assigned to a valid GL program ID when the destructor of
PipelineGL is called.
In this CL we always initialize mProgram to 0 in the constructor of
PipelineGL so that it won't be a garbage value in the destructor of
PipelineGL (0 is safe for glDeleteProgram() according to OpenGL SPEC).
Original change's description:
> OpenGL: delete shaders and pipelines when they are not used any longer
>
> Previously on OpenGL backend the GL pipelines and shaders are never
> deleted. With this patch the GL pipelines and shaders will be able to
> be destroyed correctly after they are not needed any longer.
>
> BUG=dawn:529
>
> Change-Id: I4f7f22c7b536825363fe1ecc0f5ffd1bb86fd774
> Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/65140
> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
> Reviewed-by: Stephen White <senorblanco@chromium.org>
> Commit-Queue: Jiawei Shao <jiawei.shao@intel.com>
Bug: dawn:529
Change-Id: Ie04ab069b9d26658f2b0d1b070d86bb650f3c878
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/65486
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Stephen White <senorblanco@chromium.org>
This CL renames "extension" to "feature" to follow WebGPU. It still
supports both. A future Chromium CL will pick this change, then all
"extension" occurrences will be removed.
Change-Id: I070e32d7ae042f9b846df01f200b39f6741a0a14
Bug: dawn:1149
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/65664
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: François Beaufort <beaufort.francois@gmail.com>
This patch removes a redundant creation of RenderPipelineBase object
in GetCachedRenderPipeline(). Instead, we directly compute the blueprint
hash from the uninitialized backend render pipeline object.
BUG=dawn:529
TEST=dawn_end2end_tests
Change-Id: I3c7a2acfc01ab9e1e631793030ea06c9ae908aa2
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/65000
Commit-Queue: Jiawei Shao <jiawei.shao@intel.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Previously on OpenGL backend the GL pipelines and shaders are never
deleted. With this patch the GL pipelines and shaders will be able to
be destroyed correctly after they are not needed any longer.
BUG=dawn:529
Change-Id: I4f7f22c7b536825363fe1ecc0f5ffd1bb86fd774
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/65140
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Stephen White <senorblanco@chromium.org>
Commit-Queue: Jiawei Shao <jiawei.shao@intel.com>
Every render pass which invokes DrawIndexedIndirect, either directly or
through a RenderBundle execution, is now preceded immediately by at
least one validation pass.
All indirect buffer offests used with DII are validated, and their
validated values are copied into a separate scratch buffer (or zeroed
out there, in the case of validation failure). All encoded DII commands
are rewritten to use the validated parameters instead of the original
ones.
Bug: dawn:809
Change-Id: I5eead937f19536f84f89e2c8e6fed7f18f0aee9f
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/63461
Commit-Queue: Ken Rockot <rockot@google.com>
Reviewed-by: Austin Eng <enga@chromium.org>
This is a simple RefCounted holder of a Buffer ref and offset. Encoded
commands can store a ref to this object instead of directly storing an
inline Buffer ref and offset, allowing other commands to dynamically
patch in a different buffer+offset as needed, in a memory-safe way;
as opposed to e.g. retaining an unmanaged pointer to the encoded
command itself and modifying it in-place.
Validation commands will use this to rewrite buffer references in
encoded commands, so that they execute over validated inputs rather
than over their original client-provided inputs.
No net functional changes in this CL, just some groundwork for indirect
draw/dispatch validation.
Bug: dawn:809
Change-Id: I0570521a610fe3ea08190a525b4904749d7b7f24
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/64420
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Ken Rockot <rockot@google.com>
This command copies data from host memory into a GPU buffer. It's
analogous to Queue::WriteBuffer, but executed in the context of a
command buffer, sequenced with other encoded commands. This is useful
for supporting a notion of a shared scratch buffer, with a single
allocation whose contents may need to be overwritten with new data
before each pass that uses it.
Bug: dawn:809
Change-Id: If58d49c52a41127e2980dd626fd687eb1c91fe28
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/64001
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Ken Rockot <rockot@google.com>
This patch implements the asynchronous path of CreateRenderPipelineAsync
on D3D12 backend.
1. Call the constructor of dawn_native::d3d12::RenderPipeline in main
thread.
2. Execute dawn_native::RenderPipelineBase::Initialize() (a virtual function)
asynchronously.
3. Ensure every operation in dawn_native::d3d12::RenderPipeline::Initialize()
is thread-safe.
4. Save all the return values (pipeline object or error message, userdata, etc)
in a CreateRenderPipelineAsyncWaitableCallbackTask object and insert this
callback task into CallbackTaskManager.
5. In Callback.Finish():
- Insert the pipeline object into the pipeline cache if necessary
- Call WGPUCreateRenderPipelineAsyncCallback
This patch also removes FlatRenderPipelineDescriptor as it is not needed
right now.
BUG=dawn:529
TEST=dawn_end2end_tests
Change-Id: I7fd30339ab7bea599c483dea4bd1100359982409
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/64440
Commit-Queue: Jiawei Shao <jiawei.shao@intel.com>
Reviewed-by: Austin Eng <enga@chromium.org>
Support vertex-only render pipeline on D3D12, Vulkan, Metal, OpenGL
and OpenGL ES backends. Related validation tests and end to end tests
are also implemented.
Bug: dawn:136
Change-Id: If266fd441c1d39ccd940ea26b74b405f8abb351a
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/63080
Commit-Queue: Zhaoming Jiang <zhaoming.jiang@intel.com>
Reviewed-by: Austin Eng <enga@chromium.org>
This patch removes the parameter "descriptor" in the function
ComputePipeline::Initialize() so that we don't need to define
FlatComputePipelineDescriptor right now.
For render pipeline, as descriptor->vertex is being used for vertex
pulling (passed into vertexModule->CreateFunction()), we will first
refactor the related code in vertex pulling before removing the
parameter "descriptor" in the function RenderPipeline::Initialize().
BUG=dawn:529
Change-Id: Ib172ac0c76fa24070e78c0e57c3262acad9399b9
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/64000
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Jiawei Shao <jiawei.shao@intel.com>
Prevents bind group layouts created by a default pipeline layout from
being reusable with any other pipelines or layouts, as detailed in
https://github.com/gpuweb/gpuweb/pull/2068
Change-Id: Ic398eb6c6e089ac63ce6650f125dd20a9dfc8862
Bug: dawn:1094
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/63220
Commit-Queue: Brandon Jones <bajones@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Buffer allocations in Dawn may be padded. This padding
could be visible, depending on how the backend
implements robust vertex buffer access. This commit
updates buffer creation to clear all padding bytes
immediately after creation.
It is not counted as a lazy clear.
And, add a test which reads off the end of a padded
vertex buffer to check that the padding bytes are also
initialized.
Also: Update Metal buffers to always allocate enough
space for Tint's vertex pulling, and pass the full
allocated size to Tint. While writing the test in this
commit, a bug was found where Tint assumes there is at
least 4 bytes in the vertex pulling buffer. The WebGPU
API currently allows zero-sized bindings, so Dawn needs
to always allocate enough space for one element.
Also: Update Vulkan vertex/index buffers to allocate
at least one more byte than requested because
vkCmdSetVertexBuffers and vkCmdSetIndexBuffer disallow
the offset to be equal to the buffer size. We need
at least one more byte to make zero-sized bindings
as the end of the buffer valid.
Lastly: Update helper so that a T2B copy of a single
row that touches all bytes in a buffer updates the
buffer state to be fully initialized. This is a small
optimization that is necessary to write the test.
Bug: dawn:837
Change-Id: I829f4764509c4ec784c5aeaaf40b6dcbd4be6866
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/62161
Commit-Queue: Austin Eng <enga@chromium.org>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
The only part of mGLEntryPoints actually used is the bindings, so
replace ShaderModuleGL::mGLEntryPoints with mBindings.
This required extracting BindingInfoArray from EntryPointMetadata,
making it visible in the dawn_native namespace.
Remove all non-bindings-related reflection, and MSL-specific checks
and workarounds.
Bug: dawn:1076
Change-Id: I05657c0c89f5d8a2185e55f9ad7c8f81d89a8e60
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/62180
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Stephen White <senorblanco@chromium.org>
The now-removed SPIRV-Cross path used to always do this, and the
pure Tint-only path never actually validated the SPIRV. Tint
does not run SPIRV-Tools validation on its output, so add in
validation to ensure we don't pass invalid SPIRV to the driver.
The validation can probably eventually be removed when we're more
confident that Tint's SPIRV output is always correct.
Also include various cleanups for old / unused code.
Bug: dawn:1036
Change-Id: Iaab037518965e52edbd1829f6ab6ba2af0e70143
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/61589
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
Reviewed-by: Jiawei Shao <jiawei.shao@intel.com>
Commit-Queue: Jiawei Shao <jiawei.shao@intel.com>
Auto-Submit: Austin Eng <enga@chromium.org>
Bug: dawn:1036
Change-Id: Ia60c71558bd2c47a6b7148d86b8dcedfaa79ba89
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/61585
Reviewed-by: Stephen White <senorblanco@chromium.org>
Commit-Queue: Austin Eng <enga@chromium.org>
This commit adds a getter so that Chromium can check whether an
Adapter would support external images. External images are
necessary to implement the WebGPU swapchain and interop with the
rest of the web platform. We should not expose adapters that
cannot interop with the canvas and other web platform primitives
until WebGPU gains options to request more "limited" adapters.
Bug: dawn:1056
Change-Id: Iff83ac26b0e92dccdd4e29c0d854d9276a4a37bc
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/61580
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
Reviewed-by: Brian Ho <hob@chromium.org>
Commit-Queue: Austin Eng <enga@chromium.org>
Since the other backends no longer need it, move all of the
utility functions specific to SPIRV-Cross into the OpenGL backend.
This obviates the need for the DAWN_USE_SPIRV_CROSS define, so remove it.
Bug: dawn:1036
Change-Id: I67bb5a85dc128a6f343d09876046cf559395e05f
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/61541
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Stephen White <senorblanco@chromium.org>
Older names are kept around as an alias for a while. Unfortunately we
have no mechanism for producing deprecation error messages when they
are used.
Bug: dawn:1035
Change-Id: Ic6716fd526ecbedaa8e7925ab93e3ff32f642d97
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/61382
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
Commit-Queue: Brandon Jones <bajones@chromium.org>
Buffer allocations may need padding based on backend requirements.
GetAllocatedSize returns the size of the buffer actually allocated.
This CL also updates the backends to use GetAllocatedSize for buffer
clearing and barrier operations which should operate over the
entire allocated resource.
Bug: dawn:1011
Change-Id: Ic5233214414fa090725a4f50fff70d6e178494c3
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/60701
Reviewed-by: Stephen White <senorblanco@chromium.org>
Commit-Queue: Austin Eng <enga@chromium.org>
Fix for crbug.com/dawn/1049, where setting a pipeline without drawing can
prevent bind groups from being applied later. This occurs because the mask
for the pipeline is being saved but not its layout, because the bind
groups are never applied. This changes to only save the mask if the bind
groups are applied (the pipeline is used in a draw or dispatch).
Bug: dawn:1049
Change-Id: I4c7ae1125d1b6a06af90aea49a9dd1e4883f4826
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/60740
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Austin Eng <enga@chromium.org>
Some Vulkan drivers don't handle multiple entrypoints well.
In addition, SPIRV-Cross translation can be wrong for
shader modules with multiple entrypoints. Always emit a single
SPIR-V entrypoint to workaround these issues.
This allows updating CopyTextureForBrowser to use a single
shader module, and it fixes some tests with multiple
entrypoints.
Fixed: dawn:948, dawn:959, dawn:1013
Change-Id: Ie129a32a54845316d11917331937ca44fba3d347
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/60640
Commit-Queue: Austin Eng <enga@chromium.org>
Reviewed-by: Ben Clayton <bclayton@google.com>
See https://github.com/gpuweb/gpuweb/pull/1927
Adds a typedef to make a gradual deprecation.
Bug: dawn:1023
Change-Id: Ic81a933a95556fbf5726056530788dde017a1449
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/59442
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Stephen White <senorblanco@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
This replaces multiple switches with a single indexing into a data
table, and also adds information about the VertexFormatBaseType for
later validation.
Bug: dawn:1008
Change-Id: Ic3f564b0dc571cc0008a54785613f962ce223452
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/59030
Reviewed-by: Jiawei Shao <jiawei.shao@intel.com>
Commit-Queue: Jiawei Shao <jiawei.shao@intel.com>
Auto-Submit: Corentin Wallez <cwallez@chromium.org>
--disable_workgroup_init will disable the workgroup memory zero initiailization. Useful for benchmarking.
--disable_symbol_renaming will disable tint's symbol renamer. Useful for debugging output.
Bug: tint:1003
Fixed: dawn:1016
Change-Id: I92486ef88a2c1112d9ccb40f7920947bd2011c70
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/58861
Commit-Queue: Ben Clayton <bclayton@google.com>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Auto-Submit: Ben Clayton <bclayton@google.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Tint now has a single-function API for code generation that
automatically runs the backend-specific sanitizer transforms. This API
allows for backend-specific configuration options such as the fixed
sample mask for MSL (and, in the future, information about which
compiler/version is being targeted), and returns any backend-specific
metadata such as whether a UBO of buffer sizes is required by the
shader.
This change prevents the post-sanitizer program from being exposed to
Dawn, which is a potential foot-gun (e.g. the Inspector is not
expected to be run after the sanitizer transforms, and Dawn is
currently doing this).
The old Generator class API will be removed from Tint shortly after
landing this change.
Bug: tint:697
Change-Id: I9b988d55514f810d3091ec6471731e6eb41dc27f
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/57103
Auto-Submit: James Price <jrprice@google.com>
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Austin Eng <enga@chromium.org>
This patch adds a missing check in QueueGL::WriteTexture() to return
a validation error when we try to call writeTexture() on the stencil
aspect on OpenGL backends as copying from a buffer into a stencil
texture is not supported on OpenGL.
BUG=dawn:439
Change-Id: I203d40aefe346aea663a492c5a065248b74f8ea3
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/56002
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Jiawei Shao <jiawei.shao@intel.com>
Fixes locations that explicitly have TODOs about
checking for overflow. Also fixup other locations found while
searching the code for arithmetic operations.
Fixes: dawn:830
Change-Id: I4ef6b97a9cde14439e573a1da8d569ab985efc53
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/55605
Reviewed-by: Jiawei Shao <jiawei.shao@intel.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Austin Eng <enga@chromium.org>
Loko Kung
Dawn Autoroller
Austin Eng
Corentin Wallez
Brandon Jones
Jiawei Shao
François Beaufort
Ben Clayton
Ken Rockot
Zhaoming Jiang
Stephen White
shrekshao
Aleksi Sapon
James Price