D3D/VK driver may return a size larger than requested and could cause
PowerOfTwo() to overflow. This change adds a check to ensure the size
is within the limit before attempting sub-allocation.
BUG=dawn:27
Change-Id: I2b2ce727abff953642a69b65c8f30be8e53e562d
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/23060
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Bryan Bernhart <bryan.bernhart@intel.com>
Similarly to the CL for Instance extensions, it makes each device
extension linked to an enum class and a bitset. Logic surrounding device
extensions is changed to take advantage of this to be more programmatic
and less error prone when adding support for a new extension.
Bug: dawn:457
Change-Id: Iecf623c40b890b7e00ba972d5eac0712866692b5
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22941
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
If a texture follow this rule:
* old usages of all subresources are the same
* new usages of all subresources are the same
Then we can use one barrier to transit states for all subresources.
We don't need to use one barrier per each subresource.
This patch adds a flag at front-end to capture many (but not all)
situations which follow the rule above, then we don't need to loop
over every subresource at Vulkan backend.
Bug: dawn:441
Change-Id: I6124ee2cb09c9142fefd8f057dc6d2659301e2d4
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22702
Commit-Queue: Yunchao He <yunchao.he@intel.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Adds harness for fuzzing the d3d12 backend with a WARP device.
Bug: dawn:444
Change-Id: I3e5ca325b19eb8b6cfe53e4d833c5acbbd293a26
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22880
Commit-Queue: Brandon Jones <brandon1.jones@intel.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
At the moment each instance extension needs special handling
in 6+ places when added to the Vulkan backend. This is very
error-prone and makes it difficult to do changes in how they
are extensions are handled.
This CL makes instance extensions linked with an enum class
and a bitset to know which are available (instead of individual
booleans). A table of known extensions with more information like
`versionPromoted` so that they can be handled programmatically.
Bug: dawn:457
Change-Id: I266deb730eb2b7f3ab0ee7ada1b06ff9748a60e4
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22940
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
This case verifies that reading from one read-only storage texture then
writing into another write-only storage texture in one dispatch are
supported in compute shader.
Bug: dawn:458
Change-Id: If1b4c13da067fa39b45a378b54c22a4162695c8d
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/23040
Commit-Queue: Jiawei Shao <jiawei.shao@intel.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jiawei Shao <jiawei.shao@intel.com>
This catches a few driver bugs or validation errors when doing empty
copies.
Bug: dawn:446
Change-Id: Ibef50b0c03b2fa24bc00cf5c7b6d26b8ac6fae8a
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/19602
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Stephen White <senorblanco@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
When creating a zero-sized buffer mapped, StagingBuffer creation
is skipped. This required adding a new MappedAtCreation state
since mStagingBuffer couldn't be used as a tag value for that.
Made the OpenGL backend always create non-zero-sized buffers.
Finally added tests for MapRead/WriteAsync and CreateBufferMapped
of zero-sized buffers.
Bug: dawn:446
Change-Id: I04f6fe98fd646f1867c21065cd1cd33a1595e19f
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/21481
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
This patch enables nonzero_clear_resources_on_creation_for_testing
toggle on buffer on Vulkan backends as a preparation of supporting
buffer lazy-initialization in Dawn.
BUG=dawn:414
TEST=dawn_end2end_tests
Change-Id: I7619d02ae898e30fd15438d35437802a09e959cc
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22981
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
Commit-Queue: Kai Ninomiya <kainino@chromium.org>
This patch enables the tests CopyTests_T2T/CopyWithinSameTexture* on
D3D12 after the D3D12 texture subresource tracking has been supported.
This patch also changes the related comment in Vulkan as WebGPU CG has
already decided to disallow the overlaps among source and destination
subresources in texture-to-texture copies.
BUG=dawn:453
TEST=dawn_end2end_tests
Change-Id: Ibf008c800cfea62658050f565f321a391412c54c
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/23000
Commit-Queue: Kai Ninomiya <kainino@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
Pipeline layout incorrectly indexes into a root table array
when there are more root descriptors than root tables.
To fix, the array is dynamically sized where parameters
are appended instead of indexed into the root signature.
Bug: dawn:449
Change-Id: I6d7f65fb791d323704b1c3a3af9c871a79e32a30
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22960
Commit-Queue: Bryan Bernhart <bryan.bernhart@intel.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Metal supports FP16 on both macOS and iOS, so we can
unconditionally enable this extension on Metal backend.
BUG=dawn:426
TEST=dawn_end2end_tests
Change-Id: I16c9c907bb112a66e76506ede9bc89a04313d7b3
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22882
Reviewed-by: Jiawei Shao <jiawei.shao@intel.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Xinghua Cao <xinghua.cao@intel.com>
This patch enables nonzero_clear_resources_on_creation_for_testing
toggle on buffer on Metal backends as a preparation of supporting buffer
lazy-initializations in Dawn.
BUG=dawn:414
TEST=dawn_end2end_tests
Change-Id: Ia7106482922aeb0521affa206f459d40938a0131
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22861
Commit-Queue: Kai Ninomiya <kainino@chromium.org>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
This patch makes Dawn provide more detailed error messages when the
pipeline layout is not compatible with shader module, which is helpful
to debug such errors in WebGPU applications.
BUG=dawn:456
Change-Id: Ib5a870d8e66645481434c4d3dc6fdc1a585aac36
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22881
Commit-Queue: Kai Ninomiya <kainino@chromium.org>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
When we use a texture for different purpose, we need to add proper
barrier(s) in order to make it ready. Previously, the barrier is
done per entire texture. So it is invalid to sample/read/copy from
one subresource (say a mip/array slice) and render/write/copy to
another subresource of the same texture at the same time.
With this patch, barrier is set per each texture subresource. So it is
valid to use a subresource as source and use another subresource of the
same texture as destination at the same time.
However, planar slices like depth/stencil planes are not handled
gracefully. This is a TODO task. Another task is to combine barriers
into one if they can be combined. I will do this optimization in
another patch in near future.
Bug: dawn:157
Change-Id: I783a76cb88fcdffb60c307ddfb89d50f1583201a
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22101
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Yunchao He <yunchao.he@intel.com>
This patch combines barriers in a render/compute pass into one
function call.
Previously, we need to dispatch barrier(s) for each buffer/texture
in a pass. So we may need quite a lot function calls to deliver
barriers in a pass in real web applications. One example is that
we did see that too many function calls to deliver barriers in
Aquarium (WebGPU porting) contributed to CPU usage and bottleneck.
Bug: dawn:441
Change-Id: Ibe44967fefd2e1e6e64df4587146c4fb7fbe8e73
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22700
Commit-Queue: Yunchao He <yunchao.he@intel.com>
Reviewed-by: Austin Eng <enga@chromium.org>
The defaulting when bytesPerRow is 0 has been removed so there was a
validation error.
Bug: dawn:432
Bug: dawn:455
Change-Id: I8e38eb7aeb315d93cc8a9528beb77b9285b368d7
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22800
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
This reverts commit 0357eed7de
and reland commit bdc05c3d5f.
The Vulkan-Loader has a bug where if the instance is created
with Vulkan 1.1 and not the promoted extensions, it will skip
emulation and if the ICD doesn't support Vulkan 1.1 nor the
extensions. Enable the promoted extensions, even when creating
a Vulkan 1.1 instance.
Original change's description:
> Check FP16 support on vulkan backend
>
> This patch check FP16 support on vulkan backend, and introduces
> the shader_float16 extension.
>
> BUG=dawn:426
> TEST=dawn_end2end_tests
>
> Change-Id: Ie09568a416ce9eb2c11afeede3e7da520550d5fb
> Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/21901
> Commit-Queue: Xinghua Cao <xinghua.cao@intel.com>
> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
> Reviewed-by: Jiawei Shao <jiawei.shao@intel.com>
> Reviewed-by: Kai Ninomiya <kainino@chromium.org>
> Reviewed-by: Austin Eng <enga@chromium.org>
Bug: chromium:1087896, dawn:426
Change-Id: I2c4465fb2fe957966b44d3e5840112219481c639
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22781
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
This patch adds validations on the texture-to-texture copies within the
same texture to align with the latest change in WebGPU SPEC: When the
source and destination textures are the same one, the source and the
destination subresources involved in the copy must not overlap.
Note that we don't enable the newly added end2end tests on D3D12
because when doing texture-to-texture copy within the same texture, we
need to set the source subresources into TRANSFER_SRC state and set the
destination subresources into TRANSFER_DEST state, while right now we
don't support subresource tracking on D3D12.
BUG=dawn:453
TEST=dawn_unittests
TEST=dawn_end2end_tests
Change-Id: I6408640d01beaf6ab9ef30b001e9c87cfecbdd65
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/21601
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Jiawei Shao <jiawei.shao@intel.com>
This patch changes all the tests related to write-only storage textures
in dawn_end2end_tests StorageTextureTests to make them test all the
texture formats that support being used as storage textures in WebGPU.
BUG=dawn:267
TEST=dawn_end2end_tests
Change-Id: I4e49efc21f768a9b7645bf334bd41a97fd4838e7
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22663
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Jiawei Shao <jiawei.shao@intel.com>
Zero-sized copies are invalid in a couple backends, and in follow up
CLs CreateBufferMapped will be change to handle zero-sized buffers
correctly.
Bug: chromium:1069076
Change-Id: Ieef62a13182bbe1e939a3847980c91339e42aa8f
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22460
Reviewed-by: Stephen White <senorblanco@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
This patch changes all the tests related to read-only storage textures
in dawn_end2end_tests StorageTextureTests to make them test all the
texture formats that support being used as storage textures in WebGPU.
BUG=dawn:267
TEST=dawn_end2end_tests
Change-Id: I4d6ddbee638a787a2dcfc626bd4963a9b9043772
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22640
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Jiawei Shao <jiawei.shao@intel.com>
To avoid overly ticking, we only want to tick when:
1. the last submitted serial has moved beyond the completed serial
2. or the completed serial has not reached the future command serial added
by the trackers (MapRequestTracker, FenceSignalTracker, ErrorScopeTracker).
Bug: dawn:400
Change-Id: Ie7c65acc332846ac1a27f9a18f230149d96d2189
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/19062
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Natasha Lee <natlee@microsoft.com>
Bug: dawn:433
Change-Id: I47d9ffe16d6ad7730628dc27ec6708006cf4384d
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22560
Commit-Queue: Austin Eng <enga@chromium.org>
Reviewed-by: Stephen White <senorblanco@chromium.org>
Per MSDN recommendations, Dawn should handle MakeResident failures by
evicting some more and attempting MakeResident again.
Bug: dawn:193
Change-Id: I0a9d326dcd000360f6eafb5691efb4987a77e8d5
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22280
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Brandon Jones <brandon1.jones@intel.com>
Some WebGPU CTS tests were failing with invalid WGPUTextureFormat
when really the format was simply missing. This CL updates
the validation code so it is more clear.
Bug: dawn:433
Change-Id: Ie7aacd01ea424d9f5df0078703a474e98b72ce41
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22543
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Stephen White <senorblanco@chromium.org>
Commit-Queue: Austin Eng <enga@chromium.org>
If the shader declares a storage buffer RO but uses storage buffer in
the BGL, the shader compiler will be told to treat these bindings as
UAV instead of SRV to avoid PSO mismatches.
Bug: dawn:410
Change-Id: I3be3257449de55fd2d35e914233b48c6f7121b58
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22322
Commit-Queue: Bryan Bernhart <bryan.bernhart@intel.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
This is blocking Skia from updating to newer clang
Tbr=cwallez@chromium.org
Bug: chromium:1064305
Change-Id: I7d894dba701ebe2fe0e1b78d5fb42032d88846aa
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22522
Commit-Queue: Austin Eng <enga@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Mike Klein <mtklein@google.com>
This was hitting an ASSERT because D3D12 doesn't have an option
to set None as the shader visibility.
Bug: dawn:448
Change-Id: I3e056e531e7d1bb89da1736bc609bfe97a2fa194
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22324
Reviewed-by: Bryan Bernhart <bryan.bernhart@intel.com>
Commit-Queue: Austin Eng <enga@chromium.org>
This function was validating the depth of the copy which is
incorrect for 2D array textures. Remove this check for now
since it would be only relevant for 3D textures which Dawn
does not support yet.
Bug: dawn:424
Change-Id: I756080a899a7c5effe5843a530d4db0571bc10d5
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22323
Reviewed-by: Jiawei Shao <jiawei.shao@intel.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Austin Eng <enga@chromium.org>
The windows.h macros were undefined only at the end of this platform
header previously but with the addition of d3d12sdklayers.h the
definition of ID3D12DebugQueue::GetMessage picked up the macro and
became GetMessageA or GetMessageW, but Dawn code referred to it as
GetMessage causing a compilation error.
Fix this by preemptively loading windows.h and undefing some of the
macros so that the D3D12 headers don't see them.
Bug:
Change-Id: I1985cc20a9bdec1d25619ac5088e918b2acf8ecb
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22400
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
A previous CL sorted bindings by binding number, but bindings were first
sorted by type. This means a bind group layout with mixed dynamic
storage and uniform buffers would not always have all dynamic bindings
in increasing order. Instead, it would be strictly increasing within
each section of uniform/storage buffers. This CL corrects the issue
by first sorting dynamic buffers by binding number.
Bug: dawn:408
Change-Id: I3689eb64ad8aa8768cebe266eebcba75a21894ce
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/22303
Commit-Queue: Austin Eng <enga@chromium.org>
Reviewed-by: Stephen White <senorblanco@chromium.org>
This will help remove noise in debug layer output. This patch also
promotes warnings and higher to Dawn errors.
BUG: dawn:363, dawn:418, dawn:419, dawn:421
Change-Id: I3112c94aff71fc7e76dff48c82bafe9e051ed3b4
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/21702
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Bryan Bernhart <bryan.bernhart@intel.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
When device initialization failed, DeviceBase::mCurrentErrorScope is
nullptr so we need to guard the call to UnlinkForShutdown.
Bug: chromium:1081998
Change-Id: I65a1948117fa9461f7fc8d2d1d70aef669bc375d
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/21685
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Stephen White <senorblanco@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Bryan Bernhart
Corentin Wallez
Yunchao He
Brandon Jones
Ryan Harrison
Xing Xu
Hao Li
Jiawei Shao
Natasha Lee
Xinghua Cao
Idan Raiter
Austin Eng
Enrico Galli