This member will be removed. This first requires removing code setting
it in Chromium, which will make it contain garbage value. If we kept
validating it, then we'd fail validation spuriously.
Bug: dawn:1445
Change-Id: I8daa90b949db867b89fcf955cfaec45f7845210f
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/92241
Reviewed-by: Ben Clayton <bclayton@google.com>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
In this patch NumWorkgroupsFromUniform::Config changed to storage
std::optional<sem::BindingPoint>, and if it has no value,
NumWorkgroupsFromUniform will choose a free binding group, i.e.
binding 0 of the largest used group plus 1 is used if at least one
resource is bound, otherwise group 0 binding 0 is used. Tint CLI
is also changed to provide a --hlsl-root-constant-binding-point
option allowing user to specify the binding point for num_workgroups
uniform buffer.
Bug: tint:1566
Change-Id: I3b8c22a4276bab722d901f5b07d23a268786c417
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/91980
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Commit-Queue: Zhaoming Jiang <zhaoming.jiang@intel.com>
We were missing an `UnwrapRef` when generating the return type, and
were generating invalid SPIR-V when the value being stored was a
reference. The auto-generated builtin tests only test with literal
values.
Fixed: tint:1573
Change-Id: If42280b3cc8ad3fba7355d333e02400c6db843fa
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/92144
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Auto-Submit: James Price <jrprice@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: James Price <jrprice@google.com>
Also fixed implementation of this atomic in GLSL. It was emitting code
that would not compile because, as for HLSL, we must pass in the
variable directly to atomic funcs, not via an in/out arg to a function.
Bug: tint:1185
Change-Id: Id0e9f99d6368717511ef3a94473634c512e10cb8
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/91881
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Commit-Queue: Antonio Maiorano <amaiorano@google.com>
These are tests for when there's no explicit target type for a
materialization. In this case we expect an abstract-int to
materialize to an i32 and an abstract-float to materialize to a
f32.
Fix a bug uncovered where we were creating a transposed matrix.
Bug: tint:1504
Change-Id: Ie69dd7ec47174d3d7bef20315fd3780dade3a325
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/91845
Reviewed-by: David Neto <dneto@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Previously, the device/adapter were all created on the service-side
of a test, and then injected into the client side. Injected devices
and adapters do not support querying limits and features.
This CL changes setup so that adapter and device creation is always
initiated by the client - and the implementation on the service side
may be overridden for test fixture-specific behavior.
It also adds more fuzzing coverage since the fuzzers can now also
create adapters and devices.
Bug: dawn:689
Change-Id: Ief7faa1908ceae973dcb2f600bf4dd1cf5417704
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/91680
Reviewed-by: Loko Kung <lokokung@google.com>
Commit-Queue: Austin Eng <enga@chromium.org>
SwiftShader does not export function pointer type information.
So, when fuzzing with UBSAN, fuzzers break whenever calling
a vk* function since it thinks the type of the function pointer
does not match.
Workaround this problem by proxying through a std::function
in UBSAN builds. The std::function delegates to a Call method
which does the same cast of the function pointer type, however
the Call method is tagged with
`__attribute__((no_sanitize("function")))` to silence the error.
Bug: chromium:1296934
Change-Id: I6971eecdda8ae10542a8d9bfb942f841c50227c5
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/91740
Commit-Queue: Austin Eng <enga@chromium.org>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
This patch adds the support of the experimental feature
chromium_experimental_dp4a on Vulkan. Currently this
feature is enabled on Vulkan backend only when DP4a
instructions are hardware-accelerated.
Bug: tint:1497
Test: dawn_end2end_tests
Change-Id: I5a63111a6b5972aa1934f0e7be984ebdb1e35080
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/91520
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Jiawei Shao <jiawei.shao@intel.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
This change moves the bulk of the existing GPUInfo functionality into an autogenerated source built from a JSON file that describes the GPU
vendor and device IDs, with device IDs broken down by GPU architecture.
Also adds the fields needed to implement GPUAdapterInfo in Blink to the AdapterProperties.
Bug: dawn:1427
Change-Id: I6a8b1fa7a63ec8d71556fc5bb3ae12cfe5abf28b
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/90962
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Brandon Jones <bajones@chromium.org>
When converting values between two concrete types, handle the case that
the value is unrepresentable by the target type.
For integers, the converted value will be either the maximum or minimum
value for the integer type.
For floats, the converted value will be positive or negative infinity.
Bug: tint:1504
Change-Id: Ia56fb8170c0ea994632194f166062823d9507249
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/91621
Reviewed-by: David Neto <dneto@google.com>
Commit-Queue: Ben Clayton <bclayton@chromium.org>
Currently Dawn ignores all zero-attribute vertex buffer in the given
pipeline descriptor when creating RenderPipelineBase because
zero-attribute vertex buffer is treated as unused slot, however the spec
doesn't state that zero-attribute vertex buffer should be ignored.
To support zero-attribute vertex buffer, this commit has the following
changes.
1. Add VertexBufferNotUsed enum value to wgpu::VertexStepMode to
represent unused slots
2. Ignore VertexBufferNotUsed step mode buffers when creating
RenderPipelineBase and add tests to check it
3. Record zero-attribute vertex buffers when creating RenderPipelineBase
and add tests to check it
4. Fix VertexStateTest::LastAllowedVertexBuffer broken by the above
changes
Temporarily we set the enum value of
wgpu::VertexStepMode::VertexBufferNotUsed to 0 to pass the CTS tests
because currently empty vertex buffer slots step mode can be
zero-initialized. We will make a CL to Blink to explicitly set
wgpu::VertexStepMode::VertexBufferNotUsed for empty slots and change
the enum value to 2.
Bug: dawn:1000
Change-Id: Ibd4ab87f2c922e8e460f2311547f13d58f1d5611
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/89340
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Takahiro <hogehoge@gachapin.jp>
