Offset requires the ImageGatherExtended capability, where as ConstOffset does not. WGSL enforces the offset range required by SPIRV.
Bug: tint:143
Change-Id: Ia098010ecdb8a4a01999b4e9a6faabf3f4d1251f
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/37447
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: David Neto <dneto@google.com>
All shader backend languages describe cubes as width / height, but give no way to query depth. WGSL however returns a vec3<i32> for cube textures when calling textureDimensions().
As cube textures must be square (width == height == depth), just replicate the height for the depth.
See https://github.com/gpuweb/gpuweb/issues/1345
Bug: tint:140
Bug: tint:437
Change-Id: I76ef18ee4bd8b53d5f9d9d3f1c10c3f7cb23e137
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/37446
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: David Neto <dneto@google.com>
This CL reverts the changes to use the top level namer in the various
backends. This is causing issues when rolling into Dawn in the case
where the tint generator is used to create SPIR-V which is sent to
SPIRV-Cross but then generator and inspector aren't used. The entry
points end up being incorrect as SPIRV-Cross gets the renamed entry
points.
Change-Id: I4749e1d773f2bd9edcce83e63555f07a443d5ca5
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/37342
Reviewed-by: dan sinclair <dsinclair@chromium.org>
Commit-Queue: dan sinclair <dsinclair@chromium.org>
Currently Dawn may use the tint generator or the SPIRV-Cross generator.
In the case of SPIRV-Cross, we need to generate the SPIR-V with the
original names otherwise SPIRV-Cross won't be able to match up the entry
point name with the names in the shader.
Change-Id: Ica473030009b282fee352f2d1c1acc93f1db592c
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/37222
Reviewed-by: dan sinclair <dsinclair@chromium.org>
Commit-Queue: dan sinclair <dsinclair@chromium.org>
`DetermineStorageTextureSubtype()` may add types to the module while these types are being iterated over, leading to UB.
To work around, collect all the storage texture types into a vector first, and iterate over these.
Change-Id: Ib94b1df52d6ccbbf635a6d89eeeabef46ba03416
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/37261
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: dan sinclair <dsinclair@chromium.org>
Automatically prints a newline at the end of the last diagnostic in a list. Defaults to true.
Disabled for many tests that assume no newline at end of string.
Change-Id: Id1c2f7771f03f22d926fafc2bebebcef056ac5e8
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/37260
Reviewed-by: dan sinclair <dsinclair@chromium.org>
Commit-Queue: Ben Clayton <bclayton@google.com>
This CL adds an extra constructor to the inspector to change the namer
user. The inspector tests are then updated to use the test namer.
Change-Id: Ibc91de89b52161dc125b38d65e445b5833ad6c18
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/36943
Reviewed-by: Ryan Harrison <rharrison@chromium.org>
This CL changes the generators so we don't have to move the module into
the generator. This will then allow calling the demangler at a later
point and still having access to the module.
Change-Id: Icad16ddb2b89921cbb174cf7fec520c410139285
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/36942
Reviewed-by: Ben Clayton <bclayton@google.com>
Remove the logic around `arrayLength()` being passed anything but a member accessor. Runtime arrays types cannot be used for variables nor parameters.
Add validator logic and test for runtime array parameters.
Adjust the validator error message so it doesn't include the field / variable name. This read weird, and the same information is already provided by the source.
Bug: tint:266
Bug: tint:252
Change-Id: Iecedb0524e10a67b4f8ad15635d67fe61e9d69d9
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/36420
Auto-Submit: Ben Clayton <bclayton@google.com>
Commit-Queue: dan sinclair <dsinclair@chromium.org>
Reviewed-by: dan sinclair <dsinclair@chromium.org>
Fix how handle usage is collected. OpImageQuerySize
is usable for *any* image that is either MS=1 or Sampled=0 or 2.
(Sampled=0 is not allowed by Vulkan).
So it's usable by storage images or multisampled image.
OpImageQuerySizeLod is usable by Sampled=1 MS=0 images, ie. sampled
images.
Fixed: tint:422
Change-Id: Ibfbcab2daa8b38f17758a0428a08cb9660496bac
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/36061
Auto-Submit: David Neto <dneto@google.com>
Commit-Queue: dan sinclair <dsinclair@chromium.org>
Commit-Queue: Ryan Harrison <rharrison@chromium.org>
Reviewed-by: dan sinclair <dsinclair@chromium.org>
Reviewed-by: Ryan Harrison <rharrison@chromium.org>
Helps with test grouping, and more importantly fixes a lint error about
the line length exceeding 80 characters, which cannot otherwise be
sensibly fixed as the formatter wants the fixture name to be unsplit.
Change-Id: I54b32e3516250b4064cd66a618f2239982aaf2d1
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/36381
Auto-Submit: Ben Clayton <bclayton@google.com>
Commit-Queue: dan sinclair <dsinclair@chromium.org>
Reviewed-by: dan sinclair <dsinclair@chromium.org>
Switch all remaining AST stack allocations in tests to using create<T>() which will automatically inject the current Source as the first parameter.
Most remaining uses of Source{} in the codebase are places where we need to fix.
Bug: tint:396
Change-Id: I24655800b50d6ad52e682a7339022972e9b354d9
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/36380
Commit-Queue: Ben Clayton <bclayton@google.com>
Commit-Queue: David Neto <dneto@google.com>
Auto-Submit: Ben Clayton <bclayton@google.com>
Reviewed-by: David Neto <dneto@google.com>
