Fix dependency graph traversal for bitcasts. These were not being traversed, leading to an ICE if the bitcast type was an alias, as the symbol was not resolved for later use by the resolver.
Add missing validation for bitcasts. We were permitting any bitcast that wasn't a being cast to a pointer type, when the spec only allows:
* numeric_scalar to numeric_scalar
* vecN<numeric_scalar> to vecN<numeric_scalar>
Add lots of tests.
Fixed: chromium:1276320
Change-Id: I9e5487ec7649ac543f73fc878e7e282bf932d8cb
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/71681
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Migrate some of the validation logic over to use the results of behavior
analysis.
The most significant changes are:
* Unreachable-statements now consider merge-points of control flow. For
example, if all branches of a if-statement or switch-statement either
return or discard, the next statement will be considered unreachable.
* Unreachable statements are no longer an error, but a warning. See
https://github.com/gpuweb/gpuweb/issues/2378.
* Statements that follow a loops that does not break, or have a
conditional will now be considered unreachable.
* Unreachable statements produced by the SPIR-V reader are now removed
using the new RemoveUnreachableStatements transform.
Some other new changes include additional validation for the continuing
block for for-loops, to match the rules of a loop continuing block.
The new cases this validation is testing for are not expressible in
WGSL, but some transforms may produce complex continuing statements that
might violate these rules. All the writers are able to decay these
complex for-loop continuing statements to regular loops.
Bug: tint:1302
Change-Id: I0d8a48c73d5d5c30a1cddf92cc3383a692a58e61
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/71500
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: David Neto <dneto@google.com>
writer::spirv::Function attempted to append a trailing OpReturn if the
function does not end with a terminating instruction. This wasn't
considering functions that have a non-void return type.
This has now been moved to spirv::Builder::GenerateFunction(),
where we can actually examine the function return type, and generate a
zero-expression to return if we need to.
Note: this was masked by WGSL validation that required all functions to
end with a return statement.
Bug: tint:1302
Change-Id: Iddfeda25a956622c318b8235dc6fc093a2a5c26d
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/71604
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: David Neto <dneto@google.com>
The SwitchCaseBlockStatement was bound to the BlockStatement of an ast::CaseStatement, but we had nothing that mapped to the actual ast::CaseStatement.
sem::CaseStatement replaces sem::SwitchCaseBlockStatement, and has a Block() accessor, providing a superset of the old behavior.
With this, we can now easily validate the `fallthrough` rules directly, instead of scanning the switch case. This keeps the validation more tigtly coupled to the ast / sem nodes.
Change-Id: I0f22eba37bb164b9e071a6166c7a41fc1a5ac532
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/71460
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Some versions of clang warn as error without this:
error: definition of implicit copy constructor for 'EnumSet<tint::ast::PipelineStage>' is deprecated because it has a user-declared copy assignment operator [-Werror,-Wdeprecated-copy]
Change-Id: I02cee0b1f7b70d74f29da0067e888b50acf27ee1
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/71600
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
If we encounter a global twice, the stack vector was not popped, which could trigger false cyclic-dependency errors. Because the cyclic dependency did not actually exist, later logic could break, expecting to find a global->global edge which did not exist.
Fixed: chromium:1273451
Change-Id: I9d99f1aeeaea042d9ed847a878c4717803122240
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/70980
Reviewed-by: David Neto <dneto@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
It's surprising that the fuzzers didn't find this, but I've confirmed that at least one of the fuzzers will fail with this case.
By adding a e2e test case, this will be used as a seed for the fuzzer corpus.
Bug: tint:1321
Change-Id: Ibf4bfa50bf376ae4cba401dfc31a2498fa3b9eec
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/70982
Reviewed-by: James Price <jrprice@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
The default implementation of this was generating random data for
the underlying pointers of std::unordered_map, leading to crashes
when the map was accessed. This CL populates the map in a
structured manner with pseudo-random data.
Bug: chromium:1273001
Change-Id: Ic20ecab85bedba2a59587ebe4a5016be6e53e6f8
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/70701
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Reviewed-by: Ryan Harrison <rharrison@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Austin Eng <enga@chromium.org>
Add transform::Unshadow to renamed shadowed symbols. Required by a
number of other transforms.
Replace Resolver symbol resolution with dep-graph.
The dependency graph now performs full symbol resolution before the
regular resolver pass.
Make use of this instead of duplicating the effort.
Simplfies code, and actually performs variable shadowing consistently.
Fixed: tint:819
Bug: tint:1266
Change-Id: I595d1812aebe1d79d2d32e724ff90de36e74cf4b
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/70523
Reviewed-by: David Neto <dneto@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
They are always used together, and can actually be simplified when
combined.
This transform cannot currently deal with shadowing, and will need
need to depend on a new 'unshadow' transform.
Fixes a long-standing bug where we'd get an ICE if we attempt to inline
a pointer let declaration with side-effects in a for-loop initializer.
Fixed: tint:1321
Bug: tint:819
Change-Id: I236fed688e33a4996e47310b5ece44c991b5249f
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/70661
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: James Price <jrprice@google.com>
This reverts commit 11d09f2fe7.
Reason for revert: Failing roll of Tint to Dawn: https://dawn-review.googlesource.com/c/dawn/+/70100
Original change's description:
> HLSL: force FXC to never unroll loops
>
> Emit the "[loop]" attribute on "for" and "while" so that FXC does not
> attempt to unroll them. This is to work around an FXC bug where it fails
> to unroll loops with gradient operations.
>
> FXC ostensibly unrolls such loops because gradient operations require
> uniform control flow, and loops that have varying iterations may
> possibly not be uniform. Tint will eventually validate that control flow
> is indeed uniform, so forcing FXC to avoid unrolling in these cases
> should be fine.
>
> Bug: tint:1112
> Change-Id: I10077f8b62fbbb230a0003f3864c75a8fe0e1d18
> Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/69880
> Kokoro: Kokoro <noreply+kokoro@google.com>
> Reviewed-by: Ben Clayton <bclayton@google.com>
> Commit-Queue: Antonio Maiorano <amaiorano@google.com>
# Not skipping CQ checks because original CL landed > 1 day ago.
Bug: tint:1112
Change-Id: I8e8f3c0abfa6e6bc5d0e67af9428a46ef867d5c1
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/70540
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Antonio Maiorano <amaiorano@google.com>
'__' is reserved in C++, and the 'match__' and 'build__' functions are causing OSS-fuzz builds to fail.
Add the change in tint behavior to the OT notes.
Add end to end tests for underscores. While the GLSL and MSL compilers seem to accept leading and double underscores in identifiers, the tint build failure has highlighted we have more work to do here (crbug.com/tint/1319)
Fixed: oss-fuzz:41214
Bug: tint:1292
Bug: tint:1319
Change-Id: I32b7bf4e0cff26e678b788457f90452c2503da50
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/70480
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: David Neto <dneto@google.com>
Performs a module-scope (global) declaration dependency analysis, so
that out-of-order global declarations can be re-ordered into dependency
order for consumption by the resolver.
The WGSL working group are currently debating whether out-of-order
declarations should be included in WebGPU V1, so this implementation
currently errors if module-scope declarations are declared out-of-order,
and the resolver does not currently use this sorted global list.
The analysis does however provide significantly better error diagnostics
when cyclic dependencies are formed, and when globals are declared
out-of-order.
The DependencyGraph also correctly now detects symbol collisions between
functions and types (tint:1308).
With this change, validation is duplicated between the DependencyGraph
and the Resolver. The now-unreachable validation will be removed from
the Resolver with a followup change.
Fixed: tint:1308
Bug: tint:1266
Change-Id: I809c23a069a86cf429f5ec8ef3ad9a98246766ab
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/69381
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@chromium.org>
Reviewed-by: David Neto <dneto@google.com>
Vulkan requires round-to-nearest when converting from a float
coordinate to an integer array layer. It prefers round-to-nearest-even.
Use round-to-nearest-even, instad of relying on the rounding behaviour
of the i32(f32) overload.
Fixes: tint:1316
Change-Id: I43624e25e8ea27d3a6e04841a911bbb9418810d0
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/70343
Auto-Submit: David Neto <dneto@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Use imageSize() on images, not textureSize().
In GLSL, the LOD parameter to textureSize() is mandatory for
sampled textures, so emit a default 0 if not supplied. (Also, don't pack
the level into the coords argument; that's an HLSLism.)
GLSL returns the array size of array textures in the final component
of textureSize(); remove it for WGSL.
Write the subtype of storage images correctly (uimage*, iimage*, etc).
This required a bit of cleanup to move "writeonly" ahead of subtype
emission.
Bug: tint:1298
Change-Id: Ica1cec0f833a9b684143c8b0cf6d090fb511a7d2
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/70140
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Stephen White <senorblanco@chromium.org>
Reviewed-by: Ben Clayton <bclayton@google.com>
