Rework the clamping so that it unifies the logic for arrays, matricies
and vectors. Try to preserve constant signess, and only clamp the values
if they're possibly out of bounds.
Use ConstantValue() instead of scanning for ScalarConstantExpressions.
As ConstantValue() improves, so will the performance of robustness.
Change-Id: I013a67e15f43350d0a57bcd8ba9ae0c1bcb1eaec
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/58280
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: David Neto <dneto@google.com>
Auto-Submit: Ben Clayton <bclayton@google.com>
Commit-Queue: David Neto <dneto@google.com>
For loop initializers and continuing statements do not have a BlockStatement as their parent.
Handle removal of these statements with a new Transform::RemoveStatement() helper
Fixed: tint:990
Change-Id: I24e7b18dcf71d3ef0a4d3ee68b9f68518e0eb5e8
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/58063
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: James Price <jrprice@google.com>
Some methods passed by pointer, others by reference. Standarize to pass-by-reference.
Also remove CloneWithStatementsAtStart().
CloneContext::InsertFront() is a better replacement.
Change-Id: Ibbf7caaa7a1b42c2d0a0cddaa3d6e76ca0e12a17
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/58062
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: James Price <jrprice@google.com>
CloneContext::Replace(T* what, T* with) is bug-prone, as complex transforms may want to clone `what` multiple times, or not at all. In both cases, this will likely result in an ICE as either the replacement will be reachable multiple times, or not at all.
The CTS test: webgpu:shader,execution,robust_access:linear_memory:storageClass="storage";storageMode="read_write";access="read";atomic=true;baseType="i32"
Was triggering this brokenness with DecomposeMemoryAccess's use of CloneContext::Replace(T*, T*).
Switch the usage of CloneContext::Replace(T*, T*) to the new function form.
As std::function is copyable, it cannot hold a captured std::unique_ptr.
This prevented the Replace() lambdas from capturing the necessary `BufferAccess` data, as this held a `std::unique_ptr<Offset>`.
To fix this, use a `BlockAllocator` for Offsets, and use raw pointers instead.
Because the function passed to Replace() is called just before the node is cloned, insertion of new functions will occur just before the currently evaluated module-scope entity.
This allows us to remove the "insert_after" arguments to LoadFunc(), StoreFunc(), and AtomicFunc().
We can also kill the icky InsertGlobal() and TypeDeclOf() helpers.
Bug: tint:993
Change-Id: I60972bc13a2fa819a163ee2671f61e82d0e68d2a
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/58222
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: James Price <jrprice@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Move the bulk of the constant evaulation logic out of transform::FoldConstants and into Resolver and sem::Expression.
transform::FoldConstants now replace TypeConstructor nodes that have a constant value on the expression.
This is ground work to:
* Cleaning up the HLSL uniform buffer indexing, which is `/` and `%` arithmatic heavy
* Prepares us to handle `constexpr` when it lands in the spec
* Provide a centralized place to do constant evaluation, instead of the
having similar logic scattered around the codebase.
Change-Id: I3e2f542be692046a8d243b62a82556db519953e7
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/57426
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Reviewed-by: James Price <jrprice@google.com>
Adds a new single-function API for the generators, which applies the
sanitizing transform and performs the generation in one step, and
returns a result object which contains the generated code and success
status/diagnostics.
The new APIs take an `Option` structure to control backend-specific
generation details (e.g. MSL fixed sample mask). The result objects
also provide backend-specific feedback (e.g. whether a UBO of buffer
lengths was generated).
HLSL needs a list of entry points to validate, and it's the HLSL
sanitizer that generates an entry point for programs that do not have
one. This change makes the HLSL generator return the list of
post-sanitize entry points so that the Tint executable can forward
them to the validation code.
Change-Id: I2d5aa27fda95d7c50c5bef41e206aee38f2fd2eb
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/57101
Auto-Submit: James Price <jrprice@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: James Price <jrprice@google.com>
Reviewed-by: Ben Clayton <bclayton@chromium.org>
WGSL:
* Remove vertex_idx and instance_idx.
These are now vertex_index and instance_index.
It seems this was removed once before, then reverted due to CTS
failures, but the original change never landed again.
* Remove the [[set(n)]] decoration. This has been [[group(n)]] for
months now.
API:
* Remove deprecated enums from transform::VertexFormat.
* Remove transform::Renamer constructor that takes a Config. This should
be passed by DataMap.
* Remove ast::AccessControl alias to ast::Access.
Change-Id: I988c96c4269b02a5d77163409f261fd5923188e0
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/56541
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: James Price <jrprice@google.com>
If the array size is greater than a threshold.
This is a work around for FXC stalling when initializing large arrays
with a single zero-init assignment.
Bug: tint:936
Fixed: tint:943
Fixed: tint:942
Change-Id: Ie93c8f373874b8d6d020d041fa48b38fb1352f71
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/56775
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
FXC has trouble dealing with these.
This was originally added to handle returning arrays as structures.
HLSL supports typedefs, which is a much simpiler solution, and doesn't upset FXC.
Bug: tint:848
Bug: tint:904
Change-Id: Ie841c9c454461a885a35c41476fd4d05d3f34cbf
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/56774
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Use the new transforms to try and simplify loops into for-loops.
Emit loops when the initialiser, condition and continuing are simple enough to do so.
Bug: tint:952
Change-Id: I5b3c225b245ffa72996abf6a70f52a9cd25b748e
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/56772
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: David Neto <dneto@google.com>
A Transform that attempts to convert WGSL `loop {}` statements into a for-loop statement.
For-loops cause less broken behavior with FXC than our current loop constructs.
Bug: tint:952
Change-Id: I0b7b1dbec9df4c004b0419d3feae53a195ec79bb
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/56767
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: David Neto <dneto@google.com>
Entry point function parameter can only have builtin and location
attributes (unless a disable validation attribute is present). Other
functions cannot have any decorations on their parameters.
Fix parameter creation in the CanonicalizeEntryPointIO transform for a
case where it was not stripping attributes that are not valid for
function parameters.
Change-Id: I000908aa2dc007c52d100c5f65bca051a49bec9a
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/55863
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: James Price <jrprice@google.com>
Auto-Submit: James Price <jrprice@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Use a DisableValidationDecoration to allow these storage classes only
for variables generated by the SPIR-V sanitizer. This is also used in
the Inspector tests, since the Inspector is currently run *after* the
SPIR-V sanitizer. These tests will be removed when this is no longer
the case.
Also validate that builtin/location decorations are not used on
variables (unless they have input/output storage class).
Fix or delete all of the other tests that were wrongly using these
storage classes and attributes.
Bug: tint:697
Change-Id: I8be7fb16191f5e2bed9f7dfb700e51f3b97fd1fe
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/55862
Auto-Submit: James Price <jrprice@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Describes what Tint system raised the diagnostic.
Use this information in the fuzzers to distinguish between expected and unexpected failure cases in the Transform fuzzer tests.
Fixed: chromium:1206407
Fixed: chromium:1207154
Change-Id: I3b807acafe384a2fc363d2a4165a29693450b3cf
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/55254
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: Ryan Harrison <rharrison@chromium.org>
The CanonicalizeEntryPointIO transform now takes an optional config
option for a fixed sample mask. If there was no sample mask in the
authored shader, add one and return the fixed mask, otherwise AND the
fixed mask with the authored value.
Add a config option to the MSL sanitizer to receive a fixed sample
mask and pass it through to CanonicalizeEntryPointIO.
Bug: tint:387
Change-Id: I1678d915b19a718005d5832c5d624809ee432587
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/55520
Auto-Submit: James Price <jrprice@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Generate a uniform buffer that will receive the lengths of all storage
buffers, and use this to implement calls to arrayLength(). The
transform is provided with a set of mappings from storage buffer
binding points to the corresponding index into the array of buffer
lengths. The transform reports whether it generated the uniform
buffers or not.
Use this transform from the MSL sanitizer, using the binding number as
the index into the array. This matches the behavior of spirv-cross,
and so works with how Dawn already produces this uniform buffer.
Bug: tint:256
Change-Id: I2682d2d024e8daa30f78270b8cfb6bbb32632133
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/54480
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: James Price <jrprice@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
