The TypeConstructorExpression logic that tested for splats was not considering references. This led to broken emission for the SPIR-V and HLSL backends.
Fixed: tint:992
Change-Id: I9824b71f526997f91d380c09b459f4fd73065b19
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/58397
Auto-Submit: Ben Clayton <bclayton@google.com>
Commit-Queue: David Neto <dneto@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: David Neto <dneto@google.com>
We were not printing source information for errors involving call
statements.
Change-Id: I576b4d095162333d4392d6be590e827a49918d3d
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/58520
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>
This is defensive. Without variable pointers capabilities, this is
definitely invalid, but not yet checked by the SPIRV-Tools validator.
Bug: tint:807
Change-Id: If9b0b19573b1ca14a1c55aa20c9d42784ec12568
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/56700
Auto-Submit: David Neto <dneto@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: David Neto <dneto@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
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>
Check they do not contain returns, discards
Check they do not directly contain continues, however a nested loop can have its own continue.
Bug: chromium:1229976
Change-Id: Ia3c4ac118ffdaa6cca6025366c19f9897718c930
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/58384
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: David Neto <dneto@google.com>
Commit-Queue: David Neto <dneto@google.com>
Auto-Submit: Ben Clayton <bclayton@google.com>
Arrays can be extremely large, and having the load and store functions unroll the elements can make the complier explode.
Fixed: chromium:1229233
Change-Id: Ieb5654254e16f5ce724a205d21d954ef9a0cd053
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/58382
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@chromium.org>
Reviewed-by: David Neto <dneto@google.com>
Auto-Submit: Ben Clayton <bclayton@google.com>
Have postfix_expression() use a loop instead of recursively calling itself.
Fixed chromium:1229669
Change-Id: Ied7684d00ba453a8b89ab0251d42e2a72169421f
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/58381
Auto-Submit: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: David Neto <dneto@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
NoPerspective interpolation maps to 'linear'
Centroid maps to 'centroid'
Sample maps to 'sample'
Otherwise, allow 'center' to be defaulted.
Fixed: tint:935
Change-Id: I0b040da0c57d2a363f9dc9474c1ac889e0fe2278
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/56840
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: James Price <jrprice@google.com>
Commit-Queue: James Price <jrprice@google.com>
Auto-Submit: David Neto <dneto@google.com>
Setting precision to `std::numeric_limits<float>::max_digits10` is valid
when using the `scientific` floatfield format when printing values.
However, we used `fixed` to make our floats more human-readable. This
change keeps the output in `fixed`, except if doing so loses precision,
in which case we fall back to `scientific`.
This fixes the rendering differences seen in the Babylon.js examples
(https://crbug.com/tint/944) between Dawn using Tint vs SPIRV-Cross, as
Tint's output was emitting values that had lost too much precision
(e.g. very small numbers being output as 0).
Bug: tint:944
Change-Id: I8deea23ad876825bbe390fc26907d4bbbd4b966e
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/58321
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@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>
Much like sem::Type, it greatly simplifies downstream logic if we can compare sem::Intrinsic pointers to know if they refer to the same intrinsic overload.
Change-Id: If236247cd3979bbde821d9294f304ab85ba4938e
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/58061
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: David Neto <dneto@google.com>
And call these helpers instead of inlining complex statements.
Cleans up output, and helps prevent for-loops decaying to while loops.
Change-Id: I6ac31b18ce6c5fac0e54e982f7db3bb298f7edb2
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/58060
Auto-Submit: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@chromium.org>
Reviewed-by: David Neto <dneto@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>
Before and after each transform of a transform::Manager.
This change makes the transforms use the Castable system so we can get the transform name.
Change-Id: I2cf9335960f1aca56f1d32c5ba0d104db4cc46c1
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/54581
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ryan Harrison <rharrison@chromium.org>
Reviewed-by: James Price <jrprice@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Replace(T* what, T* with) is bug-prone, as more 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.
This is the cause of some of the CTS failures reported in crbug.com/tint/993.
Bug: tint:993
Change-Id: I880ece45faab0e7f07230a1b4436f4e9846edc84
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/58221
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: James Price <jrprice@google.com>
Commit-Queue: Ben Clayton <bclayton@chromium.org>
Commit-Queue: Ben Clayton <bclayton@google.com>
This is a reland of fd5829e5ea
Original change's description:
> Validate storage class constraints
>
> As defined by https://gpuweb.github.io/gpuweb/wgsl/#storage-class-layout-constraints
>
> Bug: tint:643
> Change-Id: I9c78ba69a792a80c263a17b0a6e9b4810fdb7f30
> Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/56780
> Kokoro: Kokoro <noreply+kokoro@google.com>
> Commit-Queue: Antonio Maiorano <amaiorano@google.com>
> Reviewed-by: Ben Clayton <bclayton@google.com>
Bug: tint:643
Change-Id: I62036d615a062597339a9d130b7ccf49b5be26c7
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/58120
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Commit-Queue: Antonio Maiorano <amaiorano@google.com>
Clamp the `level` and `array_index` arguments to `textureLoad()` and `textureStore()`.
Also fix the off-by-one error for the coordinates.
See: https://github.com/gpuweb/gpuweb/pull/1906
Fixed: tint:748
Change-Id: Id7505578b632dcaf75b2a3a020fc0190c612972c
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/57700
Kokoro: Kokoro <noreply+kokoro@google.com>
Auto-Submit: Ben Clayton <bclayton@google.com>
Reviewed-by: David Neto <dneto@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
It appears that I didn't do a great job cleaning up the removal of ShareableCloneable in https://dawn-review.googlesource.com/c/tint/+/51484.
Cloning nodes shouldn't return the same pointer. Remove bad comments.
Clean up leftover logic from CloneWithoutTransform().
Change-Id: Ibbc5f625c5978e9c11da59e7aa6197f39b6f0363
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/58220
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: James Price <jrprice@google.com>
Commit-Queue: James Price <jrprice@google.com>
Auto-Submit: Ben Clayton <bclayton@google.com>
This uses FXC compilation failure mitigation for _any_ vector index assignment that has a non-constant index. FXC can still fall over if the loop calls a function that performs the dynamic index.
Use some vector swizzle logic to avoid branches in the helper.
Fixed: tint:980
Change-Id: I2a759d88a7d884bc61b4631cf57feb4acc8178de
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/57882
Auto-Submit: Ben Clayton <bclayton@google.com>
Commit-Queue: Antonio Maiorano <amaiorano@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
The UBO must have a stride that is a multiple of 16 bytes.
Note that this change was part of https://dawn-review.googlesource.com/c/tint/+/56780
but the CL was reverted because it broke Dawn. This CL relands part of
the change, and adds the macro TINT_EXPECTS_UBOS_TO_BE_MULTIPLE_OF_16 so
that Dawn can conditionally compile against it.
Bug: tint:984
Bug: tint:643
Change-Id: I303b3fe81ff97c4933c489736d5d5432a59ce9b7
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/57921
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Antonio Maiorano <amaiorano@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
This was just echoing back the source that was provided to the
validation routine, and is no longer uses by any callers.
Change-Id: I55aeb7c41405ebbe1b21b9654b94607b651ff118
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/57281
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
This is a reland of the CL:
https://dawn-review.googlesource.com/c/tint/+/55402
Now that sanitizers are no longer exposed externally, the Inspector no
longer needs to handle this post-sanitizer world.
Bug: tint:697
Change-Id: Ic02ebb9c529aa132a238285bdd0d0df8686e219b
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/57104
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Make the sanitizer transforms internal, as the new generator API
automatically runs them.
Change-Id: Ia2674762328b5d91d8370b8c18c31693936e8566
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/57102
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
This change introduces sem::CompoundStatement, a new base class for
statements that can hold other statements.
sem::BlockStatements now derives from sem::CompoundStatement, and
this change introduces the following new CompoundStatements:
* `sem::IfStatement`
* `sem::ElseStatement`
* `sem::ForLoopStatement`
* `sem::LoopStatement`
* `sem::SwitchStatement`.
These new CompoundStatements are now inserted into the semantic
tree as now documented in `docs/compound_statements.md`.
The `sem::BlockStatement::FindFirstParent()` methods have been
moved down to `sem::Statement`.
The `Resolver::BlockScope()` method has been replaced with
`Resolver::Scope()` which now maintains the `current_statement_`,
`current_compound_statement_ ` and `current_block_`. This
simplifies statement nesting.
The most significant change in behavior is that statements now
always have a parent, so calling Block() on the initializer or
continuing of a for-loop statement will now return the
BlockStatement that holds the for-loop. Before this would
return nullptr.
Fixed: tint:979
Change-Id: I90e38fd719da2a281ed9210e975ab96171cb6842
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/57707
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: James Price <jrprice@google.com>
Case statements cannot be used as regular statements. They must exist inside a switch.
Change-Id: Ica2e0573ad926c3e257f02c95397a8666dfcd547
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/57705
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: James Price <jrprice@google.com>
Commit-Queue: James Price <jrprice@google.com>
Auto-Submit: Ben Clayton <bclayton@google.com>
This class is very large, so factoring it out into its own seperate
file.
Also renaming it to InspectorBuilder, in anticipation of adding tests
that don't build their state through explicitly calling into the
ProgramBuilder.
BUG=tint:967
Change-Id: I316458a969479200edf6962ab2008180f9a1e7e2
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/57800
Auto-Submit: Ryan Harrison <rharrison@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ryan Harrison <rharrison@chromium.org>
Reviewed-by: Ben Clayton <bclayton@google.com>
ConstantValueOf() obtains the constant value from the logic in resolver_constants.cc. This is better tested, and is the foundation of Tint's constant folder.
Change-Id: I42036f3ff4ab684b4864cd69856de1715b38d246
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/57702
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: James Price <jrprice@google.com>
Use the new semantic constant value information to significantly reduce the complex indexing logic emitted for UBO accesses.
This will dramatically reduce the number of `for` loops that are decayed to `while` loops.
Change-Id: I1b0adb5edde2b4ed39c6beafc2e28106b86e0edd
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/57701
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: James Price <jrprice@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>
This reverts commit fd5829e5ea.
Reason for revert: Temporarily reverting as this is preventing a tint->dawn roll, which is needed to fix the dawn->chrome roll.
Original change's description:
> Validate storage class constraints
>
> As defined by https://gpuweb.github.io/gpuweb/wgsl/#storage-class-layout-constraints
>
> Bug: tint:643
> Change-Id: I9c78ba69a792a80c263a17b0a6e9b4810fdb7f30
> Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/56780
> Kokoro: Kokoro <noreply+kokoro@google.com>
> Commit-Queue: Antonio Maiorano <amaiorano@google.com>
> Reviewed-by: Ben Clayton <bclayton@google.com>
TBR=bclayton@google.com,amaiorano@google.com,noreply+kokoro@google.com,tint-scoped@luci-project-accounts.iam.gserviceaccount.com
Change-Id: I8dbd0e46b3e3291ef08797a196d0d9abd2a78845
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: tint:643
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/57704
Reviewed-by: Ben Clayton <bclayton@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Ben Clayton <bclayton@google.com>
... once the maximum number of errors have been reached.
https://dawn-review.googlesource.com/c/tint/+/56070 introduced maybe_set_synchronized(), which only set synchronized_ when the number of errors reported was less than max_errors_, but it seems the fuzzers have found ways to generate an excessive number of errors that keep the parser synchronized.
Revert 56070, and instead check the synchronized state along with the error count for every unbounded loop in the parser.
Fixed: chromium:1226655
Fixed: chromium:1226379
Change-Id: I178d758ac1424d4d19923fe6a3d9e123879b9eae
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/57427
Auto-Submit: Ben Clayton <bclayton@google.com>
Commit-Queue: James Price <jrprice@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: James Price <jrprice@google.com>
Report whether one was generated so that Dawn knows to use the
`-fpreserve-invariance` compiler option.
Bug: tint:772
Change-Id: Ife1eb05265646727dc864f12f983781af4df3777
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/57644
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Make sure the other backends ICE on unrecognized attributes.
Add E2E tests, currently skipped for the other backends.
Bug: tint:772
Change-Id: I4e68d111ff79b19ebb6c574058a91debcb746011
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/57642
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
This is only valid on structure members and entry point return
types. Additionally, this can only be applied to a position builtin.
Bug: tint:772
Change-Id: Iffd774ca768ab66373678ecf0eb57c766767e92b
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/57641
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
This CL removes the old token `Is<Name>()` methods which are no longer
being used in the parser or lexer. (Some of them were used in tests but
the generic `Is(<type>)` is used instead.
New `peek` methods are added which will peek if the given token comes
either next or at the given index.
Change-Id: I8d15f2a42889f981d273b6459d20c4202db1ae32
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/57540
Auto-Submit: dan sinclair <dan.sinclair@gmail.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@chromium.org>
Reviewed-by: Ben Clayton <bclayton@google.com>
Commit-Queue: Ben Clayton <bclayton@chromium.org>
By generating a helper function for these, we can keep the atomic expression pre-statement-free. This can help prevent for-loops from being transformed into while loops.
Change-Id: Id034ea5ea9be601661ddb78db973015d845c420f
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/57463
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Auto-Submit: Ben Clayton <bclayton@google.com>
Reviewed-by: James Price <jrprice@google.com>
Ben Clayton
James Price
David Neto
Sarah
Antonio Maiorano
Ryan Harrison
dan sinclair