TypeDecls (alias, structure) are not a types - they declare types.
ast::TypeName is what's used for a ast::Type.
Previously we were trying to automatically convert these to TypeNames in the builder, but having these inherit from ast::Type was extremely error prone.
reader/spirv was actually constructing ast::Structs and using them as types, which is invalid.
Change-Id: I05773ad6d488626606019015b84217a5a55a8e8a
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/53802
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Handle access control on var declarations instead of via [[access]]
decorations. This change does the minimal work to migrate the WGSL
parser over to the new syntax. Additional changes will be needed
to correctly generate defaulted access qualifiers, as well as
validating access usage.
The [[access]] decorations are still supported by the WGSL parser,
with new deprecated warnings, but not for aliases. Example:
var x : [[access(x)]] alias_to_struct;
Making this work is far more effort than I want to dedicate to backwards
compatibility, and I do not beleive any real-world usage will be doing
this.
Still TODO:
* Adding access control as the optional, third parameter to ptr<>.
* Calculating default accesses for the various storage types.
* Validating usage of variables against the different accesses.
Bug: tint:846
Change-Id: If8ca82e5d16ec319ecd01f9a2cafffd930963bde
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/53088
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: James Price <jrprice@google.com>
Reviewed-by: David Neto <dneto@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Sometimes a value can come from a block which does not appear
in the structured block order. That block will never execute,
so we can safely use the null value instead.
Bug: tint:804
Change-Id: Idc1a6c4f76f14a1d76919a95e5a65e56018ce68f
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/52840
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>
MaybeEmitCombinatorialValue() didn't check that MakeOperand() or RectifySecondOperandSignedness() didn't error, leading to ICEs.
Bug: crbug.com/tint/804
Change-Id: Ic78487a70a591e718c7b5936c6678e7a19c4b626
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/51927
Commit-Queue: Ben Clayton <bclayton@google.com>
Auto-Submit: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: David Neto <dneto@google.com>
- Translates from OpImageSampleDrefExplicitLod, but the Lod must
be a constant 0, or the reader issues an error.
The requirement for Lod 0 is a constraint from Metal, inherited
by WGSL.
Fixed: tint:425, tint:482
Change-Id: I8fdf10dbd9c5ac3e24398519a28202c84677c38d
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/51803
Commit-Queue: David Neto <dneto@google.com>
Auto-Submit: David Neto <dneto@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
This change implements pointers and references as described by the WGSL
specification change in https://github.com/gpuweb/gpuweb/pull/1569.
reader/spirv:
* Now emits address-of `&expr` and indirection `*expr` operators as
needed.
* As an identifier may now resolve to a pointer or reference type
depending on whether the declaration is a `var`, `let` or
parameter, `Function::identifier_values_` has been changed from
an ID set to an ID -> Type* map.
resolver:
* Now correctly resolves all expressions to either a value type,
reference type or pointer type.
* Validates pointer / reference rules on assignment, `var` and `let`
construction, and usage.
* Handles the address-of and indirection operators.
* No longer does any implicit loads of pointer types.
* Storage class validation is still TODO (crbug.com/tint/809)
writer/spirv:
* Correctly handles variables and expressions of pointer and
reference types, emitting OpLoads where necessary.
test:
* Lots of new test cases
Fixed: tint:727
Change-Id: I77d3281590e35e5a3122f5b74cdeb71a6fe51f74
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/50740
Commit-Queue: Ben Clayton <bclayton@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: David Neto <dneto@google.com>
Use this to check that type and expr are both not nullptr.
Cleans up the random mix of checking `expr`, `type` or both.
Change-Id: I84eb4a16e0dc22e18d6868f62c0314effe9858a2
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/51184
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: David Neto <dneto@google.com>
There's now one remainin use of the semantic info in the WGSL writer - the transformation of [[offset]] into padded fields.
This does not belong in the WGSL writer, but instead part of the transform::Wgsl sanitizer.
Bug: tint:798
Change-Id: I95ba11f022c41150cc12de84a4085cd7d42019fe
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/50822
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: David Neto <dneto@google.com>
Reviewed-by: James Price <jrprice@google.com>
To declare a local variable, we write `var name : type`, not `var<function> name : type`.
This change fixes all the places where we were feeding StorageClass::kFunction into variable declarations.
Note that the resolved, semantic variable correctly infers the `kFunction` StorageClass.
Change-Id: I6221fabae1de0435044f29b9a91808421d5cace6
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/50821
Commit-Queue: Ben Clayton <bclayton@chromium.org>
Commit-Queue: David Neto <dneto@google.com>
Auto-Submit: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: David Neto <dneto@google.com>
WGSL only has vector normalize, not scalar.
Scalar normalize is always 1.0, so return that directly.
Bug: tint:765
Change-Id: I08332c20922f5834f65284b9aaeb22995423171d
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/50603
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: David Neto <dneto@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
We don't want the WGSL parser to have to maintain type lookups.
If the WGSL language is updated to allow module-scope variables to be declared in any order, then the single-pass approach is going to fail horribly.
Instead do the check in the Resovler.
With this change, the AST nodes actually contain the correctly declared storage class.
Fix up the SPIR-V reader to generate StorageClass::kNone for handle types.
Fix all tests.
Bug: tint:724
Change-Id: I102e30c9bbef32de40e123c2676ea9a281dee74d
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/50306
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Reviewed-by: James Price <jrprice@google.com>
Don't create disjoint AST type nodes.
Instead use a new bespoke type hierarchy that can Build() the required
AST nodes.
Change-Id: I523f97054de2c553095056c0bafc17c48064cf53
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/49966
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: David Neto <dneto@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
As() was being called on a nullptr object.
Change-Id: Iedd7f1c57d12f1ad5ea2af60317817841f5d1392
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/50303
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Instead of a typ::TypePair.
Bug: tint:724
Change-Id: Ife8fac86093eb128bc98974e1f6614a73c42a9e6
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/49753
Commit-Queue: Ben Clayton <bclayton@chromium.org>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
This is a step to being able to dump the SPIR-V for
successfully converted modules.
Bug: tint:756
Change-Id: Ifa48c88835ff10824c542a4e13d8a2d3a7f7f484
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/49604
Commit-Queue: David Neto <dneto@google.com>
Auto-Submit: David Neto <dneto@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
typ::Types originating from ConvertType() will hold direct pointers to ast::Struct and ast::Aliase. These must not be used directly. Instead TypeNames should be created to refer to these.
Bug: tint:724
Change-Id: I7f511e0da56e5d3ac587cdb22b0baf38474bed1e
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/49746
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
The spirv parser now creates ast types along with sem types via
typ::Type. All sem::Type* were replaced with typ::Type, and its `ast`
member is used over the `sem` member to make it easier to migrate to
ast-only.
The parser was written to take advantage of the fact that types were
resolved to semantic types during parsing. For instance, a mapping of
spirv typeid to sem::Type* was used throughout (`id_to_type_`) to
resolve types once, and to support type aliasing. Since the goal is to
only create AST types, and to resolve only in the Resolver, I made many
changes to remove this dependency on semantic types. For instance, we
now always call ConvertType(typeid) instead of looking up via
id_to_type. Similarly, the `signed_type_for_` and `unsigned_type_for_`
maps were replaced with `UnsignedTypeFor` and `SignedTypeFor` functions.
Bug: tint:724
Change-Id: I3aee3928834febd71b473d6a8d8cb77b1ac94e21
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/49542
Reviewed-by: David Neto <dneto@google.com>
Commit-Queue: Antonio Maiorano <amaiorano@google.com>
A literal has an implicit type, so there should be no type on the AST node.
This highlighted that the resolver was nto canonicalizing TypeConstructorExpression types, which has been fixed.
This required preservation of the declared type name in order for error messages to contain aliased names.
Bug: tint:724
Change-Id: I21594a3e8a0fb1b73c6c5b46a14b8664b7f28512
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/49345
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Ben Clayton <bclayton@chromium.org>
This is a waste of memory, and now fires a TINT_ASSERT() in the resolver.
Add some more information to the resolver assertion message so that its easier to identify the node. This is especially useful when there's no source information available.
Fixed: tint:740
Bug: tint:469
Change-Id: I0cd4529db7b3906e64da6ed7290163509eb0c3f2
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/48689
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: David Neto <dneto@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Not currently called (nothing currently attaches ast::Type nodes to the AST), but implements some of the boilerplate that'll be shortly required.
This change also removes a bunch of duplicated enumerators from the sem namespace for their counterpart in the ast namespace.
Bug: tint:724
Change-Id: I0372a9f4eca2f9357ff161e7ec1b67eae1c4c8f6
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/48603
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Copy all of the type classes from src/type into ast.
Required the merging of:
* type::Struct into the existing ast::Struct - ast::Struct now has a name.
* type::AccessControl into the existing ast::AccessControl enumerator - The old ast::AccessControl enumerator is now ast::AccessControl::Access
Bug: tint:724
Change-Id: Ibb950036ed551ec769c6d3d2c8fb411809cf6931
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/48383
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Handle the case where the OpBranchConditional in a loop header
branches to two distinct blocks inside the loop construct.
This is an if-selection in disguise.
Create an kIfSelection with the same set of blocks as the kLoop,
and with the continue target as the merge.
Fixed: tint:524
Change-Id: I5150d19a2b4388da409e2da6e68ffafdc5d21a9a
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/47560
Commit-Queue: David Neto <dneto@google.com>
Auto-Submit: David Neto <dneto@google.com>
Reviewed-by: Alan Baker <alanbaker@google.com>
EmitControlBarrier() was calling MakeOperand() to create ast::Expressions that held a ScalarConstructorExpression, which held a IntLiteral.
The literal value was then taken, and the rest was discarded.
Detached AST nodes will become an ICE, so do the work to find the literal value.
Bug: tint:469
Change-Id: I522bfe8db84e853e189c714b18598feb0d49e58b
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/48049
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: James Price <jrprice@google.com>
When finding the true-head, false-head, and potentially the
premerge-head blocks of an if-selection, there was an overly
aggressive check for the true-branch or false-branch landing
on a merge block interior to the if-selection. The check was
determining if the merge block actually corresponded to the selection
header in question. If not, then it was throwing an error.
The bug was that this check must be performed only if the
target in question is actually inside the selection body.
There are cases where the target could represent a structured
exit, e.g. to an enclosing loop's merge or continue, or
an enclosing switch construct's merge.
There is still a latent bug: if either the true branch
or false branch represent such a kLoopBreak, kLoopContinue, or
kSwitchBreak edge, then those are not properly generated.
That will be fixed in a followup CL.
Bug: tint:243, tint:494
Change-Id: I141cce07fa0a1dfe5fad20dd2989315e4cd7b688
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/47482
Auto-Submit: David Neto <dneto@google.com>
Commit-Queue: Alan Baker <alanbaker@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Alan Baker <alanbaker@google.com>
Use TINT_ICE() where we have diagnostics, TINT_ASSERT() where we do not.
Change-Id: Ic6e842a7afdd957654c3461e5d03ecec7332e6f9
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/46444
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
This was updated/clarified by the SPIR WG.
Bug: tint:3
Change-Id: Ie4c503f0e5f80ffeabada9c526375588e81a5ceb
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/45740
Auto-Submit: David Neto <dneto@google.com>
Commit-Queue: Alan Baker <alanbaker@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Alan Baker <alanbaker@google.com>
Add a return type decoration list field to ast::Function.
Bug: tint:513
Change-Id: I41c1087f21a87731eb48ec7642997da5ae7f2baa
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/44601
Commit-Queue: James Price <jrprice@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
