We already compute the "first" and "last" basic block that
uses a value, so we could know when to hoist a value into
a var declaration. You have to do this sometimes to make
sure all uses are in scope of the declaration.
Until now we tracked Phis with an entirely different mechanism.
But there are cases which broke down. That's what happens
in crbug.com/tint/1649.
Additionally, GraphicsFuzz cases generarte similar weirdness.
Also, be more careful about ensuring that the assignments
generated to feed phis behave as if they occur in parallel.
Within a single batch of such assignments, generate and
use intermediate let-declarations for phis that that batch
will overwrite.
Also, unwrap-references when rectifying the signedness of
binary operators.
Skip tests that fail due to crbug.comt/tint/98:
test/tint/bug/tint/749.spvasm.*
Fixed: tint:1649
Change-Id: I7314c351b74a10bfa9a18011f3d80a520568011c
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/101220
Auto-Submit: David Neto <dneto@google.com>
Commit-Queue: David Neto <dneto@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Since GLSL ES does not support the offset= attribute, struct members
with explicit @align or @size attributes require adding explicit
padding members. This in turn requires rewriting any constructor
calls to initialize the new padding to zero, handled in the same
transform.
Note that this is currently overly-verbose, and will add padding where
GLSL doesn't technically need it (e.g., padding a vec3 out to 16 bytes).
Bug: tint:1415
Change-Id: Ia9ba513066a0e84f4c43247fcbbe02f5fadd6630
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/101720
Reviewed-by: Ben Clayton <bclayton@google.com>
Commit-Queue: Stephen White <senorblanco@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
These can be used in all sorts of places which are not tracked by sem::Function::TransitivelyReferencedGlobals().
As they're not emitted as variables by any backend, just preserve them.
Fixed: tint:1598
Change-Id: I2696486cb2ffe8408bd5dd3090d7d600ca1d170f
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/101481
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
Auto-Submit: Ben Clayton <bclayton@google.com>
Modify the AddSpirvBlockAttribute transform to fix top-level structure
access of uniform, storage and push-constant buffers for use in the
GLSL backend. The small change to the transform makes the transform
wrap host-sharable buffers, if they're also used as a
non-host-sharable structure. Also rename the transform to
AddBlockAttrbibute in order to reflect its wider applicability.
Change-Id: Ib2bf4ebf6bce72790791dbae9387032be765e4b9
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/101061
Reviewed-by: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Stephen White <senorblanco@chromium.org>
If the let initializer was an abstract numeric that was implicitly materialized to a concrete type, then we could inlining the initializer into the use without the implicit materialization cast.
This could lead to validation errors, and subtly different results in the program.
In this situation, add an explicit cast to keep the types the same before and after inlining.
Fixed: tint:1664
Change-Id: Icca980cf8af74673906ad6c681a6b07d0c1932fd
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/101160
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
This transform breaks up matNx2<f32> matrices used in uniform buffers
into column vectors, which fixes std140 layout rules.
Used by the SPIR-V and GLSL backends.
Re-enable tests that were disabled for these cases.
Bug: tint:1632
Change-Id: I596d016582b4189a0b413d762b3e7eabd3504b22
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/100907
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
Commit-Queue: Ben Clayton <bclayton@chromium.org>
This patch fix a bug in DecomposeMemoryAccess, allowing access index 0
of a member in uniform and storage buffer being recognized as constant
index.
Bug: tint:1652
Change-Id: Ia428de17c860bdafe87c3af9e46426c74fe8fd68
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/99480
Commit-Queue: Zhaoming Jiang <zhaoming.jiang@intel.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Auto-Submit: Zhaoming Jiang <zhaoming.jiang@intel.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
PromoteInitializersToLet was not handling sem::Materialize nodes.
This can happen for const arrays when they are dynamically indexed.
Fixed: tint:1653
Change-Id: I3d67d8139e481c89b31a3a30c7ef44384b7545ba
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/99500
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
This CL changes the MSL emission for struct initializers to emit the
struct name first.
`const a = {.f=float3(1)}` becomes `const a = Normals{.f=float3(1)}`.
This fixes an issues where the initialization happens inside an array
which the downstream compiler rejected without the explicit struct
naming.
Bug: tint:1641
Change-Id: I948b9ca94f4b89eac6d5bbbaa615b3d71d50c737
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/98760
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
Reviewed-by: Ben Clayton <bclayton@google.com>
If a call to atan2 with args of type AFloat and AInt is made, Resolver
would correctly select the atan2(AFloat, AFloat) overload, but if the
input args were of type (AFloat, AInt), it would attempt to constant
evaluate without first converting the AInt arg to AFloat. The same would
occur for a binary operation, say AFloat + AInt. Before constant
evaluating, the Resolver now converts AInt to AFloat if necessary.
Bug: chromium:1350147
Change-Id: I85390c5d7af7e706115278ece34b2b18b8574f9f
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/98543
Reviewed-by: Ben Clayton <bclayton@google.com>
Commit-Queue: Antonio Maiorano <amaiorano@google.com>
This will be needed by an upcoming CL that fixes call statements to
builtins that return a constant value. Without this transform, the
constant value reaches the backend, where we don't currently deal with
abstract numbers. Note that the other backends (HLSL, MSL, GLSL) already
use this transform.
Bug: chromium:1350147
Change-Id: Icc1f1416a07db228f8e3f39851a9ac079c48319f
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/98740
Commit-Queue: Antonio Maiorano <amaiorano@google.com>
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
This patch implement modf and frexp built-ins for f16 types, and also
simplify their implementation for f32 in MSL and HLSL, and clean up
deprecated code in GLSL writer. Corresponding unittests are also
implemented, but end-to-end tests for f16 are not implemented yet.
Bug: tint:1473, tint:1502
Change-Id: I12887ae5303c6dc032a51f619e1afeb19b4603b6
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/98102
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Zhaoming Jiang <zhaoming.jiang@intel.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
This CL updates the list of reserved words to match the WGSL spec. The
use of a reserved word is changed from an error to a deprecation at the
moment be cause the majority of the list would be new errors.
Bug: tint:1633 tint:1624
Change-Id: I498db41689cdd666dfb291b1a6761a1182c87ec8
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/98042
Reviewed-by: David Neto <dneto@google.com>
Reviewed-by: Ben Clayton <bclayton@chromium.org>
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
This transform was attempting to remove builtins with no side
effects, such as a call with abstract int/float args, which is unhandled
by this transform. For example, this would cause the transform to ICE:
_ = clamp(1, 2, 3);
Fixes ClusterFuzz issue crbug.com/1348739.
Bug: chromium:1348739
Change-Id: Ie355eb36c6c020417c2d93f2dc434c11dbb72d1f
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/97880
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Antonio Maiorano <amaiorano@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Remove unused enum entries.
Add a ParseStorageClass() function. Have the WGSL parser use this.
First step to standardizing the way we parse enum-backed token sets.
Change-Id: I31c02816493beeabda740ff43946edce097f5fd1
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/97148
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
Replace the temporary file name with 'shader.hlsl', so that
skip-expectations can be stably re-generated.
Change-Id: I5ead2235e6e0d84ad67c8d90f8d06b812c8fd593
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/97145
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Change tint's `--fxc` flag to take the path of the FXC compiler DLL.
Have tint attempt to validate with both FXC and DXC if `--validate` is
passed.
Fix the 'dirsWithNoPassExpectations' logic which looks like it got
broken with the tint -> dawn merge. It also incorrectly applied
filepath.FromSlash() on windows.
Change-Id: I0f46aa5c21bc48a2abc48402c41f846aff4a8633
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/96800
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Ben Clayton <bclayton@chromium.org>
If an abstract-vector or abstract-matrix is indexed with a non-constant index expression, then the resulting value is non-constant, and so cannot be abstract.
In this situation the materialization cannot be done post-index, so materialization must happen on the object before indexing.
Bug: chromium:1345468
Change-Id: I9f29dc40301779a7ff8f173724374bd845a3a5b9
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/96684
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@chromium.org>
SPIRV-Val has tightended up validation around input / output interpolation decorations.
This change ensures that the parser and writer do the right thing.
Change-Id: I29c97fdcc48c62aa77b106c42e64fbc54204d607
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/96020
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
These have not been in the spec for a long time. The read_write access
mode can be used instead.
Fixed: tint:1342
Change-Id: I01ffc343d2d2f9df9d7028bba4548c749616c65c
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/93500
Reviewed-by: Ben Clayton <bclayton@google.com>
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
Enable the parsing of 'const'.
Warn on use of module-scope 'let', and automatically replace with 'const'.
Fixed: tint:1580
Change-Id: I214aabca80686dc6b60ae21a7a57fbfb4898ea83
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/93786
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
If a module-scope private variable is only referenced within a single
function, promote it to a function scope declaration instead of
passing it as a parameter. This reduces the number of a function
parameters that are needed in some cases.
Bug: tint:1509
Change-Id: I8951f6216bc7e4cf5abfda314bea1e9ed3ded560
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/94002
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
Once 'const' is introduced, let will no longer resolve to a
creation-time constant value.
Add temporary code into each of the writers to prevent constants that
originate from a 'let' from being inlined. This will reduce the amount
of noise in later CLs.
Bug: tint:1580
Change-Id: Id3493a43ac09fe9f042ff2d517d04b2ae854d43e
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/94541
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: David Neto <dneto@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@chromium.org>
And remove the WrapArraysInStructs transform.
Wrapping arrays in structures becomes troublesome for `const` arrays, as
currently WGSL does not allow `const` structures.
MSL 2.0+ has a builtin array<> helper, but we're targetting MSL 1.2, so
we have to emit our own. Fortunately, it can be done with a few lines of
templated code.
This produces significantly cleaner output.
Change-Id: Ifc92ef21e09befa252a07c856c4b5afdc51cc2e4
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/94540
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@chromium.org>
Reviewed-by: David Neto <dneto@google.com>
https://github.com/gpuweb/gpuweb/issues/3056 proposes that creation-time
constant expressions should support arrays. We have verbal agreement to
update to spec to support const arrays.
Bug: tint:1580
Change-Id: If460f729384d700a1c2149c5b7dbe8613a960184
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/94330
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: David Neto <dneto@google.com>
Aside from more SPIR-V using OpConstantNull, this is a no-op change.
By refactoring these methods, they're easier to read, and contain less
lambda-heavy code.
Change-Id: I89c26b2b9f1cd0785d86fb3293f7cfda550bef2e
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/94331
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
Commit-Queue: Ben Clayton <bclayton@google.com>
If the object and index are both constant expressions, resolve the
index as a constant.
Note: Expectations have been updated for indexing on 'let's. Once
'const' is introduced, 'let' will no longer be treated as a constant
expression, and these expectations will be reverted.
Bug: tint:1580
Change-Id: I42793d36c1a5f82890ccaa5dbbb71b4346493bef
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/94329
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
Commit-Queue: Ben Clayton <bclayton@chromium.org>
Add the new classes:
* `ast::Let`
* `ast::Override`
* `ast::Parameter`
* `ast::Var`
Limit the fields to those that are only applicable for their type.
Note: The resolver and validator is a tangled mess for each of the
variable types. This CL tries to keep the functionality exactly the
same. I'll clean this up in another change.
Bug: tint:1582
Change-Id: Iee83324167ffd4d92ae3032b2134677629c90079
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/93780
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
Commit-Queue: Ben Clayton <bclayton@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
This CL updates all of the Tint unittests to the new @stage shorter
syntax. This also updates the WGSL writer to emit the new short forms
instead of using the deprecated form.
Bug: tint:1503
Change-Id: I8c49e5319a19cccb5b4b5078f3ab39c50f31a9a8
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/92483
Reviewed-by: Ben Clayton <bclayton@google.com>
Commit-Queue: Dan Sinclair <dsinclair@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
The abstract-numeric work has fixed this, by removing the error for
small floating point numbers.
Fixed: tint:1332
Change-Id: Ibf7382795210220d343d62fe22d201cf67db1ccc
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/92244
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
Commit-Queue: Ben Clayton <bclayton@chromium.org>
This is required to handle materialized values, and for constant
expressions.
Bug: tint:1504
Change-Id: If0a49e9b03566c06aa6e4e4c284fc427e1541e91
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/92082
Reviewed-by: Dan Sinclair <dsinclair@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@chromium.org>
This is required to handle materialized values, and for constant
expressions.
Bug: tint:1504
Change-Id: I79ad567954de2d1cfea09dda255894e4e2aa678e
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/92081
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Ben Clayton <bclayton@chromium.org>
This is required to handle materialized values, and for constant
expressions.
Bug: tint:1504
Change-Id: Ic3ac62317241fa6f7009360128f222aeb56f62e4
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/92083
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Ben Clayton <bclayton@google.com>
This is required to handle materialized values, and for constant
expressions.
Bug: tint:1504
Change-Id: Ie0177f148e08a0e1a3f4d7e06e283f121655804b
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/92080
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
We were missing an `UnwrapRef` when generating the return type, and
were generating invalid SPIR-V when the value being stored was a
reference. The auto-generated builtin tests only test with literal
values.
Fixed: tint:1573
Change-Id: If42280b3cc8ad3fba7355d333e02400c6db843fa
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/92144
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Auto-Submit: James Price <jrprice@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: James Price <jrprice@google.com>
If the literal was constructed with an 'f', make sure we print it.
Bug: tint:1504
Change-Id: I6f04e31a166919c07574db56b0a2063ce5b8ca5c
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/91965
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
David Neto
Antonio Maiorano
Stephen White
Ben Clayton
Zhaoming Jiang
dan sinclair
Loko Kung
James Price