Use the sem::Type size instead of the sem::Struct size.
Some binding numbers in MSL expected outputs changed since these
buffers were previously not being reported in the resource list that
the Tint exe uses to automatically remap bindings.
Bug: tint:1372
Change-Id: I14479ddc5129d91b91449cc2d68ee37bd99d2f7e
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/76220
Reviewed-by: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Use the AddSpirvBlockDecoration transform to do this.
Add expected results for all of the new E2E tests. The arrayLength()
tests all fail as this is not yet implemented for the GLSL
backend. The other tests all pass except two that assign whole structs
to buffers, which is also a pre-existing issue.
Bug: tint:1372
Change-Id: I230197b43a5561e619866419d642ffc1ed085aac
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/76164
Reviewed-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Stephen White <senorblanco@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
These two transforms previously assumed that the argument to
arrayLength had the form `&struct_var.array_member`. We now also need
to handle the case where it is just `&array_var`.
Bug: tint:1372
Change-Id: I173a84bd32c324445573a295b281a51e291c2ae2
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/76163
Reviewed-by: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
For SPIR-V, wrap non-struct types in structs in the
AddSpirvBlockDecoration transform.
For MSL, wrap runtime-sized arrays in structs in the
ModuleScopeVarToEntryPointParam transform.
Bug: tint:1372
Change-Id: Icced5d77b4538e816aa9fab57a634a9f4c52fdab
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/76162
Reviewed-by: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Add google benchmark to the DEPs.
Implement a basic set of benchmarks for each of the writers and the WGSL parser.
Add build rules for CMake. GN build rules TODO.
Add a simple go tool (ported from Marl) to diff two benchmarks. Less
noisy than the one provided by google benchmark.
Bug: tint:1378
Change-Id: I73cf92c5d9fd2d3bfac8f264864fd774afbd5d01
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/76840
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ryan Harrison <rharrison@chromium.org>
Commit-Queue: Ben Clayton <bclayton@chromium.org>
Remove the ICE check for expression behaviors always being either `{Next}` or `{Next, Discard}`. Unreachable code may be result in something else.
Add the RemoveUnreachableStatements transform to the SPIR-V writer sanitizer transform list. The writer cannot correctly handle unreachable statements.
Bug: tint:1369
Bug: chromium:1285622
Change-Id: I9fa54c6d2096b1ee633dd551b628c7dd3ba64fb5
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/76300
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: David Neto <dneto@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
GLSL does not support separate textures and samplers, so they must be
replaced with combined samplers. This is the first stage of that change,
where we collect the unique texture/sampler pairs. Within a function,
texture and sampler must be either a global variable or a function
parameter. At the entry point level, all references must resolve to global
variables, so by recursing the call graph we can determine all of the
global pairs required.
This information will be used by an upcoming transform to modify the AST
to be GLSL-compliant: modifying function signatures, call sites, removing
separate globals and adding combined globals. It will also eventually
replace the pair-gathering currently performed by
Inspector::GetSamplerTextureUses().
Bug: tint:1366
Change-Id: I89451b195649da26e45641ea2f6955683ae9fc66
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/75960
Reviewed-by: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Stephen White <senorblanco@chromium.org>
With C++17, std::result_of has been deprecated and superceeded with std::invoke_result.
std::result_of triggers a compiler warning about use of a deprecated feature, which is causing the ossfuzz builds to fail
Issue: oss-fuzz:43441
Change-Id: Ie4ed670dd5b25b9059d5feb8056e37a22dad3cbf
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/75981
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
The tool was getting slow, due to the number of branches being examined.
M96 is the latest stable release. We aren't applying fixes to this or older branches.
Also avoids an issue where the M95 branch was missing a cherry-pick, which the tool wanted to merge with each run.
Change-Id: I8a8b1466c10633d9539725cd4f9b1b0514e41e93
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/75067
Auto-Submit: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: James Price <jrprice@google.com>
Commit-Queue: James Price <jrprice@google.com>
Constructing a null vector, matrix or array does not need to provide an explicit list of null elements. `T()` is a more efficient and readable equivalent.
Fuzzers like to generate enormous pathological composite types, which has been triggering OOM failures.
Bug: oss-fuzz:38095
Bug: oss-fuzz:39235
Bug: oss-fuzz:39246
Bug: oss-fuzz:39874
Change-Id: I910bb04bdd0e80532c09fc038be895ec37d3d380
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/75426
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: David Neto <dneto@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
nullptrs are caused by errors. Detect and return before passing nullptrs to places they shouldn't go.
Also: Rename 'params' to 'args'. These are arguments, not parameters.
Bug: oss-fuzz:38074
Fixed: tint:1355
Change-Id: I77f97b4a8e5dab21802b300ba3eedad767ac2ad5
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/75425
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: David Neto <dneto@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
When the emit_continuing lambda always returns true, clang is able to
detect it and finds that some code is unreachable.
This commit changes places that use the lambdas directly to instead use
the stored emit_continuing_ member.
Bud: dawn:824
Change-Id: Idfd804a6ec53ac793e8988dce79991c659fbbc36
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/75060
Auto-Submit: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Just like for divide, FXC fails with the exact same error when
performing a modulo on a value that FXC determines to be zero. We
address it in the same way as we do for divide.
This also fixes a couple of the vk-gl-cts tests for which I manually
generated expectation files for.
Bug: tint:1083
Change-Id: Ia388bf002112afded542adb791d37e88e35a77ff
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/74220
Reviewed-by: James Price <jrprice@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Antonio Maiorano <amaiorano@google.com>
FXC fails to compile when it determines that the rhs of an integral
division is zero with "error X4010: Unsigned integer divide by zero".
bclayton's fix (https://dawn-review.googlesource.com/c/tint/+/60500)
addressed cases for division by an integer constant 0. This CL adds the
missing support for division by integral vectors with 0 components.
FXC also fails on division by integral expressions that it can fold to
0. To handle these cases, we now emit a runtime check for 0 and replace
by 1. In the cases I've tested, FXC seems able to optimize these checks
away.
Bug: tint:1083
Change-Id: I02f08e9077882f03c1e42b62dacb742a48fa48ba
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/73580
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: David Neto <dneto@google.com>
Reviewed-by: James Price <jrprice@google.com>
Commit-Queue: Antonio Maiorano <amaiorano@google.com>
There are a few places where a branch or return is created,
conditionally on whether a terminator was the last thing seen.
The goal is to generate a SPIR-V basic block terminator exactly
when needed, and to avoid generating a branch or return immediately
after a prior terminator.
Previously, the decision was based on the last thing seen in the AST.
But we should instead check the emitted SPIR-V instead.
This fixes cases such as a break or return inside an else-if.
That's because an if/elseif is actually a selection inside a selection.
Looking at the AST only works when trying to terminate the *inside*
selection. In the outer recursive call, the last AST node is
no longer a terminator, and we would skip generating the branch
to the merge block.
Fixed: tint:1315
Change-Id: I6b886ce85d1d681f2063997e469e0c1b4e5973a2
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/73480
Kokoro: Kokoro <noreply+kokoro@google.com>
Auto-Submit: David Neto <dneto@google.com>
Reviewed-by: James Price <jrprice@google.com>
Commit-Queue: James Price <jrprice@google.com>
* Rename GenerateNonReferenceExpression to
GenerateExpressionWithLoadIfNeeded.
This version takes an ast::Expression
* Add a variant that takes a sem::Expression, because the sem
expression already knows the resolved type, and so we can save
a lookup.
* Replace most uses of GenerateExpression ... GenerateLoadIfNeeded
with a call to one of the above.
This is a non-functional change.
Followup to the fix in tint:1343.
Bug: tint:1343
Change-Id: If19a1bc7670edd2badc1533861d8b42f0825c7b8
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/72720
Auto-Submit: David Neto <dneto@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: David Neto <dneto@google.com>
