The continuing block can exit the loop in very constrained ways:
When a break statement is placed such that it would exit from a loop’s
§ 7.3.8 Continuing Statement, then:
- The break statement must appear as either:
- The only statement in the if clause of an if statement that has:
- no else clause or an empty else clause
- no elseif clauses
- The only statement in the else clause of an if statement that has an
empty if clause and no elseif clauses.
- That if statement must appear last in the continuing clause.
By design, this allows a lossless round-trip from SPIR-V to WGSL and
back to SPIR-V. But that requires this special case construct in WGSL
to be translated to an OpBranchConditional with one target being
the loop's megre block (which is where 'break' branches to), and the
other targets the loop header (which is the loop backedge). That
OpBranchConditional takes the place of the normal case of an
unconditional backedge.
Avoids errors like this:
continue construct with the continue target X is not
post dominated by the back-edge block Y
Fixed: 1034
Change-Id: If472a179380b8d77af746a3cd8e279c8a5e56b37
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/59800
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>
Add `transform::DecomposeStridedMatrix`, which replaces matrix members of storage or uniform buffer structures, that have a [[stride]] decoration, into an array
of N column vectors.
This is required to correctly handle `mat2x2` matrices in UBOs, as std140 rules will expect a default stride of 16 bytes, when in WGSL the default structure layout expects a stride of 8 bytes.
Bug: tint:1047
Change-Id: If5ca3c6ec087bbc1ac31a8d9a657b99bf34042a4
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/59840
Reviewed-by: David Neto <dneto@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
The DefInfo structure is used for remapping storage buffer types
as well as tracking special values like builtin variables.
In the latter case, don't take the defaulted storage class value
from the DefInfo initialization.
Fixed: tint:1040, tint:1043
Change-Id: I41ee364d76e632736d51f4474c97036bcc136c93
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/59484
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>
These just print colors to stdout, and actually do not check anything.
They're pretty, but I'm fairly confident this works now, and doesn't actually give us anything other than spam output.
Disable.
Change-Id: I6144f154345ae115f0899ca7e274fdee8947c785
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/59621
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: James Price <jrprice@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
- skip dumping due to SPIR-V Tools validation error for Location
conflicts for IO variables duplicated on an OpEntryPoint interface
list (SPIRV-Tools issue #4403)
- Skip dumping for unit test for conversion of RelaxedPrecision
decoration.
Bug: tint:1043
Change-Id: I8ce3a7660d74b771e3fa62dc39e4e030e4ba6aa1
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/59560
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>
- each vertex shader must have builtin-position output.
- when testing non-vertex shader features, change to fragment shaders,
to avoid the need to have a builtin position output.
- Fix ordering of entry point declaration vs. OpName
Bug: tint:1043
Change-Id: Ic66383d42419a1ef893835ae71729ae27fc6e539
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/59482
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>
- clean up function decorations unit tests
- clean up interpolate and invariant validation and unittest
- add separate unit-tests for each shader stage input and output
- add [[builtin(position)]] tests
- add validation and test for:
structures with 'location' decorated members cannot be used as compute shaders input
Bug: tint:1007
Change-Id: I12e97e163b3a77bc76ce21faba241683eec5d917
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/58942
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Reviewed-by: James Price <jrprice@google.com>
This adds SPIR-V assembly and WGSL tests derived from VK-GL-CTS commit
571256871c2e2f03995373e1e4a02958d8cd8cf5. The following procedure was
followed:
- Those .amber files in VK-GL-CTS wholly owned by Google were
identified
- All GLSL and SPIR-V shaders were extracted from the Amber files and
converted into SPIR-V binaries
- The compact-ids pass of spirv-opt was applied to each binary
- Duplicate binaries were removed
- spirv-opt -O was used to obtain an optimized version of each remaining
binary, with duplicates discarded
- Binaries that failed validation using spirv-val with target
environment SPIR-V 1.3 were discarded
- Those binaries that tint could not successfully convert into WGSL were
put aside for further investigation
- SPIR-V assembly versions of the remaining binaries are included in
this CL
- test-runner with -generate-expected and -generate-skip was used to
generate expected .spvasm, .msl, .hlsl and .wgsl outputs for these
SPIR-V assembly tests
- Each successfully-generated .expected.wgsl is included in this CL
again, as a WGLSL test
- test-runner with -generate-expected and -generate-skip was used again,
to generate expected outputs for these WGSL tests
Change-Id: Ibe9baf2729cf97e0b633db9a426f53362a5de540
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/58842
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
The SPIR-V Tools fuzzer asserts that the binaries it receives have sizes
that are multiples of 4 bytes, as it should only ever run on valid
binaries. This is failing in ClusterFuzz, likely due to the fuzzer being
misconfigured, so for now these assertions have been replaced with early
exits. They should be reinstated once the fuzzer is correctly
configured.
Fixes: chromium:1232308
Change-Id: I1fa980d09ce9e5c349a2cfcebe0246ebad6613fb
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/59440
Reviewed-by: Ben Clayton <bclayton@google.com>
Commit-Queue: Ben Clayton <bclayton@google.com>
Kokoro: Ben Clayton <bclayton@google.com>
Kokoro: Kokoro <noreply+kokoro@google.com>
Auto-Submit: Alastair Donaldson <afdx@google.com>
These are generated by the SPIR-V reader. Structures used by the
storage and uniform storage classes are handled separately, so we can
safely ignore this attribute for the other storage classes which are
not shared with the host.
Fixed: tint:1027
Change-Id: I8df4b41c687922a97f3b0ed97804cecc17d8997a
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/59240
Kokoro: Kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Enabling particular fuzzers requires that certain other CMake options
are enabled. For example, enabling the SPIR-V Tools fuzzer requires
enabling the CMake option for building spirv-fuzz. This change uses the
FORCE parameter to force the CMake cache to be overridden in such cases.
The change also adds FORCE to other options where it seems appropriate,
and improves some doc strings associated with CMake options.
Fixes: tint:1024
Change-Id: Ia72ddc0b3af625bb21ff1cc1c7121185ec3908b6
Reviewed-on: https://dawn-review.googlesource.com/c/tint/+/59204
Auto-Submit: Alastair Donaldson <afdx@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Vasyl Teliman <vasniktel@gmail.com>
Reviewed-by: Ryan Harrison <rharrison@chromium.org>
Kokoro: Kokoro <noreply+kokoro@google.com>
Commit-Queue: Alastair Donaldson <afdx@google.com>
