Linalg vali matrixstoretodescriptor

[Ashley Coleman (V-FEXrt)]

Fixes #8497

Implements validation rules for MatrixStoreToDescriptor

[github-actions]

✅ With the latest revision this PR passed the C/C++ code formatter.

[Copilot]

Pull request overview

Adds DXIL validation rules for LinAlgMatrixStoreToDescriptor.

Changes:

• Validates layout, matrix scope, UAV resource type, and alignment.
• Adds diagnostics and negative validation tests.
• Updates existing fixtures to use valid layout and alignment values.

Reviewed changes

Copilot reviewed 15 out of 15 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
utils/hct/hctdb.py	Adds the layout validation rule.
lib/DxilValidation/DxilValidation.cpp	Implements descriptor-store validation.
docs/DXIL.rst	Documents the new diagnostic.
tools/clang/test/LitDXILValidation/LinAlgMatrix/linalgmatrix-matrixstoretodescriptor.ll	Tests invalid operands and matrix scope.
tools/clang/test/LitDXILValidation/LinAlgMatrix/linalgmatrix-vs.ll	Uses valid store arguments.
tools/clang/test/LitDXILValidation/LinAlgMatrix/linalgmatrix-ps.ll	Uses valid store arguments.
tools/clang/test/LitDXILValidation/LinAlgMatrix/linalgmatrix-cs.ll	Uses valid store arguments.
tools/clang/test/LitDXILValidation/LinAlgMatrix/linalgmatrix-hs.ll	Uses valid store arguments.
tools/clang/test/LitDXILValidation/LinAlgMatrix/linalgmatrix-ds.ll	Uses valid store arguments.
tools/clang/test/LitDXILValidation/LinAlgMatrix/linalgmatrix-gs.ll	Uses valid store arguments.
tools/clang/test/LitDXILValidation/LinAlgMatrix/linalgmatrix-ms.ll	Uses valid store arguments.
tools/clang/test/LitDXILValidation/LinAlgMatrix/linalgmatrix-as.ll	Uses valid store arguments.
tools/clang/test/LitDXILValidation/LinAlgMatrix/linalgmatrix-node.ll	Uses valid store arguments.
tools/clang/test/LitDXILValidation/LinAlgMatrix/linalgmatrix-raytracing.ll	Updates ray-tracing fixtures.
tools/clang/test/CodeGenDXIL/hlsl/linalg/builtins/matrixstoretodescriptor/nominal.hlsl	Updates nominal alignment expectations.

Suppressed comments (1)

lib/DxilValidation/DxilValidation.cpp:1188

• The nonconstant Align rejection is also untested: current cases cover zero and a non-multiple only. Add a runtime i32 alignment case and verify the immediate-constant diagnostic so this branch cannot regress.

  } else
    ValCtx.EmitInstrFormatError(CI, ValidationRule::InstrOpConst,
                                {"Align", "LinAlgMatrixStoreToDescriptor"});

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

Pull request overview

Copilot reviewed 15 out of 15 changed files in this pull request and generated no new comments.

Suppressed comments (2)

lib/DxilValidation/DxilValidation.cpp:1131

• This adds user-visible validation behavior for experimental Shader Model 6.10. Please add a single-sentence entry under ### Upcoming Preview Release in docs/ReleaseNotes.md, as required for significant compiler behavior changes by CONTRIBUTING.md.

  ValidateLinAlgOpParameters(CI, ValCtx);

tools/clang/test/LitDXILValidation/LinAlgMatrix/linalgmatrix-matrixstoretodescriptor.ll:28

• This case uses an SRV RawBuffer, so it exercises only the UAV-class half of the new resource check; it would still pass if the RawBuffer-kind check were removed. Add a second case using a UAV of a non-RawBuffer kind (for example, a typed or structured UAV) and verify the same diagnostic.

  %6 = call %dx.types.Handle @dx.op.annotateHandle(i32 216, %dx.types.Handle %2, %dx.types.ResourceProperties { i32 11, i32 0 })  ; AnnotateHandle(res,props)  resource: ByteAddressBuffer
  call void @dx.op.linAlgMatrixStoreToDescriptor.mC8M4N4U0S1(i32 -2147483628, %dx.types.LinAlgMatrixC8M4N4U0S1 %4, %dx.types.Handle %6, i32 0, i32 0, i32 0, i32 128)  ; LinAlgMatrixStoreToDescriptor(matrix,handle,offset,stride,layout,align)

[Copilot]

Pull request overview

Copilot reviewed 15 out of 15 changed files in this pull request and generated no new comments.

[Joshua Batista (bob80905)]

Only one nit comment that I'm curious about, but it looks good to me!

[Joshua Batista (bob80905)]

Should this really exit without a validation diagnostic?