[Mlir-commits] [mlir] [mlir][arith] Fix crash in SameInputOutputTensorDims with unranked tensors (PR #235349)

[llvmlistbot at llvm.org]

llvmorg-github-actions[bot] wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-mlir-arith

Author: Vinicius Silva (viniciusfdasilva)

<details>
<summary>Changes</summary>

This PR fixes #<!-- -->234633 by adding a hasRank() check first, so the verifier emits a regular error message instead of crashing.

---
Full diff: https://github.com/llvm/llvm-project/pull/235349.diff


1 Files Affected:

- (modified) mlir/include/mlir/Dialect/Arith/IR/ArithOps.td (+2-1) 


``````````diff
diff --git a/mlir/include/mlir/Dialect/Arith/IR/ArithOps.td b/mlir/include/mlir/Dialect/Arith/IR/ArithOps.td
index 4e7b85c534209c..8db9ff85703d11 100644
--- a/mlir/include/mlir/Dialect/Arith/IR/ArithOps.td
+++ b/mlir/include/mlir/Dialect/Arith/IR/ArithOps.td
@@ -135,7 +135,8 @@ class Arith_FloatBinaryOpWithRoundingMode<string mnemonic,
 def SameInputOutputTensorDims : PredOpTrait<
     "input and output have the same tensor dimensions",
     AllMatchSameOperatorPred<["in", "out"],
-      "(::llvm::isa<::mlir::TensorType>($_self.getType()) ?"
+      "(::llvm::isa<::mlir::TensorType>($_self.getType()) &&"
+      " ::llvm::cast<::mlir::TensorType>($_self.getType()).hasRank() ?"
       " ::llvm::cast<::mlir::TensorType>($_self.getType()).getShape() :"
       " ::llvm::ArrayRef<int64_t>{})">>;
 

``````````

</details>


https://github.com/llvm/llvm-project/pull/235349