[Mlir-commits] [mlir] Update `LowerContractionToSMMLAPattern` to ingnore matvec (PR #88288)

[Diego Caballero]

================
@@ -54,6 +54,8 @@ class LowerContractionToSMMLAPattern
     // Note: RHS is not transposed.
     mlir::VectorType lhsType = op.getLhsType();
     mlir::VectorType rhsType = op.getRhsType();
+    if (rhsType.getRank() < 2)
+      return failure();
----------------
dcaballe wrote:

Since 0-D vectors can also reach this point, could you add another check making sure that if any of the operands doesn't have a rank, we also bail out (+ test for that)?

https://github.com/llvm/llvm-project/pull/88288