[Mlir-commits] [mlir] [mlir][linalg] Refactor vectorization hooks to improve code reuse (PR #141244)

llvmlistbot at llvm.org llvmlistbot at llvm.org
Thu May 29 10:36:10 PDT 2025


================
@@ -1874,10 +1998,10 @@ vectorizeAsTensorPadOp(RewriterBase &rewriter, tensor::PadOp padOp,
   // Create Xfer write Op
   Value dest = rewriter.create<tensor::EmptyOp>(
       loc, reifiedReturnShapes[0], padOp.getResultType().getElementType());
-  Operation *write =
-      createWriteOrMaskedWrite(rewriter, loc, maskedRead, dest,
-                               /*inputVecSizesForLeadingDims=*/inputVectorSizes,
-                               /*useInBoundsInsteadOfMasking=*/false);
+  Operation *write = createWriteOrMaskedWrite(
+      rewriter, loc, maskedRead, dest,
+      /*inputVecSizesForLeadingDims=*/inputVectorSizes, {},
+      /*useInBoundsInsteadOfMasking=*/false);
----------------
Max191 wrote:

nit: Same here, omit default valued parameters? (or otherwise add a comment for `/*writeIndices=*/`)

https://github.com/llvm/llvm-project/pull/141244


More information about the Mlir-commits mailing list