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

[llvmlistbot at llvm.org]

================
@@ -1700,10 +1824,10 @@ vectorizeAsTensorPackOp(RewriterBase &rewriter, linalg::PackOp packOp,
   Value dest = rewriter.create<tensor::EmptyOp>(
       loc, reifiedReturnShapes[0],
       transposeOp.getResult().getType().getElementType());
-  Operation *write =
-      createWriteOrMaskedWrite(rewriter, loc, transposeOp.getResult(), dest,
-                               /*inputVecSizesForLeadingDims=*/inputVectorSizes,
-                               /*useInBoundsInsteadOfMasking=*/false);
+  Operation *write = createWriteOrMaskedWrite(
+      rewriter, loc, transposeOp.getResult(), dest,
+      /*inputVecSizesForLeadingDims=*/inputVectorSizes, /*writeIndices=*/{},
+      /*useInBoundsInsteadOfMasking=*/false);
----------------
Max191 wrote:

nit: `writeIndices` and `useInBoundsInsteadOfMasking` are using the default values. Maybe just omit them?

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