[Mlir-commits] [mlir] [MLIR][XeGPU] Lower vector.multi_reduction to vector.reduction for lane local (PR #191037)

[Nishant Patel]

================
@@ -145,7 +145,8 @@ Value subgroupReduction(Location loc, OpBuilder &builder, Value input,
 Value lowerToVectorReductions(TypedValue<VectorType> src,
                               TypedValue<VectorType> acc,
                               vector::CombiningKind kind, int64_t reductionDim,
-                              Location loc, PatternRewriter &rewriter);
+                              Location loc, PatternRewriter &rewriter,
+                              bool setLayout = true);
----------------
nbpatel wrote:

actually added a TODO there..those calls are harmless for the new pass...will remove it once we deprecate the old pass.

https://github.com/llvm/llvm-project/pull/191037