[Mlir-commits] [mlir] [MLIR][AMDGPU]Add refactoring for shared-mem optimization (PR #81791)

[Oleksandr Alex Zinenko]

================
@@ -49,7 +49,8 @@ namespace amdgpu {
 mlir::LogicalResult optimizeSharedMemoryReadsAndWrites(Operation *parentOp,
                                                        Value memrefValue);
 
-void optimizeSharedMemoryReadsAndWritesOp(mlir::func::FuncOp funcOp);
+std::optional<mlir::LogicalResult>
+optimizeSharedMemoryReadsAndWritesOp(mlir::func::FuncOp funcOp);
----------------
ftynse wrote:

Nit: there's no need to prefix with `mlir::` in the mlir code base.

https://github.com/llvm/llvm-project/pull/81791