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

[llvmlistbot at llvm.org]

================
@@ -218,7 +218,8 @@ mlir::amdgpu::optimizeSharedMemoryReadsAndWrites(Operation *parentOp,
   return success();
 }
 
-void amdgpu::optimizeSharedMemoryReadsAndWritesOp(func::FuncOp funcOp) {
+std::optional<mlir::LogicalResult>
----------------
erman-gurses wrote:

There is one more `return;` statement above after checking `hasSharedMemoryAddressSpace()`. There is no failure in this case it just checks if the shared memory relation exists or not. So I added an optional for that.   

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