[Mlir-commits] [mlir] [mlir][GPU] Extend gpu.barrier with scope and named-barrier support (PR #195692)

[Fabian Mora]

================
@@ -517,6 +518,51 @@ struct GPUShuffleOpLowering : public ConvertOpToLLVMPattern<gpu::ShuffleOp> {
   }
 };
 
+/// Emit an LLVM fence with MMRA metadata based on the given address spaces.
+/// If `addrSpaces` is nullopt, all memory is fenced (global + LDS).
+static void emitFences(std::optional<ArrayAttr> addrSpaces,
+                       ConversionPatternRewriter &rewriter, Location loc,
+                       StringRef scope, bool before) {
+  bool fenceGlobal = false;
+  bool fenceLDS = false;
+
+  if (addrSpaces) {
+    for (auto spaceAttr : addrSpaces->getAsRange<gpu::AddressSpaceAttr>()) {
+      switch (spaceAttr.getValue()) {
+      case gpu::AddressSpace::Global:
+        fenceGlobal = true;
+        break;
+      case gpu::AddressSpace::Workgroup:
+        fenceLDS = true;
+        break;
+      case gpu::AddressSpace::Private:
----------------
fabianmcg wrote:

Private usually relies in the global mem space, and constant in AMD might require the same kind of fence as LDS, so unsure about this break. Maybe an error?

https://github.com/llvm/llvm-project/pull/195692