[PATCH] D78541: [MLIR] Ensure `gpu.func` must be inside a `gpu.module`.

[River Riddle via Phabricator via llvm-commits]

rriddle added inline comments.


================
Comment at: mlir/test/lib/Transforms/TestGpuMemoryPromotion.cpp:28
     : public PassWrapper<TestGpuMemoryPromotionPass,
-                         OperationPass<gpu::GPUFuncOp>> {
+                         OperationPass<gpu::GPUModuleOp>> {
   void runOnOperation() override {
----------------
Can you just change the pipeline specification in the tests instead of changing this pass?


================
Comment at: mlir/test/lib/Transforms/TestGpuMemoryPromotion.cpp:30
   void runOnOperation() override {
-    gpu::GPUFuncOp op = getOperation();
-    for (unsigned i = 0, e = op.getNumArguments(); i < e; ++i) {
-      if (op.getArgAttrOfType<UnitAttr>(i, "gpu.test_promote_workgroup"))
-        promoteToWorkgroupMemory(op, i);
-    }
+    gpu::GPUModuleOp gpu_module = getOperation();
+    gpu_module.walk([](Operation *operation) {
----------------
MLIR uses camelCase for variable names.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D78541/new/

https://reviews.llvm.org/D78541