[Mlir-commits] [mlir] Skip address space checks for memrefs between launchOp and kernel func (PR #102925)

[Petr Kurapov]

================
@@ -401,8 +401,29 @@ LogicalResult GPUDialect::verifyOperationAttribute(Operation *op,
              << expectedNumArguments;
 
     auto functionType = kernelGPUFunction.getFunctionType();
+    auto typesMatch = [&](Type launchOpArgType, Type gpuFuncArgType) {
+      auto launchOpMemref = dyn_cast<MemRefType>(launchOpArgType);
+      auto kernelMemref = dyn_cast<MemRefType>(gpuFuncArgType);
+      // Allow address space incompatibility for OpenCL kernels: `gpu.launch`'s
+      // argument memref without address space attribute will match a kernel
+      // function's memref argument with address space `Global`.
+      if (launchOpMemref && kernelMemref) {
+        auto launchAS = llvm::dyn_cast_or_null<gpu::AddressSpaceAttr>(
+            launchOpMemref.getMemorySpace());
+        auto kernelAS = llvm::dyn_cast_or_null<gpu::AddressSpaceAttr>(
+            kernelMemref.getMemorySpace());
+        if (!launchAS && kernelAS &&
+            kernelAS.getValue() == gpu::AddressSpace::Global)
----------------
kurapov-peter wrote:

> Can't you also pass a pointer to workgroup (aka local, shared) memory here from the host?

Not sure, what would it mean? I.e. what's the expected behavior?

https://github.com/llvm/llvm-project/pull/102925