[Mlir-commits] [mlir] [MLIR][GPU-LLVM] Add in-pass signature update option for opencl kernels (PR #105664)

[Fabian Mora]

================
@@ -515,3 +517,15 @@ gpu.module @kernels {
     gpu.return
   }
 }
+
+// -----
+
+gpu.module @kernels {
+// OPENCL-LABEL:   llvm.func spir_funccc @no_address_spaces(
+// OPENCL-SAME:                                             %{{[a-zA-Z_][a-zA-Z0-9_]*}}: !llvm.ptr<1>
+// OPENCL-SAME:                                             %{{[a-zA-Z_][a-zA-Z0-9_]*}}: !llvm.ptr<1>
+// OPENCL-SAME:                                             %{{[a-zA-Z_][a-zA-Z0-9_]*}}: !llvm.ptr<1>
+  gpu.func @no_address_spaces(%arg0: memref<f32>, %arg1: memref<f32, #gpu.address_space<global>>, %arg2: memref<f32>) {
+    gpu.return
+  }
+}
----------------
fabianmcg wrote:

Can you add additional and more complex tests?
For example, adding a device function inside the module and calling it from the kernel. Loading one of the arguments, etc...

https://github.com/llvm/llvm-project/pull/105664