[Mlir-commits] [mlir] [MLIR][GPU-LLVM] Convert `gpu.func` to `llvm.func` (PR #101664)

[Victor Perez]

================
@@ -57,10 +80,18 @@ struct GPUFuncOpLowering : ConvertOpToLLVMPattern<gpu::GPUFuncOp> {
   unsigned workgroupAddrSpace;
 
   /// The attribute name to use instead of `gpu.kernel`.
-  StringAttr kernelAttributeName;
-
+  std::optional<StringAttr> kernelAttributeName;
   /// The attribute name to to set block size
   std::optional<StringAttr> kernelBlockSizeAttributeName;
+
+  /// The calling convention to use for kernel functions
+  LLVM::CConv kernelCallingConvention;
----------------
victor-eds wrote:

I think that'd make the class more cumbersome to work with. Also, I wouldn't expect the list of options to grow too much (hopefully). I don't think keeping them aligned when both definitions are like 5 lines apart is a big deal.

https://github.com/llvm/llvm-project/pull/101664