[Mlir-commits] [mlir] [MLIR][GPU] Support grid constant, byval, byref on gpu.func (PR #172037)

[Mehdi Amini]

================
@@ -369,7 +370,10 @@ GPUFuncOpLowering::matchAndRewrite(gpu::GPUFuncOp gpuFuncOp, OpAdaptor adaptor,
     }
 
     if (lowersToPointer) {
+      copyPointerAttribute(mlir::NVVM::NVVMDialect::getGridConstantAttrName());
----------------
joker-eph wrote:

Having to teach the gpu dialect about target specific dialect attributes isn't ideal, I wonder if this couldn't be served by an interface to allow target-specific dialects to hook into?

https://github.com/llvm/llvm-project/pull/172037