[Mlir-commits] [mlir] [MLIR][NVVM] Update nvvm.barrier.arrive Op (PR #202608)

[Durgadoss R]

================
@@ -1210,9 +1210,10 @@ def NVVM_BarrierReductionOp :
   }];
 }
 
-def NVVM_BarrierArriveOp : NVVM_PTXBuilder_Op<"barrier.arrive"> 
+def NVVM_BarrierArriveOp : NVVM_VoidIntrinsicOp<"barrier.arrive">
 {
-  let arguments = (ins Optional<I32>:$barrierId, I32:$numberOfThreads);
+  let arguments = (ins Optional<I32>:$barrierId, I32:$numberOfThreads,
----------------
durga4github wrote:

While we are here,
I think we should also support lowering to the `_all` variants of the intrinsics .. but thats for a separate PR..

https://github.com/llvm/llvm-project/pull/202608