[llvm] [NVPTX] Cleanup and refactor atomic lowering (PR #133781)

[Alex MacLean via llvm-commits]

================
@@ -994,6 +994,7 @@ NVPTXTargetLowering::NVPTXTargetLowering(const NVPTXTargetMachine &TM,
 
   setOperationAction(ISD::ADDRSPACECAST, {MVT::i32, MVT::i64}, Custom);
 
+  setOperationAction(ISD::ATOMIC_LOAD_SUB, {MVT::i32, MVT::i64}, Expand);
----------------
AlexMaclean wrote:

These are the only integer types we support `atom.add` for (and the only types we previously supported sub lowering for in ISel). I guess we should probably be doing something like this for `ATOMIC_LOAD_FSUB`, but that wasn't working before this change either.

https://github.com/llvm/llvm-project/pull/133781