[llvm] [NVPTX] Add syncscope support for cmpxchg (PR #140812)

[Akshay Deodhar via llvm-commits]

================
@@ -2111,9 +2132,9 @@ multiclass F_ATOMIC_2<RegTyInfo t, string sem_str, string as_str, string op_str,
 }
 
 // has 3 operands
-multiclass F_ATOMIC_3<RegTyInfo t, string sem_str, string as_str, string op_str,
-                      SDPatternOperator op, list<Predicate> preds> {
-  defvar asm_str = "atom" # sem_str # as_str # "." # op_str # " \t$dst, [$addr], $b, $c;";
+multiclass F_ATOMIC_3<RegTyInfo t, string sem_str, string scope_str, string as_str, 
----------------
akshayrdeodhar wrote:

`printLdStCode` also prints based on the "order" modifier, yet we support order for atomics in tablegen. `.scope` is a modifier explicitly defined in [atom](https://docs.nvidia.com/cuda/parallel-thread-execution/index.html#parallel-synchronization-and-communication-instructions-atom). I would prefer to keep more stuff in TableGen than in C++ source.

https://github.com/llvm/llvm-project/pull/140812