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

[Alex MacLean via llvm-commits]

================
@@ -41,6 +41,27 @@ def AS_match {
   }];
 }
 
+multiclass nvvm_ternary_atomic_op_scoped<SDPatternOperator frag> {
+  defvar frag_pat = (frag node:$ptr, node:$cmp, node:$val);
+  def NAME#_cta: PatFrag<!setdagop(frag_pat, ops),
+      (!cast<SDPatternOperator>(NAME) node:$ptr, node:$cmp, node:$val), [{
+        return Scopes[cast<MemSDNode>(N)->getSyncScopeID()] == NVPTX::Scope::Block;
+  }]>;
+  def NAME#_cluster : PatFrag<!setdagop(frag_pat, ops),
+      (!cast<SDPatternOperator>(NAME) node:$ptr, node:$cmp, node:$val), [{
+        return Scopes[cast<MemSDNode>(N)->getSyncScopeID()] == NVPTX::Scope::Cluster;
+  }]>;
+  def NAME#_gpu: PatFrag<!setdagop(frag_pat, ops),
+      (!cast<SDPatternOperator>(NAME) node:$ptr, node:$cmp, node:$val), [{
+        return Scopes[cast<MemSDNode>(N)->getSyncScopeID()] == NVPTX::Scope::Device;
+  }]>;
+  def NAME#_sys: PatFrag<!setdagop(frag_pat, ops),
+      (!cast<SDPatternOperator>(NAME) node:$ptr, node:$cmp, node:$val), [{
+        return Scopes[cast<MemSDNode>(N)->getSyncScopeID()] == NVPTX::Scope::System;
+  }]>;
+}
----------------
AlexMaclean wrote:

This looks reasonable, but I wonder if we're the first target to have added functionality like this? Have any other targets added support for scoped atomics? If so how did they implement it? I wonder if this could be simplified with a complex pattern or something similar.

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