[llvm] [NVPTX] Fixup AutoUpgrade of llvm.nvvm.atomic.load.{inc,dec}.32 (PR #138907)

[Artem Belevich via llvm-commits]

================
@@ -1353,12 +1353,12 @@ static bool upgradeIntrinsicFunction1(Function *F, Function *&NewFn,
         // nvvm.{min,max}.{i,ii,ui,ull}
         Expand = Name == "s" || Name == "i" || Name == "ll" || Name == "us" ||
                  Name == "ui" || Name == "ull";
-      else if (Name.consume_front("atomic.load.add."))
-        // nvvm.atomic.load.add.{f32.p,f64.p}
-        Expand = Name.starts_with("f32.p") || Name.starts_with("f64.p");
-      else if (Name.consume_front("atomic.load.") && Name.consume_back(".32"))
-        // nvvm.atomic.load.{inc,dec}.32
-        Expand = Name == "inc" || Name == "dec";
+      else if (Name.consume_front("atomic.load."))
+        // nvvm.atomic.load.add.{f32,f64}.p
+        // nvvm.atomic.load.{inc,dec}.32.p
+        Expand = Name.starts_with("add.f32.p") ||
----------------
Artem-B wrote:

Ack. That was a drive-by idea, not directly related to this patch.

https://github.com/llvm/llvm-project/pull/138907