[PATCH] D58632: [X86] Improve lowering of idemptotent RMW operations

[JF Bastien via Phabricator via llvm-commits]

jfb added a comment.

Overall this looks good, but I have a few comments.



================
Comment at: lib/Target/X86/X86ISelLowering.cpp:25482
   // rmw. In practice, they do not look useful, so we don't try to be
   // especially clever.
   if (SSID == SyncScope::SingleThread)
----------------
This comment needs an update.


================
Comment at: lib/Target/X86/X86ISelLowering.cpp:25492
     // is probably a small win, and x86 processors without mfence are rare
     // enough that we do not bother.
     return nullptr;
----------------
This seems to need an update too.


================
Comment at: test/CodeGen/X86/atomic-idempotent.ll:193
+  ret void
+}
+
----------------
Can you test `monotonic` as well?
Also, `i8` and `i16`.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D58632/new/

https://reviews.llvm.org/D58632