[PATCH] D106729: [AMDGPU] Support non-strictly stronger memory orderings in SIMemoryLegalizer

[Eli Friedman via Phabricator via llvm-commits]

efriedma added inline comments.


================
Comment at: llvm/lib/Target/AMDGPU/SIMemoryLegalizer.cpp:656
              MMO->getFailureOrdering() != AtomicOrdering::AcquireRelease);
       FailureOrdering =
+          getMergedAtomicOrdering(FailureOrdering, MMO->getFailureOrdering());
----------------
efriedma wrote:
> Isn't this just equivalent to `FailureOrdering = MMO->getFailureOrdering()`?
Oh, this is a loop; nevermind.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D106729/new/

https://reviews.llvm.org/D106729