[PATCH] D127609: [AtomicExpand] Make floating point conversion happens before fence insertion

[Eli Friedman via Phabricator via llvm-commits]

efriedma added inline comments.


================
Comment at: llvm/lib/CodeGen/AtomicExpandPass.cpp:226
+                  TargetLoweringBase::AtomicExpansionKind::CastToInteger) {
+      // TODO: add a TLI hook to control this so that each target can
+      // convert to lowering the original type one at a time.
----------------
I think you can drop the TODOs here, at least for load/store/atomicrmw.


================
Comment at: llvm/test/CodeGen/PowerPC/cfence-double.ll:16
 entry:
   %0 = load atomic double, double* undef acquire, align 8
   ret void
----------------
Please avoid using undef in testcases like this.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D127609/new/

https://reviews.llvm.org/D127609