[llvm-branch-commits] [llvm] [AtomicExpand] Avoid sized call when expanding load atomic vector (PR #120716)

[Matt Arsenault via llvm-branch-commits]

================
@@ -1884,7 +1884,10 @@ bool AtomicExpandImpl::expandAtomicOpToLibcall(
   IRBuilder<> Builder(I);
   IRBuilder<> AllocaBuilder(&I->getFunction()->getEntryBlock().front());
 
-  bool UseSizedLibcall = canUseSizedAtomicCall(Size, Alignment, DL);
+  const bool IsAtomic =
+      isa<LoadInst>(I) ? cast<LoadInst>(I)->isAtomic() : false;
+  const bool UseSizedLibcall = !(I->getType()->isVectorTy() && IsAtomic) &&
----------------
arsenm wrote:

This shouldn't need to consider whether it's atomic or the type. The sized call should work fine. Some casts may be necessary 

https://github.com/llvm/llvm-project/pull/120716