[llvm] [SROA] Prevent load atomic vector from being generated (PR #112432)

[Matt Arsenault via llvm-commits]

================
@@ -2853,6 +2853,11 @@ class AllocaSliceRewriter : public InstVisitor<AllocaSliceRewriter, bool> {
 
   bool visitLoadInst(LoadInst &LI) {
     LLVM_DEBUG(dbgs() << "    original: " << LI << "\n");
+
+    // load atomic vector would be generated, which is illegal
+    if (LI.isAtomic() && NewAI.getAllocatedType()->isVectorTy())
----------------
arsenm wrote:

No alloca reference, just the type. The implementation should match Verifier::visitLoadInst, the isIntOrIntPtrTy || isFloatingPointTy and checkAtomicMemAccessSize (also add some non-atomic size tests?)

https://github.com/llvm/llvm-project/pull/112432