[clang] [CodeGen] Fix volatile inst (PR #171862)

[via cfe-commits]

llvmbot wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-clang-codegen

Author: None (GkvJwa)

<details>
<summary>Changes</summary>

Fix #<!-- -->126516

Instead, use AtomicOrdering::Monotonic's atomic loading to let the backend automatically select the appropriate instruction.

---
Full diff: https://github.com/llvm/llvm-project/pull/171862.diff


1 Files Affected:

- (modified) clang/lib/CodeGen/CGBuiltin.cpp (+2) 


``````````diff
diff --git a/clang/lib/CodeGen/CGBuiltin.cpp b/clang/lib/CodeGen/CGBuiltin.cpp
index 01be374422d93..fa13240e74e87 100644
--- a/clang/lib/CodeGen/CGBuiltin.cpp
+++ b/clang/lib/CodeGen/CGBuiltin.cpp
@@ -566,6 +566,7 @@ static Value *EmitISOVolatileLoad(CodeGenFunction &CGF, const CallExpr *E) {
   llvm::Type *ITy =
       llvm::IntegerType::get(CGF.getLLVMContext(), LoadSize.getQuantity() * 8);
   llvm::LoadInst *Load = CGF.Builder.CreateAlignedLoad(ITy, Ptr, LoadSize);
+  Load->setAtomic(llvm::AtomicOrdering::Monotonic);
   Load->setVolatile(true);
   return Load;
 }
@@ -578,6 +579,7 @@ static Value *EmitISOVolatileStore(CodeGenFunction &CGF, const CallExpr *E) {
   CharUnits StoreSize = CGF.getContext().getTypeSizeInChars(ElTy);
   llvm::StoreInst *Store =
       CGF.Builder.CreateAlignedStore(Value, Ptr, StoreSize);
+  Store->setAtomic(llvm::AtomicOrdering::Monotonic);
   Store->setVolatile(true);
   return Store;
 }

``````````

</details>


https://github.com/llvm/llvm-project/pull/171862