[CIR] Cast record size to uint64 to prevent overflow

[HendrikHuebner]

llvm::TypeSize uses 64bit integers, so we should cast the recordSize before multiplying by 8 to prevent an overflow.

[llvmbot]

@llvm/pr-subscribers-clangir

@llvm/pr-subscribers-clang

Author: Hendrik Hübner (HendrikHuebner)

Changes

llvm::TypeSize uses 64bit integers, so we should cast the recordSize before multiplying by 8 to prevent an overflow.

---

Full diff: https://github.com/llvm/llvm-project/pull/167525.diff

1 Files Affected:

• (modified) clang/lib/CIR/Dialect/IR/CIRTypes.cpp (+1-1)

diff --git a/clang/lib/CIR/Dialect/IR/CIRTypes.cpp b/clang/lib/CIR/Dialect/IR/CIRTypes.cpp
index 5897352829891..f7907c76c8ccb 100644
--- a/clang/lib/CIR/Dialect/IR/CIRTypes.cpp
+++ b/clang/lib/CIR/Dialect/IR/CIRTypes.cpp
@@ -341,7 +341,7 @@ RecordType::getTypeSizeInBits(const mlir::DataLayout &dataLayout,
   if (isUnion())
     return dataLayout.getTypeSize(getLargestMember(dataLayout));
 
-  unsigned recordSize = computeStructSize(dataLayout);
+  auto recordSize = static_cast<uint64_t>(computeStructSize(dataLayout));
   return llvm::TypeSize::getFixed(recordSize * 8);
 }
 

[andykaylor]

Have you seen a case where this overflow happens? Your change seems good. Normally we require tests for every change, but in this case I don't think that's necessary.

[HendrikHuebner]

I have not tried reproducing it, I just noticed this while reading the code. Itt seems safer to add the cast.