When a function takes a variable number of pointer arguments, with a zero

pointer marking the end of the list, the zero *must* be cast to the pointer
type.  An un-cast zero is a 32-bit int, and at least on x86_64, gcc will
not extend the zero to 64 bits, thus allowing the upper 32 bits to be
random junk.

The new END_WITH_NULL macro may be used to annotate a such a function
so that GCC (version 4 or newer) will detect the use of un-casted zero
at compile time.

llvm-svn: 23888

1 files changed, 2 insertions, 1 deletions

diff --git a/llvm/lib/Transforms/Instrumentation/ProfilingUtils.cpp b/llvm/lib/Transforms/Instrumentation/ProfilingUtils.cpp
index 4093759e4d2..82e7ae7ba15 100644
--- a/llvm/lib/Transforms/Instrumentation/ProfilingUtils.cpp
+++ b/llvm/lib/Transforms/Instrumentation/ProfilingUtils.cpp
@@ -26,7 +26,8 @@ void llvm::InsertProfilingInitCall(Function *MainFn, const char *FnName,
   const PointerType *UIntPtr = PointerType::get(Type::UIntTy);
   Module &M = *MainFn->getParent();
   Function *InitFn = M.getOrInsertFunction(FnName, Type::IntTy, Type::IntTy,
-                                           ArgVTy, UIntPtr, Type::UIntTy, 0);
+                                           ArgVTy, UIntPtr, Type::UIntTy,
+                                           (Type *)0);
 
   // This could force argc and argv into programs that wouldn't otherwise have
   // them, but instead we just pass null values in.