Fix PR#226: When emitting padding, always emit it as bytes. Bytes can be
placed into any alignment situation.
llvm-svn: 11247
diff --git a/llvm/lib/Target/Sparc/EmitAssembly.cpp b/llvm/lib/Target/Sparc/EmitAssembly.cpp
index 5bb2ee6..863ffd7 100644
--- a/llvm/lib/Target/Sparc/EmitAssembly.cpp
+++ b/llvm/lib/Target/Sparc/EmitAssembly.cpp
@@ -184,14 +184,11 @@
}
void PrintZeroBytesToPad(int numBytes) {
- for (/* no init */; numBytes >= 8; numBytes -= 8)
- printSingleConstantValue(Constant::getNullValue(Type::ULongTy));
-
- if (numBytes >= 4) {
- printSingleConstantValue(Constant::getNullValue(Type::UIntTy));
- numBytes -= 4;
- }
-
+ //
+ // Always use single unsigned bytes for padding. We don't know upon
+ // what data size the beginning address is aligned, so using anything
+ // other than a byte may cause alignment errors in the assembler.
+ //
while (numBytes--)
printSingleConstantValue(Constant::getNullValue(Type::UByteTy));
}