llvm-mirror/test/CodeGen/X86/memmove-4.ll
Dan Gohman e256337a1a Expand small memmovs using inline code. Set the X86 threshold for expanding
memmove to a more plausible value, now that it's actually being used.

llvm-svn: 51696
2008-05-29 19:42:22 +00:00

13 lines
349 B
LLVM

; RUN: llvm-as < %s | llc | not grep call
target triple = "i686-pc-linux-gnu"
define void @a(i8* %a, i8* %b) nounwind {
%tmp2 = bitcast i8* %a to i8*
%tmp3 = bitcast i8* %b to i8*
tail call void @llvm.memmove.i32( i8* %tmp2, i8* %tmp3, i32 12, i32 4 )
ret void
}
declare void @llvm.memmove.i32(i8*, i8*, i32, i32)