1
0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-24 03:33:20 +01:00
llvm-mirror/test/CodeGen/X86/promote-assert-zext.ll
Bill Wendling 0b9c16295a As Dan pointed out, movzbl, movsbl, and friends are nicer than their alias
(movzx/movsx) because they give more information. Revert that part of the patch.

llvm-svn: 129498
2011-04-14 01:46:37 +00:00

23 lines
693 B
LLVM

; RUN: llc < %s | FileCheck %s
; rdar://8051990
target datalayout = "e-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:64:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64-s0:64:64-f80:128:128-n8:16:32:64"
target triple = "x86_64-apple-darwin11"
; ISel doesn't yet know how to eliminate this extra zero-extend. But until
; it knows how to do so safely, it shouldn;t eliminate it.
; CHECK: movzbl (%rdi), %eax
; CHECK: movzwl %ax, %eax
define i64 @_ZL5matchPKtPKhiR9MatchData(i8* %tmp13) nounwind {
entry:
%tmp14 = load i8* %tmp13, align 1
%tmp17 = zext i8 %tmp14 to i16
br label %bb341
bb341:
%tmp18 = add i16 %tmp17, -1
%tmp23 = sext i16 %tmp18 to i64
ret i64 %tmp23
}