1 ; RUN: opt < %s -scalarrepl -S | FileCheck %s
4 target datalayout = "e-p:32:32:32-i1:8:32-i8:8:32-i16:16:32-i32:32:32-i64:32:32-f32:32:32-f64:32:32-v64:64:64-v128:128:128-a0:0:32-n32"
5 target triple = "thumbv7-apple-darwin10"
7 %struct.__neon_int16x8x2_t = type { <8 x i16>, <8 x i16> }
8 %struct.int16x8_t = type { <8 x i16> }
9 %struct.int16x8x2_t = type { [2 x %struct.int16x8_t] }
10 %union..0anon = type { %struct.int16x8x2_t }
12 define void @test(<8 x i16> %tmp.0, %struct.int16x8x2_t* %dst) nounwind {
15 ; CHECK: "alloca point"
16 ; CHECK: store <8 x i16>
17 ; CHECK: store <8 x i16>
20 %tmp_addr = alloca %struct.int16x8_t
21 %dst_addr = alloca %struct.int16x8x2_t*
22 %__rv = alloca %union..0anon
23 %__bx = alloca %struct.int16x8_t
24 %__ax = alloca %struct.int16x8_t
25 %tmp2 = alloca %struct.int16x8x2_t
26 %0 = alloca %struct.int16x8x2_t
27 %"alloca point" = bitcast i32 0 to i32
28 %1 = getelementptr inbounds %struct.int16x8_t* %tmp_addr, i32 0, i32 0
29 store <8 x i16> %tmp.0, <8 x i16>* %1
30 store %struct.int16x8x2_t* %dst, %struct.int16x8x2_t** %dst_addr
31 %2 = getelementptr inbounds %struct.int16x8_t* %__ax, i32 0, i32 0
32 %3 = getelementptr inbounds %struct.int16x8_t* %tmp_addr, i32 0, i32 0
33 %4 = load <8 x i16>* %3, align 16
34 store <8 x i16> %4, <8 x i16>* %2, align 16
35 %5 = getelementptr inbounds %struct.int16x8_t* %__bx, i32 0, i32 0
36 %6 = getelementptr inbounds %struct.int16x8_t* %tmp_addr, i32 0, i32 0
37 %7 = load <8 x i16>* %6, align 16
38 store <8 x i16> %7, <8 x i16>* %5, align 16
39 %8 = getelementptr inbounds %struct.int16x8_t* %__ax, i32 0, i32 0
40 %9 = load <8 x i16>* %8, align 16
41 %10 = getelementptr inbounds %struct.int16x8_t* %__bx, i32 0, i32 0
42 %11 = load <8 x i16>* %10, align 16
43 %12 = getelementptr inbounds %union..0anon* %__rv, i32 0, i32 0
44 %13 = bitcast %struct.int16x8x2_t* %12 to %struct.__neon_int16x8x2_t*
45 %14 = shufflevector <8 x i16> %9, <8 x i16> %11, <8 x i32> <i32 0, i32 8, i32 2, i32 10, i32 4, i32 12, i32 6, i32 14>
46 %15 = getelementptr inbounds %struct.__neon_int16x8x2_t* %13, i32 0, i32 0
47 store <8 x i16> %14, <8 x i16>* %15
48 %16 = shufflevector <8 x i16> %9, <8 x i16> %11, <8 x i32> <i32 1, i32 9, i32 3, i32 11, i32 5, i32 13, i32 7, i32 15>
49 %17 = getelementptr inbounds %struct.__neon_int16x8x2_t* %13, i32 0, i32 1
50 store <8 x i16> %16, <8 x i16>* %17
51 %18 = getelementptr inbounds %union..0anon* %__rv, i32 0, i32 0
52 %19 = bitcast %struct.int16x8x2_t* %0 to i8*
53 %20 = bitcast %struct.int16x8x2_t* %18 to i8*
54 call void @llvm.memcpy.p0i8.p0i8.i32(i8* %19, i8* %20, i32 32, i32 16, i1 false)
55 %tmp21 = bitcast %struct.int16x8x2_t* %tmp2 to i8*
56 %21 = bitcast %struct.int16x8x2_t* %0 to i8*
57 call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp21, i8* %21, i32 32, i32 16, i1 false)
58 %22 = load %struct.int16x8x2_t** %dst_addr, align 4
59 %23 = bitcast %struct.int16x8x2_t* %22 to i8*
60 %tmp22 = bitcast %struct.int16x8x2_t* %tmp2 to i8*
61 call void @llvm.memcpy.p0i8.p0i8.i32(i8* %23, i8* %tmp22, i32 32, i32 16, i1 false)
64 return: ; preds = %entry
69 %struct._NSRange = type { i64 }
71 define void @test_memcpy_self() nounwind {
73 %range = alloca %struct._NSRange
74 br i1 undef, label %cond.true, label %cond.false
76 cond.true: ; preds = %entry
77 %tmp3 = bitcast %struct._NSRange* %range to i8*
78 %tmp4 = bitcast %struct._NSRange* %range to i8*
79 call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp3, i8* %tmp4, i32 8, i32 8, i1 false)
82 cond.false: ; preds = %entry
85 ; CHECK: @test_memcpy_self
90 declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind