1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
| ; RUN: llc -mtriple=thumb-eabi < %s -o - | FileCheck %s
; Check that stack addresses are generated using a single ADD
define void @test1(i8** %p) {
%x = alloca i8, align 1
%y = alloca i8, align 1
%z = alloca i8, align 1
; CHECK: add r1, sp, #8
; CHECK: str r1, [r0]
store volatile i8* %x, i8** %p, align 4
; CHECK: add r1, sp, #4
; CHECK: str r1, [r0]
store volatile i8* %y, i8** %p, align 4
; CHECK: mov r1, sp
; CHECK: str r1, [r0]
store volatile i8* %z, i8** %p, align 4
ret void
}
; Stack offsets larger than 1020 still need two ADDs
define void @test2([1024 x i8]** %p) {
%arr1 = alloca [1024 x i8], align 1
%arr2 = alloca [1024 x i8], align 1
; CHECK: add r1, sp, #1020
; CHECK: adds r1, #4
; CHECK: str r1, [r0]
store volatile [1024 x i8]* %arr1, [1024 x i8]** %p, align 4
; CHECK: mov r1, sp
; CHECK: str r1, [r0]
store volatile [1024 x i8]* %arr2, [1024 x i8]** %p, align 4
ret void
}
; If possible stack-based lrdb/ldrh are widened to use SP-based addressing
define i32 @test3() #0 {
%x = alloca i8, align 1
%y = alloca i8, align 1
; CHECK: ldr r0, [sp]
%1 = load i8, i8* %x, align 1
; CHECK: ldr r1, [sp, #4]
%2 = load i8, i8* %y, align 1
%3 = add nsw i8 %1, %2
%4 = zext i8 %3 to i32
ret i32 %4
}
define i32 @test4() #0 {
%x = alloca i16, align 2
%y = alloca i16, align 2
; CHECK: ldr r0, [sp]
%1 = load i16, i16* %x, align 2
; CHECK: ldr r1, [sp, #4]
%2 = load i16, i16* %y, align 2
%3 = add nsw i16 %1, %2
%4 = zext i16 %3 to i32
ret i32 %4
}
; Don't widen if the value needs to be zero-extended
define zeroext i8 @test5() {
%x = alloca i8, align 1
; CHECK: mov r0, sp
; CHECK: ldrb r0, [r0]
%1 = load i8, i8* %x, align 1
ret i8 %1
}
define zeroext i16 @test6() {
%x = alloca i16, align 2
; CHECK: mov r0, sp
; CHECK: ldrh r0, [r0]
%1 = load i16, i16* %x, align 2
ret i16 %1
}
; Accessing the bottom of a large array shouldn't require materializing a base
;
; CHECK: movs [[REG:r[0-9]+]], #1
; CHECK: str [[REG]], [sp, #16]
; CHECK: str [[REG]], [sp, #4]
define void @test7() {
%arr = alloca [200 x i32], align 4
%arrayidx = getelementptr inbounds [200 x i32], [200 x i32]* %arr, i32 0, i32 1
store i32 1, i32* %arrayidx, align 4
%arrayidx1 = getelementptr inbounds [200 x i32], [200 x i32]* %arr, i32 0, i32 4
store i32 1, i32* %arrayidx1, align 4
ret void
}
; Check that loads/stores with out-of-range offsets are handled correctly
define void @test8() {
%arr3 = alloca [224 x i32], align 4
%arr2 = alloca [224 x i32], align 4
%arr1 = alloca [224 x i32], align 4
; CHECK: movs [[REG:r[0-9]+]], #1
; CHECK-DAG: str [[REG]], [sp]
%arr1idx1 = getelementptr inbounds [224 x i32], [224 x i32]* %arr1, i32 0, i32 0
store i32 1, i32* %arr1idx1, align 4
; Offset in range for sp-based store, but not for non-sp-based store
; CHECK-DAG: str [[REG]], [sp, #128]
%arr1idx2 = getelementptr inbounds [224 x i32], [224 x i32]* %arr1, i32 0, i32 32
store i32 1, i32* %arr1idx2, align 4
; CHECK-DAG: str [[REG]], [sp, #896]
%arr2idx1 = getelementptr inbounds [224 x i32], [224 x i32]* %arr2, i32 0, i32 0
store i32 1, i32* %arr2idx1, align 4
; %arr2 is in range, but this element of it is not
; CHECK-DAG: ldr [[RA:r[0-9]+]], .LCPI7_2
; CHECK-DAG: add [[RA]], sp
; CHECK-DAG: str [[REG]], [{{r[0-9]+}}]
%arr2idx2 = getelementptr inbounds [224 x i32], [224 x i32]* %arr2, i32 0, i32 32
store i32 1, i32* %arr2idx2, align 4
; %arr3 is not in range
; CHECK-DAG: ldr [[RB:r[0-9]+]], .LCPI7_3
; CHECK-DAG: add [[RB]], sp
; CHECK-DAG: str [[REG]], [{{r[0-9]+}}]
%arr3idx1 = getelementptr inbounds [224 x i32], [224 x i32]* %arr3, i32 0, i32 0
store i32 1, i32* %arr3idx1, align 4
; CHECK-DAG: ldr [[RC:r[0-9]+]], .LCPI7_4
; CHECK-DAG: add [[RC]], sp
; CHECK-DAG: str [[REG]], [{{r[0-9]+}}]
%arr3idx2 = getelementptr inbounds [224 x i32], [224 x i32]* %arr3, i32 0, i32 32
store i32 1, i32* %arr3idx2, align 4
ret void
}
|