stack-align-memcpy.ll
9.26 KB
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
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -stackrealign -mtriple i386-apple-darwin -mcpu=i486 | FileCheck %s
%struct.foo = type { [88 x i8] }
declare void @bar(i8* nocapture, %struct.foo* align 4 byval) nounwind
declare void @baz(i8*) nounwind
; PR15249
; We can't use rep;movsl here because it clobbers the base pointer in %esi.
define void @test1(%struct.foo* nocapture %x, i32 %y) nounwind {
; CHECK-LABEL: test1:
; CHECK: ## %bb.0:
; CHECK-NEXT: pushl %ebp
; CHECK-NEXT: movl %esp, %ebp
; CHECK-NEXT: pushl %ebx
; CHECK-NEXT: pushl %edi
; CHECK-NEXT: pushl %esi
; CHECK-NEXT: andl $-16, %esp
; CHECK-NEXT: subl $80, %esp
; CHECK-NEXT: movl %esp, %esi
; CHECK-NEXT: movl 8(%ebp), %ecx
; CHECK-NEXT: movl 12(%ebp), %edx
; CHECK-NEXT: movl %esp, %eax
; CHECK-NEXT: addl $15, %edx
; CHECK-NEXT: andl $-16, %edx
; CHECK-NEXT: subl %edx, %eax
; CHECK-NEXT: movl %eax, %esp
; CHECK-NEXT: subl $4, %esp
; CHECK-NEXT: movl 84(%ecx), %edx
; CHECK-NEXT: movl %edx, 68(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 80(%ecx), %edi
; CHECK-NEXT: movl 76(%ecx), %ebx
; CHECK-NEXT: movl 72(%ecx), %edx
; CHECK-NEXT: movl %edx, 64(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 68(%ecx), %edx
; CHECK-NEXT: movl %edx, 60(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 64(%ecx), %edx
; CHECK-NEXT: movl %edx, 56(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 60(%ecx), %edx
; CHECK-NEXT: movl %edx, 52(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 56(%ecx), %edx
; CHECK-NEXT: movl %edx, 48(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 52(%ecx), %edx
; CHECK-NEXT: movl %edx, 44(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 48(%ecx), %edx
; CHECK-NEXT: movl %edx, 40(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 44(%ecx), %edx
; CHECK-NEXT: movl %edx, 36(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 40(%ecx), %edx
; CHECK-NEXT: movl %edx, 32(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 36(%ecx), %edx
; CHECK-NEXT: movl %edx, 28(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 32(%ecx), %edx
; CHECK-NEXT: movl %edx, 24(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 28(%ecx), %edx
; CHECK-NEXT: movl %edx, 20(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 24(%ecx), %edx
; CHECK-NEXT: movl %edx, 16(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 20(%ecx), %edx
; CHECK-NEXT: movl %edx, 12(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 16(%ecx), %edx
; CHECK-NEXT: movl %edx, 8(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 12(%ecx), %edx
; CHECK-NEXT: movl %edx, 4(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 8(%ecx), %edx
; CHECK-NEXT: movl %edx, (%esi) ## 4-byte Spill
; CHECK-NEXT: movl (%ecx), %edx
; CHECK-NEXT: movl %edx, 72(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 4(%ecx), %ecx
; CHECK-NEXT: pushl 68(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl %edi
; CHECK-NEXT: pushl %ebx
; CHECK-NEXT: pushl 64(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 60(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 56(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 52(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 48(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 44(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 40(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 36(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 32(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 28(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 24(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 20(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 16(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 12(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 8(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 4(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl (%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl %ecx
; CHECK-NEXT: pushl 72(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl %eax
; CHECK-NEXT: calll _bar
; CHECK-NEXT: leal -12(%ebp), %esp
; CHECK-NEXT: popl %esi
; CHECK-NEXT: popl %edi
; CHECK-NEXT: popl %ebx
; CHECK-NEXT: popl %ebp
; CHECK-NEXT: retl
%dynalloc = alloca i8, i32 %y, align 1
call void @bar(i8* %dynalloc, %struct.foo* align 4 byval %x)
ret void
}
; PR19012
; Also don't clobber %esi if the dynamic alloca comes after the memcpy.
define void @test2(%struct.foo* nocapture %x, i32 %y, i8* %z) nounwind {
; CHECK-LABEL: test2:
; CHECK: ## %bb.0:
; CHECK-NEXT: pushl %ebp
; CHECK-NEXT: movl %esp, %ebp
; CHECK-NEXT: pushl %ebx
; CHECK-NEXT: pushl %edi
; CHECK-NEXT: pushl %esi
; CHECK-NEXT: andl $-16, %esp
; CHECK-NEXT: subl $80, %esp
; CHECK-NEXT: movl %esp, %esi
; CHECK-NEXT: movl 12(%ebp), %edi
; CHECK-NEXT: movl 8(%ebp), %eax
; CHECK-NEXT: subl $4, %esp
; CHECK-NEXT: movl 84(%eax), %ecx
; CHECK-NEXT: movl %ecx, 68(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 80(%eax), %edx
; CHECK-NEXT: movl 76(%eax), %ebx
; CHECK-NEXT: movl 72(%eax), %ecx
; CHECK-NEXT: movl %ecx, 64(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 68(%eax), %ecx
; CHECK-NEXT: movl %ecx, 60(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 64(%eax), %ecx
; CHECK-NEXT: movl %ecx, 56(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 60(%eax), %ecx
; CHECK-NEXT: movl %ecx, 52(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 56(%eax), %ecx
; CHECK-NEXT: movl %ecx, 48(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 52(%eax), %ecx
; CHECK-NEXT: movl %ecx, 44(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 48(%eax), %ecx
; CHECK-NEXT: movl %ecx, 40(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 44(%eax), %ecx
; CHECK-NEXT: movl %ecx, 36(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 40(%eax), %ecx
; CHECK-NEXT: movl %ecx, 32(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 36(%eax), %ecx
; CHECK-NEXT: movl %ecx, 28(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 32(%eax), %ecx
; CHECK-NEXT: movl %ecx, 24(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 28(%eax), %ecx
; CHECK-NEXT: movl %ecx, 20(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 24(%eax), %ecx
; CHECK-NEXT: movl %ecx, 16(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 20(%eax), %ecx
; CHECK-NEXT: movl %ecx, 12(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 16(%eax), %ecx
; CHECK-NEXT: movl %ecx, 8(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 12(%eax), %ecx
; CHECK-NEXT: movl %ecx, 4(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 8(%eax), %ecx
; CHECK-NEXT: movl %ecx, (%esi) ## 4-byte Spill
; CHECK-NEXT: movl (%eax), %ecx
; CHECK-NEXT: movl %ecx, 72(%esi) ## 4-byte Spill
; CHECK-NEXT: movl 4(%eax), %eax
; CHECK-NEXT: pushl 68(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl %edx
; CHECK-NEXT: pushl %ebx
; CHECK-NEXT: pushl 64(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 60(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 56(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 52(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 48(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 44(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 40(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 36(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 32(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 28(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 24(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 20(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 16(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 12(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 8(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 4(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl (%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl %eax
; CHECK-NEXT: pushl 72(%esi) ## 4-byte Folded Reload
; CHECK-NEXT: pushl 16(%ebp)
; CHECK-NEXT: calll _bar
; CHECK-NEXT: addl $96, %esp
; CHECK-NEXT: movl %esp, %eax
; CHECK-NEXT: addl $15, %edi
; CHECK-NEXT: andl $-16, %edi
; CHECK-NEXT: subl %edi, %eax
; CHECK-NEXT: movl %eax, %esp
; CHECK-NEXT: subl $12, %esp
; CHECK-NEXT: pushl %eax
; CHECK-NEXT: calll _baz
; CHECK-NEXT: leal -12(%ebp), %esp
; CHECK-NEXT: popl %esi
; CHECK-NEXT: popl %edi
; CHECK-NEXT: popl %ebx
; CHECK-NEXT: popl %ebp
; CHECK-NEXT: retl
call void @bar(i8* %z, %struct.foo* align 4 byval %x)
%dynalloc = alloca i8, i32 %y, align 1
call void @baz(i8* %dynalloc)
ret void
}
; Check that we do use rep movs if we make the alloca static.
define void @test3(%struct.foo* nocapture %x, i32 %y, i8* %z) nounwind {
; CHECK-LABEL: test3:
; CHECK: ## %bb.0:
; CHECK-NEXT: pushl %ebp
; CHECK-NEXT: movl %esp, %ebp
; CHECK-NEXT: pushl %edi
; CHECK-NEXT: pushl %esi
; CHECK-NEXT: andl $-16, %esp
; CHECK-NEXT: subl $112, %esp
; CHECK-NEXT: movl 16(%ebp), %eax
; CHECK-NEXT: movl 8(%ebp), %esi
; CHECK-NEXT: leal {{[0-9]+}}(%esp), %edi
; CHECK-NEXT: movl $22, %ecx
; CHECK-NEXT: rep;movsl (%esi), %es:(%edi)
; CHECK-NEXT: movl %eax, (%esp)
; CHECK-NEXT: calll _bar
; CHECK-NEXT: leal {{[0-9]+}}(%esp), %eax
; CHECK-NEXT: movl %eax, (%esp)
; CHECK-NEXT: calll _baz
; CHECK-NEXT: leal -8(%ebp), %esp
; CHECK-NEXT: popl %esi
; CHECK-NEXT: popl %edi
; CHECK-NEXT: popl %ebp
; CHECK-NEXT: retl
call void @bar(i8* %z, %struct.foo* align 4 byval %x)
%statalloc = alloca i8, i32 8, align 1
call void @baz(i8* %statalloc)
ret void
}