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
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=x86_64-linux | FileCheck %s -check-prefix=CHECK -check-prefix=ENABLED
; RUN: llc --disable-x86-lea-opt < %s -mtriple=x86_64-linux | FileCheck %s -check-prefix=CHECK -check-prefix=DISABLED
%struct.anon1 = type { i32, i32, i32 }
%struct.anon2 = type { i32, [32 x i32], i32 }
@arr1 = external global [65 x %struct.anon1], align 16
@arr2 = external global [65 x %struct.anon2], align 16
define void @test1(i64 %x) nounwind {
; ENABLED-LABEL: test1:
; ENABLED: # %bb.0: # %entry
; ENABLED-NEXT: shlq $2, %rdi
; ENABLED-NEXT: movl arr1(%rdi,%rdi,2), %ecx
; ENABLED-NEXT: leaq arr1+4(%rdi,%rdi,2), %rax
; ENABLED-NEXT: subl arr1+4(%rdi,%rdi,2), %ecx
; ENABLED-NEXT: addl arr1+8(%rdi,%rdi,2), %ecx
; ENABLED-NEXT: cmpl $2, %ecx
; ENABLED-NEXT: je .LBB0_3
; ENABLED-NEXT: # %bb.1: # %entry
; ENABLED-NEXT: cmpl $1, %ecx
; ENABLED-NEXT: jne .LBB0_4
; ENABLED-NEXT: # %bb.2: # %sw.bb.1
; ENABLED-NEXT: movl $111, (%rax)
; ENABLED-NEXT: movl $222, 4(%rax)
; ENABLED-NEXT: retq
; ENABLED-NEXT: .LBB0_3: # %sw.bb.2
; ENABLED-NEXT: movl $333, (%rax) # imm = 0x14D
; ENABLED-NEXT: movl $444, 4(%rax) # imm = 0x1BC
; ENABLED-NEXT: .LBB0_4: # %sw.epilog
; ENABLED-NEXT: retq
;
; DISABLED-LABEL: test1:
; DISABLED: # %bb.0: # %entry
; DISABLED-NEXT: shlq $2, %rdi
; DISABLED-NEXT: movl arr1(%rdi,%rdi,2), %edx
; DISABLED-NEXT: leaq arr1+4(%rdi,%rdi,2), %rax
; DISABLED-NEXT: subl arr1+4(%rdi,%rdi,2), %edx
; DISABLED-NEXT: leaq arr1+8(%rdi,%rdi,2), %rcx
; DISABLED-NEXT: addl arr1+8(%rdi,%rdi,2), %edx
; DISABLED-NEXT: cmpl $2, %edx
; DISABLED-NEXT: je .LBB0_3
; DISABLED-NEXT: # %bb.1: # %entry
; DISABLED-NEXT: cmpl $1, %edx
; DISABLED-NEXT: jne .LBB0_4
; DISABLED-NEXT: # %bb.2: # %sw.bb.1
; DISABLED-NEXT: movl $111, (%rax)
; DISABLED-NEXT: movl $222, (%rcx)
; DISABLED-NEXT: retq
; DISABLED-NEXT: .LBB0_3: # %sw.bb.2
; DISABLED-NEXT: movl $333, (%rax) # imm = 0x14D
; DISABLED-NEXT: movl $444, (%rcx) # imm = 0x1BC
; DISABLED-NEXT: .LBB0_4: # %sw.epilog
; DISABLED-NEXT: retq
entry:
%a = getelementptr inbounds [65 x %struct.anon1], [65 x %struct.anon1]* @arr1, i64 0, i64 %x, i32 0
%tmp = load i32, i32* %a, align 4
%b = getelementptr inbounds [65 x %struct.anon1], [65 x %struct.anon1]* @arr1, i64 0, i64 %x, i32 1
%tmp1 = load i32, i32* %b, align 4
%sub = sub i32 %tmp, %tmp1
%c = getelementptr inbounds [65 x %struct.anon1], [65 x %struct.anon1]* @arr1, i64 0, i64 %x, i32 2
%tmp2 = load i32, i32* %c, align 4
%add = add nsw i32 %sub, %tmp2
switch i32 %add, label %sw.epilog [
i32 1, label %sw.bb.1
i32 2, label %sw.bb.2
]
sw.bb.1: ; preds = %entry
store i32 111, i32* %b, align 4
store i32 222, i32* %c, align 4
br label %sw.epilog
sw.bb.2: ; preds = %entry
store i32 333, i32* %b, align 4
store i32 444, i32* %c, align 4
br label %sw.epilog
sw.epilog: ; preds = %sw.bb.2, %sw.bb.1, %entry
ret void
}
define void @test2(i64 %x) nounwind optsize {
; ENABLED-LABEL: test2:
; ENABLED: # %bb.0: # %entry
; ENABLED-NEXT: shlq $2, %rdi
; ENABLED-NEXT: leaq arr1+4(%rdi,%rdi,2), %rax
; ENABLED-NEXT: movl -4(%rax), %ecx
; ENABLED-NEXT: subl (%rax), %ecx
; ENABLED-NEXT: addl 4(%rax), %ecx
; ENABLED-NEXT: cmpl $2, %ecx
; ENABLED-NEXT: je .LBB1_3
; ENABLED-NEXT: # %bb.1: # %entry
; ENABLED-NEXT: cmpl $1, %ecx
; ENABLED-NEXT: jne .LBB1_4
; ENABLED-NEXT: # %bb.2: # %sw.bb.1
; ENABLED-NEXT: movl $111, (%rax)
; ENABLED-NEXT: movl $222, 4(%rax)
; ENABLED-NEXT: retq
; ENABLED-NEXT: .LBB1_3: # %sw.bb.2
; ENABLED-NEXT: movl $333, (%rax) # imm = 0x14D
; ENABLED-NEXT: movl $444, 4(%rax) # imm = 0x1BC
; ENABLED-NEXT: .LBB1_4: # %sw.epilog
; ENABLED-NEXT: retq
;
; DISABLED-LABEL: test2:
; DISABLED: # %bb.0: # %entry
; DISABLED-NEXT: shlq $2, %rdi
; DISABLED-NEXT: movl arr1(%rdi,%rdi,2), %edx
; DISABLED-NEXT: leaq arr1+4(%rdi,%rdi,2), %rax
; DISABLED-NEXT: subl arr1+4(%rdi,%rdi,2), %edx
; DISABLED-NEXT: leaq arr1+8(%rdi,%rdi,2), %rcx
; DISABLED-NEXT: addl arr1+8(%rdi,%rdi,2), %edx
; DISABLED-NEXT: cmpl $2, %edx
; DISABLED-NEXT: je .LBB1_3
; DISABLED-NEXT: # %bb.1: # %entry
; DISABLED-NEXT: cmpl $1, %edx
; DISABLED-NEXT: jne .LBB1_4
; DISABLED-NEXT: # %bb.2: # %sw.bb.1
; DISABLED-NEXT: movl $111, (%rax)
; DISABLED-NEXT: movl $222, (%rcx)
; DISABLED-NEXT: retq
; DISABLED-NEXT: .LBB1_3: # %sw.bb.2
; DISABLED-NEXT: movl $333, (%rax) # imm = 0x14D
; DISABLED-NEXT: movl $444, (%rcx) # imm = 0x1BC
; DISABLED-NEXT: .LBB1_4: # %sw.epilog
; DISABLED-NEXT: retq
entry:
%a = getelementptr inbounds [65 x %struct.anon1], [65 x %struct.anon1]* @arr1, i64 0, i64 %x, i32 0
%tmp = load i32, i32* %a, align 4
%b = getelementptr inbounds [65 x %struct.anon1], [65 x %struct.anon1]* @arr1, i64 0, i64 %x, i32 1
%tmp1 = load i32, i32* %b, align 4
%sub = sub i32 %tmp, %tmp1
%c = getelementptr inbounds [65 x %struct.anon1], [65 x %struct.anon1]* @arr1, i64 0, i64 %x, i32 2
%tmp2 = load i32, i32* %c, align 4
%add = add nsw i32 %sub, %tmp2
switch i32 %add, label %sw.epilog [
i32 1, label %sw.bb.1
i32 2, label %sw.bb.2
]
sw.bb.1: ; preds = %entry
store i32 111, i32* %b, align 4
store i32 222, i32* %c, align 4
br label %sw.epilog
sw.bb.2: ; preds = %entry
store i32 333, i32* %b, align 4
store i32 444, i32* %c, align 4
br label %sw.epilog
sw.epilog: ; preds = %sw.bb.2, %sw.bb.1, %entry
ret void
}
; Check that LEA optimization pass takes into account a resultant address
; displacement when choosing a LEA instruction for replacing a redundant
; address recalculation.
define void @test3(i64 %x) nounwind optsize {
; ENABLED-LABEL: test3:
; ENABLED: # %bb.0: # %entry
; ENABLED-NEXT: movq %rdi, %rax
; ENABLED-NEXT: shlq $7, %rax
; ENABLED-NEXT: leaq arr2+132(%rax,%rdi,8), %rcx
; ENABLED-NEXT: leaq arr2(%rax,%rdi,8), %rax
; ENABLED-NEXT: movl (%rcx), %edx
; ENABLED-NEXT: addl (%rax), %edx
; ENABLED-NEXT: cmpl $2, %edx
; ENABLED-NEXT: je .LBB2_3
; ENABLED-NEXT: # %bb.1: # %entry
; ENABLED-NEXT: cmpl $1, %edx
; ENABLED-NEXT: jne .LBB2_4
; ENABLED-NEXT: # %bb.2: # %sw.bb.1
; ENABLED-NEXT: movl $111, (%rcx)
; ENABLED-NEXT: movl $222, (%rax)
; ENABLED-NEXT: retq
; ENABLED-NEXT: .LBB2_3: # %sw.bb.2
; ENABLED-NEXT: movl $333, (%rcx) # imm = 0x14D
; ENABLED-NEXT: movl %eax, (%rax)
; ENABLED-NEXT: .LBB2_4: # %sw.epilog
; ENABLED-NEXT: retq
;
; DISABLED-LABEL: test3:
; DISABLED: # %bb.0: # %entry
; DISABLED-NEXT: movq %rdi, %rsi
; DISABLED-NEXT: shlq $7, %rsi
; DISABLED-NEXT: leaq arr2+132(%rsi,%rdi,8), %rcx
; DISABLED-NEXT: leaq arr2(%rsi,%rdi,8), %rax
; DISABLED-NEXT: movl arr2+132(%rsi,%rdi,8), %edx
; DISABLED-NEXT: addl arr2(%rsi,%rdi,8), %edx
; DISABLED-NEXT: cmpl $2, %edx
; DISABLED-NEXT: je .LBB2_3
; DISABLED-NEXT: # %bb.1: # %entry
; DISABLED-NEXT: cmpl $1, %edx
; DISABLED-NEXT: jne .LBB2_4
; DISABLED-NEXT: # %bb.2: # %sw.bb.1
; DISABLED-NEXT: movl $111, (%rcx)
; DISABLED-NEXT: movl $222, (%rax)
; DISABLED-NEXT: retq
; DISABLED-NEXT: .LBB2_3: # %sw.bb.2
; DISABLED-NEXT: movl $333, (%rcx) # imm = 0x14D
; DISABLED-NEXT: movl %eax, (%rax)
; DISABLED-NEXT: .LBB2_4: # %sw.epilog
; DISABLED-NEXT: retq
entry:
%a = getelementptr inbounds [65 x %struct.anon2], [65 x %struct.anon2]* @arr2, i64 0, i64 %x, i32 2
%tmp = load i32, i32* %a, align 4
%b = getelementptr inbounds [65 x %struct.anon2], [65 x %struct.anon2]* @arr2, i64 0, i64 %x, i32 0
%tmp1 = load i32, i32* %b, align 4
%add = add nsw i32 %tmp, %tmp1
switch i32 %add, label %sw.epilog [
i32 1, label %sw.bb.1
i32 2, label %sw.bb.2
]
sw.bb.1: ; preds = %entry
store i32 111, i32* %a, align 4
store i32 222, i32* %b, align 4
br label %sw.epilog
sw.bb.2: ; preds = %entry
store i32 333, i32* %a, align 4
; Make sure the REG3's definition LEA won't be removed as redundant.
%cvt = ptrtoint i32* %b to i32
store i32 %cvt, i32* %b, align 4
br label %sw.epilog
sw.epilog: ; preds = %sw.bb.2, %sw.bb.1, %entry
ret void
; REG3's definition is closer to movl than REG2's, but the pass still chooses
; REG2 because it provides the resultant address displacement fitting 1 byte.
}
define void @test4(i64 %x) nounwind minsize {
; ENABLED-LABEL: test4:
; ENABLED: # %bb.0: # %entry
; ENABLED-NEXT: imulq $12, %rdi, %rax
; ENABLED-NEXT: leaq arr1+4(%rax), %rax
; ENABLED-NEXT: movl -4(%rax), %ecx
; ENABLED-NEXT: subl (%rax), %ecx
; ENABLED-NEXT: addl 4(%rax), %ecx
; ENABLED-NEXT: cmpl $2, %ecx
; ENABLED-NEXT: je .LBB3_3
; ENABLED-NEXT: # %bb.1: # %entry
; ENABLED-NEXT: cmpl $1, %ecx
; ENABLED-NEXT: jne .LBB3_4
; ENABLED-NEXT: # %bb.2: # %sw.bb.1
; ENABLED-NEXT: movl $111, (%rax)
; ENABLED-NEXT: movl $222, 4(%rax)
; ENABLED-NEXT: retq
; ENABLED-NEXT: .LBB3_3: # %sw.bb.2
; ENABLED-NEXT: movl $333, (%rax) # imm = 0x14D
; ENABLED-NEXT: movl $444, 4(%rax) # imm = 0x1BC
; ENABLED-NEXT: .LBB3_4: # %sw.epilog
; ENABLED-NEXT: retq
;
; DISABLED-LABEL: test4:
; DISABLED: # %bb.0: # %entry
; DISABLED-NEXT: imulq $12, %rdi, %rsi
; DISABLED-NEXT: movl arr1(%rsi), %edx
; DISABLED-NEXT: leaq arr1+4(%rsi), %rax
; DISABLED-NEXT: subl arr1+4(%rsi), %edx
; DISABLED-NEXT: leaq arr1+8(%rsi), %rcx
; DISABLED-NEXT: addl arr1+8(%rsi), %edx
; DISABLED-NEXT: cmpl $2, %edx
; DISABLED-NEXT: je .LBB3_3
; DISABLED-NEXT: # %bb.1: # %entry
; DISABLED-NEXT: cmpl $1, %edx
; DISABLED-NEXT: jne .LBB3_4
; DISABLED-NEXT: # %bb.2: # %sw.bb.1
; DISABLED-NEXT: movl $111, (%rax)
; DISABLED-NEXT: movl $222, (%rcx)
; DISABLED-NEXT: retq
; DISABLED-NEXT: .LBB3_3: # %sw.bb.2
; DISABLED-NEXT: movl $333, (%rax) # imm = 0x14D
; DISABLED-NEXT: movl $444, (%rcx) # imm = 0x1BC
; DISABLED-NEXT: .LBB3_4: # %sw.epilog
; DISABLED-NEXT: retq
entry:
%a = getelementptr inbounds [65 x %struct.anon1], [65 x %struct.anon1]* @arr1, i64 0, i64 %x, i32 0
%tmp = load i32, i32* %a, align 4
%b = getelementptr inbounds [65 x %struct.anon1], [65 x %struct.anon1]* @arr1, i64 0, i64 %x, i32 1
%tmp1 = load i32, i32* %b, align 4
%sub = sub i32 %tmp, %tmp1
%c = getelementptr inbounds [65 x %struct.anon1], [65 x %struct.anon1]* @arr1, i64 0, i64 %x, i32 2
%tmp2 = load i32, i32* %c, align 4
%add = add nsw i32 %sub, %tmp2
switch i32 %add, label %sw.epilog [
i32 1, label %sw.bb.1
i32 2, label %sw.bb.2
]
sw.bb.1: ; preds = %entry
store i32 111, i32* %b, align 4
store i32 222, i32* %c, align 4
br label %sw.epilog
sw.bb.2: ; preds = %entry
store i32 333, i32* %b, align 4
store i32 444, i32* %c, align 4
br label %sw.epilog
sw.epilog: ; preds = %sw.bb.2, %sw.bb.1, %entry
ret void
}
|
