aboutsummaryrefslogtreecommitdiff
path: root/test/CodeGen/ARM/fast-isel.ll
blob: 41fda41326326d0b61ac8f87cef59ee210b8807a (plain)
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
; RUN: llc < %s -O0 -fast-isel-abort -relocation-model=dynamic-no-pic -mtriple=armv7-apple-ios | FileCheck %s --check-prefix=ARM
; RUN: llc < %s -O0 -fast-isel-abort -relocation-model=dynamic-no-pic -mtriple=thumbv7-apple-ios | FileCheck %s --check-prefix=THUMB
; RUN: llc < %s -O0 -arm-strict-align -relocation-model=dynamic-no-pic -mtriple=armv7-apple-ios | FileCheck %s --check-prefix=ARM-STRICT-ALIGN
; RUN: llc < %s -O0 -arm-strict-align -relocation-model=dynamic-no-pic -mtriple=thumbv7-apple-ios | FileCheck %s --check-prefix=THUMB-STRICT-ALIGN

; Very basic fast-isel functionality.
define i32 @add(i32 %a, i32 %b) nounwind {
entry:
  %a.addr = alloca i32, align 4
  %b.addr = alloca i32, align 4
  store i32 %a, i32* %a.addr
  store i32 %b, i32* %b.addr
  %tmp = load i32* %a.addr
  %tmp1 = load i32* %b.addr
  %add = add nsw i32 %tmp, %tmp1
  ret i32 %add
}

; Check truncate to bool
define void @test1(i32 %tmp) nounwind {
entry:
%tobool = trunc i32 %tmp to i1
br i1 %tobool, label %if.then, label %if.end

if.then:                                          ; preds = %entry
call void @test1(i32 0)
br label %if.end

if.end:                                           ; preds = %if.then, %entry
ret void
; ARM: test1:
; ARM: tst r0, #1
; THUMB: test1:
; THUMB: tst.w r0, #1
}

; Check some simple operations with immediates
define void @test2(i32 %tmp, i32* %ptr) nounwind {
; THUMB: test2:
; ARM: test2:

b1:
  %a = add i32 %tmp, 4096
  store i32 %a, i32* %ptr
  br label %b2

; THUMB: add.w {{.*}} #4096
; ARM: add {{.*}} #4096

b2:
  %b = add i32 %tmp, 4095
  store i32 %b, i32* %ptr
  br label %b3
; THUMB: addw {{.*}} #4095
; ARM: movw {{.*}} #4095
; ARM: add

b3:
  %c = or i32 %tmp, 4
  store i32 %c, i32* %ptr
  ret void

; THUMB: orr {{.*}} #4
; ARM: orr {{.*}} #4
}

define void @test3(i32 %tmp, i32* %ptr1, i16* %ptr2, i8* %ptr3) nounwind {
; THUMB: test3:
; ARM: test3:

bb1:
  %a1 = trunc i32 %tmp to i16
  %a2 = trunc i16 %a1 to i8
  %a3 = trunc i8 %a2 to i1
  %a4 = zext i1 %a3 to i8
  store i8 %a4, i8* %ptr3
  %a5 = zext i8 %a4 to i16
  store i16 %a5, i16* %ptr2
  %a6 = zext i16 %a5 to i32
  store i32 %a6, i32* %ptr1
  br label %bb2

; THUMB: and
; THUMB: strb
; THUMB: uxtb
; THUMB: strh
; THUMB: uxth
; ARM: and
; ARM: strb
; ARM: uxtb
; ARM: strh
; ARM: uxth

bb2:
  %b1 = trunc i32 %tmp to i16
  %b2 = trunc i16 %b1 to i8
  store i8 %b2, i8* %ptr3
  %b3 = sext i8 %b2 to i16
  store i16 %b3, i16* %ptr2
  %b4 = sext i16 %b3 to i32
  store i32 %b4, i32* %ptr1
  br label %bb3

; THUMB: strb
; THUMB: sxtb
; THUMB: strh
; THUMB: sxth
; ARM: strb
; ARM: sxtb
; ARM: strh
; ARM: sxth

bb3:
  %c1 = load i8* %ptr3
  %c2 = load i16* %ptr2
  %c3 = load i32* %ptr1
  %c4 = zext i8 %c1 to i32
  %c5 = sext i16 %c2 to i32
  %c6 = add i32 %c4, %c5
  %c7 = sub i32 %c3, %c6
  store i32 %c7, i32* %ptr1
  ret void

; THUMB: ldrb
; THUMB: ldrh
; THUMB: uxtb
; THUMB: sxth
; THUMB: add
; THUMB: sub
; ARM: ldrb
; ARM: ldrh
; ARM: uxtb
; ARM: sxth
; ARM: add
; ARM: sub
}

; Check loads/stores with globals
@test4g = external global i32

define void @test4() {
  %a = load i32* @test4g
  %b = add i32 %a, 1
  store i32 %b, i32* @test4g
  ret void

; THUMB: movw r0, :lower16:L_test4g$non_lazy_ptr
; THUMB: movt r0, :upper16:L_test4g$non_lazy_ptr
; THUMB: ldr r0, [r0]
; THUMB: ldr r1, [r0]
; THUMB: adds r1, #1
; THUMB: str r1, [r0]

; ARM: movw r0, :lower16:L_test4g$non_lazy_ptr
; ARM: movt r0, :upper16:L_test4g$non_lazy_ptr
; ARM: ldr r0, [r0]
; ARM: ldr r1, [r0]
; ARM: add r1, r1, #1
; ARM: str r1, [r0]
}

; Check unaligned stores
%struct.anon = type <{ float }>

@a = common global %struct.anon* null, align 4

define void @unaligned_store(float %x, float %y) nounwind {
entry:
; ARM: @unaligned_store
; ARM: vmov r1, s0
; ARM: str r1, [r0]

; THUMB: @unaligned_store
; THUMB: vmov r1, s0
; THUMB: str r1, [r0]

  %add = fadd float %x, %y
  %0 = load %struct.anon** @a, align 4
  %x1 = getelementptr inbounds %struct.anon* %0, i32 0, i32 0
  store float %add, float* %x1, align 1
  ret void
}

; Doublewords require only word-alignment.
; rdar://10528060
%struct.anon.0 = type { double }

@foo_unpacked = common global %struct.anon.0 zeroinitializer, align 4

define void @test5(double %a, double %b) nounwind {
entry:
; ARM: @test5
; THUMB: @test5
  %add = fadd double %a, %b
  store double %add, double* getelementptr inbounds (%struct.anon.0* @foo_unpacked, i32 0, i32 0), align 4
; ARM: vstr d16, [r0]
; THUMB: vstr d16, [r0]
  ret void
}

; Check unaligned loads of floats
%class.TAlignTest = type <{ i16, float }>

define zeroext i1 @test6(%class.TAlignTest* %this) nounwind align 2 {
entry:
; ARM: @test6
; THUMB: @test6
  %0 = alloca %class.TAlignTest*, align 4
  store %class.TAlignTest* %this, %class.TAlignTest** %0, align 4
  %1 = load %class.TAlignTest** %0
  %2 = getelementptr inbounds %class.TAlignTest* %1, i32 0, i32 1
  %3 = load float* %2, align 1
  %4 = fcmp une float %3, 0.000000e+00
; ARM: ldr r0, [r0, #2]
; ARM: vmov s0, r0
; ARM: vcmpe.f32 s0, #0
; THUMB: ldr.w r0, [r0, #2]
; THUMB: vmov s0, r0
; THUMB: vcmpe.f32 s0, #0
  ret i1 %4
}

; ARM: @urem_fold
; THUMB: @urem_fold
; ARM: and r0, r0, #31
; THUMB: and r0, r0, #31
define i32 @urem_fold(i32 %a) nounwind {
  %rem = urem i32 %a, 32
  ret i32 %rem
}

define i32 @test7() noreturn nounwind  {
entry:
; ARM: @test7
; THUMB: @test7
; ARM: trap
; THUMB: trap
  tail call void @llvm.trap( )
  unreachable
}

declare void @llvm.trap() nounwind

define void @unaligned_i16_store(i16 %x, i16* %y) nounwind {
entry:
; ARM-STRICT-ALIGN: @unaligned_i16_store
; ARM-STRICT-ALIGN: strb
; ARM-STRICT-ALIGN: strb

; THUMB-STRICT-ALIGN: @unaligned_i16_store
; THUMB-STRICT-ALIGN: strb
; THUMB-STRICT-ALIGN: strb

  store i16 %x, i16* %y, align 1
  ret void
}

define i16 @unaligned_i16_load(i16* %x) nounwind {
entry:
; ARM-STRICT-ALIGN: @unaligned_i16_load
; ARM-STRICT-ALIGN: ldrb
; ARM-STRICT-ALIGN: ldrb

; THUMB-STRICT-ALIGN: @unaligned_i16_load
; THUMB-STRICT-ALIGN: ldrb
; THUMB-STRICT-ALIGN: ldrb

  %0 = load i16* %x, align 1
  ret i16 %0
}

define void @unaligned_i32_store(i32 %x, i32* %y) nounwind {
entry:
; ARM-STRICT-ALIGN: @unaligned_i32_store
; ARM-STRICT-ALIGN: strb
; ARM-STRICT-ALIGN: strb
; ARM-STRICT-ALIGN: strb
; ARM-STRICT-ALIGN: strb

; THUMB-STRICT-ALIGN: @unaligned_i32_store
; THUMB-STRICT-ALIGN: strb
; THUMB-STRICT-ALIGN: strb
; THUMB-STRICT-ALIGN: strb
; THUMB-STRICT-ALIGN: strb

  store i32 %x, i32* %y, align 1
  ret void
}

define i32 @unaligned_i32_load(i32* %x) nounwind {
entry:
; ARM-STRICT-ALIGN: @unaligned_i32_load
; ARM-STRICT-ALIGN: ldrb
; ARM-STRICT-ALIGN: ldrb
; ARM-STRICT-ALIGN: ldrb
; ARM-STRICT-ALIGN: ldrb

; THUMB-STRICT-ALIGN: @unaligned_i32_load
; THUMB-STRICT-ALIGN: ldrb
; THUMB-STRICT-ALIGN: ldrb
; THUMB-STRICT-ALIGN: ldrb
; THUMB-STRICT-ALIGN: ldrb

  %0 = load i32* %x, align 1
  ret i32 %0
}