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
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
|
// RUN: %clang_cc1 %s -emit-llvm -o - -triple=armv5-unknown-freebsd -std=c11 | FileCheck %s
// Test that we are generating atomicrmw instructions, rather than
// compare-exchange loops for common atomic ops. This makes a big difference
// on RISC platforms, where the compare-exchange loop becomes a ll/sc pair for
// the load and then another ll/sc in the loop, expanding to about 30
// instructions when it should be only 4. It has a smaller, but still
// noticeable, impact on platforms like x86 and RISC-V, where there are atomic
// RMW instructions.
//
// We currently emit cmpxchg loops for most operations on _Bools, because
// they're sufficiently rare that it's not worth making sure that the semantics
// are correct.
struct elem;
struct ptr {
struct elem *ptr;
};
// CHECK-DAG: %struct.ptr = type { ptr }
struct elem {
_Atomic(struct ptr) link;
};
struct ptr object;
// CHECK-DAG: @object ={{.*}} global %struct.ptr zeroinitializer
// CHECK-DAG: @testStructGlobal ={{.*}} global {{.*}} { i16 1, i16 2, i16 3, i16 4 }
// CHECK-DAG: @testPromotedStructGlobal ={{.*}} global {{.*}} { %{{.*}} { i16 1, i16 2, i16 3 }, [2 x i8] zeroinitializer }
typedef int __attribute__((vector_size(16))) vector;
_Atomic(_Bool) b;
_Atomic(int) i;
_Atomic(long long) l;
_Atomic(short) s;
_Atomic(char*) p;
_Atomic(float) f;
_Atomic(vector) v;
// CHECK: testinc
void testinc(void)
{
// Special case for suffix bool++, sets to true and returns the old value.
// CHECK: atomicrmw xchg ptr @b, i8 1 seq_cst, align 1
b++;
// CHECK: atomicrmw add ptr @i, i32 1 seq_cst, align 4
i++;
// CHECK: atomicrmw add ptr @l, i64 1 seq_cst, align 8
l++;
// CHECK: atomicrmw add ptr @s, i16 1 seq_cst, align 2
s++;
// Prefix increment
// Special case for bool: set to true and return true
// CHECK: store atomic i8 1, ptr @b seq_cst, align 1
++b;
// Currently, we have no variant of atomicrmw that returns the new value, so
// we have to generate an atomic add, which returns the old value, and then a
// non-atomic add.
// CHECK: atomicrmw add ptr @i, i32 1 seq_cst, align 4
// CHECK: add i32
++i;
// CHECK: atomicrmw add ptr @l, i64 1 seq_cst, align 8
// CHECK: add i64
++l;
// CHECK: atomicrmw add ptr @s, i16 1 seq_cst, align 2
// CHECK: add i16
++s;
}
// CHECK: testdec
void testdec(void)
{
// CHECK: call arm_aapcscc zeroext i1 @__atomic_compare_exchange(i32 noundef 1, ptr noundef @b
b--;
// CHECK: atomicrmw sub ptr @i, i32 1 seq_cst, align 4
i--;
// CHECK: atomicrmw sub ptr @l, i64 1 seq_cst, align 8
l--;
// CHECK: atomicrmw sub ptr @s, i16 1 seq_cst, align 2
s--;
// CHECK: call arm_aapcscc zeroext i1 @__atomic_compare_exchange(i32 noundef 1, ptr noundef @b
--b;
// CHECK: atomicrmw sub ptr @i, i32 1 seq_cst, align 4
// CHECK: sub i32
--i;
// CHECK: atomicrmw sub ptr @l, i64 1 seq_cst, align 8
// CHECK: sub i64
--l;
// CHECK: atomicrmw sub ptr @s, i16 1 seq_cst, align 2
// CHECK: sub i16
--s;
}
// CHECK: testaddeq
void testaddeq(void)
{
// CHECK: call arm_aapcscc zeroext i1 @__atomic_compare_exchange(i32 noundef 1, ptr noundef @b
// CHECK: atomicrmw add ptr @i, i32 42 seq_cst, align 4
// CHECK: atomicrmw add ptr @l, i64 42 seq_cst, align 8
// CHECK: atomicrmw add ptr @s, i16 42 seq_cst, align 2
b += 42;
i += 42;
l += 42;
s += 42;
}
// CHECK: testsubeq
void testsubeq(void)
{
// CHECK: call arm_aapcscc zeroext i1 @__atomic_compare_exchange(i32 noundef 1, ptr noundef @b
// CHECK: atomicrmw sub ptr @i, i32 42 seq_cst, align 4
// CHECK: atomicrmw sub ptr @l, i64 42 seq_cst, align 8
// CHECK: atomicrmw sub ptr @s, i16 42 seq_cst, align 2
b -= 42;
i -= 42;
l -= 42;
s -= 42;
}
// CHECK: testxoreq
void testxoreq(void)
{
// CHECK: call arm_aapcscc zeroext i1 @__atomic_compare_exchange(i32 noundef 1, ptr noundef @b
// CHECK: atomicrmw xor ptr @i, i32 42 seq_cst, align 4
// CHECK: atomicrmw xor ptr @l, i64 42 seq_cst, align 8
// CHECK: atomicrmw xor ptr @s, i16 42 seq_cst, align 2
b ^= 42;
i ^= 42;
l ^= 42;
s ^= 42;
}
// CHECK: testoreq
void testoreq(void)
{
// CHECK: call arm_aapcscc zeroext i1 @__atomic_compare_exchange(i32 noundef 1, ptr noundef @b
// CHECK: atomicrmw or ptr @i, i32 42 seq_cst, align 4
// CHECK: atomicrmw or ptr @l, i64 42 seq_cst, align 8
// CHECK: atomicrmw or ptr @s, i16 42 seq_cst, align 2
b |= 42;
i |= 42;
l |= 42;
s |= 42;
}
// CHECK: testandeq
void testandeq(void)
{
// CHECK: call arm_aapcscc zeroext i1 @__atomic_compare_exchange(i32 noundef 1, ptr noundef @b
// CHECK: atomicrmw and ptr @i, i32 42 seq_cst, align 4
// CHECK: atomicrmw and ptr @l, i64 42 seq_cst, align 8
// CHECK: atomicrmw and ptr @s, i16 42 seq_cst, align 2
b &= 42;
i &= 42;
l &= 42;
s &= 42;
}
// CHECK-LABEL: define{{.*}} arm_aapcscc void @testFloat(ptr
void testFloat(_Atomic(float) *fp) {
// CHECK: [[FP:%.*]] = alloca ptr
// CHECK-NEXT: [[X:%.*]] = alloca float
// CHECK-NEXT: [[F:%.*]] = alloca float
// CHECK-NEXT: [[TMP0:%.*]] = alloca float
// CHECK-NEXT: [[TMP1:%.*]] = alloca float
// CHECK-NEXT: store ptr {{%.*}}, ptr [[FP]]
// CHECK-NEXT: [[T0:%.*]] = load ptr, ptr [[FP]]
// CHECK-NEXT: store float 1.000000e+00, ptr [[T0]], align 4
__c11_atomic_init(fp, 1.0f);
// CHECK-NEXT: store float 2.000000e+00, ptr [[X]], align 4
_Atomic(float) x = 2.0f;
// CHECK-NEXT: [[T0:%.*]] = load ptr, ptr [[FP]]
// CHECK-NEXT: call arm_aapcscc void @__atomic_load(i32 noundef 4, ptr noundef [[T0]], ptr noundef [[TMP0]], i32 noundef 5)
// CHECK-NEXT: [[T3:%.*]] = load float, ptr [[TMP0]], align 4
// CHECK-NEXT: store float [[T3]], ptr [[F]]
float f = *fp;
// CHECK-NEXT: [[T0:%.*]] = load float, ptr [[F]], align 4
// CHECK-NEXT: [[T1:%.*]] = load ptr, ptr [[FP]], align 4
// CHECK-NEXT: store float [[T0]], ptr [[TMP1]], align 4
// CHECK-NEXT: call arm_aapcscc void @__atomic_store(i32 noundef 4, ptr noundef [[T1]], ptr noundef [[TMP1]], i32 noundef 5)
*fp = f;
// CHECK-NEXT: ret void
}
// CHECK: define{{.*}} arm_aapcscc void @testComplexFloat(ptr
void testComplexFloat(_Atomic(_Complex float) *fp) {
// CHECK: [[FP:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[X:%.*]] = alloca [[CF:{ float, float }]], align 8
// CHECK-NEXT: [[F:%.*]] = alloca [[CF]], align 4
// CHECK-NEXT: [[TMP0:%.*]] = alloca [[CF]], align 8
// CHECK-NEXT: [[TMP1:%.*]] = alloca [[CF]], align 8
// CHECK-NEXT: store ptr
// CHECK-NEXT: [[P:%.*]] = load ptr, ptr [[FP]]
// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds nuw [[CF]], ptr [[P]], i32 0, i32 0
// CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds nuw [[CF]], ptr [[P]], i32 0, i32 1
// CHECK-NEXT: store float 1.000000e+00, ptr [[T0]]
// CHECK-NEXT: store float 0.000000e+00, ptr [[T1]]
__c11_atomic_init(fp, 1.0f);
// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds nuw [[CF]], ptr [[X]], i32 0, i32 0
// CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds nuw [[CF]], ptr [[X]], i32 0, i32 1
// CHECK-NEXT: store float 2.000000e+00, ptr [[T0]]
// CHECK-NEXT: store float 0.000000e+00, ptr [[T1]]
_Atomic(_Complex float) x = 2.0f;
// CHECK-NEXT: [[T0:%.*]] = load ptr, ptr [[FP]]
// CHECK-NEXT: call arm_aapcscc void @__atomic_load(i32 noundef 8, ptr noundef [[T0]], ptr noundef [[TMP0]], i32 noundef 5)
// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds nuw [[CF]], ptr [[TMP0]], i32 0, i32 0
// CHECK-NEXT: [[R:%.*]] = load float, ptr [[T0]]
// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds nuw [[CF]], ptr [[TMP0]], i32 0, i32 1
// CHECK-NEXT: [[I:%.*]] = load float, ptr [[T0]]
// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds nuw [[CF]], ptr [[F]], i32 0, i32 0
// CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds nuw [[CF]], ptr [[F]], i32 0, i32 1
// CHECK-NEXT: store float [[R]], ptr [[T0]]
// CHECK-NEXT: store float [[I]], ptr [[T1]]
_Complex float f = *fp;
// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds nuw [[CF]], ptr [[F]], i32 0, i32 0
// CHECK-NEXT: [[R:%.*]] = load float, ptr [[T0]]
// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds nuw [[CF]], ptr [[F]], i32 0, i32 1
// CHECK-NEXT: [[I:%.*]] = load float, ptr [[T0]]
// CHECK-NEXT: [[DEST:%.*]] = load ptr, ptr [[FP]], align 4
// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds nuw [[CF]], ptr [[TMP1]], i32 0, i32 0
// CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds nuw [[CF]], ptr [[TMP1]], i32 0, i32 1
// CHECK-NEXT: store float [[R]], ptr [[T0]]
// CHECK-NEXT: store float [[I]], ptr [[T1]]
// CHECK-NEXT: call arm_aapcscc void @__atomic_store(i32 noundef 8, ptr noundef [[DEST]], ptr noundef [[TMP1]], i32 noundef 5)
*fp = f;
// CHECK-NEXT: ret void
}
typedef struct { short x, y, z, w; } S;
_Atomic S testStructGlobal = (S){1, 2, 3, 4};
// CHECK: define{{.*}} arm_aapcscc void @testStruct(ptr
void testStruct(_Atomic(S) *fp) {
// CHECK: [[FP:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[X:%.*]] = alloca [[S:.*]], align 8
// CHECK-NEXT: [[F:%.*]] = alloca [[S:%.*]], align 2
// CHECK-NEXT: [[TMP0:%.*]] = alloca [[S]], align 8
// CHECK-NEXT: store ptr
// CHECK-NEXT: [[P:%.*]] = load ptr, ptr [[FP]]
// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds nuw [[S]], ptr [[P]], i32 0, i32 0
// CHECK-NEXT: store i16 1, ptr [[T0]], align 8
// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds nuw [[S]], ptr [[P]], i32 0, i32 1
// CHECK-NEXT: store i16 2, ptr [[T0]], align 2
// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds nuw [[S]], ptr [[P]], i32 0, i32 2
// CHECK-NEXT: store i16 3, ptr [[T0]], align 4
// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds nuw [[S]], ptr [[P]], i32 0, i32 3
// CHECK-NEXT: store i16 4, ptr [[T0]], align 2
__c11_atomic_init(fp, (S){1,2,3,4});
// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds nuw [[S]], ptr [[X]], i32 0, i32 0
// CHECK-NEXT: store i16 1, ptr [[T0]], align 8
// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds nuw [[S]], ptr [[X]], i32 0, i32 1
// CHECK-NEXT: store i16 2, ptr [[T0]], align 2
// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds nuw [[S]], ptr [[X]], i32 0, i32 2
// CHECK-NEXT: store i16 3, ptr [[T0]], align 4
// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds nuw [[S]], ptr [[X]], i32 0, i32 3
// CHECK-NEXT: store i16 4, ptr [[T0]], align 2
_Atomic(S) x = (S){1,2,3,4};
// CHECK-NEXT: [[T0:%.*]] = load ptr, ptr [[FP]]
// CHECK-NEXT: call arm_aapcscc void @__atomic_load(i32 noundef 8, ptr noundef [[T0]], ptr noundef [[F]], i32 noundef 5)
S f = *fp;
// CHECK-NEXT: [[T0:%.*]] = load ptr, ptr [[FP]]
// CHECK-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 8 [[TMP0]], ptr align 2 [[F]], i32 8, i1 false)
// CHECK-NEXT: call arm_aapcscc void @__atomic_store(i32 noundef 8, ptr noundef [[T0]], ptr noundef [[TMP0]], i32 noundef 5)
*fp = f;
// CHECK-NEXT: ret void
}
typedef struct { short x, y, z; } PS;
_Atomic PS testPromotedStructGlobal = (PS){1, 2, 3};
// CHECK: define{{.*}} arm_aapcscc void @testPromotedStruct(ptr
void testPromotedStruct(_Atomic(PS) *fp) {
// CHECK: [[FP:%.*]] = alloca ptr, align 4
// CHECK-NEXT: [[X:%.*]] = alloca [[APS:.*]], align 8
// CHECK-NEXT: [[F:%.*]] = alloca [[PS:%.*]], align 2
// CHECK-NEXT: [[TMP0:%.*]] = alloca [[APS]], align 8
// CHECK-NEXT: [[TMP1:%.*]] = alloca [[APS]], align 8
// CHECK-NEXT: [[A:%.*]] = alloca i32, align 4
// CHECK-NEXT: [[TMP2:%.*]] = alloca %struct.PS, align 2
// CHECK-NEXT: [[TMP3:%.*]] = alloca [[APS]], align 8
// CHECK-NEXT: store ptr
// CHECK-NEXT: [[P:%.*]] = load ptr, ptr [[FP]]
// CHECK-NEXT: call void @llvm.memset.p0.i64(ptr align 8 [[P]], i8 0, i64 8, i1 false)
// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds nuw [[APS]], ptr [[P]], i32 0, i32 0
// CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds nuw [[PS]], ptr [[T0]], i32 0, i32 0
// CHECK-NEXT: store i16 1, ptr [[T1]], align 8
// CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds nuw [[PS]], ptr [[T0]], i32 0, i32 1
// CHECK-NEXT: store i16 2, ptr [[T1]], align 2
// CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds nuw [[PS]], ptr [[T0]], i32 0, i32 2
// CHECK-NEXT: store i16 3, ptr [[T1]], align 4
__c11_atomic_init(fp, (PS){1,2,3});
// CHECK-NEXT: call void @llvm.memset.p0.i32(ptr align 8 [[X]], i8 0, i32 8, i1 false)
// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds nuw [[APS]], ptr [[X]], i32 0, i32 0
// CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds nuw [[PS]], ptr [[T0]], i32 0, i32 0
// CHECK-NEXT: store i16 1, ptr [[T1]], align 8
// CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds nuw [[PS]], ptr [[T0]], i32 0, i32 1
// CHECK-NEXT: store i16 2, ptr [[T1]], align 2
// CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds nuw [[PS]], ptr [[T0]], i32 0, i32 2
// CHECK-NEXT: store i16 3, ptr [[T1]], align 4
_Atomic(PS) x = (PS){1,2,3};
// CHECK-NEXT: [[T0:%.*]] = load ptr, ptr [[FP]]
// CHECK-NEXT: call arm_aapcscc void @__atomic_load(i32 noundef 8, ptr noundef [[T0]], ptr noundef [[TMP0]], i32 noundef 5)
// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds nuw [[APS]], ptr [[TMP0]], i32 0, i32 0
// CHECK-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 2 [[F]], ptr align 8 [[T0]], i32 6, i1 false)
PS f = *fp;
// CHECK-NEXT: [[T0:%.*]] = load ptr, ptr [[FP]]
// CHECK-NEXT: call void @llvm.memset.p0.i32(ptr align 8 [[TMP1]], i8 0, i32 8, i1 false)
// CHECK-NEXT: [[T1:%.*]] = getelementptr inbounds nuw [[APS]], ptr [[TMP1]], i32 0, i32 0
// CHECK-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 8 [[T1]], ptr align 2 [[F]], i32 6, i1 false)
// CHECK-NEXT: call arm_aapcscc void @__atomic_store(i32 noundef 8, ptr noundef [[T0]], ptr noundef [[TMP1]], i32 noundef 5)
*fp = f;
// CHECK-NEXT: [[T0:%.*]] = load ptr, ptr [[FP]], align 4
// CHECK-NEXT: call arm_aapcscc void @__atomic_load(i32 noundef 8, ptr noundef [[T0]], ptr noundef [[TMP3]], i32 noundef 5)
// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds nuw [[APS]], ptr [[TMP3]], i32 0, i32 0
// CHECK-NEXT: call void @llvm.memcpy.p0.p0.i32(ptr align 2 [[TMP2]], ptr align 8 [[T0]], i32 6, i1 false)
// CHECK-NEXT: [[T0:%.*]] = getelementptr inbounds nuw %struct.PS, ptr [[TMP2]], i32 0, i32 0
// CHECK-NEXT: [[T1:%.*]] = load i16, ptr [[T0]], align 2
// CHECK-NEXT: [[T2:%.*]] = sext i16 [[T1]] to i32
// CHECK-NEXT: store i32 [[T2]], ptr [[A]], align 4
int a = ((PS)*fp).x;
// CHECK-NEXT: ret void
}
PS test_promoted_load(_Atomic(PS) *addr) {
// CHECK-LABEL: @test_promoted_load(ptr dead_on_unwind noalias writable sret(%struct.PS) align 2 %agg.result, ptr noundef %addr)
// CHECK: [[ADDR_ARG:%.*]] = alloca ptr, align 4
// CHECK: [[ATOMIC_RES:%.*]] = alloca { %struct.PS, [2 x i8] }, align 8
// CHECK: store ptr %addr, ptr [[ADDR_ARG]], align 4
// CHECK: [[ADDR:%.*]] = load ptr, ptr [[ADDR_ARG]], align 4
// CHECK: [[ATOMIC_RES:%.*]] = load atomic i64, ptr [[ADDR]] seq_cst, align 8
// CHECK: store i64 [[ATOMIC_RES]], ptr [[ATOMIC_RES_ADDR:%.*]], align 8
// CHECK: call void @llvm.memcpy.p0.p0.i32(ptr align 2 %agg.result, ptr align 8 [[ATOMIC_RES_ADDR]], i32 6, i1 false)
return __c11_atomic_load(addr, 5);
}
void test_promoted_store(_Atomic(PS) *addr, PS *val) {
// CHECK-LABEL: @test_promoted_store(ptr noundef %addr, ptr noundef %val)
// CHECK: [[ADDR_ARG:%.*]] = alloca ptr, align 4
// CHECK: [[VAL_ARG:%.*]] = alloca ptr, align 4
// CHECK: [[NONATOMIC_TMP:%.*]] = alloca %struct.PS, align 2
// CHECK: [[ATOMIC_VAL:%.*]] = alloca { %struct.PS, [2 x i8] }, align 8
// CHECK: store ptr %addr, ptr [[ADDR_ARG]], align 4
// CHECK: store ptr %val, ptr [[VAL_ARG]], align 4
// CHECK: [[ADDR:%.*]] = load ptr, ptr [[ADDR_ARG]], align 4
// CHECK: [[VAL:%.*]] = load ptr, ptr [[VAL_ARG]], align 4
// CHECK: call void @llvm.memcpy.p0.p0.i32(ptr align 2 [[NONATOMIC_TMP]], ptr align 2 [[VAL]], i32 6, i1 false)
// CHECK: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[ATOMIC_VAL]], ptr align 2 [[NONATOMIC_TMP]], i64 6, i1 false)
// CHECK: [[ATOMIC:%.*]] = load i64, ptr [[ATOMIC_VAL]], align 8
// CHECK: store atomic i64 [[ATOMIC]], ptr [[ADDR]] seq_cst, align 8
__c11_atomic_store(addr, *val, 5);
}
PS test_promoted_exchange(_Atomic(PS) *addr, PS *val) {
// CHECK-LABEL: @test_promoted_exchange(ptr dead_on_unwind noalias writable sret(%struct.PS) align 2 %agg.result, ptr noundef %addr, ptr noundef %val)
// CHECK: [[ADDR_ARG:%.*]] = alloca ptr, align 4
// CHECK: [[VAL_ARG:%.*]] = alloca ptr, align 4
// CHECK: [[NONATOMIC_TMP:%.*]] = alloca %struct.PS, align 2
// CHECK: [[ATOMIC_VAL:%.*]] = alloca { %struct.PS, [2 x i8] }, align 8
// CHECK: [[ATOMIC_RES:%.*]] = alloca { %struct.PS, [2 x i8] }, align 8
// CHECK: store ptr %addr, ptr [[ADDR_ARG]], align 4
// CHECK: store ptr %val, ptr [[VAL_ARG]], align 4
// CHECK: [[ADDR:%.*]] = load ptr, ptr [[ADDR_ARG]], align 4
// CHECK: [[VAL:%.*]] = load ptr, ptr [[VAL_ARG]], align 4
// CHECK: call void @llvm.memcpy.p0.p0.i32(ptr align 2 [[NONATOMIC_TMP]], ptr align 2 [[VAL]], i32 6, i1 false)
// CHECK: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[ATOMIC_VAL]], ptr align 2 [[NONATOMIC_TMP]], i64 6, i1 false)
// CHECK: [[ATOMIC:%.*]] = load i64, ptr [[ATOMIC_VAL]], align 8
// CHECK: [[ATOMIC_RES:%.*]] = atomicrmw xchg ptr [[ADDR]], i64 [[ATOMIC]] seq_cst, align 8
// CHECK: store i64 [[ATOMIC_RES]], ptr [[ATOMIC_RES_PTR:%.*]], align 8
// CHECK: call void @llvm.memcpy.p0.p0.i32(ptr align 2 %agg.result, ptr align 8 [[ATOMIC_RES_PTR]], i32 6, i1 false)
return __c11_atomic_exchange(addr, *val, 5);
}
_Bool test_promoted_cmpxchg(_Atomic(PS) *addr, PS *desired, PS *new) {
// CHECK-LABEL: i1 @test_promoted_cmpxchg(ptr noundef %addr, ptr noundef %desired, ptr noundef %new) #0 {
// CHECK: [[ADDR_ARG:%.*]] = alloca ptr, align 4
// CHECK: [[DESIRED_ARG:%.*]] = alloca ptr, align 4
// CHECK: [[NEW_ARG:%.*]] = alloca ptr, align 4
// CHECK: [[NONATOMIC_TMP:%.*]] = alloca %struct.PS, align 2
// CHECK: [[ATOMIC_DESIRED:%.*]] = alloca { %struct.PS, [2 x i8] }, align 8
// CHECK: [[ATOMIC_NEW:%.*]] = alloca { %struct.PS, [2 x i8] }, align 8
// CHECK: store ptr %addr, ptr [[ADDR_ARG]], align 4
// CHECK: store ptr %desired, ptr [[DESIRED_ARG]], align 4
// CHECK: store ptr %new, ptr [[NEW_ARG]], align 4
// CHECK: [[ADDR:%.*]] = load ptr, ptr [[ADDR_ARG]], align 4
// CHECK: [[DESIRED:%.*]] = load ptr, ptr [[DESIRED_ARG]], align 4
// CHECK: [[NEW:%.*]] = load ptr, ptr [[NEW_ARG]], align 4
// CHECK: call void @llvm.memcpy.p0.p0.i32(ptr align 2 [[NONATOMIC_TMP]], ptr align 2 [[NEW]], i32 6, i1 false)
// CHECK: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[ATOMIC_DESIRED]], ptr align 2 [[DESIRED]], i64 6, i1 false)
// CHECK: call void @llvm.memcpy.p0.p0.i64(ptr align 8 [[ATOMIC_NEW]], ptr align 2 [[NONATOMIC_TMP]], i64 6, i1 false)
// CHECK: [[VAL1:%.*]] = load i64, ptr [[ATOMIC_DESIRED]], align 8
// CHECK: [[VAL2:%.*]] = load i64, ptr [[ATOMIC_NEW]], align 8
// CHECK: [[RES_PAIR:%.*]] = cmpxchg ptr [[ADDR]], i64 [[VAL1]], i64 [[VAL2]] seq_cst seq_cst, align 8
// CHECK: [[RES:%.*]] = extractvalue { i64, i1 } [[RES_PAIR]], 1
return __c11_atomic_compare_exchange_strong(addr, desired, *new, 5, 5);
}
struct Empty {};
struct Empty test_empty_struct_load(_Atomic(struct Empty)* empty) {
// CHECK-LABEL: @test_empty_struct_load(
// CHECK: load atomic i8, ptr {{.*}}, align 1
return __c11_atomic_load(empty, 5);
}
void test_empty_struct_store(_Atomic(struct Empty)* empty, struct Empty value) {
// CHECK-LABEL: @test_empty_struct_store(
// CHECK: store atomic i8 {{.*}}, ptr {{.*}}, align 1
__c11_atomic_store(empty, value, 5);
}
|
