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
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
|
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -verify-machineinstrs < %s -mtriple=powerpc-unknown-linux-gnu -verify-machineinstrs -ppc-asm-full-reg-names | FileCheck %s --check-prefix=CHECK --check-prefix=PPC32
; This is already checked for in Atomics-64.ll
; RUN: llc -verify-machineinstrs < %s -mcpu=ppc -mtriple=powerpc64-unknown-linux-gnu -ppc-asm-full-reg-names | FileCheck %s --check-prefix=CHECK --check-prefix=PPC64
; FIXME: we don't currently check for the operations themselves with CHECK-NEXT,
; because they are implemented in a very messy way with lwarx/stwcx.
; It should be fixed soon in another patch.
; We first check loads, for all sizes from i8 to i64.
; We also vary orderings to check for barriers.
define i8 @load_i8_unordered(ptr %mem) {
; CHECK-LABEL: load_i8_unordered:
; CHECK: # %bb.0:
; CHECK-NEXT: lbz r3, 0(r3)
; CHECK-NEXT: blr
%val = load atomic i8, ptr %mem unordered, align 1
ret i8 %val
}
define i16 @load_i16_monotonic(ptr %mem) {
; CHECK-LABEL: load_i16_monotonic:
; CHECK: # %bb.0:
; CHECK-NEXT: lhz r3, 0(r3)
; CHECK-NEXT: blr
%val = load atomic i16, ptr %mem monotonic, align 2
ret i16 %val
}
define i32 @load_i32_acquire(ptr %mem) {
; PPC32-LABEL: load_i32_acquire:
; PPC32: # %bb.0:
; PPC32-NEXT: lwz r3, 0(r3)
; PPC32-NEXT: cmpw cr7, r3, r3
; PPC32-NEXT: bne- cr7, .+4
; PPC32-NEXT: isync
; PPC32-NEXT: blr
;
; PPC64-LABEL: load_i32_acquire:
; PPC64: # %bb.0:
; PPC64-NEXT: lwz r3, 0(r3)
; PPC64-NEXT: cmpd cr7, r3, r3
; PPC64-NEXT: bne- cr7, .+4
; PPC64-NEXT: isync
; PPC64-NEXT: blr
%val = load atomic i32, ptr %mem acquire, align 4
; CHECK-PPC32: lwsync
; CHECK-PPC64: cmpw [[CR:cr[0-9]+]], [[VAL]], [[VAL]]
; CHECK-PPC64: bne- [[CR]], .+4
; CHECK-PPC64: isync
ret i32 %val
}
define i64 @load_i64_seq_cst(ptr %mem) {
; PPC32-LABEL: load_i64_seq_cst:
; PPC32: # %bb.0:
; PPC32-NEXT: mflr r0
; PPC32-NEXT: stwu r1, -16(r1)
; PPC32-NEXT: stw r0, 20(r1)
; PPC32-NEXT: .cfi_def_cfa_offset 16
; PPC32-NEXT: .cfi_offset lr, 4
; PPC32-NEXT: li r4, 5
; PPC32-NEXT: bl __atomic_load_8
; PPC32-NEXT: lwz r0, 20(r1)
; PPC32-NEXT: addi r1, r1, 16
; PPC32-NEXT: mtlr r0
; PPC32-NEXT: blr
;
; PPC64-LABEL: load_i64_seq_cst:
; PPC64: # %bb.0:
; PPC64-NEXT: sync
; PPC64-NEXT: ld r3, 0(r3)
; PPC64-NEXT: cmpd cr7, r3, r3
; PPC64-NEXT: bne- cr7, .+4
; PPC64-NEXT: isync
; PPC64-NEXT: blr
%val = load atomic i64, ptr %mem seq_cst, align 8
; CHECK-PPC32: lwsync
; CHECK-PPC64: cmpw [[CR:cr[0-9]+]], [[VAL]], [[VAL]]
; CHECK-PPC64: bne- [[CR]], .+4
; CHECK-PPC64: isync
ret i64 %val
}
; Stores
define void @store_i8_unordered(ptr %mem) {
; CHECK-LABEL: store_i8_unordered:
; CHECK: # %bb.0:
; CHECK-NEXT: li r4, 42
; CHECK-NEXT: stb r4, 0(r3)
; CHECK-NEXT: blr
store atomic i8 42, ptr %mem unordered, align 1
ret void
}
define void @store_i16_monotonic(ptr %mem) {
; CHECK-LABEL: store_i16_monotonic:
; CHECK: # %bb.0:
; CHECK-NEXT: li r4, 42
; CHECK-NEXT: sth r4, 0(r3)
; CHECK-NEXT: blr
store atomic i16 42, ptr %mem monotonic, align 2
ret void
}
define void @store_i32_release(ptr %mem) {
; CHECK-LABEL: store_i32_release:
; CHECK: # %bb.0:
; CHECK-NEXT: li r4, 42
; CHECK-NEXT: lwsync
; CHECK-NEXT: stw r4, 0(r3)
; CHECK-NEXT: blr
store atomic i32 42, ptr %mem release, align 4
ret void
}
define void @store_i64_seq_cst(ptr %mem) {
; PPC32-LABEL: store_i64_seq_cst:
; PPC32: # %bb.0:
; PPC32-NEXT: mflr r0
; PPC32-NEXT: stwu r1, -16(r1)
; PPC32-NEXT: stw r0, 20(r1)
; PPC32-NEXT: .cfi_def_cfa_offset 16
; PPC32-NEXT: .cfi_offset lr, 4
; PPC32-NEXT: li r5, 0
; PPC32-NEXT: li r6, 42
; PPC32-NEXT: li r7, 5
; PPC32-NEXT: bl __atomic_store_8
; PPC32-NEXT: lwz r0, 20(r1)
; PPC32-NEXT: addi r1, r1, 16
; PPC32-NEXT: mtlr r0
; PPC32-NEXT: blr
;
; PPC64-LABEL: store_i64_seq_cst:
; PPC64: # %bb.0:
; PPC64-NEXT: li r4, 42
; PPC64-NEXT: sync
; PPC64-NEXT: std r4, 0(r3)
; PPC64-NEXT: blr
store atomic i64 42, ptr %mem seq_cst, align 8
ret void
}
; Atomic CmpXchg
define i8 @cas_strong_i8_sc_sc(ptr %mem) {
; PPC32-LABEL: cas_strong_i8_sc_sc:
; PPC32: # %bb.0:
; PPC32-NEXT: rlwinm r5, r3, 0, 0, 29
; PPC32-NEXT: lwarx r4, 0, r5
; PPC32-NEXT: not r3, r3
; PPC32-NEXT: rlwinm r3, r3, 3, 27, 28
; PPC32-NEXT: srw r6, r4, r3
; PPC32-NEXT: andi. r6, r6, 255
; PPC32-NEXT: bne cr0, .LBB8_4
; PPC32-NEXT: # %bb.1: # %cmpxchg.fencedstore
; PPC32-NEXT: li r6, 255
; PPC32-NEXT: li r7, 1
; PPC32-NEXT: slw r6, r6, r3
; PPC32-NEXT: not r6, r6
; PPC32-NEXT: slw r7, r7, r3
; PPC32-NEXT: sync
; PPC32-NEXT: .LBB8_2: # %cmpxchg.trystore
; PPC32-NEXT: # =>This Inner Loop Header: Depth=1
; PPC32-NEXT: and r8, r4, r6
; PPC32-NEXT: or r8, r8, r7
; PPC32-NEXT: stwcx. r8, 0, r5
; PPC32-NEXT: beq cr0, .LBB8_4
; PPC32-NEXT: # %bb.3: # %cmpxchg.releasedload
; PPC32-NEXT: # in Loop: Header=BB8_2 Depth=1
; PPC32-NEXT: lwarx r4, 0, r5
; PPC32-NEXT: srw r8, r4, r3
; PPC32-NEXT: andi. r8, r8, 255
; PPC32-NEXT: beq cr0, .LBB8_2
; PPC32-NEXT: .LBB8_4: # %cmpxchg.nostore
; PPC32-NEXT: srw r3, r4, r3
; PPC32-NEXT: lwsync
; PPC32-NEXT: blr
;
; PPC64-LABEL: cas_strong_i8_sc_sc:
; PPC64: # %bb.0:
; PPC64-NEXT: rldicr r5, r3, 0, 61
; PPC64-NEXT: not r3, r3
; PPC64-NEXT: lwarx r4, 0, r5
; PPC64-NEXT: rlwinm r3, r3, 3, 27, 28
; PPC64-NEXT: srw r6, r4, r3
; PPC64-NEXT: andi. r6, r6, 255
; PPC64-NEXT: bne cr0, .LBB8_4
; PPC64-NEXT: # %bb.1: # %cmpxchg.fencedstore
; PPC64-NEXT: li r6, 255
; PPC64-NEXT: li r7, 1
; PPC64-NEXT: slw r6, r6, r3
; PPC64-NEXT: not r6, r6
; PPC64-NEXT: slw r7, r7, r3
; PPC64-NEXT: sync
; PPC64-NEXT: .LBB8_2: # %cmpxchg.trystore
; PPC64-NEXT: # =>This Inner Loop Header: Depth=1
; PPC64-NEXT: and r8, r4, r6
; PPC64-NEXT: or r8, r8, r7
; PPC64-NEXT: stwcx. r8, 0, r5
; PPC64-NEXT: beq cr0, .LBB8_4
; PPC64-NEXT: # %bb.3: # %cmpxchg.releasedload
; PPC64-NEXT: # in Loop: Header=BB8_2 Depth=1
; PPC64-NEXT: lwarx r4, 0, r5
; PPC64-NEXT: srw r8, r4, r3
; PPC64-NEXT: andi. r8, r8, 255
; PPC64-NEXT: beq cr0, .LBB8_2
; PPC64-NEXT: .LBB8_4: # %cmpxchg.nostore
; PPC64-NEXT: srw r3, r4, r3
; PPC64-NEXT: lwsync
; PPC64-NEXT: blr
%val = cmpxchg ptr %mem, i8 0, i8 1 seq_cst seq_cst
%loaded = extractvalue { i8, i1} %val, 0
ret i8 %loaded
}
define i16 @cas_weak_i16_acquire_acquire(ptr %mem) {
; PPC32-LABEL: cas_weak_i16_acquire_acquire:
; PPC32: # %bb.0:
; PPC32-NEXT: rlwinm r4, r3, 0, 0, 29
; PPC32-NEXT: lwarx r5, 0, r4
; PPC32-NEXT: clrlwi r3, r3, 30
; PPC32-NEXT: xori r3, r3, 2
; PPC32-NEXT: slwi r6, r3, 3
; PPC32-NEXT: srw r3, r5, r6
; PPC32-NEXT: andi. r7, r3, 65535
; PPC32-NEXT: beq cr0, .LBB9_2
; PPC32-NEXT: # %bb.1: # %cmpxchg.failure
; PPC32-NEXT: lwsync
; PPC32-NEXT: blr
; PPC32-NEXT: .LBB9_2: # %cmpxchg.fencedstore
; PPC32-NEXT: lis r7, 0
; PPC32-NEXT: ori r7, r7, 65535
; PPC32-NEXT: slw r7, r7, r6
; PPC32-NEXT: li r8, 1
; PPC32-NEXT: not r7, r7
; PPC32-NEXT: slw r6, r8, r6
; PPC32-NEXT: and r5, r5, r7
; PPC32-NEXT: or r5, r5, r6
; PPC32-NEXT: stwcx. r5, 0, r4
; PPC32-NEXT: lwsync
; PPC32-NEXT: blr
;
; PPC64-LABEL: cas_weak_i16_acquire_acquire:
; PPC64: # %bb.0:
; PPC64-NEXT: rldicr r4, r3, 0, 61
; PPC64-NEXT: clrlwi r3, r3, 30
; PPC64-NEXT: lwarx r5, 0, r4
; PPC64-NEXT: xori r3, r3, 2
; PPC64-NEXT: slwi r6, r3, 3
; PPC64-NEXT: srw r3, r5, r6
; PPC64-NEXT: andi. r7, r3, 65535
; PPC64-NEXT: beq cr0, .LBB9_2
; PPC64-NEXT: # %bb.1: # %cmpxchg.failure
; PPC64-NEXT: lwsync
; PPC64-NEXT: blr
; PPC64-NEXT: .LBB9_2: # %cmpxchg.fencedstore
; PPC64-NEXT: lis r7, 0
; PPC64-NEXT: ori r7, r7, 65535
; PPC64-NEXT: slw r7, r7, r6
; PPC64-NEXT: li r8, 1
; PPC64-NEXT: not r7, r7
; PPC64-NEXT: slw r6, r8, r6
; PPC64-NEXT: and r5, r5, r7
; PPC64-NEXT: or r5, r5, r6
; PPC64-NEXT: stwcx. r5, 0, r4
; PPC64-NEXT: lwsync
; PPC64-NEXT: blr
%val = cmpxchg weak ptr %mem, i16 0, i16 1 acquire acquire
%loaded = extractvalue { i16, i1} %val, 0
ret i16 %loaded
}
define i32 @cas_strong_i32_acqrel_acquire(ptr %mem) {
; CHECK-LABEL: cas_strong_i32_acqrel_acquire:
; CHECK: # %bb.0:
; CHECK-NEXT: mr r4, r3
; CHECK-NEXT: lwarx r3, 0, r3
; CHECK-NEXT: cmplwi r3, 0
; CHECK-NEXT: bne cr0, .LBB10_4
; CHECK-NEXT: # %bb.1: # %cmpxchg.fencedstore
; CHECK-NEXT: li r5, 1
; CHECK-NEXT: lwsync
; CHECK-NEXT: .LBB10_2: # %cmpxchg.trystore
; CHECK-NEXT: # =>This Inner Loop Header: Depth=1
; CHECK-NEXT: stwcx. r5, 0, r4
; CHECK-NEXT: beq cr0, .LBB10_4
; CHECK-NEXT: # %bb.3: # %cmpxchg.releasedload
; CHECK-NEXT: # in Loop: Header=BB10_2 Depth=1
; CHECK-NEXT: lwarx r3, 0, r4
; CHECK-NEXT: cmplwi r3, 0
; CHECK-NEXT: beq cr0, .LBB10_2
; CHECK-NEXT: .LBB10_4: # %cmpxchg.nostore
; CHECK-NEXT: lwsync
; CHECK-NEXT: blr
%val = cmpxchg ptr %mem, i32 0, i32 1 acq_rel acquire
%loaded = extractvalue { i32, i1} %val, 0
ret i32 %loaded
}
define i64 @cas_weak_i64_release_monotonic(ptr %mem) {
; PPC32-LABEL: cas_weak_i64_release_monotonic:
; PPC32: # %bb.0:
; PPC32-NEXT: mflr r0
; PPC32-NEXT: stwu r1, -16(r1)
; PPC32-NEXT: stw r0, 20(r1)
; PPC32-NEXT: .cfi_def_cfa_offset 16
; PPC32-NEXT: .cfi_offset lr, 4
; PPC32-NEXT: li r4, 0
; PPC32-NEXT: stw r4, 12(r1)
; PPC32-NEXT: li r5, 0
; PPC32-NEXT: stw r4, 8(r1)
; PPC32-NEXT: addi r4, r1, 8
; PPC32-NEXT: li r6, 1
; PPC32-NEXT: li r7, 3
; PPC32-NEXT: li r8, 0
; PPC32-NEXT: bl __atomic_compare_exchange_8
; PPC32-NEXT: lwz r4, 12(r1)
; PPC32-NEXT: lwz r3, 8(r1)
; PPC32-NEXT: lwz r0, 20(r1)
; PPC32-NEXT: addi r1, r1, 16
; PPC32-NEXT: mtlr r0
; PPC32-NEXT: blr
;
; PPC64-LABEL: cas_weak_i64_release_monotonic:
; PPC64: # %bb.0:
; PPC64-NEXT: mr r4, r3
; PPC64-NEXT: ldarx r3, 0, r3
; PPC64-NEXT: cmpldi r3, 0
; PPC64-NEXT: bnelr cr0
; PPC64-NEXT: # %bb.1: # %cmpxchg.fencedstore
; PPC64-NEXT: li r5, 1
; PPC64-NEXT: lwsync
; PPC64-NEXT: stdcx. r5, 0, r4
; PPC64-NEXT: blr
%val = cmpxchg weak ptr %mem, i64 0, i64 1 release monotonic
%loaded = extractvalue { i64, i1} %val, 0
ret i64 %loaded
}
; AtomicRMW
define i8 @add_i8_monotonic(ptr %mem, i8 %operand) {
; PPC32-LABEL: add_i8_monotonic:
; PPC32: # %bb.0:
; PPC32-NEXT: rlwinm r7, r3, 3, 27, 28
; PPC32-NEXT: li r6, 255
; PPC32-NEXT: rlwinm r5, r3, 0, 0, 29
; PPC32-NEXT: xori r3, r7, 24
; PPC32-NEXT: slw r4, r4, r3
; PPC32-NEXT: slw r6, r6, r3
; PPC32-NEXT: .LBB12_1:
; PPC32-NEXT: lwarx r7, 0, r5
; PPC32-NEXT: add r8, r4, r7
; PPC32-NEXT: andc r9, r7, r6
; PPC32-NEXT: and r8, r8, r6
; PPC32-NEXT: or r8, r8, r9
; PPC32-NEXT: stwcx. r8, 0, r5
; PPC32-NEXT: bne cr0, .LBB12_1
; PPC32-NEXT: # %bb.2:
; PPC32-NEXT: srw r3, r7, r3
; PPC32-NEXT: clrlwi r3, r3, 24
; PPC32-NEXT: blr
;
; PPC64-LABEL: add_i8_monotonic:
; PPC64: # %bb.0:
; PPC64-NEXT: rlwinm r7, r3, 3, 27, 28
; PPC64-NEXT: li r6, 255
; PPC64-NEXT: rldicr r5, r3, 0, 61
; PPC64-NEXT: xori r3, r7, 24
; PPC64-NEXT: slw r4, r4, r3
; PPC64-NEXT: slw r6, r6, r3
; PPC64-NEXT: .LBB12_1:
; PPC64-NEXT: lwarx r7, 0, r5
; PPC64-NEXT: add r8, r4, r7
; PPC64-NEXT: andc r9, r7, r6
; PPC64-NEXT: and r8, r8, r6
; PPC64-NEXT: or r8, r8, r9
; PPC64-NEXT: stwcx. r8, 0, r5
; PPC64-NEXT: bne cr0, .LBB12_1
; PPC64-NEXT: # %bb.2:
; PPC64-NEXT: srw r3, r7, r3
; PPC64-NEXT: clrlwi r3, r3, 24
; PPC64-NEXT: blr
%val = atomicrmw add ptr %mem, i8 %operand monotonic
ret i8 %val
}
define i16 @xor_i16_seq_cst(ptr %mem, i16 %operand) {
; PPC32-LABEL: xor_i16_seq_cst:
; PPC32: # %bb.0:
; PPC32-NEXT: li r5, 0
; PPC32-NEXT: rlwinm r6, r3, 3, 27, 27
; PPC32-NEXT: ori r7, r5, 65535
; PPC32-NEXT: xori r5, r6, 16
; PPC32-NEXT: rlwinm r3, r3, 0, 0, 29
; PPC32-NEXT: slw r4, r4, r5
; PPC32-NEXT: slw r6, r7, r5
; PPC32-NEXT: sync
; PPC32-NEXT: .LBB13_1:
; PPC32-NEXT: lwarx r7, 0, r3
; PPC32-NEXT: xor r8, r4, r7
; PPC32-NEXT: andc r9, r7, r6
; PPC32-NEXT: and r8, r8, r6
; PPC32-NEXT: or r8, r8, r9
; PPC32-NEXT: stwcx. r8, 0, r3
; PPC32-NEXT: bne cr0, .LBB13_1
; PPC32-NEXT: # %bb.2:
; PPC32-NEXT: srw r3, r7, r5
; PPC32-NEXT: clrlwi r3, r3, 16
; PPC32-NEXT: lwsync
; PPC32-NEXT: blr
;
; PPC64-LABEL: xor_i16_seq_cst:
; PPC64: # %bb.0:
; PPC64-NEXT: li r5, 0
; PPC64-NEXT: rlwinm r6, r3, 3, 27, 27
; PPC64-NEXT: ori r7, r5, 65535
; PPC64-NEXT: xori r5, r6, 16
; PPC64-NEXT: rldicr r3, r3, 0, 61
; PPC64-NEXT: slw r4, r4, r5
; PPC64-NEXT: slw r6, r7, r5
; PPC64-NEXT: sync
; PPC64-NEXT: .LBB13_1:
; PPC64-NEXT: lwarx r7, 0, r3
; PPC64-NEXT: xor r8, r4, r7
; PPC64-NEXT: andc r9, r7, r6
; PPC64-NEXT: and r8, r8, r6
; PPC64-NEXT: or r8, r8, r9
; PPC64-NEXT: stwcx. r8, 0, r3
; PPC64-NEXT: bne cr0, .LBB13_1
; PPC64-NEXT: # %bb.2:
; PPC64-NEXT: srw r3, r7, r5
; PPC64-NEXT: clrlwi r3, r3, 16
; PPC64-NEXT: lwsync
; PPC64-NEXT: blr
%val = atomicrmw xor ptr %mem, i16 %operand seq_cst
ret i16 %val
}
define i32 @xchg_i32_acq_rel(ptr %mem, i32 %operand) {
; CHECK-LABEL: xchg_i32_acq_rel:
; CHECK: # %bb.0:
; CHECK-NEXT: lwsync
; CHECK-NEXT: .LBB14_1:
; CHECK-NEXT: lwarx r5, 0, r3
; CHECK-NEXT: stwcx. r4, 0, r3
; CHECK-NEXT: bne cr0, .LBB14_1
; CHECK-NEXT: # %bb.2:
; CHECK-NEXT: mr r3, r5
; CHECK-NEXT: lwsync
; CHECK-NEXT: blr
%val = atomicrmw xchg ptr %mem, i32 %operand acq_rel
ret i32 %val
}
define i64 @and_i64_release(ptr %mem, i64 %operand) {
; PPC32-LABEL: and_i64_release:
; PPC32: # %bb.0:
; PPC32-NEXT: mflr r0
; PPC32-NEXT: stwu r1, -16(r1)
; PPC32-NEXT: stw r0, 20(r1)
; PPC32-NEXT: .cfi_def_cfa_offset 16
; PPC32-NEXT: .cfi_offset lr, 4
; PPC32-NEXT: li r7, 3
; PPC32-NEXT: bl __atomic_fetch_and_8
; PPC32-NEXT: lwz r0, 20(r1)
; PPC32-NEXT: addi r1, r1, 16
; PPC32-NEXT: mtlr r0
; PPC32-NEXT: blr
;
; PPC64-LABEL: and_i64_release:
; PPC64: # %bb.0:
; PPC64-NEXT: lwsync
; PPC64-NEXT: .LBB15_1:
; PPC64-NEXT: ldarx r5, 0, r3
; PPC64-NEXT: and r6, r4, r5
; PPC64-NEXT: stdcx. r6, 0, r3
; PPC64-NEXT: bne cr0, .LBB15_1
; PPC64-NEXT: # %bb.2:
; PPC64-NEXT: mr r3, r5
; PPC64-NEXT: blr
%val = atomicrmw and ptr %mem, i64 %operand release
ret i64 %val
}
define half @load_atomic_f16__seq_cst(ptr %ptr) {
; PPC32-LABEL: load_atomic_f16__seq_cst:
; PPC32: # %bb.0:
; PPC32-NEXT: mflr r0
; PPC32-NEXT: stwu r1, -16(r1)
; PPC32-NEXT: stw r0, 20(r1)
; PPC32-NEXT: .cfi_def_cfa_offset 16
; PPC32-NEXT: .cfi_offset lr, 4
; PPC32-NEXT: sync
; PPC32-NEXT: lhz r3, 0(r3)
; PPC32-NEXT: cmpw cr7, r3, r3
; PPC32-NEXT: bne- cr7, .+4
; PPC32-NEXT: isync
; PPC32-NEXT: bl __extendhfsf2
; PPC32-NEXT: lwz r0, 20(r1)
; PPC32-NEXT: addi r1, r1, 16
; PPC32-NEXT: mtlr r0
; PPC32-NEXT: blr
;
; PPC64-LABEL: load_atomic_f16__seq_cst:
; PPC64: # %bb.0:
; PPC64-NEXT: mflr r0
; PPC64-NEXT: stdu r1, -112(r1)
; PPC64-NEXT: std r0, 128(r1)
; PPC64-NEXT: .cfi_def_cfa_offset 112
; PPC64-NEXT: .cfi_offset lr, 16
; PPC64-NEXT: sync
; PPC64-NEXT: lhz r3, 0(r3)
; PPC64-NEXT: cmpd cr7, r3, r3
; PPC64-NEXT: bne- cr7, .+4
; PPC64-NEXT: isync
; PPC64-NEXT: bl __extendhfsf2
; PPC64-NEXT: nop
; PPC64-NEXT: addi r1, r1, 112
; PPC64-NEXT: ld r0, 16(r1)
; PPC64-NEXT: mtlr r0
; PPC64-NEXT: blr
%val = load atomic half, ptr %ptr seq_cst, align 2
ret half %val
}
; FIXME: bf16_to_fp fails to select
; define bfloat @load_atomic_bf16__seq_cst(ptr %ptr) {
; %val = load atomic bfloat, ptr %ptr seq_cst, align 2
; ret bfloat %val
; }
define float @load_atomic_f32__seq_cst(ptr %ptr) {
; PPC32-LABEL: load_atomic_f32__seq_cst:
; PPC32: # %bb.0:
; PPC32-NEXT: stwu r1, -16(r1)
; PPC32-NEXT: .cfi_def_cfa_offset 16
; PPC32-NEXT: sync
; PPC32-NEXT: lwz r3, 0(r3)
; PPC32-NEXT: cmpw cr7, r3, r3
; PPC32-NEXT: bne- cr7, .+4
; PPC32-NEXT: isync
; PPC32-NEXT: stw r3, 12(r1)
; PPC32-NEXT: lfs f1, 12(r1)
; PPC32-NEXT: addi r1, r1, 16
; PPC32-NEXT: blr
;
; PPC64-LABEL: load_atomic_f32__seq_cst:
; PPC64: # %bb.0:
; PPC64-NEXT: sync
; PPC64-NEXT: lwz r3, 0(r3)
; PPC64-NEXT: cmpd cr7, r3, r3
; PPC64-NEXT: bne- cr7, .+4
; PPC64-NEXT: isync
; PPC64-NEXT: stw r3, -4(r1)
; PPC64-NEXT: lfs f1, -4(r1)
; PPC64-NEXT: blr
%val = load atomic float, ptr %ptr seq_cst, align 4
ret float %val
}
define double @load_atomic_f64__seq_cst(ptr %ptr) {
; PPC32-LABEL: load_atomic_f64__seq_cst:
; PPC32: # %bb.0:
; PPC32-NEXT: mflr r0
; PPC32-NEXT: stwu r1, -16(r1)
; PPC32-NEXT: stw r0, 20(r1)
; PPC32-NEXT: .cfi_def_cfa_offset 16
; PPC32-NEXT: .cfi_offset lr, 4
; PPC32-NEXT: li r4, 5
; PPC32-NEXT: bl __atomic_load_8
; PPC32-NEXT: stw r4, 12(r1)
; PPC32-NEXT: stw r3, 8(r1)
; PPC32-NEXT: lfd f1, 8(r1)
; PPC32-NEXT: lwz r0, 20(r1)
; PPC32-NEXT: addi r1, r1, 16
; PPC32-NEXT: mtlr r0
; PPC32-NEXT: blr
;
; PPC64-LABEL: load_atomic_f64__seq_cst:
; PPC64: # %bb.0:
; PPC64-NEXT: sync
; PPC64-NEXT: ld r3, 0(r3)
; PPC64-NEXT: cmpd cr7, r3, r3
; PPC64-NEXT: bne- cr7, .+4
; PPC64-NEXT: isync
; PPC64-NEXT: std r3, -8(r1)
; PPC64-NEXT: lfd f1, -8(r1)
; PPC64-NEXT: blr
%val = load atomic double, ptr %ptr seq_cst, align 8
ret double %val
}
define void @store_atomic_f16__seq_cst(ptr %ptr, half %val1) {
; PPC32-LABEL: store_atomic_f16__seq_cst:
; PPC32: # %bb.0:
; PPC32-NEXT: mflr r0
; PPC32-NEXT: stwu r1, -16(r1)
; PPC32-NEXT: stw r0, 20(r1)
; PPC32-NEXT: .cfi_def_cfa_offset 16
; PPC32-NEXT: .cfi_offset lr, 4
; PPC32-NEXT: .cfi_offset r30, -8
; PPC32-NEXT: stw r30, 8(r1) # 4-byte Folded Spill
; PPC32-NEXT: mr r30, r3
; PPC32-NEXT: bl __truncsfhf2
; PPC32-NEXT: sync
; PPC32-NEXT: sth r3, 0(r30)
; PPC32-NEXT: lwz r30, 8(r1) # 4-byte Folded Reload
; PPC32-NEXT: lwz r0, 20(r1)
; PPC32-NEXT: addi r1, r1, 16
; PPC32-NEXT: mtlr r0
; PPC32-NEXT: blr
;
; PPC64-LABEL: store_atomic_f16__seq_cst:
; PPC64: # %bb.0:
; PPC64-NEXT: mflr r0
; PPC64-NEXT: stdu r1, -128(r1)
; PPC64-NEXT: std r0, 144(r1)
; PPC64-NEXT: .cfi_def_cfa_offset 128
; PPC64-NEXT: .cfi_offset lr, 16
; PPC64-NEXT: .cfi_offset r30, -16
; PPC64-NEXT: std r30, 112(r1) # 8-byte Folded Spill
; PPC64-NEXT: mr r30, r3
; PPC64-NEXT: bl __truncsfhf2
; PPC64-NEXT: nop
; PPC64-NEXT: sync
; PPC64-NEXT: sth r3, 0(r30)
; PPC64-NEXT: ld r30, 112(r1) # 8-byte Folded Reload
; PPC64-NEXT: addi r1, r1, 128
; PPC64-NEXT: ld r0, 16(r1)
; PPC64-NEXT: mtlr r0
; PPC64-NEXT: blr
store atomic half %val1, ptr %ptr seq_cst, align 2
ret void
}
; FIXME: bf16_to_fp fails to select
; define void @store_atomic_bf16__seq_cst(ptr %ptr, bfloat %val1) {
; store atomic bfloat %val1, ptr %ptr seq_cst, align 2
; ret void
; }
define void @store_atomic_f32__seq_cst(ptr %ptr, float %val1) {
; PPC32-LABEL: store_atomic_f32__seq_cst:
; PPC32: # %bb.0:
; PPC32-NEXT: stwu r1, -16(r1)
; PPC32-NEXT: .cfi_def_cfa_offset 16
; PPC32-NEXT: stfs f1, 12(r1)
; PPC32-NEXT: lwz r4, 12(r1)
; PPC32-NEXT: sync
; PPC32-NEXT: stw r4, 0(r3)
; PPC32-NEXT: addi r1, r1, 16
; PPC32-NEXT: blr
;
; PPC64-LABEL: store_atomic_f32__seq_cst:
; PPC64: # %bb.0:
; PPC64-NEXT: stfs f1, -4(r1)
; PPC64-NEXT: lwz r4, -4(r1)
; PPC64-NEXT: sync
; PPC64-NEXT: stw r4, 0(r3)
; PPC64-NEXT: blr
store atomic float %val1, ptr %ptr seq_cst, align 4
ret void
}
define void @store_atomic_f64__seq_cst(ptr %ptr, double %val1) {
; PPC32-LABEL: store_atomic_f64__seq_cst:
; PPC32: # %bb.0:
; PPC32-NEXT: mflr r0
; PPC32-NEXT: stwu r1, -16(r1)
; PPC32-NEXT: stw r0, 20(r1)
; PPC32-NEXT: .cfi_def_cfa_offset 16
; PPC32-NEXT: .cfi_offset lr, 4
; PPC32-NEXT: stfd f1, 8(r1)
; PPC32-NEXT: li r7, 5
; PPC32-NEXT: lwz r5, 8(r1)
; PPC32-NEXT: lwz r6, 12(r1)
; PPC32-NEXT: bl __atomic_store_8
; PPC32-NEXT: lwz r0, 20(r1)
; PPC32-NEXT: addi r1, r1, 16
; PPC32-NEXT: mtlr r0
; PPC32-NEXT: blr
;
; PPC64-LABEL: store_atomic_f64__seq_cst:
; PPC64: # %bb.0:
; PPC64-NEXT: stfd f1, -8(r1)
; PPC64-NEXT: ld r4, -8(r1)
; PPC64-NEXT: sync
; PPC64-NEXT: std r4, 0(r3)
; PPC64-NEXT: blr
store atomic double %val1, ptr %ptr seq_cst, align 8
ret void
}
|
