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
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
|
/* Copyright (C) 2007-2015 Free Software Foundation, Inc.
Contributed by Richard Henderson <rth@redhat.com>.
This file is part of the GNU Offloading and Multi Processing Library
(libgomp).
Libgomp is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 3, or (at your option)
any later version.
Libgomp is distributed in the hope that it will be useful, but WITHOUT ANY
WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
FOR A PARTICULAR PURPOSE. See the GNU General Public License for
more details.
Under Section 7 of GPL version 3, you are granted additional
permissions described in the GCC Runtime Library Exception, version
3.1, as published by the Free Software Foundation.
You should have received a copy of the GNU General Public License and
a copy of the GCC Runtime Library Exception along with this program;
see the files COPYING3 and COPYING.RUNTIME respectively. If not, see
<http://www.gnu.org/licenses/>. */
/* This file handles the maintainence of tasks in response to task
creation and termination. */
#include "libgomp.h"
#include <stdlib.h>
#include <string.h>
#include "gomp-constants.h"
typedef struct gomp_task_depend_entry *hash_entry_type;
static inline void *
htab_alloc (size_t size)
{
return gomp_malloc (size);
}
static inline void
htab_free (void *ptr)
{
free (ptr);
}
#include "hashtab.h"
static inline hashval_t
htab_hash (hash_entry_type element)
{
return hash_pointer (element->addr);
}
static inline bool
htab_eq (hash_entry_type x, hash_entry_type y)
{
return x->addr == y->addr;
}
/* Create a new task data structure. */
void
gomp_init_task (struct gomp_task *task, struct gomp_task *parent_task,
struct gomp_task_icv *prev_icv)
{
task->parent = parent_task;
task->icv = *prev_icv;
task->kind = GOMP_TASK_IMPLICIT;
task->taskwait = NULL;
task->in_tied_task = false;
task->final_task = false;
task->copy_ctors_done = false;
task->parent_depends_on = false;
task->children = NULL;
task->taskgroup = NULL;
task->dependers = NULL;
task->depend_hash = NULL;
task->depend_count = 0;
}
/* Clean up a task, after completing it. */
void
gomp_end_task (void)
{
struct gomp_thread *thr = gomp_thread ();
struct gomp_task *task = thr->task;
gomp_finish_task (task);
thr->task = task->parent;
}
/* Orphan the task in CHILDREN and all its siblings. */
static inline void
gomp_clear_parent (struct gomp_task *children)
{
struct gomp_task *task = children;
if (task)
do
{
task->parent = NULL;
task = task->next_child;
}
while (task != children);
}
/* Helper function for GOMP_task and gomp_create_target_task. Depend clause
handling for undeferred task creation. */
static void
gomp_task_handle_depend (struct gomp_task *task, struct gomp_task *parent,
void **depend)
{
size_t ndepend = (uintptr_t) depend[0];
size_t nout = (uintptr_t) depend[1];
size_t i;
hash_entry_type ent;
task->depend_count = ndepend;
task->num_dependees = 0;
if (parent->depend_hash == NULL)
parent->depend_hash = htab_create (2 * ndepend > 12 ? 2 * ndepend : 12);
for (i = 0; i < ndepend; i++)
{
task->depend[i].addr = depend[2 + i];
task->depend[i].next = NULL;
task->depend[i].prev = NULL;
task->depend[i].task = task;
task->depend[i].is_in = i >= nout;
task->depend[i].redundant = false;
task->depend[i].redundant_out = false;
hash_entry_type *slot = htab_find_slot (&parent->depend_hash,
&task->depend[i], INSERT);
hash_entry_type out = NULL, last = NULL;
if (*slot)
{
/* If multiple depends on the same task are the same, all but the
first one are redundant. As inout/out come first, if any of them
is inout/out, it will win, which is the right semantics. */
if ((*slot)->task == task)
{
task->depend[i].redundant = true;
continue;
}
for (ent = *slot; ent; ent = ent->next)
{
if (ent->redundant_out)
break;
last = ent;
/* depend(in:...) doesn't depend on earlier depend(in:...). */
if (i >= nout && ent->is_in)
continue;
if (!ent->is_in)
out = ent;
struct gomp_task *tsk = ent->task;
if (tsk->dependers == NULL)
{
tsk->dependers
= gomp_malloc (sizeof (struct gomp_dependers_vec)
+ 6 * sizeof (struct gomp_task *));
tsk->dependers->n_elem = 1;
tsk->dependers->allocated = 6;
tsk->dependers->elem[0] = task;
task->num_dependees++;
continue;
}
/* We already have some other dependency on tsk from earlier
depend clause. */
else if (tsk->dependers->n_elem
&& (tsk->dependers->elem[tsk->dependers->n_elem - 1]
== task))
continue;
else if (tsk->dependers->n_elem == tsk->dependers->allocated)
{
tsk->dependers->allocated
= tsk->dependers->allocated * 2 + 2;
tsk->dependers
= gomp_realloc (tsk->dependers,
sizeof (struct gomp_dependers_vec)
+ (tsk->dependers->allocated
* sizeof (struct gomp_task *)));
}
tsk->dependers->elem[tsk->dependers->n_elem++] = task;
task->num_dependees++;
}
task->depend[i].next = *slot;
(*slot)->prev = &task->depend[i];
}
*slot = &task->depend[i];
/* There is no need to store more than one depend({,in}out:) task per
address in the hash table chain for the purpose of creation of
deferred tasks, because each out depends on all earlier outs, thus it
is enough to record just the last depend({,in}out:). For depend(in:),
we need to keep all of the previous ones not terminated yet, because
a later depend({,in}out:) might need to depend on all of them. So, if
the new task's clause is depend({,in}out:), we know there is at most
one other depend({,in}out:) clause in the list (out). For
non-deferred tasks we want to see all outs, so they are moved to the
end of the chain, after first redundant_out entry all following
entries should be redundant_out. */
if (!task->depend[i].is_in && out)
{
if (out != last)
{
out->next->prev = out->prev;
out->prev->next = out->next;
out->next = last->next;
out->prev = last;
last->next = out;
if (out->next)
out->next->prev = out;
}
out->redundant_out = true;
}
}
}
/* Called when encountering an explicit task directive. If IF_CLAUSE is
false, then we must not delay in executing the task. If UNTIED is true,
then the task may be executed by any member of the team.
DEPEND is an array containing:
depend[0]: number of depend elements.
depend[1]: number of depend elements of type "out".
depend[2..N+1]: address of [1..N]th depend element. */
void
GOMP_task (void (*fn) (void *), void *data, void (*cpyfn) (void *, void *),
long arg_size, long arg_align, bool if_clause, unsigned flags,
void **depend, int priority)
{
struct gomp_thread *thr = gomp_thread ();
struct gomp_team *team = thr->ts.team;
#ifdef HAVE_BROKEN_POSIX_SEMAPHORES
/* If pthread_mutex_* is used for omp_*lock*, then each task must be
tied to one thread all the time. This means UNTIED tasks must be
tied and if CPYFN is non-NULL IF(0) must be forced, as CPYFN
might be running on different thread than FN. */
if (cpyfn)
if_clause = false;
flags &= ~GOMP_TASK_FLAG_UNTIED;
#endif
/* If parallel or taskgroup has been cancelled, don't start new tasks. */
if (team
&& (gomp_team_barrier_cancelled (&team->barrier)
|| (thr->task->taskgroup && thr->task->taskgroup->cancelled)))
return;
if ((flags & GOMP_TASK_FLAG_PRIORITY) == 0)
priority = 0;
/* FIXME, use priority. */
(void) priority;
if (!if_clause || team == NULL
|| (thr->task && thr->task->final_task)
|| team->task_count > 64 * team->nthreads)
{
struct gomp_task task;
/* If there are depend clauses and earlier deferred sibling tasks
with depend clauses, check if there isn't a dependency. If there
is, we need to wait for them. There is no need to handle
depend clauses for non-deferred tasks other than this, because
the parent task is suspended until the child task finishes and thus
it can't start further child tasks. */
if ((flags & GOMP_TASK_FLAG_DEPEND)
&& thr->task && thr->task->depend_hash)
gomp_task_maybe_wait_for_dependencies (depend);
gomp_init_task (&task, thr->task, gomp_icv (false));
task.kind = GOMP_TASK_UNDEFERRED;
task.final_task = (thr->task && thr->task->final_task)
|| (flags & GOMP_TASK_FLAG_FINAL);
if (thr->task)
{
task.in_tied_task = thr->task->in_tied_task;
task.taskgroup = thr->task->taskgroup;
}
thr->task = &task;
if (__builtin_expect (cpyfn != NULL, 0))
{
char buf[arg_size + arg_align - 1];
char *arg = (char *) (((uintptr_t) buf + arg_align - 1)
& ~(uintptr_t) (arg_align - 1));
cpyfn (arg, data);
fn (arg);
}
else
fn (data);
/* Access to "children" is normally done inside a task_lock
mutex region, but the only way this particular task.children
can be set is if this thread's task work function (fn)
creates children. So since the setter is *this* thread, we
need no barriers here when testing for non-NULL. We can have
task.children set by the current thread then changed by a
child thread, but seeing a stale non-NULL value is not a
problem. Once past the task_lock acquisition, this thread
will see the real value of task.children. */
if (task.children != NULL)
{
gomp_mutex_lock (&team->task_lock);
gomp_clear_parent (task.children);
gomp_mutex_unlock (&team->task_lock);
}
gomp_end_task ();
}
else
{
struct gomp_task *task;
struct gomp_task *parent = thr->task;
struct gomp_taskgroup *taskgroup = parent->taskgroup;
char *arg;
bool do_wake;
size_t depend_size = 0;
if (flags & GOMP_TASK_FLAG_DEPEND)
depend_size = ((uintptr_t) depend[0]
* sizeof (struct gomp_task_depend_entry));
task = gomp_malloc (sizeof (*task) + depend_size
+ arg_size + arg_align - 1);
arg = (char *) (((uintptr_t) (task + 1) + depend_size + arg_align - 1)
& ~(uintptr_t) (arg_align - 1));
gomp_init_task (task, parent, gomp_icv (false));
task->kind = GOMP_TASK_UNDEFERRED;
task->in_tied_task = parent->in_tied_task;
task->taskgroup = taskgroup;
thr->task = task;
if (cpyfn)
{
cpyfn (arg, data);
task->copy_ctors_done = true;
}
else
memcpy (arg, data, arg_size);
thr->task = parent;
task->kind = GOMP_TASK_WAITING;
task->fn = fn;
task->fn_data = arg;
task->final_task = (flags & GOMP_TASK_FLAG_FINAL) >> 1;
gomp_mutex_lock (&team->task_lock);
/* If parallel or taskgroup has been cancelled, don't start new
tasks. */
if (__builtin_expect ((gomp_team_barrier_cancelled (&team->barrier)
|| (taskgroup && taskgroup->cancelled))
&& !task->copy_ctors_done, 0))
{
gomp_mutex_unlock (&team->task_lock);
gomp_finish_task (task);
free (task);
return;
}
if (taskgroup)
taskgroup->num_children++;
if (depend_size)
{
gomp_task_handle_depend (task, parent, depend);
if (task->num_dependees)
{
gomp_mutex_unlock (&team->task_lock);
return;
}
}
if (parent->children)
{
task->next_child = parent->children;
task->prev_child = parent->children->prev_child;
task->next_child->prev_child = task;
task->prev_child->next_child = task;
}
else
{
task->next_child = task;
task->prev_child = task;
}
parent->children = task;
if (taskgroup)
{
/* If applicable, place task into its taskgroup. */
if (taskgroup->children)
{
task->next_taskgroup = taskgroup->children;
task->prev_taskgroup = taskgroup->children->prev_taskgroup;
task->next_taskgroup->prev_taskgroup = task;
task->prev_taskgroup->next_taskgroup = task;
}
else
{
task->next_taskgroup = task;
task->prev_taskgroup = task;
}
taskgroup->children = task;
}
if (team->task_queue)
{
task->next_queue = team->task_queue;
task->prev_queue = team->task_queue->prev_queue;
task->next_queue->prev_queue = task;
task->prev_queue->next_queue = task;
}
else
{
task->next_queue = task;
task->prev_queue = task;
team->task_queue = task;
}
++team->task_count;
++team->task_queued_count;
gomp_team_barrier_set_task_pending (&team->barrier);
do_wake = team->task_running_count + !parent->in_tied_task
< team->nthreads;
gomp_mutex_unlock (&team->task_lock);
if (do_wake)
gomp_team_barrier_wake (&team->barrier, 1);
}
}
ialias (GOMP_taskgroup_start)
ialias (GOMP_taskgroup_end)
#define TYPE long
#define UTYPE unsigned long
#define TYPE_is_long 1
#include "taskloop.c"
#undef TYPE
#undef UTYPE
#undef TYPE_is_long
#define TYPE unsigned long long
#define UTYPE TYPE
#define GOMP_taskloop GOMP_taskloop_ull
#include "taskloop.c"
#undef TYPE
#undef UTYPE
#undef GOMP_taskloop
/* Called for nowait target tasks. */
void
gomp_create_target_task (struct gomp_device_descr *devicep,
void (*fn) (void *), size_t mapnum, void **hostaddrs,
size_t *sizes, unsigned short *kinds,
unsigned int flags, void **depend)
{
struct gomp_thread *thr = gomp_thread ();
struct gomp_team *team = thr->ts.team;
/* If parallel or taskgroup has been cancelled, don't start new tasks. */
if (team
&& (gomp_team_barrier_cancelled (&team->barrier)
|| (thr->task->taskgroup && thr->task->taskgroup->cancelled)))
return;
struct gomp_target_task *ttask;
struct gomp_task *task;
struct gomp_task *parent = thr->task;
struct gomp_taskgroup *taskgroup = parent->taskgroup;
bool do_wake;
size_t depend_size = 0;
if (depend != NULL)
depend_size = ((uintptr_t) depend[0]
* sizeof (struct gomp_task_depend_entry));
task = gomp_malloc (sizeof (*task) + depend_size
+ sizeof (*ttask)
+ mapnum * (sizeof (void *) + sizeof (size_t)
+ sizeof (unsigned short)));
gomp_init_task (task, parent, gomp_icv (false));
task->kind = GOMP_TASK_WAITING;
task->in_tied_task = parent->in_tied_task;
task->taskgroup = taskgroup;
ttask = (struct gomp_target_task *) &task->depend[(uintptr_t) depend[0]];
ttask->devicep = devicep;
ttask->fn = fn;
ttask->mapnum = mapnum;
memcpy (ttask->hostaddrs, hostaddrs, mapnum * sizeof (void *));
ttask->sizes = (size_t *) &ttask->hostaddrs[mapnum];
memcpy (ttask->sizes, sizes, mapnum * sizeof (size_t));
ttask->kinds = (unsigned short *) &ttask->sizes[mapnum];
memcpy (ttask->kinds, kinds, mapnum * sizeof (unsigned short));
ttask->flags = flags;
task->fn = gomp_target_task_fn;
task->fn_data = ttask;
task->final_task = 0;
gomp_mutex_lock (&team->task_lock);
/* If parallel or taskgroup has been cancelled, don't start new tasks. */
if (__builtin_expect (gomp_team_barrier_cancelled (&team->barrier)
|| (taskgroup && taskgroup->cancelled), 0))
{
gomp_mutex_unlock (&team->task_lock);
gomp_finish_task (task);
free (task);
return;
}
if (taskgroup)
taskgroup->num_children++;
if (depend_size)
{
gomp_task_handle_depend (task, parent, depend);
if (task->num_dependees)
{
gomp_mutex_unlock (&team->task_lock);
return;
}
}
if (parent->children)
{
task->next_child = parent->children;
task->prev_child = parent->children->prev_child;
task->next_child->prev_child = task;
task->prev_child->next_child = task;
}
else
{
task->next_child = task;
task->prev_child = task;
}
parent->children = task;
if (taskgroup)
{
/* If applicable, place task into its taskgroup. */
if (taskgroup->children)
{
task->next_taskgroup = taskgroup->children;
task->prev_taskgroup = taskgroup->children->prev_taskgroup;
task->next_taskgroup->prev_taskgroup = task;
task->prev_taskgroup->next_taskgroup = task;
}
else
{
task->next_taskgroup = task;
task->prev_taskgroup = task;
}
taskgroup->children = task;
}
if (team->task_queue)
{
task->next_queue = team->task_queue;
task->prev_queue = team->task_queue->prev_queue;
task->next_queue->prev_queue = task;
task->prev_queue->next_queue = task;
}
else
{
task->next_queue = task;
task->prev_queue = task;
team->task_queue = task;
}
++team->task_count;
++team->task_queued_count;
gomp_team_barrier_set_task_pending (&team->barrier);
do_wake = team->task_running_count + !parent->in_tied_task
< team->nthreads;
gomp_mutex_unlock (&team->task_lock);
if (do_wake)
gomp_team_barrier_wake (&team->barrier, 1);
}
#if _LIBGOMP_CHECKING
/* Sanity check TASK to make sure it is in its parent's children
queue, and that the tasks therein are in the right order.
The expected order is:
parent_depends_on WAITING tasks
!parent_depends_on WAITING tasks
TIED tasks
PARENT is the alleged parent of TASK. */
static void
verify_children_queue (struct gomp_task *task, struct gomp_task *parent)
{
if (task->parent != parent)
gomp_fatal ("verify_children_queue: incompatible parents");
/* It's OK, Annie was an orphan and she turned out all right. */
if (!parent)
return;
bool seen_tied = false;
bool seen_plain_waiting = false;
bool found = false;
struct gomp_task *t = parent->children;
while (1)
{
if (t == task)
found = true;
if (seen_tied && t->kind == GOMP_TASK_WAITING)
gomp_fatal ("verify_children_queue: WAITING task after TIED");
if (t->kind == GOMP_TASK_TIED)
seen_tied = true;
else if (t->kind == GOMP_TASK_WAITING)
{
if (t->parent_depends_on)
{
if (seen_plain_waiting)
gomp_fatal ("verify_children_queue: parent_depends_on after "
"!parent_depends_on");
}
else
seen_plain_waiting = true;
}
t = t->next_child;
if (t == parent->children)
break;
}
if (!found)
gomp_fatal ("verify_children_queue: child not found in parent queue");
}
/* Sanity check TASK to make sure it is in its taskgroup queue (if
applicable), and that the tasks therein are in the right order.
The expected order is that GOMP_TASK_WAITING tasks must come before
GOMP_TASK_TIED tasks.
TASK is the task. */
static void
verify_taskgroup_queue (struct gomp_task *task)
{
struct gomp_taskgroup *taskgroup = task->taskgroup;
if (!taskgroup)
return;
bool seen_tied = false;
bool found = false;
struct gomp_task *t = taskgroup->children;
while (1)
{
if (t == task)
found = true;
if (t->kind == GOMP_TASK_WAITING && seen_tied)
gomp_fatal ("verify_taskgroup_queue: WAITING task after TIED");
if (t->kind == GOMP_TASK_TIED)
seen_tied = true;
t = t->next_taskgroup;
if (t == taskgroup->children)
break;
}
if (!found)
gomp_fatal ("verify_taskgroup_queue: child not found in parent queue");
}
/* Verify that TASK is in the team's task queue. */
static void
verify_task_queue (struct gomp_task *task, struct gomp_team *team)
{
struct gomp_task *t = team->task_queue;
if (team)
while (1)
{
if (t == task)
return;
t = t->next_queue;
if (t == team->task_queue)
break;
}
gomp_fatal ("verify_team_queue: child not in team");
}
#endif
static inline bool
gomp_task_run_pre (struct gomp_task *child_task, struct gomp_task *parent,
struct gomp_team *team)
{
#if _LIBGOMP_CHECKING
verify_children_queue (child_task, parent);
verify_taskgroup_queue (child_task);
verify_task_queue (child_task, team);
#endif
if (parent)
{
/* Adjust children such that it will point to a next child,
while the current one is scheduled to be executed. This way,
GOMP_taskwait (and others) can schedule a next task while
waiting.
Do not remove it entirely from the circular list, as it is
still a child, though not one we should consider first (say
by GOMP_taskwait). */
if (parent->children == child_task)
parent->children = child_task->next_child;
/* TIED tasks cannot come before WAITING tasks. If we're about
to make this task TIED, rewire things appropriately.
However, a TIED task at the end is perfectly fine. */
else if (child_task->next_child->kind == GOMP_TASK_WAITING
&& child_task->next_child != parent->children)
{
/* Remove from the list. */
child_task->prev_child->next_child = child_task->next_child;
child_task->next_child->prev_child = child_task->prev_child;
/* Rewire at the end of its siblings. */
child_task->next_child = parent->children;
child_task->prev_child = parent->children->prev_child;
parent->children->prev_child->next_child = child_task;
parent->children->prev_child = child_task;
}
/* If the current task (child_task) is at the top of the
parent's last_parent_depends_on, it's about to be removed
from it. Adjust last_parent_depends_on appropriately. */
if (__builtin_expect (child_task->parent_depends_on, 0)
&& parent->taskwait->last_parent_depends_on == child_task)
{
/* The last_parent_depends_on list was built with all
parent_depends_on entries linked to the prev_child. Grab
the next last_parent_depends_on head from this prev_child if
available... */
if (child_task->prev_child->kind == GOMP_TASK_WAITING
&& child_task->prev_child->parent_depends_on)
parent->taskwait->last_parent_depends_on = child_task->prev_child;
else
{
/* ...otherwise, there are no more parent_depends_on
entries waiting to run. In which case, clear the
list. */
parent->taskwait->last_parent_depends_on = NULL;
}
}
}
/* Adjust taskgroup to point to the next taskgroup. See note above
regarding adjustment of children as to why the child_task is not
removed entirely from the circular list. */
struct gomp_taskgroup *taskgroup = child_task->taskgroup;
if (taskgroup)
{
if (taskgroup->children == child_task)
taskgroup->children = child_task->next_taskgroup;
/* TIED tasks cannot come before WAITING tasks. If we're about
to make this task TIED, rewire things appropriately.
However, a TIED task at the end is perfectly fine. */
else if (child_task->next_taskgroup->kind == GOMP_TASK_WAITING
&& child_task->next_taskgroup != taskgroup->children)
{
/* Remove from the list. */
child_task->prev_taskgroup->next_taskgroup
= child_task->next_taskgroup;
child_task->next_taskgroup->prev_taskgroup
= child_task->prev_taskgroup;
/* Rewire at the end of its taskgroup. */
child_task->next_taskgroup = taskgroup->children;
child_task->prev_taskgroup = taskgroup->children->prev_taskgroup;
taskgroup->children->prev_taskgroup->next_taskgroup = child_task;
taskgroup->children->prev_taskgroup = child_task;
}
}
/* Remove child_task from the task_queue. */
child_task->prev_queue->next_queue = child_task->next_queue;
child_task->next_queue->prev_queue = child_task->prev_queue;
if (team->task_queue == child_task)
{
if (child_task->next_queue != child_task)
team->task_queue = child_task->next_queue;
else
team->task_queue = NULL;
}
child_task->kind = GOMP_TASK_TIED;
if (--team->task_queued_count == 0)
gomp_team_barrier_clear_task_pending (&team->barrier);
if ((gomp_team_barrier_cancelled (&team->barrier)
|| (taskgroup && taskgroup->cancelled))
&& !child_task->copy_ctors_done)
return true;
return false;
}
static void
gomp_task_run_post_handle_depend_hash (struct gomp_task *child_task)
{
struct gomp_task *parent = child_task->parent;
size_t i;
for (i = 0; i < child_task->depend_count; i++)
if (!child_task->depend[i].redundant)
{
if (child_task->depend[i].next)
child_task->depend[i].next->prev = child_task->depend[i].prev;
if (child_task->depend[i].prev)
child_task->depend[i].prev->next = child_task->depend[i].next;
else
{
hash_entry_type *slot
= htab_find_slot (&parent->depend_hash, &child_task->depend[i],
NO_INSERT);
if (*slot != &child_task->depend[i])
abort ();
if (child_task->depend[i].next)
*slot = child_task->depend[i].next;
else
htab_clear_slot (parent->depend_hash, slot);
}
}
}
/* After CHILD_TASK has been run, adjust the various task queues to
give higher priority to the tasks that depend on CHILD_TASK.
TEAM is the team to which CHILD_TASK belongs to. */
static size_t
gomp_task_run_post_handle_dependers (struct gomp_task *child_task,
struct gomp_team *team)
{
struct gomp_task *parent = child_task->parent;
size_t i, count = child_task->dependers->n_elem, ret = 0;
for (i = 0; i < count; i++)
{
struct gomp_task *task = child_task->dependers->elem[i];
if (--task->num_dependees != 0)
continue;
struct gomp_taskgroup *taskgroup = task->taskgroup;
if (parent)
{
if (parent->children)
{
/* If parent is in gomp_task_maybe_wait_for_dependencies
and it doesn't need to wait for this task, put it after
all ready to run tasks it needs to wait for. */
if (parent->taskwait && parent->taskwait->last_parent_depends_on
&& !task->parent_depends_on)
{
/* Put depender in last_parent_depends_on. */
struct gomp_task *last_parent_depends_on
= parent->taskwait->last_parent_depends_on;
task->next_child = last_parent_depends_on->next_child;
task->prev_child = last_parent_depends_on;
}
else
{
/* Make depender a sibling of child_task, and place
it at the top of said sibling list. */
task->next_child = parent->children;
task->prev_child = parent->children->prev_child;
parent->children = task;
}
task->next_child->prev_child = task;
task->prev_child->next_child = task;
}
else
{
/* Make depender a sibling of child_task. */
task->next_child = task;
task->prev_child = task;
parent->children = task;
}
if (parent->taskwait)
{
if (parent->taskwait->in_taskwait)
{
parent->taskwait->in_taskwait = false;
gomp_sem_post (&parent->taskwait->taskwait_sem);
}
else if (parent->taskwait->in_depend_wait)
{
parent->taskwait->in_depend_wait = false;
gomp_sem_post (&parent->taskwait->taskwait_sem);
}
if (parent->taskwait->last_parent_depends_on == NULL
&& task->parent_depends_on)
parent->taskwait->last_parent_depends_on = task;
}
}
/* If depender is in a taskgroup, put it at the TOP of its
taskgroup. */
if (taskgroup)
{
if (taskgroup->children)
{
task->next_taskgroup = taskgroup->children;
task->prev_taskgroup = taskgroup->children->prev_taskgroup;
task->next_taskgroup->prev_taskgroup = task;
task->prev_taskgroup->next_taskgroup = task;
}
else
{
task->next_taskgroup = task;
task->prev_taskgroup = task;
}
taskgroup->children = task;
if (taskgroup->in_taskgroup_wait)
{
taskgroup->in_taskgroup_wait = false;
gomp_sem_post (&taskgroup->taskgroup_sem);
}
}
/* Put depender of child_task at the END of the team's
task_queue. */
if (team->task_queue)
{
task->next_queue = team->task_queue;
task->prev_queue = team->task_queue->prev_queue;
task->next_queue->prev_queue = task;
task->prev_queue->next_queue = task;
}
else
{
task->next_queue = task;
task->prev_queue = task;
team->task_queue = task;
}
++team->task_count;
++team->task_queued_count;
++ret;
}
free (child_task->dependers);
child_task->dependers = NULL;
if (ret > 1)
gomp_team_barrier_set_task_pending (&team->barrier);
return ret;
}
static inline size_t
gomp_task_run_post_handle_depend (struct gomp_task *child_task,
struct gomp_team *team)
{
if (child_task->depend_count == 0)
return 0;
/* If parent is gone already, the hash table is freed and nothing
will use the hash table anymore, no need to remove anything from it. */
if (child_task->parent != NULL)
gomp_task_run_post_handle_depend_hash (child_task);
if (child_task->dependers == NULL)
return 0;
return gomp_task_run_post_handle_dependers (child_task, team);
}
/* Remove CHILD_TASK from its parent. */
static inline void
gomp_task_run_post_remove_parent (struct gomp_task *child_task)
{
struct gomp_task *parent = child_task->parent;
if (parent == NULL)
return;
/* If this was the last task the parent was depending on,
synchronize with gomp_task_maybe_wait_for_dependencies so it can
clean up and return. */
if (__builtin_expect (child_task->parent_depends_on, 0)
&& --parent->taskwait->n_depend == 0
&& parent->taskwait->in_depend_wait)
{
parent->taskwait->in_depend_wait = false;
gomp_sem_post (&parent->taskwait->taskwait_sem);
}
/* Remove CHILD_TASK from its sibling list. */
child_task->prev_child->next_child = child_task->next_child;
child_task->next_child->prev_child = child_task->prev_child;
if (parent->children != child_task)
return;
if (child_task->next_child != child_task)
parent->children = child_task->next_child;
else
{
/* We access task->children in GOMP_taskwait
outside of the task lock mutex region, so
need a release barrier here to ensure memory
written by child_task->fn above is flushed
before the NULL is written. */
__atomic_store_n (&parent->children, NULL, MEMMODEL_RELEASE);
if (parent->taskwait && parent->taskwait->in_taskwait)
{
parent->taskwait->in_taskwait = false;
gomp_sem_post (&parent->taskwait->taskwait_sem);
}
}
}
/* Remove CHILD_TASK from its taskgroup. */
static inline void
gomp_task_run_post_remove_taskgroup (struct gomp_task *child_task)
{
struct gomp_taskgroup *taskgroup = child_task->taskgroup;
if (taskgroup == NULL)
return;
child_task->prev_taskgroup->next_taskgroup = child_task->next_taskgroup;
child_task->next_taskgroup->prev_taskgroup = child_task->prev_taskgroup;
if (taskgroup->num_children > 1)
--taskgroup->num_children;
else
{
/* We access taskgroup->num_children in GOMP_taskgroup_end
outside of the task lock mutex region, so
need a release barrier here to ensure memory
written by child_task->fn above is flushed
before the NULL is written. */
__atomic_store_n (&taskgroup->num_children, 0, MEMMODEL_RELEASE);
}
if (taskgroup->children != child_task)
return;
if (child_task->next_taskgroup != child_task)
taskgroup->children = child_task->next_taskgroup;
else
{
taskgroup->children = NULL;
if (taskgroup->in_taskgroup_wait)
{
taskgroup->in_taskgroup_wait = false;
gomp_sem_post (&taskgroup->taskgroup_sem);
}
}
}
void
gomp_barrier_handle_tasks (gomp_barrier_state_t state)
{
struct gomp_thread *thr = gomp_thread ();
struct gomp_team *team = thr->ts.team;
struct gomp_task *task = thr->task;
struct gomp_task *child_task = NULL;
struct gomp_task *to_free = NULL;
int do_wake = 0;
gomp_mutex_lock (&team->task_lock);
if (gomp_barrier_last_thread (state))
{
if (team->task_count == 0)
{
gomp_team_barrier_done (&team->barrier, state);
gomp_mutex_unlock (&team->task_lock);
gomp_team_barrier_wake (&team->barrier, 0);
return;
}
gomp_team_barrier_set_waiting_for_tasks (&team->barrier);
}
while (1)
{
bool cancelled = false;
if (team->task_queue != NULL)
{
child_task = team->task_queue;
cancelled = gomp_task_run_pre (child_task, child_task->parent,
team);
if (__builtin_expect (cancelled, 0))
{
if (to_free)
{
gomp_finish_task (to_free);
free (to_free);
to_free = NULL;
}
goto finish_cancelled;
}
team->task_running_count++;
child_task->in_tied_task = true;
}
gomp_mutex_unlock (&team->task_lock);
if (do_wake)
{
gomp_team_barrier_wake (&team->barrier, do_wake);
do_wake = 0;
}
if (to_free)
{
gomp_finish_task (to_free);
free (to_free);
to_free = NULL;
}
if (child_task)
{
thr->task = child_task;
child_task->fn (child_task->fn_data);
thr->task = task;
}
else
return;
gomp_mutex_lock (&team->task_lock);
if (child_task)
{
finish_cancelled:;
size_t new_tasks
= gomp_task_run_post_handle_depend (child_task, team);
gomp_task_run_post_remove_parent (child_task);
gomp_clear_parent (child_task->children);
gomp_task_run_post_remove_taskgroup (child_task);
to_free = child_task;
child_task = NULL;
if (!cancelled)
team->task_running_count--;
if (new_tasks > 1)
{
do_wake = team->nthreads - team->task_running_count;
if (do_wake > new_tasks)
do_wake = new_tasks;
}
if (--team->task_count == 0
&& gomp_team_barrier_waiting_for_tasks (&team->barrier))
{
gomp_team_barrier_done (&team->barrier, state);
gomp_mutex_unlock (&team->task_lock);
gomp_team_barrier_wake (&team->barrier, 0);
gomp_mutex_lock (&team->task_lock);
}
}
}
}
/* Called when encountering a taskwait directive.
Wait for all children of the current task. */
void
GOMP_taskwait (void)
{
struct gomp_thread *thr = gomp_thread ();
struct gomp_team *team = thr->ts.team;
struct gomp_task *task = thr->task;
struct gomp_task *child_task = NULL;
struct gomp_task *to_free = NULL;
struct gomp_taskwait taskwait;
int do_wake = 0;
/* The acquire barrier on load of task->children here synchronizes
with the write of a NULL in gomp_task_run_post_remove_parent. It is
not necessary that we synchronize with other non-NULL writes at
this point, but we must ensure that all writes to memory by a
child thread task work function are seen before we exit from
GOMP_taskwait. */
if (task == NULL
|| __atomic_load_n (&task->children, MEMMODEL_ACQUIRE) == NULL)
return;
memset (&taskwait, 0, sizeof (taskwait));
gomp_mutex_lock (&team->task_lock);
while (1)
{
bool cancelled = false;
if (task->children == NULL)
{
bool destroy_taskwait = task->taskwait != NULL;
task->taskwait = NULL;
gomp_mutex_unlock (&team->task_lock);
if (to_free)
{
gomp_finish_task (to_free);
free (to_free);
}
if (destroy_taskwait)
gomp_sem_destroy (&taskwait.taskwait_sem);
return;
}
if (task->children->kind == GOMP_TASK_WAITING)
{
child_task = task->children;
cancelled
= gomp_task_run_pre (child_task, task, team);
if (__builtin_expect (cancelled, 0))
{
if (to_free)
{
gomp_finish_task (to_free);
free (to_free);
to_free = NULL;
}
goto finish_cancelled;
}
}
else
{
/* All tasks we are waiting for are already running
in other threads. Wait for them. */
if (task->taskwait == NULL)
{
taskwait.in_depend_wait = false;
gomp_sem_init (&taskwait.taskwait_sem, 0);
task->taskwait = &taskwait;
}
taskwait.in_taskwait = true;
}
gomp_mutex_unlock (&team->task_lock);
if (do_wake)
{
gomp_team_barrier_wake (&team->barrier, do_wake);
do_wake = 0;
}
if (to_free)
{
gomp_finish_task (to_free);
free (to_free);
to_free = NULL;
}
if (child_task)
{
thr->task = child_task;
child_task->fn (child_task->fn_data);
thr->task = task;
}
else
gomp_sem_wait (&taskwait.taskwait_sem);
gomp_mutex_lock (&team->task_lock);
if (child_task)
{
finish_cancelled:;
size_t new_tasks
= gomp_task_run_post_handle_depend (child_task, team);
/* Remove child_task from children list, and set up the next
sibling to be run. */
child_task->prev_child->next_child = child_task->next_child;
child_task->next_child->prev_child = child_task->prev_child;
if (task->children == child_task)
{
if (child_task->next_child != child_task)
task->children = child_task->next_child;
else
task->children = NULL;
}
/* Orphan all the children of CHILD_TASK. */
gomp_clear_parent (child_task->children);
/* Remove CHILD_TASK from its taskgroup. */
gomp_task_run_post_remove_taskgroup (child_task);
to_free = child_task;
child_task = NULL;
team->task_count--;
if (new_tasks > 1)
{
do_wake = team->nthreads - team->task_running_count
- !task->in_tied_task;
if (do_wake > new_tasks)
do_wake = new_tasks;
}
}
}
}
/* This is like GOMP_taskwait, but we only wait for tasks that the
upcoming task depends on.
DEPEND is as in GOMP_task. */
void
gomp_task_maybe_wait_for_dependencies (void **depend)
{
struct gomp_thread *thr = gomp_thread ();
struct gomp_task *task = thr->task;
struct gomp_team *team = thr->ts.team;
struct gomp_task_depend_entry elem, *ent = NULL;
struct gomp_taskwait taskwait;
struct gomp_task *last_parent_depends_on = NULL;
size_t ndepend = (uintptr_t) depend[0];
size_t nout = (uintptr_t) depend[1];
size_t i;
size_t num_awaited = 0;
struct gomp_task *child_task = NULL;
struct gomp_task *to_free = NULL;
int do_wake = 0;
gomp_mutex_lock (&team->task_lock);
for (i = 0; i < ndepend; i++)
{
elem.addr = depend[i + 2];
ent = htab_find (task->depend_hash, &elem);
for (; ent; ent = ent->next)
if (i >= nout && ent->is_in)
continue;
else
{
struct gomp_task *tsk = ent->task;
if (!tsk->parent_depends_on)
{
tsk->parent_depends_on = true;
++num_awaited;
/* If a task we need to wait for is not already
running and is ready to be scheduled, move it to
front, so that we run it as soon as possible.
We rearrange the children queue such that all
parent_depends_on tasks are first, and
last_parent_depends_on points to the last such task
we rearranged. For example, given the following
children where PD[123] are the parent_depends_on
tasks:
task->children
|
V
C1 -> C2 -> C3 -> PD1 -> PD2 -> PD3 -> C4
We rearrange such that:
task->children
| +--- last_parent_depends_on
| |
V V
PD1 -> PD2 -> PD3 -> C1 -> C2 -> C3 -> C4
*/
if (tsk->num_dependees == 0 && tsk->kind == GOMP_TASK_WAITING)
{
if (last_parent_depends_on)
{
tsk->prev_child->next_child = tsk->next_child;
tsk->next_child->prev_child = tsk->prev_child;
tsk->prev_child = last_parent_depends_on;
tsk->next_child = last_parent_depends_on->next_child;
tsk->prev_child->next_child = tsk;
tsk->next_child->prev_child = tsk;
}
else if (tsk != task->children)
{
tsk->prev_child->next_child = tsk->next_child;
tsk->next_child->prev_child = tsk->prev_child;
tsk->prev_child = task->children->prev_child;
tsk->next_child = task->children;
task->children = tsk;
tsk->prev_child->next_child = tsk;
tsk->next_child->prev_child = tsk;
}
last_parent_depends_on = tsk;
}
}
}
}
if (num_awaited == 0)
{
gomp_mutex_unlock (&team->task_lock);
return;
}
memset (&taskwait, 0, sizeof (taskwait));
taskwait.n_depend = num_awaited;
taskwait.last_parent_depends_on = last_parent_depends_on;
gomp_sem_init (&taskwait.taskwait_sem, 0);
task->taskwait = &taskwait;
while (1)
{
bool cancelled = false;
if (taskwait.n_depend == 0)
{
task->taskwait = NULL;
gomp_mutex_unlock (&team->task_lock);
if (to_free)
{
gomp_finish_task (to_free);
free (to_free);
}
gomp_sem_destroy (&taskwait.taskwait_sem);
return;
}
if (task->children->kind == GOMP_TASK_WAITING)
{
child_task = task->children;
cancelled
= gomp_task_run_pre (child_task, task, team);
if (__builtin_expect (cancelled, 0))
{
if (to_free)
{
gomp_finish_task (to_free);
free (to_free);
to_free = NULL;
}
goto finish_cancelled;
}
}
else
/* All tasks we are waiting for are already running
in other threads. Wait for them. */
taskwait.in_depend_wait = true;
gomp_mutex_unlock (&team->task_lock);
if (do_wake)
{
gomp_team_barrier_wake (&team->barrier, do_wake);
do_wake = 0;
}
if (to_free)
{
gomp_finish_task (to_free);
free (to_free);
to_free = NULL;
}
if (child_task)
{
thr->task = child_task;
child_task->fn (child_task->fn_data);
thr->task = task;
}
else
gomp_sem_wait (&taskwait.taskwait_sem);
gomp_mutex_lock (&team->task_lock);
if (child_task)
{
finish_cancelled:;
size_t new_tasks
= gomp_task_run_post_handle_depend (child_task, team);
if (child_task->parent_depends_on)
--taskwait.n_depend;
/* Remove child_task from sibling list. */
child_task->prev_child->next_child = child_task->next_child;
child_task->next_child->prev_child = child_task->prev_child;
if (task->children == child_task)
{
if (child_task->next_child != child_task)
task->children = child_task->next_child;
else
task->children = NULL;
}
gomp_clear_parent (child_task->children);
gomp_task_run_post_remove_taskgroup (child_task);
to_free = child_task;
child_task = NULL;
team->task_count--;
if (new_tasks > 1)
{
do_wake = team->nthreads - team->task_running_count
- !task->in_tied_task;
if (do_wake > new_tasks)
do_wake = new_tasks;
}
}
}
}
/* Called when encountering a taskyield directive. */
void
GOMP_taskyield (void)
{
/* Nothing at the moment. */
}
void
GOMP_taskgroup_start (void)
{
struct gomp_thread *thr = gomp_thread ();
struct gomp_team *team = thr->ts.team;
struct gomp_task *task = thr->task;
struct gomp_taskgroup *taskgroup;
/* If team is NULL, all tasks are executed as
GOMP_TASK_UNDEFERRED tasks and thus all children tasks of
taskgroup and their descendant tasks will be finished
by the time GOMP_taskgroup_end is called. */
if (team == NULL)
return;
taskgroup = gomp_malloc (sizeof (struct gomp_taskgroup));
taskgroup->prev = task->taskgroup;
taskgroup->children = NULL;
taskgroup->in_taskgroup_wait = false;
taskgroup->cancelled = false;
taskgroup->num_children = 0;
gomp_sem_init (&taskgroup->taskgroup_sem, 0);
task->taskgroup = taskgroup;
}
void
GOMP_taskgroup_end (void)
{
struct gomp_thread *thr = gomp_thread ();
struct gomp_team *team = thr->ts.team;
struct gomp_task *task = thr->task;
struct gomp_taskgroup *taskgroup;
struct gomp_task *child_task = NULL;
struct gomp_task *to_free = NULL;
int do_wake = 0;
if (team == NULL)
return;
taskgroup = task->taskgroup;
/* The acquire barrier on load of taskgroup->num_children here
synchronizes with the write of 0 in gomp_task_run_post_remove_taskgroup.
It is not necessary that we synchronize with other non-0 writes at
this point, but we must ensure that all writes to memory by a
child thread task work function are seen before we exit from
GOMP_taskgroup_end. */
if (__atomic_load_n (&taskgroup->num_children, MEMMODEL_ACQUIRE) == 0)
goto finish;
gomp_mutex_lock (&team->task_lock);
while (1)
{
bool cancelled = false;
if (taskgroup->children == NULL)
{
if (taskgroup->num_children)
{
if (task->children == NULL)
goto do_wait;
child_task = task->children;
}
else
{
gomp_mutex_unlock (&team->task_lock);
if (to_free)
{
gomp_finish_task (to_free);
free (to_free);
}
goto finish;
}
}
else
child_task = taskgroup->children;
if (child_task->kind == GOMP_TASK_WAITING)
{
cancelled
= gomp_task_run_pre (child_task, child_task->parent, team);
if (__builtin_expect (cancelled, 0))
{
if (to_free)
{
gomp_finish_task (to_free);
free (to_free);
to_free = NULL;
}
goto finish_cancelled;
}
}
else
{
child_task = NULL;
do_wait:
/* All tasks we are waiting for are already running
in other threads. Wait for them. */
taskgroup->in_taskgroup_wait = true;
}
gomp_mutex_unlock (&team->task_lock);
if (do_wake)
{
gomp_team_barrier_wake (&team->barrier, do_wake);
do_wake = 0;
}
if (to_free)
{
gomp_finish_task (to_free);
free (to_free);
to_free = NULL;
}
if (child_task)
{
thr->task = child_task;
child_task->fn (child_task->fn_data);
thr->task = task;
}
else
gomp_sem_wait (&taskgroup->taskgroup_sem);
gomp_mutex_lock (&team->task_lock);
if (child_task)
{
finish_cancelled:;
size_t new_tasks
= gomp_task_run_post_handle_depend (child_task, team);
gomp_task_run_post_remove_parent (child_task);
gomp_clear_parent (child_task->children);
gomp_task_run_post_remove_taskgroup (child_task);
to_free = child_task;
child_task = NULL;
team->task_count--;
if (new_tasks > 1)
{
do_wake = team->nthreads - team->task_running_count
- !task->in_tied_task;
if (do_wake > new_tasks)
do_wake = new_tasks;
}
}
}
finish:
task->taskgroup = taskgroup->prev;
gomp_sem_destroy (&taskgroup->taskgroup_sem);
free (taskgroup);
}
int
omp_in_final (void)
{
struct gomp_thread *thr = gomp_thread ();
return thr->task && thr->task->final_task;
}
ialias (omp_in_final)
|