aboutsummaryrefslogtreecommitdiff
path: root/libgomp/loop.c
blob: 682df39a4ed46dadb74091f30b87f472c200dd6e (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
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
/* Copyright (C) 2005-2022 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 LOOP (FOR/DO) construct.  */

#include <limits.h>
#include <stdlib.h>
#include <string.h>
#include "libgomp.h"


ialias (GOMP_loop_runtime_next)
ialias_redirect (GOMP_taskgroup_reduction_register)

/* Initialize the given work share construct from the given arguments.  */

static inline void
gomp_loop_init (struct gomp_work_share *ws, long start, long end, long incr,
		enum gomp_schedule_type sched, long chunk_size)
{
  ws->sched = sched;
  ws->chunk_size = chunk_size;
  /* Canonicalize loops that have zero iterations to ->next == ->end.  */
  ws->end = ((incr > 0 && start > end) || (incr < 0 && start < end))
	    ? start : end;
  ws->incr = incr;
  ws->next = start;
  if (sched == GFS_DYNAMIC)
    {
      ws->chunk_size *= incr;

#ifdef HAVE_SYNC_BUILTINS
      {
	/* For dynamic scheduling prepare things to make each iteration
	   faster.  */
	struct gomp_thread *thr = gomp_thread ();
	struct gomp_team *team = thr->ts.team;
	long nthreads = team ? team->nthreads : 1;

	if (__builtin_expect (incr > 0, 1))
	  {
	    /* Cheap overflow protection.  */
	    if (__builtin_expect ((nthreads | ws->chunk_size)
				  >= 1UL << (sizeof (long)
					     * __CHAR_BIT__ / 2 - 1), 0))
	      ws->mode = 0;
	    else
	      ws->mode = ws->end < (LONG_MAX
				    - (nthreads + 1) * ws->chunk_size);
	  }
	/* Cheap overflow protection.  */
	else if (__builtin_expect ((nthreads | -ws->chunk_size)
				   >= 1UL << (sizeof (long)
					      * __CHAR_BIT__ / 2 - 1), 0))
	  ws->mode = 0;
	else
	  ws->mode = ws->end > (nthreads + 1) * -ws->chunk_size - LONG_MAX;
      }
#endif
    }
}

/* The *_start routines are called when first encountering a loop construct
   that is not bound directly to a parallel construct.  The first thread
   that arrives will create the work-share construct; subsequent threads
   will see the construct exists and allocate work from it.

   START, END, INCR are the bounds of the loop; due to the restrictions of
   OpenMP, these values must be the same in every thread.  This is not
   verified (nor is it entirely verifiable, since START is not necessarily
   retained intact in the work-share data structure).  CHUNK_SIZE is the
   scheduling parameter; again this must be identical in all threads.

   Returns true if there's any work for this thread to perform.  If so,
   *ISTART and *IEND are filled with the bounds of the iteration block
   allocated to this thread.  Returns false if all work was assigned to
   other threads prior to this thread's arrival.  */

static bool
gomp_loop_static_start (long start, long end, long incr, long chunk_size,
			long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();

  thr->ts.static_trip = 0;
  if (gomp_work_share_start (0))
    {
      gomp_loop_init (thr->ts.work_share, start, end, incr,
		      GFS_STATIC, chunk_size);
      gomp_work_share_init_done ();
    }

  return !gomp_iter_static_next (istart, iend);
}

/* The current dynamic implementation is always monotonic.  The
   entrypoints without nonmonotonic in them have to be always monotonic,
   but the nonmonotonic ones could be changed to use work-stealing for
   improved scalability.  */

static bool
gomp_loop_dynamic_start (long start, long end, long incr, long chunk_size,
			 long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();
  bool ret;

  if (gomp_work_share_start (0))
    {
      gomp_loop_init (thr->ts.work_share, start, end, incr,
		      GFS_DYNAMIC, chunk_size);
      gomp_work_share_init_done ();
    }

#ifdef HAVE_SYNC_BUILTINS
  ret = gomp_iter_dynamic_next (istart, iend);
#else
  gomp_mutex_lock (&thr->ts.work_share->lock);
  ret = gomp_iter_dynamic_next_locked (istart, iend);
  gomp_mutex_unlock (&thr->ts.work_share->lock);
#endif

  return ret;
}

/* Similarly as for dynamic, though the question is how can the chunk sizes
   be decreased without a central locking or atomics.  */

static bool
gomp_loop_guided_start (long start, long end, long incr, long chunk_size,
			long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();
  bool ret;

  if (gomp_work_share_start (0))
    {
      gomp_loop_init (thr->ts.work_share, start, end, incr,
		      GFS_GUIDED, chunk_size);
      gomp_work_share_init_done ();
    }

#ifdef HAVE_SYNC_BUILTINS
  ret = gomp_iter_guided_next (istart, iend);
#else
  gomp_mutex_lock (&thr->ts.work_share->lock);
  ret = gomp_iter_guided_next_locked (istart, iend);
  gomp_mutex_unlock (&thr->ts.work_share->lock);
#endif

  return ret;
}

bool
GOMP_loop_runtime_start (long start, long end, long incr,
			 long *istart, long *iend)
{
  struct gomp_task_icv *icv = gomp_icv (false);
  switch (icv->run_sched_var & ~GFS_MONOTONIC)
    {
    case GFS_STATIC:
      return gomp_loop_static_start (start, end, incr,
				     icv->run_sched_chunk_size,
				     istart, iend);
    case GFS_DYNAMIC:
      return gomp_loop_dynamic_start (start, end, incr,
				      icv->run_sched_chunk_size,
				      istart, iend);
    case GFS_GUIDED:
      return gomp_loop_guided_start (start, end, incr,
				     icv->run_sched_chunk_size,
				     istart, iend);
    case GFS_AUTO:
      /* For now map to schedule(static), later on we could play with feedback
	 driven choice.  */
      return gomp_loop_static_start (start, end, incr, 0, istart, iend);
    default:
      abort ();
    }
}

static long
gomp_adjust_sched (long sched, long *chunk_size)
{
  sched &= ~GFS_MONOTONIC;
  switch (sched)
    {
    case GFS_STATIC:
    case GFS_DYNAMIC:
    case GFS_GUIDED:
      return sched;
    /* GFS_RUNTIME is used for runtime schedule without monotonic
       or nonmonotonic modifiers on the clause.
       GFS_RUNTIME|GFS_MONOTONIC for runtime schedule with monotonic
       modifier.  */
    case GFS_RUNTIME:
    /* GFS_AUTO is used for runtime schedule with nonmonotonic
       modifier.  */
    case GFS_AUTO:
      {
	struct gomp_task_icv *icv = gomp_icv (false);
	sched = icv->run_sched_var & ~GFS_MONOTONIC;
	switch (sched)
	  {
	  case GFS_STATIC:
	  case GFS_DYNAMIC:
	  case GFS_GUIDED:
	    *chunk_size = icv->run_sched_chunk_size;
	    break;
	  case GFS_AUTO:
	    sched = GFS_STATIC;
	    *chunk_size = 0;
	    break;
	  default:
	    abort ();
	  }
	return sched;
      }
    default:
      abort ();
    }
}

bool
GOMP_loop_start (long start, long end, long incr, long sched,
		 long chunk_size, long *istart, long *iend,
		 uintptr_t *reductions, void **mem)
{
  struct gomp_thread *thr = gomp_thread ();

  thr->ts.static_trip = 0;
  if (reductions)
    gomp_workshare_taskgroup_start ();
  if (gomp_work_share_start (0))
    {
      sched = gomp_adjust_sched (sched, &chunk_size);
      gomp_loop_init (thr->ts.work_share, start, end, incr,
		      sched, chunk_size);
      if (reductions)
	{
	  GOMP_taskgroup_reduction_register (reductions);
	  thr->task->taskgroup->workshare = true;
	  thr->ts.work_share->task_reductions = reductions;
	}
      if (mem)
	{
	  uintptr_t size = (uintptr_t) *mem;
#define INLINE_ORDERED_TEAM_IDS_OFF \
  ((offsetof (struct gomp_work_share, inline_ordered_team_ids)		\
    + __alignof__ (long long) - 1) & ~(__alignof__ (long long) - 1))
	  if (size > (sizeof (struct gomp_work_share)
		      - INLINE_ORDERED_TEAM_IDS_OFF))
	    *mem
	      = (void *) (thr->ts.work_share->ordered_team_ids
			  = gomp_malloc_cleared (size));
	  else
	    *mem = memset (((char *) thr->ts.work_share)
			   + INLINE_ORDERED_TEAM_IDS_OFF, '\0', size);
	}
      gomp_work_share_init_done ();
    }
  else
    {
      if (reductions)
	{
	  uintptr_t *first_reductions = thr->ts.work_share->task_reductions;
	  gomp_workshare_task_reduction_register (reductions,
						  first_reductions);
	}
      if (mem)
	{
	  if ((offsetof (struct gomp_work_share, inline_ordered_team_ids)
	       & (__alignof__ (long long) - 1)) == 0)
	    *mem = (void *) thr->ts.work_share->ordered_team_ids;
	  else
	    {
	      uintptr_t p = (uintptr_t) thr->ts.work_share->ordered_team_ids;
	      p += __alignof__ (long long) - 1;
	      p &= ~(__alignof__ (long long) - 1);
	      *mem = (void *) p;
	    }
	}
    }

  if (!istart)
    return true;
  return ialias_call (GOMP_loop_runtime_next) (istart, iend);
}

/* The *_ordered_*_start routines are similar.  The only difference is that
   this work-share construct is initialized to expect an ORDERED section.  */

static bool
gomp_loop_ordered_static_start (long start, long end, long incr,
				long chunk_size, long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();

  thr->ts.static_trip = 0;
  if (gomp_work_share_start (1))
    {
      gomp_loop_init (thr->ts.work_share, start, end, incr,
		      GFS_STATIC, chunk_size);
      gomp_ordered_static_init ();
      gomp_work_share_init_done ();
    }

  return !gomp_iter_static_next (istart, iend);
}

static bool
gomp_loop_ordered_dynamic_start (long start, long end, long incr,
				 long chunk_size, long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();
  bool ret;

  if (gomp_work_share_start (1))
    {
      gomp_loop_init (thr->ts.work_share, start, end, incr,
		      GFS_DYNAMIC, chunk_size);
      gomp_mutex_lock (&thr->ts.work_share->lock);
      gomp_work_share_init_done ();
    }
  else
    gomp_mutex_lock (&thr->ts.work_share->lock);

  ret = gomp_iter_dynamic_next_locked (istart, iend);
  if (ret)
    gomp_ordered_first ();
  gomp_mutex_unlock (&thr->ts.work_share->lock);

  return ret;
}

static bool
gomp_loop_ordered_guided_start (long start, long end, long incr,
				long chunk_size, long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();
  bool ret;

  if (gomp_work_share_start (1))
    {
      gomp_loop_init (thr->ts.work_share, start, end, incr,
		      GFS_GUIDED, chunk_size);
      gomp_mutex_lock (&thr->ts.work_share->lock);
      gomp_work_share_init_done ();
    }
  else
    gomp_mutex_lock (&thr->ts.work_share->lock);

  ret = gomp_iter_guided_next_locked (istart, iend);
  if (ret)
    gomp_ordered_first ();
  gomp_mutex_unlock (&thr->ts.work_share->lock);

  return ret;
}

bool
GOMP_loop_ordered_runtime_start (long start, long end, long incr,
				 long *istart, long *iend)
{
  struct gomp_task_icv *icv = gomp_icv (false);
  switch (icv->run_sched_var & ~GFS_MONOTONIC)
    {
    case GFS_STATIC:
      return gomp_loop_ordered_static_start (start, end, incr,
					     icv->run_sched_chunk_size,
					     istart, iend);
    case GFS_DYNAMIC:
      return gomp_loop_ordered_dynamic_start (start, end, incr,
					      icv->run_sched_chunk_size,
					      istart, iend);
    case GFS_GUIDED:
      return gomp_loop_ordered_guided_start (start, end, incr,
					     icv->run_sched_chunk_size,
					     istart, iend);
    case GFS_AUTO:
      /* For now map to schedule(static), later on we could play with feedback
	 driven choice.  */
      return gomp_loop_ordered_static_start (start, end, incr,
					     0, istart, iend);
    default:
      abort ();
    }
}

bool
GOMP_loop_ordered_start (long start, long end, long incr, long sched,
			 long chunk_size, long *istart, long *iend,
			 uintptr_t *reductions, void **mem)
{
  struct gomp_thread *thr = gomp_thread ();
  size_t ordered = 1;
  bool ret;

  thr->ts.static_trip = 0;
  if (reductions)
    gomp_workshare_taskgroup_start ();
  if (mem)
    ordered += (uintptr_t) *mem;
  if (gomp_work_share_start (ordered))
    {
      sched = gomp_adjust_sched (sched, &chunk_size);
      gomp_loop_init (thr->ts.work_share, start, end, incr,
		      sched, chunk_size);
      if (reductions)
	{
	  GOMP_taskgroup_reduction_register (reductions);
	  thr->task->taskgroup->workshare = true;
	  thr->ts.work_share->task_reductions = reductions;
	}
      if (sched == GFS_STATIC)
	gomp_ordered_static_init ();
      else
	gomp_mutex_lock (&thr->ts.work_share->lock);
      gomp_work_share_init_done ();
    }
  else
    {
      if (reductions)
	{
	  uintptr_t *first_reductions = thr->ts.work_share->task_reductions;
	  gomp_workshare_task_reduction_register (reductions,
						  first_reductions);
	}
      sched = thr->ts.work_share->sched;
      if (sched != GFS_STATIC)
	gomp_mutex_lock (&thr->ts.work_share->lock);
    }

  if (mem)
    {
      uintptr_t p
	= (uintptr_t) (thr->ts.work_share->ordered_team_ids
		       + (thr->ts.team ? thr->ts.team->nthreads : 1));
      p += __alignof__ (long long) - 1;
      p &= ~(__alignof__ (long long) - 1);
      *mem = (void *) p;
    }

  switch (sched)
    {
    case GFS_STATIC:
    case GFS_AUTO:
      return !gomp_iter_static_next (istart, iend);
    case GFS_DYNAMIC:
      ret = gomp_iter_dynamic_next_locked (istart, iend);
      break;
    case GFS_GUIDED:
      ret = gomp_iter_guided_next_locked (istart, iend);
      break;
    default:
      abort ();
    }

  if (ret)
    gomp_ordered_first ();
  gomp_mutex_unlock (&thr->ts.work_share->lock);
  return ret;
}

/* The *_doacross_*_start routines are similar.  The only difference is that
   this work-share construct is initialized to expect an ORDERED(N) - DOACROSS
   section, and the worksharing loop iterates always from 0 to COUNTS[0] - 1
   and other COUNTS array elements tell the library number of iterations
   in the ordered inner loops.  */

static bool
gomp_loop_doacross_static_start (unsigned ncounts, long *counts,
				 long chunk_size, long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();

  thr->ts.static_trip = 0;
  if (gomp_work_share_start (0))
    {
      gomp_loop_init (thr->ts.work_share, 0, counts[0], 1,
		      GFS_STATIC, chunk_size);
      gomp_doacross_init (ncounts, counts, chunk_size, 0);
      gomp_work_share_init_done ();
    }

  return !gomp_iter_static_next (istart, iend);
}

static bool
gomp_loop_doacross_dynamic_start (unsigned ncounts, long *counts,
				  long chunk_size, long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();
  bool ret;

  if (gomp_work_share_start (0))
    {
      gomp_loop_init (thr->ts.work_share, 0, counts[0], 1,
		      GFS_DYNAMIC, chunk_size);
      gomp_doacross_init (ncounts, counts, chunk_size, 0);
      gomp_work_share_init_done ();
    }

#ifdef HAVE_SYNC_BUILTINS
  ret = gomp_iter_dynamic_next (istart, iend);
#else
  gomp_mutex_lock (&thr->ts.work_share->lock);
  ret = gomp_iter_dynamic_next_locked (istart, iend);
  gomp_mutex_unlock (&thr->ts.work_share->lock);
#endif

  return ret;
}

static bool
gomp_loop_doacross_guided_start (unsigned ncounts, long *counts,
				 long chunk_size, long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();
  bool ret;

  if (gomp_work_share_start (0))
    {
      gomp_loop_init (thr->ts.work_share, 0, counts[0], 1,
		      GFS_GUIDED, chunk_size);
      gomp_doacross_init (ncounts, counts, chunk_size, 0);
      gomp_work_share_init_done ();
    }

#ifdef HAVE_SYNC_BUILTINS
  ret = gomp_iter_guided_next (istart, iend);
#else
  gomp_mutex_lock (&thr->ts.work_share->lock);
  ret = gomp_iter_guided_next_locked (istart, iend);
  gomp_mutex_unlock (&thr->ts.work_share->lock);
#endif

  return ret;
}

bool
GOMP_loop_doacross_runtime_start (unsigned ncounts, long *counts,
				  long *istart, long *iend)
{
  struct gomp_task_icv *icv = gomp_icv (false);
  switch (icv->run_sched_var & ~GFS_MONOTONIC)
    {
    case GFS_STATIC:
      return gomp_loop_doacross_static_start (ncounts, counts,
					      icv->run_sched_chunk_size,
					      istart, iend);
    case GFS_DYNAMIC:
      return gomp_loop_doacross_dynamic_start (ncounts, counts,
					       icv->run_sched_chunk_size,
					       istart, iend);
    case GFS_GUIDED:
      return gomp_loop_doacross_guided_start (ncounts, counts,
					      icv->run_sched_chunk_size,
					      istart, iend);
    case GFS_AUTO:
      /* For now map to schedule(static), later on we could play with feedback
	 driven choice.  */
      return gomp_loop_doacross_static_start (ncounts, counts,
					      0, istart, iend);
    default:
      abort ();
    }
}

bool
GOMP_loop_doacross_start (unsigned ncounts, long *counts, long sched,
			  long chunk_size, long *istart, long *iend,
			  uintptr_t *reductions, void **mem)
{
  struct gomp_thread *thr = gomp_thread ();

  thr->ts.static_trip = 0;
  if (reductions)
    gomp_workshare_taskgroup_start ();
  if (gomp_work_share_start (0))
    {
      size_t extra = 0;
      if (mem)
	extra = (uintptr_t) *mem;
      sched = gomp_adjust_sched (sched, &chunk_size);
      gomp_loop_init (thr->ts.work_share, 0, counts[0], 1,
		      sched, chunk_size);
      gomp_doacross_init (ncounts, counts, chunk_size, extra);
      if (reductions)
	{
	  GOMP_taskgroup_reduction_register (reductions);
	  thr->task->taskgroup->workshare = true;
	  thr->ts.work_share->task_reductions = reductions;
	}
      gomp_work_share_init_done ();
    }
  else
    {
      if (reductions)
	{
	  uintptr_t *first_reductions = thr->ts.work_share->task_reductions;
	  gomp_workshare_task_reduction_register (reductions,
						  first_reductions);
	}
      sched = thr->ts.work_share->sched;
    }

  if (mem)
    *mem = thr->ts.work_share->doacross->extra;

  return ialias_call (GOMP_loop_runtime_next) (istart, iend);
}

/* The *_next routines are called when the thread completes processing of
   the iteration block currently assigned to it.  If the work-share
   construct is bound directly to a parallel construct, then the iteration
   bounds may have been set up before the parallel.  In which case, this
   may be the first iteration for the thread.

   Returns true if there is work remaining to be performed; *ISTART and
   *IEND are filled with a new iteration block.  Returns false if all work
   has been assigned.  */

static bool
gomp_loop_static_next (long *istart, long *iend)
{
  return !gomp_iter_static_next (istart, iend);
}

static bool
gomp_loop_dynamic_next (long *istart, long *iend)
{
  bool ret;

#ifdef HAVE_SYNC_BUILTINS
  ret = gomp_iter_dynamic_next (istart, iend);
#else
  struct gomp_thread *thr = gomp_thread ();
  gomp_mutex_lock (&thr->ts.work_share->lock);
  ret = gomp_iter_dynamic_next_locked (istart, iend);
  gomp_mutex_unlock (&thr->ts.work_share->lock);
#endif

  return ret;
}

static bool
gomp_loop_guided_next (long *istart, long *iend)
{
  bool ret;

#ifdef HAVE_SYNC_BUILTINS
  ret = gomp_iter_guided_next (istart, iend);
#else
  struct gomp_thread *thr = gomp_thread ();
  gomp_mutex_lock (&thr->ts.work_share->lock);
  ret = gomp_iter_guided_next_locked (istart, iend);
  gomp_mutex_unlock (&thr->ts.work_share->lock);
#endif

  return ret;
}

bool
GOMP_loop_runtime_next (long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();

  switch (thr->ts.work_share->sched)
    {
    case GFS_STATIC:
    case GFS_AUTO:
      return gomp_loop_static_next (istart, iend);
    case GFS_DYNAMIC:
      return gomp_loop_dynamic_next (istart, iend);
    case GFS_GUIDED:
      return gomp_loop_guided_next (istart, iend);
    default:
      abort ();
    }
}

/* The *_ordered_*_next routines are called when the thread completes
   processing of the iteration block currently assigned to it.

   Returns true if there is work remaining to be performed; *ISTART and
   *IEND are filled with a new iteration block.  Returns false if all work
   has been assigned.  */

static bool
gomp_loop_ordered_static_next (long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();
  int test;

  gomp_ordered_sync ();
  gomp_mutex_lock (&thr->ts.work_share->lock);
  test = gomp_iter_static_next (istart, iend);
  if (test >= 0)
    gomp_ordered_static_next ();
  gomp_mutex_unlock (&thr->ts.work_share->lock);

  return test == 0;
}

static bool
gomp_loop_ordered_dynamic_next (long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();
  bool ret;

  gomp_ordered_sync ();
  gomp_mutex_lock (&thr->ts.work_share->lock);
  ret = gomp_iter_dynamic_next_locked (istart, iend);
  if (ret)
    gomp_ordered_next ();
  else
    gomp_ordered_last ();
  gomp_mutex_unlock (&thr->ts.work_share->lock);

  return ret;
}

static bool
gomp_loop_ordered_guided_next (long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();
  bool ret;

  gomp_ordered_sync ();
  gomp_mutex_lock (&thr->ts.work_share->lock);
  ret = gomp_iter_guided_next_locked (istart, iend);
  if (ret)
    gomp_ordered_next ();
  else
    gomp_ordered_last ();
  gomp_mutex_unlock (&thr->ts.work_share->lock);

  return ret;
}

bool
GOMP_loop_ordered_runtime_next (long *istart, long *iend)
{
  struct gomp_thread *thr = gomp_thread ();

  switch (thr->ts.work_share->sched)
    {
    case GFS_STATIC:
    case GFS_AUTO:
      return gomp_loop_ordered_static_next (istart, iend);
    case GFS_DYNAMIC:
      return gomp_loop_ordered_dynamic_next (istart, iend);
    case GFS_GUIDED:
      return gomp_loop_ordered_guided_next (istart, iend);
    default:
      abort ();
    }
}

/* The GOMP_parallel_loop_* routines pre-initialize a work-share construct
   to avoid one synchronization once we get into the loop.  */

static void
gomp_parallel_loop_start (void (*fn) (void *), void *data,
			  unsigned num_threads, long start, long end,
			  long incr, enum gomp_schedule_type sched,
			  long chunk_size, unsigned int flags)
{
  struct gomp_team *team;

  num_threads = gomp_resolve_num_threads (num_threads, 0);
  team = gomp_new_team (num_threads);
  gomp_loop_init (&team->work_shares[0], start, end, incr, sched, chunk_size);
  gomp_team_start (fn, data, num_threads, flags, team, NULL);
}

void
GOMP_parallel_loop_static_start (void (*fn) (void *), void *data,
				 unsigned num_threads, long start, long end,
				 long incr, long chunk_size)
{
  gomp_parallel_loop_start (fn, data, num_threads, start, end, incr,
			    GFS_STATIC, chunk_size, 0);
}

void
GOMP_parallel_loop_dynamic_start (void (*fn) (void *), void *data,
				  unsigned num_threads, long start, long end,
				  long incr, long chunk_size)
{
  gomp_parallel_loop_start (fn, data, num_threads, start, end, incr,
			    GFS_DYNAMIC, chunk_size, 0);
}

void
GOMP_parallel_loop_guided_start (void (*fn) (void *), void *data,
				 unsigned num_threads, long start, long end,
				 long incr, long chunk_size)
{
  gomp_parallel_loop_start (fn, data, num_threads, start, end, incr,
			    GFS_GUIDED, chunk_size, 0);
}

void
GOMP_parallel_loop_runtime_start (void (*fn) (void *), void *data,
				  unsigned num_threads, long start, long end,
				  long incr)
{
  struct gomp_task_icv *icv = gomp_icv (false);
  gomp_parallel_loop_start (fn, data, num_threads, start, end, incr,
			    icv->run_sched_var & ~GFS_MONOTONIC,
			    icv->run_sched_chunk_size, 0);
}

ialias_redirect (GOMP_parallel_end)

void
GOMP_parallel_loop_static (void (*fn) (void *), void *data,
			   unsigned num_threads, long start, long end,
			   long incr, long chunk_size, unsigned flags)
{
  gomp_parallel_loop_start (fn, data, num_threads, start, end, incr,
			    GFS_STATIC, chunk_size, flags);
  fn (data);
  GOMP_parallel_end ();
}

void
GOMP_parallel_loop_dynamic (void (*fn) (void *), void *data,
			    unsigned num_threads, long start, long end,
			    long incr, long chunk_size, unsigned flags)
{
  gomp_parallel_loop_start (fn, data, num_threads, start, end, incr,
			    GFS_DYNAMIC, chunk_size, flags);
  fn (data);
  GOMP_parallel_end ();
}

void
GOMP_parallel_loop_guided (void (*fn) (void *), void *data,
			  unsigned num_threads, long start, long end,
			  long incr, long chunk_size, unsigned flags)
{
  gomp_parallel_loop_start (fn, data, num_threads, start, end, incr,
			    GFS_GUIDED, chunk_size, flags);
  fn (data);
  GOMP_parallel_end ();
}

void
GOMP_parallel_loop_runtime (void (*fn) (void *), void *data,
			    unsigned num_threads, long start, long end,
			    long incr, unsigned flags)
{
  struct gomp_task_icv *icv = gomp_icv (false);
  gomp_parallel_loop_start (fn, data, num_threads, start, end, incr,
			    icv->run_sched_var & ~GFS_MONOTONIC,
			    icv->run_sched_chunk_size, flags);
  fn (data);
  GOMP_parallel_end ();
}

#ifdef HAVE_ATTRIBUTE_ALIAS
extern __typeof(GOMP_parallel_loop_dynamic) GOMP_parallel_loop_nonmonotonic_dynamic
	__attribute__((alias ("GOMP_parallel_loop_dynamic")));
extern __typeof(GOMP_parallel_loop_guided) GOMP_parallel_loop_nonmonotonic_guided
	__attribute__((alias ("GOMP_parallel_loop_guided")));
extern __typeof(GOMP_parallel_loop_runtime) GOMP_parallel_loop_nonmonotonic_runtime
	__attribute__((alias ("GOMP_parallel_loop_runtime")));
extern __typeof(GOMP_parallel_loop_runtime) GOMP_parallel_loop_maybe_nonmonotonic_runtime
	__attribute__((alias ("GOMP_parallel_loop_runtime")));
#else
void
GOMP_parallel_loop_nonmonotonic_dynamic (void (*fn) (void *), void *data,
					 unsigned num_threads, long start,
					 long end, long incr, long chunk_size,
					 unsigned flags)
{
  gomp_parallel_loop_start (fn, data, num_threads, start, end, incr,
			    GFS_DYNAMIC, chunk_size, flags);
  fn (data);
  GOMP_parallel_end ();
}

void
GOMP_parallel_loop_nonmonotonic_guided (void (*fn) (void *), void *data,
					unsigned num_threads, long start,
					long end, long incr, long chunk_size,
					unsigned flags)
{
  gomp_parallel_loop_start (fn, data, num_threads, start, end, incr,
			    GFS_GUIDED, chunk_size, flags);
  fn (data);
  GOMP_parallel_end ();
}

void
GOMP_parallel_loop_nonmonotonic_runtime (void (*fn) (void *), void *data,
					 unsigned num_threads, long start,
					 long end, long incr, unsigned flags)
{
  struct gomp_task_icv *icv = gomp_icv (false);
  gomp_parallel_loop_start (fn, data, num_threads, start, end, incr,
			    icv->run_sched_var & ~GFS_MONOTONIC,
			    icv->run_sched_chunk_size, flags);
  fn (data);
  GOMP_parallel_end ();
}

void
GOMP_parallel_loop_maybe_nonmonotonic_runtime (void (*fn) (void *), void *data,
					       unsigned num_threads, long start,
					       long end, long incr,
					       unsigned flags)
{
  struct gomp_task_icv *icv = gomp_icv (false);
  gomp_parallel_loop_start (fn, data, num_threads, start, end, incr,
			    icv->run_sched_var & ~GFS_MONOTONIC,
			    icv->run_sched_chunk_size, flags);
  fn (data);
  GOMP_parallel_end ();
}
#endif

/* The GOMP_loop_end* routines are called after the thread is told that
   all loop iterations are complete.  The first two versions synchronize
   all threads; the nowait version does not.  */

void
GOMP_loop_end (void)
{
  gomp_work_share_end ();
}

bool
GOMP_loop_end_cancel (void)
{
  return gomp_work_share_end_cancel ();
}

void
GOMP_loop_end_nowait (void)
{
  gomp_work_share_end_nowait ();
}


/* We use static functions above so that we're sure that the "runtime"
   function can defer to the proper routine without interposition.  We
   export the static function with a strong alias when possible, or with
   a wrapper function otherwise.  */

#ifdef HAVE_ATTRIBUTE_ALIAS
extern __typeof(gomp_loop_static_start) GOMP_loop_static_start
	__attribute__((alias ("gomp_loop_static_start")));
extern __typeof(gomp_loop_dynamic_start) GOMP_loop_dynamic_start
	__attribute__((alias ("gomp_loop_dynamic_start")));
extern __typeof(gomp_loop_guided_start) GOMP_loop_guided_start
	__attribute__((alias ("gomp_loop_guided_start")));
extern __typeof(gomp_loop_dynamic_start) GOMP_loop_nonmonotonic_dynamic_start
	__attribute__((alias ("gomp_loop_dynamic_start")));
extern __typeof(gomp_loop_guided_start) GOMP_loop_nonmonotonic_guided_start
	__attribute__((alias ("gomp_loop_guided_start")));
extern __typeof(GOMP_loop_runtime_start) GOMP_loop_nonmonotonic_runtime_start
	__attribute__((alias ("GOMP_loop_runtime_start")));
extern __typeof(GOMP_loop_runtime_start) GOMP_loop_maybe_nonmonotonic_runtime_start
	__attribute__((alias ("GOMP_loop_runtime_start")));

extern __typeof(gomp_loop_ordered_static_start) GOMP_loop_ordered_static_start
	__attribute__((alias ("gomp_loop_ordered_static_start")));
extern __typeof(gomp_loop_ordered_dynamic_start) GOMP_loop_ordered_dynamic_start
	__attribute__((alias ("gomp_loop_ordered_dynamic_start")));
extern __typeof(gomp_loop_ordered_guided_start) GOMP_loop_ordered_guided_start
	__attribute__((alias ("gomp_loop_ordered_guided_start")));

extern __typeof(gomp_loop_doacross_static_start) GOMP_loop_doacross_static_start
	__attribute__((alias ("gomp_loop_doacross_static_start")));
extern __typeof(gomp_loop_doacross_dynamic_start) GOMP_loop_doacross_dynamic_start
	__attribute__((alias ("gomp_loop_doacross_dynamic_start")));
extern __typeof(gomp_loop_doacross_guided_start) GOMP_loop_doacross_guided_start
	__attribute__((alias ("gomp_loop_doacross_guided_start")));

extern __typeof(gomp_loop_static_next) GOMP_loop_static_next
	__attribute__((alias ("gomp_loop_static_next")));
extern __typeof(gomp_loop_dynamic_next) GOMP_loop_dynamic_next
	__attribute__((alias ("gomp_loop_dynamic_next")));
extern __typeof(gomp_loop_guided_next) GOMP_loop_guided_next
	__attribute__((alias ("gomp_loop_guided_next")));
extern __typeof(gomp_loop_dynamic_next) GOMP_loop_nonmonotonic_dynamic_next
	__attribute__((alias ("gomp_loop_dynamic_next")));
extern __typeof(gomp_loop_guided_next) GOMP_loop_nonmonotonic_guided_next
	__attribute__((alias ("gomp_loop_guided_next")));
extern __typeof(GOMP_loop_runtime_next) GOMP_loop_nonmonotonic_runtime_next
	__attribute__((alias ("GOMP_loop_runtime_next")));
extern __typeof(GOMP_loop_runtime_next) GOMP_loop_maybe_nonmonotonic_runtime_next
	__attribute__((alias ("GOMP_loop_runtime_next")));

extern __typeof(gomp_loop_ordered_static_next) GOMP_loop_ordered_static_next
	__attribute__((alias ("gomp_loop_ordered_static_next")));
extern __typeof(gomp_loop_ordered_dynamic_next) GOMP_loop_ordered_dynamic_next
	__attribute__((alias ("gomp_loop_ordered_dynamic_next")));
extern __typeof(gomp_loop_ordered_guided_next) GOMP_loop_ordered_guided_next
	__attribute__((alias ("gomp_loop_ordered_guided_next")));
#else
bool
GOMP_loop_static_start (long start, long end, long incr, long chunk_size,
			long *istart, long *iend)
{
  return gomp_loop_static_start (start, end, incr, chunk_size, istart, iend);
}

bool
GOMP_loop_dynamic_start (long start, long end, long incr, long chunk_size,
			 long *istart, long *iend)
{
  return gomp_loop_dynamic_start (start, end, incr, chunk_size, istart, iend);
}

bool
GOMP_loop_guided_start (long start, long end, long incr, long chunk_size,
			long *istart, long *iend)
{
  return gomp_loop_guided_start (start, end, incr, chunk_size, istart, iend);
}

bool
GOMP_loop_nonmonotonic_dynamic_start (long start, long end, long incr,
				      long chunk_size, long *istart,
				      long *iend)
{
  return gomp_loop_dynamic_start (start, end, incr, chunk_size, istart, iend);
}

bool
GOMP_loop_nonmonotonic_guided_start (long start, long end, long incr,
				     long chunk_size, long *istart, long *iend)
{
  return gomp_loop_guided_start (start, end, incr, chunk_size, istart, iend);
}

bool
GOMP_loop_nonmonotonic_runtime_start (long start, long end, long incr,
				      long *istart, long *iend)
{
  return GOMP_loop_runtime_start (start, end, incr, istart, iend);
}

bool
GOMP_loop_maybe_nonmonotonic_runtime_start (long start, long end, long incr,
					    long *istart, long *iend)
{
  return GOMP_loop_runtime_start (start, end, incr, istart, iend);
}

bool
GOMP_loop_ordered_static_start (long start, long end, long incr,
				long chunk_size, long *istart, long *iend)
{
  return gomp_loop_ordered_static_start (start, end, incr, chunk_size,
					 istart, iend);
}

bool
GOMP_loop_ordered_dynamic_start (long start, long end, long incr,
				 long chunk_size, long *istart, long *iend)
{
  return gomp_loop_ordered_dynamic_start (start, end, incr, chunk_size,
					  istart, iend);
}

bool
GOMP_loop_ordered_guided_start (long start, long end, long incr,
				long chunk_size, long *istart, long *iend)
{
  return gomp_loop_ordered_guided_start (start, end, incr, chunk_size,
					 istart, iend);
}

bool
GOMP_loop_doacross_static_start (unsigned ncounts, long *counts,
				 long chunk_size, long *istart, long *iend)
{
  return gomp_loop_doacross_static_start (ncounts, counts, chunk_size,
					  istart, iend);
}

bool
GOMP_loop_doacross_dynamic_start (unsigned ncounts, long *counts,
				  long chunk_size, long *istart, long *iend)
{
  return gomp_loop_doacross_dynamic_start (ncounts, counts, chunk_size,
					   istart, iend);
}

bool
GOMP_loop_doacross_guided_start (unsigned ncounts, long *counts,
				 long chunk_size, long *istart, long *iend)
{
  return gomp_loop_doacross_guided_start (ncounts, counts, chunk_size,
					  istart, iend);
}

bool
GOMP_loop_static_next (long *istart, long *iend)
{
  return gomp_loop_static_next (istart, iend);
}

bool
GOMP_loop_dynamic_next (long *istart, long *iend)
{
  return gomp_loop_dynamic_next (istart, iend);
}

bool
GOMP_loop_guided_next (long *istart, long *iend)
{
  return gomp_loop_guided_next (istart, iend);
}

bool
GOMP_loop_nonmonotonic_dynamic_next (long *istart, long *iend)
{
  return gomp_loop_dynamic_next (istart, iend);
}

bool
GOMP_loop_nonmonotonic_guided_next (long *istart, long *iend)
{
  return gomp_loop_guided_next (istart, iend);
}

bool
GOMP_loop_nonmonotonic_runtime_next (long *istart, long *iend)
{
  return GOMP_loop_runtime_next (istart, iend);
}

bool
GOMP_loop_maybe_nonmonotonic_runtime_next (long *istart, long *iend)
{
  return GOMP_loop_runtime_next (istart, iend);
}

bool
GOMP_loop_ordered_static_next (long *istart, long *iend)
{
  return gomp_loop_ordered_static_next (istart, iend);
}

bool
GOMP_loop_ordered_dynamic_next (long *istart, long *iend)
{
  return gomp_loop_ordered_dynamic_next (istart, iend);
}

bool
GOMP_loop_ordered_guided_next (long *istart, long *iend)
{
  return gomp_loop_ordered_guided_next (istart, iend);
}
#endif