summaryrefslogtreecommitdiffstats
path: root/storage/perfschema/pfs_buffer_container.h
blob: b5506fe0195182afd5d752cdc079467c6681323c (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
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
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
/* Copyright (c) 2014, 2023, Oracle and/or its affiliates.

  This program is free software; you can redistribute it and/or modify
  it under the terms of the GNU General Public License, version 2.0,
  as published by the Free Software Foundation.

  This program is also distributed with certain software (including
  but not limited to OpenSSL) that is licensed under separate terms,
  as designated in a particular file or component or in included license
  documentation.  The authors of MySQL hereby grant you an additional
  permission to link the program and your derivative works with the
  separately licensed software that they have included with MySQL.

  This program 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, version 2.0, for more details.

  You should have received a copy of the GNU General Public License
  along with this program; if not, write to the Free Software Foundation,
  51 Franklin Street, Suite 500, Boston, MA 02110-1335 USA */

#ifndef PFS_BUFFER_CONTAINER_H
#define PFS_BUFFER_CONTAINER_H

#include "my_global.h"
#include "pfs.h" // PSI_COUNT_VOLATILITY
#include "pfs_lock.h"
#include "pfs_instr.h"
#include "pfs_setup_actor.h"
#include "pfs_setup_object.h"
#include "pfs_program.h"
#include "pfs_prepared_stmt.h"
#include "pfs_builtin_memory.h"

#define USE_SCALABLE

class PFS_opaque_container_page;
class PFS_opaque_container;

struct PFS_builtin_memory_class;

template <class T>
class PFS_buffer_const_iterator;

template <class T>
class PFS_buffer_processor;

template <class T, class U, class V>
class PFS_buffer_iterator;

template <class T, int PFS_PAGE_SIZE, int PFS_PAGE_COUNT, class U, class V>
class PFS_buffer_scalable_iterator;

template <class T>
class PFS_buffer_default_array;

template <class T>
class PFS_buffer_default_allocator;

template <class T, class U, class V>
class PFS_buffer_container;

template <class T, int PFS_PAGE_SIZE, int PFS_PAGE_COUNT, class U, class V>
class PFS_buffer_scalable_container;

template <class B, int COUNT>
class PFS_partitioned_buffer_scalable_iterator;

template <class B, int COUNT>
class PFS_partitioned_buffer_scalable_container;


template <class T>
class PFS_buffer_default_array
{
public:
  typedef T value_type;

  value_type *allocate(pfs_dirty_state *dirty_state)
  {
    uint index;
    uint monotonic;
    uint monotonic_max;
    value_type *pfs;

    if (m_full)
      return NULL;

    monotonic= PFS_atomic::add_u32(& m_monotonic.m_u32, 1);
    monotonic_max= monotonic + static_cast<uint>(m_max);

    while (monotonic < monotonic_max)
    {
      index= monotonic % m_max;
      pfs= m_ptr + index;

      if (pfs->m_lock.free_to_dirty(dirty_state))
      {
        return pfs;
      }
      monotonic= PFS_atomic::add_u32(& m_monotonic.m_u32, 1);
    }

    m_full= true;
    return NULL;
  }

  void deallocate(value_type *pfs)
  {
    pfs->m_lock.allocated_to_free();
    m_full= false;
  }

  T* get_first()
  {
    return m_ptr;
  }

  T* get_last()
  {
    return m_ptr + m_max;
  }

  bool m_full;
  PFS_cacheline_uint32 m_monotonic;
  T * m_ptr;
  size_t m_max;
  /** Container. */
  PFS_opaque_container *m_container;
};

template <class T>
class PFS_buffer_default_allocator
{
public:
  typedef PFS_buffer_default_array<T> array_type;

  PFS_buffer_default_allocator(PFS_builtin_memory_class *klass)
    : m_builtin_class(klass)
  {}

  int alloc_array(array_type *array)
  {
    array->m_ptr= NULL;
    array->m_full= true;
    array->m_monotonic.m_u32= 0;

    if (array->m_max > 0)
    {
      array->m_ptr= PFS_MALLOC_ARRAY(m_builtin_class,
                                     array->m_max, sizeof(T), T, MYF(MY_ZEROFILL));
      if (array->m_ptr == NULL)
        return 1;
      array->m_full= false;
    }
    return 0;
  }

  void free_array(array_type *array)
  {
    assert(array->m_max > 0);

    PFS_FREE_ARRAY(m_builtin_class,
                   array->m_max, sizeof(T), array->m_ptr);
    array->m_ptr= NULL;
  }

private:
  PFS_builtin_memory_class *m_builtin_class;
};

template <class T,
          class U = PFS_buffer_default_array<T>,
          class V = PFS_buffer_default_allocator<T> >
class PFS_buffer_container
{
public:
  friend class PFS_buffer_iterator<T, U, V>;

  typedef T value_type;
  typedef U array_type;
  typedef V allocator_type;
  typedef PFS_buffer_const_iterator<T> const_iterator_type;
  typedef PFS_buffer_iterator<T, U, V> iterator_type;
  typedef PFS_buffer_processor<T> processor_type;
  typedef void (*function_type)(value_type *);

  PFS_buffer_container(allocator_type *allocator)
  {
    m_array.m_full= true;
    m_array.m_ptr= NULL;
    m_array.m_max= 0;
    m_array.m_monotonic.m_u32= 0;
    m_lost= 0;
    m_max= 0;
    m_allocator= allocator;
  }

  int init(ulong max_size)
  {
    if (max_size > 0)
    {
      m_array.m_max= max_size;
      int rc= m_allocator->alloc_array(& m_array);
      if (rc != 0)
      {
        m_allocator->free_array(& m_array);
        return 1;
      }
      m_max= max_size;
      m_array.m_full= false;
    }
    return 0;
  }

  void cleanup()
  {
    m_allocator->free_array(& m_array);
  }

  ulong get_row_count() const
  {
    return m_max;
  }

  ulong get_row_size() const
  {
    return sizeof(value_type);
  }

  ulong get_memory() const
  {
    return get_row_count() * get_row_size();
  }

  value_type *allocate(pfs_dirty_state *dirty_state)
  {
    value_type *pfs;

    pfs= m_array.allocate(dirty_state, m_max);
    if (pfs == NULL)
    {
      m_lost++;
    }

    return pfs;
  }

  void deallocate(value_type *pfs)
  {
    m_array.deallocate(pfs);
  }

  iterator_type iterate()
  {
    return PFS_buffer_iterator<T, U, V>(this, 0);
  }

  iterator_type iterate(uint index)
  {
    assert(index <= m_max);
    return PFS_buffer_iterator<T, U, V>(this, index);
  }

  void apply(function_type fct)
  {
    value_type *pfs= m_array.get_first();
    value_type *pfs_last= m_array.get_last();

    while (pfs < pfs_last)
    {
      if (pfs->m_lock.is_populated())
      {
        fct(pfs);
      }
      pfs++;
    }
  }

  void apply_all(function_type fct)
  {
    value_type *pfs= m_array.get_first();
    value_type *pfs_last= m_array.get_last();

    while (pfs < pfs_last)
    {
      fct(pfs);
      pfs++;
    }
  }

  void apply(processor_type & proc)
  {
    value_type *pfs= m_array.get_first();
    value_type *pfs_last= m_array.get_last();

    while (pfs < pfs_last)
    {
      if (pfs->m_lock.is_populated())
      {
        proc(pfs);
      }
      pfs++;
    }
  }

  void apply_all(processor_type & proc)
  {
    value_type *pfs= m_array.get_first();
    value_type *pfs_last= m_array.get_last();

    while (pfs < pfs_last)
    {
      proc(pfs);
      pfs++;
    }
  }

  inline value_type* get(uint index)
  {
    assert(index < m_max);

    value_type *pfs= m_array.m_ptr + index;
    if (pfs->m_lock.is_populated())
    {
      return pfs;
    }

    return NULL;
  }

  value_type* get(uint index, bool *has_more)
  {
    if (index >= m_max)
    {
      *has_more= false;
      return NULL;
    }

    *has_more= true;
    return get(index);
  }

  value_type *sanitize(value_type *unsafe)
  {
    intptr offset;
    value_type *pfs= m_array.get_first();
    value_type *pfs_last= m_array.get_last();

    if ((pfs <= unsafe) &&
        (unsafe < pfs_last))
    {
      offset= ((intptr) unsafe - (intptr) pfs) % sizeof(value_type);
      if (offset == 0)
        return unsafe;
    }

    return NULL;
  }

  ulong m_lost;

private:
  value_type* scan_next(uint & index, uint * found_index)
  {
    assert(index <= m_max);

    value_type *pfs_first= m_array.get_first();
    value_type *pfs= pfs_first + index;
    value_type *pfs_last= m_array.get_last();

    while (pfs < pfs_last)
    {
      if (pfs->m_lock.is_populated())
      {
        uint found= pfs - pfs_first;
        *found_index= found;
        index= found + 1;
        return pfs;
      }
      pfs++;
    }

    index= m_max;
    return NULL;
  }

  ulong m_max;
  array_type m_array;
  allocator_type *m_allocator;
};

template <class T,
          int PFS_PAGE_SIZE,
          int PFS_PAGE_COUNT,
          class U = PFS_buffer_default_array<T>,
          class V = PFS_buffer_default_allocator<T> >
class PFS_buffer_scalable_container
{
public:
  friend class PFS_buffer_scalable_iterator<T, PFS_PAGE_SIZE, PFS_PAGE_COUNT, U, V>;

  /**
    Type of elements in the buffer.
    The following attributes are required:
    - pfs_lock m_lock
    - PFS_opaque_container_page *m_page
  */
  typedef T value_type;
  /**
    Type of pages in the buffer.
    The following attributes are required:
    - PFS_opaque_container *m_container
  */
  typedef U array_type;
  typedef V allocator_type;
  /** This container type */
  typedef PFS_buffer_scalable_container<T, PFS_PAGE_SIZE, PFS_PAGE_COUNT, U, V> container_type;
  typedef PFS_buffer_const_iterator<T> const_iterator_type;
  typedef PFS_buffer_scalable_iterator<T, PFS_PAGE_SIZE, PFS_PAGE_COUNT, U, V> iterator_type;
  typedef PFS_buffer_processor<T> processor_type;
  typedef void (*function_type)(value_type *);

  static const size_t MAX_SIZE= PFS_PAGE_SIZE*PFS_PAGE_COUNT;

  PFS_buffer_scalable_container(allocator_type *allocator)
  {
    m_allocator= allocator;
    m_initialized= false;
    m_lost= 0;
  }

  int init(long max_size)
  {
    int i;

    m_initialized= true;
    m_full= true;
    m_max= PFS_PAGE_COUNT * PFS_PAGE_SIZE;
    m_max_page_count= PFS_PAGE_COUNT;
    m_last_page_size= PFS_PAGE_SIZE;
    m_lost= 0;
    m_monotonic.m_u32= 0;
    m_max_page_index.m_u32= 0;

    for (i=0 ; i < PFS_PAGE_COUNT; i++)
    {
      m_pages[i]= NULL;
    }

    if (max_size == 0)
    {
      /* No allocation. */
      m_max_page_count= 0;
    }
    else if (max_size > 0)
    {
      if (max_size % PFS_PAGE_SIZE == 0)
      {
        m_max_page_count= max_size / PFS_PAGE_SIZE;
      }
      else
      {
        m_max_page_count= max_size / PFS_PAGE_SIZE + 1;
        m_last_page_size= max_size % PFS_PAGE_SIZE;
      }
      /* Bounded allocation. */
      m_full= false;

      if (m_max_page_count > PFS_PAGE_COUNT)
      {
        m_max_page_count= PFS_PAGE_COUNT;
        m_last_page_size= PFS_PAGE_SIZE;
      }
    }
    else
    {
      /* max_size = -1 means unbounded allocation */
      m_full= false;
    }

    assert(m_max_page_count <= PFS_PAGE_COUNT);
    assert(0 < m_last_page_size);
    assert(m_last_page_size <= PFS_PAGE_SIZE);

    pthread_mutex_init(& m_critical_section, NULL);
    return 0;
  }

  void cleanup()
  {
    int i;
    array_type *page;

    if (! m_initialized)
      return;

    pthread_mutex_lock(& m_critical_section);

    for (i=0 ; i < PFS_PAGE_COUNT; i++)
    {
      page= m_pages[i];
      if (page != NULL)
      {
        m_allocator->free_array(page);
        delete page;
        m_pages[i]= NULL;
      }
    }
    pthread_mutex_unlock(& m_critical_section);

    pthread_mutex_destroy(& m_critical_section);

    m_initialized= false;
  }

  ulong get_row_count()
  {
    ulong page_count= PFS_atomic::load_u32(& m_max_page_index.m_u32);

    return page_count * PFS_PAGE_SIZE;
  }

  ulong get_row_size() const
  {
    return sizeof(value_type);
  }

  ulong get_memory()
  {
    return get_row_count() * get_row_size();
  }

  value_type *allocate(pfs_dirty_state *dirty_state)
  {
    if (m_full)
    {
      m_lost++;
      return NULL;
    }

    uint index;
    uint monotonic;
    uint monotonic_max;
    uint current_page_count;
    value_type *pfs;
    array_type *array;

    void *addr;
    void * volatile * typed_addr;
    void *ptr;

    /*
      1: Try to find an available record within the existing pages
    */
    current_page_count= PFS_atomic::load_u32(& m_max_page_index.m_u32);

    if (current_page_count != 0)
    {
      monotonic= PFS_atomic::load_u32(& m_monotonic.m_u32);
      monotonic_max= monotonic + current_page_count;

      while (monotonic < monotonic_max)
      {
        /*
          Scan in the [0 .. current_page_count - 1] range,
          in parallel with m_monotonic (see below)
        */
        index= monotonic % current_page_count;

        /* Atomic Load, array= m_pages[index] */
        addr= & m_pages[index];
        typed_addr= static_cast<void * volatile *>(addr);
        ptr= my_atomic_loadptr(typed_addr);
        array= static_cast<array_type *>(ptr);

        if (array != NULL)
        {
          pfs= array->allocate(dirty_state);
          if (pfs != NULL)
          {
            /* Keep a pointer to the parent page, for deallocate(). */
            pfs->m_page= reinterpret_cast<PFS_opaque_container_page *> (array);
            return pfs;
          }
        }

        /*
          Parallel scans collaborate to increase
          the common monotonic scan counter.

          Note that when all the existing page are full,
          one thread will eventually add a new page,
          and cause m_max_page_index to increase,
          which fools all the modulo logic for scans already in progress,
          because the monotonic counter is not folded to the same place
          (sometime modulo N, sometime modulo N+1).

          This is actually ok: since all the pages are full anyway,
          there is nothing to miss, so better increase the monotonic
          counter faster and then move on to the detection of new pages,
          in part 2: below.
        */
        monotonic= PFS_atomic::add_u32(& m_monotonic.m_u32, 1);
      };
    }

    /*
      2: Try to add a new page, beyond the m_max_page_index limit
    */
    while (current_page_count < m_max_page_count)
    {
      /* Peek for pages added by collaborating threads */

      /* (2-a) Atomic Load, array= m_pages[current_page_count] */
      addr= & m_pages[current_page_count];
      typed_addr= static_cast<void * volatile *>(addr);
      ptr= my_atomic_loadptr(typed_addr);
      array= static_cast<array_type *>(ptr);

      if (array == NULL)
      {
        // ==================================================================
        // BEGIN CRITICAL SECTION -- buffer expand
        // ==================================================================

        /*
          On a fresh started server, buffers are typically empty.
          When a sudden load spike is seen by the server,
          multiple threads may want to expand the buffer at the same time.

          Using a compare and swap to allow multiple pages to be added,
          possibly freeing duplicate pages on collisions,
          does not work well because the amount of code involved
          when creating a new page can be significant (PFS_thread),
          causing MANY collisions between (2-b) and (2-d).

          A huge number of collisions (which can happen when thousands
          of new connections hits the server after a restart)
          leads to a huge memory consumption, and to OOM.

          To mitigate this, we use here a mutex,
          to enforce that only ONE page is added at a time,
          so that scaling the buffer happens in a predictable
          and controlled manner.
        */
        pthread_mutex_lock(& m_critical_section);

        /*
          Peek again for pages added by collaborating threads,
          this time as the only thread allowed to expand the buffer
        */

        /* (2-b) Atomic Load, array= m_pages[current_page_count] */

        ptr= my_atomic_loadptr(typed_addr);
        array= static_cast<array_type *>(ptr);

        if (array == NULL)
        {
          /* (2-c) Found no page, allocate a new one */
          array= new array_type();
          builtin_memory_scalable_buffer.count_alloc(sizeof (array_type));

          array->m_max= get_page_logical_size(current_page_count);
          int rc= m_allocator->alloc_array(array);
          if (rc != 0)
          {
            m_allocator->free_array(array);
            delete array;
            builtin_memory_scalable_buffer.count_free(sizeof (array_type));
            m_lost++;
            pthread_mutex_unlock(& m_critical_section);
            return NULL;
          }

          /* Keep a pointer to this container, for static_deallocate(). */
          array->m_container= reinterpret_cast<PFS_opaque_container *> (this);

          /* (2-d) Atomic STORE, m_pages[current_page_count] = array  */
          ptr= array;
          my_atomic_storeptr(typed_addr, ptr);

          /* Advertise the new page */
          PFS_atomic::add_u32(& m_max_page_index.m_u32, 1);
        }

        pthread_mutex_unlock(& m_critical_section);

        // ==================================================================
        // END CRITICAL SECTION -- buffer expand
        // ==================================================================
      }

      assert(array != NULL);
      pfs= array->allocate(dirty_state);
      if (pfs != NULL)
      {
        /* Keep a pointer to the parent page, for deallocate(). */
        pfs->m_page= reinterpret_cast<PFS_opaque_container_page *> (array);
        return pfs;
      }

      current_page_count++;
    }

    m_lost++;
    m_full= true;
    return NULL;
  }

  void deallocate(value_type *safe_pfs)
  {
    /* Find the containing page */
    PFS_opaque_container_page *opaque_page= safe_pfs->m_page;
    array_type *page= reinterpret_cast<array_type *> (opaque_page);

    /* Mark the object free */
    safe_pfs->m_lock.allocated_to_free();

    /* Flag the containing page as not full. */
    page->m_full= false;

    /* Flag the overall container as not full. */
    m_full= false;
  }

  static void static_deallocate(value_type *safe_pfs)
  {
    /* Find the containing page */
    PFS_opaque_container_page *opaque_page= safe_pfs->m_page;
    array_type *page= reinterpret_cast<array_type *> (opaque_page);

    /* Mark the object free */
    safe_pfs->m_lock.allocated_to_free();

    /* Flag the containing page as not full. */
    page->m_full= false;

    /* Find the containing buffer */
    PFS_opaque_container *opaque_container= page->m_container;
    PFS_buffer_scalable_container *container;
    container= reinterpret_cast<container_type *> (opaque_container);

    /* Flag the overall container as not full. */
    container->m_full= false;
  }

  iterator_type iterate()
  {
    return PFS_buffer_scalable_iterator<T, PFS_PAGE_SIZE, PFS_PAGE_COUNT, U, V>(this, 0);
  }

  iterator_type iterate(uint index)
  {
    assert(index <= m_max);
    return PFS_buffer_scalable_iterator<T, PFS_PAGE_SIZE, PFS_PAGE_COUNT, U, V>(this, index);
  }

  void apply(function_type fct)
  {
    uint i;
    array_type *page;
    value_type *pfs;
    value_type *pfs_last;

    for (i=0 ; i < PFS_PAGE_COUNT; i++)
    {
      page= m_pages[i];
      if (page != NULL)
      {
        pfs= page->get_first();
        pfs_last= page->get_last();

        while (pfs < pfs_last)
        {
          if (pfs->m_lock.is_populated())
          {
            fct(pfs);
          }
          pfs++;
        }
      }
    }
  }

  void apply_all(function_type fct)
  {
    uint i;
    array_type *page;
    value_type *pfs;
    value_type *pfs_last;

    for (i=0 ; i < PFS_PAGE_COUNT; i++)
    {
      page= m_pages[i];
      if (page != NULL)
      {
        pfs= page->get_first();
        pfs_last= page->get_last();

        while (pfs < pfs_last)
        {
          fct(pfs);
          pfs++;
        }
      }
    }
  }

  void apply(processor_type & proc)
  {
    uint i;
    array_type *page;
    value_type *pfs;
    value_type *pfs_last;

    for (i=0 ; i < PFS_PAGE_COUNT; i++)
    {
      page= m_pages[i];
      if (page != NULL)
      {
        pfs= page->get_first();
        pfs_last= page->get_last();

        while (pfs < pfs_last)
        {
          if (pfs->m_lock.is_populated())
          {
            proc(pfs);
          }
          pfs++;
        }
      }
    }
  }

  void apply_all(processor_type & proc)
  {
    uint i;
    array_type *page;
    value_type *pfs;
    value_type *pfs_last;

    for (i=0 ; i < PFS_PAGE_COUNT; i++)
    {
      page= m_pages[i];
      if (page != NULL)
      {
        pfs= page->get_first();
        pfs_last= page->get_last();

        while (pfs < pfs_last)
        {
          proc(pfs);
          pfs++;
        }
      }
    }
  }

  value_type* get(uint index)
  {
    assert(index < m_max);

    uint index_1= index / PFS_PAGE_SIZE;
    array_type *page= m_pages[index_1];
    if (page != NULL)
    {
      uint index_2= index % PFS_PAGE_SIZE;

      if (index_2 >= page->m_max)
      {
        return NULL;
      }

      value_type *pfs= page->m_ptr + index_2;

      if (pfs->m_lock.is_populated())
      {
        return pfs;
      }
    }

    return NULL;
  }

  value_type* get(uint index, bool *has_more)
  {
    if (index >= m_max)
    {
      *has_more= false;
      return NULL;
    }

    uint index_1= index / PFS_PAGE_SIZE;
    array_type *page= m_pages[index_1];

    if (page == NULL)
    {
      *has_more= false;
      return NULL;
    }

    uint index_2= index % PFS_PAGE_SIZE;

    if (index_2 >= page->m_max)
    {
      *has_more= false;
      return NULL;
    }

    *has_more= true;
    value_type *pfs= page->m_ptr + index_2;

    if (pfs->m_lock.is_populated())
    {
      return pfs;
    }

    return NULL;
  }

  value_type *sanitize(value_type *unsafe)
  {
    intptr offset;
    uint i;
    array_type *page;
    value_type *pfs;
    value_type *pfs_last;

    for (i=0 ; i < PFS_PAGE_COUNT; i++)
    {
      page= m_pages[i];
      if (page != NULL)
      {
        pfs= page->get_first();
        pfs_last= page->get_last();

        if ((pfs <= unsafe) &&
            (unsafe < pfs_last))
        {
          offset= ((intptr) unsafe - (intptr) pfs) % sizeof(value_type);
          if (offset == 0)
            return unsafe;
        }
      }
    }

    return NULL;
  }

  ulong m_lost;

private:

  uint get_page_logical_size(uint page_index)
  {
    if (page_index + 1 < m_max_page_count)
      return PFS_PAGE_SIZE;
    assert(page_index + 1 == m_max_page_count);
    return m_last_page_size;
  }

  value_type* scan_next(uint & index, uint * found_index)
  {
    assert(index <= m_max);

    uint index_1= index / PFS_PAGE_SIZE;
    uint index_2= index % PFS_PAGE_SIZE;
    array_type *page;
    value_type *pfs_first;
    value_type *pfs;
    value_type *pfs_last;

    while (index_1 < PFS_PAGE_COUNT)
    {
      page= m_pages[index_1];

      if (page == NULL)
      {
        index= static_cast<uint>(m_max);
        return NULL;
      }

      pfs_first= page->get_first();
      pfs= pfs_first + index_2;
      pfs_last= page->get_last();

      while (pfs < pfs_last)
      {
        if (pfs->m_lock.is_populated())
        {
          uint found= index_1 * PFS_PAGE_SIZE + static_cast<uint>(pfs - pfs_first);
          *found_index= found;
          index= found + 1;
          return pfs;
        }
        pfs++;
      }

      index_1++;
      index_2= 0;
    }

    index= static_cast<uint>(m_max);
    return NULL;
  }

  bool m_initialized;
  bool m_full;
  size_t m_max;
  PFS_cacheline_uint32 m_monotonic;
  PFS_cacheline_uint32 m_max_page_index;
  ulong m_max_page_count;
  ulong m_last_page_size;
  array_type * m_pages[PFS_PAGE_COUNT];
  allocator_type *m_allocator;
  pthread_mutex_t m_critical_section;
};

template <class T, class U, class V>
class PFS_buffer_iterator
{
  friend class PFS_buffer_container<T, U, V>;

  typedef T value_type;
  typedef PFS_buffer_container<T, U, V> container_type;

public:
  value_type* scan_next()
  {
    uint unused;
    return m_container->scan_next(m_index, & unused);
  }

  value_type* scan_next(uint * found_index)
  {
    return m_container->scan_next(m_index, found_index);
  }

private:
  PFS_buffer_iterator(container_type *container, uint index)
    : m_container(container),
      m_index(index)
  {}

  container_type *m_container;
  uint m_index;
};

template <class T, int page_size, int page_count, class U, class V>
class PFS_buffer_scalable_iterator
{
  friend class PFS_buffer_scalable_container<T, page_size, page_count, U, V>;

  typedef T value_type;
  typedef PFS_buffer_scalable_container<T, page_size, page_count, U, V> container_type;

public:
  value_type* scan_next()
  {
    uint unused;
    return m_container->scan_next(m_index, & unused);
  }

  value_type* scan_next(uint * found_index)
  {
    return m_container->scan_next(m_index, found_index);
  }

private:
  PFS_buffer_scalable_iterator(container_type *container, uint index)
    : m_container(container),
      m_index(index)
  {}

  container_type *m_container;
  uint m_index;
};

template <class T>
class PFS_buffer_processor
{
public:
  virtual ~PFS_buffer_processor()= default;
  virtual void operator()(T *element) = 0;
};

template <class B, int PFS_PARTITION_COUNT>
class PFS_partitioned_buffer_scalable_container
{
public:
  friend class PFS_partitioned_buffer_scalable_iterator<B, PFS_PARTITION_COUNT>;

  typedef typename B::value_type value_type;
  typedef typename B::allocator_type allocator_type;
  typedef PFS_partitioned_buffer_scalable_iterator<B, PFS_PARTITION_COUNT> iterator_type;
  typedef typename B::iterator_type sub_iterator_type;
  typedef typename B::processor_type processor_type;
  typedef typename B::function_type function_type;

  PFS_partitioned_buffer_scalable_container(allocator_type *allocator)
  {
    for (int i=0 ; i < PFS_PARTITION_COUNT; i++)
    {
      m_partitions[i]= new B(allocator);
    }
  }

  ~PFS_partitioned_buffer_scalable_container()
  {
    for (int i=0 ; i < PFS_PARTITION_COUNT; i++)
    {
      delete m_partitions[i];
    }
  }

  int init(long max_size)
  {
    int rc= 0;
    // FIXME: we have max_size * PFS_PARTITION_COUNT here
    for (int i=0 ; i < PFS_PARTITION_COUNT; i++)
    {
      rc|= m_partitions[i]->init(max_size);
    }
    return rc;
  }

  void cleanup()
  {
    for (int i=0 ; i < PFS_PARTITION_COUNT; i++)
    {
      m_partitions[i]->cleanup();
    }
  }

  ulong get_row_count() const
  {
    ulong sum= 0;

    for (int i=0; i < PFS_PARTITION_COUNT; i++)
    {
      sum += m_partitions[i]->get_row_count();
    }

    return sum;
  }

  ulong get_row_size() const
  {
    return sizeof(value_type);
  }

  ulong get_memory() const
  {
    ulong sum= 0;

    for (int i=0; i < PFS_PARTITION_COUNT; i++)
    {
      sum += m_partitions[i]->get_memory();
    }

    return sum;
  }

  long get_lost_counter()
  {
    long sum= 0;

    for (int i=0; i < PFS_PARTITION_COUNT; i++)
    {
      sum += m_partitions[i]->m_lost;
    }

    return sum;
  }

  value_type *allocate(pfs_dirty_state *dirty_state, uint partition)
  {
    assert(partition < PFS_PARTITION_COUNT);

    return m_partitions[partition]->allocate(dirty_state);
  }

  void deallocate(value_type *safe_pfs)
  {
    /*
      One issue here is that we do not know which partition
      the record belongs to.
      Each record points to the parent page,
      and each page points to the parent buffer,
      so using static_deallocate here,
      which will find the correct partition by itself.
    */
    B::static_deallocate(safe_pfs);
  }

  iterator_type iterate()
  {
    return iterator_type(this, 0, 0);
  }

  iterator_type iterate(uint user_index)
  {
    uint partition_index;
    uint sub_index;
    unpack_index(user_index, &partition_index, &sub_index);
    return iterator_type(this, partition_index, sub_index);
  }

  void apply(function_type fct)
  {
    for (int i=0; i < PFS_PARTITION_COUNT; i++)
    {
      m_partitions[i]->apply(fct);
    }
  }

  void apply_all(function_type fct)
  {
    for (int i=0; i < PFS_PARTITION_COUNT; i++)
    {
      m_partitions[i]->apply_all(fct);
    }
  }

  void apply(processor_type & proc)
  {
    for (int i=0; i < PFS_PARTITION_COUNT; i++)
    {
      m_partitions[i]->apply(proc);
    }
  }

  void apply_all(processor_type & proc)
  {
    for (int i=0; i < PFS_PARTITION_COUNT; i++)
    {
      m_partitions[i]->apply_all(proc);
    }
  }

  value_type* get(uint user_index)
  {
    uint partition_index;
    uint sub_index;
    unpack_index(user_index, &partition_index, &sub_index);

    if (partition_index >= PFS_PARTITION_COUNT)
    {
      return NULL;
    }

    return m_partitions[partition_index]->get(sub_index);
  }

  value_type* get(uint user_index, bool *has_more)
  {
    uint partition_index;
    uint sub_index;
    unpack_index(user_index, &partition_index, &sub_index);

    if (partition_index >= PFS_PARTITION_COUNT)
    {
      *has_more= false;
      return NULL;
    }

    *has_more= true;
    return m_partitions[partition_index]->get(sub_index);
  }

  value_type *sanitize(value_type *unsafe)
  {
    value_type *safe= NULL;

    for (int i=0; i < PFS_PARTITION_COUNT; i++)
    {
      safe= m_partitions[i]->sanitize(unsafe);
      if (safe != NULL)
      {
        return safe;
      }
    }

    return safe;
  }

private:
  static void pack_index(uint partition_index, uint sub_index, uint *user_index)
  {
    /* 2^8 = 256 partitions max */
    compile_time_assert(PFS_PARTITION_COUNT <= (1 << 8));
    /* 2^24 = 16777216 max per partitioned buffer. */
    compile_time_assert((B::MAX_SIZE) <= (1 << 24));

    *user_index= (partition_index << 24) + sub_index;
  }

  static void unpack_index(uint user_index, uint *partition_index, uint *sub_index)
  {
    *partition_index= user_index >> 24;
    *sub_index= user_index & 0x00FFFFFF;
  }

  value_type* scan_next(uint & partition_index, uint & sub_index, uint * found_partition, uint * found_sub_index)
  {
    value_type *record= NULL;
    assert(partition_index < PFS_PARTITION_COUNT);

    while (partition_index < PFS_PARTITION_COUNT)
    {
      sub_iterator_type sub_iterator= m_partitions[partition_index]->iterate(sub_index);
      record= sub_iterator.scan_next(found_sub_index);
      if (record != NULL)
      {
        *found_partition= partition_index;
        sub_index= *found_sub_index + 1;
        return record;
      }

      partition_index++;
      sub_index= 0;
    }

    *found_partition= PFS_PARTITION_COUNT;
    *found_sub_index= 0;
    sub_index= 0;
    return NULL;
  }

  B *m_partitions[PFS_PARTITION_COUNT];
};

template <class B, int PFS_PARTITION_COUNT>
class PFS_partitioned_buffer_scalable_iterator
{
public:
  friend class PFS_partitioned_buffer_scalable_container<B, PFS_PARTITION_COUNT>;

  typedef typename B::value_type value_type;
  typedef PFS_partitioned_buffer_scalable_container<B, PFS_PARTITION_COUNT> container_type;

  value_type* scan_next()
  {
    uint unused_partition;
    uint unused_sub_index;
    return m_container->scan_next(m_partition, m_sub_index, & unused_partition, & unused_sub_index);
  }

  value_type* scan_next(uint *found_user_index)
  {
    uint found_partition;
    uint found_sub_index;
    value_type *record;
    record=  m_container->scan_next(m_partition, m_sub_index, &found_partition, &found_sub_index);
    container_type::pack_index(found_partition, found_sub_index, found_user_index);
    return record;
  }

private:
  PFS_partitioned_buffer_scalable_iterator(container_type *container, uint partition, uint sub_index)
    : m_container(container),
      m_partition(partition),
      m_sub_index(sub_index)
  {}

  container_type *m_container;
  uint m_partition;
  uint m_sub_index;
};

#ifdef USE_SCALABLE
typedef PFS_buffer_scalable_container<PFS_mutex, 1024, 1024> PFS_mutex_basic_container;
typedef PFS_partitioned_buffer_scalable_container<PFS_mutex_basic_container, PSI_COUNT_VOLATILITY> PFS_mutex_container;
#else
typedef PFS_buffer_container<PFS_mutex> PFS_mutex_container;
#endif
typedef PFS_mutex_container::iterator_type PFS_mutex_iterator;
extern PFS_mutex_container global_mutex_container;

#ifdef USE_SCALABLE
typedef PFS_buffer_scalable_container<PFS_rwlock, 1024, 1024> PFS_rwlock_container;
#else
typedef PFS_buffer_container<PFS_rwlock> PFS_rwlock_container;
#endif
typedef PFS_rwlock_container::iterator_type PFS_rwlock_iterator;
extern PFS_rwlock_container global_rwlock_container;

#ifdef USE_SCALABLE
typedef PFS_buffer_scalable_container<PFS_cond, 256, 256> PFS_cond_container;
#else
typedef PFS_buffer_container<PFS_cond> PFS_cond_container;
#endif
typedef PFS_cond_container::iterator_type PFS_cond_iterator;
extern PFS_cond_container global_cond_container;

#ifdef USE_SCALABLE
typedef PFS_buffer_scalable_container<PFS_file, 4 * 1024, 4 * 1024> PFS_file_container;
#else
typedef PFS_buffer_container<PFS_file> PFS_file_container;
#endif
typedef PFS_file_container::iterator_type PFS_file_iterator;
extern PFS_file_container global_file_container;

#ifdef USE_SCALABLE
typedef PFS_buffer_scalable_container<PFS_socket, 256, 256> PFS_socket_container;
#else
typedef PFS_buffer_container<PFS_socket> PFS_socket_container;
#endif
typedef PFS_socket_container::iterator_type PFS_socket_iterator;
extern PFS_socket_container global_socket_container;

#ifdef USE_SCALABLE
typedef PFS_buffer_scalable_container<PFS_metadata_lock, 1024, 1024> PFS_mdl_container;
#else
typedef PFS_buffer_container<PFS_metadata_lock> PFS_mdl_container;
#endif
typedef PFS_mdl_container::iterator_type PFS_mdl_iterator;
extern PFS_mdl_container global_mdl_container;

#ifdef USE_SCALABLE
typedef PFS_buffer_scalable_container<PFS_setup_actor, 128, 1024> PFS_setup_actor_container;
#else
typedef PFS_buffer_container<PFS_setup_actor> PFS_setup_actor_container;
#endif
typedef PFS_setup_actor_container::iterator_type PFS_setup_actor_iterator;
extern PFS_setup_actor_container global_setup_actor_container;

#ifdef USE_SCALABLE
typedef PFS_buffer_scalable_container<PFS_setup_object, 128, 1024> PFS_setup_object_container;
#else
typedef PFS_buffer_container<PFS_setup_object> PFS_setup_object_container;
#endif
typedef PFS_setup_object_container::iterator_type PFS_setup_object_iterator;
extern PFS_setup_object_container global_setup_object_container;

#ifdef USE_SCALABLE
typedef PFS_buffer_scalable_container<PFS_table, 1024, 1024> PFS_table_container;
#else
typedef PFS_buffer_container<PFS_table> PFS_table_container;
#endif
typedef PFS_table_container::iterator_type PFS_table_iterator;
extern PFS_table_container global_table_container;

#ifdef USE_SCALABLE
typedef PFS_buffer_scalable_container<PFS_table_share, 4 * 1024, 4 * 1024> PFS_table_share_container;
#else
typedef PFS_buffer_container<PFS_table_share> PFS_table_share_container;
#endif
typedef PFS_table_share_container::iterator_type PFS_table_share_iterator;
extern PFS_table_share_container global_table_share_container;

#ifdef USE_SCALABLE
typedef PFS_buffer_scalable_container<PFS_table_share_index, 8 * 1024, 8 * 1024> PFS_table_share_index_container;
#else
typedef PFS_buffer_container<PFS_table_share_index> PFS_table_share_index_container;
#endif
typedef PFS_table_share_index_container::iterator_type PFS_table_share_index_iterator;
extern PFS_table_share_index_container global_table_share_index_container;

#ifdef USE_SCALABLE
typedef PFS_buffer_scalable_container<PFS_table_share_lock, 4 * 1024, 4 * 1024> PFS_table_share_lock_container;
#else
typedef PFS_buffer_container<PFS_table_share_lock> PFS_table_share_lock_container;
#endif
typedef PFS_table_share_lock_container::iterator_type PFS_table_share_lock_iterator;
extern PFS_table_share_lock_container global_table_share_lock_container;

#ifdef USE_SCALABLE
typedef PFS_buffer_scalable_container<PFS_program, 1024, 1024> PFS_program_container;
#else
typedef PFS_buffer_container<PFS_program> PFS_program_container;
#endif
typedef PFS_program_container::iterator_type PFS_program_iterator;
extern PFS_program_container global_program_container;

#ifdef USE_SCALABLE
typedef PFS_buffer_scalable_container<PFS_prepared_stmt, 1024, 1024> PFS_prepared_stmt_container;
#else
typedef PFS_buffer_container<PFS_prepared_stmt> PFS_prepared_stmt_container;
#endif
typedef PFS_prepared_stmt_container::iterator_type PFS_prepared_stmt_iterator;
extern PFS_prepared_stmt_container global_prepared_stmt_container;

class PFS_account_array : public PFS_buffer_default_array<PFS_account>
{
public:
  PFS_single_stat *m_instr_class_waits_array;
  PFS_stage_stat *m_instr_class_stages_array;
  PFS_statement_stat *m_instr_class_statements_array;
  PFS_transaction_stat *m_instr_class_transactions_array;
  PFS_memory_stat *m_instr_class_memory_array;
};

class PFS_account_allocator
{
public:
  int alloc_array(PFS_account_array *array);
  void free_array(PFS_account_array *array);
};

#ifdef USE_SCALABLE
typedef PFS_buffer_scalable_container<PFS_account,
                                      128,
                                      128,
                                      PFS_account_array,
                                      PFS_account_allocator> PFS_account_container;
#else
typedef PFS_buffer_container<PFS_account,
                             PFS_account_array,
                             PFS_account_allocator> PFS_account_container;
#endif
typedef PFS_account_container::iterator_type PFS_account_iterator;
extern PFS_account_container global_account_container;

class PFS_host_array : public PFS_buffer_default_array<PFS_host>
{
public:
  PFS_single_stat *m_instr_class_waits_array;
  PFS_stage_stat *m_instr_class_stages_array;
  PFS_statement_stat *m_instr_class_statements_array;
  PFS_transaction_stat *m_instr_class_transactions_array;
  PFS_memory_stat *m_instr_class_memory_array;
};

class PFS_host_allocator
{
public:
  int alloc_array(PFS_host_array *array);
  void free_array(PFS_host_array *array);
};

#ifdef USE_SCALABLE
typedef PFS_buffer_scalable_container<PFS_host,
                                      128,
                                      128,
                                      PFS_host_array,
                                      PFS_host_allocator> PFS_host_container;
#else
typedef PFS_buffer_container<PFS_host,
                             PFS_host_array,
                             PFS_host_allocator> PFS_host_container;
#endif
typedef PFS_host_container::iterator_type PFS_host_iterator;
extern PFS_host_container global_host_container;

class PFS_thread_array : public PFS_buffer_default_array<PFS_thread>
{
public:
  PFS_single_stat *m_instr_class_waits_array;
  PFS_stage_stat *m_instr_class_stages_array;
  PFS_statement_stat *m_instr_class_statements_array;
  PFS_transaction_stat *m_instr_class_transactions_array;
  PFS_memory_stat *m_instr_class_memory_array;

  PFS_events_waits *m_waits_history_array;
  PFS_events_stages *m_stages_history_array;
  PFS_events_statements *m_statements_history_array;
  PFS_events_statements *m_statements_stack_array;
  PFS_events_transactions *m_transactions_history_array;
  char *m_session_connect_attrs_array;

  char *m_current_stmts_text_array;
  char *m_history_stmts_text_array;
  unsigned char *m_current_stmts_digest_token_array;
  unsigned char *m_history_stmts_digest_token_array;
};

class PFS_thread_allocator
{
public:
  int alloc_array(PFS_thread_array *array);
  void free_array(PFS_thread_array *array);
};

#ifdef USE_SCALABLE
typedef PFS_buffer_scalable_container<PFS_thread,
                                      256,
                                      256,
                                      PFS_thread_array,
                                      PFS_thread_allocator> PFS_thread_container;
#else
typedef PFS_buffer_container<PFS_thread,
                             PFS_thread_array,
                             PFS_thread_allocator> PFS_thread_container;
#endif
typedef PFS_thread_container::iterator_type PFS_thread_iterator;
extern PFS_thread_container global_thread_container;

class PFS_user_array : public PFS_buffer_default_array<PFS_user>
{
public:
  PFS_single_stat *m_instr_class_waits_array;
  PFS_stage_stat *m_instr_class_stages_array;
  PFS_statement_stat *m_instr_class_statements_array;
  PFS_transaction_stat *m_instr_class_transactions_array;
  PFS_memory_stat *m_instr_class_memory_array;
};

class PFS_user_allocator
{
public:
  int alloc_array(PFS_user_array *array);
  void free_array(PFS_user_array *array);
};

#ifdef USE_SCALABLE
typedef PFS_buffer_scalable_container<PFS_user,
                                      128,
                                      128,
                                      PFS_user_array,
                                      PFS_user_allocator> PFS_user_container;
#else
typedef PFS_buffer_container<PFS_user,
                             PFS_user_array,
                             PFS_user_allocator> PFS_user_container;
#endif
typedef PFS_user_container::iterator_type PFS_user_iterator;
extern PFS_user_container global_user_container;

#endif