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
|
/******************************************************
Transaction undo log record
(c) 1996 Innobase Oy
Created 3/26/1996 Heikki Tuuri
*******************************************************/
#include "trx0rec.h"
#ifdef UNIV_NONINL
#include "trx0rec.ic"
#endif
#include "fsp0fsp.h"
#include "mach0data.h"
#include "trx0rseg.h"
#include "trx0trx.h"
#include "trx0undo.h"
#include "dict0dict.h"
#include "ut0mem.h"
#include "row0upd.h"
#include "que0que.h"
#include "trx0purge.h"
#include "row0row.h"
/*=========== UNDO LOG RECORD CREATION AND DECODING ====================*/
/**************************************************************************
Writes the mtr log entry of the inserted undo log record on the undo log
page. */
UNIV_INLINE
void
trx_undof_page_add_undo_rec_log(
/*============================*/
page_t* undo_page, /* in: undo log page */
ulint old_free, /* in: start offset of the inserted entry */
ulint new_free, /* in: end offset of the entry */
mtr_t* mtr) /* in: mtr */
{
byte* log_ptr;
const byte* log_end;
ulint len;
log_ptr = mlog_open(mtr, 11 + 13 + MLOG_BUF_MARGIN);
if (log_ptr == NULL) {
return;
}
log_end = &log_ptr[11 + 13 + MLOG_BUF_MARGIN];
log_ptr = mlog_write_initial_log_record_fast(undo_page,
MLOG_UNDO_INSERT, log_ptr, mtr);
len = new_free - old_free - 4;
mach_write_to_2(log_ptr, len);
log_ptr += 2;
if (log_ptr + len <= log_end) {
memcpy(log_ptr, undo_page + old_free + 2, len);
mlog_close(mtr, log_ptr + len);
} else {
mlog_close(mtr, log_ptr);
mlog_catenate_string(mtr, undo_page + old_free + 2, len);
}
}
/***************************************************************
Parses a redo log record of adding an undo log record. */
byte*
trx_undo_parse_add_undo_rec(
/*========================*/
/* out: end of log record or NULL */
byte* ptr, /* in: buffer */
byte* end_ptr,/* in: buffer end */
page_t* page) /* in: page or NULL */
{
ulint len;
byte* rec;
ulint first_free;
if (end_ptr < ptr + 2) {
return(NULL);
}
len = mach_read_from_2(ptr);
ptr += 2;
if (end_ptr < ptr + len) {
return(NULL);
}
if (page == NULL) {
return(ptr + len);
}
first_free = mach_read_from_2(page + TRX_UNDO_PAGE_HDR
+ TRX_UNDO_PAGE_FREE);
rec = page + first_free;
mach_write_to_2(rec, first_free + 4 + len);
mach_write_to_2(rec + 2 + len, first_free);
mach_write_to_2(page + TRX_UNDO_PAGE_HDR + TRX_UNDO_PAGE_FREE,
first_free + 4 + len);
ut_memcpy(rec + 2, ptr, len);
return(ptr + len);
}
/**************************************************************************
Calculates the free space left for extending an undo log record. */
UNIV_INLINE
ulint
trx_undo_left(
/*==========*/
/* out: bytes left */
page_t* page, /* in: undo log page */
byte* ptr) /* in: pointer to page */
{
/* The '- 10' is a safety margin, in case we have some small
calculation error below */
return(UNIV_PAGE_SIZE - (ptr - page) - 10 - FIL_PAGE_DATA_END);
}
/**************************************************************************
Reports in the undo log of an insert of a clustered index record. */
static
ulint
trx_undo_page_report_insert(
/*========================*/
/* out: offset of the inserted entry
on the page if succeed, 0 if fail */
page_t* undo_page, /* in: undo log page */
trx_t* trx, /* in: transaction */
dict_index_t* index, /* in: clustered index */
dtuple_t* clust_entry, /* in: index entry which will be
inserted to the clustered index */
mtr_t* mtr) /* in: mtr */
{
ulint first_free;
byte* ptr;
ulint len;
dfield_t* field;
ulint flen;
ulint i;
ut_ad(mach_read_from_2(undo_page + TRX_UNDO_PAGE_HDR
+ TRX_UNDO_PAGE_TYPE) == TRX_UNDO_INSERT);
first_free = mach_read_from_2(undo_page + TRX_UNDO_PAGE_HDR
+ TRX_UNDO_PAGE_FREE);
ptr = undo_page + first_free;
ut_ad(first_free <= UNIV_PAGE_SIZE);
if (trx_undo_left(undo_page, ptr) < 30) {
/* NOTE: the value 30 must be big enough such that the general
fields written below fit on the undo log page */
return(0);
}
/* Reserve 2 bytes for the pointer to the next undo log record */
ptr += 2;
/* Store first some general parameters to the undo log */
mach_write_to_1(ptr, TRX_UNDO_INSERT_REC);
ptr++;
len = mach_dulint_write_much_compressed(ptr, trx->undo_no);
ptr += len;
len = mach_dulint_write_much_compressed(ptr, (index->table)->id);
ptr += len;
/*----------------------------------------*/
/* Store then the fields required to uniquely determine the record
to be inserted in the clustered index */
for (i = 0; i < dict_index_get_n_unique(index); i++) {
field = dtuple_get_nth_field(clust_entry, i);
flen = dfield_get_len(field);
if (trx_undo_left(undo_page, ptr) < 5) {
return(0);
}
len = mach_write_compressed(ptr, flen);
ptr += len;
if (flen != UNIV_SQL_NULL) {
if (trx_undo_left(undo_page, ptr) < flen) {
return(0);
}
ut_memcpy(ptr, dfield_get_data(field), flen);
ptr += flen;
}
}
if (trx_undo_left(undo_page, ptr) < 2) {
return(0);
}
/*----------------------------------------*/
/* Write pointers to the previous and the next undo log records */
if (trx_undo_left(undo_page, ptr) < 2) {
return(0);
}
mach_write_to_2(ptr, first_free);
ptr += 2;
mach_write_to_2(undo_page + first_free, ptr - undo_page);
mach_write_to_2(undo_page + TRX_UNDO_PAGE_HDR + TRX_UNDO_PAGE_FREE,
ptr - undo_page);
/* Write the log entry to the REDO log of this change in the UNDO
log */
trx_undof_page_add_undo_rec_log(undo_page, first_free,
ptr - undo_page, mtr);
return(first_free);
}
/**************************************************************************
Reads from an undo log record the general parameters. */
byte*
trx_undo_rec_get_pars(
/*==================*/
/* out: remaining part of undo log
record after reading these values */
trx_undo_rec_t* undo_rec, /* in: undo log record */
ulint* type, /* out: undo record type:
TRX_UNDO_INSERT_REC, ... */
ulint* cmpl_info, /* out: compiler info, relevant only
for update type records */
ibool* updated_extern, /* out: TRUE if we updated an
externally stored fild */
dulint* undo_no, /* out: undo log record number */
dulint* table_id) /* out: table id */
{
byte* ptr;
ulint len;
ulint type_cmpl;
ptr = undo_rec + 2;
type_cmpl = mach_read_from_1(ptr);
ptr++;
if (type_cmpl & TRX_UNDO_UPD_EXTERN) {
*updated_extern = TRUE;
type_cmpl -= TRX_UNDO_UPD_EXTERN;
} else {
*updated_extern = FALSE;
}
*type = type_cmpl & (TRX_UNDO_CMPL_INFO_MULT - 1);
*cmpl_info = type_cmpl / TRX_UNDO_CMPL_INFO_MULT;
*undo_no = mach_dulint_read_much_compressed(ptr);
len = mach_dulint_get_much_compressed_size(*undo_no);
ptr += len;
*table_id = mach_dulint_read_much_compressed(ptr);
len = mach_dulint_get_much_compressed_size(*table_id);
ptr += len;
return(ptr);
}
/**************************************************************************
Reads from an undo log record a stored column value. */
static
byte*
trx_undo_rec_get_col_val(
/*=====================*/
/* out: remaining part of undo log record after
reading these values */
byte* ptr, /* in: pointer to remaining part of undo log record */
byte** field, /* out: pointer to stored field */
ulint* len) /* out: length of the field, or UNIV_SQL_NULL */
{
*len = mach_read_compressed(ptr);
ptr += mach_get_compressed_size(*len);
*field = ptr;
if (*len != UNIV_SQL_NULL) {
if (*len >= UNIV_EXTERN_STORAGE_FIELD) {
ptr += (*len - UNIV_EXTERN_STORAGE_FIELD);
} else {
ptr += *len;
}
}
return(ptr);
}
/***********************************************************************
Builds a row reference from an undo log record. */
byte*
trx_undo_rec_get_row_ref(
/*=====================*/
/* out: pointer to remaining part of undo
record */
byte* ptr, /* in: remaining part of a copy of an undo log
record, at the start of the row reference;
NOTE that this copy of the undo log record must
be preserved as long as the row reference is
used, as we do NOT copy the data in the
record! */
dict_index_t* index, /* in: clustered index */
dtuple_t** ref, /* out, own: row reference */
mem_heap_t* heap) /* in: memory heap from which the memory
needed is allocated */
{
dfield_t* dfield;
byte* field;
ulint len;
ulint ref_len;
ulint i;
ut_ad(index && ptr && ref && heap);
ut_a(index->type & DICT_CLUSTERED);
ref_len = dict_index_get_n_unique(index);
*ref = dtuple_create(heap, ref_len);
dict_index_copy_types(*ref, index, ref_len);
for (i = 0; i < ref_len; i++) {
dfield = dtuple_get_nth_field(*ref, i);
ptr = trx_undo_rec_get_col_val(ptr, &field, &len);
dfield_set_data(dfield, field, len);
}
return(ptr);
}
/***********************************************************************
Skips a row reference from an undo log record. */
byte*
trx_undo_rec_skip_row_ref(
/*======================*/
/* out: pointer to remaining part of undo
record */
byte* ptr, /* in: remaining part in update undo log
record, at the start of the row reference */
dict_index_t* index) /* in: clustered index */
{
byte* field;
ulint len;
ulint ref_len;
ulint i;
ut_ad(index && ptr);
ut_a(index->type & DICT_CLUSTERED);
ref_len = dict_index_get_n_unique(index);
for (i = 0; i < ref_len; i++) {
ptr = trx_undo_rec_get_col_val(ptr, &field, &len);
}
return(ptr);
}
/**************************************************************************
Reports in the undo log of an update or delete marking of a clustered index
record. */
static
ulint
trx_undo_page_report_modify(
/*========================*/
/* out: byte offset of the inserted
undo log entry on the page if succeed,
0 if fail */
page_t* undo_page, /* in: undo log page */
trx_t* trx, /* in: transaction */
dict_index_t* index, /* in: clustered index where update or
delete marking is done */
rec_t* rec, /* in: clustered index record which
has NOT yet been modified */
const ulint* offsets, /* in: rec_get_offsets(rec, index) */
upd_t* update, /* in: update vector which tells the
columns to be updated; in the case of
a delete, this should be set to NULL */
ulint cmpl_info, /* in: compiler info on secondary
index updates */
mtr_t* mtr) /* in: mtr */
{
dict_table_t* table;
upd_field_t* upd_field;
dict_col_t* col;
ulint first_free;
byte* ptr;
ulint len;
byte* field;
ulint flen;
ulint pos;
dulint roll_ptr;
dulint trx_id;
ulint bits;
ulint col_no;
byte* old_ptr;
ulint type_cmpl;
byte* type_cmpl_ptr;
ulint i;
ut_a(index->type & DICT_CLUSTERED);
ut_ad(rec_offs_validate(rec, index, offsets));
ut_ad(mach_read_from_2(undo_page + TRX_UNDO_PAGE_HDR
+ TRX_UNDO_PAGE_TYPE) == TRX_UNDO_UPDATE);
table = index->table;
first_free = mach_read_from_2(undo_page + TRX_UNDO_PAGE_HDR
+ TRX_UNDO_PAGE_FREE);
ptr = undo_page + first_free;
ut_ad(first_free <= UNIV_PAGE_SIZE);
if (trx_undo_left(undo_page, ptr) < 50) {
/* NOTE: the value 50 must be big enough so that the general
fields written below fit on the undo log page */
return(0);
}
/* Reserve 2 bytes for the pointer to the next undo log record */
ptr += 2;
/* Store first some general parameters to the undo log */
if (update) {
if (rec_get_deleted_flag(rec, table->comp)) {
type_cmpl = TRX_UNDO_UPD_DEL_REC;
} else {
type_cmpl = TRX_UNDO_UPD_EXIST_REC;
}
} else {
type_cmpl = TRX_UNDO_DEL_MARK_REC;
}
type_cmpl = type_cmpl | (cmpl_info * TRX_UNDO_CMPL_INFO_MULT);
mach_write_to_1(ptr, type_cmpl);
type_cmpl_ptr = ptr;
ptr++;
len = mach_dulint_write_much_compressed(ptr, trx->undo_no);
ptr += len;
len = mach_dulint_write_much_compressed(ptr, table->id);
ptr += len;
/*----------------------------------------*/
/* Store the state of the info bits */
bits = rec_get_info_bits(rec, table->comp);
mach_write_to_1(ptr, bits);
ptr += 1;
/* Store the values of the system columns */
field = rec_get_nth_field(rec, offsets,
dict_index_get_sys_col_pos(index, DATA_TRX_ID), &len);
ut_ad(len == DATA_TRX_ID_LEN);
trx_id = trx_read_trx_id(field);
field = rec_get_nth_field(rec, offsets,
dict_index_get_sys_col_pos(index, DATA_ROLL_PTR), &len);
ut_ad(len == DATA_ROLL_PTR_LEN);
roll_ptr = trx_read_roll_ptr(field);
len = mach_dulint_write_compressed(ptr, trx_id);
ptr += len;
len = mach_dulint_write_compressed(ptr, roll_ptr);
ptr += len;
/*----------------------------------------*/
/* Store then the fields required to uniquely determine the
record which will be modified in the clustered index */
for (i = 0; i < dict_index_get_n_unique(index); i++) {
field = rec_get_nth_field(rec, offsets, i, &flen);
if (trx_undo_left(undo_page, ptr) < 4) {
return(0);
}
len = mach_write_compressed(ptr, flen);
ptr += len;
if (flen != UNIV_SQL_NULL) {
if (trx_undo_left(undo_page, ptr) < flen) {
return(0);
}
ut_memcpy(ptr, field, flen);
ptr += flen;
}
}
/*----------------------------------------*/
/* Save to the undo log the old values of the columns to be updated. */
if (update) {
if (trx_undo_left(undo_page, ptr) < 5) {
return(0);
}
len = mach_write_compressed(ptr, upd_get_n_fields(update));
ptr += len;
for (i = 0; i < upd_get_n_fields(update); i++) {
upd_field = upd_get_nth_field(update, i);
pos = upd_field->field_no;
/* Write field number to undo log */
if (trx_undo_left(undo_page, ptr) < 5) {
return(0);
}
len = mach_write_compressed(ptr, pos);
ptr += len;
/* Save the old value of field */
field = rec_get_nth_field(rec, offsets, pos, &flen);
if (trx_undo_left(undo_page, ptr) < 5) {
return(0);
}
if (rec_offs_nth_extern(offsets, pos)) {
/* If a field has external storage, we add to
flen the flag */
len = mach_write_compressed(ptr,
UNIV_EXTERN_STORAGE_FIELD + flen);
/* Notify purge that it eventually has to free the old
externally stored field */
trx->update_undo->del_marks = TRUE;
*type_cmpl_ptr = *type_cmpl_ptr | TRX_UNDO_UPD_EXTERN;
} else {
len = mach_write_compressed(ptr, flen);
}
ptr += len;
if (flen != UNIV_SQL_NULL) {
if (trx_undo_left(undo_page, ptr) < flen) {
return(0);
}
ut_memcpy(ptr, field, flen);
ptr += flen;
}
}
}
/*----------------------------------------*/
/* In the case of a delete marking, and also in the case of an update
where any ordering field of any index changes, store the values of all
columns which occur as ordering fields in any index. This info is used
in the purge of old versions where we use it to build and search the
delete marked index records, to look if we can remove them from the
index tree. Note that starting from 4.0.14 also externally stored
fields can be ordering in some index. But we always store at least
384 first bytes locally to the clustered index record, which means
we can construct the column prefix fields in the index from the
stored data. */
if (!update || !(cmpl_info & UPD_NODE_NO_ORD_CHANGE)) {
trx->update_undo->del_marks = TRUE;
if (trx_undo_left(undo_page, ptr) < 5) {
return(0);
}
old_ptr = ptr;
/* Reserve 2 bytes to write the number of bytes the stored fields
take in this undo record */
ptr += 2;
for (col_no = 0; col_no < dict_table_get_n_cols(table); col_no++) {
col = dict_table_get_nth_col(table, col_no);
if (col->ord_part > 0) {
pos = dict_index_get_nth_col_pos(index, col_no);
/* Write field number to undo log */
if (trx_undo_left(undo_page, ptr) < 5) {
return(0);
}
len = mach_write_compressed(ptr, pos);
ptr += len;
/* Save the old value of field */
field = rec_get_nth_field(rec, offsets, pos, &flen);
if (trx_undo_left(undo_page, ptr) < 5) {
return(0);
}
len = mach_write_compressed(ptr, flen);
ptr += len;
if (flen != UNIV_SQL_NULL) {
if (trx_undo_left(undo_page, ptr) < flen) {
return(0);
}
ut_memcpy(ptr, field, flen);
ptr += flen;
}
}
}
mach_write_to_2(old_ptr, ptr - old_ptr);
}
/*----------------------------------------*/
/* Write pointers to the previous and the next undo log records */
if (trx_undo_left(undo_page, ptr) < 2) {
return(0);
}
mach_write_to_2(ptr, first_free);
ptr += 2;
mach_write_to_2(undo_page + first_free, ptr - undo_page);
mach_write_to_2(undo_page + TRX_UNDO_PAGE_HDR + TRX_UNDO_PAGE_FREE,
ptr - undo_page);
/* Write to the REDO log about this change in the UNDO log */
trx_undof_page_add_undo_rec_log(undo_page, first_free,
ptr - undo_page, mtr);
return(first_free);
}
/**************************************************************************
Reads from an undo log update record the system field values of the old
version. */
byte*
trx_undo_update_rec_get_sys_cols(
/*=============================*/
/* out: remaining part of undo log
record after reading these values */
byte* ptr, /* in: remaining part of undo log
record after reading general
parameters */
dulint* trx_id, /* out: trx id */
dulint* roll_ptr, /* out: roll ptr */
ulint* info_bits) /* out: info bits state */
{
ulint len;
/* Read the state of the info bits */
*info_bits = mach_read_from_1(ptr);
ptr += 1;
/* Read the values of the system columns */
*trx_id = mach_dulint_read_compressed(ptr);
len = mach_dulint_get_compressed_size(*trx_id);
ptr += len;
*roll_ptr = mach_dulint_read_compressed(ptr);
len = mach_dulint_get_compressed_size(*roll_ptr);
ptr += len;
return(ptr);
}
/**************************************************************************
Reads from an update undo log record the number of updated fields. */
UNIV_INLINE
byte*
trx_undo_update_rec_get_n_upd_fields(
/*=================================*/
/* out: remaining part of undo log record after
reading this value */
byte* ptr, /* in: pointer to remaining part of undo log record */
ulint* n) /* out: number of fields */
{
*n = mach_read_compressed(ptr);
ptr += mach_get_compressed_size(*n);
return(ptr);
}
/**************************************************************************
Reads from an update undo log record a stored field number. */
UNIV_INLINE
byte*
trx_undo_update_rec_get_field_no(
/*=============================*/
/* out: remaining part of undo log record after
reading this value */
byte* ptr, /* in: pointer to remaining part of undo log record */
ulint* field_no)/* out: field number */
{
*field_no = mach_read_compressed(ptr);
ptr += mach_get_compressed_size(*field_no);
return(ptr);
}
/***********************************************************************
Builds an update vector based on a remaining part of an undo log record. */
byte*
trx_undo_update_rec_get_update(
/*===========================*/
/* out: remaining part of the record,
NULL if an error detected, which means that
the record is corrupted */
byte* ptr, /* in: remaining part in update undo log
record, after reading the row reference
NOTE that this copy of the undo log record must
be preserved as long as the update vector is
used, as we do NOT copy the data in the
record! */
dict_index_t* index, /* in: clustered index */
ulint type, /* in: TRX_UNDO_UPD_EXIST_REC,
TRX_UNDO_UPD_DEL_REC, or
TRX_UNDO_DEL_MARK_REC; in the last case,
only trx id and roll ptr fields are added to
the update vector */
dulint trx_id, /* in: transaction id from this undo record */
dulint roll_ptr,/* in: roll pointer from this undo record */
ulint info_bits,/* in: info bits from this undo record */
trx_t* trx, /* in: transaction */
mem_heap_t* heap, /* in: memory heap from which the memory
needed is allocated */
upd_t** upd) /* out, own: update vector */
{
upd_field_t* upd_field;
upd_t* update;
ulint n_fields;
byte* buf;
byte* field;
ulint len;
ulint field_no;
ulint i;
ut_a(index->type & DICT_CLUSTERED);
if (type != TRX_UNDO_DEL_MARK_REC) {
ptr = trx_undo_update_rec_get_n_upd_fields(ptr, &n_fields);
} else {
n_fields = 0;
}
update = upd_create(n_fields + 2, heap);
update->info_bits = info_bits;
/* Store first trx id and roll ptr to update vector */
upd_field = upd_get_nth_field(update, n_fields);
buf = mem_heap_alloc(heap, DATA_TRX_ID_LEN);
trx_write_trx_id(buf, trx_id);
upd_field_set_field_no(upd_field,
dict_index_get_sys_col_pos(index, DATA_TRX_ID),
index, trx);
dfield_set_data(&(upd_field->new_val), buf, DATA_TRX_ID_LEN);
upd_field = upd_get_nth_field(update, n_fields + 1);
buf = mem_heap_alloc(heap, DATA_ROLL_PTR_LEN);
trx_write_roll_ptr(buf, roll_ptr);
upd_field_set_field_no(upd_field,
dict_index_get_sys_col_pos(index, DATA_ROLL_PTR),
index, trx);
dfield_set_data(&(upd_field->new_val), buf, DATA_ROLL_PTR_LEN);
/* Store then the updated ordinary columns to the update vector */
for (i = 0; i < n_fields; i++) {
ptr = trx_undo_update_rec_get_field_no(ptr, &field_no);
if (field_no >= dict_index_get_n_fields(index)) {
fprintf(stderr,
"InnoDB: Error: trying to access update undo rec field %lu in ", (ulong) field_no);
dict_index_name_print(stderr, trx, index);
fprintf(stderr, "\n"
"InnoDB: but index has only %lu fields\n"
"InnoDB: Submit a detailed bug report to http://bugs.mysql.com\n"
"InnoDB: Run also CHECK TABLE ",
(ulong) dict_index_get_n_fields(index));
ut_print_name(stderr, trx, index->table_name);
fprintf(stderr, "\n"
"InnoDB: n_fields = %lu, i = %lu, ptr %p\n",
(ulong) n_fields, (ulong) i, ptr);
return(NULL);
}
ptr = trx_undo_rec_get_col_val(ptr, &field, &len);
upd_field = upd_get_nth_field(update, i);
upd_field_set_field_no(upd_field, field_no, index, trx);
if (len != UNIV_SQL_NULL && len >= UNIV_EXTERN_STORAGE_FIELD) {
upd_field->extern_storage = TRUE;
len -= UNIV_EXTERN_STORAGE_FIELD;
}
dfield_set_data(&(upd_field->new_val), field, len);
}
*upd = update;
return(ptr);
}
/***********************************************************************
Builds a partial row from an update undo log record. It contains the
columns which occur as ordering in any index of the table. */
byte*
trx_undo_rec_get_partial_row(
/*=========================*/
/* out: pointer to remaining part of undo
record */
byte* ptr, /* in: remaining part in update undo log
record of a suitable type, at the start of
the stored index columns;
NOTE that this copy of the undo log record must
be preserved as long as the partial row is
used, as we do NOT copy the data in the
record! */
dict_index_t* index, /* in: clustered index */
dtuple_t** row, /* out, own: partial row */
mem_heap_t* heap) /* in: memory heap from which the memory
needed is allocated */
{
dfield_t* dfield;
byte* field;
ulint len;
ulint field_no;
ulint col_no;
ulint row_len;
ulint total_len;
byte* start_ptr;
ulint i;
ut_ad(index && ptr && row && heap);
row_len = dict_table_get_n_cols(index->table);
*row = dtuple_create(heap, row_len);
dict_table_copy_types(*row, index->table);
start_ptr = ptr;
total_len = mach_read_from_2(ptr);
ptr += 2;
for (i = 0;; i++) {
if (ptr == start_ptr + total_len) {
break;
}
ptr = trx_undo_update_rec_get_field_no(ptr, &field_no);
col_no = dict_index_get_nth_col_no(index, field_no);
ptr = trx_undo_rec_get_col_val(ptr, &field, &len);
dfield = dtuple_get_nth_field(*row, col_no);
dfield_set_data(dfield, field, len);
}
return(ptr);
}
/***************************************************************************
Erases the unused undo log page end. */
static
void
trx_undo_erase_page_end(
/*====================*/
page_t* undo_page, /* in: undo page whose end to erase */
mtr_t* mtr) /* in: mtr */
{
ulint first_free;
first_free = mach_read_from_2(undo_page + TRX_UNDO_PAGE_HDR
+ TRX_UNDO_PAGE_FREE);
memset(undo_page + first_free, 0xff,
(UNIV_PAGE_SIZE - FIL_PAGE_DATA_END) - first_free);
mlog_write_initial_log_record(undo_page, MLOG_UNDO_ERASE_END, mtr);
}
/***************************************************************
Parses a redo log record of erasing of an undo page end. */
byte*
trx_undo_parse_erase_page_end(
/*==========================*/
/* out: end of log record or NULL */
byte* ptr, /* in: buffer */
byte* end_ptr __attribute__((unused)), /* in: buffer end */
page_t* page, /* in: page or NULL */
mtr_t* mtr) /* in: mtr or NULL */
{
ut_ad(ptr && end_ptr);
if (page == NULL) {
return(ptr);
}
trx_undo_erase_page_end(page, mtr);
return(ptr);
}
/***************************************************************************
Writes information to an undo log about an insert, update, or a delete marking
of a clustered index record. This information is used in a rollback of the
transaction and in consistent reads that must look to the history of this
transaction. */
ulint
trx_undo_report_row_operation(
/*==========================*/
/* out: DB_SUCCESS or error code */
ulint flags, /* in: if BTR_NO_UNDO_LOG_FLAG bit is
set, does nothing */
ulint op_type, /* in: TRX_UNDO_INSERT_OP or
TRX_UNDO_MODIFY_OP */
que_thr_t* thr, /* in: query thread */
dict_index_t* index, /* in: clustered index */
dtuple_t* clust_entry, /* in: in the case of an insert,
index entry to insert into the
clustered index, otherwise NULL */
upd_t* update, /* in: in the case of an update,
the update vector, otherwise NULL */
ulint cmpl_info, /* in: compiler info on secondary
index updates */
rec_t* rec, /* in: in case of an update or delete
marking, the record in the clustered
index, otherwise NULL */
dulint* roll_ptr) /* out: rollback pointer to the
inserted undo log record,
ut_dulint_zero if BTR_NO_UNDO_LOG
flag was specified */
{
trx_t* trx;
trx_undo_t* undo;
page_t* undo_page;
ulint offset;
ulint page_no;
ibool is_insert;
trx_rseg_t* rseg;
mtr_t mtr;
mem_heap_t* heap = NULL;
ulint offsets_[REC_OFFS_NORMAL_SIZE];
ulint* offsets = offsets_;
*offsets_ = (sizeof offsets_) / sizeof *offsets_;
ut_a(index->type & DICT_CLUSTERED);
if (flags & BTR_NO_UNDO_LOG_FLAG) {
*roll_ptr = ut_dulint_zero;
return(DB_SUCCESS);
}
ut_ad(thr);
ut_ad((op_type != TRX_UNDO_INSERT_OP)
|| (clust_entry && !update && !rec));
trx = thr_get_trx(thr);
rseg = trx->rseg;
mutex_enter(&(trx->undo_mutex));
/* If the undo log is not assigned yet, assign one */
if (op_type == TRX_UNDO_INSERT_OP) {
if (trx->insert_undo == NULL) {
trx_undo_assign_undo(trx, TRX_UNDO_INSERT);
}
undo = trx->insert_undo;
is_insert = TRUE;
} else {
ut_ad(op_type == TRX_UNDO_MODIFY_OP);
if (trx->update_undo == NULL) {
trx_undo_assign_undo(trx, TRX_UNDO_UPDATE);
}
undo = trx->update_undo;
is_insert = FALSE;
}
if (undo == NULL) {
/* Did not succeed: out of space */
mutex_exit(&(trx->undo_mutex));
return(DB_OUT_OF_FILE_SPACE);
}
page_no = undo->last_page_no;
mtr_start(&mtr);
for (;;) {
undo_page = buf_page_get_gen(undo->space, page_no,
RW_X_LATCH, undo->guess_page,
BUF_GET,
__FILE__, __LINE__,
&mtr);
#ifdef UNIV_SYNC_DEBUG
buf_page_dbg_add_level(undo_page, SYNC_TRX_UNDO_PAGE);
#endif /* UNIV_SYNC_DEBUG */
if (op_type == TRX_UNDO_INSERT_OP) {
offset = trx_undo_page_report_insert(undo_page, trx,
index, clust_entry,
&mtr);
} else {
offsets = rec_get_offsets(rec, index, offsets,
ULINT_UNDEFINED, &heap);
offset = trx_undo_page_report_modify(undo_page, trx,
index, rec, offsets, update, cmpl_info, &mtr);
}
if (offset == 0) {
/* The record did not fit on the page. We erase the
end segment of the undo log page and write a log
record of it: this is to ensure that in the debug
version the replicate page constructed using the log
records stays identical to the original page */
trx_undo_erase_page_end(undo_page, &mtr);
}
mtr_commit(&mtr);
if (offset != 0) {
/* Success */
break;
}
ut_ad(page_no == undo->last_page_no);
/* We have to extend the undo log by one page */
mtr_start(&mtr);
/* When we add a page to an undo log, this is analogous to
a pessimistic insert in a B-tree, and we must reserve the
counterpart of the tree latch, which is the rseg mutex. */
mutex_enter(&(rseg->mutex));
page_no = trx_undo_add_page(trx, undo, &mtr);
mutex_exit(&(rseg->mutex));
if (page_no == FIL_NULL) {
/* Did not succeed: out of space */
mutex_exit(&(trx->undo_mutex));
mtr_commit(&mtr);
if (UNIV_LIKELY_NULL(heap)) {
mem_heap_free(heap);
}
return(DB_OUT_OF_FILE_SPACE);
}
}
undo->empty = FALSE;
undo->top_page_no = page_no;
undo->top_offset = offset;
undo->top_undo_no = trx->undo_no;
undo->guess_page = undo_page;
UT_DULINT_INC(trx->undo_no);
mutex_exit(&(trx->undo_mutex));
*roll_ptr = trx_undo_build_roll_ptr(is_insert, rseg->id, page_no,
offset);
if (UNIV_LIKELY_NULL(heap)) {
mem_heap_free(heap);
}
return(DB_SUCCESS);
}
/*============== BUILDING PREVIOUS VERSION OF A RECORD ===============*/
/**********************************************************************
Copies an undo record to heap. This function can be called if we know that
the undo log record exists. */
trx_undo_rec_t*
trx_undo_get_undo_rec_low(
/*======================*/
/* out, own: copy of the record */
dulint roll_ptr, /* in: roll pointer to record */
mem_heap_t* heap) /* in: memory heap where copied */
{
trx_undo_rec_t* undo_rec;
ulint rseg_id;
ulint page_no;
ulint offset;
page_t* undo_page;
trx_rseg_t* rseg;
ibool is_insert;
mtr_t mtr;
trx_undo_decode_roll_ptr(roll_ptr, &is_insert, &rseg_id, &page_no,
&offset);
rseg = trx_rseg_get_on_id(rseg_id);
mtr_start(&mtr);
undo_page = trx_undo_page_get_s_latched(rseg->space, page_no, &mtr);
undo_rec = trx_undo_rec_copy(undo_page + offset, heap);
mtr_commit(&mtr);
return(undo_rec);
}
/**********************************************************************
Copies an undo record to heap. */
ulint
trx_undo_get_undo_rec(
/*==================*/
/* out: DB_SUCCESS, or
DB_MISSING_HISTORY if the undo log
has been truncated and we cannot
fetch the old version; NOTE: the
caller must have latches on the
clustered index page and purge_view */
dulint roll_ptr, /* in: roll pointer to record */
dulint trx_id, /* in: id of the trx that generated
the roll pointer: it points to an
undo log of this transaction */
trx_undo_rec_t** undo_rec, /* out, own: copy of the record */
mem_heap_t* heap) /* in: memory heap where copied */
{
#ifdef UNIV_SYNC_DEBUG
ut_ad(rw_lock_own(&(purge_sys->latch), RW_LOCK_SHARED));
#endif /* UNIV_SYNC_DEBUG */
if (!trx_purge_update_undo_must_exist(trx_id)) {
/* It may be that the necessary undo log has already been
deleted */
return(DB_MISSING_HISTORY);
}
*undo_rec = trx_undo_get_undo_rec_low(roll_ptr, heap);
return(DB_SUCCESS);
}
/***********************************************************************
Build a previous version of a clustered index record. This function checks
that the caller has a latch on the index page of the clustered index record
and an s-latch on the purge_view. This guarantees that the stack of versions
is locked. */
ulint
trx_undo_prev_version_build(
/*========================*/
/* out: DB_SUCCESS, or DB_MISSING_HISTORY if
the previous version is not >= purge_view,
which means that it may have been removed,
DB_ERROR if corrupted record */
rec_t* index_rec,/* in: clustered index record in the
index tree */
mtr_t* index_mtr __attribute__((unused)),
/* in: mtr which contains the latch to
index_rec page and purge_view */
rec_t* rec, /* in: version of a clustered index record */
dict_index_t* index, /* in: clustered index */
ulint* offsets,/* in: rec_get_offsets(rec, index) */
mem_heap_t* heap, /* in: memory heap from which the memory
needed is allocated */
rec_t** old_vers)/* out, own: previous version, or NULL if
rec is the first inserted version, or if
history data has been deleted */
{
trx_undo_rec_t* undo_rec;
dtuple_t* entry;
dulint rec_trx_id;
ulint type;
dulint undo_no;
dulint table_id;
dulint trx_id;
dulint roll_ptr;
dulint old_roll_ptr;
upd_t* update;
byte* ptr;
ulint info_bits;
ulint cmpl_info;
ibool dummy_extern;
byte* buf;
ulint err;
#ifdef UNIV_SYNC_DEBUG
ut_ad(rw_lock_own(&(purge_sys->latch), RW_LOCK_SHARED));
#endif /* UNIV_SYNC_DEBUG */
ut_ad(mtr_memo_contains(index_mtr, buf_block_align(index_rec),
MTR_MEMO_PAGE_S_FIX) ||
mtr_memo_contains(index_mtr, buf_block_align(index_rec),
MTR_MEMO_PAGE_X_FIX));
ut_ad(rec_offs_validate(rec, index, offsets));
if (!(index->type & DICT_CLUSTERED)) {
fprintf(stderr, "InnoDB: Error: trying to access"
" update undo rec for non-clustered index %s\n"
"InnoDB: Submit a detailed bug report to"
" http://bugs.mysql.com\n"
"InnoDB: index record ", index->name);
rec_print(stderr, index_rec, index);
fputs("\n"
"InnoDB: record version ", stderr);
rec_print_new(stderr, rec, offsets);
putc('\n', stderr);
return(DB_ERROR);
}
roll_ptr = row_get_rec_roll_ptr(rec, index, offsets);
old_roll_ptr = roll_ptr;
*old_vers = NULL;
if (trx_undo_roll_ptr_is_insert(roll_ptr)) {
/* The record rec is the first inserted version */
return(DB_SUCCESS);
}
rec_trx_id = row_get_rec_trx_id(rec, index, offsets);
err = trx_undo_get_undo_rec(roll_ptr, rec_trx_id, &undo_rec, heap);
if (err != DB_SUCCESS) {
return(err);
}
ptr = trx_undo_rec_get_pars(undo_rec, &type, &cmpl_info,
&dummy_extern, &undo_no, &table_id);
ptr = trx_undo_update_rec_get_sys_cols(ptr, &trx_id, &roll_ptr,
&info_bits);
ptr = trx_undo_rec_skip_row_ref(ptr, index);
ptr = trx_undo_update_rec_get_update(ptr, index, type, trx_id,
roll_ptr, info_bits, NULL, heap, &update);
if (ut_dulint_cmp(table_id, index->table->id) != 0) {
ptr = NULL;
fprintf(stderr,
"InnoDB: Error: trying to access update undo rec for table %s\n"
"InnoDB: but the table id in the undo record is wrong\n"
"InnoDB: Submit a detailed bug report to http://bugs.mysql.com\n"
"InnoDB: Run also CHECK TABLE %s\n",
index->table_name, index->table_name);
}
if (ptr == NULL) {
/* The record was corrupted, return an error; these printfs
should catch an elusive bug in row_vers_old_has_index_entry */
fprintf(stderr,
"InnoDB: table %s, index %s, n_uniq %lu\n"
"InnoDB: undo rec address %p, type %lu cmpl_info %lu\n"
"InnoDB: undo rec table id %lu %lu, index table id %lu %lu\n"
"InnoDB: dump of 150 bytes in undo rec: ",
index->table_name, index->name,
(ulong) dict_index_get_n_unique(index),
undo_rec, (ulong) type, (ulong) cmpl_info,
(ulong) ut_dulint_get_high(table_id),
(ulong) ut_dulint_get_low(table_id),
(ulong) ut_dulint_get_high(index->table->id),
(ulong) ut_dulint_get_low(index->table->id));
ut_print_buf(stderr, undo_rec, 150);
fputs("\n"
"InnoDB: index record ", stderr);
rec_print(stderr, index_rec, index);
fputs("\n"
"InnoDB: record version ", stderr);
rec_print_new(stderr, rec, offsets);
fprintf(stderr, "\n"
"InnoDB: Record trx id %lu %lu, update rec trx id %lu %lu\n"
"InnoDB: Roll ptr in rec %lu %lu, in update rec %lu %lu\n",
(ulong) ut_dulint_get_high(rec_trx_id),
(ulong) ut_dulint_get_low(rec_trx_id),
(ulong) ut_dulint_get_high(trx_id),
(ulong) ut_dulint_get_low(trx_id),
(ulong) ut_dulint_get_high(old_roll_ptr),
(ulong) ut_dulint_get_low(old_roll_ptr),
(ulong) ut_dulint_get_high(roll_ptr),
(ulong) ut_dulint_get_low(roll_ptr));
trx_purge_sys_print();
return(DB_ERROR);
}
if (row_upd_changes_field_size_or_external(index, offsets, update)) {
ulint* ext_vect;
ulint n_ext_vect;
/* We have to set the appropriate extern storage bits in the
old version of the record: the extern bits in rec for those
fields that update does NOT update, as well as the the bits for
those fields that update updates to become externally stored
fields. Store the info to ext_vect: */
ext_vect = mem_alloc(sizeof(ulint)
* rec_offs_n_fields(offsets));
n_ext_vect = btr_push_update_extern_fields(ext_vect, offsets,
update);
entry = row_rec_to_index_entry(ROW_COPY_DATA, index, rec,
heap);
row_upd_index_replace_new_col_vals(entry, index, update, heap);
buf = mem_heap_alloc(heap,
rec_get_converted_size(index, entry));
*old_vers = rec_convert_dtuple_to_rec(buf, index, entry);
/* Now set the extern bits in the old version of the record */
rec_set_field_extern_bits(*old_vers, index,
ext_vect, n_ext_vect, NULL);
mem_free(ext_vect);
} else {
buf = mem_heap_alloc(heap, rec_offs_size(offsets));
*old_vers = rec_copy(buf, rec, offsets);
rec_offs_make_valid(*old_vers, index, offsets);
row_upd_rec_in_place(*old_vers, offsets, update);
}
return(DB_SUCCESS);
}
|