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
|
/*
** 2004 May 22
**
** The author disclaims copyright to this source code. In place of
** a legal notice, here is a blessing:
**
** May you do good and not evil.
** May you find forgiveness for yourself and forgive others.
** May you share freely, never taking more than you give.
**
******************************************************************************
**
** This file contains code that is specific to Unix systems.
*/
#include "sqliteInt.h"
#include "os.h"
#if OS_UNIX /* This file is used on unix only */
#include <time.h>
#include <sys/time.h>
#include <errno.h>
#include <unistd.h>
/*
** Do not include any of the File I/O interface procedures if the
** SQLITE_OMIT_DISKIO macro is defined (indicating that there database
** will be in-memory only)
*/
#ifndef SQLITE_OMIT_DISKIO
/*
** Define various macros that are missing from some systems.
*/
#ifndef O_LARGEFILE
# define O_LARGEFILE 0
#endif
#ifdef SQLITE_DISABLE_LFS
# undef O_LARGEFILE
# define O_LARGEFILE 0
#endif
#ifndef O_NOFOLLOW
# define O_NOFOLLOW 0
#endif
#ifndef O_BINARY
# define O_BINARY 0
#endif
/*
** The DJGPP compiler environment looks mostly like Unix, but it
** lacks the fcntl() system call. So redefine fcntl() to be something
** that always succeeds. This means that locking does not occur under
** DJGPP. But its DOS - what did you expect?
*/
#ifdef __DJGPP__
# define fcntl(A,B,C) 0
#endif
/*
** Include code that is common to all os_*.c files
*/
#include "os_common.h"
/*
** The threadid macro resolves to the thread-id or to 0. Used for
** testing and debugging only.
*/
#ifdef SQLITE_UNIX_THREADS
#define threadid pthread_self()
#else
#define threadid 0
#endif
/*
** Set or check the OsFile.tid field. This field is set when an OsFile
** is first opened. All subsequent uses of the OsFile verify that the
** same thread is operating on the OsFile. Some operating systems do
** not allow locks to be overridden by other threads and that restriction
** means that sqlite3* database handles cannot be moved from one thread
** to another. This logic makes sure a user does not try to do that
** by mistake.
*/
#ifdef SQLITE_UNIX_THREADS
# define SET_THREADID(X) X->tid = pthread_self()
# define CHECK_THREADID(X) (!pthread_equal(X->tid, pthread_self()))
#else
# define SET_THREADID(X)
# define CHECK_THREADID(X) 0
#endif
/*
** Here is the dirt on POSIX advisory locks: ANSI STD 1003.1 (1996)
** section 6.5.2.2 lines 483 through 490 specify that when a process
** sets or clears a lock, that operation overrides any prior locks set
** by the same process. It does not explicitly say so, but this implies
** that it overrides locks set by the same process using a different
** file descriptor. Consider this test case:
**
** int fd1 = open("./file1", O_RDWR|O_CREAT, 0644);
** int fd2 = open("./file2", O_RDWR|O_CREAT, 0644);
**
** Suppose ./file1 and ./file2 are really the same file (because
** one is a hard or symbolic link to the other) then if you set
** an exclusive lock on fd1, then try to get an exclusive lock
** on fd2, it works. I would have expected the second lock to
** fail since there was already a lock on the file due to fd1.
** But not so. Since both locks came from the same process, the
** second overrides the first, even though they were on different
** file descriptors opened on different file names.
**
** Bummer. If you ask me, this is broken. Badly broken. It means
** that we cannot use POSIX locks to synchronize file access among
** competing threads of the same process. POSIX locks will work fine
** to synchronize access for threads in separate processes, but not
** threads within the same process.
**
** To work around the problem, SQLite has to manage file locks internally
** on its own. Whenever a new database is opened, we have to find the
** specific inode of the database file (the inode is determined by the
** st_dev and st_ino fields of the stat structure that fstat() fills in)
** and check for locks already existing on that inode. When locks are
** created or removed, we have to look at our own internal record of the
** locks to see if another thread has previously set a lock on that same
** inode.
**
** The OsFile structure for POSIX is no longer just an integer file
** descriptor. It is now a structure that holds the integer file
** descriptor and a pointer to a structure that describes the internal
** locks on the corresponding inode. There is one locking structure
** per inode, so if the same inode is opened twice, both OsFile structures
** point to the same locking structure. The locking structure keeps
** a reference count (so we will know when to delete it) and a "cnt"
** field that tells us its internal lock status. cnt==0 means the
** file is unlocked. cnt==-1 means the file has an exclusive lock.
** cnt>0 means there are cnt shared locks on the file.
**
** Any attempt to lock or unlock a file first checks the locking
** structure. The fcntl() system call is only invoked to set a
** POSIX lock if the internal lock structure transitions between
** a locked and an unlocked state.
**
** 2004-Jan-11:
** More recent discoveries about POSIX advisory locks. (The more
** I discover, the more I realize the a POSIX advisory locks are
** an abomination.)
**
** If you close a file descriptor that points to a file that has locks,
** all locks on that file that are owned by the current process are
** released. To work around this problem, each OsFile structure contains
** a pointer to an openCnt structure. There is one openCnt structure
** per open inode, which means that multiple OsFiles can point to a single
** openCnt. When an attempt is made to close an OsFile, if there are
** other OsFiles open on the same inode that are holding locks, the call
** to close() the file descriptor is deferred until all of the locks clear.
** The openCnt structure keeps a list of file descriptors that need to
** be closed and that list is walked (and cleared) when the last lock
** clears.
**
** First, under Linux threads, because each thread has a separate
** process ID, lock operations in one thread do not override locks
** to the same file in other threads. Linux threads behave like
** separate processes in this respect. But, if you close a file
** descriptor in linux threads, all locks are cleared, even locks
** on other threads and even though the other threads have different
** process IDs. Linux threads is inconsistent in this respect.
** (I'm beginning to think that linux threads is an abomination too.)
** The consequence of this all is that the hash table for the lockInfo
** structure has to include the process id as part of its key because
** locks in different threads are treated as distinct. But the
** openCnt structure should not include the process id in its
** key because close() clears lock on all threads, not just the current
** thread. Were it not for this goofiness in linux threads, we could
** combine the lockInfo and openCnt structures into a single structure.
**
** 2004-Jun-28:
** On some versions of linux, threads can override each others locks.
** On others not. Sometimes you can change the behavior on the same
** system by setting the LD_ASSUME_KERNEL environment variable. The
** POSIX standard is silent as to which behavior is correct, as far
** as I can tell, so other versions of unix might show the same
** inconsistency. There is no little doubt in my mind that posix
** advisory locks and linux threads are profoundly broken.
**
** To work around the inconsistencies, we have to test at runtime
** whether or not threads can override each others locks. This test
** is run once, the first time any lock is attempted. A static
** variable is set to record the results of this test for future
** use.
*/
/*
** An instance of the following structure serves as the key used
** to locate a particular lockInfo structure given its inode.
**
** If threads cannot override each others locks, then we set the
** lockKey.tid field to the thread ID. If threads can override
** each others locks then tid is always set to zero. tid is also
** set to zero if we compile without threading support.
*/
struct lockKey {
dev_t dev; /* Device number */
ino_t ino; /* Inode number */
#ifdef SQLITE_UNIX_THREADS
pthread_t tid; /* Thread ID or zero if threads cannot override each other */
#endif
};
/*
** An instance of the following structure is allocated for each open
** inode on each thread with a different process ID. (Threads have
** different process IDs on linux, but not on most other unixes.)
**
** A single inode can have multiple file descriptors, so each OsFile
** structure contains a pointer to an instance of this object and this
** object keeps a count of the number of OsFiles pointing to it.
*/
struct lockInfo {
struct lockKey key; /* The lookup key */
int cnt; /* Number of SHARED locks held */
int locktype; /* One of SHARED_LOCK, RESERVED_LOCK etc. */
int nRef; /* Number of pointers to this structure */
};
/*
** An instance of the following structure serves as the key used
** to locate a particular openCnt structure given its inode. This
** is the same as the lockKey except that the thread ID is omitted.
*/
struct openKey {
dev_t dev; /* Device number */
ino_t ino; /* Inode number */
};
/*
** An instance of the following structure is allocated for each open
** inode. This structure keeps track of the number of locks on that
** inode. If a close is attempted against an inode that is holding
** locks, the close is deferred until all locks clear by adding the
** file descriptor to be closed to the pending list.
*/
struct openCnt {
struct openKey key; /* The lookup key */
int nRef; /* Number of pointers to this structure */
int nLock; /* Number of outstanding locks */
int nPending; /* Number of pending close() operations */
int *aPending; /* Malloced space holding fd's awaiting a close() */
};
/*
** These hash table maps inodes and process IDs into lockInfo and openCnt
** structures. Access to these hash tables must be protected by a mutex.
*/
static Hash lockHash = { SQLITE_HASH_BINARY, 0, 0, 0, 0, 0 };
static Hash openHash = { SQLITE_HASH_BINARY, 0, 0, 0, 0, 0 };
#ifdef SQLITE_UNIX_THREADS
/*
** This variable records whether or not threads can override each others
** locks.
**
** 0: No. Threads cannot override each others locks.
** 1: Yes. Threads can override each others locks.
** -1: We don't know yet.
*/
static int threadsOverrideEachOthersLocks = -1;
/*
** This structure holds information passed into individual test
** threads by the testThreadLockingBehavior() routine.
*/
struct threadTestData {
int fd; /* File to be locked */
struct flock lock; /* The locking operation */
int result; /* Result of the locking operation */
};
#ifdef SQLITE_LOCK_TRACE
/*
** Print out information about all locking operations.
**
** This routine is used for troubleshooting locks on multithreaded
** platforms. Enable by compiling with the -DSQLITE_LOCK_TRACE
** command-line option on the compiler. This code is normally
** turnned off.
*/
static int lockTrace(int fd, int op, struct flock *p){
char *zOpName, *zType;
int s;
int savedErrno;
if( op==F_GETLK ){
zOpName = "GETLK";
}else if( op==F_SETLK ){
zOpName = "SETLK";
}else{
s = fcntl(fd, op, p);
sqlite3DebugPrintf("fcntl unknown %d %d %d\n", fd, op, s);
return s;
}
if( p->l_type==F_RDLCK ){
zType = "RDLCK";
}else if( p->l_type==F_WRLCK ){
zType = "WRLCK";
}else if( p->l_type==F_UNLCK ){
zType = "UNLCK";
}else{
assert( 0 );
}
assert( p->l_whence==SEEK_SET );
s = fcntl(fd, op, p);
savedErrno = errno;
sqlite3DebugPrintf("fcntl %d %d %s %s %d %d %d %d\n",
threadid, fd, zOpName, zType, (int)p->l_start, (int)p->l_len,
(int)p->l_pid, s);
if( s && op==F_SETLK && (p->l_type==F_RDLCK || p->l_type==F_WRLCK) ){
struct flock l2;
l2 = *p;
fcntl(fd, F_GETLK, &l2);
if( l2.l_type==F_RDLCK ){
zType = "RDLCK";
}else if( l2.l_type==F_WRLCK ){
zType = "WRLCK";
}else if( l2.l_type==F_UNLCK ){
zType = "UNLCK";
}else{
assert( 0 );
}
sqlite3DebugPrintf("fcntl-failure-reason: %s %d %d %d\n",
zType, (int)l2.l_start, (int)l2.l_len, (int)l2.l_pid);
}
errno = savedErrno;
return s;
}
#define fcntl lockTrace
#endif /* SQLITE_LOCK_TRACE */
/*
** The testThreadLockingBehavior() routine launches two separate
** threads on this routine. This routine attempts to lock a file
** descriptor then returns. The success or failure of that attempt
** allows the testThreadLockingBehavior() procedure to determine
** whether or not threads can override each others locks.
*/
static void *threadLockingTest(void *pArg){
struct threadTestData *pData = (struct threadTestData*)pArg;
pData->result = fcntl(pData->fd, F_SETLK, &pData->lock);
return pArg;
}
/*
** This procedure attempts to determine whether or not threads
** can override each others locks then sets the
** threadsOverrideEachOthersLocks variable appropriately.
*/
static void testThreadLockingBehavior(fd_orig){
int fd;
struct threadTestData d[2];
pthread_t t[2];
fd = dup(fd_orig);
if( fd<0 ) return;
memset(d, 0, sizeof(d));
d[0].fd = fd;
d[0].lock.l_type = F_RDLCK;
d[0].lock.l_len = 1;
d[0].lock.l_start = 0;
d[0].lock.l_whence = SEEK_SET;
d[1] = d[0];
d[1].lock.l_type = F_WRLCK;
pthread_create(&t[0], 0, threadLockingTest, &d[0]);
pthread_create(&t[1], 0, threadLockingTest, &d[1]);
pthread_join(t[0], 0);
pthread_join(t[1], 0);
close(fd);
threadsOverrideEachOthersLocks = d[0].result==0 && d[1].result==0;
}
#endif /* SQLITE_UNIX_THREADS */
/*
** Release a lockInfo structure previously allocated by findLockInfo().
*/
static void releaseLockInfo(struct lockInfo *pLock){
pLock->nRef--;
if( pLock->nRef==0 ){
sqlite3HashInsert(&lockHash, &pLock->key, sizeof(pLock->key), 0);
sqliteFree(pLock);
}
}
/*
** Release a openCnt structure previously allocated by findLockInfo().
*/
static void releaseOpenCnt(struct openCnt *pOpen){
pOpen->nRef--;
if( pOpen->nRef==0 ){
sqlite3HashInsert(&openHash, &pOpen->key, sizeof(pOpen->key), 0);
sqliteFree(pOpen->aPending);
sqliteFree(pOpen);
}
}
/*
** Given a file descriptor, locate lockInfo and openCnt structures that
** describes that file descriptor. Create a new ones if necessary. The
** return values might be unset if an error occurs.
**
** Return the number of errors.
*/
static int findLockInfo(
int fd, /* The file descriptor used in the key */
struct lockInfo **ppLock, /* Return the lockInfo structure here */
struct openCnt **ppOpen /* Return the openCnt structure here */
){
int rc;
struct lockKey key1;
struct openKey key2;
struct stat statbuf;
struct lockInfo *pLock;
struct openCnt *pOpen;
rc = fstat(fd, &statbuf);
if( rc!=0 ) return 1;
memset(&key1, 0, sizeof(key1));
key1.dev = statbuf.st_dev;
key1.ino = statbuf.st_ino;
#ifdef SQLITE_UNIX_THREADS
if( threadsOverrideEachOthersLocks<0 ){
testThreadLockingBehavior(fd);
}
key1.tid = threadsOverrideEachOthersLocks ? 0 : pthread_self();
#endif
memset(&key2, 0, sizeof(key2));
key2.dev = statbuf.st_dev;
key2.ino = statbuf.st_ino;
pLock = (struct lockInfo*)sqlite3HashFind(&lockHash, &key1, sizeof(key1));
if( pLock==0 ){
struct lockInfo *pOld;
pLock = sqliteMallocRaw( sizeof(*pLock) );
if( pLock==0 ) return 1;
pLock->key = key1;
pLock->nRef = 1;
pLock->cnt = 0;
pLock->locktype = 0;
pOld = sqlite3HashInsert(&lockHash, &pLock->key, sizeof(key1), pLock);
if( pOld!=0 ){
assert( pOld==pLock );
sqliteFree(pLock);
return 1;
}
}else{
pLock->nRef++;
}
*ppLock = pLock;
pOpen = (struct openCnt*)sqlite3HashFind(&openHash, &key2, sizeof(key2));
if( pOpen==0 ){
struct openCnt *pOld;
pOpen = sqliteMallocRaw( sizeof(*pOpen) );
if( pOpen==0 ){
releaseLockInfo(pLock);
return 1;
}
pOpen->key = key2;
pOpen->nRef = 1;
pOpen->nLock = 0;
pOpen->nPending = 0;
pOpen->aPending = 0;
pOld = sqlite3HashInsert(&openHash, &pOpen->key, sizeof(key2), pOpen);
if( pOld!=0 ){
assert( pOld==pOpen );
sqliteFree(pOpen);
releaseLockInfo(pLock);
return 1;
}
}else{
pOpen->nRef++;
}
*ppOpen = pOpen;
return 0;
}
/*
** Delete the named file
*/
int sqlite3OsDelete(const char *zFilename){
unlink(zFilename);
return SQLITE_OK;
}
/*
** Return TRUE if the named file exists.
*/
int sqlite3OsFileExists(const char *zFilename){
return access(zFilename, 0)==0;
}
/*
** Attempt to open a file for both reading and writing. If that
** fails, try opening it read-only. If the file does not exist,
** try to create it.
**
** On success, a handle for the open file is written to *id
** and *pReadonly is set to 0 if the file was opened for reading and
** writing or 1 if the file was opened read-only. The function returns
** SQLITE_OK.
**
** On failure, the function returns SQLITE_CANTOPEN and leaves
** *id and *pReadonly unchanged.
*/
int sqlite3OsOpenReadWrite(
const char *zFilename,
OsFile *id,
int *pReadonly,
int exclusiveFlag,
int allowReadonly
){
int rc;
assert( !id->isOpen );
id->dirfd = -1;
SET_THREADID(id);
id->h = open(zFilename, O_RDWR|O_CREAT|O_LARGEFILE|O_BINARY,
SQLITE_DEFAULT_FILE_PERMISSIONS);
if( id->h<0 ){
#ifdef EISDIR
if( errno==EISDIR ){
return SQLITE_CANTOPEN;
}
#endif
id->h = open(zFilename, O_RDONLY|O_LARGEFILE|O_BINARY);
if( id->h<0 ){
return SQLITE_CANTOPEN;
}
*pReadonly = 1;
}else{
*pReadonly = 0;
}
sqlite3OsEnterMutex();
rc = findLockInfo(id->h, &id->pLock, &id->pOpen);
sqlite3OsLeaveMutex();
if( rc ){
close(id->h);
return SQLITE_NOMEM;
}
id->locktype = 0;
id->isOpen = 1;
TRACE3("OPEN %-3d %s\n", id->h, zFilename);
OpenCounter(+1);
return SQLITE_OK;
}
/*
** Attempt to open a new file for exclusive access by this process.
** The file will be opened for both reading and writing. To avoid
** a potential security problem, we do not allow the file to have
** previously existed. Nor do we allow the file to be a symbolic
** link.
**
** If delFlag is true, then make arrangements to automatically delete
** the file when it is closed.
**
** On success, write the file handle into *id and return SQLITE_OK.
**
** On failure, return SQLITE_CANTOPEN.
*/
int sqlite3OsOpenExclusive(const char *zFilename, OsFile *id, int delFlag){
int rc;
assert( !id->isOpen );
if( access(zFilename, 0)==0 ){
return SQLITE_CANTOPEN;
}
SET_THREADID(id);
id->dirfd = -1;
id->h = open(zFilename,
O_RDWR|O_CREAT|O_EXCL|O_NOFOLLOW|O_LARGEFILE|O_BINARY,
SQLITE_DEFAULT_FILE_PERMISSIONS);
if( id->h<0 ){
return SQLITE_CANTOPEN;
}
sqlite3OsEnterMutex();
rc = findLockInfo(id->h, &id->pLock, &id->pOpen);
sqlite3OsLeaveMutex();
if( rc ){
close(id->h);
unlink(zFilename);
return SQLITE_NOMEM;
}
id->locktype = 0;
id->isOpen = 1;
if( delFlag ){
unlink(zFilename);
}
TRACE3("OPEN-EX %-3d %s\n", id->h, zFilename);
OpenCounter(+1);
return SQLITE_OK;
}
/*
** Attempt to open a new file for read-only access.
**
** On success, write the file handle into *id and return SQLITE_OK.
**
** On failure, return SQLITE_CANTOPEN.
*/
int sqlite3OsOpenReadOnly(const char *zFilename, OsFile *id){
int rc;
assert( !id->isOpen );
SET_THREADID(id);
id->dirfd = -1;
id->h = open(zFilename, O_RDONLY|O_LARGEFILE|O_BINARY);
if( id->h<0 ){
return SQLITE_CANTOPEN;
}
sqlite3OsEnterMutex();
rc = findLockInfo(id->h, &id->pLock, &id->pOpen);
sqlite3OsLeaveMutex();
if( rc ){
close(id->h);
return SQLITE_NOMEM;
}
id->locktype = 0;
id->isOpen = 1;
TRACE3("OPEN-RO %-3d %s\n", id->h, zFilename);
OpenCounter(+1);
return SQLITE_OK;
}
/*
** Attempt to open a file descriptor for the directory that contains a
** file. This file descriptor can be used to fsync() the directory
** in order to make sure the creation of a new file is actually written
** to disk.
**
** This routine is only meaningful for Unix. It is a no-op under
** windows since windows does not support hard links.
**
** On success, a handle for a previously open file is at *id is
** updated with the new directory file descriptor and SQLITE_OK is
** returned.
**
** On failure, the function returns SQLITE_CANTOPEN and leaves
** *id unchanged.
*/
int sqlite3OsOpenDirectory(
const char *zDirname,
OsFile *id
){
if( !id->isOpen ){
/* Do not open the directory if the corresponding file is not already
** open. */
return SQLITE_CANTOPEN;
}
SET_THREADID(id);
assert( id->dirfd<0 );
id->dirfd = open(zDirname, O_RDONLY|O_BINARY, 0);
if( id->dirfd<0 ){
return SQLITE_CANTOPEN;
}
TRACE3("OPENDIR %-3d %s\n", id->dirfd, zDirname);
return SQLITE_OK;
}
/*
** If the following global variable points to a string which is the
** name of a directory, then that directory will be used to store
** temporary files.
*/
char *sqlite3_temp_directory = 0;
/*
** Create a temporary file name in zBuf. zBuf must be big enough to
** hold at least SQLITE_TEMPNAME_SIZE characters.
*/
int sqlite3OsTempFileName(char *zBuf){
static const char *azDirs[] = {
0,
"/var/tmp",
"/usr/tmp",
"/tmp",
".",
};
static const unsigned char zChars[] =
"abcdefghijklmnopqrstuvwxyz"
"ABCDEFGHIJKLMNOPTQRSTUVWXYZ"
"0123456789";
int i, j;
struct stat buf;
const char *zDir = ".";
azDirs[0] = sqlite3_temp_directory;
for(i=0; i<sizeof(azDirs)/sizeof(azDirs[0]); i++){
if( azDirs[i]==0 ) continue;
if( stat(azDirs[i], &buf) ) continue;
if( !S_ISDIR(buf.st_mode) ) continue;
if( access(azDirs[i], 07) ) continue;
zDir = azDirs[i];
break;
}
do{
sprintf(zBuf, "%s/"TEMP_FILE_PREFIX, zDir);
j = strlen(zBuf);
sqlite3Randomness(15, &zBuf[j]);
for(i=0; i<15; i++, j++){
zBuf[j] = (char)zChars[ ((unsigned char)zBuf[j])%(sizeof(zChars)-1) ];
}
zBuf[j] = 0;
}while( access(zBuf,0)==0 );
return SQLITE_OK;
}
#ifndef SQLITE_OMIT_PAGER_PRAGMAS
/*
** Check that a given pathname is a directory and is writable
**
*/
int sqlite3OsIsDirWritable(char *zBuf){
struct stat buf;
if( zBuf==0 ) return 0;
if( zBuf[0]==0 ) return 0;
if( stat(zBuf, &buf) ) return 0;
if( !S_ISDIR(buf.st_mode) ) return 0;
if( access(zBuf, 07) ) return 0;
return 1;
}
#endif /* SQLITE_OMIT_PAGER_PRAGMAS */
/*
** Read data from a file into a buffer. Return SQLITE_OK if all
** bytes were read successfully and SQLITE_IOERR if anything goes
** wrong.
*/
int sqlite3OsRead(OsFile *id, void *pBuf, int amt){
int got;
assert( id->isOpen );
SimulateIOError(SQLITE_IOERR);
TIMER_START;
got = read(id->h, pBuf, amt);
TIMER_END;
TRACE5("READ %-3d %5d %7d %d\n", id->h, got, last_page, TIMER_ELAPSED);
SEEK(0);
/* if( got<0 ) got = 0; */
if( got==amt ){
return SQLITE_OK;
}else{
return SQLITE_IOERR;
}
}
/*
** Write data from a buffer into a file. Return SQLITE_OK on success
** or some other error code on failure.
*/
int sqlite3OsWrite(OsFile *id, const void *pBuf, int amt){
int wrote = 0;
assert( id->isOpen );
assert( amt>0 );
SimulateIOError(SQLITE_IOERR);
SimulateDiskfullError;
TIMER_START;
while( amt>0 && (wrote = write(id->h, pBuf, amt))>0 ){
amt -= wrote;
pBuf = &((char*)pBuf)[wrote];
}
TIMER_END;
TRACE5("WRITE %-3d %5d %7d %d\n", id->h, wrote, last_page, TIMER_ELAPSED);
SEEK(0);
if( amt>0 ){
return SQLITE_FULL;
}
return SQLITE_OK;
}
/*
** Move the read/write pointer in a file.
*/
int sqlite3OsSeek(OsFile *id, i64 offset){
assert( id->isOpen );
SEEK(offset/1024 + 1);
#ifdef SQLITE_TEST
if( offset ) SimulateDiskfullError
#endif
lseek(id->h, offset, SEEK_SET);
return SQLITE_OK;
}
#ifdef SQLITE_TEST
/*
** Count the number of fullsyncs and normal syncs. This is used to test
** that syncs and fullsyncs are occuring at the right times.
*/
int sqlite3_sync_count = 0;
int sqlite3_fullsync_count = 0;
#endif
/*
** The fsync() system call does not work as advertised on many
** unix systems. The following procedure is an attempt to make
** it work better.
**
** The SQLITE_NO_SYNC macro disables all fsync()s. This is useful
** for testing when we want to run through the test suite quickly.
** You are strongly advised *not* to deploy with SQLITE_NO_SYNC
** enabled, however, since with SQLITE_NO_SYNC enabled, an OS crash
** or power failure will likely corrupt the database file.
*/
static int full_fsync(int fd, int fullSync, int dataOnly){
int rc;
/* Record the number of times that we do a normal fsync() and
** FULLSYNC. This is used during testing to verify that this procedure
** gets called with the correct arguments.
*/
#ifdef SQLITE_TEST
if( fullSync ) sqlite3_fullsync_count++;
sqlite3_sync_count++;
#endif
/* If we compiled with the SQLITE_NO_SYNC flag, then syncing is a
** no-op
*/
#ifdef SQLITE_NO_SYNC
rc = SQLITE_OK;
#else
#ifdef F_FULLFSYNC
if( fullSync ){
rc = fcntl(fd, F_FULLFSYNC, 0);
}else{
rc = 1;
}
/* If the FULLSYNC failed, try to do a normal fsync() */
if( rc ) rc = fsync(fd);
#else /* if !defined(F_FULLSYNC) */
#if defined(_POSIX_SYNCHRONIZED_IO) && _POSIX_SYNCHRONIZED_IO>0
if( dataOnly ){
rc = fdatasync(fd);
}else
#endif /* _POSIX_SYNCHRONIZED_IO > 0 */
{
rc = fsync(fd);
}
#endif /* defined(F_FULLFSYNC) */
#endif /* defined(SQLITE_NO_SYNC) */
return rc;
}
/*
** Make sure all writes to a particular file are committed to disk.
**
** If dataOnly==0 then both the file itself and its metadata (file
** size, access time, etc) are synced. If dataOnly!=0 then only the
** file data is synced.
**
** Under Unix, also make sure that the directory entry for the file
** has been created by fsync-ing the directory that contains the file.
** If we do not do this and we encounter a power failure, the directory
** entry for the journal might not exist after we reboot. The next
** SQLite to access the file will not know that the journal exists (because
** the directory entry for the journal was never created) and the transaction
** will not roll back - possibly leading to database corruption.
*/
int sqlite3OsSync(OsFile *id, int dataOnly){
assert( id->isOpen );
SimulateIOError(SQLITE_IOERR);
TRACE2("SYNC %-3d\n", id->h);
if( full_fsync(id->h, id->fullSync, dataOnly) ){
return SQLITE_IOERR;
}
if( id->dirfd>=0 ){
TRACE2("DIRSYNC %-3d\n", id->dirfd);
full_fsync(id->dirfd, id->fullSync, 0);
close(id->dirfd); /* Only need to sync once, so close the directory */
id->dirfd = -1; /* when we are done. */
}
return SQLITE_OK;
}
/*
** Sync the directory zDirname. This is a no-op on operating systems other
** than UNIX.
**
** This is used to make sure the master journal file has truely been deleted
** before making changes to individual journals on a multi-database commit.
** The F_FULLFSYNC option is not needed here.
*/
int sqlite3OsSyncDirectory(const char *zDirname){
int fd;
int r;
SimulateIOError(SQLITE_IOERR);
fd = open(zDirname, O_RDONLY|O_BINARY, 0);
TRACE3("DIRSYNC %-3d (%s)\n", fd, zDirname);
if( fd<0 ){
return SQLITE_CANTOPEN;
}
r = fsync(fd);
close(fd);
return ((r==0)?SQLITE_OK:SQLITE_IOERR);
}
/*
** Truncate an open file to a specified size
*/
int sqlite3OsTruncate(OsFile *id, i64 nByte){
assert( id->isOpen );
SimulateIOError(SQLITE_IOERR);
return ftruncate(id->h, nByte)==0 ? SQLITE_OK : SQLITE_IOERR;
}
/*
** Determine the current size of a file in bytes
*/
int sqlite3OsFileSize(OsFile *id, i64 *pSize){
struct stat buf;
assert( id->isOpen );
SimulateIOError(SQLITE_IOERR);
if( fstat(id->h, &buf)!=0 ){
return SQLITE_IOERR;
}
*pSize = buf.st_size;
return SQLITE_OK;
}
/*
** This routine checks if there is a RESERVED lock held on the specified
** file by this or any other process. If such a lock is held, return
** non-zero. If the file is unlocked or holds only SHARED locks, then
** return zero.
*/
int sqlite3OsCheckReservedLock(OsFile *id){
int r = 0;
assert( id->isOpen );
if( CHECK_THREADID(id) ) return SQLITE_MISUSE;
sqlite3OsEnterMutex(); /* Needed because id->pLock is shared across threads */
/* Check if a thread in this process holds such a lock */
if( id->pLock->locktype>SHARED_LOCK ){
r = 1;
}
/* Otherwise see if some other process holds it.
*/
if( !r ){
struct flock lock;
lock.l_whence = SEEK_SET;
lock.l_start = RESERVED_BYTE;
lock.l_len = 1;
lock.l_type = F_WRLCK;
fcntl(id->h, F_GETLK, &lock);
if( lock.l_type!=F_UNLCK ){
r = 1;
}
}
sqlite3OsLeaveMutex();
TRACE3("TEST WR-LOCK %d %d\n", id->h, r);
return r;
}
#ifdef SQLITE_DEBUG
/*
** Helper function for printing out trace information from debugging
** binaries. This returns the string represetation of the supplied
** integer lock-type.
*/
static const char * locktypeName(int locktype){
switch( locktype ){
case NO_LOCK: return "NONE";
case SHARED_LOCK: return "SHARED";
case RESERVED_LOCK: return "RESERVED";
case PENDING_LOCK: return "PENDING";
case EXCLUSIVE_LOCK: return "EXCLUSIVE";
}
return "ERROR";
}
#endif
/*
** Lock the file with the lock specified by parameter locktype - one
** of the following:
**
** (1) SHARED_LOCK
** (2) RESERVED_LOCK
** (3) PENDING_LOCK
** (4) EXCLUSIVE_LOCK
**
** Sometimes when requesting one lock state, additional lock states
** are inserted in between. The locking might fail on one of the later
** transitions leaving the lock state different from what it started but
** still short of its goal. The following chart shows the allowed
** transitions and the inserted intermediate states:
**
** UNLOCKED -> SHARED
** SHARED -> RESERVED
** SHARED -> (PENDING) -> EXCLUSIVE
** RESERVED -> (PENDING) -> EXCLUSIVE
** PENDING -> EXCLUSIVE
**
** This routine will only increase a lock. Use the sqlite3OsUnlock()
** routine to lower a locking level.
*/
int sqlite3OsLock(OsFile *id, int locktype){
/* The following describes the implementation of the various locks and
** lock transitions in terms of the POSIX advisory shared and exclusive
** lock primitives (called read-locks and write-locks below, to avoid
** confusion with SQLite lock names). The algorithms are complicated
** slightly in order to be compatible with windows systems simultaneously
** accessing the same database file, in case that is ever required.
**
** Symbols defined in os.h indentify the 'pending byte' and the 'reserved
** byte', each single bytes at well known offsets, and the 'shared byte
** range', a range of 510 bytes at a well known offset.
**
** To obtain a SHARED lock, a read-lock is obtained on the 'pending
** byte'. If this is successful, a random byte from the 'shared byte
** range' is read-locked and the lock on the 'pending byte' released.
**
** A process may only obtain a RESERVED lock after it has a SHARED lock.
** A RESERVED lock is implemented by grabbing a write-lock on the
** 'reserved byte'.
**
** A process may only obtain a PENDING lock after it has obtained a
** SHARED lock. A PENDING lock is implemented by obtaining a write-lock
** on the 'pending byte'. This ensures that no new SHARED locks can be
** obtained, but existing SHARED locks are allowed to persist. A process
** does not have to obtain a RESERVED lock on the way to a PENDING lock.
** This property is used by the algorithm for rolling back a journal file
** after a crash.
**
** An EXCLUSIVE lock, obtained after a PENDING lock is held, is
** implemented by obtaining a write-lock on the entire 'shared byte
** range'. Since all other locks require a read-lock on one of the bytes
** within this range, this ensures that no other locks are held on the
** database.
**
** The reason a single byte cannot be used instead of the 'shared byte
** range' is that some versions of windows do not support read-locks. By
** locking a random byte from a range, concurrent SHARED locks may exist
** even if the locking primitive used is always a write-lock.
*/
int rc = SQLITE_OK;
struct lockInfo *pLock = id->pLock;
struct flock lock;
int s;
assert( id->isOpen );
TRACE7("LOCK %d %s was %s(%s,%d) pid=%d\n", id->h, locktypeName(locktype),
locktypeName(id->locktype), locktypeName(pLock->locktype), pLock->cnt
,getpid() );
if( CHECK_THREADID(id) ) return SQLITE_MISUSE;
/* If there is already a lock of this type or more restrictive on the
** OsFile, do nothing. Don't use the end_lock: exit path, as
** sqlite3OsEnterMutex() hasn't been called yet.
*/
if( id->locktype>=locktype ){
TRACE3("LOCK %d %s ok (already held)\n", id->h, locktypeName(locktype));
return SQLITE_OK;
}
/* Make sure the locking sequence is correct
*/
assert( id->locktype!=NO_LOCK || locktype==SHARED_LOCK );
assert( locktype!=PENDING_LOCK );
assert( locktype!=RESERVED_LOCK || id->locktype==SHARED_LOCK );
/* This mutex is needed because id->pLock is shared across threads
*/
sqlite3OsEnterMutex();
/* If some thread using this PID has a lock via a different OsFile*
** handle that precludes the requested lock, return BUSY.
*/
if( (id->locktype!=pLock->locktype &&
(pLock->locktype>=PENDING_LOCK || locktype>SHARED_LOCK))
){
rc = SQLITE_BUSY;
goto end_lock;
}
/* If a SHARED lock is requested, and some thread using this PID already
** has a SHARED or RESERVED lock, then increment reference counts and
** return SQLITE_OK.
*/
if( locktype==SHARED_LOCK &&
(pLock->locktype==SHARED_LOCK || pLock->locktype==RESERVED_LOCK) ){
assert( locktype==SHARED_LOCK );
assert( id->locktype==0 );
assert( pLock->cnt>0 );
id->locktype = SHARED_LOCK;
pLock->cnt++;
id->pOpen->nLock++;
goto end_lock;
}
lock.l_len = 1L;
lock.l_whence = SEEK_SET;
/* A PENDING lock is needed before acquiring a SHARED lock and before
** acquiring an EXCLUSIVE lock. For the SHARED lock, the PENDING will
** be released.
*/
if( locktype==SHARED_LOCK
|| (locktype==EXCLUSIVE_LOCK && id->locktype<PENDING_LOCK)
){
lock.l_type = (locktype==SHARED_LOCK?F_RDLCK:F_WRLCK);
lock.l_start = PENDING_BYTE;
s = fcntl(id->h, F_SETLK, &lock);
if( s ){
rc = (errno==EINVAL) ? SQLITE_NOLFS : SQLITE_BUSY;
goto end_lock;
}
}
/* If control gets to this point, then actually go ahead and make
** operating system calls for the specified lock.
*/
if( locktype==SHARED_LOCK ){
assert( pLock->cnt==0 );
assert( pLock->locktype==0 );
/* Now get the read-lock */
lock.l_start = SHARED_FIRST;
lock.l_len = SHARED_SIZE;
s = fcntl(id->h, F_SETLK, &lock);
/* Drop the temporary PENDING lock */
lock.l_start = PENDING_BYTE;
lock.l_len = 1L;
lock.l_type = F_UNLCK;
if( fcntl(id->h, F_SETLK, &lock)!=0 ){
rc = SQLITE_IOERR; /* This should never happen */
goto end_lock;
}
if( s ){
rc = (errno==EINVAL) ? SQLITE_NOLFS : SQLITE_BUSY;
}else{
id->locktype = SHARED_LOCK;
id->pOpen->nLock++;
pLock->cnt = 1;
}
}else if( locktype==EXCLUSIVE_LOCK && pLock->cnt>1 ){
/* We are trying for an exclusive lock but another thread in this
** same process is still holding a shared lock. */
rc = SQLITE_BUSY;
}else{
/* The request was for a RESERVED or EXCLUSIVE lock. It is
** assumed that there is a SHARED or greater lock on the file
** already.
*/
assert( 0!=id->locktype );
lock.l_type = F_WRLCK;
switch( locktype ){
case RESERVED_LOCK:
lock.l_start = RESERVED_BYTE;
break;
case EXCLUSIVE_LOCK:
lock.l_start = SHARED_FIRST;
lock.l_len = SHARED_SIZE;
break;
default:
assert(0);
}
s = fcntl(id->h, F_SETLK, &lock);
if( s ){
rc = (errno==EINVAL) ? SQLITE_NOLFS : SQLITE_BUSY;
}
}
if( rc==SQLITE_OK ){
id->locktype = locktype;
pLock->locktype = locktype;
}else if( locktype==EXCLUSIVE_LOCK ){
id->locktype = PENDING_LOCK;
pLock->locktype = PENDING_LOCK;
}
end_lock:
sqlite3OsLeaveMutex();
TRACE4("LOCK %d %s %s\n", id->h, locktypeName(locktype),
rc==SQLITE_OK ? "ok" : "failed");
return rc;
}
/*
** Lower the locking level on file descriptor id to locktype. locktype
** must be either NO_LOCK or SHARED_LOCK.
**
** If the locking level of the file descriptor is already at or below
** the requested locking level, this routine is a no-op.
**
** It is not possible for this routine to fail if the second argument
** is NO_LOCK. If the second argument is SHARED_LOCK, this routine
** might return SQLITE_IOERR instead of SQLITE_OK.
*/
int sqlite3OsUnlock(OsFile *id, int locktype){
struct lockInfo *pLock;
struct flock lock;
int rc = SQLITE_OK;
assert( id->isOpen );
TRACE7("UNLOCK %d %d was %d(%d,%d) pid=%d\n", id->h, locktype, id->locktype,
id->pLock->locktype, id->pLock->cnt, getpid());
if( CHECK_THREADID(id) ) return SQLITE_MISUSE;
assert( locktype<=SHARED_LOCK );
if( id->locktype<=locktype ){
return SQLITE_OK;
}
sqlite3OsEnterMutex();
pLock = id->pLock;
assert( pLock->cnt!=0 );
if( id->locktype>SHARED_LOCK ){
assert( pLock->locktype==id->locktype );
if( locktype==SHARED_LOCK ){
lock.l_type = F_RDLCK;
lock.l_whence = SEEK_SET;
lock.l_start = SHARED_FIRST;
lock.l_len = SHARED_SIZE;
if( fcntl(id->h, F_SETLK, &lock)!=0 ){
/* This should never happen */
rc = SQLITE_IOERR;
}
}
lock.l_type = F_UNLCK;
lock.l_whence = SEEK_SET;
lock.l_start = PENDING_BYTE;
lock.l_len = 2L; assert( PENDING_BYTE+1==RESERVED_BYTE );
if( fcntl(id->h, F_SETLK, &lock)==0 ){
pLock->locktype = SHARED_LOCK;
}else{
rc = SQLITE_IOERR; /* This should never happen */
}
}
if( locktype==NO_LOCK ){
struct openCnt *pOpen;
/* Decrement the shared lock counter. Release the lock using an
** OS call only when all threads in this same process have released
** the lock.
*/
pLock->cnt--;
if( pLock->cnt==0 ){
lock.l_type = F_UNLCK;
lock.l_whence = SEEK_SET;
lock.l_start = lock.l_len = 0L;
if( fcntl(id->h, F_SETLK, &lock)==0 ){
pLock->locktype = NO_LOCK;
}else{
rc = SQLITE_IOERR; /* This should never happen */
}
}
/* Decrement the count of locks against this same file. When the
** count reaches zero, close any other file descriptors whose close
** was deferred because of outstanding locks.
*/
pOpen = id->pOpen;
pOpen->nLock--;
assert( pOpen->nLock>=0 );
if( pOpen->nLock==0 && pOpen->nPending>0 ){
int i;
for(i=0; i<pOpen->nPending; i++){
close(pOpen->aPending[i]);
}
sqliteFree(pOpen->aPending);
pOpen->nPending = 0;
pOpen->aPending = 0;
}
}
sqlite3OsLeaveMutex();
id->locktype = locktype;
return rc;
}
/*
** Close a file.
*/
int sqlite3OsClose(OsFile *id){
if( !id->isOpen ) return SQLITE_OK;
if( CHECK_THREADID(id) ) return SQLITE_MISUSE;
sqlite3OsUnlock(id, NO_LOCK);
if( id->dirfd>=0 ) close(id->dirfd);
id->dirfd = -1;
sqlite3OsEnterMutex();
if( id->pOpen->nLock ){
/* If there are outstanding locks, do not actually close the file just
** yet because that would clear those locks. Instead, add the file
** descriptor to pOpen->aPending. It will be automatically closed when
** the last lock is cleared.
*/
int *aNew;
struct openCnt *pOpen = id->pOpen;
aNew = sqliteRealloc( pOpen->aPending, (pOpen->nPending+1)*sizeof(int) );
if( aNew==0 ){
/* If a malloc fails, just leak the file descriptor */
}else{
pOpen->aPending = aNew;
pOpen->aPending[pOpen->nPending] = id->h;
pOpen->nPending++;
}
}else{
/* There are no outstanding locks so we can close the file immediately */
close(id->h);
}
releaseLockInfo(id->pLock);
releaseOpenCnt(id->pOpen);
sqlite3OsLeaveMutex();
id->isOpen = 0;
TRACE2("CLOSE %-3d\n", id->h);
OpenCounter(-1);
return SQLITE_OK;
}
/*
** Turn a relative pathname into a full pathname. Return a pointer
** to the full pathname stored in space obtained from sqliteMalloc().
** The calling function is responsible for freeing this space once it
** is no longer needed.
*/
char *sqlite3OsFullPathname(const char *zRelative){
char *zFull = 0;
if( zRelative[0]=='/' ){
sqlite3SetString(&zFull, zRelative, (char*)0);
}else{
char *zBuf = sqliteMalloc(5000);
if( zBuf==0 ){
return 0;
}
zBuf[0] = 0;
sqlite3SetString(&zFull, getcwd(zBuf, 5000), "/", zRelative,
(char*)0);
sqliteFree(zBuf);
}
return zFull;
}
#endif /* SQLITE_OMIT_DISKIO */
/***************************************************************************
** Everything above deals with file I/O. Everything that follows deals
** with other miscellanous aspects of the operating system interface
****************************************************************************/
/*
** Get information to seed the random number generator. The seed
** is written into the buffer zBuf[256]. The calling function must
** supply a sufficiently large buffer.
*/
int sqlite3OsRandomSeed(char *zBuf){
/* We have to initialize zBuf to prevent valgrind from reporting
** errors. The reports issued by valgrind are incorrect - we would
** prefer that the randomness be increased by making use of the
** uninitialized space in zBuf - but valgrind errors tend to worry
** some users. Rather than argue, it seems easier just to initialize
** the whole array and silence valgrind, even if that means less randomness
** in the random seed.
**
** When testing, initializing zBuf[] to zero is all we do. That means
** that we always use the same random number sequence.* This makes the
** tests repeatable.
*/
memset(zBuf, 0, 256);
#if !defined(SQLITE_TEST)
{
int pid, fd;
fd = open("/dev/urandom", O_RDONLY);
if( fd<0 ){
time((time_t*)zBuf);
pid = getpid();
memcpy(&zBuf[sizeof(time_t)], &pid, sizeof(pid));
}else{
read(fd, zBuf, 256);
close(fd);
}
}
#endif
return SQLITE_OK;
}
/*
** Sleep for a little while. Return the amount of time slept.
*/
int sqlite3OsSleep(int ms){
#if defined(HAVE_USLEEP) && HAVE_USLEEP
usleep(ms*1000);
return ms;
#else
sleep((ms+999)/1000);
return 1000*((ms+999)/1000);
#endif
}
/*
** Static variables used for thread synchronization
*/
static int inMutex = 0;
#ifdef SQLITE_UNIX_THREADS
static pthread_mutex_t mutex = PTHREAD_MUTEX_INITIALIZER;
#endif
/*
** The following pair of routine implement mutual exclusion for
** multi-threaded processes. Only a single thread is allowed to
** executed code that is surrounded by EnterMutex() and LeaveMutex().
**
** SQLite uses only a single Mutex. There is not much critical
** code and what little there is executes quickly and without blocking.
*/
void sqlite3OsEnterMutex(){
#ifdef SQLITE_UNIX_THREADS
pthread_mutex_lock(&mutex);
#endif
assert( !inMutex );
inMutex = 1;
}
void sqlite3OsLeaveMutex(){
assert( inMutex );
inMutex = 0;
#ifdef SQLITE_UNIX_THREADS
pthread_mutex_unlock(&mutex);
#endif
}
/*
** The following variable, if set to a non-zero value, becomes the result
** returned from sqlite3OsCurrentTime(). This is used for testing.
*/
#ifdef SQLITE_TEST
int sqlite3_current_time = 0;
#endif
/*
** Find the current time (in Universal Coordinated Time). Write the
** current time and date as a Julian Day number into *prNow and
** return 0. Return 1 if the time and date cannot be found.
*/
int sqlite3OsCurrentTime(double *prNow){
#ifdef NO_GETTOD
time_t t;
time(&t);
*prNow = t/86400.0 + 2440587.5;
#else
struct timeval sNow;
struct timezone sTz; /* Not used */
gettimeofday(&sNow, &sTz);
*prNow = 2440587.5 + sNow.tv_sec/86400.0 + sNow.tv_usec/86400000000.0;
#endif
#ifdef SQLITE_TEST
if( sqlite3_current_time ){
*prNow = sqlite3_current_time/86400.0 + 2440587.5;
}
#endif
return 0;
}
#endif /* OS_UNIX */
|