forked from xen-project/xen
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtdb.c
1748 lines (1512 loc) · 46.4 KB
/
tdb.c
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
/*
Unix SMB/CIFS implementation.
trivial database library
Copyright (C) Andrew Tridgell 1999-2004
Copyright (C) Paul `Rusty' Russell 2000
Copyright (C) Jeremy Allison 2000-2003
** NOTE! The following LGPL license applies to the tdb
** library. This does NOT imply that all of Samba is released
** under the LGPL
This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2 of the License, or (at your option) any later version.
This library 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
Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public
License along with this library; If not, see <http://www.gnu.org/licenses/>.
*/
#ifndef _SAMBA_BUILD_
#ifdef HAVE_CONFIG_H
#include <config.h>
#endif
#include <stdlib.h>
#include <stdio.h>
#include <stdint.h>
#include <fcntl.h>
#include <unistd.h>
#include <string.h>
#include <fcntl.h>
#include <errno.h>
#include <sys/mman.h>
#include <sys/stat.h>
#include "tdb.h"
#include <stdarg.h>
#include "talloc.h"
#undef HAVE_MMAP
#else
#include "includes.h"
#include "lib/tdb/include/tdb.h"
#include "system/time.h"
#include "system/shmem.h"
#include "system/filesys.h"
#endif
#define TDB_MAGIC_FOOD "TDB file\n"
#define TDB_VERSION (0x26011967 + 7)
#define TDB_MAGIC (0x26011999U)
#define TDB_FREE_MAGIC (~TDB_MAGIC)
#define TDB_DEAD_MAGIC (0xFEE1DEAD)
#define TDB_ALIGNMENT 4
#define MIN_REC_SIZE (2*sizeof(struct list_struct) + TDB_ALIGNMENT)
#define DEFAULT_HASH_SIZE 131
#define TDB_PAGE_SIZE 0x2000
#define FREELIST_TOP (sizeof(struct tdb_header))
#define TDB_ALIGN(x,a) (((x) + (a)-1) & ~((a)-1))
#define TDB_BYTEREV(x) (((((x)&0xff)<<24)|((x)&0xFF00)<<8)|(((x)>>8)&0xFF00)|((x)>>24))
#define TDB_DEAD(r) ((r)->magic == TDB_DEAD_MAGIC)
#define TDB_BAD_MAGIC(r) ((r)->magic != TDB_MAGIC && !TDB_DEAD(r))
#define TDB_HASH_TOP(hash) (FREELIST_TOP + (BUCKET(hash)+1)*sizeof(tdb_off))
#define TDB_DATA_START(hash_size) (TDB_HASH_TOP(hash_size-1))
/* NB assumes there is a local variable called "tdb" that is the
* current context, also takes doubly-parenthesized print-style
* argument. */
#define TDB_LOG(x) tdb->log_fn x
/* lock offsets */
#define GLOBAL_LOCK 0
#define ACTIVE_LOCK 4
#ifndef MAP_FILE
#define MAP_FILE 0
#endif
#ifndef MAP_FAILED
#define MAP_FAILED ((void *)-1)
#endif
#ifndef discard_const_p
# if defined(__intptr_t_defined) || defined(HAVE_INTPTR_T)
# define discard_const(ptr) ((void *)((intptr_t)(ptr)))
# else
# define discard_const(ptr) ((void *)(ptr))
# endif
# define discard_const_p(type, ptr) ((type *)discard_const(ptr))
#endif
/* free memory if the pointer is valid and zero the pointer */
#ifndef SAFE_FREE
#define SAFE_FREE(x) do { if ((x) != NULL) {talloc_free(discard_const_p(void *, (x))); (x)=NULL;} } while(0)
#endif
#define BUCKET(hash) ((hash) % tdb->header.hash_size)
static TDB_DATA tdb_null;
/* all contexts, to ensure no double-opens (fcntl locks don't nest!) */
static TDB_CONTEXT *tdbs = NULL;
static int tdb_munmap(TDB_CONTEXT *tdb)
{
if (tdb->flags & TDB_INTERNAL)
return 0;
#ifdef HAVE_MMAP
if (tdb->map_ptr) {
int ret = munmap(tdb->map_ptr, tdb->map_size);
if (ret != 0)
return ret;
}
#endif
tdb->map_ptr = NULL;
return 0;
}
static void tdb_mmap(TDB_CONTEXT *tdb)
{
if (tdb->flags & TDB_INTERNAL)
return;
#ifdef HAVE_MMAP
if (!(tdb->flags & TDB_NOMMAP)) {
tdb->map_ptr = mmap(NULL, tdb->map_size,
PROT_READ|(tdb->read_only? 0:PROT_WRITE),
MAP_SHARED|MAP_FILE, tdb->fd, 0);
/*
* NB. When mmap fails it returns MAP_FAILED *NOT* NULL !!!!
*/
if (tdb->map_ptr == MAP_FAILED) {
tdb->map_ptr = NULL;
TDB_LOG((tdb, 2, "tdb_mmap failed for size %d (%s)\n",
tdb->map_size, strerror(errno)));
}
} else {
tdb->map_ptr = NULL;
}
#else
tdb->map_ptr = NULL;
#endif
}
/* Endian conversion: we only ever deal with 4 byte quantities */
static void *convert(void *buf, uint32_t size)
{
uint32_t i, *p = buf;
for (i = 0; i < size / 4; i++)
p[i] = TDB_BYTEREV(p[i]);
return buf;
}
#define DOCONV() (tdb->flags & TDB_CONVERT)
#define CONVERT(x) (DOCONV() ? convert(&x, sizeof(x)) : &x)
/* the body of the database is made of one list_struct for the free space
plus a separate data list for each hash value */
struct list_struct {
tdb_off next; /* offset of the next record in the list */
tdb_len rec_len; /* total byte length of record */
tdb_len key_len; /* byte length of key */
tdb_len data_len; /* byte length of data */
uint32_t full_hash; /* the full 32 bit hash of the key */
uint32_t magic; /* try to catch errors */
/* the following union is implied:
union {
char record[rec_len];
struct {
char key[key_len];
char data[data_len];
}
uint32_t totalsize; (tailer)
}
*/
};
/* a byte range locking function - return 0 on success
this functions locks/unlocks 1 byte at the specified offset.
On error, errno is also set so that errors are passed back properly
through tdb_open(). */
static int tdb_brlock(TDB_CONTEXT *tdb, tdb_off offset,
int rw_type, int lck_type, int probe)
{
struct flock fl;
int ret;
if (tdb->flags & TDB_NOLOCK)
return 0;
if ((rw_type == F_WRLCK) && (tdb->read_only)) {
errno = EACCES;
return -1;
}
fl.l_type = rw_type;
fl.l_whence = SEEK_SET;
fl.l_start = offset;
fl.l_len = 1;
fl.l_pid = 0;
do {
ret = fcntl(tdb->fd,lck_type,&fl);
} while (ret == -1 && errno == EINTR);
if (ret == -1) {
if (!probe && lck_type != F_SETLK) {
/* Ensure error code is set for log fun to examine. */
tdb->ecode = TDB_ERR_LOCK;
TDB_LOG((tdb, 5,"tdb_brlock failed (fd=%d) at offset %d rw_type=%d lck_type=%d\n",
tdb->fd, offset, rw_type, lck_type));
}
/* Generic lock error. errno set by fcntl.
* EAGAIN is an expected return from non-blocking
* locks. */
if (errno != EAGAIN) {
TDB_LOG((tdb, 5, "tdb_brlock failed (fd=%d) at offset %d rw_type=%d lck_type=%d: %s\n",
tdb->fd, offset, rw_type, lck_type,
strerror(errno)));
}
return TDB_ERRCODE(TDB_ERR_LOCK, -1);
}
return 0;
}
/* lock a list in the database. list -1 is the alloc list */
static int tdb_lock(TDB_CONTEXT *tdb, int list, int ltype)
{
if (list < -1 || list >= (int)tdb->header.hash_size) {
TDB_LOG((tdb, 0,"tdb_lock: invalid list %d for ltype=%d\n",
list, ltype));
return -1;
}
if (tdb->flags & TDB_NOLOCK)
return 0;
/* Since fcntl locks don't nest, we do a lock for the first one,
and simply bump the count for future ones */
if (tdb->locked[list+1].count == 0) {
if (tdb_brlock(tdb,FREELIST_TOP+4*list,ltype,F_SETLKW, 0)) {
TDB_LOG((tdb, 0,"tdb_lock failed on list %d ltype=%d (%s)\n",
list, ltype, strerror(errno)));
return -1;
}
tdb->locked[list+1].ltype = ltype;
}
tdb->locked[list+1].count++;
return 0;
}
/* unlock the database: returns void because it's too late for errors. */
/* changed to return int it may be interesting to know there
has been an error --simo */
static int tdb_unlock(TDB_CONTEXT *tdb, int list,
int ltype __attribute__((unused)))
{
int ret = -1;
if (tdb->flags & TDB_NOLOCK)
return 0;
/* Sanity checks */
if (list < -1 || list >= (int)tdb->header.hash_size) {
TDB_LOG((tdb, 0, "tdb_unlock: list %d invalid (%d)\n", list, tdb->header.hash_size));
return ret;
}
if (tdb->locked[list+1].count==0) {
TDB_LOG((tdb, 0, "tdb_unlock: count is 0\n"));
return ret;
}
if (tdb->locked[list+1].count == 1) {
/* Down to last nested lock: unlock underneath */
ret = tdb_brlock(tdb, FREELIST_TOP+4*list, F_UNLCK, F_SETLKW, 0);
} else {
ret = 0;
}
tdb->locked[list+1].count--;
if (ret)
TDB_LOG((tdb, 0,"tdb_unlock: An error occurred unlocking!\n"));
return ret;
}
/* This is based on the hash algorithm from gdbm */
static uint32_t default_tdb_hash(TDB_DATA *key)
{
uint32_t value; /* Used to compute the hash value. */
uint32_t i; /* Used to cycle through random values. */
/* Set the initial value from the key size. */
for (value = 0x238F13AF * key->dsize, i=0; i < key->dsize; i++)
value = (value + (key->dptr[i] << (i*5 % 24)));
return (1103515243 * value + 12345);
}
/* check for an out of bounds access - if it is out of bounds then
see if the database has been expanded by someone else and expand
if necessary
note that "len" is the minimum length needed for the db
*/
static int tdb_oob(TDB_CONTEXT *tdb, tdb_off len, int probe)
{
struct stat st;
if (len <= tdb->map_size)
return 0;
if (tdb->flags & TDB_INTERNAL) {
if (!probe) {
/* Ensure ecode is set for log fn. */
tdb->ecode = TDB_ERR_IO;
TDB_LOG((tdb, 0,"tdb_oob len %d beyond internal malloc size %d\n",
(int)len, (int)tdb->map_size));
}
return TDB_ERRCODE(TDB_ERR_IO, -1);
}
if (fstat(tdb->fd, &st) == -1)
return TDB_ERRCODE(TDB_ERR_IO, -1);
if (st.st_size < (off_t)len) {
if (!probe) {
/* Ensure ecode is set for log fn. */
tdb->ecode = TDB_ERR_IO;
TDB_LOG((tdb, 0,"tdb_oob len %d beyond eof at %d\n",
(int)len, (int)st.st_size));
}
return TDB_ERRCODE(TDB_ERR_IO, -1);
}
/* Unmap, update size, remap */
if (tdb_munmap(tdb) == -1)
return TDB_ERRCODE(TDB_ERR_IO, -1);
tdb->map_size = st.st_size;
tdb_mmap(tdb);
return 0;
}
/* write a lump of data at a specified offset */
static int tdb_write(TDB_CONTEXT *tdb, tdb_off off, void *buf, tdb_len len)
{
if (tdb_oob(tdb, off + len, 0) != 0)
return -1;
if (tdb->map_ptr)
memcpy(off + (char *)tdb->map_ptr, buf, len);
#ifdef HAVE_PWRITE
else if (pwrite(tdb->fd, buf, len, off) != (ssize_t)len) {
#else
else if (lseek(tdb->fd, off, SEEK_SET) != (off_t)off
|| write(tdb->fd, buf, len) != (off_t)len) {
#endif
/* Ensure ecode is set for log fn. */
tdb->ecode = TDB_ERR_IO;
TDB_LOG((tdb, 0,"tdb_write failed at %d len=%d (%s)\n",
off, len, strerror(errno)));
return TDB_ERRCODE(TDB_ERR_IO, -1);
}
return 0;
}
/* read a lump of data at a specified offset, maybe convert */
static int tdb_read(TDB_CONTEXT *tdb,tdb_off off,void *buf,tdb_len len,int cv)
{
if (tdb_oob(tdb, off + len, 0) != 0)
return -1;
if (tdb->map_ptr)
memcpy(buf, off + (char *)tdb->map_ptr, len);
#ifdef HAVE_PREAD
else if (pread(tdb->fd, buf, len, off) != (off_t)len) {
#else
else if (lseek(tdb->fd, off, SEEK_SET) != (off_t)off
|| read(tdb->fd, buf, len) != (off_t)len) {
#endif
/* Ensure ecode is set for log fn. */
tdb->ecode = TDB_ERR_IO;
TDB_LOG((tdb, 0,"tdb_read failed at %d len=%d (%s)\n",
off, len, strerror(errno)));
return TDB_ERRCODE(TDB_ERR_IO, -1);
}
if (cv)
convert(buf, len);
return 0;
}
/* don't allocate memory: used in tdb_delete path. */
static int tdb_key_eq(TDB_CONTEXT *tdb, tdb_off off, TDB_DATA key)
{
char buf[64];
uint32_t len;
if (tdb_oob(tdb, off + key.dsize, 0) != 0)
return -1;
if (tdb->map_ptr)
return !memcmp(off + (char*)tdb->map_ptr, key.dptr, key.dsize);
while (key.dsize) {
len = key.dsize;
if (len > sizeof(buf))
len = sizeof(buf);
if (tdb_read(tdb, off, buf, len, 0) != 0)
return -1;
if (memcmp(buf, key.dptr, len) != 0)
return 0;
key.dptr += len;
key.dsize -= len;
off += len;
}
return 1;
}
/* read a lump of data, allocating the space for it */
static char *tdb_alloc_read(TDB_CONTEXT *tdb, tdb_off offset, tdb_len len)
{
char *buf;
if (!(buf = talloc_size(tdb, len))) {
/* Ensure ecode is set for log fn. */
tdb->ecode = TDB_ERR_OOM;
TDB_LOG((tdb, 0,"tdb_alloc_read malloc failed len=%d (%s)\n",
len, strerror(errno)));
return TDB_ERRCODE(TDB_ERR_OOM, buf);
}
if (tdb_read(tdb, offset, buf, len, 0) == -1) {
SAFE_FREE(buf);
return NULL;
}
return buf;
}
/* read/write a tdb_off */
static int ofs_read(TDB_CONTEXT *tdb, tdb_off offset, tdb_off *d)
{
return tdb_read(tdb, offset, (char*)d, sizeof(*d), DOCONV());
}
static int ofs_write(TDB_CONTEXT *tdb, tdb_off offset, tdb_off *d)
{
tdb_off off = *d;
return tdb_write(tdb, offset, CONVERT(off), sizeof(*d));
}
/* read/write a record */
static int rec_read(TDB_CONTEXT *tdb, tdb_off offset, struct list_struct *rec)
{
if (tdb_read(tdb, offset, rec, sizeof(*rec),DOCONV()) == -1)
return -1;
if (TDB_BAD_MAGIC(rec)) {
/* Ensure ecode is set for log fn. */
tdb->ecode = TDB_ERR_CORRUPT;
TDB_LOG((tdb, 0,"rec_read bad magic 0x%x at offset=%d\n", rec->magic, offset));
return TDB_ERRCODE(TDB_ERR_CORRUPT, -1);
}
return tdb_oob(tdb, rec->next+sizeof(*rec), 0);
}
static int rec_write(TDB_CONTEXT *tdb, tdb_off offset, struct list_struct *rec)
{
struct list_struct r = *rec;
return tdb_write(tdb, offset, CONVERT(r), sizeof(r));
}
/* read a freelist record and check for simple errors */
static int rec_free_read(TDB_CONTEXT *tdb, tdb_off off, struct list_struct *rec)
{
if (tdb_read(tdb, off, rec, sizeof(*rec),DOCONV()) == -1)
return -1;
if (rec->magic == TDB_MAGIC) {
/* this happens when a app is showdown while deleting a record - we should
not completely fail when this happens */
TDB_LOG((tdb, 0,"rec_free_read non-free magic 0x%x at offset=%d - fixing\n",
rec->magic, off));
rec->magic = TDB_FREE_MAGIC;
if (tdb_write(tdb, off, rec, sizeof(*rec)) == -1)
return -1;
}
if (rec->magic != TDB_FREE_MAGIC) {
/* Ensure ecode is set for log fn. */
tdb->ecode = TDB_ERR_CORRUPT;
TDB_LOG((tdb, 0,"rec_free_read bad magic 0x%x at offset=%d\n",
rec->magic, off));
return TDB_ERRCODE(TDB_ERR_CORRUPT, -1);
}
if (tdb_oob(tdb, rec->next+sizeof(*rec), 0) != 0)
return -1;
return 0;
}
/* update a record tailer (must hold allocation lock) */
static int update_tailer(TDB_CONTEXT *tdb, tdb_off offset,
const struct list_struct *rec)
{
tdb_off totalsize;
/* Offset of tailer from record header */
totalsize = sizeof(*rec) + rec->rec_len;
return ofs_write(tdb, offset + totalsize - sizeof(tdb_off),
&totalsize);
}
/* Remove an element from the freelist. Must have alloc lock. */
static int remove_from_freelist(TDB_CONTEXT *tdb, tdb_off off, tdb_off next)
{
tdb_off last_ptr, i;
/* read in the freelist top */
last_ptr = FREELIST_TOP;
while (ofs_read(tdb, last_ptr, &i) != -1 && i != 0) {
if (i == off) {
/* We've found it! */
return ofs_write(tdb, last_ptr, &next);
}
/* Follow chain (next offset is at start of record) */
last_ptr = i;
}
TDB_LOG((tdb, 0,"remove_from_freelist: not on list at off=%d\n", off));
return TDB_ERRCODE(TDB_ERR_CORRUPT, -1);
}
/* Add an element into the freelist. Merge adjacent records if
neccessary. */
static int tdb_free(TDB_CONTEXT *tdb, tdb_off offset, struct list_struct *rec)
{
tdb_off right, left;
/* Allocation and tailer lock */
if (tdb_lock(tdb, -1, F_WRLCK) != 0)
return -1;
/* set an initial tailer, so if we fail we don't leave a bogus record */
if (update_tailer(tdb, offset, rec) != 0) {
TDB_LOG((tdb, 0, "tdb_free: upfate_tailer failed!\n"));
goto fail;
}
/* Look right first (I'm an Australian, dammit) */
right = offset + sizeof(*rec) + rec->rec_len;
if (right + sizeof(*rec) <= tdb->map_size) {
struct list_struct r;
if (tdb_read(tdb, right, &r, sizeof(r), DOCONV()) == -1) {
TDB_LOG((tdb, 0, "tdb_free: right read failed at %u\n", right));
goto left;
}
/* If it's free, expand to include it. */
if (r.magic == TDB_FREE_MAGIC) {
if (remove_from_freelist(tdb, right, r.next) == -1) {
TDB_LOG((tdb, 0, "tdb_free: right free failed at %u\n", right));
goto left;
}
rec->rec_len += sizeof(r) + r.rec_len;
}
}
left:
/* Look left */
left = offset - sizeof(tdb_off);
if (left > TDB_DATA_START(tdb->header.hash_size)) {
struct list_struct l;
tdb_off leftsize;
/* Read in tailer and jump back to header */
if (ofs_read(tdb, left, &leftsize) == -1) {
TDB_LOG((tdb, 0, "tdb_free: left offset read failed at %u\n", left));
goto update;
}
left = offset - leftsize;
/* Now read in record */
if (tdb_read(tdb, left, &l, sizeof(l), DOCONV()) == -1) {
TDB_LOG((tdb, 0, "tdb_free: left read failed at %u (%u)\n", left, leftsize));
goto update;
}
/* If it's free, expand to include it. */
if (l.magic == TDB_FREE_MAGIC) {
if (remove_from_freelist(tdb, left, l.next) == -1) {
TDB_LOG((tdb, 0, "tdb_free: left free failed at %u\n", left));
goto update;
} else {
offset = left;
rec->rec_len += leftsize;
}
}
}
update:
if (update_tailer(tdb, offset, rec) == -1) {
TDB_LOG((tdb, 0, "tdb_free: update_tailer failed at %u\n", offset));
goto fail;
}
/* Now, prepend to free list */
rec->magic = TDB_FREE_MAGIC;
if (ofs_read(tdb, FREELIST_TOP, &rec->next) == -1 ||
rec_write(tdb, offset, rec) == -1 ||
ofs_write(tdb, FREELIST_TOP, &offset) == -1) {
TDB_LOG((tdb, 0, "tdb_free record write failed at offset=%d\n", offset));
goto fail;
}
/* And we're done. */
tdb_unlock(tdb, -1, F_WRLCK);
return 0;
fail:
tdb_unlock(tdb, -1, F_WRLCK);
return -1;
}
/* expand a file. we prefer to use ftruncate, as that is what posix
says to use for mmap expansion */
static int expand_file(TDB_CONTEXT *tdb, tdb_off size, tdb_off addition)
{
char buf[1024];
#ifdef HAVE_FTRUNCATE_EXTEND
if (ftruncate(tdb->fd, size+addition) != 0) {
TDB_LOG((tdb, 0, "expand_file ftruncate to %d failed (%s)\n",
size+addition, strerror(errno)));
return -1;
}
#else
char b = 0;
#ifdef HAVE_PWRITE
if (pwrite(tdb->fd, &b, 1, (size+addition) - 1) != 1) {
#else
if (lseek(tdb->fd, (size+addition) - 1, SEEK_SET) != (off_t)(size+addition) - 1 ||
write(tdb->fd, &b, 1) != 1) {
#endif
TDB_LOG((tdb, 0, "expand_file to %d failed (%s)\n",
size+addition, strerror(errno)));
return -1;
}
#endif
/* now fill the file with something. This ensures that the file isn't sparse, which would be
very bad if we ran out of disk. This must be done with write, not via mmap */
memset(buf, 0x42, sizeof(buf));
while (addition) {
int n = addition>sizeof(buf)?sizeof(buf):addition;
#ifdef HAVE_PWRITE
int ret = pwrite(tdb->fd, buf, n, size);
#else
int ret;
if (lseek(tdb->fd, size, SEEK_SET) != (off_t)size)
return -1;
ret = write(tdb->fd, buf, n);
#endif
if (ret != n) {
TDB_LOG((tdb, 0, "expand_file write of %d failed (%s)\n",
n, strerror(errno)));
return -1;
}
addition -= n;
size += n;
}
return 0;
}
/* expand the database at least size bytes by expanding the underlying
file and doing the mmap again if necessary */
static int tdb_expand(TDB_CONTEXT *tdb, tdb_off size)
{
struct list_struct rec;
tdb_off offset;
if (tdb_lock(tdb, -1, F_WRLCK) == -1) {
TDB_LOG((tdb, 0, "lock failed in tdb_expand\n"));
return -1;
}
/* must know about any previous expansions by another process */
tdb_oob(tdb, tdb->map_size + 1, 1);
/* always make room for at least 10 more records, and round
the database up to a multiple of TDB_PAGE_SIZE */
size = TDB_ALIGN(tdb->map_size + size*10, TDB_PAGE_SIZE) - tdb->map_size;
if (!(tdb->flags & TDB_INTERNAL))
tdb_munmap(tdb);
/*
* We must ensure the file is unmapped before doing this
* to ensure consistency with systems like OpenBSD where
* writes and mmaps are not consistent.
*/
/* expand the file itself */
if (!(tdb->flags & TDB_INTERNAL)) {
if (expand_file(tdb, tdb->map_size, size) != 0)
goto fail;
}
tdb->map_size += size;
if (tdb->flags & TDB_INTERNAL) {
char *new_map_ptr = talloc_realloc_size(tdb, tdb->map_ptr,
tdb->map_size);
if (!new_map_ptr) {
tdb->map_size -= size;
goto fail;
}
tdb->map_ptr = new_map_ptr;
} else {
/*
* We must ensure the file is remapped before adding the space
* to ensure consistency with systems like OpenBSD where
* writes and mmaps are not consistent.
*/
/* We're ok if the mmap fails as we'll fallback to read/write */
tdb_mmap(tdb);
}
/* form a new freelist record */
memset(&rec,'\0',sizeof(rec));
rec.rec_len = size - sizeof(rec);
/* link it into the free list */
offset = tdb->map_size - size;
if (tdb_free(tdb, offset, &rec) == -1)
goto fail;
tdb_unlock(tdb, -1, F_WRLCK);
return 0;
fail:
tdb_unlock(tdb, -1, F_WRLCK);
return -1;
}
/*
the core of tdb_allocate - called when we have decided which
free list entry to use
*/
static tdb_off tdb_allocate_ofs(TDB_CONTEXT *tdb, tdb_len length, tdb_off rec_ptr,
struct list_struct *rec, tdb_off last_ptr)
{
struct list_struct newrec;
tdb_off newrec_ptr;
memset(&newrec, '\0', sizeof(newrec));
/* found it - now possibly split it up */
if (rec->rec_len > length + MIN_REC_SIZE) {
/* Length of left piece */
length = TDB_ALIGN(length, TDB_ALIGNMENT);
/* Right piece to go on free list */
newrec.rec_len = rec->rec_len - (sizeof(*rec) + length);
newrec_ptr = rec_ptr + sizeof(*rec) + length;
/* And left record is shortened */
rec->rec_len = length;
} else {
newrec_ptr = 0;
}
/* Remove allocated record from the free list */
if (ofs_write(tdb, last_ptr, &rec->next) == -1) {
return 0;
}
/* Update header: do this before we drop alloc
lock, otherwise tdb_free() might try to
merge with us, thinking we're free.
(Thanks Jeremy Allison). */
rec->magic = TDB_MAGIC;
if (rec_write(tdb, rec_ptr, rec) == -1) {
return 0;
}
/* Did we create new block? */
if (newrec_ptr) {
/* Update allocated record tailer (we
shortened it). */
if (update_tailer(tdb, rec_ptr, rec) == -1) {
return 0;
}
/* Free new record */
if (tdb_free(tdb, newrec_ptr, &newrec) == -1) {
return 0;
}
}
/* all done - return the new record offset */
return rec_ptr;
}
/* allocate some space from the free list. The offset returned points
to a unconnected list_struct within the database with room for at
least length bytes of total data
0 is returned if the space could not be allocated
*/
static tdb_off tdb_allocate(TDB_CONTEXT *tdb, tdb_len length,
struct list_struct *rec)
{
tdb_off rec_ptr, last_ptr, newrec_ptr;
struct {
tdb_off rec_ptr, last_ptr;
tdb_len rec_len;
} bestfit = { 0, 0, 0 };
if (tdb_lock(tdb, -1, F_WRLCK) == -1)
return 0;
/* Extra bytes required for tailer */
length += sizeof(tdb_off);
again:
last_ptr = FREELIST_TOP;
/* read in the freelist top */
if (ofs_read(tdb, FREELIST_TOP, &rec_ptr) == -1)
goto fail;
bestfit.rec_ptr = 0;
/*
this is a best fit allocation strategy. Originally we used
a first fit strategy, but it suffered from massive fragmentation
issues when faced with a slowly increasing record size.
*/
while (rec_ptr) {
if (rec_free_read(tdb, rec_ptr, rec) == -1) {
goto fail;
}
if (rec->rec_len >= length) {
if (bestfit.rec_ptr == 0 ||
rec->rec_len < bestfit.rec_len) {
bestfit.rec_len = rec->rec_len;
bestfit.rec_ptr = rec_ptr;
bestfit.last_ptr = last_ptr;
/* consider a fit to be good enough if we aren't wasting more than half the space */
if (bestfit.rec_len < 2*length) {
break;
}
}
}
/* move to the next record */
last_ptr = rec_ptr;
rec_ptr = rec->next;
}
if (bestfit.rec_ptr != 0) {
if (rec_free_read(tdb, bestfit.rec_ptr, rec) == -1) {
goto fail;
}
newrec_ptr = tdb_allocate_ofs(tdb, length, bestfit.rec_ptr, rec, bestfit.last_ptr);
tdb_unlock(tdb, -1, F_WRLCK);
return newrec_ptr;
}
/* we didn't find enough space. See if we can expand the
database and if we can then try again */
if (tdb_expand(tdb, length + sizeof(*rec)) == 0)
goto again;
fail:
tdb_unlock(tdb, -1, F_WRLCK);
return 0;
}
/* initialise a new database with a specified hash size */
static int tdb_new_database(TDB_CONTEXT *tdb, int hash_size)
{
struct tdb_header *newdb;
int size, ret = -1;
/* We make it up in memory, then write it out if not internal */
size = sizeof(struct tdb_header) + (hash_size+1)*sizeof(tdb_off);
if (!(newdb = talloc_zero_size(tdb, size)))
return TDB_ERRCODE(TDB_ERR_OOM, -1);
/* Fill in the header */
newdb->version = TDB_VERSION;
newdb->hash_size = hash_size;
if (tdb->flags & TDB_INTERNAL) {
tdb->map_size = size;
tdb->map_ptr = (char *)newdb;
memcpy(&tdb->header, newdb, sizeof(tdb->header));
/* Convert the `ondisk' version if asked. */
CONVERT(*newdb);
return 0;
}
if (lseek(tdb->fd, 0, SEEK_SET) == -1)
goto fail;
if (ftruncate(tdb->fd, 0) == -1)
goto fail;
/* This creates an endian-converted header, as if read from disk */
CONVERT(*newdb);
memcpy(&tdb->header, newdb, sizeof(tdb->header));
/* Don't endian-convert the magic food! */
memcpy(newdb->magic_food, TDB_MAGIC_FOOD, strlen(TDB_MAGIC_FOOD)+1);
if (write(tdb->fd, newdb, size) != size)
ret = -1;
else
ret = 0;
fail:
SAFE_FREE(newdb);
return ret;
}
/* Returns 0 on fail. On success, return offset of record, and fills
in rec */
static tdb_off tdb_find(TDB_CONTEXT *tdb, TDB_DATA key, uint32_t hash,
struct list_struct *r)
{
tdb_off rec_ptr;
/* read in the hash top */
if (ofs_read(tdb, TDB_HASH_TOP(hash), &rec_ptr) == -1)
return 0;
/* keep looking until we find the right record */
while (rec_ptr) {
if (rec_read(tdb, rec_ptr, r) == -1)
return 0;
if (!TDB_DEAD(r) && hash==r->full_hash && key.dsize==r->key_len) {
/* a very likely hit - read the key */
int cmp = tdb_key_eq(tdb, rec_ptr + sizeof(*r), key);
if (cmp < 0)
return 0;
else if (cmp > 0)
return rec_ptr;
}
rec_ptr = r->next;
}
return TDB_ERRCODE(TDB_ERR_NOEXIST, 0);
}
/* As tdb_find, but if you succeed, keep the lock */
static tdb_off tdb_find_lock_hash(TDB_CONTEXT *tdb, TDB_DATA key, uint32_t hash, int locktype,
struct list_struct *rec)
{
uint32_t rec_ptr;
if (tdb_lock(tdb, BUCKET(hash), locktype) == -1)
return 0;
if (!(rec_ptr = tdb_find(tdb, key, hash, rec)))
tdb_unlock(tdb, BUCKET(hash), locktype);
return rec_ptr;
}
enum TDB_ERROR tdb_error(TDB_CONTEXT *tdb)
{
return tdb->ecode;
}
static struct tdb_errname {
enum TDB_ERROR ecode; const char *estring;
} emap[] = { {TDB_SUCCESS, "Success"},
{TDB_ERR_CORRUPT, "Corrupt database"},
{TDB_ERR_IO, "IO Error"},
{TDB_ERR_LOCK, "Locking error"},
{TDB_ERR_OOM, "Out of memory"},
{TDB_ERR_EXISTS, "Record exists"},
{TDB_ERR_NOLOCK, "Lock exists on other keys"},
{TDB_ERR_NOEXIST, "Record does not exist"} };
/* Error string for the last tdb error */
const char *tdb_errorstr(TDB_CONTEXT *tdb)
{
uint32_t i;
for (i = 0; i < sizeof(emap) / sizeof(struct tdb_errname); i++)
if (tdb->ecode == emap[i].ecode)
return emap[i].estring;
return "Invalid error code";
}
/* update an entry in place - this only works if the new data size
is <= the old data size and the key exists.
on failure return -1.
*/