2 Unix SMB/CIFS implementation.
4 trivial database library
6 Copyright (C) Andrew Tridgell 2005
7 Copyright (C) Rusty Russell 2010
9 ** NOTE! The following LGPL license applies to the tdb
10 ** library. This does NOT imply that all of Samba is released
13 This library is free software; you can redistribute it and/or
14 modify it under the terms of the GNU Lesser General Public
15 License as published by the Free Software Foundation; either
16 version 3 of the License, or (at your option) any later version.
18 This library is distributed in the hope that it will be useful,
19 but WITHOUT ANY WARRANTY; without even the implied warranty of
20 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
21 Lesser General Public License for more details.
23 You should have received a copy of the GNU Lesser General Public
24 License along with this library; if not, see <http://www.gnu.org/licenses/>.
28 #define SAFE_FREE(x) do { if ((x) != NULL) {free(x); (x)=NULL;} } while(0)
33 - only allow a single transaction at a time per database. This makes
34 using the transaction API simpler, as otherwise the caller would
35 have to cope with temporary failures in transactions that conflict
36 with other current transactions
38 - keep the transaction recovery information in the same file as the
39 database, using a special 'transaction recovery' record pointed at
40 by the header. This removes the need for extra journal files as
41 used by some other databases
43 - dynamically allocated the transaction recover record, re-using it
44 for subsequent transactions. If a larger record is needed then
45 tdb_free() the old record to place it on the normal tdb freelist
46 before allocating the new record
48 - during transactions, keep a linked list of writes all that have
49 been performed by intercepting all tdb_write() calls. The hooked
50 transaction versions of tdb_read() and tdb_write() check this
51 linked list and try to use the elements of the list in preference
54 - don't allow any locks to be held when a transaction starts,
55 otherwise we can end up with deadlock (plus lack of lock nesting
56 in POSIX locks would mean the lock is lost)
58 - if the caller gains a lock during the transaction but doesn't
59 release it then fail the commit
61 - allow for nested calls to tdb_transaction_start(), re-using the
62 existing transaction record. If the inner transaction is canceled
63 then a subsequent commit will fail
65 - keep a mirrored copy of the tdb hash chain heads to allow for the
66 fast hash heads scan on traverse, updating the mirrored copy in
67 the transaction version of tdb_write
69 - allow callers to mix transaction and non-transaction use of tdb,
70 although once a transaction is started then an exclusive lock is
71 gained until the transaction is committed or canceled
73 - the commit stategy involves first saving away all modified data
74 into a linearised buffer in the transaction recovery area, then
75 marking the transaction recovery area with a magic value to
76 indicate a valid recovery record. In total 4 fsync/msync calls are
77 needed per commit to prevent race conditions. It might be possible
78 to reduce this to 3 or even 2 with some more work.
80 - check for a valid recovery record on open of the tdb, while the
81 open lock is held. Automatically recover from the transaction
82 recovery area if needed, then continue with the open as
83 usual. This allows for smooth crash recovery with no administrator
86 - if TDB_NOSYNC is passed to flags in tdb_open then transactions are
87 still available, but no transaction recovery area is used and no
88 fsync/msync calls are made.
92 hold the context of any current transaction
94 struct tdb_transaction {
95 /* the original io methods - used to do IOs to the real db */
96 const struct tdb_methods *io_methods;
98 /* the list of transaction blocks. When a block is first
99 written to, it gets created in this list */
102 size_t last_block_size; /* number of valid bytes in the last block */
104 /* non-zero when an internal transaction error has
105 occurred. All write operations will then fail until the
106 transaction is ended */
107 int transaction_error;
109 /* when inside a transaction we need to keep track of any
110 nested tdb_transaction_start() calls, as these are allowed,
111 but don't create a new transaction */
112 unsigned int nesting;
114 /* set when a prepare has already occurred */
116 tdb_off_t magic_offset;
118 /* old file size before transaction */
119 tdb_len_t old_map_size;
122 /* This doesn't really need to be pagesize, but we use it for similar reasons. */
123 #define PAGESIZE 65536
126 read while in a transaction. We need to check first if the data is in our list
127 of transaction elements, then if not do a real read
129 static enum TDB_ERROR transaction_read(struct tdb_context *tdb, tdb_off_t off,
130 void *buf, tdb_len_t len)
133 enum TDB_ERROR ecode;
135 /* break it down into block sized ops */
136 while (len + (off % PAGESIZE) > PAGESIZE) {
137 tdb_len_t len2 = PAGESIZE - (off % PAGESIZE);
138 ecode = transaction_read(tdb, off, buf, len2);
139 if (ecode != TDB_SUCCESS) {
144 buf = (void *)(len2 + (char *)buf);
151 blk = off / PAGESIZE;
153 /* see if we have it in the block list */
154 if (tdb->transaction->num_blocks <= blk ||
155 tdb->transaction->blocks[blk] == NULL) {
156 /* nope, do a real read */
157 ecode = tdb->transaction->io_methods->tread(tdb, off, buf, len);
158 if (ecode != TDB_SUCCESS) {
164 /* it is in the block list. Now check for the last block */
165 if (blk == tdb->transaction->num_blocks-1) {
166 if (len > tdb->transaction->last_block_size) {
172 /* now copy it out of this block */
173 memcpy(buf, tdb->transaction->blocks[blk] + (off % PAGESIZE), len);
177 tdb->transaction->transaction_error = 1;
178 return tdb_logerr(tdb, ecode, TDB_LOG_ERROR,
179 "transaction_read: failed at off=%zu len=%zu",
180 (size_t)off, (size_t)len);
185 write while in a transaction
187 static enum TDB_ERROR transaction_write(struct tdb_context *tdb, tdb_off_t off,
188 const void *buf, tdb_len_t len)
191 enum TDB_ERROR ecode;
193 /* Only a commit is allowed on a prepared transaction */
194 if (tdb->transaction->prepared) {
195 ecode = tdb_logerr(tdb, TDB_ERR_EINVAL, TDB_LOG_ERROR,
196 "transaction_write: transaction already"
197 " prepared, write not allowed");
201 /* break it up into block sized chunks */
202 while (len + (off % PAGESIZE) > PAGESIZE) {
203 tdb_len_t len2 = PAGESIZE - (off % PAGESIZE);
204 ecode = transaction_write(tdb, off, buf, len2);
205 if (ecode != TDB_SUCCESS) {
211 buf = (const void *)(len2 + (const char *)buf);
219 blk = off / PAGESIZE;
220 off = off % PAGESIZE;
222 if (tdb->transaction->num_blocks <= blk) {
223 uint8_t **new_blocks;
224 /* expand the blocks array */
225 if (tdb->transaction->blocks == NULL) {
226 new_blocks = (uint8_t **)malloc(
227 (blk+1)*sizeof(uint8_t *));
229 new_blocks = (uint8_t **)realloc(
230 tdb->transaction->blocks,
231 (blk+1)*sizeof(uint8_t *));
233 if (new_blocks == NULL) {
234 ecode = tdb_logerr(tdb, TDB_ERR_OOM, TDB_LOG_ERROR,
236 " failed to allocate");
239 memset(&new_blocks[tdb->transaction->num_blocks], 0,
240 (1+(blk - tdb->transaction->num_blocks))*sizeof(uint8_t *));
241 tdb->transaction->blocks = new_blocks;
242 tdb->transaction->num_blocks = blk+1;
243 tdb->transaction->last_block_size = 0;
246 /* allocate and fill a block? */
247 if (tdb->transaction->blocks[blk] == NULL) {
248 tdb->transaction->blocks[blk] = (uint8_t *)calloc(PAGESIZE, 1);
249 if (tdb->transaction->blocks[blk] == NULL) {
250 ecode = tdb_logerr(tdb, TDB_ERR_OOM, TDB_LOG_ERROR,
252 " failed to allocate");
255 if (tdb->transaction->old_map_size > blk * PAGESIZE) {
256 tdb_len_t len2 = PAGESIZE;
257 if (len2 + (blk * PAGESIZE) > tdb->transaction->old_map_size) {
258 len2 = tdb->transaction->old_map_size - (blk * PAGESIZE);
260 ecode = tdb->transaction->io_methods->tread(tdb,
262 tdb->transaction->blocks[blk],
264 if (ecode != TDB_SUCCESS) {
265 ecode = tdb_logerr(tdb, ecode,
269 " read old block: %s",
271 SAFE_FREE(tdb->transaction->blocks[blk]);
274 if (blk == tdb->transaction->num_blocks-1) {
275 tdb->transaction->last_block_size = len2;
280 /* overwrite part of an existing block */
282 memset(tdb->transaction->blocks[blk] + off, 0, len);
284 memcpy(tdb->transaction->blocks[blk] + off, buf, len);
286 if (blk == tdb->transaction->num_blocks-1) {
287 if (len + off > tdb->transaction->last_block_size) {
288 tdb->transaction->last_block_size = len + off;
295 tdb->transaction->transaction_error = 1;
301 write while in a transaction - this variant never expands the transaction blocks, it only
302 updates existing blocks. This means it cannot change the recovery size
304 static void transaction_write_existing(struct tdb_context *tdb, tdb_off_t off,
305 const void *buf, tdb_len_t len)
309 /* break it up into block sized chunks */
310 while (len + (off % PAGESIZE) > PAGESIZE) {
311 tdb_len_t len2 = PAGESIZE - (off % PAGESIZE);
312 transaction_write_existing(tdb, off, buf, len2);
316 buf = (const void *)(len2 + (const char *)buf);
324 blk = off / PAGESIZE;
325 off = off % PAGESIZE;
327 if (tdb->transaction->num_blocks <= blk ||
328 tdb->transaction->blocks[blk] == NULL) {
332 if (blk == tdb->transaction->num_blocks-1 &&
333 off + len > tdb->transaction->last_block_size) {
334 if (off >= tdb->transaction->last_block_size) {
337 len = tdb->transaction->last_block_size - off;
340 /* overwrite part of an existing block */
341 memcpy(tdb->transaction->blocks[blk] + off, buf, len);
346 out of bounds check during a transaction
348 static enum TDB_ERROR transaction_oob(struct tdb_context *tdb, tdb_off_t len,
351 if (len <= tdb->file->map_size || probe) {
355 tdb_logerr(tdb, TDB_ERR_IO, TDB_LOG_ERROR,
356 "tdb_oob len %lld beyond transaction size %lld",
358 (long long)tdb->file->map_size);
363 transaction version of tdb_expand().
365 static enum TDB_ERROR transaction_expand_file(struct tdb_context *tdb,
368 enum TDB_ERROR ecode;
370 /* add a write to the transaction elements, so subsequent
371 reads see the zero data */
372 ecode = transaction_write(tdb, tdb->file->map_size, NULL, addition);
373 if (ecode == TDB_SUCCESS) {
374 tdb->file->map_size += addition;
379 static void *transaction_direct(struct tdb_context *tdb, tdb_off_t off,
380 size_t len, bool write_mode)
382 size_t blk = off / PAGESIZE, end_blk;
384 /* This is wrong for zero-length blocks, but will fail gracefully */
385 end_blk = (off + len - 1) / PAGESIZE;
387 /* Can only do direct if in single block and we've already copied. */
389 tdb->stats.transaction_write_direct++;
391 || blk >= tdb->transaction->num_blocks
392 || tdb->transaction->blocks[blk] == NULL) {
393 tdb->stats.transaction_write_direct_fail++;
396 return tdb->transaction->blocks[blk] + off % PAGESIZE;
399 tdb->stats.transaction_read_direct++;
400 /* Single which we have copied? */
402 && blk < tdb->transaction->num_blocks
403 && tdb->transaction->blocks[blk])
404 return tdb->transaction->blocks[blk] + off % PAGESIZE;
406 /* Otherwise must be all not copied. */
407 while (blk <= end_blk) {
408 if (blk >= tdb->transaction->num_blocks)
410 if (tdb->transaction->blocks[blk]) {
411 tdb->stats.transaction_read_direct_fail++;
416 return tdb->transaction->io_methods->direct(tdb, off, len, false);
419 static const struct tdb_methods transaction_methods = {
423 transaction_expand_file,
430 static enum TDB_ERROR transaction_sync(struct tdb_context *tdb,
431 tdb_off_t offset, tdb_len_t length)
433 if (tdb->flags & TDB_NOSYNC) {
437 if (fsync(tdb->file->fd) != 0) {
438 return tdb_logerr(tdb, TDB_ERR_IO, TDB_LOG_ERROR,
439 "tdb_transaction: fsync failed: %s",
443 if (tdb->file->map_ptr) {
444 tdb_off_t moffset = offset & ~(getpagesize()-1);
445 if (msync(moffset + (char *)tdb->file->map_ptr,
446 length + (offset - moffset), MS_SYNC) != 0) {
447 return tdb_logerr(tdb, TDB_ERR_IO, TDB_LOG_ERROR,
448 "tdb_transaction: msync failed: %s",
457 static void _tdb_transaction_cancel(struct tdb_context *tdb)
460 enum TDB_ERROR ecode;
462 if (tdb->transaction == NULL) {
463 tdb_logerr(tdb, TDB_ERR_EINVAL, TDB_LOG_USE_ERROR,
464 "tdb_transaction_cancel: no transaction");
468 if (tdb->transaction->nesting != 0) {
469 tdb->transaction->transaction_error = 1;
470 tdb->transaction->nesting--;
474 tdb->file->map_size = tdb->transaction->old_map_size;
476 /* free all the transaction blocks */
477 for (i=0;i<tdb->transaction->num_blocks;i++) {
478 if (tdb->transaction->blocks[i] != NULL) {
479 free(tdb->transaction->blocks[i]);
482 SAFE_FREE(tdb->transaction->blocks);
484 if (tdb->transaction->magic_offset) {
485 const struct tdb_methods *methods = tdb->transaction->io_methods;
486 uint64_t invalid = TDB_RECOVERY_INVALID_MAGIC;
488 /* remove the recovery marker */
489 ecode = methods->twrite(tdb, tdb->transaction->magic_offset,
490 &invalid, sizeof(invalid));
491 if (ecode == TDB_SUCCESS)
492 ecode = transaction_sync(tdb,
493 tdb->transaction->magic_offset,
495 if (ecode != TDB_SUCCESS) {
496 tdb_logerr(tdb, ecode, TDB_LOG_ERROR,
497 "tdb_transaction_cancel: failed to remove"
502 if (tdb->file->allrecord_lock.count)
503 tdb_allrecord_unlock(tdb, tdb->file->allrecord_lock.ltype);
505 /* restore the normal io methods */
506 tdb->methods = tdb->transaction->io_methods;
508 tdb_transaction_unlock(tdb, F_WRLCK);
510 if (tdb_has_open_lock(tdb))
511 tdb_unlock_open(tdb, F_WRLCK);
513 SAFE_FREE(tdb->transaction);
517 start a tdb transaction. No token is returned, as only a single
518 transaction is allowed to be pending per tdb_context
520 enum TDB_ERROR tdb_transaction_start(struct tdb_context *tdb)
522 enum TDB_ERROR ecode;
524 tdb->stats.transactions++;
525 /* some sanity checks */
526 if (tdb->read_only || (tdb->flags & TDB_INTERNAL)) {
527 return tdb->last_error = tdb_logerr(tdb, TDB_ERR_EINVAL,
529 "tdb_transaction_start:"
532 "read-only or internal db");
535 /* cope with nested tdb_transaction_start() calls */
536 if (tdb->transaction != NULL) {
537 if (!(tdb->flags & TDB_ALLOW_NESTING)) {
538 return tdb->last_error
539 = tdb_logerr(tdb, TDB_ERR_IO,
541 "tdb_transaction_start:"
542 " already inside transaction");
544 tdb->transaction->nesting++;
545 tdb->stats.transaction_nest++;
549 if (tdb_has_hash_locks(tdb)) {
550 /* the caller must not have any locks when starting a
551 transaction as otherwise we'll be screwed by lack
552 of nested locks in POSIX */
553 return tdb->last_error = tdb_logerr(tdb, TDB_ERR_LOCK,
555 "tdb_transaction_start:"
557 " transaction with locks"
561 tdb->transaction = (struct tdb_transaction *)
562 calloc(sizeof(struct tdb_transaction), 1);
563 if (tdb->transaction == NULL) {
564 return tdb->last_error = tdb_logerr(tdb, TDB_ERR_OOM,
566 "tdb_transaction_start:"
570 /* get the transaction write lock. This is a blocking lock. As
571 discussed with Volker, there are a number of ways we could
572 make this async, which we will probably do in the future */
573 ecode = tdb_transaction_lock(tdb, F_WRLCK);
574 if (ecode != TDB_SUCCESS) {
575 SAFE_FREE(tdb->transaction->blocks);
576 SAFE_FREE(tdb->transaction);
577 return tdb->last_error = ecode;
580 /* get a read lock over entire file. This is upgraded to a write
581 lock during the commit */
582 ecode = tdb_allrecord_lock(tdb, F_RDLCK, TDB_LOCK_WAIT, true);
583 if (ecode != TDB_SUCCESS) {
584 goto fail_allrecord_lock;
587 /* make sure we know about any file expansions already done by
589 tdb->methods->oob(tdb, tdb->file->map_size + 1, true);
590 tdb->transaction->old_map_size = tdb->file->map_size;
592 /* finally hook the io methods, replacing them with
593 transaction specific methods */
594 tdb->transaction->io_methods = tdb->methods;
595 tdb->methods = &transaction_methods;
596 return tdb->last_error = TDB_SUCCESS;
599 tdb_transaction_unlock(tdb, F_WRLCK);
600 SAFE_FREE(tdb->transaction->blocks);
601 SAFE_FREE(tdb->transaction);
602 return tdb->last_error = ecode;
607 cancel the current transaction
609 void tdb_transaction_cancel(struct tdb_context *tdb)
611 tdb->stats.transaction_cancel++;
612 _tdb_transaction_cancel(tdb);
616 work out how much space the linearised recovery data will consume (worst case)
618 static tdb_len_t tdb_recovery_size(struct tdb_context *tdb)
620 tdb_len_t recovery_size = 0;
624 for (i=0;i<tdb->transaction->num_blocks;i++) {
625 if (i * PAGESIZE >= tdb->transaction->old_map_size) {
628 if (tdb->transaction->blocks[i] == NULL) {
631 recovery_size += 2*sizeof(tdb_off_t);
632 if (i == tdb->transaction->num_blocks-1) {
633 recovery_size += tdb->transaction->last_block_size;
635 recovery_size += PAGESIZE;
639 return recovery_size;
642 static enum TDB_ERROR tdb_recovery_area(struct tdb_context *tdb,
643 const struct tdb_methods *methods,
644 tdb_off_t *recovery_offset,
645 struct tdb_recovery_record *rec)
647 enum TDB_ERROR ecode;
649 *recovery_offset = tdb_read_off(tdb,
650 offsetof(struct tdb_header, recovery));
651 if (TDB_OFF_IS_ERR(*recovery_offset)) {
652 return *recovery_offset;
655 if (*recovery_offset == 0) {
660 ecode = methods->tread(tdb, *recovery_offset, rec, sizeof(*rec));
661 if (ecode != TDB_SUCCESS)
664 tdb_convert(tdb, rec, sizeof(*rec));
665 /* ignore invalid recovery regions: can happen in crash */
666 if (rec->magic != TDB_RECOVERY_MAGIC &&
667 rec->magic != TDB_RECOVERY_INVALID_MAGIC) {
668 *recovery_offset = 0;
674 static unsigned int same(const unsigned char *new,
675 const unsigned char *old,
680 for (i = 0; i < length; i++) {
681 if (new[i] != old[i])
687 static unsigned int different(const unsigned char *new,
688 const unsigned char *old,
690 unsigned int min_same,
691 unsigned int *samelen)
696 for (i = 0; i < length; i++) {
697 if (new[i] == old[i]) {
700 if (*samelen >= min_same) {
707 if (*samelen < min_same)
709 return length - *samelen;
712 /* Allocates recovery blob, without tdb_recovery_record at head set up. */
713 static struct tdb_recovery_record *alloc_recovery(struct tdb_context *tdb,
716 struct tdb_recovery_record *rec;
718 enum TDB_ERROR ecode;
720 const struct tdb_methods *old_methods = tdb->methods;
722 rec = malloc(sizeof(*rec) + tdb_recovery_size(tdb));
724 tdb_logerr(tdb, TDB_ERR_OOM, TDB_LOG_ERROR,
725 "transaction_setup_recovery:"
727 return TDB_ERR_PTR(TDB_ERR_OOM);
730 /* We temporarily revert to the old I/O methods, so we can use
732 tdb->methods = tdb->transaction->io_methods;
734 /* build the recovery data into a single blob to allow us to do a single
735 large write, which should be more efficient */
736 p = (unsigned char *)(rec + 1);
737 for (i=0;i<tdb->transaction->num_blocks;i++) {
741 const unsigned char *buffer;
743 if (tdb->transaction->blocks[i] == NULL) {
747 offset = i * PAGESIZE;
749 if (i == tdb->transaction->num_blocks-1) {
750 length = tdb->transaction->last_block_size;
753 if (offset >= tdb->transaction->old_map_size) {
757 if (offset + length > tdb->file->map_size) {
758 ecode = tdb_logerr(tdb, TDB_ERR_CORRUPT, TDB_LOG_ERROR,
759 "tdb_transaction_setup_recovery:"
760 " transaction data over new region"
764 if (offset + length > tdb->transaction->old_map_size) {
765 /* Short read at EOF. */
766 length = tdb->transaction->old_map_size - offset;
768 buffer = tdb_access_read(tdb, offset, length, false);
769 if (TDB_PTR_IS_ERR(buffer)) {
770 ecode = TDB_PTR_ERR(buffer);
774 /* Skip over anything the same at the start. */
775 off = same(tdb->transaction->blocks[i], buffer, length);
778 while (off < length) {
780 unsigned int samelen;
782 len = different(tdb->transaction->blocks[i] + off,
783 buffer + off, length - off,
784 sizeof(offset) + sizeof(len) + 1,
787 memcpy(p, &offset, sizeof(offset));
788 memcpy(p + sizeof(offset), &len, sizeof(len));
789 tdb_convert(tdb, p, sizeof(offset) + sizeof(len));
790 p += sizeof(offset) + sizeof(len);
791 memcpy(p, buffer + off, len);
793 off += len + samelen;
794 offset += len + samelen;
796 tdb_access_release(tdb, buffer);
799 *len = p - (unsigned char *)(rec + 1);
800 tdb->methods = old_methods;
805 tdb->methods = old_methods;
806 return TDB_ERR_PTR(ecode);
809 static tdb_off_t create_recovery_area(struct tdb_context *tdb,
810 tdb_len_t rec_length,
811 struct tdb_recovery_record *rec)
813 tdb_off_t off, recovery_off;
815 enum TDB_ERROR ecode;
816 const struct tdb_methods *methods = tdb->transaction->io_methods;
818 /* round up to a multiple of page size. Overallocate, since each
819 * such allocation forces us to expand the file. */
821 = (((sizeof(*rec) + rec_length + rec_length / 2)
822 + PAGESIZE-1) & ~(PAGESIZE-1))
824 off = tdb->file->map_size;
826 /* Restore ->map_size before calling underlying expand_file.
827 Also so that we don't try to expand the file again in the
828 transaction commit, which would destroy the recovery
830 addition = (tdb->file->map_size - tdb->transaction->old_map_size) +
831 sizeof(*rec) + rec->max_len;
832 tdb->file->map_size = tdb->transaction->old_map_size;
833 tdb->stats.transaction_expand_file++;
834 ecode = methods->expand_file(tdb, addition);
835 if (ecode != TDB_SUCCESS) {
836 return tdb_logerr(tdb, ecode, TDB_LOG_ERROR,
837 "tdb_recovery_allocate:"
838 " failed to create recovery area");
841 /* we have to reset the old map size so that we don't try to
842 expand the file again in the transaction commit, which
843 would destroy the recovery area */
844 tdb->transaction->old_map_size = tdb->file->map_size;
846 /* write the recovery header offset and sync - we can sync without a race here
847 as the magic ptr in the recovery record has not been set */
849 tdb_convert(tdb, &recovery_off, sizeof(recovery_off));
850 ecode = methods->twrite(tdb, offsetof(struct tdb_header, recovery),
851 &recovery_off, sizeof(tdb_off_t));
852 if (ecode != TDB_SUCCESS) {
853 return tdb_logerr(tdb, ecode, TDB_LOG_ERROR,
854 "tdb_recovery_allocate:"
855 " failed to write recovery head");
857 transaction_write_existing(tdb, offsetof(struct tdb_header, recovery),
864 setup the recovery data that will be used on a crash during commit
866 static enum TDB_ERROR transaction_setup_recovery(struct tdb_context *tdb)
868 tdb_len_t recovery_size = 0;
869 tdb_off_t recovery_off = 0;
870 tdb_off_t old_map_size = tdb->transaction->old_map_size;
871 struct tdb_recovery_record *recovery;
872 const struct tdb_methods *methods = tdb->transaction->io_methods;
874 enum TDB_ERROR ecode;
876 recovery = alloc_recovery(tdb, &recovery_size);
877 if (TDB_PTR_IS_ERR(recovery))
878 return TDB_PTR_ERR(recovery);
880 ecode = tdb_recovery_area(tdb, methods, &recovery_off, recovery);
886 if (recovery->max_len < recovery_size) {
887 /* Not large enough. Free up old recovery area. */
890 ecode = add_free_record(tdb, recovery_off,
893 TDB_LOCK_WAIT, true);
895 if (ecode != TDB_SUCCESS) {
896 return tdb_logerr(tdb, ecode, TDB_LOG_ERROR,
897 "tdb_recovery_allocate:"
898 " failed to free previous"
902 /* Refresh recovery after add_free_record above. */
903 recovery = alloc_recovery(tdb, &recovery_size);
904 if (TDB_PTR_IS_ERR(recovery))
905 return TDB_PTR_ERR(recovery);
908 recovery_off = create_recovery_area(tdb, recovery_size,
910 if (TDB_OFF_IS_ERR(recovery_off)) {
916 /* Now we know size, convert rec header. */
917 recovery->magic = TDB_RECOVERY_INVALID_MAGIC;
918 recovery->len = recovery_size;
919 recovery->eof = old_map_size;
920 tdb_convert(tdb, recovery, sizeof(*recovery));
922 /* write the recovery data to the recovery area */
923 ecode = methods->twrite(tdb, recovery_off, recovery, recovery_size);
924 if (ecode != TDB_SUCCESS) {
926 return tdb_logerr(tdb, ecode, TDB_LOG_ERROR,
927 "tdb_transaction_setup_recovery:"
928 " failed to write recovery data");
930 transaction_write_existing(tdb, recovery_off, recovery, recovery_size);
934 /* as we don't have ordered writes, we have to sync the recovery
935 data before we update the magic to indicate that the recovery
937 ecode = transaction_sync(tdb, recovery_off, recovery_size);
938 if (ecode != TDB_SUCCESS)
941 magic = TDB_RECOVERY_MAGIC;
942 tdb_convert(tdb, &magic, sizeof(magic));
944 tdb->transaction->magic_offset
945 = recovery_off + offsetof(struct tdb_recovery_record, magic);
947 ecode = methods->twrite(tdb, tdb->transaction->magic_offset,
948 &magic, sizeof(magic));
949 if (ecode != TDB_SUCCESS) {
950 return tdb_logerr(tdb, ecode, TDB_LOG_ERROR,
951 "tdb_transaction_setup_recovery:"
952 " failed to write recovery magic");
954 transaction_write_existing(tdb, tdb->transaction->magic_offset,
955 &magic, sizeof(magic));
957 /* ensure the recovery magic marker is on disk */
958 return transaction_sync(tdb, tdb->transaction->magic_offset,
962 static enum TDB_ERROR _tdb_transaction_prepare_commit(struct tdb_context *tdb)
964 const struct tdb_methods *methods;
965 enum TDB_ERROR ecode;
967 if (tdb->transaction == NULL) {
968 return tdb_logerr(tdb, TDB_ERR_EINVAL, TDB_LOG_USE_ERROR,
969 "tdb_transaction_prepare_commit:"
973 if (tdb->transaction->prepared) {
974 _tdb_transaction_cancel(tdb);
975 return tdb_logerr(tdb, TDB_ERR_EINVAL, TDB_LOG_USE_ERROR,
976 "tdb_transaction_prepare_commit:"
977 " transaction already prepared");
980 if (tdb->transaction->transaction_error) {
981 _tdb_transaction_cancel(tdb);
982 return tdb_logerr(tdb, TDB_ERR_EINVAL, TDB_LOG_ERROR,
983 "tdb_transaction_prepare_commit:"
984 " transaction error pending");
988 if (tdb->transaction->nesting != 0) {
992 /* check for a null transaction */
993 if (tdb->transaction->blocks == NULL) {
997 methods = tdb->transaction->io_methods;
999 /* upgrade the main transaction lock region to a write lock */
1000 ecode = tdb_allrecord_upgrade(tdb);
1001 if (ecode != TDB_SUCCESS) {
1005 /* get the open lock - this prevents new users attaching to the database
1006 during the commit */
1007 ecode = tdb_lock_open(tdb, F_WRLCK, TDB_LOCK_WAIT|TDB_LOCK_NOCHECK);
1008 if (ecode != TDB_SUCCESS) {
1012 /* Since we have whole db locked, we don't need the expansion lock. */
1013 if (!(tdb->flags & TDB_NOSYNC)) {
1014 /* Sets up tdb->transaction->recovery and
1015 * tdb->transaction->magic_offset. */
1016 ecode = transaction_setup_recovery(tdb);
1017 if (ecode != TDB_SUCCESS) {
1022 tdb->transaction->prepared = true;
1024 /* expand the file to the new size if needed */
1025 if (tdb->file->map_size != tdb->transaction->old_map_size) {
1028 add = tdb->file->map_size - tdb->transaction->old_map_size;
1029 /* Restore original map size for tdb_expand_file */
1030 tdb->file->map_size = tdb->transaction->old_map_size;
1031 ecode = methods->expand_file(tdb, add);
1032 if (ecode != TDB_SUCCESS) {
1037 /* Keep the open lock until the actual commit */
1042 prepare to commit the current transaction
1044 enum TDB_ERROR tdb_transaction_prepare_commit(struct tdb_context *tdb)
1046 return _tdb_transaction_prepare_commit(tdb);
1050 commit the current transaction
1052 enum TDB_ERROR tdb_transaction_commit(struct tdb_context *tdb)
1054 const struct tdb_methods *methods;
1056 enum TDB_ERROR ecode;
1058 if (tdb->transaction == NULL) {
1059 return tdb->last_error = tdb_logerr(tdb, TDB_ERR_EINVAL,
1061 "tdb_transaction_commit:"
1065 tdb_trace(tdb, "tdb_transaction_commit");
1067 if (tdb->transaction->nesting != 0) {
1068 tdb->transaction->nesting--;
1069 return tdb->last_error = TDB_SUCCESS;
1072 /* check for a null transaction */
1073 if (tdb->transaction->blocks == NULL) {
1074 _tdb_transaction_cancel(tdb);
1075 return tdb->last_error = TDB_SUCCESS;
1078 if (!tdb->transaction->prepared) {
1079 ecode = _tdb_transaction_prepare_commit(tdb);
1080 if (ecode != TDB_SUCCESS) {
1081 _tdb_transaction_cancel(tdb);
1082 return tdb->last_error = ecode;
1086 methods = tdb->transaction->io_methods;
1088 /* perform all the writes */
1089 for (i=0;i<tdb->transaction->num_blocks;i++) {
1093 if (tdb->transaction->blocks[i] == NULL) {
1097 offset = i * PAGESIZE;
1099 if (i == tdb->transaction->num_blocks-1) {
1100 length = tdb->transaction->last_block_size;
1103 ecode = methods->twrite(tdb, offset,
1104 tdb->transaction->blocks[i], length);
1105 if (ecode != TDB_SUCCESS) {
1106 /* we've overwritten part of the data and
1107 possibly expanded the file, so we need to
1108 run the crash recovery code */
1109 tdb->methods = methods;
1110 tdb_transaction_recover(tdb);
1112 _tdb_transaction_cancel(tdb);
1114 return tdb->last_error = ecode;
1116 SAFE_FREE(tdb->transaction->blocks[i]);
1119 SAFE_FREE(tdb->transaction->blocks);
1120 tdb->transaction->num_blocks = 0;
1122 /* ensure the new data is on disk */
1123 ecode = transaction_sync(tdb, 0, tdb->file->map_size);
1124 if (ecode != TDB_SUCCESS) {
1125 return tdb->last_error = ecode;
1129 TODO: maybe write to some dummy hdr field, or write to magic
1130 offset without mmap, before the last sync, instead of the
1134 /* on some systems (like Linux 2.6.x) changes via mmap/msync
1135 don't change the mtime of the file, this means the file may
1136 not be backed up (as tdb rounding to block sizes means that
1137 file size changes are quite rare too). The following forces
1138 mtime changes when a transaction completes */
1140 utime(tdb->name, NULL);
1143 /* use a transaction cancel to free memory and remove the
1144 transaction locks: it "restores" map_size, too. */
1145 tdb->transaction->old_map_size = tdb->file->map_size;
1146 _tdb_transaction_cancel(tdb);
1148 return tdb->last_error = TDB_SUCCESS;
1153 recover from an aborted transaction. Must be called with exclusive
1154 database write access already established (including the open
1155 lock to prevent new processes attaching)
1157 enum TDB_ERROR tdb_transaction_recover(struct tdb_context *tdb)
1159 tdb_off_t recovery_head, recovery_eof;
1160 unsigned char *data, *p;
1161 struct tdb_recovery_record rec;
1162 enum TDB_ERROR ecode;
1164 /* find the recovery area */
1165 recovery_head = tdb_read_off(tdb, offsetof(struct tdb_header,recovery));
1166 if (TDB_OFF_IS_ERR(recovery_head)) {
1167 return tdb_logerr(tdb, recovery_head, TDB_LOG_ERROR,
1168 "tdb_transaction_recover:"
1169 " failed to read recovery head");
1172 if (recovery_head == 0) {
1173 /* we have never allocated a recovery record */
1177 /* read the recovery record */
1178 ecode = tdb_read_convert(tdb, recovery_head, &rec, sizeof(rec));
1179 if (ecode != TDB_SUCCESS) {
1180 return tdb_logerr(tdb, ecode, TDB_LOG_ERROR,
1181 "tdb_transaction_recover:"
1182 " failed to read recovery record");
1185 if (rec.magic != TDB_RECOVERY_MAGIC) {
1186 /* there is no valid recovery data */
1190 if (tdb->read_only) {
1191 return tdb_logerr(tdb, TDB_ERR_CORRUPT, TDB_LOG_ERROR,
1192 "tdb_transaction_recover:"
1193 " attempt to recover read only database");
1196 recovery_eof = rec.eof;
1198 data = (unsigned char *)malloc(rec.len);
1200 return tdb_logerr(tdb, TDB_ERR_OOM, TDB_LOG_ERROR,
1201 "tdb_transaction_recover:"
1202 " failed to allocate recovery data");
1205 /* read the full recovery data */
1206 ecode = tdb->methods->tread(tdb, recovery_head + sizeof(rec), data,
1208 if (ecode != TDB_SUCCESS) {
1209 return tdb_logerr(tdb, ecode, TDB_LOG_ERROR,
1210 "tdb_transaction_recover:"
1211 " failed to read recovery data");
1214 /* recover the file data */
1216 while (p+sizeof(tdb_off_t)+sizeof(tdb_len_t) < data + rec.len) {
1219 tdb_convert(tdb, p, sizeof(ofs) + sizeof(len));
1220 memcpy(&ofs, p, sizeof(ofs));
1221 memcpy(&len, p + sizeof(ofs), sizeof(len));
1222 p += sizeof(ofs) + sizeof(len);
1224 ecode = tdb->methods->twrite(tdb, ofs, p, len);
1225 if (ecode != TDB_SUCCESS) {
1227 return tdb_logerr(tdb, ecode, TDB_LOG_ERROR,
1228 "tdb_transaction_recover:"
1229 " failed to recover %zu bytes"
1231 (size_t)len, (size_t)ofs);
1238 ecode = transaction_sync(tdb, 0, tdb->file->map_size);
1239 if (ecode != TDB_SUCCESS) {
1240 return tdb_logerr(tdb, ecode, TDB_LOG_ERROR,
1241 "tdb_transaction_recover:"
1242 " failed to sync recovery");
1245 /* if the recovery area is after the recovered eof then remove it */
1246 if (recovery_eof <= recovery_head) {
1247 ecode = tdb_write_off(tdb, offsetof(struct tdb_header,
1250 if (ecode != TDB_SUCCESS) {
1251 return tdb_logerr(tdb, ecode, TDB_LOG_ERROR,
1252 "tdb_transaction_recover:"
1253 " failed to remove recovery head");
1257 /* remove the recovery magic */
1258 ecode = tdb_write_off(tdb,
1260 + offsetof(struct tdb_recovery_record, magic),
1261 TDB_RECOVERY_INVALID_MAGIC);
1262 if (ecode != TDB_SUCCESS) {
1263 return tdb_logerr(tdb, ecode, TDB_LOG_ERROR,
1264 "tdb_transaction_recover:"
1265 " failed to remove recovery magic");
1268 ecode = transaction_sync(tdb, 0, recovery_eof);
1269 if (ecode != TDB_SUCCESS) {
1270 return tdb_logerr(tdb, ecode, TDB_LOG_ERROR,
1271 "tdb_transaction_recover:"
1272 " failed to sync2 recovery");
1275 tdb_logerr(tdb, TDB_SUCCESS, TDB_LOG_WARNING,
1276 "tdb_transaction_recover: recovered %zu byte database",
1277 (size_t)recovery_eof);
1283 tdb_bool_err tdb_needs_recovery(struct tdb_context *tdb)
1285 tdb_off_t recovery_head;
1286 struct tdb_recovery_record rec;
1287 enum TDB_ERROR ecode;
1289 /* find the recovery area */
1290 recovery_head = tdb_read_off(tdb, offsetof(struct tdb_header,recovery));
1291 if (TDB_OFF_IS_ERR(recovery_head)) {
1292 return recovery_head;
1295 if (recovery_head == 0) {
1296 /* we have never allocated a recovery record */
1300 /* read the recovery record */
1301 ecode = tdb_read_convert(tdb, recovery_head, &rec, sizeof(rec));
1302 if (ecode != TDB_SUCCESS) {
1306 return (rec.magic == TDB_RECOVERY_MAGIC);