27 #define ULONG unsigned long
36 #if (DEBUG +0) && defined(__GNUC__)
37 # define DPRINTF(fmt, ...) \
38 fprintf(stderr, "%s:%d: " fmt "\n", __func__, __LINE__, ##__VA_ARGS__)
40 # define DPRINTF(...) ((void) 0)
45 #define MDB_MAGIC 0xBEEFC0DE
47 #define MAXKEYSIZE 255
49 #define P_INVALID (~0UL)
51 #define F_ISSET(w, f) (((w) & (f)) == (f))
53 typedef uint16_t indx_t;
55 #define DEFAULT_READERS 126
56 #define DEFAULT_MAPSIZE 1048576
58 /* Lock descriptor stuff */
63 typedef struct MDB_rxbody {
68 #define CACHELINE 64 /* most CPUs. Itanium uses 128 */
71 typedef struct MDB_reader {
73 /* cache line alignment */
74 char pad[CACHELINE-sizeof(MDB_rxbody)];
79 uint32_t mt_version; \
80 pthread_mutex_t mt_mutex; \
82 uint32_t mt_numreaders
83 typedef struct MDB_txbody {
87 typedef struct MDB_txninfo {
89 char pad[CACHELINE-sizeof(MDB_txbody)];
90 pthread_mutex_t mt_wmutex;
91 char pad2[CACHELINE-sizeof(pthread_mutex_t)];
92 MDB_reader mt_readers[1];
95 /* Common header for all page types. Overflow pages
96 * occupy a number of contiguous pages with no
97 * headers on any page after the first.
99 typedef struct MDB_page { /* represents a page of storage */
100 pgno_t mp_pgno; /* page number */
101 #define P_BRANCH 0x01 /* branch page */
102 #define P_LEAF 0x02 /* leaf page */
103 #define P_OVERFLOW 0x04 /* overflow page */
104 #define P_META 0x08 /* meta page */
105 #define P_DIRTY 0x10 /* dirty page */
107 #define mp_lower mp_pb.pb.pb_lower
108 #define mp_upper mp_pb.pb.pb_upper
109 #define mp_pages mp_pb.pb_pages
112 indx_t pb_lower; /* lower bound of free space */
113 indx_t pb_upper; /* upper bound of free space */
115 uint32_t pb_pages; /* number of overflow pages */
117 indx_t mp_ptrs[1]; /* dynamic size */
120 #define PAGEHDRSZ ((unsigned) offsetof(MDB_page, mp_ptrs))
122 #define NUMKEYS(p) (((p)->mp_lower - PAGEHDRSZ) >> 1)
123 #define SIZELEFT(p) (indx_t)((p)->mp_upper - (p)->mp_lower)
124 #define PAGEFILL(env, p) (1000L * ((env)->me_meta.mm_psize - PAGEHDRSZ - SIZELEFT(p)) / \
125 ((env)->me_meta.mm_psize - PAGEHDRSZ))
126 #define IS_LEAF(p) F_ISSET((p)->mp_flags, P_LEAF)
127 #define IS_BRANCH(p) F_ISSET((p)->mp_flags, P_BRANCH)
128 #define IS_OVERFLOW(p) F_ISSET((p)->mp_flags, P_OVERFLOW)
130 #define OVPAGES(size, psize) (PAGEHDRSZ + size + psize - 1) / psize;
132 typedef struct MDB_meta { /* meta (footer) page content */
135 void *mm_address; /* address for fixed mapping */
136 size_t mm_mapsize; /* size of mmap region */
137 pgno_t mm_last_pg; /* last used page in file */
138 ULONG mm_txnid; /* txnid that committed this page */
142 ULONG mm_branch_pages;
144 ULONG mm_overflow_pages;
149 typedef struct MDB_dhead { /* a dirty page */
150 SIMPLEQ_ENTRY(MDB_dpage) md_next; /* queue of dirty pages */
152 unsigned md_pi; /* parent index */
156 typedef struct MDB_dpage {
161 SIMPLEQ_HEAD(dirty_queue, MDB_dpage);
163 typedef struct MDB_oldpages {
164 struct MDB_oldpages *mo_next;
166 pgno_t mo_pages[1]; /* dynamic */
169 typedef struct MDB_pageparent {
175 static MDB_dpage *mdb_alloc_page(MDB_txn *txn, MDB_page *parent, unsigned int parent_idx, int num);
176 static int mdb_touch(MDB_txn *txn, MDB_pageparent *mp);
178 typedef struct MDB_ppage { /* ordered list of pages */
179 SLIST_ENTRY(MDB_ppage) mp_entry;
181 unsigned int mp_ki; /* cursor index on page */
183 SLIST_HEAD(page_stack, MDB_ppage);
185 #define CURSOR_EMPTY(c) SLIST_EMPTY(&(c)->mc_stack)
186 #define CURSOR_TOP(c) SLIST_FIRST(&(c)->mc_stack)
187 #define CURSOR_POP(c) SLIST_REMOVE_HEAD(&(c)->mc_stack, mp_entry)
188 #define CURSOR_PUSH(c,p) SLIST_INSERT_HEAD(&(c)->mc_stack, p, mp_entry)
192 struct page_stack mc_stack; /* stack of parent pages */
194 short mc_initialized; /* 1 if initialized */
195 short mc_eof; /* 1 if end is reached */
198 #define METAHASHLEN offsetof(MDB_meta, mm_hash)
199 #define METADATA(p) ((void *)((char *)p + PAGEHDRSZ))
201 typedef struct MDB_node {
202 #define mn_pgno mn_p.np_pgno
203 #define mn_dsize mn_p.np_dsize
205 pgno_t np_pgno; /* child page number */
206 uint32_t np_dsize; /* leaf data size */
208 unsigned int mn_flags:4;
209 unsigned int mn_ksize:12; /* key size */
210 #define F_BIGDATA 0x01 /* data put on overflow page */
214 typedef struct MDB_dbx {
216 MDB_cmp_func *md_cmp; /* user compare function */
217 MDB_rel_func *md_rel; /* user relocate function */
220 typedef struct MDB_db {
224 ULONG md_branch_pages;
226 ULONG md_overflow_pages;
232 pgno_t mt_next_pgno; /* next unallocated page */
236 pgno_t *mt_free_pgs; /* this is an IDL */
238 struct dirty_queue *dirty_queue; /* modified pages */
241 MDB_dbx *mt_dbxs; /* array */
242 MDB_db **mt_dbs; /* array of ptrs */
243 MDB_db mt_db0; /* just for write txns */
244 unsigned int mt_numdbs;
246 #define MDB_TXN_RDONLY 0x01 /* read-only transaction */
247 #define MDB_TXN_ERROR 0x02 /* an error has occurred */
248 #define MDB_TXN_METOGGLE 0x04 /* used meta page 1 */
249 unsigned int mt_flags;
256 unsigned int me_maxreaders;
259 MDB_txninfo *me_txns;
261 MDB_txn *me_txn; /* current write transaction */
263 off_t me_size; /* current file size */
264 pthread_key_t me_txkey; /* thread-key for readers */
265 MDB_oldpages *me_pghead;
266 MDB_oldpages *me_pgtail;
267 MDB_dbx *me_dbxs; /* array */
268 MDB_db **me_dbs; /* array of ptrs */
269 unsigned int me_numdbs;
272 #define NODESIZE offsetof(MDB_node, mn_data)
274 #define INDXSIZE(k) (NODESIZE + ((k) == NULL ? 0 : (k)->mv_size))
275 #define LEAFSIZE(k, d) (NODESIZE + (k)->mv_size + (d)->mv_size)
276 #define NODEPTR(p, i) ((MDB_node *)((char *)(p) + (p)->mp_ptrs[i]))
277 #define NODEKEY(node) (void *)((node)->mn_data)
278 #define NODEDATA(node) (void *)((char *)(node)->mn_data + (node)->mn_ksize)
279 #define NODEPGNO(node) ((node)->mn_pgno)
280 #define NODEDSZ(node) ((node)->mn_dsize)
282 #define MDB_COMMIT_PAGES 64 /* max number of pages to write in one commit */
283 #define MDB_MAXCACHE_DEF 1024 /* max number of pages to keep in cache */
285 static int mdb_search_page_root(MDB_txn *txn,
286 MDB_dbi dbi, MDB_val *key,
287 MDB_cursor *cursor, int modify,
288 MDB_pageparent *mpp);
289 static int mdb_search_page(MDB_txn *txn,
290 MDB_dbi dbi, MDB_val *key,
291 MDB_cursor *cursor, int modify,
292 MDB_pageparent *mpp);
294 static int mdbenv_read_header(MDB_env *env);
295 static int mdb_check_meta_page(MDB_page *p);
296 static int mdbenv_read_meta(MDB_env *env, int *which);
297 static int mdbenv_write_meta(MDB_txn *txn);
298 static MDB_page *mdbenv_get_page(MDB_env *env, pgno_t pgno);
300 static MDB_node *mdb_search_node(MDB_txn *txn, MDB_dbi dbi, MDB_page *mp,
301 MDB_val *key, int *exactp, unsigned int *kip);
302 static int mdb_add_node(MDB_txn *txn, MDB_dbi dbi, MDB_page *mp,
303 indx_t indx, MDB_val *key, MDB_val *data,
304 pgno_t pgno, uint8_t flags);
305 static void mdb_del_node(MDB_page *mp, indx_t indx);
306 static int mdb_read_data(MDB_env *env, MDB_node *leaf, MDB_val *data);
308 static int mdb_rebalance(MDB_txn *txn, MDB_dbi dbi, MDB_pageparent *mp);
309 static int mdb_update_key(MDB_page *mp, indx_t indx, MDB_val *key);
310 static int mdb_move_node(MDB_txn *txn, MDB_dbi dbi,
311 MDB_pageparent *src, indx_t srcindx,
312 MDB_pageparent *dst, indx_t dstindx);
313 static int mdb_merge(MDB_txn *txn, MDB_dbi dbi, MDB_pageparent *src,
314 MDB_pageparent *dst);
315 static int mdb_split(MDB_txn *txn, MDB_dbi dbi, MDB_page **mpp,
316 unsigned int *newindxp, MDB_val *newkey,
317 MDB_val *newdata, pgno_t newpgno);
318 static MDB_dpage *mdb_new_page(MDB_txn *txn, MDB_dbi dbi, uint32_t flags, int num);
320 static void cursor_pop_page(MDB_cursor *cursor);
321 static MDB_ppage *cursor_push_page(MDB_cursor *cursor,
324 static int mdb_set_key(MDB_node *node, MDB_val *key);
325 static int mdb_sibling(MDB_cursor *cursor, int move_right);
326 static int mdb_cursor_next(MDB_cursor *cursor,
327 MDB_val *key, MDB_val *data);
328 static int mdb_cursor_set(MDB_cursor *cursor,
329 MDB_val *key, MDB_val *data, int *exactp);
330 static int mdb_cursor_first(MDB_cursor *cursor,
331 MDB_val *key, MDB_val *data);
333 static size_t mdb_leaf_size(MDB_env *env, MDB_val *key,
335 static size_t mdb_branch_size(MDB_env *env, MDB_val *key);
337 static int memncmp(const void *s1, size_t n1,
338 const void *s2, size_t n2);
339 static int memnrcmp(const void *s1, size_t n1,
340 const void *s2, size_t n2);
343 memncmp(const void *s1, size_t n1, const void *s2, size_t n2)
345 int diff, len_diff = -1;
348 len_diff = (n1 > n2);
351 diff = memcmp(s1, s2, n1);
352 return diff ? diff : len_diff;
356 memnrcmp(const void *s1, size_t n1, const void *s2, size_t n2)
358 const unsigned char *p1, *p2, *p1_lim;
365 p1 = (const unsigned char *)s1 + n1 - 1;
366 p2 = (const unsigned char *)s2 + n2 - 1;
368 for (p1_lim = (n1 <= n2 ? s1 : s2); *p1 == *p2; p1--, p2--) {
370 return (p1 != s1) ? (p1 != p2) : (p2 != s2) ? -1 : 0;
376 mdb_cmp(MDB_txn *txn, MDB_dbi dbi, const MDB_val *a, const MDB_val *b)
378 return txn->mt_dbxs[dbi].md_cmp(a, b);
382 _mdb_cmp(MDB_txn *txn, MDB_dbi dbi, const MDB_val *key1, const MDB_val *key2)
384 if (F_ISSET(txn->mt_dbs[dbi]->md_flags, MDB_REVERSEKEY))
385 return memnrcmp(key1->mv_data, key1->mv_size, key2->mv_data, key2->mv_size);
387 return memncmp((char *)key1->mv_data, key1->mv_size, key2->mv_data, key2->mv_size);
390 /* Allocate new page(s) for writing */
392 mdb_alloc_page(MDB_txn *txn, MDB_page *parent, unsigned int parent_idx, int num)
395 pgno_t pgno = P_INVALID;
397 if (txn->mt_env->me_pghead) {
398 ULONG oldest = txn->mt_txnid - 2;
400 for (i=0; i<txn->mt_env->me_txns->mt_numreaders; i++) {
401 if (txn->mt_env->me_txns->mt_readers[i].mr_txnid < oldest)
402 oldest = txn->mt_env->me_txns->mt_readers[i].mr_txnid;
404 if (oldest > txn->mt_env->me_pghead->mo_txnid) {
405 MDB_oldpages *mop = txn->mt_env->me_pghead;
406 txn->mt_oldest = oldest;
408 /* FIXME: For now, always use fresh pages. We
409 * really ought to search the free list for a
414 /* peel pages off tail, so we only have to truncate the list */
415 pgno = MDB_IDL_LAST(mop->mo_pages);
416 if (MDB_IDL_IS_RANGE(mop->mo_pages)) {
418 if (mop->mo_pages[2] > mop->mo_pages[1])
419 mop->mo_pages[0] = 0;
423 if (MDB_IDL_IS_ZERO(mop->mo_pages)) {
424 txn->mt_env->me_pghead = mop->mo_next;
425 if (!txn->mt_env->me_pghead)
426 txn->mt_env->me_pgtail = NULL;
433 if ((dp = malloc(txn->mt_env->me_meta.mm_psize * num + sizeof(MDB_dhead))) == NULL)
436 dp->h.md_parent = parent;
437 dp->h.md_pi = parent_idx;
438 SIMPLEQ_INSERT_TAIL(txn->mt_u.dirty_queue, dp, h.md_next);
439 if (pgno == P_INVALID) {
440 dp->p.mp_pgno = txn->mt_next_pgno;
441 txn->mt_next_pgno += num;
443 dp->p.mp_pgno = pgno;
449 /* Touch a page: make it dirty and re-insert into tree with updated pgno.
452 mdb_touch(MDB_txn *txn, MDB_pageparent *pp)
454 MDB_page *mp = pp->mp_page;
459 if (!F_ISSET(mp->mp_flags, P_DIRTY)) {
461 if ((dp = mdb_alloc_page(txn, pp->mp_parent, pp->mp_pi, 1)) == NULL)
463 DPRINTF("touched page %lu -> %lu", mp->mp_pgno, dp->p.mp_pgno);
464 mdb_idl_insert(txn->mt_free_pgs, mp->mp_pgno);
465 pgno = dp->p.mp_pgno;
466 memcpy(&dp->p, mp, txn->mt_env->me_meta.mm_psize);
469 mp->mp_flags |= P_DIRTY;
471 /* Update the page number to new touched page. */
472 if (pp->mp_parent != NULL)
473 NODEPGNO(NODEPTR(pp->mp_parent, pp->mp_pi)) = mp->mp_pgno;
480 mdbenv_sync(MDB_env *env)
483 if (!F_ISSET(env->me_flags, MDB_NOSYNC)) {
484 if (fsync(env->me_fd))
490 #define DBX_CHUNK 16 /* space for 16 DBs at a time */
493 mdb_txn_begin(MDB_env *env, int rdonly, MDB_txn **ret)
498 if ((txn = calloc(1, sizeof(*txn))) == NULL) {
499 DPRINTF("calloc: %s", strerror(errno));
504 txn->mt_flags |= MDB_TXN_RDONLY;
506 txn->mt_u.dirty_queue = calloc(1, sizeof(*txn->mt_u.dirty_queue));
507 if (txn->mt_u.dirty_queue == NULL) {
511 SIMPLEQ_INIT(txn->mt_u.dirty_queue);
513 pthread_mutex_lock(&env->me_txns->mt_wmutex);
514 env->me_txns->mt_txnid++;
515 txn->mt_free_pgs = malloc(MDB_IDL_UM_SIZEOF);
516 if (txn->mt_free_pgs == NULL) {
517 free(txn->mt_u.dirty_queue);
521 txn->mt_free_pgs[0] = 0;
524 txn->mt_txnid = env->me_txns->mt_txnid;
526 MDB_reader *r = pthread_getspecific(env->me_txkey);
529 pthread_mutex_lock(&env->me_txns->mt_mutex);
530 for (i=0; i<env->me_maxreaders; i++) {
531 if (env->me_txns->mt_readers[i].mr_pid == 0) {
532 env->me_txns->mt_readers[i].mr_pid = getpid();
533 env->me_txns->mt_readers[i].mr_tid = pthread_self();
534 r = &env->me_txns->mt_readers[i];
535 pthread_setspecific(env->me_txkey, r);
536 if (i >= env->me_txns->mt_numreaders)
537 env->me_txns->mt_numreaders = i+1;
541 pthread_mutex_unlock(&env->me_txns->mt_mutex);
542 if (i == env->me_maxreaders) {
546 r->mr_txnid = txn->mt_txnid;
547 txn->mt_u.reader = r;
554 if ((rc = mdbenv_read_meta(env, &toggle)) != MDB_SUCCESS) {
559 /* Copy the DB arrays */
560 txn->mt_numdbs = env->me_numdbs;
561 rc = (txn->mt_numdbs % DBX_CHUNK) + 1;
562 txn->mt_dbxs = malloc(rc * DBX_CHUNK * sizeof(MDB_dbx));
563 txn->mt_dbs = malloc(rc * DBX_CHUNK * sizeof(MDB_db *));
564 memcpy(txn->mt_dbxs, env->me_dbxs, txn->mt_numdbs * sizeof(MDB_dbx));
565 memcpy(txn->mt_dbs, env->me_dbs, txn->mt_numdbs * sizeof(MDB_db *));
568 memcpy(&txn->mt_db0, txn->mt_dbs[0], sizeof(txn->mt_db0));
569 txn->mt_dbs[0] = &txn->mt_db0;
571 txn->mt_flags |= MDB_TXN_METOGGLE;
572 txn->mt_next_pgno = env->me_meta.mm_last_pg+1;
575 DPRINTF("begin transaction %lu on mdbenv %p, root page %lu",
576 txn->mt_txnid, (void *) env, txn->mt_dbs[0]->md_root);
583 mdb_txn_abort(MDB_txn *txn)
592 DPRINTF("abort transaction %lu on mdbenv %p, root page %lu",
593 txn->mt_txnid, (void *) env, txn->mt_dbs[0]->md_root);
598 if (F_ISSET(txn->mt_flags, MDB_TXN_RDONLY)) {
599 txn->mt_u.reader->mr_txnid = 0;
601 /* Discard all dirty pages. Return any re-used pages
604 MDB_IDL_ZERO(txn->mt_free_pgs);
605 while (!SIMPLEQ_EMPTY(txn->mt_u.dirty_queue)) {
606 dp = SIMPLEQ_FIRST(txn->mt_u.dirty_queue);
607 SIMPLEQ_REMOVE_HEAD(txn->mt_u.dirty_queue, h.md_next);
608 if (dp->p.mp_pgno <= env->me_meta.mm_last_pg)
609 mdb_idl_insert(txn->mt_free_pgs, dp->p.mp_pgno);
612 /* put back to head of free list */
613 if (!MDB_IDL_IS_ZERO(txn->mt_free_pgs)) {
616 mop = malloc(sizeof(MDB_oldpages) + MDB_IDL_SIZEOF(txn->mt_free_pgs) - sizeof(pgno_t));
617 mop->mo_next = env->me_pghead;
618 mop->mo_txnid = txn->mt_oldest - 1;
619 if (!env->me_pghead) {
620 env->me_pgtail = mop;
622 env->me_pghead = mop;
623 memcpy(mop->mo_pages, txn->mt_free_pgs, MDB_IDL_SIZEOF(txn->mt_free_pgs));
626 free(txn->mt_free_pgs);
627 free(txn->mt_u.dirty_queue);
629 env->me_txns->mt_txnid--;
630 pthread_mutex_unlock(&env->me_txns->mt_wmutex);
637 mdb_txn_commit(MDB_txn *txn)
645 struct iovec iov[MDB_COMMIT_PAGES];
648 assert(txn->mt_env != NULL);
652 if (F_ISSET(txn->mt_flags, MDB_TXN_RDONLY)) {
653 DPRINTF("attempt to commit read-only transaction");
658 if (txn != env->me_txn) {
659 DPRINTF("attempt to commit unknown transaction");
664 if (F_ISSET(txn->mt_flags, MDB_TXN_ERROR)) {
665 DPRINTF("error flag is set, can't commit");
670 if (SIMPLEQ_EMPTY(txn->mt_u.dirty_queue))
673 DPRINTF("committing transaction %lu on mdbenv %p, root page %lu",
674 txn->mt_txnid, (void *) env, txn->mt_dbs[0]->md_root);
676 /* Commit up to MDB_COMMIT_PAGES dirty pages to disk until done.
683 SIMPLEQ_FOREACH(dp, txn->mt_u.dirty_queue, h.md_next) {
684 if (dp->p.mp_pgno != next) {
686 DPRINTF("committing %u dirty pages", n);
687 rc = writev(env->me_fd, iov, n);
691 DPRINTF("short write, filesystem full?");
693 DPRINTF("writev: %s", strerror(errno));
700 lseek(env->me_fd, dp->p.mp_pgno * env->me_meta.mm_psize, SEEK_SET);
701 next = dp->p.mp_pgno;
703 DPRINTF("committing page %lu", dp->p.mp_pgno);
704 iov[n].iov_len = env->me_meta.mm_psize * dp->h.md_num;
705 iov[n].iov_base = &dp->p;
706 size += iov[n].iov_len;
707 next = dp->p.mp_pgno + dp->h.md_num;
708 /* clear dirty flag */
709 dp->p.mp_flags &= ~P_DIRTY;
710 if (++n >= MDB_COMMIT_PAGES) {
719 DPRINTF("committing %u dirty pages", n);
720 rc = writev(env->me_fd, iov, n);
724 DPRINTF("short write, filesystem full?");
726 DPRINTF("writev: %s", strerror(errno));
733 /* Drop the dirty pages.
735 while (!SIMPLEQ_EMPTY(txn->mt_u.dirty_queue)) {
736 dp = SIMPLEQ_FIRST(txn->mt_u.dirty_queue);
737 SIMPLEQ_REMOVE_HEAD(txn->mt_u.dirty_queue, h.md_next);
741 if ((n = mdbenv_sync(env)) != 0 ||
742 (n = mdbenv_write_meta(txn)) != MDB_SUCCESS ||
743 (n = mdbenv_sync(env)) != 0) {
750 MDB_dbx *p1 = env->me_dbxs;
751 MDB_db **p2 = env->me_dbs;
753 txn->mt_dbs[0] = env->me_dbs[0];
754 env->me_dbxs = txn->mt_dbxs;
755 env->me_dbs = txn->mt_dbs;
756 env->me_numdbs = txn->mt_numdbs;
762 /* add to tail of free list */
763 if (!MDB_IDL_IS_ZERO(txn->mt_free_pgs)) {
766 mop = malloc(sizeof(MDB_oldpages) + MDB_IDL_SIZEOF(txn->mt_free_pgs) - sizeof(pgno_t));
768 if (env->me_pghead) {
769 env->me_pgtail->mo_next = mop;
771 env->me_pghead = mop;
773 env->me_pgtail = mop;
774 memcpy(mop->mo_pages, txn->mt_free_pgs, MDB_IDL_SIZEOF(txn->mt_free_pgs));
775 mop->mo_txnid = txn->mt_txnid;
778 pthread_mutex_unlock(&env->me_txns->mt_wmutex);
779 free(txn->mt_free_pgs);
780 free(txn->mt_u.dirty_queue);
791 mdbenv_read_header(MDB_env *env)
800 /* We don't know the page size yet, so use a minimum value.
803 if ((rc = pread(env->me_fd, page, PAGESIZE, 0)) == 0) {
805 } else if (rc != PAGESIZE) {
808 DPRINTF("read: %s", strerror(errno));
812 p = (MDB_page *)page;
814 if (!F_ISSET(p->mp_flags, P_META)) {
815 DPRINTF("page %lu not a meta page", p->mp_pgno);
820 if (m->mm_magic != MDB_MAGIC) {
821 DPRINTF("meta has invalid magic");
825 if (m->mm_version != MDB_VERSION) {
826 DPRINTF("database is version %u, expected version %u",
827 m->mm_version, MDB_VERSION);
831 memcpy(&env->me_meta, m, sizeof(*m));
836 mdbenv_init_meta(MDB_env *env)
843 DPRINTF("writing new meta page");
844 psize = sysconf(_SC_PAGE_SIZE);
846 env->me_meta.mm_magic = MDB_MAGIC;
847 env->me_meta.mm_version = MDB_VERSION;
848 env->me_meta.mm_psize = psize;
849 env->me_meta.mm_flags = env->me_flags & 0xffff;
850 env->me_meta.mm_root = P_INVALID;
851 env->me_meta.mm_last_pg = 1;
853 p = calloc(2, psize);
855 p->mp_flags = P_META;
858 memcpy(meta, &env->me_meta, sizeof(*meta));
860 q = (MDB_page *)((char *)p + psize);
863 q->mp_flags = P_META;
866 memcpy(meta, &env->me_meta, sizeof(*meta));
868 rc = write(env->me_fd, p, psize * 2);
870 return (rc == (int)psize * 2) ? MDB_SUCCESS : errno;
874 mdbenv_write_meta(MDB_txn *txn)
883 assert(txn->mt_env != NULL);
885 DPRINTF("writing meta page for root page %lu", txn->mt_dbs[0]->md_root);
890 off = offsetof(MDB_meta, mm_last_pg);
891 len = sizeof(MDB_meta) - off;
894 meta.mm_last_pg = txn->mt_next_pgno - 1;
895 meta.mm_txnid = txn->mt_txnid;
896 meta.mm_psize = env->me_meta.mm_psize;
897 meta.mm_flags = env->me_meta.mm_flags;
898 meta.mm_depth = txn->mt_dbs[0]->md_depth;
899 meta.mm_branch_pages = txn->mt_dbs[0]->md_branch_pages;
900 meta.mm_leaf_pages = txn->mt_dbs[0]->md_leaf_pages;
901 meta.mm_overflow_pages = txn->mt_dbs[0]->md_overflow_pages;
902 meta.mm_entries = txn->mt_dbs[0]->md_entries;
903 meta.mm_root = txn->mt_dbs[0]->md_root;
905 if (!F_ISSET(txn->mt_flags, MDB_TXN_METOGGLE))
906 off += env->me_meta.mm_psize;
909 lseek(env->me_fd, off, SEEK_SET);
910 rc = write(env->me_fd, ptr, len);
912 DPRINTF("write failed, disk error?");
919 /* Returns true if page p is a valid meta page, false otherwise.
922 mdb_check_meta_page(MDB_page *p)
924 if (!F_ISSET(p->mp_flags, P_META)) {
925 DPRINTF("page %lu not a meta page", p->mp_pgno);
933 mdbenv_read_meta(MDB_env *env, int *which)
941 if ((mp0 = mdbenv_get_page(env, 0)) == NULL ||
942 (mp1 = mdbenv_get_page(env, 1)) == NULL)
945 rc = mdb_check_meta_page(mp0);
948 rc = mdb_check_meta_page(mp1);
951 meta[0] = METADATA(mp0);
952 meta[1] = METADATA(mp1);
954 if (meta[0]->mm_txnid < meta[1]->mm_txnid)
957 if (meta[toggle]->mm_txnid > env->me_meta.mm_txnid) {
958 memcpy(&env->me_meta, meta[toggle], sizeof(env->me_meta));
963 DPRINTF("Using meta page %d", toggle);
969 mdbenv_create(MDB_env **env)
973 e = calloc(1, sizeof(*e));
974 if (!e) return ENOMEM;
976 e->me_meta.mm_mapsize = DEFAULT_MAPSIZE;
977 e->me_maxreaders = DEFAULT_READERS;
985 mdbenv_set_mapsize(MDB_env *env, size_t size)
989 env->me_mapsize = env->me_meta.mm_mapsize = size;
994 mdbenv_set_maxreaders(MDB_env *env, int readers)
996 env->me_maxreaders = readers;
1001 mdbenv_get_maxreaders(MDB_env *env, int *readers)
1003 if (!env || !readers)
1005 *readers = env->me_maxreaders;
1010 mdbenv_open2(MDB_env *env, unsigned int flags)
1014 env->me_flags = flags;
1016 if ((i = mdbenv_read_header(env)) != 0) {
1019 DPRINTF("new mdbenv");
1023 if (!env->me_mapsize)
1024 env->me_mapsize = env->me_meta.mm_mapsize;
1027 if (env->me_meta.mm_address && (flags & MDB_FIXEDMAP))
1029 env->me_map = mmap(env->me_meta.mm_address, env->me_mapsize, PROT_READ, i,
1031 if (env->me_map == MAP_FAILED)
1035 env->me_meta.mm_mapsize = env->me_mapsize;
1036 if (flags & MDB_FIXEDMAP)
1037 env->me_meta.mm_address = env->me_map;
1038 i = mdbenv_init_meta(env);
1039 if (i != MDB_SUCCESS) {
1040 munmap(env->me_map, env->me_mapsize);
1045 if ((i = mdbenv_read_meta(env, NULL)) != 0)
1048 DPRINTF("opened database version %u, pagesize %u",
1049 env->me_meta.mm_version, env->me_meta.mm_psize);
1050 DPRINTF("depth: %u", env->me_meta.mm_depth);
1051 DPRINTF("entries: %lu", env->me_meta.mm_entries);
1052 DPRINTF("branch pages: %lu", env->me_meta.mm_branch_pages);
1053 DPRINTF("leaf pages: %lu", env->me_meta.mm_leaf_pages);
1054 DPRINTF("overflow pages: %lu", env->me_meta.mm_overflow_pages);
1055 DPRINTF("root: %lu", env->me_meta.mm_root);
1061 mdbenv_reader_dest(void *ptr)
1063 MDB_reader *reader = ptr;
1065 reader->mr_txnid = 0;
1071 mdbenv_share_locks(MDB_env *env)
1073 struct flock lock_info;
1075 env->me_txns->mt_txnid = env->me_meta.mm_txnid;
1077 memset((void *)&lock_info, 0, sizeof(lock_info));
1078 lock_info.l_type = F_RDLCK;
1079 lock_info.l_whence = SEEK_SET;
1080 lock_info.l_start = 0;
1081 lock_info.l_len = 1;
1082 fcntl(env->me_lfd, F_SETLK, &lock_info);
1086 mdbenv_setup_locks(MDB_env *env, char *lpath, int mode, int *excl)
1090 struct flock lock_info;
1094 if ((env->me_lfd = open(lpath, O_RDWR|O_CREAT, mode)) == -1) {
1098 /* Try to get exclusive lock. If we succeed, then
1099 * nobody is using the lock region and we should initialize it.
1101 memset((void *)&lock_info, 0, sizeof(lock_info));
1102 lock_info.l_type = F_WRLCK;
1103 lock_info.l_whence = SEEK_SET;
1104 lock_info.l_start = 0;
1105 lock_info.l_len = 1;
1106 rc = fcntl(env->me_lfd, F_SETLK, &lock_info);
1110 lock_info.l_type = F_RDLCK;
1111 rc = fcntl(env->me_lfd, F_SETLK, &lock_info);
1117 size = lseek(env->me_lfd, 0, SEEK_END);
1118 rsize = (env->me_maxreaders-1) * sizeof(MDB_reader) + sizeof(MDB_txninfo);
1119 if (size < rsize && *excl) {
1120 if (ftruncate(env->me_lfd, rsize) != 0) {
1126 size = rsize - sizeof(MDB_txninfo);
1127 env->me_maxreaders = size/sizeof(MDB_reader) + 1;
1129 env->me_txns = mmap(0, rsize, PROT_READ|PROT_WRITE, MAP_SHARED,
1131 if (env->me_txns == MAP_FAILED) {
1136 pthread_mutexattr_t mattr;
1138 pthread_mutexattr_init(&mattr);
1139 pthread_mutexattr_setpshared(&mattr, PTHREAD_PROCESS_SHARED);
1140 pthread_mutex_init(&env->me_txns->mt_mutex, &mattr);
1141 pthread_mutex_init(&env->me_txns->mt_wmutex, &mattr);
1142 env->me_txns->mt_version = MDB_VERSION;
1143 env->me_txns->mt_magic = MDB_MAGIC;
1144 env->me_txns->mt_txnid = 0;
1145 env->me_txns->mt_numreaders = 0;
1148 if (env->me_txns->mt_magic != MDB_MAGIC) {
1149 DPRINTF("lock region has invalid magic");
1152 if (env->me_txns->mt_version != MDB_VERSION) {
1153 DPRINTF("lock region is version %u, expected version %u",
1154 env->me_txns->mt_version, MDB_VERSION);
1157 if (errno != EACCES && errno != EAGAIN) {
1171 mdbenv_open(MDB_env *env, const char *path, unsigned int flags, mode_t mode)
1173 int oflags, rc, len, excl;
1174 char *lpath, *dpath;
1177 lpath = malloc(len + sizeof("/lock.mdb") + len + sizeof("/data.db"));
1180 dpath = lpath + len + sizeof("/lock.mdb");
1181 sprintf(lpath, "%s/lock.mdb", path);
1182 sprintf(dpath, "%s/data.mdb", path);
1184 rc = mdbenv_setup_locks(env, lpath, mode, &excl);
1188 if (F_ISSET(flags, MDB_RDONLY))
1191 oflags = O_RDWR | O_CREAT;
1193 if ((env->me_fd = open(dpath, oflags, mode)) == -1)
1196 if ((rc = mdbenv_open2(env, flags)) != MDB_SUCCESS) {
1200 env->me_path = strdup(path);
1201 DPRINTF("opened dbenv %p", (void *) env);
1202 pthread_key_create(&env->me_txkey, mdbenv_reader_dest);
1204 mdbenv_share_locks(env);
1205 env->me_dbxs = calloc(DBX_CHUNK, sizeof(MDB_dbx));
1206 env->me_dbs = calloc(DBX_CHUNK, sizeof(MDB_db *));
1208 env->me_dbs[0] = (MDB_db *)&env->me_meta.mm_psize;
1218 mdbenv_close(MDB_env *env)
1228 munmap(env->me_map, env->me_mapsize);
1232 size_t size = (env->me_maxreaders-1) * sizeof(MDB_reader) + sizeof(MDB_txninfo);
1233 munmap(env->me_txns, size);
1239 /* Search for key within a leaf page, using binary search.
1240 * Returns the smallest entry larger or equal to the key.
1241 * If exactp is non-null, stores whether the found entry was an exact match
1242 * in *exactp (1 or 0).
1243 * If kip is non-null, stores the index of the found entry in *kip.
1244 * If no entry larger or equal to the key is found, returns NULL.
1247 mdb_search_node(MDB_txn *txn, MDB_dbi dbi, MDB_page *mp, MDB_val *key,
1248 int *exactp, unsigned int *kip)
1256 DPRINTF("searching %u keys in %s page %lu",
1258 IS_LEAF(mp) ? "leaf" : "branch",
1261 assert(NUMKEYS(mp) > 0);
1263 memset(&nodekey, 0, sizeof(nodekey));
1265 low = IS_LEAF(mp) ? 0 : 1;
1266 high = NUMKEYS(mp) - 1;
1267 while (low <= high) {
1268 i = (low + high) >> 1;
1269 node = NODEPTR(mp, i);
1271 nodekey.mv_size = node->mn_ksize;
1272 nodekey.mv_data = NODEKEY(node);
1274 if (txn->mt_dbxs[dbi].md_cmp)
1275 rc = txn->mt_dbxs[dbi].md_cmp(key, &nodekey);
1277 rc = _mdb_cmp(txn, dbi, key, &nodekey);
1280 DPRINTF("found leaf index %u [%.*s], rc = %i",
1281 i, (int)nodekey.mv_size, (char *)nodekey.mv_data, rc);
1283 DPRINTF("found branch index %u [%.*s -> %lu], rc = %i",
1284 i, (int)node->mn_ksize, (char *)NODEKEY(node),
1295 if (rc > 0) { /* Found entry is less than the key. */
1296 i++; /* Skip to get the smallest entry larger than key. */
1297 if (i >= NUMKEYS(mp))
1298 /* There is no entry larger or equal to the key. */
1302 *exactp = (rc == 0);
1303 if (kip) /* Store the key index if requested. */
1306 return NODEPTR(mp, i);
1310 cursor_pop_page(MDB_cursor *cursor)
1314 top = CURSOR_TOP(cursor);
1317 DPRINTF("popped page %lu off cursor %p", top->mp_page->mp_pgno, (void *) cursor);
1323 cursor_push_page(MDB_cursor *cursor, MDB_page *mp)
1327 DPRINTF("pushing page %lu on cursor %p", mp->mp_pgno, (void *) cursor);
1329 if ((ppage = calloc(1, sizeof(*ppage))) == NULL)
1331 ppage->mp_page = mp;
1332 CURSOR_PUSH(cursor, ppage);
1337 mdbenv_get_page(MDB_env *env, pgno_t pgno)
1340 MDB_txn *txn = env->me_txn;
1343 if (txn && !SIMPLEQ_EMPTY(txn->mt_u.dirty_queue)) {
1345 SIMPLEQ_FOREACH(dp, txn->mt_u.dirty_queue, h.md_next) {
1346 if (dp->p.mp_pgno == pgno) {
1354 p = (MDB_page *)(env->me_map + env->me_meta.mm_psize * pgno);
1360 mdb_search_page_root(MDB_txn *txn, MDB_dbi dbi, MDB_val *key,
1361 MDB_cursor *cursor, int modify, MDB_pageparent *mpp)
1363 MDB_page *mp = mpp->mp_page;
1366 if (cursor && cursor_push_page(cursor, mp) == NULL)
1369 while (IS_BRANCH(mp)) {
1373 DPRINTF("branch page %lu has %u keys", mp->mp_pgno, NUMKEYS(mp));
1374 assert(NUMKEYS(mp) > 1);
1375 DPRINTF("found index 0 to page %lu", NODEPGNO(NODEPTR(mp, 0)));
1377 if (key == NULL) /* Initialize cursor to first page. */
1381 node = mdb_search_node(txn, dbi, mp, key, &exact, &i);
1383 i = NUMKEYS(mp) - 1;
1391 DPRINTF("following index %u for key %.*s",
1392 i, (int)key->mv_size, (char *)key->mv_data);
1393 assert(i < NUMKEYS(mp));
1394 node = NODEPTR(mp, i);
1397 CURSOR_TOP(cursor)->mp_ki = i;
1399 mpp->mp_parent = mp;
1400 if ((mp = mdbenv_get_page(txn->mt_env, NODEPGNO(node))) == NULL)
1405 if (cursor && cursor_push_page(cursor, mp) == NULL)
1409 MDB_dhead *dh = ((MDB_dhead *)mp)-1;
1410 if ((rc = mdb_touch(txn, mpp)) != 0)
1412 dh = ((MDB_dhead *)mpp->mp_page)-1;
1413 dh->md_parent = mpp->mp_parent;
1414 dh->md_pi = mpp->mp_pi;
1421 DPRINTF("internal error, index points to a %02X page!?",
1426 DPRINTF("found leaf page %lu for key %.*s", mp->mp_pgno,
1427 key ? (int)key->mv_size : 0, key ? (char *)key->mv_data : NULL);
1432 /* Search for the page a given key should be in.
1433 * Stores a pointer to the found page in *mpp.
1434 * If key is NULL, search for the lowest page (used by mdb_cursor_first).
1435 * If cursor is non-null, pushes parent pages on the cursor stack.
1436 * If modify is true, visited pages are updated with new page numbers.
1439 mdb_search_page(MDB_txn *txn, MDB_dbi dbi, MDB_val *key,
1440 MDB_cursor *cursor, int modify, MDB_pageparent *mpp)
1445 /* Choose which root page to start with. If a transaction is given
1446 * use the root page from the transaction, otherwise read the last
1447 * committed root page.
1449 if (F_ISSET(txn->mt_flags, MDB_TXN_ERROR)) {
1450 DPRINTF("transaction has failed, must abort");
1453 root = txn->mt_dbs[dbi]->md_root;
1455 if (root == P_INVALID) { /* Tree is empty. */
1456 DPRINTF("tree is empty");
1460 if ((mpp->mp_page = mdbenv_get_page(txn->mt_env, root)) == NULL)
1463 DPRINTF("root page has flags 0x%X", mpp->mp_page->mp_flags);
1465 if (modify && !F_ISSET(mpp->mp_page->mp_flags, P_DIRTY)) {
1466 mpp->mp_parent = NULL;
1468 if ((rc = mdb_touch(txn, mpp)))
1470 txn->mt_dbs[dbi]->md_root = mpp->mp_page->mp_pgno;
1473 return mdb_search_page_root(txn, dbi, key, cursor, modify, mpp);
1477 mdb_read_data(MDB_env *env, MDB_node *leaf, MDB_val *data)
1479 MDB_page *omp; /* overflow mpage */
1482 if (!F_ISSET(leaf->mn_flags, F_BIGDATA)) {
1483 data->mv_size = leaf->mn_dsize;
1484 data->mv_data = NODEDATA(leaf);
1488 /* Read overflow data.
1490 data->mv_size = leaf->mn_dsize;
1491 memcpy(&pgno, NODEDATA(leaf), sizeof(pgno));
1492 if ((omp = mdbenv_get_page(env, pgno)) == NULL) {
1493 DPRINTF("read overflow page %lu failed", pgno);
1496 data->mv_data = omp;
1502 mdb_get(MDB_txn *txn, MDB_dbi dbi,
1503 MDB_val *key, MDB_val *data)
1511 DPRINTF("===> get key [%.*s]", (int)key->mv_size, (char *)key->mv_data);
1513 if (key->mv_size == 0 || key->mv_size > MAXKEYSIZE) {
1517 if ((rc = mdb_search_page(txn, dbi, key, NULL, 0, &mpp)) != MDB_SUCCESS)
1520 leaf = mdb_search_node(txn, dbi, mpp.mp_page, key, &exact, NULL);
1522 rc = mdb_read_data(txn->mt_env, leaf, data);
1531 mdb_sibling(MDB_cursor *cursor, int move_right)
1535 MDB_ppage *parent, *top;
1538 top = CURSOR_TOP(cursor);
1539 if ((parent = SLIST_NEXT(top, mp_entry)) == NULL) {
1540 return ENOENT; /* root has no siblings */
1543 DPRINTF("parent page is page %lu, index %u",
1544 parent->mp_page->mp_pgno, parent->mp_ki);
1546 cursor_pop_page(cursor);
1547 if (move_right ? (parent->mp_ki + 1 >= NUMKEYS(parent->mp_page))
1548 : (parent->mp_ki == 0)) {
1549 DPRINTF("no more keys left, moving to %s sibling",
1550 move_right ? "right" : "left");
1551 if ((rc = mdb_sibling(cursor, move_right)) != MDB_SUCCESS)
1553 parent = CURSOR_TOP(cursor);
1559 DPRINTF("just moving to %s index key %u",
1560 move_right ? "right" : "left", parent->mp_ki);
1562 assert(IS_BRANCH(parent->mp_page));
1564 indx = NODEPTR(parent->mp_page, parent->mp_ki);
1565 if ((mp = mdbenv_get_page(cursor->mc_txn->mt_env, indx->mn_pgno)) == NULL)
1568 mp->parent = parent->mp_page;
1569 mp->parent_index = parent->mp_ki;
1572 cursor_push_page(cursor, mp);
1578 mdb_set_key(MDB_node *node, MDB_val *key)
1583 key->mv_size = node->mn_ksize;
1584 key->mv_data = NODEKEY(node);
1590 mdb_cursor_next(MDB_cursor *cursor, MDB_val *key, MDB_val *data)
1596 if (cursor->mc_eof) {
1600 assert(cursor->mc_initialized);
1602 top = CURSOR_TOP(cursor);
1605 DPRINTF("cursor_next: top page is %lu in cursor %p", mp->mp_pgno, (void *) cursor);
1607 if (top->mp_ki + 1 >= NUMKEYS(mp)) {
1608 DPRINTF("=====> move to next sibling page");
1609 if (mdb_sibling(cursor, 1) != MDB_SUCCESS) {
1613 top = CURSOR_TOP(cursor);
1615 DPRINTF("next page is %lu, key index %u", mp->mp_pgno, top->mp_ki);
1619 DPRINTF("==> cursor points to page %lu with %u keys, key index %u",
1620 mp->mp_pgno, NUMKEYS(mp), top->mp_ki);
1622 assert(IS_LEAF(mp));
1623 leaf = NODEPTR(mp, top->mp_ki);
1625 if (data && mdb_read_data(cursor->mc_txn->mt_env, leaf, data) != MDB_SUCCESS)
1628 return mdb_set_key(leaf, key);
1632 mdb_cursor_set(MDB_cursor *cursor, MDB_val *key, MDB_val *data,
1642 assert(key->mv_size > 0);
1644 rc = mdb_search_page(cursor->mc_txn, cursor->mc_dbi, key, cursor, 0, &mpp);
1645 if (rc != MDB_SUCCESS)
1647 assert(IS_LEAF(mpp.mp_page));
1649 top = CURSOR_TOP(cursor);
1650 leaf = mdb_search_node(cursor->mc_txn, cursor->mc_dbi, mpp.mp_page, key, exactp, &top->mp_ki);
1651 if (exactp != NULL && !*exactp) {
1652 /* MDB_CURSOR_EXACT specified and not an exact match. */
1657 DPRINTF("===> inexact leaf not found, goto sibling");
1658 if ((rc = mdb_sibling(cursor, 1)) != MDB_SUCCESS)
1659 return rc; /* no entries matched */
1660 top = CURSOR_TOP(cursor);
1662 mpp.mp_page = top->mp_page;
1663 assert(IS_LEAF(mpp.mp_page));
1664 leaf = NODEPTR(mpp.mp_page, 0);
1667 cursor->mc_initialized = 1;
1670 if (data && (rc = mdb_read_data(cursor->mc_txn->mt_env, leaf, data)) != MDB_SUCCESS)
1673 rc = mdb_set_key(leaf, key);
1674 if (rc == MDB_SUCCESS) {
1675 DPRINTF("==> cursor placed on key %.*s",
1676 (int)key->mv_size, (char *)key->mv_data);
1684 mdb_cursor_first(MDB_cursor *cursor, MDB_val *key, MDB_val *data)
1690 rc = mdb_search_page(cursor->mc_txn, cursor->mc_dbi, NULL, cursor, 0, &mpp);
1691 if (rc != MDB_SUCCESS)
1693 assert(IS_LEAF(mpp.mp_page));
1695 leaf = NODEPTR(mpp.mp_page, 0);
1696 cursor->mc_initialized = 1;
1699 if (data && (rc = mdb_read_data(cursor->mc_txn->mt_env, leaf, data)) != MDB_SUCCESS)
1702 return mdb_set_key(leaf, key);
1706 mdb_cursor_get(MDB_cursor *cursor, MDB_val *key, MDB_val *data,
1716 case MDB_CURSOR_EXACT:
1717 while (CURSOR_TOP(cursor) != NULL)
1718 cursor_pop_page(cursor);
1719 if (key == NULL || key->mv_size == 0 || key->mv_size > MAXKEYSIZE) {
1721 } else if (op == MDB_CURSOR_EXACT)
1722 rc = mdb_cursor_set(cursor, key, data, &exact);
1724 rc = mdb_cursor_set(cursor, key, data, NULL);
1727 if (!cursor->mc_initialized)
1728 rc = mdb_cursor_first(cursor, key, data);
1730 rc = mdb_cursor_next(cursor, key, data);
1733 while (CURSOR_TOP(cursor) != NULL)
1734 cursor_pop_page(cursor);
1735 rc = mdb_cursor_first(cursor, key, data);
1738 DPRINTF("unhandled/unimplemented cursor operation %u", op);
1746 /* Allocate a page and initialize it
1749 mdb_new_page(MDB_txn *txn, MDB_dbi dbi, uint32_t flags, int num)
1753 if ((dp = mdb_alloc_page(txn, NULL, 0, num)) == NULL)
1755 DPRINTF("allocated new mpage %lu, page size %u",
1756 dp->p.mp_pgno, txn->mt_env->me_meta.mm_psize);
1757 dp->p.mp_flags = flags | P_DIRTY;
1758 dp->p.mp_lower = PAGEHDRSZ;
1759 dp->p.mp_upper = txn->mt_env->me_meta.mm_psize;
1761 if (IS_BRANCH(&dp->p))
1762 txn->mt_dbs[dbi]->md_branch_pages++;
1763 else if (IS_LEAF(&dp->p))
1764 txn->mt_dbs[dbi]->md_leaf_pages++;
1765 else if (IS_OVERFLOW(&dp->p)) {
1766 txn->mt_dbs[dbi]->md_overflow_pages += num;
1767 dp->p.mp_pages = num;
1774 mdb_leaf_size(MDB_env *env, MDB_val *key, MDB_val *data)
1778 sz = LEAFSIZE(key, data);
1779 if (data->mv_size >= env->me_meta.mm_psize / MDB_MINKEYS) {
1780 /* put on overflow page */
1781 sz -= data->mv_size - sizeof(pgno_t);
1784 return sz + sizeof(indx_t);
1788 mdb_branch_size(MDB_env *env, MDB_val *key)
1793 if (sz >= env->me_meta.mm_psize / MDB_MINKEYS) {
1794 /* put on overflow page */
1795 /* not implemented */
1796 /* sz -= key->size - sizeof(pgno_t); */
1799 return sz + sizeof(indx_t);
1803 mdb_add_node(MDB_txn *txn, MDB_dbi dbi, MDB_page *mp, indx_t indx,
1804 MDB_val *key, MDB_val *data, pgno_t pgno, uint8_t flags)
1807 size_t node_size = NODESIZE;
1810 MDB_dpage *ofp = NULL; /* overflow page */
1812 assert(mp->mp_upper >= mp->mp_lower);
1814 DPRINTF("add node [%.*s] to %s page %lu at index %i, key size %zu",
1815 key ? (int)key->mv_size : 0, key ? (char *)key->mv_data : NULL,
1816 IS_LEAF(mp) ? "leaf" : "branch",
1817 mp->mp_pgno, indx, key ? key->mv_size : 0);
1820 node_size += key->mv_size;
1824 if (F_ISSET(flags, F_BIGDATA)) {
1825 /* Data already on overflow page. */
1826 node_size += sizeof(pgno_t);
1827 } else if (data->mv_size >= txn->mt_env->me_meta.mm_psize / MDB_MINKEYS) {
1828 int ovpages = OVPAGES(data->mv_size, txn->mt_env->me_meta.mm_psize);
1829 /* Put data on overflow page. */
1830 DPRINTF("data size is %zu, put on overflow page",
1832 node_size += sizeof(pgno_t);
1833 if ((ofp = mdb_new_page(txn, dbi, P_OVERFLOW, ovpages)) == NULL)
1835 DPRINTF("allocated overflow page %lu", ofp->p.mp_pgno);
1838 node_size += data->mv_size;
1842 if (node_size + sizeof(indx_t) > SIZELEFT(mp)) {
1843 DPRINTF("not enough room in page %lu, got %u ptrs",
1844 mp->mp_pgno, NUMKEYS(mp));
1845 DPRINTF("upper - lower = %u - %u = %u", mp->mp_upper, mp->mp_lower,
1846 mp->mp_upper - mp->mp_lower);
1847 DPRINTF("node size = %zu", node_size);
1851 /* Move higher pointers up one slot. */
1852 for (i = NUMKEYS(mp); i > indx; i--)
1853 mp->mp_ptrs[i] = mp->mp_ptrs[i - 1];
1855 /* Adjust free space offsets. */
1856 ofs = mp->mp_upper - node_size;
1857 assert(ofs >= mp->mp_lower + sizeof(indx_t));
1858 mp->mp_ptrs[indx] = ofs;
1860 mp->mp_lower += sizeof(indx_t);
1862 /* Write the node data. */
1863 node = NODEPTR(mp, indx);
1864 node->mn_ksize = (key == NULL) ? 0 : key->mv_size;
1865 node->mn_flags = flags;
1867 node->mn_dsize = data->mv_size;
1869 node->mn_pgno = pgno;
1872 memcpy(NODEKEY(node), key->mv_data, key->mv_size);
1877 if (F_ISSET(flags, F_BIGDATA))
1878 memcpy(node->mn_data + key->mv_size, data->mv_data,
1881 memcpy(node->mn_data + key->mv_size, data->mv_data,
1884 memcpy(node->mn_data + key->mv_size, &ofp->p.mp_pgno,
1886 memcpy(METADATA(&ofp->p), data->mv_data, data->mv_size);
1894 mdb_del_node(MDB_page *mp, indx_t indx)
1897 indx_t i, j, numkeys, ptr;
1901 DPRINTF("delete node %u on %s page %lu", indx,
1902 IS_LEAF(mp) ? "leaf" : "branch", mp->mp_pgno);
1903 assert(indx < NUMKEYS(mp));
1905 node = NODEPTR(mp, indx);
1906 sz = NODESIZE + node->mn_ksize;
1908 if (F_ISSET(node->mn_flags, F_BIGDATA))
1909 sz += sizeof(pgno_t);
1911 sz += NODEDSZ(node);
1914 ptr = mp->mp_ptrs[indx];
1915 numkeys = NUMKEYS(mp);
1916 for (i = j = 0; i < numkeys; i++) {
1918 mp->mp_ptrs[j] = mp->mp_ptrs[i];
1919 if (mp->mp_ptrs[i] < ptr)
1920 mp->mp_ptrs[j] += sz;
1925 base = (char *)mp + mp->mp_upper;
1926 memmove(base + sz, base, ptr - mp->mp_upper);
1928 mp->mp_lower -= sizeof(indx_t);
1933 mdb_cursor_open(MDB_txn *txn, MDB_dbi dbi, MDB_cursor **ret)
1937 if (txn == NULL || ret == NULL)
1940 if ((cursor = calloc(1, sizeof(*cursor))) != NULL) {
1941 SLIST_INIT(&cursor->mc_stack);
1942 cursor->mc_dbi = dbi;
1943 cursor->mc_txn = txn;
1952 mdb_cursor_close(MDB_cursor *cursor)
1954 if (cursor != NULL) {
1955 while (!CURSOR_EMPTY(cursor))
1956 cursor_pop_page(cursor);
1958 /* btree_close(cursor->bt); */
1964 mdb_update_key(MDB_page *mp, indx_t indx, MDB_val *key)
1966 indx_t ptr, i, numkeys;
1972 node = NODEPTR(mp, indx);
1973 ptr = mp->mp_ptrs[indx];
1974 DPRINTF("update key %u (ofs %u) [%.*s] to [%.*s] on page %lu",
1976 (int)node->mn_ksize, (char *)NODEKEY(node),
1977 (int)key->mv_size, (char *)key->mv_data,
1980 delta = key->mv_size - node->mn_ksize;
1982 if (delta > 0 && SIZELEFT(mp) < delta) {
1983 DPRINTF("OUCH! Not enough room, delta = %d", delta);
1987 numkeys = NUMKEYS(mp);
1988 for (i = 0; i < numkeys; i++) {
1989 if (mp->mp_ptrs[i] <= ptr)
1990 mp->mp_ptrs[i] -= delta;
1993 base = (char *)mp + mp->mp_upper;
1994 len = ptr - mp->mp_upper + NODESIZE;
1995 memmove(base - delta, base, len);
1996 mp->mp_upper -= delta;
1998 node = NODEPTR(mp, indx);
1999 node->mn_ksize = key->mv_size;
2002 memcpy(NODEKEY(node), key->mv_data, key->mv_size);
2007 /* Move a node from src to dst.
2010 mdb_move_node(MDB_txn *txn, MDB_dbi dbi, MDB_pageparent *src, indx_t srcindx,
2011 MDB_pageparent *dst, indx_t dstindx)
2017 srcnode = NODEPTR(src->mp_page, srcindx);
2018 DPRINTF("moving %s node %u [%.*s] on page %lu to node %u on page %lu",
2019 IS_LEAF(src->mp_page) ? "leaf" : "branch",
2021 (int)srcnode->mn_ksize, (char *)NODEKEY(srcnode),
2022 src->mp_page->mp_pgno,
2023 dstindx, dst->mp_page->mp_pgno);
2025 /* Mark src and dst as dirty. */
2026 if ((rc = mdb_touch(txn, src)) ||
2027 (rc = mdb_touch(txn, dst)))
2030 /* Add the node to the destination page.
2032 key.mv_size = srcnode->mn_ksize;
2033 key.mv_data = NODEKEY(srcnode);
2034 data.mv_size = NODEDSZ(srcnode);
2035 data.mv_data = NODEDATA(srcnode);
2036 rc = mdb_add_node(txn, dbi, dst->mp_page, dstindx, &key, &data, NODEPGNO(srcnode),
2038 if (rc != MDB_SUCCESS)
2041 /* Delete the node from the source page.
2043 mdb_del_node(src->mp_page, srcindx);
2045 /* Update the parent separators.
2047 if (srcindx == 0 && src->mp_pi != 0) {
2048 DPRINTF("update separator for source page %lu to [%.*s]",
2049 src->mp_page->mp_pgno, (int)key.mv_size, (char *)key.mv_data);
2050 if ((rc = mdb_update_key(src->mp_parent, src->mp_pi,
2051 &key)) != MDB_SUCCESS)
2055 if (srcindx == 0 && IS_BRANCH(src->mp_page)) {
2057 nullkey.mv_size = 0;
2058 assert(mdb_update_key(src->mp_page, 0, &nullkey) == MDB_SUCCESS);
2061 if (dstindx == 0 && dst->mp_pi != 0) {
2062 DPRINTF("update separator for destination page %lu to [%.*s]",
2063 dst->mp_page->mp_pgno, (int)key.mv_size, (char *)key.mv_data);
2064 if ((rc = mdb_update_key(dst->mp_parent, dst->mp_pi,
2065 &key)) != MDB_SUCCESS)
2069 if (dstindx == 0 && IS_BRANCH(dst->mp_page)) {
2071 nullkey.mv_size = 0;
2072 assert(mdb_update_key(dst->mp_page, 0, &nullkey) == MDB_SUCCESS);
2079 mdb_merge(MDB_txn *txn, MDB_dbi dbi, MDB_pageparent *src, MDB_pageparent *dst)
2088 DPRINTF("merging page %lu and %lu", src->mp_page->mp_pgno, dst->mp_page->mp_pgno);
2090 assert(txn != NULL);
2091 assert(src->mp_parent); /* can't merge root page */
2092 assert(dst->mp_parent);
2094 /* Mark src and dst as dirty. */
2095 if ((rc = mdb_touch(txn, src)) ||
2096 (rc = mdb_touch(txn, dst)))
2099 /* Move all nodes from src to dst.
2101 for (i = 0; i < NUMKEYS(src->mp_page); i++) {
2102 srcnode = NODEPTR(src->mp_page, i);
2104 key.mv_size = srcnode->mn_ksize;
2105 key.mv_data = NODEKEY(srcnode);
2106 data.mv_size = NODEDSZ(srcnode);
2107 data.mv_data = NODEDATA(srcnode);
2108 rc = mdb_add_node(txn, dbi, dst->mp_page, NUMKEYS(dst->mp_page), &key,
2109 &data, NODEPGNO(srcnode), srcnode->mn_flags);
2110 if (rc != MDB_SUCCESS)
2114 DPRINTF("dst page %lu now has %u keys (%.1f%% filled)",
2115 dst->mp_page->mp_pgno, NUMKEYS(dst->mp_page), (float)PAGEFILL(txn->mt_env, dst->mp_page) / 10);
2117 /* Unlink the src page from parent.
2119 mdb_del_node(src->mp_parent, src->mp_pi);
2120 if (src->mp_pi == 0) {
2122 if ((rc = mdb_update_key(src->mp_parent, 0, &key)) != MDB_SUCCESS)
2126 if (IS_LEAF(src->mp_page))
2127 txn->mt_dbs[dbi]->md_leaf_pages--;
2129 txn->mt_dbs[dbi]->md_branch_pages--;
2131 mpp.mp_page = src->mp_parent;
2132 dh = (MDB_dhead *)src->mp_parent;
2134 mpp.mp_parent = dh->md_parent;
2135 mpp.mp_pi = dh->md_pi;
2137 return mdb_rebalance(txn, dbi, &mpp);
2140 #define FILL_THRESHOLD 250
2143 mdb_rebalance(MDB_txn *txn, MDB_dbi dbi, MDB_pageparent *mpp)
2148 indx_t si = 0, di = 0;
2150 assert(txn != NULL);
2151 assert(mpp != NULL);
2153 DPRINTF("rebalancing %s page %lu (has %u keys, %.1f%% full)",
2154 IS_LEAF(mpp->mp_page) ? "leaf" : "branch",
2155 mpp->mp_page->mp_pgno, NUMKEYS(mpp->mp_page), (float)PAGEFILL(txn->mt_env, mpp->mp_page) / 10);
2157 if (PAGEFILL(txn->mt_env, mpp->mp_page) >= FILL_THRESHOLD) {
2158 DPRINTF("no need to rebalance page %lu, above fill threshold",
2159 mpp->mp_page->mp_pgno);
2163 if (mpp->mp_parent == NULL) {
2164 if (NUMKEYS(mpp->mp_page) == 0) {
2165 DPRINTF("tree is completely empty");
2166 txn->mt_dbs[dbi]->md_root = P_INVALID;
2167 txn->mt_dbs[dbi]->md_depth--;
2168 txn->mt_dbs[dbi]->md_leaf_pages--;
2169 } else if (IS_BRANCH(mpp->mp_page) && NUMKEYS(mpp->mp_page) == 1) {
2170 DPRINTF("collapsing root page!");
2171 txn->mt_dbs[dbi]->md_root = NODEPGNO(NODEPTR(mpp->mp_page, 0));
2172 if ((root = mdbenv_get_page(txn->mt_env, txn->mt_dbs[dbi]->md_root)) == NULL)
2174 txn->mt_dbs[dbi]->md_depth--;
2175 txn->mt_dbs[dbi]->md_branch_pages--;
2177 DPRINTF("root page doesn't need rebalancing");
2181 /* The parent (branch page) must have at least 2 pointers,
2182 * otherwise the tree is invalid.
2184 assert(NUMKEYS(mpp->mp_parent) > 1);
2186 /* Leaf page fill factor is below the threshold.
2187 * Try to move keys from left or right neighbor, or
2188 * merge with a neighbor page.
2193 if (mpp->mp_pi == 0) {
2194 /* We're the leftmost leaf in our parent.
2196 DPRINTF("reading right neighbor");
2197 node = NODEPTR(mpp->mp_parent, mpp->mp_pi + 1);
2198 if ((npp.mp_page = mdbenv_get_page(txn->mt_env, NODEPGNO(node))) == NULL)
2200 npp.mp_pi = mpp->mp_pi + 1;
2202 di = NUMKEYS(mpp->mp_page);
2204 /* There is at least one neighbor to the left.
2206 DPRINTF("reading left neighbor");
2207 node = NODEPTR(mpp->mp_parent, mpp->mp_pi - 1);
2208 if ((npp.mp_page = mdbenv_get_page(txn->mt_env, NODEPGNO(node))) == NULL)
2210 npp.mp_pi = mpp->mp_pi - 1;
2211 si = NUMKEYS(npp.mp_page) - 1;
2214 npp.mp_parent = mpp->mp_parent;
2216 DPRINTF("found neighbor page %lu (%u keys, %.1f%% full)",
2217 npp.mp_page->mp_pgno, NUMKEYS(npp.mp_page), (float)PAGEFILL(txn->mt_env, npp.mp_page) / 10);
2219 /* If the neighbor page is above threshold and has at least two
2220 * keys, move one key from it.
2222 * Otherwise we should try to merge them.
2224 if (PAGEFILL(txn->mt_env, npp.mp_page) >= FILL_THRESHOLD && NUMKEYS(npp.mp_page) >= 2)
2225 return mdb_move_node(txn, dbi, &npp, si, mpp, di);
2226 else { /* FIXME: if (has_enough_room()) */
2227 if (mpp->mp_pi == 0)
2228 return mdb_merge(txn, dbi, &npp, mpp);
2230 return mdb_merge(txn, dbi, mpp, &npp);
2235 mdb_del(MDB_txn *txn, MDB_dbi dbi,
2236 MDB_val *key, MDB_val *data)
2243 DPRINTF("========> delete key %.*s", (int)key->mv_size, (char *)key->mv_data);
2245 assert(key != NULL);
2247 if (txn == NULL || dbi >= txn->mt_numdbs)
2250 if (F_ISSET(txn->mt_flags, MDB_TXN_RDONLY)) {
2254 if (key->mv_size == 0 || key->mv_size > MAXKEYSIZE) {
2258 if ((rc = mdb_search_page(txn, dbi, key, NULL, 1, &mpp)) != MDB_SUCCESS)
2261 leaf = mdb_search_node(txn, dbi, mpp.mp_page, key, &exact, &ki);
2262 if (leaf == NULL || !exact) {
2266 if (data && (rc = mdb_read_data(txn->mt_env, leaf, data)) != MDB_SUCCESS)
2269 mdb_del_node(mpp.mp_page, ki);
2270 /* add overflow pages to free list */
2271 if (F_ISSET(leaf->mn_flags, F_BIGDATA)) {
2275 memcpy(&pg, NODEDATA(leaf), sizeof(pg));
2276 ovpages = OVPAGES(NODEDSZ(leaf), txn->mt_env->me_meta.mm_psize);
2277 for (i=0; i<ovpages; i++) {
2278 mdb_idl_insert(txn->mt_free_pgs, pg);
2282 txn->mt_dbs[dbi]->md_entries--;
2283 rc = mdb_rebalance(txn, dbi, &mpp);
2284 if (rc != MDB_SUCCESS)
2285 txn->mt_flags |= MDB_TXN_ERROR;
2290 /* Split page <*mpp>, and insert <key,(data|newpgno)> in either left or
2291 * right sibling, at index <*newindxp> (as if unsplit). Updates *mpp and
2292 * *newindxp with the actual values after split, ie if *mpp and *newindxp
2293 * refer to a node in the new right sibling page.
2296 mdb_split(MDB_txn *txn, MDB_dbi dbi, MDB_page **mpp, unsigned int *newindxp,
2297 MDB_val *newkey, MDB_val *newdata, pgno_t newpgno)
2300 int rc = MDB_SUCCESS, ins_new = 0;
2303 unsigned int i, j, split_indx;
2305 MDB_val sepkey, rkey, rdata;
2307 MDB_dpage *mdp, *rdp, *pdp;
2310 assert(txn != NULL);
2312 dh = ((MDB_dhead *)*mpp) - 1;
2313 mdp = (MDB_dpage *)dh;
2314 newindx = *newindxp;
2316 DPRINTF("-----> splitting %s page %lu and adding [%.*s] at index %i",
2317 IS_LEAF(&mdp->p) ? "leaf" : "branch", mdp->p.mp_pgno,
2318 (int)newkey->mv_size, (char *)newkey->mv_data, *newindxp);
2320 if (mdp->h.md_parent == NULL) {
2321 if ((pdp = mdb_new_page(txn, dbi, P_BRANCH, 1)) == NULL)
2324 mdp->h.md_parent = &pdp->p;
2325 txn->mt_dbs[dbi]->md_root = pdp->p.mp_pgno;
2326 DPRINTF("root split! new root = %lu", pdp->p.mp_pgno);
2327 txn->mt_dbs[dbi]->md_depth++;
2329 /* Add left (implicit) pointer. */
2330 if (mdb_add_node(txn, dbi, &pdp->p, 0, NULL, NULL,
2331 mdp->p.mp_pgno, 0) != MDB_SUCCESS)
2334 DPRINTF("parent branch page is %lu", mdp->h.md_parent->mp_pgno);
2337 /* Create a right sibling. */
2338 if ((rdp = mdb_new_page(txn, dbi, mdp->p.mp_flags, 1)) == NULL)
2340 rdp->h.md_parent = mdp->h.md_parent;
2341 rdp->h.md_pi = mdp->h.md_pi + 1;
2342 DPRINTF("new right sibling: page %lu", rdp->p.mp_pgno);
2344 /* Move half of the keys to the right sibling. */
2345 if ((copy = malloc(txn->mt_env->me_meta.mm_psize)) == NULL)
2347 memcpy(copy, &mdp->p, txn->mt_env->me_meta.mm_psize);
2348 memset(&mdp->p.mp_ptrs, 0, txn->mt_env->me_meta.mm_psize - PAGEHDRSZ);
2349 mdp->p.mp_lower = PAGEHDRSZ;
2350 mdp->p.mp_upper = txn->mt_env->me_meta.mm_psize;
2352 split_indx = NUMKEYS(copy) / 2 + 1;
2354 /* First find the separating key between the split pages.
2356 memset(&sepkey, 0, sizeof(sepkey));
2357 if (newindx == split_indx) {
2358 sepkey.mv_size = newkey->mv_size;
2359 sepkey.mv_data = newkey->mv_data;
2361 node = NODEPTR(copy, split_indx);
2362 sepkey.mv_size = node->mn_ksize;
2363 sepkey.mv_data = NODEKEY(node);
2366 DPRINTF("separator is [%.*s]", (int)sepkey.mv_size, (char *)sepkey.mv_data);
2368 /* Copy separator key to the parent.
2370 if (SIZELEFT(rdp->h.md_parent) < mdb_branch_size(txn->mt_env, &sepkey)) {
2371 rc = mdb_split(txn, dbi, &rdp->h.md_parent, &rdp->h.md_pi,
2372 &sepkey, NULL, rdp->p.mp_pgno);
2374 /* Right page might now have changed parent.
2375 * Check if left page also changed parent.
2377 if (rdp->h.md_parent != mdp->h.md_parent &&
2378 mdp->h.md_pi >= NUMKEYS(mdp->h.md_parent)) {
2379 mdp->h.md_parent = rdp->h.md_parent;
2380 mdp->h.md_pi = rdp->h.md_pi - 1;
2383 rc = mdb_add_node(txn, dbi, rdp->h.md_parent, rdp->h.md_pi,
2384 &sepkey, NULL, rdp->p.mp_pgno, 0);
2386 if (rc != MDB_SUCCESS) {
2391 for (i = j = 0; i <= NUMKEYS(copy); j++) {
2392 if (i < split_indx) {
2393 /* Re-insert in left sibling. */
2396 /* Insert in right sibling. */
2397 if (i == split_indx)
2398 /* Reset insert index for right sibling. */
2399 j = (i == newindx && ins_new);
2403 if (i == newindx && !ins_new) {
2404 /* Insert the original entry that caused the split. */
2405 rkey.mv_data = newkey->mv_data;
2406 rkey.mv_size = newkey->mv_size;
2407 if (IS_LEAF(&mdp->p)) {
2408 rdata.mv_data = newdata->mv_data;
2409 rdata.mv_size = newdata->mv_size;
2416 /* Update page and index for the new key. */
2419 } else if (i == NUMKEYS(copy)) {
2422 node = NODEPTR(copy, i);
2423 rkey.mv_data = NODEKEY(node);
2424 rkey.mv_size = node->mn_ksize;
2425 if (IS_LEAF(&mdp->p)) {
2426 rdata.mv_data = NODEDATA(node);
2427 rdata.mv_size = node->mn_dsize;
2429 pgno = node->mn_pgno;
2430 flags = node->mn_flags;
2435 if (!IS_LEAF(&mdp->p) && j == 0) {
2436 /* First branch index doesn't need key data. */
2440 rc = mdb_add_node(txn, dbi, &pdp->p, j, &rkey, &rdata, pgno,flags);
2448 mdb_put(MDB_txn *txn, MDB_dbi dbi,
2449 MDB_val *key, MDB_val *data, unsigned int flags)
2451 int rc = MDB_SUCCESS, exact;
2456 assert(key != NULL);
2457 assert(data != NULL);
2462 if (F_ISSET(txn->mt_flags, MDB_TXN_RDONLY)) {
2466 if (txn->mt_env->me_txn != txn) {
2470 if (key->mv_size == 0 || key->mv_size > MAXKEYSIZE) {
2474 DPRINTF("==> put key %.*s, size %zu, data size %zu",
2475 (int)key->mv_size, (char *)key->mv_data, key->mv_size, data->mv_size);
2477 rc = mdb_search_page(txn, dbi, key, NULL, 1, &mpp);
2478 if (rc == MDB_SUCCESS) {
2479 leaf = mdb_search_node(txn, dbi, mpp.mp_page, key, &exact, &ki);
2480 if (leaf && exact) {
2481 if (F_ISSET(flags, MDB_NOOVERWRITE)) {
2482 DPRINTF("duplicate key %.*s",
2483 (int)key->mv_size, (char *)key->mv_data);
2486 mdb_del_node(mpp.mp_page, ki);
2488 if (leaf == NULL) { /* append if not found */
2489 ki = NUMKEYS(mpp.mp_page);
2490 DPRINTF("appending key at index %i", ki);
2492 } else if (rc == ENOENT) {
2494 /* new file, just write a root leaf page */
2495 DPRINTF("allocating new root leaf page");
2496 if ((dp = mdb_new_page(txn, dbi, P_LEAF, 1)) == NULL) {
2499 mpp.mp_page = &dp->p;
2500 txn->mt_dbs[dbi]->md_root = mpp.mp_page->mp_pgno;
2501 txn->mt_dbs[dbi]->md_depth++;
2507 assert(IS_LEAF(mpp.mp_page));
2508 DPRINTF("there are %u keys, should insert new key at index %i",
2509 NUMKEYS(mpp.mp_page), ki);
2511 if (SIZELEFT(mpp.mp_page) < mdb_leaf_size(txn->mt_env, key, data)) {
2512 rc = mdb_split(txn, dbi, &mpp.mp_page, &ki, key, data, P_INVALID);
2514 /* There is room already in this leaf page. */
2515 rc = mdb_add_node(txn, dbi, mpp.mp_page, ki, key, data, 0, 0);
2518 if (rc != MDB_SUCCESS)
2519 txn->mt_flags |= MDB_TXN_ERROR;
2521 txn->mt_dbs[dbi]->md_entries++;
2528 mdbenv_get_flags(MDB_env *env, unsigned int *arg)
2533 *arg = env->me_flags;
2538 mdbenv_get_path(MDB_env *env, const char **arg)
2543 *arg = env->me_path;
2548 mdbenv_stat(MDB_env *env, MDB_stat *arg)
2550 if (env == NULL || arg == NULL)
2553 arg->ms_psize = env->me_meta.mm_psize;
2554 arg->ms_depth = env->me_meta.mm_depth;
2555 arg->ms_branch_pages = env->me_meta.mm_branch_pages;
2556 arg->ms_leaf_pages = env->me_meta.mm_leaf_pages;
2557 arg->ms_overflow_pages = env->me_meta.mm_overflow_pages;
2558 arg->ms_entries = env->me_meta.mm_entries;
2563 int mdb_open(MDB_txn *txn, const char *name, unsigned int flags, MDB_dbi *dbi)
2575 /* Is the DB already open? */
2576 for (i=0; i<txn->mt_numdbs; i++) {
2577 if (!strcmp(name, txn->mt_dbxs[i].md_name)) {
2583 /* Find the DB info */
2584 key.mv_size = strlen(name);
2585 key.mv_data = (void *)name;
2586 rc = mdb_get(txn, 0, &key, &data);
2588 /* Create if requested */
2589 if (rc == ENOENT && (flags & MDB_CREATE)) {
2591 data.mv_size = sizeof(MDB_db);
2592 data.mv_data = &dummy;
2593 memset(&dummy, 0, sizeof(dummy));
2594 dummy.md_root = P_INVALID;
2595 rc = mdb_put(txn, 0, &key, &data, 0);
2596 if (rc == MDB_SUCCESS)
2597 rc = mdb_get(txn, 0, &key, &data);
2600 /* OK, got info, add to table */
2601 if (rc == MDB_SUCCESS) {
2602 /* Is there a free slot? */
2603 if ((txn->mt_numdbs & (DBX_CHUNK-1)) == 0) {
2607 i = txn->mt_numdbs + DBX_CHUNK;
2608 p1 = realloc(txn->mt_dbxs, i * sizeof(MDB_dbx));
2612 p2 = realloc(txn->mt_dbs, i * sizeof(MDB_db *));
2617 txn->mt_dbxs[txn->mt_numdbs].md_name = strdup(name);
2618 txn->mt_dbxs[txn->mt_numdbs].md_cmp = NULL;
2619 txn->mt_dbxs[txn->mt_numdbs].md_rel = NULL;
2620 txn->mt_dbs[txn->mt_numdbs] = data.mv_data;
2621 *dbi = txn->mt_numdbs;
2628 int mdb_stat(MDB_txn *txn, MDB_dbi dbi, MDB_stat *arg)
2630 if (txn == NULL || arg == NULL)
2633 arg->ms_psize = txn->mt_env->me_meta.mm_psize;
2634 arg->ms_depth = txn->mt_dbs[dbi]->md_depth;
2635 arg->ms_branch_pages = txn->mt_dbs[dbi]->md_branch_pages;
2636 arg->ms_leaf_pages = txn->mt_dbs[dbi]->md_leaf_pages;
2637 arg->ms_overflow_pages = txn->mt_dbs[dbi]->md_overflow_pages;
2638 arg->ms_entries = txn->mt_dbs[dbi]->md_entries;
2643 void mdb_close(MDB_txn *txn, MDB_dbi dbi)
2645 if (dbi >= txn->mt_numdbs)
2647 free(txn->mt_dbxs[dbi].md_name);
2648 txn->mt_dbxs[dbi].md_name = NULL;