bt_delete.c revision 1.7.2.1 1 1.7.2.1 jtc /* $NetBSD: bt_delete.c,v 1.7.2.1 1996/09/16 18:39:36 jtc Exp $ */
2 1.6 cgd
3 1.1 cgd /*-
4 1.5 cgd * Copyright (c) 1990, 1993, 1994
5 1.1 cgd * The Regents of the University of California. All rights reserved.
6 1.1 cgd *
7 1.1 cgd * This code is derived from software contributed to Berkeley by
8 1.1 cgd * Mike Olson.
9 1.1 cgd *
10 1.1 cgd * Redistribution and use in source and binary forms, with or without
11 1.1 cgd * modification, are permitted provided that the following conditions
12 1.1 cgd * are met:
13 1.1 cgd * 1. Redistributions of source code must retain the above copyright
14 1.1 cgd * notice, this list of conditions and the following disclaimer.
15 1.1 cgd * 2. Redistributions in binary form must reproduce the above copyright
16 1.1 cgd * notice, this list of conditions and the following disclaimer in the
17 1.1 cgd * documentation and/or other materials provided with the distribution.
18 1.1 cgd * 3. All advertising materials mentioning features or use of this software
19 1.1 cgd * must display the following acknowledgement:
20 1.1 cgd * This product includes software developed by the University of
21 1.1 cgd * California, Berkeley and its contributors.
22 1.1 cgd * 4. Neither the name of the University nor the names of its contributors
23 1.1 cgd * may be used to endorse or promote products derived from this software
24 1.1 cgd * without specific prior written permission.
25 1.1 cgd *
26 1.1 cgd * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
27 1.1 cgd * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
28 1.1 cgd * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
29 1.1 cgd * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
30 1.1 cgd * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
31 1.1 cgd * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
32 1.1 cgd * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
33 1.1 cgd * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
34 1.1 cgd * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
35 1.1 cgd * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
36 1.1 cgd * SUCH DAMAGE.
37 1.1 cgd */
38 1.1 cgd
39 1.1 cgd #if defined(LIBC_SCCS) && !defined(lint)
40 1.6 cgd #if 0
41 1.7 cgd static char sccsid[] = "@(#)bt_delete.c 8.13 (Berkeley) 7/28/94";
42 1.6 cgd #else
43 1.7.2.1 jtc static char rcsid[] = "$NetBSD: bt_delete.c,v 1.7.2.1 1996/09/16 18:39:36 jtc Exp $";
44 1.6 cgd #endif
45 1.1 cgd #endif /* LIBC_SCCS and not lint */
46 1.1 cgd
47 1.7.2.1 jtc #include "namespace.h"
48 1.1 cgd #include <sys/types.h>
49 1.1 cgd
50 1.1 cgd #include <errno.h>
51 1.1 cgd #include <stdio.h>
52 1.1 cgd #include <string.h>
53 1.1 cgd
54 1.1 cgd #include <db.h>
55 1.1 cgd #include "btree.h"
56 1.1 cgd
57 1.7 cgd static int __bt_bdelete __P((BTREE *, const DBT *));
58 1.7 cgd static int __bt_curdel __P((BTREE *, const DBT *, PAGE *, u_int));
59 1.7 cgd static int __bt_pdelete __P((BTREE *, PAGE *));
60 1.7 cgd static int __bt_relink __P((BTREE *, PAGE *));
61 1.7 cgd static int __bt_stkacq __P((BTREE *, PAGE **, CURSOR *));
62 1.1 cgd
63 1.1 cgd /*
64 1.7 cgd * __bt_delete
65 1.7 cgd * Delete the item(s) referenced by a key.
66 1.1 cgd *
67 1.7 cgd * Return RET_SPECIAL if the key is not found.
68 1.1 cgd */
69 1.1 cgd int
70 1.1 cgd __bt_delete(dbp, key, flags)
71 1.1 cgd const DB *dbp;
72 1.1 cgd const DBT *key;
73 1.1 cgd u_int flags;
74 1.1 cgd {
75 1.1 cgd BTREE *t;
76 1.7 cgd CURSOR *c;
77 1.7 cgd PAGE *h;
78 1.1 cgd int status;
79 1.1 cgd
80 1.1 cgd t = dbp->internal;
81 1.4 cgd
82 1.4 cgd /* Toss any page pinned across calls. */
83 1.4 cgd if (t->bt_pinned != NULL) {
84 1.4 cgd mpool_put(t->bt_mp, t->bt_pinned, 0);
85 1.4 cgd t->bt_pinned = NULL;
86 1.4 cgd }
87 1.4 cgd
88 1.7 cgd /* Check for change to a read-only tree. */
89 1.7 cgd if (F_ISSET(t, B_RDONLY)) {
90 1.1 cgd errno = EPERM;
91 1.1 cgd return (RET_ERROR);
92 1.1 cgd }
93 1.4 cgd
94 1.7 cgd switch (flags) {
95 1.1 cgd case 0:
96 1.7 cgd status = __bt_bdelete(t, key);
97 1.1 cgd break;
98 1.1 cgd case R_CURSOR:
99 1.1 cgd /*
100 1.7 cgd * If flags is R_CURSOR, delete the cursor. Must already
101 1.7 cgd * have started a scan and not have already deleted it.
102 1.1 cgd */
103 1.7 cgd c = &t->bt_cursor;
104 1.7 cgd if (F_ISSET(c, CURS_INIT)) {
105 1.7 cgd if (F_ISSET(c, CURS_ACQUIRE | CURS_AFTER | CURS_BEFORE))
106 1.7 cgd return (RET_SPECIAL);
107 1.7 cgd if ((h = mpool_get(t->bt_mp, c->pg.pgno, 0)) == NULL)
108 1.7 cgd return (RET_ERROR);
109 1.7 cgd
110 1.7 cgd /*
111 1.7 cgd * If the page is about to be emptied, we'll need to
112 1.7 cgd * delete it, which means we have to acquire a stack.
113 1.7 cgd */
114 1.7 cgd if (NEXTINDEX(h) == 1)
115 1.7 cgd if (__bt_stkacq(t, &h, &t->bt_cursor))
116 1.7 cgd return (RET_ERROR);
117 1.7 cgd
118 1.7 cgd status = __bt_dleaf(t, NULL, h, c->pg.index);
119 1.7 cgd
120 1.7 cgd if (NEXTINDEX(h) == 0 && status == RET_SUCCESS) {
121 1.7 cgd if (__bt_pdelete(t, h))
122 1.7 cgd return (RET_ERROR);
123 1.7 cgd } else
124 1.7 cgd mpool_put(t->bt_mp,
125 1.7 cgd h, status == RET_SUCCESS ? MPOOL_DIRTY : 0);
126 1.7 cgd break;
127 1.7 cgd }
128 1.7 cgd /* FALLTHROUGH */
129 1.1 cgd default:
130 1.7 cgd errno = EINVAL;
131 1.1 cgd return (RET_ERROR);
132 1.1 cgd }
133 1.1 cgd if (status == RET_SUCCESS)
134 1.7 cgd F_SET(t, B_MODIFIED);
135 1.1 cgd return (status);
136 1.1 cgd }
137 1.1 cgd
138 1.1 cgd /*
139 1.7 cgd * __bt_stkacq --
140 1.7 cgd * Acquire a stack so we can delete a cursor entry.
141 1.7 cgd *
142 1.7 cgd * Parameters:
143 1.7 cgd * t: tree
144 1.7 cgd * hp: pointer to current, pinned PAGE pointer
145 1.7 cgd * c: pointer to the cursor
146 1.7 cgd *
147 1.7 cgd * Returns:
148 1.7 cgd * 0 on success, 1 on failure
149 1.7 cgd */
150 1.7 cgd static int
151 1.7 cgd __bt_stkacq(t, hp, c)
152 1.7 cgd BTREE *t;
153 1.7 cgd PAGE **hp;
154 1.7 cgd CURSOR *c;
155 1.7 cgd {
156 1.7 cgd BINTERNAL *bi;
157 1.7 cgd EPG *e;
158 1.7 cgd EPGNO *parent;
159 1.7 cgd PAGE *h;
160 1.7 cgd indx_t index;
161 1.7 cgd pgno_t pgno;
162 1.7 cgd recno_t nextpg, prevpg;
163 1.7 cgd int exact, level;
164 1.7 cgd
165 1.7 cgd /*
166 1.7 cgd * Find the first occurrence of the key in the tree. Toss the
167 1.7 cgd * currently locked page so we don't hit an already-locked page.
168 1.7 cgd */
169 1.7 cgd h = *hp;
170 1.7 cgd mpool_put(t->bt_mp, h, 0);
171 1.7 cgd if ((e = __bt_search(t, &c->key, &exact)) == NULL)
172 1.7 cgd return (1);
173 1.7 cgd h = e->page;
174 1.7 cgd
175 1.7 cgd /* See if we got it in one shot. */
176 1.7 cgd if (h->pgno == c->pg.pgno)
177 1.7 cgd goto ret;
178 1.7 cgd
179 1.7 cgd /*
180 1.7 cgd * Move right, looking for the page. At each move we have to move
181 1.7 cgd * up the stack until we don't have to move to the next page. If
182 1.7 cgd * we have to change pages at an internal level, we have to fix the
183 1.7 cgd * stack back up.
184 1.7 cgd */
185 1.7 cgd while (h->pgno != c->pg.pgno) {
186 1.7 cgd if ((nextpg = h->nextpg) == P_INVALID)
187 1.7 cgd break;
188 1.7 cgd mpool_put(t->bt_mp, h, 0);
189 1.7 cgd
190 1.7 cgd /* Move up the stack. */
191 1.7 cgd for (level = 0; (parent = BT_POP(t)) != NULL; ++level) {
192 1.7 cgd /* Get the parent page. */
193 1.7 cgd if ((h = mpool_get(t->bt_mp, parent->pgno, 0)) == NULL)
194 1.7 cgd return (1);
195 1.7 cgd
196 1.7 cgd /* Move to the next index. */
197 1.7 cgd if (parent->index != NEXTINDEX(h) - 1) {
198 1.7 cgd index = parent->index + 1;
199 1.7 cgd BT_PUSH(t, h->pgno, index);
200 1.7 cgd break;
201 1.7 cgd }
202 1.7 cgd mpool_put(t->bt_mp, h, 0);
203 1.7 cgd }
204 1.7 cgd
205 1.7 cgd /* Restore the stack. */
206 1.7 cgd while (level--) {
207 1.7 cgd /* Push the next level down onto the stack. */
208 1.7 cgd bi = GETBINTERNAL(h, index);
209 1.7 cgd pgno = bi->pgno;
210 1.7 cgd BT_PUSH(t, pgno, 0);
211 1.7 cgd
212 1.7 cgd /* Lose the currently pinned page. */
213 1.7 cgd mpool_put(t->bt_mp, h, 0);
214 1.7 cgd
215 1.7 cgd /* Get the next level down. */
216 1.7 cgd if ((h = mpool_get(t->bt_mp, pgno, 0)) == NULL)
217 1.7 cgd return (1);
218 1.7 cgd index = 0;
219 1.7 cgd }
220 1.7 cgd mpool_put(t->bt_mp, h, 0);
221 1.7 cgd if ((h = mpool_get(t->bt_mp, nextpg, 0)) == NULL)
222 1.7 cgd return (1);
223 1.7 cgd }
224 1.7 cgd
225 1.7 cgd if (h->pgno == c->pg.pgno)
226 1.7 cgd goto ret;
227 1.7 cgd
228 1.7 cgd /* Reacquire the original stack. */
229 1.7 cgd mpool_put(t->bt_mp, h, 0);
230 1.7 cgd if ((e = __bt_search(t, &c->key, &exact)) == NULL)
231 1.7 cgd return (1);
232 1.7 cgd h = e->page;
233 1.7 cgd
234 1.7 cgd /*
235 1.7 cgd * Move left, looking for the page. At each move we have to move
236 1.7 cgd * up the stack until we don't have to change pages to move to the
237 1.7 cgd * next page. If we have to change pages at an internal level, we
238 1.7 cgd * have to fix the stack back up.
239 1.7 cgd */
240 1.7 cgd while (h->pgno != c->pg.pgno) {
241 1.7 cgd if ((prevpg = h->prevpg) == P_INVALID)
242 1.7 cgd break;
243 1.7 cgd mpool_put(t->bt_mp, h, 0);
244 1.7 cgd
245 1.7 cgd /* Move up the stack. */
246 1.7 cgd for (level = 0; (parent = BT_POP(t)) != NULL; ++level) {
247 1.7 cgd /* Get the parent page. */
248 1.7 cgd if ((h = mpool_get(t->bt_mp, parent->pgno, 0)) == NULL)
249 1.7 cgd return (1);
250 1.7 cgd
251 1.7 cgd /* Move to the next index. */
252 1.7 cgd if (parent->index != 0) {
253 1.7 cgd index = parent->index - 1;
254 1.7 cgd BT_PUSH(t, h->pgno, index);
255 1.7 cgd break;
256 1.7 cgd }
257 1.7 cgd mpool_put(t->bt_mp, h, 0);
258 1.7 cgd }
259 1.7 cgd
260 1.7 cgd /* Restore the stack. */
261 1.7 cgd while (level--) {
262 1.7 cgd /* Push the next level down onto the stack. */
263 1.7 cgd bi = GETBINTERNAL(h, index);
264 1.7 cgd pgno = bi->pgno;
265 1.7 cgd
266 1.7 cgd /* Lose the currently pinned page. */
267 1.7 cgd mpool_put(t->bt_mp, h, 0);
268 1.7 cgd
269 1.7 cgd /* Get the next level down. */
270 1.7 cgd if ((h = mpool_get(t->bt_mp, pgno, 0)) == NULL)
271 1.7 cgd return (1);
272 1.7 cgd
273 1.7 cgd index = NEXTINDEX(h) - 1;
274 1.7 cgd BT_PUSH(t, pgno, index);
275 1.7 cgd }
276 1.7 cgd mpool_put(t->bt_mp, h, 0);
277 1.7 cgd if ((h = mpool_get(t->bt_mp, prevpg, 0)) == NULL)
278 1.7 cgd return (1);
279 1.7 cgd }
280 1.7 cgd
281 1.7 cgd
282 1.7 cgd ret: mpool_put(t->bt_mp, h, 0);
283 1.7 cgd return ((*hp = mpool_get(t->bt_mp, c->pg.pgno, 0)) == NULL);
284 1.7 cgd }
285 1.7 cgd
286 1.7 cgd /*
287 1.7 cgd * __bt_bdelete --
288 1.7 cgd * Delete all key/data pairs matching the specified key.
289 1.1 cgd *
290 1.1 cgd * Parameters:
291 1.7 cgd * t: tree
292 1.1 cgd * key: key to delete
293 1.1 cgd *
294 1.1 cgd * Returns:
295 1.1 cgd * RET_ERROR, RET_SUCCESS and RET_SPECIAL if the key not found.
296 1.1 cgd */
297 1.1 cgd static int
298 1.7 cgd __bt_bdelete(t, key)
299 1.1 cgd BTREE *t;
300 1.1 cgd const DBT *key;
301 1.1 cgd {
302 1.7 cgd EPG *e;
303 1.1 cgd PAGE *h;
304 1.7 cgd int deleted, exact, redo;
305 1.7 cgd
306 1.7 cgd deleted = 0;
307 1.1 cgd
308 1.1 cgd /* Find any matching record; __bt_search pins the page. */
309 1.7 cgd loop: if ((e = __bt_search(t, key, &exact)) == NULL)
310 1.7 cgd return (deleted ? RET_SUCCESS : RET_ERROR);
311 1.1 cgd if (!exact) {
312 1.1 cgd mpool_put(t->bt_mp, e->page, 0);
313 1.7 cgd return (deleted ? RET_SUCCESS : RET_SPECIAL);
314 1.1 cgd }
315 1.1 cgd
316 1.1 cgd /*
317 1.7 cgd * Delete forward, then delete backward, from the found key. If
318 1.7 cgd * there are duplicates and we reach either side of the page, do
319 1.7 cgd * the key search again, so that we get them all.
320 1.1 cgd */
321 1.7 cgd redo = 0;
322 1.7 cgd h = e->page;
323 1.7 cgd do {
324 1.7 cgd if (__bt_dleaf(t, key, h, e->index)) {
325 1.7 cgd mpool_put(t->bt_mp, h, 0);
326 1.7 cgd return (RET_ERROR);
327 1.7 cgd }
328 1.7 cgd if (F_ISSET(t, B_NODUPS)) {
329 1.7 cgd if (NEXTINDEX(h) == 0) {
330 1.7 cgd if (__bt_pdelete(t, h))
331 1.1 cgd return (RET_ERROR);
332 1.7 cgd } else
333 1.7 cgd mpool_put(t->bt_mp, h, MPOOL_DIRTY);
334 1.7 cgd return (RET_SUCCESS);
335 1.7 cgd }
336 1.7 cgd deleted = 1;
337 1.7 cgd } while (e->index < NEXTINDEX(h) && __bt_cmp(t, key, e) == 0);
338 1.7 cgd
339 1.7 cgd /* Check for right-hand edge of the page. */
340 1.7 cgd if (e->index == NEXTINDEX(h))
341 1.7 cgd redo = 1;
342 1.1 cgd
343 1.7 cgd /* Delete from the key to the beginning of the page. */
344 1.7 cgd while (e->index-- > 0) {
345 1.7 cgd if (__bt_cmp(t, key, e) != 0)
346 1.1 cgd break;
347 1.7 cgd if (__bt_dleaf(t, key, h, e->index) == RET_ERROR) {
348 1.7 cgd mpool_put(t->bt_mp, h, 0);
349 1.7 cgd return (RET_ERROR);
350 1.1 cgd }
351 1.7 cgd if (e->index == 0)
352 1.7 cgd redo = 1;
353 1.7 cgd }
354 1.1 cgd
355 1.7 cgd /* Check for an empty page. */
356 1.7 cgd if (NEXTINDEX(h) == 0) {
357 1.7 cgd if (__bt_pdelete(t, h))
358 1.7 cgd return (RET_ERROR);
359 1.7 cgd goto loop;
360 1.1 cgd }
361 1.1 cgd
362 1.7 cgd /* Put the page. */
363 1.7 cgd mpool_put(t->bt_mp, h, MPOOL_DIRTY);
364 1.7 cgd
365 1.7 cgd if (redo)
366 1.7 cgd goto loop;
367 1.7 cgd return (RET_SUCCESS);
368 1.7 cgd }
369 1.7 cgd
370 1.7 cgd /*
371 1.7 cgd * __bt_pdelete --
372 1.7 cgd * Delete a single page from the tree.
373 1.7 cgd *
374 1.7 cgd * Parameters:
375 1.7 cgd * t: tree
376 1.7 cgd * h: leaf page
377 1.7 cgd *
378 1.7 cgd * Returns:
379 1.7 cgd * RET_SUCCESS, RET_ERROR.
380 1.7 cgd *
381 1.7 cgd * Side-effects:
382 1.7 cgd * mpool_put's the page
383 1.7 cgd */
384 1.7 cgd static int
385 1.7 cgd __bt_pdelete(t, h)
386 1.7 cgd BTREE *t;
387 1.7 cgd PAGE *h;
388 1.7 cgd {
389 1.7 cgd BINTERNAL *bi;
390 1.7 cgd PAGE *pg;
391 1.7 cgd EPGNO *parent;
392 1.7 cgd indx_t cnt, index, *ip, offset;
393 1.7 cgd u_int32_t nksize;
394 1.7 cgd char *from;
395 1.7 cgd
396 1.1 cgd /*
397 1.7 cgd * Walk the parent page stack -- a LIFO stack of the pages that were
398 1.7 cgd * traversed when we searched for the page where the delete occurred.
399 1.7 cgd * Each stack entry is a page number and a page index offset. The
400 1.7 cgd * offset is for the page traversed on the search. We've just deleted
401 1.7 cgd * a page, so we have to delete the key from the parent page.
402 1.7 cgd *
403 1.7 cgd * If the delete from the parent page makes it empty, this process may
404 1.7 cgd * continue all the way up the tree. We stop if we reach the root page
405 1.7 cgd * (which is never deleted, it's just not worth the effort) or if the
406 1.7 cgd * delete does not empty the page.
407 1.1 cgd */
408 1.7 cgd while ((parent = BT_POP(t)) != NULL) {
409 1.7 cgd /* Get the parent page. */
410 1.7 cgd if ((pg = mpool_get(t->bt_mp, parent->pgno, 0)) == NULL)
411 1.7 cgd return (RET_ERROR);
412 1.7 cgd
413 1.7 cgd index = parent->index;
414 1.7 cgd bi = GETBINTERNAL(pg, index);
415 1.7 cgd
416 1.7 cgd /* Free any overflow pages. */
417 1.7 cgd if (bi->flags & P_BIGKEY &&
418 1.7 cgd __ovfl_delete(t, bi->bytes) == RET_ERROR) {
419 1.7 cgd mpool_put(t->bt_mp, pg, 0);
420 1.7 cgd return (RET_ERROR);
421 1.7 cgd }
422 1.1 cgd
423 1.7 cgd /*
424 1.7 cgd * Free the parent if it has only the one key and it's not the
425 1.7 cgd * root page. If it's the rootpage, turn it back into an empty
426 1.7 cgd * leaf page.
427 1.7 cgd */
428 1.7 cgd if (NEXTINDEX(pg) == 1)
429 1.7 cgd if (pg->pgno == P_ROOT) {
430 1.7 cgd pg->lower = BTDATAOFF;
431 1.7 cgd pg->upper = t->bt_psize;
432 1.7 cgd pg->flags = P_BLEAF;
433 1.1 cgd } else {
434 1.7 cgd if (__bt_relink(t, pg) || __bt_free(t, pg))
435 1.1 cgd return (RET_ERROR);
436 1.7 cgd continue;
437 1.1 cgd }
438 1.7 cgd else {
439 1.7 cgd /* Pack remaining key items at the end of the page. */
440 1.7 cgd nksize = NBINTERNAL(bi->ksize);
441 1.7 cgd from = (char *)pg + pg->upper;
442 1.7 cgd memmove(from + nksize, from, (char *)bi - from);
443 1.7 cgd pg->upper += nksize;
444 1.7 cgd
445 1.7 cgd /* Adjust indices' offsets, shift the indices down. */
446 1.7 cgd offset = pg->linp[index];
447 1.7 cgd for (cnt = index, ip = &pg->linp[0]; cnt--; ++ip)
448 1.7 cgd if (ip[0] < offset)
449 1.7 cgd ip[0] += nksize;
450 1.7 cgd for (cnt = NEXTINDEX(pg) - index; --cnt; ++ip)
451 1.7 cgd ip[0] = ip[1] < offset ? ip[1] + nksize : ip[1];
452 1.7 cgd pg->lower -= sizeof(indx_t);
453 1.1 cgd }
454 1.1 cgd
455 1.7 cgd mpool_put(t->bt_mp, pg, MPOOL_DIRTY);
456 1.7 cgd break;
457 1.1 cgd }
458 1.1 cgd
459 1.7 cgd /* Free the leaf page, as long as it wasn't the root. */
460 1.7 cgd if (h->pgno == P_ROOT) {
461 1.7 cgd mpool_put(t->bt_mp, h, MPOOL_DIRTY);
462 1.7 cgd return (RET_SUCCESS);
463 1.7 cgd }
464 1.7 cgd return (__bt_relink(t, h) || __bt_free(t, h));
465 1.1 cgd }
466 1.1 cgd
467 1.1 cgd /*
468 1.7 cgd * __bt_dleaf --
469 1.7 cgd * Delete a single record from a leaf page.
470 1.1 cgd *
471 1.1 cgd * Parameters:
472 1.1 cgd * t: tree
473 1.7 cgd * key: referenced key
474 1.7 cgd * h: page
475 1.7 cgd * index: index on page to delete
476 1.1 cgd *
477 1.1 cgd * Returns:
478 1.1 cgd * RET_SUCCESS, RET_ERROR.
479 1.1 cgd */
480 1.1 cgd int
481 1.7 cgd __bt_dleaf(t, key, h, index)
482 1.1 cgd BTREE *t;
483 1.7 cgd const DBT *key;
484 1.1 cgd PAGE *h;
485 1.7 cgd u_int index;
486 1.1 cgd {
487 1.7 cgd BLEAF *bl;
488 1.7 cgd indx_t cnt, *ip, offset;
489 1.7 cgd u_int32_t nbytes;
490 1.7 cgd void *to;
491 1.1 cgd char *from;
492 1.1 cgd
493 1.7 cgd /* If this record is referenced by the cursor, delete the cursor. */
494 1.7 cgd if (F_ISSET(&t->bt_cursor, CURS_INIT) &&
495 1.7 cgd !F_ISSET(&t->bt_cursor, CURS_ACQUIRE) &&
496 1.7 cgd t->bt_cursor.pg.pgno == h->pgno && t->bt_cursor.pg.index == index &&
497 1.7 cgd __bt_curdel(t, key, h, index))
498 1.7 cgd return (RET_ERROR);
499 1.7 cgd
500 1.7 cgd /* If the entry uses overflow pages, make them available for reuse. */
501 1.1 cgd to = bl = GETBLEAF(h, index);
502 1.1 cgd if (bl->flags & P_BIGKEY && __ovfl_delete(t, bl->bytes) == RET_ERROR)
503 1.1 cgd return (RET_ERROR);
504 1.1 cgd if (bl->flags & P_BIGDATA &&
505 1.1 cgd __ovfl_delete(t, bl->bytes + bl->ksize) == RET_ERROR)
506 1.1 cgd return (RET_ERROR);
507 1.7 cgd
508 1.7 cgd /* Pack the remaining key/data items at the end of the page. */
509 1.1 cgd nbytes = NBLEAF(bl);
510 1.1 cgd from = (char *)h + h->upper;
511 1.1 cgd memmove(from + nbytes, from, (char *)to - from);
512 1.1 cgd h->upper += nbytes;
513 1.1 cgd
514 1.7 cgd /* Adjust the indices' offsets, shift the indices down. */
515 1.1 cgd offset = h->linp[index];
516 1.1 cgd for (cnt = index, ip = &h->linp[0]; cnt--; ++ip)
517 1.1 cgd if (ip[0] < offset)
518 1.1 cgd ip[0] += nbytes;
519 1.1 cgd for (cnt = NEXTINDEX(h) - index; --cnt; ++ip)
520 1.1 cgd ip[0] = ip[1] < offset ? ip[1] + nbytes : ip[1];
521 1.1 cgd h->lower -= sizeof(indx_t);
522 1.7 cgd
523 1.7 cgd /* If the cursor is on this page, adjust it as necessary. */
524 1.7 cgd if (F_ISSET(&t->bt_cursor, CURS_INIT) &&
525 1.7 cgd !F_ISSET(&t->bt_cursor, CURS_ACQUIRE) &&
526 1.7 cgd t->bt_cursor.pg.pgno == h->pgno && t->bt_cursor.pg.index > index)
527 1.7 cgd --t->bt_cursor.pg.index;
528 1.7 cgd
529 1.1 cgd return (RET_SUCCESS);
530 1.7 cgd }
531 1.7 cgd
532 1.7 cgd /*
533 1.7 cgd * __bt_curdel --
534 1.7 cgd * Delete the cursor.
535 1.7 cgd *
536 1.7 cgd * Parameters:
537 1.7 cgd * t: tree
538 1.7 cgd * key: referenced key (or NULL)
539 1.7 cgd * h: page
540 1.7 cgd * index: index on page to delete
541 1.7 cgd *
542 1.7 cgd * Returns:
543 1.7 cgd * RET_SUCCESS, RET_ERROR.
544 1.7 cgd */
545 1.7 cgd static int
546 1.7 cgd __bt_curdel(t, key, h, index)
547 1.7 cgd BTREE *t;
548 1.7 cgd const DBT *key;
549 1.7 cgd PAGE *h;
550 1.7 cgd u_int index;
551 1.7 cgd {
552 1.7 cgd CURSOR *c;
553 1.7 cgd EPG e;
554 1.7 cgd PAGE *pg;
555 1.7 cgd int curcopy, status;
556 1.7 cgd
557 1.7 cgd /*
558 1.7 cgd * If there are duplicates, move forward or backward to one.
559 1.7 cgd * Otherwise, copy the key into the cursor area.
560 1.7 cgd */
561 1.7 cgd c = &t->bt_cursor;
562 1.7 cgd F_CLR(c, CURS_AFTER | CURS_BEFORE | CURS_ACQUIRE);
563 1.7 cgd
564 1.7 cgd curcopy = 0;
565 1.7 cgd if (!F_ISSET(t, B_NODUPS)) {
566 1.7 cgd /*
567 1.7 cgd * We're going to have to do comparisons. If we weren't
568 1.7 cgd * provided a copy of the key, i.e. the user is deleting
569 1.7 cgd * the current cursor position, get one.
570 1.7 cgd */
571 1.7 cgd if (key == NULL) {
572 1.7 cgd e.page = h;
573 1.7 cgd e.index = index;
574 1.7 cgd if ((status = __bt_ret(t, &e,
575 1.7 cgd &c->key, &c->key, NULL, NULL, 1)) != RET_SUCCESS)
576 1.7 cgd return (status);
577 1.7 cgd curcopy = 1;
578 1.7 cgd key = &c->key;
579 1.7 cgd }
580 1.7 cgd /* Check previous key, if not at the beginning of the page. */
581 1.7 cgd if (index > 0) {
582 1.7 cgd e.page = h;
583 1.7 cgd e.index = index - 1;
584 1.7 cgd if (__bt_cmp(t, key, &e) == 0) {
585 1.7 cgd F_SET(c, CURS_BEFORE);
586 1.7 cgd goto dup2;
587 1.7 cgd }
588 1.7 cgd }
589 1.7 cgd /* Check next key, if not at the end of the page. */
590 1.7 cgd if (index < NEXTINDEX(h) - 1) {
591 1.7 cgd e.page = h;
592 1.7 cgd e.index = index + 1;
593 1.7 cgd if (__bt_cmp(t, key, &e) == 0) {
594 1.7 cgd F_SET(c, CURS_AFTER);
595 1.7 cgd goto dup2;
596 1.7 cgd }
597 1.7 cgd }
598 1.7 cgd /* Check previous key if at the beginning of the page. */
599 1.7 cgd if (index == 0 && h->prevpg != P_INVALID) {
600 1.7 cgd if ((pg = mpool_get(t->bt_mp, h->prevpg, 0)) == NULL)
601 1.7 cgd return (RET_ERROR);
602 1.7 cgd e.page = pg;
603 1.7 cgd e.index = NEXTINDEX(pg) - 1;
604 1.7 cgd if (__bt_cmp(t, key, &e) == 0) {
605 1.7 cgd F_SET(c, CURS_BEFORE);
606 1.7 cgd goto dup1;
607 1.7 cgd }
608 1.7 cgd mpool_put(t->bt_mp, pg, 0);
609 1.7 cgd }
610 1.7 cgd /* Check next key if at the end of the page. */
611 1.7 cgd if (index == NEXTINDEX(h) - 1 && h->nextpg != P_INVALID) {
612 1.7 cgd if ((pg = mpool_get(t->bt_mp, h->nextpg, 0)) == NULL)
613 1.7 cgd return (RET_ERROR);
614 1.7 cgd e.page = pg;
615 1.7 cgd e.index = 0;
616 1.7 cgd if (__bt_cmp(t, key, &e) == 0) {
617 1.7 cgd F_SET(c, CURS_AFTER);
618 1.7 cgd dup1: mpool_put(t->bt_mp, pg, 0);
619 1.7 cgd dup2: c->pg.pgno = e.page->pgno;
620 1.7 cgd c->pg.index = e.index;
621 1.7 cgd return (RET_SUCCESS);
622 1.7 cgd }
623 1.7 cgd mpool_put(t->bt_mp, pg, 0);
624 1.7 cgd }
625 1.7 cgd }
626 1.7 cgd e.page = h;
627 1.7 cgd e.index = index;
628 1.7 cgd if (curcopy || (status =
629 1.7 cgd __bt_ret(t, &e, &c->key, &c->key, NULL, NULL, 1)) == RET_SUCCESS) {
630 1.7 cgd F_SET(c, CURS_ACQUIRE);
631 1.7 cgd return (RET_SUCCESS);
632 1.7 cgd }
633 1.7 cgd return (status);
634 1.7 cgd }
635 1.7 cgd
636 1.7 cgd /*
637 1.7 cgd * __bt_relink --
638 1.7 cgd * Link around a deleted page.
639 1.7 cgd *
640 1.7 cgd * Parameters:
641 1.7 cgd * t: tree
642 1.7 cgd * h: page to be deleted
643 1.7 cgd */
644 1.7 cgd static int
645 1.7 cgd __bt_relink(t, h)
646 1.7 cgd BTREE *t;
647 1.7 cgd PAGE *h;
648 1.7 cgd {
649 1.7 cgd PAGE *pg;
650 1.7 cgd
651 1.7 cgd if (h->nextpg != P_INVALID) {
652 1.7 cgd if ((pg = mpool_get(t->bt_mp, h->nextpg, 0)) == NULL)
653 1.7 cgd return (RET_ERROR);
654 1.7 cgd pg->prevpg = h->prevpg;
655 1.7 cgd mpool_put(t->bt_mp, pg, MPOOL_DIRTY);
656 1.7 cgd }
657 1.7 cgd if (h->prevpg != P_INVALID) {
658 1.7 cgd if ((pg = mpool_get(t->bt_mp, h->prevpg, 0)) == NULL)
659 1.7 cgd return (RET_ERROR);
660 1.7 cgd pg->nextpg = h->nextpg;
661 1.7 cgd mpool_put(t->bt_mp, pg, MPOOL_DIRTY);
662 1.7 cgd }
663 1.7 cgd return (0);
664 1.1 cgd }
665