xref: /netbsd-src/lib/libc/db/btree/bt_delete.c (revision cda4f8f6ee55684e8d311b86c99ea59191e6b74f)
1 /*-
2  * Copyright (c) 1990, 1993
3  *	The Regents of the University of California.  All rights reserved.
4  *
5  * This code is derived from software contributed to Berkeley by
6  * Mike Olson.
7  *
8  * Redistribution and use in source and binary forms, with or without
9  * modification, are permitted provided that the following conditions
10  * are met:
11  * 1. Redistributions of source code must retain the above copyright
12  *    notice, this list of conditions and the following disclaimer.
13  * 2. Redistributions in binary form must reproduce the above copyright
14  *    notice, this list of conditions and the following disclaimer in the
15  *    documentation and/or other materials provided with the distribution.
16  * 3. All advertising materials mentioning features or use of this software
17  *    must display the following acknowledgement:
18  *	This product includes software developed by the University of
19  *	California, Berkeley and its contributors.
20  * 4. Neither the name of the University nor the names of its contributors
21  *    may be used to endorse or promote products derived from this software
22  *    without specific prior written permission.
23  *
24  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
25  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
26  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
27  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
28  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
29  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
30  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
31  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
32  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
33  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
34  * SUCH DAMAGE.
35  */
36 
37 #if defined(LIBC_SCCS) && !defined(lint)
38 static char sccsid[] = "@(#)bt_delete.c	8.1 (Berkeley) 6/4/93";
39 #endif /* LIBC_SCCS and not lint */
40 
41 #include <sys/types.h>
42 
43 #include <errno.h>
44 #include <stdio.h>
45 #include <string.h>
46 
47 #include <db.h>
48 #include "btree.h"
49 
50 static int bt_bdelete __P((BTREE *, const DBT *));
51 
52 /*
53  * __BT_DELETE -- Delete the item(s) referenced by a key.
54  *
55  * Parameters:
56  *	dbp:	pointer to access method
57  *	key:	key to delete
58  *	flags:	R_CURSOR if deleting what the cursor references
59  *
60  * Returns:
61  *	RET_ERROR, RET_SUCCESS and RET_SPECIAL if the key not found.
62  */
63 int
64 __bt_delete(dbp, key, flags)
65 	const DB *dbp;
66 	const DBT *key;
67 	u_int flags;
68 {
69 	BTREE *t;
70 	int status;
71 
72 	t = dbp->internal;
73 	if (ISSET(t, B_RDONLY)) {
74 		errno = EPERM;
75 		return (RET_ERROR);
76 	}
77 	switch(flags) {
78 	case 0:
79 		status = bt_bdelete(t, key);
80 		break;
81 	case R_CURSOR:
82 		/*
83 		 * If flags is R_CURSOR, delete the cursor; must already have
84 		 * started a scan and not have already deleted the record.  For
85 		 * the delete cursor bit to have been set requires that the
86 		 * scan be initialized, so no reason to check.
87 		 */
88 		if (!ISSET(t, B_SEQINIT))
89                         goto einval;
90 		status = ISSET(t, B_DELCRSR) ?
91 		    RET_SPECIAL : __bt_crsrdel(t, &t->bt_bcursor);
92 		break;
93 	default:
94 einval:		errno = EINVAL;
95 		return (RET_ERROR);
96 	}
97 	if (status == RET_SUCCESS)
98 		SET(t, B_MODIFIED);
99 	return (status);
100 }
101 
102 /*
103  * BT_BDELETE -- Delete all key/data pairs matching the specified key.
104  *
105  * Parameters:
106  *	tree:	tree
107  *	key:	key to delete
108  *
109  * Returns:
110  *	RET_ERROR, RET_SUCCESS and RET_SPECIAL if the key not found.
111  */
112 static int
113 bt_bdelete(t, key)
114 	BTREE *t;
115 	const DBT *key;
116 {
117 	EPG *e, save;
118 	PAGE *h;
119 	pgno_t cpgno, pg;
120 	indx_t cindex;
121 	int deleted, dirty1, dirty2, exact;
122 
123 	/* Find any matching record; __bt_search pins the page. */
124 	if ((e = __bt_search(t, key, &exact)) == NULL)
125 		return (RET_ERROR);
126 	if (!exact) {
127 		mpool_put(t->bt_mp, e->page, 0);
128 		return (RET_SPECIAL);
129 	}
130 
131 	/*
132 	 * Delete forward, then delete backward, from the found key.  The
133 	 * ordering is so that the deletions don't mess up the page refs.
134 	 * The first loop deletes the key from the original page, the second
135 	 * unpins the original page.  In the first loop, dirty1 is set if
136 	 * the original page is modified, and dirty2 is set if any subsequent
137 	 * pages are modified.  In the second loop, dirty1 starts off set if
138 	 * the original page has been modified, and is set if any subsequent
139 	 * pages are modified.
140 	 *
141 	 * If find the key referenced by the cursor, don't delete it, just
142 	 * flag it for future deletion.  The cursor page number is P_INVALID
143 	 * unless the sequential scan is initialized, so no reason to check.
144 	 * A special case is when the already deleted cursor record was the
145 	 * only record found.  If so, then the delete opertion fails as no
146 	 * records were deleted.
147 	 *
148 	 * Cycle in place in the current page until the current record doesn't
149 	 * match the key or the page is empty.  If the latter, walk forward,
150 	 * skipping empty pages and repeating until a record doesn't match
151 	 * the key or the end of the tree is reached.
152 	 */
153 	cpgno = t->bt_bcursor.pgno;
154 	cindex = t->bt_bcursor.index;
155 	save = *e;
156 	dirty1 = 0;
157 	for (h = e->page, deleted = 0;;) {
158 		dirty2 = 0;
159 		do {
160 			if (h->pgno == cpgno && e->index == cindex) {
161 				if (!ISSET(t, B_DELCRSR)) {
162 					SET(t, B_DELCRSR);
163 					deleted = 1;
164 				}
165 				++e->index;
166 			} else {
167 				if (__bt_dleaf(t, h, e->index)) {
168 					if (h->pgno != save.page->pgno)
169 						mpool_put(t->bt_mp, h, dirty2);
170 					mpool_put(t->bt_mp, save.page, dirty1);
171 					return (RET_ERROR);
172 				}
173 				if (h->pgno == save.page->pgno)
174 					dirty1 = MPOOL_DIRTY;
175 				else
176 					dirty2 = MPOOL_DIRTY;
177 				deleted = 1;
178 			}
179 		} while (e->index < NEXTINDEX(h) && __bt_cmp(t, key, e) == 0);
180 
181 		/*
182 		 * Quit if didn't find a match, no next page, or first key on
183 		 * the next page doesn't match.  Don't unpin the original page
184 		 * unless an error occurs.
185 		 */
186 		if (e->index < NEXTINDEX(h))
187 			break;
188 		for (;;) {
189 			if ((pg = h->nextpg) == P_INVALID)
190 				goto done1;
191 			if (h->pgno != save.page->pgno)
192 				mpool_put(t->bt_mp, h, dirty2);
193 			if ((h = mpool_get(t->bt_mp, pg, 0)) == NULL) {
194 				mpool_put(t->bt_mp, save.page, dirty1);
195 				return (RET_ERROR);
196 			}
197 			if (NEXTINDEX(h) != 0) {
198 				e->page = h;
199 				e->index = 0;
200 				break;
201 			}
202 		}
203 
204 		if (__bt_cmp(t, key, e) != 0)
205 			break;
206 	}
207 
208 	/*
209 	 * Reach here with the original page and the last page referenced
210 	 * pinned (they may be the same).  Release it if not the original.
211 	 */
212 done1:	if (h->pgno != save.page->pgno)
213 		mpool_put(t->bt_mp, h, dirty2);
214 
215 	/*
216 	 * Walk backwards from the record previous to the record returned by
217 	 * __bt_search, skipping empty pages, until a record doesn't match
218 	 * the key or reach the beginning of the tree.
219 	 */
220 	*e = save;
221 	for (;;) {
222 		if (e->index)
223 			--e->index;
224 		for (h = e->page; e->index; --e->index) {
225 			if (__bt_cmp(t, key, e) != 0)
226 				goto done2;
227 			if (h->pgno == cpgno && e->index == cindex) {
228 				if (!ISSET(t, B_DELCRSR)) {
229 					SET(t, B_DELCRSR);
230 					deleted = 1;
231 				}
232 			} else {
233 				if (__bt_dleaf(t, h, e->index) == RET_ERROR) {
234 					mpool_put(t->bt_mp, h, dirty1);
235 					return (RET_ERROR);
236 				}
237 				if (h->pgno == save.page->pgno)
238 					dirty1 = MPOOL_DIRTY;
239 				deleted = 1;
240 			}
241 		}
242 
243 		if ((pg = h->prevpg) == P_INVALID)
244 			goto done2;
245 		mpool_put(t->bt_mp, h, dirty1);
246 		dirty1 = 0;
247 		if ((e->page = mpool_get(t->bt_mp, pg, 0)) == NULL)
248 			return (RET_ERROR);
249 		e->index = NEXTINDEX(e->page);
250 	}
251 
252 	/*
253 	 * Reach here with the last page that was looked at pinned.  Release
254 	 * it.
255 	 */
256 done2:	mpool_put(t->bt_mp, h, dirty1);
257 	return (deleted ? RET_SUCCESS : RET_SPECIAL);
258 }
259 
260 /*
261  * __BT_DLEAF -- Delete a single record from a leaf page.
262  *
263  * Parameters:
264  *	t:	tree
265  *	index:	index on current page to delete
266  *
267  * Returns:
268  *	RET_SUCCESS, RET_ERROR.
269  */
270 int
271 __bt_dleaf(t, h, index)
272 	BTREE *t;
273 	PAGE *h;
274 	int index;
275 {
276 	register BLEAF *bl;
277 	register indx_t *ip, offset;
278 	register size_t nbytes;
279 	register int cnt;
280 	char *from;
281 	void *to;
282 
283 	/*
284 	 * Delete a record from a btree leaf page.  Internal records are never
285 	 * deleted from internal pages, regardless of the records that caused
286 	 * them to be added being deleted.  Pages made empty by deletion are
287 	 * not reclaimed.  They are, however, made available for reuse.
288 	 *
289 	 * Pack the remaining entries at the end of the page, shift the indices
290 	 * down, overwriting the deleted record and its index.  If the record
291 	 * uses overflow pages, make them available for reuse.
292 	 */
293 	to = bl = GETBLEAF(h, index);
294 	if (bl->flags & P_BIGKEY && __ovfl_delete(t, bl->bytes) == RET_ERROR)
295 		return (RET_ERROR);
296 	if (bl->flags & P_BIGDATA &&
297 	    __ovfl_delete(t, bl->bytes + bl->ksize) == RET_ERROR)
298 		return (RET_ERROR);
299 	nbytes = NBLEAF(bl);
300 
301 	/*
302 	 * Compress the key/data pairs.  Compress and adjust the [BR]LEAF
303 	 * offsets.  Reset the headers.
304 	 */
305 	from = (char *)h + h->upper;
306 	memmove(from + nbytes, from, (char *)to - from);
307 	h->upper += nbytes;
308 
309 	offset = h->linp[index];
310 	for (cnt = index, ip = &h->linp[0]; cnt--; ++ip)
311 		if (ip[0] < offset)
312 			ip[0] += nbytes;
313 	for (cnt = NEXTINDEX(h) - index; --cnt; ++ip)
314 		ip[0] = ip[1] < offset ? ip[1] + nbytes : ip[1];
315 	h->lower -= sizeof(indx_t);
316 	return (RET_SUCCESS);
317 }
318