Home | History | Annotate | Line # | Download | only in btree
bt_put.c revision 1.1.1.2
      1 /*-
      2  * Copyright (c) 1990, 1993, 1994
      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_put.c	8.8 (Berkeley) 7/26/94";
     39 #endif /* LIBC_SCCS and not lint */
     40 
     41 #include <sys/types.h>
     42 
     43 #include <errno.h>
     44 #include <stdio.h>
     45 #include <stdlib.h>
     46 #include <string.h>
     47 
     48 #include <db.h>
     49 #include "btree.h"
     50 
     51 static EPG *bt_fast __P((BTREE *, const DBT *, const DBT *, int *));
     52 
     53 /*
     54  * __BT_PUT -- Add a btree item to the tree.
     55  *
     56  * Parameters:
     57  *	dbp:	pointer to access method
     58  *	key:	key
     59  *	data:	data
     60  *	flag:	R_NOOVERWRITE
     61  *
     62  * Returns:
     63  *	RET_ERROR, RET_SUCCESS and RET_SPECIAL if the key is already in the
     64  *	tree and R_NOOVERWRITE specified.
     65  */
     66 int
     67 __bt_put(dbp, key, data, flags)
     68 	const DB *dbp;
     69 	DBT *key;
     70 	const DBT *data;
     71 	u_int flags;
     72 {
     73 	BTREE *t;
     74 	DBT tkey, tdata;
     75 	EPG *e;
     76 	PAGE *h;
     77 	indx_t index, nxtindex;
     78 	pgno_t pg;
     79 	u_int32_t nbytes;
     80 	int dflags, exact, status;
     81 	char *dest, db[NOVFLSIZE], kb[NOVFLSIZE];
     82 
     83 	t = dbp->internal;
     84 
     85 	/* Toss any page pinned across calls. */
     86 	if (t->bt_pinned != NULL) {
     87 		mpool_put(t->bt_mp, t->bt_pinned, 0);
     88 		t->bt_pinned = NULL;
     89 	}
     90 
     91 	/* Check for change to a read-only tree. */
     92 	if (F_ISSET(t, B_RDONLY)) {
     93 		errno = EPERM;
     94 		return (RET_ERROR);
     95 	}
     96 
     97 	switch (flags) {
     98 	case 0:
     99 	case R_NOOVERWRITE:
    100 		break;
    101 	case R_CURSOR:
    102 		/*
    103 		 * If flags is R_CURSOR, put the cursor.  Must already
    104 		 * have started a scan and not have already deleted it.
    105 		 */
    106 		if (F_ISSET(&t->bt_cursor, CURS_INIT) &&
    107 		    !F_ISSET(&t->bt_cursor,
    108 		        CURS_ACQUIRE | CURS_AFTER | CURS_BEFORE))
    109 			break;
    110 		/* FALLTHROUGH */
    111 	default:
    112 		errno = EINVAL;
    113 		return (RET_ERROR);
    114 	}
    115 
    116 	/*
    117 	 * If the key/data pair won't fit on a page, store it on overflow
    118 	 * pages.  Only put the key on the overflow page if the pair are
    119 	 * still too big after moving the data to an overflow page.
    120 	 *
    121 	 * XXX
    122 	 * If the insert fails later on, the overflow pages aren't recovered.
    123 	 */
    124 	dflags = 0;
    125 	if (key->size + data->size > t->bt_ovflsize) {
    126 		if (key->size > t->bt_ovflsize) {
    127 storekey:		if (__ovfl_put(t, key, &pg) == RET_ERROR)
    128 				return (RET_ERROR);
    129 			tkey.data = kb;
    130 			tkey.size = NOVFLSIZE;
    131 			memmove(kb, &pg, sizeof(pgno_t));
    132 			memmove(kb + sizeof(pgno_t),
    133 			    &key->size, sizeof(u_int32_t));
    134 			dflags |= P_BIGKEY;
    135 			key = &tkey;
    136 		}
    137 		if (key->size + data->size > t->bt_ovflsize) {
    138 			if (__ovfl_put(t, data, &pg) == RET_ERROR)
    139 				return (RET_ERROR);
    140 			tdata.data = db;
    141 			tdata.size = NOVFLSIZE;
    142 			memmove(db, &pg, sizeof(pgno_t));
    143 			memmove(db + sizeof(pgno_t),
    144 			    &data->size, sizeof(u_int32_t));
    145 			dflags |= P_BIGDATA;
    146 			data = &tdata;
    147 		}
    148 		if (key->size + data->size > t->bt_ovflsize)
    149 			goto storekey;
    150 	}
    151 
    152 	/* Replace the cursor. */
    153 	if (flags == R_CURSOR) {
    154 		if ((h = mpool_get(t->bt_mp, t->bt_cursor.pg.pgno, 0)) == NULL)
    155 			return (RET_ERROR);
    156 		index = t->bt_cursor.pg.index;
    157 		goto delete;
    158 	}
    159 
    160 	/*
    161 	 * Find the key to delete, or, the location at which to insert.
    162 	 * Bt_fast and __bt_search both pin the returned page.
    163 	 */
    164 	if (t->bt_order == NOT || (e = bt_fast(t, key, data, &exact)) == NULL)
    165 		if ((e = __bt_search(t, key, &exact)) == NULL)
    166 			return (RET_ERROR);
    167 	h = e->page;
    168 	index = e->index;
    169 
    170 	/*
    171 	 * Add the key/data pair to the tree.  If an identical key is already
    172 	 * in the tree, and R_NOOVERWRITE is set, an error is returned.  If
    173 	 * R_NOOVERWRITE is not set, the key is either added (if duplicates are
    174 	 * permitted) or an error is returned.
    175 	 */
    176 	switch (flags) {
    177 	case R_NOOVERWRITE:
    178 		if (!exact)
    179 			break;
    180 		mpool_put(t->bt_mp, h, 0);
    181 		return (RET_SPECIAL);
    182 	default:
    183 		if (!exact || !F_ISSET(t, B_NODUPS))
    184 			break;
    185 		/*
    186 		 * !!!
    187 		 * Note, the delete may empty the page, so we need to put a
    188 		 * new entry into the page immediately.
    189 		 */
    190 delete:		if (__bt_dleaf(t, key, h, index) == RET_ERROR) {
    191 			mpool_put(t->bt_mp, h, 0);
    192 			return (RET_ERROR);
    193 		}
    194 		break;
    195 	}
    196 
    197 	/*
    198 	 * If not enough room, or the user has put a ceiling on the number of
    199 	 * keys permitted in the page, split the page.  The split code will
    200 	 * insert the key and data and unpin the current page.  If inserting
    201 	 * into the offset array, shift the pointers up.
    202 	 */
    203 	nbytes = NBLEAFDBT(key->size, data->size);
    204 	if (h->upper - h->lower < nbytes + sizeof(indx_t)) {
    205 		if ((status = __bt_split(t, h, key,
    206 		    data, dflags, nbytes, index)) != RET_SUCCESS)
    207 			return (status);
    208 		goto success;
    209 	}
    210 
    211 	if (index < (nxtindex = NEXTINDEX(h)))
    212 		memmove(h->linp + index + 1, h->linp + index,
    213 		    (nxtindex - index) * sizeof(indx_t));
    214 	h->lower += sizeof(indx_t);
    215 
    216 	h->linp[index] = h->upper -= nbytes;
    217 	dest = (char *)h + h->upper;
    218 	WR_BLEAF(dest, key, data, dflags);
    219 
    220 	/* If the cursor is on this page, adjust it as necessary. */
    221 	if (F_ISSET(&t->bt_cursor, CURS_INIT) &&
    222 	    !F_ISSET(&t->bt_cursor, CURS_ACQUIRE) &&
    223 	    t->bt_cursor.pg.pgno == h->pgno && t->bt_cursor.pg.index >= index)
    224 		++t->bt_cursor.pg.index;
    225 
    226 	if (t->bt_order == NOT)
    227 		if (h->nextpg == P_INVALID) {
    228 			if (index == NEXTINDEX(h) - 1) {
    229 				t->bt_order = FORWARD;
    230 				t->bt_last.index = index;
    231 				t->bt_last.pgno = h->pgno;
    232 			}
    233 		} else if (h->prevpg == P_INVALID) {
    234 			if (index == 0) {
    235 				t->bt_order = BACK;
    236 				t->bt_last.index = 0;
    237 				t->bt_last.pgno = h->pgno;
    238 			}
    239 		}
    240 
    241 	mpool_put(t->bt_mp, h, MPOOL_DIRTY);
    242 
    243 success:
    244 	if (flags == R_SETCURSOR)
    245 		__bt_setcur(t, e->page->pgno, e->index);
    246 
    247 	F_SET(t, B_MODIFIED);
    248 	return (RET_SUCCESS);
    249 }
    250 
    251 #ifdef STATISTICS
    252 u_long bt_cache_hit, bt_cache_miss;
    253 #endif
    254 
    255 /*
    256  * BT_FAST -- Do a quick check for sorted data.
    257  *
    258  * Parameters:
    259  *	t:	tree
    260  *	key:	key to insert
    261  *
    262  * Returns:
    263  * 	EPG for new record or NULL if not found.
    264  */
    265 static EPG *
    266 bt_fast(t, key, data, exactp)
    267 	BTREE *t;
    268 	const DBT *key, *data;
    269 	int *exactp;
    270 {
    271 	PAGE *h;
    272 	u_int32_t nbytes;
    273 	int cmp;
    274 
    275 	if ((h = mpool_get(t->bt_mp, t->bt_last.pgno, 0)) == NULL) {
    276 		t->bt_order = NOT;
    277 		return (NULL);
    278 	}
    279 	t->bt_cur.page = h;
    280 	t->bt_cur.index = t->bt_last.index;
    281 
    282 	/*
    283 	 * If won't fit in this page or have too many keys in this page,
    284 	 * have to search to get split stack.
    285 	 */
    286 	nbytes = NBLEAFDBT(key->size, data->size);
    287 	if (h->upper - h->lower < nbytes + sizeof(indx_t))
    288 		goto miss;
    289 
    290 	if (t->bt_order == FORWARD) {
    291 		if (t->bt_cur.page->nextpg != P_INVALID)
    292 			goto miss;
    293 		if (t->bt_cur.index != NEXTINDEX(h) - 1)
    294 			goto miss;
    295 		if ((cmp = __bt_cmp(t, key, &t->bt_cur)) < 0)
    296 			goto miss;
    297 		t->bt_last.index = cmp ? ++t->bt_cur.index : t->bt_cur.index;
    298 	} else {
    299 		if (t->bt_cur.page->prevpg != P_INVALID)
    300 			goto miss;
    301 		if (t->bt_cur.index != 0)
    302 			goto miss;
    303 		if ((cmp = __bt_cmp(t, key, &t->bt_cur)) > 0)
    304 			goto miss;
    305 		t->bt_last.index = 0;
    306 	}
    307 	*exactp = cmp == 0;
    308 #ifdef STATISTICS
    309 	++bt_cache_hit;
    310 #endif
    311 	return (&t->bt_cur);
    312 
    313 miss:
    314 #ifdef STATISTICS
    315 	++bt_cache_miss;
    316 #endif
    317 	t->bt_order = NOT;
    318 	mpool_put(t->bt_mp, h, 0);
    319 	return (NULL);
    320 }
    321