1 /* $NetBSD: bt_overflow.c,v 1.16 2008/09/11 12:58:00 joerg Exp $ */ 2 3 /*- 4 * Copyright (c) 1990, 1993, 1994 5 * The Regents of the University of California. All rights reserved. 6 * 7 * This code is derived from software contributed to Berkeley by 8 * Mike Olson. 9 * 10 * Redistribution and use in source and binary forms, with or without 11 * modification, are permitted provided that the following conditions 12 * are met: 13 * 1. Redistributions of source code must retain the above copyright 14 * notice, this list of conditions and the following disclaimer. 15 * 2. Redistributions in binary form must reproduce the above copyright 16 * notice, this list of conditions and the following disclaimer in the 17 * documentation and/or other materials provided with the distribution. 18 * 3. Neither the name of the University nor the names of its contributors 19 * may be used to endorse or promote products derived from this software 20 * without specific prior written permission. 21 * 22 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 23 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 24 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 25 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 26 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 27 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 28 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 29 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 30 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 31 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 32 * SUCH DAMAGE. 33 */ 34 35 #if HAVE_NBTOOL_CONFIG_H 36 #include "nbtool_config.h" 37 #endif 38 39 #include <sys/cdefs.h> 40 #ifndef __minix 41 __RCSID("$NetBSD: bt_overflow.c,v 1.16 2008/09/11 12:58:00 joerg Exp $"); 42 #endif 43 44 #ifndef __minix 45 #include "namespace.h" 46 #endif 47 #include <sys/param.h> 48 49 #include <assert.h> 50 #include <stdio.h> 51 #include <stdlib.h> 52 #include <string.h> 53 54 #include <db.h> 55 #include "btree.h" 56 57 /* 58 * Big key/data code. 59 * 60 * Big key and data entries are stored on linked lists of pages. The initial 61 * reference is byte string stored with the key or data and is the page number 62 * and size. The actual record is stored in a chain of pages linked by the 63 * nextpg field of the PAGE header. 64 * 65 * The first page of the chain has a special property. If the record is used 66 * by an internal page, it cannot be deleted and the P_PRESERVE bit will be set 67 * in the header. 68 * 69 * XXX 70 * A single DBT is written to each chain, so a lot of space on the last page 71 * is wasted. This is a fairly major bug for some data sets. 72 */ 73 74 /* 75 * __OVFL_GET -- Get an overflow key/data item. 76 * 77 * Parameters: 78 * t: tree 79 * p: pointer to { pgno_t, uint32_t } 80 * buf: storage address 81 * bufsz: storage size 82 * 83 * Returns: 84 * RET_ERROR, RET_SUCCESS 85 */ 86 int 87 __ovfl_get(BTREE *t, void *p, size_t *ssz, void **buf, size_t *bufsz) 88 { 89 PAGE *h; 90 pgno_t pg; 91 uint32_t sz, nb, plen; 92 size_t temp; 93 94 memmove(&pg, p, sizeof(pgno_t)); 95 memmove(&sz, (char *)p + sizeof(pgno_t), sizeof(uint32_t)); 96 *ssz = sz; 97 98 #ifdef DEBUG 99 if (pg == P_INVALID || sz == 0) 100 abort(); 101 #endif 102 /* Make the buffer bigger as necessary. */ 103 if (*bufsz < sz) { 104 *buf = (char *)(*buf == NULL ? malloc(sz) : realloc(*buf, sz)); 105 if (*buf == NULL) 106 return (RET_ERROR); 107 *bufsz = sz; 108 } 109 110 /* 111 * Step through the linked list of pages, copying the data on each one 112 * into the buffer. Never copy more than the data's length. 113 */ 114 temp = t->bt_psize - BTDATAOFF; 115 _DBFIT(temp, uint32_t); 116 plen = (uint32_t)temp; 117 for (p = *buf;; p = (char *)p + nb, pg = h->nextpg) { 118 if ((h = mpool_get(t->bt_mp, pg, 0)) == NULL) 119 return (RET_ERROR); 120 121 nb = MIN(sz, plen); 122 memmove(p, (char *)(void *)h + BTDATAOFF, nb); 123 mpool_put(t->bt_mp, h, 0); 124 125 if ((sz -= nb) == 0) 126 break; 127 } 128 return (RET_SUCCESS); 129 } 130 131 /* 132 * __OVFL_PUT -- Store an overflow key/data item. 133 * 134 * Parameters: 135 * t: tree 136 * data: DBT to store 137 * pgno: storage page number 138 * 139 * Returns: 140 * RET_ERROR, RET_SUCCESS 141 */ 142 int 143 __ovfl_put(BTREE *t, const DBT *dbt, pgno_t *pg) 144 { 145 PAGE *h, *last; 146 void *p; 147 pgno_t npg; 148 uint32_t sz, nb, plen; 149 size_t temp; 150 151 /* 152 * Allocate pages and copy the key/data record into them. Store the 153 * number of the first page in the chain. 154 */ 155 temp = t->bt_psize - BTDATAOFF; 156 _DBFIT(temp, uint32_t); 157 plen = (uint32_t)temp; 158 last = NULL; 159 p = dbt->data; 160 temp = dbt->size; 161 _DBFIT(temp, uint32_t); 162 sz = temp; 163 for (;; p = (char *)p + plen, last = h) { 164 if ((h = __bt_new(t, &npg)) == NULL) 165 return (RET_ERROR); 166 167 h->pgno = npg; 168 h->nextpg = h->prevpg = P_INVALID; 169 h->flags = P_OVERFLOW; 170 h->lower = h->upper = 0; 171 172 nb = MIN(sz, plen); 173 (void)memmove((char *)(void *)h + BTDATAOFF, p, (size_t)nb); 174 175 if (last) { 176 last->nextpg = h->pgno; 177 mpool_put(t->bt_mp, last, MPOOL_DIRTY); 178 } else 179 *pg = h->pgno; 180 181 if ((sz -= nb) == 0) { 182 mpool_put(t->bt_mp, h, MPOOL_DIRTY); 183 break; 184 } 185 } 186 return (RET_SUCCESS); 187 } 188 189 /* 190 * __OVFL_DELETE -- Delete an overflow chain. 191 * 192 * Parameters: 193 * t: tree 194 * p: pointer to { pgno_t, uint32_t } 195 * 196 * Returns: 197 * RET_ERROR, RET_SUCCESS 198 */ 199 int 200 __ovfl_delete(BTREE *t, void *p) 201 { 202 PAGE *h; 203 pgno_t pg; 204 uint32_t sz, plen; 205 size_t temp; 206 207 (void)memmove(&pg, p, sizeof(pgno_t)); 208 (void)memmove(&sz, (char *)p + sizeof(pgno_t), sizeof(uint32_t)); 209 210 #ifdef DEBUG 211 if (pg == P_INVALID || sz == 0) 212 abort(); 213 #endif 214 if ((h = mpool_get(t->bt_mp, pg, 0)) == NULL) 215 return (RET_ERROR); 216 217 /* Don't delete chains used by internal pages. */ 218 if (h->flags & P_PRESERVE) { 219 mpool_put(t->bt_mp, h, 0); 220 return (RET_SUCCESS); 221 } 222 223 /* Step through the chain, calling the free routine for each page. */ 224 temp = t->bt_psize - BTDATAOFF; 225 _DBFIT(temp, uint32_t); 226 plen = (uint32_t)temp; 227 for (;; sz -= plen) { 228 pg = h->nextpg; 229 __bt_free(t, h); 230 if (sz <= plen) 231 break; 232 if ((h = mpool_get(t->bt_mp, pg, 0)) == NULL) 233 return (RET_ERROR); 234 } 235 return (RET_SUCCESS); 236 } 237