1 /* $OpenBSD: bt_put.c,v 1.13 2005/08/05 13:02:59 espie 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 #include <sys/types.h> 36 37 #include <errno.h> 38 #include <stdio.h> 39 #include <stdlib.h> 40 #include <string.h> 41 42 #include <db.h> 43 #include "btree.h" 44 45 static EPG *bt_fast(BTREE *, const DBT *, const DBT *, int *); 46 47 /* 48 * __BT_PUT -- Add a btree item to the tree. 49 * 50 * Parameters: 51 * dbp: pointer to access method 52 * key: key 53 * data: data 54 * flag: R_NOOVERWRITE 55 * 56 * Returns: 57 * RET_ERROR, RET_SUCCESS and RET_SPECIAL if the key is already in the 58 * tree and R_NOOVERWRITE specified. 59 */ 60 int 61 __bt_put(const DB *dbp, DBT *key, const DBT *data, u_int flags) 62 { 63 BTREE *t; 64 DBT tkey, tdata; 65 EPG *e; 66 PAGE *h; 67 indx_t idx, nxtindex; 68 pgno_t pg; 69 u_int32_t nbytes, size32; 70 int dflags, exact, status; 71 char *dest, db[NOVFLSIZE], kb[NOVFLSIZE]; 72 73 t = dbp->internal; 74 75 /* Toss any page pinned across calls. */ 76 if (t->bt_pinned != NULL) { 77 mpool_put(t->bt_mp, t->bt_pinned, 0); 78 t->bt_pinned = NULL; 79 } 80 81 /* Check for change to a read-only tree. */ 82 if (F_ISSET(t, B_RDONLY)) { 83 errno = EPERM; 84 return (RET_ERROR); 85 } 86 87 switch (flags) { 88 case 0: 89 case R_NOOVERWRITE: 90 break; 91 case R_CURSOR: 92 /* 93 * If flags is R_CURSOR, put the cursor. Must already 94 * have started a scan and not have already deleted it. 95 */ 96 if (F_ISSET(&t->bt_cursor, CURS_INIT) && 97 !F_ISSET(&t->bt_cursor, 98 CURS_ACQUIRE | CURS_AFTER | CURS_BEFORE)) 99 break; 100 /* FALLTHROUGH */ 101 default: 102 errno = EINVAL; 103 return (RET_ERROR); 104 } 105 106 /* 107 * If the key/data pair won't fit on a page, store it on overflow 108 * pages. Only put the key on the overflow page if the pair are 109 * still too big after moving the data to an overflow page. 110 * 111 * XXX 112 * If the insert fails later on, the overflow pages aren't recovered. 113 */ 114 dflags = 0; 115 if (key->size + data->size > t->bt_ovflsize) { 116 if (key->size > t->bt_ovflsize) { 117 storekey: if (__ovfl_put(t, key, &pg) == RET_ERROR) 118 return (RET_ERROR); 119 tkey.data = kb; 120 tkey.size = NOVFLSIZE; 121 memmove(kb, &pg, sizeof(pgno_t)); 122 size32 = key->size; 123 memmove(kb + sizeof(pgno_t), 124 &size32, sizeof(u_int32_t)); 125 dflags |= P_BIGKEY; 126 key = &tkey; 127 } 128 if (key->size + data->size > t->bt_ovflsize) { 129 if (__ovfl_put(t, data, &pg) == RET_ERROR) 130 return (RET_ERROR); 131 tdata.data = db; 132 tdata.size = NOVFLSIZE; 133 memmove(db, &pg, sizeof(pgno_t)); 134 size32 = data->size; 135 memmove(db + sizeof(pgno_t), 136 &size32, sizeof(u_int32_t)); 137 dflags |= P_BIGDATA; 138 data = &tdata; 139 } 140 if (key->size + data->size > t->bt_ovflsize) 141 goto storekey; 142 } 143 144 /* Replace the cursor. */ 145 if (flags == R_CURSOR) { 146 if ((h = mpool_get(t->bt_mp, t->bt_cursor.pg.pgno, 0)) == NULL) 147 return (RET_ERROR); 148 idx = t->bt_cursor.pg.index; 149 goto delete; 150 } 151 152 /* 153 * Find the key to delete, or, the location at which to insert. 154 * Bt_fast and __bt_search both pin the returned page. 155 */ 156 if (t->bt_order == NOT || (e = bt_fast(t, key, data, &exact)) == NULL) 157 if ((e = __bt_search(t, key, &exact)) == NULL) 158 return (RET_ERROR); 159 h = e->page; 160 idx = e->index; 161 162 /* 163 * Add the key/data pair to the tree. If an identical key is already 164 * in the tree, and R_NOOVERWRITE is set, an error is returned. If 165 * R_NOOVERWRITE is not set, the key is either added (if duplicates are 166 * permitted) or an error is returned. 167 */ 168 switch (flags) { 169 case R_NOOVERWRITE: 170 if (!exact) 171 break; 172 mpool_put(t->bt_mp, h, 0); 173 return (RET_SPECIAL); 174 default: 175 if (!exact || !F_ISSET(t, B_NODUPS)) 176 break; 177 /* 178 * !!! 179 * Note, the delete may empty the page, so we need to put a 180 * new entry into the page immediately. 181 */ 182 delete: if (__bt_dleaf(t, key, h, idx) == RET_ERROR) { 183 mpool_put(t->bt_mp, h, 0); 184 return (RET_ERROR); 185 } 186 break; 187 } 188 189 /* 190 * If not enough room, or the user has put a ceiling on the number of 191 * keys permitted in the page, split the page. The split code will 192 * insert the key and data and unpin the current page. If inserting 193 * into the offset array, shift the pointers up. 194 */ 195 nbytes = NBLEAFDBT(key->size, data->size); 196 if (h->upper - h->lower < nbytes + sizeof(indx_t)) { 197 if ((status = __bt_split(t, h, key, 198 data, dflags, nbytes, idx)) != RET_SUCCESS) 199 return (status); 200 goto success; 201 } 202 203 if (idx < (nxtindex = NEXTINDEX(h))) 204 memmove(h->linp + idx + 1, h->linp + idx, 205 (nxtindex - idx) * sizeof(indx_t)); 206 h->lower += sizeof(indx_t); 207 208 h->linp[idx] = h->upper -= nbytes; 209 dest = (char *)h + h->upper; 210 WR_BLEAF(dest, key, data, dflags); 211 212 /* If the cursor is on this page, adjust it as necessary. */ 213 if (F_ISSET(&t->bt_cursor, CURS_INIT) && 214 !F_ISSET(&t->bt_cursor, CURS_ACQUIRE) && 215 t->bt_cursor.pg.pgno == h->pgno && t->bt_cursor.pg.index >= idx) 216 ++t->bt_cursor.pg.index; 217 218 if (t->bt_order == NOT) { 219 if (h->nextpg == P_INVALID) { 220 if (idx == NEXTINDEX(h) - 1) { 221 t->bt_order = FORWARD; 222 t->bt_last.index = idx; 223 t->bt_last.pgno = h->pgno; 224 } 225 } else if (h->prevpg == P_INVALID) { 226 if (idx == 0) { 227 t->bt_order = BACK; 228 t->bt_last.index = 0; 229 t->bt_last.pgno = h->pgno; 230 } 231 } 232 } 233 234 mpool_put(t->bt_mp, h, MPOOL_DIRTY); 235 236 success: 237 if (flags == R_SETCURSOR) 238 __bt_setcur(t, e->page->pgno, e->index); 239 240 F_SET(t, B_MODIFIED); 241 return (RET_SUCCESS); 242 } 243 244 #ifdef STATISTICS 245 u_long bt_cache_hit, bt_cache_miss; 246 #endif 247 248 /* 249 * BT_FAST -- Do a quick check for sorted data. 250 * 251 * Parameters: 252 * t: tree 253 * key: key to insert 254 * 255 * Returns: 256 * EPG for new record or NULL if not found. 257 */ 258 static EPG * 259 bt_fast(BTREE *t, const DBT *key, const DBT *data, int *exactp) 260 { 261 PAGE *h; 262 u_int32_t nbytes; 263 int cmp; 264 265 if ((h = mpool_get(t->bt_mp, t->bt_last.pgno, 0)) == NULL) { 266 t->bt_order = NOT; 267 return (NULL); 268 } 269 t->bt_cur.page = h; 270 t->bt_cur.index = t->bt_last.index; 271 272 /* 273 * If won't fit in this page or have too many keys in this page, 274 * have to search to get split stack. 275 */ 276 nbytes = NBLEAFDBT(key->size, data->size); 277 if (h->upper - h->lower < nbytes + sizeof(indx_t)) 278 goto miss; 279 280 if (t->bt_order == FORWARD) { 281 if (t->bt_cur.page->nextpg != P_INVALID) 282 goto miss; 283 if (t->bt_cur.index != NEXTINDEX(h) - 1) 284 goto miss; 285 if ((cmp = __bt_cmp(t, key, &t->bt_cur)) < 0) 286 goto miss; 287 t->bt_last.index = cmp ? ++t->bt_cur.index : t->bt_cur.index; 288 } else { 289 if (t->bt_cur.page->prevpg != P_INVALID) 290 goto miss; 291 if (t->bt_cur.index != 0) 292 goto miss; 293 if ((cmp = __bt_cmp(t, key, &t->bt_cur)) > 0) 294 goto miss; 295 t->bt_last.index = 0; 296 } 297 *exactp = cmp == 0; 298 #ifdef STATISTICS 299 ++bt_cache_hit; 300 #endif 301 return (&t->bt_cur); 302 303 miss: 304 #ifdef STATISTICS 305 ++bt_cache_miss; 306 #endif 307 t->bt_order = NOT; 308 mpool_put(t->bt_mp, h, 0); 309 return (NULL); 310 } 311