db.subproj/btree.subproj/bt_put.c

   1 /*
   2  * Copyright (c) 1999 Apple Computer, Inc. All rights reserved.
   3  *
   4  * @APPLE_LICENSE_HEADER_START@
   5  *
   6  * The contents of this file constitute Original Code as defined in and
   7  * are subject to the Apple Public Source License Version 1.1 (the
   8  * "License").  You may not use this file except in compliance with the
   9  * License.  Please obtain a copy of the License at
  10  * http://www.apple.com/publicsource and read it before using this file.
  11  *
  12  * This Original Code and all software distributed under the License are
  13  * distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY KIND, EITHER
  14  * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
  15  * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
  16  * FITNESS FOR A PARTICULAR PURPOSE OR NON-INFRINGEMENT.  Please see the
  17  * License for the specific language governing rights and limitations
  18  * under the License.
  19  *
  20  * @APPLE_LICENSE_HEADER_END@
  21  */
  22 /*
  23  * Copyright (c) 1990, 1993
  24  *      The Regents of the University of California.  All rights reserved.
  25  *
  26  * This code is derived from software contributed to Berkeley by
  27  * Mike Olson.
  28  *
  29  * Redistribution and use in source and binary forms, with or without
  30  * modification, are permitted provided that the following conditions
  31  * are met:
  32  * 1. Redistributions of source code must retain the above copyright
  33  *    notice, this list of conditions and the following disclaimer.
  34  * 2. Redistributions in binary form must reproduce the above copyright
  35  *    notice, this list of conditions and the following disclaimer in the
  36  *    documentation and/or other materials provided with the distribution.
  37  * 3. All advertising materials mentioning features or use of this software
  38  *    must display the following acknowledgement:
  39  *      This product includes software developed by the University of
  40  *      California, Berkeley and its contributors.
  41  * 4. Neither the name of the University nor the names of its contributors
  42  *    may be used to endorse or promote products derived from this software
  43  *    without specific prior written permission.
  44  *
  45  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
  46  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  47  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  48  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
  49  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  50  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  51  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  52  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  53  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  54  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  55  * SUCH DAMAGE.
  56  */
  57
  58
  59 #include <sys/types.h>
  60
  61 #include <errno.h>
  62 #include <stdio.h>
  63 #include <stdlib.h>
  64 #include <string.h>
  65
  66 #include <db.h>
  67 #include "btree.h"
  68
  69 static EPG *bt_fast __P((BTREE *, const DBT *, const DBT *, int *));
  70
  71 /*
  72  * __BT_PUT -- Add a btree item to the tree.
  73  *
  74  * Parameters:
  75  *      dbp:    pointer to access method
  76  *      key:    key
  77  *      data:   data
  78  *      flag:   R_NOOVERWRITE
  79  *
  80  * Returns:
  81  *      RET_ERROR, RET_SUCCESS and RET_SPECIAL if the key is already in the
  82  *      tree and R_NOOVERWRITE specified.
  83  */
  84 int
  85 __bt_put(dbp, key, data, flags)
  86         const DB *dbp;
  87         DBT *key;
  88         const DBT *data;
  89         u_int flags;
  90 {
  91         BTREE *t;
  92         DBT tkey, tdata;
  93         EPG *e;
  94         PAGE *h;
  95         indx_t index, nxtindex;
  96         pgno_t pg;
  97         size_t nbytes;
  98         int dflags, exact, status;
  99         char *dest, db[NOVFLSIZE], kb[NOVFLSIZE];
 100
 101         t = dbp->internal;
 102
 103         /* Toss any page pinned across calls. */
 104         if (t->bt_pinned != NULL) {
 105                 mpool_put(t->bt_mp, t->bt_pinned, 0);
 106                 t->bt_pinned = NULL;
 107         }
 108
 109         switch (flags) {
 110         case R_CURSOR:
 111                 if (!ISSET(t, B_SEQINIT))
 112                         goto einval;
 113                 if (ISSET(t, B_DELCRSR))
 114                         goto einval;
 115                 break;
 116         case 0:
 117         case R_NOOVERWRITE:
 118                 break;
 119         default:
 120 einval:         errno = EINVAL;
 121                 return (RET_ERROR);
 122         }
 123
 124         if (ISSET(t, B_RDONLY)) {
 125                 errno = EPERM;
 126                 return (RET_ERROR);
 127         }
 128
 129         /*
 130          * If the key/data won't fit on a page, store it on indirect pages.
 131          * Only store the key on the overflow page if it's too big after the
 132          * data is on an overflow page.
 133          *
 134          * XXX
 135          * If the insert fails later on, these pages aren't recovered.
 136          */
 137         dflags = 0;
 138         if (key->size + data->size > t->bt_ovflsize) {
 139                 if (key->size > t->bt_ovflsize) {
 140 storekey:               if (__ovfl_put(t, key, &pg) == RET_ERROR)
 141                                 return (RET_ERROR);
 142                         tkey.data = kb;
 143                         tkey.size = NOVFLSIZE;
 144                         memmove(kb, &pg, sizeof(pgno_t));
 145                         memmove(kb + sizeof(pgno_t),
 146                             &key->size, sizeof(size_t));
 147                         dflags |= P_BIGKEY;
 148                         key = &tkey;
 149                 }
 150                 if (key->size + data->size > t->bt_ovflsize) {
 151                         if (__ovfl_put(t, data, &pg) == RET_ERROR)
 152                                 return (RET_ERROR);
 153                         tdata.data = db;
 154                         tdata.size = NOVFLSIZE;
 155                         memmove(db, &pg, sizeof(pgno_t));
 156                         memmove(db + sizeof(pgno_t),
 157                             &data->size, sizeof(size_t));
 158                         dflags |= P_BIGDATA;
 159                         data = &tdata;
 160                 }
 161                 if (key->size + data->size > t->bt_ovflsize)
 162                         goto storekey;
 163         }
 164
 165         /* Replace the cursor. */
 166         if (flags == R_CURSOR) {
 167                 if ((h = mpool_get(t->bt_mp, t->bt_bcursor.pgno, 0)) == NULL)
 168                         return (RET_ERROR);
 169                 index = t->bt_bcursor.index;
 170                 goto delete;
 171         }
 172
 173         /*
 174          * Find the key to delete, or, the location at which to insert.  Bt_fast
 175          * and __bt_search pin the returned page.
 176          */
 177         if (t->bt_order == NOT || (e = bt_fast(t, key, data, &exact)) == NULL)
 178                 if ((e = __bt_search(t, key, &exact)) == NULL)
 179                         return (RET_ERROR);
 180         h = e->page;
 181         index = e->index;
 182
 183         /*
 184          * Add the specified key/data pair to the tree.  If an identical key
 185          * is already in the tree, and R_NOOVERWRITE is set, an error is
 186          * returned.  If R_NOOVERWRITE is not set, the key is either added (if
 187          * duplicates are permitted) or an error is returned.
 188          *
 189          * Pages are split as required.
 190          */
 191         switch (flags) {
 192         case R_NOOVERWRITE:
 193                 if (!exact)
 194                         break;
 195                 /*
 196                  * One special case is if the cursor references the record and
 197                  * it's been flagged for deletion.  Then, we delete the record,
 198                  * leaving the cursor there -- this means that the inserted
 199                  * record will not be seen in a cursor scan.
 200                  */
 201                 if (ISSET(t, B_DELCRSR) && t->bt_bcursor.pgno == h->pgno &&
 202                     t->bt_bcursor.index == index) {
 203                         CLR(t, B_DELCRSR);
 204                         goto delete;
 205                 }
 206                 mpool_put(t->bt_mp, h, 0);
 207                 return (RET_SPECIAL);
 208         default:
 209                 if (!exact || !ISSET(t, B_NODUPS))
 210                         break;
 211 delete:         if (__bt_dleaf(t, h, index) == RET_ERROR) {
 212                         mpool_put(t->bt_mp, h, 0);
 213                         return (RET_ERROR);
 214                 }
 215                 break;
 216         }
 217
 218         /*
 219          * If not enough room, or the user has put a ceiling on the number of
 220          * keys permitted in the page, split the page.  The split code will
 221          * insert the key and data and unpin the current page.  If inserting
 222          * into the offset array, shift the pointers up.
 223          */
 224         nbytes = NBLEAFDBT(key->size, data->size);
 225         if (h->upper - h->lower < nbytes + sizeof(indx_t)) {
 226                 if ((status = __bt_split(t, h, key,
 227                     data, dflags, nbytes, index)) != RET_SUCCESS)
 228                         return (status);
 229                 goto success;
 230         }
 231
 232         if (index < (nxtindex = NEXTINDEX(h)))
 233                 memmove(h->linp + index + 1, h->linp + index,
 234                     (nxtindex - index) * sizeof(indx_t));
 235         h->lower += sizeof(indx_t);
 236
 237         h->linp[index] = h->upper -= nbytes;
 238         dest = (char *)h + h->upper;
 239         WR_BLEAF(dest, key, data, dflags);
 240
 241         if (t->bt_order == NOT)
 242                 if (h->nextpg == P_INVALID) {
 243                         if (index == NEXTINDEX(h) - 1) {
 244                                 t->bt_order = FORWARD;
 245                                 t->bt_last.index = index;
 246                                 t->bt_last.pgno = h->pgno;
 247                         }
 248                 } else if (h->prevpg == P_INVALID) {
 249                         if (index == 0) {
 250                                 t->bt_order = BACK;
 251                                 t->bt_last.index = 0;
 252                                 t->bt_last.pgno = h->pgno;
 253                         }
 254                 }
 255
 256         mpool_put(t->bt_mp, h, MPOOL_DIRTY);
 257
 258 success:
 259         if (flags == R_SETCURSOR) {
 260                 t->bt_bcursor.pgno = e->page->pgno;
 261                 t->bt_bcursor.index = e->index;
 262         }
 263         SET(t, B_MODIFIED);
 264         return (RET_SUCCESS);
 265 }
 266
 267 #ifdef STATISTICS
 268 u_long bt_cache_hit, bt_cache_miss;
 269 #endif
 270
 271 /*
 272  * BT_FAST -- Do a quick check for sorted data.
 273  *
 274  * Parameters:
 275  *      t:      tree
 276  *      key:    key to insert
 277  *
 278  * Returns:
 279  *      EPG for new record or NULL if not found.
 280  */
 281 static EPG *
 282 bt_fast(t, key, data, exactp)
 283         BTREE *t;
 284         const DBT *key, *data;
 285         int *exactp;
 286 {
 287         PAGE *h;
 288         size_t nbytes;
 289         int cmp;
 290
 291         if ((h = mpool_get(t->bt_mp, t->bt_last.pgno, 0)) == NULL) {
 292                 t->bt_order = NOT;
 293                 return (NULL);
 294         }
 295         t->bt_cur.page = h;
 296         t->bt_cur.index = t->bt_last.index;
 297
 298         /*
 299          * If won't fit in this page or have too many keys in this page, have
 300          * to search to get split stack.
 301          */
 302         nbytes = NBLEAFDBT(key->size, data->size);
 303         if (h->upper - h->lower < nbytes + sizeof(indx_t))
 304                 goto miss;
 305
 306         if (t->bt_order == FORWARD) {
 307                 if (t->bt_cur.page->nextpg != P_INVALID)
 308                         goto miss;
 309                 if (t->bt_cur.index != NEXTINDEX(h) - 1)
 310                         goto miss;
 311                 if ((cmp = __bt_cmp(t, key, &t->bt_cur)) < 0)
 312                         goto miss;
 313                 t->bt_last.index = cmp ? ++t->bt_cur.index : t->bt_cur.index;
 314         } else {
 315                 if (t->bt_cur.page->prevpg != P_INVALID)
 316                         goto miss;
 317                 if (t->bt_cur.index != 0)
 318                         goto miss;
 319                 if ((cmp = __bt_cmp(t, key, &t->bt_cur)) > 0)
 320                         goto miss;
 321                 t->bt_last.index = 0;
 322         }
 323         *exactp = cmp == 0;
 324 #ifdef STATISTICS
 325         ++bt_cache_hit;
 326 #endif
 327         return (&t->bt_cur);
 328
 329 miss:
 330 #ifdef STATISTICS
 331         ++bt_cache_miss;
 332 #endif
 333         t->bt_order = NOT;
 334         mpool_put(t->bt_mp, h, 0);
 335         return (NULL);
 336 }