Home | History | Annotate | Line # | Download | only in hash
hash_page.c revision 1.1
      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  * Margo Seltzer.
      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[] = "@(#)hash_page.c	8.1 (Berkeley) 6/6/93";
     39 #endif /* LIBC_SCCS and not lint */
     40 
     41 /*
     42  * PACKAGE:  hashing
     43  *
     44  * DESCRIPTION:
     45  *	Page manipulation for hashing package.
     46  *
     47  * ROUTINES:
     48  *
     49  * External
     50  *	__get_page
     51  *	__add_ovflpage
     52  * Internal
     53  *	overflow_page
     54  *	open_temp
     55  */
     56 
     57 #include <sys/types.h>
     58 
     59 #include <errno.h>
     60 #include <fcntl.h>
     61 #include <signal.h>
     62 #include <stdio.h>
     63 #include <stdlib.h>
     64 #include <string.h>
     65 #include <unistd.h>
     66 #ifdef DEBUG
     67 #include <assert.h>
     68 #endif
     69 
     70 #include <db.h>
     71 #include "hash.h"
     72 #include "page.h"
     73 #include "extern.h"
     74 
     75 static u_long	*fetch_bitmap __P((HTAB *, int));
     76 static u_long	 first_free __P((u_long));
     77 static int	 open_temp __P((HTAB *));
     78 static u_short	 overflow_page __P((HTAB *));
     79 static void	 putpair __P((char *, const DBT *, const DBT *));
     80 static void	 squeeze_key __P((u_short *, const DBT *, const DBT *));
     81 static int	 ugly_split
     82 		    __P((HTAB *, u_int, BUFHEAD *, BUFHEAD *, int, int));
     83 
     84 #define	PAGE_INIT(P) { \
     85 	((u_short *)(P))[0] = 0; \
     86 	((u_short *)(P))[1] = hashp->BSIZE - 3 * sizeof(u_short); \
     87 	((u_short *)(P))[2] = hashp->BSIZE; \
     88 }
     89 
     90 /*
     91  * This is called AFTER we have verified that there is room on the page for
     92  * the pair (PAIRFITS has returned true) so we go right ahead and start moving
     93  * stuff on.
     94  */
     95 static void
     96 putpair(p, key, val)
     97 	char *p;
     98 	const DBT *key, *val;
     99 {
    100 	register u_short *bp, n, off;
    101 
    102 	bp = (u_short *)p;
    103 
    104 	/* Enter the key first. */
    105 	n = bp[0];
    106 
    107 	off = OFFSET(bp) - key->size;
    108 	memmove(p + off, key->data, key->size);
    109 	bp[++n] = off;
    110 
    111 	/* Now the data. */
    112 	off -= val->size;
    113 	memmove(p + off, val->data, val->size);
    114 	bp[++n] = off;
    115 
    116 	/* Adjust page info. */
    117 	bp[0] = n;
    118 	bp[n + 1] = off - ((n + 3) * sizeof(u_short));
    119 	bp[n + 2] = off;
    120 }
    121 
    122 /*
    123  * Returns:
    124  *	 0 OK
    125  *	-1 error
    126  */
    127 extern int
    128 __delpair(hashp, bufp, ndx)
    129 	HTAB *hashp;
    130 	BUFHEAD *bufp;
    131 	register int ndx;
    132 {
    133 	register u_short *bp, newoff;
    134 	register int n;
    135 	u_short pairlen;
    136 
    137 	bp = (u_short *)bufp->page;
    138 	n = bp[0];
    139 
    140 	if (bp[ndx + 1] < REAL_KEY)
    141 		return (__big_delete(hashp, bufp));
    142 	if (ndx != 1)
    143 		newoff = bp[ndx - 1];
    144 	else
    145 		newoff = hashp->BSIZE;
    146 	pairlen = newoff - bp[ndx + 1];
    147 
    148 	if (ndx != (n - 1)) {
    149 		/* Hard Case -- need to shuffle keys */
    150 		register int i;
    151 		register char *src = bufp->page + (int)OFFSET(bp);
    152 		register char *dst = src + (int)pairlen;
    153 		memmove(dst, src, bp[ndx + 1] - OFFSET(bp));
    154 
    155 		/* Now adjust the pointers */
    156 		for (i = ndx + 2; i <= n; i += 2) {
    157 			if (bp[i + 1] == OVFLPAGE) {
    158 				bp[i - 2] = bp[i];
    159 				bp[i - 1] = bp[i + 1];
    160 			} else {
    161 				bp[i - 2] = bp[i] + pairlen;
    162 				bp[i - 1] = bp[i + 1] + pairlen;
    163 			}
    164 		}
    165 	}
    166 	/* Finally adjust the page data */
    167 	bp[n] = OFFSET(bp) + pairlen;
    168 	bp[n - 1] = bp[n + 1] + pairlen + 2 * sizeof(u_short);
    169 	bp[0] = n - 2;
    170 	hashp->NKEYS--;
    171 
    172 	bufp->flags |= BUF_MOD;
    173 	return (0);
    174 }
    175 /*
    176  * Returns:
    177  *	 0 ==> OK
    178  *	-1 ==> Error
    179  */
    180 extern int
    181 __split_page(hashp, obucket, nbucket)
    182 	HTAB *hashp;
    183 	u_int obucket, nbucket;
    184 {
    185 	register BUFHEAD *new_bufp, *old_bufp;
    186 	register u_short *ino;
    187 	register char *np;
    188 	DBT key, val;
    189 	int n, ndx, retval;
    190 	u_short copyto, diff, off, moved;
    191 	char *op;
    192 
    193 	copyto = (u_short)hashp->BSIZE;
    194 	off = (u_short)hashp->BSIZE;
    195 	old_bufp = __get_buf(hashp, obucket, NULL, 0);
    196 	if (old_bufp == NULL)
    197 		return (-1);
    198 	new_bufp = __get_buf(hashp, nbucket, NULL, 0);
    199 	if (new_bufp == NULL)
    200 		return (-1);
    201 
    202 	old_bufp->flags |= (BUF_MOD | BUF_PIN);
    203 	new_bufp->flags |= (BUF_MOD | BUF_PIN);
    204 
    205 	ino = (u_short *)(op = old_bufp->page);
    206 	np = new_bufp->page;
    207 
    208 	moved = 0;
    209 
    210 	for (n = 1, ndx = 1; n < ino[0]; n += 2) {
    211 		if (ino[n + 1] < REAL_KEY) {
    212 			retval = ugly_split(hashp, obucket, old_bufp, new_bufp,
    213 			    (int)copyto, (int)moved);
    214 			old_bufp->flags &= ~BUF_PIN;
    215 			new_bufp->flags &= ~BUF_PIN;
    216 			return (retval);
    217 
    218 		}
    219 		key.data = (u_char *)op + ino[n];
    220 		key.size = off - ino[n];
    221 
    222 		if (__call_hash(hashp, key.data, key.size) == obucket) {
    223 			/* Don't switch page */
    224 			diff = copyto - off;
    225 			if (diff) {
    226 				copyto = ino[n + 1] + diff;
    227 				memmove(op + copyto, op + ino[n + 1],
    228 				    off - ino[n + 1]);
    229 				ino[ndx] = copyto + ino[n] - ino[n + 1];
    230 				ino[ndx + 1] = copyto;
    231 			} else
    232 				copyto = ino[n + 1];
    233 			ndx += 2;
    234 		} else {
    235 			/* Switch page */
    236 			val.data = (u_char *)op + ino[n + 1];
    237 			val.size = ino[n] - ino[n + 1];
    238 			putpair(np, &key, &val);
    239 			moved += 2;
    240 		}
    241 
    242 		off = ino[n + 1];
    243 	}
    244 
    245 	/* Now clean up the page */
    246 	ino[0] -= moved;
    247 	FREESPACE(ino) = copyto - sizeof(u_short) * (ino[0] + 3);
    248 	OFFSET(ino) = copyto;
    249 
    250 #ifdef DEBUG3
    251 	(void)fprintf(stderr, "split %d/%d\n",
    252 	    ((u_short *)np)[0] / 2,
    253 	    ((u_short *)op)[0] / 2);
    254 #endif
    255 	/* unpin both pages */
    256 	old_bufp->flags &= ~BUF_PIN;
    257 	new_bufp->flags &= ~BUF_PIN;
    258 	return (0);
    259 }
    260 
    261 /*
    262  * Called when we encounter an overflow or big key/data page during split
    263  * handling.  This is special cased since we have to begin checking whether
    264  * the key/data pairs fit on their respective pages and because we may need
    265  * overflow pages for both the old and new pages.
    266  *
    267  * The first page might be a page with regular key/data pairs in which case
    268  * we have a regular overflow condition and just need to go on to the next
    269  * page or it might be a big key/data pair in which case we need to fix the
    270  * big key/data pair.
    271  *
    272  * Returns:
    273  *	 0 ==> success
    274  *	-1 ==> failure
    275  */
    276 static int
    277 ugly_split(hashp, obucket, old_bufp, new_bufp, copyto, moved)
    278 	HTAB *hashp;
    279 	u_int obucket;	/* Same as __split_page. */
    280 	BUFHEAD *old_bufp, *new_bufp;
    281 	int copyto;	/* First byte on page which contains key/data values. */
    282 	int moved;	/* Number of pairs moved to new page. */
    283 {
    284 	register BUFHEAD *bufp;	/* Buffer header for ino */
    285 	register u_short *ino;	/* Page keys come off of */
    286 	register u_short *np;	/* New page */
    287 	register u_short *op;	/* Page keys go on to if they aren't moving */
    288 
    289 	BUFHEAD *last_bfp;	/* Last buf header OVFL needing to be freed */
    290 	DBT key, val;
    291 	SPLIT_RETURN ret;
    292 	u_short n, off, ov_addr, scopyto;
    293 	char *cino;		/* Character value of ino */
    294 
    295 	bufp = old_bufp;
    296 	ino = (u_short *)old_bufp->page;
    297 	np = (u_short *)new_bufp->page;
    298 	op = (u_short *)old_bufp->page;
    299 	last_bfp = NULL;
    300 	scopyto = (u_short)copyto;	/* ANSI */
    301 
    302 	n = ino[0] - 1;
    303 	while (n < ino[0]) {
    304 		if (ino[2] < REAL_KEY && ino[2] != OVFLPAGE) {
    305 			/*
    306 			 * Ov_addr gets set before reaching this point; there's
    307 			 * always an overflow page before a big key/data page.
    308 			 */
    309 			if (__big_split(hashp, old_bufp,
    310 			    new_bufp, bufp, ov_addr, obucket, &ret))
    311 				return (-1);
    312 			old_bufp = ret.oldp;
    313 			if (!old_bufp)
    314 				return (-1);
    315 			op = (u_short *)old_bufp->page;
    316 			new_bufp = ret.newp;
    317 			if (!new_bufp)
    318 				return (-1);
    319 			np = (u_short *)new_bufp->page;
    320 			bufp = ret.nextp;
    321 			if (!bufp)
    322 				return (0);
    323 			cino = (char *)bufp->page;
    324 			ino = (u_short *)cino;
    325 			last_bfp = ret.nextp;
    326 		} else if (ino[n + 1] == OVFLPAGE) {
    327 			ov_addr = ino[n];
    328 			/*
    329 			 * Fix up the old page -- the extra 2 are the fields
    330 			 * which contained the overflow information.
    331 			 */
    332 			ino[0] -= (moved + 2);
    333 			FREESPACE(ino) =
    334 			    scopyto - sizeof(u_short) * (ino[0] + 3);
    335 			OFFSET(ino) = scopyto;
    336 
    337 			bufp = __get_buf(hashp, ov_addr, bufp, 0);
    338 			if (!bufp)
    339 				return (-1);
    340 
    341 			ino = (u_short *)bufp->page;
    342 			n = 1;
    343 			scopyto = hashp->BSIZE;
    344 			moved = 0;
    345 
    346 			if (last_bfp)
    347 				__free_ovflpage(hashp, last_bfp);
    348 			last_bfp = bufp;
    349 		}
    350 		/* Move regular sized pairs of there are any */
    351 		off = hashp->BSIZE;
    352 		for (n = 1; (n < ino[0]) && (ino[n + 1] >= REAL_KEY); n += 2) {
    353 			cino = (char *)ino;
    354 			key.data = (u_char *)cino + ino[n];
    355 			key.size = off - ino[n];
    356 			val.data = (u_char *)cino + ino[n + 1];
    357 			val.size = ino[n] - ino[n + 1];
    358 			off = ino[n + 1];
    359 
    360 			if (__call_hash(hashp, key.data, key.size) == obucket) {
    361 				/* Keep on old page */
    362 				if (PAIRFITS(op, (&key), (&val)))
    363 					putpair((char *)op, &key, &val);
    364 				else {
    365 					old_bufp =
    366 					    __add_ovflpage(hashp, old_bufp);
    367 					if (!old_bufp)
    368 						return (-1);
    369 					op = (u_short *)old_bufp->page;
    370 					putpair((char *)op, &key, &val);
    371 				}
    372 				old_bufp->flags |= BUF_MOD;
    373 			} else {
    374 				/* Move to new page */
    375 				if (PAIRFITS(np, (&key), (&val)))
    376 					putpair((char *)np, &key, &val);
    377 				else {
    378 					new_bufp =
    379 					    __add_ovflpage(hashp, new_bufp);
    380 					if (!new_bufp)
    381 						return (-1);
    382 					np = (u_short *)new_bufp->page;
    383 					putpair((char *)np, &key, &val);
    384 				}
    385 				new_bufp->flags |= BUF_MOD;
    386 			}
    387 		}
    388 	}
    389 	if (last_bfp)
    390 		__free_ovflpage(hashp, last_bfp);
    391 	return (0);
    392 }
    393 
    394 /*
    395  * Add the given pair to the page
    396  *
    397  * Returns:
    398  *	0 ==> OK
    399  *	1 ==> failure
    400  */
    401 extern int
    402 __addel(hashp, bufp, key, val)
    403 	HTAB *hashp;
    404 	BUFHEAD *bufp;
    405 	const DBT *key, *val;
    406 {
    407 	register u_short *bp, *sop;
    408 	int do_expand;
    409 
    410 	bp = (u_short *)bufp->page;
    411 	do_expand = 0;
    412 	while (bp[0] && (bp[bp[0]] < REAL_KEY))
    413 		/* Exception case */
    414 		if (bp[2] < REAL_KEY && bp[bp[0]] != OVFLPAGE) {
    415 			/* This is a big-keydata pair */
    416 			bufp = __add_ovflpage(hashp, bufp);
    417 			if (!bufp)
    418 				return (-1);
    419 			bp = (u_short *)bufp->page;
    420 		} else
    421 			/* Try to squeeze key on this page */
    422 			if (FREESPACE(bp) > PAIRSIZE(key, val)) {
    423 				squeeze_key(bp, key, val);
    424 				return (0);
    425 			} else {
    426 				bufp = __get_buf(hashp, bp[bp[0] - 1], bufp, 0);
    427 				if (!bufp)
    428 					return (-1);
    429 				bp = (u_short *)bufp->page;
    430 			}
    431 
    432 	if (PAIRFITS(bp, key, val))
    433 		putpair(bufp->page, key, val);
    434 	else {
    435 		do_expand = 1;
    436 		bufp = __add_ovflpage(hashp, bufp);
    437 		if (!bufp)
    438 			return (-1);
    439 		sop = (u_short *)bufp->page;
    440 
    441 		if (PAIRFITS(sop, key, val))
    442 			putpair((char *)sop, key, val);
    443 		else
    444 			if (__big_insert(hashp, bufp, key, val))
    445 				return (-1);
    446 	}
    447 	bufp->flags |= BUF_MOD;
    448 	/*
    449 	 * If the average number of keys per bucket exceeds the fill factor,
    450 	 * expand the table.
    451 	 */
    452 	hashp->NKEYS++;
    453 	if (do_expand ||
    454 	    (hashp->NKEYS / (hashp->MAX_BUCKET + 1) > hashp->FFACTOR))
    455 		return (__expand_table(hashp));
    456 	return (0);
    457 }
    458 
    459 /*
    460  *
    461  * Returns:
    462  *	pointer on success
    463  *	NULL on error
    464  */
    465 extern BUFHEAD *
    466 __add_ovflpage(hashp, bufp)
    467 	HTAB *hashp;
    468 	BUFHEAD *bufp;
    469 {
    470 	register u_short *sp;
    471 	u_short ndx, ovfl_num;
    472 #ifdef DEBUG1
    473 	int tmp1, tmp2;
    474 #endif
    475 	sp = (u_short *)bufp->page;
    476 
    477 	/* Check if we are dynamically determining the fill factor */
    478 	if (hashp->FFACTOR == DEF_FFACTOR) {
    479 		hashp->FFACTOR = sp[0] >> 1;
    480 		if (hashp->FFACTOR < MIN_FFACTOR)
    481 			hashp->FFACTOR = MIN_FFACTOR;
    482 	}
    483 	bufp->flags |= BUF_MOD;
    484 	ovfl_num = overflow_page(hashp);
    485 #ifdef DEBUG1
    486 	tmp1 = bufp->addr;
    487 	tmp2 = bufp->ovfl ? bufp->ovfl->addr : 0;
    488 #endif
    489 	if (!ovfl_num || !(bufp->ovfl = __get_buf(hashp, ovfl_num, bufp, 1)))
    490 		return (NULL);
    491 	bufp->ovfl->flags |= BUF_MOD;
    492 #ifdef DEBUG1
    493 	(void)fprintf(stderr, "ADDOVFLPAGE: %d->ovfl was %d is now %d\n",
    494 	    tmp1, tmp2, bufp->ovfl->addr);
    495 #endif
    496 	ndx = sp[0];
    497 	/*
    498 	 * Since a pair is allocated on a page only if there's room to add
    499 	 * an overflow page, we know that the OVFL information will fit on
    500 	 * the page.
    501 	 */
    502 	sp[ndx + 4] = OFFSET(sp);
    503 	sp[ndx + 3] = FREESPACE(sp) - OVFLSIZE;
    504 	sp[ndx + 1] = ovfl_num;
    505 	sp[ndx + 2] = OVFLPAGE;
    506 	sp[0] = ndx + 2;
    507 #ifdef HASH_STATISTICS
    508 	hash_overflows++;
    509 #endif
    510 	return (bufp->ovfl);
    511 }
    512 
    513 /*
    514  * Returns:
    515  *	 0 indicates SUCCESS
    516  *	-1 indicates FAILURE
    517  */
    518 extern int
    519 __get_page(hashp, p, bucket, is_bucket, is_disk, is_bitmap)
    520 	HTAB *hashp;
    521 	char *p;
    522 	u_int bucket;
    523 	int is_bucket, is_disk, is_bitmap;
    524 {
    525 	register int fd, page, size;
    526 	int rsize;
    527 	u_short *bp;
    528 
    529 	fd = hashp->fp;
    530 	size = hashp->BSIZE;
    531 
    532 	if ((fd == -1) || !is_disk) {
    533 		PAGE_INIT(p);
    534 		return (0);
    535 	}
    536 	if (is_bucket)
    537 		page = BUCKET_TO_PAGE(bucket);
    538 	else
    539 		page = OADDR_TO_PAGE(bucket);
    540 	if ((lseek(fd, (off_t)page << hashp->BSHIFT, SEEK_SET) == -1) ||
    541 	    ((rsize = read(fd, p, size)) == -1))
    542 		return (-1);
    543 	bp = (u_short *)p;
    544 	if (!rsize)
    545 		bp[0] = 0;	/* We hit the EOF, so initialize a new page */
    546 	else
    547 		if (rsize != size) {
    548 			errno = EFTYPE;
    549 			return (-1);
    550 		}
    551 	if (!is_bitmap && !bp[0]) {
    552 		PAGE_INIT(p);
    553 	} else
    554 		if (hashp->LORDER != BYTE_ORDER) {
    555 			register int i, max;
    556 
    557 			if (is_bitmap) {
    558 				max = hashp->BSIZE >> 2; /* divide by 4 */
    559 				for (i = 0; i < max; i++)
    560 					BLSWAP(((long *)p)[i]);
    561 			} else {
    562 				BSSWAP(bp[0]);
    563 				max = bp[0] + 2;
    564 				for (i = 1; i <= max; i++)
    565 					BSSWAP(bp[i]);
    566 			}
    567 		}
    568 	return (0);
    569 }
    570 
    571 /*
    572  * Write page p to disk
    573  *
    574  * Returns:
    575  *	 0 ==> OK
    576  *	-1 ==>failure
    577  */
    578 extern int
    579 __put_page(hashp, p, bucket, is_bucket, is_bitmap)
    580 	HTAB *hashp;
    581 	char *p;
    582 	u_int bucket;
    583 	int is_bucket, is_bitmap;
    584 {
    585 	register int fd, page, size;
    586 	int wsize;
    587 
    588 	size = hashp->BSIZE;
    589 	if ((hashp->fp == -1) && open_temp(hashp))
    590 		return (-1);
    591 	fd = hashp->fp;
    592 
    593 	if (hashp->LORDER != BYTE_ORDER) {
    594 		register int i;
    595 		register int max;
    596 
    597 		if (is_bitmap) {
    598 			max = hashp->BSIZE >> 2;	/* divide by 4 */
    599 			for (i = 0; i < max; i++)
    600 				BLSWAP(((long *)p)[i]);
    601 		} else {
    602 			max = ((u_short *)p)[0] + 2;
    603 			for (i = 0; i <= max; i++)
    604 				BSSWAP(((u_short *)p)[i]);
    605 		}
    606 	}
    607 	if (is_bucket)
    608 		page = BUCKET_TO_PAGE(bucket);
    609 	else
    610 		page = OADDR_TO_PAGE(bucket);
    611 	if ((lseek(fd, (off_t)page << hashp->BSHIFT, SEEK_SET) == -1) ||
    612 	    ((wsize = write(fd, p, size)) == -1))
    613 		/* Errno is set */
    614 		return (-1);
    615 	if (wsize != size) {
    616 		errno = EFTYPE;
    617 		return (-1);
    618 	}
    619 	return (0);
    620 }
    621 
    622 #define BYTE_MASK	((1 << INT_BYTE_SHIFT) -1)
    623 /*
    624  * Initialize a new bitmap page.  Bitmap pages are left in memory
    625  * once they are read in.
    626  */
    627 extern int
    628 __init_bitmap(hashp, pnum, nbits, ndx)
    629 	HTAB *hashp;
    630 	int pnum, nbits, ndx;
    631 {
    632 	u_long *ip;
    633 	int clearbytes, clearints;
    634 
    635 	if (!(ip = malloc(hashp->BSIZE)))
    636 		return (1);
    637 	hashp->nmaps++;
    638 	clearints = ((nbits - 1) >> INT_BYTE_SHIFT) + 1;
    639 	clearbytes = clearints << INT_TO_BYTE;
    640 	(void)memset((char *)ip, 0, clearbytes);
    641 	(void)memset(((char *)ip) + clearbytes, 0xFF,
    642 	    hashp->BSIZE - clearbytes);
    643 	ip[clearints - 1] = ALL_SET << (nbits & BYTE_MASK);
    644 	SETBIT(ip, 0);
    645 	hashp->BITMAPS[ndx] = (u_short)pnum;
    646 	hashp->mapp[ndx] = ip;
    647 	return (0);
    648 }
    649 
    650 static u_long
    651 first_free(map)
    652 	u_long map;
    653 {
    654 	register u_long i, mask;
    655 
    656 	mask = 0x1;
    657 	for (i = 0; i < BITS_PER_MAP; i++) {
    658 		if (!(mask & map))
    659 			return (i);
    660 		mask = mask << 1;
    661 	}
    662 	return (i);
    663 }
    664 
    665 static u_short
    666 overflow_page(hashp)
    667 	HTAB *hashp;
    668 {
    669 	register u_long *freep;
    670 	register int max_free, offset, splitnum;
    671 	u_short addr;
    672 	int bit, first_page, free_bit, free_page, i, in_use_bits, j;
    673 #ifdef DEBUG2
    674 	int tmp1, tmp2;
    675 #endif
    676 	splitnum = hashp->OVFL_POINT;
    677 	max_free = hashp->SPARES[splitnum];
    678 
    679 	free_page = (max_free - 1) >> (hashp->BSHIFT + BYTE_SHIFT);
    680 	free_bit = (max_free - 1) & ((hashp->BSIZE << BYTE_SHIFT) - 1);
    681 
    682 	/* Look through all the free maps to find the first free block */
    683 	first_page = hashp->LAST_FREED >>(hashp->BSHIFT + BYTE_SHIFT);
    684 	for ( i = first_page; i <= free_page; i++ ) {
    685 		if (!(freep = (u_long *)hashp->mapp[i]) &&
    686 		    !(freep = fetch_bitmap(hashp, i)))
    687 			return (NULL);
    688 		if (i == free_page)
    689 			in_use_bits = free_bit;
    690 		else
    691 			in_use_bits = (hashp->BSIZE << BYTE_SHIFT) - 1;
    692 
    693 		if (i == first_page) {
    694 			bit = hashp->LAST_FREED &
    695 			    ((hashp->BSIZE << BYTE_SHIFT) - 1);
    696 			j = bit / BITS_PER_MAP;
    697 			bit = bit & ~(BITS_PER_MAP - 1);
    698 		} else {
    699 			bit = 0;
    700 			j = 0;
    701 		}
    702 		for (; bit <= in_use_bits; j++, bit += BITS_PER_MAP)
    703 			if (freep[j] != ALL_SET)
    704 				goto found;
    705 	}
    706 
    707 	/* No Free Page Found */
    708 	hashp->LAST_FREED = hashp->SPARES[splitnum];
    709 	hashp->SPARES[splitnum]++;
    710 	offset = hashp->SPARES[splitnum] -
    711 	    (splitnum ? hashp->SPARES[splitnum - 1] : 0);
    712 
    713 #define	OVMSG	"HASH: Out of overflow pages.  Increase page size\n"
    714 	if (offset > SPLITMASK) {
    715 		if (++splitnum >= NCACHED) {
    716 			(void)write(STDERR_FILENO, OVMSG, sizeof(OVMSG) - 1);
    717 			return (NULL);
    718 		}
    719 		hashp->OVFL_POINT = splitnum;
    720 		hashp->SPARES[splitnum] = hashp->SPARES[splitnum-1];
    721 		hashp->SPARES[splitnum-1]--;
    722 		offset = 1;
    723 	}
    724 
    725 	/* Check if we need to allocate a new bitmap page */
    726 	if (free_bit == (hashp->BSIZE << BYTE_SHIFT) - 1) {
    727 		free_page++;
    728 		if (free_page >= NCACHED) {
    729 			(void)write(STDERR_FILENO, OVMSG, sizeof(OVMSG) - 1);
    730 			return (NULL);
    731 		}
    732 		/*
    733 		 * This is tricky.  The 1 indicates that you want the new page
    734 		 * allocated with 1 clear bit.  Actually, you are going to
    735 		 * allocate 2 pages from this map.  The first is going to be
    736 		 * the map page, the second is the overflow page we were
    737 		 * looking for.  The init_bitmap routine automatically, sets
    738 		 * the first bit of itself to indicate that the bitmap itself
    739 		 * is in use.  We would explicitly set the second bit, but
    740 		 * don't have to if we tell init_bitmap not to leave it clear
    741 		 * in the first place.
    742 		 */
    743 		if (__init_bitmap(hashp, (int)OADDR_OF(splitnum, offset),
    744 		    1, free_page))
    745 			return (NULL);
    746 		hashp->SPARES[splitnum]++;
    747 #ifdef DEBUG2
    748 		free_bit = 2;
    749 #endif
    750 		offset++;
    751 		if (offset > SPLITMASK) {
    752 			if (++splitnum >= NCACHED) {
    753 				(void)write(STDERR_FILENO, OVMSG,
    754 				    sizeof(OVMSG) - 1);
    755 				return (NULL);
    756 			}
    757 			hashp->OVFL_POINT = splitnum;
    758 			hashp->SPARES[splitnum] = hashp->SPARES[splitnum-1];
    759 			hashp->SPARES[splitnum-1]--;
    760 			offset = 0;
    761 		}
    762 	} else {
    763 		/*
    764 		 * Free_bit addresses the last used bit.  Bump it to address
    765 		 * the first available bit.
    766 		 */
    767 		free_bit++;
    768 		SETBIT(freep, free_bit);
    769 	}
    770 
    771 	/* Calculate address of the new overflow page */
    772 	addr = OADDR_OF(splitnum, offset);
    773 #ifdef DEBUG2
    774 	(void)fprintf(stderr, "OVERFLOW_PAGE: ADDR: %d BIT: %d PAGE %d\n",
    775 	    addr, free_bit, free_page);
    776 #endif
    777 	return (addr);
    778 
    779 found:
    780 	bit = bit + first_free(freep[j]);
    781 	SETBIT(freep, bit);
    782 #ifdef DEBUG2
    783 	tmp1 = bit;
    784 	tmp2 = i;
    785 #endif
    786 	/*
    787 	 * Bits are addressed starting with 0, but overflow pages are addressed
    788 	 * beginning at 1. Bit is a bit addressnumber, so we need to increment
    789 	 * it to convert it to a page number.
    790 	 */
    791 	bit = 1 + bit + (i * (hashp->BSIZE << BYTE_SHIFT));
    792 	if (bit >= hashp->LAST_FREED)
    793 		hashp->LAST_FREED = bit - 1;
    794 
    795 	/* Calculate the split number for this page */
    796 	for (i = 0; (i < splitnum) && (bit > hashp->SPARES[i]); i++);
    797 	offset = (i ? bit - hashp->SPARES[i - 1] : bit);
    798 	if (offset >= SPLITMASK)
    799 		return (NULL);	/* Out of overflow pages */
    800 	addr = OADDR_OF(i, offset);
    801 #ifdef DEBUG2
    802 	(void)fprintf(stderr, "OVERFLOW_PAGE: ADDR: %d BIT: %d PAGE %d\n",
    803 	    addr, tmp1, tmp2);
    804 #endif
    805 
    806 	/* Allocate and return the overflow page */
    807 	return (addr);
    808 }
    809 
    810 /*
    811  * Mark this overflow page as free.
    812  */
    813 extern void
    814 __free_ovflpage(hashp, obufp)
    815 	HTAB *hashp;
    816 	BUFHEAD *obufp;
    817 {
    818 	register u_short addr;
    819 	u_long *freep;
    820 	int bit_address, free_page, free_bit;
    821 	u_short ndx;
    822 
    823 	addr = obufp->addr;
    824 #ifdef DEBUG1
    825 	(void)fprintf(stderr, "Freeing %d\n", addr);
    826 #endif
    827 	ndx = (((u_short)addr) >> SPLITSHIFT);
    828 	bit_address =
    829 	    (ndx ? hashp->SPARES[ndx - 1] : 0) + (addr & SPLITMASK) - 1;
    830 	 if (bit_address < hashp->LAST_FREED)
    831 		hashp->LAST_FREED = bit_address;
    832 	free_page = (bit_address >> (hashp->BSHIFT + BYTE_SHIFT));
    833 	free_bit = bit_address & ((hashp->BSIZE << BYTE_SHIFT) - 1);
    834 
    835 	if (!(freep = hashp->mapp[free_page]))
    836 		freep = fetch_bitmap(hashp, free_page);
    837 #ifdef DEBUG
    838 	/*
    839 	 * This had better never happen.  It means we tried to read a bitmap
    840 	 * that has already had overflow pages allocated off it, and we
    841 	 * failed to read it from the file.
    842 	 */
    843 	if (!freep)
    844 		assert(0);
    845 #endif
    846 	CLRBIT(freep, free_bit);
    847 #ifdef DEBUG2
    848 	(void)fprintf(stderr, "FREE_OVFLPAGE: ADDR: %d BIT: %d PAGE %d\n",
    849 	    obufp->addr, free_bit, free_page);
    850 #endif
    851 	__reclaim_buf(hashp, obufp);
    852 }
    853 
    854 /*
    855  * Returns:
    856  *	 0 success
    857  *	-1 failure
    858  */
    859 static int
    860 open_temp(hashp)
    861 	HTAB *hashp;
    862 {
    863 	sigset_t set, oset;
    864 	static char namestr[] = "_hashXXXXXX";
    865 
    866 	/* Block signals; make sure file goes away at process exit. */
    867 	(void)sigfillset(&set);
    868 	(void)sigprocmask(SIG_BLOCK, &set, &oset);
    869 	if ((hashp->fp = mkstemp(namestr)) != -1) {
    870 		(void)unlink(namestr);
    871 		(void)fcntl(hashp->fp, F_SETFD, 1);
    872 	}
    873 	(void)sigprocmask(SIG_SETMASK, &oset, (sigset_t *)NULL);
    874 	return (hashp->fp != -1 ? 0 : -1);
    875 }
    876 
    877 /*
    878  * We have to know that the key will fit, but the last entry on the page is
    879  * an overflow pair, so we need to shift things.
    880  */
    881 static void
    882 squeeze_key(sp, key, val)
    883 	u_short *sp;
    884 	const DBT *key, *val;
    885 {
    886 	register char *p;
    887 	u_short free_space, n, off, pageno;
    888 
    889 	p = (char *)sp;
    890 	n = sp[0];
    891 	free_space = FREESPACE(sp);
    892 	off = OFFSET(sp);
    893 
    894 	pageno = sp[n - 1];
    895 	off -= key->size;
    896 	sp[n - 1] = off;
    897 	memmove(p + off, key->data, key->size);
    898 	off -= val->size;
    899 	sp[n] = off;
    900 	memmove(p + off, val->data, val->size);
    901 	sp[0] = n + 2;
    902 	sp[n + 1] = pageno;
    903 	sp[n + 2] = OVFLPAGE;
    904 	FREESPACE(sp) = free_space - PAIRSIZE(key, val);
    905 	OFFSET(sp) = off;
    906 }
    907 
    908 static u_long *
    909 fetch_bitmap(hashp, ndx)
    910 	HTAB *hashp;
    911 	int ndx;
    912 {
    913 	if (ndx >= hashp->nmaps ||
    914 	    !(hashp->mapp[ndx] = malloc(hashp->BSIZE)) ||
    915 	    __get_page(hashp, (char *)hashp->mapp[ndx],
    916 	    hashp->BITMAPS[ndx], 0, 1, 1))
    917 		return (NULL);
    918 	return (hashp->mapp[ndx]);
    919 }
    920 
    921 #ifdef DEBUG4
    922 int
    923 print_chain(addr)
    924 	int addr;
    925 {
    926 	BUFHEAD *bufp;
    927 	short *bp, oaddr;
    928 
    929 	(void)fprintf(stderr, "%d ", addr);
    930 	bufp = __get_buf(hashp, addr, NULL, 0);
    931 	bp = (short *)bufp->page;
    932 	while (bp[0] && ((bp[bp[0]] == OVFLPAGE) ||
    933 		((bp[0] > 2) && bp[2] < REAL_KEY))) {
    934 		oaddr = bp[bp[0] - 1];
    935 		(void)fprintf(stderr, "%d ", (int)oaddr);
    936 		bufp = __get_buf(hashp, (int)oaddr, bufp, 0);
    937 		bp = (short *)bufp->page;
    938 	}
    939 	(void)fprintf(stderr, "\n");
    940 }
    941 #endif
    942