dbm/src/hash.c

/*-
 * Copyright (c) 1990, 1993, 1994
 *  The Regents of the University of California.  All rights reserved.
 *
 * This code is derived from software contributed to Berkeley by
 * Margo Seltzer.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 * 3. ***REMOVED*** - see
 *    ftp://ftp.cs.berkeley.edu/pub/4bsd/README.Impt.License.Change
 * 4. Neither the name of the University nor the names of its contributors
 *    may be used to endorse or promote products derived from this software
 *    without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 */

#if defined(LIBC_SCCS) && !defined(lint)
static char sccsid[] = "@(#)hash.c  8.9 (Berkeley) 6/16/94";
#endif /* LIBC_SCCS and not lint */

#if !defined(_WIN32) && !defined(_WINDOWS) && !defined(macintosh)
#include <sys/param.h>
#endif

#if !defined(macintosh)
#ifdef XP_OS2
#include <sys/types.h>
#endif
#include <sys/stat.h>
#endif

#if defined(macintosh)
#include <unix.h>
#include <unistd.h>
#endif

#include <errno.h>
#include <fcntl.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>

#if !defined(_WIN32) && !defined(_WINDOWS) && !defined(macintosh)
#include <unistd.h>
#endif
#if defined(_WIN32) || defined(_WINDOWS)
#include <windows.h>
#endif

#include <assert.h>

#include "mcom_db.h"
#include "hash.h"
#include "page.h"

/*
#include "extern.h"
*/
static int alloc_segs(HTAB *, int);
static int flush_meta(HTAB *);
static int hash_access(HTAB *, ACTION, DBT *, DBT *);
static int hash_close(DB *);
static int hash_delete(const DB *, const DBT *, uint);
static int hash_fd(const DB *);
static int hash_get(const DB *, const DBT *, DBT *, uint);
static int hash_put(const DB *, DBT *, const DBT *, uint);
static void *hash_realloc(SEGMENT **, size_t, size_t);
static int hash_seq(const DB *, DBT *, DBT *, uint);
static int hash_sync(const DB *, uint);
static int hdestroy(HTAB *);
static HTAB *init_hash(HTAB *, const char *, HASHINFO *);
static int init_htab(HTAB *, int);
#if BYTE_ORDER == LITTLE_ENDIAN
static void swap_header(HTAB *);
static void swap_header_copy(HASHHDR *, HASHHDR *);
#endif

/* Fast arithmetic, relying on powers of 2, */
#define MOD(x, y) ((x) & ((y)-1))

#define RETURN_ERROR(ERR, LOC) \
    {                          \
        save_errno = ERR;      \
        goto LOC;              \
    }

/* Return values */
#define SUCCESS (0)
#define DBM_ERROR (-1)
#define ABNORMAL (1)

#ifdef HASH_STATISTICS
int hash_accesses, hash_collisions, hash_expansions, hash_overflows;
#endif

/* A new Lou (montulli@mozilla.com) routine.
 *
 * The database is screwed.
 *
 * This closes the file, flushing buffers as appropriate.
 */
static void
dbm_remove_database(DB *dbp)
{
    HTAB *hashp = (HTAB *)dbp->internal;

    assert(0);

    if (!hashp)
        return;
    hdestroy(hashp);
    dbp->internal = NULL;
}

/************************** INTERFACE ROUTINES ***************************/
/* OPEN/CLOSE */

extern DB *
dbm_hash_open(const char *file, int flags, int mode, const HASHINFO *info, int dflags)
{
    HTAB *hashp = NULL;
    struct stat statbuf;
    DB *dbp;
    int bpages, hdrsize, new_table, nsegs, save_errno;

    if ((flags & O_ACCMODE) == O_WRONLY) {
        errno = EINVAL;
        return NULL;
    }

    /* zero the statbuffer so that
     * we can check it for a non-zero
     * date to see if stat succeeded
     */
    memset(&statbuf, 0, sizeof(struct stat));

    if (!(hashp = (HTAB *)calloc(1, sizeof(HTAB)))) {
        errno = ENOMEM;
        return NULL;
    }
    hashp->fp = NO_FILE;
    if (file)
        hashp->filename = strdup(file);

    /*
     * Even if user wants write only, we need to be able to read
     * the actual file, so we need to open it read/write. But, the
     * field in the hashp structure needs to be accurate so that
     * we can check accesses.
     */
    hashp->flags = flags;

    new_table = 0;
    if (!file || (flags & O_TRUNC) || (stat(file, &statbuf) && (errno == ENOENT))) {
        if (errno == ENOENT)
            errno = 0; /* Just in case someone looks at errno */
        new_table = 1;
    } else if (statbuf.st_mtime && statbuf.st_size == 0) {
        /* check for a zero length file and delete it
         * if it exists
         */
        new_table = 1;
    }
    hashp->file_size = statbuf.st_size;

    if (file) {
#if defined(_WIN32) || defined(_WINDOWS) || defined(macintosh) || defined(XP_OS2)
        if ((hashp->fp = DBFILE_OPEN(file, flags | O_BINARY, mode)) == -1)
            RETURN_ERROR(errno, error1);
#else
        if ((hashp->fp = open(file, flags, mode)) == -1)
            RETURN_ERROR(errno, error1);
        (void)fcntl(hashp->fp, F_SETFD, 1);
#endif
    }
    if (new_table) {
        if (!init_hash(hashp, file, (HASHINFO *)info))
            RETURN_ERROR(errno, error1);
    } else {
        /* Table already exists */
        if (info && info->hash)
            hashp->hash = info->hash;
        else
            hashp->hash = dbm_default_hash;

        hdrsize = read(hashp->fp, (char *)&hashp->hdr, sizeof(HASHHDR));
        if (hdrsize == -1)
            RETURN_ERROR(errno, error1);
        if (hdrsize != sizeof(HASHHDR))
            RETURN_ERROR(EFTYPE, error1);
#if BYTE_ORDER == LITTLE_ENDIAN
        swap_header(hashp);
#endif
        /* Verify file type, versions and hash function */
        if (hashp->MAGIC != HASHMAGIC)
            RETURN_ERROR(EFTYPE, error1);
#define OLDHASHVERSION 1
        if (hashp->VERSION != HASHVERSION &&
            hashp->VERSION != OLDHASHVERSION)
            RETURN_ERROR(EFTYPE, error1);
        if (hashp->hash(CHARKEY, sizeof(CHARKEY)) != hashp->H_CHARKEY)
            RETURN_ERROR(EFTYPE, error1);
        if (hashp->NKEYS < 0) /* Old bad database. */
            RETURN_ERROR(EFTYPE, error1);

        /*
         * Figure out how many segments we need.  Max_Bucket is the
         * maximum bucket number, so the number of buckets is
         * max_bucket + 1.
         */
        nsegs = (hashp->MAX_BUCKET + 1 + hashp->SGSIZE - 1) /
                hashp->SGSIZE;
        hashp->nsegs = 0;
        if (alloc_segs(hashp, nsegs))
            /* If alloc_segs fails, errno will have been set.  */
            RETURN_ERROR(errno, error1);
        /* Read in bitmaps */
        bpages = (hashp->SPARES[hashp->OVFL_POINT] +
                  (hashp->BSIZE << BYTE_SHIFT) - 1) >>
                 (hashp->BSHIFT + BYTE_SHIFT);

        hashp->nmaps = bpages;
        (void)memset(&hashp->mapp[0], 0, bpages * sizeof(uint32 *));
    }

    /* Initialize Buffer Manager */
    if (info && info->cachesize)
        dbm_buf_init(hashp, (int32)info->cachesize);
    else
        dbm_buf_init(hashp, DEF_BUFSIZE);

    hashp->new_file = new_table;
#ifdef macintosh
    hashp->save_file = file && !(hashp->flags & O_RDONLY);
#else
    hashp->save_file = file && (hashp->flags & O_RDWR);
#endif
    hashp->cbucket = -1;
    if (!(dbp = (DB *)malloc(sizeof(DB)))) {
        RETURN_ERROR(ENOMEM, error1);
    }
    dbp->internal = hashp;
    dbp->close = hash_close;
    dbp->del = hash_delete;
    dbp->fd = hash_fd;
    dbp->get = hash_get;
    dbp->put = hash_put;
    dbp->seq = hash_seq;
    dbp->sync = hash_sync;
    dbp->type = DB_HASH;

#ifdef HASH_STATISTICS
    hash_overflows = hash_accesses = hash_collisions = hash_expansions = 0;
#endif
    return (dbp);

error1:
    hdestroy(hashp);
    errno = save_errno;
    return (NULL);
}

static int
hash_close(DB *dbp)
{
    HTAB *hashp;
    int retval;

    if (!dbp)
        return (DBM_ERROR);

    hashp = (HTAB *)dbp->internal;
    if (!hashp)
        return (DBM_ERROR);

    retval = hdestroy(hashp);
    free(dbp);
    return (retval);
}

static int
hash_fd(const DB *dbp)
{
    HTAB *hashp;

    if (!dbp)
        return (DBM_ERROR);

    hashp = (HTAB *)dbp->internal;
    if (!hashp)
        return (DBM_ERROR);

    if (hashp->fp == -1) {
        errno = ENOENT;
        return (-1);
    }
    return (hashp->fp);
}

/************************** LOCAL CREATION ROUTINES **********************/
static HTAB *
init_hash(HTAB *hashp, const char *file, HASHINFO *info)
{
    struct stat statbuf;
    int nelem;

    nelem = 1;
    hashp->NKEYS = 0;
    hashp->LORDER = BYTE_ORDER;
    hashp->BSIZE = DEF_BUCKET_SIZE;
    hashp->BSHIFT = DEF_BUCKET_SHIFT;
    hashp->SGSIZE = DEF_SEGSIZE;
    hashp->SSHIFT = DEF_SEGSIZE_SHIFT;
    hashp->DSIZE = DEF_DIRSIZE;
    hashp->FFACTOR = DEF_FFACTOR;
    hashp->hash = dbm_default_hash;
    memset(hashp->SPARES, 0, sizeof(hashp->SPARES));
    memset(hashp->BITMAPS, 0, sizeof(hashp->BITMAPS));

    /* Fix bucket size to be optimal for file system */
    if (file != NULL) {
        if (stat(file, &statbuf))
            return (NULL);

#if !defined(_WIN32) && !defined(_WINDOWS) && !defined(macintosh) && !defined(XP_OS2)
#if defined(__QNX__) && !defined(__QNXNTO__)
        hashp->BSIZE = 512; /* preferred blk size on qnx4 */
#else
        hashp->BSIZE = statbuf.st_blksize;
#endif

        /* new code added by Lou to reduce block
         * size down below MAX_BSIZE
         */
        if (hashp->BSIZE > MAX_BSIZE)
            hashp->BSIZE = MAX_BSIZE;
#endif
        hashp->BSHIFT = dbm_log2((uint32)hashp->BSIZE);
    }

    if (info) {
        if (info->bsize) {
            /* Round pagesize up to power of 2 */
            hashp->BSHIFT = dbm_log2(info->bsize);
            hashp->BSIZE = 1 << hashp->BSHIFT;
            if (hashp->BSIZE > MAX_BSIZE) {
                errno = EINVAL;
                return (NULL);
            }
        }
        if (info->ffactor)
            hashp->FFACTOR = info->ffactor;
        if (info->hash)
            hashp->hash = info->hash;
        if (info->nelem)
            nelem = info->nelem;
        if (info->lorder) {
            if (info->lorder != BIG_ENDIAN &&
                info->lorder != LITTLE_ENDIAN) {
                errno = EINVAL;
                return (NULL);
            }
            hashp->LORDER = info->lorder;
        }
    }
    /* init_htab sets errno if it fails */
    if (init_htab(hashp, nelem))
        return (NULL);
    else
        return (hashp);
}
/*
 * This calls alloc_segs which may run out of memory.  Alloc_segs will
 * set errno, so we just pass the error information along.
 *
 * Returns 0 on No Error
 */
static int
init_htab(HTAB *hashp, int nelem)
{
    register int nbuckets, nsegs;
    int l2;

    /*
     * Divide number of elements by the fill factor and determine a
     * desired number of buckets.  Allocate space for the next greater
     * power of two number of buckets.
     */
    nelem = (nelem - 1) / hashp->FFACTOR + 1;

    l2 = dbm_log2((uint32)PR_MAX(nelem, 2));
    nbuckets = 1 << l2;

    hashp->SPARES[l2] = l2 + 1;
    hashp->SPARES[l2 + 1] = l2 + 1;
    hashp->OVFL_POINT = l2;
    hashp->LAST_FREED = 2;

    /* First bitmap page is at: splitpoint l2 page offset 1 */
    if (dbm_ibitmap(hashp, (int)OADDR_OF(l2, 1), l2 + 1, 0))
        return (-1);

    hashp->MAX_BUCKET = hashp->LOW_MASK = nbuckets - 1;
    hashp->HIGH_MASK = (nbuckets << 1) - 1;
    hashp->HDRPAGES = ((PR_MAX(sizeof(HASHHDR), MINHDRSIZE) - 1) >>
                       hashp->BSHIFT) +
                      1;

    nsegs = (nbuckets - 1) / hashp->SGSIZE + 1;
    nsegs = 1 << dbm_log2((uint32)nsegs);

    if (nsegs > hashp->DSIZE)
        hashp->DSIZE = nsegs;
    return (alloc_segs(hashp, nsegs));
}

/********************** DESTROY/CLOSE ROUTINES ************************/

/*
 * Flushes any changes to the file if necessary and destroys the hashp
 * structure, freeing all allocated space.
 */
static int
hdestroy(HTAB *hashp)
{
    int i, save_errno;

    save_errno = 0;

#ifdef HASH_STATISTICS
    (void)fprintf(stderr, "hdestroy: accesses %ld collisions %ld\n",
                  hash_accesses, hash_collisions);
    (void)fprintf(stderr, "hdestroy: expansions %ld\n",
                  hash_expansions);
    (void)fprintf(stderr, "hdestroy: overflows %ld\n",
                  hash_overflows);
    (void)fprintf(stderr, "keys %ld maxp %d segmentcount %d\n",
                  hashp->NKEYS, hashp->MAX_BUCKET, hashp->nsegs);

    for (i = 0; i < NCACHED; i++)
        (void)fprintf(stderr,
                      "spares[%d] = %d\n", i, hashp->SPARES[i]);
#endif
    /*
     * Call on buffer manager to free buffers, and if required,
     * write them to disk.
     */
    if (dbm_buf_free(hashp, 1, hashp->save_file))
        save_errno = errno;
    if (hashp->dir) {
        free(*hashp->dir); /* Free initial segments */
        /* Free extra segments */
        while (hashp->exsegs--)
            free(hashp->dir[--hashp->nsegs]);
        free(hashp->dir);
    }
    if (flush_meta(hashp) && !save_errno)
        save_errno = errno;
    /* Free Bigmaps */
    for (i = 0; i < hashp->nmaps; i++)
        if (hashp->mapp[i])
            free(hashp->mapp[i]);

    if (hashp->fp != -1)
        (void)close(hashp->fp);

    if (hashp->filename) {
#if defined(_WIN32) || defined(_WINDOWS) || defined(XP_OS2)
        if (hashp->is_temp)
            (void)unlink(hashp->filename);
#endif
        free(hashp->filename);
    }
    if (hashp->tmp_buf)
        free(hashp->tmp_buf);
    if (hashp->tmp_key)
        free(hashp->tmp_key);
    free(hashp);
    if (save_errno) {
        errno = save_errno;
        return (DBM_ERROR);
    }
    return (SUCCESS);
}

#if defined(_WIN32) || defined(_WINDOWS)
/*
 * Close and reopen file to force file length update on windows.
 *
 * Returns:
 *   0 == OK
 *  -1 DBM_ERROR
 */
static int
update_EOF(HTAB *hashp)
{
#if defined(DBM_REOPEN_ON_FLUSH)
    char *file = hashp->filename;
    off_t file_size;
    int flags;
    int mode = -1;
    struct stat statbuf;

    memset(&statbuf, 0, sizeof statbuf);

    /* make sure we won't lose the file by closing it. */
    if (!file || (stat(file, &statbuf) && (errno == ENOENT))) {
        /* pretend we did it. */
        return 0;
    }

    (void)close(hashp->fp);

    flags = hashp->flags & ~(O_TRUNC | O_CREAT | O_EXCL);

    if ((hashp->fp = DBFILE_OPEN(file, flags | O_BINARY, mode)) == -1)
        return -1;
    file_size = lseek(hashp->fp, (off_t)0, SEEK_END);
    if (file_size == -1)
        return -1;
    hashp->file_size = file_size;
    return 0;
#else
    int fd = hashp->fp;
    off_t file_size = lseek(fd, (off_t)0, SEEK_END);
    HANDLE handle = (HANDLE)_get_osfhandle(fd);
    BOOL cool = FlushFileBuffers(handle);
#ifdef DEBUG3
    if (!cool) {
        DWORD err = GetLastError();
        (void)fprintf(stderr,
                      "FlushFileBuffers failed, last error = %d, 0x%08x\n",
                      err, err);
    }
#endif
    if (file_size == -1)
        return -1;
    hashp->file_size = file_size;
    return cool ? 0 : -1;
#endif
}
#endif

/*
 * Write modified pages to disk
 *
 * Returns:
 *   0 == OK
 *  -1 DBM_ERROR
 */
static int
hash_sync(const DB *dbp, uint flags)
{
    HTAB *hashp;

    if (flags != 0) {
        errno = EINVAL;
        return (DBM_ERROR);
    }

    if (!dbp)
        return (DBM_ERROR);

    hashp = (HTAB *)dbp->internal;
    if (!hashp)
        return (DBM_ERROR);

    if (!hashp->save_file)
        return (0);
    if (dbm_buf_free(hashp, 0, 1) || flush_meta(hashp))
        return (DBM_ERROR);
#if defined(_WIN32) || defined(_WINDOWS)
    if (hashp->updateEOF && hashp->filename && !hashp->is_temp) {
        int status = update_EOF(hashp);
        hashp->updateEOF = 0;
        if (status)
            return status;
    }
#endif
    hashp->new_file = 0;
    return (0);
}

/*
 * Returns:
 *   0 == OK
 *  -1 indicates that errno should be set
 */
static int
flush_meta(HTAB *hashp)
{
    HASHHDR *whdrp;
#if BYTE_ORDER == LITTLE_ENDIAN
    HASHHDR whdr;
#endif
    int fp, i, wsize;

    if (!hashp->save_file)
        return (0);
    hashp->MAGIC = HASHMAGIC;
    hashp->VERSION = HASHVERSION;
    hashp->H_CHARKEY = hashp->hash(CHARKEY, sizeof(CHARKEY));

    fp = hashp->fp;
    whdrp = &hashp->hdr;
#if BYTE_ORDER == LITTLE_ENDIAN
    whdrp = &whdr;
    swap_header_copy(&hashp->hdr, whdrp);
#endif
    if ((lseek(fp, (off_t)0, SEEK_SET) == -1) ||
        ((wsize = write(fp, (char *)whdrp, sizeof(HASHHDR))) == -1))
        return (-1);
    else if (wsize != sizeof(HASHHDR)) {
        errno = EFTYPE;
        hashp->dbmerrno = errno;
        return (-1);
    }
    for (i = 0; i < NCACHED; i++)
        if (hashp->mapp[i])
            if (dbm_put_page(hashp, (char *)hashp->mapp[i],
                             hashp->BITMAPS[i], 0, 1))
                return (-1);
    return (0);
}

/*******************************SEARCH ROUTINES *****************************/
/*
 * All the access routines return
 *
 * Returns:
 *   0 on SUCCESS
 *   1 to indicate an external DBM_ERROR (i.e. key not found, etc)
 *  -1 to indicate an internal DBM_ERROR (i.e. out of memory, etc)
 */
static int
hash_get(
    const DB *dbp,
    const DBT *key,
    DBT *data,
    uint flag)
{
    HTAB *hashp;
    int rv;

    hashp = (HTAB *)dbp->internal;
    if (!hashp)
        return (DBM_ERROR);

    if (flag) {
        hashp->dbmerrno = errno = EINVAL;
        return (DBM_ERROR);
    }

    rv = hash_access(hashp, HASH_GET, (DBT *)key, data);

    if (rv == DATABASE_CORRUPTED_ERROR) {
#if defined(unix) && defined(DEBUG)
        printf("\n\nDBM Database has been corrupted, tell Lou...\n\n");
#endif
        dbm_remove_database((DB *)dbp);
    }

    return (rv);
}

static int
hash_put(
    const DB *dbp,
    DBT *key,
    const DBT *data,
    uint flag)
{
    HTAB *hashp;
    int rv;

    hashp = (HTAB *)dbp->internal;
    if (!hashp)
        return (DBM_ERROR);

    if (flag && flag != R_NOOVERWRITE) {
        hashp->dbmerrno = errno = EINVAL;
        return (DBM_ERROR);
    }
    if ((hashp->flags & O_ACCMODE) == O_RDONLY) {
        hashp->dbmerrno = errno = EPERM;
        return (DBM_ERROR);
    }

    rv = hash_access(hashp, flag == R_NOOVERWRITE ? HASH_PUTNEW : HASH_PUT,
                     (DBT *)key, (DBT *)data);

    if (rv == DATABASE_CORRUPTED_ERROR) {
#if defined(unix) && defined(DEBUG)
        printf("\n\nDBM Database has been corrupted, tell Lou...\n\n");
#endif
        dbm_remove_database((DB *)dbp);
    }

    return (rv);
}

static int
hash_delete(
    const DB *dbp,
    const DBT *key,
    uint flag) /* Ignored */
{
    HTAB *hashp;
    int rv;

    hashp = (HTAB *)dbp->internal;
    if (!hashp)
        return (DBM_ERROR);

    if (flag && flag != R_CURSOR) {
        hashp->dbmerrno = errno = EINVAL;
        return (DBM_ERROR);
    }
    if ((hashp->flags & O_ACCMODE) == O_RDONLY) {
        hashp->dbmerrno = errno = EPERM;
        return (DBM_ERROR);
    }
    rv = hash_access(hashp, HASH_DELETE, (DBT *)key, NULL);

    if (rv == DATABASE_CORRUPTED_ERROR) {
#if defined(unix) && defined(DEBUG)
        printf("\n\nDBM Database has been corrupted, tell Lou...\n\n");
#endif
        dbm_remove_database((DB *)dbp);
    }

    return (rv);
}

#define MAX_OVERFLOW_HASH_ACCESS_LOOPS 2000
/*
 * Assume that hashp has been set in wrapper routine.
 */
static int
hash_access(
    HTAB *hashp,
    ACTION action,
    DBT *key, DBT *val)
{
    register BUFHEAD *rbufp;
    BUFHEAD *bufp, *save_bufp;
    register uint16 *bp;
    register long n, ndx, off;
    register size_t size;
    register char *kp;
    uint16 pageno;
    uint32 ovfl_loop_count = 0;
    int32 last_overflow_page_no = -1;

#ifdef HASH_STATISTICS
    hash_accesses++;
#endif

    off = hashp->BSIZE;
    size = key->size;
    kp = (char *)key->data;
    rbufp = dbm_get_buf(hashp, dbm_call_hash(hashp, kp, size), NULL, 0);
    if (!rbufp)
        return (DATABASE_CORRUPTED_ERROR);
    save_bufp = rbufp;

    /* Pin the bucket chain */
    rbufp->flags |= BUF_PIN;
    for (bp = (uint16 *)rbufp->page, n = *bp++, ndx = 1; ndx < n;) {

        if (bp[1] >= REAL_KEY) {
            /* Real key/data pair */
            if (size == (unsigned long)(off - *bp) &&
                memcmp(kp, rbufp->page + *bp, size) == 0)
                goto found;
            off = bp[1];
#ifdef HASH_STATISTICS
            hash_collisions++;
#endif
            bp += 2;
            ndx += 2;
        } else if (bp[1] == OVFLPAGE) {

            /* database corruption: overflow loop detection */
            if (last_overflow_page_no == (int32)*bp)
                return (DATABASE_CORRUPTED_ERROR);

            last_overflow_page_no = *bp;

            rbufp = dbm_get_buf(hashp, *bp, rbufp, 0);
            if (!rbufp) {
                save_bufp->flags &= ~BUF_PIN;
                return (DBM_ERROR);
            }

            ovfl_loop_count++;
            if (ovfl_loop_count > MAX_OVERFLOW_HASH_ACCESS_LOOPS)
                return (DATABASE_CORRUPTED_ERROR);

            /* FOR LOOP INIT */
            bp = (uint16 *)rbufp->page;
            n = *bp++;
            ndx = 1;
            off = hashp->BSIZE;
        } else if (bp[1] < REAL_KEY) {
            if ((ndx =
                     dbm_find_bigpair(hashp, rbufp, ndx, kp, (int)size)) > 0)
                goto found;
            if (ndx == -2) {
                bufp = rbufp;
                if (!(pageno =
                          dbm_find_last_page(hashp, &bufp))) {
                    ndx = 0;
                    rbufp = bufp;
                    break; /* FOR */
                }
                rbufp = dbm_get_buf(hashp, pageno, bufp, 0);
                if (!rbufp) {
                    save_bufp->flags &= ~BUF_PIN;
                    return (DBM_ERROR);
                }
                /* FOR LOOP INIT */
                bp = (uint16 *)rbufp->page;
                n = *bp++;
                ndx = 1;
                off = hashp->BSIZE;
            } else {
                save_bufp->flags &= ~BUF_PIN;
                return (DBM_ERROR);
            }
        }
    }

    /* Not found */
    switch (action) {
        case HASH_PUT:
        case HASH_PUTNEW:
            if (dbm_addel(hashp, rbufp, key, val)) {
                save_bufp->flags &= ~BUF_PIN;
                return (DBM_ERROR);
            } else {
                save_bufp->flags &= ~BUF_PIN;
                return (SUCCESS);
            }
        case HASH_GET:
        case HASH_DELETE:
        default:
            save_bufp->flags &= ~BUF_PIN;
            return (ABNORMAL);
    }

found:
    switch (action) {
        case HASH_PUTNEW:
            save_bufp->flags &= ~BUF_PIN;
            return (ABNORMAL);
        case HASH_GET:
            bp = (uint16 *)rbufp->page;
            if (bp[ndx + 1] < REAL_KEY) {
                if (dbm_big_return(hashp, rbufp, ndx, val, 0))
                    return (DBM_ERROR);
            } else {
                val->data = (uint8 *)rbufp->page + (int)bp[ndx + 1];
                val->size = bp[ndx] - bp[ndx + 1];
            }
            break;
        case HASH_PUT:
            if ((dbm_delpair(hashp, rbufp, ndx)) ||
                (dbm_addel(hashp, rbufp, key, val))) {
                save_bufp->flags &= ~BUF_PIN;
                return (DBM_ERROR);
            }
            break;
        case HASH_DELETE:
            if (dbm_delpair(hashp, rbufp, ndx))
                return (DBM_ERROR);
            break;
        default:
            abort();
    }
    save_bufp->flags &= ~BUF_PIN;
    return (SUCCESS);
}

static int
hash_seq(
    const DB *dbp,
    DBT *key, DBT *data,
    uint flag)
{
    register uint32 bucket;
    register BUFHEAD *bufp = NULL;
    HTAB *hashp;
    uint16 *bp, ndx;

    hashp = (HTAB *)dbp->internal;
    if (!hashp)
        return (DBM_ERROR);

    if (flag && flag != R_FIRST && flag != R_NEXT) {
        hashp->dbmerrno = errno = EINVAL;
        return (DBM_ERROR);
    }
#ifdef HASH_STATISTICS
    hash_accesses++;
#endif
    if ((hashp->cbucket < 0) || (flag == R_FIRST)) {
        hashp->cbucket = 0;
        hashp->cndx = 1;
        hashp->cpage = NULL;
    }

    for (bp = NULL; !bp || !bp[0];) {
        if (!(bufp = hashp->cpage)) {
            for (bucket = hashp->cbucket;
                 bucket <= (uint32)hashp->MAX_BUCKET;
                 bucket++, hashp->cndx = 1) {
                bufp = dbm_get_buf(hashp, bucket, NULL, 0);
                if (!bufp)
                    return (DBM_ERROR);
                hashp->cpage = bufp;
                bp = (uint16 *)bufp->page;
                if (bp[0])
                    break;
            }
            hashp->cbucket = bucket;
            if (hashp->cbucket > hashp->MAX_BUCKET) {
                hashp->cbucket = -1;
                return (ABNORMAL);
            }
        } else
            bp = (uint16 *)hashp->cpage->page;

#ifdef DEBUG
        assert(bp);
        assert(bufp);
#endif
        while (bp[hashp->cndx + 1] == OVFLPAGE) {
            bufp = hashp->cpage =
                dbm_get_buf(hashp, bp[hashp->cndx], bufp, 0);
            if (!bufp)
                return (DBM_ERROR);
            bp = (uint16 *)(bufp->page);
            hashp->cndx = 1;
        }
        if (!bp[0]) {
            hashp->cpage = NULL;
            ++hashp->cbucket;
        }
    }
    ndx = hashp->cndx;
    if (bp[ndx + 1] < REAL_KEY) {
        if (dbm_big_keydata(hashp, bufp, key, data, 1))
            return (DBM_ERROR);
    } else {
        key->data = (uint8 *)hashp->cpage->page + bp[ndx];
        key->size = (ndx > 1 ? bp[ndx - 1] : hashp->BSIZE) - bp[ndx];
        data->data = (uint8 *)hashp->cpage->page + bp[ndx + 1];
        data->size = bp[ndx] - bp[ndx + 1];
        ndx += 2;
        if (ndx > bp[0]) {
            hashp->cpage = NULL;
            hashp->cbucket++;
            hashp->cndx = 1;
        } else
            hashp->cndx = ndx;
    }
    return (SUCCESS);
}

/********************************* UTILITIES ************************/

/*
 * Returns:
 *   0 ==> OK
 *  -1 ==> Error
 */
extern int
dbm_expand_table(HTAB *hashp)
{
    uint32 old_bucket, new_bucket;
    int new_segnum, spare_ndx;
    size_t dirsize;

#ifdef HASH_STATISTICS
    hash_expansions++;
#endif
    new_bucket = ++hashp->MAX_BUCKET;
    old_bucket = (hashp->MAX_BUCKET & hashp->LOW_MASK);

    new_segnum = new_bucket >> hashp->SSHIFT;

    /* Check if we need a new segment */
    if (new_segnum >= hashp->nsegs) {
        /* Check if we need to expand directory */
        if (new_segnum >= hashp->DSIZE) {
            /* Reallocate directory */
            dirsize = hashp->DSIZE * sizeof(SEGMENT *);
            if (!hash_realloc(&hashp->dir, dirsize, dirsize << 1))
                return (-1);
            hashp->DSIZE = dirsize << 1;
        }
        if ((hashp->dir[new_segnum] =
                 (SEGMENT)calloc((size_t)hashp->SGSIZE, sizeof(BUFHEAD *))) == NULL)
            return (-1);
        hashp->exsegs++;
        hashp->nsegs++;
    }
    /*
     * If the split point is increasing (MAX_BUCKET's log base 2
     * * increases), we need to copy the current contents of the spare
     * split bucket to the next bucket.
     */
    spare_ndx = dbm_log2((uint32)(hashp->MAX_BUCKET + 1));
    if (spare_ndx > hashp->OVFL_POINT) {
        hashp->SPARES[spare_ndx] = hashp->SPARES[hashp->OVFL_POINT];
        hashp->OVFL_POINT = spare_ndx;
    }

    if (new_bucket > (uint32)hashp->HIGH_MASK) {
        /* Starting a new doubling */
        hashp->LOW_MASK = hashp->HIGH_MASK;
        hashp->HIGH_MASK = new_bucket | hashp->LOW_MASK;
    }
    /* Relocate records to the new bucket */
    return (dbm_split_page(hashp, old_bucket, new_bucket));
}

/*
 * If realloc guarantees that the pointer is not destroyed if the realloc
 * fails, then this routine can go away.
 */
static void *
hash_realloc(
    SEGMENT **p_ptr,
    size_t oldsize, size_t newsize)
{
    register void *p;

    if ((p = malloc(newsize))) {
        memmove(p, *p_ptr, oldsize);
        memset((char *)p + oldsize, 0, newsize - oldsize);
        free(*p_ptr);
        *p_ptr = (SEGMENT *)p;
    }
    return (p);
}

extern uint32
dbm_call_hash(HTAB *hashp, char *k, size_t len)
{
    uint32 n, bucket;

    n = hashp->hash(k, len);
    bucket = n & hashp->HIGH_MASK;
    if (bucket > (uint32)hashp->MAX_BUCKET)
        bucket = bucket & hashp->LOW_MASK;
    return (bucket);
}

/*
 * Allocate segment table.  On error, set errno.
 *
 * Returns 0 on success
 */
static int
alloc_segs(
    HTAB *hashp,
    int nsegs)
{
    register int i;
    register SEGMENT store;

    if ((hashp->dir =
             (SEGMENT *)calloc((size_t)hashp->DSIZE, sizeof(SEGMENT))) == NULL) {
        errno = ENOMEM;
        return (-1);
    }
    /* Allocate segments */
    if ((store =
             (SEGMENT)calloc((size_t)nsegs << hashp->SSHIFT, sizeof(BUFHEAD *))) == NULL) {
        errno = ENOMEM;
        return (-1);
    }
    for (i = 0; i < nsegs; i++, hashp->nsegs++)
        hashp->dir[i] = &store[i << hashp->SSHIFT];
    return (0);
}

#if BYTE_ORDER == LITTLE_ENDIAN
/*
 * Hashp->hdr needs to be byteswapped.
 */
static void
swap_header_copy(
    HASHHDR *srcp, HASHHDR *destp)
{
    int i;

    P_32_COPY(srcp->magic, destp->magic);
    P_32_COPY(srcp->version, destp->version);
    P_32_COPY(srcp->lorder, destp->lorder);
    P_32_COPY(srcp->bsize, destp->bsize);
    P_32_COPY(srcp->bshift, destp->bshift);
    P_32_COPY(srcp->dsize, destp->dsize);
    P_32_COPY(srcp->ssize, destp->ssize);
    P_32_COPY(srcp->sshift, destp->sshift);
    P_32_COPY(srcp->ovfl_point, destp->ovfl_point);
    P_32_COPY(srcp->last_freed, destp->last_freed);
    P_32_COPY(srcp->max_bucket, destp->max_bucket);
    P_32_COPY(srcp->high_mask, destp->high_mask);
    P_32_COPY(srcp->low_mask, destp->low_mask);
    P_32_COPY(srcp->ffactor, destp->ffactor);
    P_32_COPY(srcp->nkeys, destp->nkeys);
    P_32_COPY(srcp->hdrpages, destp->hdrpages);
    P_32_COPY(srcp->h_charkey, destp->h_charkey);
    for (i = 0; i < NCACHED; i++) {
        P_32_COPY(srcp->spares[i], destp->spares[i]);
        P_16_COPY(srcp->bitmaps[i], destp->bitmaps[i]);
    }
}

static void
swap_header(HTAB *hashp)
{
    HASHHDR *hdrp;
    int i;

    hdrp = &hashp->hdr;

    M_32_SWAP(hdrp->magic);
    M_32_SWAP(hdrp->version);
    M_32_SWAP(hdrp->lorder);
    M_32_SWAP(hdrp->bsize);
    M_32_SWAP(hdrp->bshift);
    M_32_SWAP(hdrp->dsize);
    M_32_SWAP(hdrp->ssize);
    M_32_SWAP(hdrp->sshift);
    M_32_SWAP(hdrp->ovfl_point);
    M_32_SWAP(hdrp->last_freed);
    M_32_SWAP(hdrp->max_bucket);
    M_32_SWAP(hdrp->high_mask);
    M_32_SWAP(hdrp->low_mask);
    M_32_SWAP(hdrp->ffactor);
    M_32_SWAP(hdrp->nkeys);
    M_32_SWAP(hdrp->hdrpages);
    M_32_SWAP(hdrp->h_charkey);
    for (i = 0; i < NCACHED; i++) {
        M_32_SWAP(hdrp->spares[i]);
        M_16_SWAP(hdrp->bitmaps[i]);
    }
}
#endif