lhash.c revision 326663 
1/* crypto/lhash/lhash.c */
2/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
3 * All rights reserved.
4 *
5 * This package is an SSL implementation written
6 * by Eric Young (eay@cryptsoft.com).
7 * The implementation was written so as to conform with Netscapes SSL.
8 *
9 * This library is free for commercial and non-commercial use as long as
10 * the following conditions are aheared to.  The following conditions
11 * apply to all code found in this distribution, be it the RC4, RSA,
12 * lhash, DES, etc., code; not just the SSL code.  The SSL documentation
13 * included with this distribution is covered by the same copyright terms
14 * except that the holder is Tim Hudson (tjh@cryptsoft.com).
15 *
16 * Copyright remains Eric Young's, and as such any Copyright notices in
17 * the code are not to be removed.
18 * If this package is used in a product, Eric Young should be given attribution
19 * as the author of the parts of the library used.
20 * This can be in the form of a textual message at program startup or
21 * in documentation (online or textual) provided with the package.
22 *
23 * Redistribution and use in source and binary forms, with or without
24 * modification, are permitted provided that the following conditions
25 * are met:
26 * 1. Redistributions of source code must retain the copyright
27 *    notice, this list of conditions and the following disclaimer.
28 * 2. Redistributions in binary form must reproduce the above copyright
29 *    notice, this list of conditions and the following disclaimer in the
30 *    documentation and/or other materials provided with the distribution.
31 * 3. All advertising materials mentioning features or use of this software
32 *    must display the following acknowledgement:
33 *    "This product includes cryptographic software written by
34 *     Eric Young (eay@cryptsoft.com)"
35 *    The word 'cryptographic' can be left out if the rouines from the library
36 *    being used are not cryptographic related :-).
37 * 4. If you include any Windows specific code (or a derivative thereof) from
38 *    the apps directory (application code) you must include an acknowledgement:
39 *    "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
40 *
41 * THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
42 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
43 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
44 * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
45 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
46 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
47 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
48 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
49 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
50 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
51 * SUCH DAMAGE.
52 *
53 * The licence and distribution terms for any publically available version or
54 * derivative of this code cannot be changed.  i.e. this code cannot simply be
55 * copied and put under another distribution licence
56 * [including the GNU Public Licence.]
57 */
58
59/*-
60 * Code for dynamic hash table routines
61 * Author - Eric Young v 2.0
62 *
63 * 2.2 eay - added #include "crypto.h" so the memory leak checking code is
64 *           present. eay 18-Jun-98
65 *
66 * 2.1 eay - Added an 'error in last operation' flag. eay 6-May-98
67 *
68 * 2.0 eay - Fixed a bug that occurred when using lh_delete
69 *           from inside lh_doall().  As entries were deleted,
70 *           the 'table' was 'contract()ed', making some entries
71 *           jump from the end of the table to the start, there by
72 *           skipping the lh_doall() processing. eay - 4/12/95
73 *
74 * 1.9 eay - Fixed a memory leak in lh_free, the LHASH_NODEs
75 *           were not being free()ed. 21/11/95
76 *
77 * 1.8 eay - Put the stats routines into a separate file, lh_stats.c
78 *           19/09/95
79 *
80 * 1.7 eay - Removed the fputs() for realloc failures - the code
81 *           should silently tolerate them.  I have also fixed things
82 *           lint complained about 04/05/95
83 *
84 * 1.6 eay - Fixed an invalid pointers in contract/expand 27/07/92
85 *
86 * 1.5 eay - Fixed a misuse of realloc in expand 02/03/1992
87 *
88 * 1.4 eay - Fixed lh_doall so the function can call lh_delete 28/05/91
89 *
90 * 1.3 eay - Fixed a few lint problems 19/3/1991
91 *
92 * 1.2 eay - Fixed lh_doall problem 13/3/1991
93 *
94 * 1.1 eay - Added lh_doall
95 *
96 * 1.0 eay - First version
97 */
98#include <stdio.h>
99#include <string.h>
100#include <stdlib.h>
101#include <openssl/crypto.h>
102#include <openssl/lhash.h>
103
104/*
105 * A hashing implementation that appears to be based on the linear hashing
106 * alogrithm:
107 * https://en.wikipedia.org/wiki/Linear_hashing
108 *
109 * Litwin, Witold (1980), "Linear hashing: A new tool for file and table
110 * addressing", Proc. 6th Conference on Very Large Databases: 212-223
111 * http://hackthology.com/pdfs/Litwin-1980-Linear_Hashing.pdf
112 *
113 * From the wikipedia article "Linear hashing is used in the BDB Berkeley
114 * database system, which in turn is used by many software systems such as
115 * OpenLDAP, using a C implementation derived from the CACM article and first
116 * published on the Usenet in 1988 by Esmond Pitt."
117 *
118 * The CACM paper is available here:
119 * https://pdfs.semanticscholar.org/ff4d/1c5deca6269cc316bfd952172284dbf610ee.pdf
120 */
121
122const char lh_version[] = "lhash" OPENSSL_VERSION_PTEXT;
123
124#undef MIN_NODES
125#define MIN_NODES       16
126#define UP_LOAD         (2*LH_LOAD_MULT) /* load times 256 (default 2) */
127#define DOWN_LOAD       (LH_LOAD_MULT) /* load times 256 (default 1) */
128
129static int expand(_LHASH *lh);
130static void contract(_LHASH *lh);
131static LHASH_NODE **getrn(_LHASH *lh, const void *data, unsigned long *rhash);
132
133_LHASH *lh_new(LHASH_HASH_FN_TYPE h, LHASH_COMP_FN_TYPE c)
134{
135    _LHASH *ret;
136    int i;
137
138    if ((ret = OPENSSL_malloc(sizeof(_LHASH))) == NULL)
139        goto err0;
140    if ((ret->b = OPENSSL_malloc(sizeof(LHASH_NODE *) * MIN_NODES)) == NULL)
141        goto err1;
142    for (i = 0; i < MIN_NODES; i++)
143        ret->b[i] = NULL;
144    ret->comp = ((c == NULL) ? (LHASH_COMP_FN_TYPE)strcmp : c);
145    ret->hash = ((h == NULL) ? (LHASH_HASH_FN_TYPE)lh_strhash : h);
146    ret->num_nodes = MIN_NODES / 2;
147    ret->num_alloc_nodes = MIN_NODES;
148    ret->p = 0;
149    ret->pmax = MIN_NODES / 2;
150    ret->up_load = UP_LOAD;
151    ret->down_load = DOWN_LOAD;
152    ret->num_items = 0;
153
154    ret->num_expands = 0;
155    ret->num_expand_reallocs = 0;
156    ret->num_contracts = 0;
157    ret->num_contract_reallocs = 0;
158    ret->num_hash_calls = 0;
159    ret->num_comp_calls = 0;
160    ret->num_insert = 0;
161    ret->num_replace = 0;
162    ret->num_delete = 0;
163    ret->num_no_delete = 0;
164    ret->num_retrieve = 0;
165    ret->num_retrieve_miss = 0;
166    ret->num_hash_comps = 0;
167
168    ret->error = 0;
169    return (ret);
170 err1:
171    OPENSSL_free(ret);
172 err0:
173    return (NULL);
174}
175
176void lh_free(_LHASH *lh)
177{
178    unsigned int i;
179    LHASH_NODE *n, *nn;
180
181    if (lh == NULL)
182        return;
183
184    for (i = 0; i < lh->num_nodes; i++) {
185        n = lh->b[i];
186        while (n != NULL) {
187            nn = n->next;
188            OPENSSL_free(n);
189            n = nn;
190        }
191    }
192    OPENSSL_free(lh->b);
193    OPENSSL_free(lh);
194}
195
196void *lh_insert(_LHASH *lh, void *data)
197{
198    unsigned long hash;
199    LHASH_NODE *nn, **rn;
200    void *ret;
201
202    lh->error = 0;
203    if (lh->up_load <= (lh->num_items * LH_LOAD_MULT / lh->num_nodes)
204            && !expand(lh))
205        return NULL;
206
207    rn = getrn(lh, data, &hash);
208
209    if (*rn == NULL) {
210        if ((nn = (LHASH_NODE *)OPENSSL_malloc(sizeof(LHASH_NODE))) == NULL) {
211            lh->error++;
212            return (NULL);
213        }
214        nn->data = data;
215        nn->next = NULL;
216#ifndef OPENSSL_NO_HASH_COMP
217        nn->hash = hash;
218#endif
219        *rn = nn;
220        ret = NULL;
221        lh->num_insert++;
222        lh->num_items++;
223    } else {                    /* replace same key */
224
225        ret = (*rn)->data;
226        (*rn)->data = data;
227        lh->num_replace++;
228    }
229    return (ret);
230}
231
232void *lh_delete(_LHASH *lh, const void *data)
233{
234    unsigned long hash;
235    LHASH_NODE *nn, **rn;
236    void *ret;
237
238    lh->error = 0;
239    rn = getrn(lh, data, &hash);
240
241    if (*rn == NULL) {
242        lh->num_no_delete++;
243        return (NULL);
244    } else {
245        nn = *rn;
246        *rn = nn->next;
247        ret = nn->data;
248        OPENSSL_free(nn);
249        lh->num_delete++;
250    }
251
252    lh->num_items--;
253    if ((lh->num_nodes > MIN_NODES) &&
254        (lh->down_load >= (lh->num_items * LH_LOAD_MULT / lh->num_nodes)))
255        contract(lh);
256
257    return (ret);
258}
259
260void *lh_retrieve(_LHASH *lh, const void *data)
261{
262    unsigned long hash;
263    LHASH_NODE **rn;
264    void *ret;
265
266    lh->error = 0;
267    rn = getrn(lh, data, &hash);
268
269    if (*rn == NULL) {
270        lh->num_retrieve_miss++;
271        return (NULL);
272    } else {
273        ret = (*rn)->data;
274        lh->num_retrieve++;
275    }
276    return (ret);
277}
278
279static void doall_util_fn(_LHASH *lh, int use_arg, LHASH_DOALL_FN_TYPE func,
280                          LHASH_DOALL_ARG_FN_TYPE func_arg, void *arg)
281{
282    int i;
283    LHASH_NODE *a, *n;
284
285    if (lh == NULL)
286        return;
287
288    /*
289     * reverse the order so we search from 'top to bottom' We were having
290     * memory leaks otherwise
291     */
292    for (i = lh->num_nodes - 1; i >= 0; i--) {
293        a = lh->b[i];
294        while (a != NULL) {
295            /*
296             * 28/05/91 - eay - n added so items can be deleted via lh_doall
297             */
298            /*
299             * 22/05/08 - ben - eh? since a is not passed, this should not be
300             * needed
301             */
302            n = a->next;
303            if (use_arg)
304                func_arg(a->data, arg);
305            else
306                func(a->data);
307            a = n;
308        }
309    }
310}
311
312void lh_doall(_LHASH *lh, LHASH_DOALL_FN_TYPE func)
313{
314    doall_util_fn(lh, 0, func, (LHASH_DOALL_ARG_FN_TYPE)0, NULL);
315}
316
317void lh_doall_arg(_LHASH *lh, LHASH_DOALL_ARG_FN_TYPE func, void *arg)
318{
319    doall_util_fn(lh, 1, (LHASH_DOALL_FN_TYPE)0, func, arg);
320}
321
322static int expand(_LHASH *lh)
323{
324    LHASH_NODE **n, **n1, **n2, *np;
325    unsigned int p, pmax, nni, j;
326    unsigned long hash;
327
328    nni = lh->num_alloc_nodes;
329    p = lh->p;
330    pmax = lh->pmax;
331    if (p + 1 >= pmax) {
332        j = nni * 2;
333        n = OPENSSL_realloc(lh->b, (int)(sizeof(LHASH_NODE *) * j));
334        if (n == NULL) {
335            lh->error++;
336            return 0;
337        }
338        lh->b = n;
339        memset(n + nni, 0, sizeof(*n) * (j - nni));
340        lh->pmax = nni;
341        lh->num_alloc_nodes = j;
342        lh->num_expand_reallocs++;
343        lh->p = 0;
344    } else {
345        lh->p++;
346    }
347
348    lh->num_nodes++;
349    lh->num_expands++;
350    n1 = &(lh->b[p]);
351    n2 = &(lh->b[p + pmax]);
352    *n2 = NULL;
353
354    for (np = *n1; np != NULL;) {
355#ifndef OPENSSL_NO_HASH_COMP
356        hash = np->hash;
357#else
358        hash = lh->hash(np->data);
359        lh->num_hash_calls++;
360#endif
361        if ((hash % nni) != p) { /* move it */
362            *n1 = (*n1)->next;
363            np->next = *n2;
364            *n2 = np;
365        } else
366            n1 = &((*n1)->next);
367        np = *n1;
368    }
369
370    return 1;
371}
372
373static void contract(_LHASH *lh)
374{
375    LHASH_NODE **n, *n1, *np;
376
377    np = lh->b[lh->p + lh->pmax - 1];
378    lh->b[lh->p + lh->pmax - 1] = NULL; /* 24/07-92 - eay - weird but :-( */
379    if (lh->p == 0) {
380        n = (LHASH_NODE **)OPENSSL_realloc(lh->b,
381                                           (unsigned int)(sizeof(LHASH_NODE *)
382                                                          * lh->pmax));
383        if (n == NULL) {
384/*                      fputs("realloc error in lhash",stderr); */
385            lh->error++;
386            return;
387        }
388        lh->num_contract_reallocs++;
389        lh->num_alloc_nodes /= 2;
390        lh->pmax /= 2;
391        lh->p = lh->pmax - 1;
392        lh->b = n;
393    } else
394        lh->p--;
395
396    lh->num_nodes--;
397    lh->num_contracts++;
398
399    n1 = lh->b[(int)lh->p];
400    if (n1 == NULL)
401        lh->b[(int)lh->p] = np;
402    else {
403        while (n1->next != NULL)
404            n1 = n1->next;
405        n1->next = np;
406    }
407}
408
409static LHASH_NODE **getrn(_LHASH *lh, const void *data, unsigned long *rhash)
410{
411    LHASH_NODE **ret, *n1;
412    unsigned long hash, nn;
413    LHASH_COMP_FN_TYPE cf;
414
415    hash = (*(lh->hash)) (data);
416    lh->num_hash_calls++;
417    *rhash = hash;
418
419    nn = hash % lh->pmax;
420    if (nn < lh->p)
421        nn = hash % lh->num_alloc_nodes;
422
423    cf = lh->comp;
424    ret = &(lh->b[(int)nn]);
425    for (n1 = *ret; n1 != NULL; n1 = n1->next) {
426#ifndef OPENSSL_NO_HASH_COMP
427        lh->num_hash_comps++;
428        if (n1->hash != hash) {
429            ret = &(n1->next);
430            continue;
431        }
432#endif
433        lh->num_comp_calls++;
434        if (cf(n1->data, data) == 0)
435            break;
436        ret = &(n1->next);
437    }
438    return (ret);
439}
440
441/*
442 * The following hash seems to work very well on normal text strings no
443 * collisions on /usr/dict/words and it distributes on %2^n quite well, not
444 * as good as MD5, but still good.
445 */
446unsigned long lh_strhash(const char *c)
447{
448    unsigned long ret = 0;
449    long n;
450    unsigned long v;
451    int r;
452
453    if ((c == NULL) || (*c == '\0'))
454        return (ret);
455/*-
456    unsigned char b[16];
457    MD5(c,strlen(c),b);
458    return(b[0]|(b[1]<<8)|(b[2]<<16)|(b[3]<<24));
459*/
460
461    n = 0x100;
462    while (*c) {
463        v = n | (*c);
464        n += 0x100;
465        r = (int)((v >> 2) ^ v) & 0x0f;
466        ret = (ret << r) | (ret >> (32 - r));
467        ret &= 0xFFFFFFFFL;
468        ret ^= v * v;
469        c++;
470    }
471    return ((ret >> 16) ^ ret);
472}
473
474unsigned long lh_num_items(const _LHASH *lh)
475{
476    return lh ? lh->num_items : 0;
477}
478