crypto/lhash/lhash.c

   1 /*
   2  * Copyright 1995-2022 The OpenSSL Project Authors. All Rights Reserved.
   3  *
   4  * Licensed under the Apache License 2.0 (the "License").  You may not use
   5  * this file except in compliance with the License.  You can obtain a copy
   6  * in the file LICENSE in the source distribution or at
   7  * https://www.openssl.org/source/license.html
   8  */
   9
  10 #include <stdio.h>
  11 #include <string.h>
  12 #include <stdlib.h>
  13 #include <openssl/crypto.h>
  14 #include <openssl/lhash.h>
  15 #include <openssl/err.h>
  16 #include "crypto/ctype.h"
  17 #include "crypto/lhash.h"
  18 #include "lhash_local.h"
  19
  20 /*
  21  * A hashing implementation that appears to be based on the linear hashing
  22  * algorithm:
  23  * https://en.wikipedia.org/wiki/Linear_hashing
  24  *
  25  * Litwin, Witold (1980), "Linear hashing: A new tool for file and table
  26  * addressing", Proc. 6th Conference on Very Large Databases: 212-223
  27  * https://hackthology.com/pdfs/Litwin-1980-Linear_Hashing.pdf
  28  *
  29  * From the Wikipedia article "Linear hashing is used in the BDB Berkeley
  30  * database system, which in turn is used by many software systems such as
  31  * OpenLDAP, using a C implementation derived from the CACM article and first
  32  * published on the Usenet in 1988 by Esmond Pitt."
  33  *
  34  * The CACM paper is available here:
  35  * https://pdfs.semanticscholar.org/ff4d/1c5deca6269cc316bfd952172284dbf610ee.pdf
  36  */
  37
  38 #undef MIN_NODES
  39 #define MIN_NODES       16
  40 #define UP_LOAD         (2*LH_LOAD_MULT) /* load times 256 (default 2) */
  41 #define DOWN_LOAD       (LH_LOAD_MULT) /* load times 256 (default 1) */
  42
  43 static int expand(OPENSSL_LHASH *lh);
  44 static void contract(OPENSSL_LHASH *lh);
  45 static OPENSSL_LH_NODE **getrn(OPENSSL_LHASH *lh, const void *data, unsigned long *rhash);
  46
  47 OPENSSL_LHASH *OPENSSL_LH_new(OPENSSL_LH_HASHFUNC h, OPENSSL_LH_COMPFUNC c)
  48 {
  49     OPENSSL_LHASH *ret;
  50
  51     if ((ret = OPENSSL_zalloc(sizeof(*ret))) == NULL) {
  52         /*
  53          * Do not set the error code, because the ERR code uses LHASH
  54          * and we want to avoid possible endless error loop.
  55          * ERR_raise(ERR_LIB_CRYPTO, ERR_R_MALLOC_FAILURE);
  56          */
  57         return NULL;
  58     }
  59     if ((ret->b = OPENSSL_zalloc(sizeof(*ret->b) * MIN_NODES)) == NULL)
  60         goto err;
  61     ret->comp = ((c == NULL) ? (OPENSSL_LH_COMPFUNC)strcmp : c);
  62     ret->hash = ((h == NULL) ? (OPENSSL_LH_HASHFUNC)OPENSSL_LH_strhash : h);
  63     ret->num_nodes = MIN_NODES / 2;
  64     ret->num_alloc_nodes = MIN_NODES;
  65     ret->pmax = MIN_NODES / 2;
  66     ret->up_load = UP_LOAD;
  67     ret->down_load = DOWN_LOAD;
  68     return ret;
  69
  70 err:
  71     OPENSSL_free(ret->b);
  72     OPENSSL_free(ret);
  73     return NULL;
  74 }
  75
  76 void OPENSSL_LH_free(OPENSSL_LHASH *lh)
  77 {
  78     if (lh == NULL)
  79         return;
  80
  81     OPENSSL_LH_flush(lh);
  82     OPENSSL_free(lh->b);
  83     OPENSSL_free(lh);
  84 }
  85
  86 void OPENSSL_LH_flush(OPENSSL_LHASH *lh)
  87 {
  88     unsigned int i;
  89     OPENSSL_LH_NODE *n, *nn;
  90
  91     if (lh == NULL)
  92         return;
  93
  94     for (i = 0; i < lh->num_nodes; i++) {
  95         n = lh->b[i];
  96         while (n != NULL) {
  97             nn = n->next;
  98             OPENSSL_free(n);
  99             n = nn;
 100         }
 101         lh->b[i] = NULL;
 102     }
 103
 104     lh->num_items = 0;
 105 }
 106
 107 void *OPENSSL_LH_insert(OPENSSL_LHASH *lh, void *data)
 108 {
 109     unsigned long hash;
 110     OPENSSL_LH_NODE *nn, **rn;
 111     void *ret;
 112
 113     lh->error = 0;
 114     if ((lh->up_load <= (lh->num_items * LH_LOAD_MULT / lh->num_nodes)) && !expand(lh))
 115         return NULL;        /* 'lh->error++' already done in 'expand' */
 116
 117     rn = getrn(lh, data, &hash);
 118
 119     if (*rn == NULL) {
 120         if ((nn = OPENSSL_malloc(sizeof(*nn))) == NULL) {
 121             lh->error++;
 122             return NULL;
 123         }
 124         nn->data = data;
 125         nn->next = NULL;
 126         nn->hash = hash;
 127         *rn = nn;
 128         ret = NULL;
 129         lh->num_items++;
 130     } else {                    /* replace same key */
 131         ret = (*rn)->data;
 132         (*rn)->data = data;
 133     }
 134     return ret;
 135 }
 136
 137 void *OPENSSL_LH_delete(OPENSSL_LHASH *lh, const void *data)
 138 {
 139     unsigned long hash;
 140     OPENSSL_LH_NODE *nn, **rn;
 141     void *ret;
 142
 143     lh->error = 0;
 144     rn = getrn(lh, data, &hash);
 145
 146     if (*rn == NULL) {
 147         return NULL;
 148     } else {
 149         nn = *rn;
 150         *rn = nn->next;
 151         ret = nn->data;
 152         OPENSSL_free(nn);
 153     }
 154
 155     lh->num_items--;
 156     if ((lh->num_nodes > MIN_NODES) &&
 157         (lh->down_load >= (lh->num_items * LH_LOAD_MULT / lh->num_nodes)))
 158         contract(lh);
 159
 160     return ret;
 161 }
 162
 163 void *OPENSSL_LH_retrieve(OPENSSL_LHASH *lh, const void *data)
 164 {
 165     unsigned long hash;
 166     OPENSSL_LH_NODE **rn;
 167
 168     if (lh->error != 0)
 169         lh->error = 0;
 170
 171     rn = getrn(lh, data, &hash);
 172
 173     return *rn == NULL ? NULL : (*rn)->data;
 174 }
 175
 176 static void doall_util_fn(OPENSSL_LHASH *lh, int use_arg,
 177                           OPENSSL_LH_DOALL_FUNC func,
 178                           OPENSSL_LH_DOALL_FUNCARG func_arg, void *arg)
 179 {
 180     int i;
 181     OPENSSL_LH_NODE *a, *n;
 182
 183     if (lh == NULL)
 184         return;
 185
 186     /*
 187      * reverse the order so we search from 'top to bottom' We were having
 188      * memory leaks otherwise
 189      */
 190     for (i = lh->num_nodes - 1; i >= 0; i--) {
 191         a = lh->b[i];
 192         while (a != NULL) {
 193             n = a->next;
 194             if (use_arg)
 195                 func_arg(a->data, arg);
 196             else
 197                 func(a->data);
 198             a = n;
 199         }
 200     }
 201 }
 202
 203 void OPENSSL_LH_doall(OPENSSL_LHASH *lh, OPENSSL_LH_DOALL_FUNC func)
 204 {
 205     doall_util_fn(lh, 0, func, (OPENSSL_LH_DOALL_FUNCARG)0, NULL);
 206 }
 207
 208 void OPENSSL_LH_doall_arg(OPENSSL_LHASH *lh, OPENSSL_LH_DOALL_FUNCARG func, void *arg)
 209 {
 210     doall_util_fn(lh, 1, (OPENSSL_LH_DOALL_FUNC)0, func, arg);
 211 }
 212
 213 static int expand(OPENSSL_LHASH *lh)
 214 {
 215     OPENSSL_LH_NODE **n, **n1, **n2, *np;
 216     unsigned int p, pmax, nni, j;
 217     unsigned long hash;
 218
 219     nni = lh->num_alloc_nodes;
 220     p = lh->p;
 221     pmax = lh->pmax;
 222     if (p + 1 >= pmax) {
 223         j = nni * 2;
 224         n = OPENSSL_realloc(lh->b, sizeof(OPENSSL_LH_NODE *) * j);
 225         if (n == NULL) {
 226             lh->error++;
 227             return 0;
 228         }
 229         lh->b = n;
 230         memset(n + nni, 0, sizeof(*n) * (j - nni));
 231         lh->pmax = nni;
 232         lh->num_alloc_nodes = j;
 233         lh->p = 0;
 234     } else {
 235         lh->p++;
 236     }
 237
 238     lh->num_nodes++;
 239     n1 = &(lh->b[p]);
 240     n2 = &(lh->b[p + pmax]);
 241     *n2 = NULL;
 242
 243     for (np = *n1; np != NULL;) {
 244         hash = np->hash;
 245         if ((hash % nni) != p) { /* move it */
 246             *n1 = (*n1)->next;
 247             np->next = *n2;
 248             *n2 = np;
 249         } else
 250             n1 = &((*n1)->next);
 251         np = *n1;
 252     }
 253
 254     return 1;
 255 }
 256
 257 static void contract(OPENSSL_LHASH *lh)
 258 {
 259     OPENSSL_LH_NODE **n, *n1, *np;
 260
 261     np = lh->b[lh->p + lh->pmax - 1];
 262     lh->b[lh->p + lh->pmax - 1] = NULL; /* 24/07-92 - eay - weird but :-( */
 263     if (lh->p == 0) {
 264         n = OPENSSL_realloc(lh->b,
 265                             (unsigned int)(sizeof(OPENSSL_LH_NODE *) * lh->pmax));
 266         if (n == NULL) {
 267             /* fputs("realloc error in lhash", stderr); */
 268             lh->error++;
 269             return;
 270         }
 271         lh->num_alloc_nodes /= 2;
 272         lh->pmax /= 2;
 273         lh->p = lh->pmax - 1;
 274         lh->b = n;
 275     } else
 276         lh->p--;
 277
 278     lh->num_nodes--;
 279
 280     n1 = lh->b[(int)lh->p];
 281     if (n1 == NULL)
 282         lh->b[(int)lh->p] = np;
 283     else {
 284         while (n1->next != NULL)
 285             n1 = n1->next;
 286         n1->next = np;
 287     }
 288 }
 289
 290 static OPENSSL_LH_NODE **getrn(OPENSSL_LHASH *lh,
 291                                const void *data, unsigned long *rhash)
 292 {
 293     OPENSSL_LH_NODE **ret, *n1;
 294     unsigned long hash, nn;
 295     OPENSSL_LH_COMPFUNC cf;
 296
 297     hash = (*(lh->hash)) (data);
 298     *rhash = hash;
 299
 300     nn = hash % lh->pmax;
 301     if (nn < lh->p)
 302         nn = hash % lh->num_alloc_nodes;
 303
 304     cf = lh->comp;
 305     ret = &(lh->b[(int)nn]);
 306     for (n1 = *ret; n1 != NULL; n1 = n1->next) {
 307         if (n1->hash != hash) {
 308             ret = &(n1->next);
 309             continue;
 310         }
 311         if (cf(n1->data, data) == 0)
 312             break;
 313         ret = &(n1->next);
 314     }
 315     return ret;
 316 }
 317
 318 /*
 319  * The following hash seems to work very well on normal text strings no
 320  * collisions on /usr/dict/words and it distributes on %2^n quite well, not
 321  * as good as MD5, but still good.
 322  */
 323 unsigned long OPENSSL_LH_strhash(const char *c)
 324 {
 325     unsigned long ret = 0;
 326     long n;
 327     unsigned long v;
 328     int r;
 329
 330     if ((c == NULL) || (*c == '\0'))
 331         return ret;
 332
 333     n = 0x100;
 334     while (*c) {
 335         v = n | (*c);
 336         n += 0x100;
 337         r = (int)((v >> 2) ^ v) & 0x0f;
 338         /* cast to uint64_t to avoid 32 bit shift of 32 bit value */
 339         ret = (ret << r) | (unsigned long)((uint64_t)ret >> (32 - r));
 340         ret &= 0xFFFFFFFFL;
 341         ret ^= v * v;
 342         c++;
 343     }
 344     return (ret >> 16) ^ ret;
 345 }
 346
 347 unsigned long ossl_lh_strcasehash(const char *c)
 348 {
 349     unsigned long ret = 0;
 350     long n;
 351     unsigned long v;
 352     int r;
 353
 354     if (c == NULL || *c == '\0')
 355         return ret;
 356
 357     for (n = 0x100; *c != '\0'; n += 0x100) {
 358         v = n | ossl_tolower(*c);
 359         r = (int)((v >> 2) ^ v) & 0x0f;
 360         /* cast to uint64_t to avoid 32 bit shift of 32 bit value */
 361         ret = (ret << r) | (unsigned long)((uint64_t)ret >> (32 - r));
 362         ret &= 0xFFFFFFFFL;
 363         ret ^= v * v;
 364         c++;
 365     }
 366     return (ret >> 16) ^ ret;
 367 }
 368
 369 unsigned long OPENSSL_LH_num_items(const OPENSSL_LHASH *lh)
 370 {
 371     return lh ? lh->num_items : 0;
 372 }
 373
 374 unsigned long OPENSSL_LH_get_down_load(const OPENSSL_LHASH *lh)
 375 {
 376     return lh->down_load;
 377 }
 378
 379 void OPENSSL_LH_set_down_load(OPENSSL_LHASH *lh, unsigned long down_load)
 380 {
 381     lh->down_load = down_load;
 382 }
 383
 384 int OPENSSL_LH_error(OPENSSL_LHASH *lh)
 385 {
 386     return lh->error;
 387 }