lib/hash.cc

   1
   2 /*
   3  * $Id$
   4  *
   5  * DEBUG: section 00    Hash Tables
   6  * AUTHOR: Harvest Derived
   7  *
   8  * SQUID Web Proxy Cache          http://www.squid-cache.org/
   9  * ----------------------------------------------------------
  10  *
  11  *  Squid is the result of efforts by numerous individuals from
  12  *  the Internet community; see the CONTRIBUTORS file for full
  13  *  details.   Many organizations have provided support for Squid's
  14  *  development; see the SPONSORS file for full details.  Squid is
  15  *  Copyrighted (C) 2001 by the Regents of the University of
  16  *  California; see the COPYRIGHT file for full details.  Squid
  17  *  incorporates software developed and/or copyrighted by other
  18  *  sources; see the CREDITS file for full details.
  19  *
  20  *  This program is free software; you can redistribute it and/or modify
  21  *  it under the terms of the GNU General Public License as published by
  22  *  the Free Software Foundation; either version 2 of the License, or
  23  *  (at your option) any later version.
  24  *
  25  *  This program is distributed in the hope that it will be useful,
  26  *  but WITHOUT ANY WARRANTY; without even the implied warranty of
  27  *  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  28  *  GNU General Public License for more details.
  29  *
  30  *  You should have received a copy of the GNU General Public License
  31  *  along with this program; if not, write to the Free Software
  32  *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.
  33  *
  34  */
  35
  36 #include "squid.h"
  37 #include "hash.h"
  38 #include "profiler/Profiler.h"
  39
  40 #if HAVE_STDIO_H
  41 #include <stdio.h>
  42 #endif
  43 #if HAVE_STDLIB_H
  44 #include <stdlib.h>
  45 #endif
  46 #if HAVE_STRING_H
  47 #include <string.h>
  48 #endif
  49 #if HAVE_UNISTD_H
  50 #include <unistd.h>
  51 #endif
  52 #if HAVE_GNUMALLLOC_H
  53 #include <gnumalloc.h>
  54 #elif HAVE_MALLOC_H
  55 #include <malloc.h>
  56 #endif
  57 #if HAVE_ASSERT_H
  58 #include <assert.h>
  59 #endif
  60 #if HAVE_MATH_H
  61 #include <math.h>
  62 #endif
  63
  64 static void hash_next_bucket(hash_table * hid);
  65
  66 unsigned int
  67 hash_string(const void *data, unsigned int size)
  68 {
  69     const unsigned char *s = static_cast<const unsigned char *>(data);
  70     unsigned int n = 0;
  71     unsigned int j = 0;
  72     unsigned int i = 0;
  73     while (*s) {
  74         ++j;
  75         n ^= 271 * (*s++);
  76     }
  77     i = n ^ (j * 271);
  78     return i % size;
  79 }
  80
  81 /* the following function(s) were adapted from
  82  *    usr/src/lib/libc/db/hash_func.c, 4.4 BSD lite */
  83
  84 /* Hash function from Chris Torek. */
  85 unsigned int
  86 hash4(const void *data, unsigned int size)
  87 {
  88     const char *key = static_cast<const char *>(data);
  89     size_t loop;
  90     unsigned int h;
  91     size_t len;
  92
  93 #define HASH4a   h = (h << 5) - h + *key++;
  94 #define HASH4b   h = (h << 5) + h + *key++;
  95 #define HASH4 HASH4b
  96
  97     h = 0;
  98     len = strlen(key);
  99     loop = len >> 3;
 100     switch (len & (8 - 1)) {
 101     case 0:
 102         break;
 103     case 7:
 104         HASH4;
 105         /* FALLTHROUGH */
 106     case 6:
 107         HASH4;
 108         /* FALLTHROUGH */
 109     case 5:
 110         HASH4;
 111         /* FALLTHROUGH */
 112     case 4:
 113         HASH4;
 114         /* FALLTHROUGH */
 115     case 3:
 116         HASH4;
 117         /* FALLTHROUGH */
 118     case 2:
 119         HASH4;
 120         /* FALLTHROUGH */
 121     case 1:
 122         HASH4;
 123     }
 124     while (loop--) {
 125         HASH4;
 126         HASH4;
 127         HASH4;
 128         HASH4;
 129         HASH4;
 130         HASH4;
 131         HASH4;
 132         HASH4;
 133     }
 134     return h % size;
 135 }
 136
 137 /**
 138  *  hash_create - creates a new hash table, uses the cmp_func
 139  *  to compare keys.  Returns the identification for the hash table;
 140  *  otherwise returns a negative number on error.
 141  */
 142 hash_table *
 143 hash_create(HASHCMP * cmp_func, int hash_sz, HASHHASH * hash_func)
 144 {
 145     hash_table *hid = (hash_table *)xcalloc(1, sizeof(hash_table));
 146     if (!hash_sz)
 147         hid->size = (unsigned int) DEFAULT_HASH_SIZE;
 148     else
 149         hid->size = (unsigned int) hash_sz;
 150     /* allocate and null the buckets */
 151     hid->buckets = (hash_link **)xcalloc(hid->size, sizeof(hash_link *));
 152     hid->cmp = cmp_func;
 153     hid->hash = hash_func;
 154     hid->next = NULL;
 155     hid->current_slot = 0;
 156     return hid;
 157 }
 158
 159 /**
 160  *  hash_join - joins a hash_link under its key lnk->key
 161  *  into the hash table 'hid'.
 162  *
 163  *  It does not copy any data into the hash table, only links pointers.
 164  */
 165 void
 166 hash_join(hash_table * hid, hash_link * lnk)
 167 {
 168     int i;
 169     i = hid->hash(lnk->key, hid->size);
 170     lnk->next = hid->buckets[i];
 171     hid->buckets[i] = lnk;
 172     ++hid->count;
 173 }
 174
 175 /**
 176  *  hash_lookup - locates the item under the key 'k' in the hash table
 177  *  'hid'.  Returns a pointer to the hash bucket on success; otherwise
 178  *  returns NULL.
 179  */
 180 hash_link *
 181 hash_lookup(hash_table * hid, const void *k)
 182 {
 183     int b;
 184     PROF_start(hash_lookup);
 185     assert(k != NULL);
 186     b = hid->hash(k, hid->size);
 187     for (hash_link *walker = hid->buckets[b]; walker != NULL; walker = walker->next) {
 188         if ((hid->cmp) (k, walker->key) == 0) {
 189             PROF_stop(hash_lookup);
 190             return (walker);
 191         }
 192         assert(walker != walker->next);
 193     }
 194     PROF_stop(hash_lookup);
 195     return NULL;
 196 }
 197
 198 static void
 199 hash_next_bucket(hash_table * hid)
 200 {
 201     while (hid->next == NULL && ++hid->current_slot < hid->size)
 202         hid->next = hid->buckets[hid->current_slot];
 203 }
 204
 205 /**
 206  *  hash_first - initializes the hash table for the hash_next()
 207  *  function.
 208  */
 209 void
 210 hash_first(hash_table * hid)
 211 {
 212     assert(NULL == hid->next);
 213     hid->current_slot = 0;
 214     hid->next = hid->buckets[hid->current_slot];
 215     if (NULL == hid->next)
 216         hash_next_bucket(hid);
 217 }
 218
 219 /**
 220  *  hash_next - returns the next item in the hash table 'hid'.
 221  *  Otherwise, returns NULL on error or end of list.
 222  *
 223  *  MUST call hash_first() before hash_next().
 224  */
 225 hash_link *
 226 hash_next(hash_table * hid)
 227 {
 228     hash_link *p = hid->next;
 229     if (NULL == p)
 230         return NULL;
 231     hid->next = p->next;
 232     if (NULL == hid->next)
 233         hash_next_bucket(hid);
 234     return p;
 235 }
 236
 237 /**
 238  *  hash_last - resets hash traversal state to NULL
 239  *
 240  */
 241 void
 242 hash_last(hash_table * hid)
 243 {
 244     assert(hid != NULL);
 245     hid->next = NULL;
 246     hid->current_slot = 0;
 247 }
 248
 249 /**
 250  *  hash_remove_link - deletes the given hash_link node from the
 251  *  hash table 'hid'.  Does not free the item, only removes it
 252  *  from the list.
 253  *
 254  *  An assertion is triggered if the hash_link is not found in the
 255  *  list.
 256  */
 257 void
 258 hash_remove_link(hash_table * hid, hash_link * hl)
 259 {
 260     assert(hl != NULL);
 261     int i = hid->hash(hl->key, hid->size);
 262     for (hash_link **P = &hid->buckets[i]; *P; P = &(*P)->next) {
 263         if (*P != hl)
 264             continue;
 265         *P = hl->next;
 266         if (hid->next == hl) {
 267             hid->next = hl->next;
 268             if (NULL == hid->next)
 269                 hash_next_bucket(hid);
 270         }
 271         --hid->count;
 272         return;
 273     }
 274     assert(0);
 275 }
 276
 277 /**
 278  *  hash_get_bucket - returns the head item of the bucket
 279  *  in the hash table 'hid'. Otherwise, returns NULL on error.
 280  */
 281 hash_link *
 282 hash_get_bucket(hash_table * hid, unsigned int bucket)
 283 {
 284     if (bucket >= hid->size)
 285         return NULL;
 286     return (hid->buckets[bucket]);
 287 }
 288
 289 void
 290 hashFreeItems(hash_table * hid, HASHFREE * free_func)
 291 {
 292     hash_link *l;
 293     int i = 0;
 294     hash_link **list = (hash_link **)xcalloc(hid->count, sizeof(hash_link *));
 295     hash_first(hid);
 296     while ((l = hash_next(hid)) && i < hid->count) {
 297         *(list + i) = l;
 298         ++i;
 299     }
 300     for (int j = 0; j < i; ++j)
 301         free_func(*(list + j));
 302     xfree(list);
 303 }
 304
 305 void
 306 hashFreeMemory(hash_table * hid)
 307 {
 308     if (hid == NULL)
 309         return;
 310     if (hid->buckets)
 311         xfree(hid->buckets);
 312     xfree(hid);
 313 }
 314
 315 static int hash_primes[] = {
 316     103,
 317     229,
 318     467,
 319     977,
 320     1979,
 321     4019,
 322     6037,
 323     7951,
 324     12149,
 325     16231,
 326     33493,
 327     65357
 328 };
 329
 330 int
 331 hashPrime(int n)
 332 {
 333     int I = sizeof(hash_primes) / sizeof(int);
 334     int best_prime = hash_primes[0];
 335     double min = fabs(log((double) n) - log((double) hash_primes[0]));
 336     double d;
 337     for (int i = 0; i < I; ++i) {
 338         d = fabs(log((double) n) - log((double) hash_primes[i]));
 339         if (d > min)
 340             continue;
 341         min = d;
 342         best_prime = hash_primes[i];
 343     }
 344     return best_prime;
 345 }
 346
 347 /**
 348  * return the key of a hash_link as a const string
 349  */
 350 const char *
 351 hashKeyStr(hash_link * hl)
 352 {
 353     return (const char *) hl->key;
 354 }
 355
 356
 357 #if USE_HASH_DRIVER
 358 /**
 359  *  hash-driver - Run with a big file as stdin to insert each line into the
 360  *  hash table, then prints the whole hash table, then deletes a random item,
 361  *  and prints the table again...
 362  */
 363 int
 364 main(void)
 365 {
 366     hash_table *hid;
 367     LOCAL_ARRAY(char, buf, BUFSIZ);
 368     LOCAL_ARRAY(char, todelete, BUFSIZ);
 369     hash_link *walker = NULL;
 370
 371     todelete[0] = '\0';
 372     printf("init\n");
 373
 374     printf("creating hash table\n");
 375     if ((hid = hash_create((HASHCMP *) strcmp, 229, hash4)) < 0) {
 376         printf("hash_create error.\n");
 377         exit(1);
 378     }
 379     printf("done creating hash table: %d\n", hid);
 380
 381     while (fgets(buf, BUFSIZ, stdin)) {
 382         buf[strlen(buf) - 1] = '\0';
 383         printf("Inserting '%s' for item %p to hash table: %d\n",
 384                buf, buf, hid);
 385         hash_insert(hid, xstrdup(buf), (void *) 0x12345678);
 386         if (random() % 17 == 0)
 387             strcpy(todelete, buf);
 388     }
 389
 390     printf("walking hash table...\n");
 391     for (int i = 0, walker = hash_first(hid); walker; walker = hash_next(hid)) {
 392         printf("item %5d: key: '%s' item: %p\n", i++, walker->key,
 393                walker->item);
 394     }
 395     printf("done walking hash table...\n");
 396
 397     if (todelete[0]) {
 398         printf("deleting %s from %d\n", todelete, hid);
 399         if (hash_delete(hid, todelete))
 400             printf("hash_delete error\n");
 401     }
 402     printf("walking hash table...\n");
 403     for (int i = 0, walker = hash_first(hid); walker; walker = hash_next(hid)) {
 404         printf("item %5d: key: '%s' item: %p\n", i++, walker->key,
 405                walker->item);
 406     }
 407     printf("done walking hash table...\n");
 408
 409
 410     printf("driver finished.\n");
 411     exit(0);
 412 }
 413 #endif