1 /* Hash table implementation.
2  *
3  * This file implements in memory hash tables with insert/del/replace/find/
4  * get-random-element operations. Hash tables will auto resize if needed
5  * tables of power of two in size are used, collisions are handled by
6  * chaining. See the source code for more information... :)
7  *
8  * Copyright (c) 2006-2010, Salvatore Sanfilippo <antirez at gmail dot com>
9  * All rights reserved.
10  *
11  * Redistribution and use in source and binary forms, with or without
12  * modification, are permitted provided that the following conditions are met:
13  *
14  *   * Redistributions of source code must retain the above copyright notice,
15  *     this list of conditions and the following disclaimer.
16  *   * Redistributions in binary form must reproduce the above copyright
17  *     notice, this list of conditions and the following disclaimer in the
18  *     documentation and/or other materials provided with the distribution.
19  *   * Neither the name of Redis nor the names of its contributors may be used
20  *     to endorse or promote products derived from this software without
21  *     specific prior written permission.
22  *
23  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
24  * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
25  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
26  * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
27  * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
28  * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
29  * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
30  * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
31  * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
32  * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
33  * POSSIBILITY OF SUCH DAMAGE.
34  */
35 
36 #include "fmacros.h"
37 #include <stdlib.h>
38 #include <assert.h>
39 #include <limits.h>
40 #include "dict.h"
41 
42 /* -------------------------- private prototypes ---------------------------- */
43 
44 static int _dictExpandIfNeeded(dict *ht);
45 static unsigned long _dictNextPower(unsigned long size);
46 static int _dictKeyIndex(dict *ht, const void *key);
47 static int _dictInit(dict *ht, dictType *type, void *privDataPtr);
48 
49 /* -------------------------- hash functions -------------------------------- */
50 
51 /* Generic hash function (a popular one from Bernstein).
52  * I tested a few and this was the best. */
dictGenHashFunction(const unsigned char * buf,int len)53 static unsigned int dictGenHashFunction(const unsigned char *buf, int len) {
54     unsigned int hash = 5381;
55 
56     while (len--)
57         hash = ((hash << 5) + hash) + (*buf++); /* hash * 33 + c */
58     return hash;
59 }
60 
61 /* ----------------------------- API implementation ------------------------- */
62 
63 /* Reset an hashtable already initialized with ht_init().
64  * NOTE: This function should only called by ht_destroy(). */
_dictReset(dict * ht)65 static void _dictReset(dict *ht) {
66     ht->table = NULL;
67     ht->size = 0;
68     ht->sizemask = 0;
69     ht->used = 0;
70 }
71 
72 /* Create a new hash table */
dictCreate(dictType * type,void * privDataPtr)73 static dict *dictCreate(dictType *type, void *privDataPtr) {
74     dict *ht = malloc(sizeof(*ht));
75     _dictInit(ht,type,privDataPtr);
76     return ht;
77 }
78 
79 /* Initialize the hash table */
_dictInit(dict * ht,dictType * type,void * privDataPtr)80 static int _dictInit(dict *ht, dictType *type, void *privDataPtr) {
81     _dictReset(ht);
82     ht->type = type;
83     ht->privdata = privDataPtr;
84     return DICT_OK;
85 }
86 
87 /* Expand or create the hashtable */
dictExpand(dict * ht,unsigned long size)88 static int dictExpand(dict *ht, unsigned long size) {
89     dict n; /* the new hashtable */
90     unsigned long realsize = _dictNextPower(size), i;
91 
92     /* the size is invalid if it is smaller than the number of
93      * elements already inside the hashtable */
94     if (ht->used > size)
95         return DICT_ERR;
96 
97     _dictInit(&n, ht->type, ht->privdata);
98     n.size = realsize;
99     n.sizemask = realsize-1;
100     n.table = calloc(realsize,sizeof(dictEntry*));
101 
102     /* Copy all the elements from the old to the new table:
103      * note that if the old hash table is empty ht->size is zero,
104      * so dictExpand just creates an hash table. */
105     n.used = ht->used;
106     for (i = 0; i < ht->size && ht->used > 0; i++) {
107         dictEntry *he, *nextHe;
108 
109         if (ht->table[i] == NULL) continue;
110 
111         /* For each hash entry on this slot... */
112         he = ht->table[i];
113         while(he) {
114             unsigned int h;
115 
116             nextHe = he->next;
117             /* Get the new element index */
118             h = dictHashKey(ht, he->key) & n.sizemask;
119             he->next = n.table[h];
120             n.table[h] = he;
121             ht->used--;
122             /* Pass to the next element */
123             he = nextHe;
124         }
125     }
126     assert(ht->used == 0);
127     free(ht->table);
128 
129     /* Remap the new hashtable in the old */
130     *ht = n;
131     return DICT_OK;
132 }
133 
134 /* Add an element to the target hash table */
dictAdd(dict * ht,void * key,void * val)135 static int dictAdd(dict *ht, void *key, void *val) {
136     int index;
137     dictEntry *entry;
138 
139     /* Get the index of the new element, or -1 if
140      * the element already exists. */
141     if ((index = _dictKeyIndex(ht, key)) == -1)
142         return DICT_ERR;
143 
144     /* Allocates the memory and stores key */
145     entry = malloc(sizeof(*entry));
146     entry->next = ht->table[index];
147     ht->table[index] = entry;
148 
149     /* Set the hash entry fields. */
150     dictSetHashKey(ht, entry, key);
151     dictSetHashVal(ht, entry, val);
152     ht->used++;
153     return DICT_OK;
154 }
155 
156 /* Add an element, discarding the old if the key already exists.
157  * Return 1 if the key was added from scratch, 0 if there was already an
158  * element with such key and dictReplace() just performed a value update
159  * operation. */
dictReplace(dict * ht,void * key,void * val)160 static int dictReplace(dict *ht, void *key, void *val) {
161     dictEntry *entry, auxentry;
162 
163     /* Try to add the element. If the key
164      * does not exists dictAdd will succeed. */
165     if (dictAdd(ht, key, val) == DICT_OK)
166         return 1;
167     /* It already exists, get the entry */
168     entry = dictFind(ht, key);
169     /* Free the old value and set the new one */
170     /* Set the new value and free the old one. Note that it is important
171      * to do that in this order, as the value may just be exactly the same
172      * as the previous one. In this context, think to reference counting,
173      * you want to increment (set), and then decrement (free), and not the
174      * reverse. */
175     auxentry = *entry;
176     dictSetHashVal(ht, entry, val);
177     dictFreeEntryVal(ht, &auxentry);
178     return 0;
179 }
180 
181 /* Search and remove an element */
dictDelete(dict * ht,const void * key)182 static int dictDelete(dict *ht, const void *key) {
183     unsigned int h;
184     dictEntry *de, *prevde;
185 
186     if (ht->size == 0)
187         return DICT_ERR;
188     h = dictHashKey(ht, key) & ht->sizemask;
189     de = ht->table[h];
190 
191     prevde = NULL;
192     while(de) {
193         if (dictCompareHashKeys(ht,key,de->key)) {
194             /* Unlink the element from the list */
195             if (prevde)
196                 prevde->next = de->next;
197             else
198                 ht->table[h] = de->next;
199 
200             dictFreeEntryKey(ht,de);
201             dictFreeEntryVal(ht,de);
202             free(de);
203             ht->used--;
204             return DICT_OK;
205         }
206         prevde = de;
207         de = de->next;
208     }
209     return DICT_ERR; /* not found */
210 }
211 
212 /* Destroy an entire hash table */
_dictClear(dict * ht)213 static int _dictClear(dict *ht) {
214     unsigned long i;
215 
216     /* Free all the elements */
217     for (i = 0; i < ht->size && ht->used > 0; i++) {
218         dictEntry *he, *nextHe;
219 
220         if ((he = ht->table[i]) == NULL) continue;
221         while(he) {
222             nextHe = he->next;
223             dictFreeEntryKey(ht, he);
224             dictFreeEntryVal(ht, he);
225             free(he);
226             ht->used--;
227             he = nextHe;
228         }
229     }
230     /* Free the table and the allocated cache structure */
231     free(ht->table);
232     /* Re-initialize the table */
233     _dictReset(ht);
234     return DICT_OK; /* never fails */
235 }
236 
237 /* Clear & Release the hash table */
dictRelease(dict * ht)238 static void dictRelease(dict *ht) {
239     _dictClear(ht);
240     free(ht);
241 }
242 
dictFind(dict * ht,const void * key)243 static dictEntry *dictFind(dict *ht, const void *key) {
244     dictEntry *he;
245     unsigned int h;
246 
247     if (ht->size == 0) return NULL;
248     h = dictHashKey(ht, key) & ht->sizemask;
249     he = ht->table[h];
250     while(he) {
251         if (dictCompareHashKeys(ht, key, he->key))
252             return he;
253         he = he->next;
254     }
255     return NULL;
256 }
257 
dictGetIterator(dict * ht)258 static dictIterator *dictGetIterator(dict *ht) {
259     dictIterator *iter = malloc(sizeof(*iter));
260 
261     iter->ht = ht;
262     iter->index = -1;
263     iter->entry = NULL;
264     iter->nextEntry = NULL;
265     return iter;
266 }
267 
dictNext(dictIterator * iter)268 static dictEntry *dictNext(dictIterator *iter) {
269     while (1) {
270         if (iter->entry == NULL) {
271             iter->index++;
272             if (iter->index >=
273                     (signed)iter->ht->size) break;
274             iter->entry = iter->ht->table[iter->index];
275         } else {
276             iter->entry = iter->nextEntry;
277         }
278         if (iter->entry) {
279             /* We need to save the 'next' here, the iterator user
280              * may delete the entry we are returning. */
281             iter->nextEntry = iter->entry->next;
282             return iter->entry;
283         }
284     }
285     return NULL;
286 }
287 
dictReleaseIterator(dictIterator * iter)288 static void dictReleaseIterator(dictIterator *iter) {
289     free(iter);
290 }
291 
292 /* ------------------------- private functions ------------------------------ */
293 
294 /* Expand the hash table if needed */
_dictExpandIfNeeded(dict * ht)295 static int _dictExpandIfNeeded(dict *ht) {
296     /* If the hash table is empty expand it to the initial size,
297      * if the table is "full" dobule its size. */
298     if (ht->size == 0)
299         return dictExpand(ht, DICT_HT_INITIAL_SIZE);
300     if (ht->used == ht->size)
301         return dictExpand(ht, ht->size*2);
302     return DICT_OK;
303 }
304 
305 /* Our hash table capability is a power of two */
_dictNextPower(unsigned long size)306 static unsigned long _dictNextPower(unsigned long size) {
307     unsigned long i = DICT_HT_INITIAL_SIZE;
308 
309     if (size >= LONG_MAX) return LONG_MAX;
310     while(1) {
311         if (i >= size)
312             return i;
313         i *= 2;
314     }
315 }
316 
317 /* Returns the index of a free slot that can be populated with
318  * an hash entry for the given 'key'.
319  * If the key already exists, -1 is returned. */
_dictKeyIndex(dict * ht,const void * key)320 static int _dictKeyIndex(dict *ht, const void *key) {
321     unsigned int h;
322     dictEntry *he;
323 
324     /* Expand the hashtable if needed */
325     if (_dictExpandIfNeeded(ht) == DICT_ERR)
326         return -1;
327     /* Compute the key hash value */
328     h = dictHashKey(ht, key) & ht->sizemask;
329     /* Search if this slot does not already contain the given key */
330     he = ht->table[h];
331     while(he) {
332         if (dictCompareHashKeys(ht, key, he->key))
333             return -1;
334         he = he->next;
335     }
336     return h;
337 }
338 
339