]> git.proxmox.com Git - mirror_frr.git/blob - lib/hash.h
*: change hash_backet to hash_bucket
[mirror_frr.git] / lib / hash.h
1 /* Hash routine.
2 * Copyright (C) 1998 Kunihiro Ishiguro
3 *
4 * This file is part of GNU Zebra.
5 *
6 * GNU Zebra is free software; you can redistribute it and/or modify
7 * it under the terms of the GNU General Public License as published
8 * by the Free Software Foundation; either version 2, or (at your
9 * option) any later version.
10 *
11 * GNU Zebra is distributed in the hope that it will be useful, but
12 * WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 * General Public License for more details.
15 *
16 * You should have received a copy of the GNU General Public License along
17 * with this program; see the file COPYING; if not, write to the Free Software
18 * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
19 */
20
21 #ifndef _ZEBRA_HASH_H
22 #define _ZEBRA_HASH_H
23
24 #include "memory.h"
25 #include "frratomic.h"
26
27 #ifdef __cplusplus
28 extern "C" {
29 #endif
30
31 /* Default hash table size. */
32 #define HASH_INITIAL_SIZE 256
33 /* Expansion threshold */
34 #define HASH_THRESHOLD(used, size) ((used) > (size))
35
36 #define HASHWALK_CONTINUE 0
37 #define HASHWALK_ABORT -1
38
39 struct hash_bucket {
40 /*
41 * if this bucket is the head of the linked listed, len denotes the
42 * number of elements in the list
43 */
44 int len;
45
46 /* Linked list. */
47 struct hash_bucket *next;
48
49 /* Hash key. */
50 unsigned int key;
51
52 /* Data. */
53 void *data;
54 };
55
56 struct hashstats {
57 /* number of empty hash buckets */
58 atomic_uint_fast32_t empty;
59 /* sum of squares of bucket length */
60 atomic_uint_fast32_t ssq;
61 };
62
63 struct hash {
64 /* Hash bucket. */
65 struct hash_bucket **index;
66
67 /* Hash table size. Must be power of 2 */
68 unsigned int size;
69
70 /* If max_size is 0 there is no limit */
71 unsigned int max_size;
72
73 /* Key make function. */
74 unsigned int (*hash_key)(const void *);
75
76 /* Data compare function. */
77 bool (*hash_cmp)(const void *, const void *);
78
79 /* Backet alloc. */
80 unsigned long count;
81
82 struct hashstats stats;
83
84 /* hash name */
85 char *name;
86 };
87
88 #define hashcount(X) ((X)->count)
89
90 /*
91 * Create a hash table.
92 *
93 * The created hash table uses chaining and a user-provided comparator function
94 * to resolve collisions. For best performance use a perfect hash function.
95 * Worst case lookup time is O(N) when using a constant hash function. Best
96 * case lookup time is O(1) when using a perfect hash function.
97 *
98 * The initial size of the created hash table is HASH_INITIAL_SIZE.
99 *
100 * hash_key
101 * hash function to use; should return a unique unsigned integer when called
102 * with a data item. Collisions are acceptable.
103 *
104 * hash_cmp
105 * comparison function used for resolving collisions; when called with two
106 * data items, should return nonzero if the two items are equal and 0
107 * otherwise
108 *
109 * name
110 * optional name for the hashtable; this is used when displaying global
111 * hashtable statistics. If this parameter is NULL the hash's name will be
112 * set to NULL and the default name will be displayed when showing
113 * statistics.
114 *
115 * Returns:
116 * a new hash table
117 */
118 extern struct hash *hash_create(unsigned int (*hash_key)(const void *),
119 bool (*hash_cmp)(const void *, const void *),
120 const char *name);
121
122 /*
123 * Create a hash table.
124 *
125 * The created hash table uses chaining and a user-provided comparator function
126 * to resolve collisions. For best performance use a perfect hash function.
127 * Worst case lookup time is O(N) when using a constant hash function. Best
128 * case lookup time is O(1) when using a perfect hash function.
129 *
130 * size
131 * initial number of hash buckets to allocate; must be a power of 2 or the
132 * program will assert
133 *
134 * hash_key
135 * hash function to use; should return a unique unsigned integer when called
136 * with a data item. Collisions are acceptable.
137 *
138 * hash_cmp
139 * comparison function used for resolving collisions; when called with two
140 * data items, should return nonzero if the two items are equal and 0
141 * otherwise
142 *
143 * name
144 * optional name for the hashtable; this is used when displaying global
145 * hashtable statistics. If this parameter is NULL the hash's name will be
146 * set to NULL and the default name will be displayed when showing
147 * statistics.
148 *
149 * Returns:
150 * a new hash table
151 */
152 extern struct hash *
153 hash_create_size(unsigned int size, unsigned int (*hash_key)(const void *),
154 bool (*hash_cmp)(const void *, const void *),
155 const char *name);
156
157 /*
158 * Retrieve or insert data from / into a hash table.
159 *
160 * This function is somewhat counterintuitive in its usage. In order to look up
161 * an element from its key, you must provide the data item itself, with the
162 * portions used in the hash function set to the same values as the data item
163 * to retrieve. To insert a data element, either provide the key as just
164 * described and provide alloc_func as descrbied below to allocate the full
165 * data element, or provide the full data element and pass 'hash_alloc_intern'
166 * to alloc_func.
167 *
168 * hash
169 * hash table to operate on
170 *
171 * data
172 * data to insert or retrieve - A hash bucket will not be created if
173 * the alloc_func returns a NULL pointer and nothing will be added to
174 * the hash. As such bucket->data will always be non-NULL.
175 *
176 * alloc_func
177 * function to call if the item is not found in the hash table. This
178 * function is called with the value of 'data' and should create the data
179 * item to insert and return a pointer to it. If the data has already been
180 * completely created and provided in the 'data' parameter, passing
181 * 'hash_alloc_intern' to this parameter will cause 'data' to be inserted.
182 * If this parameter is NULL, then this call to hash_get is equivalent to
183 * hash_lookup.
184 *
185 * Returns:
186 * the data item found or inserted, or NULL if alloc_func is NULL and the
187 * data is not found
188 */
189 extern void *hash_get(struct hash *hash, void *data,
190 void *(*alloc_func)(void *));
191
192 /*
193 * Dummy element allocation function.
194 *
195 * See hash_get for details.
196 *
197 * data
198 * data to insert into the hash table
199 *
200 * Returns:
201 * data
202 */
203 extern void *hash_alloc_intern(void *data);
204
205 /*
206 * Retrieve an item from a hash table.
207 *
208 * This function is equivalent to calling hash_get with alloc_func set to NULL.
209 *
210 * hash
211 * hash table to operate on
212 *
213 * data
214 * data element with values used for key computation set
215 *
216 * Returns:
217 * the data element if found, or NULL if not found
218 */
219 extern void *hash_lookup(struct hash *hash, void *data);
220
221 /*
222 * Remove an element from a hash table.
223 *
224 * hash
225 * hash table to operate on
226 *
227 * data
228 * data element to remove with values used for key computation set
229 *
230 * Returns:
231 * the removed element if found, or NULL if not found
232 */
233 extern void *hash_release(struct hash *hash, void *data);
234
235 /*
236 * Iterate over the elements in a hash table.
237 *
238 * It is safe to delete items passed to the iteration function from the hash
239 * table during iteration. Please note that adding entries to the hash
240 * during the walk will cause undefined behavior in that some new entries
241 * will be walked and some will not. So do not do this.
242 *
243 * The bucket passed to func will have a non-NULL data pointer.
244 *
245 * hash
246 * hash table to operate on
247 *
248 * func
249 * function to call with each data item
250 *
251 * arg
252 * arbitrary argument passed as the second parameter in each call to 'func'
253 */
254 extern void hash_iterate(struct hash *hash,
255 void (*func)(struct hash_bucket *, void *), void *arg);
256
257 /*
258 * Iterate over the elements in a hash table, stopping on condition.
259 *
260 * It is safe to delete items passed to the iteration function from the hash
261 * table during iteration. Please note that adding entries to the hash
262 * during the walk will cause undefined behavior in that some new entries
263 * will be walked and some will not. So do not do this.
264 *
265 * The bucket passed to func will have a non-NULL data pointer.
266 *
267 * hash
268 * hash table to operate on
269 *
270 * func
271 * function to call with each data item. If this function returns
272 * HASHWALK_ABORT then the iteration stops.
273 *
274 * arg
275 * arbitrary argument passed as the second parameter in each call to 'func'
276 */
277 extern void hash_walk(struct hash *hash,
278 int (*func)(struct hash_bucket *, void *), void *arg);
279
280 /*
281 * Remove all elements from a hash table.
282 *
283 * hash
284 * hash table to operate on
285 *
286 * free_func
287 * function to call with each removed item; intended to free the data
288 */
289 extern void hash_clean(struct hash *hash, void (*free_func)(void *));
290
291 /*
292 * Delete a hash table.
293 *
294 * This function assumes the table is empty. Call hash_clean to delete the
295 * hashtable contents if necessary.
296 *
297 * hash
298 * hash table to delete
299 */
300 extern void hash_free(struct hash *hash);
301
302 /*
303 * Converts a hash table to an unsorted linked list.
304 * Does not modify the hash table in any way.
305 *
306 * hash
307 * hash table to convert
308 */
309 extern struct list *hash_to_list(struct hash *hash);
310
311 /*
312 * Hash a string using the modified Bernstein hash.
313 *
314 * This is not a perfect hash function.
315 *
316 * str
317 * string to hash
318 *
319 * Returns:
320 * modified Bernstein hash of the string
321 */
322 extern unsigned int string_hash_make(const char *);
323
324 /*
325 * Install CLI commands for viewing global hash table statistics.
326 */
327 extern void hash_cmd_init(void);
328
329 #ifdef __cplusplus
330 }
331 #endif
332
333 #endif /* _ZEBRA_HASH_H */