1 /*- 2 * Copyright (c) 2005 Michael Bushkov <bushman@rsu.ru> 3 * All rights reserved. 4 * 5 * Redistribution and use in source and binary forms, with or without 6 * modification, are permitted provided that the following conditions 7 * are met: 8 * 1. Redistributions of source code must retain the above copyright 9 * notice, this list of conditions and the following disclaimer. 10 * 2. Redistributions in binary form must reproduce the above copyright 11 * notice, this list of conditions and the following disclaimer in the 12 * documentation and/or other materials provided with the distribution. 13 * 14 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 15 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 16 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 17 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 18 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 19 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 20 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 21 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 22 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 23 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 24 * SUCH DAMAGE. 25 * 26 * $FreeBSD$ 27 */ 28 29 #ifndef __CACHELIB_HASHTABLE_H__ 30 #define __CACHELIB_HASHTABLE_H__ 31 32 #include <string.h> 33 34 #define HASHTABLE_INITIAL_ENTRIES_CAPACITY 8 35 typedef unsigned int hashtable_index_t; 36 37 /* 38 * This file contains queue.h-like macro definitions for hash tables. 39 * Hash table is organized as an array of the specified size of the user 40 * defined (with HASTABLE_ENTRY_HEAD) structures. Each hash table 41 * entry (user defined structure) stores its elements in the sorted array. 42 * You can place elements into the hash table, retrieve elements with 43 * specified key, traverse through all elements, and delete them. 44 * New elements are placed into the hash table by using the compare and 45 * hashing functions, provided by the user. 46 */ 47 48 /* 49 * Defines the hash table entry structure, that uses specified type of 50 * elements. 51 */ 52 #define HASHTABLE_ENTRY_HEAD(name, type) struct name { \ 53 type *values; \ 54 size_t capacity; \ 55 size_t size; \ 56 } 57 58 /* 59 * Defines the hash table structure, which uses the specified type of entries. 60 * The only restriction for entries is that is that they should have the field, 61 * defined with HASHTABLE_ENTRY_HEAD macro. 62 */ 63 #define HASHTABLE_HEAD(name, entry) struct name { \ 64 struct entry *entries; \ 65 size_t entries_size; \ 66 } 67 68 #define HASHTABLE_ENTRIES_COUNT(table) \ 69 ((table)->entries_size) 70 71 /* 72 * Unlike most of queue.h data types, hash tables can not be initialized 73 * statically - so there is no HASHTABLE_HEAD_INITIALIZED macro. 74 */ 75 #define HASHTABLE_INIT(table, type, field, _entries_size) \ 76 do { \ 77 hashtable_index_t var; \ 78 (table)->entries = calloc(_entries_size, \ 79 sizeof(*(table)->entries)); \ 80 (table)->entries_size = (_entries_size); \ 81 for (var = 0; var < HASHTABLE_ENTRIES_COUNT(table); ++var) {\ 82 (table)->entries[var].field.capacity = \ 83 HASHTABLE_INITIAL_ENTRIES_CAPACITY; \ 84 (table)->entries[var].field.size = 0; \ 85 (table)->entries[var].field.values = malloc( \ 86 sizeof(type) * \ 87 HASHTABLE_INITIAL_ENTRIES_CAPACITY); \ 88 assert((table)->entries[var].field.values != NULL);\ 89 } \ 90 } while (0) 91 92 /* 93 * All initialized hashtables should be destroyed with this macro. 94 */ 95 #define HASHTABLE_DESTROY(table, field) \ 96 do { \ 97 hashtable_index_t var; \ 98 for (var = 0; var < HASHTABLE_ENTRIES_COUNT(table); ++var) {\ 99 free((table)->entries[var].field.values); \ 100 } \ 101 } while (0) 102 103 #define HASHTABLE_GET_ENTRY(table, hash) \ 104 (&((table)->entries[hash])) 105 106 /* 107 * Traverses through all hash table entries 108 */ 109 #define HASHTABLE_FOREACH(table, var) \ 110 for ((var) = &((table)->entries[0]); \ 111 (var) < &((table)->entries[HASHTABLE_ENTRIES_COUNT(table)]);\ 112 ++(var)) 113 114 /* 115 * Traverses through all elements of the specified hash table entry 116 */ 117 #define HASHTABLE_ENTRY_FOREACH(entry, field, var) \ 118 for ((var) = &((entry)->field.values[0]); \ 119 (var) < &((entry)->field.values[(entry)->field.size]); \ 120 ++(var)) 121 122 #define HASHTABLE_ENTRY_CLEAR(entry, field) \ 123 ((entry)->field.size = 0) 124 125 #define HASHTABLE_ENTRY_SIZE(entry, field) \ 126 ((entry)->field.size) 127 128 #define HASHTABLE_ENTRY_CAPACITY(entry, field) \ 129 ((entry)->field.capacity) 130 131 #define HASHTABLE_ENTRY_CAPACITY_INCREASE(entry, field, type) \ 132 do { \ 133 (entry)->field.capacity *= 2; \ 134 (entry)->field.values = realloc((entry)->field.values, \ 135 (entry)->field.capacity * sizeof(type)); \ 136 } while (0) 137 138 #define HASHTABLE_ENTRY_CAPACITY_DECREASE(entry, field, type) \ 139 do { \ 140 (entry)->field.capacity /= 2; \ 141 (entry)->field.values = realloc((entry)->field.values, \ 142 (entry)->field.capacity * sizeof(type)); \ 143 } while (0) 144 145 /* 146 * Generates prototypes for the hash table functions 147 */ 148 #define HASHTABLE_PROTOTYPE(name, entry_, type) \ 149 hashtable_index_t name##_CALCULATE_HASH(struct name *, type *); \ 150 void name##_ENTRY_STORE(struct entry_*, type *); \ 151 type *name##_ENTRY_FIND(struct entry_*, type *); \ 152 type *name##_ENTRY_FIND_SPECIAL(struct entry_ *, type *, \ 153 int (*) (const void *, const void *)); \ 154 void name##_ENTRY_REMOVE(struct entry_*, type *); 155 156 /* 157 * Generates implementations of the hash table functions 158 */ 159 #define HASHTABLE_GENERATE(name, entry_, type, field, HASH, CMP) \ 160 hashtable_index_t name##_CALCULATE_HASH(struct name *table, type *data) \ 161 { \ 162 \ 163 return HASH(data, table->entries_size); \ 164 } \ 165 \ 166 void name##_ENTRY_STORE(struct entry_ *the_entry, type *data) \ 167 { \ 168 \ 169 if (the_entry->field.size == the_entry->field.capacity) \ 170 HASHTABLE_ENTRY_CAPACITY_INCREASE(the_entry, field, type);\ 171 \ 172 memcpy(&(the_entry->field.values[the_entry->field.size++]), \ 173 data, \ 174 sizeof(type)); \ 175 qsort(the_entry->field.values, the_entry->field.size, \ 176 sizeof(type), CMP); \ 177 } \ 178 \ 179 type *name##_ENTRY_FIND(struct entry_ *the_entry, type *key) \ 180 { \ 181 \ 182 return ((type *)bsearch(key, the_entry->field.values, \ 183 the_entry->field.size, sizeof(type), CMP)); \ 184 } \ 185 \ 186 type *name##_ENTRY_FIND_SPECIAL(struct entry_ *the_entry, type *key, \ 187 int (*compar) (const void *, const void *)) \ 188 { \ 189 return ((type *)bsearch(key, the_entry->field.values, \ 190 the_entry->field.size, sizeof(type), compar)); \ 191 } \ 192 \ 193 void name##_ENTRY_REMOVE(struct entry_ *the_entry, type *del_elm) \ 194 { \ 195 \ 196 memmove(del_elm, del_elm + 1, \ 197 (&the_entry->field.values[--the_entry->field.size] - del_elm) *\ 198 sizeof(type)); \ 199 } 200 201 /* 202 * Macro definitions below wrap the functions, generaed with 203 * HASHTABLE_GENERATE macro. You should use them and avoid using generated 204 * functions directly. 205 */ 206 #define HASHTABLE_CALCULATE_HASH(name, table, data) \ 207 (name##_CALCULATE_HASH((table), data)) 208 209 #define HASHTABLE_ENTRY_STORE(name, entry, data) \ 210 name##_ENTRY_STORE((entry), data) 211 212 #define HASHTABLE_ENTRY_FIND(name, entry, key) \ 213 (name##_ENTRY_FIND((entry), (key))) 214 215 #define HASHTABLE_ENTRY_FIND_SPECIAL(name, entry, key, cmp) \ 216 (name##_ENTRY_FIND_SPECIAL((entry), (key), (cmp))) 217 218 #define HASHTABLE_ENTRY_REMOVE(name, entry, del_elm) \ 219 name##_ENTRY_REMOVE((entry), (del_elm)) 220 221 #endif 222