Andrew's git - gitweb.git/blob

   1/*
   2 * Generic implementation of hash-based key value mappings.
   3 */
   4#include "cache.h"
   5#include "hashmap.h"
   6
   7#define FNV32_BASE ((unsigned int) 0x811c9dc5)
   8#define FNV32_PRIME ((unsigned int) 0x01000193)
   9
  10unsigned int strhash(const char *str)
  11{
  12        unsigned int c, hash = FNV32_BASE;
  13        while ((c = (unsigned char) *str++))
  14                hash = (hash * FNV32_PRIME) ^ c;
  15        return hash;
  16}
  17
  18unsigned int strihash(const char *str)
  19{
  20        unsigned int c, hash = FNV32_BASE;
  21        while ((c = (unsigned char) *str++)) {
  22                if (c >= 'a' && c <= 'z')
  23                        c -= 'a' - 'A';
  24                hash = (hash * FNV32_PRIME) ^ c;
  25        }
  26        return hash;
  27}
  28
  29unsigned int memhash(const void *buf, size_t len)
  30{
  31        unsigned int hash = FNV32_BASE;
  32        unsigned char *ucbuf = (unsigned char *) buf;
  33        while (len--) {
  34                unsigned int c = *ucbuf++;
  35                hash = (hash * FNV32_PRIME) ^ c;
  36        }
  37        return hash;
  38}
  39
  40unsigned int memihash(const void *buf, size_t len)
  41{
  42        unsigned int hash = FNV32_BASE;
  43        unsigned char *ucbuf = (unsigned char *) buf;
  44        while (len--) {
  45                unsigned int c = *ucbuf++;
  46                if (c >= 'a' && c <= 'z')
  47                        c -= 'a' - 'A';
  48                hash = (hash * FNV32_PRIME) ^ c;
  49        }
  50        return hash;
  51}
  52
  53/*
  54 * Incoporate another chunk of data into a memihash
  55 * computation.
  56 */
  57unsigned int memihash_cont(unsigned int hash_seed, const void *buf, size_t len)
  58{
  59        unsigned int hash = hash_seed;
  60        unsigned char *ucbuf = (unsigned char *) buf;
  61        while (len--) {
  62                unsigned int c = *ucbuf++;
  63                if (c >= 'a' && c <= 'z')
  64                        c -= 'a' - 'A';
  65                hash = (hash * FNV32_PRIME) ^ c;
  66        }
  67        return hash;
  68}
  69
  70#define HASHMAP_INITIAL_SIZE 64
  71/* grow / shrink by 2^2 */
  72#define HASHMAP_RESIZE_BITS 2
  73/* load factor in percent */
  74#define HASHMAP_LOAD_FACTOR 80
  75
  76static void alloc_table(struct hashmap *map, unsigned int size)
  77{
  78        map->tablesize = size;
  79        map->table = xcalloc(size, sizeof(struct hashmap_entry *));
  80
  81        /* calculate resize thresholds for new size */
  82        map->grow_at = (unsigned int) ((uint64_t) size * HASHMAP_LOAD_FACTOR / 100);
  83        if (size <= HASHMAP_INITIAL_SIZE)
  84                map->shrink_at = 0;
  85        else
  86                /*
  87                 * The shrink-threshold must be slightly smaller than
  88                 * (grow-threshold / resize-factor) to prevent erratic resizing,
  89                 * thus we divide by (resize-factor + 1).
  90                 */
  91                map->shrink_at = map->grow_at / ((1 << HASHMAP_RESIZE_BITS) + 1);
  92}
  93
  94static inline int entry_equals(const struct hashmap *map,
  95                const struct hashmap_entry *e1, const struct hashmap_entry *e2,
  96                const void *keydata)
  97{
  98        return (e1 == e2) || (e1->hash == e2->hash && !map->cmpfn(e1, e2, keydata));
  99}
 100
 101static inline unsigned int bucket(const struct hashmap *map,
 102                const struct hashmap_entry *key)
 103{
 104        return key->hash & (map->tablesize - 1);
 105}
 106
 107int hashmap_bucket(const struct hashmap *map, unsigned int hash)
 108{
 109        return hash & (map->tablesize - 1);
 110}
 111
 112static void rehash(struct hashmap *map, unsigned int newsize)
 113{
 114        unsigned int i, oldsize = map->tablesize;
 115        struct hashmap_entry **oldtable = map->table;
 116
 117        if (map->disallow_rehash)
 118                return;
 119
 120        alloc_table(map, newsize);
 121        for (i = 0; i < oldsize; i++) {
 122                struct hashmap_entry *e = oldtable[i];
 123                while (e) {
 124                        struct hashmap_entry *next = e->next;
 125                        unsigned int b = bucket(map, e);
 126                        e->next = map->table[b];
 127                        map->table[b] = e;
 128                        e = next;
 129                }
 130        }
 131        free(oldtable);
 132}
 133
 134static inline struct hashmap_entry **find_entry_ptr(const struct hashmap *map,
 135                const struct hashmap_entry *key, const void *keydata)
 136{
 137        struct hashmap_entry **e = &map->table[bucket(map, key)];
 138        while (*e && !entry_equals(map, *e, key, keydata))
 139                e = &(*e)->next;
 140        return e;
 141}
 142
 143static int always_equal(const void *unused1, const void *unused2, const void *unused3)
 144{
 145        return 0;
 146}
 147
 148void hashmap_init(struct hashmap *map, hashmap_cmp_fn equals_function,
 149                size_t initial_size)
 150{
 151        unsigned int size = HASHMAP_INITIAL_SIZE;
 152
 153        memset(map, 0, sizeof(*map));
 154
 155        map->cmpfn = equals_function ? equals_function : always_equal;
 156
 157        /* calculate initial table size and allocate the table */
 158        initial_size = (unsigned int) ((uint64_t) initial_size * 100
 159                        / HASHMAP_LOAD_FACTOR);
 160        while (initial_size > size)
 161                size <<= HASHMAP_RESIZE_BITS;
 162        alloc_table(map, size);
 163}
 164
 165void hashmap_free(struct hashmap *map, int free_entries)
 166{
 167        if (!map || !map->table)
 168                return;
 169        if (free_entries) {
 170                struct hashmap_iter iter;
 171                struct hashmap_entry *e;
 172                hashmap_iter_init(map, &iter);
 173                while ((e = hashmap_iter_next(&iter)))
 174                        free(e);
 175        }
 176        free(map->table);
 177        memset(map, 0, sizeof(*map));
 178}
 179
 180void *hashmap_get(const struct hashmap *map, const void *key, const void *keydata)
 181{
 182        return *find_entry_ptr(map, key, keydata);
 183}
 184
 185void *hashmap_get_next(const struct hashmap *map, const void *entry)
 186{
 187        struct hashmap_entry *e = ((struct hashmap_entry *) entry)->next;
 188        for (; e; e = e->next)
 189                if (entry_equals(map, entry, e, NULL))
 190                        return e;
 191        return NULL;
 192}
 193
 194void hashmap_add(struct hashmap *map, void *entry)
 195{
 196        unsigned int b = bucket(map, entry);
 197
 198        /* add entry */
 199        ((struct hashmap_entry *) entry)->next = map->table[b];
 200        map->table[b] = entry;
 201
 202        /* fix size and rehash if appropriate */
 203        map->size++;
 204        if (map->size > map->grow_at)
 205                rehash(map, map->tablesize << HASHMAP_RESIZE_BITS);
 206}
 207
 208void *hashmap_remove(struct hashmap *map, const void *key, const void *keydata)
 209{
 210        struct hashmap_entry *old;
 211        struct hashmap_entry **e = find_entry_ptr(map, key, keydata);
 212        if (!*e)
 213                return NULL;
 214
 215        /* remove existing entry */
 216        old = *e;
 217        *e = old->next;
 218        old->next = NULL;
 219
 220        /* fix size and rehash if appropriate */
 221        map->size--;
 222        if (map->size < map->shrink_at)
 223                rehash(map, map->tablesize >> HASHMAP_RESIZE_BITS);
 224        return old;
 225}
 226
 227void *hashmap_put(struct hashmap *map, void *entry)
 228{
 229        struct hashmap_entry *old = hashmap_remove(map, entry, NULL);
 230        hashmap_add(map, entry);
 231        return old;
 232}
 233
 234void hashmap_iter_init(struct hashmap *map, struct hashmap_iter *iter)
 235{
 236        iter->map = map;
 237        iter->tablepos = 0;
 238        iter->next = NULL;
 239}
 240
 241void *hashmap_iter_next(struct hashmap_iter *iter)
 242{
 243        struct hashmap_entry *current = iter->next;
 244        for (;;) {
 245                if (current) {
 246                        iter->next = current->next;
 247                        return current;
 248                }
 249
 250                if (iter->tablepos >= iter->map->tablesize)
 251                        return NULL;
 252
 253                current = iter->map->table[iter->tablepos++];
 254        }
 255}
 256
 257struct pool_entry {
 258        struct hashmap_entry ent;
 259        size_t len;
 260        unsigned char data[FLEX_ARRAY];
 261};
 262
 263static int pool_entry_cmp(const struct pool_entry *e1,
 264                          const struct pool_entry *e2,
 265                          const unsigned char *keydata)
 266{
 267        return e1->data != keydata &&
 268               (e1->len != e2->len || memcmp(e1->data, keydata, e1->len));
 269}
 270
 271const void *memintern(const void *data, size_t len)
 272{
 273        static struct hashmap map;
 274        struct pool_entry key, *e;
 275
 276        /* initialize string pool hashmap */
 277        if (!map.tablesize)
 278                hashmap_init(&map, (hashmap_cmp_fn) pool_entry_cmp, 0);
 279
 280        /* lookup interned string in pool */
 281        hashmap_entry_init(&key, memhash(data, len));
 282        key.len = len;
 283        e = hashmap_get(&map, &key, data);
 284        if (!e) {
 285                /* not found: create it */
 286                FLEX_ALLOC_MEM(e, data, data, len);
 287                hashmap_entry_init(e, key.ent.hash);
 288                e->len = len;
 289                hashmap_add(&map, e);
 290        }
 291        return e->data;
 292}