freebsd-skq/usr.sbin/cached/hashtable.h
Hajimu UMEMOTO 06a99fe36f - Extend the nsswitch to support Services, Protocols and Rpc
databases.
- Make nsswitch support caching.

Submitted by:	Michael Bushkov <bushman__at__rsu.ru>
Sponsored by:	Google Summer of Code 2005
2006-04-28 12:03:38 +00:00

219 lines
7.5 KiB
C

/*-
* Copyright (c) 2005 Michael Bushkov <bushman@rsu.ru>
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* $FreeBSD$
*/
#ifndef __CACHELIB_HASHTABLE_H__
#define __CACHELIB_HASHTABLE_H__
#include <search.h>
#include <string.h>
#define HASHTABLE_INITIAL_ENTRIES_CAPACITY 8
typedef int hashtable_index_t;
/*
* This file contains queue.h-like macro definitions for hash tables.
* Hash table is organized as an array of the specified size of the user
* defined (with HASTABLE_ENTRY_HEAD) structures. Each hash table
* entry (user defined structure) stores its elements in the sorted array.
* You can place elements into the hash table, retrieve elements with
* specified key, traverse through all elements, and delete them.
* New elements are placed into the hash table by using the compare and
* hashing functions, provided by the user.
*/
/*
* Defines the hash table entry structure, that uses specified type of
* elements.
*/
#define HASHTABLE_ENTRY_HEAD(name, type) struct name { \
type *values; \
size_t capacity; \
size_t size; \
}
/*
* Defines the hash table structure, which uses the specified type of entries.
* The only restriction for entries is that is that they should have the field,
* defined with HASHTABLE_ENTRY_HEAD macro.
*/
#define HASHTABLE_HEAD(name, entry) struct name { \
struct entry *entries; \
size_t entries_size; \
}
#define HASHTABLE_ENTRIES_COUNT(table) ((table)->entries_size)
/*
* Unlike most of queue.h data types, hash tables can not be initialized
* statically - so there is no HASHTABLE_HEAD_INITIALIZED macro.
*/
#define HASHTABLE_INIT(table, type, field, _entries_size) \
do { \
hashtable_index_t var; \
(table)->entries = (void *)malloc( \
sizeof(*(table)->entries) * (_entries_size)); \
memset((table)->entries, 0, \
sizeof(*(table)->entries) * (_entries_size)); \
(table)->entries_size = (_entries_size); \
for (var = 0; var < HASHTABLE_ENTRIES_COUNT(table); ++var) {\
(table)->entries[var].field.capacity = \
HASHTABLE_INITIAL_ENTRIES_CAPACITY; \
(table)->entries[var].field.size = 0; \
(table)->entries[var].field.values = (type *)malloc(\
sizeof(type) * \
HASHTABLE_INITIAL_ENTRIES_CAPACITY); \
assert((table)->entries[var].field.values != NULL);\
} \
} while (0)
/*
* All initialized hashtables should be destroyed with this macro.
*/
#define HASHTABLE_DESTROY(table, field) \
do { \
hashtable_index_t var; \
for (var = 0; var < HASHTABLE_ENTRIES_COUNT(table); ++var) {\
free((table)->entries[var].field.values); \
} \
} while (0)
#define HASHTABLE_GET_ENTRY(table, hash) (&((table)->entries[hash]))
/*
* Traverses through all hash table entries
*/
#define HASHTABLE_FOREACH(table, var) \
for ((var) = &((table)->entries[0]); \
(var) < &((table)->entries[HASHTABLE_ENTRIES_COUNT(table)]);\
++(var))
/*
* Traverses through all elements of the specified hash table entry
*/
#define HASHTABLE_ENTRY_FOREACH(entry, field, var) \
for ((var) = &((entry)->field.values[0]); \
(var) < &((entry)->field.values[(entry)->field.size]); \
++(var))
#define HASHTABLE_ENTRY_CLEAR(entry, field) \
((entry)->field.size = 0)
#define HASHTABLE_ENTRY_SIZE(entry, field) \
((entry)->field.size)
#define HASHTABLE_ENTRY_CAPACITY(entry, field) \
((entry)->field.capacity)
#define HASHTABLE_ENTRY_CAPACITY_INCREASE(entry, field, type) \
(entry)->field.capacity *= 2; \
(entry)->field.values = (type *)realloc((entry)->field.values, \
(entry)->field.capacity * sizeof(type));
#define HASHTABLE_ENTRY_CAPACITY_DECREASE(entry, field, type) \
(entry)->field.capacity /= 2; \
(entry)->field.values = (type *)realloc((entry)->field.values, \
(entry)->field.capacity * sizeof(type));
/*
* Generates prototypes for the hash table functions
*/
#define HASHTABLE_PROTOTYPE(name, entry_, type) \
hashtable_index_t name##_CALCULATE_HASH(struct name *, type *); \
void name##_ENTRY_STORE(struct entry_*, type *); \
type *name##_ENTRY_FIND(struct entry_*, type *); \
type *name##_ENTRY_FIND_SPECIAL(struct entry_ *, type *, \
int (*) (const void *, const void *)); \
void name##_ENTRY_REMOVE(struct entry_*, type *);
/*
* Generates implementations of the hash table functions
*/
#define HASHTABLE_GENERATE(name, entry_, type, field, HASH, CMP) \
hashtable_index_t name##_CALCULATE_HASH(struct name *table, type *data) \
{ \
\
return HASH(data, table->entries_size); \
} \
\
void name##_ENTRY_STORE(struct entry_ *the_entry, type *data) \
{ \
\
if (the_entry->field.size == the_entry->field.capacity) \
HASHTABLE_ENTRY_CAPACITY_INCREASE(the_entry, field, type);\
\
memcpy(&(the_entry->field.values[the_entry->field.size++]), \
data, \
sizeof(type)); \
qsort(the_entry->field.values, the_entry->field.size, \
sizeof(type), CMP); \
} \
\
type *name##_ENTRY_FIND(struct entry_ *the_entry, type *key) \
{ \
\
return ((type *)bsearch(key, the_entry->field.values, \
the_entry->field.size, sizeof(type), CMP)); \
} \
\
type *name##_ENTRY_FIND_SPECIAL(struct entry_ *the_entry, type *key, \
int (*compar) (const void *, const void *)) \
{ \
return ((type *)bsearch(key, the_entry->field.values, \
the_entry->field.size, sizeof(type), compar)); \
} \
\
void name##_ENTRY_REMOVE(struct entry_ *the_entry, type *del_elm) \
{ \
\
memmove(del_elm, del_elm + 1, \
(&the_entry->field.values[--the_entry->field.size] - del_elm) *\
sizeof(type)); \
}
/*
* Macro definitions below wrap the functions, generaed with
* HASHTABLE_GENERATE macro. You should use them and avoid using generated
* functions directly.
*/
#define HASHTABLE_CALCULATE_HASH(name, table, data) \
(name##_CALCULATE_HASH((table), data))
#define HASHTABLE_ENTRY_STORE(name, entry, data) \
name##_ENTRY_STORE((entry), data)
#define HASHTABLE_ENTRY_FIND(name, entry, key) \
(name##_ENTRY_FIND((entry), (key)))
#define HASHTABLE_ENTRY_FIND_SPECIAL(name, entry, key, cmp) \
(name##_ENTRY_FIND_SPECIAL((entry), (key), (cmp)))
#define HASHTABLE_ENTRY_REMOVE(name, entry, del_elm) \
name##_ENTRY_REMOVE((entry), (del_elm))
#endif