1 .\" $OpenBSD: ohash_init.3,v 1.2 2014/05/13 14:01:41 jmc Exp $
2 .\" Copyright (c) 1999 Marc Espie <espie@openbsd.org>
4 .\" Permission to use, copy, modify, and distribute this software for any
5 .\" purpose with or without fee is hereby granted, provided that the above
6 .\" copyright notice and this permission notice appear in all copies.
8 .\" THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
9 .\" WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
10 .\" MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
11 .\" ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
12 .\" WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
13 .\" ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
14 .\" OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
22 .Nm ohash_lookup_interval ,
23 .Nm ohash_lookup_memory ,
30 .Nd light-weight open hashing
36 .Fn ohash_init "struct ohash *h" "unsigned int size" "struct ohash_info *info"
38 .Fn ohash_delete "struct ohash *h"
40 .Fn ohash_lookup_interval "struct ohash *h" "const char *start" "const char *end" "uint32_t hv"
42 .Fn ohash_lookup_memory "struct ohash *h" "const char *k" "size_t s" "uint32_t hv"
44 .Fn ohash_find "struct ohash *h" "unsigned int i"
46 .Fn ohash_remove "struct ohash *h" "unsigned int i"
48 .Fn ohash_insert "struct ohash *h" "unsigned int i" "void *p"
50 .Fn ohash_first "struct ohash *h" "unsigned int *i"
52 .Fn ohash_next "struct ohash *h" "unsigned int *i"
54 .Fn ohash_entries "struct ohash *h"
56 These functions have been designed as a fast, extensible alternative to
57 the usual hash table functions.
58 They provide storage and retrieval of records indexed by keys,
59 where a key is a contiguous sequence of bytes at a fixed position in
61 Keys can either be NUL-terminated strings or fixed-size memory areas.
62 All functions take a pointer to an ohash structure as the
65 Storage for this structure should be provided by user code.
68 initializes the table to store roughly 2 to the power
73 .Fa struct ohash_info .
74 .Bd -literal -offset indent
77 void *data; /* user data */
78 void *(*calloc)(size_t, size_t, void *);
79 void (*free)(void *, void *);
80 void *(*alloc)(size_t, void *);
86 field holds the position of the key in each record;
91 fields are pointers to
95 functions, used for managing the table internal storage;
98 field is only used by the utility function
99 .Xr ohash_create_entry 3 .
101 Each of these functions are called similarly to their standard counterpart,
104 parameter corresponding to the content of the field
106 which can be used to communicate specific information to the functions.
109 stores a copy of those fields internally, so
111 can be reclaimed after initialization.
114 frees storage internal to
116 Elements themselves should be freed by the user first, using for instance
121 .Fn ohash_lookup_interval
123 .Fn ohash_lookup_memory
124 are the basic look-up element functions.
125 The hashing function result is provided by the user as
136 This slot is only valid up to the next call to
141 .Fn ohash_lookup_interval
142 handles string-like keys.
143 .Fn ohash_lookup_interval
144 assumes the key is the interval between
149 though the actual elements stored in the table should only contain
152 .Fn ohash_lookup_memory
153 assumes the key is the memory area starting at
157 All bytes are significant in key comparison.
160 retrieves an element from a slot
167 if the slot is empty.
170 inserts a new element
176 must be empty and element
178 must have a key corresponding to the
183 removes the element at slot
185 It returns the removed element, for user code to dispose of, or
187 if the slot was empty.
192 can be used to access all elements in an ohash table, like this:
193 .Bd -literal -offset indent
194 for (n = ohash_first(h, &i); n != NULL; n = ohash_next(h, &i))
195 do_something_with(n);
199 points to an auxiliary unsigned integer used to record the current position
201 Those functions are safe to use even while entries are added to/removed
202 from the table, but in such a case they don't guarantee that new entries
204 As a special case, they can safely be used to free elements in the table.
207 returns the number of elements in the hash table.
215 may call the user-supplied memory functions:
216 .Bd -literal -offset indent
217 p = (*info->calloc)(n, sizeof_record, info->data);
218 /* copy data from old to p */
219 (*info->free)(old, info->data);
222 It is the responsibility of the user memory allocation code to verify
223 that those calls did not fail.
225 If memory allocation fails,
227 returns a useless hash table.
231 still perform the requested operation, but the returned table should be
232 considered read-only.
233 It can still be accessed by
239 to dump relevant information to disk before aborting.
241 The open hashing functions are not thread-safe by design.
242 In particular, in a threaded environment, there is no guarantee that a
244 will not move between a
253 Multi-threaded applications should explicitly protect ohash table access.
259 .%B The Art of Computer Programming
265 Those functions are completely non-standard and should be avoided in
268 Those functions were designed and written for
271 by Marc Espie in 1999.