blob: 5939549c0e7bc19b6e4afb7ccf775e2c64b91922 [file] [log] [blame]
Thomas Graf2de4ff72005-06-23 20:49:30 -07001/*
2 * lib/textsearch.c Generic text search interface
3 *
4 * This program is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU General Public License
6 * as published by the Free Software Foundation; either version
7 * 2 of the License, or (at your option) any later version.
8 *
9 * Authors: Thomas Graf <tgraf@suug.ch>
Pablo Neira Ayusoe03ba842007-12-01 00:03:52 +110010 * Pablo Neira Ayuso <pablo@netfilter.org>
Thomas Graf2de4ff72005-06-23 20:49:30 -070011 *
12 * ==========================================================================
Randy Dunlap5968a702018-04-16 12:32:55 -070013 */
14
15/**
16 * DOC: ts_intro
Thomas Graf2de4ff72005-06-23 20:49:30 -070017 * INTRODUCTION
18 *
Jesper Dangaard Brouerde0368d2011-01-24 02:41:37 +000019 * The textsearch infrastructure provides text searching facilities for
Thomas Graf2de4ff72005-06-23 20:49:30 -070020 * both linear and non-linear data. Individual search algorithms are
21 * implemented in modules and chosen by the user.
22 *
23 * ARCHITECTURE
24 *
Randy Dunlap5968a702018-04-16 12:32:55 -070025 * .. code-block:: none
26 *
27 * User
Thomas Graf2de4ff72005-06-23 20:49:30 -070028 * +----------------+
29 * | finish()|<--------------(6)-----------------+
30 * |get_next_block()|<--------------(5)---------------+ |
31 * | | Algorithm | |
32 * | | +------------------------------+
33 * | | | init() find() destroy() |
34 * | | +------------------------------+
35 * | | Core API ^ ^ ^
36 * | | +---------------+ (2) (4) (8)
37 * | (1)|----->| prepare() |---+ | |
38 * | (3)|----->| find()/next() |-----------+ |
39 * | (7)|----->| destroy() |----------------------+
40 * +----------------+ +---------------+
Randy Dunlap5968a702018-04-16 12:32:55 -070041 *
42 * (1) User configures a search by calling textsearch_prepare() specifying
43 * the search parameters such as the pattern and algorithm name.
Thomas Graf2de4ff72005-06-23 20:49:30 -070044 * (2) Core requests the algorithm to allocate and initialize a search
45 * configuration according to the specified parameters.
Randy Dunlap5968a702018-04-16 12:32:55 -070046 * (3) User starts the search(es) by calling textsearch_find() or
47 * textsearch_next() to fetch subsequent occurrences. A state variable
48 * is provided to the algorithm to store persistent variables.
Thomas Graf2de4ff72005-06-23 20:49:30 -070049 * (4) Core eventually resets the search offset and forwards the find()
50 * request to the algorithm.
Jesper Dangaard Brouerde0368d2011-01-24 02:41:37 +000051 * (5) Algorithm calls get_next_block() provided by the user continuously
Thomas Graf2de4ff72005-06-23 20:49:30 -070052 * to fetch the data to be searched in block by block.
53 * (6) Algorithm invokes finish() after the last call to get_next_block
54 * to clean up any leftovers from get_next_block. (Optional)
Randy Dunlap5968a702018-04-16 12:32:55 -070055 * (7) User destroys the configuration by calling textsearch_destroy().
Thomas Graf2de4ff72005-06-23 20:49:30 -070056 * (8) Core notifies the algorithm to destroy algorithm specific
57 * allocations. (Optional)
58 *
59 * USAGE
60 *
61 * Before a search can be performed, a configuration must be created
Joonwoo Parkb9c79672008-07-08 02:37:31 -070062 * by calling textsearch_prepare() specifying the searching algorithm,
63 * the pattern to look for and flags. As a flag, you can set TS_IGNORECASE
64 * to perform case insensitive matching. But it might slow down
65 * performance of algorithm, so you should use it at own your risk.
Jesper Dangaard Brouerde0368d2011-01-24 02:41:37 +000066 * The returned configuration may then be used for an arbitrary
Joonwoo Parkb9c79672008-07-08 02:37:31 -070067 * amount of times and even in parallel as long as a separate struct
68 * ts_state variable is provided to every instance.
Thomas Graf2de4ff72005-06-23 20:49:30 -070069 *
Randy Dunlap5968a702018-04-16 12:32:55 -070070 * The actual search is performed by either calling
71 * textsearch_find_continuous() for linear data or by providing
72 * an own get_next_block() implementation and
73 * calling textsearch_find(). Both functions return
Jesper Dangaard Brouerde0368d2011-01-24 02:41:37 +000074 * the position of the first occurrence of the pattern or UINT_MAX if
75 * no match was found. Subsequent occurrences can be found by calling
Thomas Graf2de4ff72005-06-23 20:49:30 -070076 * textsearch_next() regardless of the linearity of the data.
77 *
78 * Once you're done using a configuration it must be given back via
79 * textsearch_destroy.
80 *
Randy Dunlap5968a702018-04-16 12:32:55 -070081 * EXAMPLE::
Thomas Graf2de4ff72005-06-23 20:49:30 -070082 *
83 * int pos;
84 * struct ts_config *conf;
85 * struct ts_state state;
86 * const char *pattern = "chicken";
87 * const char *example = "We dance the funky chicken";
88 *
89 * conf = textsearch_prepare("kmp", pattern, strlen(pattern),
90 * GFP_KERNEL, TS_AUTOLOAD);
91 * if (IS_ERR(conf)) {
92 * err = PTR_ERR(conf);
93 * goto errout;
94 * }
95 *
Randy Dunlap5968a702018-04-16 12:32:55 -070096 * pos = textsearch_find_continuous(conf, \&state, example, strlen(example));
Thomas Graf2de4ff72005-06-23 20:49:30 -070097 * if (pos != UINT_MAX)
Randy Dunlap5968a702018-04-16 12:32:55 -070098 * panic("Oh my god, dancing chickens at \%d\n", pos);
Thomas Graf2de4ff72005-06-23 20:49:30 -070099 *
100 * textsearch_destroy(conf);
Thomas Graf2de4ff72005-06-23 20:49:30 -0700101 */
Randy Dunlap5968a702018-04-16 12:32:55 -0700102/* ========================================================================== */
Thomas Graf2de4ff72005-06-23 20:49:30 -0700103
Thomas Graf2de4ff72005-06-23 20:49:30 -0700104#include <linux/module.h>
105#include <linux/types.h>
106#include <linux/string.h>
107#include <linux/init.h>
Franck Bui-Huu82524742008-05-12 21:21:05 +0200108#include <linux/rculist.h>
Thomas Graf2de4ff72005-06-23 20:49:30 -0700109#include <linux/rcupdate.h>
110#include <linux/err.h>
111#include <linux/textsearch.h>
Tejun Heo5a0e3ad2010-03-24 17:04:11 +0900112#include <linux/slab.h>
Thomas Graf2de4ff72005-06-23 20:49:30 -0700113
114static LIST_HEAD(ts_ops);
115static DEFINE_SPINLOCK(ts_mod_lock);
116
117static inline struct ts_ops *lookup_ts_algo(const char *name)
118{
119 struct ts_ops *o;
120
121 rcu_read_lock();
122 list_for_each_entry_rcu(o, &ts_ops, list) {
123 if (!strcmp(name, o->name)) {
124 if (!try_module_get(o->owner))
125 o = NULL;
126 rcu_read_unlock();
127 return o;
128 }
129 }
130 rcu_read_unlock();
131
132 return NULL;
133}
134
135/**
136 * textsearch_register - register a textsearch module
137 * @ops: operations lookup table
138 *
139 * This function must be called by textsearch modules to announce
140 * their presence. The specified &@ops must have %name set to a
141 * unique identifier and the callbacks find(), init(), get_pattern(),
142 * and get_pattern_len() must be implemented.
143 *
144 * Returns 0 or -EEXISTS if another module has already registered
145 * with same name.
146 */
147int textsearch_register(struct ts_ops *ops)
148{
149 int err = -EEXIST;
150 struct ts_ops *o;
151
152 if (ops->name == NULL || ops->find == NULL || ops->init == NULL ||
153 ops->get_pattern == NULL || ops->get_pattern_len == NULL)
154 return -EINVAL;
155
156 spin_lock(&ts_mod_lock);
157 list_for_each_entry(o, &ts_ops, list) {
158 if (!strcmp(ops->name, o->name))
159 goto errout;
160 }
161
162 list_add_tail_rcu(&ops->list, &ts_ops);
163 err = 0;
164errout:
165 spin_unlock(&ts_mod_lock);
166 return err;
167}
Fabian Frederickce643a32014-06-04 16:11:58 -0700168EXPORT_SYMBOL(textsearch_register);
Thomas Graf2de4ff72005-06-23 20:49:30 -0700169
170/**
171 * textsearch_unregister - unregister a textsearch module
172 * @ops: operations lookup table
173 *
174 * This function must be called by textsearch modules to announce
175 * their disappearance for examples when the module gets unloaded.
176 * The &ops parameter must be the same as the one during the
177 * registration.
178 *
179 * Returns 0 on success or -ENOENT if no matching textsearch
180 * registration was found.
181 */
182int textsearch_unregister(struct ts_ops *ops)
183{
184 int err = 0;
185 struct ts_ops *o;
186
187 spin_lock(&ts_mod_lock);
188 list_for_each_entry(o, &ts_ops, list) {
189 if (o == ops) {
190 list_del_rcu(&o->list);
191 goto out;
192 }
193 }
194
195 err = -ENOENT;
196out:
197 spin_unlock(&ts_mod_lock);
198 return err;
199}
Fabian Frederickce643a32014-06-04 16:11:58 -0700200EXPORT_SYMBOL(textsearch_unregister);
Thomas Graf2de4ff72005-06-23 20:49:30 -0700201
202struct ts_linear_state
203{
204 unsigned int len;
205 const void *data;
206};
207
208static unsigned int get_linear_data(unsigned int consumed, const u8 **dst,
209 struct ts_config *conf,
210 struct ts_state *state)
211{
212 struct ts_linear_state *st = (struct ts_linear_state *) state->cb;
213
214 if (likely(consumed < st->len)) {
215 *dst = st->data + consumed;
216 return st->len - consumed;
217 }
218
219 return 0;
220}
221
222/**
223 * textsearch_find_continuous - search a pattern in continuous/linear data
224 * @conf: search configuration
225 * @state: search state
226 * @data: data to search in
227 * @len: length of data
228 *
229 * A simplified version of textsearch_find() for continuous/linear data.
230 * Call textsearch_next() to retrieve subsequent matches.
231 *
232 * Returns the position of first occurrence of the pattern or
Robert P. J. Day72fd4a32007-02-10 01:45:59 -0800233 * %UINT_MAX if no occurrence was found.
Randy Dunlap5968a702018-04-16 12:32:55 -0700234 */
Thomas Graf2de4ff72005-06-23 20:49:30 -0700235unsigned int textsearch_find_continuous(struct ts_config *conf,
236 struct ts_state *state,
237 const void *data, unsigned int len)
238{
239 struct ts_linear_state *st = (struct ts_linear_state *) state->cb;
240
241 conf->get_next_block = get_linear_data;
242 st->data = data;
243 st->len = len;
244
245 return textsearch_find(conf, state);
246}
Fabian Frederickce643a32014-06-04 16:11:58 -0700247EXPORT_SYMBOL(textsearch_find_continuous);
Thomas Graf2de4ff72005-06-23 20:49:30 -0700248
249/**
250 * textsearch_prepare - Prepare a search
251 * @algo: name of search algorithm
252 * @pattern: pattern data
253 * @len: length of pattern
254 * @gfp_mask: allocation mask
255 * @flags: search flags
256 *
257 * Looks up the search algorithm module and creates a new textsearch
Raphael Silvafec22902014-10-13 15:51:34 -0700258 * configuration for the specified pattern.
Thomas Graf2de4ff72005-06-23 20:49:30 -0700259 *
260 * Note: The format of the pattern may not be compatible between
261 * the various search algorithms.
262 *
263 * Returns a new textsearch configuration according to the specified
Pablo Neira Ayusoe03ba842007-12-01 00:03:52 +1100264 * parameters or a ERR_PTR(). If a zero length pattern is passed, this
265 * function returns EINVAL.
Thomas Graf2de4ff72005-06-23 20:49:30 -0700266 */
267struct ts_config *textsearch_prepare(const char *algo, const void *pattern,
Al Virofd4f2df2005-10-21 03:18:50 -0400268 unsigned int len, gfp_t gfp_mask, int flags)
Thomas Graf2de4ff72005-06-23 20:49:30 -0700269{
270 int err = -ENOENT;
271 struct ts_config *conf;
272 struct ts_ops *ops;
273
Pablo Neira Ayusoe03ba842007-12-01 00:03:52 +1100274 if (len == 0)
275 return ERR_PTR(-EINVAL);
276
Thomas Graf2de4ff72005-06-23 20:49:30 -0700277 ops = lookup_ts_algo(algo);
Johannes Berga00caa12008-07-08 19:00:24 +0200278#ifdef CONFIG_MODULES
Thomas Graf2de4ff72005-06-23 20:49:30 -0700279 /*
280 * Why not always autoload you may ask. Some users are
281 * in a situation where requesting a module may deadlock,
282 * especially when the module is located on a NFS mount.
283 */
284 if (ops == NULL && flags & TS_AUTOLOAD) {
285 request_module("ts_%s", algo);
286 ops = lookup_ts_algo(algo);
287 }
288#endif
289
290 if (ops == NULL)
291 goto errout;
292
Joonwoo Parkb9c79672008-07-08 02:37:31 -0700293 conf = ops->init(pattern, len, gfp_mask, flags);
Thomas Graf2de4ff72005-06-23 20:49:30 -0700294 if (IS_ERR(conf)) {
295 err = PTR_ERR(conf);
296 goto errout;
297 }
298
299 conf->ops = ops;
300 return conf;
301
302errout:
303 if (ops)
304 module_put(ops->owner);
305
306 return ERR_PTR(err);
307}
Fabian Frederickce643a32014-06-04 16:11:58 -0700308EXPORT_SYMBOL(textsearch_prepare);
Thomas Graf2de4ff72005-06-23 20:49:30 -0700309
310/**
311 * textsearch_destroy - destroy a search configuration
312 * @conf: search configuration
313 *
314 * Releases all references of the configuration and frees
315 * up the memory.
316 */
317void textsearch_destroy(struct ts_config *conf)
318{
319 if (conf->ops) {
320 if (conf->ops->destroy)
321 conf->ops->destroy(conf);
322 module_put(conf->ops->owner);
323 }
324
325 kfree(conf);
326}
Thomas Graf2de4ff72005-06-23 20:49:30 -0700327EXPORT_SYMBOL(textsearch_destroy);