aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
Diffstat
-rw-r--r--include/linux/textsearch.h180
-rw-r--r--lib/Kconfig8
-rw-r--r--lib/Makefile2
-rw-r--r--lib/textsearch.c317
4 files changed, 506 insertions, 1 deletions
diff --git a/include/linux/textsearch.h b/include/linux/textsearch.h
new file mode 100644
index 000000000000..941f45ac117a
--- /dev/null
+++ b/include/linux/textsearch.h
@@ -0,0 +1,180 @@
1#ifndef __LINUX_TEXTSEARCH_H
2#define __LINUX_TEXTSEARCH_H
3
4#ifdef __KERNEL__
5
6#include <linux/types.h>
7#include <linux/list.h>
8#include <linux/kernel.h>
9#include <linux/module.h>
10#include <linux/err.h>
11
12struct ts_config;
13
14/**
15 * TS_AUTOLOAD - Automatically load textsearch modules when needed
16 */
17#define TS_AUTOLOAD 1
18
19/**
20 * struct ts_state - search state
21 * @offset: offset for next match
22 * @cb: control buffer, for persistant variables of get_next_block()
23 */
24struct ts_state
25{
26 unsigned int offset;
27 char cb[40];
28};
29
30/**
31 * struct ts_ops - search module operations
32 * @name: name of search algorithm
33 * @init: initialization function to prepare a search
34 * @find: find the next occurrence of the pattern
35 * @destroy: destroy algorithm specific parts of a search configuration
36 * @get_pattern: return head of pattern
37 * @get_pattern_len: return length of pattern
38 * @owner: module reference to algorithm
39 */
40struct ts_ops
41{
42 const char *name;
43 struct ts_config * (*init)(const void *, unsigned int, int);
44 unsigned int (*find)(struct ts_config *,
45 struct ts_state *);
46 void (*destroy)(struct ts_config *);
47 void * (*get_pattern)(struct ts_config *);
48 unsigned int (*get_pattern_len)(struct ts_config *);
49 struct module *owner;
50 struct list_head list;
51};
52
53/**
54 * struct ts_config - search configuration
55 * @ops: operations of chosen algorithm
56 * @get_next_block: callback to fetch the next block to search in
57 * @finish: callback to finalize a search
58 */
59struct ts_config
60{
61 struct ts_ops *ops;
62
63 /**
64 * get_next_block - fetch next block of data
65 * @consumed: number of bytes consumed by the caller
66 * @dst: destination buffer
67 * @conf: search configuration
68 * @state: search state
69 *
70 * Called repeatedly until 0 is returned. Must assign the
71 * head of the next block of data to &*dst and return the length
72 * of the block or 0 if at the end. consumed == 0 indicates
73 * a new search. May store/read persistant values in state->cb.
74 */
75 unsigned int (*get_next_block)(unsigned int consumed,
76 const u8 **dst,
77 struct ts_config *conf,
78 struct ts_state *state);
79
80 /**
81 * finish - finalize/clean a series of get_next_block() calls
82 * @conf: search configuration
83 * @state: search state
84 *
85 * Called after the last use of get_next_block(), may be used
86 * to cleanup any leftovers.
87 */
88 void (*finish)(struct ts_config *conf,
89 struct ts_state *state);
90};
91
92/**
93 * textsearch_next - continue searching for a pattern
94 * @conf: search configuration
95 * @state: search state
96 *
97 * Continues a search looking for more occurrences of the pattern.
98 * textsearch_find() must be called to find the first occurrence
99 * in order to reset the state.
100 *
101 * Returns the position of the next occurrence of the pattern or
102 * UINT_MAX if not match was found.
103 */
104static inline unsigned int textsearch_next(struct ts_config *conf,
105 struct ts_state *state)
106{
107 unsigned int ret = conf->ops->find(conf, state);
108
109 if (conf->finish)
110 conf->finish(conf, state);
111
112 return ret;
113}
114
115/**
116 * textsearch_find - start searching for a pattern
117 * @conf: search configuration
118 * @state: search state
119 *
120 * Returns the position of first occurrence of the pattern or
121 * UINT_MAX if no match was found.
122 */
123static inline unsigned int textsearch_find(struct ts_config *conf,
124 struct ts_state *state)
125{
126 state->offset = 0;
127 return textsearch_next(conf, state);
128}
129
130/**
131 * textsearch_get_pattern - return head of the pattern
132 * @conf: search configuration
133 */
134static inline void *textsearch_get_pattern(struct ts_config *conf)
135{
136 return conf->ops->get_pattern(conf);
137}
138
139/**
140 * textsearch_get_pattern_len - return length of the pattern
141 * @conf: search configuration
142 */
143static inline unsigned int textsearch_get_pattern_len(struct ts_config *conf)
144{
145 return conf->ops->get_pattern_len(conf);
146}
147
148extern int textsearch_register(struct ts_ops *);
149extern int textsearch_unregister(struct ts_ops *);
150extern struct ts_config *textsearch_prepare(const char *, const void *,
151 unsigned int, int, int);
152extern void textsearch_destroy(struct ts_config *conf);
153extern unsigned int textsearch_find_continuous(struct ts_config *,
154 struct ts_state *,
155 const void *, unsigned int);
156
157
158#define TS_PRIV_ALIGNTO 8
159#define TS_PRIV_ALIGN(len) (((len) + TS_PRIV_ALIGNTO-1) & ~(TS_PRIV_ALIGNTO-1))
160
161static inline struct ts_config *alloc_ts_config(size_t payload, int gfp_mask)
162{
163 struct ts_config *conf;
164
165 conf = kmalloc(TS_PRIV_ALIGN(sizeof(*conf)) + payload, gfp_mask);
166 if (conf == NULL)
167 return ERR_PTR(-ENOMEM);
168
169 memset(conf, 0, TS_PRIV_ALIGN(sizeof(*conf)) + payload);
170 return conf;
171}
172
173static inline void *ts_config_priv(struct ts_config *conf)
174{
175 return ((u8 *) conf + TS_PRIV_ALIGN(sizeof(struct ts_config)));
176}
177
178#endif /* __KERNEL__ */
179
180#endif
diff --git a/lib/Kconfig b/lib/Kconfig
index 2d4d4e3bc4aa..5bc2d523e6d1 100644
--- a/lib/Kconfig
+++ b/lib/Kconfig
@@ -63,5 +63,11 @@ config REED_SOLOMON_ENC16
63config REED_SOLOMON_DEC16 63config REED_SOLOMON_DEC16
64 boolean 64 boolean
65 65
66endmenu 66config TEXTSEARCH
67 boolean "Textsearch infrastructure"
68 default y
69 help
70 Say Y here if you want to provide a textsearch infrastructure
71 to other subsystems.
67 72
73endmenu
diff --git a/lib/Makefile b/lib/Makefile
index dcb4231916e2..3e917436ad60 100644
--- a/lib/Makefile
+++ b/lib/Makefile
@@ -36,6 +36,8 @@ obj-$(CONFIG_ZLIB_INFLATE) += zlib_inflate/
36obj-$(CONFIG_ZLIB_DEFLATE) += zlib_deflate/ 36obj-$(CONFIG_ZLIB_DEFLATE) += zlib_deflate/
37obj-$(CONFIG_REED_SOLOMON) += reed_solomon/ 37obj-$(CONFIG_REED_SOLOMON) += reed_solomon/
38 38
39lib-$(CONFIG_TEXTSEARCH) += textsearch.o
40
39hostprogs-y := gen_crc32table 41hostprogs-y := gen_crc32table
40clean-files := crc32table.h 42clean-files := crc32table.h
41 43
diff --git a/lib/textsearch.c b/lib/textsearch.c
new file mode 100644
index 000000000000..1e934c196f0f
--- /dev/null
+++ b/lib/textsearch.c
@@ -0,0 +1,317 @@
1/*
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>
10 * Pablo Neira Ayuso <pablo@eurodev.net>
11 *
12 * ==========================================================================
13 *
14 * INTRODUCTION
15 *
16 * The textsearch infrastructure provides text searching facitilies for
17 * both linear and non-linear data. Individual search algorithms are
18 * implemented in modules and chosen by the user.
19 *
20 * ARCHITECTURE
21 *
22 * User
23 * +----------------+
24 * | finish()|<--------------(6)-----------------+
25 * |get_next_block()|<--------------(5)---------------+ |
26 * | | Algorithm | |
27 * | | +------------------------------+
28 * | | | init() find() destroy() |
29 * | | +------------------------------+
30 * | | Core API ^ ^ ^
31 * | | +---------------+ (2) (4) (8)
32 * | (1)|----->| prepare() |---+ | |
33 * | (3)|----->| find()/next() |-----------+ |
34 * | (7)|----->| destroy() |----------------------+
35 * +----------------+ +---------------+
36 *
37 * (1) User configures a search by calling _prepare() specifying the
38 * search parameters such as the pattern and algorithm name.
39 * (2) Core requests the algorithm to allocate and initialize a search
40 * configuration according to the specified parameters.
41 * (3) User starts the search(es) by calling _find() or _next() to
42 * fetch subsequent occurrences. A state variable is provided
43 * to the algorihtm to store persistant variables.
44 * (4) Core eventually resets the search offset and forwards the find()
45 * request to the algorithm.
46 * (5) Algorithm calls get_next_block() provided by the user continously
47 * to fetch the data to be searched in block by block.
48 * (6) Algorithm invokes finish() after the last call to get_next_block
49 * to clean up any leftovers from get_next_block. (Optional)
50 * (7) User destroys the configuration by calling _destroy().
51 * (8) Core notifies the algorithm to destroy algorithm specific
52 * allocations. (Optional)
53 *
54 * USAGE
55 *
56 * Before a search can be performed, a configuration must be created
57 * by calling textsearch_prepare() specyfing the searching algorithm and
58 * the pattern to look for. The returned configuration may then be used
59 * for an arbitary amount of times and even in parallel as long as a
60 * separate struct ts_state variable is provided to every instance.
61 *
62 * The actual search is performed by either calling textsearch_find_-
63 * continuous() for linear data or by providing an own get_next_block()
64 * implementation and calling textsearch_find(). Both functions return
65 * the position of the first occurrence of the patern or UINT_MAX if
66 * no match was found. Subsequent occurences can be found by calling
67 * textsearch_next() regardless of the linearity of the data.
68 *
69 * Once you're done using a configuration it must be given back via
70 * textsearch_destroy.
71 *
72 * EXAMPLE
73 *
74 * int pos;
75 * struct ts_config *conf;
76 * struct ts_state state;
77 * const char *pattern = "chicken";
78 * const char *example = "We dance the funky chicken";
79 *
80 * conf = textsearch_prepare("kmp", pattern, strlen(pattern),
81 * GFP_KERNEL, TS_AUTOLOAD);
82 * if (IS_ERR(conf)) {
83 * err = PTR_ERR(conf);
84 * goto errout;
85 * }
86 *
87 * pos = textsearch_find_continuous(conf, &state, example, strlen(example));
88 * if (pos != UINT_MAX)
89 * panic("Oh my god, dancing chickens at %d\n", pos);
90 *
91 * textsearch_destroy(conf);
92 *
93 * ==========================================================================
94 */
95
96#include <linux/config.h>
97#include <linux/module.h>
98#include <linux/types.h>
99#include <linux/string.h>
100#include <linux/init.h>
101#include <linux/rcupdate.h>
102#include <linux/err.h>
103#include <linux/textsearch.h>
104
105static LIST_HEAD(ts_ops);
106static DEFINE_SPINLOCK(ts_mod_lock);
107
108static inline struct ts_ops *lookup_ts_algo(const char *name)
109{
110 struct ts_ops *o;
111
112 rcu_read_lock();
113 list_for_each_entry_rcu(o, &ts_ops, list) {
114 if (!strcmp(name, o->name)) {
115 if (!try_module_get(o->owner))
116 o = NULL;
117 rcu_read_unlock();
118 return o;
119 }
120 }
121 rcu_read_unlock();
122
123 return NULL;
124}
125
126/**
127 * textsearch_register - register a textsearch module
128 * @ops: operations lookup table
129 *
130 * This function must be called by textsearch modules to announce
131 * their presence. The specified &@ops must have %name set to a
132 * unique identifier and the callbacks find(), init(), get_pattern(),
133 * and get_pattern_len() must be implemented.
134 *
135 * Returns 0 or -EEXISTS if another module has already registered
136 * with same name.
137 */
138int textsearch_register(struct ts_ops *ops)
139{
140 int err = -EEXIST;
141 struct ts_ops *o;
142
143 if (ops->name == NULL || ops->find == NULL || ops->init == NULL ||
144 ops->get_pattern == NULL || ops->get_pattern_len == NULL)
145 return -EINVAL;
146
147 spin_lock(&ts_mod_lock);
148 list_for_each_entry(o, &ts_ops, list) {
149 if (!strcmp(ops->name, o->name))
150 goto errout;
151 }
152
153 list_add_tail_rcu(&ops->list, &ts_ops);
154 err = 0;
155errout:
156 spin_unlock(&ts_mod_lock);
157 return err;
158}
159
160/**
161 * textsearch_unregister - unregister a textsearch module
162 * @ops: operations lookup table
163 *
164 * This function must be called by textsearch modules to announce
165 * their disappearance for examples when the module gets unloaded.
166 * The &ops parameter must be the same as the one during the
167 * registration.
168 *
169 * Returns 0 on success or -ENOENT if no matching textsearch
170 * registration was found.
171 */
172int textsearch_unregister(struct ts_ops *ops)
173{
174 int err = 0;
175 struct ts_ops *o;
176
177 spin_lock(&ts_mod_lock);
178 list_for_each_entry(o, &ts_ops, list) {
179 if (o == ops) {
180 list_del_rcu(&o->list);
181 goto out;
182 }
183 }
184
185 err = -ENOENT;
186out:
187 spin_unlock(&ts_mod_lock);
188 return err;
189}
190
191struct ts_linear_state
192{
193 unsigned int len;
194 const void *data;
195};
196
197static unsigned int get_linear_data(unsigned int consumed, const u8 **dst,
198 struct ts_config *conf,
199 struct ts_state *state)
200{
201 struct ts_linear_state *st = (struct ts_linear_state *) state->cb;
202
203 if (likely(consumed < st->len)) {
204 *dst = st->data + consumed;
205 return st->len - consumed;
206 }
207
208 return 0;
209}
210
211/**
212 * textsearch_find_continuous - search a pattern in continuous/linear data
213 * @conf: search configuration
214 * @state: search state
215 * @data: data to search in
216 * @len: length of data
217 *
218 * A simplified version of textsearch_find() for continuous/linear data.
219 * Call textsearch_next() to retrieve subsequent matches.
220 *
221 * Returns the position of first occurrence of the pattern or
222 * UINT_MAX if no occurrence was found.
223 */
224unsigned int textsearch_find_continuous(struct ts_config *conf,
225 struct ts_state *state,
226 const void *data, unsigned int len)
227{
228 struct ts_linear_state *st = (struct ts_linear_state *) state->cb;
229
230 conf->get_next_block = get_linear_data;
231 st->data = data;
232 st->len = len;
233
234 return textsearch_find(conf, state);
235}
236
237/**
238 * textsearch_prepare - Prepare a search
239 * @algo: name of search algorithm
240 * @pattern: pattern data
241 * @len: length of pattern
242 * @gfp_mask: allocation mask
243 * @flags: search flags
244 *
245 * Looks up the search algorithm module and creates a new textsearch
246 * configuration for the specified pattern. Upon completion all
247 * necessary refcnts are held and the configuration must be put back
248 * using textsearch_put() after usage.
249 *
250 * Note: The format of the pattern may not be compatible between
251 * the various search algorithms.
252 *
253 * Returns a new textsearch configuration according to the specified
254 * parameters or a ERR_PTR().
255 */
256struct ts_config *textsearch_prepare(const char *algo, const void *pattern,
257 unsigned int len, int gfp_mask, int flags)
258{
259 int err = -ENOENT;
260 struct ts_config *conf;
261 struct ts_ops *ops;
262
263 ops = lookup_ts_algo(algo);
264#ifdef CONFIG_KMOD
265 /*
266 * Why not always autoload you may ask. Some users are
267 * in a situation where requesting a module may deadlock,
268 * especially when the module is located on a NFS mount.
269 */
270 if (ops == NULL && flags & TS_AUTOLOAD) {
271 request_module("ts_%s", algo);
272 ops = lookup_ts_algo(algo);
273 }
274#endif
275
276 if (ops == NULL)
277 goto errout;
278
279 conf = ops->init(pattern, len, gfp_mask);
280 if (IS_ERR(conf)) {
281 err = PTR_ERR(conf);
282 goto errout;
283 }
284
285 conf->ops = ops;
286 return conf;
287
288errout:
289 if (ops)
290 module_put(ops->owner);
291
292 return ERR_PTR(err);
293}
294
295/**
296 * textsearch_destroy - destroy a search configuration
297 * @conf: search configuration
298 *
299 * Releases all references of the configuration and frees
300 * up the memory.
301 */
302void textsearch_destroy(struct ts_config *conf)
303{
304 if (conf->ops) {
305 if (conf->ops->destroy)
306 conf->ops->destroy(conf);
307 module_put(conf->ops->owner);
308 }
309
310 kfree(conf);
311}
312
313EXPORT_SYMBOL(textsearch_register);
314EXPORT_SYMBOL(textsearch_unregister);
315EXPORT_SYMBOL(textsearch_prepare);
316EXPORT_SYMBOL(textsearch_find_continuous);
317EXPORT_SYMBOL(textsearch_destroy);
generated by cgit v1.2.2 (git 2.25.0) at 2024-11-08 20:39:01 -0500