blob: 69557c74ef9f8e4dbe1994dce8d1099456717cf2 [file] [log] [blame]
Thomas Graf6408f792005-06-23 20:59:16 -07001/*
2 * lib/ts_fsm.c A naive finite state machine text search approach
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 *
11 * ==========================================================================
12 *
13 * A finite state machine consists of n states (struct ts_fsm_token)
Randy Dunlap7433a8d2017-10-20 12:15:52 -070014 * representing the pattern as a finite automaton. The data is read
Andreas Mohr1497b272006-09-29 02:01:28 -070015 * sequentially on an octet basis. Every state token specifies the number
Thomas Graf6408f792005-06-23 20:59:16 -070016 * of recurrences and the type of value accepted which can be either a
17 * specific character or ctype based set of characters. The available
18 * type of recurrences include 1, (0|1), [0 n], and [1 n].
19 *
Andreas Mohr1497b272006-09-29 02:01:28 -070020 * The algorithm differs between strict/non-strict mode specifying
21 * whether the pattern has to start at the first octet. Strict mode
Thomas Graf6408f792005-06-23 20:59:16 -070022 * is enabled by default and can be disabled by inserting
23 * TS_FSM_HEAD_IGNORE as the first token in the chain.
24 *
25 * The runtime performance of the algorithm should be around O(n),
26 * however while in strict mode the average runtime can be better.
27 */
28
Thomas Graf6408f792005-06-23 20:59:16 -070029#include <linux/module.h>
30#include <linux/types.h>
31#include <linux/string.h>
32#include <linux/ctype.h>
33#include <linux/textsearch.h>
34#include <linux/textsearch_fsm.h>
35
36struct ts_fsm
37{
38 unsigned int ntokens;
39 struct ts_fsm_token tokens[0];
40};
41
42/* other values derived from ctype.h */
43#define _A 0x100 /* ascii */
44#define _W 0x200 /* wildcard */
45
46/* Map to _ctype flags and some magic numbers */
Andreas Mohr1497b272006-09-29 02:01:28 -070047static const u16 token_map[TS_FSM_TYPE_MAX+1] = {
Thomas Graf6408f792005-06-23 20:59:16 -070048 [TS_FSM_SPECIFIC] = 0,
49 [TS_FSM_WILDCARD] = _W,
50 [TS_FSM_CNTRL] = _C,
51 [TS_FSM_LOWER] = _L,
52 [TS_FSM_UPPER] = _U,
53 [TS_FSM_PUNCT] = _P,
54 [TS_FSM_SPACE] = _S,
55 [TS_FSM_DIGIT] = _D,
56 [TS_FSM_XDIGIT] = _D | _X,
57 [TS_FSM_ALPHA] = _U | _L,
58 [TS_FSM_ALNUM] = _U | _L | _D,
59 [TS_FSM_PRINT] = _P | _U | _L | _D | _SP,
60 [TS_FSM_GRAPH] = _P | _U | _L | _D,
61 [TS_FSM_ASCII] = _A,
62};
63
Andreas Mohr1497b272006-09-29 02:01:28 -070064static const u16 token_lookup_tbl[256] = {
Thomas Graf6408f792005-06-23 20:59:16 -070065_W|_A|_C, _W|_A|_C, _W|_A|_C, _W|_A|_C, /* 0- 3 */
66_W|_A|_C, _W|_A|_C, _W|_A|_C, _W|_A|_C, /* 4- 7 */
67_W|_A|_C, _W|_A|_C|_S, _W|_A|_C|_S, _W|_A|_C|_S, /* 8- 11 */
68_W|_A|_C|_S, _W|_A|_C|_S, _W|_A|_C, _W|_A|_C, /* 12- 15 */
69_W|_A|_C, _W|_A|_C, _W|_A|_C, _W|_A|_C, /* 16- 19 */
70_W|_A|_C, _W|_A|_C, _W|_A|_C, _W|_A|_C, /* 20- 23 */
71_W|_A|_C, _W|_A|_C, _W|_A|_C, _W|_A|_C, /* 24- 27 */
72_W|_A|_C, _W|_A|_C, _W|_A|_C, _W|_A|_C, /* 28- 31 */
73_W|_A|_S|_SP, _W|_A|_P, _W|_A|_P, _W|_A|_P, /* 32- 35 */
74_W|_A|_P, _W|_A|_P, _W|_A|_P, _W|_A|_P, /* 36- 39 */
75_W|_A|_P, _W|_A|_P, _W|_A|_P, _W|_A|_P, /* 40- 43 */
76_W|_A|_P, _W|_A|_P, _W|_A|_P, _W|_A|_P, /* 44- 47 */
77_W|_A|_D, _W|_A|_D, _W|_A|_D, _W|_A|_D, /* 48- 51 */
78_W|_A|_D, _W|_A|_D, _W|_A|_D, _W|_A|_D, /* 52- 55 */
79_W|_A|_D, _W|_A|_D, _W|_A|_P, _W|_A|_P, /* 56- 59 */
80_W|_A|_P, _W|_A|_P, _W|_A|_P, _W|_A|_P, /* 60- 63 */
81_W|_A|_P, _W|_A|_U|_X, _W|_A|_U|_X, _W|_A|_U|_X, /* 64- 67 */
82_W|_A|_U|_X, _W|_A|_U|_X, _W|_A|_U|_X, _W|_A|_U, /* 68- 71 */
83_W|_A|_U, _W|_A|_U, _W|_A|_U, _W|_A|_U, /* 72- 75 */
84_W|_A|_U, _W|_A|_U, _W|_A|_U, _W|_A|_U, /* 76- 79 */
85_W|_A|_U, _W|_A|_U, _W|_A|_U, _W|_A|_U, /* 80- 83 */
86_W|_A|_U, _W|_A|_U, _W|_A|_U, _W|_A|_U, /* 84- 87 */
87_W|_A|_U, _W|_A|_U, _W|_A|_U, _W|_A|_P, /* 88- 91 */
88_W|_A|_P, _W|_A|_P, _W|_A|_P, _W|_A|_P, /* 92- 95 */
89_W|_A|_P, _W|_A|_L|_X, _W|_A|_L|_X, _W|_A|_L|_X, /* 96- 99 */
90_W|_A|_L|_X, _W|_A|_L|_X, _W|_A|_L|_X, _W|_A|_L, /* 100-103 */
91_W|_A|_L, _W|_A|_L, _W|_A|_L, _W|_A|_L, /* 104-107 */
92_W|_A|_L, _W|_A|_L, _W|_A|_L, _W|_A|_L, /* 108-111 */
93_W|_A|_L, _W|_A|_L, _W|_A|_L, _W|_A|_L, /* 112-115 */
94_W|_A|_L, _W|_A|_L, _W|_A|_L, _W|_A|_L, /* 116-119 */
95_W|_A|_L, _W|_A|_L, _W|_A|_L, _W|_A|_P, /* 120-123 */
96_W|_A|_P, _W|_A|_P, _W|_A|_P, _W|_A|_C, /* 124-127 */
97_W, _W, _W, _W, /* 128-131 */
98_W, _W, _W, _W, /* 132-135 */
99_W, _W, _W, _W, /* 136-139 */
100_W, _W, _W, _W, /* 140-143 */
101_W, _W, _W, _W, /* 144-147 */
102_W, _W, _W, _W, /* 148-151 */
103_W, _W, _W, _W, /* 152-155 */
104_W, _W, _W, _W, /* 156-159 */
105_W|_S|_SP, _W|_P, _W|_P, _W|_P, /* 160-163 */
106_W|_P, _W|_P, _W|_P, _W|_P, /* 164-167 */
107_W|_P, _W|_P, _W|_P, _W|_P, /* 168-171 */
108_W|_P, _W|_P, _W|_P, _W|_P, /* 172-175 */
109_W|_P, _W|_P, _W|_P, _W|_P, /* 176-179 */
110_W|_P, _W|_P, _W|_P, _W|_P, /* 180-183 */
111_W|_P, _W|_P, _W|_P, _W|_P, /* 184-187 */
112_W|_P, _W|_P, _W|_P, _W|_P, /* 188-191 */
113_W|_U, _W|_U, _W|_U, _W|_U, /* 192-195 */
114_W|_U, _W|_U, _W|_U, _W|_U, /* 196-199 */
115_W|_U, _W|_U, _W|_U, _W|_U, /* 200-203 */
116_W|_U, _W|_U, _W|_U, _W|_U, /* 204-207 */
117_W|_U, _W|_U, _W|_U, _W|_U, /* 208-211 */
118_W|_U, _W|_U, _W|_U, _W|_P, /* 212-215 */
119_W|_U, _W|_U, _W|_U, _W|_U, /* 216-219 */
120_W|_U, _W|_U, _W|_U, _W|_L, /* 220-223 */
121_W|_L, _W|_L, _W|_L, _W|_L, /* 224-227 */
122_W|_L, _W|_L, _W|_L, _W|_L, /* 228-231 */
123_W|_L, _W|_L, _W|_L, _W|_L, /* 232-235 */
124_W|_L, _W|_L, _W|_L, _W|_L, /* 236-239 */
125_W|_L, _W|_L, _W|_L, _W|_L, /* 240-243 */
126_W|_L, _W|_L, _W|_L, _W|_P, /* 244-247 */
127_W|_L, _W|_L, _W|_L, _W|_L, /* 248-251 */
128_W|_L, _W|_L, _W|_L, _W|_L}; /* 252-255 */
129
130static inline int match_token(struct ts_fsm_token *t, u8 d)
131{
132 if (t->type)
133 return (token_lookup_tbl[d] & t->type) != 0;
134 else
135 return t->value == d;
136}
137
138static unsigned int fsm_find(struct ts_config *conf, struct ts_state *state)
139{
140 struct ts_fsm *fsm = ts_config_priv(conf);
141 struct ts_fsm_token *cur = NULL, *next;
142 unsigned int match_start, block_idx = 0, tok_idx;
143 unsigned block_len = 0, strict, consumed = state->offset;
144 const u8 *data;
145
146#define GET_NEXT_BLOCK() \
147({ consumed += block_idx; \
148 block_idx = 0; \
149 block_len = conf->get_next_block(consumed, &data, conf, state); })
150
151#define TOKEN_MISMATCH() \
152 do { \
153 if (strict) \
154 goto no_match; \
155 block_idx++; \
156 goto startover; \
157 } while(0)
158
159#define end_of_data() unlikely(block_idx >= block_len && !GET_NEXT_BLOCK())
160
161 if (end_of_data())
162 goto no_match;
163
164 strict = fsm->tokens[0].recur != TS_FSM_HEAD_IGNORE;
165
166startover:
167 match_start = consumed + block_idx;
168
169 for (tok_idx = 0; tok_idx < fsm->ntokens; tok_idx++) {
170 cur = &fsm->tokens[tok_idx];
171
172 if (likely(tok_idx < (fsm->ntokens - 1)))
173 next = &fsm->tokens[tok_idx + 1];
174 else
175 next = NULL;
176
177 switch (cur->recur) {
178 case TS_FSM_SINGLE:
179 if (end_of_data())
180 goto no_match;
181
182 if (!match_token(cur, data[block_idx]))
183 TOKEN_MISMATCH();
184 break;
185
186 case TS_FSM_PERHAPS:
187 if (end_of_data() ||
188 !match_token(cur, data[block_idx]))
189 continue;
190 break;
191
192 case TS_FSM_MULTI:
193 if (end_of_data())
194 goto no_match;
195
196 if (!match_token(cur, data[block_idx]))
197 TOKEN_MISMATCH();
198
199 block_idx++;
200 /* fall through */
201
202 case TS_FSM_ANY:
203 if (next == NULL)
204 goto found_match;
205
206 if (end_of_data())
207 continue;
208
209 while (!match_token(next, data[block_idx])) {
210 if (!match_token(cur, data[block_idx]))
211 TOKEN_MISMATCH();
212 block_idx++;
213 if (end_of_data())
214 goto no_match;
215 }
216 continue;
217
218 /*
219 * Optimization: Prefer small local loop over jumping
220 * back and forth until garbage at head is munched.
221 */
222 case TS_FSM_HEAD_IGNORE:
223 if (end_of_data())
224 continue;
225
226 while (!match_token(next, data[block_idx])) {
227 /*
228 * Special case, don't start over upon
229 * a mismatch, give the user the
230 * chance to specify the type of data
231 * allowed to be ignored.
232 */
233 if (!match_token(cur, data[block_idx]))
234 goto no_match;
235
236 block_idx++;
237 if (end_of_data())
238 goto no_match;
239 }
240
241 match_start = consumed + block_idx;
242 continue;
243 }
244
245 block_idx++;
246 }
247
248 if (end_of_data())
249 goto found_match;
250
251no_match:
252 return UINT_MAX;
253
254found_match:
255 state->offset = consumed + block_idx;
256 return match_start;
257}
258
259static struct ts_config *fsm_init(const void *pattern, unsigned int len,
Joonwoo Park43138832008-07-08 02:38:27 -0700260 gfp_t gfp_mask, int flags)
Thomas Graf6408f792005-06-23 20:59:16 -0700261{
262 int i, err = -EINVAL;
263 struct ts_config *conf;
264 struct ts_fsm *fsm;
265 struct ts_fsm_token *tokens = (struct ts_fsm_token *) pattern;
266 unsigned int ntokens = len / sizeof(*tokens);
267 size_t priv_size = sizeof(*fsm) + len;
268
269 if (len % sizeof(struct ts_fsm_token) || ntokens < 1)
270 goto errout;
271
Joonwoo Park43138832008-07-08 02:38:27 -0700272 if (flags & TS_IGNORECASE)
273 goto errout;
274
Thomas Graf6408f792005-06-23 20:59:16 -0700275 for (i = 0; i < ntokens; i++) {
276 struct ts_fsm_token *t = &tokens[i];
277
278 if (t->type > TS_FSM_TYPE_MAX || t->recur > TS_FSM_RECUR_MAX)
279 goto errout;
280
281 if (t->recur == TS_FSM_HEAD_IGNORE &&
282 (i != 0 || i == (ntokens - 1)))
283 goto errout;
284 }
285
286 conf = alloc_ts_config(priv_size, gfp_mask);
287 if (IS_ERR(conf))
288 return conf;
289
Joonwoo Park43138832008-07-08 02:38:27 -0700290 conf->flags = flags;
Thomas Graf6408f792005-06-23 20:59:16 -0700291 fsm = ts_config_priv(conf);
292 fsm->ntokens = ntokens;
293 memcpy(fsm->tokens, pattern, len);
294
295 for (i = 0; i < fsm->ntokens; i++) {
296 struct ts_fsm_token *t = &fsm->tokens[i];
297 t->type = token_map[t->type];
298 }
299
300 return conf;
301
302errout:
303 return ERR_PTR(err);
304}
305
306static void *fsm_get_pattern(struct ts_config *conf)
307{
308 struct ts_fsm *fsm = ts_config_priv(conf);
309 return fsm->tokens;
310}
311
312static unsigned int fsm_get_pattern_len(struct ts_config *conf)
313{
314 struct ts_fsm *fsm = ts_config_priv(conf);
315 return fsm->ntokens * sizeof(struct ts_fsm_token);
316}
317
318static struct ts_ops fsm_ops = {
319 .name = "fsm",
320 .find = fsm_find,
321 .init = fsm_init,
322 .get_pattern = fsm_get_pattern,
323 .get_pattern_len = fsm_get_pattern_len,
324 .owner = THIS_MODULE,
325 .list = LIST_HEAD_INIT(fsm_ops.list)
326};
327
328static int __init init_fsm(void)
329{
330 return textsearch_register(&fsm_ops);
331}
332
333static void __exit exit_fsm(void)
334{
335 textsearch_unregister(&fsm_ops);
336}
337
338MODULE_LICENSE("GPL");
339
340module_init(init_fsm);
341module_exit(exit_fsm);