1 // SPDX-License-Identifier: GPL-2.0-or-later <<
2 /* 1 /*
3 * lib/ts_kmp.c Knuth-Morris-Pratt tex 2 * lib/ts_kmp.c Knuth-Morris-Pratt text search implementation
4 * 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 *
5 * Authors: Thomas Graf <tgraf@suug.ch> 9 * Authors: Thomas Graf <tgraf@suug.ch>
6 * 10 *
7 * =========================================== 11 * ==========================================================================
8 * 12 *
9 * Implements a linear-time string-matching 13 * Implements a linear-time string-matching algorithm due to Knuth,
10 * Morris, and Pratt [1]. Their algorithm av 14 * Morris, and Pratt [1]. Their algorithm avoids the explicit
11 * computation of the transition function DE 15 * computation of the transition function DELTA altogether. Its
12 * matching time is O(n), for n being length 16 * matching time is O(n), for n being length(text), using just an
13 * auxiliary function PI[1..m], for m being 17 * auxiliary function PI[1..m], for m being length(pattern),
14 * precomputed from the pattern in time O(m) 18 * precomputed from the pattern in time O(m). The array PI allows
15 * the transition function DELTA to be compu 19 * the transition function DELTA to be computed efficiently
16 * "on the fly" as needed. Roughly speaking, 20 * "on the fly" as needed. Roughly speaking, for any state
17 * "q" = 0,1,...,m and any character "a" in 21 * "q" = 0,1,...,m and any character "a" in SIGMA, the value
18 * PI["q"] contains the information that is 22 * PI["q"] contains the information that is independent of "a" and
19 * is needed to compute DELTA("q", "a") [2]. 23 * is needed to compute DELTA("q", "a") [2]. Since the array PI
20 * has only m entries, whereas DELTA has O(m 24 * has only m entries, whereas DELTA has O(m|SIGMA|) entries, we
21 * save a factor of |SIGMA| in the preproces 25 * save a factor of |SIGMA| in the preprocessing time by computing
22 * PI rather than DELTA. 26 * PI rather than DELTA.
23 * 27 *
24 * [1] Cormen, Leiserson, Rivest, Stein 28 * [1] Cormen, Leiserson, Rivest, Stein
25 * Introdcution to Algorithms, 2nd Editi 29 * Introdcution to Algorithms, 2nd Edition, MIT Press
26 * [2] See finite automaton theory 30 * [2] See finite automaton theory
27 */ 31 */
28 32
29 #include <linux/module.h> 33 #include <linux/module.h>
30 #include <linux/types.h> 34 #include <linux/types.h>
31 #include <linux/string.h> 35 #include <linux/string.h>
32 #include <linux/ctype.h> 36 #include <linux/ctype.h>
33 #include <linux/textsearch.h> 37 #include <linux/textsearch.h>
34 38
35 struct ts_kmp 39 struct ts_kmp
36 { 40 {
37 u8 * pattern; 41 u8 * pattern;
38 unsigned int pattern_len; 42 unsigned int pattern_len;
39 unsigned int prefix_tbl[]; !! 43 unsigned int prefix_tbl[0];
40 }; 44 };
41 45
42 static unsigned int kmp_find(struct ts_config 46 static unsigned int kmp_find(struct ts_config *conf, struct ts_state *state)
43 { 47 {
44 struct ts_kmp *kmp = ts_config_priv(co 48 struct ts_kmp *kmp = ts_config_priv(conf);
45 unsigned int i, q = 0, text_len, consu 49 unsigned int i, q = 0, text_len, consumed = state->offset;
46 const u8 *text; 50 const u8 *text;
47 const int icase = conf->flags & TS_IGN 51 const int icase = conf->flags & TS_IGNORECASE;
48 52
49 for (;;) { 53 for (;;) {
50 text_len = conf->get_next_bloc 54 text_len = conf->get_next_block(consumed, &text, conf, state);
51 55
52 if (unlikely(text_len == 0)) 56 if (unlikely(text_len == 0))
53 break; 57 break;
54 58
55 for (i = 0; i < text_len; i++) 59 for (i = 0; i < text_len; i++) {
56 while (q > 0 && kmp->p 60 while (q > 0 && kmp->pattern[q]
57 != (icase ? touppe 61 != (icase ? toupper(text[i]) : text[i]))
58 q = kmp->prefi 62 q = kmp->prefix_tbl[q - 1];
59 if (kmp->pattern[q] 63 if (kmp->pattern[q]
60 == (icase ? touppe 64 == (icase ? toupper(text[i]) : text[i]))
61 q++; 65 q++;
62 if (unlikely(q == kmp- 66 if (unlikely(q == kmp->pattern_len)) {
63 state->offset 67 state->offset = consumed + i + 1;
64 return state-> 68 return state->offset - kmp->pattern_len;
65 } 69 }
66 } 70 }
67 71
68 consumed += text_len; 72 consumed += text_len;
69 } 73 }
70 74
71 return UINT_MAX; 75 return UINT_MAX;
72 } 76 }
73 77
74 static inline void compute_prefix_tbl(const u8 78 static inline void compute_prefix_tbl(const u8 *pattern, unsigned int len,
75 unsigned 79 unsigned int *prefix_tbl, int flags)
76 { 80 {
77 unsigned int k, q; 81 unsigned int k, q;
78 const u8 icase = flags & TS_IGNORECASE 82 const u8 icase = flags & TS_IGNORECASE;
79 83
80 for (k = 0, q = 1; q < len; q++) { 84 for (k = 0, q = 1; q < len; q++) {
81 while (k > 0 && (icase ? toupp 85 while (k > 0 && (icase ? toupper(pattern[k]) : pattern[k])
82 != (icase ? toupper(patter 86 != (icase ? toupper(pattern[q]) : pattern[q]))
83 k = prefix_tbl[k-1]; 87 k = prefix_tbl[k-1];
84 if ((icase ? toupper(pattern[k 88 if ((icase ? toupper(pattern[k]) : pattern[k])
85 == (icase ? toupper(patter 89 == (icase ? toupper(pattern[q]) : pattern[q]))
86 k++; 90 k++;
87 prefix_tbl[q] = k; 91 prefix_tbl[q] = k;
88 } 92 }
89 } 93 }
90 94
91 static struct ts_config *kmp_init(const void * 95 static struct ts_config *kmp_init(const void *pattern, unsigned int len,
92 gfp_t gfp_ma 96 gfp_t gfp_mask, int flags)
93 { 97 {
94 struct ts_config *conf; 98 struct ts_config *conf;
95 struct ts_kmp *kmp; 99 struct ts_kmp *kmp;
96 int i; 100 int i;
97 unsigned int prefix_tbl_len = len * si 101 unsigned int prefix_tbl_len = len * sizeof(unsigned int);
98 size_t priv_size = sizeof(*kmp) + len 102 size_t priv_size = sizeof(*kmp) + len + prefix_tbl_len;
99 103
100 conf = alloc_ts_config(priv_size, gfp_ 104 conf = alloc_ts_config(priv_size, gfp_mask);
101 if (IS_ERR(conf)) 105 if (IS_ERR(conf))
102 return conf; 106 return conf;
103 107
104 conf->flags = flags; 108 conf->flags = flags;
105 kmp = ts_config_priv(conf); 109 kmp = ts_config_priv(conf);
106 kmp->pattern_len = len; 110 kmp->pattern_len = len;
107 compute_prefix_tbl(pattern, len, kmp-> 111 compute_prefix_tbl(pattern, len, kmp->prefix_tbl, flags);
108 kmp->pattern = (u8 *) kmp->prefix_tbl 112 kmp->pattern = (u8 *) kmp->prefix_tbl + prefix_tbl_len;
109 if (flags & TS_IGNORECASE) 113 if (flags & TS_IGNORECASE)
110 for (i = 0; i < len; i++) 114 for (i = 0; i < len; i++)
111 kmp->pattern[i] = toup 115 kmp->pattern[i] = toupper(((u8 *)pattern)[i]);
112 else 116 else
113 memcpy(kmp->pattern, pattern, 117 memcpy(kmp->pattern, pattern, len);
114 118
115 return conf; 119 return conf;
116 } 120 }
117 121
118 static void *kmp_get_pattern(struct ts_config 122 static void *kmp_get_pattern(struct ts_config *conf)
119 { 123 {
120 struct ts_kmp *kmp = ts_config_priv(co 124 struct ts_kmp *kmp = ts_config_priv(conf);
121 return kmp->pattern; 125 return kmp->pattern;
122 } 126 }
123 127
124 static unsigned int kmp_get_pattern_len(struct 128 static unsigned int kmp_get_pattern_len(struct ts_config *conf)
125 { 129 {
126 struct ts_kmp *kmp = ts_config_priv(co 130 struct ts_kmp *kmp = ts_config_priv(conf);
127 return kmp->pattern_len; 131 return kmp->pattern_len;
128 } 132 }
129 133
130 static struct ts_ops kmp_ops = { 134 static struct ts_ops kmp_ops = {
131 .name = "kmp", 135 .name = "kmp",
132 .find = kmp_find, 136 .find = kmp_find,
133 .init = kmp_init, 137 .init = kmp_init,
134 .get_pattern = kmp_get_pattern, 138 .get_pattern = kmp_get_pattern,
135 .get_pattern_len = kmp_get_pattern_le 139 .get_pattern_len = kmp_get_pattern_len,
136 .owner = THIS_MODULE, 140 .owner = THIS_MODULE,
137 .list = LIST_HEAD_INIT(kmp 141 .list = LIST_HEAD_INIT(kmp_ops.list)
138 }; 142 };
139 143
140 static int __init init_kmp(void) 144 static int __init init_kmp(void)
141 { 145 {
142 return textsearch_register(&kmp_ops); 146 return textsearch_register(&kmp_ops);
143 } 147 }
144 148
145 static void __exit exit_kmp(void) 149 static void __exit exit_kmp(void)
146 { 150 {
147 textsearch_unregister(&kmp_ops); 151 textsearch_unregister(&kmp_ops);
148 } 152 }
149 153
150 MODULE_DESCRIPTION("Knuth-Morris-Pratt text se <<
151 MODULE_LICENSE("GPL"); 154 MODULE_LICENSE("GPL");
152 155
153 module_init(init_kmp); 156 module_init(init_kmp);
154 module_exit(exit_kmp); 157 module_exit(exit_kmp);
155 158
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.