1 /*-
2 * SPDX-License-Identifier: MIT
3 *
4 * Copyright (c) 2005-2014 Rich Felker, et al.
5 *
6 * Permission is hereby granted, free of charge, to any person obtaining
7 * a copy of this software and associated documentation files (the
8 * "Software"), to deal in the Software without restriction, including
9 * without limitation the rights to use, copy, modify, merge, publish,
10 * distribute, sublicense, and/or sell copies of the Software, and to
11 * permit persons to whom the Software is furnished to do so, subject to
12 * the following conditions:
13 *
14 * The above copyright notice and this permission notice shall be
15 * included in all copies or substantial portions of the Software.
16 *
17 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
18 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
19 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
20 * IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
21 * CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
22 * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
23 * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
24 */
25 #include <sys/cdefs.h>
26 __FBSDID("$FreeBSD$");
27
28 #include <stdint.h>
29 #include <string.h>
30
31 static char *
twobyte_memmem(const unsigned char * h,size_t k,const unsigned char * n)32 twobyte_memmem(const unsigned char *h, size_t k, const unsigned char *n)
33 {
34 uint16_t nw = n[0] << 8 | n[1], hw = h[0] << 8 | h[1];
35 for (h += 2, k -= 2; k; k--, hw = hw << 8 | *h++)
36 if (hw == nw)
37 return (char *)h - 2;
38 return hw == nw ? (char *)h - 2 : 0;
39 }
40
41 static char *
threebyte_memmem(const unsigned char * h,size_t k,const unsigned char * n)42 threebyte_memmem(const unsigned char *h, size_t k, const unsigned char *n)
43 {
44 uint32_t nw = (uint32_t)n[0] << 24 | n[1] << 16 | n[2] << 8;
45 uint32_t hw = (uint32_t)h[0] << 24 | h[1] << 16 | h[2] << 8;
46 for (h += 3, k -= 3; k; k--, hw = (hw | *h++) << 8)
47 if (hw == nw)
48 return (char *)h - 3;
49 return hw == nw ? (char *)h - 3 : 0;
50 }
51
52 static char *
fourbyte_memmem(const unsigned char * h,size_t k,const unsigned char * n)53 fourbyte_memmem(const unsigned char *h, size_t k, const unsigned char *n)
54 {
55 uint32_t nw = (uint32_t)n[0] << 24 | n[1] << 16 | n[2] << 8 | n[3];
56 uint32_t hw = (uint32_t)h[0] << 24 | h[1] << 16 | h[2] << 8 | h[3];
57 for (h += 4, k -= 4; k; k--, hw = hw << 8 | *h++)
58 if (hw == nw)
59 return (char *)h - 4;
60 return hw == nw ? (char *)h - 4 : 0;
61 }
62
63 #define MAX(a, b) ((a) > (b) ? (a) : (b))
64 #define MIN(a, b) ((a) < (b) ? (a) : (b))
65
66 #define BITOP(a, b, op) \
67 ((a)[(size_t)(b) / (8 * sizeof *(a))] op \
68 (size_t)1 << ((size_t)(b) % (8 * sizeof *(a))))
69
70 /*
71 * Two Way string search algorithm, with a bad shift table applied to the last
72 * byte of the window. A bit array marks which entries in the shift table are
73 * initialized to avoid fully initializing a 1kb/2kb table.
74 *
75 * Reference: CROCHEMORE M., PERRIN D., 1991, Two-way string-matching,
76 * Journal of the ACM 38(3):651-675
77 */
78 static char *
twoway_memmem(const unsigned char * h,const unsigned char * z,const unsigned char * n,size_t l)79 twoway_memmem(const unsigned char *h, const unsigned char *z,
80 const unsigned char *n, size_t l)
81 {
82 size_t i, ip, jp, k, p, ms, p0, mem, mem0;
83 size_t byteset[32 / sizeof(size_t)] = { 0 };
84 size_t shift[256];
85
86 /* Computing length of needle and fill shift table */
87 for (i = 0; i < l; i++)
88 BITOP(byteset, n[i], |=), shift[n[i]] = i + 1;
89
90 /* Compute maximal suffix */
91 ip = -1;
92 jp = 0;
93 k = p = 1;
94 while (jp + k < l) {
95 if (n[ip + k] == n[jp + k]) {
96 if (k == p) {
97 jp += p;
98 k = 1;
99 } else
100 k++;
101 } else if (n[ip + k] > n[jp + k]) {
102 jp += k;
103 k = 1;
104 p = jp - ip;
105 } else {
106 ip = jp++;
107 k = p = 1;
108 }
109 }
110 ms = ip;
111 p0 = p;
112
113 /* And with the opposite comparison */
114 ip = -1;
115 jp = 0;
116 k = p = 1;
117 while (jp + k < l) {
118 if (n[ip + k] == n[jp + k]) {
119 if (k == p) {
120 jp += p;
121 k = 1;
122 } else
123 k++;
124 } else if (n[ip + k] < n[jp + k]) {
125 jp += k;
126 k = 1;
127 p = jp - ip;
128 } else {
129 ip = jp++;
130 k = p = 1;
131 }
132 }
133 if (ip + 1 > ms + 1)
134 ms = ip;
135 else
136 p = p0;
137
138 /* Periodic needle? */
139 if (memcmp(n, n + p, ms + 1)) {
140 mem0 = 0;
141 p = MAX(ms, l - ms - 1) + 1;
142 } else
143 mem0 = l - p;
144 mem = 0;
145
146 /* Search loop */
147 for (;;) {
148 /* If remainder of haystack is shorter than needle, done */
149 if (z - h < l)
150 return 0;
151
152 /* Check last byte first; advance by shift on mismatch */
153 if (BITOP(byteset, h[l - 1], &)) {
154 k = l - shift[h[l - 1]];
155 if (k) {
156 if (k < mem)
157 k = mem;
158 h += k;
159 mem = 0;
160 continue;
161 }
162 } else {
163 h += l;
164 mem = 0;
165 continue;
166 }
167
168 /* Compare right half */
169 for (k = MAX(ms + 1, mem); k < l && n[k] == h[k]; k++)
170 ;
171 if (k < l) {
172 h += k - ms;
173 mem = 0;
174 continue;
175 }
176 /* Compare left half */
177 for (k = ms + 1; k > mem && n[k - 1] == h[k - 1]; k--)
178 ;
179 if (k <= mem)
180 return (char *)h;
181 h += p;
182 mem = mem0;
183 }
184 }
185
186 void *
memmem(const void * h0,size_t k,const void * n0,size_t l)187 memmem(const void *h0, size_t k, const void *n0, size_t l)
188 {
189 const unsigned char *h = h0, *n = n0;
190
191 /* Return immediately on empty needle */
192 if (!l)
193 return (void *)h;
194
195 /* Return immediately when needle is longer than haystack */
196 if (k < l)
197 return 0;
198
199 /* Use faster algorithms for short needles */
200 h = memchr(h0, *n, k);
201 if (!h || l == 1)
202 return (void *)h;
203 k -= h - (const unsigned char *)h0;
204 if (k < l)
205 return 0;
206 if (l == 2)
207 return twobyte_memmem(h, k, n);
208 if (l == 3)
209 return threebyte_memmem(h, k, n);
210 if (l == 4)
211 return fourbyte_memmem(h, k, n);
212
213 return twoway_memmem(h, h + k, n, l);
214 }
215