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 <stdint.h>
26 #include <string.h>
27
28 static char *
twobyte_memmem(const unsigned char * h,size_t k,const unsigned char * n)29 twobyte_memmem(const unsigned char *h, size_t k, const unsigned char *n)
30 {
31 uint16_t nw = n[0] << 8 | n[1], hw = h[0] << 8 | h[1];
32 for (h += 2, k -= 2; k; k--, hw = hw << 8 | *h++)
33 if (hw == nw)
34 return (char *)h - 2;
35 return hw == nw ? (char *)h - 2 : 0;
36 }
37
38 static char *
threebyte_memmem(const unsigned char * h,size_t k,const unsigned char * n)39 threebyte_memmem(const unsigned char *h, size_t k, const unsigned char *n)
40 {
41 uint32_t nw = (uint32_t)n[0] << 24 | n[1] << 16 | n[2] << 8;
42 uint32_t hw = (uint32_t)h[0] << 24 | h[1] << 16 | h[2] << 8;
43 for (h += 3, k -= 3; k; k--, hw = (hw | *h++) << 8)
44 if (hw == nw)
45 return (char *)h - 3;
46 return hw == nw ? (char *)h - 3 : 0;
47 }
48
49 static char *
fourbyte_memmem(const unsigned char * h,size_t k,const unsigned char * n)50 fourbyte_memmem(const unsigned char *h, size_t k, const unsigned char *n)
51 {
52 uint32_t nw = (uint32_t)n[0] << 24 | n[1] << 16 | n[2] << 8 | n[3];
53 uint32_t hw = (uint32_t)h[0] << 24 | h[1] << 16 | h[2] << 8 | h[3];
54 for (h += 4, k -= 4; k; k--, hw = hw << 8 | *h++)
55 if (hw == nw)
56 return (char *)h - 4;
57 return hw == nw ? (char *)h - 4 : 0;
58 }
59
60 #define MAX(a, b) ((a) > (b) ? (a) : (b))
61 #define MIN(a, b) ((a) < (b) ? (a) : (b))
62
63 #define BITOP(a, b, op) \
64 ((a)[(size_t)(b) / (8 * sizeof *(a))] op \
65 (size_t)1 << ((size_t)(b) % (8 * sizeof *(a))))
66
67 /*
68 * Two Way string search algorithm, with a bad shift table applied to the last
69 * byte of the window. A bit array marks which entries in the shift table are
70 * initialized to avoid fully initializing a 1kb/2kb table.
71 *
72 * Reference: CROCHEMORE M., PERRIN D., 1991, Two-way string-matching,
73 * Journal of the ACM 38(3):651-675
74 */
75 static char *
twoway_memmem(const unsigned char * h,const unsigned char * z,const unsigned char * n,size_t l)76 twoway_memmem(const unsigned char *h, const unsigned char *z,
77 const unsigned char *n, size_t l)
78 {
79 size_t i, ip, jp, k, p, ms, p0, mem, mem0;
80 size_t byteset[32 / sizeof(size_t)] = { 0 };
81 size_t shift[256];
82
83 /* Computing length of needle and fill shift table */
84 for (i = 0; i < l; i++)
85 BITOP(byteset, n[i], |=), shift[n[i]] = i + 1;
86
87 /* Compute maximal suffix */
88 ip = -1;
89 jp = 0;
90 k = p = 1;
91 while (jp + k < l) {
92 if (n[ip + k] == n[jp + k]) {
93 if (k == p) {
94 jp += p;
95 k = 1;
96 } else
97 k++;
98 } else if (n[ip + k] > n[jp + k]) {
99 jp += k;
100 k = 1;
101 p = jp - ip;
102 } else {
103 ip = jp++;
104 k = p = 1;
105 }
106 }
107 ms = ip;
108 p0 = p;
109
110 /* And with the opposite comparison */
111 ip = -1;
112 jp = 0;
113 k = p = 1;
114 while (jp + k < l) {
115 if (n[ip + k] == n[jp + k]) {
116 if (k == p) {
117 jp += p;
118 k = 1;
119 } else
120 k++;
121 } else if (n[ip + k] < n[jp + k]) {
122 jp += k;
123 k = 1;
124 p = jp - ip;
125 } else {
126 ip = jp++;
127 k = p = 1;
128 }
129 }
130 if (ip + 1 > ms + 1)
131 ms = ip;
132 else
133 p = p0;
134
135 /* Periodic needle? */
136 if (memcmp(n, n + p, ms + 1)) {
137 mem0 = 0;
138 p = MAX(ms, l - ms - 1) + 1;
139 } else
140 mem0 = l - p;
141 mem = 0;
142
143 /* Search loop */
144 for (;;) {
145 /* If remainder of haystack is shorter than needle, done */
146 if (z - h < l)
147 return 0;
148
149 /* Check last byte first; advance by shift on mismatch */
150 if (BITOP(byteset, h[l - 1], &)) {
151 k = l - shift[h[l - 1]];
152 if (k) {
153 if (k < mem)
154 k = mem;
155 h += k;
156 mem = 0;
157 continue;
158 }
159 } else {
160 h += l;
161 mem = 0;
162 continue;
163 }
164
165 /* Compare right half */
166 for (k = MAX(ms + 1, mem); k < l && n[k] == h[k]; k++)
167 ;
168 if (k < l) {
169 h += k - ms;
170 mem = 0;
171 continue;
172 }
173 /* Compare left half */
174 for (k = ms + 1; k > mem && n[k - 1] == h[k - 1]; k--)
175 ;
176 if (k <= mem)
177 return (char *)h;
178 h += p;
179 mem = mem0;
180 }
181 }
182
183 void *
memmem(const void * h0,size_t k,const void * n0,size_t l)184 memmem(const void *h0, size_t k, const void *n0, size_t l)
185 {
186 const unsigned char *h = h0, *n = n0;
187
188 /* Return immediately on empty needle */
189 if (!l)
190 return (void *)h;
191
192 /* Return immediately when needle is longer than haystack */
193 if (k < l)
194 return 0;
195
196 /* Use faster algorithms for short needles */
197 h = memchr(h0, *n, k);
198 if (!h || l == 1)
199 return (void *)h;
200 k -= h - (const unsigned char *)h0;
201 if (k < l)
202 return 0;
203 if (l == 2)
204 return twobyte_memmem(h, k, n);
205 if (l == 3)
206 return threebyte_memmem(h, k, n);
207 if (l == 4)
208 return fourbyte_memmem(h, k, n);
209
210 return twoway_memmem(h, h + k, n, l);
211 }
212