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_strstr(const unsigned char * h,const unsigned char * n)29 twobyte_strstr(const unsigned char *h, const unsigned char *n)
30 {
31 uint16_t nw = n[0] << 8 | n[1], hw = h[0] << 8 | h[1];
32 for (h++; *h && hw != nw; hw = hw << 8 | *++h)
33 ;
34 return *h ? (char *)h - 1 : 0;
35 }
36
37 static char *
threebyte_strstr(const unsigned char * h,const unsigned char * n)38 threebyte_strstr(const unsigned char *h, const unsigned char *n)
39 {
40 uint32_t nw = (uint32_t)n[0] << 24 | n[1] << 16 | n[2] << 8;
41 uint32_t hw = (uint32_t)h[0] << 24 | h[1] << 16 | h[2] << 8;
42 for (h += 2; *h && hw != nw; hw = (hw | *++h) << 8)
43 ;
44 return *h ? (char *)h - 2 : 0;
45 }
46
47 static char *
fourbyte_strstr(const unsigned char * h,const unsigned char * n)48 fourbyte_strstr(const unsigned char *h, const unsigned char *n)
49 {
50 uint32_t nw = (uint32_t)n[0] << 24 | n[1] << 16 | n[2] << 8 | n[3];
51 uint32_t hw = (uint32_t)h[0] << 24 | h[1] << 16 | h[2] << 8 | h[3];
52 for (h += 3; *h && hw != nw; hw = hw << 8 | *++h)
53 ;
54 return *h ? (char *)h - 3 : 0;
55 }
56
57 #define MAX(a, b) ((a) > (b) ? (a) : (b))
58 #define MIN(a, b) ((a) < (b) ? (a) : (b))
59
60 #define BITOP(a, b, op) \
61 ((a)[(size_t)(b) / (8 * sizeof *(a))] op \
62 (size_t)1 << ((size_t)(b) % (8 * sizeof *(a))))
63
64 /*
65 * Two Way string search algorithm, with a bad shift table applied to the last
66 * byte of the window. A bit array marks which entries in the shift table are
67 * initialized to avoid fully initializing a 1kb/2kb table.
68 *
69 * Reference: CROCHEMORE M., PERRIN D., 1991, Two-way string-matching,
70 * Journal of the ACM 38(3):651-675
71 */
72 static char *
twoway_strstr(const unsigned char * h,const unsigned char * n)73 twoway_strstr(const unsigned char *h, const unsigned char *n)
74 {
75 const unsigned char *z;
76 size_t l, ip, jp, k, p, ms, p0, mem, mem0;
77 size_t byteset[32 / sizeof(size_t)] = { 0 };
78 size_t shift[256];
79
80 /* Computing length of needle and fill shift table */
81 for (l = 0; n[l] && h[l]; l++)
82 BITOP(byteset, n[l], |=), shift[n[l]] = l + 1;
83 if (n[l])
84 return 0; /* hit the end of h */
85
86 /* Compute maximal suffix */
87 ip = -1;
88 jp = 0;
89 k = p = 1;
90 while (jp + k < l) {
91 if (n[ip + k] == n[jp + k]) {
92 if (k == p) {
93 jp += p;
94 k = 1;
95 } else
96 k++;
97 } else if (n[ip + k] > n[jp + k]) {
98 jp += k;
99 k = 1;
100 p = jp - ip;
101 } else {
102 ip = jp++;
103 k = p = 1;
104 }
105 }
106 ms = ip;
107 p0 = p;
108
109 /* And with the opposite comparison */
110 ip = -1;
111 jp = 0;
112 k = p = 1;
113 while (jp + k < l) {
114 if (n[ip + k] == n[jp + k]) {
115 if (k == p) {
116 jp += p;
117 k = 1;
118 } else
119 k++;
120 } else if (n[ip + k] < n[jp + k]) {
121 jp += k;
122 k = 1;
123 p = jp - ip;
124 } else {
125 ip = jp++;
126 k = p = 1;
127 }
128 }
129 if (ip + 1 > ms + 1)
130 ms = ip;
131 else
132 p = p0;
133
134 /* Periodic needle? */
135 if (memcmp(n, n + p, ms + 1)) {
136 mem0 = 0;
137 p = MAX(ms, l - ms - 1) + 1;
138 } else
139 mem0 = l - p;
140 mem = 0;
141
142 /* Initialize incremental end-of-haystack pointer */
143 z = h;
144
145 /* Search loop */
146 for (;;) {
147 /* Update incremental end-of-haystack pointer */
148 if (z - h < l) {
149 /* Fast estimate for MAX(l,63) */
150 size_t grow = l | 63;
151 const unsigned char *z2 = memchr(z, 0, grow);
152 if (z2) {
153 z = z2;
154 if (z - h < l)
155 return 0;
156 } else
157 z += grow;
158 }
159
160 /* Check last byte first; advance by shift on mismatch */
161 if (BITOP(byteset, h[l - 1], &)) {
162 k = l - shift[h[l - 1]];
163 if (k) {
164 if (k < mem)
165 k = mem;
166 h += k;
167 mem = 0;
168 continue;
169 }
170 } else {
171 h += l;
172 mem = 0;
173 continue;
174 }
175
176 /* Compare right half */
177 for (k = MAX(ms + 1, mem); n[k] && n[k] == h[k]; k++)
178 ;
179 if (n[k]) {
180 h += k - ms;
181 mem = 0;
182 continue;
183 }
184 /* Compare left half */
185 for (k = ms + 1; k > mem && n[k - 1] == h[k - 1]; k--)
186 ;
187 if (k <= mem)
188 return (char *)h;
189 h += p;
190 mem = mem0;
191 }
192 }
193
194 char *
strstr(const char * h,const char * n)195 strstr(const char *h, const char *n)
196 {
197 /* Return immediately on empty needle */
198 if (!n[0])
199 return (char *)h;
200
201 /* Use faster algorithms for short needles */
202 h = strchr(h, *n);
203 if (!h || !n[1])
204 return (char *)h;
205 if (!h[1])
206 return 0;
207 if (!n[2])
208 return twobyte_strstr((void *)h, (void *)n);
209 if (!h[2])
210 return 0;
211 if (!n[3])
212 return threebyte_strstr((void *)h, (void *)n);
213 if (!h[3])
214 return 0;
215 if (!n[4])
216 return fourbyte_strstr((void *)h, (void *)n);
217
218 return twoway_strstr((void *)h, (void *)n);
219 }
220