1 0d6fb46a 2022-01-09 op /*	$OpenBSD: memmem.c,v 1.5 2020/04/16 12:39:28 claudio Exp $ */
  2 0d6fb46a 2022-01-09 op 
  3 0d6fb46a 2022-01-09 op /*
  4 0d6fb46a 2022-01-09 op  * Copyright (c) 2005-2020 Rich Felker, et al.
  5 0d6fb46a 2022-01-09 op  *
  6 0d6fb46a 2022-01-09 op  * Permission is hereby granted, free of charge, to any person obtaining
  7 0d6fb46a 2022-01-09 op  * a copy of this software and associated documentation files (the
  8 0d6fb46a 2022-01-09 op  * "Software"), to deal in the Software without restriction, including
  9 0d6fb46a 2022-01-09 op  * without limitation the rights to use, copy, modify, merge, publish,
 10 0d6fb46a 2022-01-09 op  * distribute, sublicense, and/or sell copies of the Software, and to
 11 0d6fb46a 2022-01-09 op  * permit persons to whom the Software is furnished to do so, subject to
 12 0d6fb46a 2022-01-09 op  * the following conditions:
 13 0d6fb46a 2022-01-09 op  *
 14 0d6fb46a 2022-01-09 op  * The above copyright notice and this permission notice shall be
 15 0d6fb46a 2022-01-09 op  * included in all copies or substantial portions of the Software.
 16 0d6fb46a 2022-01-09 op  *
 17 0d6fb46a 2022-01-09 op  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
 18 0d6fb46a 2022-01-09 op  * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
 19 0d6fb46a 2022-01-09 op  * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
 20 0d6fb46a 2022-01-09 op  * IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
 21 0d6fb46a 2022-01-09 op  * CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
 22 0d6fb46a 2022-01-09 op  * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
 23 0d6fb46a 2022-01-09 op  * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 24 0d6fb46a 2022-01-09 op  */
 25 0d6fb46a 2022-01-09 op 
 26 0d6fb46a 2022-01-09 op #include <string.h>
 27 0d6fb46a 2022-01-09 op #include <stdint.h>
 28 0d6fb46a 2022-01-09 op 
 29 0d6fb46a 2022-01-09 op static char *
 30 0d6fb46a 2022-01-09 op twobyte_memmem(const unsigned char *h, size_t k, const unsigned char *n)
 31 0d6fb46a 2022-01-09 op {
 32 0d6fb46a 2022-01-09 op 	uint16_t nw = n[0]<<8 | n[1], hw = h[0]<<8 | h[1];
 33 0d6fb46a 2022-01-09 op 	for (h+=2, k-=2; k; k--, hw = hw<<8 | *h++)
 34 0d6fb46a 2022-01-09 op 		if (hw == nw) return (char *)h-2;
 35 0d6fb46a 2022-01-09 op 	return hw == nw ? (char *)h-2 : 0;
 36 0d6fb46a 2022-01-09 op }
 37 0d6fb46a 2022-01-09 op 
 38 0d6fb46a 2022-01-09 op static char *
 39 0d6fb46a 2022-01-09 op threebyte_memmem(const unsigned char *h, size_t k, const unsigned char *n)
 40 0d6fb46a 2022-01-09 op {
 41 0d6fb46a 2022-01-09 op 	uint32_t nw = n[0]<<24 | n[1]<<16 | n[2]<<8;
 42 0d6fb46a 2022-01-09 op 	uint32_t hw = h[0]<<24 | h[1]<<16 | h[2]<<8;
 43 0d6fb46a 2022-01-09 op 	for (h+=3, k-=3; k; k--, hw = (hw|*h++)<<8)
 44 0d6fb46a 2022-01-09 op 		if (hw == nw) return (char *)h-3;
 45 0d6fb46a 2022-01-09 op 	return hw == nw ? (char *)h-3 : 0;
 46 0d6fb46a 2022-01-09 op }
 47 0d6fb46a 2022-01-09 op 
 48 0d6fb46a 2022-01-09 op static char *
 49 0d6fb46a 2022-01-09 op fourbyte_memmem(const unsigned char *h, size_t k, const unsigned char *n)
 50 0d6fb46a 2022-01-09 op {
 51 0d6fb46a 2022-01-09 op 	uint32_t nw = n[0]<<24 | n[1]<<16 | n[2]<<8 | n[3];
 52 0d6fb46a 2022-01-09 op 	uint32_t hw = h[0]<<24 | h[1]<<16 | h[2]<<8 | h[3];
 53 0d6fb46a 2022-01-09 op 	for (h+=4, k-=4; k; k--, hw = hw<<8 | *h++)
 54 0d6fb46a 2022-01-09 op 		if (hw == nw) return (char *)h-4;
 55 0d6fb46a 2022-01-09 op 	return hw == nw ? (char *)h-4 : 0;
 56 0d6fb46a 2022-01-09 op }
 57 0d6fb46a 2022-01-09 op 
 58 0d6fb46a 2022-01-09 op #define MAX(a,b) ((a)>(b)?(a):(b))
 59 0d6fb46a 2022-01-09 op #define MIN(a,b) ((a)<(b)?(a):(b))
 60 0d6fb46a 2022-01-09 op 
 61 0d6fb46a 2022-01-09 op #define BITOP(a,b,op) \
 62 0d6fb46a 2022-01-09 op  ((a)[(size_t)(b)/(8*sizeof *(a))] op (size_t)1<<((size_t)(b)%(8*sizeof *(a))))
 63 0d6fb46a 2022-01-09 op 
 64 0d6fb46a 2022-01-09 op /*
 65 0d6fb46a 2022-01-09 op  * Maxime Crochemore and Dominique Perrin, Two-way string-matching,
 66 0d6fb46a 2022-01-09 op  * Journal of the ACM, 38(3):651-675, July 1991.
 67 0d6fb46a 2022-01-09 op  */
 68 0d6fb46a 2022-01-09 op static char *
 69 0d6fb46a 2022-01-09 op twoway_memmem(const unsigned char *h, const unsigned char *z,
 70 0d6fb46a 2022-01-09 op     const unsigned char *n, size_t l)
 71 0d6fb46a 2022-01-09 op {
 72 0d6fb46a 2022-01-09 op 	size_t i, ip, jp, k, p, ms, p0, mem, mem0;
 73 0d6fb46a 2022-01-09 op 	size_t byteset[32 / sizeof(size_t)] = { 0 };
 74 0d6fb46a 2022-01-09 op 	size_t shift[256];
 75 0d6fb46a 2022-01-09 op 
 76 0d6fb46a 2022-01-09 op 	/* Computing length of needle and fill shift table */
 77 0d6fb46a 2022-01-09 op 	for (i=0; i<l; i++)
 78 0d6fb46a 2022-01-09 op 		BITOP(byteset, n[i], |=), shift[n[i]] = i+1;
 79 0d6fb46a 2022-01-09 op 
 80 0d6fb46a 2022-01-09 op 	/* Compute maximal suffix */
 81 0d6fb46a 2022-01-09 op 	ip = -1; jp = 0; k = p = 1;
 82 0d6fb46a 2022-01-09 op 	while (jp+k<l) {
 83 0d6fb46a 2022-01-09 op 		if (n[ip+k] == n[jp+k]) {
 84 0d6fb46a 2022-01-09 op 			if (k == p) {
 85 0d6fb46a 2022-01-09 op 				jp += p;
 86 0d6fb46a 2022-01-09 op 				k = 1;
 87 0d6fb46a 2022-01-09 op 			} else k++;
 88 0d6fb46a 2022-01-09 op 		} else if (n[ip+k] > n[jp+k]) {
 89 0d6fb46a 2022-01-09 op 			jp += k;
 90 0d6fb46a 2022-01-09 op 			k = 1;
 91 0d6fb46a 2022-01-09 op 			p = jp - ip;
 92 0d6fb46a 2022-01-09 op 		} else {
 93 0d6fb46a 2022-01-09 op 			ip = jp++;
 94 0d6fb46a 2022-01-09 op 			k = p = 1;
 95 0d6fb46a 2022-01-09 op 		}
 96 0d6fb46a 2022-01-09 op 	}
 97 0d6fb46a 2022-01-09 op 	ms = ip;
 98 0d6fb46a 2022-01-09 op 	p0 = p;
 99 0d6fb46a 2022-01-09 op 
100 0d6fb46a 2022-01-09 op 	/* And with the opposite comparison */
101 0d6fb46a 2022-01-09 op 	ip = -1; jp = 0; k = p = 1;
102 0d6fb46a 2022-01-09 op 	while (jp+k<l) {
103 0d6fb46a 2022-01-09 op 		if (n[ip+k] == n[jp+k]) {
104 0d6fb46a 2022-01-09 op 			if (k == p) {
105 0d6fb46a 2022-01-09 op 				jp += p;
106 0d6fb46a 2022-01-09 op 				k = 1;
107 0d6fb46a 2022-01-09 op 			} else k++;
108 0d6fb46a 2022-01-09 op 		} else if (n[ip+k] < n[jp+k]) {
109 0d6fb46a 2022-01-09 op 			jp += k;
110 0d6fb46a 2022-01-09 op 			k = 1;
111 0d6fb46a 2022-01-09 op 			p = jp - ip;
112 0d6fb46a 2022-01-09 op 		} else {
113 0d6fb46a 2022-01-09 op 			ip = jp++;
114 0d6fb46a 2022-01-09 op 			k = p = 1;
115 0d6fb46a 2022-01-09 op 		}
116 0d6fb46a 2022-01-09 op 	}
117 0d6fb46a 2022-01-09 op 	if (ip+1 > ms+1) ms = ip;
118 0d6fb46a 2022-01-09 op 	else p = p0;
119 0d6fb46a 2022-01-09 op 
120 0d6fb46a 2022-01-09 op 	/* Periodic needle? */
121 0d6fb46a 2022-01-09 op 	if (memcmp(n, n+p, ms+1)) {
122 0d6fb46a 2022-01-09 op 		mem0 = 0;
123 0d6fb46a 2022-01-09 op 		p = MAX(ms, l-ms-1) + 1;
124 0d6fb46a 2022-01-09 op 	} else mem0 = l-p;
125 0d6fb46a 2022-01-09 op 	mem = 0;
126 0d6fb46a 2022-01-09 op 
127 0d6fb46a 2022-01-09 op 	/* Search loop */
128 0d6fb46a 2022-01-09 op 	for (;;) {
129 0d6fb46a 2022-01-09 op 		/* If remainder of haystack is shorter than needle, done */
130 0d6fb46a 2022-01-09 op 		if (z-h < l) return 0;
131 0d6fb46a 2022-01-09 op 
132 0d6fb46a 2022-01-09 op 		/* Check last byte first; advance by shift on mismatch */
133 0d6fb46a 2022-01-09 op 		if (BITOP(byteset, h[l-1], &)) {
134 0d6fb46a 2022-01-09 op 			k = l-shift[h[l-1]];
135 0d6fb46a 2022-01-09 op 			if (k) {
136 0d6fb46a 2022-01-09 op 				if (k < mem) k = mem;
137 0d6fb46a 2022-01-09 op 				h += k;
138 0d6fb46a 2022-01-09 op 				mem = 0;
139 0d6fb46a 2022-01-09 op 				continue;
140 0d6fb46a 2022-01-09 op 			}
141 0d6fb46a 2022-01-09 op 		} else {
142 0d6fb46a 2022-01-09 op 			h += l;
143 0d6fb46a 2022-01-09 op 			mem = 0;
144 0d6fb46a 2022-01-09 op 			continue;
145 0d6fb46a 2022-01-09 op 		}
146 0d6fb46a 2022-01-09 op 
147 0d6fb46a 2022-01-09 op 		/* Compare right half */
148 0d6fb46a 2022-01-09 op 		for (k=MAX(ms+1,mem); k<l && n[k] == h[k]; k++);
149 0d6fb46a 2022-01-09 op 		if (k < l) {
150 0d6fb46a 2022-01-09 op 			h += k-ms;
151 0d6fb46a 2022-01-09 op 			mem = 0;
152 0d6fb46a 2022-01-09 op 			continue;
153 0d6fb46a 2022-01-09 op 		}
154 0d6fb46a 2022-01-09 op 		/* Compare left half */
155 0d6fb46a 2022-01-09 op 		for (k=ms+1; k>mem && n[k-1] == h[k-1]; k--);
156 0d6fb46a 2022-01-09 op 		if (k <= mem) return (char *)h;
157 0d6fb46a 2022-01-09 op 		h += p;
158 0d6fb46a 2022-01-09 op 		mem = mem0;
159 0d6fb46a 2022-01-09 op 	}
160 0d6fb46a 2022-01-09 op }
161 0d6fb46a 2022-01-09 op 
162 0d6fb46a 2022-01-09 op void *
163 0d6fb46a 2022-01-09 op memmem(const void *h0, size_t k, const void *n0, size_t l)
164 0d6fb46a 2022-01-09 op {
165 0d6fb46a 2022-01-09 op 	const unsigned char *h = h0, *n = n0;
166 0d6fb46a 2022-01-09 op 
167 0d6fb46a 2022-01-09 op 	/* Return immediately on empty needle */
168 0d6fb46a 2022-01-09 op 	if (!l) return (void *)h;
169 0d6fb46a 2022-01-09 op 
170 0d6fb46a 2022-01-09 op 	/* Return immediately when needle is longer than haystack */
171 0d6fb46a 2022-01-09 op 	if (k<l) return 0;
172 0d6fb46a 2022-01-09 op 
173 0d6fb46a 2022-01-09 op 	/* Use faster algorithms for short needles */
174 0d6fb46a 2022-01-09 op 	h = memchr(h0, *n, k);
175 0d6fb46a 2022-01-09 op 	if (!h || l==1) return (void *)h;
176 0d6fb46a 2022-01-09 op 	k -= h - (const unsigned char *)h0;
177 0d6fb46a 2022-01-09 op 	if (k<l) return 0;
178 0d6fb46a 2022-01-09 op 	if (l==2) return twobyte_memmem(h, k, n);
179 0d6fb46a 2022-01-09 op 	if (l==3) return threebyte_memmem(h, k, n);
180 0d6fb46a 2022-01-09 op 	if (l==4) return fourbyte_memmem(h, k, n);
181 0d6fb46a 2022-01-09 op 
182 0d6fb46a 2022-01-09 op 	return twoway_memmem(h, h+k, n, l);
183 0d6fb46a 2022-01-09 op }