1 /*- 2 * Copyright (c) 2005-2014 Rich Felker, et al. 3 * 4 * Permission is hereby granted, free of charge, to any person obtaining 5 * a copy of this software and associated documentation files (the 6 * "Software"), to deal in the Software without restriction, including 7 * without limitation the rights to use, copy, modify, merge, publish, 8 * distribute, sublicense, and/or sell copies of the Software, and to 9 * permit persons to whom the Software is furnished to do so, subject to 10 * the following conditions: 11 * 12 * The above copyright notice and this permission notice shall be 13 * included in all copies or substantial portions of the Software. 14 * 15 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, 16 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF 17 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. 18 * IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY 19 * CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, 20 * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE 21 * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. 22 */ 23 #include <sys/cdefs.h> 24 #if defined(LIBC_SCCS) && !defined(lint) 25 #if 0 26 __FBSDID("$FreeBSD: head/lib/libc/string/memmem.c 315468 2017-03-18 00:53:24Z emaste $"); 27 #else 28 __RCSID("$NetBSD: memmem.c,v 1.1 2018/07/08 17:53:12 christos Exp $"); 29 #endif 30 #endif /* LIBC_SCCS and not lint */ 31 32 #if !defined(_KERNEL) && !defined(_STANDALONE) 33 #include <string.h> 34 #include <stdint.h> 35 #else 36 #include <lib/libkern/libkern.h> 37 #endif 38 39 static char *twobyte_memmem(const unsigned char *h, size_t k, const unsigned char *n) 40 { 41 uint16_t nw = n[0]<<8 | n[1], hw = h[0]<<8 | h[1]; 42 for (h++, k--; k; k--, hw = hw<<8 | *++h) 43 if (hw == nw) return __UNCONST(h-1); 44 return 0; 45 } 46 47 static char *threebyte_memmem(const unsigned char *h, size_t k, const unsigned char *n) 48 { 49 uint32_t nw = n[0]<<24 | n[1]<<16 | n[2]<<8; 50 uint32_t hw = h[0]<<24 | h[1]<<16 | h[2]<<8; 51 for (h+=2, k-=2; k; k--, hw = (hw|*++h)<<8) 52 if (hw == nw) return __UNCONST(h-2); 53 return 0; 54 } 55 56 static char *fourbyte_memmem(const unsigned char *h, size_t k, const unsigned char *n) 57 { 58 uint32_t nw = n[0]<<24 | n[1]<<16 | n[2]<<8 | n[3]; 59 uint32_t hw = h[0]<<24 | h[1]<<16 | h[2]<<8 | h[3]; 60 for (h+=3, k-=3; k; k--, hw = hw<<8 | *++h) 61 if (hw == nw) return __UNCONST(h-3); 62 return 0; 63 } 64 65 #define MAX(a,b) ((a)>(b)?(a):(b)) 66 #define MIN(a,b) ((a)<(b)?(a):(b)) 67 68 #define BITOP(a,b,op) \ 69 ((a)[(size_t)(b)/(8*sizeof *(a))] op (size_t)1<<((size_t)(b)%(8*sizeof *(a)))) 70 71 /* 72 * Two Way string search algorithm, with a bad shift table applied to the last 73 * byte of the window. A bit array marks which entries in the shift table are 74 * initialized to avoid fully initializing a 1kb/2kb table. 75 * 76 * Reference: CROCHEMORE M., PERRIN D., 1991, Two-way string-matching, 77 * Journal of the ACM 38(3):651-675 78 */ 79 static char *twoway_memmem(const unsigned char *h, const unsigned char *z, const unsigned char *n, size_t l) 80 { 81 size_t i, ip, jp, k, p, ms, p0, mem, mem0; 82 size_t byteset[32 / sizeof(size_t)] = { 0 }; 83 size_t shift[256]; 84 85 /* Computing length of needle and fill shift table */ 86 for (i=0; i<l; i++) 87 BITOP(byteset, n[i], |=), shift[n[i]] = i+1; 88 89 /* Compute maximal suffix */ 90 ip = (size_t)-1; jp = 0; 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 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 = (size_t)-1; jp = 0; k = p = 1; 111 while (jp+k<l) { 112 if (n[ip+k] == n[jp+k]) { 113 if (k == p) { 114 jp += p; 115 k = 1; 116 } else k++; 117 } else if (n[ip+k] < n[jp+k]) { 118 jp += k; 119 k = 1; 120 p = jp - ip; 121 } else { 122 ip = jp++; 123 k = p = 1; 124 } 125 } 126 if (ip+1 > ms+1) ms = ip; 127 else p = p0; 128 129 /* Periodic needle? */ 130 if (memcmp(n, n+p, ms+1)) { 131 mem0 = 0; 132 p = MAX(ms, l-ms-1) + 1; 133 } else mem0 = l-p; 134 mem = 0; 135 136 /* Search loop */ 137 for (;;) { 138 /* If remainder of haystack is shorter than needle, done */ 139 if ((size_t)(z-h) < l) return 0; 140 141 /* Check last byte first; advance by shift on mismatch */ 142 if (BITOP(byteset, h[l-1], &)) { 143 k = l-shift[h[l-1]]; 144 if (k) { 145 if (mem0 && mem && k < p) k = l-p; 146 h += k; 147 mem = 0; 148 continue; 149 } 150 } else { 151 h += l; 152 mem = 0; 153 continue; 154 } 155 156 /* Compare right half */ 157 for (k=MAX(ms+1,mem); k<l && n[k] == h[k]; k++); 158 if (k < l) { 159 h += k-ms; 160 mem = 0; 161 continue; 162 } 163 /* Compare left half */ 164 for (k=ms+1; k>mem && n[k-1] == h[k-1]; k--); 165 if (k <= mem) return __UNCONST(h); 166 h += p; 167 mem = mem0; 168 } 169 } 170 171 void *memmem(const void *h0, size_t k, const void *n0, size_t l) 172 { 173 const unsigned char *h = h0, *n = n0; 174 175 /* Return immediately on empty needle */ 176 if (!l) return __UNCONST(h); 177 178 /* Return immediately when needle is longer than haystack */ 179 if (k<l) return 0; 180 181 /* Use faster algorithms for short needles */ 182 h = memchr(h0, *n, k); 183 if (!h || l==1) return __UNCONST(h); 184 k -= h - (const unsigned char *)h0; 185 if (k<l) return 0; 186 if (l==2) return twobyte_memmem(h, k, n); 187 if (l==3) return threebyte_memmem(h, k, n); 188 if (l==4) return fourbyte_memmem(h, k, n); 189 190 return twoway_memmem(h, h+k, n, l); 191 } 192