xref: /freebsd/sys/libkern/strlen.c (revision 5bb3134a8c21cb87b30e135ef168483f0333dabb)
1 /*-
2  * SPDX-License-Identifier: BSD-2-Clause-FreeBSD
3  *
4  * Copyright (c) 2009, 2010 Xin LI <delphij@FreeBSD.org>
5  *
6  * Redistribution and use in source and binary forms, with or without
7  * modification, are permitted provided that the following conditions
8  * are met:
9  * 1. Redistributions of source code must retain the above copyright
10  *    notice, this list of conditions and the following disclaimer.
11  * 2. Redistributions in binary form must reproduce the above copyright
12  *    notice, this list of conditions and the following disclaimer in the
13  *    documentation and/or other materials provided with the distribution.
14  *
15  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
16  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
17  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
18  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
19  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
20  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
21  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
22  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
23  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
24  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
25  * SUCH DAMAGE.
26  */
27 
28 #include <sys/cdefs.h>
29 __FBSDID("$FreeBSD$");
30 
31 #include <sys/libkern.h>
32 #include <sys/limits.h>
33 
34 /*
35  * Portable strlen() for 32-bit and 64-bit systems.
36  *
37  * The expression:
38  *
39  *	((x - 0x01....01) & ~x & 0x80....80)
40  *
41  * would evaluate to a non-zero value iff any of the bytes in the
42  * original word is zero.
43  *
44  * The algorithm above is found on "Hacker's Delight" by
45  * Henry S. Warren, Jr.
46  *
47  * Note: this leaves performance on the table and each architecture
48  * would be best served with a tailor made routine instead, even if
49  * using the same trick.
50  */
51 
52 /* Magic numbers for the algorithm */
53 #if LONG_BIT == 32
54 static const unsigned long mask01 = 0x01010101;
55 static const unsigned long mask80 = 0x80808080;
56 #elif LONG_BIT == 64
57 static const unsigned long mask01 = 0x0101010101010101;
58 static const unsigned long mask80 = 0x8080808080808080;
59 #else
60 #error Unsupported word size
61 #endif
62 
63 #define	LONGPTR_MASK (sizeof(long) - 1)
64 
65 /*
66  * Helper macro to return string length if we caught the zero
67  * byte.
68  */
69 #define testbyte(x)				\
70 	do {					\
71 		if (p[x] == '\0')		\
72 		    return (p - str + x);	\
73 	} while (0)
74 
75 size_t
76 (strlen)(const char *str)
77 {
78 	const char *p;
79 	const unsigned long *lp;
80 	long va, vb;
81 
82 	/*
83 	 * Before trying the hard (unaligned byte-by-byte access) way
84 	 * to figure out whether there is a nul character, try to see
85 	 * if there is a nul character is within this accessible word
86 	 * first.
87 	 *
88 	 * p and (p & ~LONGPTR_MASK) must be equally accessible since
89 	 * they always fall in the same memory page, as long as page
90 	 * boundaries is integral multiple of word size.
91 	 */
92 	lp = (const unsigned long *)((uintptr_t)str & ~LONGPTR_MASK);
93 	va = (*lp - mask01);
94 	vb = ((~*lp) & mask80);
95 	lp++;
96 	if (va & vb)
97 		/* Check if we have \0 in the first part */
98 		for (p = str; p < (const char *)lp; p++)
99 			if (*p == '\0')
100 				return (p - str);
101 
102 	/* Scan the rest of the string using word sized operation */
103 	for (; ; lp++) {
104 		va = (*lp - mask01);
105 		vb = ((~*lp) & mask80);
106 		if (va & vb) {
107 			p = (const char *)(lp);
108 			testbyte(0);
109 			testbyte(1);
110 			testbyte(2);
111 			testbyte(3);
112 #if (LONG_BIT >= 64)
113 			testbyte(4);
114 			testbyte(5);
115 			testbyte(6);
116 			testbyte(7);
117 #endif
118 		}
119 	}
120 
121 	/* NOTREACHED */
122 	return (0);
123 }
124