1/*-
2 * Copyright (c) 2009 Xin LI <delphij@FreeBSD.org>
3 * All rights reserved.
4 *
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
7 * are met:
8 * 1. Redistributions of source code must retain the above copyright
9 *    notice, this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright
11 *    notice, this list of conditions and the following disclaimer in the
12 *    documentation and/or other materials provided with the distribution.
13 *
14 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
15 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
16 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
17 * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
18 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
19 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
20 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
21 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
22 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
23 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
24 * SUCH DAMAGE.
25 */
26
27#include "string.h"
28#include <limits.h>
29
30/*
31 * Portable strlen() for 32-bit and 64-bit systems.
32 *
33 * Rationale: it is generally much more efficient to do word length
34 * operations and avoid branches on modern computer systems, as
35 * compared to byte-length operations with a lot of branches.
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.  However, we can further reduce ~1/3 of
43 * time if we consider that strlen() usually operate on 7-bit ASCII
44 * by employing the following expression, which allows false positive
45 * when high bit of 1 and use the tail case to catch these case:
46 *
47 *	((x - 0x01....01) & 0x80....80)
48 *
49 * This is more than 5.2 times as fast as the raw implementation on
50 * Intel T7300 under long mode for strings longer than word length.
51 */
52
53/* Magic numbers for the algorithm */
54#if LONG_BIT == 32
55static const unsigned long mask01 = 0x01010101;
56static const unsigned long mask80 = 0x80808080;
57#elif LONG_BIT == 64
58static const unsigned long mask01 = 0x0101010101010101;
59static const unsigned long mask80 = 0x8080808080808080;
60#else
61#error Unsupported word size
62#endif
63
64#define	LONGPTR_MASK (sizeof(long) - 1)
65
66/*
67 * Helper macro to return string length if we caught the zero
68 * byte.
69 */
70#define testbyte(x)				\
71	do {					\
72		if (p[x] == '\0')		\
73		    return (p - str + x);	\
74	} while (0)
75
76__attribute__((visibility("hidden")))
77size_t
78strlen(const char *str)
79{
80	const char *p;
81	const unsigned long *lp;
82
83	/* Skip the first few bytes until we have an aligned p */
84	for (p = str; (uintptr_t)p & LONGPTR_MASK; p++)
85	    if (*p == '\0')
86		return (p - str);
87
88	/* Scan the rest of the string using word sized operation */
89	for (lp = (const unsigned long *)p; ; lp++)
90	    if ((*lp - mask01) & mask80) {
91		p = (const char *)(lp);
92		testbyte(0);
93		testbyte(1);
94		testbyte(2);
95		testbyte(3);
96#if (LONG_BIT >= 64)
97		testbyte(4);
98		testbyte(5);
99		testbyte(6);
100		testbyte(7);
101#endif
102	    }
103
104	/* NOTREACHED */
105	return (0);
106}
107
108