1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
|
/*
* This string-include defines all string functions as inline
* functions. Use gcc. It also assumes ds=es=data space, this should be
* normal. Most of the string-functions are rather heavily hand-optimized,
* see especially strtok,strstr,str[c]spn. They should work, but are not
* very easy to understand. Everything is done entirely within the register
* set, making the functions fast and clean. String instructions have been
* used through-out, making for "slightly" unclear code :-)
*
* NO Copyright (C) 1991, 1992 Linus Torvalds,
* consider these trivial functions to be PD.
*/
/*
* Copyright (C) 2000-2005 Erik Andersen <andersen@uclibc.org>
*
* Licensed under the LGPL v2.1, see the file COPYING.LIB in this tarball.
*/
/*
* Modified for uClibc by Erik Andersen <andersen@codepoet.org>
* These make no attempt to use nifty things like mmx/3dnow/etc.
* These are not inline, and will therefore not be as fast as
* modifying the headers to use inlines (and cannot therefore
* do tricky things when dealing with const memory). But they
* should (I hope!) be faster than their generic equivalents....
*
* More importantly, these should provide a good example for
* others to follow when adding arch specific optimizations.
* -Erik
*/
#include <string.h>
/* Experimentally off - libc_hidden_proto(memcpy) */
void *memcpy(void * to, const void * from, size_t n)
{
int d0, d1, d2;
__asm__ __volatile__(
" rep; movsl\n"
" movl %4,%%ecx\n"
" andl $3,%%ecx\n"
/* jz is optional. avoids "rep; movsb" with ecx == 0,
* but adds a branch, which is currently (2008) faster */
" jz 1f\n"
" rep; movsb\n"
"1:\n"
: "=&c" (d0), "=&D" (d1), "=&S" (d2)
: "0" (n / 4), "g" (n), "1" ((long)to), "2" ((long)from)
: "memory"
);
return to;
}
libc_hidden_def(memcpy)
|
