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
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
|
/*
* Copyright (C) 2004, 2007-2010, 2011-2012 Synopsys, Inc. (www.synopsys.com)
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License version 2 as
* published by the Free Software Foundation.
*
* Joern Rennecke <joern.rennecke@embecosm.com>: Jan 2012
* -Insn Scheduling improvements to csum core routines.
* = csum_fold( ) largely derived from ARM version.
* = ip_fast_cum( ) to have module scheduling
* -gcc 4.4.x broke networking. Alias analysis needed to be primed.
* worked around by adding memory clobber to ip_fast_csum( )
*
* vineetg: May 2010
* -Rewrote ip_fast_cscum( ) and csum_fold( ) with fast inline asm
*/
#ifndef _ASM_ARC_CHECKSUM_H
#define _ASM_ARC_CHECKSUM_H
/*
* Fold a partial checksum
*
* The 2 swords comprising the 32bit sum are added, any carry to 16th bit
* added back and final sword result inverted.
*/
static inline __sum16 csum_fold(__wsum s)
{
unsigned r = s << 16 | s >> 16; /* ror */
s = ~s;
s -= r;
return s >> 16;
}
/*
* This is a version of ip_compute_csum() optimized for IP headers,
* which always checksum on 4 octet boundaries.
*/
static inline __sum16
ip_fast_csum(const void *iph, unsigned int ihl)
{
const void *ptr = iph;
unsigned int tmp, tmp2, sum;
__asm__(
" ld.ab %0, [%3, 4] \n"
" ld.ab %2, [%3, 4] \n"
" sub %1, %4, 2 \n"
" lsr.f lp_count, %1, 1 \n"
" bcc 0f \n"
" add.f %0, %0, %2 \n"
" ld.ab %2, [%3, 4] \n"
"0: lp 1f \n"
" ld.ab %1, [%3, 4] \n"
" adc.f %0, %0, %2 \n"
" ld.ab %2, [%3, 4] \n"
" adc.f %0, %0, %1 \n"
"1: adc.f %0, %0, %2 \n"
" add.cs %0,%0,1 \n"
: "=&r"(sum), "=r"(tmp), "=&r"(tmp2), "+&r" (ptr)
: "r"(ihl)
: "cc", "lp_count", "memory");
return csum_fold(sum);
}
/*
* TCP pseudo Header is 12 bytes:
* SA [4], DA [4], zeroes [1], Proto[1], TCP Seg(hdr+data) Len [2]
*/
static inline __wsum
csum_tcpudp_nofold(__be32 saddr, __be32 daddr, unsigned short len,
unsigned short proto, __wsum sum)
{
__asm__ __volatile__(
" add.f %0, %0, %1 \n"
" adc.f %0, %0, %2 \n"
" adc.f %0, %0, %3 \n"
" adc.f %0, %0, %4 \n"
" adc %0, %0, 0 \n"
: "+&r"(sum)
: "r"(saddr), "r"(daddr),
#ifdef CONFIG_CPU_BIG_ENDIAN
"r"(len),
#else
"r"(len << 8),
#endif
"r"(htons(proto))
: "cc");
return sum;
}
#define csum_fold csum_fold
#define ip_fast_csum ip_fast_csum
#define csum_tcpudp_nofold csum_tcpudp_nofold
#include <asm-generic/checksum.h>
#endif /* _ASM_ARC_CHECKSUM_H */
|
