2005-04-16 15:20:36 -07:00
/*
* S390 fast network checksum routines
*
* S390 version
2012-07-20 11:15:04 +02:00
* Copyright IBM Corp . 1999
2005-04-16 15:20:36 -07:00
* Author ( s ) : Ulrich Hild ( first version )
* Martin Schwidefsky ( heavily optimized CKSM version )
* D . J . Barrow ( third attempt )
*/
2012-07-20 11:15:04 +02:00
# ifndef _S390_CHECKSUM_H
# define _S390_CHECKSUM_H
2005-04-16 15:20:36 -07:00
# include <asm/uaccess.h>
/*
* computes the checksum of a memory block at buff , length len ,
* and adds in " sum " ( 32 - bit )
*
* returns a 32 - bit number suitable for feeding into itself
* or csum_tcpudp_magic
*
* this function must be called with even lengths , except
* for the last fragment , which may be odd
*
* it ' s best to have buff aligned on a 32 - bit boundary
*/
2006-11-14 21:22:18 -08:00
static inline __wsum
csum_partial ( const void * buff , int len , __wsum sum )
2005-04-16 15:20:36 -07:00
{
2006-09-28 16:56:43 +02:00
register unsigned long reg2 asm ( " 2 " ) = ( unsigned long ) buff ;
register unsigned long reg3 asm ( " 3 " ) = ( unsigned long ) len ;
2005-04-16 15:20:36 -07:00
2006-09-28 16:56:43 +02:00
asm volatile (
" 0: cksm %0,%1 \n " /* do checksum on longs */
" jo 0b \n "
: " +d " ( sum ) , " +d " ( reg2 ) , " +d " ( reg3 ) : : " cc " , " memory " ) ;
2005-04-16 15:20:36 -07:00
return sum ;
}
/*
* the same as csum_partial_copy , but copies from user space .
*
* here even more important to align src and dst on a 32 - bit ( or even
* better 64 - bit ) boundary
*
2014-02-24 16:33:08 +01:00
* Copy from userspace and compute checksum .
2005-04-16 15:20:36 -07:00
*/
2006-11-14 21:22:18 -08:00
static inline __wsum
csum_partial_copy_from_user ( const void __user * src , void * dst ,
int len , __wsum sum ,
2005-04-16 15:20:36 -07:00
int * err_ptr )
{
2014-02-24 16:33:08 +01:00
if ( unlikely ( copy_from_user ( dst , src , len ) ) )
2005-04-16 15:20:36 -07:00
* err_ptr = - EFAULT ;
return csum_partial ( dst , len , sum ) ;
}
2006-11-14 21:22:18 -08:00
static inline __wsum
csum_partial_copy_nocheck ( const void * src , void * dst , int len , __wsum sum )
2005-04-16 15:20:36 -07:00
{
memcpy ( dst , src , len ) ;
2006-09-28 16:56:43 +02:00
return csum_partial ( dst , len , sum ) ;
2005-04-16 15:20:36 -07:00
}
/*
* Fold a partial checksum without adding pseudo headers
*/
2006-11-14 21:22:18 -08:00
static inline __sum16 csum_fold ( __wsum sum )
2005-04-16 15:20:36 -07:00
{
[S390] convert/optimize csum_fold() to C
In the meantime gcc generates better code than the old inline
assemblies do. Original inline assembly results in:
lr %r1,%r2
sr %r3,%r3
lr %r2,%r1
srdl %r2,16
alr %r2,%r3
alr %r1,%r2
srl %r1,16
xilf %r1,65535
llghr %r2,%r1
br %r14
Out of the C code gcc generates this:
rll %r1,%r2,16
ar %r1,%r2
srl %r1,16
xilf %r1,65535
llghr %r2,%r1
br %r14
In addition we don't have any static register allocations anymore and
gcc is free to shuffle instructions around for better pipeline usage.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2009-09-11 10:28:32 +02:00
u32 csum = ( __force u32 ) sum ;
2005-04-16 15:20:36 -07:00
[S390] convert/optimize csum_fold() to C
In the meantime gcc generates better code than the old inline
assemblies do. Original inline assembly results in:
lr %r1,%r2
sr %r3,%r3
lr %r2,%r1
srdl %r2,16
alr %r2,%r3
alr %r1,%r2
srl %r1,16
xilf %r1,65535
llghr %r2,%r1
br %r14
Out of the C code gcc generates this:
rll %r1,%r2,16
ar %r1,%r2
srl %r1,16
xilf %r1,65535
llghr %r2,%r1
br %r14
In addition we don't have any static register allocations anymore and
gcc is free to shuffle instructions around for better pipeline usage.
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2009-09-11 10:28:32 +02:00
csum + = ( csum > > 16 ) + ( csum < < 16 ) ;
csum > > = 16 ;
return ( __force __sum16 ) ~ csum ;
2005-04-16 15:20:36 -07:00
}
/*
* This is a version of ip_compute_csum ( ) optimized for IP headers ,
* which always checksum on 4 octet boundaries .
*
*/
2006-11-14 21:22:18 -08:00
static inline __sum16 ip_fast_csum ( const void * iph , unsigned int ihl )
2005-04-16 15:20:36 -07:00
{
2006-09-28 16:56:43 +02:00
return csum_fold ( csum_partial ( iph , ihl * 4 , 0 ) ) ;
2005-04-16 15:20:36 -07:00
}
/*
* computes the checksum of the TCP / UDP pseudo - header
* returns a 32 - bit checksum
*/
2006-11-14 21:22:18 -08:00
static inline __wsum
csum_tcpudp_nofold ( __be32 saddr , __be32 daddr ,
2005-04-16 15:20:36 -07:00
unsigned short len , unsigned short proto ,
2006-11-14 21:22:18 -08:00
__wsum sum )
2005-04-16 15:20:36 -07:00
{
2007-03-26 20:42:39 +02:00
__u32 csum = ( __force __u32 ) sum ;
csum + = ( __force __u32 ) saddr ;
if ( csum < ( __force __u32 ) saddr )
csum + + ;
csum + = ( __force __u32 ) daddr ;
if ( csum < ( __force __u32 ) daddr )
csum + + ;
csum + = len + proto ;
if ( csum < len + proto )
csum + + ;
return ( __force __wsum ) csum ;
2005-04-16 15:20:36 -07:00
}
/*
* computes the checksum of the TCP / UDP pseudo - header
* returns a 16 - bit checksum , already complemented
*/
2006-11-14 21:22:18 -08:00
static inline __sum16
csum_tcpudp_magic ( __be32 saddr , __be32 daddr ,
2005-04-16 15:20:36 -07:00
unsigned short len , unsigned short proto ,
2006-11-14 21:22:18 -08:00
__wsum sum )
2005-04-16 15:20:36 -07:00
{
return csum_fold ( csum_tcpudp_nofold ( saddr , daddr , len , proto , sum ) ) ;
}
/*
* this routine is used for miscellaneous IP - like checksums , mainly
* in icmp . c
*/
2006-11-14 21:22:18 -08:00
static inline __sum16 ip_compute_csum ( const void * buff , int len )
2005-04-16 15:20:36 -07:00
{
return csum_fold ( csum_partial ( buff , len , 0 ) ) ;
}
# endif /* _S390_CHECKSUM_H */