* Re: SCTP checksum patch for 2.6.26.5
2008-09-22 20:12 ` SCTP checksum patch for 2.6.26.5 David Miller
@ 2008-09-22 20:36 ` Vlad Yasevich
2008-09-23 5:19 ` Willy Tarreau
1 sibling, 0 replies; 4+ messages in thread
From: Vlad Yasevich @ 2008-09-22 20:36 UTC (permalink / raw)
To: nickolay; +Cc: David Miller, linux-kernel, linux-sctp, netdev
David Miller wrote:
> From: Nickolay Vinogradov <nickolay@protei.ru>
> Date: Mon, 22 Sep 2008 19:58:26 +0400
>
> CC:'ing correct mailing lists...
>
>> SCTP checksum calculation fix for BigEndian.
>> ntohl() doesn't do anything on BigEndian.
>>
>> diff --git a/Makefile b/Makefile
>> diff --git a/include/net/sctp/checksum.h b/include/net/sctp/checksum.h
>> index ba75c67..2f0ed64 100644
>> --- a/include/net/sctp/checksum.h
>> +++ b/include/net/sctp/checksum.h
>> @@ -74,5 +74,19 @@ static inline __u32 sctp_update_cksum(__u8 *buffer, __u16 length, __u32 crc32)
>>
>> static inline __u32 sctp_end_cksum(__u32 crc32)
>> {
>> - return ntohl(~crc32);
>> + __u32 result;
>> + __u8 byte0, byte1, byte2, byte3;
>> +
>> + result = ~crc32;
>> +
>> + byte0 = result & 0xff;
>> + byte1 = (result>>8) & 0xff;
>> + byte2 = (result>>16) & 0xff;
>> + byte3 = (result>>24) & 0xff;
>> +
>> + crc32 = ((byte0 << 24) |
>> + (byte1 << 16) |
>> + (byte2 << 8) |
>> + byte3);
>> + return crc32;
>> }
>> diff --git a/net/sctp/output.c b/net/sctp/output.c
Was this tested? I remember running tests with that code between
big and little endian arches and it worked. So, the old code worked
correclty.
I changed in 2.6.27 series to do all the calculations in network order anyway.
-vlad
>>
>>
>> -- Nickolay Vinogradov
>> Protei Research and Development Center
>> St.Petersburg, 194044, Russia
>> Tel.: +7 812 449 47 27
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
>> the body of a message to majordomo@vger.kernel.org
>> More majordomo info at http://vger.kernel.org/majordomo-info.html
>> Please read the FAQ at http://www.tux.org/lkml/
> --
> To unsubscribe from this list: send the line "unsubscribe linux-sctp" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
^ permalink raw reply [flat|nested] 4+ messages in thread* Re: SCTP checksum patch for 2.6.26.5
2008-09-22 20:12 ` SCTP checksum patch for 2.6.26.5 David Miller
2008-09-22 20:36 ` Vlad Yasevich
@ 2008-09-23 5:19 ` Willy Tarreau
2008-09-23 5:45 ` Harvey Harrison
1 sibling, 1 reply; 4+ messages in thread
From: Willy Tarreau @ 2008-09-23 5:19 UTC (permalink / raw)
To: David Miller; +Cc: nickolay, linux-kernel, linux-sctp, netdev
On Mon, Sep 22, 2008 at 01:12:45PM -0700, David Miller wrote:
> From: Nickolay Vinogradov <nickolay@protei.ru>
> Date: Mon, 22 Sep 2008 19:58:26 +0400
>
> CC:'ing correct mailing lists...
>
> > SCTP checksum calculation fix for BigEndian.
> > ntohl() doesn't do anything on BigEndian.
> >
> > diff --git a/Makefile b/Makefile
> > diff --git a/include/net/sctp/checksum.h b/include/net/sctp/checksum.h
> > index ba75c67..2f0ed64 100644
> > --- a/include/net/sctp/checksum.h
> > +++ b/include/net/sctp/checksum.h
> > @@ -74,5 +74,19 @@ static inline __u32 sctp_update_cksum(__u8 *buffer, __u16 length, __u32 crc32)
> >
> > static inline __u32 sctp_end_cksum(__u32 crc32)
> > {
> > - return ntohl(~crc32);
> > + __u32 result;
> > + __u8 byte0, byte1, byte2, byte3;
> > +
> > + result = ~crc32;
> > +
> > + byte0 = result & 0xff;
> > + byte1 = (result>>8) & 0xff;
> > + byte2 = (result>>16) & 0xff;
> > + byte3 = (result>>24) & 0xff;
> > +
> > + crc32 = ((byte0 << 24) |
> > + (byte1 << 16) |
> > + (byte2 << 8) |
> > + byte3);
> > + return crc32;
> > }
Nickolay,
you should use swab32(~crc32) instead. It will make use of arch-specific
optimisations when available (eg: bswap on x86).
Willy
^ permalink raw reply [flat|nested] 4+ messages in thread