Re: UDP checksum calculation - uses lenght twice ?

I'm not exactly sure, but I believe the reason is due to an
implementation detail of the BSD network stack. The
pseudo-header corresponds to the last 12 bytes of the
original IP packet header, but modified after processing
at the IP layer. These contain the source and destination
addresses, the TTL (which is now zero), the protocol (17
for UDP), and the IP header checksum, which is apparently
replaced with the packet length before the packet is passed
to the UDP layer for processing. Notice it is _not_ in network
byte order :)...

I hope this is the answer you were looking for all along...

--
=====================================
Alexander Nickolov
Microsoft MVP [VC], MCSD
email: agnickolov@xxxxxxxx
MVP VC FAQ: http://www.mvps.org/vcfaq
=====================================

"R.Wieser" <address@xxxxxxxxxxxxx> wrote in message
news:ux58gzC9FHA.3876@xxxxxxxxxxxxxxxxxxxxxxx
> Arkady Frenkel <arkadyf@xxxxxxxxxxxxxxxx> schreef in berichtnieuws
> #WLcVnC9FHA.2264@xxxxxxxxxxxxxxxxxxxxxxx
>
> Hello Arkady,
>
>> P.S. BTW Police officer ( in this case RFC sure not me :) describe
>> that maybe not clearly enough but "This gives the TCP protection
>> against misrouted segments" ( RFC ) is the explanation.
>
> I'm sorry, but no ammount of effort on my part in trying to explain what I
> want does seem to help ....
>
> Your above answer is as correct as the "because otherwise it does not
> function" one, but allso does *not* answer my question.
>
> I'm asking why the calculation has been designed as it curently is,
> nothing
> more, nothing less.
>
> Regards,
> Rudy Wieser
>
>
>


.

Loading