Ethereal-dev: Re: [Ethereal-dev] TCP checksum over IPV6

Note: This archive is from the project's previous web site, ethereal.com. This list is no longer active.

From: "Martin Regner" <martin.regner@xxxxxxxxx>
Date: Tue, 5 Nov 2002 21:34:48 +0100
>On Sat, Nov 02, 2002 at 04:21:47PM +0100, Martin Regner wrote:
>> Right now pinfo->dst is used when calculating these checksums.  Maybe
>> there should be a pinfo->true_dst (or similar) that could be used by the
>> upper-layer dissectors when calculating checksum or maybe pinfo->dst
>> should be set to the true destination.


Guy Harris wrote:
>Doing the latter would change what displays in the destination address
>column in the display; I'd be inclined to add a new field to the
>"packet_info" structure, and perhaps allow the a column to display
>either the immediate or the ultimate destination.
>


Yes, I guessed that there could be some negative side effects of changing pinfo->dst.

I also noticed that also when there is a "Home Address" Destination Option the
checksum seems to be wrong with Ethereal. 
The IPv6 RFC 2460 doesn't mention this issue what I could see, but it seems that the "Home address" 
shall be used instead of Source Address when calculating checksum for Mobile IPv6.
http://www.irisa.fr/tipi/doc/rfc/draft-ietf-mobileip-ipv6-19.txt

I found some sample Mobile IP captures on the following page (WLAN.cap and LAN.cap):
http://siving.hia.no/ikt01/ikt6400/talund99/
Use a filter "icmpv6.checksum_bad" and you will see some packets with bad checksum
when "Routing header" or "Destination Option - Home Address" is used.

Regards,
 Martin