[Dibbler] [Bug report]Bug in Server which receives layerd Relay-agents Relay-Forward message

[Tomasz Mrugalski]

On Fri, 27 Apr 2007 somebody known as Hideshi.Enokihara at jp.yokogawa.com wrote:

> I'm a new comer in this ML and the dibbler, So, I don't know the correct
> bug report point.
It's ok to report bugs on one or the other mailing list.

> If this ML is not correct for this purpose, please let me know the
> correct point.
>
> I'm using dibbler as server in following topology.
>
> Client------network1-------Relay Agent1-----network2--------Relay
> Agnet2------Server
>
> Cleint:Windows Vista
> Relay Agent1:Sourceforge DHCPv6
> Relay Agent2:WIDE-DHCPv6
> Server:dibbler
>
> In this topology,
> Client sends a Solict, then Relay Agent1 relays this messate to Relay
> Agnet2.
> Relay Anget2 sent Relay-Forward Solicit message to ff05::1:3.
> After that, Server sends a Relay-Reply message.
>
> But the Relay-Reply message is malformed.
> The message does not include layed Relay message.
> Please refer attached file, No.25.
I think you are reffering to frame 27. Well, it seems malformed. Take note 
that the reply is there, but Ethereal does not shows it (take a look at 
hexdump, not the option list).

Anyway, that is a problem and I'll fix is soon. Luckily, we have 2 
national holidays in the upcoming week, so I'll have more time to spend on 
dibbler. Expect this to be fixed.

> I would like a patch for this.
> Could someone help me, please?
Yes, since you've sent a traffic capture, I can make a fix for this. A
here's a word of explanation. I believe the root cause of this issue is 
that one of the relay put options in this ordeer : relay-message, 
interface-id, while the second did it in reverse (interface-id, 
relay-message). Of course both approaches are valid and I confirm that 
this is a bug in dibbler-server.

Just a thought: if you are in a hurry, you could try to use following 
workaround: use only one relay type (the one, which sends interface-id 
before relay-message). I may work.

> Thank you for your cooperation!.
You are most welcome. Thanks for this bug report. I knew that there is 
some issue related to cascade relays. That's bug 156:
http://klub.com.pl/cgi-bin/bugzilla/show_bug.cgi?id=156
I have reopened it.

Best regards,
-- 
Tomasz Mrugalski,              | "I think there is a world market for     |
thomson(at)klub(dot)com(dot)pl |  about five computers."                  |
                                |     Thomas J. Watson (Chairman, IBM) 1943|