In my specific case, the custom protocol runs on the same TCP port as the h248 
MEGACO protocol and relays custom information between a media gateway its 
controller.

The custom protocol uses what I would call a "magic cookie" as the first 4 
bytes following the tpkt part of the h248 message.  These bytes are chosen 
because they would never appear if the data contained was true h248.

At the moment, I have added code to the h.248 dissector to check for this byte 
stream and call the custom dissector if they are detected.

I would suppose that code could be added to packet-h248.c that would look for 
true h248 data.  If not found, return all data back to wireshark.  This might 
be a good thing anyhow.  If this approach where followed, would wireshark then 
look for any other registered dissectors that on the h248 port?

Thanks for your assistance.

Alex Lindberg

--- On Wed, 9/30/09, Guy Harris <[email protected]> wrote:

From: Guy Harris <[email protected]>
Subject: Re: [Wireshark-dev] Two dissectors on same TCP port?
To: "Developer support list for Wireshark" <[email protected]>
Date: Wednesday, September 30, 2009, 1:43 PM


On Sep 30, 2009, at 11:30 AM, Alex Lindberg wrote:

> I am creating a custom dissector that runs on a TCP port already  
> covered by one of the standard dissectors.
>
> How do can I overload the dissector registration so that if the  
> unique condistion exist for my custom dissector my dissector will be  
> used, otherwise pass control back to Wireshark?

What is the unique condition?

Is it something in the contents of the packet, or is it a preference  
setting, or is it something else?

One way to do this would be to make your dissector a heuristic  
dissector, have it check for the port number and the unique condition  
(if there's a match, dissect and return TRUE, otherwise return FALSE),  
and set the TCP preference to run the heuristic dissectors first.
___________________________________________________________________________
Sent via:    Wireshark-dev mailing list <[email protected]>
Archives:    http://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev
             mailto:[email protected]?subject=unsubscribe



      
___________________________________________________________________________
Sent via:    Wireshark-dev mailing list <[email protected]>
Archives:    http://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev
             mailto:[email protected]?subject=unsubscribe

Reply via email to