[Michael Mann <mmann78@xxxxxxxxxxxx>]

If you create separate dissector handles with a protocol, how is a user using "Decode As" functionality supposed to know which dissection function will be picked?  That's why the check was added.

You can have the same protocol ID between TCP and UDP, but you can't have the same one twice in TCP.

 

To me there are 2 solutions

1. Create "placeholder" protocols with proto_register_protocol().  The "protocol lD" returned can be used to differentiate the different dissector function handles

2. Create a "pino" (protocol in name only) with proto_register_protocol_in_name_only.  This allows you to create "helper" protocols, with most of the "control" still with the "parent" protocol created with proto_register_protocol.

 

I believe #2 is the "preferred" solution.

 

 

 

-----Original Message-----
From: John Dill <John.Dill@xxxxxxxxxxxxxxxxx>
To: wireshark-dev <wireshark-dev@xxxxxxxxxxxxx>
Sent: Wed, Oct 25, 2017 4:44 pm
Subject: [Wireshark-dev] Duplicate dissectors (anonymous) and (anonymous) for protocol xxx

I just happened to turn on console printing to troubleshoot a different problem and I'm getting a couple of interesting messages when I change my protocol preferences.

 

Duplicate dissectors (anonymous) and (anonymous) for protocol xxx in dissector table tcp.port

Protocol <Protocol Description> is already registered in "udp" table.

 

I have a proto_reg_handoff_xxx that creates a couple of TCP port dissector handles using 'dissector_add_uint("tcp.port", MY_TCP_PORT, tcp_handle)', and a UDP heuristic dissector 'heur_dissector_add("udp", dissect_xxx_udp_heur, "<Protocol Description>", "xxx", proto_xxx, HEURISTIC_ENABLE)'.

 

My question is whether I'm doing something not recommended.  The protocol uses both TCP and UDP packets to send data, and I've lumped it under one "xxx" protocol.  For the most part, it hasn't been an issue operationally, but I'm wondering if there's a better way to do it so I don't have these messages.  I haven't looked at the console output in detail before, so I've missed this up till now.

 

Is there any advice about dissecting a protocol that appears on both TCP and UDP streams and using the same name, or what I'm doing specifically that's causing it?  The messages have the same format, but the packaging in the TCP vs the UDP streams is slightly different.

 

Best regards,

John D.

 

___________________________________________________________________________ Sent via: Wireshark-dev mailing list <wireshark-dev@xxxxxxxxxxxxx> Archives: https://www.wireshark.org/lists/wireshark-dev Unsubscribe: https://www.wireshark.org/mailman/options/wireshark-dev mailto:wireshark-dev-request@xxxxxxxxxxxxx?subject=unsubscribe