Wireshark-bugs: [Wireshark-bugs] [Bug 7902] Improved Dissection of Modbus/TCP messages and added

Date: Tue, 23 Oct 2012 08:20:22 -0700 (PDT)
https://bugs.wireshark.org/bugzilla/show_bug.cgi?id=7902

--- Comment #20 from Chris Bontje <cbontje@xxxxxxxxx> 2012-10-23 08:20:21 PDT ---
(In reply to comment #19)

> Are you talking:
> 1. Modbus RTU over "raw" Ethernet (not TCP/IP)?  If that's the case the Modbus
> RTU dissector would need an different/additional hook into the Ethernet frame,
> not a port number.  It appears some of the captures provided display this
> behavior.  I would differ to others as to the best way to do that.
> 
> 2. Modbus RTU over TCP/IP where Unit ID + Modbus PDU is sent over Port 502
> instead of the MBAP header?  I would technically consider this a "violation" as
> port 502 is reserved for Modbus/TCP.  From an implementation standpoint I would
> create a preference for the "Modbus RTU port", but default it to 0, not 502 to
> prevent inadventent collisions.  Users would then have to pick between
> Modbus/TCP and Modbus RTU by either changing ports or disabling one of the
> protocols.  There is also the possibility of not assigning it a port at all and
> just making it available to the "Decode As..." functionality.
> 
> 

#2 is what is typically seen, but port 502 is not used; most of what I see is
Modbus RTU serial data encapsulated into a standard TCP/IP message, over a
configurable port (ie: 10001 in the Modbus_Tunneled_* examples above).  I think
a separate port configuration option for Modbus RTU over TCP makes the most
sense if we are to keep the protocols completely independent.  I think
"decode-as" will not allow us seamless use of the "packet direction detection"
code, which checks if a given port number is the src or dst in a given packet.

-- 
Configure bugmail: https://bugs.wireshark.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.