Jeff,
see my comments below.
Best regards
Michael
On Aug 18, 2004, at 6:49 PM, Jeff Morriss wrote:
Michael Tuexen wrote:
Dear all,
I want to add SCTP support in the 'Decode as' dialog. The src/dst
port stuff os easy to handle, because you know the src/dst port
of of the selected packet from the packet_info structure.
For SCTP it would also be very helpful to change the dissector
for the PPID field included in every DATA chunk. But there can
be more than 1 DATA chunk in one SCTP packet. How can I know
in the 'decode as' code, which PPIDs are present in the packet?
One possibility is to add an array of PPIDs to the packet_info
structure and fill that in the SCTP dissector. This would give
the information for the first N DATA chunks. A large N would
require a lot of memory, but, let us say N = 4, would be a compromise.
So is this acceptable, or is there even a better solution to the
problem?
I frequently see more than 4 DATA chunks in a single packet BUT those
chunks always have the same PPID. (Do any of the SIGTRAN protocols
allow sharing an association
between 2 or more of them? I doubt it but I'm not sure.)
From a protocol point of view there is no reason why you can not
do that. But I doubt that it will be done by anyone... At least
I have not seen that yet. Only RSerPool (not SIGTRAN) will
multiplex ASAP and another PPID in a single association.
So possibly it would be OK to include N PPIDs in the packet_info
and let N = 2. In the packet_info structure this is not one
PPID per DATA chunk but any PPID only once.
This will work for all applications I know about and is extendable
in the future.
So is it OK to add
- a counter of PPIDs
- an array of length 2 of PPIDs
to the packet_info structure?
Best regards
Michael
_______________________________________________
Ethereal-dev mailing list
Ethereal-dev@xxxxxxxxxxxx
http://www.ethereal.com/mailman/listinfo/ethereal-dev