tcpdump mailing list archives

A puzzled maintainer with questions regarding DLT_ values


From: "M.Baris Demiray" <barisdemiray () gmail com>
Date: Wed, 19 Jan 2011 11:02:50 +0200

Hello,

I've a couple of questions regarding dissector development and DLT_
values and I hope I know enough to express myself.

First of all, we're developing radio communication software,
particularly STANAG 5066. Alongside the serial interface for the modem
interface our implementation also has a socket interface for testing
purposes.

Now we have the need of saving STANAG 5066 PDUs into .pcap files to
analyze them later using our experimental STANAG 5066 DPDU (Data
Transfer Sublayer PDU) dissector. My ex-colleague, who wrote the
dissector, wrote it for the purpose of dissecting DPDUs captured from
the socket interface with prepending TCP header. But I, as the new
maintainer of this dissector, need to dissect raw STANAG 5066 PDUs now
since I modified the Data Transfer Sublayer (of STANAG 5066)  to
maintain a .pcap file and to save incoming DPDUs into this file with
relevant record headers. But, at this point, I'm stuck since when I to
open this .pcap file using Ethereal (this is not because I'm against
updates but because the system is offline) it successfully displays
the packets yet when I attempt to "Decode " these packets "As..." I
can't because that menu item is grayed out.

By the way, I intuitively modified the dissector to make it use
dissect_s5066dts() method rather than dissect_s5066dts_tcp() upon
registering itself since I want it to dissect STANAG 5066 PDUs without
any other prepending protocol headers. So I tried to place a couple of
DLT_ values (such as DLT_USER0) into the global header's network field
with no luck. At this point, should I ask for a new DLT_ value for
PDUs which start with 0x90EB sync. sequence and without any other
prepending header?

Please feel free to ask for further information, or to correct my
terminology, or to offer a different way of doing this.

Cheers,

PS: As soon as I ensure that I have a robust-enough dissector it will
be sent as a patch.

-- 
M. Baris Demiray
-
This is the tcpdump-workers list.
Visit https://cod.sandelman.ca/ to unsubscribe.


Current thread: