Re: DLT request for SDLC (BITBUS)

[Guy Harris <guy () alum mit edu>]

On Jan 16, 2017, at 4:06 AM, Becker (ELZET80) <Becker () elzet80 de> wrote:

> we are currently working on an USB tracer which captures messages from BITBUS connected devices.
> BITBUS is based on the SDLC protocol by IBM (see: https://en.wikipedia.org/wiki/Synchronous_Data_Link_Control)

Several protocols, including HDLC, LAPB, IEEE 802.2, etc. are, directly or indirectly, "based on the SDLC protocol".

However...

> We have already seen that there is a SDLC dissector implemented in Wireshark

...the SDLC dissector is intended to dissect IBM's SDLC, and assumes that the payload of an Information frame is SNA.

According to

        http://www.bitbus.org/dnl/BITBUS%20Interconnect%20Specification.PDF

        1) "The BITBUS interconnect does not (and is not intended to) maintain strict SDLC compatibility"

and 

        2) what's described by section 5.0 "MESSAGE PROTOCOL SPECIFICATION" isn't SNA.

So we should not treat the BITBUS data link layer as SDLC, but as Yet Another SDLC-Based Protocol.

> but as we are getting our data from our tracer tool there is no ethernet layer here.
> Is there a way to use the SDLC dissector as a DLT already?

No, and if you were to do so, you probably wouldn't like the results, unless you're using SNA on the BITBUS links.

> If not we would request a DLT for SDLC communication on which we would base our BITBUS lua dissector plugin.

You should, instead, get LINKTYPE_BITBUS/DLT_BITBUS.

Would those packets begin with the Address field of the BITBUS header, followed by the Control field, followed by the 
Information, with the Flag fields not included. and with bit-stuffing removed?  Would they include the FCS, or not?
_______________________________________________
tcpdump-workers mailing list
tcpdump-workers () lists tcpdump org
https://lists.sandelman.ca/mailman/listinfo/tcpdump-workers