Hi all,

I've implemented an AX.25 connected mode 'convergence layer' for the Delay Tolerant Networking reference implementation (see dtnrg.org). I'd like to submit it upstream soon, but I've observed some quirks in testing that I'd like to understand better.

My implementation of the AX.25 Connected mode convergence layer (ax25cm-CL) uses Linux support for AX.25 sockets and has been mostly tested using a PK-232 (KISS mode) with a hard-wired loop-back cable, although a little OTA testing has been done elsewhere with no detrimental reports.

The quirk I'm observing has symptoms which prevent subsequent communications between two SSIDs on the same host. The problem shows up in netstat output as a socket in the listening state, but unusually it has call/SSID in both source and dest fields, which strikes me as being an anomaly. This entry in netstat output is in addition to the original listening socket, from which the connection threads with their own connected sockets are spawned.

When further attempts are made to set up a connection between the two call/SSIDs, an "Address already in use" error occurs. Note please that the error message is from recollection, not verbatim, as I don't have access to the TNC at the moment.

Can anyone shed any light on this behaviour. I've noticed similar error messages when dabbling with Bluetooth connections under circumstances where I already have an rfcomm connection established and erroneously try to bring up another identical one. This makes me wonder if there is a resource leak for an AX.25 socket in my ax25cm- CL implementation, when a connection 'closes' due to errors. In my loopback testing, I get quite a few errors, which seem to be mitigated somewhat by adjustment of the timing parameters.

FYI, the test system with the PK-232 is running Ubuntu 7.10.

Any help would be appreciated.

Cheers,

Darren, G0HWW


-
To unsubscribe from this list: send the line "unsubscribe linux-hams" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to