[RFC] simplify intra-mobile event passing

Harald Welte laforge at gnumonks.org
Sun Jan 30 20:12:42 UTC 2011


I've been playing with some code earlier today, trying to reuse the RR
implementation of 'mobile' but without 'MM' and '03.22' code.  Next step was to
play with RR+MM but without 03.22.

Both were not that easy to do, given the various different function calls
between those modules.

While trying to resolve it, I discovered that we have many sequences like
gsm322_msgb_alloc() followed by some error checking and a final
gsm322_plmn_senmsg().  Similar pattersn could be seen for gsm48_mmevent.

I've tried to simplify/unify this code a bit, as you can see from the
attached patches (also in laforge/mobile_event branch).  It's not tested
yet, but I would like to get some comments on it.

Initially I thought to use libosmocore signal code, but then signals are
_emitted_, and they don't fit the picture where we have some incoming events
into e.g. 03.22 code - as all the events are already labelled GSM322
and thus by the recipient, not its originator.

So now there is one gsm322_event_input() and a gsm48_mmevent_input()
function each.  Somebody who wants to reuse partial 'mobile' code can simply
provide stub functions for either one of those (or both)...

The goal of this exercise is to have a tool that can open a dedicated
channel to a user-specified cell and then send user-specified messages
while optionally taking care of loc upd / auth / ciphering in order
to make the network happy.

- Harald Welte <laforge at gnumonks.org>           http://laforge.gnumonks.org/
"Privacy in residential applications is a desirable marketing option."
                                                  (ETSI EN 300 175-7 Ch. A6)

More information about the baseband-devel mailing list