Plan for AT over RVTMUX

[Spacefalcon the Outlaw]

Well, I got my AT-over-RVTMUX idea implemented, and it seems to work.
On the target firmware side, I've only implemented it in FC GSM fw so
far, i.e., I have yet to hack it into TCS211.  On the host tools side,
there is now an fc-shell utility built in the rvinterf subtree; it
connects to rvinterf via a local UNIX socket and allows the user to
send AT commands and GPF system primitives to the target fw.  (The
underlying rvinterf pseudo-daemon has also been updated to support the
new RVTMUX channel and dynamic enabling/disabling of Rx packet
forwarding, so be sure to rebuild and install the updated version.)

Just to be clear, I have added this AT-over-RVTMUX as an additional
source channel to ATI, i.e., the "standard" UART channel (separate
MODEM UART) is still there.  With this fw running on the GTA02 modem,
I can talk AT commands to it both from the GTA02 AP and from fc-shell
running on my laptop.  The two independent AT command sources don't
seem to clash, but because our fw is not yet at a point of being able
to connect to live GSM networks properly, we don't have a way to test
yet what will happen with incoming calls etc.

My planned next step is to build a special debug version of leo2moko
with this AT-over-RVTMUX mechanism added, and see how it behaves when
used with a firmware that's more fully functional.

It also needs to be noted that for the purpose of porting our FC GSM
fw to Pirelli and Compal targets, this implementation of AT over
RVTMUX is not sufficient: we will also need to gain an ability to
build the fw with the "standard" UART channel removed before it will
make sense to port our fw to targets on which this "standard" UART
channel is inapplicable.  I prefer to leave this task for later as
there would be no immediate practical gain from doing it right now.

Happy hacking,
SF