Calypso dance (GSM)

[Werner Almesberger]

A few weeks ago, we redoubled our efforts to solve some of the
longstanding problems in the GSM subsystem. The main objective was
to finally solve the recamping problem, bug #1024.

There have been some small good news: we managed to get TI's debugging
and tracking utility PCO2 to run under Wine.

Then there was a major breakthrough: independently, Mike Westerhof
found the AT%SLEEP=2 work-around.

And then the bummer: we discovered that we had lost our Calypso
firmware sources :-( I'll skip the details to protect the guilty,
but things looked rather bleak at that moment. Fortunately, we have
since been able to recover the moko8 sources, and TI have sent us the
updates that had been integrated since then.

Around that time, Dieter Spaar, who is already familiar with that kind
of firmware has contacted Joerg, and asked if we could use some help.
Joerg then convinced the rest of us that this was indeed something
worth trying :-)

So, as a first step, we've contracted Dieter to bring that bag of code
pieces into a buildable state again. This doesn't change the status
quo regarding bugs and misfeature, but it's a precondition for
actually doing anything about the problems.

Now, this work is still in progress, but we can already look ahead at
what issues we hope to be able to tackle:

- #1024, everyone's favourite failure to use the network

- #666, 3G SIM cards not being recognized

- production may want to replace a memory chip, which may need a
  small firmware change

- suppressing/reducing the handset-mic audio feedback (which causes
  echo for the remote party)

Is there anything else directly relayed to the GSM firmware that
should be on this list ?

- Werner