Fix droid4 ofono bugs/quircks

[MerlijnWajer]

10:56 < Wizzup> tmlind: do you have a brief explanation of how those ofono kicks work for the droid 4? I want to see if I can fix the problem(s), for activating internet data, sim detection and sometimes not flushing messages out
11:07 < Wizzup> can't we just kick ofono on any message, or error if there is no handler for it at least
11:07 < Wizzup> I think it looks like for data there are messages like voice call state and maybe the voice call handler gets in the way of those
11:09 < tmlind> Wizzup: so it's based on the "feature" you noticed a few years ago where qmi usb modem sees the status briefly after a serial port event
11:10 < tmlind> i never found a way to re-enable the qmi usb interrupts, they all only trigger on the serial port ts27010 mux only
11:11 < tmlind> if some event is missing, chances are a handler needs to be added for it on the serial port mux, then kick the usb modem to see the event
11:11 < tmlind> i've been meaning to look at that stuff again but who knows when i get a chance
11:17 < Wizzup> tmlind: ok, yeah not asking for you to do it, just trying to understand how we can improve this stuff
11:19 < tmlind> Wizzup: well you can see the uart messages in dmesg with modprobe n_gsm debug=0xff, if there are some unhandled events just kick the usb modem based on those
11:23 < tmlind> mot_qmi_trigger_events() is the call that gets the usb modem to read state and do things
11:25 < Wizzup> check, thanks
11:25 < Wizzup> what I meant is what perhaps it gets routed to the wrong handler somehow, but maybe that's not possible
11:26 < tmlind> likely some unhandled events happening on the uart mux ports somewhere
11:27 < tmlind> most of the handlers just call mot_qmi_trigger_events() and let kernel qmi driver take care of things
11:28 < Wizzup> ok, I'll try to get some more dumps
11:28 < tmlind> also, the modem firmware on d4 vs bionic use different line end :( maybe some of it is not handled properly if one behaves but the other does not
11:29 < tmlind> i ended up doing most of the development against bionic firmware as that allowed a cheap t-mo sim
11:30 < Wizzup> understood, do we have a write up of what line/data goes to what tty? is that just in kernel?
11:31 < tmlind> there are some defines in ofono if you grep for dlci
11:31 < Wizzup> ok
11:32 < tmlind> switching over to use ell api instead of the gatchat would make it easy to get rid of the firmware specific reponse line end issue
11:32 < tmlind> that's only for the serial port mux communcation, no need to use ell for anything else

[MerlijnWajer]

Also see:

• https://github.com/maemo-leste/bugtracker/issues/530
• https://github.com/maemo-leste/bugtracker/issues/445

[IMbackK]

this is mostly fixed, sms hanging in output que remains a problem with a seperate bug.