[Community] [Gta04-owner] What software are we still missing on GTA04 distributions?

[Neil Jerram]

NeilBrown <neilb at suse.de> writes:

> On Thu, 26 Jul 2012 10:30:48 +0200 Radek Polak <psonek2 at seznam.cz> wrote:
>
>> On Thursday, July 26, 2012 12:52:15 AM NeilBrown wrote:
>> 
>> > Reliable firmware for the GSM module?
>> 
>> Me too ;-) 
>> 
>> > I've recently started noticing the problem where an incoming call causes
>> > the firmware to crash on occasion.  I think Radek avoided that by
>> > disabling 3G or similar (Radek: what was the command to do that?).
>> 
>> This is from QtMoko's modem plugin [1]:
>> 
>> 	chat("AT_OPSYS=0,2");       // disable UMTS, use only GSM
>
> I finally tried this.
> My phone was not accepting or making calls, though it reported a reasonable
> signal strength.
> I queried the _OPSYS setting:
>
> OK
> AT_OPSYS?
> _OPSYS: 0,2
>
> OK
> AT_OPSYS=?
> _OPSYS: (0-5),(0-6)
>
> OK
> AT_OPSYS=0,2
> OK
> AT_OPSYS?
> _OPSYS: 0,2
>
>
> found that it was already 0,2, but set it anyway.  Then I tried making a call
> and it worked.  Might have been lucky. Maybe not.
>
> So I did some reading.  Lots of places will tell me what the first number
> means:
>  0 - GSM or 2G or GPRS (opinions vary)  only
>  1 - UMTS or 3G only
>  2 - prefer 2G over 3G
>  3 - prefer 3G over 2G
>  4 - keep using whatever you are using now
>  5 - automatic selection.
>
> Then I remembered that my GSM data connection was strangely slow..
> If I set _OPSYS to 3, I get about 300KB/sec.
> If I set _OPSYS to 0, I get about 10KB/sec.
>
> So... weird.  Did setting it explicitly fix something?  Will that always fix
> something no matter what it is set to?  Why does it default to the slowest
> option?  Does it really?
>
> I think I might need apps to enforce different OPSYS setting so I can
> experiment if it fails again.
>
> But I'm happy that I now have fast mobile internet.
>
>> 
>> I also added many other workarounds in QtMoko for the GSM. I think it's 
>> working quite good now, but not perfectly. Here is list of problems and 
>> workarounds:
>> 
>> * restart QtMoko after modem crashed. AT_OPSYS=0,2 does not completely solve 
>> the crashing and after the crash the serial line stops working, so i restart 
>> whole QtMoko [2].
>
> I just close/reopen whenever the tty port stops working.
>
>> 
>> * sometimes after resume the serial line is silent - i am closing the serial 
>> port before suspend and open it after resume and doing CLCC and SMS poll after 
>> opened.
>
> I don't close over resume, but I do poll for SMS if the KEY_UNKNOWN appears
> on /dev/input/incoming.
>
>
>> 
>> * very rarely the modem after resume simply disappeared - the serial port file 
>> /dev/ttyHSO_Application was not existing at all.
>
> I had that once that I know of.  I had to reboot.  Maybe with the A4 you can
> power-cycle.
>
>> 
>> * cant detect other side hangup in outgoing call - i do CLCC polling.
>
> I open /dev/ttyHS_Modem as well as ttyHS_Application.  I get a "NO CARRIER"
> message on the _Modem port when the other end hangs up (if the network
> notices.  If I get an incoming call to my home wire-line phone, I can hang up
> then pick up again and the call is still active.  Of course I don't get NO
> CARRIER in that case, but I do for mobile-to-mobile).
>
>> 
>> I am running the phone mostly with AT logging on. But sometimes you dont have 
>> time to analyze what happend.
>> 
>> Currently the modem with above workarounds works quite good, but i think many 
>> people expect it to be 100% perfect...
>> 
>> Regards
>> 
>> Radek
>> 
>> > So if there were some way to try out a newer firmware, I'd be keen to give
>> > it a go.
>
> So if setting OPSYS turns out to reliably fix my problem, it looks like I
> don't need new firm ware.

I don't quite understand, from the above, what OPSYS values you are
recommending.  Initially (IIRC) you were trying 0,2, but later you said
that 3,X is needed for a fast data speed, so...?

Thanks,
        Neil