[Gta04-owner] QtMoko: how is long Power key press (=> shutdown/restart menu) handled?

[Dr. H. Nikolaus Schaller]

Am 06.12.2012 um 13:22 schrieb Radek Polak:

> On Wednesday, December 05, 2012 11:36:34 PM Neil Jerram wrote:
> 
>> Therefore I doubt that adopting the serial state kernel fix was a good
>> reason for removing "AT_OPSYS=0,2", and I think that people who don't
>> want "AT_OPSYS=0,2" (such as me) might be better advised to keep
>> installing restart-when-modem-stops-working.patch.  It's probably better
>> to restart QtMoko, even though that might lose some application state,
>> than to leave the phone not working and draining lots of power.
> 
> The patch is still applied when building the debian package and it looks that 
> we cant get rid of it yet.
> 
> Maybe we could rework it so that it does not touch code in serial port class, 
> move it completely to gta04 plugin and commit it.
> 
> Or we could commit - ugly as it is - in case that reworking is not possible 
> and if it would help people building from sources.
> 
>> Or have I misunderstood some part of this?
>> 
>> BTW, did we ever establish that "AT_OPSYS=0,2" 100% avoids
>> reenumerations?
> 
> I dont know way how to 100% avoid it.
> 
> 1/ without AT_OPSYS=0,2 i could make the reenumeration happen always.
> 
> 2/ with AT_OPSYS=0,2 it happened very rarely.
> 
> 3/ with kernel patch + 3G enabled it happened just once in last 3 weeks or so.
> 
> 
> Btw in latest release when the reenumeration happens, it's logged to 
> 
> 	/modem_reenumerate.log

Does this include part of dmesg or can be made so? I.e. "dmesg | tail -n 50 >>/modem_reenumerate.log" or similar?

It may be that the modem is sending some other code that the HSO driver
is not completely aware of and the HSO driver may report something.

> 
> Could people who use GTA04 as phone report contents of the file? This could 
> help us better to monitor the situation.
> 
> Regards
> 
> Radek

BR,
Nikolaus