Upgrading QtMoko kernel

Radek Polak psonek2 at seznam.cz
Tue Jan 29 15:24:06 CET 2013


On Saturday, January 26, 2013 11:01:29 AM Radek Polak wrote:

> On Friday, January 25, 2013 08:32:20 PM NeilBrown wrote:
> > On Fri, 25 Jan 2013 11:38:08 +0100 Adrien Dorsaz <adrien at adorsaz.ch> 
wrote:
> > > Hello!
> > > 
> > > I've just seen this commit :
> > > https://github.com/radekp/qtmoko/commit/683ff99d7262950da179a30b59cd6b4
> > > fa dd80454
> > 
> > Hmmm.. that patch is
> > 
> > +mount -t debugfs none /sys/kernel/debug/
> > +echo 1 > /sys/kernel/debug/pm_debug/enable_off_mode
> > 
> > which is probably a bad idea.
> > Most kernels will behave badly with off_mode enabled.
> > Those for which I've fixed the various issues have
> 
> Ahh oki, i have reverted it - i tought it works like that because i could
> see low currents even without writing 1 > enable_off_mode.
> 
> Btw whithout this patch the kernel twice didnt wake from suspend. Then i
> hacked this patch and it didnt happen again. My idea is that whithout
> debugfs mounted the timing during resume is different. Anyway i'll try
> running without the patch for a few days to check if it happens again.

I did some more tests today. It happened again - but it's not that the phone 
didnt wake. It's more likely that only the display brightness was not 
restored. I plugged USB and restarted QtMoko and everything was fine.

I spent rest of the day trying to reproduce it, but without success. I have 
also checked QtMoko resume code - it should write correct value to 
/sys/../backlight../brightness - maybe the call failed or write to the file 
failed - unfortunately i didnt have logging enabled so cant tell right now.

So suspend looks fine - i tried like hundreds of times and it's stable.

Regards

Radek



More information about the community mailing list