yaml rule to stop suspending when using gps

Klaus Kurzmann mok at mnet-online.de
Sat Feb 28 16:00:00 CET 2009


* Helge Hafting <helge.hafting at hist.no> [090227 11:16]:
> Klaus Kurzmann wrote:

> >> org.shr.ophonekitd.Usage /org/shr/ophonekitd/Usage
> >> org.shr.ophonekitd.Usage.RequestResource Display
> > which requests the resource until you release it. That way you can wrap
> > for example tangogps between a call to request display and a call to
> > release display...

> This is currently very fragile.
yep... at least it was :-) should be fixed now.

> Because it didn't work at all, I upgraded SHR last night.
> Today, the backlight stayed on, I could see my map all the way to work.

> I then quit tangogps and expected the backlight to time out.
> It did not! So no I have a phone with the backlight stuck on,
> a very power hungry condition.

> Hm - on closer inspection I see that ophonekitd has crashed.
> Its resources are not released unconditionally when the process ends?

> I didn't notice the crash at first, because the phone works so well
> without. I can send and receive SMS and call out. (the new message 
> doesn't show, but the sound is there and I can read it in the messages 
> app.)  The only thing missing is that there is no way to answer the 
> phone when it rings.

> I restarted ophonekitd and tried to release the resource again.
> ophonekitd ended without error message, and mdbus complained:

> /org/shr/ophonekitd/Usage: ReleaseResource failed: 
> org.freedesktop.DBus.Error.NoReply

> I hope this isn't too hard to fix. I plan on using scripts requesting 
> the CPU for all data-gathering apps (tangogps, cellhunter) so the phone 
> don't suspend while they are in use.

> And I also plan on requesting the display for mokomaze and tangogps. 
> Because these should be on for a long time without touch input. 
> Actually, I have two tangogps icons - one that keeps the backlight on 
> for car navigation, and one that merely reqeust the cpu for data 
> gathering without external power.

> I assume ophonekitd uses reference counting, so multiple apps can 
> request the display/cpu at the same time? And release one after another, 
> with suspend happening only after the last user has quit?
it did not... but does now. 


> Is there a way to check what resources are held and by who? Sort of like 
> a "ls" command for resources?  Is it possible to forcibly release the 
> display so the phone can work normally again - except for the obvious 
> reboot way?
in addition to the ophonekitd bugs there was also a bug in frameworkd
regarding the release of the display resource. All that should be fixed
now and hit you with opkg update & upgrade



> Helge Hafting

Klaus 'mrmoku' Kurzmann





More information about the community mailing list