[maemo-users] Dealing with N900 responsiveness (or lack of thereof)

[From: Michel Dänzer]

On Wed, 2010-02-24 at 17:10 +0100, Michel Dänzer wrote: 
> On Tue, 2010-02-23 at 11:28 -0600, Paul Hartman wrote: 
> > 2010/2/23 Michel Dänzer <michel at daenzer.net>:
> > >
> > > When the device is (supposedly) idle, make sure that e.g. htop
> > > doesn't show any significant CPU usage by any process other than itself
> > > (and Xorg and osso-xterm, if you're running it on the device display).
> > 
> > I haven't noticed anything, and my load average at the moment is 0.00,
> > 0.02, 0.00 which is good, but if I sort by time the top offenders
> > since my last reboot (uptime 1 day, 3:14) are:
> > 
> > 5:41 pulseaudio
> > [...]
> > 2:41 pulseaudio
> > 1:11 pulseaudio

(BTW, why several pulseaudio entries?)

> Maybe you're suffering from the tonegend bug discussed in the other 
> sub-thread. I encountered that once and it did cause the battery to go
> down very quickly even though pulseaudio was only using around 3-5% CPU
> (but constantly).

It actually happened again yesterday, and the battery dropped by about
20% in about 4 hours (according to Battery Eye) before I noticed, and
now the cumulative CPU time for pulseaudio looks similar to yours.

So if you ever see pulseaudio using non-0 CPU when there's no sound
playing, try killing tonegend or pulseaudio itself. The killed process
will be restarted automagically.


-- 
Earthling Michel Dänzer           |                http://www.vmware.com
Libre software enthusiast         |          Debian, X and DRI developer