Re: [wmii] wmii hg devel, increased cpu wakeups-from-idle

2009-03-16 Thread Kris Maglione

On Tue, Mar 17, 2009 at 12:21:33AM +0100, Martin Kraus wrote:

You were right. Started with just /bin/true in wmiirc and with urxvt and
powertop I'm under 10 wakeups. So the problem is with the scripts. I'm using
standard scripts that come with devel release. Guess it's time to read through
them to find out what is happening. 


My best guess is that the status script is causing the problem. 
Try disabling the 'Action status' bit of wmiirc. I'm guessing 
that one or more of of date(1), uptime(1), or wmiirc(1) aren't 
making the buffer cache. If stopping the status script solves 
your problem, try sticking wmiir on a ram disk (it's statically 
linked by default), and ideally finding a statically linked 
date(1) and uptime(1) to do the same.


--
Kris Maglione

Question with boldness even the existence of God; because if there be
one, He must approve the homage of Reason rather than that of
blindfolded Fear.
--Thomas Jefferson




[wmii] wmii hg devel, increased cpu wakeups-from-idle

2009-03-16 Thread Martin Kraus
> On Thu, Mar 12, 2009 at 11:39:06PM -0400, Kris Maglione wrote:
> > On Thu, Mar 12, 2009 at 03:37:32PM +0100, Martin Kraus wrote:
> >> Hi. I've migrated from the "stable" wmii(20080530 or similar) to 
> >> development
> >> version and I've noticed increased wakeups-from-idle when measured with
> >> powertop. It used to be something like 13 wakeups, now its 40+. I've 
> >> tested simple X(without any wm) with xterm and I get 2 wakeups in this 
> >> way,
> >> so I guess it's not a problem with the rest of my system.
> >>
> >> Also the weird thing is, powertop doesn't show any processes with many
> >> wakeups. the largest event is hrtimer_start with 3.5 wakeups. If I sum up 
> >> the
> >> rest of what powertop shows I get something around 6 wakeups so there must
> >> be either hundreds of little wakeups or there is something seriously wrong.
> >>
> >> Is anyone else experiencing similar problem? Could this really be related 
> >> to
> >> wmii or is it something else?
> >
> > That's strange. I can't really think of anything wmii would do to trigger 
> > wakeups, except perhaps in the rc code. Have you tried without any status 
> > scripts running? 

You were right. Started with just /bin/true in wmiirc and with urxvt and
powertop I'm under 10 wakeups. So the problem is with the scripts. I'm using
standard scripts that come with devel release. Guess it's time to read through
them to find out what is happening. 

mk



Re: [wmii] wmii hg devel, increased cpu wakeups-from-idle

2009-03-16 Thread Martin Kraus
On Thu, Mar 12, 2009 at 11:39:06PM -0400, Kris Maglione wrote:
> On Thu, Mar 12, 2009 at 03:37:32PM +0100, Martin Kraus wrote:
>> Hi. I've migrated from the "stable" wmii(20080530 or similar) to development
>> version and I've noticed increased wakeups-from-idle when measured with
>> powertop. It used to be something like 13 wakeups, now its 40+. I've 
>> tested simple X(without any wm) with xterm and I get 2 wakeups in this 
>> way,
>> so I guess it's not a problem with the rest of my system.
>>
>> Also the weird thing is, powertop doesn't show any processes with many
>> wakeups. the largest event is hrtimer_start with 3.5 wakeups. If I sum up the
>> rest of what powertop shows I get something around 6 wakeups so there must
>> be either hundreds of little wakeups or there is something seriously wrong.
>>
>> Is anyone else experiencing similar problem? Could this really be related to
>> wmii or is it something else?
>
> That's strange. I can't really think of anything wmii would do to trigger 
> wakeups, except perhaps in the rc code. Have you tried without any status 
> scripts running? 

you were right. ran it with only /bin/true in wmiirc and got less then 10
wakeups with only rxvt and powertop running. i'm using standard
wmiirc/wmii.ssh that comes with wmii devel. time to dig into those scripts to
finaly learn how it works.

thanks 
mk



Re: [wmii] wmii hg devel, increased cpu wakeups-from-idle

2009-03-12 Thread Kris Maglione

On Thu, Mar 12, 2009 at 03:37:32PM +0100, Martin Kraus wrote:

Hi. I've migrated from the "stable" wmii(20080530 or similar) to development
version and I've noticed increased wakeups-from-idle when measured with
powertop. It used to be something like 13 wakeups, now its 40+. 
I've tested simple X(without any wm) with xterm and I get 2 wakeups in this way,

so I guess it's not a problem with the rest of my system.

Also the weird thing is, powertop doesn't show any processes with many
wakeups. the largest event is hrtimer_start with 3.5 wakeups. If I sum up the
rest of what powertop shows I get something around 6 wakeups so there must
be either hundreds of little wakeups or there is something seriously wrong.

Is anyone else experiencing similar problem? Could this really be related to
wmii or is it something else?


That's strange. I can't really think of anything wmii would do 
to trigger wakeups, except perhaps in the rc code. Have you 
tried without any status scripts running? 


--
Kris Maglione

Beware of "the real world".  A speaker's appeal to it is always an
invitation not to challenge his tacit assumptions.
--Edsger W. Dijkstra




[wmii] wmii hg devel, increased cpu wakeups-from-idle

2009-03-12 Thread Martin Kraus
Hi. I've migrated from the "stable" wmii(20080530 or similar) to development
version and I've noticed increased wakeups-from-idle when measured with
powertop. It used to be something like 13 wakeups, now its 40+. 
I've tested simple X(without any wm) with xterm and I get 2 wakeups in this way,
so I guess it's not a problem with the rest of my system.

Also the weird thing is, powertop doesn't show any processes with many
wakeups. the largest event is hrtimer_start with 3.5 wakeups. If I sum up the
rest of what powertop shows I get something around 6 wakeups so there must
be either hundreds of little wakeups or there is something seriously wrong.

Is anyone else experiencing similar problem? Could this really be related to
wmii or is it something else?

thanks
mk