[procps] Re: C-States handling - new switch?

On Jan 25, 2012, at 7:12 AM, Jaromir Capik wrote:
> Do You consider it a good idea to introduce a new
> switch for considering the off state as idle state?

Hi Jaromir,

Some random thoughts on your email...

. Your output could not have been produced by the command line shown.

. Perhaps the procps-ng top should have been used, not the 3.2.8 version.

. Maybe the kernel should provide a new /proc/stat cpu "off" state category.

. If we had a new state, we've already exhausted an 80 column display.


Some possible approaches...

. I could force a 100% idle result for the off state using a single line of 
code subordinate to an existing conditional.  But then one would loose the 
ability to distinguish an idle cpu from an one which was turned off.

. Maybe we should keep the current 'anomaly' and simply address the off state 
potential in the man document.

. A user switch (or compile option?) could be added so if no tics were recorded 
for a particular cpu the displayed line should simply say "100.0% off" or just 
"off" with no other categories shown.  However, this approach might be 
unpleasant under a small delay interval.

Comments would be welcomed.

Regards,
Jim

Other related posts: