Re: [PATCH 0/2][KVM] guest time accounting

[Christian Borntraeger <borntraeger-tA70FqPdS9bQT0dZR+AlfA@public.gmane.org>]

Am Montag, 13. August 2007 schrieb Avi Kivity:

> Laurent's patch gives the best of both worlds: on old 'top', you get 
> guest time accounted as user time, while on new 'top' it is accounted 
> separately.  This is done by reporting user time as the sum of the real 
> user time and guest time.  A newer 'top' can subtract guest time from 
> user time to get the correct statistic.

Yes that looks promising. If I recall correctly we had some strange top 
behaviours when we introduced the steal time. Old top added the steal time to 
idle. We should check that.

> 
> 
> > My implementation uses a similar mechanism like hard and softirq. So I 
have an 
> > sie_enter an sie_exit and a task_is_in_sie function - like irq_enter and 
> > irq_exit. The main difference is based on the fact, that s390 has precise 
> > accouting for irq, steal, user and system time, and therefore my patch is 
> > based on architecture specifc code using CONFIG_VIRT_CPU_ACCOUNT. 
> >   
> 
> Okay, so the code should be under that config option, and kvm should 
> select it.

No hurry..that was specific to our implementation, not KVM :-)
Besided that, Ingo changed the accouting with his CFS scheduler, and I still 
have to figure out how CONFIG_VIRT_CPU_ACCOUNTING can be properly integrated 
in CFS.

Christian

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >>  http://get.splunk.com/