On Mon, 2011-01-31 at 00:45 +0800, lidong chen wrote:
> I think we can use performance counter.
> use unhalted core cycles event, in the nmi callback funcation, count
> which process is running .
> if the vm exit is caused by nmi,discard it.
> the system time of qemu process is the time steal by kvm.
Performance counters are a scarce resource, so I'd rather not use them,
since it will mean forcing a context switch from whoever is using it at
the moment. Which is also an expensive operation anyway.
So even though it can be possible, in theory, I don't see why use it
in this particular case.
>
>
> 2011/1/30 Avi Kivity <[email protected]>:
> > On 01/28/2011 09:52 PM, Glauber Costa wrote:
> >>
> >> This patch accounts steal time time in kernel/sched.
> >> I kept it from last proposal, because I still see advantages
> >> in it: Doing it here will give us easier access from scheduler
> >> variables such as the cpu rq. The next patch shows an example of
> >> usage for it.
> >>
> >> Since functions like account_idle_time() can be called from
> >> multiple places, not only account_process_tick(), steal time
> >> grabbing is repeated in each account function separatedely.
> >>
> >
> > I accept that steal time is worthwhile, but do you have some way to
> > demonstrate that the implementation actually works and is beneficial?
> >
> > Perhaps run two cpu-bound compute processes on one vcpu, overcommit that
> > vcpu, and see what happens to the processing rate with and without steal
> > time accounting. I'd expect a fairer response with steal time accounting.
> >
> > --
> > error compiling committee.c: too many arguments to function
> >
> > --
> > To unsubscribe from this list: send the line "unsubscribe kvm" in
> > the body of a message to [email protected]
> > More majordomo info at http://vger.kernel.org/majordomo-info.html
> >