On Mon, 30 May 2011 11:03:35 +0900
Minoru Usui <usui(a)mxm.nes.nec.co.jp> wrote:
On Mon, 30 May 2011 09:36:50 +0900
KAMEZAWA Hiroyuki <kamezawa.hiroyu(a)jp.fujitsu.com> wrote:
> On Wed, 25 May 2011 15:01:29 +0900
> Minoru Usui <usui(a)mxm.nes.nec.co.jp> wrote:
>
> > Hi,
> >
> > This is v6 of virNodeGetCPUTimeParameters() API.
> > (This time, I rename it to virNodeGetCPUStats())
> > It returns cpu utilization or
> > cumulative cpu time of the node from /proc/stat since node boots up.
> > This patch only supports linux host.
> >
>
> Can I make a question ?
>
> What "CPU' means here ?
>
> CPU usage statistics from the Guest OS's point of view ?
>
> Do we get the information by asking an daemon or module on the guest
> to gather statistics on the guest ?
No.
This API gets host cpu statistics, not guest.
It's just read statistics from /proc/stat of the host.
ok.
I have a patch to get per-domain/per-cpu cpu usage statistics by using
cpuacct cgroup.
(for fixining #virt-top -1.)
yours are per-conn, mine is per-domain.
I'd like to remake my patch to use the same style as yours and
to use the same macro/params.
As...
==
int virDomainGetCPUStats (virDomainPtr dom,
virCPUStatsPtr params,
int *nparams,
unsigned int flags);
==
I think "kernel" and "user", for linux.
Thank you for your patch.
Thanks,
-Kame