[Pkg-libvirt-maintainers] Bug#672720: Bug#672720: libvirt-bin: KVM with usb-tablet causes constant CPU load

Guido Günther agx at sigxcpu.org
Mon May 14 19:24:56 UTC 2012


On Mon, May 14, 2012 at 06:31:35PM +0300, Kari Pahula wrote:
> retitle 672720 libvirt-bin: KVM with usb-tablet causes constant CPU load
> thanks
> 
> On Sun, May 13, 2012 at 06:23:39PM +0200, Guido Günther wrote:
> > I'd check the monitor commands issued by libvirt and reproduce this with
> > stand alone kvm. Libvirt shouldn't be doing anything special here and
> > you should be able to check kvm's state. See also:
> 
> I think I found the culprit: usb-tablet.
> 
> I tested this with a minimal kvm command line and saw the cpu load
> drop to zero, even without pausing anything, with a running but idle
> virtual machine.  When I started dropping options from virsh's kvm
> command line I saw that removing "-device usb-tablet,id=input0"
> stopped kvm's cpu hogging.

Yes the usb table hogging the cpu is a known issue:

	http://lists.gnu.org/archive/html/qemu-devel/2010-04/msg00149.html

however the cpu load should drop to zero even when you have the usb
tablet enabled.

Did you create the vm using virt-manager? It's adding the tablet by
default since we can't easily distinguish between systems that need it
and those that don't.


> I'm using my virtual machines as development environments for web
> sites and can just drop <input type='tablet' bus='usb'/> from my virsh
> domain config without trouble.
> 
> I'll leave it to you to decide what to do with this information.  Can
> anyone reproduce this?

Yes. Saving cpu cycles without the tables is reproducable here.

Cheers,
 -- Guido
> 
> 
> 
> _______________________________________________
> Pkg-libvirt-maintainers mailing list
> Pkg-libvirt-maintainers at lists.alioth.debian.org
> http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-libvirt-maintainers
> 





More information about the Pkg-libvirt-maintainers mailing list