[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Xen-devel] Powertop shows xenconsoled the reason why CPU wakes up from C-sleep states


  • To: <xen-devel@xxxxxxxxxxxxxxxxxxx>
  • From: "Roger Cruz" <roger.cruz@xxxxxxxxxxxxxxxxxxx>
  • Date: Wed, 2 Mar 2011 12:08:46 -0600
  • Delivery-date: Wed, 02 Mar 2011 10:09:55 -0800
  • List-id: Xen developer discussion <xen-devel.lists.xensource.com>
  • Thread-index: AcvZBN+c1TF2ipGtQuS+yZldaUV1Hg==
  • Thread-topic: Powertop shows xenconsoled the reason why CPU wakes up from C-sleep states


I've been examining battery drain issues and when we run a PV linux kernel guest on Xen 3.4 with a dom0 2.6.32 kernel, I can see that xenconsoled is the one marked as responsible for a big percentage of the wake ups (using Powertop).  Using strace, I can see that xenconsoled is pending in select without a timeout.  that is, the application never really runs when the cpu wakes up.  Using Xentrace, I can see that VCPUs are awaked, a few hypercalls from the kernel take place and then it puts itself to sleep.  Then in about 30 pm ticks, the other CPU wakes up to run vcpu #2, which executes pretty much the same hypercalls and puts itself to sleep.  This goes back and forth causing each CPU to wake up momentarily and go to sleep again.

I did some research on the web and this thread appears to be relevant.  Even though I don't see events/0, I see xenconsoled which was mentioned this thread.

http://www.gossamer-threads.com/lists/linux/kernel/1281515

Has anyone else seen this type of ping-pong behavior?  Any idea how to fix this?

thanks
Roger

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.