WARNING - OLD ARCHIVES

This is an archived copy of the Xen.org mailing list, which we have preserved to ensure that existing links to archives are not broken. The live archive, which contains the latest emails, can be found at http://lists.xen.org/
   
 
 
Xen 
 
Home Products Support Community News
 
   
 

xen-users

AW: Re: AW: [Xen-users] cpu frequency scaling in xen 3.3.1

To: flo <flo@xxxxxxxxx>, xen-users <xen-users@xxxxxxxxxxxxxxxxxxx>
Subject: AW: Re: AW: [Xen-users] cpu frequency scaling in xen 3.3.1
From: "Carsten Schiers" <carsten@xxxxxxxxxx>
Date: Sun, 29 Mar 2009 22:45:25 +0200
Cc:
Delivery-date: Sun, 29 Mar 2009 13:46:47 -0700
Envelope-to: www-data@xxxxxxxxxxxxxxxxxxx
In-reply-to: <49CF9F83.4000201@xxxxxxxxx>
List-help: <mailto:xen-users-request@lists.xensource.com?subject=help>
List-id: Xen user discussion <xen-users.lists.xensource.com>
List-post: <mailto:xen-users@lists.xensource.com>
List-subscribe: <http://lists.xensource.com/mailman/listinfo/xen-users>, <mailto:xen-users-request@lists.xensource.com?subject=subscribe>
List-unsubscribe: <http://lists.xensource.com/mailman/listinfo/xen-users>, <mailto:xen-users-request@lists.xensource.com?subject=unsubscribe>
Sender: xen-users-bounces@xxxxxxxxxxxxxxxxxxx
Latest 3.3 Xen docs will explain the parameters, but don't give the 
background information. It's
barried deep in the Xen-Devel list. I don't think <Fam10h families will 
be ever supported diretctly
in the hypervisor, but there is no necessity to do so. The modified 
ondemand governor uses a hypercall
to get idle times and follow the policy. 

Building cpufreq management and the necessary drivers (e.g. powernow-k8, 
cpufreq_ondemand,...) is a
bit tricky. Follow the link. Eventualy you also need to patcht he 
Kconfig file a bit...

BR,
Carsten.

-----Ursprüngliche Nachricht-----
Von: Florian Gleixner [mailto:flo@xxxxxxxxx] 
Gesendet: Sonntag, 29. März 2009 18:19
An: Carsten Schiers
Cc: xen-users
Betreff: Re: AW: [Xen-users] cpu frequency scaling in xen 3.3.1

Thanks!

Is there a document that i have overseen? Or is the documentation 
missing?
Cpu frequency scaling inside the hypervisor would be great. Where can i 
vote for that my CPU model will be included?
I'll try to get the powernow-k8.ko build now. Opensuse 11.1 has it for 
the non-xen kernel but did not include it in the xen enabled kernel.

Florian

Carsten Schiers wrote:
> Florian,
> 
> handling of P-States for AMD CPUs is only possible in xen hypervisor 
> from CPU Fam 10h on.
> The 5050e is family 0Fh.
> 
> The alternative is using frequency handling in dom0, by specifying 
> cpufreq=dom0-kernel in the hypervisor start line. You need 
> powernow-k8.ko then in dom0 and can use e.g. the dom0 ondemand 
> governor (by calling cpufreq-set -g ondemand).
> 
> Assuming you are German, please see
> http://www.ctserver.org/ftopic2437.html for some details.
> It is a bit tricky and still not working 100% without log warnings for 

> me and my 4050e.
> 
> BR,
> Carsten.
> 
> -----Ursprüngliche Nachricht-----
> Von: Florian Gleixner [mailto:flo@xxxxxxxxx]
> Gesendet: Sonntag, 29. März 2009 12:03
> An: xen-users@xxxxxxxxxxxxxxxxxxx
> Betreff: [Xen-users] cpu frequency scaling in xen 3.3.1
> 
> Hi,
> 
> i read the xenpm wiki page, but i cannot use the "p-states". xenpm
> shows:
> 
> xenpm
> cpu id               : 0
> total C-states       : 2
> idle time(ms)        : 1776034
> C0                   : transition [00000000000000000000]
>                        residency  [00000000000000000000 ms]
> C1                   : transition [00000000000000260278]
>                        residency  [00000000000000000000 ms]
> 
> cpu id               : 1
> total C-states       : 0
> idle time(ms)        : 1771123
> 
> [CPU0] failed to get P-states statistics information [CPU1] failed to 
> get P-states statistics information
> 
> /proc/cpuinfo shows:
> model name      : AMD Athlon(tm) Dual Core Processor 5050e
> 
> without xen i can scale the frequency perfectly. Are p-states for this 

> CPU model unsupported? My grub line is:
>     kernel /xen.gz cpufreq=xen cpuidle ...
> 
> Thanks for the help!
> 
> Florian
> 
> 
> 
> 




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

<Prev in Thread] Current Thread [Next in Thread>