Bug#674907: shifts the clock by 50 minutes unexplicably under a Xen dom0

[Mauro]

On 1 October 2012 20:24, Jonathan Nieder  wrote:
> Mauro wrote:
>
>> The shift time happens also in dom0 not only in domUs.
>
> Thanks.
>
> Just gathering information: some references:
>
>  http://thread.gmane.org/gmane.comp.emulators.xen.user/75486
>  http://thread.gmane.org/gmane.comp.emulators.xen.user/65530
>
> What's the newest kernel and hypervisor you've tried (since
> this should make it easier to get help from upstream)?

uname -a
Linux xen-p01 2.6.32-5-xen-amd64 #1 SMP Sun Sep 23 13:49:30 UTC 2012
x86_64 GNU/Linux

xm dmesg | grep version
(XEN) Xen version 4.0.1 (Debian 4.0.1-5.4) (ultrot...@debian.org) (gcc
version 4.4.5 (Debian 4.4.5-8) ) Sat Sep  8 19:15:46 UTC 2012


-- 
To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: 
http://lists.debian.org/CAE17a0VAPvw62tXDJsg+VfeduPxiMRbgto1LhnAvHCN=5bw...@mail.gmail.com

Bug#674907: shifts the clock by 50 minutes unexplicably under a Xen dom0

[Antoine Beaupré]

On 2012-10-01, Jonathan Nieder wrote:
> Mauro wrote:
>
>> The shift time happens also in dom0 not only in domUs.
>
> Thanks.
>
> Just gathering information: some references:
>
>  http://thread.gmane.org/gmane.comp.emulators.xen.user/75486
>  http://thread.gmane.org/gmane.comp.emulators.xen.user/65530

Really interesting discussions, here's what I can extract from it.

Other people have the same problem, and have applied various
workarounds:

 * clocksource=jiffies - fails for us 

 * clocksource=tsc 

 * clocksource=pit, cpuidle=0 


We'll be trying those 2 last workarounds and report back here.

> What's the newest kernel and hypervisor you've tried (since
> this should make it easier to get help from upstream)?

We haven't tried upgrading to backports - we are running plain squeeze.

-- 
The world is a dangerous place, not because of those who do evil,
but because of those who look on and do nothing.
- Albert Einstein


pgpvIPpg2fZTq.pgp
Description: PGP signature

Bug#674907: shifts the clock by 50 minutes unexplicably under a Xen dom0

[Jonathan Nieder]

Mauro wrote:

> The shift time happens also in dom0 not only in domUs.

Thanks.

Just gathering information: some references:

 http://thread.gmane.org/gmane.comp.emulators.xen.user/75486
 http://thread.gmane.org/gmane.comp.emulators.xen.user/65530

What's the newest kernel and hypervisor you've tried (since
this should make it easier to get help from upstream)?


-- 
To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20121001182449.GC16391@elie.Belkin

Bug#674907: shifts the clock by 50 minutes unexplicably under a Xen dom0

[Mauro]

On 1 October 2012 20:01, Jonathan Nieder  wrote:

> This particular bug is known to be a domU problem, so I suspect
> Bastian made the right call.  If he has overlooked something, we'd
> welcome the missing information.

The problem happens also in dom0.


-- 
To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: 
http://lists.debian.org/CAE17a0Xd=uxc4zno4-eepfsgsudii4r3iqfu7qcvwo5ujwx...@mail.gmail.com

Bug#674907: shifts the clock by 50 minutes unexplicably under a Xen dom0

[Jonathan Nieder]

Hi Antoine,

anarcat wrote:

> Why was this reassigned to the linux kernel?
>
> I explicitely stated (twice) that machines not running Xen are *not*
> affected, so this is specifically a Xen issue.

Xen in squeeze consists of both a hypervisor and tools and a kernel
patch, and potential changes to the kernel patch are tracked in the
BTS by assigning to the kernel package.  I doubt Bastian meant to
imply that your problem affects non-Xen systems.

> I do not want to get into an edit war here, if there is proper
> justification I don't mind having it assigned to anything...

This particular bug is known to be a domU problem, so I suspect
Bastian made the right call.  If he has overlooked something, we'd
welcome the missing information.

Hoping that clarifies,
Jonathan


-- 
To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20121001180102.GB16391@elie.Belkin