sysfs on hvm guests?

sysfs on hvm guests?

[Dan Magenheimer]

[-- Attachment #1: Type: text/plain, Size: 972 bytes --]

Is sysfs expected to work fully in an hvm guest?

In RHEL5 (2.6.18-based) kernels, the following works native

/sys/devices/system/clocksource/clocksource0/...
  available_clocksource lists the possible clock sources
  current_clocksource lists the chosen clock source,
    and can be changed on the fly

but don't work for me from within an RHEL5 HVM guest.

All reads yield only "jiffies" (which I believe is the
default setting) even though there are other clock
sources available and set, and echo'ing to the
current_clocksource doesn't seem to change it.

Should I expect that it would work?  If so, does it work
on your system?  Perhaps it is something peculiar to
my environment?

Thanks,
Dan

===================================
If Xen could save time in a bottle / then clocks wouldn't virtually skew /
It would save every tick / for VMs that aren't quick /
and Xen then would send them anew
(with apologies to the late great Jim Croce)

[-- Attachment #2: Type: text/plain, Size: 138 bytes --]

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

RE: sysfs on hvm guests?

[Dan Magenheimer]

[-- Attachment #1: Type: text/plain, Size: 1583 bytes --]

Hmmm... this seems to work fine when booting a 32-bit RHEL5 on
32-bit dom0 on 32-bit xen and with 32-on-32-on-64, but fails with
64-bit RHEL5 on 32-bit dom0 on 64-bit xen.  Not sure yet if
this is a problem with the specific guest instance or what.

> -----Original Message-----
> From: xen-devel-bounces@lists.xensource.com
> [mailto:xen-devel-bounces@lists.xensource.com]On Behalf Of Dan
> Magenheimer
> Sent: Tuesday, February 19, 2008 9:48 PM
> To: xen-devel@lists.xensource.com
> Subject: [Xen-devel] sysfs on hvm guests?
> 
> 
> Is sysfs expected to work fully in an hvm guest?
> 
> In RHEL5 (2.6.18-based) kernels, the following works native
> 
> /sys/devices/system/clocksource/clocksource0/...
>   available_clocksource lists the possible clock sources
>   current_clocksource lists the chosen clock source,
>     and can be changed on the fly
> 
> but don't work for me from within an RHEL5 HVM guest.
> 
> All reads yield only "jiffies" (which I believe is the
> default setting) even though there are other clock
> sources available and set, and echo'ing to the
> current_clocksource doesn't seem to change it.
> 
> Should I expect that it would work?  If so, does it work
> on your system?  Perhaps it is something peculiar to
> my environment?
> 
> Thanks,
> Dan
> 
> ===================================
> If Xen could save time in a bottle / then clocks wouldn't 
> virtually skew /
> It would save every tick / for VMs that aren't quick /
> and Xen then would send them anew
> (with apologies to the late great Jim Croce)
> 

[-- Attachment #2: Type: text/plain, Size: 138 bytes --]

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

RE: sysfs on hvm guests?

[Dan Magenheimer]

[-- Attachment #1: Type: text/plain, Size: 2234 bytes --]

Double hmmm... it seems that even on hardware 64-bit RHEL5
reports only jiffies for these sysfs clocksource variables
even though other clocksources are available and current.

Not sure if this is a bug or a feature, but its definitely
not a Xen problem.

Sorry for the false alarm.

Dan

> -----Original Message-----
> From: Dan Magenheimer [mailto:dan.magenheimer@oracle.com]
> Sent: Wednesday, February 20, 2008 8:06 AM
> To: dan.magenheimer@oracle.com; xen-devel@lists.xensource.com
> Subject: RE: [Xen-devel] sysfs on hvm guests?
> 
> 
> Hmmm... this seems to work fine when booting a 32-bit RHEL5 on
> 32-bit dom0 on 32-bit xen and with 32-on-32-on-64, but fails with
> 64-bit RHEL5 on 32-bit dom0 on 64-bit xen.  Not sure yet if
> this is a problem with the specific guest instance or what.
> 
> > -----Original Message-----
> > From: xen-devel-bounces@lists.xensource.com
> > [mailto:xen-devel-bounces@lists.xensource.com]On Behalf Of Dan
> > Magenheimer
> > Sent: Tuesday, February 19, 2008 9:48 PM
> > To: xen-devel@lists.xensource.com
> > Subject: [Xen-devel] sysfs on hvm guests?
> > 
> > 
> > Is sysfs expected to work fully in an hvm guest?
> > 
> > In RHEL5 (2.6.18-based) kernels, the following works native
> > 
> > /sys/devices/system/clocksource/clocksource0/...
> >   available_clocksource lists the possible clock sources
> >   current_clocksource lists the chosen clock source,
> >     and can be changed on the fly
> > 
> > but don't work for me from within an RHEL5 HVM guest.
> > 
> > All reads yield only "jiffies" (which I believe is the
> > default setting) even though there are other clock
> > sources available and set, and echo'ing to the
> > current_clocksource doesn't seem to change it.
> > 
> > Should I expect that it would work?  If so, does it work
> > on your system?  Perhaps it is something peculiar to
> > my environment?
> > 
> > Thanks,
> > Dan
> > 
> > ===================================
> > If Xen could save time in a bottle / then clocks wouldn't 
> > virtually skew /
> > It would save every tick / for VMs that aren't quick /
> > and Xen then would send them anew
> > (with apologies to the late great Jim Croce)
> > 
> 

[-- Attachment #2: Type: text/plain, Size: 138 bytes --]

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

RE: sysfs on hvm guests?

[Dan Magenheimer]

[-- Attachment #1: Type: text/plain, Size: 2726 bytes --]

If anybody cares, it appears this is a 64-bit kernel bug that
has been fixed upstream post-2.6.18.

> -----Original Message-----
> From: Dan Magenheimer [mailto:dan.magenheimer@oracle.com]
> Sent: Wednesday, February 20, 2008 8:36 AM
> To: dan.magenheimer@oracle.com; xen-devel@lists.xensource.com
> Subject: RE: [Xen-devel] sysfs on hvm guests?
> 
> 
> Double hmmm... it seems that even on hardware 64-bit RHEL5
> reports only jiffies for these sysfs clocksource variables
> even though other clocksources are available and current.
> 
> Not sure if this is a bug or a feature, but its definitely
> not a Xen problem.
> 
> Sorry for the false alarm.
> 
> Dan
> 
> > -----Original Message-----
> > From: Dan Magenheimer [mailto:dan.magenheimer@oracle.com]
> > Sent: Wednesday, February 20, 2008 8:06 AM
> > To: dan.magenheimer@oracle.com; xen-devel@lists.xensource.com
> > Subject: RE: [Xen-devel] sysfs on hvm guests?
> > 
> > 
> > Hmmm... this seems to work fine when booting a 32-bit RHEL5 on
> > 32-bit dom0 on 32-bit xen and with 32-on-32-on-64, but fails with
> > 64-bit RHEL5 on 32-bit dom0 on 64-bit xen.  Not sure yet if
> > this is a problem with the specific guest instance or what.
> > 
> > > -----Original Message-----
> > > From: xen-devel-bounces@lists.xensource.com
> > > [mailto:xen-devel-bounces@lists.xensource.com]On Behalf Of Dan
> > > Magenheimer
> > > Sent: Tuesday, February 19, 2008 9:48 PM
> > > To: xen-devel@lists.xensource.com
> > > Subject: [Xen-devel] sysfs on hvm guests?
> > > 
> > > 
> > > Is sysfs expected to work fully in an hvm guest?
> > > 
> > > In RHEL5 (2.6.18-based) kernels, the following works native
> > > 
> > > /sys/devices/system/clocksource/clocksource0/...
> > >   available_clocksource lists the possible clock sources
> > >   current_clocksource lists the chosen clock source,
> > >     and can be changed on the fly
> > > 
> > > but don't work for me from within an RHEL5 HVM guest.
> > > 
> > > All reads yield only "jiffies" (which I believe is the
> > > default setting) even though there are other clock
> > > sources available and set, and echo'ing to the
> > > current_clocksource doesn't seem to change it.
> > > 
> > > Should I expect that it would work?  If so, does it work
> > > on your system?  Perhaps it is something peculiar to
> > > my environment?
> > > 
> > > Thanks,
> > > Dan
> > > 
> > > ===================================
> > > If Xen could save time in a bottle / then clocks wouldn't 
> > > virtually skew /
> > > It would save every tick / for VMs that aren't quick /
> > > and Xen then would send them anew
> > > (with apologies to the late great Jim Croce)
> > > 
> > 
> 

[-- Attachment #2: Type: text/plain, Size: 138 bytes --]

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