Re: [Xen-devel] CPU enumeration in Xen

["Daniel P. Berrange" <berrange@xxxxxxxxxx>]

On Wed, Aug 22, 2007 at 01:34:47PM +0100, Keir Fraser wrote:
> It seems a bit unlikely, since the enumeration code is from Linux. Perhaps
> the enumeration order has changed in Linux?
>
> To my knowledge the enumeration order is determined by the order we find
> CPUs in the BIOS tables, and can hence vary across different systems, and
> even across BIOS upgrades.

The HV may enumerate CPUs in the same way as Linux baremetal, but Dom0 sees
VCPUs not physical CPUs. So /proc/cpuinfo in Dom0 will bear little resemblance
to the same file on a bare metal machine. You'd need to compare the HV's view
of physical CPUs to the Linux bare metal view, which isn't something you can
do  by merely comparing /proc/cpuinfo AFAICT.

> On 22/8/07 11:34, "Dominik Klein" <dk@xxxxxxxxxxxxxxxx> wrote:
> 
> > http://lists.xensource.com/archives/html/xen-users/2007-08/msg00507.html
> > 
> > In this thread I posted some interesting experiment results which lead
> > to the assumption that Xenified Linux does not enumerate CPUs the same
> > way as a non-xen-Linux does.
> > 
> > Can someone confirm this?

Also, the Dom0 sees no (NUMA) topology - all CPUs are considered equal, so 
Dom0 doesn't see any of the hyperthread, cores, sockets, nodes relationship
that a baremetal kernel sees - the HV of course knows about this topology
but doesn't expose it to the guest kernels when assigning them vCPUs.

Dan.
-- 
|=- Red Hat, Engineering, Emerging Technologies, Boston.  +1 978 392 2496 -=|
|=-           Perl modules: http://search.cpan.org/~danberr/              -=|
|=-               Projects: http://freshmeat.net/~danielpb/               -=|
|=-  GnuPG: 7D3B9505   F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505  -=| 

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