Re: [Xen-users] Xen 3.0 Dom0 Smp enable?

[Ryan Harper <ryanh@xxxxxxxxxx>]

* Ian Pratt <m+Ian.Pratt@xxxxxxxxxxxx> [2005-08-08 08:58]:
>  
> > >> I have a 2 way machine.  Soliciting folks opinion on 
> > whether it makes 
> > >> sense to smp enable DOM0 so that IO handling can be balanced over 
> > >> both CPU's...
> > >
> > >Its generally not helpful to have an SMP dom0. In high IO situations 
> > >it's typically best to dedicate a hyperthread or core to dom0.
> 
> > How does one make the dedication of a CPU to dom0?   Seems 
> > like pincpu is 
> > the control mechanism, but if I understand pincpu correctly, 
> > it doesn't 
> > prevent other domains from sharing the chosen CPU.   It pegs 
> > the domain to 
> > the CPU, but not the reverse.   
> 
> If you create a pincpu mask for the over domains/VCPUs that doesn't
> include the CPU in question, they won't use it. I guess a shorthand
> command for this would be useful.

Something like:

xm dedicate DOM CPU

which would generate the appropriate pincpu ops to push all other doms'
vcpus off of CPU.

> 
> > Also, how can one distinguish an ht CPU vs a core CPU?
> 
> All hyperthreads of a core are enumerated before moving on to the next
> core; All cores are enumerated before the next socket. We've proposed
> changing the CPU naming convention to a '.' seperated form to make this
> more explicit.
> 
> I thought Ryan may have even knocked up a patch for this at one point
> (?)
> Ryan, if so, please can you dust off and resend. As I recall, the key
> fix it needed was that it should be xend (not xm) that does the
> translation from the dot'ed form into CPU numbers.

Certainly it will be easier to do so in xend, but I'm more concerned
with the hypervisor being responsible responsible for choosing the vcpu
to cpu mapping.  I'd like to see dom creation support the passing of a
cpumap and have the hypervisor cycle through the physical cpus marked
therein.  This would remove the logic of mapping vcpus to cpus from the
hypervisor and let the dom creation tools build whatever algorithm for
distributing vcpus across cpus as it sees fit.

If you are interested in this it means changing the hypercall interface
and so it should be done for 3.0, however I don't want to push for
significant changes so close to the testing freeze as I want to help
close features down rather than create new ones.


-- 
Ryan Harper
Software Engineer; Linux Technology Center
IBM Corp., Austin, Tx
(512) 838-9253   T/L: 678-9253
ryanh@xxxxxxxxxx

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