Re: [PATCH 1/7] xen: sched: factor affinity helpers out of sched_credit.c

[Dario Faggioli <dario.faggioli@citrix.com>]

[-- Attachment #1.1: Type: text/plain, Size: 1748 bytes --]

On Fri, 2017-06-23 at 11:02 +0100, Anshul Makkar wrote:
> On 16/06/2017 15:13, Dario Faggioli wrote:
> > diff --git a/xen/common/sched_credit.c b/xen/common/sched_credit.c
> > index efdf6bf..53773df 100644
> > --- a/xen/common/sched_credit.c
> > +++ b/xen/common/sched_credit.c
> > @@ -136,27 +136,6 @@
> 
> *
> > + * Hard affinity balancing is always necessary and must never be
> > skipped.
> > + * But soft affinity need only be considered when it has a
> > functionally
> > + * different effect than other constraints (such as hard affinity,
> > cpus
> > + * online, or cpupools).
> > + *
> > + * Soft affinity only needs to be considered if:
> > + * * The cpus in the cpupool are not a subset of soft affinity
> 
> do you mean cpus in a cpupool are not in a subset of the soft
> affinity 
> of a vcpu which is a runnable state ?
>
Err... not sure. It's probably the same. IAC, what this means is that,
if a vcpu is in cpupool C, and the cpus of cpupool C are 1,2,3 _and_
the soft affinity of the vcpu is 0,1,2,3,4,5 then, it does not make
sense to check soft-affinity.

In fact, the vcpu will only execute on either 1,2 or 3, which are all
part of its soft-affinity anyway.

And it should be exactly that the cpupool is *a* *subset* of the soft-
affinity. In fact, if cpupool is 1,2,3 and soft-affinity is 0,1,3,4,5
then it _does_ make sense to check soft affinity,, to try not to run
the vcpu on 2.

Hope this clarify things.

Regards,
Dario
-- 
<<This happens because I choose it to happen!>> (Raistlin Majere)
-----------------------------------------------------------------
Dario Faggioli, Ph.D, http://about.me/dario.faggioli
Senior Software Engineer, Citrix Systems R&D Ltd., Cambridge (UK)

[-- Attachment #1.2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

[-- Attachment #2: Type: text/plain, Size: 127 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel