Re: [for-4.8 PATCH] xen: credit2: make runqueues be per-socket by default.

[Wei Liu <wei.liu2@citrix.com>]

On Tue, Nov 29, 2016 at 12:33:24PM +0100, Dario Faggioli wrote:
> On Tue, 2016-11-29 at 11:14 +0000, Wei Liu wrote:
> > On Tue, Nov 29, 2016 at 11:57:03AM +0100, Dario Faggioli wrote:
> > Strictly speaking, this is not necessary for 4.8 because even though
> > the
> > default configuration is not optimal, there is a simple way to change
> > it.
> > 
> Absolutely true.
> 
> > On the other hand, because credit2 is declared supported in this
> > release, it is better to have a sensible default setting for credit2.
> > 
> Exactly.
> 
> > The risk is that changing the default value would cause osstest to
> > discover bugs in the scheduler logic, hence blocking the release. But
> > overall I think the chance is low, but Dario can you please state to
> > what extend did you test this patch?
> > 
> I've tried to state that in the cover letter:
> https://lists.xen.org/archives/html/xen-devel/2016-11/msg02616.html
> 

Ah, I thought that was a dupe because it had the same subject line as
the patch, so I skipped it.

> In some more details, it is _months_ that I and Anshul run benchmarks
> with Credit2 running in this configuration.
> 
> It's months that, all the time that I boot my test box, either to do
> Credit2 related development and testing, or completely unrelated things
> (e.g., recently, when testing the new Fedora 25 as HVM and PV guest), I
> use Credit2 in this configuration.
> 
> George has done some runs of his schedbench suite with this
> configuration (to compare Credit1 vs Credit2 per-core vs Credit2 per-
> socket).
> 
> So, if you ask me "how much is this tested", my answer would be "really
> really really a lot". :-)
> 

OK. This is rather convincing to me. I will wait a bit for other people
express their opinion.

Wei.

> 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)



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