Re: [PATCH 1 of 3 v5/leftover] libxl: enable automatic placement of guests on NUMA nodes

[Andre Przywara <andre.przywara@amd.com>]

On 07/20/2012 02:54 PM, Dario Faggioli wrote:
> On Fri, 2012-07-20 at 13:47 +0200, Andre Przywara wrote:
>>> That would be nice. If you happen to have time to put something like
>>> "3*nrdomains_diff+memfree_diff" and see how it goes, I'll be happy to
>>> include at least that change, even in next posting.
>>
>> I did that. Guests are 2 VCPUs/2GB RAM.
>>
> Thanks!
>
>> The results looked much better.
>>
> That's very good to hear. :-)
>
>>
>> [snip]
>>
>> So without any deeper thinking this looks much better than the original
>> version and possibly good enough for Xen 4.2.
>> Maybe the automated testing finds some leftovers.
>>
> So, here's what we can do:
>
>   1) I can resend the series with this change in it (i.e., number of
>     domains weighting 3 times more than free memory);
>
>   2) I can resend the series like in 1) _but_ while also converting the
>     domain counting in vcpu counting (I checked that and it's trivial);
>
>   3) I can resend the series like in 1) and also send a separate patch
>     turning the domain counting in vcpu counting, and we can apply that
>     after it get some testing;
>
> Which one you like better?

I'd prefer 2). To me it makes more sense and I can test it over the 
weekend and can send my acked-by on it.

Regards,
Andre.

-- 
Andre Przywara
AMD-OSRC (Dresden)
Tel: x29712