Re: [PATCH] x86: Consider multiple nodes in a single socket to be "sane"

[Dave Hansen <dave@sr71.net>]

On 09/17/2014 03:57 PM, Peter Zijlstra wrote:
>> > Another option would be to:
>> > 1. Add a new "PKG" level and actually _build_ it with phys_proc_id
>> > 2. Make sure to tie the sysfs 'core_siblings' file to PKG
>> > 3. Leave the "MC" level as it is now, but define it as being the lowest-
>> >    common-denominator of core grouping.  In other words, the "MC" group
>> >    will stop at a NUMA node or a socket boundary, whichever it sees
>> >    first.
>> > 4. Chop the "COD_NUMA" level off in sched_init_numa()
> No, we should provide an arch override for sched_domain_topology which
> has the right setup for the detected topology.
> 
> See arm,powerpc and s390, which already have his.

I _think_ the problem here is that all of the existing topologies
describe something below the smallest NUMA node.  With the
cluster-on-die stuff, we've got a node boundary smack in the middle of
the existing topology levels.

So, if we're going to do this with a new sched_domain_topology, we've
got to make sure that we eliminate all of the levels below that first
NUMA node.  I think that means just have a topology with the SMT level
alone in it.

Does that work?