Re: [PATCH v2 2/5] sched/numa: Replace runnable_load_avg by load_avg

From: Mel Gorman
Date: Tue Feb 18 2020 - 12:41:26 EST

On Tue, Feb 18, 2020 at 04:50:48PM +0000, Valentin Schneider wrote:
> On 18/02/2020 15:38, Mel Gorman wrote:
> >>
> >> Could we reuse group_type instead? The definitions are the same modulo
> >> s/group/node/.
> >>
> >
> > I kept the naming because there is the remote possibility that NUMA
> > balancing will deviate in some fashion. Right now, it's harmless.
> >
> Since it's just a subset ATM I'd go for the reuse and change that later if
> shown a split is required, but fair enough.

I would feel that I was churning code for the sake of it.

> > I didn't merge that part of the first version of my series. I was
> > waiting to see how the implementation for allowing a small degree of
> > imbalance looks like. If it's entirely confined in adjust_numa_balance
> ^^^^^^^^^^^^^^^^^^^
> Apologies if that's a newbie question, but I'm not familiar with that one.
> Would that be added in your reconciliation series? I've only had a brief
> look at it yet (it's next on the chopping block).

I should have wrote adjust_numa_imbalance but yes, it's part of the
reconciled series so that NUMA balancing and the load balancer use the
same helper.

> > Yikes, no I'd rather not do that. Basically all I did before was create
> > a common helper like __lb_has_capacity that only took basic types as
> > parameters. group_has_capacity and numa_has_capacity were simple wrappers
> > that read the correct fields from their respective stats structures.
> >
> That's more sensible indeed. It'd definitely be nice to actually reconcile
> the two balancers with these common helpers, though I guess it doesn't
> *have* to happen with this very patch.

Yep, it can happen at a later time.

As it stands, I think the reconciled series stands on its own even
though there are further improvements that could be built on top.

Mel Gorman