Re: [PATCH v9 3/7] cpuset: Add cpuset.sched.load_balance flag to v2

[Peter Zijlstra]

On Thu, May 31, 2018 at 09:54:27AM -0400, Waiman Long wrote:
> On 05/31/2018 08:26 AM, Peter Zijlstra wrote:

> > I still find all that a bit weird.
> >
> > So load_balance=0 basically changes a partition into a
> > 'fully-partitioned partition' with the seemingly random side-effect that
> > now sub-partitions are allowed to consume all CPUs.
> 
> Are you suggesting that we should allow sub-partition to consume all the
> CPUs no matter the load balance state? I can live with that if you think
> it is more logical.

I'm on the fence myself; the only thing I'm fairly sure of is that tying
this particular behaviour to the load-balance knob seems off.

> > The rationale, only given in the Changelog above, seems to be to allow
> > 'easy' emulation of isolcpus.
> >
> > I'm still not convinced this is a useful knob to have. You can do
> > fully-partitioned by simply creating a lot of 1 cpu parititions.
> 
> That is certainly true. However, I think there are some additional
> overhead in the scheduler side in maintaining those 1-cpu partitions. Right?

cpuset-controller as such doesn't have much overhead scheduler wise,
cpu-controller OTOH does, and there depth is the predominant factor, so
many sibling groups should not matter there either.

> > So this one knob does two separate things, both of which seem, to me,
> > redundant.
> >
> > Can we please get better rationale for this?
> 
> I am fine getting rid of the load_balance flag if this is the consensus.
> However, we do need to come up with a good migration story for those
> users that need the isolcpus capability. I think Mike was the one asking
> for supporting isolcpus. So Mike, what is your take on that.

So I don't strictly mind having a knob that does the 'fully-partitioned
partition' thing -- however odd that sounds -- but I feel we should have
a solid use-case for it.

I also think we should not mix the 'consume all' thing with the
'fully-partitioned' thing, as they are otherwise unrelated.