Re: [PATCH v2] memcg: first step towards hierarchical controller

[Michal Hocko <mhocko-AlSwsSmVLrQ@public.gmane.org>]

On Thu 06-09-12 16:09:20, Glauber Costa wrote:
> On 09/06/2012 04:06 PM, Michal Hocko wrote:
> > On Wed 05-09-12 13:12:38, Tejun Heo wrote:
> >> Hello, Michal.
> >>
> >> On Wed, Sep 05, 2012 at 04:49:42PM +0200, Michal Hocko wrote:
> >>> Can we settle on the following 3 steps?
> >>> 1) warn about "flat" hierarchies (give it X releases) - I will push it
> >>>    to as many Suse code streams as possible (hope other distributions
> >>>    could do the same)
> >>
> >> I think I'm just gonna trigger WARN from cgroup core if anyone tries
> >> to create hierarchy with a controller which doesn't support full
> >> hierarchy.  WARN_ON_ONCE() at first and then WARN_ON() on each
> >> creation later on.
> > 
> > How do you find out that a controller is not fully hierarchical? Memory
> > controller can be both.
> > 
> >>> 2) flip the default on the root cgroup & warn when somebody tries to
> >>>    change it to 0 (give it another X releases) that the knob will be
> >>>    removed
> >>> 3) remove the knob and the whole nonsese
> >>> 4) revert 3 if somebody really objects
> >>
> >> If we can get to 3, I don't think 4 would be a problem.
> > 
> > Agreed.
> > 
> Just so I understand it:
> 
> Michal clearly objected before folding his patch with my Kconfig patch.
> But is there still opposition to merge both?

I do not find the config option very much useful but if others feel it
really is I won't block it.

> By having it default-n, only people that are either sure that this is
> safe for them, or have more clearly defined lifecycles could set it.
-- 
Michal Hocko
SUSE Labs