[Devel] Re: [PATCH] fix bad behavior in use_hierarchy file

Tejun Heo tj at kernel.org
Tue Jun 26 10:55:13 PDT 2012


Hello,

On Tue, Jun 26, 2012 at 09:56:53AM +0200, Michal Hocko wrote:
> [Adding Ying to CC - they are using hierarchies AFAIU in their workloads]

Ooh, I'm they. :) Asking around.... okay, so google does use
.use_hierarchy but it's a tree-wide thing and would be perfectly happy
with a global switch.

> On Mon 25-06-12 13:49:08, Tejun Heo wrote:
> [...]
> > A bit of delta but is there any chance we can either deprecate
> > .use_hierarhcy or at least make it global toggle instead of subtree
> > thing?  
> 
> So what you are proposing is to have all subtrees of the root either
> hierarchical or not, right?

Yeap.  Just make it a global switch.  Probably determined on mount
time.

> > This seems needlessly complicated. :(
> 
> Toggle wouldn't help much I am afraid. We would still have to
> distinguish (non)hierarchical cases. And I am not sure we can make
> everything hierarchical easily. 

I'm kinda confused by this paragraph.  What do you mean by "wouldn't
help much"?  Do you mean in terms of complexity?

> Most users (from my experience) ignored use_hierarchy for some reasons
> and the end results might be really unexpected for them if they used
> deeper subtrees (which might be needed due to combination with other
> controller(s)).

Oh yeah, we can't change the default behavior like that.  The
transition should be a lot more gradual.  Even if making
.use_hierarchy doesn't help much in terms of reducing complexity right
now, it would at least allow us to weed out and prevent wacky woo-hoo
mom-look-at-what-I-can-do configurations which will be a lot more
difficult to deal with for both us and such users (if we end up
forcing hierarchy).

Thanks.

-- 
tejun




More information about the Devel mailing list