Re: [PATCH 06/11] cgroups: Add res counter common ancestor searching

[Frederic Weisbecker <fweisbec@gmail.com>]

On Sun, Sep 18, 2011 at 10:59:32PM +0300, Kirill A. Shutemov wrote:
> On Tue, Sep 13, 2011 at 01:11:26AM +0200, Frederic Weisbecker wrote:
> > Add a new API to find the common ancestor between two resource
> > counters. This includes the passed resource counter themselves.
> > 
> > Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
> > Acked-by: Paul Menage <paul@paulmenage.org>
> > Cc: Li Zefan <lizf@cn.fujitsu.com>
> > Cc: Johannes Weiner <hannes@cmpxchg.org>
> > Cc: Aditya Kali <adityakali@google.com>
> > Cc: Oleg Nesterov <oleg@redhat.com>
> > Cc: Andrew Morton <akpm@linux-foundation.org>
> > Cc: Kay Sievers <kay.sievers@vrfy.org>
> > Cc: Tim Hockin <thockin@hockin.org>
> > Cc: Tejun Heo <tj@kernel.org>
> > ---
> >  include/linux/res_counter.h |    3 +++
> >  kernel/res_counter.c        |   22 ++++++++++++++++++++++
> >  2 files changed, 25 insertions(+), 0 deletions(-)
> > 
> > diff --git a/include/linux/res_counter.h b/include/linux/res_counter.h
> > index de4ba29..558f39b 100644
> > --- a/include/linux/res_counter.h
> > +++ b/include/linux/res_counter.h
> > @@ -147,6 +147,9 @@ static inline void res_counter_uncharge(struct res_counter *counter,
> >  	res_counter_uncharge_until(counter, NULL, val);
> >  }
> >  
> > +struct res_counter *res_counter_common_ancestor(struct res_counter *l,
> > +						struct res_counter *r);
> > +
> >  /**
> >   * res_counter_margin - calculate chargeable space of a counter
> >   * @cnt: the counter
> > diff --git a/kernel/res_counter.c b/kernel/res_counter.c
> > index 0cc9c7e..22f07cc 100644
> > --- a/kernel/res_counter.c
> > +++ b/kernel/res_counter.c
> > @@ -91,6 +91,28 @@ void res_counter_uncharge_until(struct res_counter *counter,
> >  	local_irq_restore(flags);
> >  }
> >  
> > +/*
> > + * Walk through r1 and r2 parents and try to find the closest common one
> > + * between both. If none is found, it returns NULL.
> > + */
> > +struct res_counter *
> > +res_counter_common_ancestor(struct res_counter *r1, struct res_counter *r2)
> > +{
> > +	struct res_counter *iter;
> > +
> > +	while (r1) {
> > +		iter = r2;
> > +		while (iter) {
> > +			if (iter == r1)
> > +				return iter;
> > +			iter = iter->parent;
> > +		}
> > +
> > +		r1 = r1->parent;
> > +	}
> > +
> > +	return NULL;
> > +}
> 
> Your implementation is O(n*m).

Paul Menage said that shouldn't harm because in practice
cgroups hierarchies are not very deep.

But I can take your implementation below, it won't harm
further and might behave better on some corner cases.

> 
> What about:
> 
> struct res_counter *
> res_counter_common_ancestor(struct res_counter *r1, struct res_counter *r2)
> {
> 	struct res_counter *iter;
> 	int r1_depth = 0, r2_depth = 0;
> 
> 	for (iter = r1; iter; iter = iter->parent)
> 		r1_depth++;
> 	for (iter = r2; iter; iter = iter->parent)
> 		r2_depth++;
> 
> 	while (r1_depth > r2_depth) {
> 		r1 = r1->parent;
> 		r1_depth--;
> 	}
> 	while (r2_depth > r1_depth) {
> 		r2 = r2->parent;
> 		r2_depth--;
> 	}
> 
> 	while (r1 != r2) {
> 		r1 = r1->parent;
> 		r2 = r2->parent;
> 	}
> 
> 	return r1;
> }
> 
> Untested. It's O(n+m).
> 
> -- 
>  Kirill A. Shutemov