Re: [patch V3] percpu_counter: scalability works

[Eric Dumazet <eric.dumazet@gmail.com>]

Le lundi 16 mai 2011 à 14:37 +0800, Shaohua Li a écrit :
> On Mon, 2011-05-16 at 14:11 +0800, Eric Dumazet wrote:
> > Le lundi 16 mai 2011 à 08:58 +0800, Shaohua Li a écrit :
> > 
> > > so if _sum starts and ends here, _sum can still get deviation.
> > 
> > This makes no sense at all. If you have so many cpus 'here' right before
> > you increment fbc->sum_cnt, then no matter how precise and super
> > cautious you are in your _sum() implementation, as soon as you exit from
> > sum(), other cpus already changed the percpu counter global value.
> I don't agree here. The original implementation also just has quite
> small window we have deviation, the window only exists between the two
> lines:
> 		atomic64_add(count, &fbc->count);
> 	        __this_cpu_write(*fbc->counters, 0);
> if you think we should ignore it, we'd better not use any protection
> here.
> 

Not at all. Your version didnt forbid new cpu to come in _add() and
hitting the deviation problem.

There is a small difference, or else I wouldnt had bother.


> as I wrote in the email, the atomic and cacheline issue can be resolved
> with a per_cpu data, I just didn't post the patch. I post it this time,
> please see below. There is no cache line bounce anymore.
> 

I am afraid we make no progress at all here, if you just try to push
your patch and ignore my comments.

percpu_counter is a compromise, dont make it too slow for normal
operations. It works well if most _add() operations only go through
percpu data.

Please just move vm_committed_as to a plain atomic_t, this will solve
your problem.

Thanks