Re: [PATCH v2] sched, timer: Use atomics for thread_group_cputimer to improve scalability

[Jason Low <jason.low2@hp.com>]

On Thu, 2015-03-05 at 16:20 +0100, Frederic Weisbecker wrote:
> On Mon, Mar 02, 2015 at 01:44:04PM -0800, Linus Torvalds wrote:
> > On Mon, Mar 2, 2015 at 1:16 PM, Jason Low <jason.low2@hp.com> wrote:
> > >
> > > In original code, we set cputimer->running first so it is running while
> > > we call update_gt_cputime(). Now in this patch, we swapped the 2 calls
> > > such that we set running after calling update_gt_cputime(), so that
> > > wouldn't be an issue anymore.
> > 
> > Hmm. If you actually care about ordering, and 'running' should be
> > written to after the other things, then it might be best if you use
> > 
> >    smp_store_release(&cputimer->running, 1);
> > 
> > which makes it clear that the store happens *after* what went before it.
> > 
> > Or at least have a "smp_wmb()" between the atomic64 updates and the
> > "WRITE_ONCE()".
> 
> FWIW, perhaps it can be reduced with an smp_mb__before_atomic() on the
> account_group_*_time() side,

Hi Frederic,

I think Linus might be referring to the updates in update_gt_cputime()?
Otherwise, if the atomic updates in account_group_*_time() is already
enough for correctness, then we might not want to be adding barriers in
the hot paths if they aren't necessary.

I was thinking about the adding smp_store_release(&cputimer->running, 1)
to document that we want to write to the running field after the
operations in update_gt_cputime(). The overhead here won't be much since
it doesn't get called frequently as you mentioned.

>  paired with smp_wmb() from the thread_group_cputimer()
> side. Arming cputime->running shouldn't be too frequent while update cputime
> happens at least every tick...
> 
> Assuming smp_mb__before_atomic() is more lightweight than smp_load_acquire()
> of course. 
> 
> > 
> > I guess that since you use cmpxchg in update_gt_cputime, the accesses
> > end up being ordered anyway, but it might be better to make that thing
> > very explicit.
> > 
> >                    Linus