Re: [PATCH] rcu: Only pin GP kthread when full dynticks is actually used

["Paul E. McKenney" <paulmck@linux.vnet.ibm.com>]

On Thu, Jun 12, 2014 at 06:24:32PM -0700, Paul E. McKenney wrote:
> On Fri, Jun 13, 2014 at 02:16:59AM +0200, Frederic Weisbecker wrote:
> > CONFIG_NO_HZ_FULL may be enabled widely on distros nowadays but actual
> > users should be a tiny minority, if actually any.
> > 
> > Also there is a risk that affining the GP kthread to a single CPU could
> > end up noticeably reducing RCU performances and increasing energy
> > consumption.
> > 
> > So lets affine the GP kthread only when nohz full is actually used
> > (ie: when the nohz_full= parameter is filled or CONFIG_NO_HZ_FULL_ALL=y)
> > 
> > Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
> > Cc: Josh Triplett <josh@joshtriplett.org>
> > Cc: Steven Rostedt <rostedt@goodmis.org>
> > Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
> > ---
> >  kernel/rcu/tree_plugin.h | 10 +++++++---
> >  1 file changed, 7 insertions(+), 3 deletions(-)
> > 
> > diff --git a/kernel/rcu/tree_plugin.h b/kernel/rcu/tree_plugin.h
> > index cbc2c45..726f52c 100644
> > --- a/kernel/rcu/tree_plugin.h
> > +++ b/kernel/rcu/tree_plugin.h
> > @@ -2843,12 +2843,16 @@ static bool rcu_nohz_full_cpu(struct rcu_state *rsp)
> >   */
> >  static void rcu_bind_gp_kthread(void)
> >  {
> > -#ifdef CONFIG_NO_HZ_FULL
> > -	int cpu = ACCESS_ONCE(tick_do_timer_cpu);
> > +	int cpu;
> > +
> > +	if (!tick_nohz_full_enabled())
> > +		return;
> > +
> > +	cpu = ACCESS_ONCE(tick_do_timer_cpu);
> > 
> >  	if (cpu < 0 || cpu >= nr_cpu_ids)
> >  		return;
> > +
> >  	if (raw_smp_processor_id() != cpu)
> >  		set_cpus_allowed_ptr(current, cpumask_of(cpu));
> > -#endif /* #ifdef CONFIG_NO_HZ_FULL */
> >  }
> 
> Hello, Frederic,
> 
> I have the following queued.  Shall I port yours on top of mine, or is
> there an issue with mine?

OK, I suppose I should show you what it looks like ported on top of mine.

I didn't understand the removal of "#ifdef CONFIG_NO_HZ_FULL", so I
left that.  I also didn't understand how dumping the GP kthreads onto
a single CPU could affect energy efficiency all that much, so I omitted
that from the commit log.  If I am missing something, please enlighten me.

							Thanx, Paul

------------------------------------------------------------------------

rcu: Only pin GP kthread when full dynticks is actually used

CONFIG_NO_HZ_FULL is widely enabled in distros, but only a small fraction
of the users will be taking advantage of it.  Furthermore, there is
a risk that affining the GP kthread to a single CPU could noticeably
increase RCU grace-period latency.

Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>

diff --git a/kernel/rcu/tree_plugin.h b/kernel/rcu/tree_plugin.h
index c93c525b71fe..2f5679791594 100644
--- a/kernel/rcu/tree_plugin.h
+++ b/kernel/rcu/tree_plugin.h
@@ -2865,7 +2865,7 @@ static void rcu_bind_gp_kthread(void)
 #ifdef CONFIG_NO_HZ_FULL
 	int cpu = tick_do_timer_cpu;
 
-	if (cpu < 0 || cpu >= nr_cpu_ids)
+	if (cpu < 0 || cpu >= nr_cpu_ids || !tick_nohz_full_enabled())
 		return;
 #ifdef CONFIG_NO_HZ_FULL_SYSIDLE
 	if (raw_smp_processor_id() != cpu)