Re: [RFC][PATCH v4 3/3] sched: Periodically decay max cost of idle balance

2013-09-09 Thread Mike Galbraith
On Mon, 2013-09-09 at 14:07 -0700, Jason Low wrote: > On Mon, 2013-09-09 at 13:49 +0200, Peter Zijlstra wrote: > > On Wed, Sep 04, 2013 at 12:10:01AM -0700, Jason Low wrote: > > > On Fri, 2013-08-30 at 12:18 +0200, Peter Zijlstra wrote: > > > > On Thu, Aug 29, 2013 at 01:05:36PM -0700, Jason Low w

Re: [RFC][PATCH v4 3/3] sched: Periodically decay max cost of idle balance

2013-09-09 Thread Jason Low
On Mon, 2013-09-09 at 13:49 +0200, Peter Zijlstra wrote: > On Wed, Sep 04, 2013 at 12:10:01AM -0700, Jason Low wrote: > > On Fri, 2013-08-30 at 12:18 +0200, Peter Zijlstra wrote: > > > On Thu, Aug 29, 2013 at 01:05:36PM -0700, Jason Low wrote: > > > > diff --git a/kernel/sched/core.c b/kernel/sched

Re: [RFC][PATCH v4 3/3] sched: Periodically decay max cost of idle balance

2013-09-09 Thread Jason Low
On Mon, 2013-09-09 at 13:44 +0200, Peter Zijlstra wrote: > On Tue, Sep 03, 2013 at 11:02:59PM -0700, Jason Low wrote: > > On Fri, 2013-08-30 at 12:29 +0200, Peter Zijlstra wrote: > > > rcu_read_lock(); > > > for_each_domain(cpu, sd) { > > > + /* > > > + * Decay the newidle max

Re: [RFC][PATCH v4 3/3] sched: Periodically decay max cost of idle balance

2013-09-09 Thread Peter Zijlstra
On Wed, Sep 04, 2013 at 12:10:01AM -0700, Jason Low wrote: > On Fri, 2013-08-30 at 12:18 +0200, Peter Zijlstra wrote: > > On Thu, Aug 29, 2013 at 01:05:36PM -0700, Jason Low wrote: > > > diff --git a/kernel/sched/core.c b/kernel/sched/core.c > > > index 58b0514..bba5a07 100644 > > > --- a/kernel/sc

Re: [RFC][PATCH v4 3/3] sched: Periodically decay max cost of idle balance

2013-09-09 Thread Peter Zijlstra
On Tue, Sep 03, 2013 at 11:02:59PM -0700, Jason Low wrote: > On Fri, 2013-08-30 at 12:29 +0200, Peter Zijlstra wrote: > > rcu_read_lock(); > > for_each_domain(cpu, sd) { > > + /* > > +* Decay the newidle max times here because this is a regular > > +* visit

Re: [RFC][PATCH v4 3/3] sched: Periodically decay max cost of idle balance

2013-09-04 Thread Jason Low
On Fri, 2013-08-30 at 12:18 +0200, Peter Zijlstra wrote: > On Thu, Aug 29, 2013 at 01:05:36PM -0700, Jason Low wrote: > > diff --git a/kernel/sched/core.c b/kernel/sched/core.c > > index 58b0514..bba5a07 100644 > > --- a/kernel/sched/core.c > > +++ b/kernel/sched/core.c > > @@ -1345,7 +1345,7 @@ tt

Re: [RFC][PATCH v4 3/3] sched: Periodically decay max cost of idle balance

2013-09-03 Thread Jason Low
On Fri, 2013-08-30 at 12:29 +0200, Peter Zijlstra wrote: > rcu_read_lock(); > for_each_domain(cpu, sd) { > + /* > + * Decay the newidle max times here because this is a regular > + * visit to all the domains. Decay ~0.5% per second. > +

Re: [RFC][PATCH v4 3/3] sched: Periodically decay max cost of idle balance

2013-08-30 Thread Peter Zijlstra
On Fri, Aug 30, 2013 at 12:18:17PM +0200, Peter Zijlstra wrote: > for_each_domain(cpu, sd) { > + /* > + * Decay the newidle max times here because this is a regular > + * visit to all the domains. Decay ~0.5% per second. > + */ > +

Re: [RFC][PATCH v4 3/3] sched: Periodically decay max cost of idle balance

2013-08-30 Thread Peter Zijlstra
On Thu, Aug 29, 2013 at 01:05:36PM -0700, Jason Low wrote: > diff --git a/kernel/sched/core.c b/kernel/sched/core.c > index 58b0514..bba5a07 100644 > --- a/kernel/sched/core.c > +++ b/kernel/sched/core.c > @@ -1345,7 +1345,7 @@ ttwu_do_wakeup(struct rq *rq, struct task_struct *p, > int wake_flags)

[RFC][PATCH v4 3/3] sched: Periodically decay max cost of idle balance

2013-08-29 Thread Jason Low
This RFC patch builds on patch 2 and periodically decays that max value to do idle balancing per sched domain. Though we want to decay it fairly consistently, we may not want to lower it by too much each time, especially since avg_idle is capped based on that value. So I thought that decaying the