linux-kernel - Re: [PATCH 2/2] sched: trigger_load

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <1347488204.26695.230.camel@sbsiddha-desk.sc.intel.com>
Date:	Wed, 12 Sep 2012 15:16:44 -0700
From:	Suresh Siddha <suresh.b.siddha@...el.com>
To:	Alex Shi <alex.shi@...el.com>
Cc:	mingo@...hat.com, peterz@...radead.org, tglx@...utronix.de,
	linux-kernel@...r.kernel.org, venki@...gle.com,
	"Paul E. McKenney" <paulmck@...ux.vnet.ibm.com>
Subject: Re: [PATCH 2/2] sched: trigger_load_balance clean up

On Wed, 2012-09-12 at 15:58 +0800, Alex Shi wrote:
> On 09/12/2012 02:36 AM, Suresh Siddha wrote:
> 
> > On Mon, 2012-09-10 at 15:10 +0800, Alex Shi wrote:
> >> Remove a redundant check for on_null_domain(cpu), and rerange the code
> >> that make it more readable.
> > 
> > hmm, but we are now doing the on_null_domain() check always,
> > irrespective of whether we need the load balance or not.
> > 
> > do we really need the on_null_domain() check there? What happens if we
> > just remove it?
> 
> 
> A very very simple try can not show removing causes crash. But as to
> RCU details, I don't know. :(
> 

So the right way to approach the issue is to do couple of "git blame"
and see who added that code. My check shows 

commit 8a0be9ef8225638d26b455788f988c8f84ce9e75
Author: Frederic Weisbecker <fweisbec@...il.com>
Date:   Thu Mar 5 01:27:02 2009 +0100

    sched: don't rebalance if attached on NULL domain

And the changelog explains that it has to do with avoiding unnecessary
loadbalances (triggered by softirq) for a SMP kernel on a UP system and
actually has nothing to do with RCU per-say (as the load balancer
already takes the rcu locks etc before traversing the scheduler-domain).

And the second on_null_domain() in that function is added by Venki and
me, probably by mistake. As the nohz_kick_needed() will never return
true for the SMP kernel running on a UP system.

So please just remove the second on_null_domain() check in that function
instead of consolidating both the usages and include the above info in
the changelog.

thanks,
suresh

> > 
> > thanks,
> > suresh
> > 
> >>
> >> Signed-off-by: Alex Shi <alex.shi@...el.com>
> >> ---
> >>  kernel/sched/fair.c |    8 +++++---
> >>  1 files changed, 5 insertions(+), 3 deletions(-)
> >>
> >> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> >> index 5bbc4bf..529092d 100644
> >> --- a/kernel/sched/fair.c
> >> +++ b/kernel/sched/fair.c
> >> @@ -4934,11 +4934,13 @@ static inline int on_null_domain(int cpu)
> >>  void trigger_load_balance(struct rq *rq, int cpu)
> >>  {
> >>  	/* Don't need to rebalance while attached to NULL domain */
> >> -	if (time_after_eq(jiffies, rq->next_balance) &&
> >> -	    likely(!on_null_domain(cpu)))
> >> +	if (unlikely(on_null_domain(cpu)))
> >> +		return;
> >> +
> >> +	if (time_after_eq(jiffies, rq->next_balance))
> >>  		raise_softirq(SCHED_SOFTIRQ);
> >>  #ifdef CONFIG_NO_HZ
> >> -	if (nohz_kick_needed(rq, cpu) && likely(!on_null_domain(cpu)))
> >> +	if (nohz_kick_needed(rq, cpu))
> >>  		nohz_balancer_kick(cpu);
> >>  #endif
> >>  }
> > 
> > 
> 
> 


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/