Re: [PATCH] cpuidle: avoid using smp_processor_id() in preemptible code (nr_iowait_cpu) v4

[Andrew Morton <akpm@linux-foundation.org>]

On Fri, 25 Jun 2010 16:39:33 +0200
Peter Zijlstra <peterz@infradead.org> wrote:

> On Thu, 2010-06-17 at 09:29 +0300, Sergey Senozhatsky wrote:
> > Fix
> >    
> >  BUG: using smp_processor_id() in preemptible [00000000] code: s2disk/3392
> 
> > The initial fix was to use get_cpu/put_cpu in nr_iowait_cpu.  However,
> > Arjan stated that "the bug is that it needs to be nr_iowait_cpu(int cpu)".
> > 
> > This patch introduces nr_iowait_cpu(int cpu) and changes to its callers.
> > 
> > Arjan also pointed out that we can't use get_cpu/put_cpu in update_ts_time_stats
> > since we "pick the current cpu, rather than the one denoted by ts" in that case.
> > To match given *ts and cpu denoted by *ts we use new field in the struct tick_sched: int cpu.
> 
> 
> > diff --git a/include/linux/tick.h b/include/linux/tick.h
> > index b232ccc..db14691 100644
> > --- a/include/linux/tick.h
> > +++ b/include/linux/tick.h
> > @@ -51,6 +51,7 @@ struct tick_sched {
> >  	unsigned long			check_clocks;
> >  	enum tick_nohz_mode		nohz_mode;
> >  	ktime_t				idle_tick;
> > +	int				cpu;
> >  	int				inidle;
> >  	int				tick_stopped;
> >  	unsigned long			idle_jiffies;
> 
> > diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c
> > index 1d7b9bc..1907037 100644
> > --- a/kernel/time/tick-sched.c
> > +++ b/kernel/time/tick-sched.c
> > @@ -38,6 +38,9 @@ static ktime_t last_jiffies_update;
> >  
> >  struct tick_sched *tick_get_tick_sched(int cpu)
> >  {
> > +	/*FIXME: Arjan van de Ven:
> > +	 can we do this bit once, when the ts structure gets initialized?*/
> > +	per_cpu(tick_cpu_sched, cpu).cpu = cpu;
> >  	return &per_cpu(tick_cpu_sched, cpu);
> >  }
> 
> > @@ -161,7 +164,7 @@ update_ts_time_stats(struct tick_sched *ts, ktime_t now, u64 *last_update_time)
> >  	if (ts->idle_active) {
> >  		delta = ktime_sub(now, ts->idle_entrytime);
> >  		ts->idle_sleeptime = ktime_add(ts->idle_sleeptime, delta);
> > -		if (nr_iowait_cpu() > 0)
> > +		if (nr_iowait_cpu(ts->cpu) > 0)
> >  			ts->iowait_sleeptime = ktime_add(ts->iowait_sleeptime, delta);
> >  		ts->idle_entrytime = now;
> >  	}
> 
> 
> This all seems extremely silly, why not something like:

Does it work?

c'mon guys, it's taking us weeks and weeks to fix one simple bug.  It's
a regression!  We should be in panic mode.