Re: [patch 2/4] timers: Consolidate base->next_timer update

[Nikunj A Dadhania <nikunj@linux.vnet.ibm.com>]

On Tue, 29 May 2012 11:38:27 +0200 (CEST), Thomas Gleixner <tglx@linutronix.de> wrote:
> On Tue, 29 May 2012, Nikunj A Dadhania wrote:
> 
> > > 
> > > -static void internal_add_timer(struct tvec_base *base, struct timer_list *timer)
> > > +static void
> > > +__internal_add_timer(struct tvec_base *base, struct timer_list *timer)
> > >  {
> > >  	unsigned long expires = timer->expires;
> > >  	unsigned long idx = expires - base->timer_jiffies;
> > > @@ -372,6 +373,17 @@ static void internal_add_timer(struct tv
> > >  	list_add_tail(&timer->entry, vec);
> > >  }
> > > 
> > > +static void internal_add_timer(struct tvec_base *base, struct timer_list *timer)
> > > +{
> > > +	__internal_add_timer(base, timer);
> > > +	/*
> > > +	 * Update base->next_timer if this is the earliest one.
> > > +	 */
> > > +	if (time_before(timer->expires, base->next_timer) &&
> > > +	    !tbase_get_deferrable(timer->base))
> > > +		base->next_timer = timer->expires;
> > > +}
> > > +
> > >
> > Shouldn't this be like this?
> > 
> > +	/*
> > +	 * Update base->next_timer if this is the earliest one.
> > +	 */
> > +	if (time_before(timer->expires, base->next_timer) &&
> > +	    !tbase_get_deferrable(timer->base))
> > +		base->next_timer = timer->expires;
> > +	__internal_add_timer(base, timer);
> > 
> > As per the below code?
> 
> And why should this matter?
> 
Yes it does not matter, sorry for the noise.

Looking at the internal_add_timer(), there is no such dependency. I was
thinking that the base->next_timer is changed and would be used in
__internal_add_timer.

Nikunj