Re: [PATCH 1/3] clocksource: davinci: work around a clocksource problem on dm365 SoC

[David Lechner]

On 12/17/19 2:00 AM, Bartosz Golaszewski wrote:
> pon., 16 gru 2019 o 18:58 David Lechner <david@xxxxxxxxxxxxxx> napisaÅ(a):
> > > +static unsigned int davinci_clocksource_tim32_mode;
> >
> > static unsigned bool davinci_clocksource_initialized;
> >
> > > +
> > >    static struct davinci_clockevent *
> > >    to_davinci_clockevent(struct clock_event_device *clockevent)
> > >    {
> > > @@ -94,7 +96,7 @@ static void davinci_tim12_shutdown(void __iomem *base)
> > >         * halves. In this case TIM34 runs in periodic mode and we must
> > >         * not modify it.
> > >         */
> > > -     tcr |= DAVINCI_TIMER_ENAMODE_PERIODIC <<
> > > +     tcr |= davinci_clocksource_tim32_mode <<
> > >                DAVINCI_TIMER_ENAMODE_SHIFT_TIM34;
> >
> >          if (davinci_clocksource_initialized)
> >                  tcr |= DAVINCI_TIMER_ENAMODE_PERIODIC <<
> >                          DAVINCI_TIMER_ENAMODE_SHIFT_TIM34;
>
> I thought about doing this initially, but then figured this code would
> be called a lot, so why not avoid branching and just store the right
> value? Alternatively we can do:
>
>      if (likely(davinci_clocksource_initialized)
>          ....

Unless there is a measurable performance difference, I think it
would be better with if/likely.