Re: [PATCH v2 19/20] mips: cevt-r4k: Update the r4k-clockevent frequency in sync with CPU

[Serge Semin]

On Fri, May 08, 2020 at 05:40:46PM +0200, Thomas Bogendoerfer wrote:
> On Wed, May 06, 2020 at 08:42:37PM +0300, Sergey.Semin@xxxxxxxxxxxxxxxxxxxx wrote:
> > diff --git a/arch/mips/kernel/cevt-r4k.c b/arch/mips/kernel/cevt-r4k.c
> > index 17a9cbb8b3df..f5b72fb7d5ee 100644
> > --- a/arch/mips/kernel/cevt-r4k.c
> > +++ b/arch/mips/kernel/cevt-r4k.c
> > @@ -8,6 +8,7 @@
> >   */
> >  #include <linux/clockchips.h>
> >  #include <linux/interrupt.h>
> > +#include <linux/cpufreq.h>
> >  #include <linux/percpu.h>
> >  #include <linux/smp.h>
> >  #include <linux/irq.h>
> > @@ -250,6 +251,49 @@ unsigned int __weak get_c0_compare_int(void)
> >  	return MIPS_CPU_IRQ_BASE + cp0_compare_irq;
> >  }
> >  
> > +#ifdef CONFIG_CPU_FREQ
> > +
> > +static unsigned long mips_ref_freq;
> > +
> > +static int cpufreq_callback(struct notifier_block *nb,
> > +			    unsigned long val, void *data)
> 
> please prefix function names with r4k_ to make them different from
> the other ones you implemented in kernel/time.c. I know they are
> static, but keeping different names makes looking at crashes easier.

Agreed. I'll fix it in v3.

> 
> > +	struct cpufreq_freqs *freq = data;
> > +	struct clock_event_device *cd;
> > +	unsigned long rate;
> > +	int cpu;
> > +
> > +	if (!mips_ref_freq)
> > +		mips_ref_freq = freq->old;
> 
> isn't this the same as mips_hpt_frequency ?

No. Here I save the initial CPU frequency so use one then to scale the
mips_hpt_frequency value in accordance with the CPU clock rate change. Yes,
mips_hpt_frequency value may initially match the CPU frequency on some platforms
but normally the r4k timer is clocked with half of it while some systems may have
a complicated algorithm of the timer ref clock rate calculation.

-Sergey

> 
> Thomas.
> 
> -- 
> Crap can work. Given enough thrust pigs will fly, but it's not necessarily a
> good idea.                                                [ RFC1925, 2.3 ]