Re: [PATCH 1/4] cpufreq: stats: Defer stats update to cpufreq_stats_record_transition()

[Viresh Kumar]

On 15-09-20, 11:04, Lukasz Luba wrote:
> Hi Viresh,
> 
> On 9/2/20 8:24 AM, Viresh Kumar wrote:
> > In order to prepare for lock-less stats update, add support to defer any
> > updates to it until cpufreq_stats_record_transition() is called.
> > 
> > Signed-off-by: Viresh Kumar <viresh.kumar@xxxxxxxxxx>
> > ---
> >   drivers/cpufreq/cpufreq_stats.c | 75 ++++++++++++++++++++++++---------
> >   1 file changed, 56 insertions(+), 19 deletions(-)
> > 
> > diff --git a/drivers/cpufreq/cpufreq_stats.c b/drivers/cpufreq/cpufreq_stats.c
> > index 94d959a8e954..fdf9e8556a49 100644
> > --- a/drivers/cpufreq/cpufreq_stats.c
> > +++ b/drivers/cpufreq/cpufreq_stats.c
> > @@ -22,17 +22,22 @@ struct cpufreq_stats {
> 
> Would it be possible to move this structure in the
> linux/cpufreq.h header? Any subsystem could have access to it,
> like to the cpuidle stats.

Hmm, I am not sure why we should be doing it. In case of cpuidle many
parts of the kernel are playing with cpuidle code, like drivers/idle/,
drivers/cpuidle, etc.

Something should land in include/ only if you want others to use it,
but in case of cpufreq no one should be using cpufreq stats.

So unless you have a real case where that might be beneficial, I am
going to keep it as is.

> Apart from that (and the comment regarding the 'atomic_t' field)
> I don't see any issues.

Thanks.

-- 
viresh