[PATCH V4 07/14] cpufreq: mediatek: Add .get function

Viresh Kumar viresh.kumar at linaro.org
Thu Apr 28 04:48:35 PDT 2022


On 28-04-22, 19:16, Rex-BC Chen wrote:
> Yes, the call stack will eventually go to __cpufreq_driver_target.
> However, we can observe the mismatch between target_freq and policy-cur 
> with a tiny difference.
> e.g.
> [ 553.065356] cpufreq: target for CPU 0: 500000 kHz, relation 0,
> requested 500000 kHz
> [ 553.066366] cpufreq: target_freq/policy->cur: 500000/499999 kHz

So you are trying to set the frequency to 500 MHz now, but policy->cur says it
is 499 MHz.

> We check the assignment of policy->cur could be either from
> cpufreq_driver->get_intermediate or from cpufreq_driver->get.

policy->cur is set only at two places, in your case:
- CPUFREQ_POSTCHANGE
- cpufreq_online()

>From what I understand, it is possible that cpufreq_online() is setting your
frequency to 499999 (once at boot), but as soon as a frequency change has
happened after that, policy->cur should be set to 500 MHz and you should see
this problem only once.

>From CPUFREQ_POSTCHANGE notifier, we always set policy->cur from the table
itself, which should be 500000 MHz.

I wonder how you see policy->cur to be 499999 here. Does this happen only once ?
Or repeatedly ?

> But it is strange to have the frequency value like 499999 kHz.
> Is the result of tiny frequency difference expected from your point of
> view?

Clock driver can give this value, that is fine.

> > What do you mean by "voltage pulse" here? What actually happens which
> > you want to avoid.
> > 
> 
> When cpufreq is fixed to lowest opp, "voltage pulse" is a quick voltage
> rising and falling phenomenon which can be observed if 'cpufreq_get' is
> invoked.

Do check if the call is reaching your driver's ->target_index(), it should be
which it should not, ideally.

> Thank you for sharing the correct information.
> Is it possible to get frequency (API) a simple way, like get current
> opp frequency?

Lets dig/debug a bit further and fix this if a real problem exists.

-- 
viresh



More information about the Linux-mediatek mailing list