Re: [PATCHv2] PM / devfreq: Add dev_pm_qos support

[Leonard Crestez <leonard.crestez@nxp.com>]

On 13.08.2019 09:10, Chanwoo Choi wrote:
> In case of cpufreq, cpufreq.c replace the body of store_min_freq()
> and store_max_freq() by using struct dev_pm_qos_request instancce
> with dev_pm_qos_update_request().
> 
> If you use the new way with dev_pm_qos_update_request() for
> min_freq_store() and max_freq_store(), it doesn't need to
> get the final frequency from three candidate frequencies.

Yes, I saw that but didn't implement the equivalent for devfreq because 
it's not clear what there is to gain.

Since dev_pm_qos is measured in khz it means that min_freq/max_req on 
sysfq would lose 3 significant digits, however those digits are probably 
useless anyway.

> In result, We only consider the following two candidate frequencies
> as following:
> 1. "devfreq->scaling_min_freq" will contain the requirement
>     from devfreq thermal throttling by OPP interface.

It's a bit awkward that the OPPs enable/disable API is not obviously 
specific to "thermal".

> 2. "devfreq->min_freq" will contain the requirements
>     from both user input through sysfs and the dev_pm_qos_request.
According to a comment on a previous version it would be useful to have 
a separate files for "constraint min/max freq" and "user min/max freq":

     https://patchwork.kernel.org/patch/11078475/#22805379

Combining min/max requests from dev_pm_qos and sysfs would make this 
harder to implement. I guess user_min_freq could be implemented by 
reading back the dev_pm_qos request but there would be no way to 
implement a qos_min_freq entry.

>> +static int devfreq_qos_min_notifier_call(struct notifier_block *nb, unsigned long val, void *ptr)
> 
> Please keep the under 80 line if there are no any special reason.

OK, will check.