Re: [PATCH V2 11/13] devfreq: tegra30: Migrate to dev_pm_opp_set_opp()

[Dmitry Osipenko]

27.01.2021 13:02, Viresh Kumar пишет:
> On 27-01-21, 14:40, Viresh Kumar wrote:
>> dev_pm_opp_set_bw() is getting removed and dev_pm_opp_set_opp() should
>> be used instead. Migrate to the new API.
>>
>> We don't want the OPP core to manage the clk for this driver, migrate to
>> dev_pm_opp_of_add_table_noclk() to make sure dev_pm_opp_set_opp()
>> doesn't have any side effects.
>>
>> Signed-off-by: Viresh Kumar <viresh.kumar@xxxxxxxxxx>
>> ---
>> Dmitry,
>>
>> This is based over the patches sent here:
>>
>> https://lore.kernel.org/lkml/6c2160ff30a8f421563793020264cf9f533f293c.1611738228.git.viresh.kumar@xxxxxxxxxx/
>>
>> This should fix the problem you mentioned earlier. Will push this for
>> linux-next unless you have any issues with it.
>>
>>  drivers/devfreq/tegra30-devfreq.c | 4 ++--
>>  1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/devfreq/tegra30-devfreq.c b/drivers/devfreq/tegra30-devfreq.c
>> index 117cad7968ab..31f7dec5990b 100644
>> --- a/drivers/devfreq/tegra30-devfreq.c
>> +++ b/drivers/devfreq/tegra30-devfreq.c
>> @@ -647,7 +647,7 @@ static int tegra_devfreq_target(struct device *dev, unsigned long *freq,
>>  		return PTR_ERR(opp);
>>  	}
>>  
>> -	ret = dev_pm_opp_set_bw(dev, opp);
>> +	ret = dev_pm_opp_set_opp(dev, opp);
>>  	dev_pm_opp_put(opp);
>>  
>>  	return ret;
>> @@ -849,7 +849,7 @@ static int tegra_devfreq_probe(struct platform_device *pdev)
>>  		return err;
>>  	}
>>  
>> -	err = dev_pm_opp_of_add_table(&pdev->dev);
>> +	err = dev_pm_opp_of_add_table_noclk(&pdev->dev);
> 
> Plus this, somehow was left uncommited in my tree :(
> 
> diff --git a/drivers/devfreq/tegra30-devfreq.c b/drivers/devfreq/tegra30-devfreq.c
> index 31f7dec5990b..ce83f883ca65 100644
> --- a/drivers/devfreq/tegra30-devfreq.c
> +++ b/drivers/devfreq/tegra30-devfreq.c
> @@ -849,7 +849,7 @@ static int tegra_devfreq_probe(struct platform_device *pdev)
>                 return err;
>         }
>  
> -       err = dev_pm_opp_of_add_table_noclk(&pdev->dev);
> +       err = dev_pm_opp_of_add_table_noclk(&pdev->dev, 0);
>         if (err) {
>                 dev_err(&pdev->dev, "Failed to add OPP table: %d\n", err);
>                 goto put_hw;
> 

Sadly this doesn't work because we missed that clk is assigned to
opp_table when OPP table is allocated and not when it's added to device.

Hence we're now set back to the dev_pm_opp_clear_clk() variant.

What about to add a new OPP API which will allow OPP users to configure
behaviour that user wants from OPP core in a generic way, something like
this:

struct opp_config {
	bool no_clk;
	...
};

devm_pm_opp_set_config(dev, struct opp_config);
dev_pm_opp_set_config(dev, struct opp_config);
dev_pm_opp_unset_config(dev);

Or maybe even rename it dev_pm_opp_allocate_table(dev, struct
opp_config), which will allow users to directly allocate OPP table
instead of relying on the implicit allocations. Then there won't be a
need for drivers to use a dummy devm_pm_opp_set_clkname(dev, NULL) just
to allocate the table usable for dev_pm_opp_set_rate().