Re: [PATCH v3 1/2] clk: qcom: camcc-sm6350: add pm_runtime support

["Luca Weiss" <luca.weiss@fairphone.com>]

On Tue Mar 7, 2023 at 4:06 PM CET, Dmitry Baryshkov wrote:
> On Tue, 7 Mar 2023 at 16:54, Luca Weiss <luca.weiss@fairphone.com> wrote:
> >
> > On Tue Feb 14, 2023 at 1:32 PM CET, Dmitry Baryshkov wrote:
> > > On Tue, 14 Feb 2023 at 13:01, Luca Weiss <luca.weiss@fairphone.com> wrote:
> > > >
> > > > Make sure that we can enable and disable the power domains used for
> > > > camcc when the clocks are and aren't used.
> > > >
> > > > Signed-off-by: Luca Weiss <luca.weiss@fairphone.com>
> > > > ---
> > > >  drivers/clk/qcom/camcc-sm6350.c | 25 ++++++++++++++++++++++++-
> > > >  1 file changed, 24 insertions(+), 1 deletion(-)
> > > >
> > > > diff --git a/drivers/clk/qcom/camcc-sm6350.c b/drivers/clk/qcom/camcc-sm6350.c
> > > > index acba9f99d960..fc5532e2ee5b 100644
> > > > --- a/drivers/clk/qcom/camcc-sm6350.c
> > > > +++ b/drivers/clk/qcom/camcc-sm6350.c
> > > > @@ -7,6 +7,8 @@
> > > >  #include <linux/clk-provider.h>
> > > >  #include <linux/module.h>
> > > >  #include <linux/platform_device.h>
> > > > +#include <linux/pm_clock.h>
> > > > +#include <linux/pm_runtime.h>
> > > >  #include <linux/regmap.h>
> > > >
> > > >  #include <dt-bindings/clock/qcom,sm6350-camcc.h>
> > > > @@ -1869,6 +1871,19 @@ MODULE_DEVICE_TABLE(of, camcc_sm6350_match_table);
> > > >  static int camcc_sm6350_probe(struct platform_device *pdev)
> > > >  {
> > > >         struct regmap *regmap;
> > > > +       int ret;
> > > > +
> > > > +       ret = devm_pm_runtime_enable(&pdev->dev);
> > > > +       if (ret < 0)
> > > > +               return ret;
> > > > +
> > > > +       ret = devm_pm_clk_create(&pdev->dev);
> > > > +       if (ret < 0)
> > > > +               return ret;
> > >
> > > This makes me wonder, what is the use for the pm_clk in your case? The
> > > driver doesn't seem to use of_pm_clk_add_clk(), of_pm_clk_add_clks()
> > > or pm_clk_add_clk(). So pm_clk_suspend() and pm_clk_resume() do
> > > nothing.
> >
> > You're right that we're not using any of these functions in the driver.
> > However still when camcc is not used, the associated power domain turns
> > off correctly so that part works as expected.
> >
> > Honestly these lines have been copied from a different driver and I'm
> > not familiar enough with the pm_runtime APIs to know what to use here
> > without using the pm_clk* and pm_clk_suspend.
>
> Please don't blindly C&P code.

I normally try to avoid this.. however pm_runtime is still quite a
mystery to me.

>
> >
> > Basically we need, if any clock is being used in the driver, the
> > power-domain needs to be enabled as well, and if nothing is used the
> > power-domain can be disabled again.
>
> Adding power-domains to the camcc node and calling
> devm_pm_runtime_enable() should be enough. Please see how this is
> managed for dispcc on sm8250.

Following that driver, so just using devm_pm_runtime_enable and
pm_runtime_resume_and_get doesn't seem to enable runtime_pm for the
device..

  $ cat /sys/devices/platform/soc@0/ad00000.clock-controller/power/runtime_status 
  unsupported

Also then I don't see the device in /sys/kernel/debug/pm_genpd/pm_genpd_summary

I'm guessing we still need to set the dev_pm_ops with something? Not
sure what to use here if it should just enable/disable the power domains
that are not handled by the driver directly. I also couldn't find any
example in existing code unfortuantely.

Any pointers?

Regards
Luca

>
> -- 
> With best wishes
> Dmitry