[PATCH v2] power: domain: handle power supplies that need interrupts

[Ulf Hansson]

On Tue, 12 Jul 2022 at 14:19, Martin Kepplinger
<martin.kepplinger at puri.sm> wrote:
>
> If the power-domains' power-supply node (regulator) needs
> interrupts to work, the current setup with noirq callbacks cannot
> work; for example a pmic regulator on i2c, when suspending, usually already
> times out during suspend_noirq:
>
> [   41.024193] buck4: failed to disable: -ETIMEDOUT
>
> So fix system suspend and resume for these power-domains by using the
> "outer" suspend/resume callbacks instead. Tested on the imx8mq-librem5
> board, but by looking at the dts, this will fix imx8mq-evk and possibly
> other boards too.
>
> Possibly one can find more changes than suspend/resume for this case. They
> can be added later when testing them.
>
> Initially system suspend problems had been discussed at
> https://lore.kernel.org/linux-arm-kernel/20211002005954.1367653-8-l.stach@pengutronix.de/
> which led to discussing the pmic that contains the regulators which
> serve as power-domain power-supplies:
> https://lore.kernel.org/linux-pm/573166b75e524517782471c2b7f96e03fd93d175.camel@puri.sm/T/
>
> Signed-off-by: Martin Kepplinger <martin.kepplinger at puri.sm>
> ---
>
> revision history
> ----------------
> v2: (thank you Krzysztof)
> * rewrite: find possible regulators' interrupts property in parents
>   instead of inventing a new property.
>
> v1: (initial idea)
> https://lore.kernel.org/linux-arm-kernel/20220711094549.3445566-1-martin.kepplinger@puri.sm/T/#t
>
>
>  drivers/base/power/domain.c | 26 ++++++++++++++++++++++++++
>  1 file changed, 26 insertions(+)
>
> diff --git a/drivers/base/power/domain.c b/drivers/base/power/domain.c
> index 3e86772d5fac..ca3e3500939d 100644
> --- a/drivers/base/power/domain.c
> +++ b/drivers/base/power/domain.c
> @@ -2298,6 +2298,28 @@ static bool genpd_present(const struct generic_pm_domain *genpd)
>         return ret;
>  }
>
> +/**
> + * of_genpd_get_power_supply_irq() - Adjust if power-supply needs interrupts
> + * @genpd: Pointer to PM domain associated with the PM domain provider.
> + */
> +static void of_genpd_get_power_supply_irq(struct generic_pm_domain *pd)
> +{
> +       struct device_node *dn;
> +
> +       dn = of_parse_phandle(pd->dev.of_node, "power-supply", 0);
> +       if (!dn)
> +               return;
> +
> +       while ((dn = of_get_next_parent(dn))) {
> +               if (of_get_property(dn, "interrupts", NULL)) {
> +                       pd->domain.ops.suspend = genpd_suspend_noirq;
> +                       pd->domain.ops.resume = genpd_resume_noirq;
> +                       pd->domain.ops.suspend_noirq = NULL;
> +                       pd->domain.ops.resume_noirq = NULL;
> +               }
> +       }
> +}
> +
>  /**
>   * of_genpd_add_provider_simple() - Register a simple PM domain provider
>   * @np: Device node pointer associated with the PM domain provider.
> @@ -2343,6 +2365,8 @@ int of_genpd_add_provider_simple(struct device_node *np,
>         genpd->provider = &np->fwnode;
>         genpd->has_provider = true;
>
> +       of_genpd_get_power_supply_irq(genpd);
> +
>         return 0;
>  }
>  EXPORT_SYMBOL_GPL(of_genpd_add_provider_simple);
> @@ -2394,6 +2418,8 @@ int of_genpd_add_provider_onecell(struct device_node *np,
>
>                 genpd->provider = &np->fwnode;
>                 genpd->has_provider = true;
> +
> +               of_genpd_get_power_supply_irq(genpd);
>         }
>
>         ret = genpd_add_provider(np, data->xlate, data);

Overall I understand the need for this, but let me suggest a slightly
different approach to solve this. See below.

I think the OF parsing looks quite platform specific. Rather than
adding this in the generic layer of genpd, I suggest that we move the
OF parsing into the genpd provider code.

Moreover, to inform genpd that it should use the other set of
callbacks for system suspend/resume, let's add a new genpd
configuration bit. The genpd provider should then set the genpd->flag,
prior to calling pm_genpd_init(), to let it know that it should pick
the other callbacks.

Does it make sense?

Kind regards
Uffe