[PATCH] cpuidle: riscv-sbi: Allow cpuidle pd used by other devices

[Nick Hu]

Hi Anup

Thanks for your feedback!

On Tue, Aug 13, 2024 at 8:25 PM Anup Patel <anup at brainfault.org> wrote:
>
> On Tue, Aug 13, 2024 at 1:43 PM Nick Hu <nick.hu at sifive.com> wrote:
> >
> > To prevent the probe of consumer devices being deferred, create the
> > platform devices for each pd node under '/cpus/power-domains' and move the
> > driver initailization to the arch_initcall.
>
> To prevent the probe deferral of consumer devices, you can simply use
> fwnode_dev_initialized() instead of creating a dummy platform device.
>
You are right. I'll remove the dummy platform device.
If the fwnode_dev_initialized() was done before
fw_devlink_create_devlink(), the link will be dropped.
The fwnode_dev_initialized() is already included in the
of_genpd_add_provider_simple() and If the fwnode_dev_initialized()
was done before fw_devlink_create_devlink(), the link will be dropped.
so I think we only need to move the driver initialization
to the arch_initcall.
Thanks!

> > The consumer devices that inside the cpu/cluster power domain may register
> > the genpd notifier where their power domains are point to the pd nodes
> > under '/cpus/power-domains'. If the cpuidle.off==1, the genpd notifier
> > will fail due to sbi_cpuidle_pd_allow_domain_state is not set. We also
> > need the sbi_cpuidle_cpuhp_up/down to invoke the callbacks. Therefore,
> > add a cpuidle_disabled() check before registering the cpuidle driver to
> > address the issue.
>
> I think dealing with cpuidle.off==1 case should be a separate patch.
>
Sure, I'll update it in the next patch

> >
> > Signed-off-by: Nick Hu <nick.hu at sifive.com>
> > Link: https://lore.kernel.org/lkml/20240226065113.1690534-1-nick.hu@sifive.com/
> > Suggested-by: Anup Patel <apatel at ventanamicro.com>
> > ---
> >  drivers/cpuidle/cpuidle-riscv-sbi.c | 24 +++++++++++++++++++++---
> >  1 file changed, 21 insertions(+), 3 deletions(-)
> >
> > diff --git a/drivers/cpuidle/cpuidle-riscv-sbi.c b/drivers/cpuidle/cpuidle-riscv-sbi.c
> > index a6e123dfe394..d6b01fc64f94 100644
> > --- a/drivers/cpuidle/cpuidle-riscv-sbi.c
> > +++ b/drivers/cpuidle/cpuidle-riscv-sbi.c
> > @@ -16,6 +16,7 @@
> >  #include <linux/kernel.h>
> >  #include <linux/module.h>
> >  #include <linux/of.h>
> > +#include <linux/of_platform.h>
> >  #include <linux/slab.h>
> >  #include <linux/platform_device.h>
> >  #include <linux/pm_domain.h>
> > @@ -25,6 +26,7 @@
> >  #include <asm/smp.h>
> >  #include <asm/suspend.h>
> >
> > +#include "cpuidle.h"
> >  #include "dt_idle_states.h"
> >  #include "dt_idle_genpd.h"
> >
> > @@ -336,6 +338,9 @@ static int sbi_cpuidle_init_cpu(struct device *dev, int cpu)
> >                 return ret;
> >         }
> >
> > +       if (cpuidle_disabled())
> > +               return 0;
> > +
> >         ret = cpuidle_register(drv, NULL);
> >         if (ret)
> >                 goto deinit;
> > @@ -380,20 +385,26 @@ static int sbi_cpuidle_pd_power_off(struct generic_pm_domain *pd)
> >  struct sbi_pd_provider {
> >         struct list_head link;
> >         struct device_node *node;
> > +       struct platform_device *pdev;
> >  };
> >
> >  static LIST_HEAD(sbi_pd_providers);
> >
> >  static int sbi_pd_init(struct device_node *np)
> >  {
> > +       struct platform_device *pdev;
> >         struct generic_pm_domain *pd;
> >         struct sbi_pd_provider *pd_provider;
> >         struct dev_power_governor *pd_gov;
> >         int ret = -ENOMEM;
> >
> > +       pdev = of_platform_device_create(np, np->name, NULL);
> > +       if (!pdev)
> > +               goto out;
> > +
> >         pd = dt_idle_pd_alloc(np, sbi_dt_parse_state_node);
> >         if (!pd)
> > -               goto out;
> > +               goto free_pdev;
> >
> >         pd_provider = kzalloc(sizeof(*pd_provider), GFP_KERNEL);
> >         if (!pd_provider)
> > @@ -419,6 +430,7 @@ static int sbi_pd_init(struct device_node *np)
> >                 goto remove_pd;
> >
> >         pd_provider->node = of_node_get(np);
> > +       pd_provider->pdev = pdev;
> >         list_add(&pd_provider->link, &sbi_pd_providers);
> >
> >         pr_debug("init PM domain %s\n", pd->name);
> > @@ -430,6 +442,8 @@ static int sbi_pd_init(struct device_node *np)
> >         kfree(pd_provider);
> >  free_pd:
> >         dt_idle_pd_free(pd);
> > +free_pdev:
> > +       of_platform_device_destroy(&pdev->dev, NULL);
> >  out:
> >         pr_err("failed to init PM domain ret=%d %pOF\n", ret, np);
> >         return ret;
> > @@ -447,6 +461,7 @@ static void sbi_pd_remove(void)
> >                 if (!IS_ERR(genpd))
> >                         kfree(genpd);
> >
> > +               of_platform_device_destroy(&pd_provider->pdev->dev, NULL);
> >                 of_node_put(pd_provider->node);
> >                 list_del(&pd_provider->link);
> >                 kfree(pd_provider);
> > @@ -548,7 +563,10 @@ static int sbi_cpuidle_probe(struct platform_device *pdev)
> >         /* Setup CPU hotplut notifiers */
> >         sbi_idle_init_cpuhp();
> >
> > -       pr_info("idle driver registered for all CPUs\n");
> > +       if (cpuidle_disabled())
> > +               pr_info("cpuidle is disabled\n");
> > +       else
> > +               pr_info("idle driver registered for all CPUs\n");
> >
> >         return 0;
> >
> > @@ -592,4 +610,4 @@ static int __init sbi_cpuidle_init(void)
> >
> >         return 0;
> >  }
> > -device_initcall(sbi_cpuidle_init);
> > +arch_initcall(sbi_cpuidle_init);
> > --
> > 2.34.1
> >
>
> Regards,
> Anup

Regards,
Nick