Re: [PATCH 2/2] PM / Domains: Avoid infinite loops in attach/detach code

["Rafael J. Wysocki" <rjw@rjwysocki.net>]

On Wednesday, June 24, 2015 10:35:44 AM Geert Uytterhoeven wrote:
> On Wed, Jun 24, 2015 at 10:33 AM, Ulf Hansson <ulf.hansson@linaro.org> wrote:
> > [...]
> >
> >>>>
> >>>> @@ -2183,6 +2191,7 @@ int genpd_dev_pm_attach(struct device *dev)
> >>>>  {
> >>>>         struct of_phandle_args pd_args;
> >>>>         struct generic_pm_domain *pd;
> >>>> +       unsigned int i;
> >>>>         int ret;
> >>>>
> >>>>         if (!dev->of_node)
> >>>> @@ -2218,10 +2227,13 @@ int genpd_dev_pm_attach(struct device *dev)
> >>>>
> >>>>         dev_dbg(dev, "adding to PM domain %s\n", pd->name);
> >>>>
> >>>> -       while (1) {
> >>>> +       for (i = 0; i < GENPD_RETRIES; i++) {
> >>>>                 ret = pm_genpd_add_device(pd, dev);
> >>>>                 if (ret != -EAGAIN)
> >>>>                         break;
> >>>> +
> >>>> +               if (i > GENPD_RETRIES / 2)
> >>>> +                       udelay(GENPD_DELAY_US);
> >>>
> >>> In this execution path, we retry when getting -EAGAIN while believing
> >>> the reason to the error are only *temporary* as we are soon waiting
> >>> for all devices in the genpd to be system PM resumed. At least that's
> >>> my understanding to why we want to deal with -EAGAIN here, but I might
> >>> be wrong.
> >>>
> >>> In this regards, I wonder whether it could be better to re-try only a
> >>> few times but with a far longer interval time than a couple us. What
> >>> do you think?
> >>
> >> That's indeed viable. I have no idea for how long this temporary state can
> >> extend.
> >
> > That will depend on the system PM resume time for the devices residing
> > in the genpd. So, I guess we need a guestimate then. How about a total
> > sleep time of a few seconds?
> >
> >>
> >>> However, what if the reason to why we get -EAGAIN isn't *temporary*,
> >>> because we are about to enter system PM suspend state. Then the caller
> >>> of this function which comes via some bus' ->probe(), will hang until
> >>> the a system PM resume is completed. Is that really going to work? So,
> >>> for this case your limited re-try approach will affect this scenario
> >>> as well, have you considered that?
> >>
> >> There's a limit on the number of retries, so it won't hang indefinitely.
> >
> > What happens with the timer functions (like msleep()) during the
> > system PM suspend transition?
> 
> I guess we can no longer call msleep() after syscore suspend?

That's correct.  Time is effectively frozen at that point.

Rafael