On Fri, 4 Dec 2020 at 15:14, Arnd Bergmann <arnd@xxxxxxxxxx> wrote:
>
> On Fri, Dec 4, 2020 at 11:02 AM Ulf Hansson <ulf.hansson@xxxxxxxxxx> wrote:
> > On Thu, 3 Dec 2020 at 23:29, Arnd Bergmann <arnd@xxxxxxxxxx> wrote:
>
> > > -#ifdef CONFIG_PM
> > > static void msdc_save_reg(struct msdc_host *host)
> >
> > Shouldn't msdc_save|restore_reg() be turned into "__maybe_unused" as well?
>
> There is no need since the compiler can figure that out already when there
> is a reference to the function from dead code.
Alright, thanks for clarifying.
>
> > >
> > > -static int msdc_resume(struct device *dev)
> > > +static int __maybe_unused msdc_resume(struct device *dev)
> > > {
> > > return pm_runtime_force_resume(dev);
> > > }
> > > -#endif
> > >
> > > static const struct dev_pm_ops msdc_dev_pm_ops = {
> >
> > You may also change this to a __maybe_unused, as long as you also
> > assign the .pm pointer in the mt_msdc_driver with
> > pm_ptr(&msdc_dev_pm_ops).
> >
> > Ideally the compiler should drop these functions/datas entirely then.
>
> I don't see a lot of other instances of that yet, and it's fairly new.
> Maybe we should fix it before it gets propagated further.
>
> I would suggest we redefine pm_ptr like
>
> #define pm_ptr(_ptr) (IS_ENABLED(CONFIG_PM) ? (_ptr) : NULL)
Why is this better than the original definition?
>
> and remove the __maybe_unused annotations on those that we
> already have. This also has the effect of dropping the unused
> data from the object, but without having to an an #ifdef or
> __maybe_unused.
I didn't quite get this (sorry it's Friday afternoon... getting
tired), can you perhaps give a concrete example?
That said, I have applied your patch for fixes, but let's try to sort
out the above to make sure we are all on the same page.
Kind regards
Uffe
