[PATCH v3 2/6] mux: Add helper functions for getting optional and selected mux-state

[Ulf Hansson]

On Sun, 21 Dec 2025 at 11:38, Josua Mayer <josua at solid-run.com> wrote:
>
> Hi Ulf,
>
> Am 17.12.25 um 14:38 schrieb Ulf Hansson:
> > On Wed, 10 Dec 2025 at 18:39, Josua Mayer <josua at solid-run.com> wrote:
>
> cut
>
> >>  /*
> >>   * Using subsys_initcall instead of module_init here to try to ensure - for
> >>   * the non-modular case - that the subsystem is initialized when mux consumers
> >> diff --git a/include/linux/mux/consumer.h b/include/linux/mux/consumer.h
> >> index 2e25c838f8312..a5da2e33a45c0 100644
> >> --- a/include/linux/mux/consumer.h
> >> +++ b/include/linux/mux/consumer.h
> >> @@ -60,5 +60,9 @@ struct mux_control *devm_mux_control_get(struct device *dev,
> >>                                          const char *mux_name);
> >>  struct mux_state *devm_mux_state_get(struct device *dev,
> >>                                      const char *mux_name);
> >> +struct mux_state *devm_mux_state_get_optional(struct device *dev,
> >> +                                             const char *mux_name);
> >> +struct mux_state *devm_mux_state_get_optional_selected(struct device *dev,
> >> +                                                      const char *mux_name);
> > Seems like we need stub-functions of these too. Otherwise
> > subsystems/drivers need to have a "depends on MULTIPLEXER" in their
> > Kconfigs.
>
> Currently the drivers that can use a mux select MULTIPLEXER in Kconfig.

Yes, but that's not generally how we do this. The driver may not need
MULTIPLEXER for all platforms that driver is being used on.

>
> There already exist a few mux helpers both for mux-state and for mux-control,
> and they might all need stubs.

Correct. I think we should add subs for all of them.

>
> I'd prefer the restructuring of kconfig dependencies being independent from
> adding mux-state functionality to renesas sdhi driver.

I understand your point, but adding the stubs isn't really a big thing
- unless someone has some good arguments not to!?

Moreover, since the series changes the mux-core anyways - and
subsequent changes depend on it, I don't see an issue to fold in yet
another patch to add the stubs.

Kind regards
Uffe