Arnd Bergmann wrote at Wednesday, August 17, 2011 5:24 AM:

On Tuesday 16 August 2011, Stephen Warren wrote:

quoted

One issue here: There isn't always a single gpio/pinmux parent; as a
concrete example, the ALSA/ASoC driver for Tegra+WM8903 uses GPIOs both
from Tegra itself, and from the WM8903 audio codec.

I could imagine the same being true in basically any case where one
device uses N GPIOs (e.g. SD controller with power, change-detect,
and read-only GPIOs; some could easily come from the SoC and some
from a GPIO expander).

I'm not quite so sure that multiple parents would be useful for pinmux,
but I wouldn't say that it was impossible...

Hmm, how common do you think the example is?

It it's just a very rare case, we can still special-case it by
introducing fake subdevices for the Tegra+WM8903 driver, each of
which has only one gpio-parent. This is ugly but keeps the special
case to the rare devices.

If we can expect more than a handful of these, we probably need
a better design for the binding, which will cause a burden for
all drivers.

It looks like DT already supports the generic case for GPIO; each GPIO
reference is of the form:

<&gpio 69 0>

and so each GPIO has a "parent" specified, and hence can come from a
different controller.

For drivers, this is all completely hidden by of_get_gpio().

I imagine we'd just use the same pattern for pinmux. This would both
support the general case, and end up being consistent with how GPIOs
already work, so this seems like a win both ways.

-- 
nvpublic