Re: [PATCH] regulator: Provide dummy supply support

[Mark Brown <broonie@opensource.wolfsonmicro.com>]

On Tue, Nov 01, 2011 at 01:50:14PM -0400, Mike Frysinger wrote:
> On Friday 28 October 2011 18:47:57 Sascha Hauer wrote:

> > > We already have a dummy regulator driver and a fixed voltage regulator
> > > driver, we shouldn't be adding a third implementation of the same thing.
> > > Just use the fixed voltage regulator for this.

> > I explained in my mail why I think that the current implementation of
> > the dummy regulator is not suitable for things apart from debugging.

> your complaints seem to be specific to how the dummy regulator gets hooked in 
> and not in the specific regulator implementation.  so it seems like the right 
> thing would be to split the kconfig knobs:

Quite.  Sascha, your mail doesn't refer to the implementation of the
regulator itself at all.  Nothing in your changelog even mentions that
you're introducing a new regulator driver.

I think there's a big abstraction understanding failure here, reading
your changelog I'm not sure you understand the existing mechainsms that
are in place.  You say:

| This patch allows a board to register dummy supplies for devices
| which need a regulator but which is not software controllable
| on this board.

but this is exactly the use case the fixed voltage regulator is there
for.

>  config REGULATOR_DUMMY
> -	bool "Provide a dummy regulator if regulator lookups fail"
> +	bool "Provide a dummy regulator"
> +config REGULATOR_DUMMY_FALLBACK
> +	bool "Fallback to dummy regulator if lookup fails"
> +	depends on REGULATOR_DUMMY

As I think I said earlier I'd use the fixed regulator for this, all
Sascha's actually doing here is adding a wrapper to simplify
registration of that.