Re: [GIT PULL FOR v3.10] Camera sensors patches

[Mauro Carvalho Chehab <mchehab@redhat.com>]

Em Wed, 17 Apr 2013 14:55:03 +0100
Mark Brown <broonie@kernel.org> escreveu:

> On Tue, Apr 16, 2013 at 08:04:52PM +0200, Sylwester Nawrocki wrote:
> 
> > It's probably more clean to provide a dummy clock/regulator in a host driver
> > (platform) than to add something in a sub-device drivers that would resolve
> > which resources should be requested and which not.
> 
> Yes, that's the general theory for regulators at least - it allows the
> device driver to just trundle along and not worry about how the board is
> hooked up.  The other issue it resolves that you didn't mention is that
> it avoids just ignoring errors which isn't terribly clever.

I agree. Adding dummy clock/regulator at the host platform driver makes
sense, as the platform driver knows how the board is hooked up; keeping 
it at the I2C driver doesn't make sense, so the code needs to be moved
away from it.

Laurent,

Could you please work on a patch moving that code to the host platform
driver?

Thanks!
Mauro