On Tue, Aug 26, 2014 at 11:08:07AM +0200, Javier Martinez Canillas wrote:
> On 08/26/2014 09:17 AM, Mark Brown wrote:

> > No, this doesn't make any obvious sense to me at all.  Picking normal as
> > a default if the hardware reads back off due to overlapping
> > impelementation or something *might* make sense but not overwriting the
> > hardware state without explicit permission from the machine integration
> > is a key goal for the regulator API.

> Just to be sure I understood you correctly, what might makes sense to you
> then is to set the opmode to normal as default on probe only if off is
> read back from the hardware register but leaving the enable function as it
> is now using the opmode set on probe?

Yes.