On Mon, Oct 12, 2015 at 01:42:59PM +0200, Sascha Hauer wrote: > On Fri, Oct 02, 2015 at 06:32:56PM +0100, Mark Brown wrote:
> > One change I think we need here is only doing the propagation if either > > the device lacks a set_voltage() operation (in which case it's just a > > switch passing through the parent voltage) > Does the lack of a set_voltage() operation automatically mean it's a > switch passing through the parent voltage? What if the regulator is a > fixed regulator and the output can't be controlled because there is only > one voltage? Sorry, that was a typo for get_voltage(). > Currently we bail out int regulator_set_voltage() when we do not have a > set_voltage() or set_voltage_sel() operation. Instead of propagating the > voltage change up I would keep the current behaviour and implement voltage > propagation for switches when we need it. Then we could also introduce a > REGULATOR_IS_SWITCH flag indicating that this is a switch and not a > fixed voltage regulator. We already have people who'd like it.
signature.asc
Description: Digital signature
