On Mon, 17 Jun 2019 18:02:55 +0100
Mark Brown <broonie@kernel.org> wrote:

> On Mon, Jun 17, 2019 at 06:27:43PM +0200, Andreas Kemnade wrote:
> 
> > Citing tps65950 trm page 55:  
> 
> > The device contains three switch-mode power supplies (SMPS):
> > • VDD1: 1.2-A, buck DC/DC converter (VOUT = 0.6 V to 1.45 V, in steps of 12.5 mV)
> > • VDD2: 600-mA buck DC/DC converter (VOUT = 0.6 V to 1.45 V, in steps of 12.5 mV, and 1.5 V as a
> >    single programmable value)  
> 
> > you are right, they are not really continuous. So should I add these
> > 68 steps they have as a voltage list?  
> 
> There's helpers for linear mappings, you should be able to use those
> (see helpers.c).
> 
ok, I will send a 2 with such a list.

Thanks for the hint.

> > I think they are nearly continuous, so we should IMHO rather take that
> > not that strict. I guess there are no really continuous regulators, all
> > have steps as voltage is specified in a limited resolution. So what is
> > the exact meaning of that flag here?  
> 
> This was added for devices with extremely high resolution interfaces
> like some microcontroller interfaces that take voltage values directly
> (mirroring the regulator API) or PWM regulators - it's for cases where
> enumerating all the voltages is unreasonable.  The TWL4030 regulators
> look fairly standard in comparison.

well, VDD1 is a lot more continuous than e.g. VAUX3, but your examples
seem to be even more continuous. But maybe a comment in the api documentation
might be helpful so that people do not misinterpret the meaning.

Regards,
Andreas