On Tue, Sep 15, 2020 at 08:31:55AM -0600, Rob Herring wrote:
> On Tue, Sep 15, 2020 at 12:24 AM Andreas Kemnade <andreas@kemnade.info> wrote:
> > On Mon, 14 Sep 2020 18:54:43 -0600 Rob Herring <robh@kernel.org> wrote:
[...]
> > > Just move this to the parent and make the parent a pwm provider. There's
> > > no need for child nodes for this or the rtc.
> > >
> > hmm, there are apparently devices without rtc. If there is a child node
> > for the rtc, the corresponding devicetrees could disable rtc by not
> > having that node.
> > But maybe using the controller version is also feasible for that task.
> 
> If not probeable, then the compatible string should distinguish that.

Okay.

It's even simpler in some cases: The firmware version reported by the EC
should tell us if it's one that is known to have no RTC.

That said, I don't have a good overview of the different variants of
this device.


Thanks,
Jonathan Neuschäfer