On Tue, 8 Jun 2021 at 16:20, Mark Brown [off-list ref] wrote:

On Tue, Jun 08, 2021 at 04:08:55PM +0200, Ulf Hansson wrote:

quoted

Honestly, I am not sure about what the regulator-fixed-domain intends
to model, but I assume it's something that fits well to be modelled as
a plain regulator, to start with.

quoted

Perhaps Mark can chime in and spread some light over this?

IIRC it's for situations where there's a device that's normally built as
a separate chip that got built into a bigger SoC and wants to rear end
something onto a power domain, I guess especially if the power domain
doesn't cover the whole of a Linux device.

Alright, thanks for explaining!

Certainly something that should not be mixed up with a regular power
rail for shared resources (aka power-domain). Maybe it's worth trying
to extend the description in the DT binding a bit, so it doesn't get
abused?

Kind regards
Uffe