Re: [PATCH v2 1/6] dt-bindings: hwmon: add Microchip LAN969x fan controller
From: Rob Herring
Date: Wed Nov 11 2026 - 09:24:56 EST
On Wed, Nov 11, 2026 at 01:21:56PM +0100, Robert Marko wrote:
> On Tue, Nov 10, 2026 at 3:54 PM Rob Herring <robh@xxxxxxxxxx> wrote:
> >
> > On Mon, Nov 09, 2026 at 02:35:05PM +0100, Robert Marko wrote:
> > > The LAN969x fan controller has a PWM output and tachometer input in a
> > > register block separate from the Sparx5-compatible temperature sensor.
> > > Describe it with its own compatible and binding.
> > >
> > > The optional fan child specifies PWM period and polarity, tachometer
> > > pulses per revolution, and thermal cooling levels.
> > >
> > > Signed-off-by: Robert Marko <robert.marko@xxxxxxxxxx>
> > > ---
> > > Changes in v2:
> > > - Describe the fan controller in a separate binding and compatible.
> > > - Keep the existing temperature-sensor binding unchanged.
[...]
> > > + dependencies:
> > > + '#cooling-cells':
> > > + - cooling-levels
> > > + cooling-levels:
> > > + - '#cooling-cells'
> >
> > The dependencies should also go in fan-common.yaml. Perhaps
> > cooling-levels without #cooling-cells might make sense sometimes?
>
> I removed cooling-levels requiring #cooling-cells, as it can
> technically work without it, its not ideal, but it works.
>
> I am not sure if its feasible for fan-common.yaml though as we already
> have schemas for EMC2305, and more importantly
> DTS-es for freescale/imx95-tqma9596la-mba95xxca.dts and
> rockchip/rk3568-radxa-cm3j-rpi-cm4.dts
> that have #cooling-cells without cooling-levels.
> EMC2305 driver even calculates cooling states without a levels table.
>
> I would appreciate guidance on how to proceed with this.
Okay. Leave it as-is.
Rob