On Thu, 8 Jun 2023 13:34:43 -0600
Rob Herring <robh@xxxxxxxxxx> wrote:
> On Sat, May 20, 2023 at 04:17:53PM +0100, Jonathan Cameron wrote:
> > On Fri, 19 May 2023 19:29:15 +0100
> > Conor Dooley <conor@xxxxxxxxxx> wrote:
> >
> > > Hey Marius,
> > >
> > > On Fri, May 19, 2023 at 07:01:44PM +0300, marius.cristea@xxxxxxxxxxxxx wrote:
> > > > From: Marius Cristea <marius.cristea@xxxxxxxxxxxxx>
> > > >
> > > > This is the device tree schema for iio driver for
> > > > Microchip family of 153.6 ksps, Low-Noise 16/24-Bit
> > > > Delta-Sigma ADCs with an SPI interface.
> > >
> > > Just one quick process bit, please try to CC all of the maintainers
> > > listed by get_maintainer.pl - you unfortunately managed to miss 2 of the
> > > 3 dt-binding maintainers :/ Perhaps you ran get_maintainer.pl using our
> > > vendor tree?
> > >
> > > > Signed-off-by: Marius Cristea <marius.cristea@xxxxxxxxxxxxx>
> > > > ---
> > >
> > > > + vref-supply:
> > > > + description:
> > > > + Some devices have a specific reference voltage supplied on a different
> > > > + pin to the other supplies. Needed to be able to establish channel scaling
> > > > + unless there is also an internal reference available (e.g. mcp3564r)
> > >
> > > Should this be marked as a required property for the non-r devices that
> > > do not have an internal reference?
> > >
> > > > + microchip,hw-device-address:
> > >
> > > Hopefully Rob or Jonathan etc can chime in as to whether a common
> > > property exists for this type of thing...
> > >
> > Nope. This is a new one for me - there are devices that work on a daisy chain
> > principle but I think this one works by encoding stuff in the actual message
> > which is unusual for SPI.
>
> Not something I've seen either.
>
> >
> > > > + $ref: /schemas/types.yaml#/definitions/uint32
> > > > + minimum: 0
> > > > + maximum: 3
> > > > + description:
> > > > + The address is set on a per-device basis by fuses in the factory,
> > > > + configured on request. If not requested, the fuses are set for 0x1.
> > > > + The device address is part of the device markings to avoid
> > > > + potential confusion. This address is coded on two bits, so four possible
> > > > + addresses are available when multiple devices are present on the same
> > > > + SPI bus with only one Chip Select line for all devices.
>
> What's this going to look like with more than one device? It would need
> to be incorporated into 'reg' and the unit-address to work. Something
> like this is
>
> spi {
> device0@0 {
> reg = <0>;
> microchip,hw-device-address = <0>;
> };
>
> device1@0 {
> reg = <0>;
> microchip,hw-device-address = <1>;
> };
> };
>
> That should throw warnings because you have 2 nodes at the same address
> which is not good practice.
>
> I think you need a spi mux in here with the mux addresses being the
> microchip,hw-device-address values.
Something that looked like an spi-mux would be cute I'm not sure how
easy it would be to make it work given need to modify the messages
(rather sending extra ones before and after with a different chip select).
It would be nice if it were somewhat generic - so binding included which
bits would be replaced with the 'chip select' for the spi controller that
represents the mux.
Jonathan
>
> Rob
>
