Hit send too soon...
On Tue, Aug 21, 2018 at 8:11 AM Rob Herring <robh@xxxxxxxxxx> wrote:
>
> On Tue, Aug 21, 2018 at 4:28 AM Vinod <vkoul@xxxxxxxxxx> wrote:
> >
> > On 20-08-18, 13:18, Rob Herring wrote:
> > > On Mon, Aug 20, 2018 at 12:09:52PM +0530, Vinod Koul wrote:
> >
> > > > +PROPERTIES:
> > > > +
> > > > +- compatible:
> > > > + Usage: required
> > > > + Value type: <string>
> > > > + Definition: must be one of:
> > > > + "qcom,msm-8916-cci"
> > > > + "qcom,msm-8996-cci"
> > >
> > > I think everywhere else is 'msm8916' and 'msm8996'.
> >
> > Quick grep told me that is the case, so will update.
> >
> > > > +SUBNODES:
> > > > +
> > > > +The CCI provides I2C masters for one or two i2c busses, described as
> > > > +subdevices named "i2c-bus0" and "i2c-bus1".
> > >
> > > Use a unit-address and reg property with 0 and 1 here.
> >
> > Am not sure I understood that properly, still learning DT nuisances,
> > care to elaborate a bit please.
>
> Node names are supposed to be standard (there's a list in the DT spec)
> and i2c-bus is for cases where the controller is not the bus parent.
> So you just need it to look like this:
>
> i2c-bus@0 {
> reg = <0>;
> ...
> };
>
> i2c-bus@1 {
> reg = <1>;
> ...
> };
>
> It's similar to how i2c muxes are done where you have multiple
> downstream i2c buses. Following this will enable some i2c bus checks
> in dtc (current master, not kernel copy yet) as node names are the
> only way
...we can match i2c buses in a generic way.
Rob
