Re: [PATCH V3] dt-bindings: nvmem: convert base example to use "nvmem-layout" node

[Rafał Miłecki]

On 10.03.2023 09:59, Miquel Raynal wrote:
> Hi Rafał,
>
> zajec5@xxxxxxxxx wrote on Fri, 10 Mar 2023 08:51:45 +0100:
>
> > From: Rafał Miłecki <rafal@xxxxxxxxxx>
> >
> > With support for "fixed-layout" binding we can use now "nvmem-layout"
> > even for fixed NVMEM cells. Use that in the base example as it should be
> > preferred over placing cells directly in the device node.
> >
> > New and other bindings should follow as old binding will get deprecated
> > at some point.
> >
> > Signed-off-by: Rafał Miłecki <rafal@xxxxxxxxxx>
> > ---
> >   .../devicetree/bindings/nvmem/nvmem.yaml      | 42 +++++++++++--------
> >   1 file changed, 24 insertions(+), 18 deletions(-)
> >
> > diff --git a/Documentation/devicetree/bindings/nvmem/nvmem.yaml b/Documentation/devicetree/bindings/nvmem/nvmem.yaml
> > index 732162e9d13e..c77be1c20e47 100644
> > --- a/Documentation/devicetree/bindings/nvmem/nvmem.yaml
> > +++ b/Documentation/devicetree/bindings/nvmem/nvmem.yaml
> > @@ -67,24 +67,30 @@ examples:
> >   
> >             /* ... */
> >   
> > -          /* Data cells */
> > -          tsens_calibration: calib@404 {
> > -              reg = <0x404 0x10>;
> > -          };
> > -
> > -          tsens_calibration_bckp: calib_bckp@504 {
> > -              reg = <0x504 0x11>;
> > -              bits = <6 128>;
> > -          };
> > -
> > -          pvs_version: pvs-version@6 {
> > -              reg = <0x6 0x2>;
> > -              bits = <7 2>;
> > -          };
> > -
> > -          speed_bin: speed-bin@c{
> > -              reg = <0xc 0x1>;
> > -              bits = <2 3>;
> > +          nvmem-layout {
>
> I believe we should not introduce "intermediate state" bindings when
> this is not strictly required, in order to avoid confusion with what is
> backward and what is transitory. So I would expect this to only apply
> after the switch to:
>
> 	nvmem-layout@xxx {
> 		reg = <xxx>;
>
> I don't care who will take care of it, but I think it would be better
> to have everything in one series.
>
> Other than the "order" problematic which I think is important here, I
> agree with this series.

I fail to see how / why:
1. Adding new NVMEM layout
2. Supporting mutliple NVMEM layouts
would depend on each other.

We already have 2 NVMEM layouts bindings. I'm just adding a new (third)
one.

If having NVMEM layouts bindings puts us in any kind of intermediate
state then we're already there. I don't think my new NVMEM layout
changes this situation.


> > +              compatible = "fixed-layout";
> > +              #address-cells = <1>;
> > +              #size-cells = <1>;
> > +
> > +              /* Data cells */
> > +              tsens_calibration: calib@404 {
> > +                  reg = <0x404 0x10>;
> > +              };
> > +
> > +              tsens_calibration_bckp: calib_bckp@504 {
> > +                  reg = <0x504 0x11>;
> > +                  bits = <6 128>;
> > +              };
> > +
> > +              pvs_version: pvs-version@6 {
> > +                  reg = <0x6 0x2>;
> > +                  bits = <7 2>;
> > +              };
> > +
> > +              speed_bin: speed-bin@c{
> > +                  reg = <0xc 0x1>;
> > +                  bits = <2 3>;
> > +              };
> >             };
> >         };