Re: [PATCH 1/2] dt-bindings: phy: ti,tcan104x-can: Document Microchip ATA6561

[Conor Dooley <conor@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 5420 bytes --]

On Tue, Jul 23, 2024 at 10:55:17PM +0300, Ilya Orazov wrote:
> On Tue, 23 Jul 2024 at 21:50, Conor Dooley <conor@kernel.org> wrote:
> >
> > On Tue, Jul 23, 2024 at 08:20:04PM +0300, IlorDash wrote:
> > > On Fri, 19 Jul 2024 at 18:07, Conor Dooley <conor@kernel.org> wrote:
> > > >
> > > > On Fri, Jul 19, 2024 at 12:03:21AM +0300, Ilya Orazov wrote:
> > > > > Microchip ATA6561 is High-Speed CAN Transceiver with Standby Mode.
> > > > > It is pin-compatible with TI TCAN1042.
> > > > >
> > > > > Signed-off-by: Ilya Orazov <ilordash02@gmail.com>
> > > > > ---
> > > > >  Documentation/devicetree/bindings/phy/ti,tcan104x-can.yaml | 1 +
> > > > >  1 file changed, 1 insertion(+)
> > > > >
> > > > > diff --git a/Documentation/devicetree/bindings/phy/ti,tcan104x-can.yaml b/Documentation/devicetree/bindings/phy/ti,tcan104x-can.yaml
> > > > > index 79dad3e89aa6..03de361849d2 100644
> > > > > --- a/Documentation/devicetree/bindings/phy/ti,tcan104x-can.yaml
> > > > > +++ b/Documentation/devicetree/bindings/phy/ti,tcan104x-can.yaml
> > > > > @@ -18,6 +18,7 @@ properties:
> > > > >        - nxp,tjr1443
> > > > >        - ti,tcan1042
> > > > >        - ti,tcan1043
> > > > > +      - microchip,ata6561
> > > >
> > > > Given that your driver patch has
> > > > | diff --git a/drivers/phy/phy-can-transceiver.c b/drivers/phy/phy-can-transceiver.c
> > > > | index ee4ce4249698..dbcd99213ba1 100644
> > > > | --- a/drivers/phy/phy-can-transceiver.c
> > > > | +++ b/drivers/phy/phy-can-transceiver.c
> > > > | @@ -89,6 +89,10 @@ static const struct of_device_id can_transceiver_phy_ids[] = {
> > > > |                 .compatible = "nxp,tjr1443",
> > > > |                 .data = &tcan1043_drvdata
> > > > |         },
> > > > | +       {
> > > > | +               .compatible = "microchip,ata6561",
> > > > | +               .data = &tcan1042_drvdata
> > > > | +       },
> > > > |         { }
> > > > |  };
> > > >
> > > > the driver patch is actually not needed at all, and you just need to
> > > > allow ti,tcan1042 as fallback compatible in the binding, so something
> > > > like:
> > > >
> > > >   compatible:
> > > >     oneOf:
> > > >       - enum:
> > > >           - nxp,tjr1443
> > > >           - ti,tcan1042
> > > >           - ti,tcan1043
> > > >       - items:
> > > >           - const: microchip,ata6561
> > > >           - const: ti,tcan1042
> > > >
> > > >    '#phy-cells':
> > > >      const: 0
> > >
> > > I tested the build with fallback compatible:
> > >
> > > compatible:
> > >   oneOf:
> > >     - items:
> > >       - enum:
> > >         - microchip,ata6561
> > >       - const: ti,tcan1042
> > >     - items:
> > >       - enum:
> > >         - nxp,tjr1443
> > >       - const: ti,tcan1043
> > >
> > > and modified compatible property in DTS:
> > >
> > > compatible = "microchip,ata6561", "ti,tcan1042";
> > >
> > > Build succeeded, phy-can-transceiver driver was used. So I would like
> > > to add a fallback compatible for both "microchip,ata6561" and
> > > "nxp,tjr1443" in this binding and modify other DTS files with
> > > compatible = "nxp,tjr1443". What do you think?
> >
> > This is wrong on two counts. Firstly, were what you have correct, you
> > should
> > squash the two:
> >      - items:
> >          - enum:
> >            - nxp,tjr1443
> >            - microchip,ata6561
> >          - const: ti,tcan1042
> >
> > However, that does not allow the TI compatibles in isolation, so you
> > still need to allow that for the actual TI devices, so you need:
> >
> >    oneOf:
> >      - items:
> >          - enum:
> >            - microchip,ata6561
> >            - nxp,tjr1443
> >            - ti,tcan1043
> >          - const: ti,tcan1042
> >      - const: ti,tcan1042
> >
> > There's probably some devicetrees that would need to be fixed up. I'm
> > just not convinced that this is worth retrofitting however.
> 
> But nxp,tjr1443 is pin compatible with ti,tcan1043, so it should
> fallback only to ti,tcan1043 and not ti,tcan1042. That's why I decided
> to split them into different enums.

Ah, sorry I missed that. I misread the match data. Then you need:
  compatible:
    oneOf:
      - items:
        - enum:
          - microchip,ata6561
        - const: ti,tcan1042
      - items:
        - enum:
          - nxp,tjr1443
        - const: ti,tcan1043
      - enum:
          const: ti,tcan1042
          const: ti,tcan1043

because the TI devices exist and we still need to be able to
differentiate the TI and NXP devices. If you have
  compatible = "nxp,tjr1443", "ti,tcan1042";
that means the device is an nxp,tjr1443. If you have
  compatible = "ti,tcan1042";
then that's a tcan1042.

> I made my patch according to a similar one that adds support for
> nxp,tjr1443. You can find it's conversation on
> https://lore.kernel.org/all/6ee5e2ce00019bd3f77d6a702b38bab1a45f3bb0.1674037830.git.geert+renesas@glider.be/t/#u.

> I thought we want to hold all PHY chip names in one compatible enum
> and each in its own of_device_id struct in driver and extend them
> where appropriate.

Nah, fallbacks are preferred when the programming model is either
identical or a "compatible superset" of an existing device. New
of_device_id structs should only be used where we need to account for
differences in the programming model.

Cheers,
Conor.

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]