Re: [PATCH v10 3/4] dt-bindings: ethernet-phy: add optional brr-mode flag

[Conor Dooley]

On Thu, Jul 04, 2024 at 11:11:41PM +0200, Andrew Lunn wrote:
> > > ---
> > >  Documentation/devicetree/bindings/net/ethernet-phy.yaml | 8 ++++++++
> > >  1 file changed, 8 insertions(+)
> > > 
> > > diff --git a/Documentation/devicetree/bindings/net/ethernet-phy.yaml b/Documentation/devicetree/bindings/net/ethernet-phy.yaml
> > > index 8fb2a6ee7e5b..349ae72ebf42 100644
> > > --- a/Documentation/devicetree/bindings/net/ethernet-phy.yaml
> > > +++ b/Documentation/devicetree/bindings/net/ethernet-phy.yaml
> > > @@ -93,6 +93,14 @@ properties:
> > >        the turn around line low at end of the control phase of the
> > >        MDIO transaction.
> > >  
> > > +  brr-mode:
> > > +    $ref: /schemas/types.yaml#/definitions/flag
> > > +    description:
> > > +      If set, indicates the network cable interface is alternative one as

nit: an alternative

> > > +      defined in the BroadR-Reach link mode specification under 1BR-100 and
> > > +      1BR-10 names. The driver needs to configure the PHY to operate in
> > > +      BroadR-Reach mode.
> > 
> > I find this second sentence unclear. Does the driver need to do
> > configure the phy because this mode is not enabled by default or because
> > the device will not work outside of this mode.
> 
> BroadR-Reach uses one pair. Standard 10/100Mbps ethernet needs 2 pair,
> and 1G needs 4 pair. So any attempt to use standard 10/100/1G is going
> to fail, the socket/plug and cable just don't work for anything other
> than BroadR-Reach.
> 
> As far as i understand it, the PHY could be strapped into BroadR-Reach
> mode, but there is no guarantee it is, so we should also program it to
> BroadR-Reach mode if this property is present.

Right, I guess for people that want brr-mode they'll probably understand
that and it's just unclear to me as I don't :) The nitpicking side of my
brain would like that last sentence reworded to "The PHY must be
configured to operate in BroadR-Reach mode by software", mostly cos the
word "driver" assumes a particular model for the OS, but it's not really
a big deal.

Acked-by: Conor Dooley <conor.dooley@xxxxxxxxxxxxx>

Attachment: signature.asc
Description: PGP signature