[PATCH net-next v3 1/4] dt-bindings: net: ftgmac100: Add delay properties for AST2600
Andrew Lunn
andrew at lunn.ch
Wed Nov 5 00:40:28 AEDT 2025
On Tue, Nov 04, 2025 at 05:14:41AM +0000, Jacky Chou wrote:
> Hi Andrew,
>
> Thank you for your reply.
>
> > > + - if:
> > > + properties:
> > > + compatible:
> > > + contains:
> > > + const: aspeed,ast2600-mac01
> > > + then:
> > > + properties:
> > > + rx-internal-delay-ps:
> > > + minimum: 0
> > > + maximum: 1395
> > > + multipleOf: 45
> >
> > I would add a default: 0
> >
>
> Agreed.
> I will add it in next version.
>
> > > + tx-internal-delay-ps:
> > > + minimum: 0
> > > + maximum: 1395
> > > + multipleOf: 45
> >
> > and also here.
> >
>
> Agreed.
>
> > > + required:
> > > + - scu
> > > + - rx-internal-delay-ps
> > > + - tx-internal-delay-ps
> >
> > and then these are not required, but optional.
> >
>
> Configure the tx/rx delay in the scu register.
> At least, the scu handle must be required.
Sorry, i was unclear. By says 'and then', i was trying to chain it to
the previous comment, that the delays should default to 0. With
defaults set, rx-internal-delay-ps and tx-internal-delay-ps become
optional. I agree scu is required.
> Here I have one question.
> In v3 patches series, if the ftgmac driver cannot find one of
> tx-internal-delay-ps and rx-internal-delay-ps, it will return error in probe
> stage.
> If here is optional, does it means that just add warning and not return error when
> lack one of them and use the default to configure? Or not configure tx/rx delay just
> return success in probe stage?
Once you add the default statement, it is clear what delay should be
added if they property is not listed, 0. No warning is needed.
What you should find in the end is that most boards just set the new
compatible and 'rgmii-id', and need nothing else. Only badly designed
boards tend to need tx-internal-delay-ps and rx-internal-delay-ps.
Andrew
More information about the Linux-aspeed
mailing list