Re: [net-next: PATCH v4 2/8] net: mdio: switch fixed-link PHYs API to fwnode_

[Marcin Wojtas]

wt., 17 sty 2023 o 18:58 Russell King (Oracle) <linux@xxxxxxxxxxxxxxx>
napisał(a):
>
> Hi Marcin,
>
> On Tue, Jan 17, 2023 at 05:20:01PM +0100, Marcin Wojtas wrote:
> > Hi Russell,
> >
> >
> > pon., 16 sty 2023 o 18:50 Russell King (Oracle) <linux@xxxxxxxxxxxxxxx>
> > napisał(a):
> > >
> > > On Mon, Jan 16, 2023 at 06:34:14PM +0100, Marcin Wojtas wrote:
> > > > fixed-link PHYs API is used by DSA and a number of drivers
> > > > and was depending on of_. Switch to fwnode_ so to make it
> > > > hardware description agnostic and allow to be used in ACPI
> > > > world as well.
> > >
> > > Would it be better to let the fixed-link PHY die, and have everyone use
> > > the more flexible fixed link implementation in phylink?
> > >
> > ,
> > This patchset did not intend to introduce any functional change, simply
> > switch to a more generic HW description abstraction. Killing
> > of/fwnode_phy_(de)register_fixed_link entirely seems to be a challenge, as
> > there are a lot of users beyond the DSA. Otoh I see a value in having
> > of_/fwnode_phy_is_fixed_link check, afaik there is no equivalent in
> > phylink...
>
> Phylink provides a much improved implementation of fixed-link that is
> way more flexible than the phylib approach - it can implement speeds
> in excess of 1G. DSA already supports phylink with modern updated
> drivers that do not use the "adjust_link" implementation.
>
> What I'm proposing is that we don't bring the baggage of the phylib
> based fixed link forwards into fwnode, and leave this to be DT-only.
> I think this is what Andrew and Vladimir have also said.
>

Ok, thanks for clarifying.

Best regards,
Marcin