Re: [PATCH net-next v2 2/2] net: pcs: rzn1-miic: Add support for PHY link active-level configuration

From: Lad, Prabhakar

Date: Mon Jan 12 2026 - 07:04:23 EST


Hi Russell,

Thank you for the review.

On Fri, Jan 9, 2026 at 2:46 PM Russell King (Oracle)
<linux@xxxxxxxxxxxxxxx> wrote:
>
> On Fri, Jan 09, 2026 at 02:22:50PM +0000, Prabhakar wrote:
> > From: Lad Prabhakar <prabhakar.mahadev-lad.rj@xxxxxxxxxxxxxx>
> >
> > Add support to configure the PHY link signal active level per converter
> > using the DT property "renesas,miic-phylink-active-low".
> >
> > Introduce the MIIC_PHYLINK register definition and extend the MIIC driver
> > with a new `phylink` structure to store the mask and value for PHY link
> > configuration. Implement `miic_configure_phylink()` to determine the bit
> > position and polarity for each port based on the SoC type, such as RZ/N1
> > or RZ/T2H/N2H.
> >
> > The accumulated configuration is stored during DT parsing and applied
> > later in `miic_probe()` after hardware initialization, since the MIIC
> > registers can only be modified safely once the hardware setup is complete.
>
> Please do not re-use "phylink", we have a subsystem in the kernel named
> as such, and, for example, it too defines "struct phylink".
>
> > +/**
> > + * struct phylink - Phylink configuration
> > + * @mask: Mask of phylink bits
> > + * @val: Value of phylink bits
> > + */
> > +struct phylink {
> > + u32 mask;
> > + u32 val;
> > +};
> > +
>
> You don't get a warning for this, because, although you have:
>
> #include <linux/phylink.h>
>
> which delares "struct phylink" as:
>
> struct phylink;
>
> The definition of this structure is entirely private to
> drivers/net/phy/phylink.c and is intentionally not exposed.
>
> By redefining "struct phylink" here, it means that anyone using gdb
> is going to run into problems - which version of this struct is the
> right one for any particular pointer.
>
Ack, I will rename the struct to `miic_phy_link_cfg`.

> You describe this feature as "PHY-link" and "PHY link" in your commit
> and cover messages. Please use "phy_link" and "PHY_LINK" as identifies
> for this so that grep can distinguish between your PHY link feature
> and the phylink infrastructure.
>
Ok, I will use "phy_link"/ "PHY_LINK".

Cheers,
Prabhakar