Re: [PATCH net-next 6/7] net: txgbe: support copper NIC with external PHY

["Russell King (Oracle)" <linux@armlinux.org.uk>]

On Tue, Jul 25, 2023 at 10:41:46AM +0800, Jiawen Wu wrote:
> On Monday, July 24, 2023 6:43 PM, Russell King (Oracle) wrote:
> > On Mon, Jul 24, 2023 at 06:23:40PM +0800, Jiawen Wu wrote:
> > > @@ -22,6 +25,9 @@ static int txgbe_get_link_ksettings(struct net_device *netdev,
> > >  {
> > >  	struct txgbe *txgbe = netdev_to_txgbe(netdev);
> > >
> > > +	if (txgbe->wx->media_type == sp_media_copper)
> > > +		return phy_ethtool_get_link_ksettings(netdev, cmd);
> > 
> > Why? If a PHY is attached via phylink, then phylink will automatically
> > forward the call below to phylib.
> 
> No, there is no phylink implemented for sp_media_copper.
> 
> > > +
> > >  	return phylink_ethtool_ksettings_get(txgbe->phylink, cmd);
> > 
> > If you implement it correctly, you also don't need two entirely
> > separate paths to configure the MAC/PCS for the results of the PHY's
> > negotiation, because phylink gives you a _generic_ set of interfaces
> > between whatever is downstream from the MAC and the MAC.
> 
> For sp_media_copper, only mii bus is registered for attaching PHY.
> Most MAC/PCS configuration is done in firmware, so it is not necessary
> to implement phylink as sp_media_fiber.

If you do implement phylink for copper, then you don't need all these
conditionals and the additional adjust_link implementation. In other
words, you can re-use a lot of the code you've already added.

You don't have to provide a PCS to phylink provided you don't tell
phylink that it's "in-band".

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 80Mbps down 10Mbps up. Decent connectivity at last!