Re: [RFC PATCH 2/3] net: sparx5: Add Sparx5 switchdev driver

[Steen Hegelund]

On 28.11.2020 20:03, Andrew Lunn wrote:
> EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe
>
> > +static int sparx5_port_open(struct net_device *ndev)
> > +{
> > +     struct sparx5_port *port = netdev_priv(ndev);
> > +     int err = 0;
> > +
> > +     sparx5_port_enable(port, true);
> > +     if (port->conf.phy_mode != PHY_INTERFACE_MODE_NA) {
> > +             err = phylink_of_phy_connect(port->phylink, port->of_node, 0);
> > +             if (err) {
> > +                     netdev_err(ndev, "Could not attach to PHY\n");
> > +                     return err;
> > +             }
> > +     }
>
> This looks a bit odd. PHY_INTERFACE_MODE_NA means don't touch,
> something else has already configured the MAC-PHY mode in the PHY.
> You should not not connect the PHY because of this.

Hmm.  I will have to revisit this again.  The intent was to be able to
destinguish between regular PHYs and SFPs (as read from the DT).
But maybe the phylink_of_phy_connect function handles this
automatically...
> > +void sparx5_destroy_netdev(struct sparx5 *sparx5, struct sparx5_port *port)
> > +{
> > +     if (port->phylink) {
> > +             /* Disconnect the phy */
> > +             if (rtnl_trylock()) {
>
> Why do you use rtnl_trylock()?

The sparx5_port_stop() in turn calls phylink_stop() that expects the lock
to be taken.  Should I rather just call rtnl_lock()?

Thanks for your comments

/Steen


>    Andrew

BR
Steen

---------------------------------------
Steen Hegelund
steen.hegelund@xxxxxxxxxxxxx