On Sun, Jun 04, 2017 at 09:01:33PM +0200, Andrew Lunn wrote:
> > diff --git a/drivers/net/cris/eth_v10.c b/drivers/net/cris/eth_v10.c
> > index da02041..017f48c 100644
> > --- a/drivers/net/cris/eth_v10.c
> > +++ b/drivers/net/cris/eth_v10.c
> > @@ -1417,10 +1417,9 @@ static int e100_get_link_ksettings(struct net_device *dev,
> > {
> > struct net_local *np = netdev_priv(dev);
> > u32 supported;
> > - int err;
> >
> > spin_lock_irq(&np->lock);
> > - err = mii_ethtool_get_link_ksettings(&np->mii_if, cmd);
> > + mii_ethtool_get_link_ksettings(&np->mii_if, cmd);
> > spin_unlock_irq(&np->lock);
> >
> > /* The PHY may support 1000baseT, but the Etrax100 does not. */
> > @@ -1432,7 +1431,7 @@ static int e100_get_link_ksettings(struct net_device *dev,
> > ethtool_convert_legacy_u32_to_link_mode(cmd->link_modes.supported,
> > supported);
> >
> > - return err;
> > + return 0;
> > }
>
> How far are going planning on going? It seems like
> *_get_link_ksettings() now all return a useless 0. Do you plan to
> change ethtool_ops and make if void all the way up?
It is not always correct, see for example how xgene_get_link_ksettings
returns non-zero value so i assume that ethtool_ops should remain as it is.
Also, looking at ethtool_get_settings it seems that returned value is
checked.
>
> Andrew
