Re: [PATCH net-next 2/2] net: dsa: mv88e6xxx: Set link down when changing speed

[Russell King - ARM Linux admin <linux@armlinux.org.uk>]

On Mon, Mar 23, 2020 at 11:39:34PM +0100, Andrew Lunn wrote:
> On Mon, Mar 23, 2020 at 10:01:13PM +0000, Russell King - ARM Linux admin wrote:
> > On Mon, Mar 23, 2020 at 10:49:00PM +0100, Andrew Lunn wrote:
> > > The MAC control register must not be changed unless the link is down.
> > > Add the necassary call into mv88e6xxx_mac_link_up. Without it, the MAC
> > > does not change state, the link remains at the wrong speed.
> > > 
> > > Fixes: 30c4a5b0aad8 ("net: mv88e6xxx: use resolved link config in mac_link_up()")
> > > Signed-off-by: Andrew Lunn <andrew@lunn.ch>
> > > ---
> > >  drivers/net/dsa/mv88e6xxx/chip.c | 8 ++++++++
> > >  1 file changed, 8 insertions(+)
> > > 
> > > diff --git a/drivers/net/dsa/mv88e6xxx/chip.c b/drivers/net/dsa/mv88e6xxx/chip.c
> > > index dd8a5666a584..24ce17503950 100644
> > > --- a/drivers/net/dsa/mv88e6xxx/chip.c
> > > +++ b/drivers/net/dsa/mv88e6xxx/chip.c
> > > @@ -733,6 +733,14 @@ static void mv88e6xxx_mac_link_up(struct dsa_switch *ds, int port,
> > >  
> > >  	mv88e6xxx_reg_lock(chip);
> > >  	if (!mv88e6xxx_port_ppu_updates(chip, port) || mode == MLO_AN_FIXED) {
> > > +		/* Port's MAC control must not be changed unless the
> > > +		 * link is down
> > > +		 */
> > > +		err = chip->info->ops->port_set_link(chip, port,
> > > +						     LINK_FORCED_DOWN);
> > > +		if (err)
> > > +			goto error;
> > > +
> > 
> > The port should be down at this point, otherwise the link state is not
> > matching phylink's idea of the state.  Your patch merely works around
> > that.  I think it needs solving properly.
> 
> Hi Russell
> 
> So the problem here is that CPU and DSA ports should default to up and
> at their fastest speed. During setup, the driver is setting the CPU
> port to 1G and up. Later on, phylink finds the fixed-link node in DT,
> and then sets the port to 100Mbps as requested.
> 
> How do you suggest fixing this? If we find a fixed-link, configure it
> first down and then up?

I think this is another example of DSA fighting phylink in terms of
what's expected.

The only suggestion I've come up so far with is to avoid calling
mv88e6xxx_port_setup_mac() with forced-link-up in
mv88e6xxx_setup_port() if we have phylink attached.

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line in suburbia: sync at 10.2Mbps down 587kbps up