On Fri, May 24, 2019 at 01:04:01AM +0300, Vladimir Oltean wrote:
> On Fri, 24 May 2019 at 00:55, Russell King - ARM Linux admin
> <li...@armlinux.org.uk> wrote:
> >
> > On Thu, May 23, 2019 at 01:20:40AM +0000, Ioana Ciornei wrote:
> > > + if (pl->ops) {
> > > + pl->ops->mac_link_up(ndev, pl->link_an_mode,
> > > pl->phy_state.interface,
> > > pl->phydev);
> > >
> > > + netif_carrier_on(ndev);
> > >
> > > + netdev_info(ndev,
> > > + "Link is Up - %s/%s - flow control %s\n",
> > > + phy_speed_to_str(link_state.speed),
> > > + phy_duplex_to_str(link_state.duplex),
> > > + phylink_pause_to_str(link_state.pause));
> > > + } else {
> > > + blocking_notifier_call_chain(&pl->notifier_chain,
> > > + PHYLINK_MAC_LINK_UP, &info);
> > > + phydev_info(pl->phydev,
> > > + "Link is Up - %s/%s - flow control %s\n",
> > > + phy_speed_to_str(link_state.speed),
> > > + phy_duplex_to_str(link_state.duplex),
> > > + phylink_pause_to_str(link_state.pause));
> > > + }
> >
> > So if we don't have pl->ops, what happens when we call phydev_info()
> > with a NULL phydev, which is a very real possibility: one of phylink's
> > whole points is to support dynamic presence of a PHY.
> >
> > What will happen in that case is this will oops, due to dereferencing
> > an offset NULL pointer via:
> >
> > #define phydev_info(_phydev, format, args...) \
> > dev_info(&_phydev->mdio.dev, format, ##args)
> >
> > You can't just decide that if there's no netdev, we will be guaranteed
> > a phy.
> >
> > --
> > RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
> > FTTC broadband for 0.8mile line in suburbia: sync at 12.1Mbps down 622kbps
> > up
> > According to speedtest.net: 11.9Mbps down 500kbps up
>
> True, however it does not crash:
>
> [ 2.539949] (NULL device *): Link is Up - 1Gbps/Full - flow control off
>
> I agree that a better printing system has to be established though.
The only reason that happens is because struct mdio_device is at the
start of struct phy_device, and struct device is at the start of
struct mdio_device.
Should either of these move, that breaks and we get an oops. Sorry,
that's way too fragile.
Plus, of course, do we think that printing "(NULL device *):" is
really acceptable? We completely lose any information about _what_
link came up or went down.
--
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTC broadband for 0.8mile line in suburbia: sync at 12.1Mbps down 622kbps up
According to speedtest.net: 11.9Mbps down 500kbps up
