On Sun, 20 Dec 2020 18:08:19 +0100 Marcin Wojtas wrote:
> > > > > >> > > This patch fixes a regression that was introduced in v5.3:
> > > > > >> > > Commit 44cc27e43fa3 ("net: phylink: Add struct phylink_config to PHYLINK API")
> > > > > >> > >
> > > > > >> > > Above results in a NULL pointer dereference when booting the
> > > > > >> > > Armada7k8k/CN913x with ACPI between 5.3 and 5.8, which will be
> > > > > >> > > problematic especially for the distros using LTSv5.4 and above (the
> > > > > >> > > issue was reported on Fedora 32).
> > > > > >> > >
> > > > > >> > > Please help with backporting to the stable v5.3+ branches (it applies
> > > > > >> > > smoothly on v5.4/v5.6/v5.8).
> > > > > >> > >
> > > > > >> >
> > > > > >> > Any chances to backport this patch to relevant v5.3+ stable branches?
> > > > > >>
> > > > > >> What patch? What git commit id needs to be backported?
> > > > > >>
> > > > > >
> > > > > >The actual patch is:
> > > > > >Commit 6c2b49eb9671 ("net: mvpp2: add mvpp2_phylink_to_port() helper").
> > > > > >
> > > > > >URL for reference:
> > > > > >https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/drivers/net/ethernet/marvell/mvpp2?h=v5.10-rc7&id=6c2b49eb96716e91f202756bfbd3f5fea3b2b172
> > > > > >
> > > > > >Do you think it would be possible to get it merged to v5.3+ stable branches?
> > > > >
> > > > > Could you explain how that patch fixes anything? It reads like a
> > > > > cleanup.
> > > > >
> > > >
> > > > Indeed, I am aware of it, but I'm not sure about the best way to fix
> > > > it. In fact the mentioned patch is an unintentional fix. Commit
> > > > 44cc27e43fa3 ("net: phylink: Add struct phylink_config to PHYLINK
> > > > API") reworked an argument list of mvpp2_mac_config() routine in a way
> > > > that resulted in a NULL pointer dereference when booting the
> > > > Armada7k8k/CN913x with ACPI between 5.3 and 5.8. Part of Russell's
> > > > patch resolves this issue.
> > >
> > > What part fixes the issue? I can't see it...
> > >
> >
> > I re-checked in my setup and here's the smallest part of the original
> > patch, that fixes previously described issue:
> > diff --git a/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c
> > b/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c
> > index e98be8372780..9d71a4fe1750 100644
> > --- a/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c
> > +++ b/drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c
> > @@ -4767,6 +4767,11 @@ static void mvpp2_port_copy_mac_addr(struct
> > net_device *dev, struct mvpp2 *priv,
> > eth_hw_addr_random(dev);
> > }
> >
> > +static struct mvpp2_port *mvpp2_phylink_to_port(struct phylink_config *config)
> > +{
> > + return container_of(config, struct mvpp2_port, phylink_config);
> > +}
> > +
> > static void mvpp2_phylink_validate(struct phylink_config *config,
> > unsigned long *supported,
> > struct phylink_link_state *state)
> > @@ -5105,13 +5110,12 @@ static void mvpp2_gmac_config(struct
> > mvpp2_port *port, unsigned int mode,
> > static void mvpp2_mac_config(struct phylink_config *config, unsigned int mode,
> > const struct phylink_link_state *state)
> > {
> > - struct net_device *dev = to_net_dev(config->dev);
> > - struct mvpp2_port *port = netdev_priv(dev);
> > + struct mvpp2_port *port = mvpp2_phylink_to_port(config);
> > bool change_interface = port->phy_interface != state->interface;
> >
> > /* Check for invalid configuration */
> > if (mvpp2_is_xlg(state->interface) && port->gop_id != 0) {
> > - netdev_err(dev, "Invalid mode on %s\n", dev->name);
> > + netdev_err(port->dev, "Invalid mode on %s\n", port->dev->name);
> > return;
> > }
> >
> > @@ -5151,8 +5155,7 @@ static void mvpp2_mac_link_up(struct
> > phylink_config *config,
> > int speed, int duplex,
> > bool tx_pause, bool rx_pause)
> > {
> > - struct net_device *dev = to_net_dev(config->dev);
> > - struct mvpp2_port *port = netdev_priv(dev);
> > + struct mvpp2_port *port = mvpp2_phylink_to_port(config);
> > u32 val;
> >
> > if (mvpp2_is_xlg(interface)) {
> > @@ -5199,7 +5202,7 @@ static void mvpp2_mac_link_up(struct
> > phylink_config *config,
> >
> > mvpp2_egress_enable(port);
> > mvpp2_ingress_enable(port);
> > - netif_tx_wake_all_queues(dev);
> > + netif_tx_wake_all_queues(port->dev);
> > }
> >
> > static void mvpp2_mac_link_down(struct phylink_config *config,
> > --
> >
> > Do you think there is a point of slicing the original patch or rather
> > cherry-pick as-is?
> >
>
> Do you think there is a chance to get the above fix merged to stable (v5.3+)?
We need to work with stable maintainers on this, let's see..
Greg asked for a clear description of what happens, from your
previous response it sounds like a null-deref in mvpp2_mac_config().
Is the netdev -> config -> netdev linking not ready by the time
mvpp2_mac_config() is called?
