mlxsw: spectrum_switchdev: Fix VLAN device deletion via ioctl

When deleting a VLAN device using an ioctl the netdev is unregistered
before the VLAN filter is updated via ndo_vlan_rx_kill_vid(). It can
lead to a use-after-free in mlxsw in case the VLAN device is deleted
while being enslaved to a bridge.

The reason for the above is that when mlxsw receives the CHANGEUPPER
event, it wrongly assumes that the VLAN device is no longer its upper
and thus destroys the internal representation of the bridge port despite
the reference count being non-zero.

Fix this by checking if the VLAN device is our upper using its real
device. In net-next I'm going to remove this trick and instead make
mlxsw completely agnostic to the order of the events.

Fixes: c57529e ("mlxsw: spectrum: Replace vPorts with Port-VLAN")
Signed-off-by: Ido Schimmel <[email protected]>
Reviewed-by: Petr Machata <[email protected]>
Signed-off-by: David S. Miller <[email protected]>

Author: idosch

drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c

@@ -296,7 +296,13 @@ static bool
 mlxsw_sp_bridge_port_should_destroy(const struct mlxsw_sp_bridge_port *
 				    bridge_port)
 {
-	struct mlxsw_sp *mlxsw_sp = mlxsw_sp_lower_get(bridge_port->dev);
+	struct net_device *dev = bridge_port->dev;
+	struct mlxsw_sp *mlxsw_sp;
+
+	if (is_vlan_dev(dev))
+		mlxsw_sp = mlxsw_sp_lower_get(vlan_dev_real_dev(dev));
+	else
+		mlxsw_sp = mlxsw_sp_lower_get(dev);
 
 	/* In case ports were pulled from out of a bridged LAG, then
 	 * it's possible the reference count isn't zero, yet the bridge
@@ -2109,7 +2115,7 @@ mlxsw_sp_bridge_8021d_port_leave(struct mlxsw_sp_bridge_device *bridge_device,
 
 	vid = is_vlan_dev(dev) ? vlan_dev_vlan_id(dev) : 1;
 	mlxsw_sp_port_vlan = mlxsw_sp_port_vlan_find_by_vid(mlxsw_sp_port, vid);
-	if (WARN_ON(!mlxsw_sp_port_vlan))
+	if (!mlxsw_sp_port_vlan)
 		return;
 
 	mlxsw_sp_port_vlan_bridge_leave(mlxsw_sp_port_vlan);