switchdev: Require RTNL mutex to be held when sending FDB notifications

When switchdev drivers process FDB notifications from the underlying
device they resolve the netdev to which the entry points to and notify
the bridge using the switchdev notifier.

However, since the RTNL mutex is not held there is nothing preventing
the netdev from disappearing in the middle, which will cause
br_switchdev_event() to dereference a non-existing netdev.

Make switchdev drivers hold the lock at the beginning of the
notification processing session and release it once it ends, after
notifying the bridge.

Also, remove switchdev_mutex and fdb_lock, as they are no longer needed
when RTNL mutex is held.

Fixes: 03bf0c281234 ("switchdev: introduce switchdev notifier")
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 3 insertions, 5 deletions

diff --git a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
index f2ab7cd09cf3..e492ca2cdecd 100644
--- a/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
+++ b/drivers/net/ethernet/mellanox/mlxsw/spectrum_switchdev.c
@@ -45,6 +45,7 @@
 #include <linux/if_bridge.h>
 #include <linux/workqueue.h>
 #include <linux/jiffies.h>
+#include <linux/rtnetlink.h>
 #include <net/switchdev.h>
 
 #include "spectrum.h"
@@ -1068,7 +1069,6 @@ static int mlxsw_sp_port_fdb_dump(struct mlxsw_sp_port *mlxsw_sp_port,
 	if (!sfd_pl)
 		return -ENOMEM;
 
-	mutex_lock(&mlxsw_sp_port->mlxsw_sp->fdb_lock);
 	if (mlxsw_sp_port_is_vport(mlxsw_sp_port)) {
 		u16 tmp;
 
@@ -1142,7 +1142,6 @@ static int mlxsw_sp_port_fdb_dump(struct mlxsw_sp_port *mlxsw_sp_port,
 	} while (num_rec == MLXSW_REG_SFD_REC_MAX_COUNT);
 
 out:
-	mutex_unlock(&mlxsw_sp_port->mlxsw_sp->fdb_lock);
 	kfree(sfd_pl);
 	return stored_err ? stored_err : err;
 }
@@ -1393,7 +1392,7 @@ static void mlxsw_sp_fdb_notify_work(struct work_struct *work)
 
 	mlxsw_sp = container_of(work, struct mlxsw_sp, fdb_notify.dw.work);
 
-	mutex_lock(&mlxsw_sp->fdb_lock);
+	rtnl_lock();
 	do {
 		mlxsw_reg_sfn_pack(sfn_pl);
 		err = mlxsw_reg_query(mlxsw_sp->core, MLXSW_REG(sfn), sfn_pl);
@@ -1406,7 +1405,7 @@ static void mlxsw_sp_fdb_notify_work(struct work_struct *work)
 			mlxsw_sp_fdb_notify_rec_process(mlxsw_sp, sfn_pl, i);
 
 	} while (num_rec);
-	mutex_unlock(&mlxsw_sp->fdb_lock);
+	rtnl_unlock();
 
 	kfree(sfn_pl);
 	mlxsw_sp_fdb_notify_work_schedule(mlxsw_sp);
@@ -1421,7 +1420,6 @@ static int mlxsw_sp_fdb_init(struct mlxsw_sp *mlxsw_sp)
 		dev_err(mlxsw_sp->bus_info->dev, "Failed to set default ageing time\n");
 		return err;
 	}
-	mutex_init(&mlxsw_sp->fdb_lock);
 	INIT_DELAYED_WORK(&mlxsw_sp->fdb_notify.dw, mlxsw_sp_fdb_notify_work);
 	mlxsw_sp->fdb_notify.interval = MLXSW_SP_DEFAULT_LEARNING_INTERVAL;
 	mlxsw_sp_fdb_notify_work_schedule(mlxsw_sp);