Skip to contentSkip to conversation

bonding: fix netdev event NULL pointer dereference

PATCH Superseded Nikolay Aleksandrov 3 participants 5 messages Apr 9, 2013 – Apr 11, 2013

Subsystems (3)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/2 bonding: fix netdev event NULL pointer dereference

In commit 471cb5a33dcbd7c529684a2ac7ba4451414ee4a7 ("bonding: remove usage of dev->master") a bug was introduced which causes a NULL pointer dereference. If a bond device is in mode 6 (ALB) and a slave is added it will dereference a NULL pointer in bond_slave_netdev_event(). This is because in bond_enslave we have bond_alb_init_slave() which changes the MAC address of the slave and causes a NETDEV_CHANGEADDR. Then we have in bond_slave_netdev_event():

        struct slave *slave = bond_slave_get_rtnl(slave_dev);
        struct bonding *bond = slave->bond;

bond_slave_get_rtnl() dereferences slave_dev->rx_handler_data which at that time is NULL since netdev_rx_handler_register() is called later.

This is fixed by checking if slave is NULL before dereferencing it.

---
 drivers/net/bonding/bond_main.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)
drivers/net/bonding/bond_main.c +12 -2
index 171b10f..9995ddc 100644
@@ -3168,11 +3168,21 @@ static int bond_slave_netdev_event(unsigned long event,
 				   struct net_device *slave_dev)
 {
 	struct slave *slave = bond_slave_get_rtnl(slave_dev);
-	struct bonding *bond = slave->bond;
-	struct net_device *bond_dev = slave->bond->dev;
+	struct bonding *bond;
+	struct net_device *bond_dev;
 	u32 old_speed;
 	u8 old_duplex;
 
+	/*
+	 * A netdev event can be generated while enslaving a device
+	 * before netdev_rx_handler_register is called in which case
+	 * slave will be NULL
+	 */
+	if (!slave)
+		return NOTIFY_DONE;
+	bond_dev = slave->bond->dev;
+	bond = slave->bond;
+
 	switch (event) {
 	case NETDEV_UNREGISTER:
 		if (bond->setup_by_slave)
Signature · 1 line
1.8.1.4
#1 lore raw

PATCH 2/2 bonding: IFF_BONDING is not stripped on enslave failure

While enslaving a new device and after IFF_BONDING flag is set, in case of failure it is not stripped from the device's priv_flags while cleaning up, which could lead to other problems.
Cleaning at err_close because the flag is set after dev_open().

---
 drivers/net/bonding/bond_main.c | 1 +
 1 file changed, 1 insertion(+)
drivers/net/bonding/bond_main.c +1 -0
index 9995ddc..0d9ee48 100644
@@ -1906,6 +1906,7 @@ err_detach:
 	write_unlock_bh(&bond->lock);
 
 err_close:
+	slave_dev->priv_flags &= ~IFF_BONDING;
 	dev_close(slave_dev);
 
 err_unset_master:
Signature · 1 line
1.8.1.4
#2 lore raw
In reply to Nikolay Aleksandrov

Hello.

Quoted 7 lines · Nikolay Aleksandrov

In commit 471cb5a33dcbd7c529684a2ac7ba4451414ee4a7 ("bonding: remove usage of dev->master") a bug was introduced which causes a NULL pointer dereference. If a bond device is in mode 6 (ALB) and a slave is added it will dereference a NULL pointer in bond_slave_netdev_event(). This is because in bond_enslave we have bond_alb_init_slave() which changes the MAC address of the slave and causes a NETDEV_CHANGEADDR. Then we have in bond_slave_netdev_event():

Quoted 8 lines
         struct slave *slave = bond_slave_get_rtnl(slave_dev);
         struct bonding *bond = slave->bond;

bond_slave_get_rtnl() dereferences slave_dev->rx_handler_data which at that time is NULL since netdev_rx_handler_register() is called later.

This is fixed by checking if slave is NULL before dereferencing it.

[...]

Minor formatting nit.

Quoted 10 lines
drivers/net/bonding/bond_main.c +0 -2
index 171b10f..9995ddc 100644
@@ -3168,11 +3168,21 @@ static int bond_slave_netdev_event(unsigned long event,
  				   struct net_device *slave_dev)
  {
  	struct slave *slave = bond_slave_get_rtnl(slave_dev);
-	struct bonding *bond = slave->bond;
-	struct net_device *bond_dev = slave->bond->dev;

drivers/net/bonding/bond_main.c static int bond_slave_netdev_event(unsigned long event, · line 3180

Quoted 10 lines
drivers/net/bonding/bond_main.c +7 -0
+	struct bonding *bond;
+	struct net_device *bond_dev;
  	u32 old_speed;
  	u8 old_duplex;
 
+	/*
+	 * A netdev event can be generated while enslaving a device
+	 * before netdev_rx_handler_register is called in which case
+	 * slave will be NULL
+	 */

The preferred multi-line comment style in the networking code is:

/* bla
* bla
*/

WBR, Sergei

#3 lore raw

PATCH v2 1/2 bonding: fix netdev event NULL pointer dereference

In reply to Nikolay Aleksandrov

In commit 471cb5a33dcbd7c529684a2ac7ba4451414ee4a7 ("bonding: remove usage of dev->master") a bug was introduced which causes a NULL pointer dereference. If a bond device is in mode 6 (ALB) and a slave is added it will dereference a NULL pointer in bond_slave_netdev_event(). This is because in bond_enslave we have bond_alb_init_slave() which changes the MAC address of the slave and causes a NETDEV_CHANGEADDR. Then we have in bond_slave_netdev_event():

        struct slave *slave = bond_slave_get_rtnl(slave_dev);
        struct bonding *bond = slave->bond;

bond_slave_get_rtnl() dereferences slave_dev->rx_handler_data which at that time is NULL since netdev_rx_handler_register() is called later.

This is fixed by checking if slave is NULL before dereferencing it.

v2: Fix the comment styling

---
 drivers/net/bonding/bond_main.c | 13 +++++++++++--
 1 file changed, 11 insertions(+), 2 deletions(-)
drivers/net/bonding/bond_main.c +11 -2
index 171b10f..15c675c 100644
@@ -3168,11 +3168,20 @@ static int bond_slave_netdev_event(unsigned long event,
 				   struct net_device *slave_dev)
 {
 	struct slave *slave = bond_slave_get_rtnl(slave_dev);
-	struct bonding *bond = slave->bond;
-	struct net_device *bond_dev = slave->bond->dev;
+	struct bonding *bond;
+	struct net_device *bond_dev;
 	u32 old_speed;
 	u8 old_duplex;
 
+	/* A netdev event can be generated while enslaving a device
+	 * before netdev_rx_handler_register is called in which case
+	 * slave will be NULL
+	 */
+	if (!slave)
+		return NOTIFY_DONE;
+	bond_dev = slave->bond->dev;
+	bond = slave->bond;
+
 	switch (event) {
 	case NETDEV_UNREGISTER:
 		if (bond->setup_by_slave)
Signature · 1 line
1.8.1.4
maintainer
#4 lore raw

When you repost a patch that's part of a series, you must resubmit all patches in that series, not just the one that changed.

You should also provide a header posting "[PATCH 0/N] ..." that explains the high level purpose and intent of the patch set.