vrf: Fix a potential NPD when removing a port from a VRF
Subsystems (3)
PATCH 1/1 vrf: Fix a potential NPD when removing a port from a VRF
RCU readers that identified a net device as a VRF port using netif_is_l3_slave() assume that a subsequent call to netdev_master_upper_dev_get_rcu() will return a VRF device. They then continue to dereference its l3mdev operations.
This assumption is not always correct and can result in a NPD [1]. There is no RCU synchronization when removing a port from a VRF, so it is possible for an RCU reader to see a new master device (e.g., a bridge) that does not have l3mdev operations.
Fix by adding RCU synchronization after clearing the IFF_L3MDEV_SLAVE flag. Skip this synchronization when a net device is removed from a VRF as part of its deletion and when the VRF device itself is deleted. In the latter case an RCU grace period will pass by the time RTNL is released.
[1]
BUG: kernel NULL pointer dereference, address: 0000000000000000 [...] RIP: 0010:l3mdev_fib_table_rcu (net/l3mdev/l3mdev.c:181) [...] Call Trace: <TASK> l3mdev_fib_table_by_index (net/l3mdev/l3mdev.c:201 net/l3mdev/l3mdev.c:189) __inet_bind (net/ipv4/af_inet.c:499 (discriminator 3)) inet_bind_sk (net/ipv4/af_inet.c:469) __sys_bind (./include/linux/file.h:62 (discriminator 1) ./include/linux/file.h:83 (discriminator 1) net/socket.c:1951 (discriminator 1)) __x64_sys_bind (net/socket.c:1969 (discriminator 1) net/socket.c:1967 (discriminator 1) net/socket.c:1967 (discriminator 1)) do_syscall_64 (arch/x86/entry/syscall_64.c:63 (discriminator 1) arch/x86/entry/syscall_64.c:94 (discriminator 1))
entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:130)
--- drivers/net/vrf.c | 15 +++++++++++---- 1 file changed, 11 insertions(+), 4 deletions(-)
drivers/net/vrf.c +11 -4
@@ -1034,6 +1034,7 @@ static int do_vrf_add_slave(struct net_device *dev, struct net_device *port_dev, err: port_dev->priv_flags &= ~IFF_L3MDEV_SLAVE ; +synchronize_net ();return ret; }
@@ -1053,10 +1054,16 @@ static int vrf_add_slave(struct net_device *dev, struct net_device *port_dev, }/* inverse of do_vrf_add_slave */ -static int do_vrf_del_slave (struct net_device *dev,struct net_device *port_dev) +static int do_vrf_del_slave (struct net_device *dev,struct net_device *port_dev, +bool needs_sync) {netdev_upper_dev_unlink (port_dev, dev); port_dev->priv_flags &= ~IFF_L3MDEV_SLAVE ; +/* Make sure that concurrent RCU readers that identified the device +* as a VRF port see a VRF master or no master at all. +*/ +if (needs_sync) +synchronize_net ();cycle_netdev (port_dev,NULL );
@@ -1065,7 +1072,7 @@ static int do_vrf_del_slave(struct net_device *dev, struct net_device *port_dev)static int vrf_del_slave (struct net_device *dev,struct net_device *port_dev) { -return do_vrf_del_slave (dev, port_dev); +return do_vrf_del_slave (dev, port_dev, true); }static void vrf_dev_uninit (struct net_device *dev)
@@ -1619,7 +1626,7 @@ static void vrf_dellink(struct net_device *dev, struct list_head *head)struct list_head *iter;netdev_for_each_lower_dev (dev, port_dev, iter) -vrf_del_slave (dev, port_dev); +do_vrf_del_slave (dev, port_dev, false);vrf_map_unregister_dev (dev);
@@ -1751,7 +1758,7 @@ static int vrf_device_event(struct notifier_block *unused, goto out; vrf_dev =netdev_master_upper_dev_get (dev); -vrf_del_slave (vrf_dev, dev); +do_vrf_del_slave (vrf_dev, dev, false); } out:return NOTIFY_DONE ;
Signature · 1 line
2.53.0
Quoted 38 lines · Ido Schimmel
RCU readers that identified a net device as a VRF port using
netif_is_l3_slave()assume that a subsequent call tonetdev_master_upper_dev_get_rcu()will return a VRF device. They then continue to dereference its l3mdev operations.This assumption is not always correct and can result in a NPD [1]. There is no RCU synchronization when removing a port from a VRF, so it is possible for an RCU reader to see a new master device (e.g., a bridge) that does not have l3mdev operations.
Fix by adding RCU synchronization after clearing the
IFF_L3MDEV_SLAVEflag. Skip this synchronization when a net device is removed from a VRF as part of its deletion and when the VRF device itself is deleted. In the latter case an RCU grace period will pass by the time RTNL is released.[1]
BUG: kernel NULL pointer dereference, address: 0000000000000000 [...] RIP: 0010:l3mdev_fib_table_rcu (net/l3mdev/l3mdev.c:181) [...] Call Trace: <TASK> l3mdev_fib_table_by_index (net/l3mdev/l3mdev.c:201 net/l3mdev/l3mdev.c:189) __inet_bind (net/ipv4/af_inet.c:499 (discriminator 3)) inet_bind_sk (net/ipv4/af_inet.c:469) __sys_bind (./include/linux/file.h:62 (discriminator 1) ./include/linux/file.h:83 (discriminator 1) net/socket.c:1951 (discriminator 1)) __x64_sys_bind (net/socket.c:1969 (discriminator 1) net/socket.c:1967 (discriminator 1) net/socket.c:1967 (discriminator 1)) do_syscall_64 (arch/x86/entry/syscall_64.c:63 (discriminator 1) arch/x86/entry/syscall_64.c:94 (discriminator 1))
entry_SYSCALL_64_after_hwframe(arch/x86/entry/entry_64.S:130)Fixes fdeea7be88b1 ("net: vrf: Set slave's private flag before linking")Reported-by Haoze Xie <[email protected]>Reported-by Yifan Wu <[email protected]>Reported-by Juefei Pu <[email protected]>Reported-by Yuan Tan <[email protected]>Signed-off-by Ido Schimmel <[email protected]>
Quoted 4 lines
--- drivers/net/vrf.c | 15 +++++++++++---- 1 file changed, 11 insertions(+), 4 deletions(-)
This patch was applied to netdev/net.git (main) by Jakub Kicinski <[email protected]>:
Hello:
This patch was applied to netdev/net.git (main)
by Jakub Kicinski <[email protected]>:
Quoted 11 lines · On Thu, 23 Apr 2026 09:36:07 +0300 you wrote:
RCU readers that identified a net device as a VRF port using
netif_is_l3_slave()assume that a subsequent call tonetdev_master_upper_dev_get_rcu()will return a VRF device. They then continue to dereference its l3mdev operations.This assumption is not always correct and can result in a NPD [1]. There is no RCU synchronization when removing a port from a VRF, so it is possible for an RCU reader to see a new master device (e.g., a bridge) that does not have l3mdev operations.
[...]
Here is the summary with links:
- [net] vrf: Fix a potential NPD when removing a port from a VRF
https://git.kernel.org/netdev/net/c/2674d603a9e6
You are awesome, thank you!
Signature · 2 lines
Deet-doot-dot, I am a bot. https://korg.docs.kernel.org/patchwork/pwbot.html