[PATCH net-next 06/12] ipvlan: Take addr_lock in ipvlan_open()
From: Dmitry Skorodumov <hidden>
Date: 2025-11-20 17:50:31
Also in:
lkml
Subsystem:
networking drivers, the rest · Maintainers:
Andrew Lunn, "David S. Miller", Eric Dumazet, Jakub Kicinski, Paolo Abeni, Linus Torvalds
It was forgotten to lock addrs in ipvlan_open(). Seems that code was initially written in assumption that any address change occurs under rtnl_lock(). But it's not true for the ipv6 case. So, we have to take addr_lock in ipvlan_open(). Signed-off-by: Dmitry Skorodumov <redacted> --- drivers/net/ipvlan/ipvlan_main.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/drivers/net/ipvlan/ipvlan_main.c b/drivers/net/ipvlan/ipvlan_main.c
index c1df97a88a40..27d289aadef1 100644
--- a/drivers/net/ipvlan/ipvlan_main.c
+++ b/drivers/net/ipvlan/ipvlan_main.c@@ -262,20 +262,20 @@ static void ipvlan_uninit(struct net_device *dev) static int ipvlan_open(struct net_device *dev) { struct ipvl_dev *ipvlan = netdev_priv(dev); + struct ipvl_port *port = ipvlan->port; struct ipvl_addr *addr; - if (ipvlan->port->mode == IPVLAN_MODE_L3 || - ipvlan->port->mode == IPVLAN_MODE_L3S) + if (port->mode == IPVLAN_MODE_L3 || port->mode == IPVLAN_MODE_L3S) dev->flags |= IFF_NOARP; else dev->flags &= ~IFF_NOARP; /* for learnable, addresses will be obtained from tx-packets. */ - if (!ipvlan_is_macnat(ipvlan->port)) { - rcu_read_lock(); + if (!ipvlan_is_macnat(port)) { + spin_lock_bh(&port->addrs_lock); list_for_each_entry_rcu(addr, &ipvlan->addrs, anode) ipvlan_ht_addr_add(ipvlan, addr); - rcu_read_unlock(); + spin_unlock_bh(&port->addrs_lock); } return 0;
--
2.25.1