]> git.proxmox.com Git - mirror_ubuntu-bionic-kernel.git/commitdiff
vxlan: test dev->flags & IFF_UP before calling gro_cells_receive()
authorEric Dumazet <edumazet@google.com>
Sun, 10 Mar 2019 17:36:40 +0000 (10:36 -0700)
committerKleber Sacilotto de Souza <kleber.souza@canonical.com>
Wed, 14 Aug 2019 09:18:49 +0000 (11:18 +0200)
BugLink: https://bugs.launchpad.net/bugs/1837952
[ Upstream commit 59cbf56fcd98ba2a715b6e97c4e43f773f956393 ]

Same reasons than the ones explained in commit 4179cb5a4c92
("vxlan: test dev->flags & IFF_UP before calling netif_rx()")

netif_rx() or gro_cells_receive() must be called under a strict contract.

At device dismantle phase, core networking clears IFF_UP
and flush_all_backlogs() is called after rcu grace period
to make sure no incoming packet might be in a cpu backlog
and still referencing the device.

A similar protocol is used for gro_cells infrastructure, as
gro_cells_destroy() will be called only after a full rcu
grace period is observed after IFF_UP has been cleared.

Most drivers call netif_rx() from their interrupt handler,
and since the interrupts are disabled at device dismantle,
netif_rx() does not have to check dev->flags & IFF_UP

Virtual drivers do not have this guarantee, and must
therefore make the check themselves.

Otherwise we risk use-after-free and/or crashes.

Fixes: d342894c5d2f ("vxlan: virtual extensible lan")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Andrea Righi <andrea.righi@canonical.com>
Signed-off-by: Khalid Elmously <khalid.elmously@canonical.com>
drivers/net/vxlan.c

index f09ced0bfebaf63b59c1fcce1bca5e03e745f5af..c26fecc933ca6463c17f03cdf9eb0b2d0ef8ce03 100644 (file)
@@ -1630,6 +1630,14 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
                goto drop;
        }
 
+       rcu_read_lock();
+
+       if (unlikely(!(vxlan->dev->flags & IFF_UP))) {
+               rcu_read_unlock();
+               atomic_long_inc(&vxlan->dev->rx_dropped);
+               goto drop;
+       }
+
        stats = this_cpu_ptr(vxlan->dev->tstats);
        u64_stats_update_begin(&stats->syncp);
        stats->rx_packets++;
@@ -1637,6 +1645,9 @@ static int vxlan_rcv(struct sock *sk, struct sk_buff *skb)
        u64_stats_update_end(&stats->syncp);
 
        gro_cells_receive(&vxlan->gro_cells, skb);
+
+       rcu_read_unlock();
+
        return 0;
 
 drop: