The patch below does not apply to the 4.4-stable tree. If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to stable@vger.kernel.org.
thanks,
greg k-h
------------------ original commit in Linus's tree ------------------
From 610f8c0fc8d46e0933955ce13af3d64484a4630a Mon Sep 17 00:00:00 2001
From: Hristo Venev hristo@venev.name Date: Mon, 12 Apr 2021 20:41:16 +0300 Subject: [PATCH] net: sit: Unregister catch-all devices
A sit interface created without a local or a remote address is linked into the `sit_net::tunnels_wc` list of its original namespace. When deleting a network namespace, delete the devices that have been moved.
The following script triggers a null pointer dereference if devices linked in a deleted `sit_net` remain:
for i in `seq 1 30`; do ip netns add ns-test ip netns exec ns-test ip link add dev veth0 type veth peer veth1 ip netns exec ns-test ip link add dev sit$i type sit dev veth0 ip netns exec ns-test ip link set dev sit$i netns $$ ip netns del ns-test done for i in `seq 1 30`; do ip link del dev sit$i done
Fixes: 5e6700b3bf98f ("sit: add support of x-netns") Signed-off-by: Hristo Venev hristo@venev.name Signed-off-by: David S. Miller davem@davemloft.net
diff --git a/net/ipv6/sit.c b/net/ipv6/sit.c index 63ccd9f2dccc..9fdccf0718b5 100644 --- a/net/ipv6/sit.c +++ b/net/ipv6/sit.c @@ -1867,9 +1867,9 @@ static void __net_exit sit_destroy_tunnels(struct net *net, if (dev->rtnl_link_ops == &sit_link_ops) unregister_netdevice_queue(dev, head);
- for (prio = 1; prio < 4; prio++) { + for (prio = 0; prio < 4; prio++) { int h; - for (h = 0; h < IP6_SIT_HASH_SIZE; h++) { + for (h = 0; h < (prio ? IP6_SIT_HASH_SIZE : 1); h++) { struct ip_tunnel *t;
t = rtnl_dereference(sitn->tunnels[prio][h]);