[ upstream commit 82dd0d2a9a76fc8fa2b18d80b987d455728bf83a ]
Miguel reported an skb use after free / double free in vrf_finish_output when neigh_output returns an error. The vrf driver should return after the call to neigh_output as it takes over the skb on error path as well.
Patch is a simplified version of Miguel's patch which was written for 4.9, and updated to top of tree.
Fixes: 8f58336d3f78a ("net: Add ethernet header for pass through VRF device") Signed-off-by: Miguel Fadon Perlines mfadon@teldat.com Signed-off-by: David Ahern dsahern@gmail.com Signed-off-by: David S. Miller davem@davemloft.net [ backport to 4.4 and 4.9 dropped the sock_confirm_neigh and changed neigh_output to dst_neigh_output ] --- note to stable: this patch applies to both 4.9 and 4.4 (the latter has an offset but still applies cleanly
drivers/net/vrf.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-)
diff --git a/drivers/net/vrf.c b/drivers/net/vrf.c index 346e48698555..42c9480acdc7 100644 --- a/drivers/net/vrf.c +++ b/drivers/net/vrf.c @@ -585,13 +585,15 @@ static int vrf_finish_output(struct net *net, struct sock *sk, struct sk_buff *s neigh = __ipv4_neigh_lookup_noref(dev, nexthop); if (unlikely(!neigh)) neigh = __neigh_create(&arp_tbl, &nexthop, dev, false); - if (!IS_ERR(neigh)) + if (!IS_ERR(neigh)) { ret = dst_neigh_output(dst, neigh, skb); + rcu_read_unlock_bh(); + return ret; + }
rcu_read_unlock_bh(); err: - if (unlikely(ret < 0)) - vrf_tx_error(skb->dev, skb); + vrf_tx_error(skb->dev, skb); return ret; }