[PATCH net] dccp/tcp: fix routing redirect race

From: Jon Maxwell
Date: Wed Mar 08 2017 - 20:16:19 EST

We have seen a few incidents lately where a dst_enty has been freed
with a dangling TCP socket reference (sk->sk_dst_cache) pointing to that
dst_entry. If the conditions/timings are right a crash then ensues when the
freed dst_entry is referenced later on. A Common crashing back trace is:

#8 [] page_fault at ffffffff8163e648
[exception RIP: __tcp_ack_snd_check+74]
#9 [] tcp_rcv_established at ffffffff81580b64
#10 [] tcp_v4_do_rcv at ffffffff8158b54a
#11 [] tcp_v4_rcv at ffffffff8158cd02
#12 [] ip_local_deliver_finish at ffffffff815668f4
#13 [] ip_local_deliver at ffffffff81566bd9
#14 [] ip_rcv_finish at ffffffff8156656d
#15 [] ip_rcv at ffffffff81566f06
#16 [] __netif_receive_skb_core at ffffffff8152b3a2
#17 [] __netif_receive_skb at ffffffff8152b608
#18 [] netif_receive_skb at ffffffff8152b690
#19 [] vmxnet3_rq_rx_complete at ffffffffa015eeaf [vmxnet3]
#20 [] vmxnet3_poll_rx_only at ffffffffa015f32a [vmxnet3]
#21 [] net_rx_action at ffffffff8152bac2
#22 [] __do_softirq at ffffffff81084b4f
#23 [] call_softirq at ffffffff8164845c
#24 [] do_softirq at ffffffff81016fc5
#25 [] irq_exit at ffffffff81084ee5
#26 [] do_IRQ at ffffffff81648ff8

Of course it may happen with other NIC drivers as well.

It's found the freed dst_entry here:

224 static bool tcp_in_quickack_mode(struct sock *sk)â
225 {â
226 â const struct inet_connection_sock *icsk = inet_csk(sk);â
227 â const struct dst_entry *dst = __sk_dst_get(sk);â
228 â
229 â return (dst && dst_metric(dst, RTAX_QUICKACK)) ||â
230 â â (icsk->icsk_ack.quick && !icsk->icsk_ack.pingpong);â
231 }â

But there are other backtraces attributed to the same freed dst_entry in
netfilter code as well.

All the vmcores showed 2 significant clues:

- Remote hosts behind the default gateway had always been redirected to a
different gateway. A rtable/dst_entry will be added for that host. Making
more dst_entrys with lower reference counts. Making this more probable.

- All vmcores showed a postitive LockDroppedIcmps value, e.g:

LockDroppedIcmps 267

A closer look at the tcp_v4_err() handler revealed that do_redirect() will run
regardless of whether user space has the socket locked. This can result in a
race condition where the same dst_entry cached in sk->sk_dst_entry can be
decremented twice for the same socket via:

do_redirect()->__sk_dst_check()-> dst_release().

Which leads to the dst_entry being prematurely freed with another socket
pointing to it via sk->sk_dst_cache and a subsequent crash.

To fix this skip do_redirect() if usespace has the socket locked. Instead let
the redirect take place later when user space does not have the socket

The dccp code is very similar in this respect, so fixing it there too.

As Eric Garver pointed out the following commit now invalidates routes. Which
can set the dst->obsolete flag so that ipv4_dst_check() returns null and
triggers the dst_release().

Fixes: ceb3320610d6 ("ipv4: Kill routes during PMTU/redirect updates.")
Cc: Eric Garver <egarver@xxxxxxxxxx>
Cc: Hannes Sowa <hsowa@xxxxxxxxxx>
Signed-off-by: Jon Maxwell <jmaxwell37@xxxxxxxxx>
net/dccp/ipv4.c | 3 ++-
net/ipv4/tcp_ipv4.c | 3 ++-
2 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/net/dccp/ipv4.c b/net/dccp/ipv4.c
index 409d0cf..b99168b 100644
--- a/net/dccp/ipv4.c
+++ b/net/dccp/ipv4.c
@@ -289,7 +289,8 @@ static void dccp_v4_err(struct sk_buff *skb, u32 info)

switch (type) {
- dccp_do_redirect(skb, sk);
+ if (!sock_owned_by_user(sk))
+ dccp_do_redirect(skb, sk);
goto out;
/* Just silently ignore these. */
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 8f3ec13..575e19d 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -431,7 +431,8 @@ void tcp_v4_err(struct sk_buff *icmp_skb, u32 info)

switch (type) {
- do_redirect(icmp_skb, sk);
+ if (!sock_owned_by_user(sk))
+ do_redirect(icmp_skb, sk);
goto out;
/* Just silently ignore these. */