netfilter: ctnetlink: fix incorrect nf_ct_put during hash resize
authorLiping Zhang <zlpnobody@gmail.com>
Sat, 20 May 2017 23:22:49 +0000 (07:22 +0800)
committerPablo Neira Ayuso <pablo@netfilter.org>
Wed, 24 May 2017 09:26:01 +0000 (11:26 +0200)
commitfefa92679dbe0c613e62b6c27235dcfbe9640ad1
tree3b3eb1b92d592116594a647a8cedabf3a8e6e5fe
parent124dffea9e8e372509e055aebd118e85518fd644
netfilter: ctnetlink: fix incorrect nf_ct_put during hash resize

If nf_conntrack_htable_size was adjusted by the user during the ct
dump operation, we may invoke nf_ct_put twice for the same ct, i.e.
the "last" ct. This will cause the ct will be freed but still linked
in hash buckets.

It's very easy to reproduce the problem by the following commands:
  # while : ; do
  echo $RANDOM > /proc/sys/net/netfilter/nf_conntrack_buckets
  done
  # while : ; do
  conntrack -L
  done
  # iperf -s 127.0.0.1 &
  # iperf -c 127.0.0.1 -P 60 -t 36000

After a while, the system will hang like this:
  NMI watchdog: BUG: soft lockup - CPU#1 stuck for 22s! [bash:20184]
  NMI watchdog: BUG: soft lockup - CPU#0 stuck for 22s! [iperf:20382]
  ...

So at last if we find cb->args[1] is equal to "last", this means hash
resize happened, then we can set cb->args[1] to 0 to fix the above
issue.

Fixes: d205dc40798d ("[NETFILTER]: ctnetlink: fix deadlock in table dumping")
Signed-off-by: Liping Zhang <zlpnobody@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
net/netfilter/nf_conntrack_netlink.c