On Wed, Sep 14, 2016 at 03:48:45PM -0400, Johannes Weiner wrote:
> From: Johannes Weiner <jwei...@fb.com>
>
> When a socket is cloned, the associated sock_cgroup_data is duplicated
> but not its reference on the cgroup. As a result, the cgroup reference
> count will underflow when both sockets are destroyed later on.
>
> Fixes: bd1060a1d671 ("sock, cgroup: add sock->sk_cgroup")
> Cc: <sta...@vger.kernel.org> # 4.5+
> Signed-off-by: Johannes Weiner <han...@cmpxchg.org>
Reviewed-by: Vladimir Davydov <vdavydov....@gmail.com>
