ipv6: ip6_sk_dst_check() must not assume ipv6 dst
authorEric Dumazet <edumazet@google.com>
Wed, 26 Jun 2013 11:15:07 +0000 (04:15 -0700)
committerDavid S. Miller <davem@davemloft.net>
Wed, 26 Jun 2013 22:13:47 +0000 (15:13 -0700)
It's possible to use AF_INET6 sockets and to connect to an IPv4
destination. After this, socket dst cache is a pointer to a rtable,
not rt6_info.

ip6_sk_dst_check() should check the socket dst cache is IPv6, or else
various corruptions/crashes can happen.

Dave Jones can reproduce immediate crash with
trinity -q -l off -n -c sendmsg -c connect

With help from Hannes Frederic Sowa

Reported-by: Dave Jones <davej@redhat.com>
Reported-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
net/ipv6/ip6_output.c

index 95703ba2c084de31c8de3c41d830bb789f2f416e..d5d20cde8d92808387e171fb25b2fa701a06f275 100644 (file)
@@ -821,11 +821,17 @@ static struct dst_entry *ip6_sk_dst_check(struct sock *sk,
                                          const struct flowi6 *fl6)
 {
        struct ipv6_pinfo *np = inet6_sk(sk);
-       struct rt6_info *rt = (struct rt6_info *)dst;
+       struct rt6_info *rt;
 
        if (!dst)
                goto out;
 
+       if (dst->ops->family != AF_INET6) {
+               dst_release(dst);
+               return NULL;
+       }
+
+       rt = (struct rt6_info *)dst;
        /* Yes, checking route validity in not connected
         * case is not very simple. Take into account,
         * that we do not support routing by source, TOS,