ipv6: prevent possible UAF in ip6_xmit()

If skb_expand_head() returns NULL, skb has been freed
and the associated dst/idev could also have been freed.

We must use rcu_read_lock() to prevent a possible UAF.

Fixes: 0c9f227bee11 ("ipv6: use skb_expand_head in ip6_xmit")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Vasily Averin <vasily.averin@linux.dev>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://patch.msgid.link/20240820160859.3786976-4-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

diff --git a/net/ipv6/ip6_output.c b/net/ipv6/ip6_output.c
index 1b9ebee7308f02a626c766de1794e6b114ae8554..f26841f1490f5c31e3b09bce5773391e779128e1 100644 (file)
--- a/net/ipv6/ip6_output.c
+++ b/net/ipv6/ip6_output.c
@@ -287,11 +287,15 @@ int ip6_xmit(const struct sock *sk, struct sk_buff *skb, struct flowi6 *fl6,
                head_room += opt->opt_nflen + opt->opt_flen;
 
        if (unlikely(head_room > skb_headroom(skb))) {
+               /* Make sure idev stays alive */
+               rcu_read_lock();
                skb = skb_expand_head(skb, head_room);
                if (!skb) {
                        IP6_INC_STATS(net, idev, IPSTATS_MIB_OUTDISCARDS);
+                       rcu_read_unlock();
                        return -ENOBUFS;
                }
+               rcu_read_unlock();
        }
 
        if (opt) {