netfilter: nf_queue: reject NF_STOLEN verdicts from userspace

A userspace listener may send (bogus) NF_STOLEN verdict, which causes skb leak.

This problem was previously fixed via
64507fdbc2 (netfilter:
nf_queue: fix NF_STOLEN skb leak) but this had to be reverted because
NF_STOLEN can also be returned by a netfilter hook when iterating the
rules in nf_reinject.

Reject userspace NF_STOLEN verdict, as suggested by Michal Miroslaw.

This is complementary to commit fad5444043
(netfilter: avoid double free in nf_reinject).

Cc: Julian Anastasov <[email protected]>
Cc: Eric Dumazet <[email protected]>
Signed-off-by: Florian Westphal <[email protected]>
Signed-off-by: Patrick McHardy <[email protected]>
This commit is contained in:
Florian Westphal 2011-08-30 15:01:20 +02:00 • committed by Patrick McHardy
commit c6675233f9
3 changed files with 10 additions and 16 deletions

View file

@ -314,7 +314,7 @@ ipq_set_verdict(struct ipq_verdict_msg *vmsg, unsigned int len)
{
struct nf_queue_entry *entry;
if (vmsg->value > NF_MAX_VERDICT)
if (vmsg->value > NF_MAX_VERDICT || vmsg->value == NF_STOLEN)
return -EINVAL;
entry = ipq_find_dequeue_entry(vmsg->id);
@ -359,12 +359,9 @@ ipq_receive_peer(struct ipq_peer_msg *pmsg,
break;
case IPQM_VERDICT:
if (pmsg->msg.verdict.value > NF_MAX_VERDICT)
status = -EINVAL;
else
status = ipq_set_verdict(&pmsg->msg.verdict,
len - sizeof(*pmsg));
break;
status = ipq_set_verdict(&pmsg->msg.verdict,
len - sizeof(*pmsg));
break;
default:
status = -EINVAL;
}