When get_ct_or_tuple_from_skb() falls back to looking u a connection via nf_conntrack_find_get(), a successful lookup sets ct but leaves tuple and zone unupdated. If the packet belongs to a reply flow, tuple will remain in the reply direction. As conncount relies on the original direction tuple to count the connections consistenly, passing an unnormalized reply tuple could lead to problems. Fix this by making sure that tuple and zone are normalized. Suggested-by: Florian Westphal Signed-off-by: Fernando Fernandez Mancera --- Note: I am sending it to nf-next because I couldn't reproduce this issue pointed out by Sashiko. Having said that, I still think the change should be done because it makes sense and for consistency reasons (aboved we are using the original direction too). --- net/netfilter/nf_conncount.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/net/netfilter/nf_conncount.c b/net/netfilter/nf_conncount.c index 91582069f6d2..4a62d3bef711 100644 --- a/net/netfilter/nf_conncount.c +++ b/net/netfilter/nf_conncount.c @@ -158,6 +158,8 @@ static bool get_ct_or_tuple_from_skb(struct net *net, return true; found_ct = nf_ct_tuplehash_to_ctrack(h); + *tuple = found_ct->tuplehash[IP_CT_DIR_ORIGINAL].tuple; + *zone = nf_ct_zone(found_ct); *refcounted = true; *ct = found_ct; -- 2.55.0