If the first packet conntrack sees after a re-register is an outgoing
keepalive packet with no data (SEG.SEQ = SND.NXT-1), td_end is set to
SND.NXT-1.
When the peer correctly acknowledges SND.NXT, tcp_in_window fails
check III (Upper bound for valid (s)ack: sack <= receiver.td_end) and
returns false, which cascades into nf_conntrack_in setting
skb->_nfct = 0 and in later conntrack iptables rules not matching.
In cases where iptables are dropping packets that do not match
conntrack rules this can result in idle tcp connections to time out.
v2: adjust td_end when getting the reply rather than when sending out
the keepalive packet.
Fixes: f94e63801ab2 ("netfilter: conntrack: reset tcp maxwin on re-register")
Signed-off-by: Francesco Ruggeri <frugg...@arista.com>
---
net/netfilter/nf_conntrack_proto_tcp.c | 19 +++++++++++++------
1 file changed, 13 insertions(+), 6 deletions(-)
diff --git a/net/netfilter/nf_conntrack_proto_tcp.c
b/net/netfilter/nf_conntrack_proto_tcp.c
index e8c86ee4c1c4..c8fb2187ad4b 100644
--- a/net/netfilter/nf_conntrack_proto_tcp.c
+++ b/net/netfilter/nf_conntrack_proto_tcp.c
@@ -541,13 +541,20 @@ static bool tcp_in_window(const struct nf_conn *ct,
swin = win << sender->td_scale;
sender->td_maxwin = (swin == 0 ? 1 : swin);
sender->td_maxend = end + sender->td_maxwin;
- /*
- * We haven't seen traffic in the other direction yet
- * but we have to tweak window tracking to pass III
- * and IV until that happens.
- */
- if (receiver->td_maxwin == 0)
+ if (receiver->td_maxwin == 0) {
+ /* We haven't seen traffic in the other
+ * direction yet but we have to tweak window
+ * tracking to pass III and IV until that
+ * happens.
+ */
receiver->td_end = receiver->td_maxend = sack;
+ } else if (sack == receiver->td_end + 1) {
+ /* Likely a reply to a keepalive.
+ * Needed for III.
+ */
+ receiver->td_end++;
+ }
+
}
} else if (((state->state == TCP_CONNTRACK_SYN_SENT
&& dir == IP_CT_DIR_ORIGINAL)
--
2.28.0
