netfilter: nfnetlink: allow to detect if ctnetlink listeners exist

At this time, every new conntrack gets the 'event cache extension'
enabled for it.

This is because the 'net.netfilter.nf_conntrack_events' sysctl defaults
to 1.

Changing the default to 0 means that commands that rely on the event
notification extension, e.g. 'conntrack -E' or conntrackd, stop working.

We COULD detect if there is a listener by means of
'nfnetlink_has_listeners()' and only add the extension if this is true.

The downside is a dependency from conntrack module to nfnetlink module.

This adds a different way: inc/dec a counter whenever a ctnetlink group
is being (un)subscribed and toggle a flag in struct net.

Next patches will take advantage of this and will only add the event
extension if the flag is set.

Signed-off-by: Florian Westphal <[email protected]>
Signed-off-by: Pablo Neira Ayuso <[email protected]>

include/net/netns/conntrack.h

@@ -95,6 +95,7 @@ struct nf_ip_net {
 
 struct netns_ct {
 #ifdef CONFIG_NF_CONNTRACK_EVENTS
+	bool ctnetlink_has_listener;
 	bool ecache_dwork_pending;
 #endif
 	u8			sysctl_log_invalid; /* Log invalid packets */

net/netfilter/nfnetlink.c

@@ -45,6 +45,7 @@ MODULE_DESCRIPTION("Netfilter messages via netlink socket");
 static unsigned int nfnetlink_pernet_id __read_mostly;
 
 struct nfnl_net {
+	unsigned int ctnetlink_listeners;
 	struct sock *nfnl;
 };
 
@@ -654,7 +655,6 @@ static void nfnetlink_rcv(struct sk_buff *skb)
 		netlink_rcv_skb(skb, nfnetlink_rcv_msg);
 }
 
-#ifdef CONFIG_MODULES
 static int nfnetlink_bind(struct net *net, int group)
 {
 	const struct nfnetlink_subsystem *ss;
@@ -670,19 +670,53 @@ static int nfnetlink_bind(struct net *net, int group)
 	rcu_read_unlock();
 	if (!ss)
 		request_module_nowait("nfnetlink-subsys-%d", type);
+
+#ifdef CONFIG_NF_CONNTRACK_EVENTS
+	if (type == NFNL_SUBSYS_CTNETLINK) {
+		struct nfnl_net *nfnlnet = nfnl_pernet(net);
+
+		nfnl_lock(NFNL_SUBSYS_CTNETLINK);
+
+		if (WARN_ON_ONCE(nfnlnet->ctnetlink_listeners == UINT_MAX)) {
+			nfnl_unlock(NFNL_SUBSYS_CTNETLINK);
+			return -EOVERFLOW;
+		}
+
+		nfnlnet->ctnetlink_listeners++;
+		if (nfnlnet->ctnetlink_listeners == 1)
+			WRITE_ONCE(net->ct.ctnetlink_has_listener, true);
+		nfnl_unlock(NFNL_SUBSYS_CTNETLINK);
+	}
+#endif
 	return 0;
 }
+
+static void nfnetlink_unbind(struct net *net, int group)
+{
+#ifdef CONFIG_NF_CONNTRACK_EVENTS
+	int type = nfnl_group2type[group];
+
+	if (type == NFNL_SUBSYS_CTNETLINK) {
+		struct nfnl_net *nfnlnet = nfnl_pernet(net);
+
+		nfnl_lock(NFNL_SUBSYS_CTNETLINK);
+		WARN_ON_ONCE(nfnlnet->ctnetlink_listeners == 0);
+		nfnlnet->ctnetlink_listeners--;
+		if (nfnlnet->ctnetlink_listeners == 0)
+			WRITE_ONCE(net->ct.ctnetlink_has_listener, false);
+		nfnl_unlock(NFNL_SUBSYS_CTNETLINK);
+	}
 #endif
+}
 
 static int __net_init nfnetlink_net_init(struct net *net)
 {
 	struct nfnl_net *nfnlnet = nfnl_pernet(net);
 	struct netlink_kernel_cfg cfg = {
 		.groups	= NFNLGRP_MAX,
 		.input	= nfnetlink_rcv,
-#ifdef CONFIG_MODULES
 		.bind	= nfnetlink_bind,
-#endif
+		.unbind	= nfnetlink_unbind,
 	};
 
 	nfnlnet->nfnl = netlink_kernel_create(net, NETLINK_NETFILTER, &cfg);