tcp/udp: Make early_demux back namespacified.

Commit e21145a9871a ("ipv4: namespacify ip_early_demux sysctl knob") made
it possible to enable/disable early_demux on a per-netns basis.  Then, we
introduced two knobs, tcp_early_demux and udp_early_demux, to switch it for
TCP/UDP in commit dddb64bcb346 ("net: Add sysctl to toggle early demux for
tcp and udp").  However, the .proc_handler() was wrong and actually
disabled us from changing the behaviour in each netns.

We can execute early_demux if net.ipv4.ip_early_demux is on and each proto
.early_demux() handler is not NULL.  When we toggle (tcp|udp)_early_demux,
the change itself is saved in each netns variable, but the .early_demux()
handler is a global variable, so the handler is switched based on the
init_net's sysctl variable.  Thus, netns (tcp|udp)_early_demux knobs have
nothing to do with the logic.  Whether we CAN execute proto .early_demux()
is always decided by init_net's sysctl knob, and whether we DO it or not is
by each netns ip_early_demux knob.

This patch namespacifies (tcp|udp)_early_demux again.  For now, the users
of the .early_demux() handler are TCP and UDP only, and they are called
directly to avoid retpoline.  So, we can remove the .early_demux() handler
from inet6?_protos and need not dereference them in ip6?_rcv_finish_core().
If another proto needs .early_demux(), we can restore it at that time.

Fixes: dddb64bcb346 ("net: Add sysctl to toggle early demux for tcp and udp")
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://lore.kernel.org/r/20220713175207.7727-1-kuniyu@amazon.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

6 files changed, 43 insertions, 108 deletions

diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c
index 4bc24f9e38b3..8baef2f3fc4b 100644
--- a/net/ipv4/af_inet.c
+++ b/net/ipv4/af_inet.c
@@ -1710,24 +1710,14 @@ static const struct net_protocol igmp_protocol = {
 };
 #endif
 
-/* thinking of making this const? Don't.
- * early_demux can change based on sysctl.
- */
-static struct net_protocol tcp_protocol = {
-	.early_demux	=	tcp_v4_early_demux,
-	.early_demux_handler =  tcp_v4_early_demux,
+static const struct net_protocol tcp_protocol = {
 	.handler	=	tcp_v4_rcv,
 	.err_handler	=	tcp_v4_err,
 	.no_policy	=	1,
 	.icmp_strict_tag_validation = 1,
 };
 
-/* thinking of making this const? Don't.
- * early_demux can change based on sysctl.
- */
-static struct net_protocol udp_protocol = {
-	.early_demux =	udp_v4_early_demux,
-	.early_demux_handler =	udp_v4_early_demux,
+static const struct net_protocol udp_protocol = {
 	.handler =	udp_rcv,
 	.err_handler =	udp_err,
 	.no_policy =	1,
diff --git a/net/ipv4/ip_input.c b/net/ipv4/ip_input.c
index b1165f717cd1..1b512390b3cf 100644
--- a/net/ipv4/ip_input.c
+++ b/net/ipv4/ip_input.c
@@ -312,14 +312,13 @@ static bool ip_can_use_hint(const struct sk_buff *skb, const struct iphdr *iph,
 	       ip_hdr(hint)->tos == iph->tos;
 }
 
-INDIRECT_CALLABLE_DECLARE(int udp_v4_early_demux(struct sk_buff *));
-INDIRECT_CALLABLE_DECLARE(int tcp_v4_early_demux(struct sk_buff *));
+int tcp_v4_early_demux(struct sk_buff *skb);
+int udp_v4_early_demux(struct sk_buff *skb);
 static int ip_rcv_finish_core(struct net *net, struct sock *sk,
 			      struct sk_buff *skb, struct net_device *dev,
 			      const struct sk_buff *hint)
 {
 	const struct iphdr *iph = ip_hdr(skb);
-	int (*edemux)(struct sk_buff *skb);
 	int err, drop_reason;
 	struct rtable *rt;
 
@@ -332,21 +331,29 @@ static int ip_rcv_finish_core(struct net *net, struct sock *sk,
 			goto drop_error;
 	}
 
-	if (net->ipv4.sysctl_ip_early_demux &&
+	if (READ_ONCE(net->ipv4.sysctl_ip_early_demux) &&
 	    !skb_dst(skb) &&
 	    !skb->sk &&
 	    !ip_is_fragment(iph)) {
-		const struct net_protocol *ipprot;
-		int protocol = iph->protocol;
-
-		ipprot = rcu_dereference(inet_protos[protocol]);
-		if (ipprot && (edemux = READ_ONCE(ipprot->early_demux))) {
-			err = INDIRECT_CALL_2(edemux, tcp_v4_early_demux,
-					      udp_v4_early_demux, skb);
-			if (unlikely(err))
-				goto drop_error;
-			/* must reload iph, skb->head might have changed */
-			iph = ip_hdr(skb);
+		switch (iph->protocol) {
+		case IPPROTO_TCP:
+			if (READ_ONCE(net->ipv4.sysctl_tcp_early_demux)) {
+				tcp_v4_early_demux(skb);
+
+				/* must reload iph, skb->head might have changed */
+				iph = ip_hdr(skb);
+			}
+			break;
+		case IPPROTO_UDP:
+			if (READ_ONCE(net->ipv4.sysctl_udp_early_demux)) {
+				err = udp_v4_early_demux(skb);
+				if (unlikely(err))
+					goto drop_error;
+
+				/* must reload iph, skb->head might have changed */
+				iph = ip_hdr(skb);
+			}
+			break;
 		}
 	}
 
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index 108fd86f2718..130e9c130311 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -350,61 +350,6 @@ bad_key:
 	return ret;
 }
 
-static void proc_configure_early_demux(int enabled, int protocol)
-{
-	struct net_protocol *ipprot;
-#if IS_ENABLED(CONFIG_IPV6)
-	struct inet6_protocol *ip6prot;
-#endif
-
-	rcu_read_lock();
-
-	ipprot = rcu_dereference(inet_protos[protocol]);
-	if (ipprot)
-		ipprot->early_demux = enabled ? ipprot->early_demux_handler :
-						NULL;
-
-#if IS_ENABLED(CONFIG_IPV6)
-	ip6prot = rcu_dereference(inet6_protos[protocol]);
-	if (ip6prot)
-		ip6prot->early_demux = enabled ? ip6prot->early_demux_handler :
-						 NULL;
-#endif
-	rcu_read_unlock();
-}
-
-static int proc_tcp_early_demux(struct ctl_table *table, int write,
-				void *buffer, size_t *lenp, loff_t *ppos)
-{
-	int ret = 0;
-
-	ret = proc_dou8vec_minmax(table, write, buffer, lenp, ppos);
-
-	if (write && !ret) {
-		int enabled = init_net.ipv4.sysctl_tcp_early_demux;
-
-		proc_configure_early_demux(enabled, IPPROTO_TCP);
-	}
-
-	return ret;
-}
-
-static int proc_udp_early_demux(struct ctl_table *table, int write,
-				void *buffer, size_t *lenp, loff_t *ppos)
-{
-	int ret = 0;
-
-	ret = proc_dou8vec_minmax(table, write, buffer, lenp, ppos);
-
-	if (write && !ret) {
-		int enabled = init_net.ipv4.sysctl_udp_early_demux;
-
-		proc_configure_early_demux(enabled, IPPROTO_UDP);
-	}
-
-	return ret;
-}
-
 static int proc_tfo_blackhole_detect_timeout(struct ctl_table *table,
 					     int write, void *buffer,
 					     size_t *lenp, loff_t *ppos)
@@ -707,14 +652,14 @@ static struct ctl_table ipv4_net_table[] = {
 		.data           = &init_net.ipv4.sysctl_udp_early_demux,
 		.maxlen         = sizeof(u8),
 		.mode           = 0644,
-		.proc_handler   = proc_udp_early_demux
+		.proc_handler   = proc_dou8vec_minmax,
 	},
 	{
 		.procname       = "tcp_early_demux",
 		.data           = &init_net.ipv4.sysctl_tcp_early_demux,
 		.maxlen         = sizeof(u8),
 		.mode           = 0644,
-		.proc_handler   = proc_tcp_early_demux
+		.proc_handler   = proc_dou8vec_minmax,
 	},
 	{
 		.procname       = "nexthop_compat_mode",
diff --git a/net/ipv6/ip6_input.c b/net/ipv6/ip6_input.c
index 0322cc86b84e..e1ebf5e42ebe 100644
--- a/net/ipv6/ip6_input.c
+++ b/net/ipv6/ip6_input.c
@@ -45,20 +45,23 @@
 #include <net/inet_ecn.h>
 #include <net/dst_metadata.h>
 
-INDIRECT_CALLABLE_DECLARE(void tcp_v6_early_demux(struct sk_buff *));
 static void ip6_rcv_finish_core(struct net *net, struct sock *sk,
 				struct sk_buff *skb)
 {
-	void (*edemux)(struct sk_buff *skb);
-
-	if (net->ipv4.sysctl_ip_early_demux && !skb_dst(skb) && skb->sk == NULL) {
-		const struct inet6_protocol *ipprot;
-
-		ipprot = rcu_dereference(inet6_protos[ipv6_hdr(skb)->nexthdr]);
-		if (ipprot && (edemux = READ_ONCE(ipprot->early_demux)))
-			INDIRECT_CALL_2(edemux, tcp_v6_early_demux,
-					udp_v6_early_demux, skb);
+	if (READ_ONCE(net->ipv4.sysctl_ip_early_demux) &&
+	    !skb_dst(skb) && !skb->sk) {
+		switch (ipv6_hdr(skb)->nexthdr) {
+		case IPPROTO_TCP:
+			if (READ_ONCE(net->ipv4.sysctl_tcp_early_demux))
+				tcp_v6_early_demux(skb);
+			break;
+		case IPPROTO_UDP:
+			if (READ_ONCE(net->ipv4.sysctl_udp_early_demux))
+				udp_v6_early_demux(skb);
+			break;
+		}
 	}
+
 	if (!skb_valid_dst(skb))
 		ip6_route_input(skb);
 }
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index f37dd4aa91c6..9d3ede293258 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -1822,7 +1822,7 @@ do_time_wait:
 	goto discard_it;
 }
 
-INDIRECT_CALLABLE_SCOPE void tcp_v6_early_demux(struct sk_buff *skb)
+void tcp_v6_early_demux(struct sk_buff *skb)
 {
 	const struct ipv6hdr *hdr;
 	const struct tcphdr *th;
@@ -2176,12 +2176,7 @@ struct proto tcpv6_prot = {
 };
 EXPORT_SYMBOL_GPL(tcpv6_prot);
 
-/* thinking of making this const? Don't.
- * early_demux can change based on sysctl.
- */
-static struct inet6_protocol tcpv6_protocol = {
-	.early_demux	=	tcp_v6_early_demux,
-	.early_demux_handler =  tcp_v6_early_demux,
+static const struct inet6_protocol tcpv6_protocol = {
 	.handler	=	tcp_v6_rcv,
 	.err_handler	=	tcp_v6_err,
 	.flags		=	INET6_PROTO_NOPOLICY|INET6_PROTO_FINAL,
diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c
index 55afd7f39c04..e2f2e087a753 100644
--- a/net/ipv6/udp.c
+++ b/net/ipv6/udp.c
@@ -1052,7 +1052,7 @@ static struct sock *__udp6_lib_demux_lookup(struct net *net,
 	return NULL;
 }
 
-INDIRECT_CALLABLE_SCOPE void udp_v6_early_demux(struct sk_buff *skb)
+void udp_v6_early_demux(struct sk_buff *skb)
 {
 	struct net *net = dev_net(skb->dev);
 	const struct udphdr *uh;
@@ -1660,12 +1660,7 @@ int udpv6_getsockopt(struct sock *sk, int level, int optname,
 	return ipv6_getsockopt(sk, level, optname, optval, optlen);
 }
 
-/* thinking of making this const? Don't.
- * early_demux can change based on sysctl.
- */
-static struct inet6_protocol udpv6_protocol = {
-	.early_demux	=	udp_v6_early_demux,
-	.early_demux_handler =  udp_v6_early_demux,
+static const struct inet6_protocol udpv6_protocol = {
 	.handler	=	udpv6_rcv,
 	.err_handler	=	udpv6_err,
 	.flags		=	INET6_PROTO_NOPOLICY|INET6_PROTO_FINAL,