net: Avoid receiving packets with an l3mdev on unbound UDP sockets

Packets arriving in a VRF currently are delivered to UDP sockets that
aren't bound to any interface. TCP defaults to not delivering packets
arriving in a VRF to unbound sockets. IP route lookup and socket
transmit both assume that unbound means using the default table and
UDP applications that haven't been changed to be aware of VRFs may not
function correctly in this case since they may not be able to handle
overlapping IP address ranges, or be able to send packets back to the
original sender if required.

So add a sysctl, udp_l3mdev_accept, to control this behaviour with it
being analgous to the existing tcp_l3mdev_accept, namely to allow a
process to have a VRF-global listen socket. Have this default to off
as this is the behaviour that users will expect, given that there is
no explicit mechanism to set unmodified VRF-unaware application into a
default VRF.

Signed-off-by: Robert Shearman <rshearma@brocade.com>
Acked-by: David Ahern <dsa@cumulusnetworks.com>
Tested-by: David Ahern <dsa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

6 files changed, 66 insertions, 17 deletions

diff --git a/Documentation/networking/ip-sysctl.txt b/Documentation/networking/ip-sysctl.txt
index 17f2e7791042..fc73eeb7b3b8 100644
--- a/Documentation/networking/ip-sysctl.txt
+++ b/Documentation/networking/ip-sysctl.txt
@@ -721,6 +721,13 @@ tcp_challenge_ack_limit - INTEGER
 
 UDP variables:
 
+udp_l3mdev_accept - BOOLEAN
+	Enabling this option allows a "global" bound socket to work
+	across L3 master domains (e.g., VRFs) with packets capable of
+	being received regardless of the L3 domain in which they
+	originated. Only valid when the kernel was compiled with
+	CONFIG_NET_L3_MASTER_DEV.
+
 udp_mem - vector of 3 INTEGERs: min, pressure, max
 	Number of pages allowed for queueing by all UDP sockets.
 
diff --git a/Documentation/networking/vrf.txt b/Documentation/networking/vrf.txt
index 755dab856392..3918dae964d4 100644
--- a/Documentation/networking/vrf.txt
+++ b/Documentation/networking/vrf.txt
@@ -98,10 +98,11 @@ VRF device:
 
 or to specify the output device using cmsg and IP_PKTINFO.
 
-TCP services running in the default VRF context (ie., not bound to any VRF
-device) can work across all VRF domains by enabling the tcp_l3mdev_accept
-sysctl option:
+TCP & UDP services running in the default VRF context (ie., not bound
+to any VRF device) can work across all VRF domains by enabling the
+tcp_l3mdev_accept and udp_l3mdev_accept sysctl options:
     sysctl -w net.ipv4.tcp_l3mdev_accept=1
+    sysctl -w net.ipv4.udp_l3mdev_accept=1
 
 netfilter rules on the VRF device can be used to limit access to services
 running in the default VRF context as well.
diff --git a/include/net/netns/ipv4.h b/include/net/netns/ipv4.h
index e365732b8051..622d2da27135 100644
--- a/include/net/netns/ipv4.h
+++ b/include/net/netns/ipv4.h
@@ -124,6 +124,10 @@ struct netns_ipv4 {
 	struct inet_timewait_death_row tcp_death_row;
 	int sysctl_max_syn_backlog;
 
+#ifdef CONFIG_NET_L3_MASTER_DEV
+	int sysctl_udp_l3mdev_accept;
+#endif
+
 	int sysctl_igmp_max_memberships;
 	int sysctl_igmp_max_msf;
 	int sysctl_igmp_llm_reports;
diff --git a/net/ipv4/sysctl_net_ipv4.c b/net/ipv4/sysctl_net_ipv4.c
index 1b861997fdc5..d6880a6149ee 100644
--- a/net/ipv4/sysctl_net_ipv4.c
+++ b/net/ipv4/sysctl_net_ipv4.c
@@ -1012,6 +1012,17 @@ static struct ctl_table ipv4_net_table[] = {
 		.mode		= 0644,
 		.proc_handler	= ipv4_privileged_ports,
 	},
+#ifdef CONFIG_NET_L3_MASTER_DEV
+	{
+		.procname	= "udp_l3mdev_accept",
+		.data		= &init_net.ipv4.sysctl_udp_l3mdev_accept,
+		.maxlen		= sizeof(int),
+		.mode		= 0644,
+		.proc_handler	= proc_dointvec_minmax,
+		.extra1		= &zero,
+		.extra2		= &one,
+	},
+#endif
 	{ }
 };
 
diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c
index d6dddcf59e79..cf6ba3387401 100644
--- a/net/ipv4/udp.c
+++ b/net/ipv4/udp.c
@@ -134,6 +134,17 @@ EXPORT_SYMBOL(udp_memory_allocated);
 #define MAX_UDP_PORTS 65536
 #define PORTS_PER_CHAIN (MAX_UDP_PORTS / UDP_HTABLE_SIZE_MIN)
 
+/* IPCB reference means this can not be used from early demux */
+static bool udp_lib_exact_dif_match(struct net *net, struct sk_buff *skb)
+{
+#if IS_ENABLED(CONFIG_NET_L3_MASTER_DEV)
+	if (!net->ipv4.sysctl_udp_l3mdev_accept &&
+	    skb && ipv4_l3mdev_skb(IPCB(skb)->flags))
+		return true;
+#endif
+	return false;
+}
+
 static int udp_lib_lport_inuse(struct net *net, __u16 num,
 			       const struct udp_hslot *hslot,
 			       unsigned long *bitmap,
@@ -369,7 +380,8 @@ int udp_v4_get_port(struct sock *sk, unsigned short snum)
 
 static int compute_score(struct sock *sk, struct net *net,
 			 __be32 saddr, __be16 sport,
-			 __be32 daddr, unsigned short hnum, int dif)
+			 __be32 daddr, unsigned short hnum, int dif,
+			 bool exact_dif)
 {
 	int score;
 	struct inet_sock *inet;
@@ -400,7 +412,7 @@ static int compute_score(struct sock *sk, struct net *net,
 		score += 4;
 	}
 
-	if (sk->sk_bound_dev_if) {
+	if (sk->sk_bound_dev_if || exact_dif) {
 		if (sk->sk_bound_dev_if != dif)
 			return -1;
 		score += 4;
@@ -425,7 +437,7 @@ static u32 udp_ehashfn(const struct net *net, const __be32 laddr,
 /* called with rcu_read_lock() */
 static struct sock *udp4_lib_lookup2(struct net *net,
 		__be32 saddr, __be16 sport,
-		__be32 daddr, unsigned int hnum, int dif,
+		__be32 daddr, unsigned int hnum, int dif, bool exact_dif,
 		struct udp_hslot *hslot2,
 		struct sk_buff *skb)
 {
@@ -437,7 +449,7 @@ static struct sock *udp4_lib_lookup2(struct net *net,
 	badness = 0;
 	udp_portaddr_for_each_entry_rcu(sk, &hslot2->head) {
 		score = compute_score(sk, net, saddr, sport,
-				      daddr, hnum, dif);
+				      daddr, hnum, dif, exact_dif);
 		if (score > badness) {
 			reuseport = sk->sk_reuseport;
 			if (reuseport) {
@@ -472,6 +484,7 @@ struct sock *__udp4_lib_lookup(struct net *net, __be32 saddr,
 	unsigned short hnum = ntohs(dport);
 	unsigned int hash2, slot2, slot = udp_hashfn(net, hnum, udptable->mask);
 	struct udp_hslot *hslot2, *hslot = &udptable->hash[slot];
+	bool exact_dif = udp_lib_exact_dif_match(net, skb);
 	int score, badness, matches = 0, reuseport = 0;
 	u32 hash = 0;
 
@@ -484,7 +497,7 @@ struct sock *__udp4_lib_lookup(struct net *net, __be32 saddr,
 
 		result = udp4_lib_lookup2(net, saddr, sport,
 					  daddr, hnum, dif,
-					  hslot2, skb);
+					  exact_dif, hslot2, skb);
 		if (!result) {
 			unsigned int old_slot2 = slot2;
 			hash2 = udp4_portaddr_hash(net, htonl(INADDR_ANY), hnum);
@@ -499,7 +512,7 @@ struct sock *__udp4_lib_lookup(struct net *net, __be32 saddr,
 
 			result = udp4_lib_lookup2(net, saddr, sport,
 						  daddr, hnum, dif,
-						  hslot2, skb);
+						  exact_dif, hslot2, skb);
 		}
 		return result;
 	}
@@ -508,7 +521,7 @@ begin:
 	badness = 0;
 	sk_for_each_rcu(sk, &hslot->head) {
 		score = compute_score(sk, net, saddr, sport,
-				      daddr, hnum, dif);
+				      daddr, hnum, dif, exact_dif);
 		if (score > badness) {
 			reuseport = sk->sk_reuseport;
 			if (reuseport) {
diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c
index 05d69324862e..b4c6516a3a0c 100644
--- a/net/ipv6/udp.c
+++ b/net/ipv6/udp.c
@@ -55,6 +55,16 @@
 #include <trace/events/skb.h>
 #include "udp_impl.h"
 
+static bool udp6_lib_exact_dif_match(struct net *net, struct sk_buff *skb)
+{
+#if defined(CONFIG_NET_L3_MASTER_DEV)
+	if (!net->ipv4.sysctl_udp_l3mdev_accept &&
+	    skb && ipv6_l3mdev_skb(IP6CB(skb)->flags))
+		return true;
+#endif
+	return false;
+}
+
 static u32 udp6_ehashfn(const struct net *net,
 			const struct in6_addr *laddr,
 			const u16 lport,
@@ -118,7 +128,7 @@ static void udp_v6_rehash(struct sock *sk)
 static int compute_score(struct sock *sk, struct net *net,
 			 const struct in6_addr *saddr, __be16 sport,
 			 const struct in6_addr *daddr, unsigned short hnum,
-			 int dif)
+			 int dif, bool exact_dif)
 {
 	int score;
 	struct inet_sock *inet;
@@ -149,7 +159,7 @@ static int compute_score(struct sock *sk, struct net *net,
 		score++;
 	}
 
-	if (sk->sk_bound_dev_if) {
+	if (sk->sk_bound_dev_if || exact_dif) {
 		if (sk->sk_bound_dev_if != dif)
 			return -1;
 		score++;
@@ -165,7 +175,7 @@ static int compute_score(struct sock *sk, struct net *net,
 static struct sock *udp6_lib_lookup2(struct net *net,
 		const struct in6_addr *saddr, __be16 sport,
 		const struct in6_addr *daddr, unsigned int hnum, int dif,
-		struct udp_hslot *hslot2,
+		bool exact_dif, struct udp_hslot *hslot2,
 		struct sk_buff *skb)
 {
 	struct sock *sk, *result;
@@ -176,7 +186,7 @@ static struct sock *udp6_lib_lookup2(struct net *net,
 	badness = -1;
 	udp_portaddr_for_each_entry_rcu(sk, &hslot2->head) {
 		score = compute_score(sk, net, saddr, sport,
-				      daddr, hnum, dif);
+				      daddr, hnum, dif, exact_dif);
 		if (score > badness) {
 			reuseport = sk->sk_reuseport;
 			if (reuseport) {
@@ -212,6 +222,7 @@ struct sock *__udp6_lib_lookup(struct net *net,
 	unsigned short hnum = ntohs(dport);
 	unsigned int hash2, slot2, slot = udp_hashfn(net, hnum, udptable->mask);
 	struct udp_hslot *hslot2, *hslot = &udptable->hash[slot];
+	bool exact_dif = udp6_lib_exact_dif_match(net, skb);
 	int score, badness, matches = 0, reuseport = 0;
 	u32 hash = 0;
 
@@ -223,7 +234,7 @@ struct sock *__udp6_lib_lookup(struct net *net,
 			goto begin;
 
 		result = udp6_lib_lookup2(net, saddr, sport,
-					  daddr, hnum, dif,
+					  daddr, hnum, dif, exact_dif,
 					  hslot2, skb);
 		if (!result) {
 			unsigned int old_slot2 = slot2;
@@ -239,7 +250,8 @@ struct sock *__udp6_lib_lookup(struct net *net,
 
 			result = udp6_lib_lookup2(net, saddr, sport,
 						  daddr, hnum, dif,
-						  hslot2, skb);
+						  exact_dif, hslot2,
+						  skb);
 		}
 		return result;
 	}
@@ -247,7 +259,8 @@ begin:
 	result = NULL;
 	badness = -1;
 	sk_for_each_rcu(sk, &hslot->head) {
-		score = compute_score(sk, net, saddr, sport, daddr, hnum, dif);
+		score = compute_score(sk, net, saddr, sport, daddr, hnum, dif,
+				      exact_dif);
 		if (score > badness) {
 			reuseport = sk->sk_reuseport;
 			if (reuseport) {