Skip to contentSkip to conversation

udp: resubmit encapsulation packets on all multicast listeners

PATCH In review Mariano Baragiola 3 participants 4 messages Aug 15 – Aug 24

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 udp: resubmit encapsulation packets on all multicast listeners

UDP encapsulation handlers (FOU/GUE and similar) return a positive protocol number from udp_queue_rcv_skb()/udpv6_queue_rcv_skb() when the UDP header has been consumed and the packet must be resubmitted to the IP protocol handler. Unicast paths already propagate that return value.

Multicast delivery called consume_skb() on every positive return, so the inner packet was dropped instead of being reinjected.

Commit 3cb8d4b9bfeb ("udp: fix encapsulation packet resubmit in multicast deliver") fixed only the primary ("first") socket path on net-next and is not yet in net. Secondary listeners still clone the skb and drop it on a positive return, and net itself still drops the primary socket path too.

Resubmit secondary clones inline via ip_protocol_deliver_rcu() / ip6_protocol_deliver_rcu() (matching the GSO segment path in udp_queue_rcv_skb()/udpv6_queue_rcv_skb()), and propagate the primary socket return value with the same IPv4/IPv6 sign convention as the unicast helpers.

Fixes ca065d0cf80f ("udp: no longer use SLAB_DESTROY_BY_RCU")
---
 net/ipv4/udp.c | 16 ++++++++++++----
 net/ipv6/udp.c | 16 ++++++++++++----
 2 files changed, 24 insertions(+), 8 deletions(-)
net/ipv4/udp.c +12 -4
index 70f6cbd4ef73..d6a17b0462b6 100644
@@ -2475,6 +2475,7 @@ static int __udp4_lib_mcast_deliver(struct net *net, struct sk_buff *skb,
 	struct udp_hslot *hslot;
 	struct sk_buff *nskb;
 	bool use_hash2;
+	int ret;
 
 	hash2_any = 0;
 	hash2 = 0;
@@ -2508,8 +2509,13 @@ static int __udp4_lib_mcast_deliver(struct net *net, struct sk_buff *skb,
 			__UDP_INC_STATS(net, UDP_MIB_INERRORS);
 			continue;
 		}
-		if (udp_queue_rcv_skb(sk, nskb) > 0)
-			consume_skb(nskb);
+		/* >0 means the encap handler wants IP-level resubmit. Do that
+		 * inline for secondary listeners; only the first socket can
+		 * propagate the protocol number to the caller.
+		 */
+		ret = udp_queue_rcv_skb(sk, nskb);
+		if (ret > 0)
+			ip_protocol_deliver_rcu(net, nskb, ret);
 	}
 
 	/* Also lookup *:port if we are using hash2 and haven't done so yet. */
@@ -2519,8 +2525,10 @@ static int __udp4_lib_mcast_deliver(struct net *net, struct sk_buff *skb,
 	}
 
 	if (first) {
-		if (udp_queue_rcv_skb(first, skb) > 0)
-			consume_skb(skb);
+		ret = udp_queue_rcv_skb(first, skb);
+		/* Match udp_unicast_rcv_skb(): return -protocol for IPv4. */
+		if (ret > 0)
+			return -ret;
 	} else {
 		kfree_skb(skb);
 		__UDP_INC_STATS(net, UDP_MIB_IGNOREDMULTI);
net/ipv6/udp.c +12 -4
index 15e032194ecc..65dd944211e9 100644
@@ -949,6 +949,7 @@ static int __udp6_lib_mcast_deliver(struct net *net, struct sk_buff *skb,
 	struct udp_hslot *hslot;
 	struct sk_buff *nskb;
 	bool use_hash2;
+	int ret;
 
 	hash2_any = 0;
 	hash2 = 0;
@@ -987,8 +988,13 @@ static int __udp6_lib_mcast_deliver(struct net *net, struct sk_buff *skb,
 			continue;
 		}
 
-		if (udpv6_queue_rcv_skb(sk, nskb) > 0)
-			consume_skb(nskb);
+		/* >0 means the encap handler wants IP-level resubmit. Do that
+		 * inline for secondary listeners; only the first socket can
+		 * propagate the nexthdr to the caller.
+		 */
+		ret = udpv6_queue_rcv_skb(sk, nskb);
+		if (ret > 0)
+			ip6_protocol_deliver_rcu(net, nskb, ret, true);
 	}
 
 	/* Also lookup *:port if we are using hash2 and haven't done so yet. */
@@ -998,8 +1004,10 @@ static int __udp6_lib_mcast_deliver(struct net *net, struct sk_buff *skb,
 	}
 
 	if (first) {
-		if (udpv6_queue_rcv_skb(first, skb) > 0)
-			consume_skb(skb);
+		ret = udpv6_queue_rcv_skb(first, skb);
+		/* Match udp6_unicast_rcv_skb(): return protocol for IPv6. */
+		if (ret > 0)
+			return ret;
 	} else {
 		kfree_skb(skb);
 		__UDP6_INC_STATS(net, UDP_MIB_IGNOREDMULTI);
Signature · 1 line
2.55.0
maintainer
#1 lore raw
Quoted 5 lines · On Sat, 15 Aug 2026 11:11:28 -0300 Mariano Baragiola wrote:

UDP encapsulation handlers (FOU/GUE and similar) return a positive protocol number from udp_queue_rcv_skb()/udpv6_queue_rcv_skb() when the UDP header has been consumed and the packet must be resubmitted to the IP protocol handler. Unicast paths already propagate that return value.

Quoted 7 lines

Multicast delivery called consume_skb() on every positive return, so the inner packet was dropped instead of being reinjected.

Commit 3cb8d4b9bfeb ("udp: fix encapsulation packet resubmit in multicast deliver") fixed only the primary ("first") socket path on net-next and is not yet in net. Secondary listeners still clone the skb and drop it on a positive return, and net itself still drops the primary socket path too.

Wait a week, rebase and repost against net.

Quoted 5 lines

Resubmit secondary clones inline via ip_protocol_deliver_rcu() / ip6_protocol_deliver_rcu() (matching the GSO segment path in udp_queue_rcv_skb()/udpv6_queue_rcv_skb()), and propagate the primary socket return value with the same IPv4/IPv6 sign convention as the unicast helpers.

Please include in the commit description how the issue was found.

maintainer
#2 lore raw
In reply to Mariano Baragiola
Quoted 11 lines · Mariano Baragiola

UDP encapsulation handlers (FOU/GUE and similar) return a positive protocol number from udp_queue_rcv_skb()/udpv6_queue_rcv_skb() when the UDP header has been consumed and the packet must be resubmitted to the IP protocol handler. Unicast paths already propagate that return value.

Multicast delivery called consume_skb() on every positive return, so the inner packet was dropped instead of being reinjected.

Commit 3cb8d4b9bfeb ("udp: fix encapsulation packet resubmit in multicast deliver") fixed only the primary ("first") socket path on net-next and is not yet in net.

It is now, so this will conflict.

Quoted 2 lines

Secondary listeners still clone the skb and drop it on a positive return, and net itself still drops the primary socket path too.

When I asked about additional listeners Anton responded:

"
The clone loop is not reachable for encapsulation sockets, so there is no remaining gap.

FOU/GUE tunnel sockets are created via udp_sock_create() / setup_udp_tunnel_sock() and do not set SO_REUSEADDR or SO_REUSEPORT. Without either, a UDP socket cannot share its port, so an encap socket is always the only socket bound to its port. In __udp[46]_lib_mcast_deliver() it is therefore always delivered as 'first', and the clone loop -- which handles the second and subsequent sockets in the group -- never runs for it. The positive (resubmit) return from udp_queue_rcv_skb() only happens for encap sockets; plain UDP sockets return 0 or a negative value there. So the resubmit case in the clone loop cannot occur.
"
https://lore.kernel.org/netdev/alAAXv60KUZ9KYx1@dau-home-pc/

Quoted 1 line

Resubmit secondary clones inline via ip_protocol_deliver_rcu() /

So is this a real path?

Small caveat on the SO_REUSEADDR/SO_REUSEPORT: SO_BINDTODEVICE is another way to skip the equality check in udp_lib_lport_inuse2.

Quoted 6 lines

ip6_protocol_deliver_rcu() (matching the GSO segment path in udp_queue_rcv_skb()/udpv6_queue_rcv_skb()), and propagate the primary socket return value with the same IPv4/IPv6 sign convention as the unicast helpers.

Fixes ca065d0cf80f ("udp: no longer use SLAB_DESTROY_BY_RCU")

Fixes should be the above patch if that is incomplete.

Quoted 32 lines
---
 net/ipv4/udp.c | 16 ++++++++++++----
 net/ipv6/udp.c | 16 ++++++++++++----
 2 files changed, 24 insertions(+), 8 deletions(-)
net/ipv4/udp.c +8 -2
index 70f6cbd4ef73..d6a17b0462b6 100644
@@ -2475,6 +2475,7 @@ static int __udp4_lib_mcast_deliver(struct net *net, struct sk_buff *skb,
 	struct udp_hslot *hslot;
 	struct sk_buff *nskb;
 	bool use_hash2;
+	int ret;
 
 	hash2_any = 0;
 	hash2 = 0;
@@ -2508,8 +2509,13 @@ static int __udp4_lib_mcast_deliver(struct net *net, struct sk_buff *skb,
 			__UDP_INC_STATS(net, UDP_MIB_INERRORS);
 			continue;
 		}
-		if (udp_queue_rcv_skb(sk, nskb) > 0)
-			consume_skb(nskb);
+		/* >0 means the encap handler wants IP-level resubmit. Do that
+		 * inline for secondary listeners; only the first socket can
+		 * propagate the protocol number to the caller.
+		 */
+		ret = udp_queue_rcv_skb(sk, nskb);
+		if (ret > 0)
+			ip_protocol_deliver_rcu(net, nskb, ret);
 	}
#3 lore raw

Thanks for checking this.

I fetched current net and confirmed that 3cb8d4b9bfeb is now present, so the primary-path issue is fixed there.

I also checked the SO_BINDTODEVICE case. Although different bound devices can bypass the port-use equality check, multicast delivery requires the socket's bound device to match the packet's ingress device. Those sockets therefore cannot become multiple matching listeners for the same multicast packet. FOU/GUE sockets also do not enable SO_REUSEADDR or SO_REUSEPORT.

The secondary clone path is therefore not reachable for the positive encapsulation return. I will withdraw this patch rather than repost it.

For completeness, I found the original issue while comparing the multicast receive path with the unicast and GSO paths, then confirmed it with a focused FOU probe against clean and patched kernels. The clean kernel dropped the multicast encapsulated packets; the patched kernel resubmitted them to the IP protocol handler.

Thanks again for the review.