Skip to contentSkip to conversation

refactor MPTCP socket option handling

PATCH In review Geliang Tang 2 participants 8 messages Aug 12, 2025

Subsystems (2)
lore
Share & tools
lore

Geliang Tang

author reviewer
#0 lore raw

cover refactor MPTCP socket option handling

From: Geliang Tang <[email protected]>

This series restructures MPTCP socket option implementation to improve maintainability and consistency. We introduce dedicated helper functions for different option levels (SOL_SOCKET, SOL_IP, SOL_IPV6, SOL_TCP) and propagation scopes (first subflow vs. all subflows), addressing current code fragmentation.

Geliang Tang (6):
mptcp: add helper for setsockopt on first subflow only
mptcp: handle SOL_SOCKET options on first subflow only
mptcp: refactor IP socket option handling for first subflow
mptcp: extract IPv6 socket option handling for first subflow
mptcp: handle SOL_TCP options on all subflows
mptcp: handle IP socket options on all subflows

 net/mptcp/sockopt.c | 215 +++++++++++++++++++++++++++++---------------
 1 file changed, 145 insertions(+), 70 deletions(-)
Signature · 1 line
2.48.1
author reviewer
#1 lore raw

PATCH 1/6 mptcp: add helper for setsockopt on first subflow only

From: Geliang Tang <[email protected]>

Implement mptcp_setsockopt_first_sf_only() helper to manage setsockopt operations applied solely to the first subflow. This centralizes handling of TCP-level options that require synchronization only with the initial subflow.

---
 net/mptcp/sockopt.c | 19 ++++++++++++++++---
 1 file changed, 16 insertions(+), 3 deletions(-)
net/mptcp/sockopt.c +16 -3
index 2c267aff95be..6cc073fab17c 100644
@@ -776,8 +776,8 @@ static int mptcp_setsockopt_v4(struct mptcp_sock *msk, int optname,
 	return -EOPNOTSUPP;
 }
 
-static int mptcp_setsockopt_first_sf_only(struct mptcp_sock *msk, int level, int optname,
-					  sockptr_t optval, unsigned int optlen)
+static int mptcp_setsockopt_sol_tcp_first_sf_only(struct mptcp_sock *msk, int optname,
+						  sockptr_t optval, unsigned int optlen)
 {
 	struct sock *sk = (struct sock *)msk;
 	struct sock *ssk;
@@ -791,13 +791,26 @@ static int mptcp_setsockopt_first_sf_only(struct mptcp_sock *msk, int level, int
 		goto unlock;
 	}
 
-	ret = tcp_setsockopt(ssk, level, optname, optval, optlen);
+	ret = tcp_setsockopt(ssk, SOL_TCP, optname, optval, optlen);
 
 unlock:
 	release_sock(sk);
 	return ret;
 }
 
+static int mptcp_setsockopt_first_sf_only(struct mptcp_sock *msk, int level,
+					  int optname, sockptr_t optval,
+					  unsigned int optlen)
+{
+	switch (level) {
+	case SOL_TCP:
+		return mptcp_setsockopt_sol_tcp_first_sf_only(msk, optname,
+							      optval, optlen);
+	}
+
+	return -EOPNOTSUPP;
+}
+
 static int mptcp_setsockopt_all_sf(struct mptcp_sock *msk, int level,
 				   int optname, sockptr_t optval,
 				   unsigned int optlen)
Signature · 1 line
2.48.1
author reviewer
#2 lore raw

PATCH 2/6 mptcp: handle SOL_SOCKET options on first subflow only

In reply to Geliang Tang

From: Geliang Tang <[email protected]>

Refactor SOL_SOCKET option handling using the common handler mptcp_setsockopt_sol_socket_first_sf_only(). This helper applies socket options (SO_REUSEPORT, SO_REUSEADDR, SO_BINDTODEVICE, SO_BINDTOIFINDEX) to the first subflow and synchronizes values back to the MPTCP socket.

---
 net/mptcp/sockopt.c | 64 +++++++++++++++++++++++++++++----------------
 1 file changed, 42 insertions(+), 22 deletions(-)
net/mptcp/sockopt.c +42 -22
index 6cc073fab17c..8017cbad0597 100644
@@ -17,6 +17,10 @@
 #define MIN_INFO_OPTLEN_SIZE		16
 #define MIN_FULL_INFO_OPTLEN_SIZE	40
 
+static int mptcp_setsockopt_first_sf_only(struct mptcp_sock *msk, int level,
+					  int optname, sockptr_t optval,
+					  unsigned int optlen);
+
 static struct sock *__mptcp_tcp_fallback(struct mptcp_sock *msk)
 {
 	msk_owned_by_me(msk);
@@ -287,38 +291,51 @@ static int mptcp_setsockopt_sol_socket_linger(struct mptcp_sock *msk, sockptr_t
 	return 0;
 }
 
-static int mptcp_setsockopt_sol_socket(struct mptcp_sock *msk, int optname,
-				       sockptr_t optval, unsigned int optlen)
+static int mptcp_setsockopt_sol_socket_first_sf_only(struct mptcp_sock *msk, int optname,
+						     sockptr_t optval, unsigned int optlen)
 {
 	struct sock *sk = (struct sock *)msk;
 	struct sock *ssk;
 	int ret;
 
+	if (optname != SO_REUSEPORT && optname != SO_REUSEADDR &&
+	    optname != SO_BINDTODEVICE && optname != SO_BINDTOIFINDEX)
+		return -EOPNOTSUPP;
+
+	lock_sock(sk);
+	ssk = __mptcp_nmpc_sk(msk);
+	if (IS_ERR(ssk)) {
+		release_sock(sk);
+		return PTR_ERR(ssk);
+	}
+
+	ret = sk_setsockopt(ssk, SOL_SOCKET, optname, optval, optlen);
+	if (ret == 0) {
+		if (optname == SO_REUSEPORT)
+			sk->sk_reuseport = ssk->sk_reuseport;
+		else if (optname == SO_REUSEADDR)
+			sk->sk_reuse = ssk->sk_reuse;
+		else if (optname == SO_BINDTODEVICE)
+			sk->sk_bound_dev_if = ssk->sk_bound_dev_if;
+		else if (optname == SO_BINDTOIFINDEX)
+			sk->sk_bound_dev_if = ssk->sk_bound_dev_if;
+	}
+	release_sock(sk);
+	return ret;
+}
+
+static int mptcp_setsockopt_sol_socket(struct mptcp_sock *msk, int optname,
+				       sockptr_t optval, unsigned int optlen)
+{
+	struct sock *sk = (struct sock *)msk;
+
 	switch (optname) {
 	case SO_REUSEPORT:
 	case SO_REUSEADDR:
 	case SO_BINDTODEVICE:
 	case SO_BINDTOIFINDEX:
-		lock_sock(sk);
-		ssk = __mptcp_nmpc_sk(msk);
-		if (IS_ERR(ssk)) {
-			release_sock(sk);
-			return PTR_ERR(ssk);
-		}
-
-		ret = sk_setsockopt(ssk, SOL_SOCKET, optname, optval, optlen);
-		if (ret == 0) {
-			if (optname == SO_REUSEPORT)
-				sk->sk_reuseport = ssk->sk_reuseport;
-			else if (optname == SO_REUSEADDR)
-				sk->sk_reuse = ssk->sk_reuse;
-			else if (optname == SO_BINDTODEVICE)
-				sk->sk_bound_dev_if = ssk->sk_bound_dev_if;
-			else if (optname == SO_BINDTOIFINDEX)
-				sk->sk_bound_dev_if = ssk->sk_bound_dev_if;
-		}
-		release_sock(sk);
-		return ret;
+		return mptcp_setsockopt_first_sf_only(msk, SOL_SOCKET, optname,
+						      optval, optlen);
 	case SO_KEEPALIVE:
 	case SO_PRIORITY:
 	case SO_SNDBUF:
@@ -806,6 +823,9 @@ static int mptcp_setsockopt_first_sf_only(struct mptcp_sock *msk, int level,
 	case SOL_TCP:
 		return mptcp_setsockopt_sol_tcp_first_sf_only(msk, optname,
 							      optval, optlen);
+	case SOL_SOCKET:
+		return mptcp_setsockopt_sol_socket_first_sf_only(msk, optname,
+								 optval, optlen);
 	}
 
 	return -EOPNOTSUPP;
Signature · 1 line
2.48.1
author reviewer
#3 lore raw

PATCH 3/6 mptcp: refactor IP socket option handling for first subflow

In reply to Geliang Tang

From: Geliang Tang <[email protected]>

Consolidate IP socket option logic with mptcp_setsockopt_first_sf_only()
- the common setsockopt handler for first subflow. This function applies options (IP_TRANSPARENT, IP_BIND_ADDRESS_NO_PORT, IP_LOCAL_PORT_RANGE) to the first subflow and synchronizes values to the MPTCP socket. Improves code organization for future socket option implementations.

---
 net/mptcp/sockopt.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)
net/mptcp/sockopt.c +6 -3
index 8017cbad0597..1ab3dbe5124e 100644
@@ -703,8 +703,8 @@ static int __mptcp_setsockopt_sol_tcp_nodelay(struct mptcp_sock *msk, int val)
 	return 0;
 }
 
-static int mptcp_setsockopt_sol_ip_set(struct mptcp_sock *msk, int optname,
-				       sockptr_t optval, unsigned int optlen)
+static int mptcp_setsockopt_sol_ip_first_sf_only(struct mptcp_sock *msk, int optname,
+						 sockptr_t optval, unsigned int optlen)
 {
 	struct sock *sk = (struct sock *)msk;
 	struct sock *ssk;
@@ -785,7 +785,7 @@ static int mptcp_setsockopt_v4(struct mptcp_sock *msk, int optname,
 	case IP_TRANSPARENT:
 	case IP_BIND_ADDRESS_NO_PORT:
 	case IP_LOCAL_PORT_RANGE:
-		return mptcp_setsockopt_sol_ip_set(msk, optname, optval, optlen);
+		return mptcp_setsockopt_first_sf_only(msk, SOL_IP, optname, optval, optlen);
 	case IP_TOS:
 		return mptcp_setsockopt_v4_set_tos(msk, optname, optval, optlen);
 	}
@@ -826,6 +826,9 @@ static int mptcp_setsockopt_first_sf_only(struct mptcp_sock *msk, int level,
 	case SOL_SOCKET:
 		return mptcp_setsockopt_sol_socket_first_sf_only(msk, optname,
 								 optval, optlen);
+	case SOL_IP:
+		return mptcp_setsockopt_sol_ip_first_sf_only(msk, optname,
+							     optval, optlen);
 	}
 
 	return -EOPNOTSUPP;
Signature · 1 line
2.48.1
author reviewer
#4 lore raw

PATCH 4/6 mptcp: extract IPv6 socket option handling for first subflow

In reply to Geliang Tang

From: Geliang Tang <[email protected]>

Extract IPv6 socket option logic into dedicated helper function mptcp_setsockopt_sol_ipv6_first_sf_only(). This helper handles:

1. Options requiring first-subflow-only application:
- IPV6_V6ONLY
- IPV6_TRANSPARENT
- IPV6_FREEBIND

2. Synchronization flow:
a) Lock MPTCP socket
b) Retrieve first subflow
c) Apply option to subflow
d) Update MPTCP socket state
e) Increment option sequence counter

The refactoring provides:
- Clear separation of IPv6-specific logic
- Consistent error handling
- Reusable pattern for future IPv6 options
- Improved code maintainability

---
 net/mptcp/sockopt.c | 80 ++++++++++++++++++++++++++-------------------
 1 file changed, 47 insertions(+), 33 deletions(-)
net/mptcp/sockopt.c +47 -33
index 1ab3dbe5124e..964131ef82d8 100644
@@ -401,51 +401,62 @@ static int mptcp_setsockopt_sol_socket(struct mptcp_sock *msk, int optname,
 	return -EOPNOTSUPP;
 }
 
-static int mptcp_setsockopt_v6(struct mptcp_sock *msk, int optname,
-			       sockptr_t optval, unsigned int optlen)
+static int mptcp_setsockopt_sol_ipv6_first_sf_only(struct mptcp_sock *msk, int optname,
+						   sockptr_t optval, unsigned int optlen)
 {
 	struct sock *sk = (struct sock *)msk;
-	int ret = -EOPNOTSUPP;
 	struct sock *ssk;
+	int ret;
+
+	if (optname != IPV6_V6ONLY && optname != IPV6_TRANSPARENT &&
+	    optname != IPV6_FREEBIND)
+		return -EOPNOTSUPP;
+
+	lock_sock(sk);
+	ssk = __mptcp_nmpc_sk(msk);
+	if (IS_ERR(ssk)) {
+		release_sock(sk);
+		return PTR_ERR(ssk);
+	}
+
+	ret = tcp_setsockopt(ssk, SOL_IPV6, optname, optval, optlen);
+	if (ret != 0) {
+		release_sock(sk);
+		return ret;
+	}
+
+	sockopt_seq_inc(msk);
 
 	switch (optname) {
 	case IPV6_V6ONLY:
+		sk->sk_ipv6only = ssk->sk_ipv6only;
+		break;
 	case IPV6_TRANSPARENT:
+		inet_assign_bit(TRANSPARENT, sk,
+				inet_test_bit(TRANSPARENT, ssk));
+		break;
 	case IPV6_FREEBIND:
-		lock_sock(sk);
-		ssk = __mptcp_nmpc_sk(msk);
-		if (IS_ERR(ssk)) {
-			release_sock(sk);
-			return PTR_ERR(ssk);
-		}
-
-		ret = tcp_setsockopt(ssk, SOL_IPV6, optname, optval, optlen);
-		if (ret != 0) {
-			release_sock(sk);
-			return ret;
-		}
-
-		sockopt_seq_inc(msk);
+		inet_assign_bit(FREEBIND, sk,
+				inet_test_bit(FREEBIND, ssk));
+		break;
+	}
 
-		switch (optname) {
-		case IPV6_V6ONLY:
-			sk->sk_ipv6only = ssk->sk_ipv6only;
-			break;
-		case IPV6_TRANSPARENT:
-			inet_assign_bit(TRANSPARENT, sk,
-					inet_test_bit(TRANSPARENT, ssk));
-			break;
-		case IPV6_FREEBIND:
-			inet_assign_bit(FREEBIND, sk,
-					inet_test_bit(FREEBIND, ssk));
-			break;
-		}
+	release_sock(sk);
+	return 0;
+}
 
-		release_sock(sk);
-		break;
+static int mptcp_setsockopt_v6(struct mptcp_sock *msk, int optname,
+			       sockptr_t optval, unsigned int optlen)
+{
+	switch (optname) {
+	case IPV6_V6ONLY:
+	case IPV6_TRANSPARENT:
+	case IPV6_FREEBIND:
+		return mptcp_setsockopt_first_sf_only(msk, SOL_IPV6, optname,
+						      optval, optlen);
 	}
 
-	return ret;
+	return -EOPNOTSUPP;
 }
 
 static bool mptcp_supported_sockopt(int level, int optname)
@@ -829,6 +840,9 @@ static int mptcp_setsockopt_first_sf_only(struct mptcp_sock *msk, int level,
 	case SOL_IP:
 		return mptcp_setsockopt_sol_ip_first_sf_only(msk, optname,
 							     optval, optlen);
+	case SOL_IPV6:
+		return mptcp_setsockopt_sol_ipv6_first_sf_only(msk, optname,
+							       optval, optlen);
 	}
 
 	return -EOPNOTSUPP;
Signature · 1 line
2.48.1
author reviewer
#5 lore raw

PATCH 5/6 mptcp: handle SOL_TCP options on all subflows

In reply to Geliang Tang

From: Geliang Tang <[email protected]>

Implement mptcp_setsockopt_sol_tcp_all_sf() helper to apply TCP socket options to all existing subflows. This ensures consistent configuration across every subflow in the MPTCP connection.

---
 net/mptcp/sockopt.c | 21 +++++++++++++++++----
 1 file changed, 17 insertions(+), 4 deletions(-)
net/mptcp/sockopt.c +17 -4
index 964131ef82d8..81cc24125c29 100644
@@ -848,9 +848,9 @@ static int mptcp_setsockopt_first_sf_only(struct mptcp_sock *msk, int level,
 	return -EOPNOTSUPP;
 }
 
-static int mptcp_setsockopt_all_sf(struct mptcp_sock *msk, int level,
-				   int optname, sockptr_t optval,
-				   unsigned int optlen)
+static int mptcp_setsockopt_sol_tcp_all_sf(struct mptcp_sock *msk,
+					   int optname, sockptr_t optval,
+					   unsigned int optlen)
 {
 	struct mptcp_subflow_context *subflow;
 	int ret = 0;
@@ -858,13 +858,26 @@ static int mptcp_setsockopt_all_sf(struct mptcp_sock *msk, int level,
 	mptcp_for_each_subflow(msk, subflow) {
 		struct sock *ssk = mptcp_subflow_tcp_sock(subflow);
 
-		ret = tcp_setsockopt(ssk, level, optname, optval, optlen);
+		ret = tcp_setsockopt(ssk, SOL_TCP, optname, optval, optlen);
 		if (ret)
 			break;
 	}
 	return ret;
 }
 
+static int mptcp_setsockopt_all_sf(struct mptcp_sock *msk, int level,
+				   int optname, sockptr_t optval,
+				   unsigned int optlen)
+{
+	switch (level) {
+	case SOL_TCP:
+		return mptcp_setsockopt_sol_tcp_all_sf(msk, optname,
+						       optval, optlen);
+	}
+
+	return -EOPNOTSUPP;
+}
+
 static int mptcp_setsockopt_sol_tcp(struct mptcp_sock *msk, int optname,
 				    sockptr_t optval, unsigned int optlen)
 {
Signature · 1 line
2.48.1
author reviewer
#6 lore raw

PATCH 6/6 mptcp: handle IP socket options on all subflows

In reply to Geliang Tang

From: Geliang Tang <[email protected]>

Introduce mptcp_ip_setsockopt_all_sf() helper to consistently apply IP socket options across all MPTCP subflows. This helper:
1. Sets options on the main MPTCP socket
2. Propagates values to all existing subflows
3. Uses lock synchronization for atomic updates
4. Maintains option sequence counters

Centralizes IP-level option handling for future options requiring subflow propagation.

---
 net/mptcp/sockopt.c | 22 +++++++++++++++++-----
 1 file changed, 17 insertions(+), 5 deletions(-)
net/mptcp/sockopt.c +17 -5
index 81cc24125c29..f5579fd8b70d 100644
@@ -20,6 +20,9 @@
 static int mptcp_setsockopt_first_sf_only(struct mptcp_sock *msk, int level,
 					  int optname, sockptr_t optval,
 					  unsigned int optlen);
+static int mptcp_setsockopt_all_sf(struct mptcp_sock *msk, int level,
+				   int optname, sockptr_t optval,
+				   unsigned int optlen);
 
 static struct sock *__mptcp_tcp_fallback(struct mptcp_sock *msk)
 {
@@ -760,13 +763,17 @@ static int mptcp_setsockopt_sol_ip_first_sf_only(struct mptcp_sock *msk, int opt
 	return 0;
 }
 
-static int mptcp_setsockopt_v4_set_tos(struct mptcp_sock *msk, int optname,
-				       sockptr_t optval, unsigned int optlen)
+static int mptcp_setsockopt_sol_ip_all_sf(struct mptcp_sock *msk,
+					  int optname, sockptr_t optval,
+					  unsigned int optlen)
 {
 	struct mptcp_subflow_context *subflow;
 	struct sock *sk = (struct sock *)msk;
 	int err, val;
 
+	if (optname != IP_TOS)
+		return -EOPNOTSUPP;
+
 	err = ip_setsockopt(sk, SOL_IP, optname, optval, optlen);
 
 	if (err != 0)
@@ -774,13 +781,15 @@ static int mptcp_setsockopt_v4_set_tos(struct mptcp_sock *msk, int optname,
 
 	lock_sock(sk);
 	sockopt_seq_inc(msk);
-	val = READ_ONCE(inet_sk(sk)->tos);
+	if (optname == IP_TOS)
+		val = READ_ONCE(inet_sk(sk)->tos);
 	mptcp_for_each_subflow(msk, subflow) {
 		struct sock *ssk = mptcp_subflow_tcp_sock(subflow);
 		bool slow;
 
 		slow = lock_sock_fast(ssk);
-		__ip_sock_set_tos(ssk, val);
+		if (optname == IP_TOS)
+			__ip_sock_set_tos(ssk, val);
 		unlock_sock_fast(ssk, slow);
 	}
 	release_sock(sk);
@@ -798,7 +807,7 @@ static int mptcp_setsockopt_v4(struct mptcp_sock *msk, int optname,
 	case IP_LOCAL_PORT_RANGE:
 		return mptcp_setsockopt_first_sf_only(msk, SOL_IP, optname, optval, optlen);
 	case IP_TOS:
-		return mptcp_setsockopt_v4_set_tos(msk, optname, optval, optlen);
+		return mptcp_setsockopt_all_sf(msk, SOL_IP, optname, optval, optlen);
 	}
 
 	return -EOPNOTSUPP;
@@ -873,6 +882,9 @@ static int mptcp_setsockopt_all_sf(struct mptcp_sock *msk, int level,
 	case SOL_TCP:
 		return mptcp_setsockopt_sol_tcp_all_sf(msk, optname,
 						       optval, optlen);
+	case SOL_IP:
+		return mptcp_setsockopt_sol_ip_all_sf(msk, optname,
+						      optval, optlen);
 	}
 
 	return -EOPNOTSUPP;
Signature · 1 line
2.48.1
#7 lore raw
In reply to Geliang Tang

Hi Geliang,

Thank you for your modifications, that's great!

Our CI did some validations and here is its report:

- KVM Validation: normal: Unstable: 1 failed test(s): selftest_mptcp_connect 🔴
- KVM Validation: debug: Unstable: 1 failed test(s): packetdrill_mp_join 🔴
- KVM Validation: btf-normal (only bpftest_all): Success! ✅
- KVM Validation: btf-debug (only bpftest_all): Success! ✅
- Task: https://github.com/multipath-tcp/mptcp_net-next/actions/runs/16903519178

Initiator: Patchew Applier
Commits: https://github.com/multipath-tcp/mptcp_net-next/commits/54fd10adb571
Patchwork: https://patchwork.kernel.org/project/mptcp/list/?series=990414

If there are some issues, you can reproduce them using the same environment as the one used by the CI thanks to a docker image, e.g.:

    $ cd [kernel source code]
    $ docker run -v "${PWD}:${PWD}:rw" -w "${PWD}" --privileged --rm -it \
        --pull always mptcp/mptcp-upstream-virtme-docker:latest \
        auto-normal

For more details:

https://github.com/multipath-tcp/mptcp-upstream-virtme-docker

Please note that despite all the efforts that have been already done to have a stable tests suite when executed on a public CI like here, it is possible some reported issues are not due to your modifications. Still, do not hesitate to help us improve that ;-)

Cheers,
MPTCP GH Action bot
Bot operated by Matthieu Baerts (NGI0 Core)