Some middleboxes might be nasty with MPTCP, and decide to drop packets
with MPTCP options, instead of just dropping the MPTCP options (or
letting them pass...).

In this case, it sounds better to fallback to "plain" TCP after 2
retransmissions, and try again.

Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/477
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
---
 include/net/mptcp.h  |  4 ++++
 net/ipv4/tcp_timer.c |  1 +
 net/mptcp/ctrl.c     | 20 ++++++++++++++++++++
 net/mptcp/mib.c      |  1 +
 net/mptcp/mib.h      |  1 +
 5 files changed, 27 insertions(+)

diff --git a/include/net/mptcp.h b/include/net/mptcp.h
index 0bc4ab03f487..814b5f2e3ed5 100644
--- a/include/net/mptcp.h
+++ b/include/net/mptcp.h
@@ -223,6 +223,8 @@ static inline __be32 mptcp_reset_option(const struct sk_buff *skb)
 
 	return htonl(0u);
 }
+
+void mptcp_active_detect_blackhole(struct sock *sk, bool expired);
 #else
 
 static inline void mptcp_init(void)
@@ -307,6 +309,8 @@ static inline struct request_sock *mptcp_subflow_reqsk_alloc(const struct reques
 }
 
 static inline __be32 mptcp_reset_option(const struct sk_buff *skb)  { return htonl(0u); }
+
+static inline void mptcp_active_detect_blackhole(struct sock *sk, bool expired) { }
 #endif /* CONFIG_MPTCP */
 
 #if IS_ENABLED(CONFIG_MPTCP_IPV6)
diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
index 86169127e4d1..79064580c8c0 100644
--- a/net/ipv4/tcp_timer.c
+++ b/net/ipv4/tcp_timer.c
@@ -282,6 +282,7 @@ static int tcp_write_timeout(struct sock *sk)
 		expired = retransmits_timed_out(sk, retry_until,
 						READ_ONCE(icsk->icsk_user_timeout));
 	tcp_fastopen_active_detect_blackhole(sk, expired);
+	mptcp_active_detect_blackhole(sk, expired);
 
 	if (BPF_SOCK_OPS_TEST_FLAG(tp, BPF_SOCK_OPS_RTO_CB_FLAG))
 		tcp_call_bpf_3arg(sk, BPF_SOCK_OPS_RTO_CB,
diff --git a/net/mptcp/ctrl.c b/net/mptcp/ctrl.c
index 99382c317ebb..0b23e3c5e8ff 100644
--- a/net/mptcp/ctrl.c
+++ b/net/mptcp/ctrl.c
@@ -12,6 +12,7 @@
 #include <net/netns/generic.h>
 
 #include "protocol.h"
+#include "mib.h"
 
 #define MPTCP_SYSCTL_PATH "net/mptcp"
 
@@ -277,6 +278,25 @@ static void mptcp_pernet_del_table(struct mptcp_pernet *pernet) {}
 
 #endif /* CONFIG_SYSCTL */
 
+/* Check the number of retransmissions, and fallback to TCP if needed */
+void mptcp_active_detect_blackhole(struct sock *ssk, bool expired)
+{
+	struct mptcp_subflow_context *subflow;
+	u32 timeouts;
+
+	if (!sk_is_mptcp(ssk))
+		return;
+
+	timeouts = inet_csk(ssk)->icsk_retransmits;
+	subflow = mptcp_subflow_ctx(ssk);
+
+	if (subflow->request_mptcp && ssk->sk_state == TCP_SYN_SENT &&
+	    (timeouts == 2 || (timeouts < 2 && expired))) {
+		MPTCP_INC_STATS(sock_net(ssk), MPTCP_MIB_MPCAPABLEACTIVEDROP);
+		mptcp_subflow_early_fallback(mptcp_sk(subflow->conn), subflow);
+	}
+}
+
 static int __net_init mptcp_net_init(struct net *net)
 {
 	struct mptcp_pernet *pernet = mptcp_get_pernet(net);
diff --git a/net/mptcp/mib.c b/net/mptcp/mib.c
index ec0d461cb921..d70a3e2bfad6 100644
--- a/net/mptcp/mib.c
+++ b/net/mptcp/mib.c
@@ -15,6 +15,7 @@ static const struct snmp_mib mptcp_snmp_list[] = {
 	SNMP_MIB_ITEM("MPCapableACKRX", MPTCP_MIB_MPCAPABLEPASSIVEACK),
 	SNMP_MIB_ITEM("MPCapableFallbackACK", MPTCP_MIB_MPCAPABLEPASSIVEFALLBACK),
 	SNMP_MIB_ITEM("MPCapableFallbackSYNACK", MPTCP_MIB_MPCAPABLEACTIVEFALLBACK),
+	SNMP_MIB_ITEM("MPCapableSYNTXDrop", MPTCP_MIB_MPCAPABLEACTIVEDROP),
 	SNMP_MIB_ITEM("MPFallbackTokenInit", MPTCP_MIB_TOKENFALLBACKINIT),
 	SNMP_MIB_ITEM("MPTCPRetrans", MPTCP_MIB_RETRANSSEGS),
 	SNMP_MIB_ITEM("MPJoinNoTokenFound", MPTCP_MIB_JOINNOTOKEN),
diff --git a/net/mptcp/mib.h b/net/mptcp/mib.h
index d68136f93dac..062775700b63 100644
--- a/net/mptcp/mib.h
+++ b/net/mptcp/mib.h
@@ -10,6 +10,7 @@ enum linux_mptcp_mib_field {
 	MPTCP_MIB_MPCAPABLEPASSIVEACK,	/* Received third ACK with MP_CAPABLE */
 	MPTCP_MIB_MPCAPABLEPASSIVEFALLBACK,/* Server-side fallback during 3-way handshake */
 	MPTCP_MIB_MPCAPABLEACTIVEFALLBACK, /* Client-side fallback during 3-way handshake */
+	MPTCP_MIB_MPCAPABLEACTIVEDROP,	/* Client-side fallback due to a MPC drop */
 	MPTCP_MIB_TOKENFALLBACKINIT,	/* Could not init/allocate token */
 	MPTCP_MIB_RETRANSSEGS,		/* Segments retransmitted at the MPTCP-level */
 	MPTCP_MIB_JOINNOTOKEN,		/* Received MP_JOIN but the token was not found */

-- 
2.45.2

On Mon, Sep 9, 2024 at 10:10 PM Matthieu Baerts (NGI0)
<matttbe@kernel.org> wrote:
>
> Some middleboxes might be nasty with MPTCP, and decide to drop packets
> with MPTCP options, instead of just dropping the MPTCP options (or
> letting them pass...).
>
> In this case, it sounds better to fallback to "plain" TCP after 2
> retransmissions, and try again.
>
> Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/477
> Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>

Reviewed-by: Eric Dumazet <edumazet@google.com>