Merge branch 'mptcp-more-fixes-for-v6-5'

Matthieu Baerts says:

====================
mptcp: more fixes for v6.5

Here is a new batch of fixes related to MPTCP for v6.5 and older.

Patches 1 and 2 fix issues with MPTCP Join selftest when manually
launched with '-i' parameter to use 'ip mptcp' tool instead of the
dedicated one (pm_nl_ctl). The issues have been there since v5.18.

Thank you Andrea for your first contributions to MPTCP code in the
upstream kernel!

Patch 3 avoids corrupting the data stream when trying to reset
connections that have fallen back to TCP. This can happen from v6.1.

Patch 4 fixes a race when doing a disconnect() and an accept() in
parallel on a listener socket. The issue only happens in rare cases if
the user is really unlucky since a fix that landed in v6.3 but
backported up to v6.1.
====================

Link: https://lore.kernel.org/r/20230803-upstream-net-20230803-misc-fixes-6-5-v1-0-6671b1ab11cc@tessares.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

4 files changed, 34 insertions, 33 deletions

diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c
index be5f6f4b94c7..d80658547836 100644
--- a/net/mptcp/protocol.c
+++ b/net/mptcp/protocol.c
@@ -2335,7 +2335,7 @@ static void __mptcp_close_ssk(struct sock *sk, struct sock *ssk,
 
 	lock_sock_nested(ssk, SINGLE_DEPTH_NESTING);
 
-	if (flags & MPTCP_CF_FASTCLOSE) {
+	if ((flags & MPTCP_CF_FASTCLOSE) && !__mptcp_check_fallback(msk)) {
 		/* be sure to force the tcp_disconnect() path,
 		 * to generate the egress reset
 		 */
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index 37fbe22e2433..ba2a873a4d2e 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -325,7 +325,6 @@ struct mptcp_sock {
 	u32		subflow_id;
 	u32		setsockopt_seq;
 	char		ca_name[TCP_CA_NAME_MAX];
-	struct mptcp_sock	*dl_next;
 };
 
 #define mptcp_data_lock(sk) spin_lock_bh(&(sk)->sk_lock.slock)
diff --git a/net/mptcp/subflow.c b/net/mptcp/subflow.c
index 9ee3b7abbaf6..94ae7dd01c65 100644
--- a/net/mptcp/subflow.c
+++ b/net/mptcp/subflow.c
@@ -1793,16 +1793,31 @@ static void subflow_state_change(struct sock *sk)
 void mptcp_subflow_queue_clean(struct sock *listener_sk, struct sock *listener_ssk)
 {
 	struct request_sock_queue *queue = &inet_csk(listener_ssk)->icsk_accept_queue;
-	struct mptcp_sock *msk, *next, *head = NULL;
-	struct request_sock *req;
-	struct sock *sk;
+	struct request_sock *req, *head, *tail;
+	struct mptcp_subflow_context *subflow;
+	struct sock *sk, *ssk;
 
-	/* build a list of all unaccepted mptcp sockets */
+	/* Due to lock dependencies no relevant lock can be acquired under rskq_lock.
+	 * Splice the req list, so that accept() can not reach the pending ssk after
+	 * the listener socket is released below.
+	 */
 	spin_lock_bh(&queue->rskq_lock);
-	for (req = queue->rskq_accept_head; req; req = req->dl_next) {
-		struct mptcp_subflow_context *subflow;
-		struct sock *ssk = req->sk;
+	head = queue->rskq_accept_head;
+	tail = queue->rskq_accept_tail;
+	queue->rskq_accept_head = NULL;
+	queue->rskq_accept_tail = NULL;
+	spin_unlock_bh(&queue->rskq_lock);
+
+	if (!head)
+		return;
 
+	/* can't acquire the msk socket lock under the subflow one,
+	 * or will cause ABBA deadlock
+	 */
+	release_sock(listener_ssk);
+
+	for (req = head; req; req = req->dl_next) {
+		ssk = req->sk;
 		if (!sk_is_mptcp(ssk))
 			continue;
 
@@ -1810,32 +1825,10 @@ void mptcp_subflow_queue_clean(struct sock *listener_sk, struct sock *listener_s
 		if (!subflow || !subflow->conn)
 			continue;
 
-		/* skip if already in list */
 		sk = subflow->conn;
-		msk = mptcp_sk(sk);
-		if (msk->dl_next || msk == head)
-			continue;
-
 		sock_hold(sk);
-		msk->dl_next = head;
-		head = msk;
-	}
-	spin_unlock_bh(&queue->rskq_lock);
-	if (!head)
-		return;
-
-	/* can't acquire the msk socket lock under the subflow one,
-	 * or will cause ABBA deadlock
-	 */
-	release_sock(listener_ssk);
-
-	for (msk = head; msk; msk = next) {
-		sk = (struct sock *)msk;
 
 		lock_sock_nested(sk, SINGLE_DEPTH_NESTING);
-		next = msk->dl_next;
-		msk->dl_next = NULL;
-
 		__mptcp_unaccepted_force_close(sk);
 		release_sock(sk);
 
@@ -1859,6 +1852,13 @@ void mptcp_subflow_queue_clean(struct sock *listener_sk, struct sock *listener_s
 
 	/* we are still under the listener msk socket lock */
 	lock_sock_nested(listener_ssk, SINGLE_DEPTH_NESTING);
+
+	/* restore the listener queue, to let the TCP code clean it up */
+	spin_lock_bh(&queue->rskq_lock);
+	WARN_ON_ONCE(queue->rskq_accept_head);
+	queue->rskq_accept_head = head;
+	queue->rskq_accept_tail = tail;
+	spin_unlock_bh(&queue->rskq_lock);
 }
 
 static int subflow_ulp_init(struct sock *sk)
diff --git a/tools/testing/selftests/net/mptcp/mptcp_join.sh b/tools/testing/selftests/net/mptcp/mptcp_join.sh
index 3c2096ac97ef..d01b73a8ed0f 100755
--- a/tools/testing/selftests/net/mptcp/mptcp_join.sh
+++ b/tools/testing/selftests/net/mptcp/mptcp_join.sh
@@ -705,6 +705,7 @@ pm_nl_del_endpoint()
 	local addr=$3
 
 	if [ $ip_mptcp -eq 1 ]; then
+		[ $id -ne 0 ] && addr=''
 		ip -n $ns mptcp endpoint delete id $id $addr
 	else
 		ip netns exec $ns ./pm_nl_ctl del $id $addr
@@ -795,10 +796,11 @@ pm_nl_check_endpoint()
 	fi
 
 	if [ $ip_mptcp -eq 1 ]; then
+		# get line and trim trailing whitespace
 		line=$(ip -n $ns mptcp endpoint show $id)
+		line="${line% }"
 		# the dump order is: address id flags port dev
-		expected_line="$addr"
-		[ -n "$addr" ] && expected_line="$expected_line $addr"
+		[ -n "$addr" ] && expected_line="$addr"
 		expected_line="$expected_line $id"
 		[ -n "$_flags" ] && expected_line="$expected_line ${_flags//","/" "}"
 		[ -n "$dev" ] && expected_line="$expected_line $dev"