1 files changed, 314 insertions, 162 deletions

diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c
index 54613f5b7521..f60f01b14fac 100644
--- a/net/mptcp/protocol.c
+++ b/net/mptcp/protocol.c
@@ -22,6 +22,7 @@
 #endif
 #include <net/mptcp.h>
 #include <net/xfrm.h>
+#include <asm/ioctls.h>
 #include "protocol.h"
 #include "mib.h"
 
@@ -46,9 +47,10 @@ struct mptcp_skb_cb {
 
 enum {
 	MPTCP_CMSG_TS = BIT(0),
+	MPTCP_CMSG_INQ = BIT(1),
 };
 
-static struct percpu_counter mptcp_sockets_allocated;
+static struct percpu_counter mptcp_sockets_allocated ____cacheline_aligned_in_smp;
 
 static void __mptcp_destroy_sock(struct sock *sk);
 static void __mptcp_check_send_data_fin(struct sock *sk);
@@ -738,6 +740,7 @@ static bool __mptcp_ofo_queue(struct mptcp_sock *msk)
 				 MPTCP_SKB_CB(skb)->map_seq, msk->ack_seq,
 				 delta);
 			MPTCP_SKB_CB(skb)->offset += delta;
+			MPTCP_SKB_CB(skb)->map_seq += delta;
 			__skb_queue_tail(&sk->sk_receive_queue, skb);
 		}
 		msk->ack_seq = end_seq;
@@ -760,7 +763,7 @@ static bool move_skbs_to_msk(struct mptcp_sock *msk, struct sock *ssk)
 		if (!sock_owned_by_user(sk))
 			__mptcp_error_report(sk);
 		else
-			set_bit(MPTCP_ERROR_REPORT,  &msk->flags);
+			__set_bit(MPTCP_ERROR_REPORT,  &msk->cb_flags);
 	}
 
 	/* If the moves have caught up with the DATA_FIN sequence number
@@ -805,47 +808,38 @@ void mptcp_data_ready(struct sock *sk, struct sock *ssk)
 	mptcp_data_unlock(sk);
 }
 
-static bool mptcp_do_flush_join_list(struct mptcp_sock *msk)
+static bool __mptcp_finish_join(struct mptcp_sock *msk, struct sock *ssk)
 {
-	struct mptcp_subflow_context *subflow;
-	bool ret = false;
+	struct sock *sk = (struct sock *)msk;
 
-	if (likely(list_empty(&msk->join_list)))
+	if (sk->sk_state != TCP_ESTABLISHED)
 		return false;
 
-	spin_lock_bh(&msk->join_list_lock);
-	list_for_each_entry(subflow, &msk->join_list, node) {
-		u32 sseq = READ_ONCE(subflow->setsockopt_seq);
-
-		mptcp_propagate_sndbuf((struct sock *)msk, mptcp_subflow_tcp_sock(subflow));
-		if (READ_ONCE(msk->setsockopt_seq) != sseq)
-			ret = true;
-	}
-	list_splice_tail_init(&msk->join_list, &msk->conn_list);
-	spin_unlock_bh(&msk->join_list_lock);
-
-	return ret;
-}
-
-void __mptcp_flush_join_list(struct mptcp_sock *msk)
-{
-	if (likely(!mptcp_do_flush_join_list(msk)))
-		return;
+	/* attach to msk socket only after we are sure we will deal with it
+	 * at close time
+	 */
+	if (sk->sk_socket && !ssk->sk_socket)
+		mptcp_sock_graft(ssk, sk->sk_socket);
 
-	if (!test_and_set_bit(MPTCP_WORK_SYNC_SETSOCKOPT, &msk->flags))
-		mptcp_schedule_work((struct sock *)msk);
+	mptcp_propagate_sndbuf((struct sock *)msk, ssk);
+	mptcp_sockopt_sync_locked(msk, ssk);
+	return true;
 }
 
-static void mptcp_flush_join_list(struct mptcp_sock *msk)
+static void __mptcp_flush_join_list(struct sock *sk)
 {
-	bool sync_needed = test_and_clear_bit(MPTCP_WORK_SYNC_SETSOCKOPT, &msk->flags);
-
-	might_sleep();
+	struct mptcp_subflow_context *tmp, *subflow;
+	struct mptcp_sock *msk = mptcp_sk(sk);
 
-	if (!mptcp_do_flush_join_list(msk) && !sync_needed)
-		return;
+	list_for_each_entry_safe(subflow, tmp, &msk->join_list, node) {
+		struct sock *ssk = mptcp_subflow_tcp_sock(subflow);
+		bool slow = lock_sock_fast(ssk);
 
-	mptcp_sockopt_sync_all(msk);
+		list_move_tail(&subflow->node, &msk->conn_list);
+		if (!__mptcp_finish_join(msk, ssk))
+			mptcp_subflow_reset(ssk);
+		unlock_sock_fast(ssk, slow);
+	}
 }
 
 static bool mptcp_timer_pending(struct sock *sk)
@@ -972,7 +966,9 @@ static void __mptcp_mem_reclaim_partial(struct sock *sk)
 
 	lockdep_assert_held_once(&sk->sk_lock.slock);
 
-	__mptcp_rmem_reclaim(sk, reclaimable - 1);
+	if (reclaimable > SK_MEM_QUANTUM)
+		__mptcp_rmem_reclaim(sk, reclaimable - 1);
+
 	sk_mem_reclaim_partial(sk);
 }
 
@@ -1369,7 +1365,7 @@ out:
 
 struct subflow_send_info {
 	struct sock *ssk;
-	u64 ratio;
+	u64 linger_time;
 };
 
 void mptcp_subflow_set_active(struct mptcp_subflow_context *subflow)
@@ -1394,20 +1390,24 @@ bool mptcp_subflow_active(struct mptcp_subflow_context *subflow)
 	return __mptcp_subflow_active(subflow);
 }
 
+#define SSK_MODE_ACTIVE	0
+#define SSK_MODE_BACKUP	1
+#define SSK_MODE_MAX	2
+
 /* implement the mptcp packet scheduler;
  * returns the subflow that will transmit the next DSS
  * additionally updates the rtx timeout
  */
 static struct sock *mptcp_subflow_get_send(struct mptcp_sock *msk)
 {
-	struct subflow_send_info send_info[2];
+	struct subflow_send_info send_info[SSK_MODE_MAX];
 	struct mptcp_subflow_context *subflow;
 	struct sock *sk = (struct sock *)msk;
+	u32 pace, burst, wmem;
 	int i, nr_active = 0;
 	struct sock *ssk;
+	u64 linger_time;
 	long tout = 0;
-	u64 ratio;
-	u32 pace;
 
 	sock_owned_by_me(sk);
 
@@ -1426,10 +1426,11 @@ static struct sock *mptcp_subflow_get_send(struct mptcp_sock *msk)
 	}
 
 	/* pick the subflow with the lower wmem/wspace ratio */
-	for (i = 0; i < 2; ++i) {
+	for (i = 0; i < SSK_MODE_MAX; ++i) {
 		send_info[i].ssk = NULL;
-		send_info[i].ratio = -1;
+		send_info[i].linger_time = -1;
 	}
+
 	mptcp_for_each_subflow(msk, subflow) {
 		trace_mptcp_subflow_get_send(subflow);
 		ssk =  mptcp_subflow_tcp_sock(subflow);
@@ -1438,34 +1439,51 @@ static struct sock *mptcp_subflow_get_send(struct mptcp_sock *msk)
 
 		tout = max(tout, mptcp_timeout_from_subflow(subflow));
 		nr_active += !subflow->backup;
-		if (!sk_stream_memory_free(subflow->tcp_sock) || !tcp_sk(ssk)->snd_wnd)
-			continue;
-
-		pace = READ_ONCE(ssk->sk_pacing_rate);
-		if (!pace)
-			continue;
+		pace = subflow->avg_pacing_rate;
+		if (unlikely(!pace)) {
+			/* init pacing rate from socket */
+			subflow->avg_pacing_rate = READ_ONCE(ssk->sk_pacing_rate);
+			pace = subflow->avg_pacing_rate;
+			if (!pace)
+				continue;
+		}
 
-		ratio = div_u64((u64)READ_ONCE(ssk->sk_wmem_queued) << 32,
-				pace);
-		if (ratio < send_info[subflow->backup].ratio) {
+		linger_time = div_u64((u64)READ_ONCE(ssk->sk_wmem_queued) << 32, pace);
+		if (linger_time < send_info[subflow->backup].linger_time) {
 			send_info[subflow->backup].ssk = ssk;
-			send_info[subflow->backup].ratio = ratio;
+			send_info[subflow->backup].linger_time = linger_time;
 		}
 	}
 	__mptcp_set_timeout(sk, tout);
 
 	/* pick the best backup if no other subflow is active */
 	if (!nr_active)
-		send_info[0].ssk = send_info[1].ssk;
-
-	if (send_info[0].ssk) {
-		msk->last_snd = send_info[0].ssk;
-		msk->snd_burst = min_t(int, MPTCP_SEND_BURST_SIZE,
-				       tcp_sk(msk->last_snd)->snd_wnd);
-		return msk->last_snd;
-	}
+		send_info[SSK_MODE_ACTIVE].ssk = send_info[SSK_MODE_BACKUP].ssk;
+
+	/* According to the blest algorithm, to avoid HoL blocking for the
+	 * faster flow, we need to:
+	 * - estimate the faster flow linger time
+	 * - use the above to estimate the amount of byte transferred
+	 *   by the faster flow
+	 * - check that the amount of queued data is greter than the above,
+	 *   otherwise do not use the picked, slower, subflow
+	 * We select the subflow with the shorter estimated time to flush
+	 * the queued mem, which basically ensure the above. We just need
+	 * to check that subflow has a non empty cwin.
+	 */
+	ssk = send_info[SSK_MODE_ACTIVE].ssk;
+	if (!ssk || !sk_stream_memory_free(ssk) || !tcp_sk(ssk)->snd_wnd)
+		return NULL;
 
-	return NULL;
+	burst = min_t(int, MPTCP_SEND_BURST_SIZE, tcp_sk(ssk)->snd_wnd);
+	wmem = READ_ONCE(ssk->sk_wmem_queued);
+	subflow = mptcp_subflow_ctx(ssk);
+	subflow->avg_pacing_rate = div_u64((u64)subflow->avg_pacing_rate * wmem +
+					   READ_ONCE(ssk->sk_pacing_rate) * burst,
+					   burst + wmem);
+	msk->last_snd = ssk;
+	msk->snd_burst = burst;
+	return ssk;
 }
 
 static void mptcp_push_release(struct sock *ssk, struct mptcp_sendmsg_info *info)
@@ -1499,11 +1517,10 @@ static void mptcp_update_post_push(struct mptcp_sock *msk,
 		msk->snd_nxt = snd_nxt_new;
 }
 
-static void mptcp_check_and_set_pending(struct sock *sk)
+void mptcp_check_and_set_pending(struct sock *sk)
 {
-	if (mptcp_send_head(sk) &&
-	    !test_bit(MPTCP_PUSH_PENDING, &mptcp_sk(sk)->flags))
-		set_bit(MPTCP_PUSH_PENDING, &mptcp_sk(sk)->flags);
+	if (mptcp_send_head(sk))
+		mptcp_sk(sk)->push_pending |= BIT(MPTCP_PUSH_PENDING);
 }
 
 void __mptcp_push_pending(struct sock *sk, unsigned int flags)
@@ -1524,7 +1541,6 @@ void __mptcp_push_pending(struct sock *sk, unsigned int flags)
 			int ret = 0;
 
 			prev_ssk = ssk;
-			__mptcp_flush_join_list(msk);
 			ssk = mptcp_subflow_get_send(msk);
 
 			/* First check. If the ssk has changed since
@@ -1784,8 +1800,10 @@ static int __mptcp_recvmsg_mskq(struct mptcp_sock *msk,
 		copied += count;
 
 		if (count < data_len) {
-			if (!(flags & MSG_PEEK))
+			if (!(flags & MSG_PEEK)) {
 				MPTCP_SKB_CB(skb)->offset += count;
+				MPTCP_SKB_CB(skb)->map_seq += count;
+			}
 			break;
 		}
 
@@ -1927,7 +1945,6 @@ static bool __mptcp_move_skbs(struct mptcp_sock *msk)
 	unsigned int moved = 0;
 	bool ret, done;
 
-	mptcp_flush_join_list(msk);
 	do {
 		struct sock *ssk = mptcp_subflow_recv_lookup(msk);
 		bool slowpath;
@@ -1965,6 +1982,27 @@ static bool __mptcp_move_skbs(struct mptcp_sock *msk)
 	return !skb_queue_empty(&msk->receive_queue);
 }
 
+static unsigned int mptcp_inq_hint(const struct sock *sk)
+{
+	const struct mptcp_sock *msk = mptcp_sk(sk);
+	const struct sk_buff *skb;
+
+	skb = skb_peek(&msk->receive_queue);
+	if (skb) {
+		u64 hint_val = msk->ack_seq - MPTCP_SKB_CB(skb)->map_seq;
+
+		if (hint_val >= INT_MAX)
+			return INT_MAX;
+
+		return (unsigned int)hint_val;
+	}
+
+	if (sk->sk_state == TCP_CLOSE || (sk->sk_shutdown & RCV_SHUTDOWN))
+		return 1;
+
+	return 0;
+}
+
 static int mptcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len,
 			 int nonblock, int flags, int *addr_len)
 {
@@ -1989,6 +2027,9 @@ static int mptcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len,
 	len = min_t(size_t, len, INT_MAX);
 	target = sock_rcvlowat(sk, flags & MSG_WAITALL, len);
 
+	if (unlikely(msk->recvmsg_inq))
+		cmsg_flags = MPTCP_CMSG_INQ;
+
 	while (copied < len) {
 		int bytes_read;
 
@@ -2062,6 +2103,12 @@ out_err:
 	if (cmsg_flags && copied >= 0) {
 		if (cmsg_flags & MPTCP_CMSG_TS)
 			tcp_recv_timestamp(msg, sk, &tss);
+
+		if (cmsg_flags & MPTCP_CMSG_INQ) {
+			unsigned int inq = mptcp_inq_hint(sk);
+
+			put_cmsg(msg, SOL_TCP, TCP_CM_INQ, sizeof(inq), &inq);
+		}
 	}
 
 	pr_debug("msk=%p rx queue empty=%d:%d copied=%d",
@@ -2088,7 +2135,7 @@ static void mptcp_retransmit_timer(struct timer_list *t)
 			mptcp_schedule_work(sk);
 	} else {
 		/* delegate our work to tcp_release_cb() */
-		set_bit(MPTCP_RETRANSMIT, &msk->flags);
+		__set_bit(MPTCP_RETRANSMIT, &msk->cb_flags);
 	}
 	bh_unlock_sock(sk);
 	sock_put(sk);
@@ -2196,6 +2243,10 @@ bool __mptcp_retransmit_pending_data(struct sock *sk)
 	return true;
 }
 
+/* flags for __mptcp_close_ssk() */
+#define MPTCP_CF_PUSH		BIT(1)
+#define MPTCP_CF_FASTCLOSE	BIT(2)
+
 /* subflow sockets can be either outgoing (connect) or incoming
  * (accept).
  *
@@ -2205,22 +2256,37 @@ bool __mptcp_retransmit_pending_data(struct sock *sk)
  * parent socket.
  */
 static void __mptcp_close_ssk(struct sock *sk, struct sock *ssk,
-			      struct mptcp_subflow_context *subflow)
+			      struct mptcp_subflow_context *subflow,
+			      unsigned int flags)
 {
 	struct mptcp_sock *msk = mptcp_sk(sk);
-	bool need_push;
+	bool need_push, dispose_it;
 
-	list_del(&subflow->node);
+	dispose_it = !msk->subflow || ssk != msk->subflow->sk;
+	if (dispose_it)
+		list_del(&subflow->node);
 
 	lock_sock_nested(ssk, SINGLE_DEPTH_NESTING);
 
+	if (flags & MPTCP_CF_FASTCLOSE)
+		subflow->send_fastclose = 1;
+
+	need_push = (flags & MPTCP_CF_PUSH) && __mptcp_retransmit_pending_data(sk);
+	if (!dispose_it) {
+		tcp_disconnect(ssk, 0);
+		msk->subflow->state = SS_UNCONNECTED;
+		mptcp_subflow_ctx_reset(subflow);
+		release_sock(ssk);
+
+		goto out;
+	}
+
 	/* if we are invoked by the msk cleanup code, the subflow is
 	 * already orphaned
 	 */
 	if (ssk->sk_socket)
 		sock_orphan(ssk);
 
-	need_push = __mptcp_retransmit_pending_data(sk);
 	subflow->disposable = 1;
 
 	/* if ssk hit tcp_done(), tcp_cleanup_ulp() cleared the related ops
@@ -2240,14 +2306,12 @@ static void __mptcp_close_ssk(struct sock *sk, struct sock *ssk,
 
 	sock_put(ssk);
 
-	if (ssk == msk->last_snd)
-		msk->last_snd = NULL;
-
 	if (ssk == msk->first)
 		msk->first = NULL;
 
-	if (msk->subflow && ssk == msk->subflow->sk)
-		mptcp_dispose_initial_subflow(msk);
+out:
+	if (ssk == msk->last_snd)
+		msk->last_snd = NULL;
 
 	if (need_push)
 		__mptcp_push_pending(sk, 0);
@@ -2258,7 +2322,13 @@ void mptcp_close_ssk(struct sock *sk, struct sock *ssk,
 {
 	if (sk->sk_state == TCP_ESTABLISHED)
 		mptcp_event(MPTCP_EVENT_SUB_CLOSED, mptcp_sk(sk), ssk, GFP_KERNEL);
-	__mptcp_close_ssk(sk, ssk, subflow);
+
+	/* subflow aborted before reaching the fully_established status
+	 * attempt the creation of the next subflow
+	 */
+	mptcp_pm_subflow_check_next(mptcp_sk(sk), ssk, subflow);
+
+	__mptcp_close_ssk(sk, ssk, subflow, MPTCP_CF_PUSH);
 }
 
 static unsigned int mptcp_sync_mss(struct sock *sk, u32 pmtu)
@@ -2410,12 +2480,10 @@ static void mptcp_worker(struct work_struct *work)
 		goto unlock;
 
 	mptcp_check_data_fin_ack(sk);
-	mptcp_flush_join_list(msk);
 
 	mptcp_check_fastclose(msk);
 
-	if (msk->pm.status)
-		mptcp_pm_nl_work(msk);
+	mptcp_pm_nl_work(msk);
 
 	if (test_and_clear_bit(MPTCP_WORK_EOF, &msk->flags))
 		mptcp_check_for_eof(msk);
@@ -2449,8 +2517,6 @@ static int __mptcp_init_sock(struct sock *sk)
 {
 	struct mptcp_sock *msk = mptcp_sk(sk);
 
-	spin_lock_init(&msk->join_list_lock);
-
 	INIT_LIST_HEAD(&msk->conn_list);
 	INIT_LIST_HEAD(&msk->join_list);
 	INIT_LIST_HEAD(&msk->rtx_queue);
@@ -2476,9 +2542,20 @@ static int __mptcp_init_sock(struct sock *sk)
 	return 0;
 }
 
-static int mptcp_init_sock(struct sock *sk)
+static void mptcp_ca_reset(struct sock *sk)
 {
 	struct inet_connection_sock *icsk = inet_csk(sk);
+
+	tcp_assign_congestion_control(sk);
+	strcpy(mptcp_sk(sk)->ca_name, icsk->icsk_ca_ops->name);
+
+	/* no need to keep a reference to the ops, the name will suffice */
+	tcp_cleanup_congestion_control(sk);
+	icsk->icsk_ca_ops = NULL;
+}
+
+static int mptcp_init_sock(struct sock *sk)
+{
 	struct net *net = sock_net(sk);
 	int ret;
 
@@ -2499,12 +2576,7 @@ static int mptcp_init_sock(struct sock *sk)
 	/* fetch the ca name; do it outside __mptcp_init_sock(), so that clone will
 	 * propagate the correct value
 	 */
-	tcp_assign_congestion_control(sk);
-	strcpy(mptcp_sk(sk)->ca_name, icsk->icsk_ca_ops->name);
-
-	/* no need to keep a reference to the ops, the name will suffice */
-	tcp_cleanup_congestion_control(sk);
-	icsk->icsk_ca_ops = NULL;
+	mptcp_ca_reset(sk);
 
 	sk_sockets_allocated_inc(sk);
 	sk->sk_rcvbuf = sock_net(sk)->ipv4.sysctl_tcp_rmem[1];
@@ -2609,6 +2681,7 @@ static void __mptcp_check_send_data_fin(struct sock *sk)
 	 * state now
 	 */
 	if (__mptcp_check_fallback(msk)) {
+		WRITE_ONCE(msk->snd_una, msk->write_seq);
 		if ((1 << sk->sk_state) & (TCPF_CLOSING | TCPF_LAST_ACK)) {
 			inet_sk_state_store(sk, TCP_CLOSE);
 			mptcp_close_wake_up(sk);
@@ -2617,7 +2690,6 @@ static void __mptcp_check_send_data_fin(struct sock *sk)
 		}
 	}
 
-	mptcp_flush_join_list(msk);
 	mptcp_for_each_subflow(msk, subflow) {
 		struct sock *tcp_sk = mptcp_subflow_tcp_sock(subflow);
 
@@ -2650,21 +2722,20 @@ static void __mptcp_destroy_sock(struct sock *sk)
 
 	might_sleep();
 
-	/* be sure to always acquire the join list lock, to sync vs
-	 * mptcp_finish_join().
-	 */
-	spin_lock_bh(&msk->join_list_lock);
-	list_splice_tail_init(&msk->join_list, &msk->conn_list);
-	spin_unlock_bh(&msk->join_list_lock);
+	/* join list will be eventually flushed (with rst) at sock lock release time*/
 	list_splice_init(&msk->conn_list, &conn_list);
 
 	sk_stop_timer(sk, &msk->sk.icsk_retransmit_timer);
 	sk_stop_timer(sk, &sk->sk_timer);
 	msk->pm.status = 0;
 
+	/* clears msk->subflow, allowing the following loop to close
+	 * even the initial subflow
+	 */
+	mptcp_dispose_initial_subflow(msk);
 	list_for_each_entry_safe(subflow, tmp, &conn_list, node) {
 		struct sock *ssk = mptcp_subflow_tcp_sock(subflow);
-		__mptcp_close_ssk(sk, ssk, subflow);
+		__mptcp_close_ssk(sk, ssk, subflow, 0);
 	}
 
 	sk->sk_prot->destroy(sk);
@@ -2675,7 +2746,6 @@ static void __mptcp_destroy_sock(struct sock *sk)
 	xfrm_sk_free_policy(sk);
 
 	sk_refcnt_debug_release(sk);
-	mptcp_dispose_initial_subflow(msk);
 	sock_put(sk);
 }
 
@@ -2711,6 +2781,9 @@ cleanup:
 
 	sock_hold(sk);
 	pr_debug("msk=%p state=%d", sk, sk->sk_state);
+	if (mptcp_sk(sk)->token)
+		mptcp_event(MPTCP_EVENT_CLOSED, mptcp_sk(sk), NULL, GFP_KERNEL);
+
 	if (sk->sk_state == TCP_CLOSE) {
 		__mptcp_destroy_sock(sk);
 		do_cancel_work = true;
@@ -2721,9 +2794,6 @@ cleanup:
 	if (do_cancel_work)
 		mptcp_cancel_work(sk);
 
-	if (mptcp_sk(sk)->token)
-		mptcp_event(MPTCP_EVENT_CLOSED, mptcp_sk(sk), NULL, GFP_KERNEL);
-
 	sock_put(sk);
 }
 
@@ -2755,15 +2825,38 @@ static int mptcp_disconnect(struct sock *sk, int flags)
 	struct mptcp_subflow_context *subflow;
 	struct mptcp_sock *msk = mptcp_sk(sk);
 
-	mptcp_do_flush_join_list(msk);
+	inet_sk_state_store(sk, TCP_CLOSE);
 
 	mptcp_for_each_subflow(msk, subflow) {
 		struct sock *ssk = mptcp_subflow_tcp_sock(subflow);
 
-		lock_sock(ssk);
-		tcp_disconnect(ssk, flags);
-		release_sock(ssk);
+		__mptcp_close_ssk(sk, ssk, subflow, MPTCP_CF_FASTCLOSE);
 	}
+
+	sk_stop_timer(sk, &msk->sk.icsk_retransmit_timer);
+	sk_stop_timer(sk, &sk->sk_timer);
+
+	if (mptcp_sk(sk)->token)
+		mptcp_event(MPTCP_EVENT_CLOSED, mptcp_sk(sk), NULL, GFP_KERNEL);
+
+	mptcp_destroy_common(msk);
+	msk->last_snd = NULL;
+	WRITE_ONCE(msk->flags, 0);
+	msk->cb_flags = 0;
+	msk->push_pending = 0;
+	msk->recovery = false;
+	msk->can_ack = false;
+	msk->fully_established = false;
+	msk->rcv_data_fin = false;
+	msk->snd_data_fin_enable = false;
+	msk->rcv_fastclose = false;
+	msk->use_64bit_ack = false;
+	WRITE_ONCE(msk->csum_enabled, mptcp_is_checksum_enabled(sock_net(sk)));
+	mptcp_pm_data_reset(msk);
+	mptcp_ca_reset(sk);
+
+	sk->sk_shutdown = 0;
+	sk_error_report(sk);
 	return 0;
 }
 
@@ -2903,9 +2996,11 @@ void mptcp_destroy_common(struct mptcp_sock *msk)
 	__mptcp_clear_xmit(sk);
 
 	/* move to sk_receive_queue, sk_stream_kill_queues will purge it */
+	mptcp_data_lock(sk);
 	skb_queue_splice_tail_init(&msk->receive_queue, &sk->sk_receive_queue);
 	__skb_queue_purge(&sk->sk_receive_queue);
 	skb_rbtree_purge(&msk->out_of_order_queue);
+	mptcp_data_unlock(sk);
 
 	/* move all the rx fwd alloc into the sk_mem_reclaim_final in
 	 * inet_sock_destruct() will dispose it
@@ -2929,7 +3024,7 @@ void __mptcp_data_acked(struct sock *sk)
 	if (!sock_owned_by_user(sk))
 		__mptcp_clean_una(sk);
 	else
-		set_bit(MPTCP_CLEAN_UNA, &mptcp_sk(sk)->flags);
+		__set_bit(MPTCP_CLEAN_UNA, &mptcp_sk(sk)->cb_flags);
 
 	if (mptcp_pending_data_fin_ack(sk))
 		mptcp_schedule_work(sk);
@@ -2948,20 +3043,23 @@ void __mptcp_check_push(struct sock *sk, struct sock *ssk)
 		else if (xmit_ssk)
 			mptcp_subflow_delegate(mptcp_subflow_ctx(xmit_ssk), MPTCP_DELEGATE_SEND);
 	} else {
-		set_bit(MPTCP_PUSH_PENDING, &mptcp_sk(sk)->flags);
+		__set_bit(MPTCP_PUSH_PENDING, &mptcp_sk(sk)->cb_flags);
 	}
 }
 
+#define MPTCP_FLAGS_PROCESS_CTX_NEED (BIT(MPTCP_PUSH_PENDING) | \
+				      BIT(MPTCP_RETRANSMIT) | \
+				      BIT(MPTCP_FLUSH_JOIN_LIST))
+
 /* processes deferred events and flush wmem */
 static void mptcp_release_cb(struct sock *sk)
+	__must_hold(&sk->sk_lock.slock)
 {
-	for (;;) {
-		unsigned long flags = 0;
+	struct mptcp_sock *msk = mptcp_sk(sk);
 
-		if (test_and_clear_bit(MPTCP_PUSH_PENDING, &mptcp_sk(sk)->flags))
-			flags |= BIT(MPTCP_PUSH_PENDING);
-		if (test_and_clear_bit(MPTCP_RETRANSMIT, &mptcp_sk(sk)->flags))
-			flags |= BIT(MPTCP_RETRANSMIT);
+	for (;;) {
+		unsigned long flags = (msk->cb_flags & MPTCP_FLAGS_PROCESS_CTX_NEED) |
+				      msk->push_pending;
 		if (!flags)
 			break;
 
@@ -2972,8 +3070,11 @@ static void mptcp_release_cb(struct sock *sk)
 		 *    datapath acquires the msk socket spinlock while helding
 		 *    the subflow socket lock
 		 */
-
+		msk->push_pending = 0;
+		msk->cb_flags &= ~flags;
 		spin_unlock_bh(&sk->sk_lock.slock);
+		if (flags & BIT(MPTCP_FLUSH_JOIN_LIST))
+			__mptcp_flush_join_list(sk);
 		if (flags & BIT(MPTCP_PUSH_PENDING))
 			__mptcp_push_pending(sk, 0);
 		if (flags & BIT(MPTCP_RETRANSMIT))
@@ -2986,11 +3087,11 @@ static void mptcp_release_cb(struct sock *sk)
 	/* be sure to set the current sk state before tacking actions
 	 * depending on sk_state
 	 */
-	if (test_and_clear_bit(MPTCP_CONNECTED, &mptcp_sk(sk)->flags))
+	if (__test_and_clear_bit(MPTCP_CONNECTED, &msk->cb_flags))
 		__mptcp_set_connected(sk);
-	if (test_and_clear_bit(MPTCP_CLEAN_UNA, &mptcp_sk(sk)->flags))
+	if (__test_and_clear_bit(MPTCP_CLEAN_UNA, &msk->cb_flags))
 		__mptcp_clean_una_wakeup(sk);
-	if (test_and_clear_bit(MPTCP_ERROR_REPORT, &mptcp_sk(sk)->flags))
+	if (__test_and_clear_bit(MPTCP_ERROR_REPORT, &msk->cb_flags))
 		__mptcp_error_report(sk);
 
 	__mptcp_update_rmem(sk);
@@ -3032,7 +3133,7 @@ void mptcp_subflow_process_delegated(struct sock *ssk)
 		if (!sock_owned_by_user(sk))
 			__mptcp_subflow_push_pending(sk, ssk);
 		else
-			set_bit(MPTCP_PUSH_PENDING, &mptcp_sk(sk)->flags);
+			__set_bit(MPTCP_PUSH_PENDING, &mptcp_sk(sk)->cb_flags);
 		mptcp_data_unlock(sk);
 		mptcp_subflow_delegated_done(subflow, MPTCP_DELEGATE_SEND);
 	}
@@ -3118,8 +3219,7 @@ bool mptcp_finish_join(struct sock *ssk)
 	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(ssk);
 	struct mptcp_sock *msk = mptcp_sk(subflow->conn);
 	struct sock *parent = (void *)msk;
-	struct socket *parent_sock;
-	bool ret;
+	bool ret = true;
 
 	pr_debug("msk=%p, subflow=%p", msk, subflow);
 
@@ -3132,35 +3232,38 @@ bool mptcp_finish_join(struct sock *ssk)
 	if (!msk->pm.server_side)
 		goto out;
 
-	if (!mptcp_pm_allow_new_subflow(msk)) {
-		subflow->reset_reason = MPTCP_RST_EPROHIBIT;
-		return false;
-	}
+	if (!mptcp_pm_allow_new_subflow(msk))
+		goto err_prohibited;
+
+	if (WARN_ON_ONCE(!list_empty(&subflow->node)))
+		goto err_prohibited;
 
-	/* active connections are already on conn_list, and we can't acquire
-	 * msk lock here.
-	 * use the join list lock as synchronization point and double-check
-	 * msk status to avoid racing with __mptcp_destroy_sock()
+	/* active connections are already on conn_list.
+	 * If we can't acquire msk socket lock here, let the release callback
+	 * handle it
 	 */
-	spin_lock_bh(&msk->join_list_lock);
-	ret = inet_sk_state_load(parent) == TCP_ESTABLISHED;
-	if (ret && !WARN_ON_ONCE(!list_empty(&subflow->node))) {
-		list_add_tail(&subflow->node, &msk->join_list);
+	mptcp_data_lock(parent);
+	if (!sock_owned_by_user(parent)) {
+		ret = __mptcp_finish_join(msk, ssk);
+		if (ret) {
+			sock_hold(ssk);
+			list_add_tail(&subflow->node, &msk->conn_list);
+		}
+	} else {
 		sock_hold(ssk);
+		list_add_tail(&subflow->node, &msk->join_list);
+		__set_bit(MPTCP_FLUSH_JOIN_LIST, &msk->cb_flags);
 	}
-	spin_unlock_bh(&msk->join_list_lock);
+	mptcp_data_unlock(parent);
+
 	if (!ret) {
+err_prohibited:
 		subflow->reset_reason = MPTCP_RST_EPROHIBIT;
 		return false;
 	}
 
-	/* attach to msk socket only after we are sure he will deal with us
-	 * at close time
-	 */
-	parent_sock = READ_ONCE(parent->sk_socket);
-	if (parent_sock && !ssk->sk_socket)
-		mptcp_sock_graft(ssk, parent_sock);
 	subflow->map_seq = READ_ONCE(msk->ack_seq);
+
 out:
 	mptcp_event(MPTCP_EVENT_SUB_ESTABLISHED, msk, ssk, GFP_ATOMIC);
 	return true;
@@ -3179,6 +3282,57 @@ static int mptcp_forward_alloc_get(const struct sock *sk)
 	return sk->sk_forward_alloc + mptcp_sk(sk)->rmem_fwd_alloc;
 }
 
+static int mptcp_ioctl_outq(const struct mptcp_sock *msk, u64 v)
+{
+	const struct sock *sk = (void *)msk;
+	u64 delta;
+
+	if (sk->sk_state == TCP_LISTEN)
+		return -EINVAL;
+
+	if ((1 << sk->sk_state) & (TCPF_SYN_SENT | TCPF_SYN_RECV))
+		return 0;
+
+	delta = msk->write_seq - v;
+	if (delta > INT_MAX)
+		delta = INT_MAX;
+
+	return (int)delta;
+}
+
+static int mptcp_ioctl(struct sock *sk, int cmd, unsigned long arg)
+{
+	struct mptcp_sock *msk = mptcp_sk(sk);
+	bool slow;
+	int answ;
+
+	switch (cmd) {
+	case SIOCINQ:
+		if (sk->sk_state == TCP_LISTEN)
+			return -EINVAL;
+
+		lock_sock(sk);
+		__mptcp_move_skbs(msk);
+		answ = mptcp_inq_hint(sk);
+		release_sock(sk);
+		break;
+	case SIOCOUTQ:
+		slow = lock_sock_fast(sk);
+		answ = mptcp_ioctl_outq(msk, READ_ONCE(msk->snd_una));
+		unlock_sock_fast(sk, slow);
+		break;
+	case SIOCOUTQNSD:
+		slow = lock_sock_fast(sk);
+		answ = mptcp_ioctl_outq(msk, msk->snd_nxt);
+		unlock_sock_fast(sk, slow);
+		break;
+	default:
+		return -ENOIOCTLCMD;
+	}
+
+	return put_user(answ, (int __user *)arg);
+}
+
 static struct proto mptcp_prot = {
 	.name		= "MPTCP",
 	.owner		= THIS_MODULE,
@@ -3191,6 +3345,7 @@ static struct proto mptcp_prot = {
 	.shutdown	= mptcp_shutdown,
 	.destroy	= mptcp_destroy,
 	.sendmsg	= mptcp_sendmsg,
+	.ioctl		= mptcp_ioctl,
 	.recvmsg	= mptcp_recvmsg,
 	.release_cb	= mptcp_release_cb,
 	.hash		= mptcp_hash,
@@ -3243,9 +3398,20 @@ static int mptcp_stream_connect(struct socket *sock, struct sockaddr *uaddr,
 	struct mptcp_sock *msk = mptcp_sk(sock->sk);
 	struct mptcp_subflow_context *subflow;
 	struct socket *ssock;
-	int err;
+	int err = -EINVAL;
 
 	lock_sock(sock->sk);
+	if (uaddr) {
+		if (addr_len < sizeof(uaddr->sa_family))
+			goto unlock;
+
+		if (uaddr->sa_family == AF_UNSPEC) {
+			err = mptcp_disconnect(sock->sk, flags);
+			sock->state = err ? SS_DISCONNECTING : SS_UNCONNECTED;
+			goto unlock;
+		}
+	}
+
 	if (sock->state != SS_UNCONNECTED && msk->subflow) {
 		/* pending connection or invalid state, let existing subflow
 		 * cope with that
@@ -3255,10 +3421,8 @@ static int mptcp_stream_connect(struct socket *sock, struct sockaddr *uaddr,
 	}
 
 	ssock = __mptcp_nmpc_socket(msk);
-	if (!ssock) {
-		err = -EINVAL;
+	if (!ssock)
 		goto unlock;
-	}
 
 	mptcp_token_destroy(msk);
 	inet_sk_state_store(sock->sk, TCP_SYN_SENT);
@@ -3332,17 +3496,9 @@ static int mptcp_stream_accept(struct socket *sock, struct socket *newsock,
 
 	pr_debug("msk=%p", msk);
 
-	lock_sock(sock->sk);
-	if (sock->sk->sk_state != TCP_LISTEN)
-		goto unlock_fail;
-
 	ssock = __mptcp_nmpc_socket(msk);
 	if (!ssock)
-		goto unlock_fail;
-
-	clear_bit(MPTCP_DATA_READY, &msk->flags);
-	sock_hold(ssock->sk);
-	release_sock(sock->sk);
+		return -EINVAL;
 
 	err = ssock->ops->accept(sock, newsock, flags, kern);
 	if (err == 0 && !mptcp_is_tcpsk(newsock->sk)) {
@@ -3372,7 +3528,6 @@ static int mptcp_stream_accept(struct socket *sock, struct socket *newsock,
 		/* set ssk->sk_socket of accept()ed flows to mptcp socket.
 		 * This is needed so NOSPACE flag can be set from tcp stack.
 		 */
-		mptcp_flush_join_list(msk);
 		mptcp_for_each_subflow(msk, subflow) {
 			struct sock *ssk = mptcp_subflow_tcp_sock(subflow);
 
@@ -3382,14 +3537,7 @@ static int mptcp_stream_accept(struct socket *sock, struct socket *newsock,
 		release_sock(newsk);
 	}
 
-	if (inet_csk_listen_poll(ssock->sk))
-		set_bit(MPTCP_DATA_READY, &msk->flags);
-	sock_put(ssock->sk);
 	return err;
-
-unlock_fail:
-	release_sock(sock->sk);
-	return -EINVAL;
 }
 
 static __poll_t mptcp_check_readable(struct mptcp_sock *msk)
@@ -3435,8 +3583,12 @@ static __poll_t mptcp_poll(struct file *file, struct socket *sock,
 
 	state = inet_sk_state_load(sk);
 	pr_debug("msk=%p state=%d flags=%lx", msk, state, msk->flags);
-	if (state == TCP_LISTEN)
-		return test_bit(MPTCP_DATA_READY, &msk->flags) ? EPOLLIN | EPOLLRDNORM : 0;
+	if (state == TCP_LISTEN) {
+		if (WARN_ON_ONCE(!msk->subflow || !msk->subflow->sk))
+			return 0;
+
+		return inet_csk_listen_poll(msk->subflow->sk);
+	}
 
 	if (state != TCP_SYN_SENT && state != TCP_SYN_RECV) {
 		mask |= mptcp_check_readable(msk);