Skip to contentSkip to conversation

mptcp: re-enable snd buf autotune

PATCH In review Paolo Abeni 3 participants 7 messages Jan 12, 2021 – Jan 13, 2021

Subsystems (2)
lore
Share & tools
lore

Paolo Abeni

author
#0 lore raw

v3 cover mptcp: re-enable snd buf autotune

WRT to the previous iteration, this clean-up patch 2/4 dropping the left over 'use_backup' field, as suggested by Mat. No other change

Relevant bits from the previous cover letter:

There are still some sporadic self-test failures for assimetric links. Looks like they are caused by retransmissions/receiver dropping packets due to MPTCP level OoO check.

I'll try to address that asap, but I still consider this series for inclusion, so that it can stage a bit in the export branch.

Note: the last patch re-introduces the workqueue usage for the datapath (yes, it makes me feel sick, sad and tired :(. I was unable to avoid it.

Paolo Abeni (4):
mptcp: always graft subflow socket to parent.
mptcp: re-enable sndbuf autotune
mptcp: do not queue excessive data on subflows.
mptcp: schedule work for better snd subflow selection

 net/mptcp/protocol.c | 104 +++++++++++++++++--------------------------
 net/mptcp/protocol.h |  20 +++++++++
 net/mptcp/subflow.c  |  13 +++++-
 3 files changed, 73 insertions(+), 64 deletions(-)
Signature · 1 line
2.26.2
author
#1 lore raw

PATCH v3 1/4 mptcp: always graft subflow socket to parent.

Currently incoming subflow uses link to the parent socket, while outgoing one link to a per subflow socket. The latter is not really needed, except at initial connect() time and for the first subflow.

Always graft the outgoing subflow to the parent socket and free the unneeded ones early.

This allows some code cleanup, reduces the amount of memory used and will simplify the next patch

Signed-off-by Paolo Abeni <pabeni(a)redhat.com>
---
 net/mptcp/protocol.c | 36 ++++++++++--------------------------
 net/mptcp/protocol.h |  1 +
 net/mptcp/subflow.c  |  3 +++
 3 files changed, 14 insertions(+), 26 deletions(-)
net/mptcp/protocol.c +10 -26
index 5977e9e083be..ada69495a51b 100644
@@ -114,11 +114,7 @@ static int __mptcp_socket_create(struct mptcp_sock *msk)
 	list_add(&subflow->node, &msk->conn_list);
 	sock_hold(ssock->sk);
 	subflow->request_mptcp = 1;
-
-	/* accept() will wait on first subflow sk_wq, and we always wakes up
-	 * via msk->sk_socket
-	 */
-	RCU_INIT_POINTER(msk->first->sk_wq, &sk->sk_socket->wq);
+	mptcp_sock_graft(msk->first, sk->sk_socket);
 
 	return 0;
 }
@@ -2116,9 +2112,6 @@ static struct sock *mptcp_subflow_get_retrans(const struct mptcp_sock *msk)
 void __mptcp_close_ssk(struct sock *sk, struct sock *ssk,
 		       struct mptcp_subflow_context *subflow)
 {
-	bool dispose_socket = false;
-	struct socket *sock;
-
 	list_del(&subflow->node);
 
 	lock_sock_nested(ssk, SINGLE_DEPTH_NESTING);
@@ -2126,11 +2119,8 @@ void __mptcp_close_ssk(struct sock *sk, struct sock *ssk,
 	/* if we are invoked by the msk cleanup code, the subflow is
 	 * already orphaned
 	 */
-	sock = ssk->sk_socket;
-	if (sock) {
-		dispose_socket = sock != sk->sk_socket;
+	if (ssk->sk_socket)
 		sock_orphan(ssk);
-	}
 
 	subflow->disposable = 1;
 
@@ -2148,8 +2138,6 @@ void __mptcp_close_ssk(struct sock *sk, struct sock *ssk,
 		__sock_put(ssk);
 	}
 	release_sock(ssk);
-	if (dispose_socket)
-		iput(SOCK_INODE(sock));
 
 	sock_put(ssk);
 }
@@ -2540,6 +2528,12 @@ static void __mptcp_destroy_sock(struct sock *sk)
 
 	pr_debug("msk=%p", msk);
 
+	/* dispose the ancillatory tcp socket, if any */
+	if (msk->subflow) {
+		iput(SOCK_INODE(msk->subflow));
+		msk->subflow = NULL;
+	}
+
 	/* be sure to always acquire the join list lock, to sync vs
 	 * mptcp_finish_join().
 	 */
@@ -2590,20 +2584,10 @@ static void mptcp_close(struct sock *sk, long timeout)
 	inet_csk(sk)->icsk_mtup.probe_timestamp = tcp_jiffies32;
 	list_for_each_entry(subflow, &mptcp_sk(sk)->conn_list, node) {
 		struct sock *ssk = mptcp_subflow_tcp_sock(subflow);
-		bool slow, dispose_socket;
-		struct socket *sock;
+		bool slow = lock_sock_fast(ssk);
 
-		slow = lock_sock_fast(ssk);
-		sock = ssk->sk_socket;
-		dispose_socket = sock && sock != sk->sk_socket;
 		sock_orphan(ssk);
 		unlock_sock_fast(ssk, slow);
-
-		/* for the outgoing subflows we additionally need to free
-		 * the associated socket
-		 */
-		if (dispose_socket)
-			iput(SOCK_INODE(sock));
 	}
 	sock_orphan(sk);
 
@@ -3042,7 +3026,7 @@ void mptcp_finish_connect(struct sock *ssk)
 	mptcp_rcv_space_init(msk, ssk);
 }
 
-static void mptcp_sock_graft(struct sock *sk, struct socket *parent)
+void mptcp_sock_graft(struct sock *sk, struct socket *parent)
 {
 	write_lock_bh(&sk->sk_callback_lock);
 	rcu_assign_pointer(sk->sk_wq, &parent->wq);
net/mptcp/protocol.h +1 -0
index d6400ad2d615..65d200a1072b 100644
@@ -473,6 +473,7 @@ void mptcp_subflow_shutdown(struct sock *sk, struct sock *ssk, int how);
 void __mptcp_close_ssk(struct sock *sk, struct sock *ssk,
 		       struct mptcp_subflow_context *subflow);
 void mptcp_subflow_reset(struct sock *ssk);
+void mptcp_sock_graft(struct sock *sk, struct socket *parent);
 
 /* called with sk socket lock held */
 int __mptcp_subflow_connect(struct sock *sk, const struct mptcp_addr_info *loc,
net/mptcp/subflow.c +3 -0
index 278cbe3e539e..22313710d769 100644
@@ -1159,6 +1159,9 @@ int __mptcp_subflow_connect(struct sock *sk, const struct mptcp_addr_info *loc,
 	if (err && err != -EINPROGRESS)
 		goto failed_unlink;
 
+	/* discard the subflow socket */
+	mptcp_sock_graft(ssk, sk->sk_socket);
+	iput(SOCK_INODE(sf));
 	return err;
 
 failed_unlink:
Signature · 1 line
2.26.2
author
#2 lore raw

PATCH v3 2/4 mptcp: re-enable sndbuf autotune

In reply to Paolo Abeni

After commit 6e628cd3a8f7 ("mptcp: use mptcp release_cb for delayed tasks"), MPTCP never sets the flag bit SOCK_NOSPACE on its subflow. As a side effect, autotune never takes place, as it happens inside tcp_new_space(), which in turn is called only when the mentioned bit is set.

Let's sendmsg() set the subflows NOSPACE bit when looking for more memory. Additionally, cleanup the sndbuf propagation from subflow into the msk, leveraging the subflow write space callback and dropping a bunch of duplicate code.

This also makes the SNDBUF_LIMITED chrono relevant again for MPTCP subflows.

Fixes 6e628cd3a8f7 ("mptcp: use mptcp release_cb for delayed tasks")
Signed-off-by Paolo Abeni <pabeni(a)redhat.com>

---
v2 -> v3:
- drop left-over 'use_backup' flag

v1 -> v2:
- leverage the previous commit, so we can always simply set the nospace bit on the msk/parent socket

---
 net/mptcp/protocol.c | 53 ++++++++++++++++++--------------------------
 net/mptcp/protocol.h | 19 ++++++++++++++++
 net/mptcp/subflow.c  | 10 ++++++++-
 3 files changed, 49 insertions(+), 33 deletions(-)
net/mptcp/protocol.c +21 -32
index ada69495a51b..ab623e9690f3 100644
@@ -730,10 +730,14 @@ void mptcp_data_ready(struct sock *sk, struct sock *ssk)
 
 void __mptcp_flush_join_list(struct mptcp_sock *msk)
 {
+	struct mptcp_subflow_context *subflow;
+
 	if (likely(list_empty(&msk->join_list)))
 		return;
 
 	spin_lock_bh(&msk->join_list_lock);
+	list_for_each_entry(subflow, &msk->join_list, node)
+		mptcp_propagate_sndbuf((struct sock *)msk, mptcp_subflow_tcp_sock(subflow));
 	list_splice_tail_init(&msk->join_list, &msk->conn_list);
 	spin_unlock_bh(&msk->join_list_lock);
 }
@@ -1033,13 +1037,7 @@ static void __mptcp_clean_una(struct sock *sk)
 			__mptcp_update_wmem(sk);
 			sk_mem_reclaim_partial(sk);
 		}
-
-		if (sk_stream_is_writeable(sk)) {
-			/* pairs with memory barrier in mptcp_poll */
-			smp_mb();
-			if (test_and_clear_bit(MPTCP_NOSPACE, &msk->flags))
-				sk_stream_write_space(sk);
-		}
+		mptcp_write_space(sk);
 	}
 
 	if (snd_una == READ_ONCE(msk->snd_nxt)) {
@@ -1358,8 +1356,7 @@ struct subflow_send_info {
 	u64 ratio;
 };
 
-static struct sock *mptcp_subflow_get_send(struct mptcp_sock *msk,
-					   u32 *sndbuf)
+static struct sock *mptcp_subflow_get_send(struct mptcp_sock *msk)
 {
 	struct subflow_send_info send_info[2];
 	struct mptcp_subflow_context *subflow;
@@ -1370,24 +1367,17 @@ static struct sock *mptcp_subflow_get_send(struct mptcp_sock *msk,
 
 	sock_owned_by_me((struct sock *)msk);
 
-	*sndbuf = 0;
 	if (__mptcp_check_fallback(msk)) {
 		if (!msk->first)
 			return NULL;
-		*sndbuf = msk->first->sk_sndbuf;
 		return sk_stream_memory_free(msk->first) ? msk->first : NULL;
 	}
 
 	/* re-use last subflow, if the burst allow that */
 	if (msk->last_snd && msk->snd_burst > 0 &&
 	    sk_stream_memory_free(msk->last_snd) &&
-	    mptcp_subflow_active(mptcp_subflow_ctx(msk->last_snd))) {
-		mptcp_for_each_subflow(msk, subflow) {
-			ssk =  mptcp_subflow_tcp_sock(subflow);
-			*sndbuf = max(tcp_sk(ssk)->snd_wnd, *sndbuf);
-		}
+	    mptcp_subflow_active(mptcp_subflow_ctx(msk->last_snd)))
 		return msk->last_snd;
-	}
 
 	/* pick the subflow with the lower wmem/wspace ratio */
 	for (i = 0; i < 2; ++i) {
@@ -1400,7 +1390,6 @@ static struct sock *mptcp_subflow_get_send(struct mptcp_sock *msk,
 			continue;
 
 		nr_active += !subflow->backup;
-		*sndbuf = max(tcp_sk(ssk)->snd_wnd, *sndbuf);
 		if (!sk_stream_memory_free(subflow->tcp_sock))
 			continue;
 
@@ -1430,6 +1419,7 @@ static struct sock *mptcp_subflow_get_send(struct mptcp_sock *msk,
 				       sk_stream_wspace(msk->last_snd));
 		return msk->last_snd;
 	}
+
 	return NULL;
 }
 
@@ -1450,7 +1440,6 @@ static void mptcp_push_pending(struct sock *sk, unsigned int flags)
 	};
 	struct mptcp_data_frag *dfrag;
 	int len, copied = 0;
-	u32 sndbuf;
 
 	while ((dfrag = mptcp_send_head(sk))) {
 		info.sent = dfrag->already_sent;
@@ -1461,12 +1450,7 @@ static void mptcp_push_pending(struct sock *sk, unsigned int flags)
 
 			prev_ssk = ssk;
 			__mptcp_flush_join_list(msk);
-			ssk = mptcp_subflow_get_send(msk, &sndbuf);
-
-			/* do auto tuning */
-			if (!(sk->sk_userlocks & SOCK_SNDBUF_LOCK) &&
-			    sndbuf > READ_ONCE(sk->sk_sndbuf))
-				WRITE_ONCE(sk->sk_sndbuf, sndbuf);
+			ssk = mptcp_subflow_get_send(msk);
 
 			/* try to keep the subflow socket lock across
 			 * consecutive xmit on the same socket
@@ -1533,11 +1517,6 @@ static void __mptcp_subflow_push_pending(struct sock *sk, struct sock *ssk)
 		while (len > 0) {
 			int ret = 0;
 
-			/* do auto tuning */
-			if (!(sk->sk_userlocks & SOCK_SNDBUF_LOCK) &&
-			    ssk->sk_sndbuf > READ_ONCE(sk->sk_sndbuf))
-				WRITE_ONCE(sk->sk_sndbuf, ssk->sk_sndbuf);
-
 			if (unlikely(mptcp_must_reclaim_memory(sk, ssk))) {
 				__mptcp_update_wmem(sk);
 				sk_mem_reclaim_partial(sk);
@@ -1575,6 +1554,15 @@ static void __mptcp_subflow_push_pending(struct sock *sk, struct sock *ssk)
 	}
 }
 
+static void mptcp_set_nospace(struct sock *sk)
+{
+	/* enable autotune */
+	set_bit(SOCK_NOSPACE, &sk->sk_socket->flags);
+
+	/* will be cleared on avail space */
+	set_bit(MPTCP_NOSPACE, &mptcp_sk(sk)->flags);
+}
+
 static int mptcp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
 {
 	struct mptcp_sock *msk = mptcp_sk(sk);
@@ -1676,7 +1664,7 @@ static int mptcp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
 		continue;
 
 wait_for_memory:
-		set_bit(MPTCP_NOSPACE, &msk->flags);
+		mptcp_set_nospace(sk);
 		mptcp_push_pending(sk, msg->msg_flags);
 		ret = sk_stream_wait_memory(sk, &timeo);
 		if (ret)
@@ -3269,6 +3257,7 @@ static int mptcp_stream_accept(struct socket *sock, struct socket *newsock,
 
 		mptcp_copy_inaddrs(newsk, msk->first);
 		mptcp_rcv_space_init(msk, msk->first);
+		mptcp_propagate_sndbuf(newsk, msk->first);
 
 		/* set ssk->sk_socket of accept()ed flows to mptcp socket.
 		 * This is needed so NOSPACE flag can be set from tcp stack.
@@ -3309,7 +3298,7 @@ static __poll_t mptcp_check_writeable(struct mptcp_sock *msk)
 	if (sk_stream_is_writeable(sk))
 		return EPOLLOUT | EPOLLWRNORM;
 
-	set_bit(MPTCP_NOSPACE, &msk->flags);
+	mptcp_set_nospace(sk);
 	smp_mb__after_atomic(); /* msk->flags is changed by write_space cb */
 	if (sk_stream_is_writeable(sk))
 		return EPOLLOUT | EPOLLWRNORM;
net/mptcp/protocol.h +19 -0
index 65d200a1072b..871534df6140 100644
@@ -522,6 +522,25 @@ static inline bool mptcp_data_fin_enabled(const struct mptcp_sock *msk)
 	       READ_ONCE(msk->write_seq) == READ_ONCE(msk->snd_nxt);
 }
 
+static inline bool mptcp_propagate_sndbuf(struct sock *sk, struct sock *ssk)
+{
+	if ((sk->sk_userlocks & SOCK_SNDBUF_LOCK) || ssk->sk_sndbuf <= READ_ONCE(sk->sk_sndbuf))
+		return false;
+
+	WRITE_ONCE(sk->sk_sndbuf, ssk->sk_sndbuf);
+	return true;
+}
+
+static inline void mptcp_write_space(struct sock *sk)
+{
+	if (sk_stream_is_writeable(sk)) {
+		/* pairs with memory barrier in mptcp_poll */
+		smp_mb();
+		if (test_and_clear_bit(MPTCP_NOSPACE, &mptcp_sk(sk)->flags))
+			sk_stream_write_space(sk);
+	}
+}
+
 void mptcp_destroy_common(struct mptcp_sock *msk);
 
 void __init mptcp_token_init(void);
net/mptcp/subflow.c +9 -1
index 22313710d769..31cc362a4638 100644
@@ -343,6 +343,7 @@ static void subflow_finish_connect(struct sock *sk, const struct sk_buff *skb)
 	if (subflow->conn_finished)
 		return;
 
+	mptcp_propagate_sndbuf(parent, sk);
 	subflow->rel_write_seq = 1;
 	subflow->conn_finished = 1;
 	subflow->ssn_offset = TCP_SKB_CB(skb)->seq;
@@ -1040,7 +1041,13 @@ static void subflow_data_ready(struct sock *sk)
 
 static void subflow_write_space(struct sock *ssk)
 {
-	/* we take action in __mptcp_clean_una() */
+	struct socket *sock = ssk->sk_socket;
+
+	if (mptcp_propagate_sndbuf(mptcp_subflow_ctx(ssk)->conn, ssk))
+		mptcp_write_space(ssk);
+
+	if (sk_stream_is_writeable(ssk) && sock)
+		clear_bit(SOCK_NOSPACE, &sock->flags);
 }
 
 static struct inet_connection_sock_af_ops *
@@ -1302,6 +1309,7 @@ static void subflow_state_change(struct sock *sk)
 	__subflow_state_change(sk);
 
 	if (subflow_simultaneous_connect(sk)) {
+		mptcp_propagate_sndbuf(parent, sk);
 		mptcp_do_fallback(sk);
 		mptcp_rcv_space_init(mptcp_sk(parent), sk);
 		pr_fallback(mptcp_sk(parent));
Signature · 1 line
2.26.2
author
#3 lore raw

PATCH v3 3/4 mptcp: do not queue excessive data on subflows.

In reply to Paolo Abeni

The current packet scheduler can enqueue up to sndbuf data on each subflow. If the send buffer is large and the subflows are not symmetric, this could lead to suboptimal aggregate bandwidth utilization.

Limit the amount of queued data to the maximum cwnd.

Signed-off-by Paolo Abeni <pabeni(a)redhat.com>
---
 net/mptcp/protocol.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
net/mptcp/protocol.c +2 -2
index ab623e9690f3..f2e440d40fbd 100644
@@ -1390,7 +1390,7 @@ static struct sock *mptcp_subflow_get_send(struct mptcp_sock *msk)
 			continue;
 
 		nr_active += !subflow->backup;
-		if (!sk_stream_memory_free(subflow->tcp_sock))
+		if (!sk_stream_memory_free(subflow->tcp_sock) || !tcp_sk(ssk)->snd_wnd)
 			continue;
 
 		pace = READ_ONCE(ssk->sk_pacing_rate);
@@ -1416,7 +1416,7 @@ static struct sock *mptcp_subflow_get_send(struct mptcp_sock *msk)
 	if (send_info[0].ssk) {
 		msk->last_snd = send_info[0].ssk;
 		msk->snd_burst = min_t(int, MPTCP_SEND_BURST_SIZE,
-				       sk_stream_wspace(msk->last_snd));
+				       tcp_sk(msk->last_snd)->snd_wnd);
 		return msk->last_snd;
 	}
 
Signature · 1 line
2.26.2
author
#4 lore raw

PATCH v3 4/4 mptcp: schedule work for better snd subflow selection

In reply to Paolo Abeni

Otherwise the packet scheduler policy will not be enforced when pushing pending data at MPTCP-level ack reception time.

Signed-off-by Paolo Abeni <pabeni(a)redhat.com>

---
I'm very sad to re-introduce the worker usage for the datapath, but I could not find an easy way to avoid it.

I'm thinking of some weird schema involving a dummy percpu napi instance serving/processing a queue of mptcp sockets instead of actual packets. Any BH user can enqueue an MPTCP subflow there to delegate some action to the latter.

The above will require also overriding the tcp_release_cb() with something able to process this delegated events.

Yep, crazy, but possibly doable without any core change and possibly could be used to avoid the worker usage even for data fin processing/sending.

---
 net/mptcp/protocol.c | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)
net/mptcp/protocol.c +8 -3
index f2e440d40fbd..d538f5a6810c 100644
@@ -2243,6 +2243,7 @@ static void mptcp_worker(struct work_struct *work)
 	if (unlikely(state == TCP_CLOSE))
 		goto unlock;
 
+	mptcp_push_pending(sk, 0);
 	mptcp_check_data_fin_ack(sk);
 	__mptcp_flush_join_list(msk);
 
@@ -2901,10 +2902,14 @@ void __mptcp_check_push(struct sock *sk, struct sock *ssk)
 	if (!mptcp_send_head(sk))
 		return;
 
-	if (!sock_owned_by_user(sk))
-		__mptcp_subflow_push_pending(sk, ssk);
-	else
+	if (!sock_owned_by_user(sk)) {
+		if (mptcp_subflow_get_send(mptcp_sk(sk)) == ssk)
+			__mptcp_subflow_push_pending(sk, ssk);
+		else
+			mptcp_schedule_work(sk);
+	} else {
 		set_bit(MPTCP_PUSH_PENDING, &mptcp_sk(sk)->flags);
+	}
 }
 
 #define MPTCP_DEFERRED_ALL (TCPF_WRITE_TIMER_DEFERRED)
Signature · 1 line
2.26.2
#5 lore raw
In reply to Paolo Abeni
Quoted 22 lines · Paolo Abeni

WRT to the previous iteration, this clean-up patch 2/4 dropping the left over 'use_backup' field, as suggested by Mat. No other change

Relevant bits from the previous cover letter:

There are still some sporadic self-test failures for assimetric links. Looks like they are caused by retransmissions/receiver dropping packets due to MPTCP level OoO check.

I'll try to address that asap, but I still consider this series for inclusion, so that it can stage a bit in the export branch.

Note: the last patch re-introduces the workqueue usage for the datapath (yes, it makes me feel sick, sad and tired :(. I was unable to avoid it.

Paolo Abeni (4):
mptcp: always graft subflow socket to parent.
mptcp: re-enable sndbuf autotune
mptcp: do not queue excessive data on subflows.
mptcp: schedule work for better snd subflow selection

Quoted 7 lines
net/mptcp/protocol.c | 104 +++++++++++++++++--------------------------
net/mptcp/protocol.h |  20 +++++++++
net/mptcp/subflow.c  |  13 +++++-
3 files changed, 73 insertions(+), 64 deletions(-)
Signature · 1 line
2.26.2

Thanks Paolo. I think v3 of the series looks good to stage on the export branch.

--
Mat Martineau
Intel

#6 lore raw
In reply to Paolo Abeni

Hi Paolo, Mat,

Quoted 9 lines · Paolo Abeni

WRT to the previous iteration, this clean-up patch 2/4 dropping the left over 'use_backup' field, as suggested by Mat. No other change

Relevant bits from the previous cover letter:

There are still some sporadic self-test failures for assimetric links. Looks like they are caused by retransmissions/receiver dropping packets due to MPTCP level OoO check.

Quoted 5 lines

I'll try to address that asap, but I still consider this series for inclusion, so that it can stage a bit in the export branch.

Note: the last patch re-introduces the workqueue usage for the datapath (yes, it makes me feel sick, sad and tired :(. I was unable to avoid it.

Thank you for these new patches and the reviews!

Just added in the tree without dot at the end of the commit titles and Mat's Reviewed-by tag:

- 19c774f1761a: mptcp: always graft subflow socket to parent
- 3d489e89a9d6: mptcp: re-enable sndbuf autotune
- b8a4e1a17113: mptcp: do not queue excessive data on subflows
- 2e4cfdb68e07: mptcp: schedule work for better snd subflow selection
- Results: 591445962c81..bdff08d2ce0e

Tests + export have been scheduled!

Cheers,
Matt

Signature · 2 lines
Tessares | Belgium | Hybrid Access Solutions
www.tessares.net