Skip to contentSkip to conversation

refactor push pending

PATCH Superseded Geliang Tang 4 participants 19 messages Oct 11, 2022 – Oct 18, 2022

Subsystems (5)

v6 of 9 Superseded Latest is v9 →

lore
Share & tools
lore
#0 lore raw

v6 cover refactor push pending

v6:
- drop all msk->last_snd, add last_snd variable instead in patch 13
- fix lock_sock issue in patch 3
- merge squash-to patches
- This series should be merged between the commit "mptcp: add get_subflow wrappers" and "bpf: Add bpf_mptcp_sched_ops", except the last squash-to patch.

v5:
- address Mat's comments in v4.

v4:
- update __mptcp_subflow_push_pending as Mat suggested.
- add more patches from "BPF redundant scheduler" series.

v3:
- add a cleanup patch.
- remove msk->last_snd in mptcp_subflow_get_send().
- add the loop that calls the scheduler again in __mptcp_push_pending().

v2:
- add snd_burst check in dfrags loop as Mat suggested.

Refactor __mptcp_push_pending() and __mptcp_subflow_push_pending() to remove duplicate code and support redundant scheduler more easily in __mptcp_subflow_push_pending().

Geliang Tang (13):
Squash to "mptcp: add get_subflow wrappers"
mptcp: change 'first' as a parameter
mptcp: refactor push_pending logic
mptcp: drop last_snd of struct mptcp_sock
mptcp: simplify push_pending
mptcp: multi subflows push_pending
mptcp: don't delegate sending repeatedly
mptcp: use msk instead of mptcp_sk
mptcp: refactor subflow_push_pending logic
mptcp: simplify subflow_push_pending
mptcp: multi subflows subflow_push_pending
mptcp: multi subflows retrans support
Squash to "selftests/bpf: Add bpf_rr scheduler"

 net/mptcp/pm.c                                |   9 +-
 net/mptcp/pm_netlink.c                        |   3 -
 net/mptcp/protocol.c                          | 280 +++++++++---------
 net/mptcp/protocol.h                          |   6 +-
 net/mptcp/sched.c                             |  61 ++--
 tools/testing/selftests/bpf/bpf_tcp_helpers.h |   1 -
 .../selftests/bpf/progs/mptcp_bpf_rr.c        |   6 +-
 7 files changed, 182 insertions(+), 184 deletions(-)
Signature · 1 line
2.35.3
#1 lore raw

PATCH v6 1/13 Squash to "mptcp: add get_subflow wrappers"

Please update the commit log:

'''
This patch defines two new wrappers mptcp_sched_get_send() and mptcp_sched_get_retrans(), invoke get_subflow() of msk->sched in them.

Set the subflow pointers array in struct mptcp_sched_data before invoking get_subflow(), then it can be used in get_subflow() in the BPF contexts.

Check the subflow scheduled flags to test which subflow or subflows are picked by the scheduler.

Move sock_owned_by_me() and the fallback check code from mptcp_subflow_get_send/retrans() into the wrappers. '''

---
 net/mptcp/protocol.c |  8 +++---
 net/mptcp/protocol.h |  4 +--
 net/mptcp/sched.c    | 61 +++++++++++++++++++++-----------------------
 3 files changed, 35 insertions(+), 38 deletions(-)
net/mptcp/protocol.c +4 -4
index 1aa940928b4f..9a2253436f50 100644
@@ -1547,7 +1547,7 @@ void __mptcp_push_pending(struct sock *sk, unsigned int flags)
 			int ret = 0;
 
 			prev_ssk = ssk;
-			ssk = mptcp_sched_get_send(msk);
+			ssk = mptcp_subflow_get_send(msk);
 
 			/* First check. If the ssk has changed since
 			 * the last round, release prev_ssk
@@ -1616,7 +1616,7 @@ static void __mptcp_subflow_push_pending(struct sock *sk, struct sock *ssk)
 			 * check for a different subflow usage only after
 			 * spooling the first chunk of data
 			 */
-			xmit_ssk = first ? ssk : mptcp_sched_get_send(mptcp_sk(sk));
+			xmit_ssk = first ? ssk : mptcp_subflow_get_send(mptcp_sk(sk));
 			if (!xmit_ssk)
 				goto out;
 			if (xmit_ssk != ssk) {
@@ -2481,7 +2481,7 @@ static void __mptcp_retrans(struct sock *sk)
 	mptcp_clean_una_wakeup(sk);
 
 	/* first check ssk: need to kick "stale" logic */
-	ssk = mptcp_sched_get_retrans(msk);
+	ssk = mptcp_subflow_get_retrans(msk);
 	dfrag = mptcp_rtx_head(sk);
 	if (!dfrag) {
 		if (mptcp_data_fin_enabled(msk)) {
@@ -3199,7 +3199,7 @@ void __mptcp_check_push(struct sock *sk, struct sock *ssk)
 		return;
 
 	if (!sock_owned_by_user(sk)) {
-		struct sock *xmit_ssk = mptcp_sched_get_send(mptcp_sk(sk));
+		struct sock *xmit_ssk = mptcp_subflow_get_send(mptcp_sk(sk));
 
 		if (xmit_ssk == ssk)
 			__mptcp_subflow_push_pending(sk, ssk);
net/mptcp/protocol.h +2 -2
index 18f866b1afda..d44c97bb41db 100644
@@ -642,8 +642,8 @@ void mptcp_subflow_set_scheduled(struct mptcp_subflow_context *subflow,
 				 bool scheduled);
 struct sock *mptcp_subflow_get_send(struct mptcp_sock *msk);
 struct sock *mptcp_subflow_get_retrans(struct mptcp_sock *msk);
-struct sock *mptcp_sched_get_send(struct mptcp_sock *msk);
-struct sock *mptcp_sched_get_retrans(struct mptcp_sock *msk);
+int mptcp_sched_get_send(struct mptcp_sock *msk);
+int mptcp_sched_get_retrans(struct mptcp_sock *msk);
 
 static inline bool __tcp_can_send(const struct sock *ssk)
 {
net/mptcp/sched.c +29 -32

96 lines not on this page · raw message

Signature · 1 line
2.35.3
#2 lore raw

PATCH v6 2/13 mptcp: change 'first' as a parameter

In reply to Geliang Tang

The function mptcp_subflow_process_delegated() uses the input ssk first, while __mptcp_check_push() invokes the packet scheduler first.

So this patch adds a new parameter named 'first' for the function __mptcp_subflow_push_pending() to deal with these two cases separately.

With this change, the code that invokes the packet scheduler in the function __mptcp_check_push() can be removed, and replaced by invoking __mptcp_subflow_push_pending() directly.

---
 net/mptcp/protocol.c | 21 +++++++--------------
 1 file changed, 7 insertions(+), 14 deletions(-)
net/mptcp/protocol.c +7 -14
index 9a2253436f50..0285b21ff912 100644
@@ -1593,7 +1593,8 @@ void __mptcp_push_pending(struct sock *sk, unsigned int flags)
 		__mptcp_check_send_data_fin(sk);
 }
 
-static void __mptcp_subflow_push_pending(struct sock *sk, struct sock *ssk)
+static void __mptcp_subflow_push_pending(struct sock *sk, struct sock *ssk,
+					 bool first)
 {
 	struct mptcp_sock *msk = mptcp_sk(sk);
 	struct mptcp_sendmsg_info info = {
@@ -1602,7 +1603,6 @@ static void __mptcp_subflow_push_pending(struct sock *sk, struct sock *ssk)
 	struct mptcp_data_frag *dfrag;
 	struct sock *xmit_ssk;
 	int len, copied = 0;
-	bool first = true;
 
 	info.flags = 0;
 	while ((dfrag = mptcp_send_head(sk))) {
@@ -1612,8 +1612,7 @@ static void __mptcp_subflow_push_pending(struct sock *sk, struct sock *ssk)
 		while (len > 0) {
 			int ret = 0;
 
-			/* the caller already invoked the packet scheduler,
-			 * check for a different subflow usage only after
+			/* check for a different subflow usage only after
 			 * spooling the first chunk of data
 			 */
 			xmit_ssk = first ? ssk : mptcp_subflow_get_send(mptcp_sk(sk));
@@ -3198,16 +3197,10 @@ void __mptcp_check_push(struct sock *sk, struct sock *ssk)
 	if (!mptcp_send_head(sk))
 		return;
 
-	if (!sock_owned_by_user(sk)) {
-		struct sock *xmit_ssk = mptcp_subflow_get_send(mptcp_sk(sk));
-
-		if (xmit_ssk == ssk)
-			__mptcp_subflow_push_pending(sk, ssk);
-		else if (xmit_ssk)
-			mptcp_subflow_delegate(mptcp_subflow_ctx(xmit_ssk), MPTCP_DELEGATE_SEND);
-	} else {
+	if (!sock_owned_by_user(sk))
+		__mptcp_subflow_push_pending(sk, ssk, false);
+	else
 		__set_bit(MPTCP_PUSH_PENDING, &mptcp_sk(sk)->cb_flags);
-	}
 }
 
 #define MPTCP_FLAGS_PROCESS_CTX_NEED (BIT(MPTCP_PUSH_PENDING) | \
@@ -3298,7 +3291,7 @@ void mptcp_subflow_process_delegated(struct sock *ssk)
 	if (test_bit(MPTCP_DELEGATE_SEND, &subflow->delegated_status)) {
 		mptcp_data_lock(sk);
 		if (!sock_owned_by_user(sk))
-			__mptcp_subflow_push_pending(sk, ssk);
+			__mptcp_subflow_push_pending(sk, ssk, true);
 		else
 			__set_bit(MPTCP_PUSH_PENDING, &mptcp_sk(sk)->cb_flags);
 		mptcp_data_unlock(sk);
Signature · 1 line
2.35.3
#3 lore raw

PATCH v6 3/13 mptcp: refactor push_pending logic

In reply to Geliang Tang

To support redundant package schedulers more easily, this patch refactors __mptcp_push_pending() logic from:

For each dfrag:
While sends succeed:
Call the scheduler (selects subflow and msk->snd_burst) Update subflow locks (push/release/acquire as needed) Send the dfrag data with mptcp_sendmsg_frag() Update already_sent, snd_nxt, snd_burst
Update msk->first_pending
Push/release on final subflow

->

While the scheduler selects one subflow:
Lock the subflow
For each pending dfrag:
While sends succeed:
Send the dfrag data with mptcp_sendmsg_frag() Update already_sent, snd_nxt, snd_burst
Update msk->first_pending
Break if required by msk->snd_burst / etc
Push and release the subflow

This patch also moves the burst check conditions out of the function mptcp_subflow_get_send(), check them in __mptcp_push_pending() and __mptcp_subflow_push_pending() in the inner "for each pending dfrag" loop.

---
 net/mptcp/protocol.c | 86 ++++++++++++++++++++------------------------
 1 file changed, 39 insertions(+), 47 deletions(-)
net/mptcp/protocol.c +39 -47
index 0285b21ff912..52ac57fd8c27 100644
@@ -1417,14 +1417,6 @@ struct sock *mptcp_subflow_get_send(struct mptcp_sock *msk)
 	u64 linger_time;
 	long tout = 0;
 
-	/* 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_set_timeout(sk);
-		return msk->last_snd;
-	}
-
 	/* pick the subflow with the lower wmem/wspace ratio */
 	for (i = 0; i < SSK_MODE_MAX; ++i) {
 		send_info[i].ssk = NULL;
@@ -1530,60 +1522,53 @@ void mptcp_check_and_set_pending(struct sock *sk)
 
 void __mptcp_push_pending(struct sock *sk, unsigned int flags)
 {
-	struct sock *prev_ssk = NULL, *ssk = NULL;
 	struct mptcp_sock *msk = mptcp_sk(sk);
 	struct mptcp_sendmsg_info info = {
 				.flags = flags,
 	};
 	bool do_check_data_fin = false;
 	struct mptcp_data_frag *dfrag;
+	struct sock *ssk;
 	int len;
 
-	while ((dfrag = mptcp_send_head(sk))) {
-		info.sent = dfrag->already_sent;
-		info.limit = dfrag->data_len;
-		len = dfrag->data_len - dfrag->already_sent;
-		while (len > 0) {
-			int ret = 0;
-
-			prev_ssk = ssk;
-			ssk = mptcp_subflow_get_send(msk);
-
-			/* First check. If the ssk has changed since
-			 * the last round, release prev_ssk
-			 */
-			if (ssk != prev_ssk && prev_ssk)
-				mptcp_push_release(prev_ssk, &info);
-			if (!ssk)
-				goto out;
+again:
+	while (mptcp_send_head(sk) && (ssk = mptcp_subflow_get_send(msk))) {
+		lock_sock(ssk);
 
-			/* Need to lock the new subflow only if different
-			 * from the previous one, otherwise we are still
-			 * helding the relevant lock
-			 */
-			if (ssk != prev_ssk)
-				lock_sock(ssk);
+		while ((dfrag = mptcp_send_head(sk))) {
+			info.sent = dfrag->already_sent;
+			info.limit = dfrag->data_len;
+			len = dfrag->data_len - dfrag->already_sent;
+			while (len > 0) {
+				int ret = 0;
+
+				ret = mptcp_sendmsg_frag(sk, ssk, dfrag, &info);
+				if (ret <= 0) {
+					mptcp_push_release(ssk, &info);
+					if (ret == -EAGAIN)
+						goto again;
+					goto out;
+				}
+
+				do_check_data_fin = true;
+				info.sent += ret;
+				len -= ret;
+
+				mptcp_update_post_push(msk, dfrag, ret);
+			}
+			WRITE_ONCE(msk->first_pending, mptcp_send_next(sk));
 
-			ret = mptcp_sendmsg_frag(sk, ssk, dfrag, &info);
-			if (ret <= 0) {
-				if (ret == -EAGAIN)
-					continue;
+			if (msk->snd_burst <= 0 ||
+			    !sk_stream_memory_free(ssk) ||
+			    !mptcp_subflow_active(mptcp_subflow_ctx(ssk))) {
 				mptcp_push_release(ssk, &info);
-				goto out;
+				goto again;
 			}
-
-			do_check_data_fin = true;
-			info.sent += ret;
-			len -= ret;
-
-			mptcp_update_post_push(msk, dfrag, ret);
+			mptcp_set_timeout(sk);
 		}
-		WRITE_ONCE(msk->first_pending, mptcp_send_next(sk));
-	}
 
-	/* at this point we held the socket lock for the last subflow we used */
-	if (ssk)
 		mptcp_push_release(ssk, &info);
+	}
 
 out:
 	/* ensure the rtx timer is running */
@@ -1636,6 +1621,13 @@ static void __mptcp_subflow_push_pending(struct sock *sk, struct sock *ssk,
 			mptcp_update_post_push(msk, dfrag, ret);
 		}
 		WRITE_ONCE(msk->first_pending, mptcp_send_next(sk));
+
+		if (msk->snd_burst <= 0 ||
+		    !sk_stream_memory_free(ssk) ||
+		    !mptcp_subflow_active(mptcp_subflow_ctx(ssk))) {
+			goto out;
+		}
+		mptcp_set_timeout(sk);
 	}
 
 out:
Signature · 1 line
2.35.3
#4 lore raw

PATCH v6 4/13 mptcp: drop last_snd of struct mptcp_sock

In reply to Geliang Tang

msk->last_snd is no longer used anymore, drop it as well as the macro MPTCP_RESET_SCHEDULER.

---
 net/mptcp/pm.c         |  9 +--------
 net/mptcp/pm_netlink.c |  3 ---
 net/mptcp/protocol.c   | 11 +----------
 net/mptcp/protocol.h   |  2 --
 4 files changed, 2 insertions(+), 23 deletions(-)
net/mptcp/pm.c +1 -8
index 45e2a48397b9..cdeb7280ac76 100644
@@ -282,15 +282,8 @@ void mptcp_pm_mp_prio_received(struct sock *ssk, u8 bkup)
 
 	pr_debug("subflow->backup=%d, bkup=%d\n", subflow->backup, bkup);
 	msk = mptcp_sk(sk);
-	if (subflow->backup != bkup) {
+	if (subflow->backup != bkup)
 		subflow->backup = bkup;
-		mptcp_data_lock(sk);
-		if (!sock_owned_by_user(sk))
-			msk->last_snd = NULL;
-		else
-			__set_bit(MPTCP_RESET_SCHEDULER,  &msk->cb_flags);
-		mptcp_data_unlock(sk);
-	}
 
 	mptcp_event(MPTCP_EVENT_SUB_PRIORITY, msk, ssk, GFP_ATOMIC);
 }
net/mptcp/pm_netlink.c +0 -3
index 9813ed0fde9b..1f2da4aedcb4 100644
@@ -475,9 +475,6 @@ static void __mptcp_pm_send_ack(struct mptcp_sock *msk, struct mptcp_subflow_con
 
 	slow = lock_sock_fast(ssk);
 	if (prio) {
-		if (subflow->backup != backup)
-			msk->last_snd = NULL;
-
 		subflow->send_mp_prio = 1;
 		subflow->backup = backup;
 		subflow->request_bkup = backup;
net/mptcp/protocol.c +1 -10
index 52ac57fd8c27..7d5f89799c9a 100644
@@ -1469,16 +1469,13 @@ struct sock *mptcp_subflow_get_send(struct mptcp_sock *msk)
 
 	burst = min_t(int, MPTCP_SEND_BURST_SIZE, mptcp_wnd_end(msk) - msk->snd_nxt);
 	wmem = READ_ONCE(ssk->sk_wmem_queued);
-	if (!burst) {
-		msk->last_snd = NULL;
+	if (!burst)
 		return ssk;
-	}
 
 	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;
 }
@@ -2346,9 +2343,6 @@ static void __mptcp_close_ssk(struct sock *sk, struct sock *ssk,
 		msk->first = NULL;
 
 out:
-	if (ssk == msk->last_snd)
-		msk->last_snd = NULL;
-
 	if (need_push)
 		__mptcp_push_pending(sk, 0);
 }
@@ -2981,7 +2975,6 @@ static int mptcp_disconnect(struct sock *sk, int flags)
 	 * subflow
 	 */
 	mptcp_destroy_common(msk, MPTCP_CF_FASTCLOSE);
-	msk->last_snd = NULL;
 	WRITE_ONCE(msk->flags, 0);
 	msk->cb_flags = 0;
 	msk->push_pending = 0;
@@ -3242,8 +3235,6 @@ static void mptcp_release_cb(struct sock *sk)
 			__mptcp_set_connected(sk);
 		if (__test_and_clear_bit(MPTCP_ERROR_REPORT, &msk->cb_flags))
 			__mptcp_error_report(sk);
-		if (__test_and_clear_bit(MPTCP_RESET_SCHEDULER, &msk->cb_flags))
-			msk->last_snd = NULL;
 	}
 
 	__mptcp_update_rmem(sk);
net/mptcp/protocol.h +0 -2
index d44c97bb41db..68a2b22a47eb 100644
@@ -124,7 +124,6 @@
 #define MPTCP_RETRANSMIT	4
 #define MPTCP_FLUSH_JOIN_LIST	5
 #define MPTCP_CONNECTED		6
-#define MPTCP_RESET_SCHEDULER	7
 
 static inline bool before64(__u64 seq1, __u64 seq2)
 {
@@ -258,7 +257,6 @@ struct mptcp_sock {
 	atomic64_t	rcv_wnd_sent;
 	u64		rcv_data_fin_seq;
 	int		rmem_fwd_alloc;
-	struct sock	*last_snd;
 	int		snd_burst;
 	int		old_wspace;
 	u64		recovery_snd_nxt;	/* in recovery mode accept up to this seq;
Signature · 1 line
2.35.3
#5 lore raw

PATCH v6 5/13 mptcp: simplify push_pending

In reply to Geliang Tang

This patch moves the duplicate code from __mptcp_push_pending() and __mptcp_subflow_push_pending() into a new helper function, named __subflow_push_pending(). And simplify __mptcp_push_pending() by invoking this helper.

---
 net/mptcp/protocol.c | 95 +++++++++++++++++++++++++-------------------
 1 file changed, 54 insertions(+), 41 deletions(-)
net/mptcp/protocol.c +54 -41
index 7d5f89799c9a..817e539d1d12 100644
@@ -1480,12 +1480,6 @@ struct sock *mptcp_subflow_get_send(struct mptcp_sock *msk)
 	return ssk;
 }
 
-static void mptcp_push_release(struct sock *ssk, struct mptcp_sendmsg_info *info)
-{
-	tcp_push(ssk, 0, info->mss_now, tcp_sk(ssk)->nonagle, info->size_goal);
-	release_sock(ssk);
-}
-
 static void mptcp_update_post_push(struct mptcp_sock *msk,
 				   struct mptcp_data_frag *dfrag,
 				   u32 sent)
@@ -1517,61 +1511,80 @@ void mptcp_check_and_set_pending(struct sock *sk)
 		mptcp_sk(sk)->push_pending |= BIT(MPTCP_PUSH_PENDING);
 }
 
+static int __subflow_push_pending(struct sock *sk, struct sock *ssk,
+				  struct mptcp_sendmsg_info *info)
+{
+	struct mptcp_sock *msk = mptcp_sk(sk);
+	struct mptcp_data_frag *dfrag;
+	int len, copied = 0, err = 0;
+
+	while ((dfrag = mptcp_send_head(sk))) {
+		info->sent = dfrag->already_sent;
+		info->limit = dfrag->data_len;
+		len = dfrag->data_len - dfrag->already_sent;
+		while (len > 0) {
+			int ret = 0;
+
+			ret = mptcp_sendmsg_frag(sk, ssk, dfrag, info);
+			if (ret <= 0) {
+				err = copied ? : ret;
+				goto out;
+			}
+
+			info->sent += ret;
+			copied += ret;
+			len -= ret;
+
+			mptcp_update_post_push(msk, dfrag, ret);
+		}
+		WRITE_ONCE(msk->first_pending, mptcp_send_next(sk));
+
+		if (msk->snd_burst <= 0 ||
+		    !sk_stream_memory_free(ssk) ||
+		    !mptcp_subflow_active(mptcp_subflow_ctx(ssk))) {
+			err = copied ? : -EAGAIN;
+			goto out;
+		}
+		mptcp_set_timeout(sk);
+	}
+	err = copied;
+
+out:
+	if (copied) {
+		tcp_push(ssk, 0, info->mss_now, tcp_sk(ssk)->nonagle,
+			 info->size_goal);
+	}
+
+	return err;
+}
+
 void __mptcp_push_pending(struct sock *sk, unsigned int flags)
 {
 	struct mptcp_sock *msk = mptcp_sk(sk);
 	struct mptcp_sendmsg_info info = {
 				.flags = flags,
 	};
-	bool do_check_data_fin = false;
-	struct mptcp_data_frag *dfrag;
 	struct sock *ssk;
-	int len;
+	int ret = 0;
 
 again:
 	while (mptcp_send_head(sk) && (ssk = mptcp_subflow_get_send(msk))) {
 		lock_sock(ssk);
+		ret = __subflow_push_pending(sk, ssk, &info);
+		release_sock(ssk);
 
-		while ((dfrag = mptcp_send_head(sk))) {
-			info.sent = dfrag->already_sent;
-			info.limit = dfrag->data_len;
-			len = dfrag->data_len - dfrag->already_sent;
-			while (len > 0) {
-				int ret = 0;
-
-				ret = mptcp_sendmsg_frag(sk, ssk, dfrag, &info);
-				if (ret <= 0) {
-					mptcp_push_release(ssk, &info);
-					if (ret == -EAGAIN)
-						goto again;
-					goto out;
-				}
-
-				do_check_data_fin = true;
-				info.sent += ret;
-				len -= ret;
-
-				mptcp_update_post_push(msk, dfrag, ret);
-			}
-			WRITE_ONCE(msk->first_pending, mptcp_send_next(sk));
-
-			if (msk->snd_burst <= 0 ||
-			    !sk_stream_memory_free(ssk) ||
-			    !mptcp_subflow_active(mptcp_subflow_ctx(ssk))) {
-				mptcp_push_release(ssk, &info);
+		if (ret <= 0) {
+			if (ret == -EAGAIN)
 				goto again;
-			}
-			mptcp_set_timeout(sk);
+			goto out;
 		}
-
-		mptcp_push_release(ssk, &info);
 	}
 
 out:
 	/* ensure the rtx timer is running */
 	if (!mptcp_timer_pending(sk))
 		mptcp_reset_timer(sk);
-	if (do_check_data_fin)
+	if (ret > 0)
 		__mptcp_check_send_data_fin(sk);
 }
 
Signature · 1 line
2.35.3
#6 lore raw

PATCH v6 6/13 mptcp: multi subflows push_pending

In reply to Geliang Tang

This patch adds the multiple subflows support for __mptcp_push_pending(). Use mptcp_sched_get_send() wrapper instead of mptcp_subflow_get_send() in it.

Check the subflow scheduled flags to test which subflow or subflows are picked by the scheduler, use them to send data.

---
 net/mptcp/protocol.c | 32 ++++++++++++++++++++------------
 1 file changed, 20 insertions(+), 12 deletions(-)
net/mptcp/protocol.c +20 -12
index 817e539d1d12..86ac38d10bc4 100644
@@ -1561,22 +1561,30 @@ static int __subflow_push_pending(struct sock *sk, struct sock *ssk,
 void __mptcp_push_pending(struct sock *sk, unsigned int flags)
 {
 	struct mptcp_sock *msk = mptcp_sk(sk);
-	struct mptcp_sendmsg_info info = {
-				.flags = flags,
-	};
-	struct sock *ssk;
 	int ret = 0;
 
 again:
-	while (mptcp_send_head(sk) && (ssk = mptcp_subflow_get_send(msk))) {
-		lock_sock(ssk);
-		ret = __subflow_push_pending(sk, ssk, &info);
-		release_sock(ssk);
+	while (mptcp_send_head(sk) && !mptcp_sched_get_send(msk)) {
+		struct mptcp_subflow_context *subflow;
+		struct mptcp_sendmsg_info info = {
+			.flags = flags,
+		};
 
-		if (ret <= 0) {
-			if (ret == -EAGAIN)
-				goto again;
-			goto out;
+		mptcp_for_each_subflow(msk, subflow) {
+			if (READ_ONCE(subflow->scheduled)) {
+				struct sock *ssk = mptcp_subflow_tcp_sock(subflow);
+
+				lock_sock(ssk);
+				ret = __subflow_push_pending(sk, ssk, &info);
+				release_sock(ssk);
+
+				if (ret <= 0) {
+					if (ret == -EAGAIN)
+						goto again;
+					goto out;
+				}
+				mptcp_subflow_set_scheduled(subflow, false);
+			}
 		}
 	}
 
Signature · 1 line
2.35.3
#7 lore raw

Re: [PATCH mptcp-next v6 06/13] mptcp: multi subflows push_pending

Quoted 38 lines · Geliang Tang

This patch adds the multiple subflows support for __mptcp_push_pending(). Use mptcp_sched_get_send() wrapper instead of mptcp_subflow_get_send() in it.

Check the subflow scheduled flags to test which subflow or subflows are picked by the scheduler, use them to send data.

---
net/mptcp/protocol.c | 32 ++++++++++++++++++++------------

1 file changed, 20 insertions(+), 12 deletions(-)
net/mptcp/protocol.c +0 -4
index 817e539d1d12..86ac38d10bc4 100644
@@ -1561,22 +1561,30 @@ static int __subflow_push_pending(struct sock *sk, struct sock *ssk,
 void __mptcp_push_pending(struct sock *sk, unsigned int flags)
 {
 	struct mptcp_sock *msk = mptcp_sk(sk);
-	struct mptcp_sendmsg_info info = {
-				.flags = flags,
-	};
-	struct sock *ssk;
 	int ret = 0;
 

again:

net/mptcp/protocol.c +5 -5
-	while (mptcp_send_head(sk) && (ssk = mptcp_subflow_get_send(msk))) {
-		lock_sock(ssk);
-		ret = __subflow_push_pending(sk, ssk, &info);
-		release_sock(ssk);
+	while (mptcp_send_head(sk) && !mptcp_sched_get_send(msk)) {
+		struct mptcp_subflow_context *subflow;
+		struct mptcp_sendmsg_info info = {
+			.flags = flags,
+		};
 
-		if (ret <= 0) {

net/mptcp/protocol.c static int __subflow_push_pending(struct sock *sk, struct sock *ssk, · line 1579

Quoted 10 lines
net/mptcp/protocol.c +7 -3
-			if (ret == -EAGAIN)
-				goto again;
-			goto out;
+		mptcp_for_each_subflow(msk, subflow) {
+			if (READ_ONCE(subflow->scheduled)) {
+				struct sock *ssk = mptcp_subflow_tcp_sock(subflow);
+
+				lock_sock(ssk);
+				ret = __subflow_push_pending(sk, ssk, &info);
+				release_sock(ssk);

Geliang -

It seems like this doesn't allow for redundant transmits, since __subflow_push_pending() updates msk->first_pending. Did you find that redundant sends were working correctly (maybe I've misunderstood)?

- Mat

Quoted 8 lines
net/mptcp/protocol.c +8 -0
+
+				if (ret <= 0) {
+					if (ret == -EAGAIN)
+						goto again;
+					goto out;
+				}
+				mptcp_subflow_set_scheduled(subflow, false);
+			}
Quoted 8 lines
		}
	}
Signature · 1 line
2.35.3

--
Mat Martineau
Intel

#8 lore raw

Re: [PATCH mptcp-next v6 06/13] mptcp: multi subflows push_pending

Quoted 53 lines · Mat Martineau
Quoted 38 lines · Geliang Tang

This patch adds the multiple subflows support for __mptcp_push_pending(). Use mptcp_sched_get_send() wrapper instead of mptcp_subflow_get_send() in it.

Check the subflow scheduled flags to test which subflow or subflows are picked by the scheduler, use them to send data.

---
net/mptcp/protocol.c | 32 ++++++++++++++++++++------------

1 file changed, 20 insertions(+), 12 deletions(-)
net/mptcp/protocol.c +0 -4
index 817e539d1d12..86ac38d10bc4 100644
@@ -1561,22 +1561,30 @@ static int __subflow_push_pending(struct sock *sk, struct sock *ssk,
 void __mptcp_push_pending(struct sock *sk, unsigned int flags)
 {
 	struct mptcp_sock *msk = mptcp_sk(sk);
-	struct mptcp_sendmsg_info info = {
-				.flags = flags,
-	};
-	struct sock *ssk;
 	int ret = 0;
 

again:

net/mptcp/protocol.c +5 -5
-	while (mptcp_send_head(sk) && (ssk = mptcp_subflow_get_send(msk))) {
-		lock_sock(ssk);
-		ret = __subflow_push_pending(sk, ssk, &info);
-		release_sock(ssk);
+	while (mptcp_send_head(sk) && !mptcp_sched_get_send(msk)) {
+		struct mptcp_subflow_context *subflow;
+		struct mptcp_sendmsg_info info = {
+			.flags = flags,
+		};
 
-		if (ret <= 0) {

net/mptcp/protocol.c static int __subflow_push_pending(struct sock *sk, struct sock *ssk, · line 1579

Quoted 10 lines
net/mptcp/protocol.c +7 -3
-			if (ret == -EAGAIN)
-				goto again;
-			goto out;
+		mptcp_for_each_subflow(msk, subflow) {
+			if (READ_ONCE(subflow->scheduled)) {
+				struct sock *ssk = mptcp_subflow_tcp_sock(subflow);
+
+				lock_sock(ssk);
+				ret = __subflow_push_pending(sk, ssk, &info);
+				release_sock(ssk);

Geliang -

Quoted 3 lines

It seems like this doesn't allow for redundant transmits, since __subflow_push_pending() updates msk->first_pending. Did you find that redundant sends were working correctly (maybe I've misunderstood)?

Hi Mat,

Redundant sends are not support in this series yet. It will be added in this __subflow_push_pending() function in the next series with the selftests patches.

There are still some issues that haven't been solved in the redundant support patches. I'm still working on them.

Thanks,
-Geliang

Quoted 23 lines

- Mat

Quoted 8 lines
net/mptcp/protocol.c +8 -0
+
+				if (ret <= 0) {
+					if (ret == -EAGAIN)
+						goto again;
+					goto out;
+				}
+				mptcp_subflow_set_scheduled(subflow, false);
+			}
Quoted 8 lines
		}
	}
Signature · 1 line
2.35.3

--
Mat Martineau
Intel

#9 lore raw

Re: [PATCH mptcp-next v6 06/13] mptcp: multi subflows push_pending

Quoted 59 lines · Geliang Tang <[email protected]> 于2022年10月12日周三 18:37写道:
Quoted 53 lines · Mat Martineau
Quoted 38 lines · Geliang Tang

This patch adds the multiple subflows support for __mptcp_push_pending(). Use mptcp_sched_get_send() wrapper instead of mptcp_subflow_get_send() in it.

Check the subflow scheduled flags to test which subflow or subflows are picked by the scheduler, use them to send data.

---
net/mptcp/protocol.c | 32 ++++++++++++++++++++------------

1 file changed, 20 insertions(+), 12 deletions(-)
net/mptcp/protocol.c +0 -4
index 817e539d1d12..86ac38d10bc4 100644
@@ -1561,22 +1561,30 @@ static int __subflow_push_pending(struct sock *sk, struct sock *ssk,
 void __mptcp_push_pending(struct sock *sk, unsigned int flags)
 {
    struct mptcp_sock *msk = mptcp_sk(sk);
-   struct mptcp_sendmsg_info info = {
-                           .flags = flags,
-   };
-   struct sock *ssk;
    int ret = 0;
 

again:

net/mptcp/protocol.c +5 -5
-   while (mptcp_send_head(sk) && (ssk = mptcp_subflow_get_send(msk))) {
-           lock_sock(ssk);
-           ret = __subflow_push_pending(sk, ssk, &info);
-           release_sock(ssk);
+   while (mptcp_send_head(sk) && !mptcp_sched_get_send(msk)) {
+           struct mptcp_subflow_context *subflow;
+           struct mptcp_sendmsg_info info = {
+                   .flags = flags,
+           };
 
-           if (ret <= 0) {

net/mptcp/protocol.c static int __subflow_push_pending(struct sock *sk, struct sock *ssk, · line 1579

Quoted 10 lines
net/mptcp/protocol.c +7 -3
-                   if (ret == -EAGAIN)
-                           goto again;
-                   goto out;
+           mptcp_for_each_subflow(msk, subflow) {
+                   if (READ_ONCE(subflow->scheduled)) {
+                           struct sock *ssk = mptcp_subflow_tcp_sock(subflow);
+
+                           lock_sock(ssk);
+                           ret = __subflow_push_pending(sk, ssk, &info);
+                           release_sock(ssk);

Geliang -

Quoted 3 lines

It seems like this doesn't allow for redundant transmits, since __subflow_push_pending() updates msk->first_pending. Did you find that redundant sends were working correctly (maybe I've misunderstood)?

Quoted 8 lines

Hi Mat,

Redundant sends are not support in this series yet. It will be added in this __subflow_push_pending() function in the next series with the selftests patches.

There are still some issues that haven't been solved in the redundant support patches. I'm still working on them.

Hi Mat, I have good news. I finally solved these issues today and sent a new version, "BPF redundant scheduler" v13 . Now redundant sends are support in it, and all selftests (mptcp_connect.sh, mptcp_join.sh and simult_flows.sh) passed.

Quoted 28 lines

Thanks,
-Geliang

Quoted 23 lines

- Mat

Quoted 12 lines
net/mptcp/protocol.c +8 -1
+
+                           if (ret <= 0) {
+                                   if (ret == -EAGAIN)
+                                           goto again;
+                                   goto out;
+                           }
+                           mptcp_subflow_set_scheduled(subflow, false);
+                   }
            }
    }
 
--
Quoted 4 lines

2.35.3

--
Mat Martineau
Intel

#10 lore raw

PATCH v6 7/13 mptcp: don't delegate sending repeatedly

In reply to Geliang Tang

Do not delegate sending to a subflow repeatedly. Only do it to the subflow without delegated action.

---
 net/mptcp/protocol.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
net/mptcp/protocol.c +5 -1
index 86ac38d10bc4..6f55b51e10eb 100644
@@ -1600,6 +1600,7 @@ static void __mptcp_subflow_push_pending(struct sock *sk, struct sock *ssk,
 					 bool first)
 {
 	struct mptcp_sock *msk = mptcp_sk(sk);
+	struct mptcp_subflow_context *subflow;
 	struct mptcp_sendmsg_info info = {
 		.data_lock_held = true,
 	};
@@ -1622,7 +1623,10 @@ static void __mptcp_subflow_push_pending(struct sock *sk, struct sock *ssk,
 			if (!xmit_ssk)
 				goto out;
 			if (xmit_ssk != ssk) {
-				mptcp_subflow_delegate(mptcp_subflow_ctx(xmit_ssk),
+				subflow = mptcp_subflow_ctx(xmit_ssk);
+				if (mptcp_subflow_has_delegated_action(subflow))
+					goto out;
+				mptcp_subflow_delegate(subflow,
 						       MPTCP_DELEGATE_SEND);
 				goto out;
 			}
Signature · 1 line
2.35.3
#11 lore raw

Re: [PATCH mptcp-next v6 07/13] mptcp: don't delegate sending repeatedly

Quoted 21 lines · Geliang Tang

Do not delegate sending to a subflow repeatedly. Only do it to the subflow without delegated action.

---
net/mptcp/protocol.c | 6 +++++-

1 file changed, 5 insertions(+), 1 deletion(-)
net/mptcp/protocol.c +1 -0
index 86ac38d10bc4..6f55b51e10eb 100644
@@ -1600,6 +1600,7 @@ static void __mptcp_subflow_push_pending(struct sock *sk, struct sock *ssk,
 					 bool first)
 {
 	struct mptcp_sock *msk = mptcp_sk(sk);
+	struct mptcp_subflow_context *subflow;
 	struct mptcp_sendmsg_info info = {
 		.data_lock_held = true,
 	};

net/mptcp/protocol.c static void __mptcp_subflow_push_pending(struct sock *sk, struct sock *ssk, · line 1630

Quoted 9 lines
net/mptcp/protocol.c +4 -1
@@ -1622,7 +1623,10 @@ static void __mptcp_subflow_push_pending(struct sock *sk, struct sock *ssk,
 			if (!xmit_ssk)
 				goto out;
 			if (xmit_ssk != ssk) {
-				mptcp_subflow_delegate(mptcp_subflow_ctx(xmit_ssk),
+				subflow = mptcp_subflow_ctx(xmit_ssk);
+				if (mptcp_subflow_has_delegated_action(subflow))
+					goto out;
+				mptcp_subflow_delegate(subflow,
 						       MPTCP_DELEGATE_SEND);

Hi Geliang -

Did this patch fix a problem you were seeing? It seems like it isn't needed, and might introduce a bug.

mptcp_subflow_delegate() does a test_and_set_bit() to see if the specific action is already scheduled. This new code will skip the delegated send if *any* delegated action is set on the subflow.

Quoted 7 lines
net/mptcp/protocol.c +0 -0
 				goto out;
 			}
Signature · 1 line
2.35.3

--
Mat Martineau
Intel

#12 lore raw

PATCH v6 8/13 mptcp: use msk instead of mptcp_sk

In reply to Geliang Tang

Use msk instead of mptcp_sk(sk) in the functions where the variable "msk = mptcp_sk(sk)" has been defined.

---
 net/mptcp/protocol.c | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)
net/mptcp/protocol.c +5 -5
index 6f55b51e10eb..e3a5936df0fc 100644
@@ -1619,7 +1619,7 @@ static void __mptcp_subflow_push_pending(struct sock *sk, struct sock *ssk,
 			/* check for a different subflow usage only after
 			 * spooling the first chunk of data
 			 */
-			xmit_ssk = first ? ssk : mptcp_subflow_get_send(mptcp_sk(sk));
+			xmit_ssk = first ? ssk : mptcp_subflow_get_send(msk);
 			if (!xmit_ssk)
 				goto out;
 			if (xmit_ssk != ssk) {
@@ -2255,7 +2255,7 @@ bool __mptcp_retransmit_pending_data(struct sock *sk)
 	struct mptcp_data_frag *cur, *rtx_head;
 	struct mptcp_sock *msk = mptcp_sk(sk);
 
-	if (__mptcp_check_fallback(mptcp_sk(sk)))
+	if (__mptcp_check_fallback(msk))
 		return false;
 
 	if (tcp_rtx_and_write_queues_empty(sk))
@@ -2934,7 +2934,7 @@ bool __mptcp_close(struct sock *sk, long timeout)
 
 	sock_hold(sk);
 	pr_debug("msk=%p state=%d", sk, sk->sk_state);
-	if (mptcp_sk(sk)->token)
+	if (msk->token)
 		mptcp_event(MPTCP_EVENT_CLOSED, msk, NULL, GFP_KERNEL);
 
 	if (sk->sk_state == TCP_CLOSE) {
@@ -2993,8 +2993,8 @@ static int mptcp_disconnect(struct sock *sk, int flags)
 	mptcp_stop_timer(sk);
 	sk_stop_timer(sk, &sk->sk_timer);
 
-	if (mptcp_sk(sk)->token)
-		mptcp_event(MPTCP_EVENT_CLOSED, mptcp_sk(sk), NULL, GFP_KERNEL);
+	if (msk->token)
+		mptcp_event(MPTCP_EVENT_CLOSED, msk, NULL, GFP_KERNEL);
 
 	/* msk->subflow is still intact, the following will not free the first
 	 * subflow
Signature · 1 line
2.35.3
#13 lore raw

PATCH v6 9/13 mptcp: refactor subflow_push_pending logic

In reply to Geliang Tang

This patch refactors __mptcp_subflow_push_pending logic from:

For each dfrag:

	While sends succeed:
		Call the scheduler (selects subflow and msk->snd_burst)
		Send the dfrag data with mptcp_subflow_delegate(), break
		Send the dfrag data with mptcp_sendmsg_frag()
		Update dfrag->already_sent, msk->snd_nxt, msk->snd_burst
	Update msk->first_pending

->

While first_pending isn't empty:
Call the scheduler (selects subflow and msk->snd_burst) Send the dfrag data with mptcp_subflow_delegate(), break Send the dfrag data with mptcp_sendmsg_frag()
For each pending dfrag:
While sends succeed:
Send the dfrag data with mptcp_sendmsg_frag() Update already_sent, snd_nxt, snd_burst
Update msk->first_pending
Break if required by msk->snd_burst / etc

---
 net/mptcp/protocol.c | 72 +++++++++++++++++++++++---------------------
 1 file changed, 37 insertions(+), 35 deletions(-)
net/mptcp/protocol.c +37 -35
index e3a5936df0fc..676f3bff988b 100644
@@ -1609,47 +1609,49 @@ static void __mptcp_subflow_push_pending(struct sock *sk, struct sock *ssk,
 	int len, copied = 0;
 
 	info.flags = 0;
-	while ((dfrag = mptcp_send_head(sk))) {
-		info.sent = dfrag->already_sent;
-		info.limit = dfrag->data_len;
-		len = dfrag->data_len - dfrag->already_sent;
-		while (len > 0) {
-			int ret = 0;
-
-			/* check for a different subflow usage only after
-			 * spooling the first chunk of data
-			 */
-			xmit_ssk = first ? ssk : mptcp_subflow_get_send(msk);
-			if (!xmit_ssk)
-				goto out;
-			if (xmit_ssk != ssk) {
-				subflow = mptcp_subflow_ctx(xmit_ssk);
-				if (mptcp_subflow_has_delegated_action(subflow))
-					goto out;
-				mptcp_subflow_delegate(subflow,
-						       MPTCP_DELEGATE_SEND);
+	while (mptcp_send_head(sk)) {
+		/* check for a different subflow usage only after
+		 * spooling the first chunk of data
+		 */
+		xmit_ssk = first ? ssk : mptcp_subflow_get_send(msk);
+		if (!xmit_ssk)
+			goto out;
+		if (xmit_ssk != ssk) {
+			subflow = mptcp_subflow_ctx(xmit_ssk);
+			if (mptcp_subflow_has_delegated_action(subflow))
 				goto out;
-			}
+			mptcp_subflow_delegate(subflow,
+					       MPTCP_DELEGATE_SEND);
+			goto out;
+		}
 
-			ret = mptcp_sendmsg_frag(sk, ssk, dfrag, &info);
-			if (ret <= 0)
-				goto out;
+		while ((dfrag = mptcp_send_head(sk))) {
+			info.sent = dfrag->already_sent;
+			info.limit = dfrag->data_len;
+			len = dfrag->data_len - dfrag->already_sent;
+			while (len > 0) {
+				int ret = 0;
 
-			info.sent += ret;
-			copied += ret;
-			len -= ret;
-			first = false;
+				ret = mptcp_sendmsg_frag(sk, ssk, dfrag, &info);
+				if (ret <= 0)
+					goto out;
 
-			mptcp_update_post_push(msk, dfrag, ret);
-		}
-		WRITE_ONCE(msk->first_pending, mptcp_send_next(sk));
+				info.sent += ret;
+				copied += ret;
+				len -= ret;
+				first = false;
 
-		if (msk->snd_burst <= 0 ||
-		    !sk_stream_memory_free(ssk) ||
-		    !mptcp_subflow_active(mptcp_subflow_ctx(ssk))) {
-			goto out;
+				mptcp_update_post_push(msk, dfrag, ret);
+			}
+			WRITE_ONCE(msk->first_pending, mptcp_send_next(sk));
+
+			if (msk->snd_burst <= 0 ||
+			    !sk_stream_memory_free(ssk) ||
+			    !mptcp_subflow_active(mptcp_subflow_ctx(ssk))) {
+				goto out;
+			}
+			mptcp_set_timeout(sk);
 		}
-		mptcp_set_timeout(sk);
 	}
 
 out:
Signature · 1 line
2.35.3
#14 lore raw

PATCH v6 10/13 mptcp: simplify subflow_push_pending

In reply to Geliang Tang

This patch simplifies __mptcp_subflow_push_pending() by invoking __subflow_push_pending() helper.

---
 net/mptcp/protocol.c | 39 ++++++++-------------------------------
 1 file changed, 8 insertions(+), 31 deletions(-)
net/mptcp/protocol.c +8 -31
index 676f3bff988b..e9eb1da42418 100644
@@ -1604,11 +1604,11 @@ static void __mptcp_subflow_push_pending(struct sock *sk, struct sock *ssk,
 	struct mptcp_sendmsg_info info = {
 		.data_lock_held = true,
 	};
-	struct mptcp_data_frag *dfrag;
 	struct sock *xmit_ssk;
-	int len, copied = 0;
+	int ret = 0;
 
 	info.flags = 0;
+again:
 	while (mptcp_send_head(sk)) {
 		/* check for a different subflow usage only after
 		 * spooling the first chunk of data
@@ -1625,32 +1625,11 @@ static void __mptcp_subflow_push_pending(struct sock *sk, struct sock *ssk,
 			goto out;
 		}
 
-		while ((dfrag = mptcp_send_head(sk))) {
-			info.sent = dfrag->already_sent;
-			info.limit = dfrag->data_len;
-			len = dfrag->data_len - dfrag->already_sent;
-			while (len > 0) {
-				int ret = 0;
-
-				ret = mptcp_sendmsg_frag(sk, ssk, dfrag, &info);
-				if (ret <= 0)
-					goto out;
-
-				info.sent += ret;
-				copied += ret;
-				len -= ret;
-				first = false;
-
-				mptcp_update_post_push(msk, dfrag, ret);
-			}
-			WRITE_ONCE(msk->first_pending, mptcp_send_next(sk));
-
-			if (msk->snd_burst <= 0 ||
-			    !sk_stream_memory_free(ssk) ||
-			    !mptcp_subflow_active(mptcp_subflow_ctx(ssk))) {
-				goto out;
-			}
-			mptcp_set_timeout(sk);
+		ret = __subflow_push_pending(sk, ssk, &info);
+		if (ret <= 0) {
+			if (ret == -EAGAIN)
+				goto again;
+			break;
 		}
 	}
 
@@ -1658,9 +1637,7 @@ static void __mptcp_subflow_push_pending(struct sock *sk, struct sock *ssk,
 	/* __mptcp_alloc_tx_skb could have released some wmem and we are
 	 * not going to flush it via release_sock()
 	 */
-	if (copied) {
-		tcp_push(ssk, 0, info.mss_now, tcp_sk(ssk)->nonagle,
-			 info.size_goal);
+	if (ret > 0) {
 		if (!mptcp_timer_pending(sk))
 			mptcp_reset_timer(sk);
 
Signature · 1 line
2.35.3
#15 lore raw

PATCH v6 11/13 mptcp: multi subflows subflow_push_pending

In reply to Geliang Tang

This patch adds the multiple subflows support for __mptcp_subflow_push_pending(). Use mptcp_sched_get_send() wrapper instead of mptcp_subflow_get_send() in it.

Check the subflow scheduled flags to test which subflow or subflows are picked by the scheduler, use them to send data.

---
 net/mptcp/protocol.c | 49 +++++++++++++++++++++++++++++---------------
 1 file changed, 33 insertions(+), 16 deletions(-)
net/mptcp/protocol.c +33 -16
index e9eb1da42418..6a2d9972dec1 100644
@@ -1604,7 +1604,6 @@ static void __mptcp_subflow_push_pending(struct sock *sk, struct sock *ssk,
 	struct mptcp_sendmsg_info info = {
 		.data_lock_held = true,
 	};
-	struct sock *xmit_ssk;
 	int ret = 0;
 
 	info.flags = 0;
@@ -1613,23 +1612,41 @@ static void __mptcp_subflow_push_pending(struct sock *sk, struct sock *ssk,
 		/* check for a different subflow usage only after
 		 * spooling the first chunk of data
 		 */
-		xmit_ssk = first ? ssk : mptcp_subflow_get_send(msk);
-		if (!xmit_ssk)
-			goto out;
-		if (xmit_ssk != ssk) {
-			subflow = mptcp_subflow_ctx(xmit_ssk);
-			if (mptcp_subflow_has_delegated_action(subflow))
-				goto out;
-			mptcp_subflow_delegate(subflow,
-					       MPTCP_DELEGATE_SEND);
-			goto out;
+		if (first) {
+			ret = __subflow_push_pending(sk, ssk, &info);
+			if (ret <= 0) {
+				if (ret == -EAGAIN)
+					goto again;
+				break;
+			}
+			first = false;
+			continue;
 		}
 
-		ret = __subflow_push_pending(sk, ssk, &info);
-		if (ret <= 0) {
-			if (ret == -EAGAIN)
-				goto again;
-			break;
+		if (mptcp_sched_get_send(msk))
+			goto out;
+
+		mptcp_for_each_subflow(msk, subflow) {
+			if (READ_ONCE(subflow->scheduled)) {
+				struct sock *xmit_ssk = mptcp_subflow_tcp_sock(subflow);
+
+				if (xmit_ssk != ssk) {
+					if (mptcp_subflow_has_delegated_action(subflow))
+						goto out;
+					mptcp_subflow_delegate(subflow,
+							       MPTCP_DELEGATE_SEND);
+					mptcp_subflow_set_scheduled(subflow, false);
+					continue;
+				}
+
+				ret = __subflow_push_pending(sk, ssk, &info);
+				if (ret <= 0) {
+					if (ret == -EAGAIN)
+						goto again;
+					goto out;
+				}
+				mptcp_subflow_set_scheduled(subflow, false);
+			}
 		}
 	}
 
Signature · 1 line
2.35.3
#16 lore raw

PATCH v6 12/13 mptcp: multi subflows retrans support

In reply to Geliang Tang

This patch adds the multiple subflows support for __mptcp_retrans(). In it, use sched_get_retrans() wrapper instead of mptcp_subflow_get_retrans().

Iterate each subflow of msk, check the scheduled flag to test if it is picked by the scheduler. If so, use it to retrans data.

---
 net/mptcp/protocol.c | 61 +++++++++++++++++++++++++++-----------------
 1 file changed, 38 insertions(+), 23 deletions(-)
net/mptcp/protocol.c +38 -23
index 6a2d9972dec1..33ae11ca6c3c 100644
@@ -2478,16 +2478,17 @@ static void mptcp_check_fastclose(struct mptcp_sock *msk)
 static void __mptcp_retrans(struct sock *sk)
 {
 	struct mptcp_sock *msk = mptcp_sk(sk);
+	struct mptcp_subflow_context *subflow;
 	struct mptcp_sendmsg_info info = {};
 	struct mptcp_data_frag *dfrag;
-	size_t copied = 0;
 	struct sock *ssk;
-	int ret;
+	int ret, err;
+	u16 len = 0;
 
 	mptcp_clean_una_wakeup(sk);
 
 	/* first check ssk: need to kick "stale" logic */
-	ssk = mptcp_subflow_get_retrans(msk);
+	err = mptcp_sched_get_retrans(msk);
 	dfrag = mptcp_rtx_head(sk);
 	if (!dfrag) {
 		if (mptcp_data_fin_enabled(msk)) {
@@ -2506,31 +2507,45 @@ static void __mptcp_retrans(struct sock *sk)
 		goto reset_timer;
 	}
 
-	if (!ssk)
+	if (err)
 		goto reset_timer;
 
-	lock_sock(ssk);
+	mptcp_for_each_subflow(msk, subflow) {
+		if (READ_ONCE(subflow->scheduled)) {
+			u16 copied = 0;
+
+			ssk = mptcp_subflow_tcp_sock(subflow);
+			if (!ssk)
+				goto reset_timer;
+
+			lock_sock(ssk);
+
+			/* limit retransmission to the bytes already sent on some subflows */
+			info.sent = 0;
+			info.limit = READ_ONCE(msk->csum_enabled) ? dfrag->data_len :
+								    dfrag->already_sent;
+			while (info.sent < info.limit) {
+				ret = mptcp_sendmsg_frag(sk, ssk, dfrag, &info);
+				if (ret <= 0)
+					break;
+
+				MPTCP_INC_STATS(sock_net(sk), MPTCP_MIB_RETRANSSEGS);
+				copied += ret;
+				info.sent += ret;
+			}
+			if (copied) {
+				len = max(copied, len);
+				tcp_push(ssk, 0, info.mss_now, tcp_sk(ssk)->nonagle,
+					 info.size_goal);
+				WRITE_ONCE(msk->allow_infinite_fallback, false);
+			}
 
-	/* limit retransmission to the bytes already sent on some subflows */
-	info.sent = 0;
-	info.limit = READ_ONCE(msk->csum_enabled) ? dfrag->data_len : dfrag->already_sent;
-	while (info.sent < info.limit) {
-		ret = mptcp_sendmsg_frag(sk, ssk, dfrag, &info);
-		if (ret <= 0)
-			break;
+			release_sock(ssk);
 
-		MPTCP_INC_STATS(sock_net(sk), MPTCP_MIB_RETRANSSEGS);
-		copied += ret;
-		info.sent += ret;
-	}
-	if (copied) {
-		dfrag->already_sent = max(dfrag->already_sent, info.sent);
-		tcp_push(ssk, 0, info.mss_now, tcp_sk(ssk)->nonagle,
-			 info.size_goal);
-		WRITE_ONCE(msk->allow_infinite_fallback, false);
+			mptcp_subflow_set_scheduled(subflow, false);
+		}
 	}
-
-	release_sock(ssk);
+	dfrag->already_sent = max(dfrag->already_sent, len);
 
 reset_timer:
 	mptcp_check_and_set_pending(sk);
Signature · 1 line
2.35.3
#17 lore raw

PATCH v6 13/13 Squash to "selftests/bpf: Add bpf_rr scheduler"

In reply to Geliang Tang

Use last_snd instead of msk->last_snd, then last_snd of struct mptcp_sock could be removed.

---
 tools/testing/selftests/bpf/bpf_tcp_helpers.h    | 1 -
 tools/testing/selftests/bpf/progs/mptcp_bpf_rr.c | 6 ++++--
 2 files changed, 4 insertions(+), 3 deletions(-)
tools/testing/selftests/bpf/bpf_tcp_helpers.h +0 -1
index c7d4a9a69cfc..c8792e6f125a 100644
@@ -257,7 +257,6 @@ struct mptcp_sched_ops {
 struct mptcp_sock {
 	struct inet_connection_sock	sk;
 
-	struct sock	*last_snd;
 	__u32		token;
 	struct sock	*first;
 	char		ca_name[TCP_CA_NAME_MAX];
tools/testing/selftests/bpf/progs/mptcp_bpf_rr.c +4 -2
index ce4e98f83e43..b7156f6aae8b 100644
@@ -5,6 +5,7 @@
 #include "bpf_tcp_helpers.h"
 
 char _license[] SEC("license") = "GPL";
+struct sock *last_snd = NULL;
 
 SEC("struct_ops/mptcp_sched_rr_init")
 void BPF_PROG(mptcp_sched_rr_init, const struct mptcp_sock *msk)
@@ -22,10 +23,10 @@ void BPF_STRUCT_OPS(bpf_rr_get_subflow, const struct mptcp_sock *msk,
 	int nr = 0;
 
 	for (int i = 0; i < MPTCP_SUBFLOWS_MAX; i++) {
-		if (!msk->last_snd || !data->contexts[i])
+		if (!last_snd || !data->contexts[i])
 			break;
 
-		if (data->contexts[i]->tcp_sock == msk->last_snd) {
+		if (data->contexts[i]->tcp_sock == last_snd) {
 			if (i + 1 == MPTCP_SUBFLOWS_MAX || !data->contexts[i + 1])
 				break;
 
@@ -35,6 +36,7 @@ void BPF_STRUCT_OPS(bpf_rr_get_subflow, const struct mptcp_sock *msk,
 	}
 
 	mptcp_subflow_set_scheduled(data->contexts[nr], true);
+	last_snd = data->contexts[nr]->tcp_sock;
 }
 
 SEC(".struct_ops")
Signature · 1 line
2.35.3
#18 lore raw

Re: Squash to "selftests/bpf: Add bpf_rr scheduler": Tests Results

Hi Geliang,

Thank you for your modifications, that's great!

Our CI did some validations and here is its report:

- KVM Validation: normal:
- Unstable: 2 failed test(s): packetdrill_add_addr selftest_mptcp_join 🔴:
- Task: https://cirrus-ci.com/task/6091106109620224
- Summary: https://api.cirrus-ci.com/v1/artifact/task/6091106109620224/summary/summary.txt

- KVM Validation: debug:
- Success! ✅:
- Task: https://cirrus-ci.com/task/5528156156198912
- Summary: https://api.cirrus-ci.com/v1/artifact/task/5528156156198912/summary/summary.txt

Initiator: Patchew Applier
Commits: https://github.com/multipath-tcp/mptcp_net-next/commits/329ea0dac7da

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-debug

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 (Tessares)