[PATCH mptcp-next 0/6] mptcp: Add userspace PM mode to bypass kernel PM

[PATCH mptcp-next 0/6] mptcp: Add userspace PM mode to bypass kernel PM

[Mat Martineau]

One part of supporting userspace path managers is to prevent the
in-kernel PM from acting on userspace-managed MPTCP connections.

These patches:

 * Add a per-MPTCP-socket 'pm_type'. A mix of kernel-managed and
   userspace-managed connections are supported within each namespace.

 * Conditionally decouple incoming ADD_ADDR/RM_ADDR and subflow changes
   from the in-kernel PM. Netlink events are still triggered, and ADD_ADDR
   echo handling is still in kernel code even if path management is
   otherwise handled in userspace.

 * Add a sysctl for setting the per-namespace default for in-kernel vs
   userspace path management of new MPTCP sockets. This is an integer
   value to allow extensibility.

 * Add selftests to confirm that the in-kernel PM is bypassed.

RFC -> v1: Changed sysctl from a bool to an integer, added patch 1
(cleanup) and patch 6 (selftests), fixed ADD_ADDR echo and initial
pm->subflows_allowed settings.


Mat Martineau (6):
  mptcp: Remove redundant assignments in path manager init
  mptcp: Add a member to mptcp_pm_data to track kernel vs userspace mode
  mptcp: Bypass kernel PM when userspace PM is enabled
  mptcp: Make kernel path manager check for userspace-managed sockets
  mptcp: Add a per-namespace sysctl to set the default path manager type
  selftests: mptcp: Add tests for userspace PM type

 Documentation/networking/mptcp-sysctl.rst     | 18 +++++
 net/mptcp/ctrl.c                              | 18 +++++
 net/mptcp/pm.c                                | 53 +++++++++-----
 net/mptcp/pm_netlink.c                        | 32 ++++-----
 net/mptcp/protocol.h                          | 11 ++-
 .../testing/selftests/net/mptcp/mptcp_join.sh | 70 ++++++++++++++++++-
 6 files changed, 164 insertions(+), 38 deletions(-)


base-commit: 98fb1b1ee27f9173f03e2787422ad1045b988a96
-- 
2.34.0

[PATCH mptcp-next 1/6] mptcp: Remove redundant assignments in path manager init

[Mat Martineau]

A few members of the mptcp_pm_data struct were assigned to hard-coded
values in mptcp_pm_data_init(), and then immediately changed in
mptcp_pm_nl_data_init().

Instead, flatten all the initialization in to mptcp_pm_data_init().

Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
---
 net/mptcp/pm.c         | 38 +++++++++++++++++++++-----------------
 net/mptcp/pm_netlink.c | 12 ------------
 net/mptcp/protocol.h   |  1 -
 3 files changed, 21 insertions(+), 30 deletions(-)

diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
index 86b38a830b4c..467fbb2a4eba 100644
--- a/net/mptcp/pm.c
+++ b/net/mptcp/pm.c
@@ -364,23 +364,27 @@ void mptcp_pm_subflow_chk_stale(const struct mptcp_sock *msk, struct sock *ssk)
 
 void mptcp_pm_data_init(struct mptcp_sock *msk)
 {
-	msk->pm.add_addr_signaled = 0;
-	msk->pm.add_addr_accepted = 0;
-	msk->pm.local_addr_used = 0;
-	msk->pm.subflows = 0;
-	msk->pm.rm_list_tx.nr = 0;
-	msk->pm.rm_list_rx.nr = 0;
-	WRITE_ONCE(msk->pm.work_pending, false);
-	WRITE_ONCE(msk->pm.addr_signal, 0);
-	WRITE_ONCE(msk->pm.accept_addr, false);
-	WRITE_ONCE(msk->pm.accept_subflow, false);
-	WRITE_ONCE(msk->pm.remote_deny_join_id0, false);
-	msk->pm.status = 0;
-
-	spin_lock_init(&msk->pm.lock);
-	INIT_LIST_HEAD(&msk->pm.anno_list);
-
-	mptcp_pm_nl_data_init(msk);
+	bool subflows_allowed = !!mptcp_pm_get_subflows_max(msk);
+	struct mptcp_pm_data *pm = &msk->pm;
+
+	pm->add_addr_signaled = 0;
+	pm->add_addr_accepted = 0;
+	pm->local_addr_used = 0;
+	pm->subflows = 0;
+	pm->rm_list_tx.nr = 0;
+	pm->rm_list_rx.nr = 0;
+	WRITE_ONCE(pm->work_pending,
+		   (!!mptcp_pm_get_local_addr_max(msk) && subflows_allowed) ||
+		   !!mptcp_pm_get_add_addr_signal_max(msk));
+	WRITE_ONCE(pm->addr_signal, 0);
+	WRITE_ONCE(pm->accept_addr,
+		   !!mptcp_pm_get_add_addr_accept_max(msk) && subflows_allowed);
+	WRITE_ONCE(pm->accept_subflow, subflows_allowed);
+	WRITE_ONCE(pm->remote_deny_join_id0, false);
+	pm->status = 0;
+
+	spin_lock_init(&pm->lock);
+	INIT_LIST_HEAD(&pm->anno_list);
 }
 
 void __init mptcp_pm_init(void)
diff --git a/net/mptcp/pm_netlink.c b/net/mptcp/pm_netlink.c
index 7b96be1e9f14..bab78e2f5299 100644
--- a/net/mptcp/pm_netlink.c
+++ b/net/mptcp/pm_netlink.c
@@ -954,18 +954,6 @@ int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk, struct sock_common *skc)
 	return ret;
 }
 
-void mptcp_pm_nl_data_init(struct mptcp_sock *msk)
-{
-	struct mptcp_pm_data *pm = &msk->pm;
-	bool subflows;
-
-	subflows = !!mptcp_pm_get_subflows_max(msk);
-	WRITE_ONCE(pm->work_pending, (!!mptcp_pm_get_local_addr_max(msk) && subflows) ||
-		   !!mptcp_pm_get_add_addr_signal_max(msk));
-	WRITE_ONCE(pm->accept_addr, !!mptcp_pm_get_add_addr_accept_max(msk) && subflows);
-	WRITE_ONCE(pm->accept_subflow, subflows);
-}
-
 #define MPTCP_PM_CMD_GRP_OFFSET       0
 #define MPTCP_PM_EV_GRP_OFFSET        1
 
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index 906509c6cde5..092cb77a64d8 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -803,7 +803,6 @@ bool mptcp_pm_rm_addr_signal(struct mptcp_sock *msk, unsigned int remaining,
 int mptcp_pm_get_local_id(struct mptcp_sock *msk, struct sock_common *skc);
 
 void __init mptcp_pm_nl_init(void);
-void mptcp_pm_nl_data_init(struct mptcp_sock *msk);
 void mptcp_pm_nl_work(struct mptcp_sock *msk);
 void mptcp_pm_nl_rm_subflow_received(struct mptcp_sock *msk,
 				     const struct mptcp_rm_list *rm_list);
-- 
2.34.0

[PATCH mptcp-next 2/6] mptcp: Add a member to mptcp_pm_data to track kernel vs userspace mode

[Mat Martineau]

When adding support for netlink path management commands, the kernel
needs to know whether paths are being controlled by the in-kernel path
manager or a userspace PM.

Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
---
 net/mptcp/pm.c       | 4 ++++
 net/mptcp/protocol.h | 9 +++++++++
 2 files changed, 13 insertions(+)

diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
index 467fbb2a4eba..3a3800771aca 100644
--- a/net/mptcp/pm.c
+++ b/net/mptcp/pm.c
@@ -373,6 +373,10 @@ void mptcp_pm_data_init(struct mptcp_sock *msk)
 	pm->subflows = 0;
 	pm->rm_list_tx.nr = 0;
 	pm->rm_list_rx.nr = 0;
+	WRITE_ONCE(pm->pm_type, MPTCP_PM_TYPE_KERNEL);
+	/* pm->work_pending must be only be set to 'true' when
+	 * pm->pm_type is set to MPTCP_PM_TYPE_KERNEL
+	 */
 	WRITE_ONCE(pm->work_pending,
 		   (!!mptcp_pm_get_local_addr_max(msk) && subflows_allowed) ||
 		   !!mptcp_pm_get_add_addr_signal_max(msk));
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index 092cb77a64d8..edaf0cc75474 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -178,6 +178,14 @@ enum mptcp_pm_status {
 	MPTCP_PM_SUBFLOW_ESTABLISHED,
 };
 
+enum mptcp_pm_type {
+	MPTCP_PM_TYPE_KERNEL = 0,
+	MPTCP_PM_TYPE_USERSPACE,
+
+	__MPTCP_PM_TYPE_NR,
+	__MPTCP_PM_TYPE_MAX = __MPTCP_PM_TYPE_NR - 1,
+};
+
 enum mptcp_addr_signal_status {
 	MPTCP_ADD_ADDR_SIGNAL,
 	MPTCP_ADD_ADDR_ECHO,
@@ -200,6 +208,7 @@ struct mptcp_pm_data {
 	u8		add_addr_signaled;
 	u8		add_addr_accepted;
 	u8		local_addr_used;
+	u8		pm_type;
 	u8		subflows;
 	u8		status;
 	struct mptcp_rm_list rm_list_tx;
-- 
2.34.0

[PATCH mptcp-next 3/6] mptcp: Bypass kernel PM when userspace PM is enabled

[Mat Martineau]

When a MPTCP connection is managed by a userspace PM, bypass the kernel
PM for incoming advertisements and subflow events. Netlink events are
still sent to userspace.

Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
---
 net/mptcp/pm.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
index 3a3800771aca..7f96a888e2d5 100644
--- a/net/mptcp/pm.c
+++ b/net/mptcp/pm.c
@@ -189,7 +189,8 @@ void mptcp_pm_add_addr_received(struct mptcp_sock *msk,
 
 	spin_lock_bh(&pm->lock);
 
-	if (!READ_ONCE(pm->accept_addr)) {
+	if (!READ_ONCE(pm->accept_addr) ||
+	    READ_ONCE(pm->pm_type) != MPTCP_PM_TYPE_KERNEL) {
 		mptcp_pm_announce_addr(msk, addr, true);
 		mptcp_pm_add_addr_send_ack(msk);
 	} else if (mptcp_pm_schedule_work(msk, MPTCP_PM_ADD_ADDR_RECEIVED)) {
@@ -233,6 +234,9 @@ void mptcp_pm_rm_addr_received(struct mptcp_sock *msk,
 	for (i = 0; i < rm_list->nr; i++)
 		mptcp_event_addr_removed(msk, rm_list->ids[i]);
 
+	if (READ_ONCE(pm->pm_type) != MPTCP_PM_TYPE_KERNEL)
+		return;
+
 	spin_lock_bh(&pm->lock);
 	mptcp_pm_schedule_work(msk, MPTCP_PM_RM_ADDR_RECEIVED);
 	pm->rm_list_rx = *rm_list;
-- 
2.34.0

[PATCH mptcp-next 4/6] mptcp: Make kernel path manager check for userspace-managed sockets

[Mat Martineau]

Userspace-managed sockets should not have their subflows or
advertisements changed by the kernel path manager.

Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
---
 net/mptcp/pm_netlink.c | 20 ++++++++++++++------
 1 file changed, 14 insertions(+), 6 deletions(-)

diff --git a/net/mptcp/pm_netlink.c b/net/mptcp/pm_netlink.c
index bab78e2f5299..95880cff9764 100644
--- a/net/mptcp/pm_netlink.c
+++ b/net/mptcp/pm_netlink.c
@@ -1119,7 +1119,8 @@ static int mptcp_nl_add_subflow_or_signal_addr(struct net *net)
 	while ((msk = mptcp_token_iter_next(net, &s_slot, &s_num)) != NULL) {
 		struct sock *sk = (struct sock *)msk;
 
-		if (!READ_ONCE(msk->fully_established))
+		if (!READ_ONCE(msk->fully_established) ||
+		    (READ_ONCE(msk->pm.pm_type) != MPTCP_PM_TYPE_KERNEL))
 			goto next;
 
 		lock_sock(sk);
@@ -1257,6 +1258,9 @@ static int mptcp_nl_remove_subflow_and_signal_addr(struct net *net,
 		struct sock *sk = (struct sock *)msk;
 		bool remove_subflow;
 
+		if (READ_ONCE(msk->pm.pm_type) != MPTCP_PM_TYPE_KERNEL)
+			goto next;
+
 		if (list_empty(&msk->conn_list)) {
 			mptcp_pm_remove_anno_addr(msk, addr, false);
 			goto next;
@@ -1298,7 +1302,8 @@ static int mptcp_nl_remove_id_zero_address(struct net *net,
 		struct sock *sk = (struct sock *)msk;
 		struct mptcp_addr_info msk_local;
 
-		if (list_empty(&msk->conn_list))
+		if (list_empty(&msk->conn_list) ||
+		    (READ_ONCE(msk->pm.pm_type) != MPTCP_PM_TYPE_KERNEL))
 			goto next;
 
 		local_address((struct sock_common *)msk, &msk_local);
@@ -1407,9 +1412,11 @@ static void mptcp_nl_remove_addrs_list(struct net *net,
 	while ((msk = mptcp_token_iter_next(net, &s_slot, &s_num)) != NULL) {
 		struct sock *sk = (struct sock *)msk;
 
-		lock_sock(sk);
-		mptcp_pm_remove_addrs_and_subflows(msk, rm_list);
-		release_sock(sk);
+		if (READ_ONCE(msk->pm.pm_type) == MPTCP_PM_TYPE_KERNEL) {
+			lock_sock(sk);
+			mptcp_pm_remove_addrs_and_subflows(msk, rm_list);
+			release_sock(sk);
+		}
 
 		sock_put(sk);
 		cond_resched();
@@ -1671,7 +1678,8 @@ static int mptcp_nl_addr_backup(struct net *net,
 	while ((msk = mptcp_token_iter_next(net, &s_slot, &s_num)) != NULL) {
 		struct sock *sk = (struct sock *)msk;
 
-		if (list_empty(&msk->conn_list))
+		if (list_empty(&msk->conn_list) ||
+		    (READ_ONCE(msk->pm.pm_type) != MPTCP_PM_TYPE_KERNEL))
 			goto next;
 
 		lock_sock(sk);
-- 
2.34.0

[PATCH mptcp-next 5/6] mptcp: Add a per-namespace sysctl to set the default path manager type

[Mat Martineau]

The new net.mptcp.pm_type sysctl determines which path manager will be
used by each newly-created MPTCP socket.

Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
---
 Documentation/networking/mptcp-sysctl.rst | 18 ++++++++++++++++++
 net/mptcp/ctrl.c                          | 18 ++++++++++++++++++
 net/mptcp/pm.c                            | 13 +++++++++----
 net/mptcp/protocol.h                      |  1 +
 4 files changed, 46 insertions(+), 4 deletions(-)

diff --git a/Documentation/networking/mptcp-sysctl.rst b/Documentation/networking/mptcp-sysctl.rst
index b0d4da71e68e..e263dfcc4b40 100644
--- a/Documentation/networking/mptcp-sysctl.rst
+++ b/Documentation/networking/mptcp-sysctl.rst
@@ -46,6 +46,24 @@ allow_join_initial_addr_port - BOOLEAN
 
 	Default: 1
 
+pm_type - INTEGER
+
+	Set the default path manager type to use for each new MPTCP
+	socket. In-kernel path management will control subflow
+	connections and address advertisements according to
+	per-namespace values configured over the MPTCP netlink
+	API. Userspace path management puts per-MPTCP-connection subflow
+	connection decisions and address advertisements under control of
+	a privileged userspace program, at the cost of more netlink
+	traffic to propagate all of the related events and commands.
+
+	This is a per-namespace sysctl.
+
+	* 0 - In-kernel path manager
+	* 1 - Userspace path manager
+
+	Default: 0
+
 stale_loss_cnt - INTEGER
 	The number of MPTCP-level retransmission intervals with no traffic and
 	pending outstanding data on a given subflow required to declare it stale.
diff --git a/net/mptcp/ctrl.c b/net/mptcp/ctrl.c
index 8b235468c88f..64f1366aec5e 100644
--- a/net/mptcp/ctrl.c
+++ b/net/mptcp/ctrl.c
@@ -16,6 +16,8 @@
 #define MPTCP_SYSCTL_PATH "net/mptcp"
 
 static int mptcp_pernet_id;
+static int mptcp_pm_type_max = __MPTCP_PM_TYPE_MAX;
+
 struct mptcp_pernet {
 #ifdef CONFIG_SYSCTL
 	struct ctl_table_header *ctl_table_hdr;
@@ -26,6 +28,7 @@ struct mptcp_pernet {
 	u8 mptcp_enabled;
 	u8 checksum_enabled;
 	u8 allow_join_initial_addr_port;
+	u8 pm_type;
 };
 
 static struct mptcp_pernet *mptcp_get_pernet(const struct net *net)
@@ -58,6 +61,11 @@ unsigned int mptcp_stale_loss_cnt(const struct net *net)
 	return mptcp_get_pernet(net)->stale_loss_cnt;
 }
 
+int mptcp_get_pm_type(const struct net *net)
+{
+	return mptcp_get_pernet(net)->pm_type;
+}
+
 static void mptcp_pernet_set_defaults(struct mptcp_pernet *pernet)
 {
 	pernet->mptcp_enabled = 1;
@@ -65,6 +73,7 @@ static void mptcp_pernet_set_defaults(struct mptcp_pernet *pernet)
 	pernet->checksum_enabled = 0;
 	pernet->allow_join_initial_addr_port = 1;
 	pernet->stale_loss_cnt = 4;
+	pernet->pm_type = MPTCP_PM_TYPE_KERNEL;
 }
 
 #ifdef CONFIG_SYSCTL
@@ -108,6 +117,14 @@ static struct ctl_table mptcp_sysctl_table[] = {
 		.mode = 0644,
 		.proc_handler = proc_douintvec_minmax,
 	},
+	{
+		.procname = "pm_type",
+		.maxlen = sizeof(u8),
+		.mode = 0644,
+		.proc_handler = proc_dou8vec_minmax,
+		.extra1       = SYSCTL_ZERO,
+		.extra2       = &mptcp_pm_type_max
+	},
 	{}
 };
 
@@ -128,6 +145,7 @@ static int mptcp_pernet_new_table(struct net *net, struct mptcp_pernet *pernet)
 	table[2].data = &pernet->checksum_enabled;
 	table[3].data = &pernet->allow_join_initial_addr_port;
 	table[4].data = &pernet->stale_loss_cnt;
+	table[5].data = &pernet->pm_type;
 
 	hdr = register_net_sysctl(net, MPTCP_SYSCTL_PATH, table);
 	if (!hdr)
diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
index 7f96a888e2d5..7ef27d41ad56 100644
--- a/net/mptcp/pm.c
+++ b/net/mptcp/pm.c
@@ -368,8 +368,12 @@ void mptcp_pm_subflow_chk_stale(const struct mptcp_sock *msk, struct sock *ssk)
 
 void mptcp_pm_data_init(struct mptcp_sock *msk)
 {
-	bool subflows_allowed = !!mptcp_pm_get_subflows_max(msk);
+	u8 pm_type = mptcp_get_pm_type(sock_net((struct sock *)msk));
 	struct mptcp_pm_data *pm = &msk->pm;
+	bool subflows_allowed;
+
+	subflows_allowed = !!mptcp_pm_get_subflows_max(msk) &&
+		pm_type == MPTCP_PM_TYPE_KERNEL;
 
 	pm->add_addr_signaled = 0;
 	pm->add_addr_accepted = 0;
@@ -377,13 +381,14 @@ void mptcp_pm_data_init(struct mptcp_sock *msk)
 	pm->subflows = 0;
 	pm->rm_list_tx.nr = 0;
 	pm->rm_list_rx.nr = 0;
-	WRITE_ONCE(pm->pm_type, MPTCP_PM_TYPE_KERNEL);
+	WRITE_ONCE(pm->pm_type, pm_type);
 	/* pm->work_pending must be only be set to 'true' when
 	 * pm->pm_type is set to MPTCP_PM_TYPE_KERNEL
 	 */
 	WRITE_ONCE(pm->work_pending,
-		   (!!mptcp_pm_get_local_addr_max(msk) && subflows_allowed) ||
-		   !!mptcp_pm_get_add_addr_signal_max(msk));
+		   pm_type == MPTCP_PM_TYPE_KERNEL &&
+		   ((!!mptcp_pm_get_local_addr_max(msk) && subflows_allowed) ||
+		    !!mptcp_pm_get_add_addr_signal_max(msk)));
 	WRITE_ONCE(pm->addr_signal, 0);
 	WRITE_ONCE(pm->accept_addr,
 		   !!mptcp_pm_get_add_addr_accept_max(msk) && subflows_allowed);
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index edaf0cc75474..b56ad45703dd 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -563,6 +563,7 @@ unsigned int mptcp_get_add_addr_timeout(const struct net *net);
 int mptcp_is_checksum_enabled(const struct net *net);
 int mptcp_allow_join_id0(const struct net *net);
 unsigned int mptcp_stale_loss_cnt(const struct net *net);
+int mptcp_get_pm_type(const struct net *net);
 void mptcp_subflow_fully_established(struct mptcp_subflow_context *subflow,
 				     struct mptcp_options_received *mp_opt);
 bool __mptcp_retransmit_pending_data(struct sock *sk);
-- 
2.34.0

[PATCH mptcp-next 6/6] selftests: mptcp: Add tests for userspace PM type

[Mat Martineau]

These tests ensure that the in-kernel path manager is bypassed when
the userspace path manager is configured. Kernel code is still
responsible for ADD_ADDR echo, so also make sure that's working.

Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
---
 .../testing/selftests/net/mptcp/mptcp_join.sh | 70 ++++++++++++++++++-
 1 file changed, 69 insertions(+), 1 deletion(-)

diff --git a/tools/testing/selftests/net/mptcp/mptcp_join.sh b/tools/testing/selftests/net/mptcp/mptcp_join.sh
index 2684ef9c0d42..7df9ddb307a8 100755
--- a/tools/testing/selftests/net/mptcp/mptcp_join.sh
+++ b/tools/testing/selftests/net/mptcp/mptcp_join.sh
@@ -50,6 +50,7 @@ init()
 		ip netns add $netns || exit $ksft_skip
 		ip -net $netns link set lo up
 		ip netns exec $netns sysctl -q net.mptcp.enabled=1
+		ip netns exec $netns sysctl -q net.mptcp.pm_type=0
 		ip netns exec $netns sysctl -q net.ipv4.conf.all.rp_filter=0
 		ip netns exec $netns sysctl -q net.ipv4.conf.default.rp_filter=0
 		if [ $checksum -eq 1 ]; then
@@ -1837,6 +1838,68 @@ fullmesh_tests()
 	chk_add_nr 1 1
 }
 
+userspace_tests()
+{
+	# userspace pm type prevents add_addr
+	reset
+	ip netns exec $ns1 sysctl -q net.mptcp.pm_type=1
+	ip netns exec $ns1 ./pm_nl_ctl limits 0 2
+	ip netns exec $ns2 ./pm_nl_ctl limits 0 2
+	ip netns exec $ns1 ./pm_nl_ctl add 10.0.2.1 flags signal
+	run_tests $ns1 $ns2 10.0.1.1
+	chk_join_nr "userspace pm type prevents add_addr" 0 0 0
+	chk_add_nr 0 0
+
+	# userspace pm type echoes add_addr
+	reset
+	ip netns exec $ns2 sysctl -q net.mptcp.pm_type=1
+	ip netns exec $ns1 ./pm_nl_ctl limits 0 2
+	ip netns exec $ns2 ./pm_nl_ctl limits 0 2
+	ip netns exec $ns1 ./pm_nl_ctl add 10.0.2.1 flags signal
+	run_tests $ns1 $ns2 10.0.1.1
+	chk_join_nr "userspace pm type echoes add_addr" 0 0 0
+	chk_add_nr 1 1
+
+	# userspace pm type rejects join
+	reset
+	ip netns exec $ns1 sysctl -q net.mptcp.pm_type=1
+	ip netns exec $ns1 ./pm_nl_ctl limits 1 1
+	ip netns exec $ns2 ./pm_nl_ctl limits 1 1
+	ip netns exec $ns2 ./pm_nl_ctl add 10.0.3.2 flags subflow
+	run_tests $ns1 $ns2 10.0.1.1
+	chk_join_nr "userspace pm type rejects join" 1 1 0
+
+	# userspace pm type does not send join
+	reset
+	ip netns exec $ns2 sysctl -q net.mptcp.pm_type=1
+	ip netns exec $ns1 ./pm_nl_ctl limits 1 1
+	ip netns exec $ns2 ./pm_nl_ctl limits 1 1
+	ip netns exec $ns2 ./pm_nl_ctl add 10.0.3.2 flags subflow
+	run_tests $ns1 $ns2 10.0.1.1
+	chk_join_nr "userspace pm type does not send join" 0 0 0
+
+	# userspace pm type prevents mp_prio
+	reset
+	ip netns exec $ns1 sysctl -q net.mptcp.pm_type=1
+	ip netns exec $ns1 ./pm_nl_ctl limits 1 1
+	ip netns exec $ns2 ./pm_nl_ctl limits 1 1
+	ip netns exec $ns2 ./pm_nl_ctl add 10.0.3.2 flags subflow
+	run_tests $ns1 $ns2 10.0.1.1 0 0 0 slow backup
+	chk_join_nr "userspace pm type prevents mp_prio" 1 1 0
+	chk_prio_nr 0 0
+
+	# userspace pm type prevents rm_addr
+	reset
+	ip netns exec $ns1 sysctl -q net.mptcp.pm_type=1
+	ip netns exec $ns2 sysctl -q net.mptcp.pm_type=1
+	ip netns exec $ns1 ./pm_nl_ctl limits 0 1
+	ip netns exec $ns2 ./pm_nl_ctl limits 0 1
+	ip netns exec $ns2 ./pm_nl_ctl add 10.0.3.2 flags subflow
+	run_tests $ns1 $ns2 10.0.1.1 0 0 -1 slow
+	chk_join_nr "userspace pm type prevents rm_addr" 0 0 0
+	chk_rm_nr 0 0
+}
+
 all_tests()
 {
 	subflows_tests
@@ -1853,6 +1916,7 @@ all_tests()
 	checksum_tests
 	deny_join_id0_tests
 	fullmesh_tests
+	userspace_tests
 }
 
 usage()
@@ -1872,6 +1936,7 @@ usage()
 	echo "  -S checksum_tests"
 	echo "  -d deny_join_id0_tests"
 	echo "  -m fullmesh_tests"
+	echo "  -u userspace_tests"
 	echo "  -c capture pcap files"
 	echo "  -C enable data checksum"
 	echo "  -h help"
@@ -1907,7 +1972,7 @@ if [ $do_all_tests -eq 1 ]; then
 	exit $ret
 fi
 
-while getopts 'fsltra64bpkdmchCS' opt; do
+while getopts 'fsltra64bpkdmuchCS' opt; do
 	case $opt in
 		f)
 			subflows_tests
@@ -1951,6 +2016,9 @@ while getopts 'fsltra64bpkdmchCS' opt; do
 		m)
 			fullmesh_tests
 			;;
+		u)
+			userspace_tests
+			;;
 		c)
 			;;
 		C)
-- 
2.34.0

Re: selftests: mptcp: Add tests for userspace PM type: Build Failure

[MPTCP CI]

Hi Mat,

Thank you for your modifications, that's great!

But sadly, our CI spotted some issues with it when trying to build it.

You can find more details there:

  https://patchwork.kernel.org/project/mptcp/patch/20211118013005.263660-7-mathew.j.martineau@linux.intel.com/
  https://github.com/multipath-tcp/mptcp_net-next/actions/runs/1476796774

Status: failure
Initiator: MPTCPimporter
Commits: https://github.com/multipath-tcp/mptcp_net-next/commits/7a651678e619

Feel free to reply to this email if you cannot access logs, if you need
some support to fix the error, if this doesn't seem to be caused by your
modifications or if the error is a false positive one.

Cheers,
MPTCP GH Action bot

Re: [PATCH mptcp-next 5/6] mptcp: Add a per-namespace sysctl to set the default path manager type

[Mat Martineau]

On Wed, 17 Nov 2021, Mat Martineau wrote:

> The new net.mptcp.pm_type sysctl determines which path manager will be
> used by each newly-created MPTCP socket.
>
> Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
> ---
> Documentation/networking/mptcp-sysctl.rst | 18 ++++++++++++++++++
> net/mptcp/ctrl.c                          | 18 ++++++++++++++++++
> net/mptcp/pm.c                            | 13 +++++++++----
> net/mptcp/protocol.h                      |  1 +
> 4 files changed, 46 insertions(+), 4 deletions(-)
>
> diff --git a/Documentation/networking/mptcp-sysctl.rst b/Documentation/networking/mptcp-sysctl.rst
> index b0d4da71e68e..e263dfcc4b40 100644
> --- a/Documentation/networking/mptcp-sysctl.rst
> +++ b/Documentation/networking/mptcp-sysctl.rst
> @@ -46,6 +46,24 @@ allow_join_initial_addr_port - BOOLEAN
>
> 	Default: 1
>
> +pm_type - INTEGER
> +
> +	Set the default path manager type to use for each new MPTCP
> +	socket. In-kernel path management will control subflow
> +	connections and address advertisements according to
> +	per-namespace values configured over the MPTCP netlink
> +	API. Userspace path management puts per-MPTCP-connection subflow
> +	connection decisions and address advertisements under control of
> +	a privileged userspace program, at the cost of more netlink
> +	traffic to propagate all of the related events and commands.
> +
> +	This is a per-namespace sysctl.
> +
> +	* 0 - In-kernel path manager
> +	* 1 - Userspace path manager
> +
> +	Default: 0
> +
> stale_loss_cnt - INTEGER
> 	The number of MPTCP-level retransmission intervals with no traffic and
> 	pending outstanding data on a given subflow required to declare it stale.
> diff --git a/net/mptcp/ctrl.c b/net/mptcp/ctrl.c
> index 8b235468c88f..64f1366aec5e 100644
> --- a/net/mptcp/ctrl.c
> +++ b/net/mptcp/ctrl.c
> @@ -16,6 +16,8 @@
> #define MPTCP_SYSCTL_PATH "net/mptcp"
>
> static int mptcp_pernet_id;
> +static int mptcp_pm_type_max = __MPTCP_PM_TYPE_MAX;
> +

CI found an unused variable warning here - above line needs "#ifdef 
CONFIG_SYSCTL" around it.

- Mat

> struct mptcp_pernet {
> #ifdef CONFIG_SYSCTL
> 	struct ctl_table_header *ctl_table_hdr;
> @@ -26,6 +28,7 @@ struct mptcp_pernet {
> 	u8 mptcp_enabled;
> 	u8 checksum_enabled;
> 	u8 allow_join_initial_addr_port;
> +	u8 pm_type;
> };
>
> static struct mptcp_pernet *mptcp_get_pernet(const struct net *net)
> @@ -58,6 +61,11 @@ unsigned int mptcp_stale_loss_cnt(const struct net *net)
> 	return mptcp_get_pernet(net)->stale_loss_cnt;
> }
>
> +int mptcp_get_pm_type(const struct net *net)
> +{
> +	return mptcp_get_pernet(net)->pm_type;
> +}
> +
> static void mptcp_pernet_set_defaults(struct mptcp_pernet *pernet)
> {
> 	pernet->mptcp_enabled = 1;
> @@ -65,6 +73,7 @@ static void mptcp_pernet_set_defaults(struct mptcp_pernet *pernet)
> 	pernet->checksum_enabled = 0;
> 	pernet->allow_join_initial_addr_port = 1;
> 	pernet->stale_loss_cnt = 4;
> +	pernet->pm_type = MPTCP_PM_TYPE_KERNEL;
> }
>
> #ifdef CONFIG_SYSCTL
> @@ -108,6 +117,14 @@ static struct ctl_table mptcp_sysctl_table[] = {
> 		.mode = 0644,
> 		.proc_handler = proc_douintvec_minmax,
> 	},
> +	{
> +		.procname = "pm_type",
> +		.maxlen = sizeof(u8),
> +		.mode = 0644,
> +		.proc_handler = proc_dou8vec_minmax,
> +		.extra1       = SYSCTL_ZERO,
> +		.extra2       = &mptcp_pm_type_max
> +	},
> 	{}
> };
>
> @@ -128,6 +145,7 @@ static int mptcp_pernet_new_table(struct net *net, struct mptcp_pernet *pernet)
> 	table[2].data = &pernet->checksum_enabled;
> 	table[3].data = &pernet->allow_join_initial_addr_port;
> 	table[4].data = &pernet->stale_loss_cnt;
> +	table[5].data = &pernet->pm_type;
>
> 	hdr = register_net_sysctl(net, MPTCP_SYSCTL_PATH, table);
> 	if (!hdr)
> diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
> index 7f96a888e2d5..7ef27d41ad56 100644
> --- a/net/mptcp/pm.c
> +++ b/net/mptcp/pm.c
> @@ -368,8 +368,12 @@ void mptcp_pm_subflow_chk_stale(const struct mptcp_sock *msk, struct sock *ssk)
>
> void mptcp_pm_data_init(struct mptcp_sock *msk)
> {
> -	bool subflows_allowed = !!mptcp_pm_get_subflows_max(msk);
> +	u8 pm_type = mptcp_get_pm_type(sock_net((struct sock *)msk));
> 	struct mptcp_pm_data *pm = &msk->pm;
> +	bool subflows_allowed;
> +
> +	subflows_allowed = !!mptcp_pm_get_subflows_max(msk) &&
> +		pm_type == MPTCP_PM_TYPE_KERNEL;
>
> 	pm->add_addr_signaled = 0;
> 	pm->add_addr_accepted = 0;
> @@ -377,13 +381,14 @@ void mptcp_pm_data_init(struct mptcp_sock *msk)
> 	pm->subflows = 0;
> 	pm->rm_list_tx.nr = 0;
> 	pm->rm_list_rx.nr = 0;
> -	WRITE_ONCE(pm->pm_type, MPTCP_PM_TYPE_KERNEL);
> +	WRITE_ONCE(pm->pm_type, pm_type);
> 	/* pm->work_pending must be only be set to 'true' when
> 	 * pm->pm_type is set to MPTCP_PM_TYPE_KERNEL
> 	 */
> 	WRITE_ONCE(pm->work_pending,
> -		   (!!mptcp_pm_get_local_addr_max(msk) && subflows_allowed) ||
> -		   !!mptcp_pm_get_add_addr_signal_max(msk));
> +		   pm_type == MPTCP_PM_TYPE_KERNEL &&
> +		   ((!!mptcp_pm_get_local_addr_max(msk) && subflows_allowed) ||
> +		    !!mptcp_pm_get_add_addr_signal_max(msk)));
> 	WRITE_ONCE(pm->addr_signal, 0);
> 	WRITE_ONCE(pm->accept_addr,
> 		   !!mptcp_pm_get_add_addr_accept_max(msk) && subflows_allowed);
> diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
> index edaf0cc75474..b56ad45703dd 100644
> --- a/net/mptcp/protocol.h
> +++ b/net/mptcp/protocol.h
> @@ -563,6 +563,7 @@ unsigned int mptcp_get_add_addr_timeout(const struct net *net);
> int mptcp_is_checksum_enabled(const struct net *net);
> int mptcp_allow_join_id0(const struct net *net);
> unsigned int mptcp_stale_loss_cnt(const struct net *net);
> +int mptcp_get_pm_type(const struct net *net);
> void mptcp_subflow_fully_established(struct mptcp_subflow_context *subflow,
> 				     struct mptcp_options_received *mp_opt);
> bool __mptcp_retransmit_pending_data(struct sock *sk);
> -- 
> 2.34.0
>
>

--
Mat Martineau
Intel

Re: [PATCH mptcp-next 3/6] mptcp: Bypass kernel PM when userspace PM is enabled

[Kishen Maloor]

On 11/17/21 5:30 PM, Mat Martineau wrote:
> When a MPTCP connection is managed by a userspace PM, bypass the kernel
> PM for incoming advertisements and subflow events. Netlink events are
> still sent to userspace.
> 
> Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
> ---
>  net/mptcp/pm.c | 6 +++++-
>  1 file changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
> index 3a3800771aca..7f96a888e2d5 100644
> --- a/net/mptcp/pm.c
> +++ b/net/mptcp/pm.c
> @@ -189,7 +189,8 @@ void mptcp_pm_add_addr_received(struct mptcp_sock *msk,
>  
>  	spin_lock_bh(&pm->lock);
>  
> -	if (!READ_ONCE(pm->accept_addr)) {
> +	if (!READ_ONCE(pm->accept_addr) ||
> +	    READ_ONCE(pm->pm_type) != MPTCP_PM_TYPE_KERNEL) {
>  		mptcp_pm_announce_addr(msk, addr, true);
>  		mptcp_pm_add_addr_send_ack(msk);
>  	} else if (mptcp_pm_schedule_work(msk, MPTCP_PM_ADD_ADDR_RECEIVED)) {
> @@ -233,6 +234,9 @@ void mptcp_pm_rm_addr_received(struct mptcp_sock *msk,
>  	for (i = 0; i < rm_list->nr; i++)
>  		mptcp_event_addr_removed(msk, rm_list->ids[i]);
>  
> +	if (READ_ONCE(pm->pm_type) != MPTCP_PM_TYPE_KERNEL)
> +		return;
> +

The flow which ensues from MPTCP_PM_RM_ADDR_RECEIVED can be fully utilized with userspace PMs, so the above check for pm_type can be removed.

>  	spin_lock_bh(&pm->lock);
>  	mptcp_pm_schedule_work(msk, MPTCP_PM_RM_ADDR_RECEIVED);
>  	pm->rm_list_rx = *rm_list;
>