* [PATCH mptcp-next v4 0/6] mptcp: Add userspace PM mode to bypass kernel PM
@ 2021-12-16 0:33 Mat Martineau
2021-12-16 0:33 ` [PATCH mptcp-next v4 1/6] mptcp: Remove redundant assignments in path manager init Mat Martineau
` (7 more replies)
0 siblings, 8 replies; 9+ messages in thread
From: Mat Martineau @ 2021-12-16 0:33 UTC (permalink / raw)
To: mptcp; +Cc: 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.
v1 -> v2: Rebased on latest export branch, removed extra kernel-mode
check when receiving ADD_ADDR, and fix !CONFIG_SYSCTL build.
v2 -> v3: Rebased on latest export branch, use helper function for PM
mode, clarify initialization logic for userspace vs. kernel PM,
add a Tested-by tag from v2 review.
v3 -> v4: Rebased on latest export branch and resolved conflicts
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 | 21 ++++++
net/mptcp/pm.c | 50 +++++++++----
net/mptcp/pm_netlink.c | 30 ++++----
net/mptcp/protocol.h | 16 ++++-
.../testing/selftests/net/mptcp/mptcp_join.sh | 70 ++++++++++++++++++-
6 files changed, 170 insertions(+), 35 deletions(-)
base-commit: 81706660508b6caa94d0fbeb8875882f155c73a9
--
2.34.1
^ permalink raw reply [flat|nested] 9+ messages in thread
* [PATCH mptcp-next v4 1/6] mptcp: Remove redundant assignments in path manager init
2021-12-16 0:33 [PATCH mptcp-next v4 0/6] mptcp: Add userspace PM mode to bypass kernel PM Mat Martineau
@ 2021-12-16 0:33 ` Mat Martineau
2021-12-16 0:33 ` [PATCH mptcp-next v4 2/6] mptcp: Add a member to mptcp_pm_data to track kernel vs userspace mode Mat Martineau
` (6 subsequent siblings)
7 siblings, 0 replies; 9+ messages in thread
From: Mat Martineau @ 2021-12-16 0:33 UTC (permalink / raw)
To: mptcp; +Cc: Mat Martineau
A few members of the mptcp_pm_data struct were assigned to hard-coded
values in mptcp_pm_data_reset(), and then immediately changed in
mptcp_pm_nl_data_init().
Instead, flatten all the assignments in to mptcp_pm_data_reset().
v2: Resolve conflicts due to rename of mptcp_pm_data_reset()
v4: Resolve conflict in mptcp_pm_data_reset()
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
---
net/mptcp/pm.c | 32 ++++++++++++++++++--------------
net/mptcp/pm_netlink.c | 12 ------------
net/mptcp/protocol.h | 1 -
3 files changed, 18 insertions(+), 27 deletions(-)
diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
index ec20be3b43bd..49402556b4ec 100644
--- a/net/mptcp/pm.c
+++ b/net/mptcp/pm.c
@@ -383,21 +383,25 @@ void mptcp_pm_subflow_chk_stale(const struct mptcp_sock *msk, struct sock *ssk)
void mptcp_pm_data_reset(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;
- bitmap_fill(msk->pm.id_avail_bitmap, MPTCP_PM_MAX_ADDR_ID + 1);
+ bool subflows_allowed = !!mptcp_pm_get_subflows_max(msk);
+ struct mptcp_pm_data *pm = &msk->pm;
- mptcp_pm_nl_data_init(msk);
+ 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;
+ bitmap_fill(msk->pm.id_avail_bitmap, MPTCP_PM_MAX_ADDR_ID + 1);
}
void mptcp_pm_data_init(struct mptcp_sock *msk)
diff --git a/net/mptcp/pm_netlink.c b/net/mptcp/pm_netlink.c
index 75af1f701e1d..c8977b0c6cb0 100644
--- a/net/mptcp/pm_netlink.c
+++ b/net/mptcp/pm_netlink.c
@@ -1008,18 +1008,6 @@ 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 e1cf4f70ae9c..b22370878c86 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -822,7 +822,6 @@ 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);
--
2.34.1
^ permalink raw reply related [flat|nested] 9+ messages in thread
* [PATCH mptcp-next v4 2/6] mptcp: Add a member to mptcp_pm_data to track kernel vs userspace mode
2021-12-16 0:33 [PATCH mptcp-next v4 0/6] mptcp: Add userspace PM mode to bypass kernel PM Mat Martineau
2021-12-16 0:33 ` [PATCH mptcp-next v4 1/6] mptcp: Remove redundant assignments in path manager init Mat Martineau
@ 2021-12-16 0:33 ` Mat Martineau
2021-12-16 0:33 ` [PATCH mptcp-next v4 3/6] mptcp: Bypass kernel PM when userspace PM is enabled Mat Martineau
` (5 subsequent siblings)
7 siblings, 0 replies; 9+ messages in thread
From: Mat Martineau @ 2021-12-16 0:33 UTC (permalink / raw)
To: mptcp; +Cc: 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 49402556b4ec..e72e6010e323 100644
--- a/net/mptcp/pm.c
+++ b/net/mptcp/pm.c
@@ -392,6 +392,10 @@ void mptcp_pm_data_reset(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 b22370878c86..2fe59769b037 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -180,6 +180,14 @@ enum mptcp_pm_status {
*/
};
+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,
+};
+
/* Status bits below MPTCP_PM_ALREADY_ESTABLISHED need pm worker actions */
#define MPTCP_PM_WORK_MASK ((1 << MPTCP_PM_ALREADY_ESTABLISHED) - 1)
@@ -208,6 +216,7 @@ struct mptcp_pm_data {
u8 add_addr_signaled;
u8 add_addr_accepted;
u8 local_addr_used;
+ u8 pm_type;
u8 subflows;
u8 status;
DECLARE_BITMAP(id_avail_bitmap, MPTCP_PM_MAX_ADDR_ID + 1);
--
2.34.1
^ permalink raw reply related [flat|nested] 9+ messages in thread
* [PATCH mptcp-next v4 3/6] mptcp: Bypass kernel PM when userspace PM is enabled
2021-12-16 0:33 [PATCH mptcp-next v4 0/6] mptcp: Add userspace PM mode to bypass kernel PM Mat Martineau
2021-12-16 0:33 ` [PATCH mptcp-next v4 1/6] mptcp: Remove redundant assignments in path manager init Mat Martineau
2021-12-16 0:33 ` [PATCH mptcp-next v4 2/6] mptcp: Add a member to mptcp_pm_data to track kernel vs userspace mode Mat Martineau
@ 2021-12-16 0:33 ` Mat Martineau
2021-12-16 0:33 ` [PATCH mptcp-next v4 4/6] mptcp: Make kernel path manager check for userspace-managed sockets Mat Martineau
` (4 subsequent siblings)
7 siblings, 0 replies; 9+ messages in thread
From: Mat Martineau @ 2021-12-16 0:33 UTC (permalink / raw)
To: mptcp; +Cc: 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.
v2: Remove unneeded check in mptcp_pm_rm_addr_received() (Kishen Maloor)
v3: Add and use helper function for PM mode (Paolo Abeni)
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
---
net/mptcp/pm.c | 2 +-
net/mptcp/protocol.h | 5 +++++
2 files changed, 6 insertions(+), 1 deletion(-)
diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
index e72e6010e323..47ac9651e12c 100644
--- a/net/mptcp/pm.c
+++ b/net/mptcp/pm.c
@@ -208,7 +208,7 @@ 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) || mptcp_pm_is_userspace(msk)) {
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)) {
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index 2fe59769b037..0c680a5177eb 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -799,6 +799,11 @@ static inline bool mptcp_pm_should_rm_signal(struct mptcp_sock *msk)
return READ_ONCE(msk->pm.addr_signal) & BIT(MPTCP_RM_ADDR_SIGNAL);
}
+static inline bool mptcp_pm_is_userspace(struct mptcp_sock *msk)
+{
+ return READ_ONCE(msk->pm.pm_type) != MPTCP_PM_TYPE_KERNEL;
+}
+
static inline unsigned int mptcp_add_addr_len(int family, bool echo, bool port)
{
u8 len = TCPOLEN_MPTCP_ADD_ADDR_BASE;
--
2.34.1
^ permalink raw reply related [flat|nested] 9+ messages in thread
* [PATCH mptcp-next v4 4/6] mptcp: Make kernel path manager check for userspace-managed sockets
2021-12-16 0:33 [PATCH mptcp-next v4 0/6] mptcp: Add userspace PM mode to bypass kernel PM Mat Martineau
` (2 preceding siblings ...)
2021-12-16 0:33 ` [PATCH mptcp-next v4 3/6] mptcp: Bypass kernel PM when userspace PM is enabled Mat Martineau
@ 2021-12-16 0:33 ` Mat Martineau
2021-12-16 0:33 ` [PATCH mptcp-next v4 5/6] mptcp: Add a per-namespace sysctl to set the default path manager type Mat Martineau
` (3 subsequent siblings)
7 siblings, 0 replies; 9+ messages in thread
From: Mat Martineau @ 2021-12-16 0:33 UTC (permalink / raw)
To: mptcp; +Cc: Mat Martineau
Userspace-managed sockets should not have their subflows or
advertisements changed by the kernel path manager.
v3: Use helper function for PM mode (Paolo Abeni)
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
---
net/mptcp/pm_netlink.c | 18 ++++++++++++------
1 file changed, 12 insertions(+), 6 deletions(-)
diff --git a/net/mptcp/pm_netlink.c b/net/mptcp/pm_netlink.c
index c8977b0c6cb0..f12effa71942 100644
--- a/net/mptcp/pm_netlink.c
+++ b/net/mptcp/pm_netlink.c
@@ -1173,7 +1173,8 @@ 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) ||
+ mptcp_pm_is_userspace(msk))
goto next;
lock_sock(sk);
@@ -1299,6 +1300,9 @@ static int mptcp_nl_remove_subflow_and_signal_addr(struct net *net,
struct sock *sk = (struct sock *)msk;
bool remove_subflow;
+ if (mptcp_pm_is_userspace(msk))
+ goto next;
+
if (list_empty(&msk->conn_list)) {
mptcp_pm_remove_anno_addr(msk, addr, false);
goto next;
@@ -1340,7 +1344,7 @@ 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) || mptcp_pm_is_userspace(msk))
goto next;
local_address((struct sock_common *)msk, &msk_local);
@@ -1449,9 +1453,11 @@ 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 (!mptcp_pm_is_userspace(msk)) {
+ lock_sock(sk);
+ mptcp_pm_remove_addrs_and_subflows(msk, rm_list);
+ release_sock(sk);
+ }
sock_put(sk);
cond_resched();
@@ -1713,7 +1719,7 @@ 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) || mptcp_pm_is_userspace(msk))
goto next;
lock_sock(sk);
--
2.34.1
^ permalink raw reply related [flat|nested] 9+ messages in thread
* [PATCH mptcp-next v4 5/6] mptcp: Add a per-namespace sysctl to set the default path manager type
2021-12-16 0:33 [PATCH mptcp-next v4 0/6] mptcp: Add userspace PM mode to bypass kernel PM Mat Martineau
` (3 preceding siblings ...)
2021-12-16 0:33 ` [PATCH mptcp-next v4 4/6] mptcp: Make kernel path manager check for userspace-managed sockets Mat Martineau
@ 2021-12-16 0:33 ` Mat Martineau
2021-12-16 0:33 ` [PATCH mptcp-next v4 6/6] selftests: mptcp: Add tests for userspace PM type Mat Martineau
` (2 subsequent siblings)
7 siblings, 0 replies; 9+ messages in thread
From: Mat Martineau @ 2021-12-16 0:33 UTC (permalink / raw)
To: mptcp; +Cc: Mat Martineau
The new net.mptcp.pm_type sysctl determines which path manager will be
used by each newly-created MPTCP socket.
v2: Handle builds without CONFIG_SYSCTL
v3: Clarify logic for type-specific PM init (Geliang Tang and Paolo Abeni)
Signed-off-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
---
Documentation/networking/mptcp-sysctl.rst | 18 ++++++++++++
net/mptcp/ctrl.c | 21 ++++++++++++++
net/mptcp/pm.c | 34 +++++++++++++++--------
net/mptcp/protocol.h | 1 +
4 files changed, 63 insertions(+), 11 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..ae20b7d92e28 100644
--- a/net/mptcp/ctrl.c
+++ b/net/mptcp/ctrl.c
@@ -16,6 +16,11 @@
#define MPTCP_SYSCTL_PATH "net/mptcp"
static int mptcp_pernet_id;
+
+#ifdef CONFIG_SYSCTL
+static int mptcp_pm_type_max = __MPTCP_PM_TYPE_MAX;
+#endif
+
struct mptcp_pernet {
#ifdef CONFIG_SYSCTL
struct ctl_table_header *ctl_table_hdr;
@@ -26,6 +31,7 @@ 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 +64,11 @@ 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 +76,7 @@ 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 +120,14 @@ 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 +148,7 @@ 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 47ac9651e12c..1f8878cc29e3 100644
--- a/net/mptcp/pm.c
+++ b/net/mptcp/pm.c
@@ -383,7 +383,7 @@ void mptcp_pm_subflow_chk_stale(const struct mptcp_sock *msk, struct sock *ssk)
void mptcp_pm_data_reset(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;
pm->add_addr_signaled = 0;
@@ -392,17 +392,29 @@ void mptcp_pm_data_reset(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));
+ WRITE_ONCE(pm->pm_type, pm_type);
+
+ if (pm_type == MPTCP_PM_TYPE_KERNEL) {
+ bool subflows_allowed = !!mptcp_pm_get_subflows_max(msk);
+
+ /* 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));
+ WRITE_ONCE(pm->accept_addr,
+ !!mptcp_pm_get_add_addr_accept_max(msk) &&
+ subflows_allowed);
+ WRITE_ONCE(pm->accept_subflow, subflows_allowed);
+ } else {
+ WRITE_ONCE(pm->work_pending, 0);
+ WRITE_ONCE(pm->accept_addr, 0);
+ WRITE_ONCE(pm->accept_subflow, 0);
+ }
+
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;
bitmap_fill(msk->pm.id_avail_bitmap, MPTCP_PM_MAX_ADDR_ID + 1);
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index 0c680a5177eb..517f087f2b3a 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -579,6 +579,7 @@ 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);
--
2.34.1
^ permalink raw reply related [flat|nested] 9+ messages in thread
* [PATCH mptcp-next v4 6/6] selftests: mptcp: Add tests for userspace PM type
2021-12-16 0:33 [PATCH mptcp-next v4 0/6] mptcp: Add userspace PM mode to bypass kernel PM Mat Martineau
` (4 preceding siblings ...)
2021-12-16 0:33 ` [PATCH mptcp-next v4 5/6] mptcp: Add a per-namespace sysctl to set the default path manager type Mat Martineau
@ 2021-12-16 0:33 ` Mat Martineau
2021-12-16 9:50 ` [PATCH mptcp-next v4 0/6] mptcp: Add userspace PM mode to bypass kernel PM Paolo Abeni
2021-12-16 15:29 ` Matthieu Baerts
7 siblings, 0 replies; 9+ messages in thread
From: Mat Martineau @ 2021-12-16 0:33 UTC (permalink / raw)
To: mptcp; +Cc: Mat Martineau, Geliang Tang
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.
Tested-by: Geliang Tang <geliang.tang@suse.com>
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 8cbff7ce35d6..dfff017a08f6 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
@@ -1911,6 +1912,68 @@ 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
@@ -1928,6 +1991,7 @@ all_tests()
checksum_tests
deny_join_id0_tests
fullmesh_tests
+ userspace_tests
}
usage()
@@ -1948,6 +2012,7 @@ 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"
@@ -1983,7 +2048,7 @@ if [ $do_all_tests -eq 1 ]; then
exit $ret
fi
-while getopts 'fesltra64bpkdmchCS' opt; do
+while getopts 'fesltra64bpkdmuchCS' opt; do
case $opt in
f)
subflows_tests
@@ -2030,6 +2095,9 @@ while getopts 'fesltra64bpkdmchCS' opt; do
m)
fullmesh_tests
;;
+ u)
+ userspace_tests
+ ;;
c)
;;
C)
--
2.34.1
^ permalink raw reply related [flat|nested] 9+ messages in thread
* Re: [PATCH mptcp-next v4 0/6] mptcp: Add userspace PM mode to bypass kernel PM
2021-12-16 0:33 [PATCH mptcp-next v4 0/6] mptcp: Add userspace PM mode to bypass kernel PM Mat Martineau
` (5 preceding siblings ...)
2021-12-16 0:33 ` [PATCH mptcp-next v4 6/6] selftests: mptcp: Add tests for userspace PM type Mat Martineau
@ 2021-12-16 9:50 ` Paolo Abeni
2021-12-16 15:29 ` Matthieu Baerts
7 siblings, 0 replies; 9+ messages in thread
From: Paolo Abeni @ 2021-12-16 9:50 UTC (permalink / raw)
To: Mat Martineau, mptcp
On Wed, 2021-12-15 at 16:33 -0800, Mat Martineau wrote:
> 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.
>
> v1 -> v2: Rebased on latest export branch, removed extra kernel-mode
> check when receiving ADD_ADDR, and fix !CONFIG_SYSCTL build.
>
> v2 -> v3: Rebased on latest export branch, use helper function for PM
> mode, clarify initialization logic for userspace vs. kernel PM,
> add a Tested-by tag from v2 review.
>
> v3 -> v4: Rebased on latest export branch and resolved conflicts
>
> 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 | 21 ++++++
> net/mptcp/pm.c | 50 +++++++++----
> net/mptcp/pm_netlink.c | 30 ++++----
> net/mptcp/protocol.h | 16 ++++-
> .../testing/selftests/net/mptcp/mptcp_join.sh | 70 ++++++++++++++++++-
> 6 files changed, 170 insertions(+), 35 deletions(-)
>
>
> base-commit: 81706660508b6caa94d0fbeb8875882f155c73a9
LGTM, thanks Mat!
Acked-by: Paolo Abeni <pabeni@redhat.com>
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [PATCH mptcp-next v4 0/6] mptcp: Add userspace PM mode to bypass kernel PM
2021-12-16 0:33 [PATCH mptcp-next v4 0/6] mptcp: Add userspace PM mode to bypass kernel PM Mat Martineau
` (6 preceding siblings ...)
2021-12-16 9:50 ` [PATCH mptcp-next v4 0/6] mptcp: Add userspace PM mode to bypass kernel PM Paolo Abeni
@ 2021-12-16 15:29 ` Matthieu Baerts
7 siblings, 0 replies; 9+ messages in thread
From: Matthieu Baerts @ 2021-12-16 15:29 UTC (permalink / raw)
To: Mat Martineau, Paolo Abeni; +Cc: mptcp
Hi Mat, Paolo,
On 16/12/2021 01:33, Mat Martineau wrote:
> 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.
Thank you for the patches and the reviews!
Just applied in our tree (new feat. for net-next) with Paolo's ACK:
- b27a8a51fc93: mptcp: Remove redundant assignments in path manager init
- d9fae9ab22a1: mptcp: Add a member to mptcp_pm_data to track kernel vs
userspace mode
- 136ded889aa6: mptcp: Bypass kernel PM when userspace PM is enabled
- 08c305f2bca4: mptcp: Make kernel path manager check for
userspace-managed sockets
- 9864091e402f: mptcp: Add a per-namespace sysctl to set the default
path manager type
- 693a40231edf: selftests: mptcp: Add tests for userspace PM type
- Results: 1fdc9db07613..f83839879847
Builds and tests are now in progress:
https://cirrus-ci.com/github/multipath-tcp/mptcp_net-next/export/20211216T152900
https://github.com/multipath-tcp/mptcp_net-next/actions/workflows/build-validation.yml?query=branch:export
Cheers,
Matt
--
Tessares | Belgium | Hybrid Access Solutions
www.tessares.net
^ permalink raw reply [flat|nested] 9+ messages in thread
end of thread, other threads:[~2021-12-16 15:29 UTC | newest]
Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-12-16 0:33 [PATCH mptcp-next v4 0/6] mptcp: Add userspace PM mode to bypass kernel PM Mat Martineau
2021-12-16 0:33 ` [PATCH mptcp-next v4 1/6] mptcp: Remove redundant assignments in path manager init Mat Martineau
2021-12-16 0:33 ` [PATCH mptcp-next v4 2/6] mptcp: Add a member to mptcp_pm_data to track kernel vs userspace mode Mat Martineau
2021-12-16 0:33 ` [PATCH mptcp-next v4 3/6] mptcp: Bypass kernel PM when userspace PM is enabled Mat Martineau
2021-12-16 0:33 ` [PATCH mptcp-next v4 4/6] mptcp: Make kernel path manager check for userspace-managed sockets Mat Martineau
2021-12-16 0:33 ` [PATCH mptcp-next v4 5/6] mptcp: Add a per-namespace sysctl to set the default path manager type Mat Martineau
2021-12-16 0:33 ` [PATCH mptcp-next v4 6/6] selftests: mptcp: Add tests for userspace PM type Mat Martineau
2021-12-16 9:50 ` [PATCH mptcp-next v4 0/6] mptcp: Add userspace PM mode to bypass kernel PM Paolo Abeni
2021-12-16 15:29 ` Matthieu Baerts
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.