Series comparison

-[PATCH mptcp-next v4 0/4] BPF path manager, part 4
+[PATCH mptcp-next v6 00/13] BPF path manager, part 4
 From: Geliang Tang <tanggeliang@kylinos.cn>
+v6:
+ - add "name" in struct mptcp_pm_ops.
+ - add some "sysctl" patches.
+ - drop "struct mptcp_pm_param".
+ - drop "pm_type" in mptcp_pm_data.
+v5:
+ - use "struct mptcp_pm_param *param" as unified parameters for all
+   interfaces.
+ - register in-kernel mptcp_pm_ops too.
+ - only implement two interfaces "get_local_id" and "get_priority" in
+   this set.
 v4:
  - include a new patch "define BPF path manager type".
  - add new interfaces:
 ...
  - update get_local_id interface in patch 2.
 get_addr() and dump_addr() interfaces of BPF userspace pm are dropped
 as Matt suggested.
-In order to implement BPF userspace path manager, it is necessary to
+In order to implement BPF path manager, it's necessary to unify the
-unify the interfaces of the path manager. This set contains some
+interfaces of the path manager. This set contains some cleanups and
-cleanups and refactoring to unify the interfaces in kernel space.
+refactoring to unify the interfaces in kernel space. Finally, define
-Finally, define a struct mptcp_pm_ops for a userspace path manager
+a struct mptcp_pm_ops for a path manager.
 like this:
-struct mptcp_pm_ops {
+Geliang Tang (13):
-        int (*created)(struct mptcp_sock *msk);
+  mptcp: sysctl: use index for sysctl table
-        int (*established)(struct mptcp_sock *msk);
+  mptcp: sysctl: add path_manager to set pm name
-        int (*closed)(struct mptcp_sock *msk);
+  mptcp: pm: use addr entry for get_local_id
-        int (*address_announced)(struct mptcp_sock *msk,
+  mptcp: pm: define struct mptcp_pm_ops
-                                 struct mptcp_pm_addr_entry *local);
+  mptcp: pm: in-kernel: register mptcp_in_kernel_pm
-        int (*address_removed)(struct mptcp_sock *msk, u8 id);
+  mptcp: pm: userspace: register mptcp_userspace_pm
-        int (*subflow_established)(struct mptcp_sock *msk,
+  mptcp: pm: initialize and release mptcp_pm_ops
-                                   struct mptcp_pm_addr_entry *local,
+  mptcp: pm: drop pm_type in mptcp_pm_data
-                                   struct mptcp_addr_info *remote);
+  mptcp: sysctl: drop old pm_type sysctl
-        int (*subflow_closed)(struct mptcp_sock *msk,
+  mptcp: sysctl: add available_path_managers
-                              struct mptcp_pm_addr_entry *local,
+  mptcp: pm: drop get_local_id helpers
-                              struct mptcp_addr_info *remote);
+  mptcp: pm: drop is_backup helpers
-        int (*get_local_id)(struct mptcp_sock *msk,
+  selftests/mptcp: use path_manager instead of pm_type
                             struct mptcp_pm_addr_entry *skc);
         bool (*get_priority)(struct mptcp_sock *msk,
                              struct mptcp_addr_info *skc);
         int (*set_priority)(struct mptcp_sock *msk,
                             struct mptcp_pm_addr_entry *local,
                             struct mptcp_addr_info *remote);
         int (*listener_created)(struct mptcp_sock *msk);
         int (*listener_closed)(struct mptcp_sock *msk);
-        u8                      type;
+ Documentation/networking/mptcp-sysctl.rst     |  14 +-
-        struct module           *owner;
+ include/net/mptcp.h                           |  39 +++++
-        struct list_head        list;
+ net/mptcp/ctrl.c                              | 104 ++++++++++---
+ net/mptcp/pm.c                                | 141 ++++++++++++++++--
-        void (*init)(struct mptcp_sock *msk);
+ net/mptcp/pm_netlink.c                        |  24 ++-
-        void (*release)(struct mptcp_sock *msk);
+ net/mptcp/pm_userspace.c                      |  36 +++--
-} ____cacheline_aligned_in_smp;
+ net/mptcp/protocol.c                          |   1 +
+ net/mptcp/protocol.h                          |  23 ++-
-Geliang Tang (4):
+ .../testing/selftests/net/mptcp/mptcp_join.sh |  26 ++--
-  mptcp: define struct mptcp_pm_ops
+ .../selftests/net/mptcp/userspace_pm.sh       |   4 +-
-  mptcp: define BPF path manager type
+files changed, 324 insertions(+), 88 deletions(-)
   mptcp: register default userspace pm
   mptcp: initialize and release mptcp_pm_ops
  include/net/mptcp.h      |  32 +++++
  net/mptcp/pm.c           | 109 ++++++++++++++-
  net/mptcp/pm_netlink.c   |  11 +-
  net/mptcp/pm_userspace.c | 294 ++++++++++++++++++++++++---------------
  net/mptcp/protocol.c     |  10 +-
  net/mptcp/protocol.h     |  15 +-
 files changed, 355 insertions(+), 116 deletions(-)
 --
 .43.0

-New patch
+[PATCH mptcp-next v6 01/13] mptcp: sysctl: use index for sysctl table
+From: Geliang Tang <tanggeliang@kylinos.cn>
+The sysctl table needs to be updated when a new sysctl is added or an old
+sysctl is deleted. It's inconvenient to use fixed index numbers, since the
+index number that needs to be changed and all subsequent index numbers need
+to be updated accordingly.
+This patch adds an "i" variable to access each sysctl in the sysctl table
+incrementally.
+Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
+---
+ net/mptcp/ctrl.c | 23 ++++++++++++-----------
+file changed, 12 insertions(+), 11 deletions(-)
+diff --git a/net/mptcp/ctrl.c b/net/mptcp/ctrl.c
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/ctrl.c
++++ b/net/mptcp/ctrl.c
+@@ -XXX,XX +XXX,XX @@ static int mptcp_pernet_new_table(struct net *net, struct mptcp_pernet *pernet)
+ {
+     struct ctl_table_header *hdr;
+     struct ctl_table *table;
++    int i = 0;
+     table = mptcp_sysctl_table;
+     if (!net_eq(net, &init_net)) {
+@@ -XXX,XX +XXX,XX @@ static int mptcp_pernet_new_table(struct net *net, struct mptcp_pernet *pernet)
+             goto err_alloc;
+     }
+-    table[0].data = &pernet->mptcp_enabled;
+-    table[1].data = &pernet->add_addr_timeout;
+-    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;
+-    table[6].data = &pernet->scheduler;
+-    /* table[7] is for available_schedulers which is read-only info */
+-    table[8].data = &pernet->close_timeout;
+-    table[9].data = &pernet->blackhole_timeout;
+-    table[10].data = &pernet->syn_retrans_before_tcp_fallback;
++    table[i++].data = &pernet->mptcp_enabled;
++    table[i++].data = &pernet->add_addr_timeout;
++    table[i++].data = &pernet->checksum_enabled;
++    table[i++].data = &pernet->allow_join_initial_addr_port;
++    table[i++].data = &pernet->stale_loss_cnt;
++    table[i++].data = &pernet->pm_type;
++    table[i++].data = &pernet->scheduler;
++    i++; /* table[i] is for available_schedulers which is read-only info */
++    table[i++].data = &pernet->close_timeout;
++    table[i++].data = &pernet->blackhole_timeout;
++    table[i++].data = &pernet->syn_retrans_before_tcp_fallback;
+     hdr = register_net_sysctl_sz(net, MPTCP_SYSCTL_PATH, table,
+                      ARRAY_SIZE(mptcp_sysctl_table));
+--
+.43.0

-New patch
+[PATCH mptcp-next v6 02/13] mptcp: sysctl: add path_manager to set pm name
+From: Geliang Tang <tanggeliang@kylinos.cn>
+A new net.mptcp.path_manager sysctl is added to determine which path
+manager will be used by each newly-created MPTCP socket by setting the
+name of it. This sysctl will replace the old one pm_type.
+Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
+---
+ Documentation/networking/mptcp-sysctl.rst | 17 ++++++++
+ net/mptcp/ctrl.c                          | 50 +++++++++++++++++++++++
+ net/mptcp/protocol.h                      |  1 +
+files changed, 68 insertions(+)
+diff --git a/Documentation/networking/mptcp-sysctl.rst b/Documentation/networking/mptcp-sysctl.rst
+index XXXXXXX..XXXXXXX 100644
+--- a/Documentation/networking/mptcp-sysctl.rst
++++ b/Documentation/networking/mptcp-sysctl.rst
+@@ -XXX,XX +XXX,XX @@ pm_type - INTEGER
+     Default: 0
++path_manager - STRING
++    Set the default path manager name 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.
++
++    * "in-kernel" - In-kernel path manager
++    * "userspace" - Userspace path manager
++
++    Default: "in-kernel"
++
+ scheduler - STRING
+     Select the scheduler of your choice.
+diff --git a/net/mptcp/ctrl.c b/net/mptcp/ctrl.c
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/ctrl.c
++++ b/net/mptcp/ctrl.c
+@@ -XXX,XX +XXX,XX @@ struct mptcp_pernet {
+     u8 checksum_enabled;
+     u8 allow_join_initial_addr_port;
+     u8 pm_type;
++    char path_manager[MPTCP_PM_NAME_MAX];
+     char scheduler[MPTCP_SCHED_NAME_MAX];
+ };
+@@ -XXX,XX +XXX,XX @@ int mptcp_get_pm_type(const struct net *net)
+     return mptcp_get_pernet(net)->pm_type;
+ }
++const char *mptcp_get_path_manager(const struct net *net)
++{
++    return mptcp_get_pernet(net)->path_manager;
++}
++
+ const char *mptcp_get_scheduler(const struct net *net)
+ {
+     return mptcp_get_pernet(net)->scheduler;
+@@ -XXX,XX +XXX,XX @@ static void mptcp_pernet_set_defaults(struct mptcp_pernet *pernet)
+     pernet->allow_join_initial_addr_port = 1;
+     pernet->stale_loss_cnt = 4;
+     pernet->pm_type = MPTCP_PM_TYPE_KERNEL;
++    strscpy(pernet->path_manager, "in-kernel", sizeof(pernet->path_manager));
+     strscpy(pernet->scheduler, "default", sizeof(pernet->scheduler));
+ }
+ #ifdef CONFIG_SYSCTL
++static int mptcp_set_path_manager(char *path_manager, const char *name)
++{
++    struct mptcp_pm_ops *pm;
++    int ret = 0;
++
++    rcu_read_lock();
++    pm = mptcp_pm_find(name);
++    if (pm)
++        strscpy(path_manager, name, MPTCP_PM_NAME_MAX);
++    else
++        ret = -ENOENT;
++    rcu_read_unlock();
++
++    return ret;
++}
++
++static int proc_path_manager(const struct ctl_table *ctl, int write,
++                 void *buffer, size_t *lenp, loff_t *ppos)
++{
++    char (*path_manager)[MPTCP_PM_NAME_MAX] = ctl->data;
++    char val[MPTCP_PM_NAME_MAX];
++    const struct ctl_table tbl = {
++        .data = val,
++        .maxlen = MPTCP_PM_NAME_MAX,
++    };
++    int ret;
++
++    strscpy(val, *path_manager, MPTCP_PM_NAME_MAX);
++
++    ret = proc_dostring(&tbl, write, buffer, lenp, ppos);
++    if (write && ret == 0)
++        ret = mptcp_set_path_manager(*path_manager, val);
++
++    return ret;
++}
++
+ static int mptcp_set_scheduler(char *scheduler, const char *name)
+ {
+     struct mptcp_sched_ops *sched;
+@@ -XXX,XX +XXX,XX @@ static struct ctl_table mptcp_sysctl_table[] = {
+         .extra1       = SYSCTL_ZERO,
+         .extra2       = &mptcp_pm_type_max
+     },
++    {
++        .procname = "path_manager",
++        .maxlen    = MPTCP_PM_NAME_MAX,
++        .mode = 0644,
++        .proc_handler = proc_path_manager,
++    },
+     {
+         .procname = "scheduler",
+         .maxlen    = MPTCP_SCHED_NAME_MAX,
+@@ -XXX,XX +XXX,XX @@ static int mptcp_pernet_new_table(struct net *net, struct mptcp_pernet *pernet)
+     table[i++].data = &pernet->allow_join_initial_addr_port;
+     table[i++].data = &pernet->stale_loss_cnt;
+     table[i++].data = &pernet->pm_type;
++    table[i++].data = &pernet->path_manager;
+     table[i++].data = &pernet->scheduler;
+     i++; /* table[i] is for available_schedulers which is read-only info */
+     table[i++].data = &pernet->close_timeout;
+diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/protocol.h
++++ b/net/mptcp/protocol.h
+@@ -XXX,XX +XXX,XX @@ int mptcp_allow_join_id0(const struct net *net);
+ unsigned int mptcp_stale_loss_cnt(const struct net *net);
+ unsigned int mptcp_close_timeout(const struct sock *sk);
+ int mptcp_get_pm_type(const struct net *net);
++const char *mptcp_get_path_manager(const struct net *net);
+ const char *mptcp_get_scheduler(const struct net *net);
+ void mptcp_active_disable(struct sock *sk);
+--
+.43.0

-New patch
+[PATCH mptcp-next v6 03/13] mptcp: pm: use addr entry for get_local_id
+From: Geliang Tang <tanggeliang@kylinos.cn>
+The following code in mptcp_userspace_pm_get_local_id() that assigns "skc"
+to "new_entry" is not allowed in BPF if we use the same code to implement
+the get_local_id() interface of a BFP path manager:
+    memset(&new_entry, 0, sizeof(struct mptcp_pm_addr_entry));
+    new_entry.addr = *skc;
+    new_entry.addr.id = 0;
+    new_entry.flags = MPTCP_PM_ADDR_FLAG_IMPLICIT;
+To solve the issue, this patch moves this assignment to "new_entry" forward
+to mptcp_pm_get_local_id(), and then passing "new_entry" as a parameter to
+both mptcp_pm_nl_get_local_id() and mptcp_userspace_pm_get_local_id().
+Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
+---
+ net/mptcp/pm.c           |  9 ++++++---
+ net/mptcp/pm_netlink.c   | 11 ++++-------
+ net/mptcp/pm_userspace.c | 17 ++++++-----------
+ net/mptcp/protocol.h     |  6 ++++--
+files changed, 20 insertions(+), 23 deletions(-)
+diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/pm.c
++++ b/net/mptcp/pm.c
+@@ -XXX,XX +XXX,XX @@ 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)
+ {
+-    struct mptcp_addr_info skc_local;
++    struct mptcp_pm_addr_entry skc_local = { 0 };
+     struct mptcp_addr_info msk_local;
+     if (WARN_ON_ONCE(!msk))
+@@ -XXX,XX +XXX,XX @@ int mptcp_pm_get_local_id(struct mptcp_sock *msk, struct sock_common *skc)
+      * addr
+      */
+     mptcp_local_address((struct sock_common *)msk, &msk_local);
+-    mptcp_local_address((struct sock_common *)skc, &skc_local);
+-    if (mptcp_addresses_equal(&msk_local, &skc_local, false))
++    mptcp_local_address((struct sock_common *)skc, &skc_local.addr);
++    if (mptcp_addresses_equal(&msk_local, &skc_local.addr, false))
+         return 0;
++    skc_local.addr.id = 0;
++    skc_local.flags = MPTCP_PM_ADDR_FLAG_IMPLICIT;
++
+     if (mptcp_pm_is_userspace(msk))
+         return mptcp_userspace_pm_get_local_id(msk, &skc_local);
+     return mptcp_pm_nl_get_local_id(msk, &skc_local);
+diff --git a/net/mptcp/pm_netlink.c b/net/mptcp/pm_netlink.c
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/pm_netlink.c
++++ b/net/mptcp/pm_netlink.c
+@@ -XXX,XX +XXX,XX @@ static int mptcp_pm_nl_create_listen_socket(struct sock *sk,
+     return err;
+ }
+-int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk, struct mptcp_addr_info *skc)
++int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk,
++                 struct mptcp_pm_addr_entry *skc)
+ {
+     struct mptcp_pm_addr_entry *entry;
+     struct pm_nl_pernet *pernet;
+@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk, struct mptcp_addr_info *skc
+     pernet = pm_nl_get_pernet_from_msk(msk);
+     rcu_read_lock();
+-    entry = __lookup_addr(pernet, skc);
++    entry = __lookup_addr(pernet, &skc->addr);
+     ret = entry ? entry->addr.id : -1;
+     rcu_read_unlock();
+     if (ret >= 0)
+@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk, struct mptcp_addr_info *skc
+     if (!entry)
+         return -ENOMEM;
+-    entry->addr = *skc;
+-    entry->addr.id = 0;
++    *entry = *skc;
+     entry->addr.port = 0;
+-    entry->ifindex = 0;
+-    entry->flags = MPTCP_PM_ADDR_FLAG_IMPLICIT;
+-    entry->lsk = NULL;
+     ret = mptcp_pm_nl_append_new_local_addr(pernet, entry, true);
+     if (ret < 0)
+         kfree(entry);
+diff --git a/net/mptcp/pm_userspace.c b/net/mptcp/pm_userspace.c
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/pm_userspace.c
++++ b/net/mptcp/pm_userspace.c
+@@ -XXX,XX +XXX,XX @@ mptcp_userspace_pm_lookup_addr_by_id(struct mptcp_sock *msk, unsigned int id)
+ }
+ int mptcp_userspace_pm_get_local_id(struct mptcp_sock *msk,
+-                    struct mptcp_addr_info *skc)
++                    struct mptcp_pm_addr_entry *skc)
+ {
+-    struct mptcp_pm_addr_entry *entry = NULL, new_entry;
+     __be16 msk_sport =  ((struct inet_sock *)
+                  inet_sk((struct sock *)msk))->inet_sport;
++    struct mptcp_pm_addr_entry *entry;
+     spin_lock_bh(&msk->pm.lock);
+-    entry = mptcp_userspace_pm_lookup_addr(msk, skc);
++    entry = mptcp_userspace_pm_lookup_addr(msk, &skc->addr);
+     spin_unlock_bh(&msk->pm.lock);
+     if (entry)
+         return entry->addr.id;
+-    memset(&new_entry, 0, sizeof(struct mptcp_pm_addr_entry));
+-    new_entry.addr = *skc;
+-    new_entry.addr.id = 0;
+-    new_entry.flags = MPTCP_PM_ADDR_FLAG_IMPLICIT;
+-
+-    if (new_entry.addr.port == msk_sport)
+-        new_entry.addr.port = 0;
++    if (skc->addr.port == msk_sport)
++        skc->addr.port = 0;
+-    return mptcp_userspace_pm_append_new_local_addr(msk, &new_entry, true);
++    return mptcp_userspace_pm_append_new_local_addr(msk, skc, true);
+ }
+ bool mptcp_userspace_pm_is_backup(struct mptcp_sock *msk,
+diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/protocol.h
++++ b/net/mptcp/protocol.h
+@@ -XXX,XX +XXX,XX @@ bool mptcp_pm_add_addr_signal(struct mptcp_sock *msk, const struct sk_buff *skb,
+ bool mptcp_pm_rm_addr_signal(struct mptcp_sock *msk, unsigned int remaining,
+                  struct mptcp_rm_list *rm_list);
+ int mptcp_pm_get_local_id(struct mptcp_sock *msk, struct sock_common *skc);
+-int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk, struct mptcp_addr_info *skc);
+-int mptcp_userspace_pm_get_local_id(struct mptcp_sock *msk, struct mptcp_addr_info *skc);
++int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk,
++                 struct mptcp_pm_addr_entry *skc);
++int mptcp_userspace_pm_get_local_id(struct mptcp_sock *msk,
++                    struct mptcp_pm_addr_entry *skc);
+ bool mptcp_pm_is_backup(struct mptcp_sock *msk, struct sock_common *skc);
+ bool mptcp_pm_nl_is_backup(struct mptcp_sock *msk, struct mptcp_addr_info *skc);
+ bool mptcp_userspace_pm_is_backup(struct mptcp_sock *msk, struct mptcp_addr_info *skc);
+--
+.43.0

-[PATCH mptcp-next v4 1/4] mptcp: define struct mptcp_pm_ops
+[PATCH mptcp-next v6 04/13] mptcp: pm: define struct mptcp_pm_ops
 ...
 Add a set of functions to register, unregister, find and validate a
 given struct ops.
 Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
 ---
- include/net/mptcp.h  | 32 ++++++++++++++++++++++++
+ include/net/mptcp.h  | 37 +++++++++++++++++++++++++++++
- net/mptcp/pm.c       | 59 ++++++++++++++++++++++++++++++++++++++++++++
+ net/mptcp/pm.c       | 55 ++++++++++++++++++++++++++++++++++++++++++++
  net/mptcp/protocol.h |  5 ++++
-files changed, 96 insertions(+)
+files changed, 97 insertions(+)
 diff --git a/include/net/mptcp.h b/include/net/mptcp.h
 index XXXXXXX..XXXXXXX 100644
 --- a/include/net/mptcp.h
 +++ b/include/net/mptcp.h
 ...
  /* MPTCP sk_buff extension data */
 @@ -XXX,XX +XXX,XX @@ struct mptcp_sched_ops {
      void (*release)(struct mptcp_sock *msk);
  } ____cacheline_aligned_in_smp;
++#define MPTCP_PM_NAME_MAX    16
++
 +struct mptcp_pm_ops {
 +    int (*created)(struct mptcp_sock *msk);
 +    int (*established)(struct mptcp_sock *msk);
 +    int (*closed)(struct mptcp_sock *msk);
++    int (*get_local_id)(struct mptcp_sock *msk,
++                struct mptcp_pm_addr_entry *skc);
++    bool (*get_priority)(struct mptcp_sock *msk,
++                 struct mptcp_addr_info *skc);
 +    int (*address_announced)(struct mptcp_sock *msk,
 +                 struct mptcp_pm_addr_entry *local);
-+    int (*address_removed)(struct mptcp_sock *msk, u8 id);
++    int (*address_removed)(struct mptcp_sock *msk,
 +                   struct mptcp_pm_addr_entry *local);
 +    int (*subflow_established)(struct mptcp_sock *msk,
 +                   struct mptcp_pm_addr_entry *local,
 +                   struct mptcp_addr_info *remote);
 +    int (*subflow_closed)(struct mptcp_sock *msk,
 +                  struct mptcp_pm_addr_entry *local,
 +                  struct mptcp_addr_info *remote);
-+    int (*get_local_id)(struct mptcp_sock *msk,
-+                struct mptcp_pm_addr_entry *skc);
-+    bool (*get_priority)(struct mptcp_sock *msk,
-+                 struct mptcp_addr_info *skc);
 +    int (*set_priority)(struct mptcp_sock *msk,
 +                struct mptcp_pm_addr_entry *local,
-+                struct mptcp_addr_info *remote);
++                struct mptcp_addr_info *remote,
 +                u8 changed);
 +    int (*listener_created)(struct mptcp_sock *msk);
 +    int (*listener_closed)(struct mptcp_sock *msk);
 +
 +    u8            type;
++    char            name[MPTCP_PM_NAME_MAX];
 +    struct module        *owner;
 +    struct list_head    list;
 +
 +    void (*init)(struct mptcp_sock *msk);
 +    void (*release)(struct mptcp_sock *msk);
 ...
  {
      mptcp_pm_nl_init();
  }
 +
 +/* Must be called with rcu read lock held */
-+struct mptcp_pm_ops *mptcp_pm_find(enum mptcp_pm_type type)
++struct mptcp_pm_ops *mptcp_pm_find(const char *name)
 +{
 +    struct mptcp_pm_ops *pm;
 +
 +    list_for_each_entry_rcu(pm, &mptcp_pm_list, list) {
-+        if (pm->type == type)
++        if (!strcmp(pm->name, name))
 +            return pm;
 +    }
 +
 +    return NULL;
 +}
 +
 +int mptcp_pm_validate(struct mptcp_pm_ops *pm)
 +{
-+    if (!pm->created && !pm->established && !pm->closed &&
++    if (!pm->get_local_id || !pm->get_priority) {
 +        !pm->address_announced && !pm->address_removed &&
 +        !pm->subflow_established && !pm->subflow_closed &&
 +        !pm->get_local_id && !pm->get_priority && !pm->set_priority &&
 +        !pm->listener_created && !pm->listener_closed) {
 +        pr_err("%u does not implement required ops\n", pm->type);
 +        return -EINVAL;
 +    }
 +
 +    return 0;
 ...
 +    ret = mptcp_pm_validate(pm);
 +    if (ret)
 +        return ret;
 +
 +    spin_lock(&mptcp_pm_list_lock);
-+    if (mptcp_pm_find(pm->type)) {
++    if (mptcp_pm_find(pm->name)) {
 +        spin_unlock(&mptcp_pm_list_lock);
 +        return -EEXIST;
 +    }
 +    list_add_tail_rcu(&pm->list, &mptcp_pm_list);
 +    spin_unlock(&mptcp_pm_list_lock);
 +
-+    pr_debug("userspace_pm type %u registered\n", pm->type);
++    pr_debug("%s registered\n", pm->name);
 +    return 0;
 +}
 +
 +void mptcp_pm_unregister(struct mptcp_pm_ops *pm)
 +{
 ...
 +++ b/net/mptcp/protocol.h
 @@ -XXX,XX +XXX,XX @@ int mptcp_pm_remove_addr(struct mptcp_sock *msk, const struct mptcp_rm_list *rm_
  void mptcp_pm_remove_addr_entry(struct mptcp_sock *msk,
                  struct mptcp_pm_addr_entry *entry);
-+struct mptcp_pm_ops *mptcp_pm_find(enum mptcp_pm_type type);
++struct mptcp_pm_ops *mptcp_pm_find(const char *name);
 +int mptcp_pm_validate(struct mptcp_pm_ops *pm);
 +int mptcp_pm_register(struct mptcp_pm_ops *pm);
 +void mptcp_pm_unregister(struct mptcp_pm_ops *pm);
 +
  void mptcp_free_local_addr_list(struct mptcp_sock *msk);
  void mptcp_event(enum mptcp_event_type type, const struct mptcp_sock *msk,
 --
 .43.0

-New patch
+[PATCH mptcp-next v6 05/13] mptcp: pm: in-kernel: register mptcp_in_kernel_pm
+From: Geliang Tang <tanggeliang@kylinos.cn>
+This patch defines the original in-kernel netlink path manager as a new
+struct mptcp_pm_ops named "mptcp_in_kernel_pm", and register it in
+mptcp_pm_nl_init().
+This mptcp_pm_ops will be skipped in mptcp_pm_unregister().
+Only get_local_id() and get_priority() interfaces are implemented here.
+mptcp_pm_nl_is_backup() becomes a wrapper of get_priority().
+Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
+---
+ net/mptcp/pm.c         |  3 +++
+ net/mptcp/pm_netlink.c | 18 +++++++++++++++++-
+ net/mptcp/protocol.h   |  2 ++
+files changed, 22 insertions(+), 1 deletion(-)
+diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/pm.c
++++ b/net/mptcp/pm.c
+@@ -XXX,XX +XXX,XX @@ int mptcp_pm_register(struct mptcp_pm_ops *pm)
+ void mptcp_pm_unregister(struct mptcp_pm_ops *pm)
+ {
++    if (pm == &mptcp_in_kernel_pm)
++        return;
++
+     spin_lock(&mptcp_pm_list_lock);
+     list_del_rcu(&pm->list);
+     spin_unlock(&mptcp_pm_list_lock);
+diff --git a/net/mptcp/pm_netlink.c b/net/mptcp/pm_netlink.c
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/pm_netlink.c
++++ b/net/mptcp/pm_netlink.c
+@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk,
+     return ret;
+ }
+-bool mptcp_pm_nl_is_backup(struct mptcp_sock *msk, struct mptcp_addr_info *skc)
++static bool mptcp_pm_nl_get_priority(struct mptcp_sock *msk,
++                     struct mptcp_addr_info *skc)
+ {
+     struct pm_nl_pernet *pernet = pm_nl_get_pernet_from_msk(msk);
+     struct mptcp_pm_addr_entry *entry;
+@@ -XXX,XX +XXX,XX @@ bool mptcp_pm_nl_is_backup(struct mptcp_sock *msk, struct mptcp_addr_info *skc)
+     return backup;
+ }
++bool mptcp_pm_nl_is_backup(struct mptcp_sock *msk, struct mptcp_addr_info *skc)
++{
++    return mptcp_pm_nl_get_priority(msk, skc);
++}
++
+ #define MPTCP_PM_CMD_GRP_OFFSET       0
+ #define MPTCP_PM_EV_GRP_OFFSET        1
+@@ -XXX,XX +XXX,XX @@ static struct pernet_operations mptcp_pm_pernet_ops = {
+     .size = sizeof(struct pm_nl_pernet),
+ };
++struct mptcp_pm_ops mptcp_in_kernel_pm = {
++    .get_local_id        = mptcp_pm_nl_get_local_id,
++    .get_priority        = mptcp_pm_nl_get_priority,
++    .type            = MPTCP_PM_TYPE_KERNEL,
++    .name            = "in-kernel",
++    .owner            = THIS_MODULE,
++};
++
+ void __init mptcp_pm_nl_init(void)
+ {
+     if (register_pernet_subsys(&mptcp_pm_pernet_ops) < 0)
+@@ -XXX,XX +XXX,XX @@ void __init mptcp_pm_nl_init(void)
+     if (genl_register_family(&mptcp_genl_family))
+         panic("Failed to register MPTCP PM netlink family\n");
++
++    mptcp_pm_register(&mptcp_in_kernel_pm);
+ }
+diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/protocol.h
++++ b/net/mptcp/protocol.h
+@@ -XXX,XX +XXX,XX @@ int mptcp_pm_remove_addr(struct mptcp_sock *msk, const struct mptcp_rm_list *rm_
+ void mptcp_pm_remove_addr_entry(struct mptcp_sock *msk,
+                 struct mptcp_pm_addr_entry *entry);
++extern struct mptcp_pm_ops mptcp_in_kernel_pm;
++
+ struct mptcp_pm_ops *mptcp_pm_find(const char *name);
+ int mptcp_pm_validate(struct mptcp_pm_ops *pm);
+ int mptcp_pm_register(struct mptcp_pm_ops *pm);
+--
+.43.0

-[PATCH mptcp-next v4 3/4] mptcp: register default userspace pm
+[PATCH mptcp-next v6 06/13] mptcp: pm: userspace: register mptcp_userspace_pm
 From: Geliang Tang <tanggeliang@kylinos.cn>
-This patch defines the original userspace pm as the default path manager,
+This patch defines the original userspace path manager as a new struct
-named mptcp_userspace_pm, and register it in mptcp_pm_data_init().
+mptcp_pm_ops named "mptcp_userspace_pm", and register it in
 mptcp_pm_data_init().
-Extract address_announce() interface of the userspace PM from the handler
+Only get_local_id() and get_priority() interfaces are implemented here.
-of netlink command MPTCP_PM_CMD_ANNOUNCE mptcp_pm_nl_announce_doit(), only
+mptcp_userspace_pm_is_backup() becomes a wrapper of get_priority().
 leave the code for obtaining msk through "info" and parsing address entry
 in the handler.
 Extract address_remove() interface of the userspace PM from the handler
 of netlink command MPTCP_PM_CMD_REMOVE mptcp_pm_nl_remove_doit(), only
 leave the code for parsing address id and obtaining msk through "info"
 in the handler.
 Extract subflow_create() interface of the userspace PM from the handler of
 netlink command MPTCP_PM_CMD_SUBFLOW_CREATE
     mptcp_pm_nl_subflow_create_doit(),
 only leave the code for obtaining msk through "info", parsing local address
 entry and parsing remote address info in the handler.
 Extract subflow_destroy() interface of the userspace PM from the handler of
 netlink command MPTCP_PM_CMD_SUBFLOW_DESTROY
     mptcp_pm_nl_subflow_destroy_doit(),
 only leave the code for obtaining msk through "info", parsing local address
 entry and parsing remote address info in the handler.
 Extract set_flags() interface of the userspace PM from function
 mptcp_userspace_pm_set_flags(), only leave the code for obtaining
 msk through "info" in this function.
 All fives interfaces are invoked under holding the msk socket lock.
 Extract get_local_id() interface of the userspace PM from function
 mptcp_userspace_pm_get_local_id(), only leave the code for setting
 new address entry in this function.
 get_flags() interface of the userspace PM is defined as the same as
 the function mptcp_userspace_pm_get_flags(), then this function
 becomes a wrapper.
 Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
 ---
- net/mptcp/pm.c           |   1 +
+ net/mptcp/pm.c           |  1 +
- net/mptcp/pm_userspace.c | 280 +++++++++++++++++++++++----------------
+ net/mptcp/pm_userspace.c | 23 +++++++++++++++++++++--
- net/mptcp/protocol.h     |   1 +
+ net/mptcp/protocol.h     |  1 +
-files changed, 171 insertions(+), 111 deletions(-)
+files changed, 23 insertions(+), 2 deletions(-)
 diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
 index XXXXXXX..XXXXXXX 100644
 --- a/net/mptcp/pm.c
 +++ b/net/mptcp/pm.c
 ...
  /* Must be called with rcu read lock held */
 diff --git a/net/mptcp/pm_userspace.c b/net/mptcp/pm_userspace.c
 index XXXXXXX..XXXXXXX 100644
 --- a/net/mptcp/pm_userspace.c
 +++ b/net/mptcp/pm_userspace.c
-@@ -XXX,XX +XXX,XX @@ mptcp_userspace_pm_lookup_addr_by_id(struct mptcp_sock *msk, unsigned int id)
+@@ -XXX,XX +XXX,XX @@ int mptcp_userspace_pm_get_local_id(struct mptcp_sock *msk,
-     return NULL;
+     return mptcp_userspace_pm_append_new_local_addr(msk, skc, true);
  }
--int mptcp_userspace_pm_get_local_id(struct mptcp_sock *msk,
--                    struct mptcp_addr_info *skc)
-+static int userspace_pm_get_local_id(struct mptcp_sock *msk,
-+                     struct mptcp_pm_addr_entry *skc)
- {
--    struct mptcp_pm_addr_entry *entry = NULL, new_entry;
--    __be16 msk_sport =  ((struct inet_sock *)
--                 inet_sk((struct sock *)msk))->inet_sport;
-+    struct mptcp_pm_addr_entry *entry;
-     spin_lock_bh(&msk->pm.lock);
--    entry = mptcp_userspace_pm_lookup_addr(msk, skc);
-+    entry = mptcp_userspace_pm_lookup_addr(msk, &skc->addr);
-     spin_unlock_bh(&msk->pm.lock);
-     if (entry)
-         return entry->addr.id;
-+    return mptcp_userspace_pm_append_new_local_addr(msk, skc, true);
-+}
-+
-+int mptcp_userspace_pm_get_local_id(struct mptcp_sock *msk,
-+                    struct mptcp_addr_info *skc)
-+{
-+    __be16 msk_sport = ((struct inet_sock *)
-+                inet_sk((struct sock *)msk))->inet_sport;
-+    struct mptcp_pm_addr_entry new_entry;
-+
-     memset(&new_entry, 0, sizeof(struct mptcp_pm_addr_entry));
-     new_entry.addr = *skc;
-     new_entry.addr.id = 0;
-@@ -XXX,XX +XXX,XX @@ int mptcp_userspace_pm_get_local_id(struct mptcp_sock *msk,
-     if (new_entry.addr.port == msk_sport)
-         new_entry.addr.port = 0;
--    return mptcp_userspace_pm_append_new_local_addr(msk, &new_entry, true);
-+    return userspace_pm_get_local_id(msk, &new_entry);
- }
 -bool mptcp_userspace_pm_is_backup(struct mptcp_sock *msk,
 -                  struct mptcp_addr_info *skc)
-+static bool userspace_pm_get_priority(struct mptcp_sock *msk,
++static bool mptcp_userspace_pm_get_priority(struct mptcp_sock *msk,
-+                      struct mptcp_addr_info *skc)
++                        struct mptcp_addr_info *skc)
  {
      struct mptcp_pm_addr_entry *entry;
      bool backup;
 @@ -XXX,XX +XXX,XX @@ bool mptcp_userspace_pm_is_backup(struct mptcp_sock *msk,
      return backup;
  }
 +bool mptcp_userspace_pm_is_backup(struct mptcp_sock *msk,
 +                  struct mptcp_addr_info *skc)
 +{
-+    return userspace_pm_get_priority(msk, skc);
++    return mptcp_userspace_pm_get_priority(msk, skc);
 +}
 +
  static struct mptcp_sock *mptcp_userspace_pm_get_sock(const struct genl_info *info)
  {
      struct mptcp_sock *msk;
-@@ -XXX,XX +XXX,XX @@ static struct mptcp_sock *mptcp_userspace_pm_get_sock(const struct genl_info *in
-     return msk;
- }
-+static int userspace_pm_address_announced(struct mptcp_sock *msk,
-+                      struct mptcp_pm_addr_entry *local)
-+{
-+    int err;
-+
-+    err = mptcp_userspace_pm_append_new_local_addr(msk, local, false);
-+    if (err < 0)
-+        return err;
-+
-+    spin_lock_bh(&msk->pm.lock);
-+
-+    if (mptcp_pm_alloc_anno_list(msk, &local->addr)) {
-+        msk->pm.add_addr_signaled++;
-+        mptcp_pm_announce_addr(msk, &local->addr, false);
-+        mptcp_pm_nl_addr_send_ack(msk);
-+    }
-+
-+    spin_unlock_bh(&msk->pm.lock);
-+
-+    return 0;
-+}
-+
- int mptcp_pm_nl_announce_doit(struct sk_buff *skb, struct genl_info *info)
- {
-     struct mptcp_pm_addr_entry addr_val;
-@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_announce_doit(struct sk_buff *skb, struct genl_info *info)
-         goto announce_err;
-     }
--    err = mptcp_userspace_pm_append_new_local_addr(msk, &addr_val, false);
--    if (err < 0) {
--        NL_SET_ERR_MSG_ATTR(info->extack, addr,
--                    "did not match address and id");
--        goto announce_err;
--    }
--
-     lock_sock(sk);
--    spin_lock_bh(&msk->pm.lock);
--
--    if (mptcp_pm_alloc_anno_list(msk, &addr_val.addr)) {
--        msk->pm.add_addr_signaled++;
--        mptcp_pm_announce_addr(msk, &addr_val.addr, false);
--        mptcp_pm_nl_addr_send_ack(msk);
--    }
--
--    spin_unlock_bh(&msk->pm.lock);
-+    err = userspace_pm_address_announced(msk, &addr_val);
-     release_sock(sk);
-+    if (err)
-+        NL_SET_ERR_MSG_ATTR(info->extack, addr,
-+                    "did not match address and id");
--    err = 0;
-  announce_err:
-     sock_put(sk);
-     return err;
-@@ -XXX,XX +XXX,XX @@ static int mptcp_userspace_pm_remove_id_zero_address(struct mptcp_sock *msk)
- {
-     struct mptcp_rm_list list = { .nr = 0 };
-     struct mptcp_subflow_context *subflow;
--    struct sock *sk = (struct sock *)msk;
-     bool has_id_0 = false;
--    int err = -EINVAL;
--    lock_sock(sk);
-     mptcp_for_each_subflow(msk, subflow) {
-         if (READ_ONCE(subflow->local_id) == 0) {
-             has_id_0 = true;
-@@ -XXX,XX +XXX,XX @@ static int mptcp_userspace_pm_remove_id_zero_address(struct mptcp_sock *msk)
-         }
-     }
-     if (!has_id_0)
--        goto remove_err;
-+        return -EINVAL;
-     list.ids[list.nr++] = 0;
-@@ -XXX,XX +XXX,XX @@ static int mptcp_userspace_pm_remove_id_zero_address(struct mptcp_sock *msk)
-     mptcp_pm_remove_addr(msk, &list);
-     spin_unlock_bh(&msk->pm.lock);
--    err = 0;
--
--remove_err:
--    release_sock(sk);
--    return err;
-+    return 0;
- }
- void mptcp_pm_remove_addr_entry(struct mptcp_sock *msk,
-@@ -XXX,XX +XXX,XX @@ void mptcp_pm_remove_addr_entry(struct mptcp_sock *msk,
-     spin_unlock_bh(&msk->pm.lock);
- }
-+static int userspace_pm_address_removed(struct mptcp_sock *msk, u8 id)
-+{
-+    struct mptcp_pm_addr_entry *entry;
-+
-+    if (id == 0)
-+        return mptcp_userspace_pm_remove_id_zero_address(msk);
-+
-+    spin_lock_bh(&msk->pm.lock);
-+    entry = mptcp_userspace_pm_lookup_addr_by_id(msk, id);
-+    if (!entry) {
-+        spin_unlock_bh(&msk->pm.lock);
-+        return -EINVAL;
-+    }
-+
-+    list_del_rcu(&entry->list);
-+    spin_unlock_bh(&msk->pm.lock);
-+
-+    mptcp_pm_remove_addr_entry(msk, entry);
-+
-+    sock_kfree_s((struct sock *)msk, entry, sizeof(*entry));
-+
-+    return 0;
-+}
-+
- int mptcp_pm_nl_remove_doit(struct sk_buff *skb, struct genl_info *info)
- {
--    struct mptcp_pm_addr_entry *match;
-     struct mptcp_sock *msk;
-     struct nlattr *id;
-     int err = -EINVAL;
-@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_remove_doit(struct sk_buff *skb, struct genl_info *info)
-     sk = (struct sock *)msk;
--    if (id_val == 0) {
--        err = mptcp_userspace_pm_remove_id_zero_address(msk);
--        goto out;
--    }
--
-     lock_sock(sk);
--
--    spin_lock_bh(&msk->pm.lock);
--    match = mptcp_userspace_pm_lookup_addr_by_id(msk, id_val);
--    if (!match) {
--        spin_unlock_bh(&msk->pm.lock);
--        release_sock(sk);
--        goto out;
--    }
--
--    list_del_rcu(&match->list);
--    spin_unlock_bh(&msk->pm.lock);
--
--    mptcp_pm_remove_addr_entry(msk, match);
--
-+    err = userspace_pm_address_removed(msk, id_val);
-     release_sock(sk);
--
--    sock_kfree_s(sk, match, sizeof(*match));
--
--    err = 0;
--out:
-     if (err)
-         NL_SET_ERR_MSG_ATTR_FMT(info->extack, id,
-                     "address with id %u not found",
-@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_remove_doit(struct sk_buff *skb, struct genl_info *info)
-     return err;
- }
-+static int userspace_pm_subflow_established(struct mptcp_sock *msk,
-+                        struct mptcp_pm_addr_entry *entry,
-+                        struct mptcp_addr_info *remote)
-+{
-+    struct sock *sk = (struct sock *)msk;
-+    struct mptcp_pm_local local;
-+    int err;
-+
-+    err = mptcp_userspace_pm_append_new_local_addr(msk, entry, false);
-+    if (err < 0)
-+        return err;
-+
-+    local.addr = entry->addr;
-+    local.flags = entry->flags;
-+    local.ifindex = entry->ifindex;
-+
-+    err = __mptcp_subflow_connect(sk, &local, remote);
-+    spin_lock_bh(&msk->pm.lock);
-+    if (err)
-+        mptcp_userspace_pm_delete_local_addr(msk, entry);
-+    else
-+        msk->pm.subflows++;
-+    spin_unlock_bh(&msk->pm.lock);
-+
-+    return err;
-+}
-+
- int mptcp_pm_nl_subflow_create_doit(struct sk_buff *skb, struct genl_info *info)
- {
-     struct mptcp_pm_addr_entry entry = { 0 };
-     struct mptcp_addr_info addr_r;
-     struct nlattr *raddr, *laddr;
--    struct mptcp_pm_local local;
-     struct mptcp_sock *msk;
-     int err = -EINVAL;
-     struct sock *sk;
-@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_subflow_create_doit(struct sk_buff *skb, struct genl_info *info)
-         goto create_err;
-     }
--    err = mptcp_userspace_pm_append_new_local_addr(msk, &entry, false);
--    if (err < 0) {
--        NL_SET_ERR_MSG_ATTR(info->extack, laddr,
--                    "did not match address and id");
--        goto create_err;
--    }
--
--    local.addr = entry.addr;
--    local.flags = entry.flags;
--    local.ifindex = entry.ifindex;
--
-     lock_sock(sk);
--    err = __mptcp_subflow_connect(sk, &local, &addr_r);
-+    err = userspace_pm_subflow_established(msk, &entry, &addr_r);
-     release_sock(sk);
-     if (err)
-         GENL_SET_ERR_MSG_FMT(info, "connect error: %d", err);
--    spin_lock_bh(&msk->pm.lock);
--    if (err)
--        mptcp_userspace_pm_delete_local_addr(msk, &entry);
--    else
--        msk->pm.subflows++;
--    spin_unlock_bh(&msk->pm.lock);
--
-  create_err:
-     sock_put(sk);
-     return err;
-@@ -XXX,XX +XXX,XX @@ static struct sock *mptcp_nl_find_ssk(struct mptcp_sock *msk,
-     return NULL;
- }
-+static int userspace_pm_subflow_closed(struct mptcp_sock *msk,
-+                       struct mptcp_pm_addr_entry *local,
-+                       struct mptcp_addr_info *remote)
-+{
-+    struct sock *ssk, *sk = (struct sock *)msk;
-+
-+    ssk = mptcp_nl_find_ssk(msk, &local->addr, remote);
-+    if (!ssk)
-+        return -ESRCH;
-+
-+    spin_lock_bh(&msk->pm.lock);
-+    mptcp_userspace_pm_delete_local_addr(msk, local);
-+    spin_unlock_bh(&msk->pm.lock);
-+    mptcp_subflow_shutdown(sk, ssk, RCV_SHUTDOWN | SEND_SHUTDOWN);
-+    mptcp_close_ssk(sk, ssk, mptcp_subflow_ctx(ssk));
-+    MPTCP_INC_STATS(sock_net(sk), MPTCP_MIB_RMSUBFLOW);
-+
-+    return 0;
-+}
-+
- int mptcp_pm_nl_subflow_destroy_doit(struct sk_buff *skb, struct genl_info *info)
- {
-     struct mptcp_pm_addr_entry addr_l;
-     struct mptcp_addr_info addr_r;
-     struct nlattr *raddr, *laddr;
-     struct mptcp_sock *msk;
--    struct sock *sk, *ssk;
-     int err = -EINVAL;
-+    struct sock *sk;
-     if (GENL_REQ_ATTR_CHECK(info, MPTCP_PM_ATTR_ADDR) ||
-         GENL_REQ_ATTR_CHECK(info, MPTCP_PM_ATTR_ADDR_REMOTE))
-@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_subflow_destroy_doit(struct sk_buff *skb, struct genl_info *info
-     }
-     lock_sock(sk);
--    ssk = mptcp_nl_find_ssk(msk, &addr_l.addr, &addr_r);
--    if (!ssk) {
--        GENL_SET_ERR_MSG(info, "subflow not found");
--        err = -ESRCH;
--        goto release_sock;
--    }
--
--    spin_lock_bh(&msk->pm.lock);
--    mptcp_userspace_pm_delete_local_addr(msk, &addr_l);
--    spin_unlock_bh(&msk->pm.lock);
--    mptcp_subflow_shutdown(sk, ssk, RCV_SHUTDOWN | SEND_SHUTDOWN);
--    mptcp_close_ssk(sk, ssk, mptcp_subflow_ctx(ssk));
--    MPTCP_INC_STATS(sock_net(sk), MPTCP_MIB_RMSUBFLOW);
--release_sock:
-+    err = userspace_pm_subflow_closed(msk, &addr_l, &addr_r);
-     release_sock(sk);
-+    if (err)
-+        GENL_SET_ERR_MSG(info, "subflow not found");
- destroy_err:
-     sock_put(sk);
-     return err;
- }
-+static int userspace_pm_set_priority(struct mptcp_sock *msk,
-+                     struct mptcp_pm_addr_entry *local,
-+                     struct mptcp_addr_info *remote)
-+{
-+    struct mptcp_pm_addr_entry *entry;
-+    u8 bkup = 0;
-+
-+    if (local->flags & MPTCP_PM_ADDR_FLAG_BACKUP)
-+        bkup = 1;
-+
-+    spin_lock_bh(&msk->pm.lock);
-+    entry = mptcp_userspace_pm_lookup_addr(msk, &local->addr);
-+    if (entry) {
-+        if (bkup)
-+            entry->flags |= MPTCP_PM_ADDR_FLAG_BACKUP;
-+        else
-+            entry->flags &= ~MPTCP_PM_ADDR_FLAG_BACKUP;
-+    }
-+    spin_unlock_bh(&msk->pm.lock);
-+
-+    return mptcp_pm_nl_mp_prio_send_ack(msk, &local->addr, remote, bkup);
-+}
-+
- int mptcp_userspace_pm_set_flags(struct mptcp_pm_addr_entry *local,
-                  struct genl_info *info)
- {
-     struct mptcp_addr_info rem = { .family = AF_UNSPEC, };
--    struct mptcp_pm_addr_entry *entry;
-     struct nlattr *attr, *attr_rem;
-     struct mptcp_sock *msk;
-     int ret = -EINVAL;
-     struct sock *sk;
--    u8 bkup = 0;
-     if (GENL_REQ_ATTR_CHECK(info, MPTCP_PM_ATTR_ADDR_REMOTE))
-         return ret;
-@@ -XXX,XX +XXX,XX @@ int mptcp_userspace_pm_set_flags(struct mptcp_pm_addr_entry *local,
-         goto set_flags_err;
-     }
--    if (local->flags & MPTCP_PM_ADDR_FLAG_BACKUP)
--        bkup = 1;
--
--    spin_lock_bh(&msk->pm.lock);
--    entry = mptcp_userspace_pm_lookup_addr(msk, &local->addr);
--    if (entry) {
--        if (bkup)
--            entry->flags |= MPTCP_PM_ADDR_FLAG_BACKUP;
--        else
--            entry->flags &= ~MPTCP_PM_ADDR_FLAG_BACKUP;
--    }
--    spin_unlock_bh(&msk->pm.lock);
--
-     lock_sock(sk);
--    ret = mptcp_pm_nl_mp_prio_send_ack(msk, &local->addr, &rem, bkup);
-+    ret = userspace_pm_set_priority(msk, local, &rem);
-     release_sock(sk);
-     /* mptcp_pm_nl_mp_prio_send_ack() only fails in one case */
 @@ -XXX,XX +XXX,XX @@ int mptcp_userspace_pm_get_addr(u8 id, struct mptcp_pm_addr_entry *addr,
      sock_put(sk);
      return ret;
  }
 +
 +static struct mptcp_pm_ops mptcp_userspace_pm = {
-+    .address_announced    = userspace_pm_address_announced,
++    .get_local_id        = mptcp_userspace_pm_get_local_id,
-+    .address_removed    = userspace_pm_address_removed,
++    .get_priority        = mptcp_userspace_pm_get_priority,
 +    .subflow_established    = userspace_pm_subflow_established,
 +    .subflow_closed        = userspace_pm_subflow_closed,
 +    .get_local_id        = userspace_pm_get_local_id,
 +    .get_priority        = userspace_pm_get_priority,
 +    .set_priority        = userspace_pm_set_priority,
 +    .type            = MPTCP_PM_TYPE_USERSPACE,
++    .name            = "userspace",
 +    .owner            = THIS_MODULE,
 +};
 +
 +void __init mptcp_userspace_pm_init(void)
 +{
 ...

-[PATCH mptcp-next v4 4/4] mptcp: initialize and release mptcp_pm_ops
+[PATCH mptcp-next v6 07/13] mptcp: pm: initialize and release mptcp_pm_ops
 From: Geliang Tang <tanggeliang@kylinos.cn>
 Add a struct mptcp_pm_ops pointer "ops" in struct mptcp_pm_data, and two
 functions mptcp_pm_initialize() and mptcp_pm_release(), to set and release
 this pointer. mptcp_pm_initialize() is invoked in mptcp_pm_data_reset(),
-while mptcp_pm_release() is invoked in __mptcp_destroy_sock().
+while mptcp_pm_release() is invoked in mptcp_destroy_common().
 In this way, different userspace path managers can be initialized through
 the pm_type sysctl, and then called into their respective interfaces
 through "ops" of "msk->pm".
 Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
 ---
- net/mptcp/pm.c           | 49 ++++++++++++++++++++++++++++++++++++++--
+ net/mptcp/pm.c       | 38 ++++++++++++++++++++++++++++++++++++++
- net/mptcp/pm_netlink.c   | 11 ++++++++-
+ net/mptcp/protocol.c |  1 +
- net/mptcp/pm_userspace.c | 28 +++++++++++++++++------
+ net/mptcp/protocol.h |  3 +++
- net/mptcp/protocol.c     | 10 +++++++-
+files changed, 42 insertions(+)
  net/mptcp/protocol.h     |  3 +++
 files changed, 90 insertions(+), 11 deletions(-)
 diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
 index XXXXXXX..XXXXXXX 100644
 --- a/net/mptcp/pm.c
 +++ b/net/mptcp/pm.c
-@@ -XXX,XX +XXX,XX @@ void mptcp_pm_new_connection(struct mptcp_sock *msk, const struct sock *ssk, int
+@@ -XXX,XX +XXX,XX @@ bool mptcp_pm_addr_families_match(const struct sock *sk,
-     WRITE_ONCE(pm->server_side, server_side);
+ void mptcp_pm_data_reset(struct mptcp_sock *msk)
-     mptcp_event(MPTCP_EVENT_CREATED, msk, ssk, GFP_ATOMIC);
+ {
 +    const char *path_manager = mptcp_get_path_manager(sock_net((struct sock *)msk));
      u8 pm_type = mptcp_get_pm_type(sock_net((struct sock *)msk));
      struct mptcp_pm_data *pm = &msk->pm;
 +    int ret;
      pm->add_addr_signaled = 0;
      pm->add_addr_accepted = 0;
@@ -XXX,XX +XXX,XX @@ void mptcp_pm_data_reset(struct mptcp_sock *msk)
      pm->rm_list_rx.nr = 0;
      WRITE_ONCE(pm->pm_type, pm_type);
 +    rcu_read_lock();
 +    ret = mptcp_pm_initialize(msk, mptcp_pm_find(path_manager));
 +    rcu_read_unlock();
 +    if (ret)
 +        return;
 +
-+    if (pm->ops && pm->ops->created)
+     if (pm_type == MPTCP_PM_TYPE_KERNEL) {
-+        pm->ops->created(msk);
+         bool subflows_allowed = !!mptcp_pm_get_subflows_max(msk);
- }
  bool mptcp_pm_allow_new_subflow(struct mptcp_sock *msk)
@@ -XXX,XX +XXX,XX @@ void mptcp_pm_fully_established(struct mptcp_sock *msk, const struct sock *ssk)
      msk->pm.status |= BIT(MPTCP_PM_ALREADY_ESTABLISHED);
      spin_unlock_bh(&pm->lock);
 -    if (announce)
 +    if (announce) {
          mptcp_event(MPTCP_EVENT_ESTABLISHED, msk, ssk, GFP_ATOMIC);
 +
 +        if (pm->ops && pm->ops->established)
 +            pm->ops->established(msk);
 +    }
  }
  void mptcp_pm_connection_closed(struct mptcp_sock *msk)
  {
      pr_debug("msk=%p\n", msk);
 -    if (msk->token)
 +    if (msk->token) {
          mptcp_event(MPTCP_EVENT_CLOSED, msk, NULL, GFP_KERNEL);
 +
 +        if (msk->pm.ops && msk->pm.ops->closed)
 +            msk->pm.ops->closed(msk);
 +    }
  }
  void mptcp_pm_subflow_established(struct mptcp_sock *msk)
@@ -XXX,XX +XXX,XX @@ void mptcp_pm_data_reset(struct mptcp_sock *msk)
          WRITE_ONCE(pm->work_pending, 0);
          WRITE_ONCE(pm->accept_addr, 0);
          WRITE_ONCE(pm->accept_subflow, 0);
 +
 +        rcu_read_lock();
 +        mptcp_pm_initialize(msk, mptcp_pm_find(pm_type));
 +        rcu_read_unlock();
      }
      WRITE_ONCE(pm->addr_signal, 0);
 @@ -XXX,XX +XXX,XX @@ void mptcp_pm_unregister(struct mptcp_pm_ops *pm)
      list_del_rcu(&pm->list);
      spin_unlock(&mptcp_pm_list_lock);
  }
 +
 +int mptcp_pm_initialize(struct mptcp_sock *msk, struct mptcp_pm_ops *pm)
 +{
 +    if (!pm)
-+        return -EINVAL;
++        pm = &mptcp_in_kernel_pm;
 +
 +    if (!bpf_try_module_get(pm, pm->owner))
 +        return -EBUSY;
 +
 +    msk->pm.ops = pm;
 +    if (msk->pm.ops->init)
 +        msk->pm.ops->init(msk);
 +
-+    pr_debug("userspace_pm type %u initialized\n", msk->pm.ops->type);
++    pr_debug("pm %s initialized\n", pm->name);
 +    return 0;
 +}
 +
 +void mptcp_pm_release(struct mptcp_sock *msk)
 +{
 ...
 +    if (pm->release)
 +        pm->release(msk);
 +
 +    bpf_module_put(pm, pm->owner);
 +}
-diff --git a/net/mptcp/pm_netlink.c b/net/mptcp/pm_netlink.c
-index XXXXXXX..XXXXXXX 100644
---- a/net/mptcp/pm_netlink.c
-+++ b/net/mptcp/pm_netlink.c
-@@ -XXX,XX +XXX,XX @@ static int mptcp_pm_nl_create_listen_socket(struct sock *sk,
-     int addrlen = sizeof(struct sockaddr_in);
-     struct sockaddr_storage addr;
-     struct sock *newsk, *ssk;
-+    struct mptcp_sock *msk;
-     int backlog = 1024;
-     int err;
-@@ -XXX,XX +XXX,XX @@ static int mptcp_pm_nl_create_listen_socket(struct sock *sk,
-                       is_ipv6 ? "msk_lock-AF_INET6" : "msk_lock-AF_INET",
-                       &mptcp_keys[is_ipv6]);
-+    msk = mptcp_sk(newsk);
-     lock_sock(newsk);
--    ssk = __mptcp_nmpc_sk(mptcp_sk(newsk));
-+    ssk = __mptcp_nmpc_sk(msk);
-     release_sock(newsk);
-     if (IS_ERR(ssk))
-         return PTR_ERR(ssk);
-@@ -XXX,XX +XXX,XX @@ static int mptcp_pm_nl_create_listen_socket(struct sock *sk,
-     if (!err)
-         mptcp_event_pm_listener(ssk, MPTCP_EVENT_LISTENER_CREATED);
-     release_sock(ssk);
-+
-+    if (!err) {
-+        lock_sock(newsk);
-+        if (msk->pm.ops && msk->pm.ops->listener_created)
-+            msk->pm.ops->listener_created(msk);
-+        release_sock(newsk);
-+    }
-     return err;
- }
-diff --git a/net/mptcp/pm_userspace.c b/net/mptcp/pm_userspace.c
-index XXXXXXX..XXXXXXX 100644
---- a/net/mptcp/pm_userspace.c
-+++ b/net/mptcp/pm_userspace.c
-@@ -XXX,XX +XXX,XX @@ int mptcp_userspace_pm_get_local_id(struct mptcp_sock *msk,
-     if (new_entry.addr.port == msk_sport)
-         new_entry.addr.port = 0;
--    return userspace_pm_get_local_id(msk, &new_entry);
-+    return msk->pm.ops && msk->pm.ops->get_local_id ?
-+           msk->pm.ops->get_local_id(msk, &new_entry) :
-+           userspace_pm_get_local_id(msk, &new_entry);
- }
- static bool userspace_pm_get_priority(struct mptcp_sock *msk,
-@@ -XXX,XX +XXX,XX @@ static bool userspace_pm_get_priority(struct mptcp_sock *msk,
- bool mptcp_userspace_pm_is_backup(struct mptcp_sock *msk,
-                   struct mptcp_addr_info *skc)
- {
--    return userspace_pm_get_priority(msk, skc);
-+    return msk->pm.ops && msk->pm.ops->get_priority ?
-+           msk->pm.ops->get_priority(msk, skc) :
-+           userspace_pm_get_priority(msk, skc);
- }
- static struct mptcp_sock *mptcp_userspace_pm_get_sock(const struct genl_info *info)
-@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_announce_doit(struct sk_buff *skb, struct genl_info *info)
-     }
-     lock_sock(sk);
--    err = userspace_pm_address_announced(msk, &addr_val);
-+    err = msk->pm.ops && msk->pm.ops->address_announced ?
-+          msk->pm.ops->address_announced(msk, &addr_val) :
-+          userspace_pm_address_announced(msk, &addr_val);
-     release_sock(sk);
-     if (err)
-         NL_SET_ERR_MSG_ATTR(info->extack, addr,
-@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_remove_doit(struct sk_buff *skb, struct genl_info *info)
-     sk = (struct sock *)msk;
-     lock_sock(sk);
--    err = userspace_pm_address_removed(msk, id_val);
-+    err = msk->pm.ops && msk->pm.ops->address_removed ?
-+          msk->pm.ops->address_removed(msk, id_val) :
-+          userspace_pm_address_removed(msk, id_val);
-     release_sock(sk);
-     if (err)
-         NL_SET_ERR_MSG_ATTR_FMT(info->extack, id,
-@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_subflow_create_doit(struct sk_buff *skb, struct genl_info *info)
-     }
-     lock_sock(sk);
--    err = userspace_pm_subflow_established(msk, &entry, &addr_r);
-+    err = msk->pm.ops && msk->pm.ops->subflow_established ?
-+          msk->pm.ops->subflow_established(msk, &entry, &addr_r) :
-+          userspace_pm_subflow_established(msk, &entry, &addr_r);
-     release_sock(sk);
-     if (err)
-@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_subflow_destroy_doit(struct sk_buff *skb, struct genl_info *info
-     }
-     lock_sock(sk);
--    err = userspace_pm_subflow_closed(msk, &addr_l, &addr_r);
-+    err = msk->pm.ops && msk->pm.ops->subflow_closed ?
-+          msk->pm.ops->subflow_closed(msk, &addr_l, &addr_r) :
-+          userspace_pm_subflow_closed(msk, &addr_l, &addr_r);
-     release_sock(sk);
-     if (err)
-         GENL_SET_ERR_MSG(info, "subflow not found");
-@@ -XXX,XX +XXX,XX @@ int mptcp_userspace_pm_set_flags(struct mptcp_pm_addr_entry *local,
-     }
-     lock_sock(sk);
--    ret = userspace_pm_set_priority(msk, local, &rem);
-+    ret = msk->pm.ops && msk->pm.ops->set_priority ?
-+          msk->pm.ops->set_priority(msk, local, &rem) :
-+          userspace_pm_set_priority(msk, local, &rem);
-     release_sock(sk);
-     /* mptcp_pm_nl_mp_prio_send_ack() only fails in one case */
 diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c
 index XXXXXXX..XXXXXXX 100644
 --- a/net/mptcp/protocol.c
 +++ b/net/mptcp/protocol.c
-@@ -XXX,XX +XXX,XX @@ static void __mptcp_destroy_sock(struct sock *sk)
+@@ -XXX,XX +XXX,XX @@ void mptcp_destroy_common(struct mptcp_sock *msk, unsigned int flags)
-     sk_stop_timer(sk, &sk->sk_timer);
+     mptcp_token_destroy(msk);
-     msk->pm.status = 0;
+     mptcp_pm_free_anno_list(msk);
-     mptcp_release_sched(msk);
+     mptcp_free_local_addr_list(msk);
 +    mptcp_pm_release(msk);
-     sk->sk_prot->destroy(sk);
-@@ -XXX,XX +XXX,XX @@ static __poll_t mptcp_check_readable(struct sock *sk)
- static void mptcp_check_listen_stop(struct sock *sk)
- {
-+    struct mptcp_sock *msk = mptcp_sk(sk);
-     struct sock *ssk;
-     if (inet_sk_state_load(sk) != TCP_LISTEN)
-         return;
-     sock_prot_inuse_add(sock_net(sk), sk->sk_prot, -1);
--    ssk = mptcp_sk(sk)->first;
-+    ssk = msk->first;
-     if (WARN_ON_ONCE(!ssk || inet_sk_state_load(ssk) != TCP_LISTEN))
-         return;
-@@ -XXX,XX +XXX,XX @@ static void mptcp_check_listen_stop(struct sock *sk)
-     inet_csk_listen_stop(ssk);
-     mptcp_event_pm_listener(ssk, MPTCP_EVENT_LISTENER_CLOSED);
-     release_sock(ssk);
-+
-+    if (msk->pm.ops && msk->pm.ops->listener_closed)
-+        msk->pm.ops->listener_closed(msk);
  }
- bool __mptcp_close(struct sock *sk, long timeout)
+ static void mptcp_destroy(struct sock *sk)
@@ -XXX,XX +XXX,XX @@ static int mptcp_listen(struct socket *sock, int backlog)
          sock_prot_inuse_add(sock_net(sk), sk->sk_prot, 1);
          mptcp_copy_inaddrs(sk, ssk);
          mptcp_event_pm_listener(ssk, MPTCP_EVENT_LISTENER_CREATED);
 +
 +        if (msk->pm.ops && msk->pm.ops->listener_created)
 +            msk->pm.ops->listener_created(msk);
      }
  unlock:
 diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
 index XXXXXXX..XXXXXXX 100644
 --- a/net/mptcp/protocol.h
 +++ b/net/mptcp/protocol.h
 @@ -XXX,XX +XXX,XX @@ struct mptcp_pm_data {
 ...
      struct list_head userspace_pm_local_addr_list;
 +    struct mptcp_pm_ops *ops;
      spinlock_t    lock;        /*protects the whole PM data */
-@@ -XXX,XX +XXX,XX @@ struct mptcp_pm_ops *mptcp_pm_find(enum mptcp_pm_type type);
+@@ -XXX,XX +XXX,XX @@ struct mptcp_pm_ops *mptcp_pm_find(const char *name);
  int mptcp_pm_validate(struct mptcp_pm_ops *pm);
  int mptcp_pm_register(struct mptcp_pm_ops *pm);
  void mptcp_pm_unregister(struct mptcp_pm_ops *pm);
 +int mptcp_pm_initialize(struct mptcp_sock *msk, struct mptcp_pm_ops *pm);
 +void mptcp_pm_release(struct mptcp_sock *msk);
  void mptcp_free_local_addr_list(struct mptcp_sock *msk);
 --
 .43.0

-[PATCH mptcp-next v4 2/4] mptcp: define BPF path manager type
+[PATCH mptcp-next v6 08/13] mptcp: pm: drop pm_type in mptcp_pm_data
 From: Geliang Tang <tanggeliang@kylinos.cn>
-This patch defines a new path manager type: MPTCP_PM_TYPE_BPF_USERSPACE,
+Now pm->pm_type can be replaced by pm->ops->type, then "pm_type" filed
-which is used for the userspace path manager implemented in BPF. This
+of struct mptcp_pm_data can be dropped.
 path manager is also a userspace path manager, so add this new type in
 mptcp_pm_is_userspace() too.
 Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
 ---
- net/mptcp/protocol.h | 6 +++++-
+ net/mptcp/pm.c       | 6 ++----
-file changed, 5 insertions(+), 1 deletion(-)
+ net/mptcp/protocol.h | 5 ++---
 files changed, 4 insertions(+), 7 deletions(-)
+diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/pm.c
++++ b/net/mptcp/pm.c
+@@ -XXX,XX +XXX,XX @@ bool mptcp_pm_addr_families_match(const struct sock *sk,
+ void mptcp_pm_data_reset(struct mptcp_sock *msk)
+ {
+     const char *path_manager = mptcp_get_path_manager(sock_net((struct sock *)msk));
+-    u8 pm_type = mptcp_get_pm_type(sock_net((struct sock *)msk));
+     struct mptcp_pm_data *pm = &msk->pm;
+     int ret;
+@@ -XXX,XX +XXX,XX @@ 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, pm_type);
+     rcu_read_lock();
+     ret = mptcp_pm_initialize(msk, mptcp_pm_find(path_manager));
+@@ -XXX,XX +XXX,XX @@ void mptcp_pm_data_reset(struct mptcp_sock *msk)
+     if (ret)
+         return;
+-    if (pm_type == MPTCP_PM_TYPE_KERNEL) {
++    if (pm->ops->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
++         * pm->ops->type is set to MPTCP_PM_TYPE_KERNEL
+          */
+         WRITE_ONCE(pm->work_pending,
+                (!!mptcp_pm_get_local_addr_max(msk) &&
 diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
 index XXXXXXX..XXXXXXX 100644
 --- a/net/mptcp/protocol.h
 +++ b/net/mptcp/protocol.h
-@@ -XXX,XX +XXX,XX @@ enum mptcp_pm_status {
+@@ -XXX,XX +XXX,XX @@ struct mptcp_pm_data {
- enum mptcp_pm_type {
+     u8        add_addr_signaled;
-     MPTCP_PM_TYPE_KERNEL = 0,
+     u8        add_addr_accepted;
-     MPTCP_PM_TYPE_USERSPACE,
+     u8        local_addr_used;
-+    MPTCP_PM_TYPE_BPF_USERSPACE,
+-    u8        pm_type;
+     u8        subflows;
-     __MPTCP_PM_TYPE_NR,
+     u8        status;
-     __MPTCP_PM_TYPE_MAX = __MPTCP_PM_TYPE_NR - 1,
+     DECLARE_BITMAP(id_avail_bitmap, MPTCP_PM_MAX_ADDR_ID + 1);
 @@ -XXX,XX +XXX,XX @@ static inline bool mptcp_pm_should_rm_signal(struct mptcp_sock *msk)
  static inline bool mptcp_pm_is_userspace(const struct mptcp_sock *msk)
  {
 -    return READ_ONCE(msk->pm.pm_type) == MPTCP_PM_TYPE_USERSPACE;
-+    enum mptcp_pm_type pm_type = READ_ONCE(msk->pm.pm_type);
++    return msk->pm.ops->type == MPTCP_PM_TYPE_USERSPACE;
 +
 +    return (pm_type == MPTCP_PM_TYPE_USERSPACE ||
 +        pm_type == MPTCP_PM_TYPE_BPF_USERSPACE);
  }
  static inline bool mptcp_pm_is_kernel(const struct mptcp_sock *msk)
+ {
+-    return READ_ONCE(msk->pm.pm_type) == MPTCP_PM_TYPE_KERNEL;
++    return msk->pm.ops->type == MPTCP_PM_TYPE_KERNEL;
+ }
+ static inline unsigned int mptcp_add_addr_len(int family, bool echo, bool port)
 --
 .43.0

-New patch
+[PATCH mptcp-next v6 09/13] mptcp: sysctl: drop old pm_type sysctl
+From: Geliang Tang <tanggeliang@kylinos.cn>
+Old path manager sysctl "pm_type" can be replaced by the newly added
+"path_manager". Drop it together with mptcp_get_pm_type() helper.
+Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
+---
+ Documentation/networking/mptcp-sysctl.rst | 17 -----------------
+ net/mptcp/ctrl.c                          | 20 --------------------
+ net/mptcp/protocol.h                      |  1 -
+files changed, 38 deletions(-)
+diff --git a/Documentation/networking/mptcp-sysctl.rst b/Documentation/networking/mptcp-sysctl.rst
+index XXXXXXX..XXXXXXX 100644
+--- a/Documentation/networking/mptcp-sysctl.rst
++++ b/Documentation/networking/mptcp-sysctl.rst
+@@ -XXX,XX +XXX,XX @@ enabled - BOOLEAN
+     Default: 1 (enabled)
+-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
+-
+ path_manager - STRING
+     Set the default path manager name to use for each new MPTCP
+     socket. In-kernel path management will control subflow
+diff --git a/net/mptcp/ctrl.c b/net/mptcp/ctrl.c
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/ctrl.c
++++ b/net/mptcp/ctrl.c
+@@ -XXX,XX +XXX,XX @@
+ 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;
+@@ -XXX,XX +XXX,XX @@ struct mptcp_pernet {
+     u8 mptcp_enabled;
+     u8 checksum_enabled;
+     u8 allow_join_initial_addr_port;
+-    u8 pm_type;
+     char path_manager[MPTCP_PM_NAME_MAX];
+     char scheduler[MPTCP_SCHED_NAME_MAX];
+ };
+@@ -XXX,XX +XXX,XX @@ unsigned int mptcp_close_timeout(const struct sock *sk)
+     return mptcp_get_pernet(sock_net(sk))->close_timeout;
+ }
+-int mptcp_get_pm_type(const struct net *net)
+-{
+-    return mptcp_get_pernet(net)->pm_type;
+-}
+-
+ const char *mptcp_get_path_manager(const struct net *net)
+ {
+     return mptcp_get_pernet(net)->path_manager;
+@@ -XXX,XX +XXX,XX @@ 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;
+     strscpy(pernet->path_manager, "in-kernel", sizeof(pernet->path_manager));
+     strscpy(pernet->scheduler, "default", sizeof(pernet->scheduler));
+ }
+@@ -XXX,XX +XXX,XX @@ 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
+-    },
+     {
+         .procname = "path_manager",
+         .maxlen    = MPTCP_PM_NAME_MAX,
+@@ -XXX,XX +XXX,XX @@ static int mptcp_pernet_new_table(struct net *net, struct mptcp_pernet *pernet)
+     table[i++].data = &pernet->checksum_enabled;
+     table[i++].data = &pernet->allow_join_initial_addr_port;
+     table[i++].data = &pernet->stale_loss_cnt;
+-    table[i++].data = &pernet->pm_type;
+     table[i++].data = &pernet->path_manager;
+     table[i++].data = &pernet->scheduler;
+     i++; /* table[i] is for available_schedulers which is read-only info */
+diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/protocol.h
++++ b/net/mptcp/protocol.h
+@@ -XXX,XX +XXX,XX @@ 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);
+ unsigned int mptcp_close_timeout(const struct sock *sk);
+-int mptcp_get_pm_type(const struct net *net);
+ const char *mptcp_get_path_manager(const struct net *net);
+ const char *mptcp_get_scheduler(const struct net *net);
+--
+.43.0

-New patch
+[PATCH mptcp-next v6 10/13] mptcp: sysctl: add available_path_managers
+From: Geliang Tang <tanggeliang@kylinos.cn>
+Similarly to net.mptcp.available_schedulers, this patch adds a new one
+net.mptcp.available_path_managers to list the available path mangers.
+Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
+---
+ Documentation/networking/mptcp-sysctl.rst |  4 ++++
+ include/net/mptcp.h                       |  2 ++
+ net/mptcp/ctrl.c                          | 25 +++++++++++++++++++++++
+ net/mptcp/pm.c                            | 20 ++++++++++++++++++
+ net/mptcp/protocol.h                      |  1 +
+files changed, 52 insertions(+)
+diff --git a/Documentation/networking/mptcp-sysctl.rst b/Documentation/networking/mptcp-sysctl.rst
+index XXXXXXX..XXXXXXX 100644
+--- a/Documentation/networking/mptcp-sysctl.rst
++++ b/Documentation/networking/mptcp-sysctl.rst
+@@ -XXX,XX +XXX,XX @@ path_manager - STRING
+     Default: "in-kernel"
++available_path_managers - STRING
++    Shows the available path managers choices that are registered. More
++        path managers may be available, but not loaded.
++
+ scheduler - STRING
+     Select the scheduler of your choice.
+diff --git a/include/net/mptcp.h b/include/net/mptcp.h
+index XXXXXXX..XXXXXXX 100644
+--- a/include/net/mptcp.h
++++ b/include/net/mptcp.h
+@@ -XXX,XX +XXX,XX @@ struct mptcp_sched_ops {
+ } ____cacheline_aligned_in_smp;
+ #define MPTCP_PM_NAME_MAX    16
++#define MPTCP_PM_MAX        128
++#define MPTCP_PM_BUF_MAX    (MPTCP_PM_NAME_MAX * MPTCP_PM_MAX)
+ struct mptcp_pm_ops {
+     int (*created)(struct mptcp_sock *msk);
+diff --git a/net/mptcp/ctrl.c b/net/mptcp/ctrl.c
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/ctrl.c
++++ b/net/mptcp/ctrl.c
+@@ -XXX,XX +XXX,XX @@ static int proc_path_manager(const struct ctl_table *ctl, int write,
+     return ret;
+ }
++static int proc_available_path_managers(const struct ctl_table *ctl,
++                    int write, void *buffer,
++                    size_t *lenp, loff_t *ppos)
++{
++    struct ctl_table tbl = { .maxlen = MPTCP_PM_BUF_MAX, };
++    int ret;
++
++    tbl.data = kmalloc(tbl.maxlen, GFP_USER);
++    if (!tbl.data)
++        return -ENOMEM;
++
++    mptcp_pm_get_available(tbl.data, MPTCP_PM_BUF_MAX);
++    ret = proc_dostring(&tbl, write, buffer, lenp, ppos);
++    kfree(tbl.data);
++
++    return ret;
++}
++
+ static int mptcp_set_scheduler(char *scheduler, const char *name)
+ {
+     struct mptcp_sched_ops *sched;
+@@ -XXX,XX +XXX,XX @@ static struct ctl_table mptcp_sysctl_table[] = {
+         .mode = 0644,
+         .proc_handler = proc_path_manager,
+     },
++    {
++        .procname = "available_path_managers",
++        .maxlen    = MPTCP_PM_BUF_MAX,
++        .mode = 0444,
++        .proc_handler = proc_available_path_managers,
++    },
+     {
+         .procname = "scheduler",
+         .maxlen    = MPTCP_SCHED_NAME_MAX,
+@@ -XXX,XX +XXX,XX @@ static int mptcp_pernet_new_table(struct net *net, struct mptcp_pernet *pernet)
+     table[i++].data = &pernet->allow_join_initial_addr_port;
+     table[i++].data = &pernet->stale_loss_cnt;
+     table[i++].data = &pernet->path_manager;
++    i++; /* table[i] is for available_path_managers which is read-only info */
+     table[i++].data = &pernet->scheduler;
+     i++; /* table[i] is for available_schedulers which is read-only info */
+     table[i++].data = &pernet->close_timeout;
+diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/pm.c
++++ b/net/mptcp/pm.c
+@@ -XXX,XX +XXX,XX @@ struct mptcp_pm_ops *mptcp_pm_find(const char *name)
+     return NULL;
+ }
++/* Build string with list of available path manager values.
++ * Similar to tcp_get_available_congestion_control()
++ */
++void mptcp_pm_get_available(char *buf, size_t maxlen)
++{
++    struct mptcp_pm_ops *pm;
++    size_t offs = 0;
++
++    rcu_read_lock();
++    list_for_each_entry_rcu(pm, &mptcp_pm_list, list) {
++        offs += snprintf(buf + offs, maxlen - offs,
++                 "%s%s",
++                 offs == 0 ? "" : " ", pm->name);
++
++        if (WARN_ON_ONCE(offs >= maxlen))
++            break;
++    }
++    rcu_read_unlock();
++}
++
+ int mptcp_pm_validate(struct mptcp_pm_ops *pm)
+ {
+     if (!pm->get_local_id || !pm->get_priority) {
+diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/protocol.h
++++ b/net/mptcp/protocol.h
+@@ -XXX,XX +XXX,XX @@ void mptcp_pm_remove_addr_entry(struct mptcp_sock *msk,
+ extern struct mptcp_pm_ops mptcp_in_kernel_pm;
+ struct mptcp_pm_ops *mptcp_pm_find(const char *name);
++void mptcp_pm_get_available(char *buf, size_t maxlen);
+ int mptcp_pm_validate(struct mptcp_pm_ops *pm);
+ int mptcp_pm_register(struct mptcp_pm_ops *pm);
+ void mptcp_pm_unregister(struct mptcp_pm_ops *pm);
+--
+.43.0

-New patch
+[PATCH mptcp-next v6 11/13] mptcp: pm: drop get_local_id helpers
+From: Geliang Tang <tanggeliang@kylinos.cn>
+Now mptcp_pm_nl_get_local_id() and mptcp_userspace_pm_get_local_id()
+helpers can be dropped, and mptcp_pm_get_local_id() can directly invoke
+get_local_id() interface through "ops" of "msk->pm".
+Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
+---
+ net/mptcp/pm.c           | 4 +---
+ net/mptcp/pm_netlink.c   | 4 ++--
+ net/mptcp/pm_userspace.c | 4 ++--
+ net/mptcp/protocol.h     | 4 ----
+files changed, 5 insertions(+), 11 deletions(-)
+diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/pm.c
++++ b/net/mptcp/pm.c
+@@ -XXX,XX +XXX,XX @@ int mptcp_pm_get_local_id(struct mptcp_sock *msk, struct sock_common *skc)
+     skc_local.addr.id = 0;
+     skc_local.flags = MPTCP_PM_ADDR_FLAG_IMPLICIT;
+-    if (mptcp_pm_is_userspace(msk))
+-        return mptcp_userspace_pm_get_local_id(msk, &skc_local);
+-    return mptcp_pm_nl_get_local_id(msk, &skc_local);
++    return msk->pm.ops->get_local_id(msk, &skc_local);
+ }
+ bool mptcp_pm_is_backup(struct mptcp_sock *msk, struct sock_common *skc)
+diff --git a/net/mptcp/pm_netlink.c b/net/mptcp/pm_netlink.c
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/pm_netlink.c
++++ b/net/mptcp/pm_netlink.c
+@@ -XXX,XX +XXX,XX @@ static int mptcp_pm_nl_create_listen_socket(struct sock *sk,
+     return err;
+ }
+-int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk,
+-                 struct mptcp_pm_addr_entry *skc)
++static int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk,
++                    struct mptcp_pm_addr_entry *skc)
+ {
+     struct mptcp_pm_addr_entry *entry;
+     struct pm_nl_pernet *pernet;
+diff --git a/net/mptcp/pm_userspace.c b/net/mptcp/pm_userspace.c
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/pm_userspace.c
++++ b/net/mptcp/pm_userspace.c
+@@ -XXX,XX +XXX,XX @@ mptcp_userspace_pm_lookup_addr_by_id(struct mptcp_sock *msk, unsigned int id)
+     return NULL;
+ }
+-int mptcp_userspace_pm_get_local_id(struct mptcp_sock *msk,
+-                    struct mptcp_pm_addr_entry *skc)
++static int mptcp_userspace_pm_get_local_id(struct mptcp_sock *msk,
++                       struct mptcp_pm_addr_entry *skc)
+ {
+     __be16 msk_sport =  ((struct inet_sock *)
+                  inet_sk((struct sock *)msk))->inet_sport;
+diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/protocol.h
++++ b/net/mptcp/protocol.h
+@@ -XXX,XX +XXX,XX @@ bool mptcp_pm_add_addr_signal(struct mptcp_sock *msk, const struct sk_buff *skb,
+ bool mptcp_pm_rm_addr_signal(struct mptcp_sock *msk, unsigned int remaining,
+                  struct mptcp_rm_list *rm_list);
+ int mptcp_pm_get_local_id(struct mptcp_sock *msk, struct sock_common *skc);
+-int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk,
+-                 struct mptcp_pm_addr_entry *skc);
+-int mptcp_userspace_pm_get_local_id(struct mptcp_sock *msk,
+-                    struct mptcp_pm_addr_entry *skc);
+ bool mptcp_pm_is_backup(struct mptcp_sock *msk, struct sock_common *skc);
+ bool mptcp_pm_nl_is_backup(struct mptcp_sock *msk, struct mptcp_addr_info *skc);
+ bool mptcp_userspace_pm_is_backup(struct mptcp_sock *msk, struct mptcp_addr_info *skc);
+--
+.43.0

-New patch
+[PATCH mptcp-next v6 12/13] mptcp: pm: drop is_backup helpers
+From: Geliang Tang <tanggeliang@kylinos.cn>
+Now mptcp_pm_nl_is_backup() and mptcp_userspace_pm_is_backup()
+helpers can be dropped, and mptcp_pm_is_backup() can directly
+invoke get_priority() interface through "ops" of "msk->pm".
+Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
+---
+ net/mptcp/pm.c           | 5 +----
+ net/mptcp/pm_netlink.c   | 5 -----
+ net/mptcp/pm_userspace.c | 6 ------
+ net/mptcp/protocol.h     | 2 --
+files changed, 1 insertion(+), 17 deletions(-)
+diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/pm.c
++++ b/net/mptcp/pm.c
+@@ -XXX,XX +XXX,XX @@ bool mptcp_pm_is_backup(struct mptcp_sock *msk, struct sock_common *skc)
+     mptcp_local_address((struct sock_common *)skc, &skc_local);
+-    if (mptcp_pm_is_userspace(msk))
+-        return mptcp_userspace_pm_is_backup(msk, &skc_local);
+-
+-    return mptcp_pm_nl_is_backup(msk, &skc_local);
++    return msk->pm.ops->get_priority(msk, &skc_local);
+ }
+ static int mptcp_pm_get_addr(u8 id, struct mptcp_pm_addr_entry *addr,
+diff --git a/net/mptcp/pm_netlink.c b/net/mptcp/pm_netlink.c
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/pm_netlink.c
++++ b/net/mptcp/pm_netlink.c
+@@ -XXX,XX +XXX,XX @@ static bool mptcp_pm_nl_get_priority(struct mptcp_sock *msk,
+     return backup;
+ }
+-bool mptcp_pm_nl_is_backup(struct mptcp_sock *msk, struct mptcp_addr_info *skc)
+-{
+-    return mptcp_pm_nl_get_priority(msk, skc);
+-}
+-
+ #define MPTCP_PM_CMD_GRP_OFFSET       0
+ #define MPTCP_PM_EV_GRP_OFFSET        1
+diff --git a/net/mptcp/pm_userspace.c b/net/mptcp/pm_userspace.c
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/pm_userspace.c
++++ b/net/mptcp/pm_userspace.c
+@@ -XXX,XX +XXX,XX @@ static bool mptcp_userspace_pm_get_priority(struct mptcp_sock *msk,
+     return backup;
+ }
+-bool mptcp_userspace_pm_is_backup(struct mptcp_sock *msk,
+-                  struct mptcp_addr_info *skc)
+-{
+-    return mptcp_userspace_pm_get_priority(msk, skc);
+-}
+-
+ static struct mptcp_sock *mptcp_userspace_pm_get_sock(const struct genl_info *info)
+ {
+     struct mptcp_sock *msk;
+diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
+index XXXXXXX..XXXXXXX 100644
+--- a/net/mptcp/protocol.h
++++ b/net/mptcp/protocol.h
+@@ -XXX,XX +XXX,XX @@ bool mptcp_pm_rm_addr_signal(struct mptcp_sock *msk, unsigned int remaining,
+                  struct mptcp_rm_list *rm_list);
+ int mptcp_pm_get_local_id(struct mptcp_sock *msk, struct sock_common *skc);
+ bool mptcp_pm_is_backup(struct mptcp_sock *msk, struct sock_common *skc);
+-bool mptcp_pm_nl_is_backup(struct mptcp_sock *msk, struct mptcp_addr_info *skc);
+-bool mptcp_userspace_pm_is_backup(struct mptcp_sock *msk, struct mptcp_addr_info *skc);
+ int mptcp_pm_nl_dump_addr(struct sk_buff *msg,
+               struct netlink_callback *cb);
+ int mptcp_userspace_pm_dump_addr(struct sk_buff *msg,
+--
+.43.0

-New patch
+[PATCH mptcp-next v6 13/13] selftests/mptcp: use path_manager instead of pm_type
+From: Geliang Tang <tanggeliang@kylinos.cn>
+This patch uses the newly added net.mptcp.path_manager instead of the
+old net.mptcp.pm_type for all path manager selftests in mptcp_join.sh
+and userspace_pm.sh.
+Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
+---
+ .../testing/selftests/net/mptcp/mptcp_join.sh | 26 +++++++++----------
+ .../selftests/net/mptcp/userspace_pm.sh       |  4 +--
+files changed, 15 insertions(+), 15 deletions(-)
+diff --git a/tools/testing/selftests/net/mptcp/mptcp_join.sh b/tools/testing/selftests/net/mptcp/mptcp_join.sh
+index XXXXXXX..XXXXXXX 100755
+--- a/tools/testing/selftests/net/mptcp/mptcp_join.sh
++++ b/tools/testing/selftests/net/mptcp/mptcp_join.sh
+@@ -XXX,XX +XXX,XX @@ init_partial()
+     local netns
+     for netns in "$ns1" "$ns2"; do
+-        ip netns exec $netns sysctl -q net.mptcp.pm_type=0 2>/dev/null || true
++        ip netns exec $netns sysctl -q net.mptcp.path_manager="in-kernel" 2>/dev/null || true
+         if $checksum; then
+             ip netns exec $netns sysctl -q net.mptcp.checksum_enabled=1
+         fi
+@@ -XXX,XX +XXX,XX @@ set_userspace_pm()
+ {
+     local ns=$1
+-    ip netns exec $ns sysctl -q net.mptcp.pm_type=1
++    ip netns exec $ns sysctl -q net.mptcp.path_manager="userspace"
+ }
+ subflows_tests()
+@@ -XXX,XX +XXX,XX @@ userspace_tests()
+ {
+     # userspace pm type prevents add_addr
+     if reset "userspace pm type prevents add_addr" &&
+-       continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
++       continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
+         set_userspace_pm $ns1
+         pm_nl_set_limits $ns1 0 2
+         pm_nl_set_limits $ns2 0 2
+@@ -XXX,XX +XXX,XX @@ userspace_tests()
+     # userspace pm type does not echo add_addr without daemon
+     if reset "userspace pm no echo w/o daemon" &&
+-       continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
++       continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
+         set_userspace_pm $ns2
+         pm_nl_set_limits $ns1 0 2
+         pm_nl_set_limits $ns2 0 2
+@@ -XXX,XX +XXX,XX @@ userspace_tests()
+     # userspace pm type rejects join
+     if reset "userspace pm type rejects join" &&
+-       continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
++       continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
+         set_userspace_pm $ns1
+         pm_nl_set_limits $ns1 1 1
+         pm_nl_set_limits $ns2 1 1
+@@ -XXX,XX +XXX,XX @@ userspace_tests()
+     # userspace pm type does not send join
+     if reset "userspace pm type does not send join" &&
+-       continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
++       continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
+         set_userspace_pm $ns2
+         pm_nl_set_limits $ns1 1 1
+         pm_nl_set_limits $ns2 1 1
+@@ -XXX,XX +XXX,XX @@ userspace_tests()
+     # userspace pm type prevents mp_prio
+     if reset "userspace pm type prevents mp_prio" &&
+-       continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
++       continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
+         set_userspace_pm $ns1
+         pm_nl_set_limits $ns1 1 1
+         pm_nl_set_limits $ns2 1 1
+@@ -XXX,XX +XXX,XX @@ userspace_tests()
+     # userspace pm type prevents rm_addr
+     if reset "userspace pm type prevents rm_addr" &&
+-       continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
++       continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
+         set_userspace_pm $ns1
+         set_userspace_pm $ns2
+         pm_nl_set_limits $ns1 0 1
+@@ -XXX,XX +XXX,XX @@ userspace_tests()
+     # userspace pm add & remove address
+     if reset_with_events "userspace pm add & remove address" &&
+-       continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
++       continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
+         set_userspace_pm $ns1
+         pm_nl_set_limits $ns2 2 2
+         { speed=5 \
+@@ -XXX,XX +XXX,XX @@ userspace_tests()
+     # userspace pm create destroy subflow
+     if reset_with_events "userspace pm create destroy subflow" &&
+-       continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
++       continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
+         set_userspace_pm $ns2
+         pm_nl_set_limits $ns1 0 1
+         { speed=5 \
+@@ -XXX,XX +XXX,XX @@ userspace_tests()
+     # userspace pm create id 0 subflow
+     if reset_with_events "userspace pm create id 0 subflow" &&
+-       continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
++       continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
+         set_userspace_pm $ns2
+         pm_nl_set_limits $ns1 0 1
+         { speed=5 \
+@@ -XXX,XX +XXX,XX @@ userspace_tests()
+     # userspace pm remove initial subflow
+     if reset_with_events "userspace pm remove initial subflow" &&
+-       continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
++       continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
+         set_userspace_pm $ns2
+         pm_nl_set_limits $ns1 0 1
+         { speed=5 \
+@@ -XXX,XX +XXX,XX @@ userspace_tests()
+     # userspace pm send RM_ADDR for ID 0
+     if reset_with_events "userspace pm send RM_ADDR for ID 0" &&
+-       continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
++       continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
+         set_userspace_pm $ns1
+         pm_nl_set_limits $ns2 1 1
+         { speed=5 \
+diff --git a/tools/testing/selftests/net/mptcp/userspace_pm.sh b/tools/testing/selftests/net/mptcp/userspace_pm.sh
+index XXXXXXX..XXXXXXX 100755
+--- a/tools/testing/selftests/net/mptcp/userspace_pm.sh
++++ b/tools/testing/selftests/net/mptcp/userspace_pm.sh
+@@ -XXX,XX +XXX,XX @@
+ mptcp_lib_check_mptcp
+ mptcp_lib_check_kallsyms
+-if ! mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
++if ! mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
+     echo "userspace pm tests are not supported by the kernel: SKIP"
+     exit ${KSFT_SKIP}
+ fi
+@@ -XXX,XX +XXX,XX @@ trap cleanup EXIT
+ # Create and configure network namespaces for testing
+ mptcp_lib_ns_init ns1 ns2
+ for i in "$ns1" "$ns2" ;do
+-    ip netns exec "$i" sysctl -q net.mptcp.pm_type=1
++    ip netns exec "$i" sysctl -q net.mptcp.path_manager="userspace"
+ done
+ #  "$ns1"              ns2
+--
+.43.0

From: Geliang Tang <tanggeliang@kylinos.cn>

v4:
 - include a new patch "define BPF path manager type".

- add new interfaces:
	created established closed
	listerner_created listener_closed

- rename interfaces as:
	address_announced address_removed
	subflow_established subflow_closed
	get_priority set_priority

- rename functions as:
	mptcp_pm_validate
	mptcp_pm_register
	mptcp_pm_unregister
	mptcp_pm_initialize
	mptcp_pm_release

v3:
 - rename the 2nd parameter of get_local_id() from 'local' to 'skc'.
 - keep the 'msk_sport' check in mptcp_userspace_pm_get_local_id().
 - return 'err' instead of '0' in userspace_pm_subflow_create().
 - drop 'ret' variable inmptcp_pm_data_reset().
 - fix typos in commit log.

v2:
 - update get_local_id interface in patch 2.

get_addr() and dump_addr() interfaces of BPF userspace pm are dropped
as Matt suggested.

In order to implement BPF userspace path manager, it is necessary to
unify the interfaces of the path manager. This set contains some
cleanups and refactoring to unify the interfaces in kernel space.
Finally, define a struct mptcp_pm_ops for a userspace path manager
like this:

struct mptcp_pm_ops {
        int (*created)(struct mptcp_sock *msk);
        int (*established)(struct mptcp_sock *msk);
        int (*closed)(struct mptcp_sock *msk);
        int (*address_announced)(struct mptcp_sock *msk,
                                 struct mptcp_pm_addr_entry *local);
        int (*address_removed)(struct mptcp_sock *msk, u8 id);
        int (*subflow_established)(struct mptcp_sock *msk,
                                   struct mptcp_pm_addr_entry *local,
                                   struct mptcp_addr_info *remote);
        int (*subflow_closed)(struct mptcp_sock *msk,
                              struct mptcp_pm_addr_entry *local,
                              struct mptcp_addr_info *remote);
        int (*get_local_id)(struct mptcp_sock *msk,
                            struct mptcp_pm_addr_entry *skc);
        bool (*get_priority)(struct mptcp_sock *msk,
                             struct mptcp_addr_info *skc);
        int (*set_priority)(struct mptcp_sock *msk,
                            struct mptcp_pm_addr_entry *local,
                            struct mptcp_addr_info *remote);
        int (*listener_created)(struct mptcp_sock *msk);
        int (*listener_closed)(struct mptcp_sock *msk);

u8                      type;
        struct module           *owner;
        struct list_head        list;

void (*init)(struct mptcp_sock *msk);
        void (*release)(struct mptcp_sock *msk);
} ____cacheline_aligned_in_smp;

Geliang Tang (4):
  mptcp: define struct mptcp_pm_ops
  mptcp: define BPF path manager type
  mptcp: register default userspace pm
  mptcp: initialize and release mptcp_pm_ops

-- 
2.43.0

From: Geliang Tang <tanggeliang@kylinos.cn>

In order to allow users to develop their own BPF-based path manager,
this patch defines a struct ops "mptcp_pm_ops" for a userspace path
manager, which contains a set of interfaces.

Add a set of functions to register, unregister, find and validate a
given struct ops.

Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 include/net/mptcp.h  | 32 ++++++++++++++++++++++++
 net/mptcp/pm.c       | 59 ++++++++++++++++++++++++++++++++++++++++++++
 net/mptcp/protocol.h |  5 ++++
 3 files changed, 96 insertions(+)

diff --git a/include/net/mptcp.h b/include/net/mptcp.h
index XXXXXXX..XXXXXXX 100644
--- a/include/net/mptcp.h
+++ b/include/net/mptcp.h
@@ -XXX,XX +XXX,XX @@
 
 struct mptcp_info;
 struct mptcp_sock;
+struct mptcp_pm_addr_entry;
 struct seq_file;
 
 /* MPTCP sk_buff extension data */
@@ -XXX,XX +XXX,XX @@ struct mptcp_sched_ops {
 	void (*release)(struct mptcp_sock *msk);
 } ____cacheline_aligned_in_smp;
 
+struct mptcp_pm_ops {
+	int (*created)(struct mptcp_sock *msk);
+	int (*established)(struct mptcp_sock *msk);
+	int (*closed)(struct mptcp_sock *msk);
+	int (*address_announced)(struct mptcp_sock *msk,
+				 struct mptcp_pm_addr_entry *local);
+	int (*address_removed)(struct mptcp_sock *msk, u8 id);
+	int (*subflow_established)(struct mptcp_sock *msk,
+				   struct mptcp_pm_addr_entry *local,
+				   struct mptcp_addr_info *remote);
+	int (*subflow_closed)(struct mptcp_sock *msk,
+			      struct mptcp_pm_addr_entry *local,
+			      struct mptcp_addr_info *remote);
+	int (*get_local_id)(struct mptcp_sock *msk,
+			    struct mptcp_pm_addr_entry *skc);
+	bool (*get_priority)(struct mptcp_sock *msk,
+			     struct mptcp_addr_info *skc);
+	int (*set_priority)(struct mptcp_sock *msk,
+			    struct mptcp_pm_addr_entry *local,
+			    struct mptcp_addr_info *remote);
+	int (*listener_created)(struct mptcp_sock *msk);
+	int (*listener_closed)(struct mptcp_sock *msk);
+
+	u8			type;
+	struct module		*owner;
+	struct list_head	list;
+
+	void (*init)(struct mptcp_sock *msk);
+	void (*release)(struct mptcp_sock *msk);
+} ____cacheline_aligned_in_smp;
+
 #ifdef CONFIG_MPTCP
 void mptcp_init(void);
 
diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/pm.c
+++ b/net/mptcp/pm.c
@@ -XXX,XX +XXX,XX @@
 #define pr_fmt(fmt) "MPTCP: " fmt
 
 #include <linux/kernel.h>
+#include <linux/rculist.h>
+#include <linux/spinlock.h>
 #include <net/mptcp.h>
 #include "protocol.h"
 
 #include "mib.h"
 #include "mptcp_pm_gen.h"
 
+static DEFINE_SPINLOCK(mptcp_pm_list_lock);
+static LIST_HEAD(mptcp_pm_list);
+
 /* path manager command handlers */
 
 int mptcp_pm_announce_addr(struct mptcp_sock *msk,
@@ -XXX,XX +XXX,XX @@ void __init mptcp_pm_init(void)
 {
 	mptcp_pm_nl_init();
 }
+
+/* Must be called with rcu read lock held */
+struct mptcp_pm_ops *mptcp_pm_find(enum mptcp_pm_type type)
+{
+	struct mptcp_pm_ops *pm;
+
+	list_for_each_entry_rcu(pm, &mptcp_pm_list, list) {
+		if (pm->type == type)
+			return pm;
+	}
+
+	return NULL;
+}
+
+int mptcp_pm_validate(struct mptcp_pm_ops *pm)
+{
+	if (!pm->created && !pm->established && !pm->closed &&
+	    !pm->address_announced && !pm->address_removed &&
+	    !pm->subflow_established && !pm->subflow_closed &&
+	    !pm->get_local_id && !pm->get_priority && !pm->set_priority &&
+	    !pm->listener_created && !pm->listener_closed) {
+		pr_err("%u does not implement required ops\n", pm->type);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+int mptcp_pm_register(struct mptcp_pm_ops *pm)
+{
+	int ret;
+
+	ret = mptcp_pm_validate(pm);
+	if (ret)
+		return ret;
+
+	spin_lock(&mptcp_pm_list_lock);
+	if (mptcp_pm_find(pm->type)) {
+		spin_unlock(&mptcp_pm_list_lock);
+		return -EEXIST;
+	}
+	list_add_tail_rcu(&pm->list, &mptcp_pm_list);
+	spin_unlock(&mptcp_pm_list_lock);
+
+	pr_debug("userspace_pm type %u registered\n", pm->type);
+	return 0;
+}
+
+void mptcp_pm_unregister(struct mptcp_pm_ops *pm)
+{
+	spin_lock(&mptcp_pm_list_lock);
+	list_del_rcu(&pm->list);
+	spin_unlock(&mptcp_pm_list_lock);
+}
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -XXX,XX +XXX,XX @@ int mptcp_pm_remove_addr(struct mptcp_sock *msk, const struct mptcp_rm_list *rm_
 void mptcp_pm_remove_addr_entry(struct mptcp_sock *msk,
 				struct mptcp_pm_addr_entry *entry);
 
+struct mptcp_pm_ops *mptcp_pm_find(enum mptcp_pm_type type);
+int mptcp_pm_validate(struct mptcp_pm_ops *pm);
+int mptcp_pm_register(struct mptcp_pm_ops *pm);
+void mptcp_pm_unregister(struct mptcp_pm_ops *pm);
+
 void mptcp_free_local_addr_list(struct mptcp_sock *msk);
 
 void mptcp_event(enum mptcp_event_type type, const struct mptcp_sock *msk,
-- 
2.43.0

From: Geliang Tang <tanggeliang@kylinos.cn>

This patch defines a new path manager type: MPTCP_PM_TYPE_BPF_USERSPACE,
which is used for the userspace path manager implemented in BPF. This
path manager is also a userspace path manager, so add this new type in
mptcp_pm_is_userspace() too.

Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 net/mptcp/protocol.h | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -XXX,XX +XXX,XX @@ enum mptcp_pm_status {
 enum mptcp_pm_type {
 	MPTCP_PM_TYPE_KERNEL = 0,
 	MPTCP_PM_TYPE_USERSPACE,
+	MPTCP_PM_TYPE_BPF_USERSPACE,
 
 	__MPTCP_PM_TYPE_NR,
 	__MPTCP_PM_TYPE_MAX = __MPTCP_PM_TYPE_NR - 1,
@@ -XXX,XX +XXX,XX @@ static inline bool mptcp_pm_should_rm_signal(struct mptcp_sock *msk)
 
 static inline bool mptcp_pm_is_userspace(const struct mptcp_sock *msk)
 {
-	return READ_ONCE(msk->pm.pm_type) == MPTCP_PM_TYPE_USERSPACE;
+	enum mptcp_pm_type pm_type = READ_ONCE(msk->pm.pm_type);
+
+	return (pm_type == MPTCP_PM_TYPE_USERSPACE ||
+		pm_type == MPTCP_PM_TYPE_BPF_USERSPACE);
 }
 
 static inline bool mptcp_pm_is_kernel(const struct mptcp_sock *msk)
-- 
2.43.0

From: Geliang Tang <tanggeliang@kylinos.cn>

This patch defines the original userspace pm as the default path manager,
named mptcp_userspace_pm, and register it in mptcp_pm_data_init().

Extract address_announce() interface of the userspace PM from the handler
of netlink command MPTCP_PM_CMD_ANNOUNCE mptcp_pm_nl_announce_doit(), only
leave the code for obtaining msk through "info" and parsing address entry
in the handler.

Extract address_remove() interface of the userspace PM from the handler
of netlink command MPTCP_PM_CMD_REMOVE mptcp_pm_nl_remove_doit(), only
leave the code for parsing address id and obtaining msk through "info"
in the handler.

Extract subflow_create() interface of the userspace PM from the handler of
netlink command MPTCP_PM_CMD_SUBFLOW_CREATE

mptcp_pm_nl_subflow_create_doit(),

only leave the code for obtaining msk through "info", parsing local address
entry and parsing remote address info in the handler.

Extract subflow_destroy() interface of the userspace PM from the handler of
netlink command MPTCP_PM_CMD_SUBFLOW_DESTROY

mptcp_pm_nl_subflow_destroy_doit(),

only leave the code for obtaining msk through "info", parsing local address
entry and parsing remote address info in the handler.

Extract set_flags() interface of the userspace PM from function
mptcp_userspace_pm_set_flags(), only leave the code for obtaining
msk through "info" in this function.

All fives interfaces are invoked under holding the msk socket lock.

Extract get_local_id() interface of the userspace PM from function
mptcp_userspace_pm_get_local_id(), only leave the code for setting
new address entry in this function.

get_flags() interface of the userspace PM is defined as the same as
the function mptcp_userspace_pm_get_flags(), then this function
becomes a wrapper.

Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 net/mptcp/pm.c           |   1 +
 net/mptcp/pm_userspace.c | 280 +++++++++++++++++++++++----------------
 net/mptcp/protocol.h     |   1 +
 3 files changed, 171 insertions(+), 111 deletions(-)

diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/pm.c
+++ b/net/mptcp/pm.c
@@ -XXX,XX +XXX,XX @@ void mptcp_pm_data_init(struct mptcp_sock *msk)
 void __init mptcp_pm_init(void)
 {
 	mptcp_pm_nl_init();
+	mptcp_userspace_pm_init();
 }
 
 /* Must be called with rcu read lock held */
diff --git a/net/mptcp/pm_userspace.c b/net/mptcp/pm_userspace.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/pm_userspace.c
+++ b/net/mptcp/pm_userspace.c
@@ -XXX,XX +XXX,XX @@ mptcp_userspace_pm_lookup_addr_by_id(struct mptcp_sock *msk, unsigned int id)
 	return NULL;
 }
 
-int mptcp_userspace_pm_get_local_id(struct mptcp_sock *msk,
-				    struct mptcp_addr_info *skc)
+static int userspace_pm_get_local_id(struct mptcp_sock *msk,
+				     struct mptcp_pm_addr_entry *skc)
 {
-	struct mptcp_pm_addr_entry *entry = NULL, new_entry;
-	__be16 msk_sport =  ((struct inet_sock *)
-			     inet_sk((struct sock *)msk))->inet_sport;
+	struct mptcp_pm_addr_entry *entry;
 
 	spin_lock_bh(&msk->pm.lock);
-	entry = mptcp_userspace_pm_lookup_addr(msk, skc);
+	entry = mptcp_userspace_pm_lookup_addr(msk, &skc->addr);
 	spin_unlock_bh(&msk->pm.lock);
 	if (entry)
 		return entry->addr.id;
 
+	return mptcp_userspace_pm_append_new_local_addr(msk, skc, true);
+}
+
+int mptcp_userspace_pm_get_local_id(struct mptcp_sock *msk,
+				    struct mptcp_addr_info *skc)
+{
+	__be16 msk_sport = ((struct inet_sock *)
+			    inet_sk((struct sock *)msk))->inet_sport;
+	struct mptcp_pm_addr_entry new_entry;
+
 	memset(&new_entry, 0, sizeof(struct mptcp_pm_addr_entry));
 	new_entry.addr = *skc;
 	new_entry.addr.id = 0;
@@ -XXX,XX +XXX,XX @@ int mptcp_userspace_pm_get_local_id(struct mptcp_sock *msk,
 	if (new_entry.addr.port == msk_sport)
 		new_entry.addr.port = 0;
 
-	return mptcp_userspace_pm_append_new_local_addr(msk, &new_entry, true);
+	return userspace_pm_get_local_id(msk, &new_entry);
 }
 
-bool mptcp_userspace_pm_is_backup(struct mptcp_sock *msk,
-				  struct mptcp_addr_info *skc)
+static bool userspace_pm_get_priority(struct mptcp_sock *msk,
+				      struct mptcp_addr_info *skc)
 {
 	struct mptcp_pm_addr_entry *entry;
 	bool backup;
@@ -XXX,XX +XXX,XX @@ bool mptcp_userspace_pm_is_backup(struct mptcp_sock *msk,
 	return backup;
 }
 
+bool mptcp_userspace_pm_is_backup(struct mptcp_sock *msk,
+				  struct mptcp_addr_info *skc)
+{
+	return userspace_pm_get_priority(msk, skc);
+}
+
 static struct mptcp_sock *mptcp_userspace_pm_get_sock(const struct genl_info *info)
 {
 	struct mptcp_sock *msk;
@@ -XXX,XX +XXX,XX @@ static struct mptcp_sock *mptcp_userspace_pm_get_sock(const struct genl_info *in
 	return msk;
 }
 
+static int userspace_pm_address_announced(struct mptcp_sock *msk,
+					  struct mptcp_pm_addr_entry *local)
+{
+	int err;
+
+	err = mptcp_userspace_pm_append_new_local_addr(msk, local, false);
+	if (err < 0)
+		return err;
+
+	spin_lock_bh(&msk->pm.lock);
+
+	if (mptcp_pm_alloc_anno_list(msk, &local->addr)) {
+		msk->pm.add_addr_signaled++;
+		mptcp_pm_announce_addr(msk, &local->addr, false);
+		mptcp_pm_nl_addr_send_ack(msk);
+	}
+
+	spin_unlock_bh(&msk->pm.lock);
+
+	return 0;
+}
+
 int mptcp_pm_nl_announce_doit(struct sk_buff *skb, struct genl_info *info)
 {
 	struct mptcp_pm_addr_entry addr_val;
@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_announce_doit(struct sk_buff *skb, struct genl_info *info)
 		goto announce_err;
 	}
 
-	err = mptcp_userspace_pm_append_new_local_addr(msk, &addr_val, false);
-	if (err < 0) {
-		NL_SET_ERR_MSG_ATTR(info->extack, addr,
-				    "did not match address and id");
-		goto announce_err;
-	}
-
 	lock_sock(sk);
-	spin_lock_bh(&msk->pm.lock);
-
-	if (mptcp_pm_alloc_anno_list(msk, &addr_val.addr)) {
-		msk->pm.add_addr_signaled++;
-		mptcp_pm_announce_addr(msk, &addr_val.addr, false);
-		mptcp_pm_nl_addr_send_ack(msk);
-	}
-
-	spin_unlock_bh(&msk->pm.lock);
+	err = userspace_pm_address_announced(msk, &addr_val);
 	release_sock(sk);
+	if (err)
+		NL_SET_ERR_MSG_ATTR(info->extack, addr,
+				    "did not match address and id");
 
-	err = 0;
  announce_err:
 	sock_put(sk);
 	return err;
@@ -XXX,XX +XXX,XX @@ static int mptcp_userspace_pm_remove_id_zero_address(struct mptcp_sock *msk)
 {
 	struct mptcp_rm_list list = { .nr = 0 };
 	struct mptcp_subflow_context *subflow;
-	struct sock *sk = (struct sock *)msk;
 	bool has_id_0 = false;
-	int err = -EINVAL;
 
-	lock_sock(sk);
 	mptcp_for_each_subflow(msk, subflow) {
 		if (READ_ONCE(subflow->local_id) == 0) {
 			has_id_0 = true;
@@ -XXX,XX +XXX,XX @@ static int mptcp_userspace_pm_remove_id_zero_address(struct mptcp_sock *msk)
 		}
 	}
 	if (!has_id_0)
-		goto remove_err;
+		return -EINVAL;
 
 	list.ids[list.nr++] = 0;
 
@@ -XXX,XX +XXX,XX @@ static int mptcp_userspace_pm_remove_id_zero_address(struct mptcp_sock *msk)
 	mptcp_pm_remove_addr(msk, &list);
 	spin_unlock_bh(&msk->pm.lock);
 
-	err = 0;
-
-remove_err:
-	release_sock(sk);
-	return err;
+	return 0;
 }
 
 void mptcp_pm_remove_addr_entry(struct mptcp_sock *msk,
@@ -XXX,XX +XXX,XX @@ void mptcp_pm_remove_addr_entry(struct mptcp_sock *msk,
 	spin_unlock_bh(&msk->pm.lock);
 }
 
+static int userspace_pm_address_removed(struct mptcp_sock *msk, u8 id)
+{
+	struct mptcp_pm_addr_entry *entry;
+
+	if (id == 0)
+		return mptcp_userspace_pm_remove_id_zero_address(msk);
+
+	spin_lock_bh(&msk->pm.lock);
+	entry = mptcp_userspace_pm_lookup_addr_by_id(msk, id);
+	if (!entry) {
+		spin_unlock_bh(&msk->pm.lock);
+		return -EINVAL;
+	}
+
+	list_del_rcu(&entry->list);
+	spin_unlock_bh(&msk->pm.lock);
+
+	mptcp_pm_remove_addr_entry(msk, entry);
+
+	sock_kfree_s((struct sock *)msk, entry, sizeof(*entry));
+
+	return 0;
+}
+
 int mptcp_pm_nl_remove_doit(struct sk_buff *skb, struct genl_info *info)
 {
-	struct mptcp_pm_addr_entry *match;
 	struct mptcp_sock *msk;
 	struct nlattr *id;
 	int err = -EINVAL;
@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_remove_doit(struct sk_buff *skb, struct genl_info *info)
 
 	sk = (struct sock *)msk;
 
-	if (id_val == 0) {
-		err = mptcp_userspace_pm_remove_id_zero_address(msk);
-		goto out;
-	}
-
 	lock_sock(sk);
-
-	spin_lock_bh(&msk->pm.lock);
-	match = mptcp_userspace_pm_lookup_addr_by_id(msk, id_val);
-	if (!match) {
-		spin_unlock_bh(&msk->pm.lock);
-		release_sock(sk);
-		goto out;
-	}
-
-	list_del_rcu(&match->list);
-	spin_unlock_bh(&msk->pm.lock);
-
-	mptcp_pm_remove_addr_entry(msk, match);
-
+	err = userspace_pm_address_removed(msk, id_val);
 	release_sock(sk);
-
-	sock_kfree_s(sk, match, sizeof(*match));
-
-	err = 0;
-out:
 	if (err)
 		NL_SET_ERR_MSG_ATTR_FMT(info->extack, id,
 					"address with id %u not found",
@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_remove_doit(struct sk_buff *skb, struct genl_info *info)
 	return err;
 }
 
+static int userspace_pm_subflow_established(struct mptcp_sock *msk,
+					    struct mptcp_pm_addr_entry *entry,
+					    struct mptcp_addr_info *remote)
+{
+	struct sock *sk = (struct sock *)msk;
+	struct mptcp_pm_local local;
+	int err;
+
+	err = mptcp_userspace_pm_append_new_local_addr(msk, entry, false);
+	if (err < 0)
+		return err;
+
+	local.addr = entry->addr;
+	local.flags = entry->flags;
+	local.ifindex = entry->ifindex;
+
+	err = __mptcp_subflow_connect(sk, &local, remote);
+	spin_lock_bh(&msk->pm.lock);
+	if (err)
+		mptcp_userspace_pm_delete_local_addr(msk, entry);
+	else
+		msk->pm.subflows++;
+	spin_unlock_bh(&msk->pm.lock);
+
+	return err;
+}
+
 int mptcp_pm_nl_subflow_create_doit(struct sk_buff *skb, struct genl_info *info)
 {
 	struct mptcp_pm_addr_entry entry = { 0 };
 	struct mptcp_addr_info addr_r;
 	struct nlattr *raddr, *laddr;
-	struct mptcp_pm_local local;
 	struct mptcp_sock *msk;
 	int err = -EINVAL;
 	struct sock *sk;
@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_subflow_create_doit(struct sk_buff *skb, struct genl_info *info)
 		goto create_err;
 	}
 
-	err = mptcp_userspace_pm_append_new_local_addr(msk, &entry, false);
-	if (err < 0) {
-		NL_SET_ERR_MSG_ATTR(info->extack, laddr,
-				    "did not match address and id");
-		goto create_err;
-	}
-
-	local.addr = entry.addr;
-	local.flags = entry.flags;
-	local.ifindex = entry.ifindex;
-
 	lock_sock(sk);
-	err = __mptcp_subflow_connect(sk, &local, &addr_r);
+	err = userspace_pm_subflow_established(msk, &entry, &addr_r);
 	release_sock(sk);
 
 	if (err)
 		GENL_SET_ERR_MSG_FMT(info, "connect error: %d", err);
 
-	spin_lock_bh(&msk->pm.lock);
-	if (err)
-		mptcp_userspace_pm_delete_local_addr(msk, &entry);
-	else
-		msk->pm.subflows++;
-	spin_unlock_bh(&msk->pm.lock);
-
  create_err:
 	sock_put(sk);
 	return err;
@@ -XXX,XX +XXX,XX @@ static struct sock *mptcp_nl_find_ssk(struct mptcp_sock *msk,
 	return NULL;
 }
 
+static int userspace_pm_subflow_closed(struct mptcp_sock *msk,
+				       struct mptcp_pm_addr_entry *local,
+				       struct mptcp_addr_info *remote)
+{
+	struct sock *ssk, *sk = (struct sock *)msk;
+
+	ssk = mptcp_nl_find_ssk(msk, &local->addr, remote);
+	if (!ssk)
+		return -ESRCH;
+
+	spin_lock_bh(&msk->pm.lock);
+	mptcp_userspace_pm_delete_local_addr(msk, local);
+	spin_unlock_bh(&msk->pm.lock);
+	mptcp_subflow_shutdown(sk, ssk, RCV_SHUTDOWN | SEND_SHUTDOWN);
+	mptcp_close_ssk(sk, ssk, mptcp_subflow_ctx(ssk));
+	MPTCP_INC_STATS(sock_net(sk), MPTCP_MIB_RMSUBFLOW);
+
+	return 0;
+}
+
 int mptcp_pm_nl_subflow_destroy_doit(struct sk_buff *skb, struct genl_info *info)
 {
 	struct mptcp_pm_addr_entry addr_l;
 	struct mptcp_addr_info addr_r;
 	struct nlattr *raddr, *laddr;
 	struct mptcp_sock *msk;
-	struct sock *sk, *ssk;
 	int err = -EINVAL;
+	struct sock *sk;
 
 	if (GENL_REQ_ATTR_CHECK(info, MPTCP_PM_ATTR_ADDR) ||
 	    GENL_REQ_ATTR_CHECK(info, MPTCP_PM_ATTR_ADDR_REMOTE))
@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_subflow_destroy_doit(struct sk_buff *skb, struct genl_info *info
 	}
 
 	lock_sock(sk);
-	ssk = mptcp_nl_find_ssk(msk, &addr_l.addr, &addr_r);
-	if (!ssk) {
-		GENL_SET_ERR_MSG(info, "subflow not found");
-		err = -ESRCH;
-		goto release_sock;
-	}
-
-	spin_lock_bh(&msk->pm.lock);
-	mptcp_userspace_pm_delete_local_addr(msk, &addr_l);
-	spin_unlock_bh(&msk->pm.lock);
-	mptcp_subflow_shutdown(sk, ssk, RCV_SHUTDOWN | SEND_SHUTDOWN);
-	mptcp_close_ssk(sk, ssk, mptcp_subflow_ctx(ssk));
-	MPTCP_INC_STATS(sock_net(sk), MPTCP_MIB_RMSUBFLOW);
-release_sock:
+	err = userspace_pm_subflow_closed(msk, &addr_l, &addr_r);
 	release_sock(sk);
+	if (err)
+		GENL_SET_ERR_MSG(info, "subflow not found");
 
 destroy_err:
 	sock_put(sk);
 	return err;
 }
 
+static int userspace_pm_set_priority(struct mptcp_sock *msk,
+				     struct mptcp_pm_addr_entry *local,
+				     struct mptcp_addr_info *remote)
+{
+	struct mptcp_pm_addr_entry *entry;
+	u8 bkup = 0;
+
+	if (local->flags & MPTCP_PM_ADDR_FLAG_BACKUP)
+		bkup = 1;
+
+	spin_lock_bh(&msk->pm.lock);
+	entry = mptcp_userspace_pm_lookup_addr(msk, &local->addr);
+	if (entry) {
+		if (bkup)
+			entry->flags |= MPTCP_PM_ADDR_FLAG_BACKUP;
+		else
+			entry->flags &= ~MPTCP_PM_ADDR_FLAG_BACKUP;
+	}
+	spin_unlock_bh(&msk->pm.lock);
+
+	return mptcp_pm_nl_mp_prio_send_ack(msk, &local->addr, remote, bkup);
+}
+
 int mptcp_userspace_pm_set_flags(struct mptcp_pm_addr_entry *local,
 				 struct genl_info *info)
 {
 	struct mptcp_addr_info rem = { .family = AF_UNSPEC, };
-	struct mptcp_pm_addr_entry *entry;
 	struct nlattr *attr, *attr_rem;
 	struct mptcp_sock *msk;
 	int ret = -EINVAL;
 	struct sock *sk;
-	u8 bkup = 0;
 
 	if (GENL_REQ_ATTR_CHECK(info, MPTCP_PM_ATTR_ADDR_REMOTE))
 		return ret;
@@ -XXX,XX +XXX,XX @@ int mptcp_userspace_pm_set_flags(struct mptcp_pm_addr_entry *local,
 		goto set_flags_err;
 	}
 
-	if (local->flags & MPTCP_PM_ADDR_FLAG_BACKUP)
-		bkup = 1;
-
-	spin_lock_bh(&msk->pm.lock);
-	entry = mptcp_userspace_pm_lookup_addr(msk, &local->addr);
-	if (entry) {
-		if (bkup)
-			entry->flags |= MPTCP_PM_ADDR_FLAG_BACKUP;
-		else
-			entry->flags &= ~MPTCP_PM_ADDR_FLAG_BACKUP;
-	}
-	spin_unlock_bh(&msk->pm.lock);
-
 	lock_sock(sk);
-	ret = mptcp_pm_nl_mp_prio_send_ack(msk, &local->addr, &rem, bkup);
+	ret = userspace_pm_set_priority(msk, local, &rem);
 	release_sock(sk);
 
 	/* mptcp_pm_nl_mp_prio_send_ack() only fails in one case */
@@ -XXX,XX +XXX,XX @@ int mptcp_userspace_pm_get_addr(u8 id, struct mptcp_pm_addr_entry *addr,
 	sock_put(sk);
 	return ret;
 }
+
+static struct mptcp_pm_ops mptcp_userspace_pm = {
+	.address_announced	= userspace_pm_address_announced,
+	.address_removed	= userspace_pm_address_removed,
+	.subflow_established	= userspace_pm_subflow_established,
+	.subflow_closed		= userspace_pm_subflow_closed,
+	.get_local_id		= userspace_pm_get_local_id,
+	.get_priority		= userspace_pm_get_priority,
+	.set_priority		= userspace_pm_set_priority,
+	.type			= MPTCP_PM_TYPE_USERSPACE,
+	.owner			= THIS_MODULE,
+};
+
+void __init mptcp_userspace_pm_init(void)
+{
+	mptcp_pm_register(&mptcp_userspace_pm);
+}
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -XXX,XX +XXX,XX @@ static inline u8 subflow_get_local_id(const struct mptcp_subflow_context *subflo
 }
 
 void __init mptcp_pm_nl_init(void);
+void __init mptcp_userspace_pm_init(void);
 void mptcp_pm_nl_work(struct mptcp_sock *msk);
 unsigned int mptcp_pm_get_add_addr_signal_max(const struct mptcp_sock *msk);
 unsigned int mptcp_pm_get_add_addr_accept_max(const struct mptcp_sock *msk);
-- 
2.43.0

From: Geliang Tang <tanggeliang@kylinos.cn>

Add a struct mptcp_pm_ops pointer "ops" in struct mptcp_pm_data, and two
functions mptcp_pm_initialize() and mptcp_pm_release(), to set and release
this pointer. mptcp_pm_initialize() is invoked in mptcp_pm_data_reset(),
while mptcp_pm_release() is invoked in __mptcp_destroy_sock().

In this way, different userspace path managers can be initialized through
the pm_type sysctl, and then called into their respective interfaces
through "ops" of "msk->pm".

Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 net/mptcp/pm.c           | 49 ++++++++++++++++++++++++++++++++++++++--
 net/mptcp/pm_netlink.c   | 11 ++++++++-
 net/mptcp/pm_userspace.c | 28 +++++++++++++++++------
 net/mptcp/protocol.c     | 10 +++++++-
 net/mptcp/protocol.h     |  3 +++
 5 files changed, 90 insertions(+), 11 deletions(-)

diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/pm.c
+++ b/net/mptcp/pm.c
@@ -XXX,XX +XXX,XX @@ void mptcp_pm_new_connection(struct mptcp_sock *msk, const struct sock *ssk, int
 
 	WRITE_ONCE(pm->server_side, server_side);
 	mptcp_event(MPTCP_EVENT_CREATED, msk, ssk, GFP_ATOMIC);
+
+	if (pm->ops && pm->ops->created)
+		pm->ops->created(msk);
 }
 
 bool mptcp_pm_allow_new_subflow(struct mptcp_sock *msk)
@@ -XXX,XX +XXX,XX @@ void mptcp_pm_fully_established(struct mptcp_sock *msk, const struct sock *ssk)
 	msk->pm.status |= BIT(MPTCP_PM_ALREADY_ESTABLISHED);
 	spin_unlock_bh(&pm->lock);
 
-	if (announce)
+	if (announce) {
 		mptcp_event(MPTCP_EVENT_ESTABLISHED, msk, ssk, GFP_ATOMIC);
+
+		if (pm->ops && pm->ops->established)
+			pm->ops->established(msk);
+	}
 }
 
 void mptcp_pm_connection_closed(struct mptcp_sock *msk)
 {
 	pr_debug("msk=%p\n", msk);
 
-	if (msk->token)
+	if (msk->token) {
 		mptcp_event(MPTCP_EVENT_CLOSED, msk, NULL, GFP_KERNEL);
+
+		if (msk->pm.ops && msk->pm.ops->closed)
+			msk->pm.ops->closed(msk);
+	}
 }
 
 void mptcp_pm_subflow_established(struct mptcp_sock *msk)
@@ -XXX,XX +XXX,XX @@ void mptcp_pm_data_reset(struct mptcp_sock *msk)
 		WRITE_ONCE(pm->work_pending, 0);
 		WRITE_ONCE(pm->accept_addr, 0);
 		WRITE_ONCE(pm->accept_subflow, 0);
+
+		rcu_read_lock();
+		mptcp_pm_initialize(msk, mptcp_pm_find(pm_type));
+		rcu_read_unlock();
 	}
 
 	WRITE_ONCE(pm->addr_signal, 0);
@@ -XXX,XX +XXX,XX @@ void mptcp_pm_unregister(struct mptcp_pm_ops *pm)
 	list_del_rcu(&pm->list);
 	spin_unlock(&mptcp_pm_list_lock);
 }
+
+int mptcp_pm_initialize(struct mptcp_sock *msk, struct mptcp_pm_ops *pm)
+{
+	if (!pm)
+		return -EINVAL;
+
+	if (!bpf_try_module_get(pm, pm->owner))
+		return -EBUSY;
+
+	msk->pm.ops = pm;
+	if (msk->pm.ops->init)
+		msk->pm.ops->init(msk);
+
+	pr_debug("userspace_pm type %u initialized\n", msk->pm.ops->type);
+	return 0;
+}
+
+void mptcp_pm_release(struct mptcp_sock *msk)
+{
+	struct mptcp_pm_ops *pm = msk->pm.ops;
+
+	if (!pm)
+		return;
+
+	msk->pm.ops = NULL;
+	if (pm->release)
+		pm->release(msk);
+
+	bpf_module_put(pm, pm->owner);
+}
diff --git a/net/mptcp/pm_netlink.c b/net/mptcp/pm_netlink.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/pm_netlink.c
+++ b/net/mptcp/pm_netlink.c
@@ -XXX,XX +XXX,XX @@ static int mptcp_pm_nl_create_listen_socket(struct sock *sk,
 	int addrlen = sizeof(struct sockaddr_in);
 	struct sockaddr_storage addr;
 	struct sock *newsk, *ssk;
+	struct mptcp_sock *msk;
 	int backlog = 1024;
 	int err;
 
@@ -XXX,XX +XXX,XX @@ static int mptcp_pm_nl_create_listen_socket(struct sock *sk,
 				      is_ipv6 ? "msk_lock-AF_INET6" : "msk_lock-AF_INET",
 				      &mptcp_keys[is_ipv6]);
 
+	msk = mptcp_sk(newsk);
 	lock_sock(newsk);
-	ssk = __mptcp_nmpc_sk(mptcp_sk(newsk));
+	ssk = __mptcp_nmpc_sk(msk);
 	release_sock(newsk);
 	if (IS_ERR(ssk))
 		return PTR_ERR(ssk);
@@ -XXX,XX +XXX,XX @@ static int mptcp_pm_nl_create_listen_socket(struct sock *sk,
 	if (!err)
 		mptcp_event_pm_listener(ssk, MPTCP_EVENT_LISTENER_CREATED);
 	release_sock(ssk);
+
+	if (!err) {
+		lock_sock(newsk);
+		if (msk->pm.ops && msk->pm.ops->listener_created)
+			msk->pm.ops->listener_created(msk);
+		release_sock(newsk);
+	}
 	return err;
 }
 
diff --git a/net/mptcp/pm_userspace.c b/net/mptcp/pm_userspace.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/pm_userspace.c
+++ b/net/mptcp/pm_userspace.c
@@ -XXX,XX +XXX,XX @@ int mptcp_userspace_pm_get_local_id(struct mptcp_sock *msk,
 	if (new_entry.addr.port == msk_sport)
 		new_entry.addr.port = 0;
 
-	return userspace_pm_get_local_id(msk, &new_entry);
+	return msk->pm.ops && msk->pm.ops->get_local_id ?
+	       msk->pm.ops->get_local_id(msk, &new_entry) :
+	       userspace_pm_get_local_id(msk, &new_entry);
 }
 
 static bool userspace_pm_get_priority(struct mptcp_sock *msk,
@@ -XXX,XX +XXX,XX @@ static bool userspace_pm_get_priority(struct mptcp_sock *msk,
 bool mptcp_userspace_pm_is_backup(struct mptcp_sock *msk,
 				  struct mptcp_addr_info *skc)
 {
-	return userspace_pm_get_priority(msk, skc);
+	return msk->pm.ops && msk->pm.ops->get_priority ?
+	       msk->pm.ops->get_priority(msk, skc) :
+	       userspace_pm_get_priority(msk, skc);
 }
 
 static struct mptcp_sock *mptcp_userspace_pm_get_sock(const struct genl_info *info)
@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_announce_doit(struct sk_buff *skb, struct genl_info *info)
 	}
 
 	lock_sock(sk);
-	err = userspace_pm_address_announced(msk, &addr_val);
+	err = msk->pm.ops && msk->pm.ops->address_announced ?
+	      msk->pm.ops->address_announced(msk, &addr_val) :
+	      userspace_pm_address_announced(msk, &addr_val);
 	release_sock(sk);
 	if (err)
 		NL_SET_ERR_MSG_ATTR(info->extack, addr,
@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_remove_doit(struct sk_buff *skb, struct genl_info *info)
 	sk = (struct sock *)msk;
 
 	lock_sock(sk);
-	err = userspace_pm_address_removed(msk, id_val);
+	err = msk->pm.ops && msk->pm.ops->address_removed ?
+	      msk->pm.ops->address_removed(msk, id_val) :
+	      userspace_pm_address_removed(msk, id_val);
 	release_sock(sk);
 	if (err)
 		NL_SET_ERR_MSG_ATTR_FMT(info->extack, id,
@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_subflow_create_doit(struct sk_buff *skb, struct genl_info *info)
 	}
 
 	lock_sock(sk);
-	err = userspace_pm_subflow_established(msk, &entry, &addr_r);
+	err = msk->pm.ops && msk->pm.ops->subflow_established ?
+	      msk->pm.ops->subflow_established(msk, &entry, &addr_r) :
+	      userspace_pm_subflow_established(msk, &entry, &addr_r);
 	release_sock(sk);
 
 	if (err)
@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_subflow_destroy_doit(struct sk_buff *skb, struct genl_info *info
 	}
 
 	lock_sock(sk);
-	err = userspace_pm_subflow_closed(msk, &addr_l, &addr_r);
+	err = msk->pm.ops && msk->pm.ops->subflow_closed ?
+	      msk->pm.ops->subflow_closed(msk, &addr_l, &addr_r) :
+	      userspace_pm_subflow_closed(msk, &addr_l, &addr_r);
 	release_sock(sk);
 	if (err)
 		GENL_SET_ERR_MSG(info, "subflow not found");
@@ -XXX,XX +XXX,XX @@ int mptcp_userspace_pm_set_flags(struct mptcp_pm_addr_entry *local,
 	}
 
 	lock_sock(sk);
-	ret = userspace_pm_set_priority(msk, local, &rem);
+	ret = msk->pm.ops && msk->pm.ops->set_priority ?
+	      msk->pm.ops->set_priority(msk, local, &rem) :
+	      userspace_pm_set_priority(msk, local, &rem);
 	release_sock(sk);
 
 	/* mptcp_pm_nl_mp_prio_send_ack() only fails in one case */
diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/protocol.c
+++ b/net/mptcp/protocol.c
@@ -XXX,XX +XXX,XX @@ static void __mptcp_destroy_sock(struct sock *sk)
 	sk_stop_timer(sk, &sk->sk_timer);
 	msk->pm.status = 0;
 	mptcp_release_sched(msk);
+	mptcp_pm_release(msk);
 
 	sk->sk_prot->destroy(sk);
 
@@ -XXX,XX +XXX,XX @@ static __poll_t mptcp_check_readable(struct sock *sk)
 
 static void mptcp_check_listen_stop(struct sock *sk)
 {
+	struct mptcp_sock *msk = mptcp_sk(sk);
 	struct sock *ssk;
 
 	if (inet_sk_state_load(sk) != TCP_LISTEN)
 		return;
 
 	sock_prot_inuse_add(sock_net(sk), sk->sk_prot, -1);
-	ssk = mptcp_sk(sk)->first;
+	ssk = msk->first;
 	if (WARN_ON_ONCE(!ssk || inet_sk_state_load(ssk) != TCP_LISTEN))
 		return;
 
@@ -XXX,XX +XXX,XX @@ static void mptcp_check_listen_stop(struct sock *sk)
 	inet_csk_listen_stop(ssk);
 	mptcp_event_pm_listener(ssk, MPTCP_EVENT_LISTENER_CLOSED);
 	release_sock(ssk);
+
+	if (msk->pm.ops && msk->pm.ops->listener_closed)
+		msk->pm.ops->listener_closed(msk);
 }
 
 bool __mptcp_close(struct sock *sk, long timeout)
@@ -XXX,XX +XXX,XX @@ static int mptcp_listen(struct socket *sock, int backlog)
 		sock_prot_inuse_add(sock_net(sk), sk->sk_prot, 1);
 		mptcp_copy_inaddrs(sk, ssk);
 		mptcp_event_pm_listener(ssk, MPTCP_EVENT_LISTENER_CREATED);
+
+		if (msk->pm.ops && msk->pm.ops->listener_created)
+			msk->pm.ops->listener_created(msk);
 	}
 
 unlock:
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -XXX,XX +XXX,XX @@ struct mptcp_pm_data {
 	struct mptcp_addr_info remote;
 	struct list_head anno_list;
 	struct list_head userspace_pm_local_addr_list;
+	struct mptcp_pm_ops *ops;
 
 	spinlock_t	lock;		/*protects the whole PM data */
 
@@ -XXX,XX +XXX,XX @@ struct mptcp_pm_ops *mptcp_pm_find(enum mptcp_pm_type type);
 int mptcp_pm_validate(struct mptcp_pm_ops *pm);
 int mptcp_pm_register(struct mptcp_pm_ops *pm);
 void mptcp_pm_unregister(struct mptcp_pm_ops *pm);
+int mptcp_pm_initialize(struct mptcp_sock *msk, struct mptcp_pm_ops *pm);
+void mptcp_pm_release(struct mptcp_sock *msk);
 
 void mptcp_free_local_addr_list(struct mptcp_sock *msk);
 
-- 
2.43.0

From: Geliang Tang <tanggeliang@kylinos.cn>

v6:
 - add "name" in struct mptcp_pm_ops.
 - add some "sysctl" patches.
 - drop "struct mptcp_pm_param".
 - drop "pm_type" in mptcp_pm_data.

v5:
 - use "struct mptcp_pm_param *param" as unified parameters for all
   interfaces.
 - register in-kernel mptcp_pm_ops too.
 - only implement two interfaces "get_local_id" and "get_priority" in
   this set.

v4:
 - include a new patch "define BPF path manager type".

- add new interfaces:
	created established closed
	listerner_created listener_closed

- rename interfaces as:
	address_announced address_removed
	subflow_established subflow_closed
	get_priority set_priority

- rename functions as:
	mptcp_pm_validate
	mptcp_pm_register
	mptcp_pm_unregister
	mptcp_pm_initialize
	mptcp_pm_release

v2:
 - update get_local_id interface in patch 2.

get_addr() and dump_addr() interfaces of BPF userspace pm are dropped
as Matt suggested.

In order to implement BPF path manager, it's necessary to unify the
interfaces of the path manager. This set contains some cleanups and
refactoring to unify the interfaces in kernel space. Finally, define
a struct mptcp_pm_ops for a path manager.

Geliang Tang (13):
  mptcp: sysctl: use index for sysctl table
  mptcp: sysctl: add path_manager to set pm name
  mptcp: pm: use addr entry for get_local_id
  mptcp: pm: define struct mptcp_pm_ops
  mptcp: pm: in-kernel: register mptcp_in_kernel_pm
  mptcp: pm: userspace: register mptcp_userspace_pm
  mptcp: pm: initialize and release mptcp_pm_ops
  mptcp: pm: drop pm_type in mptcp_pm_data
  mptcp: sysctl: drop old pm_type sysctl
  mptcp: sysctl: add available_path_managers
  mptcp: pm: drop get_local_id helpers
  mptcp: pm: drop is_backup helpers
  selftests/mptcp: use path_manager instead of pm_type

-- 
2.43.0

From: Geliang Tang <tanggeliang@kylinos.cn>

The sysctl table needs to be updated when a new sysctl is added or an old
sysctl is deleted. It's inconvenient to use fixed index numbers, since the
index number that needs to be changed and all subsequent index numbers need
to be updated accordingly.

This patch adds an "i" variable to access each sysctl in the sysctl table
incrementally.

Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 net/mptcp/ctrl.c | 23 ++++++++++++-----------
 1 file changed, 12 insertions(+), 11 deletions(-)

diff --git a/net/mptcp/ctrl.c b/net/mptcp/ctrl.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/ctrl.c
+++ b/net/mptcp/ctrl.c
@@ -XXX,XX +XXX,XX @@ static int mptcp_pernet_new_table(struct net *net, struct mptcp_pernet *pernet)
 {
 	struct ctl_table_header *hdr;
 	struct ctl_table *table;
+	int i = 0;
 
 	table = mptcp_sysctl_table;
 	if (!net_eq(net, &init_net)) {
@@ -XXX,XX +XXX,XX @@ static int mptcp_pernet_new_table(struct net *net, struct mptcp_pernet *pernet)
 			goto err_alloc;
 	}
 
-	table[0].data = &pernet->mptcp_enabled;
-	table[1].data = &pernet->add_addr_timeout;
-	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;
-	table[6].data = &pernet->scheduler;
-	/* table[7] is for available_schedulers which is read-only info */
-	table[8].data = &pernet->close_timeout;
-	table[9].data = &pernet->blackhole_timeout;
-	table[10].data = &pernet->syn_retrans_before_tcp_fallback;
+	table[i++].data = &pernet->mptcp_enabled;
+	table[i++].data = &pernet->add_addr_timeout;
+	table[i++].data = &pernet->checksum_enabled;
+	table[i++].data = &pernet->allow_join_initial_addr_port;
+	table[i++].data = &pernet->stale_loss_cnt;
+	table[i++].data = &pernet->pm_type;
+	table[i++].data = &pernet->scheduler;
+	i++; /* table[i] is for available_schedulers which is read-only info */
+	table[i++].data = &pernet->close_timeout;
+	table[i++].data = &pernet->blackhole_timeout;
+	table[i++].data = &pernet->syn_retrans_before_tcp_fallback;
 
 	hdr = register_net_sysctl_sz(net, MPTCP_SYSCTL_PATH, table,
 				     ARRAY_SIZE(mptcp_sysctl_table));
-- 
2.43.0

From: Geliang Tang <tanggeliang@kylinos.cn>

A new net.mptcp.path_manager sysctl is added to determine which path
manager will be used by each newly-created MPTCP socket by setting the
name of it. This sysctl will replace the old one pm_type.

Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 Documentation/networking/mptcp-sysctl.rst | 17 ++++++++
 net/mptcp/ctrl.c                          | 50 +++++++++++++++++++++++
 net/mptcp/protocol.h                      |  1 +
 3 files changed, 68 insertions(+)

diff --git a/Documentation/networking/mptcp-sysctl.rst b/Documentation/networking/mptcp-sysctl.rst
index XXXXXXX..XXXXXXX 100644
--- a/Documentation/networking/mptcp-sysctl.rst
+++ b/Documentation/networking/mptcp-sysctl.rst
@@ -XXX,XX +XXX,XX @@ pm_type - INTEGER
 
 	Default: 0
 
+path_manager - STRING
+	Set the default path manager name 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.
+
+	* "in-kernel" - In-kernel path manager
+	* "userspace" - Userspace path manager
+
+	Default: "in-kernel"
+
 scheduler - STRING
 	Select the scheduler of your choice.
 
diff --git a/net/mptcp/ctrl.c b/net/mptcp/ctrl.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/ctrl.c
+++ b/net/mptcp/ctrl.c
@@ -XXX,XX +XXX,XX @@ struct mptcp_pernet {
 	u8 checksum_enabled;
 	u8 allow_join_initial_addr_port;
 	u8 pm_type;
+	char path_manager[MPTCP_PM_NAME_MAX];
 	char scheduler[MPTCP_SCHED_NAME_MAX];
 };
 
@@ -XXX,XX +XXX,XX @@ int mptcp_get_pm_type(const struct net *net)
 	return mptcp_get_pernet(net)->pm_type;
 }
 
+const char *mptcp_get_path_manager(const struct net *net)
+{
+	return mptcp_get_pernet(net)->path_manager;
+}
+
 const char *mptcp_get_scheduler(const struct net *net)
 {
 	return mptcp_get_pernet(net)->scheduler;
@@ -XXX,XX +XXX,XX @@ static void mptcp_pernet_set_defaults(struct mptcp_pernet *pernet)
 	pernet->allow_join_initial_addr_port = 1;
 	pernet->stale_loss_cnt = 4;
 	pernet->pm_type = MPTCP_PM_TYPE_KERNEL;
+	strscpy(pernet->path_manager, "in-kernel", sizeof(pernet->path_manager));
 	strscpy(pernet->scheduler, "default", sizeof(pernet->scheduler));
 }
 
 #ifdef CONFIG_SYSCTL
+static int mptcp_set_path_manager(char *path_manager, const char *name)
+{
+	struct mptcp_pm_ops *pm;
+	int ret = 0;
+
+	rcu_read_lock();
+	pm = mptcp_pm_find(name);
+	if (pm)
+		strscpy(path_manager, name, MPTCP_PM_NAME_MAX);
+	else
+		ret = -ENOENT;
+	rcu_read_unlock();
+
+	return ret;
+}
+
+static int proc_path_manager(const struct ctl_table *ctl, int write,
+			     void *buffer, size_t *lenp, loff_t *ppos)
+{
+	char (*path_manager)[MPTCP_PM_NAME_MAX] = ctl->data;
+	char val[MPTCP_PM_NAME_MAX];
+	const struct ctl_table tbl = {
+		.data = val,
+		.maxlen = MPTCP_PM_NAME_MAX,
+	};
+	int ret;
+
+	strscpy(val, *path_manager, MPTCP_PM_NAME_MAX);
+
+	ret = proc_dostring(&tbl, write, buffer, lenp, ppos);
+	if (write && ret == 0)
+		ret = mptcp_set_path_manager(*path_manager, val);
+
+	return ret;
+}
+
 static int mptcp_set_scheduler(char *scheduler, const char *name)
 {
 	struct mptcp_sched_ops *sched;
@@ -XXX,XX +XXX,XX @@ static struct ctl_table mptcp_sysctl_table[] = {
 		.extra1       = SYSCTL_ZERO,
 		.extra2       = &mptcp_pm_type_max
 	},
+	{
+		.procname = "path_manager",
+		.maxlen	= MPTCP_PM_NAME_MAX,
+		.mode = 0644,
+		.proc_handler = proc_path_manager,
+	},
 	{
 		.procname = "scheduler",
 		.maxlen	= MPTCP_SCHED_NAME_MAX,
@@ -XXX,XX +XXX,XX @@ static int mptcp_pernet_new_table(struct net *net, struct mptcp_pernet *pernet)
 	table[i++].data = &pernet->allow_join_initial_addr_port;
 	table[i++].data = &pernet->stale_loss_cnt;
 	table[i++].data = &pernet->pm_type;
+	table[i++].data = &pernet->path_manager;
 	table[i++].data = &pernet->scheduler;
 	i++; /* table[i] is for available_schedulers which is read-only info */
 	table[i++].data = &pernet->close_timeout;
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -XXX,XX +XXX,XX @@ int mptcp_allow_join_id0(const struct net *net);
 unsigned int mptcp_stale_loss_cnt(const struct net *net);
 unsigned int mptcp_close_timeout(const struct sock *sk);
 int mptcp_get_pm_type(const struct net *net);
+const char *mptcp_get_path_manager(const struct net *net);
 const char *mptcp_get_scheduler(const struct net *net);
 
 void mptcp_active_disable(struct sock *sk);
-- 
2.43.0

From: Geliang Tang <tanggeliang@kylinos.cn>

The following code in mptcp_userspace_pm_get_local_id() that assigns "skc"
to "new_entry" is not allowed in BPF if we use the same code to implement
the get_local_id() interface of a BFP path manager:

memset(&new_entry, 0, sizeof(struct mptcp_pm_addr_entry));
	new_entry.addr = *skc;
	new_entry.addr.id = 0;
	new_entry.flags = MPTCP_PM_ADDR_FLAG_IMPLICIT;

To solve the issue, this patch moves this assignment to "new_entry" forward
to mptcp_pm_get_local_id(), and then passing "new_entry" as a parameter to
both mptcp_pm_nl_get_local_id() and mptcp_userspace_pm_get_local_id().

Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 net/mptcp/pm.c           |  9 ++++++---
 net/mptcp/pm_netlink.c   | 11 ++++-------
 net/mptcp/pm_userspace.c | 17 ++++++-----------
 net/mptcp/protocol.h     |  6 ++++--
 4 files changed, 20 insertions(+), 23 deletions(-)

diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/pm.c
+++ b/net/mptcp/pm.c
@@ -XXX,XX +XXX,XX @@ 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)
 {
-	struct mptcp_addr_info skc_local;
+	struct mptcp_pm_addr_entry skc_local = { 0 };
 	struct mptcp_addr_info msk_local;
 
 	if (WARN_ON_ONCE(!msk))
@@ -XXX,XX +XXX,XX @@ int mptcp_pm_get_local_id(struct mptcp_sock *msk, struct sock_common *skc)
 	 * addr
 	 */
 	mptcp_local_address((struct sock_common *)msk, &msk_local);
-	mptcp_local_address((struct sock_common *)skc, &skc_local);
-	if (mptcp_addresses_equal(&msk_local, &skc_local, false))
+	mptcp_local_address((struct sock_common *)skc, &skc_local.addr);
+	if (mptcp_addresses_equal(&msk_local, &skc_local.addr, false))
 		return 0;
 
+	skc_local.addr.id = 0;
+	skc_local.flags = MPTCP_PM_ADDR_FLAG_IMPLICIT;
+
 	if (mptcp_pm_is_userspace(msk))
 		return mptcp_userspace_pm_get_local_id(msk, &skc_local);
 	return mptcp_pm_nl_get_local_id(msk, &skc_local);
diff --git a/net/mptcp/pm_netlink.c b/net/mptcp/pm_netlink.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/pm_netlink.c
+++ b/net/mptcp/pm_netlink.c
@@ -XXX,XX +XXX,XX @@ static int mptcp_pm_nl_create_listen_socket(struct sock *sk,
 	return err;
 }
 
-int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk, struct mptcp_addr_info *skc)
+int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk,
+			     struct mptcp_pm_addr_entry *skc)
 {
 	struct mptcp_pm_addr_entry *entry;
 	struct pm_nl_pernet *pernet;
@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk, struct mptcp_addr_info *skc
 	pernet = pm_nl_get_pernet_from_msk(msk);
 
 	rcu_read_lock();
-	entry = __lookup_addr(pernet, skc);
+	entry = __lookup_addr(pernet, &skc->addr);
 	ret = entry ? entry->addr.id : -1;
 	rcu_read_unlock();
 	if (ret >= 0)
@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk, struct mptcp_addr_info *skc
 	if (!entry)
 		return -ENOMEM;
 
-	entry->addr = *skc;
-	entry->addr.id = 0;
+	*entry = *skc;
 	entry->addr.port = 0;
-	entry->ifindex = 0;
-	entry->flags = MPTCP_PM_ADDR_FLAG_IMPLICIT;
-	entry->lsk = NULL;
 	ret = mptcp_pm_nl_append_new_local_addr(pernet, entry, true);
 	if (ret < 0)
 		kfree(entry);
diff --git a/net/mptcp/pm_userspace.c b/net/mptcp/pm_userspace.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/pm_userspace.c
+++ b/net/mptcp/pm_userspace.c
@@ -XXX,XX +XXX,XX @@ mptcp_userspace_pm_lookup_addr_by_id(struct mptcp_sock *msk, unsigned int id)
 }
 
 int mptcp_userspace_pm_get_local_id(struct mptcp_sock *msk,
-				    struct mptcp_addr_info *skc)
+				    struct mptcp_pm_addr_entry *skc)
 {
-	struct mptcp_pm_addr_entry *entry = NULL, new_entry;
 	__be16 msk_sport =  ((struct inet_sock *)
 			     inet_sk((struct sock *)msk))->inet_sport;
+	struct mptcp_pm_addr_entry *entry;
 
 	spin_lock_bh(&msk->pm.lock);
-	entry = mptcp_userspace_pm_lookup_addr(msk, skc);
+	entry = mptcp_userspace_pm_lookup_addr(msk, &skc->addr);
 	spin_unlock_bh(&msk->pm.lock);
 	if (entry)
 		return entry->addr.id;
 
-	memset(&new_entry, 0, sizeof(struct mptcp_pm_addr_entry));
-	new_entry.addr = *skc;
-	new_entry.addr.id = 0;
-	new_entry.flags = MPTCP_PM_ADDR_FLAG_IMPLICIT;
-
-	if (new_entry.addr.port == msk_sport)
-		new_entry.addr.port = 0;
+	if (skc->addr.port == msk_sport)
+		skc->addr.port = 0;
 
-	return mptcp_userspace_pm_append_new_local_addr(msk, &new_entry, true);
+	return mptcp_userspace_pm_append_new_local_addr(msk, skc, true);
 }
 
 bool mptcp_userspace_pm_is_backup(struct mptcp_sock *msk,
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -XXX,XX +XXX,XX @@ bool mptcp_pm_add_addr_signal(struct mptcp_sock *msk, const struct sk_buff *skb,
 bool mptcp_pm_rm_addr_signal(struct mptcp_sock *msk, unsigned int remaining,
 			     struct mptcp_rm_list *rm_list);
 int mptcp_pm_get_local_id(struct mptcp_sock *msk, struct sock_common *skc);
-int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk, struct mptcp_addr_info *skc);
-int mptcp_userspace_pm_get_local_id(struct mptcp_sock *msk, struct mptcp_addr_info *skc);
+int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk,
+			     struct mptcp_pm_addr_entry *skc);
+int mptcp_userspace_pm_get_local_id(struct mptcp_sock *msk,
+				    struct mptcp_pm_addr_entry *skc);
 bool mptcp_pm_is_backup(struct mptcp_sock *msk, struct sock_common *skc);
 bool mptcp_pm_nl_is_backup(struct mptcp_sock *msk, struct mptcp_addr_info *skc);
 bool mptcp_userspace_pm_is_backup(struct mptcp_sock *msk, struct mptcp_addr_info *skc);
-- 
2.43.0

From: Geliang Tang <tanggeliang@kylinos.cn>

In order to allow users to develop their own BPF-based path manager,
this patch defines a struct ops "mptcp_pm_ops" for a userspace path
manager, which contains a set of interfaces.

Add a set of functions to register, unregister, find and validate a
given struct ops.

Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 include/net/mptcp.h  | 37 +++++++++++++++++++++++++++++
 net/mptcp/pm.c       | 55 ++++++++++++++++++++++++++++++++++++++++++++
 net/mptcp/protocol.h |  5 ++++
 3 files changed, 97 insertions(+)

diff --git a/include/net/mptcp.h b/include/net/mptcp.h
index XXXXXXX..XXXXXXX 100644
--- a/include/net/mptcp.h
+++ b/include/net/mptcp.h
@@ -XXX,XX +XXX,XX @@
 
 struct mptcp_info;
 struct mptcp_sock;
+struct mptcp_pm_addr_entry;
 struct seq_file;
 
 /* MPTCP sk_buff extension data */
@@ -XXX,XX +XXX,XX @@ struct mptcp_sched_ops {
 	void (*release)(struct mptcp_sock *msk);
 } ____cacheline_aligned_in_smp;
 
+#define MPTCP_PM_NAME_MAX	16
+
+struct mptcp_pm_ops {
+	int (*created)(struct mptcp_sock *msk);
+	int (*established)(struct mptcp_sock *msk);
+	int (*closed)(struct mptcp_sock *msk);
+	int (*get_local_id)(struct mptcp_sock *msk,
+			    struct mptcp_pm_addr_entry *skc);
+	bool (*get_priority)(struct mptcp_sock *msk,
+			     struct mptcp_addr_info *skc);
+	int (*address_announced)(struct mptcp_sock *msk,
+				 struct mptcp_pm_addr_entry *local);
+	int (*address_removed)(struct mptcp_sock *msk,
+			       struct mptcp_pm_addr_entry *local);
+	int (*subflow_established)(struct mptcp_sock *msk,
+				   struct mptcp_pm_addr_entry *local,
+				   struct mptcp_addr_info *remote);
+	int (*subflow_closed)(struct mptcp_sock *msk,
+			      struct mptcp_pm_addr_entry *local,
+			      struct mptcp_addr_info *remote);
+	int (*set_priority)(struct mptcp_sock *msk,
+			    struct mptcp_pm_addr_entry *local,
+			    struct mptcp_addr_info *remote,
+			    u8 changed);
+	int (*listener_created)(struct mptcp_sock *msk);
+	int (*listener_closed)(struct mptcp_sock *msk);
+
+	u8			type;
+	char			name[MPTCP_PM_NAME_MAX];
+	struct module		*owner;
+	struct list_head	list;
+
+	void (*init)(struct mptcp_sock *msk);
+	void (*release)(struct mptcp_sock *msk);
+} ____cacheline_aligned_in_smp;
+
 #ifdef CONFIG_MPTCP
 void mptcp_init(void);
 
diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/pm.c
+++ b/net/mptcp/pm.c
@@ -XXX,XX +XXX,XX @@
 #define pr_fmt(fmt) "MPTCP: " fmt
 
 #include <linux/kernel.h>
+#include <linux/rculist.h>
+#include <linux/spinlock.h>
 #include <net/mptcp.h>
 #include "protocol.h"
 
 #include "mib.h"
 #include "mptcp_pm_gen.h"
 
+static DEFINE_SPINLOCK(mptcp_pm_list_lock);
+static LIST_HEAD(mptcp_pm_list);
+
 /* path manager command handlers */
 
 int mptcp_pm_announce_addr(struct mptcp_sock *msk,
@@ -XXX,XX +XXX,XX @@ void __init mptcp_pm_init(void)
 {
 	mptcp_pm_nl_init();
 }
+
+/* Must be called with rcu read lock held */
+struct mptcp_pm_ops *mptcp_pm_find(const char *name)
+{
+	struct mptcp_pm_ops *pm;
+
+	list_for_each_entry_rcu(pm, &mptcp_pm_list, list) {
+		if (!strcmp(pm->name, name))
+			return pm;
+	}
+
+	return NULL;
+}
+
+int mptcp_pm_validate(struct mptcp_pm_ops *pm)
+{
+	if (!pm->get_local_id || !pm->get_priority) {
+		pr_err("%u does not implement required ops\n", pm->type);
+		return -EINVAL;
+	}
+
+	return 0;
+}
+
+int mptcp_pm_register(struct mptcp_pm_ops *pm)
+{
+	int ret;
+
+	ret = mptcp_pm_validate(pm);
+	if (ret)
+		return ret;
+
+	spin_lock(&mptcp_pm_list_lock);
+	if (mptcp_pm_find(pm->name)) {
+		spin_unlock(&mptcp_pm_list_lock);
+		return -EEXIST;
+	}
+	list_add_tail_rcu(&pm->list, &mptcp_pm_list);
+	spin_unlock(&mptcp_pm_list_lock);
+
+	pr_debug("%s registered\n", pm->name);
+	return 0;
+}
+
+void mptcp_pm_unregister(struct mptcp_pm_ops *pm)
+{
+	spin_lock(&mptcp_pm_list_lock);
+	list_del_rcu(&pm->list);
+	spin_unlock(&mptcp_pm_list_lock);
+}
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -XXX,XX +XXX,XX @@ int mptcp_pm_remove_addr(struct mptcp_sock *msk, const struct mptcp_rm_list *rm_
 void mptcp_pm_remove_addr_entry(struct mptcp_sock *msk,
 				struct mptcp_pm_addr_entry *entry);
 
+struct mptcp_pm_ops *mptcp_pm_find(const char *name);
+int mptcp_pm_validate(struct mptcp_pm_ops *pm);
+int mptcp_pm_register(struct mptcp_pm_ops *pm);
+void mptcp_pm_unregister(struct mptcp_pm_ops *pm);
+
 void mptcp_free_local_addr_list(struct mptcp_sock *msk);
 
 void mptcp_event(enum mptcp_event_type type, const struct mptcp_sock *msk,
-- 
2.43.0

From: Geliang Tang <tanggeliang@kylinos.cn>

This patch defines the original in-kernel netlink path manager as a new
struct mptcp_pm_ops named "mptcp_in_kernel_pm", and register it in
mptcp_pm_nl_init().

This mptcp_pm_ops will be skipped in mptcp_pm_unregister().

Only get_local_id() and get_priority() interfaces are implemented here.
mptcp_pm_nl_is_backup() becomes a wrapper of get_priority().

Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 net/mptcp/pm.c         |  3 +++
 net/mptcp/pm_netlink.c | 18 +++++++++++++++++-
 net/mptcp/protocol.h   |  2 ++
 3 files changed, 22 insertions(+), 1 deletion(-)

diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/pm.c
+++ b/net/mptcp/pm.c
@@ -XXX,XX +XXX,XX @@ int mptcp_pm_register(struct mptcp_pm_ops *pm)
 
 void mptcp_pm_unregister(struct mptcp_pm_ops *pm)
 {
+	if (pm == &mptcp_in_kernel_pm)
+		return;
+
 	spin_lock(&mptcp_pm_list_lock);
 	list_del_rcu(&pm->list);
 	spin_unlock(&mptcp_pm_list_lock);
diff --git a/net/mptcp/pm_netlink.c b/net/mptcp/pm_netlink.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/pm_netlink.c
+++ b/net/mptcp/pm_netlink.c
@@ -XXX,XX +XXX,XX @@ int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk,
 	return ret;
 }
 
-bool mptcp_pm_nl_is_backup(struct mptcp_sock *msk, struct mptcp_addr_info *skc)
+static bool mptcp_pm_nl_get_priority(struct mptcp_sock *msk,
+				     struct mptcp_addr_info *skc)
 {
 	struct pm_nl_pernet *pernet = pm_nl_get_pernet_from_msk(msk);
 	struct mptcp_pm_addr_entry *entry;
@@ -XXX,XX +XXX,XX @@ bool mptcp_pm_nl_is_backup(struct mptcp_sock *msk, struct mptcp_addr_info *skc)
 	return backup;
 }
 
+bool mptcp_pm_nl_is_backup(struct mptcp_sock *msk, struct mptcp_addr_info *skc)
+{
+	return mptcp_pm_nl_get_priority(msk, skc);
+}
+
 #define MPTCP_PM_CMD_GRP_OFFSET       0
 #define MPTCP_PM_EV_GRP_OFFSET        1
 
@@ -XXX,XX +XXX,XX @@ static struct pernet_operations mptcp_pm_pernet_ops = {
 	.size = sizeof(struct pm_nl_pernet),
 };
 
+struct mptcp_pm_ops mptcp_in_kernel_pm = {
+	.get_local_id		= mptcp_pm_nl_get_local_id,
+	.get_priority		= mptcp_pm_nl_get_priority,
+	.type			= MPTCP_PM_TYPE_KERNEL,
+	.name			= "in-kernel",
+	.owner			= THIS_MODULE,
+};
+
 void __init mptcp_pm_nl_init(void)
 {
 	if (register_pernet_subsys(&mptcp_pm_pernet_ops) < 0)
@@ -XXX,XX +XXX,XX @@ void __init mptcp_pm_nl_init(void)
 
 	if (genl_register_family(&mptcp_genl_family))
 		panic("Failed to register MPTCP PM netlink family\n");
+
+	mptcp_pm_register(&mptcp_in_kernel_pm);
 }
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -XXX,XX +XXX,XX @@ int mptcp_pm_remove_addr(struct mptcp_sock *msk, const struct mptcp_rm_list *rm_
 void mptcp_pm_remove_addr_entry(struct mptcp_sock *msk,
 				struct mptcp_pm_addr_entry *entry);
 
+extern struct mptcp_pm_ops mptcp_in_kernel_pm;
+
 struct mptcp_pm_ops *mptcp_pm_find(const char *name);
 int mptcp_pm_validate(struct mptcp_pm_ops *pm);
 int mptcp_pm_register(struct mptcp_pm_ops *pm);
-- 
2.43.0

From: Geliang Tang <tanggeliang@kylinos.cn>

This patch defines the original userspace path manager as a new struct
mptcp_pm_ops named "mptcp_userspace_pm", and register it in
mptcp_pm_data_init().

Only get_local_id() and get_priority() interfaces are implemented here.
mptcp_userspace_pm_is_backup() becomes a wrapper of get_priority().

Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 net/mptcp/pm.c           |  1 +
 net/mptcp/pm_userspace.c | 23 +++++++++++++++++++++--
 net/mptcp/protocol.h     |  1 +
 3 files changed, 23 insertions(+), 2 deletions(-)

From: Geliang Tang <tanggeliang@kylinos.cn>

Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 net/mptcp/pm.c       | 38 ++++++++++++++++++++++++++++++++++++++
 net/mptcp/protocol.c |  1 +
 net/mptcp/protocol.h |  3 +++
 3 files changed, 42 insertions(+)

diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/pm.c
+++ b/net/mptcp/pm.c
@@ -XXX,XX +XXX,XX @@ bool mptcp_pm_addr_families_match(const struct sock *sk,
 
 void mptcp_pm_data_reset(struct mptcp_sock *msk)
 {
+	const char *path_manager = mptcp_get_path_manager(sock_net((struct sock *)msk));
 	u8 pm_type = mptcp_get_pm_type(sock_net((struct sock *)msk));
 	struct mptcp_pm_data *pm = &msk->pm;
+	int ret;
 
 	pm->add_addr_signaled = 0;
 	pm->add_addr_accepted = 0;
@@ -XXX,XX +XXX,XX @@ void mptcp_pm_data_reset(struct mptcp_sock *msk)
 	pm->rm_list_rx.nr = 0;
 	WRITE_ONCE(pm->pm_type, pm_type);
 
+	rcu_read_lock();
+	ret = mptcp_pm_initialize(msk, mptcp_pm_find(path_manager));
+	rcu_read_unlock();
+	if (ret)
+		return;
+
 	if (pm_type == MPTCP_PM_TYPE_KERNEL) {
 		bool subflows_allowed = !!mptcp_pm_get_subflows_max(msk);
 
@@ -XXX,XX +XXX,XX @@ void mptcp_pm_unregister(struct mptcp_pm_ops *pm)
 	list_del_rcu(&pm->list);
 	spin_unlock(&mptcp_pm_list_lock);
 }
+
+int mptcp_pm_initialize(struct mptcp_sock *msk, struct mptcp_pm_ops *pm)
+{
+	if (!pm)
+		pm = &mptcp_in_kernel_pm;
+
+	if (!bpf_try_module_get(pm, pm->owner))
+		return -EBUSY;
+
+	msk->pm.ops = pm;
+	if (msk->pm.ops->init)
+		msk->pm.ops->init(msk);
+
+	pr_debug("pm %s initialized\n", pm->name);
+	return 0;
+}
+
+void mptcp_pm_release(struct mptcp_sock *msk)
+{
+	struct mptcp_pm_ops *pm = msk->pm.ops;
+
+	if (!pm)
+		return;
+
+	msk->pm.ops = NULL;
+	if (pm->release)
+		pm->release(msk);
+
+	bpf_module_put(pm, pm->owner);
+}
diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/protocol.c
+++ b/net/mptcp/protocol.c
@@ -XXX,XX +XXX,XX @@ void mptcp_destroy_common(struct mptcp_sock *msk, unsigned int flags)
 	mptcp_token_destroy(msk);
 	mptcp_pm_free_anno_list(msk);
 	mptcp_free_local_addr_list(msk);
+	mptcp_pm_release(msk);
 }
 
 static void mptcp_destroy(struct sock *sk)
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -XXX,XX +XXX,XX @@ struct mptcp_pm_data {
 	struct mptcp_addr_info remote;
 	struct list_head anno_list;
 	struct list_head userspace_pm_local_addr_list;
+	struct mptcp_pm_ops *ops;
 
 	spinlock_t	lock;		/*protects the whole PM data */
 
@@ -XXX,XX +XXX,XX @@ struct mptcp_pm_ops *mptcp_pm_find(const char *name);
 int mptcp_pm_validate(struct mptcp_pm_ops *pm);
 int mptcp_pm_register(struct mptcp_pm_ops *pm);
 void mptcp_pm_unregister(struct mptcp_pm_ops *pm);
+int mptcp_pm_initialize(struct mptcp_sock *msk, struct mptcp_pm_ops *pm);
+void mptcp_pm_release(struct mptcp_sock *msk);
 
 void mptcp_free_local_addr_list(struct mptcp_sock *msk);
 
-- 
2.43.0

From: Geliang Tang <tanggeliang@kylinos.cn>

Now pm->pm_type can be replaced by pm->ops->type, then "pm_type" filed
of struct mptcp_pm_data can be dropped.

Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 net/mptcp/pm.c       | 6 ++----
 net/mptcp/protocol.h | 5 ++---
 2 files changed, 4 insertions(+), 7 deletions(-)

diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/pm.c
+++ b/net/mptcp/pm.c
@@ -XXX,XX +XXX,XX @@ bool mptcp_pm_addr_families_match(const struct sock *sk,
 void mptcp_pm_data_reset(struct mptcp_sock *msk)
 {
 	const char *path_manager = mptcp_get_path_manager(sock_net((struct sock *)msk));
-	u8 pm_type = mptcp_get_pm_type(sock_net((struct sock *)msk));
 	struct mptcp_pm_data *pm = &msk->pm;
 	int ret;
 
@@ -XXX,XX +XXX,XX @@ 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, pm_type);
 
 	rcu_read_lock();
 	ret = mptcp_pm_initialize(msk, mptcp_pm_find(path_manager));
@@ -XXX,XX +XXX,XX @@ void mptcp_pm_data_reset(struct mptcp_sock *msk)
 	if (ret)
 		return;
 
-	if (pm_type == MPTCP_PM_TYPE_KERNEL) {
+	if (pm->ops->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
+		 * pm->ops->type is set to MPTCP_PM_TYPE_KERNEL
 		 */
 		WRITE_ONCE(pm->work_pending,
 			   (!!mptcp_pm_get_local_addr_max(msk) &&
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -XXX,XX +XXX,XX @@ 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);
@@ -XXX,XX +XXX,XX @@ static inline bool mptcp_pm_should_rm_signal(struct mptcp_sock *msk)
 
 static inline bool mptcp_pm_is_userspace(const struct mptcp_sock *msk)
 {
-	return READ_ONCE(msk->pm.pm_type) == MPTCP_PM_TYPE_USERSPACE;
+	return msk->pm.ops->type == MPTCP_PM_TYPE_USERSPACE;
 }
 
 static inline bool mptcp_pm_is_kernel(const struct mptcp_sock *msk)
 {
-	return READ_ONCE(msk->pm.pm_type) == MPTCP_PM_TYPE_KERNEL;
+	return msk->pm.ops->type == MPTCP_PM_TYPE_KERNEL;
 }
 
 static inline unsigned int mptcp_add_addr_len(int family, bool echo, bool port)
-- 
2.43.0

From: Geliang Tang <tanggeliang@kylinos.cn>

Old path manager sysctl "pm_type" can be replaced by the newly added
"path_manager". Drop it together with mptcp_get_pm_type() helper.

Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 Documentation/networking/mptcp-sysctl.rst | 17 -----------------
 net/mptcp/ctrl.c                          | 20 --------------------
 net/mptcp/protocol.h                      |  1 -
 3 files changed, 38 deletions(-)

diff --git a/Documentation/networking/mptcp-sysctl.rst b/Documentation/networking/mptcp-sysctl.rst
index XXXXXXX..XXXXXXX 100644
--- a/Documentation/networking/mptcp-sysctl.rst
+++ b/Documentation/networking/mptcp-sysctl.rst
@@ -XXX,XX +XXX,XX @@ enabled - BOOLEAN
 
 	Default: 1 (enabled)
 
-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
-
 path_manager - STRING
 	Set the default path manager name to use for each new MPTCP
 	socket. In-kernel path management will control subflow
diff --git a/net/mptcp/ctrl.c b/net/mptcp/ctrl.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/ctrl.c
+++ b/net/mptcp/ctrl.c
@@ -XXX,XX +XXX,XX @@
 
 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;
@@ -XXX,XX +XXX,XX @@ struct mptcp_pernet {
 	u8 mptcp_enabled;
 	u8 checksum_enabled;
 	u8 allow_join_initial_addr_port;
-	u8 pm_type;
 	char path_manager[MPTCP_PM_NAME_MAX];
 	char scheduler[MPTCP_SCHED_NAME_MAX];
 };
@@ -XXX,XX +XXX,XX @@ unsigned int mptcp_close_timeout(const struct sock *sk)
 	return mptcp_get_pernet(sock_net(sk))->close_timeout;
 }
 
-int mptcp_get_pm_type(const struct net *net)
-{
-	return mptcp_get_pernet(net)->pm_type;
-}
-
 const char *mptcp_get_path_manager(const struct net *net)
 {
 	return mptcp_get_pernet(net)->path_manager;
@@ -XXX,XX +XXX,XX @@ 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;
 	strscpy(pernet->path_manager, "in-kernel", sizeof(pernet->path_manager));
 	strscpy(pernet->scheduler, "default", sizeof(pernet->scheduler));
 }
@@ -XXX,XX +XXX,XX @@ 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
-	},
 	{
 		.procname = "path_manager",
 		.maxlen	= MPTCP_PM_NAME_MAX,
@@ -XXX,XX +XXX,XX @@ static int mptcp_pernet_new_table(struct net *net, struct mptcp_pernet *pernet)
 	table[i++].data = &pernet->checksum_enabled;
 	table[i++].data = &pernet->allow_join_initial_addr_port;
 	table[i++].data = &pernet->stale_loss_cnt;
-	table[i++].data = &pernet->pm_type;
 	table[i++].data = &pernet->path_manager;
 	table[i++].data = &pernet->scheduler;
 	i++; /* table[i] is for available_schedulers which is read-only info */
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -XXX,XX +XXX,XX @@ 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);
 unsigned int mptcp_close_timeout(const struct sock *sk);
-int mptcp_get_pm_type(const struct net *net);
 const char *mptcp_get_path_manager(const struct net *net);
 const char *mptcp_get_scheduler(const struct net *net);
 
-- 
2.43.0

From: Geliang Tang <tanggeliang@kylinos.cn>

Similarly to net.mptcp.available_schedulers, this patch adds a new one
net.mptcp.available_path_managers to list the available path mangers.

Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 Documentation/networking/mptcp-sysctl.rst |  4 ++++
 include/net/mptcp.h                       |  2 ++
 net/mptcp/ctrl.c                          | 25 +++++++++++++++++++++++
 net/mptcp/pm.c                            | 20 ++++++++++++++++++
 net/mptcp/protocol.h                      |  1 +
 5 files changed, 52 insertions(+)

diff --git a/Documentation/networking/mptcp-sysctl.rst b/Documentation/networking/mptcp-sysctl.rst
index XXXXXXX..XXXXXXX 100644
--- a/Documentation/networking/mptcp-sysctl.rst
+++ b/Documentation/networking/mptcp-sysctl.rst
@@ -XXX,XX +XXX,XX @@ path_manager - STRING
 
 	Default: "in-kernel"
 
+available_path_managers - STRING
+	Shows the available path managers choices that are registered. More
+        path managers may be available, but not loaded.
+
 scheduler - STRING
 	Select the scheduler of your choice.
 
diff --git a/include/net/mptcp.h b/include/net/mptcp.h
index XXXXXXX..XXXXXXX 100644
--- a/include/net/mptcp.h
+++ b/include/net/mptcp.h
@@ -XXX,XX +XXX,XX @@ struct mptcp_sched_ops {
 } ____cacheline_aligned_in_smp;
 
 #define MPTCP_PM_NAME_MAX	16
+#define MPTCP_PM_MAX		128
+#define MPTCP_PM_BUF_MAX	(MPTCP_PM_NAME_MAX * MPTCP_PM_MAX)
 
 struct mptcp_pm_ops {
 	int (*created)(struct mptcp_sock *msk);
diff --git a/net/mptcp/ctrl.c b/net/mptcp/ctrl.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/ctrl.c
+++ b/net/mptcp/ctrl.c
@@ -XXX,XX +XXX,XX @@ static int proc_path_manager(const struct ctl_table *ctl, int write,
 	return ret;
 }
 
+static int proc_available_path_managers(const struct ctl_table *ctl,
+					int write, void *buffer,
+					size_t *lenp, loff_t *ppos)
+{
+	struct ctl_table tbl = { .maxlen = MPTCP_PM_BUF_MAX, };
+	int ret;
+
+	tbl.data = kmalloc(tbl.maxlen, GFP_USER);
+	if (!tbl.data)
+		return -ENOMEM;
+
+	mptcp_pm_get_available(tbl.data, MPTCP_PM_BUF_MAX);
+	ret = proc_dostring(&tbl, write, buffer, lenp, ppos);
+	kfree(tbl.data);
+
+	return ret;
+}
+
 static int mptcp_set_scheduler(char *scheduler, const char *name)
 {
 	struct mptcp_sched_ops *sched;
@@ -XXX,XX +XXX,XX @@ static struct ctl_table mptcp_sysctl_table[] = {
 		.mode = 0644,
 		.proc_handler = proc_path_manager,
 	},
+	{
+		.procname = "available_path_managers",
+		.maxlen	= MPTCP_PM_BUF_MAX,
+		.mode = 0444,
+		.proc_handler = proc_available_path_managers,
+	},
 	{
 		.procname = "scheduler",
 		.maxlen	= MPTCP_SCHED_NAME_MAX,
@@ -XXX,XX +XXX,XX @@ static int mptcp_pernet_new_table(struct net *net, struct mptcp_pernet *pernet)
 	table[i++].data = &pernet->allow_join_initial_addr_port;
 	table[i++].data = &pernet->stale_loss_cnt;
 	table[i++].data = &pernet->path_manager;
+	i++; /* table[i] is for available_path_managers which is read-only info */
 	table[i++].data = &pernet->scheduler;
 	i++; /* table[i] is for available_schedulers which is read-only info */
 	table[i++].data = &pernet->close_timeout;
diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/pm.c
+++ b/net/mptcp/pm.c
@@ -XXX,XX +XXX,XX @@ struct mptcp_pm_ops *mptcp_pm_find(const char *name)
 	return NULL;
 }
 
+/* Build string with list of available path manager values.
+ * Similar to tcp_get_available_congestion_control()
+ */
+void mptcp_pm_get_available(char *buf, size_t maxlen)
+{
+	struct mptcp_pm_ops *pm;
+	size_t offs = 0;
+
+	rcu_read_lock();
+	list_for_each_entry_rcu(pm, &mptcp_pm_list, list) {
+		offs += snprintf(buf + offs, maxlen - offs,
+				 "%s%s",
+				 offs == 0 ? "" : " ", pm->name);
+
+		if (WARN_ON_ONCE(offs >= maxlen))
+			break;
+	}
+	rcu_read_unlock();
+}
+
 int mptcp_pm_validate(struct mptcp_pm_ops *pm)
 {
 	if (!pm->get_local_id || !pm->get_priority) {
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -XXX,XX +XXX,XX @@ void mptcp_pm_remove_addr_entry(struct mptcp_sock *msk,
 extern struct mptcp_pm_ops mptcp_in_kernel_pm;
 
 struct mptcp_pm_ops *mptcp_pm_find(const char *name);
+void mptcp_pm_get_available(char *buf, size_t maxlen);
 int mptcp_pm_validate(struct mptcp_pm_ops *pm);
 int mptcp_pm_register(struct mptcp_pm_ops *pm);
 void mptcp_pm_unregister(struct mptcp_pm_ops *pm);
-- 
2.43.0

From: Geliang Tang <tanggeliang@kylinos.cn>

Now mptcp_pm_nl_get_local_id() and mptcp_userspace_pm_get_local_id()
helpers can be dropped, and mptcp_pm_get_local_id() can directly invoke
get_local_id() interface through "ops" of "msk->pm".

Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 net/mptcp/pm.c           | 4 +---
 net/mptcp/pm_netlink.c   | 4 ++--
 net/mptcp/pm_userspace.c | 4 ++--
 net/mptcp/protocol.h     | 4 ----
 4 files changed, 5 insertions(+), 11 deletions(-)

diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/pm.c
+++ b/net/mptcp/pm.c
@@ -XXX,XX +XXX,XX @@ int mptcp_pm_get_local_id(struct mptcp_sock *msk, struct sock_common *skc)
 	skc_local.addr.id = 0;
 	skc_local.flags = MPTCP_PM_ADDR_FLAG_IMPLICIT;
 
-	if (mptcp_pm_is_userspace(msk))
-		return mptcp_userspace_pm_get_local_id(msk, &skc_local);
-	return mptcp_pm_nl_get_local_id(msk, &skc_local);
+	return msk->pm.ops->get_local_id(msk, &skc_local);
 }
 
 bool mptcp_pm_is_backup(struct mptcp_sock *msk, struct sock_common *skc)
diff --git a/net/mptcp/pm_netlink.c b/net/mptcp/pm_netlink.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/pm_netlink.c
+++ b/net/mptcp/pm_netlink.c
@@ -XXX,XX +XXX,XX @@ static int mptcp_pm_nl_create_listen_socket(struct sock *sk,
 	return err;
 }
 
-int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk,
-			     struct mptcp_pm_addr_entry *skc)
+static int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk,
+				    struct mptcp_pm_addr_entry *skc)
 {
 	struct mptcp_pm_addr_entry *entry;
 	struct pm_nl_pernet *pernet;
diff --git a/net/mptcp/pm_userspace.c b/net/mptcp/pm_userspace.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/pm_userspace.c
+++ b/net/mptcp/pm_userspace.c
@@ -XXX,XX +XXX,XX @@ mptcp_userspace_pm_lookup_addr_by_id(struct mptcp_sock *msk, unsigned int id)
 	return NULL;
 }
 
-int mptcp_userspace_pm_get_local_id(struct mptcp_sock *msk,
-				    struct mptcp_pm_addr_entry *skc)
+static int mptcp_userspace_pm_get_local_id(struct mptcp_sock *msk,
+					   struct mptcp_pm_addr_entry *skc)
 {
 	__be16 msk_sport =  ((struct inet_sock *)
 			     inet_sk((struct sock *)msk))->inet_sport;
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -XXX,XX +XXX,XX @@ bool mptcp_pm_add_addr_signal(struct mptcp_sock *msk, const struct sk_buff *skb,
 bool mptcp_pm_rm_addr_signal(struct mptcp_sock *msk, unsigned int remaining,
 			     struct mptcp_rm_list *rm_list);
 int mptcp_pm_get_local_id(struct mptcp_sock *msk, struct sock_common *skc);
-int mptcp_pm_nl_get_local_id(struct mptcp_sock *msk,
-			     struct mptcp_pm_addr_entry *skc);
-int mptcp_userspace_pm_get_local_id(struct mptcp_sock *msk,
-				    struct mptcp_pm_addr_entry *skc);
 bool mptcp_pm_is_backup(struct mptcp_sock *msk, struct sock_common *skc);
 bool mptcp_pm_nl_is_backup(struct mptcp_sock *msk, struct mptcp_addr_info *skc);
 bool mptcp_userspace_pm_is_backup(struct mptcp_sock *msk, struct mptcp_addr_info *skc);
-- 
2.43.0

From: Geliang Tang <tanggeliang@kylinos.cn>

Now mptcp_pm_nl_is_backup() and mptcp_userspace_pm_is_backup()
helpers can be dropped, and mptcp_pm_is_backup() can directly
invoke get_priority() interface through "ops" of "msk->pm".

Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 net/mptcp/pm.c           | 5 +----
 net/mptcp/pm_netlink.c   | 5 -----
 net/mptcp/pm_userspace.c | 6 ------
 net/mptcp/protocol.h     | 2 --
 4 files changed, 1 insertion(+), 17 deletions(-)

diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/pm.c
+++ b/net/mptcp/pm.c
@@ -XXX,XX +XXX,XX @@ bool mptcp_pm_is_backup(struct mptcp_sock *msk, struct sock_common *skc)
 
 	mptcp_local_address((struct sock_common *)skc, &skc_local);
 
-	if (mptcp_pm_is_userspace(msk))
-		return mptcp_userspace_pm_is_backup(msk, &skc_local);
-
-	return mptcp_pm_nl_is_backup(msk, &skc_local);
+	return msk->pm.ops->get_priority(msk, &skc_local);
 }
 
 static int mptcp_pm_get_addr(u8 id, struct mptcp_pm_addr_entry *addr,
diff --git a/net/mptcp/pm_netlink.c b/net/mptcp/pm_netlink.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/pm_netlink.c
+++ b/net/mptcp/pm_netlink.c
@@ -XXX,XX +XXX,XX @@ static bool mptcp_pm_nl_get_priority(struct mptcp_sock *msk,
 	return backup;
 }
 
-bool mptcp_pm_nl_is_backup(struct mptcp_sock *msk, struct mptcp_addr_info *skc)
-{
-	return mptcp_pm_nl_get_priority(msk, skc);
-}
-
 #define MPTCP_PM_CMD_GRP_OFFSET       0
 #define MPTCP_PM_EV_GRP_OFFSET        1
 
diff --git a/net/mptcp/pm_userspace.c b/net/mptcp/pm_userspace.c
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/pm_userspace.c
+++ b/net/mptcp/pm_userspace.c
@@ -XXX,XX +XXX,XX @@ static bool mptcp_userspace_pm_get_priority(struct mptcp_sock *msk,
 	return backup;
 }
 
-bool mptcp_userspace_pm_is_backup(struct mptcp_sock *msk,
-				  struct mptcp_addr_info *skc)
-{
-	return mptcp_userspace_pm_get_priority(msk, skc);
-}
-
 static struct mptcp_sock *mptcp_userspace_pm_get_sock(const struct genl_info *info)
 {
 	struct mptcp_sock *msk;
diff --git a/net/mptcp/protocol.h b/net/mptcp/protocol.h
index XXXXXXX..XXXXXXX 100644
--- a/net/mptcp/protocol.h
+++ b/net/mptcp/protocol.h
@@ -XXX,XX +XXX,XX @@ bool mptcp_pm_rm_addr_signal(struct mptcp_sock *msk, unsigned int remaining,
 			     struct mptcp_rm_list *rm_list);
 int mptcp_pm_get_local_id(struct mptcp_sock *msk, struct sock_common *skc);
 bool mptcp_pm_is_backup(struct mptcp_sock *msk, struct sock_common *skc);
-bool mptcp_pm_nl_is_backup(struct mptcp_sock *msk, struct mptcp_addr_info *skc);
-bool mptcp_userspace_pm_is_backup(struct mptcp_sock *msk, struct mptcp_addr_info *skc);
 int mptcp_pm_nl_dump_addr(struct sk_buff *msg,
 			  struct netlink_callback *cb);
 int mptcp_userspace_pm_dump_addr(struct sk_buff *msg,
-- 
2.43.0

From: Geliang Tang <tanggeliang@kylinos.cn>

This patch uses the newly added net.mptcp.path_manager instead of the
old net.mptcp.pm_type for all path manager selftests in mptcp_join.sh
and userspace_pm.sh.

Signed-off-by: Geliang Tang <tanggeliang@kylinos.cn>
---
 .../testing/selftests/net/mptcp/mptcp_join.sh | 26 +++++++++----------
 .../selftests/net/mptcp/userspace_pm.sh       |  4 +--
 2 files changed, 15 insertions(+), 15 deletions(-)

diff --git a/tools/testing/selftests/net/mptcp/mptcp_join.sh b/tools/testing/selftests/net/mptcp/mptcp_join.sh
index XXXXXXX..XXXXXXX 100755
--- a/tools/testing/selftests/net/mptcp/mptcp_join.sh
+++ b/tools/testing/selftests/net/mptcp/mptcp_join.sh
@@ -XXX,XX +XXX,XX @@ init_partial()
 
 	local netns
 	for netns in "$ns1" "$ns2"; do
-		ip netns exec $netns sysctl -q net.mptcp.pm_type=0 2>/dev/null || true
+		ip netns exec $netns sysctl -q net.mptcp.path_manager="in-kernel" 2>/dev/null || true
 		if $checksum; then
 			ip netns exec $netns sysctl -q net.mptcp.checksum_enabled=1
 		fi
@@ -XXX,XX +XXX,XX @@ set_userspace_pm()
 {
 	local ns=$1
 
-	ip netns exec $ns sysctl -q net.mptcp.pm_type=1
+	ip netns exec $ns sysctl -q net.mptcp.path_manager="userspace"
 }
 
 subflows_tests()
@@ -XXX,XX +XXX,XX @@ userspace_tests()
 {
 	# userspace pm type prevents add_addr
 	if reset "userspace pm type prevents add_addr" &&
-	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
+	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
 		set_userspace_pm $ns1
 		pm_nl_set_limits $ns1 0 2
 		pm_nl_set_limits $ns2 0 2
@@ -XXX,XX +XXX,XX @@ userspace_tests()
 
 	# userspace pm type does not echo add_addr without daemon
 	if reset "userspace pm no echo w/o daemon" &&
-	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
+	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
 		set_userspace_pm $ns2
 		pm_nl_set_limits $ns1 0 2
 		pm_nl_set_limits $ns2 0 2
@@ -XXX,XX +XXX,XX @@ userspace_tests()
 
 	# userspace pm type rejects join
 	if reset "userspace pm type rejects join" &&
-	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
+	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
 		set_userspace_pm $ns1
 		pm_nl_set_limits $ns1 1 1
 		pm_nl_set_limits $ns2 1 1
@@ -XXX,XX +XXX,XX @@ userspace_tests()
 
 	# userspace pm type does not send join
 	if reset "userspace pm type does not send join" &&
-	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
+	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
 		set_userspace_pm $ns2
 		pm_nl_set_limits $ns1 1 1
 		pm_nl_set_limits $ns2 1 1
@@ -XXX,XX +XXX,XX @@ userspace_tests()
 
 	# userspace pm type prevents mp_prio
 	if reset "userspace pm type prevents mp_prio" &&
-	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
+	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
 		set_userspace_pm $ns1
 		pm_nl_set_limits $ns1 1 1
 		pm_nl_set_limits $ns2 1 1
@@ -XXX,XX +XXX,XX @@ userspace_tests()
 
 	# userspace pm type prevents rm_addr
 	if reset "userspace pm type prevents rm_addr" &&
-	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
+	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
 		set_userspace_pm $ns1
 		set_userspace_pm $ns2
 		pm_nl_set_limits $ns1 0 1
@@ -XXX,XX +XXX,XX @@ userspace_tests()
 
 	# userspace pm add & remove address
 	if reset_with_events "userspace pm add & remove address" &&
-	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
+	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
 		set_userspace_pm $ns1
 		pm_nl_set_limits $ns2 2 2
 		{ speed=5 \
@@ -XXX,XX +XXX,XX @@ userspace_tests()
 
 	# userspace pm create destroy subflow
 	if reset_with_events "userspace pm create destroy subflow" &&
-	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
+	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
 		set_userspace_pm $ns2
 		pm_nl_set_limits $ns1 0 1
 		{ speed=5 \
@@ -XXX,XX +XXX,XX @@ userspace_tests()
 
 	# userspace pm create id 0 subflow
 	if reset_with_events "userspace pm create id 0 subflow" &&
-	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
+	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
 		set_userspace_pm $ns2
 		pm_nl_set_limits $ns1 0 1
 		{ speed=5 \
@@ -XXX,XX +XXX,XX @@ userspace_tests()
 
 	# userspace pm remove initial subflow
 	if reset_with_events "userspace pm remove initial subflow" &&
-	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
+	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
 		set_userspace_pm $ns2
 		pm_nl_set_limits $ns1 0 1
 		{ speed=5 \
@@ -XXX,XX +XXX,XX @@ userspace_tests()
 
 	# userspace pm send RM_ADDR for ID 0
 	if reset_with_events "userspace pm send RM_ADDR for ID 0" &&
-	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
+	   continue_if mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
 		set_userspace_pm $ns1
 		pm_nl_set_limits $ns2 1 1
 		{ speed=5 \
diff --git a/tools/testing/selftests/net/mptcp/userspace_pm.sh b/tools/testing/selftests/net/mptcp/userspace_pm.sh
index XXXXXXX..XXXXXXX 100755
--- a/tools/testing/selftests/net/mptcp/userspace_pm.sh
+++ b/tools/testing/selftests/net/mptcp/userspace_pm.sh
@@ -XXX,XX +XXX,XX @@
 mptcp_lib_check_mptcp
 mptcp_lib_check_kallsyms
 
-if ! mptcp_lib_has_file '/proc/sys/net/mptcp/pm_type'; then
+if ! mptcp_lib_has_file '/proc/sys/net/mptcp/path_manager'; then
 	echo "userspace pm tests are not supported by the kernel: SKIP"
 	exit ${KSFT_SKIP}
 fi
@@ -XXX,XX +XXX,XX @@ trap cleanup EXIT
 # Create and configure network namespaces for testing
 mptcp_lib_ns_init ns1 ns2
 for i in "$ns1" "$ns2" ;do
-	ip netns exec "$i" sysctl -q net.mptcp.pm_type=1
+	ip netns exec "$i" sysctl -q net.mptcp.path_manager="userspace"
 done
 
 #  "$ns1"              ns2
-- 
2.43.0