IPVS: netns awareness to ip_vs_sync
authorHans Schillstrom <hans.schillstrom@ericsson.com>
Mon, 3 Jan 2011 13:44:55 +0000 (14:44 +0100)
committerSimon Horman <horms@verge.net.au>
Thu, 13 Jan 2011 01:30:28 +0000 (10:30 +0900)
All global variables moved to struct ipvs,
most external changes fixed (i.e. init_net removed)
in sync_buf create  + 4 replaced by sizeof(struct..)

Signed-off-by: Hans Schillstrom <hans.schillstrom@ericsson.com>
Acked-by: Julian Anastasov <ja@ssi.bg>
Signed-off-by: Simon Horman <horms@verge.net.au>
include/net/ip_vs.h
include/net/netns/ip_vs.h
net/netfilter/ipvs/ip_vs_core.c
net/netfilter/ipvs/ip_vs_ctl.c
net/netfilter/ipvs/ip_vs_sync.c

index c08927bb17289aa2ac47c88af32061d28c9666e3..4265b5e00c94699c877c9f8783cc29df5741ab30 100644 (file)
@@ -958,7 +958,7 @@ extern struct ip_vs_stats ip_vs_stats;
 extern const struct ctl_path net_vs_ctl_path[];
 extern int sysctl_ip_vs_sync_ver;
 
-extern void ip_vs_sync_switch_mode(int mode);
+extern void ip_vs_sync_switch_mode(struct net *net, int mode);
 extern struct ip_vs_service *
 ip_vs_service_get(struct net *net, int af, __u32 fwmark, __u16 protocol,
                  const union nf_inet_addr *vaddr, __be16 vport);
@@ -987,14 +987,10 @@ extern struct ip_vs_dest *ip_vs_try_bind_dest(struct ip_vs_conn *cp);
  *      IPVS sync daemon data and function prototypes
  *      (from ip_vs_sync.c)
  */
-extern volatile int ip_vs_sync_state;
-extern volatile int ip_vs_master_syncid;
-extern volatile int ip_vs_backup_syncid;
-extern char ip_vs_master_mcast_ifn[IP_VS_IFNAME_MAXLEN];
-extern char ip_vs_backup_mcast_ifn[IP_VS_IFNAME_MAXLEN];
-extern int start_sync_thread(int state, char *mcast_ifn, __u8 syncid);
-extern int stop_sync_thread(int state);
-extern void ip_vs_sync_conn(struct ip_vs_conn *cp);
+extern int start_sync_thread(struct net *net, int state, char *mcast_ifn,
+                            __u8 syncid);
+extern int stop_sync_thread(struct net *net, int state);
+extern void ip_vs_sync_conn(struct net *net, struct ip_vs_conn *cp);
 extern int ip_vs_sync_init(void);
 extern void ip_vs_sync_cleanup(void);
 
index db0240198339cd537679ea3df8c16f4bd61c0e4e..aba78f3c83411eb96d6428fc2e8edfd69964d242 100644 (file)
@@ -74,6 +74,22 @@ struct netns_ipvs {
        struct list_head        est_list;       /* estimator list */
        spinlock_t              est_lock;
        struct timer_list       est_timer;      /* Estimation timer */
+       /* ip_vs_sync */
+       struct list_head        sync_queue;
+       spinlock_t              sync_lock;
+       struct ip_vs_sync_buff  *sync_buff;
+       spinlock_t              sync_buff_lock;
+       struct sockaddr_in      sync_mcast_addr;
+       struct task_struct      *master_thread;
+       struct task_struct      *backup_thread;
+       int                     send_mesg_maxlen;
+       int                     recv_mesg_maxlen;
+       volatile int            sync_state;
+       volatile int            master_syncid;
+       volatile int            backup_syncid;
+       /* multicast interface name */
+       char                    master_mcast_ifn[IP_VS_IFNAME_MAXLEN];
+       char                    backup_mcast_ifn[IP_VS_IFNAME_MAXLEN];
 };
 
 #endif /* IP_VS_H_ */
index 9317affc5ea1435989aefe4dceeaef3bd965d34f..5531d569aa5e39eef5e8fb7f937b6417d515ac0d 100644 (file)
@@ -1471,12 +1471,13 @@ ip_vs_in_icmp_v6(struct sk_buff *skb, int *related, unsigned int hooknum)
 static unsigned int
 ip_vs_in(unsigned int hooknum, struct sk_buff *skb, int af)
 {
-       struct net *net = NULL;
+       struct net *net;
        struct ip_vs_iphdr iph;
        struct ip_vs_protocol *pp;
        struct ip_vs_proto_data *pd;
        struct ip_vs_conn *cp;
        int ret, restart, pkts;
+       struct netns_ipvs *ipvs;
 
        /* Already marked as IPVS request or reply? */
        if (skb->ipvs_property)
@@ -1556,7 +1557,8 @@ ip_vs_in(unsigned int hooknum, struct sk_buff *skb, int af)
        }
 
        IP_VS_DBG_PKT(11, af, pp, skb, 0, "Incoming packet");
-
+       net = skb_net(skb);
+       ipvs = net_ipvs(net);
        /* Check the server status */
        if (cp->dest && !(cp->dest->flags & IP_VS_DEST_F_AVAILABLE)) {
                /* the destination server is not available */
@@ -1589,12 +1591,13 @@ ip_vs_in(unsigned int hooknum, struct sk_buff *skb, int af)
         *
         * For ONE_PKT let ip_vs_sync_conn() do the filter work.
         */
+
        if (cp->flags & IP_VS_CONN_F_ONE_PACKET)
                pkts = sysctl_ip_vs_sync_threshold[0];
        else
                pkts = atomic_add_return(1, &cp->in_pkts);
 
-       if ((ip_vs_sync_state & IP_VS_STATE_MASTER) &&
+       if ((ipvs->sync_state & IP_VS_STATE_MASTER) &&
            cp->protocol == IPPROTO_SCTP) {
                if ((cp->state == IP_VS_SCTP_S_ESTABLISHED &&
                        (pkts % sysctl_ip_vs_sync_threshold[1]
@@ -1603,13 +1606,13 @@ ip_vs_in(unsigned int hooknum, struct sk_buff *skb, int af)
                                 ((cp->state == IP_VS_SCTP_S_CLOSED) ||
                                  (cp->state == IP_VS_SCTP_S_SHUT_ACK_CLI) ||
                                  (cp->state == IP_VS_SCTP_S_SHUT_ACK_SER)))) {
-                       ip_vs_sync_conn(cp);
+                       ip_vs_sync_conn(net, cp);
                        goto out;
                }
        }
 
        /* Keep this block last: TCP and others with pp->num_states <= 1 */
-       else if ((ip_vs_sync_state & IP_VS_STATE_MASTER) &&
+       else if ((ipvs->sync_state & IP_VS_STATE_MASTER) &&
            (((cp->protocol != IPPROTO_TCP ||
               cp->state == IP_VS_TCP_S_ESTABLISHED) &&
              (pkts % sysctl_ip_vs_sync_threshold[1]
@@ -1619,7 +1622,7 @@ ip_vs_in(unsigned int hooknum, struct sk_buff *skb, int af)
               (cp->state == IP_VS_TCP_S_CLOSE) ||
               (cp->state == IP_VS_TCP_S_CLOSE_WAIT) ||
               (cp->state == IP_VS_TCP_S_TIME_WAIT)))))
-               ip_vs_sync_conn(cp);
+               ip_vs_sync_conn(net, cp);
 out:
        cp->old_state = cp->state;
 
index c89beb8eafbb4bf65dbd0911298e55ce4555bb83..03f86312b4bbb6bdaaad592e11354cf56446f36b 100644 (file)
@@ -1559,7 +1559,8 @@ proc_do_sync_mode(ctl_table *table, int write,
                        /* Restore the correct value */
                        *valp = val;
                } else {
-                       ip_vs_sync_switch_mode(val);
+                       struct net *net = current->nsproxy->net_ns;
+                       ip_vs_sync_switch_mode(net, val);
                }
        }
        return rc;
@@ -2174,11 +2175,12 @@ do_ip_vs_set_ctl(struct sock *sk, int cmd, void __user *user, unsigned int len)
                goto out_unlock;
        } else if (cmd == IP_VS_SO_SET_STARTDAEMON) {
                struct ip_vs_daemon_user *dm = (struct ip_vs_daemon_user *)arg;
-               ret = start_sync_thread(dm->state, dm->mcast_ifn, dm->syncid);
+               ret = start_sync_thread(net, dm->state, dm->mcast_ifn,
+                                       dm->syncid);
                goto out_unlock;
        } else if (cmd == IP_VS_SO_SET_STOPDAEMON) {
                struct ip_vs_daemon_user *dm = (struct ip_vs_daemon_user *)arg;
-               ret = stop_sync_thread(dm->state);
+               ret = stop_sync_thread(net, dm->state);
                goto out_unlock;
        }
 
@@ -2424,6 +2426,7 @@ do_ip_vs_get_ctl(struct sock *sk, int cmd, void __user *user, int *len)
        int ret = 0;
        unsigned int copylen;
        struct net *net = sock_net(sk);
+       struct netns_ipvs *ipvs = net_ipvs(net);
 
        BUG_ON(!net);
        if (!capable(CAP_NET_ADMIN))
@@ -2546,15 +2549,17 @@ do_ip_vs_get_ctl(struct sock *sk, int cmd, void __user *user, int *len)
                struct ip_vs_daemon_user d[2];
 
                memset(&d, 0, sizeof(d));
-               if (ip_vs_sync_state & IP_VS_STATE_MASTER) {
+               if (ipvs->sync_state & IP_VS_STATE_MASTER) {
                        d[0].state = IP_VS_STATE_MASTER;
-                       strlcpy(d[0].mcast_ifn, ip_vs_master_mcast_ifn, sizeof(d[0].mcast_ifn));
-                       d[0].syncid = ip_vs_master_syncid;
+                       strlcpy(d[0].mcast_ifn, ipvs->master_mcast_ifn,
+                               sizeof(d[0].mcast_ifn));
+                       d[0].syncid = ipvs->master_syncid;
                }
-               if (ip_vs_sync_state & IP_VS_STATE_BACKUP) {
+               if (ipvs->sync_state & IP_VS_STATE_BACKUP) {
                        d[1].state = IP_VS_STATE_BACKUP;
-                       strlcpy(d[1].mcast_ifn, ip_vs_backup_mcast_ifn, sizeof(d[1].mcast_ifn));
-                       d[1].syncid = ip_vs_backup_syncid;
+                       strlcpy(d[1].mcast_ifn, ipvs->backup_mcast_ifn,
+                               sizeof(d[1].mcast_ifn));
+                       d[1].syncid = ipvs->backup_syncid;
                }
                if (copy_to_user(user, &d, sizeof(d)) != 0)
                        ret = -EFAULT;
@@ -3061,20 +3066,23 @@ nla_put_failure:
 static int ip_vs_genl_dump_daemons(struct sk_buff *skb,
                                   struct netlink_callback *cb)
 {
+       struct net *net = skb_net(skb);
+       struct netns_ipvs *ipvs = net_ipvs(net);
+
        mutex_lock(&__ip_vs_mutex);
-       if ((ip_vs_sync_state & IP_VS_STATE_MASTER) && !cb->args[0]) {
+       if ((ipvs->sync_state & IP_VS_STATE_MASTER) && !cb->args[0]) {
                if (ip_vs_genl_dump_daemon(skb, IP_VS_STATE_MASTER,
-                                          ip_vs_master_mcast_ifn,
-                                          ip_vs_master_syncid, cb) < 0)
+                                          ipvs->master_mcast_ifn,
+                                          ipvs->master_syncid, cb) < 0)
                        goto nla_put_failure;
 
                cb->args[0] = 1;
        }
 
-       if ((ip_vs_sync_state & IP_VS_STATE_BACKUP) && !cb->args[1]) {
+       if ((ipvs->sync_state & IP_VS_STATE_BACKUP) && !cb->args[1]) {
                if (ip_vs_genl_dump_daemon(skb, IP_VS_STATE_BACKUP,
-                                          ip_vs_backup_mcast_ifn,
-                                          ip_vs_backup_syncid, cb) < 0)
+                                          ipvs->backup_mcast_ifn,
+                                          ipvs->backup_syncid, cb) < 0)
                        goto nla_put_failure;
 
                cb->args[1] = 1;
@@ -3086,24 +3094,26 @@ nla_put_failure:
        return skb->len;
 }
 
-static int ip_vs_genl_new_daemon(struct nlattr **attrs)
+static int ip_vs_genl_new_daemon(struct net *net, struct nlattr **attrs)
 {
        if (!(attrs[IPVS_DAEMON_ATTR_STATE] &&
              attrs[IPVS_DAEMON_ATTR_MCAST_IFN] &&
              attrs[IPVS_DAEMON_ATTR_SYNC_ID]))
                return -EINVAL;
 
-       return start_sync_thread(nla_get_u32(attrs[IPVS_DAEMON_ATTR_STATE]),
+       return start_sync_thread(net,
+                                nla_get_u32(attrs[IPVS_DAEMON_ATTR_STATE]),
                                 nla_data(attrs[IPVS_DAEMON_ATTR_MCAST_IFN]),
                                 nla_get_u32(attrs[IPVS_DAEMON_ATTR_SYNC_ID]));
 }
 
-static int ip_vs_genl_del_daemon(struct nlattr **attrs)
+static int ip_vs_genl_del_daemon(struct net *net, struct nlattr **attrs)
 {
        if (!attrs[IPVS_DAEMON_ATTR_STATE])
                return -EINVAL;
 
-       return stop_sync_thread(nla_get_u32(attrs[IPVS_DAEMON_ATTR_STATE]));
+       return stop_sync_thread(net,
+                               nla_get_u32(attrs[IPVS_DAEMON_ATTR_STATE]));
 }
 
 static int ip_vs_genl_set_config(struct net *net, struct nlattr **attrs)
@@ -3159,9 +3169,9 @@ static int ip_vs_genl_set_cmd(struct sk_buff *skb, struct genl_info *info)
                }
 
                if (cmd == IPVS_CMD_NEW_DAEMON)
-                       ret = ip_vs_genl_new_daemon(daemon_attrs);
+                       ret = ip_vs_genl_new_daemon(net, daemon_attrs);
                else
-                       ret = ip_vs_genl_del_daemon(daemon_attrs);
+                       ret = ip_vs_genl_del_daemon(net, daemon_attrs);
                goto out;
        } else if (cmd == IPVS_CMD_ZERO &&
                   !info->attrs[IPVS_CMD_ATTR_SERVICE]) {
index 6831e8fac8db9a9eda8f0a9f8054eedfaba60e62..c29e73d686fb07a6923b9ce395df1c800eae021f 100644 (file)
@@ -192,6 +192,7 @@ union ip_vs_sync_conn {
 #define IPVS_OPT_F_PARAM       (1 << (IPVS_OPT_PARAM-1))
 
 struct ip_vs_sync_thread_data {
+       struct net *net;
        struct socket *sock;
        char *buf;
 };
@@ -259,10 +260,6 @@ struct ip_vs_sync_mesg {
        /* ip_vs_sync_conn entries start here */
 };
 
-/* the maximum length of sync (sending/receiving) message */
-static int sync_send_mesg_maxlen;
-static int sync_recv_mesg_maxlen;
-
 struct ip_vs_sync_buff {
        struct list_head        list;
        unsigned long           firstuse;
@@ -273,28 +270,6 @@ struct ip_vs_sync_buff {
        unsigned char           *end;
 };
 
-
-/* the sync_buff list head and the lock */
-static LIST_HEAD(ip_vs_sync_queue);
-static DEFINE_SPINLOCK(ip_vs_sync_lock);
-
-/* current sync_buff for accepting new conn entries */
-static struct ip_vs_sync_buff   *curr_sb = NULL;
-static DEFINE_SPINLOCK(curr_sb_lock);
-
-/* ipvs sync daemon state */
-volatile int ip_vs_sync_state = IP_VS_STATE_NONE;
-volatile int ip_vs_master_syncid = 0;
-volatile int ip_vs_backup_syncid = 0;
-
-/* multicast interface name */
-char ip_vs_master_mcast_ifn[IP_VS_IFNAME_MAXLEN];
-char ip_vs_backup_mcast_ifn[IP_VS_IFNAME_MAXLEN];
-
-/* sync daemon tasks */
-static struct task_struct *sync_master_thread;
-static struct task_struct *sync_backup_thread;
-
 /* multicast addr */
 static struct sockaddr_in mcast_addr = {
        .sin_family             = AF_INET,
@@ -324,20 +299,20 @@ static void hton_seq(struct ip_vs_seq *ho, struct ip_vs_seq *no)
        put_unaligned_be32(ho->previous_delta, &no->previous_delta);
 }
 
-static inline struct ip_vs_sync_buff *sb_dequeue(void)
+static inline struct ip_vs_sync_buff *sb_dequeue(struct netns_ipvs *ipvs)
 {
        struct ip_vs_sync_buff *sb;
 
-       spin_lock_bh(&ip_vs_sync_lock);
-       if (list_empty(&ip_vs_sync_queue)) {
+       spin_lock_bh(&ipvs->sync_lock);
+       if (list_empty(&ipvs->sync_queue)) {
                sb = NULL;
        } else {
-               sb = list_entry(ip_vs_sync_queue.next,
+               sb = list_entry(ipvs->sync_queue.next,
                                struct ip_vs_sync_buff,
                                list);
                list_del(&sb->list);
        }
-       spin_unlock_bh(&ip_vs_sync_lock);
+       spin_unlock_bh(&ipvs->sync_lock);
 
        return sb;
 }
@@ -345,25 +320,27 @@ static inline struct ip_vs_sync_buff *sb_dequeue(void)
 /*
  * Create a new sync buffer for Version 1 proto.
  */
-static inline struct ip_vs_sync_buff * ip_vs_sync_buff_create(void)
+static inline struct ip_vs_sync_buff *
+ip_vs_sync_buff_create(struct netns_ipvs *ipvs)
 {
        struct ip_vs_sync_buff *sb;
 
        if (!(sb=kmalloc(sizeof(struct ip_vs_sync_buff), GFP_ATOMIC)))
                return NULL;
 
-       if (!(sb->mesg=kmalloc(sync_send_mesg_maxlen, GFP_ATOMIC))) {
+       sb->mesg = kmalloc(ipvs->send_mesg_maxlen, GFP_ATOMIC);
+       if (!sb->mesg) {
                kfree(sb);
                return NULL;
        }
        sb->mesg->reserved = 0;  /* old nr_conns i.e. must be zeo now */
        sb->mesg->version = SYNC_PROTO_VER;
-       sb->mesg->syncid = ip_vs_master_syncid;
+       sb->mesg->syncid = ipvs->master_syncid;
        sb->mesg->size = sizeof(struct ip_vs_sync_mesg);
        sb->mesg->nr_conns = 0;
        sb->mesg->spare = 0;
        sb->head = (unsigned char *)sb->mesg + sizeof(struct ip_vs_sync_mesg);
-       sb->end = (unsigned char *)sb->mesg + sync_send_mesg_maxlen;
+       sb->end = (unsigned char *)sb->mesg + ipvs->send_mesg_maxlen;
 
        sb->firstuse = jiffies;
        return sb;
@@ -375,14 +352,16 @@ static inline void ip_vs_sync_buff_release(struct ip_vs_sync_buff *sb)
        kfree(sb);
 }
 
-static inline void sb_queue_tail(struct ip_vs_sync_buff *sb)
+static inline void sb_queue_tail(struct netns_ipvs *ipvs)
 {
-       spin_lock(&ip_vs_sync_lock);
-       if (ip_vs_sync_state & IP_VS_STATE_MASTER)
-               list_add_tail(&sb->list, &ip_vs_sync_queue);
+       struct ip_vs_sync_buff *sb = ipvs->sync_buff;
+
+       spin_lock(&ipvs->sync_lock);
+       if (ipvs->sync_state & IP_VS_STATE_MASTER)
+               list_add_tail(&sb->list, &ipvs->sync_queue);
        else
                ip_vs_sync_buff_release(sb);
-       spin_unlock(&ip_vs_sync_lock);
+       spin_unlock(&ipvs->sync_lock);
 }
 
 /*
@@ -390,18 +369,18 @@ static inline void sb_queue_tail(struct ip_vs_sync_buff *sb)
  *     than the specified time or the specified time is zero.
  */
 static inline struct ip_vs_sync_buff *
-get_curr_sync_buff(unsigned long time)
+get_curr_sync_buff(struct netns_ipvs *ipvs, unsigned long time)
 {
        struct ip_vs_sync_buff *sb;
 
-       spin_lock_bh(&curr_sb_lock);
-       if (curr_sb && (time == 0 ||
-                       time_before(jiffies - curr_sb->firstuse, time))) {
-               sb = curr_sb;
-               curr_sb = NULL;
+       spin_lock_bh(&ipvs->sync_buff_lock);
+       if (ipvs->sync_buff && (time == 0 ||
+           time_before(jiffies - ipvs->sync_buff->firstuse, time))) {
+               sb = ipvs->sync_buff;
+               ipvs->sync_buff = NULL;
        } else
                sb = NULL;
-       spin_unlock_bh(&curr_sb_lock);
+       spin_unlock_bh(&ipvs->sync_buff_lock);
        return sb;
 }
 
@@ -409,33 +388,37 @@ get_curr_sync_buff(unsigned long time)
  * Switch mode from sending version 0 or 1
  *  - must handle sync_buf
  */
-void ip_vs_sync_switch_mode(int mode) {
+void ip_vs_sync_switch_mode(struct net *net, int mode)
+{
+       struct netns_ipvs *ipvs = net_ipvs(net);
 
-       if (!ip_vs_sync_state & IP_VS_STATE_MASTER)
+       if (!ipvs->sync_state & IP_VS_STATE_MASTER)
                return;
-       if (mode == sysctl_ip_vs_sync_ver || !curr_sb)
+       if (mode == sysctl_ip_vs_sync_ver || !ipvs->sync_buff)
                return;
 
-       spin_lock_bh(&curr_sb_lock);
+       spin_lock_bh(&ipvs->sync_buff_lock);
        /* Buffer empty ? then let buf_create do the job  */
-       if ( curr_sb->mesg->size <=  sizeof(struct ip_vs_sync_mesg)) {
-               kfree(curr_sb);
-               curr_sb = NULL;
+       if (ipvs->sync_buff->mesg->size <=  sizeof(struct ip_vs_sync_mesg)) {
+               kfree(ipvs->sync_buff);
+               ipvs->sync_buff = NULL;
        } else {
-               spin_lock_bh(&ip_vs_sync_lock);
-               if (ip_vs_sync_state & IP_VS_STATE_MASTER)
-                       list_add_tail(&curr_sb->list, &ip_vs_sync_queue);
+               spin_lock_bh(&ipvs->sync_lock);
+               if (ipvs->sync_state & IP_VS_STATE_MASTER)
+                       list_add_tail(&ipvs->sync_buff->list,
+                                     &ipvs->sync_queue);
                else
-                       ip_vs_sync_buff_release(curr_sb);
-               spin_unlock_bh(&ip_vs_sync_lock);
+                       ip_vs_sync_buff_release(ipvs->sync_buff);
+               spin_unlock_bh(&ipvs->sync_lock);
        }
-       spin_unlock_bh(&curr_sb_lock);
+       spin_unlock_bh(&ipvs->sync_buff_lock);
 }
 
 /*
  * Create a new sync buffer for Version 0 proto.
  */
-static inline struct ip_vs_sync_buff * ip_vs_sync_buff_create_v0(void)
+static inline struct ip_vs_sync_buff *
+ip_vs_sync_buff_create_v0(struct netns_ipvs *ipvs)
 {
        struct ip_vs_sync_buff *sb;
        struct ip_vs_sync_mesg_v0 *mesg;
@@ -443,16 +426,17 @@ static inline struct ip_vs_sync_buff * ip_vs_sync_buff_create_v0(void)
        if (!(sb=kmalloc(sizeof(struct ip_vs_sync_buff), GFP_ATOMIC)))
                return NULL;
 
-       if (!(sb->mesg=kmalloc(sync_send_mesg_maxlen, GFP_ATOMIC))) {
+       sb->mesg = kmalloc(ipvs->send_mesg_maxlen, GFP_ATOMIC);
+       if (!sb->mesg) {
                kfree(sb);
                return NULL;
        }
        mesg = (struct ip_vs_sync_mesg_v0 *)sb->mesg;
        mesg->nr_conns = 0;
-       mesg->syncid = ip_vs_master_syncid;
-       mesg->size = 4;
-       sb->head = (unsigned char *)mesg + 4;
-       sb->end = (unsigned char *)mesg + sync_send_mesg_maxlen;
+       mesg->syncid = ipvs->master_syncid;
+       mesg->size = sizeof(struct ip_vs_sync_mesg_v0);
+       sb->head = (unsigned char *)mesg + sizeof(struct ip_vs_sync_mesg_v0);
+       sb->end = (unsigned char *)mesg + ipvs->send_mesg_maxlen;
        sb->firstuse = jiffies;
        return sb;
 }
@@ -461,8 +445,9 @@ static inline struct ip_vs_sync_buff * ip_vs_sync_buff_create_v0(void)
  *      Version 0 , could be switched in by sys_ctl.
  *      Add an ip_vs_conn information into the current sync_buff.
  */
-void ip_vs_sync_conn_v0(struct ip_vs_conn *cp)
+void ip_vs_sync_conn_v0(struct net *net, struct ip_vs_conn *cp)
 {
+       struct netns_ipvs *ipvs = net_ipvs(net);
        struct ip_vs_sync_mesg_v0 *m;
        struct ip_vs_sync_conn_v0 *s;
        int len;
@@ -473,10 +458,12 @@ void ip_vs_sync_conn_v0(struct ip_vs_conn *cp)
        if (cp->flags & IP_VS_CONN_F_ONE_PACKET)
                return;
 
-       spin_lock(&curr_sb_lock);
-       if (!curr_sb) {
-               if (!(curr_sb=ip_vs_sync_buff_create_v0())) {
-                       spin_unlock(&curr_sb_lock);
+       spin_lock(&ipvs->sync_buff_lock);
+       if (!ipvs->sync_buff) {
+               ipvs->sync_buff =
+                       ip_vs_sync_buff_create_v0(ipvs);
+               if (!ipvs->sync_buff) {
+                       spin_unlock(&ipvs->sync_buff_lock);
                        pr_err("ip_vs_sync_buff_create failed.\n");
                        return;
                }
@@ -484,8 +471,8 @@ void ip_vs_sync_conn_v0(struct ip_vs_conn *cp)
 
        len = (cp->flags & IP_VS_CONN_F_SEQ_MASK) ? FULL_CONN_SIZE :
                SIMPLE_CONN_SIZE;
-       m = (struct ip_vs_sync_mesg_v0 *)curr_sb->mesg;
-       s = (struct ip_vs_sync_conn_v0 *)curr_sb->head;
+       m = (struct ip_vs_sync_mesg_v0 *)ipvs->sync_buff->mesg;
+       s = (struct ip_vs_sync_conn_v0 *)ipvs->sync_buff->head;
 
        /* copy members */
        s->reserved = 0;
@@ -506,18 +493,18 @@ void ip_vs_sync_conn_v0(struct ip_vs_conn *cp)
 
        m->nr_conns++;
        m->size += len;
-       curr_sb->head += len;
+       ipvs->sync_buff->head += len;
 
        /* check if there is a space for next one */
-       if (curr_sb->head + FULL_CONN_SIZE > curr_sb->end) {
-               sb_queue_tail(curr_sb);
-               curr_sb = NULL;
+       if (ipvs->sync_buff->head + FULL_CONN_SIZE > ipvs->sync_buff->end) {
+               sb_queue_tail(ipvs);
+               ipvs->sync_buff = NULL;
        }
-       spin_unlock(&curr_sb_lock);
+       spin_unlock(&ipvs->sync_buff_lock);
 
        /* synchronize its controller if it has */
        if (cp->control)
-               ip_vs_sync_conn(cp->control);
+               ip_vs_sync_conn(net, cp->control);
 }
 
 /*
@@ -525,8 +512,9 @@ void ip_vs_sync_conn_v0(struct ip_vs_conn *cp)
  *      Called by ip_vs_in.
  *      Sending Version 1 messages
  */
-void ip_vs_sync_conn(struct ip_vs_conn *cp)
+void ip_vs_sync_conn(struct net *net, struct ip_vs_conn *cp)
 {
+       struct netns_ipvs *ipvs = net_ipvs(net);
        struct ip_vs_sync_mesg *m;
        union ip_vs_sync_conn *s;
        __u8 *p;
@@ -534,7 +522,7 @@ void ip_vs_sync_conn(struct ip_vs_conn *cp)
 
        /* Handle old version of the protocol */
        if (sysctl_ip_vs_sync_ver == 0) {
-               ip_vs_sync_conn_v0(cp);
+               ip_vs_sync_conn_v0(net, cp);
                return;
        }
        /* Do not sync ONE PACKET */
@@ -551,7 +539,7 @@ sloop:
                pe_name_len = strnlen(cp->pe->name, IP_VS_PENAME_MAXLEN);
        }
 
-       spin_lock(&curr_sb_lock);
+       spin_lock(&ipvs->sync_buff_lock);
 
 #ifdef CONFIG_IP_VS_IPV6
        if (cp->af == AF_INET6)
@@ -570,26 +558,27 @@ sloop:
 
        /* check if there is a space for this one  */
        pad = 0;
-       if (curr_sb) {
-               pad = (4 - (size_t)curr_sb->head) & 3;
-               if (curr_sb->head + len + pad > curr_sb->end) {
-                       sb_queue_tail(curr_sb);
-                       curr_sb = NULL;
+       if (ipvs->sync_buff) {
+               pad = (4 - (size_t)ipvs->sync_buff->head) & 3;
+               if (ipvs->sync_buff->head + len + pad > ipvs->sync_buff->end) {
+                       sb_queue_tail(ipvs);
+                       ipvs->sync_buff = NULL;
                        pad = 0;
                }
        }
 
-       if (!curr_sb) {
-               if (!(curr_sb=ip_vs_sync_buff_create())) {
-                       spin_unlock(&curr_sb_lock);
+       if (!ipvs->sync_buff) {
+               ipvs->sync_buff = ip_vs_sync_buff_create(ipvs);
+               if (!ipvs->sync_buff) {
+                       spin_unlock(&ipvs->sync_buff_lock);
                        pr_err("ip_vs_sync_buff_create failed.\n");
                        return;
                }
        }
 
-       m = curr_sb->mesg;
-       p = curr_sb->head;
-       curr_sb->head += pad + len;
+       m = ipvs->sync_buff->mesg;
+       p = ipvs->sync_buff->head;
+       ipvs->sync_buff->head += pad + len;
        m->size += pad + len;
        /* Add ev. padding from prev. sync_conn */
        while (pad--)
@@ -647,7 +636,7 @@ sloop:
                }
        }
 
-       spin_unlock(&curr_sb_lock);
+       spin_unlock(&ipvs->sync_buff_lock);
 
 control:
        /* synchronize its controller if it has */
@@ -699,7 +688,8 @@ ip_vs_conn_fill_param_sync(int af, union ip_vs_sync_conn *sc,
                        buff[pe_name_len]=0;
                        p->pe = __ip_vs_pe_getbyname(buff);
                        if (!p->pe) {
-                               IP_VS_DBG(3, "BACKUP, no %s engine found/loaded\n", buff);
+                               IP_VS_DBG(3, "BACKUP, no %s engine found/loaded\n",
+                                            buff);
                                return 1;
                        }
                } else {
@@ -748,7 +738,7 @@ static void ip_vs_proc_conn(struct net *net, struct ip_vs_conn_param *param,
                 * If it is not found the connection will remain unbound
                 * but still handled.
                 */
-               dest = ip_vs_find_dest(&init_net, type, daddr, dport, param->vaddr,
+               dest = ip_vs_find_dest(net, type, daddr, dport, param->vaddr,
                                       param->vport, protocol, fwmark);
 
                /*  Set the approprite ativity flag */
@@ -1089,6 +1079,7 @@ out:
 static void ip_vs_process_message(struct net *net, __u8 *buffer,
                                  const size_t buflen)
 {
+       struct netns_ipvs *ipvs = net_ipvs(net);
        struct ip_vs_sync_mesg *m2 = (struct ip_vs_sync_mesg *)buffer;
        __u8 *p, *msg_end;
        int i, nr_conns;
@@ -1105,7 +1096,7 @@ static void ip_vs_process_message(struct net *net, __u8 *buffer,
                return;
        }
        /* SyncID sanity check */
-       if (ip_vs_backup_syncid != 0 && m2->syncid != ip_vs_backup_syncid) {
+       if (ipvs->backup_syncid != 0 && m2->syncid != ipvs->backup_syncid) {
                IP_VS_DBG(7, "BACKUP, Ignoring syncid = %d\n", m2->syncid);
                return;
        }
@@ -1190,8 +1181,10 @@ static int set_mcast_if(struct sock *sk, char *ifname)
 {
        struct net_device *dev;
        struct inet_sock *inet = inet_sk(sk);
+       struct net *net = sock_net(sk);
 
-       if ((dev = __dev_get_by_name(&init_net, ifname)) == NULL)
+       dev = __dev_get_by_name(net, ifname);
+       if (!dev)
                return -ENODEV;
 
        if (sk->sk_bound_dev_if && dev->ifindex != sk->sk_bound_dev_if)
@@ -1210,30 +1203,33 @@ static int set_mcast_if(struct sock *sk, char *ifname)
  *     Set the maximum length of sync message according to the
  *     specified interface's MTU.
  */
-static int set_sync_mesg_maxlen(int sync_state)
+static int set_sync_mesg_maxlen(struct net *net, int sync_state)
 {
+       struct netns_ipvs *ipvs = net_ipvs(net);
        struct net_device *dev;
        int num;
 
        if (sync_state == IP_VS_STATE_MASTER) {
-               if ((dev = __dev_get_by_name(&init_net, ip_vs_master_mcast_ifn)) == NULL)
+               dev = __dev_get_by_name(net, ipvs->master_mcast_ifn);
+               if (!dev)
                        return -ENODEV;
 
                num = (dev->mtu - sizeof(struct iphdr) -
                       sizeof(struct udphdr) -
                       SYNC_MESG_HEADER_LEN - 20) / SIMPLE_CONN_SIZE;
-               sync_send_mesg_maxlen = SYNC_MESG_HEADER_LEN +
+               ipvs->send_mesg_maxlen = SYNC_MESG_HEADER_LEN +
                        SIMPLE_CONN_SIZE * min(num, MAX_CONNS_PER_SYNCBUFF);
                IP_VS_DBG(7, "setting the maximum length of sync sending "
-                         "message %d.\n", sync_send_mesg_maxlen);
+                         "message %d.\n", ipvs->send_mesg_maxlen);
        } else if (sync_state == IP_VS_STATE_BACKUP) {
-               if ((dev = __dev_get_by_name(&init_net, ip_vs_backup_mcast_ifn)) == NULL)
+               dev = __dev_get_by_name(net, ipvs->backup_mcast_ifn);
+               if (!dev)
                        return -ENODEV;
 
-               sync_recv_mesg_maxlen = dev->mtu -
+               ipvs->recv_mesg_maxlen = dev->mtu -
                        sizeof(struct iphdr) - sizeof(struct udphdr);
                IP_VS_DBG(7, "setting the maximum length of sync receiving "
-                         "message %d.\n", sync_recv_mesg_maxlen);
+                         "message %d.\n", ipvs->recv_mesg_maxlen);
        }
 
        return 0;
@@ -1248,6 +1244,7 @@ static int set_sync_mesg_maxlen(int sync_state)
 static int
 join_mcast_group(struct sock *sk, struct in_addr *addr, char *ifname)
 {
+       struct net *net = sock_net(sk);
        struct ip_mreqn mreq;
        struct net_device *dev;
        int ret;
@@ -1255,7 +1252,8 @@ join_mcast_group(struct sock *sk, struct in_addr *addr, char *ifname)
        memset(&mreq, 0, sizeof(mreq));
        memcpy(&mreq.imr_multiaddr, addr, sizeof(struct in_addr));
 
-       if ((dev = __dev_get_by_name(&init_net, ifname)) == NULL)
+       dev = __dev_get_by_name(net, ifname);
+       if (!dev)
                return -ENODEV;
        if (sk->sk_bound_dev_if && dev->ifindex != sk->sk_bound_dev_if)
                return -EINVAL;
@@ -1272,11 +1270,13 @@ join_mcast_group(struct sock *sk, struct in_addr *addr, char *ifname)
 
 static int bind_mcastif_addr(struct socket *sock, char *ifname)
 {
+       struct net *net = sock_net(sock->sk);
        struct net_device *dev;
        __be32 addr;
        struct sockaddr_in sin;
 
-       if ((dev = __dev_get_by_name(&init_net, ifname)) == NULL)
+       dev = __dev_get_by_name(net, ifname);
+       if (!dev)
                return -ENODEV;
 
        addr = inet_select_addr(dev, 0, RT_SCOPE_UNIVERSE);
@@ -1298,8 +1298,9 @@ static int bind_mcastif_addr(struct socket *sock, char *ifname)
 /*
  *      Set up sending multicast socket over UDP
  */
-static struct socket * make_send_sock(void)
+static struct socket *make_send_sock(struct net *net)
 {
+       struct netns_ipvs *ipvs = net_ipvs(net);
        struct socket *sock;
        int result;
 
@@ -1310,7 +1311,7 @@ static struct socket * make_send_sock(void)
                return ERR_PTR(result);
        }
 
-       result = set_mcast_if(sock->sk, ip_vs_master_mcast_ifn);
+       result = set_mcast_if(sock->sk, ipvs->master_mcast_ifn);
        if (result < 0) {
                pr_err("Error setting outbound mcast interface\n");
                goto error;
@@ -1319,7 +1320,7 @@ static struct socket * make_send_sock(void)
        set_mcast_loop(sock->sk, 0);
        set_mcast_ttl(sock->sk, 1);
 
-       result = bind_mcastif_addr(sock, ip_vs_master_mcast_ifn);
+       result = bind_mcastif_addr(sock, ipvs->master_mcast_ifn);
        if (result < 0) {
                pr_err("Error binding address of the mcast interface\n");
                goto error;
@@ -1343,8 +1344,9 @@ static struct socket * make_send_sock(void)
 /*
  *      Set up receiving multicast socket over UDP
  */
-static struct socket * make_receive_sock(void)
+static struct socket *make_receive_sock(struct net *net)
 {
+       struct netns_ipvs *ipvs = net_ipvs(net);
        struct socket *sock;
        int result;
 
@@ -1368,7 +1370,7 @@ static struct socket * make_receive_sock(void)
        /* join the multicast group */
        result = join_mcast_group(sock->sk,
                        (struct in_addr *) &mcast_addr.sin_addr,
-                       ip_vs_backup_mcast_ifn);
+                       ipvs->backup_mcast_ifn);
        if (result < 0) {
                pr_err("Error joining to the multicast group\n");
                goto error;
@@ -1439,20 +1441,21 @@ ip_vs_receive(struct socket *sock, char *buffer, const size_t buflen)
 static int sync_thread_master(void *data)
 {
        struct ip_vs_sync_thread_data *tinfo = data;
+       struct netns_ipvs *ipvs = net_ipvs(tinfo->net);
        struct ip_vs_sync_buff *sb;
 
        pr_info("sync thread started: state = MASTER, mcast_ifn = %s, "
                "syncid = %d\n",
-               ip_vs_master_mcast_ifn, ip_vs_master_syncid);
+               ipvs->master_mcast_ifn, ipvs->master_syncid);
 
        while (!kthread_should_stop()) {
-               while ((sb = sb_dequeue())) {
+               while ((sb = sb_dequeue(ipvs))) {
                        ip_vs_send_sync_msg(tinfo->sock, sb->mesg);
                        ip_vs_sync_buff_release(sb);
                }
 
-               /* check if entries stay in curr_sb for 2 seconds */
-               sb = get_curr_sync_buff(2 * HZ);
+               /* check if entries stay in ipvs->sync_buff for 2 seconds */
+               sb = get_curr_sync_buff(ipvs, 2 * HZ);
                if (sb) {
                        ip_vs_send_sync_msg(tinfo->sock, sb->mesg);
                        ip_vs_sync_buff_release(sb);
@@ -1462,14 +1465,13 @@ static int sync_thread_master(void *data)
        }
 
        /* clean up the sync_buff queue */
-       while ((sb=sb_dequeue())) {
+       while ((sb = sb_dequeue(ipvs)))
                ip_vs_sync_buff_release(sb);
-       }
 
        /* clean up the current sync_buff */
-       if ((sb = get_curr_sync_buff(0))) {
+       sb = get_curr_sync_buff(ipvs, 0);
+       if (sb)
                ip_vs_sync_buff_release(sb);
-       }
 
        /* release the sending multicast socket */
        sock_release(tinfo->sock);
@@ -1482,11 +1484,12 @@ static int sync_thread_master(void *data)
 static int sync_thread_backup(void *data)
 {
        struct ip_vs_sync_thread_data *tinfo = data;
+       struct netns_ipvs *ipvs = net_ipvs(tinfo->net);
        int len;
 
        pr_info("sync thread started: state = BACKUP, mcast_ifn = %s, "
                "syncid = %d\n",
-               ip_vs_backup_mcast_ifn, ip_vs_backup_syncid);
+               ipvs->backup_mcast_ifn, ipvs->backup_syncid);
 
        while (!kthread_should_stop()) {
                wait_event_interruptible(*sk_sleep(tinfo->sock->sk),
@@ -1496,7 +1499,7 @@ static int sync_thread_backup(void *data)
                /* do we have data now? */
                while (!skb_queue_empty(&(tinfo->sock->sk->sk_receive_queue))) {
                        len = ip_vs_receive(tinfo->sock, tinfo->buf,
-                                       sync_recv_mesg_maxlen);
+                                       ipvs->recv_mesg_maxlen);
                        if (len <= 0) {
                                pr_err("receiving message error\n");
                                break;
@@ -1505,7 +1508,7 @@ static int sync_thread_backup(void *data)
                        /* disable bottom half, because it accesses the data
                           shared by softirq while getting/creating conns */
                        local_bh_disable();
-                       ip_vs_process_message(&init_net, tinfo->buf, len);
+                       ip_vs_process_message(tinfo->net, tinfo->buf, len);
                        local_bh_enable();
                }
        }
@@ -1519,11 +1522,12 @@ static int sync_thread_backup(void *data)
 }
 
 
-int start_sync_thread(int state, char *mcast_ifn, __u8 syncid)
+int start_sync_thread(struct net *net, int state, char *mcast_ifn, __u8 syncid)
 {
        struct ip_vs_sync_thread_data *tinfo;
        struct task_struct **realtask, *task;
        struct socket *sock;
+       struct netns_ipvs *ipvs = net_ipvs(net);
        char *name, *buf = NULL;
        int (*threadfn)(void *data);
        int result = -ENOMEM;
@@ -1533,27 +1537,27 @@ int start_sync_thread(int state, char *mcast_ifn, __u8 syncid)
                  sizeof(struct ip_vs_sync_conn_v0));
 
        if (state == IP_VS_STATE_MASTER) {
-               if (sync_master_thread)
+               if (ipvs->master_thread)
                        return -EEXIST;
 
-               strlcpy(ip_vs_master_mcast_ifn, mcast_ifn,
-                       sizeof(ip_vs_master_mcast_ifn));
-               ip_vs_master_syncid = syncid;
-               realtask = &sync_master_thread;
-               name = "ipvs_syncmaster";
+               strlcpy(ipvs->master_mcast_ifn, mcast_ifn,
+                       sizeof(ipvs->master_mcast_ifn));
+               ipvs->master_syncid = syncid;
+               realtask = &ipvs->master_thread;
+               name = "ipvs_master:%d";
                threadfn = sync_thread_master;
-               sock = make_send_sock();
+               sock = make_send_sock(net);
        } else if (state == IP_VS_STATE_BACKUP) {
-               if (sync_backup_thread)
+               if (ipvs->backup_thread)
                        return -EEXIST;
 
-               strlcpy(ip_vs_backup_mcast_ifn, mcast_ifn,
-                       sizeof(ip_vs_backup_mcast_ifn));
-               ip_vs_backup_syncid = syncid;
-               realtask = &sync_backup_thread;
-               name = "ipvs_syncbackup";
+               strlcpy(ipvs->backup_mcast_ifn, mcast_ifn,
+                       sizeof(ipvs->backup_mcast_ifn));
+               ipvs->backup_syncid = syncid;
+               realtask = &ipvs->backup_thread;
+               name = "ipvs_backup:%d";
                threadfn = sync_thread_backup;
-               sock = make_receive_sock();
+               sock = make_receive_sock(net);
        } else {
                return -EINVAL;
        }
@@ -1563,9 +1567,9 @@ int start_sync_thread(int state, char *mcast_ifn, __u8 syncid)
                goto out;
        }
 
-       set_sync_mesg_maxlen(state);
+       set_sync_mesg_maxlen(net, state);
        if (state == IP_VS_STATE_BACKUP) {
-               buf = kmalloc(sync_recv_mesg_maxlen, GFP_KERNEL);
+               buf = kmalloc(ipvs->recv_mesg_maxlen, GFP_KERNEL);
                if (!buf)
                        goto outsocket;
        }
@@ -1574,10 +1578,11 @@ int start_sync_thread(int state, char *mcast_ifn, __u8 syncid)
        if (!tinfo)
                goto outbuf;
 
+       tinfo->net = net;
        tinfo->sock = sock;
        tinfo->buf = buf;
 
-       task = kthread_run(threadfn, tinfo, name);
+       task = kthread_run(threadfn, tinfo, name, ipvs->gen);
        if (IS_ERR(task)) {
                result = PTR_ERR(task);
                goto outtinfo;
@@ -1585,7 +1590,7 @@ int start_sync_thread(int state, char *mcast_ifn, __u8 syncid)
 
        /* mark as active */
        *realtask = task;
-       ip_vs_sync_state |= state;
+       ipvs->sync_state |= state;
 
        /* increase the module use count */
        ip_vs_use_count_inc();
@@ -1603,16 +1608,18 @@ out:
 }
 
 
-int stop_sync_thread(int state)
+int stop_sync_thread(struct net *net, int state)
 {
+       struct netns_ipvs *ipvs = net_ipvs(net);
+
        IP_VS_DBG(7, "%s(): pid %d\n", __func__, task_pid_nr(current));
 
        if (state == IP_VS_STATE_MASTER) {
-               if (!sync_master_thread)
+               if (!ipvs->master_thread)
                        return -ESRCH;
 
                pr_info("stopping master sync thread %d ...\n",
-                       task_pid_nr(sync_master_thread));
+                       task_pid_nr(ipvs->master_thread));
 
                /*
                 * The lock synchronizes with sb_queue_tail(), so that we don't
@@ -1620,21 +1627,21 @@ int stop_sync_thread(int state)
                 * progress of stopping the master sync daemon.
                 */
 
-               spin_lock_bh(&ip_vs_sync_lock);
-               ip_vs_sync_state &= ~IP_VS_STATE_MASTER;
-               spin_unlock_bh(&ip_vs_sync_lock);
-               kthread_stop(sync_master_thread);
-               sync_master_thread = NULL;
+               spin_lock_bh(&ipvs->sync_lock);
+               ipvs->sync_state &= ~IP_VS_STATE_MASTER;
+               spin_unlock_bh(&ipvs->sync_lock);
+               kthread_stop(ipvs->master_thread);
+               ipvs->master_thread = NULL;
        } else if (state == IP_VS_STATE_BACKUP) {
-               if (!sync_backup_thread)
+               if (!ipvs->backup_thread)
                        return -ESRCH;
 
                pr_info("stopping backup sync thread %d ...\n",
-                       task_pid_nr(sync_backup_thread));
+                       task_pid_nr(ipvs->backup_thread));
 
-               ip_vs_sync_state &= ~IP_VS_STATE_BACKUP;
-               kthread_stop(sync_backup_thread);
-               sync_backup_thread = NULL;
+               ipvs->sync_state &= ~IP_VS_STATE_BACKUP;
+               kthread_stop(ipvs->backup_thread);
+               ipvs->backup_thread = NULL;
        } else {
                return -EINVAL;
        }
@@ -1650,12 +1657,29 @@ int stop_sync_thread(int state)
  */
 static int __net_init __ip_vs_sync_init(struct net *net)
 {
+       struct netns_ipvs *ipvs = net_ipvs(net);
+
+       if (!net_eq(net, &init_net))    /* netns not enabled yet */
+               return -EPERM;
+
+       INIT_LIST_HEAD(&ipvs->sync_queue);
+       spin_lock_init(&ipvs->sync_lock);
+       spin_lock_init(&ipvs->sync_buff_lock);
+
+       ipvs->sync_mcast_addr.sin_family = AF_INET;
+       ipvs->sync_mcast_addr.sin_port = cpu_to_be16(IP_VS_SYNC_PORT);
+       ipvs->sync_mcast_addr.sin_addr.s_addr = cpu_to_be32(IP_VS_SYNC_GROUP);
        return 0;
 }
 
 static void __ip_vs_sync_cleanup(struct net *net)
 {
+       if (!net_eq(net, &init_net))    /* netns not enabled yet */
+               return;
+       stop_sync_thread(net, IP_VS_STATE_MASTER);
+       stop_sync_thread(net, IP_VS_STATE_BACKUP);
 }
+
 static struct pernet_operations ipvs_sync_ops = {
        .init = __ip_vs_sync_init,
        .exit = __ip_vs_sync_cleanup,