Merge tag 'ipvs2-for-v4.3' of https://git.kernel.org/pub/scm/linux/kernel/git/horms/ipvs-next

Simon Horman says:

====================
Second Round of IPVS Updates for v4.3

I realise these are a little late in the cycle, so if you would prefer
me to repost them for v4.4 then just let me know.

The updates include:
* A new scheduler from Raducu Deaconu
* Enhanced configurability of the sync daemon from Julian Anastasov
====================
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>

include/net/ip_vs.h

@@ -846,6 +846,17 @@ struct ipvs_master_sync_state {
 /* How much time to keep dests in trash */
 #define IP_VS_DEST_TRASH_PERIOD		(120 * HZ)
 
+struct ipvs_sync_daemon_cfg {
+	union nf_inet_addr	mcast_group;
+	int			syncid;
+	u16			sync_maxlen;
+	u16			mcast_port;
+	u8			mcast_af;
+	u8			mcast_ttl;
+	/* multicast interface name */
+	char			mcast_ifn[IP_VS_IFNAME_MAXLEN];
+};
+
 /* IPVS in network namespace */
 struct netns_ipvs {
 	int			gen;		/* Generation */
@@ -961,15 +972,10 @@ struct netns_ipvs {
 	spinlock_t		sync_buff_lock;
 	struct task_struct	**backup_threads;
 	int			threads_mask;
-	int			send_mesg_maxlen;
-	int			recv_mesg_maxlen;
 	volatile int		sync_state;
-	volatile int		master_syncid;
-	volatile int		backup_syncid;
 	struct mutex		sync_mutex;
-	/* multicast interface name */
-	char			master_mcast_ifn[IP_VS_IFNAME_MAXLEN];
-	char			backup_mcast_ifn[IP_VS_IFNAME_MAXLEN];
+	struct ipvs_sync_daemon_cfg	mcfg;	/* Master Configuration */
+	struct ipvs_sync_daemon_cfg	bcfg;	/* Backup Configuration */
 	/* net name space ptr */
 	struct net		*net;            /* Needed by timer routines */
 	/* Number of heterogeneous destinations, needed becaus heterogeneous
@@ -1408,7 +1414,8 @@ static inline void ip_vs_dest_put_and_free(struct ip_vs_dest *dest)
 /* IPVS sync daemon data and function prototypes
  * (from ip_vs_sync.c)
  */
-int start_sync_thread(struct net *net, int state, char *mcast_ifn, __u8 syncid);
+int start_sync_thread(struct net *net, struct ipvs_sync_daemon_cfg *cfg,
+		      int state);
 int stop_sync_thread(struct net *net, int state);
 void ip_vs_sync_conn(struct net *net, struct ip_vs_conn *cp, int pkts);
 

include/uapi/linux/ip_vs.h

@@ -406,6 +406,11 @@ enum {
 	IPVS_DAEMON_ATTR_STATE,		/* sync daemon state (master/backup) */
 	IPVS_DAEMON_ATTR_MCAST_IFN,	/* multicast interface name */
 	IPVS_DAEMON_ATTR_SYNC_ID,	/* SyncID we belong to */
+	IPVS_DAEMON_ATTR_SYNC_MAXLEN,	/* UDP Payload Size */
+	IPVS_DAEMON_ATTR_MCAST_GROUP,	/* IPv4 Multicast Address */
+	IPVS_DAEMON_ATTR_MCAST_GROUP6,	/* IPv6 Multicast Address */
+	IPVS_DAEMON_ATTR_MCAST_PORT,	/* Multicast Port (base) */
+	IPVS_DAEMON_ATTR_MCAST_TTL,	/* Multicast TTL */
 	__IPVS_DAEMON_ATTR_MAX,
 };
 

net/netfilter/ipvs/Kconfig

@@ -162,6 +162,17 @@ config  IP_VS_FO
 	  If you want to compile it in kernel, say Y. To compile it as a
 	  module, choose M here. If unsure, say N.
 
+config  IP_VS_OVF
+	tristate "weighted overflow scheduling"
+	---help---
+	  The weighted overflow scheduling algorithm directs network
+	  connections to the server with the highest weight that is
+	  currently available and overflows to the next when active
+	  connections exceed the node's weight.
+
+	  If you want to compile it in kernel, say Y. To compile it as a
+	  module, choose M here. If unsure, say N.
+
 config	IP_VS_LBLC
 	tristate "locality-based least-connection scheduling"
 	---help---

net/netfilter/ipvs/Makefile

@@ -27,6 +27,7 @@ obj-$(CONFIG_IP_VS_WRR) += ip_vs_wrr.o
 obj-$(CONFIG_IP_VS_LC) += ip_vs_lc.o
 obj-$(CONFIG_IP_VS_WLC) += ip_vs_wlc.o
 obj-$(CONFIG_IP_VS_FO) += ip_vs_fo.o
+obj-$(CONFIG_IP_VS_OVF) += ip_vs_ovf.o
 obj-$(CONFIG_IP_VS_LBLC) += ip_vs_lblc.o
 obj-$(CONFIG_IP_VS_LBLCR) += ip_vs_lblcr.o
 obj-$(CONFIG_IP_VS_DH) += ip_vs_dh.o

net/netfilter/ipvs/ip_vs_ctl.c

@@ -2335,13 +2335,23 @@ do_ip_vs_set_ctl(struct sock *sk, int cmd, void __user *user, unsigned int len)
 	    cmd == IP_VS_SO_SET_STOPDAEMON) {
 		struct ip_vs_daemon_user *dm = (struct ip_vs_daemon_user *)arg;
 
-		mutex_lock(&ipvs->sync_mutex);
-		if (cmd == IP_VS_SO_SET_STARTDAEMON)
-			ret = start_sync_thread(net, dm->state, dm->mcast_ifn,
-						dm->syncid);
-		else
+		if (cmd == IP_VS_SO_SET_STARTDAEMON) {
+			struct ipvs_sync_daemon_cfg cfg;
+
+			memset(&cfg, 0, sizeof(cfg));
+			strlcpy(cfg.mcast_ifn, dm->mcast_ifn,
+				sizeof(cfg.mcast_ifn));
+			cfg.syncid = dm->syncid;
+			rtnl_lock();
+			mutex_lock(&ipvs->sync_mutex);
+			ret = start_sync_thread(net, &cfg, dm->state);
+			mutex_unlock(&ipvs->sync_mutex);
+			rtnl_unlock();
+		} else {
+			mutex_lock(&ipvs->sync_mutex);
 			ret = stop_sync_thread(net, dm->state);
-		mutex_unlock(&ipvs->sync_mutex);
+			mutex_unlock(&ipvs->sync_mutex);
+		}
 		goto out_dec;
 	}
 
@@ -2645,15 +2655,15 @@ do_ip_vs_get_ctl(struct sock *sk, int cmd, void __user *user, int *len)
 		mutex_lock(&ipvs->sync_mutex);
 		if (ipvs->sync_state & IP_VS_STATE_MASTER) {
 			d[0].state = IP_VS_STATE_MASTER;
-			strlcpy(d[0].mcast_ifn, ipvs->master_mcast_ifn,
+			strlcpy(d[0].mcast_ifn, ipvs->mcfg.mcast_ifn,
 				sizeof(d[0].mcast_ifn));
-			d[0].syncid = ipvs->master_syncid;
+			d[0].syncid = ipvs->mcfg.syncid;
 		}
 		if (ipvs->sync_state & IP_VS_STATE_BACKUP) {
 			d[1].state = IP_VS_STATE_BACKUP;
-			strlcpy(d[1].mcast_ifn, ipvs->backup_mcast_ifn,
+			strlcpy(d[1].mcast_ifn, ipvs->bcfg.mcast_ifn,
 				sizeof(d[1].mcast_ifn));
-			d[1].syncid = ipvs->backup_syncid;
+			d[1].syncid = ipvs->bcfg.syncid;
 		}
 		if (copy_to_user(user, &d, sizeof(d)) != 0)
 			ret = -EFAULT;
@@ -2808,6 +2818,11 @@ static const struct nla_policy ip_vs_daemon_policy[IPVS_DAEMON_ATTR_MAX + 1] = {
 	[IPVS_DAEMON_ATTR_MCAST_IFN]	= { .type = NLA_NUL_STRING,
 					    .len = IP_VS_IFNAME_MAXLEN },
 	[IPVS_DAEMON_ATTR_SYNC_ID]	= { .type = NLA_U32 },
+	[IPVS_DAEMON_ATTR_SYNC_MAXLEN]	= { .type = NLA_U16 },
+	[IPVS_DAEMON_ATTR_MCAST_GROUP]	= { .type = NLA_U32 },
+	[IPVS_DAEMON_ATTR_MCAST_GROUP6]	= { .len = sizeof(struct in6_addr) },
+	[IPVS_DAEMON_ATTR_MCAST_PORT]	= { .type = NLA_U16 },
+	[IPVS_DAEMON_ATTR_MCAST_TTL]	= { .type = NLA_U8 },
 };
 
 /* Policy used for attributes in nested attribute IPVS_CMD_ATTR_SERVICE */
@@ -3266,7 +3281,7 @@ static int ip_vs_genl_parse_dest(struct ip_vs_dest_user_kern *udest,
 }
 
 static int ip_vs_genl_fill_daemon(struct sk_buff *skb, __u32 state,
-				  const char *mcast_ifn, __u32 syncid)
+				  struct ipvs_sync_daemon_cfg *c)
 {
 	struct nlattr *nl_daemon;
 
@@ -3275,9 +3290,23 @@ static int ip_vs_genl_fill_daemon(struct sk_buff *skb, __u32 state,
 		return -EMSGSIZE;
 
 	if (nla_put_u32(skb, IPVS_DAEMON_ATTR_STATE, state) ||
-	    nla_put_string(skb, IPVS_DAEMON_ATTR_MCAST_IFN, mcast_ifn) ||
-	    nla_put_u32(skb, IPVS_DAEMON_ATTR_SYNC_ID, syncid))
+	    nla_put_string(skb, IPVS_DAEMON_ATTR_MCAST_IFN, c->mcast_ifn) ||
+	    nla_put_u32(skb, IPVS_DAEMON_ATTR_SYNC_ID, c->syncid) ||
+	    nla_put_u16(skb, IPVS_DAEMON_ATTR_SYNC_MAXLEN, c->sync_maxlen) ||
+	    nla_put_u16(skb, IPVS_DAEMON_ATTR_MCAST_PORT, c->mcast_port) ||
+	    nla_put_u8(skb, IPVS_DAEMON_ATTR_MCAST_TTL, c->mcast_ttl))
 		goto nla_put_failure;
+#ifdef CONFIG_IP_VS_IPV6
+	if (c->mcast_af == AF_INET6) {
+		if (nla_put_in6_addr(skb, IPVS_DAEMON_ATTR_MCAST_GROUP6,
+				     &c->mcast_group.in6))
+			goto nla_put_failure;
+	} else
+#endif
+		if (c->mcast_af == AF_INET &&
+		    nla_put_in_addr(skb, IPVS_DAEMON_ATTR_MCAST_GROUP,
+				    c->mcast_group.ip))
+			goto nla_put_failure;
 	nla_nest_end(skb, nl_daemon);
 
 	return 0;
@@ -3288,7 +3317,7 @@ static int ip_vs_genl_fill_daemon(struct sk_buff *skb, __u32 state,
 }
 
 static int ip_vs_genl_dump_daemon(struct sk_buff *skb, __u32 state,
-				  const char *mcast_ifn, __u32 syncid,
+				  struct ipvs_sync_daemon_cfg *c,
 				  struct netlink_callback *cb)
 {
 	void *hdr;
@@ -3298,7 +3327,7 @@ static int ip_vs_genl_dump_daemon(struct sk_buff *skb, __u32 state,
 	if (!hdr)
 		return -EMSGSIZE;
 
-	if (ip_vs_genl_fill_daemon(skb, state, mcast_ifn, syncid))
+	if (ip_vs_genl_fill_daemon(skb, state, c))
 		goto nla_put_failure;
 
 	genlmsg_end(skb, hdr);
@@ -3318,8 +3347,7 @@ static int ip_vs_genl_dump_daemons(struct sk_buff *skb,
 	mutex_lock(&ipvs->sync_mutex);
 	if ((ipvs->sync_state & IP_VS_STATE_MASTER) && !cb->args[0]) {
 		if (ip_vs_genl_dump_daemon(skb, IP_VS_STATE_MASTER,
-					   ipvs->master_mcast_ifn,
-					   ipvs->master_syncid, cb) < 0)
+					   &ipvs->mcfg, cb) < 0)
 			goto nla_put_failure;
 
 		cb->args[0] = 1;
@@ -3327,8 +3355,7 @@ static int ip_vs_genl_dump_daemons(struct sk_buff *skb,
 
 	if ((ipvs->sync_state & IP_VS_STATE_BACKUP) && !cb->args[1]) {
 		if (ip_vs_genl_dump_daemon(skb, IP_VS_STATE_BACKUP,
-					   ipvs->backup_mcast_ifn,
-					   ipvs->backup_syncid, cb) < 0)
+					   &ipvs->bcfg, cb) < 0)
 			goto nla_put_failure;
 
 		cb->args[1] = 1;
@@ -3342,30 +3369,83 @@ static int ip_vs_genl_dump_daemons(struct sk_buff *skb,
 
 static int ip_vs_genl_new_daemon(struct net *net, struct nlattr **attrs)
 {
+	struct netns_ipvs *ipvs = net_ipvs(net);
+	struct ipvs_sync_daemon_cfg c;
+	struct nlattr *a;
+	int ret;
+
+	memset(&c, 0, sizeof(c));
 	if (!(attrs[IPVS_DAEMON_ATTR_STATE] &&
 	      attrs[IPVS_DAEMON_ATTR_MCAST_IFN] &&
 	      attrs[IPVS_DAEMON_ATTR_SYNC_ID]))
 		return -EINVAL;
+	strlcpy(c.mcast_ifn, nla_data(attrs[IPVS_DAEMON_ATTR_MCAST_IFN]),
+		sizeof(c.mcast_ifn));
+	c.syncid = nla_get_u32(attrs[IPVS_DAEMON_ATTR_SYNC_ID]);
+
+	a = attrs[IPVS_DAEMON_ATTR_SYNC_MAXLEN];
+	if (a)
+		c.sync_maxlen = nla_get_u16(a);
+
+	a = attrs[IPVS_DAEMON_ATTR_MCAST_GROUP];
+	if (a) {
+		c.mcast_af = AF_INET;
+		c.mcast_group.ip = nla_get_in_addr(a);
+		if (!ipv4_is_multicast(c.mcast_group.ip))
+			return -EINVAL;
+	} else {
+		a = attrs[IPVS_DAEMON_ATTR_MCAST_GROUP6];
+		if (a) {
+#ifdef CONFIG_IP_VS_IPV6
+			int addr_type;
+
+			c.mcast_af = AF_INET6;
+			c.mcast_group.in6 = nla_get_in6_addr(a);
+			addr_type = ipv6_addr_type(&c.mcast_group.in6);
+			if (!(addr_type & IPV6_ADDR_MULTICAST))
+				return -EINVAL;
+#else
+			return -EAFNOSUPPORT;
+#endif
+		}
+	}
+
+	a = attrs[IPVS_DAEMON_ATTR_MCAST_PORT];
+	if (a)
+		c.mcast_port = nla_get_u16(a);
+
+	a = attrs[IPVS_DAEMON_ATTR_MCAST_TTL];
+	if (a)
+		c.mcast_ttl = nla_get_u8(a);
 
 	/* The synchronization protocol is incompatible with mixed family
 	 * services
 	 */
-	if (net_ipvs(net)->mixed_address_family_dests > 0)
+	if (ipvs->mixed_address_family_dests > 0)
 		return -EINVAL;
 
-	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]));
+	rtnl_lock();
+	mutex_lock(&ipvs->sync_mutex);
+	ret = start_sync_thread(net, &c,
+				nla_get_u32(attrs[IPVS_DAEMON_ATTR_STATE]));
+	mutex_unlock(&ipvs->sync_mutex);
+	rtnl_unlock();
+	return ret;
 }
 
 static int ip_vs_genl_del_daemon(struct net *net, struct nlattr **attrs)
 {
+	struct netns_ipvs *ipvs = net_ipvs(net);
+	int ret;
+
 	if (!attrs[IPVS_DAEMON_ATTR_STATE])
 		return -EINVAL;
 
-	return stop_sync_thread(net,
-				nla_get_u32(attrs[IPVS_DAEMON_ATTR_STATE]));
+	mutex_lock(&ipvs->sync_mutex);
+	ret = stop_sync_thread(net,
+			       nla_get_u32(attrs[IPVS_DAEMON_ATTR_STATE]));
+	mutex_unlock(&ipvs->sync_mutex);
+	return ret;
 }
 
 static int ip_vs_genl_set_config(struct net *net, struct nlattr **attrs)
@@ -3389,7 +3469,7 @@ static int ip_vs_genl_set_config(struct net *net, struct nlattr **attrs)
 
 static int ip_vs_genl_set_daemon(struct sk_buff *skb, struct genl_info *info)
 {
-	int ret = 0, cmd;
+	int ret = -EINVAL, cmd;
 	struct net *net;
 	struct netns_ipvs *ipvs;
 
@@ -3400,22 +3480,19 @@ static int ip_vs_genl_set_daemon(struct sk_buff *skb, struct genl_info *info)
 	if (cmd == IPVS_CMD_NEW_DAEMON || cmd == IPVS_CMD_DEL_DAEMON) {
 		struct nlattr *daemon_attrs[IPVS_DAEMON_ATTR_MAX + 1];
 
-		mutex_lock(&ipvs->sync_mutex);
 		if (!info->attrs[IPVS_CMD_ATTR_DAEMON] ||
 		    nla_parse_nested(daemon_attrs, IPVS_DAEMON_ATTR_MAX,
 				     info->attrs[IPVS_CMD_ATTR_DAEMON],
-				     ip_vs_daemon_policy)) {
-			ret = -EINVAL;
+				     ip_vs_daemon_policy))
 			goto out;
-		}
 
 		if (cmd == IPVS_CMD_NEW_DAEMON)
 			ret = ip_vs_genl_new_daemon(net, daemon_attrs);
 		else
 			ret = ip_vs_genl_del_daemon(net, daemon_attrs);
-out:
-		mutex_unlock(&ipvs->sync_mutex);
 	}
+
+out:
 	return ret;
 }
 

net/netfilter/ipvs/ip_vs_ovf.c

+/*
+ * IPVS:        Overflow-Connection Scheduling module
+ *
+ * Authors:     Raducu Deaconu <rhadoo_io@yahoo.com>
+ *
+ *              This program is free software; you can redistribute it and/or
+ *              modify it under the terms of the GNU General Public License
+ *              as published by the Free Software Foundation; either version
+ *              2 of the License, or (at your option) any later version.
+ *
+ * Scheduler implements "overflow" loadbalancing according to number of active
+ * connections , will keep all conections to the node with the highest weight
+ * and overflow to the next node if the number of connections exceeds the node's
+ * weight.
+ * Note that this scheduler might not be suitable for UDP because it only uses
+ * active connections
+ *
+ */
+
+#define KMSG_COMPONENT "IPVS"
+#define pr_fmt(fmt) KMSG_COMPONENT ": " fmt
+
+#include <linux/module.h>
+#include <linux/kernel.h>
+
+#include <net/ip_vs.h>
+
+/* OVF Connection scheduling  */
+static struct ip_vs_dest *
+ip_vs_ovf_schedule(struct ip_vs_service *svc, const struct sk_buff *skb,
+		   struct ip_vs_iphdr *iph)
+{
+	struct ip_vs_dest *dest, *h = NULL;
+	int hw = 0, w;
+
+	IP_VS_DBG(6, "ip_vs_ovf_schedule(): Scheduling...\n");
+	/* select the node with highest weight, go to next in line if active
+	* connections exceed weight
+	*/
+	list_for_each_entry_rcu(dest, &svc->destinations, n_list) {
+		w = atomic_read(&dest->weight);
+		if ((dest->flags & IP_VS_DEST_F_OVERLOAD) ||
+		    atomic_read(&dest->activeconns) > w ||
+		    w == 0)
+			continue;
+		if (!h || w > hw) {
+			h = dest;
+			hw = w;
+		}
+	}
+
+	if (h) {
+		IP_VS_DBG_BUF(6, "OVF: server %s:%u active %d w %d\n",
+			      IP_VS_DBG_ADDR(h->af, &h->addr),
+			      ntohs(h->port),
+			      atomic_read(&h->activeconns),
+			      atomic_read(&h->weight));
+		return h;
+	}
+
+	ip_vs_scheduler_err(svc, "no destination available");
+	return NULL;
+}
+
+static struct ip_vs_scheduler ip_vs_ovf_scheduler = {
+	.name =			"ovf",
+	.refcnt =		ATOMIC_INIT(0),
+	.module =		THIS_MODULE,
+	.n_list =		LIST_HEAD_INIT(ip_vs_ovf_scheduler.n_list),
+	.schedule =		ip_vs_ovf_schedule,
+};
+
+static int __init ip_vs_ovf_init(void)
+{
+	return register_ip_vs_scheduler(&ip_vs_ovf_scheduler);
+}
+
+static void __exit ip_vs_ovf_cleanup(void)
+{
+	unregister_ip_vs_scheduler(&ip_vs_ovf_scheduler);
+	synchronize_rcu();
+}
+
+module_init(ip_vs_ovf_init);
+module_exit(ip_vs_ovf_cleanup);
+MODULE_LICENSE("GPL");