diff --git a/include/net/ip_vs.h b/include/net/ip_vs.h index 4e3731ee4eac..9b9ca87a4210 100644 --- a/include/net/ip_vs.h +++ b/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); diff --git a/include/uapi/linux/ip_vs.h b/include/uapi/linux/ip_vs.h index 3199243f2028..391395c06c7e 100644 --- a/include/uapi/linux/ip_vs.h +++ b/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, }; diff --git a/net/netfilter/ipvs/Kconfig b/net/netfilter/ipvs/Kconfig index 3b6929dec748..b32fb0dbe237 100644 --- a/net/netfilter/ipvs/Kconfig +++ b/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--- diff --git a/net/netfilter/ipvs/Makefile b/net/netfilter/ipvs/Makefile index 38b2723b2e3d..67f3f4389602 100644 --- a/net/netfilter/ipvs/Makefile +++ b/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 diff --git a/net/netfilter/ipvs/ip_vs_ctl.c b/net/netfilter/ipvs/ip_vs_ctl.c index 24c554201a76..1a23e91d50d8 100644 --- a/net/netfilter/ipvs/ip_vs_ctl.c +++ b/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 @@ nla_put_failure: } 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 @@ nla_put_failure: 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; } diff --git a/net/netfilter/ipvs/ip_vs_ovf.c b/net/netfilter/ipvs/ip_vs_ovf.c new file mode 100644 index 000000000000..f7d62c3b7329 --- /dev/null +++ b/net/netfilter/ipvs/ip_vs_ovf.c @@ -0,0 +1,86 @@ +/* + * IPVS: Overflow-Connection Scheduling module + * + * Authors: Raducu Deaconu + * + * 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 +#include + +#include + +/* 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"); diff --git a/net/netfilter/ipvs/ip_vs_sync.c b/net/netfilter/ipvs/ip_vs_sync.c index d99ad93eb855..43f140950075 100644 --- a/net/netfilter/ipvs/ip_vs_sync.c +++ b/net/netfilter/ipvs/ip_vs_sync.c @@ -262,6 +262,11 @@ struct ip_vs_sync_mesg { /* ip_vs_sync_conn entries start here */ }; +union ipvs_sockaddr { + struct sockaddr_in in; + struct sockaddr_in6 in6; +}; + struct ip_vs_sync_buff { struct list_head list; unsigned long firstuse; @@ -320,26 +325,28 @@ sb_dequeue(struct netns_ipvs *ipvs, struct ipvs_master_sync_state *ms) * Create a new sync buffer for Version 1 proto. */ static inline struct ip_vs_sync_buff * -ip_vs_sync_buff_create(struct netns_ipvs *ipvs) +ip_vs_sync_buff_create(struct netns_ipvs *ipvs, unsigned int len) { struct ip_vs_sync_buff *sb; if (!(sb=kmalloc(sizeof(struct ip_vs_sync_buff), GFP_ATOMIC))) return NULL; - sb->mesg = kmalloc(ipvs->send_mesg_maxlen, GFP_ATOMIC); + len = max_t(unsigned int, len + sizeof(struct ip_vs_sync_mesg), + ipvs->mcfg.sync_maxlen); + sb->mesg = kmalloc(len, GFP_ATOMIC); if (!sb->mesg) { kfree(sb); return NULL; } sb->mesg->reserved = 0; /* old nr_conns i.e. must be zero now */ sb->mesg->version = SYNC_PROTO_VER; - sb->mesg->syncid = ipvs->master_syncid; + sb->mesg->syncid = ipvs->mcfg.syncid; sb->mesg->size = htons(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 + ipvs->send_mesg_maxlen; + sb->end = (unsigned char *)sb->mesg + len; sb->firstuse = jiffies; return sb; @@ -402,7 +409,7 @@ select_master_thread_id(struct netns_ipvs *ipvs, struct ip_vs_conn *cp) * Create a new sync buffer for Version 0 proto. */ static inline struct ip_vs_sync_buff * -ip_vs_sync_buff_create_v0(struct netns_ipvs *ipvs) +ip_vs_sync_buff_create_v0(struct netns_ipvs *ipvs, unsigned int len) { struct ip_vs_sync_buff *sb; struct ip_vs_sync_mesg_v0 *mesg; @@ -410,17 +417,19 @@ ip_vs_sync_buff_create_v0(struct netns_ipvs *ipvs) if (!(sb=kmalloc(sizeof(struct ip_vs_sync_buff), GFP_ATOMIC))) return NULL; - sb->mesg = kmalloc(ipvs->send_mesg_maxlen, GFP_ATOMIC); + len = max_t(unsigned int, len + sizeof(struct ip_vs_sync_mesg_v0), + ipvs->mcfg.sync_maxlen); + sb->mesg = kmalloc(len, GFP_ATOMIC); if (!sb->mesg) { kfree(sb); return NULL; } mesg = (struct ip_vs_sync_mesg_v0 *)sb->mesg; mesg->nr_conns = 0; - mesg->syncid = ipvs->master_syncid; + mesg->syncid = ipvs->mcfg.syncid; mesg->size = htons(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->end = (unsigned char *)mesg + len; sb->firstuse = jiffies; return sb; } @@ -533,7 +542,7 @@ static void ip_vs_sync_conn_v0(struct net *net, struct ip_vs_conn *cp, struct ip_vs_sync_buff *buff; struct ipvs_master_sync_state *ms; int id; - int len; + unsigned int len; if (unlikely(cp->af != AF_INET)) return; @@ -553,17 +562,19 @@ static void ip_vs_sync_conn_v0(struct net *net, struct ip_vs_conn *cp, id = select_master_thread_id(ipvs, cp); ms = &ipvs->ms[id]; buff = ms->sync_buff; + len = (cp->flags & IP_VS_CONN_F_SEQ_MASK) ? FULL_CONN_SIZE : + SIMPLE_CONN_SIZE; if (buff) { m = (struct ip_vs_sync_mesg_v0 *) buff->mesg; /* Send buffer if it is for v1 */ - if (!m->nr_conns) { + if (buff->head + len > buff->end || !m->nr_conns) { sb_queue_tail(ipvs, ms); ms->sync_buff = NULL; buff = NULL; } } if (!buff) { - buff = ip_vs_sync_buff_create_v0(ipvs); + buff = ip_vs_sync_buff_create_v0(ipvs, len); if (!buff) { spin_unlock_bh(&ipvs->sync_buff_lock); pr_err("ip_vs_sync_buff_create failed.\n"); @@ -572,8 +583,6 @@ static void ip_vs_sync_conn_v0(struct net *net, struct ip_vs_conn *cp, ms->sync_buff = buff; } - len = (cp->flags & IP_VS_CONN_F_SEQ_MASK) ? FULL_CONN_SIZE : - SIMPLE_CONN_SIZE; m = (struct ip_vs_sync_mesg_v0 *) buff->mesg; s = (struct ip_vs_sync_conn_v0 *) buff->head; @@ -597,12 +606,6 @@ static void ip_vs_sync_conn_v0(struct net *net, struct ip_vs_conn *cp, m->nr_conns++; m->size = htons(ntohs(m->size) + len); buff->head += len; - - /* check if there is a space for next one */ - if (buff->head + FULL_CONN_SIZE > buff->end) { - sb_queue_tail(ipvs, ms); - ms->sync_buff = NULL; - } spin_unlock_bh(&ipvs->sync_buff_lock); /* synchronize its controller if it has */ @@ -694,7 +697,7 @@ sloop: } if (!buff) { - buff = ip_vs_sync_buff_create(ipvs); + buff = ip_vs_sync_buff_create(ipvs, len); if (!buff) { spin_unlock_bh(&ipvs->sync_buff_lock); pr_err("ip_vs_sync_buff_create failed.\n"); @@ -1219,7 +1222,7 @@ static void ip_vs_process_message(struct net *net, __u8 *buffer, return; } /* SyncID sanity check */ - if (ipvs->backup_syncid != 0 && m2->syncid != ipvs->backup_syncid) { + if (ipvs->bcfg.syncid != 0 && m2->syncid != ipvs->bcfg.syncid) { IP_VS_DBG(7, "BACKUP, Ignoring syncid = %d\n", m2->syncid); return; } @@ -1303,6 +1306,14 @@ static void set_mcast_loop(struct sock *sk, u_char loop) /* setsockopt(sock, SOL_IP, IP_MULTICAST_LOOP, &loop, sizeof(loop)); */ lock_sock(sk); inet->mc_loop = loop ? 1 : 0; +#ifdef CONFIG_IP_VS_IPV6 + if (sk->sk_family == AF_INET6) { + struct ipv6_pinfo *np = inet6_sk(sk); + + /* IPV6_MULTICAST_LOOP */ + np->mc_loop = loop ? 1 : 0; + } +#endif release_sock(sk); } @@ -1316,6 +1327,33 @@ static void set_mcast_ttl(struct sock *sk, u_char ttl) /* setsockopt(sock, SOL_IP, IP_MULTICAST_TTL, &ttl, sizeof(ttl)); */ lock_sock(sk); inet->mc_ttl = ttl; +#ifdef CONFIG_IP_VS_IPV6 + if (sk->sk_family == AF_INET6) { + struct ipv6_pinfo *np = inet6_sk(sk); + + /* IPV6_MULTICAST_HOPS */ + np->mcast_hops = ttl; + } +#endif + release_sock(sk); +} + +/* Control fragmentation of messages */ +static void set_mcast_pmtudisc(struct sock *sk, int val) +{ + struct inet_sock *inet = inet_sk(sk); + + /* setsockopt(sock, SOL_IP, IP_MTU_DISCOVER, &val, sizeof(val)); */ + lock_sock(sk); + inet->pmtudisc = val; +#ifdef CONFIG_IP_VS_IPV6 + if (sk->sk_family == AF_INET6) { + struct ipv6_pinfo *np = inet6_sk(sk); + + /* IPV6_MTU_DISCOVER */ + np->pmtudisc = val; + } +#endif release_sock(sk); } @@ -1338,44 +1376,15 @@ static int set_mcast_if(struct sock *sk, char *ifname) lock_sock(sk); inet->mc_index = dev->ifindex; /* inet->mc_addr = 0; */ - release_sock(sk); +#ifdef CONFIG_IP_VS_IPV6 + if (sk->sk_family == AF_INET6) { + struct ipv6_pinfo *np = inet6_sk(sk); - return 0; -} - - -/* - * Set the maximum length of sync message according to the - * specified interface's MTU. - */ -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) { - 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; - 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", ipvs->send_mesg_maxlen); - } else if (sync_state == IP_VS_STATE_BACKUP) { - dev = __dev_get_by_name(net, ipvs->backup_mcast_ifn); - if (!dev) - return -ENODEV; - - 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", ipvs->recv_mesg_maxlen); + /* IPV6_MULTICAST_IF */ + np->mcast_oif = dev->ifindex; } +#endif + release_sock(sk); return 0; } @@ -1405,15 +1414,34 @@ join_mcast_group(struct sock *sk, struct in_addr *addr, char *ifname) mreq.imr_ifindex = dev->ifindex; - rtnl_lock(); lock_sock(sk); ret = ip_mc_join_group(sk, &mreq); release_sock(sk); - rtnl_unlock(); return ret; } +#ifdef CONFIG_IP_VS_IPV6 +static int join_mcast_group6(struct sock *sk, struct in6_addr *addr, + char *ifname) +{ + struct net *net = sock_net(sk); + struct net_device *dev; + int ret; + + 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; + + lock_sock(sk); + ret = ipv6_sock_mc_join(sk, dev->ifindex, addr); + release_sock(sk); + + return ret; +} +#endif static int bind_mcastif_addr(struct socket *sock, char *ifname) { @@ -1442,6 +1470,26 @@ static int bind_mcastif_addr(struct socket *sock, char *ifname) return sock->ops->bind(sock, (struct sockaddr*)&sin, sizeof(sin)); } +static void get_mcast_sockaddr(union ipvs_sockaddr *sa, int *salen, + struct ipvs_sync_daemon_cfg *c, int id) +{ + if (AF_INET6 == c->mcast_af) { + sa->in6 = (struct sockaddr_in6) { + .sin6_family = AF_INET6, + .sin6_port = htons(c->mcast_port + id), + }; + sa->in6.sin6_addr = c->mcast_group.in6; + *salen = sizeof(sa->in6); + } else { + sa->in = (struct sockaddr_in) { + .sin_family = AF_INET, + .sin_port = htons(c->mcast_port + id), + }; + sa->in.sin_addr = c->mcast_group.in; + *salen = sizeof(sa->in); + } +} + /* * Set up sending multicast socket over UDP */ @@ -1449,40 +1497,43 @@ static struct socket *make_send_sock(struct net *net, int id) { struct netns_ipvs *ipvs = net_ipvs(net); /* multicast addr */ - struct sockaddr_in mcast_addr = { - .sin_family = AF_INET, - .sin_port = cpu_to_be16(IP_VS_SYNC_PORT + id), - .sin_addr.s_addr = cpu_to_be32(IP_VS_SYNC_GROUP), - }; + union ipvs_sockaddr mcast_addr; struct socket *sock; - int result; + int result, salen; /* First create a socket */ - result = sock_create_kern(net, PF_INET, SOCK_DGRAM, IPPROTO_UDP, &sock); + result = sock_create_kern(net, ipvs->mcfg.mcast_af, SOCK_DGRAM, + IPPROTO_UDP, &sock); if (result < 0) { pr_err("Error during creation of socket; terminating\n"); return ERR_PTR(result); } - result = set_mcast_if(sock->sk, ipvs->master_mcast_ifn); + result = set_mcast_if(sock->sk, ipvs->mcfg.mcast_ifn); if (result < 0) { pr_err("Error setting outbound mcast interface\n"); goto error; } set_mcast_loop(sock->sk, 0); - set_mcast_ttl(sock->sk, 1); + set_mcast_ttl(sock->sk, ipvs->mcfg.mcast_ttl); + /* Allow fragmentation if MTU changes */ + set_mcast_pmtudisc(sock->sk, IP_PMTUDISC_DONT); result = sysctl_sync_sock_size(ipvs); if (result > 0) set_sock_size(sock->sk, 1, result); - result = bind_mcastif_addr(sock, ipvs->master_mcast_ifn); + if (AF_INET == ipvs->mcfg.mcast_af) + result = bind_mcastif_addr(sock, ipvs->mcfg.mcast_ifn); + else + result = 0; if (result < 0) { pr_err("Error binding address of the mcast interface\n"); goto error; } + get_mcast_sockaddr(&mcast_addr, &salen, &ipvs->mcfg, id); result = sock->ops->connect(sock, (struct sockaddr *) &mcast_addr, - sizeof(struct sockaddr), 0); + salen, 0); if (result < 0) { pr_err("Error connecting to the multicast addr\n"); goto error; @@ -1503,16 +1554,13 @@ static struct socket *make_receive_sock(struct net *net, int id) { struct netns_ipvs *ipvs = net_ipvs(net); /* multicast addr */ - struct sockaddr_in mcast_addr = { - .sin_family = AF_INET, - .sin_port = cpu_to_be16(IP_VS_SYNC_PORT + id), - .sin_addr.s_addr = cpu_to_be32(IP_VS_SYNC_GROUP), - }; + union ipvs_sockaddr mcast_addr; struct socket *sock; - int result; + int result, salen; /* First create a socket */ - result = sock_create_kern(net, PF_INET, SOCK_DGRAM, IPPROTO_UDP, &sock); + result = sock_create_kern(net, ipvs->bcfg.mcast_af, SOCK_DGRAM, + IPPROTO_UDP, &sock); if (result < 0) { pr_err("Error during creation of socket; terminating\n"); return ERR_PTR(result); @@ -1523,17 +1571,22 @@ static struct socket *make_receive_sock(struct net *net, int id) if (result > 0) set_sock_size(sock->sk, 0, result); - result = sock->ops->bind(sock, (struct sockaddr *) &mcast_addr, - sizeof(struct sockaddr)); + get_mcast_sockaddr(&mcast_addr, &salen, &ipvs->bcfg, id); + result = sock->ops->bind(sock, (struct sockaddr *)&mcast_addr, salen); if (result < 0) { pr_err("Error binding to the multicast addr\n"); goto error; } /* join the multicast group */ - result = join_mcast_group(sock->sk, - (struct in_addr *) &mcast_addr.sin_addr, - ipvs->backup_mcast_ifn); +#ifdef CONFIG_IP_VS_IPV6 + if (ipvs->bcfg.mcast_af == AF_INET6) + result = join_mcast_group6(sock->sk, &mcast_addr.in6.sin6_addr, + ipvs->bcfg.mcast_ifn); + else +#endif + result = join_mcast_group(sock->sk, &mcast_addr.in.sin_addr, + ipvs->bcfg.mcast_ifn); if (result < 0) { pr_err("Error joining to the multicast group\n"); goto error; @@ -1641,7 +1694,7 @@ static int sync_thread_master(void *data) pr_info("sync thread started: state = MASTER, mcast_ifn = %s, " "syncid = %d, id = %d\n", - ipvs->master_mcast_ifn, ipvs->master_syncid, tinfo->id); + ipvs->mcfg.mcast_ifn, ipvs->mcfg.syncid, tinfo->id); for (;;) { sb = next_sync_buff(ipvs, ms); @@ -1695,7 +1748,7 @@ static int sync_thread_backup(void *data) pr_info("sync thread started: state = BACKUP, mcast_ifn = %s, " "syncid = %d, id = %d\n", - ipvs->backup_mcast_ifn, ipvs->backup_syncid, tinfo->id); + ipvs->bcfg.mcast_ifn, ipvs->bcfg.syncid, tinfo->id); while (!kthread_should_stop()) { wait_event_interruptible(*sk_sleep(tinfo->sock->sk), @@ -1705,7 +1758,7 @@ 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, - ipvs->recv_mesg_maxlen); + ipvs->bcfg.sync_maxlen); if (len <= 0) { if (len != -EAGAIN) pr_err("receiving message error\n"); @@ -1725,16 +1778,19 @@ static int sync_thread_backup(void *data) } -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 *c, + int state) { struct ip_vs_sync_thread_data *tinfo; struct task_struct **array = NULL, *task; struct socket *sock; struct netns_ipvs *ipvs = net_ipvs(net); + struct net_device *dev; char *name; int (*threadfn)(void *data); - int id, count; + int id, count, hlen; int result = -ENOMEM; + u16 mtu, min_mtu; IP_VS_DBG(7, "%s(): pid %d\n", __func__, task_pid_nr(current)); IP_VS_DBG(7, "Each ip_vs_sync_conn entry needs %Zd bytes\n", @@ -1746,22 +1802,46 @@ int start_sync_thread(struct net *net, int state, char *mcast_ifn, __u8 syncid) } else count = ipvs->threads_mask + 1; + if (c->mcast_af == AF_UNSPEC) { + c->mcast_af = AF_INET; + c->mcast_group.ip = cpu_to_be32(IP_VS_SYNC_GROUP); + } + if (!c->mcast_port) + c->mcast_port = IP_VS_SYNC_PORT; + if (!c->mcast_ttl) + c->mcast_ttl = 1; + + dev = __dev_get_by_name(net, c->mcast_ifn); + if (!dev) { + pr_err("Unknown mcast interface: %s\n", c->mcast_ifn); + return -ENODEV; + } + hlen = (AF_INET6 == c->mcast_af) ? + sizeof(struct ipv6hdr) + sizeof(struct udphdr) : + sizeof(struct iphdr) + sizeof(struct udphdr); + mtu = (state == IP_VS_STATE_BACKUP) ? + clamp(dev->mtu, 1500U, 65535U) : 1500U; + min_mtu = (state == IP_VS_STATE_BACKUP) ? 1024 : 1; + + if (c->sync_maxlen) + c->sync_maxlen = clamp_t(unsigned int, + c->sync_maxlen, min_mtu, + 65535 - hlen); + else + c->sync_maxlen = mtu - hlen; + if (state == IP_VS_STATE_MASTER) { if (ipvs->ms) return -EEXIST; - strlcpy(ipvs->master_mcast_ifn, mcast_ifn, - sizeof(ipvs->master_mcast_ifn)); - ipvs->master_syncid = syncid; + ipvs->mcfg = *c; name = "ipvs-m:%d:%d"; threadfn = sync_thread_master; } else if (state == IP_VS_STATE_BACKUP) { if (ipvs->backup_threads) return -EEXIST; - strlcpy(ipvs->backup_mcast_ifn, mcast_ifn, - sizeof(ipvs->backup_mcast_ifn)); - ipvs->backup_syncid = syncid; + ipvs->bcfg = *c; name = "ipvs-b:%d:%d"; threadfn = sync_thread_backup; } else { @@ -1789,7 +1869,6 @@ int start_sync_thread(struct net *net, int state, char *mcast_ifn, __u8 syncid) if (!array) goto out; } - set_sync_mesg_maxlen(net, state); tinfo = NULL; for (id = 0; id < count; id++) { @@ -1807,7 +1886,7 @@ int start_sync_thread(struct net *net, int state, char *mcast_ifn, __u8 syncid) tinfo->net = net; tinfo->sock = sock; if (state == IP_VS_STATE_BACKUP) { - tinfo->buf = kmalloc(ipvs->recv_mesg_maxlen, + tinfo->buf = kmalloc(ipvs->bcfg.sync_maxlen, GFP_KERNEL); if (!tinfo->buf) goto outtinfo;