29 files changed, 272 insertions, 5 deletions

diff --git a/net/ipv4/Kconfig b/net/ipv4/Kconfig
index 591ea23639ca..691268f3a359 100644
--- a/net/ipv4/Kconfig
+++ b/net/ipv4/Kconfig
@@ -630,5 +630,3 @@ config TCP_MD5SIG
 
           If unsure, say N.
 
-source "net/ipv4/ipvs/Kconfig"
-
diff --git a/net/ipv4/Makefile b/net/ipv4/Makefile
index ad40ef3f9ebc..80ff87ce43aa 100644
--- a/net/ipv4/Makefile
+++ b/net/ipv4/Makefile
@@ -33,7 +33,6 @@ obj-$(CONFIG_INET_XFRM_MODE_TRANSPORT) += xfrm4_mode_transport.o
 obj-$(CONFIG_INET_XFRM_MODE_TUNNEL) += xfrm4_mode_tunnel.o
 obj-$(CONFIG_IP_PNP) += ipconfig.o
 obj-$(CONFIG_NETFILTER) += netfilter.o netfilter/
-obj-$(CONFIG_IP_VS) += ipvs/
 obj-$(CONFIG_INET_DIAG) += inet_diag.o 
 obj-$(CONFIG_INET_TCP_DIAG) += tcp_diag.o
 obj-$(CONFIG_NET_TCPPROBE) += tcp_probe.o
diff --git a/net/netfilter/Kconfig b/net/netfilter/Kconfig
index ee898e74808d..73f9378e1bfe 100644
--- a/net/netfilter/Kconfig
+++ b/net/netfilter/Kconfig
@@ -838,3 +838,5 @@ config NETFILTER_XT_MATCH_HASHLIMIT
 
 endmenu
 
+source "net/netfilter/ipvs/Kconfig"
+
diff --git a/net/netfilter/Makefile b/net/netfilter/Makefile
index 3bd2cc556aea..cf75055be830 100644
--- a/net/netfilter/Makefile
+++ b/net/netfilter/Makefile
@@ -83,3 +83,6 @@ obj-$(CONFIG_NETFILTER_XT_MATCH_STRING) += xt_string.o
 obj-$(CONFIG_NETFILTER_XT_MATCH_TCPMSS) += xt_tcpmss.o
 obj-$(CONFIG_NETFILTER_XT_MATCH_TIME) += xt_time.o
 obj-$(CONFIG_NETFILTER_XT_MATCH_U32) += xt_u32.o
+
+# IPVS
+obj-$(CONFIG_IP_VS) += ipvs/
diff --git a/net/ipv4/ipvs/Kconfig b/net/netfilter/ipvs/Kconfig
index de6004de80bc..de6004de80bc 100644
--- a/net/ipv4/ipvs/Kconfig
+++ b/net/netfilter/ipvs/Kconfig
diff --git a/net/ipv4/ipvs/Makefile b/net/netfilter/ipvs/Makefile
index 73a46fe1fe4c..73a46fe1fe4c 100644
--- a/net/ipv4/ipvs/Makefile
+++ b/net/netfilter/ipvs/Makefile
diff --git a/net/ipv4/ipvs/ip_vs_app.c b/net/netfilter/ipvs/ip_vs_app.c
index 201b8ea3020d..201b8ea3020d 100644
--- a/net/ipv4/ipvs/ip_vs_app.c
+++ b/net/netfilter/ipvs/ip_vs_app.c
diff --git a/net/ipv4/ipvs/ip_vs_conn.c b/net/netfilter/ipvs/ip_vs_conn.c
index 9a24332fbed8..9a24332fbed8 100644
--- a/net/ipv4/ipvs/ip_vs_conn.c
+++ b/net/netfilter/ipvs/ip_vs_conn.c
diff --git a/net/ipv4/ipvs/ip_vs_core.c b/net/netfilter/ipvs/ip_vs_core.c
index 958abf3e5f8c..958abf3e5f8c 100644
--- a/net/ipv4/ipvs/ip_vs_core.c
+++ b/net/netfilter/ipvs/ip_vs_core.c
diff --git a/net/ipv4/ipvs/ip_vs_ctl.c b/net/netfilter/ipvs/ip_vs_ctl.c
index 771551d8fba9..0302cf3e5039 100644
--- a/net/ipv4/ipvs/ip_vs_ctl.c
+++ b/net/netfilter/ipvs/ip_vs_ctl.c
@@ -1330,7 +1330,9 @@ ip_vs_edit_service(struct ip_vs_service *svc, struct ip_vs_service_user_kern *u)
 
   out_unlock:
         write_unlock_bh(&__ip_vs_svc_lock);
+#ifdef CONFIG_IP_VS_IPV6
   out:
+#endif
 
         if (old_sched)
                 ip_vs_scheduler_put(old_sched);
diff --git a/net/netfilter/ipvs/ip_vs_dh.c b/net/netfilter/ipvs/ip_vs_dh.c
new file mode 100644
index 000000000000..a16943fd72f1
--- /dev/null
+++ b/net/netfilter/ipvs/ip_vs_dh.c
@@ -0,0 +1,261 @@
+/*
+ * IPVS:        Destination Hashing scheduling module
+ *
+ * Authors:     Wensong Zhang <wensong@gnuchina.org>
+ *
+ *              Inspired by the consistent hashing scheduler patch from
+ *              Thomas Proell <proellt@gmx.de>
+ *
+ *              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.
+ *
+ * Changes:
+ *
+ */
+
+/*
+ * The dh algorithm is to select server by the hash key of destination IP
+ * address. The pseudo code is as follows:
+ *
+ *       n <- servernode[dest_ip];
+ *       if (n is dead) OR
+ *          (n is overloaded) OR (n.weight <= 0) then
+ *                 return NULL;
+ *
+ *       return n;
+ *
+ * Notes that servernode is a 256-bucket hash table that maps the hash
+ * index derived from packet destination IP address to the current server
+ * array. If the dh scheduler is used in cache cluster, it is good to
+ * combine it with cache_bypass feature. When the statically assigned
+ * server is dead or overloaded, the load balancer can bypass the cache
+ * server and send requests to the original server directly.
+ *
+ */
+
+#include <linux/ip.h>
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/skbuff.h>
+
+#include <net/ip_vs.h>
+
+
+/*
+ *      IPVS DH bucket
+ */
+struct ip_vs_dh_bucket {
+        struct ip_vs_dest       *dest;          /* real server (cache) */
+};
+
+/*
+ *     for IPVS DH entry hash table
+ */
+#ifndef CONFIG_IP_VS_DH_TAB_BITS
+#define CONFIG_IP_VS_DH_TAB_BITS        8
+#endif
+#define IP_VS_DH_TAB_BITS               CONFIG_IP_VS_DH_TAB_BITS
+#define IP_VS_DH_TAB_SIZE               (1 << IP_VS_DH_TAB_BITS)
+#define IP_VS_DH_TAB_MASK               (IP_VS_DH_TAB_SIZE - 1)
+
+
+/*
+ *      Returns hash value for IPVS DH entry
+ */
+static inline unsigned ip_vs_dh_hashkey(__be32 addr)
+{
+        return (ntohl(addr)*2654435761UL) & IP_VS_DH_TAB_MASK;
+}
+
+
+/*
+ *      Get ip_vs_dest associated with supplied parameters.
+ */
+static inline struct ip_vs_dest *
+ip_vs_dh_get(struct ip_vs_dh_bucket *tbl, __be32 addr)
+{
+        return (tbl[ip_vs_dh_hashkey(addr)]).dest;
+}
+
+
+/*
+ *      Assign all the hash buckets of the specified table with the service.
+ */
+static int
+ip_vs_dh_assign(struct ip_vs_dh_bucket *tbl, struct ip_vs_service *svc)
+{
+        int i;
+        struct ip_vs_dh_bucket *b;
+        struct list_head *p;
+        struct ip_vs_dest *dest;
+
+        b = tbl;
+        p = &svc->destinations;
+        for (i=0; i<IP_VS_DH_TAB_SIZE; i++) {
+                if (list_empty(p)) {
+                        b->dest = NULL;
+                } else {
+                        if (p == &svc->destinations)
+                                p = p->next;
+
+                        dest = list_entry(p, struct ip_vs_dest, n_list);
+                        atomic_inc(&dest->refcnt);
+                        b->dest = dest;
+
+                        p = p->next;
+                }
+                b++;
+        }
+        return 0;
+}
+
+
+/*
+ *      Flush all the hash buckets of the specified table.
+ */
+static void ip_vs_dh_flush(struct ip_vs_dh_bucket *tbl)
+{
+        int i;
+        struct ip_vs_dh_bucket *b;
+
+        b = tbl;
+        for (i=0; i<IP_VS_DH_TAB_SIZE; i++) {
+                if (b->dest) {
+                        atomic_dec(&b->dest->refcnt);
+                        b->dest = NULL;
+                }
+                b++;
+        }
+}
+
+
+static int ip_vs_dh_init_svc(struct ip_vs_service *svc)
+{
+        struct ip_vs_dh_bucket *tbl;
+
+        /* allocate the DH table for this service */
+        tbl = kmalloc(sizeof(struct ip_vs_dh_bucket)*IP_VS_DH_TAB_SIZE,
+                      GFP_ATOMIC);
+        if (tbl == NULL) {
+                IP_VS_ERR("ip_vs_dh_init_svc(): no memory\n");
+                return -ENOMEM;
+        }
+        svc->sched_data = tbl;
+        IP_VS_DBG(6, "DH hash table (memory=%Zdbytes) allocated for "
+                  "current service\n",
+                  sizeof(struct ip_vs_dh_bucket)*IP_VS_DH_TAB_SIZE);
+
+        /* assign the hash buckets with the updated service */
+        ip_vs_dh_assign(tbl, svc);
+
+        return 0;
+}
+
+
+static int ip_vs_dh_done_svc(struct ip_vs_service *svc)
+{
+        struct ip_vs_dh_bucket *tbl = svc->sched_data;
+
+        /* got to clean up hash buckets here */
+        ip_vs_dh_flush(tbl);
+
+        /* release the table itself */
+        kfree(svc->sched_data);
+        IP_VS_DBG(6, "DH hash table (memory=%Zdbytes) released\n",
+                  sizeof(struct ip_vs_dh_bucket)*IP_VS_DH_TAB_SIZE);
+
+        return 0;
+}
+
+
+static int ip_vs_dh_update_svc(struct ip_vs_service *svc)
+{
+        struct ip_vs_dh_bucket *tbl = svc->sched_data;
+
+        /* got to clean up hash buckets here */
+        ip_vs_dh_flush(tbl);
+
+        /* assign the hash buckets with the updated service */
+        ip_vs_dh_assign(tbl, svc);
+
+        return 0;
+}
+
+
+/*
+ *      If the dest flags is set with IP_VS_DEST_F_OVERLOAD,
+ *      consider that the server is overloaded here.
+ */
+static inline int is_overloaded(struct ip_vs_dest *dest)
+{
+        return dest->flags & IP_VS_DEST_F_OVERLOAD;
+}
+
+
+/*
+ *      Destination hashing scheduling
+ */
+static struct ip_vs_dest *
+ip_vs_dh_schedule(struct ip_vs_service *svc, const struct sk_buff *skb)
+{
+        struct ip_vs_dest *dest;
+        struct ip_vs_dh_bucket *tbl;
+        struct iphdr *iph = ip_hdr(skb);
+
+        IP_VS_DBG(6, "ip_vs_dh_schedule(): Scheduling...\n");
+
+        tbl = (struct ip_vs_dh_bucket *)svc->sched_data;
+        dest = ip_vs_dh_get(tbl, iph->daddr);
+        if (!dest
+            || !(dest->flags & IP_VS_DEST_F_AVAILABLE)
+            || atomic_read(&dest->weight) <= 0
+            || is_overloaded(dest)) {
+                return NULL;
+        }
+
+        IP_VS_DBG(6, "DH: destination IP address %u.%u.%u.%u "
+                  "--> server %u.%u.%u.%u:%d\n",
+                  NIPQUAD(iph->daddr),
+                  NIPQUAD(dest->addr.ip),
+                  ntohs(dest->port));
+
+        return dest;
+}
+
+
+/*
+ *      IPVS DH Scheduler structure
+ */
+static struct ip_vs_scheduler ip_vs_dh_scheduler =
+{
+        .name =                 "dh",
+        .refcnt =               ATOMIC_INIT(0),
+        .module =               THIS_MODULE,
+        .n_list =               LIST_HEAD_INIT(ip_vs_dh_scheduler.n_list),
+#ifdef CONFIG_IP_VS_IPV6
+        .supports_ipv6 =        0,
+#endif
+        .init_service =         ip_vs_dh_init_svc,
+        .done_service =         ip_vs_dh_done_svc,
+        .update_service =       ip_vs_dh_update_svc,
+        .schedule =             ip_vs_dh_schedule,
+};
+
+
+static int __init ip_vs_dh_init(void)
+{
+        return register_ip_vs_scheduler(&ip_vs_dh_scheduler);
+}
+
+
+static void __exit ip_vs_dh_cleanup(void)
+{
+        unregister_ip_vs_scheduler(&ip_vs_dh_scheduler);
+}
+
+
+module_init(ip_vs_dh_init);
+module_exit(ip_vs_dh_cleanup);
+MODULE_LICENSE("GPL");
diff --git a/net/ipv4/ipvs/ip_vs_est.c b/net/netfilter/ipvs/ip_vs_est.c
index 2eb2860dabb5..2eb2860dabb5 100644
--- a/net/ipv4/ipvs/ip_vs_est.c
+++ b/net/netfilter/ipvs/ip_vs_est.c
diff --git a/net/ipv4/ipvs/ip_vs_ftp.c b/net/netfilter/ipvs/ip_vs_ftp.c
index 2e7dbd8b73a4..2e7dbd8b73a4 100644
--- a/net/ipv4/ipvs/ip_vs_ftp.c
+++ b/net/netfilter/ipvs/ip_vs_ftp.c
diff --git a/net/ipv4/ipvs/ip_vs_lblc.c b/net/netfilter/ipvs/ip_vs_lblc.c
index 6ecef3518cac..6ecef3518cac 100644
--- a/net/ipv4/ipvs/ip_vs_lblc.c
+++ b/net/netfilter/ipvs/ip_vs_lblc.c
diff --git a/net/ipv4/ipvs/ip_vs_lblcr.c b/net/netfilter/ipvs/ip_vs_lblcr.c
index 1f75ea83bcf8..1f75ea83bcf8 100644
--- a/net/ipv4/ipvs/ip_vs_lblcr.c
+++ b/net/netfilter/ipvs/ip_vs_lblcr.c
diff --git a/net/ipv4/ipvs/ip_vs_lc.c b/net/netfilter/ipvs/ip_vs_lc.c
index b69f808ac461..b69f808ac461 100644
--- a/net/ipv4/ipvs/ip_vs_lc.c
+++ b/net/netfilter/ipvs/ip_vs_lc.c
diff --git a/net/ipv4/ipvs/ip_vs_nq.c b/net/netfilter/ipvs/ip_vs_nq.c
index 9a2d8033f08f..9a2d8033f08f 100644
--- a/net/ipv4/ipvs/ip_vs_nq.c
+++ b/net/netfilter/ipvs/ip_vs_nq.c
diff --git a/net/ipv4/ipvs/ip_vs_proto.c b/net/netfilter/ipvs/ip_vs_proto.c
index 0791f9e08feb..0791f9e08feb 100644
--- a/net/ipv4/ipvs/ip_vs_proto.c
+++ b/net/netfilter/ipvs/ip_vs_proto.c
diff --git a/net/ipv4/ipvs/ip_vs_proto_ah_esp.c b/net/netfilter/ipvs/ip_vs_proto_ah_esp.c
index 80ab0c8e5b4a..80ab0c8e5b4a 100644
--- a/net/ipv4/ipvs/ip_vs_proto_ah_esp.c
+++ b/net/netfilter/ipvs/ip_vs_proto_ah_esp.c
diff --git a/net/ipv4/ipvs/ip_vs_proto_tcp.c b/net/netfilter/ipvs/ip_vs_proto_tcp.c
index dd4566ea2bff..dd4566ea2bff 100644
--- a/net/ipv4/ipvs/ip_vs_proto_tcp.c
+++ b/net/netfilter/ipvs/ip_vs_proto_tcp.c
diff --git a/net/ipv4/ipvs/ip_vs_proto_udp.c b/net/netfilter/ipvs/ip_vs_proto_udp.c
index 6eb6039d6343..6eb6039d6343 100644
--- a/net/ipv4/ipvs/ip_vs_proto_udp.c
+++ b/net/netfilter/ipvs/ip_vs_proto_udp.c
diff --git a/net/ipv4/ipvs/ip_vs_rr.c b/net/netfilter/ipvs/ip_vs_rr.c
index a22195f68ac4..a22195f68ac4 100644
--- a/net/ipv4/ipvs/ip_vs_rr.c
+++ b/net/netfilter/ipvs/ip_vs_rr.c
diff --git a/net/ipv4/ipvs/ip_vs_sched.c b/net/netfilter/ipvs/ip_vs_sched.c
index a46ad9e35016..a46ad9e35016 100644
--- a/net/ipv4/ipvs/ip_vs_sched.c
+++ b/net/netfilter/ipvs/ip_vs_sched.c
diff --git a/net/ipv4/ipvs/ip_vs_sed.c b/net/netfilter/ipvs/ip_vs_sed.c
index 7d2f22f04b83..7d2f22f04b83 100644
--- a/net/ipv4/ipvs/ip_vs_sed.c
+++ b/net/netfilter/ipvs/ip_vs_sed.c
diff --git a/net/ipv4/ipvs/ip_vs_sh.c b/net/netfilter/ipvs/ip_vs_sh.c
index 1d96de27fefd..1d96de27fefd 100644
--- a/net/ipv4/ipvs/ip_vs_sh.c
+++ b/net/netfilter/ipvs/ip_vs_sh.c
diff --git a/net/ipv4/ipvs/ip_vs_sync.c b/net/netfilter/ipvs/ip_vs_sync.c
index 28237a5f62e2..de5e7e118eed 100644
--- a/net/ipv4/ipvs/ip_vs_sync.c
+++ b/net/netfilter/ipvs/ip_vs_sync.c
@@ -30,6 +30,7 @@
 #include <linux/err.h>
 #include <linux/kthread.h>
 #include <linux/wait.h>
+#include <linux/kernel.h>
 
 #include <net/ip.h>
 #include <net/sock.h>
@@ -99,6 +100,7 @@ struct ip_vs_sync_thread_data {
 */
 
 #define SYNC_MESG_HEADER_LEN    4
+#define MAX_CONNS_PER_SYNCBUFF  255 /* nr_conns in ip_vs_sync_mesg is 8 bit */
 
 struct ip_vs_sync_mesg {
         __u8                    nr_conns;
@@ -516,8 +518,8 @@ static int set_sync_mesg_maxlen(int sync_state)
                 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 + SIMPLE_CONN_SIZE * num;
+                sync_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);
         } else if (sync_state == IP_VS_STATE_BACKUP) {
diff --git a/net/ipv4/ipvs/ip_vs_wlc.c b/net/netfilter/ipvs/ip_vs_wlc.c
index 8c596e712599..8c596e712599 100644
--- a/net/ipv4/ipvs/ip_vs_wlc.c
+++ b/net/netfilter/ipvs/ip_vs_wlc.c
diff --git a/net/ipv4/ipvs/ip_vs_wrr.c b/net/netfilter/ipvs/ip_vs_wrr.c
index 7ea92fed50bf..7ea92fed50bf 100644
--- a/net/ipv4/ipvs/ip_vs_wrr.c
+++ b/net/netfilter/ipvs/ip_vs_wrr.c
diff --git a/net/ipv4/ipvs/ip_vs_xmit.c b/net/netfilter/ipvs/ip_vs_xmit.c
index 02ddc2b3ce2e..02ddc2b3ce2e 100644
--- a/net/ipv4/ipvs/ip_vs_xmit.c
+++ b/net/netfilter/ipvs/ip_vs_xmit.c