ipvs: SCTP Trasport Loadbalancing Support

Enhance IPVS to load balance SCTP transport protocol packets. This is done
based on the SCTP rfc 4960. All possible control chunks have been taken
care. The state machine used in this code looks some what lengthy. I tried
to make the state machine easy to understand.

Signed-off-by: Venkata Mohan Reddy Koppula <mohanreddykv@gmail.com>
Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: Patrick McHardy <kaber@trash.net>

include/net/ip_vs.h

@@ -224,6 +224,26 @@ enum {
 	IP_VS_ICMP_S_LAST,
 };
 
+/*
+ *	SCTP State Values
+ */
+enum ip_vs_sctp_states {
+	IP_VS_SCTP_S_NONE,
+	IP_VS_SCTP_S_INIT_CLI,
+	IP_VS_SCTP_S_INIT_SER,
+	IP_VS_SCTP_S_INIT_ACK_CLI,
+	IP_VS_SCTP_S_INIT_ACK_SER,
+	IP_VS_SCTP_S_ECHO_CLI,
+	IP_VS_SCTP_S_ECHO_SER,
+	IP_VS_SCTP_S_ESTABLISHED,
+	IP_VS_SCTP_S_SHUT_CLI,
+	IP_VS_SCTP_S_SHUT_SER,
+	IP_VS_SCTP_S_SHUT_ACK_CLI,
+	IP_VS_SCTP_S_SHUT_ACK_SER,
+	IP_VS_SCTP_S_CLOSED,
+	IP_VS_SCTP_S_LAST
+};
+
 /*
  *	Delta sequence info structure
  *	Each ip_vs_conn has 2 (output AND input seq. changes).
@@ -741,7 +761,7 @@ extern struct ip_vs_protocol ip_vs_protocol_udp;
 extern struct ip_vs_protocol ip_vs_protocol_icmp;
 extern struct ip_vs_protocol ip_vs_protocol_esp;
 extern struct ip_vs_protocol ip_vs_protocol_ah;
-
+extern struct ip_vs_protocol ip_vs_protocol_sctp;
 
 /*
  *      Registering/unregistering scheduler functions

net/netfilter/ipvs/Kconfig

@@ -104,6 +104,13 @@ config	IP_VS_PROTO_AH
 	  This option enables support for load balancing AH (Authentication
 	  Header) transport protocol. Say Y if unsure.
 
+config  IP_VS_PROTO_SCTP
+	bool "SCTP load balancing support"
+	select LIBCRC32C
+	---help---
+	  This option enables support for load balancing SCTP transport
+	  protocol. Say Y if unsure.
+
 comment "IPVS scheduler"
 
 config	IP_VS_RR

net/netfilter/ipvs/Makefile

@@ -7,6 +7,7 @@ ip_vs_proto-objs-y :=
 ip_vs_proto-objs-$(CONFIG_IP_VS_PROTO_TCP) += ip_vs_proto_tcp.o
 ip_vs_proto-objs-$(CONFIG_IP_VS_PROTO_UDP) += ip_vs_proto_udp.o
 ip_vs_proto-objs-$(CONFIG_IP_VS_PROTO_AH_ESP) += ip_vs_proto_ah_esp.o
+ip_vs_proto-objs-$(CONFIG_IP_VS_PROTO_SCTP) += ip_vs_proto_sctp.o
 
 ip_vs-objs :=	ip_vs_conn.o ip_vs_core.o ip_vs_ctl.o ip_vs_sched.o	   \
 		ip_vs_xmit.o ip_vs_app.o ip_vs_sync.o	   		   \

net/netfilter/ipvs/ip_vs_core.c

@@ -31,6 +31,7 @@
 #include <linux/kernel.h>
 #include <linux/ip.h>
 #include <linux/tcp.h>
+#include <linux/sctp.h>
 #include <linux/icmp.h>
 
 #include <net/ip.h>
@@ -81,6 +82,8 @@ const char *ip_vs_proto_name(unsigned proto)
 		return "UDP";
 	case IPPROTO_TCP:
 		return "TCP";
+	case IPPROTO_SCTP:
+		return "SCTP";
 	case IPPROTO_ICMP:
 		return "ICMP";
 #ifdef CONFIG_IP_VS_IPV6
@@ -589,8 +592,9 @@ void ip_vs_nat_icmp(struct sk_buff *skb, struct ip_vs_protocol *pp,
 		ip_send_check(ciph);
 	}
 
-	/* the TCP/UDP port */
+	/* the TCP/UDP/SCTP port */
-	if (IPPROTO_TCP == ciph->protocol || IPPROTO_UDP == ciph->protocol) {
+	if (IPPROTO_TCP == ciph->protocol || IPPROTO_UDP == ciph->protocol ||
+	    IPPROTO_SCTP == ciph->protocol) {
 		__be16 *ports = (void *)ciph + ciph->ihl*4;
 
 		if (inout)
@@ -630,8 +634,9 @@ void ip_vs_nat_icmp_v6(struct sk_buff *skb, struct ip_vs_protocol *pp,
 		ciph->saddr = cp->daddr.in6;
 	}
 
-	/* the TCP/UDP port */
+	/* the TCP/UDP/SCTP port */
-	if (IPPROTO_TCP == ciph->nexthdr || IPPROTO_UDP == ciph->nexthdr) {
+	if (IPPROTO_TCP == ciph->nexthdr || IPPROTO_UDP == ciph->nexthdr ||
+	    IPPROTO_SCTP == ciph->nexthdr) {
 		__be16 *ports = (void *)ciph + sizeof(struct ipv6hdr);
 
 		if (inout)
@@ -679,7 +684,8 @@ static int handle_response_icmp(int af, struct sk_buff *skb,
 		goto out;
 	}
 
-	if (IPPROTO_TCP == protocol || IPPROTO_UDP == protocol)
+	if (IPPROTO_TCP == protocol || IPPROTO_UDP == protocol ||
+	    IPPROTO_SCTP == protocol)
 		offset += 2 * sizeof(__u16);
 	if (!skb_make_writable(skb, offset))
 		goto out;
@@ -857,6 +863,21 @@ static int ip_vs_out_icmp_v6(struct sk_buff *skb, int *related)
 }
 #endif
 
+/*
+ * Check if sctp chunc is ABORT chunk
+ */
+static inline int is_sctp_abort(const struct sk_buff *skb, int nh_len)
+{
+	sctp_chunkhdr_t *sch, schunk;
+	sch = skb_header_pointer(skb, nh_len + sizeof(sctp_sctphdr_t),
+			sizeof(schunk), &schunk);
+	if (sch == NULL)
+		return 0;
+	if (sch->type == SCTP_CID_ABORT)
+		return 1;
+	return 0;
+}
+
 static inline int is_tcp_reset(const struct sk_buff *skb, int nh_len)
 {
 	struct tcphdr _tcph, *th;
@@ -999,7 +1020,8 @@ ip_vs_out(unsigned int hooknum, struct sk_buff *skb,
 	if (unlikely(!cp)) {
 		if (sysctl_ip_vs_nat_icmp_send &&
 		    (pp->protocol == IPPROTO_TCP ||
-		     pp->protocol == IPPROTO_UDP)) {
+		     pp->protocol == IPPROTO_UDP ||
+		     pp->protocol == IPPROTO_SCTP)) {
 			__be16 _ports[2], *pptr;
 
 			pptr = skb_header_pointer(skb, iph.len,
@@ -1014,8 +1036,13 @@ ip_vs_out(unsigned int hooknum, struct sk_buff *skb,
 				 * existing entry if it is not RST
 				 * packet or not TCP packet.
 				 */
-				if (iph.protocol != IPPROTO_TCP
+				if ((iph.protocol != IPPROTO_TCP &&
-				    || !is_tcp_reset(skb, iph.len)) {
+				     iph.protocol != IPPROTO_SCTP)
+				     || ((iph.protocol == IPPROTO_TCP
+					  && !is_tcp_reset(skb, iph.len))
+					 || (iph.protocol == IPPROTO_SCTP
+						&& !is_sctp_abort(skb,
+							iph.len)))) {
 #ifdef CONFIG_IP_VS_IPV6
 					if (af == AF_INET6)
 						icmpv6_send(skb,
@@ -1235,7 +1262,8 @@ ip_vs_in_icmp_v6(struct sk_buff *skb, int *related, unsigned int hooknum)
 
 	/* do the statistics and put it back */
 	ip_vs_in_stats(cp, skb);
-	if (IPPROTO_TCP == cih->nexthdr || IPPROTO_UDP == cih->nexthdr)
+	if (IPPROTO_TCP == cih->nexthdr || IPPROTO_UDP == cih->nexthdr ||
+	    IPPROTO_SCTP == cih->nexthdr)
 		offset += 2 * sizeof(__u16);
 	verdict = ip_vs_icmp_xmit_v6(skb, cp, pp, offset);
 	/* do not touch skb anymore */
@@ -1358,6 +1386,21 @@ ip_vs_in(unsigned int hooknum, struct sk_buff *skb,
 	 * encorage the standby servers to update the connections timeout
 	 */
 	pkts = atomic_add_return(1, &cp->in_pkts);
+	if (af == AF_INET && (ip_vs_sync_state & IP_VS_STATE_MASTER) &&
+	    cp->protocol == IPPROTO_SCTP) {
+		if ((cp->state == IP_VS_SCTP_S_ESTABLISHED &&
+			(atomic_read(&cp->in_pkts) %
+			 sysctl_ip_vs_sync_threshold[1]
+			 == sysctl_ip_vs_sync_threshold[0])) ||
+				(cp->old_state != cp->state &&
+				 ((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);
+			goto out;
+		}
+	}
+
 	if (af == AF_INET &&
 	    (ip_vs_sync_state & IP_VS_STATE_MASTER) &&
 	    (((cp->protocol != IPPROTO_TCP ||
@@ -1370,6 +1413,7 @@ ip_vs_in(unsigned int hooknum, struct sk_buff *skb,
 	       (cp->state == IP_VS_TCP_S_CLOSE_WAIT) ||
 	       (cp->state == IP_VS_TCP_S_TIME_WAIT)))))
 		ip_vs_sync_conn(cp);
+out:
 	cp->old_state = cp->state;
 
 	ip_vs_conn_put(cp);

net/netfilter/ipvs/ip_vs_ctl.c

@@ -2132,8 +2132,9 @@ do_ip_vs_set_ctl(struct sock *sk, int cmd, void __user *user, unsigned int len)
 		}
 	}
 
-	/* Check for valid protocol: TCP or UDP, even for fwmark!=0 */
+	/* Check for valid protocol: TCP or UDP or SCTP, even for fwmark!=0 */
-	if (usvc.protocol != IPPROTO_TCP && usvc.protocol != IPPROTO_UDP) {
+	if (usvc.protocol != IPPROTO_TCP && usvc.protocol != IPPROTO_UDP &&
+	    usvc.protocol != IPPROTO_SCTP) {
 		pr_err("set_ctl: invalid protocol: %d %pI4:%d %s\n",
 		       usvc.protocol, &usvc.addr.ip,
 		       ntohs(usvc.port), usvc.sched_name);

net/netfilter/ipvs/ip_vs_proto.c

@@ -257,6 +257,9 @@ int __init ip_vs_protocol_init(void)
 #ifdef CONFIG_IP_VS_PROTO_UDP
 	REGISTER_PROTOCOL(&ip_vs_protocol_udp);
 #endif
+#ifdef CONFIG_IP_VS_PROTO_SCTP
+	REGISTER_PROTOCOL(&ip_vs_protocol_sctp);
+#endif
 #ifdef CONFIG_IP_VS_PROTO_AH
 	REGISTER_PROTOCOL(&ip_vs_protocol_ah);
 #endif

net/netfilter/ipvs/ip_vs_sync.c

@@ -400,6 +400,11 @@ static void ip_vs_process_message(const char *buffer, const size_t buflen)
 					flags |= IP_VS_CONN_F_INACTIVE;
 				else
 					flags &= ~IP_VS_CONN_F_INACTIVE;
+			} else if (s->protocol == IPPROTO_SCTP) {
+				if (state != IP_VS_SCTP_S_ESTABLISHED)
+					flags |= IP_VS_CONN_F_INACTIVE;
+				else
+					flags &= ~IP_VS_CONN_F_INACTIVE;
 			}
 			cp = ip_vs_conn_new(AF_INET, s->protocol,
 					    (union nf_inet_addr *)&s->caddr,
@@ -434,6 +439,15 @@ static void ip_vs_process_message(const char *buffer, const size_t buflen)
 				atomic_dec(&dest->inactconns);
 				cp->flags &= ~IP_VS_CONN_F_INACTIVE;
 			}
+		} else if ((cp->dest) && (cp->protocol == IPPROTO_SCTP) &&
+			   (cp->state != state)) {
+			dest = cp->dest;
+			if (!(cp->flags & IP_VS_CONN_F_INACTIVE) &&
+			     (state != IP_VS_SCTP_S_ESTABLISHED)) {
+			    atomic_dec(&dest->activeconns);
+			    atomic_inc(&dest->inactconns);
+			    cp->flags &= ~IP_VS_CONN_F_INACTIVE;
+			}
 		}
 
 		if (opt)