Implement a new IP_SENDSRCADDR ancillary message type that permits

a server process bound to a wildcard UDP socket to select the IP
address from which outgoing packets are sent on a per-datagram
basis. When combined with IP_RECVDSTADDR, such a server process can
guarantee to reply to an incoming request using the same source IP
address as the destination IP address of the request, without having
to open one socket per server IP address.

Discussed on:	-net
Approved by:	re

3 files changed, 91 insertions, 4 deletions

diff --git a/share/man/man4/ip.4 b/share/man/man4/ip.4
index 69b0f2a..cae7061 100644
--- a/share/man/man4/ip.4
+++ b/share/man/man4/ip.4
@@ -136,6 +136,34 @@ cmsg_level = IPPROTO_IP
 cmsg_type = IP_RECVDSTADDR
 .Ed
 .Pp
+The source address to be used for outgoing
+.Tn UDP
+datagrams on a socket that is not bound to a specific
+.Tn IP
+address can be specified as ancillary data with a type code of
+.Dv IP_SENDSRCADDR .
+The msg_control field in the msghdr structure should point to a buffer
+that contains a cmsghdr structure followed by the
+.Tn IP
+address.
+The cmsghdr fields should have the following values:
+.Bd -literal
+cmsg_len = sizeof(struct in_addr)
+cmsg_level = IPPROTO_IP
+cmsg_type = IP_SENDSRCADDR
+.Ed
+.Pp
+For convenience,
+.Dv IP_SENDSRCADDR
+is defined to have the same value as
+.Dv IP_RECVDSTADDR ,
+so the
+.Dv IP_RECVDSTADDR
+control message from
+.Xr recvmsg 2
+can be used directly as a control message for
+.Xr sendmsg 2 .
+.Pp
 .Dv IP_PORTRANGE
 may be used to set the port range used for selecting a local port number
 on a socket with an unspecified (zero) port number.
diff --git a/sys/netinet/in.h b/sys/netinet/in.h
index 09fb134..09fd395 100644
--- a/sys/netinet/in.h
+++ b/sys/netinet/in.h
@@ -366,6 +366,7 @@ __END_DECLS
 #define	IP_RECVOPTS		5    /* bool; receive all IP opts w/dgram */
 #define	IP_RECVRETOPTS		6    /* bool; receive IP opts for response */
 #define	IP_RECVDSTADDR		7    /* bool; receive IP dst addr w/dgram */
+#define	IP_SENDSRCADDR		IP_RECVDSTADDR /* cmsg_type to set src addr */
 #define	IP_RETOPTS		8    /* ip_opts; set/get IP options */
 #define	IP_MULTICAST_IF		9    /* u_char; set/get IP multicast i/f  */
 #define	IP_MULTICAST_TTL	10   /* u_char; set/get IP multicast ttl */
diff --git a/sys/netinet/udp_usrreq.c b/sys/netinet/udp_usrreq.c
index 263a92e..437ee4e 100644
--- a/sys/netinet/udp_usrreq.c
+++ b/sys/netinet/udp_usrreq.c
@@ -689,7 +689,8 @@ udp_output(inp, m, addr, control, td)
 	register struct udpiphdr *ui;
 	register int len = m->m_pkthdr.len;
 	struct in_addr faddr, laddr;
-	struct sockaddr_in *sin;
+	struct cmsghdr *cm;
+	struct sockaddr_in *sin, src;
 	int error = 0;
 	u_short fport, lport;
 
@@ -697,16 +698,73 @@ udp_output(inp, m, addr, control, td)
 	mac_create_mbuf_from_socket(inp->inp_socket, m);
 #endif
 
-	if (control)
-		m_freem(control);		/* XXX */
-
 	if (len + sizeof(struct udpiphdr) > IP_MAXPACKET) {
 		error = EMSGSIZE;
+		if (control)
+			m_freem(control);
 		goto release;
 	}
 
+	src.sin_addr.s_addr = INADDR_ANY;
+	if (control != NULL) {
+		/*
+		 * XXX: Currently, we assume all the optional information
+		 * is stored in a single mbuf.
+		 */
+		if (control->m_next) {
+			error = EINVAL;
+			m_freem(control);
+			goto release;
+		}
+		for (; control->m_len > 0;
+		    control->m_data += CMSG_ALIGN(cm->cmsg_len),
+		    control->m_len -= CMSG_ALIGN(cm->cmsg_len)) {
+			cm = mtod(control, struct cmsghdr *);
+			if (control->m_len < sizeof(*cm) || cm->cmsg_len == 0 ||
+			    cm->cmsg_len > control->m_len) {
+				error = EINVAL;
+				break;
+			}
+			if (cm->cmsg_level != IPPROTO_IP)
+				continue;
+
+			switch (cm->cmsg_type) {
+			case IP_SENDSRCADDR:
+				if (cm->cmsg_len !=
+				    CMSG_LEN(sizeof(struct in_addr))) {
+					error = EINVAL;
+					break;
+				}
+				bzero(&src, sizeof(src));
+				src.sin_family = AF_INET;
+				src.sin_len = sizeof(src);
+				src.sin_port = inp->inp_lport;
+				src.sin_addr = *(struct in_addr *)CMSG_DATA(cm);
+				break;
+			default:
+				error = ENOPROTOOPT;
+				break;
+			}
+			if (error)
+				break;
+		}
+		m_freem(control);
+	}
+	if (error)
+		goto release;
 	laddr = inp->inp_laddr;
 	lport = inp->inp_lport;
+	if (src.sin_addr.s_addr != INADDR_ANY) {
+		if (lport == 0) {
+			error = EINVAL;
+			goto release;
+		}
+		error = in_pcbbind_setup(inp, (struct sockaddr *)&src,
+		    &laddr.s_addr, &lport, td);
+		if (error)
+			goto release;
+	}
+
 	if (addr) {
 		sin = (struct sockaddr_in *)addr;
 		if (td && jailed(td->td_ucred))