svn commit: r191063 - in stable/6/sys: . contrib/pf dev/cxgb netinet

[Ed Maste]

Author: emaste
Date: Tue Apr 14 16:45:17 2009
New Revision: 191063
URL: http://svn.freebsd.org/changeset/base/191063

Log:
  MFC r171746 by csjp
  
  Summary:
  
  - We disallow multicast operations on divert sockets.  It really doesn't make
    semantic sense to allow this, since typically you would set multicast
    parameters on multicast end points.
  
    NOTE: As a part of this change, we actually dis-allow multicast options on
    any socket that IS a divert socket OR IS NOT a SOCK_RAW or SOCK_DGRAM family
  
  - We check to see if there are any socket options that have been specified on
    the socket, and if there was (which is very un-common and also probably
    doesnt make sense to support) we duplicate the mbuf carrying the options.
  
  - We then drop the INP/INFO locks over the call to ip_output().  It should be
    noted that since we no longer support multicast operations on divert sockets
    and we have duplicated any socket options, we no longer need the reference
    to the pcb to be coherent.
  
  - Finally, we replaced the call to ip_input() to use netisr queuing.  This
    should remove the recursive entry into the IP stack from divert.
  
  (The ip_output.c changes come from in_mcast.c in head.)

Modified:
  stable/6/sys/   (props changed)
  stable/6/sys/contrib/pf/   (props changed)
  stable/6/sys/dev/cxgb/   (props changed)
  stable/6/sys/netinet/ip_divert.c
  stable/6/sys/netinet/ip_output.c

Modified: stable/6/sys/netinet/ip_divert.c
==============================================================================
--- stable/6/sys/netinet/ip_divert.c	Tue Apr 14 16:24:56 2009	(r191062)
+++ stable/6/sys/netinet/ip_divert.c	Tue Apr 14 16:45:17 2009	(r191063)
@@ -61,6 +61,7 @@
 #include <vm/uma.h>
 
 #include <net/if.h>
+#include <net/netisr.h> 
 #include <net/route.h>
 
 #include <netinet/in.h>
@@ -303,6 +304,7 @@ div_output(struct socket *so, struct mbu
 	struct m_tag *mtag;
 	struct divert_tag *dt;
 	int error = 0;
+	struct mbuf *options;
 
 	/*
 	 * An mbuf may hasn't come from userland, but we pretend
@@ -359,6 +361,8 @@ div_output(struct socket *so, struct mbu
 		if (((ip->ip_hl != (sizeof (*ip) >> 2)) && inp->inp_options) ||
 		     ((u_short)ntohs(ip->ip_len) > m->m_pkthdr.len)) {
 			error = EINVAL;
+			INP_UNLOCK(inp);
+			INP_INFO_WUNLOCK(&divcbinfo);
 			m_freem(m);
 		} else {
 			/* Convert fields to host order for ip_output() */
@@ -371,15 +375,46 @@ div_output(struct socket *so, struct mbu
 #ifdef MAC
 			mac_create_mbuf_from_inpcb(inp, m);
 #endif
-			error = ip_output(m,
-				    inp->inp_options, NULL,
-				    ((so->so_options & SO_DONTROUTE) ?
-				    IP_ROUTETOIF : 0) |
-				    IP_ALLOWBROADCAST | IP_RAWOUTPUT,
-				    inp->inp_moptions, NULL);
+			/*
+			 * Get ready to inject the packet into ip_output().
+			 * Just in case socket options were specified on the
+			 * divert socket, we duplicate them.  This is done
+			 * to avoid having to hold the PCB locks over the call
+			 * to ip_output(), as doing this results in a number of
+			 * lock ordering complexities.
+			 *
+			 * Note that we set the multicast options argument for
+			 * ip_output() to NULL since it should be invariant that
+			 * they are not present.
+			 */
+			KASSERT(inp->inp_moptions == NULL,
+			    ("multicast options set on a divert socket"));
+			options = NULL;
+			/*
+			 * XXXCSJP: It is unclear to me whether or not it makes
+			 * sense for divert sockets to have options.  However,
+			 * for now we will duplicate them with the INP locks
+			 * held so we can use them in ip_output() without
+			 * requring a reference to the pcb.
+			 */
+			if (inp->inp_options != NULL) {
+				options = m_dup(inp->inp_options, M_DONTWAIT);
+				if (options == NULL)
+					error = ENOBUFS;
+			}
+			INP_UNLOCK(inp);
+			INP_INFO_WUNLOCK(&divcbinfo);
+			if (error == ENOBUFS) {
+				m_freem(m);
+				return (error);
+			}
+			error = ip_output(m, options, NULL,
+			    ((so->so_options & SO_DONTROUTE) ?
+			    IP_ROUTETOIF : 0) | IP_ALLOWBROADCAST |
+			    IP_RAWOUTPUT, NULL, NULL);
+			if (options != NULL)
+				m_freem(options);
 		}
-		INP_UNLOCK(inp);
-		INP_INFO_WUNLOCK(&divcbinfo);
 	} else {
 		dt->info |= IP_FW_DIVERT_LOOPBACK_FLAG;
 		if (m->m_pkthdr.rcvif == NULL) {
@@ -404,8 +439,8 @@ div_output(struct socket *so, struct mbu
 		mac_create_mbuf_from_socket(so, m);
 		SOCK_UNLOCK(so);
 #endif
-		/* Send packet to input processing */
-		ip_input(m);
+		/* Send packet to input processing via netisr */
+		netisr_queue(NETISR_IP, m);
 	}
 
 	return error;

Modified: stable/6/sys/netinet/ip_output.c
==============================================================================
--- stable/6/sys/netinet/ip_output.c	Tue Apr 14 16:24:56 2009	(r191062)
+++ stable/6/sys/netinet/ip_output.c	Tue Apr 14 16:45:17 2009	(r191063)
@@ -1710,6 +1710,16 @@ ip_setmoptions(struct inpcb *inp, struct
 	int ifindex;
 	int s;
 
+	/*
+	 * If socket is neither of type SOCK_RAW or SOCK_DGRAM,
+	 * or is a divert socket, reject it.
+	 * XXX Unlocked read of inp_socket believed OK.
+	 */
+	if (inp->inp_socket->so_proto->pr_protocol == IPPROTO_DIVERT ||
+	    (inp->inp_socket->so_proto->pr_type != SOCK_RAW &&
+	    inp->inp_socket->so_proto->pr_type != SOCK_DGRAM))
+		return (EOPNOTSUPP);
+
 	switch (sopt->sopt_name) {
 	/* store an index number for the vif you wanna use in the send */
 	case IP_MULTICAST_VIF:
@@ -1995,6 +2005,16 @@ ip_getmoptions(struct inpcb *inp, struct
 
 	INP_LOCK(inp);
 	imo = inp->inp_moptions;
+	/*
+	 * If socket is neither of type SOCK_RAW or SOCK_DGRAM,
+	 * or is a divert socket, reject it.
+	 */
+	if (inp->inp_socket->so_proto->pr_protocol == IPPROTO_DIVERT ||
+	    (inp->inp_socket->so_proto->pr_type != SOCK_RAW &&
+	    inp->inp_socket->so_proto->pr_type != SOCK_DGRAM)) {
+		INP_UNLOCK(inp);
+		return (EOPNOTSUPP);
+	}
 
 	error = 0;
 	switch (sopt->sopt_name) {