1 files changed, 64 insertions, 90 deletions

diff --git a/freebsd/sys/netinet6/in6_src.c b/freebsd/sys/netinet6/in6_src.c
index f7cd642e..4f20ee54 100644
--- a/freebsd/sys/netinet6/in6_src.c
+++ b/freebsd/sys/netinet6/in6_src.c
@@ -131,9 +131,10 @@ VNET_DEFINE(int, ip6_prefer_tempaddr) = 0;
 
 static int selectroute __P((struct sockaddr_in6 *, struct ip6_pktopts *,
 	struct ip6_moptions *, struct route_in6 *, struct ifnet **,
-	struct rtentry **, int));
+	struct rtentry **, int, u_int));
 static int in6_selectif __P((struct sockaddr_in6 *, struct ip6_pktopts *,
-	struct ip6_moptions *, struct route_in6 *ro, struct ifnet **));
+	struct ip6_moptions *, struct route_in6 *ro, struct ifnet **,
+	struct ifnet *, u_int));
 
 static struct in6_addrpolicy *lookup_addrsel_policy(struct sockaddr_in6 *);
 
@@ -184,7 +185,7 @@ in6_selectsrc(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts,
     struct ifnet **ifpp, struct in6_addr *srcp)
 {
 	struct in6_addr dst, tmp;
-	struct ifnet *ifp = NULL;
+	struct ifnet *ifp = NULL, *oifp = NULL;
 	struct in6_ifaddr *ia = NULL, *ia_best = NULL;
 	struct in6_pktinfo *pi = NULL;
 	int dst_scope = -1, best_scope = -1, best_matchlen = -1;
@@ -197,8 +198,18 @@ in6_selectsrc(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts,
 	KASSERT(srcp != NULL, ("%s: srcp is NULL", __func__));
 
 	dst = dstsock->sin6_addr; /* make a copy for local operation */
-	if (ifpp)
+	if (ifpp) {
+		/*
+		 * Save a possibly passed in ifp for in6_selectsrc. Only
+		 * neighbor discovery code should use this feature, where
+		 * we may know the interface but not the FIB number holding
+		 * the connected subnet in case someone deleted it from the
+		 * default FIB and we need to check the interface.
+		 */
+		if (*ifpp != NULL)
+			oifp = *ifpp;
 		*ifpp = NULL;
+	}
 
 	if (inp != NULL) {
 		INP_LOCK_ASSERT(inp);
@@ -219,7 +230,9 @@ in6_selectsrc(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts,
 		struct in6_ifaddr *ia6;
 
 		/* get the outgoing interface */
-		if ((error = in6_selectif(dstsock, opts, mopts, ro, &ifp)) != 0)
+		if ((error = in6_selectif(dstsock, opts, mopts, ro, &ifp, oifp,
+		    (inp != NULL) ? inp->inp_inc.inc_fibnum : RT_DEFAULT_FIB))
+		    != 0)
 			return (error);
 
 		/*
@@ -283,7 +296,8 @@ in6_selectsrc(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts,
 	 * the outgoing interface and the destination address.
 	 */
 	/* get the outgoing interface */
-	if ((error = in6_selectif(dstsock, opts, mopts, ro, &ifp)) != 0)
+	if ((error = in6_selectif(dstsock, opts, mopts, ro, &ifp, oifp,
+	    (inp != NULL) ? inp->inp_inc.inc_fibnum : RT_DEFAULT_FIB)) != 0)
 		return (error);
 
 #ifdef DIAGNOSTIC
@@ -506,7 +520,7 @@ in6_selectsrc(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts,
 static int
 selectroute(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts,
     struct ip6_moptions *mopts, struct route_in6 *ro,
-    struct ifnet **retifp, struct rtentry **retrt, int norouteok)
+    struct ifnet **retifp, struct rtentry **retrt, int norouteok, u_int fibnum)
 {
 	int error = 0;
 	struct ifnet *ifp = NULL;
@@ -583,7 +597,7 @@ selectroute(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts,
 
 
 		if (ron->ro_rt == NULL) {
-			rtalloc((struct route *)ron); /* multi path case? */
+			in6_rtalloc(ron, fibnum); /* multi path case? */
 			if (ron->ro_rt == NULL) {
 				if (ron->ro_rt) {
 					RTFREE(ron->ro_rt);
@@ -596,9 +610,9 @@ selectroute(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts,
 
 		rt = ron->ro_rt;
 		ifp = rt->rt_ifp;
-		IF_AFDATA_LOCK(ifp);
+		IF_AFDATA_RLOCK(ifp);
 		la = lla_lookup(LLTABLE6(ifp), 0, (struct sockaddr *)&sin6_next->sin6_addr);
-		IF_AFDATA_UNLOCK(ifp);
+		IF_AFDATA_RUNLOCK(ifp);
 		if (la != NULL) 
 			LLE_RUNLOCK(la);
 		else {
@@ -618,7 +632,7 @@ selectroute(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts,
 			*satosin6(&ron->ro_dst) = *sin6_next;
 		}
 		if (ron->ro_rt == NULL) {
-			rtalloc((struct route *)ron); /* multi path case? */
+			in6_rtalloc(ron, fibnum); /* multi path case? */
 			if (ron->ro_rt == NULL ||
 			    !(ron->ro_rt->rt_flags & RTF_LLINFO)) {
 				if (ron->ro_rt) {
@@ -663,11 +677,11 @@ selectroute(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts,
 			sa6->sin6_scope_id = 0;
 
 #ifdef RADIX_MPATH
-				rtalloc_mpath((struct route *)ro,
-				    ntohl(sa6->sin6_addr.s6_addr32[3]));
+				rtalloc_mpath_fib((struct route *)ro,
+				    ntohl(sa6->sin6_addr.s6_addr32[3]), fibnum);
 #else			
-				ro->ro_rt = rtalloc1(&((struct route *)ro)
-				    ->ro_dst, 0, 0UL);
+				ro->ro_rt = in6_rtalloc1((struct sockaddr *)
+				    &ro->ro_dst, 0, 0UL, fibnum);
 				if (ro->ro_rt)
 					RT_UNLOCK(ro->ro_rt);
 #endif
@@ -748,21 +762,29 @@ selectroute(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts,
 
 static int
 in6_selectif(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts,
-    struct ip6_moptions *mopts, struct route_in6 *ro, struct ifnet **retifp)
+    struct ip6_moptions *mopts, struct route_in6 *ro, struct ifnet **retifp,
+    struct ifnet *oifp, u_int fibnum)
 {
 	int error;
 	struct route_in6 sro;
 	struct rtentry *rt = NULL;
 
+	KASSERT(retifp != NULL, ("%s: retifp is NULL", __func__));
+
 	if (ro == NULL) {
 		bzero(&sro, sizeof(sro));
 		ro = &sro;
 	}
 
 	if ((error = selectroute(dstsock, opts, mopts, ro, retifp,
-				     &rt, 1)) != 0) {
+	    &rt, 1, fibnum)) != 0) {
 		if (ro == &sro && rt && rt == sro.ro_rt)
 			RTFREE(rt);
+		/* Help ND. See oifp comment in in6_selectsrc(). */
+		if (oifp != NULL && fibnum == RT_DEFAULT_FIB) {
+			*retifp = oifp;
+			error = 0;
+		}
 		return (error);
 	}
 
@@ -797,7 +819,10 @@ in6_selectif(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts,
 }
 
 /*
- * clone - meaningful only for bsdi and freebsd
+ * Public wrapper function to selectroute().
+ *
+ * XXX-BZ in6_selectroute() should and will grow the FIB argument. The
+ * in6_selectroute_fib() function is only there for backward compat on stable.
  */
 int
 in6_selectroute(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts,
@@ -806,9 +831,21 @@ in6_selectroute(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts,
 {
 
 	return (selectroute(dstsock, opts, mopts, ro, retifp,
-	    retrt, 0));
+	    retrt, 0, RT_DEFAULT_FIB));
 }
 
+#ifndef BURN_BRIDGES
+int
+in6_selectroute_fib(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts,
+    struct ip6_moptions *mopts, struct route_in6 *ro,
+    struct ifnet **retifp, struct rtentry **retrt, u_int fibnum)
+{
+
+	return (selectroute(dstsock, opts, mopts, ro, retifp,
+	    retrt, 0, fibnum));
+}
+#endif
+
 /*
  * Default hop limit selection. The precedence is as follows:
  * 1. Hoplimit value specified via ioctl.
@@ -832,7 +869,7 @@ in6_selecthlim(struct inpcb *in6p, struct ifnet *ifp)
 		ro6.ro_dst.sin6_family = AF_INET6;
 		ro6.ro_dst.sin6_len = sizeof(struct sockaddr_in6);
 		ro6.ro_dst.sin6_addr = in6p->in6p_faddr;
-		rtalloc((struct route *)&ro6);
+		in6_rtalloc(&ro6, in6p->inp_inc.inc_fibnum);
 		if (ro6.ro_rt) {
 			lifp = ro6.ro_rt->rt_ifp;
 			RTFREE(ro6.ro_rt);
@@ -852,9 +889,11 @@ int
 in6_pcbsetport(struct in6_addr *laddr, struct inpcb *inp, struct ucred *cred)
 {
 	struct socket *so = inp->inp_socket;
-	u_int16_t lport = 0, first, last, *lastport;
-	int count, error, wild = 0, dorandom;
+	u_int16_t lport = 0;
+	int error, wild = 0;
+#ifdef INVARIANTS
 	struct inpcbinfo *pcbinfo = inp->inp_pcbinfo;
+#endif
 
 	INP_INFO_WLOCK_ASSERT(pcbinfo);
 	INP_WLOCK_ASSERT(inp);
@@ -870,74 +909,9 @@ in6_pcbsetport(struct in6_addr *laddr, struct inpcb *inp, struct ucred *cred)
 
 	inp->inp_flags |= INP_ANONPORT;
 
-	if (inp->inp_flags & INP_HIGHPORT) {
-		first = V_ipport_hifirstauto;	/* sysctl */
-		last  = V_ipport_hilastauto;
-		lastport = &pcbinfo->ipi_lasthi;
-	} else if (inp->inp_flags & INP_LOWPORT) {
-		error = priv_check_cred(cred, PRIV_NETINET_RESERVEDPORT, 0);
-		if (error)
-			return error;
-		first = V_ipport_lowfirstauto;	/* 1023 */
-		last  = V_ipport_lowlastauto;	/* 600 */
-		lastport = &pcbinfo->ipi_lastlow;
-	} else {
-		first = V_ipport_firstauto;	/* sysctl */
-		last  = V_ipport_lastauto;
-		lastport = &pcbinfo->ipi_lastport;
-	}
-
-	/*
-	 * For UDP, use random port allocation as long as the user
-	 * allows it.  For TCP (and as of yet unknown) connections,
-	 * use random port allocation only if the user allows it AND
-	 * ipport_tick() allows it.
-	 */
-	if (V_ipport_randomized &&
-	    (!V_ipport_stoprandom || pcbinfo == &V_udbinfo))
-		dorandom = 1;
-	else
-		dorandom = 0;
-	/*
-	 * It makes no sense to do random port allocation if
-	 * we have the only port available.
-	 */
-	if (first == last)
-		dorandom = 0;
-	/* Make sure to not include UDP packets in the count. */
-	if (pcbinfo != &V_udbinfo)
-		V_ipport_tcpallocs++;
-
-	/*
-	 * Instead of having two loops further down counting up or down
-	 * make sure that first is always <= last and go with only one
-	 * code path implementing all logic.
-	 */
-	if (first > last) {
-		u_int16_t aux;
-
-		aux = first;
-		first = last;
-		last = aux;
-	}
-
-	if (dorandom)
-		*lastport = first + (arc4random() % (last - first));
-
-	count = last - first;
-
-	do {
-		if (count-- < 0) {	/* completely used? */
-			/* Undo an address bind that may have occurred. */
-			inp->in6p_laddr = in6addr_any;
-			return (EADDRNOTAVAIL);
-		}
-		++*lastport;
-		if (*lastport < first || *lastport > last)
-			*lastport = first;
-		lport = htons(*lastport);
-	} while (in6_pcblookup_local(pcbinfo, &inp->in6p_laddr,
-	    lport, wild, cred));
+	error = in_pcb_lport(inp, NULL, &lport, cred, wild);
+	if (error != 0)
+		return (error);
 
 	inp->inp_lport = lport;
 	if (in_pcbinshash(inp) != 0) {