Merge nuts.davemloft.net:/disk1/BK/ip6route-2.6

into nuts.davemloft.net:/disk1/BK/net-2.6

include/net/if_inet6.h

@@ -51,6 +51,7 @@ struct inet6_ifaddr
 	struct timer_list	timer;
 
 	struct inet6_dev	*idev;
+	struct rt6_info		*rt;
 
 	struct inet6_ifaddr	*lst_next;      /* next addr in addr_lst */
 	struct inet6_ifaddr	*if_next;       /* next addr in inet6_dev */
@@ -133,6 +134,7 @@ struct ifacaddr6
 {
 	struct in6_addr		aca_addr;
 	struct inet6_dev	*aca_idev;
+	struct rt6_info		*aca_rt;
 	struct ifacaddr6	*aca_next;
 	int			aca_users;
 	atomic_t		aca_refcnt;

include/net/ip6_route.h

@@ -42,6 +42,9 @@ extern int			ipv6_route_ioctl(unsigned int cmd, void __user *arg);
 extern int			ip6_route_add(struct in6_rtmsg *rtmsg,
 					      struct nlmsghdr *,
 					      void *rtattr);
+extern int			ip6_ins_rt(struct rt6_info *,
+					   struct nlmsghdr *,
+					   void *rtattr);
 extern int			ip6_del_rt(struct rt6_info *,
 					   struct nlmsghdr *,
 					   void *rtattr);
@@ -71,6 +74,10 @@ extern struct dst_entry *ndisc_dst_alloc(struct net_device *dev,
 extern int ndisc_dst_gc(int *more);
 extern void fib6_force_start_gc(void);
 
+extern struct rt6_info *addrconf_dst_alloc(struct inet6_dev *idev,
+					   const struct in6_addr *addr,
+					   int anycast);
+
 /*
  *	support functions for ND
  *

net/ipv4/route.c

@@ -1342,10 +1342,13 @@ static void ipv4_dst_ifdown(struct dst_entry *dst, int how)
 {
 	struct rtable *rt = (struct rtable *) dst;
 	struct in_device *idev = rt->idev;
-	if (idev) {
-		rt->idev = NULL;
+	if (idev && idev->dev != &loopback_dev) {
+		struct in_device *loopback_idev = in_dev_get(&loopback_dev);
+		if (loopback_idev) {
+			rt->idev = loopback_idev;
 			in_dev_put(idev);
 		}
+	}
 }
 
 static void ipv4_link_failure(struct sk_buff *skb)

net/ipv6/addrconf.c

@@ -472,6 +472,8 @@ void inet6_ifa_finish_destroy(struct inet6_ifaddr *ifp)
 		printk("Freeing alive inet6 address %p\n", ifp);
 		return;
 	}
+	dst_release(&ifp->rt->u.dst);
+
 	inet6_ifa_count--;
 	kfree(ifp);
 }
@@ -482,25 +484,33 @@ static struct inet6_ifaddr *
 ipv6_add_addr(struct inet6_dev *idev, const struct in6_addr *addr, int pfxlen,
 	      int scope, unsigned flags)
 {
-	struct inet6_ifaddr *ifa;
+	struct inet6_ifaddr *ifa = NULL;
+	struct rt6_info *rt;
 	int hash;
 	static spinlock_t lock = SPIN_LOCK_UNLOCKED;
+	int err = 0;
 
 	spin_lock_bh(&lock);
 
 	/* Ignore adding duplicate addresses on an interface */
 	if (ipv6_chk_same_addr(addr, idev->dev)) {
-		spin_unlock_bh(&lock);
 		ADBG(("ipv6_add_addr: already assigned\n"));
-		return ERR_PTR(-EEXIST);
+		err = -EEXIST;
+		goto out;
 	}
 
 	ifa = kmalloc(sizeof(struct inet6_ifaddr), GFP_ATOMIC);
 
 	if (ifa == NULL) {
-		spin_unlock_bh(&lock);
 		ADBG(("ipv6_add_addr: malloc failed\n"));
-		return ERR_PTR(-ENOBUFS);
+		err = -ENOBUFS;
+		goto out;
+	}
+
+	rt = addrconf_dst_alloc(idev, addr, 0);
+	if (IS_ERR(rt)) {
+		err = PTR_ERR(rt);
+		goto out;
 	}
 
 	memset(ifa, 0, sizeof(struct inet6_ifaddr));
@@ -517,9 +527,8 @@ ipv6_add_addr(struct inet6_dev *idev, const struct in6_addr *addr, int pfxlen,
 	read_lock(&addrconf_lock);
 	if (idev->dead) {
 		read_unlock(&addrconf_lock);
-		spin_unlock_bh(&lock);
-		kfree(ifa);
-		return ERR_PTR(-ENODEV);	/*XXX*/
+		err = -ENODEV;	/*XXX*/
+		goto out;
 	}
 
 	inet6_ifa_count++;
@@ -553,12 +562,20 @@ ipv6_add_addr(struct inet6_dev *idev, const struct in6_addr *addr, int pfxlen,
 	}
 #endif
 
+	ifa->rt = rt;
+
 	in6_ifa_hold(ifa);
 	write_unlock_bh(&idev->lock);
 	read_unlock(&addrconf_lock);
+out:
 	spin_unlock_bh(&lock);
 
-	notifier_call_chain(&inet6addr_chain,NETDEV_UP,ifa);
+	if (unlikely(err == 0))
+		notifier_call_chain(&inet6addr_chain, NETDEV_UP, ifa);
+	else {
+		kfree(ifa);
+		ifa = ERR_PTR(err);
+	}
 
 	return ifa;
 }
@@ -1457,8 +1474,7 @@ void addrconf_prefix_rcv(struct net_device *dev, u8 *opt, int len)
 				spin_unlock(&ifp->lock);
 
 				if (!(flags&IFA_F_TENTATIVE))
-					ipv6_ifa_notify((flags&IFA_F_DEPRECATED) ?
-							0 : RTM_NEWADDR, ifp);
+					ipv6_ifa_notify(0, ifp);
 			} else
 				spin_unlock(&ifp->lock);
 
@@ -2982,7 +2998,9 @@ static void ipv6_ifa_notify(int event, struct inet6_ifaddr *ifp)
 
 	switch (event) {
 	case RTM_NEWADDR:
-		ip6_rt_addr_add(&ifp->addr, ifp->idev->dev, 0);
+		dst_hold(&ifp->rt->u.dst);
+		if (ip6_ins_rt(ifp->rt, NULL, NULL))
+			dst_release(&ifp->rt->u.dst);
 		break;
 	case RTM_DELADDR:
 		addrconf_leave_solict(ifp->idev->dev, &ifp->addr);
@@ -2993,8 +3011,11 @@ static void ipv6_ifa_notify(int event, struct inet6_ifaddr *ifp)
 			if (!ipv6_addr_any(&addr))
 				ipv6_dev_ac_dec(ifp->idev->dev, &addr);
 		}
-		if (!ipv6_chk_addr(&ifp->addr, ifp->idev->dev, 1))
-			ip6_rt_addr_del(&ifp->addr, ifp->idev->dev);
+		dst_hold(&ifp->rt->u.dst);
+		if (ip6_del_rt(ifp->rt, NULL, NULL))
+			dst_free(&ifp->rt->u.dst);
+		else
+			dst_release(&ifp->rt->u.dst);
 		break;
 	}
 }

net/ipv6/anycast.c

@@ -293,6 +293,7 @@ static void aca_put(struct ifacaddr6 *ac)
 {
 	if (atomic_dec_and_test(&ac->aca_refcnt)) {
 		in6_dev_put(ac->aca_idev);
+		dst_release(&ac->aca_rt->u.dst);
 		kfree(ac);
 	}
 }
@@ -304,6 +305,8 @@ int ipv6_dev_ac_inc(struct net_device *dev, struct in6_addr *addr)
 {
 	struct ifacaddr6 *aca;
 	struct inet6_dev *idev;
+	struct rt6_info *rt;
+	int err;
 
 	idev = in6_dev_get(dev);
 
@@ -312,17 +315,15 @@ int ipv6_dev_ac_inc(struct net_device *dev, struct in6_addr *addr)
 
 	write_lock_bh(&idev->lock);
 	if (idev->dead) {
-		write_unlock_bh(&idev->lock);
-		in6_dev_put(idev);
-		return -ENODEV;
+		err = -ENODEV;
+		goto out;
 	}
 
 	for (aca = idev->ac_list; aca; aca = aca->aca_next) {
 		if (ipv6_addr_cmp(&aca->aca_addr, addr) == 0) {
 			aca->aca_users++;
-			write_unlock_bh(&idev->lock);
-			in6_dev_put(idev);
-			return 0;
+			err = 0;
+			goto out;
 		}
 	}
 
@@ -333,15 +334,22 @@ int ipv6_dev_ac_inc(struct net_device *dev, struct in6_addr *addr)
 	aca = kmalloc(sizeof(struct ifacaddr6), GFP_ATOMIC);
 
 	if (aca == NULL) {
-		write_unlock_bh(&idev->lock);
-		in6_dev_put(idev);
-		return -ENOMEM;
+		err = -ENOMEM;
+		goto out;
+	}
+
+	rt = addrconf_dst_alloc(idev, addr, 1);
+	if (IS_ERR(rt)) {
+		kfree(aca);
+		err = PTR_ERR(rt);
+		goto out;
 	}
 
 	memset(aca, 0, sizeof(struct ifacaddr6));
 
 	ipv6_addr_copy(&aca->aca_addr, addr);
 	aca->aca_idev = idev;
+	aca->aca_rt = rt;
 	aca->aca_users = 1;
 	/* aca_tstamp should be updated upon changes */
 	aca->aca_cstamp = aca->aca_tstamp = jiffies;
@@ -352,12 +360,18 @@ int ipv6_dev_ac_inc(struct net_device *dev, struct in6_addr *addr)
 	idev->ac_list = aca;
 	write_unlock_bh(&idev->lock);
 
-	ip6_rt_addr_add(&aca->aca_addr, dev, 1);
+	dst_hold(&rt->u.dst);
+	if (ip6_ins_rt(rt, NULL, NULL))
+		dst_release(&rt->u.dst);
 
 	addrconf_join_solict(dev, &aca->aca_addr);
 
 	aca_put(aca);
 	return 0;
+out:
+	write_unlock_bh(&idev->lock);
+	in6_dev_put(idev);
+	return err;
 }
 
 /*
@@ -396,7 +410,11 @@ int ipv6_dev_ac_dec(struct net_device *dev, struct in6_addr *addr)
 	write_unlock_bh(&idev->lock);
 	addrconf_leave_solict(dev, &aca->aca_addr);
 
-	ip6_rt_addr_del(&aca->aca_addr, dev);
+	dst_hold(&aca->aca_rt->u.dst);
+	if (ip6_del_rt(aca->aca_rt, NULL, NULL))
+		dst_free(&aca->aca_rt->u.dst);
+	else
+		dst_release(&aca->aca_rt->u.dst);
 
 	aca_put(aca);
 	in6_dev_put(idev);

net/ipv6/ip6_fib.c

@@ -449,9 +449,10 @@ static int fib6_add_rt2node(struct fib6_node *fn, struct rt6_info *rt,
 			 *	Same priority level
 			 */
 
-			if ((iter->rt6i_dev == rt->rt6i_dev) &&
-			    (ipv6_addr_cmp(&iter->rt6i_gateway,
-					   &rt->rt6i_gateway) == 0)) {
+			if (iter->rt6i_dev == rt->rt6i_dev &&
+			    iter->rt6i_idev == rt->rt6i_idev &&
+			    ipv6_addr_cmp(&iter->rt6i_gateway,
+					   &rt->rt6i_gateway) == 0) {
 				if (!(iter->rt6i_flags&RTF_EXPIRES))
 					return -EEXIST;
 				iter->rt6i_expires = rt->rt6i_expires;

net/ipv6/route.c

@@ -155,7 +155,16 @@ static void ip6_dst_destroy(struct dst_entry *dst)
 
 static void ip6_dst_ifdown(struct dst_entry *dst, int how)
 {
-	ip6_dst_destroy(dst);
+	struct rt6_info *rt = (struct rt6_info *)dst;
+	struct inet6_dev *idev = rt->rt6i_idev;
+
+	if (idev != NULL && idev->dev != &loopback_dev) {
+		struct inet6_dev *loopback_idev = in6_dev_get(&loopback_dev);
+		if (loopback_idev != NULL) {
+			rt->rt6i_idev = loopback_idev;
+			in6_dev_put(idev);
+		}
+	}
 }
 
 /*
@@ -174,9 +183,17 @@ static __inline__ struct rt6_info *rt6_device_match(struct rt6_info *rt,
 			struct net_device *dev = sprt->rt6i_dev;
 			if (dev->ifindex == oif)
 				return sprt;
-			if (dev->flags&IFF_LOOPBACK)
+			if (dev->flags & IFF_LOOPBACK) {
+				if (sprt->rt6i_idev->dev->ifindex != oif) {
+					if (strict && oif)
+						continue;
+					if (local && (!oif || 
+						      local->rt6i_idev->dev->ifindex == oif))
+						continue;
+				}
 				local = sprt;
 			}
+		}
 
 		if (local)
 			return local;
@@ -336,13 +353,13 @@ struct rt6_info *rt6_lookup(struct in6_addr *daddr, struct in6_addr *saddr,
 	return NULL;
 }
 
-/* rt6_ins is called with FREE rt6_lock.
+/* ip6_ins_rt is called with FREE rt6_lock.
    It takes new route entry, the addition fails by any reason the
    route is freed. In any case, if caller does not hold it, it may
    be destroyed.
  */
 
-static int rt6_ins(struct rt6_info *rt, struct nlmsghdr *nlh, void *_rtattr)
+int ip6_ins_rt(struct rt6_info *rt, struct nlmsghdr *nlh, void *_rtattr)
 {
 	int err;
 
@@ -390,7 +407,7 @@ static struct rt6_info *rt6_cow(struct rt6_info *ort, struct in6_addr *daddr,
 
 		dst_hold(&rt->u.dst);
 
-		err = rt6_ins(rt, NULL, NULL);
+		err = ip6_ins_rt(rt, NULL, NULL);
 		if (err == 0)
 			return rt;
 
@@ -608,8 +625,13 @@ struct dst_entry *ndisc_dst_alloc(struct net_device *dev,
 				  struct in6_addr *addr,
 				  int (*output)(struct sk_buff **))
 {
-	struct rt6_info *rt = ip6_dst_alloc();
+	struct rt6_info *rt;
+	struct inet6_dev *idev = in6_dev_get(dev);
 
+	if (unlikely(idev == NULL))
+		return NULL;
+
+	rt = ip6_dst_alloc();
 	if (unlikely(rt == NULL))
 		goto out;
 
@@ -620,7 +642,7 @@ struct dst_entry *ndisc_dst_alloc(struct net_device *dev,
 		neigh = ndisc_get_neigh(dev, addr);
 
 	rt->rt6i_dev	  = dev;
-	rt->rt6i_idev     = in6_dev_get(dev);
+	rt->rt6i_idev     = idev;
 	rt->rt6i_nexthop  = neigh;
 	atomic_set(&rt->u.dst.__refcnt, 1);
 	rt->u.dst.metrics[RTAX_HOPLIMIT-1] = 255;
@@ -731,8 +753,9 @@ int ip6_route_add(struct in6_rtmsg *rtmsg, struct nlmsghdr *nlh, void *_rtattr)
 	int err;
 	struct rtmsg *r;
 	struct rtattr **rta;
-	struct rt6_info *rt;
+	struct rt6_info *rt = NULL;
 	struct net_device *dev = NULL;
+	struct inet6_dev *idev = NULL;
 	int addr_type;
 
 	rta = (struct rtattr **) _rtattr;
@@ -744,9 +767,13 @@ int ip6_route_add(struct in6_rtmsg *rtmsg, struct nlmsghdr *nlh, void *_rtattr)
 		return -EINVAL;
 #endif
 	if (rtmsg->rtmsg_ifindex) {
+		err = -ENODEV;
 		dev = dev_get_by_index(rtmsg->rtmsg_ifindex);
 		if (!dev)
-			return -ENODEV;
+			goto out;
+		idev = in6_dev_get(dev);
+		if (!idev)
+			goto out;
 	}
 
 	if (rtmsg->rtmsg_metric == 0)
@@ -793,10 +820,17 @@ int ip6_route_add(struct in6_rtmsg *rtmsg, struct nlmsghdr *nlh, void *_rtattr)
 	 */
 	if ((rtmsg->rtmsg_flags&RTF_REJECT) ||
 	    (dev && (dev->flags&IFF_LOOPBACK) && !(addr_type&IPV6_ADDR_LOOPBACK))) {
-		if (dev)
+		if (dev && dev != &loopback_dev) {
 			dev_put(dev);
+			in6_dev_put(idev);
 			dev = &loopback_dev;
 			dev_hold(dev);
+			idev = in6_dev_get(dev);
+			if (!idev) {
+				err = -ENODEV;
+				goto out;
+			}
+		}
 		rt->u.dst.output = ip6_pkt_discard_out;
 		rt->u.dst.input = ip6_pkt_discard;
 		rt->u.dst.error = -ENETUNREACH;
@@ -838,7 +872,9 @@ int ip6_route_add(struct in6_rtmsg *rtmsg, struct nlmsghdr *nlh, void *_rtattr)
 				}
 			} else {
 				dev = grt->rt6i_dev;
+				idev = grt->rt6i_idev;
 				dev_hold(dev);
+				in6_dev_hold(grt->rt6i_idev);
 			}
 			if (!(grt->rt6i_flags&RTF_GATEWAY))
 				err = 0;
@@ -900,8 +936,8 @@ int ip6_route_add(struct in6_rtmsg *rtmsg, struct nlmsghdr *nlh, void *_rtattr)
 	if (!rt->u.dst.metrics[RTAX_ADVMSS-1])
 		rt->u.dst.metrics[RTAX_ADVMSS-1] = ipv6_advmss(dst_pmtu(&rt->u.dst));
 	rt->u.dst.dev = dev;
-	rt->rt6i_idev = in6_dev_get(dev);
-	return rt6_ins(rt, nlh, _rtattr);
+	rt->rt6i_idev = idev;
+	return ip6_ins_rt(rt, nlh, _rtattr);
 
 out:
 	if (dev)
@@ -1054,7 +1090,7 @@ void rt6_redirect(struct in6_addr *dest, struct in6_addr *saddr,
 	nrt->u.dst.metrics[RTAX_MTU-1] = ipv6_get_mtu(neigh->dev);
 	nrt->u.dst.metrics[RTAX_ADVMSS-1] = ipv6_advmss(dst_pmtu(&nrt->u.dst));
 
-	if (rt6_ins(nrt, NULL, NULL))
+	if (ip6_ins_rt(nrt, NULL, NULL))
 		goto out;
 
 	if (rt->rt6i_flags&RTF_CACHE) {
@@ -1144,7 +1180,7 @@ void rt6_pmtu_discovery(struct in6_addr *daddr, struct in6_addr *saddr,
 		dst_set_expires(&nrt->u.dst, ip6_rt_mtu_expires);
 		nrt->rt6i_flags |= RTF_DYNAMIC|RTF_CACHE|RTF_EXPIRES;
 		nrt->u.dst.metrics[RTAX_MTU-1] = pmtu;
-		rt6_ins(nrt, NULL, NULL);
+		ip6_ins_rt(nrt, NULL, NULL);
 	}
 
 out:
@@ -1303,23 +1339,26 @@ int ip6_pkt_discard_out(struct sk_buff **pskb)
 }
 
 /*
- *	Add address
+ *	Allocate a dst for local (unicast / anycast) address.
  */
 
-int ip6_rt_addr_add(struct in6_addr *addr, struct net_device *dev, int anycast)
+struct rt6_info *addrconf_dst_alloc(struct inet6_dev *idev,
+				    const struct in6_addr *addr,
+				    int anycast)
 {
 	struct rt6_info *rt = ip6_dst_alloc();
 
 	if (rt == NULL)
-		return -ENOMEM;
+		return ERR_PTR(-ENOMEM);
 
 	dev_hold(&loopback_dev);
+	in6_dev_hold(idev);
 
 	rt->u.dst.flags = DST_HOST;
 	rt->u.dst.input = ip6_input;
 	rt->u.dst.output = ip6_output;
 	rt->rt6i_dev = &loopback_dev;
-	rt->rt6i_idev = in6_dev_get(&loopback_dev);
+	rt->rt6i_idev = idev;
 	rt->u.dst.metrics[RTAX_MTU-1] = ipv6_get_mtu(rt->rt6i_dev);
 	rt->u.dst.metrics[RTAX_ADVMSS-1] = ipv6_advmss(dst_pmtu(&rt->u.dst));
 	rt->u.dst.metrics[RTAX_HOPLIMIT-1] = ipv6_get_hoplimit(rt->rt6i_dev);
@@ -1331,34 +1370,15 @@ int ip6_rt_addr_add(struct in6_addr *addr, struct net_device *dev, int anycast)
 	rt->rt6i_nexthop = ndisc_get_neigh(rt->rt6i_dev, &rt->rt6i_gateway);
 	if (rt->rt6i_nexthop == NULL) {
 		dst_free((struct dst_entry *) rt);
-		return -ENOMEM;
+		return ERR_PTR(-ENOMEM);
 	}
 
 	ipv6_addr_copy(&rt->rt6i_dst.addr, addr);
 	rt->rt6i_dst.plen = 128;
-	rt6_ins(rt, NULL, NULL);
-
-	return 0;
-}
-
-/* Delete address. Warning: you should check that this address
-   disappeared before calling this function.
- */
 
-int ip6_rt_addr_del(struct in6_addr *addr, struct net_device *dev)
-{
-	struct rt6_info *rt;
-	int err = -ENOENT;
-
-	rt = rt6_lookup(addr, NULL, loopback_dev.ifindex, 1);
-	if (rt) {
-		if (rt->rt6i_dst.plen == 128)
-			err = ip6_del_rt(rt, NULL, NULL);
-		else
-			dst_release(&rt->u.dst);
-	}
+	atomic_set(&rt->u.dst.__refcnt, 1);
 
-	return err;
+	return rt;
 }
 
 static int fib6_ifdown(struct rt6_info *rt, void *arg)