Merge branch 'vrf-next'

David Ahern says:

====================
net: vrf: Fix ipv6 source address selection

IPv6 address selection is currently messed up for several use cases such
as unnumbered deployments with global addresses on the VRF device and none
on the enslaved devices.

Update the source address selection to consider the real output route as
opposed to the VRF route that sends packets to the VRF device first (ie.,
implement get_saddr6 similar to the IPv4 method) and update the IPv6
address selection to consider L3 domains and preference for addresses on
the VRF device).
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 48 insertions, 0 deletions

diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c
index 6c8fc3f96b11..a1f6b7b31531 100644
--- a/net/ipv6/addrconf.c
+++ b/net/ipv6/addrconf.c
@@ -1524,6 +1524,28 @@ out:
 	return hiscore_idx;
 }
 
+static int ipv6_get_saddr_master(struct net *net,
+				 const struct net_device *dst_dev,
+				 const struct net_device *master,
+				 struct ipv6_saddr_dst *dst,
+				 struct ipv6_saddr_score *scores,
+				 int hiscore_idx)
+{
+	struct inet6_dev *idev;
+
+	idev = __in6_dev_get(dst_dev);
+	if (idev)
+		hiscore_idx = __ipv6_dev_get_saddr(net, dst, idev,
+						   scores, hiscore_idx);
+
+	idev = __in6_dev_get(master);
+	if (idev)
+		hiscore_idx = __ipv6_dev_get_saddr(net, dst, idev,
+						   scores, hiscore_idx);
+
+	return hiscore_idx;
+}
+
 int ipv6_dev_get_saddr(struct net *net, const struct net_device *dst_dev,
 		       const struct in6_addr *daddr, unsigned int prefs,
 		       struct in6_addr *saddr)
@@ -1577,13 +1599,39 @@ int ipv6_dev_get_saddr(struct net *net, const struct net_device *dst_dev,
 		if (idev)
 			hiscore_idx = __ipv6_dev_get_saddr(net, &dst, idev, scores, hiscore_idx);
 	} else {
+		const struct net_device *master;
+		int master_idx = 0;
+
+		/* if dst_dev exists and is enslaved to an L3 device, then
+		 * prefer addresses from dst_dev and then the master over
+		 * any other enslaved devices in the L3 domain.
+		 */
+		master = l3mdev_master_dev_rcu(dst_dev);
+		if (master) {
+			master_idx = master->ifindex;
+
+			hiscore_idx = ipv6_get_saddr_master(net, dst_dev,
+							    master, &dst,
+							    scores, hiscore_idx);
+
+			if (scores[hiscore_idx].ifa)
+				goto out;
+		}
+
 		for_each_netdev_rcu(net, dev) {
+			/* only consider addresses on devices in the
+			 * same L3 domain
+			 */
+			if (l3mdev_master_ifindex_rcu(dev) != master_idx)
+				continue;
 			idev = __in6_dev_get(dev);
 			if (!idev)
 				continue;
 			hiscore_idx = __ipv6_dev_get_saddr(net, &dst, idev, scores, hiscore_idx);
 		}
 	}
+
+out:
 	rcu_read_unlock();
 
 	hiscore = &scores[hiscore_idx];