From: David Ahern <dsahern@gmail.com>
Date: Fri, 4 Jan 2019 16:58:15 -0800
Subject: ipv6: Consider sk_bound_dev_if when binding a socket to a v4 mapped
 address
Git-commit: ec90ad334986fa5856d11dd272f7f22fa86c55c4
Patch-mainline: v5.0-rc3
References: networking-stable-19_01_22

Similar to c5ee066333eb ("ipv6: Consider sk_bound_dev_if when binding a
socket to an address"), binding a socket to v4 mapped addresses needs to
consider if the socket is bound to a device.

This problem also exists from the beginning of git history.

Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
---
 net/ipv6/af_inet6.c |   11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)

--- a/net/ipv6/af_inet6.c
+++ b/net/ipv6/af_inet6.c
@@ -316,6 +316,7 @@ int inet6_bind(struct socket *sock, stru
 
 	/* Check if the address belongs to the host. */
 	if (addr_type == IPV6_ADDR_MAPPED) {
+		struct net_device *dev = NULL;
 		int chk_addr_ret;
 
 		/* Binding to v4-mapped address on a v6-only socket
@@ -326,9 +327,17 @@ int inet6_bind(struct socket *sock, stru
 			goto out;
 		}
 
+		if (sk->sk_bound_dev_if) {
+			dev = dev_get_by_index_rcu(net, sk->sk_bound_dev_if);
+			if (!dev) {
+				err = -ENODEV;
+				goto out;
+			}
+		}
+
 		/* Reproduce AF_INET checks to make the bindings consistent */
 		v4addr = addr->sin6_addr.s6_addr32[3];
-		chk_addr_ret = inet_addr_type(net, v4addr);
+		chk_addr_ret = inet_addr_type_dev_table(net, dev, v4addr);
 		if (!net->ipv4.sysctl_ip_nonlocal_bind &&
 		    !(inet->freebind || inet->transparent) &&
 		    v4addr != htonl(INADDR_ANY) &&