[PATCH net] ipv6: Consider sk_bound_dev_if when binding a socket to an address

[PATCH net] ipv6: Consider sk_bound_dev_if when binding a socket to an address

[David Ahern]

From: David Ahern <dsahern@gmail.com>

IPv6 does not consider if the socket is bound to a device when binding
to an address. The result is that a socket can be bound to eth0 and then
bound to the address of eth1. If the device is a VRF, the result is that
a socket can only be bound to an address in the default VRF.

Resolve by considering the device if sk_bound_dev_if is set.

This problem exists from the beginning of git history.

Signed-off-by: David Ahern <dsahern@gmail.com>
---
 net/ipv6/af_inet6.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/net/ipv6/af_inet6.c b/net/ipv6/af_inet6.c
index f0cd291034f0..0bfb6cc0a30a 100644
--- a/net/ipv6/af_inet6.c
+++ b/net/ipv6/af_inet6.c
@@ -350,6 +350,9 @@ static int __inet6_bind(struct sock *sk, struct sockaddr *uaddr, int addr_len,
 					err = -EINVAL;
 					goto out_unlock;
 				}
+			}
+
+			if (sk->sk_bound_dev_if) {
 				dev = dev_get_by_index_rcu(net, sk->sk_bound_dev_if);
 				if (!dev) {
 					err = -ENODEV;
-- 
2.11.0

Re: [PATCH net] ipv6: Consider sk_bound_dev_if when binding a socket to an address

[David Miller]

From: David Ahern <dsahern@kernel.org>
Date: Wed,  2 Jan 2019 18:57:09 -0800

> From: David Ahern <dsahern@gmail.com>
> 
> IPv6 does not consider if the socket is bound to a device when binding
> to an address. The result is that a socket can be bound to eth0 and then
> bound to the address of eth1. If the device is a VRF, the result is that
> a socket can only be bound to an address in the default VRF.
> 
> Resolve by considering the device if sk_bound_dev_if is set.
> 
> This problem exists from the beginning of git history.
> 
> Signed-off-by: David Ahern <dsahern@gmail.com>

Applied and queued up for -stable.

Re: [PATCH net] ipv6: Consider sk_bound_dev_if when binding a socket to an address

[Ralf Jung]

Hi all,

thanks a lot!  I can confirm that this fixes my test script.

Kind regards,
Ralf

On 03.01.19 03:57, David Ahern wrote:
> From: David Ahern <dsahern@gmail.com>
> 
> IPv6 does not consider if the socket is bound to a device when binding
> to an address. The result is that a socket can be bound to eth0 and then
> bound to the address of eth1. If the device is a VRF, the result is that
> a socket can only be bound to an address in the default VRF.
> 
> Resolve by considering the device if sk_bound_dev_if is set.
> 
> This problem exists from the beginning of git history.
> 
> Signed-off-by: David Ahern <dsahern@gmail.com>
> ---
>  net/ipv6/af_inet6.c | 3 +++
>  1 file changed, 3 insertions(+)
> 
> diff --git a/net/ipv6/af_inet6.c b/net/ipv6/af_inet6.c
> index f0cd291034f0..0bfb6cc0a30a 100644
> --- a/net/ipv6/af_inet6.c
> +++ b/net/ipv6/af_inet6.c
> @@ -350,6 +350,9 @@ static int __inet6_bind(struct sock *sk, struct sockaddr *uaddr, int addr_len,
>  					err = -EINVAL;
>  					goto out_unlock;
>  				}
> +			}
> +
> +			if (sk->sk_bound_dev_if) {
>  				dev = dev_get_by_index_rcu(net, sk->sk_bound_dev_if);
>  				if (!dev) {
>  					err = -ENODEV;
>