ipv6: Take rcu_read_lock in __inet6_bind for mapped addresses
authorDavid Ahern <dsahern@gmail.com>
Sat, 5 Jan 2019 15:35:04 +0000 (07:35 -0800)
committerDavid S. Miller <davem@davemloft.net>
Sat, 5 Jan 2019 22:17:07 +0000 (14:17 -0800)
commitd4a7e9bb74b5aaf07b89f6531c080b1130bdf019
tree82ae9e9a5c557f6d765aea8eaedb778cc45c1bfd
parentec90ad334986fa5856d11dd272f7f22fa86c55c4
ipv6: Take rcu_read_lock in __inet6_bind for mapped addresses

I realized the last patch calls dev_get_by_index_rcu in a branch not
holding the rcu lock. Add the calls to rcu_read_lock and rcu_read_unlock.

Fixes: ec90ad334986 ("ipv6: Consider sk_bound_dev_if when binding a socket to a v4 mapped address")
Signed-off-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
net/ipv6/af_inet6.c