bonding: Prevent IPv6 link local address on enslaved devices
authorKarl Heiss <kheiss@gmail.com>
Mon, 11 Jan 2016 13:28:43 +0000 (08:28 -0500)
committerDavid S. Miller <davem@davemloft.net>
Mon, 11 Jan 2016 22:55:41 +0000 (17:55 -0500)
commit03d84a5f83a67e692af00a3d3901e7820e3e84d5
tree37e212b0b29af57ebc79f518c5ac76ffce541c9d
parent40ba330227ad00b8c0cdf2f425736ff9549cc423
bonding: Prevent IPv6 link local address on enslaved devices

Commit 1f718f0f4f97 ("bonding: populate neighbour's private on enslave")
undoes the fix provided by commit c2edacf80e15 ("bonding / ipv6: no addrconf
for slaves separately from master") by effectively setting the slave flag
after the slave has been opened.  If the slave comes up quickly enough, it
will go through the IPv6 addrconf before the slave flag has been set and
will get a link local IPv6 address.

In order to ensure that addrconf knows to ignore the slave devices on state
change, set IFF_SLAVE before dev_open() during bonding enslavement.

Fixes: 1f718f0f4f97 ("bonding: populate neighbour's private on enslave")
Signed-off-by: Karl Heiss <kheiss@gmail.com>
Signed-off-by: Jay Vosburgh <jay.vosburgh@canonical.com>
Reviewed-by: Jarod Wilson <jarod@redhat.com>
Signed-off-by: Andy Gospodarek <gospo@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
drivers/net/bonding/bond_main.c