Re: [PATCH] ipv6: no addrconf for slave devices

On Fri, Oct 16, 2015 at 6:14 PM, David Ahern [off-list ref] wrote:

On 10/16/15 10:12 AM, Jan Blunck wrote:

quoted

On Fri, Oct 16, 2015 at 6:02 PM, David Ahern [off-list ref]
wrote:

quoted

On 10/16/15 9:57 AM, Jan Blunck wrote:

quoted

I don't think that enslaved ports should get network layer addresses.
This is one example with a team device:

for VRF devices we do want the enslaved links to have link local
addresses.

That is interesting. As far I can see you are setting IFF_SLAVE in
do_vrf_add_slave() and therefore already stop IPv6 addrconf.

Check net-next. That had to be removed to get IPv6 working.

Thanks for the pointer.

So it would be better to differentiate between L2 and L3 ports and
only start addrconf on later ones? I don't think there is a flag that
allows for that though.