Re: unregister_netdevice: waiting for eth0 to become free. Usage count = 1

[Wei Wang]

On Fri, Aug 11, 2017 at 5:19 PM, David Ahern <dsahern@xxxxxxxxx> wrote:
> On 8/11/17 6:10 PM, Wei Wang wrote:
>> I think we have a potential fix for this issue.
>> Martin and I found that when addrconf_dst_alloc() creates a rt6, it is
>> possible that rt6->dst.dev points to loopback device while
>> rt6->rt6i_idev->dev points to a real device.
>> When the real device goes down, the current fib6 clean up code only
>> checks for rt6->dst.dev and assumes rt6->rt6i_idev->dev is the same.
>> That leaves unreleased refcnt on the real device if rt6->dst.dev
>> points to loopback dev.
>
> Yes, host routes and anycast routes.
>
> I have a patch to fix that but it is held up on a few VRF test cases
> failing. Hopefully I can get that figured out next week. These unrelated
> routes against the loopback device have been a source of a number of
> problems (e.g. take down 'lo' and all of IPv6 networking stops for that
> namespace).

Thanks David.
By "a patch to fix that" do you mean after your patch, for every rt6,
rt6->rt6i_idev will be the same as rt6->dst.dev?