Re: IPv6 routes inserted into the kernel with 'route' end up with invalid type

Re: IPv6 routes inserted into the kernel with 'route' end up with invalid type

[David Ahern]

On 8/23/19 8:43 AM, Toke Høiland-Jørgensen wrote:
> Hi David
> 
> Tom noticed[0] that on newer kernels, the Bird routing daemon rejects IPv6
> routes received from the kernel if those routes were inserted with the
> old 'route' utility (i.e., when they're inserted through the ioctl
> interface).
> 
> We tracked it down to the routes having an rtm_type of RTN_UNKNOWN, and
> a bit of git archaeology points suggestively at this commit:
> 
> e8478e80e5a ("net/ipv6: Save route type in rt6_info")
> 
> The same setup works with older kernels, so this seems like it's a
> regression, the age of 'route' notwithstanding. Any good ideas for the
> proper way to fix this?
> 

Should be fixed by:

commit c7036d97acd2527cef145b5ef9ad1a37ed21bbe6
Author: David Ahern <dsahern@gmail.com>
Date:   Wed Jun 19 10:50:24 2019 -0700

    ipv6: Default fib6_type to RTN_UNICAST when not set

    A user reported that routes are getting installed with type 0
(RTN_UNSPEC)
    where before the routes were RTN_UNICAST. One example is from accel-ppp
    which apparently still uses the ioctl interface and does not set
    rtmsg_type. Another is the netlink interface where ipv6 does not require
    rtm_type to be set (v4 does). Prior to the commit in the Fixes tag the
    ipv6 stack converted type 0 to RTN_UNICAST, so restore that behavior.

    Fixes: e8478e80e5a7 ("net/ipv6: Save route type in rt6_info")
    Signed-off-by: David Ahern <dsahern@gmail.com>
    Signed-off-by: David S. Miller <davem@davemloft.net>

Re: IPv6 routes inserted into the kernel with 'route' end up with invalid type

[Toke Høiland-Jørgensen]

David Ahern <dsahern@gmail.com> writes:

> On 8/23/19 8:43 AM, Toke Høiland-Jørgensen wrote:
>> Hi David
>> 
>> Tom noticed[0] that on newer kernels, the Bird routing daemon rejects IPv6
>> routes received from the kernel if those routes were inserted with the
>> old 'route' utility (i.e., when they're inserted through the ioctl
>> interface).
>> 
>> We tracked it down to the routes having an rtm_type of RTN_UNKNOWN, and
>> a bit of git archaeology points suggestively at this commit:
>> 
>> e8478e80e5a ("net/ipv6: Save route type in rt6_info")
>> 
>> The same setup works with older kernels, so this seems like it's a
>> regression, the age of 'route' notwithstanding. Any good ideas for the
>> proper way to fix this?
>> 
>
> Should be fixed by:
>
> commit c7036d97acd2527cef145b5ef9ad1a37ed21bbe6
> Author: David Ahern <dsahern@gmail.com>
> Date:   Wed Jun 19 10:50:24 2019 -0700
>
>     ipv6: Default fib6_type to RTN_UNICAST when not set
>
>     A user reported that routes are getting installed with type 0
> (RTN_UNSPEC)
>     where before the routes were RTN_UNICAST. One example is from accel-ppp
>     which apparently still uses the ioctl interface and does not set
>     rtmsg_type. Another is the netlink interface where ipv6 does not require
>     rtm_type to be set (v4 does). Prior to the commit in the Fixes tag the
>     ipv6 stack converted type 0 to RTN_UNICAST, so restore that behavior.
>
>     Fixes: e8478e80e5a7 ("net/ipv6: Save route type in rt6_info")
>     Signed-off-by: David Ahern <dsahern@gmail.com>
>     Signed-off-by: David S. Miller <davem@davemloft.net>


Ah, great! Guess that hasn't made its way to the stable and distribution
kernels yet. Thanks for the pointer! :)

-Toke