Re: IPv6 prefix route lifetime

[Roy Marples <roy%marples.name@localhost>]

On 2016-08-09 11:45, Ryota Ozaki wrote:
> On Tue, Aug 9, 2016 at 6:58 PM, Kengo NAKAHARA <k-nakahara%iij.ad.jp@localhost> 
> wrote:
> > I think the packet might not be routed to gif(4) interface. To ensure
> > that, could you try to apply below patch and show me the ping output
> > message?
> > ====================
> > diff --git a/tests/net/if_gif/t_gif.sh b/tests/net/if_gif/t_gif.sh
> > index 6337f23..28ef7bf 100644
> > --- a/tests/net/if_gif/t_gif.sh
> > +++ b/tests/net/if_gif/t_gif.sh
> > @@ -467,6 +467,12 @@ test_recursive_check()
> >                         rump.ping -n -w $TIMEOUT -c 1 
> > $ROUTER2_GIFIP_RECURSIVE2
> >         fi
> >
> > +       if [ ${mode} = "ipv6" ]; then
> > +           rump.ping6 -n -X $TIMEOUT -c 1 $ROUTER2_GIFIP6_RECURSIVE2
> > +       else
> > +           rump.ping -n -w $TIMEOUT -c 1 $ROUTER2_GIFIP_RECURSIVE2
> > +       fi
> > +
> >         atf_check -o match:'gif0: recursively called too many times' \
> >                 -x "$HIJACKING dmesg"
> > ====================
>
> I run tests w/o and w/ the patch. The diff between netstat -nr -f inet6
> of the two runs at that point is below. There are less routes to
> gif interfaces.

So there are!
I refreshed the patch for -current to try and address this.
http://www.netbsd.org/~roy/in6_ifa_route.patch

Can you test it please as rump still doens't work for me?

Thanks

Roy