[quagga-dev 5585] Bug 445 redux: bgpd picks wrong nexthop on point-to-point interfaces

John Morrissey jwm at horde.net
Sat Jul 26 01:12:27 BST 2008

I've posted before, but I'd like to make a public plea for someone to have a
quick look at Bug 445 (http://bugzilla.quagga.net/show_bug.cgi?id=445).

AFAICT, it looks like bgpd(8) is choosing the wrong end of GRE tunnels when
determining valid next hops (it decides that the *local* end of the tunnel
is the only acceptable next hop for received announcements, not the *remote*
end of the tunnel).

I dug around the bgpd source a bit more, but I'm failing at my attempts to
find where the nexthop detection logic goes wrong. This isn't the most
widespread showstopper in the world, but I've done a lot of the legwork and
it seems all that's left is for someone with more knowledge of the source to
look over this bug to see if my diagnosis is sane and check out what's
happening during nexthop selection.

If someone with that knowledge could take a quick look at this bug, I'd
really appreciate it.

John Morrissey          _o            /\         ----  __o
jwm at horde.net        _-< \_          /  \       ----  <  \,
www.horde.net/    __(_)/_(_)________/    \_______(_) /_(_)__

More information about the Quagga-dev mailing list