This was causing the channel list to remain stuck at the old value
when we were rerouting.

Conflicts:
	CHANGES

Recent linux kernels (roughly 3.x) keep a routing cache for IPv6: each
time a host is contacted, a temporary /128 route is added. These
routes are visible using “ip -6 route show cache”.

Asking babeld to redistribute all IPv6 routes known to the host, using
"redistribute ip ::/0", causes it to announce and de-announce these
/128 routes frequently, leading other nodes to mark them as
unreachable (thus preventing communication with the host described by
the /128 route, even though less specific routes are available).

Interface_up can be called before an IPv6 address has been assigned to
the interface.  In this case, we need to down the interface again, wait
for a few seconds, then attempt to up it again.

Thanks to Julien Cristau.

Thanks to Julien Cristau.

This reverts commit 4229bea0.

The logic that attempted to optimise this away was buggy since it
wasn't taking IHU expiration times into account.  At any rate,
this optimisation is probably not worth the complexity.

This was requested by people who want to parse the local socket output
using sed/awk scripts.

Of course rp_filter should be reset to its old value, and not to the old
accept_redirects value.
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>

When a route is retracted, we want to switch to an alternate route
straight away even if the smoothed metric hasn't converged yet.
Setting the smoothed metric to infinity in this case makes this happen.

Use the hostname instead of "alamakota" when announcing the router
on the local interface.  It would make it possible for visualisation
tools, such as babelweb, to display a meaningful name automatically.

Off-by-one in error checking.  Reported by Gioacchino Mazzurco.

This used to mean "parasitic", but has been removed a long time ago.

We're working with deltas, not with absolute values, so we need to
subtract 1.  Oh, and it's multiplication, not division.

This would cause the smoothed metric to instantaneously jump to
the true metric when route_smoothed_metric was called twice in the
same second.

This caused a minor loss of accuracy.

The netlink API doesn't natively support atomic route changes.  We used
to attempt to install the new route before removing the old route, but that
would fail in mysterious ways on non-multipath kernels, leading to "stuck"
routes.  Avoid the pain, just take the native approach.

This makes Babel work (to the extent possible) over BATMAN interfaces, which
apparently some people are doing.  Thanks to Gioacchino Mazzurco and Sven
Eckelmann.

This caused empty lines to cause parse failures.

Reported by Antoine Beaupré.

This is a systematic way of implementing hyseresis in route
selection, and should solve the issues we're seeing with ETX
jitter and feedback.  The default half-time is 4s, which might
be too small.