resolving routes externally

[Martin Eugen]

On Tue, 23 Nov 2004 14:52:36 +0100, Joerg Sonnenberger
<joerg at britannica.bec.de> wrote:
> On Tue, Nov 23, 2004 at 11:42:39AM -0200, Jo?o Carlos Mendes Lu?s wrote:
> > >So I started to look at the ARP
> > >code, but it of course lacks the kernel - userland communication
> > >interface. I would appreciate any ideas about what would be the easier
> > >way to implement such a thing where the kernel could wait (up to some
> > >reasonable time-out) a userland daemon to install a new route.
> >
> > Why don't you simply discard the packet and wait for the next retry?
>

Because the network is not like the internet, packet error correction
and so on is done at lower layers, I mean... if there are some packets
that are equivalent to the TCP SYNs, the 'SYN' timeout in our case is
in minutes (because it is believed the host or a link is down or
something else that could take longer time to resolve). This is bad,
because connections will be established within some minutes...
 
> Or alternatively use an internal queue of limited size to keep track of
> those packages.

This is probably the only solution I can think of right now, but I
think poking a queue at regular, short intervals seems to me quite
expensive, isn't it? Or perhaps there could be a netgraph node that
handles the queue and connects to the userland daemon... but this
could make things much more complicated... ?

> 
> Joerg
>