Re: Re: xPL bridging loop prevention

["Mark Hindess" <xpl@xxxxxxxxxxxxxxxxx>]

[ Answer your own messages is always a bad sign. ;-) ]

On 18 November 2005 at 20:34, Mark Hindess <xpl@xxxxxxx> wrote:
>
> On 17 November 2005 at 22:53, Mal Lansell <mal@xxxxxxx> wrote:
> >
> > Actually, no :-)  The hop count should be left as it is.  I asked
this
> > very same question a few weeks ago, and the answer was that the
hop is
> > incremented only when the underlying network changes - such as
when a
> > message moves from an RS232 to an IP network.
>
> Oh no, that's another decision, I struggle to understand. :-(
> I'll have to read the mailing list archive.

Ok.  I've read the previous discussion at:

http://www.ukha-archive.com/ml/xpl/2005-oct/msg00080.html

specifically statements like:

it's there are a safety feature given that xpl was always intended to
run across multiple network systems (and therefore encounter possible
bridging issues)

and the statement in the spec that says:

This is incremented each time the xPL message is transferred from one
physical network to another, for instance by a bridge application
passing traffic between an RS485 Serial bus and Ethernet.

It seems to me that my bridge is transferring messages from one physical
network to another - wifi to ethernet for instance or just different
physical ethernets.

I think (and my unit testing has shown!) that if anything it is easier
to create catastrophic loops with "virtual" links like a software
bridge
than it will be with bridges that link physical networks.

With this in mind, any bridge code I release will be increasing the hop
count by default.  I'd not feel happy releasing code that didn't take
this simple step to protect users.  Users will have the option to turn
this behaviour off but I'd certainly not recommend it.

I'm really not trying to be deliberately controversial.  If anyone
feels that I am please tell me and I'll stop this discussion and, with
regrets, keep my bridge code private.

Regards,
Mark.