>> Then it's the IP encapsulation layer that's broken, [...].

>> Any provider that tries to sell a thing such as you describe as an
>> IP transport should be prosecuted for fraud.

> Why ? It's not really his fault.  [...]  The problem is on *my*
> equipement.

Ah.  That wasn't clear to me at the start.

Then, well, it's your fault for trying to send IP over a link not
suited to it, without a suitable impedance-matching layer.

While that's fine - sending IP over all sorts of things that one
wouldn't normally expect it over is a long-standing tradition and in
many cases a lot of fun - I see no need to complicate everyone else's
setups to support it.  Research is fine, and NetSBD is a great vehicle
for it, but until it makes the jump from research to engineering - from
experiment to practice - I don't see it as appropriate for the tree.

Not that I'm an arbiter of such appropriateness, of course.

I've also rather lost track of the connection to the original subject
matter of the discussion.

> And I can't change it (where can I buy an ATMF25 sbus adapter, with
> documentation to write a driver ?)

Sure you can change it.  You're just not willing to pay the price it
would entail.  (The cheapest price would probably be to get a low-end
x86 box for your end of it.  Other alternatives might include such
things as switching to a different grade of service.)

/~\ The ASCII				der Mouse
\ / Ribbon Campaign
 X  Against HTML	       mouse@rodents.montreal.qc.ca
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B