Jason,
    I think, and so does Dave, that you're closer to the source of the
problem with this line of thought!  more....

>Remember the "ping of doom"? Pings sent to a machine that were larger 
>
Yep.
>
>Likewise, it could be that something that NetBSD is sending something 
>(either valid or invalid) that exposes a bug in linux / pppd. 

Right on!  And Dave suggested a sysctl regarding RFC1323 to disable
something that it may be doing.....precisely!  Oh, its not the Linux
pppd that I think is being affected, but it is the IP_masquerading 
code (the NAT functionality) that is breaking, if anything, that leads
to this fault.  I think that is the weak link here.

The reason is, that there are limits on, and the need to provide
special mechanisms for some protocols and/or UDP's for example.  
Plus, my IP_masq code is old, and needs to be updated.  Thats why
I think its the weak link in this, and probably what is breaking.

In light of that, the question becomes what is it, if anything, that
might be happening during the setup of the session being originated
on NetBSD that is exposing the fault.  Dave may have id'd it!

Dave and I have been having a spirited discussion on the side on this,
and I'm hoping it has led to us  finding something significant on it.
I missed it the first time he said it....guess I was tired and more
than just a little frustrated...but I'll be able to try it later and
I'll be sure to let everyone know what the outcome is.

Thanks!
Paul