Greg A. Woods writes:
> What you've said (about the protocol being fine vs. needing to pay
> attention to ICMP issues) is somewhat self contradictory.  It doesn't
> really matter how the ICMP "needs frag" packet gets lost if its loss can
> cause something like a TCP/IP connection to fail.

Greg, this is an IETF issue, not a NetBSD issue. We're doing what we
need to. Firewall and NAT vendors are starting to figure out that they 
can't be lazy about this. The problem is not ours and will correct
itself.

Perry