fixing out of order first fragment processing?

[othermark]

Maxim Konovalov wrote:
>> 2. are there workarounds/patches?
> 
> First of all, I can't reproduce this behaviour.  "ping -c 1 -s 4500
> freebsd" from the linux produces a pseudo tcpdump liek below:

This turned out to a problem with having 'pf' installed on the 
freebsd target.  I mis-assumed it had to do with the stack itself, but
turning on fragment re-assembly in pf 'fixed' the problem:

scrub in all random-id no-df fragment reassemble

Another -current box I just tested against linux re-assembles fine.

I was reasonably sure from doing some searches on the net that this
was a freebsd problem, but it turned out not to be.

Thanks for looking!
-- 
othermark
atkin901 at nospam dot yahoo dot com
(!wired)?(coffee++):(wired);