FreeBSD, IPv6 and World IPv6 Day

[Matthew Luckie]

>>> I found measurement results on this website:
>>> http://hide.dnsalias.net/aaaa/worldipv6day.cgi
>>
>> Some more results:
>>
>> http://www.wand.net.nz/~mluckie/ipv6day/
>
> Interesting.  Did you post on the v6 lists as well?

Just on ipv6-techsig at listserv.internetnz.net.nz
Feel free to forward on to other lists if you think its useful.

>> http://www.freebsd.org/cgi/query-pr.cgi?pr=145733
>
> That's the frag6 ipfw iusse, right? Well actually 3 issues. I have this
> is open in my other window today after having stared at it for too long.
>
> BTW. does anyone have any idea what kind of systems generate these packets?
> Does anyone have packet dumps of them?  A couple of people would be
> curious to have/see that.

In terms of systems that will respond to a PTB by sending IPv6 fragments 
rather than sending smaller TCP packets:

http://www.kddi.com
http://dream.jp/
http://www.ubc.ca/
http://www.goneo.de/
http://www.nict.go.jp/

(I can go on further if you want)

If I had to guess at what OS sends these packets I'd guess netbsd or 
openbsd but I've never looked.  To reproduce above, use

http://www.wand.net.nz/scamper/scamper-cvs-20110608.tar.gz
http://www.wand.net.nz/scamper/pmtud

Use -M 1480, that will cause the last fragment to be small and trigger 
one of the bugs in the PR.

In terms of systems that will respond to a PTB with nhmtu < 1280, it 
seems most follow the advice in RFC2460 and send packets with a 
fragmentation header without fragmenting the packet.  Use -M 576 with 
your scamper command line.