On 06/05/15 23:12, Toke Høiland-Jørgensen wrote: > Simon Kelley <si...@thekelleys.org.uk> writes: > >> The MTU if the SIXXs IPv6 network interface is 1428. Failure to >> receive UDP packets larger than the MTU is a bigger bug than DNS, but >> I don't know if it's a SIXXS problem or a wider IPv6 one. > > Well, IPv6 doesn't fragment packets; hosts are supposed to do PMTU > discovery and transmit at the MTU that works end-to-end.
It's difficult to see how that would work in practise for DNS. Take the Google-public-DNS example. It's clearly not sane for Google's servers to do PMTU on the address of every client, just to send one UDP packet, and caching PMTU for clients is going to get hard, fast. If the reply-size is bigger than the PMTU what happens anyway? Presumably a truncated response, to force fallback to TCP. I guess it could work the other way around, and the client finds the PMTU to the server, and puts that in the EDNS0 max packet field. Are PMTU symetric? In any case, good luck updating every stub resolver in the world to do this...... Cheers, Simon. > I've found this > to be broken way too often for comfort... Even encountered operators who > just told me to increase the MTU at my end rather than fix their > discovery mechanism (though this was when I was running with the minimum > 1280 MTU and they were doing ethernet max-sized packets). > > -Toke > _______________________________________________ Dnsmasq-discuss mailing list Dnsmasq-discuss@lists.thekelleys.org.uk http://lists.thekelleys.org.uk/mailman/listinfo/dnsmasq-discuss
