Hi, > -----Original Message----- > From: ietf [mailto:ietf-bounces@xxxxxxxx] On Behalf Of Masataka Ohta > Sent: Friday, April 10, 2015 5:26 AM > To: ietf@xxxxxxxx > Cc: int-area@xxxxxxxx > Subject: Re: Last Call: <draft-ietf-intarea-gre-mtu-02.txt> (A Widely-Deployed Solution To The Generic Routing Encapsulation (GRE) > Fragmentation Problem) to Informational RFC > > As the draft says; > > o When the GRE ingress node receives a non-fragmentable packet with > length greater than the GMTU, it discards the packet and send an > ICMP PTB message to the packet's source. > > the draft should clearly state that, if GMTU<1280B, it is a violation > of the following requirement of RFC2460: > > IPv6 requires that every link in the internet have an MTU of 1280 > octets or greater. On any link that cannot convey a 1280-octet > packet in one piece, link-specific fragmentation and reassembly must > be provided at a layer below IPv6. > > and that 1280B IPv6 packets can not be carried over IPv6 with the > default GRE configuration. We have been through this already. They want to say that widely deployed implementations already ignore this requirement and so they want to document the behavior to make it all OK. The non-robustness principle in action. Thanks - Fred fred.l.templin@xxxxxxxxxx > It is especially so, because, according to the draft: > > Typically, GRE ingress nodes further refine their GMTU estimate by > executing PMTUD procedures. However, if an implementation supports > PMTUD for GRE tunnels, it also includes a configuration option that > disables PMTUD. This configuration option is required to mitigate > certain denial of service attacks (see Section 5). > > PMTUD is often turned off and, then, RFC2460 requires GMTU<1280B. > > Also, I think the paragraph above is not very honest on the reason > why PMTUD is often turned off. > > Masataka Ohta