> >> OK, but if those extra payloads are disguised as some notification
> >> (there is no payload actually called “info”), then responders do tend
> to ignore notifications they don’t recognize.
> >
> > True, but in this case the inputs to the hash function will be
> > different (you need to insert Notification payload header in the m`),
> so the attack will fail.
> 
> I must correct myself - if attacker takes care and puts the Notify
> payload header at the end of ck it sends to initiator (and he must
> correctly guess the length of info`), then it will work - all the
> original payloads from the initiator will appear inside Notification
> payload and will be ignored by responder.
> 

[HJ] how would this work? even attack append a notification header in ck, and 
return a ck= C1|SAi'|g^x'|ni| notify_header, then m1(from real 
initiator)=HDR|(C1|SAi'|g^x'|ni| notify_header)_as_ck|SAi|g^x|ni|nat-t, notify 
header is part of ck, won't be parsed as actual notify payload header.
_______________________________________________
IPsec mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/ipsec

Reply via email to