>
> On 16-Nov-00 JINMEI Tatuya wrote:
> > Applications can dynamically grab the path MTU by the IPV6_RECVPATHMTU
> > socket option (upon receiving an ICMPv6 packet toobig error).
> >
>
> Is that mean, when IPV6_RECVPATHMTU option is setted (UDP/raw socket), that
> fragmentation is disabled ?
No, I don't think so.
IPV6_RECVPATHMTU is so that e.g. UDP applications that do packetization
can determine when the path MTU changes and adjust the size of the
packets they are sending.
Should we just define the semantic equivalent of setting IP_DF on
a raw socket with IP_HDRINCL set for IPv6? (IPV6_DONTFRAG?)
This would allow a traceroute to detect the MTU, and it would allow
the above UDP example application to avoid "accidental" fragmentation
by a too helpful IP/UDP layer. (The IP/UDP layer needs to be helpful for
UDP applications which are not involved in PMTU discovery.)
Erik
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive: ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------
