[krbdev.mit.edu #6566] UDP datagrams > 4K do not work.

[Ken Raeburn via RT]

On Sep 15, 2009, at 22:02, elric at mournblade.imrryr.org via RT wrote:
> Unfortunately, if you receive a datagram of over sizeof(pktbuf)
> you will succeed with cc == sizeof(pktbuf) not detecting the fact
> that there was additional data.  This results in an ASN.1 parse
> error.  What should happen is that the KDC should return an
> appropriate error to the client indicating that TCP should be used.

Regardless of other options, it sounds like cc==sizeof(pktbuf) should  
trigger the use-TCP error, since we can't distinguish between a packet  
equal in size to the buffer and a packet that was larger but got  
truncated.  Either that, or we could peek at the size of the next  
datagram and grow the buffer as needed, but I'm not sure that peeking  
can be done portably.

> Or maybe the buffer size should be increased to the maximum allowable
> for UDP.  I prefer the second option as there is nothing inherently
> wrong with 64K UDP packets.

With jumbograms, UDP messages larger than 64K are possible. (RFC  
2675)  Still, 64K does seem like a reasonable limit (i.e., way larger  
than we would normally expect).

> I noticed this while debugging a JGSS problem.  Apparently, the
> Java Kerberos libraries do not fail over from UDP to TCP unless
> the KDC specifically tells them to.  And they have no default
> setting for udp_preference_limit.  And so, if you are constructing
> tickets of over 4K because, let's say, a user is in a lot of groups
> in Windows, JGSS will just fail against an MIT KDC.

 From what I've read, the common wisdom still seems to be that some  
gateways/routers/NAT boxes/firewalls/whatever will not properly  
process UDP fragments, so UDP traffic over ~1500 bytes (or less) may  
never get to the KDC.  So this sounds like a bug in the Java Kerberos  
libraries.

Ken

-- 
Ken Raeburn / raeburn at mit.edu / no longer at MIT Kerberos Consortium