Re: Path MTU Discovery

[[e-mail] (Dan McDonald)]

> >
> >An encrypting router often has a "tunnel interface" that'll have properties
> >like:
> >
> >	tun0:	10.69.0.0/16  --> 10.9.1.25
> 
> Not necessarilly. Virtual interfaces might not be used, or you might
> have more than one SPIs using the same interface. Per SPI(-chain) MTUs
> need to be kept. If you have a router that behaves as you describe,
> then you don't need any other provision.

Hmmm.  In a router-to-host tunneling case, you MIGHT be right.

Let's consider one other property about router-to-* tunneling.  Since the
router is the _originator_ of the outermost packet, it does not need to obey
the inner packet's "Don't fragment" bit.  Remember, I'm originating the
packet.

There's one faulty assumption you make below.  I'll point it out in a bit.
This may explain why we aren't agreeing here.

> Which is the same as keeping that information on a per-SPI(-chain)
> basis, if you don't use ViFs. 

Could you illustrate a non-VIF case?  Granted, in the NRL code, there's an
"encrypting route," but even there, I just don't see any of the MTU
problems.

> Another point is that fragmentation checking should be done before any
> IPsec handling takes place (easier and faster).

WRONG FOR OUTBOUND PACKETS!!!  This is in clear violation of RFC 1825.  Lemme
quote:

>> 3.1 AUTHENTICATION HEADER
 
<SNIP!>

>>   Fragmentation occurs after the Authentication Header processing for
>>   outbound packets and prior to Authentication Header processing for
>>   inbound packets.  The receiver verifies the correctness of the

There actually isn't text in the ESP section, but I'll bet small sums that
either Ran A. or Steve K. will back me up on this one.

If you meant inbound packets, my bad.

> Only if you can associate a packet "flow" with the ICMP message; easy
> to do if you use a ViF for each tunnel.

Any implementors out there not doing this?  I can only see this in the
router-to-host tunnel.

> So TCP has to "forsee" whether a packet will go through IPsec, and if
> so what will be the size overhead.

It'll know.  Before that SYN or SYN/ACK gets sent, it can consult endpoint
state (socket, pcb, whatever) and make an appropriate guess.

Dan