Re: IPv6 payload length check?

On Wed, 5 Feb 2020, 00:45 Michael Richardson, <mcr@sandelman.ca> wrote:

>
> Mark Smith <markzzzsmith@gmail.com> wrote:
>     >> Thanks Eric.
>     >>
>     >> In this case the packet is *not* larger than the MTU of the
> outgoing link..
>     >> Rather a packet has been received (or possibly generated locally)
> with an
>     >> IPv6 payload length greater than the link layer frame size.
>     >>
>     >> Probably this packet should not be forwarded.
>     >>
>
>     > It's better to forward it without checking.
>
>     > These packets with this fault are going to be rare. Checking each
> and every
>     > packet for this rare problem at each forwarding hop would impact
> forwarding
>     > performance.
>
>     > The fault with the packet will be detected at the final destination
> and
>     > dealt with there.
>
> Yes, I agree that it will get discarded.  Maybe... Probably?
>

The upper transport layer checksum catches it at the destination because
the transport layer payload has been truncated at the fault point.

> When the hardware sends the packet, will it send it with the correct L2
> frame
> size? (i.e. equal to the L3 packet length)
> If so, what will it put at the end of the packet?
> Will it clear those bytes?

> If it doesn't, then it might send stuff from a previous transmission.
>

Probably. Those bytes cause the transport layer checksum to fail.

That is a potential information leak, which is why you shouldn't trust the
network and you should encrypt at least your transport layer payloads.

> Note that the inverse case, where L3.length < L2.length is also
> interesting,
> because if forwarded according to L2.length, then there is a covert channel
> opportunity.
>

Yes it would be.

Who polices Ethernet padding bytes being all zeros?

Regards,
Mark.

> --
> ]               Never tell me the odds!                 | ipv6 mesh
> networks [
> ]   Michael Richardson, Sandelman Software Works        | network
> architect  [
> ]     mcr@sandelman.ca  http://www.sandelman.ca/        |   ruby on
> rails    [
>
>
>
> --------------------------------------------------------------------
> IETF IPv6 working group mailing list
> ipv6@ietf.org
> Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
> --------------------------------------------------------------------
>