Re: Proposal for changing how IPv6 Hop-by-Hop Options are processed

["Joel M. Halpern" <[e-mail]>]

Heavily trimmed

On 12/7/2020 12:32 PM, Templin (US), Fred L wrote:
...
> I am not concerned about a Linux host; I am concerned about IPv6 routers.
> Do IPv6 routers look at extension headers beyond the first one trying to see
> if there is an out-of-place HbH header and then drop the packet if there is?
> If so, I don't think they should.
> 
> Fred

This sounds like the classic IP Header checksum case.
The practical answer was that while the standard said that an invalid 
IPv4 header checksum required dropping the packet, most routers did not 
bother checking.  They simply made sure that any changes they made to 
the checksum did not turn an invalid checksum valid.

There is a practical tension, in that it would be nice to drop bad 
packets (of whatever kind) as early as possible.  So it would be nice to 
drop a packet that the far end will silently reject early.
But doing extra work to drop packets is generally frowned on.
And if the rejection is not supposed to be silent, it is much better to 
let the far end entity do the work of responding.

None of which has much to do with the topic at hand, but is related to 
the concern you expressed.

Yours,
Joel