Re: [v6ops] A common problem with SLAAC in "renumbering" scenarios

[Fernando Gont <[e-mail]>]

On 4/2/19 02:44, Christian Huitema wrote:
> 
> On 2/3/2019 1:51 PM, Fernando Gont wrote:
>> On 3/2/19 17:19, Ole Troan wrote:
>>> Christian,
>>>
>>> The router cannot in its RPF check determine the reason for the wrong source address. If its spoofed or a previously valid address. Therefore code 5. 
>> Before anybody proposes reactions to code 5, please remember that GTSM
>> is *not* enforced on ICMP errors, nor dor they ned to use any specific
>> source address. i.e., they are trivial to spoof, and make an obvious
>> attack vector unless simply considered "soft errors".
> 
> Sure. But then, devices can the code 5 Dest Unreachable ICMP as a hint.
> They could for example solicit an RA, and check what prefixes are
> currently advertised. If they learn a new prefix, they might want to
> configure an additional address.

Isn't this a more complicated way of doing what we're proposing in our
document?

Thanks,
-- 
Fernando Gont
SI6 Networks
e-mail: [e-mail]
PGP Fingerprint: 6666 31C6 D484 63B2 8FB1 E3C4 AE25 0D55 1D4E 7492