Re: dhcpcd: re1: DHCPv6 REPLY: iana not found

[Roy Marples <roy%marples.name@localhost>]

On 05/02/2016 14:13, Rhialto wrote:
> On Fri 05 Feb 2016 at 11:38:46 +0000, Roy Marples wrote:
>>> Feb  4 20:51:48 murthe dhcpcd[10390]: re1: DHCPv6 REPLY: iana not found
>>
>> That message is from the DHCPv6 server.
> 
> So I guess I'll have to go to AVM, the maker of the Fritz!Box, probably.
> Could it refer to some option that dhcpcd is sending, or failing to
> send? I see these few references to iana in the dhcpcd source:
> 
> dhcp-common.c:vivso_find(uint32_t iana_en, const void *arg)
> dhcp-common.c:          if (opt->option == iana_en)
> dhcp-common.c:          if (opt->option == iana_en)
> 
> (I didn't look at that in context yet so this may be barking up the
> wrong tree.)

It's perfectly possible.
dhcpcd encodes a vendor class option in each DHCPv6 message.
dhcpcd has an iana enterprise number of 40712:
https://www.iana.org/assignments/enterprise-numbers/enterprise-numbers

The real question is why the Fritz! box is trying to do anything with
that field at all - it's there just for informative purposes.

> (There seem some other weird things with the Fritz!Box and IPv6 so it
> may well be its problem!)
> 
>> The kernel markes the address as deprecated when the preferred lifetime
>> expires but the valid lifetime is still ticking.
> 
> Makes sense. But is it possible to simply ignore the lifetime and set
> the addresses as fixed? Because in practice they are supposed to be,
> with this ISP.

So maybe the ISP should note they have an infinite lifetime in the RA or
DHCPv6 message?
The protocols supports this.

Roy