dhclient -6 handling of IPv6 Duplicate Address Detection

[Patrik Lahti]

On 27/04/10 06:08 AM, Jiri Popelka wrote:
> On 04/27/2010 12:09 AM, Patrik Lahti wrote:
>> Hi all,
>>
>> When assigning an IPv6 address a node performs Duplicate Address 
>> Detection (DAD, RFC 4862). And as expected this is triggered by the 
>> OS stack when dhclient -6 configures an address via the 
>> dhclient-script. However, if the address is found to be duplicate, 
>> i.e. when DAD fails, I don't see a Decline and an attempt to retry 
>> (RFC 3315, section 18.1.8). It seems the dhclient goes on unaware of 
>> what occurred...
>>
>> Is this a known problem? With known solutions? Should dhclient handle 
>> this, or is the intent that an outside entity (e.g. the 
>> dhclient-script) do it? (is it ok to sleep in the script while DAD is 
>> performed or will that mess up some dhclient timing)
>>
>> Regards and Thanks in advance,
>> /P
>>
>> PS. I saw 
>> https://lists.isc.org/pipermail/dhcp-workers/2010-February/000108.html but 
>> couldn't find a response...
> Hi,
>
> There's no response for my question(s), so I decided to do it my own way.
>
> I implemented start_decline6() and do_decline6() functions which are 
> called when
> dhclient-script returns value indicating that DAD failed.
> I put some sleep in the dhclient-script while DAD is performed and it
> looks like it didn't break any timing.
>
> I didn't send my patch to ISC yet, but I'm going to do it.
> I just want to know if I should send it to dhcp-bugs at isc.org or 
> dhcp-suggest at isc.org
> Thanks
>
> Jiri Popelka

Thanks Jiri, that's great, looking forward to your patch!

I also don't know where to suggest patches (I sent one yesterday to this 
list but maybe that wasn't the right place for it)

Cheers!
/P