[Dnsmasq-discuss] Router Advertisement: Prefix-Specific Options?

[Simon Kelley]

I just committed

http://thekelleys.org.uk/gitweb/?p=dnsmasq.git;a=commit;h=c488b68e75ee5304007eef37203c4fc10193d191

which suppresses construction of a dhcp-range if there's an explict
dhcp-range already.

Testing would be very useful.


Cheers,

Simon.


On 19/04/18 03:38, Luis Marsano wrote:
> Simon Kelley <simon at thekelleys.org.uk> wrote:
> 
>> Apologies, there's no way to sue the solution I suggested in current
>> dnsmasq, it was a possible future enhancement.
>>
>> On 17/04/18 00:16, Luis Marsano wrote:
>>> Yes, I’d expect that to work, though I’d only know after testing.
>>> Is there a way to do that?
>>> I was using the constructor option to handle dynamic prefixes, which
>>> also need to be advertised.
>>> The option
>>>> dhcp-range=set:lan,::,constructor:br-lan,ra-stateless,ra-names,12h
>>> advertises dynamic prefixes **and** static prefixes: whatever is bound
>>> to the interface, which seems an all or none proposition to me.
>>> If I could exclude the static prefix from the constructed
>>> advertisements, that would work.
>>
>> How would you tell which prefixes were static, and which dynamic?
> 
> I'd know from having to explicitly setup the static prefixes myself rather than getting them automatically.
> IPv6 prefixes from the 6in4 tunnel broker are static: 6in4 is a static mechanism and the tunnel broker gave me static addresses and prefixes to set up.
> The other global IPv6 addresses and prefixes are potentially dynamic, and automatically appear by enabling IPv6 on the WAN interface: DHCPv6-PD gets a prefix from my ISP, and the openWRT/LEDE router automatically assigns a subnet from that to its LAN interfaces.
> Though I could write out the current prefixes, I have no assurance they'll remain the same later.
> 
>>> If I could simply pass an additional option for the static prefix, that
>>> would also work.
>>> Is there a way to do either?
>>> I’m sorry if I missed it in the manual.
>>
>> You didn't. I don't think there's any way to do what you want in the
>> current release of dnsmasq. We have to invent a new function to do it.
> 
> With the approach you postulated, I might try something like
> 
> dhcp-range=set:lan,::,constructor:br-lan,exclude:2001:db8::,ra-stateless,ra-names,12h
> dhcp-range=set:lan,2001:db8::,ra-stateless,ra-names,deprecated
> 
> I was imagining an approach like
> 
> dhcp-range=set:lan,::,constructor:br-lan,ra-stateless,ra-names,12h
> dhcp-range=set:lan,2001:db8::,ra-stateless,ra-names,deprecated
> 
> that merges options and overrides according to some precedence like order or specificity.
> I'm not sure about the best design for a new feature: according to https://tools.ietf.org/html/rfc4861#section-6.2.3 router advertisements allow each prefix to have its own options, so either design might suffice.
> I was also considering an alternative based on the tag system, though I'm not sure it's meant for that.
> 
> Your project is great: whatever solution you think is best is probably a good one.
> 
>> Cheers,
>>
>> Simon.
> 
> Thanks
> 
>>> Thanks,
>>> Luis
>>> *From: *Simon Kelley <mailto:simon at thekelleys.org.uk>
>>> *Sent: *Monday, April 16, 2018 6:37 PM
>>> *To: *dnsmasq-discuss at lists.thekelleys.org.uk
>>> <mailto:dnsmasq-discuss at lists.thekelleys.org.uk>
>>> *Subject: *Re: [Dnsmasq-discuss] Router Advertisement: Prefix-Specific
>>> Options?
>>>  
>>> Would this be solved by not constructing a prefix advertisement for
>>> 2001:db8:: when it's already explicitly configured?
>>>  
>>>  
>>> Cheers,
>>>  
>>> Simon.
> 
> _______________________________________________
> Dnsmasq-discuss mailing list
> Dnsmasq-discuss at lists.thekelleys.org.uk
> http://lists.thekelleys.org.uk/mailman/listinfo/dnsmasq-discuss
>