On 05/07/2021 12:34, Rockwell, Dennis wrote:
> I have a situation for which extending those features would be the exact
> solution.
> 


The code is there at the bleeding edge now.

https://thekelleys.org.uk/gitweb/?p=dnsmasq.git;a=commit;h=5bcca1219af8bad328352d7a656bc9b1e9d61b92


Simon.

> Dennis
> 
> On Jul 4, 2021 5:21 PM, Simon Kelley <si...@thekelleys.org.uk> wrote:
> On 04/07/2021 21:32, Simon Kelley wrote:
>> On 30/06/2021 10:40, Kevin Darbyshire-Bryant wrote:
>>> As an ‘experiment’ I tried switching from my own local ‘adblocking’ 
>>> solution to using an upstream adblocking resolver, eg. cloudflare’s 1.1.1.2 
>>> or 1.1.1.3 service.
>>>
>>> The local adblock solution uses (multiple!) ‘—address/naughtydomain.foo/‘ 
>>> lines that cause dnsmasq to return ’NXDOMAIN’ - fair enough.
>>>
>>> Cloudflare (& others I’ve tested) return ‘0.0.0.0’ or ‘::’ instead, not 
>>> NXDOMAIN.  With rebind protection enabled (--stop-dns-rebind), even with 
>>> --rebind-localhost-ok I get log ’spam’ warning of possible rebind attacks 
>>> due to the ‘0.0.0.0’ address response.
>>>
>>> I can turn ‘0.0.0.0’ into NXDOMAIN by using --bogus-nxdomain=0.0.0.0 and 
>>> that works fine and stops the rebind warnings.  However ‘::’ still gets 
>>> through if an AAAA is specifically requested.  There is no equivalent 
>>> bogus-nxdomain for ipv6.
>>>
>>> The dnsmasq manpage (under —address) advised "Note that NULL addresses 
>>> [0.0.0.0 & ::] normally work in the same way as localhost, so beware that 
>>> clients looking up these names are likely to end up talking to themselves.” 
>>>  Ideally then 0.0.0.0 & :: would both be turned into NXDOMAIN.
>>>
>>> Should ‘0.0.0.0/32’ be excluded from the rebind checks/accepted by the 
>>> ‘—rebind-localhost-ok’ option.  It’s currently being caught by a 
>>> ‘0.0.0.0/8’ check.
>>>
>> 
>> I looked at the code that determines private addresses for --bogus-priv
>> and rebind: It's a bit unruly for IPv6, so I've rationalised things and
>> included :: and 0.0.0.0 in the --rebind-localhost-ok coverage, which at
>> least avoids the log spam.
>> 
>> 
>> I wonder if bogus-nxdomain should be extended to IPv6, or we could add
>> another option which is the equivalent of
>> 
>> bogus-nxdomain=0.0.0.0,::
>> 
>> Or both.
>> 
>> Simon.
>> 
> 
> AT the least, bogus-nxdomain should be extended to IPv6, that would
> extend --ignore-address too, for free.
> 
> 
> In progress.
> 
> Simon.
> 
> _______________________________________________
> Dnsmasq-discuss mailing list
> Dnsmasq-discuss@lists.thekelleys.org.uk
> https://urldefense.com/v3/__https://lists.thekelleys.org.uk/cgi-bin/mailman/listinfo/dnsmasq-discuss__;!!GjvTz_vk!G5VhBaG2LcDjkUOkXosk2wo1PHeuWlbg5rEhJreyBTz0RI4-Cn81DdAnrqJqq6o$
> <https://urldefense.com/v3/__https://lists.thekelleys.org.uk/cgi-bin/mailman/listinfo/dnsmasq-discuss__;!!GjvTz_vk!G5VhBaG2LcDjkUOkXosk2wo1PHeuWlbg5rEhJreyBTz0RI4-Cn81DdAnrqJqq6o$>
> 


_______________________________________________
Dnsmasq-discuss mailing list
Dnsmasq-discuss@lists.thekelleys.org.uk
https://lists.thekelleys.org.uk/cgi-bin/mailman/listinfo/dnsmasq-discuss

Reply via email to