[Dnsmasq-discuss] force --local/--server protocol

[Simon Kelley]

On 09/01/13 16:00, Mr Dash Four wrote:
> 
>>> Thanks, I'll seek to implement that, though I have another query:
>>> if dnsmasq receives a request from a client, which is directed to
>>> dns1 and then receives a response from that server
>>> "not-implemented" what happens next: does that error gets
>>> returned to the client,
>> Yes, this.
> Damn, this wasn't the answer I was hoping for...
> 
>> This is the behavior for SERVFAIL and REFUSED rcodes, but not 
>> "not-implemented".
>> 
>> Not that the above reflects what actually happens, which might not
>> be the same as what _should_ happen.
> Is there any chance I could solve this? If not, I am looking at
> running the "full" dns server on a different port, which is going to
> be a massive headache for me as I have to, among other things, reset
> half the configuration on that machine as otherwise there won't be a
> way in which I could direct dns requests properly using dnsmasq.
> 

If you want to patch things so that NOTIMP is treated the same way as
SERVFAIL, that's pretty simple. All the code that needs changing is in
src/forward.c, three instances found by searching for SERVFAIL. I'd
seriously consider making this change in future releases, but there is
no released code out there now for you to use -  a patch is necessary.

Cheers,

Simon.