Re: [v6ops] IPREF as a transitioning tool

["Soni \"They/Them\" L." <[e-mail]>]

On 2023-11-12 08:26, Gert Doering wrote:
> Hi,
>
> On Sun, Nov 12, 2023 at 08:23:03AM -0300, Soni "They/Them" L. wrote:
> > > So, asking again - why do you think it would be beneficial if the server
> > > had to care?
> > 
> > Because DNS64 is a crime (sometimes literally, but often at least breach of
> > contract, because it breaks DNSSEC) and 464XLAT lets you avoid it. And on a
> > server, you might wanna run NixOS, which depends on github and that's still
> > IPv4-only.
>
> ok, I understand your use-case.  I'd still go for "no IPv4 anywhere near
> my servers" - and DNS64 synthesis from a local (and trusted) resolver that
> honours upstream DNSSEC is just fine in my book.
>
> Removing avoidable code and configuration complexity from servers is a plus -
> "perfection has been reached when there is nothing left to take away",
> not when every possible variant has been added.

We mean we still think moving the CLAT to libc would be a good idea. 
Because, yeah, remove that in-kernel IPv4 stack. Why should we need it?

(Moving the CLAT to libc also has an obvious solution to the whole "what 
does binding to IPv6 do with IPv4" issue around IPv4-mapped IPv6 that 
eventually led to the IPV6_V6ONLY socket option. And it's far more 
flexible about which interfaces/addresses to accept both IPv6 and IPv4 
connections on, if you ever wanted multiple IPv4s on the same server for 
some reason.)

>
> Gert Doering
>          -- NetMaster