> On 27 Jun 2024, at 06:51, Ole Trøan <otroan=40employees....@dmarc.ietf.org> > wrote: > > > >> On 26 Jun 2024, at 22:47, Brian Candler <br...@nsrc.org> wrote: >> >> On 26/06/2024 21:26, Ole Trøan wrote: >>> I would still like the option of having an IPv6 only host. (Which 464XLAT >>> doesn’t give). >> That depends what you mean by "IPv6 only host". Do you mean it only has an >> IPv6 address on its external interface? Or do you want to disable the IPv4 >> stack entirely in the kernel? > > Yes, no ipv4 stack at all. >> >> >> The CLAT-in-libc approach gives an interesting middle ground, where the >> application can still ask to open an AF_INET socket, but this gets >> translated to AF_INET6 before it hits the kernel. Of course, it would be >> good to see running code first. > Agree, that would be quite neat.
If there is no IPv4 stack then a CLAT is not needed. Only address mapping is needed. socket(PF_INET,…) should fail with EAFNOSUPPORT. CLAT in libc is only needed to support IPv4 only applications and for there to be no IPv4 stack these do not exist by definition. Dual stack applications don’t need CLAT. Add a flag to getaddrinfo to say to use PREF64 when looking for addresses and don’t return IPv4 addresses if a PREF64 address is present and AI_ADDRCONFIG is set. One could also fallback to looking at ipv4-only.arpa if that flag is set. Run address literals though getaddrinfo. > Cheers > Ole > _______________________________________________ > v6ops mailing list -- v6...@ietf.org > To unsubscribe send an email to v6ops-le...@ietf.org -- Mark Andrews, ISC 1 Seymour St., Dundas Valley, NSW 2117, Australia PHONE: +61 2 9871 4742 INTERNET: ma...@isc.org _______________________________________________ DNSOP mailing list -- dnsop@ietf.org To unsubscribe send an email to dnsop-le...@ietf.org
