2018-02-26, 12:11:27 -0500, David Miller wrote: > From: Sabrina Dubroca <s...@queasysnail.net> > Date: Mon, 26 Feb 2018 17:56:19 +0100 > > > 2018-02-26, 10:57:11 -0500, David Miller wrote: > >> Userland is now repsonsible for implementing correct behavior when it > >> takes over this task, and therefore the kernel has no say in the > >> matter of proper ipv6 neighbor discovery and addrconf behavior. > > > > As an aside, that's also the case whenever userland uses packet > > sockets. > > When you use packet sockets, all bets are off and it is clearly the > case that the user gets to keep the broken pieces when things go > wrong.
Ok, I see your point. > That's completely different to this case, which is a bonfide explicit > allowance for userspace to take over these fundamental protocol tasks > from the kernel. This is not letting userspace take over. On the contrary, it allows userspace to take advantage of the kernel's DAD, without suffering the delay. The alternative, without optimistic DAD, would be to completely disable DAD done by the kernel. This follows RFC 4429, which explicitly allows optimistic DAD for addresses generated by (among other mechanisms) DHCPv6. -- Sabrina