Add tests.

In D56130#1283954, @pouria wrote:

For reviewers:
The whole revision stack will reduce PI code complexity and fix multiple issues along the way.
I also wrote additional tests in D56128 to make sure everything works as intended.

Also, there are some references in carp(4) man page and ifconfig code.

In D55558#1271208, @bms wrote:

Just my 2c: I think we should be discouraging users from implementing CARP at all.

LGTM.

In D55398#1267078, @pouria wrote:

I don't understand why we need a separate data structure (priv).

In D54443#1260647, @pouria wrote:

I see no problems with this patch alongside my changes.
We have a minor conflict, which I resolved at github.
Also, AFAICU, I don't need to use priv instead of softc for netlink because I didn't touch the data plane.
Your code will synchronize priv and softc together so there should be no problem to directly using the softc.

Should I do anything else?

Can you take a look at this patch? Maybe you can reuse some pieces, it should fix 275474 and also make access to softc in a safer way.

Do we still need this? I think scope modification ability was not used/implemented since KAME import.

LGTM.

I think you should note that this fixes some rare panics, that are reported in derived projects: panic: esp_input_cb: Unexpected address family

LGTM.

LGTM.

It seems to have affected MLDv6: https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=290407

Yes, I think they were replaced with handlers from ipfw_sopt_handler.

LGTM.