ipv6: update @ipv6_stable_privacy_dad scenarios

Recent CI failures showed problems with these scenarios, based on those:

• code comment turned into a clearer Commentary step
• the other endpoint of the veth pair only gets the duplicate address after the con_ipv6 gets down
• race conditions should be removed
• addresses are read in more robust way
• the timeouts and ordering of the checks is adjusted

---

@runtests:pass,ipv6_stable_privacy_dad

@os:rhel10.0
@os:c9s
@os:rhel9.2