just few thoughts:

1. Why is ck library used and not atomic(9) ?
2. yndx_vlan_set(): is that fence really required? May be store with release semantic is enough?

AFAIR those changes were done for _callout_stop_safe() aka callout_(stop|drain).
callout_reset_sbt_on() is not affected.

Please correct me if i'm wrong but there is ref leakage. In particular:

1. fix issues reported by sbahra_repnop.org here: https://reviews.freebsd.org/D1945
2. fix synchronization chain (producer -> producer -> consumer) noted by alc.
3. massive comments.

In D8637#180625, @alc wrote:

Have you looked at D1945, in particular, the most recent postings by sbahra_repnop.org? It's not clear to me that these changes will address the problem described in sbahra_repnop.org's postings. That said, your proposed changes do correct the most obvious remaining issues with the use of acquires and releases in this code.