The current hostcache hashing mechanism apparently
ends up with an very uneven distribution of the
entries into the hostcache hash array.
This was observed by adding a histogram output
Also, very rarely, the hostcache accounting seems
to go off rails. Rather than KASSERT and panicing,
provide a way to sanitize these accounting variables
to reflect the current state, when purging all
hostcache entries, as a less disruptive means to
resume proper operations.
(is there a better way to globally lock the hostcache
rather than iterating over each hashbucket mutex?)