I found a performance issue in Knot Resolver 6.5.0 when reloading an RPZ file.

Thanks for this great report :-)

We just didn't anticipate that the fallocate syscall could get so expensive.  Let's only call it when it's needed - once a cache file is newly created (or on the enlarged part of it):
https://gitlab.nic.cz/knot/knot-resolver/-/merge_requests/1924

--Vladimir