Revert "client: optimize rename operation under different quota root"

Revert "client: optimize rename operation under different quota root"

This reverts commit b8954e5734b3c53f90c47fbdbbbdb8a23747ef07.

The possibility of doing a cross quota realms rename in the kernel client
has been reverted because there was a corner case where it wasn't being
handled correctly:

  mkdir files limit
  truncate files/file -s 10G
  setfattr limit -n ceph.quota.max_bytes -v 1000000
  mv files limit/

The above would succeed because the kernel client won't immediately notify
the MDSs with the new file size when doing a truncate(2), and thus the
quota realms stats won't be updated.

Although the fuse client doesn't have this issue, multi-client scenarios
may be impacted.

URL: https://tracker.ceph.com/issues/48203
Signed-off-by: Luis Henriques <lhenriques@suse.de>