Duperemove is a tool for finding duplicated extents and submitting them for
 deduplication. When given a list of files it will hash their contents on a
 block by block basis and compare those hashes to each other, finding and
 categorizing extents that match each other.
 .
 On BTRFS and, experimentally, XFS, it can then reflink such extents in a
 race-free way. Unlike hardlink-based solutions, affected files appear
 independent in any way other than reduced disk space used.