I recognised that squashfs is not really supporting
Of course all other hard links of a file will be
detected as a duplicate of the first and so the file is
just saved once.
But this duplicates are saves as seperate files with a
This has two main disadvantages:
1) Time: If you have many hard links to big files it
needs a lot of time to scan all. I don't think
mksquashfs is looking yet to the original inodes to
detect hard links more easily and faster.
This should be changed/added.
2) Restoration: If you make a backup with mksquashfs
you would like to be able to restore the exact original
state (i.e. if you copy the files from the mounted
squashfs to the original place). The current squashfs
will brake all hard links into single files. This can
be very bad. You can run out of disk space while a
So, mksquashfs should store hard links as hard links
(if all hard links are supposed to get saved into the