there is one line that makes me itch a bit ....
after we mount $rootmnt from $writable_mnt we move $writable_mnt underneath $rootmnt, so we are effectively pulling out the floor under ourselves.
while i'm not sure what the implications of this actually are (since run-init makes everything above $rootmnt invisible anyway once we switch to the real root) I wonder if we shouldnt use --bind instead of --move (thats line 197 below)
mount --move "$writable_mnt" "${rootmnt}/writable"
there is one line that makes me itch a bit ....
after we mount $rootmnt from $writable_mnt we move $writable_mnt underneath $rootmnt, so we are effectively pulling out the floor under ourselves.
while i'm not sure what the implications of this actually are (since run-init makes everything above $rootmnt invisible anyway once we switch to the real root) I wonder if we shouldnt use --bind instead of --move (thats line 197 below)
mount --move "$writable_mnt" "${rootmnt} /writable"