Bug#765594: systemd: Attempting to hibernate permanently breaks encrypted swap partition

[Maximilian Gaukler]

In one instance, which I cant exactly reproduce, systemd killed my swap 
partition's LUKS header even without hibernating.
I was using a keyscript which failed (and there are some huge bugs 
regarding keyscripts and systemd), rebooted and somehow it had killed 
the header.

Is there a documentation what the "swap" keyword in crypttab stands for 
from the systemd point of view?

cryptdisks_start seems to see it differently, as man crypttab states 
only: "run mkswap on the created device"