Note that this requires using dm-crypt in plain mode, which allows only one passphrase, doesn't perform any passphrase mangling or salting, and performs no passphrase checking on decryption. It requires one to re-enter exact parameters, and an improper passphrase will happily give bunk data.
Secondly, high-entropy data is evident at even a courtesy glance - normal computer and filesystem operations do not produce high-entropy data on disk, therefore a large portion of high-entropy disk data is highly suspect. The author discusses this in detail in sections 2.4 and 5.2 of https://gitlab.com/cryptsetup/cryptsetup/wikis/FrequentlyAsk...
True, but the default parameters of cryptsetup are quite secure with a high entropy password and they are seldom changed. A volume opened with the wrong password will fail to mount with no ill effects, you would just crypysetup delete and then re-create it again with the right password. I've addressed at length the other issues above, initializing an encrypted drive with random data is a widely recommended countermeasure against attacks that exploit the way the filesystem allocates sectors, VeraCrypt does it by default and many graphical setups for Linux offer the option for encrypted drives.
Secondly, high-entropy data is evident at even a courtesy glance - normal computer and filesystem operations do not produce high-entropy data on disk, therefore a large portion of high-entropy disk data is highly suspect. The author discusses this in detail in sections 2.4 and 5.2 of https://gitlab.com/cryptsetup/cryptsetup/wikis/FrequentlyAsk...