singulink / singulink.cryptography.passwordhasher.argon2 Goto Github PK

Hi!

First of all, let me say that I found your PasswordHasher library today and it's just perfect for my use case.
I'm rather surprised that this library is not very popular.
I love the rehash and update functionality, which seems like a textbook implementation of the guidelines on the OWASP Password Storage Cheat Sheet.
Also, the documentation provided in the README is understandable, straight to the point, and it explains all available features.

However, I have found a very serious bug in your Argon2 implementation, which unfortunately is not detected by a unit test.

In short, the Argon2HashAlgorithm.Hash(...) method always returns a byte array containing only zeroes.
This effectively makes the PasswordHasher believe that all passwords are the same.
The cause for this is that the SecureArray<byte>.Buffer is returned directly by the method, but at the same time the SecureArray<byte> is disposed through a using statement, which in turn zeroes out the buffer before it is actually returned to the caller.

I would suggest two changes:

1. return a copy of the SecureArray<byte>.Buffer contents
2. add a unit test which ensures that the PasswordHasher.Verify(...) method returns false for the wrong password

If you want I could provide a pull request for that.