2

I notice that encfs fails to mount a filesystem when a wrong password is entered.

However, this capability of detecting whether a password is right or wrong can allow brute-force attacks to be carried.

Is there a way to disable password checking in encfs?

I realize this will have some drawbacks such as silent failures when a genuine user enters a wrong password. However, this could be mitigated by the user by a custom script. For example, a script that mounts encfs and then does a cat mountPoint/welcomeMessage.txt. The user would know from the output whether mounting was successful or not. This I believe is much safer because brute-force attacks are not possible without prior-knowledge of the contents of welcomeMessage.txt.

If it is not possible to disable password checking in encfs, is it a vulnerability?

Improve this question
  • Why shouldn't the attacker be able to run cat mountPoint/welcomeMessage.txt automatically? – CL. Apr 30 '14 at 8:42
  • @CL he might run be able to run it, but wouldn't know what it contains. A brute force attack would require some knowledge of the contents to know whether they succeeded. Though DouglasLeeder's answer has a good counter-argument. – HRJ Apr 30 '14 at 18:03
1

Unless you are storing entirely random data inside the encfs-protected files, an attacker should be able to easily see if the files are reasonable when decrypted. So an early check doesn't weaken the defences much, and makes the product much easier to use.

Improve this answer
  • I agree that brute-force attacks are possible by checking if decrypted files look reasonable. But I guess the cost of doing that would be an order of magnitude higher than just checking the password directly. – HRJ Apr 30 '14 at 18:07

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.