[RESOLVED] Mcrypt Problems

King_Kovifor

Hey,

I'm creating a vBulletin Modification (update actually) and as of right now I am storing passwords as plain text. I was told about mcrypt, and decided to use it. But I'm having a problem. Since I need an unencrypted password in several cases, I decrypt the password to edit and authenticate. Now, I have the following code to encrypt:

$size = mcrypt_get_iv_size(MCRYPT_3DES, MCRYPT_MODE_CBC);
$iv = mcrypt_create_iv($size, MCRYPT_DEV_RANDOM);
$vbulletin->GPC['twitter_password'] = mcrypt_encrypt(MCRYPT_3DES, COOKIE_SALT, $vbulletin->GPC['twitter_password'], MCRYPT_MODE_CBC, $iv);

Where $vbulletin->GPC['twitter_password'] starts out unencrypted, but that variable is overwritten using the encrypted value. Now, this works just fine, but when I use this for decryption it comes out exactly the same as the encrypted value:

$size = mcrypt_get_iv_size(MCRYPT_3DES, MCRYPT_MODE_CBC);
$iv = mcrypt_create_iv($size, MCRYPT_DEV_RANDOM);
$bbuserinfo['twitter_password'] = mcrypt_decrypt(MCRYPT_3DES, COOKIE_SALT, $vbulletin->userinfo['twitter_password'], MCRYPT_MODE_CBC, $iv);

Where values are the same. I just happen to assign it to a new variable. I cannot figure this out (and doing a quick search of this forum hasn't helped) and since I have fixed several crucial bugs and need to get this out ASAP, it would be much appreciated!

King_Kovifor

Do I need to store the IV I created? Doesn't that defeat the purpose?

King_Kovifor

It seems I need to store and use the same IV. Now that I'm doing that, I have 1 more problem that I'll to solve that one on my own.

Weedpacket

King Kovifor wrote:
Do I need to store the IV I created? Doesn't that defeat the purpose?

No; the IV isn't that much of a secret; it can be sent along with the ciphertext.

It seems I need to store and use the same IV.

The idea is to use a different IV every time. If necessary it can be stored alongside the corresponding ciphertext.

I decrypt the password to edit and authenticate

Which defeats the whole purpose of encrypting the password: to decrypt the password the decryption key needs to be available and anyone who gets their hands on the database will get their hands on the key as well. ("The front door key is hidden under the doormat.")

The idea is to store a one-way hash of the password: a password is supplied, hashed (with the appropriate salt) and compared against the stored hash. If you still need the plaintext password after that, don't throw away the one you were supplied with in the first place.

King_Kovifor

Weedpacket;10903346 wrote:
No; the IV isn't that much of a secret; it can be sent along with the ciphertext.

The idea is to use a different IV every time. If necessary it can be stored alongside the corresponding ciphertext.

Which defeats the whole purpose of encrypting the password: to decrypt the password the decryption key needs to be available and anyone who gets their hands on the database will get their hands on the key as well. ("The front door key is hidden under the doormat.")

The idea is to store a one-way hash of the password: a password is supplied, hashed (with the appropriate salt) and compared against the stored hash. If you still need the plaintext password after that, don't throw away the one you were supplied with in the first place.

I generate a new IV every time it is updated. The key is unique to each and every installation of the mod, and is not located within the database. They must know where to find the key on the file system. And it's for a twitter integration. I blame twitter's API for that. 😉

CaptSaltyJack

Weedpacket is totally right on. Hashing is the way to go for password storage. Don't use SHA1 or MD5, they're on their way out. Whirlpool is pretty rock solid (I've looked into this in depth).

$encryptedPass = hash('whirlpool', $salt . $pass);

Weedpacket

The U.S. Government uses something from the SHA2 family (SHA-256 etc., but not SHA1) for nonclassified material; at least until the new hash algorithm competition throws up a winner.

CaptSaltyJack

Article comparing hash algorithms:
http://www.kellermansoftware.com/t-articlestrongesthash.aspx