Skip to content

Operating system password encryption

Unix/Linux Password File

Unix and its various clones have traditionally used the /etc/passwd file to store user account information, including passwords. Because the /etc/password file needs to be world-readable in order for utilities such as `ls` and `finger` to work modern Unix operating systems store the encrypted passwords in 'shadow' file named /etc/shadow.

Username The user's username
Password Older Unixes store the password crypt here, more modern ones use an 'x' character to denote that a shadow file is in use.
UID The numeric user ID of the user
GID The primary numeric group ID of the user
GECOS Field This is a text field which may contain information about the user such as name and contact details
Home directory The user's home directory
Shell The user's Unix shell
user1:x:600:600:User 1:/home/user1:/bin/bash
user2:x:601:601:User 2:/home/user2:/bin/bash
admin:x:602:602:Admin Account:/home/admin:/bin/bash
apache:x:603:603:Apache HTTP User:/var/www:/bin/bash

The password is stored as an encrypted one-way hash of the original password. When a user attempts to authenticate the password supplied is encrypted using the same algorithm and compared to the stored password crypt.

Unix Crypt

The most commonly used password encryption in Unix for many year was crypt(). The Unix crypt command can be used to generate the Unix crypt value for a given string.

jim@localhost ~
$ crypt hello

In modern computing Unix crypt is severly limited. Passwords are restricted to 8 character passwords, and any trailing character as ignored. This puts brute force attacks on Unix crypts well within the realms of possibility.

jim@localhost ~
$ crypt xx hellohel

jim@localhost ~
$ crypt xx hellohello


Unix passwords usually use what is know as a salt to help make pre-computation of password hashes more difficult. A salt is a string which is prepended to the password before it is encrypted and stored along with the password in /etc/passwd. You cannot simply pre-compute crypt() values for a list of dictionary words, you would need to pre-compute the hash for each word along with every possible salt to produce a rainbow table of Unix password hashes. The result is a number of different hashes for any given password.

If we use the Unix crypt command to encrypt a password and do not specify a salt then a random salt value is chosen.

jim@localhost ~
$ crypt hello

jim@localhost ~
$ crypt hello

The first two characters of the resulting hash are the salt and must be used when subsequently comparing a supplied password with the stored crypt.

jim@localhost ~
$ crypt v3 hello

Salts can be of any length but is typically 2 characters on Unix systems, which helps to ensure compatibility across systems.


More modern Unix systems and just about all Linux systems avoid the limitations of Unix crypt() by using a message digest algorithm such as MD5 or SHA1. Like Unix crypts the stored passwords are usually salted. These message digest algorithms not only allow for longer passwords but are also compuationally more expensive, making brute force attacks harder.


NIS (Yellow Pages) also uses the Unix password format to store user credentials. `ypcat passwd` will show the password file on a Unix server configured to user NIS.