Bug #64449 [NEW]: crypt doesn't fail on "$" in CRYPT_DES salt

From: brad at bradconte dot com Operating system: Linux PHP version: 5.4.13 Package: *Encryption and hash functions Bug Type: Bug Bug description:crypt doesn't fail on "$" in CRYPT_DES salt Description: ------------ --- From manual page: http://www.php.net/function.crypt#refsect1-function.crypt-description --- The PHP documentation states that crypt() will "return a failure" if the salt contains an invalid character. For CRYPT_DES, the character "$" (ie, dollar-sign) is not listed as a valid character. However, an input salt with the "$" character still results in valid CRYPT_DES hash output instead of a failure. The documentation and the function's behavior are inconsistent. Note that the glibc crypt() function fails when "$" is provided, so the documentation is probably correct and it is probably the implementation that is at fault. Note that allowing "$" for CRYPT_DES has an interesting security implication: Since other hash algorithm salt formats begin with "$", a mistake on the programmer's part in selecting the desired hash algorithm can cause a silent downgrade to the undesirable hash algorithm CRYPT_DES. Examples: * Valid salt strings for unsupported salt algorithms may get downgraded to CRYPT_DES. For example, in PHP 5.2 (pre-CRYPT_BLOW_FISH support) the CRYPT_BLOWFISH salt string '$2a$abcdefghijklmnopqrstuv' will be silently downgraded to CRYPT_DES. * An invalid salt string for another algorithm may get downgraded to CRYPT_DES. For example the attempted CRYPT_SHA256 string '$6 $1234567890123456789012345678901234567890123' (notice the space) will get downgraded to CRYPT_DES. This downgrade will only be caught by the programmer comparing output format and length to the expected format. Since the documentation says CRYPT_DES will return a failure on a salt with a "$", the programmer could be led astray and not thoroughly examine the result. If crypt() failed instead, the programmer's mistake would be much more obvious. Test script: --------------- /* PHP 5.2 */ if (crypt('password', '$2') == crypt('password', '$2a$11$07fc1ae880d703a7a83424) echo "TRIPLE_DES selection"; /* PHP 5.4 */ crypt('password', '$2'); /* Outputs same as above. */ Expected result: ---------------- I expect crypt('password', '$2') to return a failure. Actual result: -------------- crypt('password', '$2') returns a valid hash. -- Edit bug report at https://bugs.php.net/bug.php?id=64449&edit=1 -- Try a snapshot (PHP 5.4): https://bugs.php.net/fix.php?id=64449&r=trysnapshot54 Try a snapshot (PHP 5.3): https://bugs.php.net/fix.php?id=64449&r=trysnapshot53 Try a snapshot (trunk): https://bugs.php.net/fix.php?id=64449&r=trysnapshottrunk Fixed in SVN: https://bugs.php.net/fix.php?id=64449&r=fixed Fixed in release: https://bugs.php.net/fix.php?id=64449&r=alreadyfixed Need backtrace: https://bugs.php.net/fix.php?id=64449&r=needtrace Need Reproduce Script: https://bugs.php.net/fix.php?id=64449&r=needscript Try newer version: https://bugs.php.net/fix.php?id=64449&r=oldversion Not developer issue: https://bugs.php.net/fix.php?id=64449&r=support Expected behavior: https://bugs.php.net/fix.php?id=64449&r=notwrong Not enough info: https://bugs.php.net/fix.php?id=64449&r=notenoughinfo Submitted twice: https://bugs.php.net/fix.php?id=64449&r=submittedtwice register_globals: https://bugs.php.net/fix.php?id=64449&r=globals PHP 4 support discontinued: https://bugs.php.net/fix.php?id=64449&r=php4 Daylight Savings: https://bugs.php.net/fix.php?id=64449&r=dst IIS Stability: https://bugs.php.net/fix.php?id=64449&r=isapi Install GNU Sed: https://bugs.php.net/fix.php?id=64449&r=gnused Floating point limitations: https://bugs.php.net/fix.php?id=64449&r=float No Zend Extensions: https://bugs.php.net/fix.php?id=64449&r=nozend MySQL Configuration Error: https://bugs.php.net/fix.php?id=64449&r=mysqlcfg