Re: Random number generation security problem

Hello all, > See > > http://www.suspekt.org/2008/08/17/mt_srand-and-not-so-random-numbers/ > > Quote: > > I just grepped through PEAR. 11 times srand() dangerously used and 4 > > times mt_srand() dangerously used. So if srand() would become an > > alias of mt_srand() the situation is obviously worse than before. > > If someone has time, please check which packages this are and how we > can fix this. I grepped through pear cvs and the following files seed the random number generator: ./Auth_SASL/SASL/DigestMD5.php: mt_srand((double)microtime()*10000000); ./Crypt_CBC/CBC.php: srand((double)microtime()*1000000); ./Cache/Cache.php: srand((double) microtime() * 1000000); ./Crypt_CHAP/CHAP.php: mt_srand(hexdec(substr(md5(microtime()), -8)) & 0x7fffffff); ./HTML_Crypt/HTML/Crypt.php: srand((float) microtime() * 10000000); ./HTML_Table_Matrix/Matrix/Filler/Random.php: srand(microtime()); ./HTTP_Upload/Upload.php: srand((double) microtime() * 1000000); ./Mail_Mime/mimePart.php: srand((double)microtime()*1000000); ./Math_BigInteger/demo/ssh_demo.php: mt_srand((float) $sec + ((float) $usec * 100000)); ./Payment_Process/Process/Dummy.php: srand(microtime()); ./XML_Transformer/Transformer/Namespace/Anchor.php: srand((double)microtime()*1000000); The short version of Stefan Esser's post is that seeding the random number generator makes your application more vulnerable since the random tokens used can be predicted better. Since php 4.2.0 we don't need to seed anymore since that's automatically done. Removing mt_srand() and srand() calls should suffice in most/all cases. Now we need bug reports for those packages. -- Regards/Mit freundlichen Grüßen Christian Weiske -= Geeking around in the name of science since 1982 =-

Attachment: [application/pgp-signature] signature.asc