$v ) { if ( is_numeric( $k ) ) { unset( $k ); } } // The absolute filename itself will differ from install to install so don't leave it out if ( ( $path = realpath( $file ) ) !== false ) { $state .= $path; } else { $state .= $file; } $state .= implode( '', $stat ); } else { // The fact that the file isn't there is worth at least a // minuscule amount of entropy. $state .= '0'; } } // Try and make this a little more unstable by including the varying process // id of the php process we are running inside of if we are able to access it if ( function_exists( 'getmypid' ) ) { $state .= getmypid(); } // If available try to increase the instability of the data by throwing in // the precise amount of memory that we happen to be using at the moment. if ( function_exists( 'memory_get_usage' ) ) { $state .= memory_get_usage( true ); } // It's mostly worthless but throw the wiki's id into the data for a little more variance $state .= wfWikiID(); // If we have a secret key set then throw it into the state as well global $wgSecretKey; if ( $wgSecretKey ) { $state .= $wgSecretKey; } return $state; } /** * Randomly hash data while mixing in clock drift data for randomness * * @param string $data The data to randomly hash. * @return string The hashed bytes * @author Tim Starling */ protected function driftHash( $data ) { // Minimum number of iterations (to avoid slow operations causing the // loop to gather little entropy) $minIterations = self::MIN_ITERATIONS; // Duration of time to spend doing calculations (in seconds) $duration = ( self::MSEC_PER_BYTE / 1000 ) * $this->hashLength(); // Create a buffer to use to trigger memory operations $bufLength = 10000000; $buffer = str_repeat( ' ', $bufLength ); $bufPos = 0; // Iterate for $duration seconds or at least $minIterations number of iterations $iterations = 0; $startTime = microtime( true ); $currentTime = $startTime; while ( $iterations < $minIterations || $currentTime - $startTime < $duration ) { // Trigger some memory writing to trigger some bus activity // This may create variance in the time between iterations $bufPos = ( $bufPos + 13 ) % $bufLength; $buffer[$bufPos] = ' '; // Add the drift between this iteration and the last in as entropy $nextTime = microtime( true ); $delta = (int)( ( $nextTime - $currentTime ) * 1000000 ); $data .= $delta; // Every 100 iterations hash the data and entropy if ( $iterations % 100 === 0 ) { $data = sha1( $data ); } $currentTime = $nextTime; $iterations++; } $timeTaken = $currentTime - $startTime; $data = $this->hash( $data ); wfDebug( __METHOD__ . ": Clock drift calculation " . "(time-taken=" . ( $timeTaken * 1000 ) . "ms, " . "iterations=$iterations, " . "time-per-iteration=" . ( $timeTaken / $iterations * 1e6 ) . "us)\n" ); return $data; } /** * Return a rolling random state initially build using data from unstable sources * @return string A new weak random state */ protected function randomState() { static $state = null; if ( is_null( $state ) ) { // Initialize the state with whatever unstable data we can find // It's important that this data is hashed right afterwards to prevent // it from being leaked into the output stream $state = $this->hash( $this->initialRandomState() ); } // Generate a new random state based on the initial random state or previous // random state by combining it with clock drift $state = $this->driftHash( $state ); return $state; } /** * Decide on the best acceptable hash algorithm we have available for hash() * @throws MWException * @return string A hash algorithm */ protected function hashAlgo() { if ( !is_null( $this->algo ) ) { return $this->algo; } $algos = hash_algos(); $preference = array( 'whirlpool', 'sha256', 'sha1', 'md5' ); foreach ( $preference as $algorithm ) { if ( in_array( $algorithm, $algos ) ) { $this->algo = $algorithm; wfDebug( __METHOD__ . ": Using the {$this->algo} hash algorithm.\n" ); return $this->algo; } } // We only reach here if no acceptable hash is found in the list, this should // be a technical impossibility since most of php's hash list is fixed and // some of the ones we list are available as their own native functions // But since we already require at least 5.2 and hash() was default in // 5.1.2 we don't bother falling back to methods like sha1 and md5. throw new MWException( "Could not find an acceptable hashing function in hash_algos()" ); } /** * Return the byte-length output of the hash algorithm we are * using in self::hash and self::hmac. * * @return int Number of bytes the hash outputs */ protected function hashLength() { if ( is_null( $this->hashLength ) ) { $this->hashLength = strlen( $this->hash( '' ) ); } return $this->hashLength; } /** * Generate an acceptably unstable one-way-hash of some text * making use of the best hash algorithm that we have available. * * @param string $data * @return string A raw hash of the data */ protected function hash( $data ) { return hash( $this->hashAlgo(), $data, true ); } /** * Generate an acceptably unstable one-way-hmac of some text * making use of the best hash algorithm that we have available. * * @param string $data * @param string $key * @return string A raw hash of the data */ protected function hmac( $data, $key ) { return hash_hmac( $this->hashAlgo(), $data, $key, true ); } /** * @see self::wasStrong() */ public function realWasStrong() { if ( is_null( $this->strong ) ) { throw new MWException( __METHOD__ . ' called before generation of random data' ); } return $this->strong; } /** * @see self::generate() */ public function realGenerate( $bytes, $forceStrong = false ) { wfProfileIn( __METHOD__ ); wfDebug( __METHOD__ . ": Generating cryptographic random bytes for " . wfGetAllCallers( 5 ) . "\n" ); $bytes = floor( $bytes ); static $buffer = ''; if ( is_null( $this->strong ) ) { // Set strength to false initially until we know what source data is coming from $this->strong = true; } if ( strlen( $buffer ) < $bytes ) { // If available make use of mcrypt_create_iv URANDOM source to generate randomness // On unix-like systems this reads from /dev/urandom but does it without any buffering // and bypasses openbasedir restrictions, so it's preferable to reading directly // On Windows starting in PHP 5.3.0 Windows' native CryptGenRandom is used to generate // entropy so this is also preferable to just trying to read urandom because it may work // on Windows systems as well. if ( function_exists( 'mcrypt_create_iv' ) ) { wfProfileIn( __METHOD__ . '-mcrypt' ); $rem = $bytes - strlen( $buffer ); $iv = mcrypt_create_iv( $rem, MCRYPT_DEV_URANDOM ); if ( $iv === false ) { wfDebug( __METHOD__ . ": mcrypt_create_iv returned false.\n" ); } else { $buffer .= $iv; wfDebug( __METHOD__ . ": mcrypt_create_iv generated " . strlen( $iv ) . " bytes of randomness.\n" ); } wfProfileOut( __METHOD__ . '-mcrypt' ); } } if ( strlen( $buffer ) < $bytes ) { // If available make use of openssl's random_pseudo_bytes method to // attempt to generate randomness. However don't do this on Windows // with PHP < 5.3.4 due to a bug: // http://stackoverflow.com/questions/1940168/openssl-random-pseudo-bytes-is-slow-php // http://git.php.net/?p=php-src.git;a=commitdiff;h=cd62a70863c261b07f6dadedad9464f7e213cad5 if ( function_exists( 'openssl_random_pseudo_bytes' ) && ( !wfIsWindows() || version_compare( PHP_VERSION, '5.3.4', '>=' ) ) ) { wfProfileIn( __METHOD__ . '-openssl' ); $rem = $bytes - strlen( $buffer ); $openssl_bytes = openssl_random_pseudo_bytes( $rem, $openssl_strong ); if ( $openssl_bytes === false ) { wfDebug( __METHOD__ . ": openssl_random_pseudo_bytes returned false.\n" ); } else { $buffer .= $openssl_bytes; wfDebug( __METHOD__ . ": openssl_random_pseudo_bytes generated " . strlen( $openssl_bytes ) . " bytes of " . ( $openssl_strong ? "strong" : "weak" ) . " randomness.\n" ); } if ( strlen( $buffer ) >= $bytes ) { // openssl tells us if the random source was strong, if some of our data was generated // using it use it's say on whether the randomness is strong $this->strong = !!$openssl_strong; } wfProfileOut( __METHOD__ . '-openssl' ); } } // Only read from urandom if we can control the buffer size or were passed forceStrong if ( strlen( $buffer ) < $bytes && ( function_exists( 'stream_set_read_buffer' ) || $forceStrong ) ) { wfProfileIn( __METHOD__ . '-fopen-urandom' ); $rem = $bytes - strlen( $buffer ); if ( !function_exists( 'stream_set_read_buffer' ) && $forceStrong ) { wfDebug( __METHOD__ . ": Was forced to read from /dev/urandom " . "without control over the buffer size.\n" ); } // /dev/urandom is generally considered the best possible commonly // available random source, and is available on most *nix systems. wfSuppressWarnings(); $urandom = fopen( "/dev/urandom", "rb" ); wfRestoreWarnings(); // Attempt to read all our random data from urandom // php's fread always does buffered reads based on the stream's chunk_size // so in reality it will usually read more than the amount of data we're // asked for and not storing that risks depleting the system's random pool. // If stream_set_read_buffer is available set the chunk_size to the amount // of data we need. Otherwise read 8k, php's default chunk_size. if ( $urandom ) { // php's default chunk_size is 8k $chunk_size = 1024 * 8; if ( function_exists( 'stream_set_read_buffer' ) ) { // If possible set the chunk_size to the amount of data we need stream_set_read_buffer( $urandom, $rem ); $chunk_size = $rem; } $random_bytes = fread( $urandom, max( $chunk_size, $rem ) ); $buffer .= $random_bytes; fclose( $urandom ); wfDebug( __METHOD__ . ": /dev/urandom generated " . strlen( $random_bytes ) . " bytes of randomness.\n" ); if ( strlen( $buffer ) >= $bytes ) { // urandom is always strong, set to true if all our data was generated using it $this->strong = true; } } else { wfDebug( __METHOD__ . ": /dev/urandom could not be opened.\n" ); } wfProfileOut( __METHOD__ . '-fopen-urandom' ); } // If we cannot use or generate enough data from a secure source // use this loop to generate a good set of pseudo random data. // This works by initializing a random state using a pile of unstable data // and continually shoving it through a hash along with a variable salt. // We hash the random state with more salt to avoid the state from leaking // out and being used to predict the /randomness/ that follows. if ( strlen( $buffer ) < $bytes ) { wfDebug( __METHOD__ . ": Falling back to using a pseudo random state to generate randomness.\n" ); } while ( strlen( $buffer ) < $bytes ) { wfProfileIn( __METHOD__ . '-fallback' ); $buffer .= $this->hmac( $this->randomState(), mt_rand() ); // This code is never really cryptographically strong, if we use it // at all, then set strong to false. $this->strong = false; wfProfileOut( __METHOD__ . '-fallback' ); } // Once the buffer has been filled up with enough random data to fulfill // the request shift off enough data to handle the request and leave the // unused portion left inside the buffer for the next request for random data $generated = substr( $buffer, 0, $bytes ); $buffer = substr( $buffer, $bytes ); wfDebug( __METHOD__ . ": " . strlen( $buffer ) . " bytes of randomness leftover in the buffer.\n" ); wfProfileOut( __METHOD__ ); return $generated; } /** * @see self::generateHex() */ public function realGenerateHex( $chars, $forceStrong = false ) { // hex strings are 2x the length of raw binary so we divide the length in half // odd numbers will result in a .5 that leads the generate() being 1 character // short, so we use ceil() to ensure that we always have enough bytes $bytes = ceil( $chars / 2 ); // Generate the data and then convert it to a hex string $hex = bin2hex( $this->generate( $bytes, $forceStrong ) ); // A bit of paranoia here, the caller asked for a specific length of string // here, and it's possible (eg when given an odd number) that we may actually // have at least 1 char more than they asked for. Just in case they made this // call intending to insert it into a database that does truncation we don't // want to give them too much and end up with their database and their live // code having two different values because part of what we gave them is truncated // hence, we strip out any run of characters longer than what we were asked for. return substr( $hex, 0, $chars ); } /** Publicly exposed static methods **/ /** * Return a singleton instance of MWCryptRand * @return MWCryptRand */ protected static function singleton() { if ( is_null( self::$singleton ) ) { self::$singleton = new self; } return self::$singleton; } /** * Return a boolean indicating whether or not the source used for cryptographic * random bytes generation in the previously run generate* call * was cryptographically strong. * * @return bool Returns true if the source was strong, false if not. */ public static function wasStrong() { return self::singleton()->realWasStrong(); } /** * Generate a run of (ideally) cryptographically random data and return * it in raw binary form. * You can use MWCryptRand::wasStrong() if you wish to know if the source used * was cryptographically strong. * * @param int $bytes The number of bytes of random data to generate * @param bool $forceStrong Pass true if you want generate to prefer cryptographically * strong sources of entropy even if reading from them may steal * more entropy from the system than optimal. * @return string Raw binary random data */ public static function generate( $bytes, $forceStrong = false ) { return self::singleton()->realGenerate( $bytes, $forceStrong ); } /** * Generate a run of (ideally) cryptographically random data and return * it in hexadecimal string format. * You can use MWCryptRand::wasStrong() if you wish to know if the source used * was cryptographically strong. * * @param int $chars The number of hex chars of random data to generate * @param bool $forceStrong Pass true if you want generate to prefer cryptographically * strong sources of entropy even if reading from them may steal * more entropy from the system than optimal. * @return string Hexadecimal random data */ public static function generateHex( $chars, $forceStrong = false ) { return self::singleton()->realGenerateHex( $chars, $forceStrong ); } }