includes/utils/UIDGenerator.php

   1 <?php
   2 /**
   3  * This file deals with UID generation.
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  * @author Aaron Schulz
  22  */
  23
  24 /**
  25  * Class for getting statistically unique IDs
  26  *
  27  * @since 1.21
  28  */
  29 class UIDGenerator {
  30         /** @var UIDGenerator */
  31         protected static $instance = null;
  32
  33         protected $nodeId32; // string; node ID in binary (32 bits)
  34         protected $nodeId48; // string; node ID in binary (48 bits)
  35
  36         protected $lockFile88; // string; local file path
  37         protected $lockFile128; // string; local file path
  38
  39         /** @var Array */
  40         protected $fileHandles = array(); // cache file handles
  41
  42         const QUICK_RAND = 1; // get randomness from fast and insecure sources
  43         const QUICK_VOLATILE = 2; // use an APC like in-memory counter if available
  44
  45         protected function __construct() {
  46                 $idFile = wfTempDir() . '/mw-' . __CLASS__ . '-UID-nodeid';
  47                 $nodeId = is_file( $idFile ) ? file_get_contents( $idFile ) : '';
  48                 // Try to get some ID that uniquely identifies this machine (RFC 4122)...
  49                 if ( !preg_match( '/^[0-9a-f]{12}$/i', $nodeId ) ) {
  50                         wfSuppressWarnings();
  51                         if ( wfIsWindows() ) {
  52                                 // http://technet.microsoft.com/en-us/library/bb490913.aspx
  53                                 $csv = trim( wfShellExec( 'getmac /NH /FO CSV' ) );
  54                                 $line = substr( $csv, 0, strcspn( $csv, "\n" ) );
  55                                 $info = str_getcsv( $line );
  56                                 $nodeId = isset( $info[0] ) ? str_replace( '-', '', $info[0] ) : '';
  57                         } elseif ( is_executable( '/sbin/ifconfig' ) ) { // Linux/BSD/Solaris/OS X
  58                                 // See http://linux.die.net/man/8/ifconfig
  59                                 $m = array();
  60                                 preg_match( '/\s([0-9a-f]{2}(:[0-9a-f]{2}){5})\s/',
  61                                         wfShellExec( '/sbin/ifconfig -a' ), $m );
  62                                 $nodeId = isset( $m[1] ) ? str_replace( ':', '', $m[1] ) : '';
  63                         }
  64                         wfRestoreWarnings();
  65                         if ( !preg_match( '/^[0-9a-f]{12}$/i', $nodeId ) ) {
  66                                 $nodeId = MWCryptRand::generateHex( 12, true );
  67                                 $nodeId[1] = dechex( hexdec( $nodeId[1] ) | 0x1 ); // set multicast bit
  68                         }
  69                         file_put_contents( $idFile, $nodeId ); // cache
  70                 }
  71                 $this->nodeId32 = wfBaseConvert( substr( sha1( $nodeId ), 0, 8 ), 16, 2, 32 );
  72                 $this->nodeId48 = wfBaseConvert( $nodeId, 16, 2, 48 );
  73                 // If different processes run as different users, they may have different temp dirs.
  74                 // This is dealt with by initializing the clock sequence number and counters randomly.
  75                 $this->lockFile88 = wfTempDir() . '/mw-' . __CLASS__ . '-UID-88';
  76                 $this->lockFile128 = wfTempDir() . '/mw-' . __CLASS__ . '-UID-128';
  77         }
  78
  79         /**
  80          * @return UIDGenerator
  81          */
  82         protected static function singleton() {
  83                 if ( self::$instance === null ) {
  84                         self::$instance = new self();
  85                 }
  86
  87                 return self::$instance;
  88         }
  89
  90         /**
  91          * Get a statistically unique 88-bit unsigned integer ID string.
  92          * The bits of the UID are prefixed with the time (down to the millisecond).
  93          *
  94          * These IDs are suitable as values for the shard key of distributed data.
  95          * If a column uses these as values, it should be declared UNIQUE to handle collisions.
  96          * New rows almost always have higher UIDs, which makes B-TREE updates on INSERT fast.
  97          * They can also be stored "DECIMAL(27) UNSIGNED" or BINARY(11) in MySQL.
  98          *
  99          * UID generation is serialized on each server (as the node ID is for the whole machine).
 100          *
 101          * @param $base integer Specifies a base other than 10
 102          * @return string Number
 103          * @throws MWException
 104          */
 105         public static function newTimestampedUID88( $base = 10 ) {
 106                 if ( !is_integer( $base ) || $base > 36 || $base < 2 ) {
 107                         throw new MWException( "Base must an integer be between 2 and 36" );
 108                 }
 109                 $gen = self::singleton();
 110                 $time = $gen->getTimestampAndDelay( 'lockFile88', 1, 1024 );
 111
 112                 return wfBaseConvert( $gen->getTimestampedID88( $time ), 2, $base );
 113         }
 114
 115         /**
 116          * @param array $time (UIDGenerator::millitime(), clock sequence)
 117          * @return string 88 bits
 118          */
 119         protected function getTimestampedID88( array $info ) {
 120                 list( $time, $counter ) = $info;
 121                 // Take the 46 MSBs of "milliseconds since epoch"
 122                 $id_bin = $this->millisecondsSinceEpochBinary( $time );
 123                 // Add a 10 bit counter resulting in 56 bits total
 124                 $id_bin .= str_pad( decbin( $counter ), 10, '0', STR_PAD_LEFT );
 125                 // Add the 32 bit node ID resulting in 88 bits total
 126                 $id_bin .= $this->nodeId32;
 127                 // Convert to a 1-27 digit integer string
 128                 if ( strlen( $id_bin ) !== 88 ) {
 129                         throw new MWException( "Detected overflow for millisecond timestamp." );
 130                 }
 131
 132                 return $id_bin;
 133         }
 134
 135         /**
 136          * Get a statistically unique 128-bit unsigned integer ID string.
 137          * The bits of the UID are prefixed with the time (down to the millisecond).
 138          *
 139          * These IDs are suitable as globally unique IDs, without any enforced uniqueness.
 140          * New rows almost always have higher UIDs, which makes B-TREE updates on INSERT fast.
 141          * They can also be stored as "DECIMAL(39) UNSIGNED" or BINARY(16) in MySQL.
 142          *
 143          * UID generation is serialized on each server (as the node ID is for the whole machine).
 144          *
 145          * @param $base integer Specifies a base other than 10
 146          * @return string Number
 147          * @throws MWException
 148          */
 149         public static function newTimestampedUID128( $base = 10 ) {
 150                 if ( !is_integer( $base ) || $base > 36 || $base < 2 ) {
 151                         throw new MWException( "Base must be an integer between 2 and 36" );
 152                 }
 153                 $gen = self::singleton();
 154                 $time = $gen->getTimestampAndDelay( 'lockFile128', 16384, 1048576 );
 155
 156                 return wfBaseConvert( $gen->getTimestampedID128( $time ), 2, $base );
 157         }
 158
 159         /**
 160          * @param array $info (UIDGenerator::millitime(), counter, clock sequence)
 161          * @return string 128 bits
 162          */
 163         protected function getTimestampedID128( array $info ) {
 164                 list( $time, $counter, $clkSeq ) = $info;
 165                 // Take the 46 MSBs of "milliseconds since epoch"
 166                 $id_bin = $this->millisecondsSinceEpochBinary( $time );
 167                 // Add a 20 bit counter resulting in 66 bits total
 168                 $id_bin .= str_pad( decbin( $counter ), 20, '0', STR_PAD_LEFT );
 169                 // Add a 14 bit clock sequence number resulting in 80 bits total
 170                 $id_bin .= str_pad( decbin( $clkSeq ), 14, '0', STR_PAD_LEFT );
 171                 // Add the 48 bit node ID resulting in 128 bits total
 172                 $id_bin .= $this->nodeId48;
 173                 // Convert to a 1-39 digit integer string
 174                 if ( strlen( $id_bin ) !== 128 ) {
 175                         throw new MWException( "Detected overflow for millisecond timestamp." );
 176                 }
 177
 178                 return $id_bin;
 179         }
 180
 181         /**
 182          * Return an RFC4122 compliant v4 UUID
 183          *
 184          * @param $flags integer Bitfield (supports UIDGenerator::QUICK_RAND)
 185          * @return string
 186          * @throws MWException
 187          */
 188         public static function newUUIDv4( $flags = 0 ) {
 189                 $hex = ( $flags & self::QUICK_RAND )
 190                         ? wfRandomString( 31 )
 191                         : MWCryptRand::generateHex( 31 );
 192
 193                 return sprintf( '%s-%s-%s-%s-%s',
 194                         // "time_low" (32 bits)
 195                         substr( $hex, 0, 8 ),
 196                         // "time_mid" (16 bits)
 197                         substr( $hex, 8, 4 ),
 198                         // "time_hi_and_version" (16 bits)
 199                         '4' . substr( $hex, 12, 3 ),
 200                         // "clk_seq_hi_res (8 bits, variant is binary 10x) and "clk_seq_low" (8 bits)
 201                         dechex( 0x8 | ( hexdec( $hex[15] ) & 0x3 ) ) . $hex[16] . substr( $hex, 17, 2 ),
 202                         // "node" (48 bits)
 203                         substr( $hex, 19, 12 )
 204                 );
 205         }
 206
 207         /**
 208          * Return an RFC4122 compliant v4 UUID
 209          *
 210          * @param $flags integer Bitfield (supports UIDGenerator::QUICK_RAND)
 211          * @return string 32 hex characters with no hyphens
 212          * @throws MWException
 213          */
 214         public static function newRawUUIDv4( $flags = 0 ) {
 215                 return str_replace( '-', '', self::newUUIDv4( $flags ) );
 216         }
 217
 218         /**
 219          * Return an ID that is sequential *only* for this node and bucket
 220          *
 221          * These IDs are suitable for per-host sequence numbers, e.g. for some packet protocols.
 222          * If UIDGenerator::QUICK_VOLATILE is used the counter might reset on server restart.
 223          *
 224          * @param string $bucket Arbitrary bucket name (should be ASCII)
 225          * @param integer $bits Bit size (<=48) of resulting numbers before wrap-around
 226          * @param integer $flags (supports UIDGenerator::QUICK_VOLATILE)
 227          * @return float Integer value as float
 228          * @since 1.23
 229          */
 230         public static function newSequentialPerNodeID( $bucket, $bits = 48, $flags = 0 ) {
 231                 return current( self::newSequentialPerNodeIDs( $bucket, $bits, 1, $flags ) );
 232         }
 233
 234         /**
 235          * Return IDs that are sequential *only* for this node and bucket
 236          *
 237          * @see UIDGenerator::newSequentialPerNodeID()
 238          * @param string $bucket Arbitrary bucket name (should be ASCII)
 239          * @param integer $bits Bit size (16 to 48) of resulting numbers before wrap-around
 240          * @param integer $count Number of IDs to return (1 to 10000)
 241          * @param integer $flags (supports UIDGenerator::QUICK_VOLATILE)
 242          * @return array Ordered list of float integer values
 243          * @since 1.23
 244          */
 245         public static function newSequentialPerNodeIDs( $bucket, $bits, $count, $flags = 0 ) {
 246                 if ( $count <= 0 ) {
 247                         return array(); // nothing to do
 248                 } elseif ( $count > 10000 ) {
 249                         throw new MWException( "Number of requested IDs ($count) is too high." );
 250                 } elseif ( $bits < 16 || $bits > 48 ) {
 251                         throw new MWException( "Requested bit size ($bits) is out of range." );
 252                 }
 253
 254                 $counter = null; // post-increment persistent counter value
 255
 256                 // Use APC/eAccelerator/xcache if requested, available, and not in CLI mode;
 257                 // Counter values would not survive accross script instances in CLI mode.
 258                 $cache = null;
 259                 if ( ( $flags & self::QUICK_VOLATILE ) && PHP_SAPI !== 'cli' ) {
 260                         try {
 261                                 $cache = ObjectCache::newAccelerator( array() );
 262                         } catch ( MWException $e ) {} // not supported
 263                 }
 264                 if ( $cache ) {
 265                         $counter = $cache->incr( $bucket, $count );
 266                         if ( $counter === false ) {
 267                                 if ( !$cache->add( $bucket, $count ) ) {
 268                                         throw new MWException( 'Unable to set value to ' . get_class( $cache ) );
 269                                 }
 270                                 $counter = $count;
 271                         }
 272                 }
 273
 274                 // Note: use of fmod() avoids "division by zero" on 32 bit machines
 275                 if ( $counter === null ) {
 276                         $path = wfTempDir() . '/mw-' . __CLASS__ . '-' . rawurlencode( $bucket ) . '-48';
 277                         $handle = fopen( $path, 'cb+' );
 278                         // Acquire the UID lock file
 279                         if ( $handle === false ) {
 280                                 throw new MWException( "Could not open '{$path}'." );
 281                         } elseif ( !flock( $handle, LOCK_EX ) ) {
 282                                 fclose( $handle );
 283                                 throw new MWException( "Could not acquire '{$path}'." );
 284                         }
 285                         // Fetch the counter value and increment it...
 286                         rewind( $handle );
 287                         $counter = floor( trim( fgets( $handle ) ) ) + $count; // fetch as float
 288                         // Write back the new counter value
 289                         ftruncate( $handle, 0 );
 290                         rewind( $handle );
 291                         fwrite( $handle, fmod( $counter, pow( 2, 48 ) ) ); // warp-around as needed
 292                         fflush( $handle );
 293                         // Release the UID lock file
 294                         flock( $handle, LOCK_UN );
 295                         fclose( $handle );
 296                 }
 297                 $ids = array();
 298                 $divisor = pow( 2, $bits );
 299                 $currentId = floor( $counter - $count ); // pre-increment counter value
 300                 for ( $i = 0; $i < $count; ++$i ) {
 301                         $ids[] = fmod( ++$currentId, $divisor );
 302                 }
 303
 304                 return $ids;
 305         }
 306
 307         /**
 308          * Get a (time,counter,clock sequence) where (time,counter) is higher
 309          * than any previous (time,counter) value for the given clock sequence.
 310          * This is useful for making UIDs sequential on a per-node bases.
 311          *
 312          * @param string $lockFile Name of a local lock file
 313          * @param $clockSeqSize integer The number of possible clock sequence values
 314          * @param $counterSize integer The number of possible counter values
 315          * @return Array (result of UIDGenerator::millitime(), counter, clock sequence)
 316          * @throws MWException
 317          */
 318         protected function getTimestampAndDelay( $lockFile, $clockSeqSize, $counterSize ) {
 319                 // Get the UID lock file handle
 320                 if ( isset( $this->fileHandles[$lockFile] ) ) {
 321                         $handle = $this->fileHandles[$lockFile];
 322                 } else {
 323                         $handle = fopen( $this->$lockFile, 'cb+' );
 324                         $this->fileHandles[$lockFile] = $handle ?: null; // cache
 325                 }
 326                 // Acquire the UID lock file
 327                 if ( $handle === false ) {
 328                         throw new MWException( "Could not open '{$this->$lockFile}'." );
 329                 } elseif ( !flock( $handle, LOCK_EX ) ) {
 330                         fclose( $handle );
 331                         throw new MWException( "Could not acquire '{$this->$lockFile}'." );
 332                 }
 333                 // Get the current timestamp, clock sequence number, last time, and counter
 334                 rewind( $handle );
 335                 $data = explode( ' ', fgets( $handle ) ); // "<clk seq> <sec> <msec> <counter> <offset>"
 336                 $clockChanged = false; // clock set back significantly?
 337                 if ( count( $data ) == 5 ) { // last UID info already initialized
 338                         $clkSeq = (int)$data[0] % $clockSeqSize;
 339                         $prevTime = array( (int)$data[1], (int)$data[2] );
 340                         $offset = (int)$data[4] % $counterSize; // random counter offset
 341                         $counter = 0; // counter for UIDs with the same timestamp
 342                         // Delay until the clock reaches the time of the last ID.
 343                         // This detects any microtime() drift among processes.
 344                         $time = $this->timeWaitUntil( $prevTime );
 345                         if ( !$time ) { // too long to delay?
 346                                 $clockChanged = true; // bump clock sequence number
 347                                 $time = self::millitime();
 348                         } elseif ( $time == $prevTime ) {
 349                                 // Bump the counter if there are timestamp collisions
 350                                 $counter = (int)$data[3] % $counterSize;
 351                                 if ( ++$counter >= $counterSize ) { // sanity (starts at 0)
 352                                         flock( $handle, LOCK_UN ); // abort
 353                                         throw new MWException( "Counter overflow for timestamp value." );
 354                                 }
 355                         }
 356                 } else { // last UID info not initialized
 357                         $clkSeq = mt_rand( 0, $clockSeqSize - 1 );
 358                         $counter = 0;
 359                         $offset = mt_rand( 0, $counterSize - 1 );
 360                         $time = self::millitime();
 361                 }
 362                 // microtime() and gettimeofday() can drift from time() at least on Windows.
 363                 // The drift is immediate for processes running while the system clock changes.
 364                 // time() does not have this problem. See https://bugs.php.net/bug.php?id=42659.
 365                 if ( abs( time() - $time[0] ) >= 2 ) {
 366                         // We don't want processes using too high or low timestamps to avoid duplicate
 367                         // UIDs and clock sequence number churn. This process should just be restarted.
 368                         flock( $handle, LOCK_UN ); // abort
 369                         throw new MWException( "Process clock is outdated or drifted." );
 370                 }
 371                 // If microtime() is synced and a clock change was detected, then the clock went back
 372                 if ( $clockChanged ) {
 373                         // Bump the clock sequence number and also randomize the counter offset,
 374                         // which is useful for UIDs that do not include the clock sequence number.
 375                         $clkSeq = ( $clkSeq + 1 ) % $clockSeqSize;
 376                         $offset = mt_rand( 0, $counterSize - 1 );
 377                         trigger_error( "Clock was set back; sequence number incremented." );
 378                 }
 379                 // Update the (clock sequence number, timestamp, counter)
 380                 ftruncate( $handle, 0 );
 381                 rewind( $handle );
 382                 fwrite( $handle, "{$clkSeq} {$time[0]} {$time[1]} {$counter} {$offset}" );
 383                 fflush( $handle );
 384                 // Release the UID lock file
 385                 flock( $handle, LOCK_UN );
 386
 387                 return array( $time, ( $counter + $offset ) % $counterSize, $clkSeq );
 388         }
 389
 390         /**
 391          * Wait till the current timestamp reaches $time and return the current
 392          * timestamp. This returns false if it would have to wait more than 10ms.
 393          *
 394          * @param array $time Result of UIDGenerator::millitime()
 395          * @return Array|bool UIDGenerator::millitime() result or false
 396          */
 397         protected function timeWaitUntil( array $time ) {
 398                 do {
 399                         $ct = self::millitime();
 400                         if ( $ct >= $time ) { // http://php.net/manual/en/language.operators.comparison.php
 401                                 return $ct; // current timestamp is higher than $time
 402                         }
 403                 } while ( ( ( $time[0] - $ct[0] ) * 1000 + ( $time[1] - $ct[1] ) ) <= 10 );
 404
 405                 return false;
 406         }
 407
 408         /**
 409          * @param array $time Result of UIDGenerator::millitime()
 410          * @return string 46 MSBs of "milliseconds since epoch" in binary (rolls over in 4201)
 411          */
 412         protected function millisecondsSinceEpochBinary( array $time ) {
 413                 list( $sec, $msec ) = $time;
 414                 $ts = 1000 * $sec + $msec;
 415                 if ( $ts > pow( 2, 52 ) ) {
 416                         throw new MWException( __METHOD__ .
 417                                 ': sorry, this function doesn\'t work after the year 144680' );
 418                 }
 419
 420                 return substr( wfBaseConvert( $ts, 10, 2, 46 ), -46 );
 421         }
 422
 423         /**
 424          * @return Array (current time in seconds, milliseconds since then)
 425          */
 426         protected static function millitime() {
 427                 list( $msec, $sec ) = explode( ' ', microtime() );
 428
 429                 return array( (int)$sec, (int)( $msec * 1000 ) );
 430         }
 431
 432         function __destruct() {
 433                 array_map( 'fclose', $this->fileHandles );
 434         }
 435 }