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 $nodeIdFile; // string; local file path
  34         protected $nodeId32; // string; node ID in binary (32 bits)
  35         protected $nodeId48; // string; node ID in binary (48 bits)
  36
  37         protected $lockFile88; // string; local file path
  38         protected $lockFile128; // string; local file path
  39
  40         /** @var array */
  41         protected $fileHandles = array(); // cache file handles
  42
  43         const QUICK_RAND = 1; // get randomness from fast and insecure sources
  44         const QUICK_VOLATILE = 2; // use an APC like in-memory counter if available
  45
  46         protected function __construct() {
  47                 $this->nodeIdFile = wfTempDir() . '/mw-' . __CLASS__ . '-UID-nodeid';
  48                 $nodeId = '';
  49                 if ( is_file( $this->nodeIdFile ) ) {
  50                         $nodeId = file_get_contents( $this->nodeIdFile );
  51                 }
  52                 // Try to get some ID that uniquely identifies this machine (RFC 4122)...
  53                 if ( !preg_match( '/^[0-9a-f]{12}$/i', $nodeId ) ) {
  54                         wfSuppressWarnings();
  55                         if ( wfIsWindows() ) {
  56                                 // http://technet.microsoft.com/en-us/library/bb490913.aspx
  57                                 $csv = trim( wfShellExec( 'getmac /NH /FO CSV' ) );
  58                                 $line = substr( $csv, 0, strcspn( $csv, "\n" ) );
  59                                 $info = str_getcsv( $line );
  60                                 $nodeId = isset( $info[0] ) ? str_replace( '-', '', $info[0] ) : '';
  61                         } elseif ( is_executable( '/sbin/ifconfig' ) ) { // Linux/BSD/Solaris/OS X
  62                                 // See http://linux.die.net/man/8/ifconfig
  63                                 $m = array();
  64                                 preg_match( '/\s([0-9a-f]{2}(:[0-9a-f]{2}){5})\s/',
  65                                         wfShellExec( '/sbin/ifconfig -a' ), $m );
  66                                 $nodeId = isset( $m[1] ) ? str_replace( ':', '', $m[1] ) : '';
  67                         }
  68                         wfRestoreWarnings();
  69                         if ( !preg_match( '/^[0-9a-f]{12}$/i', $nodeId ) ) {
  70                                 $nodeId = MWCryptRand::generateHex( 12, true );
  71                                 $nodeId[1] = dechex( hexdec( $nodeId[1] ) | 0x1 ); // set multicast bit
  72                         }
  73                         file_put_contents( $this->nodeIdFile, $nodeId ); // cache
  74                 }
  75                 $this->nodeId32 = wfBaseConvert( substr( sha1( $nodeId ), 0, 8 ), 16, 2, 32 );
  76                 $this->nodeId48 = wfBaseConvert( $nodeId, 16, 2, 48 );
  77                 // If different processes run as different users, they may have different temp dirs.
  78                 // This is dealt with by initializing the clock sequence number and counters randomly.
  79                 $this->lockFile88 = wfTempDir() . '/mw-' . __CLASS__ . '-UID-88';
  80                 $this->lockFile128 = wfTempDir() . '/mw-' . __CLASS__ . '-UID-128';
  81         }
  82
  83         /**
  84          * @return UIDGenerator
  85          */
  86         protected static function singleton() {
  87                 if ( self::$instance === null ) {
  88                         self::$instance = new self();
  89                 }
  90
  91                 return self::$instance;
  92         }
  93
  94         /**
  95          * Get a statistically unique 88-bit unsigned integer ID string.
  96          * The bits of the UID are prefixed with the time (down to the millisecond).
  97          *
  98          * These IDs are suitable as values for the shard key of distributed data.
  99          * If a column uses these as values, it should be declared UNIQUE to handle collisions.
 100          * New rows almost always have higher UIDs, which makes B-TREE updates on INSERT fast.
 101          * They can also be stored "DECIMAL(27) UNSIGNED" or BINARY(11) in MySQL.
 102          *
 103          * UID generation is serialized on each server (as the node ID is for the whole machine).
 104          *
 105          * @param int $base Specifies a base other than 10
 106          * @return string Number
 107          * @throws MWException
 108          */
 109         public static function newTimestampedUID88( $base = 10 ) {
 110                 if ( !is_integer( $base ) || $base > 36 || $base < 2 ) {
 111                         throw new MWException( "Base must an integer be between 2 and 36" );
 112                 }
 113                 $gen = self::singleton();
 114                 $time = $gen->getTimestampAndDelay( 'lockFile88', 1, 1024 );
 115
 116                 return wfBaseConvert( $gen->getTimestampedID88( $time ), 2, $base );
 117         }
 118
 119         /**
 120          * @param array $info (UIDGenerator::millitime(), counter, clock sequence)
 121          * @return string 88 bits
 122          * @throws MWException
 123          */
 124         protected function getTimestampedID88( array $info ) {
 125                 list( $time, $counter ) = $info;
 126                 // Take the 46 MSBs of "milliseconds since epoch"
 127                 $id_bin = $this->millisecondsSinceEpochBinary( $time );
 128                 // Add a 10 bit counter resulting in 56 bits total
 129                 $id_bin .= str_pad( decbin( $counter ), 10, '0', STR_PAD_LEFT );
 130                 // Add the 32 bit node ID resulting in 88 bits total
 131                 $id_bin .= $this->nodeId32;
 132                 // Convert to a 1-27 digit integer string
 133                 if ( strlen( $id_bin ) !== 88 ) {
 134                         throw new MWException( "Detected overflow for millisecond timestamp." );
 135                 }
 136
 137                 return $id_bin;
 138         }
 139
 140         /**
 141          * Get a statistically unique 128-bit unsigned integer ID string.
 142          * The bits of the UID are prefixed with the time (down to the millisecond).
 143          *
 144          * These IDs are suitable as globally unique IDs, without any enforced uniqueness.
 145          * New rows almost always have higher UIDs, which makes B-TREE updates on INSERT fast.
 146          * They can also be stored as "DECIMAL(39) UNSIGNED" or BINARY(16) in MySQL.
 147          *
 148          * UID generation is serialized on each server (as the node ID is for the whole machine).
 149          *
 150          * @param int $base Specifies a base other than 10
 151          * @return string Number
 152          * @throws MWException
 153          */
 154         public static function newTimestampedUID128( $base = 10 ) {
 155                 if ( !is_integer( $base ) || $base > 36 || $base < 2 ) {
 156                         throw new MWException( "Base must be an integer between 2 and 36" );
 157                 }
 158                 $gen = self::singleton();
 159                 $time = $gen->getTimestampAndDelay( 'lockFile128', 16384, 1048576 );
 160
 161                 return wfBaseConvert( $gen->getTimestampedID128( $time ), 2, $base );
 162         }
 163
 164         /**
 165          * @param array $info (UIDGenerator::millitime(), counter, clock sequence)
 166          * @return string 128 bits
 167          * @throws MWException
 168          */
 169         protected function getTimestampedID128( array $info ) {
 170                 list( $time, $counter, $clkSeq ) = $info;
 171                 // Take the 46 MSBs of "milliseconds since epoch"
 172                 $id_bin = $this->millisecondsSinceEpochBinary( $time );
 173                 // Add a 20 bit counter resulting in 66 bits total
 174                 $id_bin .= str_pad( decbin( $counter ), 20, '0', STR_PAD_LEFT );
 175                 // Add a 14 bit clock sequence number resulting in 80 bits total
 176                 $id_bin .= str_pad( decbin( $clkSeq ), 14, '0', STR_PAD_LEFT );
 177                 // Add the 48 bit node ID resulting in 128 bits total
 178                 $id_bin .= $this->nodeId48;
 179                 // Convert to a 1-39 digit integer string
 180                 if ( strlen( $id_bin ) !== 128 ) {
 181                         throw new MWException( "Detected overflow for millisecond timestamp." );
 182                 }
 183
 184                 return $id_bin;
 185         }
 186
 187         /**
 188          * Return an RFC4122 compliant v4 UUID
 189          *
 190          * @param int $flags Bitfield (supports UIDGenerator::QUICK_RAND)
 191          * @return string
 192          * @throws MWException
 193          */
 194         public static function newUUIDv4( $flags = 0 ) {
 195                 $hex = ( $flags & self::QUICK_RAND )
 196                         ? wfRandomString( 31 )
 197                         : MWCryptRand::generateHex( 31 );
 198
 199                 return sprintf( '%s-%s-%s-%s-%s',
 200                         // "time_low" (32 bits)
 201                         substr( $hex, 0, 8 ),
 202                         // "time_mid" (16 bits)
 203                         substr( $hex, 8, 4 ),
 204                         // "time_hi_and_version" (16 bits)
 205                         '4' . substr( $hex, 12, 3 ),
 206                         // "clk_seq_hi_res (8 bits, variant is binary 10x) and "clk_seq_low" (8 bits)
 207                         dechex( 0x8 | ( hexdec( $hex[15] ) & 0x3 ) ) . $hex[16] . substr( $hex, 17, 2 ),
 208                         // "node" (48 bits)
 209                         substr( $hex, 19, 12 )
 210                 );
 211         }
 212
 213         /**
 214          * Return an RFC4122 compliant v4 UUID
 215          *
 216          * @param int $flags Bitfield (supports UIDGenerator::QUICK_RAND)
 217          * @return string 32 hex characters with no hyphens
 218          * @throws MWException
 219          */
 220         public static function newRawUUIDv4( $flags = 0 ) {
 221                 return str_replace( '-', '', self::newUUIDv4( $flags ) );
 222         }
 223
 224         /**
 225          * Return an ID that is sequential *only* for this node and bucket
 226          *
 227          * These IDs are suitable for per-host sequence numbers, e.g. for some packet protocols.
 228          * If UIDGenerator::QUICK_VOLATILE is used the counter might reset on server restart.
 229          *
 230          * @param string $bucket Arbitrary bucket name (should be ASCII)
 231          * @param int $bits Bit size (<=48) of resulting numbers before wrap-around
 232          * @param int $flags (supports UIDGenerator::QUICK_VOLATILE)
 233          * @return float Integer value as float
 234          * @since 1.23
 235          */
 236         public static function newSequentialPerNodeID( $bucket, $bits = 48, $flags = 0 ) {
 237                 return current( self::newSequentialPerNodeIDs( $bucket, $bits, 1, $flags ) );
 238         }
 239
 240         /**
 241          * Return IDs that are sequential *only* for this node and bucket
 242          *
 243          * @see UIDGenerator::newSequentialPerNodeID()
 244          * @param string $bucket Arbitrary bucket name (should be ASCII)
 245          * @param int $bits Bit size (16 to 48) of resulting numbers before wrap-around
 246          * @param int $count Number of IDs to return (1 to 10000)
 247          * @param int $flags (supports UIDGenerator::QUICK_VOLATILE)
 248          * @return array Ordered list of float integer values
 249          * @since 1.23
 250          */
 251         public static function newSequentialPerNodeIDs( $bucket, $bits, $count, $flags = 0 ) {
 252                 $gen = self::singleton();
 253                 return $gen->getSequentialPerNodeIDs( $bucket, $bits, $count, $flags );
 254         }
 255
 256         /**
 257          * Return IDs that are sequential *only* for this node and bucket
 258          *
 259          * @see UIDGenerator::newSequentialPerNodeID()
 260          * @param string $bucket Arbitrary bucket name (should be ASCII)
 261          * @param int $bits Bit size (16 to 48) of resulting numbers before wrap-around
 262          * @param int $count Number of IDs to return (1 to 10000)
 263          * @param int $flags (supports UIDGenerator::QUICK_VOLATILE)
 264          * @return array Ordered list of float integer values
 265          * @throws MWException
 266          */
 267         protected function getSequentialPerNodeIDs( $bucket, $bits, $count, $flags ) {
 268                 if ( $count <= 0 ) {
 269                         return array(); // nothing to do
 270                 } elseif ( $count > 10000 ) {
 271                         throw new MWException( "Number of requested IDs ($count) is too high." );
 272                 } elseif ( $bits < 16 || $bits > 48 ) {
 273                         throw new MWException( "Requested bit size ($bits) is out of range." );
 274                 }
 275
 276                 $counter = null; // post-increment persistent counter value
 277
 278                 // Use APC/eAccelerator/xcache if requested, available, and not in CLI mode;
 279                 // Counter values would not survive accross script instances in CLI mode.
 280                 $cache = null;
 281                 if ( ( $flags & self::QUICK_VOLATILE ) && PHP_SAPI !== 'cli' ) {
 282                         try {
 283                                 $cache = ObjectCache::newAccelerator( array() );
 284                         } catch ( Exception $e ) {
 285                                 // not supported
 286                         }
 287                 }
 288                 if ( $cache ) {
 289                         $counter = $cache->incr( $bucket, $count );
 290                         if ( $counter === false ) {
 291                                 if ( !$cache->add( $bucket, (int)$count ) ) {
 292                                         throw new MWException( 'Unable to set value to ' . get_class( $cache ) );
 293                                 }
 294                                 $counter = $count;
 295                         }
 296                 }
 297
 298                 // Note: use of fmod() avoids "division by zero" on 32 bit machines
 299                 if ( $counter === null ) {
 300                         $path = wfTempDir() . '/mw-' . __CLASS__ . '-' . rawurlencode( $bucket ) . '-48';
 301                         // Get the UID lock file handle
 302                         if ( isset( $this->fileHandles[$path] ) ) {
 303                                 $handle = $this->fileHandles[$path];
 304                         } else {
 305                                 $handle = fopen( $path, 'cb+' );
 306                                 $this->fileHandles[$path] = $handle ?: null; // cache
 307                         }
 308                         // Acquire the UID lock file
 309                         if ( $handle === false ) {
 310                                 throw new MWException( "Could not open '{$path}'." );
 311                         } elseif ( !flock( $handle, LOCK_EX ) ) {
 312                                 fclose( $handle );
 313                                 throw new MWException( "Could not acquire '{$path}'." );
 314                         }
 315                         // Fetch the counter value and increment it...
 316                         rewind( $handle );
 317                         $counter = floor( trim( fgets( $handle ) ) ) + $count; // fetch as float
 318                         // Write back the new counter value
 319                         ftruncate( $handle, 0 );
 320                         rewind( $handle );
 321                         fwrite( $handle, fmod( $counter, pow( 2, 48 ) ) ); // warp-around as needed
 322                         fflush( $handle );
 323                         // Release the UID lock file
 324                         flock( $handle, LOCK_UN );
 325                 }
 326
 327                 $ids = array();
 328                 $divisor = pow( 2, $bits );
 329                 $currentId = floor( $counter - $count ); // pre-increment counter value
 330                 for ( $i = 0; $i < $count; ++$i ) {
 331                         $ids[] = fmod( ++$currentId, $divisor );
 332                 }
 333
 334                 return $ids;
 335         }
 336
 337         /**
 338          * Get a (time,counter,clock sequence) where (time,counter) is higher
 339          * than any previous (time,counter) value for the given clock sequence.
 340          * This is useful for making UIDs sequential on a per-node bases.
 341          *
 342          * @param string $lockFile Name of a local lock file
 343          * @param int $clockSeqSize The number of possible clock sequence values
 344          * @param int $counterSize The number of possible counter values
 345          * @return array (result of UIDGenerator::millitime(), counter, clock sequence)
 346          * @throws MWException
 347          */
 348         protected function getTimestampAndDelay( $lockFile, $clockSeqSize, $counterSize ) {
 349                 // Get the UID lock file handle
 350                 $path = $this->$lockFile;
 351                 if ( isset( $this->fileHandles[$path] ) ) {
 352                         $handle = $this->fileHandles[$path];
 353                 } else {
 354                         $handle = fopen( $path, 'cb+' );
 355                         $this->fileHandles[$path] = $handle ?: null; // cache
 356                 }
 357                 // Acquire the UID lock file
 358                 if ( $handle === false ) {
 359                         throw new MWException( "Could not open '{$this->$lockFile}'." );
 360                 } elseif ( !flock( $handle, LOCK_EX ) ) {
 361                         fclose( $handle );
 362                         throw new MWException( "Could not acquire '{$this->$lockFile}'." );
 363                 }
 364                 // Get the current timestamp, clock sequence number, last time, and counter
 365                 rewind( $handle );
 366                 $data = explode( ' ', fgets( $handle ) ); // "<clk seq> <sec> <msec> <counter> <offset>"
 367                 $clockChanged = false; // clock set back significantly?
 368                 if ( count( $data ) == 5 ) { // last UID info already initialized
 369                         $clkSeq = (int)$data[0] % $clockSeqSize;
 370                         $prevTime = array( (int)$data[1], (int)$data[2] );
 371                         $offset = (int)$data[4] % $counterSize; // random counter offset
 372                         $counter = 0; // counter for UIDs with the same timestamp
 373                         // Delay until the clock reaches the time of the last ID.
 374                         // This detects any microtime() drift among processes.
 375                         $time = $this->timeWaitUntil( $prevTime );
 376                         if ( !$time ) { // too long to delay?
 377                                 $clockChanged = true; // bump clock sequence number
 378                                 $time = self::millitime();
 379                         } elseif ( $time == $prevTime ) {
 380                                 // Bump the counter if there are timestamp collisions
 381                                 $counter = (int)$data[3] % $counterSize;
 382                                 if ( ++$counter >= $counterSize ) { // sanity (starts at 0)
 383                                         flock( $handle, LOCK_UN ); // abort
 384                                         throw new MWException( "Counter overflow for timestamp value." );
 385                                 }
 386                         }
 387                 } else { // last UID info not initialized
 388                         $clkSeq = mt_rand( 0, $clockSeqSize - 1 );
 389                         $counter = 0;
 390                         $offset = mt_rand( 0, $counterSize - 1 );
 391                         $time = self::millitime();
 392                 }
 393                 // microtime() and gettimeofday() can drift from time() at least on Windows.
 394                 // The drift is immediate for processes running while the system clock changes.
 395                 // time() does not have this problem. See https://bugs.php.net/bug.php?id=42659.
 396                 if ( abs( time() - $time[0] ) >= 2 ) {
 397                         // We don't want processes using too high or low timestamps to avoid duplicate
 398                         // UIDs and clock sequence number churn. This process should just be restarted.
 399                         flock( $handle, LOCK_UN ); // abort
 400                         throw new MWException( "Process clock is outdated or drifted." );
 401                 }
 402                 // If microtime() is synced and a clock change was detected, then the clock went back
 403                 if ( $clockChanged ) {
 404                         // Bump the clock sequence number and also randomize the counter offset,
 405                         // which is useful for UIDs that do not include the clock sequence number.
 406                         $clkSeq = ( $clkSeq + 1 ) % $clockSeqSize;
 407                         $offset = mt_rand( 0, $counterSize - 1 );
 408                         trigger_error( "Clock was set back; sequence number incremented." );
 409                 }
 410                 // Update the (clock sequence number, timestamp, counter)
 411                 ftruncate( $handle, 0 );
 412                 rewind( $handle );
 413                 fwrite( $handle, "{$clkSeq} {$time[0]} {$time[1]} {$counter} {$offset}" );
 414                 fflush( $handle );
 415                 // Release the UID lock file
 416                 flock( $handle, LOCK_UN );
 417
 418                 return array( $time, ( $counter + $offset ) % $counterSize, $clkSeq );
 419         }
 420
 421         /**
 422          * Wait till the current timestamp reaches $time and return the current
 423          * timestamp. This returns false if it would have to wait more than 10ms.
 424          *
 425          * @param array $time Result of UIDGenerator::millitime()
 426          * @return array|bool UIDGenerator::millitime() result or false
 427          */
 428         protected function timeWaitUntil( array $time ) {
 429                 do {
 430                         $ct = self::millitime();
 431                         if ( $ct >= $time ) { // http://php.net/manual/en/language.operators.comparison.php
 432                                 return $ct; // current timestamp is higher than $time
 433                         }
 434                 } while ( ( ( $time[0] - $ct[0] ) * 1000 + ( $time[1] - $ct[1] ) ) <= 10 );
 435
 436                 return false;
 437         }
 438
 439         /**
 440          * @param array $time Result of UIDGenerator::millitime()
 441          * @return string 46 MSBs of "milliseconds since epoch" in binary (rolls over in 4201)
 442          * @throws MWException
 443          */
 444         protected function millisecondsSinceEpochBinary( array $time ) {
 445                 list( $sec, $msec ) = $time;
 446                 $ts = 1000 * $sec + $msec;
 447                 if ( $ts > pow( 2, 52 ) ) {
 448                         throw new MWException( __METHOD__ .
 449                                 ': sorry, this function doesn\'t work after the year 144680' );
 450                 }
 451
 452                 return substr( wfBaseConvert( $ts, 10, 2, 46 ), -46 );
 453         }
 454
 455         /**
 456          * @return array (current time in seconds, milliseconds since then)
 457          */
 458         protected static function millitime() {
 459                 list( $msec, $sec ) = explode( ' ', microtime() );
 460
 461                 return array( (int)$sec, (int)( $msec * 1000 ) );
 462         }
 463
 464         /**
 465          * Delete all cache files that have been created.
 466          *
 467          * This is a cleanup method primarily meant to be used from unit tests to
 468          * avoid poluting the local filesystem. If used outside of a unit test
 469          * environment it should be used with caution as it may destroy state saved
 470          * in the files.
 471          *
 472          * @see unitTestTearDown
 473          * @since 1.23
 474          */
 475         protected function deleteCacheFiles() {
 476                 // Bug: 44850
 477                 foreach ( $this->fileHandles as $path => $handle ) {
 478                         if ( $handle !== null ) {
 479                                 fclose( $handle );
 480                         }
 481                         if ( is_file( $path ) ) {
 482                                 unlink( $path );
 483                         }
 484                         unset( $this->fileHandles[$path] );
 485                 }
 486                 if ( is_file( $this->nodeIdFile ) ) {
 487                         unlink( $this->nodeIdFile );
 488                 }
 489         }
 490
 491         /**
 492          * Cleanup resources when tearing down after a unit test.
 493          *
 494          * This is a cleanup method primarily meant to be used from unit tests to
 495          * avoid poluting the local filesystem. If used outside of a unit test
 496          * environment it should be used with caution as it may destroy state saved
 497          * in the files.
 498          *
 499          * @see deleteCacheFiles
 500          * @since 1.23
 501          */
 502         public static function unitTestTearDown() {
 503                 // Bug: 44850
 504                 $gen = self::singleton();
 505                 $gen->deleteCacheFiles();
 506         }
 507
 508         function __destruct() {
 509                 array_map( 'fclose', array_filter( $this->fileHandles ) );
 510         }
 511 }