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 use Wikimedia\Assert\Assert;
  24
  25 /**
  26  * Class for getting statistically unique IDs
  27  *
  28  * @since 1.21
  29  */
  30 class UIDGenerator {
  31         /** @var UIDGenerator */
  32         protected static $instance = null;
  33
  34         protected $nodeIdFile; // string; local file path
  35         protected $nodeId32; // string; node ID in binary (32 bits)
  36         protected $nodeId48; // string; node ID in binary (48 bits)
  37
  38         protected $lockFile88; // string; local file path
  39         protected $lockFile128; // string; local file path
  40
  41         /** @var array */
  42         protected $fileHandles = array(); // cache file handles
  43
  44         const QUICK_RAND = 1; // get randomness from fast and insecure sources
  45         const QUICK_VOLATILE = 2; // use an APC like in-memory counter if available
  46
  47         protected function __construct() {
  48                 $this->nodeIdFile = wfTempDir() . '/mw-' . __CLASS__ . '-UID-nodeid';
  49                 $nodeId = '';
  50                 if ( is_file( $this->nodeIdFile ) ) {
  51                         $nodeId = file_get_contents( $this->nodeIdFile );
  52                 }
  53                 // Try to get some ID that uniquely identifies this machine (RFC 4122)...
  54                 if ( !preg_match( '/^[0-9a-f]{12}$/i', $nodeId ) ) {
  55                         MediaWiki\suppressWarnings();
  56                         if ( wfIsWindows() ) {
  57                                 // http://technet.microsoft.com/en-us/library/bb490913.aspx
  58                                 $csv = trim( wfShellExec( 'getmac /NH /FO CSV' ) );
  59                                 $line = substr( $csv, 0, strcspn( $csv, "\n" ) );
  60                                 $info = str_getcsv( $line );
  61                                 $nodeId = isset( $info[0] ) ? str_replace( '-', '', $info[0] ) : '';
  62                         } elseif ( is_executable( '/sbin/ifconfig' ) ) { // Linux/BSD/Solaris/OS X
  63                                 // See http://linux.die.net/man/8/ifconfig
  64                                 $m = array();
  65                                 preg_match( '/\s([0-9a-f]{2}(:[0-9a-f]{2}){5})\s/',
  66                                         wfShellExec( '/sbin/ifconfig -a' ), $m );
  67                                 $nodeId = isset( $m[1] ) ? str_replace( ':', '', $m[1] ) : '';
  68                         }
  69                         MediaWiki\restoreWarnings();
  70                         if ( !preg_match( '/^[0-9a-f]{12}$/i', $nodeId ) ) {
  71                                 $nodeId = MWCryptRand::generateHex( 12, true );
  72                                 $nodeId[1] = dechex( hexdec( $nodeId[1] ) | 0x1 ); // set multicast bit
  73                         }
  74                         file_put_contents( $this->nodeIdFile, $nodeId ); // cache
  75                 }
  76                 $this->nodeId32 = Wikimedia\base_convert( substr( sha1( $nodeId ), 0, 8 ), 16, 2, 32 );
  77                 $this->nodeId48 = Wikimedia\base_convert( $nodeId, 16, 2, 48 );
  78                 // If different processes run as different users, they may have different temp dirs.
  79                 // This is dealt with by initializing the clock sequence number and counters randomly.
  80                 $this->lockFile88 = wfTempDir() . '/mw-' . __CLASS__ . '-UID-88';
  81                 $this->lockFile128 = wfTempDir() . '/mw-' . __CLASS__ . '-UID-128';
  82         }
  83
  84         /**
  85          * @todo: move to MW-specific factory class and inject temp dir
  86          * @return UIDGenerator
  87          */
  88         protected static function singleton() {
  89                 if ( self::$instance === null ) {
  90                         self::$instance = new self();
  91                 }
  92
  93                 return self::$instance;
  94         }
  95
  96         /**
  97          * Get a statistically unique 88-bit unsigned integer ID string.
  98          * The bits of the UID are prefixed with the time (down to the millisecond).
  99          *
 100          * These IDs are suitable as values for the shard key of distributed data.
 101          * If a column uses these as values, it should be declared UNIQUE to handle collisions.
 102          * New rows almost always have higher UIDs, which makes B-TREE updates on INSERT fast.
 103          * They can also be stored "DECIMAL(27) UNSIGNED" or BINARY(11) in MySQL.
 104          *
 105          * UID generation is serialized on each server (as the node ID is for the whole machine).
 106          *
 107          * @param int $base Specifies a base other than 10
 108          * @return string Number
 109          * @throws RuntimeException
 110          */
 111         public static function newTimestampedUID88( $base = 10 ) {
 112                 Assert::parameterType( 'integer', $base, '$base' );
 113                 Assert::parameter( $base <= 36, '$base', 'must be <= 36' );
 114                 Assert::parameter( $base >= 2, '$base', 'must be >= 2' );
 115
 116                 $gen = self::singleton();
 117                 $time = $gen->getTimestampAndDelay( 'lockFile88', 1, 1024 );
 118
 119                 return Wikimedia\base_convert( $gen->getTimestampedID88( $time ), 2, $base );
 120         }
 121
 122         /**
 123          * @param array $info (UIDGenerator::millitime(), counter, clock sequence)
 124          * @return string 88 bits
 125          * @throws RuntimeException
 126          */
 127         protected function getTimestampedID88( array $info ) {
 128                 list( $time, $counter ) = $info;
 129                 // Take the 46 MSBs of "milliseconds since epoch"
 130                 $id_bin = $this->millisecondsSinceEpochBinary( $time );
 131                 // Add a 10 bit counter resulting in 56 bits total
 132                 $id_bin .= str_pad( decbin( $counter ), 10, '0', STR_PAD_LEFT );
 133                 // Add the 32 bit node ID resulting in 88 bits total
 134                 $id_bin .= $this->nodeId32;
 135                 // Convert to a 1-27 digit integer string
 136                 if ( strlen( $id_bin ) !== 88 ) {
 137                         throw new RuntimeException( "Detected overflow for millisecond timestamp." );
 138                 }
 139
 140                 return $id_bin;
 141         }
 142
 143         /**
 144          * Get a statistically unique 128-bit unsigned integer ID string.
 145          * The bits of the UID are prefixed with the time (down to the millisecond).
 146          *
 147          * These IDs are suitable as globally unique IDs, without any enforced uniqueness.
 148          * New rows almost always have higher UIDs, which makes B-TREE updates on INSERT fast.
 149          * They can also be stored as "DECIMAL(39) UNSIGNED" or BINARY(16) in MySQL.
 150          *
 151          * UID generation is serialized on each server (as the node ID is for the whole machine).
 152          *
 153          * @param int $base Specifies a base other than 10
 154          * @return string Number
 155          * @throws RuntimeException
 156          */
 157         public static function newTimestampedUID128( $base = 10 ) {
 158                 Assert::parameterType( 'integer', $base, '$base' );
 159                 Assert::parameter( $base <= 36, '$base', 'must be <= 36' );
 160                 Assert::parameter( $base >= 2, '$base', 'must be >= 2' );
 161
 162                 $gen = self::singleton();
 163                 $time = $gen->getTimestampAndDelay( 'lockFile128', 16384, 1048576 );
 164
 165                 return Wikimedia\base_convert( $gen->getTimestampedID128( $time ), 2, $base );
 166         }
 167
 168         /**
 169          * @param array $info (UIDGenerator::millitime(), counter, clock sequence)
 170          * @return string 128 bits
 171          * @throws RuntimeException
 172          */
 173         protected function getTimestampedID128( array $info ) {
 174                 list( $time, $counter, $clkSeq ) = $info;
 175                 // Take the 46 MSBs of "milliseconds since epoch"
 176                 $id_bin = $this->millisecondsSinceEpochBinary( $time );
 177                 // Add a 20 bit counter resulting in 66 bits total
 178                 $id_bin .= str_pad( decbin( $counter ), 20, '0', STR_PAD_LEFT );
 179                 // Add a 14 bit clock sequence number resulting in 80 bits total
 180                 $id_bin .= str_pad( decbin( $clkSeq ), 14, '0', STR_PAD_LEFT );
 181                 // Add the 48 bit node ID resulting in 128 bits total
 182                 $id_bin .= $this->nodeId48;
 183                 // Convert to a 1-39 digit integer string
 184                 if ( strlen( $id_bin ) !== 128 ) {
 185                         throw new RuntimeException( "Detected overflow for millisecond timestamp." );
 186                 }
 187
 188                 return $id_bin;
 189         }
 190
 191         /**
 192          * Return an RFC4122 compliant v4 UUID
 193          *
 194          * @param int $flags Bitfield (supports UIDGenerator::QUICK_RAND)
 195          * @return string
 196          * @throws RuntimeException
 197          */
 198         public static function newUUIDv4( $flags = 0 ) {
 199                 $hex = ( $flags & self::QUICK_RAND )
 200                         ? wfRandomString( 31 )
 201                         : MWCryptRand::generateHex( 31 );
 202
 203                 return sprintf( '%s-%s-%s-%s-%s',
 204                         // "time_low" (32 bits)
 205                         substr( $hex, 0, 8 ),
 206                         // "time_mid" (16 bits)
 207                         substr( $hex, 8, 4 ),
 208                         // "time_hi_and_version" (16 bits)
 209                         '4' . substr( $hex, 12, 3 ),
 210                         // "clk_seq_hi_res (8 bits, variant is binary 10x) and "clk_seq_low" (8 bits)
 211                         dechex( 0x8 | ( hexdec( $hex[15] ) & 0x3 ) ) . $hex[16] . substr( $hex, 17, 2 ),
 212                         // "node" (48 bits)
 213                         substr( $hex, 19, 12 )
 214                 );
 215         }
 216
 217         /**
 218          * Return an RFC4122 compliant v4 UUID
 219          *
 220          * @param int $flags Bitfield (supports UIDGenerator::QUICK_RAND)
 221          * @return string 32 hex characters with no hyphens
 222          * @throws RuntimeException
 223          */
 224         public static function newRawUUIDv4( $flags = 0 ) {
 225                 return str_replace( '-', '', self::newUUIDv4( $flags ) );
 226         }
 227
 228         /**
 229          * Return an ID that is sequential *only* for this node and bucket
 230          *
 231          * These IDs are suitable for per-host sequence numbers, e.g. for some packet protocols.
 232          * If UIDGenerator::QUICK_VOLATILE is used the counter might reset on server restart.
 233          *
 234          * @param string $bucket Arbitrary bucket name (should be ASCII)
 235          * @param int $bits Bit size (<=48) of resulting numbers before wrap-around
 236          * @param int $flags (supports UIDGenerator::QUICK_VOLATILE)
 237          * @return float Integer value as float
 238          * @since 1.23
 239          */
 240         public static function newSequentialPerNodeID( $bucket, $bits = 48, $flags = 0 ) {
 241                 return current( self::newSequentialPerNodeIDs( $bucket, $bits, 1, $flags ) );
 242         }
 243
 244         /**
 245          * Return IDs that are sequential *only* for this node and bucket
 246          *
 247          * @see UIDGenerator::newSequentialPerNodeID()
 248          * @param string $bucket Arbitrary bucket name (should be ASCII)
 249          * @param int $bits Bit size (16 to 48) of resulting numbers before wrap-around
 250          * @param int $count Number of IDs to return (1 to 10000)
 251          * @param int $flags (supports UIDGenerator::QUICK_VOLATILE)
 252          * @return array Ordered list of float integer values
 253          * @since 1.23
 254          */
 255         public static function newSequentialPerNodeIDs( $bucket, $bits, $count, $flags = 0 ) {
 256                 $gen = self::singleton();
 257                 return $gen->getSequentialPerNodeIDs( $bucket, $bits, $count, $flags );
 258         }
 259
 260         /**
 261          * Return IDs that are sequential *only* for this node and bucket
 262          *
 263          * @see UIDGenerator::newSequentialPerNodeID()
 264          * @param string $bucket Arbitrary bucket name (should be ASCII)
 265          * @param int $bits Bit size (16 to 48) of resulting numbers before wrap-around
 266          * @param int $count Number of IDs to return (1 to 10000)
 267          * @param int $flags (supports UIDGenerator::QUICK_VOLATILE)
 268          * @return array Ordered list of float integer values
 269          * @throws RuntimeException
 270          */
 271         protected function getSequentialPerNodeIDs( $bucket, $bits, $count, $flags ) {
 272                 if ( $count <= 0 ) {
 273                         return array(); // nothing to do
 274                 } elseif ( $count > 10000 ) {
 275                         throw new RuntimeException( "Number of requested IDs ($count) is too high." );
 276                 } elseif ( $bits < 16 || $bits > 48 ) {
 277                         throw new RuntimeException( "Requested bit size ($bits) is out of range." );
 278                 }
 279
 280                 $counter = null; // post-increment persistent counter value
 281
 282                 // Use APC/eAccelerator/xcache if requested, available, and not in CLI mode;
 283                 // Counter values would not survive accross script instances in CLI mode.
 284                 $cache = null;
 285                 if ( ( $flags & self::QUICK_VOLATILE ) && PHP_SAPI !== 'cli' ) {
 286                         $cache = ObjectCache::getLocalServerInstance();
 287                 }
 288                 if ( $cache ) {
 289                         $counter = $cache->incr( $bucket, $count );
 290                         if ( $counter === false ) {
 291                                 if ( !$cache->add( $bucket, (int)$count ) ) {
 292                                         throw new RuntimeException( '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 ( $handle === false ) {
 310                                 throw new RuntimeException( "Could not open '{$path}'." );
 311                         } elseif ( !flock( $handle, LOCK_EX ) ) {
 312                                 fclose( $handle );
 313                                 throw new RuntimeException( "Could not acquire '{$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 RuntimeException
 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 ( $handle === false ) {
 359                         throw new RuntimeException( "Could not open '{$this->$lockFile}'." );
 360                 } elseif ( !flock( $handle, LOCK_EX ) ) {
 361                         fclose( $handle );
 362                         throw new RuntimeException( "Could not acquire '{$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 RuntimeException( "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 RuntimeException( "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 RuntimeException
 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 RuntimeException( __METHOD__ .
 449                                 ': sorry, this function doesn\'t work after the year 144680' );
 450                 }
 451
 452                 return substr( Wikimedia\base_convert( $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 }