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