includes/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
  44         protected function __construct() {
  45                 $idFile = wfTempDir() . '/mw-' . __CLASS__ . '-UID-nodeid';
  46                 $nodeId = is_file( $idFile ) ? file_get_contents( $idFile ) : '';
  47                 // Try to get some ID that uniquely identifies this machine (RFC 4122)...
  48                 if ( !preg_match( '/^[0-9a-f]{12}$/i', $nodeId ) ) {
  49                         wfSuppressWarnings();
  50                         if ( wfIsWindows() ) {
  51                                 // http://technet.microsoft.com/en-us/library/bb490913.aspx
  52                                 $csv = trim( wfShellExec( 'getmac /NH /FO CSV' ) );
  53                                 $line = substr( $csv, 0, strcspn( $csv, "\n" ) );
  54                                 $info = str_getcsv( $line );
  55                                 $nodeId = isset( $info[0] ) ? str_replace( '-', '', $info[0] ) : '';
  56                         } elseif ( is_executable( '/sbin/ifconfig' ) ) { // Linux/BSD/Solaris/OS X
  57                                 // See http://linux.die.net/man/8/ifconfig
  58                                 $m = array();
  59                                 preg_match( '/\s([0-9a-f]{2}(:[0-9a-f]{2}){5})\s/',
  60                                         wfShellExec( '/sbin/ifconfig -a' ), $m );
  61                                 $nodeId = isset( $m[1] ) ? str_replace( ':', '', $m[1] ) : '';
  62                         }
  63                         wfRestoreWarnings();
  64                         if ( !preg_match( '/^[0-9a-f]{12}$/i', $nodeId ) ) {
  65                                 $nodeId = MWCryptRand::generateHex( 12, true );
  66                                 $nodeId[1] = dechex( hexdec( $nodeId[1] ) | 0x1 ); // set multicast bit
  67                         }
  68                         file_put_contents( $idFile, $nodeId ); // cache
  69                 }
  70                 $this->nodeId32 = wfBaseConvert( substr( sha1( $nodeId ), 0, 8 ), 16, 2, 32 );
  71                 $this->nodeId48 = wfBaseConvert( $nodeId, 16, 2, 48 );
  72                 // If different processes run as different users, they may have different temp dirs.
  73                 // This is dealt with by initializing the clock sequence number and counters randomly.
  74                 $this->lockFile88 = wfTempDir() . '/mw-' . __CLASS__ . '-UID-88';
  75                 $this->lockFile128 = wfTempDir() . '/mw-' . __CLASS__ . '-UID-128';
  76         }
  77
  78         /**
  79          * @return UIDGenerator
  80          */
  81         protected static function singleton() {
  82                 if ( self::$instance === null ) {
  83                         self::$instance = new self();
  84                 }
  85                 return self::$instance;
  86         }
  87
  88         /**
  89          * Get a statistically unique 88-bit unsigned integer ID string.
  90          * The bits of the UID are prefixed with the time (down to the millisecond).
  91          *
  92          * These IDs are suitable as values for the shard key of distributed data.
  93          * If a column uses these as values, it should be declared UNIQUE to handle collisions.
  94          * New rows almost always have higher UIDs, which makes B-TREE updates on INSERT fast.
  95          * They can also be stored "DECIMAL(27) UNSIGNED" or BINARY(11) in MySQL.
  96          *
  97          * UID generation is serialized on each server (as the node ID is for the whole machine).
  98          *
  99          * @param $base integer Specifies a base other than 10
 100          * @return string Number
 101          * @throws MWException
 102          */
 103         public static function newTimestampedUID88( $base = 10 ) {
 104                 if ( !is_integer( $base ) || $base > 36 || $base < 2 ) {
 105                         throw new MWException( "Base must an integer be between 2 and 36" );
 106                 }
 107                 $gen = self::singleton();
 108                 $time = $gen->getTimestampAndDelay( 'lockFile88', 1, 1024 );
 109                 return wfBaseConvert( $gen->getTimestampedID88( $time ), 2, $base );
 110         }
 111
 112         /**
 113          * @param array $time (UIDGenerator::millitime(), clock sequence)
 114          * @return string 88 bits
 115          */
 116         protected function getTimestampedID88( array $info ) {
 117                 list( $time, $counter ) = $info;
 118                 // Take the 46 MSBs of "milliseconds since epoch"
 119                 $id_bin = $this->millisecondsSinceEpochBinary( $time );
 120                 // Add a 10 bit counter resulting in 56 bits total
 121                 $id_bin .= str_pad( decbin( $counter ), 10, '0', STR_PAD_LEFT );
 122                 // Add the 32 bit node ID resulting in 88 bits total
 123                 $id_bin .= $this->nodeId32;
 124                 // Convert to a 1-27 digit integer string
 125                 if ( strlen( $id_bin ) !== 88 ) {
 126                         throw new MWException( "Detected overflow for millisecond timestamp." );
 127                 }
 128                 return $id_bin;
 129         }
 130
 131         /**
 132          * Get a statistically unique 128-bit unsigned integer ID string.
 133          * The bits of the UID are prefixed with the time (down to the millisecond).
 134          *
 135          * These IDs are suitable as globally unique IDs, without any enforced uniqueness.
 136          * New rows almost always have higher UIDs, which makes B-TREE updates on INSERT fast.
 137          * They can also be stored as "DECIMAL(39) UNSIGNED" or BINARY(16) in MySQL.
 138          *
 139          * UID generation is serialized on each server (as the node ID is for the whole machine).
 140          *
 141          * @param $base integer Specifies a base other than 10
 142          * @return string Number
 143          * @throws MWException
 144          */
 145         public static function newTimestampedUID128( $base = 10 ) {
 146                 if ( !is_integer( $base ) || $base > 36 || $base < 2 ) {
 147                         throw new MWException( "Base must be an integer between 2 and 36" );
 148                 }
 149                 $gen = self::singleton();
 150                 $time = $gen->getTimestampAndDelay( 'lockFile128', 16384, 1048576 );
 151                 return wfBaseConvert( $gen->getTimestampedID128( $time ), 2, $base );
 152         }
 153
 154         /**
 155          * @param array $info (UIDGenerator::millitime(), counter, clock sequence)
 156          * @return string 128 bits
 157          */
 158         protected function getTimestampedID128( array $info ) {
 159                 list( $time, $counter, $clkSeq ) = $info;
 160                 // Take the 46 MSBs of "milliseconds since epoch"
 161                 $id_bin = $this->millisecondsSinceEpochBinary( $time );
 162                 // Add a 20 bit counter resulting in 66 bits total
 163                 $id_bin .= str_pad( decbin( $counter ), 20, '0', STR_PAD_LEFT );
 164                 // Add a 14 bit clock sequence number resulting in 80 bits total
 165                 $id_bin .= str_pad( decbin( $clkSeq ), 14, '0', STR_PAD_LEFT );
 166                 // Add the 48 bit node ID resulting in 128 bits total
 167                 $id_bin .= $this->nodeId48;
 168                 // Convert to a 1-39 digit integer string
 169                 if ( strlen( $id_bin ) !== 128 ) {
 170                         throw new MWException( "Detected overflow for millisecond timestamp." );
 171                 }
 172                 return $id_bin;
 173         }
 174
 175         /**
 176          * Return an RFC4122 compliant v4 UUID
 177          *
 178          * @param $flags integer Bitfield (supports UIDGenerator::QUICK_RAND)
 179          * @return string
 180          * @throws MWException
 181          */
 182         public static function newUUIDv4( $flags = 0 ) {
 183                 $hex = ( $flags & self::QUICK_RAND )
 184                         ? wfRandomString( 31 )
 185                         : MWCryptRand::generateHex( 31 );
 186
 187                 return sprintf( '%s-%s-%s-%s-%s',
 188                         // "time_low" (32 bits)
 189                         substr( $hex, 0, 8 ),
 190                         // "time_mid" (16 bits)
 191                         substr( $hex, 8, 4 ),
 192                         // "time_hi_and_version" (16 bits)
 193                         '4' . substr( $hex, 12, 3 ),
 194                         // "clk_seq_hi_res (8 bits, variant is binary 10x) and "clk_seq_low" (8 bits)
 195                         dechex( 0x8 | ( hexdec( $hex[15] ) & 0x3 ) ) . $hex[16] . substr( $hex, 17, 2 ),
 196                         // "node" (48 bits)
 197                         substr( $hex, 19, 12 )
 198                 );
 199         }
 200
 201         /**
 202          * Return an RFC4122 compliant v4 UUID
 203          *
 204          * @param $flags integer Bitfield (supports UIDGenerator::QUICK_RAND)
 205          * @return string 32 hex characters with no hyphens
 206          * @throws MWException
 207          */
 208         public static function newRawUUIDv4( $flags = 0 ) {
 209                 return str_replace( '-', '', self::newUUIDv4( $flags ) );
 210         }
 211
 212         /**
 213          * Get a (time,counter,clock sequence) where (time,counter) is higher
 214          * than any previous (time,counter) value for the given clock sequence.
 215          * This is useful for making UIDs sequential on a per-node bases.
 216          *
 217          * @param string $lockFile Name of a local lock file
 218          * @param $clockSeqSize integer The number of possible clock sequence values
 219          * @param $counterSize integer The number of possible counter values
 220          * @return Array (result of UIDGenerator::millitime(), counter, clock sequence)
 221          * @throws MWException
 222          */
 223         protected function getTimestampAndDelay( $lockFile, $clockSeqSize, $counterSize ) {
 224                 // Get the UID lock file handle
 225                 if ( isset( $this->fileHandles[$lockFile] ) ) {
 226                         $handle = $this->fileHandles[$lockFile];
 227                 } else {
 228                         $handle = fopen( $this->$lockFile, 'cb+' );
 229                         $this->fileHandles[$lockFile] = $handle ?: null; // cache
 230                 }
 231                 // Acquire the UID lock file
 232                 if ( $handle === false ) {
 233                         throw new MWException( "Could not open '{$this->$lockFile}'." );
 234                 } elseif ( !flock( $handle, LOCK_EX ) ) {
 235                         throw new MWException( "Could not acquire '{$this->$lockFile}'." );
 236                 }
 237                 // Get the current timestamp, clock sequence number, last time, and counter
 238                 rewind( $handle );
 239                 $data = explode( ' ', fgets( $handle ) ); // "<clk seq> <sec> <msec> <counter> <offset>"
 240                 $clockChanged = false; // clock set back significantly?
 241                 if ( count( $data ) == 5 ) { // last UID info already initialized
 242                         $clkSeq = (int) $data[0] % $clockSeqSize;
 243                         $prevTime = array( (int) $data[1], (int) $data[2] );
 244                         $offset = (int) $data[4] % $counterSize; // random counter offset
 245                         $counter = 0; // counter for UIDs with the same timestamp
 246                         // Delay until the clock reaches the time of the last ID.
 247                         // This detects any microtime() drift among processes.
 248                         $time = $this->timeWaitUntil( $prevTime );
 249                         if ( !$time ) { // too long to delay?
 250                                 $clockChanged = true; // bump clock sequence number
 251                                 $time = self::millitime();
 252                         } elseif ( $time == $prevTime ) {
 253                                 // Bump the counter if there are timestamp collisions
 254                                 $counter = (int) $data[3] % $counterSize;
 255                                 if ( ++$counter >= $counterSize ) { // sanity (starts at 0)
 256                                         flock( $handle, LOCK_UN ); // abort
 257                                         throw new MWException( "Counter overflow for timestamp value." );
 258                                 }
 259                         }
 260                 } else { // last UID info not initialized
 261                         $clkSeq = mt_rand( 0, $clockSeqSize - 1 );
 262                         $counter = 0;
 263                         $offset = mt_rand( 0, $counterSize - 1 );
 264                         $time = self::millitime();
 265                 }
 266                 // microtime() and gettimeofday() can drift from time() at least on Windows.
 267                 // The drift is immediate for processes running while the system clock changes.
 268                 // time() does not have this problem. See https://bugs.php.net/bug.php?id=42659.
 269                 if ( abs( time() - $time[0] ) >= 2 ) {
 270                         // We don't want processes using too high or low timestamps to avoid duplicate
 271                         // UIDs and clock sequence number churn. This process should just be restarted.
 272                         flock( $handle, LOCK_UN ); // abort
 273                         throw new MWException( "Process clock is outdated or drifted." );
 274                 }
 275                 // If microtime() is synced and a clock change was detected, then the clock went back
 276                 if ( $clockChanged ) {
 277                         // Bump the clock sequence number and also randomize the counter offset,
 278                         // which is useful for UIDs that do not include the clock sequence number.
 279                         $clkSeq = ( $clkSeq + 1 ) % $clockSeqSize;
 280                         $offset = mt_rand( 0, $counterSize - 1 );
 281                         trigger_error( "Clock was set back; sequence number incremented." );
 282                 }
 283                 // Update the (clock sequence number, timestamp, counter)
 284                 ftruncate( $handle, 0 );
 285                 rewind( $handle );
 286                 fwrite( $handle, "{$clkSeq} {$time[0]} {$time[1]} {$counter} {$offset}" );
 287                 fflush( $handle );
 288                 // Release the UID lock file
 289                 flock( $handle, LOCK_UN );
 290
 291                 return array( $time, ( $counter + $offset ) % $counterSize, $clkSeq );
 292         }
 293
 294         /**
 295          * Wait till the current timestamp reaches $time and return the current
 296          * timestamp. This returns false if it would have to wait more than 10ms.
 297          *
 298          * @param array $time Result of UIDGenerator::millitime()
 299          * @return Array|bool UIDGenerator::millitime() result or false
 300          */
 301         protected function timeWaitUntil( array $time ) {
 302                 do {
 303                         $ct = self::millitime();
 304                         if ( $ct >= $time ) { // http://php.net/manual/en/language.operators.comparison.php
 305                                 return $ct; // current timestamp is higher than $time
 306                         }
 307                 } while ( ( ( $time[0] - $ct[0] ) * 1000 + ( $time[1] - $ct[1] ) ) <= 10 );
 308
 309                 return false;
 310         }
 311
 312         /**
 313          * @param array $time Result of UIDGenerator::millitime()
 314          * @return string 46 MSBs of "milliseconds since epoch" in binary (rolls over in 4201)
 315          */
 316         protected function millisecondsSinceEpochBinary( array $time ) {
 317                 list( $sec, $msec ) = $time;
 318                 $ts = 1000 * $sec + $msec;
 319                 if ( $ts > pow( 2, 52 ) ) {
 320                         throw new MWException( __METHOD__ .
 321                                 ': sorry, this function doesn\'t work after the year 144680' );
 322                 }
 323                 return substr( wfBaseConvert( $ts, 10, 2, 46 ), -46 );
 324         }
 325
 326         /**
 327          * @return Array (current time in seconds, milliseconds since then)
 328          */
 329         protected static function millitime() {
 330                 list( $msec, $sec ) = explode( ' ', microtime() );
 331                 return array( (int) $sec, (int) ( $msec * 1000 ) );
 332         }
 333
 334         function __destruct() {
 335                 array_map( 'fclose', $this->fileHandles );
 336         }
 337 }