includes/ProxyTools.php

   1 <?php
   2 /**
   3  * Functions for dealing with proxies
   4  * @file
   5  */
   6
   7 /**
   8  * Extracts the XFF string from the request header
   9  * Checks first for "X-Forwarded-For", then "Client-ip"
  10  * Note: headers are spoofable
  11  * @return string
  12  */
  13 function wfGetForwardedFor() {
  14         if( function_exists( 'apache_request_headers' ) ) {
  15                 // More reliable than $_SERVER due to case and -/_ folding
  16                 $set = array ();
  17                 foreach ( apache_request_headers() as $tempName => $tempValue ) {
  18                         $set[ strtoupper( $tempName ) ] = $tempValue;
  19                 }
  20                 $index = strtoupper ( 'X-Forwarded-For' );
  21                 $index2 = strtoupper ( 'Client-ip' );
  22         } else {
  23                 // Subject to spoofing with headers like X_Forwarded_For
  24                 $set = $_SERVER;
  25                 $index = 'HTTP_X_FORWARDED_FOR';
  26                 $index2 = 'CLIENT-IP';
  27         }
  28
  29         #Try a couple of headers
  30         if( isset( $set[$index] ) ) {
  31                 return $set[$index];
  32         } else if( isset( $set[$index2] ) ) {
  33                 return $set[$index2];
  34         } else {
  35                 return null;
  36         }
  37 }
  38
  39 /**
  40  * Returns the browser/OS data from the request header
  41  * Note: headers are spoofable
  42  * @return string
  43  */
  44 function wfGetAgent() {
  45         if( function_exists( 'apache_request_headers' ) ) {
  46                 // More reliable than $_SERVER due to case and -/_ folding
  47                 $set = array ();
  48                 foreach ( apache_request_headers() as $tempName => $tempValue ) {
  49                         $set[ strtoupper( $tempName ) ] = $tempValue;
  50                 }
  51                 $index = strtoupper ( 'User-Agent' );
  52         } else {
  53                 // Subject to spoofing with headers like X_Forwarded_For
  54                 $set = $_SERVER;
  55                 $index = 'HTTP_USER_AGENT';
  56         }
  57         if( isset( $set[$index] ) ) {
  58                 return $set[$index];
  59         } else {
  60                 return '';
  61         }
  62 }
  63
  64 /**
  65  * Work out the IP address based on various globals
  66  * For trusted proxies, use the XFF client IP (first of the chain)
  67  * @return string
  68  */
  69 function wfGetIP() {
  70         global $wgIP;
  71
  72         # Return cached result
  73         if ( !empty( $wgIP ) ) {
  74                 return $wgIP;
  75         }
  76
  77         /* collect the originating ips */
  78         # Client connecting to this webserver
  79         if ( isset( $_SERVER['REMOTE_ADDR'] ) ) {
  80                 $ipchain = array( IP::canonicalize( $_SERVER['REMOTE_ADDR'] ) );
  81         } else {
  82                 # Running on CLI?
  83                 $ipchain = array( '127.0.0.1' );
  84         }
  85         $ip = $ipchain[0];
  86
  87         # Append XFF on to $ipchain
  88         $forwardedFor = wfGetForwardedFor();
  89         if ( isset( $forwardedFor ) ) {
  90                 $xff = array_map( 'trim', explode( ',', $forwardedFor ) );
  91                 $xff = array_reverse( $xff );
  92                 $ipchain = array_merge( $ipchain, $xff );
  93         }
  94
  95         # Step through XFF list and find the last address in the list which is a trusted server
  96         # Set $ip to the IP address given by that trusted server, unless the address is not sensible (e.g. private)
  97         foreach ( $ipchain as $i => $curIP ) {
  98                 $curIP = IP::canonicalize( $curIP );
  99                 if ( wfIsTrustedProxy( $curIP ) ) {
 100                         if ( isset( $ipchain[$i + 1] ) && IP::isPublic( $ipchain[$i + 1] ) ) {
 101                                 $ip = $ipchain[$i + 1];
 102                         }
 103                 } else {
 104                         break;
 105                 }
 106         }
 107
 108         wfDebug( "IP: $ip\n" );
 109         $wgIP = $ip;
 110         return $ip;
 111 }
 112
 113 /**
 114  * Checks if an IP is a trusted proxy providor
 115  * Useful to tell if X-Fowarded-For data is possibly bogus
 116  * Squid cache servers for the site and AOL are whitelisted
 117  * @param string $ip
 118  * @return bool
 119  */
 120 function wfIsTrustedProxy( $ip ) {
 121         global $wgSquidServers, $wgSquidServersNoPurge;
 122
 123         if ( in_array( $ip, $wgSquidServers ) ||
 124                 in_array( $ip, $wgSquidServersNoPurge ) ||
 125                 wfIsAOLProxy( $ip )
 126         ) {
 127                 $trusted = true;
 128         } else {
 129                 $trusted = false;
 130         }
 131         wfRunHooks( 'IsTrustedProxy', array( &$ip, &$trusted ) );
 132         return $trusted;
 133 }
 134
 135 /**
 136  * Forks processes to scan the originating IP for an open proxy server
 137  * MemCached can be used to skip IPs that have already been scanned
 138  */
 139 function wfProxyCheck() {
 140         global $wgBlockOpenProxies, $wgProxyPorts, $wgProxyScriptPath;
 141         global $wgMemc, $wgProxyMemcExpiry;
 142         global $wgProxyKey;
 143
 144         if ( !$wgBlockOpenProxies ) {
 145                 return;
 146         }
 147
 148         $ip = wfGetIP();
 149
 150         # Get MemCached key
 151         $mcKey = wfMemcKey( 'proxy', 'ip', $ip );
 152         $mcValue = $wgMemc->get( $mcKey );
 153         $skip = (bool)$mcValue;
 154
 155         # Fork the processes
 156         if ( !$skip ) {
 157                 $title = SpecialPage::getTitleFor( 'Blockme' );
 158                 $iphash = md5( $ip . $wgProxyKey );
 159                 $url = $title->getFullURL( 'ip='.$iphash );
 160
 161                 foreach ( $wgProxyPorts as $port ) {
 162                         $params = implode( ' ', array(
 163                                                 escapeshellarg( $wgProxyScriptPath ),
 164                                                 escapeshellarg( $ip ),
 165                                                 escapeshellarg( $port ),
 166                                                 escapeshellarg( $url )
 167                                                 ));
 168                         exec( "php $params &>/dev/null &" );
 169                 }
 170                 # Set MemCached key
 171                 $wgMemc->set( $mcKey, 1, $wgProxyMemcExpiry );
 172         }
 173 }
 174
 175 /**
 176  * Convert a network specification in CIDR notation to an integer network and a number of bits
 177  * @return array(string, int)
 178  */
 179 function wfParseCIDR( $range ) {
 180         return IP::parseCIDR( $range );
 181 }
 182
 183 /**
 184  * Check if an IP address is in the local proxy list
 185  * @return bool
 186  */
 187 function wfIsLocallyBlockedProxy( $ip ) {
 188         global $wgProxyList;
 189         $fname = 'wfIsLocallyBlockedProxy';
 190
 191         if ( !$wgProxyList ) {
 192                 return false;
 193         }
 194         wfProfileIn( $fname );
 195
 196         if ( !is_array( $wgProxyList ) ) {
 197                 # Load from the specified file
 198                 $wgProxyList = array_map( 'trim', file( $wgProxyList ) );
 199         }
 200
 201         if ( !is_array( $wgProxyList ) ) {
 202                 $ret = false;
 203         } elseif ( array_search( $ip, $wgProxyList ) !== false ) {
 204                 $ret = true;
 205         } elseif ( array_key_exists( $ip, $wgProxyList ) ) {
 206                 # Old-style flipped proxy list
 207                 $ret = true;
 208         } else {
 209                 $ret = false;
 210         }
 211         wfProfileOut( $fname );
 212         return $ret;
 213 }
 214
 215 /**
 216  * TODO: move this list to the database in a global IP info table incorporating
 217  * trusted ISP proxies, blocked IP addresses and open proxies.
 218  * @return bool
 219  */
 220 function wfIsAOLProxy( $ip ) {
 221         # From http://webmaster.info.aol.com/proxyinfo.html
 222         $ranges = array(
 223                 '64.12.96.0/19',
 224                 '149.174.160.0/20',
 225                 '152.163.240.0/21',
 226                 '152.163.248.0/22',
 227                 '152.163.252.0/23',
 228                 '152.163.96.0/22',
 229                 '152.163.100.0/23',
 230                 '195.93.32.0/22',
 231                 '195.93.48.0/22',
 232                 '195.93.64.0/19',
 233                 '195.93.96.0/19',
 234                 '195.93.16.0/20',
 235                 '198.81.0.0/22',
 236                 '198.81.16.0/20',
 237                 '198.81.8.0/23',
 238                 '202.67.64.128/25',
 239                 '205.188.192.0/20',
 240                 '205.188.208.0/23',
 241                 '205.188.112.0/20',
 242                 '205.188.146.144/30',
 243                 '207.200.112.0/21',
 244         );
 245
 246         static $parsedRanges;
 247         if ( is_null( $parsedRanges ) ) {
 248                 $parsedRanges = array();
 249                 foreach ( $ranges as $range ) {
 250                         $parsedRanges[] =  IP::parseRange( $range );
 251                 }
 252         }
 253
 254         $hex = IP::toHex( $ip );
 255         foreach ( $parsedRanges as $range ) {
 256                 if ( $hex >= $range[0] && $hex <= $range[1] ) {
 257                         return true;
 258                 }
 259         }
 260         return false;
 261 }