IP.php

Go to the documentation of this file.

<?php
// Some regex definition to "play" with IP address and IP address blocks

// An IPv4 address is made of 4 bytes from x00 to xFF which is d0 to d255
define( 'RE_IP_BYTE', '(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|0?[0-9]?[0-9])' );
define( 'RE_IP_ADD', RE_IP_BYTE . '\.' . RE_IP_BYTE . '\.' . RE_IP_BYTE . '\.' . RE_IP_BYTE );
// An IPv4 block is an IP address and a prefix (d1 to d32)
define( 'RE_IP_PREFIX', '(3[0-2]|[12]?\d)' );
define( 'RE_IP_BLOCK', RE_IP_ADD . '\/' . RE_IP_PREFIX );

// An IPv6 address is made up of 8 words (each x0000 to xFFFF).
// However, the "::" abbreviation can be used on consecutive x0000 words.
define( 'RE_IPV6_WORD', '([0-9A-Fa-f]{1,4})' );
define( 'RE_IPV6_PREFIX', '(12[0-8]|1[01][0-9]|[1-9]?\d)' );
define( 'RE_IPV6_ADD',
    '(?:' . // starts with "::" (including "::")
        ':(?::|(?::' . RE_IPV6_WORD . '){1,7})' .
    '|' . // ends with "::" (except "::")
        RE_IPV6_WORD . '(?::' . RE_IPV6_WORD . '){0,6}::' .
    '|' . // contains one "::" in the middle (the ^ makes the test fail if none found)
        RE_IPV6_WORD . '(?::((?(-1)|:))?' . RE_IPV6_WORD . '){1,6}(?(-2)|^)' .
    '|' . // contains no "::"
        RE_IPV6_WORD . '(?::' . RE_IPV6_WORD . '){7}' .
    ')'
);
// An IPv6 block is an IP address and a prefix (d1 to d128)
define( 'RE_IPV6_BLOCK', RE_IPV6_ADD . '\/' . RE_IPV6_PREFIX );
// For IPv6 canonicalization (NOT for strict validation; these are quite lax!)
define( 'RE_IPV6_GAP', ':(?:0+:)*(?::(?:0+:)*)?' );
define( 'RE_IPV6_V4_PREFIX', '0*' . RE_IPV6_GAP . '(?:ffff:)?' );

// This might be useful for regexps used elsewhere, matches any IPv6 or IPv6 address or network
define( 'IP_ADDRESS_STRING',
    '(?:' .
        RE_IP_ADD . '(?:\/' . RE_IP_PREFIX . ')?' . // IPv4
    '|' .
        RE_IPV6_ADD . '(?:\/' . RE_IPV6_PREFIX . ')?' . // IPv6
    ')'
);

class IP {
    private static $proxyIpSet = null;

    public static function isIPAddress( $ip ) {
        return (bool)preg_match( '/^' . IP_ADDRESS_STRING . '$/', $ip );
    }

    public static function isIPv6( $ip ) {
        return (bool)preg_match( '/^' . RE_IPV6_ADD . '(?:\/' . RE_IPV6_PREFIX . ')?$/', $ip );
    }

    public static function isIPv4( $ip ) {
        return (bool)preg_match( '/^' . RE_IP_ADD . '(?:\/' . RE_IP_PREFIX . ')?$/', $ip );
    }

    public static function isValid( $ip ) {
        return ( preg_match( '/^' . RE_IP_ADD . '$/', $ip )
            || preg_match( '/^' . RE_IPV6_ADD . '$/', $ip ) );
    }

    public static function isValidBlock( $ipblock ) {
        return ( preg_match( '/^' . RE_IPV6_BLOCK . '$/', $ipblock )
            || preg_match( '/^' . RE_IP_BLOCK . '$/', $ipblock ) );
    }

    public static function sanitizeIP( $ip ) {
        $ip = trim( $ip );
        if ( $ip === '' ) {
            return null;
        }
        if ( self::isIPv4( $ip ) || !self::isIPv6( $ip ) ) {
            return $ip; // nothing else to do for IPv4 addresses or invalid ones
        }
        // Remove any whitespaces, convert to upper case
        $ip = strtoupper( $ip );
        // Expand zero abbreviations
        $abbrevPos = strpos( $ip, '::' );
        if ( $abbrevPos !== false ) {
            // We know this is valid IPv6. Find the last index of the
            // address before any CIDR number (e.g. "a:b:c::/24").
            $CIDRStart = strpos( $ip, "/" );
            $addressEnd = ( $CIDRStart !== false )
                ? $CIDRStart - 1
                : strlen( $ip ) - 1;
            // If the '::' is at the beginning...
            if ( $abbrevPos == 0 ) {
                $repeat = '0:';
                $extra = ( $ip == '::' ) ? '0' : ''; // for the address '::'
                $pad = 9; // 7+2 (due to '::')
            // If the '::' is at the end...
            } elseif ( $abbrevPos == ( $addressEnd - 1 ) ) {
                $repeat = ':0';
                $extra = '';
                $pad = 9; // 7+2 (due to '::')
            // If the '::' is in the middle...
            } else {
                $repeat = ':0';
                $extra = ':';
                $pad = 8; // 6+2 (due to '::')
            }
            $ip = str_replace( '::',
                str_repeat( $repeat, $pad - substr_count( $ip, ':' ) ) . $extra,
                $ip
            );
        }
        // Remove leading zeros from each bloc as needed
        $ip = preg_replace( '/(^|:)0+(' . RE_IPV6_WORD . ')/', '$1$2', $ip );

        return $ip;
    }

    public static function prettifyIP( $ip ) {
        $ip = self::sanitizeIP( $ip ); // normalize (removes '::')
        if ( self::isIPv6( $ip ) ) {
            // Split IP into an address and a CIDR
            if ( strpos( $ip, '/' ) !== false ) {
                list( $ip, $cidr ) = explode( '/', $ip, 2 );
            } else {
                list( $ip, $cidr ) = array( $ip, '' );
            }
            // Get the largest slice of words with multiple zeros
            $offset = 0;
            $longest = $longestPos = false;
            while ( preg_match(
                '!(?:^|:)0(?::0)+(?:$|:)!', $ip, $m, PREG_OFFSET_CAPTURE, $offset
            ) ) {
                list( $match, $pos ) = $m[0]; // full match
                if ( strlen( $match ) > strlen( $longest ) ) {
                    $longest = $match;
                    $longestPos = $pos;
                }
                $offset = ( $pos + strlen( $match ) ); // advance
            }
            if ( $longest !== false ) {
                // Replace this portion of the string with the '::' abbreviation
                $ip = substr_replace( $ip, '::', $longestPos, strlen( $longest ) );
            }
            // Add any CIDR back on
            if ( $cidr !== '' ) {
                $ip = "{$ip}/{$cidr}";
            }
            // Convert to lower case to make it more readable
            $ip = strtolower( $ip );
        }

        return $ip;
    }

    public static function splitHostAndPort( $both ) {
        if ( substr( $both, 0, 1 ) === '[' ) {
            if ( preg_match( '/^\[(' . RE_IPV6_ADD . ')\](?::(?P<port>\d+))?$/', $both, $m ) ) {
                if ( isset( $m['port'] ) ) {
                    return array( $m[1], intval( $m['port'] ) );
                } else {
                    return array( $m[1], false );
                }
            } else {
                // Square bracket found but no IPv6
                return false;
            }
        }
        $numColons = substr_count( $both, ':' );
        if ( $numColons >= 2 ) {
            // Is it a bare IPv6 address?
            if ( preg_match( '/^' . RE_IPV6_ADD . '$/', $both ) ) {
                return array( $both, false );
            } else {
                // Not valid IPv6, but too many colons for anything else
                return false;
            }
        }
        if ( $numColons >= 1 ) {
            // Host:port?
            $bits = explode( ':', $both );
            if ( preg_match( '/^\d+/', $bits[1] ) ) {
                return array( $bits[0], intval( $bits[1] ) );
            } else {
                // Not a valid port
                return false;
            }
        }

        // Plain hostname
        return array( $both, false );
    }

    public static function combineHostAndPort( $host, $port, $defaultPort = false ) {
        if ( strpos( $host, ':' ) !== false ) {
            $host = "[$host]";
        }
        if ( $defaultPort !== false && $port == $defaultPort ) {
            return $host;
        } else {
            return "$host:$port";
        }
    }

    public static function formatHex( $hex ) {
        if ( substr( $hex, 0, 3 ) == 'v6-' ) { // IPv6
            return self::hexToOctet( substr( $hex, 3 ) );
        } else { // IPv4
            return self::hexToQuad( $hex );
        }
    }

    public static function hexToOctet( $ip_hex ) {
        // Pad hex to 32 chars (128 bits)
        $ip_hex = str_pad( strtoupper( $ip_hex ), 32, '0', STR_PAD_LEFT );
        // Separate into 8 words
        $ip_oct = substr( $ip_hex, 0, 4 );
        for ( $n = 1; $n < 8; $n++ ) {
            $ip_oct .= ':' . substr( $ip_hex, 4 * $n, 4 );
        }
        // NO leading zeroes
        $ip_oct = preg_replace( '/(^|:)0+(' . RE_IPV6_WORD . ')/', '$1$2', $ip_oct );

        return $ip_oct;
    }

    public static function hexToQuad( $ip_hex ) {
        // Pad hex to 8 chars (32 bits)
        $ip_hex = str_pad( strtoupper( $ip_hex ), 8, '0', STR_PAD_LEFT );
        // Separate into four quads
        $s = '';
        for ( $i = 0; $i < 4; $i++ ) {
            if ( $s !== '' ) {
                $s .= '.';
            }
            $s .= base_convert( substr( $ip_hex, $i * 2, 2 ), 16, 10 );
        }

        return $s;
    }

    public static function isPublic( $ip ) {
        static $privateSet = null;
        if ( !$privateSet ) {
            $privateSet = new IPSet( array(
                '10.0.0.0/8', # RFC 1918 (private)
                '172.16.0.0/12', # RFC 1918 (private)
                '192.168.0.0/16', # RFC 1918 (private)
                '0.0.0.0/8', # this network
                '127.0.0.0/8', # loopback
                'fc00::/7', # RFC 4193 (local)
                '0:0:0:0:0:0:0:1', # loopback
            ) );
        }
        return !$privateSet->match( $ip );
    }

    public static function toHex( $ip ) {
        if ( self::isIPv6( $ip ) ) {
            $n = 'v6-' . self::IPv6ToRawHex( $ip );
        } elseif ( self::isIPv4( $ip ) ) {
            // Bug 60035: an IP with leading 0's fails in ip2long sometimes (e.g. *.08)
            $ip = preg_replace( '/(?<=\.)0+(?=[1-9])/', '', $ip );
            $n = ip2long( $ip );
            if ( $n < 0 ) {
                $n += pow( 2, 32 );
                # On 32-bit platforms (and on Windows), 2^32 does not fit into an int,
                # so $n becomes a float. We convert it to string instead.
                if ( is_float( $n ) ) {
                    $n = (string)$n;
                }
            }
            if ( $n !== false ) {
                # Floating points can handle the conversion; faster than wfBaseConvert()
                $n = strtoupper( str_pad( base_convert( $n, 10, 16 ), 8, '0', STR_PAD_LEFT ) );
            }
        } else {
            $n = false;
        }

        return $n;
    }

    private static function IPv6ToRawHex( $ip ) {
        $ip = self::sanitizeIP( $ip );
        if ( !$ip ) {
            return false;
        }
        $r_ip = '';
        foreach ( explode( ':', $ip ) as $v ) {
            $r_ip .= str_pad( $v, 4, 0, STR_PAD_LEFT );
        }

        return $r_ip;
    }

    public static function parseCIDR( $range ) {
        if ( self::isIPv6( $range ) ) {
            return self::parseCIDR6( $range );
        }
        $parts = explode( '/', $range, 2 );
        if ( count( $parts ) != 2 ) {
            return array( false, false );
        }
        list( $network, $bits ) = $parts;
        $network = ip2long( $network );
        if ( $network !== false && is_numeric( $bits ) && $bits >= 0 && $bits <= 32 ) {
            if ( $bits == 0 ) {
                $network = 0;
            } else {
                $network &= ~( ( 1 << ( 32 - $bits ) ) - 1 );
            }
            # Convert to unsigned
            if ( $network < 0 ) {
                $network += pow( 2, 32 );
            }
        } else {
            $network = false;
            $bits = false;
        }

        return array( $network, $bits );
    }

    public static function parseRange( $range ) {
        // CIDR notation
        if ( strpos( $range, '/' ) !== false ) {
            if ( self::isIPv6( $range ) ) {
                return self::parseRange6( $range );
            }
            list( $network, $bits ) = self::parseCIDR( $range );
            if ( $network === false ) {
                $start = $end = false;
            } else {
                $start = sprintf( '%08X', $network );
                $end = sprintf( '%08X', $network + pow( 2, ( 32 - $bits ) ) - 1 );
            }
        // Explicit range
        } elseif ( strpos( $range, '-' ) !== false ) {
            list( $start, $end ) = array_map( 'trim', explode( '-', $range, 2 ) );
            if ( self::isIPv6( $start ) && self::isIPv6( $end ) ) {
                return self::parseRange6( $range );
            }
            if ( self::isIPv4( $start ) && self::isIPv4( $end ) ) {
                $start = self::toHex( $start );
                $end = self::toHex( $end );
                if ( $start > $end ) {
                    $start = $end = false;
                }
            } else {
                $start = $end = false;
            }
        } else {
            # Single IP
            $start = $end = self::toHex( $range );
        }
        if ( $start === false || $end === false ) {
            return array( false, false );
        } else {
            return array( $start, $end );
        }
    }

    private static function parseCIDR6( $range ) {
        # Explode into <expanded IP,range>
        $parts = explode( '/', IP::sanitizeIP( $range ), 2 );
        if ( count( $parts ) != 2 ) {
            return array( false, false );
        }
        list( $network, $bits ) = $parts;
        $network = self::IPv6ToRawHex( $network );
        if ( $network !== false && is_numeric( $bits ) && $bits >= 0 && $bits <= 128 ) {
            if ( $bits == 0 ) {
                $network = "0";
            } else {
                # Native 32 bit functions WONT work here!!!
                # Convert to a padded binary number
                $network = wfBaseConvert( $network, 16, 2, 128 );
                # Truncate the last (128-$bits) bits and replace them with zeros
                $network = str_pad( substr( $network, 0, $bits ), 128, 0, STR_PAD_RIGHT );
                # Convert back to an integer
                $network = wfBaseConvert( $network, 2, 10 );
            }
        } else {
            $network = false;
            $bits = false;
        }

        return array( $network, (int)$bits );
    }

    private static function parseRange6( $range ) {
        # Expand any IPv6 IP
        $range = IP::sanitizeIP( $range );
        // CIDR notation...
        if ( strpos( $range, '/' ) !== false ) {
            list( $network, $bits ) = self::parseCIDR6( $range );
            if ( $network === false ) {
                $start = $end = false;
            } else {
                $start = wfBaseConvert( $network, 10, 16, 32, false );
                # Turn network to binary (again)
                $end = wfBaseConvert( $network, 10, 2, 128 );
                # Truncate the last (128-$bits) bits and replace them with ones
                $end = str_pad( substr( $end, 0, $bits ), 128, 1, STR_PAD_RIGHT );
                # Convert to hex
                $end = wfBaseConvert( $end, 2, 16, 32, false );
                # see toHex() comment
                $start = "v6-$start";
                $end = "v6-$end";
            }
        // Explicit range notation...
        } elseif ( strpos( $range, '-' ) !== false ) {
            list( $start, $end ) = array_map( 'trim', explode( '-', $range, 2 ) );
            $start = self::toHex( $start );
            $end = self::toHex( $end );
            if ( $start > $end ) {
                $start = $end = false;
            }
        } else {
            # Single IP
            $start = $end = self::toHex( $range );
        }
        if ( $start === false || $end === false ) {
            return array( false, false );
        } else {
            return array( $start, $end );
        }
    }

    public static function isInRange( $addr, $range ) {
        $hexIP = self::toHex( $addr );
        list( $start, $end ) = self::parseRange( $range );

        return ( strcmp( $hexIP, $start ) >= 0 &&
            strcmp( $hexIP, $end ) <= 0 );
    }

    public static function canonicalize( $addr ) {
        // remove zone info (bug 35738)
        $addr = preg_replace( '/\%.*/', '', $addr );

        if ( self::isValid( $addr ) ) {
            return $addr;
        }
        // Turn mapped addresses from ::ce:ffff:1.2.3.4 to 1.2.3.4
        if ( strpos( $addr, ':' ) !== false && strpos( $addr, '.' ) !== false ) {
            $addr = substr( $addr, strrpos( $addr, ':' ) + 1 );
            if ( self::isIPv4( $addr ) ) {
                return $addr;
            }
        }
        // IPv6 loopback address
        $m = array();
        if ( preg_match( '/^0*' . RE_IPV6_GAP . '1$/', $addr, $m ) ) {
            return '127.0.0.1';
        }
        // IPv4-mapped and IPv4-compatible IPv6 addresses
        if ( preg_match( '/^' . RE_IPV6_V4_PREFIX . '(' . RE_IP_ADD . ')$/i', $addr, $m ) ) {
            return $m[1];
        }
        if ( preg_match( '/^' . RE_IPV6_V4_PREFIX . RE_IPV6_WORD .
            ':' . RE_IPV6_WORD . '$/i', $addr, $m )
        ) {
            return long2ip( ( hexdec( $m[1] ) << 16 ) + hexdec( $m[2] ) );
        }

        return null; // give up
    }

    public static function sanitizeRange( $range ) {
        list( /*...*/, $bits ) = self::parseCIDR( $range );
        list( $start, /*...*/ ) = self::parseRange( $range );
        $start = self::formatHex( $start );
        if ( $bits === false ) {
            return $start; // wasn't actually a range
        }

        return "$start/$bits";
    }

    public static function isTrustedProxy( $ip ) {
        $trusted = self::isConfiguredProxy( $ip );
        wfRunHooks( 'IsTrustedProxy', array( &$ip, &$trusted ) );
        return $trusted;
    }

    public static function isConfiguredProxy( $ip ) {
        global $wgSquidServers, $wgSquidServersNoPurge;

        wfProfileIn( __METHOD__ );
        // Quick check of known singular proxy servers
        $trusted = in_array( $ip, $wgSquidServers );

        // Check against addresses and CIDR nets in the NoPurge list
        if ( !$trusted ) {
            if ( !self::$proxyIpSet ) {
                self::$proxyIpSet = new IPSet( $wgSquidServersNoPurge );
            }
            $trusted = self::$proxyIpSet->match( $ip );
        }
        wfProfileOut( __METHOD__ );

        return $trusted;
    }

    public static function clearCaches() {
        self::$proxyIpSet = null;
    }
}