3 * Functions and constants to play with IP addresses and ranges
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.
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.
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
21 * @author Antoine Musso "<hashar at free dot fr>"
26 // Some regex definition to "play" with IP address and IP address ranges
28 // An IPv4 address is made of 4 bytes from x00 to xFF which is d0 to d255
29 define( 'RE_IP_BYTE', '(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|0?[0-9]?[0-9])' );
30 define( 'RE_IP_ADD', RE_IP_BYTE
. '\.' . RE_IP_BYTE
. '\.' . RE_IP_BYTE
. '\.' . RE_IP_BYTE
);
31 // An IPv4 range is an IP address and a prefix (d1 to d32)
32 define( 'RE_IP_PREFIX', '(3[0-2]|[12]?\d)' );
33 define( 'RE_IP_RANGE', RE_IP_ADD
. '\/' . RE_IP_PREFIX
);
35 // An IPv6 address is made up of 8 words (each x0000 to xFFFF).
36 // However, the "::" abbreviation can be used on consecutive x0000 words.
37 define( 'RE_IPV6_WORD', '([0-9A-Fa-f]{1,4})' );
38 define( 'RE_IPV6_PREFIX', '(12[0-8]|1[01][0-9]|[1-9]?\d)' );
39 define( 'RE_IPV6_ADD',
40 '(?:' . // starts with "::" (including "::")
41 ':(?::|(?::' . RE_IPV6_WORD
. '){1,7})' .
42 '|' . // ends with "::" (except "::")
43 RE_IPV6_WORD
. '(?::' . RE_IPV6_WORD
. '){0,6}::' .
44 '|' . // contains one "::" in the middle (the ^ makes the test fail if none found)
45 RE_IPV6_WORD
. '(?::((?(-1)|:))?' . RE_IPV6_WORD
. '){1,6}(?(-2)|^)' .
46 '|' . // contains no "::"
47 RE_IPV6_WORD
. '(?::' . RE_IPV6_WORD
. '){7}' .
50 // An IPv6 range is an IP address and a prefix (d1 to d128)
51 define( 'RE_IPV6_RANGE', RE_IPV6_ADD
. '\/' . RE_IPV6_PREFIX
);
52 // For IPv6 canonicalization (NOT for strict validation; these are quite lax!)
53 define( 'RE_IPV6_GAP', ':(?:0+:)*(?::(?:0+:)*)?' );
54 define( 'RE_IPV6_V4_PREFIX', '0*' . RE_IPV6_GAP
. '(?:ffff:)?' );
56 // This might be useful for regexps used elsewhere, matches any IPv4 or IPv6 address or network
57 define( 'IP_ADDRESS_STRING',
59 RE_IP_ADD
. '(?:\/' . RE_IP_PREFIX
. ')?' . // IPv4
61 RE_IPV6_ADD
. '(?:\/' . RE_IPV6_PREFIX
. ')?' . // IPv6
66 * A collection of public static functions to play with IP address
72 * Determine if a string is as valid IP address or network (CIDR prefix).
73 * SIIT IPv4-translated addresses are rejected.
74 * @note canonicalize() tries to convert translated addresses to IPv4.
76 * @param string $ip Possible IP address
79 public static function isIPAddress( $ip ) {
80 return (bool)preg_match( '/^' . IP_ADDRESS_STRING
. '$/', $ip );
84 * Given a string, determine if it as valid IP in IPv6 only.
85 * @note Unlike isValid(), this looks for networks too.
87 * @param string $ip Possible IP address
90 public static function isIPv6( $ip ) {
91 return (bool)preg_match( '/^' . RE_IPV6_ADD
. '(?:\/' . RE_IPV6_PREFIX
. ')?$/', $ip );
95 * Given a string, determine if it as valid IP in IPv4 only.
96 * @note Unlike isValid(), this looks for networks too.
98 * @param string $ip Possible IP address
101 public static function isIPv4( $ip ) {
102 return (bool)preg_match( '/^' . RE_IP_ADD
. '(?:\/' . RE_IP_PREFIX
. ')?$/', $ip );
106 * Validate an IP address. Ranges are NOT considered valid.
107 * SIIT IPv4-translated addresses are rejected.
108 * @note canonicalize() tries to convert translated addresses to IPv4.
111 * @return bool True if it is valid
113 public static function isValid( $ip ) {
114 return ( preg_match( '/^' . RE_IP_ADD
. '$/', $ip )
115 ||
preg_match( '/^' . RE_IPV6_ADD
. '$/', $ip ) );
119 * Validate an IP range (valid address with a valid CIDR prefix).
120 * SIIT IPv4-translated addresses are rejected.
121 * @note canonicalize() tries to convert translated addresses to IPv4.
123 * @deprecated since 1.30. Use the equivalent IP::isValidRange().
124 * @param string $ipRange
125 * @return bool True if it is valid
127 public static function isValidBlock( $ipRange ) {
128 wfDeprecated( __METHOD__
, '1.30' );
129 return self
::isValidRange( $ipRange );
133 * Validate an IP range (valid address with a valid CIDR prefix).
134 * SIIT IPv4-translated addresses are rejected.
135 * @note canonicalize() tries to convert translated addresses to IPv4.
137 * @param string $ipRange
138 * @return bool True if it is valid
141 public static function isValidRange( $ipRange ) {
142 return ( preg_match( '/^' . RE_IPV6_RANGE
. '$/', $ipRange )
143 ||
preg_match( '/^' . RE_IP_RANGE
. '$/', $ipRange ) );
147 * Convert an IP into a verbose, uppercase, normalized form.
148 * Both IPv4 and IPv6 addresses are trimmed. Additionally,
149 * IPv6 addresses in octet notation are expanded to 8 words;
150 * IPv4 addresses have leading zeros, in each octet, removed.
152 * @param string $ip IP address in quad or octet form (CIDR or not).
155 public static function sanitizeIP( $ip ) {
160 /* If not an IP, just return trimmed value, since sanitizeIP() is called
161 * in a number of contexts where usernames are supplied as input.
163 if ( !self
::isIPAddress( $ip ) ) {
166 if ( self
::isIPv4( $ip ) ) {
167 // Remove leading 0's from octet representation of IPv4 address
168 $ip = preg_replace( '!(?:^|(?<=\.))0+(?=[1-9]|0[./]|0$)!', '', $ip );
171 // Remove any whitespaces, convert to upper case
172 $ip = strtoupper( $ip );
173 // Expand zero abbreviations
174 $abbrevPos = strpos( $ip, '::' );
175 if ( $abbrevPos !== false ) {
176 // We know this is valid IPv6. Find the last index of the
177 // address before any CIDR number (e.g. "a:b:c::/24").
178 $CIDRStart = strpos( $ip, "/" );
179 $addressEnd = ( $CIDRStart !== false )
182 // If the '::' is at the beginning...
183 if ( $abbrevPos == 0 ) {
185 $extra = ( $ip == '::' ) ?
'0' : ''; // for the address '::'
186 $pad = 9; // 7+2 (due to '::')
187 // If the '::' is at the end...
188 } elseif ( $abbrevPos == ( $addressEnd - 1 ) ) {
191 $pad = 9; // 7+2 (due to '::')
192 // If the '::' is in the middle...
196 $pad = 8; // 6+2 (due to '::')
198 $ip = str_replace( '::',
199 str_repeat( $repeat, $pad - substr_count( $ip, ':' ) ) . $extra,
203 // Remove leading zeros from each bloc as needed
204 $ip = preg_replace( '/(^|:)0+(' . RE_IPV6_WORD
. ')/', '$1$2', $ip );
210 * Prettify an IP for display to end users.
211 * This will make it more compact and lower-case.
216 public static function prettifyIP( $ip ) {
217 $ip = self
::sanitizeIP( $ip ); // normalize (removes '::')
218 if ( self
::isIPv6( $ip ) ) {
219 // Split IP into an address and a CIDR
220 if ( strpos( $ip, '/' ) !== false ) {
221 list( $ip, $cidr ) = explode( '/', $ip, 2 );
223 list( $ip, $cidr ) = [ $ip, '' ];
225 // Get the largest slice of words with multiple zeros
227 $longest = $longestPos = false;
229 '!(?:^|:)0(?::0)+(?:$|:)!', $ip, $m, PREG_OFFSET_CAPTURE
, $offset
231 list( $match, $pos ) = $m[0]; // full match
232 if ( strlen( $match ) > strlen( $longest ) ) {
236 $offset = ( $pos +
strlen( $match ) ); // advance
238 if ( $longest !== false ) {
239 // Replace this portion of the string with the '::' abbreviation
240 $ip = substr_replace( $ip, '::', $longestPos, strlen( $longest ) );
242 // Add any CIDR back on
243 if ( $cidr !== '' ) {
244 $ip = "{$ip}/{$cidr}";
246 // Convert to lower case to make it more readable
247 $ip = strtolower( $ip );
254 * Given a host/port string, like one might find in the host part of a URL
255 * per RFC 2732, split the hostname part and the port part and return an
256 * array with an element for each. If there is no port part, the array will
257 * have false in place of the port. If the string was invalid in some way,
260 * This was easy with IPv4 and was generally done in an ad-hoc way, but
261 * with IPv6 it's somewhat more complicated due to the need to parse the
262 * square brackets and colons.
264 * A bare IPv6 address is accepted despite the lack of square brackets.
266 * @param string $both The string with the host and port
267 * @return array|false Array normally, false on certain failures
269 public static function splitHostAndPort( $both ) {
270 if ( substr( $both, 0, 1 ) === '[' ) {
271 if ( preg_match( '/^\[(' . RE_IPV6_ADD
. ')\](?::(?P<port>\d+))?$/', $both, $m ) ) {
272 if ( isset( $m['port'] ) ) {
273 return [ $m[1], intval( $m['port'] ) ];
275 return [ $m[1], false ];
278 // Square bracket found but no IPv6
282 $numColons = substr_count( $both, ':' );
283 if ( $numColons >= 2 ) {
284 // Is it a bare IPv6 address?
285 if ( preg_match( '/^' . RE_IPV6_ADD
. '$/', $both ) ) {
286 return [ $both, false ];
288 // Not valid IPv6, but too many colons for anything else
292 if ( $numColons >= 1 ) {
294 $bits = explode( ':', $both );
295 if ( preg_match( '/^\d+/', $bits[1] ) ) {
296 return [ $bits[0], intval( $bits[1] ) ];
304 return [ $both, false ];
308 * Given a host name and a port, combine them into host/port string like
309 * you might find in a URL. If the host contains a colon, wrap it in square
310 * brackets like in RFC 2732. If the port matches the default port, omit
311 * the port specification
313 * @param string $host
315 * @param bool|int $defaultPort
318 public static function combineHostAndPort( $host, $port, $defaultPort = false ) {
319 if ( strpos( $host, ':' ) !== false ) {
322 if ( $defaultPort !== false && $port == $defaultPort ) {
325 return "$host:$port";
330 * Convert an IPv4 or IPv6 hexadecimal representation back to readable format
332 * @param string $hex Number, with "v6-" prefix if it is IPv6
333 * @return string Quad-dotted (IPv4) or octet notation (IPv6)
335 public static function formatHex( $hex ) {
336 if ( substr( $hex, 0, 3 ) == 'v6-' ) { // IPv6
337 return self
::hexToOctet( substr( $hex, 3 ) );
339 return self
::hexToQuad( $hex );
344 * Converts a hexadecimal number to an IPv6 address in octet notation
346 * @param string $ip_hex Pure hex (no v6- prefix)
347 * @return string (of format a:b:c:d:e:f:g:h)
349 public static function hexToOctet( $ip_hex ) {
350 // Pad hex to 32 chars (128 bits)
351 $ip_hex = str_pad( strtoupper( $ip_hex ), 32, '0', STR_PAD_LEFT
);
352 // Separate into 8 words
353 $ip_oct = substr( $ip_hex, 0, 4 );
354 for ( $n = 1; $n < 8; $n++
) {
355 $ip_oct .= ':' . substr( $ip_hex, 4 * $n, 4 );
358 $ip_oct = preg_replace( '/(^|:)0+(' . RE_IPV6_WORD
. ')/', '$1$2', $ip_oct );
364 * Converts a hexadecimal number to an IPv4 address in quad-dotted notation
366 * @param string $ip_hex Pure hex
367 * @return string (of format a.b.c.d)
369 public static function hexToQuad( $ip_hex ) {
370 // Pad hex to 8 chars (32 bits)
371 $ip_hex = str_pad( strtoupper( $ip_hex ), 8, '0', STR_PAD_LEFT
);
372 // Separate into four quads
374 for ( $i = 0; $i < 4; $i++
) {
378 $s .= base_convert( substr( $ip_hex, $i * 2, 2 ), 16, 10 );
385 * Determine if an IP address really is an IP address, and if it is public,
386 * i.e. not RFC 1918 or similar
391 public static function isPublic( $ip ) {
392 static $privateSet = null;
393 if ( !$privateSet ) {
394 $privateSet = new IPSet( [
395 '10.0.0.0/8', # RFC 1918 (private)
396 '172.16.0.0/12', # RFC 1918 (private)
397 '192.168.0.0/16', # RFC 1918 (private)
398 '0.0.0.0/8', # this network
399 '127.0.0.0/8', # loopback
400 'fc00::/7', # RFC 4193 (local)
401 '0:0:0:0:0:0:0:1', # loopback
402 '169.254.0.0/16', # link-local
403 'fe80::/10', # link-local
406 return !$privateSet->match( $ip );
410 * Return a zero-padded upper case hexadecimal representation of an IP address.
412 * Hexadecimal addresses are used because they can easily be extended to
413 * IPv6 support. To separate the ranges, the return value from this
414 * function for an IPv6 address will be prefixed with "v6-", a non-
415 * hexadecimal string which sorts after the IPv4 addresses.
417 * @param string $ip Quad dotted/octet IP address.
418 * @return string|bool False on failure
420 public static function toHex( $ip ) {
421 if ( self
::isIPv6( $ip ) ) {
422 $n = 'v6-' . self
::IPv6ToRawHex( $ip );
423 } elseif ( self
::isIPv4( $ip ) ) {
424 // T62035/T97897: An IP with leading 0's fails in ip2long sometimes (e.g. *.08),
425 // also double/triple 0 needs to be changed to just a single 0 for ip2long.
426 $ip = self
::sanitizeIP( $ip );
430 # On 32-bit platforms (and on Windows), 2^32 does not fit into an int,
431 # so $n becomes a float. We convert it to string instead.
432 if ( is_float( $n ) ) {
436 if ( $n !== false ) {
437 # Floating points can handle the conversion; faster than Wikimedia\base_convert()
438 $n = strtoupper( str_pad( base_convert( $n, 10, 16 ), 8, '0', STR_PAD_LEFT
) );
448 * Given an IPv6 address in octet notation, returns a pure hex string.
450 * @param string $ip Octet ipv6 IP address.
451 * @return string|bool Pure hex (uppercase); false on failure
453 private static function IPv6ToRawHex( $ip ) {
454 $ip = self
::sanitizeIP( $ip );
459 foreach ( explode( ':', $ip ) as $v ) {
460 $r_ip .= str_pad( $v, 4, 0, STR_PAD_LEFT
);
467 * Convert a network specification in CIDR notation
468 * to an integer network and a number of bits
470 * @param string $range IP with CIDR prefix
471 * @return array [int or string, int]
473 public static function parseCIDR( $range ) {
474 if ( self
::isIPv6( $range ) ) {
475 return self
::parseCIDR6( $range );
477 $parts = explode( '/', $range, 2 );
478 if ( count( $parts ) != 2 ) {
479 return [ false, false ];
481 list( $network, $bits ) = $parts;
482 $network = ip2long( $network );
483 if ( $network !== false && is_numeric( $bits ) && $bits >= 0 && $bits <= 32 ) {
487 $network &= ~
( ( 1 << ( 32 - $bits ) ) - 1 );
489 # Convert to unsigned
490 if ( $network < 0 ) {
498 return [ $network, $bits ];
502 * Given a string range in a number of formats,
503 * return the start and end of the range in hexadecimal.
507 * 1.2.3.4 - 1.2.3.5 Explicit range
510 * 2001:0db8:85a3::7344/96 CIDR
511 * 2001:0db8:85a3::7344 - 2001:0db8:85a3::7344 Explicit range
512 * 2001:0db8:85a3::7344 Single IP
513 * @param string $range IP range
514 * @return array [ string, string ]
516 public static function parseRange( $range ) {
518 if ( strpos( $range, '/' ) !== false ) {
519 if ( self
::isIPv6( $range ) ) {
520 return self
::parseRange6( $range );
522 list( $network, $bits ) = self
::parseCIDR( $range );
523 if ( $network === false ) {
524 $start = $end = false;
526 $start = sprintf( '%08X', $network );
527 $end = sprintf( '%08X', $network +
2 ** ( 32 - $bits ) - 1 );
530 } elseif ( strpos( $range, '-' ) !== false ) {
531 list( $start, $end ) = array_map( 'trim', explode( '-', $range, 2 ) );
532 if ( self
::isIPv6( $start ) && self
::isIPv6( $end ) ) {
533 return self
::parseRange6( $range );
535 if ( self
::isIPv4( $start ) && self
::isIPv4( $end ) ) {
536 $start = self
::toHex( $start );
537 $end = self
::toHex( $end );
538 if ( $start > $end ) {
539 $start = $end = false;
542 $start = $end = false;
546 $start = $end = self
::toHex( $range );
548 if ( $start === false ||
$end === false ) {
549 return [ false, false ];
551 return [ $start, $end ];
556 * Convert a network specification in IPv6 CIDR notation to an
557 * integer network and a number of bits
559 * @param string $range
561 * @return array [string, int]
563 private static function parseCIDR6( $range ) {
564 # Explode into <expanded IP,range>
565 $parts = explode( '/', self
::sanitizeIP( $range ), 2 );
566 if ( count( $parts ) != 2 ) {
567 return [ false, false ];
569 list( $network, $bits ) = $parts;
570 $network = self
::IPv6ToRawHex( $network );
571 if ( $network !== false && is_numeric( $bits ) && $bits >= 0 && $bits <= 128 ) {
575 # Native 32 bit functions WONT work here!!!
576 # Convert to a padded binary number
577 $network = Wikimedia\base_convert
( $network, 16, 2, 128 );
578 # Truncate the last (128-$bits) bits and replace them with zeros
579 $network = str_pad( substr( $network, 0, $bits ), 128, 0, STR_PAD_RIGHT
);
580 # Convert back to an integer
581 $network = Wikimedia\base_convert
( $network, 2, 10 );
588 return [ $network, (int)$bits ];
592 * Given a string range in a number of formats, return the
593 * start and end of the range in hexadecimal. For IPv6.
596 * 2001:0db8:85a3::7344/96 CIDR
597 * 2001:0db8:85a3::7344 - 2001:0db8:85a3::7344 Explicit range
598 * 2001:0db8:85a3::7344/96 Single IP
600 * @param string $range
602 * @return array [string, string]
604 private static function parseRange6( $range ) {
606 $range = self
::sanitizeIP( $range );
608 if ( strpos( $range, '/' ) !== false ) {
609 list( $network, $bits ) = self
::parseCIDR6( $range );
610 if ( $network === false ) {
611 $start = $end = false;
613 $start = Wikimedia\base_convert
( $network, 10, 16, 32, false );
614 # Turn network to binary (again)
615 $end = Wikimedia\base_convert
( $network, 10, 2, 128 );
616 # Truncate the last (128-$bits) bits and replace them with ones
617 $end = str_pad( substr( $end, 0, $bits ), 128, 1, STR_PAD_RIGHT
);
619 $end = Wikimedia\base_convert
( $end, 2, 16, 32, false );
620 # see toHex() comment
621 $start = "v6-$start";
624 // Explicit range notation...
625 } elseif ( strpos( $range, '-' ) !== false ) {
626 list( $start, $end ) = array_map( 'trim', explode( '-', $range, 2 ) );
627 $start = self
::toHex( $start );
628 $end = self
::toHex( $end );
629 if ( $start > $end ) {
630 $start = $end = false;
634 $start = $end = self
::toHex( $range );
636 if ( $start === false ||
$end === false ) {
637 return [ false, false ];
639 return [ $start, $end ];
644 * Determine if a given IPv4/IPv6 address is in a given CIDR network
646 * @param string $addr The address to check against the given range.
647 * @param string $range The range to check the given address against.
648 * @return bool Whether or not the given address is in the given range.
650 * @note This can return unexpected results for invalid arguments!
651 * Make sure you pass a valid IP address and IP range.
653 public static function isInRange( $addr, $range ) {
654 $hexIP = self
::toHex( $addr );
655 list( $start, $end ) = self
::parseRange( $range );
657 return ( strcmp( $hexIP, $start ) >= 0 &&
658 strcmp( $hexIP, $end ) <= 0 );
662 * Determines if an IP address is a list of CIDR a.b.c.d/n ranges.
666 * @param string $ip the IP to check
667 * @param array $ranges the IP ranges, each element a range
669 * @return bool true if the specified adress belongs to the specified range; otherwise, false.
671 public static function isInRanges( $ip, $ranges ) {
672 foreach ( $ranges as $range ) {
673 if ( self
::isInRange( $ip, $range ) ) {
681 * Convert some unusual representations of IPv4 addresses to their
682 * canonical dotted quad representation.
684 * This currently only checks a few IPV4-to-IPv6 related cases. More
685 * unusual representations may be added later.
687 * @param string $addr Something that might be an IP address
688 * @return string|null Valid dotted quad IPv4 address or null
690 public static function canonicalize( $addr ) {
691 // remove zone info (T37738)
692 $addr = preg_replace( '/\%.*/', '', $addr );
694 if ( self
::isValid( $addr ) ) {
697 // Turn mapped addresses from ::ce:ffff:1.2.3.4 to 1.2.3.4
698 if ( strpos( $addr, ':' ) !== false && strpos( $addr, '.' ) !== false ) {
699 $addr = substr( $addr, strrpos( $addr, ':' ) +
1 );
700 if ( self
::isIPv4( $addr ) ) {
704 // IPv6 loopback address
706 if ( preg_match( '/^0*' . RE_IPV6_GAP
. '1$/', $addr, $m ) ) {
709 // IPv4-mapped and IPv4-compatible IPv6 addresses
710 if ( preg_match( '/^' . RE_IPV6_V4_PREFIX
. '(' . RE_IP_ADD
. ')$/i', $addr, $m ) ) {
713 if ( preg_match( '/^' . RE_IPV6_V4_PREFIX
. RE_IPV6_WORD
.
714 ':' . RE_IPV6_WORD
. '$/i', $addr, $m )
716 return long2ip( ( hexdec( $m[1] ) << 16 ) +
hexdec( $m[2] ) );
719 return null; // give up
723 * Gets rid of unneeded numbers in quad-dotted/octet IP strings
724 * For example, 127.111.113.151/24 -> 127.111.113.0/24
725 * @param string $range IP address to normalize
728 public static function sanitizeRange( $range ) {
729 list( /*...*/, $bits ) = self
::parseCIDR( $range );
730 list( $start, /*...*/ ) = self
::parseRange( $range );
731 $start = self
::formatHex( $start );
732 if ( $bits === false ) {
733 return $start; // wasn't actually a range
736 return "$start/$bits";
740 * Returns the subnet of a given IP
743 * @return string|false
745 public static function getSubnet( $ip ) {
748 if ( self
::isIPv6( $ip ) ) {
749 $parts = self
::parseRange( "$ip/64" );
751 } elseif ( preg_match( '/^(\d+\.\d+\.\d+)\.\d+$/', $ip, $matches ) ) {
753 $subnet = $matches[1];