* Followed-up r76267:
[lhc/web/wiklou.git] / includes / IP.php
1 <?php
2 /**
3 * Functions and constants to play with IP addresses and ranges
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 Ashar Voultoiz <hashar at free dot fr>
22 */
23
24 // Some regex definition to "play" with IP address and IP address blocks
25
26 // An IP is made of 4 bytes from x00 to xFF which is d0 to d255
27 define( 'RE_IP_BYTE', '(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|0?[0-9]?[0-9])' );
28 define( 'RE_IP_ADD' , RE_IP_BYTE . '\.' . RE_IP_BYTE . '\.' . RE_IP_BYTE . '\.' . RE_IP_BYTE );
29 // An IPv4 block is an IP address and a prefix (d1 to d32)
30 define( 'RE_IP_PREFIX', '(3[0-2]|[12]?\d)' );
31 define( 'RE_IP_BLOCK', RE_IP_ADD . '\/' . RE_IP_PREFIX );
32 // For IPv6 canonicalization (NOT for strict validation; these are quite lax!)
33 define( 'RE_IPV6_WORD', '([0-9A-Fa-f]{1,4})' );
34 define( 'RE_IPV6_GAP', ':(?:0+:)*(?::(?:0+:)*)?' );
35 define( 'RE_IPV6_V4_PREFIX', '0*' . RE_IPV6_GAP . '(?:ffff:)?' );
36 // An IPv6 block is an IP address and a prefix (d1 to d128)
37 define( 'RE_IPV6_PREFIX', '(12[0-8]|1[01][0-9]|[1-9]?\d)');
38 // An IPv6 address is made up of 8 octets. However, the "::" abbreviations can be used.
39 define( 'RE_IPV6_ADD',
40 '(' . // starts with "::" (includes the address "::")
41 '(::|:(:' . RE_IPV6_WORD . '){1,7})' .
42 '|' . // ends with "::" (not including the address "::")
43 RE_IPV6_WORD . '(:' . RE_IPV6_WORD . '){0,6}::' .
44 '|' . // has no "::"
45 RE_IPV6_WORD . '(:' . RE_IPV6_WORD . '){7}' .
46 '|' . // contains one "::" in the middle ("^" check always fails if no "::" found)
47 RE_IPV6_WORD . '(:(?P<abbr>(?(abbr)|:))?' . RE_IPV6_WORD . '){1,6}(?(abbr)|^)' .
48 ')'
49 );
50 define( 'RE_IPV6_BLOCK', RE_IPV6_ADD . '\/' . RE_IPV6_PREFIX );
51 // This might be useful for regexps used elsewhere, matches any IPv6 or IPv6 address or network
52 define( 'IP_ADDRESS_STRING',
53 '(?:' .
54 RE_IP_ADD . '(\/' . RE_IP_PREFIX . '|)' . // IPv4
55 '|' .
56 RE_IPV6_ADD . '(\/' . RE_IPV6_PREFIX . '|)' . // IPv6
57 ')'
58 );
59
60 /**
61 * A collection of public static functions to play with IP address
62 * and IP blocks.
63 */
64 class IP {
65 /**
66 * Given a string, determine if it as valid IP.
67 * Note: Unlike isValid(), this looks for networks too.
68 * @param $ip string possible IP address
69 * @return string
70 */
71 public static function isIPAddress( $ip ) {
72 if ( !$ip ) {
73 return false;
74 }
75 return preg_match( '/^' . IP_ADDRESS_STRING . '$/', $ip );
76 }
77
78 /**
79 * Given a string, determine if it as valid IP in IPv6 only.
80 * Note: Unlike isValid(), this looks for networks too.
81 * @param $ip string possible IP address
82 * @return string
83 */
84 public static function isIPv6( $ip ) {
85 if ( !$ip ) {
86 return false;
87 }
88 return preg_match( '/^' . RE_IPV6_ADD . '(\/' . RE_IPV6_PREFIX . '|)$/', $ip );
89 }
90
91 /**
92 * Given a string, determine if it as valid IP in IPv4 only.
93 * Note: Unlike isValid(), this looks for networks too.
94 * @param $ip string possible IP address
95 * @return string
96 */
97 public static function isIPv4( $ip ) {
98 if ( !$ip ) {
99 return false;
100 }
101 return preg_match( '/^' . RE_IP_ADD . '(\/' . RE_IP_PREFIX . '|)$/', $ip );
102 }
103
104 /**
105 * Given an IP address in dotted-quad notation, returns an IPv6 octet.
106 * See http://www.answers.com/topic/ipv4-compatible-address
107 * IPs with the first 92 bits as zeros are reserved from IPv6
108 * @param $ip quad-dotted IP address.
109 * @return string
110 */
111 public static function IPv4toIPv6( $ip ) {
112 if ( !$ip ) {
113 return null;
114 }
115 // Convert only if needed
116 if ( self::isIPv6( $ip ) ) {
117 return $ip;
118 }
119 // IPv4 CIDRs
120 if ( strpos( $ip, '/' ) !== false ) {
121 $parts = explode( '/', $ip, 2 );
122 if ( count( $parts ) != 2 ) {
123 return false;
124 }
125 list( $network, $bits ) = $parts;
126 $network = self::toUnsigned( $network );
127 if ( $network !== false && is_numeric( $bits ) && $bits >= 0 && $bits <= 32 ) {
128 $bits += 96;
129 return self::toOctet( $network ) . "/$bits";
130 } else {
131 return false;
132 }
133 }
134 return self::toOctet( self::toUnsigned( $ip ) );
135 }
136
137 /**
138 * Given an IPv6 address in octet notation, returns an unsigned integer.
139 * @param $ip octet ipv6 IP address.
140 * @return string
141 */
142 public static function toUnsigned6( $ip ) {
143 if ( !$ip ) {
144 return null;
145 }
146 $ip = explode( ':', self::sanitizeIP( $ip ) );
147 $r_ip = '';
148 foreach ( $ip as $v ) {
149 $r_ip .= str_pad( $v, 4, 0, STR_PAD_LEFT );
150 }
151 $r_ip = wfBaseConvert( $r_ip, 16, 10 );
152 return $r_ip;
153 }
154
155 /**
156 * Given an IPv6 address in octet notation, returns the expanded octet.
157 * IPv4 IPs will be trimmed, thats it...
158 * @param $ip string IP address in quad or octet form (CIDR or not).
159 * @return string
160 */
161 public static function sanitizeIP( $ip ) {
162 $ip = trim( $ip );
163 if ( $ip === '' ) {
164 return null;
165 }
166 if ( self::isIPv4( $ip ) || !self::isIPv6( $ip ) ) {
167 return $ip; // nothing else to do for IPv4 addresses or invalid ones
168 }
169 // Remove any whitespaces, convert to upper case
170 $ip = strtoupper( $ip );
171 // Expand zero abbreviations
172 $abbrevPos = strpos( $ip, '::' );
173 if ( $abbrevPos !== false ) {
174 // We know this is valid IPv6. Find the last index of the
175 // address before any CIDR number (e.g. "a:b:c::/24").
176 $CIDRStart = strpos( $ip, "/" );
177 $addressEnd = ( $CIDRStart !== false )
178 ? $CIDRStart - 1
179 : strlen( $ip ) - 1;
180 // If the '::' is at the beginning...
181 if( $abbrevPos == 0 ) {
182 $repeat = '0:';
183 $extra = ( $ip == '::' ) ? '0' : ''; // for the address '::'
184 $pad = 9; // 7+2 (due to '::')
185 // If the '::' is at the end...
186 } elseif( $abbrevPos == ( $addressEnd - 1 ) ) {
187 $repeat = ':0';
188 $extra = '';
189 $pad = 9; // 7+2 (due to '::')
190 // If the '::' is in the middle...
191 } else {
192 $repeat = ':0';
193 $extra = ':';
194 $pad = 8; // 6+2 (due to '::')
195 }
196 $ip = str_replace( '::',
197 str_repeat( $repeat, $pad - substr_count( $ip, ':' ) ) . $extra,
198 $ip
199 );
200 }
201 // Remove leading zereos from each bloc as needed
202 $ip = preg_replace( '/(^|:)0+' . RE_IPV6_WORD . '/', '$1$2', $ip );
203 return $ip;
204 }
205
206 /**
207 * Given an unsigned integer, returns an IPv6 address in octet notation
208 * @param $ip_int integer IP address.
209 * @return string
210 */
211 public static function toOctet( $ip_int ) {
212 // Convert to padded uppercase hex
213 $ip_hex = wfBaseConvert( $ip_int, 10, 16, 32, false );
214 // Separate into 8 octets
215 $ip_oct = substr( $ip_hex, 0, 4 );
216 for ( $n = 1; $n < 8; $n++ ) {
217 $ip_oct .= ':' . substr( $ip_hex, 4 * $n, 4 );
218 }
219 // NO leading zeroes
220 $ip_oct = preg_replace( '/(^|:)0+' . RE_IPV6_WORD . '/', '$1$2', $ip_oct );
221 return $ip_oct;
222 }
223
224 /**
225 * Convert an IPv4 or IPv6 hexadecimal representation back to readable format
226 * @param $hex string number, with "v6-" prefix if it is IPv6
227 * @return string quad-dotted (IPv4) or octet notation (IPv6)
228 */
229 public static function formatHex( $hex ) {
230 if ( substr( $hex, 0, 3 ) == 'v6-' ) { // IPv6
231 return self::hexToOctet( substr( $hex, 3 ) );
232 } else { // IPv4
233 return self::hexToQuad( $hex );
234 }
235 }
236
237 /**
238 * Converts a hexadecimal number to an IPv6 address in octet notation
239 * @param $ip_hex string hex IP
240 * @return string (of format a:b:c:d:e:f:g:h)
241 */
242 public static function hexToOctet( $ip_hex ) {
243 // Convert to padded uppercase hex
244 $ip_hex = str_pad( strtoupper( $ip_hex ), 32, '0' );
245 // Separate into 8 octets
246 $ip_oct = substr( $ip_hex, 0, 4 );
247 for ( $n = 1; $n < 8; $n++ ) {
248 $ip_oct .= ':' . substr( $ip_hex, 4 * $n, 4 );
249 }
250 // NO leading zeroes
251 $ip_oct = preg_replace( '/(^|:)0+' . RE_IPV6_WORD . '/', '$1$2', $ip_oct );
252 return $ip_oct;
253 }
254
255 /**
256 * Converts a hexadecimal number to an IPv4 address in quad-dotted notation
257 * @param $ip string Hex IP
258 * @return string (of format a.b.c.d)
259 */
260 public static function hexToQuad( $ip ) {
261 $s = '';
262 for ( $i = 0; $i < 4; $i++ ) {
263 if ( $s !== '' ) {
264 $s .= '.';
265 }
266 $s .= base_convert( substr( $ip, $i * 2, 2 ), 16, 10 );
267 }
268 return $s;
269 }
270
271 /**
272 * Convert a network specification in IPv6 CIDR notation to an
273 * integer network and a number of bits
274 * @return array(string, int)
275 */
276 public static function parseCIDR6( $range ) {
277 # Explode into <expanded IP,range>
278 $parts = explode( '/', IP::sanitizeIP( $range ), 2 );
279 if ( count( $parts ) != 2 ) {
280 return array( false, false );
281 }
282 list( $network, $bits ) = $parts;
283 $network = self::toUnsigned6( $network );
284 if ( $network !== false && is_numeric( $bits ) && $bits >= 0 && $bits <= 128 ) {
285 if ( $bits == 0 ) {
286 $network = "0";
287 } else {
288 # Native 32 bit functions WONT work here!!!
289 # Convert to a padded binary number
290 $network = wfBaseConvert( $network, 10, 2, 128 );
291 # Truncate the last (128-$bits) bits and replace them with zeros
292 $network = str_pad( substr( $network, 0, $bits ), 128, 0, STR_PAD_RIGHT );
293 # Convert back to an integer
294 $network = wfBaseConvert( $network, 2, 10 );
295 }
296 } else {
297 $network = false;
298 $bits = false;
299 }
300 return array( $network, (int)$bits );
301 }
302
303 /**
304 * Given a string range in a number of formats, return the start and end of
305 * the range in hexadecimal. For IPv6.
306 *
307 * Formats are:
308 * 2001:0db8:85a3::7344/96 CIDR
309 * 2001:0db8:85a3::7344 - 2001:0db8:85a3::7344 Explicit range
310 * 2001:0db8:85a3::7344/96 Single IP
311 * @return array(string, int)
312 */
313 public static function parseRange6( $range ) {
314 # Expand any IPv6 IP
315 $range = IP::sanitizeIP( $range );
316 // CIDR notation...
317 if ( strpos( $range, '/' ) !== false ) {
318 list( $network, $bits ) = self::parseCIDR6( $range );
319 if ( $network === false ) {
320 $start = $end = false;
321 } else {
322 $start = wfBaseConvert( $network, 10, 16, 32, false );
323 # Turn network to binary (again)
324 $end = wfBaseConvert( $network, 10, 2, 128 );
325 # Truncate the last (128-$bits) bits and replace them with ones
326 $end = str_pad( substr( $end, 0, $bits ), 128, 1, STR_PAD_RIGHT );
327 # Convert to hex
328 $end = wfBaseConvert( $end, 2, 16, 32, false );
329 # see toHex() comment
330 $start = "v6-$start";
331 $end = "v6-$end";
332 }
333 // Explicit range notation...
334 } elseif ( strpos( $range, '-' ) !== false ) {
335 list( $start, $end ) = array_map( 'trim', explode( '-', $range, 2 ) );
336 $start = self::toUnsigned6( $start );
337 $end = self::toUnsigned6( $end );
338 if ( $start > $end ) {
339 $start = $end = false;
340 } else {
341 $start = wfBaseConvert( $start, 10, 16, 32, false );
342 $end = wfBaseConvert( $end, 10, 16, 32, false );
343 }
344 # see toHex() comment
345 $start = "v6-$start";
346 $end = "v6-$end";
347 } else {
348 # Single IP
349 $start = $end = self::toHex( $range );
350 }
351 if ( $start === false || $end === false ) {
352 return array( false, false );
353 } else {
354 return array( $start, $end );
355 }
356 }
357
358 /**
359 * Validate an IP address.
360 * @return boolean True if it is valid.
361 */
362 public static function isValid( $ip ) {
363 return ( preg_match( '/^' . RE_IP_ADD . '$/', $ip ) || preg_match( '/^' . RE_IPV6_ADD . '$/', $ip ) );
364 }
365
366 /**
367 * Validate an IP Block.
368 * @return boolean True if it is valid.
369 */
370 public static function isValidBlock( $ipblock ) {
371 return ( count( self::toArray( $ipblock ) ) == 1 + 5 );
372 }
373
374 /**
375 * Determine if an IP address really is an IP address, and if it is public,
376 * i.e. not RFC 1918 or similar
377 * Comes from ProxyTools.php
378 */
379 public static function isPublic( $ip ) {
380 $n = self::toUnsigned( $ip );
381 if ( !$n ) {
382 return false;
383 }
384
385 // ip2long accepts incomplete addresses, as well as some addresses
386 // followed by garbage characters. Check that it's really valid.
387 if( $ip != long2ip( $n ) ) {
388 return false;
389 }
390
391 static $privateRanges = false;
392 if ( !$privateRanges ) {
393 $privateRanges = array(
394 array( '10.0.0.0', '10.255.255.255' ), # RFC 1918 (private)
395 array( '172.16.0.0', '172.31.255.255' ), # "
396 array( '192.168.0.0', '192.168.255.255' ), # "
397 array( '0.0.0.0', '0.255.255.255' ), # this network
398 array( '127.0.0.0', '127.255.255.255' ), # loopback
399 );
400 }
401
402 foreach ( $privateRanges as $r ) {
403 $start = self::toUnsigned( $r[0] );
404 $end = self::toUnsigned( $r[1] );
405 if ( $n >= $start && $n <= $end ) {
406 return false;
407 }
408 }
409 return true;
410 }
411
412 /**
413 * Split out an IP block as an array of 4 bytes and a mask,
414 * return false if it can't be determined
415 *
416 * @param $ipblock string A quad dotted/octet IP address
417 * @return array
418 */
419 public static function toArray( $ipblock ) {
420 $matches = array();
421 if( preg_match( '/^' . RE_IP_ADD . '(?:\/(?:' . RE_IP_PREFIX . '))?' . '$/', $ipblock, $matches ) ) {
422 return $matches;
423 } elseif ( preg_match( '/^' . RE_IPV6_ADD . '(?:\/(?:' . RE_IPV6_PREFIX . '))?' . '$/', $ipblock, $matches ) ) {
424 return $matches;
425 } else {
426 return false;
427 }
428 }
429
430 /**
431 * Return a zero-padded hexadecimal representation of an IP address.
432 *
433 * Hexadecimal addresses are used because they can easily be extended to
434 * IPv6 support. To separate the ranges, the return value from this
435 * function for an IPv6 address will be prefixed with "v6-", a non-
436 * hexadecimal string which sorts after the IPv4 addresses.
437 *
438 * @param $ip Quad dotted/octet IP address.
439 * @return hexidecimal
440 */
441 public static function toHex( $ip ) {
442 $n = self::toUnsigned( $ip );
443 if ( $n !== false ) {
444 $n = self::isIPv6( $ip )
445 ? 'v6-' . wfBaseConvert( $n, 10, 16, 32, false )
446 : wfBaseConvert( $n, 10, 16, 8, false );
447 }
448 return $n;
449 }
450
451 /**
452 * Given an IP address in dotted-quad/octet notation, returns an unsigned integer.
453 * Like ip2long() except that it actually works and has a consistent error return value.
454 * Comes from ProxyTools.php
455 * @param $ip Quad dotted IP address.
456 * @return integer
457 */
458 public static function toUnsigned( $ip ) {
459 // Use IPv6 functions if needed
460 if ( self::isIPv6( $ip ) ) {
461 return self::toUnsigned6( $ip );
462 }
463 if ( $ip == '255.255.255.255' ) {
464 $n = -1;
465 } else {
466 $n = ip2long( $ip );
467 if ( $n == -1 || $n === false ) { # Return value on error depends on PHP version
468 $n = false;
469 }
470 }
471 if ( $n < 0 ) {
472 $n += pow( 2, 32 );
473 }
474 return $n;
475 }
476
477 /**
478 * Convert a dotted-quad IP to a signed integer
479 * Returns false on failure
480 */
481 public static function toSigned( $ip ) {
482 if ( $ip == '255.255.255.255' ) {
483 $n = -1;
484 } else {
485 $n = ip2long( $ip );
486 if ( $n == -1 ) {
487 $n = false;
488 }
489 }
490 return $n;
491 }
492
493 /**
494 * Convert a network specification in CIDR notation to an integer network and a number of bits
495 * @return array(int, int)
496 */
497 public static function parseCIDR( $range ) {
498 $parts = explode( '/', $range, 2 );
499 if ( count( $parts ) != 2 ) {
500 return array( false, false );
501 }
502 list( $network, $bits ) = $parts;
503 $network = self::toSigned( $network );
504 if ( $network !== false && is_numeric( $bits ) && $bits >= 0 && $bits <= 32 ) {
505 if ( $bits == 0 ) {
506 $network = 0;
507 } else {
508 $network &= ~( ( 1 << ( 32 - $bits ) ) - 1);
509 }
510 # Convert to unsigned
511 if ( $network < 0 ) {
512 $network += pow( 2, 32 );
513 }
514 } else {
515 $network = false;
516 $bits = false;
517 }
518 return array( $network, $bits );
519 }
520
521 /**
522 * Given a string range in a number of formats, return the start and end of
523 * the range in hexadecimal.
524 *
525 * Formats are:
526 * 1.2.3.4/24 CIDR
527 * 1.2.3.4 - 1.2.3.5 Explicit range
528 * 1.2.3.4 Single IP
529 *
530 * 2001:0db8:85a3::7344/96 CIDR
531 * 2001:0db8:85a3::7344 - 2001:0db8:85a3::7344 Explicit range
532 * 2001:0db8:85a3::7344 Single IP
533 * @return array(string, int)
534 */
535 public static function parseRange( $range ) {
536 // Use IPv6 functions if needed
537 if ( self::isIPv6( $range ) ) {
538 return self::parseRange6( $range );
539 }
540 if ( strpos( $range, '/' ) !== false ) {
541 # CIDR
542 list( $network, $bits ) = self::parseCIDR( $range );
543 if ( $network === false ) {
544 $start = $end = false;
545 } else {
546 $start = sprintf( '%08X', $network );
547 $end = sprintf( '%08X', $network + pow( 2, ( 32 - $bits ) ) - 1 );
548 }
549 } elseif ( strpos( $range, '-' ) !== false ) {
550 # Explicit range
551 list( $start, $end ) = array_map( 'trim', explode( '-', $range, 2 ) );
552 if( self::isIPAddress( $start ) && self::isIPAddress( $end ) ) {
553 $start = self::toUnsigned( $start );
554 $end = self::toUnsigned( $end );
555 if ( $start > $end ) {
556 $start = $end = false;
557 } else {
558 $start = sprintf( '%08X', $start );
559 $end = sprintf( '%08X', $end );
560 }
561 } else {
562 $start = $end = false;
563 }
564 } else {
565 # Single IP
566 $start = $end = self::toHex( $range );
567 }
568 if ( $start === false || $end === false ) {
569 return array( false, false );
570 } else {
571 return array( $start, $end );
572 }
573 }
574
575 /**
576 * Determine if a given IPv4/IPv6 address is in a given CIDR network
577 * @param $addr The address to check against the given range.
578 * @param $range The range to check the given address against.
579 * @return bool Whether or not the given address is in the given range.
580 */
581 public static function isInRange( $addr, $range ) {
582 // Convert to IPv6 if needed
583 $hexIP = self::toHex( $addr );
584 list( $start, $end ) = self::parseRange( $range );
585 return ( strcmp( $hexIP, $start ) >= 0 &&
586 strcmp( $hexIP, $end ) <= 0 );
587 }
588
589 /**
590 * Convert some unusual representations of IPv4 addresses to their
591 * canonical dotted quad representation.
592 *
593 * This currently only checks a few IPV4-to-IPv6 related cases. More
594 * unusual representations may be added later.
595 *
596 * @param $addr something that might be an IP address
597 * @return valid dotted quad IPv4 address or null
598 */
599 public static function canonicalize( $addr ) {
600 if ( self::isValid( $addr ) ) {
601 return $addr;
602 }
603
604 // Turn mapped addresses from ::ce:ffff:1.2.3.4 to 1.2.3.4
605 if ( strpos( $addr, ':' ) !== false && strpos( $addr, '.' ) !== false ) {
606 $addr = substr( $addr, strrpos( $addr, ':' ) + 1 );
607 if( self::isIPv4( $addr ) ) {
608 return $addr;
609 }
610 }
611
612 // IPv6 loopback address
613 $m = array();
614 if ( preg_match( '/^0*' . RE_IPV6_GAP . '1$/', $addr, $m ) ) {
615 return '127.0.0.1';
616 }
617
618 // IPv4-mapped and IPv4-compatible IPv6 addresses
619 if ( preg_match( '/^' . RE_IPV6_V4_PREFIX . '(' . RE_IP_ADD . ')$/i', $addr, $m ) ) {
620 return $m[1];
621 }
622 if ( preg_match( '/^' . RE_IPV6_V4_PREFIX . RE_IPV6_WORD . ':' . RE_IPV6_WORD . '$/i', $addr, $m ) ) {
623 return long2ip( ( hexdec( $m[1] ) << 16 ) + hexdec( $m[2] ) );
624 }
625
626 return null; // give up
627 }
628 }