1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971
|
# Copyright 2012-2014, Damian Johnson and The Tor Project
# See LICENSE for licensing information
"""
Parsing for Tor server descriptors, which contains the infrequently changing
information about a Tor relay (contact information, exit policy, public keys,
etc). This information is provided from a few sources...
* control port via 'GETINFO desc/\*' queries
* the 'cached-descriptors' file in tor's data directory
* tor metrics, at https://metrics.torproject.org/data.html
* directory authorities and mirrors via their DirPort
**Module Overview:**
::
ServerDescriptor - Tor server descriptor.
|- RelayDescriptor - Server descriptor for a relay.
|
|- BridgeDescriptor - Scrubbed server descriptor for a bridge.
| |- is_scrubbed - checks if our content has been properly scrubbed
| +- get_scrubbing_issues - description of issues with our scrubbing
|
|- digest - calculates the upper-case hex digest value for our content
|- get_unrecognized_lines - lines with unrecognized content
|- get_annotations - dictionary of content prior to the descriptor entry
+- get_annotation_lines - lines that provided the annotations
"""
import base64
import codecs
import datetime
import hashlib
import re
import stem.descriptor.extrainfo_descriptor
import stem.exit_policy
import stem.prereq
import stem.util.connection
import stem.util.str_tools
import stem.util.tor_tools
import stem.version
from stem.util import log
from stem.descriptor import (
PGP_BLOCK_END,
Descriptor,
_get_bytes_field,
_get_descriptor_components,
_read_until_keywords,
)
try:
# added in python 3.2
from functools import lru_cache
except ImportError:
from stem.util.lru_cache import lru_cache
# relay descriptors must have exactly one of the following
REQUIRED_FIELDS = (
'router',
'bandwidth',
'published',
'onion-key',
'signing-key',
'router-signature',
)
# optional entries that can appear at most once
SINGLE_FIELDS = (
'platform',
'fingerprint',
'hibernating',
'uptime',
'contact',
'read-history',
'write-history',
'eventdns',
'family',
'caches-extra-info',
'extra-info-digest',
'hidden-service-dir',
'protocols',
'allow-single-hop-exits',
'ntor-onion-key',
)
DEFAULT_IPV6_EXIT_POLICY = stem.exit_policy.MicroExitPolicy('reject 1-65535')
REJECT_ALL_POLICY = stem.exit_policy.ExitPolicy('reject *:*')
def _parse_file(descriptor_file, is_bridge = False, validate = True, **kwargs):
"""
Iterates over the server descriptors in a file.
:param file descriptor_file: file with descriptor content
:param bool is_bridge: parses the file as being a bridge descriptor
:param bool validate: checks the validity of the descriptor's content if
**True**, skips these checks otherwise
:param dict kwargs: additional arguments for the descriptor constructor
:returns: iterator for ServerDescriptor instances in the file
:raises:
* **ValueError** if the contents is malformed and validate is True
* **IOError** if the file can't be read
"""
# Handler for relay descriptors
#
# Cached descriptors consist of annotations followed by the descriptor
# itself. For instance...
#
# @downloaded-at 2012-03-14 16:31:05
# @source "145.53.65.130"
# router caerSidi 71.35.143.157 9001 0 0
# platform Tor 0.2.1.30 on Linux x86_64
# <rest of the descriptor content>
# router-signature
# -----BEGIN SIGNATURE-----
# <signature for the above descriptor>
# -----END SIGNATURE-----
#
# Metrics descriptor files are the same, but lack any annotations. The
# following simply does the following...
#
# - parse as annotations until we get to 'router'
# - parse as descriptor content until we get to 'router-signature' followed
# by the end of the signature block
# - construct a descriptor and provide it back to the caller
#
# Any annotations after the last server descriptor is ignored (never provided
# to the caller).
while True:
annotations = _read_until_keywords('router', descriptor_file)
descriptor_content = _read_until_keywords('router-signature', descriptor_file)
# we've reached the 'router-signature', now include the pgp style block
block_end_prefix = PGP_BLOCK_END.split(' ', 1)[0]
descriptor_content += _read_until_keywords(block_end_prefix, descriptor_file, True)
if descriptor_content:
if descriptor_content[0].startswith(b'@type'):
descriptor_content = descriptor_content[1:]
# strip newlines from annotations
annotations = map(bytes.strip, annotations)
descriptor_text = bytes.join(b'', descriptor_content)
if is_bridge:
yield BridgeDescriptor(descriptor_text, validate, annotations, **kwargs)
else:
yield RelayDescriptor(descriptor_text, validate, annotations, **kwargs)
else:
if validate and annotations:
orphaned_annotations = stem.util.str_tools._to_unicode(b'\n'.join(annotations))
raise ValueError('Content conform to being a server descriptor:\n%s' % orphaned_annotations)
break # done parsing descriptors
class ServerDescriptor(Descriptor):
"""
Common parent for server descriptors.
:var str nickname: **\*** relay's nickname
:var str fingerprint: identity key fingerprint
:var datetime published: **\*** time in UTC when this descriptor was made
:var str address: **\*** IPv4 address of the relay
:var int or_port: **\*** port used for relaying
:var int socks_port: **\*** port used as client (deprecated, always **None**)
:var int dir_port: **\*** port used for descriptor mirroring
:var bytes platform: line with operating system and tor version
:var stem.version.Version tor_version: version of tor
:var str operating_system: operating system
:var int uptime: uptime when published in seconds
:var bytes contact: contact information
:var stem.exit_policy.ExitPolicy exit_policy: **\*** stated exit policy
:var stem.exit_policy.MicroExitPolicy exit_policy_v6: **\*** exit policy for IPv6
:var set family: **\*** nicknames or fingerprints of declared family
:var int average_bandwidth: **\*** average rate it's willing to relay in bytes/s
:var int burst_bandwidth: **\*** burst rate it's willing to relay in bytes/s
:var int observed_bandwidth: **\*** estimated capacity based on usage in bytes/s
:var list link_protocols: link protocols supported by the relay
:var list circuit_protocols: circuit protocols supported by the relay
:var bool hibernating: **\*** hibernating when published
:var bool allow_single_hop_exits: **\*** flag if single hop exiting is allowed
:var bool extra_info_cache: **\*** flag if a mirror for extra-info documents
:var str extra_info_digest: upper-case hex encoded digest of our extra-info document
:var bool eventdns: flag for evdns backend (deprecated, always unset)
:var list or_addresses: **\*** alternative for our address/or_port
attributes, each entry is a tuple of the form (address (**str**), port
(**int**), is_ipv6 (**bool**))
Deprecated, moved to extra-info descriptor...
:var datetime read_history_end: end of the sampling interval
:var int read_history_interval: seconds per interval
:var list read_history_values: bytes read during each interval
:var datetime write_history_end: end of the sampling interval
:var int write_history_interval: seconds per interval
:var list write_history_values: bytes written during each interval
**\*** attribute is either required when we're parsed with validation or has
a default value, others are left as **None** if undefined
"""
def __init__(self, raw_contents, validate = True, annotations = None):
"""
Server descriptor constructor, created from an individual relay's
descriptor content (as provided by 'GETINFO desc/*', cached descriptors,
and metrics).
By default this validates the descriptor's content as it's parsed. This
validation can be disables to either improve performance or be accepting of
malformed data.
:param str raw_contents: descriptor content provided by the relay
:param bool validate: checks the validity of the descriptor's content if
**True**, skips these checks otherwise
:param list annotations: lines that appeared prior to the descriptor
:raises: **ValueError** if the contents is malformed and validate is True
"""
super(ServerDescriptor, self).__init__(raw_contents)
# Only a few things can be arbitrary bytes according to the dir-spec, so
# parsing them separately.
self.platform = _get_bytes_field('platform', raw_contents)
self.contact = _get_bytes_field('contact', raw_contents)
raw_contents = stem.util.str_tools._to_unicode(raw_contents)
self.nickname = None
self.fingerprint = None
self.published = None
self.address = None
self.or_port = None
self.socks_port = None
self.dir_port = None
self.tor_version = None
self.operating_system = None
self.uptime = None
self.exit_policy = None
self.exit_policy_v6 = DEFAULT_IPV6_EXIT_POLICY
self.family = set()
self.average_bandwidth = None
self.burst_bandwidth = None
self.observed_bandwidth = None
self.link_protocols = None
self.circuit_protocols = None
self.hibernating = False
self.allow_single_hop_exits = False
self.extra_info_cache = False
self.extra_info_digest = None
self.hidden_service_dir = None
self.eventdns = None
self.or_addresses = []
self.read_history_end = None
self.read_history_interval = None
self.read_history_values = None
self.write_history_end = None
self.write_history_interval = None
self.write_history_values = None
self._unrecognized_lines = []
self._annotation_lines = annotations if annotations else []
# A descriptor contains a series of 'keyword lines' which are simply a
# keyword followed by an optional value. Lines can also be followed by a
# signature block.
#
# We care about the ordering of 'accept' and 'reject' entries because this
# influences the resulting exit policy, but for everything else the order
# does not matter so breaking it into key / value pairs.
entries, policy = _get_descriptor_components(raw_contents, validate, ('accept', 'reject'))
if policy == [u'reject *:*']:
self.exit_policy = REJECT_ALL_POLICY
else:
self.exit_policy = stem.exit_policy.ExitPolicy(*policy)
self._parse(entries, validate)
if validate:
self._check_constraints(entries)
def digest(self):
"""
Provides the hex encoded sha1 of our content. This value is part of the
network status entry for this relay.
:returns: **unicode** with the upper-case hex digest value for this server descriptor
"""
raise NotImplementedError('Unsupported Operation: this should be implemented by the ServerDescriptor subclass')
def get_unrecognized_lines(self):
return list(self._unrecognized_lines)
@lru_cache()
def get_annotations(self):
"""
Provides content that appeared prior to the descriptor. If this comes from
the cached-descriptors file then this commonly contains content like...
::
@downloaded-at 2012-03-18 21:18:29
@source "173.254.216.66"
:returns: **dict** with the key/value pairs in our annotations
"""
annotation_dict = {}
for line in self._annotation_lines:
if b' ' in line:
key, value = line.split(b' ', 1)
annotation_dict[key] = value
else:
annotation_dict[line] = None
return annotation_dict
def get_annotation_lines(self):
"""
Provides the lines of content that appeared prior to the descriptor. This
is the same as the
:func:`~stem.descriptor.server_descriptor.ServerDescriptor.get_annotations`
results, but with the unparsed lines and ordering retained.
:returns: **list** with the lines of annotation that came before this descriptor
"""
return self._annotation_lines
def _parse(self, entries, validate):
"""
Parses a series of 'keyword => (value, pgp block)' mappings and applies
them as attributes.
:param dict entries: descriptor contents to be applied
:param bool validate: checks the validity of descriptor content if **True**
:raises: **ValueError** if an error occurs in validation
"""
for keyword, values in entries.items():
# most just work with the first (and only) value
value, block_contents = values[0]
line = '%s %s' % (keyword, value) # original line
if block_contents:
line += '\n%s' % block_contents
if keyword == 'router':
# "router" nickname address ORPort SocksPort DirPort
router_comp = value.split()
if len(router_comp) < 5:
if not validate:
continue
raise ValueError('Router line must have five values: %s' % line)
if validate:
if not stem.util.tor_tools.is_valid_nickname(router_comp[0]):
raise ValueError("Router line entry isn't a valid nickname: %s" % router_comp[0])
elif not stem.util.connection.is_valid_ipv4_address(router_comp[1]):
raise ValueError("Router line entry isn't a valid IPv4 address: %s" % router_comp[1])
elif not stem.util.connection.is_valid_port(router_comp[2], allow_zero = True):
raise ValueError("Router line's ORPort is invalid: %s" % router_comp[2])
elif not stem.util.connection.is_valid_port(router_comp[3], allow_zero = True):
raise ValueError("Router line's SocksPort is invalid: %s" % router_comp[3])
elif not stem.util.connection.is_valid_port(router_comp[4], allow_zero = True):
raise ValueError("Router line's DirPort is invalid: %s" % router_comp[4])
elif not (router_comp[2].isdigit() and router_comp[3].isdigit() and router_comp[4].isdigit()):
continue
self.nickname = router_comp[0]
self.address = router_comp[1]
self.or_port = int(router_comp[2])
self.socks_port = None if router_comp[3] == '0' else int(router_comp[3])
self.dir_port = None if router_comp[4] == '0' else int(router_comp[4])
elif keyword == 'bandwidth':
# "bandwidth" bandwidth-avg bandwidth-burst bandwidth-observed
bandwidth_comp = value.split()
if len(bandwidth_comp) < 3:
if not validate:
continue
raise ValueError('Bandwidth line must have three values: %s' % line)
elif not bandwidth_comp[0].isdigit():
if not validate:
continue
raise ValueError("Bandwidth line's average rate isn't numeric: %s" % bandwidth_comp[0])
elif not bandwidth_comp[1].isdigit():
if not validate:
continue
raise ValueError("Bandwidth line's burst rate isn't numeric: %s" % bandwidth_comp[1])
elif not bandwidth_comp[2].isdigit():
if not validate:
continue
raise ValueError("Bandwidth line's observed rate isn't numeric: %s" % bandwidth_comp[2])
self.average_bandwidth = int(bandwidth_comp[0])
self.burst_bandwidth = int(bandwidth_comp[1])
self.observed_bandwidth = int(bandwidth_comp[2])
elif keyword == 'platform':
# "platform" string
# The platform attribute was set earlier. This line can contain any
# arbitrary data, but tor seems to report its version followed by the
# os like the following...
#
# platform Tor 0.2.2.35 (git-73ff13ab3cc9570d) on Linux x86_64
#
# There's no guarantee that we'll be able to pick these out the
# version, but might as well try to save our caller the effort.
platform_match = re.match('^(?:node-)?Tor (\S*).* on (.*)$', value)
if platform_match:
version_str, self.operating_system = platform_match.groups()
try:
self.tor_version = stem.version._get_version(version_str)
except ValueError:
pass
elif keyword == 'published':
# "published" YYYY-MM-DD HH:MM:SS
try:
self.published = datetime.datetime.strptime(value, '%Y-%m-%d %H:%M:%S')
except ValueError:
if validate:
raise ValueError("Published line's time wasn't parsable: %s" % line)
elif keyword == 'fingerprint':
# This is forty hex digits split into space separated groups of four.
# Checking that we match this pattern.
fingerprint = value.replace(' ', '')
if validate:
for grouping in value.split(' '):
if len(grouping) != 4:
raise ValueError('Fingerprint line should have groupings of four hex digits: %s' % value)
if not stem.util.tor_tools.is_valid_fingerprint(fingerprint):
raise ValueError('Tor relay fingerprints consist of forty hex digits: %s' % value)
self.fingerprint = fingerprint
elif keyword == 'hibernating':
# "hibernating" 0|1 (in practice only set if one)
if validate and not value in ('0', '1'):
raise ValueError('Hibernating line had an invalid value, must be zero or one: %s' % value)
self.hibernating = value == '1'
elif keyword == 'allow-single-hop-exits':
self.allow_single_hop_exits = True
elif keyword == 'caches-extra-info':
self.extra_info_cache = True
elif keyword == 'extra-info-digest':
# this is forty hex digits which just so happens to be the same a
# fingerprint
if validate and not stem.util.tor_tools.is_valid_fingerprint(value):
raise ValueError('Extra-info digests should consist of forty hex digits: %s' % value)
self.extra_info_digest = value
elif keyword == 'hidden-service-dir':
if value:
self.hidden_service_dir = value.split(' ')
else:
self.hidden_service_dir = ['2']
elif keyword == 'uptime':
# We need to be tolerant of negative uptimes to accommodate a past tor
# bug...
#
# Changes in version 0.1.2.7-alpha - 2007-02-06
# - If our system clock jumps back in time, don't publish a negative
# uptime in the descriptor. Also, don't let the global rate limiting
# buckets go absurdly negative.
#
# After parsing all of the attributes we'll double check that negative
# uptimes only occurred prior to this fix.
try:
self.uptime = int(value)
except ValueError:
if not validate:
continue
raise ValueError('Uptime line must have an integer value: %s' % value)
elif keyword == 'contact':
pass # parsed as a bytes field earlier
elif keyword == 'protocols':
protocols_match = re.match('^Link (.*) Circuit (.*)$', value)
if protocols_match:
link_versions, circuit_versions = protocols_match.groups()
self.link_protocols = link_versions.split(' ')
self.circuit_protocols = circuit_versions.split(' ')
elif validate:
raise ValueError('Protocols line did not match the expected pattern: %s' % line)
elif keyword == 'family':
self.family = set(value.split(' '))
elif keyword == 'eventdns':
self.eventdns = value == '1'
elif keyword == 'ipv6-policy':
self.exit_policy_v6 = stem.exit_policy.MicroExitPolicy(value)
elif keyword == 'or-address':
or_address_entries = [value for (value, _) in values]
for entry in or_address_entries:
line = '%s %s' % (keyword, entry)
if not ':' in entry:
if not validate:
continue
else:
raise ValueError('or-address line missing a colon: %s' % line)
address, port = entry.rsplit(':', 1)
is_ipv6 = address.startswith('[') and address.endswith(']')
if is_ipv6:
address = address[1:-1] # remove brackets
if not ((not is_ipv6 and stem.util.connection.is_valid_ipv4_address(address)) or
(is_ipv6 and stem.util.connection.is_valid_ipv6_address(address))):
if not validate:
continue
else:
raise ValueError('or-address line has a malformed address: %s' % line)
if stem.util.connection.is_valid_port(port):
self.or_addresses.append((address, int(port), is_ipv6))
elif validate:
raise ValueError('or-address line has a malformed port: %s' % line)
elif keyword in ('read-history', 'write-history'):
try:
timestamp, interval, remainder = \
stem.descriptor.extrainfo_descriptor._parse_timestamp_and_interval(keyword, value)
try:
if remainder:
history_values = [int(entry) for entry in remainder.split(',')]
else:
history_values = []
except ValueError:
raise ValueError('%s line has non-numeric values: %s' % (keyword, line))
if keyword == 'read-history':
self.read_history_end = timestamp
self.read_history_interval = interval
self.read_history_values = history_values
else:
self.write_history_end = timestamp
self.write_history_interval = interval
self.write_history_values = history_values
except ValueError as exc:
if validate:
raise exc
else:
self._unrecognized_lines.append(line)
# if we have a negative uptime and a tor version that shouldn't exhibit
# this bug then fail validation
if validate and self.uptime and self.tor_version:
if self.uptime < 0 and self.tor_version >= stem.version.Version('0.1.2.7'):
raise ValueError("Descriptor for version '%s' had a negative uptime value: %i" % (self.tor_version, self.uptime))
def _check_constraints(self, entries):
"""
Does a basic check that the entries conform to this descriptor type's
constraints.
:param dict entries: keyword => (value, pgp key) entries
:raises: **ValueError** if an issue arises in validation
"""
for keyword in self._required_fields():
if not keyword in entries:
raise ValueError("Descriptor must have a '%s' entry" % keyword)
for keyword in self._single_fields():
if keyword in entries and len(entries[keyword]) > 1:
raise ValueError("The '%s' entry can only appear once in a descriptor" % keyword)
expected_first_keyword = self._first_keyword()
if expected_first_keyword and expected_first_keyword != entries.keys()[0]:
raise ValueError("Descriptor must start with a '%s' entry" % expected_first_keyword)
expected_last_keyword = self._last_keyword()
if expected_last_keyword and expected_last_keyword != entries.keys()[-1]:
raise ValueError("Descriptor must end with a '%s' entry" % expected_last_keyword)
if not self.exit_policy:
raise ValueError("Descriptor must have at least one 'accept' or 'reject' entry")
# Constraints that the descriptor must meet to be valid. These can be None if
# not applicable.
def _required_fields(self):
return REQUIRED_FIELDS
def _single_fields(self):
return REQUIRED_FIELDS + SINGLE_FIELDS
def _first_keyword(self):
return 'router'
def _last_keyword(self):
return 'router-signature'
class RelayDescriptor(ServerDescriptor):
"""
Server descriptor (`descriptor specification
<https://gitweb.torproject.org/torspec.git/blob/HEAD:/dir-spec.txt>`_)
:var str onion_key: **\*** key used to encrypt EXTEND cells
:var str ntor_onion_key: base64 key used to encrypt EXTEND in the ntor protocol
:var str signing_key: **\*** relay's long-term identity key
:var str signature: **\*** signature for this descriptor
**\*** attribute is required when we're parsed with validation
"""
def __init__(self, raw_contents, validate = True, annotations = None):
self.onion_key = None
self.ntor_onion_key = None
self.signing_key = None
self.signature = None
super(RelayDescriptor, self).__init__(raw_contents, validate, annotations)
# validate the descriptor if required
if validate:
self._validate_content()
@lru_cache()
def digest(self):
"""
Provides the digest of our descriptor's content.
:returns: the digest string encoded in uppercase hex
:raises: ValueError if the digest canot be calculated
"""
# Digest is calculated from everything in the
# descriptor except the router-signature.
raw_descriptor = self.get_bytes()
start_token = b'router '
sig_token = b'\nrouter-signature\n'
start = raw_descriptor.find(start_token)
sig_start = raw_descriptor.find(sig_token)
end = sig_start + len(sig_token)
if start >= 0 and sig_start > 0 and end > start:
for_digest = raw_descriptor[start:end]
digest_hash = hashlib.sha1(stem.util.str_tools._to_bytes(for_digest))
return stem.util.str_tools._to_unicode(digest_hash.hexdigest().upper())
else:
raise ValueError('unable to calculate digest for descriptor')
def _validate_content(self):
"""
Validates that the descriptor content matches the signature.
:raises: ValueError if the signature does not match the content
"""
key_as_bytes = RelayDescriptor._get_key_bytes(self.signing_key)
# ensure the fingerprint is a hash of the signing key
if self.fingerprint:
# calculate the signing key hash
key_der_as_hash = hashlib.sha1(stem.util.str_tools._to_bytes(key_as_bytes)).hexdigest()
if key_der_as_hash != self.fingerprint.lower():
log.warn('Signing key hash: %s != fingerprint: %s' % (key_der_as_hash, self.fingerprint.lower()))
raise ValueError('Fingerprint does not match hash')
self._verify_digest(key_as_bytes)
def _verify_digest(self, key_as_der):
# check that our digest matches what was signed
if not stem.prereq.is_crypto_available():
return
from Crypto.Util import asn1
from Crypto.Util.number import bytes_to_long, long_to_bytes
# get the ASN.1 sequence
seq = asn1.DerSequence()
seq.decode(key_as_der)
modulus = seq[0]
public_exponent = seq[1] # should always be 65537
sig_as_bytes = RelayDescriptor._get_key_bytes(self.signature)
# convert the descriptor signature to an int
sig_as_long = bytes_to_long(sig_as_bytes)
# use the public exponent[e] & the modulus[n] to decrypt the int
decrypted_int = pow(sig_as_long, public_exponent, modulus)
# block size will always be 128 for a 1024 bit key
blocksize = 128
# convert the int to a byte array.
decrypted_bytes = long_to_bytes(decrypted_int, blocksize)
############################################################################
## The decrypted bytes should have a structure exactly along these lines.
## 1 byte - [null '\x00']
## 1 byte - [block type identifier '\x01'] - Should always be 1
## N bytes - [padding '\xFF' ]
## 1 byte - [separator '\x00' ]
## M bytes - [message]
## Total - 128 bytes
## More info here http://www.ietf.org/rfc/rfc2313.txt
## esp the Notes in section 8.1
############################################################################
try:
if decrypted_bytes.index(b'\x00\x01') != 0:
raise ValueError('Verification failed, identifier missing')
except ValueError:
raise ValueError('Verification failed, malformed data')
try:
identifier_offset = 2
# find the separator
seperator_index = decrypted_bytes.index(b'\x00', identifier_offset)
except ValueError:
raise ValueError('Verification failed, seperator not found')
digest_hex = codecs.encode(decrypted_bytes[seperator_index + 1:], 'hex_codec')
digest = stem.util.str_tools._to_unicode(digest_hex.upper())
local_digest = self.digest()
if digest != local_digest:
raise ValueError('Decrypted digest does not match local digest (calculated: %s, local: %s)' % (digest, local_digest))
def _parse(self, entries, validate):
entries = dict(entries) # shallow copy since we're destructive
# handles fields only in server descriptors
for keyword, values in entries.items():
value, block_contents = values[0]
line = '%s %s' % (keyword, value)
if keyword == 'onion-key':
if validate and not block_contents:
raise ValueError('Onion key line must be followed by a public key: %s' % line)
self.onion_key = block_contents
del entries['onion-key']
elif keyword == 'ntor-onion-key':
self.ntor_onion_key = value
del entries['ntor-onion-key']
elif keyword == 'signing-key':
if validate and not block_contents:
raise ValueError('Signing key line must be followed by a public key: %s' % line)
self.signing_key = block_contents
del entries['signing-key']
elif keyword == 'router-signature':
if validate and not block_contents:
raise ValueError('Router signature line must be followed by a signature block: %s' % line)
self.signature = block_contents
del entries['router-signature']
ServerDescriptor._parse(self, entries, validate)
def _compare(self, other, method):
if not isinstance(other, RelayDescriptor):
return False
return method(str(self).strip(), str(other).strip())
def __hash__(self):
return hash(str(self).strip())
def __eq__(self, other):
return self._compare(other, lambda s, o: s == o)
def __lt__(self, other):
return self._compare(other, lambda s, o: s < o)
def __le__(self, other):
return self._compare(other, lambda s, o: s <= o)
@staticmethod
def _get_key_bytes(key_string):
# Remove the newlines from the key string & strip off the
# '-----BEGIN RSA PUBLIC KEY-----' header and
# '-----END RSA PUBLIC KEY-----' footer
key_as_string = ''.join(key_string.split('\n')[1:4])
# get the key representation in bytes
key_bytes = base64.b64decode(stem.util.str_tools._to_bytes(key_as_string))
return key_bytes
class BridgeDescriptor(ServerDescriptor):
"""
Bridge descriptor (`bridge descriptor specification
<https://metrics.torproject.org/formats.html#bridgedesc>`_)
"""
def __init__(self, raw_contents, validate = True, annotations = None):
self._digest = None
super(BridgeDescriptor, self).__init__(raw_contents, validate, annotations)
def digest(self):
return self._digest
def _parse(self, entries, validate):
entries = dict(entries)
# handles fields only in bridge descriptors
for keyword, values in entries.items():
value, block_contents = values[0]
line = '%s %s' % (keyword, value)
if keyword == 'router-digest':
if validate and not stem.util.tor_tools.is_hex_digits(value, 40):
raise ValueError('Router digest line had an invalid sha1 digest: %s' % line)
self._digest = stem.util.str_tools._to_unicode(value)
del entries['router-digest']
ServerDescriptor._parse(self, entries, validate)
def is_scrubbed(self):
"""
Checks if we've been properly scrubbed in accordance with the `bridge
descriptor specification
<https://metrics.torproject.org/formats.html#bridgedesc>`_. Validation is a
moving target so this may not
be fully up to date.
:returns: **True** if we're scrubbed, **False** otherwise
"""
return self.get_scrubbing_issues() == []
@lru_cache()
def get_scrubbing_issues(self):
"""
Provides issues with our scrubbing.
:returns: **list** of strings which describe issues we have with our
scrubbing, this list is empty if we're properly scrubbed
"""
issues = []
if not self.address.startswith('10.'):
issues.append("Router line's address should be scrubbed to be '10.x.x.x': %s" % self.address)
if self.contact and self.contact != "somebody":
issues.append("Contact line should be scrubbed to be 'somebody', but instead had '%s'" % self.contact)
for address, _, is_ipv6 in self.or_addresses:
if not is_ipv6 and not address.startswith('10.'):
issues.append("or-address line's address should be scrubbed to be '10.x.x.x': %s" % address)
elif is_ipv6 and not address.startswith('fd9f:2e19:3bcf::'):
# TODO: this check isn't quite right because we aren't checking that
# the next grouping of hex digits contains 1-2 digits
issues.append("or-address line's address should be scrubbed to be 'fd9f:2e19:3bcf::xx:xxxx': %s" % address)
for line in self.get_unrecognized_lines():
if line.startswith('onion-key '):
issues.append('Bridge descriptors should have their onion-key scrubbed: %s' % line)
elif line.startswith('signing-key '):
issues.append('Bridge descriptors should have their signing-key scrubbed: %s' % line)
elif line.startswith('router-signature '):
issues.append('Bridge descriptors should have their signature scrubbed: %s' % line)
return issues
def _required_fields(self):
# bridge required fields are the same as a relay descriptor, minus items
# excluded according to the format page
excluded_fields = [
'onion-key',
'signing-key',
'router-signature',
]
included_fields = [
'router-digest',
]
return tuple(included_fields + [f for f in REQUIRED_FIELDS if not f in excluded_fields])
def _single_fields(self):
return self._required_fields() + SINGLE_FIELDS
def _last_keyword(self):
return None
def _compare(self, other, method):
if not isinstance(other, BridgeDescriptor):
return False
return method(str(self).strip(), str(other).strip())
def __hash__(self):
return hash(str(self).strip())
def __eq__(self, other):
return self._compare(other, lambda s, o: s == o)
def __lt__(self, other):
return self._compare(other, lambda s, o: s < o)
def __le__(self, other):
return self._compare(other, lambda s, o: s <= o)
|