[204] | 1 | #!/usr/bin/env python |
---|
| 2 | |
---|
[210] | 3 | ## @package pyregfi |
---|
| 4 | # Python interface to the regfi library. |
---|
| 5 | # |
---|
| 6 | |
---|
[221] | 7 | ## @mainpage API Documentation |
---|
| 8 | # |
---|
| 9 | # The pyregfi module provides a Python interface to the @ref regfi Windows |
---|
| 10 | # registry library. |
---|
| 11 | # |
---|
| 12 | # The library operates on registry hives, each of which is contained within a |
---|
| 13 | # single file. To get started, one must first open the registry hive file with |
---|
| 14 | # the open() or file() Python built-in functions (or equivalent) and then pass |
---|
| 15 | # the resulting file object to pyregfi. For example: |
---|
| 16 | # @code |
---|
| 17 | # >>> import pyregfi |
---|
| 18 | # >>> fh = open('/mnt/win/c/WINDOWS/system32/config/system', 'rb') |
---|
| 19 | # >>> myHive = pyregfi.Hive(fh) |
---|
| 20 | # @endcode |
---|
| 21 | # |
---|
| 22 | # Using this Hive object, one can begin investigating what top-level keys |
---|
| 23 | # exist by starting with the root Key attribute: |
---|
| 24 | # @code |
---|
| 25 | # >>> for key in myHive.root.subkeys: |
---|
| 26 | # ... print(key.name) |
---|
| 27 | # ControlSet001 |
---|
| 28 | # ControlSet003 |
---|
| 29 | # LastKnownGoodRecovery |
---|
| 30 | # MountedDevices |
---|
| 31 | # Select |
---|
| 32 | # Setup |
---|
| 33 | # WPA |
---|
| 34 | # @endcode |
---|
| 35 | # |
---|
| 36 | # From there, accessing subkeys and values by name is a simple matter of: |
---|
| 37 | # @code |
---|
| 38 | # >>> myKey = myHive.root.subkeys['Select'] |
---|
| 39 | # >>> myValue = myKey.values['Current'] |
---|
| 40 | # @endcode |
---|
| 41 | # |
---|
| 42 | # The data associated with a Value can be obtained through the fetch_data() |
---|
| 43 | # method: |
---|
| 44 | # @code |
---|
| 45 | # >>> print(myValue.fetch_data()) |
---|
| 46 | # 1 |
---|
| 47 | # @endcode |
---|
| 48 | # |
---|
| 49 | # While useful for simple exercises, using the subkeys object for deeply nested |
---|
| 50 | # paths is not efficient and doesn't make for particularly attractive code. |
---|
| 51 | # Instead, a special-purpose HiveIterator class is provided for simplicity of |
---|
| 52 | # use and fast access to specific known paths: |
---|
| 53 | # @code |
---|
| 54 | # >>> myIter = pyregfi.HiveIterator(myHive) |
---|
| 55 | # >>> myIter.descend(['ControlSet001','Control','NetworkProvider','HwOrder']) |
---|
| 56 | # >>> myKey = myIter.current_key() |
---|
| 57 | # >>> print(myKey.values['ProviderOrder'].fetch_data()) |
---|
| 58 | # RDPNP,LanmanWorkstation,WebClient |
---|
| 59 | # @endcode |
---|
| 60 | # |
---|
| 61 | # The first two lines above can be simplified in some "syntactic sugar" provided |
---|
| 62 | # by the Hive.subtree() method. Also, as one might expect, the HiveIterator |
---|
| 63 | # also acts as an iterator, producing keys in a depth-first order. |
---|
| 64 | # For instance, to traverse all keys under the ControlSet003\\Services key, |
---|
| 65 | # printing their names as we go, we could do: |
---|
| 66 | # @code |
---|
| 67 | # >>> for key in Hive.subtree(['ControlSet003','Services']): |
---|
| 68 | # >>> print(key.name) |
---|
| 69 | # Services |
---|
| 70 | # Abiosdsk |
---|
| 71 | # abp480n5 |
---|
| 72 | # Parameters |
---|
| 73 | # PnpInterface |
---|
| 74 | # ACPI |
---|
| 75 | # [...] |
---|
| 76 | # @endcode |
---|
| 77 | # |
---|
| 78 | # Note that "Services" was printed first, since the subtree is traversed as a |
---|
| 79 | # "preordering depth-first" search starting with the HiveIterator's current_key(). |
---|
| 80 | # As one might expect, traversals of subtrees stops when all elements in a |
---|
| 81 | # specific subtree (and none outside of it) have been traversed. |
---|
| 82 | # |
---|
| 83 | # For more information, peruse the various attributes and methods available on |
---|
| 84 | # the Hive, HiveIterator, Key, Value, and Security classes. |
---|
| 85 | # |
---|
| 86 | # @note @ref regfi is a read-only library by design and there |
---|
| 87 | # are no plans to implement write support. |
---|
| 88 | # |
---|
| 89 | # @note At present, pyregfi has been tested with Python versions 2.6 and 3.1 |
---|
| 90 | # |
---|
| 91 | # @note Developers strive to make pyregfi thread-safe. |
---|
| 92 | # |
---|
| 93 | # @note Key and Value names are case-sensitive in regfi and pyregfi |
---|
| 94 | # |
---|
[204] | 95 | import sys |
---|
[219] | 96 | import time |
---|
[214] | 97 | import weakref |
---|
[204] | 98 | from pyregfi.structures import * |
---|
| 99 | |
---|
| 100 | import ctypes |
---|
| 101 | import ctypes.util |
---|
| 102 | |
---|
[221] | 103 | ## An enumeration of registry Value data types |
---|
[210] | 104 | # |
---|
[221] | 105 | # @note This is a static class, there is no need to instantiate it. |
---|
| 106 | # Just access its attributes directly as DATA_TYPES.SZ, etc |
---|
| 107 | class DATA_TYPES(object): |
---|
| 108 | ## None / Unknown |
---|
| 109 | NONE = 0 |
---|
| 110 | ## String |
---|
| 111 | SZ = 1 |
---|
| 112 | ## String with %...% expansions |
---|
| 113 | EXPAND_SZ = 2 |
---|
| 114 | ## Binary buffer |
---|
| 115 | BINARY = 3 |
---|
| 116 | ## 32 bit integer (little endian) |
---|
| 117 | DWORD = 4 # DWORD, little endian |
---|
| 118 | ## 32 bit integer (little endian) |
---|
| 119 | DWORD_LE = 4 |
---|
| 120 | ## 32 bit integer (big endian) |
---|
| 121 | DWORD_BE = 5 # DWORD, big endian |
---|
| 122 | ## Symbolic link |
---|
| 123 | LINK = 6 |
---|
| 124 | ## List of strings |
---|
| 125 | MULTI_SZ = 7 |
---|
| 126 | ## Unknown structure |
---|
| 127 | RESOURCE_LIST = 8 |
---|
| 128 | ## Unknown structure |
---|
| 129 | FULL_RESOURCE_DESCRIPTOR = 9 |
---|
| 130 | ## Unknown structure |
---|
| 131 | RESOURCE_REQUIREMENTS_LIST = 10 |
---|
| 132 | ## 64 bit integer |
---|
| 133 | QWORD = 11 # 64-bit little endian |
---|
[205] | 134 | |
---|
| 135 | |
---|
[208] | 136 | def _buffer2bytearray(char_pointer, length): |
---|
| 137 | if length == 0 or char_pointer == None: |
---|
| 138 | return None |
---|
| 139 | |
---|
| 140 | ret_val = bytearray(length) |
---|
| 141 | for i in range(0,length): |
---|
| 142 | ret_val[i] = char_pointer[i][0] |
---|
| 143 | |
---|
| 144 | return ret_val |
---|
| 145 | |
---|
| 146 | |
---|
[215] | 147 | def _strlist2charss(str_list): |
---|
| 148 | ret_val = [] |
---|
| 149 | for s in str_list: |
---|
| 150 | ret_val.append(s.encode('utf-8', 'replace')) |
---|
| 151 | |
---|
[220] | 152 | ret_val = (ctypes.c_char_p*(len(str_list)+1))(*ret_val) |
---|
[215] | 153 | # Terminate the char** with a NULL pointer |
---|
| 154 | ret_val[-1] = 0 |
---|
| 155 | |
---|
| 156 | return ret_val |
---|
| 157 | |
---|
| 158 | |
---|
[209] | 159 | def _charss2strlist(chars_pointer): |
---|
| 160 | ret_val = [] |
---|
| 161 | i = 0 |
---|
| 162 | s = chars_pointer[i] |
---|
| 163 | while s != None: |
---|
[213] | 164 | ret_val.append(s.decode('utf-8', 'replace')) |
---|
[209] | 165 | i += 1 |
---|
| 166 | s = chars_pointer[i] |
---|
[208] | 167 | |
---|
[209] | 168 | return ret_val |
---|
[208] | 169 | |
---|
[210] | 170 | |
---|
[221] | 171 | ## Retrieves messages produced by regfi during parsing and interpretation |
---|
| 172 | # |
---|
| 173 | # The regfi C library may generate log messages stored in a special thread-safe |
---|
| 174 | # global data structure. These messages should be retrieved periodically or |
---|
| 175 | # after each major operation by callers to determine if any errors or warnings |
---|
| 176 | # should be reported to the user. Failure to retrieve these could result in |
---|
| 177 | # excessive memory consumption. |
---|
| 178 | def GetLogMessages(): |
---|
| 179 | msgs = regfi.regfi_log_get_str() |
---|
| 180 | if msgs == None: |
---|
| 181 | return '' |
---|
| 182 | return msgs.decode('utf-8') |
---|
| 183 | |
---|
| 184 | |
---|
| 185 | ## Abstract class for most objects returned by the library |
---|
[212] | 186 | class _StructureWrapper(object): |
---|
[214] | 187 | _hive = None |
---|
| 188 | _base = None |
---|
[206] | 189 | |
---|
[207] | 190 | def __init__(self, hive, base): |
---|
[215] | 191 | if not hive: |
---|
| 192 | raise Exception("Could not create _StructureWrapper," |
---|
| 193 | + " hive is NULL. Current log:\n" |
---|
| 194 | + GetLogMessages()) |
---|
| 195 | if not base: |
---|
| 196 | raise Exception("Could not create _StructureWrapper," |
---|
| 197 | + " base is NULL. Current log:\n" |
---|
| 198 | + GetLogMessages()) |
---|
[214] | 199 | self._hive = hive |
---|
| 200 | self._base = base |
---|
[206] | 201 | |
---|
[221] | 202 | # Memory management for most regfi structures is taken care of here |
---|
[206] | 203 | def __del__(self): |
---|
[214] | 204 | regfi.regfi_free_record(self._base) |
---|
[206] | 205 | |
---|
[221] | 206 | # Any attribute requests not explicitly defined in subclasses gets passed |
---|
| 207 | # to the equivalent REGFI_* structure defined in structures.py |
---|
[206] | 208 | def __getattr__(self, name): |
---|
[214] | 209 | return getattr(self._base.contents, name) |
---|
[221] | 210 | |
---|
| 211 | ## Test for equality |
---|
| 212 | # |
---|
| 213 | # Records returned by pyregfi may be compared with one another. For example: |
---|
| 214 | # @code |
---|
| 215 | # >>> key2 = key1.subkeys['child'] |
---|
| 216 | # >>> key1 == key2 |
---|
| 217 | # False |
---|
| 218 | # >>> key1 != key2 |
---|
| 219 | # True |
---|
| 220 | # >>> key1 == key2.get_parent() |
---|
| 221 | # True |
---|
| 222 | # @endcode |
---|
[206] | 223 | def __eq__(self, other): |
---|
| 224 | return (type(self) == type(other)) and (self.offset == other.offset) |
---|
| 225 | |
---|
| 226 | def __ne__(self, other): |
---|
| 227 | return (not self.__eq__(other)) |
---|
| 228 | |
---|
[208] | 229 | |
---|
[221] | 230 | class Key(): |
---|
[206] | 231 | pass |
---|
| 232 | |
---|
[221] | 233 | |
---|
| 234 | class Value(): |
---|
[206] | 235 | pass |
---|
| 236 | |
---|
[221] | 237 | |
---|
| 238 | ## Registry security record and descriptor |
---|
| 239 | # XXX: Access to security descriptors not yet implemented |
---|
[206] | 240 | class Security(_StructureWrapper): |
---|
| 241 | pass |
---|
| 242 | |
---|
[221] | 243 | ## Abstract class for ValueList and SubkeyList |
---|
[212] | 244 | class _GenericList(object): |
---|
[214] | 245 | _hive = None |
---|
| 246 | _key = None |
---|
| 247 | _length = None |
---|
| 248 | _current = None |
---|
[207] | 249 | |
---|
[221] | 250 | # implementation-specific functions for SubkeyList and ValueList |
---|
[214] | 251 | _fetch_num = None |
---|
| 252 | _find_element = None |
---|
| 253 | _get_element = None |
---|
| 254 | _constructor = None |
---|
[208] | 255 | |
---|
[207] | 256 | def __init__(self, key): |
---|
[214] | 257 | self._hive = key._hive |
---|
| 258 | |
---|
| 259 | # Normally it's good to avoid cyclic references like this |
---|
| 260 | # (key.list.key...) but in this case it makes ctypes memory |
---|
| 261 | # management easier to reference the Key instead of the base |
---|
| 262 | # structure. We use a weak reference in order to allow for garbage |
---|
| 263 | # collection, since value/subkey lists should not be usable if their |
---|
| 264 | # parent Key is freed anyway. |
---|
| 265 | |
---|
[207] | 266 | # XXX: check for NULL here, throw an exception if so. |
---|
[214] | 267 | self._key = weakref.proxy(key) |
---|
| 268 | self._length = self._fetch_num(key._base) |
---|
[207] | 269 | |
---|
[221] | 270 | |
---|
| 271 | ## Length of list |
---|
[207] | 272 | def __len__(self): |
---|
[214] | 273 | return self._length |
---|
[207] | 274 | |
---|
[221] | 275 | |
---|
| 276 | ## Retrieves a list element by name |
---|
| 277 | # |
---|
| 278 | # @return the first element whose name matches, or None if the element |
---|
| 279 | # could not be found |
---|
[207] | 280 | def __getitem__(self, name): |
---|
[220] | 281 | index = ctypes.c_uint32() |
---|
[208] | 282 | if isinstance(name, str): |
---|
| 283 | name = name.encode('utf-8') |
---|
| 284 | |
---|
[209] | 285 | if name != None: |
---|
| 286 | name = create_string_buffer(bytes(name)) |
---|
| 287 | |
---|
[220] | 288 | if self._find_element(self._hive.file, self._key._base, |
---|
| 289 | name, byref(index)): |
---|
| 290 | return self._constructor(self._hive, |
---|
[214] | 291 | self._get_element(self._hive.file, |
---|
[216] | 292 | self._key._base, |
---|
[214] | 293 | index)) |
---|
[207] | 294 | raise KeyError('') |
---|
| 295 | |
---|
[209] | 296 | def get(self, name, default): |
---|
| 297 | try: |
---|
| 298 | return self[name] |
---|
| 299 | except KeyError: |
---|
| 300 | return default |
---|
| 301 | |
---|
[207] | 302 | def __iter__(self): |
---|
[214] | 303 | self._current = 0 |
---|
[207] | 304 | return self |
---|
| 305 | |
---|
| 306 | def __next__(self): |
---|
[214] | 307 | if self._current >= self._length: |
---|
[207] | 308 | raise StopIteration('') |
---|
| 309 | |
---|
[214] | 310 | elem = self._get_element(self._hive.file, self._key._base, |
---|
[220] | 311 | ctypes.c_uint32(self._current)) |
---|
[214] | 312 | self._current += 1 |
---|
| 313 | return self._constructor(self._hive, elem) |
---|
[207] | 314 | |
---|
[212] | 315 | # For Python 2.x |
---|
[214] | 316 | next = __next__ |
---|
[207] | 317 | |
---|
[212] | 318 | |
---|
[221] | 319 | ## The list of subkeys associated with a Key |
---|
| 320 | # |
---|
| 321 | # This attribute is both iterable: |
---|
| 322 | # @code |
---|
| 323 | # for k in myKey.subkeys: |
---|
| 324 | # ... |
---|
| 325 | # @endcode |
---|
| 326 | # and accessible as a dictionary: |
---|
| 327 | # @code |
---|
| 328 | # mySubkey = myKey.subkeys["keyName"] |
---|
| 329 | # @endcode |
---|
| 330 | # |
---|
| 331 | # @note SubkeyLists should never be accessed directly and only exist |
---|
| 332 | # in association with a parent Key object. Do not retain references to |
---|
| 333 | # SubkeyLists. Instead, access them via their parent Key at all times. |
---|
| 334 | class SubkeyList(_GenericList): |
---|
[214] | 335 | _fetch_num = regfi.regfi_fetch_num_subkeys |
---|
| 336 | _find_element = regfi.regfi_find_subkey |
---|
| 337 | _get_element = regfi.regfi_get_subkey |
---|
[208] | 338 | |
---|
| 339 | |
---|
[221] | 340 | ## The list of values associated with a Key |
---|
| 341 | # |
---|
| 342 | # This attribute is both iterable: |
---|
| 343 | # @code |
---|
| 344 | # for v in myKey.values: |
---|
| 345 | # ... |
---|
| 346 | # @endcode |
---|
| 347 | # and accessible as a dictionary: |
---|
| 348 | # @code |
---|
| 349 | # myValue = myKey.values["valueName"] |
---|
| 350 | # @endcode |
---|
| 351 | # |
---|
| 352 | # @note ValueLists should never be accessed directly and only exist |
---|
| 353 | # in association with a parent Key object. Do not retain references to |
---|
| 354 | # ValueLists. Instead, access them via their parent Key at all times. |
---|
| 355 | class ValueList(_GenericList): |
---|
[214] | 356 | _fetch_num = regfi.regfi_fetch_num_values |
---|
| 357 | _find_element = regfi.regfi_find_value |
---|
| 358 | _get_element = regfi.regfi_get_value |
---|
[208] | 359 | |
---|
| 360 | |
---|
[215] | 361 | ## Registry key |
---|
[221] | 362 | # These represent registry keys (@ref REGFI_NK records) and provide |
---|
| 363 | # access to their subkeys, values, and other metadata. |
---|
| 364 | # |
---|
| 365 | # @note Value instances may provide access to more than the attributes |
---|
| 366 | # documented here. However, undocumented attributes may change over time |
---|
| 367 | # and are not officially supported. If you need access to an attribute |
---|
| 368 | # not shown here, see pyregfi.structures. |
---|
[207] | 369 | class Key(_StructureWrapper): |
---|
[221] | 370 | ## A @ref ValueList object representing the list of Values |
---|
| 371 | # stored on this Key |
---|
[207] | 372 | values = None |
---|
[221] | 373 | |
---|
| 374 | ## A @ref SubkeyList object representing the list of subkeys |
---|
| 375 | # stored on this Key |
---|
[208] | 376 | subkeys = None |
---|
[207] | 377 | |
---|
[221] | 378 | ## The raw Key name as an uninterpreted bytearray |
---|
| 379 | name_raw = (b"...") |
---|
| 380 | |
---|
| 381 | ## The name of the Key as a (unicode) string |
---|
| 382 | name = "..." |
---|
| 383 | |
---|
| 384 | ## The absolute file offset of the Key record's cell in the Hive file |
---|
| 385 | offset = 0xCAFEBABE |
---|
| 386 | |
---|
| 387 | ## This Key's last modified time represented as the number of seconds |
---|
| 388 | # since the UNIX epoch in UTC; similar to what time.time() returns |
---|
| 389 | modified = 1300000000.123456 |
---|
| 390 | |
---|
| 391 | ## The NK record's flags field |
---|
| 392 | flags = 0x10110001 |
---|
| 393 | |
---|
[207] | 394 | def __init__(self, hive, base): |
---|
| 395 | super(Key, self).__init__(hive, base) |
---|
[221] | 396 | self.values = ValueList(self) |
---|
| 397 | self.subkeys = SubkeyList(self) |
---|
[207] | 398 | |
---|
[208] | 399 | def __getattr__(self, name): |
---|
| 400 | if name == "name": |
---|
[219] | 401 | ret_val = super(Key, self).__getattr__(name) |
---|
| 402 | |
---|
[209] | 403 | if ret_val == None: |
---|
| 404 | ret_val = self.name_raw |
---|
| 405 | else: |
---|
[213] | 406 | ret_val = ret_val.decode('utf-8', 'replace') |
---|
[209] | 407 | |
---|
[208] | 408 | elif name == "name_raw": |
---|
[219] | 409 | ret_val = super(Key, self).__getattr__(name) |
---|
[208] | 410 | length = super(Key, self).__getattr__('name_length') |
---|
| 411 | ret_val = _buffer2bytearray(ret_val, length) |
---|
| 412 | |
---|
[219] | 413 | elif name == "modified": |
---|
| 414 | ret_val = regfi.regfi_nt2unix_time(byref(self._base.contents.mtime)) |
---|
| 415 | |
---|
| 416 | else: |
---|
| 417 | ret_val = super(Key, self).__getattr__(name) |
---|
| 418 | |
---|
[208] | 419 | return ret_val |
---|
| 420 | |
---|
[221] | 421 | |
---|
| 422 | ## Retrieves the Security properties for this key |
---|
[207] | 423 | def fetch_security(self): |
---|
[214] | 424 | return Security(self._hive, |
---|
[215] | 425 | regfi.regfi_fetch_sk(self._hive.file, self._base)) |
---|
[207] | 426 | |
---|
[221] | 427 | |
---|
| 428 | ## Retrieves the class name for this key |
---|
| 429 | # |
---|
| 430 | # Class names are typically stored as UTF-16LE strings, so these are decoded |
---|
| 431 | # into proper python (unicode) strings. However, if this fails, a bytearray |
---|
| 432 | # is instead returned containing the raw buffer stored for the class name. |
---|
| 433 | # |
---|
| 434 | # @return The class name as a string or bytearray. None if a class name |
---|
| 435 | # doesn't exist or an unrecoverable error occurred during retrieval. |
---|
[219] | 436 | def fetch_classname(self): |
---|
| 437 | ret_val = None |
---|
| 438 | cn_p = regfi.regfi_fetch_classname(self._hive.file, self._base) |
---|
| 439 | if cn_p: |
---|
| 440 | cn_struct = cn_p.contents |
---|
| 441 | if cn_struct.interpreted: |
---|
| 442 | ret_val = cn_struct.interpreted.decode('utf-8', 'replace') |
---|
| 443 | else: |
---|
| 444 | ret_val = _buffer2bytearray(cn_struct.raw, |
---|
| 445 | cn_struct.size) |
---|
| 446 | regfi.regfi_free_record(cn_p) |
---|
| 447 | |
---|
| 448 | return ret_val |
---|
| 449 | |
---|
[221] | 450 | |
---|
| 451 | ## Retrieves this key's parent key |
---|
| 452 | # |
---|
| 453 | # @return The parent's Key instance or None if current key is root |
---|
| 454 | # (or an error occured) |
---|
[215] | 455 | def get_parent(self): |
---|
[218] | 456 | if self.is_root(): |
---|
| 457 | return None |
---|
[215] | 458 | parent_base = regfi.regfi_get_parentkey(self._hive.file, self._base) |
---|
| 459 | if parent_base: |
---|
| 460 | return Key(self._hive, parent_base) |
---|
| 461 | return None |
---|
| 462 | |
---|
| 463 | def is_root(self): |
---|
[218] | 464 | return (self._hive.root == self) |
---|
[215] | 465 | |
---|
| 466 | |
---|
[210] | 467 | ## Registry value (metadata) |
---|
| 468 | # |
---|
| 469 | # These represent registry values (@ref REGFI_VK records) and provide |
---|
| 470 | # access to their associated data. |
---|
[221] | 471 | # |
---|
| 472 | # @note Value instances may provide access to more than the attributes |
---|
| 473 | # documented here. However, undocumented attributes may change over time |
---|
| 474 | # and are not officially supported. If you need access to an attribute |
---|
| 475 | # not shown here, see pyregfi.structures. |
---|
[208] | 476 | class Value(_StructureWrapper): |
---|
[221] | 477 | ## The raw Value name as an uninterpreted bytearray |
---|
| 478 | name_raw = (b"...") |
---|
| 479 | |
---|
| 480 | ## The name of the Value as a (unicode) string |
---|
| 481 | name = "..." |
---|
| 482 | |
---|
| 483 | ## The absolute file offset of the Value record's cell in the Hive file |
---|
| 484 | offset = 0xCAFEBABE |
---|
| 485 | |
---|
| 486 | ## The length of data advertised in the VK record |
---|
| 487 | data_size = 0xCAFEBABE |
---|
| 488 | |
---|
| 489 | ## An integer which represents the data type for this Value's data |
---|
| 490 | # Typically this value is one of 12 types defined in @ref DATA_TYPES, |
---|
| 491 | # but in some cases (the SAM hive) it may be used for other purposes |
---|
| 492 | type = DATA_TYPES.NONE |
---|
| 493 | |
---|
| 494 | ## The VK record's flags field |
---|
| 495 | flags = 0x10110001 |
---|
| 496 | |
---|
| 497 | ## Retrieves the Value's data according to advertised type |
---|
| 498 | # |
---|
| 499 | # Data is loaded from its cell(s) and then interpreted based on the data |
---|
| 500 | # type recorded in the Value. It is not uncommon for data to be stored with |
---|
| 501 | # the wrong type or even with invalid types. If you have difficulty |
---|
| 502 | # obtaining desired data here, use @ref fetch_raw_data(). |
---|
| 503 | # |
---|
| 504 | # @return The interpreted representation of the data as one of several |
---|
| 505 | # possible Python types, as listed below. None if any failure |
---|
| 506 | # occurred during extraction or conversion. |
---|
| 507 | # |
---|
| 508 | # @retval string for SZ, EXPAND_SZ, and LINK |
---|
| 509 | # @retval int for DWORD, DWORD_BE, and QWORD |
---|
| 510 | # @retval list(string) for MULTI_SZ |
---|
| 511 | # @retval bytearray for NONE, BINARY, RESOURCE_LIST, |
---|
| 512 | # FULL_RESOURCE_DESCRIPTOR, and RESOURCE_REQUIREMENTS_LIST |
---|
| 513 | # |
---|
[219] | 514 | def fetch_data(self): |
---|
[209] | 515 | ret_val = None |
---|
[219] | 516 | data_p = regfi.regfi_fetch_data(self._hive.file, self._base) |
---|
| 517 | if not data_p: |
---|
| 518 | return None |
---|
| 519 | data_struct = data_p.contents |
---|
[208] | 520 | |
---|
[219] | 521 | if data_struct.interpreted_size == 0: |
---|
| 522 | ret_val = None |
---|
[221] | 523 | elif data_struct.type in (DATA_TYPES.SZ, DATA_TYPES.EXPAND_SZ, DATA_TYPES.LINK): |
---|
[219] | 524 | # Unicode strings |
---|
| 525 | ret_val = data_struct.interpreted.string.decode('utf-8', 'replace') |
---|
[221] | 526 | elif data_struct.type in (DATA_TYPES.DWORD, DATA_TYPES.DWORD_BE): |
---|
[219] | 527 | # 32 bit integers |
---|
| 528 | ret_val = data_struct.interpreted.dword |
---|
[221] | 529 | elif data_struct.type == DATA_TYPES.QWORD: |
---|
[219] | 530 | # 64 bit integers |
---|
| 531 | ret_val = data_struct.interpreted.qword |
---|
[221] | 532 | elif data_struct.type == DATA_TYPES.MULTI_SZ: |
---|
[219] | 533 | ret_val = _charss2strlist(data_struct.interpreted.multiple_string) |
---|
[221] | 534 | elif data_struct.type in (DATA_TYPES.NONE, DATA_TYPES.RESOURCE_LIST, |
---|
| 535 | DATA_TYPES.FULL_RESOURCE_DESCRIPTOR, |
---|
| 536 | DATA_TYPES.RESOURCE_REQUIREMENTS_LIST, |
---|
| 537 | DATA_TYPES.BINARY): |
---|
[219] | 538 | ret_val = _buffer2bytearray(data_struct.interpreted.none, |
---|
| 539 | data_struct.interpreted_size) |
---|
[209] | 540 | |
---|
[219] | 541 | regfi.regfi_free_record(data_p) |
---|
| 542 | return ret_val |
---|
[221] | 543 | |
---|
| 544 | |
---|
| 545 | ## Retrieves raw representation of Value's data |
---|
| 546 | # |
---|
| 547 | # @return A bytearray containing the data |
---|
| 548 | # |
---|
[219] | 549 | def fetch_raw_data(self): |
---|
| 550 | ret_val = None |
---|
| 551 | # XXX: should we load the data without interpretation instead? |
---|
| 552 | data_p = regfi.regfi_fetch_data(self._hive.file, self._base) |
---|
| 553 | if not data_p: |
---|
| 554 | return None |
---|
[209] | 555 | |
---|
[219] | 556 | data_struct = data_p.contents |
---|
| 557 | ret_val = _buffer2bytearray(data_struct.raw, |
---|
| 558 | data_struct.size) |
---|
| 559 | regfi.regfi_free_record(data_p) |
---|
[208] | 560 | return ret_val |
---|
| 561 | |
---|
[221] | 562 | |
---|
[219] | 563 | def __getattr__(self, name): |
---|
| 564 | ret_val = super(Value, self).__getattr__(name) |
---|
| 565 | if name == "name": |
---|
| 566 | if ret_val == None: |
---|
| 567 | ret_val = self.name_raw |
---|
| 568 | else: |
---|
| 569 | ret_val = ret_val.decode('utf-8', 'replace') |
---|
[208] | 570 | |
---|
[219] | 571 | elif name == "name_raw": |
---|
| 572 | length = super(Value, self).__getattr__('name_length') |
---|
| 573 | ret_val = _buffer2bytearray(ret_val, length) |
---|
| 574 | |
---|
| 575 | return ret_val |
---|
| 576 | |
---|
| 577 | |
---|
[208] | 578 | # Avoids chicken/egg class definitions. |
---|
| 579 | # Also makes for convenient code reuse in these lists' parent classes. |
---|
[221] | 580 | SubkeyList._constructor = Key |
---|
| 581 | ValueList._constructor = Value |
---|
[208] | 582 | |
---|
| 583 | |
---|
| 584 | |
---|
[210] | 585 | ## Represents a single registry hive (file) |
---|
| 586 | class Hive(): |
---|
[204] | 587 | file = None |
---|
| 588 | raw_file = None |
---|
[218] | 589 | _root = None |
---|
| 590 | |
---|
[221] | 591 | ## The root Key of this Hive |
---|
| 592 | root = None |
---|
| 593 | |
---|
| 594 | ## This Hives's last modified time represented as the number of seconds |
---|
| 595 | # since the UNIX epoch in UTC; similar to what time.time() returns |
---|
| 596 | modified = 1300000000.123456 |
---|
| 597 | |
---|
| 598 | ## First sequence number |
---|
| 599 | sequence1 = 12345678 |
---|
| 600 | |
---|
| 601 | ## Second sequence number |
---|
| 602 | sequence2 = 12345678 |
---|
| 603 | |
---|
| 604 | ## Major version |
---|
| 605 | major_version = 1 |
---|
| 606 | |
---|
| 607 | ## Minor version |
---|
| 608 | minor_version = 5 |
---|
| 609 | |
---|
| 610 | # XXX: Possibly add a second or factory function which opens a |
---|
| 611 | # hive file for you |
---|
| 612 | |
---|
| 613 | ## Constructor |
---|
| 614 | # |
---|
| 615 | # @param fh A Python file object. The constructor first looks for a valid |
---|
| 616 | # fileno attribute on this object and uses it if possible. |
---|
| 617 | # Otherwise, the seek and read methods are used for file |
---|
| 618 | # access. |
---|
| 619 | # |
---|
| 620 | # @note Supplied file must be seekable |
---|
[204] | 621 | def __init__(self, fh): |
---|
[205] | 622 | try: |
---|
[221] | 623 | # The fileno method may not exist, or it may throw an exception |
---|
| 624 | # when called if the file isn't backed with a descriptor. |
---|
[205] | 625 | if hasattr(fh, 'fileno'): |
---|
[213] | 626 | self.file = regfi.regfi_alloc(fh.fileno(), REGFI_ENCODING_UTF8) |
---|
[205] | 627 | return |
---|
| 628 | except: |
---|
| 629 | pass |
---|
[204] | 630 | |
---|
[205] | 631 | self.raw_file = structures.REGFI_RAW_FILE() |
---|
| 632 | self.raw_file.fh = fh |
---|
| 633 | self.raw_file.seek = seek_cb_type(self.raw_file.cb_seek) |
---|
| 634 | self.raw_file.read = read_cb_type(self.raw_file.cb_read) |
---|
[213] | 635 | self.file = regfi.regfi_alloc_cb(self.raw_file, REGFI_ENCODING_UTF8) |
---|
[204] | 636 | |
---|
[221] | 637 | |
---|
[204] | 638 | def __getattr__(self, name): |
---|
[218] | 639 | if name == "root": |
---|
| 640 | if self._root == None: |
---|
| 641 | self._root = Key(self, regfi.regfi_get_rootkey(self.file)) |
---|
| 642 | return self._root |
---|
| 643 | |
---|
[221] | 644 | elif name == "modified": |
---|
| 645 | return regfi.regfi_nt2unix_time(byref(self._base.contents.mtime)) |
---|
| 646 | |
---|
[204] | 647 | return getattr(self.file.contents, name) |
---|
[221] | 648 | |
---|
[205] | 649 | |
---|
[210] | 650 | def __del__(self): |
---|
[205] | 651 | regfi.regfi_free(self.file) |
---|
| 652 | if self.raw_file != None: |
---|
[213] | 653 | self.raw_file = None |
---|
[204] | 654 | |
---|
[221] | 655 | |
---|
[205] | 656 | def __iter__(self): |
---|
| 657 | return HiveIterator(self) |
---|
[204] | 658 | |
---|
[215] | 659 | |
---|
[210] | 660 | ## Creates a @ref HiveIterator initialized at the specified path in |
---|
[221] | 661 | # the hive. |
---|
[210] | 662 | # |
---|
[221] | 663 | # @param path A list of Key names which represent an absolute path within |
---|
| 664 | # the Hive |
---|
| 665 | # |
---|
| 666 | # @return A @ref HiveIterator which is positioned at the specified path. |
---|
| 667 | # |
---|
| 668 | # @exception Exception If the path could not be found/traversed |
---|
[206] | 669 | def subtree(self, path): |
---|
| 670 | hi = HiveIterator(self) |
---|
| 671 | hi.descend(path) |
---|
| 672 | return hi |
---|
[205] | 673 | |
---|
[206] | 674 | |
---|
[210] | 675 | ## A special purpose iterator for registry hives |
---|
| 676 | # |
---|
| 677 | # Iterating over an object of this type causes all keys in a specific |
---|
| 678 | # hive subtree to be returned in a depth-first manner. These iterators |
---|
| 679 | # are typically created using the @ref Hive.subtree() function on a @ref Hive |
---|
| 680 | # object. |
---|
| 681 | # |
---|
| 682 | # HiveIterators can also be used to manually traverse up and down a |
---|
| 683 | # registry hive as they retain information about the current position in |
---|
| 684 | # the hive, along with which iteration state for subkeys and values for |
---|
| 685 | # every parent key. See the @ref up and @ref down methods for more |
---|
| 686 | # information. |
---|
[205] | 687 | class HiveIterator(): |
---|
[220] | 688 | _hive = None |
---|
| 689 | _iter = None |
---|
| 690 | _iteration_root = None |
---|
[205] | 691 | |
---|
| 692 | def __init__(self, hive): |
---|
[220] | 693 | self._iter = regfi.regfi_iterator_new(hive.file, REGFI_ENCODING_UTF8) |
---|
| 694 | if self._iter == None: |
---|
[205] | 695 | raise Exception("Could not create iterator. Current log:\n" |
---|
| 696 | + GetLogMessages()) |
---|
[214] | 697 | self._hive = hive |
---|
[205] | 698 | |
---|
| 699 | def __getattr__(self, name): |
---|
| 700 | return getattr(self.file.contents, name) |
---|
| 701 | |
---|
| 702 | def __del__(self): |
---|
[220] | 703 | regfi.regfi_iterator_free(self._iter) |
---|
[205] | 704 | |
---|
| 705 | def __iter__(self): |
---|
[220] | 706 | self._iteration_root = None |
---|
[205] | 707 | return self |
---|
| 708 | |
---|
| 709 | def __next__(self): |
---|
[220] | 710 | if self._iteration_root == None: |
---|
| 711 | self._iteration_root = self.current_key() |
---|
| 712 | elif not regfi.regfi_iterator_down(self._iter): |
---|
| 713 | up_ret = regfi.regfi_iterator_up(self._iter) |
---|
[206] | 714 | while (up_ret and |
---|
[220] | 715 | not regfi.regfi_iterator_next_subkey(self._iter)): |
---|
| 716 | if self._iteration_root == self.current_key(): |
---|
| 717 | self._iteration_root = None |
---|
[206] | 718 | raise StopIteration('') |
---|
[220] | 719 | up_ret = regfi.regfi_iterator_up(self._iter) |
---|
[205] | 720 | |
---|
| 721 | if not up_ret: |
---|
[221] | 722 | self._iteration_root = None |
---|
[205] | 723 | raise StopIteration('') |
---|
| 724 | |
---|
[210] | 725 | # XXX: Use non-generic exception |
---|
[220] | 726 | if not regfi.regfi_iterator_down(self._iter): |
---|
[205] | 727 | raise Exception('Error traversing iterator downward.'+ |
---|
| 728 | ' Current log:\n'+ GetLogMessages()) |
---|
| 729 | |
---|
[220] | 730 | regfi.regfi_iterator_first_subkey(self._iter) |
---|
[206] | 731 | return self.current_key() |
---|
[205] | 732 | |
---|
[212] | 733 | # For Python 2.x |
---|
[214] | 734 | next = __next__ |
---|
[212] | 735 | |
---|
[221] | 736 | # XXX: Should add sanity checks on some of these traversal functions |
---|
| 737 | # to throw exceptions if a traversal/retrieval *should* have worked |
---|
| 738 | # but failed for some reason. |
---|
| 739 | |
---|
| 740 | ## Descends the iterator to a subkey |
---|
| 741 | # |
---|
| 742 | # Descends the iterator one level to the current subkey, or a subkey |
---|
| 743 | # specified by name. |
---|
| 744 | # |
---|
| 745 | # @param subkey_name If specified, locates specified subkey by name |
---|
| 746 | # (via find_subkey()) and descends to it. |
---|
| 747 | # |
---|
| 748 | # @return True if successful, False otherwise |
---|
[220] | 749 | def down(self, subkey_name=None): |
---|
| 750 | if subkey_name == None: |
---|
| 751 | return regfi.regfi_iterator_down(self._iter) |
---|
| 752 | else: |
---|
| 753 | if name != None: |
---|
| 754 | name = name.encode('utf-8') |
---|
| 755 | return (regfi.regfi_iterator_find_subkey(self._iter, name) |
---|
| 756 | and regfi.regfi_iterator_down(self._iter)) |
---|
[206] | 757 | |
---|
[221] | 758 | |
---|
| 759 | ## Causes the iterator to ascend to the current Key's parent |
---|
| 760 | # |
---|
| 761 | # @return True if successful, False otherwise |
---|
| 762 | # |
---|
| 763 | # @note The state of current subkeys and values at this level in the tree |
---|
| 764 | # is lost as a side effect. That is, if you go up() and then back |
---|
| 765 | # down() again, current_subkey() and current_value() will return |
---|
| 766 | # default selections. |
---|
[206] | 767 | def up(self): |
---|
[220] | 768 | return regfi.regfi_iterator_up(self._iter) |
---|
[206] | 769 | |
---|
[221] | 770 | |
---|
| 771 | ## Selects first subkey of current key |
---|
| 772 | # |
---|
| 773 | # @return A Key instance for the first subkey. |
---|
| 774 | # None on error or if the current key has no subkeys. |
---|
[220] | 775 | def first_subkey(self): |
---|
| 776 | if regfi.regfi_iterator_first_subkey(self._iter): |
---|
| 777 | return self.current_subkey() |
---|
| 778 | return None |
---|
| 779 | |
---|
[221] | 780 | |
---|
| 781 | ## Selects first value of current Key |
---|
| 782 | # |
---|
| 783 | # @return A Value instance for the first value. |
---|
| 784 | # None on error or if the current key has no values. |
---|
[220] | 785 | def first_value(self): |
---|
| 786 | if regfi.regfi_iterator_first_value(self._iter): |
---|
| 787 | return self.current_value() |
---|
| 788 | return None |
---|
| 789 | |
---|
[221] | 790 | |
---|
| 791 | ## Selects the next subkey in the current Key's list |
---|
| 792 | # |
---|
| 793 | # @return A Key instance for the next subkey. |
---|
| 794 | # None if there are no remaining subkeys or an error occurred. |
---|
[220] | 795 | def next_subkey(self): |
---|
| 796 | if regfi.regfi_iterator_next_subkey(self._iter): |
---|
| 797 | return self.current_subkey() |
---|
| 798 | return None |
---|
| 799 | |
---|
[221] | 800 | |
---|
| 801 | ## Selects the next value in the current Key's list |
---|
| 802 | # |
---|
| 803 | # @return A Value instance for the next value. |
---|
| 804 | # None if there are no remaining values or an error occurred. |
---|
[220] | 805 | def next_value(self): |
---|
| 806 | if regfi.regfi_iterator_next_value(self._iter): |
---|
| 807 | return self.current_value() |
---|
| 808 | return None |
---|
| 809 | |
---|
[221] | 810 | |
---|
| 811 | ## Selects the first subkey which has the specified name |
---|
| 812 | # |
---|
| 813 | # @return A Key instance for the selected key. |
---|
| 814 | # None if it could not be located or an error occurred. |
---|
[220] | 815 | def find_subkey(self, name): |
---|
| 816 | if name != None: |
---|
| 817 | name = name.encode('utf-8') |
---|
| 818 | if regfi.regfi_iterator_find_subkey(self._iter, name): |
---|
| 819 | return self.current_subkey() |
---|
| 820 | return None |
---|
| 821 | |
---|
[221] | 822 | |
---|
| 823 | ## Selects the first value which has the specified name |
---|
| 824 | # |
---|
| 825 | # @return A Value instance for the selected value. |
---|
| 826 | # None if it could not be located or an error occurred. |
---|
[220] | 827 | def find_value(self, name): |
---|
| 828 | if name != None: |
---|
| 829 | name = name.encode('utf-8') |
---|
| 830 | if regfi.regfi_iterator_find_value(self._iter, name): |
---|
| 831 | return self.current_value() |
---|
| 832 | return None |
---|
| 833 | |
---|
[221] | 834 | ## Retrieves the currently selected subkey |
---|
| 835 | # |
---|
| 836 | # @return A Key instance of the current subkey |
---|
[220] | 837 | def current_subkey(self): |
---|
| 838 | return Key(self._hive, regfi.regfi_iterator_cur_subkey(self._iter)) |
---|
| 839 | |
---|
[221] | 840 | ## Retrieves the currently selected value |
---|
| 841 | # |
---|
| 842 | # @return A Value instance of the current value |
---|
[220] | 843 | def current_value(self): |
---|
| 844 | return Value(self._hive, regfi.regfi_iterator_cur_value(self._iter)) |
---|
| 845 | |
---|
[221] | 846 | ## Retrieves the current key |
---|
| 847 | # |
---|
| 848 | # @return A Key instance of the current position of the iterator |
---|
[220] | 849 | def current_key(self): |
---|
| 850 | return Key(self._hive, regfi.regfi_iterator_cur_key(self._iter)) |
---|
| 851 | |
---|
[221] | 852 | |
---|
| 853 | ## Traverse downward multiple levels |
---|
| 854 | # |
---|
| 855 | # This is more efficient than calling down() multiple times |
---|
| 856 | # |
---|
| 857 | # @param path A list of Key names which represent the path to descend |
---|
| 858 | # |
---|
| 859 | # @exception Exception If path could not be located |
---|
[206] | 860 | def descend(self, path): |
---|
[215] | 861 | cpath = _strlist2charss(path) |
---|
[206] | 862 | |
---|
[210] | 863 | # XXX: Use non-generic exception |
---|
[220] | 864 | if not regfi.regfi_iterator_walk_path(self._iter, cpath): |
---|
[206] | 865 | raise Exception('Could not locate path.\n'+GetLogMessages()) |
---|
[221] | 866 | |
---|
| 867 | |
---|
| 868 | # Freeing symbols defined for the sake of documentation |
---|
| 869 | del Value.name,Value.name_raw,Value.offset,Value.data_size,Value.type,Value.flags |
---|
| 870 | del Key.name,Key.name_raw,Key.offset,Key.modified,Key.flags |
---|
| 871 | del Hive.root,Hive.modified,Hive.sequence1,Hive.sequence2,Hive.major_version,Hive.minor_version |
---|