Commit | Line | Data |
---|---|---|
00b3cd10 PG |
1 | #!/usr/bin/env python3 |
2 | ||
3 | """ | |
83f2d71e | 4 | Intra2net 2017 |
00b3cd10 PG |
5 | |
6 | =============================================================================== | |
704ceaa5 | 7 | crypto -- Encryption Layer for the Deltatar Backup |
00b3cd10 PG |
8 | =============================================================================== |
9 | ||
10 | Crypto stack: | |
11 | ||
12 | - AES-GCM for the symmetric encryption; | |
13 | - Scrypt as KDF. | |
14 | ||
15 | References: | |
16 | ||
17 | - NIST Recommendation for Block Cipher Modes of Operation: Galois/Counter | |
18 | Mode (GCM) and GMAC | |
19 | http://nvlpubs.nist.gov/nistpubs/Legacy/SP/nistspecialpublication800-38d.pdf | |
20 | ||
21 | - AES-GCM v1: | |
22 | https://cryptome.org/2014/01/aes-gcm-v1.pdf | |
23 | ||
24 | - Authentication weaknesses in GCM | |
25 | http://csrc.nist.gov/groups/ST/toolkit/BCM/documents/comments/CWC-GCM/Ferguson2.pdf | |
26 | ||
83f2d71e PG |
27 | Trouble with python-cryptography packages: authentication tags can only be |
28 | passed in advance: https://github.com/pyca/cryptography/pull/3421 | |
29 | ||
6d08915c PG |
30 | Errors |
31 | ------------------------------------------------------------------------------- | |
32 | ||
33 | Errors fall into roughly three categories: | |
34 | ||
704ceaa5 | 35 | - Cryptographical errors or invalid data. |
6d08915c PG |
36 | |
37 | - ``InvalidGCMTag`` (decryption failed on account of an invalid GCM | |
38 | tag), | |
39 | - ``InvalidIVFixedPart`` (IV fixed part of object not found in list), | |
f6cd676f | 40 | - ``DuplicateIV`` (the IV of an encrypted object already occurred), |
704ceaa5 PG |
41 | - ``DecryptionError`` (used in CLI decryption for presenting error |
42 | conditions to the user). | |
6d08915c PG |
43 | |
44 | - Incorrect usage of the library. | |
45 | ||
46 | - ``InvalidParameter`` (non-conforming user supplied parameter), | |
47 | - ``InvalidHeader`` (data passed for reading not parsable into header), | |
48 | - ``FormatError`` (cannot handle header or parameter version), | |
49 | - ``RuntimeError``. | |
50 | ||
51 | - Bad internal state. If one of these is encountered it means that a state | |
52 | was reached that shouldn’t occur during normal processing. | |
53 | ||
54 | - ``InternalError``, | |
55 | - ``Unreachable``. | |
56 | ||
57 | Also, ``EndOfFile`` is used as a sentinel to communicate that a stream supplied | |
58 | for reading is exhausted. | |
59 | ||
f6cd676f PG |
60 | Initialization Vectors |
61 | ------------------------------------------------------------------------------- | |
62 | ||
63 | Initialization vectors are checked reuse during the lifetime of a decryptor. | |
704ceaa5 PG |
64 | The fixed counters for metadata files cannot be reused and attempts to do so |
65 | will cause a DuplicateIV error. This means the length of objects encrypted with | |
66 | a metadata counter is capped at 63 GB. | |
67 | ||
68 | For ordinary, non-metadata payload, there is an optional mode with strict IV | |
69 | checking that causes a crypto context to fail if an IV encountered or created | |
70 | was already used for decrypting or encrypting, respectively, an earlier object. | |
71 | Note that this mode can trigger false positives when decrypting non-linearly, | |
72 | e. g. when traversing the same object multiple times. Since the crypto context | |
73 | has no notion of a position in a PDT encrypted archive, this condition must be | |
74 | sorted out downstream. | |
75 | ||
76 | Command Line Utility | |
77 | ------------------------------------------------------------------------------- | |
78 | ||
79 | ``crypto.py`` may be invoked as a script for decrypting, validating, and | |
80 | splitting PDT encrypted files. Consult the usage message for details. | |
81 | ||
82 | Usage examples: | |
83 | ||
84 | Decrypt from stdin using the password ‘foo’: :: | |
85 | ||
86 | $ crypto.py process foo -i - -o - <some-file.tar.gz.pdtcrypt >some-file.tar.gz | |
87 | ||
88 | Output verbose information about the encrypted objects in the archive: :: | |
89 | ||
90 | $ crypto.py process foo -v -i some-file.tar.gz.pdtcrypt -o /dev/null | |
91 | PDT: decrypt from some-file.tar.gz.pdtcrypt | |
92 | PDT: decrypt to /dev/null | |
93 | PDT: source: file some-file.tar.gz.pdtcrypt | |
94 | PDT: sink: file /dev/null | |
95 | PDT: 0 hdr | |
96 | PDT: · version = 1 : 0100 | |
97 | PDT: · paramversion = 1 : 0100 | |
98 | PDT: · nacl : d270 b031 00d1 87e2 c946 610d 7b7f 7e5f | |
99 | PDT: · iv : 02ee 3dd7 a963 1eb1 0100 0000 | |
100 | PDT: · ctsize = 591 : 4f02 0000 0000 0000 | |
101 | PDT: · tag : 5b2d 6d8b 8f82 4842 12fd 0b10 b6e3 369b | |
102 | PDT: 64 decrypt obj no. 1, 591 B | |
103 | PDT: · [64] 0% done, read block (591 B of 591 B remaining) | |
104 | PDT: · decrypt ciphertext 591 B | |
105 | PDT: · decrypt plaintext 591 B | |
106 | PDT: 655 finalize | |
107 | … | |
108 | ||
109 | Also, the mode *scrypt* allows deriving encryption keys. To calculate the | |
110 | encryption key from the password ‘foo’ and the salt of the first object in a | |
111 | PDT encrypted file: :: | |
112 | ||
113 | $ crypto.py scrypt foo -i some-file.pdtcrypt | |
4f6405d6 | 114 | {"paramversion": 1, "salt": "Cqzbk48e3peEjzWto8D0yA==", "key": "JH9EkMwaM4x9F5aim5gK/Q=="} |
704ceaa5 PG |
115 | |
116 | The computed 16 byte key is given in hexadecimal notation in the value to | |
117 | ``hash`` and can be fed into Python’s ``binascii.unhexlify()`` to obtain the | |
118 | corresponding binary representation. | |
119 | ||
120 | Note that in Scrypt hashing mode, no data integrity checks are being performed. | |
121 | If the wrong password is given, a wrong key will be derived. Whether the password | |
122 | was indeed correct can only be determined by decrypting. Note that since PDT | |
123 | archives essentially consist of a stream of independent objects, the salt and | |
124 | other parameters may change. Thus a key derived using above method from the | |
125 | first object doesn’t necessarily apply to any of the subsequent objects. | |
f6cd676f | 126 | |
00b3cd10 PG |
127 | """ |
128 | ||
7b3940e5 | 129 | import base64 |
00b3cd10 | 130 | import binascii |
50710d86 | 131 | import bisect |
00b3cd10 PG |
132 | import ctypes |
133 | import io | |
c46c8670 | 134 | from functools import reduce, partial |
00b3cd10 PG |
135 | import os |
136 | import struct | |
137 | import sys | |
138 | import time | |
da82bc58 | 139 | import types |
00b3cd10 PG |
140 | try: |
141 | import enum34 | |
142 | except ImportError as exn: | |
143 | pass | |
144 | ||
145 | if __name__ == "__main__": ## Work around the import mechanism’s lest Python’s | |
146 | pwd = os.getcwd() ## preference for local imports causes a cyclical | |
147 | ## import (crypto → pylibscrypt → […] → ./tarfile → crypto). | |
148 | sys.path = [ p for p in sys.path if p.find ("deltatar") < 0 ] | |
149 | ||
150 | import pylibscrypt | |
151 | from cryptography.hazmat.primitives.ciphers import Cipher, algorithms, modes | |
152 | from cryptography.hazmat.backends import default_backend | |
15d3eefd | 153 | import cryptography |
00b3cd10 PG |
154 | |
155 | ||
a64085a8 | 156 | __all__ = [ "hdr_make", "hdr_read", "hdr_fmt", "hdr_fmt_pretty" |
b360b772 | 157 | , "scrypt_hashfile" |
3031b7ae PG |
158 | , "PDTCRYPT_HDR_SIZE", "AES_GCM_IV_CNT_DATA" |
159 | , "AES_GCM_IV_CNT_INFOFILE", "AES_GCM_IV_CNT_INDEX" | |
2d6fd8c8 | 160 | ] |
00b3cd10 | 161 | |
a393d9cb PG |
162 | |
163 | ############################################################################### | |
15d3eefd PG |
164 | ## exceptions |
165 | ############################################################################### | |
166 | ||
167 | class EndOfFile (Exception): | |
168 | """Reached EOF.""" | |
ae3d0f2a PG |
169 | remainder = 0 |
170 | msg = 0 | |
8a8ac469 | 171 | def __init__ (self, n=None, msg=None): |
5d394c0d PG |
172 | if n is not None: |
173 | self.remainder = n | |
174 | self.msg = msg | |
15d3eefd | 175 | |
b0078f26 | 176 | |
b12110dd PG |
177 | class InvalidParameter (Exception): |
178 | """Inputs not valid for PDT encryption.""" | |
179 | pass | |
180 | ||
b0078f26 | 181 | |
15d3eefd PG |
182 | class InvalidHeader (Exception): |
183 | """Header not valid.""" | |
184 | pass | |
185 | ||
b0078f26 PG |
186 | |
187 | class InvalidGCMTag (Exception): | |
188 | """ | |
189 | The GCM tag calculated during decryption differs from that in the object | |
190 | header. | |
191 | """ | |
192 | pass | |
193 | ||
194 | ||
26b42ad4 | 195 | class InvalidIVFixedPart (Exception): |
89ec6e2f PG |
196 | """ |
197 | IV fixed part not in supplied list: either the backup is corrupt or the | |
198 | current object does not belong to it. | |
199 | """ | |
26b42ad4 PG |
200 | pass |
201 | ||
b0078f26 | 202 | |
be124bca | 203 | class IVFixedPartError (Exception): |
89ec6e2f PG |
204 | """ |
205 | Error creating a unique IV fixed part: repeated calls to system RNG yielded | |
206 | the same sequence of bytes as the last IV used. | |
207 | """ | |
be124bca PG |
208 | pass |
209 | ||
210 | ||
fac2cfe1 | 211 | class InvalidFileCounter (Exception): |
89ec6e2f PG |
212 | """ |
213 | When encrypting, an attempted reuse of a dedicated counter (info file, | |
214 | index file) was caught. | |
215 | """ | |
fac2cfe1 PG |
216 | pass |
217 | ||
218 | ||
ee6aa239 | 219 | class DuplicateIV (Exception): |
89ec6e2f PG |
220 | """ |
221 | During encryption, the current IV fixed part is identical to an already | |
222 | existing IV (same prefix and file counter). This indicates tampering or | |
223 | programmer error and cannot be recovered from. | |
224 | """ | |
ee6aa239 PG |
225 | pass |
226 | ||
227 | ||
228 | class NonConsecutiveIV (Exception): | |
89ec6e2f PG |
229 | """ |
230 | IVs not numbered consecutively. This is a hard error with strict IV | |
231 | checking. Precludes random access to the encrypted objects. | |
232 | """ | |
ee6aa239 PG |
233 | pass |
234 | ||
235 | ||
b12110dd PG |
236 | class FormatError (Exception): |
237 | """Unusable parameters in header.""" | |
238 | pass | |
239 | ||
b0078f26 | 240 | |
15d3eefd | 241 | class DecryptionError (Exception): |
89ec6e2f | 242 | """Error during decryption with ``crypto.py`` on the command line.""" |
15d3eefd PG |
243 | pass |
244 | ||
b0078f26 | 245 | |
70ad9458 | 246 | class Unreachable (Exception): |
89ec6e2f PG |
247 | """ |
248 | Makeshift __builtin_unreachable(); always a programmer error if | |
249 | thrown. | |
250 | """ | |
70ad9458 PG |
251 | pass |
252 | ||
b0078f26 | 253 | |
b12110dd PG |
254 | class InternalError (Exception): |
255 | """Errors not ascribable to bad user inputs or cryptography.""" | |
256 | pass | |
257 | ||
15d3eefd PG |
258 | |
259 | ############################################################################### | |
a393d9cb PG |
260 | ## crypto layer version |
261 | ############################################################################### | |
262 | ||
263 | ENCRYPTION_PARAMETERS = \ | |
c46c8670 | 264 | { 0: \ |
dd23cbc9 PG |
265 | { "kdf": ("dummy", 16) |
266 | , "enc": "passthrough" } | |
c46c8670 | 267 | , 1: \ |
dd23cbc9 PG |
268 | { "kdf": ( "scrypt" |
269 | , { "dkLen" : 16 | |
270 | , "N" : 1 << 16 | |
271 | , "r" : 8 | |
272 | , "p" : 1 | |
273 | , "NaCl_LEN" : 16 }) | |
274 | , "enc": "aes-gcm" } } | |
a393d9cb | 275 | |
00b3cd10 PG |
276 | ############################################################################### |
277 | ## constants | |
278 | ############################################################################### | |
279 | ||
dd47d6a2 | 280 | PDTCRYPT_HDR_MAGIC = b"PDTCRYPT" |
00b3cd10 | 281 | |
dd47d6a2 PG |
282 | PDTCRYPT_HDR_SIZE_MAGIC = 8 # 8 |
283 | PDTCRYPT_HDR_SIZE_VERSION = 2 # 10 | |
284 | PDTCRYPT_HDR_SIZE_PARAMVERSION = 2 # 12 | |
285 | PDTCRYPT_HDR_SIZE_NACL = 16 # 28 | |
286 | PDTCRYPT_HDR_SIZE_IV = 12 # 40 | |
287 | PDTCRYPT_HDR_SIZE_CTSIZE = 8 # 48 | |
288 | PDTCRYPT_HDR_SIZE_TAG = 16 # 64 GCM auth tag | |
00b3cd10 | 289 | |
dd47d6a2 PG |
290 | PDTCRYPT_HDR_SIZE = PDTCRYPT_HDR_SIZE_MAGIC + PDTCRYPT_HDR_SIZE_VERSION \ |
291 | + PDTCRYPT_HDR_SIZE_PARAMVERSION + PDTCRYPT_HDR_SIZE_NACL \ | |
292 | + PDTCRYPT_HDR_SIZE_IV + PDTCRYPT_HDR_SIZE_CTSIZE \ | |
293 | + PDTCRYPT_HDR_SIZE_TAG # = 64 | |
00b3cd10 PG |
294 | |
295 | # precalculate offsets since Python can’t do constant folding over names | |
dd47d6a2 PG |
296 | HDR_OFF_VERSION = PDTCRYPT_HDR_SIZE_MAGIC |
297 | HDR_OFF_PARAMVERSION = HDR_OFF_VERSION + PDTCRYPT_HDR_SIZE_VERSION | |
298 | HDR_OFF_NACL = HDR_OFF_PARAMVERSION + PDTCRYPT_HDR_SIZE_PARAMVERSION | |
299 | HDR_OFF_IV = HDR_OFF_NACL + PDTCRYPT_HDR_SIZE_NACL | |
300 | HDR_OFF_CTSIZE = HDR_OFF_IV + PDTCRYPT_HDR_SIZE_IV | |
301 | HDR_OFF_TAG = HDR_OFF_CTSIZE + PDTCRYPT_HDR_SIZE_CTSIZE | |
00b3cd10 PG |
302 | |
303 | FMT_UINT16_LE = "<H" | |
304 | FMT_UINT64_LE = "<Q" | |
50710d86 | 305 | FMT_I2N_IV = "<8sL" # 8 random bytes ‖ 32 bit counter |
83f2d71e PG |
306 | FMT_I2N_HDR = ("<" # host byte order |
307 | "8s" # magic | |
308 | "H" # version | |
309 | "H" # paramversion | |
310 | "16s" # sodium chloride | |
311 | "12s" # iv | |
3b53fb98 PG |
312 | "Q" # size |
313 | "16s") # GCM tag | |
00b3cd10 PG |
314 | |
315 | # aes+gcm | |
cb7a3911 PG |
316 | AES_GCM_MAX_SIZE = (1 << 36) - (1 << 5) # 2^39 - 2^8 b ≅ 64 GB |
317 | PDTCRYPT_MAX_OBJ_SIZE_DEFAULT = 63 * (1 << 30) # 63 GB | |
318 | PDTCRYPT_MAX_OBJ_SIZE = PDTCRYPT_MAX_OBJ_SIZE_DEFAULT | |
00b3cd10 | 319 | |
3031b7ae PG |
320 | # index and info files are written on-the fly while encrypting so their |
321 | # counters must be available inadvance | |
cb7a3911 PG |
322 | AES_GCM_IV_CNT_INFOFILE = 1 # constant |
323 | AES_GCM_IV_CNT_INDEX = AES_GCM_IV_CNT_INFOFILE + 1 | |
324 | AES_GCM_IV_CNT_DATA = AES_GCM_IV_CNT_INDEX + 1 # also for multivolume | |
325 | AES_GCM_IV_CNT_MAX_DEFAULT = 0xffFFffFF | |
326 | AES_GCM_IV_CNT_MAX = AES_GCM_IV_CNT_MAX_DEFAULT | |
2d6fd8c8 | 327 | |
be124bca PG |
328 | # IV structure and generation |
329 | PDTCRYPT_IV_GEN_MAX_RETRIES = 10 # × | |
330 | PDTCRYPT_IV_FIXEDPART_SIZE = 8 # B | |
331 | PDTCRYPT_IV_COUNTER_SIZE = 4 # B | |
39accaaa | 332 | |
00b3cd10 | 333 | ############################################################################### |
39accaaa | 334 | ## header, trailer |
00b3cd10 PG |
335 | ############################################################################### |
336 | # | |
337 | # Interface: | |
338 | # | |
339 | # struct hdrinfo | |
340 | # { version : u16 | |
341 | # , paramversion : u16 | |
342 | # , nacl : [u8; 16] | |
343 | # , iv : [u8; 12] | |
704ceaa5 PG |
344 | # , ctsize : usize |
345 | # , tag : [u8; 16] } | |
83f2d71e | 346 | # |
00b3cd10 | 347 | # fn hdr_read (f : handle) -> hdrinfo; |
c2d1c3ec | 348 | # fn hdr_make (f : handle, h : hdrinfo) -> IOResult<usize>; |
00b3cd10 PG |
349 | # fn hdr_fmt (h : hdrinfo) -> String; |
350 | # | |
351 | ||
83f2d71e | 352 | def hdr_read (data): |
704ceaa5 PG |
353 | """ |
354 | Read bytes as header structure. | |
355 | ||
356 | If the input could not be interpreted as a header, fail with | |
357 | ``InvalidHeader``. | |
358 | """ | |
83f2d71e | 359 | |
00b3cd10 | 360 | try: |
3b53fb98 | 361 | mag, version, paramversion, nacl, iv, ctsize, tag = \ |
83f2d71e PG |
362 | struct.unpack (FMT_I2N_HDR, data) |
363 | except Exception as exn: | |
15d3eefd PG |
364 | raise InvalidHeader ("error unpacking header from [%r]: %s" |
365 | % (binascii.hexlify (data), str (exn))) | |
00b3cd10 | 366 | |
dd47d6a2 | 367 | if mag != PDTCRYPT_HDR_MAGIC: |
15d3eefd | 368 | raise InvalidHeader ("bad magic in header: expected [%s], got [%s]" |
dd47d6a2 | 369 | % (PDTCRYPT_HDR_MAGIC, mag)) |
00b3cd10 | 370 | |
15d3eefd | 371 | return \ |
00b3cd10 PG |
372 | { "version" : version |
373 | , "paramversion" : paramversion | |
374 | , "nacl" : nacl | |
375 | , "iv" : iv | |
376 | , "ctsize" : ctsize | |
3b53fb98 | 377 | , "tag" : tag |
00b3cd10 PG |
378 | } |
379 | ||
380 | ||
39accaaa | 381 | def hdr_read_stream (instr): |
704ceaa5 PG |
382 | """ |
383 | Read header from stream at the current position. | |
384 | ||
385 | Fail with ``InvalidHeader`` if insufficient bytes were read from the | |
386 | stream, or if the content could not be interpreted as a header. | |
387 | """ | |
dd47d6a2 | 388 | data = instr.read(PDTCRYPT_HDR_SIZE) |
ae3d0f2a | 389 | ldata = len (data) |
8a8ac469 PG |
390 | if ldata == 0: |
391 | raise EndOfFile | |
392 | elif ldata != PDTCRYPT_HDR_SIZE: | |
393 | raise InvalidHeader ("hdr_read_stream: expected %d B, received %d B" | |
394 | % (PDTCRYPT_HDR_SIZE, ldata)) | |
47e27926 | 395 | return hdr_read (data) |
39accaaa PG |
396 | |
397 | ||
3b53fb98 | 398 | def hdr_from_params (version, paramversion, nacl, iv, ctsize, tag): |
704ceaa5 PG |
399 | """ |
400 | Assemble the necessary values into a PDTCRYPT header. | |
401 | ||
402 | :type version: int to fit uint16_t | |
403 | :type paramversion: int to fit uint16_t | |
404 | :type nacl: bytes to fit uint8_t[16] | |
405 | :type iv: bytes to fit uint8_t[12] | |
406 | :type size: int to fit uint64_t | |
407 | :type tag: bytes to fit uint8_t[16] | |
408 | """ | |
dd47d6a2 | 409 | buf = bytearray (PDTCRYPT_HDR_SIZE) |
83f2d71e | 410 | bufv = memoryview (buf) |
00b3cd10 | 411 | |
00b3cd10 | 412 | try: |
83f2d71e | 413 | struct.pack_into (FMT_I2N_HDR, bufv, 0, |
dd47d6a2 | 414 | PDTCRYPT_HDR_MAGIC, |
3b53fb98 | 415 | version, paramversion, nacl, iv, ctsize, tag) |
83f2d71e | 416 | except Exception as exn: |
a83fa4ed | 417 | return False, "error assembling header: %s" % str (exn) |
00b3cd10 | 418 | |
83f2d71e | 419 | return True, bytes (buf) |
00b3cd10 | 420 | |
00b3cd10 | 421 | |
8a990744 PG |
422 | def hdr_make_dummy (s): |
423 | """ | |
424 | Create a header sized block of bytes initialized to a value derived from a | |
425 | string. Used to verify we’ve jumped back correctly to the actual position | |
426 | of the object header. | |
427 | """ | |
428 | c = reduce (lambda a, c: a + ord(c), s, 0) % 0xFF | |
dd47d6a2 | 429 | return bytes (bytearray (struct.pack ("B", c)) * PDTCRYPT_HDR_SIZE) |
8a990744 PG |
430 | |
431 | ||
a393d9cb | 432 | def hdr_make (hdr): |
704ceaa5 PG |
433 | """ |
434 | Assemble a header from the given header structure. | |
435 | """ | |
a393d9cb PG |
436 | return hdr_from_params (version=hdr.get("version"), |
437 | paramversion=hdr.get("paramversion"), | |
438 | nacl=hdr.get("nacl"), iv=hdr.get("iv"), | |
3b53fb98 | 439 | ctsize=hdr.get("ctsize"), tag=hdr.get("tag")) |
a393d9cb PG |
440 | |
441 | ||
83f2d71e | 442 | HDR_FMT = "I2n_header { version: %d, paramversion: %d, nacl: %s[%d]," \ |
89131745 | 443 | " iv: %s[%d], ctsize: %d, tag: %s[%d] }" |
00b3cd10 | 444 | |
83f2d71e | 445 | def hdr_fmt (h): |
704ceaa5 | 446 | """Format a header structure into readable output.""" |
83f2d71e PG |
447 | return HDR_FMT % (h["version"], h["paramversion"], |
448 | binascii.hexlify (h["nacl"]), len(h["nacl"]), | |
449 | binascii.hexlify (h["iv"]), len(h["iv"]), | |
db1f3ac7 PG |
450 | h["ctsize"], |
451 | binascii.hexlify (h["tag"]), len(h["tag"])) | |
00b3cd10 | 452 | |
00b3cd10 | 453 | |
83f2d71e | 454 | def hex_spaced_of_bytes (b): |
704ceaa5 | 455 | """Format bytes object, hexdump style.""" |
83f2d71e PG |
456 | return " ".join ([ "%.2x%.2x" % (c1, c2) |
457 | for c1, c2 in zip (b[0::2], b[1::2]) ]) \ | |
458 | + (len (b) | 1 == len (b) and " %.2x" % b[-1] or "") # odd lengths | |
00b3cd10 | 459 | |
591a722f | 460 | |
3031b7ae PG |
461 | def hdr_iv_counter (h): |
462 | """Extract the variable part of the IV of the given header.""" | |
463 | _fixed, cnt = struct.unpack (FMT_I2N_IV, h ["iv"]) | |
464 | return cnt | |
465 | ||
466 | ||
467 | def hdr_iv_fixed (h): | |
468 | """Extract the fixed part of the IV of the given header.""" | |
469 | fixed, _cnt = struct.unpack (FMT_I2N_IV, h ["iv"]) | |
470 | return fixed | |
471 | ||
472 | ||
83f2d71e | 473 | hdr_dump = hex_spaced_of_bytes |
00b3cd10 | 474 | |
00b3cd10 | 475 | |
15d3eefd PG |
476 | HDR_FMT_PRETTY = \ |
477 | """version = %-4d : %s | |
478 | paramversion = %-4d : %s | |
479 | nacl : %s | |
480 | iv : %s | |
481 | ctsize = %-20d : %s | |
482 | tag : %s | |
83f2d71e | 483 | """ |
00b3cd10 | 484 | |
83f2d71e | 485 | def hdr_fmt_pretty (h): |
704ceaa5 PG |
486 | """ |
487 | Format header structure into multi-line representation of its contents and | |
488 | their raw representation. (Omit the implicit “PDTCRYPT” magic bytes that | |
489 | precede every header.) | |
490 | """ | |
83f2d71e PG |
491 | return HDR_FMT_PRETTY \ |
492 | % (h["version"], | |
493 | hex_spaced_of_bytes (struct.pack (FMT_UINT16_LE, h["version"])), | |
494 | h["paramversion"], | |
495 | hex_spaced_of_bytes (struct.pack (FMT_UINT16_LE, h["paramversion"])), | |
496 | hex_spaced_of_bytes (h["nacl"]), | |
497 | hex_spaced_of_bytes (h["iv"]), | |
498 | h["ctsize"], | |
15d3eefd PG |
499 | hex_spaced_of_bytes (struct.pack (FMT_UINT64_LE, h["ctsize"])), |
500 | hex_spaced_of_bytes (h["tag"])) | |
00b3cd10 | 501 | |
f6cd676f PG |
502 | IV_FMT = "((f %s) (c %d))" |
503 | ||
504 | def iv_fmt (iv): | |
704ceaa5 | 505 | """Format the two components of an IV in a readable fashion.""" |
f6cd676f PG |
506 | fixed, cnt = struct.unpack (FMT_I2N_IV, iv) |
507 | return IV_FMT % (binascii.hexlify (fixed), cnt) | |
508 | ||
00b3cd10 | 509 | |
00b3cd10 | 510 | ############################################################################### |
6178061e PG |
511 | ## passthrough / null encryption |
512 | ############################################################################### | |
513 | ||
514 | class PassthroughCipher (object): | |
515 | ||
516 | tag = struct.pack ("<QQ", 0, 0) | |
517 | ||
518 | def __init__ (self) : pass | |
519 | ||
520 | def update (self, b) : return b | |
521 | ||
50710d86 | 522 | def finalize (self) : return b"" |
6178061e PG |
523 | |
524 | def finalize_with_tag (self, _) : return b"" | |
525 | ||
526 | ############################################################################### | |
a393d9cb | 527 | ## convenience wrapper |
00b3cd10 PG |
528 | ############################################################################### |
529 | ||
c46c8670 PG |
530 | |
531 | def kdf_dummy (klen, password, _nacl): | |
704ceaa5 PG |
532 | """ |
533 | Fake KDF for testing purposes that is called when parameter version zero is | |
534 | encountered. | |
535 | """ | |
c46c8670 PG |
536 | q, r = divmod (klen, len (password)) |
537 | if isinstance (password, bytes) is False: | |
538 | password = password.encode () | |
539 | return password * q + password [:r], b"" | |
540 | ||
541 | ||
542 | SCRYPT_KEY_MEMO = { } # static because needed for both the info file and the archive | |
543 | ||
544 | ||
545 | def kdf_scrypt (params, password, nacl): | |
704ceaa5 PG |
546 | """ |
547 | Wrapper for the Scrypt KDF, corresponds to parameter version one. The | |
548 | computation result is memoized based on the inputs to facilitate spawning | |
549 | multiple encryption contexts. | |
550 | """ | |
c46c8670 PG |
551 | N = params["N"] |
552 | r = params["r"] | |
553 | p = params["p"] | |
554 | dkLen = params["dkLen"] | |
555 | ||
556 | if nacl is None: | |
557 | nacl = os.urandom (params["NaCl_LEN"]) | |
558 | ||
559 | key_parms = (password, nacl, N, r, p, dkLen) | |
560 | global SCRYPT_KEY_MEMO | |
561 | if key_parms not in SCRYPT_KEY_MEMO: | |
562 | SCRYPT_KEY_MEMO [key_parms] = \ | |
563 | pylibscrypt.scrypt (password, nacl, N, r, p, dkLen) | |
564 | return SCRYPT_KEY_MEMO [key_parms], nacl | |
a64085a8 PG |
565 | |
566 | ||
da82bc58 | 567 | def kdf_by_version (paramversion=None, defs=None): |
704ceaa5 PG |
568 | """ |
569 | Pick the KDF handler corresponding to the parameter version or the | |
570 | definition set. | |
571 | ||
572 | :rtype: function (password : str, nacl : str) -> str | |
573 | """ | |
da82bc58 PG |
574 | if paramversion is not None: |
575 | defs = ENCRYPTION_PARAMETERS.get(paramversion, None) | |
a64085a8 | 576 | if defs is None: |
1ed44e7b PG |
577 | raise InvalidParameter ("no encryption parameters for version %r" |
578 | % paramversion) | |
a64085a8 | 579 | (kdf, params) = defs["kdf"] |
c46c8670 PG |
580 | fn = None |
581 | if kdf == "scrypt" : fn = kdf_scrypt | |
582 | if kdf == "dummy" : fn = kdf_dummy | |
583 | if fn is None: | |
a64085a8 | 584 | raise ValueError ("key derivation method %r unknown" % kdf) |
c46c8670 | 585 | return partial (fn, params) |
a64085a8 PG |
586 | |
587 | ||
b360b772 PG |
588 | ############################################################################### |
589 | ## SCRYPT hashing | |
590 | ############################################################################### | |
591 | ||
592 | def scrypt_hashsource (pw, ins): | |
593 | """ | |
594 | Calculate the SCRYPT hash from the password and the information contained | |
595 | in the first header found in ``ins``. | |
596 | ||
597 | This does not validate whether the first object is encrypted correctly. | |
598 | """ | |
c1ecc2e2 PG |
599 | if isinstance (pw, str) is True: |
600 | pw = str.encode (pw) | |
601 | elif isinstance (pw, bytes) is False: | |
602 | raise InvalidParameter ("password must be a string, not %s" | |
603 | % type (password)) | |
604 | if isinstance (ins, io.BufferedReader) is False and \ | |
605 | isinstance (ins, io.FileIO) is False: | |
606 | raise InvalidParameter ("file to hash must be opened in “binary” mode") | |
b360b772 PG |
607 | hdr = None |
608 | try: | |
609 | hdr = hdr_read_stream (ins) | |
610 | except EndOfFile as exn: | |
611 | noise ("PDT: malformed input: end of file reading first object header") | |
612 | noise ("PDT:") | |
613 | return 1 | |
614 | ||
615 | nacl = hdr ["nacl"] | |
616 | pver = hdr ["paramversion"] | |
617 | if PDTCRYPT_VERBOSE is True: | |
618 | noise ("PDT: salt of first object : %s" % binascii.hexlify (nacl)) | |
619 | noise ("PDT: parameter version of archive : %d" % pver) | |
620 | ||
621 | try: | |
622 | defs = ENCRYPTION_PARAMETERS.get(pver, None) | |
623 | kdfname, params = defs ["kdf"] | |
624 | if kdfname != "scrypt": | |
625 | noise ("PDT: input is not an SCRYPT archive") | |
626 | noise ("") | |
627 | return 1 | |
628 | kdf = kdf_by_version (None, defs) | |
629 | except ValueError as exn: | |
630 | noise ("PDT: object has unknown parameter version %d" % pver) | |
631 | ||
632 | hsh, _void = kdf (pw, nacl) | |
633 | ||
c1ecc2e2 | 634 | return hsh, nacl, hdr ["version"], pver |
b360b772 PG |
635 | |
636 | ||
637 | def scrypt_hashfile (pw, fname): | |
704ceaa5 PG |
638 | """ |
639 | Calculate the SCRYPT hash from the password and the information contained | |
640 | in the first header found in the given file. The header is read only at | |
641 | offset zero. | |
642 | """ | |
b360b772 | 643 | with deptdcrypt_mk_stream (PDTCRYPT_SOURCE, fname or "-") as ins: |
c1ecc2e2 | 644 | hsh, _void, _void, _void = scrypt_hashsource (pw, ins) |
b360b772 PG |
645 | return hsh |
646 | ||
647 | ||
648 | ############################################################################### | |
649 | ## AES-GCM context | |
650 | ############################################################################### | |
651 | ||
a393d9cb PG |
652 | class Crypto (object): |
653 | """ | |
654 | Encryption context to remain alive throughout an entire tarfile pass. | |
655 | """ | |
6178061e | 656 | enc = None |
a393d9cb PG |
657 | nacl = None |
658 | key = None | |
50710d86 PG |
659 | cnt = None # file counter (uint32_t != 0) |
660 | iv = None # current IV | |
30019abf PG |
661 | fixed = None # accu for 64 bit fixed parts of IV |
662 | used_ivs = None # tracks IVs | |
663 | strict_ivs = False # if True, panic on duplicate object IV | |
48db09ba PG |
664 | password = None |
665 | paramversion = None | |
633b18a9 PG |
666 | stats = { "in" : 0 |
667 | , "out" : 0 | |
668 | , "obj" : 0 } | |
fa47412e | 669 | |
fa47412e PG |
670 | ctsize = -1 |
671 | ptsize = -1 | |
3031b7ae PG |
672 | info_counter_used = False |
673 | index_counter_used = False | |
a393d9cb | 674 | |
a64085a8 | 675 | def __init__ (self, *al, **akv): |
30019abf | 676 | self.used_ivs = set () |
a64085a8 | 677 | self.set_parameters (*al, **akv) |
39accaaa PG |
678 | |
679 | ||
704ceaa5 | 680 | def next_fixed (self): |
be124bca | 681 | # NOP for decryption |
50710d86 PG |
682 | pass |
683 | ||
684 | ||
685 | def set_object_counter (self, cnt=None): | |
704ceaa5 PG |
686 | """ |
687 | Safely set the internal counter of encrypted objects. Numerous | |
688 | constraints apply: | |
689 | ||
690 | The same counter may not be reused in combination with one IV fixed | |
691 | part. This is validated elsewhere in the IV handling. | |
692 | ||
693 | Counter zero is invalid. The first two counters are reserved for | |
694 | metadata. The implementation does not allow for splitting metadata | |
695 | files over multiple encrypted objects. (This would be possible by | |
696 | assigning new fixed parts.) Thus in a Deltatar backup there is at most | |
697 | one object with a counter value of one and two. On creation of a | |
698 | context, the initial counter may be chosen. The globals | |
699 | ``AES_GCM_IV_CNT_INFOFILE`` and ``AES_GCM_IV_CNT_INDEX`` can be used to | |
700 | request one of the reserved values. If one of these values has been | |
701 | used, any further attempt of setting the counter to that value will | |
702 | be rejected with an ``InvalidFileCounter`` exception. | |
703 | ||
704 | Out of bounds values (i. e. below one and more than the maximum of 2³²) | |
705 | cause an ``InvalidParameter`` exception to be thrown. | |
706 | """ | |
50710d86 PG |
707 | if cnt is None: |
708 | self.cnt = AES_GCM_IV_CNT_DATA | |
709 | return | |
710 | if cnt == 0 or cnt > AES_GCM_IV_CNT_MAX + 1: | |
b12110dd PG |
711 | raise InvalidParameter ("invalid counter value %d requested: " |
712 | "acceptable values are from 1 to %d" | |
713 | % (cnt, AES_GCM_IV_CNT_MAX)) | |
50710d86 PG |
714 | if cnt == AES_GCM_IV_CNT_INFOFILE: |
715 | if self.info_counter_used is True: | |
fac2cfe1 PG |
716 | raise InvalidFileCounter ("attempted to reuse info file " |
717 | "counter %d: must be unique" % cnt) | |
50710d86 | 718 | self.info_counter_used = True |
3031b7ae PG |
719 | elif cnt == AES_GCM_IV_CNT_INDEX: |
720 | if self.index_counter_used is True: | |
fac2cfe1 PG |
721 | raise InvalidFileCounter ("attempted to reuse index file " |
722 | " counter %d: must be unique" % cnt) | |
3031b7ae | 723 | self.index_counter_used = True |
50710d86 PG |
724 | if cnt <= AES_GCM_IV_CNT_MAX: |
725 | self.cnt = cnt | |
726 | return | |
727 | # cnt == AES_GCM_IV_CNT_MAX + 1 → wrap | |
728 | self.cnt = AES_GCM_IV_CNT_DATA | |
704ceaa5 | 729 | self.next_fixed () |
50710d86 PG |
730 | |
731 | ||
1f3fd7b0 | 732 | def set_parameters (self, password=None, key=None, paramversion=None, |
be124bca | 733 | nacl=None, counter=None, strict_ivs=False): |
704ceaa5 PG |
734 | """ |
735 | Configure the internal state of a crypto context. Not intended for | |
736 | external use. | |
737 | """ | |
be124bca | 738 | self.next_fixed () |
50710d86 | 739 | self.set_object_counter (counter) |
30019abf PG |
740 | self.strict_ivs = strict_ivs |
741 | ||
a83fa4ed PG |
742 | if paramversion is not None: |
743 | self.paramversion = paramversion | |
744 | ||
1f3fd7b0 PG |
745 | if key is not None: |
746 | self.key, self.nacl = key, nacl | |
747 | return | |
748 | ||
a83fa4ed PG |
749 | if password is not None: |
750 | if isinstance (password, bytes) is False: | |
751 | password = str.encode (password) | |
752 | self.password = password | |
753 | if paramversion is None and nacl is None: | |
754 | # postpone key setup until first header is available | |
755 | return | |
756 | kdf = kdf_by_version (paramversion) | |
757 | if kdf is not None: | |
758 | self.key, self.nacl = kdf (password, nacl) | |
fa47412e | 759 | |
39accaaa | 760 | |
39accaaa | 761 | def process (self, buf): |
704ceaa5 PG |
762 | """ |
763 | Encrypt / decrypt a buffer. Invokes the ``.update()`` method on the | |
764 | wrapped encryptor or decryptor, respectively. | |
765 | ||
766 | The Cryptography exception ``AlreadyFinalized`` is translated to an | |
767 | ``InternalError`` at this point. It may occur in sound code when the GC | |
768 | closes an encrypting stream after an error. Everywhere else it must be | |
769 | treated as a bug. | |
770 | """ | |
cb7a3911 PG |
771 | if self.enc is None: |
772 | raise RuntimeError ("process: context not initialized") | |
773 | self.stats ["in"] += len (buf) | |
fac2cfe1 PG |
774 | try: |
775 | out = self.enc.update (buf) | |
776 | except cryptography.exceptions.AlreadyFinalized as exn: | |
777 | raise InternalError (exn) | |
cb7a3911 PG |
778 | self.stats ["out"] += len (out) |
779 | return out | |
39accaaa PG |
780 | |
781 | ||
30019abf | 782 | def next (self, password, paramversion, nacl, iv): |
704ceaa5 PG |
783 | """ |
784 | Prepare for encrypting another object: Reset the data counters and | |
785 | change the configuration in case one of the variable parameters differs | |
786 | from the last object. Also check the IV for duplicates and error out | |
787 | if strict checking was requested. | |
788 | """ | |
fa47412e PG |
789 | self.ctsize = 0 |
790 | self.ptsize = 0 | |
791 | self.stats ["obj"] += 1 | |
30019abf PG |
792 | |
793 | self.check_duplicate_iv (iv) | |
794 | ||
6178061e PG |
795 | if ( self.paramversion != paramversion |
796 | or self.password != password | |
797 | or self.nacl != nacl): | |
1f3fd7b0 | 798 | self.set_parameters (password=password, paramversion=paramversion, |
30019abf PG |
799 | nacl=nacl, strict_ivs=self.strict_ivs) |
800 | ||
801 | ||
802 | def check_duplicate_iv (self, iv): | |
704ceaa5 PG |
803 | """ |
804 | Add an IV (the 12 byte representation as in the header) to the list. With | |
805 | strict checking enabled, this will throw a ``DuplicateIV``. Depending on | |
806 | the context, this may indicate a serious error (IV reuse). | |
807 | """ | |
30019abf PG |
808 | if self.strict_ivs is True and iv in self.used_ivs: |
809 | raise DuplicateIV ("iv %s was reused" % iv_fmt (iv)) | |
810 | # vi has not been used before; add to collection | |
811 | self.used_ivs.add (iv) | |
fa47412e PG |
812 | |
813 | ||
633b18a9 | 814 | def counters (self): |
704ceaa5 PG |
815 | """ |
816 | Access the data counters. | |
817 | """ | |
633b18a9 PG |
818 | return self.stats ["obj"], self.stats ["in"], self.stats ["out"] |
819 | ||
820 | ||
8de91f4f PG |
821 | def drop (self): |
822 | """ | |
823 | Clear the current context regardless of its finalization state. The | |
824 | next operation must be ``.next()``. | |
825 | """ | |
826 | self.enc = None | |
827 | ||
828 | ||
39accaaa PG |
829 | class Encrypt (Crypto): |
830 | ||
48db09ba PG |
831 | lastinfo = None |
832 | version = None | |
72a42219 | 833 | paramenc = None |
50710d86 | 834 | |
1f3fd7b0 | 835 | def __init__ (self, version, paramversion, password=None, key=None, nacl=None, |
30019abf | 836 | counter=AES_GCM_IV_CNT_DATA, strict_ivs=True): |
704ceaa5 PG |
837 | """ |
838 | The ctor will throw immediately if one of the parameters does not conform | |
839 | to our expectations. | |
840 | ||
841 | counter=AES_GCM_IV_CNT_DATA, strict_ivs=True): | |
842 | :type version: int to fit uint16_t | |
843 | :type paramversion: int to fit uint16_t | |
844 | :param password: mutually exclusive with ``key`` | |
845 | :type password: bytes | |
846 | :param key: mutually exclusive with ``password`` | |
847 | :type key: bytes | |
848 | :type nacl: bytes | |
849 | :type counter: initial object counter the values | |
850 | ``AES_GCM_IV_CNT_INFOFILE`` and | |
851 | ``AES_GCM_IV_CNT_INDEX`` are unique in each backup set | |
852 | and cannot be reused even with different fixed parts. | |
853 | :type strict_ivs: bool | |
854 | """ | |
1f3fd7b0 PG |
855 | if password is None and key is None \ |
856 | or password is not None and key is not None : | |
857 | raise InvalidParameter ("__init__: need either key or password") | |
858 | ||
859 | if key is not None: | |
860 | if isinstance (key, bytes) is False: | |
861 | raise InvalidParameter ("__init__: key must be provided as " | |
862 | "bytes, not %s" % type (key)) | |
863 | if nacl is None: | |
864 | raise InvalidParameter ("__init__: salt must be provided along " | |
865 | "with encryption key") | |
866 | else: # password, no key | |
867 | if isinstance (password, str) is False: | |
868 | raise InvalidParameter ("__init__: password must be a string, not %s" | |
869 | % type (password)) | |
870 | if len (password) == 0: | |
871 | raise InvalidParameter ("__init__: supplied empty password but not " | |
872 | "permitted for PDT encrypted files") | |
36b9932a PG |
873 | # version |
874 | if isinstance (version, int) is False: | |
875 | raise InvalidParameter ("__init__: version number must be an " | |
876 | "integer, not %s" % type (version)) | |
877 | if version < 0: | |
878 | raise InvalidParameter ("__init__: version number must be a " | |
879 | "nonnegative integer, not %d" % version) | |
880 | # paramversion | |
881 | if isinstance (paramversion, int) is False: | |
882 | raise InvalidParameter ("__init__: crypto parameter version number " | |
883 | "must be an integer, not %s" | |
884 | % type (paramversion)) | |
885 | if paramversion < 0: | |
886 | raise InvalidParameter ("__init__: crypto parameter version number " | |
887 | "must be a nonnegative integer, not %d" | |
888 | % paramversion) | |
889 | # salt | |
890 | if nacl is not None: | |
891 | if isinstance (nacl, bytes) is False: | |
892 | raise InvalidParameter ("__init__: salt given, but of type %s " | |
893 | "instead of bytes" % type (nacl)) | |
894 | # salt length would depend on the actual encryption so it can’t be | |
895 | # validated at this point | |
b12110dd | 896 | self.fixed = [ ] |
48db09ba PG |
897 | self.version = version |
898 | self.paramenc = ENCRYPTION_PARAMETERS.get (paramversion) ["enc"] | |
72a42219 | 899 | |
1f3fd7b0 | 900 | super().__init__ (password, key, paramversion, nacl, counter=counter, |
30019abf | 901 | strict_ivs=strict_ivs) |
a393d9cb PG |
902 | |
903 | ||
be124bca PG |
904 | def next_fixed (self, retries=PDTCRYPT_IV_GEN_MAX_RETRIES): |
905 | """ | |
906 | Generate the next IV fixed part by reading eight bytes from | |
907 | ``/dev/urandom``. The buffer so obtained is tested against the fixed | |
908 | parts used so far to prevent accidental reuse of IVs. After a | |
909 | configurable number of attempts to create a unique fixed part, it will | |
910 | refuse to continue with an ``IVFixedPartError``. This is unlikely to | |
911 | ever happen on a normal system but may detect an issue with the random | |
912 | generator. | |
913 | ||
914 | The list of fixed parts that were used by the context at hand can be | |
915 | accessed through the ``.fixed`` list. Its last element is the fixed | |
916 | part currently in use. | |
917 | """ | |
918 | i = 0 | |
919 | while i < retries: | |
920 | fp = os.urandom (PDTCRYPT_IV_FIXEDPART_SIZE) | |
921 | if fp not in self.fixed: | |
922 | self.fixed.append (fp) | |
923 | return | |
924 | i += 1 | |
925 | raise IVFixedPartError ("error obtaining a unique IV fixed part from " | |
926 | "/dev/urandom; giving up after %d tries" % i) | |
927 | ||
928 | ||
a393d9cb | 929 | def iv_make (self): |
704ceaa5 PG |
930 | """ |
931 | Construct a 12-bytes IV from the current fixed part and the object | |
932 | counter. | |
933 | """ | |
b12110dd | 934 | return struct.pack(FMT_I2N_IV, self.fixed [-1], self.cnt) |
a393d9cb PG |
935 | |
936 | ||
cb7a3911 | 937 | def next (self, filename=None, counter=None): |
704ceaa5 PG |
938 | """ |
939 | Prepare for encrypting the next incoming object. Update the counter | |
940 | and put together the IV, possibly changing prefixes. Then create the | |
941 | new encryptor. | |
942 | ||
943 | The argument ``counter`` can be used to specify a file counter for this | |
944 | object. Unless it is one of the reserved values, the counter of | |
945 | subsequent objects will be computed from this one. | |
946 | ||
947 | If this is the first object in a series, ``filename`` is required, | |
948 | otherwise it is reused if not present. The value is used to derive a | |
949 | header sized placeholder to use until after encryption when all the | |
950 | inputs to construct the final header are available. This is then | |
951 | matched in ``.done()`` against the value found at the position of the | |
952 | header. The motivation for this extra check is primarily to assist | |
953 | format debugging: It makes stray headers easy to spot in malformed | |
954 | PDTCRYPT files. | |
955 | """ | |
cb7a3911 PG |
956 | if filename is None: |
957 | if self.lastinfo is None: | |
958 | raise InvalidParameter ("next: filename is mandatory for " | |
959 | "first object") | |
960 | filename, _dummy = self.lastinfo | |
961 | else: | |
962 | if isinstance (filename, str) is False: | |
963 | raise InvalidParameter ("next: filename must be a string, no %s" | |
964 | % type (filename)) | |
3031b7ae PG |
965 | if counter is not None: |
966 | if isinstance (counter, int) is False: | |
967 | raise InvalidParameter ("next: the supplied counter is of " | |
968 | "invalid type %s; please pass an " | |
969 | "integer instead" % type (counter)) | |
970 | self.set_object_counter (counter) | |
fac2cfe1 | 971 | |
50710d86 | 972 | self.iv = self.iv_make () |
72a42219 | 973 | if self.paramenc == "aes-gcm": |
6178061e PG |
974 | self.enc = Cipher \ |
975 | ( algorithms.AES (self.key) | |
976 | , modes.GCM (self.iv) | |
977 | , backend = default_backend ()) \ | |
978 | .encryptor () | |
72a42219 | 979 | elif self.paramenc == "passthrough": |
6178061e PG |
980 | self.enc = PassthroughCipher () |
981 | else: | |
b12110dd PG |
982 | raise InvalidParameter ("next: parameter version %d not known" |
983 | % self.paramversion) | |
48db09ba PG |
984 | hdrdum = hdr_make_dummy (filename) |
985 | self.lastinfo = (filename, hdrdum) | |
30019abf | 986 | super().next (self.password, self.paramversion, self.nacl, self.iv) |
72a42219 | 987 | |
3031b7ae | 988 | self.set_object_counter (self.cnt + 1) |
48db09ba | 989 | return hdrdum |
a393d9cb | 990 | |
a393d9cb | 991 | |
cd77dadb | 992 | def done (self, cmpdata): |
704ceaa5 PG |
993 | """ |
994 | Complete encryption of an object. After this has been called, attempts | |
995 | of encrypting further data will cause an error until ``.next()`` is | |
996 | invoked properly. | |
997 | ||
998 | Returns a 64 bytes buffer containing the object header including all | |
999 | values including the “late” ones e. g. the ciphertext size and the | |
1000 | GCM tag. | |
1001 | """ | |
36b9932a PG |
1002 | if isinstance (cmpdata, bytes) is False: |
1003 | raise InvalidParameter ("done: comparison input expected as bytes, " | |
1004 | "not %s" % type (cmpdata)) | |
cb7a3911 PG |
1005 | if self.lastinfo is None: |
1006 | raise RuntimeError ("done: encryption context not initialized") | |
48db09ba PG |
1007 | filename, hdrdum = self.lastinfo |
1008 | if cmpdata != hdrdum: | |
b12110dd PG |
1009 | raise RuntimeError ("done: bad sync of header for object %d: " |
1010 | "preliminary data does not match; this likely " | |
1011 | "indicates a wrongly repositioned stream" | |
1012 | % self.cnt) | |
6178061e | 1013 | data = self.enc.finalize () |
633b18a9 | 1014 | self.stats ["out"] += len (data) |
cd77dadb | 1015 | self.ctsize += len (data) |
48db09ba PG |
1016 | ok, hdr = hdr_from_params (self.version, self.paramversion, self.nacl, |
1017 | self.iv, self.ctsize, self.enc.tag) | |
8a990744 | 1018 | if ok is False: |
b12110dd PG |
1019 | raise InternalError ("error constructing header: %r" % hdr) |
1020 | return data, hdr, self.fixed | |
a393d9cb | 1021 | |
a393d9cb | 1022 | |
cd77dadb | 1023 | def process (self, buf): |
704ceaa5 PG |
1024 | """ |
1025 | Encrypt a chunk of plaintext with the active encryptor. Returns the | |
1026 | size of the input consumed. This **must** be checked downstream. If the | |
1027 | maximum possible object size has been reached, the current context must | |
1028 | be finalized and a new one established before any further data can be | |
1029 | encrypted. The second argument is the remainder of the plaintext that | |
1030 | was not encrypted for the caller to use immediately after the new | |
1031 | context is ready. | |
1032 | """ | |
36b9932a PG |
1033 | if isinstance (buf, bytes) is False: |
1034 | raise InvalidParameter ("process: expected byte buffer, not %s" | |
1035 | % type (buf)) | |
cb7a3911 PG |
1036 | bsize = len (buf) |
1037 | newptsize = self.ptsize + bsize | |
1038 | diff = newptsize - PDTCRYPT_MAX_OBJ_SIZE | |
1039 | if diff > 0: | |
1040 | bsize -= diff | |
1041 | newptsize = PDTCRYPT_MAX_OBJ_SIZE | |
1042 | self.ptsize = newptsize | |
1043 | data = super().process (buf [:bsize]) | |
cd77dadb | 1044 | self.ctsize += len (data) |
cb7a3911 | 1045 | return bsize, data |
cd77dadb PG |
1046 | |
1047 | ||
39accaaa | 1048 | class Decrypt (Crypto): |
a393d9cb | 1049 | |
3031b7ae | 1050 | tag = None # GCM tag, part of header |
3031b7ae | 1051 | last_iv = None # check consecutive ivs in strict mode |
39accaaa | 1052 | |
1f3fd7b0 | 1053 | def __init__ (self, password=None, key=None, counter=None, fixedparts=None, |
ee6aa239 | 1054 | strict_ivs=False): |
704ceaa5 PG |
1055 | """ |
1056 | Sanitizing ctor for the decryption context. ``fixedparts`` specifies a | |
1057 | list of IV fixed parts accepted during decryption. If a fixed part is | |
1058 | encountered that is not in the list, decryption will fail. | |
1059 | ||
1060 | :param password: mutually exclusive with ``key`` | |
1061 | :type password: bytes | |
1062 | :param key: mutually exclusive with ``password`` | |
1063 | :type key: bytes | |
1064 | :type counter: initial object counter the values | |
1065 | ``AES_GCM_IV_CNT_INFOFILE`` and | |
1066 | ``AES_GCM_IV_CNT_INDEX`` are unique in each backup set | |
1067 | and cannot be reused even with different fixed parts. | |
1068 | :type fixedparts: bytes list | |
1069 | """ | |
1f3fd7b0 PG |
1070 | if password is None and key is None \ |
1071 | or password is not None and key is not None : | |
1072 | raise InvalidParameter ("__init__: need either key or password") | |
1073 | ||
1074 | if key is not None: | |
1075 | if isinstance (key, bytes) is False: | |
1076 | raise InvalidParameter ("__init__: key must be provided as " | |
1077 | "bytes, not %s" % type (key)) | |
1078 | else: # password, no key | |
1079 | if isinstance (password, str) is False: | |
1080 | raise InvalidParameter ("__init__: password must be a string, not %s" | |
1081 | % type (password)) | |
1082 | if len (password) == 0: | |
1083 | raise InvalidParameter ("__init__: supplied empty password but not " | |
1084 | "permitted for PDT encrypted files") | |
36b9932a | 1085 | # fixed parts |
50710d86 | 1086 | if fixedparts is not None: |
36b9932a PG |
1087 | if isinstance (fixedparts, list) is False: |
1088 | raise InvalidParameter ("__init__: IV fixed parts must be " | |
1089 | "supplied as list, not %s" | |
1090 | % type (fixedparts)) | |
b12110dd PG |
1091 | self.fixed = fixedparts |
1092 | self.fixed.sort () | |
ee6aa239 | 1093 | |
a83fa4ed PG |
1094 | super().__init__ (password=password, key=key, counter=counter, |
1095 | strict_ivs=strict_ivs) | |
39accaaa PG |
1096 | |
1097 | ||
b12110dd | 1098 | def valid_fixed_part (self, iv): |
704ceaa5 PG |
1099 | """ |
1100 | Check if a fixed part was already seen. | |
1101 | """ | |
50710d86 | 1102 | # check if fixed part is known |
b12110dd PG |
1103 | fixed, _cnt = struct.unpack (FMT_I2N_IV, iv) |
1104 | i = bisect.bisect_left (self.fixed, fixed) | |
1105 | return i != len (self.fixed) and self.fixed [i] == fixed | |
50710d86 PG |
1106 | |
1107 | ||
ee6aa239 | 1108 | def check_consecutive_iv (self, iv): |
704ceaa5 PG |
1109 | """ |
1110 | Check whether the counter part of the given IV is indeed the successor | |
1111 | of the currently present counter. This should always be the case for | |
1112 | the objects in a well formed PDT archive but should not be enforced | |
1113 | when decrypting out-of-order. | |
1114 | """ | |
ee6aa239 | 1115 | fixed, cnt = struct.unpack (FMT_I2N_IV, iv) |
3031b7ae PG |
1116 | if self.strict_ivs is True \ |
1117 | and self.last_iv is not None \ | |
ee6aa239 PG |
1118 | and self.last_iv [0] == fixed \ |
1119 | and self.last_iv [1] != cnt - 1: | |
f6cd676f | 1120 | raise NonConsecutiveIV ("iv %s counter not successor of " |
ee6aa239 | 1121 | "last object (expected %d, found %d)" |
f6cd676f | 1122 | % (iv_fmt (self.last_iv [1]), cnt)) |
ee6aa239 PG |
1123 | self.last_iv = (iv, cnt) |
1124 | ||
1125 | ||
79782fa9 | 1126 | def next (self, hdr): |
704ceaa5 PG |
1127 | """ |
1128 | Start decrypting the next object. The PDTCRYPT header for the object | |
1129 | can be given either as already parsed object or as bytes. | |
1130 | """ | |
dccfe104 PG |
1131 | if isinstance (hdr, bytes) is True: |
1132 | hdr = hdr_read (hdr) | |
36b9932a PG |
1133 | elif isinstance (hdr, dict) is False: |
1134 | # this won’t catch malformed specs though | |
1135 | raise InvalidParameter ("next: wrong type of parameter hdr: " | |
1136 | "expected bytes or spec, got %s" | |
fbfda3d4 | 1137 | % type (hdr)) |
36b9932a PG |
1138 | try: |
1139 | paramversion = hdr ["paramversion"] | |
1140 | nacl = hdr ["nacl"] | |
1141 | iv = hdr ["iv"] | |
1142 | tag = hdr ["tag"] | |
1143 | except KeyError: | |
1144 | raise InvalidHeader ("next: not a header %r" % hdr) | |
1145 | ||
30019abf | 1146 | super().next (self.password, paramversion, nacl, iv) |
b12110dd | 1147 | if self.fixed is not None and self.valid_fixed_part (iv) is False: |
f6cd676f PG |
1148 | raise InvalidIVFixedPart ("iv %s has invalid fixed part" |
1149 | % iv_fmt (iv)) | |
3031b7ae | 1150 | self.check_consecutive_iv (iv) |
ee6aa239 | 1151 | |
36b9932a | 1152 | self.tag = tag |
b12110dd PG |
1153 | defs = ENCRYPTION_PARAMETERS.get (paramversion, None) |
1154 | if defs is None: | |
1155 | raise FormatError ("header contains unknown parameter version %d; " | |
1156 | "maybe the file was created by a more recent " | |
1157 | "version of Deltatar" % paramversion) | |
50710d86 | 1158 | enc = defs ["enc"] |
6178061e PG |
1159 | if enc == "aes-gcm": |
1160 | self.enc = Cipher \ | |
1161 | ( algorithms.AES (self.key) | |
36b9932a | 1162 | , modes.GCM (iv, tag=self.tag) |
6178061e PG |
1163 | , backend = default_backend ()) \ |
1164 | . decryptor () | |
1165 | elif enc == "passthrough": | |
1166 | self.enc = PassthroughCipher () | |
1167 | else: | |
b12110dd PG |
1168 | raise InternalError ("encryption parameter set %d refers to unknown " |
1169 | "mode %r" % (paramversion, enc)) | |
f484f2d1 | 1170 | self.set_object_counter (self.cnt + 1) |
39accaaa PG |
1171 | |
1172 | ||
db1f3ac7 | 1173 | def done (self, tag=None): |
704ceaa5 PG |
1174 | """ |
1175 | Stop decryption of the current object and finalize it with the active | |
1176 | context. This will throw an *InvalidGCMTag* exception to indicate that | |
1177 | the authentication tag does not match the data. If the tag is correct, | |
1178 | the rest of the plaintext is returned. | |
1179 | """ | |
633b18a9 | 1180 | data = b"" |
db1f3ac7 PG |
1181 | try: |
1182 | if tag is None: | |
f484f2d1 | 1183 | data = self.enc.finalize () |
db1f3ac7 | 1184 | else: |
36b9932a PG |
1185 | if isinstance (tag, bytes) is False: |
1186 | raise InvalidParameter ("done: wrong type of parameter " | |
1187 | "tag: expected bytes, got %s" | |
1188 | % type (tag)) | |
f484f2d1 | 1189 | data = self.enc.finalize_with_tag (self.tag) |
b0078f26 | 1190 | except cryptography.exceptions.InvalidTag: |
f08c604b | 1191 | raise InvalidGCMTag ("done: tag mismatch of object %d: %s " |
b0078f26 | 1192 | "rejected by finalize ()" |
f08c604b | 1193 | % (self.cnt, binascii.hexlify (self.tag))) |
50710d86 | 1194 | self.ctsize += len (data) |
633b18a9 | 1195 | self.stats ["out"] += len (data) |
b0078f26 | 1196 | return data |
00b3cd10 PG |
1197 | |
1198 | ||
47e27926 | 1199 | def process (self, buf): |
704ceaa5 PG |
1200 | """ |
1201 | Decrypt the bytes object *buf* with the active decryptor. | |
1202 | """ | |
36b9932a PG |
1203 | if isinstance (buf, bytes) is False: |
1204 | raise InvalidParameter ("process: expected byte buffer, not %s" | |
1205 | % type (buf)) | |
47e27926 PG |
1206 | self.ctsize += len (buf) |
1207 | data = super().process (buf) | |
1208 | self.ptsize += len (data) | |
1209 | return data | |
1210 | ||
1211 | ||
00b3cd10 | 1212 | ############################################################################### |
770173c5 PG |
1213 | ## testing helpers |
1214 | ############################################################################### | |
1215 | ||
cb7a3911 | 1216 | def _patch_global (glob, vow, n=None): |
770173c5 PG |
1217 | """ |
1218 | Adapt upper file counter bound for testing IV logic. Completely unsafe. | |
1219 | """ | |
1220 | assert vow == "I am fully aware that this will void my warranty." | |
cb7a3911 PG |
1221 | r = globals () [glob] |
1222 | if n is None: | |
1223 | n = globals () [glob + "_DEFAULT"] | |
1224 | globals () [glob] = n | |
770173c5 PG |
1225 | return r |
1226 | ||
cb7a3911 PG |
1227 | _testing_set_AES_GCM_IV_CNT_MAX = \ |
1228 | partial (_patch_global, "AES_GCM_IV_CNT_MAX") | |
1229 | ||
1230 | _testing_set_PDTCRYPT_MAX_OBJ_SIZE = \ | |
1231 | partial (_patch_global, "PDTCRYPT_MAX_OBJ_SIZE") | |
1232 | ||
770173c5 | 1233 | ############################################################################### |
00b3cd10 PG |
1234 | ## freestanding invocation |
1235 | ############################################################################### | |
1236 | ||
da82bc58 PG |
1237 | PDTCRYPT_SUB_PROCESS = 0 |
1238 | PDTCRYPT_SUB_SCRYPT = 1 | |
1239 | ||
1240 | PDTCRYPT_SUB = \ | |
1241 | { "process" : PDTCRYPT_SUB_PROCESS | |
1242 | , "scrypt" : PDTCRYPT_SUB_SCRYPT } | |
1243 | ||
a83fa4ed PG |
1244 | PDTCRYPT_SECRET_PW = 0 |
1245 | PDTCRYPT_SECRET_KEY = 1 | |
1246 | ||
e3abcdf0 PG |
1247 | PDTCRYPT_DECRYPT = 1 << 0 # decrypt archive with password |
1248 | PDTCRYPT_SPLIT = 1 << 1 # split archive into individual objects | |
da82bc58 | 1249 | PDTCRYPT_HASH = 1 << 2 # output scrypt hash for file and given password |
e3abcdf0 PG |
1250 | |
1251 | PDTCRYPT_SPLITNAME = "pdtcrypt-object-%d.bin" | |
1252 | ||
70ad9458 | 1253 | PDTCRYPT_VERBOSE = False |
ee6aa239 | 1254 | PDTCRYPT_STRICTIVS = False |
b07633d3 | 1255 | PDTCRYPT_OVERWRITE = False |
15d3eefd | 1256 | PDTCRYPT_BLOCKSIZE = 1 << 12 |
70ad9458 PG |
1257 | PDTCRYPT_SINK = 0 |
1258 | PDTCRYPT_SOURCE = 1 | |
1259 | SELF = None | |
1260 | ||
77058bab PG |
1261 | PDTCRYPT_DEFAULT_VER = 1 |
1262 | PDTCRYPT_DEFAULT_PVER = 1 | |
1263 | ||
7b3940e5 PG |
1264 | # scrypt hashing output control |
1265 | PDTCRYPT_SCRYPT_INTRANATOR = 0 | |
1266 | PDTCRYPT_SCRYPT_PARAMETERS = 1 | |
4f6405d6 | 1267 | PDTCRYPT_SCRYPT_DEFAULT = PDTCRYPT_SCRYPT_INTRANATOR |
7b3940e5 PG |
1268 | |
1269 | PDTCRYPT_SCRYPT_FORMAT = \ | |
1270 | { "i2n" : PDTCRYPT_SCRYPT_INTRANATOR | |
1271 | , "params" : PDTCRYPT_SCRYPT_PARAMETERS } | |
1272 | ||
15d3eefd PG |
1273 | |
1274 | class PDTDecryptionError (Exception): | |
1275 | """Decryption failed.""" | |
1276 | ||
e3abcdf0 PG |
1277 | class PDTSplitError (Exception): |
1278 | """Decryption failed.""" | |
1279 | ||
15d3eefd PG |
1280 | |
1281 | def noise (*a, **b): | |
591a722f | 1282 | print (file=sys.stderr, *a, **b) |
15d3eefd PG |
1283 | |
1284 | ||
89e1073c PG |
1285 | class PassthroughDecryptor (object): |
1286 | ||
1287 | curhdr = None # write current header on first data write | |
1288 | ||
1289 | def __init__ (self): | |
1290 | if PDTCRYPT_VERBOSE is True: | |
1291 | noise ("PDT: no encryption; data passthrough") | |
1292 | ||
1293 | def next (self, hdr): | |
1294 | ok, curhdr = hdr_make (hdr) | |
1295 | if ok is False: | |
1296 | raise PDTDecryptionError ("bad header %r" % hdr) | |
1297 | self.curhdr = curhdr | |
1298 | ||
1299 | def done (self): | |
1300 | if self.curhdr is not None: | |
1301 | return self.curhdr | |
1302 | return b"" | |
1303 | ||
1304 | def process (self, d): | |
1305 | if self.curhdr is not None: | |
1306 | d = self.curhdr + d | |
1307 | self.curhdr = None | |
1308 | return d | |
1309 | ||
1310 | ||
a83fa4ed | 1311 | def depdtcrypt (mode, secret, ins, outs): |
15d3eefd | 1312 | """ |
a83fa4ed PG |
1313 | Remove PDTCRYPT layer from all objects encrypted with the secret. Used on a |
1314 | Deltatar backup this will yield a (possibly Gzip compressed) tarball. | |
15d3eefd PG |
1315 | """ |
1316 | ctleft = -1 # length of ciphertext to consume | |
1317 | ctcurrent = 0 # total ciphertext of current object | |
15d3eefd PG |
1318 | total_obj = 0 # total number of objects read |
1319 | total_pt = 0 # total plaintext bytes | |
1320 | total_ct = 0 # total ciphertext bytes | |
1321 | total_read = 0 # total bytes read | |
e3abcdf0 PG |
1322 | outfile = None # Python file object for output |
1323 | ||
89e1073c | 1324 | if mode & PDTCRYPT_DECRYPT: # decryptor |
a83fa4ed PG |
1325 | ks = secret [0] |
1326 | if ks == PDTCRYPT_SECRET_PW: | |
1327 | decr = Decrypt (password=secret [1], strict_ivs=PDTCRYPT_STRICTIVS) | |
1328 | elif ks == PDTCRYPT_SECRET_KEY: | |
1329 | key = binascii.unhexlify (secret [1]) | |
1330 | decr = Decrypt (key=key, strict_ivs=PDTCRYPT_STRICTIVS) | |
1331 | else: | |
1332 | raise InternalError ("‘%d’ does not specify a valid kind of secret" | |
1333 | % ks) | |
89e1073c PG |
1334 | else: |
1335 | decr = PassthroughDecryptor () | |
1336 | ||
e3abcdf0 PG |
1337 | def nextout (_): |
1338 | """Dummy for non-split mode: output file does not vary.""" | |
1339 | return outs | |
1340 | ||
1341 | if mode & PDTCRYPT_SPLIT: | |
1342 | def nextout (outfile): | |
1343 | """ | |
1344 | We were passed an fd as outs for accessing the destination | |
1345 | directory where extracted archive components are supposed | |
1346 | to end up in. | |
1347 | """ | |
1348 | ||
1349 | if outfile is None: | |
1350 | if PDTCRYPT_VERBOSE is True: | |
1351 | noise ("PDT: no output file to close at this point") | |
77058bab PG |
1352 | else: |
1353 | if PDTCRYPT_VERBOSE is True: | |
1354 | noise ("PDT: release output file %r" % outfile) | |
e3abcdf0 PG |
1355 | # cleanup happens automatically by the GC; the next |
1356 | # line will error out on account of an invalid fd | |
1357 | #outfile.close () | |
1358 | ||
1359 | assert total_obj > 0 | |
1360 | fname = PDTCRYPT_SPLITNAME % total_obj | |
1361 | try: | |
b07633d3 PG |
1362 | oflags = os.O_CREAT | os.O_WRONLY |
1363 | if PDTCRYPT_OVERWRITE is True: | |
1364 | oflags |= os.O_TRUNC | |
1365 | else: | |
1366 | oflags |= os.O_EXCL | |
1367 | outfd = os.open (fname, oflags, 0o600, dir_fd=outs) | |
e3abcdf0 PG |
1368 | if PDTCRYPT_VERBOSE is True: |
1369 | noise ("PDT: new output file %s → %d" % (fname, outfd)) | |
1370 | except FileExistsError as exn: | |
1371 | noise ("PDT: refusing to overwrite existing file %s" % fname) | |
1372 | noise ("") | |
1373 | raise PDTSplitError ("destination file %s already exists" | |
1374 | % fname) | |
1375 | ||
1376 | return os.fdopen (outfd, "wb", closefd=True) | |
1377 | ||
15d3eefd | 1378 | |
47d22679 | 1379 | def tell (s): |
b09a99eb | 1380 | """ESPIPE is normal on non-seekable stdio stream.""" |
47d22679 PG |
1381 | try: |
1382 | return s.tell () | |
1383 | except OSError as exn: | |
b09a99eb | 1384 | if exn.errno == os.errno.ESPIPE: |
47d22679 PG |
1385 | return -1 |
1386 | ||
e3abcdf0 | 1387 | def out (pt, outfile): |
15d3eefd PG |
1388 | npt = len (pt) |
1389 | nonlocal total_pt | |
1390 | total_pt += npt | |
70ad9458 | 1391 | if PDTCRYPT_VERBOSE is True: |
15d3eefd PG |
1392 | noise ("PDT:\t· decrypt plaintext %d B" % (npt)) |
1393 | try: | |
e3abcdf0 | 1394 | nn = outfile.write (pt) |
15d3eefd PG |
1395 | except OSError as exn: # probably ENOSPC |
1396 | raise DecryptionError ("error (%s)" % exn) | |
1397 | if nn != npt: | |
1398 | raise DecryptionError ("write aborted after %d of %d B" % (nn, npt)) | |
1399 | ||
1400 | while True: | |
1401 | if ctleft <= 0: | |
1402 | # current object completed; in a valid archive this marks either | |
1403 | # the start of a new header or the end of the input | |
1404 | if ctleft == 0: # current object requires finalization | |
70ad9458 | 1405 | if PDTCRYPT_VERBOSE is True: |
47d22679 | 1406 | noise ("PDT: %d finalize" % tell (ins)) |
5d394c0d PG |
1407 | try: |
1408 | pt = decr.done () | |
1409 | except InvalidGCMTag as exn: | |
f08c604b PG |
1410 | raise DecryptionError ("error finalizing object %d (%d B): " |
1411 | "%r" % (total_obj, len (pt), exn)) \ | |
1412 | from exn | |
e3abcdf0 | 1413 | out (pt, outfile) |
70ad9458 | 1414 | if PDTCRYPT_VERBOSE is True: |
15d3eefd PG |
1415 | noise ("PDT:\t· object validated") |
1416 | ||
70ad9458 | 1417 | if PDTCRYPT_VERBOSE is True: |
47d22679 | 1418 | noise ("PDT: %d hdr" % tell (ins)) |
15d3eefd PG |
1419 | try: |
1420 | hdr = hdr_read_stream (ins) | |
dd47d6a2 | 1421 | total_read += PDTCRYPT_HDR_SIZE |
ae3d0f2a PG |
1422 | except EndOfFile as exn: |
1423 | total_read += exn.remainder | |
dd47d6a2 | 1424 | if total_ct + total_obj * PDTCRYPT_HDR_SIZE != total_read: |
15d3eefd PG |
1425 | raise PDTDecryptionError ("ciphertext processed (%d B) plus " |
1426 | "overhead (%d × %d B) does not match " | |
1427 | "the number of bytes read (%d )" | |
dd47d6a2 | 1428 | % (total_ct, total_obj, PDTCRYPT_HDR_SIZE, |
15d3eefd PG |
1429 | total_read)) |
1430 | # the single good exit | |
1431 | return total_read, total_obj, total_ct, total_pt | |
1432 | except InvalidHeader as exn: | |
1433 | raise PDTDecryptionError ("invalid header at position %d in %r " | |
ee6aa239 | 1434 | "(%s)" % (tell (ins), exn, ins)) |
70ad9458 | 1435 | if PDTCRYPT_VERBOSE is True: |
15d3eefd PG |
1436 | pretty = hdr_fmt_pretty (hdr) |
1437 | noise (reduce (lambda a, e: (a + "\n" if a else "") + "PDT:\t· " + e, | |
1438 | pretty.splitlines (), "")) | |
1439 | ctcurrent = ctleft = hdr ["ctsize"] | |
89e1073c | 1440 | |
15d3eefd | 1441 | decr.next (hdr) |
e3abcdf0 PG |
1442 | |
1443 | total_obj += 1 # used in file counter with split mode | |
1444 | ||
1445 | # finalization complete or skipped in case of first object in | |
1446 | # stream; create a new output file if necessary | |
1447 | outfile = nextout (outfile) | |
15d3eefd | 1448 | |
70ad9458 | 1449 | if PDTCRYPT_VERBOSE is True: |
15d3eefd | 1450 | noise ("PDT: %d decrypt obj no. %d, %d B" |
47d22679 | 1451 | % (tell (ins), total_obj, ctleft)) |
15d3eefd PG |
1452 | |
1453 | # always allocate a new buffer since python-cryptography doesn’t allow | |
1454 | # passing a bytearray :/ | |
1455 | nexpect = min (ctleft, PDTCRYPT_BLOCKSIZE) | |
70ad9458 | 1456 | if PDTCRYPT_VERBOSE is True: |
15d3eefd | 1457 | noise ("PDT:\t· [%d] %d%% done, read block (%d B of %d B remaining)" |
47d22679 | 1458 | % (tell (ins), |
15d3eefd PG |
1459 | 100 - ctleft * 100 / (ctcurrent > 0 and ctcurrent or 1), |
1460 | nexpect, ctleft)) | |
1461 | ct = ins.read (nexpect) | |
1462 | nct = len (ct) | |
1463 | if nct < nexpect: | |
47d22679 | 1464 | off = tell (ins) |
ae3d0f2a PG |
1465 | raise EndOfFile (nct, |
1466 | "hit EOF after %d of %d B in block [%d:%d); " | |
15d3eefd PG |
1467 | "%d B ciphertext remaining for object no %d" |
1468 | % (nct, nexpect, off, off + nexpect, ctleft, | |
1469 | total_obj)) | |
1470 | ctleft -= nct | |
1471 | total_ct += nct | |
1472 | total_read += nct | |
1473 | ||
70ad9458 | 1474 | if PDTCRYPT_VERBOSE is True: |
15d3eefd PG |
1475 | noise ("PDT:\t· decrypt ciphertext %d B" % (nct)) |
1476 | pt = decr.process (ct) | |
e3abcdf0 | 1477 | out (pt, outfile) |
15d3eefd | 1478 | |
d6c15a52 | 1479 | |
70ad9458 | 1480 | def deptdcrypt_mk_stream (kind, path): |
d6c15a52 | 1481 | """Create stream from file or stdio descriptor.""" |
70ad9458 | 1482 | if kind == PDTCRYPT_SINK: |
d6c15a52 | 1483 | if path == "-": |
70ad9458 | 1484 | if PDTCRYPT_VERBOSE is True: noise ("PDT: sink: stdout") |
d6c15a52 PG |
1485 | return sys.stdout.buffer |
1486 | else: | |
70ad9458 | 1487 | if PDTCRYPT_VERBOSE is True: noise ("PDT: sink: file %s" % path) |
d6c15a52 | 1488 | return io.FileIO (path, "w") |
70ad9458 | 1489 | if kind == PDTCRYPT_SOURCE: |
d6c15a52 | 1490 | if path == "-": |
70ad9458 | 1491 | if PDTCRYPT_VERBOSE is True: noise ("PDT: source: stdin") |
d6c15a52 PG |
1492 | return sys.stdin.buffer |
1493 | else: | |
70ad9458 | 1494 | if PDTCRYPT_VERBOSE is True: noise ("PDT: source: file %s" % path) |
d6c15a52 PG |
1495 | return io.FileIO (path, "r") |
1496 | ||
1497 | raise ValueError ("bogus stream “%s” / %s" % (kind, path)) | |
1498 | ||
15d3eefd | 1499 | |
a83fa4ed | 1500 | def mode_depdtcrypt (mode, secret, ins, outs): |
da82bc58 PG |
1501 | try: |
1502 | total_read, total_obj, total_ct, total_pt = \ | |
a83fa4ed | 1503 | depdtcrypt (mode, secret, ins, outs) |
da82bc58 PG |
1504 | except DecryptionError as exn: |
1505 | noise ("PDT: Decryption failed:") | |
1506 | noise ("PDT:") | |
1507 | noise ("PDT: “%s”" % exn) | |
1508 | noise ("PDT:") | |
a83fa4ed | 1509 | noise ("PDT: Did you specify the correct key / password?") |
da82bc58 PG |
1510 | noise ("") |
1511 | return 1 | |
1512 | except PDTSplitError as exn: | |
1513 | noise ("PDT: Split operation failed:") | |
1514 | noise ("PDT:") | |
1515 | noise ("PDT: “%s”" % exn) | |
1516 | noise ("PDT:") | |
a83fa4ed | 1517 | noise ("PDT: Hint: target directory should be empty.") |
da82bc58 PG |
1518 | noise ("") |
1519 | return 1 | |
1520 | ||
1521 | if PDTCRYPT_VERBOSE is True: | |
1522 | noise ("PDT: decryption successful" ) | |
1523 | noise ("PDT: %.10d bytes read" % total_read) | |
1524 | noise ("PDT: %.10d objects decrypted" % total_obj ) | |
1525 | noise ("PDT: %.10d bytes ciphertext" % total_ct ) | |
1526 | noise ("PDT: %.10d bytes plaintext" % total_pt ) | |
1527 | noise ("" ) | |
1528 | ||
1529 | return 0 | |
1530 | ||
1531 | ||
7b3940e5 | 1532 | def mode_scrypt (pw, ins=None, nacl=None, fmt=PDTCRYPT_SCRYPT_INTRANATOR): |
77058bab | 1533 | hsh = None |
7b3940e5 | 1534 | paramversion = PDTCRYPT_DEFAULT_PVER |
77058bab PG |
1535 | if ins is not None: |
1536 | hsh, nacl, version, paramversion = scrypt_hashsource (pw, ins) | |
1537 | defs = ENCRYPTION_PARAMETERS.get(paramversion, None) | |
1538 | else: | |
1539 | nacl = binascii.unhexlify (nacl) | |
7b3940e5 | 1540 | defs = ENCRYPTION_PARAMETERS.get(paramversion, None) |
77058bab PG |
1541 | version = PDTCRYPT_DEFAULT_VER |
1542 | ||
1543 | kdfname, params = defs ["kdf"] | |
1544 | if hsh is None: | |
1545 | kdf = kdf_by_version (None, defs) | |
1546 | hsh, _void = kdf (pw, nacl) | |
da82bc58 PG |
1547 | |
1548 | import json | |
7b3940e5 PG |
1549 | |
1550 | if fmt == PDTCRYPT_SCRYPT_INTRANATOR: | |
1551 | out = json.dumps ({ "salt" : base64.b64encode (nacl).decode () | |
1552 | , "key" : base64.b64encode (hsh) .decode () | |
1553 | , "paramversion" : paramversion }) | |
1554 | elif fmt == PDTCRYPT_SCRYPT_PARAMETERS: | |
1555 | out = json.dumps ({ "salt" : binascii.hexlify (nacl).decode () | |
1556 | , "key" : binascii.hexlify (hsh) .decode () | |
1557 | , "version" : version | |
1558 | , "scrypt_params" : { "N" : params ["N"] | |
1559 | , "r" : params ["r"] | |
1560 | , "p" : params ["p"] | |
1561 | , "dkLen" : params ["dkLen"] } }) | |
1562 | else: | |
1563 | raise RuntimeError ("bad scrypt output scheme %r" % fmt) | |
1564 | ||
da82bc58 PG |
1565 | print (out) |
1566 | ||
1567 | ||
7b3940e5 | 1568 | |
70ad9458 PG |
1569 | def usage (err=False): |
1570 | out = print | |
1571 | if err is True: | |
1572 | out = noise | |
5afcb45d | 1573 | indent = ' ' * len (SELF) |
da82bc58 | 1574 | out ("usage: %s SUBCOMMAND { --help" % SELF) |
5afcb45d | 1575 | out (" %s | [ -v ] { -p PASSWORD | -k KEY }" % indent) |
77058bab PG |
1576 | out (" %s [ { -i | --in } { - | SOURCE } ]" % indent) |
1577 | out (" %s [ { -n | --nacl } { SALT } ]" % indent) | |
1578 | out (" %s [ { -o | --out } { - | DESTINATION } ]" % indent) | |
1579 | out (" %s [ -D | --no-decrypt ] [ -S | --split ]" % indent) | |
7b3940e5 | 1580 | out (" %s [ -f | --format ]" % indent) |
70ad9458 PG |
1581 | out ("") |
1582 | out ("\twhere") | |
da82bc58 PG |
1583 | out ("\t\tSUBCOMMAND main mode: { process | scrypt }") |
1584 | out ("\t\t where:") | |
1585 | out ("\t\t process: extract objects from PDT archive") | |
1586 | out ("\t\t scrypt: calculate hash from password and first object") | |
a83fa4ed PG |
1587 | out ("\t\t-p PASSWORD password to derive the encryption key from") |
1588 | out ("\t\t-k KEY encryption key as 16 bytes in hexadecimal notation") | |
e3abcdf0 | 1589 | out ("\t\t-s enforce strict handling of initialization vectors") |
70ad9458 PG |
1590 | out ("\t\t-i SOURCE file name to read from") |
1591 | out ("\t\t-o DESTINATION file to write output to") | |
77058bab | 1592 | out ("\t\t-n SALT provide salt for scrypt mode in hex encoding") |
70ad9458 | 1593 | out ("\t\t-v print extra info") |
e3abcdf0 PG |
1594 | out ("\t\t-S split into files at object boundaries; this") |
1595 | out ("\t\t requires DESTINATION to refer to directory") | |
1596 | out ("\t\t-D PDT header and ciphertext passthrough") | |
7b3940e5 | 1597 | out ("\t\t-f format of SCRYPT hash output (“default” or “parameters”)") |
70ad9458 PG |
1598 | out ("") |
1599 | out ("\tinstead of filenames, “-” may used to specify stdin / stdout") | |
1600 | out ("") | |
1601 | sys.exit ((err is True) and 42 or 0) | |
1602 | ||
1603 | ||
a83fa4ed PG |
1604 | def bail (msg): |
1605 | noise (msg) | |
1606 | noise ("") | |
1607 | usage (err=True) | |
1608 | raise Unreachable | |
1609 | ||
1610 | ||
70ad9458 PG |
1611 | def parse_argv (argv): |
1612 | global SELF | |
7b3940e5 PG |
1613 | mode = PDTCRYPT_DECRYPT |
1614 | secret = None | |
1615 | insspec = None | |
1616 | outsspec = None | |
1617 | nacl = None | |
4f6405d6 | 1618 | scrypt_format = PDTCRYPT_SCRYPT_DEFAULT |
70ad9458 PG |
1619 | |
1620 | argvi = iter (argv) | |
1621 | SELF = os.path.basename (next (argvi)) | |
1622 | ||
da82bc58 PG |
1623 | try: |
1624 | rawsubcmd = next (argvi) | |
1625 | subcommand = PDTCRYPT_SUB [rawsubcmd] | |
1626 | except StopIteration: | |
a83fa4ed | 1627 | bail ("ERROR: subcommand required") |
da82bc58 | 1628 | except KeyError: |
a83fa4ed | 1629 | bail ("ERROR: invalid subcommand “%s” specified" % rawsubcmd) |
da82bc58 | 1630 | |
59d74e2b PG |
1631 | def checked_arg (): |
1632 | nonlocal argvi | |
1633 | try: | |
1634 | return next (argvi) | |
1635 | except StopIteration: | |
1636 | bail ("ERROR: argument list incomplete") | |
1637 | ||
a83fa4ed PG |
1638 | def checked_secret (t, arg): |
1639 | nonlocal secret | |
1640 | if secret is None: | |
1641 | secret = (t, arg) | |
da82bc58 | 1642 | else: |
a83fa4ed | 1643 | bail ("ERROR: encountered “%s” but secret already given" % arg) |
da82bc58 | 1644 | |
70ad9458 PG |
1645 | for arg in argvi: |
1646 | if arg in [ "-h", "--help" ]: | |
1647 | usage () | |
1648 | raise Unreachable | |
1649 | elif arg in [ "-v", "--verbose", "--wtf" ]: | |
1650 | global PDTCRYPT_VERBOSE | |
1651 | PDTCRYPT_VERBOSE = True | |
1652 | elif arg in [ "-i", "--in", "--source" ]: | |
59d74e2b | 1653 | insspec = checked_arg () |
70ad9458 | 1654 | if PDTCRYPT_VERBOSE is True: noise ("PDT: decrypt from %s" % insspec) |
a83fa4ed | 1655 | elif arg in [ "-p", "--password" ]: |
59d74e2b | 1656 | arg = checked_arg () |
a83fa4ed PG |
1657 | checked_secret (PDTCRYPT_SECRET_PW, arg) |
1658 | if PDTCRYPT_VERBOSE is True: noise ("PDT: decrypting with password") | |
70ad9458 | 1659 | else: |
da82bc58 PG |
1660 | if subcommand == PDTCRYPT_SUB_PROCESS: |
1661 | if arg in [ "-s", "--strict-ivs" ]: | |
1662 | global PDTCRYPT_STRICTIVS | |
1663 | PDTCRYPT_STRICTIVS = True | |
77058bab PG |
1664 | elif arg in [ "-o", "--out", "--dest", "--sink" ]: |
1665 | outsspec = checked_arg () | |
1666 | if PDTCRYPT_VERBOSE is True: noise ("PDT: decrypt to %s" % outsspec) | |
da82bc58 PG |
1667 | elif arg in [ "-f", "--force" ]: |
1668 | global PDTCRYPT_OVERWRITE | |
1669 | PDTCRYPT_OVERWRITE = True | |
1670 | if PDTCRYPT_VERBOSE is True: noise ("PDT: overwrite existing files") | |
1671 | elif arg in [ "-S", "--split" ]: | |
1672 | mode |= PDTCRYPT_SPLIT | |
1673 | if PDTCRYPT_VERBOSE is True: noise ("PDT: split files") | |
1674 | elif arg in [ "-D", "--no-decrypt" ]: | |
1675 | mode &= ~PDTCRYPT_DECRYPT | |
1676 | if PDTCRYPT_VERBOSE is True: noise ("PDT: not decrypting") | |
a83fa4ed | 1677 | elif arg in [ "-k", "--key" ]: |
59d74e2b | 1678 | arg = checked_arg () |
a83fa4ed PG |
1679 | checked_secret (PDTCRYPT_SECRET_KEY, arg) |
1680 | if PDTCRYPT_VERBOSE is True: noise ("PDT: decrypting with key") | |
da82bc58 | 1681 | else: |
a83fa4ed | 1682 | bail ("ERROR: unexpected positional argument “%s”" % arg) |
da82bc58 | 1683 | elif subcommand == PDTCRYPT_SUB_SCRYPT: |
77058bab PG |
1684 | if arg in [ "-n", "--nacl", "--salt" ]: |
1685 | nacl = checked_arg () | |
1686 | if PDTCRYPT_VERBOSE is True: noise ("PDT: salt key with %s" % nacl) | |
7b3940e5 PG |
1687 | elif arg in [ "-f", "--format" ]: |
1688 | arg = checked_arg () | |
1689 | try: | |
1690 | scrypt_format = PDTCRYPT_SCRYPT_FORMAT [arg] | |
1691 | except KeyError: | |
1692 | bail ("ERROR: invalid scrypt output format %s" % arg) | |
1693 | if PDTCRYPT_VERBOSE is True: | |
1694 | noise ("PDT: scrypt output format “%s”" % scrypt_format) | |
77058bab PG |
1695 | else: |
1696 | bail ("ERROR: unexpected positional argument “%s”" % arg) | |
70ad9458 | 1697 | |
a83fa4ed | 1698 | if secret is None: |
ecb9676d | 1699 | if PDTCRYPT_VERBOSE is True: |
a83fa4ed | 1700 | noise ("ERROR: no password or key specified, trying $PDTCRYPT_PASSWORD") |
ecb9676d PG |
1701 | epw = os.getenv ("PDTCRYPT_PASSWORD") |
1702 | if epw is not None: | |
a83fa4ed PG |
1703 | checked_secret (PDTCRYPT_SECRET_PW, epw.strip ()) |
1704 | ||
1705 | if secret is None: | |
1706 | if PDTCRYPT_VERBOSE is True: | |
1707 | noise ("ERROR: no password or key specified, trying $PDTCRYPT_KEY") | |
1708 | ek = os.getenv ("PDTCRYPT_KEY") | |
1709 | if ek is not None: | |
1710 | checked_secret (PDTCRYPT_SECRET_KEY, ek.strip ()) | |
ecb9676d | 1711 | |
a83fa4ed | 1712 | if secret is None: |
da82bc58 | 1713 | if subcommand == PDTCRYPT_SUB_SCRYPT: |
a83fa4ed | 1714 | bail ("ERROR: scrypt hash mode requested but no password given") |
da82bc58 | 1715 | elif mode & PDTCRYPT_DECRYPT: |
a83fa4ed PG |
1716 | bail ("ERROR: encryption requested but no password given") |
1717 | ||
77058bab PG |
1718 | if subcommand == PDTCRYPT_SUB_SCRYPT: |
1719 | if secret [0] == PDTCRYPT_SECRET_KEY: | |
1720 | bail ("ERROR: scrypt mode requires a password") | |
1721 | if insspec is not None and nacl is not None \ | |
1722 | or insspec is None and nacl is None : | |
1723 | bail ("ERROR: please supply either an input file or " | |
1724 | "the salt") | |
70ad9458 PG |
1725 | |
1726 | # default to stdout | |
77058bab PG |
1727 | ins = None |
1728 | if insspec is not None or subcommand != PDTCRYPT_SUB_SCRYPT: | |
1729 | ins = deptdcrypt_mk_stream (PDTCRYPT_SOURCE, insspec or "-") | |
da82bc58 PG |
1730 | |
1731 | if subcommand == PDTCRYPT_SUB_SCRYPT: | |
7b3940e5 PG |
1732 | return True, partial (mode_scrypt, secret [1].encode (), ins, nacl, |
1733 | fmt=scrypt_format) | |
da82bc58 | 1734 | |
e3abcdf0 PG |
1735 | if mode & PDTCRYPT_SPLIT: # destination must be directory |
1736 | if outsspec is None or outsspec == "-": | |
a83fa4ed | 1737 | bail ("ERROR: split mode is incompatible with stdout sink") |
e3abcdf0 PG |
1738 | |
1739 | try: | |
1740 | try: | |
1741 | os.makedirs (outsspec, 0o700) | |
1742 | except FileExistsError: | |
1743 | # if it’s a directory with appropriate perms, everything is | |
1744 | # good; otherwise, below invocation of open(2) will fail | |
1745 | pass | |
1746 | outs = os.open (outsspec, os.O_DIRECTORY, 0o600) | |
1747 | except FileNotFoundError as exn: | |
a83fa4ed | 1748 | bail ("ERROR: cannot create target directory “%s”" % outsspec) |
e3abcdf0 | 1749 | except NotADirectoryError as exn: |
a83fa4ed | 1750 | bail ("ERROR: target path “%s” is not a directory" % outsspec) |
da82bc58 | 1751 | |
e3abcdf0 | 1752 | else: |
89e1073c | 1753 | outs = deptdcrypt_mk_stream (PDTCRYPT_SINK, outsspec or "-") |
da82bc58 | 1754 | |
a83fa4ed | 1755 | return True, partial (mode_depdtcrypt, mode, secret, ins, outs) |
15d3eefd PG |
1756 | |
1757 | ||
00b3cd10 | 1758 | def main (argv): |
da82bc58 | 1759 | ok, runner = parse_argv (argv) |
f08c604b | 1760 | |
da82bc58 | 1761 | if ok is True: return runner () |
15d3eefd | 1762 | |
da82bc58 | 1763 | return 1 |
f08c604b | 1764 | |
00b3cd10 PG |
1765 | |
1766 | if __name__ == "__main__": | |
1767 | sys.exit (main (sys.argv)) | |
1768 |