]> git.ipfire.org Git - thirdparty/openssl.git/blob - CHANGES
Ensure EVP_get_digestbyname() and EVP_get_cipherbyname() know all aliases
[thirdparty/openssl.git] / CHANGES
1
2 OpenSSL CHANGES
3 _______________
4
5 This is a high-level summary of the most important changes.
6 For a full list of changes, see the git commit log; for example,
7 https://github.com/openssl/openssl/commits/ and pick the appropriate
8 release branch.
9
10 Changes between 1.1.1 and 3.0.0 [xx XXX xxxx]
11
12 *) Added the X509_LOOKUP_METHOD called X509_LOOKUP_store, to allow
13 access to certificate and CRL stores via URIs and OSSL_STORE
14 loaders.
15
16 This adds the following functions:
17
18 X509_LOOKUP_store()
19 X509_STORE_load_file()
20 X509_STORE_load_path()
21 X509_STORE_load_store()
22 SSL_add_store_cert_subjects_to_stack()
23 SSL_CTX_set_default_verify_store()
24 SSL_CTX_load_verify_file()
25 SSL_CTX_load_verify_dir()
26 SSL_CTX_load_verify_store()
27
28 Also, the following functions are now deprecated:
29
30 - X509_STORE_load_locations() (use X509_STORE_load_file(),
31 X509_STORE_load_path() or X509_STORE_load_store() instead)
32 - SSL_CTX_load_verify_locations() (use SSL_CTX_load_verify_file(),
33 SSL_CTX_load_verify_dir() or SSL_CTX_load_verify_store() instead)
34 [Richard Levitte]
35
36 *) Added a new method to gather entropy on VMS, based on SYS$GET_ENTROPY.
37 The presence of this system service is determined at run-time.
38 [Richard Levitte]
39
40 *) Added functionality to create an EVP_PKEY context based on data
41 for methods from providers. This takes an algorithm name and a
42 property query string and simply stores them, with the intent
43 that any operation that uses this context will use those strings
44 to fetch the needed methods implicitly, thereby making the port
45 of application written for pre-3.0 OpenSSL easier.
46 [Richard Levitte]
47
48 *) The undocumented function NCONF_WIN32() has been deprecated; for
49 conversion details see the HISTORY section of doc/man5/config.pod
50 [Rich Salz]
51
52 *) Introduced the new functions EVP_DigestSignInit_ex() and
53 EVP_DigestVerifyInit_ex(). The macros EVP_DigestSignUpdate() and
54 EVP_DigestVerifyUpdate() have been converted to functions. See the man
55 pages for further details.
56 [Matt Caswell]
57
58 *) Over two thousand fixes were made to the documentation, including:
59 adding missing command flags, better style conformance, documentation
60 of internals, etc.
61 [Rich Salz, Richard Levitte]
62
63 *) s390x assembly pack: add hardware-support for P-256, P-384, P-521,
64 X25519, X448, Ed25519 and Ed448.
65 [Patrick Steuer]
66
67 *) Print all values for a PKCS#12 attribute with 'openssl pkcs12', not just
68 the first value.
69 [Jon Spillett]
70
71 *) Deprecated the public definition of ERR_STATE as well as the function
72 ERR_get_state(). This is done in preparation of making ERR_STATE an
73 opaque type.
74 [Richard Levitte]
75
76 *) Added ERR functionality to give callers access to the stored function
77 names that have replaced the older function code based functions.
78
79 New functions are ERR_get_error_func(), ERR_peek_error_func(),
80 ERR_peek_last_error_func(), ERR_get_error_data(), ERR_peek_error_data(),
81 ERR_peek_last_error_data(), ERR_get_error_all(), ERR_peek_error_all()
82 and ERR_peek_last_error_all().
83
84 These functions have become deprecated: ERR_get_error_line_data(),
85 ERR_peek_error_line_data(), ERR_peek_last_error_line_data() and
86 ERR_func_error_string().
87 [Richard Levitte]
88
89 *) Extended testing to be verbose for failing tests only. The make variables
90 VERBOSE_FAILURE or VF can be used to enable this:
91
92 $ make VF=1 test # Unix
93 $ mms /macro=(VF=1) test ! OpenVMS
94 $ nmake VF=1 test # Windows
95
96 [Richard Levitte]
97
98 *) For built-in EC curves, ensure an EC_GROUP built from the curve name is
99 used even when parsing explicit parameters, when loading a serialized key
100 or calling `EC_GROUP_new_from_ecpkparameters()`/
101 `EC_GROUP_new_from_ecparameters()`.
102 This prevents bypass of security hardening and performance gains,
103 especially for curves with specialized EC_METHODs.
104 By default, if a key encoded with explicit parameters is loaded and later
105 serialized, the output is still encoded with explicit parameters, even if
106 internally a "named" EC_GROUP is used for computation.
107 [Nicola Tuveri]
108
109 *) Compute ECC cofactors if not provided during EC_GROUP construction. Before
110 this change, EC_GROUP_set_generator would accept order and/or cofactor as
111 NULL. After this change, only the cofactor parameter can be NULL. It also
112 does some minimal sanity checks on the passed order.
113 (CVE-2019-1547)
114 [Billy Bob Brumley]
115
116 *) Fixed a padding oracle in PKCS7_dataDecode and CMS_decrypt_set1_pkey.
117 An attack is simple, if the first CMS_recipientInfo is valid but the
118 second CMS_recipientInfo is chosen ciphertext. If the second
119 recipientInfo decodes to PKCS #1 v1.5 form plaintext, the correct
120 encryption key will be replaced by garbage, and the message cannot be
121 decoded, but if the RSA decryption fails, the correct encryption key is
122 used and the recipient will not notice the attack.
123 As a work around for this potential attack the length of the decrypted
124 key must be equal to the cipher default key length, in case the
125 certifiate is not given and all recipientInfo are tried out.
126 The old behaviour can be re-enabled in the CMS code by setting the
127 CMS_DEBUG_DECRYPT flag.
128 [Bernd Edlinger]
129
130 *) Early start up entropy quality from the DEVRANDOM seed source has been
131 improved for older Linux systems. The RAND subsystem will wait for
132 /dev/random to be producing output before seeding from /dev/urandom.
133 The seeded state is stored for future library initialisations using
134 a system global shared memory segment. The shared memory identifier
135 can be configured by defining OPENSSL_RAND_SEED_DEVRANDOM_SHM_ID to
136 the desired value. The default identifier is 114.
137 [Paul Dale]
138
139 *) Revised BN_generate_prime_ex to not avoid factors 2..17863 in p-1
140 when primes for RSA keys are computed.
141 Since we previously always generated primes == 2 (mod 3) for RSA keys,
142 the 2-prime and 3-prime RSA modules were easy to distinguish, since
143 N = p*q = 1 (mod 3), but N = p*q*r = 2 (mod 3). Therefore fingerprinting
144 2-prime vs. 3-prime RSA keys was possible by computing N mod 3.
145 This avoids possible fingerprinting of newly generated RSA modules.
146 [Bernd Edlinger]
147
148 *) Correct the extended master secret constant on EBCDIC systems. Without this
149 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
150 negotiate EMS will fail. Unfortunately this also means that TLS connections
151 between EBCDIC systems with this fix, and EBCDIC systems without this
152 fix will fail if they negotiate EMS.
153 [Matt Caswell]
154
155 *) Changed the library initialisation so that the config file is now loaded
156 by default. This was already the case for libssl. It now occurs for both
157 libcrypto and libssl. Use the OPENSSL_INIT_NO_LOAD_CONFIG option to
158 OPENSSL_init_crypto() to suppress automatic loading of a config file.
159 [Matt Caswell]
160
161 *) Introduced new error raising macros, ERR_raise() and ERR_raise_data(),
162 where the former acts as a replacement for ERR_put_error(), and the
163 latter replaces the combination ERR_put_error()+ERR_add_error_data().
164 ERR_raise_data() adds more flexibility by taking a format string and
165 an arbitrary number of arguments following it, to be processed with
166 BIO_snprintf().
167 [Richard Levitte]
168
169 *) Introduced a new function, OSSL_PROVIDER_available(), which can be used
170 to check if a named provider is loaded and available. When called, it
171 will also activate all fallback providers if such are still present.
172 [Richard Levitte]
173
174 *) Enforce a minimum DH modulus size of 512 bits.
175 [Bernd Edlinger]
176
177 *) Changed DH parameters to generate the order q subgroup instead of 2q.
178 Previously generated DH parameters are still accepted by DH_check
179 but DH_generate_key works around that by clearing bit 0 of the
180 private key for those. This avoids leaking bit 0 of the private key.
181 [Bernd Edlinger]
182
183 *) Significantly reduce secure memory usage by the randomness pools.
184 [Paul Dale]
185
186 *) {CRYPTO,OPENSSL}_mem_debug_{push,pop} are now no-ops and have been
187 deprecated.
188 [Rich Salz]
189
190 *) A new type, EVP_KEYEXCH, has been introduced to represent key exchange
191 algorithms. An implementation of a key exchange algorithm can be obtained
192 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
193 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
194 the older EVP_PKEY_derive_init() function. See the man pages for the new
195 functions for further details.
196 [Matt Caswell]
197
198 *) The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
199 [Matt Caswell]
200
201 *) Removed the function names from error messages and deprecated the
202 xxx_F_xxx define's.
203
204 *) Removed NextStep support and the macro OPENSSL_UNISTD
205 [Rich Salz]
206
207 *) Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
208 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
209 Also removed "export var as function" capability; we do not export
210 variables, only functions.
211 [Rich Salz]
212
213 *) RC5_32_set_key has been changed to return an int type, with 0 indicating
214 an error and 1 indicating success. In previous versions of OpenSSL this
215 was a void type. If a key was set longer than the maximum possible this
216 would crash.
217 [Matt Caswell]
218
219 *) Support SM2 signing and verification schemes with X509 certificate.
220 [Paul Yang]
221
222 *) Use SHA256 as the default digest for TS query in the ts app.
223 [Tomas Mraz]
224
225 *) Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
226 This checks that the salt length is at least 128 bits, the derived key
227 length is at least 112 bits, and that the iteration count is at least 1000.
228 For backwards compatibility these checks are disabled by default in the
229 default provider, but are enabled by default in the fips provider.
230 To enable or disable these checks use the control
231 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
232 [Shane Lontis]
233
234 *) Default cipher lists/suites are now available via a function, the
235 #defines are deprecated.
236 [Todd Short]
237
238 *) Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
239 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
240 for Windows Store apps easier. Also, the "no-uplink" option has been added.
241 [Kenji Mouri]
242
243 *) Join the directories crypto/x509 and crypto/x509v3
244 [Richard Levitte]
245
246 *) Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
247 This changes the size when using the genpkey app when no size is given. It
248 fixes an omission in earlier changes that changed all RSA, DSA and DH
249 generation apps to use 2048 bits by default.
250 [Kurt Roeckx]
251
252 *) Added command 'openssl kdf' that uses the EVP_KDF API.
253 [Shane Lontis]
254
255 *) Added command 'openssl mac' that uses the EVP_MAC API.
256 [Shane Lontis]
257
258 *) Added OPENSSL_info() to get diverse built-in OpenSSL data, such
259 as default directories. Also added the command 'openssl info'
260 for scripting purposes.
261 [Richard Levitte]
262
263 *) The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
264 deprecated. These undocumented functions were never integrated into the EVP
265 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
266 Bi-directional IGE mode. These modes were never formally standardised and
267 usage of these functions is believed to be very small. In particular
268 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
269 is ever used. The security implications are believed to be minimal, but
270 this issue was never fixed for backwards compatibility reasons. New code
271 should not use these modes.
272 [Matt Caswell]
273
274 *) Add prediction resistance to the DRBG reseeding process.
275 [Paul Dale]
276
277 *) Limit the number of blocks in a data unit for AES-XTS to 2^20 as
278 mandated by IEEE Std 1619-2018.
279 [Paul Dale]
280
281 *) Added newline escaping functionality to a filename when using openssl dgst.
282 This output format is to replicate the output format found in the '*sum'
283 checksum programs. This aims to preserve backward compatibility.
284 [Matt Eaton, Richard Levitte, and Paul Dale]
285
286 *) Removed the heartbeat message in DTLS feature, as it has very
287 little usage and doesn't seem to fulfill a valuable purpose.
288 The configuration option is now deprecated.
289 [Richard Levitte]
290
291 *) Changed the output of 'openssl {digestname} < file' to display the
292 digest name in its output.
293 [Richard Levitte]
294
295 *) Added a new generic trace API which provides support for enabling
296 instrumentation through trace output. This feature is mainly intended
297 as an aid for developers and is disabled by default. To utilize it,
298 OpenSSL needs to be configured with the `enable-trace` option.
299
300 If the tracing API is enabled, the application can activate trace output
301 by registering BIOs as trace channels for a number of tracing and debugging
302 categories.
303
304 The 'openssl' application has been expanded to enable any of the types
305 available via environment variables defined by the user, and serves as
306 one possible example on how to use this functionality.
307 [Richard Levitte & Matthias St. Pierre]
308
309 *) Added build tests for C++. These are generated files that only do one
310 thing, to include one public OpenSSL head file each. This tests that
311 the public header files can be usefully included in a C++ application.
312
313 This test isn't enabled by default. It can be enabled with the option
314 'enable-buildtest-c++'.
315 [Richard Levitte]
316
317 *) Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
318 [Shane Lontis]
319
320 *) Add KMAC to EVP_MAC.
321 [Shane Lontis]
322
323 *) Added property based algorithm implementation selection framework to
324 the core.
325 [Paul Dale]
326
327 *) Added SCA hardening for modular field inversion in EC_GROUP through
328 a new dedicated field_inv() pointer in EC_METHOD.
329 This also addresses a leakage affecting conversions from projective
330 to affine coordinates.
331 [Billy Bob Brumley, Nicola Tuveri]
332
333 *) Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
334 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
335 those algorithms that were already supported through the EVP_PKEY API
336 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
337 and scrypt are now wrappers that call EVP_KDF.
338 [David Makepeace]
339
340 *) Build devcrypto engine as a dynamic engine.
341 [Eneas U de Queiroz]
342
343 *) Add keyed BLAKE2 to EVP_MAC.
344 [Antoine Salon]
345
346 *) Fix a bug in the computation of the endpoint-pair shared secret used
347 by DTLS over SCTP. This breaks interoperability with older versions
348 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
349 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
350 interoperability with such broken implementations. However, enabling
351 this switch breaks interoperability with correct implementations.
352
353 *) Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
354 re-used X509_PUBKEY object if the second PUBKEY is malformed.
355 [Bernd Edlinger]
356
357 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
358 [Richard Levitte]
359
360 *) Change the license to the Apache License v2.0.
361 [Richard Levitte]
362
363 *) Change the possible version information given with OPENSSL_API_COMPAT.
364 It may be a pre-3.0.0 style numerical version number as it was defined
365 in 1.1.0, and it may also simply take the major version number.
366
367 Because of the version numbering of pre-3.0.0 releases, the values 0,
368 1 and 2 are equivalent to 0x00908000L (0.9.8), 0x10000000L (1.0.0) and
369 0x10100000L (1.1.0), respectively.
370 [Richard Levitte]
371
372 *) Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
373
374 o Major releases (indicated by incrementing the MAJOR release number)
375 may introduce incompatible API/ABI changes.
376 o Minor releases (indicated by incrementing the MINOR release number)
377 may introduce new features but retain API/ABI compatibility.
378 o Patch releases (indicated by incrementing the PATCH number)
379 are intended for bug fixes and other improvements of existing
380 features only (like improving performance or adding documentation)
381 and retain API/ABI compatibility.
382 [Richard Levitte]
383
384 *) Add support for RFC5297 SIV mode (siv128), including AES-SIV.
385 [Todd Short]
386
387 *) Remove the 'dist' target and add a tarball building script. The
388 'dist' target has fallen out of use, and it shouldn't be
389 necessary to configure just to create a source distribution.
390 [Richard Levitte]
391
392 *) Recreate the OS390-Unix config target. It no longer relies on a
393 special script like it did for OpenSSL pre-1.1.0.
394 [Richard Levitte]
395
396 *) Instead of having the source directories listed in Configure, add
397 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
398 look into.
399 [Richard Levitte]
400
401 *) Add GMAC to EVP_MAC.
402 [Paul Dale]
403
404 *) Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
405 [Richard Levitte]
406
407 *) Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
408 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
409 to facilitate the continued use of MACs through raw private keys in
410 functionality such as EVP_DigestSign* and EVP_DigestVerify*.
411 [Richard Levitte]
412
413 *) Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
414 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
415 [Antoine Salon]
416
417 *) Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
418 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
419 are retained for backwards compatibility.
420 [Antoine Salon]
421
422 *) AES-XTS mode now enforces that its two keys are different to mitigate
423 the attacked described in "Efficient Instantiations of Tweakable
424 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
425 Details of this attack can be obtained from:
426 http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf
427 [Paul Dale]
428
429 *) Rename the object files, i.e. give them other names than in previous
430 versions. Their names now include the name of the final product, as
431 well as its type mnemonic (bin, lib, shlib).
432 [Richard Levitte]
433
434 *) Added new option for 'openssl list', '-objects', which will display the
435 list of built in objects, i.e. OIDs with names.
436 [Richard Levitte]
437
438 *) Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
439 improves application performance by removing data copies and providing
440 applications with zero-copy system calls such as sendfile and splice.
441 [Boris Pismenny]
442
443 Changes between 1.1.1a and 1.1.1b [xx XXX xxxx]
444
445 *) Change the info callback signals for the start and end of a post-handshake
446 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
447 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
448 confused by this and assume that a TLSv1.2 renegotiation has started. This
449 can break KeyUpdate handling. Instead we no longer signal the start and end
450 of a post handshake message exchange (although the messages themselves are
451 still signalled). This could break some applications that were expecting
452 the old signals. However without this KeyUpdate is not usable for many
453 applications.
454 [Matt Caswell]
455
456 Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
457
458 *) Timing vulnerability in DSA signature generation
459
460 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
461 timing side channel attack. An attacker could use variations in the signing
462 algorithm to recover the private key.
463
464 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
465 (CVE-2018-0734)
466 [Paul Dale]
467
468 *) Timing vulnerability in ECDSA signature generation
469
470 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
471 timing side channel attack. An attacker could use variations in the signing
472 algorithm to recover the private key.
473
474 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
475 (CVE-2018-0735)
476 [Paul Dale]
477
478 *) Fixed the issue that RAND_add()/RAND_seed() silently discards random input
479 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
480 of two gigabytes and the error handling improved.
481
482 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
483 categorized as a normal bug, not a security issue, because the DRBG reseeds
484 automatically and is fully functional even without additional randomness
485 provided by the application.
486
487 Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
488
489 *) Add a new ClientHello callback. Provides a callback interface that gives
490 the application the ability to adjust the nascent SSL object at the
491 earliest stage of ClientHello processing, immediately after extensions have
492 been collected but before they have been processed. In particular, this
493 callback can adjust the supported TLS versions in response to the contents
494 of the ClientHello
495 [Benjamin Kaduk]
496
497 *) Add SM2 base algorithm support.
498 [Jack Lloyd]
499
500 *) s390x assembly pack: add (improved) hardware-support for the following
501 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
502 aes-cfb/cfb8, aes-ecb.
503 [Patrick Steuer]
504
505 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
506 parameter is no longer accepted, as it leads to a corrupt table. NULL
507 pem_str is reserved for alias entries only.
508 [Richard Levitte]
509
510 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
511 step for prime curves. The new implementation is based on formulae from
512 differential addition-and-doubling in homogeneous projective coordinates
513 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
514 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
515 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
516 to work in projective coordinates.
517 [Billy Bob Brumley, Nicola Tuveri]
518
519 *) Change generating and checking of primes so that the error rate of not
520 being prime depends on the intended use based on the size of the input.
521 For larger primes this will result in more rounds of Miller-Rabin.
522 The maximal error rate for primes with more than 1080 bits is lowered
523 to 2^-128.
524 [Kurt Roeckx, Annie Yousar]
525
526 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
527 [Kurt Roeckx]
528
529 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
530 moving between systems, and to avoid confusion when a Windows build is
531 done with mingw vs with MSVC. For POSIX installs, there's still a
532 symlink or copy named 'tsget' to avoid that confusion as well.
533 [Richard Levitte]
534
535 *) Revert blinding in ECDSA sign and instead make problematic addition
536 length-invariant. Switch even to fixed-length Montgomery multiplication.
537 [Andy Polyakov]
538
539 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
540 step for binary curves. The new implementation is based on formulae from
541 differential addition-and-doubling in mixed Lopez-Dahab projective
542 coordinates, modified to independently blind the operands.
543 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
544
545 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
546 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
547 EC_METHODs to implement their own specialized "ladder step", to take
548 advantage of more favorable coordinate systems or more efficient
549 differential addition-and-doubling algorithms.
550 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
551
552 *) Modified the random device based seed sources to keep the relevant
553 file descriptors open rather than reopening them on each access.
554 This allows such sources to operate in a chroot() jail without
555 the associated device nodes being available. This behaviour can be
556 controlled using RAND_keep_random_devices_open().
557 [Paul Dale]
558
559 *) Numerous side-channel attack mitigations have been applied. This may have
560 performance impacts for some algorithms for the benefit of improved
561 security. Specific changes are noted in this change log by their respective
562 authors.
563 [Matt Caswell]
564
565 *) AIX shared library support overhaul. Switch to AIX "natural" way of
566 handling shared libraries, which means collecting shared objects of
567 different versions and bitnesses in one common archive. This allows to
568 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
569 doesn't affect the way 3rd party applications are linked, only how
570 multi-version installation is managed.
571 [Andy Polyakov]
572
573 *) Make ec_group_do_inverse_ord() more robust and available to other
574 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
575 mitigations are applied to the fallback BN_mod_inverse().
576 When using this function rather than BN_mod_inverse() directly, new
577 EC cryptosystem implementations are then safer-by-default.
578 [Billy Bob Brumley]
579
580 *) Add coordinate blinding for EC_POINT and implement projective
581 coordinate blinding for generic prime curves as a countermeasure to
582 chosen point SCA attacks.
583 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
584
585 *) Add blinding to ECDSA and DSA signatures to protect against side channel
586 attacks discovered by Keegan Ryan (NCC Group).
587 [Matt Caswell]
588
589 *) Enforce checking in the pkeyutl command line app to ensure that the input
590 length does not exceed the maximum supported digest length when performing
591 a sign, verify or verifyrecover operation.
592 [Matt Caswell]
593
594 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
595 I/O in combination with something like select() or poll() will hang. This
596 can be turned off again using SSL_CTX_clear_mode().
597 Many applications do not properly handle non-application data records, and
598 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
599 around the problems in those applications, but can also break some.
600 It's recommended to read the manpages about SSL_read(), SSL_write(),
601 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
602 SSL_CTX_set_read_ahead() again.
603 [Kurt Roeckx]
604
605 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
606 now allow empty (zero character) pass phrases.
607 [Richard Levitte]
608
609 *) Apply blinding to binary field modular inversion and remove patent
610 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
611 [Billy Bob Brumley]
612
613 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
614 binary and prime elliptic curves.
615 [Billy Bob Brumley]
616
617 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
618 constant time fixed point multiplication.
619 [Billy Bob Brumley]
620
621 *) Revise elliptic curve scalar multiplication with timing attack
622 defenses: ec_wNAF_mul redirects to a constant time implementation
623 when computing fixed point and variable point multiplication (which
624 in OpenSSL are mostly used with secret scalars in keygen, sign,
625 ECDH derive operations).
626 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
627 Sohaib ul Hassan]
628
629 *) Updated CONTRIBUTING
630 [Rich Salz]
631
632 *) Updated DRBG / RAND to request nonce and additional low entropy
633 randomness from the system.
634 [Matthias St. Pierre]
635
636 *) Updated 'openssl rehash' to use OpenSSL consistent default.
637 [Richard Levitte]
638
639 *) Moved the load of the ssl_conf module to libcrypto, which helps
640 loading engines that libssl uses before libssl is initialised.
641 [Matt Caswell]
642
643 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
644 [Matt Caswell]
645
646 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
647 [Ingo Schwarze, Rich Salz]
648
649 *) Added output of accepting IP address and port for 'openssl s_server'
650 [Richard Levitte]
651
652 *) Added a new API for TLSv1.3 ciphersuites:
653 SSL_CTX_set_ciphersuites()
654 SSL_set_ciphersuites()
655 [Matt Caswell]
656
657 *) Memory allocation failures consistently add an error to the error
658 stack.
659 [Rich Salz]
660
661 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
662 in libcrypto when run as setuid/setgid.
663 [Bernd Edlinger]
664
665 *) Load any config file by default when libssl is used.
666 [Matt Caswell]
667
668 *) Added new public header file <openssl/rand_drbg.h> and documentation
669 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
670 [Matthias St. Pierre]
671
672 *) QNX support removed (cannot find contributors to get their approval
673 for the license change).
674 [Rich Salz]
675
676 *) TLSv1.3 replay protection for early data has been implemented. See the
677 SSL_read_early_data() man page for further details.
678 [Matt Caswell]
679
680 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
681 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
682 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
683 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
684 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
685 configuration has been separated out. See the ciphers man page or the
686 SSL_CTX_set_ciphersuites() man page for more information.
687 [Matt Caswell]
688
689 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
690 in responder mode now supports the new "-multi" option, which
691 spawns the specified number of child processes to handle OCSP
692 requests. The "-timeout" option now also limits the OCSP
693 responder's patience to wait to receive the full client request
694 on a newly accepted connection. Child processes are respawned
695 as needed, and the CA index file is automatically reloaded
696 when changed. This makes it possible to run the "ocsp" responder
697 as a long-running service, making the OpenSSL CA somewhat more
698 feature-complete. In this mode, most diagnostic messages logged
699 after entering the event loop are logged via syslog(3) rather than
700 written to stderr.
701 [Viktor Dukhovni]
702
703 *) Added support for X448 and Ed448. Heavily based on original work by
704 Mike Hamburg.
705 [Matt Caswell]
706
707 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
708 objects loaded. This adds the functions OSSL_STORE_expect() and
709 OSSL_STORE_find() as well as needed tools to construct searches and
710 get the search data out of them.
711 [Richard Levitte]
712
713 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
714 version of OpenSSL should review their configuration settings to ensure
715 that they are still appropriate for TLSv1.3. For further information see:
716 https://wiki.openssl.org/index.php/TLS1.3
717 [Matt Caswell]
718
719 *) Grand redesign of the OpenSSL random generator
720
721 The default RAND method now utilizes an AES-CTR DRBG according to
722 NIST standard SP 800-90Ar1. The new random generator is essentially
723 a port of the default random generator from the OpenSSL FIPS 2.0
724 object module. It is a hybrid deterministic random bit generator
725 using an AES-CTR bit stream and which seeds and reseeds itself
726 automatically using trusted system entropy sources.
727
728 Some of its new features are:
729 o Support for multiple DRBG instances with seed chaining.
730 o The default RAND method makes use of a DRBG.
731 o There is a public and private DRBG instance.
732 o The DRBG instances are fork-safe.
733 o Keep all global DRBG instances on the secure heap if it is enabled.
734 o The public and private DRBG instance are per thread for lock free
735 operation
736 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
737
738 *) Changed Configure so it only says what it does and doesn't dump
739 so much data. Instead, ./configdata.pm should be used as a script
740 to display all sorts of configuration data.
741 [Richard Levitte]
742
743 *) Added processing of "make variables" to Configure.
744 [Richard Levitte]
745
746 *) Added SHA512/224 and SHA512/256 algorithm support.
747 [Paul Dale]
748
749 *) The last traces of Netware support, first removed in 1.1.0, have
750 now been removed.
751 [Rich Salz]
752
753 *) Get rid of Makefile.shared, and in the process, make the processing
754 of certain files (rc.obj, or the .def/.map/.opt files produced from
755 the ordinal files) more visible and hopefully easier to trace and
756 debug (or make silent).
757 [Richard Levitte]
758
759 *) Make it possible to have environment variable assignments as
760 arguments to config / Configure.
761 [Richard Levitte]
762
763 *) Add multi-prime RSA (RFC 8017) support.
764 [Paul Yang]
765
766 *) Add SM3 implemented according to GB/T 32905-2016
767 [ Jack Lloyd <jack.lloyd@ribose.com>,
768 Ronald Tse <ronald.tse@ribose.com>,
769 Erick Borsboom <erick.borsboom@ribose.com> ]
770
771 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
772 as documented in RFC6066.
773 Based on a patch from Tomasz Moń
774 [Filipe Raimundo da Silva]
775
776 *) Add SM4 implemented according to GB/T 32907-2016.
777 [ Jack Lloyd <jack.lloyd@ribose.com>,
778 Ronald Tse <ronald.tse@ribose.com>,
779 Erick Borsboom <erick.borsboom@ribose.com> ]
780
781 *) Reimplement -newreq-nodes and ERR_error_string_n; the
782 original author does not agree with the license change.
783 [Rich Salz]
784
785 *) Add ARIA AEAD TLS support.
786 [Jon Spillett]
787
788 *) Some macro definitions to support VS6 have been removed. Visual
789 Studio 6 has not worked since 1.1.0
790 [Rich Salz]
791
792 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
793 without clearing the errors.
794 [Richard Levitte]
795
796 *) Add "atfork" functions. If building on a system that without
797 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
798 requirements. The RAND facility now uses/requires this.
799 [Rich Salz]
800
801 *) Add SHA3.
802 [Andy Polyakov]
803
804 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
805 not possible to disable entirely. However, it's still possible to
806 disable the console reading UI method, UI_OpenSSL() (use UI_null()
807 as a fallback).
808
809 To disable, configure with 'no-ui-console'. 'no-ui' is still
810 possible to use as an alias. Check at compile time with the
811 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
812 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
813 [Richard Levitte]
814
815 *) Add a STORE module, which implements a uniform and URI based reader of
816 stores that can contain keys, certificates, CRLs and numerous other
817 objects. The main API is loosely based on a few stdio functions,
818 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
819 OSSL_STORE_error and OSSL_STORE_close.
820 The implementation uses backends called "loaders" to implement arbitrary
821 URI schemes. There is one built in "loader" for the 'file' scheme.
822 [Richard Levitte]
823
824 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
825 then adjusted to work on FreeBSD 8.4 as well.
826 Enable by configuring with 'enable-devcryptoeng'. This is done by default
827 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
828 [Richard Levitte]
829
830 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
831 util/mkerr.pl, which is adapted to allow those prefixes, leading to
832 error code calls like this:
833
834 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
835
836 With this change, we claim the namespaces OSSL and OPENSSL in a manner
837 that can be encoded in C. For the foreseeable future, this will only
838 affect new modules.
839 [Richard Levitte and Tim Hudson]
840
841 *) Removed BSD cryptodev engine.
842 [Rich Salz]
843
844 *) Add a build target 'build_all_generated', to build all generated files
845 and only that. This can be used to prepare everything that requires
846 things like perl for a system that lacks perl and then move everything
847 to that system and do the rest of the build there.
848 [Richard Levitte]
849
850 *) In the UI interface, make it possible to duplicate the user data. This
851 can be used by engines that need to retain the data for a longer time
852 than just the call where this user data is passed.
853 [Richard Levitte]
854
855 *) Ignore the '-named_curve auto' value for compatibility of applications
856 with OpenSSL 1.0.2.
857 [Tomas Mraz <tmraz@fedoraproject.org>]
858
859 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
860 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
861 alerts across multiple records (some of which could be empty). In practice
862 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
863 prohibts this altogether and other libraries (BoringSSL, NSS) do not
864 support this at all. Supporting it adds significant complexity to the
865 record layer, and its removal is unlikely to cause inter-operability
866 issues.
867 [Matt Caswell]
868
869 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
870 with Z. These are meant to replace LONG and ZLONG and to be size safe.
871 The use of LONG and ZLONG is discouraged and scheduled for deprecation
872 in OpenSSL 1.2.0.
873 [Richard Levitte]
874
875 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
876 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
877 [Richard Levitte, Andy Polyakov]
878
879 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
880 does for RSA, etc.
881 [Richard Levitte]
882
883 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
884 platform rather than 'mingw'.
885 [Richard Levitte]
886
887 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
888 success if they are asked to add an object which already exists
889 in the store. This change cascades to other functions which load
890 certificates and CRLs.
891 [Paul Dale]
892
893 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
894 facilitate stack unwinding even from assembly subroutines.
895 [Andy Polyakov]
896
897 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
898 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
899 [Richard Levitte]
900
901 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
902 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
903 which is the minimum version we support.
904 [Richard Levitte]
905
906 *) Certificate time validation (X509_cmp_time) enforces stricter
907 compliance with RFC 5280. Fractional seconds and timezone offsets
908 are no longer allowed.
909 [Emilia Käsper]
910
911 *) Add support for ARIA
912 [Paul Dale]
913
914 *) s_client will now send the Server Name Indication (SNI) extension by
915 default unless the new "-noservername" option is used. The server name is
916 based on the host provided to the "-connect" option unless overridden by
917 using "-servername".
918 [Matt Caswell]
919
920 *) Add support for SipHash
921 [Todd Short]
922
923 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
924 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
925 prevent issues where no progress is being made and the peer continually
926 sends unrecognised record types, using up resources processing them.
927 [Matt Caswell]
928
929 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
930 using the algorithm defined in
931 https://www.akkadia.org/drepper/SHA-crypt.txt
932 [Richard Levitte]
933
934 *) Heartbeat support has been removed; the ABI is changed for now.
935 [Richard Levitte, Rich Salz]
936
937 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
938 [Emilia Käsper]
939
940 *) The RSA "null" method, which was partially supported to avoid patent
941 issues, has been replaced to always returns NULL.
942 [Rich Salz]
943
944
945 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
946
947 *) Client DoS due to large DH parameter
948
949 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
950 malicious server can send a very large prime value to the client. This will
951 cause the client to spend an unreasonably long period of time generating a
952 key for this prime resulting in a hang until the client has finished. This
953 could be exploited in a Denial Of Service attack.
954
955 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
956 (CVE-2018-0732)
957 [Guido Vranken]
958
959 *) Cache timing vulnerability in RSA Key Generation
960
961 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
962 a cache timing side channel attack. An attacker with sufficient access to
963 mount cache timing attacks during the RSA key generation process could
964 recover the private key.
965
966 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
967 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
968 (CVE-2018-0737)
969 [Billy Brumley]
970
971 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
972 parameter is no longer accepted, as it leads to a corrupt table. NULL
973 pem_str is reserved for alias entries only.
974 [Richard Levitte]
975
976 *) Revert blinding in ECDSA sign and instead make problematic addition
977 length-invariant. Switch even to fixed-length Montgomery multiplication.
978 [Andy Polyakov]
979
980 *) Change generating and checking of primes so that the error rate of not
981 being prime depends on the intended use based on the size of the input.
982 For larger primes this will result in more rounds of Miller-Rabin.
983 The maximal error rate for primes with more than 1080 bits is lowered
984 to 2^-128.
985 [Kurt Roeckx, Annie Yousar]
986
987 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
988 [Kurt Roeckx]
989
990 *) Add blinding to ECDSA and DSA signatures to protect against side channel
991 attacks discovered by Keegan Ryan (NCC Group).
992 [Matt Caswell]
993
994 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
995 now allow empty (zero character) pass phrases.
996 [Richard Levitte]
997
998 *) Certificate time validation (X509_cmp_time) enforces stricter
999 compliance with RFC 5280. Fractional seconds and timezone offsets
1000 are no longer allowed.
1001 [Emilia Käsper]
1002
1003 *) Fixed a text canonicalisation bug in CMS
1004
1005 Where a CMS detached signature is used with text content the text goes
1006 through a canonicalisation process first prior to signing or verifying a
1007 signature. This process strips trailing space at the end of lines, converts
1008 line terminators to CRLF and removes additional trailing line terminators
1009 at the end of a file. A bug in the canonicalisation process meant that
1010 some characters, such as form-feed, were incorrectly treated as whitespace
1011 and removed. This is contrary to the specification (RFC5485). This fix
1012 could mean that detached text data signed with an earlier version of
1013 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
1014 signed with a fixed OpenSSL may fail to verify with an earlier version of
1015 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
1016 and use the "-binary" flag (for the "cms" command line application) or set
1017 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
1018 [Matt Caswell]
1019
1020 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
1021
1022 *) Constructed ASN.1 types with a recursive definition could exceed the stack
1023
1024 Constructed ASN.1 types with a recursive definition (such as can be found
1025 in PKCS7) could eventually exceed the stack given malicious input with
1026 excessive recursion. This could result in a Denial Of Service attack. There
1027 are no such structures used within SSL/TLS that come from untrusted sources
1028 so this is considered safe.
1029
1030 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
1031 project.
1032 (CVE-2018-0739)
1033 [Matt Caswell]
1034
1035 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
1036
1037 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
1038 effectively reduced to only comparing the least significant bit of each
1039 byte. This allows an attacker to forge messages that would be considered as
1040 authenticated in an amount of tries lower than that guaranteed by the
1041 security claims of the scheme. The module can only be compiled by the
1042 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
1043
1044 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
1045 (IBM).
1046 (CVE-2018-0733)
1047 [Andy Polyakov]
1048
1049 *) Add a build target 'build_all_generated', to build all generated files
1050 and only that. This can be used to prepare everything that requires
1051 things like perl for a system that lacks perl and then move everything
1052 to that system and do the rest of the build there.
1053 [Richard Levitte]
1054
1055 *) Backport SSL_OP_NO_RENGOTIATION
1056
1057 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
1058 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
1059 changes this is no longer possible in 1.1.0. Therefore the new
1060 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
1061 1.1.0 to provide equivalent functionality.
1062
1063 Note that if an application built against 1.1.0h headers (or above) is run
1064 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
1065 accepted but nothing will happen, i.e. renegotiation will not be prevented.
1066 [Matt Caswell]
1067
1068 *) Removed the OS390-Unix config target. It relied on a script that doesn't
1069 exist.
1070 [Rich Salz]
1071
1072 *) rsaz_1024_mul_avx2 overflow bug on x86_64
1073
1074 There is an overflow bug in the AVX2 Montgomery multiplication procedure
1075 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
1076 Analysis suggests that attacks against RSA and DSA as a result of this
1077 defect would be very difficult to perform and are not believed likely.
1078 Attacks against DH1024 are considered just feasible, because most of the
1079 work necessary to deduce information about a private key may be performed
1080 offline. The amount of resources required for such an attack would be
1081 significant. However, for an attack on TLS to be meaningful, the server
1082 would have to share the DH1024 private key among multiple clients, which is
1083 no longer an option since CVE-2016-0701.
1084
1085 This only affects processors that support the AVX2 but not ADX extensions
1086 like Intel Haswell (4th generation).
1087
1088 This issue was reported to OpenSSL by David Benjamin (Google). The issue
1089 was originally found via the OSS-Fuzz project.
1090 (CVE-2017-3738)
1091 [Andy Polyakov]
1092
1093 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
1094
1095 *) bn_sqrx8x_internal carry bug on x86_64
1096
1097 There is a carry propagating bug in the x86_64 Montgomery squaring
1098 procedure. No EC algorithms are affected. Analysis suggests that attacks
1099 against RSA and DSA as a result of this defect would be very difficult to
1100 perform and are not believed likely. Attacks against DH are considered just
1101 feasible (although very difficult) because most of the work necessary to
1102 deduce information about a private key may be performed offline. The amount
1103 of resources required for such an attack would be very significant and
1104 likely only accessible to a limited number of attackers. An attacker would
1105 additionally need online access to an unpatched system using the target
1106 private key in a scenario with persistent DH parameters and a private
1107 key that is shared between multiple clients.
1108
1109 This only affects processors that support the BMI1, BMI2 and ADX extensions
1110 like Intel Broadwell (5th generation) and later or AMD Ryzen.
1111
1112 This issue was reported to OpenSSL by the OSS-Fuzz project.
1113 (CVE-2017-3736)
1114 [Andy Polyakov]
1115
1116 *) Malformed X.509 IPAddressFamily could cause OOB read
1117
1118 If an X.509 certificate has a malformed IPAddressFamily extension,
1119 OpenSSL could do a one-byte buffer overread. The most likely result
1120 would be an erroneous display of the certificate in text format.
1121
1122 This issue was reported to OpenSSL by the OSS-Fuzz project.
1123 (CVE-2017-3735)
1124 [Rich Salz]
1125
1126 Changes between 1.1.0e and 1.1.0f [25 May 2017]
1127
1128 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
1129 platform rather than 'mingw'.
1130 [Richard Levitte]
1131
1132 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
1133 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
1134 which is the minimum version we support.
1135 [Richard Levitte]
1136
1137 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
1138
1139 *) Encrypt-Then-Mac renegotiation crash
1140
1141 During a renegotiation handshake if the Encrypt-Then-Mac extension is
1142 negotiated where it was not in the original handshake (or vice-versa) then
1143 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
1144 and servers are affected.
1145
1146 This issue was reported to OpenSSL by Joe Orton (Red Hat).
1147 (CVE-2017-3733)
1148 [Matt Caswell]
1149
1150 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
1151
1152 *) Truncated packet could crash via OOB read
1153
1154 If one side of an SSL/TLS path is running on a 32-bit host and a specific
1155 cipher is being used, then a truncated packet can cause that host to
1156 perform an out-of-bounds read, usually resulting in a crash.
1157
1158 This issue was reported to OpenSSL by Robert Święcki of Google.
1159 (CVE-2017-3731)
1160 [Andy Polyakov]
1161
1162 *) Bad (EC)DHE parameters cause a client crash
1163
1164 If a malicious server supplies bad parameters for a DHE or ECDHE key
1165 exchange then this can result in the client attempting to dereference a
1166 NULL pointer leading to a client crash. This could be exploited in a Denial
1167 of Service attack.
1168
1169 This issue was reported to OpenSSL by Guido Vranken.
1170 (CVE-2017-3730)
1171 [Matt Caswell]
1172
1173 *) BN_mod_exp may produce incorrect results on x86_64
1174
1175 There is a carry propagating bug in the x86_64 Montgomery squaring
1176 procedure. No EC algorithms are affected. Analysis suggests that attacks
1177 against RSA and DSA as a result of this defect would be very difficult to
1178 perform and are not believed likely. Attacks against DH are considered just
1179 feasible (although very difficult) because most of the work necessary to
1180 deduce information about a private key may be performed offline. The amount
1181 of resources required for such an attack would be very significant and
1182 likely only accessible to a limited number of attackers. An attacker would
1183 additionally need online access to an unpatched system using the target
1184 private key in a scenario with persistent DH parameters and a private
1185 key that is shared between multiple clients. For example this can occur by
1186 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
1187 similar to CVE-2015-3193 but must be treated as a separate problem.
1188
1189 This issue was reported to OpenSSL by the OSS-Fuzz project.
1190 (CVE-2017-3732)
1191 [Andy Polyakov]
1192
1193 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
1194
1195 *) ChaCha20/Poly1305 heap-buffer-overflow
1196
1197 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
1198 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
1199 crash. This issue is not considered to be exploitable beyond a DoS.
1200
1201 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
1202 (CVE-2016-7054)
1203 [Richard Levitte]
1204
1205 *) CMS Null dereference
1206
1207 Applications parsing invalid CMS structures can crash with a NULL pointer
1208 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
1209 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
1210 structure callback if an attempt is made to free certain invalid encodings.
1211 Only CHOICE structures using a callback which do not handle NULL value are
1212 affected.
1213
1214 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
1215 (CVE-2016-7053)
1216 [Stephen Henson]
1217
1218 *) Montgomery multiplication may produce incorrect results
1219
1220 There is a carry propagating bug in the Broadwell-specific Montgomery
1221 multiplication procedure that handles input lengths divisible by, but
1222 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
1223 and DH private keys are impossible. This is because the subroutine in
1224 question is not used in operations with the private key itself and an input
1225 of the attacker's direct choice. Otherwise the bug can manifest itself as
1226 transient authentication and key negotiation failures or reproducible
1227 erroneous outcome of public-key operations with specially crafted input.
1228 Among EC algorithms only Brainpool P-512 curves are affected and one
1229 presumably can attack ECDH key negotiation. Impact was not analyzed in
1230 detail, because pre-requisites for attack are considered unlikely. Namely
1231 multiple clients have to choose the curve in question and the server has to
1232 share the private key among them, neither of which is default behaviour.
1233 Even then only clients that chose the curve will be affected.
1234
1235 This issue was publicly reported as transient failures and was not
1236 initially recognized as a security issue. Thanks to Richard Morgan for
1237 providing reproducible case.
1238 (CVE-2016-7055)
1239 [Andy Polyakov]
1240
1241 *) Removed automatic addition of RPATH in shared libraries and executables,
1242 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
1243 [Richard Levitte]
1244
1245 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
1246
1247 *) Fix Use After Free for large message sizes
1248
1249 The patch applied to address CVE-2016-6307 resulted in an issue where if a
1250 message larger than approx 16k is received then the underlying buffer to
1251 store the incoming message is reallocated and moved. Unfortunately a
1252 dangling pointer to the old location is left which results in an attempt to
1253 write to the previously freed location. This is likely to result in a
1254 crash, however it could potentially lead to execution of arbitrary code.
1255
1256 This issue only affects OpenSSL 1.1.0a.
1257
1258 This issue was reported to OpenSSL by Robert Święcki.
1259 (CVE-2016-6309)
1260 [Matt Caswell]
1261
1262 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
1263
1264 *) OCSP Status Request extension unbounded memory growth
1265
1266 A malicious client can send an excessively large OCSP Status Request
1267 extension. If that client continually requests renegotiation, sending a
1268 large OCSP Status Request extension each time, then there will be unbounded
1269 memory growth on the server. This will eventually lead to a Denial Of
1270 Service attack through memory exhaustion. Servers with a default
1271 configuration are vulnerable even if they do not support OCSP. Builds using
1272 the "no-ocsp" build time option are not affected.
1273
1274 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1275 (CVE-2016-6304)
1276 [Matt Caswell]
1277
1278 *) SSL_peek() hang on empty record
1279
1280 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
1281 sends an empty record. This could be exploited by a malicious peer in a
1282 Denial Of Service attack.
1283
1284 This issue was reported to OpenSSL by Alex Gaynor.
1285 (CVE-2016-6305)
1286 [Matt Caswell]
1287
1288 *) Excessive allocation of memory in tls_get_message_header() and
1289 dtls1_preprocess_fragment()
1290
1291 A (D)TLS message includes 3 bytes for its length in the header for the
1292 message. This would allow for messages up to 16Mb in length. Messages of
1293 this length are excessive and OpenSSL includes a check to ensure that a
1294 peer is sending reasonably sized messages in order to avoid too much memory
1295 being consumed to service a connection. A flaw in the logic of version
1296 1.1.0 means that memory for the message is allocated too early, prior to
1297 the excessive message length check. Due to way memory is allocated in
1298 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
1299 to service a connection. This could lead to a Denial of Service through
1300 memory exhaustion. However, the excessive message length check still takes
1301 place, and this would cause the connection to immediately fail. Assuming
1302 that the application calls SSL_free() on the failed connection in a timely
1303 manner then the 21Mb of allocated memory will then be immediately freed
1304 again. Therefore the excessive memory allocation will be transitory in
1305 nature. This then means that there is only a security impact if:
1306
1307 1) The application does not call SSL_free() in a timely manner in the event
1308 that the connection fails
1309 or
1310 2) The application is working in a constrained environment where there is
1311 very little free memory
1312 or
1313 3) The attacker initiates multiple connection attempts such that there are
1314 multiple connections in a state where memory has been allocated for the
1315 connection; SSL_free() has not yet been called; and there is insufficient
1316 memory to service the multiple requests.
1317
1318 Except in the instance of (1) above any Denial Of Service is likely to be
1319 transitory because as soon as the connection fails the memory is
1320 subsequently freed again in the SSL_free() call. However there is an
1321 increased risk during this period of application crashes due to the lack of
1322 memory - which would then mean a more serious Denial of Service.
1323
1324 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1325 (CVE-2016-6307 and CVE-2016-6308)
1326 [Matt Caswell]
1327
1328 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
1329 had to be removed. Primary reason is that vendor assembler can't
1330 assemble our modules with -KPIC flag. As result it, assembly
1331 support, was not even available as option. But its lack means
1332 lack of side-channel resistant code, which is incompatible with
1333 security by todays standards. Fortunately gcc is readily available
1334 prepackaged option, which we firmly point at...
1335 [Andy Polyakov]
1336
1337 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
1338
1339 *) Windows command-line tool supports UTF-8 opt-in option for arguments
1340 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
1341 (to any value) allows Windows user to access PKCS#12 file generated
1342 with Windows CryptoAPI and protected with non-ASCII password, as well
1343 as files generated under UTF-8 locale on Linux also protected with
1344 non-ASCII password.
1345 [Andy Polyakov]
1346
1347 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
1348 have been disabled by default and removed from DEFAULT, just like RC4.
1349 See the RC4 item below to re-enable both.
1350 [Rich Salz]
1351
1352 *) The method for finding the storage location for the Windows RAND seed file
1353 has changed. First we check %RANDFILE%. If that is not set then we check
1354 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
1355 all else fails we fall back to C:\.
1356 [Matt Caswell]
1357
1358 *) The EVP_EncryptUpdate() function has had its return type changed from void
1359 to int. A return of 0 indicates and error while a return of 1 indicates
1360 success.
1361 [Matt Caswell]
1362
1363 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
1364 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
1365 off the constant time implementation for RSA, DSA and DH have been made
1366 no-ops and deprecated.
1367 [Matt Caswell]
1368
1369 *) Windows RAND implementation was simplified to only get entropy by
1370 calling CryptGenRandom(). Various other RAND-related tickets
1371 were also closed.
1372 [Joseph Wylie Yandle, Rich Salz]
1373
1374 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
1375 and OPENSSL_LH_, respectively. The old names are available
1376 with API compatibility. They new names are now completely documented.
1377 [Rich Salz]
1378
1379 *) Unify TYPE_up_ref(obj) methods signature.
1380 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
1381 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
1382 int (instead of void) like all others TYPE_up_ref() methods.
1383 So now these methods also check the return value of CRYPTO_atomic_add(),
1384 and the validity of object reference counter.
1385 [fdasilvayy@gmail.com]
1386
1387 *) With Windows Visual Studio builds, the .pdb files are installed
1388 alongside the installed libraries and executables. For a static
1389 library installation, ossl_static.pdb is the associate compiler
1390 generated .pdb file to be used when linking programs.
1391 [Richard Levitte]
1392
1393 *) Remove openssl.spec. Packaging files belong with the packagers.
1394 [Richard Levitte]
1395
1396 *) Automatic Darwin/OSX configuration has had a refresh, it will now
1397 recognise x86_64 architectures automatically. You can still decide
1398 to build for a different bitness with the environment variable
1399 KERNEL_BITS (can be 32 or 64), for example:
1400
1401 KERNEL_BITS=32 ./config
1402
1403 [Richard Levitte]
1404
1405 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
1406 256 bit AES and HMAC with SHA256.
1407 [Steve Henson]
1408
1409 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
1410 [Andy Polyakov]
1411
1412 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
1413 [Rich Salz]
1414
1415 *) To enable users to have their own config files and build file templates,
1416 Configure looks in the directory indicated by the environment variable
1417 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
1418 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
1419 name and is used as is.
1420 [Richard Levitte]
1421
1422 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
1423 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
1424 X509_CERT_FILE_CTX was removed.
1425 [Rich Salz]
1426
1427 *) "shared" builds are now the default. To create only static libraries use
1428 the "no-shared" Configure option.
1429 [Matt Caswell]
1430
1431 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
1432 All of these option have not worked for some while and are fundamental
1433 algorithms.
1434 [Matt Caswell]
1435
1436 *) Make various cleanup routines no-ops and mark them as deprecated. Most
1437 global cleanup functions are no longer required because they are handled
1438 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
1439 Explicitly de-initing can cause problems (e.g. where a library that uses
1440 OpenSSL de-inits, but an application is still using it). The affected
1441 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
1442 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
1443 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
1444 COMP_zlib_cleanup().
1445 [Matt Caswell]
1446
1447 *) --strict-warnings no longer enables runtime debugging options
1448 such as REF_DEBUG. Instead, debug options are automatically
1449 enabled with '--debug' builds.
1450 [Andy Polyakov, Emilia Käsper]
1451
1452 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
1453 have been moved out of the public header files. New functions for managing
1454 these have been added.
1455 [Matt Caswell]
1456
1457 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
1458 objects have been moved out of the public header files. New
1459 functions for managing these have been added.
1460 [Richard Levitte]
1461
1462 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1463 have been moved out of the public header files. New functions for managing
1464 these have been added.
1465 [Matt Caswell]
1466
1467 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1468 moved out of the public header files. New functions for managing these
1469 have been added.
1470 [Matt Caswell]
1471
1472 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
1473 [Matt Caswell]
1474
1475 *) Removed the mk1mf build scripts.
1476 [Richard Levitte]
1477
1478 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1479 it is always safe to #include a header now.
1480 [Rich Salz]
1481
1482 *) Removed the aged BC-32 config and all its supporting scripts
1483 [Richard Levitte]
1484
1485 *) Removed support for Ultrix, Netware, and OS/2.
1486 [Rich Salz]
1487
1488 *) Add support for HKDF.
1489 [Alessandro Ghedini]
1490
1491 *) Add support for blake2b and blake2s
1492 [Bill Cox]
1493
1494 *) Added support for "pipelining". Ciphers that have the
1495 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1496 encryptions/decryptions simultaneously. There are currently no built-in
1497 ciphers with this property but the expectation is that engines will be able
1498 to offer it to significantly improve throughput. Support has been extended
1499 into libssl so that multiple records for a single connection can be
1500 processed in one go (for >=TLS 1.1).
1501 [Matt Caswell]
1502
1503 *) Added the AFALG engine. This is an async capable engine which is able to
1504 offload work to the Linux kernel. In this initial version it only supports
1505 AES128-CBC. The kernel must be version 4.1.0 or greater.
1506 [Catriona Lucey]
1507
1508 *) OpenSSL now uses a new threading API. It is no longer necessary to
1509 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1510 are two supported threading models: pthreads and windows threads. It is
1511 also possible to configure OpenSSL at compile time for "no-threads". The
1512 old threading API should no longer be used. The functions have been
1513 replaced with "no-op" compatibility macros.
1514 [Alessandro Ghedini, Matt Caswell]
1515
1516 *) Modify behavior of ALPN to invoke callback after SNI/servername
1517 callback, such that updates to the SSL_CTX affect ALPN.
1518 [Todd Short]
1519
1520 *) Add SSL_CIPHER queries for authentication and key-exchange.
1521 [Todd Short]
1522
1523 *) Changes to the DEFAULT cipherlist:
1524 - Prefer (EC)DHE handshakes over plain RSA.
1525 - Prefer AEAD ciphers over legacy ciphers.
1526 - Prefer ECDSA over RSA when both certificates are available.
1527 - Prefer TLSv1.2 ciphers/PRF.
1528 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1529 default cipherlist.
1530 [Emilia Käsper]
1531
1532 *) Change the ECC default curve list to be this, in order: x25519,
1533 secp256r1, secp521r1, secp384r1.
1534 [Rich Salz]
1535
1536 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1537 disabled by default. They can be re-enabled using the
1538 enable-weak-ssl-ciphers option to Configure.
1539 [Matt Caswell]
1540
1541 *) If the server has ALPN configured, but supports no protocols that the
1542 client advertises, send a fatal "no_application_protocol" alert.
1543 This behaviour is SHALL in RFC 7301, though it isn't universally
1544 implemented by other servers.
1545 [Emilia Käsper]
1546
1547 *) Add X25519 support.
1548 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1549 for public and private key encoding using the format documented in
1550 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1551 key generation and key derivation.
1552
1553 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1554 X25519(29).
1555 [Steve Henson]
1556
1557 *) Deprecate SRP_VBASE_get_by_user.
1558 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1559 In order to fix an unavoidable memory leak (CVE-2016-0798),
1560 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1561 seed, even if the seed is configured.
1562
1563 Users should use SRP_VBASE_get1_by_user instead. Note that in
1564 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1565 also that even though configuring the SRP seed attempts to hide
1566 invalid usernames by continuing the handshake with fake
1567 credentials, this behaviour is not constant time and no strong
1568 guarantees are made that the handshake is indistinguishable from
1569 that of a valid user.
1570 [Emilia Käsper]
1571
1572 *) Configuration change; it's now possible to build dynamic engines
1573 without having to build shared libraries and vice versa. This
1574 only applies to the engines in engines/, those in crypto/engine/
1575 will always be built into libcrypto (i.e. "static").
1576
1577 Building dynamic engines is enabled by default; to disable, use
1578 the configuration option "disable-dynamic-engine".
1579
1580 The only requirements for building dynamic engines are the
1581 presence of the DSO module and building with position independent
1582 code, so they will also automatically be disabled if configuring
1583 with "disable-dso" or "disable-pic".
1584
1585 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1586 are also taken away from openssl/opensslconf.h, as they are
1587 irrelevant.
1588 [Richard Levitte]
1589
1590 *) Configuration change; if there is a known flag to compile
1591 position independent code, it will always be applied on the
1592 libcrypto and libssl object files, and never on the application
1593 object files. This means other libraries that use routines from
1594 libcrypto / libssl can be made into shared libraries regardless
1595 of how OpenSSL was configured.
1596
1597 If this isn't desirable, the configuration options "disable-pic"
1598 or "no-pic" can be used to disable the use of PIC. This will
1599 also disable building shared libraries and dynamic engines.
1600 [Richard Levitte]
1601
1602 *) Removed JPAKE code. It was experimental and has no wide use.
1603 [Rich Salz]
1604
1605 *) The INSTALL_PREFIX Makefile variable has been renamed to
1606 DESTDIR. That makes for less confusion on what this variable
1607 is for. Also, the configuration option --install_prefix is
1608 removed.
1609 [Richard Levitte]
1610
1611 *) Heartbeat for TLS has been removed and is disabled by default
1612 for DTLS; configure with enable-heartbeats. Code that uses the
1613 old #define's might need to be updated.
1614 [Emilia Käsper, Rich Salz]
1615
1616 *) Rename REF_CHECK to REF_DEBUG.
1617 [Rich Salz]
1618
1619 *) New "unified" build system
1620
1621 The "unified" build system is aimed to be a common system for all
1622 platforms we support. With it comes new support for VMS.
1623
1624 This system builds supports building in a different directory tree
1625 than the source tree. It produces one Makefile (for unix family
1626 or lookalikes), or one descrip.mms (for VMS).
1627
1628 The source of information to make the Makefile / descrip.mms is
1629 small files called 'build.info', holding the necessary
1630 information for each directory with source to compile, and a
1631 template in Configurations, like unix-Makefile.tmpl or
1632 descrip.mms.tmpl.
1633
1634 With this change, the library names were also renamed on Windows
1635 and on VMS. They now have names that are closer to the standard
1636 on Unix, and include the major version number, and in certain
1637 cases, the architecture they are built for. See "Notes on shared
1638 libraries" in INSTALL.
1639
1640 We rely heavily on the perl module Text::Template.
1641 [Richard Levitte]
1642
1643 *) Added support for auto-initialisation and de-initialisation of the library.
1644 OpenSSL no longer requires explicit init or deinit routines to be called,
1645 except in certain circumstances. See the OPENSSL_init_crypto() and
1646 OPENSSL_init_ssl() man pages for further information.
1647 [Matt Caswell]
1648
1649 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1650 "peer" argument is now expected to be a BIO_ADDR object.
1651
1652 *) Rewrite of BIO networking library. The BIO library lacked consistent
1653 support of IPv6, and adding it required some more extensive
1654 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1655 which hold all types of addresses and chains of address information.
1656 It also introduces a new API, with functions like BIO_socket,
1657 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1658 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1659 have been adapted accordingly.
1660 [Richard Levitte]
1661
1662 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1663 the leading 0-byte.
1664 [Emilia Käsper]
1665
1666 *) CRIME protection: disable compression by default, even if OpenSSL is
1667 compiled with zlib enabled. Applications can still enable compression
1668 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1669 using the SSL_CONF library to configure compression.
1670 [Emilia Käsper]
1671
1672 *) The signature of the session callback configured with
1673 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1674 was explicitly marked as 'const unsigned char*' instead of
1675 'unsigned char*'.
1676 [Emilia Käsper]
1677
1678 *) Always DPURIFY. Remove the use of uninitialized memory in the
1679 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1680 [Emilia Käsper]
1681
1682 *) Removed many obsolete configuration items, including
1683 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1684 MD2_CHAR, MD2_INT, MD2_LONG
1685 BF_PTR, BF_PTR2
1686 IDEA_SHORT, IDEA_LONG
1687 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1688 [Rich Salz, with advice from Andy Polyakov]
1689
1690 *) Many BN internals have been moved to an internal header file.
1691 [Rich Salz with help from Andy Polyakov]
1692
1693 *) Configuration and writing out the results from it has changed.
1694 Files such as Makefile include/openssl/opensslconf.h and are now
1695 produced through general templates, such as Makefile.in and
1696 crypto/opensslconf.h.in and some help from the perl module
1697 Text::Template.
1698
1699 Also, the center of configuration information is no longer
1700 Makefile. Instead, Configure produces a perl module in
1701 configdata.pm which holds most of the config data (in the hash
1702 table %config), the target data that comes from the target
1703 configuration in one of the Configurations/*.conf files (in
1704 %target).
1705 [Richard Levitte]
1706
1707 *) To clarify their intended purposes, the Configure options
1708 --prefix and --openssldir change their semantics, and become more
1709 straightforward and less interdependent.
1710
1711 --prefix shall be used exclusively to give the location INSTALLTOP
1712 where programs, scripts, libraries, include files and manuals are
1713 going to be installed. The default is now /usr/local.
1714
1715 --openssldir shall be used exclusively to give the default
1716 location OPENSSLDIR where certificates, private keys, CRLs are
1717 managed. This is also where the default openssl.cnf gets
1718 installed.
1719 If the directory given with this option is a relative path, the
1720 values of both the --prefix value and the --openssldir value will
1721 be combined to become OPENSSLDIR.
1722 The default for --openssldir is INSTALLTOP/ssl.
1723
1724 Anyone who uses --openssldir to specify where OpenSSL is to be
1725 installed MUST change to use --prefix instead.
1726 [Richard Levitte]
1727
1728 *) The GOST engine was out of date and therefore it has been removed. An up
1729 to date GOST engine is now being maintained in an external repository.
1730 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1731 support for GOST ciphersuites (these are only activated if a GOST engine
1732 is present).
1733 [Matt Caswell]
1734
1735 *) EGD is no longer supported by default; use enable-egd when
1736 configuring.
1737 [Ben Kaduk and Rich Salz]
1738
1739 *) The distribution now has Makefile.in files, which are used to
1740 create Makefile's when Configure is run. *Configure must be run
1741 before trying to build now.*
1742 [Rich Salz]
1743
1744 *) The return value for SSL_CIPHER_description() for error conditions
1745 has changed.
1746 [Rich Salz]
1747
1748 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1749
1750 Obtaining and performing DNSSEC validation of TLSA records is
1751 the application's responsibility. The application provides
1752 the TLSA records of its choice to OpenSSL, and these are then
1753 used to authenticate the peer.
1754
1755 The TLSA records need not even come from DNS. They can, for
1756 example, be used to implement local end-entity certificate or
1757 trust-anchor "pinning", where the "pin" data takes the form
1758 of TLSA records, which can augment or replace verification
1759 based on the usual WebPKI public certification authorities.
1760 [Viktor Dukhovni]
1761
1762 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1763 continues to support deprecated interfaces in default builds.
1764 However, applications are strongly advised to compile their
1765 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1766 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1767 or the 1.1.0 releases.
1768
1769 In environments in which all applications have been ported to
1770 not use any deprecated interfaces OpenSSL's Configure script
1771 should be used with the --api=1.1.0 option to entirely remove
1772 support for the deprecated features from the library and
1773 unconditionally disable them in the installed headers.
1774 Essentially the same effect can be achieved with the "no-deprecated"
1775 argument to Configure, except that this will always restrict
1776 the build to just the latest API, rather than a fixed API
1777 version.
1778
1779 As applications are ported to future revisions of the API,
1780 they should update their compile-time OPENSSL_API_COMPAT define
1781 accordingly, but in most cases should be able to continue to
1782 compile with later releases.
1783
1784 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1785 0x10000000L and 0x00908000L, respectively. However those
1786 versions did not support the OPENSSL_API_COMPAT feature, and
1787 so applications are not typically tested for explicit support
1788 of just the undeprecated features of either release.
1789 [Viktor Dukhovni]
1790
1791 *) Add support for setting the minimum and maximum supported protocol.
1792 It can bet set via the SSL_set_min_proto_version() and
1793 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1794 MaxProtocol. It's recommended to use the new APIs to disable
1795 protocols instead of disabling individual protocols using
1796 SSL_set_options() or SSL_CONF's Protocol. This change also
1797 removes support for disabling TLS 1.2 in the OpenSSL TLS
1798 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1799 [Kurt Roeckx]
1800
1801 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1802 [Andy Polyakov]
1803
1804 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1805 and integrates ECDSA and ECDH functionality into EC. Implementations can
1806 now redirect key generation and no longer need to convert to or from
1807 ECDSA_SIG format.
1808
1809 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1810 include the ec.h header file instead.
1811 [Steve Henson]
1812
1813 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1814 ciphers who are no longer supported and drops support the ephemeral RSA key
1815 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1816 [Kurt Roeckx]
1817
1818 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1819 opaque. For HMAC_CTX, the following constructors and destructors
1820 were added:
1821
1822 HMAC_CTX *HMAC_CTX_new(void);
1823 void HMAC_CTX_free(HMAC_CTX *ctx);
1824
1825 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1826 destroy such methods has been added. See EVP_MD_meth_new(3) and
1827 EVP_CIPHER_meth_new(3) for documentation.
1828
1829 Additional changes:
1830 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1831 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1832 EVP_MD_CTX_reset() should be called instead to reinitialise
1833 an already created structure.
1834 2) For consistency with the majority of our object creators and
1835 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1836 EVP_MD_CTX_(new|free). The old names are retained as macros
1837 for deprecated builds.
1838 [Richard Levitte]
1839
1840 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1841 cryptographic operations to be performed asynchronously as long as an
1842 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1843 further details. Libssl has also had this capability integrated with the
1844 introduction of the new mode SSL_MODE_ASYNC and associated error
1845 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1846 pages. This work was developed in partnership with Intel Corp.
1847 [Matt Caswell]
1848
1849 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1850 always enabled now. If you want to disable the support you should
1851 exclude it using the list of supported ciphers. This also means that the
1852 "-no_ecdhe" option has been removed from s_server.
1853 [Kurt Roeckx]
1854
1855 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1856 SSL_{CTX_}set1_curves() which can set a list.
1857 [Kurt Roeckx]
1858
1859 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1860 curve you want to support using SSL_{CTX_}set1_curves().
1861 [Kurt Roeckx]
1862
1863 *) State machine rewrite. The state machine code has been significantly
1864 refactored in order to remove much duplication of code and solve issues
1865 with the old code (see ssl/statem/README for further details). This change
1866 does have some associated API changes. Notably the SSL_state() function
1867 has been removed and replaced by SSL_get_state which now returns an
1868 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1869 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1870 also been removed.
1871 [Matt Caswell]
1872
1873 *) All instances of the string "ssleay" in the public API were replaced
1874 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1875 Some error codes related to internal RSA_eay API's were renamed.
1876 [Rich Salz]
1877
1878 *) The demo files in crypto/threads were moved to demo/threads.
1879 [Rich Salz]
1880
1881 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1882 sureware and ubsec.
1883 [Matt Caswell, Rich Salz]
1884
1885 *) New ASN.1 embed macro.
1886
1887 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1888 structure is not allocated: it is part of the parent. That is instead of
1889
1890 FOO *x;
1891
1892 it must be:
1893
1894 FOO x;
1895
1896 This reduces memory fragmentation and make it impossible to accidentally
1897 set a mandatory field to NULL.
1898
1899 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1900 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1901 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1902 SEQUENCE OF.
1903 [Steve Henson]
1904
1905 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1906 [Emilia Käsper]
1907
1908 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1909 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1910 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1911 DES and RC4 ciphersuites.
1912 [Matt Caswell]
1913
1914 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1915 This changes the decoding behaviour for some invalid messages,
1916 though the change is mostly in the more lenient direction, and
1917 legacy behaviour is preserved as much as possible.
1918 [Emilia Käsper]
1919
1920 *) Fix no-stdio build.
1921 [ David Woodhouse <David.Woodhouse@intel.com> and also
1922 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1923
1924 *) New testing framework
1925 The testing framework has been largely rewritten and is now using
1926 perl and the perl modules Test::Harness and an extended variant of
1927 Test::More called OpenSSL::Test to do its work. All test scripts in
1928 test/ have been rewritten into test recipes, and all direct calls to
1929 executables in test/Makefile have become individual recipes using the
1930 simplified testing OpenSSL::Test::Simple.
1931
1932 For documentation on our testing modules, do:
1933
1934 perldoc test/testlib/OpenSSL/Test/Simple.pm
1935 perldoc test/testlib/OpenSSL/Test.pm
1936
1937 [Richard Levitte]
1938
1939 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1940 are used; the latter aborts on memory leaks (usually checked on exit).
1941 Some undocumented "set malloc, etc., hooks" functions were removed
1942 and others were changed. All are now documented.
1943 [Rich Salz]
1944
1945 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1946 return an error
1947 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1948
1949 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1950 from RFC4279, RFC4785, RFC5487, RFC5489.
1951
1952 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1953 original RSA_PSK patch.
1954 [Steve Henson]
1955
1956 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1957 era flag was never set throughout the codebase (only read). Also removed
1958 SSL3_FLAGS_POP_BUFFER which was only used if
1959 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1960 [Matt Caswell]
1961
1962 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1963 to be "oneline" instead of "compat".
1964 [Richard Levitte]
1965
1966 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1967 not aware of clients that still exhibit this bug, and the workaround
1968 hasn't been working properly for a while.
1969 [Emilia Käsper]
1970
1971 *) The return type of BIO_number_read() and BIO_number_written() as well as
1972 the corresponding num_read and num_write members in the BIO structure has
1973 changed from unsigned long to uint64_t. On platforms where an unsigned
1974 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1975 transferred.
1976 [Matt Caswell]
1977
1978 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1979 OpenSSL without support for them. It also means that maintaining
1980 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1981 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1982 [Matt Caswell]
1983
1984 *) Removed support for the two export grade static DH ciphersuites
1985 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1986 were newly added (along with a number of other static DH ciphersuites) to
1987 1.0.2. However the two export ones have *never* worked since they were
1988 introduced. It seems strange in any case to be adding new export
1989 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1990 [Matt Caswell]
1991
1992 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1993 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1994 and turned into macros which simply call the new preferred function names
1995 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1996 should use the new names instead. Also as part of this change the ssl23.h
1997 header file has been removed.
1998 [Matt Caswell]
1999
2000 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
2001 code and the associated standard is no longer considered fit-for-purpose.
2002 [Matt Caswell]
2003
2004 *) RT2547 was closed. When generating a private key, try to make the
2005 output file readable only by the owner. This behavior change might
2006 be noticeable when interacting with other software.
2007
2008 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
2009 Added a test.
2010 [Rich Salz]
2011
2012 *) Added HTTP GET support to the ocsp command.
2013 [Rich Salz]
2014
2015 *) Changed default digest for the dgst and enc commands from MD5 to
2016 sha256
2017 [Rich Salz]
2018
2019 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
2020 [Matt Caswell]
2021
2022 *) Added support for TLS extended master secret from
2023 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
2024 initial patch which was a great help during development.
2025 [Steve Henson]
2026
2027 *) All libssl internal structures have been removed from the public header
2028 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
2029 now redundant). Users should not attempt to access internal structures
2030 directly. Instead they should use the provided API functions.
2031 [Matt Caswell]
2032
2033 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
2034 Access to deprecated functions can be re-enabled by running config with
2035 "enable-deprecated". In addition applications wishing to use deprecated
2036 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
2037 will, by default, disable some transitive includes that previously existed
2038 in the header files (e.g. ec.h will no longer, by default, include bn.h)
2039 [Matt Caswell]
2040
2041 *) Added support for OCB mode. OpenSSL has been granted a patent license
2042 compatible with the OpenSSL license for use of OCB. Details are available
2043 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
2044 for OCB can be removed by calling config with no-ocb.
2045 [Matt Caswell]
2046
2047 *) SSLv2 support has been removed. It still supports receiving a SSLv2
2048 compatible client hello.
2049 [Kurt Roeckx]
2050
2051 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
2052 done while fixing the error code for the key-too-small case.
2053 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
2054
2055 *) CA.sh has been removed; use CA.pl instead.
2056 [Rich Salz]
2057
2058 *) Removed old DES API.
2059 [Rich Salz]
2060
2061 *) Remove various unsupported platforms:
2062 Sony NEWS4
2063 BEOS and BEOS_R5
2064 NeXT
2065 SUNOS
2066 MPE/iX
2067 Sinix/ReliantUNIX RM400
2068 DGUX
2069 NCR
2070 Tandem
2071 Cray
2072 16-bit platforms such as WIN16
2073 [Rich Salz]
2074
2075 *) Clean up OPENSSL_NO_xxx #define's
2076 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
2077 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
2078 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
2079 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
2080 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
2081 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
2082 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
2083 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
2084 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
2085 Remove MS_STATIC; it's a relic from platforms <32 bits.
2086 [Rich Salz]
2087
2088 *) Cleaned up dead code
2089 Remove all but one '#ifdef undef' which is to be looked at.
2090 [Rich Salz]
2091
2092 *) Clean up calling of xxx_free routines.
2093 Just like free(), fix most of the xxx_free routines to accept
2094 NULL. Remove the non-null checks from callers. Save much code.
2095 [Rich Salz]
2096
2097 *) Add secure heap for storage of private keys (when possible).
2098 Add BIO_s_secmem(), CBIGNUM, etc.
2099 Contributed by Akamai Technologies under our Corporate CLA.
2100 [Rich Salz]
2101
2102 *) Experimental support for a new, fast, unbiased prime candidate generator,
2103 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
2104 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
2105
2106 *) New output format NSS in the sess_id command line tool. This allows
2107 exporting the session id and the master key in NSS keylog format.
2108 [Martin Kaiser <martin@kaiser.cx>]
2109
2110 *) Harmonize version and its documentation. -f flag is used to display
2111 compilation flags.
2112 [mancha <mancha1@zoho.com>]
2113
2114 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
2115 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
2116 [mancha <mancha1@zoho.com>]
2117
2118 *) Fix some double frees. These are not thought to be exploitable.
2119 [mancha <mancha1@zoho.com>]
2120
2121 *) A missing bounds check in the handling of the TLS heartbeat extension
2122 can be used to reveal up to 64k of memory to a connected client or
2123 server.
2124
2125 Thanks for Neel Mehta of Google Security for discovering this bug and to
2126 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
2127 preparing the fix (CVE-2014-0160)
2128 [Adam Langley, Bodo Moeller]
2129
2130 *) Fix for the attack described in the paper "Recovering OpenSSL
2131 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
2132 by Yuval Yarom and Naomi Benger. Details can be obtained from:
2133 http://eprint.iacr.org/2014/140
2134
2135 Thanks to Yuval Yarom and Naomi Benger for discovering this
2136 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
2137 [Yuval Yarom and Naomi Benger]
2138
2139 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2140 this fixes a limitation in previous versions of OpenSSL.
2141 [Steve Henson]
2142
2143 *) Experimental encrypt-then-mac support.
2144
2145 Experimental support for encrypt then mac from
2146 draft-gutmann-tls-encrypt-then-mac-02.txt
2147
2148 To enable it set the appropriate extension number (0x42 for the test
2149 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
2150
2151 For non-compliant peers (i.e. just about everything) this should have no
2152 effect.
2153
2154 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
2155
2156 [Steve Henson]
2157
2158 *) Add EVP support for key wrapping algorithms, to avoid problems with
2159 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2160 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2161 algorithms and include tests cases.
2162 [Steve Henson]
2163
2164 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
2165 enveloped data.
2166 [Steve Henson]
2167
2168 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2169 MGF1 digest and OAEP label.
2170 [Steve Henson]
2171
2172 *) Make openssl verify return errors.
2173 [Chris Palmer <palmer@google.com> and Ben Laurie]
2174
2175 *) New function ASN1_TIME_diff to calculate the difference between two
2176 ASN1_TIME structures or one structure and the current time.
2177 [Steve Henson]
2178
2179 *) Update fips_test_suite to support multiple command line options. New
2180 test to induce all self test errors in sequence and check expected
2181 failures.
2182 [Steve Henson]
2183
2184 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
2185 sign or verify all in one operation.
2186 [Steve Henson]
2187
2188 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
2189 test programs and fips_test_suite. Includes functionality to parse
2190 the minimal script output of fipsalgest.pl directly.
2191 [Steve Henson]
2192
2193 *) Add authorisation parameter to FIPS_module_mode_set().
2194 [Steve Henson]
2195
2196 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
2197 [Steve Henson]
2198
2199 *) Use separate DRBG fields for internal and external flags. New function
2200 FIPS_drbg_health_check() to perform on demand health checking. Add
2201 generation tests to fips_test_suite with reduced health check interval to
2202 demonstrate periodic health checking. Add "nodh" option to
2203 fips_test_suite to skip very slow DH test.
2204 [Steve Henson]
2205
2206 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
2207 based on NID.
2208 [Steve Henson]
2209
2210 *) More extensive health check for DRBG checking many more failure modes.
2211 New function FIPS_selftest_drbg_all() to handle every possible DRBG
2212 combination: call this in fips_test_suite.
2213 [Steve Henson]
2214
2215 *) Add support for canonical generation of DSA parameter 'g'. See
2216 FIPS 186-3 A.2.3.
2217
2218 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
2219 POST to handle HMAC cases.
2220 [Steve Henson]
2221
2222 *) Add functions FIPS_module_version() and FIPS_module_version_text()
2223 to return numerical and string versions of the FIPS module number.
2224 [Steve Henson]
2225
2226 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
2227 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
2228 outside the validated module in the FIPS capable OpenSSL.
2229 [Steve Henson]
2230
2231 *) Minor change to DRBG entropy callback semantics. In some cases
2232 there is no multiple of the block length between min_len and
2233 max_len. Allow the callback to return more than max_len bytes
2234 of entropy but discard any extra: it is the callback's responsibility
2235 to ensure that the extra data discarded does not impact the
2236 requested amount of entropy.
2237 [Steve Henson]
2238
2239 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
2240 information in FIPS186-3, SP800-57 and SP800-131A.
2241 [Steve Henson]
2242
2243 *) CCM support via EVP. Interface is very similar to GCM case except we
2244 must supply all data in one chunk (i.e. no update, final) and the
2245 message length must be supplied if AAD is used. Add algorithm test
2246 support.
2247 [Steve Henson]
2248
2249 *) Initial version of POST overhaul. Add POST callback to allow the status
2250 of POST to be monitored and/or failures induced. Modify fips_test_suite
2251 to use callback. Always run all selftests even if one fails.
2252 [Steve Henson]
2253
2254 *) XTS support including algorithm test driver in the fips_gcmtest program.
2255 Note: this does increase the maximum key length from 32 to 64 bytes but
2256 there should be no binary compatibility issues as existing applications
2257 will never use XTS mode.
2258 [Steve Henson]
2259
2260 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
2261 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
2262 performs algorithm blocking for unapproved PRNG types. Also do not
2263 set PRNG type in FIPS_mode_set(): leave this to the application.
2264 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
2265 the standard OpenSSL PRNG: set additional data to a date time vector.
2266 [Steve Henson]
2267
2268 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
2269 This shouldn't present any incompatibility problems because applications
2270 shouldn't be using these directly and any that are will need to rethink
2271 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
2272 [Steve Henson]
2273
2274 *) Extensive self tests and health checking required by SP800-90 DRBG.
2275 Remove strength parameter from FIPS_drbg_instantiate and always
2276 instantiate at maximum supported strength.
2277 [Steve Henson]
2278
2279 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
2280 [Steve Henson]
2281
2282 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
2283 [Steve Henson]
2284
2285 *) New function DH_compute_key_padded() to compute a DH key and pad with
2286 leading zeroes if needed: this complies with SP800-56A et al.
2287 [Steve Henson]
2288
2289 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
2290 anything, incomplete, subject to change and largely untested at present.
2291 [Steve Henson]
2292
2293 *) Modify fipscanisteronly build option to only build the necessary object
2294 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
2295 [Steve Henson]
2296
2297 *) Add experimental option FIPSSYMS to give all symbols in
2298 fipscanister.o and FIPS or fips prefix. This will avoid
2299 conflicts with future versions of OpenSSL. Add perl script
2300 util/fipsas.pl to preprocess assembly language source files
2301 and rename any affected symbols.
2302 [Steve Henson]
2303
2304 *) Add selftest checks and algorithm block of non-fips algorithms in
2305 FIPS mode. Remove DES2 from selftests.
2306 [Steve Henson]
2307
2308 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
2309 return internal method without any ENGINE dependencies. Add new
2310 tiny fips sign and verify functions.
2311 [Steve Henson]
2312
2313 *) New build option no-ec2m to disable characteristic 2 code.
2314 [Steve Henson]
2315
2316 *) New build option "fipscanisteronly". This only builds fipscanister.o
2317 and (currently) associated fips utilities. Uses the file Makefile.fips
2318 instead of Makefile.org as the prototype.
2319 [Steve Henson]
2320
2321 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
2322 Update fips_gcmtest to use IV generator.
2323 [Steve Henson]
2324
2325 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
2326 setting output buffer to NULL. The *Final function must be
2327 called although it will not retrieve any additional data. The tag
2328 can be set or retrieved with a ctrl. The IV length is by default 12
2329 bytes (96 bits) but can be set to an alternative value. If the IV
2330 length exceeds the maximum IV length (currently 16 bytes) it cannot be
2331 set before the key.
2332 [Steve Henson]
2333
2334 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
2335 underlying do_cipher function handles all cipher semantics itself
2336 including padding and finalisation. This is useful if (for example)
2337 an ENGINE cipher handles block padding itself. The behaviour of
2338 do_cipher is subtly changed if this flag is set: the return value
2339 is the number of characters written to the output buffer (zero is
2340 no longer an error code) or a negative error code. Also if the
2341 input buffer is NULL and length 0 finalisation should be performed.
2342 [Steve Henson]
2343
2344 *) If a candidate issuer certificate is already part of the constructed
2345 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
2346 [Steve Henson]
2347
2348 *) Improve forward-security support: add functions
2349
2350 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
2351 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
2352
2353 for use by SSL/TLS servers; the callback function will be called whenever a
2354 new session is created, and gets to decide whether the session may be
2355 cached to make it resumable (return 0) or not (return 1). (As by the
2356 SSL/TLS protocol specifications, the session_id sent by the server will be
2357 empty to indicate that the session is not resumable; also, the server will
2358 not generate RFC 4507 (RFC 5077) session tickets.)
2359
2360 A simple reasonable callback implementation is to return is_forward_secure.
2361 This parameter will be set to 1 or 0 depending on the ciphersuite selected
2362 by the SSL/TLS server library, indicating whether it can provide forward
2363 security.
2364 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2365
2366 *) New -verify_name option in command line utilities to set verification
2367 parameters by name.
2368 [Steve Henson]
2369
2370 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
2371 Add CMAC pkey methods.
2372 [Steve Henson]
2373
2374 *) Experimental renegotiation in s_server -www mode. If the client
2375 browses /reneg connection is renegotiated. If /renegcert it is
2376 renegotiated requesting a certificate.
2377 [Steve Henson]
2378
2379 *) Add an "external" session cache for debugging purposes to s_server. This
2380 should help trace issues which normally are only apparent in deployed
2381 multi-process servers.
2382 [Steve Henson]
2383
2384 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
2385 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
2386 BIO_set_cipher() and some obscure PEM functions were changed so they
2387 can now return an error. The RAND changes required a change to the
2388 RAND_METHOD structure.
2389 [Steve Henson]
2390
2391 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
2392 a gcc attribute to warn if the result of a function is ignored. This
2393 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
2394 whose return value is often ignored.
2395 [Steve Henson]
2396
2397 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
2398 These allow SCTs (signed certificate timestamps) to be requested and
2399 validated when establishing a connection.
2400 [Rob Percival <robpercival@google.com>]
2401
2402 Changes between 1.0.2g and 1.0.2h [3 May 2016]
2403
2404 *) Prevent padding oracle in AES-NI CBC MAC check
2405
2406 A MITM attacker can use a padding oracle attack to decrypt traffic
2407 when the connection uses an AES CBC cipher and the server support
2408 AES-NI.
2409
2410 This issue was introduced as part of the fix for Lucky 13 padding
2411 attack (CVE-2013-0169). The padding check was rewritten to be in
2412 constant time by making sure that always the same bytes are read and
2413 compared against either the MAC or padding bytes. But it no longer
2414 checked that there was enough data to have both the MAC and padding
2415 bytes.
2416
2417 This issue was reported by Juraj Somorovsky using TLS-Attacker.
2418 (CVE-2016-2107)
2419 [Kurt Roeckx]
2420
2421 *) Fix EVP_EncodeUpdate overflow
2422
2423 An overflow can occur in the EVP_EncodeUpdate() function which is used for
2424 Base64 encoding of binary data. If an attacker is able to supply very large
2425 amounts of input data then a length check can overflow resulting in a heap
2426 corruption.
2427
2428 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
2429 the PEM_write_bio* family of functions. These are mainly used within the
2430 OpenSSL command line applications, so any application which processes data
2431 from an untrusted source and outputs it as a PEM file should be considered
2432 vulnerable to this issue. User applications that call these APIs directly
2433 with large amounts of untrusted data may also be vulnerable.
2434
2435 This issue was reported by Guido Vranken.
2436 (CVE-2016-2105)
2437 [Matt Caswell]
2438
2439 *) Fix EVP_EncryptUpdate overflow
2440
2441 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
2442 is able to supply very large amounts of input data after a previous call to
2443 EVP_EncryptUpdate() with a partial block then a length check can overflow
2444 resulting in a heap corruption. Following an analysis of all OpenSSL
2445 internal usage of the EVP_EncryptUpdate() function all usage is one of two
2446 forms. The first form is where the EVP_EncryptUpdate() call is known to be
2447 the first called function after an EVP_EncryptInit(), and therefore that
2448 specific call must be safe. The second form is where the length passed to
2449 EVP_EncryptUpdate() can be seen from the code to be some small value and
2450 therefore there is no possibility of an overflow. Since all instances are
2451 one of these two forms, it is believed that there can be no overflows in
2452 internal code due to this problem. It should be noted that
2453 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
2454 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
2455 of these calls have also been analysed too and it is believed there are no
2456 instances in internal usage where an overflow could occur.
2457
2458 This issue was reported by Guido Vranken.
2459 (CVE-2016-2106)
2460 [Matt Caswell]
2461
2462 *) Prevent ASN.1 BIO excessive memory allocation
2463
2464 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2465 a short invalid encoding can cause allocation of large amounts of memory
2466 potentially consuming excessive resources or exhausting memory.
2467
2468 Any application parsing untrusted data through d2i BIO functions is
2469 affected. The memory based functions such as d2i_X509() are *not* affected.
2470 Since the memory based functions are used by the TLS library, TLS
2471 applications are not affected.
2472
2473 This issue was reported by Brian Carpenter.
2474 (CVE-2016-2109)
2475 [Stephen Henson]
2476
2477 *) EBCDIC overread
2478
2479 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2480 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2481 in arbitrary stack data being returned in the buffer.
2482
2483 This issue was reported by Guido Vranken.
2484 (CVE-2016-2176)
2485 [Matt Caswell]
2486
2487 *) Modify behavior of ALPN to invoke callback after SNI/servername
2488 callback, such that updates to the SSL_CTX affect ALPN.
2489 [Todd Short]
2490
2491 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2492 default.
2493 [Kurt Roeckx]
2494
2495 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2496 methods are enabled and ssl2 is disabled the methods return NULL.
2497 [Kurt Roeckx]
2498
2499 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2500
2501 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2502 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2503 provide any "EXPORT" or "LOW" strength ciphers.
2504 [Viktor Dukhovni]
2505
2506 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2507 is by default disabled at build-time. Builds that are not configured with
2508 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2509 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2510 will need to explicitly call either of:
2511
2512 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2513 or
2514 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2515
2516 as appropriate. Even if either of those is used, or the application
2517 explicitly uses the version-specific SSLv2_method() or its client and
2518 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2519 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2520 ciphers, and SSLv2 56-bit DES are no longer available.
2521 (CVE-2016-0800)
2522 [Viktor Dukhovni]
2523
2524 *) Fix a double-free in DSA code
2525
2526 A double free bug was discovered when OpenSSL parses malformed DSA private
2527 keys and could lead to a DoS attack or memory corruption for applications
2528 that receive DSA private keys from untrusted sources. This scenario is
2529 considered rare.
2530
2531 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2532 libFuzzer.
2533 (CVE-2016-0705)
2534 [Stephen Henson]
2535
2536 *) Disable SRP fake user seed to address a server memory leak.
2537
2538 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2539
2540 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2541 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2542 was changed to ignore the "fake user" SRP seed, even if the seed
2543 is configured.
2544
2545 Users should use SRP_VBASE_get1_by_user instead. Note that in
2546 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2547 also that even though configuring the SRP seed attempts to hide
2548 invalid usernames by continuing the handshake with fake
2549 credentials, this behaviour is not constant time and no strong
2550 guarantees are made that the handshake is indistinguishable from
2551 that of a valid user.
2552 (CVE-2016-0798)
2553 [Emilia Käsper]
2554
2555 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2556
2557 In the BN_hex2bn function the number of hex digits is calculated using an
2558 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2559 large values of |i| this can result in |bn_expand| not allocating any
2560 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2561 field as NULL leading to a subsequent NULL ptr deref. For very large values
2562 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2563 In this case memory is allocated to the internal BIGNUM data field, but it
2564 is insufficiently sized leading to heap corruption. A similar issue exists
2565 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2566 is ever called by user applications with very large untrusted hex/dec data.
2567 This is anticipated to be a rare occurrence.
2568
2569 All OpenSSL internal usage of these functions use data that is not expected
2570 to be untrusted, e.g. config file data or application command line
2571 arguments. If user developed applications generate config file data based
2572 on untrusted data then it is possible that this could also lead to security
2573 consequences. This is also anticipated to be rare.
2574
2575 This issue was reported to OpenSSL by Guido Vranken.
2576 (CVE-2016-0797)
2577 [Matt Caswell]
2578
2579 *) Fix memory issues in BIO_*printf functions
2580
2581 The internal |fmtstr| function used in processing a "%s" format string in
2582 the BIO_*printf functions could overflow while calculating the length of a
2583 string and cause an OOB read when printing very long strings.
2584
2585 Additionally the internal |doapr_outch| function can attempt to write to an
2586 OOB memory location (at an offset from the NULL pointer) in the event of a
2587 memory allocation failure. In 1.0.2 and below this could be caused where
2588 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2589 could be in processing a very long "%s" format string. Memory leaks can
2590 also occur.
2591
2592 The first issue may mask the second issue dependent on compiler behaviour.
2593 These problems could enable attacks where large amounts of untrusted data
2594 is passed to the BIO_*printf functions. If applications use these functions
2595 in this way then they could be vulnerable. OpenSSL itself uses these
2596 functions when printing out human-readable dumps of ASN.1 data. Therefore
2597 applications that print this data could be vulnerable if the data is from
2598 untrusted sources. OpenSSL command line applications could also be
2599 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2600 as command line arguments.
2601
2602 Libssl is not considered directly vulnerable. Additionally certificates etc
2603 received via remote connections via libssl are also unlikely to be able to
2604 trigger these issues because of message size limits enforced within libssl.
2605
2606 This issue was reported to OpenSSL Guido Vranken.
2607 (CVE-2016-0799)
2608 [Matt Caswell]
2609
2610 *) Side channel attack on modular exponentiation
2611
2612 A side-channel attack was found which makes use of cache-bank conflicts on
2613 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2614 of RSA keys. The ability to exploit this issue is limited as it relies on
2615 an attacker who has control of code in a thread running on the same
2616 hyper-threaded core as the victim thread which is performing decryptions.
2617
2618 This issue was reported to OpenSSL by Yuval Yarom, The University of
2619 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2620 Nadia Heninger, University of Pennsylvania with more information at
2621 http://cachebleed.info.
2622 (CVE-2016-0702)
2623 [Andy Polyakov]
2624
2625 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2626 if no keysize is specified with default_bits. This fixes an
2627 omission in an earlier change that changed all RSA/DSA key generation
2628 apps to use 2048 bits by default.
2629 [Emilia Käsper]
2630
2631 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2632 *) DH small subgroups
2633
2634 Historically OpenSSL only ever generated DH parameters based on "safe"
2635 primes. More recently (in version 1.0.2) support was provided for
2636 generating X9.42 style parameter files such as those required for RFC 5114
2637 support. The primes used in such files may not be "safe". Where an
2638 application is using DH configured with parameters based on primes that are
2639 not "safe" then an attacker could use this fact to find a peer's private
2640 DH exponent. This attack requires that the attacker complete multiple
2641 handshakes in which the peer uses the same private DH exponent. For example
2642 this could be used to discover a TLS server's private DH exponent if it's
2643 reusing the private DH exponent or it's using a static DH ciphersuite.
2644
2645 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2646 TLS. It is not on by default. If the option is not set then the server
2647 reuses the same private DH exponent for the life of the server process and
2648 would be vulnerable to this attack. It is believed that many popular
2649 applications do set this option and would therefore not be at risk.
2650
2651 The fix for this issue adds an additional check where a "q" parameter is
2652 available (as is the case in X9.42 based parameters). This detects the
2653 only known attack, and is the only possible defense for static DH
2654 ciphersuites. This could have some performance impact.
2655
2656 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2657 default and cannot be disabled. This could have some performance impact.
2658
2659 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2660 (CVE-2016-0701)
2661 [Matt Caswell]
2662
2663 *) SSLv2 doesn't block disabled ciphers
2664
2665 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2666 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2667 been disabled, provided that the SSLv2 protocol was not also disabled via
2668 SSL_OP_NO_SSLv2.
2669
2670 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2671 and Sebastian Schinzel.
2672 (CVE-2015-3197)
2673 [Viktor Dukhovni]
2674
2675 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2676
2677 *) BN_mod_exp may produce incorrect results on x86_64
2678
2679 There is a carry propagating bug in the x86_64 Montgomery squaring
2680 procedure. No EC algorithms are affected. Analysis suggests that attacks
2681 against RSA and DSA as a result of this defect would be very difficult to
2682 perform and are not believed likely. Attacks against DH are considered just
2683 feasible (although very difficult) because most of the work necessary to
2684 deduce information about a private key may be performed offline. The amount
2685 of resources required for such an attack would be very significant and
2686 likely only accessible to a limited number of attackers. An attacker would
2687 additionally need online access to an unpatched system using the target
2688 private key in a scenario with persistent DH parameters and a private
2689 key that is shared between multiple clients. For example this can occur by
2690 default in OpenSSL DHE based SSL/TLS ciphersuites.
2691
2692 This issue was reported to OpenSSL by Hanno Böck.
2693 (CVE-2015-3193)
2694 [Andy Polyakov]
2695
2696 *) Certificate verify crash with missing PSS parameter
2697
2698 The signature verification routines will crash with a NULL pointer
2699 dereference if presented with an ASN.1 signature using the RSA PSS
2700 algorithm and absent mask generation function parameter. Since these
2701 routines are used to verify certificate signature algorithms this can be
2702 used to crash any certificate verification operation and exploited in a
2703 DoS attack. Any application which performs certificate verification is
2704 vulnerable including OpenSSL clients and servers which enable client
2705 authentication.
2706
2707 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2708 (CVE-2015-3194)
2709 [Stephen Henson]
2710
2711 *) X509_ATTRIBUTE memory leak
2712
2713 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2714 memory. This structure is used by the PKCS#7 and CMS routines so any
2715 application which reads PKCS#7 or CMS data from untrusted sources is
2716 affected. SSL/TLS is not affected.
2717
2718 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2719 libFuzzer.
2720 (CVE-2015-3195)
2721 [Stephen Henson]
2722
2723 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2724 This changes the decoding behaviour for some invalid messages,
2725 though the change is mostly in the more lenient direction, and
2726 legacy behaviour is preserved as much as possible.
2727 [Emilia Käsper]
2728
2729 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2730 return an error
2731 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2732
2733 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2734
2735 *) Alternate chains certificate forgery
2736
2737 During certificate verification, OpenSSL will attempt to find an
2738 alternative certificate chain if the first attempt to build such a chain
2739 fails. An error in the implementation of this logic can mean that an
2740 attacker could cause certain checks on untrusted certificates to be
2741 bypassed, such as the CA flag, enabling them to use a valid leaf
2742 certificate to act as a CA and "issue" an invalid certificate.
2743
2744 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2745 (Google/BoringSSL).
2746 [Matt Caswell]
2747
2748 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2749
2750 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2751 incompatibility in the handling of HMAC. The previous ABI has now been
2752 restored.
2753 [Matt Caswell]
2754
2755 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2756
2757 *) Malformed ECParameters causes infinite loop
2758
2759 When processing an ECParameters structure OpenSSL enters an infinite loop
2760 if the curve specified is over a specially malformed binary polynomial
2761 field.
2762
2763 This can be used to perform denial of service against any
2764 system which processes public keys, certificate requests or
2765 certificates. This includes TLS clients and TLS servers with
2766 client authentication enabled.
2767
2768 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2769 (CVE-2015-1788)
2770 [Andy Polyakov]
2771
2772 *) Exploitable out-of-bounds read in X509_cmp_time
2773
2774 X509_cmp_time does not properly check the length of the ASN1_TIME
2775 string and can read a few bytes out of bounds. In addition,
2776 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2777 time string.
2778
2779 An attacker can use this to craft malformed certificates and CRLs of
2780 various sizes and potentially cause a segmentation fault, resulting in
2781 a DoS on applications that verify certificates or CRLs. TLS clients
2782 that verify CRLs are affected. TLS clients and servers with client
2783 authentication enabled may be affected if they use custom verification
2784 callbacks.
2785
2786 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2787 independently by Hanno Böck.
2788 (CVE-2015-1789)
2789 [Emilia Käsper]
2790
2791 *) PKCS7 crash with missing EnvelopedContent
2792
2793 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2794 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2795 with missing content and trigger a NULL pointer dereference on parsing.
2796
2797 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2798 structures from untrusted sources are affected. OpenSSL clients and
2799 servers are not affected.
2800
2801 This issue was reported to OpenSSL by Michal Zalewski (Google).
2802 (CVE-2015-1790)
2803 [Emilia Käsper]
2804
2805 *) CMS verify infinite loop with unknown hash function
2806
2807 When verifying a signedData message the CMS code can enter an infinite loop
2808 if presented with an unknown hash function OID. This can be used to perform
2809 denial of service against any system which verifies signedData messages using
2810 the CMS code.
2811 This issue was reported to OpenSSL by Johannes Bauer.
2812 (CVE-2015-1792)
2813 [Stephen Henson]
2814
2815 *) Race condition handling NewSessionTicket
2816
2817 If a NewSessionTicket is received by a multi-threaded client when attempting to
2818 reuse a previous ticket then a race condition can occur potentially leading to
2819 a double free of the ticket data.
2820 (CVE-2015-1791)
2821 [Matt Caswell]
2822
2823 *) Only support 256-bit or stronger elliptic curves with the
2824 'ecdh_auto' setting (server) or by default (client). Of supported
2825 curves, prefer P-256 (both).
2826 [Emilia Kasper]
2827
2828 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2829
2830 *) ClientHello sigalgs DoS fix
2831
2832 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2833 invalid signature algorithms extension a NULL pointer dereference will
2834 occur. This can be exploited in a DoS attack against the server.
2835
2836 This issue was was reported to OpenSSL by David Ramos of Stanford
2837 University.
2838 (CVE-2015-0291)
2839 [Stephen Henson and Matt Caswell]
2840
2841 *) Multiblock corrupted pointer fix
2842
2843 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2844 feature only applies on 64 bit x86 architecture platforms that support AES
2845 NI instructions. A defect in the implementation of "multiblock" can cause
2846 OpenSSL's internal write buffer to become incorrectly set to NULL when
2847 using non-blocking IO. Typically, when the user application is using a
2848 socket BIO for writing, this will only result in a failed connection.
2849 However if some other BIO is used then it is likely that a segmentation
2850 fault will be triggered, thus enabling a potential DoS attack.
2851
2852 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2853 (CVE-2015-0290)
2854 [Matt Caswell]
2855
2856 *) Segmentation fault in DTLSv1_listen fix
2857
2858 The DTLSv1_listen function is intended to be stateless and processes the
2859 initial ClientHello from many peers. It is common for user code to loop
2860 over the call to DTLSv1_listen until a valid ClientHello is received with
2861 an associated cookie. A defect in the implementation of DTLSv1_listen means
2862 that state is preserved in the SSL object from one invocation to the next
2863 that can lead to a segmentation fault. Errors processing the initial
2864 ClientHello can trigger this scenario. An example of such an error could be
2865 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2866 server.
2867
2868 This issue was reported to OpenSSL by Per Allansson.
2869 (CVE-2015-0207)
2870 [Matt Caswell]
2871
2872 *) Segmentation fault in ASN1_TYPE_cmp fix
2873
2874 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2875 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2876 certificate signature algorithm consistency this can be used to crash any
2877 certificate verification operation and exploited in a DoS attack. Any
2878 application which performs certificate verification is vulnerable including
2879 OpenSSL clients and servers which enable client authentication.
2880 (CVE-2015-0286)
2881 [Stephen Henson]
2882
2883 *) Segmentation fault for invalid PSS parameters fix
2884
2885 The signature verification routines will crash with a NULL pointer
2886 dereference if presented with an ASN.1 signature using the RSA PSS
2887 algorithm and invalid parameters. Since these routines are used to verify
2888 certificate signature algorithms this can be used to crash any
2889 certificate verification operation and exploited in a DoS attack. Any
2890 application which performs certificate verification is vulnerable including
2891 OpenSSL clients and servers which enable client authentication.
2892
2893 This issue was was reported to OpenSSL by Brian Carpenter.
2894 (CVE-2015-0208)
2895 [Stephen Henson]
2896
2897 *) ASN.1 structure reuse memory corruption fix
2898
2899 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2900 memory corruption via an invalid write. Such reuse is and has been
2901 strongly discouraged and is believed to be rare.
2902
2903 Applications that parse structures containing CHOICE or ANY DEFINED BY
2904 components may be affected. Certificate parsing (d2i_X509 and related
2905 functions) are however not affected. OpenSSL clients and servers are
2906 not affected.
2907 (CVE-2015-0287)
2908 [Stephen Henson]
2909
2910 *) PKCS7 NULL pointer dereferences fix
2911
2912 The PKCS#7 parsing code does not handle missing outer ContentInfo
2913 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2914 missing content and trigger a NULL pointer dereference on parsing.
2915
2916 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2917 otherwise parse PKCS#7 structures from untrusted sources are
2918 affected. OpenSSL clients and servers are not affected.
2919
2920 This issue was reported to OpenSSL by Michal Zalewski (Google).
2921 (CVE-2015-0289)
2922 [Emilia Käsper]
2923
2924 *) DoS via reachable assert in SSLv2 servers fix
2925
2926 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2927 servers that both support SSLv2 and enable export cipher suites by sending
2928 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2929
2930 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2931 (OpenSSL development team).
2932 (CVE-2015-0293)
2933 [Emilia Käsper]
2934
2935 *) Empty CKE with client auth and DHE fix
2936
2937 If client auth is used then a server can seg fault in the event of a DHE
2938 ciphersuite being selected and a zero length ClientKeyExchange message
2939 being sent by the client. This could be exploited in a DoS attack.
2940 (CVE-2015-1787)
2941 [Matt Caswell]
2942
2943 *) Handshake with unseeded PRNG fix
2944
2945 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2946 with an unseeded PRNG. The conditions are:
2947 - The client is on a platform where the PRNG has not been seeded
2948 automatically, and the user has not seeded manually
2949 - A protocol specific client method version has been used (i.e. not
2950 SSL_client_methodv23)
2951 - A ciphersuite is used that does not require additional random data from
2952 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2953
2954 If the handshake succeeds then the client random that has been used will
2955 have been generated from a PRNG with insufficient entropy and therefore the
2956 output may be predictable.
2957
2958 For example using the following command with an unseeded openssl will
2959 succeed on an unpatched platform:
2960
2961 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2962 (CVE-2015-0285)
2963 [Matt Caswell]
2964
2965 *) Use After Free following d2i_ECPrivatekey error fix
2966
2967 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2968 could cause a use after free condition. This, in turn, could cause a double
2969 free in several private key parsing functions (such as d2i_PrivateKey
2970 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2971 for applications that receive EC private keys from untrusted
2972 sources. This scenario is considered rare.
2973
2974 This issue was discovered by the BoringSSL project and fixed in their
2975 commit 517073cd4b.
2976 (CVE-2015-0209)
2977 [Matt Caswell]
2978
2979 *) X509_to_X509_REQ NULL pointer deref fix
2980
2981 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2982 the certificate key is invalid. This function is rarely used in practice.
2983
2984 This issue was discovered by Brian Carpenter.
2985 (CVE-2015-0288)
2986 [Stephen Henson]
2987
2988 *) Removed the export ciphers from the DEFAULT ciphers
2989 [Kurt Roeckx]
2990
2991 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2992
2993 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2994 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2995 So far those who have to target multiple platforms would compromise
2996 and argue that binary targeting say ARMv5 would still execute on
2997 ARMv8. "Universal" build resolves this compromise by providing
2998 near-optimal performance even on newer platforms.
2999 [Andy Polyakov]
3000
3001 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
3002 (other platforms pending).
3003 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
3004
3005 *) Add support for the SignedCertificateTimestampList certificate and
3006 OCSP response extensions from RFC6962.
3007 [Rob Stradling]
3008
3009 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3010 for corner cases. (Certain input points at infinity could lead to
3011 bogus results, with non-infinity inputs mapped to infinity too.)
3012 [Bodo Moeller]
3013
3014 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
3015 This covers AES, SHA256/512 and GHASH. "Initial" means that most
3016 common cases are optimized and there still is room for further
3017 improvements. Vector Permutation AES for Altivec is also added.
3018 [Andy Polyakov]
3019
3020 *) Add support for little-endian ppc64 Linux target.
3021 [Marcelo Cerri (IBM)]
3022
3023 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
3024 SHA1, SHA256 and GHASH. "Initial" means that most common cases
3025 are optimized and there still is room for further improvements.
3026 Both 32- and 64-bit modes are supported.
3027 [Andy Polyakov, Ard Biesheuvel (Linaro)]
3028
3029 *) Improved ARMv7 NEON support.
3030 [Andy Polyakov]
3031
3032 *) Support for SPARC Architecture 2011 crypto extensions, first
3033 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
3034 SHA256/512, MD5, GHASH and modular exponentiation.
3035 [Andy Polyakov, David Miller]
3036
3037 *) Accelerated modular exponentiation for Intel processors, a.k.a.
3038 RSAZ.
3039 [Shay Gueron & Vlad Krasnov (Intel Corp)]
3040
3041 *) Support for new and upcoming Intel processors, including AVX2,
3042 BMI and SHA ISA extensions. This includes additional "stitched"
3043 implementations, AESNI-SHA256 and GCM, and multi-buffer support
3044 for TLS encrypt.
3045
3046 This work was sponsored by Intel Corp.
3047 [Andy Polyakov]
3048
3049 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
3050 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
3051 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
3052 [Steve Henson]
3053
3054 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
3055 this fixes a limitation in previous versions of OpenSSL.
3056 [Steve Henson]
3057
3058 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
3059 MGF1 digest and OAEP label.
3060 [Steve Henson]
3061
3062 *) Add EVP support for key wrapping algorithms, to avoid problems with
3063 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
3064 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
3065 algorithms and include tests cases.
3066 [Steve Henson]
3067
3068 *) Add functions to allocate and set the fields of an ECDSA_METHOD
3069 structure.
3070 [Douglas E. Engert, Steve Henson]
3071
3072 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
3073 difference in days and seconds between two tm or ASN1_TIME structures.
3074 [Steve Henson]
3075
3076 *) Add -rev test option to s_server to just reverse order of characters
3077 received by client and send back to server. Also prints an abbreviated
3078 summary of the connection parameters.
3079 [Steve Henson]
3080
3081 *) New option -brief for s_client and s_server to print out a brief summary
3082 of connection parameters.
3083 [Steve Henson]
3084
3085 *) Add callbacks for arbitrary TLS extensions.
3086 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
3087
3088 *) New option -crl_download in several openssl utilities to download CRLs
3089 from CRLDP extension in certificates.
3090 [Steve Henson]
3091
3092 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
3093 [Steve Henson]
3094
3095 *) New function X509_CRL_diff to generate a delta CRL from the difference
3096 of two full CRLs. Add support to "crl" utility.
3097 [Steve Henson]
3098
3099 *) New functions to set lookup_crls function and to retrieve
3100 X509_STORE from X509_STORE_CTX.
3101 [Steve Henson]
3102
3103 *) Print out deprecated issuer and subject unique ID fields in
3104 certificates.
3105 [Steve Henson]
3106
3107 *) Extend OCSP I/O functions so they can be used for simple general purpose
3108 HTTP as well as OCSP. New wrapper function which can be used to download
3109 CRLs using the OCSP API.
3110 [Steve Henson]
3111
3112 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
3113 [Steve Henson]
3114
3115 *) SSL_CONF* functions. These provide a common framework for application
3116 configuration using configuration files or command lines.
3117 [Steve Henson]
3118
3119 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
3120 message callback and prints the results. Needs compile time option
3121 "enable-ssl-trace". New options to s_client and s_server to enable
3122 tracing.
3123 [Steve Henson]
3124
3125 *) New ctrl and macro to retrieve supported points extensions.
3126 Print out extension in s_server and s_client.
3127 [Steve Henson]
3128
3129 *) New functions to retrieve certificate signature and signature
3130 OID NID.
3131 [Steve Henson]
3132
3133 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
3134 client to OpenSSL.
3135 [Steve Henson]
3136
3137 *) New Suite B modes for TLS code. These use and enforce the requirements
3138 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
3139 only use Suite B curves. The Suite B modes can be set by using the
3140 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
3141 [Steve Henson]
3142
3143 *) New chain verification flags for Suite B levels of security. Check
3144 algorithms are acceptable when flags are set in X509_verify_cert.
3145 [Steve Henson]
3146
3147 *) Make tls1_check_chain return a set of flags indicating checks passed
3148 by a certificate chain. Add additional tests to handle client
3149 certificates: checks for matching certificate type and issuer name
3150 comparison.
3151 [Steve Henson]
3152
3153 *) If an attempt is made to use a signature algorithm not in the peer
3154 preference list abort the handshake. If client has no suitable
3155 signature algorithms in response to a certificate request do not
3156 use the certificate.
3157 [Steve Henson]
3158
3159 *) If server EC tmp key is not in client preference list abort handshake.
3160 [Steve Henson]
3161
3162 *) Add support for certificate stores in CERT structure. This makes it
3163 possible to have different stores per SSL structure or one store in
3164 the parent SSL_CTX. Include distinct stores for certificate chain
3165 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
3166 to build and store a certificate chain in CERT structure: returning
3167 an error if the chain cannot be built: this will allow applications
3168 to test if a chain is correctly configured.
3169
3170 Note: if the CERT based stores are not set then the parent SSL_CTX
3171 store is used to retain compatibility with existing behaviour.
3172
3173 [Steve Henson]
3174
3175 *) New function ssl_set_client_disabled to set a ciphersuite disabled
3176 mask based on the current session, check mask when sending client
3177 hello and checking the requested ciphersuite.
3178 [Steve Henson]
3179
3180 *) New ctrls to retrieve and set certificate types in a certificate
3181 request message. Print out received values in s_client. If certificate
3182 types is not set with custom values set sensible values based on
3183 supported signature algorithms.
3184 [Steve Henson]
3185
3186 *) Support for distinct client and server supported signature algorithms.
3187 [Steve Henson]
3188
3189 *) Add certificate callback. If set this is called whenever a certificate
3190 is required by client or server. An application can decide which
3191 certificate chain to present based on arbitrary criteria: for example
3192 supported signature algorithms. Add very simple example to s_server.
3193 This fixes many of the problems and restrictions of the existing client
3194 certificate callback: for example you can now clear an existing
3195 certificate and specify the whole chain.
3196 [Steve Henson]
3197
3198 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
3199 the certificate can be used for (if anything). Set valid_flags field
3200 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
3201 to have similar checks in it.
3202
3203 Add new "cert_flags" field to CERT structure and include a "strict mode".
3204 This enforces some TLS certificate requirements (such as only permitting
3205 certificate signature algorithms contained in the supported algorithms
3206 extension) which some implementations ignore: this option should be used
3207 with caution as it could cause interoperability issues.
3208 [Steve Henson]
3209
3210 *) Update and tidy signature algorithm extension processing. Work out
3211 shared signature algorithms based on preferences and peer algorithms
3212 and print them out in s_client and s_server. Abort handshake if no
3213 shared signature algorithms.
3214 [Steve Henson]
3215
3216 *) Add new functions to allow customised supported signature algorithms
3217 for SSL and SSL_CTX structures. Add options to s_client and s_server
3218 to support them.
3219 [Steve Henson]
3220
3221 *) New function SSL_certs_clear() to delete all references to certificates
3222 from an SSL structure. Before this once a certificate had been added
3223 it couldn't be removed.
3224 [Steve Henson]
3225
3226 *) Integrate hostname, email address and IP address checking with certificate
3227 verification. New verify options supporting checking in openssl utility.
3228 [Steve Henson]
3229
3230 *) Fixes and wildcard matching support to hostname and email checking
3231 functions. Add manual page.
3232 [Florian Weimer (Red Hat Product Security Team)]
3233
3234 *) New functions to check a hostname email or IP address against a
3235 certificate. Add options x509 utility to print results of checks against
3236 a certificate.
3237 [Steve Henson]
3238
3239 *) Fix OCSP checking.
3240 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
3241
3242 *) Initial experimental support for explicitly trusted non-root CAs.
3243 OpenSSL still tries to build a complete chain to a root but if an
3244 intermediate CA has a trust setting included that is used. The first
3245 setting is used: whether to trust (e.g., -addtrust option to the x509
3246 utility) or reject.
3247 [Steve Henson]
3248
3249 *) Add -trusted_first option which attempts to find certificates in the
3250 trusted store even if an untrusted chain is also supplied.
3251 [Steve Henson]
3252
3253 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
3254 platform support for Linux and Android.
3255 [Andy Polyakov]
3256
3257 *) Support for linux-x32, ILP32 environment in x86_64 framework.
3258 [Andy Polyakov]
3259
3260 *) Experimental multi-implementation support for FIPS capable OpenSSL.
3261 When in FIPS mode the approved implementations are used as normal,
3262 when not in FIPS mode the internal unapproved versions are used instead.
3263 This means that the FIPS capable OpenSSL isn't forced to use the
3264 (often lower performance) FIPS implementations outside FIPS mode.
3265 [Steve Henson]
3266
3267 *) Transparently support X9.42 DH parameters when calling
3268 PEM_read_bio_DHparameters. This means existing applications can handle
3269 the new parameter format automatically.
3270 [Steve Henson]
3271
3272 *) Initial experimental support for X9.42 DH parameter format: mainly
3273 to support use of 'q' parameter for RFC5114 parameters.
3274 [Steve Henson]
3275
3276 *) Add DH parameters from RFC5114 including test data to dhtest.
3277 [Steve Henson]
3278
3279 *) Support for automatic EC temporary key parameter selection. If enabled
3280 the most preferred EC parameters are automatically used instead of
3281 hardcoded fixed parameters. Now a server just has to call:
3282 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
3283 support ECDH and use the most appropriate parameters.
3284 [Steve Henson]
3285
3286 *) Enhance and tidy EC curve and point format TLS extension code. Use
3287 static structures instead of allocation if default values are used.
3288 New ctrls to set curves we wish to support and to retrieve shared curves.
3289 Print out shared curves in s_server. New options to s_server and s_client
3290 to set list of supported curves.
3291 [Steve Henson]
3292
3293 *) New ctrls to retrieve supported signature algorithms and
3294 supported curve values as an array of NIDs. Extend openssl utility
3295 to print out received values.
3296 [Steve Henson]
3297
3298 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
3299 between NIDs and the more common NIST names such as "P-256". Enhance
3300 ecparam utility and ECC method to recognise the NIST names for curves.
3301 [Steve Henson]
3302
3303 *) Enhance SSL/TLS certificate chain handling to support different
3304 chains for each certificate instead of one chain in the parent SSL_CTX.
3305 [Steve Henson]
3306
3307 *) Support for fixed DH ciphersuite client authentication: where both
3308 server and client use DH certificates with common parameters.
3309 [Steve Henson]
3310
3311 *) Support for fixed DH ciphersuites: those requiring DH server
3312 certificates.
3313 [Steve Henson]
3314
3315 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
3316 the certificate.
3317 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
3318 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
3319 X509_CINF_get_signature were reverted post internal team review.
3320
3321 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
3322
3323 *) Build fixes for the Windows and OpenVMS platforms
3324 [Matt Caswell and Richard Levitte]
3325
3326 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
3327
3328 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
3329 message can cause a segmentation fault in OpenSSL due to a NULL pointer
3330 dereference. This could lead to a Denial Of Service attack. Thanks to
3331 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
3332 (CVE-2014-3571)
3333 [Steve Henson]
3334
3335 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
3336 dtls1_buffer_record function under certain conditions. In particular this
3337 could occur if an attacker sent repeated DTLS records with the same
3338 sequence number but for the next epoch. The memory leak could be exploited
3339 by an attacker in a Denial of Service attack through memory exhaustion.
3340 Thanks to Chris Mueller for reporting this issue.
3341 (CVE-2015-0206)
3342 [Matt Caswell]
3343
3344 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
3345 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
3346 method would be set to NULL which could later result in a NULL pointer
3347 dereference. Thanks to Frank Schmirler for reporting this issue.
3348 (CVE-2014-3569)
3349 [Kurt Roeckx]
3350
3351 *) Abort handshake if server key exchange message is omitted for ephemeral
3352 ECDH ciphersuites.
3353
3354 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
3355 reporting this issue.
3356 (CVE-2014-3572)
3357 [Steve Henson]
3358
3359 *) Remove non-export ephemeral RSA code on client and server. This code
3360 violated the TLS standard by allowing the use of temporary RSA keys in
3361 non-export ciphersuites and could be used by a server to effectively
3362 downgrade the RSA key length used to a value smaller than the server
3363 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
3364 INRIA or reporting this issue.
3365 (CVE-2015-0204)
3366 [Steve Henson]
3367
3368 *) Fixed issue where DH client certificates are accepted without verification.
3369 An OpenSSL server will accept a DH certificate for client authentication
3370 without the certificate verify message. This effectively allows a client to
3371 authenticate without the use of a private key. This only affects servers
3372 which trust a client certificate authority which issues certificates
3373 containing DH keys: these are extremely rare and hardly ever encountered.
3374 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
3375 this issue.
3376 (CVE-2015-0205)
3377 [Steve Henson]
3378
3379 *) Ensure that the session ID context of an SSL is updated when its
3380 SSL_CTX is updated via SSL_set_SSL_CTX.
3381
3382 The session ID context is typically set from the parent SSL_CTX,
3383 and can vary with the CTX.
3384 [Adam Langley]
3385
3386 *) Fix various certificate fingerprint issues.
3387
3388 By using non-DER or invalid encodings outside the signed portion of a
3389 certificate the fingerprint can be changed without breaking the signature.
3390 Although no details of the signed portion of the certificate can be changed
3391 this can cause problems with some applications: e.g. those using the
3392 certificate fingerprint for blacklists.
3393
3394 1. Reject signatures with non zero unused bits.
3395
3396 If the BIT STRING containing the signature has non zero unused bits reject
3397 the signature. All current signature algorithms require zero unused bits.
3398
3399 2. Check certificate algorithm consistency.
3400
3401 Check the AlgorithmIdentifier inside TBS matches the one in the
3402 certificate signature. NB: this will result in signature failure
3403 errors for some broken certificates.
3404
3405 Thanks to Konrad Kraszewski from Google for reporting this issue.
3406
3407 3. Check DSA/ECDSA signatures use DER.
3408
3409 Re-encode DSA/ECDSA signatures and compare with the original received
3410 signature. Return an error if there is a mismatch.
3411
3412 This will reject various cases including garbage after signature
3413 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
3414 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
3415 (negative or with leading zeroes).
3416
3417 Further analysis was conducted and fixes were developed by Stephen Henson
3418 of the OpenSSL core team.
3419
3420 (CVE-2014-8275)
3421 [Steve Henson]
3422
3423 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
3424 results on some platforms, including x86_64. This bug occurs at random
3425 with a very low probability, and is not known to be exploitable in any
3426 way, though its exact impact is difficult to determine. Thanks to Pieter
3427 Wuille (Blockstream) who reported this issue and also suggested an initial
3428 fix. Further analysis was conducted by the OpenSSL development team and
3429 Adam Langley of Google. The final fix was developed by Andy Polyakov of
3430 the OpenSSL core team.
3431 (CVE-2014-3570)
3432 [Andy Polyakov]
3433
3434 *) Do not resume sessions on the server if the negotiated protocol
3435 version does not match the session's version. Resuming with a different
3436 version, while not strictly forbidden by the RFC, is of questionable
3437 sanity and breaks all known clients.
3438 [David Benjamin, Emilia Käsper]
3439
3440 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
3441 early CCS messages during renegotiation. (Note that because
3442 renegotiation is encrypted, this early CCS was not exploitable.)
3443 [Emilia Käsper]
3444
3445 *) Tighten client-side session ticket handling during renegotiation:
3446 ensure that the client only accepts a session ticket if the server sends
3447 the extension anew in the ServerHello. Previously, a TLS client would
3448 reuse the old extension state and thus accept a session ticket if one was
3449 announced in the initial ServerHello.
3450
3451 Similarly, ensure that the client requires a session ticket if one
3452 was advertised in the ServerHello. Previously, a TLS client would
3453 ignore a missing NewSessionTicket message.
3454 [Emilia Käsper]
3455
3456 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
3457
3458 *) SRTP Memory Leak.
3459
3460 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
3461 sends a carefully crafted handshake message, to cause OpenSSL to fail
3462 to free up to 64k of memory causing a memory leak. This could be
3463 exploited in a Denial Of Service attack. This issue affects OpenSSL
3464 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
3465 whether SRTP is used or configured. Implementations of OpenSSL that
3466 have been compiled with OPENSSL_NO_SRTP defined are not affected.
3467
3468 The fix was developed by the OpenSSL team.
3469 (CVE-2014-3513)
3470 [OpenSSL team]
3471
3472 *) Session Ticket Memory Leak.
3473
3474 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
3475 integrity of that ticket is first verified. In the event of a session
3476 ticket integrity check failing, OpenSSL will fail to free memory
3477 causing a memory leak. By sending a large number of invalid session
3478 tickets an attacker could exploit this issue in a Denial Of Service
3479 attack.
3480 (CVE-2014-3567)
3481 [Steve Henson]
3482
3483 *) Build option no-ssl3 is incomplete.
3484
3485 When OpenSSL is configured with "no-ssl3" as a build option, servers
3486 could accept and complete a SSL 3.0 handshake, and clients could be
3487 configured to send them.
3488 (CVE-2014-3568)
3489 [Akamai and the OpenSSL team]
3490
3491 *) Add support for TLS_FALLBACK_SCSV.
3492 Client applications doing fallback retries should call
3493 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3494 (CVE-2014-3566)
3495 [Adam Langley, Bodo Moeller]
3496
3497 *) Add additional DigestInfo checks.
3498
3499 Re-encode DigestInto in DER and check against the original when
3500 verifying RSA signature: this will reject any improperly encoded
3501 DigestInfo structures.
3502
3503 Note: this is a precautionary measure and no attacks are currently known.
3504
3505 [Steve Henson]
3506
3507 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3508
3509 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3510 SRP code can be overrun an internal buffer. Add sanity check that
3511 g, A, B < N to SRP code.
3512
3513 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3514 Group for discovering this issue.
3515 (CVE-2014-3512)
3516 [Steve Henson]
3517
3518 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3519 TLS 1.0 instead of higher protocol versions when the ClientHello message
3520 is badly fragmented. This allows a man-in-the-middle attacker to force a
3521 downgrade to TLS 1.0 even if both the server and the client support a
3522 higher protocol version, by modifying the client's TLS records.
3523
3524 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3525 researching this issue.
3526 (CVE-2014-3511)
3527 [David Benjamin]
3528
3529 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3530 to a denial of service attack. A malicious server can crash the client
3531 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3532 ciphersuite and sending carefully crafted handshake messages.
3533
3534 Thanks to Felix Gröbert (Google) for discovering and researching this
3535 issue.
3536 (CVE-2014-3510)
3537 [Emilia Käsper]
3538
3539 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3540 to leak memory. This can be exploited through a Denial of Service attack.
3541 Thanks to Adam Langley for discovering and researching this issue.
3542 (CVE-2014-3507)
3543 [Adam Langley]
3544
3545 *) An attacker can force openssl to consume large amounts of memory whilst
3546 processing DTLS handshake messages. This can be exploited through a
3547 Denial of Service attack.
3548 Thanks to Adam Langley for discovering and researching this issue.
3549 (CVE-2014-3506)
3550 [Adam Langley]
3551
3552 *) An attacker can force an error condition which causes openssl to crash
3553 whilst processing DTLS packets due to memory being freed twice. This
3554 can be exploited through a Denial of Service attack.
3555 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3556 this issue.
3557 (CVE-2014-3505)
3558 [Adam Langley]
3559
3560 *) If a multithreaded client connects to a malicious server using a resumed
3561 session and the server sends an ec point format extension it could write
3562 up to 255 bytes to freed memory.
3563
3564 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3565 issue.
3566 (CVE-2014-3509)
3567 [Gabor Tyukasz]
3568
3569 *) A malicious server can crash an OpenSSL client with a null pointer
3570 dereference (read) by specifying an SRP ciphersuite even though it was not
3571 properly negotiated with the client. This can be exploited through a
3572 Denial of Service attack.
3573
3574 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3575 discovering and researching this issue.
3576 (CVE-2014-5139)
3577 [Steve Henson]
3578
3579 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3580 X509_name_oneline, X509_name_print_ex et al. to leak some information
3581 from the stack. Applications may be affected if they echo pretty printing
3582 output to the attacker.
3583
3584 Thanks to Ivan Fratric (Google) for discovering this issue.
3585 (CVE-2014-3508)
3586 [Emilia Käsper, and Steve Henson]
3587
3588 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3589 for corner cases. (Certain input points at infinity could lead to
3590 bogus results, with non-infinity inputs mapped to infinity too.)
3591 [Bodo Moeller]
3592
3593 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3594
3595 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3596 handshake can force the use of weak keying material in OpenSSL
3597 SSL/TLS clients and servers.
3598
3599 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3600 researching this issue. (CVE-2014-0224)
3601 [KIKUCHI Masashi, Steve Henson]
3602
3603 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3604 OpenSSL DTLS client the code can be made to recurse eventually crashing
3605 in a DoS attack.
3606
3607 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3608 (CVE-2014-0221)
3609 [Imre Rad, Steve Henson]
3610
3611 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3612 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3613 client or server. This is potentially exploitable to run arbitrary
3614 code on a vulnerable client or server.
3615
3616 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3617 [Jüri Aedla, Steve Henson]
3618
3619 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3620 are subject to a denial of service attack.
3621
3622 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3623 this issue. (CVE-2014-3470)
3624 [Felix Gröbert, Ivan Fratric, Steve Henson]
3625
3626 *) Harmonize version and its documentation. -f flag is used to display
3627 compilation flags.
3628 [mancha <mancha1@zoho.com>]
3629
3630 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3631 in i2d_ECPrivateKey.
3632 [mancha <mancha1@zoho.com>]
3633
3634 *) Fix some double frees. These are not thought to be exploitable.
3635 [mancha <mancha1@zoho.com>]
3636
3637 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3638
3639 *) A missing bounds check in the handling of the TLS heartbeat extension
3640 can be used to reveal up to 64k of memory to a connected client or
3641 server.
3642
3643 Thanks for Neel Mehta of Google Security for discovering this bug and to
3644 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3645 preparing the fix (CVE-2014-0160)
3646 [Adam Langley, Bodo Moeller]
3647
3648 *) Fix for the attack described in the paper "Recovering OpenSSL
3649 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3650 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3651 http://eprint.iacr.org/2014/140
3652
3653 Thanks to Yuval Yarom and Naomi Benger for discovering this
3654 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3655 [Yuval Yarom and Naomi Benger]
3656
3657 *) TLS pad extension: draft-agl-tls-padding-03
3658
3659 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3660 TLS client Hello record length value would otherwise be > 255 and
3661 less that 512 pad with a dummy extension containing zeroes so it
3662 is at least 512 bytes long.
3663
3664 [Adam Langley, Steve Henson]
3665
3666 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3667
3668 *) Fix for TLS record tampering bug. A carefully crafted invalid
3669 handshake could crash OpenSSL with a NULL pointer exception.
3670 Thanks to Anton Johansson for reporting this issues.
3671 (CVE-2013-4353)
3672
3673 *) Keep original DTLS digest and encryption contexts in retransmission
3674 structures so we can use the previous session parameters if they need
3675 to be resent. (CVE-2013-6450)
3676 [Steve Henson]
3677
3678 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3679 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3680 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3681 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3682 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3683 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3684 [Rob Stradling, Adam Langley]
3685
3686 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3687
3688 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3689 supporting platforms or when small records were transferred.
3690 [Andy Polyakov, Steve Henson]
3691
3692 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3693
3694 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3695
3696 This addresses the flaw in CBC record processing discovered by
3697 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3698 at: http://www.isg.rhul.ac.uk/tls/
3699
3700 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3701 Security Group at Royal Holloway, University of London
3702 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3703 Emilia Käsper for the initial patch.
3704 (CVE-2013-0169)
3705 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3706
3707 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3708 ciphersuites which can be exploited in a denial of service attack.
3709 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3710 and detecting this bug and to Wolfgang Ettlinger
3711 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3712 (CVE-2012-2686)
3713 [Adam Langley]
3714
3715 *) Return an error when checking OCSP signatures when key is NULL.
3716 This fixes a DoS attack. (CVE-2013-0166)
3717 [Steve Henson]
3718
3719 *) Make openssl verify return errors.
3720 [Chris Palmer <palmer@google.com> and Ben Laurie]
3721
3722 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3723 the right response is stapled. Also change SSL_get_certificate()
3724 so it returns the certificate actually sent.
3725 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3726 [Rob Stradling <rob.stradling@comodo.com>]
3727
3728 *) Fix possible deadlock when decoding public keys.
3729 [Steve Henson]
3730
3731 *) Don't use TLS 1.0 record version number in initial client hello
3732 if renegotiating.
3733 [Steve Henson]
3734
3735 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3736
3737 *) Sanity check record length before skipping explicit IV in TLS
3738 1.2, 1.1 and DTLS to fix DoS attack.
3739
3740 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3741 fuzzing as a service testing platform.
3742 (CVE-2012-2333)
3743 [Steve Henson]
3744
3745 *) Initialise tkeylen properly when encrypting CMS messages.
3746 Thanks to Solar Designer of Openwall for reporting this issue.
3747 [Steve Henson]
3748
3749 *) In FIPS mode don't try to use composite ciphers as they are not
3750 approved.
3751 [Steve Henson]
3752
3753 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3754
3755 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3756 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3757 mean any application compiled against OpenSSL 1.0.0 headers setting
3758 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3759 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3760 0x10000000L Any application which was previously compiled against
3761 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3762 will need to be recompiled as a result. Letting be results in
3763 inability to disable specifically TLS 1.1 and in client context,
3764 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3765 [Steve Henson]
3766
3767 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3768 disable just protocol X, but all protocols above X *if* there are
3769 protocols *below* X still enabled. In more practical terms it means
3770 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3771 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3772 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3773 client side.
3774 [Andy Polyakov]
3775
3776 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3777
3778 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3779 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3780 in CRYPTO_realloc_clean.
3781
3782 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3783 issue and to Adam Langley <agl@chromium.org> for fixing it.
3784 (CVE-2012-2110)
3785 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3786
3787 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3788 [Adam Langley]
3789
3790 *) Workarounds for some broken servers that "hang" if a client hello
3791 record length exceeds 255 bytes.
3792
3793 1. Do not use record version number > TLS 1.0 in initial client
3794 hello: some (but not all) hanging servers will now work.
3795 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3796 the number of ciphers sent in the client hello. This should be
3797 set to an even number, such as 50, for example by passing:
3798 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3799 Most broken servers should now work.
3800 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3801 TLS 1.2 client support entirely.
3802 [Steve Henson]
3803
3804 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3805 [Andy Polyakov]
3806
3807 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3808
3809 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3810 STRING form instead of a DigestInfo.
3811 [Steve Henson]
3812
3813 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3814 and the RSA_sign/RSA_verify functions. This was made more apparent when
3815 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3816 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3817 the correct format in RSA_verify so both forms transparently work.
3818 [Steve Henson]
3819
3820 *) Some servers which support TLS 1.0 can choke if we initially indicate
3821 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3822 encrypted premaster secret. As a workaround use the maximum permitted
3823 client version in client hello, this should keep such servers happy
3824 and still work with previous versions of OpenSSL.
3825 [Steve Henson]
3826
3827 *) Add support for TLS/DTLS heartbeats.
3828 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3829
3830 *) Add support for SCTP.
3831 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3832
3833 *) Improved PRNG seeding for VOS.
3834 [Paul Green <Paul.Green@stratus.com>]
3835
3836 *) Extensive assembler packs updates, most notably:
3837
3838 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3839 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3840 - x86_64: bit-sliced AES implementation;
3841 - ARM: NEON support, contemporary platforms optimizations;
3842 - s390x: z196 support;
3843 - *: GHASH and GF(2^m) multiplication implementations;
3844
3845 [Andy Polyakov]
3846
3847 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3848 (removal of unnecessary code)
3849 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3850
3851 *) Add TLS key material exporter from RFC 5705.
3852 [Eric Rescorla]
3853
3854 *) Add DTLS-SRTP negotiation from RFC 5764.
3855 [Eric Rescorla]
3856
3857 *) Add Next Protocol Negotiation,
3858 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3859 disabled with a no-npn flag to config or Configure. Code donated
3860 by Google.
3861 [Adam Langley <agl@google.com> and Ben Laurie]
3862
3863 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3864 NIST-P256, NIST-P521, with constant-time single point multiplication on
3865 typical inputs. Compiler support for the nonstandard type __uint128_t is
3866 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3867 Code made available under Apache License version 2.0.
3868
3869 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3870 line to include this in your build of OpenSSL, and run "make depend" (or
3871 "make update"). This enables the following EC_METHODs:
3872
3873 EC_GFp_nistp224_method()
3874 EC_GFp_nistp256_method()
3875 EC_GFp_nistp521_method()
3876
3877 EC_GROUP_new_by_curve_name() will automatically use these (while
3878 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3879 implementations).
3880 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3881
3882 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3883 all platforms. Move ssize_t definition from e_os.h to the public
3884 header file e_os2.h as it now appears in public header file cms.h
3885 [Steve Henson]
3886
3887 *) New -sigopt option to the ca, req and x509 utilities. Additional
3888 signature parameters can be passed using this option and in
3889 particular PSS.
3890 [Steve Henson]
3891
3892 *) Add RSA PSS signing function. This will generate and set the
3893 appropriate AlgorithmIdentifiers for PSS based on those in the
3894 corresponding EVP_MD_CTX structure. No application support yet.
3895 [Steve Henson]
3896
3897 *) Support for companion algorithm specific ASN1 signing routines.
3898 New function ASN1_item_sign_ctx() signs a pre-initialised
3899 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3900 the appropriate parameters.
3901 [Steve Henson]
3902
3903 *) Add new algorithm specific ASN1 verification initialisation function
3904 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3905 handling will be the same no matter what EVP_PKEY_METHOD is used.
3906 Add a PSS handler to support verification of PSS signatures: checked
3907 against a number of sample certificates.
3908 [Steve Henson]
3909
3910 *) Add signature printing for PSS. Add PSS OIDs.
3911 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3912
3913 *) Add algorithm specific signature printing. An individual ASN1 method
3914 can now print out signatures instead of the standard hex dump.
3915
3916 More complex signatures (e.g. PSS) can print out more meaningful
3917 information. Include DSA version that prints out the signature
3918 parameters r, s.
3919 [Steve Henson]
3920
3921 *) Password based recipient info support for CMS library: implementing
3922 RFC3211.
3923 [Steve Henson]
3924
3925 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3926 neatly separates the code into cipher and PBE sections and is required
3927 for some algorithms that split PBES2 into separate pieces (such as
3928 password based CMS).
3929 [Steve Henson]
3930
3931 *) Session-handling fixes:
3932 - Fix handling of connections that are resuming with a session ID,
3933 but also support Session Tickets.
3934 - Fix a bug that suppressed issuing of a new ticket if the client
3935 presented a ticket with an expired session.
3936 - Try to set the ticket lifetime hint to something reasonable.
3937 - Make tickets shorter by excluding irrelevant information.
3938 - On the client side, don't ignore renewed tickets.
3939 [Adam Langley, Bodo Moeller (Google)]
3940
3941 *) Fix PSK session representation.
3942 [Bodo Moeller]
3943
3944 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3945
3946 This work was sponsored by Intel.
3947 [Andy Polyakov]
3948
3949 *) Add GCM support to TLS library. Some custom code is needed to split
3950 the IV between the fixed (from PRF) and explicit (from TLS record)
3951 portions. This adds all GCM ciphersuites supported by RFC5288 and
3952 RFC5289. Generalise some AES* cipherstrings to include GCM and
3953 add a special AESGCM string for GCM only.
3954 [Steve Henson]
3955
3956 *) Expand range of ctrls for AES GCM. Permit setting invocation
3957 field on decrypt and retrieval of invocation field only on encrypt.
3958 [Steve Henson]
3959
3960 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3961 As required by RFC5289 these ciphersuites cannot be used if for
3962 versions of TLS earlier than 1.2.
3963 [Steve Henson]
3964
3965 *) For FIPS capable OpenSSL interpret a NULL default public key method
3966 as unset and return the appropriate default but do *not* set the default.
3967 This means we can return the appropriate method in applications that
3968 switch between FIPS and non-FIPS modes.
3969 [Steve Henson]
3970
3971 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3972 ENGINE is used then we cannot handle that in the FIPS module so we
3973 keep original code iff non-FIPS operations are allowed.
3974 [Steve Henson]
3975
3976 *) Add -attime option to openssl utilities.
3977 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3978
3979 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3980 [Steve Henson]
3981
3982 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3983 FIPS EC methods unconditionally for now.
3984 [Steve Henson]
3985
3986 *) New build option no-ec2m to disable characteristic 2 code.
3987 [Steve Henson]
3988
3989 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3990 all cases can be covered as some introduce binary incompatibilities.
3991 [Steve Henson]
3992
3993 *) Redirect RSA operations to FIPS module including keygen,
3994 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3995 [Steve Henson]
3996
3997 *) Add similar low level API blocking to ciphers.
3998 [Steve Henson]
3999
4000 *) Low level digest APIs are not approved in FIPS mode: any attempt
4001 to use these will cause a fatal error. Applications that *really* want
4002 to use them can use the private_* version instead.
4003 [Steve Henson]
4004
4005 *) Redirect cipher operations to FIPS module for FIPS builds.
4006 [Steve Henson]
4007
4008 *) Redirect digest operations to FIPS module for FIPS builds.
4009 [Steve Henson]
4010
4011 *) Update build system to add "fips" flag which will link in fipscanister.o
4012 for static and shared library builds embedding a signature if needed.
4013 [Steve Henson]
4014
4015 *) Output TLS supported curves in preference order instead of numerical
4016 order. This is currently hardcoded for the highest order curves first.
4017 This should be configurable so applications can judge speed vs strength.
4018 [Steve Henson]
4019
4020 *) Add TLS v1.2 server support for client authentication.
4021 [Steve Henson]
4022
4023 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
4024 and enable MD5.
4025 [Steve Henson]
4026
4027 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
4028 FIPS modules versions.
4029 [Steve Henson]
4030
4031 *) Add TLS v1.2 client side support for client authentication. Keep cache
4032 of handshake records longer as we don't know the hash algorithm to use
4033 until after the certificate request message is received.
4034 [Steve Henson]
4035
4036 *) Initial TLS v1.2 client support. Add a default signature algorithms
4037 extension including all the algorithms we support. Parse new signature
4038 format in client key exchange. Relax some ECC signing restrictions for
4039 TLS v1.2 as indicated in RFC5246.
4040 [Steve Henson]
4041
4042 *) Add server support for TLS v1.2 signature algorithms extension. Switch
4043 to new signature format when needed using client digest preference.
4044 All server ciphersuites should now work correctly in TLS v1.2. No client
4045 support yet and no support for client certificates.
4046 [Steve Henson]
4047
4048 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
4049 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
4050 ciphersuites. At present only RSA key exchange ciphersuites work with
4051 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
4052 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
4053 and version checking.
4054 [Steve Henson]
4055
4056 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
4057 with this defined it will not be affected by any changes to ssl internal
4058 structures. Add several utility functions to allow openssl application
4059 to work with OPENSSL_NO_SSL_INTERN defined.
4060 [Steve Henson]
4061
4062 *) A long standing patch to add support for SRP from EdelWeb (Peter
4063 Sylvester and Christophe Renou) was integrated.
4064 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
4065 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
4066 Ben Laurie]
4067
4068 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
4069 [Steve Henson]
4070
4071 *) Permit abbreviated handshakes when renegotiating using the function
4072 SSL_renegotiate_abbreviated().
4073 [Robin Seggelmann <seggelmann@fh-muenster.de>]
4074
4075 *) Add call to ENGINE_register_all_complete() to
4076 ENGINE_load_builtin_engines(), so some implementations get used
4077 automatically instead of needing explicit application support.
4078 [Steve Henson]
4079
4080 *) Add support for TLS key exporter as described in RFC5705.
4081 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
4082
4083 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
4084 a few changes are required:
4085
4086 Add SSL_OP_NO_TLSv1_1 flag.
4087 Add TLSv1_1 methods.
4088 Update version checking logic to handle version 1.1.
4089 Add explicit IV handling (ported from DTLS code).
4090 Add command line options to s_client/s_server.
4091 [Steve Henson]
4092
4093 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
4094
4095 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
4096 in CMS and PKCS7 code. When RSA decryption fails use a random key for
4097 content decryption and always return the same error. Note: this attack
4098 needs on average 2^20 messages so it only affects automated senders. The
4099 old behaviour can be re-enabled in the CMS code by setting the
4100 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
4101 an MMA defence is not necessary.
4102 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
4103 this issue. (CVE-2012-0884)
4104 [Steve Henson]
4105
4106 *) Fix CVE-2011-4619: make sure we really are receiving a
4107 client hello before rejecting multiple SGC restarts. Thanks to
4108 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
4109 [Steve Henson]
4110
4111 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
4112
4113 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
4114 Thanks to Antonio Martin, Enterprise Secure Access Research and
4115 Development, Cisco Systems, Inc. for discovering this bug and
4116 preparing a fix. (CVE-2012-0050)
4117 [Antonio Martin]
4118
4119 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
4120
4121 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
4122 of the Vaudenay padding oracle attack on CBC mode encryption
4123 which enables an efficient plaintext recovery attack against
4124 the OpenSSL implementation of DTLS. Their attack exploits timing
4125 differences arising during decryption processing. A research
4126 paper describing this attack can be found at:
4127 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
4128 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
4129 Security Group at Royal Holloway, University of London
4130 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
4131 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
4132 for preparing the fix. (CVE-2011-4108)
4133 [Robin Seggelmann, Michael Tuexen]
4134
4135 *) Clear bytes used for block padding of SSL 3.0 records.
4136 (CVE-2011-4576)
4137 [Adam Langley (Google)]
4138
4139 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
4140 Kadianakis <desnacked@gmail.com> for discovering this issue and
4141 Adam Langley for preparing the fix. (CVE-2011-4619)
4142 [Adam Langley (Google)]
4143
4144 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
4145 [Andrey Kulikov <amdeich@gmail.com>]
4146
4147 *) Prevent malformed RFC3779 data triggering an assertion failure.
4148 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
4149 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
4150 [Rob Austein <sra@hactrn.net>]
4151
4152 *) Improved PRNG seeding for VOS.
4153 [Paul Green <Paul.Green@stratus.com>]
4154
4155 *) Fix ssl_ciph.c set-up race.
4156 [Adam Langley (Google)]
4157
4158 *) Fix spurious failures in ecdsatest.c.
4159 [Emilia Käsper (Google)]
4160
4161 *) Fix the BIO_f_buffer() implementation (which was mixing different
4162 interpretations of the '..._len' fields).
4163 [Adam Langley (Google)]
4164
4165 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
4166 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
4167 threads won't reuse the same blinding coefficients.
4168
4169 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
4170 lock to call BN_BLINDING_invert_ex, and avoids one use of
4171 BN_BLINDING_update for each BN_BLINDING structure (previously,
4172 the last update always remained unused).
4173 [Emilia Käsper (Google)]
4174
4175 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
4176 [Bob Buckholz (Google)]
4177
4178 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
4179
4180 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
4181 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
4182 [Kaspar Brand <ossl@velox.ch>]
4183
4184 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
4185 for multi-threaded use of ECDH. (CVE-2011-3210)
4186 [Adam Langley (Google)]
4187
4188 *) Fix x509_name_ex_d2i memory leak on bad inputs.
4189 [Bodo Moeller]
4190
4191 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
4192 signature public key algorithm by using OID xref utilities instead.
4193 Before this you could only use some ECC ciphersuites with SHA1 only.
4194 [Steve Henson]
4195
4196 *) Add protection against ECDSA timing attacks as mentioned in the paper
4197 by Billy Bob Brumley and Nicola Tuveri, see:
4198
4199 http://eprint.iacr.org/2011/232.pdf
4200
4201 [Billy Bob Brumley and Nicola Tuveri]
4202
4203 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
4204
4205 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
4206 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
4207
4208 *) Fix bug in string printing code: if *any* escaping is enabled we must
4209 escape the escape character (backslash) or the resulting string is
4210 ambiguous.
4211 [Steve Henson]
4212
4213 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
4214
4215 *) Disable code workaround for ancient and obsolete Netscape browsers
4216 and servers: an attacker can use it in a ciphersuite downgrade attack.
4217 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
4218 [Steve Henson]
4219
4220 *) Fixed J-PAKE implementation error, originally discovered by
4221 Sebastien Martini, further info and confirmation from Stefan
4222 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
4223 [Ben Laurie]
4224
4225 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
4226
4227 *) Fix extension code to avoid race conditions which can result in a buffer
4228 overrun vulnerability: resumed sessions must not be modified as they can
4229 be shared by multiple threads. CVE-2010-3864
4230 [Steve Henson]
4231
4232 *) Fix WIN32 build system to correctly link an ENGINE directory into
4233 a DLL.
4234 [Steve Henson]
4235
4236 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
4237
4238 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
4239 (CVE-2010-1633)
4240 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
4241
4242 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
4243
4244 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
4245 context. The operation can be customised via the ctrl mechanism in
4246 case ENGINEs want to include additional functionality.
4247 [Steve Henson]
4248
4249 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
4250 [Steve Henson]
4251
4252 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
4253 output hashes compatible with older versions of OpenSSL.
4254 [Willy Weisz <weisz@vcpc.univie.ac.at>]
4255
4256 *) Fix compression algorithm handling: if resuming a session use the
4257 compression algorithm of the resumed session instead of determining
4258 it from client hello again. Don't allow server to change algorithm.
4259 [Steve Henson]
4260
4261 *) Add load_crls() function to apps tidying load_certs() too. Add option
4262 to verify utility to allow additional CRLs to be included.
4263 [Steve Henson]
4264
4265 *) Update OCSP request code to permit adding custom headers to the request:
4266 some responders need this.
4267 [Steve Henson]
4268
4269 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
4270 correctly.
4271 [Julia Lawall <julia@diku.dk>]
4272
4273 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
4274 needlessly dereferenced structures, used obsolete functions and
4275 didn't handle all updated verify codes correctly.
4276 [Steve Henson]
4277
4278 *) Disable MD2 in the default configuration.
4279 [Steve Henson]
4280
4281 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
4282 indicate the initial BIO being pushed or popped. This makes it possible
4283 to determine whether the BIO is the one explicitly called or as a result
4284 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
4285 it handles reference counts correctly and doesn't zero out the I/O bio
4286 when it is not being explicitly popped. WARNING: applications which
4287 included workarounds for the old buggy behaviour will need to be modified
4288 or they could free up already freed BIOs.
4289 [Steve Henson]
4290
4291 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
4292 renaming to all platforms (within the 0.9.8 branch, this was
4293 done conditionally on Netware platforms to avoid a name clash).
4294 [Guenter <lists@gknw.net>]
4295
4296 *) Add ECDHE and PSK support to DTLS.
4297 [Michael Tuexen <tuexen@fh-muenster.de>]
4298
4299 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
4300 be used on C++.
4301 [Steve Henson]
4302
4303 *) Add "missing" function EVP_MD_flags() (without this the only way to
4304 retrieve a digest flags is by accessing the structure directly. Update
4305 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
4306 or cipher is registered as in the "from" argument. Print out all
4307 registered digests in the dgst usage message instead of manually
4308 attempting to work them out.
4309 [Steve Henson]
4310
4311 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
4312 this allows the use of compression and extensions. Change default cipher
4313 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
4314 by default unless an application cipher string requests it.
4315 [Steve Henson]
4316
4317 *) Alter match criteria in PKCS12_parse(). It used to try to use local
4318 key ids to find matching certificates and keys but some PKCS#12 files
4319 don't follow the (somewhat unwritten) rules and this strategy fails.
4320 Now just gather all certificates together and the first private key
4321 then look for the first certificate that matches the key.
4322 [Steve Henson]
4323
4324 *) Support use of registered digest and cipher names for dgst and cipher
4325 commands instead of having to add each one as a special case. So now
4326 you can do:
4327
4328 openssl sha256 foo
4329
4330 as well as:
4331
4332 openssl dgst -sha256 foo
4333
4334 and this works for ENGINE based algorithms too.
4335
4336 [Steve Henson]
4337
4338 *) Update Gost ENGINE to support parameter files.
4339 [Victor B. Wagner <vitus@cryptocom.ru>]
4340
4341 *) Support GeneralizedTime in ca utility.
4342 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
4343
4344 *) Enhance the hash format used for certificate directory links. The new
4345 form uses the canonical encoding (meaning equivalent names will work
4346 even if they aren't identical) and uses SHA1 instead of MD5. This form
4347 is incompatible with the older format and as a result c_rehash should
4348 be used to rebuild symbolic links.
4349 [Steve Henson]
4350
4351 *) Make PKCS#8 the default write format for private keys, replacing the
4352 traditional format. This form is standardised, more secure and doesn't
4353 include an implicit MD5 dependency.
4354 [Steve Henson]
4355
4356 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
4357 committed to OpenSSL should pass this lot as a minimum.
4358 [Steve Henson]
4359
4360 *) Add session ticket override functionality for use by EAP-FAST.
4361 [Jouni Malinen <j@w1.fi>]
4362
4363 *) Modify HMAC functions to return a value. Since these can be implemented
4364 in an ENGINE errors can occur.
4365 [Steve Henson]
4366
4367 *) Type-checked OBJ_bsearch_ex.
4368 [Ben Laurie]
4369
4370 *) Type-checked OBJ_bsearch. Also some constification necessitated
4371 by type-checking. Still to come: TXT_DB, bsearch(?),
4372 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
4373 CONF_VALUE.
4374 [Ben Laurie]
4375
4376 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
4377 seconds to a tm structure directly, instead of going through OS
4378 specific date routines. This avoids any issues with OS routines such
4379 as the year 2038 bug. New *_adj() functions for ASN1 time structures
4380 and X509_time_adj_ex() to cover the extended range. The existing
4381 X509_time_adj() is still usable and will no longer have any date issues.
4382 [Steve Henson]
4383
4384 *) Delta CRL support. New use deltas option which will attempt to locate
4385 and search any appropriate delta CRLs available.
4386
4387 This work was sponsored by Google.
4388 [Steve Henson]
4389
4390 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
4391 code and add additional score elements. Validate alternate CRL paths
4392 as part of the CRL checking and indicate a new error "CRL path validation
4393 error" in this case. Applications wanting additional details can use
4394 the verify callback and check the new "parent" field. If this is not
4395 NULL CRL path validation is taking place. Existing applications won't
4396 see this because it requires extended CRL support which is off by
4397 default.
4398
4399 This work was sponsored by Google.
4400 [Steve Henson]
4401
4402 *) Support for freshest CRL extension.
4403
4404 This work was sponsored by Google.
4405 [Steve Henson]
4406
4407 *) Initial indirect CRL support. Currently only supported in the CRLs
4408 passed directly and not via lookup. Process certificate issuer
4409 CRL entry extension and lookup CRL entries by bother issuer name
4410 and serial number. Check and process CRL issuer entry in IDP extension.
4411
4412 This work was sponsored by Google.
4413 [Steve Henson]
4414
4415 *) Add support for distinct certificate and CRL paths. The CRL issuer
4416 certificate is validated separately in this case. Only enabled if
4417 an extended CRL support flag is set: this flag will enable additional
4418 CRL functionality in future.
4419
4420 This work was sponsored by Google.
4421 [Steve Henson]
4422
4423 *) Add support for policy mappings extension.
4424
4425 This work was sponsored by Google.
4426 [Steve Henson]
4427
4428 *) Fixes to pathlength constraint, self issued certificate handling,
4429 policy processing to align with RFC3280 and PKITS tests.
4430
4431 This work was sponsored by Google.
4432 [Steve Henson]
4433
4434 *) Support for name constraints certificate extension. DN, email, DNS
4435 and URI types are currently supported.
4436
4437 This work was sponsored by Google.
4438 [Steve Henson]
4439
4440 *) To cater for systems that provide a pointer-based thread ID rather
4441 than numeric, deprecate the current numeric thread ID mechanism and
4442 replace it with a structure and associated callback type. This
4443 mechanism allows a numeric "hash" to be extracted from a thread ID in
4444 either case, and on platforms where pointers are larger than 'long',
4445 mixing is done to help ensure the numeric 'hash' is usable even if it
4446 can't be guaranteed unique. The default mechanism is to use "&errno"
4447 as a pointer-based thread ID to distinguish between threads.
4448
4449 Applications that want to provide their own thread IDs should now use
4450 CRYPTO_THREADID_set_callback() to register a callback that will call
4451 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
4452
4453 Note that ERR_remove_state() is now deprecated, because it is tied
4454 to the assumption that thread IDs are numeric. ERR_remove_state(0)
4455 to free the current thread's error state should be replaced by
4456 ERR_remove_thread_state(NULL).
4457
4458 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
4459 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
4460 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
4461 application was previously providing a numeric thread callback that
4462 was inappropriate for distinguishing threads, then uniqueness might
4463 have been obtained with &errno that happened immediately in the
4464 intermediate development versions of OpenSSL; this is no longer the
4465 case, the numeric thread callback will now override the automatic use
4466 of &errno.)
4467 [Geoff Thorpe, with help from Bodo Moeller]
4468
4469 *) Initial support for different CRL issuing certificates. This covers a
4470 simple case where the self issued certificates in the chain exist and
4471 the real CRL issuer is higher in the existing chain.
4472
4473 This work was sponsored by Google.
4474 [Steve Henson]
4475
4476 *) Removed effectively defunct crypto/store from the build.
4477 [Ben Laurie]
4478
4479 *) Revamp of STACK to provide stronger type-checking. Still to come:
4480 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
4481 ASN1_STRING, CONF_VALUE.
4482 [Ben Laurie]
4483
4484 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4485 RAM on SSL connections. This option can save about 34k per idle SSL.
4486 [Nick Mathewson]
4487
4488 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4489 STACK, TXT_DB, bsearch, qsort.
4490 [Ben Laurie]
4491
4492 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4493 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4494 support for data, signedData, compressedData, digestedData and
4495 encryptedData, envelopedData types included. Scripts to check against
4496 RFC4134 examples draft and interop and consistency checks of many
4497 content types and variants.
4498 [Steve Henson]
4499
4500 *) Add options to enc utility to support use of zlib compression BIO.
4501 [Steve Henson]
4502
4503 *) Extend mk1mf to support importing of options and assembly language
4504 files from Configure script, currently only included in VC-WIN32.
4505 The assembly language rules can now optionally generate the source
4506 files from the associated perl scripts.
4507 [Steve Henson]
4508
4509 *) Implement remaining functionality needed to support GOST ciphersuites.
4510 Interop testing has been performed using CryptoPro implementations.
4511 [Victor B. Wagner <vitus@cryptocom.ru>]
4512
4513 *) s390x assembler pack.
4514 [Andy Polyakov]
4515
4516 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4517 "family."
4518 [Andy Polyakov]
4519
4520 *) Implement Opaque PRF Input TLS extension as specified in
4521 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4522 official specification yet and no extension type assignment by
4523 IANA exists, this extension (for now) will have to be explicitly
4524 enabled when building OpenSSL by providing the extension number
4525 to use. For example, specify an option
4526
4527 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4528
4529 to the "config" or "Configure" script to enable the extension,
4530 assuming extension number 0x9527 (which is a completely arbitrary
4531 and unofficial assignment based on the MD5 hash of the Internet
4532 Draft). Note that by doing so, you potentially lose
4533 interoperability with other TLS implementations since these might
4534 be using the same extension number for other purposes.
4535
4536 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4537 opaque PRF input value to use in the handshake. This will create
4538 an internal copy of the length-'len' string at 'src', and will
4539 return non-zero for success.
4540
4541 To get more control and flexibility, provide a callback function
4542 by using
4543
4544 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4545 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4546
4547 where
4548
4549 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4550 void *arg;
4551
4552 Callback function 'cb' will be called in handshakes, and is
4553 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4554 Argument 'arg' is for application purposes (the value as given to
4555 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4556 be provided to the callback function). The callback function
4557 has to return non-zero to report success: usually 1 to use opaque
4558 PRF input just if possible, or 2 to enforce use of the opaque PRF
4559 input. In the latter case, the library will abort the handshake
4560 if opaque PRF input is not successfully negotiated.
4561
4562 Arguments 'peerinput' and 'len' given to the callback function
4563 will always be NULL and 0 in the case of a client. A server will
4564 see the client's opaque PRF input through these variables if
4565 available (NULL and 0 otherwise). Note that if the server
4566 provides an opaque PRF input, the length must be the same as the
4567 length of the client's opaque PRF input.
4568
4569 Note that the callback function will only be called when creating
4570 a new session (session resumption can resume whatever was
4571 previously negotiated), and will not be called in SSL 2.0
4572 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4573 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4574 for applications that need to enforce opaque PRF input.
4575
4576 [Bodo Moeller]
4577
4578 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4579 MAC.
4580
4581 [Victor B. Wagner <vitus@cryptocom.ru>]
4582
4583 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4584 RFC4507bis. The encrypted ticket format is an encrypted encoded
4585 SSL_SESSION structure, that way new session features are automatically
4586 supported.
4587
4588 If a client application caches session in an SSL_SESSION structure
4589 support is transparent because tickets are now stored in the encoded
4590 SSL_SESSION.
4591
4592 The SSL_CTX structure automatically generates keys for ticket
4593 protection in servers so again support should be possible
4594 with no application modification.
4595
4596 If a client or server wishes to disable RFC4507 support then the option
4597 SSL_OP_NO_TICKET can be set.
4598
4599 Add a TLS extension debugging callback to allow the contents of any client
4600 or server extensions to be examined.
4601
4602 This work was sponsored by Google.
4603 [Steve Henson]
4604
4605 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4606 OpenSSL should now compile cleanly on gcc 4.2
4607 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4608
4609 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4610 support including streaming MAC support: this is required for GOST
4611 ciphersuite support.
4612 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4613
4614 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4615 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4616 to output in BER and PEM format.
4617 [Steve Henson]
4618
4619 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4620 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4621 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4622 ENGINE support for HMAC keys which are unextractable. New -mac and
4623 -macopt options to dgst utility.
4624 [Steve Henson]
4625
4626 *) New option -sigopt to dgst utility. Update dgst to use
4627 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4628 alternative signing parameters such as X9.31 or PSS in the dgst
4629 utility.
4630 [Steve Henson]
4631
4632 *) Change ssl_cipher_apply_rule(), the internal function that does
4633 the work each time a ciphersuite string requests enabling
4634 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4635 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4636 the order of disabled ciphersuites such that those ciphersuites
4637 that most recently went from enabled to disabled not only stay
4638 in order with respect to each other, but also have higher priority
4639 than other disabled ciphersuites the next time ciphersuites are
4640 enabled again.
4641
4642 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4643 the same ciphersuites as with "HIGH" alone, but in a specific
4644 order where the PSK ciphersuites come first (since they are the
4645 most recently disabled ciphersuites when "HIGH" is parsed).
4646
4647 Also, change ssl_create_cipher_list() (using this new
4648 functionality) such that between otherwise identical
4649 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4650 the default order.
4651 [Bodo Moeller]
4652
4653 *) Change ssl_create_cipher_list() so that it automatically
4654 arranges the ciphersuites in reasonable order before starting
4655 to process the rule string. Thus, the definition for "DEFAULT"
4656 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4657 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4658 This makes it much easier to arrive at a reasonable default order
4659 in applications for which anonymous ciphers are OK (meaning
4660 that you can't actually use DEFAULT).
4661 [Bodo Moeller; suggested by Victor Duchovni]
4662
4663 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4664 processing) into multiple integers instead of setting
4665 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
4666 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
4667 (These masks as well as the individual bit definitions are hidden
4668 away into the non-exported interface ssl/ssl_locl.h, so this
4669 change to the definition of the SSL_CIPHER structure shouldn't
4670 affect applications.) This give us more bits for each of these
4671 categories, so there is no longer a need to coagulate AES128 and
4672 AES256 into a single algorithm bit, and to coagulate Camellia128
4673 and Camellia256 into a single algorithm bit, which has led to all
4674 kinds of kludges.
4675
4676 Thus, among other things, the kludge introduced in 0.9.7m and
4677 0.9.8e for masking out AES256 independently of AES128 or masking
4678 out Camellia256 independently of AES256 is not needed here in 0.9.9.
4679
4680 With the change, we also introduce new ciphersuite aliases that
4681 so far were missing: "AES128", "AES256", "CAMELLIA128", and
4682 "CAMELLIA256".
4683 [Bodo Moeller]
4684
4685 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
4686 Use the leftmost N bytes of the signature input if the input is
4687 larger than the prime q (with N being the size in bytes of q).
4688 [Nils Larsch]
4689
4690 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
4691 it yet and it is largely untested.
4692 [Steve Henson]
4693
4694 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
4695 [Nils Larsch]
4696
4697 *) Initial incomplete changes to avoid need for function casts in OpenSSL
4698 some compilers (gcc 4.2 and later) reject their use. Safestack is
4699 reimplemented. Update ASN1 to avoid use of legacy functions.
4700 [Steve Henson]
4701
4702 *) Win32/64 targets are linked with Winsock2.
4703 [Andy Polyakov]
4704
4705 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
4706 to external functions. This can be used to increase CRL handling
4707 efficiency especially when CRLs are very large by (for example) storing
4708 the CRL revoked certificates in a database.
4709 [Steve Henson]
4710
4711 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
4712 new CRLs added to a directory can be used. New command line option
4713 -verify_return_error to s_client and s_server. This causes real errors
4714 to be returned by the verify callback instead of carrying on no matter
4715 what. This reflects the way a "real world" verify callback would behave.
4716 [Steve Henson]
4717
4718 *) GOST engine, supporting several GOST algorithms and public key formats.
4719 Kindly donated by Cryptocom.
4720 [Cryptocom]
4721
4722 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4723 partitioned by DP are handled but no indirect CRL or reason partitioning
4724 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4725 selected via a scoring technique which handles IDP and AKID in CRLs.
4726 [Steve Henson]
4727
4728 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4729 will ultimately be used for all verify operations: this will remove the
4730 X509_STORE dependency on certificate verification and allow alternative
4731 lookup methods. X509_STORE based implementations of these two callbacks.
4732 [Steve Henson]
4733
4734 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4735 Modify get_crl() to find a valid (unexpired) CRL if possible.
4736 [Steve Henson]
4737
4738 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4739 this would be called X509_CRL_cmp() but that name is already used by
4740 a function that just compares CRL issuer names. Cache several CRL
4741 extensions in X509_CRL structure and cache CRLDP in X509.
4742 [Steve Henson]
4743
4744 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4745 this maps equivalent X509_NAME structures into a consistent structure.
4746 Name comparison can then be performed rapidly using memcmp().
4747 [Steve Henson]
4748
4749 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4750 utility.
4751 [Steve Henson]
4752
4753 *) Allow digests to supply their own micalg string for S/MIME type using
4754 the ctrl EVP_MD_CTRL_MICALG.
4755 [Steve Henson]
4756
4757 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4758 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4759 ctrl. It can then customise the structure before and/or after signing
4760 if necessary.
4761 [Steve Henson]
4762
4763 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4764 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4765 to free up any added signature OIDs.
4766 [Steve Henson]
4767
4768 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4769 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4770 digest and cipher tables. New options added to openssl utility:
4771 list-message-digest-algorithms and list-cipher-algorithms.
4772 [Steve Henson]
4773
4774 *) Change the array representation of binary polynomials: the list
4775 of degrees of non-zero coefficients is now terminated with -1.
4776 Previously it was terminated with 0, which was also part of the
4777 value; thus, the array representation was not applicable to
4778 polynomials where t^0 has coefficient zero. This change makes
4779 the array representation useful in a more general context.
4780 [Douglas Stebila]
4781
4782 *) Various modifications and fixes to SSL/TLS cipher string
4783 handling. For ECC, the code now distinguishes between fixed ECDH
4784 with RSA certificates on the one hand and with ECDSA certificates
4785 on the other hand, since these are separate ciphersuites. The
4786 unused code for Fortezza ciphersuites has been removed.
4787
4788 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4789 (not "ECDHE"). For consistency with the code for DH
4790 certificates, use of ECDH certificates is now considered ECDH
4791 authentication, not RSA or ECDSA authentication (the latter is
4792 merely the CA's signing algorithm and not actively used in the
4793 protocol).
4794
4795 The temporary ciphersuite alias "ECCdraft" is no longer
4796 available, and ECC ciphersuites are no longer excluded from "ALL"
4797 and "DEFAULT". The following aliases now exist for RFC 4492
4798 ciphersuites, most of these by analogy with the DH case:
4799
4800 kECDHr - ECDH cert, signed with RSA
4801 kECDHe - ECDH cert, signed with ECDSA
4802 kECDH - ECDH cert (signed with either RSA or ECDSA)
4803 kEECDH - ephemeral ECDH
4804 ECDH - ECDH cert or ephemeral ECDH
4805
4806 aECDH - ECDH cert
4807 aECDSA - ECDSA cert
4808 ECDSA - ECDSA cert
4809
4810 AECDH - anonymous ECDH
4811 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4812
4813 [Bodo Moeller]
4814
4815 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
4816 Use correct micalg parameters depending on digest(s) in signed message.
4817 [Steve Henson]
4818
4819 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
4820 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
4821 [Steve Henson]
4822
4823 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
4824 an engine to register a method. Add ENGINE lookups for methods and
4825 functional reference processing.
4826 [Steve Henson]
4827
4828 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
4829 EVP_{Sign,Verify}* which allow an application to customise the signature
4830 process.
4831 [Steve Henson]
4832
4833 *) New -resign option to smime utility. This adds one or more signers
4834 to an existing PKCS#7 signedData structure. Also -md option to use an
4835 alternative message digest algorithm for signing.
4836 [Steve Henson]
4837
4838 *) Tidy up PKCS#7 routines and add new functions to make it easier to
4839 create PKCS7 structures containing multiple signers. Update smime
4840 application to support multiple signers.
4841 [Steve Henson]
4842
4843 *) New -macalg option to pkcs12 utility to allow setting of an alternative
4844 digest MAC.
4845 [Steve Henson]
4846
4847 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
4848 Reorganize PBE internals to lookup from a static table using NIDs,
4849 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
4850 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
4851 PRF which will be automatically used with PBES2.
4852 [Steve Henson]
4853
4854 *) Replace the algorithm specific calls to generate keys in "req" with the
4855 new API.
4856 [Steve Henson]
4857
4858 *) Update PKCS#7 enveloped data routines to use new API. This is now
4859 supported by any public key method supporting the encrypt operation. A
4860 ctrl is added to allow the public key algorithm to examine or modify
4861 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
4862 a no op.
4863 [Steve Henson]
4864
4865 *) Add a ctrl to asn1 method to allow a public key algorithm to express
4866 a default digest type to use. In most cases this will be SHA1 but some
4867 algorithms (such as GOST) need to specify an alternative digest. The
4868 return value indicates how strong the preference is 1 means optional and
4869 2 is mandatory (that is it is the only supported type). Modify
4870 ASN1_item_sign() to accept a NULL digest argument to indicate it should
4871 use the default md. Update openssl utilities to use the default digest
4872 type for signing if it is not explicitly indicated.
4873 [Steve Henson]
4874
4875 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
4876 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
4877 signing method from the key type. This effectively removes the link
4878 between digests and public key types.
4879 [Steve Henson]
4880
4881 *) Add an OID cross reference table and utility functions. Its purpose is to
4882 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
4883 rsaEncryption. This will allow some of the algorithm specific hackery
4884 needed to use the correct OID to be removed.
4885 [Steve Henson]
4886
4887 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
4888 structures for PKCS7_sign(). They are now set up by the relevant public
4889 key ASN1 method.
4890 [Steve Henson]
4891
4892 *) Add provisional EC pkey method with support for ECDSA and ECDH.
4893 [Steve Henson]
4894
4895 *) Add support for key derivation (agreement) in the API, DH method and
4896 pkeyutl.
4897 [Steve Henson]
4898
4899 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
4900 public and private key formats. As a side effect these add additional
4901 command line functionality not previously available: DSA signatures can be
4902 generated and verified using pkeyutl and DH key support and generation in
4903 pkey, genpkey.
4904 [Steve Henson]
4905
4906 *) BeOS support.
4907 [Oliver Tappe <zooey@hirschkaefer.de>]
4908
4909 *) New make target "install_html_docs" installs HTML renditions of the
4910 manual pages.
4911 [Oliver Tappe <zooey@hirschkaefer.de>]
4912
4913 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4914 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4915 support key and parameter generation and add initial key generation
4916 functionality for RSA.
4917 [Steve Henson]
4918
4919 *) Add functions for main EVP_PKEY_method operations. The undocumented
4920 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4921 EVP_PKEY_{encrypt,decrypt}_old.
4922 [Steve Henson]
4923
4924 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4925 key API, doesn't do much yet.
4926 [Steve Henson]
4927
4928 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4929 public key algorithms. New option to openssl utility:
4930 "list-public-key-algorithms" to print out info.
4931 [Steve Henson]
4932
4933 *) Implement the Supported Elliptic Curves Extension for
4934 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4935 [Douglas Stebila]
4936
4937 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4938 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4939 [Steve Henson]
4940
4941 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4942 utilities such as rsa, dsa, dsaparam etc except they process any key
4943 type.
4944 [Steve Henson]
4945
4946 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4947 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4948 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4949 structure.
4950 [Steve Henson]
4951
4952 *) Initial support for pluggable public key ASN1.
4953 De-spaghettify the public key ASN1 handling. Move public and private
4954 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4955 algorithm specific handling to a single module within the relevant
4956 algorithm directory. Add functions to allow (near) opaque processing
4957 of public and private key structures.
4958 [Steve Henson]
4959
4960 *) Implement the Supported Point Formats Extension for
4961 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4962 [Douglas Stebila]
4963
4964 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4965 for the psk identity [hint] and the psk callback functions to the
4966 SSL_SESSION, SSL and SSL_CTX structure.
4967
4968 New ciphersuites:
4969 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4970 PSK-AES256-CBC-SHA
4971
4972 New functions:
4973 SSL_CTX_use_psk_identity_hint
4974 SSL_get_psk_identity_hint
4975 SSL_get_psk_identity
4976 SSL_use_psk_identity_hint
4977
4978 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4979
4980 *) Add RFC 3161 compliant time stamp request creation, response generation
4981 and response verification functionality.
4982 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4983
4984 *) Add initial support for TLS extensions, specifically for the server_name
4985 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4986 have new members for a host name. The SSL data structure has an
4987 additional member SSL_CTX *initial_ctx so that new sessions can be
4988 stored in that context to allow for session resumption, even after the
4989 SSL has been switched to a new SSL_CTX in reaction to a client's
4990 server_name extension.
4991
4992 New functions (subject to change):
4993
4994 SSL_get_servername()
4995 SSL_get_servername_type()
4996 SSL_set_SSL_CTX()
4997
4998 New CTRL codes and macros (subject to change):
4999
5000 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5001 - SSL_CTX_set_tlsext_servername_callback()
5002 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5003 - SSL_CTX_set_tlsext_servername_arg()
5004 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5005
5006 openssl s_client has a new '-servername ...' option.
5007
5008 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5009 '-key2 ...', '-servername_fatal' (subject to change). This allows
5010 testing the HostName extension for a specific single host name ('-cert'
5011 and '-key' remain fallbacks for handshakes without HostName
5012 negotiation). If the unrecognized_name alert has to be sent, this by
5013 default is a warning; it becomes fatal with the '-servername_fatal'
5014 option.
5015
5016 [Peter Sylvester, Remy Allais, Christophe Renou]
5017
5018 *) Whirlpool hash implementation is added.
5019 [Andy Polyakov]
5020
5021 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
5022 bn(64,32). Because of instruction set limitations it doesn't have
5023 any negative impact on performance. This was done mostly in order
5024 to make it possible to share assembler modules, such as bn_mul_mont
5025 implementations, between 32- and 64-bit builds without hassle.
5026 [Andy Polyakov]
5027
5028 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
5029 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
5030 macro.
5031 [Bodo Moeller]
5032
5033 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
5034 dedicated Montgomery multiplication procedure, is introduced.
5035 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
5036 "64-bit" performance on certain 32-bit targets.
5037 [Andy Polyakov]
5038
5039 *) New option SSL_OP_NO_COMP to disable use of compression selectively
5040 in SSL structures. New SSL ctrl to set maximum send fragment size.
5041 Save memory by setting the I/O buffer sizes dynamically instead of
5042 using the maximum available value.
5043 [Steve Henson]
5044
5045 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
5046 in addition to the text details.
5047 [Bodo Moeller]
5048
5049 *) Very, very preliminary EXPERIMENTAL support for printing of general
5050 ASN1 structures. This currently produces rather ugly output and doesn't
5051 handle several customised structures at all.
5052 [Steve Henson]
5053
5054 *) Integrated support for PVK file format and some related formats such
5055 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
5056 these in the 'rsa' and 'dsa' utilities.
5057 [Steve Henson]
5058
5059 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
5060 [Steve Henson]
5061
5062 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
5063 place for the (very old) "NETSCAPE" format certificates which are now
5064 handled using new ASN1 code equivalents.
5065 [Steve Henson]
5066
5067 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
5068 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
5069 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
5070 [Nils Larsch]
5071
5072 *) Modify CRL distribution points extension code to print out previously
5073 unsupported fields. Enhance extension setting code to allow setting of
5074 all fields.
5075 [Steve Henson]
5076
5077 *) Add print and set support for Issuing Distribution Point CRL extension.
5078 [Steve Henson]
5079
5080 *) Change 'Configure' script to enable Camellia by default.
5081 [NTT]
5082
5083 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
5084
5085 *) When rejecting SSL/TLS records due to an incorrect version number, never
5086 update s->server with a new major version number. As of
5087 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
5088 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
5089 the previous behavior could result in a read attempt at NULL when
5090 receiving specific incorrect SSL/TLS records once record payload
5091 protection is active. (CVE-2010-0740)
5092 [Bodo Moeller, Adam Langley <agl@chromium.org>]
5093
5094 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
5095 could be crashed if the relevant tables were not present (e.g. chrooted).
5096 [Tomas Hoger <thoger@redhat.com>]
5097
5098 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
5099
5100 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
5101 [Martin Olsson, Neel Mehta]
5102
5103 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
5104 accommodate for stack sorting, always a write lock!).
5105 [Bodo Moeller]
5106
5107 *) On some versions of WIN32 Heap32Next is very slow. This can cause
5108 excessive delays in the RAND_poll(): over a minute. As a workaround
5109 include a time check in the inner Heap32Next loop too.
5110 [Steve Henson]
5111
5112 *) The code that handled flushing of data in SSL/TLS originally used the
5113 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
5114 the problem outlined in PR#1949. The fix suggested there however can
5115 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
5116 of Apache). So instead simplify the code to flush unconditionally.
5117 This should be fine since flushing with no data to flush is a no op.
5118 [Steve Henson]
5119
5120 *) Handle TLS versions 2.0 and later properly and correctly use the
5121 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
5122 off ancient servers have a habit of sticking around for a while...
5123 [Steve Henson]
5124
5125 *) Modify compression code so it frees up structures without using the
5126 ex_data callbacks. This works around a problem where some applications
5127 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
5128 restarting) then use compression (e.g. SSL with compression) later.
5129 This results in significant per-connection memory leaks and
5130 has caused some security issues including CVE-2008-1678 and
5131 CVE-2009-4355.
5132 [Steve Henson]
5133
5134 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
5135 change when encrypting or decrypting.
5136 [Bodo Moeller]
5137
5138 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
5139 connect and renegotiate with servers which do not support RI.
5140 Until RI is more widely deployed this option is enabled by default.
5141 [Steve Henson]
5142
5143 *) Add "missing" ssl ctrls to clear options and mode.
5144 [Steve Henson]
5145
5146 *) If client attempts to renegotiate and doesn't support RI respond with
5147 a no_renegotiation alert as required by RFC5746. Some renegotiating
5148 TLS clients will continue a connection gracefully when they receive
5149 the alert. Unfortunately OpenSSL mishandled this alert and would hang
5150 waiting for a server hello which it will never receive. Now we treat a
5151 received no_renegotiation alert as a fatal error. This is because
5152 applications requesting a renegotiation might well expect it to succeed
5153 and would have no code in place to handle the server denying it so the
5154 only safe thing to do is to terminate the connection.
5155 [Steve Henson]
5156
5157 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
5158 peer supports secure renegotiation and 0 otherwise. Print out peer
5159 renegotiation support in s_client/s_server.
5160 [Steve Henson]
5161
5162 *) Replace the highly broken and deprecated SPKAC certification method with
5163 the updated NID creation version. This should correctly handle UTF8.
5164 [Steve Henson]
5165
5166 *) Implement RFC5746. Re-enable renegotiation but require the extension
5167 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
5168 turns out to be a bad idea. It has been replaced by
5169 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
5170 SSL_CTX_set_options(). This is really not recommended unless you
5171 know what you are doing.
5172 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
5173
5174 *) Fixes to stateless session resumption handling. Use initial_ctx when
5175 issuing and attempting to decrypt tickets in case it has changed during
5176 servername handling. Use a non-zero length session ID when attempting
5177 stateless session resumption: this makes it possible to determine if
5178 a resumption has occurred immediately after receiving server hello
5179 (several places in OpenSSL subtly assume this) instead of later in
5180 the handshake.
5181 [Steve Henson]
5182
5183 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
5184 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
5185 fixes for a few places where the return code is not checked
5186 correctly.
5187 [Julia Lawall <julia@diku.dk>]
5188
5189 *) Add --strict-warnings option to Configure script to include devteam
5190 warnings in other configurations.
5191 [Steve Henson]
5192
5193 *) Add support for --libdir option and LIBDIR variable in makefiles. This
5194 makes it possible to install openssl libraries in locations which
5195 have names other than "lib", for example "/usr/lib64" which some
5196 systems need.
5197 [Steve Henson, based on patch from Jeremy Utley]
5198
5199 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
5200 X690 8.9.12 and can produce some misleading textual output of OIDs.
5201 [Steve Henson, reported by Dan Kaminsky]
5202
5203 *) Delete MD2 from algorithm tables. This follows the recommendation in
5204 several standards that it is not used in new applications due to
5205 several cryptographic weaknesses. For binary compatibility reasons
5206 the MD2 API is still compiled in by default.
5207 [Steve Henson]
5208
5209 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
5210 and restored.
5211 [Steve Henson]
5212
5213 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
5214 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
5215 clash.
5216 [Guenter <lists@gknw.net>]
5217
5218 *) Fix the server certificate chain building code to use X509_verify_cert(),
5219 it used to have an ad-hoc builder which was unable to cope with anything
5220 other than a simple chain.
5221 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
5222
5223 *) Don't check self signed certificate signatures in X509_verify_cert()
5224 by default (a flag can override this): it just wastes time without
5225 adding any security. As a useful side effect self signed root CAs
5226 with non-FIPS digests are now usable in FIPS mode.
5227 [Steve Henson]
5228
5229 *) In dtls1_process_out_of_seq_message() the check if the current message
5230 is already buffered was missing. For every new message was memory
5231 allocated, allowing an attacker to perform an denial of service attack
5232 with sending out of seq handshake messages until there is no memory
5233 left. Additionally every future message was buffered, even if the
5234 sequence number made no sense and would be part of another handshake.
5235 So only messages with sequence numbers less than 10 in advance will be
5236 buffered. (CVE-2009-1378)
5237 [Robin Seggelmann, discovered by Daniel Mentz]
5238
5239 *) Records are buffered if they arrive with a future epoch to be
5240 processed after finishing the corresponding handshake. There is
5241 currently no limitation to this buffer allowing an attacker to perform
5242 a DOS attack with sending records with future epochs until there is no
5243 memory left. This patch adds the pqueue_size() function to determine
5244 the size of a buffer and limits the record buffer to 100 entries.
5245 (CVE-2009-1377)
5246 [Robin Seggelmann, discovered by Daniel Mentz]
5247
5248 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
5249 parent structure is freed. (CVE-2009-1379)
5250 [Daniel Mentz]
5251
5252 *) Handle non-blocking I/O properly in SSL_shutdown() call.
5253 [Darryl Miles <darryl-mailinglists@netbauds.net>]
5254
5255 *) Add 2.5.4.* OIDs
5256 [Ilya O. <vrghost@gmail.com>]
5257
5258 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
5259
5260 *) Disable renegotiation completely - this fixes a severe security
5261 problem (CVE-2009-3555) at the cost of breaking all
5262 renegotiation. Renegotiation can be re-enabled by setting
5263 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
5264 run-time. This is really not recommended unless you know what
5265 you're doing.
5266 [Ben Laurie]
5267
5268 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
5269
5270 *) Don't set val to NULL when freeing up structures, it is freed up by
5271 underlying code. If sizeof(void *) > sizeof(long) this can result in
5272 zeroing past the valid field. (CVE-2009-0789)
5273 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
5274
5275 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
5276 checked correctly. This would allow some invalid signed attributes to
5277 appear to verify correctly. (CVE-2009-0591)
5278 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5279
5280 *) Reject UniversalString and BMPString types with invalid lengths. This
5281 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
5282 a legal length. (CVE-2009-0590)
5283 [Steve Henson]
5284
5285 *) Set S/MIME signing as the default purpose rather than setting it
5286 unconditionally. This allows applications to override it at the store
5287 level.
5288 [Steve Henson]
5289
5290 *) Permit restricted recursion of ASN1 strings. This is needed in practice
5291 to handle some structures.
5292 [Steve Henson]
5293
5294 *) Improve efficiency of mem_gets: don't search whole buffer each time
5295 for a '\n'
5296 [Jeremy Shapiro <jnshapir@us.ibm.com>]
5297
5298 *) New -hex option for openssl rand.
5299 [Matthieu Herrb]
5300
5301 *) Print out UTF8String and NumericString when parsing ASN1.
5302 [Steve Henson]
5303
5304 *) Support NumericString type for name components.
5305 [Steve Henson]
5306
5307 *) Allow CC in the environment to override the automatically chosen
5308 compiler. Note that nothing is done to ensure flags work with the
5309 chosen compiler.
5310 [Ben Laurie]
5311
5312 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
5313
5314 *) Properly check EVP_VerifyFinal() and similar return values
5315 (CVE-2008-5077).
5316 [Ben Laurie, Bodo Moeller, Google Security Team]
5317
5318 *) Enable TLS extensions by default.
5319 [Ben Laurie]
5320
5321 *) Allow the CHIL engine to be loaded, whether the application is
5322 multithreaded or not. (This does not release the developer from the
5323 obligation to set up the dynamic locking callbacks.)
5324 [Sander Temme <sander@temme.net>]
5325
5326 *) Use correct exit code if there is an error in dgst command.
5327 [Steve Henson; problem pointed out by Roland Dirlewanger]
5328
5329 *) Tweak Configure so that you need to say "experimental-jpake" to enable
5330 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
5331 [Bodo Moeller]
5332
5333 *) Add experimental JPAKE support, including demo authentication in
5334 s_client and s_server.
5335 [Ben Laurie]
5336
5337 *) Set the comparison function in v3_addr_canonize().
5338 [Rob Austein <sra@hactrn.net>]
5339
5340 *) Add support for XMPP STARTTLS in s_client.
5341 [Philip Paeps <philip@freebsd.org>]
5342
5343 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
5344 to ensure that even with this option, only ciphersuites in the
5345 server's preference list will be accepted. (Note that the option
5346 applies only when resuming a session, so the earlier behavior was
5347 just about the algorithm choice for symmetric cryptography.)
5348 [Bodo Moeller]
5349
5350 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
5351
5352 *) Fix NULL pointer dereference if a DTLS server received
5353 ChangeCipherSpec as first record (CVE-2009-1386).
5354 [PR #1679]
5355
5356 *) Fix a state transition in s3_srvr.c and d1_srvr.c
5357 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
5358 [Nagendra Modadugu]
5359
5360 *) The fix in 0.9.8c that supposedly got rid of unsafe
5361 double-checked locking was incomplete for RSA blinding,
5362 addressing just one layer of what turns out to have been
5363 doubly unsafe triple-checked locking.
5364
5365 So now fix this for real by retiring the MONT_HELPER macro
5366 in crypto/rsa/rsa_eay.c.
5367
5368 [Bodo Moeller; problem pointed out by Marius Schilder]
5369
5370 *) Various precautionary measures:
5371
5372 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
5373
5374 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
5375 (NB: This would require knowledge of the secret session ticket key
5376 to exploit, in which case you'd be SOL either way.)
5377
5378 - Change bn_nist.c so that it will properly handle input BIGNUMs
5379 outside the expected range.
5380
5381 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
5382 builds.
5383
5384 [Neel Mehta, Bodo Moeller]
5385
5386 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
5387 the load fails. Useful for distros.
5388 [Ben Laurie and the FreeBSD team]
5389
5390 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
5391 [Steve Henson]
5392
5393 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
5394 [Huang Ying]
5395
5396 *) Expand ENGINE to support engine supplied SSL client certificate functions.
5397
5398 This work was sponsored by Logica.
5399 [Steve Henson]
5400
5401 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
5402 keystores. Support for SSL/TLS client authentication too.
5403 Not compiled unless enable-capieng specified to Configure.
5404
5405 This work was sponsored by Logica.
5406 [Steve Henson]
5407
5408 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
5409 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
5410 attribute creation routines such as certificate requests and PKCS#12
5411 files.
5412 [Steve Henson]
5413
5414 Changes between 0.9.8g and 0.9.8h [28 May 2008]
5415
5416 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
5417 handshake which could lead to a client crash as found using the
5418 Codenomicon TLS test suite (CVE-2008-1672)
5419 [Steve Henson, Mark Cox]
5420
5421 *) Fix double free in TLS server name extensions which could lead to
5422 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
5423 [Joe Orton]
5424
5425 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
5426
5427 Clear the error queue to ensure that error entries left from
5428 older function calls do not interfere with the correct operation.
5429 [Lutz Jaenicke, Erik de Castro Lopo]
5430
5431 *) Remove root CA certificates of commercial CAs:
5432
5433 The OpenSSL project does not recommend any specific CA and does not
5434 have any policy with respect to including or excluding any CA.
5435 Therefore it does not make any sense to ship an arbitrary selection
5436 of root CA certificates with the OpenSSL software.
5437 [Lutz Jaenicke]
5438
5439 *) RSA OAEP patches to fix two separate invalid memory reads.
5440 The first one involves inputs when 'lzero' is greater than
5441 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
5442 before the beginning of from). The second one involves inputs where
5443 the 'db' section contains nothing but zeroes (there is a one-byte
5444 invalid read after the end of 'db').
5445 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5446
5447 *) Partial backport from 0.9.9-dev:
5448
5449 Introduce bn_mul_mont (dedicated Montgomery multiplication
5450 procedure) as a candidate for BIGNUM assembler implementation.
5451 While 0.9.9-dev uses assembler for various architectures, only
5452 x86_64 is available by default here in the 0.9.8 branch, and
5453 32-bit x86 is available through a compile-time setting.
5454
5455 To try the 32-bit x86 assembler implementation, use Configure
5456 option "enable-montasm" (which exists only for this backport).
5457
5458 As "enable-montasm" for 32-bit x86 disclaims code stability
5459 anyway, in this constellation we activate additional code
5460 backported from 0.9.9-dev for further performance improvements,
5461 namely BN_from_montgomery_word. (To enable this otherwise,
5462 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
5463
5464 [Andy Polyakov (backport partially by Bodo Moeller)]
5465
5466 *) Add TLS session ticket callback. This allows an application to set
5467 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
5468 values. This is useful for key rollover for example where several key
5469 sets may exist with different names.
5470 [Steve Henson]
5471
5472 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
5473 This was broken until now in 0.9.8 releases, such that the only way
5474 a registered ENGINE could be used (assuming it initialises
5475 successfully on the host) was to explicitly set it as the default
5476 for the relevant algorithms. This is in contradiction with 0.9.7
5477 behaviour and the documentation. With this fix, when an ENGINE is
5478 registered into a given algorithm's table of implementations, the
5479 'uptodate' flag is reset so that auto-discovery will be used next
5480 time a new context for that algorithm attempts to select an
5481 implementation.
5482 [Ian Lister (tweaked by Geoff Thorpe)]
5483
5484 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5485 implementation in the following ways:
5486
5487 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5488 hard coded.
5489
5490 Lack of BER streaming support means one pass streaming processing is
5491 only supported if data is detached: setting the streaming flag is
5492 ignored for embedded content.
5493
5494 CMS support is disabled by default and must be explicitly enabled
5495 with the enable-cms configuration option.
5496 [Steve Henson]
5497
5498 *) Update the GMP engine glue to do direct copies between BIGNUM and
5499 mpz_t when openssl and GMP use the same limb size. Otherwise the
5500 existing "conversion via a text string export" trick is still used.
5501 [Paul Sheer <paulsheer@gmail.com>]
5502
5503 *) Zlib compression BIO. This is a filter BIO which compressed and
5504 uncompresses any data passed through it.
5505 [Steve Henson]
5506
5507 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5508 RFC3394 compatible AES key wrapping.
5509 [Steve Henson]
5510
5511 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5512 sets string data without copying. X509_ALGOR_set0() and
5513 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5514 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5515 from an X509_ATTRIBUTE structure optionally checking it occurs only
5516 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5517 data.
5518 [Steve Henson]
5519
5520 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5521 to get the expected BN_FLG_CONSTTIME behavior.
5522 [Bodo Moeller (Google)]
5523
5524 *) Netware support:
5525
5526 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5527 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5528 - added some more tests to do_tests.pl
5529 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5530 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5531 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5532 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5533 - various changes to netware.pl to enable gcc-cross builds on Win32
5534 platform
5535 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5536 - various changes to fix missing prototype warnings
5537 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5538 - added AES, WHIRLPOOL and CPUID assembler code to build files
5539 - added missing AES assembler make rules to mk1mf.pl
5540 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5541 [Guenter Knauf <eflash@gmx.net>]
5542
5543 *) Implement certificate status request TLS extension defined in RFC3546.
5544 A client can set the appropriate parameters and receive the encoded
5545 OCSP response via a callback. A server can query the supplied parameters
5546 and set the encoded OCSP response in the callback. Add simplified examples
5547 to s_client and s_server.
5548 [Steve Henson]
5549
5550 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5551
5552 *) Fix various bugs:
5553 + Binary incompatibility of ssl_ctx_st structure
5554 + DTLS interoperation with non-compliant servers
5555 + Don't call get_session_cb() without proposed session
5556 + Fix ia64 assembler code
5557 [Andy Polyakov, Steve Henson]
5558
5559 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5560
5561 *) DTLS Handshake overhaul. There were longstanding issues with
5562 OpenSSL DTLS implementation, which were making it impossible for
5563 RFC 4347 compliant client to communicate with OpenSSL server.
5564 Unfortunately just fixing these incompatibilities would "cut off"
5565 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5566 server keeps tolerating non RFC compliant syntax. The opposite is
5567 not true, 0.9.8f client can not communicate with earlier server.
5568 This update even addresses CVE-2007-4995.
5569 [Andy Polyakov]
5570
5571 *) Changes to avoid need for function casts in OpenSSL: some compilers
5572 (gcc 4.2 and later) reject their use.
5573 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5574 Steve Henson]
5575
5576 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5577 RFC4507bis. The encrypted ticket format is an encrypted encoded
5578 SSL_SESSION structure, that way new session features are automatically
5579 supported.
5580
5581 If a client application caches session in an SSL_SESSION structure
5582 support is transparent because tickets are now stored in the encoded
5583 SSL_SESSION.
5584
5585 The SSL_CTX structure automatically generates keys for ticket
5586 protection in servers so again support should be possible
5587 with no application modification.
5588
5589 If a client or server wishes to disable RFC4507 support then the option
5590 SSL_OP_NO_TICKET can be set.
5591
5592 Add a TLS extension debugging callback to allow the contents of any client
5593 or server extensions to be examined.
5594
5595 This work was sponsored by Google.
5596 [Steve Henson]
5597
5598 *) Add initial support for TLS extensions, specifically for the server_name
5599 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5600 have new members for a host name. The SSL data structure has an
5601 additional member SSL_CTX *initial_ctx so that new sessions can be
5602 stored in that context to allow for session resumption, even after the
5603 SSL has been switched to a new SSL_CTX in reaction to a client's
5604 server_name extension.
5605
5606 New functions (subject to change):
5607
5608 SSL_get_servername()
5609 SSL_get_servername_type()
5610 SSL_set_SSL_CTX()
5611
5612 New CTRL codes and macros (subject to change):
5613
5614 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5615 - SSL_CTX_set_tlsext_servername_callback()
5616 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5617 - SSL_CTX_set_tlsext_servername_arg()
5618 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5619
5620 openssl s_client has a new '-servername ...' option.
5621
5622 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5623 '-key2 ...', '-servername_fatal' (subject to change). This allows
5624 testing the HostName extension for a specific single host name ('-cert'
5625 and '-key' remain fallbacks for handshakes without HostName
5626 negotiation). If the unrecognized_name alert has to be sent, this by
5627 default is a warning; it becomes fatal with the '-servername_fatal'
5628 option.
5629
5630 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
5631
5632 *) Add AES and SSE2 assembly language support to VC++ build.
5633 [Steve Henson]
5634
5635 *) Mitigate attack on final subtraction in Montgomery reduction.
5636 [Andy Polyakov]
5637
5638 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
5639 (which previously caused an internal error).
5640 [Bodo Moeller]
5641
5642 *) Squeeze another 10% out of IGE mode when in != out.
5643 [Ben Laurie]
5644
5645 *) AES IGE mode speedup.
5646 [Dean Gaudet (Google)]
5647
5648 *) Add the Korean symmetric 128-bit cipher SEED (see
5649 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
5650 add SEED ciphersuites from RFC 4162:
5651
5652 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
5653 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
5654 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
5655 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
5656
5657 To minimize changes between patchlevels in the OpenSSL 0.9.8
5658 series, SEED remains excluded from compilation unless OpenSSL
5659 is configured with 'enable-seed'.
5660 [KISA, Bodo Moeller]
5661
5662 *) Mitigate branch prediction attacks, which can be practical if a
5663 single processor is shared, allowing a spy process to extract
5664 information. For detailed background information, see
5665 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
5666 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
5667 and Necessary Software Countermeasures"). The core of the change
5668 are new versions BN_div_no_branch() and
5669 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
5670 respectively, which are slower, but avoid the security-relevant
5671 conditional branches. These are automatically called by BN_div()
5672 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
5673 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
5674 remove a conditional branch.
5675
5676 BN_FLG_CONSTTIME is the new name for the previous
5677 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
5678 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
5679 in the exponent causes BN_mod_exp_mont() to use the alternative
5680 implementation in BN_mod_exp_mont_consttime().) The old name
5681 remains as a deprecated alias.
5682
5683 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
5684 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
5685 constant-time implementations for more than just exponentiation.
5686 Here too the old name is kept as a deprecated alias.
5687
5688 BN_BLINDING_new() will now use BN_dup() for the modulus so that
5689 the BN_BLINDING structure gets an independent copy of the
5690 modulus. This means that the previous "BIGNUM *m" argument to
5691 BN_BLINDING_new() and to BN_BLINDING_create_param() now
5692 essentially becomes "const BIGNUM *m", although we can't actually
5693 change this in the header file before 0.9.9. It allows
5694 RSA_setup_blinding() to use BN_with_flags() on the modulus to
5695 enable BN_FLG_CONSTTIME.
5696
5697 [Matthew D Wood (Intel Corp)]
5698
5699 *) In the SSL/TLS server implementation, be strict about session ID
5700 context matching (which matters if an application uses a single
5701 external cache for different purposes). Previously,
5702 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
5703 set. This did ensure strict client verification, but meant that,
5704 with applications using a single external cache for quite
5705 different requirements, clients could circumvent ciphersuite
5706 restrictions for a given session ID context by starting a session
5707 in a different context.
5708 [Bodo Moeller]
5709
5710 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5711 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5712 authentication-only ciphersuites.
5713 [Bodo Moeller]
5714
5715 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
5716 not complete and could lead to a possible single byte overflow
5717 (CVE-2007-5135) [Ben Laurie]
5718
5719 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
5720
5721 *) Since AES128 and AES256 (and similarly Camellia128 and
5722 Camellia256) share a single mask bit in the logic of
5723 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5724 kludge to work properly if AES128 is available and AES256 isn't
5725 (or if Camellia128 is available and Camellia256 isn't).
5726 [Victor Duchovni]
5727
5728 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
5729 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
5730 When a point or a seed is encoded in a BIT STRING, we need to
5731 prevent the removal of trailing zero bits to get the proper DER
5732 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
5733 of a NamedBitList, for which trailing 0 bits need to be removed.)
5734 [Bodo Moeller]
5735
5736 *) Have SSL/TLS server implementation tolerate "mismatched" record
5737 protocol version while receiving ClientHello even if the
5738 ClientHello is fragmented. (The server can't insist on the
5739 particular protocol version it has chosen before the ServerHello
5740 message has informed the client about his choice.)
5741 [Bodo Moeller]
5742
5743 *) Add RFC 3779 support.
5744 [Rob Austein for ARIN, Ben Laurie]
5745
5746 *) Load error codes if they are not already present instead of using a
5747 static variable. This allows them to be cleanly unloaded and reloaded.
5748 Improve header file function name parsing.
5749 [Steve Henson]
5750
5751 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
5752 or CAPABILITY handshake as required by RFCs.
5753 [Goetz Babin-Ebell]
5754
5755 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5756
5757 *) Introduce limits to prevent malicious keys being able to
5758 cause a denial of service. (CVE-2006-2940)
5759 [Steve Henson, Bodo Moeller]
5760
5761 *) Fix ASN.1 parsing of certain invalid structures that can result
5762 in a denial of service. (CVE-2006-2937) [Steve Henson]
5763
5764 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5765 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5766
5767 *) Fix SSL client code which could crash if connecting to a
5768 malicious SSLv2 server. (CVE-2006-4343)
5769 [Tavis Ormandy and Will Drewry, Google Security Team]
5770
5771 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
5772 match only those. Before that, "AES256-SHA" would be interpreted
5773 as a pattern and match "AES128-SHA" too (since AES128-SHA got
5774 the same strength classification in 0.9.7h) as we currently only
5775 have a single AES bit in the ciphersuite description bitmap.
5776 That change, however, also applied to ciphersuite strings such as
5777 "RC4-MD5" that intentionally matched multiple ciphersuites --
5778 namely, SSL 2.0 ciphersuites in addition to the more common ones
5779 from SSL 3.0/TLS 1.0.
5780
5781 So we change the selection algorithm again: Naming an explicit
5782 ciphersuite selects this one ciphersuite, and any other similar
5783 ciphersuite (same bitmap) from *other* protocol versions.
5784 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5785 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5786
5787 Since SSL 2.0 does not have any ciphersuites for which the
5788 128/256 bit distinction would be relevant, this works for now.
5789 The proper fix will be to use different bits for AES128 and
5790 AES256, which would have avoided the problems from the beginning;
5791 however, bits are scarce, so we can only do this in a new release
5792 (not just a patchlevel) when we can change the SSL_CIPHER
5793 definition to split the single 'unsigned long mask' bitmap into
5794 multiple values to extend the available space.
5795
5796 [Bodo Moeller]
5797
5798 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5799
5800 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5801 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5802
5803 *) Add AES IGE and biIGE modes.
5804 [Ben Laurie]
5805
5806 *) Change the Unix randomness entropy gathering to use poll() when
5807 possible instead of select(), since the latter has some
5808 undesirable limitations.
5809 [Darryl Miles via Richard Levitte and Bodo Moeller]
5810
5811 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
5812 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
5813 cannot be implicitly activated as part of, e.g., the "AES" alias.
5814 However, please upgrade to OpenSSL 0.9.9[-dev] for
5815 non-experimental use of the ECC ciphersuites to get TLS extension
5816 support, which is required for curve and point format negotiation
5817 to avoid potential handshake problems.
5818 [Bodo Moeller]
5819
5820 *) Disable rogue ciphersuites:
5821
5822 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5823 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5824 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5825
5826 The latter two were purportedly from
5827 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5828 appear there.
5829
5830 Also deactivate the remaining ciphersuites from
5831 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5832 unofficial, and the ID has long expired.
5833 [Bodo Moeller]
5834
5835 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5836 dual-core machines) and other potential thread-safety issues.
5837 [Bodo Moeller]
5838
5839 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
5840 versions), which is now available for royalty-free use
5841 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
5842 Also, add Camellia TLS ciphersuites from RFC 4132.
5843
5844 To minimize changes between patchlevels in the OpenSSL 0.9.8
5845 series, Camellia remains excluded from compilation unless OpenSSL
5846 is configured with 'enable-camellia'.
5847 [NTT]
5848
5849 *) Disable the padding bug check when compression is in use. The padding
5850 bug check assumes the first packet is of even length, this is not
5851 necessarily true if compression is enabled and can result in false
5852 positives causing handshake failure. The actual bug test is ancient
5853 code so it is hoped that implementations will either have fixed it by
5854 now or any which still have the bug do not support compression.
5855 [Steve Henson]
5856
5857 Changes between 0.9.8a and 0.9.8b [04 May 2006]
5858
5859 *) When applying a cipher rule check to see if string match is an explicit
5860 cipher suite and only match that one cipher suite if it is.
5861 [Steve Henson]
5862
5863 *) Link in manifests for VC++ if needed.
5864 [Austin Ziegler <halostatue@gmail.com>]
5865
5866 *) Update support for ECC-based TLS ciphersuites according to
5867 draft-ietf-tls-ecc-12.txt with proposed changes (but without
5868 TLS extensions, which are supported starting with the 0.9.9
5869 branch, not in the OpenSSL 0.9.8 branch).
5870 [Douglas Stebila]
5871
5872 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
5873 opaque EVP_CIPHER_CTX handling.
5874 [Steve Henson]
5875
5876 *) Fixes and enhancements to zlib compression code. We now only use
5877 "zlib1.dll" and use the default __cdecl calling convention on Win32
5878 to conform with the standards mentioned here:
5879 http://www.zlib.net/DLL_FAQ.txt
5880 Static zlib linking now works on Windows and the new --with-zlib-include
5881 --with-zlib-lib options to Configure can be used to supply the location
5882 of the headers and library. Gracefully handle case where zlib library
5883 can't be loaded.
5884 [Steve Henson]
5885
5886 *) Several fixes and enhancements to the OID generation code. The old code
5887 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
5888 handle numbers larger than ULONG_MAX, truncated printing and had a
5889 non standard OBJ_obj2txt() behaviour.
5890 [Steve Henson]
5891
5892 *) Add support for building of engines under engine/ as shared libraries
5893 under VC++ build system.
5894 [Steve Henson]
5895
5896 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
5897 Hopefully, we will not see any false combination of paths any more.
5898 [Richard Levitte]
5899
5900 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5901
5902 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5903 (part of SSL_OP_ALL). This option used to disable the
5904 countermeasure against man-in-the-middle protocol-version
5905 rollback in the SSL 2.0 server implementation, which is a bad
5906 idea. (CVE-2005-2969)
5907
5908 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5909 for Information Security, National Institute of Advanced Industrial
5910 Science and Technology [AIST], Japan)]
5911
5912 *) Add two function to clear and return the verify parameter flags.
5913 [Steve Henson]
5914
5915 *) Keep cipherlists sorted in the source instead of sorting them at
5916 runtime, thus removing the need for a lock.
5917 [Nils Larsch]
5918
5919 *) Avoid some small subgroup attacks in Diffie-Hellman.
5920 [Nick Mathewson and Ben Laurie]
5921
5922 *) Add functions for well-known primes.
5923 [Nick Mathewson]
5924
5925 *) Extended Windows CE support.
5926 [Satoshi Nakamura and Andy Polyakov]
5927
5928 *) Initialize SSL_METHOD structures at compile time instead of during
5929 runtime, thus removing the need for a lock.
5930 [Steve Henson]
5931
5932 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5933 attempting to decrypt each encrypted key in turn. Add support to
5934 smime utility.
5935 [Steve Henson]
5936
5937 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5938
5939 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5940 OpenSSL 0.9.8.]
5941
5942 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5943 [Richard Levitte]
5944
5945 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5946 key into the same file any more.
5947 [Richard Levitte]
5948
5949 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5950 [Andy Polyakov]
5951
5952 *) Add -utf8 command line and config file option to 'ca'.
5953 [Stefan <stf@udoma.org]
5954
5955 *) Removed the macro des_crypt(), as it seems to conflict with some
5956 libraries. Use DES_crypt().
5957 [Richard Levitte]
5958
5959 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5960 involves renaming the source and generated shared-libs for
5961 both. The engines will accept the corrected or legacy ids
5962 ('ncipher' and '4758_cca' respectively) when binding. NB,
5963 this only applies when building 'shared'.
5964 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5965
5966 *) Add attribute functions to EVP_PKEY structure. Modify
5967 PKCS12_create() to recognize a CSP name attribute and
5968 use it. Make -CSP option work again in pkcs12 utility.
5969 [Steve Henson]
5970
5971 *) Add new functionality to the bn blinding code:
5972 - automatic re-creation of the BN_BLINDING parameters after
5973 a fixed number of uses (currently 32)
5974 - add new function for parameter creation
5975 - introduce flags to control the update behaviour of the
5976 BN_BLINDING parameters
5977 - hide BN_BLINDING structure
5978 Add a second BN_BLINDING slot to the RSA structure to improve
5979 performance when a single RSA object is shared among several
5980 threads.
5981 [Nils Larsch]
5982
5983 *) Add support for DTLS.
5984 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5985
5986 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5987 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5988 [Walter Goulet]
5989
5990 *) Remove buggy and incomplete DH cert support from
5991 ssl/ssl_rsa.c and ssl/s3_both.c
5992 [Nils Larsch]
5993
5994 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5995 the apps/openssl applications.
5996 [Nils Larsch]
5997
5998 *) Compile clean with "-Wall -Wmissing-prototypes
5999 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
6000 DEBUG_SAFESTACK must also be set.
6001 [Ben Laurie]
6002
6003 *) Change ./Configure so that certain algorithms can be disabled by default.
6004 The new counterpiece to "no-xxx" is "enable-xxx".
6005
6006 The patented RC5 and MDC2 algorithms will now be disabled unless
6007 "enable-rc5" and "enable-mdc2", respectively, are specified.
6008
6009 (IDEA remains enabled despite being patented. This is because IDEA
6010 is frequently required for interoperability, and there is no license
6011 fee for non-commercial use. As before, "no-idea" can be used to
6012 avoid this algorithm.)
6013
6014 [Bodo Moeller]
6015
6016 *) Add processing of proxy certificates (see RFC 3820). This work was
6017 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
6018 EGEE (Enabling Grids for E-science in Europe).
6019 [Richard Levitte]
6020
6021 *) RC4 performance overhaul on modern architectures/implementations, such
6022 as Intel P4, IA-64 and AMD64.
6023 [Andy Polyakov]
6024
6025 *) New utility extract-section.pl. This can be used specify an alternative
6026 section number in a pod file instead of having to treat each file as
6027 a separate case in Makefile. This can be done by adding two lines to the
6028 pod file:
6029
6030 =for comment openssl_section:XXX
6031
6032 The blank line is mandatory.
6033
6034 [Steve Henson]
6035
6036 *) New arguments -certform, -keyform and -pass for s_client and s_server
6037 to allow alternative format key and certificate files and passphrase
6038 sources.
6039 [Steve Henson]
6040
6041 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
6042 update associated structures and add various utility functions.
6043
6044 Add new policy related verify parameters, include policy checking in
6045 standard verify code. Enhance 'smime' application with extra parameters
6046 to support policy checking and print out.
6047 [Steve Henson]
6048
6049 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
6050 Nehemiah processors. These extensions support AES encryption in hardware
6051 as well as RNG (though RNG support is currently disabled).
6052 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
6053
6054 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
6055 [Geoff Thorpe]
6056
6057 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
6058 [Andy Polyakov and a number of other people]
6059
6060 *) Improved PowerPC platform support. Most notably BIGNUM assembler
6061 implementation contributed by IBM.
6062 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
6063
6064 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
6065 exponent rather than 'unsigned long'. There is a corresponding change to
6066 the new 'rsa_keygen' element of the RSA_METHOD structure.
6067 [Jelte Jansen, Geoff Thorpe]
6068
6069 *) Functionality for creating the initial serial number file is now
6070 moved from CA.pl to the 'ca' utility with a new option -create_serial.
6071
6072 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
6073 number file to 1, which is bound to cause problems. To avoid
6074 the problems while respecting compatibility between different 0.9.7
6075 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
6076 CA.pl for serial number initialization. With the new release 0.9.8,
6077 we can fix the problem directly in the 'ca' utility.)
6078 [Steve Henson]
6079
6080 *) Reduced header interdependencies by declaring more opaque objects in
6081 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
6082 give fewer recursive includes, which could break lazy source code - so
6083 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
6084 developers should define this symbol when building and using openssl to
6085 ensure they track the recommended behaviour, interfaces, [etc], but
6086 backwards-compatible behaviour prevails when this isn't defined.
6087 [Geoff Thorpe]
6088
6089 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
6090 [Steve Henson]
6091
6092 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
6093 This will generate a random key of the appropriate length based on the
6094 cipher context. The EVP_CIPHER can provide its own random key generation
6095 routine to support keys of a specific form. This is used in the des and
6096 3des routines to generate a key of the correct parity. Update S/MIME
6097 code to use new functions and hence generate correct parity DES keys.
6098 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
6099 valid (weak or incorrect parity).
6100 [Steve Henson]
6101
6102 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
6103 as looking them up. This is useful when the verified structure may contain
6104 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
6105 present unless the new PKCS7_NO_CRL flag is asserted.
6106 [Steve Henson]
6107
6108 *) Extend ASN1 oid configuration module. It now additionally accepts the
6109 syntax:
6110
6111 shortName = some long name, 1.2.3.4
6112 [Steve Henson]
6113
6114 *) Reimplemented the BN_CTX implementation. There is now no more static
6115 limitation on the number of variables it can handle nor the depth of the
6116 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
6117 information can now expand as required, and rather than having a single
6118 static array of bignums, BN_CTX now uses a linked-list of such arrays
6119 allowing it to expand on demand whilst maintaining the usefulness of
6120 BN_CTX's "bundling".
6121 [Geoff Thorpe]
6122
6123 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
6124 to allow all RSA operations to function using a single BN_CTX.
6125 [Geoff Thorpe]
6126
6127 *) Preliminary support for certificate policy evaluation and checking. This
6128 is initially intended to pass the tests outlined in "Conformance Testing
6129 of Relying Party Client Certificate Path Processing Logic" v1.07.
6130 [Steve Henson]
6131
6132 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
6133 remained unused and not that useful. A variety of other little bignum
6134 tweaks and fixes have also been made continuing on from the audit (see
6135 below).
6136 [Geoff Thorpe]
6137
6138 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
6139 associated ASN1, EVP and SSL functions and old ASN1 macros.
6140 [Richard Levitte]
6141
6142 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
6143 and this should never fail. So the return value from the use of
6144 BN_set_word() (which can fail due to needless expansion) is now deprecated;
6145 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
6146 [Geoff Thorpe]
6147
6148 *) BN_CTX_get() should return zero-valued bignums, providing the same
6149 initialised value as BN_new().
6150 [Geoff Thorpe, suggested by Ulf Möller]
6151
6152 *) Support for inhibitAnyPolicy certificate extension.
6153 [Steve Henson]
6154
6155 *) An audit of the BIGNUM code is underway, for which debugging code is
6156 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
6157 is considered valid when processing BIGNUMs, and causes execution to
6158 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
6159 further steps are taken to deliberately pollute unused data in BIGNUM
6160 structures to try and expose faulty code further on. For now, openssl will
6161 (in its default mode of operation) continue to tolerate the inconsistent
6162 forms that it has tolerated in the past, but authors and packagers should
6163 consider trying openssl and their own applications when compiled with
6164 these debugging symbols defined. It will help highlight potential bugs in
6165 their own code, and will improve the test coverage for OpenSSL itself. At
6166 some point, these tighter rules will become openssl's default to improve
6167 maintainability, though the assert()s and other overheads will remain only
6168 in debugging configurations. See bn.h for more details.
6169 [Geoff Thorpe, Nils Larsch, Ulf Möller]
6170
6171 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
6172 that can only be obtained through BN_CTX_new() (which implicitly
6173 initialises it). The presence of this function only made it possible
6174 to overwrite an existing structure (and cause memory leaks).
6175 [Geoff Thorpe]
6176
6177 *) Because of the callback-based approach for implementing LHASH as a
6178 template type, lh_insert() adds opaque objects to hash-tables and
6179 lh_doall() or lh_doall_arg() are typically used with a destructor callback
6180 to clean up those corresponding objects before destroying the hash table
6181 (and losing the object pointers). So some over-zealous constifications in
6182 LHASH have been relaxed so that lh_insert() does not take (nor store) the
6183 objects as "const" and the lh_doall[_arg] callback wrappers are not
6184 prototyped to have "const" restrictions on the object pointers they are
6185 given (and so aren't required to cast them away any more).
6186 [Geoff Thorpe]
6187
6188 *) The tmdiff.h API was so ugly and minimal that our own timing utility
6189 (speed) prefers to use its own implementation. The two implementations
6190 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
6191 its object type properly exposed (MS_TM) instead of casting to/from "char
6192 *". This may still change yet if someone realises MS_TM and "ms_time_***"
6193 aren't necessarily the greatest nomenclatures - but this is what was used
6194 internally to the implementation so I've used that for now.
6195 [Geoff Thorpe]
6196
6197 *) Ensure that deprecated functions do not get compiled when
6198 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
6199 the self-tests were still using deprecated key-generation functions so
6200 these have been updated also.
6201 [Geoff Thorpe]
6202
6203 *) Reorganise PKCS#7 code to separate the digest location functionality
6204 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
6205 New function PKCS7_set_digest() to set the digest type for PKCS#7
6206 digestedData type. Add additional code to correctly generate the
6207 digestedData type and add support for this type in PKCS7 initialization
6208 functions.
6209 [Steve Henson]
6210
6211 *) New function PKCS7_set0_type_other() this initializes a PKCS7
6212 structure of type "other".
6213 [Steve Henson]
6214
6215 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
6216 sure the loop does correctly stop and breaking ("division by zero")
6217 modulus operations are not performed. The (pre-generated) prime
6218 table crypto/bn/bn_prime.h was already correct, but it could not be
6219 re-generated on some platforms because of the "division by zero"
6220 situation in the script.
6221 [Ralf S. Engelschall]
6222
6223 *) Update support for ECC-based TLS ciphersuites according to
6224 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
6225 SHA-1 now is only used for "small" curves (where the
6226 representation of a field element takes up to 24 bytes); for
6227 larger curves, the field element resulting from ECDH is directly
6228 used as premaster secret.
6229 [Douglas Stebila (Sun Microsystems Laboratories)]
6230
6231 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
6232 curve secp160r1 to the tests.
6233 [Douglas Stebila (Sun Microsystems Laboratories)]
6234
6235 *) Add the possibility to load symbols globally with DSO.
6236 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
6237
6238 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
6239 control of the error stack.
6240 [Richard Levitte]
6241
6242 *) Add support for STORE in ENGINE.
6243 [Richard Levitte]
6244
6245 *) Add the STORE type. The intention is to provide a common interface
6246 to certificate and key stores, be they simple file-based stores, or
6247 HSM-type store, or LDAP stores, or...
6248 NOTE: The code is currently UNTESTED and isn't really used anywhere.
6249 [Richard Levitte]
6250
6251 *) Add a generic structure called OPENSSL_ITEM. This can be used to
6252 pass a list of arguments to any function as well as provide a way
6253 for a function to pass data back to the caller.
6254 [Richard Levitte]
6255
6256 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
6257 works like BUF_strdup() but can be used to duplicate a portion of
6258 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
6259 a memory area.
6260 [Richard Levitte]
6261
6262 *) Add the function sk_find_ex() which works like sk_find(), but will
6263 return an index to an element even if an exact match couldn't be
6264 found. The index is guaranteed to point at the element where the
6265 searched-for key would be inserted to preserve sorting order.
6266 [Richard Levitte]
6267
6268 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
6269 takes an extra flags argument for optional functionality. Currently,
6270 the following flags are defined:
6271
6272 OBJ_BSEARCH_VALUE_ON_NOMATCH
6273 This one gets OBJ_bsearch_ex() to return a pointer to the first
6274 element where the comparing function returns a negative or zero
6275 number.
6276
6277 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
6278 This one gets OBJ_bsearch_ex() to return a pointer to the first
6279 element where the comparing function returns zero. This is useful
6280 if there are more than one element where the comparing function
6281 returns zero.
6282 [Richard Levitte]
6283
6284 *) Make it possible to create self-signed certificates with 'openssl ca'
6285 in such a way that the self-signed certificate becomes part of the
6286 CA database and uses the same mechanisms for serial number generation
6287 as all other certificate signing. The new flag '-selfsign' enables
6288 this functionality. Adapt CA.sh and CA.pl.in.
6289 [Richard Levitte]
6290
6291 *) Add functionality to check the public key of a certificate request
6292 against a given private. This is useful to check that a certificate
6293 request can be signed by that key (self-signing).
6294 [Richard Levitte]
6295
6296 *) Make it possible to have multiple active certificates with the same
6297 subject in the CA index file. This is done only if the keyword
6298 'unique_subject' is set to 'no' in the main CA section (default
6299 if 'CA_default') of the configuration file. The value is saved
6300 with the database itself in a separate index attribute file,
6301 named like the index file with '.attr' appended to the name.
6302 [Richard Levitte]
6303
6304 *) Generate multi-valued AVAs using '+' notation in config files for
6305 req and dirName.
6306 [Steve Henson]
6307
6308 *) Support for nameConstraints certificate extension.
6309 [Steve Henson]
6310
6311 *) Support for policyConstraints certificate extension.
6312 [Steve Henson]
6313
6314 *) Support for policyMappings certificate extension.
6315 [Steve Henson]
6316
6317 *) Make sure the default DSA_METHOD implementation only uses its
6318 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
6319 and change its own handlers to be NULL so as to remove unnecessary
6320 indirection. This lets alternative implementations fallback to the
6321 default implementation more easily.
6322 [Geoff Thorpe]
6323
6324 *) Support for directoryName in GeneralName related extensions
6325 in config files.
6326 [Steve Henson]
6327
6328 *) Make it possible to link applications using Makefile.shared.
6329 Make that possible even when linking against static libraries!
6330 [Richard Levitte]
6331
6332 *) Support for single pass processing for S/MIME signing. This now
6333 means that S/MIME signing can be done from a pipe, in addition
6334 cleartext signing (multipart/signed type) is effectively streaming
6335 and the signed data does not need to be all held in memory.
6336
6337 This is done with a new flag PKCS7_STREAM. When this flag is set
6338 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
6339 is done after the data is output (and digests calculated) in
6340 SMIME_write_PKCS7().
6341 [Steve Henson]
6342
6343 *) Add full support for -rpath/-R, both in shared libraries and
6344 applications, at least on the platforms where it's known how
6345 to do it.
6346 [Richard Levitte]
6347
6348 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
6349 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
6350 will now compute a table of multiples of the generator that
6351 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
6352 faster (notably in the case of a single point multiplication,
6353 scalar * generator).
6354 [Nils Larsch, Bodo Moeller]
6355
6356 *) IPv6 support for certificate extensions. The various extensions
6357 which use the IP:a.b.c.d can now take IPv6 addresses using the
6358 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
6359 correctly.
6360 [Steve Henson]
6361
6362 *) Added an ENGINE that implements RSA by performing private key
6363 exponentiations with the GMP library. The conversions to and from
6364 GMP's mpz_t format aren't optimised nor are any montgomery forms
6365 cached, and on x86 it appears OpenSSL's own performance has caught up.
6366 However there are likely to be other architectures where GMP could
6367 provide a boost. This ENGINE is not built in by default, but it can be
6368 specified at Configure time and should be accompanied by the necessary
6369 linker additions, eg;
6370 ./config -DOPENSSL_USE_GMP -lgmp
6371 [Geoff Thorpe]
6372
6373 *) "openssl engine" will not display ENGINE/DSO load failure errors when
6374 testing availability of engines with "-t" - the old behaviour is
6375 produced by increasing the feature's verbosity with "-tt".
6376 [Geoff Thorpe]
6377
6378 *) ECDSA routines: under certain error conditions uninitialized BN objects
6379 could be freed. Solution: make sure initialization is performed early
6380 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
6381 via PR#459)
6382 [Lutz Jaenicke]
6383
6384 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
6385 and DH_METHOD (eg. by ENGINE implementations) to override the normal
6386 software implementations. For DSA and DH, parameter generation can
6387 also be overridden by providing the appropriate method callbacks.
6388 [Geoff Thorpe]
6389
6390 *) Change the "progress" mechanism used in key-generation and
6391 primality testing to functions that take a new BN_GENCB pointer in
6392 place of callback/argument pairs. The new API functions have "_ex"
6393 postfixes and the older functions are reimplemented as wrappers for
6394 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
6395 declarations of the old functions to help (graceful) attempts to
6396 migrate to the new functions. Also, the new key-generation API
6397 functions operate on a caller-supplied key-structure and return
6398 success/failure rather than returning a key or NULL - this is to
6399 help make "keygen" another member function of RSA_METHOD etc.
6400
6401 Example for using the new callback interface:
6402
6403 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
6404 void *my_arg = ...;
6405 BN_GENCB my_cb;
6406
6407 BN_GENCB_set(&my_cb, my_callback, my_arg);
6408
6409 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
6410 /* For the meaning of a, b in calls to my_callback(), see the
6411 * documentation of the function that calls the callback.
6412 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
6413 * my_callback should return 1 if it wants BN_is_prime_ex()
6414 * to continue, or 0 to stop.
6415 */
6416
6417 [Geoff Thorpe]
6418
6419 *) Change the ZLIB compression method to be stateful, and make it
6420 available to TLS with the number defined in
6421 draft-ietf-tls-compression-04.txt.
6422 [Richard Levitte]
6423
6424 *) Add the ASN.1 structures and functions for CertificatePair, which
6425 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
6426
6427 CertificatePair ::= SEQUENCE {
6428 forward [0] Certificate OPTIONAL,
6429 reverse [1] Certificate OPTIONAL,
6430 -- at least one of the pair shall be present -- }
6431
6432 Also implement the PEM functions to read and write certificate
6433 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
6434
6435 This needed to be defined, mostly for the sake of the LDAP
6436 attribute crossCertificatePair, but may prove useful elsewhere as
6437 well.
6438 [Richard Levitte]
6439
6440 *) Make it possible to inhibit symlinking of shared libraries in
6441 Makefile.shared, for Cygwin's sake.
6442 [Richard Levitte]
6443
6444 *) Extend the BIGNUM API by creating a function
6445 void BN_set_negative(BIGNUM *a, int neg);
6446 and a macro that behave like
6447 int BN_is_negative(const BIGNUM *a);
6448
6449 to avoid the need to access 'a->neg' directly in applications.
6450 [Nils Larsch]
6451
6452 *) Implement fast modular reduction for pseudo-Mersenne primes
6453 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
6454 EC_GROUP_new_curve_GFp() will now automatically use this
6455 if applicable.
6456 [Nils Larsch <nla@trustcenter.de>]
6457
6458 *) Add new lock type (CRYPTO_LOCK_BN).
6459 [Bodo Moeller]
6460
6461 *) Change the ENGINE framework to automatically load engines
6462 dynamically from specific directories unless they could be
6463 found to already be built in or loaded. Move all the
6464 current engines except for the cryptodev one to a new
6465 directory engines/.
6466 The engines in engines/ are built as shared libraries if
6467 the "shared" options was given to ./Configure or ./config.
6468 Otherwise, they are inserted in libcrypto.a.
6469 /usr/local/ssl/engines is the default directory for dynamic
6470 engines, but that can be overridden at configure time through
6471 the usual use of --prefix and/or --openssldir, and at run
6472 time with the environment variable OPENSSL_ENGINES.
6473 [Geoff Thorpe and Richard Levitte]
6474
6475 *) Add Makefile.shared, a helper makefile to build shared
6476 libraries. Adapt Makefile.org.
6477 [Richard Levitte]
6478
6479 *) Add version info to Win32 DLLs.
6480 [Peter 'Luna' Runestig" <peter@runestig.com>]
6481
6482 *) Add new 'medium level' PKCS#12 API. Certificates and keys
6483 can be added using this API to created arbitrary PKCS#12
6484 files while avoiding the low level API.
6485
6486 New options to PKCS12_create(), key or cert can be NULL and
6487 will then be omitted from the output file. The encryption
6488 algorithm NIDs can be set to -1 for no encryption, the mac
6489 iteration count can be set to 0 to omit the mac.
6490
6491 Enhance pkcs12 utility by making the -nokeys and -nocerts
6492 options work when creating a PKCS#12 file. New option -nomac
6493 to omit the mac, NONE can be set for an encryption algorithm.
6494 New code is modified to use the enhanced PKCS12_create()
6495 instead of the low level API.
6496 [Steve Henson]
6497
6498 *) Extend ASN1 encoder to support indefinite length constructed
6499 encoding. This can output sequences tags and octet strings in
6500 this form. Modify pk7_asn1.c to support indefinite length
6501 encoding. This is experimental and needs additional code to
6502 be useful, such as an ASN1 bio and some enhanced streaming
6503 PKCS#7 code.
6504
6505 Extend template encode functionality so that tagging is passed
6506 down to the template encoder.
6507 [Steve Henson]
6508
6509 *) Let 'openssl req' fail if an argument to '-newkey' is not
6510 recognized instead of using RSA as a default.
6511 [Bodo Moeller]
6512
6513 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6514 As these are not official, they are not included in "ALL";
6515 the "ECCdraft" ciphersuite group alias can be used to select them.
6516 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6517
6518 *) Add ECDH engine support.
6519 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6520
6521 *) Add ECDH in new directory crypto/ecdh/.
6522 [Douglas Stebila (Sun Microsystems Laboratories)]
6523
6524 *) Let BN_rand_range() abort with an error after 100 iterations
6525 without success (which indicates a broken PRNG).
6526 [Bodo Moeller]
6527
6528 *) Change BN_mod_sqrt() so that it verifies that the input value
6529 is really the square of the return value. (Previously,
6530 BN_mod_sqrt would show GIGO behaviour.)
6531 [Bodo Moeller]
6532
6533 *) Add named elliptic curves over binary fields from X9.62, SECG,
6534 and WAP/WTLS; add OIDs that were still missing.
6535
6536 [Sheueling Chang Shantz and Douglas Stebila
6537 (Sun Microsystems Laboratories)]
6538
6539 *) Extend the EC library for elliptic curves over binary fields
6540 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6541 New EC_METHOD:
6542
6543 EC_GF2m_simple_method
6544
6545 New API functions:
6546
6547 EC_GROUP_new_curve_GF2m
6548 EC_GROUP_set_curve_GF2m
6549 EC_GROUP_get_curve_GF2m
6550 EC_POINT_set_affine_coordinates_GF2m
6551 EC_POINT_get_affine_coordinates_GF2m
6552 EC_POINT_set_compressed_coordinates_GF2m
6553
6554 Point compression for binary fields is disabled by default for
6555 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6556 enable it).
6557
6558 As binary polynomials are represented as BIGNUMs, various members
6559 of the EC_GROUP and EC_POINT data structures can be shared
6560 between the implementations for prime fields and binary fields;
6561 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6562 are essentially identical to their ..._GFp counterparts.
6563 (For simplicity, the '..._GFp' prefix has been dropped from
6564 various internal method names.)
6565
6566 An internal 'field_div' method (similar to 'field_mul' and
6567 'field_sqr') has been added; this is used only for binary fields.
6568
6569 [Sheueling Chang Shantz and Douglas Stebila
6570 (Sun Microsystems Laboratories)]
6571
6572 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6573 through methods ('mul', 'precompute_mult').
6574
6575 The generic implementations (now internally called 'ec_wNAF_mul'
6576 and 'ec_wNAF_precomputed_mult') remain the default if these
6577 methods are undefined.
6578
6579 [Sheueling Chang Shantz and Douglas Stebila
6580 (Sun Microsystems Laboratories)]
6581
6582 *) New function EC_GROUP_get_degree, which is defined through
6583 EC_METHOD. For curves over prime fields, this returns the bit
6584 length of the modulus.
6585
6586 [Sheueling Chang Shantz and Douglas Stebila
6587 (Sun Microsystems Laboratories)]
6588
6589 *) New functions EC_GROUP_dup, EC_POINT_dup.
6590 (These simply call ..._new and ..._copy).
6591
6592 [Sheueling Chang Shantz and Douglas Stebila
6593 (Sun Microsystems Laboratories)]
6594
6595 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6596 Polynomials are represented as BIGNUMs (where the sign bit is not
6597 used) in the following functions [macros]:
6598
6599 BN_GF2m_add
6600 BN_GF2m_sub [= BN_GF2m_add]
6601 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6602 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6603 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6604 BN_GF2m_mod_inv
6605 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6606 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6607 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6608 BN_GF2m_cmp [= BN_ucmp]
6609
6610 (Note that only the 'mod' functions are actually for fields GF(2^m).
6611 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6612
6613 For some functions, an the irreducible polynomial defining a
6614 field can be given as an 'unsigned int[]' with strictly
6615 decreasing elements giving the indices of those bits that are set;
6616 i.e., p[] represents the polynomial
6617 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6618 where
6619 p[0] > p[1] > ... > p[k] = 0.
6620 This applies to the following functions:
6621
6622 BN_GF2m_mod_arr
6623 BN_GF2m_mod_mul_arr
6624 BN_GF2m_mod_sqr_arr
6625 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
6626 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
6627 BN_GF2m_mod_exp_arr
6628 BN_GF2m_mod_sqrt_arr
6629 BN_GF2m_mod_solve_quad_arr
6630 BN_GF2m_poly2arr
6631 BN_GF2m_arr2poly
6632
6633 Conversion can be performed by the following functions:
6634
6635 BN_GF2m_poly2arr
6636 BN_GF2m_arr2poly
6637
6638 bntest.c has additional tests for binary polynomial arithmetic.
6639
6640 Two implementations for BN_GF2m_mod_div() are available.
6641 The default algorithm simply uses BN_GF2m_mod_inv() and
6642 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
6643 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
6644 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
6645
6646 [Sheueling Chang Shantz and Douglas Stebila
6647 (Sun Microsystems Laboratories)]
6648
6649 *) Add new error code 'ERR_R_DISABLED' that can be used when some
6650 functionality is disabled at compile-time.
6651 [Douglas Stebila <douglas.stebila@sun.com>]
6652
6653 *) Change default behaviour of 'openssl asn1parse' so that more
6654 information is visible when viewing, e.g., a certificate:
6655
6656 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
6657 mode the content of non-printable OCTET STRINGs is output in a
6658 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
6659 avoid the appearance of a printable string.
6660 [Nils Larsch <nla@trustcenter.de>]
6661
6662 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
6663 functions
6664 EC_GROUP_set_asn1_flag()
6665 EC_GROUP_get_asn1_flag()
6666 EC_GROUP_set_point_conversion_form()
6667 EC_GROUP_get_point_conversion_form()
6668 These control ASN1 encoding details:
6669 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
6670 has been set to OPENSSL_EC_NAMED_CURVE.
6671 - Points are encoded in uncompressed form by default; options for
6672 asn1_for are as for point2oct, namely
6673 POINT_CONVERSION_COMPRESSED
6674 POINT_CONVERSION_UNCOMPRESSED
6675 POINT_CONVERSION_HYBRID
6676
6677 Also add 'seed' and 'seed_len' members to EC_GROUP with access
6678 functions
6679 EC_GROUP_set_seed()
6680 EC_GROUP_get0_seed()
6681 EC_GROUP_get_seed_len()
6682 This is used only for ASN1 purposes (so far).
6683 [Nils Larsch <nla@trustcenter.de>]
6684
6685 *) Add 'field_type' member to EC_METHOD, which holds the NID
6686 of the appropriate field type OID. The new function
6687 EC_METHOD_get_field_type() returns this value.
6688 [Nils Larsch <nla@trustcenter.de>]
6689
6690 *) Add functions
6691 EC_POINT_point2bn()
6692 EC_POINT_bn2point()
6693 EC_POINT_point2hex()
6694 EC_POINT_hex2point()
6695 providing useful interfaces to EC_POINT_point2oct() and
6696 EC_POINT_oct2point().
6697 [Nils Larsch <nla@trustcenter.de>]
6698
6699 *) Change internals of the EC library so that the functions
6700 EC_GROUP_set_generator()
6701 EC_GROUP_get_generator()
6702 EC_GROUP_get_order()
6703 EC_GROUP_get_cofactor()
6704 are implemented directly in crypto/ec/ec_lib.c and not dispatched
6705 to methods, which would lead to unnecessary code duplication when
6706 adding different types of curves.
6707 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
6708
6709 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
6710 arithmetic, and such that modified wNAFs are generated
6711 (which avoid length expansion in many cases).
6712 [Bodo Moeller]
6713
6714 *) Add a function EC_GROUP_check_discriminant() (defined via
6715 EC_METHOD) that verifies that the curve discriminant is non-zero.
6716
6717 Add a function EC_GROUP_check() that makes some sanity tests
6718 on a EC_GROUP, its generator and order. This includes
6719 EC_GROUP_check_discriminant().
6720 [Nils Larsch <nla@trustcenter.de>]
6721
6722 *) Add ECDSA in new directory crypto/ecdsa/.
6723
6724 Add applications 'openssl ecparam' and 'openssl ecdsa'
6725 (these are based on 'openssl dsaparam' and 'openssl dsa').
6726
6727 ECDSA support is also included in various other files across the
6728 library. Most notably,
6729 - 'openssl req' now has a '-newkey ecdsa:file' option;
6730 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
6731 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
6732 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
6733 them suitable for ECDSA where domain parameters must be
6734 extracted before the specific public key;
6735 - ECDSA engine support has been added.
6736 [Nils Larsch <nla@trustcenter.de>]
6737
6738 *) Include some named elliptic curves, and add OIDs from X9.62,
6739 SECG, and WAP/WTLS. Each curve can be obtained from the new
6740 function
6741 EC_GROUP_new_by_curve_name(),
6742 and the list of available named curves can be obtained with
6743 EC_get_builtin_curves().
6744 Also add a 'curve_name' member to EC_GROUP objects, which can be
6745 accessed via
6746 EC_GROUP_set_curve_name()
6747 EC_GROUP_get_curve_name()
6748 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
6749
6750 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6751 was actually never needed) and in BN_mul(). The removal in BN_mul()
6752 required a small change in bn_mul_part_recursive() and the addition
6753 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6754 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6755 bn_sub_words() and bn_add_words() except they take arrays with
6756 differing sizes.
6757 [Richard Levitte]
6758
6759 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
6760
6761 *) Cleanse PEM buffers before freeing them since they may contain
6762 sensitive data.
6763 [Benjamin Bennett <ben@psc.edu>]
6764
6765 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6766 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6767 authentication-only ciphersuites.
6768 [Bodo Moeller]
6769
6770 *) Since AES128 and AES256 share a single mask bit in the logic of
6771 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6772 kludge to work properly if AES128 is available and AES256 isn't.
6773 [Victor Duchovni]
6774
6775 *) Expand security boundary to match 1.1.1 module.
6776 [Steve Henson]
6777
6778 *) Remove redundant features: hash file source, editing of test vectors
6779 modify fipsld to use external fips_premain.c signature.
6780 [Steve Henson]
6781
6782 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6783 run algorithm test programs.
6784 [Steve Henson]
6785
6786 *) Make algorithm test programs more tolerant of whitespace.
6787 [Steve Henson]
6788
6789 *) Have SSL/TLS server implementation tolerate "mismatched" record
6790 protocol version while receiving ClientHello even if the
6791 ClientHello is fragmented. (The server can't insist on the
6792 particular protocol version it has chosen before the ServerHello
6793 message has informed the client about his choice.)
6794 [Bodo Moeller]
6795
6796 *) Load error codes if they are not already present instead of using a
6797 static variable. This allows them to be cleanly unloaded and reloaded.
6798 [Steve Henson]
6799
6800 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6801
6802 *) Introduce limits to prevent malicious keys being able to
6803 cause a denial of service. (CVE-2006-2940)
6804 [Steve Henson, Bodo Moeller]
6805
6806 *) Fix ASN.1 parsing of certain invalid structures that can result
6807 in a denial of service. (CVE-2006-2937) [Steve Henson]
6808
6809 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6810 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6811
6812 *) Fix SSL client code which could crash if connecting to a
6813 malicious SSLv2 server. (CVE-2006-4343)
6814 [Tavis Ormandy and Will Drewry, Google Security Team]
6815
6816 *) Change ciphersuite string processing so that an explicit
6817 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
6818 will no longer include "AES128-SHA"), and any other similar
6819 ciphersuite (same bitmap) from *other* protocol versions (so that
6820 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
6821 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
6822 changes from 0.9.8b and 0.9.8d.
6823 [Bodo Moeller]
6824
6825 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
6826
6827 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6828 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6829
6830 *) Change the Unix randomness entropy gathering to use poll() when
6831 possible instead of select(), since the latter has some
6832 undesirable limitations.
6833 [Darryl Miles via Richard Levitte and Bodo Moeller]
6834
6835 *) Disable rogue ciphersuites:
6836
6837 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6838 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6839 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6840
6841 The latter two were purportedly from
6842 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6843 appear there.
6844
6845 Also deactivate the remaining ciphersuites from
6846 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6847 unofficial, and the ID has long expired.
6848 [Bodo Moeller]
6849
6850 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6851 dual-core machines) and other potential thread-safety issues.
6852 [Bodo Moeller]
6853
6854 Changes between 0.9.7i and 0.9.7j [04 May 2006]
6855
6856 *) Adapt fipsld and the build system to link against the validated FIPS
6857 module in FIPS mode.
6858 [Steve Henson]
6859
6860 *) Fixes for VC++ 2005 build under Windows.
6861 [Steve Henson]
6862
6863 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
6864 from a Windows bash shell such as MSYS. It is autodetected from the
6865 "config" script when run from a VC++ environment. Modify standard VC++
6866 build to use fipscanister.o from the GNU make build.
6867 [Steve Henson]
6868
6869 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
6870
6871 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
6872 The value now differs depending on if you build for FIPS or not.
6873 BEWARE! A program linked with a shared FIPSed libcrypto can't be
6874 safely run with a non-FIPSed libcrypto, as it may crash because of
6875 the difference induced by this change.
6876 [Andy Polyakov]
6877
6878 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
6879
6880 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6881 (part of SSL_OP_ALL). This option used to disable the
6882 countermeasure against man-in-the-middle protocol-version
6883 rollback in the SSL 2.0 server implementation, which is a bad
6884 idea. (CVE-2005-2969)
6885
6886 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6887 for Information Security, National Institute of Advanced Industrial
6888 Science and Technology [AIST], Japan)]
6889
6890 *) Minimal support for X9.31 signatures and PSS padding modes. This is
6891 mainly for FIPS compliance and not fully integrated at this stage.
6892 [Steve Henson]
6893
6894 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
6895 the exponentiation using a fixed-length exponent. (Otherwise,
6896 the information leaked through timing could expose the secret key
6897 after many signatures; cf. Bleichenbacher's attack on DSA with
6898 biased k.)
6899 [Bodo Moeller]
6900
6901 *) Make a new fixed-window mod_exp implementation the default for
6902 RSA, DSA, and DH private-key operations so that the sequence of
6903 squares and multiplies and the memory access pattern are
6904 independent of the particular secret key. This will mitigate
6905 cache-timing and potential related attacks.
6906
6907 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6908 and this is automatically used by BN_mod_exp_mont() if the new flag
6909 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6910 will use this BN flag for private exponents unless the flag
6911 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6912 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6913
6914 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6915
6916 *) Change the client implementation for SSLv23_method() and
6917 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6918 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6919 (Previously, the SSL 2.0 backwards compatible Client Hello
6920 message format would be used even with SSL_OP_NO_SSLv2.)
6921 [Bodo Moeller]
6922
6923 *) Add support for smime-type MIME parameter in S/MIME messages which some
6924 clients need.
6925 [Steve Henson]
6926
6927 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6928 a threadsafe manner. Modify rsa code to use new function and add calls
6929 to dsa and dh code (which had race conditions before).
6930 [Steve Henson]
6931
6932 *) Include the fixed error library code in the C error file definitions
6933 instead of fixing them up at runtime. This keeps the error code
6934 structures constant.
6935 [Steve Henson]
6936
6937 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6938
6939 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6940 OpenSSL 0.9.8.]
6941
6942 *) Fixes for newer kerberos headers. NB: the casts are needed because
6943 the 'length' field is signed on one version and unsigned on another
6944 with no (?) obvious way to tell the difference, without these VC++
6945 complains. Also the "definition" of FAR (blank) is no longer included
6946 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6947 some needed definitions.
6948 [Steve Henson]
6949
6950 *) Undo Cygwin change.
6951 [Ulf Möller]
6952
6953 *) Added support for proxy certificates according to RFC 3820.
6954 Because they may be a security thread to unaware applications,
6955 they must be explicitly allowed in run-time. See
6956 docs/HOWTO/proxy_certificates.txt for further information.
6957 [Richard Levitte]
6958
6959 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6960
6961 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6962 server and client random values. Previously
6963 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6964 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6965
6966 This change has negligible security impact because:
6967
6968 1. Server and client random values still have 24 bytes of pseudo random
6969 data.
6970
6971 2. Server and client random values are sent in the clear in the initial
6972 handshake.
6973
6974 3. The master secret is derived using the premaster secret (48 bytes in
6975 size for static RSA ciphersuites) as well as client server and random
6976 values.
6977
6978 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6979 to our attention.
6980
6981 [Stephen Henson, reported by UK NISCC]
6982
6983 *) Use Windows randomness collection on Cygwin.
6984 [Ulf Möller]
6985
6986 *) Fix hang in EGD/PRNGD query when communication socket is closed
6987 prematurely by EGD/PRNGD.
6988 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6989
6990 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6991 [Steve Henson]
6992
6993 *) Back-port of selected performance improvements from development
6994 branch, as well as improved support for PowerPC platforms.
6995 [Andy Polyakov]
6996
6997 *) Add lots of checks for memory allocation failure, error codes to indicate
6998 failure and freeing up memory if a failure occurs.
6999 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
7000
7001 *) Add new -passin argument to dgst.
7002 [Steve Henson]
7003
7004 *) Perform some character comparisons of different types in X509_NAME_cmp:
7005 this is needed for some certificates that re-encode DNs into UTF8Strings
7006 (in violation of RFC3280) and can't or won't issue name rollover
7007 certificates.
7008 [Steve Henson]
7009
7010 *) Make an explicit check during certificate validation to see that
7011 the CA setting in each certificate on the chain is correct. As a
7012 side effect always do the following basic checks on extensions,
7013 not just when there's an associated purpose to the check:
7014
7015 - if there is an unhandled critical extension (unless the user
7016 has chosen to ignore this fault)
7017 - if the path length has been exceeded (if one is set at all)
7018 - that certain extensions fit the associated purpose (if one has
7019 been given)
7020 [Richard Levitte]
7021
7022 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
7023
7024 *) Avoid a race condition when CRLs are checked in a multi threaded
7025 environment. This would happen due to the reordering of the revoked
7026 entries during signature checking and serial number lookup. Now the
7027 encoding is cached and the serial number sort performed under a lock.
7028 Add new STACK function sk_is_sorted().
7029 [Steve Henson]
7030
7031 *) Add Delta CRL to the extension code.
7032 [Steve Henson]
7033
7034 *) Various fixes to s3_pkt.c so alerts are sent properly.
7035 [David Holmes <d.holmes@f5.com>]
7036
7037 *) Reduce the chances of duplicate issuer name and serial numbers (in
7038 violation of RFC3280) using the OpenSSL certificate creation utilities.
7039 This is done by creating a random 64 bit value for the initial serial
7040 number when a serial number file is created or when a self signed
7041 certificate is created using 'openssl req -x509'. The initial serial
7042 number file is created using 'openssl x509 -next_serial' in CA.pl
7043 rather than being initialized to 1.
7044 [Steve Henson]
7045
7046 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
7047
7048 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
7049 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
7050 [Joe Orton, Steve Henson]
7051
7052 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
7053 (CVE-2004-0112)
7054 [Joe Orton, Steve Henson]
7055
7056 *) Make it possible to have multiple active certificates with the same
7057 subject in the CA index file. This is done only if the keyword
7058 'unique_subject' is set to 'no' in the main CA section (default
7059 if 'CA_default') of the configuration file. The value is saved
7060 with the database itself in a separate index attribute file,
7061 named like the index file with '.attr' appended to the name.
7062 [Richard Levitte]
7063
7064 *) X509 verify fixes. Disable broken certificate workarounds when
7065 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
7066 keyUsage extension present. Don't accept CRLs with unhandled critical
7067 extensions: since verify currently doesn't process CRL extensions this
7068 rejects a CRL with *any* critical extensions. Add new verify error codes
7069 for these cases.
7070 [Steve Henson]
7071
7072 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
7073 A clarification of RFC2560 will require the use of OCTET STRINGs and
7074 some implementations cannot handle the current raw format. Since OpenSSL
7075 copies and compares OCSP nonces as opaque blobs without any attempt at
7076 parsing them this should not create any compatibility issues.
7077 [Steve Henson]
7078
7079 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
7080 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
7081 this HMAC (and other) operations are several times slower than OpenSSL
7082 < 0.9.7.
7083 [Steve Henson]
7084
7085 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
7086 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7087
7088 *) Use the correct content when signing type "other".
7089 [Steve Henson]
7090
7091 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
7092
7093 *) Fix various bugs revealed by running the NISCC test suite:
7094
7095 Stop out of bounds reads in the ASN1 code when presented with
7096 invalid tags (CVE-2003-0543 and CVE-2003-0544).
7097
7098 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
7099
7100 If verify callback ignores invalid public key errors don't try to check
7101 certificate signature with the NULL public key.
7102
7103 [Steve Henson]
7104
7105 *) New -ignore_err option in ocsp application to stop the server
7106 exiting on the first error in a request.
7107 [Steve Henson]
7108
7109 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
7110 if the server requested one: as stated in TLS 1.0 and SSL 3.0
7111 specifications.
7112 [Steve Henson]
7113
7114 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
7115 extra data after the compression methods not only for TLS 1.0
7116 but also for SSL 3.0 (as required by the specification).
7117 [Bodo Moeller; problem pointed out by Matthias Loepfe]
7118
7119 *) Change X509_certificate_type() to mark the key as exported/exportable
7120 when it's 512 *bits* long, not 512 bytes.
7121 [Richard Levitte]
7122
7123 *) Change AES_cbc_encrypt() so it outputs exact multiple of
7124 blocks during encryption.
7125 [Richard Levitte]
7126
7127 *) Various fixes to base64 BIO and non blocking I/O. On write
7128 flushes were not handled properly if the BIO retried. On read
7129 data was not being buffered properly and had various logic bugs.
7130 This also affects blocking I/O when the data being decoded is a
7131 certain size.
7132 [Steve Henson]
7133
7134 *) Various S/MIME bugfixes and compatibility changes:
7135 output correct application/pkcs7 MIME type if
7136 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
7137 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
7138 of files as .eml work). Correctly handle very long lines in MIME
7139 parser.
7140 [Steve Henson]
7141
7142 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
7143
7144 *) Countermeasure against the Klima-Pokorny-Rosa extension of
7145 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
7146 a protocol version number mismatch like a decryption error
7147 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
7148 [Bodo Moeller]
7149
7150 *) Turn on RSA blinding by default in the default implementation
7151 to avoid a timing attack. Applications that don't want it can call
7152 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
7153 They would be ill-advised to do so in most cases.
7154 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
7155
7156 *) Change RSA blinding code so that it works when the PRNG is not
7157 seeded (in this case, the secret RSA exponent is abused as
7158 an unpredictable seed -- if it is not unpredictable, there
7159 is no point in blinding anyway). Make RSA blinding thread-safe
7160 by remembering the creator's thread ID in rsa->blinding and
7161 having all other threads use local one-time blinding factors
7162 (this requires more computation than sharing rsa->blinding, but
7163 avoids excessive locking; and if an RSA object is not shared
7164 between threads, blinding will still be very fast).
7165 [Bodo Moeller]
7166
7167 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
7168 ENGINE as defaults for all supported algorithms irrespective of
7169 the 'flags' parameter. 'flags' is now honoured, so applications
7170 should make sure they are passing it correctly.
7171 [Geoff Thorpe]
7172
7173 *) Target "mingw" now allows native Windows code to be generated in
7174 the Cygwin environment as well as with the MinGW compiler.
7175 [Ulf Moeller]
7176
7177 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
7178
7179 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7180 via timing by performing a MAC computation even if incorrect
7181 block cipher padding has been found. This is a countermeasure
7182 against active attacks where the attacker has to distinguish
7183 between bad padding and a MAC verification error. (CVE-2003-0078)
7184
7185 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7186 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7187 Martin Vuagnoux (EPFL, Ilion)]
7188
7189 *) Make the no-err option work as intended. The intention with no-err
7190 is not to have the whole error stack handling routines removed from
7191 libcrypto, it's only intended to remove all the function name and
7192 reason texts, thereby removing some of the footprint that may not
7193 be interesting if those errors aren't displayed anyway.
7194
7195 NOTE: it's still possible for any application or module to have its
7196 own set of error texts inserted. The routines are there, just not
7197 used by default when no-err is given.
7198 [Richard Levitte]
7199
7200 *) Add support for FreeBSD on IA64.
7201 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
7202
7203 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
7204 Kerberos function mit_des_cbc_cksum(). Before this change,
7205 the value returned by DES_cbc_cksum() was like the one from
7206 mit_des_cbc_cksum(), except the bytes were swapped.
7207 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
7208
7209 *) Allow an application to disable the automatic SSL chain building.
7210 Before this a rather primitive chain build was always performed in
7211 ssl3_output_cert_chain(): an application had no way to send the
7212 correct chain if the automatic operation produced an incorrect result.
7213
7214 Now the chain builder is disabled if either:
7215
7216 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
7217
7218 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
7219
7220 The reasoning behind this is that an application would not want the
7221 auto chain building to take place if extra chain certificates are
7222 present and it might also want a means of sending no additional
7223 certificates (for example the chain has two certificates and the
7224 root is omitted).
7225 [Steve Henson]
7226
7227 *) Add the possibility to build without the ENGINE framework.
7228 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7229
7230 *) Under Win32 gmtime() can return NULL: check return value in
7231 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
7232 [Steve Henson]
7233
7234 *) DSA routines: under certain error conditions uninitialized BN objects
7235 could be freed. Solution: make sure initialization is performed early
7236 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
7237 Nils Larsch <nla@trustcenter.de> via PR#459)
7238 [Lutz Jaenicke]
7239
7240 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
7241 checked on reconnect on the client side, therefore session resumption
7242 could still fail with a "ssl session id is different" error. This
7243 behaviour is masked when SSL_OP_ALL is used due to
7244 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
7245 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7246 followup to PR #377.
7247 [Lutz Jaenicke]
7248
7249 *) IA-32 assembler support enhancements: unified ELF targets, support
7250 for SCO/Caldera platforms, fix for Cygwin shared build.
7251 [Andy Polyakov]
7252
7253 *) Add support for FreeBSD on sparc64. As a consequence, support for
7254 FreeBSD on non-x86 processors is separate from x86 processors on
7255 the config script, much like the NetBSD support.
7256 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
7257
7258 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
7259
7260 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
7261 OpenSSL 0.9.7.]
7262
7263 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
7264 code (06) was taken as the first octet of the session ID and the last
7265 octet was ignored consequently. As a result SSLv2 client side session
7266 caching could not have worked due to the session ID mismatch between
7267 client and server.
7268 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7269 PR #377.
7270 [Lutz Jaenicke]
7271
7272 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
7273 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
7274 removed entirely.
7275 [Richard Levitte]
7276
7277 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
7278 seems that in spite of existing for more than a year, many application
7279 author have done nothing to provide the necessary callbacks, which
7280 means that this particular engine will not work properly anywhere.
7281 This is a very unfortunate situation which forces us, in the name
7282 of usability, to give the hw_ncipher.c a static lock, which is part
7283 of libcrypto.
7284 NOTE: This is for the 0.9.7 series ONLY. This hack will never
7285 appear in 0.9.8 or later. We EXPECT application authors to have
7286 dealt properly with this when 0.9.8 is released (unless we actually
7287 make such changes in the libcrypto locking code that changes will
7288 have to be made anyway).
7289 [Richard Levitte]
7290
7291 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
7292 octets have been read, EOF or an error occurs. Without this change
7293 some truncated ASN1 structures will not produce an error.
7294 [Steve Henson]
7295
7296 *) Disable Heimdal support, since it hasn't been fully implemented.
7297 Still give the possibility to force the use of Heimdal, but with
7298 warnings and a request that patches get sent to openssl-dev.
7299 [Richard Levitte]
7300
7301 *) Add the VC-CE target, introduce the WINCE sysname, and add
7302 INSTALL.WCE and appropriate conditionals to make it build.
7303 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7304
7305 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
7306 cygssl-x.y.z.dll, where x, y and z are the major, minor and
7307 edit numbers of the version.
7308 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7309
7310 *) Introduce safe string copy and catenation functions
7311 (BUF_strlcpy() and BUF_strlcat()).
7312 [Ben Laurie (CHATS) and Richard Levitte]
7313
7314 *) Avoid using fixed-size buffers for one-line DNs.
7315 [Ben Laurie (CHATS)]
7316
7317 *) Add BUF_MEM_grow_clean() to avoid information leakage when
7318 resizing buffers containing secrets, and use where appropriate.
7319 [Ben Laurie (CHATS)]
7320
7321 *) Avoid using fixed size buffers for configuration file location.
7322 [Ben Laurie (CHATS)]
7323
7324 *) Avoid filename truncation for various CA files.
7325 [Ben Laurie (CHATS)]
7326
7327 *) Use sizeof in preference to magic numbers.
7328 [Ben Laurie (CHATS)]
7329
7330 *) Avoid filename truncation in cert requests.
7331 [Ben Laurie (CHATS)]
7332
7333 *) Add assertions to check for (supposedly impossible) buffer
7334 overflows.
7335 [Ben Laurie (CHATS)]
7336
7337 *) Don't cache truncated DNS entries in the local cache (this could
7338 potentially lead to a spoofing attack).
7339 [Ben Laurie (CHATS)]
7340
7341 *) Fix various buffers to be large enough for hex/decimal
7342 representations in a platform independent manner.
7343 [Ben Laurie (CHATS)]
7344
7345 *) Add CRYPTO_realloc_clean() to avoid information leakage when
7346 resizing buffers containing secrets, and use where appropriate.
7347 [Ben Laurie (CHATS)]
7348
7349 *) Add BIO_indent() to avoid much slightly worrying code to do
7350 indents.
7351 [Ben Laurie (CHATS)]
7352
7353 *) Convert sprintf()/BIO_puts() to BIO_printf().
7354 [Ben Laurie (CHATS)]
7355
7356 *) buffer_gets() could terminate with the buffer only half
7357 full. Fixed.
7358 [Ben Laurie (CHATS)]
7359
7360 *) Add assertions to prevent user-supplied crypto functions from
7361 overflowing internal buffers by having large block sizes, etc.
7362 [Ben Laurie (CHATS)]
7363
7364 *) New OPENSSL_assert() macro (similar to assert(), but enabled
7365 unconditionally).
7366 [Ben Laurie (CHATS)]
7367
7368 *) Eliminate unused copy of key in RC4.
7369 [Ben Laurie (CHATS)]
7370
7371 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
7372 [Ben Laurie (CHATS)]
7373
7374 *) Fix off-by-one error in EGD path.
7375 [Ben Laurie (CHATS)]
7376
7377 *) If RANDFILE path is too long, ignore instead of truncating.
7378 [Ben Laurie (CHATS)]
7379
7380 *) Eliminate unused and incorrectly sized X.509 structure
7381 CBCParameter.
7382 [Ben Laurie (CHATS)]
7383
7384 *) Eliminate unused and dangerous function knumber().
7385 [Ben Laurie (CHATS)]
7386
7387 *) Eliminate unused and dangerous structure, KSSL_ERR.
7388 [Ben Laurie (CHATS)]
7389
7390 *) Protect against overlong session ID context length in an encoded
7391 session object. Since these are local, this does not appear to be
7392 exploitable.
7393 [Ben Laurie (CHATS)]
7394
7395 *) Change from security patch (see 0.9.6e below) that did not affect
7396 the 0.9.6 release series:
7397
7398 Remote buffer overflow in SSL3 protocol - an attacker could
7399 supply an oversized master key in Kerberos-enabled versions.
7400 (CVE-2002-0657)
7401 [Ben Laurie (CHATS)]
7402
7403 *) Change the SSL kerb5 codes to match RFC 2712.
7404 [Richard Levitte]
7405
7406 *) Make -nameopt work fully for req and add -reqopt switch.
7407 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
7408
7409 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
7410 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
7411
7412 *) Make sure tests can be performed even if the corresponding algorithms
7413 have been removed entirely. This was also the last step to make
7414 OpenSSL compilable with DJGPP under all reasonable conditions.
7415 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
7416
7417 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
7418 to allow version independent disabling of normally unselected ciphers,
7419 which may be activated as a side-effect of selecting a single cipher.
7420
7421 (E.g., cipher list string "RSA" enables ciphersuites that are left
7422 out of "ALL" because they do not provide symmetric encryption.
7423 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
7424 [Lutz Jaenicke, Bodo Moeller]
7425
7426 *) Add appropriate support for separate platform-dependent build
7427 directories. The recommended way to make a platform-dependent
7428 build directory is the following (tested on Linux), maybe with
7429 some local tweaks:
7430
7431 # Place yourself outside of the OpenSSL source tree. In
7432 # this example, the environment variable OPENSSL_SOURCE
7433 # is assumed to contain the absolute OpenSSL source directory.
7434 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
7435 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
7436 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
7437 mkdir -p `dirname $F`
7438 ln -s $OPENSSL_SOURCE/$F $F
7439 done
7440
7441 To be absolutely sure not to disturb the source tree, a "make clean"
7442 is a good thing. If it isn't successful, don't worry about it,
7443 it probably means the source directory is very clean.
7444 [Richard Levitte]
7445
7446 *) Make sure any ENGINE control commands make local copies of string
7447 pointers passed to them whenever necessary. Otherwise it is possible
7448 the caller may have overwritten (or deallocated) the original string
7449 data when a later ENGINE operation tries to use the stored values.
7450 [Götz Babin-Ebell <babinebell@trustcenter.de>]
7451
7452 *) Improve diagnostics in file reading and command-line digests.
7453 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
7454
7455 *) Add AES modes CFB and OFB to the object database. Correct an
7456 error in AES-CFB decryption.
7457 [Richard Levitte]
7458
7459 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
7460 allows existing EVP_CIPHER_CTX structures to be reused after
7461 calling EVP_*Final(). This behaviour is used by encryption
7462 BIOs and some applications. This has the side effect that
7463 applications must explicitly clean up cipher contexts with
7464 EVP_CIPHER_CTX_cleanup() or they will leak memory.
7465 [Steve Henson]
7466
7467 *) Check the values of dna and dnb in bn_mul_recursive before calling
7468 bn_mul_comba (a non zero value means the a or b arrays do not contain
7469 n2 elements) and fallback to bn_mul_normal if either is not zero.
7470 [Steve Henson]
7471
7472 *) Fix escaping of non-ASCII characters when using the -subj option
7473 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
7474 [Lutz Jaenicke]
7475
7476 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
7477 form for "surname", serialNumber has no short form.
7478 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
7479 therefore remove "mail" short name for "internet 7".
7480 The OID for unique identifiers in X509 certificates is
7481 x500UniqueIdentifier, not uniqueIdentifier.
7482 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7483 [Lutz Jaenicke]
7484
7485 *) Add an "init" command to the ENGINE config module and auto initialize
7486 ENGINEs. Without any "init" command the ENGINE will be initialized
7487 after all ctrl commands have been executed on it. If init=1 the
7488 ENGINE is initialized at that point (ctrls before that point are run
7489 on the uninitialized ENGINE and after on the initialized one). If
7490 init=0 then the ENGINE will not be initialized at all.
7491 [Steve Henson]
7492
7493 *) Fix the 'app_verify_callback' interface so that the user-defined
7494 argument is actually passed to the callback: In the
7495 SSL_CTX_set_cert_verify_callback() prototype, the callback
7496 declaration has been changed from
7497 int (*cb)()
7498 into
7499 int (*cb)(X509_STORE_CTX *,void *);
7500 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7501 i=s->ctx->app_verify_callback(&ctx)
7502 has been changed into
7503 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7504
7505 To update applications using SSL_CTX_set_cert_verify_callback(),
7506 a dummy argument can be added to their callback functions.
7507 [D. K. Smetters <smetters@parc.xerox.com>]
7508
7509 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7510 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7511
7512 *) Add and OPENSSL_LOAD_CONF define which will cause
7513 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7514 This allows older applications to transparently support certain
7515 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7516 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7517 load the config file and OPENSSL_add_all_algorithms_conf() which will
7518 always load it have also been added.
7519 [Steve Henson]
7520
7521 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7522 Adjust NIDs and EVP layer.
7523 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7524
7525 *) Config modules support in openssl utility.
7526
7527 Most commands now load modules from the config file,
7528 though in a few (such as version) this isn't done
7529 because it couldn't be used for anything.
7530
7531 In the case of ca and req the config file used is
7532 the same as the utility itself: that is the -config
7533 command line option can be used to specify an
7534 alternative file.
7535 [Steve Henson]
7536
7537 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7538 use "openssl_conf" if filename is NULL use default openssl config file.
7539 [Steve Henson]
7540
7541 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7542 config section name. Add a new flag to tolerate a missing config file
7543 and move code to CONF_modules_load_file().
7544 [Steve Henson]
7545
7546 *) Support for crypto accelerator cards from Accelerated Encryption
7547 Processing, www.aep.ie. (Use engine 'aep')
7548 The support was copied from 0.9.6c [engine] and adapted/corrected
7549 to work with the new engine framework.
7550 [AEP Inc. and Richard Levitte]
7551
7552 *) Support for SureWare crypto accelerator cards from Baltimore
7553 Technologies. (Use engine 'sureware')
7554 The support was copied from 0.9.6c [engine] and adapted
7555 to work with the new engine framework.
7556 [Richard Levitte]
7557
7558 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7559 make the newer ENGINE framework commands for the CHIL engine work.
7560 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7561
7562 *) Make it possible to produce shared libraries on ReliantUNIX.
7563 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7564
7565 *) Add the configuration target debug-linux-ppro.
7566 Make 'openssl rsa' use the general key loading routines
7567 implemented in apps.c, and make those routines able to
7568 handle the key format FORMAT_NETSCAPE and the variant
7569 FORMAT_IISSGC.
7570 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7571
7572 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7573 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7574
7575 *) Add -keyform to rsautl, and document -engine.
7576 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7577
7578 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7579 BIO_R_NO_SUCH_FILE error code rather than the generic
7580 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7581 [Ben Laurie]
7582
7583 *) Add new functions
7584 ERR_peek_last_error
7585 ERR_peek_last_error_line
7586 ERR_peek_last_error_line_data.
7587 These are similar to
7588 ERR_peek_error
7589 ERR_peek_error_line
7590 ERR_peek_error_line_data,
7591 but report on the latest error recorded rather than the first one
7592 still in the error queue.
7593 [Ben Laurie, Bodo Moeller]
7594
7595 *) default_algorithms option in ENGINE config module. This allows things
7596 like:
7597 default_algorithms = ALL
7598 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7599 [Steve Henson]
7600
7601 *) Preliminary ENGINE config module.
7602 [Steve Henson]
7603
7604 *) New experimental application configuration code.
7605 [Steve Henson]
7606
7607 *) Change the AES code to follow the same name structure as all other
7608 symmetric ciphers, and behave the same way. Move everything to
7609 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7610 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7611
7612 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7613 [Ben Laurie and Theo de Raadt]
7614
7615 *) Add option to output public keys in req command.
7616 [Massimiliano Pala madwolf@openca.org]
7617
7618 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7619 (up to about 10% better than before for P-192 and P-224).
7620 [Bodo Moeller]
7621
7622 *) New functions/macros
7623
7624 SSL_CTX_set_msg_callback(ctx, cb)
7625 SSL_CTX_set_msg_callback_arg(ctx, arg)
7626 SSL_set_msg_callback(ssl, cb)
7627 SSL_set_msg_callback_arg(ssl, arg)
7628
7629 to request calling a callback function
7630
7631 void cb(int write_p, int version, int content_type,
7632 const void *buf, size_t len, SSL *ssl, void *arg)
7633
7634 whenever a protocol message has been completely received
7635 (write_p == 0) or sent (write_p == 1). Here 'version' is the
7636 protocol version according to which the SSL library interprets
7637 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
7638 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
7639 the content type as defined in the SSL 3.0/TLS 1.0 protocol
7640 specification (change_cipher_spec(20), alert(21), handshake(22)).
7641 'buf' and 'len' point to the actual message, 'ssl' to the
7642 SSL object, and 'arg' is the application-defined value set by
7643 SSL[_CTX]_set_msg_callback_arg().
7644
7645 'openssl s_client' and 'openssl s_server' have new '-msg' options
7646 to enable a callback that displays all protocol messages.
7647 [Bodo Moeller]
7648
7649 *) Change the shared library support so shared libraries are built as
7650 soon as the corresponding static library is finished, and thereby get
7651 openssl and the test programs linked against the shared library.
7652 This still only happens when the keyword "shard" has been given to
7653 the configuration scripts.
7654
7655 NOTE: shared library support is still an experimental thing, and
7656 backward binary compatibility is still not guaranteed.
7657 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
7658
7659 *) Add support for Subject Information Access extension.
7660 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7661
7662 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
7663 additional bytes when new memory had to be allocated, not just
7664 when reusing an existing buffer.
7665 [Bodo Moeller]
7666
7667 *) New command line and configuration option 'utf8' for the req command.
7668 This allows field values to be specified as UTF8 strings.
7669 [Steve Henson]
7670
7671 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
7672 runs for the former and machine-readable output for the latter.
7673 [Ben Laurie]
7674
7675 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
7676 of the e-mail address in the DN (i.e., it will go into a certificate
7677 extension only). The new configuration file option 'email_in_dn = no'
7678 has the same effect.
7679 [Massimiliano Pala madwolf@openca.org]
7680
7681 *) Change all functions with names starting with des_ to be starting
7682 with DES_ instead. Add wrappers that are compatible with libdes,
7683 but are named _ossl_old_des_*. Finally, add macros that map the
7684 des_* symbols to the corresponding _ossl_old_des_* if libdes
7685 compatibility is desired. If OpenSSL 0.9.6c compatibility is
7686 desired, the des_* symbols will be mapped to DES_*, with one
7687 exception.
7688
7689 Since we provide two compatibility mappings, the user needs to
7690 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
7691 compatibility is desired. The default (i.e., when that macro
7692 isn't defined) is OpenSSL 0.9.6c compatibility.
7693
7694 There are also macros that enable and disable the support of old
7695 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
7696 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
7697 are defined, the default will apply: to support the old des routines.
7698
7699 In either case, one must include openssl/des.h to get the correct
7700 definitions. Do not try to just include openssl/des_old.h, that
7701 won't work.
7702
7703 NOTE: This is a major break of an old API into a new one. Software
7704 authors are encouraged to switch to the DES_ style functions. Some
7705 time in the future, des_old.h and the libdes compatibility functions
7706 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
7707 default), and then completely removed.
7708 [Richard Levitte]
7709
7710 *) Test for certificates which contain unsupported critical extensions.
7711 If such a certificate is found during a verify operation it is
7712 rejected by default: this behaviour can be overridden by either
7713 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
7714 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
7715 X509_supported_extension() has also been added which returns 1 if a
7716 particular extension is supported.
7717 [Steve Henson]
7718
7719 *) Modify the behaviour of EVP cipher functions in similar way to digests
7720 to retain compatibility with existing code.
7721 [Steve Henson]
7722
7723 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
7724 compatibility with existing code. In particular the 'ctx' parameter does
7725 not have to be to be initialized before the call to EVP_DigestInit() and
7726 it is tidied up after a call to EVP_DigestFinal(). New function
7727 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
7728 EVP_MD_CTX_copy() changed to not require the destination to be
7729 initialized valid and new function EVP_MD_CTX_copy_ex() added which
7730 requires the destination to be valid.
7731
7732 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
7733 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
7734 [Steve Henson]
7735
7736 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
7737 so that complete 'Handshake' protocol structures are kept in memory
7738 instead of overwriting 'msg_type' and 'length' with 'body' data.
7739 [Bodo Moeller]
7740
7741 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
7742 [Massimo Santin via Richard Levitte]
7743
7744 *) Major restructuring to the underlying ENGINE code. This includes
7745 reduction of linker bloat, separation of pure "ENGINE" manipulation
7746 (initialisation, etc) from functionality dealing with implementations
7747 of specific crypto interfaces. This change also introduces integrated
7748 support for symmetric ciphers and digest implementations - so ENGINEs
7749 can now accelerate these by providing EVP_CIPHER and EVP_MD
7750 implementations of their own. This is detailed in crypto/engine/README
7751 as it couldn't be adequately described here. However, there are a few
7752 API changes worth noting - some RSA, DSA, DH, and RAND functions that
7753 were changed in the original introduction of ENGINE code have now
7754 reverted back - the hooking from this code to ENGINE is now a good
7755 deal more passive and at run-time, operations deal directly with
7756 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
7757 dereferencing through an ENGINE pointer any more. Also, the ENGINE
7758 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
7759 they were not being used by the framework as there is no concept of a
7760 BIGNUM_METHOD and they could not be generalised to the new
7761 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
7762 ENGINE_cpy() has been removed as it cannot be consistently defined in
7763 the new code.
7764 [Geoff Thorpe]
7765
7766 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
7767 [Steve Henson]
7768
7769 *) Change mkdef.pl to sort symbols that get the same entry number,
7770 and make sure the automatically generated functions ERR_load_*
7771 become part of libeay.num as well.
7772 [Richard Levitte]
7773
7774 *) New function SSL_renegotiate_pending(). This returns true once
7775 renegotiation has been requested (either SSL_renegotiate() call
7776 or HelloRequest/ClientHello received from the peer) and becomes
7777 false once a handshake has been completed.
7778 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
7779 sends a HelloRequest, but does not ensure that a handshake takes
7780 place. SSL_renegotiate_pending() is useful for checking if the
7781 client has followed the request.)
7782 [Bodo Moeller]
7783
7784 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7785 By default, clients may request session resumption even during
7786 renegotiation (if session ID contexts permit); with this option,
7787 session resumption is possible only in the first handshake.
7788
7789 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7790 more bits available for options that should not be part of
7791 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7792 [Bodo Moeller]
7793
7794 *) Add some demos for certificate and certificate request creation.
7795 [Steve Henson]
7796
7797 *) Make maximum certificate chain size accepted from the peer application
7798 settable (SSL*_get/set_max_cert_list()), as proposed by
7799 "Douglas E. Engert" <deengert@anl.gov>.
7800 [Lutz Jaenicke]
7801
7802 *) Add support for shared libraries for Unixware-7
7803 (Boyd Lynn Gerber <gerberb@zenez.com>).
7804 [Lutz Jaenicke]
7805
7806 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7807 be done prior to destruction. Use this to unload error strings from
7808 ENGINEs that load their own error strings. NB: This adds two new API
7809 functions to "get" and "set" this destroy handler in an ENGINE.
7810 [Geoff Thorpe]
7811
7812 *) Alter all existing ENGINE implementations (except "openssl" and
7813 "openbsd") to dynamically instantiate their own error strings. This
7814 makes them more flexible to be built both as statically-linked ENGINEs
7815 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
7816 Also, add stub code to each that makes building them as self-contained
7817 shared-libraries easier (see README.ENGINE).
7818 [Geoff Thorpe]
7819
7820 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
7821 implementations into applications that are completely implemented in
7822 self-contained shared-libraries. The "dynamic" ENGINE exposes control
7823 commands that can be used to configure what shared-library to load and
7824 to control aspects of the way it is handled. Also, made an update to
7825 the README.ENGINE file that brings its information up-to-date and
7826 provides some information and instructions on the "dynamic" ENGINE
7827 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
7828 [Geoff Thorpe]
7829
7830 *) Make it possible to unload ranges of ERR strings with a new
7831 "ERR_unload_strings" function.
7832 [Geoff Thorpe]
7833
7834 *) Add a copy() function to EVP_MD.
7835 [Ben Laurie]
7836
7837 *) Make EVP_MD routines take a context pointer instead of just the
7838 md_data void pointer.
7839 [Ben Laurie]
7840
7841 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
7842 that the digest can only process a single chunk of data
7843 (typically because it is provided by a piece of
7844 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
7845 is only going to provide a single chunk of data, and hence the
7846 framework needn't accumulate the data for oneshot drivers.
7847 [Ben Laurie]
7848
7849 *) As with "ERR", make it possible to replace the underlying "ex_data"
7850 functions. This change also alters the storage and management of global
7851 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
7852 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
7853 index counters. The API functions that use this state have been changed
7854 to take a "class_index" rather than pointers to the class's local STACK
7855 and counter, and there is now an API function to dynamically create new
7856 classes. This centralisation allows us to (a) plug a lot of the
7857 thread-safety problems that existed, and (b) makes it possible to clean
7858 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
7859 such data would previously have always leaked in application code and
7860 workarounds were in place to make the memory debugging turn a blind eye
7861 to it. Application code that doesn't use this new function will still
7862 leak as before, but their memory debugging output will announce it now
7863 rather than letting it slide.
7864
7865 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
7866 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
7867 has a return value to indicate success or failure.
7868 [Geoff Thorpe]
7869
7870 *) Make it possible to replace the underlying "ERR" functions such that the
7871 global state (2 LHASH tables and 2 locks) is only used by the "default"
7872 implementation. This change also adds two functions to "get" and "set"
7873 the implementation prior to it being automatically set the first time
7874 any other ERR function takes place. Ie. an application can call "get",
7875 pass the return value to a module it has just loaded, and that module
7876 can call its own "set" function using that value. This means the
7877 module's "ERR" operations will use (and modify) the error state in the
7878 application and not in its own statically linked copy of OpenSSL code.
7879 [Geoff Thorpe]
7880
7881 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
7882 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
7883 the operation, and provides a more encapsulated way for external code
7884 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
7885 to use these functions rather than manually incrementing the counts.
7886
7887 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
7888 [Geoff Thorpe]
7889
7890 *) Add EVP test program.
7891 [Ben Laurie]
7892
7893 *) Add symmetric cipher support to ENGINE. Expect the API to change!
7894 [Ben Laurie]
7895
7896 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
7897 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
7898 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
7899 These allow a CRL to be built without having to access X509_CRL fields
7900 directly. Modify 'ca' application to use new functions.
7901 [Steve Henson]
7902
7903 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
7904 bug workarounds. Rollback attack detection is a security feature.
7905 The problem will only arise on OpenSSL servers when TLSv1 is not
7906 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7907 Software authors not wanting to support TLSv1 will have special reasons
7908 for their choice and can explicitly enable this option.
7909 [Bodo Moeller, Lutz Jaenicke]
7910
7911 *) Rationalise EVP so it can be extended: don't include a union of
7912 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7913 (similar to those existing for EVP_CIPHER_CTX).
7914 Usage example:
7915
7916 EVP_MD_CTX md;
7917
7918 EVP_MD_CTX_init(&md); /* new function call */
7919 EVP_DigestInit(&md, EVP_sha1());
7920 EVP_DigestUpdate(&md, in, len);
7921 EVP_DigestFinal(&md, out, NULL);
7922 EVP_MD_CTX_cleanup(&md); /* new function call */
7923
7924 [Ben Laurie]
7925
7926 *) Make DES key schedule conform to the usual scheme, as well as
7927 correcting its structure. This means that calls to DES functions
7928 now have to pass a pointer to a des_key_schedule instead of a
7929 plain des_key_schedule (which was actually always a pointer
7930 anyway): E.g.,
7931
7932 des_key_schedule ks;
7933
7934 des_set_key_checked(..., &ks);
7935 des_ncbc_encrypt(..., &ks, ...);
7936
7937 (Note that a later change renames 'des_...' into 'DES_...'.)
7938 [Ben Laurie]
7939
7940 *) Initial reduction of linker bloat: the use of some functions, such as
7941 PEM causes large amounts of unused functions to be linked in due to
7942 poor organisation. For example pem_all.c contains every PEM function
7943 which has a knock on effect of linking in large amounts of (unused)
7944 ASN1 code. Grouping together similar functions and splitting unrelated
7945 functions prevents this.
7946 [Steve Henson]
7947
7948 *) Cleanup of EVP macros.
7949 [Ben Laurie]
7950
7951 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7952 correct _ecb suffix.
7953 [Ben Laurie]
7954
7955 *) Add initial OCSP responder support to ocsp application. The
7956 revocation information is handled using the text based index
7957 use by the ca application. The responder can either handle
7958 requests generated internally, supplied in files (for example
7959 via a CGI script) or using an internal minimal server.
7960 [Steve Henson]
7961
7962 *) Add configuration choices to get zlib compression for TLS.
7963 [Richard Levitte]
7964
7965 *) Changes to Kerberos SSL for RFC 2712 compliance:
7966 1. Implemented real KerberosWrapper, instead of just using
7967 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7968 2. Implemented optional authenticator field of KerberosWrapper.
7969
7970 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7971 and authenticator structs; see crypto/krb5/.
7972
7973 Generalized Kerberos calls to support multiple Kerberos libraries.
7974 [Vern Staats <staatsvr@asc.hpc.mil>,
7975 Jeffrey Altman <jaltman@columbia.edu>
7976 via Richard Levitte]
7977
7978 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7979 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7980 values for each of the key sizes rather than having just
7981 parameters (and 'speed' generating keys each time).
7982 [Geoff Thorpe]
7983
7984 *) Speed up EVP routines.
7985 Before:
7986 encrypt
7987 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7988 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7989 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7990 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7991 decrypt
7992 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7993 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7994 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7995 After:
7996 encrypt
7997 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7998 decrypt
7999 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
8000 [Ben Laurie]
8001
8002 *) Added the OS2-EMX target.
8003 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
8004
8005 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
8006 to support NCONF routines in extension code. New function CONF_set_nconf()
8007 to allow functions which take an NCONF to also handle the old LHASH
8008 structure: this means that the old CONF compatible routines can be
8009 retained (in particular wrt extensions) without having to duplicate the
8010 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
8011 [Steve Henson]
8012
8013 *) Enhance the general user interface with mechanisms for inner control
8014 and with possibilities to have yes/no kind of prompts.
8015 [Richard Levitte]
8016
8017 *) Change all calls to low level digest routines in the library and
8018 applications to use EVP. Add missing calls to HMAC_cleanup() and
8019 don't assume HMAC_CTX can be copied using memcpy().
8020 [Verdon Walker <VWalker@novell.com>, Steve Henson]
8021
8022 *) Add the possibility to control engines through control names but with
8023 arbitrary arguments instead of just a string.
8024 Change the key loaders to take a UI_METHOD instead of a callback
8025 function pointer. NOTE: this breaks binary compatibility with earlier
8026 versions of OpenSSL [engine].
8027 Adapt the nCipher code for these new conditions and add a card insertion
8028 callback.
8029 [Richard Levitte]
8030
8031 *) Enhance the general user interface with mechanisms to better support
8032 dialog box interfaces, application-defined prompts, the possibility
8033 to use defaults (for example default passwords from somewhere else)
8034 and interrupts/cancellations.
8035 [Richard Levitte]
8036
8037 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
8038 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
8039 [Steve Henson]
8040
8041 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
8042 tidy up some unnecessarily weird code in 'sk_new()').
8043 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
8044
8045 *) Change the key loading routines for ENGINEs to use the same kind
8046 callback (pem_password_cb) as all other routines that need this
8047 kind of callback.
8048 [Richard Levitte]
8049
8050 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
8051 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
8052 than this minimum value is recommended.
8053 [Lutz Jaenicke]
8054
8055 *) New random seeder for OpenVMS, using the system process statistics
8056 that are easily reachable.
8057 [Richard Levitte]
8058
8059 *) Windows apparently can't transparently handle global
8060 variables defined in DLLs. Initialisations such as:
8061
8062 const ASN1_ITEM *it = &ASN1_INTEGER_it;
8063
8064 won't compile. This is used by the any applications that need to
8065 declare their own ASN1 modules. This was fixed by adding the option
8066 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
8067 needed for static libraries under Win32.
8068 [Steve Henson]
8069
8070 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
8071 setting of purpose and trust fields. New X509_STORE trust and
8072 purpose functions and tidy up setting in other SSL functions.
8073 [Steve Henson]
8074
8075 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
8076 structure. These are inherited by X509_STORE_CTX when it is
8077 initialised. This allows various defaults to be set in the
8078 X509_STORE structure (such as flags for CRL checking and custom
8079 purpose or trust settings) for functions which only use X509_STORE_CTX
8080 internally such as S/MIME.
8081
8082 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
8083 trust settings if they are not set in X509_STORE. This allows X509_STORE
8084 purposes and trust (in S/MIME for example) to override any set by default.
8085
8086 Add command line options for CRL checking to smime, s_client and s_server
8087 applications.
8088 [Steve Henson]
8089
8090 *) Initial CRL based revocation checking. If the CRL checking flag(s)
8091 are set then the CRL is looked up in the X509_STORE structure and
8092 its validity and signature checked, then if the certificate is found
8093 in the CRL the verify fails with a revoked error.
8094
8095 Various new CRL related callbacks added to X509_STORE_CTX structure.
8096
8097 Command line options added to 'verify' application to support this.
8098
8099 This needs some additional work, such as being able to handle multiple
8100 CRLs with different times, extension based lookup (rather than just
8101 by subject name) and ultimately more complete V2 CRL extension
8102 handling.
8103 [Steve Henson]
8104
8105 *) Add a general user interface API (crypto/ui/). This is designed
8106 to replace things like des_read_password and friends (backward
8107 compatibility functions using this new API are provided).
8108 The purpose is to remove prompting functions from the DES code
8109 section as well as provide for prompting through dialog boxes in
8110 a window system and the like.
8111 [Richard Levitte]
8112
8113 *) Add "ex_data" support to ENGINE so implementations can add state at a
8114 per-structure level rather than having to store it globally.
8115 [Geoff]
8116
8117 *) Make it possible for ENGINE structures to be copied when retrieved by
8118 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
8119 This causes the "original" ENGINE structure to act like a template,
8120 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
8121 operational state can be localised to each ENGINE structure, despite the
8122 fact they all share the same "methods". New ENGINE structures returned in
8123 this case have no functional references and the return value is the single
8124 structural reference. This matches the single structural reference returned
8125 by ENGINE_by_id() normally, when it is incremented on the pre-existing
8126 ENGINE structure.
8127 [Geoff]
8128
8129 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
8130 needs to match any other type at all we need to manually clear the
8131 tag cache.
8132 [Steve Henson]
8133
8134 *) Changes to the "openssl engine" utility to include;
8135 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
8136 about an ENGINE's available control commands.
8137 - executing control commands from command line arguments using the
8138 '-pre' and '-post' switches. '-post' is only used if '-t' is
8139 specified and the ENGINE is successfully initialised. The syntax for
8140 the individual commands are colon-separated, for example;
8141 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
8142 [Geoff]
8143
8144 *) New dynamic control command support for ENGINEs. ENGINEs can now
8145 declare their own commands (numbers), names (strings), descriptions,
8146 and input types for run-time discovery by calling applications. A
8147 subset of these commands are implicitly classed as "executable"
8148 depending on their input type, and only these can be invoked through
8149 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
8150 can be based on user input, config files, etc). The distinction is
8151 that "executable" commands cannot return anything other than a boolean
8152 result and can only support numeric or string input, whereas some
8153 discoverable commands may only be for direct use through
8154 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
8155 pointers, or other custom uses. The "executable" commands are to
8156 support parameterisations of ENGINE behaviour that can be
8157 unambiguously defined by ENGINEs and used consistently across any
8158 OpenSSL-based application. Commands have been added to all the
8159 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
8160 control over shared-library paths without source code alterations.
8161 [Geoff]
8162
8163 *) Changed all ENGINE implementations to dynamically allocate their
8164 ENGINEs rather than declaring them statically. Apart from this being
8165 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
8166 this also allows the implementations to compile without using the
8167 internal engine_int.h header.
8168 [Geoff]
8169
8170 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
8171 'const' value. Any code that should be able to modify a RAND_METHOD
8172 should already have non-const pointers to it (ie. they should only
8173 modify their own ones).
8174 [Geoff]
8175
8176 *) Made a variety of little tweaks to the ENGINE code.
8177 - "atalla" and "ubsec" string definitions were moved from header files
8178 to C code. "nuron" string definitions were placed in variables
8179 rather than hard-coded - allowing parameterisation of these values
8180 later on via ctrl() commands.
8181 - Removed unused "#if 0"'d code.
8182 - Fixed engine list iteration code so it uses ENGINE_free() to release
8183 structural references.
8184 - Constified the RAND_METHOD element of ENGINE structures.
8185 - Constified various get/set functions as appropriate and added
8186 missing functions (including a catch-all ENGINE_cpy that duplicates
8187 all ENGINE values onto a new ENGINE except reference counts/state).
8188 - Removed NULL parameter checks in get/set functions. Setting a method
8189 or function to NULL is a way of cancelling out a previously set
8190 value. Passing a NULL ENGINE parameter is just plain stupid anyway
8191 and doesn't justify the extra error symbols and code.
8192 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
8193 flags from engine_int.h to engine.h.
8194 - Changed prototypes for ENGINE handler functions (init(), finish(),
8195 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
8196 [Geoff]
8197
8198 *) Implement binary inversion algorithm for BN_mod_inverse in addition
8199 to the algorithm using long division. The binary algorithm can be
8200 used only if the modulus is odd. On 32-bit systems, it is faster
8201 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
8202 roughly 5-15% for 256-bit moduli), so we use it only for moduli
8203 up to 450 bits. In 64-bit environments, the binary algorithm
8204 appears to be advantageous for much longer moduli; here we use it
8205 for moduli up to 2048 bits.
8206 [Bodo Moeller]
8207
8208 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
8209 could not support the combine flag in choice fields.
8210 [Steve Henson]
8211
8212 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
8213 extensions from a certificate request to the certificate.
8214 [Steve Henson]
8215
8216 *) Allow multiple 'certopt' and 'nameopt' options to be separated
8217 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
8218 file: this allows the display of the certificate about to be
8219 signed to be customised, to allow certain fields to be included
8220 or excluded and extension details. The old system didn't display
8221 multicharacter strings properly, omitted fields not in the policy
8222 and couldn't display additional details such as extensions.
8223 [Steve Henson]
8224
8225 *) Function EC_POINTs_mul for multiple scalar multiplication
8226 of an arbitrary number of elliptic curve points
8227 \sum scalars[i]*points[i],
8228 optionally including the generator defined for the EC_GROUP:
8229 scalar*generator + \sum scalars[i]*points[i].
8230
8231 EC_POINT_mul is a simple wrapper function for the typical case
8232 that the point list has just one item (besides the optional
8233 generator).
8234 [Bodo Moeller]
8235
8236 *) First EC_METHODs for curves over GF(p):
8237
8238 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
8239 operations and provides various method functions that can also
8240 operate with faster implementations of modular arithmetic.
8241
8242 EC_GFp_mont_method() reuses most functions that are part of
8243 EC_GFp_simple_method, but uses Montgomery arithmetic.
8244
8245 [Bodo Moeller; point addition and point doubling
8246 implementation directly derived from source code provided by
8247 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
8248
8249 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
8250 crypto/ec/ec_lib.c):
8251
8252 Curves are EC_GROUP objects (with an optional group generator)
8253 based on EC_METHODs that are built into the library.
8254
8255 Points are EC_POINT objects based on EC_GROUP objects.
8256
8257 Most of the framework would be able to handle curves over arbitrary
8258 finite fields, but as there are no obvious types for fields other
8259 than GF(p), some functions are limited to that for now.
8260 [Bodo Moeller]
8261
8262 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
8263 that the file contains a complete HTTP response.
8264 [Richard Levitte]
8265
8266 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
8267 change the def and num file printf format specifier from "%-40sXXX"
8268 to "%-39s XXX". The latter will always guarantee a space after the
8269 field while the former will cause them to run together if the field
8270 is 40 of more characters long.
8271 [Steve Henson]
8272
8273 *) Constify the cipher and digest 'method' functions and structures
8274 and modify related functions to take constant EVP_MD and EVP_CIPHER
8275 pointers.
8276 [Steve Henson]
8277
8278 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
8279 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
8280 [Bodo Moeller]
8281
8282 *) Modify EVP_Digest*() routines so they now return values. Although the
8283 internal software routines can never fail additional hardware versions
8284 might.
8285 [Steve Henson]
8286
8287 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
8288
8289 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
8290 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
8291
8292 ASN1 error codes
8293 ERR_R_NESTED_ASN1_ERROR
8294 ...
8295 ERR_R_MISSING_ASN1_EOS
8296 were 4 .. 9, conflicting with
8297 ERR_LIB_RSA (= ERR_R_RSA_LIB)
8298 ...
8299 ERR_LIB_PEM (= ERR_R_PEM_LIB).
8300 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
8301
8302 Add new error code 'ERR_R_INTERNAL_ERROR'.
8303 [Bodo Moeller]
8304
8305 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
8306 suffices.
8307 [Bodo Moeller]
8308
8309 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
8310 sets the subject name for a new request or supersedes the
8311 subject name in a given request. Formats that can be parsed are
8312 'CN=Some Name, OU=myOU, C=IT'
8313 and
8314 'CN=Some Name/OU=myOU/C=IT'.
8315
8316 Add options '-batch' and '-verbose' to 'openssl req'.
8317 [Massimiliano Pala <madwolf@hackmasters.net>]
8318
8319 *) Introduce the possibility to access global variables through
8320 functions on platform were that's the best way to handle exporting
8321 global variables in shared libraries. To enable this functionality,
8322 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
8323 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
8324 is normally done by Configure or something similar).
8325
8326 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
8327 in the source file (foo.c) like this:
8328
8329 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
8330 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
8331
8332 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
8333 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
8334
8335 OPENSSL_DECLARE_GLOBAL(int,foo);
8336 #define foo OPENSSL_GLOBAL_REF(foo)
8337 OPENSSL_DECLARE_GLOBAL(double,bar);
8338 #define bar OPENSSL_GLOBAL_REF(bar)
8339
8340 The #defines are very important, and therefore so is including the
8341 header file everywhere where the defined globals are used.
8342
8343 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
8344 of ASN.1 items, but that structure is a bit different.
8345
8346 The largest change is in util/mkdef.pl which has been enhanced with
8347 better and easier to understand logic to choose which symbols should
8348 go into the Windows .def files as well as a number of fixes and code
8349 cleanup (among others, algorithm keywords are now sorted
8350 lexicographically to avoid constant rewrites).
8351 [Richard Levitte]
8352
8353 *) In BN_div() keep a copy of the sign of 'num' before writing the
8354 result to 'rm' because if rm==num the value will be overwritten
8355 and produce the wrong result if 'num' is negative: this caused
8356 problems with BN_mod() and BN_nnmod().
8357 [Steve Henson]
8358
8359 *) Function OCSP_request_verify(). This checks the signature on an
8360 OCSP request and verifies the signer certificate. The signer
8361 certificate is just checked for a generic purpose and OCSP request
8362 trust settings.
8363 [Steve Henson]
8364
8365 *) Add OCSP_check_validity() function to check the validity of OCSP
8366 responses. OCSP responses are prepared in real time and may only
8367 be a few seconds old. Simply checking that the current time lies
8368 between thisUpdate and nextUpdate max reject otherwise valid responses
8369 caused by either OCSP responder or client clock inaccuracy. Instead
8370 we allow thisUpdate and nextUpdate to fall within a certain period of
8371 the current time. The age of the response can also optionally be
8372 checked. Two new options -validity_period and -status_age added to
8373 ocsp utility.
8374 [Steve Henson]
8375
8376 *) If signature or public key algorithm is unrecognized print out its
8377 OID rather that just UNKNOWN.
8378 [Steve Henson]
8379
8380 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
8381 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
8382 ID to be generated from the issuer certificate alone which can then be
8383 passed to OCSP_id_issuer_cmp().
8384 [Steve Henson]
8385
8386 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
8387 ASN1 modules to export functions returning ASN1_ITEM pointers
8388 instead of the ASN1_ITEM structures themselves. This adds several
8389 new macros which allow the underlying ASN1 function/structure to
8390 be accessed transparently. As a result code should not use ASN1_ITEM
8391 references directly (such as &X509_it) but instead use the relevant
8392 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
8393 use of the new ASN1 code on platforms where exporting structures
8394 is problematical (for example in shared libraries) but exporting
8395 functions returning pointers to structures is not.
8396 [Steve Henson]
8397
8398 *) Add support for overriding the generation of SSL/TLS session IDs.
8399 These callbacks can be registered either in an SSL_CTX or per SSL.
8400 The purpose of this is to allow applications to control, if they wish,
8401 the arbitrary values chosen for use as session IDs, particularly as it
8402 can be useful for session caching in multiple-server environments. A
8403 command-line switch for testing this (and any client code that wishes
8404 to use such a feature) has been added to "s_server".
8405 [Geoff Thorpe, Lutz Jaenicke]
8406
8407 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
8408 of the form '#if defined(...) || defined(...) || ...' and
8409 '#if !defined(...) && !defined(...) && ...'. This also avoids
8410 the growing number of special cases it was previously handling.
8411 [Richard Levitte]
8412
8413 *) Make all configuration macros available for application by making
8414 sure they are available in opensslconf.h, by giving them names starting
8415 with "OPENSSL_" to avoid conflicts with other packages and by making
8416 sure e_os2.h will cover all platform-specific cases together with
8417 opensslconf.h.
8418 Additionally, it is now possible to define configuration/platform-
8419 specific names (called "system identities"). In the C code, these
8420 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
8421 macro with the name beginning with "OPENSSL_SYS_", which is determined
8422 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
8423 what is available.
8424 [Richard Levitte]
8425
8426 *) New option -set_serial to 'req' and 'x509' this allows the serial
8427 number to use to be specified on the command line. Previously self
8428 signed certificates were hard coded with serial number 0 and the
8429 CA options of 'x509' had to use a serial number in a file which was
8430 auto incremented.
8431 [Steve Henson]
8432
8433 *) New options to 'ca' utility to support V2 CRL entry extensions.
8434 Currently CRL reason, invalidity date and hold instruction are
8435 supported. Add new CRL extensions to V3 code and some new objects.
8436 [Steve Henson]
8437
8438 *) New function EVP_CIPHER_CTX_set_padding() this is used to
8439 disable standard block padding (aka PKCS#5 padding) in the EVP
8440 API, which was previously mandatory. This means that the data is
8441 not padded in any way and so the total length much be a multiple
8442 of the block size, otherwise an error occurs.
8443 [Steve Henson]
8444
8445 *) Initial (incomplete) OCSP SSL support.
8446 [Steve Henson]
8447
8448 *) New function OCSP_parse_url(). This splits up a URL into its host,
8449 port and path components: primarily to parse OCSP URLs. New -url
8450 option to ocsp utility.
8451 [Steve Henson]
8452
8453 *) New nonce behavior. The return value of OCSP_check_nonce() now
8454 reflects the various checks performed. Applications can decide
8455 whether to tolerate certain situations such as an absent nonce
8456 in a response when one was present in a request: the ocsp application
8457 just prints out a warning. New function OCSP_add1_basic_nonce()
8458 this is to allow responders to include a nonce in a response even if
8459 the request is nonce-less.
8460 [Steve Henson]
8461
8462 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
8463 skipped when using openssl x509 multiple times on a single input file,
8464 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
8465 [Bodo Moeller]
8466
8467 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
8468 set string type: to handle setting ASN1_TIME structures. Fix ca
8469 utility to correctly initialize revocation date of CRLs.
8470 [Steve Henson]
8471
8472 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
8473 the clients preferred ciphersuites and rather use its own preferences.
8474 Should help to work around M$ SGC (Server Gated Cryptography) bug in
8475 Internet Explorer by ensuring unchanged hash method during stepup.
8476 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
8477 [Lutz Jaenicke]
8478
8479 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
8480 to aes and add a new 'exist' option to print out symbols that don't
8481 appear to exist.
8482 [Steve Henson]
8483
8484 *) Additional options to ocsp utility to allow flags to be set and
8485 additional certificates supplied.
8486 [Steve Henson]
8487
8488 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8489 OCSP client a number of certificate to only verify the response
8490 signature against.
8491 [Richard Levitte]
8492
8493 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8494 handle the new API. Currently only ECB, CBC modes supported. Add new
8495 AES OIDs.
8496
8497 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8498 Encryption Standard (AES) Ciphersuites for Transport Layer
8499 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8500 not enabled by default and were not part of the "ALL" ciphersuite
8501 alias because they were not yet official; they could be
8502 explicitly requested by specifying the "AESdraft" ciphersuite
8503 group alias. In the final release of OpenSSL 0.9.7, the group
8504 alias is called "AES" and is part of "ALL".)
8505 [Ben Laurie, Steve Henson, Bodo Moeller]
8506
8507 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8508 request to response.
8509 [Steve Henson]
8510
8511 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8512 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8513 extract information from a certificate request. OCSP_response_create()
8514 creates a response and optionally adds a basic response structure.
8515 OCSP_basic_add1_status() adds a complete single response to a basic
8516 response and returns the OCSP_SINGLERESP structure just added (to allow
8517 extensions to be included for example). OCSP_basic_add1_cert() adds a
8518 certificate to a basic response and OCSP_basic_sign() signs a basic
8519 response with various flags. New helper functions ASN1_TIME_check()
8520 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8521 (converts ASN1_TIME to GeneralizedTime).
8522 [Steve Henson]
8523
8524 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8525 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8526 structure from a certificate. X509_pubkey_digest() digests the public_key
8527 contents: this is used in various key identifiers.
8528 [Steve Henson]
8529
8530 *) Make sk_sort() tolerate a NULL argument.
8531 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8532
8533 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8534 passed by the function are trusted implicitly. If any of them signed the
8535 response then it is assumed to be valid and is not verified.
8536 [Steve Henson]
8537
8538 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8539 to data. This was previously part of the PKCS7 ASN1 code. This
8540 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8541 [Steve Henson, reported by Kenneth R. Robinette
8542 <support@securenetterm.com>]
8543
8544 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8545 routines: without these tracing memory leaks is very painful.
8546 Fix leaks in PKCS12 and PKCS7 routines.
8547 [Steve Henson]
8548
8549 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8550 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8551 effectively meant GeneralizedTime would never be used. Now it
8552 is initialised to -1 but X509_time_adj() now has to check the value
8553 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8554 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8555 [Steve Henson, reported by Kenneth R. Robinette
8556 <support@securenetterm.com>]
8557
8558 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8559 result in a zero length in the ASN1_INTEGER structure which was
8560 not consistent with the structure when d2i_ASN1_INTEGER() was used
8561 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8562 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8563 where it did not print out a minus for negative ASN1_INTEGER.
8564 [Steve Henson]
8565
8566 *) Add summary printout to ocsp utility. The various functions which
8567 convert status values to strings have been renamed to:
8568 OCSP_response_status_str(), OCSP_cert_status_str() and
8569 OCSP_crl_reason_str() and are no longer static. New options
8570 to verify nonce values and to disable verification. OCSP response
8571 printout format cleaned up.
8572 [Steve Henson]
8573
8574 *) Add additional OCSP certificate checks. These are those specified
8575 in RFC2560. This consists of two separate checks: the CA of the
8576 certificate being checked must either be the OCSP signer certificate
8577 or the issuer of the OCSP signer certificate. In the latter case the
8578 OCSP signer certificate must contain the OCSP signing extended key
8579 usage. This check is performed by attempting to match the OCSP
8580 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8581 in the OCSP_CERTID structures of the response.
8582 [Steve Henson]
8583
8584 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8585 and related routines. This uses the standard OpenSSL certificate
8586 verify routines to perform initial checks (just CA validity) and
8587 to obtain the certificate chain. Then additional checks will be
8588 performed on the chain. Currently the root CA is checked to see
8589 if it is explicitly trusted for OCSP signing. This is used to set
8590 a root CA as a global signing root: that is any certificate that
8591 chains to that CA is an acceptable OCSP signing certificate.
8592 [Steve Henson]
8593
8594 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8595 extensions from a separate configuration file.
8596 As when reading extensions from the main configuration file,
8597 the '-extensions ...' option may be used for specifying the
8598 section to use.
8599 [Massimiliano Pala <madwolf@comune.modena.it>]
8600
8601 *) New OCSP utility. Allows OCSP requests to be generated or
8602 read. The request can be sent to a responder and the output
8603 parsed, outputed or printed in text form. Not complete yet:
8604 still needs to check the OCSP response validity.
8605 [Steve Henson]
8606
8607 *) New subcommands for 'openssl ca':
8608 'openssl ca -status <serial>' prints the status of the cert with
8609 the given serial number (according to the index file).
8610 'openssl ca -updatedb' updates the expiry status of certificates
8611 in the index file.
8612 [Massimiliano Pala <madwolf@comune.modena.it>]
8613
8614 *) New '-newreq-nodes' command option to CA.pl. This is like
8615 '-newreq', but calls 'openssl req' with the '-nodes' option
8616 so that the resulting key is not encrypted.
8617 [Damien Miller <djm@mindrot.org>]
8618
8619 *) New configuration for the GNU Hurd.
8620 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8621
8622 *) Initial code to implement OCSP basic response verify. This
8623 is currently incomplete. Currently just finds the signer's
8624 certificate and verifies the signature on the response.
8625 [Steve Henson]
8626
8627 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
8628 value of OPENSSLDIR. This is available via the new '-d' option
8629 to 'openssl version', and is also included in 'openssl version -a'.
8630 [Bodo Moeller]
8631
8632 *) Allowing defining memory allocation callbacks that will be given
8633 file name and line number information in additional arguments
8634 (a const char* and an int). The basic functionality remains, as
8635 well as the original possibility to just replace malloc(),
8636 realloc() and free() by functions that do not know about these
8637 additional arguments. To register and find out the current
8638 settings for extended allocation functions, the following
8639 functions are provided:
8640
8641 CRYPTO_set_mem_ex_functions
8642 CRYPTO_set_locked_mem_ex_functions
8643 CRYPTO_get_mem_ex_functions
8644 CRYPTO_get_locked_mem_ex_functions
8645
8646 These work the same way as CRYPTO_set_mem_functions and friends.
8647 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
8648 extended allocation function is enabled.
8649 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
8650 a conventional allocation function is enabled.
8651 [Richard Levitte, Bodo Moeller]
8652
8653 *) Finish off removing the remaining LHASH function pointer casts.
8654 There should no longer be any prototype-casting required when using
8655 the LHASH abstraction, and any casts that remain are "bugs". See
8656 the callback types and macros at the head of lhash.h for details
8657 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
8658 [Geoff Thorpe]
8659
8660 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
8661 If /dev/[u]random devices are not available or do not return enough
8662 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
8663 be queried.
8664 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
8665 /etc/entropy will be queried once each in this sequence, querying stops
8666 when enough entropy was collected without querying more sockets.
8667 [Lutz Jaenicke]
8668
8669 *) Change the Unix RAND_poll() variant to be able to poll several
8670 random devices, as specified by DEVRANDOM, until a sufficient amount
8671 of data has been collected. We spend at most 10 ms on each file
8672 (select timeout) and read in non-blocking mode. DEVRANDOM now
8673 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
8674 (previously it was just the string "/dev/urandom"), so on typical
8675 platforms the 10 ms delay will never occur.
8676 Also separate out the Unix variant to its own file, rand_unix.c.
8677 For VMS, there's a currently-empty rand_vms.c.
8678 [Richard Levitte]
8679
8680 *) Move OCSP client related routines to ocsp_cl.c. These
8681 provide utility functions which an application needing
8682 to issue a request to an OCSP responder and analyse the
8683 response will typically need: as opposed to those which an
8684 OCSP responder itself would need which will be added later.
8685
8686 OCSP_request_sign() signs an OCSP request with an API similar
8687 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
8688 response. OCSP_response_get1_basic() extracts basic response
8689 from response. OCSP_resp_find_status(): finds and extracts status
8690 information from an OCSP_CERTID structure (which will be created
8691 when the request structure is built). These are built from lower
8692 level functions which work on OCSP_SINGLERESP structures but
8693 won't normally be used unless the application wishes to examine
8694 extensions in the OCSP response for example.
8695
8696 Replace nonce routines with a pair of functions.
8697 OCSP_request_add1_nonce() adds a nonce value and optionally
8698 generates a random value. OCSP_check_nonce() checks the
8699 validity of the nonce in an OCSP response.
8700 [Steve Henson]
8701
8702 *) Change function OCSP_request_add() to OCSP_request_add0_id().
8703 This doesn't copy the supplied OCSP_CERTID and avoids the
8704 need to free up the newly created id. Change return type
8705 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
8706 This can then be used to add extensions to the request.
8707 Deleted OCSP_request_new(), since most of its functionality
8708 is now in OCSP_REQUEST_new() (and the case insensitive name
8709 clash) apart from the ability to set the request name which
8710 will be added elsewhere.
8711 [Steve Henson]
8712
8713 *) Update OCSP API. Remove obsolete extensions argument from
8714 various functions. Extensions are now handled using the new
8715 OCSP extension code. New simple OCSP HTTP function which
8716 can be used to send requests and parse the response.
8717 [Steve Henson]
8718
8719 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
8720 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
8721 uses the special reorder version of SET OF to sort the attributes
8722 and reorder them to match the encoded order. This resolves a long
8723 standing problem: a verify on a PKCS7 structure just after signing
8724 it used to fail because the attribute order did not match the
8725 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
8726 it uses the received order. This is necessary to tolerate some broken
8727 software that does not order SET OF. This is handled by encoding
8728 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
8729 to produce the required SET OF.
8730 [Steve Henson]
8731
8732 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
8733 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8734 files to get correct declarations of the ASN.1 item variables.
8735 [Richard Levitte]
8736
8737 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
8738 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
8739 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
8740 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
8741 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
8742 ASN1_ITEM and no wrapper functions.
8743 [Steve Henson]
8744
8745 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
8746 replace the old function pointer based I/O routines. Change most of
8747 the *_d2i_bio() and *_d2i_fp() functions to use these.
8748 [Steve Henson]
8749
8750 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
8751 lines, recognize more "algorithms" that can be deselected, and make
8752 it complain about algorithm deselection that isn't recognised.
8753 [Richard Levitte]
8754
8755 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
8756 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
8757 to use new functions. Add NO_ASN1_OLD which can be set to remove
8758 some old style ASN1 functions: this can be used to determine if old
8759 code will still work when these eventually go away.
8760 [Steve Henson]
8761
8762 *) New extension functions for OCSP structures, these follow the
8763 same conventions as certificates and CRLs.
8764 [Steve Henson]
8765
8766 *) New function X509V3_add1_i2d(). This automatically encodes and
8767 adds an extension. Its behaviour can be customised with various
8768 flags to append, replace or delete. Various wrappers added for
8769 certificates and CRLs.
8770 [Steve Henson]
8771
8772 *) Fix to avoid calling the underlying ASN1 print routine when
8773 an extension cannot be parsed. Correct a typo in the
8774 OCSP_SERVICELOC extension. Tidy up print OCSP format.
8775 [Steve Henson]
8776
8777 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
8778 entries for variables.
8779 [Steve Henson]
8780
8781 *) Add functionality to apps/openssl.c for detecting locking
8782 problems: As the program is single-threaded, all we have
8783 to do is register a locking callback using an array for
8784 storing which locks are currently held by the program.
8785 [Bodo Moeller]
8786
8787 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8788 SSL_get_ex_data_X509_STORE_idx(), which is used in
8789 ssl_verify_cert_chain() and thus can be called at any time
8790 during TLS/SSL handshakes so that thread-safety is essential.
8791 Unfortunately, the ex_data design is not at all suited
8792 for multi-threaded use, so it probably should be abolished.
8793 [Bodo Moeller]
8794
8795 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8796 [Broadcom, tweaked and integrated by Geoff Thorpe]
8797
8798 *) Move common extension printing code to new function
8799 X509V3_print_extensions(). Reorganise OCSP print routines and
8800 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8801 [Steve Henson]
8802
8803 *) New function X509_signature_print() to remove duplication in some
8804 print routines.
8805 [Steve Henson]
8806
8807 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8808 set (this was treated exactly the same as SET OF previously). This
8809 is used to reorder the STACK representing the structure to match the
8810 encoding. This will be used to get round a problem where a PKCS7
8811 structure which was signed could not be verified because the STACK
8812 order did not reflect the encoded order.
8813 [Steve Henson]
8814
8815 *) Reimplement the OCSP ASN1 module using the new code.
8816 [Steve Henson]
8817
8818 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
8819 for its ASN1 operations. The old style function pointers still exist
8820 for now but they will eventually go away.
8821 [Steve Henson]
8822
8823 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
8824 completely replaces the old ASN1 functionality with a table driven
8825 encoder and decoder which interprets an ASN1_ITEM structure describing
8826 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
8827 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
8828 has also been converted to the new form.
8829 [Steve Henson]
8830
8831 *) Change BN_mod_exp_recp so that negative moduli are tolerated
8832 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
8833 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
8834 for negative moduli.
8835 [Bodo Moeller]
8836
8837 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
8838 of not touching the result's sign bit.
8839 [Bodo Moeller]
8840
8841 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
8842 set.
8843 [Bodo Moeller]
8844
8845 *) Changed the LHASH code to use prototypes for callbacks, and created
8846 macros to declare and implement thin (optionally static) functions
8847 that provide type-safety and avoid function pointer casting for the
8848 type-specific callbacks.
8849 [Geoff Thorpe]
8850
8851 *) Added Kerberos Cipher Suites to be used with TLS, as written in
8852 RFC 2712.
8853 [Veers Staats <staatsvr@asc.hpc.mil>,
8854 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
8855
8856 *) Reformat the FAQ so the different questions and answers can be divided
8857 in sections depending on the subject.
8858 [Richard Levitte]
8859
8860 *) Have the zlib compression code load ZLIB.DLL dynamically under
8861 Windows.
8862 [Richard Levitte]
8863
8864 *) New function BN_mod_sqrt for computing square roots modulo a prime
8865 (using the probabilistic Tonelli-Shanks algorithm unless
8866 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
8867 be handled deterministically).
8868 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8869
8870 *) Make BN_mod_inverse faster by explicitly handling small quotients
8871 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
8872 512 bits], about 30% for larger ones [1024 or 2048 bits].)
8873 [Bodo Moeller]
8874
8875 *) New function BN_kronecker.
8876 [Bodo Moeller]
8877
8878 *) Fix BN_gcd so that it works on negative inputs; the result is
8879 positive unless both parameters are zero.
8880 Previously something reasonably close to an infinite loop was
8881 possible because numbers could be growing instead of shrinking
8882 in the implementation of Euclid's algorithm.
8883 [Bodo Moeller]
8884
8885 *) Fix BN_is_word() and BN_is_one() macros to take into account the
8886 sign of the number in question.
8887
8888 Fix BN_is_word(a,w) to work correctly for w == 0.
8889
8890 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
8891 because its test if the absolute value of 'a' equals 'w'.
8892 Note that BN_abs_is_word does *not* handle w == 0 reliably;
8893 it exists mostly for use in the implementations of BN_is_zero(),
8894 BN_is_one(), and BN_is_word().
8895 [Bodo Moeller]
8896
8897 *) New function BN_swap.
8898 [Bodo Moeller]
8899
8900 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
8901 the exponentiation functions are more likely to produce reasonable
8902 results on negative inputs.
8903 [Bodo Moeller]
8904
8905 *) Change BN_mod_mul so that the result is always non-negative.
8906 Previously, it could be negative if one of the factors was negative;
8907 I don't think anyone really wanted that behaviour.
8908 [Bodo Moeller]
8909
8910 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8911 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8912 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8913 and add new functions:
8914
8915 BN_nnmod
8916 BN_mod_sqr
8917 BN_mod_add
8918 BN_mod_add_quick
8919 BN_mod_sub
8920 BN_mod_sub_quick
8921 BN_mod_lshift1
8922 BN_mod_lshift1_quick
8923 BN_mod_lshift
8924 BN_mod_lshift_quick
8925
8926 These functions always generate non-negative results.
8927
8928 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8929 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8930
8931 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8932 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8933 be reduced modulo m.
8934 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8935
8936 #if 0
8937 The following entry accidentally appeared in the CHANGES file
8938 distributed with OpenSSL 0.9.7. The modifications described in
8939 it do *not* apply to OpenSSL 0.9.7.
8940
8941 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8942 was actually never needed) and in BN_mul(). The removal in BN_mul()
8943 required a small change in bn_mul_part_recursive() and the addition
8944 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8945 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8946 bn_sub_words() and bn_add_words() except they take arrays with
8947 differing sizes.
8948 [Richard Levitte]
8949 #endif
8950
8951 *) In 'openssl passwd', verify passwords read from the terminal
8952 unless the '-salt' option is used (which usually means that
8953 verification would just waste user's time since the resulting
8954 hash is going to be compared with some given password hash)
8955 or the new '-noverify' option is used.
8956
8957 This is an incompatible change, but it does not affect
8958 non-interactive use of 'openssl passwd' (passwords on the command
8959 line, '-stdin' option, '-in ...' option) and thus should not
8960 cause any problems.
8961 [Bodo Moeller]
8962
8963 *) Remove all references to RSAref, since there's no more need for it.
8964 [Richard Levitte]
8965
8966 *) Make DSO load along a path given through an environment variable
8967 (SHLIB_PATH) with shl_load().
8968 [Richard Levitte]
8969
8970 *) Constify the ENGINE code as a result of BIGNUM constification.
8971 Also constify the RSA code and most things related to it. In a
8972 few places, most notable in the depth of the ASN.1 code, ugly
8973 casts back to non-const were required (to be solved at a later
8974 time)
8975 [Richard Levitte]
8976
8977 *) Make it so the openssl application has all engines loaded by default.
8978 [Richard Levitte]
8979
8980 *) Constify the BIGNUM routines a little more.
8981 [Richard Levitte]
8982
8983 *) Add the following functions:
8984
8985 ENGINE_load_cswift()
8986 ENGINE_load_chil()
8987 ENGINE_load_atalla()
8988 ENGINE_load_nuron()
8989 ENGINE_load_builtin_engines()
8990
8991 That way, an application can itself choose if external engines that
8992 are built-in in OpenSSL shall ever be used or not. The benefit is
8993 that applications won't have to be linked with libdl or other dso
8994 libraries unless it's really needed.
8995
8996 Changed 'openssl engine' to load all engines on demand.
8997 Changed the engine header files to avoid the duplication of some
8998 declarations (they differed!).
8999 [Richard Levitte]
9000
9001 *) 'openssl engine' can now list capabilities.
9002 [Richard Levitte]
9003
9004 *) Better error reporting in 'openssl engine'.
9005 [Richard Levitte]
9006
9007 *) Never call load_dh_param(NULL) in s_server.
9008 [Bodo Moeller]
9009
9010 *) Add engine application. It can currently list engines by name and
9011 identity, and test if they are actually available.
9012 [Richard Levitte]
9013
9014 *) Improve RPM specification file by forcing symbolic linking and making
9015 sure the installed documentation is also owned by root.root.
9016 [Damien Miller <djm@mindrot.org>]
9017
9018 *) Give the OpenSSL applications more possibilities to make use of
9019 keys (public as well as private) handled by engines.
9020 [Richard Levitte]
9021
9022 *) Add OCSP code that comes from CertCo.
9023 [Richard Levitte]
9024
9025 *) Add VMS support for the Rijndael code.
9026 [Richard Levitte]
9027
9028 *) Added untested support for Nuron crypto accelerator.
9029 [Ben Laurie]
9030
9031 *) Add support for external cryptographic devices. This code was
9032 previously distributed separately as the "engine" branch.
9033 [Geoff Thorpe, Richard Levitte]
9034
9035 *) Rework the filename-translation in the DSO code. It is now possible to
9036 have far greater control over how a "name" is turned into a filename
9037 depending on the operating environment and any oddities about the
9038 different shared library filenames on each system.
9039 [Geoff Thorpe]
9040
9041 *) Support threads on FreeBSD-elf in Configure.
9042 [Richard Levitte]
9043
9044 *) Fix for SHA1 assembly problem with MASM: it produces
9045 warnings about corrupt line number information when assembling
9046 with debugging information. This is caused by the overlapping
9047 of two sections.
9048 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
9049
9050 *) NCONF changes.
9051 NCONF_get_number() has no error checking at all. As a replacement,
9052 NCONF_get_number_e() is defined (_e for "error checking") and is
9053 promoted strongly. The old NCONF_get_number is kept around for
9054 binary backward compatibility.
9055 Make it possible for methods to load from something other than a BIO,
9056 by providing a function pointer that is given a name instead of a BIO.
9057 For example, this could be used to load configuration data from an
9058 LDAP server.
9059 [Richard Levitte]
9060
9061 *) Fix for non blocking accept BIOs. Added new I/O special reason
9062 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
9063 with non blocking I/O was not possible because no retry code was
9064 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
9065 this case.
9066 [Steve Henson]
9067
9068 *) Added the beginnings of Rijndael support.
9069 [Ben Laurie]
9070
9071 *) Fix for bug in DirectoryString mask setting. Add support for
9072 X509_NAME_print_ex() in 'req' and X509_print_ex() function
9073 to allow certificate printing to more controllable, additional
9074 'certopt' option to 'x509' to allow new printing options to be
9075 set.
9076 [Steve Henson]
9077
9078 *) Clean old EAY MD5 hack from e_os.h.
9079 [Richard Levitte]
9080
9081 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
9082
9083 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
9084 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
9085 [Joe Orton, Steve Henson]
9086
9087 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
9088
9089 *) Fix additional bug revealed by the NISCC test suite:
9090
9091 Stop bug triggering large recursion when presented with
9092 certain ASN.1 tags (CVE-2003-0851)
9093 [Steve Henson]
9094
9095 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
9096
9097 *) Fix various bugs revealed by running the NISCC test suite:
9098
9099 Stop out of bounds reads in the ASN1 code when presented with
9100 invalid tags (CVE-2003-0543 and CVE-2003-0544).
9101
9102 If verify callback ignores invalid public key errors don't try to check
9103 certificate signature with the NULL public key.
9104
9105 [Steve Henson]
9106
9107 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
9108 if the server requested one: as stated in TLS 1.0 and SSL 3.0
9109 specifications.
9110 [Steve Henson]
9111
9112 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
9113 extra data after the compression methods not only for TLS 1.0
9114 but also for SSL 3.0 (as required by the specification).
9115 [Bodo Moeller; problem pointed out by Matthias Loepfe]
9116
9117 *) Change X509_certificate_type() to mark the key as exported/exportable
9118 when it's 512 *bits* long, not 512 bytes.
9119 [Richard Levitte]
9120
9121 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
9122
9123 *) Countermeasure against the Klima-Pokorny-Rosa extension of
9124 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
9125 a protocol version number mismatch like a decryption error
9126 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
9127 [Bodo Moeller]
9128
9129 *) Turn on RSA blinding by default in the default implementation
9130 to avoid a timing attack. Applications that don't want it can call
9131 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
9132 They would be ill-advised to do so in most cases.
9133 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
9134
9135 *) Change RSA blinding code so that it works when the PRNG is not
9136 seeded (in this case, the secret RSA exponent is abused as
9137 an unpredictable seed -- if it is not unpredictable, there
9138 is no point in blinding anyway). Make RSA blinding thread-safe
9139 by remembering the creator's thread ID in rsa->blinding and
9140 having all other threads use local one-time blinding factors
9141 (this requires more computation than sharing rsa->blinding, but
9142 avoids excessive locking; and if an RSA object is not shared
9143 between threads, blinding will still be very fast).
9144 [Bodo Moeller]
9145
9146 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
9147
9148 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
9149 via timing by performing a MAC computation even if incorrect
9150 block cipher padding has been found. This is a countermeasure
9151 against active attacks where the attacker has to distinguish
9152 between bad padding and a MAC verification error. (CVE-2003-0078)
9153
9154 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
9155 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
9156 Martin Vuagnoux (EPFL, Ilion)]
9157
9158 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
9159
9160 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
9161 memory from its contents. This is done with a counter that will
9162 place alternating values in each byte. This can be used to solve
9163 two issues: 1) the removal of calls to memset() by highly optimizing
9164 compilers, and 2) cleansing with other values than 0, since those can
9165 be read through on certain media, for example a swap space on disk.
9166 [Geoff Thorpe]
9167
9168 *) Bugfix: client side session caching did not work with external caching,
9169 because the session->cipher setting was not restored when reloading
9170 from the external cache. This problem was masked, when
9171 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
9172 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
9173 [Lutz Jaenicke]
9174
9175 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
9176 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
9177 [Zeev Lieber <zeev-l@yahoo.com>]
9178
9179 *) Undo an undocumented change introduced in 0.9.6e which caused
9180 repeated calls to OpenSSL_add_all_ciphers() and
9181 OpenSSL_add_all_digests() to be ignored, even after calling
9182 EVP_cleanup().
9183 [Richard Levitte]
9184
9185 *) Change the default configuration reader to deal with last line not
9186 being properly terminated.
9187 [Richard Levitte]
9188
9189 *) Change X509_NAME_cmp() so it applies the special rules on handling
9190 DN values that are of type PrintableString, as well as RDNs of type
9191 emailAddress where the value has the type ia5String.
9192 [stefank@valicert.com via Richard Levitte]
9193
9194 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
9195 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
9196 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
9197 the bitwise-OR of the two for use by the majority of applications
9198 wanting this behaviour, and update the docs. The documented
9199 behaviour and actual behaviour were inconsistent and had been
9200 changing anyway, so this is more a bug-fix than a behavioural
9201 change.
9202 [Geoff Thorpe, diagnosed by Nadav Har'El]
9203
9204 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
9205 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
9206 [Bodo Moeller]
9207
9208 *) Fix initialization code race conditions in
9209 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
9210 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
9211 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
9212 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
9213 ssl2_get_cipher_by_char(),
9214 ssl3_get_cipher_by_char().
9215 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
9216
9217 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
9218 the cached sessions are flushed, as the remove_cb() might use ex_data
9219 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
9220 (see [openssl.org #212]).
9221 [Geoff Thorpe, Lutz Jaenicke]
9222
9223 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
9224 length, instead of the encoding length to d2i_ASN1_OBJECT.
9225 [Steve Henson]
9226
9227 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
9228
9229 *) [In 0.9.6g-engine release:]
9230 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
9231 [Lynn Gazis <lgazis@rainbow.com>]
9232
9233 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
9234
9235 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
9236 and get fix the header length calculation.
9237 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
9238 Alon Kantor <alonk@checkpoint.com> (and others),
9239 Steve Henson]
9240
9241 *) Use proper error handling instead of 'assertions' in buffer
9242 overflow checks added in 0.9.6e. This prevents DoS (the
9243 assertions could call abort()).
9244 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
9245
9246 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
9247
9248 *) Add various sanity checks to asn1_get_length() to reject
9249 the ASN1 length bytes if they exceed sizeof(long), will appear
9250 negative or the content length exceeds the length of the
9251 supplied buffer.
9252 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9253
9254 *) Fix cipher selection routines: ciphers without encryption had no flags
9255 for the cipher strength set and where therefore not handled correctly
9256 by the selection routines (PR #130).
9257 [Lutz Jaenicke]
9258
9259 *) Fix EVP_dsa_sha macro.
9260 [Nils Larsch]
9261
9262 *) New option
9263 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
9264 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
9265 that was added in OpenSSL 0.9.6d.
9266
9267 As the countermeasure turned out to be incompatible with some
9268 broken SSL implementations, the new option is part of SSL_OP_ALL.
9269 SSL_OP_ALL is usually employed when compatibility with weird SSL
9270 implementations is desired (e.g. '-bugs' option to 's_client' and
9271 's_server'), so the new option is automatically set in many
9272 applications.
9273 [Bodo Moeller]
9274
9275 *) Changes in security patch:
9276
9277 Changes marked "(CHATS)" were sponsored by the Defense Advanced
9278 Research Projects Agency (DARPA) and Air Force Research Laboratory,
9279 Air Force Materiel Command, USAF, under agreement number
9280 F30602-01-2-0537.
9281
9282 *) Add various sanity checks to asn1_get_length() to reject
9283 the ASN1 length bytes if they exceed sizeof(long), will appear
9284 negative or the content length exceeds the length of the
9285 supplied buffer. (CVE-2002-0659)
9286 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9287
9288 *) Assertions for various potential buffer overflows, not known to
9289 happen in practice.
9290 [Ben Laurie (CHATS)]
9291
9292 *) Various temporary buffers to hold ASCII versions of integers were
9293 too small for 64 bit platforms. (CVE-2002-0655)
9294 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
9295
9296 *) Remote buffer overflow in SSL3 protocol - an attacker could
9297 supply an oversized session ID to a client. (CVE-2002-0656)
9298 [Ben Laurie (CHATS)]
9299
9300 *) Remote buffer overflow in SSL2 protocol - an attacker could
9301 supply an oversized client master key. (CVE-2002-0656)
9302 [Ben Laurie (CHATS)]
9303
9304 Changes between 0.9.6c and 0.9.6d [9 May 2002]
9305
9306 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
9307 encoded as NULL) with id-dsa-with-sha1.
9308 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
9309
9310 *) Check various X509_...() return values in apps/req.c.
9311 [Nils Larsch <nla@trustcenter.de>]
9312
9313 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
9314 an end-of-file condition would erroneously be flagged, when the CRLF
9315 was just at the end of a processed block. The bug was discovered when
9316 processing data through a buffering memory BIO handing the data to a
9317 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
9318 <ptsekov@syntrex.com> and Nedelcho Stanev.
9319 [Lutz Jaenicke]
9320
9321 *) Implement a countermeasure against a vulnerability recently found
9322 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
9323 before application data chunks to avoid the use of known IVs
9324 with data potentially chosen by the attacker.
9325 [Bodo Moeller]
9326
9327 *) Fix length checks in ssl3_get_client_hello().
9328 [Bodo Moeller]
9329
9330 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
9331 to prevent ssl3_read_internal() from incorrectly assuming that
9332 ssl3_read_bytes() found application data while handshake
9333 processing was enabled when in fact s->s3->in_read_app_data was
9334 merely automatically cleared during the initial handshake.
9335 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
9336
9337 *) Fix object definitions for Private and Enterprise: they were not
9338 recognized in their shortname (=lowercase) representation. Extend
9339 obj_dat.pl to issue an error when using undefined keywords instead
9340 of silently ignoring the problem (Svenning Sorensen
9341 <sss@sss.dnsalias.net>).
9342 [Lutz Jaenicke]
9343
9344 *) Fix DH_generate_parameters() so that it works for 'non-standard'
9345 generators, i.e. generators other than 2 and 5. (Previously, the
9346 code did not properly initialise the 'add' and 'rem' values to
9347 BN_generate_prime().)
9348
9349 In the new general case, we do not insist that 'generator' is
9350 actually a primitive root: This requirement is rather pointless;
9351 a generator of the order-q subgroup is just as good, if not
9352 better.
9353 [Bodo Moeller]
9354
9355 *) Map new X509 verification errors to alerts. Discovered and submitted by
9356 Tom Wu <tom@arcot.com>.
9357 [Lutz Jaenicke]
9358
9359 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
9360 returning non-zero before the data has been completely received
9361 when using non-blocking I/O.
9362 [Bodo Moeller; problem pointed out by John Hughes]
9363
9364 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
9365 [Ben Laurie, Lutz Jaenicke]
9366
9367 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
9368 Yoram Zahavi <YoramZ@gilian.com>).
9369 [Lutz Jaenicke]
9370
9371 *) Add information about CygWin 1.3 and on, and preserve proper
9372 configuration for the versions before that.
9373 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
9374
9375 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
9376 check whether we deal with a copy of a session and do not delete from
9377 the cache in this case. Problem reported by "Izhar Shoshani Levi"
9378 <izhar@checkpoint.com>.
9379 [Lutz Jaenicke]
9380
9381 *) Do not store session data into the internal session cache, if it
9382 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
9383 flag is set). Proposed by Aslam <aslam@funk.com>.
9384 [Lutz Jaenicke]
9385
9386 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
9387 value is 0.
9388 [Richard Levitte]
9389
9390 *) [In 0.9.6d-engine release:]
9391 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
9392 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
9393
9394 *) Add the configuration target linux-s390x.
9395 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
9396
9397 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
9398 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
9399 variable as an indication that a ClientHello message has been
9400 received. As the flag value will be lost between multiple
9401 invocations of ssl3_accept when using non-blocking I/O, the
9402 function may not be aware that a handshake has actually taken
9403 place, thus preventing a new session from being added to the
9404 session cache.
9405
9406 To avoid this problem, we now set s->new_session to 2 instead of
9407 using a local variable.
9408 [Lutz Jaenicke, Bodo Moeller]
9409
9410 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
9411 if the SSL_R_LENGTH_MISMATCH error is detected.
9412 [Geoff Thorpe, Bodo Moeller]
9413
9414 *) New 'shared_ldflag' column in Configure platform table.
9415 [Richard Levitte]
9416
9417 *) Fix EVP_CIPHER_mode macro.
9418 ["Dan S. Camper" <dan@bti.net>]
9419
9420 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
9421 type, we must throw them away by setting rr->length to 0.
9422 [D P Chang <dpc@qualys.com>]
9423
9424 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
9425
9426 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
9427 <Dominikus.Scherkl@biodata.com>. (The previous implementation
9428 worked incorrectly for those cases where range = 10..._2 and
9429 3*range is two bits longer than range.)
9430 [Bodo Moeller]
9431
9432 *) Only add signing time to PKCS7 structures if it is not already
9433 present.
9434 [Steve Henson]
9435
9436 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
9437 OBJ_ld_ce should be OBJ_id_ce.
9438 Also some ip-pda OIDs in crypto/objects/objects.txt were
9439 incorrect (cf. RFC 3039).
9440 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
9441
9442 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
9443 returns early because it has nothing to do.
9444 [Andy Schneider <andy.schneider@bjss.co.uk>]
9445
9446 *) [In 0.9.6c-engine release:]
9447 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
9448 [Andy Schneider <andy.schneider@bjss.co.uk>]
9449
9450 *) [In 0.9.6c-engine release:]
9451 Add support for Cryptographic Appliance's keyserver technology.
9452 (Use engine 'keyclient')
9453 [Cryptographic Appliances and Geoff Thorpe]
9454
9455 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
9456 is called via tools/c89.sh because arguments have to be
9457 rearranged (all '-L' options must appear before the first object
9458 modules).
9459 [Richard Shapiro <rshapiro@abinitio.com>]
9460
9461 *) [In 0.9.6c-engine release:]
9462 Add support for Broadcom crypto accelerator cards, backported
9463 from 0.9.7.
9464 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
9465
9466 *) [In 0.9.6c-engine release:]
9467 Add support for SureWare crypto accelerator cards from
9468 Baltimore Technologies. (Use engine 'sureware')
9469 [Baltimore Technologies and Mark Cox]
9470
9471 *) [In 0.9.6c-engine release:]
9472 Add support for crypto accelerator cards from Accelerated
9473 Encryption Processing, www.aep.ie. (Use engine 'aep')
9474 [AEP Inc. and Mark Cox]
9475
9476 *) Add a configuration entry for gcc on UnixWare.
9477 [Gary Benson <gbenson@redhat.com>]
9478
9479 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
9480 messages are stored in a single piece (fixed-length part and
9481 variable-length part combined) and fix various bugs found on the way.
9482 [Bodo Moeller]
9483
9484 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9485 instead. BIO_gethostbyname() does not know what timeouts are
9486 appropriate, so entries would stay in cache even when they have
9487 become invalid.
9488 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9489
9490 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9491 faced with a pathologically small ClientHello fragment that does
9492 not contain client_version: Instead of aborting with an error,
9493 simply choose the highest available protocol version (i.e.,
9494 TLS 1.0 unless it is disabled). In practice, ClientHello
9495 messages are never sent like this, but this change gives us
9496 strictly correct behaviour at least for TLS.
9497 [Bodo Moeller]
9498
9499 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9500 never resets s->method to s->ctx->method when called from within
9501 one of the SSL handshake functions.
9502 [Bodo Moeller; problem pointed out by Niko Baric]
9503
9504 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9505 (sent using the client's version number) if client_version is
9506 smaller than the protocol version in use. Also change
9507 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9508 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9509 the client will at least see that alert.
9510 [Bodo Moeller]
9511
9512 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9513 correctly.
9514 [Bodo Moeller]
9515
9516 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9517 client receives HelloRequest while in a handshake.
9518 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9519
9520 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9521 should end in 'break', not 'goto end' which circumvents various
9522 cleanups done in state SSL_ST_OK. But session related stuff
9523 must be disabled for SSL_ST_OK in the case that we just sent a
9524 HelloRequest.
9525
9526 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9527 before just sending a HelloRequest.
9528 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9529
9530 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9531 reveal whether illegal block cipher padding was found or a MAC
9532 verification error occurred. (Neither SSLerr() codes nor alerts
9533 are directly visible to potential attackers, but the information
9534 may leak via logfiles.)
9535
9536 Similar changes are not required for the SSL 2.0 implementation
9537 because the number of padding bytes is sent in clear for SSL 2.0,
9538 and the extra bytes are just ignored. However ssl/s2_pkt.c
9539 failed to verify that the purported number of padding bytes is in
9540 the legal range.
9541 [Bodo Moeller]
9542
9543 *) Add OpenUNIX-8 support including shared libraries
9544 (Boyd Lynn Gerber <gerberb@zenez.com>).
9545 [Lutz Jaenicke]
9546
9547 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9548 'wristwatch attack' using huge encoding parameters (cf.
9549 James H. Manger's CRYPTO 2001 paper). Note that the
9550 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9551 encoding parameters and hence was not vulnerable.
9552 [Bodo Moeller]
9553
9554 *) BN_sqr() bug fix.
9555 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9556
9557 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9558 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9559 followed by modular reduction.
9560 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9561
9562 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9563 equivalent based on BN_pseudo_rand() instead of BN_rand().
9564 [Bodo Moeller]
9565
9566 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9567 This function was broken, as the check for a new client hello message
9568 to handle SGC did not allow these large messages.
9569 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9570 [Lutz Jaenicke]
9571
9572 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9573 [Lutz Jaenicke]
9574
9575 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9576 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9577 [Lutz Jaenicke]
9578
9579 *) Rework the configuration and shared library support for Tru64 Unix.
9580 The configuration part makes use of modern compiler features and
9581 still retains old compiler behavior for those that run older versions
9582 of the OS. The shared library support part includes a variant that
9583 uses the RPATH feature, and is available through the special
9584 configuration target "alpha-cc-rpath", which will never be selected
9585 automatically.
9586 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9587
9588 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9589 with the same message size as in ssl3_get_certificate_request().
9590 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9591 messages might inadvertently be reject as too long.
9592 [Petr Lampa <lampa@fee.vutbr.cz>]
9593
9594 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9595 [Andy Polyakov]
9596
9597 *) Modified SSL library such that the verify_callback that has been set
9598 specificly for an SSL object with SSL_set_verify() is actually being
9599 used. Before the change, a verify_callback set with this function was
9600 ignored and the verify_callback() set in the SSL_CTX at the time of
9601 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9602 to allow the necessary settings.
9603 [Lutz Jaenicke]
9604
9605 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9606 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9607 done automatically (in contradiction to the requirements of the C
9608 standard). This made problems when used from OpenSSH.
9609 [Lutz Jaenicke]
9610
9611 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9612 dh->length and always used
9613
9614 BN_rand_range(priv_key, dh->p).
9615
9616 BN_rand_range() is not necessary for Diffie-Hellman, and this
9617 specific range makes Diffie-Hellman unnecessarily inefficient if
9618 dh->length (recommended exponent length) is much smaller than the
9619 length of dh->p. We could use BN_rand_range() if the order of
9620 the subgroup was stored in the DH structure, but we only have
9621 dh->length.
9622
9623 So switch back to
9624
9625 BN_rand(priv_key, l, ...)
9626
9627 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
9628 otherwise.
9629 [Bodo Moeller]
9630
9631 *) In
9632
9633 RSA_eay_public_encrypt
9634 RSA_eay_private_decrypt
9635 RSA_eay_private_encrypt (signing)
9636 RSA_eay_public_decrypt (signature verification)
9637
9638 (default implementations for RSA_public_encrypt,
9639 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
9640 always reject numbers >= n.
9641 [Bodo Moeller]
9642
9643 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
9644 to synchronize access to 'locking_thread'. This is necessary on
9645 systems where access to 'locking_thread' (an 'unsigned long'
9646 variable) is not atomic.
9647 [Bodo Moeller]
9648
9649 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
9650 *before* setting the 'crypto_lock_rand' flag. The previous code had
9651 a race condition if 0 is a valid thread ID.
9652 [Travis Vitek <vitek@roguewave.com>]
9653
9654 *) Add support for shared libraries under Irix.
9655 [Albert Chin-A-Young <china@thewrittenword.com>]
9656
9657 *) Add configuration option to build on Linux on both big-endian and
9658 little-endian MIPS.
9659 [Ralf Baechle <ralf@uni-koblenz.de>]
9660
9661 *) Add the possibility to create shared libraries on HP-UX.
9662 [Richard Levitte]
9663
9664 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
9665
9666 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
9667 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
9668 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
9669 PRNG state recovery was possible based on the output of
9670 one PRNG request appropriately sized to gain knowledge on
9671 'md' followed by enough consecutive 1-byte PRNG requests
9672 to traverse all of 'state'.
9673
9674 1. When updating 'md_local' (the current thread's copy of 'md')
9675 during PRNG output generation, hash all of the previous
9676 'md_local' value, not just the half used for PRNG output.
9677
9678 2. Make the number of bytes from 'state' included into the hash
9679 independent from the number of PRNG bytes requested.
9680
9681 The first measure alone would be sufficient to avoid
9682 Markku-Juhani's attack. (Actually it had never occurred
9683 to me that the half of 'md_local' used for chaining was the
9684 half from which PRNG output bytes were taken -- I had always
9685 assumed that the secret half would be used.) The second
9686 measure makes sure that additional data from 'state' is never
9687 mixed into 'md_local' in small portions; this heuristically
9688 further strengthens the PRNG.
9689 [Bodo Moeller]
9690
9691 *) Fix crypto/bn/asm/mips3.s.
9692 [Andy Polyakov]
9693
9694 *) When only the key is given to "enc", the IV is undefined. Print out
9695 an error message in this case.
9696 [Lutz Jaenicke]
9697
9698 *) Handle special case when X509_NAME is empty in X509 printing routines.
9699 [Steve Henson]
9700
9701 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
9702 positive and less than q.
9703 [Bodo Moeller]
9704
9705 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
9706 used: it isn't thread safe and the add_lock_callback should handle
9707 that itself.
9708 [Paul Rose <Paul.Rose@bridge.com>]
9709
9710 *) Verify that incoming data obeys the block size in
9711 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
9712 [Bodo Moeller]
9713
9714 *) Fix OAEP check.
9715 [Ulf Möller, Bodo Möller]
9716
9717 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
9718 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
9719 when fixing the server behaviour for backwards-compatible 'client
9720 hello' messages. (Note that the attack is impractical against
9721 SSL 3.0 and TLS 1.0 anyway because length and version checking
9722 means that the probability of guessing a valid ciphertext is
9723 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
9724 paper.)
9725
9726 Before 0.9.5, the countermeasure (hide the error by generating a
9727 random 'decryption result') did not work properly because
9728 ERR_clear_error() was missing, meaning that SSL_get_error() would
9729 detect the supposedly ignored error.
9730
9731 Both problems are now fixed.
9732 [Bodo Moeller]
9733
9734 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
9735 (previously it was 1024).
9736 [Bodo Moeller]
9737
9738 *) Fix for compatibility mode trust settings: ignore trust settings
9739 unless some valid trust or reject settings are present.
9740 [Steve Henson]
9741
9742 *) Fix for blowfish EVP: its a variable length cipher.
9743 [Steve Henson]
9744
9745 *) Fix various bugs related to DSA S/MIME verification. Handle missing
9746 parameters in DSA public key structures and return an error in the
9747 DSA routines if parameters are absent.
9748 [Steve Henson]
9749
9750 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
9751 in the current directory if neither $RANDFILE nor $HOME was set.
9752 RAND_file_name() in 0.9.6a returned NULL in this case. This has
9753 caused some confusion to Windows users who haven't defined $HOME.
9754 Thus RAND_file_name() is changed again: e_os.h can define a
9755 DEFAULT_HOME, which will be used if $HOME is not set.
9756 For Windows, we use "C:"; on other platforms, we still require
9757 environment variables.
9758
9759 *) Move 'if (!initialized) RAND_poll()' into regions protected by
9760 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
9761 having multiple threads call RAND_poll() concurrently.
9762 [Bodo Moeller]
9763
9764 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
9765 combination of a flag and a thread ID variable.
9766 Otherwise while one thread is in ssleay_rand_bytes (which sets the
9767 flag), *other* threads can enter ssleay_add_bytes without obeying
9768 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
9769 that they do not hold after the first thread unsets add_do_not_lock).
9770 [Bodo Moeller]
9771
9772 *) Change bctest again: '-x' expressions are not available in all
9773 versions of 'test'.
9774 [Bodo Moeller]
9775
9776 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
9777
9778 *) Fix a couple of memory leaks in PKCS7_dataDecode()
9779 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9780
9781 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9782 the default extension for executables, if any. Also, make the perl
9783 scripts that use symlink() to test if it really exists and use "cp"
9784 if it doesn't. All this made OpenSSL compilable and installable in
9785 CygWin.
9786 [Richard Levitte]
9787
9788 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9789 If SEQUENCE is length is indefinite just set c->slen to the total
9790 amount of data available.
9791 [Steve Henson, reported by shige@FreeBSD.org]
9792 [This change does not apply to 0.9.7.]
9793
9794 *) Change bctest to avoid here-documents inside command substitution
9795 (workaround for FreeBSD /bin/sh bug).
9796 For compatibility with Ultrix, avoid shell functions (introduced
9797 in the bctest version that searches along $PATH).
9798 [Bodo Moeller]
9799
9800 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9801 with des_encrypt() defined on some operating systems, like Solaris
9802 and UnixWare.
9803 [Richard Levitte]
9804
9805 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9806 On the Importance of Eliminating Errors in Cryptographic
9807 Computations, J. Cryptology 14 (2001) 2, 101-119,
9808 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9809 [Ulf Moeller]
9810
9811 *) MIPS assembler BIGNUM division bug fix.
9812 [Andy Polyakov]
9813
9814 *) Disabled incorrect Alpha assembler code.
9815 [Richard Levitte]
9816
9817 *) Fix PKCS#7 decode routines so they correctly update the length
9818 after reading an EOC for the EXPLICIT tag.
9819 [Steve Henson]
9820 [This change does not apply to 0.9.7.]
9821
9822 *) Fix bug in PKCS#12 key generation routines. This was triggered
9823 if a 3DES key was generated with a 0 initial byte. Include
9824 PKCS12_BROKEN_KEYGEN compilation option to retain the old
9825 (but broken) behaviour.
9826 [Steve Henson]
9827
9828 *) Enhance bctest to search for a working bc along $PATH and print
9829 it when found.
9830 [Tim Rice <tim@multitalents.net> via Richard Levitte]
9831
9832 *) Fix memory leaks in err.c: free err_data string if necessary;
9833 don't write to the wrong index in ERR_set_error_data.
9834 [Bodo Moeller]
9835
9836 *) Implement ssl23_peek (analogous to ssl23_read), which previously
9837 did not exist.
9838 [Bodo Moeller]
9839
9840 *) Replace rdtsc with _emit statements for VC++ version 5.
9841 [Jeremy Cooper <jeremy@baymoo.org>]
9842
9843 *) Make it possible to reuse SSLv2 sessions.
9844 [Richard Levitte]
9845
9846 *) In copy_email() check for >= 0 as a return value for
9847 X509_NAME_get_index_by_NID() since 0 is a valid index.
9848 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
9849
9850 *) Avoid coredump with unsupported or invalid public keys by checking if
9851 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
9852 PKCS7_verify() fails with non detached data.
9853 [Steve Henson]
9854
9855 *) Don't use getenv in library functions when run as setuid/setgid.
9856 New function OPENSSL_issetugid().
9857 [Ulf Moeller]
9858
9859 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
9860 due to incorrect handling of multi-threading:
9861
9862 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
9863
9864 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
9865
9866 3. Count how many times MemCheck_off() has been called so that
9867 nested use can be treated correctly. This also avoids
9868 inband-signalling in the previous code (which relied on the
9869 assumption that thread ID 0 is impossible).
9870 [Bodo Moeller]
9871
9872 *) Add "-rand" option also to s_client and s_server.
9873 [Lutz Jaenicke]
9874
9875 *) Fix CPU detection on Irix 6.x.
9876 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
9877 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9878
9879 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
9880 was empty.
9881 [Steve Henson]
9882 [This change does not apply to 0.9.7.]
9883
9884 *) Use the cached encoding of an X509_NAME structure rather than
9885 copying it. This is apparently the reason for the libsafe "errors"
9886 but the code is actually correct.
9887 [Steve Henson]
9888
9889 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
9890 Bleichenbacher's DSA attack.
9891 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
9892 to be set and top=0 forces the highest bit to be set; top=-1 is new
9893 and leaves the highest bit random.
9894 [Ulf Moeller, Bodo Moeller]
9895
9896 *) In the NCONF_...-based implementations for CONF_... queries
9897 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
9898 a temporary CONF structure with the data component set to NULL
9899 (which gives segmentation faults in lh_retrieve).
9900 Instead, use NULL for the CONF pointer in CONF_get_string and
9901 CONF_get_number (which may use environment variables) and directly
9902 return NULL from CONF_get_section.
9903 [Bodo Moeller]
9904
9905 *) Fix potential buffer overrun for EBCDIC.
9906 [Ulf Moeller]
9907
9908 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9909 keyUsage if basicConstraints absent for a CA.
9910 [Steve Henson]
9911
9912 *) Make SMIME_write_PKCS7() write mail header values with a format that
9913 is more generally accepted (no spaces before the semicolon), since
9914 some programs can't parse those values properly otherwise. Also make
9915 sure BIO's that break lines after each write do not create invalid
9916 headers.
9917 [Richard Levitte]
9918
9919 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9920 macros previously used would not encode an empty SEQUENCE OF
9921 and break the signature.
9922 [Steve Henson]
9923 [This change does not apply to 0.9.7.]
9924
9925 *) Zero the premaster secret after deriving the master secret in
9926 DH ciphersuites.
9927 [Steve Henson]
9928
9929 *) Add some EVP_add_digest_alias registrations (as found in
9930 OpenSSL_add_all_digests()) to SSL_library_init()
9931 aka OpenSSL_add_ssl_algorithms(). This provides improved
9932 compatibility with peers using X.509 certificates
9933 with unconventional AlgorithmIdentifier OIDs.
9934 [Bodo Moeller]
9935
9936 *) Fix for Irix with NO_ASM.
9937 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9938
9939 *) ./config script fixes.
9940 [Ulf Moeller, Richard Levitte]
9941
9942 *) Fix 'openssl passwd -1'.
9943 [Bodo Moeller]
9944
9945 *) Change PKCS12_key_gen_asc() so it can cope with non null
9946 terminated strings whose length is passed in the passlen
9947 parameter, for example from PEM callbacks. This was done
9948 by adding an extra length parameter to asc2uni().
9949 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9950
9951 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9952 call failed, free the DSA structure.
9953 [Bodo Moeller]
9954
9955 *) Fix to uni2asc() to cope with zero length Unicode strings.
9956 These are present in some PKCS#12 files.
9957 [Steve Henson]
9958
9959 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9960 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9961 when writing a 32767 byte record.
9962 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9963
9964 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9965 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9966
9967 (RSA objects have a reference count access to which is protected
9968 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9969 so they are meant to be shared between threads.)
9970 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9971 "Reddie, Steven" <Steven.Reddie@ca.com>]
9972
9973 *) Fix a deadlock in CRYPTO_mem_leaks().
9974 [Bodo Moeller]
9975
9976 *) Use better test patterns in bntest.
9977 [Ulf Möller]
9978
9979 *) rand_win.c fix for Borland C.
9980 [Ulf Möller]
9981
9982 *) BN_rshift bugfix for n == 0.
9983 [Bodo Moeller]
9984
9985 *) Add a 'bctest' script that checks for some known 'bc' bugs
9986 so that 'make test' does not abort just because 'bc' is broken.
9987 [Bodo Moeller]
9988
9989 *) Store verify_result within SSL_SESSION also for client side to
9990 avoid potential security hole. (Re-used sessions on the client side
9991 always resulted in verify_result==X509_V_OK, not using the original
9992 result of the server certificate verification.)
9993 [Lutz Jaenicke]
9994
9995 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9996 SSL3_RT_APPLICATION_DATA, return 0.
9997 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9998 [Bodo Moeller]
9999
10000 *) Fix SSL_peek:
10001 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
10002 releases, have been re-implemented by renaming the previous
10003 implementations of ssl2_read and ssl3_read to ssl2_read_internal
10004 and ssl3_read_internal, respectively, and adding 'peek' parameters
10005 to them. The new ssl[23]_{read,peek} functions are calls to
10006 ssl[23]_read_internal with the 'peek' flag set appropriately.
10007 A 'peek' parameter has also been added to ssl3_read_bytes, which
10008 does the actual work for ssl3_read_internal.
10009 [Bodo Moeller]
10010
10011 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
10012 the method-specific "init()" handler. Also clean up ex_data after
10013 calling the method-specific "finish()" handler. Previously, this was
10014 happening the other way round.
10015 [Geoff Thorpe]
10016
10017 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
10018 The previous value, 12, was not always sufficient for BN_mod_exp().
10019 [Bodo Moeller]
10020
10021 *) Make sure that shared libraries get the internal name engine with
10022 the full version number and not just 0. This should mark the
10023 shared libraries as not backward compatible. Of course, this should
10024 be changed again when we can guarantee backward binary compatibility.
10025 [Richard Levitte]
10026
10027 *) Fix typo in get_cert_by_subject() in by_dir.c
10028 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
10029
10030 *) Rework the system to generate shared libraries:
10031
10032 - Make note of the expected extension for the shared libraries and
10033 if there is a need for symbolic links from for example libcrypto.so.0
10034 to libcrypto.so.0.9.7. There is extended info in Configure for
10035 that.
10036
10037 - Make as few rebuilds of the shared libraries as possible.
10038
10039 - Still avoid linking the OpenSSL programs with the shared libraries.
10040
10041 - When installing, install the shared libraries separately from the
10042 static ones.
10043 [Richard Levitte]
10044
10045 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
10046
10047 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
10048 and not in SSL_clear because the latter is also used by the
10049 accept/connect functions; previously, the settings made by
10050 SSL_set_read_ahead would be lost during the handshake.
10051 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
10052
10053 *) Correct util/mkdef.pl to be selective about disabled algorithms.
10054 Previously, it would create entries for disabled algorithms no
10055 matter what.
10056 [Richard Levitte]
10057
10058 *) Added several new manual pages for SSL_* function.
10059 [Lutz Jaenicke]
10060
10061 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
10062
10063 *) In ssl23_get_client_hello, generate an error message when faced
10064 with an initial SSL 3.0/TLS record that is too small to contain the
10065 first two bytes of the ClientHello message, i.e. client_version.
10066 (Note that this is a pathologic case that probably has never happened
10067 in real life.) The previous approach was to use the version number
10068 from the record header as a substitute; but our protocol choice
10069 should not depend on that one because it is not authenticated
10070 by the Finished messages.
10071 [Bodo Moeller]
10072
10073 *) More robust randomness gathering functions for Windows.
10074 [Jeffrey Altman <jaltman@columbia.edu>]
10075
10076 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
10077 not set then we don't setup the error code for issuer check errors
10078 to avoid possibly overwriting other errors which the callback does
10079 handle. If an application does set the flag then we assume it knows
10080 what it is doing and can handle the new informational codes
10081 appropriately.
10082 [Steve Henson]
10083
10084 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
10085 a general "ANY" type, as such it should be able to decode anything
10086 including tagged types. However it didn't check the class so it would
10087 wrongly interpret tagged types in the same way as their universal
10088 counterpart and unknown types were just rejected. Changed so that the
10089 tagged and unknown types are handled in the same way as a SEQUENCE:
10090 that is the encoding is stored intact. There is also a new type
10091 "V_ASN1_OTHER" which is used when the class is not universal, in this
10092 case we have no idea what the actual type is so we just lump them all
10093 together.
10094 [Steve Henson]
10095
10096 *) On VMS, stdout may very well lead to a file that is written to
10097 in a record-oriented fashion. That means that every write() will
10098 write a separate record, which will be read separately by the
10099 programs trying to read from it. This can be very confusing.
10100
10101 The solution is to put a BIO filter in the way that will buffer
10102 text until a linefeed is reached, and then write everything a
10103 line at a time, so every record written will be an actual line,
10104 not chunks of lines and not (usually doesn't happen, but I've
10105 seen it once) several lines in one record. BIO_f_linebuffer() is
10106 the answer.
10107
10108 Currently, it's a VMS-only method, because that's where it has
10109 been tested well enough.
10110 [Richard Levitte]
10111
10112 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
10113 it can return incorrect results.
10114 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
10115 but it was in 0.9.6-beta[12].)
10116 [Bodo Moeller]
10117
10118 *) Disable the check for content being present when verifying detached
10119 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
10120 include zero length content when signing messages.
10121 [Steve Henson]
10122
10123 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
10124 BIO_ctrl (for BIO pairs).
10125 [Bodo Möller]
10126
10127 *) Add DSO method for VMS.
10128 [Richard Levitte]
10129
10130 *) Bug fix: Montgomery multiplication could produce results with the
10131 wrong sign.
10132 [Ulf Möller]
10133
10134 *) Add RPM specification openssl.spec and modify it to build three
10135 packages. The default package contains applications, application
10136 documentation and run-time libraries. The devel package contains
10137 include files, static libraries and function documentation. The
10138 doc package contains the contents of the doc directory. The original
10139 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
10140 [Richard Levitte]
10141
10142 *) Add a large number of documentation files for many SSL routines.
10143 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10144
10145 *) Add a configuration entry for Sony News 4.
10146 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
10147
10148 *) Don't set the two most significant bits to one when generating a
10149 random number < q in the DSA library.
10150 [Ulf Möller]
10151
10152 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
10153 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
10154 the underlying transport is blocking) if a handshake took place.
10155 (The default behaviour is needed by applications such as s_client
10156 and s_server that use select() to determine when to use SSL_read;
10157 but for applications that know in advance when to expect data, it
10158 just makes things more complicated.)
10159 [Bodo Moeller]
10160
10161 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
10162 from EGD.
10163 [Ben Laurie]
10164
10165 *) Add a few more EBCDIC conditionals that make `req' and `x509'
10166 work better on such systems.
10167 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
10168
10169 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
10170 Update PKCS12_parse() so it copies the friendlyName and the
10171 keyid to the certificates aux info.
10172 [Steve Henson]
10173
10174 *) Fix bug in PKCS7_verify() which caused an infinite loop
10175 if there was more than one signature.
10176 [Sven Uszpelkat <su@celocom.de>]
10177
10178 *) Major change in util/mkdef.pl to include extra information
10179 about each symbol, as well as presenting variables as well
10180 as functions. This change means that there's n more need
10181 to rebuild the .num files when some algorithms are excluded.
10182 [Richard Levitte]
10183
10184 *) Allow the verify time to be set by an application,
10185 rather than always using the current time.
10186 [Steve Henson]
10187
10188 *) Phase 2 verify code reorganisation. The certificate
10189 verify code now looks up an issuer certificate by a
10190 number of criteria: subject name, authority key id
10191 and key usage. It also verifies self signed certificates
10192 by the same criteria. The main comparison function is
10193 X509_check_issued() which performs these checks.
10194
10195 Lot of changes were necessary in order to support this
10196 without completely rewriting the lookup code.
10197
10198 Authority and subject key identifier are now cached.
10199
10200 The LHASH 'certs' is X509_STORE has now been replaced
10201 by a STACK_OF(X509_OBJECT). This is mainly because an
10202 LHASH can't store or retrieve multiple objects with
10203 the same hash value.
10204
10205 As a result various functions (which were all internal
10206 use only) have changed to handle the new X509_STORE
10207 structure. This will break anything that messed round
10208 with X509_STORE internally.
10209
10210 The functions X509_STORE_add_cert() now checks for an
10211 exact match, rather than just subject name.
10212
10213 The X509_STORE API doesn't directly support the retrieval
10214 of multiple certificates matching a given criteria, however
10215 this can be worked round by performing a lookup first
10216 (which will fill the cache with candidate certificates)
10217 and then examining the cache for matches. This is probably
10218 the best we can do without throwing out X509_LOOKUP
10219 entirely (maybe later...).
10220
10221 The X509_VERIFY_CTX structure has been enhanced considerably.
10222
10223 All certificate lookup operations now go via a get_issuer()
10224 callback. Although this currently uses an X509_STORE it
10225 can be replaced by custom lookups. This is a simple way
10226 to bypass the X509_STORE hackery necessary to make this
10227 work and makes it possible to use more efficient techniques
10228 in future. A very simple version which uses a simple
10229 STACK for its trusted certificate store is also provided
10230 using X509_STORE_CTX_trusted_stack().
10231
10232 The verify_cb() and verify() callbacks now have equivalents
10233 in the X509_STORE_CTX structure.
10234
10235 X509_STORE_CTX also has a 'flags' field which can be used
10236 to customise the verify behaviour.
10237 [Steve Henson]
10238
10239 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
10240 excludes S/MIME capabilities.
10241 [Steve Henson]
10242
10243 *) When a certificate request is read in keep a copy of the
10244 original encoding of the signed data and use it when outputting
10245 again. Signatures then use the original encoding rather than
10246 a decoded, encoded version which may cause problems if the
10247 request is improperly encoded.
10248 [Steve Henson]
10249
10250 *) For consistency with other BIO_puts implementations, call
10251 buffer_write(b, ...) directly in buffer_puts instead of calling
10252 BIO_write(b, ...).
10253
10254 In BIO_puts, increment b->num_write as in BIO_write.
10255 [Peter.Sylvester@EdelWeb.fr]
10256
10257 *) Fix BN_mul_word for the case where the word is 0. (We have to use
10258 BN_zero, we may not return a BIGNUM with an array consisting of
10259 words set to zero.)
10260 [Bodo Moeller]
10261
10262 *) Avoid calling abort() from within the library when problems are
10263 detected, except if preprocessor symbols have been defined
10264 (such as REF_CHECK, BN_DEBUG etc.).
10265 [Bodo Moeller]
10266
10267 *) New openssl application 'rsautl'. This utility can be
10268 used for low level RSA operations. DER public key
10269 BIO/fp routines also added.
10270 [Steve Henson]
10271
10272 *) New Configure entry and patches for compiling on QNX 4.
10273 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
10274
10275 *) A demo state-machine implementation was sponsored by
10276 Nuron (http://www.nuron.com/) and is now available in
10277 demos/state_machine.
10278 [Ben Laurie]
10279
10280 *) New options added to the 'dgst' utility for signature
10281 generation and verification.
10282 [Steve Henson]
10283
10284 *) Unrecognized PKCS#7 content types are now handled via a
10285 catch all ASN1_TYPE structure. This allows unsupported
10286 types to be stored as a "blob" and an application can
10287 encode and decode it manually.
10288 [Steve Henson]
10289
10290 *) Fix various signed/unsigned issues to make a_strex.c
10291 compile under VC++.
10292 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
10293
10294 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
10295 length if passed a buffer. ASN1_INTEGER_to_BN failed
10296 if passed a NULL BN and its argument was negative.
10297 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
10298
10299 *) Modification to PKCS#7 encoding routines to output definite
10300 length encoding. Since currently the whole structures are in
10301 memory there's not real point in using indefinite length
10302 constructed encoding. However if OpenSSL is compiled with
10303 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
10304 [Steve Henson]
10305
10306 *) Added BIO_vprintf() and BIO_vsnprintf().
10307 [Richard Levitte]
10308
10309 *) Added more prefixes to parse for in the strings written
10310 through a logging bio, to cover all the levels that are available
10311 through syslog. The prefixes are now:
10312
10313 PANIC, EMERG, EMR => LOG_EMERG
10314 ALERT, ALR => LOG_ALERT
10315 CRIT, CRI => LOG_CRIT
10316 ERROR, ERR => LOG_ERR
10317 WARNING, WARN, WAR => LOG_WARNING
10318 NOTICE, NOTE, NOT => LOG_NOTICE
10319 INFO, INF => LOG_INFO
10320 DEBUG, DBG => LOG_DEBUG
10321
10322 and as before, if none of those prefixes are present at the
10323 beginning of the string, LOG_ERR is chosen.
10324
10325 On Win32, the LOG_* levels are mapped according to this:
10326
10327 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
10328 LOG_WARNING => EVENTLOG_WARNING_TYPE
10329 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
10330
10331 [Richard Levitte]
10332
10333 *) Made it possible to reconfigure with just the configuration
10334 argument "reconf" or "reconfigure". The command line arguments
10335 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
10336 and are retrieved from there when reconfiguring.
10337 [Richard Levitte]
10338
10339 *) MD4 implemented.
10340 [Assar Westerlund <assar@sics.se>, Richard Levitte]
10341
10342 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
10343 [Richard Levitte]
10344
10345 *) The obj_dat.pl script was messing up the sorting of object
10346 names. The reason was that it compared the quoted version
10347 of strings as a result "OCSP" > "OCSP Signing" because
10348 " > SPACE. Changed script to store unquoted versions of
10349 names and add quotes on output. It was also omitting some
10350 names from the lookup table if they were given a default
10351 value (that is if SN is missing it is given the same
10352 value as LN and vice versa), these are now added on the
10353 grounds that if an object has a name we should be able to
10354 look it up. Finally added warning output when duplicate
10355 short or long names are found.
10356 [Steve Henson]
10357
10358 *) Changes needed for Tandem NSK.
10359 [Scott Uroff <scott@xypro.com>]
10360
10361 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
10362 RSA_padding_check_SSLv23(), special padding was never detected
10363 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
10364 version rollback attacks was not effective.
10365
10366 In s23_clnt.c, don't use special rollback-attack detection padding
10367 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
10368 client; similarly, in s23_srvr.c, don't do the rollback check if
10369 SSL 2.0 is the only protocol enabled in the server.
10370 [Bodo Moeller]
10371
10372 *) Make it possible to get hexdumps of unprintable data with 'openssl
10373 asn1parse'. By implication, the functions ASN1_parse_dump() and
10374 BIO_dump_indent() are added.
10375 [Richard Levitte]
10376
10377 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
10378 these print out strings and name structures based on various
10379 flags including RFC2253 support and proper handling of
10380 multibyte characters. Added options to the 'x509' utility
10381 to allow the various flags to be set.
10382 [Steve Henson]
10383
10384 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
10385 Also change the functions X509_cmp_current_time() and
10386 X509_gmtime_adj() work with an ASN1_TIME structure,
10387 this will enable certificates using GeneralizedTime in validity
10388 dates to be checked.
10389 [Steve Henson]
10390
10391 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
10392 negative public key encodings) on by default,
10393 NO_NEG_PUBKEY_BUG can be set to disable it.
10394 [Steve Henson]
10395
10396 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
10397 content octets. An i2c_ASN1_OBJECT is unnecessary because
10398 the encoding can be trivially obtained from the structure.
10399 [Steve Henson]
10400
10401 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
10402 not read locks (CRYPTO_r_[un]lock).
10403 [Bodo Moeller]
10404
10405 *) A first attempt at creating official support for shared
10406 libraries through configuration. I've kept it so the
10407 default is static libraries only, and the OpenSSL programs
10408 are always statically linked for now, but there are
10409 preparations for dynamic linking in place.
10410 This has been tested on Linux and Tru64.
10411 [Richard Levitte]
10412
10413 *) Randomness polling function for Win9x, as described in:
10414 Peter Gutmann, Software Generation of Practically Strong
10415 Random Numbers.
10416 [Ulf Möller]
10417
10418 *) Fix so PRNG is seeded in req if using an already existing
10419 DSA key.
10420 [Steve Henson]
10421
10422 *) New options to smime application. -inform and -outform
10423 allow alternative formats for the S/MIME message including
10424 PEM and DER. The -content option allows the content to be
10425 specified separately. This should allow things like Netscape
10426 form signing output easier to verify.
10427 [Steve Henson]
10428
10429 *) Fix the ASN1 encoding of tags using the 'long form'.
10430 [Steve Henson]
10431
10432 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
10433 STRING types. These convert content octets to and from the
10434 underlying type. The actual tag and length octets are
10435 already assumed to have been read in and checked. These
10436 are needed because all other string types have virtually
10437 identical handling apart from the tag. By having versions
10438 of the ASN1 functions that just operate on content octets
10439 IMPLICIT tagging can be handled properly. It also allows
10440 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
10441 and ASN1_INTEGER are identical apart from the tag.
10442 [Steve Henson]
10443
10444 *) Change the handling of OID objects as follows:
10445
10446 - New object identifiers are inserted in objects.txt, following
10447 the syntax given in objects.README.
10448 - objects.pl is used to process obj_mac.num and create a new
10449 obj_mac.h.
10450 - obj_dat.pl is used to create a new obj_dat.h, using the data in
10451 obj_mac.h.
10452
10453 This is currently kind of a hack, and the perl code in objects.pl
10454 isn't very elegant, but it works as I intended. The simplest way
10455 to check that it worked correctly is to look in obj_dat.h and
10456 check the array nid_objs and make sure the objects haven't moved
10457 around (this is important!). Additions are OK, as well as
10458 consistent name changes.
10459 [Richard Levitte]
10460
10461 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
10462 [Bodo Moeller]
10463
10464 *) Addition of the command line parameter '-rand file' to 'openssl req'.
10465 The given file adds to whatever has already been seeded into the
10466 random pool through the RANDFILE configuration file option or
10467 environment variable, or the default random state file.
10468 [Richard Levitte]
10469
10470 *) mkstack.pl now sorts each macro group into lexical order.
10471 Previously the output order depended on the order the files
10472 appeared in the directory, resulting in needless rewriting
10473 of safestack.h .
10474 [Steve Henson]
10475
10476 *) Patches to make OpenSSL compile under Win32 again. Mostly
10477 work arounds for the VC++ problem that it treats func() as
10478 func(void). Also stripped out the parts of mkdef.pl that
10479 added extra typesafe functions: these no longer exist.
10480 [Steve Henson]
10481
10482 *) Reorganisation of the stack code. The macros are now all
10483 collected in safestack.h . Each macro is defined in terms of
10484 a "stack macro" of the form SKM_<name>(type, a, b). The
10485 DEBUG_SAFESTACK is now handled in terms of function casts,
10486 this has the advantage of retaining type safety without the
10487 use of additional functions. If DEBUG_SAFESTACK is not defined
10488 then the non typesafe macros are used instead. Also modified the
10489 mkstack.pl script to handle the new form. Needs testing to see
10490 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10491 the default if no major problems. Similar behaviour for ASN1_SET_OF
10492 and PKCS12_STACK_OF.
10493 [Steve Henson]
10494
10495 *) When some versions of IIS use the 'NET' form of private key the
10496 key derivation algorithm is different. Normally MD5(password) is
10497 used as a 128 bit RC4 key. In the modified case
10498 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10499 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10500 as the old Netscape_RSA functions except they have an additional
10501 'sgckey' parameter which uses the modified algorithm. Also added
10502 an -sgckey command line option to the rsa utility. Thanks to
10503 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10504 algorithm to openssl-dev.
10505 [Steve Henson]
10506
10507 *) The evp_local.h macros were using 'c.##kname' which resulted in
10508 invalid expansion on some systems (SCO 5.0.5 for example).
10509 Corrected to 'c.kname'.
10510 [Phillip Porch <root@theporch.com>]
10511
10512 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10513 a STACK of email addresses from a certificate or request, these look
10514 in the subject name and the subject alternative name extensions and
10515 omit any duplicate addresses.
10516 [Steve Henson]
10517
10518 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10519 This makes DSA verification about 2 % faster.
10520 [Bodo Moeller]
10521
10522 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10523 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10524 plus overhead for 1024 bit moduli).
10525 This makes exponentiations about 0.5 % faster for 1024 bit
10526 exponents (as measured by "openssl speed rsa2048").
10527 [Bodo Moeller]
10528
10529 *) Rename memory handling macros to avoid conflicts with other
10530 software:
10531 Malloc => OPENSSL_malloc
10532 Malloc_locked => OPENSSL_malloc_locked
10533 Realloc => OPENSSL_realloc
10534 Free => OPENSSL_free
10535 [Richard Levitte]
10536
10537 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10538 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10539 [Bodo Moeller]
10540
10541 *) CygWin32 support.
10542 [John Jarvie <jjarvie@newsguy.com>]
10543
10544 *) The type-safe stack code has been rejigged. It is now only compiled
10545 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10546 by default all type-specific stack functions are "#define"d back to
10547 standard stack functions. This results in more streamlined output
10548 but retains the type-safety checking possibilities of the original
10549 approach.
10550 [Geoff Thorpe]
10551
10552 *) The STACK code has been cleaned up, and certain type declarations
10553 that didn't make a lot of sense have been brought in line. This has
10554 also involved a cleanup of sorts in safestack.h to more correctly
10555 map type-safe stack functions onto their plain stack counterparts.
10556 This work has also resulted in a variety of "const"ifications of
10557 lots of the code, especially "_cmp" operations which should normally
10558 be prototyped with "const" parameters anyway.
10559 [Geoff Thorpe]
10560
10561 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10562 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10563 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10564 where all of 'md' is used each time the PRNG is used, but 'state'
10565 is used only indexed by a cyclic counter. As entropy may not be
10566 well distributed from the beginning, 'md' is important as a
10567 chaining variable. However, the output function chains only half
10568 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10569 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10570 in all of 'state' being rewritten, with the new values depending
10571 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10572 [Bodo Moeller]
10573
10574 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10575 the handshake is continued after ssl_verify_cert_chain();
10576 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10577 can lead to 'unexplainable' connection aborts later.
10578 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10579
10580 *) Major EVP API cipher revision.
10581 Add hooks for extra EVP features. This allows various cipher
10582 parameters to be set in the EVP interface. Support added for variable
10583 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10584 setting of RC2 and RC5 parameters.
10585
10586 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10587 ciphers.
10588
10589 Remove lots of duplicated code from the EVP library. For example *every*
10590 cipher init() function handles the 'iv' in the same way according to the
10591 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10592 for CFB and OFB modes they zero ctx->num.
10593
10594 New functionality allows removal of S/MIME code RC2 hack.
10595
10596 Most of the routines have the same form and so can be declared in terms
10597 of macros.
10598
10599 By shifting this to the top level EVP_CipherInit() it can be removed from
10600 all individual ciphers. If the cipher wants to handle IVs or keys
10601 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10602 flags.
10603
10604 Change lots of functions like EVP_EncryptUpdate() to now return a
10605 value: although software versions of the algorithms cannot fail
10606 any installed hardware versions can.
10607 [Steve Henson]
10608
10609 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10610 this option is set, tolerate broken clients that send the negotiated
10611 protocol version number instead of the requested protocol version
10612 number.
10613 [Bodo Moeller]
10614
10615 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10616 i.e. non-zero for export ciphersuites, zero otherwise.
10617 Previous versions had this flag inverted, inconsistent with
10618 rsa_tmp_cb (..._TMP_RSA_CB).
10619 [Bodo Moeller; problem reported by Amit Chopra]
10620
10621 *) Add missing DSA library text string. Work around for some IIS
10622 key files with invalid SEQUENCE encoding.
10623 [Steve Henson]
10624
10625 *) Add a document (doc/standards.txt) that list all kinds of standards
10626 and so on that are implemented in OpenSSL.
10627 [Richard Levitte]
10628
10629 *) Enhance c_rehash script. Old version would mishandle certificates
10630 with the same subject name hash and wouldn't handle CRLs at all.
10631 Added -fingerprint option to crl utility, to support new c_rehash
10632 features.
10633 [Steve Henson]
10634
10635 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
10636 [Ulf Möller]
10637
10638 *) Fix for SSL server purpose checking. Server checking was
10639 rejecting certificates which had extended key usage present
10640 but no ssl client purpose.
10641 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
10642
10643 *) Make PKCS#12 code work with no password. The PKCS#12 spec
10644 is a little unclear about how a blank password is handled.
10645 Since the password in encoded as a BMPString with terminating
10646 double NULL a zero length password would end up as just the
10647 double NULL. However no password at all is different and is
10648 handled differently in the PKCS#12 key generation code. NS
10649 treats a blank password as zero length. MSIE treats it as no
10650 password on export: but it will try both on import. We now do
10651 the same: PKCS12_parse() tries zero length and no password if
10652 the password is set to "" or NULL (NULL is now a valid password:
10653 it wasn't before) as does the pkcs12 application.
10654 [Steve Henson]
10655
10656 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
10657 perror when PEM_read_bio_X509_REQ fails, the error message must
10658 be obtained from the error queue.
10659 [Bodo Moeller]
10660
10661 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
10662 it in ERR_remove_state if appropriate, and change ERR_get_state
10663 accordingly to avoid race conditions (this is necessary because
10664 thread_hash is no longer constant once set).
10665 [Bodo Moeller]
10666
10667 *) Bugfix for linux-elf makefile.one.
10668 [Ulf Möller]
10669
10670 *) RSA_get_default_method() will now cause a default
10671 RSA_METHOD to be chosen if one doesn't exist already.
10672 Previously this was only set during a call to RSA_new()
10673 or RSA_new_method(NULL) meaning it was possible for
10674 RSA_get_default_method() to return NULL.
10675 [Geoff Thorpe]
10676
10677 *) Added native name translation to the existing DSO code
10678 that will convert (if the flag to do so is set) filenames
10679 that are sufficiently small and have no path information
10680 into a canonical native form. Eg. "blah" converted to
10681 "libblah.so" or "blah.dll" etc.
10682 [Geoff Thorpe]
10683
10684 *) New function ERR_error_string_n(e, buf, len) which is like
10685 ERR_error_string(e, buf), but writes at most 'len' bytes
10686 including the 0 terminator. For ERR_error_string_n, 'buf'
10687 may not be NULL.
10688 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
10689
10690 *) CONF library reworked to become more general. A new CONF
10691 configuration file reader "class" is implemented as well as a
10692 new functions (NCONF_*, for "New CONF") to handle it. The now
10693 old CONF_* functions are still there, but are reimplemented to
10694 work in terms of the new functions. Also, a set of functions
10695 to handle the internal storage of the configuration data is
10696 provided to make it easier to write new configuration file
10697 reader "classes" (I can definitely see something reading a
10698 configuration file in XML format, for example), called _CONF_*,
10699 or "the configuration storage API"...
10700
10701 The new configuration file reading functions are:
10702
10703 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
10704 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
10705
10706 NCONF_default, NCONF_WIN32
10707
10708 NCONF_dump_fp, NCONF_dump_bio
10709
10710 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
10711 NCONF_new creates a new CONF object. This works in the same way
10712 as other interfaces in OpenSSL, like the BIO interface.
10713 NCONF_dump_* dump the internal storage of the configuration file,
10714 which is useful for debugging. All other functions take the same
10715 arguments as the old CONF_* functions wth the exception of the
10716 first that must be a `CONF *' instead of a `LHASH *'.
10717
10718 To make it easer to use the new classes with the old CONF_* functions,
10719 the function CONF_set_default_method is provided.
10720 [Richard Levitte]
10721
10722 *) Add '-tls1' option to 'openssl ciphers', which was already
10723 mentioned in the documentation but had not been implemented.
10724 (This option is not yet really useful because even the additional
10725 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
10726 [Bodo Moeller]
10727
10728 *) Initial DSO code added into libcrypto for letting OpenSSL (and
10729 OpenSSL-based applications) load shared libraries and bind to
10730 them in a portable way.
10731 [Geoff Thorpe, with contributions from Richard Levitte]
10732
10733 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
10734
10735 *) Make sure _lrotl and _lrotr are only used with MSVC.
10736
10737 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
10738 (the default implementation of RAND_status).
10739
10740 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
10741 to '-clrext' (= clear extensions), as intended and documented.
10742 [Bodo Moeller; inconsistency pointed out by Michael Attili
10743 <attili@amaxo.com>]
10744
10745 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
10746 was larger than the MD block size.
10747 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
10748
10749 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
10750 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
10751 using the passed key: if the passed key was a private key the result
10752 of X509_print(), for example, would be to print out all the private key
10753 components.
10754 [Steve Henson]
10755
10756 *) des_quad_cksum() byte order bug fix.
10757 [Ulf Möller, using the problem description in krb4-0.9.7, where
10758 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
10759
10760 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
10761 discouraged.
10762 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
10763
10764 *) For easily testing in shell scripts whether some command
10765 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
10766 returns with exit code 0 iff no command of the given name is available.
10767 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
10768 the output goes to stdout and nothing is printed to stderr.
10769 Additional arguments are always ignored.
10770
10771 Since for each cipher there is a command of the same name,
10772 the 'no-cipher' compilation switches can be tested this way.
10773
10774 ('openssl no-XXX' is not able to detect pseudo-commands such
10775 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
10776 [Bodo Moeller]
10777
10778 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10779 [Bodo Moeller]
10780
10781 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10782 is set; it will be thrown away anyway because each handshake creates
10783 its own key.
10784 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10785 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10786 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10787 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10788 [Bodo Moeller]
10789
10790 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10791 'Q' and 'R' lose their special meanings (quit/renegotiate).
10792 This is part of what -quiet does; unlike -quiet, -ign_eof
10793 does not suppress any output.
10794 [Richard Levitte]
10795
10796 *) Add compatibility options to the purpose and trust code. The
10797 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10798 accepts a certificate or CA, this was the previous behaviour,
10799 with all the associated security issues.
10800
10801 X509_TRUST_COMPAT is the old trust behaviour: only and
10802 automatically trust self signed roots in certificate store. A
10803 new trust setting X509_TRUST_DEFAULT is used to specify that
10804 a purpose has no associated trust setting and it should instead
10805 use the value in the default purpose.
10806 [Steve Henson]
10807
10808 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10809 and fix a memory leak.
10810 [Steve Henson]
10811
10812 *) In util/mkerr.pl (which implements 'make errors'), preserve
10813 reason strings from the previous version of the .c file, as
10814 the default to have only downcase letters (and digits) in
10815 automatically generated reasons codes is not always appropriate.
10816 [Bodo Moeller]
10817
10818 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
10819 using strerror. Previously, ERR_reason_error_string() returned
10820 library names as reason strings for SYSerr; but SYSerr is a special
10821 case where small numbers are errno values, not library numbers.
10822 [Bodo Moeller]
10823
10824 *) Add '-dsaparam' option to 'openssl dhparam' application. This
10825 converts DSA parameters into DH parameters. (When creating parameters,
10826 DSA_generate_parameters is used.)
10827 [Bodo Moeller]
10828
10829 *) Include 'length' (recommended exponent length) in C code generated
10830 by 'openssl dhparam -C'.
10831 [Bodo Moeller]
10832
10833 *) The second argument to set_label in perlasm was already being used
10834 so couldn't be used as a "file scope" flag. Moved to third argument
10835 which was free.
10836 [Steve Henson]
10837
10838 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
10839 instead of RAND_bytes for encryption IVs and salts.
10840 [Bodo Moeller]
10841
10842 *) Include RAND_status() into RAND_METHOD instead of implementing
10843 it only for md_rand.c Otherwise replacing the PRNG by calling
10844 RAND_set_rand_method would be impossible.
10845 [Bodo Moeller]
10846
10847 *) Don't let DSA_generate_key() enter an infinite loop if the random
10848 number generation fails.
10849 [Bodo Moeller]
10850
10851 *) New 'rand' application for creating pseudo-random output.
10852 [Bodo Moeller]
10853
10854 *) Added configuration support for Linux/IA64
10855 [Rolf Haberrecker <rolf@suse.de>]
10856
10857 *) Assembler module support for Mingw32.
10858 [Ulf Möller]
10859
10860 *) Shared library support for HPUX (in shlib/).
10861 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
10862
10863 *) Shared library support for Solaris gcc.
10864 [Lutz Behnke <behnke@trustcenter.de>]
10865
10866 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
10867
10868 *) PKCS7_encrypt() was adding text MIME headers twice because they
10869 were added manually and by SMIME_crlf_copy().
10870 [Steve Henson]
10871
10872 *) In bntest.c don't call BN_rand with zero bits argument.
10873 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
10874
10875 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
10876 case was implemented. This caused BN_div_recp() to fail occasionally.
10877 [Ulf Möller]
10878
10879 *) Add an optional second argument to the set_label() in the perl
10880 assembly language builder. If this argument exists and is set
10881 to 1 it signals that the assembler should use a symbol whose
10882 scope is the entire file, not just the current function. This
10883 is needed with MASM which uses the format label:: for this scope.
10884 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
10885
10886 *) Change the ASN1 types so they are typedefs by default. Before
10887 almost all types were #define'd to ASN1_STRING which was causing
10888 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
10889 for example.
10890 [Steve Henson]
10891
10892 *) Change names of new functions to the new get1/get0 naming
10893 convention: After 'get1', the caller owns a reference count
10894 and has to call ..._free; 'get0' returns a pointer to some
10895 data structure without incrementing reference counters.
10896 (Some of the existing 'get' functions increment a reference
10897 counter, some don't.)
10898 Similarly, 'set1' and 'add1' functions increase reference
10899 counters or duplicate objects.
10900 [Steve Henson]
10901
10902 *) Allow for the possibility of temp RSA key generation failure:
10903 the code used to assume it always worked and crashed on failure.
10904 [Steve Henson]
10905
10906 *) Fix potential buffer overrun problem in BIO_printf().
10907 [Ulf Möller, using public domain code by Patrick Powell; problem
10908 pointed out by David Sacerdote <das33@cornell.edu>]
10909
10910 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10911 RAND_egd() and RAND_status(). In the command line application,
10912 the EGD socket can be specified like a seed file using RANDFILE
10913 or -rand.
10914 [Ulf Möller]
10915
10916 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10917 Some CAs (e.g. Verisign) distribute certificates in this form.
10918 [Steve Henson]
10919
10920 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10921 list to exclude them. This means that no special compilation option
10922 is needed to use anonymous DH: it just needs to be included in the
10923 cipher list.
10924 [Steve Henson]
10925
10926 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10927 EVP_MD_type. The old functionality is available in a new macro called
10928 EVP_MD_md(). Change code that uses it and update docs.
10929 [Steve Henson]
10930
10931 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10932 where the 'void *' argument is replaced by a function pointer argument.
10933 Previously 'void *' was abused to point to functions, which works on
10934 many platforms, but is not correct. As these functions are usually
10935 called by macros defined in OpenSSL header files, most source code
10936 should work without changes.
10937 [Richard Levitte]
10938
10939 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10940 sections with information on -D... compiler switches used for
10941 compiling the library so that applications can see them. To enable
10942 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10943 must be defined. E.g.,
10944 #define OPENSSL_ALGORITHM_DEFINES
10945 #include <openssl/opensslconf.h>
10946 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10947 [Richard Levitte, Ulf and Bodo Möller]
10948
10949 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10950 record layer.
10951 [Bodo Moeller]
10952
10953 *) Change the 'other' type in certificate aux info to a STACK_OF
10954 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10955 the required ASN1 format: arbitrary types determined by an OID.
10956 [Steve Henson]
10957
10958 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10959 argument to 'req'. This is not because the function is newer or
10960 better than others it just uses the work 'NEW' in the certificate
10961 request header lines. Some software needs this.
10962 [Steve Henson]
10963
10964 *) Reorganise password command line arguments: now passwords can be
10965 obtained from various sources. Delete the PEM_cb function and make
10966 it the default behaviour: i.e. if the callback is NULL and the
10967 usrdata argument is not NULL interpret it as a null terminated pass
10968 phrase. If usrdata and the callback are NULL then the pass phrase
10969 is prompted for as usual.
10970 [Steve Henson]
10971
10972 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10973 the support is automatically enabled. The resulting binaries will
10974 autodetect the card and use it if present.
10975 [Ben Laurie and Compaq Inc.]
10976
10977 *) Work around for Netscape hang bug. This sends certificate request
10978 and server done in one record. Since this is perfectly legal in the
10979 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10980 the bugs/SSLv3 entry for more info.
10981 [Steve Henson]
10982
10983 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10984 [Andy Polyakov]
10985
10986 *) Add -rand argument to smime and pkcs12 applications and read/write
10987 of seed file.
10988 [Steve Henson]
10989
10990 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10991 [Bodo Moeller]
10992
10993 *) Add command line password options to the remaining applications.
10994 [Steve Henson]
10995
10996 *) Bug fix for BN_div_recp() for numerators with an even number of
10997 bits.
10998 [Ulf Möller]
10999
11000 *) More tests in bntest.c, and changed test_bn output.
11001 [Ulf Möller]
11002
11003 *) ./config recognizes MacOS X now.
11004 [Andy Polyakov]
11005
11006 *) Bug fix for BN_div() when the first words of num and divisor are
11007 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
11008 [Ulf Möller]
11009
11010 *) Add support for various broken PKCS#8 formats, and command line
11011 options to produce them.
11012 [Steve Henson]
11013
11014 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
11015 get temporary BIGNUMs from a BN_CTX.
11016 [Ulf Möller]
11017
11018 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
11019 for p == 0.
11020 [Ulf Möller]
11021
11022 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
11023 include a #define from the old name to the new. The original intent
11024 was that statically linked binaries could for example just call
11025 SSLeay_add_all_ciphers() to just add ciphers to the table and not
11026 link with digests. This never worked because SSLeay_add_all_digests()
11027 and SSLeay_add_all_ciphers() were in the same source file so calling
11028 one would link with the other. They are now in separate source files.
11029 [Steve Henson]
11030
11031 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
11032 [Steve Henson]
11033
11034 *) Use a less unusual form of the Miller-Rabin primality test (it used
11035 a binary algorithm for exponentiation integrated into the Miller-Rabin
11036 loop, our standard modexp algorithms are faster).
11037 [Bodo Moeller]
11038
11039 *) Support for the EBCDIC character set completed.
11040 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
11041
11042 *) Source code cleanups: use const where appropriate, eliminate casts,
11043 use void * instead of char * in lhash.
11044 [Ulf Möller]
11045
11046 *) Bugfix: ssl3_send_server_key_exchange was not restartable
11047 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
11048 this the server could overwrite ephemeral keys that the client
11049 has already seen).
11050 [Bodo Moeller]
11051
11052 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
11053 using 50 iterations of the Rabin-Miller test.
11054
11055 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
11056 iterations of the Rabin-Miller test as required by the appendix
11057 to FIPS PUB 186[-1]) instead of DSA_is_prime.
11058 As BN_is_prime_fasttest includes trial division, DSA parameter
11059 generation becomes much faster.
11060
11061 This implies a change for the callback functions in DSA_is_prime
11062 and DSA_generate_parameters: The callback function is called once
11063 for each positive witness in the Rabin-Miller test, not just
11064 occasionally in the inner loop; and the parameters to the
11065 callback function now provide an iteration count for the outer
11066 loop rather than for the current invocation of the inner loop.
11067 DSA_generate_parameters additionally can call the callback
11068 function with an 'iteration count' of -1, meaning that a
11069 candidate has passed the trial division test (when q is generated
11070 from an application-provided seed, trial division is skipped).
11071 [Bodo Moeller]
11072
11073 *) New function BN_is_prime_fasttest that optionally does trial
11074 division before starting the Rabin-Miller test and has
11075 an additional BN_CTX * argument (whereas BN_is_prime always
11076 has to allocate at least one BN_CTX).
11077 'callback(1, -1, cb_arg)' is called when a number has passed the
11078 trial division stage.
11079 [Bodo Moeller]
11080
11081 *) Fix for bug in CRL encoding. The validity dates weren't being handled
11082 as ASN1_TIME.
11083 [Steve Henson]
11084
11085 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
11086 [Steve Henson]
11087
11088 *) New function BN_pseudo_rand().
11089 [Ulf Möller]
11090
11091 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
11092 bignum version of BN_from_montgomery() with the working code from
11093 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
11094 the comments.
11095 [Ulf Möller]
11096
11097 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
11098 made it impossible to use the same SSL_SESSION data structure in
11099 SSL2 clients in multiple threads.
11100 [Bodo Moeller]
11101
11102 *) The return value of RAND_load_file() no longer counts bytes obtained
11103 by stat(). RAND_load_file(..., -1) is new and uses the complete file
11104 to seed the PRNG (previously an explicit byte count was required).
11105 [Ulf Möller, Bodo Möller]
11106
11107 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
11108 used (char *) instead of (void *) and had casts all over the place.
11109 [Steve Henson]
11110
11111 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
11112 [Ulf Möller]
11113
11114 *) Retain source code compatibility for BN_prime_checks macro:
11115 BN_is_prime(..., BN_prime_checks, ...) now uses
11116 BN_prime_checks_for_size to determine the appropriate number of
11117 Rabin-Miller iterations.
11118 [Ulf Möller]
11119
11120 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
11121 DH_CHECK_P_NOT_SAFE_PRIME.
11122 (Check if this is true? OpenPGP calls them "strong".)
11123 [Ulf Möller]
11124
11125 *) Merge the functionality of "dh" and "gendh" programs into a new program
11126 "dhparam". The old programs are retained for now but will handle DH keys
11127 (instead of parameters) in future.
11128 [Steve Henson]
11129
11130 *) Make the ciphers, s_server and s_client programs check the return values
11131 when a new cipher list is set.
11132 [Steve Henson]
11133
11134 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
11135 ciphers. Before when the 56bit ciphers were enabled the sorting was
11136 wrong.
11137
11138 The syntax for the cipher sorting has been extended to support sorting by
11139 cipher-strength (using the strength_bits hard coded in the tables).
11140 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
11141
11142 Fix a bug in the cipher-command parser: when supplying a cipher command
11143 string with an "undefined" symbol (neither command nor alphanumeric
11144 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
11145 an error is flagged.
11146
11147 Due to the strength-sorting extension, the code of the
11148 ssl_create_cipher_list() function was completely rearranged. I hope that
11149 the readability was also increased :-)
11150 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
11151
11152 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
11153 for the first serial number and places 2 in the serial number file. This
11154 avoids problems when the root CA is created with serial number zero and
11155 the first user certificate has the same issuer name and serial number
11156 as the root CA.
11157 [Steve Henson]
11158
11159 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
11160 the new code. Add documentation for this stuff.
11161 [Steve Henson]
11162
11163 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
11164 X509_*() to X509at_*() on the grounds that they don't handle X509
11165 structures and behave in an analogous way to the X509v3 functions:
11166 they shouldn't be called directly but wrapper functions should be used
11167 instead.
11168
11169 So we also now have some wrapper functions that call the X509at functions
11170 when passed certificate requests. (TO DO: similar things can be done with
11171 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
11172 things. Some of these need some d2i or i2d and print functionality
11173 because they handle more complex structures.)
11174 [Steve Henson]
11175
11176 *) Add missing #ifndefs that caused missing symbols when building libssl
11177 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
11178 NO_RSA in ssl/s2*.c.
11179 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
11180
11181 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
11182 has a return value which indicates the quality of the random data
11183 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
11184 error queue. New function RAND_pseudo_bytes() generates output that is
11185 guaranteed to be unique but not unpredictable. RAND_add is like
11186 RAND_seed, but takes an extra argument for an entropy estimate
11187 (RAND_seed always assumes full entropy).
11188 [Ulf Möller]
11189
11190 *) Do more iterations of Rabin-Miller probable prime test (specifically,
11191 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
11192 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
11193 in crypto/bn/bn_prime.c for the complete table). This guarantees a
11194 false-positive rate of at most 2^-80 for random input.
11195 [Bodo Moeller]
11196
11197 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
11198 [Bodo Moeller]
11199
11200 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
11201 in the 0.9.5 release), this returns the chain
11202 from an X509_CTX structure with a dup of the stack and all
11203 the X509 reference counts upped: so the stack will exist
11204 after X509_CTX_cleanup() has been called. Modify pkcs12.c
11205 to use this.
11206
11207 Also make SSL_SESSION_print() print out the verify return
11208 code.
11209 [Steve Henson]
11210
11211 *) Add manpage for the pkcs12 command. Also change the default
11212 behaviour so MAC iteration counts are used unless the new
11213 -nomaciter option is used. This improves file security and
11214 only older versions of MSIE (4.0 for example) need it.
11215 [Steve Henson]
11216
11217 *) Honor the no-xxx Configure options when creating .DEF files.
11218 [Ulf Möller]
11219
11220 *) Add PKCS#10 attributes to field table: challengePassword,
11221 unstructuredName and unstructuredAddress. These are taken from
11222 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
11223 international characters are used.
11224
11225 More changes to X509_ATTRIBUTE code: allow the setting of types
11226 based on strings. Remove the 'loc' parameter when adding
11227 attributes because these will be a SET OF encoding which is sorted
11228 in ASN1 order.
11229 [Steve Henson]
11230
11231 *) Initial changes to the 'req' utility to allow request generation
11232 automation. This will allow an application to just generate a template
11233 file containing all the field values and have req construct the
11234 request.
11235
11236 Initial support for X509_ATTRIBUTE handling. Stacks of these are
11237 used all over the place including certificate requests and PKCS#7
11238 structures. They are currently handled manually where necessary with
11239 some primitive wrappers for PKCS#7. The new functions behave in a
11240 manner analogous to the X509 extension functions: they allow
11241 attributes to be looked up by NID and added.
11242
11243 Later something similar to the X509V3 code would be desirable to
11244 automatically handle the encoding, decoding and printing of the
11245 more complex types. The string types like challengePassword can
11246 be handled by the string table functions.
11247
11248 Also modified the multi byte string table handling. Now there is
11249 a 'global mask' which masks out certain types. The table itself
11250 can use the flag STABLE_NO_MASK to ignore the mask setting: this
11251 is useful when for example there is only one permissible type
11252 (as in countryName) and using the mask might result in no valid
11253 types at all.
11254 [Steve Henson]
11255
11256 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
11257 SSL_get_peer_finished to allow applications to obtain the latest
11258 Finished messages sent to the peer or expected from the peer,
11259 respectively. (SSL_get_peer_finished is usually the Finished message
11260 actually received from the peer, otherwise the protocol will be aborted.)
11261
11262 As the Finished message are message digests of the complete handshake
11263 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
11264 be used for external authentication procedures when the authentication
11265 provided by SSL/TLS is not desired or is not enough.
11266 [Bodo Moeller]
11267
11268 *) Enhanced support for Alpha Linux is added. Now ./config checks if
11269 the host supports BWX extension and if Compaq C is present on the
11270 $PATH. Just exploiting of the BWX extension results in 20-30%
11271 performance kick for some algorithms, e.g. DES and RC4 to mention
11272 a couple. Compaq C in turn generates ~20% faster code for MD5 and
11273 SHA1.
11274 [Andy Polyakov]
11275
11276 *) Add support for MS "fast SGC". This is arguably a violation of the
11277 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
11278 weak crypto and after checking the certificate is SGC a second one
11279 with strong crypto. MS SGC stops the first handshake after receiving
11280 the server certificate message and sends a second client hello. Since
11281 a server will typically do all the time consuming operations before
11282 expecting any further messages from the client (server key exchange
11283 is the most expensive) there is little difference between the two.
11284
11285 To get OpenSSL to support MS SGC we have to permit a second client
11286 hello message after we have sent server done. In addition we have to
11287 reset the MAC if we do get this second client hello.
11288 [Steve Henson]
11289
11290 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
11291 if a DER encoded private key is RSA or DSA traditional format. Changed
11292 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
11293 format DER encoded private key. Newer code should use PKCS#8 format which
11294 has the key type encoded in the ASN1 structure. Added DER private key
11295 support to pkcs8 application.
11296 [Steve Henson]
11297
11298 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
11299 ciphersuites has been selected (as required by the SSL 3/TLS 1
11300 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
11301 is set, we interpret this as a request to violate the specification
11302 (the worst that can happen is a handshake failure, and 'correct'
11303 behaviour would result in a handshake failure anyway).
11304 [Bodo Moeller]
11305
11306 *) In SSL_CTX_add_session, take into account that there might be multiple
11307 SSL_SESSION structures with the same session ID (e.g. when two threads
11308 concurrently obtain them from an external cache).
11309 The internal cache can handle only one SSL_SESSION with a given ID,
11310 so if there's a conflict, we now throw out the old one to achieve
11311 consistency.
11312 [Bodo Moeller]
11313
11314 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
11315 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
11316 some routines that use cipher OIDs: some ciphers do not have OIDs
11317 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
11318 example.
11319 [Steve Henson]
11320
11321 *) Simplify the trust setting structure and code. Now we just have
11322 two sequences of OIDs for trusted and rejected settings. These will
11323 typically have values the same as the extended key usage extension
11324 and any application specific purposes.
11325
11326 The trust checking code now has a default behaviour: it will just
11327 check for an object with the same NID as the passed id. Functions can
11328 be provided to override either the default behaviour or the behaviour
11329 for a given id. SSL client, server and email already have functions
11330 in place for compatibility: they check the NID and also return "trusted"
11331 if the certificate is self signed.
11332 [Steve Henson]
11333
11334 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
11335 traditional format into an EVP_PKEY structure.
11336 [Steve Henson]
11337
11338 *) Add a password callback function PEM_cb() which either prompts for
11339 a password if usr_data is NULL or otherwise assumes it is a null
11340 terminated password. Allow passwords to be passed on command line
11341 environment or config files in a few more utilities.
11342 [Steve Henson]
11343
11344 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
11345 keys. Add some short names for PKCS#8 PBE algorithms and allow them
11346 to be specified on the command line for the pkcs8 and pkcs12 utilities.
11347 Update documentation.
11348 [Steve Henson]
11349
11350 *) Support for ASN1 "NULL" type. This could be handled before by using
11351 ASN1_TYPE but there wasn't any function that would try to read a NULL
11352 and produce an error if it couldn't. For compatibility we also have
11353 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
11354 don't allocate anything because they don't need to.
11355 [Steve Henson]
11356
11357 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
11358 for details.
11359 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
11360
11361 *) Rebuild of the memory allocation routines used by OpenSSL code and
11362 possibly others as well. The purpose is to make an interface that
11363 provide hooks so anyone can build a separate set of allocation and
11364 deallocation routines to be used by OpenSSL, for example memory
11365 pool implementations, or something else, which was previously hard
11366 since Malloc(), Realloc() and Free() were defined as macros having
11367 the values malloc, realloc and free, respectively (except for Win32
11368 compilations). The same is provided for memory debugging code.
11369 OpenSSL already comes with functionality to find memory leaks, but
11370 this gives people a chance to debug other memory problems.
11371
11372 With these changes, a new set of functions and macros have appeared:
11373
11374 CRYPTO_set_mem_debug_functions() [F]
11375 CRYPTO_get_mem_debug_functions() [F]
11376 CRYPTO_dbg_set_options() [F]
11377 CRYPTO_dbg_get_options() [F]
11378 CRYPTO_malloc_debug_init() [M]
11379
11380 The memory debug functions are NULL by default, unless the library
11381 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
11382 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
11383 gives the standard debugging functions that come with OpenSSL) or
11384 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
11385 provided by the library user) must be used. When the standard
11386 debugging functions are used, CRYPTO_dbg_set_options can be used to
11387 request additional information:
11388 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
11389 the CRYPTO_MDEBUG_xxx macro when compiling the library.
11390
11391 Also, things like CRYPTO_set_mem_functions will always give the
11392 expected result (the new set of functions is used for allocation
11393 and deallocation) at all times, regardless of platform and compiler
11394 options.
11395
11396 To finish it up, some functions that were never use in any other
11397 way than through macros have a new API and new semantic:
11398
11399 CRYPTO_dbg_malloc()
11400 CRYPTO_dbg_realloc()
11401 CRYPTO_dbg_free()
11402
11403 All macros of value have retained their old syntax.
11404 [Richard Levitte and Bodo Moeller]
11405
11406 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
11407 ordering of SMIMECapabilities wasn't in "strength order" and there
11408 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
11409 algorithm.
11410 [Steve Henson]
11411
11412 *) Some ASN1 types with illegal zero length encoding (INTEGER,
11413 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
11414 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
11415
11416 *) Merge in my S/MIME library for OpenSSL. This provides a simple
11417 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
11418 functionality to handle multipart/signed properly) and a utility
11419 called 'smime' to call all this stuff. This is based on code I
11420 originally wrote for Celo who have kindly allowed it to be
11421 included in OpenSSL.
11422 [Steve Henson]
11423
11424 *) Add variants des_set_key_checked and des_set_key_unchecked of
11425 des_set_key (aka des_key_sched). Global variable des_check_key
11426 decides which of these is called by des_set_key; this way
11427 des_check_key behaves as it always did, but applications and
11428 the library itself, which was buggy for des_check_key == 1,
11429 have a cleaner way to pick the version they need.
11430 [Bodo Moeller]
11431
11432 *) New function PKCS12_newpass() which changes the password of a
11433 PKCS12 structure.
11434 [Steve Henson]
11435
11436 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
11437 dynamic mix. In both cases the ids can be used as an index into the
11438 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
11439 functions so they accept a list of the field values and the
11440 application doesn't need to directly manipulate the X509_TRUST
11441 structure.
11442 [Steve Henson]
11443
11444 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
11445 need initialising.
11446 [Steve Henson]
11447
11448 *) Modify the way the V3 extension code looks up extensions. This now
11449 works in a similar way to the object code: we have some "standard"
11450 extensions in a static table which is searched with OBJ_bsearch()
11451 and the application can add dynamic ones if needed. The file
11452 crypto/x509v3/ext_dat.h now has the info: this file needs to be
11453 updated whenever a new extension is added to the core code and kept
11454 in ext_nid order. There is a simple program 'tabtest.c' which checks
11455 this. New extensions are not added too often so this file can readily
11456 be maintained manually.
11457
11458 There are two big advantages in doing things this way. The extensions
11459 can be looked up immediately and no longer need to be "added" using
11460 X509V3_add_standard_extensions(): this function now does nothing.
11461 [Side note: I get *lots* of email saying the extension code doesn't
11462 work because people forget to call this function]
11463 Also no dynamic allocation is done unless new extensions are added:
11464 so if we don't add custom extensions there is no need to call
11465 X509V3_EXT_cleanup().
11466 [Steve Henson]
11467
11468 *) Modify enc utility's salting as follows: make salting the default. Add a
11469 magic header, so unsalted files fail gracefully instead of just decrypting
11470 to garbage. This is because not salting is a big security hole, so people
11471 should be discouraged from doing it.
11472 [Ben Laurie]
11473
11474 *) Fixes and enhancements to the 'x509' utility. It allowed a message
11475 digest to be passed on the command line but it only used this
11476 parameter when signing a certificate. Modified so all relevant
11477 operations are affected by the digest parameter including the
11478 -fingerprint and -x509toreq options. Also -x509toreq choked if a
11479 DSA key was used because it didn't fix the digest.
11480 [Steve Henson]
11481
11482 *) Initial certificate chain verify code. Currently tests the untrusted
11483 certificates for consistency with the verify purpose (which is set
11484 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11485
11486 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11487 this is because it will reject chains with invalid extensions whereas
11488 every previous version of OpenSSL and SSLeay made no checks at all.
11489
11490 Trust code: checks the root CA for the relevant trust settings. Trust
11491 settings have an initial value consistent with the verify purpose: e.g.
11492 if the verify purpose is for SSL client use it expects the CA to be
11493 trusted for SSL client use. However the default value can be changed to
11494 permit custom trust settings: one example of this would be to only trust
11495 certificates from a specific "secure" set of CAs.
11496
11497 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11498 which should be used for version portability: especially since the
11499 verify structure is likely to change more often now.
11500
11501 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11502 to set them. If not set then assume SSL clients will verify SSL servers
11503 and vice versa.
11504
11505 Two new options to the verify program: -untrusted allows a set of
11506 untrusted certificates to be passed in and -purpose which sets the
11507 intended purpose of the certificate. If a purpose is set then the
11508 new chain verify code is used to check extension consistency.
11509 [Steve Henson]
11510
11511 *) Support for the authority information access extension.
11512 [Steve Henson]
11513
11514 *) Modify RSA and DSA PEM read routines to transparently handle
11515 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11516 public keys in a format compatible with certificate
11517 SubjectPublicKeyInfo structures. Unfortunately there were already
11518 functions called *_PublicKey_* which used various odd formats so
11519 these are retained for compatibility: however the DSA variants were
11520 never in a public release so they have been deleted. Changed dsa/rsa
11521 utilities to handle the new format: note no releases ever handled public
11522 keys so we should be OK.
11523
11524 The primary motivation for this change is to avoid the same fiasco
11525 that dogs private keys: there are several incompatible private key
11526 formats some of which are standard and some OpenSSL specific and
11527 require various evil hacks to allow partial transparent handling and
11528 even then it doesn't work with DER formats. Given the option anything
11529 other than PKCS#8 should be dumped: but the other formats have to
11530 stay in the name of compatibility.
11531
11532 With public keys and the benefit of hindsight one standard format
11533 is used which works with EVP_PKEY, RSA or DSA structures: though
11534 it clearly returns an error if you try to read the wrong kind of key.
11535
11536 Added a -pubkey option to the 'x509' utility to output the public key.
11537 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11538 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11539 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11540 that do the same as the EVP_PKEY_assign_*() except they up the
11541 reference count of the added key (they don't "swallow" the
11542 supplied key).
11543 [Steve Henson]
11544
11545 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11546 CRLs would fail if the file contained no certificates or no CRLs:
11547 added a new function to read in both types and return the number
11548 read: this means that if none are read it will be an error. The
11549 DER versions of the certificate and CRL reader would always fail
11550 because it isn't possible to mix certificates and CRLs in DER format
11551 without choking one or the other routine. Changed this to just read
11552 a certificate: this is the best we can do. Also modified the code
11553 in apps/verify.c to take notice of return codes: it was previously
11554 attempting to read in certificates from NULL pointers and ignoring
11555 any errors: this is one reason why the cert and CRL reader seemed
11556 to work. It doesn't check return codes from the default certificate
11557 routines: these may well fail if the certificates aren't installed.
11558 [Steve Henson]
11559
11560 *) Code to support otherName option in GeneralName.
11561 [Steve Henson]
11562
11563 *) First update to verify code. Change the verify utility
11564 so it warns if it is passed a self signed certificate:
11565 for consistency with the normal behaviour. X509_verify
11566 has been modified to it will now verify a self signed
11567 certificate if *exactly* the same certificate appears
11568 in the store: it was previously impossible to trust a
11569 single self signed certificate. This means that:
11570 openssl verify ss.pem
11571 now gives a warning about a self signed certificate but
11572 openssl verify -CAfile ss.pem ss.pem
11573 is OK.
11574 [Steve Henson]
11575
11576 *) For servers, store verify_result in SSL_SESSION data structure
11577 (and add it to external session representation).
11578 This is needed when client certificate verifications fails,
11579 but an application-provided verification callback (set by
11580 SSL_CTX_set_cert_verify_callback) allows accepting the session
11581 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11582 but returns 1): When the session is reused, we have to set
11583 ssl->verify_result to the appropriate error code to avoid
11584 security holes.
11585 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11586
11587 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11588 case in PKCS7_dataInit() where the signed PKCS7 structure
11589 didn't contain any existing data because it was being created.
11590 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11591
11592 *) Add a salt to the key derivation routines in enc.c. This
11593 forms the first 8 bytes of the encrypted file. Also add a
11594 -S option to allow a salt to be input on the command line.
11595 [Steve Henson]
11596
11597 *) New function X509_cmp(). Oddly enough there wasn't a function
11598 to compare two certificates. We do this by working out the SHA1
11599 hash and comparing that. X509_cmp() will be needed by the trust
11600 code.
11601 [Steve Henson]
11602
11603 *) SSL_get1_session() is like SSL_get_session(), but increments
11604 the reference count in the SSL_SESSION returned.
11605 [Geoff Thorpe <geoff@eu.c2.net>]
11606
11607 *) Fix for 'req': it was adding a null to request attributes.
11608 Also change the X509_LOOKUP and X509_INFO code to handle
11609 certificate auxiliary information.
11610 [Steve Henson]
11611
11612 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11613 the 'enc' command.
11614 [Steve Henson]
11615
11616 *) Add the possibility to add extra information to the memory leak
11617 detecting output, to form tracebacks, showing from where each
11618 allocation was originated: CRYPTO_push_info("constant string") adds
11619 the string plus current file name and line number to a per-thread
11620 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11621 is like calling CYRPTO_pop_info() until the stack is empty.
11622 Also updated memory leak detection code to be multi-thread-safe.
11623 [Richard Levitte]
11624
11625 *) Add options -text and -noout to pkcs7 utility and delete the
11626 encryption options which never did anything. Update docs.
11627 [Steve Henson]
11628
11629 *) Add options to some of the utilities to allow the pass phrase
11630 to be included on either the command line (not recommended on
11631 OSes like Unix) or read from the environment. Update the
11632 manpages and fix a few bugs.
11633 [Steve Henson]
11634
11635 *) Add a few manpages for some of the openssl commands.
11636 [Steve Henson]
11637
11638 *) Fix the -revoke option in ca. It was freeing up memory twice,
11639 leaking and not finding already revoked certificates.
11640 [Steve Henson]
11641
11642 *) Extensive changes to support certificate auxiliary information.
11643 This involves the use of X509_CERT_AUX structure and X509_AUX
11644 functions. An X509_AUX function such as PEM_read_X509_AUX()
11645 can still read in a certificate file in the usual way but it
11646 will also read in any additional "auxiliary information". By
11647 doing things this way a fair degree of compatibility can be
11648 retained: existing certificates can have this information added
11649 using the new 'x509' options.
11650
11651 Current auxiliary information includes an "alias" and some trust
11652 settings. The trust settings will ultimately be used in enhanced
11653 certificate chain verification routines: currently a certificate
11654 can only be trusted if it is self signed and then it is trusted
11655 for all purposes.
11656 [Steve Henson]
11657
11658 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
11659 The problem was that one of the replacement routines had not been working
11660 since SSLeay releases. For now the offending routine has been replaced
11661 with non-optimised assembler. Even so, this now gives around 95%
11662 performance improvement for 1024 bit RSA signs.
11663 [Mark Cox]
11664
11665 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
11666 handling. Most clients have the effective key size in bits equal to
11667 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
11668 A few however don't do this and instead use the size of the decrypted key
11669 to determine the RC2 key length and the AlgorithmIdentifier to determine
11670 the effective key length. In this case the effective key length can still
11671 be 40 bits but the key length can be 168 bits for example. This is fixed
11672 by manually forcing an RC2 key into the EVP_PKEY structure because the
11673 EVP code can't currently handle unusual RC2 key sizes: it always assumes
11674 the key length and effective key length are equal.
11675 [Steve Henson]
11676
11677 *) Add a bunch of functions that should simplify the creation of
11678 X509_NAME structures. Now you should be able to do:
11679 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
11680 and have it automatically work out the correct field type and fill in
11681 the structures. The more adventurous can try:
11682 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
11683 and it will (hopefully) work out the correct multibyte encoding.
11684 [Steve Henson]
11685
11686 *) Change the 'req' utility to use the new field handling and multibyte
11687 copy routines. Before the DN field creation was handled in an ad hoc
11688 way in req, ca, and x509 which was rather broken and didn't support
11689 BMPStrings or UTF8Strings. Since some software doesn't implement
11690 BMPStrings or UTF8Strings yet, they can be enabled using the config file
11691 using the dirstring_type option. See the new comment in the default
11692 openssl.cnf for more info.
11693 [Steve Henson]
11694
11695 *) Make crypto/rand/md_rand.c more robust:
11696 - Assure unique random numbers after fork().
11697 - Make sure that concurrent threads access the global counter and
11698 md serializably so that we never lose entropy in them
11699 or use exactly the same state in multiple threads.
11700 Access to the large state is not always serializable because
11701 the additional locking could be a performance killer, and
11702 md should be large enough anyway.
11703 [Bodo Moeller]
11704
11705 *) New file apps/app_rand.c with commonly needed functionality
11706 for handling the random seed file.
11707
11708 Use the random seed file in some applications that previously did not:
11709 ca,
11710 dsaparam -genkey (which also ignored its '-rand' option),
11711 s_client,
11712 s_server,
11713 x509 (when signing).
11714 Except on systems with /dev/urandom, it is crucial to have a random
11715 seed file at least for key creation, DSA signing, and for DH exchanges;
11716 for RSA signatures we could do without one.
11717
11718 gendh and gendsa (unlike genrsa) used to read only the first byte
11719 of each file listed in the '-rand' option. The function as previously
11720 found in genrsa is now in app_rand.c and is used by all programs
11721 that support '-rand'.
11722 [Bodo Moeller]
11723
11724 *) In RAND_write_file, use mode 0600 for creating files;
11725 don't just chmod when it may be too late.
11726 [Bodo Moeller]
11727
11728 *) Report an error from X509_STORE_load_locations
11729 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
11730 [Bill Perry]
11731
11732 *) New function ASN1_mbstring_copy() this copies a string in either
11733 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
11734 into an ASN1_STRING type. A mask of permissible types is passed
11735 and it chooses the "minimal" type to use or an error if not type
11736 is suitable.
11737 [Steve Henson]
11738
11739 *) Add function equivalents to the various macros in asn1.h. The old
11740 macros are retained with an M_ prefix. Code inside the library can
11741 use the M_ macros. External code (including the openssl utility)
11742 should *NOT* in order to be "shared library friendly".
11743 [Steve Henson]
11744
11745 *) Add various functions that can check a certificate's extensions
11746 to see if it usable for various purposes such as SSL client,
11747 server or S/MIME and CAs of these types. This is currently
11748 VERY EXPERIMENTAL but will ultimately be used for certificate chain
11749 verification. Also added a -purpose flag to x509 utility to
11750 print out all the purposes.
11751 [Steve Henson]
11752
11753 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
11754 functions.
11755 [Steve Henson]
11756
11757 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
11758 for, obtain and decode and extension and obtain its critical flag.
11759 This allows all the necessary extension code to be handled in a
11760 single function call.
11761 [Steve Henson]
11762
11763 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
11764 platforms. See crypto/rc4/rc4_enc.c for further details.
11765 [Andy Polyakov]
11766
11767 *) New -noout option to asn1parse. This causes no output to be produced
11768 its main use is when combined with -strparse and -out to extract data
11769 from a file (which may not be in ASN.1 format).
11770 [Steve Henson]
11771
11772 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
11773 when producing the local key id.
11774 [Richard Levitte <levitte@stacken.kth.se>]
11775
11776 *) New option -dhparam in s_server. This allows a DH parameter file to be
11777 stated explicitly. If it is not stated then it tries the first server
11778 certificate file. The previous behaviour hard coded the filename
11779 "server.pem".
11780 [Steve Henson]
11781
11782 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11783 a public key to be input or output. For example:
11784 openssl rsa -in key.pem -pubout -out pubkey.pem
11785 Also added necessary DSA public key functions to handle this.
11786 [Steve Henson]
11787
11788 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11789 in the message. This was handled by allowing
11790 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11791 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11792
11793 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11794 to the end of the strings whereas this didn't. This would cause problems
11795 if strings read with d2i_ASN1_bytes() were later modified.
11796 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11797
11798 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11799 data and it contains EOF it will end up returning an error. This is
11800 caused by input 46 bytes long. The cause is due to the way base64
11801 BIOs find the start of base64 encoded data. They do this by trying a
11802 trial decode on each line until they find one that works. When they
11803 do a flag is set and it starts again knowing it can pass all the
11804 data directly through the decoder. Unfortunately it doesn't reset
11805 the context it uses. This means that if EOF is reached an attempt
11806 is made to pass two EOFs through the context and this causes the
11807 resulting error. This can also cause other problems as well. As is
11808 usual with these problems it takes *ages* to find and the fix is
11809 trivial: move one line.
11810 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
11811
11812 *) Ugly workaround to get s_client and s_server working under Windows. The
11813 old code wouldn't work because it needed to select() on sockets and the
11814 tty (for keypresses and to see if data could be written). Win32 only
11815 supports select() on sockets so we select() with a 1s timeout on the
11816 sockets and then see if any characters are waiting to be read, if none
11817 are present then we retry, we also assume we can always write data to
11818 the tty. This isn't nice because the code then blocks until we've
11819 received a complete line of data and it is effectively polling the
11820 keyboard at 1s intervals: however it's quite a bit better than not
11821 working at all :-) A dedicated Windows application might handle this
11822 with an event loop for example.
11823 [Steve Henson]
11824
11825 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
11826 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
11827 will be called when RSA_sign() and RSA_verify() are used. This is useful
11828 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
11829 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
11830 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
11831 This necessitated the support of an extra signature type NID_md5_sha1
11832 for SSL signatures and modifications to the SSL library to use it instead
11833 of calling RSA_public_decrypt() and RSA_private_encrypt().
11834 [Steve Henson]
11835
11836 *) Add new -verify -CAfile and -CApath options to the crl program, these
11837 will lookup a CRL issuers certificate and verify the signature in a
11838 similar way to the verify program. Tidy up the crl program so it
11839 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
11840 less strict. It will now permit CRL extensions even if it is not
11841 a V2 CRL: this will allow it to tolerate some broken CRLs.
11842 [Steve Henson]
11843
11844 *) Initialize all non-automatic variables each time one of the openssl
11845 sub-programs is started (this is necessary as they may be started
11846 multiple times from the "OpenSSL>" prompt).
11847 [Lennart Bang, Bodo Moeller]
11848
11849 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
11850 removing all other RSA functionality (this is what NO_RSA does). This
11851 is so (for example) those in the US can disable those operations covered
11852 by the RSA patent while allowing storage and parsing of RSA keys and RSA
11853 key generation.
11854 [Steve Henson]
11855
11856 *) Non-copying interface to BIO pairs.
11857 (still largely untested)
11858 [Bodo Moeller]
11859
11860 *) New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
11861 ASCII string. This was handled independently in various places before.
11862 [Steve Henson]
11863
11864 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
11865 UTF8 strings a character at a time.
11866 [Steve Henson]
11867
11868 *) Use client_version from client hello to select the protocol
11869 (s23_srvr.c) and for RSA client key exchange verification
11870 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
11871 [Bodo Moeller]
11872
11873 *) Add various utility functions to handle SPKACs, these were previously
11874 handled by poking round in the structure internals. Added new function
11875 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
11876 print, verify and generate SPKACs. Based on an original idea from
11877 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
11878 [Steve Henson]
11879
11880 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
11881 [Andy Polyakov]
11882
11883 *) Allow the config file extension section to be overwritten on the
11884 command line. Based on an original idea from Massimiliano Pala
11885 <madwolf@comune.modena.it>. The new option is called -extensions
11886 and can be applied to ca, req and x509. Also -reqexts to override
11887 the request extensions in req and -crlexts to override the crl extensions
11888 in ca.
11889 [Steve Henson]
11890
11891 *) Add new feature to the SPKAC handling in ca. Now you can include
11892 the same field multiple times by preceding it by "XXXX." for example:
11893 1.OU="Unit name 1"
11894 2.OU="Unit name 2"
11895 this is the same syntax as used in the req config file.
11896 [Steve Henson]
11897
11898 *) Allow certificate extensions to be added to certificate requests. These
11899 are specified in a 'req_extensions' option of the req section of the
11900 config file. They can be printed out with the -text option to req but
11901 are otherwise ignored at present.
11902 [Steve Henson]
11903
11904 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11905 data read consists of only the final block it would not decrypted because
11906 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11907 A misplaced 'break' also meant the decrypted final block might not be
11908 copied until the next read.
11909 [Steve Henson]
11910
11911 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11912 a few extra parameters to the DH structure: these will be useful if
11913 for example we want the value of 'q' or implement X9.42 DH.
11914 [Steve Henson]
11915
11916 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11917 provides hooks that allow the default DSA functions or functions on a
11918 "per key" basis to be replaced. This allows hardware acceleration and
11919 hardware key storage to be handled without major modification to the
11920 library. Also added low level modexp hooks and CRYPTO_EX structure and
11921 associated functions.
11922 [Steve Henson]
11923
11924 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11925 as "read only": it can't be written to and the buffer it points to will
11926 not be freed. Reading from a read only BIO is much more efficient than
11927 a normal memory BIO. This was added because there are several times when
11928 an area of memory needs to be read from a BIO. The previous method was
11929 to create a memory BIO and write the data to it, this results in two
11930 copies of the data and an O(n^2) reading algorithm. There is a new
11931 function BIO_new_mem_buf() which creates a read only memory BIO from
11932 an area of memory. Also modified the PKCS#7 routines to use read only
11933 memory BIOs.
11934 [Steve Henson]
11935
11936 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11937 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11938 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11939 but a retry condition occurred while trying to read the rest.
11940 [Bodo Moeller]
11941
11942 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11943 NID_pkcs7_encrypted by default: this was wrong since this should almost
11944 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11945 the encrypted data type: this is a more sensible place to put it and it
11946 allows the PKCS#12 code to be tidied up that duplicated this
11947 functionality.
11948 [Steve Henson]
11949
11950 *) Changed obj_dat.pl script so it takes its input and output files on
11951 the command line. This should avoid shell escape redirection problems
11952 under Win32.
11953 [Steve Henson]
11954
11955 *) Initial support for certificate extension requests, these are included
11956 in things like Xenroll certificate requests. Included functions to allow
11957 extensions to be obtained and added.
11958 [Steve Henson]
11959
11960 *) -crlf option to s_client and s_server for sending newlines as
11961 CRLF (as required by many protocols).
11962 [Bodo Moeller]
11963
11964 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11965
11966 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11967 [Ralf S. Engelschall]
11968
11969 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11970 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11971
11972 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11973 program.
11974 [Steve Henson]
11975
11976 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11977 DH parameters/keys (q is lost during that conversion, but the resulting
11978 DH parameters contain its length).
11979
11980 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11981 much faster than DH_generate_parameters (which creates parameters
11982 where p = 2*q + 1), and also the smaller q makes DH computations
11983 much more efficient (160-bit exponentiation instead of 1024-bit
11984 exponentiation); so this provides a convenient way to support DHE
11985 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11986 utter importance to use
11987 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11988 or
11989 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11990 when such DH parameters are used, because otherwise small subgroup
11991 attacks may become possible!
11992 [Bodo Moeller]
11993
11994 *) Avoid memory leak in i2d_DHparams.
11995 [Bodo Moeller]
11996
11997 *) Allow the -k option to be used more than once in the enc program:
11998 this allows the same encrypted message to be read by multiple recipients.
11999 [Steve Henson]
12000
12001 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
12002 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
12003 it will always use the numerical form of the OID, even if it has a short
12004 or long name.
12005 [Steve Henson]
12006
12007 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
12008 method only got called if p,q,dmp1,dmq1,iqmp components were present,
12009 otherwise bn_mod_exp was called. In the case of hardware keys for example
12010 no private key components need be present and it might store extra data
12011 in the RSA structure, which cannot be accessed from bn_mod_exp.
12012 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
12013 private key operations.
12014 [Steve Henson]
12015
12016 *) Added support for SPARC Linux.
12017 [Andy Polyakov]
12018
12019 *) pem_password_cb function type incompatibly changed from
12020 typedef int pem_password_cb(char *buf, int size, int rwflag);
12021 to
12022 ....(char *buf, int size, int rwflag, void *userdata);
12023 so that applications can pass data to their callbacks:
12024 The PEM[_ASN1]_{read,write}... functions and macros now take an
12025 additional void * argument, which is just handed through whenever
12026 the password callback is called.
12027 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
12028
12029 New function SSL_CTX_set_default_passwd_cb_userdata.
12030
12031 Compatibility note: As many C implementations push function arguments
12032 onto the stack in reverse order, the new library version is likely to
12033 interoperate with programs that have been compiled with the old
12034 pem_password_cb definition (PEM_whatever takes some data that
12035 happens to be on the stack as its last argument, and the callback
12036 just ignores this garbage); but there is no guarantee whatsoever that
12037 this will work.
12038
12039 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
12040 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
12041 problems not only on Windows, but also on some Unix platforms.
12042 To avoid problematic command lines, these definitions are now in an
12043 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
12044 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
12045 [Bodo Moeller]
12046
12047 *) MIPS III/IV assembler module is reimplemented.
12048 [Andy Polyakov]
12049
12050 *) More DES library cleanups: remove references to srand/rand and
12051 delete an unused file.
12052 [Ulf Möller]
12053
12054 *) Add support for the free Netwide assembler (NASM) under Win32,
12055 since not many people have MASM (ml) and it can be hard to obtain.
12056 This is currently experimental but it seems to work OK and pass all
12057 the tests. Check out INSTALL.W32 for info.
12058 [Steve Henson]
12059
12060 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
12061 without temporary keys kept an extra copy of the server key,
12062 and connections with temporary keys did not free everything in case
12063 of an error.
12064 [Bodo Moeller]
12065
12066 *) New function RSA_check_key and new openssl rsa option -check
12067 for verifying the consistency of RSA keys.
12068 [Ulf Moeller, Bodo Moeller]
12069
12070 *) Various changes to make Win32 compile work:
12071 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
12072 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
12073 comparison" warnings.
12074 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
12075 [Steve Henson]
12076
12077 *) Add a debugging option to PKCS#5 v2 key generation function: when
12078 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
12079 derived keys are printed to stderr.
12080 [Steve Henson]
12081
12082 *) Copy the flags in ASN1_STRING_dup().
12083 [Roman E. Pavlov <pre@mo.msk.ru>]
12084
12085 *) The x509 application mishandled signing requests containing DSA
12086 keys when the signing key was also DSA and the parameters didn't match.
12087
12088 It was supposed to omit the parameters when they matched the signing key:
12089 the verifying software was then supposed to automatically use the CA's
12090 parameters if they were absent from the end user certificate.
12091
12092 Omitting parameters is no longer recommended. The test was also
12093 the wrong way round! This was probably due to unusual behaviour in
12094 EVP_cmp_parameters() which returns 1 if the parameters match.
12095 This meant that parameters were omitted when they *didn't* match and
12096 the certificate was useless. Certificates signed with 'ca' didn't have
12097 this bug.
12098 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
12099
12100 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
12101 The interface is as follows:
12102 Applications can use
12103 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
12104 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
12105 "off" is now the default.
12106 The library internally uses
12107 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
12108 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
12109 to disable memory-checking temporarily.
12110
12111 Some inconsistent states that previously were possible (and were
12112 even the default) are now avoided.
12113
12114 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
12115 with each memory chunk allocated; this is occasionally more helpful
12116 than just having a counter.
12117
12118 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
12119
12120 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
12121 extensions.
12122 [Bodo Moeller]
12123
12124 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
12125 which largely parallels "options", but is for changing API behaviour,
12126 whereas "options" are about protocol behaviour.
12127 Initial "mode" flags are:
12128
12129 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
12130 a single record has been written.
12131 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
12132 retries use the same buffer location.
12133 (But all of the contents must be
12134 copied!)
12135 [Bodo Moeller]
12136
12137 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
12138 worked.
12139
12140 *) Fix problems with no-hmac etc.
12141 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
12142
12143 *) New functions RSA_get_default_method(), RSA_set_method() and
12144 RSA_get_method(). These allows replacement of RSA_METHODs without having
12145 to mess around with the internals of an RSA structure.
12146 [Steve Henson]
12147
12148 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
12149 Also really enable memory leak checks in openssl.c and in some
12150 test programs.
12151 [Chad C. Mulligan, Bodo Moeller]
12152
12153 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
12154 up the length of negative integers. This has now been simplified to just
12155 store the length when it is first determined and use it later, rather
12156 than trying to keep track of where data is copied and updating it to
12157 point to the end.
12158 [Steve Henson, reported by Brien Wheeler
12159 <bwheeler@authentica-security.com>]
12160
12161 *) Add a new function PKCS7_signatureVerify. This allows the verification
12162 of a PKCS#7 signature but with the signing certificate passed to the
12163 function itself. This contrasts with PKCS7_dataVerify which assumes the
12164 certificate is present in the PKCS#7 structure. This isn't always the
12165 case: certificates can be omitted from a PKCS#7 structure and be
12166 distributed by "out of band" means (such as a certificate database).
12167 [Steve Henson]
12168
12169 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
12170 function prototypes in pem.h, also change util/mkdef.pl to add the
12171 necessary function names.
12172 [Steve Henson]
12173
12174 *) mk1mf.pl (used by Windows builds) did not properly read the
12175 options set by Configure in the top level Makefile, and Configure
12176 was not even able to write more than one option correctly.
12177 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
12178 [Bodo Moeller]
12179
12180 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
12181 file to be loaded from a BIO or FILE pointer. The BIO version will
12182 for example allow memory BIOs to contain config info.
12183 [Steve Henson]
12184
12185 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
12186 Whoever hopes to achieve shared-library compatibility across versions
12187 must use this, not the compile-time macro.
12188 (Exercise 0.9.4: Which is the minimum library version required by
12189 such programs?)
12190 Note: All this applies only to multi-threaded programs, others don't
12191 need locks.
12192 [Bodo Moeller]
12193
12194 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
12195 through a BIO pair triggered the default case, i.e.
12196 SSLerr(...,SSL_R_UNKNOWN_STATE).
12197 [Bodo Moeller]
12198
12199 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
12200 can use the SSL library even if none of the specific BIOs is
12201 appropriate.
12202 [Bodo Moeller]
12203
12204 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
12205 for the encoded length.
12206 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
12207
12208 *) Add initial documentation of the X509V3 functions.
12209 [Steve Henson]
12210
12211 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
12212 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
12213 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
12214 secure PKCS#8 private key format with a high iteration count.
12215 [Steve Henson]
12216
12217 *) Fix determination of Perl interpreter: A perl or perl5
12218 _directory_ in $PATH was also accepted as the interpreter.
12219 [Ralf S. Engelschall]
12220
12221 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
12222 wrong with it but it was very old and did things like calling
12223 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
12224 unusual formatting.
12225 [Steve Henson]
12226
12227 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
12228 to use the new extension code.
12229 [Steve Henson]
12230
12231 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
12232 with macros. This should make it easier to change their form, add extra
12233 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
12234 constant.
12235 [Steve Henson]
12236
12237 *) Add to configuration table a new entry that can specify an alternative
12238 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
12239 according to Mark Crispin <MRC@Panda.COM>.
12240 [Bodo Moeller]
12241
12242 #if 0
12243 *) DES CBC did not update the IV. Weird.
12244 [Ben Laurie]
12245 #else
12246 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
12247 Changing the behaviour of the former might break existing programs --
12248 where IV updating is needed, des_ncbc_encrypt can be used.
12249 #endif
12250
12251 *) When bntest is run from "make test" it drives bc to check its
12252 calculations, as well as internally checking them. If an internal check
12253 fails, it needs to cause bc to give a non-zero result or make test carries
12254 on without noticing the failure. Fixed.
12255 [Ben Laurie]
12256
12257 *) DES library cleanups.
12258 [Ulf Möller]
12259
12260 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
12261 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
12262 ciphers. NOTE: although the key derivation function has been verified
12263 against some published test vectors it has not been extensively tested
12264 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
12265 of v2.0.
12266 [Steve Henson]
12267
12268 *) Instead of "mkdir -p", which is not fully portable, use new
12269 Perl script "util/mkdir-p.pl".
12270 [Bodo Moeller]
12271
12272 *) Rewrite the way password based encryption (PBE) is handled. It used to
12273 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
12274 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
12275 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
12276 the 'parameter' field of the AlgorithmIdentifier is passed to the
12277 underlying key generation function so it must do its own ASN1 parsing.
12278 This has also changed the EVP_PBE_CipherInit() function which now has a
12279 'parameter' argument instead of literal salt and iteration count values
12280 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
12281 [Steve Henson]
12282
12283 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
12284 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
12285 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
12286 KEY" because this clashed with PKCS#8 unencrypted string. Since this
12287 value was just used as a "magic string" and not used directly its
12288 value doesn't matter.
12289 [Steve Henson]
12290
12291 *) Introduce some semblance of const correctness to BN. Shame C doesn't
12292 support mutable.
12293 [Ben Laurie]
12294
12295 *) "linux-sparc64" configuration (ultrapenguin).
12296 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
12297 "linux-sparc" configuration.
12298 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
12299
12300 *) config now generates no-xxx options for missing ciphers.
12301 [Ulf Möller]
12302
12303 *) Support the EBCDIC character set (work in progress).
12304 File ebcdic.c not yet included because it has a different license.
12305 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12306
12307 *) Support BS2000/OSD-POSIX.
12308 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12309
12310 *) Make callbacks for key generation use void * instead of char *.
12311 [Ben Laurie]
12312
12313 *) Make S/MIME samples compile (not yet tested).
12314 [Ben Laurie]
12315
12316 *) Additional typesafe stacks.
12317 [Ben Laurie]
12318
12319 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
12320 [Bodo Moeller]
12321
12322
12323 Changes between 0.9.3 and 0.9.3a [29 May 1999]
12324
12325 *) New configuration variant "sco5-gcc".
12326
12327 *) Updated some demos.
12328 [Sean O Riordain, Wade Scholine]
12329
12330 *) Add missing BIO_free at exit of pkcs12 application.
12331 [Wu Zhigang]
12332
12333 *) Fix memory leak in conf.c.
12334 [Steve Henson]
12335
12336 *) Updates for Win32 to assembler version of MD5.
12337 [Steve Henson]
12338
12339 *) Set #! path to perl in apps/der_chop to where we found it
12340 instead of using a fixed path.
12341 [Bodo Moeller]
12342
12343 *) SHA library changes for irix64-mips4-cc.
12344 [Andy Polyakov]
12345
12346 *) Improvements for VMS support.
12347 [Richard Levitte]
12348
12349
12350 Changes between 0.9.2b and 0.9.3 [24 May 1999]
12351
12352 *) Bignum library bug fix. IRIX 6 passes "make test" now!
12353 This also avoids the problems with SC4.2 and unpatched SC5.
12354 [Andy Polyakov <appro@fy.chalmers.se>]
12355
12356 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
12357 These are required because of the typesafe stack would otherwise break
12358 existing code. If old code used a structure member which used to be STACK
12359 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
12360 sk_num or sk_value it would produce an error because the num, data members
12361 are not present in STACK_OF. Now it just produces a warning. sk_set
12362 replaces the old method of assigning a value to sk_value
12363 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
12364 that does this will no longer work (and should use sk_set instead) but
12365 this could be regarded as a "questionable" behaviour anyway.
12366 [Steve Henson]
12367
12368 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
12369 correctly handle encrypted S/MIME data.
12370 [Steve Henson]
12371
12372 *) Change type of various DES function arguments from des_cblock
12373 (which means, in function argument declarations, pointer to char)
12374 to des_cblock * (meaning pointer to array with 8 char elements),
12375 which allows the compiler to do more typechecking; it was like
12376 that back in SSLeay, but with lots of ugly casts.
12377
12378 Introduce new type const_des_cblock.
12379 [Bodo Moeller]
12380
12381 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
12382 problems: find RecipientInfo structure that matches recipient certificate
12383 and initialise the ASN1 structures properly based on passed cipher.
12384 [Steve Henson]
12385
12386 *) Belatedly make the BN tests actually check the results.
12387 [Ben Laurie]
12388
12389 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
12390 to and from BNs: it was completely broken. New compilation option
12391 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
12392 key elements as negative integers.
12393 [Steve Henson]
12394
12395 *) Reorganize and speed up MD5.
12396 [Andy Polyakov <appro@fy.chalmers.se>]
12397
12398 *) VMS support.
12399 [Richard Levitte <richard@levitte.org>]
12400
12401 *) New option -out to asn1parse to allow the parsed structure to be
12402 output to a file. This is most useful when combined with the -strparse
12403 option to examine the output of things like OCTET STRINGS.
12404 [Steve Henson]
12405
12406 *) Make SSL library a little more fool-proof by not requiring any longer
12407 that SSL_set_{accept,connect}_state be called before
12408 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
12409 in many applications because usually everything *appeared* to work as
12410 intended anyway -- now it really works as intended).
12411 [Bodo Moeller]
12412
12413 *) Move openssl.cnf out of lib/.
12414 [Ulf Möller]
12415
12416 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
12417 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
12418 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
12419 [Ralf S. Engelschall]
12420
12421 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
12422 handle PKCS#7 enveloped data properly.
12423 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
12424
12425 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
12426 copying pointers. The cert_st handling is changed by this in
12427 various ways (and thus what used to be known as ctx->default_cert
12428 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
12429 any longer when s->cert does not give us what we need).
12430 ssl_cert_instantiate becomes obsolete by this change.
12431 As soon as we've got the new code right (possibly it already is?),
12432 we have solved a couple of bugs of the earlier code where s->cert
12433 was used as if it could not have been shared with other SSL structures.
12434
12435 Note that using the SSL API in certain dirty ways now will result
12436 in different behaviour than observed with earlier library versions:
12437 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
12438 does not influence s as it used to.
12439
12440 In order to clean up things more thoroughly, inside SSL_SESSION
12441 we don't use CERT any longer, but a new structure SESS_CERT
12442 that holds per-session data (if available); currently, this is
12443 the peer's certificate chain and, for clients, the server's certificate
12444 and temporary key. CERT holds only those values that can have
12445 meaningful defaults in an SSL_CTX.
12446 [Bodo Moeller]
12447
12448 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
12449 from the internal representation. Various PKCS#7 fixes: remove some
12450 evil casts and set the enc_dig_alg field properly based on the signing
12451 key type.
12452 [Steve Henson]
12453
12454 *) Allow PKCS#12 password to be set from the command line or the
12455 environment. Let 'ca' get its config file name from the environment
12456 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
12457 and 'x509').
12458 [Steve Henson]
12459
12460 *) Allow certificate policies extension to use an IA5STRING for the
12461 organization field. This is contrary to the PKIX definition but
12462 VeriSign uses it and IE5 only recognises this form. Document 'x509'
12463 extension option.
12464 [Steve Henson]
12465
12466 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
12467 without disallowing inline assembler and the like for non-pedantic builds.
12468 [Ben Laurie]
12469
12470 *) Support Borland C++ builder.
12471 [Janez Jere <jj@void.si>, modified by Ulf Möller]
12472
12473 *) Support Mingw32.
12474 [Ulf Möller]
12475
12476 *) SHA-1 cleanups and performance enhancements.
12477 [Andy Polyakov <appro@fy.chalmers.se>]
12478
12479 *) Sparc v8plus assembler for the bignum library.
12480 [Andy Polyakov <appro@fy.chalmers.se>]
12481
12482 *) Accept any -xxx and +xxx compiler options in Configure.
12483 [Ulf Möller]
12484
12485 *) Update HPUX configuration.
12486 [Anonymous]
12487
12488 *) Add missing sk_<type>_unshift() function to safestack.h
12489 [Ralf S. Engelschall]
12490
12491 *) New function SSL_CTX_use_certificate_chain_file that sets the
12492 "extra_cert"s in addition to the certificate. (This makes sense
12493 only for "PEM" format files, as chains as a whole are not
12494 DER-encoded.)
12495 [Bodo Moeller]
12496
12497 *) Support verify_depth from the SSL API.
12498 x509_vfy.c had what can be considered an off-by-one-error:
12499 Its depth (which was not part of the external interface)
12500 was actually counting the number of certificates in a chain;
12501 now it really counts the depth.
12502 [Bodo Moeller]
12503
12504 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12505 instead of X509err, which often resulted in confusing error
12506 messages since the error codes are not globally unique
12507 (e.g. an alleged error in ssl3_accept when a certificate
12508 didn't match the private key).
12509
12510 *) New function SSL_CTX_set_session_id_context that allows to set a default
12511 value (so that you don't need SSL_set_session_id_context for each
12512 connection using the SSL_CTX).
12513 [Bodo Moeller]
12514
12515 *) OAEP decoding bug fix.
12516 [Ulf Möller]
12517
12518 *) Support INSTALL_PREFIX for package builders, as proposed by
12519 David Harris.
12520 [Bodo Moeller]
12521
12522 *) New Configure options "threads" and "no-threads". For systems
12523 where the proper compiler options are known (currently Solaris
12524 and Linux), "threads" is the default.
12525 [Bodo Moeller]
12526
12527 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12528 [Bodo Moeller]
12529
12530 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12531 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12532 such as /usr/local/bin.
12533 [Bodo Moeller]
12534
12535 *) "make linux-shared" to build shared libraries.
12536 [Niels Poppe <niels@netbox.org>]
12537
12538 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12539 [Ulf Möller]
12540
12541 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12542 extension adding in x509 utility.
12543 [Steve Henson]
12544
12545 *) Remove NOPROTO sections and error code comments.
12546 [Ulf Möller]
12547
12548 *) Partial rewrite of the DEF file generator to now parse the ANSI
12549 prototypes.
12550 [Steve Henson]
12551
12552 *) New Configure options --prefix=DIR and --openssldir=DIR.
12553 [Ulf Möller]
12554
12555 *) Complete rewrite of the error code script(s). It is all now handled
12556 by one script at the top level which handles error code gathering,
12557 header rewriting and C source file generation. It should be much better
12558 than the old method: it now uses a modified version of Ulf's parser to
12559 read the ANSI prototypes in all header files (thus the old K&R definitions
12560 aren't needed for error creation any more) and do a better job of
12561 translating function codes into names. The old 'ASN1 error code imbedded
12562 in a comment' is no longer necessary and it doesn't use .err files which
12563 have now been deleted. Also the error code call doesn't have to appear all
12564 on one line (which resulted in some large lines...).
12565 [Steve Henson]
12566
12567 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12568 [Bodo Moeller]
12569
12570 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12571 0 (which usually indicates a closed connection), but continue reading.
12572 [Bodo Moeller]
12573
12574 *) Fix some race conditions.
12575 [Bodo Moeller]
12576
12577 *) Add support for CRL distribution points extension. Add Certificate
12578 Policies and CRL distribution points documentation.
12579 [Steve Henson]
12580
12581 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12582 [Ulf Möller]
12583
12584 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12585 8 of keying material. Merlin has also confirmed interop with this fix
12586 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12587 [Merlin Hughes <merlin@baltimore.ie>]
12588
12589 *) Fix lots of warnings.
12590 [Richard Levitte <levitte@stacken.kth.se>]
12591
12592 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12593 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12594 [Richard Levitte <levitte@stacken.kth.se>]
12595
12596 *) Fix problems with sizeof(long) == 8.
12597 [Andy Polyakov <appro@fy.chalmers.se>]
12598
12599 *) Change functions to ANSI C.
12600 [Ulf Möller]
12601
12602 *) Fix typos in error codes.
12603 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12604
12605 *) Remove defunct assembler files from Configure.
12606 [Ulf Möller]
12607
12608 *) SPARC v8 assembler BIGNUM implementation.
12609 [Andy Polyakov <appro@fy.chalmers.se>]
12610
12611 *) Support for Certificate Policies extension: both print and set.
12612 Various additions to support the r2i method this uses.
12613 [Steve Henson]
12614
12615 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12616 return a const string when you are expecting an allocated buffer.
12617 [Ben Laurie]
12618
12619 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12620 types DirectoryString and DisplayText.
12621 [Steve Henson]
12622
12623 *) Add code to allow r2i extensions to access the configuration database,
12624 add an LHASH database driver and add several ctx helper functions.
12625 [Steve Henson]
12626
12627 *) Fix an evil bug in bn_expand2() which caused various BN functions to
12628 fail when they extended the size of a BIGNUM.
12629 [Steve Henson]
12630
12631 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
12632 support typesafe stack.
12633 [Steve Henson]
12634
12635 *) Fix typo in SSL_[gs]et_options().
12636 [Nils Frostberg <nils@medcom.se>]
12637
12638 *) Delete various functions and files that belonged to the (now obsolete)
12639 old X509V3 handling code.
12640 [Steve Henson]
12641
12642 *) New Configure option "rsaref".
12643 [Ulf Möller]
12644
12645 *) Don't auto-generate pem.h.
12646 [Bodo Moeller]
12647
12648 *) Introduce type-safe ASN.1 SETs.
12649 [Ben Laurie]
12650
12651 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
12652 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
12653
12654 *) Introduce type-safe STACKs. This will almost certainly break lots of code
12655 that links with OpenSSL (well at least cause lots of warnings), but fear
12656 not: the conversion is trivial, and it eliminates loads of evil casts. A
12657 few STACKed things have been converted already. Feel free to convert more.
12658 In the fullness of time, I'll do away with the STACK type altogether.
12659 [Ben Laurie]
12660
12661 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
12662 specified in <certfile> by updating the entry in the index.txt file.
12663 This way one no longer has to edit the index.txt file manually for
12664 revoking a certificate. The -revoke option does the gory details now.
12665 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
12666
12667 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
12668 `-text' option at all and this way the `-noout -text' combination was
12669 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
12670 [Ralf S. Engelschall]
12671
12672 *) Make sure a corresponding plain text error message exists for the
12673 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
12674 verify callback function determined that a certificate was revoked.
12675 [Ralf S. Engelschall]
12676
12677 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
12678 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
12679 all available ciphers including rc5, which was forgotten until now.
12680 In order to let the testing shell script know which algorithms
12681 are available, a new (up to now undocumented) command
12682 "openssl list-cipher-commands" is used.
12683 [Bodo Moeller]
12684
12685 *) Bugfix: s_client occasionally would sleep in select() when
12686 it should have checked SSL_pending() first.
12687 [Bodo Moeller]
12688
12689 *) New functions DSA_do_sign and DSA_do_verify to provide access to
12690 the raw DSA values prior to ASN.1 encoding.
12691 [Ulf Möller]
12692
12693 *) Tweaks to Configure
12694 [Niels Poppe <niels@netbox.org>]
12695
12696 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
12697 yet...
12698 [Steve Henson]
12699
12700 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
12701 [Ulf Möller]
12702
12703 *) New config option to avoid instructions that are illegal on the 80386.
12704 The default code is faster, but requires at least a 486.
12705 [Ulf Möller]
12706
12707 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
12708 SSL2_SERVER_VERSION (not used at all) macros, which are now the
12709 same as SSL2_VERSION anyway.
12710 [Bodo Moeller]
12711
12712 *) New "-showcerts" option for s_client.
12713 [Bodo Moeller]
12714
12715 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
12716 application. Various cleanups and fixes.
12717 [Steve Henson]
12718
12719 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
12720 modify error routines to work internally. Add error codes and PBE init
12721 to library startup routines.
12722 [Steve Henson]
12723
12724 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
12725 packing functions to asn1 and evp. Changed function names and error
12726 codes along the way.
12727 [Steve Henson]
12728
12729 *) PKCS12 integration: and so it begins... First of several patches to
12730 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
12731 objects to objects.h
12732 [Steve Henson]
12733
12734 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
12735 and display support for Thawte strong extranet extension.
12736 [Steve Henson]
12737
12738 *) Add LinuxPPC support.
12739 [Jeff Dubrule <igor@pobox.org>]
12740
12741 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
12742 bn_div_words in alpha.s.
12743 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
12744
12745 *) Make sure the RSA OAEP test is skipped under -DRSAref because
12746 OAEP isn't supported when OpenSSL is built with RSAref.
12747 [Ulf Moeller <ulf@fitug.de>]
12748
12749 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
12750 so they no longer are missing under -DNOPROTO.
12751 [Soren S. Jorvang <soren@t.dk>]
12752
12753
12754 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
12755
12756 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
12757 doesn't work when the session is reused. Coming soon!
12758 [Ben Laurie]
12759
12760 *) Fix a security hole, that allows sessions to be reused in the wrong
12761 context thus bypassing client cert protection! All software that uses
12762 client certs and session caches in multiple contexts NEEDS PATCHING to
12763 allow session reuse! A fuller solution is in the works.
12764 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
12765
12766 *) Some more source tree cleanups (removed obsolete files
12767 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
12768 permission on "config" script to be executable) and a fix for the INSTALL
12769 document.
12770 [Ulf Moeller <ulf@fitug.de>]
12771
12772 *) Remove some legacy and erroneous uses of malloc, free instead of
12773 Malloc, Free.
12774 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
12775
12776 *) Make rsa_oaep_test return non-zero on error.
12777 [Ulf Moeller <ulf@fitug.de>]
12778
12779 *) Add support for native Solaris shared libraries. Configure
12780 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12781 if someone would make that last step automatic.
12782 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12783
12784 *) ctx_size was not built with the right compiler during "make links". Fixed.
12785 [Ben Laurie]
12786
12787 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12788 except NULL ciphers". This means the default cipher list will no longer
12789 enable NULL ciphers. They need to be specifically enabled e.g. with
12790 the string "DEFAULT:eNULL".
12791 [Steve Henson]
12792
12793 *) Fix to RSA private encryption routines: if p < q then it would
12794 occasionally produce an invalid result. This will only happen with
12795 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12796 [Steve Henson]
12797
12798 *) Be less restrictive and allow also `perl util/perlpath.pl
12799 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12800 because this way one can also use an interpreter named `perl5' (which is
12801 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12802 installed as `perl').
12803 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12804
12805 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12806 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12807
12808 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12809 advapi32.lib to Win32 build and change the pem test comparison
12810 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12811 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
12812 and crypto/des/ede_cbcm_enc.c.
12813 [Steve Henson]
12814
12815 *) DES quad checksum was broken on big-endian architectures. Fixed.
12816 [Ben Laurie]
12817
12818 *) Comment out two functions in bio.h that aren't implemented. Fix up the
12819 Win32 test batch file so it (might) work again. The Win32 test batch file
12820 is horrible: I feel ill....
12821 [Steve Henson]
12822
12823 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
12824 in e_os.h. Audit of header files to check ANSI and non ANSI
12825 sections: 10 functions were absent from non ANSI section and not exported
12826 from Windows DLLs. Fixed up libeay.num for new functions.
12827 [Steve Henson]
12828
12829 *) Make `openssl version' output lines consistent.
12830 [Ralf S. Engelschall]
12831
12832 *) Fix Win32 symbol export lists for BIO functions: Added
12833 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
12834 to ms/libeay{16,32}.def.
12835 [Ralf S. Engelschall]
12836
12837 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
12838 fine under Unix and passes some trivial tests I've now added. But the
12839 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
12840 added to make sure no one expects that this stuff really works in the
12841 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
12842 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
12843 openssl_bio.xs.
12844 [Ralf S. Engelschall]
12845
12846 *) Fix the generation of two part addresses in perl.
12847 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
12848
12849 *) Add config entry for Linux on MIPS.
12850 [John Tobey <jtobey@channel1.com>]
12851
12852 *) Make links whenever Configure is run, unless we are on Windoze.
12853 [Ben Laurie]
12854
12855 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
12856 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
12857 in CRLs.
12858 [Steve Henson]
12859
12860 *) Add a useful kludge to allow package maintainers to specify compiler and
12861 other platforms details on the command line without having to patch the
12862 Configure script everytime: One now can use ``perl Configure
12863 <id>:<details>'', i.e. platform ids are allowed to have details appended
12864 to them (separated by colons). This is treated as there would be a static
12865 pre-configured entry in Configure's %table under key <id> with value
12866 <details> and ``perl Configure <id>'' is called. So, when you want to
12867 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
12868 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
12869 now, which overrides the FreeBSD-elf entry on-the-fly.
12870 [Ralf S. Engelschall]
12871
12872 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
12873 [Ben Laurie]
12874
12875 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
12876 on the `perl Configure ...' command line. This way one can compile
12877 OpenSSL libraries with Position Independent Code (PIC) which is needed
12878 for linking it into DSOs.
12879 [Ralf S. Engelschall]
12880
12881 *) Remarkably, export ciphers were totally broken and no-one had noticed!
12882 Fixed.
12883 [Ben Laurie]
12884
12885 *) Cleaned up the LICENSE document: The official contact for any license
12886 questions now is the OpenSSL core team under openssl-core@openssl.org.
12887 And add a paragraph about the dual-license situation to make sure people
12888 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
12889 to the OpenSSL toolkit.
12890 [Ralf S. Engelschall]
12891
12892 *) General source tree makefile cleanups: Made `making xxx in yyy...'
12893 display consistent in the source tree and replaced `/bin/rm' by `rm'.
12894 Additionally cleaned up the `make links' target: Remove unnecessary
12895 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
12896 to speed processing and no longer clutter the display with confusing
12897 stuff. Instead only the actually done links are displayed.
12898 [Ralf S. Engelschall]
12899
12900 *) Permit null encryption ciphersuites, used for authentication only. It used
12901 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
12902 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12903 encryption.
12904 [Ben Laurie]
12905
12906 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12907 signed attributes when verifying signatures (this would break them),
12908 the detached data encoding was wrong and public keys obtained using
12909 X509_get_pubkey() weren't freed.
12910 [Steve Henson]
12911
12912 *) Add text documentation for the BUFFER functions. Also added a work around
12913 to a Win95 console bug. This was triggered by the password read stuff: the
12914 last character typed gets carried over to the next fread(). If you were
12915 generating a new cert request using 'req' for example then the last
12916 character of the passphrase would be CR which would then enter the first
12917 field as blank.
12918 [Steve Henson]
12919
12920 *) Added the new `Includes OpenSSL Cryptography Software' button as
12921 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12922 button and can be used by applications based on OpenSSL to show the
12923 relationship to the OpenSSL project.
12924 [Ralf S. Engelschall]
12925
12926 *) Remove confusing variables in function signatures in files
12927 ssl/ssl_lib.c and ssl/ssl.h.
12928 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12929
12930 *) Don't install bss_file.c under PREFIX/include/
12931 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12932
12933 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12934 functions that return function pointers and has support for NT specific
12935 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12936 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12937 unsigned to signed types: this was killing the Win32 compile.
12938 [Steve Henson]
12939
12940 *) Add new certificate file to stack functions,
12941 SSL_add_dir_cert_subjects_to_stack() and
12942 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12943 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12944 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12945 This means that Apache-SSL and similar packages don't have to mess around
12946 to add as many CAs as they want to the preferred list.
12947 [Ben Laurie]
12948
12949 *) Experiment with doxygen documentation. Currently only partially applied to
12950 ssl/ssl_lib.c.
12951 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12952 openssl.doxy as the configuration file.
12953 [Ben Laurie]
12954
12955 *) Get rid of remaining C++-style comments which strict C compilers hate.
12956 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12957
12958 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12959 compiled in by default: it has problems with large keys.
12960 [Steve Henson]
12961
12962 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12963 DH private keys and/or callback functions which directly correspond to
12964 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12965 is needed for applications which have to configure certificates on a
12966 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12967 (e.g. s_server).
12968 For the RSA certificate situation is makes no difference, but
12969 for the DSA certificate situation this fixes the "no shared cipher"
12970 problem where the OpenSSL cipher selection procedure failed because the
12971 temporary keys were not overtaken from the context and the API provided
12972 no way to reconfigure them.
12973 The new functions now let applications reconfigure the stuff and they
12974 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12975 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12976 non-public-API function ssl_cert_instantiate() is used as a helper
12977 function and also to reduce code redundancy inside ssl_rsa.c.
12978 [Ralf S. Engelschall]
12979
12980 *) Move s_server -dcert and -dkey options out of the undocumented feature
12981 area because they are useful for the DSA situation and should be
12982 recognized by the users.
12983 [Ralf S. Engelschall]
12984
12985 *) Fix the cipher decision scheme for export ciphers: the export bits are
12986 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12987 SSL_EXP_MASK. So, the original variable has to be used instead of the
12988 already masked variable.
12989 [Richard Levitte <levitte@stacken.kth.se>]
12990
12991 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12992 [Richard Levitte <levitte@stacken.kth.se>]
12993
12994 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12995 from `int' to `unsigned int' because it's a length and initialized by
12996 EVP_DigestFinal() which expects an `unsigned int *'.
12997 [Richard Levitte <levitte@stacken.kth.se>]
12998
12999 *) Don't hard-code path to Perl interpreter on shebang line of Configure
13000 script. Instead use the usual Shell->Perl transition trick.
13001 [Ralf S. Engelschall]
13002
13003 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
13004 (in addition to RSA certificates) to match the behaviour of `openssl dsa
13005 -noout -modulus' as it's already the case for `openssl rsa -noout
13006 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
13007 currently the public key is printed (a decision which was already done by
13008 `openssl dsa -modulus' in the past) which serves a similar purpose.
13009 Additionally the NO_RSA no longer completely removes the whole -modulus
13010 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
13011 now, too.
13012 [Ralf S. Engelschall]
13013
13014 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
13015 BIO. See the source (crypto/evp/bio_ok.c) for more info.
13016 [Arne Ansper <arne@ats.cyber.ee>]
13017
13018 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
13019 to be added. Now both 'req' and 'ca' can use new objects defined in the
13020 config file.
13021 [Steve Henson]
13022
13023 *) Add cool BIO that does syslog (or event log on NT).
13024 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
13025
13026 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
13027 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
13028 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
13029 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
13030 [Ben Laurie]
13031
13032 *) Add preliminary config info for new extension code.
13033 [Steve Henson]
13034
13035 *) Make RSA_NO_PADDING really use no padding.
13036 [Ulf Moeller <ulf@fitug.de>]
13037
13038 *) Generate errors when private/public key check is done.
13039 [Ben Laurie]
13040
13041 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
13042 for some CRL extensions and new objects added.
13043 [Steve Henson]
13044
13045 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
13046 key usage extension and fuller support for authority key id.
13047 [Steve Henson]
13048
13049 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
13050 padding method for RSA, which is recommended for new applications in PKCS
13051 #1 v2.0 (RFC 2437, October 1998).
13052 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
13053 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
13054 against Bleichbacher's attack on RSA.
13055 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
13056 Ben Laurie]
13057
13058 *) Updates to the new SSL compression code
13059 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13060
13061 *) Fix so that the version number in the master secret, when passed
13062 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
13063 (because the server will not accept higher), that the version number
13064 is 0x03,0x01, not 0x03,0x00
13065 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13066
13067 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
13068 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
13069 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
13070 [Steve Henson]
13071
13072 *) Support for RAW extensions where an arbitrary extension can be
13073 created by including its DER encoding. See apps/openssl.cnf for
13074 an example.
13075 [Steve Henson]
13076
13077 *) Make sure latest Perl versions don't interpret some generated C array
13078 code as Perl array code in the crypto/err/err_genc.pl script.
13079 [Lars Weber <3weber@informatik.uni-hamburg.de>]
13080
13081 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
13082 not many people have the assembler. Various Win32 compilation fixes and
13083 update to the INSTALL.W32 file with (hopefully) more accurate Win32
13084 build instructions.
13085 [Steve Henson]
13086
13087 *) Modify configure script 'Configure' to automatically create crypto/date.h
13088 file under Win32 and also build pem.h from pem.org. New script
13089 util/mkfiles.pl to create the MINFO file on environments that can't do a
13090 'make files': perl util/mkfiles.pl >MINFO should work.
13091 [Steve Henson]
13092
13093 *) Major rework of DES function declarations, in the pursuit of correctness
13094 and purity. As a result, many evil casts evaporated, and some weirdness,
13095 too. You may find this causes warnings in your code. Zapping your evil
13096 casts will probably fix them. Mostly.
13097 [Ben Laurie]
13098
13099 *) Fix for a typo in asn1.h. Bug fix to object creation script
13100 obj_dat.pl. It considered a zero in an object definition to mean
13101 "end of object": none of the objects in objects.h have any zeros
13102 so it wasn't spotted.
13103 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
13104
13105 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
13106 Masking (CBCM). In the absence of test vectors, the best I have been able
13107 to do is check that the decrypt undoes the encrypt, so far. Send me test
13108 vectors if you have them.
13109 [Ben Laurie]
13110
13111 *) Correct calculation of key length for export ciphers (too much space was
13112 allocated for null ciphers). This has not been tested!
13113 [Ben Laurie]
13114
13115 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
13116 message is now correct (it understands "crypto" and "ssl" on its
13117 command line). There is also now an "update" option. This will update
13118 the util/ssleay.num and util/libeay.num files with any new functions.
13119 If you do a:
13120 perl util/mkdef.pl crypto ssl update
13121 it will update them.
13122 [Steve Henson]
13123
13124 *) Overhauled the Perl interface (perl/*):
13125 - ported BN stuff to OpenSSL's different BN library
13126 - made the perl/ source tree CVS-aware
13127 - renamed the package from SSLeay to OpenSSL (the files still contain
13128 their history because I've copied them in the repository)
13129 - removed obsolete files (the test scripts will be replaced
13130 by better Test::Harness variants in the future)
13131 [Ralf S. Engelschall]
13132
13133 *) First cut for a very conservative source tree cleanup:
13134 1. merge various obsolete readme texts into doc/ssleay.txt
13135 where we collect the old documents and readme texts.
13136 2. remove the first part of files where I'm already sure that we no
13137 longer need them because of three reasons: either they are just temporary
13138 files which were left by Eric or they are preserved original files where
13139 I've verified that the diff is also available in the CVS via "cvs diff
13140 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
13141 the crypto/md/ stuff).
13142 [Ralf S. Engelschall]
13143
13144 *) More extension code. Incomplete support for subject and issuer alt
13145 name, issuer and authority key id. Change the i2v function parameters
13146 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
13147 what that's for :-) Fix to ASN1 macro which messed up
13148 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
13149 [Steve Henson]
13150
13151 *) Preliminary support for ENUMERATED type. This is largely copied from the
13152 INTEGER code.
13153 [Steve Henson]
13154
13155 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
13156 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13157
13158 *) Make sure `make rehash' target really finds the `openssl' program.
13159 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
13160
13161 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
13162 like to hear about it if this slows down other processors.
13163 [Ben Laurie]
13164
13165 *) Add CygWin32 platform information to Configure script.
13166 [Alan Batie <batie@aahz.jf.intel.com>]
13167
13168 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
13169 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
13170
13171 *) New program nseq to manipulate netscape certificate sequences
13172 [Steve Henson]
13173
13174 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
13175 few typos.
13176 [Steve Henson]
13177
13178 *) Fixes to BN code. Previously the default was to define BN_RECURSION
13179 but the BN code had some problems that would cause failures when
13180 doing certificate verification and some other functions.
13181 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13182
13183 *) Add ASN1 and PEM code to support netscape certificate sequences.
13184 [Steve Henson]
13185
13186 *) Add ASN1 and PEM code to support netscape certificate sequences.
13187 [Steve Henson]
13188
13189 *) Add several PKIX and private extended key usage OIDs.
13190 [Steve Henson]
13191
13192 *) Modify the 'ca' program to handle the new extension code. Modify
13193 openssl.cnf for new extension format, add comments.
13194 [Steve Henson]
13195
13196 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
13197 and add a sample to openssl.cnf so req -x509 now adds appropriate
13198 CA extensions.
13199 [Steve Henson]
13200
13201 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
13202 error code, add initial support to X509_print() and x509 application.
13203 [Steve Henson]
13204
13205 *) Takes a deep breath and start adding X509 V3 extension support code. Add
13206 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
13207 stuff is currently isolated and isn't even compiled yet.
13208 [Steve Henson]
13209
13210 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
13211 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
13212 Removed the versions check from X509 routines when loading extensions:
13213 this allows certain broken certificates that don't set the version
13214 properly to be processed.
13215 [Steve Henson]
13216
13217 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
13218 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
13219 can still be regenerated with "make depend".
13220 [Ben Laurie]
13221
13222 *) Spelling mistake in C version of CAST-128.
13223 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
13224
13225 *) Changes to the error generation code. The perl script err-code.pl
13226 now reads in the old error codes and retains the old numbers, only
13227 adding new ones if necessary. It also only changes the .err files if new
13228 codes are added. The makefiles have been modified to only insert errors
13229 when needed (to avoid needlessly modifying header files). This is done
13230 by only inserting errors if the .err file is newer than the auto generated
13231 C file. To rebuild all the error codes from scratch (the old behaviour)
13232 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
13233 or delete all the .err files.
13234 [Steve Henson]
13235
13236 *) CAST-128 was incorrectly implemented for short keys. The C version has
13237 been fixed, but is untested. The assembler versions are also fixed, but
13238 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
13239 to regenerate it if needed.
13240 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
13241 Hagino <itojun@kame.net>]
13242
13243 *) File was opened incorrectly in randfile.c.
13244 [Ulf Möller <ulf@fitug.de>]
13245
13246 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
13247 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
13248 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
13249 al: it's just almost always a UTCTime. Note this patch adds new error
13250 codes so do a "make errors" if there are problems.
13251 [Steve Henson]
13252
13253 *) Correct Linux 1 recognition in config.
13254 [Ulf Möller <ulf@fitug.de>]
13255
13256 *) Remove pointless MD5 hash when using DSA keys in ca.
13257 [Anonymous <nobody@replay.com>]
13258
13259 *) Generate an error if given an empty string as a cert directory. Also
13260 generate an error if handed NULL (previously returned 0 to indicate an
13261 error, but didn't set one).
13262 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
13263
13264 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
13265 [Ben Laurie]
13266
13267 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
13268 parameters. This was causing a warning which killed off the Win32 compile.
13269 [Steve Henson]
13270
13271 *) Remove C++ style comments from crypto/bn/bn_local.h.
13272 [Neil Costigan <neil.costigan@celocom.com>]
13273
13274 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
13275 based on a text string, looking up short and long names and finally
13276 "dot" format. The "dot" format stuff didn't work. Added new function
13277 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
13278 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
13279 OID is not part of the table.
13280 [Steve Henson]
13281
13282 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
13283 X509_LOOKUP_by_alias().
13284 [Ben Laurie]
13285
13286 *) Sort openssl functions by name.
13287 [Ben Laurie]
13288
13289 *) Get the gendsa program working (hopefully) and add it to app list. Remove
13290 encryption from sample DSA keys (in case anyone is interested the password
13291 was "1234").
13292 [Steve Henson]
13293
13294 *) Make _all_ *_free functions accept a NULL pointer.
13295 [Frans Heymans <fheymans@isaserver.be>]
13296
13297 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
13298 NULL pointers.
13299 [Anonymous <nobody@replay.com>]
13300
13301 *) s_server should send the CAfile as acceptable CAs, not its own cert.
13302 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13303
13304 *) Don't blow it for numeric -newkey arguments to apps/req.
13305 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13306
13307 *) Temp key "for export" tests were wrong in s3_srvr.c.
13308 [Anonymous <nobody@replay.com>]
13309
13310 *) Add prototype for temp key callback functions
13311 SSL_CTX_set_tmp_{rsa,dh}_callback().
13312 [Ben Laurie]
13313
13314 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
13315 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
13316 [Steve Henson]
13317
13318 *) X509_name_add_entry() freed the wrong thing after an error.
13319 [Arne Ansper <arne@ats.cyber.ee>]
13320
13321 *) rsa_eay.c would attempt to free a NULL context.
13322 [Arne Ansper <arne@ats.cyber.ee>]
13323
13324 *) BIO_s_socket() had a broken should_retry() on Windoze.
13325 [Arne Ansper <arne@ats.cyber.ee>]
13326
13327 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
13328 [Arne Ansper <arne@ats.cyber.ee>]
13329
13330 *) Make sure the already existing X509_STORE->depth variable is initialized
13331 in X509_STORE_new(), but document the fact that this variable is still
13332 unused in the certificate verification process.
13333 [Ralf S. Engelschall]
13334
13335 *) Fix the various library and apps files to free up pkeys obtained from
13336 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
13337 [Steve Henson]
13338
13339 *) Fix reference counting in X509_PUBKEY_get(). This makes
13340 demos/maurice/example2.c work, amongst others, probably.
13341 [Steve Henson and Ben Laurie]
13342
13343 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
13344 `openssl' and second, the shortcut symlinks for the `openssl <command>'
13345 are no longer created. This way we have a single and consistent command
13346 line interface `openssl <command>', similar to `cvs <command>'.
13347 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
13348
13349 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
13350 BIT STRING wrapper always have zero unused bits.
13351 [Steve Henson]
13352
13353 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
13354 [Steve Henson]
13355
13356 *) Make the top-level INSTALL documentation easier to understand.
13357 [Paul Sutton]
13358
13359 *) Makefiles updated to exit if an error occurs in a sub-directory
13360 make (including if user presses ^C) [Paul Sutton]
13361
13362 *) Make Montgomery context stuff explicit in RSA data structure.
13363 [Ben Laurie]
13364
13365 *) Fix build order of pem and err to allow for generated pem.h.
13366 [Ben Laurie]
13367
13368 *) Fix renumbering bug in X509_NAME_delete_entry().
13369 [Ben Laurie]
13370
13371 *) Enhanced the err-ins.pl script so it makes the error library number
13372 global and can add a library name. This is needed for external ASN1 and
13373 other error libraries.
13374 [Steve Henson]
13375
13376 *) Fixed sk_insert which never worked properly.
13377 [Steve Henson]
13378
13379 *) Fix ASN1 macros so they can handle indefinite length constructed
13380 EXPLICIT tags. Some non standard certificates use these: they can now
13381 be read in.
13382 [Steve Henson]
13383
13384 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
13385 into a single doc/ssleay.txt bundle. This way the information is still
13386 preserved but no longer messes up this directory. Now it's new room for
13387 the new set of documentation files.
13388 [Ralf S. Engelschall]
13389
13390 *) SETs were incorrectly DER encoded. This was a major pain, because they
13391 shared code with SEQUENCEs, which aren't coded the same. This means that
13392 almost everything to do with SETs or SEQUENCEs has either changed name or
13393 number of arguments.
13394 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
13395
13396 *) Fix test data to work with the above.
13397 [Ben Laurie]
13398
13399 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
13400 was already fixed by Eric for 0.9.1 it seems.
13401 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
13402
13403 *) Autodetect FreeBSD3.
13404 [Ben Laurie]
13405
13406 *) Fix various bugs in Configure. This affects the following platforms:
13407 nextstep
13408 ncr-scde
13409 unixware-2.0
13410 unixware-2.0-pentium
13411 sco5-cc.
13412 [Ben Laurie]
13413
13414 *) Eliminate generated files from CVS. Reorder tests to regenerate files
13415 before they are needed.
13416 [Ben Laurie]
13417
13418 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
13419 [Ben Laurie]
13420
13421
13422 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
13423
13424 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
13425 changed SSLeay to OpenSSL in version strings.
13426 [Ralf S. Engelschall]
13427
13428 *) Some fixups to the top-level documents.
13429 [Paul Sutton]
13430
13431 *) Fixed the nasty bug where rsaref.h was not found under compile-time
13432 because the symlink to include/ was missing.
13433 [Ralf S. Engelschall]
13434
13435 *) Incorporated the popular no-RSA/DSA-only patches
13436 which allow to compile a RSA-free SSLeay.
13437 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
13438
13439 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
13440 when "ssleay" is still not found.
13441 [Ralf S. Engelschall]
13442
13443 *) Added more platforms to Configure: Cray T3E, HPUX 11,
13444 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
13445
13446 *) Updated the README file.
13447 [Ralf S. Engelschall]
13448
13449 *) Added various .cvsignore files in the CVS repository subdirs
13450 to make a "cvs update" really silent.
13451 [Ralf S. Engelschall]
13452
13453 *) Recompiled the error-definition header files and added
13454 missing symbols to the Win32 linker tables.
13455 [Ralf S. Engelschall]
13456
13457 *) Cleaned up the top-level documents;
13458 o new files: CHANGES and LICENSE
13459 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
13460 o merged COPYRIGHT into LICENSE
13461 o removed obsolete TODO file
13462 o renamed MICROSOFT to INSTALL.W32
13463 [Ralf S. Engelschall]
13464
13465 *) Removed dummy files from the 0.9.1b source tree:
13466 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
13467 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
13468 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
13469 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
13470 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
13471 [Ralf S. Engelschall]
13472
13473 *) Added various platform portability fixes.
13474 [Mark J. Cox]
13475
13476 *) The Genesis of the OpenSSL rpject:
13477 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
13478 Young and Tim J. Hudson created while they were working for C2Net until
13479 summer 1998.
13480 [The OpenSSL Project]
13481
13482
13483 Changes between 0.9.0b and 0.9.1b [not released]
13484
13485 *) Updated a few CA certificates under certs/
13486 [Eric A. Young]
13487
13488 *) Changed some BIGNUM api stuff.
13489 [Eric A. Young]
13490
13491 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13492 DGUX x86, Linux Alpha, etc.
13493 [Eric A. Young]
13494
13495 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13496 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13497 available).
13498 [Eric A. Young]
13499
13500 *) Add -strparse option to asn1pars program which parses nested
13501 binary structures
13502 [Dr Stephen Henson <shenson@bigfoot.com>]
13503
13504 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13505 [Eric A. Young]
13506
13507 *) DSA fix for "ca" program.
13508 [Eric A. Young]
13509
13510 *) Added "-genkey" option to "dsaparam" program.
13511 [Eric A. Young]
13512
13513 *) Added RIPE MD160 (rmd160) message digest.
13514 [Eric A. Young]
13515
13516 *) Added -a (all) option to "ssleay version" command.
13517 [Eric A. Young]
13518
13519 *) Added PLATFORM define which is the id given to Configure.
13520 [Eric A. Young]
13521
13522 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13523 [Eric A. Young]
13524
13525 *) Extended the ASN.1 parser routines.
13526 [Eric A. Young]
13527
13528 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13529 [Eric A. Young]
13530
13531 *) Added a BN_CTX to the BN library.
13532 [Eric A. Young]
13533
13534 *) Fixed the weak key values in DES library
13535 [Eric A. Young]
13536
13537 *) Changed API in EVP library for cipher aliases.
13538 [Eric A. Young]
13539
13540 *) Added support for RC2/64bit cipher.
13541 [Eric A. Young]
13542
13543 *) Converted the lhash library to the crypto/mem.c functions.
13544 [Eric A. Young]
13545
13546 *) Added more recognized ASN.1 object ids.
13547 [Eric A. Young]
13548
13549 *) Added more RSA padding checks for SSL/TLS.
13550 [Eric A. Young]
13551
13552 *) Added BIO proxy/filter functionality.
13553 [Eric A. Young]
13554
13555 *) Added extra_certs to SSL_CTX which can be used
13556 send extra CA certificates to the client in the CA cert chain sending
13557 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13558 [Eric A. Young]
13559
13560 *) Now Fortezza is denied in the authentication phase because
13561 this is key exchange mechanism is not supported by SSLeay at all.
13562 [Eric A. Young]
13563
13564 *) Additional PKCS1 checks.
13565 [Eric A. Young]
13566
13567 *) Support the string "TLSv1" for all TLS v1 ciphers.
13568 [Eric A. Young]
13569
13570 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13571 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13572 [Eric A. Young]
13573
13574 *) Fixed a few memory leaks.
13575 [Eric A. Young]
13576
13577 *) Fixed various code and comment typos.
13578 [Eric A. Young]
13579
13580 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13581 bytes sent in the client random.
13582 [Edward Bishop <ebishop@spyglass.com>]