]> git.ipfire.org Git - thirdparty/openssl.git/blob - CHANGES
Add CHANGES entry for blake2mac
[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 *) Add keyed BLAKE2 to EVP_MAC.
13 [Antoine Salon]
14
15 *) Fix a bug in the computation of the endpoint-pair shared secret used
16 by DTLS over SCTP. This breaks interoperability with older versions
17 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
18 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
19 interoperability with such broken implementations. However, enabling
20 this switch breaks interoperability with correct implementations.
21
22 *) Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
23 re-used X509_PUBKEY object if the second PUBKEY is malformed.
24 [Bernd Edlinger]
25
26 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
27 [Richard Levitte]
28
29 *) Change the license to the Apache License v2.0.
30 [Richard Levitte]
31
32 *) Change the possible version information given with OPENSSL_API_COMPAT.
33 It may be a pre-3.0.0 style numerical version number as it was defined
34 in 1.1.0, and it may also simply take the major version number.
35
36 Because of the version numbering of pre-3.0.0 releases, the values 0,
37 1 and 2 are equivalent to 0x00908000L (0.9.8), 0x10000000L (1.0.0) and
38 0x10100000L (1.1.0), respectively.
39 [Richard Levitte]
40
41 *) Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
42
43 o Major releases (indicated by incrementing the MAJOR release number)
44 may introduce incompatible API/ABI changes.
45 o Minor releases (indicated by incrementing the MINOR release number)
46 may introduce new features but retain API/ABI compatibility.
47 o Patch releases (indicated by incrementing the PATCH number)
48 are intended for bug fixes and other improvements of existing
49 features only (like improving performance or adding documentation)
50 and retain API/ABI compatibility.
51 [Richard Levitte]
52
53 *) Add support for RFC5297 SIV mode (siv128), including AES-SIV.
54 [Todd Short]
55
56 *) Remove the 'dist' target and add a tarball building script. The
57 'dist' target has fallen out of use, and it shouldn't be
58 necessary to configure just to create a source distribution.
59 [Richard Levitte]
60
61 *) Recreate the OS390-Unix config target. It no longer relies on a
62 special script like it did for OpenSSL pre-1.1.0.
63 [Richard Levitte]
64
65 *) Instead of having the source directories listed in Configure, add
66 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
67 look into.
68 [Richard Levitte]
69
70 *) Add GMAC to EVP_MAC.
71 [Paul Dale]
72
73 *) Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
74 [Richard Levitte]
75
76 *) Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
77 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
78 to facilitate the continued use of MACs through raw private keys in
79 functionality such as EVP_DigestSign* and EVP_DigestVerify*.
80 [Richard Levitte]
81
82 *) Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
83 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
84 [Antoine Salon]
85
86 *) Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
87 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
88 are retained for backwards compatibility.
89 [Antoine Salon]
90
91 *) AES-XTS mode now enforces that its two keys are different to mitigate
92 the attacked described in "Efficient Instantiations of Tweakable
93 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
94 Details of this attack can be obtained from:
95 http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf
96 [Paul Dale]
97
98 *) Rename the object files, i.e. give them other names than in previous
99 versions. Their names now include the name of the final product, as
100 well as its type mnemonic (bin, lib, shlib).
101 [Richard Levitte]
102
103 *) Added new option for 'openssl list', '-objects', which will display the
104 list of built in objects, i.e. OIDs with names.
105 [Richard Levitte]
106
107 *) Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
108 improves application performance by removing data copies and providing
109 applications with zero-copy system calls such as sendfile and splice.
110 [Boris Pismenny]
111
112 Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
113
114 *) Timing vulnerability in DSA signature generation
115
116 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
117 timing side channel attack. An attacker could use variations in the signing
118 algorithm to recover the private key.
119
120 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
121 (CVE-2018-0734)
122 [Paul Dale]
123
124 *) Timing vulnerability in ECDSA signature generation
125
126 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
127 timing side channel attack. An attacker could use variations in the signing
128 algorithm to recover the private key.
129
130 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
131 (CVE-2018-0735)
132 [Paul Dale]
133
134 *) Fixed the issue that RAND_add()/RAND_seed() silently discards random input
135 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
136 of two gigabytes and the error handling improved.
137
138 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
139 categorized as a normal bug, not a security issue, because the DRBG reseeds
140 automatically and is fully functional even without additional randomness
141 provided by the application.
142
143 Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
144
145 *) Add a new ClientHello callback. Provides a callback interface that gives
146 the application the ability to adjust the nascent SSL object at the
147 earliest stage of ClientHello processing, immediately after extensions have
148 been collected but before they have been processed. In particular, this
149 callback can adjust the supported TLS versions in response to the contents
150 of the ClientHello
151 [Benjamin Kaduk]
152
153 *) Add SM2 base algorithm support.
154 [Jack Lloyd]
155
156 *) s390x assembly pack: add (improved) hardware-support for the following
157 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
158 aes-cfb/cfb8, aes-ecb.
159 [Patrick Steuer]
160
161 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
162 parameter is no longer accepted, as it leads to a corrupt table. NULL
163 pem_str is reserved for alias entries only.
164 [Richard Levitte]
165
166 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
167 step for prime curves. The new implementation is based on formulae from
168 differential addition-and-doubling in homogeneous projective coordinates
169 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
170 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
171 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
172 to work in projective coordinates.
173 [Billy Bob Brumley, Nicola Tuveri]
174
175 *) Change generating and checking of primes so that the error rate of not
176 being prime depends on the intended use based on the size of the input.
177 For larger primes this will result in more rounds of Miller-Rabin.
178 The maximal error rate for primes with more than 1080 bits is lowered
179 to 2^-128.
180 [Kurt Roeckx, Annie Yousar]
181
182 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
183 [Kurt Roeckx]
184
185 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
186 moving between systems, and to avoid confusion when a Windows build is
187 done with mingw vs with MSVC. For POSIX installs, there's still a
188 symlink or copy named 'tsget' to avoid that confusion as well.
189 [Richard Levitte]
190
191 *) Revert blinding in ECDSA sign and instead make problematic addition
192 length-invariant. Switch even to fixed-length Montgomery multiplication.
193 [Andy Polyakov]
194
195 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
196 step for binary curves. The new implementation is based on formulae from
197 differential addition-and-doubling in mixed Lopez-Dahab projective
198 coordinates, modified to independently blind the operands.
199 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
200
201 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
202 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
203 EC_METHODs to implement their own specialized "ladder step", to take
204 advantage of more favorable coordinate systems or more efficient
205 differential addition-and-doubling algorithms.
206 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
207
208 *) Modified the random device based seed sources to keep the relevant
209 file descriptors open rather than reopening them on each access.
210 This allows such sources to operate in a chroot() jail without
211 the associated device nodes being available. This behaviour can be
212 controlled using RAND_keep_random_devices_open().
213 [Paul Dale]
214
215 *) Numerous side-channel attack mitigations have been applied. This may have
216 performance impacts for some algorithms for the benefit of improved
217 security. Specific changes are noted in this change log by their respective
218 authors.
219 [Matt Caswell]
220
221 *) AIX shared library support overhaul. Switch to AIX "natural" way of
222 handling shared libraries, which means collecting shared objects of
223 different versions and bitnesses in one common archive. This allows to
224 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
225 doesn't affect the way 3rd party applications are linked, only how
226 multi-version installation is managed.
227 [Andy Polyakov]
228
229 *) Make ec_group_do_inverse_ord() more robust and available to other
230 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
231 mitigations are applied to the fallback BN_mod_inverse().
232 When using this function rather than BN_mod_inverse() directly, new
233 EC cryptosystem implementations are then safer-by-default.
234 [Billy Bob Brumley]
235
236 *) Add coordinate blinding for EC_POINT and implement projective
237 coordinate blinding for generic prime curves as a countermeasure to
238 chosen point SCA attacks.
239 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
240
241 *) Add blinding to ECDSA and DSA signatures to protect against side channel
242 attacks discovered by Keegan Ryan (NCC Group).
243 [Matt Caswell]
244
245 *) Enforce checking in the pkeyutl command line app to ensure that the input
246 length does not exceed the maximum supported digest length when performing
247 a sign, verify or verifyrecover operation.
248 [Matt Caswell]
249
250 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
251 I/O in combination with something like select() or poll() will hang. This
252 can be turned off again using SSL_CTX_clear_mode().
253 Many applications do not properly handle non-application data records, and
254 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
255 around the problems in those applications, but can also break some.
256 It's recommended to read the manpages about SSL_read(), SSL_write(),
257 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
258 SSL_CTX_set_read_ahead() again.
259 [Kurt Roeckx]
260
261 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
262 now allow empty (zero character) pass phrases.
263 [Richard Levitte]
264
265 *) Apply blinding to binary field modular inversion and remove patent
266 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
267 [Billy Bob Brumley]
268
269 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
270 binary and prime elliptic curves.
271 [Billy Bob Brumley]
272
273 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
274 constant time fixed point multiplication.
275 [Billy Bob Brumley]
276
277 *) Revise elliptic curve scalar multiplication with timing attack
278 defenses: ec_wNAF_mul redirects to a constant time implementation
279 when computing fixed point and variable point multiplication (which
280 in OpenSSL are mostly used with secret scalars in keygen, sign,
281 ECDH derive operations).
282 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
283 Sohaib ul Hassan]
284
285 *) Updated CONTRIBUTING
286 [Rich Salz]
287
288 *) Updated DRBG / RAND to request nonce and additional low entropy
289 randomness from the system.
290 [Matthias St. Pierre]
291
292 *) Updated 'openssl rehash' to use OpenSSL consistent default.
293 [Richard Levitte]
294
295 *) Moved the load of the ssl_conf module to libcrypto, which helps
296 loading engines that libssl uses before libssl is initialised.
297 [Matt Caswell]
298
299 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
300 [Matt Caswell]
301
302 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
303 [Ingo Schwarze, Rich Salz]
304
305 *) Added output of accepting IP address and port for 'openssl s_server'
306 [Richard Levitte]
307
308 *) Added a new API for TLSv1.3 ciphersuites:
309 SSL_CTX_set_ciphersuites()
310 SSL_set_ciphersuites()
311 [Matt Caswell]
312
313 *) Memory allocation failures consistenly add an error to the error
314 stack.
315 [Rich Salz]
316
317 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
318 in libcrypto when run as setuid/setgid.
319 [Bernd Edlinger]
320
321 *) Load any config file by default when libssl is used.
322 [Matt Caswell]
323
324 *) Added new public header file <openssl/rand_drbg.h> and documentation
325 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
326 [Matthias St. Pierre]
327
328 *) QNX support removed (cannot find contributors to get their approval
329 for the license change).
330 [Rich Salz]
331
332 *) TLSv1.3 replay protection for early data has been implemented. See the
333 SSL_read_early_data() man page for further details.
334 [Matt Caswell]
335
336 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
337 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
338 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
339 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
340 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
341 configuration has been separated out. See the ciphers man page or the
342 SSL_CTX_set_ciphersuites() man page for more information.
343 [Matt Caswell]
344
345 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
346 in responder mode now supports the new "-multi" option, which
347 spawns the specified number of child processes to handle OCSP
348 requests. The "-timeout" option now also limits the OCSP
349 responder's patience to wait to receive the full client request
350 on a newly accepted connection. Child processes are respawned
351 as needed, and the CA index file is automatically reloaded
352 when changed. This makes it possible to run the "ocsp" responder
353 as a long-running service, making the OpenSSL CA somewhat more
354 feature-complete. In this mode, most diagnostic messages logged
355 after entering the event loop are logged via syslog(3) rather than
356 written to stderr.
357 [Viktor Dukhovni]
358
359 *) Added support for X448 and Ed448. Heavily based on original work by
360 Mike Hamburg.
361 [Matt Caswell]
362
363 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
364 objects loaded. This adds the functions OSSL_STORE_expect() and
365 OSSL_STORE_find() as well as needed tools to construct searches and
366 get the search data out of them.
367 [Richard Levitte]
368
369 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
370 version of OpenSSL should review their configuration settings to ensure
371 that they are still appropriate for TLSv1.3. For further information see:
372 https://wiki.openssl.org/index.php/TLS1.3
373 [Matt Caswell]
374
375 *) Grand redesign of the OpenSSL random generator
376
377 The default RAND method now utilizes an AES-CTR DRBG according to
378 NIST standard SP 800-90Ar1. The new random generator is essentially
379 a port of the default random generator from the OpenSSL FIPS 2.0
380 object module. It is a hybrid deterministic random bit generator
381 using an AES-CTR bit stream and which seeds and reseeds itself
382 automatically using trusted system entropy sources.
383
384 Some of its new features are:
385 o Support for multiple DRBG instances with seed chaining.
386 o The default RAND method makes use of a DRBG.
387 o There is a public and private DRBG instance.
388 o The DRBG instances are fork-safe.
389 o Keep all global DRBG instances on the secure heap if it is enabled.
390 o The public and private DRBG instance are per thread for lock free
391 operation
392 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
393
394 *) Changed Configure so it only says what it does and doesn't dump
395 so much data. Instead, ./configdata.pm should be used as a script
396 to display all sorts of configuration data.
397 [Richard Levitte]
398
399 *) Added processing of "make variables" to Configure.
400 [Richard Levitte]
401
402 *) Added SHA512/224 and SHA512/256 algorithm support.
403 [Paul Dale]
404
405 *) The last traces of Netware support, first removed in 1.1.0, have
406 now been removed.
407 [Rich Salz]
408
409 *) Get rid of Makefile.shared, and in the process, make the processing
410 of certain files (rc.obj, or the .def/.map/.opt files produced from
411 the ordinal files) more visible and hopefully easier to trace and
412 debug (or make silent).
413 [Richard Levitte]
414
415 *) Make it possible to have environment variable assignments as
416 arguments to config / Configure.
417 [Richard Levitte]
418
419 *) Add multi-prime RSA (RFC 8017) support.
420 [Paul Yang]
421
422 *) Add SM3 implemented according to GB/T 32905-2016
423 [ Jack Lloyd <jack.lloyd@ribose.com>,
424 Ronald Tse <ronald.tse@ribose.com>,
425 Erick Borsboom <erick.borsboom@ribose.com> ]
426
427 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
428 as documented in RFC6066.
429 Based on a patch from Tomasz Moń
430 [Filipe Raimundo da Silva]
431
432 *) Add SM4 implemented according to GB/T 32907-2016.
433 [ Jack Lloyd <jack.lloyd@ribose.com>,
434 Ronald Tse <ronald.tse@ribose.com>,
435 Erick Borsboom <erick.borsboom@ribose.com> ]
436
437 *) Reimplement -newreq-nodes and ERR_error_string_n; the
438 original author does not agree with the license change.
439 [Rich Salz]
440
441 *) Add ARIA AEAD TLS support.
442 [Jon Spillett]
443
444 *) Some macro definitions to support VS6 have been removed. Visual
445 Studio 6 has not worked since 1.1.0
446 [Rich Salz]
447
448 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
449 without clearing the errors.
450 [Richard Levitte]
451
452 *) Add "atfork" functions. If building on a system that without
453 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
454 requirements. The RAND facility now uses/requires this.
455 [Rich Salz]
456
457 *) Add SHA3.
458 [Andy Polyakov]
459
460 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
461 not possible to disable entirely. However, it's still possible to
462 disable the console reading UI method, UI_OpenSSL() (use UI_null()
463 as a fallback).
464
465 To disable, configure with 'no-ui-console'. 'no-ui' is still
466 possible to use as an alias. Check at compile time with the
467 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
468 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
469 [Richard Levitte]
470
471 *) Add a STORE module, which implements a uniform and URI based reader of
472 stores that can contain keys, certificates, CRLs and numerous other
473 objects. The main API is loosely based on a few stdio functions,
474 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
475 OSSL_STORE_error and OSSL_STORE_close.
476 The implementation uses backends called "loaders" to implement arbitrary
477 URI schemes. There is one built in "loader" for the 'file' scheme.
478 [Richard Levitte]
479
480 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
481 then adjusted to work on FreeBSD 8.4 as well.
482 Enable by configuring with 'enable-devcryptoeng'. This is done by default
483 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
484 [Richard Levitte]
485
486 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
487 util/mkerr.pl, which is adapted to allow those prefixes, leading to
488 error code calls like this:
489
490 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
491
492 With this change, we claim the namespaces OSSL and OPENSSL in a manner
493 that can be encoded in C. For the foreseeable future, this will only
494 affect new modules.
495 [Richard Levitte and Tim Hudson]
496
497 *) Removed BSD cryptodev engine.
498 [Rich Salz]
499
500 *) Add a build target 'build_all_generated', to build all generated files
501 and only that. This can be used to prepare everything that requires
502 things like perl for a system that lacks perl and then move everything
503 to that system and do the rest of the build there.
504 [Richard Levitte]
505
506 *) In the UI interface, make it possible to duplicate the user data. This
507 can be used by engines that need to retain the data for a longer time
508 than just the call where this user data is passed.
509 [Richard Levitte]
510
511 *) Ignore the '-named_curve auto' value for compatibility of applications
512 with OpenSSL 1.0.2.
513 [Tomas Mraz <tmraz@fedoraproject.org>]
514
515 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
516 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
517 alerts across multiple records (some of which could be empty). In practice
518 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
519 prohibts this altogether and other libraries (BoringSSL, NSS) do not
520 support this at all. Supporting it adds significant complexity to the
521 record layer, and its removal is unlikely to cause inter-operability
522 issues.
523 [Matt Caswell]
524
525 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
526 with Z. These are meant to replace LONG and ZLONG and to be size safe.
527 The use of LONG and ZLONG is discouraged and scheduled for deprecation
528 in OpenSSL 1.2.0.
529 [Richard Levitte]
530
531 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
532 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
533 [Richard Levitte, Andy Polyakov]
534
535 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
536 does for RSA, etc.
537 [Richard Levitte]
538
539 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
540 platform rather than 'mingw'.
541 [Richard Levitte]
542
543 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
544 success if they are asked to add an object which already exists
545 in the store. This change cascades to other functions which load
546 certificates and CRLs.
547 [Paul Dale]
548
549 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
550 facilitate stack unwinding even from assembly subroutines.
551 [Andy Polyakov]
552
553 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
554 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
555 [Richard Levitte]
556
557 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
558 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
559 which is the minimum version we support.
560 [Richard Levitte]
561
562 *) Certificate time validation (X509_cmp_time) enforces stricter
563 compliance with RFC 5280. Fractional seconds and timezone offsets
564 are no longer allowed.
565 [Emilia Käsper]
566
567 *) Add support for ARIA
568 [Paul Dale]
569
570 *) s_client will now send the Server Name Indication (SNI) extension by
571 default unless the new "-noservername" option is used. The server name is
572 based on the host provided to the "-connect" option unless overridden by
573 using "-servername".
574 [Matt Caswell]
575
576 *) Add support for SipHash
577 [Todd Short]
578
579 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
580 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
581 prevent issues where no progress is being made and the peer continually
582 sends unrecognised record types, using up resources processing them.
583 [Matt Caswell]
584
585 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
586 using the algorithm defined in
587 https://www.akkadia.org/drepper/SHA-crypt.txt
588 [Richard Levitte]
589
590 *) Heartbeat support has been removed; the ABI is changed for now.
591 [Richard Levitte, Rich Salz]
592
593 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
594 [Emilia Käsper]
595
596 *) The RSA "null" method, which was partially supported to avoid patent
597 issues, has been replaced to always returns NULL.
598 [Rich Salz]
599
600
601 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
602
603 *) Client DoS due to large DH parameter
604
605 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
606 malicious server can send a very large prime value to the client. This will
607 cause the client to spend an unreasonably long period of time generating a
608 key for this prime resulting in a hang until the client has finished. This
609 could be exploited in a Denial Of Service attack.
610
611 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
612 (CVE-2018-0732)
613 [Guido Vranken]
614
615 *) Cache timing vulnerability in RSA Key Generation
616
617 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
618 a cache timing side channel attack. An attacker with sufficient access to
619 mount cache timing attacks during the RSA key generation process could
620 recover the private key.
621
622 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
623 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
624 (CVE-2018-0737)
625 [Billy Brumley]
626
627 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
628 parameter is no longer accepted, as it leads to a corrupt table. NULL
629 pem_str is reserved for alias entries only.
630 [Richard Levitte]
631
632 *) Revert blinding in ECDSA sign and instead make problematic addition
633 length-invariant. Switch even to fixed-length Montgomery multiplication.
634 [Andy Polyakov]
635
636 *) Change generating and checking of primes so that the error rate of not
637 being prime depends on the intended use based on the size of the input.
638 For larger primes this will result in more rounds of Miller-Rabin.
639 The maximal error rate for primes with more than 1080 bits is lowered
640 to 2^-128.
641 [Kurt Roeckx, Annie Yousar]
642
643 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
644 [Kurt Roeckx]
645
646 *) Add blinding to ECDSA and DSA signatures to protect against side channel
647 attacks discovered by Keegan Ryan (NCC Group).
648 [Matt Caswell]
649
650 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
651 now allow empty (zero character) pass phrases.
652 [Richard Levitte]
653
654 *) Certificate time validation (X509_cmp_time) enforces stricter
655 compliance with RFC 5280. Fractional seconds and timezone offsets
656 are no longer allowed.
657 [Emilia Käsper]
658
659 *) Fixed a text canonicalisation bug in CMS
660
661 Where a CMS detached signature is used with text content the text goes
662 through a canonicalisation process first prior to signing or verifying a
663 signature. This process strips trailing space at the end of lines, converts
664 line terminators to CRLF and removes additional trailing line terminators
665 at the end of a file. A bug in the canonicalisation process meant that
666 some characters, such as form-feed, were incorrectly treated as whitespace
667 and removed. This is contrary to the specification (RFC5485). This fix
668 could mean that detached text data signed with an earlier version of
669 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
670 signed with a fixed OpenSSL may fail to verify with an earlier version of
671 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
672 and use the "-binary" flag (for the "cms" command line application) or set
673 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
674 [Matt Caswell]
675
676 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
677
678 *) Constructed ASN.1 types with a recursive definition could exceed the stack
679
680 Constructed ASN.1 types with a recursive definition (such as can be found
681 in PKCS7) could eventually exceed the stack given malicious input with
682 excessive recursion. This could result in a Denial Of Service attack. There
683 are no such structures used within SSL/TLS that come from untrusted sources
684 so this is considered safe.
685
686 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
687 project.
688 (CVE-2018-0739)
689 [Matt Caswell]
690
691 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
692
693 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
694 effectively reduced to only comparing the least significant bit of each
695 byte. This allows an attacker to forge messages that would be considered as
696 authenticated in an amount of tries lower than that guaranteed by the
697 security claims of the scheme. The module can only be compiled by the
698 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
699
700 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
701 (IBM).
702 (CVE-2018-0733)
703 [Andy Polyakov]
704
705 *) Add a build target 'build_all_generated', to build all generated files
706 and only that. This can be used to prepare everything that requires
707 things like perl for a system that lacks perl and then move everything
708 to that system and do the rest of the build there.
709 [Richard Levitte]
710
711 *) Backport SSL_OP_NO_RENGOTIATION
712
713 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
714 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
715 changes this is no longer possible in 1.1.0. Therefore the new
716 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
717 1.1.0 to provide equivalent functionality.
718
719 Note that if an application built against 1.1.0h headers (or above) is run
720 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
721 accepted but nothing will happen, i.e. renegotiation will not be prevented.
722 [Matt Caswell]
723
724 *) Removed the OS390-Unix config target. It relied on a script that doesn't
725 exist.
726 [Rich Salz]
727
728 *) rsaz_1024_mul_avx2 overflow bug on x86_64
729
730 There is an overflow bug in the AVX2 Montgomery multiplication procedure
731 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
732 Analysis suggests that attacks against RSA and DSA as a result of this
733 defect would be very difficult to perform and are not believed likely.
734 Attacks against DH1024 are considered just feasible, because most of the
735 work necessary to deduce information about a private key may be performed
736 offline. The amount of resources required for such an attack would be
737 significant. However, for an attack on TLS to be meaningful, the server
738 would have to share the DH1024 private key among multiple clients, which is
739 no longer an option since CVE-2016-0701.
740
741 This only affects processors that support the AVX2 but not ADX extensions
742 like Intel Haswell (4th generation).
743
744 This issue was reported to OpenSSL by David Benjamin (Google). The issue
745 was originally found via the OSS-Fuzz project.
746 (CVE-2017-3738)
747 [Andy Polyakov]
748
749 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
750
751 *) bn_sqrx8x_internal carry bug on x86_64
752
753 There is a carry propagating bug in the x86_64 Montgomery squaring
754 procedure. No EC algorithms are affected. Analysis suggests that attacks
755 against RSA and DSA as a result of this defect would be very difficult to
756 perform and are not believed likely. Attacks against DH are considered just
757 feasible (although very difficult) because most of the work necessary to
758 deduce information about a private key may be performed offline. The amount
759 of resources required for such an attack would be very significant and
760 likely only accessible to a limited number of attackers. An attacker would
761 additionally need online access to an unpatched system using the target
762 private key in a scenario with persistent DH parameters and a private
763 key that is shared between multiple clients.
764
765 This only affects processors that support the BMI1, BMI2 and ADX extensions
766 like Intel Broadwell (5th generation) and later or AMD Ryzen.
767
768 This issue was reported to OpenSSL by the OSS-Fuzz project.
769 (CVE-2017-3736)
770 [Andy Polyakov]
771
772 *) Malformed X.509 IPAddressFamily could cause OOB read
773
774 If an X.509 certificate has a malformed IPAddressFamily extension,
775 OpenSSL could do a one-byte buffer overread. The most likely result
776 would be an erroneous display of the certificate in text format.
777
778 This issue was reported to OpenSSL by the OSS-Fuzz project.
779 (CVE-2017-3735)
780 [Rich Salz]
781
782 Changes between 1.1.0e and 1.1.0f [25 May 2017]
783
784 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
785 platform rather than 'mingw'.
786 [Richard Levitte]
787
788 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
789 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
790 which is the minimum version we support.
791 [Richard Levitte]
792
793 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
794
795 *) Encrypt-Then-Mac renegotiation crash
796
797 During a renegotiation handshake if the Encrypt-Then-Mac extension is
798 negotiated where it was not in the original handshake (or vice-versa) then
799 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
800 and servers are affected.
801
802 This issue was reported to OpenSSL by Joe Orton (Red Hat).
803 (CVE-2017-3733)
804 [Matt Caswell]
805
806 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
807
808 *) Truncated packet could crash via OOB read
809
810 If one side of an SSL/TLS path is running on a 32-bit host and a specific
811 cipher is being used, then a truncated packet can cause that host to
812 perform an out-of-bounds read, usually resulting in a crash.
813
814 This issue was reported to OpenSSL by Robert Święcki of Google.
815 (CVE-2017-3731)
816 [Andy Polyakov]
817
818 *) Bad (EC)DHE parameters cause a client crash
819
820 If a malicious server supplies bad parameters for a DHE or ECDHE key
821 exchange then this can result in the client attempting to dereference a
822 NULL pointer leading to a client crash. This could be exploited in a Denial
823 of Service attack.
824
825 This issue was reported to OpenSSL by Guido Vranken.
826 (CVE-2017-3730)
827 [Matt Caswell]
828
829 *) BN_mod_exp may produce incorrect results on x86_64
830
831 There is a carry propagating bug in the x86_64 Montgomery squaring
832 procedure. No EC algorithms are affected. Analysis suggests that attacks
833 against RSA and DSA as a result of this defect would be very difficult to
834 perform and are not believed likely. Attacks against DH are considered just
835 feasible (although very difficult) because most of the work necessary to
836 deduce information about a private key may be performed offline. The amount
837 of resources required for such an attack would be very significant and
838 likely only accessible to a limited number of attackers. An attacker would
839 additionally need online access to an unpatched system using the target
840 private key in a scenario with persistent DH parameters and a private
841 key that is shared between multiple clients. For example this can occur by
842 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
843 similar to CVE-2015-3193 but must be treated as a separate problem.
844
845 This issue was reported to OpenSSL by the OSS-Fuzz project.
846 (CVE-2017-3732)
847 [Andy Polyakov]
848
849 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
850
851 *) ChaCha20/Poly1305 heap-buffer-overflow
852
853 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
854 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
855 crash. This issue is not considered to be exploitable beyond a DoS.
856
857 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
858 (CVE-2016-7054)
859 [Richard Levitte]
860
861 *) CMS Null dereference
862
863 Applications parsing invalid CMS structures can crash with a NULL pointer
864 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
865 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
866 structure callback if an attempt is made to free certain invalid encodings.
867 Only CHOICE structures using a callback which do not handle NULL value are
868 affected.
869
870 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
871 (CVE-2016-7053)
872 [Stephen Henson]
873
874 *) Montgomery multiplication may produce incorrect results
875
876 There is a carry propagating bug in the Broadwell-specific Montgomery
877 multiplication procedure that handles input lengths divisible by, but
878 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
879 and DH private keys are impossible. This is because the subroutine in
880 question is not used in operations with the private key itself and an input
881 of the attacker's direct choice. Otherwise the bug can manifest itself as
882 transient authentication and key negotiation failures or reproducible
883 erroneous outcome of public-key operations with specially crafted input.
884 Among EC algorithms only Brainpool P-512 curves are affected and one
885 presumably can attack ECDH key negotiation. Impact was not analyzed in
886 detail, because pre-requisites for attack are considered unlikely. Namely
887 multiple clients have to choose the curve in question and the server has to
888 share the private key among them, neither of which is default behaviour.
889 Even then only clients that chose the curve will be affected.
890
891 This issue was publicly reported as transient failures and was not
892 initially recognized as a security issue. Thanks to Richard Morgan for
893 providing reproducible case.
894 (CVE-2016-7055)
895 [Andy Polyakov]
896
897 *) Removed automatic addition of RPATH in shared libraries and executables,
898 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
899 [Richard Levitte]
900
901 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
902
903 *) Fix Use After Free for large message sizes
904
905 The patch applied to address CVE-2016-6307 resulted in an issue where if a
906 message larger than approx 16k is received then the underlying buffer to
907 store the incoming message is reallocated and moved. Unfortunately a
908 dangling pointer to the old location is left which results in an attempt to
909 write to the previously freed location. This is likely to result in a
910 crash, however it could potentially lead to execution of arbitrary code.
911
912 This issue only affects OpenSSL 1.1.0a.
913
914 This issue was reported to OpenSSL by Robert Święcki.
915 (CVE-2016-6309)
916 [Matt Caswell]
917
918 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
919
920 *) OCSP Status Request extension unbounded memory growth
921
922 A malicious client can send an excessively large OCSP Status Request
923 extension. If that client continually requests renegotiation, sending a
924 large OCSP Status Request extension each time, then there will be unbounded
925 memory growth on the server. This will eventually lead to a Denial Of
926 Service attack through memory exhaustion. Servers with a default
927 configuration are vulnerable even if they do not support OCSP. Builds using
928 the "no-ocsp" build time option are not affected.
929
930 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
931 (CVE-2016-6304)
932 [Matt Caswell]
933
934 *) SSL_peek() hang on empty record
935
936 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
937 sends an empty record. This could be exploited by a malicious peer in a
938 Denial Of Service attack.
939
940 This issue was reported to OpenSSL by Alex Gaynor.
941 (CVE-2016-6305)
942 [Matt Caswell]
943
944 *) Excessive allocation of memory in tls_get_message_header() and
945 dtls1_preprocess_fragment()
946
947 A (D)TLS message includes 3 bytes for its length in the header for the
948 message. This would allow for messages up to 16Mb in length. Messages of
949 this length are excessive and OpenSSL includes a check to ensure that a
950 peer is sending reasonably sized messages in order to avoid too much memory
951 being consumed to service a connection. A flaw in the logic of version
952 1.1.0 means that memory for the message is allocated too early, prior to
953 the excessive message length check. Due to way memory is allocated in
954 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
955 to service a connection. This could lead to a Denial of Service through
956 memory exhaustion. However, the excessive message length check still takes
957 place, and this would cause the connection to immediately fail. Assuming
958 that the application calls SSL_free() on the failed connection in a timely
959 manner then the 21Mb of allocated memory will then be immediately freed
960 again. Therefore the excessive memory allocation will be transitory in
961 nature. This then means that there is only a security impact if:
962
963 1) The application does not call SSL_free() in a timely manner in the event
964 that the connection fails
965 or
966 2) The application is working in a constrained environment where there is
967 very little free memory
968 or
969 3) The attacker initiates multiple connection attempts such that there are
970 multiple connections in a state where memory has been allocated for the
971 connection; SSL_free() has not yet been called; and there is insufficient
972 memory to service the multiple requests.
973
974 Except in the instance of (1) above any Denial Of Service is likely to be
975 transitory because as soon as the connection fails the memory is
976 subsequently freed again in the SSL_free() call. However there is an
977 increased risk during this period of application crashes due to the lack of
978 memory - which would then mean a more serious Denial of Service.
979
980 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
981 (CVE-2016-6307 and CVE-2016-6308)
982 [Matt Caswell]
983
984 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
985 had to be removed. Primary reason is that vendor assembler can't
986 assemble our modules with -KPIC flag. As result it, assembly
987 support, was not even available as option. But its lack means
988 lack of side-channel resistant code, which is incompatible with
989 security by todays standards. Fortunately gcc is readily available
990 prepackaged option, which we firmly point at...
991 [Andy Polyakov]
992
993 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
994
995 *) Windows command-line tool supports UTF-8 opt-in option for arguments
996 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
997 (to any value) allows Windows user to access PKCS#12 file generated
998 with Windows CryptoAPI and protected with non-ASCII password, as well
999 as files generated under UTF-8 locale on Linux also protected with
1000 non-ASCII password.
1001 [Andy Polyakov]
1002
1003 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
1004 have been disabled by default and removed from DEFAULT, just like RC4.
1005 See the RC4 item below to re-enable both.
1006 [Rich Salz]
1007
1008 *) The method for finding the storage location for the Windows RAND seed file
1009 has changed. First we check %RANDFILE%. If that is not set then we check
1010 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
1011 all else fails we fall back to C:\.
1012 [Matt Caswell]
1013
1014 *) The EVP_EncryptUpdate() function has had its return type changed from void
1015 to int. A return of 0 indicates and error while a return of 1 indicates
1016 success.
1017 [Matt Caswell]
1018
1019 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
1020 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
1021 off the constant time implementation for RSA, DSA and DH have been made
1022 no-ops and deprecated.
1023 [Matt Caswell]
1024
1025 *) Windows RAND implementation was simplified to only get entropy by
1026 calling CryptGenRandom(). Various other RAND-related tickets
1027 were also closed.
1028 [Joseph Wylie Yandle, Rich Salz]
1029
1030 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
1031 and OPENSSL_LH_, respectively. The old names are available
1032 with API compatibility. They new names are now completely documented.
1033 [Rich Salz]
1034
1035 *) Unify TYPE_up_ref(obj) methods signature.
1036 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
1037 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
1038 int (instead of void) like all others TYPE_up_ref() methods.
1039 So now these methods also check the return value of CRYPTO_atomic_add(),
1040 and the validity of object reference counter.
1041 [fdasilvayy@gmail.com]
1042
1043 *) With Windows Visual Studio builds, the .pdb files are installed
1044 alongside the installed libraries and executables. For a static
1045 library installation, ossl_static.pdb is the associate compiler
1046 generated .pdb file to be used when linking programs.
1047 [Richard Levitte]
1048
1049 *) Remove openssl.spec. Packaging files belong with the packagers.
1050 [Richard Levitte]
1051
1052 *) Automatic Darwin/OSX configuration has had a refresh, it will now
1053 recognise x86_64 architectures automatically. You can still decide
1054 to build for a different bitness with the environment variable
1055 KERNEL_BITS (can be 32 or 64), for example:
1056
1057 KERNEL_BITS=32 ./config
1058
1059 [Richard Levitte]
1060
1061 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
1062 256 bit AES and HMAC with SHA256.
1063 [Steve Henson]
1064
1065 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
1066 [Andy Polyakov]
1067
1068 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
1069 [Rich Salz]
1070
1071 *) To enable users to have their own config files and build file templates,
1072 Configure looks in the directory indicated by the environment variable
1073 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
1074 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
1075 name and is used as is.
1076 [Richard Levitte]
1077
1078 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
1079 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
1080 X509_CERT_FILE_CTX was removed.
1081 [Rich Salz]
1082
1083 *) "shared" builds are now the default. To create only static libraries use
1084 the "no-shared" Configure option.
1085 [Matt Caswell]
1086
1087 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
1088 All of these option have not worked for some while and are fundamental
1089 algorithms.
1090 [Matt Caswell]
1091
1092 *) Make various cleanup routines no-ops and mark them as deprecated. Most
1093 global cleanup functions are no longer required because they are handled
1094 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
1095 Explicitly de-initing can cause problems (e.g. where a library that uses
1096 OpenSSL de-inits, but an application is still using it). The affected
1097 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
1098 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
1099 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
1100 COMP_zlib_cleanup().
1101 [Matt Caswell]
1102
1103 *) --strict-warnings no longer enables runtime debugging options
1104 such as REF_DEBUG. Instead, debug options are automatically
1105 enabled with '--debug' builds.
1106 [Andy Polyakov, Emilia Käsper]
1107
1108 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
1109 have been moved out of the public header files. New functions for managing
1110 these have been added.
1111 [Matt Caswell]
1112
1113 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
1114 objects have been moved out of the public header files. New
1115 functions for managing these have been added.
1116 [Richard Levitte]
1117
1118 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1119 have been moved out of the public header files. New functions for managing
1120 these have been added.
1121 [Matt Caswell]
1122
1123 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1124 moved out of the public header files. New functions for managing these
1125 have been added.
1126 [Matt Caswell]
1127
1128 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
1129 [Matt Caswell]
1130
1131 *) Removed the mk1mf build scripts.
1132 [Richard Levitte]
1133
1134 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1135 it is always safe to #include a header now.
1136 [Rich Salz]
1137
1138 *) Removed the aged BC-32 config and all its supporting scripts
1139 [Richard Levitte]
1140
1141 *) Removed support for Ultrix, Netware, and OS/2.
1142 [Rich Salz]
1143
1144 *) Add support for HKDF.
1145 [Alessandro Ghedini]
1146
1147 *) Add support for blake2b and blake2s
1148 [Bill Cox]
1149
1150 *) Added support for "pipelining". Ciphers that have the
1151 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1152 encryptions/decryptions simultaneously. There are currently no built-in
1153 ciphers with this property but the expectation is that engines will be able
1154 to offer it to significantly improve throughput. Support has been extended
1155 into libssl so that multiple records for a single connection can be
1156 processed in one go (for >=TLS 1.1).
1157 [Matt Caswell]
1158
1159 *) Added the AFALG engine. This is an async capable engine which is able to
1160 offload work to the Linux kernel. In this initial version it only supports
1161 AES128-CBC. The kernel must be version 4.1.0 or greater.
1162 [Catriona Lucey]
1163
1164 *) OpenSSL now uses a new threading API. It is no longer necessary to
1165 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1166 are two supported threading models: pthreads and windows threads. It is
1167 also possible to configure OpenSSL at compile time for "no-threads". The
1168 old threading API should no longer be used. The functions have been
1169 replaced with "no-op" compatibility macros.
1170 [Alessandro Ghedini, Matt Caswell]
1171
1172 *) Modify behavior of ALPN to invoke callback after SNI/servername
1173 callback, such that updates to the SSL_CTX affect ALPN.
1174 [Todd Short]
1175
1176 *) Add SSL_CIPHER queries for authentication and key-exchange.
1177 [Todd Short]
1178
1179 *) Changes to the DEFAULT cipherlist:
1180 - Prefer (EC)DHE handshakes over plain RSA.
1181 - Prefer AEAD ciphers over legacy ciphers.
1182 - Prefer ECDSA over RSA when both certificates are available.
1183 - Prefer TLSv1.2 ciphers/PRF.
1184 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1185 default cipherlist.
1186 [Emilia Käsper]
1187
1188 *) Change the ECC default curve list to be this, in order: x25519,
1189 secp256r1, secp521r1, secp384r1.
1190 [Rich Salz]
1191
1192 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1193 disabled by default. They can be re-enabled using the
1194 enable-weak-ssl-ciphers option to Configure.
1195 [Matt Caswell]
1196
1197 *) If the server has ALPN configured, but supports no protocols that the
1198 client advertises, send a fatal "no_application_protocol" alert.
1199 This behaviour is SHALL in RFC 7301, though it isn't universally
1200 implemented by other servers.
1201 [Emilia Käsper]
1202
1203 *) Add X25519 support.
1204 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1205 for public and private key encoding using the format documented in
1206 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1207 key generation and key derivation.
1208
1209 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1210 X25519(29).
1211 [Steve Henson]
1212
1213 *) Deprecate SRP_VBASE_get_by_user.
1214 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1215 In order to fix an unavoidable memory leak (CVE-2016-0798),
1216 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1217 seed, even if the seed is configured.
1218
1219 Users should use SRP_VBASE_get1_by_user instead. Note that in
1220 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1221 also that even though configuring the SRP seed attempts to hide
1222 invalid usernames by continuing the handshake with fake
1223 credentials, this behaviour is not constant time and no strong
1224 guarantees are made that the handshake is indistinguishable from
1225 that of a valid user.
1226 [Emilia Käsper]
1227
1228 *) Configuration change; it's now possible to build dynamic engines
1229 without having to build shared libraries and vice versa. This
1230 only applies to the engines in engines/, those in crypto/engine/
1231 will always be built into libcrypto (i.e. "static").
1232
1233 Building dynamic engines is enabled by default; to disable, use
1234 the configuration option "disable-dynamic-engine".
1235
1236 The only requirements for building dynamic engines are the
1237 presence of the DSO module and building with position independent
1238 code, so they will also automatically be disabled if configuring
1239 with "disable-dso" or "disable-pic".
1240
1241 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1242 are also taken away from openssl/opensslconf.h, as they are
1243 irrelevant.
1244 [Richard Levitte]
1245
1246 *) Configuration change; if there is a known flag to compile
1247 position independent code, it will always be applied on the
1248 libcrypto and libssl object files, and never on the application
1249 object files. This means other libraries that use routines from
1250 libcrypto / libssl can be made into shared libraries regardless
1251 of how OpenSSL was configured.
1252
1253 If this isn't desirable, the configuration options "disable-pic"
1254 or "no-pic" can be used to disable the use of PIC. This will
1255 also disable building shared libraries and dynamic engines.
1256 [Richard Levitte]
1257
1258 *) Removed JPAKE code. It was experimental and has no wide use.
1259 [Rich Salz]
1260
1261 *) The INSTALL_PREFIX Makefile variable has been renamed to
1262 DESTDIR. That makes for less confusion on what this variable
1263 is for. Also, the configuration option --install_prefix is
1264 removed.
1265 [Richard Levitte]
1266
1267 *) Heartbeat for TLS has been removed and is disabled by default
1268 for DTLS; configure with enable-heartbeats. Code that uses the
1269 old #define's might need to be updated.
1270 [Emilia Käsper, Rich Salz]
1271
1272 *) Rename REF_CHECK to REF_DEBUG.
1273 [Rich Salz]
1274
1275 *) New "unified" build system
1276
1277 The "unified" build system is aimed to be a common system for all
1278 platforms we support. With it comes new support for VMS.
1279
1280 This system builds supports building in a different directory tree
1281 than the source tree. It produces one Makefile (for unix family
1282 or lookalikes), or one descrip.mms (for VMS).
1283
1284 The source of information to make the Makefile / descrip.mms is
1285 small files called 'build.info', holding the necessary
1286 information for each directory with source to compile, and a
1287 template in Configurations, like unix-Makefile.tmpl or
1288 descrip.mms.tmpl.
1289
1290 With this change, the library names were also renamed on Windows
1291 and on VMS. They now have names that are closer to the standard
1292 on Unix, and include the major version number, and in certain
1293 cases, the architecture they are built for. See "Notes on shared
1294 libraries" in INSTALL.
1295
1296 We rely heavily on the perl module Text::Template.
1297 [Richard Levitte]
1298
1299 *) Added support for auto-initialisation and de-initialisation of the library.
1300 OpenSSL no longer requires explicit init or deinit routines to be called,
1301 except in certain circumstances. See the OPENSSL_init_crypto() and
1302 OPENSSL_init_ssl() man pages for further information.
1303 [Matt Caswell]
1304
1305 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1306 "peer" argument is now expected to be a BIO_ADDR object.
1307
1308 *) Rewrite of BIO networking library. The BIO library lacked consistent
1309 support of IPv6, and adding it required some more extensive
1310 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1311 which hold all types of addresses and chains of address information.
1312 It also introduces a new API, with functions like BIO_socket,
1313 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1314 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1315 have been adapted accordingly.
1316 [Richard Levitte]
1317
1318 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1319 the leading 0-byte.
1320 [Emilia Käsper]
1321
1322 *) CRIME protection: disable compression by default, even if OpenSSL is
1323 compiled with zlib enabled. Applications can still enable compression
1324 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1325 using the SSL_CONF library to configure compression.
1326 [Emilia Käsper]
1327
1328 *) The signature of the session callback configured with
1329 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1330 was explicitly marked as 'const unsigned char*' instead of
1331 'unsigned char*'.
1332 [Emilia Käsper]
1333
1334 *) Always DPURIFY. Remove the use of uninitialized memory in the
1335 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1336 [Emilia Käsper]
1337
1338 *) Removed many obsolete configuration items, including
1339 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1340 MD2_CHAR, MD2_INT, MD2_LONG
1341 BF_PTR, BF_PTR2
1342 IDEA_SHORT, IDEA_LONG
1343 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1344 [Rich Salz, with advice from Andy Polyakov]
1345
1346 *) Many BN internals have been moved to an internal header file.
1347 [Rich Salz with help from Andy Polyakov]
1348
1349 *) Configuration and writing out the results from it has changed.
1350 Files such as Makefile include/openssl/opensslconf.h and are now
1351 produced through general templates, such as Makefile.in and
1352 crypto/opensslconf.h.in and some help from the perl module
1353 Text::Template.
1354
1355 Also, the center of configuration information is no longer
1356 Makefile. Instead, Configure produces a perl module in
1357 configdata.pm which holds most of the config data (in the hash
1358 table %config), the target data that comes from the target
1359 configuration in one of the Configurations/*.conf files (in
1360 %target).
1361 [Richard Levitte]
1362
1363 *) To clarify their intended purposes, the Configure options
1364 --prefix and --openssldir change their semantics, and become more
1365 straightforward and less interdependent.
1366
1367 --prefix shall be used exclusively to give the location INSTALLTOP
1368 where programs, scripts, libraries, include files and manuals are
1369 going to be installed. The default is now /usr/local.
1370
1371 --openssldir shall be used exclusively to give the default
1372 location OPENSSLDIR where certificates, private keys, CRLs are
1373 managed. This is also where the default openssl.cnf gets
1374 installed.
1375 If the directory given with this option is a relative path, the
1376 values of both the --prefix value and the --openssldir value will
1377 be combined to become OPENSSLDIR.
1378 The default for --openssldir is INSTALLTOP/ssl.
1379
1380 Anyone who uses --openssldir to specify where OpenSSL is to be
1381 installed MUST change to use --prefix instead.
1382 [Richard Levitte]
1383
1384 *) The GOST engine was out of date and therefore it has been removed. An up
1385 to date GOST engine is now being maintained in an external repository.
1386 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1387 support for GOST ciphersuites (these are only activated if a GOST engine
1388 is present).
1389 [Matt Caswell]
1390
1391 *) EGD is no longer supported by default; use enable-egd when
1392 configuring.
1393 [Ben Kaduk and Rich Salz]
1394
1395 *) The distribution now has Makefile.in files, which are used to
1396 create Makefile's when Configure is run. *Configure must be run
1397 before trying to build now.*
1398 [Rich Salz]
1399
1400 *) The return value for SSL_CIPHER_description() for error conditions
1401 has changed.
1402 [Rich Salz]
1403
1404 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1405
1406 Obtaining and performing DNSSEC validation of TLSA records is
1407 the application's responsibility. The application provides
1408 the TLSA records of its choice to OpenSSL, and these are then
1409 used to authenticate the peer.
1410
1411 The TLSA records need not even come from DNS. They can, for
1412 example, be used to implement local end-entity certificate or
1413 trust-anchor "pinning", where the "pin" data takes the form
1414 of TLSA records, which can augment or replace verification
1415 based on the usual WebPKI public certification authorities.
1416 [Viktor Dukhovni]
1417
1418 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1419 continues to support deprecated interfaces in default builds.
1420 However, applications are strongly advised to compile their
1421 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1422 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1423 or the 1.1.0 releases.
1424
1425 In environments in which all applications have been ported to
1426 not use any deprecated interfaces OpenSSL's Configure script
1427 should be used with the --api=1.1.0 option to entirely remove
1428 support for the deprecated features from the library and
1429 unconditionally disable them in the installed headers.
1430 Essentially the same effect can be achieved with the "no-deprecated"
1431 argument to Configure, except that this will always restrict
1432 the build to just the latest API, rather than a fixed API
1433 version.
1434
1435 As applications are ported to future revisions of the API,
1436 they should update their compile-time OPENSSL_API_COMPAT define
1437 accordingly, but in most cases should be able to continue to
1438 compile with later releases.
1439
1440 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1441 0x10000000L and 0x00908000L, respectively. However those
1442 versions did not support the OPENSSL_API_COMPAT feature, and
1443 so applications are not typically tested for explicit support
1444 of just the undeprecated features of either release.
1445 [Viktor Dukhovni]
1446
1447 *) Add support for setting the minimum and maximum supported protocol.
1448 It can bet set via the SSL_set_min_proto_version() and
1449 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1450 MaxProtocol. It's recommended to use the new APIs to disable
1451 protocols instead of disabling individual protocols using
1452 SSL_set_options() or SSL_CONF's Protocol. This change also
1453 removes support for disabling TLS 1.2 in the OpenSSL TLS
1454 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1455 [Kurt Roeckx]
1456
1457 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1458 [Andy Polyakov]
1459
1460 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1461 and integrates ECDSA and ECDH functionality into EC. Implementations can
1462 now redirect key generation and no longer need to convert to or from
1463 ECDSA_SIG format.
1464
1465 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1466 include the ec.h header file instead.
1467 [Steve Henson]
1468
1469 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1470 ciphers who are no longer supported and drops support the ephemeral RSA key
1471 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1472 [Kurt Roeckx]
1473
1474 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1475 opaque. For HMAC_CTX, the following constructors and destructors
1476 were added:
1477
1478 HMAC_CTX *HMAC_CTX_new(void);
1479 void HMAC_CTX_free(HMAC_CTX *ctx);
1480
1481 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1482 destroy such methods has been added. See EVP_MD_meth_new(3) and
1483 EVP_CIPHER_meth_new(3) for documentation.
1484
1485 Additional changes:
1486 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1487 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1488 EVP_MD_CTX_reset() should be called instead to reinitialise
1489 an already created structure.
1490 2) For consistency with the majority of our object creators and
1491 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1492 EVP_MD_CTX_(new|free). The old names are retained as macros
1493 for deprecated builds.
1494 [Richard Levitte]
1495
1496 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1497 cryptographic operations to be performed asynchronously as long as an
1498 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1499 further details. Libssl has also had this capability integrated with the
1500 introduction of the new mode SSL_MODE_ASYNC and associated error
1501 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1502 pages. This work was developed in partnership with Intel Corp.
1503 [Matt Caswell]
1504
1505 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1506 always enabled now. If you want to disable the support you should
1507 exclude it using the list of supported ciphers. This also means that the
1508 "-no_ecdhe" option has been removed from s_server.
1509 [Kurt Roeckx]
1510
1511 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1512 SSL_{CTX_}set1_curves() which can set a list.
1513 [Kurt Roeckx]
1514
1515 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1516 curve you want to support using SSL_{CTX_}set1_curves().
1517 [Kurt Roeckx]
1518
1519 *) State machine rewrite. The state machine code has been significantly
1520 refactored in order to remove much duplication of code and solve issues
1521 with the old code (see ssl/statem/README for further details). This change
1522 does have some associated API changes. Notably the SSL_state() function
1523 has been removed and replaced by SSL_get_state which now returns an
1524 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1525 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1526 also been removed.
1527 [Matt Caswell]
1528
1529 *) All instances of the string "ssleay" in the public API were replaced
1530 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1531 Some error codes related to internal RSA_eay API's were renamed.
1532 [Rich Salz]
1533
1534 *) The demo files in crypto/threads were moved to demo/threads.
1535 [Rich Salz]
1536
1537 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1538 sureware and ubsec.
1539 [Matt Caswell, Rich Salz]
1540
1541 *) New ASN.1 embed macro.
1542
1543 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1544 structure is not allocated: it is part of the parent. That is instead of
1545
1546 FOO *x;
1547
1548 it must be:
1549
1550 FOO x;
1551
1552 This reduces memory fragmentation and make it impossible to accidentally
1553 set a mandatory field to NULL.
1554
1555 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1556 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1557 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1558 SEQUENCE OF.
1559 [Steve Henson]
1560
1561 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1562 [Emilia Käsper]
1563
1564 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1565 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1566 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1567 DES and RC4 ciphersuites.
1568 [Matt Caswell]
1569
1570 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1571 This changes the decoding behaviour for some invalid messages,
1572 though the change is mostly in the more lenient direction, and
1573 legacy behaviour is preserved as much as possible.
1574 [Emilia Käsper]
1575
1576 *) Fix no-stdio build.
1577 [ David Woodhouse <David.Woodhouse@intel.com> and also
1578 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1579
1580 *) New testing framework
1581 The testing framework has been largely rewritten and is now using
1582 perl and the perl modules Test::Harness and an extended variant of
1583 Test::More called OpenSSL::Test to do its work. All test scripts in
1584 test/ have been rewritten into test recipes, and all direct calls to
1585 executables in test/Makefile have become individual recipes using the
1586 simplified testing OpenSSL::Test::Simple.
1587
1588 For documentation on our testing modules, do:
1589
1590 perldoc test/testlib/OpenSSL/Test/Simple.pm
1591 perldoc test/testlib/OpenSSL/Test.pm
1592
1593 [Richard Levitte]
1594
1595 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1596 are used; the latter aborts on memory leaks (usually checked on exit).
1597 Some undocumented "set malloc, etc., hooks" functions were removed
1598 and others were changed. All are now documented.
1599 [Rich Salz]
1600
1601 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1602 return an error
1603 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1604
1605 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1606 from RFC4279, RFC4785, RFC5487, RFC5489.
1607
1608 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1609 original RSA_PSK patch.
1610 [Steve Henson]
1611
1612 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1613 era flag was never set throughout the codebase (only read). Also removed
1614 SSL3_FLAGS_POP_BUFFER which was only used if
1615 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1616 [Matt Caswell]
1617
1618 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1619 to be "oneline" instead of "compat".
1620 [Richard Levitte]
1621
1622 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1623 not aware of clients that still exhibit this bug, and the workaround
1624 hasn't been working properly for a while.
1625 [Emilia Käsper]
1626
1627 *) The return type of BIO_number_read() and BIO_number_written() as well as
1628 the corresponding num_read and num_write members in the BIO structure has
1629 changed from unsigned long to uint64_t. On platforms where an unsigned
1630 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1631 transferred.
1632 [Matt Caswell]
1633
1634 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1635 OpenSSL without support for them. It also means that maintaining
1636 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1637 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1638 [Matt Caswell]
1639
1640 *) Removed support for the two export grade static DH ciphersuites
1641 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1642 were newly added (along with a number of other static DH ciphersuites) to
1643 1.0.2. However the two export ones have *never* worked since they were
1644 introduced. It seems strange in any case to be adding new export
1645 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1646 [Matt Caswell]
1647
1648 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1649 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1650 and turned into macros which simply call the new preferred function names
1651 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1652 should use the new names instead. Also as part of this change the ssl23.h
1653 header file has been removed.
1654 [Matt Caswell]
1655
1656 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1657 code and the associated standard is no longer considered fit-for-purpose.
1658 [Matt Caswell]
1659
1660 *) RT2547 was closed. When generating a private key, try to make the
1661 output file readable only by the owner. This behavior change might
1662 be noticeable when interacting with other software.
1663
1664 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1665 Added a test.
1666 [Rich Salz]
1667
1668 *) Added HTTP GET support to the ocsp command.
1669 [Rich Salz]
1670
1671 *) Changed default digest for the dgst and enc commands from MD5 to
1672 sha256
1673 [Rich Salz]
1674
1675 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1676 [Matt Caswell]
1677
1678 *) Added support for TLS extended master secret from
1679 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1680 initial patch which was a great help during development.
1681 [Steve Henson]
1682
1683 *) All libssl internal structures have been removed from the public header
1684 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1685 now redundant). Users should not attempt to access internal structures
1686 directly. Instead they should use the provided API functions.
1687 [Matt Caswell]
1688
1689 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1690 Access to deprecated functions can be re-enabled by running config with
1691 "enable-deprecated". In addition applications wishing to use deprecated
1692 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1693 will, by default, disable some transitive includes that previously existed
1694 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1695 [Matt Caswell]
1696
1697 *) Added support for OCB mode. OpenSSL has been granted a patent license
1698 compatible with the OpenSSL license for use of OCB. Details are available
1699 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1700 for OCB can be removed by calling config with no-ocb.
1701 [Matt Caswell]
1702
1703 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1704 compatible client hello.
1705 [Kurt Roeckx]
1706
1707 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1708 done while fixing the error code for the key-too-small case.
1709 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1710
1711 *) CA.sh has been removed; use CA.pl instead.
1712 [Rich Salz]
1713
1714 *) Removed old DES API.
1715 [Rich Salz]
1716
1717 *) Remove various unsupported platforms:
1718 Sony NEWS4
1719 BEOS and BEOS_R5
1720 NeXT
1721 SUNOS
1722 MPE/iX
1723 Sinix/ReliantUNIX RM400
1724 DGUX
1725 NCR
1726 Tandem
1727 Cray
1728 16-bit platforms such as WIN16
1729 [Rich Salz]
1730
1731 *) Clean up OPENSSL_NO_xxx #define's
1732 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1733 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1734 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1735 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1736 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1737 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1738 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1739 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1740 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1741 Remove MS_STATIC; it's a relic from platforms <32 bits.
1742 [Rich Salz]
1743
1744 *) Cleaned up dead code
1745 Remove all but one '#ifdef undef' which is to be looked at.
1746 [Rich Salz]
1747
1748 *) Clean up calling of xxx_free routines.
1749 Just like free(), fix most of the xxx_free routines to accept
1750 NULL. Remove the non-null checks from callers. Save much code.
1751 [Rich Salz]
1752
1753 *) Add secure heap for storage of private keys (when possible).
1754 Add BIO_s_secmem(), CBIGNUM, etc.
1755 Contributed by Akamai Technologies under our Corporate CLA.
1756 [Rich Salz]
1757
1758 *) Experimental support for a new, fast, unbiased prime candidate generator,
1759 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1760 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1761
1762 *) New output format NSS in the sess_id command line tool. This allows
1763 exporting the session id and the master key in NSS keylog format.
1764 [Martin Kaiser <martin@kaiser.cx>]
1765
1766 *) Harmonize version and its documentation. -f flag is used to display
1767 compilation flags.
1768 [mancha <mancha1@zoho.com>]
1769
1770 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1771 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1772 [mancha <mancha1@zoho.com>]
1773
1774 *) Fix some double frees. These are not thought to be exploitable.
1775 [mancha <mancha1@zoho.com>]
1776
1777 *) A missing bounds check in the handling of the TLS heartbeat extension
1778 can be used to reveal up to 64k of memory to a connected client or
1779 server.
1780
1781 Thanks for Neel Mehta of Google Security for discovering this bug and to
1782 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1783 preparing the fix (CVE-2014-0160)
1784 [Adam Langley, Bodo Moeller]
1785
1786 *) Fix for the attack described in the paper "Recovering OpenSSL
1787 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1788 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1789 http://eprint.iacr.org/2014/140
1790
1791 Thanks to Yuval Yarom and Naomi Benger for discovering this
1792 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1793 [Yuval Yarom and Naomi Benger]
1794
1795 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1796 this fixes a limitation in previous versions of OpenSSL.
1797 [Steve Henson]
1798
1799 *) Experimental encrypt-then-mac support.
1800
1801 Experimental support for encrypt then mac from
1802 draft-gutmann-tls-encrypt-then-mac-02.txt
1803
1804 To enable it set the appropriate extension number (0x42 for the test
1805 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1806
1807 For non-compliant peers (i.e. just about everything) this should have no
1808 effect.
1809
1810 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1811
1812 [Steve Henson]
1813
1814 *) Add EVP support for key wrapping algorithms, to avoid problems with
1815 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1816 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1817 algorithms and include tests cases.
1818 [Steve Henson]
1819
1820 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1821 enveloped data.
1822 [Steve Henson]
1823
1824 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1825 MGF1 digest and OAEP label.
1826 [Steve Henson]
1827
1828 *) Make openssl verify return errors.
1829 [Chris Palmer <palmer@google.com> and Ben Laurie]
1830
1831 *) New function ASN1_TIME_diff to calculate the difference between two
1832 ASN1_TIME structures or one structure and the current time.
1833 [Steve Henson]
1834
1835 *) Update fips_test_suite to support multiple command line options. New
1836 test to induce all self test errors in sequence and check expected
1837 failures.
1838 [Steve Henson]
1839
1840 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1841 sign or verify all in one operation.
1842 [Steve Henson]
1843
1844 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1845 test programs and fips_test_suite. Includes functionality to parse
1846 the minimal script output of fipsalgest.pl directly.
1847 [Steve Henson]
1848
1849 *) Add authorisation parameter to FIPS_module_mode_set().
1850 [Steve Henson]
1851
1852 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1853 [Steve Henson]
1854
1855 *) Use separate DRBG fields for internal and external flags. New function
1856 FIPS_drbg_health_check() to perform on demand health checking. Add
1857 generation tests to fips_test_suite with reduced health check interval to
1858 demonstrate periodic health checking. Add "nodh" option to
1859 fips_test_suite to skip very slow DH test.
1860 [Steve Henson]
1861
1862 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1863 based on NID.
1864 [Steve Henson]
1865
1866 *) More extensive health check for DRBG checking many more failure modes.
1867 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1868 combination: call this in fips_test_suite.
1869 [Steve Henson]
1870
1871 *) Add support for canonical generation of DSA parameter 'g'. See
1872 FIPS 186-3 A.2.3.
1873
1874 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1875 POST to handle HMAC cases.
1876 [Steve Henson]
1877
1878 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1879 to return numerical and string versions of the FIPS module number.
1880 [Steve Henson]
1881
1882 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1883 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1884 outside the validated module in the FIPS capable OpenSSL.
1885 [Steve Henson]
1886
1887 *) Minor change to DRBG entropy callback semantics. In some cases
1888 there is no multiple of the block length between min_len and
1889 max_len. Allow the callback to return more than max_len bytes
1890 of entropy but discard any extra: it is the callback's responsibility
1891 to ensure that the extra data discarded does not impact the
1892 requested amount of entropy.
1893 [Steve Henson]
1894
1895 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1896 information in FIPS186-3, SP800-57 and SP800-131A.
1897 [Steve Henson]
1898
1899 *) CCM support via EVP. Interface is very similar to GCM case except we
1900 must supply all data in one chunk (i.e. no update, final) and the
1901 message length must be supplied if AAD is used. Add algorithm test
1902 support.
1903 [Steve Henson]
1904
1905 *) Initial version of POST overhaul. Add POST callback to allow the status
1906 of POST to be monitored and/or failures induced. Modify fips_test_suite
1907 to use callback. Always run all selftests even if one fails.
1908 [Steve Henson]
1909
1910 *) XTS support including algorithm test driver in the fips_gcmtest program.
1911 Note: this does increase the maximum key length from 32 to 64 bytes but
1912 there should be no binary compatibility issues as existing applications
1913 will never use XTS mode.
1914 [Steve Henson]
1915
1916 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1917 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1918 performs algorithm blocking for unapproved PRNG types. Also do not
1919 set PRNG type in FIPS_mode_set(): leave this to the application.
1920 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1921 the standard OpenSSL PRNG: set additional data to a date time vector.
1922 [Steve Henson]
1923
1924 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1925 This shouldn't present any incompatibility problems because applications
1926 shouldn't be using these directly and any that are will need to rethink
1927 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1928 [Steve Henson]
1929
1930 *) Extensive self tests and health checking required by SP800-90 DRBG.
1931 Remove strength parameter from FIPS_drbg_instantiate and always
1932 instantiate at maximum supported strength.
1933 [Steve Henson]
1934
1935 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1936 [Steve Henson]
1937
1938 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1939 [Steve Henson]
1940
1941 *) New function DH_compute_key_padded() to compute a DH key and pad with
1942 leading zeroes if needed: this complies with SP800-56A et al.
1943 [Steve Henson]
1944
1945 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1946 anything, incomplete, subject to change and largely untested at present.
1947 [Steve Henson]
1948
1949 *) Modify fipscanisteronly build option to only build the necessary object
1950 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
1951 [Steve Henson]
1952
1953 *) Add experimental option FIPSSYMS to give all symbols in
1954 fipscanister.o and FIPS or fips prefix. This will avoid
1955 conflicts with future versions of OpenSSL. Add perl script
1956 util/fipsas.pl to preprocess assembly language source files
1957 and rename any affected symbols.
1958 [Steve Henson]
1959
1960 *) Add selftest checks and algorithm block of non-fips algorithms in
1961 FIPS mode. Remove DES2 from selftests.
1962 [Steve Henson]
1963
1964 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
1965 return internal method without any ENGINE dependencies. Add new
1966 tiny fips sign and verify functions.
1967 [Steve Henson]
1968
1969 *) New build option no-ec2m to disable characteristic 2 code.
1970 [Steve Henson]
1971
1972 *) New build option "fipscanisteronly". This only builds fipscanister.o
1973 and (currently) associated fips utilities. Uses the file Makefile.fips
1974 instead of Makefile.org as the prototype.
1975 [Steve Henson]
1976
1977 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
1978 Update fips_gcmtest to use IV generator.
1979 [Steve Henson]
1980
1981 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
1982 setting output buffer to NULL. The *Final function must be
1983 called although it will not retrieve any additional data. The tag
1984 can be set or retrieved with a ctrl. The IV length is by default 12
1985 bytes (96 bits) but can be set to an alternative value. If the IV
1986 length exceeds the maximum IV length (currently 16 bytes) it cannot be
1987 set before the key.
1988 [Steve Henson]
1989
1990 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
1991 underlying do_cipher function handles all cipher semantics itself
1992 including padding and finalisation. This is useful if (for example)
1993 an ENGINE cipher handles block padding itself. The behaviour of
1994 do_cipher is subtly changed if this flag is set: the return value
1995 is the number of characters written to the output buffer (zero is
1996 no longer an error code) or a negative error code. Also if the
1997 input buffer is NULL and length 0 finalisation should be performed.
1998 [Steve Henson]
1999
2000 *) If a candidate issuer certificate is already part of the constructed
2001 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
2002 [Steve Henson]
2003
2004 *) Improve forward-security support: add functions
2005
2006 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
2007 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
2008
2009 for use by SSL/TLS servers; the callback function will be called whenever a
2010 new session is created, and gets to decide whether the session may be
2011 cached to make it resumable (return 0) or not (return 1). (As by the
2012 SSL/TLS protocol specifications, the session_id sent by the server will be
2013 empty to indicate that the session is not resumable; also, the server will
2014 not generate RFC 4507 (RFC 5077) session tickets.)
2015
2016 A simple reasonable callback implementation is to return is_forward_secure.
2017 This parameter will be set to 1 or 0 depending on the ciphersuite selected
2018 by the SSL/TLS server library, indicating whether it can provide forward
2019 security.
2020 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2021
2022 *) New -verify_name option in command line utilities to set verification
2023 parameters by name.
2024 [Steve Henson]
2025
2026 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
2027 Add CMAC pkey methods.
2028 [Steve Henson]
2029
2030 *) Experimental renegotiation in s_server -www mode. If the client
2031 browses /reneg connection is renegotiated. If /renegcert it is
2032 renegotiated requesting a certificate.
2033 [Steve Henson]
2034
2035 *) Add an "external" session cache for debugging purposes to s_server. This
2036 should help trace issues which normally are only apparent in deployed
2037 multi-process servers.
2038 [Steve Henson]
2039
2040 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
2041 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
2042 BIO_set_cipher() and some obscure PEM functions were changed so they
2043 can now return an error. The RAND changes required a change to the
2044 RAND_METHOD structure.
2045 [Steve Henson]
2046
2047 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
2048 a gcc attribute to warn if the result of a function is ignored. This
2049 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
2050 whose return value is often ignored.
2051 [Steve Henson]
2052
2053 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
2054 These allow SCTs (signed certificate timestamps) to be requested and
2055 validated when establishing a connection.
2056 [Rob Percival <robpercival@google.com>]
2057
2058 Changes between 1.0.2g and 1.0.2h [3 May 2016]
2059
2060 *) Prevent padding oracle in AES-NI CBC MAC check
2061
2062 A MITM attacker can use a padding oracle attack to decrypt traffic
2063 when the connection uses an AES CBC cipher and the server support
2064 AES-NI.
2065
2066 This issue was introduced as part of the fix for Lucky 13 padding
2067 attack (CVE-2013-0169). The padding check was rewritten to be in
2068 constant time by making sure that always the same bytes are read and
2069 compared against either the MAC or padding bytes. But it no longer
2070 checked that there was enough data to have both the MAC and padding
2071 bytes.
2072
2073 This issue was reported by Juraj Somorovsky using TLS-Attacker.
2074 (CVE-2016-2107)
2075 [Kurt Roeckx]
2076
2077 *) Fix EVP_EncodeUpdate overflow
2078
2079 An overflow can occur in the EVP_EncodeUpdate() function which is used for
2080 Base64 encoding of binary data. If an attacker is able to supply very large
2081 amounts of input data then a length check can overflow resulting in a heap
2082 corruption.
2083
2084 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
2085 the PEM_write_bio* family of functions. These are mainly used within the
2086 OpenSSL command line applications, so any application which processes data
2087 from an untrusted source and outputs it as a PEM file should be considered
2088 vulnerable to this issue. User applications that call these APIs directly
2089 with large amounts of untrusted data may also be vulnerable.
2090
2091 This issue was reported by Guido Vranken.
2092 (CVE-2016-2105)
2093 [Matt Caswell]
2094
2095 *) Fix EVP_EncryptUpdate overflow
2096
2097 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
2098 is able to supply very large amounts of input data after a previous call to
2099 EVP_EncryptUpdate() with a partial block then a length check can overflow
2100 resulting in a heap corruption. Following an analysis of all OpenSSL
2101 internal usage of the EVP_EncryptUpdate() function all usage is one of two
2102 forms. The first form is where the EVP_EncryptUpdate() call is known to be
2103 the first called function after an EVP_EncryptInit(), and therefore that
2104 specific call must be safe. The second form is where the length passed to
2105 EVP_EncryptUpdate() can be seen from the code to be some small value and
2106 therefore there is no possibility of an overflow. Since all instances are
2107 one of these two forms, it is believed that there can be no overflows in
2108 internal code due to this problem. It should be noted that
2109 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
2110 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
2111 of these calls have also been analysed too and it is believed there are no
2112 instances in internal usage where an overflow could occur.
2113
2114 This issue was reported by Guido Vranken.
2115 (CVE-2016-2106)
2116 [Matt Caswell]
2117
2118 *) Prevent ASN.1 BIO excessive memory allocation
2119
2120 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2121 a short invalid encoding can cause allocation of large amounts of memory
2122 potentially consuming excessive resources or exhausting memory.
2123
2124 Any application parsing untrusted data through d2i BIO functions is
2125 affected. The memory based functions such as d2i_X509() are *not* affected.
2126 Since the memory based functions are used by the TLS library, TLS
2127 applications are not affected.
2128
2129 This issue was reported by Brian Carpenter.
2130 (CVE-2016-2109)
2131 [Stephen Henson]
2132
2133 *) EBCDIC overread
2134
2135 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2136 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2137 in arbitrary stack data being returned in the buffer.
2138
2139 This issue was reported by Guido Vranken.
2140 (CVE-2016-2176)
2141 [Matt Caswell]
2142
2143 *) Modify behavior of ALPN to invoke callback after SNI/servername
2144 callback, such that updates to the SSL_CTX affect ALPN.
2145 [Todd Short]
2146
2147 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2148 default.
2149 [Kurt Roeckx]
2150
2151 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2152 methods are enabled and ssl2 is disabled the methods return NULL.
2153 [Kurt Roeckx]
2154
2155 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2156
2157 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2158 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2159 provide any "EXPORT" or "LOW" strength ciphers.
2160 [Viktor Dukhovni]
2161
2162 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2163 is by default disabled at build-time. Builds that are not configured with
2164 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2165 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2166 will need to explicitly call either of:
2167
2168 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2169 or
2170 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2171
2172 as appropriate. Even if either of those is used, or the application
2173 explicitly uses the version-specific SSLv2_method() or its client and
2174 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2175 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2176 ciphers, and SSLv2 56-bit DES are no longer available.
2177 (CVE-2016-0800)
2178 [Viktor Dukhovni]
2179
2180 *) Fix a double-free in DSA code
2181
2182 A double free bug was discovered when OpenSSL parses malformed DSA private
2183 keys and could lead to a DoS attack or memory corruption for applications
2184 that receive DSA private keys from untrusted sources. This scenario is
2185 considered rare.
2186
2187 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2188 libFuzzer.
2189 (CVE-2016-0705)
2190 [Stephen Henson]
2191
2192 *) Disable SRP fake user seed to address a server memory leak.
2193
2194 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2195
2196 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2197 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2198 was changed to ignore the "fake user" SRP seed, even if the seed
2199 is configured.
2200
2201 Users should use SRP_VBASE_get1_by_user instead. Note that in
2202 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2203 also that even though configuring the SRP seed attempts to hide
2204 invalid usernames by continuing the handshake with fake
2205 credentials, this behaviour is not constant time and no strong
2206 guarantees are made that the handshake is indistinguishable from
2207 that of a valid user.
2208 (CVE-2016-0798)
2209 [Emilia Käsper]
2210
2211 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2212
2213 In the BN_hex2bn function the number of hex digits is calculated using an
2214 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2215 large values of |i| this can result in |bn_expand| not allocating any
2216 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2217 field as NULL leading to a subsequent NULL ptr deref. For very large values
2218 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2219 In this case memory is allocated to the internal BIGNUM data field, but it
2220 is insufficiently sized leading to heap corruption. A similar issue exists
2221 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2222 is ever called by user applications with very large untrusted hex/dec data.
2223 This is anticipated to be a rare occurrence.
2224
2225 All OpenSSL internal usage of these functions use data that is not expected
2226 to be untrusted, e.g. config file data or application command line
2227 arguments. If user developed applications generate config file data based
2228 on untrusted data then it is possible that this could also lead to security
2229 consequences. This is also anticipated to be rare.
2230
2231 This issue was reported to OpenSSL by Guido Vranken.
2232 (CVE-2016-0797)
2233 [Matt Caswell]
2234
2235 *) Fix memory issues in BIO_*printf functions
2236
2237 The internal |fmtstr| function used in processing a "%s" format string in
2238 the BIO_*printf functions could overflow while calculating the length of a
2239 string and cause an OOB read when printing very long strings.
2240
2241 Additionally the internal |doapr_outch| function can attempt to write to an
2242 OOB memory location (at an offset from the NULL pointer) in the event of a
2243 memory allocation failure. In 1.0.2 and below this could be caused where
2244 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2245 could be in processing a very long "%s" format string. Memory leaks can
2246 also occur.
2247
2248 The first issue may mask the second issue dependent on compiler behaviour.
2249 These problems could enable attacks where large amounts of untrusted data
2250 is passed to the BIO_*printf functions. If applications use these functions
2251 in this way then they could be vulnerable. OpenSSL itself uses these
2252 functions when printing out human-readable dumps of ASN.1 data. Therefore
2253 applications that print this data could be vulnerable if the data is from
2254 untrusted sources. OpenSSL command line applications could also be
2255 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2256 as command line arguments.
2257
2258 Libssl is not considered directly vulnerable. Additionally certificates etc
2259 received via remote connections via libssl are also unlikely to be able to
2260 trigger these issues because of message size limits enforced within libssl.
2261
2262 This issue was reported to OpenSSL Guido Vranken.
2263 (CVE-2016-0799)
2264 [Matt Caswell]
2265
2266 *) Side channel attack on modular exponentiation
2267
2268 A side-channel attack was found which makes use of cache-bank conflicts on
2269 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2270 of RSA keys. The ability to exploit this issue is limited as it relies on
2271 an attacker who has control of code in a thread running on the same
2272 hyper-threaded core as the victim thread which is performing decryptions.
2273
2274 This issue was reported to OpenSSL by Yuval Yarom, The University of
2275 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2276 Nadia Heninger, University of Pennsylvania with more information at
2277 http://cachebleed.info.
2278 (CVE-2016-0702)
2279 [Andy Polyakov]
2280
2281 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2282 if no keysize is specified with default_bits. This fixes an
2283 omission in an earlier change that changed all RSA/DSA key generation
2284 apps to use 2048 bits by default.
2285 [Emilia Käsper]
2286
2287 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2288 *) DH small subgroups
2289
2290 Historically OpenSSL only ever generated DH parameters based on "safe"
2291 primes. More recently (in version 1.0.2) support was provided for
2292 generating X9.42 style parameter files such as those required for RFC 5114
2293 support. The primes used in such files may not be "safe". Where an
2294 application is using DH configured with parameters based on primes that are
2295 not "safe" then an attacker could use this fact to find a peer's private
2296 DH exponent. This attack requires that the attacker complete multiple
2297 handshakes in which the peer uses the same private DH exponent. For example
2298 this could be used to discover a TLS server's private DH exponent if it's
2299 reusing the private DH exponent or it's using a static DH ciphersuite.
2300
2301 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2302 TLS. It is not on by default. If the option is not set then the server
2303 reuses the same private DH exponent for the life of the server process and
2304 would be vulnerable to this attack. It is believed that many popular
2305 applications do set this option and would therefore not be at risk.
2306
2307 The fix for this issue adds an additional check where a "q" parameter is
2308 available (as is the case in X9.42 based parameters). This detects the
2309 only known attack, and is the only possible defense for static DH
2310 ciphersuites. This could have some performance impact.
2311
2312 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2313 default and cannot be disabled. This could have some performance impact.
2314
2315 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2316 (CVE-2016-0701)
2317 [Matt Caswell]
2318
2319 *) SSLv2 doesn't block disabled ciphers
2320
2321 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2322 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2323 been disabled, provided that the SSLv2 protocol was not also disabled via
2324 SSL_OP_NO_SSLv2.
2325
2326 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2327 and Sebastian Schinzel.
2328 (CVE-2015-3197)
2329 [Viktor Dukhovni]
2330
2331 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2332
2333 *) BN_mod_exp may produce incorrect results on x86_64
2334
2335 There is a carry propagating bug in the x86_64 Montgomery squaring
2336 procedure. No EC algorithms are affected. Analysis suggests that attacks
2337 against RSA and DSA as a result of this defect would be very difficult to
2338 perform and are not believed likely. Attacks against DH are considered just
2339 feasible (although very difficult) because most of the work necessary to
2340 deduce information about a private key may be performed offline. The amount
2341 of resources required for such an attack would be very significant and
2342 likely only accessible to a limited number of attackers. An attacker would
2343 additionally need online access to an unpatched system using the target
2344 private key in a scenario with persistent DH parameters and a private
2345 key that is shared between multiple clients. For example this can occur by
2346 default in OpenSSL DHE based SSL/TLS ciphersuites.
2347
2348 This issue was reported to OpenSSL by Hanno Böck.
2349 (CVE-2015-3193)
2350 [Andy Polyakov]
2351
2352 *) Certificate verify crash with missing PSS parameter
2353
2354 The signature verification routines will crash with a NULL pointer
2355 dereference if presented with an ASN.1 signature using the RSA PSS
2356 algorithm and absent mask generation function parameter. Since these
2357 routines are used to verify certificate signature algorithms this can be
2358 used to crash any certificate verification operation and exploited in a
2359 DoS attack. Any application which performs certificate verification is
2360 vulnerable including OpenSSL clients and servers which enable client
2361 authentication.
2362
2363 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2364 (CVE-2015-3194)
2365 [Stephen Henson]
2366
2367 *) X509_ATTRIBUTE memory leak
2368
2369 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2370 memory. This structure is used by the PKCS#7 and CMS routines so any
2371 application which reads PKCS#7 or CMS data from untrusted sources is
2372 affected. SSL/TLS is not affected.
2373
2374 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2375 libFuzzer.
2376 (CVE-2015-3195)
2377 [Stephen Henson]
2378
2379 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2380 This changes the decoding behaviour for some invalid messages,
2381 though the change is mostly in the more lenient direction, and
2382 legacy behaviour is preserved as much as possible.
2383 [Emilia Käsper]
2384
2385 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2386 return an error
2387 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2388
2389 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2390
2391 *) Alternate chains certificate forgery
2392
2393 During certificate verification, OpenSSL will attempt to find an
2394 alternative certificate chain if the first attempt to build such a chain
2395 fails. An error in the implementation of this logic can mean that an
2396 attacker could cause certain checks on untrusted certificates to be
2397 bypassed, such as the CA flag, enabling them to use a valid leaf
2398 certificate to act as a CA and "issue" an invalid certificate.
2399
2400 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2401 (Google/BoringSSL).
2402 [Matt Caswell]
2403
2404 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2405
2406 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2407 incompatibility in the handling of HMAC. The previous ABI has now been
2408 restored.
2409 [Matt Caswell]
2410
2411 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2412
2413 *) Malformed ECParameters causes infinite loop
2414
2415 When processing an ECParameters structure OpenSSL enters an infinite loop
2416 if the curve specified is over a specially malformed binary polynomial
2417 field.
2418
2419 This can be used to perform denial of service against any
2420 system which processes public keys, certificate requests or
2421 certificates. This includes TLS clients and TLS servers with
2422 client authentication enabled.
2423
2424 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2425 (CVE-2015-1788)
2426 [Andy Polyakov]
2427
2428 *) Exploitable out-of-bounds read in X509_cmp_time
2429
2430 X509_cmp_time does not properly check the length of the ASN1_TIME
2431 string and can read a few bytes out of bounds. In addition,
2432 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2433 time string.
2434
2435 An attacker can use this to craft malformed certificates and CRLs of
2436 various sizes and potentially cause a segmentation fault, resulting in
2437 a DoS on applications that verify certificates or CRLs. TLS clients
2438 that verify CRLs are affected. TLS clients and servers with client
2439 authentication enabled may be affected if they use custom verification
2440 callbacks.
2441
2442 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2443 independently by Hanno Böck.
2444 (CVE-2015-1789)
2445 [Emilia Käsper]
2446
2447 *) PKCS7 crash with missing EnvelopedContent
2448
2449 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2450 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2451 with missing content and trigger a NULL pointer dereference on parsing.
2452
2453 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2454 structures from untrusted sources are affected. OpenSSL clients and
2455 servers are not affected.
2456
2457 This issue was reported to OpenSSL by Michal Zalewski (Google).
2458 (CVE-2015-1790)
2459 [Emilia Käsper]
2460
2461 *) CMS verify infinite loop with unknown hash function
2462
2463 When verifying a signedData message the CMS code can enter an infinite loop
2464 if presented with an unknown hash function OID. This can be used to perform
2465 denial of service against any system which verifies signedData messages using
2466 the CMS code.
2467 This issue was reported to OpenSSL by Johannes Bauer.
2468 (CVE-2015-1792)
2469 [Stephen Henson]
2470
2471 *) Race condition handling NewSessionTicket
2472
2473 If a NewSessionTicket is received by a multi-threaded client when attempting to
2474 reuse a previous ticket then a race condition can occur potentially leading to
2475 a double free of the ticket data.
2476 (CVE-2015-1791)
2477 [Matt Caswell]
2478
2479 *) Only support 256-bit or stronger elliptic curves with the
2480 'ecdh_auto' setting (server) or by default (client). Of supported
2481 curves, prefer P-256 (both).
2482 [Emilia Kasper]
2483
2484 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2485
2486 *) ClientHello sigalgs DoS fix
2487
2488 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2489 invalid signature algorithms extension a NULL pointer dereference will
2490 occur. This can be exploited in a DoS attack against the server.
2491
2492 This issue was was reported to OpenSSL by David Ramos of Stanford
2493 University.
2494 (CVE-2015-0291)
2495 [Stephen Henson and Matt Caswell]
2496
2497 *) Multiblock corrupted pointer fix
2498
2499 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2500 feature only applies on 64 bit x86 architecture platforms that support AES
2501 NI instructions. A defect in the implementation of "multiblock" can cause
2502 OpenSSL's internal write buffer to become incorrectly set to NULL when
2503 using non-blocking IO. Typically, when the user application is using a
2504 socket BIO for writing, this will only result in a failed connection.
2505 However if some other BIO is used then it is likely that a segmentation
2506 fault will be triggered, thus enabling a potential DoS attack.
2507
2508 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2509 (CVE-2015-0290)
2510 [Matt Caswell]
2511
2512 *) Segmentation fault in DTLSv1_listen fix
2513
2514 The DTLSv1_listen function is intended to be stateless and processes the
2515 initial ClientHello from many peers. It is common for user code to loop
2516 over the call to DTLSv1_listen until a valid ClientHello is received with
2517 an associated cookie. A defect in the implementation of DTLSv1_listen means
2518 that state is preserved in the SSL object from one invocation to the next
2519 that can lead to a segmentation fault. Errors processing the initial
2520 ClientHello can trigger this scenario. An example of such an error could be
2521 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2522 server.
2523
2524 This issue was reported to OpenSSL by Per Allansson.
2525 (CVE-2015-0207)
2526 [Matt Caswell]
2527
2528 *) Segmentation fault in ASN1_TYPE_cmp fix
2529
2530 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2531 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2532 certificate signature algorithm consistency this can be used to crash any
2533 certificate verification operation and exploited in a DoS attack. Any
2534 application which performs certificate verification is vulnerable including
2535 OpenSSL clients and servers which enable client authentication.
2536 (CVE-2015-0286)
2537 [Stephen Henson]
2538
2539 *) Segmentation fault for invalid PSS parameters fix
2540
2541 The signature verification routines will crash with a NULL pointer
2542 dereference if presented with an ASN.1 signature using the RSA PSS
2543 algorithm and invalid parameters. Since these routines are used to verify
2544 certificate signature algorithms this can be used to crash any
2545 certificate verification operation and exploited in a DoS attack. Any
2546 application which performs certificate verification is vulnerable including
2547 OpenSSL clients and servers which enable client authentication.
2548
2549 This issue was was reported to OpenSSL by Brian Carpenter.
2550 (CVE-2015-0208)
2551 [Stephen Henson]
2552
2553 *) ASN.1 structure reuse memory corruption fix
2554
2555 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2556 memory corruption via an invalid write. Such reuse is and has been
2557 strongly discouraged and is believed to be rare.
2558
2559 Applications that parse structures containing CHOICE or ANY DEFINED BY
2560 components may be affected. Certificate parsing (d2i_X509 and related
2561 functions) are however not affected. OpenSSL clients and servers are
2562 not affected.
2563 (CVE-2015-0287)
2564 [Stephen Henson]
2565
2566 *) PKCS7 NULL pointer dereferences fix
2567
2568 The PKCS#7 parsing code does not handle missing outer ContentInfo
2569 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2570 missing content and trigger a NULL pointer dereference on parsing.
2571
2572 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2573 otherwise parse PKCS#7 structures from untrusted sources are
2574 affected. OpenSSL clients and servers are not affected.
2575
2576 This issue was reported to OpenSSL by Michal Zalewski (Google).
2577 (CVE-2015-0289)
2578 [Emilia Käsper]
2579
2580 *) DoS via reachable assert in SSLv2 servers fix
2581
2582 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2583 servers that both support SSLv2 and enable export cipher suites by sending
2584 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2585
2586 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2587 (OpenSSL development team).
2588 (CVE-2015-0293)
2589 [Emilia Käsper]
2590
2591 *) Empty CKE with client auth and DHE fix
2592
2593 If client auth is used then a server can seg fault in the event of a DHE
2594 ciphersuite being selected and a zero length ClientKeyExchange message
2595 being sent by the client. This could be exploited in a DoS attack.
2596 (CVE-2015-1787)
2597 [Matt Caswell]
2598
2599 *) Handshake with unseeded PRNG fix
2600
2601 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2602 with an unseeded PRNG. The conditions are:
2603 - The client is on a platform where the PRNG has not been seeded
2604 automatically, and the user has not seeded manually
2605 - A protocol specific client method version has been used (i.e. not
2606 SSL_client_methodv23)
2607 - A ciphersuite is used that does not require additional random data from
2608 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2609
2610 If the handshake succeeds then the client random that has been used will
2611 have been generated from a PRNG with insufficient entropy and therefore the
2612 output may be predictable.
2613
2614 For example using the following command with an unseeded openssl will
2615 succeed on an unpatched platform:
2616
2617 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2618 (CVE-2015-0285)
2619 [Matt Caswell]
2620
2621 *) Use After Free following d2i_ECPrivatekey error fix
2622
2623 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2624 could cause a use after free condition. This, in turn, could cause a double
2625 free in several private key parsing functions (such as d2i_PrivateKey
2626 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2627 for applications that receive EC private keys from untrusted
2628 sources. This scenario is considered rare.
2629
2630 This issue was discovered by the BoringSSL project and fixed in their
2631 commit 517073cd4b.
2632 (CVE-2015-0209)
2633 [Matt Caswell]
2634
2635 *) X509_to_X509_REQ NULL pointer deref fix
2636
2637 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2638 the certificate key is invalid. This function is rarely used in practice.
2639
2640 This issue was discovered by Brian Carpenter.
2641 (CVE-2015-0288)
2642 [Stephen Henson]
2643
2644 *) Removed the export ciphers from the DEFAULT ciphers
2645 [Kurt Roeckx]
2646
2647 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2648
2649 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2650 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2651 So far those who have to target multiple platforms would compromise
2652 and argue that binary targeting say ARMv5 would still execute on
2653 ARMv8. "Universal" build resolves this compromise by providing
2654 near-optimal performance even on newer platforms.
2655 [Andy Polyakov]
2656
2657 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2658 (other platforms pending).
2659 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2660
2661 *) Add support for the SignedCertificateTimestampList certificate and
2662 OCSP response extensions from RFC6962.
2663 [Rob Stradling]
2664
2665 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2666 for corner cases. (Certain input points at infinity could lead to
2667 bogus results, with non-infinity inputs mapped to infinity too.)
2668 [Bodo Moeller]
2669
2670 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2671 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2672 common cases are optimized and there still is room for further
2673 improvements. Vector Permutation AES for Altivec is also added.
2674 [Andy Polyakov]
2675
2676 *) Add support for little-endian ppc64 Linux target.
2677 [Marcelo Cerri (IBM)]
2678
2679 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2680 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2681 are optimized and there still is room for further improvements.
2682 Both 32- and 64-bit modes are supported.
2683 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2684
2685 *) Improved ARMv7 NEON support.
2686 [Andy Polyakov]
2687
2688 *) Support for SPARC Architecture 2011 crypto extensions, first
2689 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2690 SHA256/512, MD5, GHASH and modular exponentiation.
2691 [Andy Polyakov, David Miller]
2692
2693 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2694 RSAZ.
2695 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2696
2697 *) Support for new and upcoming Intel processors, including AVX2,
2698 BMI and SHA ISA extensions. This includes additional "stitched"
2699 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2700 for TLS encrypt.
2701
2702 This work was sponsored by Intel Corp.
2703 [Andy Polyakov]
2704
2705 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2706 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2707 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2708 [Steve Henson]
2709
2710 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2711 this fixes a limitation in previous versions of OpenSSL.
2712 [Steve Henson]
2713
2714 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2715 MGF1 digest and OAEP label.
2716 [Steve Henson]
2717
2718 *) Add EVP support for key wrapping algorithms, to avoid problems with
2719 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2720 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2721 algorithms and include tests cases.
2722 [Steve Henson]
2723
2724 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2725 structure.
2726 [Douglas E. Engert, Steve Henson]
2727
2728 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2729 difference in days and seconds between two tm or ASN1_TIME structures.
2730 [Steve Henson]
2731
2732 *) Add -rev test option to s_server to just reverse order of characters
2733 received by client and send back to server. Also prints an abbreviated
2734 summary of the connection parameters.
2735 [Steve Henson]
2736
2737 *) New option -brief for s_client and s_server to print out a brief summary
2738 of connection parameters.
2739 [Steve Henson]
2740
2741 *) Add callbacks for arbitrary TLS extensions.
2742 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2743
2744 *) New option -crl_download in several openssl utilities to download CRLs
2745 from CRLDP extension in certificates.
2746 [Steve Henson]
2747
2748 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2749 [Steve Henson]
2750
2751 *) New function X509_CRL_diff to generate a delta CRL from the difference
2752 of two full CRLs. Add support to "crl" utility.
2753 [Steve Henson]
2754
2755 *) New functions to set lookup_crls function and to retrieve
2756 X509_STORE from X509_STORE_CTX.
2757 [Steve Henson]
2758
2759 *) Print out deprecated issuer and subject unique ID fields in
2760 certificates.
2761 [Steve Henson]
2762
2763 *) Extend OCSP I/O functions so they can be used for simple general purpose
2764 HTTP as well as OCSP. New wrapper function which can be used to download
2765 CRLs using the OCSP API.
2766 [Steve Henson]
2767
2768 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2769 [Steve Henson]
2770
2771 *) SSL_CONF* functions. These provide a common framework for application
2772 configuration using configuration files or command lines.
2773 [Steve Henson]
2774
2775 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2776 message callback and prints the results. Needs compile time option
2777 "enable-ssl-trace". New options to s_client and s_server to enable
2778 tracing.
2779 [Steve Henson]
2780
2781 *) New ctrl and macro to retrieve supported points extensions.
2782 Print out extension in s_server and s_client.
2783 [Steve Henson]
2784
2785 *) New functions to retrieve certificate signature and signature
2786 OID NID.
2787 [Steve Henson]
2788
2789 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2790 client to OpenSSL.
2791 [Steve Henson]
2792
2793 *) New Suite B modes for TLS code. These use and enforce the requirements
2794 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2795 only use Suite B curves. The Suite B modes can be set by using the
2796 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2797 [Steve Henson]
2798
2799 *) New chain verification flags for Suite B levels of security. Check
2800 algorithms are acceptable when flags are set in X509_verify_cert.
2801 [Steve Henson]
2802
2803 *) Make tls1_check_chain return a set of flags indicating checks passed
2804 by a certificate chain. Add additional tests to handle client
2805 certificates: checks for matching certificate type and issuer name
2806 comparison.
2807 [Steve Henson]
2808
2809 *) If an attempt is made to use a signature algorithm not in the peer
2810 preference list abort the handshake. If client has no suitable
2811 signature algorithms in response to a certificate request do not
2812 use the certificate.
2813 [Steve Henson]
2814
2815 *) If server EC tmp key is not in client preference list abort handshake.
2816 [Steve Henson]
2817
2818 *) Add support for certificate stores in CERT structure. This makes it
2819 possible to have different stores per SSL structure or one store in
2820 the parent SSL_CTX. Include distinct stores for certificate chain
2821 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2822 to build and store a certificate chain in CERT structure: returning
2823 an error if the chain cannot be built: this will allow applications
2824 to test if a chain is correctly configured.
2825
2826 Note: if the CERT based stores are not set then the parent SSL_CTX
2827 store is used to retain compatibility with existing behaviour.
2828
2829 [Steve Henson]
2830
2831 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2832 mask based on the current session, check mask when sending client
2833 hello and checking the requested ciphersuite.
2834 [Steve Henson]
2835
2836 *) New ctrls to retrieve and set certificate types in a certificate
2837 request message. Print out received values in s_client. If certificate
2838 types is not set with custom values set sensible values based on
2839 supported signature algorithms.
2840 [Steve Henson]
2841
2842 *) Support for distinct client and server supported signature algorithms.
2843 [Steve Henson]
2844
2845 *) Add certificate callback. If set this is called whenever a certificate
2846 is required by client or server. An application can decide which
2847 certificate chain to present based on arbitrary criteria: for example
2848 supported signature algorithms. Add very simple example to s_server.
2849 This fixes many of the problems and restrictions of the existing client
2850 certificate callback: for example you can now clear an existing
2851 certificate and specify the whole chain.
2852 [Steve Henson]
2853
2854 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2855 the certificate can be used for (if anything). Set valid_flags field
2856 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2857 to have similar checks in it.
2858
2859 Add new "cert_flags" field to CERT structure and include a "strict mode".
2860 This enforces some TLS certificate requirements (such as only permitting
2861 certificate signature algorithms contained in the supported algorithms
2862 extension) which some implementations ignore: this option should be used
2863 with caution as it could cause interoperability issues.
2864 [Steve Henson]
2865
2866 *) Update and tidy signature algorithm extension processing. Work out
2867 shared signature algorithms based on preferences and peer algorithms
2868 and print them out in s_client and s_server. Abort handshake if no
2869 shared signature algorithms.
2870 [Steve Henson]
2871
2872 *) Add new functions to allow customised supported signature algorithms
2873 for SSL and SSL_CTX structures. Add options to s_client and s_server
2874 to support them.
2875 [Steve Henson]
2876
2877 *) New function SSL_certs_clear() to delete all references to certificates
2878 from an SSL structure. Before this once a certificate had been added
2879 it couldn't be removed.
2880 [Steve Henson]
2881
2882 *) Integrate hostname, email address and IP address checking with certificate
2883 verification. New verify options supporting checking in openssl utility.
2884 [Steve Henson]
2885
2886 *) Fixes and wildcard matching support to hostname and email checking
2887 functions. Add manual page.
2888 [Florian Weimer (Red Hat Product Security Team)]
2889
2890 *) New functions to check a hostname email or IP address against a
2891 certificate. Add options x509 utility to print results of checks against
2892 a certificate.
2893 [Steve Henson]
2894
2895 *) Fix OCSP checking.
2896 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2897
2898 *) Initial experimental support for explicitly trusted non-root CAs.
2899 OpenSSL still tries to build a complete chain to a root but if an
2900 intermediate CA has a trust setting included that is used. The first
2901 setting is used: whether to trust (e.g., -addtrust option to the x509
2902 utility) or reject.
2903 [Steve Henson]
2904
2905 *) Add -trusted_first option which attempts to find certificates in the
2906 trusted store even if an untrusted chain is also supplied.
2907 [Steve Henson]
2908
2909 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2910 platform support for Linux and Android.
2911 [Andy Polyakov]
2912
2913 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2914 [Andy Polyakov]
2915
2916 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2917 When in FIPS mode the approved implementations are used as normal,
2918 when not in FIPS mode the internal unapproved versions are used instead.
2919 This means that the FIPS capable OpenSSL isn't forced to use the
2920 (often lower performance) FIPS implementations outside FIPS mode.
2921 [Steve Henson]
2922
2923 *) Transparently support X9.42 DH parameters when calling
2924 PEM_read_bio_DHparameters. This means existing applications can handle
2925 the new parameter format automatically.
2926 [Steve Henson]
2927
2928 *) Initial experimental support for X9.42 DH parameter format: mainly
2929 to support use of 'q' parameter for RFC5114 parameters.
2930 [Steve Henson]
2931
2932 *) Add DH parameters from RFC5114 including test data to dhtest.
2933 [Steve Henson]
2934
2935 *) Support for automatic EC temporary key parameter selection. If enabled
2936 the most preferred EC parameters are automatically used instead of
2937 hardcoded fixed parameters. Now a server just has to call:
2938 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2939 support ECDH and use the most appropriate parameters.
2940 [Steve Henson]
2941
2942 *) Enhance and tidy EC curve and point format TLS extension code. Use
2943 static structures instead of allocation if default values are used.
2944 New ctrls to set curves we wish to support and to retrieve shared curves.
2945 Print out shared curves in s_server. New options to s_server and s_client
2946 to set list of supported curves.
2947 [Steve Henson]
2948
2949 *) New ctrls to retrieve supported signature algorithms and
2950 supported curve values as an array of NIDs. Extend openssl utility
2951 to print out received values.
2952 [Steve Henson]
2953
2954 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
2955 between NIDs and the more common NIST names such as "P-256". Enhance
2956 ecparam utility and ECC method to recognise the NIST names for curves.
2957 [Steve Henson]
2958
2959 *) Enhance SSL/TLS certificate chain handling to support different
2960 chains for each certificate instead of one chain in the parent SSL_CTX.
2961 [Steve Henson]
2962
2963 *) Support for fixed DH ciphersuite client authentication: where both
2964 server and client use DH certificates with common parameters.
2965 [Steve Henson]
2966
2967 *) Support for fixed DH ciphersuites: those requiring DH server
2968 certificates.
2969 [Steve Henson]
2970
2971 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
2972 the certificate.
2973 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
2974 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
2975 X509_CINF_get_signature were reverted post internal team review.
2976
2977 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
2978
2979 *) Build fixes for the Windows and OpenVMS platforms
2980 [Matt Caswell and Richard Levitte]
2981
2982 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
2983
2984 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
2985 message can cause a segmentation fault in OpenSSL due to a NULL pointer
2986 dereference. This could lead to a Denial Of Service attack. Thanks to
2987 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
2988 (CVE-2014-3571)
2989 [Steve Henson]
2990
2991 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
2992 dtls1_buffer_record function under certain conditions. In particular this
2993 could occur if an attacker sent repeated DTLS records with the same
2994 sequence number but for the next epoch. The memory leak could be exploited
2995 by an attacker in a Denial of Service attack through memory exhaustion.
2996 Thanks to Chris Mueller for reporting this issue.
2997 (CVE-2015-0206)
2998 [Matt Caswell]
2999
3000 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
3001 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
3002 method would be set to NULL which could later result in a NULL pointer
3003 dereference. Thanks to Frank Schmirler for reporting this issue.
3004 (CVE-2014-3569)
3005 [Kurt Roeckx]
3006
3007 *) Abort handshake if server key exchange message is omitted for ephemeral
3008 ECDH ciphersuites.
3009
3010 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
3011 reporting this issue.
3012 (CVE-2014-3572)
3013 [Steve Henson]
3014
3015 *) Remove non-export ephemeral RSA code on client and server. This code
3016 violated the TLS standard by allowing the use of temporary RSA keys in
3017 non-export ciphersuites and could be used by a server to effectively
3018 downgrade the RSA key length used to a value smaller than the server
3019 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
3020 INRIA or reporting this issue.
3021 (CVE-2015-0204)
3022 [Steve Henson]
3023
3024 *) Fixed issue where DH client certificates are accepted without verification.
3025 An OpenSSL server will accept a DH certificate for client authentication
3026 without the certificate verify message. This effectively allows a client to
3027 authenticate without the use of a private key. This only affects servers
3028 which trust a client certificate authority which issues certificates
3029 containing DH keys: these are extremely rare and hardly ever encountered.
3030 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
3031 this issue.
3032 (CVE-2015-0205)
3033 [Steve Henson]
3034
3035 *) Ensure that the session ID context of an SSL is updated when its
3036 SSL_CTX is updated via SSL_set_SSL_CTX.
3037
3038 The session ID context is typically set from the parent SSL_CTX,
3039 and can vary with the CTX.
3040 [Adam Langley]
3041
3042 *) Fix various certificate fingerprint issues.
3043
3044 By using non-DER or invalid encodings outside the signed portion of a
3045 certificate the fingerprint can be changed without breaking the signature.
3046 Although no details of the signed portion of the certificate can be changed
3047 this can cause problems with some applications: e.g. those using the
3048 certificate fingerprint for blacklists.
3049
3050 1. Reject signatures with non zero unused bits.
3051
3052 If the BIT STRING containing the signature has non zero unused bits reject
3053 the signature. All current signature algorithms require zero unused bits.
3054
3055 2. Check certificate algorithm consistency.
3056
3057 Check the AlgorithmIdentifier inside TBS matches the one in the
3058 certificate signature. NB: this will result in signature failure
3059 errors for some broken certificates.
3060
3061 Thanks to Konrad Kraszewski from Google for reporting this issue.
3062
3063 3. Check DSA/ECDSA signatures use DER.
3064
3065 Re-encode DSA/ECDSA signatures and compare with the original received
3066 signature. Return an error if there is a mismatch.
3067
3068 This will reject various cases including garbage after signature
3069 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
3070 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
3071 (negative or with leading zeroes).
3072
3073 Further analysis was conducted and fixes were developed by Stephen Henson
3074 of the OpenSSL core team.
3075
3076 (CVE-2014-8275)
3077 [Steve Henson]
3078
3079 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
3080 results on some platforms, including x86_64. This bug occurs at random
3081 with a very low probability, and is not known to be exploitable in any
3082 way, though its exact impact is difficult to determine. Thanks to Pieter
3083 Wuille (Blockstream) who reported this issue and also suggested an initial
3084 fix. Further analysis was conducted by the OpenSSL development team and
3085 Adam Langley of Google. The final fix was developed by Andy Polyakov of
3086 the OpenSSL core team.
3087 (CVE-2014-3570)
3088 [Andy Polyakov]
3089
3090 *) Do not resume sessions on the server if the negotiated protocol
3091 version does not match the session's version. Resuming with a different
3092 version, while not strictly forbidden by the RFC, is of questionable
3093 sanity and breaks all known clients.
3094 [David Benjamin, Emilia Käsper]
3095
3096 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
3097 early CCS messages during renegotiation. (Note that because
3098 renegotiation is encrypted, this early CCS was not exploitable.)
3099 [Emilia Käsper]
3100
3101 *) Tighten client-side session ticket handling during renegotiation:
3102 ensure that the client only accepts a session ticket if the server sends
3103 the extension anew in the ServerHello. Previously, a TLS client would
3104 reuse the old extension state and thus accept a session ticket if one was
3105 announced in the initial ServerHello.
3106
3107 Similarly, ensure that the client requires a session ticket if one
3108 was advertised in the ServerHello. Previously, a TLS client would
3109 ignore a missing NewSessionTicket message.
3110 [Emilia Käsper]
3111
3112 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
3113
3114 *) SRTP Memory Leak.
3115
3116 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
3117 sends a carefully crafted handshake message, to cause OpenSSL to fail
3118 to free up to 64k of memory causing a memory leak. This could be
3119 exploited in a Denial Of Service attack. This issue affects OpenSSL
3120 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
3121 whether SRTP is used or configured. Implementations of OpenSSL that
3122 have been compiled with OPENSSL_NO_SRTP defined are not affected.
3123
3124 The fix was developed by the OpenSSL team.
3125 (CVE-2014-3513)
3126 [OpenSSL team]
3127
3128 *) Session Ticket Memory Leak.
3129
3130 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
3131 integrity of that ticket is first verified. In the event of a session
3132 ticket integrity check failing, OpenSSL will fail to free memory
3133 causing a memory leak. By sending a large number of invalid session
3134 tickets an attacker could exploit this issue in a Denial Of Service
3135 attack.
3136 (CVE-2014-3567)
3137 [Steve Henson]
3138
3139 *) Build option no-ssl3 is incomplete.
3140
3141 When OpenSSL is configured with "no-ssl3" as a build option, servers
3142 could accept and complete a SSL 3.0 handshake, and clients could be
3143 configured to send them.
3144 (CVE-2014-3568)
3145 [Akamai and the OpenSSL team]
3146
3147 *) Add support for TLS_FALLBACK_SCSV.
3148 Client applications doing fallback retries should call
3149 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3150 (CVE-2014-3566)
3151 [Adam Langley, Bodo Moeller]
3152
3153 *) Add additional DigestInfo checks.
3154
3155 Re-encode DigestInto in DER and check against the original when
3156 verifying RSA signature: this will reject any improperly encoded
3157 DigestInfo structures.
3158
3159 Note: this is a precautionary measure and no attacks are currently known.
3160
3161 [Steve Henson]
3162
3163 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3164
3165 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3166 SRP code can be overrun an internal buffer. Add sanity check that
3167 g, A, B < N to SRP code.
3168
3169 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3170 Group for discovering this issue.
3171 (CVE-2014-3512)
3172 [Steve Henson]
3173
3174 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3175 TLS 1.0 instead of higher protocol versions when the ClientHello message
3176 is badly fragmented. This allows a man-in-the-middle attacker to force a
3177 downgrade to TLS 1.0 even if both the server and the client support a
3178 higher protocol version, by modifying the client's TLS records.
3179
3180 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3181 researching this issue.
3182 (CVE-2014-3511)
3183 [David Benjamin]
3184
3185 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3186 to a denial of service attack. A malicious server can crash the client
3187 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3188 ciphersuite and sending carefully crafted handshake messages.
3189
3190 Thanks to Felix Gröbert (Google) for discovering and researching this
3191 issue.
3192 (CVE-2014-3510)
3193 [Emilia Käsper]
3194
3195 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3196 to leak memory. This can be exploited through a Denial of Service attack.
3197 Thanks to Adam Langley for discovering and researching this issue.
3198 (CVE-2014-3507)
3199 [Adam Langley]
3200
3201 *) An attacker can force openssl to consume large amounts of memory whilst
3202 processing DTLS handshake messages. This can be exploited through a
3203 Denial of Service attack.
3204 Thanks to Adam Langley for discovering and researching this issue.
3205 (CVE-2014-3506)
3206 [Adam Langley]
3207
3208 *) An attacker can force an error condition which causes openssl to crash
3209 whilst processing DTLS packets due to memory being freed twice. This
3210 can be exploited through a Denial of Service attack.
3211 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3212 this issue.
3213 (CVE-2014-3505)
3214 [Adam Langley]
3215
3216 *) If a multithreaded client connects to a malicious server using a resumed
3217 session and the server sends an ec point format extension it could write
3218 up to 255 bytes to freed memory.
3219
3220 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3221 issue.
3222 (CVE-2014-3509)
3223 [Gabor Tyukasz]
3224
3225 *) A malicious server can crash an OpenSSL client with a null pointer
3226 dereference (read) by specifying an SRP ciphersuite even though it was not
3227 properly negotiated with the client. This can be exploited through a
3228 Denial of Service attack.
3229
3230 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3231 discovering and researching this issue.
3232 (CVE-2014-5139)
3233 [Steve Henson]
3234
3235 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3236 X509_name_oneline, X509_name_print_ex et al. to leak some information
3237 from the stack. Applications may be affected if they echo pretty printing
3238 output to the attacker.
3239
3240 Thanks to Ivan Fratric (Google) for discovering this issue.
3241 (CVE-2014-3508)
3242 [Emilia Käsper, and Steve Henson]
3243
3244 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3245 for corner cases. (Certain input points at infinity could lead to
3246 bogus results, with non-infinity inputs mapped to infinity too.)
3247 [Bodo Moeller]
3248
3249 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3250
3251 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3252 handshake can force the use of weak keying material in OpenSSL
3253 SSL/TLS clients and servers.
3254
3255 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3256 researching this issue. (CVE-2014-0224)
3257 [KIKUCHI Masashi, Steve Henson]
3258
3259 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3260 OpenSSL DTLS client the code can be made to recurse eventually crashing
3261 in a DoS attack.
3262
3263 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3264 (CVE-2014-0221)
3265 [Imre Rad, Steve Henson]
3266
3267 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3268 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3269 client or server. This is potentially exploitable to run arbitrary
3270 code on a vulnerable client or server.
3271
3272 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3273 [Jüri Aedla, Steve Henson]
3274
3275 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3276 are subject to a denial of service attack.
3277
3278 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3279 this issue. (CVE-2014-3470)
3280 [Felix Gröbert, Ivan Fratric, Steve Henson]
3281
3282 *) Harmonize version and its documentation. -f flag is used to display
3283 compilation flags.
3284 [mancha <mancha1@zoho.com>]
3285
3286 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3287 in i2d_ECPrivateKey.
3288 [mancha <mancha1@zoho.com>]
3289
3290 *) Fix some double frees. These are not thought to be exploitable.
3291 [mancha <mancha1@zoho.com>]
3292
3293 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3294
3295 *) A missing bounds check in the handling of the TLS heartbeat extension
3296 can be used to reveal up to 64k of memory to a connected client or
3297 server.
3298
3299 Thanks for Neel Mehta of Google Security for discovering this bug and to
3300 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3301 preparing the fix (CVE-2014-0160)
3302 [Adam Langley, Bodo Moeller]
3303
3304 *) Fix for the attack described in the paper "Recovering OpenSSL
3305 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3306 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3307 http://eprint.iacr.org/2014/140
3308
3309 Thanks to Yuval Yarom and Naomi Benger for discovering this
3310 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3311 [Yuval Yarom and Naomi Benger]
3312
3313 *) TLS pad extension: draft-agl-tls-padding-03
3314
3315 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3316 TLS client Hello record length value would otherwise be > 255 and
3317 less that 512 pad with a dummy extension containing zeroes so it
3318 is at least 512 bytes long.
3319
3320 [Adam Langley, Steve Henson]
3321
3322 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3323
3324 *) Fix for TLS record tampering bug. A carefully crafted invalid
3325 handshake could crash OpenSSL with a NULL pointer exception.
3326 Thanks to Anton Johansson for reporting this issues.
3327 (CVE-2013-4353)
3328
3329 *) Keep original DTLS digest and encryption contexts in retransmission
3330 structures so we can use the previous session parameters if they need
3331 to be resent. (CVE-2013-6450)
3332 [Steve Henson]
3333
3334 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3335 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3336 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3337 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3338 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3339 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3340 [Rob Stradling, Adam Langley]
3341
3342 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3343
3344 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3345 supporting platforms or when small records were transferred.
3346 [Andy Polyakov, Steve Henson]
3347
3348 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3349
3350 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3351
3352 This addresses the flaw in CBC record processing discovered by
3353 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3354 at: http://www.isg.rhul.ac.uk/tls/
3355
3356 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3357 Security Group at Royal Holloway, University of London
3358 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3359 Emilia Käsper for the initial patch.
3360 (CVE-2013-0169)
3361 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3362
3363 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3364 ciphersuites which can be exploited in a denial of service attack.
3365 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3366 and detecting this bug and to Wolfgang Ettlinger
3367 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3368 (CVE-2012-2686)
3369 [Adam Langley]
3370
3371 *) Return an error when checking OCSP signatures when key is NULL.
3372 This fixes a DoS attack. (CVE-2013-0166)
3373 [Steve Henson]
3374
3375 *) Make openssl verify return errors.
3376 [Chris Palmer <palmer@google.com> and Ben Laurie]
3377
3378 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3379 the right response is stapled. Also change SSL_get_certificate()
3380 so it returns the certificate actually sent.
3381 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3382 [Rob Stradling <rob.stradling@comodo.com>]
3383
3384 *) Fix possible deadlock when decoding public keys.
3385 [Steve Henson]
3386
3387 *) Don't use TLS 1.0 record version number in initial client hello
3388 if renegotiating.
3389 [Steve Henson]
3390
3391 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3392
3393 *) Sanity check record length before skipping explicit IV in TLS
3394 1.2, 1.1 and DTLS to fix DoS attack.
3395
3396 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3397 fuzzing as a service testing platform.
3398 (CVE-2012-2333)
3399 [Steve Henson]
3400
3401 *) Initialise tkeylen properly when encrypting CMS messages.
3402 Thanks to Solar Designer of Openwall for reporting this issue.
3403 [Steve Henson]
3404
3405 *) In FIPS mode don't try to use composite ciphers as they are not
3406 approved.
3407 [Steve Henson]
3408
3409 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3410
3411 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3412 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3413 mean any application compiled against OpenSSL 1.0.0 headers setting
3414 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3415 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3416 0x10000000L Any application which was previously compiled against
3417 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3418 will need to be recompiled as a result. Letting be results in
3419 inability to disable specifically TLS 1.1 and in client context,
3420 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3421 [Steve Henson]
3422
3423 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3424 disable just protocol X, but all protocols above X *if* there are
3425 protocols *below* X still enabled. In more practical terms it means
3426 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3427 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3428 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3429 client side.
3430 [Andy Polyakov]
3431
3432 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3433
3434 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3435 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3436 in CRYPTO_realloc_clean.
3437
3438 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3439 issue and to Adam Langley <agl@chromium.org> for fixing it.
3440 (CVE-2012-2110)
3441 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3442
3443 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3444 [Adam Langley]
3445
3446 *) Workarounds for some broken servers that "hang" if a client hello
3447 record length exceeds 255 bytes.
3448
3449 1. Do not use record version number > TLS 1.0 in initial client
3450 hello: some (but not all) hanging servers will now work.
3451 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3452 the number of ciphers sent in the client hello. This should be
3453 set to an even number, such as 50, for example by passing:
3454 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3455 Most broken servers should now work.
3456 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3457 TLS 1.2 client support entirely.
3458 [Steve Henson]
3459
3460 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3461 [Andy Polyakov]
3462
3463 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3464
3465 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3466 STRING form instead of a DigestInfo.
3467 [Steve Henson]
3468
3469 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3470 and the RSA_sign/RSA_verify functions. This was made more apparent when
3471 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3472 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3473 the correct format in RSA_verify so both forms transparently work.
3474 [Steve Henson]
3475
3476 *) Some servers which support TLS 1.0 can choke if we initially indicate
3477 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3478 encrypted premaster secret. As a workaround use the maximum permitted
3479 client version in client hello, this should keep such servers happy
3480 and still work with previous versions of OpenSSL.
3481 [Steve Henson]
3482
3483 *) Add support for TLS/DTLS heartbeats.
3484 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3485
3486 *) Add support for SCTP.
3487 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3488
3489 *) Improved PRNG seeding for VOS.
3490 [Paul Green <Paul.Green@stratus.com>]
3491
3492 *) Extensive assembler packs updates, most notably:
3493
3494 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3495 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3496 - x86_64: bit-sliced AES implementation;
3497 - ARM: NEON support, contemporary platforms optimizations;
3498 - s390x: z196 support;
3499 - *: GHASH and GF(2^m) multiplication implementations;
3500
3501 [Andy Polyakov]
3502
3503 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3504 (removal of unnecessary code)
3505 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3506
3507 *) Add TLS key material exporter from RFC 5705.
3508 [Eric Rescorla]
3509
3510 *) Add DTLS-SRTP negotiation from RFC 5764.
3511 [Eric Rescorla]
3512
3513 *) Add Next Protocol Negotiation,
3514 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3515 disabled with a no-npn flag to config or Configure. Code donated
3516 by Google.
3517 [Adam Langley <agl@google.com> and Ben Laurie]
3518
3519 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3520 NIST-P256, NIST-P521, with constant-time single point multiplication on
3521 typical inputs. Compiler support for the nonstandard type __uint128_t is
3522 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3523 Code made available under Apache License version 2.0.
3524
3525 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3526 line to include this in your build of OpenSSL, and run "make depend" (or
3527 "make update"). This enables the following EC_METHODs:
3528
3529 EC_GFp_nistp224_method()
3530 EC_GFp_nistp256_method()
3531 EC_GFp_nistp521_method()
3532
3533 EC_GROUP_new_by_curve_name() will automatically use these (while
3534 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3535 implementations).
3536 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3537
3538 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3539 all platforms. Move ssize_t definition from e_os.h to the public
3540 header file e_os2.h as it now appears in public header file cms.h
3541 [Steve Henson]
3542
3543 *) New -sigopt option to the ca, req and x509 utilities. Additional
3544 signature parameters can be passed using this option and in
3545 particular PSS.
3546 [Steve Henson]
3547
3548 *) Add RSA PSS signing function. This will generate and set the
3549 appropriate AlgorithmIdentifiers for PSS based on those in the
3550 corresponding EVP_MD_CTX structure. No application support yet.
3551 [Steve Henson]
3552
3553 *) Support for companion algorithm specific ASN1 signing routines.
3554 New function ASN1_item_sign_ctx() signs a pre-initialised
3555 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3556 the appropriate parameters.
3557 [Steve Henson]
3558
3559 *) Add new algorithm specific ASN1 verification initialisation function
3560 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3561 handling will be the same no matter what EVP_PKEY_METHOD is used.
3562 Add a PSS handler to support verification of PSS signatures: checked
3563 against a number of sample certificates.
3564 [Steve Henson]
3565
3566 *) Add signature printing for PSS. Add PSS OIDs.
3567 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3568
3569 *) Add algorithm specific signature printing. An individual ASN1 method
3570 can now print out signatures instead of the standard hex dump.
3571
3572 More complex signatures (e.g. PSS) can print out more meaningful
3573 information. Include DSA version that prints out the signature
3574 parameters r, s.
3575 [Steve Henson]
3576
3577 *) Password based recipient info support for CMS library: implementing
3578 RFC3211.
3579 [Steve Henson]
3580
3581 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3582 neatly separates the code into cipher and PBE sections and is required
3583 for some algorithms that split PBES2 into separate pieces (such as
3584 password based CMS).
3585 [Steve Henson]
3586
3587 *) Session-handling fixes:
3588 - Fix handling of connections that are resuming with a session ID,
3589 but also support Session Tickets.
3590 - Fix a bug that suppressed issuing of a new ticket if the client
3591 presented a ticket with an expired session.
3592 - Try to set the ticket lifetime hint to something reasonable.
3593 - Make tickets shorter by excluding irrelevant information.
3594 - On the client side, don't ignore renewed tickets.
3595 [Adam Langley, Bodo Moeller (Google)]
3596
3597 *) Fix PSK session representation.
3598 [Bodo Moeller]
3599
3600 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3601
3602 This work was sponsored by Intel.
3603 [Andy Polyakov]
3604
3605 *) Add GCM support to TLS library. Some custom code is needed to split
3606 the IV between the fixed (from PRF) and explicit (from TLS record)
3607 portions. This adds all GCM ciphersuites supported by RFC5288 and
3608 RFC5289. Generalise some AES* cipherstrings to include GCM and
3609 add a special AESGCM string for GCM only.
3610 [Steve Henson]
3611
3612 *) Expand range of ctrls for AES GCM. Permit setting invocation
3613 field on decrypt and retrieval of invocation field only on encrypt.
3614 [Steve Henson]
3615
3616 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3617 As required by RFC5289 these ciphersuites cannot be used if for
3618 versions of TLS earlier than 1.2.
3619 [Steve Henson]
3620
3621 *) For FIPS capable OpenSSL interpret a NULL default public key method
3622 as unset and return the appropriate default but do *not* set the default.
3623 This means we can return the appropriate method in applications that
3624 switch between FIPS and non-FIPS modes.
3625 [Steve Henson]
3626
3627 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3628 ENGINE is used then we cannot handle that in the FIPS module so we
3629 keep original code iff non-FIPS operations are allowed.
3630 [Steve Henson]
3631
3632 *) Add -attime option to openssl utilities.
3633 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3634
3635 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3636 [Steve Henson]
3637
3638 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3639 FIPS EC methods unconditionally for now.
3640 [Steve Henson]
3641
3642 *) New build option no-ec2m to disable characteristic 2 code.
3643 [Steve Henson]
3644
3645 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3646 all cases can be covered as some introduce binary incompatibilities.
3647 [Steve Henson]
3648
3649 *) Redirect RSA operations to FIPS module including keygen,
3650 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3651 [Steve Henson]
3652
3653 *) Add similar low level API blocking to ciphers.
3654 [Steve Henson]
3655
3656 *) Low level digest APIs are not approved in FIPS mode: any attempt
3657 to use these will cause a fatal error. Applications that *really* want
3658 to use them can use the private_* version instead.
3659 [Steve Henson]
3660
3661 *) Redirect cipher operations to FIPS module for FIPS builds.
3662 [Steve Henson]
3663
3664 *) Redirect digest operations to FIPS module for FIPS builds.
3665 [Steve Henson]
3666
3667 *) Update build system to add "fips" flag which will link in fipscanister.o
3668 for static and shared library builds embedding a signature if needed.
3669 [Steve Henson]
3670
3671 *) Output TLS supported curves in preference order instead of numerical
3672 order. This is currently hardcoded for the highest order curves first.
3673 This should be configurable so applications can judge speed vs strength.
3674 [Steve Henson]
3675
3676 *) Add TLS v1.2 server support for client authentication.
3677 [Steve Henson]
3678
3679 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3680 and enable MD5.
3681 [Steve Henson]
3682
3683 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3684 FIPS modules versions.
3685 [Steve Henson]
3686
3687 *) Add TLS v1.2 client side support for client authentication. Keep cache
3688 of handshake records longer as we don't know the hash algorithm to use
3689 until after the certificate request message is received.
3690 [Steve Henson]
3691
3692 *) Initial TLS v1.2 client support. Add a default signature algorithms
3693 extension including all the algorithms we support. Parse new signature
3694 format in client key exchange. Relax some ECC signing restrictions for
3695 TLS v1.2 as indicated in RFC5246.
3696 [Steve Henson]
3697
3698 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3699 to new signature format when needed using client digest preference.
3700 All server ciphersuites should now work correctly in TLS v1.2. No client
3701 support yet and no support for client certificates.
3702 [Steve Henson]
3703
3704 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3705 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3706 ciphersuites. At present only RSA key exchange ciphersuites work with
3707 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3708 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3709 and version checking.
3710 [Steve Henson]
3711
3712 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3713 with this defined it will not be affected by any changes to ssl internal
3714 structures. Add several utility functions to allow openssl application
3715 to work with OPENSSL_NO_SSL_INTERN defined.
3716 [Steve Henson]
3717
3718 *) A long standing patch to add support for SRP from EdelWeb (Peter
3719 Sylvester and Christophe Renou) was integrated.
3720 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3721 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3722 Ben Laurie]
3723
3724 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3725 [Steve Henson]
3726
3727 *) Permit abbreviated handshakes when renegotiating using the function
3728 SSL_renegotiate_abbreviated().
3729 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3730
3731 *) Add call to ENGINE_register_all_complete() to
3732 ENGINE_load_builtin_engines(), so some implementations get used
3733 automatically instead of needing explicit application support.
3734 [Steve Henson]
3735
3736 *) Add support for TLS key exporter as described in RFC5705.
3737 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3738
3739 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3740 a few changes are required:
3741
3742 Add SSL_OP_NO_TLSv1_1 flag.
3743 Add TLSv1_1 methods.
3744 Update version checking logic to handle version 1.1.
3745 Add explicit IV handling (ported from DTLS code).
3746 Add command line options to s_client/s_server.
3747 [Steve Henson]
3748
3749 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3750
3751 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3752 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3753 content decryption and always return the same error. Note: this attack
3754 needs on average 2^20 messages so it only affects automated senders. The
3755 old behaviour can be re-enabled in the CMS code by setting the
3756 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3757 an MMA defence is not necessary.
3758 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3759 this issue. (CVE-2012-0884)
3760 [Steve Henson]
3761
3762 *) Fix CVE-2011-4619: make sure we really are receiving a
3763 client hello before rejecting multiple SGC restarts. Thanks to
3764 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3765 [Steve Henson]
3766
3767 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3768
3769 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3770 Thanks to Antonio Martin, Enterprise Secure Access Research and
3771 Development, Cisco Systems, Inc. for discovering this bug and
3772 preparing a fix. (CVE-2012-0050)
3773 [Antonio Martin]
3774
3775 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3776
3777 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3778 of the Vaudenay padding oracle attack on CBC mode encryption
3779 which enables an efficient plaintext recovery attack against
3780 the OpenSSL implementation of DTLS. Their attack exploits timing
3781 differences arising during decryption processing. A research
3782 paper describing this attack can be found at:
3783 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3784 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3785 Security Group at Royal Holloway, University of London
3786 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3787 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3788 for preparing the fix. (CVE-2011-4108)
3789 [Robin Seggelmann, Michael Tuexen]
3790
3791 *) Clear bytes used for block padding of SSL 3.0 records.
3792 (CVE-2011-4576)
3793 [Adam Langley (Google)]
3794
3795 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3796 Kadianakis <desnacked@gmail.com> for discovering this issue and
3797 Adam Langley for preparing the fix. (CVE-2011-4619)
3798 [Adam Langley (Google)]
3799
3800 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3801 [Andrey Kulikov <amdeich@gmail.com>]
3802
3803 *) Prevent malformed RFC3779 data triggering an assertion failure.
3804 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3805 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3806 [Rob Austein <sra@hactrn.net>]
3807
3808 *) Improved PRNG seeding for VOS.
3809 [Paul Green <Paul.Green@stratus.com>]
3810
3811 *) Fix ssl_ciph.c set-up race.
3812 [Adam Langley (Google)]
3813
3814 *) Fix spurious failures in ecdsatest.c.
3815 [Emilia Käsper (Google)]
3816
3817 *) Fix the BIO_f_buffer() implementation (which was mixing different
3818 interpretations of the '..._len' fields).
3819 [Adam Langley (Google)]
3820
3821 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3822 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3823 threads won't reuse the same blinding coefficients.
3824
3825 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3826 lock to call BN_BLINDING_invert_ex, and avoids one use of
3827 BN_BLINDING_update for each BN_BLINDING structure (previously,
3828 the last update always remained unused).
3829 [Emilia Käsper (Google)]
3830
3831 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3832 [Bob Buckholz (Google)]
3833
3834 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3835
3836 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3837 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3838 [Kaspar Brand <ossl@velox.ch>]
3839
3840 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3841 for multi-threaded use of ECDH. (CVE-2011-3210)
3842 [Adam Langley (Google)]
3843
3844 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3845 [Bodo Moeller]
3846
3847 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3848 signature public key algorithm by using OID xref utilities instead.
3849 Before this you could only use some ECC ciphersuites with SHA1 only.
3850 [Steve Henson]
3851
3852 *) Add protection against ECDSA timing attacks as mentioned in the paper
3853 by Billy Bob Brumley and Nicola Tuveri, see:
3854
3855 http://eprint.iacr.org/2011/232.pdf
3856
3857 [Billy Bob Brumley and Nicola Tuveri]
3858
3859 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3860
3861 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3862 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3863
3864 *) Fix bug in string printing code: if *any* escaping is enabled we must
3865 escape the escape character (backslash) or the resulting string is
3866 ambiguous.
3867 [Steve Henson]
3868
3869 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3870
3871 *) Disable code workaround for ancient and obsolete Netscape browsers
3872 and servers: an attacker can use it in a ciphersuite downgrade attack.
3873 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3874 [Steve Henson]
3875
3876 *) Fixed J-PAKE implementation error, originally discovered by
3877 Sebastien Martini, further info and confirmation from Stefan
3878 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3879 [Ben Laurie]
3880
3881 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3882
3883 *) Fix extension code to avoid race conditions which can result in a buffer
3884 overrun vulnerability: resumed sessions must not be modified as they can
3885 be shared by multiple threads. CVE-2010-3864
3886 [Steve Henson]
3887
3888 *) Fix WIN32 build system to correctly link an ENGINE directory into
3889 a DLL.
3890 [Steve Henson]
3891
3892 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3893
3894 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3895 (CVE-2010-1633)
3896 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3897
3898 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3899
3900 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3901 context. The operation can be customised via the ctrl mechanism in
3902 case ENGINEs want to include additional functionality.
3903 [Steve Henson]
3904
3905 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3906 [Steve Henson]
3907
3908 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3909 output hashes compatible with older versions of OpenSSL.
3910 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3911
3912 *) Fix compression algorithm handling: if resuming a session use the
3913 compression algorithm of the resumed session instead of determining
3914 it from client hello again. Don't allow server to change algorithm.
3915 [Steve Henson]
3916
3917 *) Add load_crls() function to apps tidying load_certs() too. Add option
3918 to verify utility to allow additional CRLs to be included.
3919 [Steve Henson]
3920
3921 *) Update OCSP request code to permit adding custom headers to the request:
3922 some responders need this.
3923 [Steve Henson]
3924
3925 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3926 correctly.
3927 [Julia Lawall <julia@diku.dk>]
3928
3929 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3930 needlessly dereferenced structures, used obsolete functions and
3931 didn't handle all updated verify codes correctly.
3932 [Steve Henson]
3933
3934 *) Disable MD2 in the default configuration.
3935 [Steve Henson]
3936
3937 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3938 indicate the initial BIO being pushed or popped. This makes it possible
3939 to determine whether the BIO is the one explicitly called or as a result
3940 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3941 it handles reference counts correctly and doesn't zero out the I/O bio
3942 when it is not being explicitly popped. WARNING: applications which
3943 included workarounds for the old buggy behaviour will need to be modified
3944 or they could free up already freed BIOs.
3945 [Steve Henson]
3946
3947 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
3948 renaming to all platforms (within the 0.9.8 branch, this was
3949 done conditionally on Netware platforms to avoid a name clash).
3950 [Guenter <lists@gknw.net>]
3951
3952 *) Add ECDHE and PSK support to DTLS.
3953 [Michael Tuexen <tuexen@fh-muenster.de>]
3954
3955 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
3956 be used on C++.
3957 [Steve Henson]
3958
3959 *) Add "missing" function EVP_MD_flags() (without this the only way to
3960 retrieve a digest flags is by accessing the structure directly. Update
3961 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
3962 or cipher is registered as in the "from" argument. Print out all
3963 registered digests in the dgst usage message instead of manually
3964 attempting to work them out.
3965 [Steve Henson]
3966
3967 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
3968 this allows the use of compression and extensions. Change default cipher
3969 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
3970 by default unless an application cipher string requests it.
3971 [Steve Henson]
3972
3973 *) Alter match criteria in PKCS12_parse(). It used to try to use local
3974 key ids to find matching certificates and keys but some PKCS#12 files
3975 don't follow the (somewhat unwritten) rules and this strategy fails.
3976 Now just gather all certificates together and the first private key
3977 then look for the first certificate that matches the key.
3978 [Steve Henson]
3979
3980 *) Support use of registered digest and cipher names for dgst and cipher
3981 commands instead of having to add each one as a special case. So now
3982 you can do:
3983
3984 openssl sha256 foo
3985
3986 as well as:
3987
3988 openssl dgst -sha256 foo
3989
3990 and this works for ENGINE based algorithms too.
3991
3992 [Steve Henson]
3993
3994 *) Update Gost ENGINE to support parameter files.
3995 [Victor B. Wagner <vitus@cryptocom.ru>]
3996
3997 *) Support GeneralizedTime in ca utility.
3998 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
3999
4000 *) Enhance the hash format used for certificate directory links. The new
4001 form uses the canonical encoding (meaning equivalent names will work
4002 even if they aren't identical) and uses SHA1 instead of MD5. This form
4003 is incompatible with the older format and as a result c_rehash should
4004 be used to rebuild symbolic links.
4005 [Steve Henson]
4006
4007 *) Make PKCS#8 the default write format for private keys, replacing the
4008 traditional format. This form is standardised, more secure and doesn't
4009 include an implicit MD5 dependency.
4010 [Steve Henson]
4011
4012 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
4013 committed to OpenSSL should pass this lot as a minimum.
4014 [Steve Henson]
4015
4016 *) Add session ticket override functionality for use by EAP-FAST.
4017 [Jouni Malinen <j@w1.fi>]
4018
4019 *) Modify HMAC functions to return a value. Since these can be implemented
4020 in an ENGINE errors can occur.
4021 [Steve Henson]
4022
4023 *) Type-checked OBJ_bsearch_ex.
4024 [Ben Laurie]
4025
4026 *) Type-checked OBJ_bsearch. Also some constification necessitated
4027 by type-checking. Still to come: TXT_DB, bsearch(?),
4028 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
4029 CONF_VALUE.
4030 [Ben Laurie]
4031
4032 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
4033 seconds to a tm structure directly, instead of going through OS
4034 specific date routines. This avoids any issues with OS routines such
4035 as the year 2038 bug. New *_adj() functions for ASN1 time structures
4036 and X509_time_adj_ex() to cover the extended range. The existing
4037 X509_time_adj() is still usable and will no longer have any date issues.
4038 [Steve Henson]
4039
4040 *) Delta CRL support. New use deltas option which will attempt to locate
4041 and search any appropriate delta CRLs available.
4042
4043 This work was sponsored by Google.
4044 [Steve Henson]
4045
4046 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
4047 code and add additional score elements. Validate alternate CRL paths
4048 as part of the CRL checking and indicate a new error "CRL path validation
4049 error" in this case. Applications wanting additional details can use
4050 the verify callback and check the new "parent" field. If this is not
4051 NULL CRL path validation is taking place. Existing applications won't
4052 see this because it requires extended CRL support which is off by
4053 default.
4054
4055 This work was sponsored by Google.
4056 [Steve Henson]
4057
4058 *) Support for freshest CRL extension.
4059
4060 This work was sponsored by Google.
4061 [Steve Henson]
4062
4063 *) Initial indirect CRL support. Currently only supported in the CRLs
4064 passed directly and not via lookup. Process certificate issuer
4065 CRL entry extension and lookup CRL entries by bother issuer name
4066 and serial number. Check and process CRL issuer entry in IDP extension.
4067
4068 This work was sponsored by Google.
4069 [Steve Henson]
4070
4071 *) Add support for distinct certificate and CRL paths. The CRL issuer
4072 certificate is validated separately in this case. Only enabled if
4073 an extended CRL support flag is set: this flag will enable additional
4074 CRL functionality in future.
4075
4076 This work was sponsored by Google.
4077 [Steve Henson]
4078
4079 *) Add support for policy mappings extension.
4080
4081 This work was sponsored by Google.
4082 [Steve Henson]
4083
4084 *) Fixes to pathlength constraint, self issued certificate handling,
4085 policy processing to align with RFC3280 and PKITS tests.
4086
4087 This work was sponsored by Google.
4088 [Steve Henson]
4089
4090 *) Support for name constraints certificate extension. DN, email, DNS
4091 and URI types are currently supported.
4092
4093 This work was sponsored by Google.
4094 [Steve Henson]
4095
4096 *) To cater for systems that provide a pointer-based thread ID rather
4097 than numeric, deprecate the current numeric thread ID mechanism and
4098 replace it with a structure and associated callback type. This
4099 mechanism allows a numeric "hash" to be extracted from a thread ID in
4100 either case, and on platforms where pointers are larger than 'long',
4101 mixing is done to help ensure the numeric 'hash' is usable even if it
4102 can't be guaranteed unique. The default mechanism is to use "&errno"
4103 as a pointer-based thread ID to distinguish between threads.
4104
4105 Applications that want to provide their own thread IDs should now use
4106 CRYPTO_THREADID_set_callback() to register a callback that will call
4107 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
4108
4109 Note that ERR_remove_state() is now deprecated, because it is tied
4110 to the assumption that thread IDs are numeric. ERR_remove_state(0)
4111 to free the current thread's error state should be replaced by
4112 ERR_remove_thread_state(NULL).
4113
4114 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
4115 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
4116 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
4117 application was previously providing a numeric thread callback that
4118 was inappropriate for distinguishing threads, then uniqueness might
4119 have been obtained with &errno that happened immediately in the
4120 intermediate development versions of OpenSSL; this is no longer the
4121 case, the numeric thread callback will now override the automatic use
4122 of &errno.)
4123 [Geoff Thorpe, with help from Bodo Moeller]
4124
4125 *) Initial support for different CRL issuing certificates. This covers a
4126 simple case where the self issued certificates in the chain exist and
4127 the real CRL issuer is higher in the existing chain.
4128
4129 This work was sponsored by Google.
4130 [Steve Henson]
4131
4132 *) Removed effectively defunct crypto/store from the build.
4133 [Ben Laurie]
4134
4135 *) Revamp of STACK to provide stronger type-checking. Still to come:
4136 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
4137 ASN1_STRING, CONF_VALUE.
4138 [Ben Laurie]
4139
4140 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4141 RAM on SSL connections. This option can save about 34k per idle SSL.
4142 [Nick Mathewson]
4143
4144 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4145 STACK, TXT_DB, bsearch, qsort.
4146 [Ben Laurie]
4147
4148 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4149 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4150 support for data, signedData, compressedData, digestedData and
4151 encryptedData, envelopedData types included. Scripts to check against
4152 RFC4134 examples draft and interop and consistency checks of many
4153 content types and variants.
4154 [Steve Henson]
4155
4156 *) Add options to enc utility to support use of zlib compression BIO.
4157 [Steve Henson]
4158
4159 *) Extend mk1mf to support importing of options and assembly language
4160 files from Configure script, currently only included in VC-WIN32.
4161 The assembly language rules can now optionally generate the source
4162 files from the associated perl scripts.
4163 [Steve Henson]
4164
4165 *) Implement remaining functionality needed to support GOST ciphersuites.
4166 Interop testing has been performed using CryptoPro implementations.
4167 [Victor B. Wagner <vitus@cryptocom.ru>]
4168
4169 *) s390x assembler pack.
4170 [Andy Polyakov]
4171
4172 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4173 "family."
4174 [Andy Polyakov]
4175
4176 *) Implement Opaque PRF Input TLS extension as specified in
4177 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4178 official specification yet and no extension type assignment by
4179 IANA exists, this extension (for now) will have to be explicitly
4180 enabled when building OpenSSL by providing the extension number
4181 to use. For example, specify an option
4182
4183 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4184
4185 to the "config" or "Configure" script to enable the extension,
4186 assuming extension number 0x9527 (which is a completely arbitrary
4187 and unofficial assignment based on the MD5 hash of the Internet
4188 Draft). Note that by doing so, you potentially lose
4189 interoperability with other TLS implementations since these might
4190 be using the same extension number for other purposes.
4191
4192 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4193 opaque PRF input value to use in the handshake. This will create
4194 an internal copy of the length-'len' string at 'src', and will
4195 return non-zero for success.
4196
4197 To get more control and flexibility, provide a callback function
4198 by using
4199
4200 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4201 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4202
4203 where
4204
4205 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4206 void *arg;
4207
4208 Callback function 'cb' will be called in handshakes, and is
4209 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4210 Argument 'arg' is for application purposes (the value as given to
4211 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4212 be provided to the callback function). The callback function
4213 has to return non-zero to report success: usually 1 to use opaque
4214 PRF input just if possible, or 2 to enforce use of the opaque PRF
4215 input. In the latter case, the library will abort the handshake
4216 if opaque PRF input is not successfully negotiated.
4217
4218 Arguments 'peerinput' and 'len' given to the callback function
4219 will always be NULL and 0 in the case of a client. A server will
4220 see the client's opaque PRF input through these variables if
4221 available (NULL and 0 otherwise). Note that if the server
4222 provides an opaque PRF input, the length must be the same as the
4223 length of the client's opaque PRF input.
4224
4225 Note that the callback function will only be called when creating
4226 a new session (session resumption can resume whatever was
4227 previously negotiated), and will not be called in SSL 2.0
4228 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4229 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4230 for applications that need to enforce opaque PRF input.
4231
4232 [Bodo Moeller]
4233
4234 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4235 MAC.
4236
4237 [Victor B. Wagner <vitus@cryptocom.ru>]
4238
4239 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4240 RFC4507bis. The encrypted ticket format is an encrypted encoded
4241 SSL_SESSION structure, that way new session features are automatically
4242 supported.
4243
4244 If a client application caches session in an SSL_SESSION structure
4245 support is transparent because tickets are now stored in the encoded
4246 SSL_SESSION.
4247
4248 The SSL_CTX structure automatically generates keys for ticket
4249 protection in servers so again support should be possible
4250 with no application modification.
4251
4252 If a client or server wishes to disable RFC4507 support then the option
4253 SSL_OP_NO_TICKET can be set.
4254
4255 Add a TLS extension debugging callback to allow the contents of any client
4256 or server extensions to be examined.
4257
4258 This work was sponsored by Google.
4259 [Steve Henson]
4260
4261 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4262 OpenSSL should now compile cleanly on gcc 4.2
4263 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4264
4265 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4266 support including streaming MAC support: this is required for GOST
4267 ciphersuite support.
4268 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4269
4270 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4271 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4272 to output in BER and PEM format.
4273 [Steve Henson]
4274
4275 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4276 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4277 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4278 ENGINE support for HMAC keys which are unextractable. New -mac and
4279 -macopt options to dgst utility.
4280 [Steve Henson]
4281
4282 *) New option -sigopt to dgst utility. Update dgst to use
4283 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4284 alternative signing parameters such as X9.31 or PSS in the dgst
4285 utility.
4286 [Steve Henson]
4287
4288 *) Change ssl_cipher_apply_rule(), the internal function that does
4289 the work each time a ciphersuite string requests enabling
4290 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4291 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4292 the order of disabled ciphersuites such that those ciphersuites
4293 that most recently went from enabled to disabled not only stay
4294 in order with respect to each other, but also have higher priority
4295 than other disabled ciphersuites the next time ciphersuites are
4296 enabled again.
4297
4298 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4299 the same ciphersuites as with "HIGH" alone, but in a specific
4300 order where the PSK ciphersuites come first (since they are the
4301 most recently disabled ciphersuites when "HIGH" is parsed).
4302
4303 Also, change ssl_create_cipher_list() (using this new
4304 functionality) such that between otherwise identical
4305 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4306 the default order.
4307 [Bodo Moeller]
4308
4309 *) Change ssl_create_cipher_list() so that it automatically
4310 arranges the ciphersuites in reasonable order before starting
4311 to process the rule string. Thus, the definition for "DEFAULT"
4312 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4313 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4314 This makes it much easier to arrive at a reasonable default order
4315 in applications for which anonymous ciphers are OK (meaning
4316 that you can't actually use DEFAULT).
4317 [Bodo Moeller; suggested by Victor Duchovni]
4318
4319 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4320 processing) into multiple integers instead of setting
4321 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
4322 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
4323 (These masks as well as the individual bit definitions are hidden
4324 away into the non-exported interface ssl/ssl_locl.h, so this
4325 change to the definition of the SSL_CIPHER structure shouldn't
4326 affect applications.) This give us more bits for each of these
4327 categories, so there is no longer a need to coagulate AES128 and
4328 AES256 into a single algorithm bit, and to coagulate Camellia128
4329 and Camellia256 into a single algorithm bit, which has led to all
4330 kinds of kludges.
4331
4332 Thus, among other things, the kludge introduced in 0.9.7m and
4333 0.9.8e for masking out AES256 independently of AES128 or masking
4334 out Camellia256 independently of AES256 is not needed here in 0.9.9.
4335
4336 With the change, we also introduce new ciphersuite aliases that
4337 so far were missing: "AES128", "AES256", "CAMELLIA128", and
4338 "CAMELLIA256".
4339 [Bodo Moeller]
4340
4341 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
4342 Use the leftmost N bytes of the signature input if the input is
4343 larger than the prime q (with N being the size in bytes of q).
4344 [Nils Larsch]
4345
4346 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
4347 it yet and it is largely untested.
4348 [Steve Henson]
4349
4350 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
4351 [Nils Larsch]
4352
4353 *) Initial incomplete changes to avoid need for function casts in OpenSSL
4354 some compilers (gcc 4.2 and later) reject their use. Safestack is
4355 reimplemented. Update ASN1 to avoid use of legacy functions.
4356 [Steve Henson]
4357
4358 *) Win32/64 targets are linked with Winsock2.
4359 [Andy Polyakov]
4360
4361 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
4362 to external functions. This can be used to increase CRL handling
4363 efficiency especially when CRLs are very large by (for example) storing
4364 the CRL revoked certificates in a database.
4365 [Steve Henson]
4366
4367 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
4368 new CRLs added to a directory can be used. New command line option
4369 -verify_return_error to s_client and s_server. This causes real errors
4370 to be returned by the verify callback instead of carrying on no matter
4371 what. This reflects the way a "real world" verify callback would behave.
4372 [Steve Henson]
4373
4374 *) GOST engine, supporting several GOST algorithms and public key formats.
4375 Kindly donated by Cryptocom.
4376 [Cryptocom]
4377
4378 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4379 partitioned by DP are handled but no indirect CRL or reason partitioning
4380 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4381 selected via a scoring technique which handles IDP and AKID in CRLs.
4382 [Steve Henson]
4383
4384 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4385 will ultimately be used for all verify operations: this will remove the
4386 X509_STORE dependency on certificate verification and allow alternative
4387 lookup methods. X509_STORE based implementations of these two callbacks.
4388 [Steve Henson]
4389
4390 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4391 Modify get_crl() to find a valid (unexpired) CRL if possible.
4392 [Steve Henson]
4393
4394 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4395 this would be called X509_CRL_cmp() but that name is already used by
4396 a function that just compares CRL issuer names. Cache several CRL
4397 extensions in X509_CRL structure and cache CRLDP in X509.
4398 [Steve Henson]
4399
4400 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4401 this maps equivalent X509_NAME structures into a consistent structure.
4402 Name comparison can then be performed rapidly using memcmp().
4403 [Steve Henson]
4404
4405 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4406 utility.
4407 [Steve Henson]
4408
4409 *) Allow digests to supply their own micalg string for S/MIME type using
4410 the ctrl EVP_MD_CTRL_MICALG.
4411 [Steve Henson]
4412
4413 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4414 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4415 ctrl. It can then customise the structure before and/or after signing
4416 if necessary.
4417 [Steve Henson]
4418
4419 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4420 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4421 to free up any added signature OIDs.
4422 [Steve Henson]
4423
4424 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4425 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4426 digest and cipher tables. New options added to openssl utility:
4427 list-message-digest-algorithms and list-cipher-algorithms.
4428 [Steve Henson]
4429
4430 *) Change the array representation of binary polynomials: the list
4431 of degrees of non-zero coefficients is now terminated with -1.
4432 Previously it was terminated with 0, which was also part of the
4433 value; thus, the array representation was not applicable to
4434 polynomials where t^0 has coefficient zero. This change makes
4435 the array representation useful in a more general context.
4436 [Douglas Stebila]
4437
4438 *) Various modifications and fixes to SSL/TLS cipher string
4439 handling. For ECC, the code now distinguishes between fixed ECDH
4440 with RSA certificates on the one hand and with ECDSA certificates
4441 on the other hand, since these are separate ciphersuites. The
4442 unused code for Fortezza ciphersuites has been removed.
4443
4444 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4445 (not "ECDHE"). For consistency with the code for DH
4446 certificates, use of ECDH certificates is now considered ECDH
4447 authentication, not RSA or ECDSA authentication (the latter is
4448 merely the CA's signing algorithm and not actively used in the
4449 protocol).
4450
4451 The temporary ciphersuite alias "ECCdraft" is no longer
4452 available, and ECC ciphersuites are no longer excluded from "ALL"
4453 and "DEFAULT". The following aliases now exist for RFC 4492
4454 ciphersuites, most of these by analogy with the DH case:
4455
4456 kECDHr - ECDH cert, signed with RSA
4457 kECDHe - ECDH cert, signed with ECDSA
4458 kECDH - ECDH cert (signed with either RSA or ECDSA)
4459 kEECDH - ephemeral ECDH
4460 ECDH - ECDH cert or ephemeral ECDH
4461
4462 aECDH - ECDH cert
4463 aECDSA - ECDSA cert
4464 ECDSA - ECDSA cert
4465
4466 AECDH - anonymous ECDH
4467 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4468
4469 [Bodo Moeller]
4470
4471 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
4472 Use correct micalg parameters depending on digest(s) in signed message.
4473 [Steve Henson]
4474
4475 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
4476 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
4477 [Steve Henson]
4478
4479 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
4480 an engine to register a method. Add ENGINE lookups for methods and
4481 functional reference processing.
4482 [Steve Henson]
4483
4484 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
4485 EVP_{Sign,Verify}* which allow an application to customise the signature
4486 process.
4487 [Steve Henson]
4488
4489 *) New -resign option to smime utility. This adds one or more signers
4490 to an existing PKCS#7 signedData structure. Also -md option to use an
4491 alternative message digest algorithm for signing.
4492 [Steve Henson]
4493
4494 *) Tidy up PKCS#7 routines and add new functions to make it easier to
4495 create PKCS7 structures containing multiple signers. Update smime
4496 application to support multiple signers.
4497 [Steve Henson]
4498
4499 *) New -macalg option to pkcs12 utility to allow setting of an alternative
4500 digest MAC.
4501 [Steve Henson]
4502
4503 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
4504 Reorganize PBE internals to lookup from a static table using NIDs,
4505 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
4506 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
4507 PRF which will be automatically used with PBES2.
4508 [Steve Henson]
4509
4510 *) Replace the algorithm specific calls to generate keys in "req" with the
4511 new API.
4512 [Steve Henson]
4513
4514 *) Update PKCS#7 enveloped data routines to use new API. This is now
4515 supported by any public key method supporting the encrypt operation. A
4516 ctrl is added to allow the public key algorithm to examine or modify
4517 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
4518 a no op.
4519 [Steve Henson]
4520
4521 *) Add a ctrl to asn1 method to allow a public key algorithm to express
4522 a default digest type to use. In most cases this will be SHA1 but some
4523 algorithms (such as GOST) need to specify an alternative digest. The
4524 return value indicates how strong the preference is 1 means optional and
4525 2 is mandatory (that is it is the only supported type). Modify
4526 ASN1_item_sign() to accept a NULL digest argument to indicate it should
4527 use the default md. Update openssl utilities to use the default digest
4528 type for signing if it is not explicitly indicated.
4529 [Steve Henson]
4530
4531 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
4532 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
4533 signing method from the key type. This effectively removes the link
4534 between digests and public key types.
4535 [Steve Henson]
4536
4537 *) Add an OID cross reference table and utility functions. Its purpose is to
4538 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
4539 rsaEncryption. This will allow some of the algorithm specific hackery
4540 needed to use the correct OID to be removed.
4541 [Steve Henson]
4542
4543 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
4544 structures for PKCS7_sign(). They are now set up by the relevant public
4545 key ASN1 method.
4546 [Steve Henson]
4547
4548 *) Add provisional EC pkey method with support for ECDSA and ECDH.
4549 [Steve Henson]
4550
4551 *) Add support for key derivation (agreement) in the API, DH method and
4552 pkeyutl.
4553 [Steve Henson]
4554
4555 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
4556 public and private key formats. As a side effect these add additional
4557 command line functionality not previously available: DSA signatures can be
4558 generated and verified using pkeyutl and DH key support and generation in
4559 pkey, genpkey.
4560 [Steve Henson]
4561
4562 *) BeOS support.
4563 [Oliver Tappe <zooey@hirschkaefer.de>]
4564
4565 *) New make target "install_html_docs" installs HTML renditions of the
4566 manual pages.
4567 [Oliver Tappe <zooey@hirschkaefer.de>]
4568
4569 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4570 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4571 support key and parameter generation and add initial key generation
4572 functionality for RSA.
4573 [Steve Henson]
4574
4575 *) Add functions for main EVP_PKEY_method operations. The undocumented
4576 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4577 EVP_PKEY_{encrypt,decrypt}_old.
4578 [Steve Henson]
4579
4580 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4581 key API, doesn't do much yet.
4582 [Steve Henson]
4583
4584 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4585 public key algorithms. New option to openssl utility:
4586 "list-public-key-algorithms" to print out info.
4587 [Steve Henson]
4588
4589 *) Implement the Supported Elliptic Curves Extension for
4590 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4591 [Douglas Stebila]
4592
4593 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4594 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4595 [Steve Henson]
4596
4597 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4598 utilities such as rsa, dsa, dsaparam etc except they process any key
4599 type.
4600 [Steve Henson]
4601
4602 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4603 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4604 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4605 structure.
4606 [Steve Henson]
4607
4608 *) Initial support for pluggable public key ASN1.
4609 De-spaghettify the public key ASN1 handling. Move public and private
4610 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4611 algorithm specific handling to a single module within the relevant
4612 algorithm directory. Add functions to allow (near) opaque processing
4613 of public and private key structures.
4614 [Steve Henson]
4615
4616 *) Implement the Supported Point Formats Extension for
4617 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4618 [Douglas Stebila]
4619
4620 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4621 for the psk identity [hint] and the psk callback functions to the
4622 SSL_SESSION, SSL and SSL_CTX structure.
4623
4624 New ciphersuites:
4625 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4626 PSK-AES256-CBC-SHA
4627
4628 New functions:
4629 SSL_CTX_use_psk_identity_hint
4630 SSL_get_psk_identity_hint
4631 SSL_get_psk_identity
4632 SSL_use_psk_identity_hint
4633
4634 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4635
4636 *) Add RFC 3161 compliant time stamp request creation, response generation
4637 and response verification functionality.
4638 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4639
4640 *) Add initial support for TLS extensions, specifically for the server_name
4641 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4642 have new members for a host name. The SSL data structure has an
4643 additional member SSL_CTX *initial_ctx so that new sessions can be
4644 stored in that context to allow for session resumption, even after the
4645 SSL has been switched to a new SSL_CTX in reaction to a client's
4646 server_name extension.
4647
4648 New functions (subject to change):
4649
4650 SSL_get_servername()
4651 SSL_get_servername_type()
4652 SSL_set_SSL_CTX()
4653
4654 New CTRL codes and macros (subject to change):
4655
4656 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4657 - SSL_CTX_set_tlsext_servername_callback()
4658 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4659 - SSL_CTX_set_tlsext_servername_arg()
4660 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4661
4662 openssl s_client has a new '-servername ...' option.
4663
4664 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4665 '-key2 ...', '-servername_fatal' (subject to change). This allows
4666 testing the HostName extension for a specific single host name ('-cert'
4667 and '-key' remain fallbacks for handshakes without HostName
4668 negotiation). If the unrecognized_name alert has to be sent, this by
4669 default is a warning; it becomes fatal with the '-servername_fatal'
4670 option.
4671
4672 [Peter Sylvester, Remy Allais, Christophe Renou]
4673
4674 *) Whirlpool hash implementation is added.
4675 [Andy Polyakov]
4676
4677 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4678 bn(64,32). Because of instruction set limitations it doesn't have
4679 any negative impact on performance. This was done mostly in order
4680 to make it possible to share assembler modules, such as bn_mul_mont
4681 implementations, between 32- and 64-bit builds without hassle.
4682 [Andy Polyakov]
4683
4684 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4685 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4686 macro.
4687 [Bodo Moeller]
4688
4689 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4690 dedicated Montgomery multiplication procedure, is introduced.
4691 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4692 "64-bit" performance on certain 32-bit targets.
4693 [Andy Polyakov]
4694
4695 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4696 in SSL structures. New SSL ctrl to set maximum send fragment size.
4697 Save memory by setting the I/O buffer sizes dynamically instead of
4698 using the maximum available value.
4699 [Steve Henson]
4700
4701 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
4702 in addition to the text details.
4703 [Bodo Moeller]
4704
4705 *) Very, very preliminary EXPERIMENTAL support for printing of general
4706 ASN1 structures. This currently produces rather ugly output and doesn't
4707 handle several customised structures at all.
4708 [Steve Henson]
4709
4710 *) Integrated support for PVK file format and some related formats such
4711 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4712 these in the 'rsa' and 'dsa' utilities.
4713 [Steve Henson]
4714
4715 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4716 [Steve Henson]
4717
4718 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4719 place for the (very old) "NETSCAPE" format certificates which are now
4720 handled using new ASN1 code equivalents.
4721 [Steve Henson]
4722
4723 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4724 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4725 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4726 [Nils Larsch]
4727
4728 *) Modify CRL distribution points extension code to print out previously
4729 unsupported fields. Enhance extension setting code to allow setting of
4730 all fields.
4731 [Steve Henson]
4732
4733 *) Add print and set support for Issuing Distribution Point CRL extension.
4734 [Steve Henson]
4735
4736 *) Change 'Configure' script to enable Camellia by default.
4737 [NTT]
4738
4739 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4740
4741 *) When rejecting SSL/TLS records due to an incorrect version number, never
4742 update s->server with a new major version number. As of
4743 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
4744 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
4745 the previous behavior could result in a read attempt at NULL when
4746 receiving specific incorrect SSL/TLS records once record payload
4747 protection is active. (CVE-2010-0740)
4748 [Bodo Moeller, Adam Langley <agl@chromium.org>]
4749
4750 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
4751 could be crashed if the relevant tables were not present (e.g. chrooted).
4752 [Tomas Hoger <thoger@redhat.com>]
4753
4754 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
4755
4756 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
4757 [Martin Olsson, Neel Mehta]
4758
4759 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
4760 accommodate for stack sorting, always a write lock!).
4761 [Bodo Moeller]
4762
4763 *) On some versions of WIN32 Heap32Next is very slow. This can cause
4764 excessive delays in the RAND_poll(): over a minute. As a workaround
4765 include a time check in the inner Heap32Next loop too.
4766 [Steve Henson]
4767
4768 *) The code that handled flushing of data in SSL/TLS originally used the
4769 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
4770 the problem outlined in PR#1949. The fix suggested there however can
4771 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4772 of Apache). So instead simplify the code to flush unconditionally.
4773 This should be fine since flushing with no data to flush is a no op.
4774 [Steve Henson]
4775
4776 *) Handle TLS versions 2.0 and later properly and correctly use the
4777 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4778 off ancient servers have a habit of sticking around for a while...
4779 [Steve Henson]
4780
4781 *) Modify compression code so it frees up structures without using the
4782 ex_data callbacks. This works around a problem where some applications
4783 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4784 restarting) then use compression (e.g. SSL with compression) later.
4785 This results in significant per-connection memory leaks and
4786 has caused some security issues including CVE-2008-1678 and
4787 CVE-2009-4355.
4788 [Steve Henson]
4789
4790 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4791 change when encrypting or decrypting.
4792 [Bodo Moeller]
4793
4794 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4795 connect and renegotiate with servers which do not support RI.
4796 Until RI is more widely deployed this option is enabled by default.
4797 [Steve Henson]
4798
4799 *) Add "missing" ssl ctrls to clear options and mode.
4800 [Steve Henson]
4801
4802 *) If client attempts to renegotiate and doesn't support RI respond with
4803 a no_renegotiation alert as required by RFC5746. Some renegotiating
4804 TLS clients will continue a connection gracefully when they receive
4805 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4806 waiting for a server hello which it will never receive. Now we treat a
4807 received no_renegotiation alert as a fatal error. This is because
4808 applications requesting a renegotiation might well expect it to succeed
4809 and would have no code in place to handle the server denying it so the
4810 only safe thing to do is to terminate the connection.
4811 [Steve Henson]
4812
4813 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4814 peer supports secure renegotiation and 0 otherwise. Print out peer
4815 renegotiation support in s_client/s_server.
4816 [Steve Henson]
4817
4818 *) Replace the highly broken and deprecated SPKAC certification method with
4819 the updated NID creation version. This should correctly handle UTF8.
4820 [Steve Henson]
4821
4822 *) Implement RFC5746. Re-enable renegotiation but require the extension
4823 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
4824 turns out to be a bad idea. It has been replaced by
4825 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
4826 SSL_CTX_set_options(). This is really not recommended unless you
4827 know what you are doing.
4828 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
4829
4830 *) Fixes to stateless session resumption handling. Use initial_ctx when
4831 issuing and attempting to decrypt tickets in case it has changed during
4832 servername handling. Use a non-zero length session ID when attempting
4833 stateless session resumption: this makes it possible to determine if
4834 a resumption has occurred immediately after receiving server hello
4835 (several places in OpenSSL subtly assume this) instead of later in
4836 the handshake.
4837 [Steve Henson]
4838
4839 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
4840 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
4841 fixes for a few places where the return code is not checked
4842 correctly.
4843 [Julia Lawall <julia@diku.dk>]
4844
4845 *) Add --strict-warnings option to Configure script to include devteam
4846 warnings in other configurations.
4847 [Steve Henson]
4848
4849 *) Add support for --libdir option and LIBDIR variable in makefiles. This
4850 makes it possible to install openssl libraries in locations which
4851 have names other than "lib", for example "/usr/lib64" which some
4852 systems need.
4853 [Steve Henson, based on patch from Jeremy Utley]
4854
4855 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
4856 X690 8.9.12 and can produce some misleading textual output of OIDs.
4857 [Steve Henson, reported by Dan Kaminsky]
4858
4859 *) Delete MD2 from algorithm tables. This follows the recommendation in
4860 several standards that it is not used in new applications due to
4861 several cryptographic weaknesses. For binary compatibility reasons
4862 the MD2 API is still compiled in by default.
4863 [Steve Henson]
4864
4865 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
4866 and restored.
4867 [Steve Henson]
4868
4869 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
4870 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
4871 clash.
4872 [Guenter <lists@gknw.net>]
4873
4874 *) Fix the server certificate chain building code to use X509_verify_cert(),
4875 it used to have an ad-hoc builder which was unable to cope with anything
4876 other than a simple chain.
4877 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
4878
4879 *) Don't check self signed certificate signatures in X509_verify_cert()
4880 by default (a flag can override this): it just wastes time without
4881 adding any security. As a useful side effect self signed root CAs
4882 with non-FIPS digests are now usable in FIPS mode.
4883 [Steve Henson]
4884
4885 *) In dtls1_process_out_of_seq_message() the check if the current message
4886 is already buffered was missing. For every new message was memory
4887 allocated, allowing an attacker to perform an denial of service attack
4888 with sending out of seq handshake messages until there is no memory
4889 left. Additionally every future message was buffered, even if the
4890 sequence number made no sense and would be part of another handshake.
4891 So only messages with sequence numbers less than 10 in advance will be
4892 buffered. (CVE-2009-1378)
4893 [Robin Seggelmann, discovered by Daniel Mentz]
4894
4895 *) Records are buffered if they arrive with a future epoch to be
4896 processed after finishing the corresponding handshake. There is
4897 currently no limitation to this buffer allowing an attacker to perform
4898 a DOS attack with sending records with future epochs until there is no
4899 memory left. This patch adds the pqueue_size() function to determine
4900 the size of a buffer and limits the record buffer to 100 entries.
4901 (CVE-2009-1377)
4902 [Robin Seggelmann, discovered by Daniel Mentz]
4903
4904 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
4905 parent structure is freed. (CVE-2009-1379)
4906 [Daniel Mentz]
4907
4908 *) Handle non-blocking I/O properly in SSL_shutdown() call.
4909 [Darryl Miles <darryl-mailinglists@netbauds.net>]
4910
4911 *) Add 2.5.4.* OIDs
4912 [Ilya O. <vrghost@gmail.com>]
4913
4914 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
4915
4916 *) Disable renegotiation completely - this fixes a severe security
4917 problem (CVE-2009-3555) at the cost of breaking all
4918 renegotiation. Renegotiation can be re-enabled by setting
4919 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
4920 run-time. This is really not recommended unless you know what
4921 you're doing.
4922 [Ben Laurie]
4923
4924 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
4925
4926 *) Don't set val to NULL when freeing up structures, it is freed up by
4927 underlying code. If sizeof(void *) > sizeof(long) this can result in
4928 zeroing past the valid field. (CVE-2009-0789)
4929 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
4930
4931 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
4932 checked correctly. This would allow some invalid signed attributes to
4933 appear to verify correctly. (CVE-2009-0591)
4934 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4935
4936 *) Reject UniversalString and BMPString types with invalid lengths. This
4937 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
4938 a legal length. (CVE-2009-0590)
4939 [Steve Henson]
4940
4941 *) Set S/MIME signing as the default purpose rather than setting it
4942 unconditionally. This allows applications to override it at the store
4943 level.
4944 [Steve Henson]
4945
4946 *) Permit restricted recursion of ASN1 strings. This is needed in practice
4947 to handle some structures.
4948 [Steve Henson]
4949
4950 *) Improve efficiency of mem_gets: don't search whole buffer each time
4951 for a '\n'
4952 [Jeremy Shapiro <jnshapir@us.ibm.com>]
4953
4954 *) New -hex option for openssl rand.
4955 [Matthieu Herrb]
4956
4957 *) Print out UTF8String and NumericString when parsing ASN1.
4958 [Steve Henson]
4959
4960 *) Support NumericString type for name components.
4961 [Steve Henson]
4962
4963 *) Allow CC in the environment to override the automatically chosen
4964 compiler. Note that nothing is done to ensure flags work with the
4965 chosen compiler.
4966 [Ben Laurie]
4967
4968 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
4969
4970 *) Properly check EVP_VerifyFinal() and similar return values
4971 (CVE-2008-5077).
4972 [Ben Laurie, Bodo Moeller, Google Security Team]
4973
4974 *) Enable TLS extensions by default.
4975 [Ben Laurie]
4976
4977 *) Allow the CHIL engine to be loaded, whether the application is
4978 multithreaded or not. (This does not release the developer from the
4979 obligation to set up the dynamic locking callbacks.)
4980 [Sander Temme <sander@temme.net>]
4981
4982 *) Use correct exit code if there is an error in dgst command.
4983 [Steve Henson; problem pointed out by Roland Dirlewanger]
4984
4985 *) Tweak Configure so that you need to say "experimental-jpake" to enable
4986 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
4987 [Bodo Moeller]
4988
4989 *) Add experimental JPAKE support, including demo authentication in
4990 s_client and s_server.
4991 [Ben Laurie]
4992
4993 *) Set the comparison function in v3_addr_canonize().
4994 [Rob Austein <sra@hactrn.net>]
4995
4996 *) Add support for XMPP STARTTLS in s_client.
4997 [Philip Paeps <philip@freebsd.org>]
4998
4999 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
5000 to ensure that even with this option, only ciphersuites in the
5001 server's preference list will be accepted. (Note that the option
5002 applies only when resuming a session, so the earlier behavior was
5003 just about the algorithm choice for symmetric cryptography.)
5004 [Bodo Moeller]
5005
5006 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
5007
5008 *) Fix NULL pointer dereference if a DTLS server received
5009 ChangeCipherSpec as first record (CVE-2009-1386).
5010 [PR #1679]
5011
5012 *) Fix a state transition in s3_srvr.c and d1_srvr.c
5013 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
5014 [Nagendra Modadugu]
5015
5016 *) The fix in 0.9.8c that supposedly got rid of unsafe
5017 double-checked locking was incomplete for RSA blinding,
5018 addressing just one layer of what turns out to have been
5019 doubly unsafe triple-checked locking.
5020
5021 So now fix this for real by retiring the MONT_HELPER macro
5022 in crypto/rsa/rsa_eay.c.
5023
5024 [Bodo Moeller; problem pointed out by Marius Schilder]
5025
5026 *) Various precautionary measures:
5027
5028 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
5029
5030 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
5031 (NB: This would require knowledge of the secret session ticket key
5032 to exploit, in which case you'd be SOL either way.)
5033
5034 - Change bn_nist.c so that it will properly handle input BIGNUMs
5035 outside the expected range.
5036
5037 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
5038 builds.
5039
5040 [Neel Mehta, Bodo Moeller]
5041
5042 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
5043 the load fails. Useful for distros.
5044 [Ben Laurie and the FreeBSD team]
5045
5046 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
5047 [Steve Henson]
5048
5049 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
5050 [Huang Ying]
5051
5052 *) Expand ENGINE to support engine supplied SSL client certificate functions.
5053
5054 This work was sponsored by Logica.
5055 [Steve Henson]
5056
5057 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
5058 keystores. Support for SSL/TLS client authentication too.
5059 Not compiled unless enable-capieng specified to Configure.
5060
5061 This work was sponsored by Logica.
5062 [Steve Henson]
5063
5064 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
5065 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
5066 attribute creation routines such as certificate requests and PKCS#12
5067 files.
5068 [Steve Henson]
5069
5070 Changes between 0.9.8g and 0.9.8h [28 May 2008]
5071
5072 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
5073 handshake which could lead to a client crash as found using the
5074 Codenomicon TLS test suite (CVE-2008-1672)
5075 [Steve Henson, Mark Cox]
5076
5077 *) Fix double free in TLS server name extensions which could lead to
5078 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
5079 [Joe Orton]
5080
5081 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
5082
5083 Clear the error queue to ensure that error entries left from
5084 older function calls do not interfere with the correct operation.
5085 [Lutz Jaenicke, Erik de Castro Lopo]
5086
5087 *) Remove root CA certificates of commercial CAs:
5088
5089 The OpenSSL project does not recommend any specific CA and does not
5090 have any policy with respect to including or excluding any CA.
5091 Therefore it does not make any sense to ship an arbitrary selection
5092 of root CA certificates with the OpenSSL software.
5093 [Lutz Jaenicke]
5094
5095 *) RSA OAEP patches to fix two separate invalid memory reads.
5096 The first one involves inputs when 'lzero' is greater than
5097 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
5098 before the beginning of from). The second one involves inputs where
5099 the 'db' section contains nothing but zeroes (there is a one-byte
5100 invalid read after the end of 'db').
5101 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5102
5103 *) Partial backport from 0.9.9-dev:
5104
5105 Introduce bn_mul_mont (dedicated Montgomery multiplication
5106 procedure) as a candidate for BIGNUM assembler implementation.
5107 While 0.9.9-dev uses assembler for various architectures, only
5108 x86_64 is available by default here in the 0.9.8 branch, and
5109 32-bit x86 is available through a compile-time setting.
5110
5111 To try the 32-bit x86 assembler implementation, use Configure
5112 option "enable-montasm" (which exists only for this backport).
5113
5114 As "enable-montasm" for 32-bit x86 disclaims code stability
5115 anyway, in this constellation we activate additional code
5116 backported from 0.9.9-dev for further performance improvements,
5117 namely BN_from_montgomery_word. (To enable this otherwise,
5118 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
5119
5120 [Andy Polyakov (backport partially by Bodo Moeller)]
5121
5122 *) Add TLS session ticket callback. This allows an application to set
5123 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
5124 values. This is useful for key rollover for example where several key
5125 sets may exist with different names.
5126 [Steve Henson]
5127
5128 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
5129 This was broken until now in 0.9.8 releases, such that the only way
5130 a registered ENGINE could be used (assuming it initialises
5131 successfully on the host) was to explicitly set it as the default
5132 for the relevant algorithms. This is in contradiction with 0.9.7
5133 behaviour and the documentation. With this fix, when an ENGINE is
5134 registered into a given algorithm's table of implementations, the
5135 'uptodate' flag is reset so that auto-discovery will be used next
5136 time a new context for that algorithm attempts to select an
5137 implementation.
5138 [Ian Lister (tweaked by Geoff Thorpe)]
5139
5140 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5141 implementation in the following ways:
5142
5143 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5144 hard coded.
5145
5146 Lack of BER streaming support means one pass streaming processing is
5147 only supported if data is detached: setting the streaming flag is
5148 ignored for embedded content.
5149
5150 CMS support is disabled by default and must be explicitly enabled
5151 with the enable-cms configuration option.
5152 [Steve Henson]
5153
5154 *) Update the GMP engine glue to do direct copies between BIGNUM and
5155 mpz_t when openssl and GMP use the same limb size. Otherwise the
5156 existing "conversion via a text string export" trick is still used.
5157 [Paul Sheer <paulsheer@gmail.com>]
5158
5159 *) Zlib compression BIO. This is a filter BIO which compressed and
5160 uncompresses any data passed through it.
5161 [Steve Henson]
5162
5163 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5164 RFC3394 compatible AES key wrapping.
5165 [Steve Henson]
5166
5167 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5168 sets string data without copying. X509_ALGOR_set0() and
5169 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5170 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5171 from an X509_ATTRIBUTE structure optionally checking it occurs only
5172 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5173 data.
5174 [Steve Henson]
5175
5176 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5177 to get the expected BN_FLG_CONSTTIME behavior.
5178 [Bodo Moeller (Google)]
5179
5180 *) Netware support:
5181
5182 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5183 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5184 - added some more tests to do_tests.pl
5185 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5186 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5187 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5188 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5189 - various changes to netware.pl to enable gcc-cross builds on Win32
5190 platform
5191 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5192 - various changes to fix missing prototype warnings
5193 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5194 - added AES, WHIRLPOOL and CPUID assembler code to build files
5195 - added missing AES assembler make rules to mk1mf.pl
5196 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5197 [Guenter Knauf <eflash@gmx.net>]
5198
5199 *) Implement certificate status request TLS extension defined in RFC3546.
5200 A client can set the appropriate parameters and receive the encoded
5201 OCSP response via a callback. A server can query the supplied parameters
5202 and set the encoded OCSP response in the callback. Add simplified examples
5203 to s_client and s_server.
5204 [Steve Henson]
5205
5206 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5207
5208 *) Fix various bugs:
5209 + Binary incompatibility of ssl_ctx_st structure
5210 + DTLS interoperation with non-compliant servers
5211 + Don't call get_session_cb() without proposed session
5212 + Fix ia64 assembler code
5213 [Andy Polyakov, Steve Henson]
5214
5215 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5216
5217 *) DTLS Handshake overhaul. There were longstanding issues with
5218 OpenSSL DTLS implementation, which were making it impossible for
5219 RFC 4347 compliant client to communicate with OpenSSL server.
5220 Unfortunately just fixing these incompatibilities would "cut off"
5221 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5222 server keeps tolerating non RFC compliant syntax. The opposite is
5223 not true, 0.9.8f client can not communicate with earlier server.
5224 This update even addresses CVE-2007-4995.
5225 [Andy Polyakov]
5226
5227 *) Changes to avoid need for function casts in OpenSSL: some compilers
5228 (gcc 4.2 and later) reject their use.
5229 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5230 Steve Henson]
5231
5232 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5233 RFC4507bis. The encrypted ticket format is an encrypted encoded
5234 SSL_SESSION structure, that way new session features are automatically
5235 supported.
5236
5237 If a client application caches session in an SSL_SESSION structure
5238 support is transparent because tickets are now stored in the encoded
5239 SSL_SESSION.
5240
5241 The SSL_CTX structure automatically generates keys for ticket
5242 protection in servers so again support should be possible
5243 with no application modification.
5244
5245 If a client or server wishes to disable RFC4507 support then the option
5246 SSL_OP_NO_TICKET can be set.
5247
5248 Add a TLS extension debugging callback to allow the contents of any client
5249 or server extensions to be examined.
5250
5251 This work was sponsored by Google.
5252 [Steve Henson]
5253
5254 *) Add initial support for TLS extensions, specifically for the server_name
5255 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5256 have new members for a host name. The SSL data structure has an
5257 additional member SSL_CTX *initial_ctx so that new sessions can be
5258 stored in that context to allow for session resumption, even after the
5259 SSL has been switched to a new SSL_CTX in reaction to a client's
5260 server_name extension.
5261
5262 New functions (subject to change):
5263
5264 SSL_get_servername()
5265 SSL_get_servername_type()
5266 SSL_set_SSL_CTX()
5267
5268 New CTRL codes and macros (subject to change):
5269
5270 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5271 - SSL_CTX_set_tlsext_servername_callback()
5272 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5273 - SSL_CTX_set_tlsext_servername_arg()
5274 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5275
5276 openssl s_client has a new '-servername ...' option.
5277
5278 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5279 '-key2 ...', '-servername_fatal' (subject to change). This allows
5280 testing the HostName extension for a specific single host name ('-cert'
5281 and '-key' remain fallbacks for handshakes without HostName
5282 negotiation). If the unrecognized_name alert has to be sent, this by
5283 default is a warning; it becomes fatal with the '-servername_fatal'
5284 option.
5285
5286 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
5287
5288 *) Add AES and SSE2 assembly language support to VC++ build.
5289 [Steve Henson]
5290
5291 *) Mitigate attack on final subtraction in Montgomery reduction.
5292 [Andy Polyakov]
5293
5294 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
5295 (which previously caused an internal error).
5296 [Bodo Moeller]
5297
5298 *) Squeeze another 10% out of IGE mode when in != out.
5299 [Ben Laurie]
5300
5301 *) AES IGE mode speedup.
5302 [Dean Gaudet (Google)]
5303
5304 *) Add the Korean symmetric 128-bit cipher SEED (see
5305 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
5306 add SEED ciphersuites from RFC 4162:
5307
5308 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
5309 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
5310 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
5311 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
5312
5313 To minimize changes between patchlevels in the OpenSSL 0.9.8
5314 series, SEED remains excluded from compilation unless OpenSSL
5315 is configured with 'enable-seed'.
5316 [KISA, Bodo Moeller]
5317
5318 *) Mitigate branch prediction attacks, which can be practical if a
5319 single processor is shared, allowing a spy process to extract
5320 information. For detailed background information, see
5321 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
5322 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
5323 and Necessary Software Countermeasures"). The core of the change
5324 are new versions BN_div_no_branch() and
5325 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
5326 respectively, which are slower, but avoid the security-relevant
5327 conditional branches. These are automatically called by BN_div()
5328 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
5329 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
5330 remove a conditional branch.
5331
5332 BN_FLG_CONSTTIME is the new name for the previous
5333 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
5334 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
5335 in the exponent causes BN_mod_exp_mont() to use the alternative
5336 implementation in BN_mod_exp_mont_consttime().) The old name
5337 remains as a deprecated alias.
5338
5339 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
5340 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
5341 constant-time implementations for more than just exponentiation.
5342 Here too the old name is kept as a deprecated alias.
5343
5344 BN_BLINDING_new() will now use BN_dup() for the modulus so that
5345 the BN_BLINDING structure gets an independent copy of the
5346 modulus. This means that the previous "BIGNUM *m" argument to
5347 BN_BLINDING_new() and to BN_BLINDING_create_param() now
5348 essentially becomes "const BIGNUM *m", although we can't actually
5349 change this in the header file before 0.9.9. It allows
5350 RSA_setup_blinding() to use BN_with_flags() on the modulus to
5351 enable BN_FLG_CONSTTIME.
5352
5353 [Matthew D Wood (Intel Corp)]
5354
5355 *) In the SSL/TLS server implementation, be strict about session ID
5356 context matching (which matters if an application uses a single
5357 external cache for different purposes). Previously,
5358 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
5359 set. This did ensure strict client verification, but meant that,
5360 with applications using a single external cache for quite
5361 different requirements, clients could circumvent ciphersuite
5362 restrictions for a given session ID context by starting a session
5363 in a different context.
5364 [Bodo Moeller]
5365
5366 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5367 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5368 authentication-only ciphersuites.
5369 [Bodo Moeller]
5370
5371 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
5372 not complete and could lead to a possible single byte overflow
5373 (CVE-2007-5135) [Ben Laurie]
5374
5375 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
5376
5377 *) Since AES128 and AES256 (and similarly Camellia128 and
5378 Camellia256) share a single mask bit in the logic of
5379 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5380 kludge to work properly if AES128 is available and AES256 isn't
5381 (or if Camellia128 is available and Camellia256 isn't).
5382 [Victor Duchovni]
5383
5384 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
5385 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
5386 When a point or a seed is encoded in a BIT STRING, we need to
5387 prevent the removal of trailing zero bits to get the proper DER
5388 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
5389 of a NamedBitList, for which trailing 0 bits need to be removed.)
5390 [Bodo Moeller]
5391
5392 *) Have SSL/TLS server implementation tolerate "mismatched" record
5393 protocol version while receiving ClientHello even if the
5394 ClientHello is fragmented. (The server can't insist on the
5395 particular protocol version it has chosen before the ServerHello
5396 message has informed the client about his choice.)
5397 [Bodo Moeller]
5398
5399 *) Add RFC 3779 support.
5400 [Rob Austein for ARIN, Ben Laurie]
5401
5402 *) Load error codes if they are not already present instead of using a
5403 static variable. This allows them to be cleanly unloaded and reloaded.
5404 Improve header file function name parsing.
5405 [Steve Henson]
5406
5407 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
5408 or CAPABILITY handshake as required by RFCs.
5409 [Goetz Babin-Ebell]
5410
5411 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5412
5413 *) Introduce limits to prevent malicious keys being able to
5414 cause a denial of service. (CVE-2006-2940)
5415 [Steve Henson, Bodo Moeller]
5416
5417 *) Fix ASN.1 parsing of certain invalid structures that can result
5418 in a denial of service. (CVE-2006-2937) [Steve Henson]
5419
5420 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5421 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5422
5423 *) Fix SSL client code which could crash if connecting to a
5424 malicious SSLv2 server. (CVE-2006-4343)
5425 [Tavis Ormandy and Will Drewry, Google Security Team]
5426
5427 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
5428 match only those. Before that, "AES256-SHA" would be interpreted
5429 as a pattern and match "AES128-SHA" too (since AES128-SHA got
5430 the same strength classification in 0.9.7h) as we currently only
5431 have a single AES bit in the ciphersuite description bitmap.
5432 That change, however, also applied to ciphersuite strings such as
5433 "RC4-MD5" that intentionally matched multiple ciphersuites --
5434 namely, SSL 2.0 ciphersuites in addition to the more common ones
5435 from SSL 3.0/TLS 1.0.
5436
5437 So we change the selection algorithm again: Naming an explicit
5438 ciphersuite selects this one ciphersuite, and any other similar
5439 ciphersuite (same bitmap) from *other* protocol versions.
5440 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5441 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5442
5443 Since SSL 2.0 does not have any ciphersuites for which the
5444 128/256 bit distinction would be relevant, this works for now.
5445 The proper fix will be to use different bits for AES128 and
5446 AES256, which would have avoided the problems from the beginning;
5447 however, bits are scarce, so we can only do this in a new release
5448 (not just a patchlevel) when we can change the SSL_CIPHER
5449 definition to split the single 'unsigned long mask' bitmap into
5450 multiple values to extend the available space.
5451
5452 [Bodo Moeller]
5453
5454 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5455
5456 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5457 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5458
5459 *) Add AES IGE and biIGE modes.
5460 [Ben Laurie]
5461
5462 *) Change the Unix randomness entropy gathering to use poll() when
5463 possible instead of select(), since the latter has some
5464 undesirable limitations.
5465 [Darryl Miles via Richard Levitte and Bodo Moeller]
5466
5467 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
5468 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
5469 cannot be implicitly activated as part of, e.g., the "AES" alias.
5470 However, please upgrade to OpenSSL 0.9.9[-dev] for
5471 non-experimental use of the ECC ciphersuites to get TLS extension
5472 support, which is required for curve and point format negotiation
5473 to avoid potential handshake problems.
5474 [Bodo Moeller]
5475
5476 *) Disable rogue ciphersuites:
5477
5478 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5479 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5480 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5481
5482 The latter two were purportedly from
5483 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5484 appear there.
5485
5486 Also deactivate the remaining ciphersuites from
5487 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5488 unofficial, and the ID has long expired.
5489 [Bodo Moeller]
5490
5491 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5492 dual-core machines) and other potential thread-safety issues.
5493 [Bodo Moeller]
5494
5495 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
5496 versions), which is now available for royalty-free use
5497 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
5498 Also, add Camellia TLS ciphersuites from RFC 4132.
5499
5500 To minimize changes between patchlevels in the OpenSSL 0.9.8
5501 series, Camellia remains excluded from compilation unless OpenSSL
5502 is configured with 'enable-camellia'.
5503 [NTT]
5504
5505 *) Disable the padding bug check when compression is in use. The padding
5506 bug check assumes the first packet is of even length, this is not
5507 necessarily true if compression is enabled and can result in false
5508 positives causing handshake failure. The actual bug test is ancient
5509 code so it is hoped that implementations will either have fixed it by
5510 now or any which still have the bug do not support compression.
5511 [Steve Henson]
5512
5513 Changes between 0.9.8a and 0.9.8b [04 May 2006]
5514
5515 *) When applying a cipher rule check to see if string match is an explicit
5516 cipher suite and only match that one cipher suite if it is.
5517 [Steve Henson]
5518
5519 *) Link in manifests for VC++ if needed.
5520 [Austin Ziegler <halostatue@gmail.com>]
5521
5522 *) Update support for ECC-based TLS ciphersuites according to
5523 draft-ietf-tls-ecc-12.txt with proposed changes (but without
5524 TLS extensions, which are supported starting with the 0.9.9
5525 branch, not in the OpenSSL 0.9.8 branch).
5526 [Douglas Stebila]
5527
5528 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
5529 opaque EVP_CIPHER_CTX handling.
5530 [Steve Henson]
5531
5532 *) Fixes and enhancements to zlib compression code. We now only use
5533 "zlib1.dll" and use the default __cdecl calling convention on Win32
5534 to conform with the standards mentioned here:
5535 http://www.zlib.net/DLL_FAQ.txt
5536 Static zlib linking now works on Windows and the new --with-zlib-include
5537 --with-zlib-lib options to Configure can be used to supply the location
5538 of the headers and library. Gracefully handle case where zlib library
5539 can't be loaded.
5540 [Steve Henson]
5541
5542 *) Several fixes and enhancements to the OID generation code. The old code
5543 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
5544 handle numbers larger than ULONG_MAX, truncated printing and had a
5545 non standard OBJ_obj2txt() behaviour.
5546 [Steve Henson]
5547
5548 *) Add support for building of engines under engine/ as shared libraries
5549 under VC++ build system.
5550 [Steve Henson]
5551
5552 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
5553 Hopefully, we will not see any false combination of paths any more.
5554 [Richard Levitte]
5555
5556 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5557
5558 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5559 (part of SSL_OP_ALL). This option used to disable the
5560 countermeasure against man-in-the-middle protocol-version
5561 rollback in the SSL 2.0 server implementation, which is a bad
5562 idea. (CVE-2005-2969)
5563
5564 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5565 for Information Security, National Institute of Advanced Industrial
5566 Science and Technology [AIST], Japan)]
5567
5568 *) Add two function to clear and return the verify parameter flags.
5569 [Steve Henson]
5570
5571 *) Keep cipherlists sorted in the source instead of sorting them at
5572 runtime, thus removing the need for a lock.
5573 [Nils Larsch]
5574
5575 *) Avoid some small subgroup attacks in Diffie-Hellman.
5576 [Nick Mathewson and Ben Laurie]
5577
5578 *) Add functions for well-known primes.
5579 [Nick Mathewson]
5580
5581 *) Extended Windows CE support.
5582 [Satoshi Nakamura and Andy Polyakov]
5583
5584 *) Initialize SSL_METHOD structures at compile time instead of during
5585 runtime, thus removing the need for a lock.
5586 [Steve Henson]
5587
5588 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5589 attempting to decrypt each encrypted key in turn. Add support to
5590 smime utility.
5591 [Steve Henson]
5592
5593 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5594
5595 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5596 OpenSSL 0.9.8.]
5597
5598 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5599 [Richard Levitte]
5600
5601 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5602 key into the same file any more.
5603 [Richard Levitte]
5604
5605 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5606 [Andy Polyakov]
5607
5608 *) Add -utf8 command line and config file option to 'ca'.
5609 [Stefan <stf@udoma.org]
5610
5611 *) Removed the macro des_crypt(), as it seems to conflict with some
5612 libraries. Use DES_crypt().
5613 [Richard Levitte]
5614
5615 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5616 involves renaming the source and generated shared-libs for
5617 both. The engines will accept the corrected or legacy ids
5618 ('ncipher' and '4758_cca' respectively) when binding. NB,
5619 this only applies when building 'shared'.
5620 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5621
5622 *) Add attribute functions to EVP_PKEY structure. Modify
5623 PKCS12_create() to recognize a CSP name attribute and
5624 use it. Make -CSP option work again in pkcs12 utility.
5625 [Steve Henson]
5626
5627 *) Add new functionality to the bn blinding code:
5628 - automatic re-creation of the BN_BLINDING parameters after
5629 a fixed number of uses (currently 32)
5630 - add new function for parameter creation
5631 - introduce flags to control the update behaviour of the
5632 BN_BLINDING parameters
5633 - hide BN_BLINDING structure
5634 Add a second BN_BLINDING slot to the RSA structure to improve
5635 performance when a single RSA object is shared among several
5636 threads.
5637 [Nils Larsch]
5638
5639 *) Add support for DTLS.
5640 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5641
5642 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5643 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5644 [Walter Goulet]
5645
5646 *) Remove buggy and incomplete DH cert support from
5647 ssl/ssl_rsa.c and ssl/s3_both.c
5648 [Nils Larsch]
5649
5650 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5651 the apps/openssl applications.
5652 [Nils Larsch]
5653
5654 *) Compile clean with "-Wall -Wmissing-prototypes
5655 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5656 DEBUG_SAFESTACK must also be set.
5657 [Ben Laurie]
5658
5659 *) Change ./Configure so that certain algorithms can be disabled by default.
5660 The new counterpiece to "no-xxx" is "enable-xxx".
5661
5662 The patented RC5 and MDC2 algorithms will now be disabled unless
5663 "enable-rc5" and "enable-mdc2", respectively, are specified.
5664
5665 (IDEA remains enabled despite being patented. This is because IDEA
5666 is frequently required for interoperability, and there is no license
5667 fee for non-commercial use. As before, "no-idea" can be used to
5668 avoid this algorithm.)
5669
5670 [Bodo Moeller]
5671
5672 *) Add processing of proxy certificates (see RFC 3820). This work was
5673 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5674 EGEE (Enabling Grids for E-science in Europe).
5675 [Richard Levitte]
5676
5677 *) RC4 performance overhaul on modern architectures/implementations, such
5678 as Intel P4, IA-64 and AMD64.
5679 [Andy Polyakov]
5680
5681 *) New utility extract-section.pl. This can be used specify an alternative
5682 section number in a pod file instead of having to treat each file as
5683 a separate case in Makefile. This can be done by adding two lines to the
5684 pod file:
5685
5686 =for comment openssl_section:XXX
5687
5688 The blank line is mandatory.
5689
5690 [Steve Henson]
5691
5692 *) New arguments -certform, -keyform and -pass for s_client and s_server
5693 to allow alternative format key and certificate files and passphrase
5694 sources.
5695 [Steve Henson]
5696
5697 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5698 update associated structures and add various utility functions.
5699
5700 Add new policy related verify parameters, include policy checking in
5701 standard verify code. Enhance 'smime' application with extra parameters
5702 to support policy checking and print out.
5703 [Steve Henson]
5704
5705 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
5706 Nehemiah processors. These extensions support AES encryption in hardware
5707 as well as RNG (though RNG support is currently disabled).
5708 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
5709
5710 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
5711 [Geoff Thorpe]
5712
5713 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
5714 [Andy Polyakov and a number of other people]
5715
5716 *) Improved PowerPC platform support. Most notably BIGNUM assembler
5717 implementation contributed by IBM.
5718 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
5719
5720 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
5721 exponent rather than 'unsigned long'. There is a corresponding change to
5722 the new 'rsa_keygen' element of the RSA_METHOD structure.
5723 [Jelte Jansen, Geoff Thorpe]
5724
5725 *) Functionality for creating the initial serial number file is now
5726 moved from CA.pl to the 'ca' utility with a new option -create_serial.
5727
5728 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
5729 number file to 1, which is bound to cause problems. To avoid
5730 the problems while respecting compatibility between different 0.9.7
5731 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
5732 CA.pl for serial number initialization. With the new release 0.9.8,
5733 we can fix the problem directly in the 'ca' utility.)
5734 [Steve Henson]
5735
5736 *) Reduced header interdependencies by declaring more opaque objects in
5737 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
5738 give fewer recursive includes, which could break lazy source code - so
5739 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
5740 developers should define this symbol when building and using openssl to
5741 ensure they track the recommended behaviour, interfaces, [etc], but
5742 backwards-compatible behaviour prevails when this isn't defined.
5743 [Geoff Thorpe]
5744
5745 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
5746 [Steve Henson]
5747
5748 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
5749 This will generate a random key of the appropriate length based on the
5750 cipher context. The EVP_CIPHER can provide its own random key generation
5751 routine to support keys of a specific form. This is used in the des and
5752 3des routines to generate a key of the correct parity. Update S/MIME
5753 code to use new functions and hence generate correct parity DES keys.
5754 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
5755 valid (weak or incorrect parity).
5756 [Steve Henson]
5757
5758 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
5759 as looking them up. This is useful when the verified structure may contain
5760 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
5761 present unless the new PKCS7_NO_CRL flag is asserted.
5762 [Steve Henson]
5763
5764 *) Extend ASN1 oid configuration module. It now additionally accepts the
5765 syntax:
5766
5767 shortName = some long name, 1.2.3.4
5768 [Steve Henson]
5769
5770 *) Reimplemented the BN_CTX implementation. There is now no more static
5771 limitation on the number of variables it can handle nor the depth of the
5772 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
5773 information can now expand as required, and rather than having a single
5774 static array of bignums, BN_CTX now uses a linked-list of such arrays
5775 allowing it to expand on demand whilst maintaining the usefulness of
5776 BN_CTX's "bundling".
5777 [Geoff Thorpe]
5778
5779 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5780 to allow all RSA operations to function using a single BN_CTX.
5781 [Geoff Thorpe]
5782
5783 *) Preliminary support for certificate policy evaluation and checking. This
5784 is initially intended to pass the tests outlined in "Conformance Testing
5785 of Relying Party Client Certificate Path Processing Logic" v1.07.
5786 [Steve Henson]
5787
5788 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
5789 remained unused and not that useful. A variety of other little bignum
5790 tweaks and fixes have also been made continuing on from the audit (see
5791 below).
5792 [Geoff Thorpe]
5793
5794 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
5795 associated ASN1, EVP and SSL functions and old ASN1 macros.
5796 [Richard Levitte]
5797
5798 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
5799 and this should never fail. So the return value from the use of
5800 BN_set_word() (which can fail due to needless expansion) is now deprecated;
5801 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
5802 [Geoff Thorpe]
5803
5804 *) BN_CTX_get() should return zero-valued bignums, providing the same
5805 initialised value as BN_new().
5806 [Geoff Thorpe, suggested by Ulf Möller]
5807
5808 *) Support for inhibitAnyPolicy certificate extension.
5809 [Steve Henson]
5810
5811 *) An audit of the BIGNUM code is underway, for which debugging code is
5812 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
5813 is considered valid when processing BIGNUMs, and causes execution to
5814 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
5815 further steps are taken to deliberately pollute unused data in BIGNUM
5816 structures to try and expose faulty code further on. For now, openssl will
5817 (in its default mode of operation) continue to tolerate the inconsistent
5818 forms that it has tolerated in the past, but authors and packagers should
5819 consider trying openssl and their own applications when compiled with
5820 these debugging symbols defined. It will help highlight potential bugs in
5821 their own code, and will improve the test coverage for OpenSSL itself. At
5822 some point, these tighter rules will become openssl's default to improve
5823 maintainability, though the assert()s and other overheads will remain only
5824 in debugging configurations. See bn.h for more details.
5825 [Geoff Thorpe, Nils Larsch, Ulf Möller]
5826
5827 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
5828 that can only be obtained through BN_CTX_new() (which implicitly
5829 initialises it). The presence of this function only made it possible
5830 to overwrite an existing structure (and cause memory leaks).
5831 [Geoff Thorpe]
5832
5833 *) Because of the callback-based approach for implementing LHASH as a
5834 template type, lh_insert() adds opaque objects to hash-tables and
5835 lh_doall() or lh_doall_arg() are typically used with a destructor callback
5836 to clean up those corresponding objects before destroying the hash table
5837 (and losing the object pointers). So some over-zealous constifications in
5838 LHASH have been relaxed so that lh_insert() does not take (nor store) the
5839 objects as "const" and the lh_doall[_arg] callback wrappers are not
5840 prototyped to have "const" restrictions on the object pointers they are
5841 given (and so aren't required to cast them away any more).
5842 [Geoff Thorpe]
5843
5844 *) The tmdiff.h API was so ugly and minimal that our own timing utility
5845 (speed) prefers to use its own implementation. The two implementations
5846 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
5847 its object type properly exposed (MS_TM) instead of casting to/from "char
5848 *". This may still change yet if someone realises MS_TM and "ms_time_***"
5849 aren't necessarily the greatest nomenclatures - but this is what was used
5850 internally to the implementation so I've used that for now.
5851 [Geoff Thorpe]
5852
5853 *) Ensure that deprecated functions do not get compiled when
5854 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
5855 the self-tests were still using deprecated key-generation functions so
5856 these have been updated also.
5857 [Geoff Thorpe]
5858
5859 *) Reorganise PKCS#7 code to separate the digest location functionality
5860 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
5861 New function PKCS7_set_digest() to set the digest type for PKCS#7
5862 digestedData type. Add additional code to correctly generate the
5863 digestedData type and add support for this type in PKCS7 initialization
5864 functions.
5865 [Steve Henson]
5866
5867 *) New function PKCS7_set0_type_other() this initializes a PKCS7
5868 structure of type "other".
5869 [Steve Henson]
5870
5871 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
5872 sure the loop does correctly stop and breaking ("division by zero")
5873 modulus operations are not performed. The (pre-generated) prime
5874 table crypto/bn/bn_prime.h was already correct, but it could not be
5875 re-generated on some platforms because of the "division by zero"
5876 situation in the script.
5877 [Ralf S. Engelschall]
5878
5879 *) Update support for ECC-based TLS ciphersuites according to
5880 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
5881 SHA-1 now is only used for "small" curves (where the
5882 representation of a field element takes up to 24 bytes); for
5883 larger curves, the field element resulting from ECDH is directly
5884 used as premaster secret.
5885 [Douglas Stebila (Sun Microsystems Laboratories)]
5886
5887 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
5888 curve secp160r1 to the tests.
5889 [Douglas Stebila (Sun Microsystems Laboratories)]
5890
5891 *) Add the possibility to load symbols globally with DSO.
5892 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
5893
5894 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
5895 control of the error stack.
5896 [Richard Levitte]
5897
5898 *) Add support for STORE in ENGINE.
5899 [Richard Levitte]
5900
5901 *) Add the STORE type. The intention is to provide a common interface
5902 to certificate and key stores, be they simple file-based stores, or
5903 HSM-type store, or LDAP stores, or...
5904 NOTE: The code is currently UNTESTED and isn't really used anywhere.
5905 [Richard Levitte]
5906
5907 *) Add a generic structure called OPENSSL_ITEM. This can be used to
5908 pass a list of arguments to any function as well as provide a way
5909 for a function to pass data back to the caller.
5910 [Richard Levitte]
5911
5912 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
5913 works like BUF_strdup() but can be used to duplicate a portion of
5914 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
5915 a memory area.
5916 [Richard Levitte]
5917
5918 *) Add the function sk_find_ex() which works like sk_find(), but will
5919 return an index to an element even if an exact match couldn't be
5920 found. The index is guaranteed to point at the element where the
5921 searched-for key would be inserted to preserve sorting order.
5922 [Richard Levitte]
5923
5924 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
5925 takes an extra flags argument for optional functionality. Currently,
5926 the following flags are defined:
5927
5928 OBJ_BSEARCH_VALUE_ON_NOMATCH
5929 This one gets OBJ_bsearch_ex() to return a pointer to the first
5930 element where the comparing function returns a negative or zero
5931 number.
5932
5933 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
5934 This one gets OBJ_bsearch_ex() to return a pointer to the first
5935 element where the comparing function returns zero. This is useful
5936 if there are more than one element where the comparing function
5937 returns zero.
5938 [Richard Levitte]
5939
5940 *) Make it possible to create self-signed certificates with 'openssl ca'
5941 in such a way that the self-signed certificate becomes part of the
5942 CA database and uses the same mechanisms for serial number generation
5943 as all other certificate signing. The new flag '-selfsign' enables
5944 this functionality. Adapt CA.sh and CA.pl.in.
5945 [Richard Levitte]
5946
5947 *) Add functionality to check the public key of a certificate request
5948 against a given private. This is useful to check that a certificate
5949 request can be signed by that key (self-signing).
5950 [Richard Levitte]
5951
5952 *) Make it possible to have multiple active certificates with the same
5953 subject in the CA index file. This is done only if the keyword
5954 'unique_subject' is set to 'no' in the main CA section (default
5955 if 'CA_default') of the configuration file. The value is saved
5956 with the database itself in a separate index attribute file,
5957 named like the index file with '.attr' appended to the name.
5958 [Richard Levitte]
5959
5960 *) Generate multi-valued AVAs using '+' notation in config files for
5961 req and dirName.
5962 [Steve Henson]
5963
5964 *) Support for nameConstraints certificate extension.
5965 [Steve Henson]
5966
5967 *) Support for policyConstraints certificate extension.
5968 [Steve Henson]
5969
5970 *) Support for policyMappings certificate extension.
5971 [Steve Henson]
5972
5973 *) Make sure the default DSA_METHOD implementation only uses its
5974 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
5975 and change its own handlers to be NULL so as to remove unnecessary
5976 indirection. This lets alternative implementations fallback to the
5977 default implementation more easily.
5978 [Geoff Thorpe]
5979
5980 *) Support for directoryName in GeneralName related extensions
5981 in config files.
5982 [Steve Henson]
5983
5984 *) Make it possible to link applications using Makefile.shared.
5985 Make that possible even when linking against static libraries!
5986 [Richard Levitte]
5987
5988 *) Support for single pass processing for S/MIME signing. This now
5989 means that S/MIME signing can be done from a pipe, in addition
5990 cleartext signing (multipart/signed type) is effectively streaming
5991 and the signed data does not need to be all held in memory.
5992
5993 This is done with a new flag PKCS7_STREAM. When this flag is set
5994 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
5995 is done after the data is output (and digests calculated) in
5996 SMIME_write_PKCS7().
5997 [Steve Henson]
5998
5999 *) Add full support for -rpath/-R, both in shared libraries and
6000 applications, at least on the platforms where it's known how
6001 to do it.
6002 [Richard Levitte]
6003
6004 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
6005 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
6006 will now compute a table of multiples of the generator that
6007 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
6008 faster (notably in the case of a single point multiplication,
6009 scalar * generator).
6010 [Nils Larsch, Bodo Moeller]
6011
6012 *) IPv6 support for certificate extensions. The various extensions
6013 which use the IP:a.b.c.d can now take IPv6 addresses using the
6014 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
6015 correctly.
6016 [Steve Henson]
6017
6018 *) Added an ENGINE that implements RSA by performing private key
6019 exponentiations with the GMP library. The conversions to and from
6020 GMP's mpz_t format aren't optimised nor are any montgomery forms
6021 cached, and on x86 it appears OpenSSL's own performance has caught up.
6022 However there are likely to be other architectures where GMP could
6023 provide a boost. This ENGINE is not built in by default, but it can be
6024 specified at Configure time and should be accompanied by the necessary
6025 linker additions, eg;
6026 ./config -DOPENSSL_USE_GMP -lgmp
6027 [Geoff Thorpe]
6028
6029 *) "openssl engine" will not display ENGINE/DSO load failure errors when
6030 testing availability of engines with "-t" - the old behaviour is
6031 produced by increasing the feature's verbosity with "-tt".
6032 [Geoff Thorpe]
6033
6034 *) ECDSA routines: under certain error conditions uninitialized BN objects
6035 could be freed. Solution: make sure initialization is performed early
6036 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
6037 via PR#459)
6038 [Lutz Jaenicke]
6039
6040 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
6041 and DH_METHOD (eg. by ENGINE implementations) to override the normal
6042 software implementations. For DSA and DH, parameter generation can
6043 also be overridden by providing the appropriate method callbacks.
6044 [Geoff Thorpe]
6045
6046 *) Change the "progress" mechanism used in key-generation and
6047 primality testing to functions that take a new BN_GENCB pointer in
6048 place of callback/argument pairs. The new API functions have "_ex"
6049 postfixes and the older functions are reimplemented as wrappers for
6050 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
6051 declarations of the old functions to help (graceful) attempts to
6052 migrate to the new functions. Also, the new key-generation API
6053 functions operate on a caller-supplied key-structure and return
6054 success/failure rather than returning a key or NULL - this is to
6055 help make "keygen" another member function of RSA_METHOD etc.
6056
6057 Example for using the new callback interface:
6058
6059 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
6060 void *my_arg = ...;
6061 BN_GENCB my_cb;
6062
6063 BN_GENCB_set(&my_cb, my_callback, my_arg);
6064
6065 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
6066 /* For the meaning of a, b in calls to my_callback(), see the
6067 * documentation of the function that calls the callback.
6068 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
6069 * my_callback should return 1 if it wants BN_is_prime_ex()
6070 * to continue, or 0 to stop.
6071 */
6072
6073 [Geoff Thorpe]
6074
6075 *) Change the ZLIB compression method to be stateful, and make it
6076 available to TLS with the number defined in
6077 draft-ietf-tls-compression-04.txt.
6078 [Richard Levitte]
6079
6080 *) Add the ASN.1 structures and functions for CertificatePair, which
6081 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
6082
6083 CertificatePair ::= SEQUENCE {
6084 forward [0] Certificate OPTIONAL,
6085 reverse [1] Certificate OPTIONAL,
6086 -- at least one of the pair shall be present -- }
6087
6088 Also implement the PEM functions to read and write certificate
6089 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
6090
6091 This needed to be defined, mostly for the sake of the LDAP
6092 attribute crossCertificatePair, but may prove useful elsewhere as
6093 well.
6094 [Richard Levitte]
6095
6096 *) Make it possible to inhibit symlinking of shared libraries in
6097 Makefile.shared, for Cygwin's sake.
6098 [Richard Levitte]
6099
6100 *) Extend the BIGNUM API by creating a function
6101 void BN_set_negative(BIGNUM *a, int neg);
6102 and a macro that behave like
6103 int BN_is_negative(const BIGNUM *a);
6104
6105 to avoid the need to access 'a->neg' directly in applications.
6106 [Nils Larsch]
6107
6108 *) Implement fast modular reduction for pseudo-Mersenne primes
6109 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
6110 EC_GROUP_new_curve_GFp() will now automatically use this
6111 if applicable.
6112 [Nils Larsch <nla@trustcenter.de>]
6113
6114 *) Add new lock type (CRYPTO_LOCK_BN).
6115 [Bodo Moeller]
6116
6117 *) Change the ENGINE framework to automatically load engines
6118 dynamically from specific directories unless they could be
6119 found to already be built in or loaded. Move all the
6120 current engines except for the cryptodev one to a new
6121 directory engines/.
6122 The engines in engines/ are built as shared libraries if
6123 the "shared" options was given to ./Configure or ./config.
6124 Otherwise, they are inserted in libcrypto.a.
6125 /usr/local/ssl/engines is the default directory for dynamic
6126 engines, but that can be overridden at configure time through
6127 the usual use of --prefix and/or --openssldir, and at run
6128 time with the environment variable OPENSSL_ENGINES.
6129 [Geoff Thorpe and Richard Levitte]
6130
6131 *) Add Makefile.shared, a helper makefile to build shared
6132 libraries. Adapt Makefile.org.
6133 [Richard Levitte]
6134
6135 *) Add version info to Win32 DLLs.
6136 [Peter 'Luna' Runestig" <peter@runestig.com>]
6137
6138 *) Add new 'medium level' PKCS#12 API. Certificates and keys
6139 can be added using this API to created arbitrary PKCS#12
6140 files while avoiding the low level API.
6141
6142 New options to PKCS12_create(), key or cert can be NULL and
6143 will then be omitted from the output file. The encryption
6144 algorithm NIDs can be set to -1 for no encryption, the mac
6145 iteration count can be set to 0 to omit the mac.
6146
6147 Enhance pkcs12 utility by making the -nokeys and -nocerts
6148 options work when creating a PKCS#12 file. New option -nomac
6149 to omit the mac, NONE can be set for an encryption algorithm.
6150 New code is modified to use the enhanced PKCS12_create()
6151 instead of the low level API.
6152 [Steve Henson]
6153
6154 *) Extend ASN1 encoder to support indefinite length constructed
6155 encoding. This can output sequences tags and octet strings in
6156 this form. Modify pk7_asn1.c to support indefinite length
6157 encoding. This is experimental and needs additional code to
6158 be useful, such as an ASN1 bio and some enhanced streaming
6159 PKCS#7 code.
6160
6161 Extend template encode functionality so that tagging is passed
6162 down to the template encoder.
6163 [Steve Henson]
6164
6165 *) Let 'openssl req' fail if an argument to '-newkey' is not
6166 recognized instead of using RSA as a default.
6167 [Bodo Moeller]
6168
6169 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6170 As these are not official, they are not included in "ALL";
6171 the "ECCdraft" ciphersuite group alias can be used to select them.
6172 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6173
6174 *) Add ECDH engine support.
6175 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6176
6177 *) Add ECDH in new directory crypto/ecdh/.
6178 [Douglas Stebila (Sun Microsystems Laboratories)]
6179
6180 *) Let BN_rand_range() abort with an error after 100 iterations
6181 without success (which indicates a broken PRNG).
6182 [Bodo Moeller]
6183
6184 *) Change BN_mod_sqrt() so that it verifies that the input value
6185 is really the square of the return value. (Previously,
6186 BN_mod_sqrt would show GIGO behaviour.)
6187 [Bodo Moeller]
6188
6189 *) Add named elliptic curves over binary fields from X9.62, SECG,
6190 and WAP/WTLS; add OIDs that were still missing.
6191
6192 [Sheueling Chang Shantz and Douglas Stebila
6193 (Sun Microsystems Laboratories)]
6194
6195 *) Extend the EC library for elliptic curves over binary fields
6196 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6197 New EC_METHOD:
6198
6199 EC_GF2m_simple_method
6200
6201 New API functions:
6202
6203 EC_GROUP_new_curve_GF2m
6204 EC_GROUP_set_curve_GF2m
6205 EC_GROUP_get_curve_GF2m
6206 EC_POINT_set_affine_coordinates_GF2m
6207 EC_POINT_get_affine_coordinates_GF2m
6208 EC_POINT_set_compressed_coordinates_GF2m
6209
6210 Point compression for binary fields is disabled by default for
6211 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6212 enable it).
6213
6214 As binary polynomials are represented as BIGNUMs, various members
6215 of the EC_GROUP and EC_POINT data structures can be shared
6216 between the implementations for prime fields and binary fields;
6217 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6218 are essentially identical to their ..._GFp counterparts.
6219 (For simplicity, the '..._GFp' prefix has been dropped from
6220 various internal method names.)
6221
6222 An internal 'field_div' method (similar to 'field_mul' and
6223 'field_sqr') has been added; this is used only for binary fields.
6224
6225 [Sheueling Chang Shantz and Douglas Stebila
6226 (Sun Microsystems Laboratories)]
6227
6228 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6229 through methods ('mul', 'precompute_mult').
6230
6231 The generic implementations (now internally called 'ec_wNAF_mul'
6232 and 'ec_wNAF_precomputed_mult') remain the default if these
6233 methods are undefined.
6234
6235 [Sheueling Chang Shantz and Douglas Stebila
6236 (Sun Microsystems Laboratories)]
6237
6238 *) New function EC_GROUP_get_degree, which is defined through
6239 EC_METHOD. For curves over prime fields, this returns the bit
6240 length of the modulus.
6241
6242 [Sheueling Chang Shantz and Douglas Stebila
6243 (Sun Microsystems Laboratories)]
6244
6245 *) New functions EC_GROUP_dup, EC_POINT_dup.
6246 (These simply call ..._new and ..._copy).
6247
6248 [Sheueling Chang Shantz and Douglas Stebila
6249 (Sun Microsystems Laboratories)]
6250
6251 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6252 Polynomials are represented as BIGNUMs (where the sign bit is not
6253 used) in the following functions [macros]:
6254
6255 BN_GF2m_add
6256 BN_GF2m_sub [= BN_GF2m_add]
6257 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6258 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6259 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6260 BN_GF2m_mod_inv
6261 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6262 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6263 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6264 BN_GF2m_cmp [= BN_ucmp]
6265
6266 (Note that only the 'mod' functions are actually for fields GF(2^m).
6267 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6268
6269 For some functions, an the irreducible polynomial defining a
6270 field can be given as an 'unsigned int[]' with strictly
6271 decreasing elements giving the indices of those bits that are set;
6272 i.e., p[] represents the polynomial
6273 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6274 where
6275 p[0] > p[1] > ... > p[k] = 0.
6276 This applies to the following functions:
6277
6278 BN_GF2m_mod_arr
6279 BN_GF2m_mod_mul_arr
6280 BN_GF2m_mod_sqr_arr
6281 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
6282 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
6283 BN_GF2m_mod_exp_arr
6284 BN_GF2m_mod_sqrt_arr
6285 BN_GF2m_mod_solve_quad_arr
6286 BN_GF2m_poly2arr
6287 BN_GF2m_arr2poly
6288
6289 Conversion can be performed by the following functions:
6290
6291 BN_GF2m_poly2arr
6292 BN_GF2m_arr2poly
6293
6294 bntest.c has additional tests for binary polynomial arithmetic.
6295
6296 Two implementations for BN_GF2m_mod_div() are available.
6297 The default algorithm simply uses BN_GF2m_mod_inv() and
6298 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
6299 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
6300 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
6301
6302 [Sheueling Chang Shantz and Douglas Stebila
6303 (Sun Microsystems Laboratories)]
6304
6305 *) Add new error code 'ERR_R_DISABLED' that can be used when some
6306 functionality is disabled at compile-time.
6307 [Douglas Stebila <douglas.stebila@sun.com>]
6308
6309 *) Change default behaviour of 'openssl asn1parse' so that more
6310 information is visible when viewing, e.g., a certificate:
6311
6312 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
6313 mode the content of non-printable OCTET STRINGs is output in a
6314 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
6315 avoid the appearance of a printable string.
6316 [Nils Larsch <nla@trustcenter.de>]
6317
6318 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
6319 functions
6320 EC_GROUP_set_asn1_flag()
6321 EC_GROUP_get_asn1_flag()
6322 EC_GROUP_set_point_conversion_form()
6323 EC_GROUP_get_point_conversion_form()
6324 These control ASN1 encoding details:
6325 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
6326 has been set to OPENSSL_EC_NAMED_CURVE.
6327 - Points are encoded in uncompressed form by default; options for
6328 asn1_for are as for point2oct, namely
6329 POINT_CONVERSION_COMPRESSED
6330 POINT_CONVERSION_UNCOMPRESSED
6331 POINT_CONVERSION_HYBRID
6332
6333 Also add 'seed' and 'seed_len' members to EC_GROUP with access
6334 functions
6335 EC_GROUP_set_seed()
6336 EC_GROUP_get0_seed()
6337 EC_GROUP_get_seed_len()
6338 This is used only for ASN1 purposes (so far).
6339 [Nils Larsch <nla@trustcenter.de>]
6340
6341 *) Add 'field_type' member to EC_METHOD, which holds the NID
6342 of the appropriate field type OID. The new function
6343 EC_METHOD_get_field_type() returns this value.
6344 [Nils Larsch <nla@trustcenter.de>]
6345
6346 *) Add functions
6347 EC_POINT_point2bn()
6348 EC_POINT_bn2point()
6349 EC_POINT_point2hex()
6350 EC_POINT_hex2point()
6351 providing useful interfaces to EC_POINT_point2oct() and
6352 EC_POINT_oct2point().
6353 [Nils Larsch <nla@trustcenter.de>]
6354
6355 *) Change internals of the EC library so that the functions
6356 EC_GROUP_set_generator()
6357 EC_GROUP_get_generator()
6358 EC_GROUP_get_order()
6359 EC_GROUP_get_cofactor()
6360 are implemented directly in crypto/ec/ec_lib.c and not dispatched
6361 to methods, which would lead to unnecessary code duplication when
6362 adding different types of curves.
6363 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
6364
6365 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
6366 arithmetic, and such that modified wNAFs are generated
6367 (which avoid length expansion in many cases).
6368 [Bodo Moeller]
6369
6370 *) Add a function EC_GROUP_check_discriminant() (defined via
6371 EC_METHOD) that verifies that the curve discriminant is non-zero.
6372
6373 Add a function EC_GROUP_check() that makes some sanity tests
6374 on a EC_GROUP, its generator and order. This includes
6375 EC_GROUP_check_discriminant().
6376 [Nils Larsch <nla@trustcenter.de>]
6377
6378 *) Add ECDSA in new directory crypto/ecdsa/.
6379
6380 Add applications 'openssl ecparam' and 'openssl ecdsa'
6381 (these are based on 'openssl dsaparam' and 'openssl dsa').
6382
6383 ECDSA support is also included in various other files across the
6384 library. Most notably,
6385 - 'openssl req' now has a '-newkey ecdsa:file' option;
6386 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
6387 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
6388 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
6389 them suitable for ECDSA where domain parameters must be
6390 extracted before the specific public key;
6391 - ECDSA engine support has been added.
6392 [Nils Larsch <nla@trustcenter.de>]
6393
6394 *) Include some named elliptic curves, and add OIDs from X9.62,
6395 SECG, and WAP/WTLS. Each curve can be obtained from the new
6396 function
6397 EC_GROUP_new_by_curve_name(),
6398 and the list of available named curves can be obtained with
6399 EC_get_builtin_curves().
6400 Also add a 'curve_name' member to EC_GROUP objects, which can be
6401 accessed via
6402 EC_GROUP_set_curve_name()
6403 EC_GROUP_get_curve_name()
6404 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
6405
6406 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6407 was actually never needed) and in BN_mul(). The removal in BN_mul()
6408 required a small change in bn_mul_part_recursive() and the addition
6409 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6410 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6411 bn_sub_words() and bn_add_words() except they take arrays with
6412 differing sizes.
6413 [Richard Levitte]
6414
6415 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
6416
6417 *) Cleanse PEM buffers before freeing them since they may contain
6418 sensitive data.
6419 [Benjamin Bennett <ben@psc.edu>]
6420
6421 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6422 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6423 authentication-only ciphersuites.
6424 [Bodo Moeller]
6425
6426 *) Since AES128 and AES256 share a single mask bit in the logic of
6427 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6428 kludge to work properly if AES128 is available and AES256 isn't.
6429 [Victor Duchovni]
6430
6431 *) Expand security boundary to match 1.1.1 module.
6432 [Steve Henson]
6433
6434 *) Remove redundant features: hash file source, editing of test vectors
6435 modify fipsld to use external fips_premain.c signature.
6436 [Steve Henson]
6437
6438 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6439 run algorithm test programs.
6440 [Steve Henson]
6441
6442 *) Make algorithm test programs more tolerant of whitespace.
6443 [Steve Henson]
6444
6445 *) Have SSL/TLS server implementation tolerate "mismatched" record
6446 protocol version while receiving ClientHello even if the
6447 ClientHello is fragmented. (The server can't insist on the
6448 particular protocol version it has chosen before the ServerHello
6449 message has informed the client about his choice.)
6450 [Bodo Moeller]
6451
6452 *) Load error codes if they are not already present instead of using a
6453 static variable. This allows them to be cleanly unloaded and reloaded.
6454 [Steve Henson]
6455
6456 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6457
6458 *) Introduce limits to prevent malicious keys being able to
6459 cause a denial of service. (CVE-2006-2940)
6460 [Steve Henson, Bodo Moeller]
6461
6462 *) Fix ASN.1 parsing of certain invalid structures that can result
6463 in a denial of service. (CVE-2006-2937) [Steve Henson]
6464
6465 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6466 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6467
6468 *) Fix SSL client code which could crash if connecting to a
6469 malicious SSLv2 server. (CVE-2006-4343)
6470 [Tavis Ormandy and Will Drewry, Google Security Team]
6471
6472 *) Change ciphersuite string processing so that an explicit
6473 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
6474 will no longer include "AES128-SHA"), and any other similar
6475 ciphersuite (same bitmap) from *other* protocol versions (so that
6476 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
6477 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
6478 changes from 0.9.8b and 0.9.8d.
6479 [Bodo Moeller]
6480
6481 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
6482
6483 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6484 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6485
6486 *) Change the Unix randomness entropy gathering to use poll() when
6487 possible instead of select(), since the latter has some
6488 undesirable limitations.
6489 [Darryl Miles via Richard Levitte and Bodo Moeller]
6490
6491 *) Disable rogue ciphersuites:
6492
6493 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6494 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6495 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6496
6497 The latter two were purportedly from
6498 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6499 appear there.
6500
6501 Also deactivate the remaining ciphersuites from
6502 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6503 unofficial, and the ID has long expired.
6504 [Bodo Moeller]
6505
6506 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6507 dual-core machines) and other potential thread-safety issues.
6508 [Bodo Moeller]
6509
6510 Changes between 0.9.7i and 0.9.7j [04 May 2006]
6511
6512 *) Adapt fipsld and the build system to link against the validated FIPS
6513 module in FIPS mode.
6514 [Steve Henson]
6515
6516 *) Fixes for VC++ 2005 build under Windows.
6517 [Steve Henson]
6518
6519 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
6520 from a Windows bash shell such as MSYS. It is autodetected from the
6521 "config" script when run from a VC++ environment. Modify standard VC++
6522 build to use fipscanister.o from the GNU make build.
6523 [Steve Henson]
6524
6525 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
6526
6527 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
6528 The value now differs depending on if you build for FIPS or not.
6529 BEWARE! A program linked with a shared FIPSed libcrypto can't be
6530 safely run with a non-FIPSed libcrypto, as it may crash because of
6531 the difference induced by this change.
6532 [Andy Polyakov]
6533
6534 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
6535
6536 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6537 (part of SSL_OP_ALL). This option used to disable the
6538 countermeasure against man-in-the-middle protocol-version
6539 rollback in the SSL 2.0 server implementation, which is a bad
6540 idea. (CVE-2005-2969)
6541
6542 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6543 for Information Security, National Institute of Advanced Industrial
6544 Science and Technology [AIST], Japan)]
6545
6546 *) Minimal support for X9.31 signatures and PSS padding modes. This is
6547 mainly for FIPS compliance and not fully integrated at this stage.
6548 [Steve Henson]
6549
6550 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
6551 the exponentiation using a fixed-length exponent. (Otherwise,
6552 the information leaked through timing could expose the secret key
6553 after many signatures; cf. Bleichenbacher's attack on DSA with
6554 biased k.)
6555 [Bodo Moeller]
6556
6557 *) Make a new fixed-window mod_exp implementation the default for
6558 RSA, DSA, and DH private-key operations so that the sequence of
6559 squares and multiplies and the memory access pattern are
6560 independent of the particular secret key. This will mitigate
6561 cache-timing and potential related attacks.
6562
6563 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6564 and this is automatically used by BN_mod_exp_mont() if the new flag
6565 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6566 will use this BN flag for private exponents unless the flag
6567 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6568 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6569
6570 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6571
6572 *) Change the client implementation for SSLv23_method() and
6573 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6574 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6575 (Previously, the SSL 2.0 backwards compatible Client Hello
6576 message format would be used even with SSL_OP_NO_SSLv2.)
6577 [Bodo Moeller]
6578
6579 *) Add support for smime-type MIME parameter in S/MIME messages which some
6580 clients need.
6581 [Steve Henson]
6582
6583 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6584 a threadsafe manner. Modify rsa code to use new function and add calls
6585 to dsa and dh code (which had race conditions before).
6586 [Steve Henson]
6587
6588 *) Include the fixed error library code in the C error file definitions
6589 instead of fixing them up at runtime. This keeps the error code
6590 structures constant.
6591 [Steve Henson]
6592
6593 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6594
6595 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6596 OpenSSL 0.9.8.]
6597
6598 *) Fixes for newer kerberos headers. NB: the casts are needed because
6599 the 'length' field is signed on one version and unsigned on another
6600 with no (?) obvious way to tell the difference, without these VC++
6601 complains. Also the "definition" of FAR (blank) is no longer included
6602 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6603 some needed definitions.
6604 [Steve Henson]
6605
6606 *) Undo Cygwin change.
6607 [Ulf Möller]
6608
6609 *) Added support for proxy certificates according to RFC 3820.
6610 Because they may be a security thread to unaware applications,
6611 they must be explicitly allowed in run-time. See
6612 docs/HOWTO/proxy_certificates.txt for further information.
6613 [Richard Levitte]
6614
6615 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6616
6617 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6618 server and client random values. Previously
6619 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6620 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6621
6622 This change has negligible security impact because:
6623
6624 1. Server and client random values still have 24 bytes of pseudo random
6625 data.
6626
6627 2. Server and client random values are sent in the clear in the initial
6628 handshake.
6629
6630 3. The master secret is derived using the premaster secret (48 bytes in
6631 size for static RSA ciphersuites) as well as client server and random
6632 values.
6633
6634 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6635 to our attention.
6636
6637 [Stephen Henson, reported by UK NISCC]
6638
6639 *) Use Windows randomness collection on Cygwin.
6640 [Ulf Möller]
6641
6642 *) Fix hang in EGD/PRNGD query when communication socket is closed
6643 prematurely by EGD/PRNGD.
6644 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6645
6646 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6647 [Steve Henson]
6648
6649 *) Back-port of selected performance improvements from development
6650 branch, as well as improved support for PowerPC platforms.
6651 [Andy Polyakov]
6652
6653 *) Add lots of checks for memory allocation failure, error codes to indicate
6654 failure and freeing up memory if a failure occurs.
6655 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6656
6657 *) Add new -passin argument to dgst.
6658 [Steve Henson]
6659
6660 *) Perform some character comparisons of different types in X509_NAME_cmp:
6661 this is needed for some certificates that re-encode DNs into UTF8Strings
6662 (in violation of RFC3280) and can't or won't issue name rollover
6663 certificates.
6664 [Steve Henson]
6665
6666 *) Make an explicit check during certificate validation to see that
6667 the CA setting in each certificate on the chain is correct. As a
6668 side effect always do the following basic checks on extensions,
6669 not just when there's an associated purpose to the check:
6670
6671 - if there is an unhandled critical extension (unless the user
6672 has chosen to ignore this fault)
6673 - if the path length has been exceeded (if one is set at all)
6674 - that certain extensions fit the associated purpose (if one has
6675 been given)
6676 [Richard Levitte]
6677
6678 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6679
6680 *) Avoid a race condition when CRLs are checked in a multi threaded
6681 environment. This would happen due to the reordering of the revoked
6682 entries during signature checking and serial number lookup. Now the
6683 encoding is cached and the serial number sort performed under a lock.
6684 Add new STACK function sk_is_sorted().
6685 [Steve Henson]
6686
6687 *) Add Delta CRL to the extension code.
6688 [Steve Henson]
6689
6690 *) Various fixes to s3_pkt.c so alerts are sent properly.
6691 [David Holmes <d.holmes@f5.com>]
6692
6693 *) Reduce the chances of duplicate issuer name and serial numbers (in
6694 violation of RFC3280) using the OpenSSL certificate creation utilities.
6695 This is done by creating a random 64 bit value for the initial serial
6696 number when a serial number file is created or when a self signed
6697 certificate is created using 'openssl req -x509'. The initial serial
6698 number file is created using 'openssl x509 -next_serial' in CA.pl
6699 rather than being initialized to 1.
6700 [Steve Henson]
6701
6702 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
6703
6704 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6705 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6706 [Joe Orton, Steve Henson]
6707
6708 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
6709 (CVE-2004-0112)
6710 [Joe Orton, Steve Henson]
6711
6712 *) Make it possible to have multiple active certificates with the same
6713 subject in the CA index file. This is done only if the keyword
6714 'unique_subject' is set to 'no' in the main CA section (default
6715 if 'CA_default') of the configuration file. The value is saved
6716 with the database itself in a separate index attribute file,
6717 named like the index file with '.attr' appended to the name.
6718 [Richard Levitte]
6719
6720 *) X509 verify fixes. Disable broken certificate workarounds when
6721 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
6722 keyUsage extension present. Don't accept CRLs with unhandled critical
6723 extensions: since verify currently doesn't process CRL extensions this
6724 rejects a CRL with *any* critical extensions. Add new verify error codes
6725 for these cases.
6726 [Steve Henson]
6727
6728 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
6729 A clarification of RFC2560 will require the use of OCTET STRINGs and
6730 some implementations cannot handle the current raw format. Since OpenSSL
6731 copies and compares OCSP nonces as opaque blobs without any attempt at
6732 parsing them this should not create any compatibility issues.
6733 [Steve Henson]
6734
6735 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
6736 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
6737 this HMAC (and other) operations are several times slower than OpenSSL
6738 < 0.9.7.
6739 [Steve Henson]
6740
6741 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
6742 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6743
6744 *) Use the correct content when signing type "other".
6745 [Steve Henson]
6746
6747 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
6748
6749 *) Fix various bugs revealed by running the NISCC test suite:
6750
6751 Stop out of bounds reads in the ASN1 code when presented with
6752 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6753
6754 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
6755
6756 If verify callback ignores invalid public key errors don't try to check
6757 certificate signature with the NULL public key.
6758
6759 [Steve Henson]
6760
6761 *) New -ignore_err option in ocsp application to stop the server
6762 exiting on the first error in a request.
6763 [Steve Henson]
6764
6765 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6766 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6767 specifications.
6768 [Steve Henson]
6769
6770 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6771 extra data after the compression methods not only for TLS 1.0
6772 but also for SSL 3.0 (as required by the specification).
6773 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6774
6775 *) Change X509_certificate_type() to mark the key as exported/exportable
6776 when it's 512 *bits* long, not 512 bytes.
6777 [Richard Levitte]
6778
6779 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6780 blocks during encryption.
6781 [Richard Levitte]
6782
6783 *) Various fixes to base64 BIO and non blocking I/O. On write
6784 flushes were not handled properly if the BIO retried. On read
6785 data was not being buffered properly and had various logic bugs.
6786 This also affects blocking I/O when the data being decoded is a
6787 certain size.
6788 [Steve Henson]
6789
6790 *) Various S/MIME bugfixes and compatibility changes:
6791 output correct application/pkcs7 MIME type if
6792 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
6793 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
6794 of files as .eml work). Correctly handle very long lines in MIME
6795 parser.
6796 [Steve Henson]
6797
6798 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
6799
6800 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6801 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6802 a protocol version number mismatch like a decryption error
6803 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6804 [Bodo Moeller]
6805
6806 *) Turn on RSA blinding by default in the default implementation
6807 to avoid a timing attack. Applications that don't want it can call
6808 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6809 They would be ill-advised to do so in most cases.
6810 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6811
6812 *) Change RSA blinding code so that it works when the PRNG is not
6813 seeded (in this case, the secret RSA exponent is abused as
6814 an unpredictable seed -- if it is not unpredictable, there
6815 is no point in blinding anyway). Make RSA blinding thread-safe
6816 by remembering the creator's thread ID in rsa->blinding and
6817 having all other threads use local one-time blinding factors
6818 (this requires more computation than sharing rsa->blinding, but
6819 avoids excessive locking; and if an RSA object is not shared
6820 between threads, blinding will still be very fast).
6821 [Bodo Moeller]
6822
6823 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
6824 ENGINE as defaults for all supported algorithms irrespective of
6825 the 'flags' parameter. 'flags' is now honoured, so applications
6826 should make sure they are passing it correctly.
6827 [Geoff Thorpe]
6828
6829 *) Target "mingw" now allows native Windows code to be generated in
6830 the Cygwin environment as well as with the MinGW compiler.
6831 [Ulf Moeller]
6832
6833 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
6834
6835 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6836 via timing by performing a MAC computation even if incorrect
6837 block cipher padding has been found. This is a countermeasure
6838 against active attacks where the attacker has to distinguish
6839 between bad padding and a MAC verification error. (CVE-2003-0078)
6840
6841 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6842 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6843 Martin Vuagnoux (EPFL, Ilion)]
6844
6845 *) Make the no-err option work as intended. The intention with no-err
6846 is not to have the whole error stack handling routines removed from
6847 libcrypto, it's only intended to remove all the function name and
6848 reason texts, thereby removing some of the footprint that may not
6849 be interesting if those errors aren't displayed anyway.
6850
6851 NOTE: it's still possible for any application or module to have it's
6852 own set of error texts inserted. The routines are there, just not
6853 used by default when no-err is given.
6854 [Richard Levitte]
6855
6856 *) Add support for FreeBSD on IA64.
6857 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
6858
6859 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
6860 Kerberos function mit_des_cbc_cksum(). Before this change,
6861 the value returned by DES_cbc_cksum() was like the one from
6862 mit_des_cbc_cksum(), except the bytes were swapped.
6863 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
6864
6865 *) Allow an application to disable the automatic SSL chain building.
6866 Before this a rather primitive chain build was always performed in
6867 ssl3_output_cert_chain(): an application had no way to send the
6868 correct chain if the automatic operation produced an incorrect result.
6869
6870 Now the chain builder is disabled if either:
6871
6872 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
6873
6874 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
6875
6876 The reasoning behind this is that an application would not want the
6877 auto chain building to take place if extra chain certificates are
6878 present and it might also want a means of sending no additional
6879 certificates (for example the chain has two certificates and the
6880 root is omitted).
6881 [Steve Henson]
6882
6883 *) Add the possibility to build without the ENGINE framework.
6884 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6885
6886 *) Under Win32 gmtime() can return NULL: check return value in
6887 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
6888 [Steve Henson]
6889
6890 *) DSA routines: under certain error conditions uninitialized BN objects
6891 could be freed. Solution: make sure initialization is performed early
6892 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
6893 Nils Larsch <nla@trustcenter.de> via PR#459)
6894 [Lutz Jaenicke]
6895
6896 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
6897 checked on reconnect on the client side, therefore session resumption
6898 could still fail with a "ssl session id is different" error. This
6899 behaviour is masked when SSL_OP_ALL is used due to
6900 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
6901 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6902 followup to PR #377.
6903 [Lutz Jaenicke]
6904
6905 *) IA-32 assembler support enhancements: unified ELF targets, support
6906 for SCO/Caldera platforms, fix for Cygwin shared build.
6907 [Andy Polyakov]
6908
6909 *) Add support for FreeBSD on sparc64. As a consequence, support for
6910 FreeBSD on non-x86 processors is separate from x86 processors on
6911 the config script, much like the NetBSD support.
6912 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
6913
6914 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
6915
6916 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
6917 OpenSSL 0.9.7.]
6918
6919 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
6920 code (06) was taken as the first octet of the session ID and the last
6921 octet was ignored consequently. As a result SSLv2 client side session
6922 caching could not have worked due to the session ID mismatch between
6923 client and server.
6924 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6925 PR #377.
6926 [Lutz Jaenicke]
6927
6928 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
6929 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
6930 removed entirely.
6931 [Richard Levitte]
6932
6933 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
6934 seems that in spite of existing for more than a year, many application
6935 author have done nothing to provide the necessary callbacks, which
6936 means that this particular engine will not work properly anywhere.
6937 This is a very unfortunate situation which forces us, in the name
6938 of usability, to give the hw_ncipher.c a static lock, which is part
6939 of libcrypto.
6940 NOTE: This is for the 0.9.7 series ONLY. This hack will never
6941 appear in 0.9.8 or later. We EXPECT application authors to have
6942 dealt properly with this when 0.9.8 is released (unless we actually
6943 make such changes in the libcrypto locking code that changes will
6944 have to be made anyway).
6945 [Richard Levitte]
6946
6947 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
6948 octets have been read, EOF or an error occurs. Without this change
6949 some truncated ASN1 structures will not produce an error.
6950 [Steve Henson]
6951
6952 *) Disable Heimdal support, since it hasn't been fully implemented.
6953 Still give the possibility to force the use of Heimdal, but with
6954 warnings and a request that patches get sent to openssl-dev.
6955 [Richard Levitte]
6956
6957 *) Add the VC-CE target, introduce the WINCE sysname, and add
6958 INSTALL.WCE and appropriate conditionals to make it build.
6959 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6960
6961 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
6962 cygssl-x.y.z.dll, where x, y and z are the major, minor and
6963 edit numbers of the version.
6964 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
6965
6966 *) Introduce safe string copy and catenation functions
6967 (BUF_strlcpy() and BUF_strlcat()).
6968 [Ben Laurie (CHATS) and Richard Levitte]
6969
6970 *) Avoid using fixed-size buffers for one-line DNs.
6971 [Ben Laurie (CHATS)]
6972
6973 *) Add BUF_MEM_grow_clean() to avoid information leakage when
6974 resizing buffers containing secrets, and use where appropriate.
6975 [Ben Laurie (CHATS)]
6976
6977 *) Avoid using fixed size buffers for configuration file location.
6978 [Ben Laurie (CHATS)]
6979
6980 *) Avoid filename truncation for various CA files.
6981 [Ben Laurie (CHATS)]
6982
6983 *) Use sizeof in preference to magic numbers.
6984 [Ben Laurie (CHATS)]
6985
6986 *) Avoid filename truncation in cert requests.
6987 [Ben Laurie (CHATS)]
6988
6989 *) Add assertions to check for (supposedly impossible) buffer
6990 overflows.
6991 [Ben Laurie (CHATS)]
6992
6993 *) Don't cache truncated DNS entries in the local cache (this could
6994 potentially lead to a spoofing attack).
6995 [Ben Laurie (CHATS)]
6996
6997 *) Fix various buffers to be large enough for hex/decimal
6998 representations in a platform independent manner.
6999 [Ben Laurie (CHATS)]
7000
7001 *) Add CRYPTO_realloc_clean() to avoid information leakage when
7002 resizing buffers containing secrets, and use where appropriate.
7003 [Ben Laurie (CHATS)]
7004
7005 *) Add BIO_indent() to avoid much slightly worrying code to do
7006 indents.
7007 [Ben Laurie (CHATS)]
7008
7009 *) Convert sprintf()/BIO_puts() to BIO_printf().
7010 [Ben Laurie (CHATS)]
7011
7012 *) buffer_gets() could terminate with the buffer only half
7013 full. Fixed.
7014 [Ben Laurie (CHATS)]
7015
7016 *) Add assertions to prevent user-supplied crypto functions from
7017 overflowing internal buffers by having large block sizes, etc.
7018 [Ben Laurie (CHATS)]
7019
7020 *) New OPENSSL_assert() macro (similar to assert(), but enabled
7021 unconditionally).
7022 [Ben Laurie (CHATS)]
7023
7024 *) Eliminate unused copy of key in RC4.
7025 [Ben Laurie (CHATS)]
7026
7027 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
7028 [Ben Laurie (CHATS)]
7029
7030 *) Fix off-by-one error in EGD path.
7031 [Ben Laurie (CHATS)]
7032
7033 *) If RANDFILE path is too long, ignore instead of truncating.
7034 [Ben Laurie (CHATS)]
7035
7036 *) Eliminate unused and incorrectly sized X.509 structure
7037 CBCParameter.
7038 [Ben Laurie (CHATS)]
7039
7040 *) Eliminate unused and dangerous function knumber().
7041 [Ben Laurie (CHATS)]
7042
7043 *) Eliminate unused and dangerous structure, KSSL_ERR.
7044 [Ben Laurie (CHATS)]
7045
7046 *) Protect against overlong session ID context length in an encoded
7047 session object. Since these are local, this does not appear to be
7048 exploitable.
7049 [Ben Laurie (CHATS)]
7050
7051 *) Change from security patch (see 0.9.6e below) that did not affect
7052 the 0.9.6 release series:
7053
7054 Remote buffer overflow in SSL3 protocol - an attacker could
7055 supply an oversized master key in Kerberos-enabled versions.
7056 (CVE-2002-0657)
7057 [Ben Laurie (CHATS)]
7058
7059 *) Change the SSL kerb5 codes to match RFC 2712.
7060 [Richard Levitte]
7061
7062 *) Make -nameopt work fully for req and add -reqopt switch.
7063 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
7064
7065 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
7066 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
7067
7068 *) Make sure tests can be performed even if the corresponding algorithms
7069 have been removed entirely. This was also the last step to make
7070 OpenSSL compilable with DJGPP under all reasonable conditions.
7071 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
7072
7073 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
7074 to allow version independent disabling of normally unselected ciphers,
7075 which may be activated as a side-effect of selecting a single cipher.
7076
7077 (E.g., cipher list string "RSA" enables ciphersuites that are left
7078 out of "ALL" because they do not provide symmetric encryption.
7079 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
7080 [Lutz Jaenicke, Bodo Moeller]
7081
7082 *) Add appropriate support for separate platform-dependent build
7083 directories. The recommended way to make a platform-dependent
7084 build directory is the following (tested on Linux), maybe with
7085 some local tweaks:
7086
7087 # Place yourself outside of the OpenSSL source tree. In
7088 # this example, the environment variable OPENSSL_SOURCE
7089 # is assumed to contain the absolute OpenSSL source directory.
7090 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
7091 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
7092 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
7093 mkdir -p `dirname $F`
7094 ln -s $OPENSSL_SOURCE/$F $F
7095 done
7096
7097 To be absolutely sure not to disturb the source tree, a "make clean"
7098 is a good thing. If it isn't successful, don't worry about it,
7099 it probably means the source directory is very clean.
7100 [Richard Levitte]
7101
7102 *) Make sure any ENGINE control commands make local copies of string
7103 pointers passed to them whenever necessary. Otherwise it is possible
7104 the caller may have overwritten (or deallocated) the original string
7105 data when a later ENGINE operation tries to use the stored values.
7106 [Götz Babin-Ebell <babinebell@trustcenter.de>]
7107
7108 *) Improve diagnostics in file reading and command-line digests.
7109 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
7110
7111 *) Add AES modes CFB and OFB to the object database. Correct an
7112 error in AES-CFB decryption.
7113 [Richard Levitte]
7114
7115 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
7116 allows existing EVP_CIPHER_CTX structures to be reused after
7117 calling EVP_*Final(). This behaviour is used by encryption
7118 BIOs and some applications. This has the side effect that
7119 applications must explicitly clean up cipher contexts with
7120 EVP_CIPHER_CTX_cleanup() or they will leak memory.
7121 [Steve Henson]
7122
7123 *) Check the values of dna and dnb in bn_mul_recursive before calling
7124 bn_mul_comba (a non zero value means the a or b arrays do not contain
7125 n2 elements) and fallback to bn_mul_normal if either is not zero.
7126 [Steve Henson]
7127
7128 *) Fix escaping of non-ASCII characters when using the -subj option
7129 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
7130 [Lutz Jaenicke]
7131
7132 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
7133 form for "surname", serialNumber has no short form.
7134 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
7135 therefore remove "mail" short name for "internet 7".
7136 The OID for unique identifiers in X509 certificates is
7137 x500UniqueIdentifier, not uniqueIdentifier.
7138 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7139 [Lutz Jaenicke]
7140
7141 *) Add an "init" command to the ENGINE config module and auto initialize
7142 ENGINEs. Without any "init" command the ENGINE will be initialized
7143 after all ctrl commands have been executed on it. If init=1 the
7144 ENGINE is initialized at that point (ctrls before that point are run
7145 on the uninitialized ENGINE and after on the initialized one). If
7146 init=0 then the ENGINE will not be initialized at all.
7147 [Steve Henson]
7148
7149 *) Fix the 'app_verify_callback' interface so that the user-defined
7150 argument is actually passed to the callback: In the
7151 SSL_CTX_set_cert_verify_callback() prototype, the callback
7152 declaration has been changed from
7153 int (*cb)()
7154 into
7155 int (*cb)(X509_STORE_CTX *,void *);
7156 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7157 i=s->ctx->app_verify_callback(&ctx)
7158 has been changed into
7159 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7160
7161 To update applications using SSL_CTX_set_cert_verify_callback(),
7162 a dummy argument can be added to their callback functions.
7163 [D. K. Smetters <smetters@parc.xerox.com>]
7164
7165 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7166 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7167
7168 *) Add and OPENSSL_LOAD_CONF define which will cause
7169 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7170 This allows older applications to transparently support certain
7171 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7172 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7173 load the config file and OPENSSL_add_all_algorithms_conf() which will
7174 always load it have also been added.
7175 [Steve Henson]
7176
7177 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7178 Adjust NIDs and EVP layer.
7179 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7180
7181 *) Config modules support in openssl utility.
7182
7183 Most commands now load modules from the config file,
7184 though in a few (such as version) this isn't done
7185 because it couldn't be used for anything.
7186
7187 In the case of ca and req the config file used is
7188 the same as the utility itself: that is the -config
7189 command line option can be used to specify an
7190 alternative file.
7191 [Steve Henson]
7192
7193 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7194 use "openssl_conf" if filename is NULL use default openssl config file.
7195 [Steve Henson]
7196
7197 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7198 config section name. Add a new flag to tolerate a missing config file
7199 and move code to CONF_modules_load_file().
7200 [Steve Henson]
7201
7202 *) Support for crypto accelerator cards from Accelerated Encryption
7203 Processing, www.aep.ie. (Use engine 'aep')
7204 The support was copied from 0.9.6c [engine] and adapted/corrected
7205 to work with the new engine framework.
7206 [AEP Inc. and Richard Levitte]
7207
7208 *) Support for SureWare crypto accelerator cards from Baltimore
7209 Technologies. (Use engine 'sureware')
7210 The support was copied from 0.9.6c [engine] and adapted
7211 to work with the new engine framework.
7212 [Richard Levitte]
7213
7214 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7215 make the newer ENGINE framework commands for the CHIL engine work.
7216 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7217
7218 *) Make it possible to produce shared libraries on ReliantUNIX.
7219 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7220
7221 *) Add the configuration target debug-linux-ppro.
7222 Make 'openssl rsa' use the general key loading routines
7223 implemented in apps.c, and make those routines able to
7224 handle the key format FORMAT_NETSCAPE and the variant
7225 FORMAT_IISSGC.
7226 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7227
7228 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7229 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7230
7231 *) Add -keyform to rsautl, and document -engine.
7232 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7233
7234 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7235 BIO_R_NO_SUCH_FILE error code rather than the generic
7236 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7237 [Ben Laurie]
7238
7239 *) Add new functions
7240 ERR_peek_last_error
7241 ERR_peek_last_error_line
7242 ERR_peek_last_error_line_data.
7243 These are similar to
7244 ERR_peek_error
7245 ERR_peek_error_line
7246 ERR_peek_error_line_data,
7247 but report on the latest error recorded rather than the first one
7248 still in the error queue.
7249 [Ben Laurie, Bodo Moeller]
7250
7251 *) default_algorithms option in ENGINE config module. This allows things
7252 like:
7253 default_algorithms = ALL
7254 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7255 [Steve Henson]
7256
7257 *) Preliminary ENGINE config module.
7258 [Steve Henson]
7259
7260 *) New experimental application configuration code.
7261 [Steve Henson]
7262
7263 *) Change the AES code to follow the same name structure as all other
7264 symmetric ciphers, and behave the same way. Move everything to
7265 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7266 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7267
7268 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7269 [Ben Laurie and Theo de Raadt]
7270
7271 *) Add option to output public keys in req command.
7272 [Massimiliano Pala madwolf@openca.org]
7273
7274 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7275 (up to about 10% better than before for P-192 and P-224).
7276 [Bodo Moeller]
7277
7278 *) New functions/macros
7279
7280 SSL_CTX_set_msg_callback(ctx, cb)
7281 SSL_CTX_set_msg_callback_arg(ctx, arg)
7282 SSL_set_msg_callback(ssl, cb)
7283 SSL_set_msg_callback_arg(ssl, arg)
7284
7285 to request calling a callback function
7286
7287 void cb(int write_p, int version, int content_type,
7288 const void *buf, size_t len, SSL *ssl, void *arg)
7289
7290 whenever a protocol message has been completely received
7291 (write_p == 0) or sent (write_p == 1). Here 'version' is the
7292 protocol version according to which the SSL library interprets
7293 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
7294 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
7295 the content type as defined in the SSL 3.0/TLS 1.0 protocol
7296 specification (change_cipher_spec(20), alert(21), handshake(22)).
7297 'buf' and 'len' point to the actual message, 'ssl' to the
7298 SSL object, and 'arg' is the application-defined value set by
7299 SSL[_CTX]_set_msg_callback_arg().
7300
7301 'openssl s_client' and 'openssl s_server' have new '-msg' options
7302 to enable a callback that displays all protocol messages.
7303 [Bodo Moeller]
7304
7305 *) Change the shared library support so shared libraries are built as
7306 soon as the corresponding static library is finished, and thereby get
7307 openssl and the test programs linked against the shared library.
7308 This still only happens when the keyword "shard" has been given to
7309 the configuration scripts.
7310
7311 NOTE: shared library support is still an experimental thing, and
7312 backward binary compatibility is still not guaranteed.
7313 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
7314
7315 *) Add support for Subject Information Access extension.
7316 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7317
7318 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
7319 additional bytes when new memory had to be allocated, not just
7320 when reusing an existing buffer.
7321 [Bodo Moeller]
7322
7323 *) New command line and configuration option 'utf8' for the req command.
7324 This allows field values to be specified as UTF8 strings.
7325 [Steve Henson]
7326
7327 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
7328 runs for the former and machine-readable output for the latter.
7329 [Ben Laurie]
7330
7331 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
7332 of the e-mail address in the DN (i.e., it will go into a certificate
7333 extension only). The new configuration file option 'email_in_dn = no'
7334 has the same effect.
7335 [Massimiliano Pala madwolf@openca.org]
7336
7337 *) Change all functions with names starting with des_ to be starting
7338 with DES_ instead. Add wrappers that are compatible with libdes,
7339 but are named _ossl_old_des_*. Finally, add macros that map the
7340 des_* symbols to the corresponding _ossl_old_des_* if libdes
7341 compatibility is desired. If OpenSSL 0.9.6c compatibility is
7342 desired, the des_* symbols will be mapped to DES_*, with one
7343 exception.
7344
7345 Since we provide two compatibility mappings, the user needs to
7346 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
7347 compatibility is desired. The default (i.e., when that macro
7348 isn't defined) is OpenSSL 0.9.6c compatibility.
7349
7350 There are also macros that enable and disable the support of old
7351 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
7352 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
7353 are defined, the default will apply: to support the old des routines.
7354
7355 In either case, one must include openssl/des.h to get the correct
7356 definitions. Do not try to just include openssl/des_old.h, that
7357 won't work.
7358
7359 NOTE: This is a major break of an old API into a new one. Software
7360 authors are encouraged to switch to the DES_ style functions. Some
7361 time in the future, des_old.h and the libdes compatibility functions
7362 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
7363 default), and then completely removed.
7364 [Richard Levitte]
7365
7366 *) Test for certificates which contain unsupported critical extensions.
7367 If such a certificate is found during a verify operation it is
7368 rejected by default: this behaviour can be overridden by either
7369 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
7370 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
7371 X509_supported_extension() has also been added which returns 1 if a
7372 particular extension is supported.
7373 [Steve Henson]
7374
7375 *) Modify the behaviour of EVP cipher functions in similar way to digests
7376 to retain compatibility with existing code.
7377 [Steve Henson]
7378
7379 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
7380 compatibility with existing code. In particular the 'ctx' parameter does
7381 not have to be to be initialized before the call to EVP_DigestInit() and
7382 it is tidied up after a call to EVP_DigestFinal(). New function
7383 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
7384 EVP_MD_CTX_copy() changed to not require the destination to be
7385 initialized valid and new function EVP_MD_CTX_copy_ex() added which
7386 requires the destination to be valid.
7387
7388 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
7389 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
7390 [Steve Henson]
7391
7392 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
7393 so that complete 'Handshake' protocol structures are kept in memory
7394 instead of overwriting 'msg_type' and 'length' with 'body' data.
7395 [Bodo Moeller]
7396
7397 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
7398 [Massimo Santin via Richard Levitte]
7399
7400 *) Major restructuring to the underlying ENGINE code. This includes
7401 reduction of linker bloat, separation of pure "ENGINE" manipulation
7402 (initialisation, etc) from functionality dealing with implementations
7403 of specific crypto interfaces. This change also introduces integrated
7404 support for symmetric ciphers and digest implementations - so ENGINEs
7405 can now accelerate these by providing EVP_CIPHER and EVP_MD
7406 implementations of their own. This is detailed in crypto/engine/README
7407 as it couldn't be adequately described here. However, there are a few
7408 API changes worth noting - some RSA, DSA, DH, and RAND functions that
7409 were changed in the original introduction of ENGINE code have now
7410 reverted back - the hooking from this code to ENGINE is now a good
7411 deal more passive and at run-time, operations deal directly with
7412 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
7413 dereferencing through an ENGINE pointer any more. Also, the ENGINE
7414 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
7415 they were not being used by the framework as there is no concept of a
7416 BIGNUM_METHOD and they could not be generalised to the new
7417 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
7418 ENGINE_cpy() has been removed as it cannot be consistently defined in
7419 the new code.
7420 [Geoff Thorpe]
7421
7422 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
7423 [Steve Henson]
7424
7425 *) Change mkdef.pl to sort symbols that get the same entry number,
7426 and make sure the automatically generated functions ERR_load_*
7427 become part of libeay.num as well.
7428 [Richard Levitte]
7429
7430 *) New function SSL_renegotiate_pending(). This returns true once
7431 renegotiation has been requested (either SSL_renegotiate() call
7432 or HelloRequest/ClientHello received from the peer) and becomes
7433 false once a handshake has been completed.
7434 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
7435 sends a HelloRequest, but does not ensure that a handshake takes
7436 place. SSL_renegotiate_pending() is useful for checking if the
7437 client has followed the request.)
7438 [Bodo Moeller]
7439
7440 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7441 By default, clients may request session resumption even during
7442 renegotiation (if session ID contexts permit); with this option,
7443 session resumption is possible only in the first handshake.
7444
7445 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7446 more bits available for options that should not be part of
7447 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7448 [Bodo Moeller]
7449
7450 *) Add some demos for certificate and certificate request creation.
7451 [Steve Henson]
7452
7453 *) Make maximum certificate chain size accepted from the peer application
7454 settable (SSL*_get/set_max_cert_list()), as proposed by
7455 "Douglas E. Engert" <deengert@anl.gov>.
7456 [Lutz Jaenicke]
7457
7458 *) Add support for shared libraries for Unixware-7
7459 (Boyd Lynn Gerber <gerberb@zenez.com>).
7460 [Lutz Jaenicke]
7461
7462 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7463 be done prior to destruction. Use this to unload error strings from
7464 ENGINEs that load their own error strings. NB: This adds two new API
7465 functions to "get" and "set" this destroy handler in an ENGINE.
7466 [Geoff Thorpe]
7467
7468 *) Alter all existing ENGINE implementations (except "openssl" and
7469 "openbsd") to dynamically instantiate their own error strings. This
7470 makes them more flexible to be built both as statically-linked ENGINEs
7471 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
7472 Also, add stub code to each that makes building them as self-contained
7473 shared-libraries easier (see README.ENGINE).
7474 [Geoff Thorpe]
7475
7476 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
7477 implementations into applications that are completely implemented in
7478 self-contained shared-libraries. The "dynamic" ENGINE exposes control
7479 commands that can be used to configure what shared-library to load and
7480 to control aspects of the way it is handled. Also, made an update to
7481 the README.ENGINE file that brings its information up-to-date and
7482 provides some information and instructions on the "dynamic" ENGINE
7483 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
7484 [Geoff Thorpe]
7485
7486 *) Make it possible to unload ranges of ERR strings with a new
7487 "ERR_unload_strings" function.
7488 [Geoff Thorpe]
7489
7490 *) Add a copy() function to EVP_MD.
7491 [Ben Laurie]
7492
7493 *) Make EVP_MD routines take a context pointer instead of just the
7494 md_data void pointer.
7495 [Ben Laurie]
7496
7497 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
7498 that the digest can only process a single chunk of data
7499 (typically because it is provided by a piece of
7500 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
7501 is only going to provide a single chunk of data, and hence the
7502 framework needn't accumulate the data for oneshot drivers.
7503 [Ben Laurie]
7504
7505 *) As with "ERR", make it possible to replace the underlying "ex_data"
7506 functions. This change also alters the storage and management of global
7507 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
7508 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
7509 index counters. The API functions that use this state have been changed
7510 to take a "class_index" rather than pointers to the class's local STACK
7511 and counter, and there is now an API function to dynamically create new
7512 classes. This centralisation allows us to (a) plug a lot of the
7513 thread-safety problems that existed, and (b) makes it possible to clean
7514 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
7515 such data would previously have always leaked in application code and
7516 workarounds were in place to make the memory debugging turn a blind eye
7517 to it. Application code that doesn't use this new function will still
7518 leak as before, but their memory debugging output will announce it now
7519 rather than letting it slide.
7520
7521 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
7522 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
7523 has a return value to indicate success or failure.
7524 [Geoff Thorpe]
7525
7526 *) Make it possible to replace the underlying "ERR" functions such that the
7527 global state (2 LHASH tables and 2 locks) is only used by the "default"
7528 implementation. This change also adds two functions to "get" and "set"
7529 the implementation prior to it being automatically set the first time
7530 any other ERR function takes place. Ie. an application can call "get",
7531 pass the return value to a module it has just loaded, and that module
7532 can call its own "set" function using that value. This means the
7533 module's "ERR" operations will use (and modify) the error state in the
7534 application and not in its own statically linked copy of OpenSSL code.
7535 [Geoff Thorpe]
7536
7537 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
7538 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
7539 the operation, and provides a more encapsulated way for external code
7540 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
7541 to use these functions rather than manually incrementing the counts.
7542
7543 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
7544 [Geoff Thorpe]
7545
7546 *) Add EVP test program.
7547 [Ben Laurie]
7548
7549 *) Add symmetric cipher support to ENGINE. Expect the API to change!
7550 [Ben Laurie]
7551
7552 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
7553 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
7554 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
7555 These allow a CRL to be built without having to access X509_CRL fields
7556 directly. Modify 'ca' application to use new functions.
7557 [Steve Henson]
7558
7559 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
7560 bug workarounds. Rollback attack detection is a security feature.
7561 The problem will only arise on OpenSSL servers when TLSv1 is not
7562 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7563 Software authors not wanting to support TLSv1 will have special reasons
7564 for their choice and can explicitly enable this option.
7565 [Bodo Moeller, Lutz Jaenicke]
7566
7567 *) Rationalise EVP so it can be extended: don't include a union of
7568 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7569 (similar to those existing for EVP_CIPHER_CTX).
7570 Usage example:
7571
7572 EVP_MD_CTX md;
7573
7574 EVP_MD_CTX_init(&md); /* new function call */
7575 EVP_DigestInit(&md, EVP_sha1());
7576 EVP_DigestUpdate(&md, in, len);
7577 EVP_DigestFinal(&md, out, NULL);
7578 EVP_MD_CTX_cleanup(&md); /* new function call */
7579
7580 [Ben Laurie]
7581
7582 *) Make DES key schedule conform to the usual scheme, as well as
7583 correcting its structure. This means that calls to DES functions
7584 now have to pass a pointer to a des_key_schedule instead of a
7585 plain des_key_schedule (which was actually always a pointer
7586 anyway): E.g.,
7587
7588 des_key_schedule ks;
7589
7590 des_set_key_checked(..., &ks);
7591 des_ncbc_encrypt(..., &ks, ...);
7592
7593 (Note that a later change renames 'des_...' into 'DES_...'.)
7594 [Ben Laurie]
7595
7596 *) Initial reduction of linker bloat: the use of some functions, such as
7597 PEM causes large amounts of unused functions to be linked in due to
7598 poor organisation. For example pem_all.c contains every PEM function
7599 which has a knock on effect of linking in large amounts of (unused)
7600 ASN1 code. Grouping together similar functions and splitting unrelated
7601 functions prevents this.
7602 [Steve Henson]
7603
7604 *) Cleanup of EVP macros.
7605 [Ben Laurie]
7606
7607 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7608 correct _ecb suffix.
7609 [Ben Laurie]
7610
7611 *) Add initial OCSP responder support to ocsp application. The
7612 revocation information is handled using the text based index
7613 use by the ca application. The responder can either handle
7614 requests generated internally, supplied in files (for example
7615 via a CGI script) or using an internal minimal server.
7616 [Steve Henson]
7617
7618 *) Add configuration choices to get zlib compression for TLS.
7619 [Richard Levitte]
7620
7621 *) Changes to Kerberos SSL for RFC 2712 compliance:
7622 1. Implemented real KerberosWrapper, instead of just using
7623 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7624 2. Implemented optional authenticator field of KerberosWrapper.
7625
7626 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7627 and authenticator structs; see crypto/krb5/.
7628
7629 Generalized Kerberos calls to support multiple Kerberos libraries.
7630 [Vern Staats <staatsvr@asc.hpc.mil>,
7631 Jeffrey Altman <jaltman@columbia.edu>
7632 via Richard Levitte]
7633
7634 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7635 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7636 values for each of the key sizes rather than having just
7637 parameters (and 'speed' generating keys each time).
7638 [Geoff Thorpe]
7639
7640 *) Speed up EVP routines.
7641 Before:
7642 encrypt
7643 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7644 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7645 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7646 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7647 decrypt
7648 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7649 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7650 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7651 After:
7652 encrypt
7653 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7654 decrypt
7655 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7656 [Ben Laurie]
7657
7658 *) Added the OS2-EMX target.
7659 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7660
7661 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7662 to support NCONF routines in extension code. New function CONF_set_nconf()
7663 to allow functions which take an NCONF to also handle the old LHASH
7664 structure: this means that the old CONF compatible routines can be
7665 retained (in particular wrt extensions) without having to duplicate the
7666 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7667 [Steve Henson]
7668
7669 *) Enhance the general user interface with mechanisms for inner control
7670 and with possibilities to have yes/no kind of prompts.
7671 [Richard Levitte]
7672
7673 *) Change all calls to low level digest routines in the library and
7674 applications to use EVP. Add missing calls to HMAC_cleanup() and
7675 don't assume HMAC_CTX can be copied using memcpy().
7676 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7677
7678 *) Add the possibility to control engines through control names but with
7679 arbitrary arguments instead of just a string.
7680 Change the key loaders to take a UI_METHOD instead of a callback
7681 function pointer. NOTE: this breaks binary compatibility with earlier
7682 versions of OpenSSL [engine].
7683 Adapt the nCipher code for these new conditions and add a card insertion
7684 callback.
7685 [Richard Levitte]
7686
7687 *) Enhance the general user interface with mechanisms to better support
7688 dialog box interfaces, application-defined prompts, the possibility
7689 to use defaults (for example default passwords from somewhere else)
7690 and interrupts/cancellations.
7691 [Richard Levitte]
7692
7693 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7694 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7695 [Steve Henson]
7696
7697 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7698 tidy up some unnecessarily weird code in 'sk_new()').
7699 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
7700
7701 *) Change the key loading routines for ENGINEs to use the same kind
7702 callback (pem_password_cb) as all other routines that need this
7703 kind of callback.
7704 [Richard Levitte]
7705
7706 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
7707 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
7708 than this minimum value is recommended.
7709 [Lutz Jaenicke]
7710
7711 *) New random seeder for OpenVMS, using the system process statistics
7712 that are easily reachable.
7713 [Richard Levitte]
7714
7715 *) Windows apparently can't transparently handle global
7716 variables defined in DLLs. Initialisations such as:
7717
7718 const ASN1_ITEM *it = &ASN1_INTEGER_it;
7719
7720 won't compile. This is used by the any applications that need to
7721 declare their own ASN1 modules. This was fixed by adding the option
7722 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
7723 needed for static libraries under Win32.
7724 [Steve Henson]
7725
7726 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
7727 setting of purpose and trust fields. New X509_STORE trust and
7728 purpose functions and tidy up setting in other SSL functions.
7729 [Steve Henson]
7730
7731 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
7732 structure. These are inherited by X509_STORE_CTX when it is
7733 initialised. This allows various defaults to be set in the
7734 X509_STORE structure (such as flags for CRL checking and custom
7735 purpose or trust settings) for functions which only use X509_STORE_CTX
7736 internally such as S/MIME.
7737
7738 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
7739 trust settings if they are not set in X509_STORE. This allows X509_STORE
7740 purposes and trust (in S/MIME for example) to override any set by default.
7741
7742 Add command line options for CRL checking to smime, s_client and s_server
7743 applications.
7744 [Steve Henson]
7745
7746 *) Initial CRL based revocation checking. If the CRL checking flag(s)
7747 are set then the CRL is looked up in the X509_STORE structure and
7748 its validity and signature checked, then if the certificate is found
7749 in the CRL the verify fails with a revoked error.
7750
7751 Various new CRL related callbacks added to X509_STORE_CTX structure.
7752
7753 Command line options added to 'verify' application to support this.
7754
7755 This needs some additional work, such as being able to handle multiple
7756 CRLs with different times, extension based lookup (rather than just
7757 by subject name) and ultimately more complete V2 CRL extension
7758 handling.
7759 [Steve Henson]
7760
7761 *) Add a general user interface API (crypto/ui/). This is designed
7762 to replace things like des_read_password and friends (backward
7763 compatibility functions using this new API are provided).
7764 The purpose is to remove prompting functions from the DES code
7765 section as well as provide for prompting through dialog boxes in
7766 a window system and the like.
7767 [Richard Levitte]
7768
7769 *) Add "ex_data" support to ENGINE so implementations can add state at a
7770 per-structure level rather than having to store it globally.
7771 [Geoff]
7772
7773 *) Make it possible for ENGINE structures to be copied when retrieved by
7774 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
7775 This causes the "original" ENGINE structure to act like a template,
7776 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7777 operational state can be localised to each ENGINE structure, despite the
7778 fact they all share the same "methods". New ENGINE structures returned in
7779 this case have no functional references and the return value is the single
7780 structural reference. This matches the single structural reference returned
7781 by ENGINE_by_id() normally, when it is incremented on the pre-existing
7782 ENGINE structure.
7783 [Geoff]
7784
7785 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
7786 needs to match any other type at all we need to manually clear the
7787 tag cache.
7788 [Steve Henson]
7789
7790 *) Changes to the "openssl engine" utility to include;
7791 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
7792 about an ENGINE's available control commands.
7793 - executing control commands from command line arguments using the
7794 '-pre' and '-post' switches. '-post' is only used if '-t' is
7795 specified and the ENGINE is successfully initialised. The syntax for
7796 the individual commands are colon-separated, for example;
7797 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
7798 [Geoff]
7799
7800 *) New dynamic control command support for ENGINEs. ENGINEs can now
7801 declare their own commands (numbers), names (strings), descriptions,
7802 and input types for run-time discovery by calling applications. A
7803 subset of these commands are implicitly classed as "executable"
7804 depending on their input type, and only these can be invoked through
7805 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
7806 can be based on user input, config files, etc). The distinction is
7807 that "executable" commands cannot return anything other than a boolean
7808 result and can only support numeric or string input, whereas some
7809 discoverable commands may only be for direct use through
7810 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
7811 pointers, or other custom uses. The "executable" commands are to
7812 support parameterisations of ENGINE behaviour that can be
7813 unambiguously defined by ENGINEs and used consistently across any
7814 OpenSSL-based application. Commands have been added to all the
7815 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
7816 control over shared-library paths without source code alterations.
7817 [Geoff]
7818
7819 *) Changed all ENGINE implementations to dynamically allocate their
7820 ENGINEs rather than declaring them statically. Apart from this being
7821 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
7822 this also allows the implementations to compile without using the
7823 internal engine_int.h header.
7824 [Geoff]
7825
7826 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
7827 'const' value. Any code that should be able to modify a RAND_METHOD
7828 should already have non-const pointers to it (ie. they should only
7829 modify their own ones).
7830 [Geoff]
7831
7832 *) Made a variety of little tweaks to the ENGINE code.
7833 - "atalla" and "ubsec" string definitions were moved from header files
7834 to C code. "nuron" string definitions were placed in variables
7835 rather than hard-coded - allowing parameterisation of these values
7836 later on via ctrl() commands.
7837 - Removed unused "#if 0"'d code.
7838 - Fixed engine list iteration code so it uses ENGINE_free() to release
7839 structural references.
7840 - Constified the RAND_METHOD element of ENGINE structures.
7841 - Constified various get/set functions as appropriate and added
7842 missing functions (including a catch-all ENGINE_cpy that duplicates
7843 all ENGINE values onto a new ENGINE except reference counts/state).
7844 - Removed NULL parameter checks in get/set functions. Setting a method
7845 or function to NULL is a way of cancelling out a previously set
7846 value. Passing a NULL ENGINE parameter is just plain stupid anyway
7847 and doesn't justify the extra error symbols and code.
7848 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
7849 flags from engine_int.h to engine.h.
7850 - Changed prototypes for ENGINE handler functions (init(), finish(),
7851 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
7852 [Geoff]
7853
7854 *) Implement binary inversion algorithm for BN_mod_inverse in addition
7855 to the algorithm using long division. The binary algorithm can be
7856 used only if the modulus is odd. On 32-bit systems, it is faster
7857 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
7858 roughly 5-15% for 256-bit moduli), so we use it only for moduli
7859 up to 450 bits. In 64-bit environments, the binary algorithm
7860 appears to be advantageous for much longer moduli; here we use it
7861 for moduli up to 2048 bits.
7862 [Bodo Moeller]
7863
7864 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
7865 could not support the combine flag in choice fields.
7866 [Steve Henson]
7867
7868 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
7869 extensions from a certificate request to the certificate.
7870 [Steve Henson]
7871
7872 *) Allow multiple 'certopt' and 'nameopt' options to be separated
7873 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
7874 file: this allows the display of the certificate about to be
7875 signed to be customised, to allow certain fields to be included
7876 or excluded and extension details. The old system didn't display
7877 multicharacter strings properly, omitted fields not in the policy
7878 and couldn't display additional details such as extensions.
7879 [Steve Henson]
7880
7881 *) Function EC_POINTs_mul for multiple scalar multiplication
7882 of an arbitrary number of elliptic curve points
7883 \sum scalars[i]*points[i],
7884 optionally including the generator defined for the EC_GROUP:
7885 scalar*generator + \sum scalars[i]*points[i].
7886
7887 EC_POINT_mul is a simple wrapper function for the typical case
7888 that the point list has just one item (besides the optional
7889 generator).
7890 [Bodo Moeller]
7891
7892 *) First EC_METHODs for curves over GF(p):
7893
7894 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
7895 operations and provides various method functions that can also
7896 operate with faster implementations of modular arithmetic.
7897
7898 EC_GFp_mont_method() reuses most functions that are part of
7899 EC_GFp_simple_method, but uses Montgomery arithmetic.
7900
7901 [Bodo Moeller; point addition and point doubling
7902 implementation directly derived from source code provided by
7903 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
7904
7905 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
7906 crypto/ec/ec_lib.c):
7907
7908 Curves are EC_GROUP objects (with an optional group generator)
7909 based on EC_METHODs that are built into the library.
7910
7911 Points are EC_POINT objects based on EC_GROUP objects.
7912
7913 Most of the framework would be able to handle curves over arbitrary
7914 finite fields, but as there are no obvious types for fields other
7915 than GF(p), some functions are limited to that for now.
7916 [Bodo Moeller]
7917
7918 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
7919 that the file contains a complete HTTP response.
7920 [Richard Levitte]
7921
7922 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
7923 change the def and num file printf format specifier from "%-40sXXX"
7924 to "%-39s XXX". The latter will always guarantee a space after the
7925 field while the former will cause them to run together if the field
7926 is 40 of more characters long.
7927 [Steve Henson]
7928
7929 *) Constify the cipher and digest 'method' functions and structures
7930 and modify related functions to take constant EVP_MD and EVP_CIPHER
7931 pointers.
7932 [Steve Henson]
7933
7934 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
7935 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
7936 [Bodo Moeller]
7937
7938 *) Modify EVP_Digest*() routines so they now return values. Although the
7939 internal software routines can never fail additional hardware versions
7940 might.
7941 [Steve Henson]
7942
7943 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
7944
7945 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
7946 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
7947
7948 ASN1 error codes
7949 ERR_R_NESTED_ASN1_ERROR
7950 ...
7951 ERR_R_MISSING_ASN1_EOS
7952 were 4 .. 9, conflicting with
7953 ERR_LIB_RSA (= ERR_R_RSA_LIB)
7954 ...
7955 ERR_LIB_PEM (= ERR_R_PEM_LIB).
7956 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
7957
7958 Add new error code 'ERR_R_INTERNAL_ERROR'.
7959 [Bodo Moeller]
7960
7961 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
7962 suffices.
7963 [Bodo Moeller]
7964
7965 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
7966 sets the subject name for a new request or supersedes the
7967 subject name in a given request. Formats that can be parsed are
7968 'CN=Some Name, OU=myOU, C=IT'
7969 and
7970 'CN=Some Name/OU=myOU/C=IT'.
7971
7972 Add options '-batch' and '-verbose' to 'openssl req'.
7973 [Massimiliano Pala <madwolf@hackmasters.net>]
7974
7975 *) Introduce the possibility to access global variables through
7976 functions on platform were that's the best way to handle exporting
7977 global variables in shared libraries. To enable this functionality,
7978 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
7979 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
7980 is normally done by Configure or something similar).
7981
7982 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
7983 in the source file (foo.c) like this:
7984
7985 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
7986 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
7987
7988 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
7989 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
7990
7991 OPENSSL_DECLARE_GLOBAL(int,foo);
7992 #define foo OPENSSL_GLOBAL_REF(foo)
7993 OPENSSL_DECLARE_GLOBAL(double,bar);
7994 #define bar OPENSSL_GLOBAL_REF(bar)
7995
7996 The #defines are very important, and therefore so is including the
7997 header file everywhere where the defined globals are used.
7998
7999 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
8000 of ASN.1 items, but that structure is a bit different.
8001
8002 The largest change is in util/mkdef.pl which has been enhanced with
8003 better and easier to understand logic to choose which symbols should
8004 go into the Windows .def files as well as a number of fixes and code
8005 cleanup (among others, algorithm keywords are now sorted
8006 lexicographically to avoid constant rewrites).
8007 [Richard Levitte]
8008
8009 *) In BN_div() keep a copy of the sign of 'num' before writing the
8010 result to 'rm' because if rm==num the value will be overwritten
8011 and produce the wrong result if 'num' is negative: this caused
8012 problems with BN_mod() and BN_nnmod().
8013 [Steve Henson]
8014
8015 *) Function OCSP_request_verify(). This checks the signature on an
8016 OCSP request and verifies the signer certificate. The signer
8017 certificate is just checked for a generic purpose and OCSP request
8018 trust settings.
8019 [Steve Henson]
8020
8021 *) Add OCSP_check_validity() function to check the validity of OCSP
8022 responses. OCSP responses are prepared in real time and may only
8023 be a few seconds old. Simply checking that the current time lies
8024 between thisUpdate and nextUpdate max reject otherwise valid responses
8025 caused by either OCSP responder or client clock inaccuracy. Instead
8026 we allow thisUpdate and nextUpdate to fall within a certain period of
8027 the current time. The age of the response can also optionally be
8028 checked. Two new options -validity_period and -status_age added to
8029 ocsp utility.
8030 [Steve Henson]
8031
8032 *) If signature or public key algorithm is unrecognized print out its
8033 OID rather that just UNKNOWN.
8034 [Steve Henson]
8035
8036 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
8037 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
8038 ID to be generated from the issuer certificate alone which can then be
8039 passed to OCSP_id_issuer_cmp().
8040 [Steve Henson]
8041
8042 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
8043 ASN1 modules to export functions returning ASN1_ITEM pointers
8044 instead of the ASN1_ITEM structures themselves. This adds several
8045 new macros which allow the underlying ASN1 function/structure to
8046 be accessed transparently. As a result code should not use ASN1_ITEM
8047 references directly (such as &X509_it) but instead use the relevant
8048 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
8049 use of the new ASN1 code on platforms where exporting structures
8050 is problematical (for example in shared libraries) but exporting
8051 functions returning pointers to structures is not.
8052 [Steve Henson]
8053
8054 *) Add support for overriding the generation of SSL/TLS session IDs.
8055 These callbacks can be registered either in an SSL_CTX or per SSL.
8056 The purpose of this is to allow applications to control, if they wish,
8057 the arbitrary values chosen for use as session IDs, particularly as it
8058 can be useful for session caching in multiple-server environments. A
8059 command-line switch for testing this (and any client code that wishes
8060 to use such a feature) has been added to "s_server".
8061 [Geoff Thorpe, Lutz Jaenicke]
8062
8063 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
8064 of the form '#if defined(...) || defined(...) || ...' and
8065 '#if !defined(...) && !defined(...) && ...'. This also avoids
8066 the growing number of special cases it was previously handling.
8067 [Richard Levitte]
8068
8069 *) Make all configuration macros available for application by making
8070 sure they are available in opensslconf.h, by giving them names starting
8071 with "OPENSSL_" to avoid conflicts with other packages and by making
8072 sure e_os2.h will cover all platform-specific cases together with
8073 opensslconf.h.
8074 Additionally, it is now possible to define configuration/platform-
8075 specific names (called "system identities"). In the C code, these
8076 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
8077 macro with the name beginning with "OPENSSL_SYS_", which is determined
8078 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
8079 what is available.
8080 [Richard Levitte]
8081
8082 *) New option -set_serial to 'req' and 'x509' this allows the serial
8083 number to use to be specified on the command line. Previously self
8084 signed certificates were hard coded with serial number 0 and the
8085 CA options of 'x509' had to use a serial number in a file which was
8086 auto incremented.
8087 [Steve Henson]
8088
8089 *) New options to 'ca' utility to support V2 CRL entry extensions.
8090 Currently CRL reason, invalidity date and hold instruction are
8091 supported. Add new CRL extensions to V3 code and some new objects.
8092 [Steve Henson]
8093
8094 *) New function EVP_CIPHER_CTX_set_padding() this is used to
8095 disable standard block padding (aka PKCS#5 padding) in the EVP
8096 API, which was previously mandatory. This means that the data is
8097 not padded in any way and so the total length much be a multiple
8098 of the block size, otherwise an error occurs.
8099 [Steve Henson]
8100
8101 *) Initial (incomplete) OCSP SSL support.
8102 [Steve Henson]
8103
8104 *) New function OCSP_parse_url(). This splits up a URL into its host,
8105 port and path components: primarily to parse OCSP URLs. New -url
8106 option to ocsp utility.
8107 [Steve Henson]
8108
8109 *) New nonce behavior. The return value of OCSP_check_nonce() now
8110 reflects the various checks performed. Applications can decide
8111 whether to tolerate certain situations such as an absent nonce
8112 in a response when one was present in a request: the ocsp application
8113 just prints out a warning. New function OCSP_add1_basic_nonce()
8114 this is to allow responders to include a nonce in a response even if
8115 the request is nonce-less.
8116 [Steve Henson]
8117
8118 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
8119 skipped when using openssl x509 multiple times on a single input file,
8120 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
8121 [Bodo Moeller]
8122
8123 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
8124 set string type: to handle setting ASN1_TIME structures. Fix ca
8125 utility to correctly initialize revocation date of CRLs.
8126 [Steve Henson]
8127
8128 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
8129 the clients preferred ciphersuites and rather use its own preferences.
8130 Should help to work around M$ SGC (Server Gated Cryptography) bug in
8131 Internet Explorer by ensuring unchanged hash method during stepup.
8132 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
8133 [Lutz Jaenicke]
8134
8135 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
8136 to aes and add a new 'exist' option to print out symbols that don't
8137 appear to exist.
8138 [Steve Henson]
8139
8140 *) Additional options to ocsp utility to allow flags to be set and
8141 additional certificates supplied.
8142 [Steve Henson]
8143
8144 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8145 OCSP client a number of certificate to only verify the response
8146 signature against.
8147 [Richard Levitte]
8148
8149 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8150 handle the new API. Currently only ECB, CBC modes supported. Add new
8151 AES OIDs.
8152
8153 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8154 Encryption Standard (AES) Ciphersuites for Transport Layer
8155 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8156 not enabled by default and were not part of the "ALL" ciphersuite
8157 alias because they were not yet official; they could be
8158 explicitly requested by specifying the "AESdraft" ciphersuite
8159 group alias. In the final release of OpenSSL 0.9.7, the group
8160 alias is called "AES" and is part of "ALL".)
8161 [Ben Laurie, Steve Henson, Bodo Moeller]
8162
8163 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8164 request to response.
8165 [Steve Henson]
8166
8167 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8168 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8169 extract information from a certificate request. OCSP_response_create()
8170 creates a response and optionally adds a basic response structure.
8171 OCSP_basic_add1_status() adds a complete single response to a basic
8172 response and returns the OCSP_SINGLERESP structure just added (to allow
8173 extensions to be included for example). OCSP_basic_add1_cert() adds a
8174 certificate to a basic response and OCSP_basic_sign() signs a basic
8175 response with various flags. New helper functions ASN1_TIME_check()
8176 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8177 (converts ASN1_TIME to GeneralizedTime).
8178 [Steve Henson]
8179
8180 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8181 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8182 structure from a certificate. X509_pubkey_digest() digests the public_key
8183 contents: this is used in various key identifiers.
8184 [Steve Henson]
8185
8186 *) Make sk_sort() tolerate a NULL argument.
8187 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8188
8189 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8190 passed by the function are trusted implicitly. If any of them signed the
8191 response then it is assumed to be valid and is not verified.
8192 [Steve Henson]
8193
8194 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8195 to data. This was previously part of the PKCS7 ASN1 code. This
8196 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8197 [Steve Henson, reported by Kenneth R. Robinette
8198 <support@securenetterm.com>]
8199
8200 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8201 routines: without these tracing memory leaks is very painful.
8202 Fix leaks in PKCS12 and PKCS7 routines.
8203 [Steve Henson]
8204
8205 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8206 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8207 effectively meant GeneralizedTime would never be used. Now it
8208 is initialised to -1 but X509_time_adj() now has to check the value
8209 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8210 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8211 [Steve Henson, reported by Kenneth R. Robinette
8212 <support@securenetterm.com>]
8213
8214 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8215 result in a zero length in the ASN1_INTEGER structure which was
8216 not consistent with the structure when d2i_ASN1_INTEGER() was used
8217 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8218 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8219 where it did not print out a minus for negative ASN1_INTEGER.
8220 [Steve Henson]
8221
8222 *) Add summary printout to ocsp utility. The various functions which
8223 convert status values to strings have been renamed to:
8224 OCSP_response_status_str(), OCSP_cert_status_str() and
8225 OCSP_crl_reason_str() and are no longer static. New options
8226 to verify nonce values and to disable verification. OCSP response
8227 printout format cleaned up.
8228 [Steve Henson]
8229
8230 *) Add additional OCSP certificate checks. These are those specified
8231 in RFC2560. This consists of two separate checks: the CA of the
8232 certificate being checked must either be the OCSP signer certificate
8233 or the issuer of the OCSP signer certificate. In the latter case the
8234 OCSP signer certificate must contain the OCSP signing extended key
8235 usage. This check is performed by attempting to match the OCSP
8236 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8237 in the OCSP_CERTID structures of the response.
8238 [Steve Henson]
8239
8240 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8241 and related routines. This uses the standard OpenSSL certificate
8242 verify routines to perform initial checks (just CA validity) and
8243 to obtain the certificate chain. Then additional checks will be
8244 performed on the chain. Currently the root CA is checked to see
8245 if it is explicitly trusted for OCSP signing. This is used to set
8246 a root CA as a global signing root: that is any certificate that
8247 chains to that CA is an acceptable OCSP signing certificate.
8248 [Steve Henson]
8249
8250 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8251 extensions from a separate configuration file.
8252 As when reading extensions from the main configuration file,
8253 the '-extensions ...' option may be used for specifying the
8254 section to use.
8255 [Massimiliano Pala <madwolf@comune.modena.it>]
8256
8257 *) New OCSP utility. Allows OCSP requests to be generated or
8258 read. The request can be sent to a responder and the output
8259 parsed, outputed or printed in text form. Not complete yet:
8260 still needs to check the OCSP response validity.
8261 [Steve Henson]
8262
8263 *) New subcommands for 'openssl ca':
8264 'openssl ca -status <serial>' prints the status of the cert with
8265 the given serial number (according to the index file).
8266 'openssl ca -updatedb' updates the expiry status of certificates
8267 in the index file.
8268 [Massimiliano Pala <madwolf@comune.modena.it>]
8269
8270 *) New '-newreq-nodes' command option to CA.pl. This is like
8271 '-newreq', but calls 'openssl req' with the '-nodes' option
8272 so that the resulting key is not encrypted.
8273 [Damien Miller <djm@mindrot.org>]
8274
8275 *) New configuration for the GNU Hurd.
8276 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8277
8278 *) Initial code to implement OCSP basic response verify. This
8279 is currently incomplete. Currently just finds the signer's
8280 certificate and verifies the signature on the response.
8281 [Steve Henson]
8282
8283 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
8284 value of OPENSSLDIR. This is available via the new '-d' option
8285 to 'openssl version', and is also included in 'openssl version -a'.
8286 [Bodo Moeller]
8287
8288 *) Allowing defining memory allocation callbacks that will be given
8289 file name and line number information in additional arguments
8290 (a const char* and an int). The basic functionality remains, as
8291 well as the original possibility to just replace malloc(),
8292 realloc() and free() by functions that do not know about these
8293 additional arguments. To register and find out the current
8294 settings for extended allocation functions, the following
8295 functions are provided:
8296
8297 CRYPTO_set_mem_ex_functions
8298 CRYPTO_set_locked_mem_ex_functions
8299 CRYPTO_get_mem_ex_functions
8300 CRYPTO_get_locked_mem_ex_functions
8301
8302 These work the same way as CRYPTO_set_mem_functions and friends.
8303 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
8304 extended allocation function is enabled.
8305 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
8306 a conventional allocation function is enabled.
8307 [Richard Levitte, Bodo Moeller]
8308
8309 *) Finish off removing the remaining LHASH function pointer casts.
8310 There should no longer be any prototype-casting required when using
8311 the LHASH abstraction, and any casts that remain are "bugs". See
8312 the callback types and macros at the head of lhash.h for details
8313 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
8314 [Geoff Thorpe]
8315
8316 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
8317 If /dev/[u]random devices are not available or do not return enough
8318 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
8319 be queried.
8320 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
8321 /etc/entropy will be queried once each in this sequence, querying stops
8322 when enough entropy was collected without querying more sockets.
8323 [Lutz Jaenicke]
8324
8325 *) Change the Unix RAND_poll() variant to be able to poll several
8326 random devices, as specified by DEVRANDOM, until a sufficient amount
8327 of data has been collected. We spend at most 10 ms on each file
8328 (select timeout) and read in non-blocking mode. DEVRANDOM now
8329 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
8330 (previously it was just the string "/dev/urandom"), so on typical
8331 platforms the 10 ms delay will never occur.
8332 Also separate out the Unix variant to its own file, rand_unix.c.
8333 For VMS, there's a currently-empty rand_vms.c.
8334 [Richard Levitte]
8335
8336 *) Move OCSP client related routines to ocsp_cl.c. These
8337 provide utility functions which an application needing
8338 to issue a request to an OCSP responder and analyse the
8339 response will typically need: as opposed to those which an
8340 OCSP responder itself would need which will be added later.
8341
8342 OCSP_request_sign() signs an OCSP request with an API similar
8343 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
8344 response. OCSP_response_get1_basic() extracts basic response
8345 from response. OCSP_resp_find_status(): finds and extracts status
8346 information from an OCSP_CERTID structure (which will be created
8347 when the request structure is built). These are built from lower
8348 level functions which work on OCSP_SINGLERESP structures but
8349 won't normally be used unless the application wishes to examine
8350 extensions in the OCSP response for example.
8351
8352 Replace nonce routines with a pair of functions.
8353 OCSP_request_add1_nonce() adds a nonce value and optionally
8354 generates a random value. OCSP_check_nonce() checks the
8355 validity of the nonce in an OCSP response.
8356 [Steve Henson]
8357
8358 *) Change function OCSP_request_add() to OCSP_request_add0_id().
8359 This doesn't copy the supplied OCSP_CERTID and avoids the
8360 need to free up the newly created id. Change return type
8361 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
8362 This can then be used to add extensions to the request.
8363 Deleted OCSP_request_new(), since most of its functionality
8364 is now in OCSP_REQUEST_new() (and the case insensitive name
8365 clash) apart from the ability to set the request name which
8366 will be added elsewhere.
8367 [Steve Henson]
8368
8369 *) Update OCSP API. Remove obsolete extensions argument from
8370 various functions. Extensions are now handled using the new
8371 OCSP extension code. New simple OCSP HTTP function which
8372 can be used to send requests and parse the response.
8373 [Steve Henson]
8374
8375 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
8376 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
8377 uses the special reorder version of SET OF to sort the attributes
8378 and reorder them to match the encoded order. This resolves a long
8379 standing problem: a verify on a PKCS7 structure just after signing
8380 it used to fail because the attribute order did not match the
8381 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
8382 it uses the received order. This is necessary to tolerate some broken
8383 software that does not order SET OF. This is handled by encoding
8384 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
8385 to produce the required SET OF.
8386 [Steve Henson]
8387
8388 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
8389 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8390 files to get correct declarations of the ASN.1 item variables.
8391 [Richard Levitte]
8392
8393 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
8394 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
8395 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
8396 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
8397 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
8398 ASN1_ITEM and no wrapper functions.
8399 [Steve Henson]
8400
8401 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
8402 replace the old function pointer based I/O routines. Change most of
8403 the *_d2i_bio() and *_d2i_fp() functions to use these.
8404 [Steve Henson]
8405
8406 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
8407 lines, recognize more "algorithms" that can be deselected, and make
8408 it complain about algorithm deselection that isn't recognised.
8409 [Richard Levitte]
8410
8411 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
8412 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
8413 to use new functions. Add NO_ASN1_OLD which can be set to remove
8414 some old style ASN1 functions: this can be used to determine if old
8415 code will still work when these eventually go away.
8416 [Steve Henson]
8417
8418 *) New extension functions for OCSP structures, these follow the
8419 same conventions as certificates and CRLs.
8420 [Steve Henson]
8421
8422 *) New function X509V3_add1_i2d(). This automatically encodes and
8423 adds an extension. Its behaviour can be customised with various
8424 flags to append, replace or delete. Various wrappers added for
8425 certificates and CRLs.
8426 [Steve Henson]
8427
8428 *) Fix to avoid calling the underlying ASN1 print routine when
8429 an extension cannot be parsed. Correct a typo in the
8430 OCSP_SERVICELOC extension. Tidy up print OCSP format.
8431 [Steve Henson]
8432
8433 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
8434 entries for variables.
8435 [Steve Henson]
8436
8437 *) Add functionality to apps/openssl.c for detecting locking
8438 problems: As the program is single-threaded, all we have
8439 to do is register a locking callback using an array for
8440 storing which locks are currently held by the program.
8441 [Bodo Moeller]
8442
8443 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8444 SSL_get_ex_data_X509_STORE_idx(), which is used in
8445 ssl_verify_cert_chain() and thus can be called at any time
8446 during TLS/SSL handshakes so that thread-safety is essential.
8447 Unfortunately, the ex_data design is not at all suited
8448 for multi-threaded use, so it probably should be abolished.
8449 [Bodo Moeller]
8450
8451 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8452 [Broadcom, tweaked and integrated by Geoff Thorpe]
8453
8454 *) Move common extension printing code to new function
8455 X509V3_print_extensions(). Reorganise OCSP print routines and
8456 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8457 [Steve Henson]
8458
8459 *) New function X509_signature_print() to remove duplication in some
8460 print routines.
8461 [Steve Henson]
8462
8463 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8464 set (this was treated exactly the same as SET OF previously). This
8465 is used to reorder the STACK representing the structure to match the
8466 encoding. This will be used to get round a problem where a PKCS7
8467 structure which was signed could not be verified because the STACK
8468 order did not reflect the encoded order.
8469 [Steve Henson]
8470
8471 *) Reimplement the OCSP ASN1 module using the new code.
8472 [Steve Henson]
8473
8474 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
8475 for its ASN1 operations. The old style function pointers still exist
8476 for now but they will eventually go away.
8477 [Steve Henson]
8478
8479 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
8480 completely replaces the old ASN1 functionality with a table driven
8481 encoder and decoder which interprets an ASN1_ITEM structure describing
8482 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
8483 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
8484 has also been converted to the new form.
8485 [Steve Henson]
8486
8487 *) Change BN_mod_exp_recp so that negative moduli are tolerated
8488 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
8489 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
8490 for negative moduli.
8491 [Bodo Moeller]
8492
8493 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
8494 of not touching the result's sign bit.
8495 [Bodo Moeller]
8496
8497 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
8498 set.
8499 [Bodo Moeller]
8500
8501 *) Changed the LHASH code to use prototypes for callbacks, and created
8502 macros to declare and implement thin (optionally static) functions
8503 that provide type-safety and avoid function pointer casting for the
8504 type-specific callbacks.
8505 [Geoff Thorpe]
8506
8507 *) Added Kerberos Cipher Suites to be used with TLS, as written in
8508 RFC 2712.
8509 [Veers Staats <staatsvr@asc.hpc.mil>,
8510 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
8511
8512 *) Reformat the FAQ so the different questions and answers can be divided
8513 in sections depending on the subject.
8514 [Richard Levitte]
8515
8516 *) Have the zlib compression code load ZLIB.DLL dynamically under
8517 Windows.
8518 [Richard Levitte]
8519
8520 *) New function BN_mod_sqrt for computing square roots modulo a prime
8521 (using the probabilistic Tonelli-Shanks algorithm unless
8522 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
8523 be handled deterministically).
8524 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8525
8526 *) Make BN_mod_inverse faster by explicitly handling small quotients
8527 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
8528 512 bits], about 30% for larger ones [1024 or 2048 bits].)
8529 [Bodo Moeller]
8530
8531 *) New function BN_kronecker.
8532 [Bodo Moeller]
8533
8534 *) Fix BN_gcd so that it works on negative inputs; the result is
8535 positive unless both parameters are zero.
8536 Previously something reasonably close to an infinite loop was
8537 possible because numbers could be growing instead of shrinking
8538 in the implementation of Euclid's algorithm.
8539 [Bodo Moeller]
8540
8541 *) Fix BN_is_word() and BN_is_one() macros to take into account the
8542 sign of the number in question.
8543
8544 Fix BN_is_word(a,w) to work correctly for w == 0.
8545
8546 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
8547 because its test if the absolute value of 'a' equals 'w'.
8548 Note that BN_abs_is_word does *not* handle w == 0 reliably;
8549 it exists mostly for use in the implementations of BN_is_zero(),
8550 BN_is_one(), and BN_is_word().
8551 [Bodo Moeller]
8552
8553 *) New function BN_swap.
8554 [Bodo Moeller]
8555
8556 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
8557 the exponentiation functions are more likely to produce reasonable
8558 results on negative inputs.
8559 [Bodo Moeller]
8560
8561 *) Change BN_mod_mul so that the result is always non-negative.
8562 Previously, it could be negative if one of the factors was negative;
8563 I don't think anyone really wanted that behaviour.
8564 [Bodo Moeller]
8565
8566 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8567 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8568 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8569 and add new functions:
8570
8571 BN_nnmod
8572 BN_mod_sqr
8573 BN_mod_add
8574 BN_mod_add_quick
8575 BN_mod_sub
8576 BN_mod_sub_quick
8577 BN_mod_lshift1
8578 BN_mod_lshift1_quick
8579 BN_mod_lshift
8580 BN_mod_lshift_quick
8581
8582 These functions always generate non-negative results.
8583
8584 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8585 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8586
8587 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8588 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8589 be reduced modulo m.
8590 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8591
8592 #if 0
8593 The following entry accidentally appeared in the CHANGES file
8594 distributed with OpenSSL 0.9.7. The modifications described in
8595 it do *not* apply to OpenSSL 0.9.7.
8596
8597 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8598 was actually never needed) and in BN_mul(). The removal in BN_mul()
8599 required a small change in bn_mul_part_recursive() and the addition
8600 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8601 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8602 bn_sub_words() and bn_add_words() except they take arrays with
8603 differing sizes.
8604 [Richard Levitte]
8605 #endif
8606
8607 *) In 'openssl passwd', verify passwords read from the terminal
8608 unless the '-salt' option is used (which usually means that
8609 verification would just waste user's time since the resulting
8610 hash is going to be compared with some given password hash)
8611 or the new '-noverify' option is used.
8612
8613 This is an incompatible change, but it does not affect
8614 non-interactive use of 'openssl passwd' (passwords on the command
8615 line, '-stdin' option, '-in ...' option) and thus should not
8616 cause any problems.
8617 [Bodo Moeller]
8618
8619 *) Remove all references to RSAref, since there's no more need for it.
8620 [Richard Levitte]
8621
8622 *) Make DSO load along a path given through an environment variable
8623 (SHLIB_PATH) with shl_load().
8624 [Richard Levitte]
8625
8626 *) Constify the ENGINE code as a result of BIGNUM constification.
8627 Also constify the RSA code and most things related to it. In a
8628 few places, most notable in the depth of the ASN.1 code, ugly
8629 casts back to non-const were required (to be solved at a later
8630 time)
8631 [Richard Levitte]
8632
8633 *) Make it so the openssl application has all engines loaded by default.
8634 [Richard Levitte]
8635
8636 *) Constify the BIGNUM routines a little more.
8637 [Richard Levitte]
8638
8639 *) Add the following functions:
8640
8641 ENGINE_load_cswift()
8642 ENGINE_load_chil()
8643 ENGINE_load_atalla()
8644 ENGINE_load_nuron()
8645 ENGINE_load_builtin_engines()
8646
8647 That way, an application can itself choose if external engines that
8648 are built-in in OpenSSL shall ever be used or not. The benefit is
8649 that applications won't have to be linked with libdl or other dso
8650 libraries unless it's really needed.
8651
8652 Changed 'openssl engine' to load all engines on demand.
8653 Changed the engine header files to avoid the duplication of some
8654 declarations (they differed!).
8655 [Richard Levitte]
8656
8657 *) 'openssl engine' can now list capabilities.
8658 [Richard Levitte]
8659
8660 *) Better error reporting in 'openssl engine'.
8661 [Richard Levitte]
8662
8663 *) Never call load_dh_param(NULL) in s_server.
8664 [Bodo Moeller]
8665
8666 *) Add engine application. It can currently list engines by name and
8667 identity, and test if they are actually available.
8668 [Richard Levitte]
8669
8670 *) Improve RPM specification file by forcing symbolic linking and making
8671 sure the installed documentation is also owned by root.root.
8672 [Damien Miller <djm@mindrot.org>]
8673
8674 *) Give the OpenSSL applications more possibilities to make use of
8675 keys (public as well as private) handled by engines.
8676 [Richard Levitte]
8677
8678 *) Add OCSP code that comes from CertCo.
8679 [Richard Levitte]
8680
8681 *) Add VMS support for the Rijndael code.
8682 [Richard Levitte]
8683
8684 *) Added untested support for Nuron crypto accelerator.
8685 [Ben Laurie]
8686
8687 *) Add support for external cryptographic devices. This code was
8688 previously distributed separately as the "engine" branch.
8689 [Geoff Thorpe, Richard Levitte]
8690
8691 *) Rework the filename-translation in the DSO code. It is now possible to
8692 have far greater control over how a "name" is turned into a filename
8693 depending on the operating environment and any oddities about the
8694 different shared library filenames on each system.
8695 [Geoff Thorpe]
8696
8697 *) Support threads on FreeBSD-elf in Configure.
8698 [Richard Levitte]
8699
8700 *) Fix for SHA1 assembly problem with MASM: it produces
8701 warnings about corrupt line number information when assembling
8702 with debugging information. This is caused by the overlapping
8703 of two sections.
8704 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
8705
8706 *) NCONF changes.
8707 NCONF_get_number() has no error checking at all. As a replacement,
8708 NCONF_get_number_e() is defined (_e for "error checking") and is
8709 promoted strongly. The old NCONF_get_number is kept around for
8710 binary backward compatibility.
8711 Make it possible for methods to load from something other than a BIO,
8712 by providing a function pointer that is given a name instead of a BIO.
8713 For example, this could be used to load configuration data from an
8714 LDAP server.
8715 [Richard Levitte]
8716
8717 *) Fix for non blocking accept BIOs. Added new I/O special reason
8718 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
8719 with non blocking I/O was not possible because no retry code was
8720 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
8721 this case.
8722 [Steve Henson]
8723
8724 *) Added the beginnings of Rijndael support.
8725 [Ben Laurie]
8726
8727 *) Fix for bug in DirectoryString mask setting. Add support for
8728 X509_NAME_print_ex() in 'req' and X509_print_ex() function
8729 to allow certificate printing to more controllable, additional
8730 'certopt' option to 'x509' to allow new printing options to be
8731 set.
8732 [Steve Henson]
8733
8734 *) Clean old EAY MD5 hack from e_os.h.
8735 [Richard Levitte]
8736
8737 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
8738
8739 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
8740 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8741 [Joe Orton, Steve Henson]
8742
8743 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
8744
8745 *) Fix additional bug revealed by the NISCC test suite:
8746
8747 Stop bug triggering large recursion when presented with
8748 certain ASN.1 tags (CVE-2003-0851)
8749 [Steve Henson]
8750
8751 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
8752
8753 *) Fix various bugs revealed by running the NISCC test suite:
8754
8755 Stop out of bounds reads in the ASN1 code when presented with
8756 invalid tags (CVE-2003-0543 and CVE-2003-0544).
8757
8758 If verify callback ignores invalid public key errors don't try to check
8759 certificate signature with the NULL public key.
8760
8761 [Steve Henson]
8762
8763 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
8764 if the server requested one: as stated in TLS 1.0 and SSL 3.0
8765 specifications.
8766 [Steve Henson]
8767
8768 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
8769 extra data after the compression methods not only for TLS 1.0
8770 but also for SSL 3.0 (as required by the specification).
8771 [Bodo Moeller; problem pointed out by Matthias Loepfe]
8772
8773 *) Change X509_certificate_type() to mark the key as exported/exportable
8774 when it's 512 *bits* long, not 512 bytes.
8775 [Richard Levitte]
8776
8777 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8778
8779 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8780 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8781 a protocol version number mismatch like a decryption error
8782 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8783 [Bodo Moeller]
8784
8785 *) Turn on RSA blinding by default in the default implementation
8786 to avoid a timing attack. Applications that don't want it can call
8787 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8788 They would be ill-advised to do so in most cases.
8789 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
8790
8791 *) Change RSA blinding code so that it works when the PRNG is not
8792 seeded (in this case, the secret RSA exponent is abused as
8793 an unpredictable seed -- if it is not unpredictable, there
8794 is no point in blinding anyway). Make RSA blinding thread-safe
8795 by remembering the creator's thread ID in rsa->blinding and
8796 having all other threads use local one-time blinding factors
8797 (this requires more computation than sharing rsa->blinding, but
8798 avoids excessive locking; and if an RSA object is not shared
8799 between threads, blinding will still be very fast).
8800 [Bodo Moeller]
8801
8802 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
8803
8804 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8805 via timing by performing a MAC computation even if incorrect
8806 block cipher padding has been found. This is a countermeasure
8807 against active attacks where the attacker has to distinguish
8808 between bad padding and a MAC verification error. (CVE-2003-0078)
8809
8810 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8811 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8812 Martin Vuagnoux (EPFL, Ilion)]
8813
8814 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
8815
8816 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
8817 memory from it's contents. This is done with a counter that will
8818 place alternating values in each byte. This can be used to solve
8819 two issues: 1) the removal of calls to memset() by highly optimizing
8820 compilers, and 2) cleansing with other values than 0, since those can
8821 be read through on certain media, for example a swap space on disk.
8822 [Geoff Thorpe]
8823
8824 *) Bugfix: client side session caching did not work with external caching,
8825 because the session->cipher setting was not restored when reloading
8826 from the external cache. This problem was masked, when
8827 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
8828 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
8829 [Lutz Jaenicke]
8830
8831 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
8832 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
8833 [Zeev Lieber <zeev-l@yahoo.com>]
8834
8835 *) Undo an undocumented change introduced in 0.9.6e which caused
8836 repeated calls to OpenSSL_add_all_ciphers() and
8837 OpenSSL_add_all_digests() to be ignored, even after calling
8838 EVP_cleanup().
8839 [Richard Levitte]
8840
8841 *) Change the default configuration reader to deal with last line not
8842 being properly terminated.
8843 [Richard Levitte]
8844
8845 *) Change X509_NAME_cmp() so it applies the special rules on handling
8846 DN values that are of type PrintableString, as well as RDNs of type
8847 emailAddress where the value has the type ia5String.
8848 [stefank@valicert.com via Richard Levitte]
8849
8850 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
8851 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
8852 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
8853 the bitwise-OR of the two for use by the majority of applications
8854 wanting this behaviour, and update the docs. The documented
8855 behaviour and actual behaviour were inconsistent and had been
8856 changing anyway, so this is more a bug-fix than a behavioural
8857 change.
8858 [Geoff Thorpe, diagnosed by Nadav Har'El]
8859
8860 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
8861 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
8862 [Bodo Moeller]
8863
8864 *) Fix initialization code race conditions in
8865 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
8866 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
8867 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
8868 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
8869 ssl2_get_cipher_by_char(),
8870 ssl3_get_cipher_by_char().
8871 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
8872
8873 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
8874 the cached sessions are flushed, as the remove_cb() might use ex_data
8875 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
8876 (see [openssl.org #212]).
8877 [Geoff Thorpe, Lutz Jaenicke]
8878
8879 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
8880 length, instead of the encoding length to d2i_ASN1_OBJECT.
8881 [Steve Henson]
8882
8883 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
8884
8885 *) [In 0.9.6g-engine release:]
8886 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
8887 [Lynn Gazis <lgazis@rainbow.com>]
8888
8889 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
8890
8891 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
8892 and get fix the header length calculation.
8893 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
8894 Alon Kantor <alonk@checkpoint.com> (and others),
8895 Steve Henson]
8896
8897 *) Use proper error handling instead of 'assertions' in buffer
8898 overflow checks added in 0.9.6e. This prevents DoS (the
8899 assertions could call abort()).
8900 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
8901
8902 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
8903
8904 *) Add various sanity checks to asn1_get_length() to reject
8905 the ASN1 length bytes if they exceed sizeof(long), will appear
8906 negative or the content length exceeds the length of the
8907 supplied buffer.
8908 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8909
8910 *) Fix cipher selection routines: ciphers without encryption had no flags
8911 for the cipher strength set and where therefore not handled correctly
8912 by the selection routines (PR #130).
8913 [Lutz Jaenicke]
8914
8915 *) Fix EVP_dsa_sha macro.
8916 [Nils Larsch]
8917
8918 *) New option
8919 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
8920 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
8921 that was added in OpenSSL 0.9.6d.
8922
8923 As the countermeasure turned out to be incompatible with some
8924 broken SSL implementations, the new option is part of SSL_OP_ALL.
8925 SSL_OP_ALL is usually employed when compatibility with weird SSL
8926 implementations is desired (e.g. '-bugs' option to 's_client' and
8927 's_server'), so the new option is automatically set in many
8928 applications.
8929 [Bodo Moeller]
8930
8931 *) Changes in security patch:
8932
8933 Changes marked "(CHATS)" were sponsored by the Defense Advanced
8934 Research Projects Agency (DARPA) and Air Force Research Laboratory,
8935 Air Force Materiel Command, USAF, under agreement number
8936 F30602-01-2-0537.
8937
8938 *) Add various sanity checks to asn1_get_length() to reject
8939 the ASN1 length bytes if they exceed sizeof(long), will appear
8940 negative or the content length exceeds the length of the
8941 supplied buffer. (CVE-2002-0659)
8942 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8943
8944 *) Assertions for various potential buffer overflows, not known to
8945 happen in practice.
8946 [Ben Laurie (CHATS)]
8947
8948 *) Various temporary buffers to hold ASCII versions of integers were
8949 too small for 64 bit platforms. (CVE-2002-0655)
8950 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
8951
8952 *) Remote buffer overflow in SSL3 protocol - an attacker could
8953 supply an oversized session ID to a client. (CVE-2002-0656)
8954 [Ben Laurie (CHATS)]
8955
8956 *) Remote buffer overflow in SSL2 protocol - an attacker could
8957 supply an oversized client master key. (CVE-2002-0656)
8958 [Ben Laurie (CHATS)]
8959
8960 Changes between 0.9.6c and 0.9.6d [9 May 2002]
8961
8962 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
8963 encoded as NULL) with id-dsa-with-sha1.
8964 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
8965
8966 *) Check various X509_...() return values in apps/req.c.
8967 [Nils Larsch <nla@trustcenter.de>]
8968
8969 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
8970 an end-of-file condition would erroneously be flagged, when the CRLF
8971 was just at the end of a processed block. The bug was discovered when
8972 processing data through a buffering memory BIO handing the data to a
8973 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
8974 <ptsekov@syntrex.com> and Nedelcho Stanev.
8975 [Lutz Jaenicke]
8976
8977 *) Implement a countermeasure against a vulnerability recently found
8978 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
8979 before application data chunks to avoid the use of known IVs
8980 with data potentially chosen by the attacker.
8981 [Bodo Moeller]
8982
8983 *) Fix length checks in ssl3_get_client_hello().
8984 [Bodo Moeller]
8985
8986 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
8987 to prevent ssl3_read_internal() from incorrectly assuming that
8988 ssl3_read_bytes() found application data while handshake
8989 processing was enabled when in fact s->s3->in_read_app_data was
8990 merely automatically cleared during the initial handshake.
8991 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
8992
8993 *) Fix object definitions for Private and Enterprise: they were not
8994 recognized in their shortname (=lowercase) representation. Extend
8995 obj_dat.pl to issue an error when using undefined keywords instead
8996 of silently ignoring the problem (Svenning Sorensen
8997 <sss@sss.dnsalias.net>).
8998 [Lutz Jaenicke]
8999
9000 *) Fix DH_generate_parameters() so that it works for 'non-standard'
9001 generators, i.e. generators other than 2 and 5. (Previously, the
9002 code did not properly initialise the 'add' and 'rem' values to
9003 BN_generate_prime().)
9004
9005 In the new general case, we do not insist that 'generator' is
9006 actually a primitive root: This requirement is rather pointless;
9007 a generator of the order-q subgroup is just as good, if not
9008 better.
9009 [Bodo Moeller]
9010
9011 *) Map new X509 verification errors to alerts. Discovered and submitted by
9012 Tom Wu <tom@arcot.com>.
9013 [Lutz Jaenicke]
9014
9015 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
9016 returning non-zero before the data has been completely received
9017 when using non-blocking I/O.
9018 [Bodo Moeller; problem pointed out by John Hughes]
9019
9020 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
9021 [Ben Laurie, Lutz Jaenicke]
9022
9023 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
9024 Yoram Zahavi <YoramZ@gilian.com>).
9025 [Lutz Jaenicke]
9026
9027 *) Add information about CygWin 1.3 and on, and preserve proper
9028 configuration for the versions before that.
9029 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
9030
9031 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
9032 check whether we deal with a copy of a session and do not delete from
9033 the cache in this case. Problem reported by "Izhar Shoshani Levi"
9034 <izhar@checkpoint.com>.
9035 [Lutz Jaenicke]
9036
9037 *) Do not store session data into the internal session cache, if it
9038 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
9039 flag is set). Proposed by Aslam <aslam@funk.com>.
9040 [Lutz Jaenicke]
9041
9042 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
9043 value is 0.
9044 [Richard Levitte]
9045
9046 *) [In 0.9.6d-engine release:]
9047 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
9048 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
9049
9050 *) Add the configuration target linux-s390x.
9051 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
9052
9053 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
9054 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
9055 variable as an indication that a ClientHello message has been
9056 received. As the flag value will be lost between multiple
9057 invocations of ssl3_accept when using non-blocking I/O, the
9058 function may not be aware that a handshake has actually taken
9059 place, thus preventing a new session from being added to the
9060 session cache.
9061
9062 To avoid this problem, we now set s->new_session to 2 instead of
9063 using a local variable.
9064 [Lutz Jaenicke, Bodo Moeller]
9065
9066 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
9067 if the SSL_R_LENGTH_MISMATCH error is detected.
9068 [Geoff Thorpe, Bodo Moeller]
9069
9070 *) New 'shared_ldflag' column in Configure platform table.
9071 [Richard Levitte]
9072
9073 *) Fix EVP_CIPHER_mode macro.
9074 ["Dan S. Camper" <dan@bti.net>]
9075
9076 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
9077 type, we must throw them away by setting rr->length to 0.
9078 [D P Chang <dpc@qualys.com>]
9079
9080 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
9081
9082 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
9083 <Dominikus.Scherkl@biodata.com>. (The previous implementation
9084 worked incorrectly for those cases where range = 10..._2 and
9085 3*range is two bits longer than range.)
9086 [Bodo Moeller]
9087
9088 *) Only add signing time to PKCS7 structures if it is not already
9089 present.
9090 [Steve Henson]
9091
9092 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
9093 OBJ_ld_ce should be OBJ_id_ce.
9094 Also some ip-pda OIDs in crypto/objects/objects.txt were
9095 incorrect (cf. RFC 3039).
9096 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
9097
9098 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
9099 returns early because it has nothing to do.
9100 [Andy Schneider <andy.schneider@bjss.co.uk>]
9101
9102 *) [In 0.9.6c-engine release:]
9103 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
9104 [Andy Schneider <andy.schneider@bjss.co.uk>]
9105
9106 *) [In 0.9.6c-engine release:]
9107 Add support for Cryptographic Appliance's keyserver technology.
9108 (Use engine 'keyclient')
9109 [Cryptographic Appliances and Geoff Thorpe]
9110
9111 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
9112 is called via tools/c89.sh because arguments have to be
9113 rearranged (all '-L' options must appear before the first object
9114 modules).
9115 [Richard Shapiro <rshapiro@abinitio.com>]
9116
9117 *) [In 0.9.6c-engine release:]
9118 Add support for Broadcom crypto accelerator cards, backported
9119 from 0.9.7.
9120 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
9121
9122 *) [In 0.9.6c-engine release:]
9123 Add support for SureWare crypto accelerator cards from
9124 Baltimore Technologies. (Use engine 'sureware')
9125 [Baltimore Technologies and Mark Cox]
9126
9127 *) [In 0.9.6c-engine release:]
9128 Add support for crypto accelerator cards from Accelerated
9129 Encryption Processing, www.aep.ie. (Use engine 'aep')
9130 [AEP Inc. and Mark Cox]
9131
9132 *) Add a configuration entry for gcc on UnixWare.
9133 [Gary Benson <gbenson@redhat.com>]
9134
9135 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
9136 messages are stored in a single piece (fixed-length part and
9137 variable-length part combined) and fix various bugs found on the way.
9138 [Bodo Moeller]
9139
9140 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9141 instead. BIO_gethostbyname() does not know what timeouts are
9142 appropriate, so entries would stay in cache even when they have
9143 become invalid.
9144 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9145
9146 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9147 faced with a pathologically small ClientHello fragment that does
9148 not contain client_version: Instead of aborting with an error,
9149 simply choose the highest available protocol version (i.e.,
9150 TLS 1.0 unless it is disabled). In practice, ClientHello
9151 messages are never sent like this, but this change gives us
9152 strictly correct behaviour at least for TLS.
9153 [Bodo Moeller]
9154
9155 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9156 never resets s->method to s->ctx->method when called from within
9157 one of the SSL handshake functions.
9158 [Bodo Moeller; problem pointed out by Niko Baric]
9159
9160 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9161 (sent using the client's version number) if client_version is
9162 smaller than the protocol version in use. Also change
9163 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9164 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9165 the client will at least see that alert.
9166 [Bodo Moeller]
9167
9168 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9169 correctly.
9170 [Bodo Moeller]
9171
9172 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9173 client receives HelloRequest while in a handshake.
9174 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9175
9176 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9177 should end in 'break', not 'goto end' which circumvents various
9178 cleanups done in state SSL_ST_OK. But session related stuff
9179 must be disabled for SSL_ST_OK in the case that we just sent a
9180 HelloRequest.
9181
9182 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9183 before just sending a HelloRequest.
9184 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9185
9186 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9187 reveal whether illegal block cipher padding was found or a MAC
9188 verification error occurred. (Neither SSLerr() codes nor alerts
9189 are directly visible to potential attackers, but the information
9190 may leak via logfiles.)
9191
9192 Similar changes are not required for the SSL 2.0 implementation
9193 because the number of padding bytes is sent in clear for SSL 2.0,
9194 and the extra bytes are just ignored. However ssl/s2_pkt.c
9195 failed to verify that the purported number of padding bytes is in
9196 the legal range.
9197 [Bodo Moeller]
9198
9199 *) Add OpenUNIX-8 support including shared libraries
9200 (Boyd Lynn Gerber <gerberb@zenez.com>).
9201 [Lutz Jaenicke]
9202
9203 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9204 'wristwatch attack' using huge encoding parameters (cf.
9205 James H. Manger's CRYPTO 2001 paper). Note that the
9206 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9207 encoding parameters and hence was not vulnerable.
9208 [Bodo Moeller]
9209
9210 *) BN_sqr() bug fix.
9211 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9212
9213 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9214 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9215 followed by modular reduction.
9216 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9217
9218 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9219 equivalent based on BN_pseudo_rand() instead of BN_rand().
9220 [Bodo Moeller]
9221
9222 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9223 This function was broken, as the check for a new client hello message
9224 to handle SGC did not allow these large messages.
9225 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9226 [Lutz Jaenicke]
9227
9228 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9229 [Lutz Jaenicke]
9230
9231 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9232 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9233 [Lutz Jaenicke]
9234
9235 *) Rework the configuration and shared library support for Tru64 Unix.
9236 The configuration part makes use of modern compiler features and
9237 still retains old compiler behavior for those that run older versions
9238 of the OS. The shared library support part includes a variant that
9239 uses the RPATH feature, and is available through the special
9240 configuration target "alpha-cc-rpath", which will never be selected
9241 automatically.
9242 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9243
9244 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9245 with the same message size as in ssl3_get_certificate_request().
9246 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9247 messages might inadvertently be reject as too long.
9248 [Petr Lampa <lampa@fee.vutbr.cz>]
9249
9250 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9251 [Andy Polyakov]
9252
9253 *) Modified SSL library such that the verify_callback that has been set
9254 specificly for an SSL object with SSL_set_verify() is actually being
9255 used. Before the change, a verify_callback set with this function was
9256 ignored and the verify_callback() set in the SSL_CTX at the time of
9257 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9258 to allow the necessary settings.
9259 [Lutz Jaenicke]
9260
9261 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9262 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9263 done automatically (in contradiction to the requirements of the C
9264 standard). This made problems when used from OpenSSH.
9265 [Lutz Jaenicke]
9266
9267 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9268 dh->length and always used
9269
9270 BN_rand_range(priv_key, dh->p).
9271
9272 BN_rand_range() is not necessary for Diffie-Hellman, and this
9273 specific range makes Diffie-Hellman unnecessarily inefficient if
9274 dh->length (recommended exponent length) is much smaller than the
9275 length of dh->p. We could use BN_rand_range() if the order of
9276 the subgroup was stored in the DH structure, but we only have
9277 dh->length.
9278
9279 So switch back to
9280
9281 BN_rand(priv_key, l, ...)
9282
9283 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
9284 otherwise.
9285 [Bodo Moeller]
9286
9287 *) In
9288
9289 RSA_eay_public_encrypt
9290 RSA_eay_private_decrypt
9291 RSA_eay_private_encrypt (signing)
9292 RSA_eay_public_decrypt (signature verification)
9293
9294 (default implementations for RSA_public_encrypt,
9295 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
9296 always reject numbers >= n.
9297 [Bodo Moeller]
9298
9299 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
9300 to synchronize access to 'locking_thread'. This is necessary on
9301 systems where access to 'locking_thread' (an 'unsigned long'
9302 variable) is not atomic.
9303 [Bodo Moeller]
9304
9305 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
9306 *before* setting the 'crypto_lock_rand' flag. The previous code had
9307 a race condition if 0 is a valid thread ID.
9308 [Travis Vitek <vitek@roguewave.com>]
9309
9310 *) Add support for shared libraries under Irix.
9311 [Albert Chin-A-Young <china@thewrittenword.com>]
9312
9313 *) Add configuration option to build on Linux on both big-endian and
9314 little-endian MIPS.
9315 [Ralf Baechle <ralf@uni-koblenz.de>]
9316
9317 *) Add the possibility to create shared libraries on HP-UX.
9318 [Richard Levitte]
9319
9320 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
9321
9322 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
9323 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
9324 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
9325 PRNG state recovery was possible based on the output of
9326 one PRNG request appropriately sized to gain knowledge on
9327 'md' followed by enough consecutive 1-byte PRNG requests
9328 to traverse all of 'state'.
9329
9330 1. When updating 'md_local' (the current thread's copy of 'md')
9331 during PRNG output generation, hash all of the previous
9332 'md_local' value, not just the half used for PRNG output.
9333
9334 2. Make the number of bytes from 'state' included into the hash
9335 independent from the number of PRNG bytes requested.
9336
9337 The first measure alone would be sufficient to avoid
9338 Markku-Juhani's attack. (Actually it had never occurred
9339 to me that the half of 'md_local' used for chaining was the
9340 half from which PRNG output bytes were taken -- I had always
9341 assumed that the secret half would be used.) The second
9342 measure makes sure that additional data from 'state' is never
9343 mixed into 'md_local' in small portions; this heuristically
9344 further strengthens the PRNG.
9345 [Bodo Moeller]
9346
9347 *) Fix crypto/bn/asm/mips3.s.
9348 [Andy Polyakov]
9349
9350 *) When only the key is given to "enc", the IV is undefined. Print out
9351 an error message in this case.
9352 [Lutz Jaenicke]
9353
9354 *) Handle special case when X509_NAME is empty in X509 printing routines.
9355 [Steve Henson]
9356
9357 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
9358 positive and less than q.
9359 [Bodo Moeller]
9360
9361 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
9362 used: it isn't thread safe and the add_lock_callback should handle
9363 that itself.
9364 [Paul Rose <Paul.Rose@bridge.com>]
9365
9366 *) Verify that incoming data obeys the block size in
9367 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
9368 [Bodo Moeller]
9369
9370 *) Fix OAEP check.
9371 [Ulf Möller, Bodo Möller]
9372
9373 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
9374 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
9375 when fixing the server behaviour for backwards-compatible 'client
9376 hello' messages. (Note that the attack is impractical against
9377 SSL 3.0 and TLS 1.0 anyway because length and version checking
9378 means that the probability of guessing a valid ciphertext is
9379 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
9380 paper.)
9381
9382 Before 0.9.5, the countermeasure (hide the error by generating a
9383 random 'decryption result') did not work properly because
9384 ERR_clear_error() was missing, meaning that SSL_get_error() would
9385 detect the supposedly ignored error.
9386
9387 Both problems are now fixed.
9388 [Bodo Moeller]
9389
9390 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
9391 (previously it was 1024).
9392 [Bodo Moeller]
9393
9394 *) Fix for compatibility mode trust settings: ignore trust settings
9395 unless some valid trust or reject settings are present.
9396 [Steve Henson]
9397
9398 *) Fix for blowfish EVP: its a variable length cipher.
9399 [Steve Henson]
9400
9401 *) Fix various bugs related to DSA S/MIME verification. Handle missing
9402 parameters in DSA public key structures and return an error in the
9403 DSA routines if parameters are absent.
9404 [Steve Henson]
9405
9406 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
9407 in the current directory if neither $RANDFILE nor $HOME was set.
9408 RAND_file_name() in 0.9.6a returned NULL in this case. This has
9409 caused some confusion to Windows users who haven't defined $HOME.
9410 Thus RAND_file_name() is changed again: e_os.h can define a
9411 DEFAULT_HOME, which will be used if $HOME is not set.
9412 For Windows, we use "C:"; on other platforms, we still require
9413 environment variables.
9414
9415 *) Move 'if (!initialized) RAND_poll()' into regions protected by
9416 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
9417 having multiple threads call RAND_poll() concurrently.
9418 [Bodo Moeller]
9419
9420 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
9421 combination of a flag and a thread ID variable.
9422 Otherwise while one thread is in ssleay_rand_bytes (which sets the
9423 flag), *other* threads can enter ssleay_add_bytes without obeying
9424 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
9425 that they do not hold after the first thread unsets add_do_not_lock).
9426 [Bodo Moeller]
9427
9428 *) Change bctest again: '-x' expressions are not available in all
9429 versions of 'test'.
9430 [Bodo Moeller]
9431
9432 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
9433
9434 *) Fix a couple of memory leaks in PKCS7_dataDecode()
9435 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9436
9437 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9438 the default extension for executables, if any. Also, make the perl
9439 scripts that use symlink() to test if it really exists and use "cp"
9440 if it doesn't. All this made OpenSSL compilable and installable in
9441 CygWin.
9442 [Richard Levitte]
9443
9444 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9445 If SEQUENCE is length is indefinite just set c->slen to the total
9446 amount of data available.
9447 [Steve Henson, reported by shige@FreeBSD.org]
9448 [This change does not apply to 0.9.7.]
9449
9450 *) Change bctest to avoid here-documents inside command substitution
9451 (workaround for FreeBSD /bin/sh bug).
9452 For compatibility with Ultrix, avoid shell functions (introduced
9453 in the bctest version that searches along $PATH).
9454 [Bodo Moeller]
9455
9456 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9457 with des_encrypt() defined on some operating systems, like Solaris
9458 and UnixWare.
9459 [Richard Levitte]
9460
9461 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9462 On the Importance of Eliminating Errors in Cryptographic
9463 Computations, J. Cryptology 14 (2001) 2, 101-119,
9464 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9465 [Ulf Moeller]
9466
9467 *) MIPS assembler BIGNUM division bug fix.
9468 [Andy Polyakov]
9469
9470 *) Disabled incorrect Alpha assembler code.
9471 [Richard Levitte]
9472
9473 *) Fix PKCS#7 decode routines so they correctly update the length
9474 after reading an EOC for the EXPLICIT tag.
9475 [Steve Henson]
9476 [This change does not apply to 0.9.7.]
9477
9478 *) Fix bug in PKCS#12 key generation routines. This was triggered
9479 if a 3DES key was generated with a 0 initial byte. Include
9480 PKCS12_BROKEN_KEYGEN compilation option to retain the old
9481 (but broken) behaviour.
9482 [Steve Henson]
9483
9484 *) Enhance bctest to search for a working bc along $PATH and print
9485 it when found.
9486 [Tim Rice <tim@multitalents.net> via Richard Levitte]
9487
9488 *) Fix memory leaks in err.c: free err_data string if necessary;
9489 don't write to the wrong index in ERR_set_error_data.
9490 [Bodo Moeller]
9491
9492 *) Implement ssl23_peek (analogous to ssl23_read), which previously
9493 did not exist.
9494 [Bodo Moeller]
9495
9496 *) Replace rdtsc with _emit statements for VC++ version 5.
9497 [Jeremy Cooper <jeremy@baymoo.org>]
9498
9499 *) Make it possible to reuse SSLv2 sessions.
9500 [Richard Levitte]
9501
9502 *) In copy_email() check for >= 0 as a return value for
9503 X509_NAME_get_index_by_NID() since 0 is a valid index.
9504 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
9505
9506 *) Avoid coredump with unsupported or invalid public keys by checking if
9507 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
9508 PKCS7_verify() fails with non detached data.
9509 [Steve Henson]
9510
9511 *) Don't use getenv in library functions when run as setuid/setgid.
9512 New function OPENSSL_issetugid().
9513 [Ulf Moeller]
9514
9515 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
9516 due to incorrect handling of multi-threading:
9517
9518 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
9519
9520 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
9521
9522 3. Count how many times MemCheck_off() has been called so that
9523 nested use can be treated correctly. This also avoids
9524 inband-signalling in the previous code (which relied on the
9525 assumption that thread ID 0 is impossible).
9526 [Bodo Moeller]
9527
9528 *) Add "-rand" option also to s_client and s_server.
9529 [Lutz Jaenicke]
9530
9531 *) Fix CPU detection on Irix 6.x.
9532 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
9533 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9534
9535 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
9536 was empty.
9537 [Steve Henson]
9538 [This change does not apply to 0.9.7.]
9539
9540 *) Use the cached encoding of an X509_NAME structure rather than
9541 copying it. This is apparently the reason for the libsafe "errors"
9542 but the code is actually correct.
9543 [Steve Henson]
9544
9545 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
9546 Bleichenbacher's DSA attack.
9547 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
9548 to be set and top=0 forces the highest bit to be set; top=-1 is new
9549 and leaves the highest bit random.
9550 [Ulf Moeller, Bodo Moeller]
9551
9552 *) In the NCONF_...-based implementations for CONF_... queries
9553 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
9554 a temporary CONF structure with the data component set to NULL
9555 (which gives segmentation faults in lh_retrieve).
9556 Instead, use NULL for the CONF pointer in CONF_get_string and
9557 CONF_get_number (which may use environment variables) and directly
9558 return NULL from CONF_get_section.
9559 [Bodo Moeller]
9560
9561 *) Fix potential buffer overrun for EBCDIC.
9562 [Ulf Moeller]
9563
9564 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9565 keyUsage if basicConstraints absent for a CA.
9566 [Steve Henson]
9567
9568 *) Make SMIME_write_PKCS7() write mail header values with a format that
9569 is more generally accepted (no spaces before the semicolon), since
9570 some programs can't parse those values properly otherwise. Also make
9571 sure BIO's that break lines after each write do not create invalid
9572 headers.
9573 [Richard Levitte]
9574
9575 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9576 macros previously used would not encode an empty SEQUENCE OF
9577 and break the signature.
9578 [Steve Henson]
9579 [This change does not apply to 0.9.7.]
9580
9581 *) Zero the premaster secret after deriving the master secret in
9582 DH ciphersuites.
9583 [Steve Henson]
9584
9585 *) Add some EVP_add_digest_alias registrations (as found in
9586 OpenSSL_add_all_digests()) to SSL_library_init()
9587 aka OpenSSL_add_ssl_algorithms(). This provides improved
9588 compatibility with peers using X.509 certificates
9589 with unconventional AlgorithmIdentifier OIDs.
9590 [Bodo Moeller]
9591
9592 *) Fix for Irix with NO_ASM.
9593 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9594
9595 *) ./config script fixes.
9596 [Ulf Moeller, Richard Levitte]
9597
9598 *) Fix 'openssl passwd -1'.
9599 [Bodo Moeller]
9600
9601 *) Change PKCS12_key_gen_asc() so it can cope with non null
9602 terminated strings whose length is passed in the passlen
9603 parameter, for example from PEM callbacks. This was done
9604 by adding an extra length parameter to asc2uni().
9605 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9606
9607 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9608 call failed, free the DSA structure.
9609 [Bodo Moeller]
9610
9611 *) Fix to uni2asc() to cope with zero length Unicode strings.
9612 These are present in some PKCS#12 files.
9613 [Steve Henson]
9614
9615 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9616 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9617 when writing a 32767 byte record.
9618 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9619
9620 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9621 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9622
9623 (RSA objects have a reference count access to which is protected
9624 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9625 so they are meant to be shared between threads.)
9626 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9627 "Reddie, Steven" <Steven.Reddie@ca.com>]
9628
9629 *) Fix a deadlock in CRYPTO_mem_leaks().
9630 [Bodo Moeller]
9631
9632 *) Use better test patterns in bntest.
9633 [Ulf Möller]
9634
9635 *) rand_win.c fix for Borland C.
9636 [Ulf Möller]
9637
9638 *) BN_rshift bugfix for n == 0.
9639 [Bodo Moeller]
9640
9641 *) Add a 'bctest' script that checks for some known 'bc' bugs
9642 so that 'make test' does not abort just because 'bc' is broken.
9643 [Bodo Moeller]
9644
9645 *) Store verify_result within SSL_SESSION also for client side to
9646 avoid potential security hole. (Re-used sessions on the client side
9647 always resulted in verify_result==X509_V_OK, not using the original
9648 result of the server certificate verification.)
9649 [Lutz Jaenicke]
9650
9651 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9652 SSL3_RT_APPLICATION_DATA, return 0.
9653 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9654 [Bodo Moeller]
9655
9656 *) Fix SSL_peek:
9657 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9658 releases, have been re-implemented by renaming the previous
9659 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9660 and ssl3_read_internal, respectively, and adding 'peek' parameters
9661 to them. The new ssl[23]_{read,peek} functions are calls to
9662 ssl[23]_read_internal with the 'peek' flag set appropriately.
9663 A 'peek' parameter has also been added to ssl3_read_bytes, which
9664 does the actual work for ssl3_read_internal.
9665 [Bodo Moeller]
9666
9667 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9668 the method-specific "init()" handler. Also clean up ex_data after
9669 calling the method-specific "finish()" handler. Previously, this was
9670 happening the other way round.
9671 [Geoff Thorpe]
9672
9673 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9674 The previous value, 12, was not always sufficient for BN_mod_exp().
9675 [Bodo Moeller]
9676
9677 *) Make sure that shared libraries get the internal name engine with
9678 the full version number and not just 0. This should mark the
9679 shared libraries as not backward compatible. Of course, this should
9680 be changed again when we can guarantee backward binary compatibility.
9681 [Richard Levitte]
9682
9683 *) Fix typo in get_cert_by_subject() in by_dir.c
9684 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9685
9686 *) Rework the system to generate shared libraries:
9687
9688 - Make note of the expected extension for the shared libraries and
9689 if there is a need for symbolic links from for example libcrypto.so.0
9690 to libcrypto.so.0.9.7. There is extended info in Configure for
9691 that.
9692
9693 - Make as few rebuilds of the shared libraries as possible.
9694
9695 - Still avoid linking the OpenSSL programs with the shared libraries.
9696
9697 - When installing, install the shared libraries separately from the
9698 static ones.
9699 [Richard Levitte]
9700
9701 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
9702
9703 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
9704 and not in SSL_clear because the latter is also used by the
9705 accept/connect functions; previously, the settings made by
9706 SSL_set_read_ahead would be lost during the handshake.
9707 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
9708
9709 *) Correct util/mkdef.pl to be selective about disabled algorithms.
9710 Previously, it would create entries for disabled algorithms no
9711 matter what.
9712 [Richard Levitte]
9713
9714 *) Added several new manual pages for SSL_* function.
9715 [Lutz Jaenicke]
9716
9717 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
9718
9719 *) In ssl23_get_client_hello, generate an error message when faced
9720 with an initial SSL 3.0/TLS record that is too small to contain the
9721 first two bytes of the ClientHello message, i.e. client_version.
9722 (Note that this is a pathologic case that probably has never happened
9723 in real life.) The previous approach was to use the version number
9724 from the record header as a substitute; but our protocol choice
9725 should not depend on that one because it is not authenticated
9726 by the Finished messages.
9727 [Bodo Moeller]
9728
9729 *) More robust randomness gathering functions for Windows.
9730 [Jeffrey Altman <jaltman@columbia.edu>]
9731
9732 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
9733 not set then we don't setup the error code for issuer check errors
9734 to avoid possibly overwriting other errors which the callback does
9735 handle. If an application does set the flag then we assume it knows
9736 what it is doing and can handle the new informational codes
9737 appropriately.
9738 [Steve Henson]
9739
9740 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
9741 a general "ANY" type, as such it should be able to decode anything
9742 including tagged types. However it didn't check the class so it would
9743 wrongly interpret tagged types in the same way as their universal
9744 counterpart and unknown types were just rejected. Changed so that the
9745 tagged and unknown types are handled in the same way as a SEQUENCE:
9746 that is the encoding is stored intact. There is also a new type
9747 "V_ASN1_OTHER" which is used when the class is not universal, in this
9748 case we have no idea what the actual type is so we just lump them all
9749 together.
9750 [Steve Henson]
9751
9752 *) On VMS, stdout may very well lead to a file that is written to
9753 in a record-oriented fashion. That means that every write() will
9754 write a separate record, which will be read separately by the
9755 programs trying to read from it. This can be very confusing.
9756
9757 The solution is to put a BIO filter in the way that will buffer
9758 text until a linefeed is reached, and then write everything a
9759 line at a time, so every record written will be an actual line,
9760 not chunks of lines and not (usually doesn't happen, but I've
9761 seen it once) several lines in one record. BIO_f_linebuffer() is
9762 the answer.
9763
9764 Currently, it's a VMS-only method, because that's where it has
9765 been tested well enough.
9766 [Richard Levitte]
9767
9768 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
9769 it can return incorrect results.
9770 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
9771 but it was in 0.9.6-beta[12].)
9772 [Bodo Moeller]
9773
9774 *) Disable the check for content being present when verifying detached
9775 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9776 include zero length content when signing messages.
9777 [Steve Henson]
9778
9779 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9780 BIO_ctrl (for BIO pairs).
9781 [Bodo Möller]
9782
9783 *) Add DSO method for VMS.
9784 [Richard Levitte]
9785
9786 *) Bug fix: Montgomery multiplication could produce results with the
9787 wrong sign.
9788 [Ulf Möller]
9789
9790 *) Add RPM specification openssl.spec and modify it to build three
9791 packages. The default package contains applications, application
9792 documentation and run-time libraries. The devel package contains
9793 include files, static libraries and function documentation. The
9794 doc package contains the contents of the doc directory. The original
9795 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
9796 [Richard Levitte]
9797
9798 *) Add a large number of documentation files for many SSL routines.
9799 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9800
9801 *) Add a configuration entry for Sony News 4.
9802 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
9803
9804 *) Don't set the two most significant bits to one when generating a
9805 random number < q in the DSA library.
9806 [Ulf Möller]
9807
9808 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
9809 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
9810 the underlying transport is blocking) if a handshake took place.
9811 (The default behaviour is needed by applications such as s_client
9812 and s_server that use select() to determine when to use SSL_read;
9813 but for applications that know in advance when to expect data, it
9814 just makes things more complicated.)
9815 [Bodo Moeller]
9816
9817 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
9818 from EGD.
9819 [Ben Laurie]
9820
9821 *) Add a few more EBCDIC conditionals that make `req' and `x509'
9822 work better on such systems.
9823 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9824
9825 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
9826 Update PKCS12_parse() so it copies the friendlyName and the
9827 keyid to the certificates aux info.
9828 [Steve Henson]
9829
9830 *) Fix bug in PKCS7_verify() which caused an infinite loop
9831 if there was more than one signature.
9832 [Sven Uszpelkat <su@celocom.de>]
9833
9834 *) Major change in util/mkdef.pl to include extra information
9835 about each symbol, as well as presenting variables as well
9836 as functions. This change means that there's n more need
9837 to rebuild the .num files when some algorithms are excluded.
9838 [Richard Levitte]
9839
9840 *) Allow the verify time to be set by an application,
9841 rather than always using the current time.
9842 [Steve Henson]
9843
9844 *) Phase 2 verify code reorganisation. The certificate
9845 verify code now looks up an issuer certificate by a
9846 number of criteria: subject name, authority key id
9847 and key usage. It also verifies self signed certificates
9848 by the same criteria. The main comparison function is
9849 X509_check_issued() which performs these checks.
9850
9851 Lot of changes were necessary in order to support this
9852 without completely rewriting the lookup code.
9853
9854 Authority and subject key identifier are now cached.
9855
9856 The LHASH 'certs' is X509_STORE has now been replaced
9857 by a STACK_OF(X509_OBJECT). This is mainly because an
9858 LHASH can't store or retrieve multiple objects with
9859 the same hash value.
9860
9861 As a result various functions (which were all internal
9862 use only) have changed to handle the new X509_STORE
9863 structure. This will break anything that messed round
9864 with X509_STORE internally.
9865
9866 The functions X509_STORE_add_cert() now checks for an
9867 exact match, rather than just subject name.
9868
9869 The X509_STORE API doesn't directly support the retrieval
9870 of multiple certificates matching a given criteria, however
9871 this can be worked round by performing a lookup first
9872 (which will fill the cache with candidate certificates)
9873 and then examining the cache for matches. This is probably
9874 the best we can do without throwing out X509_LOOKUP
9875 entirely (maybe later...).
9876
9877 The X509_VERIFY_CTX structure has been enhanced considerably.
9878
9879 All certificate lookup operations now go via a get_issuer()
9880 callback. Although this currently uses an X509_STORE it
9881 can be replaced by custom lookups. This is a simple way
9882 to bypass the X509_STORE hackery necessary to make this
9883 work and makes it possible to use more efficient techniques
9884 in future. A very simple version which uses a simple
9885 STACK for its trusted certificate store is also provided
9886 using X509_STORE_CTX_trusted_stack().
9887
9888 The verify_cb() and verify() callbacks now have equivalents
9889 in the X509_STORE_CTX structure.
9890
9891 X509_STORE_CTX also has a 'flags' field which can be used
9892 to customise the verify behaviour.
9893 [Steve Henson]
9894
9895 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
9896 excludes S/MIME capabilities.
9897 [Steve Henson]
9898
9899 *) When a certificate request is read in keep a copy of the
9900 original encoding of the signed data and use it when outputting
9901 again. Signatures then use the original encoding rather than
9902 a decoded, encoded version which may cause problems if the
9903 request is improperly encoded.
9904 [Steve Henson]
9905
9906 *) For consistency with other BIO_puts implementations, call
9907 buffer_write(b, ...) directly in buffer_puts instead of calling
9908 BIO_write(b, ...).
9909
9910 In BIO_puts, increment b->num_write as in BIO_write.
9911 [Peter.Sylvester@EdelWeb.fr]
9912
9913 *) Fix BN_mul_word for the case where the word is 0. (We have to use
9914 BN_zero, we may not return a BIGNUM with an array consisting of
9915 words set to zero.)
9916 [Bodo Moeller]
9917
9918 *) Avoid calling abort() from within the library when problems are
9919 detected, except if preprocessor symbols have been defined
9920 (such as REF_CHECK, BN_DEBUG etc.).
9921 [Bodo Moeller]
9922
9923 *) New openssl application 'rsautl'. This utility can be
9924 used for low level RSA operations. DER public key
9925 BIO/fp routines also added.
9926 [Steve Henson]
9927
9928 *) New Configure entry and patches for compiling on QNX 4.
9929 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
9930
9931 *) A demo state-machine implementation was sponsored by
9932 Nuron (http://www.nuron.com/) and is now available in
9933 demos/state_machine.
9934 [Ben Laurie]
9935
9936 *) New options added to the 'dgst' utility for signature
9937 generation and verification.
9938 [Steve Henson]
9939
9940 *) Unrecognized PKCS#7 content types are now handled via a
9941 catch all ASN1_TYPE structure. This allows unsupported
9942 types to be stored as a "blob" and an application can
9943 encode and decode it manually.
9944 [Steve Henson]
9945
9946 *) Fix various signed/unsigned issues to make a_strex.c
9947 compile under VC++.
9948 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
9949
9950 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
9951 length if passed a buffer. ASN1_INTEGER_to_BN failed
9952 if passed a NULL BN and its argument was negative.
9953 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
9954
9955 *) Modification to PKCS#7 encoding routines to output definite
9956 length encoding. Since currently the whole structures are in
9957 memory there's not real point in using indefinite length
9958 constructed encoding. However if OpenSSL is compiled with
9959 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
9960 [Steve Henson]
9961
9962 *) Added BIO_vprintf() and BIO_vsnprintf().
9963 [Richard Levitte]
9964
9965 *) Added more prefixes to parse for in the strings written
9966 through a logging bio, to cover all the levels that are available
9967 through syslog. The prefixes are now:
9968
9969 PANIC, EMERG, EMR => LOG_EMERG
9970 ALERT, ALR => LOG_ALERT
9971 CRIT, CRI => LOG_CRIT
9972 ERROR, ERR => LOG_ERR
9973 WARNING, WARN, WAR => LOG_WARNING
9974 NOTICE, NOTE, NOT => LOG_NOTICE
9975 INFO, INF => LOG_INFO
9976 DEBUG, DBG => LOG_DEBUG
9977
9978 and as before, if none of those prefixes are present at the
9979 beginning of the string, LOG_ERR is chosen.
9980
9981 On Win32, the LOG_* levels are mapped according to this:
9982
9983 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
9984 LOG_WARNING => EVENTLOG_WARNING_TYPE
9985 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
9986
9987 [Richard Levitte]
9988
9989 *) Made it possible to reconfigure with just the configuration
9990 argument "reconf" or "reconfigure". The command line arguments
9991 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
9992 and are retrieved from there when reconfiguring.
9993 [Richard Levitte]
9994
9995 *) MD4 implemented.
9996 [Assar Westerlund <assar@sics.se>, Richard Levitte]
9997
9998 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
9999 [Richard Levitte]
10000
10001 *) The obj_dat.pl script was messing up the sorting of object
10002 names. The reason was that it compared the quoted version
10003 of strings as a result "OCSP" > "OCSP Signing" because
10004 " > SPACE. Changed script to store unquoted versions of
10005 names and add quotes on output. It was also omitting some
10006 names from the lookup table if they were given a default
10007 value (that is if SN is missing it is given the same
10008 value as LN and vice versa), these are now added on the
10009 grounds that if an object has a name we should be able to
10010 look it up. Finally added warning output when duplicate
10011 short or long names are found.
10012 [Steve Henson]
10013
10014 *) Changes needed for Tandem NSK.
10015 [Scott Uroff <scott@xypro.com>]
10016
10017 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
10018 RSA_padding_check_SSLv23(), special padding was never detected
10019 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
10020 version rollback attacks was not effective.
10021
10022 In s23_clnt.c, don't use special rollback-attack detection padding
10023 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
10024 client; similarly, in s23_srvr.c, don't do the rollback check if
10025 SSL 2.0 is the only protocol enabled in the server.
10026 [Bodo Moeller]
10027
10028 *) Make it possible to get hexdumps of unprintable data with 'openssl
10029 asn1parse'. By implication, the functions ASN1_parse_dump() and
10030 BIO_dump_indent() are added.
10031 [Richard Levitte]
10032
10033 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
10034 these print out strings and name structures based on various
10035 flags including RFC2253 support and proper handling of
10036 multibyte characters. Added options to the 'x509' utility
10037 to allow the various flags to be set.
10038 [Steve Henson]
10039
10040 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
10041 Also change the functions X509_cmp_current_time() and
10042 X509_gmtime_adj() work with an ASN1_TIME structure,
10043 this will enable certificates using GeneralizedTime in validity
10044 dates to be checked.
10045 [Steve Henson]
10046
10047 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
10048 negative public key encodings) on by default,
10049 NO_NEG_PUBKEY_BUG can be set to disable it.
10050 [Steve Henson]
10051
10052 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
10053 content octets. An i2c_ASN1_OBJECT is unnecessary because
10054 the encoding can be trivially obtained from the structure.
10055 [Steve Henson]
10056
10057 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
10058 not read locks (CRYPTO_r_[un]lock).
10059 [Bodo Moeller]
10060
10061 *) A first attempt at creating official support for shared
10062 libraries through configuration. I've kept it so the
10063 default is static libraries only, and the OpenSSL programs
10064 are always statically linked for now, but there are
10065 preparations for dynamic linking in place.
10066 This has been tested on Linux and Tru64.
10067 [Richard Levitte]
10068
10069 *) Randomness polling function for Win9x, as described in:
10070 Peter Gutmann, Software Generation of Practically Strong
10071 Random Numbers.
10072 [Ulf Möller]
10073
10074 *) Fix so PRNG is seeded in req if using an already existing
10075 DSA key.
10076 [Steve Henson]
10077
10078 *) New options to smime application. -inform and -outform
10079 allow alternative formats for the S/MIME message including
10080 PEM and DER. The -content option allows the content to be
10081 specified separately. This should allow things like Netscape
10082 form signing output easier to verify.
10083 [Steve Henson]
10084
10085 *) Fix the ASN1 encoding of tags using the 'long form'.
10086 [Steve Henson]
10087
10088 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
10089 STRING types. These convert content octets to and from the
10090 underlying type. The actual tag and length octets are
10091 already assumed to have been read in and checked. These
10092 are needed because all other string types have virtually
10093 identical handling apart from the tag. By having versions
10094 of the ASN1 functions that just operate on content octets
10095 IMPLICIT tagging can be handled properly. It also allows
10096 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
10097 and ASN1_INTEGER are identical apart from the tag.
10098 [Steve Henson]
10099
10100 *) Change the handling of OID objects as follows:
10101
10102 - New object identifiers are inserted in objects.txt, following
10103 the syntax given in objects.README.
10104 - objects.pl is used to process obj_mac.num and create a new
10105 obj_mac.h.
10106 - obj_dat.pl is used to create a new obj_dat.h, using the data in
10107 obj_mac.h.
10108
10109 This is currently kind of a hack, and the perl code in objects.pl
10110 isn't very elegant, but it works as I intended. The simplest way
10111 to check that it worked correctly is to look in obj_dat.h and
10112 check the array nid_objs and make sure the objects haven't moved
10113 around (this is important!). Additions are OK, as well as
10114 consistent name changes.
10115 [Richard Levitte]
10116
10117 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
10118 [Bodo Moeller]
10119
10120 *) Addition of the command line parameter '-rand file' to 'openssl req'.
10121 The given file adds to whatever has already been seeded into the
10122 random pool through the RANDFILE configuration file option or
10123 environment variable, or the default random state file.
10124 [Richard Levitte]
10125
10126 *) mkstack.pl now sorts each macro group into lexical order.
10127 Previously the output order depended on the order the files
10128 appeared in the directory, resulting in needless rewriting
10129 of safestack.h .
10130 [Steve Henson]
10131
10132 *) Patches to make OpenSSL compile under Win32 again. Mostly
10133 work arounds for the VC++ problem that it treats func() as
10134 func(void). Also stripped out the parts of mkdef.pl that
10135 added extra typesafe functions: these no longer exist.
10136 [Steve Henson]
10137
10138 *) Reorganisation of the stack code. The macros are now all
10139 collected in safestack.h . Each macro is defined in terms of
10140 a "stack macro" of the form SKM_<name>(type, a, b). The
10141 DEBUG_SAFESTACK is now handled in terms of function casts,
10142 this has the advantage of retaining type safety without the
10143 use of additional functions. If DEBUG_SAFESTACK is not defined
10144 then the non typesafe macros are used instead. Also modified the
10145 mkstack.pl script to handle the new form. Needs testing to see
10146 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10147 the default if no major problems. Similar behaviour for ASN1_SET_OF
10148 and PKCS12_STACK_OF.
10149 [Steve Henson]
10150
10151 *) When some versions of IIS use the 'NET' form of private key the
10152 key derivation algorithm is different. Normally MD5(password) is
10153 used as a 128 bit RC4 key. In the modified case
10154 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10155 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10156 as the old Netscape_RSA functions except they have an additional
10157 'sgckey' parameter which uses the modified algorithm. Also added
10158 an -sgckey command line option to the rsa utility. Thanks to
10159 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10160 algorithm to openssl-dev.
10161 [Steve Henson]
10162
10163 *) The evp_local.h macros were using 'c.##kname' which resulted in
10164 invalid expansion on some systems (SCO 5.0.5 for example).
10165 Corrected to 'c.kname'.
10166 [Phillip Porch <root@theporch.com>]
10167
10168 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10169 a STACK of email addresses from a certificate or request, these look
10170 in the subject name and the subject alternative name extensions and
10171 omit any duplicate addresses.
10172 [Steve Henson]
10173
10174 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10175 This makes DSA verification about 2 % faster.
10176 [Bodo Moeller]
10177
10178 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10179 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10180 plus overhead for 1024 bit moduli).
10181 This makes exponentiations about 0.5 % faster for 1024 bit
10182 exponents (as measured by "openssl speed rsa2048").
10183 [Bodo Moeller]
10184
10185 *) Rename memory handling macros to avoid conflicts with other
10186 software:
10187 Malloc => OPENSSL_malloc
10188 Malloc_locked => OPENSSL_malloc_locked
10189 Realloc => OPENSSL_realloc
10190 Free => OPENSSL_free
10191 [Richard Levitte]
10192
10193 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10194 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10195 [Bodo Moeller]
10196
10197 *) CygWin32 support.
10198 [John Jarvie <jjarvie@newsguy.com>]
10199
10200 *) The type-safe stack code has been rejigged. It is now only compiled
10201 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10202 by default all type-specific stack functions are "#define"d back to
10203 standard stack functions. This results in more streamlined output
10204 but retains the type-safety checking possibilities of the original
10205 approach.
10206 [Geoff Thorpe]
10207
10208 *) The STACK code has been cleaned up, and certain type declarations
10209 that didn't make a lot of sense have been brought in line. This has
10210 also involved a cleanup of sorts in safestack.h to more correctly
10211 map type-safe stack functions onto their plain stack counterparts.
10212 This work has also resulted in a variety of "const"ifications of
10213 lots of the code, especially "_cmp" operations which should normally
10214 be prototyped with "const" parameters anyway.
10215 [Geoff Thorpe]
10216
10217 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10218 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10219 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10220 where all of 'md' is used each time the PRNG is used, but 'state'
10221 is used only indexed by a cyclic counter. As entropy may not be
10222 well distributed from the beginning, 'md' is important as a
10223 chaining variable. However, the output function chains only half
10224 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10225 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10226 in all of 'state' being rewritten, with the new values depending
10227 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10228 [Bodo Moeller]
10229
10230 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10231 the handshake is continued after ssl_verify_cert_chain();
10232 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10233 can lead to 'unexplainable' connection aborts later.
10234 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10235
10236 *) Major EVP API cipher revision.
10237 Add hooks for extra EVP features. This allows various cipher
10238 parameters to be set in the EVP interface. Support added for variable
10239 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10240 setting of RC2 and RC5 parameters.
10241
10242 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10243 ciphers.
10244
10245 Remove lots of duplicated code from the EVP library. For example *every*
10246 cipher init() function handles the 'iv' in the same way according to the
10247 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10248 for CFB and OFB modes they zero ctx->num.
10249
10250 New functionality allows removal of S/MIME code RC2 hack.
10251
10252 Most of the routines have the same form and so can be declared in terms
10253 of macros.
10254
10255 By shifting this to the top level EVP_CipherInit() it can be removed from
10256 all individual ciphers. If the cipher wants to handle IVs or keys
10257 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10258 flags.
10259
10260 Change lots of functions like EVP_EncryptUpdate() to now return a
10261 value: although software versions of the algorithms cannot fail
10262 any installed hardware versions can.
10263 [Steve Henson]
10264
10265 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10266 this option is set, tolerate broken clients that send the negotiated
10267 protocol version number instead of the requested protocol version
10268 number.
10269 [Bodo Moeller]
10270
10271 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10272 i.e. non-zero for export ciphersuites, zero otherwise.
10273 Previous versions had this flag inverted, inconsistent with
10274 rsa_tmp_cb (..._TMP_RSA_CB).
10275 [Bodo Moeller; problem reported by Amit Chopra]
10276
10277 *) Add missing DSA library text string. Work around for some IIS
10278 key files with invalid SEQUENCE encoding.
10279 [Steve Henson]
10280
10281 *) Add a document (doc/standards.txt) that list all kinds of standards
10282 and so on that are implemented in OpenSSL.
10283 [Richard Levitte]
10284
10285 *) Enhance c_rehash script. Old version would mishandle certificates
10286 with the same subject name hash and wouldn't handle CRLs at all.
10287 Added -fingerprint option to crl utility, to support new c_rehash
10288 features.
10289 [Steve Henson]
10290
10291 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
10292 [Ulf Möller]
10293
10294 *) Fix for SSL server purpose checking. Server checking was
10295 rejecting certificates which had extended key usage present
10296 but no ssl client purpose.
10297 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
10298
10299 *) Make PKCS#12 code work with no password. The PKCS#12 spec
10300 is a little unclear about how a blank password is handled.
10301 Since the password in encoded as a BMPString with terminating
10302 double NULL a zero length password would end up as just the
10303 double NULL. However no password at all is different and is
10304 handled differently in the PKCS#12 key generation code. NS
10305 treats a blank password as zero length. MSIE treats it as no
10306 password on export: but it will try both on import. We now do
10307 the same: PKCS12_parse() tries zero length and no password if
10308 the password is set to "" or NULL (NULL is now a valid password:
10309 it wasn't before) as does the pkcs12 application.
10310 [Steve Henson]
10311
10312 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
10313 perror when PEM_read_bio_X509_REQ fails, the error message must
10314 be obtained from the error queue.
10315 [Bodo Moeller]
10316
10317 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
10318 it in ERR_remove_state if appropriate, and change ERR_get_state
10319 accordingly to avoid race conditions (this is necessary because
10320 thread_hash is no longer constant once set).
10321 [Bodo Moeller]
10322
10323 *) Bugfix for linux-elf makefile.one.
10324 [Ulf Möller]
10325
10326 *) RSA_get_default_method() will now cause a default
10327 RSA_METHOD to be chosen if one doesn't exist already.
10328 Previously this was only set during a call to RSA_new()
10329 or RSA_new_method(NULL) meaning it was possible for
10330 RSA_get_default_method() to return NULL.
10331 [Geoff Thorpe]
10332
10333 *) Added native name translation to the existing DSO code
10334 that will convert (if the flag to do so is set) filenames
10335 that are sufficiently small and have no path information
10336 into a canonical native form. Eg. "blah" converted to
10337 "libblah.so" or "blah.dll" etc.
10338 [Geoff Thorpe]
10339
10340 *) New function ERR_error_string_n(e, buf, len) which is like
10341 ERR_error_string(e, buf), but writes at most 'len' bytes
10342 including the 0 terminator. For ERR_error_string_n, 'buf'
10343 may not be NULL.
10344 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
10345
10346 *) CONF library reworked to become more general. A new CONF
10347 configuration file reader "class" is implemented as well as a
10348 new functions (NCONF_*, for "New CONF") to handle it. The now
10349 old CONF_* functions are still there, but are reimplemented to
10350 work in terms of the new functions. Also, a set of functions
10351 to handle the internal storage of the configuration data is
10352 provided to make it easier to write new configuration file
10353 reader "classes" (I can definitely see something reading a
10354 configuration file in XML format, for example), called _CONF_*,
10355 or "the configuration storage API"...
10356
10357 The new configuration file reading functions are:
10358
10359 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
10360 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
10361
10362 NCONF_default, NCONF_WIN32
10363
10364 NCONF_dump_fp, NCONF_dump_bio
10365
10366 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
10367 NCONF_new creates a new CONF object. This works in the same way
10368 as other interfaces in OpenSSL, like the BIO interface.
10369 NCONF_dump_* dump the internal storage of the configuration file,
10370 which is useful for debugging. All other functions take the same
10371 arguments as the old CONF_* functions wth the exception of the
10372 first that must be a `CONF *' instead of a `LHASH *'.
10373
10374 To make it easer to use the new classes with the old CONF_* functions,
10375 the function CONF_set_default_method is provided.
10376 [Richard Levitte]
10377
10378 *) Add '-tls1' option to 'openssl ciphers', which was already
10379 mentioned in the documentation but had not been implemented.
10380 (This option is not yet really useful because even the additional
10381 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
10382 [Bodo Moeller]
10383
10384 *) Initial DSO code added into libcrypto for letting OpenSSL (and
10385 OpenSSL-based applications) load shared libraries and bind to
10386 them in a portable way.
10387 [Geoff Thorpe, with contributions from Richard Levitte]
10388
10389 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
10390
10391 *) Make sure _lrotl and _lrotr are only used with MSVC.
10392
10393 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
10394 (the default implementation of RAND_status).
10395
10396 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
10397 to '-clrext' (= clear extensions), as intended and documented.
10398 [Bodo Moeller; inconsistency pointed out by Michael Attili
10399 <attili@amaxo.com>]
10400
10401 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
10402 was larger than the MD block size.
10403 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
10404
10405 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
10406 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
10407 using the passed key: if the passed key was a private key the result
10408 of X509_print(), for example, would be to print out all the private key
10409 components.
10410 [Steve Henson]
10411
10412 *) des_quad_cksum() byte order bug fix.
10413 [Ulf Möller, using the problem description in krb4-0.9.7, where
10414 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
10415
10416 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
10417 discouraged.
10418 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
10419
10420 *) For easily testing in shell scripts whether some command
10421 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
10422 returns with exit code 0 iff no command of the given name is available.
10423 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
10424 the output goes to stdout and nothing is printed to stderr.
10425 Additional arguments are always ignored.
10426
10427 Since for each cipher there is a command of the same name,
10428 the 'no-cipher' compilation switches can be tested this way.
10429
10430 ('openssl no-XXX' is not able to detect pseudo-commands such
10431 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
10432 [Bodo Moeller]
10433
10434 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10435 [Bodo Moeller]
10436
10437 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10438 is set; it will be thrown away anyway because each handshake creates
10439 its own key.
10440 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10441 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10442 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10443 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10444 [Bodo Moeller]
10445
10446 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10447 'Q' and 'R' lose their special meanings (quit/renegotiate).
10448 This is part of what -quiet does; unlike -quiet, -ign_eof
10449 does not suppress any output.
10450 [Richard Levitte]
10451
10452 *) Add compatibility options to the purpose and trust code. The
10453 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10454 accepts a certificate or CA, this was the previous behaviour,
10455 with all the associated security issues.
10456
10457 X509_TRUST_COMPAT is the old trust behaviour: only and
10458 automatically trust self signed roots in certificate store. A
10459 new trust setting X509_TRUST_DEFAULT is used to specify that
10460 a purpose has no associated trust setting and it should instead
10461 use the value in the default purpose.
10462 [Steve Henson]
10463
10464 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10465 and fix a memory leak.
10466 [Steve Henson]
10467
10468 *) In util/mkerr.pl (which implements 'make errors'), preserve
10469 reason strings from the previous version of the .c file, as
10470 the default to have only downcase letters (and digits) in
10471 automatically generated reasons codes is not always appropriate.
10472 [Bodo Moeller]
10473
10474 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
10475 using strerror. Previously, ERR_reason_error_string() returned
10476 library names as reason strings for SYSerr; but SYSerr is a special
10477 case where small numbers are errno values, not library numbers.
10478 [Bodo Moeller]
10479
10480 *) Add '-dsaparam' option to 'openssl dhparam' application. This
10481 converts DSA parameters into DH parameters. (When creating parameters,
10482 DSA_generate_parameters is used.)
10483 [Bodo Moeller]
10484
10485 *) Include 'length' (recommended exponent length) in C code generated
10486 by 'openssl dhparam -C'.
10487 [Bodo Moeller]
10488
10489 *) The second argument to set_label in perlasm was already being used
10490 so couldn't be used as a "file scope" flag. Moved to third argument
10491 which was free.
10492 [Steve Henson]
10493
10494 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
10495 instead of RAND_bytes for encryption IVs and salts.
10496 [Bodo Moeller]
10497
10498 *) Include RAND_status() into RAND_METHOD instead of implementing
10499 it only for md_rand.c Otherwise replacing the PRNG by calling
10500 RAND_set_rand_method would be impossible.
10501 [Bodo Moeller]
10502
10503 *) Don't let DSA_generate_key() enter an infinite loop if the random
10504 number generation fails.
10505 [Bodo Moeller]
10506
10507 *) New 'rand' application for creating pseudo-random output.
10508 [Bodo Moeller]
10509
10510 *) Added configuration support for Linux/IA64
10511 [Rolf Haberrecker <rolf@suse.de>]
10512
10513 *) Assembler module support for Mingw32.
10514 [Ulf Möller]
10515
10516 *) Shared library support for HPUX (in shlib/).
10517 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
10518
10519 *) Shared library support for Solaris gcc.
10520 [Lutz Behnke <behnke@trustcenter.de>]
10521
10522 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
10523
10524 *) PKCS7_encrypt() was adding text MIME headers twice because they
10525 were added manually and by SMIME_crlf_copy().
10526 [Steve Henson]
10527
10528 *) In bntest.c don't call BN_rand with zero bits argument.
10529 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
10530
10531 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
10532 case was implemented. This caused BN_div_recp() to fail occasionally.
10533 [Ulf Möller]
10534
10535 *) Add an optional second argument to the set_label() in the perl
10536 assembly language builder. If this argument exists and is set
10537 to 1 it signals that the assembler should use a symbol whose
10538 scope is the entire file, not just the current function. This
10539 is needed with MASM which uses the format label:: for this scope.
10540 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
10541
10542 *) Change the ASN1 types so they are typedefs by default. Before
10543 almost all types were #define'd to ASN1_STRING which was causing
10544 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
10545 for example.
10546 [Steve Henson]
10547
10548 *) Change names of new functions to the new get1/get0 naming
10549 convention: After 'get1', the caller owns a reference count
10550 and has to call ..._free; 'get0' returns a pointer to some
10551 data structure without incrementing reference counters.
10552 (Some of the existing 'get' functions increment a reference
10553 counter, some don't.)
10554 Similarly, 'set1' and 'add1' functions increase reference
10555 counters or duplicate objects.
10556 [Steve Henson]
10557
10558 *) Allow for the possibility of temp RSA key generation failure:
10559 the code used to assume it always worked and crashed on failure.
10560 [Steve Henson]
10561
10562 *) Fix potential buffer overrun problem in BIO_printf().
10563 [Ulf Möller, using public domain code by Patrick Powell; problem
10564 pointed out by David Sacerdote <das33@cornell.edu>]
10565
10566 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10567 RAND_egd() and RAND_status(). In the command line application,
10568 the EGD socket can be specified like a seed file using RANDFILE
10569 or -rand.
10570 [Ulf Möller]
10571
10572 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10573 Some CAs (e.g. Verisign) distribute certificates in this form.
10574 [Steve Henson]
10575
10576 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10577 list to exclude them. This means that no special compilation option
10578 is needed to use anonymous DH: it just needs to be included in the
10579 cipher list.
10580 [Steve Henson]
10581
10582 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10583 EVP_MD_type. The old functionality is available in a new macro called
10584 EVP_MD_md(). Change code that uses it and update docs.
10585 [Steve Henson]
10586
10587 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10588 where the 'void *' argument is replaced by a function pointer argument.
10589 Previously 'void *' was abused to point to functions, which works on
10590 many platforms, but is not correct. As these functions are usually
10591 called by macros defined in OpenSSL header files, most source code
10592 should work without changes.
10593 [Richard Levitte]
10594
10595 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10596 sections with information on -D... compiler switches used for
10597 compiling the library so that applications can see them. To enable
10598 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10599 must be defined. E.g.,
10600 #define OPENSSL_ALGORITHM_DEFINES
10601 #include <openssl/opensslconf.h>
10602 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10603 [Richard Levitte, Ulf and Bodo Möller]
10604
10605 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10606 record layer.
10607 [Bodo Moeller]
10608
10609 *) Change the 'other' type in certificate aux info to a STACK_OF
10610 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10611 the required ASN1 format: arbitrary types determined by an OID.
10612 [Steve Henson]
10613
10614 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10615 argument to 'req'. This is not because the function is newer or
10616 better than others it just uses the work 'NEW' in the certificate
10617 request header lines. Some software needs this.
10618 [Steve Henson]
10619
10620 *) Reorganise password command line arguments: now passwords can be
10621 obtained from various sources. Delete the PEM_cb function and make
10622 it the default behaviour: i.e. if the callback is NULL and the
10623 usrdata argument is not NULL interpret it as a null terminated pass
10624 phrase. If usrdata and the callback are NULL then the pass phrase
10625 is prompted for as usual.
10626 [Steve Henson]
10627
10628 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10629 the support is automatically enabled. The resulting binaries will
10630 autodetect the card and use it if present.
10631 [Ben Laurie and Compaq Inc.]
10632
10633 *) Work around for Netscape hang bug. This sends certificate request
10634 and server done in one record. Since this is perfectly legal in the
10635 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10636 the bugs/SSLv3 entry for more info.
10637 [Steve Henson]
10638
10639 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10640 [Andy Polyakov]
10641
10642 *) Add -rand argument to smime and pkcs12 applications and read/write
10643 of seed file.
10644 [Steve Henson]
10645
10646 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10647 [Bodo Moeller]
10648
10649 *) Add command line password options to the remaining applications.
10650 [Steve Henson]
10651
10652 *) Bug fix for BN_div_recp() for numerators with an even number of
10653 bits.
10654 [Ulf Möller]
10655
10656 *) More tests in bntest.c, and changed test_bn output.
10657 [Ulf Möller]
10658
10659 *) ./config recognizes MacOS X now.
10660 [Andy Polyakov]
10661
10662 *) Bug fix for BN_div() when the first words of num and divisor are
10663 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10664 [Ulf Möller]
10665
10666 *) Add support for various broken PKCS#8 formats, and command line
10667 options to produce them.
10668 [Steve Henson]
10669
10670 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10671 get temporary BIGNUMs from a BN_CTX.
10672 [Ulf Möller]
10673
10674 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10675 for p == 0.
10676 [Ulf Möller]
10677
10678 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10679 include a #define from the old name to the new. The original intent
10680 was that statically linked binaries could for example just call
10681 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10682 link with digests. This never worked because SSLeay_add_all_digests()
10683 and SSLeay_add_all_ciphers() were in the same source file so calling
10684 one would link with the other. They are now in separate source files.
10685 [Steve Henson]
10686
10687 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10688 [Steve Henson]
10689
10690 *) Use a less unusual form of the Miller-Rabin primality test (it used
10691 a binary algorithm for exponentiation integrated into the Miller-Rabin
10692 loop, our standard modexp algorithms are faster).
10693 [Bodo Moeller]
10694
10695 *) Support for the EBCDIC character set completed.
10696 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10697
10698 *) Source code cleanups: use const where appropriate, eliminate casts,
10699 use void * instead of char * in lhash.
10700 [Ulf Möller]
10701
10702 *) Bugfix: ssl3_send_server_key_exchange was not restartable
10703 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
10704 this the server could overwrite ephemeral keys that the client
10705 has already seen).
10706 [Bodo Moeller]
10707
10708 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
10709 using 50 iterations of the Rabin-Miller test.
10710
10711 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
10712 iterations of the Rabin-Miller test as required by the appendix
10713 to FIPS PUB 186[-1]) instead of DSA_is_prime.
10714 As BN_is_prime_fasttest includes trial division, DSA parameter
10715 generation becomes much faster.
10716
10717 This implies a change for the callback functions in DSA_is_prime
10718 and DSA_generate_parameters: The callback function is called once
10719 for each positive witness in the Rabin-Miller test, not just
10720 occasionally in the inner loop; and the parameters to the
10721 callback function now provide an iteration count for the outer
10722 loop rather than for the current invocation of the inner loop.
10723 DSA_generate_parameters additionally can call the callback
10724 function with an 'iteration count' of -1, meaning that a
10725 candidate has passed the trial division test (when q is generated
10726 from an application-provided seed, trial division is skipped).
10727 [Bodo Moeller]
10728
10729 *) New function BN_is_prime_fasttest that optionally does trial
10730 division before starting the Rabin-Miller test and has
10731 an additional BN_CTX * argument (whereas BN_is_prime always
10732 has to allocate at least one BN_CTX).
10733 'callback(1, -1, cb_arg)' is called when a number has passed the
10734 trial division stage.
10735 [Bodo Moeller]
10736
10737 *) Fix for bug in CRL encoding. The validity dates weren't being handled
10738 as ASN1_TIME.
10739 [Steve Henson]
10740
10741 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
10742 [Steve Henson]
10743
10744 *) New function BN_pseudo_rand().
10745 [Ulf Möller]
10746
10747 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
10748 bignum version of BN_from_montgomery() with the working code from
10749 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
10750 the comments.
10751 [Ulf Möller]
10752
10753 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
10754 made it impossible to use the same SSL_SESSION data structure in
10755 SSL2 clients in multiple threads.
10756 [Bodo Moeller]
10757
10758 *) The return value of RAND_load_file() no longer counts bytes obtained
10759 by stat(). RAND_load_file(..., -1) is new and uses the complete file
10760 to seed the PRNG (previously an explicit byte count was required).
10761 [Ulf Möller, Bodo Möller]
10762
10763 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
10764 used (char *) instead of (void *) and had casts all over the place.
10765 [Steve Henson]
10766
10767 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
10768 [Ulf Möller]
10769
10770 *) Retain source code compatibility for BN_prime_checks macro:
10771 BN_is_prime(..., BN_prime_checks, ...) now uses
10772 BN_prime_checks_for_size to determine the appropriate number of
10773 Rabin-Miller iterations.
10774 [Ulf Möller]
10775
10776 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10777 DH_CHECK_P_NOT_SAFE_PRIME.
10778 (Check if this is true? OpenPGP calls them "strong".)
10779 [Ulf Möller]
10780
10781 *) Merge the functionality of "dh" and "gendh" programs into a new program
10782 "dhparam". The old programs are retained for now but will handle DH keys
10783 (instead of parameters) in future.
10784 [Steve Henson]
10785
10786 *) Make the ciphers, s_server and s_client programs check the return values
10787 when a new cipher list is set.
10788 [Steve Henson]
10789
10790 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
10791 ciphers. Before when the 56bit ciphers were enabled the sorting was
10792 wrong.
10793
10794 The syntax for the cipher sorting has been extended to support sorting by
10795 cipher-strength (using the strength_bits hard coded in the tables).
10796 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
10797
10798 Fix a bug in the cipher-command parser: when supplying a cipher command
10799 string with an "undefined" symbol (neither command nor alphanumeric
10800 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
10801 an error is flagged.
10802
10803 Due to the strength-sorting extension, the code of the
10804 ssl_create_cipher_list() function was completely rearranged. I hope that
10805 the readability was also increased :-)
10806 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10807
10808 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
10809 for the first serial number and places 2 in the serial number file. This
10810 avoids problems when the root CA is created with serial number zero and
10811 the first user certificate has the same issuer name and serial number
10812 as the root CA.
10813 [Steve Henson]
10814
10815 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
10816 the new code. Add documentation for this stuff.
10817 [Steve Henson]
10818
10819 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
10820 X509_*() to X509at_*() on the grounds that they don't handle X509
10821 structures and behave in an analogous way to the X509v3 functions:
10822 they shouldn't be called directly but wrapper functions should be used
10823 instead.
10824
10825 So we also now have some wrapper functions that call the X509at functions
10826 when passed certificate requests. (TO DO: similar things can be done with
10827 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
10828 things. Some of these need some d2i or i2d and print functionality
10829 because they handle more complex structures.)
10830 [Steve Henson]
10831
10832 *) Add missing #ifndefs that caused missing symbols when building libssl
10833 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
10834 NO_RSA in ssl/s2*.c.
10835 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
10836
10837 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
10838 has a return value which indicates the quality of the random data
10839 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
10840 error queue. New function RAND_pseudo_bytes() generates output that is
10841 guaranteed to be unique but not unpredictable. RAND_add is like
10842 RAND_seed, but takes an extra argument for an entropy estimate
10843 (RAND_seed always assumes full entropy).
10844 [Ulf Möller]
10845
10846 *) Do more iterations of Rabin-Miller probable prime test (specifically,
10847 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
10848 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
10849 in crypto/bn/bn_prime.c for the complete table). This guarantees a
10850 false-positive rate of at most 2^-80 for random input.
10851 [Bodo Moeller]
10852
10853 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
10854 [Bodo Moeller]
10855
10856 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
10857 in the 0.9.5 release), this returns the chain
10858 from an X509_CTX structure with a dup of the stack and all
10859 the X509 reference counts upped: so the stack will exist
10860 after X509_CTX_cleanup() has been called. Modify pkcs12.c
10861 to use this.
10862
10863 Also make SSL_SESSION_print() print out the verify return
10864 code.
10865 [Steve Henson]
10866
10867 *) Add manpage for the pkcs12 command. Also change the default
10868 behaviour so MAC iteration counts are used unless the new
10869 -nomaciter option is used. This improves file security and
10870 only older versions of MSIE (4.0 for example) need it.
10871 [Steve Henson]
10872
10873 *) Honor the no-xxx Configure options when creating .DEF files.
10874 [Ulf Möller]
10875
10876 *) Add PKCS#10 attributes to field table: challengePassword,
10877 unstructuredName and unstructuredAddress. These are taken from
10878 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
10879 international characters are used.
10880
10881 More changes to X509_ATTRIBUTE code: allow the setting of types
10882 based on strings. Remove the 'loc' parameter when adding
10883 attributes because these will be a SET OF encoding which is sorted
10884 in ASN1 order.
10885 [Steve Henson]
10886
10887 *) Initial changes to the 'req' utility to allow request generation
10888 automation. This will allow an application to just generate a template
10889 file containing all the field values and have req construct the
10890 request.
10891
10892 Initial support for X509_ATTRIBUTE handling. Stacks of these are
10893 used all over the place including certificate requests and PKCS#7
10894 structures. They are currently handled manually where necessary with
10895 some primitive wrappers for PKCS#7. The new functions behave in a
10896 manner analogous to the X509 extension functions: they allow
10897 attributes to be looked up by NID and added.
10898
10899 Later something similar to the X509V3 code would be desirable to
10900 automatically handle the encoding, decoding and printing of the
10901 more complex types. The string types like challengePassword can
10902 be handled by the string table functions.
10903
10904 Also modified the multi byte string table handling. Now there is
10905 a 'global mask' which masks out certain types. The table itself
10906 can use the flag STABLE_NO_MASK to ignore the mask setting: this
10907 is useful when for example there is only one permissible type
10908 (as in countryName) and using the mask might result in no valid
10909 types at all.
10910 [Steve Henson]
10911
10912 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
10913 SSL_get_peer_finished to allow applications to obtain the latest
10914 Finished messages sent to the peer or expected from the peer,
10915 respectively. (SSL_get_peer_finished is usually the Finished message
10916 actually received from the peer, otherwise the protocol will be aborted.)
10917
10918 As the Finished message are message digests of the complete handshake
10919 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
10920 be used for external authentication procedures when the authentication
10921 provided by SSL/TLS is not desired or is not enough.
10922 [Bodo Moeller]
10923
10924 *) Enhanced support for Alpha Linux is added. Now ./config checks if
10925 the host supports BWX extension and if Compaq C is present on the
10926 $PATH. Just exploiting of the BWX extension results in 20-30%
10927 performance kick for some algorithms, e.g. DES and RC4 to mention
10928 a couple. Compaq C in turn generates ~20% faster code for MD5 and
10929 SHA1.
10930 [Andy Polyakov]
10931
10932 *) Add support for MS "fast SGC". This is arguably a violation of the
10933 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
10934 weak crypto and after checking the certificate is SGC a second one
10935 with strong crypto. MS SGC stops the first handshake after receiving
10936 the server certificate message and sends a second client hello. Since
10937 a server will typically do all the time consuming operations before
10938 expecting any further messages from the client (server key exchange
10939 is the most expensive) there is little difference between the two.
10940
10941 To get OpenSSL to support MS SGC we have to permit a second client
10942 hello message after we have sent server done. In addition we have to
10943 reset the MAC if we do get this second client hello.
10944 [Steve Henson]
10945
10946 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
10947 if a DER encoded private key is RSA or DSA traditional format. Changed
10948 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
10949 format DER encoded private key. Newer code should use PKCS#8 format which
10950 has the key type encoded in the ASN1 structure. Added DER private key
10951 support to pkcs8 application.
10952 [Steve Henson]
10953
10954 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
10955 ciphersuites has been selected (as required by the SSL 3/TLS 1
10956 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
10957 is set, we interpret this as a request to violate the specification
10958 (the worst that can happen is a handshake failure, and 'correct'
10959 behaviour would result in a handshake failure anyway).
10960 [Bodo Moeller]
10961
10962 *) In SSL_CTX_add_session, take into account that there might be multiple
10963 SSL_SESSION structures with the same session ID (e.g. when two threads
10964 concurrently obtain them from an external cache).
10965 The internal cache can handle only one SSL_SESSION with a given ID,
10966 so if there's a conflict, we now throw out the old one to achieve
10967 consistency.
10968 [Bodo Moeller]
10969
10970 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
10971 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
10972 some routines that use cipher OIDs: some ciphers do not have OIDs
10973 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
10974 example.
10975 [Steve Henson]
10976
10977 *) Simplify the trust setting structure and code. Now we just have
10978 two sequences of OIDs for trusted and rejected settings. These will
10979 typically have values the same as the extended key usage extension
10980 and any application specific purposes.
10981
10982 The trust checking code now has a default behaviour: it will just
10983 check for an object with the same NID as the passed id. Functions can
10984 be provided to override either the default behaviour or the behaviour
10985 for a given id. SSL client, server and email already have functions
10986 in place for compatibility: they check the NID and also return "trusted"
10987 if the certificate is self signed.
10988 [Steve Henson]
10989
10990 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
10991 traditional format into an EVP_PKEY structure.
10992 [Steve Henson]
10993
10994 *) Add a password callback function PEM_cb() which either prompts for
10995 a password if usr_data is NULL or otherwise assumes it is a null
10996 terminated password. Allow passwords to be passed on command line
10997 environment or config files in a few more utilities.
10998 [Steve Henson]
10999
11000 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
11001 keys. Add some short names for PKCS#8 PBE algorithms and allow them
11002 to be specified on the command line for the pkcs8 and pkcs12 utilities.
11003 Update documentation.
11004 [Steve Henson]
11005
11006 *) Support for ASN1 "NULL" type. This could be handled before by using
11007 ASN1_TYPE but there wasn't any function that would try to read a NULL
11008 and produce an error if it couldn't. For compatibility we also have
11009 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
11010 don't allocate anything because they don't need to.
11011 [Steve Henson]
11012
11013 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
11014 for details.
11015 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
11016
11017 *) Rebuild of the memory allocation routines used by OpenSSL code and
11018 possibly others as well. The purpose is to make an interface that
11019 provide hooks so anyone can build a separate set of allocation and
11020 deallocation routines to be used by OpenSSL, for example memory
11021 pool implementations, or something else, which was previously hard
11022 since Malloc(), Realloc() and Free() were defined as macros having
11023 the values malloc, realloc and free, respectively (except for Win32
11024 compilations). The same is provided for memory debugging code.
11025 OpenSSL already comes with functionality to find memory leaks, but
11026 this gives people a chance to debug other memory problems.
11027
11028 With these changes, a new set of functions and macros have appeared:
11029
11030 CRYPTO_set_mem_debug_functions() [F]
11031 CRYPTO_get_mem_debug_functions() [F]
11032 CRYPTO_dbg_set_options() [F]
11033 CRYPTO_dbg_get_options() [F]
11034 CRYPTO_malloc_debug_init() [M]
11035
11036 The memory debug functions are NULL by default, unless the library
11037 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
11038 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
11039 gives the standard debugging functions that come with OpenSSL) or
11040 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
11041 provided by the library user) must be used. When the standard
11042 debugging functions are used, CRYPTO_dbg_set_options can be used to
11043 request additional information:
11044 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
11045 the CRYPTO_MDEBUG_xxx macro when compiling the library.
11046
11047 Also, things like CRYPTO_set_mem_functions will always give the
11048 expected result (the new set of functions is used for allocation
11049 and deallocation) at all times, regardless of platform and compiler
11050 options.
11051
11052 To finish it up, some functions that were never use in any other
11053 way than through macros have a new API and new semantic:
11054
11055 CRYPTO_dbg_malloc()
11056 CRYPTO_dbg_realloc()
11057 CRYPTO_dbg_free()
11058
11059 All macros of value have retained their old syntax.
11060 [Richard Levitte and Bodo Moeller]
11061
11062 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
11063 ordering of SMIMECapabilities wasn't in "strength order" and there
11064 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
11065 algorithm.
11066 [Steve Henson]
11067
11068 *) Some ASN1 types with illegal zero length encoding (INTEGER,
11069 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
11070 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
11071
11072 *) Merge in my S/MIME library for OpenSSL. This provides a simple
11073 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
11074 functionality to handle multipart/signed properly) and a utility
11075 called 'smime' to call all this stuff. This is based on code I
11076 originally wrote for Celo who have kindly allowed it to be
11077 included in OpenSSL.
11078 [Steve Henson]
11079
11080 *) Add variants des_set_key_checked and des_set_key_unchecked of
11081 des_set_key (aka des_key_sched). Global variable des_check_key
11082 decides which of these is called by des_set_key; this way
11083 des_check_key behaves as it always did, but applications and
11084 the library itself, which was buggy for des_check_key == 1,
11085 have a cleaner way to pick the version they need.
11086 [Bodo Moeller]
11087
11088 *) New function PKCS12_newpass() which changes the password of a
11089 PKCS12 structure.
11090 [Steve Henson]
11091
11092 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
11093 dynamic mix. In both cases the ids can be used as an index into the
11094 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
11095 functions so they accept a list of the field values and the
11096 application doesn't need to directly manipulate the X509_TRUST
11097 structure.
11098 [Steve Henson]
11099
11100 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
11101 need initialising.
11102 [Steve Henson]
11103
11104 *) Modify the way the V3 extension code looks up extensions. This now
11105 works in a similar way to the object code: we have some "standard"
11106 extensions in a static table which is searched with OBJ_bsearch()
11107 and the application can add dynamic ones if needed. The file
11108 crypto/x509v3/ext_dat.h now has the info: this file needs to be
11109 updated whenever a new extension is added to the core code and kept
11110 in ext_nid order. There is a simple program 'tabtest.c' which checks
11111 this. New extensions are not added too often so this file can readily
11112 be maintained manually.
11113
11114 There are two big advantages in doing things this way. The extensions
11115 can be looked up immediately and no longer need to be "added" using
11116 X509V3_add_standard_extensions(): this function now does nothing.
11117 [Side note: I get *lots* of email saying the extension code doesn't
11118 work because people forget to call this function]
11119 Also no dynamic allocation is done unless new extensions are added:
11120 so if we don't add custom extensions there is no need to call
11121 X509V3_EXT_cleanup().
11122 [Steve Henson]
11123
11124 *) Modify enc utility's salting as follows: make salting the default. Add a
11125 magic header, so unsalted files fail gracefully instead of just decrypting
11126 to garbage. This is because not salting is a big security hole, so people
11127 should be discouraged from doing it.
11128 [Ben Laurie]
11129
11130 *) Fixes and enhancements to the 'x509' utility. It allowed a message
11131 digest to be passed on the command line but it only used this
11132 parameter when signing a certificate. Modified so all relevant
11133 operations are affected by the digest parameter including the
11134 -fingerprint and -x509toreq options. Also -x509toreq choked if a
11135 DSA key was used because it didn't fix the digest.
11136 [Steve Henson]
11137
11138 *) Initial certificate chain verify code. Currently tests the untrusted
11139 certificates for consistency with the verify purpose (which is set
11140 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11141
11142 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11143 this is because it will reject chains with invalid extensions whereas
11144 every previous version of OpenSSL and SSLeay made no checks at all.
11145
11146 Trust code: checks the root CA for the relevant trust settings. Trust
11147 settings have an initial value consistent with the verify purpose: e.g.
11148 if the verify purpose is for SSL client use it expects the CA to be
11149 trusted for SSL client use. However the default value can be changed to
11150 permit custom trust settings: one example of this would be to only trust
11151 certificates from a specific "secure" set of CAs.
11152
11153 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11154 which should be used for version portability: especially since the
11155 verify structure is likely to change more often now.
11156
11157 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11158 to set them. If not set then assume SSL clients will verify SSL servers
11159 and vice versa.
11160
11161 Two new options to the verify program: -untrusted allows a set of
11162 untrusted certificates to be passed in and -purpose which sets the
11163 intended purpose of the certificate. If a purpose is set then the
11164 new chain verify code is used to check extension consistency.
11165 [Steve Henson]
11166
11167 *) Support for the authority information access extension.
11168 [Steve Henson]
11169
11170 *) Modify RSA and DSA PEM read routines to transparently handle
11171 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11172 public keys in a format compatible with certificate
11173 SubjectPublicKeyInfo structures. Unfortunately there were already
11174 functions called *_PublicKey_* which used various odd formats so
11175 these are retained for compatibility: however the DSA variants were
11176 never in a public release so they have been deleted. Changed dsa/rsa
11177 utilities to handle the new format: note no releases ever handled public
11178 keys so we should be OK.
11179
11180 The primary motivation for this change is to avoid the same fiasco
11181 that dogs private keys: there are several incompatible private key
11182 formats some of which are standard and some OpenSSL specific and
11183 require various evil hacks to allow partial transparent handling and
11184 even then it doesn't work with DER formats. Given the option anything
11185 other than PKCS#8 should be dumped: but the other formats have to
11186 stay in the name of compatibility.
11187
11188 With public keys and the benefit of hindsight one standard format
11189 is used which works with EVP_PKEY, RSA or DSA structures: though
11190 it clearly returns an error if you try to read the wrong kind of key.
11191
11192 Added a -pubkey option to the 'x509' utility to output the public key.
11193 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11194 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11195 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11196 that do the same as the EVP_PKEY_assign_*() except they up the
11197 reference count of the added key (they don't "swallow" the
11198 supplied key).
11199 [Steve Henson]
11200
11201 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11202 CRLs would fail if the file contained no certificates or no CRLs:
11203 added a new function to read in both types and return the number
11204 read: this means that if none are read it will be an error. The
11205 DER versions of the certificate and CRL reader would always fail
11206 because it isn't possible to mix certificates and CRLs in DER format
11207 without choking one or the other routine. Changed this to just read
11208 a certificate: this is the best we can do. Also modified the code
11209 in apps/verify.c to take notice of return codes: it was previously
11210 attempting to read in certificates from NULL pointers and ignoring
11211 any errors: this is one reason why the cert and CRL reader seemed
11212 to work. It doesn't check return codes from the default certificate
11213 routines: these may well fail if the certificates aren't installed.
11214 [Steve Henson]
11215
11216 *) Code to support otherName option in GeneralName.
11217 [Steve Henson]
11218
11219 *) First update to verify code. Change the verify utility
11220 so it warns if it is passed a self signed certificate:
11221 for consistency with the normal behaviour. X509_verify
11222 has been modified to it will now verify a self signed
11223 certificate if *exactly* the same certificate appears
11224 in the store: it was previously impossible to trust a
11225 single self signed certificate. This means that:
11226 openssl verify ss.pem
11227 now gives a warning about a self signed certificate but
11228 openssl verify -CAfile ss.pem ss.pem
11229 is OK.
11230 [Steve Henson]
11231
11232 *) For servers, store verify_result in SSL_SESSION data structure
11233 (and add it to external session representation).
11234 This is needed when client certificate verifications fails,
11235 but an application-provided verification callback (set by
11236 SSL_CTX_set_cert_verify_callback) allows accepting the session
11237 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11238 but returns 1): When the session is reused, we have to set
11239 ssl->verify_result to the appropriate error code to avoid
11240 security holes.
11241 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11242
11243 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11244 case in PKCS7_dataInit() where the signed PKCS7 structure
11245 didn't contain any existing data because it was being created.
11246 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11247
11248 *) Add a salt to the key derivation routines in enc.c. This
11249 forms the first 8 bytes of the encrypted file. Also add a
11250 -S option to allow a salt to be input on the command line.
11251 [Steve Henson]
11252
11253 *) New function X509_cmp(). Oddly enough there wasn't a function
11254 to compare two certificates. We do this by working out the SHA1
11255 hash and comparing that. X509_cmp() will be needed by the trust
11256 code.
11257 [Steve Henson]
11258
11259 *) SSL_get1_session() is like SSL_get_session(), but increments
11260 the reference count in the SSL_SESSION returned.
11261 [Geoff Thorpe <geoff@eu.c2.net>]
11262
11263 *) Fix for 'req': it was adding a null to request attributes.
11264 Also change the X509_LOOKUP and X509_INFO code to handle
11265 certificate auxiliary information.
11266 [Steve Henson]
11267
11268 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11269 the 'enc' command.
11270 [Steve Henson]
11271
11272 *) Add the possibility to add extra information to the memory leak
11273 detecting output, to form tracebacks, showing from where each
11274 allocation was originated: CRYPTO_push_info("constant string") adds
11275 the string plus current file name and line number to a per-thread
11276 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11277 is like calling CYRPTO_pop_info() until the stack is empty.
11278 Also updated memory leak detection code to be multi-thread-safe.
11279 [Richard Levitte]
11280
11281 *) Add options -text and -noout to pkcs7 utility and delete the
11282 encryption options which never did anything. Update docs.
11283 [Steve Henson]
11284
11285 *) Add options to some of the utilities to allow the pass phrase
11286 to be included on either the command line (not recommended on
11287 OSes like Unix) or read from the environment. Update the
11288 manpages and fix a few bugs.
11289 [Steve Henson]
11290
11291 *) Add a few manpages for some of the openssl commands.
11292 [Steve Henson]
11293
11294 *) Fix the -revoke option in ca. It was freeing up memory twice,
11295 leaking and not finding already revoked certificates.
11296 [Steve Henson]
11297
11298 *) Extensive changes to support certificate auxiliary information.
11299 This involves the use of X509_CERT_AUX structure and X509_AUX
11300 functions. An X509_AUX function such as PEM_read_X509_AUX()
11301 can still read in a certificate file in the usual way but it
11302 will also read in any additional "auxiliary information". By
11303 doing things this way a fair degree of compatibility can be
11304 retained: existing certificates can have this information added
11305 using the new 'x509' options.
11306
11307 Current auxiliary information includes an "alias" and some trust
11308 settings. The trust settings will ultimately be used in enhanced
11309 certificate chain verification routines: currently a certificate
11310 can only be trusted if it is self signed and then it is trusted
11311 for all purposes.
11312 [Steve Henson]
11313
11314 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
11315 The problem was that one of the replacement routines had not been working
11316 since SSLeay releases. For now the offending routine has been replaced
11317 with non-optimised assembler. Even so, this now gives around 95%
11318 performance improvement for 1024 bit RSA signs.
11319 [Mark Cox]
11320
11321 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
11322 handling. Most clients have the effective key size in bits equal to
11323 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
11324 A few however don't do this and instead use the size of the decrypted key
11325 to determine the RC2 key length and the AlgorithmIdentifier to determine
11326 the effective key length. In this case the effective key length can still
11327 be 40 bits but the key length can be 168 bits for example. This is fixed
11328 by manually forcing an RC2 key into the EVP_PKEY structure because the
11329 EVP code can't currently handle unusual RC2 key sizes: it always assumes
11330 the key length and effective key length are equal.
11331 [Steve Henson]
11332
11333 *) Add a bunch of functions that should simplify the creation of
11334 X509_NAME structures. Now you should be able to do:
11335 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
11336 and have it automatically work out the correct field type and fill in
11337 the structures. The more adventurous can try:
11338 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
11339 and it will (hopefully) work out the correct multibyte encoding.
11340 [Steve Henson]
11341
11342 *) Change the 'req' utility to use the new field handling and multibyte
11343 copy routines. Before the DN field creation was handled in an ad hoc
11344 way in req, ca, and x509 which was rather broken and didn't support
11345 BMPStrings or UTF8Strings. Since some software doesn't implement
11346 BMPStrings or UTF8Strings yet, they can be enabled using the config file
11347 using the dirstring_type option. See the new comment in the default
11348 openssl.cnf for more info.
11349 [Steve Henson]
11350
11351 *) Make crypto/rand/md_rand.c more robust:
11352 - Assure unique random numbers after fork().
11353 - Make sure that concurrent threads access the global counter and
11354 md serializably so that we never lose entropy in them
11355 or use exactly the same state in multiple threads.
11356 Access to the large state is not always serializable because
11357 the additional locking could be a performance killer, and
11358 md should be large enough anyway.
11359 [Bodo Moeller]
11360
11361 *) New file apps/app_rand.c with commonly needed functionality
11362 for handling the random seed file.
11363
11364 Use the random seed file in some applications that previously did not:
11365 ca,
11366 dsaparam -genkey (which also ignored its '-rand' option),
11367 s_client,
11368 s_server,
11369 x509 (when signing).
11370 Except on systems with /dev/urandom, it is crucial to have a random
11371 seed file at least for key creation, DSA signing, and for DH exchanges;
11372 for RSA signatures we could do without one.
11373
11374 gendh and gendsa (unlike genrsa) used to read only the first byte
11375 of each file listed in the '-rand' option. The function as previously
11376 found in genrsa is now in app_rand.c and is used by all programs
11377 that support '-rand'.
11378 [Bodo Moeller]
11379
11380 *) In RAND_write_file, use mode 0600 for creating files;
11381 don't just chmod when it may be too late.
11382 [Bodo Moeller]
11383
11384 *) Report an error from X509_STORE_load_locations
11385 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
11386 [Bill Perry]
11387
11388 *) New function ASN1_mbstring_copy() this copies a string in either
11389 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
11390 into an ASN1_STRING type. A mask of permissible types is passed
11391 and it chooses the "minimal" type to use or an error if not type
11392 is suitable.
11393 [Steve Henson]
11394
11395 *) Add function equivalents to the various macros in asn1.h. The old
11396 macros are retained with an M_ prefix. Code inside the library can
11397 use the M_ macros. External code (including the openssl utility)
11398 should *NOT* in order to be "shared library friendly".
11399 [Steve Henson]
11400
11401 *) Add various functions that can check a certificate's extensions
11402 to see if it usable for various purposes such as SSL client,
11403 server or S/MIME and CAs of these types. This is currently
11404 VERY EXPERIMENTAL but will ultimately be used for certificate chain
11405 verification. Also added a -purpose flag to x509 utility to
11406 print out all the purposes.
11407 [Steve Henson]
11408
11409 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
11410 functions.
11411 [Steve Henson]
11412
11413 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
11414 for, obtain and decode and extension and obtain its critical flag.
11415 This allows all the necessary extension code to be handled in a
11416 single function call.
11417 [Steve Henson]
11418
11419 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
11420 platforms. See crypto/rc4/rc4_enc.c for further details.
11421 [Andy Polyakov]
11422
11423 *) New -noout option to asn1parse. This causes no output to be produced
11424 its main use is when combined with -strparse and -out to extract data
11425 from a file (which may not be in ASN.1 format).
11426 [Steve Henson]
11427
11428 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
11429 when producing the local key id.
11430 [Richard Levitte <levitte@stacken.kth.se>]
11431
11432 *) New option -dhparam in s_server. This allows a DH parameter file to be
11433 stated explicitly. If it is not stated then it tries the first server
11434 certificate file. The previous behaviour hard coded the filename
11435 "server.pem".
11436 [Steve Henson]
11437
11438 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11439 a public key to be input or output. For example:
11440 openssl rsa -in key.pem -pubout -out pubkey.pem
11441 Also added necessary DSA public key functions to handle this.
11442 [Steve Henson]
11443
11444 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11445 in the message. This was handled by allowing
11446 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11447 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11448
11449 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11450 to the end of the strings whereas this didn't. This would cause problems
11451 if strings read with d2i_ASN1_bytes() were later modified.
11452 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11453
11454 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11455 data and it contains EOF it will end up returning an error. This is
11456 caused by input 46 bytes long. The cause is due to the way base64
11457 BIOs find the start of base64 encoded data. They do this by trying a
11458 trial decode on each line until they find one that works. When they
11459 do a flag is set and it starts again knowing it can pass all the
11460 data directly through the decoder. Unfortunately it doesn't reset
11461 the context it uses. This means that if EOF is reached an attempt
11462 is made to pass two EOFs through the context and this causes the
11463 resulting error. This can also cause other problems as well. As is
11464 usual with these problems it takes *ages* to find and the fix is
11465 trivial: move one line.
11466 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
11467
11468 *) Ugly workaround to get s_client and s_server working under Windows. The
11469 old code wouldn't work because it needed to select() on sockets and the
11470 tty (for keypresses and to see if data could be written). Win32 only
11471 supports select() on sockets so we select() with a 1s timeout on the
11472 sockets and then see if any characters are waiting to be read, if none
11473 are present then we retry, we also assume we can always write data to
11474 the tty. This isn't nice because the code then blocks until we've
11475 received a complete line of data and it is effectively polling the
11476 keyboard at 1s intervals: however it's quite a bit better than not
11477 working at all :-) A dedicated Windows application might handle this
11478 with an event loop for example.
11479 [Steve Henson]
11480
11481 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
11482 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
11483 will be called when RSA_sign() and RSA_verify() are used. This is useful
11484 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
11485 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
11486 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
11487 This necessitated the support of an extra signature type NID_md5_sha1
11488 for SSL signatures and modifications to the SSL library to use it instead
11489 of calling RSA_public_decrypt() and RSA_private_encrypt().
11490 [Steve Henson]
11491
11492 *) Add new -verify -CAfile and -CApath options to the crl program, these
11493 will lookup a CRL issuers certificate and verify the signature in a
11494 similar way to the verify program. Tidy up the crl program so it
11495 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
11496 less strict. It will now permit CRL extensions even if it is not
11497 a V2 CRL: this will allow it to tolerate some broken CRLs.
11498 [Steve Henson]
11499
11500 *) Initialize all non-automatic variables each time one of the openssl
11501 sub-programs is started (this is necessary as they may be started
11502 multiple times from the "OpenSSL>" prompt).
11503 [Lennart Bang, Bodo Moeller]
11504
11505 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
11506 removing all other RSA functionality (this is what NO_RSA does). This
11507 is so (for example) those in the US can disable those operations covered
11508 by the RSA patent while allowing storage and parsing of RSA keys and RSA
11509 key generation.
11510 [Steve Henson]
11511
11512 *) Non-copying interface to BIO pairs.
11513 (still largely untested)
11514 [Bodo Moeller]
11515
11516 *) New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
11517 ASCII string. This was handled independently in various places before.
11518 [Steve Henson]
11519
11520 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
11521 UTF8 strings a character at a time.
11522 [Steve Henson]
11523
11524 *) Use client_version from client hello to select the protocol
11525 (s23_srvr.c) and for RSA client key exchange verification
11526 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
11527 [Bodo Moeller]
11528
11529 *) Add various utility functions to handle SPKACs, these were previously
11530 handled by poking round in the structure internals. Added new function
11531 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
11532 print, verify and generate SPKACs. Based on an original idea from
11533 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
11534 [Steve Henson]
11535
11536 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
11537 [Andy Polyakov]
11538
11539 *) Allow the config file extension section to be overwritten on the
11540 command line. Based on an original idea from Massimiliano Pala
11541 <madwolf@comune.modena.it>. The new option is called -extensions
11542 and can be applied to ca, req and x509. Also -reqexts to override
11543 the request extensions in req and -crlexts to override the crl extensions
11544 in ca.
11545 [Steve Henson]
11546
11547 *) Add new feature to the SPKAC handling in ca. Now you can include
11548 the same field multiple times by preceding it by "XXXX." for example:
11549 1.OU="Unit name 1"
11550 2.OU="Unit name 2"
11551 this is the same syntax as used in the req config file.
11552 [Steve Henson]
11553
11554 *) Allow certificate extensions to be added to certificate requests. These
11555 are specified in a 'req_extensions' option of the req section of the
11556 config file. They can be printed out with the -text option to req but
11557 are otherwise ignored at present.
11558 [Steve Henson]
11559
11560 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11561 data read consists of only the final block it would not decrypted because
11562 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11563 A misplaced 'break' also meant the decrypted final block might not be
11564 copied until the next read.
11565 [Steve Henson]
11566
11567 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11568 a few extra parameters to the DH structure: these will be useful if
11569 for example we want the value of 'q' or implement X9.42 DH.
11570 [Steve Henson]
11571
11572 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11573 provides hooks that allow the default DSA functions or functions on a
11574 "per key" basis to be replaced. This allows hardware acceleration and
11575 hardware key storage to be handled without major modification to the
11576 library. Also added low level modexp hooks and CRYPTO_EX structure and
11577 associated functions.
11578 [Steve Henson]
11579
11580 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11581 as "read only": it can't be written to and the buffer it points to will
11582 not be freed. Reading from a read only BIO is much more efficient than
11583 a normal memory BIO. This was added because there are several times when
11584 an area of memory needs to be read from a BIO. The previous method was
11585 to create a memory BIO and write the data to it, this results in two
11586 copies of the data and an O(n^2) reading algorithm. There is a new
11587 function BIO_new_mem_buf() which creates a read only memory BIO from
11588 an area of memory. Also modified the PKCS#7 routines to use read only
11589 memory BIOs.
11590 [Steve Henson]
11591
11592 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11593 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11594 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11595 but a retry condition occurred while trying to read the rest.
11596 [Bodo Moeller]
11597
11598 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11599 NID_pkcs7_encrypted by default: this was wrong since this should almost
11600 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11601 the encrypted data type: this is a more sensible place to put it and it
11602 allows the PKCS#12 code to be tidied up that duplicated this
11603 functionality.
11604 [Steve Henson]
11605
11606 *) Changed obj_dat.pl script so it takes its input and output files on
11607 the command line. This should avoid shell escape redirection problems
11608 under Win32.
11609 [Steve Henson]
11610
11611 *) Initial support for certificate extension requests, these are included
11612 in things like Xenroll certificate requests. Included functions to allow
11613 extensions to be obtained and added.
11614 [Steve Henson]
11615
11616 *) -crlf option to s_client and s_server for sending newlines as
11617 CRLF (as required by many protocols).
11618 [Bodo Moeller]
11619
11620 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11621
11622 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11623 [Ralf S. Engelschall]
11624
11625 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11626 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11627
11628 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11629 program.
11630 [Steve Henson]
11631
11632 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11633 DH parameters/keys (q is lost during that conversion, but the resulting
11634 DH parameters contain its length).
11635
11636 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11637 much faster than DH_generate_parameters (which creates parameters
11638 where p = 2*q + 1), and also the smaller q makes DH computations
11639 much more efficient (160-bit exponentiation instead of 1024-bit
11640 exponentiation); so this provides a convenient way to support DHE
11641 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11642 utter importance to use
11643 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11644 or
11645 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11646 when such DH parameters are used, because otherwise small subgroup
11647 attacks may become possible!
11648 [Bodo Moeller]
11649
11650 *) Avoid memory leak in i2d_DHparams.
11651 [Bodo Moeller]
11652
11653 *) Allow the -k option to be used more than once in the enc program:
11654 this allows the same encrypted message to be read by multiple recipients.
11655 [Steve Henson]
11656
11657 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11658 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11659 it will always use the numerical form of the OID, even if it has a short
11660 or long name.
11661 [Steve Henson]
11662
11663 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11664 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11665 otherwise bn_mod_exp was called. In the case of hardware keys for example
11666 no private key components need be present and it might store extra data
11667 in the RSA structure, which cannot be accessed from bn_mod_exp.
11668 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11669 private key operations.
11670 [Steve Henson]
11671
11672 *) Added support for SPARC Linux.
11673 [Andy Polyakov]
11674
11675 *) pem_password_cb function type incompatibly changed from
11676 typedef int pem_password_cb(char *buf, int size, int rwflag);
11677 to
11678 ....(char *buf, int size, int rwflag, void *userdata);
11679 so that applications can pass data to their callbacks:
11680 The PEM[_ASN1]_{read,write}... functions and macros now take an
11681 additional void * argument, which is just handed through whenever
11682 the password callback is called.
11683 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11684
11685 New function SSL_CTX_set_default_passwd_cb_userdata.
11686
11687 Compatibility note: As many C implementations push function arguments
11688 onto the stack in reverse order, the new library version is likely to
11689 interoperate with programs that have been compiled with the old
11690 pem_password_cb definition (PEM_whatever takes some data that
11691 happens to be on the stack as its last argument, and the callback
11692 just ignores this garbage); but there is no guarantee whatsoever that
11693 this will work.
11694
11695 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11696 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11697 problems not only on Windows, but also on some Unix platforms.
11698 To avoid problematic command lines, these definitions are now in an
11699 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11700 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
11701 [Bodo Moeller]
11702
11703 *) MIPS III/IV assembler module is reimplemented.
11704 [Andy Polyakov]
11705
11706 *) More DES library cleanups: remove references to srand/rand and
11707 delete an unused file.
11708 [Ulf Möller]
11709
11710 *) Add support for the free Netwide assembler (NASM) under Win32,
11711 since not many people have MASM (ml) and it can be hard to obtain.
11712 This is currently experimental but it seems to work OK and pass all
11713 the tests. Check out INSTALL.W32 for info.
11714 [Steve Henson]
11715
11716 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
11717 without temporary keys kept an extra copy of the server key,
11718 and connections with temporary keys did not free everything in case
11719 of an error.
11720 [Bodo Moeller]
11721
11722 *) New function RSA_check_key and new openssl rsa option -check
11723 for verifying the consistency of RSA keys.
11724 [Ulf Moeller, Bodo Moeller]
11725
11726 *) Various changes to make Win32 compile work:
11727 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
11728 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
11729 comparison" warnings.
11730 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
11731 [Steve Henson]
11732
11733 *) Add a debugging option to PKCS#5 v2 key generation function: when
11734 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
11735 derived keys are printed to stderr.
11736 [Steve Henson]
11737
11738 *) Copy the flags in ASN1_STRING_dup().
11739 [Roman E. Pavlov <pre@mo.msk.ru>]
11740
11741 *) The x509 application mishandled signing requests containing DSA
11742 keys when the signing key was also DSA and the parameters didn't match.
11743
11744 It was supposed to omit the parameters when they matched the signing key:
11745 the verifying software was then supposed to automatically use the CA's
11746 parameters if they were absent from the end user certificate.
11747
11748 Omitting parameters is no longer recommended. The test was also
11749 the wrong way round! This was probably due to unusual behaviour in
11750 EVP_cmp_parameters() which returns 1 if the parameters match.
11751 This meant that parameters were omitted when they *didn't* match and
11752 the certificate was useless. Certificates signed with 'ca' didn't have
11753 this bug.
11754 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
11755
11756 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
11757 The interface is as follows:
11758 Applications can use
11759 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
11760 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
11761 "off" is now the default.
11762 The library internally uses
11763 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
11764 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
11765 to disable memory-checking temporarily.
11766
11767 Some inconsistent states that previously were possible (and were
11768 even the default) are now avoided.
11769
11770 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
11771 with each memory chunk allocated; this is occasionally more helpful
11772 than just having a counter.
11773
11774 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11775
11776 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11777 extensions.
11778 [Bodo Moeller]
11779
11780 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11781 which largely parallels "options", but is for changing API behaviour,
11782 whereas "options" are about protocol behaviour.
11783 Initial "mode" flags are:
11784
11785 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
11786 a single record has been written.
11787 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
11788 retries use the same buffer location.
11789 (But all of the contents must be
11790 copied!)
11791 [Bodo Moeller]
11792
11793 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
11794 worked.
11795
11796 *) Fix problems with no-hmac etc.
11797 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
11798
11799 *) New functions RSA_get_default_method(), RSA_set_method() and
11800 RSA_get_method(). These allows replacement of RSA_METHODs without having
11801 to mess around with the internals of an RSA structure.
11802 [Steve Henson]
11803
11804 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
11805 Also really enable memory leak checks in openssl.c and in some
11806 test programs.
11807 [Chad C. Mulligan, Bodo Moeller]
11808
11809 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
11810 up the length of negative integers. This has now been simplified to just
11811 store the length when it is first determined and use it later, rather
11812 than trying to keep track of where data is copied and updating it to
11813 point to the end.
11814 [Steve Henson, reported by Brien Wheeler
11815 <bwheeler@authentica-security.com>]
11816
11817 *) Add a new function PKCS7_signatureVerify. This allows the verification
11818 of a PKCS#7 signature but with the signing certificate passed to the
11819 function itself. This contrasts with PKCS7_dataVerify which assumes the
11820 certificate is present in the PKCS#7 structure. This isn't always the
11821 case: certificates can be omitted from a PKCS#7 structure and be
11822 distributed by "out of band" means (such as a certificate database).
11823 [Steve Henson]
11824
11825 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
11826 function prototypes in pem.h, also change util/mkdef.pl to add the
11827 necessary function names.
11828 [Steve Henson]
11829
11830 *) mk1mf.pl (used by Windows builds) did not properly read the
11831 options set by Configure in the top level Makefile, and Configure
11832 was not even able to write more than one option correctly.
11833 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
11834 [Bodo Moeller]
11835
11836 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
11837 file to be loaded from a BIO or FILE pointer. The BIO version will
11838 for example allow memory BIOs to contain config info.
11839 [Steve Henson]
11840
11841 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
11842 Whoever hopes to achieve shared-library compatibility across versions
11843 must use this, not the compile-time macro.
11844 (Exercise 0.9.4: Which is the minimum library version required by
11845 such programs?)
11846 Note: All this applies only to multi-threaded programs, others don't
11847 need locks.
11848 [Bodo Moeller]
11849
11850 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
11851 through a BIO pair triggered the default case, i.e.
11852 SSLerr(...,SSL_R_UNKNOWN_STATE).
11853 [Bodo Moeller]
11854
11855 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
11856 can use the SSL library even if none of the specific BIOs is
11857 appropriate.
11858 [Bodo Moeller]
11859
11860 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
11861 for the encoded length.
11862 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
11863
11864 *) Add initial documentation of the X509V3 functions.
11865 [Steve Henson]
11866
11867 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
11868 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
11869 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
11870 secure PKCS#8 private key format with a high iteration count.
11871 [Steve Henson]
11872
11873 *) Fix determination of Perl interpreter: A perl or perl5
11874 _directory_ in $PATH was also accepted as the interpreter.
11875 [Ralf S. Engelschall]
11876
11877 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
11878 wrong with it but it was very old and did things like calling
11879 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
11880 unusual formatting.
11881 [Steve Henson]
11882
11883 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
11884 to use the new extension code.
11885 [Steve Henson]
11886
11887 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
11888 with macros. This should make it easier to change their form, add extra
11889 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
11890 constant.
11891 [Steve Henson]
11892
11893 *) Add to configuration table a new entry that can specify an alternative
11894 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
11895 according to Mark Crispin <MRC@Panda.COM>.
11896 [Bodo Moeller]
11897
11898 #if 0
11899 *) DES CBC did not update the IV. Weird.
11900 [Ben Laurie]
11901 #else
11902 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
11903 Changing the behaviour of the former might break existing programs --
11904 where IV updating is needed, des_ncbc_encrypt can be used.
11905 #endif
11906
11907 *) When bntest is run from "make test" it drives bc to check its
11908 calculations, as well as internally checking them. If an internal check
11909 fails, it needs to cause bc to give a non-zero result or make test carries
11910 on without noticing the failure. Fixed.
11911 [Ben Laurie]
11912
11913 *) DES library cleanups.
11914 [Ulf Möller]
11915
11916 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
11917 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
11918 ciphers. NOTE: although the key derivation function has been verified
11919 against some published test vectors it has not been extensively tested
11920 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
11921 of v2.0.
11922 [Steve Henson]
11923
11924 *) Instead of "mkdir -p", which is not fully portable, use new
11925 Perl script "util/mkdir-p.pl".
11926 [Bodo Moeller]
11927
11928 *) Rewrite the way password based encryption (PBE) is handled. It used to
11929 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
11930 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
11931 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
11932 the 'parameter' field of the AlgorithmIdentifier is passed to the
11933 underlying key generation function so it must do its own ASN1 parsing.
11934 This has also changed the EVP_PBE_CipherInit() function which now has a
11935 'parameter' argument instead of literal salt and iteration count values
11936 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
11937 [Steve Henson]
11938
11939 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
11940 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
11941 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
11942 KEY" because this clashed with PKCS#8 unencrypted string. Since this
11943 value was just used as a "magic string" and not used directly its
11944 value doesn't matter.
11945 [Steve Henson]
11946
11947 *) Introduce some semblance of const correctness to BN. Shame C doesn't
11948 support mutable.
11949 [Ben Laurie]
11950
11951 *) "linux-sparc64" configuration (ultrapenguin).
11952 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
11953 "linux-sparc" configuration.
11954 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
11955
11956 *) config now generates no-xxx options for missing ciphers.
11957 [Ulf Möller]
11958
11959 *) Support the EBCDIC character set (work in progress).
11960 File ebcdic.c not yet included because it has a different license.
11961 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11962
11963 *) Support BS2000/OSD-POSIX.
11964 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
11965
11966 *) Make callbacks for key generation use void * instead of char *.
11967 [Ben Laurie]
11968
11969 *) Make S/MIME samples compile (not yet tested).
11970 [Ben Laurie]
11971
11972 *) Additional typesafe stacks.
11973 [Ben Laurie]
11974
11975 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
11976 [Bodo Moeller]
11977
11978
11979 Changes between 0.9.3 and 0.9.3a [29 May 1999]
11980
11981 *) New configuration variant "sco5-gcc".
11982
11983 *) Updated some demos.
11984 [Sean O Riordain, Wade Scholine]
11985
11986 *) Add missing BIO_free at exit of pkcs12 application.
11987 [Wu Zhigang]
11988
11989 *) Fix memory leak in conf.c.
11990 [Steve Henson]
11991
11992 *) Updates for Win32 to assembler version of MD5.
11993 [Steve Henson]
11994
11995 *) Set #! path to perl in apps/der_chop to where we found it
11996 instead of using a fixed path.
11997 [Bodo Moeller]
11998
11999 *) SHA library changes for irix64-mips4-cc.
12000 [Andy Polyakov]
12001
12002 *) Improvements for VMS support.
12003 [Richard Levitte]
12004
12005
12006 Changes between 0.9.2b and 0.9.3 [24 May 1999]
12007
12008 *) Bignum library bug fix. IRIX 6 passes "make test" now!
12009 This also avoids the problems with SC4.2 and unpatched SC5.
12010 [Andy Polyakov <appro@fy.chalmers.se>]
12011
12012 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
12013 These are required because of the typesafe stack would otherwise break
12014 existing code. If old code used a structure member which used to be STACK
12015 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
12016 sk_num or sk_value it would produce an error because the num, data members
12017 are not present in STACK_OF. Now it just produces a warning. sk_set
12018 replaces the old method of assigning a value to sk_value
12019 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
12020 that does this will no longer work (and should use sk_set instead) but
12021 this could be regarded as a "questionable" behaviour anyway.
12022 [Steve Henson]
12023
12024 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
12025 correctly handle encrypted S/MIME data.
12026 [Steve Henson]
12027
12028 *) Change type of various DES function arguments from des_cblock
12029 (which means, in function argument declarations, pointer to char)
12030 to des_cblock * (meaning pointer to array with 8 char elements),
12031 which allows the compiler to do more typechecking; it was like
12032 that back in SSLeay, but with lots of ugly casts.
12033
12034 Introduce new type const_des_cblock.
12035 [Bodo Moeller]
12036
12037 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
12038 problems: find RecipientInfo structure that matches recipient certificate
12039 and initialise the ASN1 structures properly based on passed cipher.
12040 [Steve Henson]
12041
12042 *) Belatedly make the BN tests actually check the results.
12043 [Ben Laurie]
12044
12045 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
12046 to and from BNs: it was completely broken. New compilation option
12047 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
12048 key elements as negative integers.
12049 [Steve Henson]
12050
12051 *) Reorganize and speed up MD5.
12052 [Andy Polyakov <appro@fy.chalmers.se>]
12053
12054 *) VMS support.
12055 [Richard Levitte <richard@levitte.org>]
12056
12057 *) New option -out to asn1parse to allow the parsed structure to be
12058 output to a file. This is most useful when combined with the -strparse
12059 option to examine the output of things like OCTET STRINGS.
12060 [Steve Henson]
12061
12062 *) Make SSL library a little more fool-proof by not requiring any longer
12063 that SSL_set_{accept,connect}_state be called before
12064 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
12065 in many applications because usually everything *appeared* to work as
12066 intended anyway -- now it really works as intended).
12067 [Bodo Moeller]
12068
12069 *) Move openssl.cnf out of lib/.
12070 [Ulf Möller]
12071
12072 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
12073 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
12074 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
12075 [Ralf S. Engelschall]
12076
12077 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
12078 handle PKCS#7 enveloped data properly.
12079 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
12080
12081 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
12082 copying pointers. The cert_st handling is changed by this in
12083 various ways (and thus what used to be known as ctx->default_cert
12084 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
12085 any longer when s->cert does not give us what we need).
12086 ssl_cert_instantiate becomes obsolete by this change.
12087 As soon as we've got the new code right (possibly it already is?),
12088 we have solved a couple of bugs of the earlier code where s->cert
12089 was used as if it could not have been shared with other SSL structures.
12090
12091 Note that using the SSL API in certain dirty ways now will result
12092 in different behaviour than observed with earlier library versions:
12093 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
12094 does not influence s as it used to.
12095
12096 In order to clean up things more thoroughly, inside SSL_SESSION
12097 we don't use CERT any longer, but a new structure SESS_CERT
12098 that holds per-session data (if available); currently, this is
12099 the peer's certificate chain and, for clients, the server's certificate
12100 and temporary key. CERT holds only those values that can have
12101 meaningful defaults in an SSL_CTX.
12102 [Bodo Moeller]
12103
12104 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
12105 from the internal representation. Various PKCS#7 fixes: remove some
12106 evil casts and set the enc_dig_alg field properly based on the signing
12107 key type.
12108 [Steve Henson]
12109
12110 *) Allow PKCS#12 password to be set from the command line or the
12111 environment. Let 'ca' get its config file name from the environment
12112 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
12113 and 'x509').
12114 [Steve Henson]
12115
12116 *) Allow certificate policies extension to use an IA5STRING for the
12117 organization field. This is contrary to the PKIX definition but
12118 VeriSign uses it and IE5 only recognises this form. Document 'x509'
12119 extension option.
12120 [Steve Henson]
12121
12122 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
12123 without disallowing inline assembler and the like for non-pedantic builds.
12124 [Ben Laurie]
12125
12126 *) Support Borland C++ builder.
12127 [Janez Jere <jj@void.si>, modified by Ulf Möller]
12128
12129 *) Support Mingw32.
12130 [Ulf Möller]
12131
12132 *) SHA-1 cleanups and performance enhancements.
12133 [Andy Polyakov <appro@fy.chalmers.se>]
12134
12135 *) Sparc v8plus assembler for the bignum library.
12136 [Andy Polyakov <appro@fy.chalmers.se>]
12137
12138 *) Accept any -xxx and +xxx compiler options in Configure.
12139 [Ulf Möller]
12140
12141 *) Update HPUX configuration.
12142 [Anonymous]
12143
12144 *) Add missing sk_<type>_unshift() function to safestack.h
12145 [Ralf S. Engelschall]
12146
12147 *) New function SSL_CTX_use_certificate_chain_file that sets the
12148 "extra_cert"s in addition to the certificate. (This makes sense
12149 only for "PEM" format files, as chains as a whole are not
12150 DER-encoded.)
12151 [Bodo Moeller]
12152
12153 *) Support verify_depth from the SSL API.
12154 x509_vfy.c had what can be considered an off-by-one-error:
12155 Its depth (which was not part of the external interface)
12156 was actually counting the number of certificates in a chain;
12157 now it really counts the depth.
12158 [Bodo Moeller]
12159
12160 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12161 instead of X509err, which often resulted in confusing error
12162 messages since the error codes are not globally unique
12163 (e.g. an alleged error in ssl3_accept when a certificate
12164 didn't match the private key).
12165
12166 *) New function SSL_CTX_set_session_id_context that allows to set a default
12167 value (so that you don't need SSL_set_session_id_context for each
12168 connection using the SSL_CTX).
12169 [Bodo Moeller]
12170
12171 *) OAEP decoding bug fix.
12172 [Ulf Möller]
12173
12174 *) Support INSTALL_PREFIX for package builders, as proposed by
12175 David Harris.
12176 [Bodo Moeller]
12177
12178 *) New Configure options "threads" and "no-threads". For systems
12179 where the proper compiler options are known (currently Solaris
12180 and Linux), "threads" is the default.
12181 [Bodo Moeller]
12182
12183 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12184 [Bodo Moeller]
12185
12186 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12187 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12188 such as /usr/local/bin.
12189 [Bodo Moeller]
12190
12191 *) "make linux-shared" to build shared libraries.
12192 [Niels Poppe <niels@netbox.org>]
12193
12194 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12195 [Ulf Möller]
12196
12197 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12198 extension adding in x509 utility.
12199 [Steve Henson]
12200
12201 *) Remove NOPROTO sections and error code comments.
12202 [Ulf Möller]
12203
12204 *) Partial rewrite of the DEF file generator to now parse the ANSI
12205 prototypes.
12206 [Steve Henson]
12207
12208 *) New Configure options --prefix=DIR and --openssldir=DIR.
12209 [Ulf Möller]
12210
12211 *) Complete rewrite of the error code script(s). It is all now handled
12212 by one script at the top level which handles error code gathering,
12213 header rewriting and C source file generation. It should be much better
12214 than the old method: it now uses a modified version of Ulf's parser to
12215 read the ANSI prototypes in all header files (thus the old K&R definitions
12216 aren't needed for error creation any more) and do a better job of
12217 translating function codes into names. The old 'ASN1 error code imbedded
12218 in a comment' is no longer necessary and it doesn't use .err files which
12219 have now been deleted. Also the error code call doesn't have to appear all
12220 on one line (which resulted in some large lines...).
12221 [Steve Henson]
12222
12223 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12224 [Bodo Moeller]
12225
12226 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12227 0 (which usually indicates a closed connection), but continue reading.
12228 [Bodo Moeller]
12229
12230 *) Fix some race conditions.
12231 [Bodo Moeller]
12232
12233 *) Add support for CRL distribution points extension. Add Certificate
12234 Policies and CRL distribution points documentation.
12235 [Steve Henson]
12236
12237 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12238 [Ulf Möller]
12239
12240 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12241 8 of keying material. Merlin has also confirmed interop with this fix
12242 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12243 [Merlin Hughes <merlin@baltimore.ie>]
12244
12245 *) Fix lots of warnings.
12246 [Richard Levitte <levitte@stacken.kth.se>]
12247
12248 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12249 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12250 [Richard Levitte <levitte@stacken.kth.se>]
12251
12252 *) Fix problems with sizeof(long) == 8.
12253 [Andy Polyakov <appro@fy.chalmers.se>]
12254
12255 *) Change functions to ANSI C.
12256 [Ulf Möller]
12257
12258 *) Fix typos in error codes.
12259 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12260
12261 *) Remove defunct assembler files from Configure.
12262 [Ulf Möller]
12263
12264 *) SPARC v8 assembler BIGNUM implementation.
12265 [Andy Polyakov <appro@fy.chalmers.se>]
12266
12267 *) Support for Certificate Policies extension: both print and set.
12268 Various additions to support the r2i method this uses.
12269 [Steve Henson]
12270
12271 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12272 return a const string when you are expecting an allocated buffer.
12273 [Ben Laurie]
12274
12275 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12276 types DirectoryString and DisplayText.
12277 [Steve Henson]
12278
12279 *) Add code to allow r2i extensions to access the configuration database,
12280 add an LHASH database driver and add several ctx helper functions.
12281 [Steve Henson]
12282
12283 *) Fix an evil bug in bn_expand2() which caused various BN functions to
12284 fail when they extended the size of a BIGNUM.
12285 [Steve Henson]
12286
12287 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
12288 support typesafe stack.
12289 [Steve Henson]
12290
12291 *) Fix typo in SSL_[gs]et_options().
12292 [Nils Frostberg <nils@medcom.se>]
12293
12294 *) Delete various functions and files that belonged to the (now obsolete)
12295 old X509V3 handling code.
12296 [Steve Henson]
12297
12298 *) New Configure option "rsaref".
12299 [Ulf Möller]
12300
12301 *) Don't auto-generate pem.h.
12302 [Bodo Moeller]
12303
12304 *) Introduce type-safe ASN.1 SETs.
12305 [Ben Laurie]
12306
12307 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
12308 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
12309
12310 *) Introduce type-safe STACKs. This will almost certainly break lots of code
12311 that links with OpenSSL (well at least cause lots of warnings), but fear
12312 not: the conversion is trivial, and it eliminates loads of evil casts. A
12313 few STACKed things have been converted already. Feel free to convert more.
12314 In the fullness of time, I'll do away with the STACK type altogether.
12315 [Ben Laurie]
12316
12317 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
12318 specified in <certfile> by updating the entry in the index.txt file.
12319 This way one no longer has to edit the index.txt file manually for
12320 revoking a certificate. The -revoke option does the gory details now.
12321 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
12322
12323 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
12324 `-text' option at all and this way the `-noout -text' combination was
12325 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
12326 [Ralf S. Engelschall]
12327
12328 *) Make sure a corresponding plain text error message exists for the
12329 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
12330 verify callback function determined that a certificate was revoked.
12331 [Ralf S. Engelschall]
12332
12333 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
12334 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
12335 all available ciphers including rc5, which was forgotten until now.
12336 In order to let the testing shell script know which algorithms
12337 are available, a new (up to now undocumented) command
12338 "openssl list-cipher-commands" is used.
12339 [Bodo Moeller]
12340
12341 *) Bugfix: s_client occasionally would sleep in select() when
12342 it should have checked SSL_pending() first.
12343 [Bodo Moeller]
12344
12345 *) New functions DSA_do_sign and DSA_do_verify to provide access to
12346 the raw DSA values prior to ASN.1 encoding.
12347 [Ulf Möller]
12348
12349 *) Tweaks to Configure
12350 [Niels Poppe <niels@netbox.org>]
12351
12352 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
12353 yet...
12354 [Steve Henson]
12355
12356 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
12357 [Ulf Möller]
12358
12359 *) New config option to avoid instructions that are illegal on the 80386.
12360 The default code is faster, but requires at least a 486.
12361 [Ulf Möller]
12362
12363 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
12364 SSL2_SERVER_VERSION (not used at all) macros, which are now the
12365 same as SSL2_VERSION anyway.
12366 [Bodo Moeller]
12367
12368 *) New "-showcerts" option for s_client.
12369 [Bodo Moeller]
12370
12371 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
12372 application. Various cleanups and fixes.
12373 [Steve Henson]
12374
12375 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
12376 modify error routines to work internally. Add error codes and PBE init
12377 to library startup routines.
12378 [Steve Henson]
12379
12380 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
12381 packing functions to asn1 and evp. Changed function names and error
12382 codes along the way.
12383 [Steve Henson]
12384
12385 *) PKCS12 integration: and so it begins... First of several patches to
12386 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
12387 objects to objects.h
12388 [Steve Henson]
12389
12390 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
12391 and display support for Thawte strong extranet extension.
12392 [Steve Henson]
12393
12394 *) Add LinuxPPC support.
12395 [Jeff Dubrule <igor@pobox.org>]
12396
12397 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
12398 bn_div_words in alpha.s.
12399 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
12400
12401 *) Make sure the RSA OAEP test is skipped under -DRSAref because
12402 OAEP isn't supported when OpenSSL is built with RSAref.
12403 [Ulf Moeller <ulf@fitug.de>]
12404
12405 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
12406 so they no longer are missing under -DNOPROTO.
12407 [Soren S. Jorvang <soren@t.dk>]
12408
12409
12410 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
12411
12412 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
12413 doesn't work when the session is reused. Coming soon!
12414 [Ben Laurie]
12415
12416 *) Fix a security hole, that allows sessions to be reused in the wrong
12417 context thus bypassing client cert protection! All software that uses
12418 client certs and session caches in multiple contexts NEEDS PATCHING to
12419 allow session reuse! A fuller solution is in the works.
12420 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
12421
12422 *) Some more source tree cleanups (removed obsolete files
12423 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
12424 permission on "config" script to be executable) and a fix for the INSTALL
12425 document.
12426 [Ulf Moeller <ulf@fitug.de>]
12427
12428 *) Remove some legacy and erroneous uses of malloc, free instead of
12429 Malloc, Free.
12430 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
12431
12432 *) Make rsa_oaep_test return non-zero on error.
12433 [Ulf Moeller <ulf@fitug.de>]
12434
12435 *) Add support for native Solaris shared libraries. Configure
12436 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12437 if someone would make that last step automatic.
12438 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12439
12440 *) ctx_size was not built with the right compiler during "make links". Fixed.
12441 [Ben Laurie]
12442
12443 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12444 except NULL ciphers". This means the default cipher list will no longer
12445 enable NULL ciphers. They need to be specifically enabled e.g. with
12446 the string "DEFAULT:eNULL".
12447 [Steve Henson]
12448
12449 *) Fix to RSA private encryption routines: if p < q then it would
12450 occasionally produce an invalid result. This will only happen with
12451 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12452 [Steve Henson]
12453
12454 *) Be less restrictive and allow also `perl util/perlpath.pl
12455 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12456 because this way one can also use an interpreter named `perl5' (which is
12457 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12458 installed as `perl').
12459 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12460
12461 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12462 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12463
12464 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12465 advapi32.lib to Win32 build and change the pem test comparison
12466 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12467 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
12468 and crypto/des/ede_cbcm_enc.c.
12469 [Steve Henson]
12470
12471 *) DES quad checksum was broken on big-endian architectures. Fixed.
12472 [Ben Laurie]
12473
12474 *) Comment out two functions in bio.h that aren't implemented. Fix up the
12475 Win32 test batch file so it (might) work again. The Win32 test batch file
12476 is horrible: I feel ill....
12477 [Steve Henson]
12478
12479 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
12480 in e_os.h. Audit of header files to check ANSI and non ANSI
12481 sections: 10 functions were absent from non ANSI section and not exported
12482 from Windows DLLs. Fixed up libeay.num for new functions.
12483 [Steve Henson]
12484
12485 *) Make `openssl version' output lines consistent.
12486 [Ralf S. Engelschall]
12487
12488 *) Fix Win32 symbol export lists for BIO functions: Added
12489 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
12490 to ms/libeay{16,32}.def.
12491 [Ralf S. Engelschall]
12492
12493 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
12494 fine under Unix and passes some trivial tests I've now added. But the
12495 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
12496 added to make sure no one expects that this stuff really works in the
12497 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
12498 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
12499 openssl_bio.xs.
12500 [Ralf S. Engelschall]
12501
12502 *) Fix the generation of two part addresses in perl.
12503 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
12504
12505 *) Add config entry for Linux on MIPS.
12506 [John Tobey <jtobey@channel1.com>]
12507
12508 *) Make links whenever Configure is run, unless we are on Windoze.
12509 [Ben Laurie]
12510
12511 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
12512 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
12513 in CRLs.
12514 [Steve Henson]
12515
12516 *) Add a useful kludge to allow package maintainers to specify compiler and
12517 other platforms details on the command line without having to patch the
12518 Configure script everytime: One now can use ``perl Configure
12519 <id>:<details>'', i.e. platform ids are allowed to have details appended
12520 to them (separated by colons). This is treated as there would be a static
12521 pre-configured entry in Configure's %table under key <id> with value
12522 <details> and ``perl Configure <id>'' is called. So, when you want to
12523 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
12524 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
12525 now, which overrides the FreeBSD-elf entry on-the-fly.
12526 [Ralf S. Engelschall]
12527
12528 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
12529 [Ben Laurie]
12530
12531 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
12532 on the `perl Configure ...' command line. This way one can compile
12533 OpenSSL libraries with Position Independent Code (PIC) which is needed
12534 for linking it into DSOs.
12535 [Ralf S. Engelschall]
12536
12537 *) Remarkably, export ciphers were totally broken and no-one had noticed!
12538 Fixed.
12539 [Ben Laurie]
12540
12541 *) Cleaned up the LICENSE document: The official contact for any license
12542 questions now is the OpenSSL core team under openssl-core@openssl.org.
12543 And add a paragraph about the dual-license situation to make sure people
12544 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
12545 to the OpenSSL toolkit.
12546 [Ralf S. Engelschall]
12547
12548 *) General source tree makefile cleanups: Made `making xxx in yyy...'
12549 display consistent in the source tree and replaced `/bin/rm' by `rm'.
12550 Additionally cleaned up the `make links' target: Remove unnecessary
12551 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
12552 to speed processing and no longer clutter the display with confusing
12553 stuff. Instead only the actually done links are displayed.
12554 [Ralf S. Engelschall]
12555
12556 *) Permit null encryption ciphersuites, used for authentication only. It used
12557 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
12558 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12559 encryption.
12560 [Ben Laurie]
12561
12562 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12563 signed attributes when verifying signatures (this would break them),
12564 the detached data encoding was wrong and public keys obtained using
12565 X509_get_pubkey() weren't freed.
12566 [Steve Henson]
12567
12568 *) Add text documentation for the BUFFER functions. Also added a work around
12569 to a Win95 console bug. This was triggered by the password read stuff: the
12570 last character typed gets carried over to the next fread(). If you were
12571 generating a new cert request using 'req' for example then the last
12572 character of the passphrase would be CR which would then enter the first
12573 field as blank.
12574 [Steve Henson]
12575
12576 *) Added the new `Includes OpenSSL Cryptography Software' button as
12577 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12578 button and can be used by applications based on OpenSSL to show the
12579 relationship to the OpenSSL project.
12580 [Ralf S. Engelschall]
12581
12582 *) Remove confusing variables in function signatures in files
12583 ssl/ssl_lib.c and ssl/ssl.h.
12584 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12585
12586 *) Don't install bss_file.c under PREFIX/include/
12587 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12588
12589 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12590 functions that return function pointers and has support for NT specific
12591 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12592 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12593 unsigned to signed types: this was killing the Win32 compile.
12594 [Steve Henson]
12595
12596 *) Add new certificate file to stack functions,
12597 SSL_add_dir_cert_subjects_to_stack() and
12598 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12599 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12600 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12601 This means that Apache-SSL and similar packages don't have to mess around
12602 to add as many CAs as they want to the preferred list.
12603 [Ben Laurie]
12604
12605 *) Experiment with doxygen documentation. Currently only partially applied to
12606 ssl/ssl_lib.c.
12607 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12608 openssl.doxy as the configuration file.
12609 [Ben Laurie]
12610
12611 *) Get rid of remaining C++-style comments which strict C compilers hate.
12612 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12613
12614 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12615 compiled in by default: it has problems with large keys.
12616 [Steve Henson]
12617
12618 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12619 DH private keys and/or callback functions which directly correspond to
12620 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12621 is needed for applications which have to configure certificates on a
12622 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12623 (e.g. s_server).
12624 For the RSA certificate situation is makes no difference, but
12625 for the DSA certificate situation this fixes the "no shared cipher"
12626 problem where the OpenSSL cipher selection procedure failed because the
12627 temporary keys were not overtaken from the context and the API provided
12628 no way to reconfigure them.
12629 The new functions now let applications reconfigure the stuff and they
12630 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12631 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12632 non-public-API function ssl_cert_instantiate() is used as a helper
12633 function and also to reduce code redundancy inside ssl_rsa.c.
12634 [Ralf S. Engelschall]
12635
12636 *) Move s_server -dcert and -dkey options out of the undocumented feature
12637 area because they are useful for the DSA situation and should be
12638 recognized by the users.
12639 [Ralf S. Engelschall]
12640
12641 *) Fix the cipher decision scheme for export ciphers: the export bits are
12642 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12643 SSL_EXP_MASK. So, the original variable has to be used instead of the
12644 already masked variable.
12645 [Richard Levitte <levitte@stacken.kth.se>]
12646
12647 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12648 [Richard Levitte <levitte@stacken.kth.se>]
12649
12650 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12651 from `int' to `unsigned int' because it's a length and initialized by
12652 EVP_DigestFinal() which expects an `unsigned int *'.
12653 [Richard Levitte <levitte@stacken.kth.se>]
12654
12655 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12656 script. Instead use the usual Shell->Perl transition trick.
12657 [Ralf S. Engelschall]
12658
12659 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12660 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12661 -noout -modulus' as it's already the case for `openssl rsa -noout
12662 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12663 currently the public key is printed (a decision which was already done by
12664 `openssl dsa -modulus' in the past) which serves a similar purpose.
12665 Additionally the NO_RSA no longer completely removes the whole -modulus
12666 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12667 now, too.
12668 [Ralf S. Engelschall]
12669
12670 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12671 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12672 [Arne Ansper <arne@ats.cyber.ee>]
12673
12674 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12675 to be added. Now both 'req' and 'ca' can use new objects defined in the
12676 config file.
12677 [Steve Henson]
12678
12679 *) Add cool BIO that does syslog (or event log on NT).
12680 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12681
12682 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12683 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12684 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12685 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12686 [Ben Laurie]
12687
12688 *) Add preliminary config info for new extension code.
12689 [Steve Henson]
12690
12691 *) Make RSA_NO_PADDING really use no padding.
12692 [Ulf Moeller <ulf@fitug.de>]
12693
12694 *) Generate errors when private/public key check is done.
12695 [Ben Laurie]
12696
12697 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12698 for some CRL extensions and new objects added.
12699 [Steve Henson]
12700
12701 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
12702 key usage extension and fuller support for authority key id.
12703 [Steve Henson]
12704
12705 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
12706 padding method for RSA, which is recommended for new applications in PKCS
12707 #1 v2.0 (RFC 2437, October 1998).
12708 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
12709 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
12710 against Bleichbacher's attack on RSA.
12711 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
12712 Ben Laurie]
12713
12714 *) Updates to the new SSL compression code
12715 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12716
12717 *) Fix so that the version number in the master secret, when passed
12718 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
12719 (because the server will not accept higher), that the version number
12720 is 0x03,0x01, not 0x03,0x00
12721 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12722
12723 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
12724 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
12725 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
12726 [Steve Henson]
12727
12728 *) Support for RAW extensions where an arbitrary extension can be
12729 created by including its DER encoding. See apps/openssl.cnf for
12730 an example.
12731 [Steve Henson]
12732
12733 *) Make sure latest Perl versions don't interpret some generated C array
12734 code as Perl array code in the crypto/err/err_genc.pl script.
12735 [Lars Weber <3weber@informatik.uni-hamburg.de>]
12736
12737 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
12738 not many people have the assembler. Various Win32 compilation fixes and
12739 update to the INSTALL.W32 file with (hopefully) more accurate Win32
12740 build instructions.
12741 [Steve Henson]
12742
12743 *) Modify configure script 'Configure' to automatically create crypto/date.h
12744 file under Win32 and also build pem.h from pem.org. New script
12745 util/mkfiles.pl to create the MINFO file on environments that can't do a
12746 'make files': perl util/mkfiles.pl >MINFO should work.
12747 [Steve Henson]
12748
12749 *) Major rework of DES function declarations, in the pursuit of correctness
12750 and purity. As a result, many evil casts evaporated, and some weirdness,
12751 too. You may find this causes warnings in your code. Zapping your evil
12752 casts will probably fix them. Mostly.
12753 [Ben Laurie]
12754
12755 *) Fix for a typo in asn1.h. Bug fix to object creation script
12756 obj_dat.pl. It considered a zero in an object definition to mean
12757 "end of object": none of the objects in objects.h have any zeros
12758 so it wasn't spotted.
12759 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
12760
12761 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
12762 Masking (CBCM). In the absence of test vectors, the best I have been able
12763 to do is check that the decrypt undoes the encrypt, so far. Send me test
12764 vectors if you have them.
12765 [Ben Laurie]
12766
12767 *) Correct calculation of key length for export ciphers (too much space was
12768 allocated for null ciphers). This has not been tested!
12769 [Ben Laurie]
12770
12771 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
12772 message is now correct (it understands "crypto" and "ssl" on its
12773 command line). There is also now an "update" option. This will update
12774 the util/ssleay.num and util/libeay.num files with any new functions.
12775 If you do a:
12776 perl util/mkdef.pl crypto ssl update
12777 it will update them.
12778 [Steve Henson]
12779
12780 *) Overhauled the Perl interface (perl/*):
12781 - ported BN stuff to OpenSSL's different BN library
12782 - made the perl/ source tree CVS-aware
12783 - renamed the package from SSLeay to OpenSSL (the files still contain
12784 their history because I've copied them in the repository)
12785 - removed obsolete files (the test scripts will be replaced
12786 by better Test::Harness variants in the future)
12787 [Ralf S. Engelschall]
12788
12789 *) First cut for a very conservative source tree cleanup:
12790 1. merge various obsolete readme texts into doc/ssleay.txt
12791 where we collect the old documents and readme texts.
12792 2. remove the first part of files where I'm already sure that we no
12793 longer need them because of three reasons: either they are just temporary
12794 files which were left by Eric or they are preserved original files where
12795 I've verified that the diff is also available in the CVS via "cvs diff
12796 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
12797 the crypto/md/ stuff).
12798 [Ralf S. Engelschall]
12799
12800 *) More extension code. Incomplete support for subject and issuer alt
12801 name, issuer and authority key id. Change the i2v function parameters
12802 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
12803 what that's for :-) Fix to ASN1 macro which messed up
12804 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
12805 [Steve Henson]
12806
12807 *) Preliminary support for ENUMERATED type. This is largely copied from the
12808 INTEGER code.
12809 [Steve Henson]
12810
12811 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
12812 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12813
12814 *) Make sure `make rehash' target really finds the `openssl' program.
12815 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12816
12817 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
12818 like to hear about it if this slows down other processors.
12819 [Ben Laurie]
12820
12821 *) Add CygWin32 platform information to Configure script.
12822 [Alan Batie <batie@aahz.jf.intel.com>]
12823
12824 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
12825 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
12826
12827 *) New program nseq to manipulate netscape certificate sequences
12828 [Steve Henson]
12829
12830 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
12831 few typos.
12832 [Steve Henson]
12833
12834 *) Fixes to BN code. Previously the default was to define BN_RECURSION
12835 but the BN code had some problems that would cause failures when
12836 doing certificate verification and some other functions.
12837 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12838
12839 *) Add ASN1 and PEM code to support netscape certificate sequences.
12840 [Steve Henson]
12841
12842 *) Add ASN1 and PEM code to support netscape certificate sequences.
12843 [Steve Henson]
12844
12845 *) Add several PKIX and private extended key usage OIDs.
12846 [Steve Henson]
12847
12848 *) Modify the 'ca' program to handle the new extension code. Modify
12849 openssl.cnf for new extension format, add comments.
12850 [Steve Henson]
12851
12852 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
12853 and add a sample to openssl.cnf so req -x509 now adds appropriate
12854 CA extensions.
12855 [Steve Henson]
12856
12857 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
12858 error code, add initial support to X509_print() and x509 application.
12859 [Steve Henson]
12860
12861 *) Takes a deep breath and start adding X509 V3 extension support code. Add
12862 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
12863 stuff is currently isolated and isn't even compiled yet.
12864 [Steve Henson]
12865
12866 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
12867 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
12868 Removed the versions check from X509 routines when loading extensions:
12869 this allows certain broken certificates that don't set the version
12870 properly to be processed.
12871 [Steve Henson]
12872
12873 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
12874 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
12875 can still be regenerated with "make depend".
12876 [Ben Laurie]
12877
12878 *) Spelling mistake in C version of CAST-128.
12879 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
12880
12881 *) Changes to the error generation code. The perl script err-code.pl
12882 now reads in the old error codes and retains the old numbers, only
12883 adding new ones if necessary. It also only changes the .err files if new
12884 codes are added. The makefiles have been modified to only insert errors
12885 when needed (to avoid needlessly modifying header files). This is done
12886 by only inserting errors if the .err file is newer than the auto generated
12887 C file. To rebuild all the error codes from scratch (the old behaviour)
12888 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
12889 or delete all the .err files.
12890 [Steve Henson]
12891
12892 *) CAST-128 was incorrectly implemented for short keys. The C version has
12893 been fixed, but is untested. The assembler versions are also fixed, but
12894 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
12895 to regenerate it if needed.
12896 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
12897 Hagino <itojun@kame.net>]
12898
12899 *) File was opened incorrectly in randfile.c.
12900 [Ulf Möller <ulf@fitug.de>]
12901
12902 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
12903 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
12904 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
12905 al: it's just almost always a UTCTime. Note this patch adds new error
12906 codes so do a "make errors" if there are problems.
12907 [Steve Henson]
12908
12909 *) Correct Linux 1 recognition in config.
12910 [Ulf Möller <ulf@fitug.de>]
12911
12912 *) Remove pointless MD5 hash when using DSA keys in ca.
12913 [Anonymous <nobody@replay.com>]
12914
12915 *) Generate an error if given an empty string as a cert directory. Also
12916 generate an error if handed NULL (previously returned 0 to indicate an
12917 error, but didn't set one).
12918 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
12919
12920 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
12921 [Ben Laurie]
12922
12923 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
12924 parameters. This was causing a warning which killed off the Win32 compile.
12925 [Steve Henson]
12926
12927 *) Remove C++ style comments from crypto/bn/bn_local.h.
12928 [Neil Costigan <neil.costigan@celocom.com>]
12929
12930 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
12931 based on a text string, looking up short and long names and finally
12932 "dot" format. The "dot" format stuff didn't work. Added new function
12933 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
12934 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
12935 OID is not part of the table.
12936 [Steve Henson]
12937
12938 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
12939 X509_LOOKUP_by_alias().
12940 [Ben Laurie]
12941
12942 *) Sort openssl functions by name.
12943 [Ben Laurie]
12944
12945 *) Get the gendsa program working (hopefully) and add it to app list. Remove
12946 encryption from sample DSA keys (in case anyone is interested the password
12947 was "1234").
12948 [Steve Henson]
12949
12950 *) Make _all_ *_free functions accept a NULL pointer.
12951 [Frans Heymans <fheymans@isaserver.be>]
12952
12953 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
12954 NULL pointers.
12955 [Anonymous <nobody@replay.com>]
12956
12957 *) s_server should send the CAfile as acceptable CAs, not its own cert.
12958 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12959
12960 *) Don't blow it for numeric -newkey arguments to apps/req.
12961 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
12962
12963 *) Temp key "for export" tests were wrong in s3_srvr.c.
12964 [Anonymous <nobody@replay.com>]
12965
12966 *) Add prototype for temp key callback functions
12967 SSL_CTX_set_tmp_{rsa,dh}_callback().
12968 [Ben Laurie]
12969
12970 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
12971 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
12972 [Steve Henson]
12973
12974 *) X509_name_add_entry() freed the wrong thing after an error.
12975 [Arne Ansper <arne@ats.cyber.ee>]
12976
12977 *) rsa_eay.c would attempt to free a NULL context.
12978 [Arne Ansper <arne@ats.cyber.ee>]
12979
12980 *) BIO_s_socket() had a broken should_retry() on Windoze.
12981 [Arne Ansper <arne@ats.cyber.ee>]
12982
12983 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
12984 [Arne Ansper <arne@ats.cyber.ee>]
12985
12986 *) Make sure the already existing X509_STORE->depth variable is initialized
12987 in X509_STORE_new(), but document the fact that this variable is still
12988 unused in the certificate verification process.
12989 [Ralf S. Engelschall]
12990
12991 *) Fix the various library and apps files to free up pkeys obtained from
12992 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
12993 [Steve Henson]
12994
12995 *) Fix reference counting in X509_PUBKEY_get(). This makes
12996 demos/maurice/example2.c work, amongst others, probably.
12997 [Steve Henson and Ben Laurie]
12998
12999 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
13000 `openssl' and second, the shortcut symlinks for the `openssl <command>'
13001 are no longer created. This way we have a single and consistent command
13002 line interface `openssl <command>', similar to `cvs <command>'.
13003 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
13004
13005 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
13006 BIT STRING wrapper always have zero unused bits.
13007 [Steve Henson]
13008
13009 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
13010 [Steve Henson]
13011
13012 *) Make the top-level INSTALL documentation easier to understand.
13013 [Paul Sutton]
13014
13015 *) Makefiles updated to exit if an error occurs in a sub-directory
13016 make (including if user presses ^C) [Paul Sutton]
13017
13018 *) Make Montgomery context stuff explicit in RSA data structure.
13019 [Ben Laurie]
13020
13021 *) Fix build order of pem and err to allow for generated pem.h.
13022 [Ben Laurie]
13023
13024 *) Fix renumbering bug in X509_NAME_delete_entry().
13025 [Ben Laurie]
13026
13027 *) Enhanced the err-ins.pl script so it makes the error library number
13028 global and can add a library name. This is needed for external ASN1 and
13029 other error libraries.
13030 [Steve Henson]
13031
13032 *) Fixed sk_insert which never worked properly.
13033 [Steve Henson]
13034
13035 *) Fix ASN1 macros so they can handle indefinite length constructed
13036 EXPLICIT tags. Some non standard certificates use these: they can now
13037 be read in.
13038 [Steve Henson]
13039
13040 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
13041 into a single doc/ssleay.txt bundle. This way the information is still
13042 preserved but no longer messes up this directory. Now it's new room for
13043 the new set of documentation files.
13044 [Ralf S. Engelschall]
13045
13046 *) SETs were incorrectly DER encoded. This was a major pain, because they
13047 shared code with SEQUENCEs, which aren't coded the same. This means that
13048 almost everything to do with SETs or SEQUENCEs has either changed name or
13049 number of arguments.
13050 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
13051
13052 *) Fix test data to work with the above.
13053 [Ben Laurie]
13054
13055 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
13056 was already fixed by Eric for 0.9.1 it seems.
13057 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
13058
13059 *) Autodetect FreeBSD3.
13060 [Ben Laurie]
13061
13062 *) Fix various bugs in Configure. This affects the following platforms:
13063 nextstep
13064 ncr-scde
13065 unixware-2.0
13066 unixware-2.0-pentium
13067 sco5-cc.
13068 [Ben Laurie]
13069
13070 *) Eliminate generated files from CVS. Reorder tests to regenerate files
13071 before they are needed.
13072 [Ben Laurie]
13073
13074 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
13075 [Ben Laurie]
13076
13077
13078 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
13079
13080 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
13081 changed SSLeay to OpenSSL in version strings.
13082 [Ralf S. Engelschall]
13083
13084 *) Some fixups to the top-level documents.
13085 [Paul Sutton]
13086
13087 *) Fixed the nasty bug where rsaref.h was not found under compile-time
13088 because the symlink to include/ was missing.
13089 [Ralf S. Engelschall]
13090
13091 *) Incorporated the popular no-RSA/DSA-only patches
13092 which allow to compile a RSA-free SSLeay.
13093 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
13094
13095 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
13096 when "ssleay" is still not found.
13097 [Ralf S. Engelschall]
13098
13099 *) Added more platforms to Configure: Cray T3E, HPUX 11,
13100 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
13101
13102 *) Updated the README file.
13103 [Ralf S. Engelschall]
13104
13105 *) Added various .cvsignore files in the CVS repository subdirs
13106 to make a "cvs update" really silent.
13107 [Ralf S. Engelschall]
13108
13109 *) Recompiled the error-definition header files and added
13110 missing symbols to the Win32 linker tables.
13111 [Ralf S. Engelschall]
13112
13113 *) Cleaned up the top-level documents;
13114 o new files: CHANGES and LICENSE
13115 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
13116 o merged COPYRIGHT into LICENSE
13117 o removed obsolete TODO file
13118 o renamed MICROSOFT to INSTALL.W32
13119 [Ralf S. Engelschall]
13120
13121 *) Removed dummy files from the 0.9.1b source tree:
13122 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
13123 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
13124 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
13125 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
13126 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
13127 [Ralf S. Engelschall]
13128
13129 *) Added various platform portability fixes.
13130 [Mark J. Cox]
13131
13132 *) The Genesis of the OpenSSL rpject:
13133 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
13134 Young and Tim J. Hudson created while they were working for C2Net until
13135 summer 1998.
13136 [The OpenSSL Project]
13137
13138
13139 Changes between 0.9.0b and 0.9.1b [not released]
13140
13141 *) Updated a few CA certificates under certs/
13142 [Eric A. Young]
13143
13144 *) Changed some BIGNUM api stuff.
13145 [Eric A. Young]
13146
13147 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13148 DGUX x86, Linux Alpha, etc.
13149 [Eric A. Young]
13150
13151 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13152 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13153 available).
13154 [Eric A. Young]
13155
13156 *) Add -strparse option to asn1pars program which parses nested
13157 binary structures
13158 [Dr Stephen Henson <shenson@bigfoot.com>]
13159
13160 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13161 [Eric A. Young]
13162
13163 *) DSA fix for "ca" program.
13164 [Eric A. Young]
13165
13166 *) Added "-genkey" option to "dsaparam" program.
13167 [Eric A. Young]
13168
13169 *) Added RIPE MD160 (rmd160) message digest.
13170 [Eric A. Young]
13171
13172 *) Added -a (all) option to "ssleay version" command.
13173 [Eric A. Young]
13174
13175 *) Added PLATFORM define which is the id given to Configure.
13176 [Eric A. Young]
13177
13178 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13179 [Eric A. Young]
13180
13181 *) Extended the ASN.1 parser routines.
13182 [Eric A. Young]
13183
13184 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13185 [Eric A. Young]
13186
13187 *) Added a BN_CTX to the BN library.
13188 [Eric A. Young]
13189
13190 *) Fixed the weak key values in DES library
13191 [Eric A. Young]
13192
13193 *) Changed API in EVP library for cipher aliases.
13194 [Eric A. Young]
13195
13196 *) Added support for RC2/64bit cipher.
13197 [Eric A. Young]
13198
13199 *) Converted the lhash library to the crypto/mem.c functions.
13200 [Eric A. Young]
13201
13202 *) Added more recognized ASN.1 object ids.
13203 [Eric A. Young]
13204
13205 *) Added more RSA padding checks for SSL/TLS.
13206 [Eric A. Young]
13207
13208 *) Added BIO proxy/filter functionality.
13209 [Eric A. Young]
13210
13211 *) Added extra_certs to SSL_CTX which can be used
13212 send extra CA certificates to the client in the CA cert chain sending
13213 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13214 [Eric A. Young]
13215
13216 *) Now Fortezza is denied in the authentication phase because
13217 this is key exchange mechanism is not supported by SSLeay at all.
13218 [Eric A. Young]
13219
13220 *) Additional PKCS1 checks.
13221 [Eric A. Young]
13222
13223 *) Support the string "TLSv1" for all TLS v1 ciphers.
13224 [Eric A. Young]
13225
13226 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13227 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13228 [Eric A. Young]
13229
13230 *) Fixed a few memory leaks.
13231 [Eric A. Young]
13232
13233 *) Fixed various code and comment typos.
13234 [Eric A. Young]
13235
13236 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13237 bytes sent in the client random.
13238 [Edward Bishop <ebishop@spyglass.com>]