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
10 Changes between 1.1.1c and 1.1.1d [xx XXX xxxx]
12 *) Use Windows installation paths in the mingw builds
14 Mingw isn't a POSIX environment per se, which means that Windows
15 paths should be used for installation.
19 *) Changed DH parameters to generate the order q subgroup instead of 2q.
20 Previously generated DH parameters are still accepted by DH_check
21 but DH_generate_key works around that by clearing bit 0 of the
22 private key for those. This avoids leaking bit 0 of the private key.
25 *) Significantly reduce secure memory usage by the randomness pools.
28 *) Revert the DEVRANDOM_WAIT feature for Linux systems
30 The DEVRANDOM_WAIT feature added a select() call to wait for the
31 /dev/random device to become readable before reading from the
34 It turned out that this change had negative side effects on
35 performance which were not acceptable. After some discussion it
36 was decided to revert this feature and leave it up to the OS
37 resp. the platform maintainer to ensure a proper initialization
38 during early boot time.
40 Changes between 1.1.1b and 1.1.1c [28 May 2019]
42 *) Add build tests for C++. These are generated files that only do one
43 thing, to include one public OpenSSL head file each. This tests that
44 the public header files can be usefully included in a C++ application.
46 This test isn't enabled by default. It can be enabled with the option
47 'enable-buildtest-c++'.
50 *) Enable SHA3 pre-hashing for ECDSA and DSA.
53 *) Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
54 This changes the size when using the genpkey app when no size is given. It
55 fixes an omission in earlier changes that changed all RSA, DSA and DH
56 generation apps to use 2048 bits by default.
59 *) Reorganize the manual pages to consistently have RETURN VALUES,
60 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
61 util/fix-doc-nits accordingly.
62 [Paul Yang, Joshua Lock]
64 *) Add the missing accessor EVP_PKEY_get0_engine()
67 *) Have apps like 's_client' and 's_server' output the signature scheme
68 along with other cipher suite parameters when debugging.
71 *) Make OPENSSL_config() error agnostic again.
74 *) Do the error handling in RSA decryption constant time.
77 *) Prevent over long nonces in ChaCha20-Poly1305.
79 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
80 for every encryption operation. RFC 7539 specifies that the nonce value
81 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
82 and front pads the nonce with 0 bytes if it is less than 12
83 bytes. However it also incorrectly allows a nonce to be set of up to 16
84 bytes. In this case only the last 12 bytes are significant and any
85 additional leading bytes are ignored.
87 It is a requirement of using this cipher that nonce values are
88 unique. Messages encrypted using a reused nonce value are susceptible to
89 serious confidentiality and integrity attacks. If an application changes
90 the default nonce length to be longer than 12 bytes and then makes a
91 change to the leading bytes of the nonce expecting the new value to be a
92 new unique nonce then such an application could inadvertently encrypt
93 messages with a reused nonce.
95 Additionally the ignored bytes in a long nonce are not covered by the
96 integrity guarantee of this cipher. Any application that relies on the
97 integrity of these ignored leading bytes of a long nonce may be further
98 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
99 is safe because no such use sets such a long nonce value. However user
100 applications that use this cipher directly and set a non-default nonce
101 length to be longer than 12 bytes may be vulnerable.
103 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
108 *) Add DEVRANDOM_WAIT feature for Linux systems
110 On older Linux systems where the getrandom() system call is not available,
111 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
112 Contrary to getrandom(), the /dev/urandom device will not block during
113 early boot when the kernel CSPRNG has not been seeded yet.
115 To mitigate this known weakness, use select() to wait for /dev/random to
116 become readable before reading from /dev/urandom.
118 *) Ensure that SM2 only uses SM3 as digest algorithm
121 Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
123 *) Added SCA hardening for modular field inversion in EC_GROUP through
124 a new dedicated field_inv() pointer in EC_METHOD.
125 This also addresses a leakage affecting conversions from projective
126 to affine coordinates.
127 [Billy Bob Brumley, Nicola Tuveri]
129 *) Change the info callback signals for the start and end of a post-handshake
130 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
131 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
132 confused by this and assume that a TLSv1.2 renegotiation has started. This
133 can break KeyUpdate handling. Instead we no longer signal the start and end
134 of a post handshake message exchange (although the messages themselves are
135 still signalled). This could break some applications that were expecting
136 the old signals. However without this KeyUpdate is not usable for many
140 *) Fix a bug in the computation of the endpoint-pair shared secret used
141 by DTLS over SCTP. This breaks interoperability with older versions
142 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
143 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
144 interoperability with such broken implementations. However, enabling
145 this switch breaks interoperability with correct implementations.
147 *) Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
148 re-used X509_PUBKEY object if the second PUBKEY is malformed.
151 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
154 *) Remove the 'dist' target and add a tarball building script. The
155 'dist' target has fallen out of use, and it shouldn't be
156 necessary to configure just to create a source distribution.
159 Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
161 *) Timing vulnerability in DSA signature generation
163 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
164 timing side channel attack. An attacker could use variations in the signing
165 algorithm to recover the private key.
167 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
171 *) Timing vulnerability in ECDSA signature generation
173 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
174 timing side channel attack. An attacker could use variations in the signing
175 algorithm to recover the private key.
177 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
181 *) Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
182 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
183 are retained for backwards compatibility.
186 *) Fixed the issue that RAND_add()/RAND_seed() silently discards random input
187 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
188 of two gigabytes and the error handling improved.
190 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
191 categorized as a normal bug, not a security issue, because the DRBG reseeds
192 automatically and is fully functional even without additional randomness
193 provided by the application.
195 Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
197 *) Add a new ClientHello callback. Provides a callback interface that gives
198 the application the ability to adjust the nascent SSL object at the
199 earliest stage of ClientHello processing, immediately after extensions have
200 been collected but before they have been processed. In particular, this
201 callback can adjust the supported TLS versions in response to the contents
205 *) Add SM2 base algorithm support.
208 *) s390x assembly pack: add (improved) hardware-support for the following
209 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
210 aes-cfb/cfb8, aes-ecb.
213 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
214 parameter is no longer accepted, as it leads to a corrupt table. NULL
215 pem_str is reserved for alias entries only.
218 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
219 step for prime curves. The new implementation is based on formulae from
220 differential addition-and-doubling in homogeneous projective coordinates
221 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
222 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
223 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
224 to work in projective coordinates.
225 [Billy Bob Brumley, Nicola Tuveri]
227 *) Change generating and checking of primes so that the error rate of not
228 being prime depends on the intended use based on the size of the input.
229 For larger primes this will result in more rounds of Miller-Rabin.
230 The maximal error rate for primes with more than 1080 bits is lowered
232 [Kurt Roeckx, Annie Yousar]
234 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
237 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
238 moving between systems, and to avoid confusion when a Windows build is
239 done with mingw vs with MSVC. For POSIX installs, there's still a
240 symlink or copy named 'tsget' to avoid that confusion as well.
243 *) Revert blinding in ECDSA sign and instead make problematic addition
244 length-invariant. Switch even to fixed-length Montgomery multiplication.
247 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
248 step for binary curves. The new implementation is based on formulae from
249 differential addition-and-doubling in mixed Lopez-Dahab projective
250 coordinates, modified to independently blind the operands.
251 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
253 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
254 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
255 EC_METHODs to implement their own specialized "ladder step", to take
256 advantage of more favorable coordinate systems or more efficient
257 differential addition-and-doubling algorithms.
258 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
260 *) Modified the random device based seed sources to keep the relevant
261 file descriptors open rather than reopening them on each access.
262 This allows such sources to operate in a chroot() jail without
263 the associated device nodes being available. This behaviour can be
264 controlled using RAND_keep_random_devices_open().
267 *) Numerous side-channel attack mitigations have been applied. This may have
268 performance impacts for some algorithms for the benefit of improved
269 security. Specific changes are noted in this change log by their respective
273 *) AIX shared library support overhaul. Switch to AIX "natural" way of
274 handling shared libraries, which means collecting shared objects of
275 different versions and bitnesses in one common archive. This allows to
276 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
277 doesn't affect the way 3rd party applications are linked, only how
278 multi-version installation is managed.
281 *) Make ec_group_do_inverse_ord() more robust and available to other
282 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
283 mitigations are applied to the fallback BN_mod_inverse().
284 When using this function rather than BN_mod_inverse() directly, new
285 EC cryptosystem implementations are then safer-by-default.
288 *) Add coordinate blinding for EC_POINT and implement projective
289 coordinate blinding for generic prime curves as a countermeasure to
290 chosen point SCA attacks.
291 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
293 *) Add blinding to ECDSA and DSA signatures to protect against side channel
294 attacks discovered by Keegan Ryan (NCC Group).
297 *) Enforce checking in the pkeyutl command line app to ensure that the input
298 length does not exceed the maximum supported digest length when performing
299 a sign, verify or verifyrecover operation.
302 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
303 I/O in combination with something like select() or poll() will hang. This
304 can be turned off again using SSL_CTX_clear_mode().
305 Many applications do not properly handle non-application data records, and
306 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
307 around the problems in those applications, but can also break some.
308 It's recommended to read the manpages about SSL_read(), SSL_write(),
309 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
310 SSL_CTX_set_read_ahead() again.
313 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
314 now allow empty (zero character) pass phrases.
317 *) Apply blinding to binary field modular inversion and remove patent
318 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
321 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
322 binary and prime elliptic curves.
325 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
326 constant time fixed point multiplication.
329 *) Revise elliptic curve scalar multiplication with timing attack
330 defenses: ec_wNAF_mul redirects to a constant time implementation
331 when computing fixed point and variable point multiplication (which
332 in OpenSSL are mostly used with secret scalars in keygen, sign,
333 ECDH derive operations).
334 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
337 *) Updated CONTRIBUTING
340 *) Updated DRBG / RAND to request nonce and additional low entropy
341 randomness from the system.
342 [Matthias St. Pierre]
344 *) Updated 'openssl rehash' to use OpenSSL consistent default.
347 *) Moved the load of the ssl_conf module to libcrypto, which helps
348 loading engines that libssl uses before libssl is initialised.
351 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
354 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
355 [Ingo Schwarze, Rich Salz]
357 *) Added output of accepting IP address and port for 'openssl s_server'
360 *) Added a new API for TLSv1.3 ciphersuites:
361 SSL_CTX_set_ciphersuites()
362 SSL_set_ciphersuites()
365 *) Memory allocation failures consistenly add an error to the error
369 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
370 in libcrypto when run as setuid/setgid.
373 *) Load any config file by default when libssl is used.
376 *) Added new public header file <openssl/rand_drbg.h> and documentation
377 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
378 [Matthias St. Pierre]
380 *) QNX support removed (cannot find contributors to get their approval
381 for the license change).
384 *) TLSv1.3 replay protection for early data has been implemented. See the
385 SSL_read_early_data() man page for further details.
388 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
389 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
390 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
391 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
392 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
393 configuration has been separated out. See the ciphers man page or the
394 SSL_CTX_set_ciphersuites() man page for more information.
397 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
398 in responder mode now supports the new "-multi" option, which
399 spawns the specified number of child processes to handle OCSP
400 requests. The "-timeout" option now also limits the OCSP
401 responder's patience to wait to receive the full client request
402 on a newly accepted connection. Child processes are respawned
403 as needed, and the CA index file is automatically reloaded
404 when changed. This makes it possible to run the "ocsp" responder
405 as a long-running service, making the OpenSSL CA somewhat more
406 feature-complete. In this mode, most diagnostic messages logged
407 after entering the event loop are logged via syslog(3) rather than
411 *) Added support for X448 and Ed448. Heavily based on original work by
415 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
416 objects loaded. This adds the functions OSSL_STORE_expect() and
417 OSSL_STORE_find() as well as needed tools to construct searches and
418 get the search data out of them.
421 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
422 version of OpenSSL should review their configuration settings to ensure
423 that they are still appropriate for TLSv1.3. For further information see:
424 https://wiki.openssl.org/index.php/TLS1.3
427 *) Grand redesign of the OpenSSL random generator
429 The default RAND method now utilizes an AES-CTR DRBG according to
430 NIST standard SP 800-90Ar1. The new random generator is essentially
431 a port of the default random generator from the OpenSSL FIPS 2.0
432 object module. It is a hybrid deterministic random bit generator
433 using an AES-CTR bit stream and which seeds and reseeds itself
434 automatically using trusted system entropy sources.
436 Some of its new features are:
437 o Support for multiple DRBG instances with seed chaining.
438 o The default RAND method makes use of a DRBG.
439 o There is a public and private DRBG instance.
440 o The DRBG instances are fork-safe.
441 o Keep all global DRBG instances on the secure heap if it is enabled.
442 o The public and private DRBG instance are per thread for lock free
444 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
446 *) Changed Configure so it only says what it does and doesn't dump
447 so much data. Instead, ./configdata.pm should be used as a script
448 to display all sorts of configuration data.
451 *) Added processing of "make variables" to Configure.
454 *) Added SHA512/224 and SHA512/256 algorithm support.
457 *) The last traces of Netware support, first removed in 1.1.0, have
461 *) Get rid of Makefile.shared, and in the process, make the processing
462 of certain files (rc.obj, or the .def/.map/.opt files produced from
463 the ordinal files) more visible and hopefully easier to trace and
464 debug (or make silent).
467 *) Make it possible to have environment variable assignments as
468 arguments to config / Configure.
471 *) Add multi-prime RSA (RFC 8017) support.
474 *) Add SM3 implemented according to GB/T 32905-2016
475 [ Jack Lloyd <jack.lloyd@ribose.com>,
476 Ronald Tse <ronald.tse@ribose.com>,
477 Erick Borsboom <erick.borsboom@ribose.com> ]
479 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
480 as documented in RFC6066.
481 Based on a patch from Tomasz Moń
482 [Filipe Raimundo da Silva]
484 *) Add SM4 implemented according to GB/T 32907-2016.
485 [ Jack Lloyd <jack.lloyd@ribose.com>,
486 Ronald Tse <ronald.tse@ribose.com>,
487 Erick Borsboom <erick.borsboom@ribose.com> ]
489 *) Reimplement -newreq-nodes and ERR_error_string_n; the
490 original author does not agree with the license change.
493 *) Add ARIA AEAD TLS support.
496 *) Some macro definitions to support VS6 have been removed. Visual
497 Studio 6 has not worked since 1.1.0
500 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
501 without clearing the errors.
504 *) Add "atfork" functions. If building on a system that without
505 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
506 requirements. The RAND facility now uses/requires this.
512 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
513 not possible to disable entirely. However, it's still possible to
514 disable the console reading UI method, UI_OpenSSL() (use UI_null()
517 To disable, configure with 'no-ui-console'. 'no-ui' is still
518 possible to use as an alias. Check at compile time with the
519 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
520 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
523 *) Add a STORE module, which implements a uniform and URI based reader of
524 stores that can contain keys, certificates, CRLs and numerous other
525 objects. The main API is loosely based on a few stdio functions,
526 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
527 OSSL_STORE_error and OSSL_STORE_close.
528 The implementation uses backends called "loaders" to implement arbitrary
529 URI schemes. There is one built in "loader" for the 'file' scheme.
532 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
533 then adjusted to work on FreeBSD 8.4 as well.
534 Enable by configuring with 'enable-devcryptoeng'. This is done by default
535 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
538 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
539 util/mkerr.pl, which is adapted to allow those prefixes, leading to
540 error code calls like this:
542 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
544 With this change, we claim the namespaces OSSL and OPENSSL in a manner
545 that can be encoded in C. For the foreseeable future, this will only
547 [Richard Levitte and Tim Hudson]
549 *) Removed BSD cryptodev engine.
552 *) Add a build target 'build_all_generated', to build all generated files
553 and only that. This can be used to prepare everything that requires
554 things like perl for a system that lacks perl and then move everything
555 to that system and do the rest of the build there.
558 *) In the UI interface, make it possible to duplicate the user data. This
559 can be used by engines that need to retain the data for a longer time
560 than just the call where this user data is passed.
563 *) Ignore the '-named_curve auto' value for compatibility of applications
565 [Tomas Mraz <tmraz@fedoraproject.org>]
567 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
568 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
569 alerts across multiple records (some of which could be empty). In practice
570 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
571 prohibts this altogether and other libraries (BoringSSL, NSS) do not
572 support this at all. Supporting it adds significant complexity to the
573 record layer, and its removal is unlikely to cause inter-operability
577 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
578 with Z. These are meant to replace LONG and ZLONG and to be size safe.
579 The use of LONG and ZLONG is discouraged and scheduled for deprecation
583 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
584 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
585 [Richard Levitte, Andy Polyakov]
587 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
591 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
592 platform rather than 'mingw'.
595 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
596 success if they are asked to add an object which already exists
597 in the store. This change cascades to other functions which load
598 certificates and CRLs.
601 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
602 facilitate stack unwinding even from assembly subroutines.
605 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
606 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
609 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
610 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
611 which is the minimum version we support.
614 *) Certificate time validation (X509_cmp_time) enforces stricter
615 compliance with RFC 5280. Fractional seconds and timezone offsets
616 are no longer allowed.
619 *) Add support for ARIA
622 *) s_client will now send the Server Name Indication (SNI) extension by
623 default unless the new "-noservername" option is used. The server name is
624 based on the host provided to the "-connect" option unless overridden by
628 *) Add support for SipHash
631 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
632 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
633 prevent issues where no progress is being made and the peer continually
634 sends unrecognised record types, using up resources processing them.
637 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
638 using the algorithm defined in
639 https://www.akkadia.org/drepper/SHA-crypt.txt
642 *) Heartbeat support has been removed; the ABI is changed for now.
643 [Richard Levitte, Rich Salz]
645 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
648 *) The RSA "null" method, which was partially supported to avoid patent
649 issues, has been replaced to always returns NULL.
653 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
655 *) Client DoS due to large DH parameter
657 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
658 malicious server can send a very large prime value to the client. This will
659 cause the client to spend an unreasonably long period of time generating a
660 key for this prime resulting in a hang until the client has finished. This
661 could be exploited in a Denial Of Service attack.
663 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
667 *) Cache timing vulnerability in RSA Key Generation
669 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
670 a cache timing side channel attack. An attacker with sufficient access to
671 mount cache timing attacks during the RSA key generation process could
672 recover the private key.
674 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
675 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
679 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
680 parameter is no longer accepted, as it leads to a corrupt table. NULL
681 pem_str is reserved for alias entries only.
684 *) Revert blinding in ECDSA sign and instead make problematic addition
685 length-invariant. Switch even to fixed-length Montgomery multiplication.
688 *) Change generating and checking of primes so that the error rate of not
689 being prime depends on the intended use based on the size of the input.
690 For larger primes this will result in more rounds of Miller-Rabin.
691 The maximal error rate for primes with more than 1080 bits is lowered
693 [Kurt Roeckx, Annie Yousar]
695 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
698 *) Add blinding to ECDSA and DSA signatures to protect against side channel
699 attacks discovered by Keegan Ryan (NCC Group).
702 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
703 now allow empty (zero character) pass phrases.
706 *) Certificate time validation (X509_cmp_time) enforces stricter
707 compliance with RFC 5280. Fractional seconds and timezone offsets
708 are no longer allowed.
711 *) Fixed a text canonicalisation bug in CMS
713 Where a CMS detached signature is used with text content the text goes
714 through a canonicalisation process first prior to signing or verifying a
715 signature. This process strips trailing space at the end of lines, converts
716 line terminators to CRLF and removes additional trailing line terminators
717 at the end of a file. A bug in the canonicalisation process meant that
718 some characters, such as form-feed, were incorrectly treated as whitespace
719 and removed. This is contrary to the specification (RFC5485). This fix
720 could mean that detached text data signed with an earlier version of
721 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
722 signed with a fixed OpenSSL may fail to verify with an earlier version of
723 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
724 and use the "-binary" flag (for the "cms" command line application) or set
725 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
728 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
730 *) Constructed ASN.1 types with a recursive definition could exceed the stack
732 Constructed ASN.1 types with a recursive definition (such as can be found
733 in PKCS7) could eventually exceed the stack given malicious input with
734 excessive recursion. This could result in a Denial Of Service attack. There
735 are no such structures used within SSL/TLS that come from untrusted sources
736 so this is considered safe.
738 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
743 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
745 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
746 effectively reduced to only comparing the least significant bit of each
747 byte. This allows an attacker to forge messages that would be considered as
748 authenticated in an amount of tries lower than that guaranteed by the
749 security claims of the scheme. The module can only be compiled by the
750 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
752 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
757 *) Add a build target 'build_all_generated', to build all generated files
758 and only that. This can be used to prepare everything that requires
759 things like perl for a system that lacks perl and then move everything
760 to that system and do the rest of the build there.
763 *) Backport SSL_OP_NO_RENGOTIATION
765 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
766 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
767 changes this is no longer possible in 1.1.0. Therefore the new
768 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
769 1.1.0 to provide equivalent functionality.
771 Note that if an application built against 1.1.0h headers (or above) is run
772 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
773 accepted but nothing will happen, i.e. renegotiation will not be prevented.
776 *) Removed the OS390-Unix config target. It relied on a script that doesn't
780 *) rsaz_1024_mul_avx2 overflow bug on x86_64
782 There is an overflow bug in the AVX2 Montgomery multiplication procedure
783 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
784 Analysis suggests that attacks against RSA and DSA as a result of this
785 defect would be very difficult to perform and are not believed likely.
786 Attacks against DH1024 are considered just feasible, because most of the
787 work necessary to deduce information about a private key may be performed
788 offline. The amount of resources required for such an attack would be
789 significant. However, for an attack on TLS to be meaningful, the server
790 would have to share the DH1024 private key among multiple clients, which is
791 no longer an option since CVE-2016-0701.
793 This only affects processors that support the AVX2 but not ADX extensions
794 like Intel Haswell (4th generation).
796 This issue was reported to OpenSSL by David Benjamin (Google). The issue
797 was originally found via the OSS-Fuzz project.
801 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
803 *) bn_sqrx8x_internal carry bug on x86_64
805 There is a carry propagating bug in the x86_64 Montgomery squaring
806 procedure. No EC algorithms are affected. Analysis suggests that attacks
807 against RSA and DSA as a result of this defect would be very difficult to
808 perform and are not believed likely. Attacks against DH are considered just
809 feasible (although very difficult) because most of the work necessary to
810 deduce information about a private key may be performed offline. The amount
811 of resources required for such an attack would be very significant and
812 likely only accessible to a limited number of attackers. An attacker would
813 additionally need online access to an unpatched system using the target
814 private key in a scenario with persistent DH parameters and a private
815 key that is shared between multiple clients.
817 This only affects processors that support the BMI1, BMI2 and ADX extensions
818 like Intel Broadwell (5th generation) and later or AMD Ryzen.
820 This issue was reported to OpenSSL by the OSS-Fuzz project.
824 *) Malformed X.509 IPAddressFamily could cause OOB read
826 If an X.509 certificate has a malformed IPAddressFamily extension,
827 OpenSSL could do a one-byte buffer overread. The most likely result
828 would be an erroneous display of the certificate in text format.
830 This issue was reported to OpenSSL by the OSS-Fuzz project.
834 Changes between 1.1.0e and 1.1.0f [25 May 2017]
836 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
837 platform rather than 'mingw'.
840 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
841 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
842 which is the minimum version we support.
845 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
847 *) Encrypt-Then-Mac renegotiation crash
849 During a renegotiation handshake if the Encrypt-Then-Mac extension is
850 negotiated where it was not in the original handshake (or vice-versa) then
851 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
852 and servers are affected.
854 This issue was reported to OpenSSL by Joe Orton (Red Hat).
858 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
860 *) Truncated packet could crash via OOB read
862 If one side of an SSL/TLS path is running on a 32-bit host and a specific
863 cipher is being used, then a truncated packet can cause that host to
864 perform an out-of-bounds read, usually resulting in a crash.
866 This issue was reported to OpenSSL by Robert Święcki of Google.
870 *) Bad (EC)DHE parameters cause a client crash
872 If a malicious server supplies bad parameters for a DHE or ECDHE key
873 exchange then this can result in the client attempting to dereference a
874 NULL pointer leading to a client crash. This could be exploited in a Denial
877 This issue was reported to OpenSSL by Guido Vranken.
881 *) BN_mod_exp may produce incorrect results on x86_64
883 There is a carry propagating bug in the x86_64 Montgomery squaring
884 procedure. No EC algorithms are affected. Analysis suggests that attacks
885 against RSA and DSA as a result of this defect would be very difficult to
886 perform and are not believed likely. Attacks against DH are considered just
887 feasible (although very difficult) because most of the work necessary to
888 deduce information about a private key may be performed offline. The amount
889 of resources required for such an attack would be very significant and
890 likely only accessible to a limited number of attackers. An attacker would
891 additionally need online access to an unpatched system using the target
892 private key in a scenario with persistent DH parameters and a private
893 key that is shared between multiple clients. For example this can occur by
894 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
895 similar to CVE-2015-3193 but must be treated as a separate problem.
897 This issue was reported to OpenSSL by the OSS-Fuzz project.
901 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
903 *) ChaCha20/Poly1305 heap-buffer-overflow
905 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
906 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
907 crash. This issue is not considered to be exploitable beyond a DoS.
909 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
913 *) CMS Null dereference
915 Applications parsing invalid CMS structures can crash with a NULL pointer
916 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
917 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
918 structure callback if an attempt is made to free certain invalid encodings.
919 Only CHOICE structures using a callback which do not handle NULL value are
922 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
926 *) Montgomery multiplication may produce incorrect results
928 There is a carry propagating bug in the Broadwell-specific Montgomery
929 multiplication procedure that handles input lengths divisible by, but
930 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
931 and DH private keys are impossible. This is because the subroutine in
932 question is not used in operations with the private key itself and an input
933 of the attacker's direct choice. Otherwise the bug can manifest itself as
934 transient authentication and key negotiation failures or reproducible
935 erroneous outcome of public-key operations with specially crafted input.
936 Among EC algorithms only Brainpool P-512 curves are affected and one
937 presumably can attack ECDH key negotiation. Impact was not analyzed in
938 detail, because pre-requisites for attack are considered unlikely. Namely
939 multiple clients have to choose the curve in question and the server has to
940 share the private key among them, neither of which is default behaviour.
941 Even then only clients that chose the curve will be affected.
943 This issue was publicly reported as transient failures and was not
944 initially recognized as a security issue. Thanks to Richard Morgan for
945 providing reproducible case.
949 *) Removed automatic addition of RPATH in shared libraries and executables,
950 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
953 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
955 *) Fix Use After Free for large message sizes
957 The patch applied to address CVE-2016-6307 resulted in an issue where if a
958 message larger than approx 16k is received then the underlying buffer to
959 store the incoming message is reallocated and moved. Unfortunately a
960 dangling pointer to the old location is left which results in an attempt to
961 write to the previously freed location. This is likely to result in a
962 crash, however it could potentially lead to execution of arbitrary code.
964 This issue only affects OpenSSL 1.1.0a.
966 This issue was reported to OpenSSL by Robert Święcki.
970 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
972 *) OCSP Status Request extension unbounded memory growth
974 A malicious client can send an excessively large OCSP Status Request
975 extension. If that client continually requests renegotiation, sending a
976 large OCSP Status Request extension each time, then there will be unbounded
977 memory growth on the server. This will eventually lead to a Denial Of
978 Service attack through memory exhaustion. Servers with a default
979 configuration are vulnerable even if they do not support OCSP. Builds using
980 the "no-ocsp" build time option are not affected.
982 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
986 *) SSL_peek() hang on empty record
988 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
989 sends an empty record. This could be exploited by a malicious peer in a
990 Denial Of Service attack.
992 This issue was reported to OpenSSL by Alex Gaynor.
996 *) Excessive allocation of memory in tls_get_message_header() and
997 dtls1_preprocess_fragment()
999 A (D)TLS message includes 3 bytes for its length in the header for the
1000 message. This would allow for messages up to 16Mb in length. Messages of
1001 this length are excessive and OpenSSL includes a check to ensure that a
1002 peer is sending reasonably sized messages in order to avoid too much memory
1003 being consumed to service a connection. A flaw in the logic of version
1004 1.1.0 means that memory for the message is allocated too early, prior to
1005 the excessive message length check. Due to way memory is allocated in
1006 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
1007 to service a connection. This could lead to a Denial of Service through
1008 memory exhaustion. However, the excessive message length check still takes
1009 place, and this would cause the connection to immediately fail. Assuming
1010 that the application calls SSL_free() on the failed connection in a timely
1011 manner then the 21Mb of allocated memory will then be immediately freed
1012 again. Therefore the excessive memory allocation will be transitory in
1013 nature. This then means that there is only a security impact if:
1015 1) The application does not call SSL_free() in a timely manner in the event
1016 that the connection fails
1018 2) The application is working in a constrained environment where there is
1019 very little free memory
1021 3) The attacker initiates multiple connection attempts such that there are
1022 multiple connections in a state where memory has been allocated for the
1023 connection; SSL_free() has not yet been called; and there is insufficient
1024 memory to service the multiple requests.
1026 Except in the instance of (1) above any Denial Of Service is likely to be
1027 transitory because as soon as the connection fails the memory is
1028 subsequently freed again in the SSL_free() call. However there is an
1029 increased risk during this period of application crashes due to the lack of
1030 memory - which would then mean a more serious Denial of Service.
1032 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1033 (CVE-2016-6307 and CVE-2016-6308)
1036 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
1037 had to be removed. Primary reason is that vendor assembler can't
1038 assemble our modules with -KPIC flag. As result it, assembly
1039 support, was not even available as option. But its lack means
1040 lack of side-channel resistant code, which is incompatible with
1041 security by todays standards. Fortunately gcc is readily available
1042 prepackaged option, which we firmly point at...
1045 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
1047 *) Windows command-line tool supports UTF-8 opt-in option for arguments
1048 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
1049 (to any value) allows Windows user to access PKCS#12 file generated
1050 with Windows CryptoAPI and protected with non-ASCII password, as well
1051 as files generated under UTF-8 locale on Linux also protected with
1055 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
1056 have been disabled by default and removed from DEFAULT, just like RC4.
1057 See the RC4 item below to re-enable both.
1060 *) The method for finding the storage location for the Windows RAND seed file
1061 has changed. First we check %RANDFILE%. If that is not set then we check
1062 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
1063 all else fails we fall back to C:\.
1066 *) The EVP_EncryptUpdate() function has had its return type changed from void
1067 to int. A return of 0 indicates and error while a return of 1 indicates
1071 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
1072 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
1073 off the constant time implementation for RSA, DSA and DH have been made
1074 no-ops and deprecated.
1077 *) Windows RAND implementation was simplified to only get entropy by
1078 calling CryptGenRandom(). Various other RAND-related tickets
1080 [Joseph Wylie Yandle, Rich Salz]
1082 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
1083 and OPENSSL_LH_, respectively. The old names are available
1084 with API compatibility. They new names are now completely documented.
1087 *) Unify TYPE_up_ref(obj) methods signature.
1088 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
1089 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
1090 int (instead of void) like all others TYPE_up_ref() methods.
1091 So now these methods also check the return value of CRYPTO_atomic_add(),
1092 and the validity of object reference counter.
1093 [fdasilvayy@gmail.com]
1095 *) With Windows Visual Studio builds, the .pdb files are installed
1096 alongside the installed libraries and executables. For a static
1097 library installation, ossl_static.pdb is the associate compiler
1098 generated .pdb file to be used when linking programs.
1101 *) Remove openssl.spec. Packaging files belong with the packagers.
1104 *) Automatic Darwin/OSX configuration has had a refresh, it will now
1105 recognise x86_64 architectures automatically. You can still decide
1106 to build for a different bitness with the environment variable
1107 KERNEL_BITS (can be 32 or 64), for example:
1109 KERNEL_BITS=32 ./config
1113 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
1114 256 bit AES and HMAC with SHA256.
1117 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
1120 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
1123 *) To enable users to have their own config files and build file templates,
1124 Configure looks in the directory indicated by the environment variable
1125 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
1126 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
1127 name and is used as is.
1130 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
1131 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
1132 X509_CERT_FILE_CTX was removed.
1135 *) "shared" builds are now the default. To create only static libraries use
1136 the "no-shared" Configure option.
1139 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
1140 All of these option have not worked for some while and are fundamental
1144 *) Make various cleanup routines no-ops and mark them as deprecated. Most
1145 global cleanup functions are no longer required because they are handled
1146 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
1147 Explicitly de-initing can cause problems (e.g. where a library that uses
1148 OpenSSL de-inits, but an application is still using it). The affected
1149 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
1150 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
1151 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
1152 COMP_zlib_cleanup().
1155 *) --strict-warnings no longer enables runtime debugging options
1156 such as REF_DEBUG. Instead, debug options are automatically
1157 enabled with '--debug' builds.
1158 [Andy Polyakov, Emilia Käsper]
1160 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
1161 have been moved out of the public header files. New functions for managing
1162 these have been added.
1165 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
1166 objects have been moved out of the public header files. New
1167 functions for managing these have been added.
1170 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1171 have been moved out of the public header files. New functions for managing
1172 these have been added.
1175 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1176 moved out of the public header files. New functions for managing these
1180 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
1183 *) Removed the mk1mf build scripts.
1186 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1187 it is always safe to #include a header now.
1190 *) Removed the aged BC-32 config and all its supporting scripts
1193 *) Removed support for Ultrix, Netware, and OS/2.
1196 *) Add support for HKDF.
1197 [Alessandro Ghedini]
1199 *) Add support for blake2b and blake2s
1202 *) Added support for "pipelining". Ciphers that have the
1203 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1204 encryptions/decryptions simultaneously. There are currently no built-in
1205 ciphers with this property but the expectation is that engines will be able
1206 to offer it to significantly improve throughput. Support has been extended
1207 into libssl so that multiple records for a single connection can be
1208 processed in one go (for >=TLS 1.1).
1211 *) Added the AFALG engine. This is an async capable engine which is able to
1212 offload work to the Linux kernel. In this initial version it only supports
1213 AES128-CBC. The kernel must be version 4.1.0 or greater.
1216 *) OpenSSL now uses a new threading API. It is no longer necessary to
1217 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1218 are two supported threading models: pthreads and windows threads. It is
1219 also possible to configure OpenSSL at compile time for "no-threads". The
1220 old threading API should no longer be used. The functions have been
1221 replaced with "no-op" compatibility macros.
1222 [Alessandro Ghedini, Matt Caswell]
1224 *) Modify behavior of ALPN to invoke callback after SNI/servername
1225 callback, such that updates to the SSL_CTX affect ALPN.
1228 *) Add SSL_CIPHER queries for authentication and key-exchange.
1231 *) Changes to the DEFAULT cipherlist:
1232 - Prefer (EC)DHE handshakes over plain RSA.
1233 - Prefer AEAD ciphers over legacy ciphers.
1234 - Prefer ECDSA over RSA when both certificates are available.
1235 - Prefer TLSv1.2 ciphers/PRF.
1236 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1240 *) Change the ECC default curve list to be this, in order: x25519,
1241 secp256r1, secp521r1, secp384r1.
1244 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1245 disabled by default. They can be re-enabled using the
1246 enable-weak-ssl-ciphers option to Configure.
1249 *) If the server has ALPN configured, but supports no protocols that the
1250 client advertises, send a fatal "no_application_protocol" alert.
1251 This behaviour is SHALL in RFC 7301, though it isn't universally
1252 implemented by other servers.
1255 *) Add X25519 support.
1256 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1257 for public and private key encoding using the format documented in
1258 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1259 key generation and key derivation.
1261 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1265 *) Deprecate SRP_VBASE_get_by_user.
1266 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1267 In order to fix an unavoidable memory leak (CVE-2016-0798),
1268 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1269 seed, even if the seed is configured.
1271 Users should use SRP_VBASE_get1_by_user instead. Note that in
1272 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1273 also that even though configuring the SRP seed attempts to hide
1274 invalid usernames by continuing the handshake with fake
1275 credentials, this behaviour is not constant time and no strong
1276 guarantees are made that the handshake is indistinguishable from
1277 that of a valid user.
1280 *) Configuration change; it's now possible to build dynamic engines
1281 without having to build shared libraries and vice versa. This
1282 only applies to the engines in engines/, those in crypto/engine/
1283 will always be built into libcrypto (i.e. "static").
1285 Building dynamic engines is enabled by default; to disable, use
1286 the configuration option "disable-dynamic-engine".
1288 The only requirements for building dynamic engines are the
1289 presence of the DSO module and building with position independent
1290 code, so they will also automatically be disabled if configuring
1291 with "disable-dso" or "disable-pic".
1293 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1294 are also taken away from openssl/opensslconf.h, as they are
1298 *) Configuration change; if there is a known flag to compile
1299 position independent code, it will always be applied on the
1300 libcrypto and libssl object files, and never on the application
1301 object files. This means other libraries that use routines from
1302 libcrypto / libssl can be made into shared libraries regardless
1303 of how OpenSSL was configured.
1305 If this isn't desirable, the configuration options "disable-pic"
1306 or "no-pic" can be used to disable the use of PIC. This will
1307 also disable building shared libraries and dynamic engines.
1310 *) Removed JPAKE code. It was experimental and has no wide use.
1313 *) The INSTALL_PREFIX Makefile variable has been renamed to
1314 DESTDIR. That makes for less confusion on what this variable
1315 is for. Also, the configuration option --install_prefix is
1319 *) Heartbeat for TLS has been removed and is disabled by default
1320 for DTLS; configure with enable-heartbeats. Code that uses the
1321 old #define's might need to be updated.
1322 [Emilia Käsper, Rich Salz]
1324 *) Rename REF_CHECK to REF_DEBUG.
1327 *) New "unified" build system
1329 The "unified" build system is aimed to be a common system for all
1330 platforms we support. With it comes new support for VMS.
1332 This system builds supports building in a different directory tree
1333 than the source tree. It produces one Makefile (for unix family
1334 or lookalikes), or one descrip.mms (for VMS).
1336 The source of information to make the Makefile / descrip.mms is
1337 small files called 'build.info', holding the necessary
1338 information for each directory with source to compile, and a
1339 template in Configurations, like unix-Makefile.tmpl or
1342 With this change, the library names were also renamed on Windows
1343 and on VMS. They now have names that are closer to the standard
1344 on Unix, and include the major version number, and in certain
1345 cases, the architecture they are built for. See "Notes on shared
1346 libraries" in INSTALL.
1348 We rely heavily on the perl module Text::Template.
1351 *) Added support for auto-initialisation and de-initialisation of the library.
1352 OpenSSL no longer requires explicit init or deinit routines to be called,
1353 except in certain circumstances. See the OPENSSL_init_crypto() and
1354 OPENSSL_init_ssl() man pages for further information.
1357 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1358 "peer" argument is now expected to be a BIO_ADDR object.
1360 *) Rewrite of BIO networking library. The BIO library lacked consistent
1361 support of IPv6, and adding it required some more extensive
1362 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1363 which hold all types of addresses and chains of address information.
1364 It also introduces a new API, with functions like BIO_socket,
1365 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1366 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1367 have been adapted accordingly.
1370 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1374 *) CRIME protection: disable compression by default, even if OpenSSL is
1375 compiled with zlib enabled. Applications can still enable compression
1376 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1377 using the SSL_CONF library to configure compression.
1380 *) The signature of the session callback configured with
1381 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1382 was explicitly marked as 'const unsigned char*' instead of
1386 *) Always DPURIFY. Remove the use of uninitialized memory in the
1387 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1390 *) Removed many obsolete configuration items, including
1391 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1392 MD2_CHAR, MD2_INT, MD2_LONG
1394 IDEA_SHORT, IDEA_LONG
1395 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1396 [Rich Salz, with advice from Andy Polyakov]
1398 *) Many BN internals have been moved to an internal header file.
1399 [Rich Salz with help from Andy Polyakov]
1401 *) Configuration and writing out the results from it has changed.
1402 Files such as Makefile include/openssl/opensslconf.h and are now
1403 produced through general templates, such as Makefile.in and
1404 crypto/opensslconf.h.in and some help from the perl module
1407 Also, the center of configuration information is no longer
1408 Makefile. Instead, Configure produces a perl module in
1409 configdata.pm which holds most of the config data (in the hash
1410 table %config), the target data that comes from the target
1411 configuration in one of the Configurations/*.conf files (in
1415 *) To clarify their intended purposes, the Configure options
1416 --prefix and --openssldir change their semantics, and become more
1417 straightforward and less interdependent.
1419 --prefix shall be used exclusively to give the location INSTALLTOP
1420 where programs, scripts, libraries, include files and manuals are
1421 going to be installed. The default is now /usr/local.
1423 --openssldir shall be used exclusively to give the default
1424 location OPENSSLDIR where certificates, private keys, CRLs are
1425 managed. This is also where the default openssl.cnf gets
1427 If the directory given with this option is a relative path, the
1428 values of both the --prefix value and the --openssldir value will
1429 be combined to become OPENSSLDIR.
1430 The default for --openssldir is INSTALLTOP/ssl.
1432 Anyone who uses --openssldir to specify where OpenSSL is to be
1433 installed MUST change to use --prefix instead.
1436 *) The GOST engine was out of date and therefore it has been removed. An up
1437 to date GOST engine is now being maintained in an external repository.
1438 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1439 support for GOST ciphersuites (these are only activated if a GOST engine
1443 *) EGD is no longer supported by default; use enable-egd when
1445 [Ben Kaduk and Rich Salz]
1447 *) The distribution now has Makefile.in files, which are used to
1448 create Makefile's when Configure is run. *Configure must be run
1449 before trying to build now.*
1452 *) The return value for SSL_CIPHER_description() for error conditions
1456 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1458 Obtaining and performing DNSSEC validation of TLSA records is
1459 the application's responsibility. The application provides
1460 the TLSA records of its choice to OpenSSL, and these are then
1461 used to authenticate the peer.
1463 The TLSA records need not even come from DNS. They can, for
1464 example, be used to implement local end-entity certificate or
1465 trust-anchor "pinning", where the "pin" data takes the form
1466 of TLSA records, which can augment or replace verification
1467 based on the usual WebPKI public certification authorities.
1470 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1471 continues to support deprecated interfaces in default builds.
1472 However, applications are strongly advised to compile their
1473 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1474 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1475 or the 1.1.0 releases.
1477 In environments in which all applications have been ported to
1478 not use any deprecated interfaces OpenSSL's Configure script
1479 should be used with the --api=1.1.0 option to entirely remove
1480 support for the deprecated features from the library and
1481 unconditionally disable them in the installed headers.
1482 Essentially the same effect can be achieved with the "no-deprecated"
1483 argument to Configure, except that this will always restrict
1484 the build to just the latest API, rather than a fixed API
1487 As applications are ported to future revisions of the API,
1488 they should update their compile-time OPENSSL_API_COMPAT define
1489 accordingly, but in most cases should be able to continue to
1490 compile with later releases.
1492 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1493 0x10000000L and 0x00908000L, respectively. However those
1494 versions did not support the OPENSSL_API_COMPAT feature, and
1495 so applications are not typically tested for explicit support
1496 of just the undeprecated features of either release.
1499 *) Add support for setting the minimum and maximum supported protocol.
1500 It can bet set via the SSL_set_min_proto_version() and
1501 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1502 MaxProtocol. It's recommended to use the new APIs to disable
1503 protocols instead of disabling individual protocols using
1504 SSL_set_options() or SSL_CONF's Protocol. This change also
1505 removes support for disabling TLS 1.2 in the OpenSSL TLS
1506 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1509 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1512 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1513 and integrates ECDSA and ECDH functionality into EC. Implementations can
1514 now redirect key generation and no longer need to convert to or from
1517 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1518 include the ec.h header file instead.
1521 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1522 ciphers who are no longer supported and drops support the ephemeral RSA key
1523 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1526 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1527 opaque. For HMAC_CTX, the following constructors and destructors
1530 HMAC_CTX *HMAC_CTX_new(void);
1531 void HMAC_CTX_free(HMAC_CTX *ctx);
1533 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1534 destroy such methods has been added. See EVP_MD_meth_new(3) and
1535 EVP_CIPHER_meth_new(3) for documentation.
1538 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1539 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1540 EVP_MD_CTX_reset() should be called instead to reinitialise
1541 an already created structure.
1542 2) For consistency with the majority of our object creators and
1543 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1544 EVP_MD_CTX_(new|free). The old names are retained as macros
1545 for deprecated builds.
1548 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1549 cryptographic operations to be performed asynchronously as long as an
1550 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1551 further details. Libssl has also had this capability integrated with the
1552 introduction of the new mode SSL_MODE_ASYNC and associated error
1553 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1554 pages. This work was developed in partnership with Intel Corp.
1557 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1558 always enabled now. If you want to disable the support you should
1559 exclude it using the list of supported ciphers. This also means that the
1560 "-no_ecdhe" option has been removed from s_server.
1563 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1564 SSL_{CTX_}set1_curves() which can set a list.
1567 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1568 curve you want to support using SSL_{CTX_}set1_curves().
1571 *) State machine rewrite. The state machine code has been significantly
1572 refactored in order to remove much duplication of code and solve issues
1573 with the old code (see ssl/statem/README for further details). This change
1574 does have some associated API changes. Notably the SSL_state() function
1575 has been removed and replaced by SSL_get_state which now returns an
1576 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1577 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1581 *) All instances of the string "ssleay" in the public API were replaced
1582 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1583 Some error codes related to internal RSA_eay API's were renamed.
1586 *) The demo files in crypto/threads were moved to demo/threads.
1589 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1591 [Matt Caswell, Rich Salz]
1593 *) New ASN.1 embed macro.
1595 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1596 structure is not allocated: it is part of the parent. That is instead of
1604 This reduces memory fragmentation and make it impossible to accidentally
1605 set a mandatory field to NULL.
1607 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1608 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1609 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1613 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1616 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1617 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1618 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1619 DES and RC4 ciphersuites.
1622 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1623 This changes the decoding behaviour for some invalid messages,
1624 though the change is mostly in the more lenient direction, and
1625 legacy behaviour is preserved as much as possible.
1628 *) Fix no-stdio build.
1629 [ David Woodhouse <David.Woodhouse@intel.com> and also
1630 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1632 *) New testing framework
1633 The testing framework has been largely rewritten and is now using
1634 perl and the perl modules Test::Harness and an extended variant of
1635 Test::More called OpenSSL::Test to do its work. All test scripts in
1636 test/ have been rewritten into test recipes, and all direct calls to
1637 executables in test/Makefile have become individual recipes using the
1638 simplified testing OpenSSL::Test::Simple.
1640 For documentation on our testing modules, do:
1642 perldoc test/testlib/OpenSSL/Test/Simple.pm
1643 perldoc test/testlib/OpenSSL/Test.pm
1647 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1648 are used; the latter aborts on memory leaks (usually checked on exit).
1649 Some undocumented "set malloc, etc., hooks" functions were removed
1650 and others were changed. All are now documented.
1653 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1655 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1657 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1658 from RFC4279, RFC4785, RFC5487, RFC5489.
1660 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1661 original RSA_PSK patch.
1664 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1665 era flag was never set throughout the codebase (only read). Also removed
1666 SSL3_FLAGS_POP_BUFFER which was only used if
1667 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1670 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1671 to be "oneline" instead of "compat".
1674 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1675 not aware of clients that still exhibit this bug, and the workaround
1676 hasn't been working properly for a while.
1679 *) The return type of BIO_number_read() and BIO_number_written() as well as
1680 the corresponding num_read and num_write members in the BIO structure has
1681 changed from unsigned long to uint64_t. On platforms where an unsigned
1682 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1686 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1687 OpenSSL without support for them. It also means that maintaining
1688 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1689 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1692 *) Removed support for the two export grade static DH ciphersuites
1693 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1694 were newly added (along with a number of other static DH ciphersuites) to
1695 1.0.2. However the two export ones have *never* worked since they were
1696 introduced. It seems strange in any case to be adding new export
1697 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1700 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1701 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1702 and turned into macros which simply call the new preferred function names
1703 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1704 should use the new names instead. Also as part of this change the ssl23.h
1705 header file has been removed.
1708 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1709 code and the associated standard is no longer considered fit-for-purpose.
1712 *) RT2547 was closed. When generating a private key, try to make the
1713 output file readable only by the owner. This behavior change might
1714 be noticeable when interacting with other software.
1716 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1720 *) Added HTTP GET support to the ocsp command.
1723 *) Changed default digest for the dgst and enc commands from MD5 to
1727 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1730 *) Added support for TLS extended master secret from
1731 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1732 initial patch which was a great help during development.
1735 *) All libssl internal structures have been removed from the public header
1736 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1737 now redundant). Users should not attempt to access internal structures
1738 directly. Instead they should use the provided API functions.
1741 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1742 Access to deprecated functions can be re-enabled by running config with
1743 "enable-deprecated". In addition applications wishing to use deprecated
1744 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1745 will, by default, disable some transitive includes that previously existed
1746 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1749 *) Added support for OCB mode. OpenSSL has been granted a patent license
1750 compatible with the OpenSSL license for use of OCB. Details are available
1751 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1752 for OCB can be removed by calling config with no-ocb.
1755 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1756 compatible client hello.
1759 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1760 done while fixing the error code for the key-too-small case.
1761 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1763 *) CA.sh has been removed; use CA.pl instead.
1766 *) Removed old DES API.
1769 *) Remove various unsupported platforms:
1775 Sinix/ReliantUNIX RM400
1780 16-bit platforms such as WIN16
1783 *) Clean up OPENSSL_NO_xxx #define's
1784 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1785 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1786 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1787 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1788 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1789 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1790 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1791 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1792 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1793 Remove MS_STATIC; it's a relic from platforms <32 bits.
1796 *) Cleaned up dead code
1797 Remove all but one '#ifdef undef' which is to be looked at.
1800 *) Clean up calling of xxx_free routines.
1801 Just like free(), fix most of the xxx_free routines to accept
1802 NULL. Remove the non-null checks from callers. Save much code.
1805 *) Add secure heap for storage of private keys (when possible).
1806 Add BIO_s_secmem(), CBIGNUM, etc.
1807 Contributed by Akamai Technologies under our Corporate CLA.
1810 *) Experimental support for a new, fast, unbiased prime candidate generator,
1811 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1812 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1814 *) New output format NSS in the sess_id command line tool. This allows
1815 exporting the session id and the master key in NSS keylog format.
1816 [Martin Kaiser <martin@kaiser.cx>]
1818 *) Harmonize version and its documentation. -f flag is used to display
1820 [mancha <mancha1@zoho.com>]
1822 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1823 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1824 [mancha <mancha1@zoho.com>]
1826 *) Fix some double frees. These are not thought to be exploitable.
1827 [mancha <mancha1@zoho.com>]
1829 *) A missing bounds check in the handling of the TLS heartbeat extension
1830 can be used to reveal up to 64k of memory to a connected client or
1833 Thanks for Neel Mehta of Google Security for discovering this bug and to
1834 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1835 preparing the fix (CVE-2014-0160)
1836 [Adam Langley, Bodo Moeller]
1838 *) Fix for the attack described in the paper "Recovering OpenSSL
1839 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1840 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1841 http://eprint.iacr.org/2014/140
1843 Thanks to Yuval Yarom and Naomi Benger for discovering this
1844 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1845 [Yuval Yarom and Naomi Benger]
1847 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1848 this fixes a limitation in previous versions of OpenSSL.
1851 *) Experimental encrypt-then-mac support.
1853 Experimental support for encrypt then mac from
1854 draft-gutmann-tls-encrypt-then-mac-02.txt
1856 To enable it set the appropriate extension number (0x42 for the test
1857 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1859 For non-compliant peers (i.e. just about everything) this should have no
1862 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1866 *) Add EVP support for key wrapping algorithms, to avoid problems with
1867 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1868 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1869 algorithms and include tests cases.
1872 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1876 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1877 MGF1 digest and OAEP label.
1880 *) Make openssl verify return errors.
1881 [Chris Palmer <palmer@google.com> and Ben Laurie]
1883 *) New function ASN1_TIME_diff to calculate the difference between two
1884 ASN1_TIME structures or one structure and the current time.
1887 *) Update fips_test_suite to support multiple command line options. New
1888 test to induce all self test errors in sequence and check expected
1892 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1893 sign or verify all in one operation.
1896 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1897 test programs and fips_test_suite. Includes functionality to parse
1898 the minimal script output of fipsalgest.pl directly.
1901 *) Add authorisation parameter to FIPS_module_mode_set().
1904 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1907 *) Use separate DRBG fields for internal and external flags. New function
1908 FIPS_drbg_health_check() to perform on demand health checking. Add
1909 generation tests to fips_test_suite with reduced health check interval to
1910 demonstrate periodic health checking. Add "nodh" option to
1911 fips_test_suite to skip very slow DH test.
1914 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1918 *) More extensive health check for DRBG checking many more failure modes.
1919 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1920 combination: call this in fips_test_suite.
1923 *) Add support for canonical generation of DSA parameter 'g'. See
1926 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1927 POST to handle HMAC cases.
1930 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1931 to return numerical and string versions of the FIPS module number.
1934 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1935 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1936 outside the validated module in the FIPS capable OpenSSL.
1939 *) Minor change to DRBG entropy callback semantics. In some cases
1940 there is no multiple of the block length between min_len and
1941 max_len. Allow the callback to return more than max_len bytes
1942 of entropy but discard any extra: it is the callback's responsibility
1943 to ensure that the extra data discarded does not impact the
1944 requested amount of entropy.
1947 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1948 information in FIPS186-3, SP800-57 and SP800-131A.
1951 *) CCM support via EVP. Interface is very similar to GCM case except we
1952 must supply all data in one chunk (i.e. no update, final) and the
1953 message length must be supplied if AAD is used. Add algorithm test
1957 *) Initial version of POST overhaul. Add POST callback to allow the status
1958 of POST to be monitored and/or failures induced. Modify fips_test_suite
1959 to use callback. Always run all selftests even if one fails.
1962 *) XTS support including algorithm test driver in the fips_gcmtest program.
1963 Note: this does increase the maximum key length from 32 to 64 bytes but
1964 there should be no binary compatibility issues as existing applications
1965 will never use XTS mode.
1968 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1969 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1970 performs algorithm blocking for unapproved PRNG types. Also do not
1971 set PRNG type in FIPS_mode_set(): leave this to the application.
1972 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1973 the standard OpenSSL PRNG: set additional data to a date time vector.
1976 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1977 This shouldn't present any incompatibility problems because applications
1978 shouldn't be using these directly and any that are will need to rethink
1979 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1982 *) Extensive self tests and health checking required by SP800-90 DRBG.
1983 Remove strength parameter from FIPS_drbg_instantiate and always
1984 instantiate at maximum supported strength.
1987 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1990 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
1993 *) New function DH_compute_key_padded() to compute a DH key and pad with
1994 leading zeroes if needed: this complies with SP800-56A et al.
1997 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
1998 anything, incomplete, subject to change and largely untested at present.
2001 *) Modify fipscanisteronly build option to only build the necessary object
2002 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
2005 *) Add experimental option FIPSSYMS to give all symbols in
2006 fipscanister.o and FIPS or fips prefix. This will avoid
2007 conflicts with future versions of OpenSSL. Add perl script
2008 util/fipsas.pl to preprocess assembly language source files
2009 and rename any affected symbols.
2012 *) Add selftest checks and algorithm block of non-fips algorithms in
2013 FIPS mode. Remove DES2 from selftests.
2016 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
2017 return internal method without any ENGINE dependencies. Add new
2018 tiny fips sign and verify functions.
2021 *) New build option no-ec2m to disable characteristic 2 code.
2024 *) New build option "fipscanisteronly". This only builds fipscanister.o
2025 and (currently) associated fips utilities. Uses the file Makefile.fips
2026 instead of Makefile.org as the prototype.
2029 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
2030 Update fips_gcmtest to use IV generator.
2033 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
2034 setting output buffer to NULL. The *Final function must be
2035 called although it will not retrieve any additional data. The tag
2036 can be set or retrieved with a ctrl. The IV length is by default 12
2037 bytes (96 bits) but can be set to an alternative value. If the IV
2038 length exceeds the maximum IV length (currently 16 bytes) it cannot be
2042 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
2043 underlying do_cipher function handles all cipher semantics itself
2044 including padding and finalisation. This is useful if (for example)
2045 an ENGINE cipher handles block padding itself. The behaviour of
2046 do_cipher is subtly changed if this flag is set: the return value
2047 is the number of characters written to the output buffer (zero is
2048 no longer an error code) or a negative error code. Also if the
2049 input buffer is NULL and length 0 finalisation should be performed.
2052 *) If a candidate issuer certificate is already part of the constructed
2053 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
2056 *) Improve forward-security support: add functions
2058 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
2059 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
2061 for use by SSL/TLS servers; the callback function will be called whenever a
2062 new session is created, and gets to decide whether the session may be
2063 cached to make it resumable (return 0) or not (return 1). (As by the
2064 SSL/TLS protocol specifications, the session_id sent by the server will be
2065 empty to indicate that the session is not resumable; also, the server will
2066 not generate RFC 4507 (RFC 5077) session tickets.)
2068 A simple reasonable callback implementation is to return is_forward_secure.
2069 This parameter will be set to 1 or 0 depending on the ciphersuite selected
2070 by the SSL/TLS server library, indicating whether it can provide forward
2072 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2074 *) New -verify_name option in command line utilities to set verification
2078 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
2079 Add CMAC pkey methods.
2082 *) Experimental renegotiation in s_server -www mode. If the client
2083 browses /reneg connection is renegotiated. If /renegcert it is
2084 renegotiated requesting a certificate.
2087 *) Add an "external" session cache for debugging purposes to s_server. This
2088 should help trace issues which normally are only apparent in deployed
2089 multi-process servers.
2092 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
2093 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
2094 BIO_set_cipher() and some obscure PEM functions were changed so they
2095 can now return an error. The RAND changes required a change to the
2096 RAND_METHOD structure.
2099 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
2100 a gcc attribute to warn if the result of a function is ignored. This
2101 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
2102 whose return value is often ignored.
2105 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
2106 These allow SCTs (signed certificate timestamps) to be requested and
2107 validated when establishing a connection.
2108 [Rob Percival <robpercival@google.com>]
2110 Changes between 1.0.2g and 1.0.2h [3 May 2016]
2112 *) Prevent padding oracle in AES-NI CBC MAC check
2114 A MITM attacker can use a padding oracle attack to decrypt traffic
2115 when the connection uses an AES CBC cipher and the server support
2118 This issue was introduced as part of the fix for Lucky 13 padding
2119 attack (CVE-2013-0169). The padding check was rewritten to be in
2120 constant time by making sure that always the same bytes are read and
2121 compared against either the MAC or padding bytes. But it no longer
2122 checked that there was enough data to have both the MAC and padding
2125 This issue was reported by Juraj Somorovsky using TLS-Attacker.
2129 *) Fix EVP_EncodeUpdate overflow
2131 An overflow can occur in the EVP_EncodeUpdate() function which is used for
2132 Base64 encoding of binary data. If an attacker is able to supply very large
2133 amounts of input data then a length check can overflow resulting in a heap
2136 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
2137 the PEM_write_bio* family of functions. These are mainly used within the
2138 OpenSSL command line applications, so any application which processes data
2139 from an untrusted source and outputs it as a PEM file should be considered
2140 vulnerable to this issue. User applications that call these APIs directly
2141 with large amounts of untrusted data may also be vulnerable.
2143 This issue was reported by Guido Vranken.
2147 *) Fix EVP_EncryptUpdate overflow
2149 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
2150 is able to supply very large amounts of input data after a previous call to
2151 EVP_EncryptUpdate() with a partial block then a length check can overflow
2152 resulting in a heap corruption. Following an analysis of all OpenSSL
2153 internal usage of the EVP_EncryptUpdate() function all usage is one of two
2154 forms. The first form is where the EVP_EncryptUpdate() call is known to be
2155 the first called function after an EVP_EncryptInit(), and therefore that
2156 specific call must be safe. The second form is where the length passed to
2157 EVP_EncryptUpdate() can be seen from the code to be some small value and
2158 therefore there is no possibility of an overflow. Since all instances are
2159 one of these two forms, it is believed that there can be no overflows in
2160 internal code due to this problem. It should be noted that
2161 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
2162 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
2163 of these calls have also been analysed too and it is believed there are no
2164 instances in internal usage where an overflow could occur.
2166 This issue was reported by Guido Vranken.
2170 *) Prevent ASN.1 BIO excessive memory allocation
2172 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2173 a short invalid encoding can cause allocation of large amounts of memory
2174 potentially consuming excessive resources or exhausting memory.
2176 Any application parsing untrusted data through d2i BIO functions is
2177 affected. The memory based functions such as d2i_X509() are *not* affected.
2178 Since the memory based functions are used by the TLS library, TLS
2179 applications are not affected.
2181 This issue was reported by Brian Carpenter.
2187 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2188 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2189 in arbitrary stack data being returned in the buffer.
2191 This issue was reported by Guido Vranken.
2195 *) Modify behavior of ALPN to invoke callback after SNI/servername
2196 callback, such that updates to the SSL_CTX affect ALPN.
2199 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2203 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2204 methods are enabled and ssl2 is disabled the methods return NULL.
2207 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2209 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2210 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2211 provide any "EXPORT" or "LOW" strength ciphers.
2214 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2215 is by default disabled at build-time. Builds that are not configured with
2216 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2217 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2218 will need to explicitly call either of:
2220 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2222 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2224 as appropriate. Even if either of those is used, or the application
2225 explicitly uses the version-specific SSLv2_method() or its client and
2226 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2227 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2228 ciphers, and SSLv2 56-bit DES are no longer available.
2232 *) Fix a double-free in DSA code
2234 A double free bug was discovered when OpenSSL parses malformed DSA private
2235 keys and could lead to a DoS attack or memory corruption for applications
2236 that receive DSA private keys from untrusted sources. This scenario is
2239 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2244 *) Disable SRP fake user seed to address a server memory leak.
2246 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2248 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2249 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2250 was changed to ignore the "fake user" SRP seed, even if the seed
2253 Users should use SRP_VBASE_get1_by_user instead. Note that in
2254 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2255 also that even though configuring the SRP seed attempts to hide
2256 invalid usernames by continuing the handshake with fake
2257 credentials, this behaviour is not constant time and no strong
2258 guarantees are made that the handshake is indistinguishable from
2259 that of a valid user.
2263 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2265 In the BN_hex2bn function the number of hex digits is calculated using an
2266 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2267 large values of |i| this can result in |bn_expand| not allocating any
2268 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2269 field as NULL leading to a subsequent NULL ptr deref. For very large values
2270 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2271 In this case memory is allocated to the internal BIGNUM data field, but it
2272 is insufficiently sized leading to heap corruption. A similar issue exists
2273 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2274 is ever called by user applications with very large untrusted hex/dec data.
2275 This is anticipated to be a rare occurrence.
2277 All OpenSSL internal usage of these functions use data that is not expected
2278 to be untrusted, e.g. config file data or application command line
2279 arguments. If user developed applications generate config file data based
2280 on untrusted data then it is possible that this could also lead to security
2281 consequences. This is also anticipated to be rare.
2283 This issue was reported to OpenSSL by Guido Vranken.
2287 *) Fix memory issues in BIO_*printf functions
2289 The internal |fmtstr| function used in processing a "%s" format string in
2290 the BIO_*printf functions could overflow while calculating the length of a
2291 string and cause an OOB read when printing very long strings.
2293 Additionally the internal |doapr_outch| function can attempt to write to an
2294 OOB memory location (at an offset from the NULL pointer) in the event of a
2295 memory allocation failure. In 1.0.2 and below this could be caused where
2296 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2297 could be in processing a very long "%s" format string. Memory leaks can
2300 The first issue may mask the second issue dependent on compiler behaviour.
2301 These problems could enable attacks where large amounts of untrusted data
2302 is passed to the BIO_*printf functions. If applications use these functions
2303 in this way then they could be vulnerable. OpenSSL itself uses these
2304 functions when printing out human-readable dumps of ASN.1 data. Therefore
2305 applications that print this data could be vulnerable if the data is from
2306 untrusted sources. OpenSSL command line applications could also be
2307 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2308 as command line arguments.
2310 Libssl is not considered directly vulnerable. Additionally certificates etc
2311 received via remote connections via libssl are also unlikely to be able to
2312 trigger these issues because of message size limits enforced within libssl.
2314 This issue was reported to OpenSSL Guido Vranken.
2318 *) Side channel attack on modular exponentiation
2320 A side-channel attack was found which makes use of cache-bank conflicts on
2321 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2322 of RSA keys. The ability to exploit this issue is limited as it relies on
2323 an attacker who has control of code in a thread running on the same
2324 hyper-threaded core as the victim thread which is performing decryptions.
2326 This issue was reported to OpenSSL by Yuval Yarom, The University of
2327 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2328 Nadia Heninger, University of Pennsylvania with more information at
2329 http://cachebleed.info.
2333 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2334 if no keysize is specified with default_bits. This fixes an
2335 omission in an earlier change that changed all RSA/DSA key generation
2336 apps to use 2048 bits by default.
2339 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2340 *) DH small subgroups
2342 Historically OpenSSL only ever generated DH parameters based on "safe"
2343 primes. More recently (in version 1.0.2) support was provided for
2344 generating X9.42 style parameter files such as those required for RFC 5114
2345 support. The primes used in such files may not be "safe". Where an
2346 application is using DH configured with parameters based on primes that are
2347 not "safe" then an attacker could use this fact to find a peer's private
2348 DH exponent. This attack requires that the attacker complete multiple
2349 handshakes in which the peer uses the same private DH exponent. For example
2350 this could be used to discover a TLS server's private DH exponent if it's
2351 reusing the private DH exponent or it's using a static DH ciphersuite.
2353 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2354 TLS. It is not on by default. If the option is not set then the server
2355 reuses the same private DH exponent for the life of the server process and
2356 would be vulnerable to this attack. It is believed that many popular
2357 applications do set this option and would therefore not be at risk.
2359 The fix for this issue adds an additional check where a "q" parameter is
2360 available (as is the case in X9.42 based parameters). This detects the
2361 only known attack, and is the only possible defense for static DH
2362 ciphersuites. This could have some performance impact.
2364 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2365 default and cannot be disabled. This could have some performance impact.
2367 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2371 *) SSLv2 doesn't block disabled ciphers
2373 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2374 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2375 been disabled, provided that the SSLv2 protocol was not also disabled via
2378 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2379 and Sebastian Schinzel.
2383 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2385 *) BN_mod_exp may produce incorrect results on x86_64
2387 There is a carry propagating bug in the x86_64 Montgomery squaring
2388 procedure. No EC algorithms are affected. Analysis suggests that attacks
2389 against RSA and DSA as a result of this defect would be very difficult to
2390 perform and are not believed likely. Attacks against DH are considered just
2391 feasible (although very difficult) because most of the work necessary to
2392 deduce information about a private key may be performed offline. The amount
2393 of resources required for such an attack would be very significant and
2394 likely only accessible to a limited number of attackers. An attacker would
2395 additionally need online access to an unpatched system using the target
2396 private key in a scenario with persistent DH parameters and a private
2397 key that is shared between multiple clients. For example this can occur by
2398 default in OpenSSL DHE based SSL/TLS ciphersuites.
2400 This issue was reported to OpenSSL by Hanno Böck.
2404 *) Certificate verify crash with missing PSS parameter
2406 The signature verification routines will crash with a NULL pointer
2407 dereference if presented with an ASN.1 signature using the RSA PSS
2408 algorithm and absent mask generation function parameter. Since these
2409 routines are used to verify certificate signature algorithms this can be
2410 used to crash any certificate verification operation and exploited in a
2411 DoS attack. Any application which performs certificate verification is
2412 vulnerable including OpenSSL clients and servers which enable client
2415 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2419 *) X509_ATTRIBUTE memory leak
2421 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2422 memory. This structure is used by the PKCS#7 and CMS routines so any
2423 application which reads PKCS#7 or CMS data from untrusted sources is
2424 affected. SSL/TLS is not affected.
2426 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2431 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2432 This changes the decoding behaviour for some invalid messages,
2433 though the change is mostly in the more lenient direction, and
2434 legacy behaviour is preserved as much as possible.
2437 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2439 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2441 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2443 *) Alternate chains certificate forgery
2445 During certificate verification, OpenSSL will attempt to find an
2446 alternative certificate chain if the first attempt to build such a chain
2447 fails. An error in the implementation of this logic can mean that an
2448 attacker could cause certain checks on untrusted certificates to be
2449 bypassed, such as the CA flag, enabling them to use a valid leaf
2450 certificate to act as a CA and "issue" an invalid certificate.
2452 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2456 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2458 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2459 incompatibility in the handling of HMAC. The previous ABI has now been
2463 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2465 *) Malformed ECParameters causes infinite loop
2467 When processing an ECParameters structure OpenSSL enters an infinite loop
2468 if the curve specified is over a specially malformed binary polynomial
2471 This can be used to perform denial of service against any
2472 system which processes public keys, certificate requests or
2473 certificates. This includes TLS clients and TLS servers with
2474 client authentication enabled.
2476 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2480 *) Exploitable out-of-bounds read in X509_cmp_time
2482 X509_cmp_time does not properly check the length of the ASN1_TIME
2483 string and can read a few bytes out of bounds. In addition,
2484 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2487 An attacker can use this to craft malformed certificates and CRLs of
2488 various sizes and potentially cause a segmentation fault, resulting in
2489 a DoS on applications that verify certificates or CRLs. TLS clients
2490 that verify CRLs are affected. TLS clients and servers with client
2491 authentication enabled may be affected if they use custom verification
2494 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2495 independently by Hanno Böck.
2499 *) PKCS7 crash with missing EnvelopedContent
2501 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2502 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2503 with missing content and trigger a NULL pointer dereference on parsing.
2505 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2506 structures from untrusted sources are affected. OpenSSL clients and
2507 servers are not affected.
2509 This issue was reported to OpenSSL by Michal Zalewski (Google).
2513 *) CMS verify infinite loop with unknown hash function
2515 When verifying a signedData message the CMS code can enter an infinite loop
2516 if presented with an unknown hash function OID. This can be used to perform
2517 denial of service against any system which verifies signedData messages using
2519 This issue was reported to OpenSSL by Johannes Bauer.
2523 *) Race condition handling NewSessionTicket
2525 If a NewSessionTicket is received by a multi-threaded client when attempting to
2526 reuse a previous ticket then a race condition can occur potentially leading to
2527 a double free of the ticket data.
2531 *) Only support 256-bit or stronger elliptic curves with the
2532 'ecdh_auto' setting (server) or by default (client). Of supported
2533 curves, prefer P-256 (both).
2536 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2538 *) ClientHello sigalgs DoS fix
2540 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2541 invalid signature algorithms extension a NULL pointer dereference will
2542 occur. This can be exploited in a DoS attack against the server.
2544 This issue was was reported to OpenSSL by David Ramos of Stanford
2547 [Stephen Henson and Matt Caswell]
2549 *) Multiblock corrupted pointer fix
2551 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2552 feature only applies on 64 bit x86 architecture platforms that support AES
2553 NI instructions. A defect in the implementation of "multiblock" can cause
2554 OpenSSL's internal write buffer to become incorrectly set to NULL when
2555 using non-blocking IO. Typically, when the user application is using a
2556 socket BIO for writing, this will only result in a failed connection.
2557 However if some other BIO is used then it is likely that a segmentation
2558 fault will be triggered, thus enabling a potential DoS attack.
2560 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2564 *) Segmentation fault in DTLSv1_listen fix
2566 The DTLSv1_listen function is intended to be stateless and processes the
2567 initial ClientHello from many peers. It is common for user code to loop
2568 over the call to DTLSv1_listen until a valid ClientHello is received with
2569 an associated cookie. A defect in the implementation of DTLSv1_listen means
2570 that state is preserved in the SSL object from one invocation to the next
2571 that can lead to a segmentation fault. Errors processing the initial
2572 ClientHello can trigger this scenario. An example of such an error could be
2573 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2576 This issue was reported to OpenSSL by Per Allansson.
2580 *) Segmentation fault in ASN1_TYPE_cmp fix
2582 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2583 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2584 certificate signature algorithm consistency this can be used to crash any
2585 certificate verification operation and exploited in a DoS attack. Any
2586 application which performs certificate verification is vulnerable including
2587 OpenSSL clients and servers which enable client authentication.
2591 *) Segmentation fault for invalid PSS parameters fix
2593 The signature verification routines will crash with a NULL pointer
2594 dereference if presented with an ASN.1 signature using the RSA PSS
2595 algorithm and invalid parameters. Since these routines are used to verify
2596 certificate signature algorithms this can be used to crash any
2597 certificate verification operation and exploited in a DoS attack. Any
2598 application which performs certificate verification is vulnerable including
2599 OpenSSL clients and servers which enable client authentication.
2601 This issue was was reported to OpenSSL by Brian Carpenter.
2605 *) ASN.1 structure reuse memory corruption fix
2607 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2608 memory corruption via an invalid write. Such reuse is and has been
2609 strongly discouraged and is believed to be rare.
2611 Applications that parse structures containing CHOICE or ANY DEFINED BY
2612 components may be affected. Certificate parsing (d2i_X509 and related
2613 functions) are however not affected. OpenSSL clients and servers are
2618 *) PKCS7 NULL pointer dereferences fix
2620 The PKCS#7 parsing code does not handle missing outer ContentInfo
2621 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2622 missing content and trigger a NULL pointer dereference on parsing.
2624 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2625 otherwise parse PKCS#7 structures from untrusted sources are
2626 affected. OpenSSL clients and servers are not affected.
2628 This issue was reported to OpenSSL by Michal Zalewski (Google).
2632 *) DoS via reachable assert in SSLv2 servers fix
2634 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2635 servers that both support SSLv2 and enable export cipher suites by sending
2636 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2638 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2639 (OpenSSL development team).
2643 *) Empty CKE with client auth and DHE fix
2645 If client auth is used then a server can seg fault in the event of a DHE
2646 ciphersuite being selected and a zero length ClientKeyExchange message
2647 being sent by the client. This could be exploited in a DoS attack.
2651 *) Handshake with unseeded PRNG fix
2653 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2654 with an unseeded PRNG. The conditions are:
2655 - The client is on a platform where the PRNG has not been seeded
2656 automatically, and the user has not seeded manually
2657 - A protocol specific client method version has been used (i.e. not
2658 SSL_client_methodv23)
2659 - A ciphersuite is used that does not require additional random data from
2660 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2662 If the handshake succeeds then the client random that has been used will
2663 have been generated from a PRNG with insufficient entropy and therefore the
2664 output may be predictable.
2666 For example using the following command with an unseeded openssl will
2667 succeed on an unpatched platform:
2669 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2673 *) Use After Free following d2i_ECPrivatekey error fix
2675 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2676 could cause a use after free condition. This, in turn, could cause a double
2677 free in several private key parsing functions (such as d2i_PrivateKey
2678 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2679 for applications that receive EC private keys from untrusted
2680 sources. This scenario is considered rare.
2682 This issue was discovered by the BoringSSL project and fixed in their
2687 *) X509_to_X509_REQ NULL pointer deref fix
2689 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2690 the certificate key is invalid. This function is rarely used in practice.
2692 This issue was discovered by Brian Carpenter.
2696 *) Removed the export ciphers from the DEFAULT ciphers
2699 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2701 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2702 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2703 So far those who have to target multiple platforms would compromise
2704 and argue that binary targeting say ARMv5 would still execute on
2705 ARMv8. "Universal" build resolves this compromise by providing
2706 near-optimal performance even on newer platforms.
2709 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2710 (other platforms pending).
2711 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2713 *) Add support for the SignedCertificateTimestampList certificate and
2714 OCSP response extensions from RFC6962.
2717 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2718 for corner cases. (Certain input points at infinity could lead to
2719 bogus results, with non-infinity inputs mapped to infinity too.)
2722 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2723 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2724 common cases are optimized and there still is room for further
2725 improvements. Vector Permutation AES for Altivec is also added.
2728 *) Add support for little-endian ppc64 Linux target.
2729 [Marcelo Cerri (IBM)]
2731 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2732 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2733 are optimized and there still is room for further improvements.
2734 Both 32- and 64-bit modes are supported.
2735 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2737 *) Improved ARMv7 NEON support.
2740 *) Support for SPARC Architecture 2011 crypto extensions, first
2741 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2742 SHA256/512, MD5, GHASH and modular exponentiation.
2743 [Andy Polyakov, David Miller]
2745 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2747 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2749 *) Support for new and upcoming Intel processors, including AVX2,
2750 BMI and SHA ISA extensions. This includes additional "stitched"
2751 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2754 This work was sponsored by Intel Corp.
2757 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2758 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2759 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2762 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2763 this fixes a limitation in previous versions of OpenSSL.
2766 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2767 MGF1 digest and OAEP label.
2770 *) Add EVP support for key wrapping algorithms, to avoid problems with
2771 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2772 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2773 algorithms and include tests cases.
2776 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2778 [Douglas E. Engert, Steve Henson]
2780 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2781 difference in days and seconds between two tm or ASN1_TIME structures.
2784 *) Add -rev test option to s_server to just reverse order of characters
2785 received by client and send back to server. Also prints an abbreviated
2786 summary of the connection parameters.
2789 *) New option -brief for s_client and s_server to print out a brief summary
2790 of connection parameters.
2793 *) Add callbacks for arbitrary TLS extensions.
2794 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2796 *) New option -crl_download in several openssl utilities to download CRLs
2797 from CRLDP extension in certificates.
2800 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2803 *) New function X509_CRL_diff to generate a delta CRL from the difference
2804 of two full CRLs. Add support to "crl" utility.
2807 *) New functions to set lookup_crls function and to retrieve
2808 X509_STORE from X509_STORE_CTX.
2811 *) Print out deprecated issuer and subject unique ID fields in
2815 *) Extend OCSP I/O functions so they can be used for simple general purpose
2816 HTTP as well as OCSP. New wrapper function which can be used to download
2817 CRLs using the OCSP API.
2820 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2823 *) SSL_CONF* functions. These provide a common framework for application
2824 configuration using configuration files or command lines.
2827 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2828 message callback and prints the results. Needs compile time option
2829 "enable-ssl-trace". New options to s_client and s_server to enable
2833 *) New ctrl and macro to retrieve supported points extensions.
2834 Print out extension in s_server and s_client.
2837 *) New functions to retrieve certificate signature and signature
2841 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2845 *) New Suite B modes for TLS code. These use and enforce the requirements
2846 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2847 only use Suite B curves. The Suite B modes can be set by using the
2848 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2851 *) New chain verification flags for Suite B levels of security. Check
2852 algorithms are acceptable when flags are set in X509_verify_cert.
2855 *) Make tls1_check_chain return a set of flags indicating checks passed
2856 by a certificate chain. Add additional tests to handle client
2857 certificates: checks for matching certificate type and issuer name
2861 *) If an attempt is made to use a signature algorithm not in the peer
2862 preference list abort the handshake. If client has no suitable
2863 signature algorithms in response to a certificate request do not
2864 use the certificate.
2867 *) If server EC tmp key is not in client preference list abort handshake.
2870 *) Add support for certificate stores in CERT structure. This makes it
2871 possible to have different stores per SSL structure or one store in
2872 the parent SSL_CTX. Include distinct stores for certificate chain
2873 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2874 to build and store a certificate chain in CERT structure: returning
2875 an error if the chain cannot be built: this will allow applications
2876 to test if a chain is correctly configured.
2878 Note: if the CERT based stores are not set then the parent SSL_CTX
2879 store is used to retain compatibility with existing behaviour.
2883 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2884 mask based on the current session, check mask when sending client
2885 hello and checking the requested ciphersuite.
2888 *) New ctrls to retrieve and set certificate types in a certificate
2889 request message. Print out received values in s_client. If certificate
2890 types is not set with custom values set sensible values based on
2891 supported signature algorithms.
2894 *) Support for distinct client and server supported signature algorithms.
2897 *) Add certificate callback. If set this is called whenever a certificate
2898 is required by client or server. An application can decide which
2899 certificate chain to present based on arbitrary criteria: for example
2900 supported signature algorithms. Add very simple example to s_server.
2901 This fixes many of the problems and restrictions of the existing client
2902 certificate callback: for example you can now clear an existing
2903 certificate and specify the whole chain.
2906 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2907 the certificate can be used for (if anything). Set valid_flags field
2908 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2909 to have similar checks in it.
2911 Add new "cert_flags" field to CERT structure and include a "strict mode".
2912 This enforces some TLS certificate requirements (such as only permitting
2913 certificate signature algorithms contained in the supported algorithms
2914 extension) which some implementations ignore: this option should be used
2915 with caution as it could cause interoperability issues.
2918 *) Update and tidy signature algorithm extension processing. Work out
2919 shared signature algorithms based on preferences and peer algorithms
2920 and print them out in s_client and s_server. Abort handshake if no
2921 shared signature algorithms.
2924 *) Add new functions to allow customised supported signature algorithms
2925 for SSL and SSL_CTX structures. Add options to s_client and s_server
2929 *) New function SSL_certs_clear() to delete all references to certificates
2930 from an SSL structure. Before this once a certificate had been added
2931 it couldn't be removed.
2934 *) Integrate hostname, email address and IP address checking with certificate
2935 verification. New verify options supporting checking in openssl utility.
2938 *) Fixes and wildcard matching support to hostname and email checking
2939 functions. Add manual page.
2940 [Florian Weimer (Red Hat Product Security Team)]
2942 *) New functions to check a hostname email or IP address against a
2943 certificate. Add options x509 utility to print results of checks against
2947 *) Fix OCSP checking.
2948 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2950 *) Initial experimental support for explicitly trusted non-root CAs.
2951 OpenSSL still tries to build a complete chain to a root but if an
2952 intermediate CA has a trust setting included that is used. The first
2953 setting is used: whether to trust (e.g., -addtrust option to the x509
2957 *) Add -trusted_first option which attempts to find certificates in the
2958 trusted store even if an untrusted chain is also supplied.
2961 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2962 platform support for Linux and Android.
2965 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2968 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2969 When in FIPS mode the approved implementations are used as normal,
2970 when not in FIPS mode the internal unapproved versions are used instead.
2971 This means that the FIPS capable OpenSSL isn't forced to use the
2972 (often lower performance) FIPS implementations outside FIPS mode.
2975 *) Transparently support X9.42 DH parameters when calling
2976 PEM_read_bio_DHparameters. This means existing applications can handle
2977 the new parameter format automatically.
2980 *) Initial experimental support for X9.42 DH parameter format: mainly
2981 to support use of 'q' parameter for RFC5114 parameters.
2984 *) Add DH parameters from RFC5114 including test data to dhtest.
2987 *) Support for automatic EC temporary key parameter selection. If enabled
2988 the most preferred EC parameters are automatically used instead of
2989 hardcoded fixed parameters. Now a server just has to call:
2990 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2991 support ECDH and use the most appropriate parameters.
2994 *) Enhance and tidy EC curve and point format TLS extension code. Use
2995 static structures instead of allocation if default values are used.
2996 New ctrls to set curves we wish to support and to retrieve shared curves.
2997 Print out shared curves in s_server. New options to s_server and s_client
2998 to set list of supported curves.
3001 *) New ctrls to retrieve supported signature algorithms and
3002 supported curve values as an array of NIDs. Extend openssl utility
3003 to print out received values.
3006 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
3007 between NIDs and the more common NIST names such as "P-256". Enhance
3008 ecparam utility and ECC method to recognise the NIST names for curves.
3011 *) Enhance SSL/TLS certificate chain handling to support different
3012 chains for each certificate instead of one chain in the parent SSL_CTX.
3015 *) Support for fixed DH ciphersuite client authentication: where both
3016 server and client use DH certificates with common parameters.
3019 *) Support for fixed DH ciphersuites: those requiring DH server
3023 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
3025 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
3026 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
3027 X509_CINF_get_signature were reverted post internal team review.
3029 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
3031 *) Build fixes for the Windows and OpenVMS platforms
3032 [Matt Caswell and Richard Levitte]
3034 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
3036 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
3037 message can cause a segmentation fault in OpenSSL due to a NULL pointer
3038 dereference. This could lead to a Denial Of Service attack. Thanks to
3039 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
3043 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
3044 dtls1_buffer_record function under certain conditions. In particular this
3045 could occur if an attacker sent repeated DTLS records with the same
3046 sequence number but for the next epoch. The memory leak could be exploited
3047 by an attacker in a Denial of Service attack through memory exhaustion.
3048 Thanks to Chris Mueller for reporting this issue.
3052 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
3053 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
3054 method would be set to NULL which could later result in a NULL pointer
3055 dereference. Thanks to Frank Schmirler for reporting this issue.
3059 *) Abort handshake if server key exchange message is omitted for ephemeral
3062 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
3063 reporting this issue.
3067 *) Remove non-export ephemeral RSA code on client and server. This code
3068 violated the TLS standard by allowing the use of temporary RSA keys in
3069 non-export ciphersuites and could be used by a server to effectively
3070 downgrade the RSA key length used to a value smaller than the server
3071 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
3072 INRIA or reporting this issue.
3076 *) Fixed issue where DH client certificates are accepted without verification.
3077 An OpenSSL server will accept a DH certificate for client authentication
3078 without the certificate verify message. This effectively allows a client to
3079 authenticate without the use of a private key. This only affects servers
3080 which trust a client certificate authority which issues certificates
3081 containing DH keys: these are extremely rare and hardly ever encountered.
3082 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
3087 *) Ensure that the session ID context of an SSL is updated when its
3088 SSL_CTX is updated via SSL_set_SSL_CTX.
3090 The session ID context is typically set from the parent SSL_CTX,
3091 and can vary with the CTX.
3094 *) Fix various certificate fingerprint issues.
3096 By using non-DER or invalid encodings outside the signed portion of a
3097 certificate the fingerprint can be changed without breaking the signature.
3098 Although no details of the signed portion of the certificate can be changed
3099 this can cause problems with some applications: e.g. those using the
3100 certificate fingerprint for blacklists.
3102 1. Reject signatures with non zero unused bits.
3104 If the BIT STRING containing the signature has non zero unused bits reject
3105 the signature. All current signature algorithms require zero unused bits.
3107 2. Check certificate algorithm consistency.
3109 Check the AlgorithmIdentifier inside TBS matches the one in the
3110 certificate signature. NB: this will result in signature failure
3111 errors for some broken certificates.
3113 Thanks to Konrad Kraszewski from Google for reporting this issue.
3115 3. Check DSA/ECDSA signatures use DER.
3117 Re-encode DSA/ECDSA signatures and compare with the original received
3118 signature. Return an error if there is a mismatch.
3120 This will reject various cases including garbage after signature
3121 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
3122 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
3123 (negative or with leading zeroes).
3125 Further analysis was conducted and fixes were developed by Stephen Henson
3126 of the OpenSSL core team.
3131 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
3132 results on some platforms, including x86_64. This bug occurs at random
3133 with a very low probability, and is not known to be exploitable in any
3134 way, though its exact impact is difficult to determine. Thanks to Pieter
3135 Wuille (Blockstream) who reported this issue and also suggested an initial
3136 fix. Further analysis was conducted by the OpenSSL development team and
3137 Adam Langley of Google. The final fix was developed by Andy Polyakov of
3138 the OpenSSL core team.
3142 *) Do not resume sessions on the server if the negotiated protocol
3143 version does not match the session's version. Resuming with a different
3144 version, while not strictly forbidden by the RFC, is of questionable
3145 sanity and breaks all known clients.
3146 [David Benjamin, Emilia Käsper]
3148 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
3149 early CCS messages during renegotiation. (Note that because
3150 renegotiation is encrypted, this early CCS was not exploitable.)
3153 *) Tighten client-side session ticket handling during renegotiation:
3154 ensure that the client only accepts a session ticket if the server sends
3155 the extension anew in the ServerHello. Previously, a TLS client would
3156 reuse the old extension state and thus accept a session ticket if one was
3157 announced in the initial ServerHello.
3159 Similarly, ensure that the client requires a session ticket if one
3160 was advertised in the ServerHello. Previously, a TLS client would
3161 ignore a missing NewSessionTicket message.
3164 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
3166 *) SRTP Memory Leak.
3168 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
3169 sends a carefully crafted handshake message, to cause OpenSSL to fail
3170 to free up to 64k of memory causing a memory leak. This could be
3171 exploited in a Denial Of Service attack. This issue affects OpenSSL
3172 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
3173 whether SRTP is used or configured. Implementations of OpenSSL that
3174 have been compiled with OPENSSL_NO_SRTP defined are not affected.
3176 The fix was developed by the OpenSSL team.
3180 *) Session Ticket Memory Leak.
3182 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
3183 integrity of that ticket is first verified. In the event of a session
3184 ticket integrity check failing, OpenSSL will fail to free memory
3185 causing a memory leak. By sending a large number of invalid session
3186 tickets an attacker could exploit this issue in a Denial Of Service
3191 *) Build option no-ssl3 is incomplete.
3193 When OpenSSL is configured with "no-ssl3" as a build option, servers
3194 could accept and complete a SSL 3.0 handshake, and clients could be
3195 configured to send them.
3197 [Akamai and the OpenSSL team]
3199 *) Add support for TLS_FALLBACK_SCSV.
3200 Client applications doing fallback retries should call
3201 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3203 [Adam Langley, Bodo Moeller]
3205 *) Add additional DigestInfo checks.
3207 Re-encode DigestInto in DER and check against the original when
3208 verifying RSA signature: this will reject any improperly encoded
3209 DigestInfo structures.
3211 Note: this is a precautionary measure and no attacks are currently known.
3215 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3217 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3218 SRP code can be overrun an internal buffer. Add sanity check that
3219 g, A, B < N to SRP code.
3221 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3222 Group for discovering this issue.
3226 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3227 TLS 1.0 instead of higher protocol versions when the ClientHello message
3228 is badly fragmented. This allows a man-in-the-middle attacker to force a
3229 downgrade to TLS 1.0 even if both the server and the client support a
3230 higher protocol version, by modifying the client's TLS records.
3232 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3233 researching this issue.
3237 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3238 to a denial of service attack. A malicious server can crash the client
3239 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3240 ciphersuite and sending carefully crafted handshake messages.
3242 Thanks to Felix Gröbert (Google) for discovering and researching this
3247 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3248 to leak memory. This can be exploited through a Denial of Service attack.
3249 Thanks to Adam Langley for discovering and researching this issue.
3253 *) An attacker can force openssl to consume large amounts of memory whilst
3254 processing DTLS handshake messages. This can be exploited through a
3255 Denial of Service attack.
3256 Thanks to Adam Langley for discovering and researching this issue.
3260 *) An attacker can force an error condition which causes openssl to crash
3261 whilst processing DTLS packets due to memory being freed twice. This
3262 can be exploited through a Denial of Service attack.
3263 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3268 *) If a multithreaded client connects to a malicious server using a resumed
3269 session and the server sends an ec point format extension it could write
3270 up to 255 bytes to freed memory.
3272 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3277 *) A malicious server can crash an OpenSSL client with a null pointer
3278 dereference (read) by specifying an SRP ciphersuite even though it was not
3279 properly negotiated with the client. This can be exploited through a
3280 Denial of Service attack.
3282 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3283 discovering and researching this issue.
3287 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3288 X509_name_oneline, X509_name_print_ex et al. to leak some information
3289 from the stack. Applications may be affected if they echo pretty printing
3290 output to the attacker.
3292 Thanks to Ivan Fratric (Google) for discovering this issue.
3294 [Emilia Käsper, and Steve Henson]
3296 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3297 for corner cases. (Certain input points at infinity could lead to
3298 bogus results, with non-infinity inputs mapped to infinity too.)
3301 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3303 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3304 handshake can force the use of weak keying material in OpenSSL
3305 SSL/TLS clients and servers.
3307 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3308 researching this issue. (CVE-2014-0224)
3309 [KIKUCHI Masashi, Steve Henson]
3311 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3312 OpenSSL DTLS client the code can be made to recurse eventually crashing
3315 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3317 [Imre Rad, Steve Henson]
3319 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3320 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3321 client or server. This is potentially exploitable to run arbitrary
3322 code on a vulnerable client or server.
3324 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3325 [Jüri Aedla, Steve Henson]
3327 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3328 are subject to a denial of service attack.
3330 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3331 this issue. (CVE-2014-3470)
3332 [Felix Gröbert, Ivan Fratric, Steve Henson]
3334 *) Harmonize version and its documentation. -f flag is used to display
3336 [mancha <mancha1@zoho.com>]
3338 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3339 in i2d_ECPrivateKey.
3340 [mancha <mancha1@zoho.com>]
3342 *) Fix some double frees. These are not thought to be exploitable.
3343 [mancha <mancha1@zoho.com>]
3345 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3347 *) A missing bounds check in the handling of the TLS heartbeat extension
3348 can be used to reveal up to 64k of memory to a connected client or
3351 Thanks for Neel Mehta of Google Security for discovering this bug and to
3352 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3353 preparing the fix (CVE-2014-0160)
3354 [Adam Langley, Bodo Moeller]
3356 *) Fix for the attack described in the paper "Recovering OpenSSL
3357 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3358 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3359 http://eprint.iacr.org/2014/140
3361 Thanks to Yuval Yarom and Naomi Benger for discovering this
3362 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3363 [Yuval Yarom and Naomi Benger]
3365 *) TLS pad extension: draft-agl-tls-padding-03
3367 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3368 TLS client Hello record length value would otherwise be > 255 and
3369 less that 512 pad with a dummy extension containing zeroes so it
3370 is at least 512 bytes long.
3372 [Adam Langley, Steve Henson]
3374 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3376 *) Fix for TLS record tampering bug. A carefully crafted invalid
3377 handshake could crash OpenSSL with a NULL pointer exception.
3378 Thanks to Anton Johansson for reporting this issues.
3381 *) Keep original DTLS digest and encryption contexts in retransmission
3382 structures so we can use the previous session parameters if they need
3383 to be resent. (CVE-2013-6450)
3386 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3387 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3388 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3389 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3390 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3391 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3392 [Rob Stradling, Adam Langley]
3394 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3396 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3397 supporting platforms or when small records were transferred.
3398 [Andy Polyakov, Steve Henson]
3400 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3402 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3404 This addresses the flaw in CBC record processing discovered by
3405 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3406 at: http://www.isg.rhul.ac.uk/tls/
3408 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3409 Security Group at Royal Holloway, University of London
3410 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3411 Emilia Käsper for the initial patch.
3413 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3415 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3416 ciphersuites which can be exploited in a denial of service attack.
3417 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3418 and detecting this bug and to Wolfgang Ettlinger
3419 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3423 *) Return an error when checking OCSP signatures when key is NULL.
3424 This fixes a DoS attack. (CVE-2013-0166)
3427 *) Make openssl verify return errors.
3428 [Chris Palmer <palmer@google.com> and Ben Laurie]
3430 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3431 the right response is stapled. Also change SSL_get_certificate()
3432 so it returns the certificate actually sent.
3433 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3434 [Rob Stradling <rob.stradling@comodo.com>]
3436 *) Fix possible deadlock when decoding public keys.
3439 *) Don't use TLS 1.0 record version number in initial client hello
3443 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3445 *) Sanity check record length before skipping explicit IV in TLS
3446 1.2, 1.1 and DTLS to fix DoS attack.
3448 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3449 fuzzing as a service testing platform.
3453 *) Initialise tkeylen properly when encrypting CMS messages.
3454 Thanks to Solar Designer of Openwall for reporting this issue.
3457 *) In FIPS mode don't try to use composite ciphers as they are not
3461 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3463 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3464 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3465 mean any application compiled against OpenSSL 1.0.0 headers setting
3466 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3467 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3468 0x10000000L Any application which was previously compiled against
3469 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3470 will need to be recompiled as a result. Letting be results in
3471 inability to disable specifically TLS 1.1 and in client context,
3472 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3475 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3476 disable just protocol X, but all protocols above X *if* there are
3477 protocols *below* X still enabled. In more practical terms it means
3478 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3479 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3480 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3484 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3486 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3487 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3488 in CRYPTO_realloc_clean.
3490 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3491 issue and to Adam Langley <agl@chromium.org> for fixing it.
3493 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3495 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3498 *) Workarounds for some broken servers that "hang" if a client hello
3499 record length exceeds 255 bytes.
3501 1. Do not use record version number > TLS 1.0 in initial client
3502 hello: some (but not all) hanging servers will now work.
3503 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3504 the number of ciphers sent in the client hello. This should be
3505 set to an even number, such as 50, for example by passing:
3506 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3507 Most broken servers should now work.
3508 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3509 TLS 1.2 client support entirely.
3512 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3515 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3517 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3518 STRING form instead of a DigestInfo.
3521 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3522 and the RSA_sign/RSA_verify functions. This was made more apparent when
3523 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3524 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3525 the correct format in RSA_verify so both forms transparently work.
3528 *) Some servers which support TLS 1.0 can choke if we initially indicate
3529 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3530 encrypted premaster secret. As a workaround use the maximum permitted
3531 client version in client hello, this should keep such servers happy
3532 and still work with previous versions of OpenSSL.
3535 *) Add support for TLS/DTLS heartbeats.
3536 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3538 *) Add support for SCTP.
3539 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3541 *) Improved PRNG seeding for VOS.
3542 [Paul Green <Paul.Green@stratus.com>]
3544 *) Extensive assembler packs updates, most notably:
3546 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3547 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3548 - x86_64: bit-sliced AES implementation;
3549 - ARM: NEON support, contemporary platforms optimizations;
3550 - s390x: z196 support;
3551 - *: GHASH and GF(2^m) multiplication implementations;
3555 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3556 (removal of unnecessary code)
3557 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3559 *) Add TLS key material exporter from RFC 5705.
3562 *) Add DTLS-SRTP negotiation from RFC 5764.
3565 *) Add Next Protocol Negotiation,
3566 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3567 disabled with a no-npn flag to config or Configure. Code donated
3569 [Adam Langley <agl@google.com> and Ben Laurie]
3571 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3572 NIST-P256, NIST-P521, with constant-time single point multiplication on
3573 typical inputs. Compiler support for the nonstandard type __uint128_t is
3574 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3575 Code made available under Apache License version 2.0.
3577 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3578 line to include this in your build of OpenSSL, and run "make depend" (or
3579 "make update"). This enables the following EC_METHODs:
3581 EC_GFp_nistp224_method()
3582 EC_GFp_nistp256_method()
3583 EC_GFp_nistp521_method()
3585 EC_GROUP_new_by_curve_name() will automatically use these (while
3586 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3588 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3590 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3591 all platforms. Move ssize_t definition from e_os.h to the public
3592 header file e_os2.h as it now appears in public header file cms.h
3595 *) New -sigopt option to the ca, req and x509 utilities. Additional
3596 signature parameters can be passed using this option and in
3600 *) Add RSA PSS signing function. This will generate and set the
3601 appropriate AlgorithmIdentifiers for PSS based on those in the
3602 corresponding EVP_MD_CTX structure. No application support yet.
3605 *) Support for companion algorithm specific ASN1 signing routines.
3606 New function ASN1_item_sign_ctx() signs a pre-initialised
3607 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3608 the appropriate parameters.
3611 *) Add new algorithm specific ASN1 verification initialisation function
3612 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3613 handling will be the same no matter what EVP_PKEY_METHOD is used.
3614 Add a PSS handler to support verification of PSS signatures: checked
3615 against a number of sample certificates.
3618 *) Add signature printing for PSS. Add PSS OIDs.
3619 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3621 *) Add algorithm specific signature printing. An individual ASN1 method
3622 can now print out signatures instead of the standard hex dump.
3624 More complex signatures (e.g. PSS) can print out more meaningful
3625 information. Include DSA version that prints out the signature
3629 *) Password based recipient info support for CMS library: implementing
3633 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3634 neatly separates the code into cipher and PBE sections and is required
3635 for some algorithms that split PBES2 into separate pieces (such as
3636 password based CMS).
3639 *) Session-handling fixes:
3640 - Fix handling of connections that are resuming with a session ID,
3641 but also support Session Tickets.
3642 - Fix a bug that suppressed issuing of a new ticket if the client
3643 presented a ticket with an expired session.
3644 - Try to set the ticket lifetime hint to something reasonable.
3645 - Make tickets shorter by excluding irrelevant information.
3646 - On the client side, don't ignore renewed tickets.
3647 [Adam Langley, Bodo Moeller (Google)]
3649 *) Fix PSK session representation.
3652 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3654 This work was sponsored by Intel.
3657 *) Add GCM support to TLS library. Some custom code is needed to split
3658 the IV between the fixed (from PRF) and explicit (from TLS record)
3659 portions. This adds all GCM ciphersuites supported by RFC5288 and
3660 RFC5289. Generalise some AES* cipherstrings to include GCM and
3661 add a special AESGCM string for GCM only.
3664 *) Expand range of ctrls for AES GCM. Permit setting invocation
3665 field on decrypt and retrieval of invocation field only on encrypt.
3668 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3669 As required by RFC5289 these ciphersuites cannot be used if for
3670 versions of TLS earlier than 1.2.
3673 *) For FIPS capable OpenSSL interpret a NULL default public key method
3674 as unset and return the appropriate default but do *not* set the default.
3675 This means we can return the appropriate method in applications that
3676 switch between FIPS and non-FIPS modes.
3679 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3680 ENGINE is used then we cannot handle that in the FIPS module so we
3681 keep original code iff non-FIPS operations are allowed.
3684 *) Add -attime option to openssl utilities.
3685 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3687 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3690 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3691 FIPS EC methods unconditionally for now.
3694 *) New build option no-ec2m to disable characteristic 2 code.
3697 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3698 all cases can be covered as some introduce binary incompatibilities.
3701 *) Redirect RSA operations to FIPS module including keygen,
3702 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3705 *) Add similar low level API blocking to ciphers.
3708 *) Low level digest APIs are not approved in FIPS mode: any attempt
3709 to use these will cause a fatal error. Applications that *really* want
3710 to use them can use the private_* version instead.
3713 *) Redirect cipher operations to FIPS module for FIPS builds.
3716 *) Redirect digest operations to FIPS module for FIPS builds.
3719 *) Update build system to add "fips" flag which will link in fipscanister.o
3720 for static and shared library builds embedding a signature if needed.
3723 *) Output TLS supported curves in preference order instead of numerical
3724 order. This is currently hardcoded for the highest order curves first.
3725 This should be configurable so applications can judge speed vs strength.
3728 *) Add TLS v1.2 server support for client authentication.
3731 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3735 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3736 FIPS modules versions.
3739 *) Add TLS v1.2 client side support for client authentication. Keep cache
3740 of handshake records longer as we don't know the hash algorithm to use
3741 until after the certificate request message is received.
3744 *) Initial TLS v1.2 client support. Add a default signature algorithms
3745 extension including all the algorithms we support. Parse new signature
3746 format in client key exchange. Relax some ECC signing restrictions for
3747 TLS v1.2 as indicated in RFC5246.
3750 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3751 to new signature format when needed using client digest preference.
3752 All server ciphersuites should now work correctly in TLS v1.2. No client
3753 support yet and no support for client certificates.
3756 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3757 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3758 ciphersuites. At present only RSA key exchange ciphersuites work with
3759 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3760 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3761 and version checking.
3764 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3765 with this defined it will not be affected by any changes to ssl internal
3766 structures. Add several utility functions to allow openssl application
3767 to work with OPENSSL_NO_SSL_INTERN defined.
3770 *) A long standing patch to add support for SRP from EdelWeb (Peter
3771 Sylvester and Christophe Renou) was integrated.
3772 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3773 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3776 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3779 *) Permit abbreviated handshakes when renegotiating using the function
3780 SSL_renegotiate_abbreviated().
3781 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3783 *) Add call to ENGINE_register_all_complete() to
3784 ENGINE_load_builtin_engines(), so some implementations get used
3785 automatically instead of needing explicit application support.
3788 *) Add support for TLS key exporter as described in RFC5705.
3789 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3791 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3792 a few changes are required:
3794 Add SSL_OP_NO_TLSv1_1 flag.
3795 Add TLSv1_1 methods.
3796 Update version checking logic to handle version 1.1.
3797 Add explicit IV handling (ported from DTLS code).
3798 Add command line options to s_client/s_server.
3801 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3803 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3804 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3805 content decryption and always return the same error. Note: this attack
3806 needs on average 2^20 messages so it only affects automated senders. The
3807 old behaviour can be re-enabled in the CMS code by setting the
3808 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3809 an MMA defence is not necessary.
3810 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3811 this issue. (CVE-2012-0884)
3814 *) Fix CVE-2011-4619: make sure we really are receiving a
3815 client hello before rejecting multiple SGC restarts. Thanks to
3816 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3819 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3821 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3822 Thanks to Antonio Martin, Enterprise Secure Access Research and
3823 Development, Cisco Systems, Inc. for discovering this bug and
3824 preparing a fix. (CVE-2012-0050)
3827 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3829 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3830 of the Vaudenay padding oracle attack on CBC mode encryption
3831 which enables an efficient plaintext recovery attack against
3832 the OpenSSL implementation of DTLS. Their attack exploits timing
3833 differences arising during decryption processing. A research
3834 paper describing this attack can be found at:
3835 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3836 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3837 Security Group at Royal Holloway, University of London
3838 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3839 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3840 for preparing the fix. (CVE-2011-4108)
3841 [Robin Seggelmann, Michael Tuexen]
3843 *) Clear bytes used for block padding of SSL 3.0 records.
3845 [Adam Langley (Google)]
3847 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3848 Kadianakis <desnacked@gmail.com> for discovering this issue and
3849 Adam Langley for preparing the fix. (CVE-2011-4619)
3850 [Adam Langley (Google)]
3852 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3853 [Andrey Kulikov <amdeich@gmail.com>]
3855 *) Prevent malformed RFC3779 data triggering an assertion failure.
3856 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3857 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3858 [Rob Austein <sra@hactrn.net>]
3860 *) Improved PRNG seeding for VOS.
3861 [Paul Green <Paul.Green@stratus.com>]
3863 *) Fix ssl_ciph.c set-up race.
3864 [Adam Langley (Google)]
3866 *) Fix spurious failures in ecdsatest.c.
3867 [Emilia Käsper (Google)]
3869 *) Fix the BIO_f_buffer() implementation (which was mixing different
3870 interpretations of the '..._len' fields).
3871 [Adam Langley (Google)]
3873 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3874 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3875 threads won't reuse the same blinding coefficients.
3877 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3878 lock to call BN_BLINDING_invert_ex, and avoids one use of
3879 BN_BLINDING_update for each BN_BLINDING structure (previously,
3880 the last update always remained unused).
3881 [Emilia Käsper (Google)]
3883 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3884 [Bob Buckholz (Google)]
3886 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3888 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3889 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3890 [Kaspar Brand <ossl@velox.ch>]
3892 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3893 for multi-threaded use of ECDH. (CVE-2011-3210)
3894 [Adam Langley (Google)]
3896 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3899 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3900 signature public key algorithm by using OID xref utilities instead.
3901 Before this you could only use some ECC ciphersuites with SHA1 only.
3904 *) Add protection against ECDSA timing attacks as mentioned in the paper
3905 by Billy Bob Brumley and Nicola Tuveri, see:
3907 http://eprint.iacr.org/2011/232.pdf
3909 [Billy Bob Brumley and Nicola Tuveri]
3911 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3913 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3914 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3916 *) Fix bug in string printing code: if *any* escaping is enabled we must
3917 escape the escape character (backslash) or the resulting string is
3921 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3923 *) Disable code workaround for ancient and obsolete Netscape browsers
3924 and servers: an attacker can use it in a ciphersuite downgrade attack.
3925 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3928 *) Fixed J-PAKE implementation error, originally discovered by
3929 Sebastien Martini, further info and confirmation from Stefan
3930 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3933 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3935 *) Fix extension code to avoid race conditions which can result in a buffer
3936 overrun vulnerability: resumed sessions must not be modified as they can
3937 be shared by multiple threads. CVE-2010-3864
3940 *) Fix WIN32 build system to correctly link an ENGINE directory into
3944 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3946 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3948 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3950 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3952 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3953 context. The operation can be customised via the ctrl mechanism in
3954 case ENGINEs want to include additional functionality.
3957 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3960 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3961 output hashes compatible with older versions of OpenSSL.
3962 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3964 *) Fix compression algorithm handling: if resuming a session use the
3965 compression algorithm of the resumed session instead of determining
3966 it from client hello again. Don't allow server to change algorithm.
3969 *) Add load_crls() function to apps tidying load_certs() too. Add option
3970 to verify utility to allow additional CRLs to be included.
3973 *) Update OCSP request code to permit adding custom headers to the request:
3974 some responders need this.
3977 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3979 [Julia Lawall <julia@diku.dk>]
3981 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3982 needlessly dereferenced structures, used obsolete functions and
3983 didn't handle all updated verify codes correctly.
3986 *) Disable MD2 in the default configuration.
3989 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3990 indicate the initial BIO being pushed or popped. This makes it possible
3991 to determine whether the BIO is the one explicitly called or as a result
3992 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
3993 it handles reference counts correctly and doesn't zero out the I/O bio
3994 when it is not being explicitly popped. WARNING: applications which
3995 included workarounds for the old buggy behaviour will need to be modified
3996 or they could free up already freed BIOs.
3999 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
4000 renaming to all platforms (within the 0.9.8 branch, this was
4001 done conditionally on Netware platforms to avoid a name clash).
4002 [Guenter <lists@gknw.net>]
4004 *) Add ECDHE and PSK support to DTLS.
4005 [Michael Tuexen <tuexen@fh-muenster.de>]
4007 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
4011 *) Add "missing" function EVP_MD_flags() (without this the only way to
4012 retrieve a digest flags is by accessing the structure directly. Update
4013 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
4014 or cipher is registered as in the "from" argument. Print out all
4015 registered digests in the dgst usage message instead of manually
4016 attempting to work them out.
4019 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
4020 this allows the use of compression and extensions. Change default cipher
4021 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
4022 by default unless an application cipher string requests it.
4025 *) Alter match criteria in PKCS12_parse(). It used to try to use local
4026 key ids to find matching certificates and keys but some PKCS#12 files
4027 don't follow the (somewhat unwritten) rules and this strategy fails.
4028 Now just gather all certificates together and the first private key
4029 then look for the first certificate that matches the key.
4032 *) Support use of registered digest and cipher names for dgst and cipher
4033 commands instead of having to add each one as a special case. So now
4040 openssl dgst -sha256 foo
4042 and this works for ENGINE based algorithms too.
4046 *) Update Gost ENGINE to support parameter files.
4047 [Victor B. Wagner <vitus@cryptocom.ru>]
4049 *) Support GeneralizedTime in ca utility.
4050 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
4052 *) Enhance the hash format used for certificate directory links. The new
4053 form uses the canonical encoding (meaning equivalent names will work
4054 even if they aren't identical) and uses SHA1 instead of MD5. This form
4055 is incompatible with the older format and as a result c_rehash should
4056 be used to rebuild symbolic links.
4059 *) Make PKCS#8 the default write format for private keys, replacing the
4060 traditional format. This form is standardised, more secure and doesn't
4061 include an implicit MD5 dependency.
4064 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
4065 committed to OpenSSL should pass this lot as a minimum.
4068 *) Add session ticket override functionality for use by EAP-FAST.
4069 [Jouni Malinen <j@w1.fi>]
4071 *) Modify HMAC functions to return a value. Since these can be implemented
4072 in an ENGINE errors can occur.
4075 *) Type-checked OBJ_bsearch_ex.
4078 *) Type-checked OBJ_bsearch. Also some constification necessitated
4079 by type-checking. Still to come: TXT_DB, bsearch(?),
4080 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
4084 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
4085 seconds to a tm structure directly, instead of going through OS
4086 specific date routines. This avoids any issues with OS routines such
4087 as the year 2038 bug. New *_adj() functions for ASN1 time structures
4088 and X509_time_adj_ex() to cover the extended range. The existing
4089 X509_time_adj() is still usable and will no longer have any date issues.
4092 *) Delta CRL support. New use deltas option which will attempt to locate
4093 and search any appropriate delta CRLs available.
4095 This work was sponsored by Google.
4098 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
4099 code and add additional score elements. Validate alternate CRL paths
4100 as part of the CRL checking and indicate a new error "CRL path validation
4101 error" in this case. Applications wanting additional details can use
4102 the verify callback and check the new "parent" field. If this is not
4103 NULL CRL path validation is taking place. Existing applications won't
4104 see this because it requires extended CRL support which is off by
4107 This work was sponsored by Google.
4110 *) Support for freshest CRL extension.
4112 This work was sponsored by Google.
4115 *) Initial indirect CRL support. Currently only supported in the CRLs
4116 passed directly and not via lookup. Process certificate issuer
4117 CRL entry extension and lookup CRL entries by bother issuer name
4118 and serial number. Check and process CRL issuer entry in IDP extension.
4120 This work was sponsored by Google.
4123 *) Add support for distinct certificate and CRL paths. The CRL issuer
4124 certificate is validated separately in this case. Only enabled if
4125 an extended CRL support flag is set: this flag will enable additional
4126 CRL functionality in future.
4128 This work was sponsored by Google.
4131 *) Add support for policy mappings extension.
4133 This work was sponsored by Google.
4136 *) Fixes to pathlength constraint, self issued certificate handling,
4137 policy processing to align with RFC3280 and PKITS tests.
4139 This work was sponsored by Google.
4142 *) Support for name constraints certificate extension. DN, email, DNS
4143 and URI types are currently supported.
4145 This work was sponsored by Google.
4148 *) To cater for systems that provide a pointer-based thread ID rather
4149 than numeric, deprecate the current numeric thread ID mechanism and
4150 replace it with a structure and associated callback type. This
4151 mechanism allows a numeric "hash" to be extracted from a thread ID in
4152 either case, and on platforms where pointers are larger than 'long',
4153 mixing is done to help ensure the numeric 'hash' is usable even if it
4154 can't be guaranteed unique. The default mechanism is to use "&errno"
4155 as a pointer-based thread ID to distinguish between threads.
4157 Applications that want to provide their own thread IDs should now use
4158 CRYPTO_THREADID_set_callback() to register a callback that will call
4159 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
4161 Note that ERR_remove_state() is now deprecated, because it is tied
4162 to the assumption that thread IDs are numeric. ERR_remove_state(0)
4163 to free the current thread's error state should be replaced by
4164 ERR_remove_thread_state(NULL).
4166 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
4167 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
4168 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
4169 application was previously providing a numeric thread callback that
4170 was inappropriate for distinguishing threads, then uniqueness might
4171 have been obtained with &errno that happened immediately in the
4172 intermediate development versions of OpenSSL; this is no longer the
4173 case, the numeric thread callback will now override the automatic use
4175 [Geoff Thorpe, with help from Bodo Moeller]
4177 *) Initial support for different CRL issuing certificates. This covers a
4178 simple case where the self issued certificates in the chain exist and
4179 the real CRL issuer is higher in the existing chain.
4181 This work was sponsored by Google.
4184 *) Removed effectively defunct crypto/store from the build.
4187 *) Revamp of STACK to provide stronger type-checking. Still to come:
4188 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
4189 ASN1_STRING, CONF_VALUE.
4192 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4193 RAM on SSL connections. This option can save about 34k per idle SSL.
4196 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4197 STACK, TXT_DB, bsearch, qsort.
4200 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4201 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4202 support for data, signedData, compressedData, digestedData and
4203 encryptedData, envelopedData types included. Scripts to check against
4204 RFC4134 examples draft and interop and consistency checks of many
4205 content types and variants.
4208 *) Add options to enc utility to support use of zlib compression BIO.
4211 *) Extend mk1mf to support importing of options and assembly language
4212 files from Configure script, currently only included in VC-WIN32.
4213 The assembly language rules can now optionally generate the source
4214 files from the associated perl scripts.
4217 *) Implement remaining functionality needed to support GOST ciphersuites.
4218 Interop testing has been performed using CryptoPro implementations.
4219 [Victor B. Wagner <vitus@cryptocom.ru>]
4221 *) s390x assembler pack.
4224 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4228 *) Implement Opaque PRF Input TLS extension as specified in
4229 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4230 official specification yet and no extension type assignment by
4231 IANA exists, this extension (for now) will have to be explicitly
4232 enabled when building OpenSSL by providing the extension number
4233 to use. For example, specify an option
4235 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4237 to the "config" or "Configure" script to enable the extension,
4238 assuming extension number 0x9527 (which is a completely arbitrary
4239 and unofficial assignment based on the MD5 hash of the Internet
4240 Draft). Note that by doing so, you potentially lose
4241 interoperability with other TLS implementations since these might
4242 be using the same extension number for other purposes.
4244 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4245 opaque PRF input value to use in the handshake. This will create
4246 an internal copy of the length-'len' string at 'src', and will
4247 return non-zero for success.
4249 To get more control and flexibility, provide a callback function
4252 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4253 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4257 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4260 Callback function 'cb' will be called in handshakes, and is
4261 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4262 Argument 'arg' is for application purposes (the value as given to
4263 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4264 be provided to the callback function). The callback function
4265 has to return non-zero to report success: usually 1 to use opaque
4266 PRF input just if possible, or 2 to enforce use of the opaque PRF
4267 input. In the latter case, the library will abort the handshake
4268 if opaque PRF input is not successfully negotiated.
4270 Arguments 'peerinput' and 'len' given to the callback function
4271 will always be NULL and 0 in the case of a client. A server will
4272 see the client's opaque PRF input through these variables if
4273 available (NULL and 0 otherwise). Note that if the server
4274 provides an opaque PRF input, the length must be the same as the
4275 length of the client's opaque PRF input.
4277 Note that the callback function will only be called when creating
4278 a new session (session resumption can resume whatever was
4279 previously negotiated), and will not be called in SSL 2.0
4280 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4281 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4282 for applications that need to enforce opaque PRF input.
4286 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4289 [Victor B. Wagner <vitus@cryptocom.ru>]
4291 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4292 RFC4507bis. The encrypted ticket format is an encrypted encoded
4293 SSL_SESSION structure, that way new session features are automatically
4296 If a client application caches session in an SSL_SESSION structure
4297 support is transparent because tickets are now stored in the encoded
4300 The SSL_CTX structure automatically generates keys for ticket
4301 protection in servers so again support should be possible
4302 with no application modification.
4304 If a client or server wishes to disable RFC4507 support then the option
4305 SSL_OP_NO_TICKET can be set.
4307 Add a TLS extension debugging callback to allow the contents of any client
4308 or server extensions to be examined.
4310 This work was sponsored by Google.
4313 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4314 OpenSSL should now compile cleanly on gcc 4.2
4315 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4317 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4318 support including streaming MAC support: this is required for GOST
4319 ciphersuite support.
4320 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4322 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4323 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4324 to output in BER and PEM format.
4327 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4328 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4329 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4330 ENGINE support for HMAC keys which are unextractable. New -mac and
4331 -macopt options to dgst utility.
4334 *) New option -sigopt to dgst utility. Update dgst to use
4335 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4336 alternative signing parameters such as X9.31 or PSS in the dgst
4340 *) Change ssl_cipher_apply_rule(), the internal function that does
4341 the work each time a ciphersuite string requests enabling
4342 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4343 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4344 the order of disabled ciphersuites such that those ciphersuites
4345 that most recently went from enabled to disabled not only stay
4346 in order with respect to each other, but also have higher priority
4347 than other disabled ciphersuites the next time ciphersuites are
4350 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4351 the same ciphersuites as with "HIGH" alone, but in a specific
4352 order where the PSK ciphersuites come first (since they are the
4353 most recently disabled ciphersuites when "HIGH" is parsed).
4355 Also, change ssl_create_cipher_list() (using this new
4356 functionality) such that between otherwise identical
4357 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4361 *) Change ssl_create_cipher_list() so that it automatically
4362 arranges the ciphersuites in reasonable order before starting
4363 to process the rule string. Thus, the definition for "DEFAULT"
4364 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4365 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4366 This makes it much easier to arrive at a reasonable default order
4367 in applications for which anonymous ciphers are OK (meaning
4368 that you can't actually use DEFAULT).
4369 [Bodo Moeller; suggested by Victor Duchovni]
4371 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4372 processing) into multiple integers instead of setting
4373 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
4374 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
4375 (These masks as well as the individual bit definitions are hidden
4376 away into the non-exported interface ssl/ssl_locl.h, so this
4377 change to the definition of the SSL_CIPHER structure shouldn't
4378 affect applications.) This give us more bits for each of these
4379 categories, so there is no longer a need to coagulate AES128 and
4380 AES256 into a single algorithm bit, and to coagulate Camellia128
4381 and Camellia256 into a single algorithm bit, which has led to all
4384 Thus, among other things, the kludge introduced in 0.9.7m and
4385 0.9.8e for masking out AES256 independently of AES128 or masking
4386 out Camellia256 independently of AES256 is not needed here in 0.9.9.
4388 With the change, we also introduce new ciphersuite aliases that
4389 so far were missing: "AES128", "AES256", "CAMELLIA128", and
4393 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
4394 Use the leftmost N bytes of the signature input if the input is
4395 larger than the prime q (with N being the size in bytes of q).
4398 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
4399 it yet and it is largely untested.
4402 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
4405 *) Initial incomplete changes to avoid need for function casts in OpenSSL
4406 some compilers (gcc 4.2 and later) reject their use. Safestack is
4407 reimplemented. Update ASN1 to avoid use of legacy functions.
4410 *) Win32/64 targets are linked with Winsock2.
4413 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
4414 to external functions. This can be used to increase CRL handling
4415 efficiency especially when CRLs are very large by (for example) storing
4416 the CRL revoked certificates in a database.
4419 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
4420 new CRLs added to a directory can be used. New command line option
4421 -verify_return_error to s_client and s_server. This causes real errors
4422 to be returned by the verify callback instead of carrying on no matter
4423 what. This reflects the way a "real world" verify callback would behave.
4426 *) GOST engine, supporting several GOST algorithms and public key formats.
4427 Kindly donated by Cryptocom.
4430 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4431 partitioned by DP are handled but no indirect CRL or reason partitioning
4432 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4433 selected via a scoring technique which handles IDP and AKID in CRLs.
4436 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4437 will ultimately be used for all verify operations: this will remove the
4438 X509_STORE dependency on certificate verification and allow alternative
4439 lookup methods. X509_STORE based implementations of these two callbacks.
4442 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4443 Modify get_crl() to find a valid (unexpired) CRL if possible.
4446 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4447 this would be called X509_CRL_cmp() but that name is already used by
4448 a function that just compares CRL issuer names. Cache several CRL
4449 extensions in X509_CRL structure and cache CRLDP in X509.
4452 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4453 this maps equivalent X509_NAME structures into a consistent structure.
4454 Name comparison can then be performed rapidly using memcmp().
4457 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4461 *) Allow digests to supply their own micalg string for S/MIME type using
4462 the ctrl EVP_MD_CTRL_MICALG.
4465 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4466 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4467 ctrl. It can then customise the structure before and/or after signing
4471 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4472 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4473 to free up any added signature OIDs.
4476 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4477 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4478 digest and cipher tables. New options added to openssl utility:
4479 list-message-digest-algorithms and list-cipher-algorithms.
4482 *) Change the array representation of binary polynomials: the list
4483 of degrees of non-zero coefficients is now terminated with -1.
4484 Previously it was terminated with 0, which was also part of the
4485 value; thus, the array representation was not applicable to
4486 polynomials where t^0 has coefficient zero. This change makes
4487 the array representation useful in a more general context.
4490 *) Various modifications and fixes to SSL/TLS cipher string
4491 handling. For ECC, the code now distinguishes between fixed ECDH
4492 with RSA certificates on the one hand and with ECDSA certificates
4493 on the other hand, since these are separate ciphersuites. The
4494 unused code for Fortezza ciphersuites has been removed.
4496 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4497 (not "ECDHE"). For consistency with the code for DH
4498 certificates, use of ECDH certificates is now considered ECDH
4499 authentication, not RSA or ECDSA authentication (the latter is
4500 merely the CA's signing algorithm and not actively used in the
4503 The temporary ciphersuite alias "ECCdraft" is no longer
4504 available, and ECC ciphersuites are no longer excluded from "ALL"
4505 and "DEFAULT". The following aliases now exist for RFC 4492
4506 ciphersuites, most of these by analogy with the DH case:
4508 kECDHr - ECDH cert, signed with RSA
4509 kECDHe - ECDH cert, signed with ECDSA
4510 kECDH - ECDH cert (signed with either RSA or ECDSA)
4511 kEECDH - ephemeral ECDH
4512 ECDH - ECDH cert or ephemeral ECDH
4518 AECDH - anonymous ECDH
4519 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4523 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
4524 Use correct micalg parameters depending on digest(s) in signed message.
4527 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
4528 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
4531 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
4532 an engine to register a method. Add ENGINE lookups for methods and
4533 functional reference processing.
4536 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
4537 EVP_{Sign,Verify}* which allow an application to customise the signature
4541 *) New -resign option to smime utility. This adds one or more signers
4542 to an existing PKCS#7 signedData structure. Also -md option to use an
4543 alternative message digest algorithm for signing.
4546 *) Tidy up PKCS#7 routines and add new functions to make it easier to
4547 create PKCS7 structures containing multiple signers. Update smime
4548 application to support multiple signers.
4551 *) New -macalg option to pkcs12 utility to allow setting of an alternative
4555 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
4556 Reorganize PBE internals to lookup from a static table using NIDs,
4557 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
4558 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
4559 PRF which will be automatically used with PBES2.
4562 *) Replace the algorithm specific calls to generate keys in "req" with the
4566 *) Update PKCS#7 enveloped data routines to use new API. This is now
4567 supported by any public key method supporting the encrypt operation. A
4568 ctrl is added to allow the public key algorithm to examine or modify
4569 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
4573 *) Add a ctrl to asn1 method to allow a public key algorithm to express
4574 a default digest type to use. In most cases this will be SHA1 but some
4575 algorithms (such as GOST) need to specify an alternative digest. The
4576 return value indicates how strong the preference is 1 means optional and
4577 2 is mandatory (that is it is the only supported type). Modify
4578 ASN1_item_sign() to accept a NULL digest argument to indicate it should
4579 use the default md. Update openssl utilities to use the default digest
4580 type for signing if it is not explicitly indicated.
4583 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
4584 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
4585 signing method from the key type. This effectively removes the link
4586 between digests and public key types.
4589 *) Add an OID cross reference table and utility functions. Its purpose is to
4590 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
4591 rsaEncryption. This will allow some of the algorithm specific hackery
4592 needed to use the correct OID to be removed.
4595 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
4596 structures for PKCS7_sign(). They are now set up by the relevant public
4600 *) Add provisional EC pkey method with support for ECDSA and ECDH.
4603 *) Add support for key derivation (agreement) in the API, DH method and
4607 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
4608 public and private key formats. As a side effect these add additional
4609 command line functionality not previously available: DSA signatures can be
4610 generated and verified using pkeyutl and DH key support and generation in
4615 [Oliver Tappe <zooey@hirschkaefer.de>]
4617 *) New make target "install_html_docs" installs HTML renditions of the
4619 [Oliver Tappe <zooey@hirschkaefer.de>]
4621 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4622 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4623 support key and parameter generation and add initial key generation
4624 functionality for RSA.
4627 *) Add functions for main EVP_PKEY_method operations. The undocumented
4628 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4629 EVP_PKEY_{encrypt,decrypt}_old.
4632 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4633 key API, doesn't do much yet.
4636 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4637 public key algorithms. New option to openssl utility:
4638 "list-public-key-algorithms" to print out info.
4641 *) Implement the Supported Elliptic Curves Extension for
4642 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4645 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4646 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4649 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4650 utilities such as rsa, dsa, dsaparam etc except they process any key
4654 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4655 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4656 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4660 *) Initial support for pluggable public key ASN1.
4661 De-spaghettify the public key ASN1 handling. Move public and private
4662 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4663 algorithm specific handling to a single module within the relevant
4664 algorithm directory. Add functions to allow (near) opaque processing
4665 of public and private key structures.
4668 *) Implement the Supported Point Formats Extension for
4669 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4672 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4673 for the psk identity [hint] and the psk callback functions to the
4674 SSL_SESSION, SSL and SSL_CTX structure.
4677 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4681 SSL_CTX_use_psk_identity_hint
4682 SSL_get_psk_identity_hint
4683 SSL_get_psk_identity
4684 SSL_use_psk_identity_hint
4686 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4688 *) Add RFC 3161 compliant time stamp request creation, response generation
4689 and response verification functionality.
4690 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4692 *) Add initial support for TLS extensions, specifically for the server_name
4693 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4694 have new members for a host name. The SSL data structure has an
4695 additional member SSL_CTX *initial_ctx so that new sessions can be
4696 stored in that context to allow for session resumption, even after the
4697 SSL has been switched to a new SSL_CTX in reaction to a client's
4698 server_name extension.
4700 New functions (subject to change):
4702 SSL_get_servername()
4703 SSL_get_servername_type()
4706 New CTRL codes and macros (subject to change):
4708 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4709 - SSL_CTX_set_tlsext_servername_callback()
4710 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4711 - SSL_CTX_set_tlsext_servername_arg()
4712 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4714 openssl s_client has a new '-servername ...' option.
4716 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4717 '-key2 ...', '-servername_fatal' (subject to change). This allows
4718 testing the HostName extension for a specific single host name ('-cert'
4719 and '-key' remain fallbacks for handshakes without HostName
4720 negotiation). If the unrecognized_name alert has to be sent, this by
4721 default is a warning; it becomes fatal with the '-servername_fatal'
4724 [Peter Sylvester, Remy Allais, Christophe Renou]
4726 *) Whirlpool hash implementation is added.
4729 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4730 bn(64,32). Because of instruction set limitations it doesn't have
4731 any negative impact on performance. This was done mostly in order
4732 to make it possible to share assembler modules, such as bn_mul_mont
4733 implementations, between 32- and 64-bit builds without hassle.
4736 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4737 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4741 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4742 dedicated Montgomery multiplication procedure, is introduced.
4743 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4744 "64-bit" performance on certain 32-bit targets.
4747 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4748 in SSL structures. New SSL ctrl to set maximum send fragment size.
4749 Save memory by setting the I/O buffer sizes dynamically instead of
4750 using the maximum available value.
4753 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
4754 in addition to the text details.
4757 *) Very, very preliminary EXPERIMENTAL support for printing of general
4758 ASN1 structures. This currently produces rather ugly output and doesn't
4759 handle several customised structures at all.
4762 *) Integrated support for PVK file format and some related formats such
4763 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4764 these in the 'rsa' and 'dsa' utilities.
4767 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4770 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4771 place for the (very old) "NETSCAPE" format certificates which are now
4772 handled using new ASN1 code equivalents.
4775 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4776 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4777 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4780 *) Modify CRL distribution points extension code to print out previously
4781 unsupported fields. Enhance extension setting code to allow setting of
4785 *) Add print and set support for Issuing Distribution Point CRL extension.
4788 *) Change 'Configure' script to enable Camellia by default.
4791 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4793 *) When rejecting SSL/TLS records due to an incorrect version number, never
4794 update s->server with a new major version number. As of
4795 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
4796 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
4797 the previous behavior could result in a read attempt at NULL when
4798 receiving specific incorrect SSL/TLS records once record payload
4799 protection is active. (CVE-2010-0740)
4800 [Bodo Moeller, Adam Langley <agl@chromium.org>]
4802 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
4803 could be crashed if the relevant tables were not present (e.g. chrooted).
4804 [Tomas Hoger <thoger@redhat.com>]
4806 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
4808 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
4809 [Martin Olsson, Neel Mehta]
4811 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
4812 accommodate for stack sorting, always a write lock!).
4815 *) On some versions of WIN32 Heap32Next is very slow. This can cause
4816 excessive delays in the RAND_poll(): over a minute. As a workaround
4817 include a time check in the inner Heap32Next loop too.
4820 *) The code that handled flushing of data in SSL/TLS originally used the
4821 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
4822 the problem outlined in PR#1949. The fix suggested there however can
4823 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4824 of Apache). So instead simplify the code to flush unconditionally.
4825 This should be fine since flushing with no data to flush is a no op.
4828 *) Handle TLS versions 2.0 and later properly and correctly use the
4829 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4830 off ancient servers have a habit of sticking around for a while...
4833 *) Modify compression code so it frees up structures without using the
4834 ex_data callbacks. This works around a problem where some applications
4835 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4836 restarting) then use compression (e.g. SSL with compression) later.
4837 This results in significant per-connection memory leaks and
4838 has caused some security issues including CVE-2008-1678 and
4842 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4843 change when encrypting or decrypting.
4846 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4847 connect and renegotiate with servers which do not support RI.
4848 Until RI is more widely deployed this option is enabled by default.
4851 *) Add "missing" ssl ctrls to clear options and mode.
4854 *) If client attempts to renegotiate and doesn't support RI respond with
4855 a no_renegotiation alert as required by RFC5746. Some renegotiating
4856 TLS clients will continue a connection gracefully when they receive
4857 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4858 waiting for a server hello which it will never receive. Now we treat a
4859 received no_renegotiation alert as a fatal error. This is because
4860 applications requesting a renegotiation might well expect it to succeed
4861 and would have no code in place to handle the server denying it so the
4862 only safe thing to do is to terminate the connection.
4865 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4866 peer supports secure renegotiation and 0 otherwise. Print out peer
4867 renegotiation support in s_client/s_server.
4870 *) Replace the highly broken and deprecated SPKAC certification method with
4871 the updated NID creation version. This should correctly handle UTF8.
4874 *) Implement RFC5746. Re-enable renegotiation but require the extension
4875 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
4876 turns out to be a bad idea. It has been replaced by
4877 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
4878 SSL_CTX_set_options(). This is really not recommended unless you
4879 know what you are doing.
4880 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
4882 *) Fixes to stateless session resumption handling. Use initial_ctx when
4883 issuing and attempting to decrypt tickets in case it has changed during
4884 servername handling. Use a non-zero length session ID when attempting
4885 stateless session resumption: this makes it possible to determine if
4886 a resumption has occurred immediately after receiving server hello
4887 (several places in OpenSSL subtly assume this) instead of later in
4891 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
4892 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
4893 fixes for a few places where the return code is not checked
4895 [Julia Lawall <julia@diku.dk>]
4897 *) Add --strict-warnings option to Configure script to include devteam
4898 warnings in other configurations.
4901 *) Add support for --libdir option and LIBDIR variable in makefiles. This
4902 makes it possible to install openssl libraries in locations which
4903 have names other than "lib", for example "/usr/lib64" which some
4905 [Steve Henson, based on patch from Jeremy Utley]
4907 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
4908 X690 8.9.12 and can produce some misleading textual output of OIDs.
4909 [Steve Henson, reported by Dan Kaminsky]
4911 *) Delete MD2 from algorithm tables. This follows the recommendation in
4912 several standards that it is not used in new applications due to
4913 several cryptographic weaknesses. For binary compatibility reasons
4914 the MD2 API is still compiled in by default.
4917 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
4921 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
4922 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
4924 [Guenter <lists@gknw.net>]
4926 *) Fix the server certificate chain building code to use X509_verify_cert(),
4927 it used to have an ad-hoc builder which was unable to cope with anything
4928 other than a simple chain.
4929 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
4931 *) Don't check self signed certificate signatures in X509_verify_cert()
4932 by default (a flag can override this): it just wastes time without
4933 adding any security. As a useful side effect self signed root CAs
4934 with non-FIPS digests are now usable in FIPS mode.
4937 *) In dtls1_process_out_of_seq_message() the check if the current message
4938 is already buffered was missing. For every new message was memory
4939 allocated, allowing an attacker to perform an denial of service attack
4940 with sending out of seq handshake messages until there is no memory
4941 left. Additionally every future message was buffered, even if the
4942 sequence number made no sense and would be part of another handshake.
4943 So only messages with sequence numbers less than 10 in advance will be
4944 buffered. (CVE-2009-1378)
4945 [Robin Seggelmann, discovered by Daniel Mentz]
4947 *) Records are buffered if they arrive with a future epoch to be
4948 processed after finishing the corresponding handshake. There is
4949 currently no limitation to this buffer allowing an attacker to perform
4950 a DOS attack with sending records with future epochs until there is no
4951 memory left. This patch adds the pqueue_size() function to determine
4952 the size of a buffer and limits the record buffer to 100 entries.
4954 [Robin Seggelmann, discovered by Daniel Mentz]
4956 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
4957 parent structure is freed. (CVE-2009-1379)
4960 *) Handle non-blocking I/O properly in SSL_shutdown() call.
4961 [Darryl Miles <darryl-mailinglists@netbauds.net>]
4964 [Ilya O. <vrghost@gmail.com>]
4966 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
4968 *) Disable renegotiation completely - this fixes a severe security
4969 problem (CVE-2009-3555) at the cost of breaking all
4970 renegotiation. Renegotiation can be re-enabled by setting
4971 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
4972 run-time. This is really not recommended unless you know what
4976 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
4978 *) Don't set val to NULL when freeing up structures, it is freed up by
4979 underlying code. If sizeof(void *) > sizeof(long) this can result in
4980 zeroing past the valid field. (CVE-2009-0789)
4981 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
4983 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
4984 checked correctly. This would allow some invalid signed attributes to
4985 appear to verify correctly. (CVE-2009-0591)
4986 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4988 *) Reject UniversalString and BMPString types with invalid lengths. This
4989 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
4990 a legal length. (CVE-2009-0590)
4993 *) Set S/MIME signing as the default purpose rather than setting it
4994 unconditionally. This allows applications to override it at the store
4998 *) Permit restricted recursion of ASN1 strings. This is needed in practice
4999 to handle some structures.
5002 *) Improve efficiency of mem_gets: don't search whole buffer each time
5004 [Jeremy Shapiro <jnshapir@us.ibm.com>]
5006 *) New -hex option for openssl rand.
5009 *) Print out UTF8String and NumericString when parsing ASN1.
5012 *) Support NumericString type for name components.
5015 *) Allow CC in the environment to override the automatically chosen
5016 compiler. Note that nothing is done to ensure flags work with the
5020 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
5022 *) Properly check EVP_VerifyFinal() and similar return values
5024 [Ben Laurie, Bodo Moeller, Google Security Team]
5026 *) Enable TLS extensions by default.
5029 *) Allow the CHIL engine to be loaded, whether the application is
5030 multithreaded or not. (This does not release the developer from the
5031 obligation to set up the dynamic locking callbacks.)
5032 [Sander Temme <sander@temme.net>]
5034 *) Use correct exit code if there is an error in dgst command.
5035 [Steve Henson; problem pointed out by Roland Dirlewanger]
5037 *) Tweak Configure so that you need to say "experimental-jpake" to enable
5038 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
5041 *) Add experimental JPAKE support, including demo authentication in
5042 s_client and s_server.
5045 *) Set the comparison function in v3_addr_canonize().
5046 [Rob Austein <sra@hactrn.net>]
5048 *) Add support for XMPP STARTTLS in s_client.
5049 [Philip Paeps <philip@freebsd.org>]
5051 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
5052 to ensure that even with this option, only ciphersuites in the
5053 server's preference list will be accepted. (Note that the option
5054 applies only when resuming a session, so the earlier behavior was
5055 just about the algorithm choice for symmetric cryptography.)
5058 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
5060 *) Fix NULL pointer dereference if a DTLS server received
5061 ChangeCipherSpec as first record (CVE-2009-1386).
5064 *) Fix a state transition in s3_srvr.c and d1_srvr.c
5065 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
5068 *) The fix in 0.9.8c that supposedly got rid of unsafe
5069 double-checked locking was incomplete for RSA blinding,
5070 addressing just one layer of what turns out to have been
5071 doubly unsafe triple-checked locking.
5073 So now fix this for real by retiring the MONT_HELPER macro
5074 in crypto/rsa/rsa_eay.c.
5076 [Bodo Moeller; problem pointed out by Marius Schilder]
5078 *) Various precautionary measures:
5080 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
5082 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
5083 (NB: This would require knowledge of the secret session ticket key
5084 to exploit, in which case you'd be SOL either way.)
5086 - Change bn_nist.c so that it will properly handle input BIGNUMs
5087 outside the expected range.
5089 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
5092 [Neel Mehta, Bodo Moeller]
5094 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
5095 the load fails. Useful for distros.
5096 [Ben Laurie and the FreeBSD team]
5098 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
5101 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
5104 *) Expand ENGINE to support engine supplied SSL client certificate functions.
5106 This work was sponsored by Logica.
5109 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
5110 keystores. Support for SSL/TLS client authentication too.
5111 Not compiled unless enable-capieng specified to Configure.
5113 This work was sponsored by Logica.
5116 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
5117 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
5118 attribute creation routines such as certificate requests and PKCS#12
5122 Changes between 0.9.8g and 0.9.8h [28 May 2008]
5124 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
5125 handshake which could lead to a client crash as found using the
5126 Codenomicon TLS test suite (CVE-2008-1672)
5127 [Steve Henson, Mark Cox]
5129 *) Fix double free in TLS server name extensions which could lead to
5130 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
5133 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
5135 Clear the error queue to ensure that error entries left from
5136 older function calls do not interfere with the correct operation.
5137 [Lutz Jaenicke, Erik de Castro Lopo]
5139 *) Remove root CA certificates of commercial CAs:
5141 The OpenSSL project does not recommend any specific CA and does not
5142 have any policy with respect to including or excluding any CA.
5143 Therefore it does not make any sense to ship an arbitrary selection
5144 of root CA certificates with the OpenSSL software.
5147 *) RSA OAEP patches to fix two separate invalid memory reads.
5148 The first one involves inputs when 'lzero' is greater than
5149 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
5150 before the beginning of from). The second one involves inputs where
5151 the 'db' section contains nothing but zeroes (there is a one-byte
5152 invalid read after the end of 'db').
5153 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5155 *) Partial backport from 0.9.9-dev:
5157 Introduce bn_mul_mont (dedicated Montgomery multiplication
5158 procedure) as a candidate for BIGNUM assembler implementation.
5159 While 0.9.9-dev uses assembler for various architectures, only
5160 x86_64 is available by default here in the 0.9.8 branch, and
5161 32-bit x86 is available through a compile-time setting.
5163 To try the 32-bit x86 assembler implementation, use Configure
5164 option "enable-montasm" (which exists only for this backport).
5166 As "enable-montasm" for 32-bit x86 disclaims code stability
5167 anyway, in this constellation we activate additional code
5168 backported from 0.9.9-dev for further performance improvements,
5169 namely BN_from_montgomery_word. (To enable this otherwise,
5170 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
5172 [Andy Polyakov (backport partially by Bodo Moeller)]
5174 *) Add TLS session ticket callback. This allows an application to set
5175 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
5176 values. This is useful for key rollover for example where several key
5177 sets may exist with different names.
5180 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
5181 This was broken until now in 0.9.8 releases, such that the only way
5182 a registered ENGINE could be used (assuming it initialises
5183 successfully on the host) was to explicitly set it as the default
5184 for the relevant algorithms. This is in contradiction with 0.9.7
5185 behaviour and the documentation. With this fix, when an ENGINE is
5186 registered into a given algorithm's table of implementations, the
5187 'uptodate' flag is reset so that auto-discovery will be used next
5188 time a new context for that algorithm attempts to select an
5190 [Ian Lister (tweaked by Geoff Thorpe)]
5192 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5193 implementation in the following ways:
5195 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5198 Lack of BER streaming support means one pass streaming processing is
5199 only supported if data is detached: setting the streaming flag is
5200 ignored for embedded content.
5202 CMS support is disabled by default and must be explicitly enabled
5203 with the enable-cms configuration option.
5206 *) Update the GMP engine glue to do direct copies between BIGNUM and
5207 mpz_t when openssl and GMP use the same limb size. Otherwise the
5208 existing "conversion via a text string export" trick is still used.
5209 [Paul Sheer <paulsheer@gmail.com>]
5211 *) Zlib compression BIO. This is a filter BIO which compressed and
5212 uncompresses any data passed through it.
5215 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5216 RFC3394 compatible AES key wrapping.
5219 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5220 sets string data without copying. X509_ALGOR_set0() and
5221 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5222 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5223 from an X509_ATTRIBUTE structure optionally checking it occurs only
5224 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5228 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5229 to get the expected BN_FLG_CONSTTIME behavior.
5230 [Bodo Moeller (Google)]
5234 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5235 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5236 - added some more tests to do_tests.pl
5237 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5238 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5239 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5240 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5241 - various changes to netware.pl to enable gcc-cross builds on Win32
5243 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5244 - various changes to fix missing prototype warnings
5245 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5246 - added AES, WHIRLPOOL and CPUID assembler code to build files
5247 - added missing AES assembler make rules to mk1mf.pl
5248 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5249 [Guenter Knauf <eflash@gmx.net>]
5251 *) Implement certificate status request TLS extension defined in RFC3546.
5252 A client can set the appropriate parameters and receive the encoded
5253 OCSP response via a callback. A server can query the supplied parameters
5254 and set the encoded OCSP response in the callback. Add simplified examples
5255 to s_client and s_server.
5258 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5260 *) Fix various bugs:
5261 + Binary incompatibility of ssl_ctx_st structure
5262 + DTLS interoperation with non-compliant servers
5263 + Don't call get_session_cb() without proposed session
5264 + Fix ia64 assembler code
5265 [Andy Polyakov, Steve Henson]
5267 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5269 *) DTLS Handshake overhaul. There were longstanding issues with
5270 OpenSSL DTLS implementation, which were making it impossible for
5271 RFC 4347 compliant client to communicate with OpenSSL server.
5272 Unfortunately just fixing these incompatibilities would "cut off"
5273 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5274 server keeps tolerating non RFC compliant syntax. The opposite is
5275 not true, 0.9.8f client can not communicate with earlier server.
5276 This update even addresses CVE-2007-4995.
5279 *) Changes to avoid need for function casts in OpenSSL: some compilers
5280 (gcc 4.2 and later) reject their use.
5281 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5284 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5285 RFC4507bis. The encrypted ticket format is an encrypted encoded
5286 SSL_SESSION structure, that way new session features are automatically
5289 If a client application caches session in an SSL_SESSION structure
5290 support is transparent because tickets are now stored in the encoded
5293 The SSL_CTX structure automatically generates keys for ticket
5294 protection in servers so again support should be possible
5295 with no application modification.
5297 If a client or server wishes to disable RFC4507 support then the option
5298 SSL_OP_NO_TICKET can be set.
5300 Add a TLS extension debugging callback to allow the contents of any client
5301 or server extensions to be examined.
5303 This work was sponsored by Google.
5306 *) Add initial support for TLS extensions, specifically for the server_name
5307 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5308 have new members for a host name. The SSL data structure has an
5309 additional member SSL_CTX *initial_ctx so that new sessions can be
5310 stored in that context to allow for session resumption, even after the
5311 SSL has been switched to a new SSL_CTX in reaction to a client's
5312 server_name extension.
5314 New functions (subject to change):
5316 SSL_get_servername()
5317 SSL_get_servername_type()
5320 New CTRL codes and macros (subject to change):
5322 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5323 - SSL_CTX_set_tlsext_servername_callback()
5324 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5325 - SSL_CTX_set_tlsext_servername_arg()
5326 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5328 openssl s_client has a new '-servername ...' option.
5330 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5331 '-key2 ...', '-servername_fatal' (subject to change). This allows
5332 testing the HostName extension for a specific single host name ('-cert'
5333 and '-key' remain fallbacks for handshakes without HostName
5334 negotiation). If the unrecognized_name alert has to be sent, this by
5335 default is a warning; it becomes fatal with the '-servername_fatal'
5338 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
5340 *) Add AES and SSE2 assembly language support to VC++ build.
5343 *) Mitigate attack on final subtraction in Montgomery reduction.
5346 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
5347 (which previously caused an internal error).
5350 *) Squeeze another 10% out of IGE mode when in != out.
5353 *) AES IGE mode speedup.
5354 [Dean Gaudet (Google)]
5356 *) Add the Korean symmetric 128-bit cipher SEED (see
5357 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
5358 add SEED ciphersuites from RFC 4162:
5360 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
5361 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
5362 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
5363 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
5365 To minimize changes between patchlevels in the OpenSSL 0.9.8
5366 series, SEED remains excluded from compilation unless OpenSSL
5367 is configured with 'enable-seed'.
5368 [KISA, Bodo Moeller]
5370 *) Mitigate branch prediction attacks, which can be practical if a
5371 single processor is shared, allowing a spy process to extract
5372 information. For detailed background information, see
5373 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
5374 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
5375 and Necessary Software Countermeasures"). The core of the change
5376 are new versions BN_div_no_branch() and
5377 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
5378 respectively, which are slower, but avoid the security-relevant
5379 conditional branches. These are automatically called by BN_div()
5380 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
5381 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
5382 remove a conditional branch.
5384 BN_FLG_CONSTTIME is the new name for the previous
5385 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
5386 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
5387 in the exponent causes BN_mod_exp_mont() to use the alternative
5388 implementation in BN_mod_exp_mont_consttime().) The old name
5389 remains as a deprecated alias.
5391 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
5392 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
5393 constant-time implementations for more than just exponentiation.
5394 Here too the old name is kept as a deprecated alias.
5396 BN_BLINDING_new() will now use BN_dup() for the modulus so that
5397 the BN_BLINDING structure gets an independent copy of the
5398 modulus. This means that the previous "BIGNUM *m" argument to
5399 BN_BLINDING_new() and to BN_BLINDING_create_param() now
5400 essentially becomes "const BIGNUM *m", although we can't actually
5401 change this in the header file before 0.9.9. It allows
5402 RSA_setup_blinding() to use BN_with_flags() on the modulus to
5403 enable BN_FLG_CONSTTIME.
5405 [Matthew D Wood (Intel Corp)]
5407 *) In the SSL/TLS server implementation, be strict about session ID
5408 context matching (which matters if an application uses a single
5409 external cache for different purposes). Previously,
5410 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
5411 set. This did ensure strict client verification, but meant that,
5412 with applications using a single external cache for quite
5413 different requirements, clients could circumvent ciphersuite
5414 restrictions for a given session ID context by starting a session
5415 in a different context.
5418 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5419 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5420 authentication-only ciphersuites.
5423 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
5424 not complete and could lead to a possible single byte overflow
5425 (CVE-2007-5135) [Ben Laurie]
5427 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
5429 *) Since AES128 and AES256 (and similarly Camellia128 and
5430 Camellia256) share a single mask bit in the logic of
5431 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5432 kludge to work properly if AES128 is available and AES256 isn't
5433 (or if Camellia128 is available and Camellia256 isn't).
5436 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
5437 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
5438 When a point or a seed is encoded in a BIT STRING, we need to
5439 prevent the removal of trailing zero bits to get the proper DER
5440 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
5441 of a NamedBitList, for which trailing 0 bits need to be removed.)
5444 *) Have SSL/TLS server implementation tolerate "mismatched" record
5445 protocol version while receiving ClientHello even if the
5446 ClientHello is fragmented. (The server can't insist on the
5447 particular protocol version it has chosen before the ServerHello
5448 message has informed the client about his choice.)
5451 *) Add RFC 3779 support.
5452 [Rob Austein for ARIN, Ben Laurie]
5454 *) Load error codes if they are not already present instead of using a
5455 static variable. This allows them to be cleanly unloaded and reloaded.
5456 Improve header file function name parsing.
5459 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
5460 or CAPABILITY handshake as required by RFCs.
5463 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5465 *) Introduce limits to prevent malicious keys being able to
5466 cause a denial of service. (CVE-2006-2940)
5467 [Steve Henson, Bodo Moeller]
5469 *) Fix ASN.1 parsing of certain invalid structures that can result
5470 in a denial of service. (CVE-2006-2937) [Steve Henson]
5472 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5473 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5475 *) Fix SSL client code which could crash if connecting to a
5476 malicious SSLv2 server. (CVE-2006-4343)
5477 [Tavis Ormandy and Will Drewry, Google Security Team]
5479 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
5480 match only those. Before that, "AES256-SHA" would be interpreted
5481 as a pattern and match "AES128-SHA" too (since AES128-SHA got
5482 the same strength classification in 0.9.7h) as we currently only
5483 have a single AES bit in the ciphersuite description bitmap.
5484 That change, however, also applied to ciphersuite strings such as
5485 "RC4-MD5" that intentionally matched multiple ciphersuites --
5486 namely, SSL 2.0 ciphersuites in addition to the more common ones
5487 from SSL 3.0/TLS 1.0.
5489 So we change the selection algorithm again: Naming an explicit
5490 ciphersuite selects this one ciphersuite, and any other similar
5491 ciphersuite (same bitmap) from *other* protocol versions.
5492 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5493 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5495 Since SSL 2.0 does not have any ciphersuites for which the
5496 128/256 bit distinction would be relevant, this works for now.
5497 The proper fix will be to use different bits for AES128 and
5498 AES256, which would have avoided the problems from the beginning;
5499 however, bits are scarce, so we can only do this in a new release
5500 (not just a patchlevel) when we can change the SSL_CIPHER
5501 definition to split the single 'unsigned long mask' bitmap into
5502 multiple values to extend the available space.
5506 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5508 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5509 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5511 *) Add AES IGE and biIGE modes.
5514 *) Change the Unix randomness entropy gathering to use poll() when
5515 possible instead of select(), since the latter has some
5516 undesirable limitations.
5517 [Darryl Miles via Richard Levitte and Bodo Moeller]
5519 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
5520 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
5521 cannot be implicitly activated as part of, e.g., the "AES" alias.
5522 However, please upgrade to OpenSSL 0.9.9[-dev] for
5523 non-experimental use of the ECC ciphersuites to get TLS extension
5524 support, which is required for curve and point format negotiation
5525 to avoid potential handshake problems.
5528 *) Disable rogue ciphersuites:
5530 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5531 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5532 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5534 The latter two were purportedly from
5535 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5538 Also deactivate the remaining ciphersuites from
5539 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5540 unofficial, and the ID has long expired.
5543 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5544 dual-core machines) and other potential thread-safety issues.
5547 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
5548 versions), which is now available for royalty-free use
5549 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
5550 Also, add Camellia TLS ciphersuites from RFC 4132.
5552 To minimize changes between patchlevels in the OpenSSL 0.9.8
5553 series, Camellia remains excluded from compilation unless OpenSSL
5554 is configured with 'enable-camellia'.
5557 *) Disable the padding bug check when compression is in use. The padding
5558 bug check assumes the first packet is of even length, this is not
5559 necessarily true if compression is enabled and can result in false
5560 positives causing handshake failure. The actual bug test is ancient
5561 code so it is hoped that implementations will either have fixed it by
5562 now or any which still have the bug do not support compression.
5565 Changes between 0.9.8a and 0.9.8b [04 May 2006]
5567 *) When applying a cipher rule check to see if string match is an explicit
5568 cipher suite and only match that one cipher suite if it is.
5571 *) Link in manifests for VC++ if needed.
5572 [Austin Ziegler <halostatue@gmail.com>]
5574 *) Update support for ECC-based TLS ciphersuites according to
5575 draft-ietf-tls-ecc-12.txt with proposed changes (but without
5576 TLS extensions, which are supported starting with the 0.9.9
5577 branch, not in the OpenSSL 0.9.8 branch).
5580 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
5581 opaque EVP_CIPHER_CTX handling.
5584 *) Fixes and enhancements to zlib compression code. We now only use
5585 "zlib1.dll" and use the default __cdecl calling convention on Win32
5586 to conform with the standards mentioned here:
5587 http://www.zlib.net/DLL_FAQ.txt
5588 Static zlib linking now works on Windows and the new --with-zlib-include
5589 --with-zlib-lib options to Configure can be used to supply the location
5590 of the headers and library. Gracefully handle case where zlib library
5594 *) Several fixes and enhancements to the OID generation code. The old code
5595 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
5596 handle numbers larger than ULONG_MAX, truncated printing and had a
5597 non standard OBJ_obj2txt() behaviour.
5600 *) Add support for building of engines under engine/ as shared libraries
5601 under VC++ build system.
5604 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
5605 Hopefully, we will not see any false combination of paths any more.
5608 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5610 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5611 (part of SSL_OP_ALL). This option used to disable the
5612 countermeasure against man-in-the-middle protocol-version
5613 rollback in the SSL 2.0 server implementation, which is a bad
5614 idea. (CVE-2005-2969)
5616 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5617 for Information Security, National Institute of Advanced Industrial
5618 Science and Technology [AIST], Japan)]
5620 *) Add two function to clear and return the verify parameter flags.
5623 *) Keep cipherlists sorted in the source instead of sorting them at
5624 runtime, thus removing the need for a lock.
5627 *) Avoid some small subgroup attacks in Diffie-Hellman.
5628 [Nick Mathewson and Ben Laurie]
5630 *) Add functions for well-known primes.
5633 *) Extended Windows CE support.
5634 [Satoshi Nakamura and Andy Polyakov]
5636 *) Initialize SSL_METHOD structures at compile time instead of during
5637 runtime, thus removing the need for a lock.
5640 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5641 attempting to decrypt each encrypted key in turn. Add support to
5645 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5647 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5650 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5653 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5654 key into the same file any more.
5657 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5660 *) Add -utf8 command line and config file option to 'ca'.
5661 [Stefan <stf@udoma.org]
5663 *) Removed the macro des_crypt(), as it seems to conflict with some
5664 libraries. Use DES_crypt().
5667 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5668 involves renaming the source and generated shared-libs for
5669 both. The engines will accept the corrected or legacy ids
5670 ('ncipher' and '4758_cca' respectively) when binding. NB,
5671 this only applies when building 'shared'.
5672 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5674 *) Add attribute functions to EVP_PKEY structure. Modify
5675 PKCS12_create() to recognize a CSP name attribute and
5676 use it. Make -CSP option work again in pkcs12 utility.
5679 *) Add new functionality to the bn blinding code:
5680 - automatic re-creation of the BN_BLINDING parameters after
5681 a fixed number of uses (currently 32)
5682 - add new function for parameter creation
5683 - introduce flags to control the update behaviour of the
5684 BN_BLINDING parameters
5685 - hide BN_BLINDING structure
5686 Add a second BN_BLINDING slot to the RSA structure to improve
5687 performance when a single RSA object is shared among several
5691 *) Add support for DTLS.
5692 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5694 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5695 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5698 *) Remove buggy and incomplete DH cert support from
5699 ssl/ssl_rsa.c and ssl/s3_both.c
5702 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5703 the apps/openssl applications.
5706 *) Compile clean with "-Wall -Wmissing-prototypes
5707 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5708 DEBUG_SAFESTACK must also be set.
5711 *) Change ./Configure so that certain algorithms can be disabled by default.
5712 The new counterpiece to "no-xxx" is "enable-xxx".
5714 The patented RC5 and MDC2 algorithms will now be disabled unless
5715 "enable-rc5" and "enable-mdc2", respectively, are specified.
5717 (IDEA remains enabled despite being patented. This is because IDEA
5718 is frequently required for interoperability, and there is no license
5719 fee for non-commercial use. As before, "no-idea" can be used to
5720 avoid this algorithm.)
5724 *) Add processing of proxy certificates (see RFC 3820). This work was
5725 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5726 EGEE (Enabling Grids for E-science in Europe).
5729 *) RC4 performance overhaul on modern architectures/implementations, such
5730 as Intel P4, IA-64 and AMD64.
5733 *) New utility extract-section.pl. This can be used specify an alternative
5734 section number in a pod file instead of having to treat each file as
5735 a separate case in Makefile. This can be done by adding two lines to the
5738 =for comment openssl_section:XXX
5740 The blank line is mandatory.
5744 *) New arguments -certform, -keyform and -pass for s_client and s_server
5745 to allow alternative format key and certificate files and passphrase
5749 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5750 update associated structures and add various utility functions.
5752 Add new policy related verify parameters, include policy checking in
5753 standard verify code. Enhance 'smime' application with extra parameters
5754 to support policy checking and print out.
5757 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
5758 Nehemiah processors. These extensions support AES encryption in hardware
5759 as well as RNG (though RNG support is currently disabled).
5760 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
5762 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
5765 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
5766 [Andy Polyakov and a number of other people]
5768 *) Improved PowerPC platform support. Most notably BIGNUM assembler
5769 implementation contributed by IBM.
5770 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
5772 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
5773 exponent rather than 'unsigned long'. There is a corresponding change to
5774 the new 'rsa_keygen' element of the RSA_METHOD structure.
5775 [Jelte Jansen, Geoff Thorpe]
5777 *) Functionality for creating the initial serial number file is now
5778 moved from CA.pl to the 'ca' utility with a new option -create_serial.
5780 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
5781 number file to 1, which is bound to cause problems. To avoid
5782 the problems while respecting compatibility between different 0.9.7
5783 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
5784 CA.pl for serial number initialization. With the new release 0.9.8,
5785 we can fix the problem directly in the 'ca' utility.)
5788 *) Reduced header interdependencies by declaring more opaque objects in
5789 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
5790 give fewer recursive includes, which could break lazy source code - so
5791 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
5792 developers should define this symbol when building and using openssl to
5793 ensure they track the recommended behaviour, interfaces, [etc], but
5794 backwards-compatible behaviour prevails when this isn't defined.
5797 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
5800 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
5801 This will generate a random key of the appropriate length based on the
5802 cipher context. The EVP_CIPHER can provide its own random key generation
5803 routine to support keys of a specific form. This is used in the des and
5804 3des routines to generate a key of the correct parity. Update S/MIME
5805 code to use new functions and hence generate correct parity DES keys.
5806 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
5807 valid (weak or incorrect parity).
5810 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
5811 as looking them up. This is useful when the verified structure may contain
5812 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
5813 present unless the new PKCS7_NO_CRL flag is asserted.
5816 *) Extend ASN1 oid configuration module. It now additionally accepts the
5819 shortName = some long name, 1.2.3.4
5822 *) Reimplemented the BN_CTX implementation. There is now no more static
5823 limitation on the number of variables it can handle nor the depth of the
5824 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
5825 information can now expand as required, and rather than having a single
5826 static array of bignums, BN_CTX now uses a linked-list of such arrays
5827 allowing it to expand on demand whilst maintaining the usefulness of
5828 BN_CTX's "bundling".
5831 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5832 to allow all RSA operations to function using a single BN_CTX.
5835 *) Preliminary support for certificate policy evaluation and checking. This
5836 is initially intended to pass the tests outlined in "Conformance Testing
5837 of Relying Party Client Certificate Path Processing Logic" v1.07.
5840 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
5841 remained unused and not that useful. A variety of other little bignum
5842 tweaks and fixes have also been made continuing on from the audit (see
5846 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
5847 associated ASN1, EVP and SSL functions and old ASN1 macros.
5850 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
5851 and this should never fail. So the return value from the use of
5852 BN_set_word() (which can fail due to needless expansion) is now deprecated;
5853 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
5856 *) BN_CTX_get() should return zero-valued bignums, providing the same
5857 initialised value as BN_new().
5858 [Geoff Thorpe, suggested by Ulf Möller]
5860 *) Support for inhibitAnyPolicy certificate extension.
5863 *) An audit of the BIGNUM code is underway, for which debugging code is
5864 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
5865 is considered valid when processing BIGNUMs, and causes execution to
5866 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
5867 further steps are taken to deliberately pollute unused data in BIGNUM
5868 structures to try and expose faulty code further on. For now, openssl will
5869 (in its default mode of operation) continue to tolerate the inconsistent
5870 forms that it has tolerated in the past, but authors and packagers should
5871 consider trying openssl and their own applications when compiled with
5872 these debugging symbols defined. It will help highlight potential bugs in
5873 their own code, and will improve the test coverage for OpenSSL itself. At
5874 some point, these tighter rules will become openssl's default to improve
5875 maintainability, though the assert()s and other overheads will remain only
5876 in debugging configurations. See bn.h for more details.
5877 [Geoff Thorpe, Nils Larsch, Ulf Möller]
5879 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
5880 that can only be obtained through BN_CTX_new() (which implicitly
5881 initialises it). The presence of this function only made it possible
5882 to overwrite an existing structure (and cause memory leaks).
5885 *) Because of the callback-based approach for implementing LHASH as a
5886 template type, lh_insert() adds opaque objects to hash-tables and
5887 lh_doall() or lh_doall_arg() are typically used with a destructor callback
5888 to clean up those corresponding objects before destroying the hash table
5889 (and losing the object pointers). So some over-zealous constifications in
5890 LHASH have been relaxed so that lh_insert() does not take (nor store) the
5891 objects as "const" and the lh_doall[_arg] callback wrappers are not
5892 prototyped to have "const" restrictions on the object pointers they are
5893 given (and so aren't required to cast them away any more).
5896 *) The tmdiff.h API was so ugly and minimal that our own timing utility
5897 (speed) prefers to use its own implementation. The two implementations
5898 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
5899 its object type properly exposed (MS_TM) instead of casting to/from "char
5900 *". This may still change yet if someone realises MS_TM and "ms_time_***"
5901 aren't necessarily the greatest nomenclatures - but this is what was used
5902 internally to the implementation so I've used that for now.
5905 *) Ensure that deprecated functions do not get compiled when
5906 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
5907 the self-tests were still using deprecated key-generation functions so
5908 these have been updated also.
5911 *) Reorganise PKCS#7 code to separate the digest location functionality
5912 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
5913 New function PKCS7_set_digest() to set the digest type for PKCS#7
5914 digestedData type. Add additional code to correctly generate the
5915 digestedData type and add support for this type in PKCS7 initialization
5919 *) New function PKCS7_set0_type_other() this initializes a PKCS7
5920 structure of type "other".
5923 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
5924 sure the loop does correctly stop and breaking ("division by zero")
5925 modulus operations are not performed. The (pre-generated) prime
5926 table crypto/bn/bn_prime.h was already correct, but it could not be
5927 re-generated on some platforms because of the "division by zero"
5928 situation in the script.
5929 [Ralf S. Engelschall]
5931 *) Update support for ECC-based TLS ciphersuites according to
5932 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
5933 SHA-1 now is only used for "small" curves (where the
5934 representation of a field element takes up to 24 bytes); for
5935 larger curves, the field element resulting from ECDH is directly
5936 used as premaster secret.
5937 [Douglas Stebila (Sun Microsystems Laboratories)]
5939 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
5940 curve secp160r1 to the tests.
5941 [Douglas Stebila (Sun Microsystems Laboratories)]
5943 *) Add the possibility to load symbols globally with DSO.
5944 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
5946 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
5947 control of the error stack.
5950 *) Add support for STORE in ENGINE.
5953 *) Add the STORE type. The intention is to provide a common interface
5954 to certificate and key stores, be they simple file-based stores, or
5955 HSM-type store, or LDAP stores, or...
5956 NOTE: The code is currently UNTESTED and isn't really used anywhere.
5959 *) Add a generic structure called OPENSSL_ITEM. This can be used to
5960 pass a list of arguments to any function as well as provide a way
5961 for a function to pass data back to the caller.
5964 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
5965 works like BUF_strdup() but can be used to duplicate a portion of
5966 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
5970 *) Add the function sk_find_ex() which works like sk_find(), but will
5971 return an index to an element even if an exact match couldn't be
5972 found. The index is guaranteed to point at the element where the
5973 searched-for key would be inserted to preserve sorting order.
5976 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
5977 takes an extra flags argument for optional functionality. Currently,
5978 the following flags are defined:
5980 OBJ_BSEARCH_VALUE_ON_NOMATCH
5981 This one gets OBJ_bsearch_ex() to return a pointer to the first
5982 element where the comparing function returns a negative or zero
5985 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
5986 This one gets OBJ_bsearch_ex() to return a pointer to the first
5987 element where the comparing function returns zero. This is useful
5988 if there are more than one element where the comparing function
5992 *) Make it possible to create self-signed certificates with 'openssl ca'
5993 in such a way that the self-signed certificate becomes part of the
5994 CA database and uses the same mechanisms for serial number generation
5995 as all other certificate signing. The new flag '-selfsign' enables
5996 this functionality. Adapt CA.sh and CA.pl.in.
5999 *) Add functionality to check the public key of a certificate request
6000 against a given private. This is useful to check that a certificate
6001 request can be signed by that key (self-signing).
6004 *) Make it possible to have multiple active certificates with the same
6005 subject in the CA index file. This is done only if the keyword
6006 'unique_subject' is set to 'no' in the main CA section (default
6007 if 'CA_default') of the configuration file. The value is saved
6008 with the database itself in a separate index attribute file,
6009 named like the index file with '.attr' appended to the name.
6012 *) Generate multi-valued AVAs using '+' notation in config files for
6016 *) Support for nameConstraints certificate extension.
6019 *) Support for policyConstraints certificate extension.
6022 *) Support for policyMappings certificate extension.
6025 *) Make sure the default DSA_METHOD implementation only uses its
6026 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
6027 and change its own handlers to be NULL so as to remove unnecessary
6028 indirection. This lets alternative implementations fallback to the
6029 default implementation more easily.
6032 *) Support for directoryName in GeneralName related extensions
6036 *) Make it possible to link applications using Makefile.shared.
6037 Make that possible even when linking against static libraries!
6040 *) Support for single pass processing for S/MIME signing. This now
6041 means that S/MIME signing can be done from a pipe, in addition
6042 cleartext signing (multipart/signed type) is effectively streaming
6043 and the signed data does not need to be all held in memory.
6045 This is done with a new flag PKCS7_STREAM. When this flag is set
6046 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
6047 is done after the data is output (and digests calculated) in
6048 SMIME_write_PKCS7().
6051 *) Add full support for -rpath/-R, both in shared libraries and
6052 applications, at least on the platforms where it's known how
6056 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
6057 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
6058 will now compute a table of multiples of the generator that
6059 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
6060 faster (notably in the case of a single point multiplication,
6061 scalar * generator).
6062 [Nils Larsch, Bodo Moeller]
6064 *) IPv6 support for certificate extensions. The various extensions
6065 which use the IP:a.b.c.d can now take IPv6 addresses using the
6066 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
6070 *) Added an ENGINE that implements RSA by performing private key
6071 exponentiations with the GMP library. The conversions to and from
6072 GMP's mpz_t format aren't optimised nor are any montgomery forms
6073 cached, and on x86 it appears OpenSSL's own performance has caught up.
6074 However there are likely to be other architectures where GMP could
6075 provide a boost. This ENGINE is not built in by default, but it can be
6076 specified at Configure time and should be accompanied by the necessary
6077 linker additions, eg;
6078 ./config -DOPENSSL_USE_GMP -lgmp
6081 *) "openssl engine" will not display ENGINE/DSO load failure errors when
6082 testing availability of engines with "-t" - the old behaviour is
6083 produced by increasing the feature's verbosity with "-tt".
6086 *) ECDSA routines: under certain error conditions uninitialized BN objects
6087 could be freed. Solution: make sure initialization is performed early
6088 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
6092 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
6093 and DH_METHOD (eg. by ENGINE implementations) to override the normal
6094 software implementations. For DSA and DH, parameter generation can
6095 also be overridden by providing the appropriate method callbacks.
6098 *) Change the "progress" mechanism used in key-generation and
6099 primality testing to functions that take a new BN_GENCB pointer in
6100 place of callback/argument pairs. The new API functions have "_ex"
6101 postfixes and the older functions are reimplemented as wrappers for
6102 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
6103 declarations of the old functions to help (graceful) attempts to
6104 migrate to the new functions. Also, the new key-generation API
6105 functions operate on a caller-supplied key-structure and return
6106 success/failure rather than returning a key or NULL - this is to
6107 help make "keygen" another member function of RSA_METHOD etc.
6109 Example for using the new callback interface:
6111 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
6115 BN_GENCB_set(&my_cb, my_callback, my_arg);
6117 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
6118 /* For the meaning of a, b in calls to my_callback(), see the
6119 * documentation of the function that calls the callback.
6120 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
6121 * my_callback should return 1 if it wants BN_is_prime_ex()
6122 * to continue, or 0 to stop.
6127 *) Change the ZLIB compression method to be stateful, and make it
6128 available to TLS with the number defined in
6129 draft-ietf-tls-compression-04.txt.
6132 *) Add the ASN.1 structures and functions for CertificatePair, which
6133 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
6135 CertificatePair ::= SEQUENCE {
6136 forward [0] Certificate OPTIONAL,
6137 reverse [1] Certificate OPTIONAL,
6138 -- at least one of the pair shall be present -- }
6140 Also implement the PEM functions to read and write certificate
6141 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
6143 This needed to be defined, mostly for the sake of the LDAP
6144 attribute crossCertificatePair, but may prove useful elsewhere as
6148 *) Make it possible to inhibit symlinking of shared libraries in
6149 Makefile.shared, for Cygwin's sake.
6152 *) Extend the BIGNUM API by creating a function
6153 void BN_set_negative(BIGNUM *a, int neg);
6154 and a macro that behave like
6155 int BN_is_negative(const BIGNUM *a);
6157 to avoid the need to access 'a->neg' directly in applications.
6160 *) Implement fast modular reduction for pseudo-Mersenne primes
6161 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
6162 EC_GROUP_new_curve_GFp() will now automatically use this
6164 [Nils Larsch <nla@trustcenter.de>]
6166 *) Add new lock type (CRYPTO_LOCK_BN).
6169 *) Change the ENGINE framework to automatically load engines
6170 dynamically from specific directories unless they could be
6171 found to already be built in or loaded. Move all the
6172 current engines except for the cryptodev one to a new
6174 The engines in engines/ are built as shared libraries if
6175 the "shared" options was given to ./Configure or ./config.
6176 Otherwise, they are inserted in libcrypto.a.
6177 /usr/local/ssl/engines is the default directory for dynamic
6178 engines, but that can be overridden at configure time through
6179 the usual use of --prefix and/or --openssldir, and at run
6180 time with the environment variable OPENSSL_ENGINES.
6181 [Geoff Thorpe and Richard Levitte]
6183 *) Add Makefile.shared, a helper makefile to build shared
6184 libraries. Adapt Makefile.org.
6187 *) Add version info to Win32 DLLs.
6188 [Peter 'Luna' Runestig" <peter@runestig.com>]
6190 *) Add new 'medium level' PKCS#12 API. Certificates and keys
6191 can be added using this API to created arbitrary PKCS#12
6192 files while avoiding the low level API.
6194 New options to PKCS12_create(), key or cert can be NULL and
6195 will then be omitted from the output file. The encryption
6196 algorithm NIDs can be set to -1 for no encryption, the mac
6197 iteration count can be set to 0 to omit the mac.
6199 Enhance pkcs12 utility by making the -nokeys and -nocerts
6200 options work when creating a PKCS#12 file. New option -nomac
6201 to omit the mac, NONE can be set for an encryption algorithm.
6202 New code is modified to use the enhanced PKCS12_create()
6203 instead of the low level API.
6206 *) Extend ASN1 encoder to support indefinite length constructed
6207 encoding. This can output sequences tags and octet strings in
6208 this form. Modify pk7_asn1.c to support indefinite length
6209 encoding. This is experimental and needs additional code to
6210 be useful, such as an ASN1 bio and some enhanced streaming
6213 Extend template encode functionality so that tagging is passed
6214 down to the template encoder.
6217 *) Let 'openssl req' fail if an argument to '-newkey' is not
6218 recognized instead of using RSA as a default.
6221 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6222 As these are not official, they are not included in "ALL";
6223 the "ECCdraft" ciphersuite group alias can be used to select them.
6224 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6226 *) Add ECDH engine support.
6227 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6229 *) Add ECDH in new directory crypto/ecdh/.
6230 [Douglas Stebila (Sun Microsystems Laboratories)]
6232 *) Let BN_rand_range() abort with an error after 100 iterations
6233 without success (which indicates a broken PRNG).
6236 *) Change BN_mod_sqrt() so that it verifies that the input value
6237 is really the square of the return value. (Previously,
6238 BN_mod_sqrt would show GIGO behaviour.)
6241 *) Add named elliptic curves over binary fields from X9.62, SECG,
6242 and WAP/WTLS; add OIDs that were still missing.
6244 [Sheueling Chang Shantz and Douglas Stebila
6245 (Sun Microsystems Laboratories)]
6247 *) Extend the EC library for elliptic curves over binary fields
6248 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6251 EC_GF2m_simple_method
6255 EC_GROUP_new_curve_GF2m
6256 EC_GROUP_set_curve_GF2m
6257 EC_GROUP_get_curve_GF2m
6258 EC_POINT_set_affine_coordinates_GF2m
6259 EC_POINT_get_affine_coordinates_GF2m
6260 EC_POINT_set_compressed_coordinates_GF2m
6262 Point compression for binary fields is disabled by default for
6263 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6266 As binary polynomials are represented as BIGNUMs, various members
6267 of the EC_GROUP and EC_POINT data structures can be shared
6268 between the implementations for prime fields and binary fields;
6269 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6270 are essentially identical to their ..._GFp counterparts.
6271 (For simplicity, the '..._GFp' prefix has been dropped from
6272 various internal method names.)
6274 An internal 'field_div' method (similar to 'field_mul' and
6275 'field_sqr') has been added; this is used only for binary fields.
6277 [Sheueling Chang Shantz and Douglas Stebila
6278 (Sun Microsystems Laboratories)]
6280 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6281 through methods ('mul', 'precompute_mult').
6283 The generic implementations (now internally called 'ec_wNAF_mul'
6284 and 'ec_wNAF_precomputed_mult') remain the default if these
6285 methods are undefined.
6287 [Sheueling Chang Shantz and Douglas Stebila
6288 (Sun Microsystems Laboratories)]
6290 *) New function EC_GROUP_get_degree, which is defined through
6291 EC_METHOD. For curves over prime fields, this returns the bit
6292 length of the modulus.
6294 [Sheueling Chang Shantz and Douglas Stebila
6295 (Sun Microsystems Laboratories)]
6297 *) New functions EC_GROUP_dup, EC_POINT_dup.
6298 (These simply call ..._new and ..._copy).
6300 [Sheueling Chang Shantz and Douglas Stebila
6301 (Sun Microsystems Laboratories)]
6303 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6304 Polynomials are represented as BIGNUMs (where the sign bit is not
6305 used) in the following functions [macros]:
6308 BN_GF2m_sub [= BN_GF2m_add]
6309 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6310 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6311 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6313 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6314 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6315 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6316 BN_GF2m_cmp [= BN_ucmp]
6318 (Note that only the 'mod' functions are actually for fields GF(2^m).
6319 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6321 For some functions, an the irreducible polynomial defining a
6322 field can be given as an 'unsigned int[]' with strictly
6323 decreasing elements giving the indices of those bits that are set;
6324 i.e., p[] represents the polynomial
6325 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6327 p[0] > p[1] > ... > p[k] = 0.
6328 This applies to the following functions:
6333 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
6334 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
6336 BN_GF2m_mod_sqrt_arr
6337 BN_GF2m_mod_solve_quad_arr
6341 Conversion can be performed by the following functions:
6346 bntest.c has additional tests for binary polynomial arithmetic.
6348 Two implementations for BN_GF2m_mod_div() are available.
6349 The default algorithm simply uses BN_GF2m_mod_inv() and
6350 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
6351 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
6352 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
6354 [Sheueling Chang Shantz and Douglas Stebila
6355 (Sun Microsystems Laboratories)]
6357 *) Add new error code 'ERR_R_DISABLED' that can be used when some
6358 functionality is disabled at compile-time.
6359 [Douglas Stebila <douglas.stebila@sun.com>]
6361 *) Change default behaviour of 'openssl asn1parse' so that more
6362 information is visible when viewing, e.g., a certificate:
6364 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
6365 mode the content of non-printable OCTET STRINGs is output in a
6366 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
6367 avoid the appearance of a printable string.
6368 [Nils Larsch <nla@trustcenter.de>]
6370 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
6372 EC_GROUP_set_asn1_flag()
6373 EC_GROUP_get_asn1_flag()
6374 EC_GROUP_set_point_conversion_form()
6375 EC_GROUP_get_point_conversion_form()
6376 These control ASN1 encoding details:
6377 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
6378 has been set to OPENSSL_EC_NAMED_CURVE.
6379 - Points are encoded in uncompressed form by default; options for
6380 asn1_for are as for point2oct, namely
6381 POINT_CONVERSION_COMPRESSED
6382 POINT_CONVERSION_UNCOMPRESSED
6383 POINT_CONVERSION_HYBRID
6385 Also add 'seed' and 'seed_len' members to EC_GROUP with access
6388 EC_GROUP_get0_seed()
6389 EC_GROUP_get_seed_len()
6390 This is used only for ASN1 purposes (so far).
6391 [Nils Larsch <nla@trustcenter.de>]
6393 *) Add 'field_type' member to EC_METHOD, which holds the NID
6394 of the appropriate field type OID. The new function
6395 EC_METHOD_get_field_type() returns this value.
6396 [Nils Larsch <nla@trustcenter.de>]
6401 EC_POINT_point2hex()
6402 EC_POINT_hex2point()
6403 providing useful interfaces to EC_POINT_point2oct() and
6404 EC_POINT_oct2point().
6405 [Nils Larsch <nla@trustcenter.de>]
6407 *) Change internals of the EC library so that the functions
6408 EC_GROUP_set_generator()
6409 EC_GROUP_get_generator()
6410 EC_GROUP_get_order()
6411 EC_GROUP_get_cofactor()
6412 are implemented directly in crypto/ec/ec_lib.c and not dispatched
6413 to methods, which would lead to unnecessary code duplication when
6414 adding different types of curves.
6415 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
6417 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
6418 arithmetic, and such that modified wNAFs are generated
6419 (which avoid length expansion in many cases).
6422 *) Add a function EC_GROUP_check_discriminant() (defined via
6423 EC_METHOD) that verifies that the curve discriminant is non-zero.
6425 Add a function EC_GROUP_check() that makes some sanity tests
6426 on a EC_GROUP, its generator and order. This includes
6427 EC_GROUP_check_discriminant().
6428 [Nils Larsch <nla@trustcenter.de>]
6430 *) Add ECDSA in new directory crypto/ecdsa/.
6432 Add applications 'openssl ecparam' and 'openssl ecdsa'
6433 (these are based on 'openssl dsaparam' and 'openssl dsa').
6435 ECDSA support is also included in various other files across the
6436 library. Most notably,
6437 - 'openssl req' now has a '-newkey ecdsa:file' option;
6438 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
6439 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
6440 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
6441 them suitable for ECDSA where domain parameters must be
6442 extracted before the specific public key;
6443 - ECDSA engine support has been added.
6444 [Nils Larsch <nla@trustcenter.de>]
6446 *) Include some named elliptic curves, and add OIDs from X9.62,
6447 SECG, and WAP/WTLS. Each curve can be obtained from the new
6449 EC_GROUP_new_by_curve_name(),
6450 and the list of available named curves can be obtained with
6451 EC_get_builtin_curves().
6452 Also add a 'curve_name' member to EC_GROUP objects, which can be
6454 EC_GROUP_set_curve_name()
6455 EC_GROUP_get_curve_name()
6456 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
6458 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6459 was actually never needed) and in BN_mul(). The removal in BN_mul()
6460 required a small change in bn_mul_part_recursive() and the addition
6461 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6462 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6463 bn_sub_words() and bn_add_words() except they take arrays with
6467 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
6469 *) Cleanse PEM buffers before freeing them since they may contain
6471 [Benjamin Bennett <ben@psc.edu>]
6473 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6474 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6475 authentication-only ciphersuites.
6478 *) Since AES128 and AES256 share a single mask bit in the logic of
6479 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6480 kludge to work properly if AES128 is available and AES256 isn't.
6483 *) Expand security boundary to match 1.1.1 module.
6486 *) Remove redundant features: hash file source, editing of test vectors
6487 modify fipsld to use external fips_premain.c signature.
6490 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6491 run algorithm test programs.
6494 *) Make algorithm test programs more tolerant of whitespace.
6497 *) Have SSL/TLS server implementation tolerate "mismatched" record
6498 protocol version while receiving ClientHello even if the
6499 ClientHello is fragmented. (The server can't insist on the
6500 particular protocol version it has chosen before the ServerHello
6501 message has informed the client about his choice.)
6504 *) Load error codes if they are not already present instead of using a
6505 static variable. This allows them to be cleanly unloaded and reloaded.
6508 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6510 *) Introduce limits to prevent malicious keys being able to
6511 cause a denial of service. (CVE-2006-2940)
6512 [Steve Henson, Bodo Moeller]
6514 *) Fix ASN.1 parsing of certain invalid structures that can result
6515 in a denial of service. (CVE-2006-2937) [Steve Henson]
6517 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6518 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6520 *) Fix SSL client code which could crash if connecting to a
6521 malicious SSLv2 server. (CVE-2006-4343)
6522 [Tavis Ormandy and Will Drewry, Google Security Team]
6524 *) Change ciphersuite string processing so that an explicit
6525 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
6526 will no longer include "AES128-SHA"), and any other similar
6527 ciphersuite (same bitmap) from *other* protocol versions (so that
6528 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
6529 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
6530 changes from 0.9.8b and 0.9.8d.
6533 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
6535 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6536 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6538 *) Change the Unix randomness entropy gathering to use poll() when
6539 possible instead of select(), since the latter has some
6540 undesirable limitations.
6541 [Darryl Miles via Richard Levitte and Bodo Moeller]
6543 *) Disable rogue ciphersuites:
6545 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6546 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6547 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6549 The latter two were purportedly from
6550 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6553 Also deactivate the remaining ciphersuites from
6554 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6555 unofficial, and the ID has long expired.
6558 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6559 dual-core machines) and other potential thread-safety issues.
6562 Changes between 0.9.7i and 0.9.7j [04 May 2006]
6564 *) Adapt fipsld and the build system to link against the validated FIPS
6565 module in FIPS mode.
6568 *) Fixes for VC++ 2005 build under Windows.
6571 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
6572 from a Windows bash shell such as MSYS. It is autodetected from the
6573 "config" script when run from a VC++ environment. Modify standard VC++
6574 build to use fipscanister.o from the GNU make build.
6577 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
6579 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
6580 The value now differs depending on if you build for FIPS or not.
6581 BEWARE! A program linked with a shared FIPSed libcrypto can't be
6582 safely run with a non-FIPSed libcrypto, as it may crash because of
6583 the difference induced by this change.
6586 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
6588 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6589 (part of SSL_OP_ALL). This option used to disable the
6590 countermeasure against man-in-the-middle protocol-version
6591 rollback in the SSL 2.0 server implementation, which is a bad
6592 idea. (CVE-2005-2969)
6594 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6595 for Information Security, National Institute of Advanced Industrial
6596 Science and Technology [AIST], Japan)]
6598 *) Minimal support for X9.31 signatures and PSS padding modes. This is
6599 mainly for FIPS compliance and not fully integrated at this stage.
6602 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
6603 the exponentiation using a fixed-length exponent. (Otherwise,
6604 the information leaked through timing could expose the secret key
6605 after many signatures; cf. Bleichenbacher's attack on DSA with
6609 *) Make a new fixed-window mod_exp implementation the default for
6610 RSA, DSA, and DH private-key operations so that the sequence of
6611 squares and multiplies and the memory access pattern are
6612 independent of the particular secret key. This will mitigate
6613 cache-timing and potential related attacks.
6615 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6616 and this is automatically used by BN_mod_exp_mont() if the new flag
6617 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6618 will use this BN flag for private exponents unless the flag
6619 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6620 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6622 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6624 *) Change the client implementation for SSLv23_method() and
6625 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6626 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6627 (Previously, the SSL 2.0 backwards compatible Client Hello
6628 message format would be used even with SSL_OP_NO_SSLv2.)
6631 *) Add support for smime-type MIME parameter in S/MIME messages which some
6635 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6636 a threadsafe manner. Modify rsa code to use new function and add calls
6637 to dsa and dh code (which had race conditions before).
6640 *) Include the fixed error library code in the C error file definitions
6641 instead of fixing them up at runtime. This keeps the error code
6642 structures constant.
6645 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6647 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6650 *) Fixes for newer kerberos headers. NB: the casts are needed because
6651 the 'length' field is signed on one version and unsigned on another
6652 with no (?) obvious way to tell the difference, without these VC++
6653 complains. Also the "definition" of FAR (blank) is no longer included
6654 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6655 some needed definitions.
6658 *) Undo Cygwin change.
6661 *) Added support for proxy certificates according to RFC 3820.
6662 Because they may be a security thread to unaware applications,
6663 they must be explicitly allowed in run-time. See
6664 docs/HOWTO/proxy_certificates.txt for further information.
6667 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6669 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6670 server and client random values. Previously
6671 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6672 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6674 This change has negligible security impact because:
6676 1. Server and client random values still have 24 bytes of pseudo random
6679 2. Server and client random values are sent in the clear in the initial
6682 3. The master secret is derived using the premaster secret (48 bytes in
6683 size for static RSA ciphersuites) as well as client server and random
6686 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6689 [Stephen Henson, reported by UK NISCC]
6691 *) Use Windows randomness collection on Cygwin.
6694 *) Fix hang in EGD/PRNGD query when communication socket is closed
6695 prematurely by EGD/PRNGD.
6696 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6698 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6701 *) Back-port of selected performance improvements from development
6702 branch, as well as improved support for PowerPC platforms.
6705 *) Add lots of checks for memory allocation failure, error codes to indicate
6706 failure and freeing up memory if a failure occurs.
6707 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6709 *) Add new -passin argument to dgst.
6712 *) Perform some character comparisons of different types in X509_NAME_cmp:
6713 this is needed for some certificates that re-encode DNs into UTF8Strings
6714 (in violation of RFC3280) and can't or won't issue name rollover
6718 *) Make an explicit check during certificate validation to see that
6719 the CA setting in each certificate on the chain is correct. As a
6720 side effect always do the following basic checks on extensions,
6721 not just when there's an associated purpose to the check:
6723 - if there is an unhandled critical extension (unless the user
6724 has chosen to ignore this fault)
6725 - if the path length has been exceeded (if one is set at all)
6726 - that certain extensions fit the associated purpose (if one has
6730 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6732 *) Avoid a race condition when CRLs are checked in a multi threaded
6733 environment. This would happen due to the reordering of the revoked
6734 entries during signature checking and serial number lookup. Now the
6735 encoding is cached and the serial number sort performed under a lock.
6736 Add new STACK function sk_is_sorted().
6739 *) Add Delta CRL to the extension code.
6742 *) Various fixes to s3_pkt.c so alerts are sent properly.
6743 [David Holmes <d.holmes@f5.com>]
6745 *) Reduce the chances of duplicate issuer name and serial numbers (in
6746 violation of RFC3280) using the OpenSSL certificate creation utilities.
6747 This is done by creating a random 64 bit value for the initial serial
6748 number when a serial number file is created or when a self signed
6749 certificate is created using 'openssl req -x509'. The initial serial
6750 number file is created using 'openssl x509 -next_serial' in CA.pl
6751 rather than being initialized to 1.
6754 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
6756 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6757 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6758 [Joe Orton, Steve Henson]
6760 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
6762 [Joe Orton, Steve Henson]
6764 *) Make it possible to have multiple active certificates with the same
6765 subject in the CA index file. This is done only if the keyword
6766 'unique_subject' is set to 'no' in the main CA section (default
6767 if 'CA_default') of the configuration file. The value is saved
6768 with the database itself in a separate index attribute file,
6769 named like the index file with '.attr' appended to the name.
6772 *) X509 verify fixes. Disable broken certificate workarounds when
6773 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
6774 keyUsage extension present. Don't accept CRLs with unhandled critical
6775 extensions: since verify currently doesn't process CRL extensions this
6776 rejects a CRL with *any* critical extensions. Add new verify error codes
6780 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
6781 A clarification of RFC2560 will require the use of OCTET STRINGs and
6782 some implementations cannot handle the current raw format. Since OpenSSL
6783 copies and compares OCSP nonces as opaque blobs without any attempt at
6784 parsing them this should not create any compatibility issues.
6787 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
6788 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
6789 this HMAC (and other) operations are several times slower than OpenSSL
6793 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
6794 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6796 *) Use the correct content when signing type "other".
6799 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
6801 *) Fix various bugs revealed by running the NISCC test suite:
6803 Stop out of bounds reads in the ASN1 code when presented with
6804 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6806 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
6808 If verify callback ignores invalid public key errors don't try to check
6809 certificate signature with the NULL public key.
6813 *) New -ignore_err option in ocsp application to stop the server
6814 exiting on the first error in a request.
6817 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6818 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6822 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6823 extra data after the compression methods not only for TLS 1.0
6824 but also for SSL 3.0 (as required by the specification).
6825 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6827 *) Change X509_certificate_type() to mark the key as exported/exportable
6828 when it's 512 *bits* long, not 512 bytes.
6831 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6832 blocks during encryption.
6835 *) Various fixes to base64 BIO and non blocking I/O. On write
6836 flushes were not handled properly if the BIO retried. On read
6837 data was not being buffered properly and had various logic bugs.
6838 This also affects blocking I/O when the data being decoded is a
6842 *) Various S/MIME bugfixes and compatibility changes:
6843 output correct application/pkcs7 MIME type if
6844 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
6845 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
6846 of files as .eml work). Correctly handle very long lines in MIME
6850 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
6852 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6853 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6854 a protocol version number mismatch like a decryption error
6855 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6858 *) Turn on RSA blinding by default in the default implementation
6859 to avoid a timing attack. Applications that don't want it can call
6860 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6861 They would be ill-advised to do so in most cases.
6862 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6864 *) Change RSA blinding code so that it works when the PRNG is not
6865 seeded (in this case, the secret RSA exponent is abused as
6866 an unpredictable seed -- if it is not unpredictable, there
6867 is no point in blinding anyway). Make RSA blinding thread-safe
6868 by remembering the creator's thread ID in rsa->blinding and
6869 having all other threads use local one-time blinding factors
6870 (this requires more computation than sharing rsa->blinding, but
6871 avoids excessive locking; and if an RSA object is not shared
6872 between threads, blinding will still be very fast).
6875 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
6876 ENGINE as defaults for all supported algorithms irrespective of
6877 the 'flags' parameter. 'flags' is now honoured, so applications
6878 should make sure they are passing it correctly.
6881 *) Target "mingw" now allows native Windows code to be generated in
6882 the Cygwin environment as well as with the MinGW compiler.
6885 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
6887 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6888 via timing by performing a MAC computation even if incorrect
6889 block cipher padding has been found. This is a countermeasure
6890 against active attacks where the attacker has to distinguish
6891 between bad padding and a MAC verification error. (CVE-2003-0078)
6893 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6894 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6895 Martin Vuagnoux (EPFL, Ilion)]
6897 *) Make the no-err option work as intended. The intention with no-err
6898 is not to have the whole error stack handling routines removed from
6899 libcrypto, it's only intended to remove all the function name and
6900 reason texts, thereby removing some of the footprint that may not
6901 be interesting if those errors aren't displayed anyway.
6903 NOTE: it's still possible for any application or module to have its
6904 own set of error texts inserted. The routines are there, just not
6905 used by default when no-err is given.
6908 *) Add support for FreeBSD on IA64.
6909 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
6911 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
6912 Kerberos function mit_des_cbc_cksum(). Before this change,
6913 the value returned by DES_cbc_cksum() was like the one from
6914 mit_des_cbc_cksum(), except the bytes were swapped.
6915 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
6917 *) Allow an application to disable the automatic SSL chain building.
6918 Before this a rather primitive chain build was always performed in
6919 ssl3_output_cert_chain(): an application had no way to send the
6920 correct chain if the automatic operation produced an incorrect result.
6922 Now the chain builder is disabled if either:
6924 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
6926 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
6928 The reasoning behind this is that an application would not want the
6929 auto chain building to take place if extra chain certificates are
6930 present and it might also want a means of sending no additional
6931 certificates (for example the chain has two certificates and the
6935 *) Add the possibility to build without the ENGINE framework.
6936 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6938 *) Under Win32 gmtime() can return NULL: check return value in
6939 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
6942 *) DSA routines: under certain error conditions uninitialized BN objects
6943 could be freed. Solution: make sure initialization is performed early
6944 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
6945 Nils Larsch <nla@trustcenter.de> via PR#459)
6948 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
6949 checked on reconnect on the client side, therefore session resumption
6950 could still fail with a "ssl session id is different" error. This
6951 behaviour is masked when SSL_OP_ALL is used due to
6952 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
6953 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6954 followup to PR #377.
6957 *) IA-32 assembler support enhancements: unified ELF targets, support
6958 for SCO/Caldera platforms, fix for Cygwin shared build.
6961 *) Add support for FreeBSD on sparc64. As a consequence, support for
6962 FreeBSD on non-x86 processors is separate from x86 processors on
6963 the config script, much like the NetBSD support.
6964 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
6966 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
6968 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
6971 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
6972 code (06) was taken as the first octet of the session ID and the last
6973 octet was ignored consequently. As a result SSLv2 client side session
6974 caching could not have worked due to the session ID mismatch between
6976 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6980 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
6981 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
6985 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
6986 seems that in spite of existing for more than a year, many application
6987 author have done nothing to provide the necessary callbacks, which
6988 means that this particular engine will not work properly anywhere.
6989 This is a very unfortunate situation which forces us, in the name
6990 of usability, to give the hw_ncipher.c a static lock, which is part
6992 NOTE: This is for the 0.9.7 series ONLY. This hack will never
6993 appear in 0.9.8 or later. We EXPECT application authors to have
6994 dealt properly with this when 0.9.8 is released (unless we actually
6995 make such changes in the libcrypto locking code that changes will
6996 have to be made anyway).
6999 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
7000 octets have been read, EOF or an error occurs. Without this change
7001 some truncated ASN1 structures will not produce an error.
7004 *) Disable Heimdal support, since it hasn't been fully implemented.
7005 Still give the possibility to force the use of Heimdal, but with
7006 warnings and a request that patches get sent to openssl-dev.
7009 *) Add the VC-CE target, introduce the WINCE sysname, and add
7010 INSTALL.WCE and appropriate conditionals to make it build.
7011 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7013 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
7014 cygssl-x.y.z.dll, where x, y and z are the major, minor and
7015 edit numbers of the version.
7016 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7018 *) Introduce safe string copy and catenation functions
7019 (BUF_strlcpy() and BUF_strlcat()).
7020 [Ben Laurie (CHATS) and Richard Levitte]
7022 *) Avoid using fixed-size buffers for one-line DNs.
7023 [Ben Laurie (CHATS)]
7025 *) Add BUF_MEM_grow_clean() to avoid information leakage when
7026 resizing buffers containing secrets, and use where appropriate.
7027 [Ben Laurie (CHATS)]
7029 *) Avoid using fixed size buffers for configuration file location.
7030 [Ben Laurie (CHATS)]
7032 *) Avoid filename truncation for various CA files.
7033 [Ben Laurie (CHATS)]
7035 *) Use sizeof in preference to magic numbers.
7036 [Ben Laurie (CHATS)]
7038 *) Avoid filename truncation in cert requests.
7039 [Ben Laurie (CHATS)]
7041 *) Add assertions to check for (supposedly impossible) buffer
7043 [Ben Laurie (CHATS)]
7045 *) Don't cache truncated DNS entries in the local cache (this could
7046 potentially lead to a spoofing attack).
7047 [Ben Laurie (CHATS)]
7049 *) Fix various buffers to be large enough for hex/decimal
7050 representations in a platform independent manner.
7051 [Ben Laurie (CHATS)]
7053 *) Add CRYPTO_realloc_clean() to avoid information leakage when
7054 resizing buffers containing secrets, and use where appropriate.
7055 [Ben Laurie (CHATS)]
7057 *) Add BIO_indent() to avoid much slightly worrying code to do
7059 [Ben Laurie (CHATS)]
7061 *) Convert sprintf()/BIO_puts() to BIO_printf().
7062 [Ben Laurie (CHATS)]
7064 *) buffer_gets() could terminate with the buffer only half
7066 [Ben Laurie (CHATS)]
7068 *) Add assertions to prevent user-supplied crypto functions from
7069 overflowing internal buffers by having large block sizes, etc.
7070 [Ben Laurie (CHATS)]
7072 *) New OPENSSL_assert() macro (similar to assert(), but enabled
7074 [Ben Laurie (CHATS)]
7076 *) Eliminate unused copy of key in RC4.
7077 [Ben Laurie (CHATS)]
7079 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
7080 [Ben Laurie (CHATS)]
7082 *) Fix off-by-one error in EGD path.
7083 [Ben Laurie (CHATS)]
7085 *) If RANDFILE path is too long, ignore instead of truncating.
7086 [Ben Laurie (CHATS)]
7088 *) Eliminate unused and incorrectly sized X.509 structure
7090 [Ben Laurie (CHATS)]
7092 *) Eliminate unused and dangerous function knumber().
7093 [Ben Laurie (CHATS)]
7095 *) Eliminate unused and dangerous structure, KSSL_ERR.
7096 [Ben Laurie (CHATS)]
7098 *) Protect against overlong session ID context length in an encoded
7099 session object. Since these are local, this does not appear to be
7101 [Ben Laurie (CHATS)]
7103 *) Change from security patch (see 0.9.6e below) that did not affect
7104 the 0.9.6 release series:
7106 Remote buffer overflow in SSL3 protocol - an attacker could
7107 supply an oversized master key in Kerberos-enabled versions.
7109 [Ben Laurie (CHATS)]
7111 *) Change the SSL kerb5 codes to match RFC 2712.
7114 *) Make -nameopt work fully for req and add -reqopt switch.
7115 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
7117 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
7118 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
7120 *) Make sure tests can be performed even if the corresponding algorithms
7121 have been removed entirely. This was also the last step to make
7122 OpenSSL compilable with DJGPP under all reasonable conditions.
7123 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
7125 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
7126 to allow version independent disabling of normally unselected ciphers,
7127 which may be activated as a side-effect of selecting a single cipher.
7129 (E.g., cipher list string "RSA" enables ciphersuites that are left
7130 out of "ALL" because they do not provide symmetric encryption.
7131 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
7132 [Lutz Jaenicke, Bodo Moeller]
7134 *) Add appropriate support for separate platform-dependent build
7135 directories. The recommended way to make a platform-dependent
7136 build directory is the following (tested on Linux), maybe with
7139 # Place yourself outside of the OpenSSL source tree. In
7140 # this example, the environment variable OPENSSL_SOURCE
7141 # is assumed to contain the absolute OpenSSL source directory.
7142 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
7143 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
7144 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
7145 mkdir -p `dirname $F`
7146 ln -s $OPENSSL_SOURCE/$F $F
7149 To be absolutely sure not to disturb the source tree, a "make clean"
7150 is a good thing. If it isn't successful, don't worry about it,
7151 it probably means the source directory is very clean.
7154 *) Make sure any ENGINE control commands make local copies of string
7155 pointers passed to them whenever necessary. Otherwise it is possible
7156 the caller may have overwritten (or deallocated) the original string
7157 data when a later ENGINE operation tries to use the stored values.
7158 [Götz Babin-Ebell <babinebell@trustcenter.de>]
7160 *) Improve diagnostics in file reading and command-line digests.
7161 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
7163 *) Add AES modes CFB and OFB to the object database. Correct an
7164 error in AES-CFB decryption.
7167 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
7168 allows existing EVP_CIPHER_CTX structures to be reused after
7169 calling EVP_*Final(). This behaviour is used by encryption
7170 BIOs and some applications. This has the side effect that
7171 applications must explicitly clean up cipher contexts with
7172 EVP_CIPHER_CTX_cleanup() or they will leak memory.
7175 *) Check the values of dna and dnb in bn_mul_recursive before calling
7176 bn_mul_comba (a non zero value means the a or b arrays do not contain
7177 n2 elements) and fallback to bn_mul_normal if either is not zero.
7180 *) Fix escaping of non-ASCII characters when using the -subj option
7181 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
7184 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
7185 form for "surname", serialNumber has no short form.
7186 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
7187 therefore remove "mail" short name for "internet 7".
7188 The OID for unique identifiers in X509 certificates is
7189 x500UniqueIdentifier, not uniqueIdentifier.
7190 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7193 *) Add an "init" command to the ENGINE config module and auto initialize
7194 ENGINEs. Without any "init" command the ENGINE will be initialized
7195 after all ctrl commands have been executed on it. If init=1 the
7196 ENGINE is initialized at that point (ctrls before that point are run
7197 on the uninitialized ENGINE and after on the initialized one). If
7198 init=0 then the ENGINE will not be initialized at all.
7201 *) Fix the 'app_verify_callback' interface so that the user-defined
7202 argument is actually passed to the callback: In the
7203 SSL_CTX_set_cert_verify_callback() prototype, the callback
7204 declaration has been changed from
7207 int (*cb)(X509_STORE_CTX *,void *);
7208 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7209 i=s->ctx->app_verify_callback(&ctx)
7210 has been changed into
7211 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7213 To update applications using SSL_CTX_set_cert_verify_callback(),
7214 a dummy argument can be added to their callback functions.
7215 [D. K. Smetters <smetters@parc.xerox.com>]
7217 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7218 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7220 *) Add and OPENSSL_LOAD_CONF define which will cause
7221 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7222 This allows older applications to transparently support certain
7223 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7224 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7225 load the config file and OPENSSL_add_all_algorithms_conf() which will
7226 always load it have also been added.
7229 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7230 Adjust NIDs and EVP layer.
7231 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7233 *) Config modules support in openssl utility.
7235 Most commands now load modules from the config file,
7236 though in a few (such as version) this isn't done
7237 because it couldn't be used for anything.
7239 In the case of ca and req the config file used is
7240 the same as the utility itself: that is the -config
7241 command line option can be used to specify an
7245 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7246 use "openssl_conf" if filename is NULL use default openssl config file.
7249 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7250 config section name. Add a new flag to tolerate a missing config file
7251 and move code to CONF_modules_load_file().
7254 *) Support for crypto accelerator cards from Accelerated Encryption
7255 Processing, www.aep.ie. (Use engine 'aep')
7256 The support was copied from 0.9.6c [engine] and adapted/corrected
7257 to work with the new engine framework.
7258 [AEP Inc. and Richard Levitte]
7260 *) Support for SureWare crypto accelerator cards from Baltimore
7261 Technologies. (Use engine 'sureware')
7262 The support was copied from 0.9.6c [engine] and adapted
7263 to work with the new engine framework.
7266 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7267 make the newer ENGINE framework commands for the CHIL engine work.
7268 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7270 *) Make it possible to produce shared libraries on ReliantUNIX.
7271 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7273 *) Add the configuration target debug-linux-ppro.
7274 Make 'openssl rsa' use the general key loading routines
7275 implemented in apps.c, and make those routines able to
7276 handle the key format FORMAT_NETSCAPE and the variant
7278 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7280 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7281 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7283 *) Add -keyform to rsautl, and document -engine.
7284 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7286 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7287 BIO_R_NO_SUCH_FILE error code rather than the generic
7288 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7291 *) Add new functions
7293 ERR_peek_last_error_line
7294 ERR_peek_last_error_line_data.
7295 These are similar to
7298 ERR_peek_error_line_data,
7299 but report on the latest error recorded rather than the first one
7300 still in the error queue.
7301 [Ben Laurie, Bodo Moeller]
7303 *) default_algorithms option in ENGINE config module. This allows things
7305 default_algorithms = ALL
7306 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7309 *) Preliminary ENGINE config module.
7312 *) New experimental application configuration code.
7315 *) Change the AES code to follow the same name structure as all other
7316 symmetric ciphers, and behave the same way. Move everything to
7317 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7318 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7320 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7321 [Ben Laurie and Theo de Raadt]
7323 *) Add option to output public keys in req command.
7324 [Massimiliano Pala madwolf@openca.org]
7326 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7327 (up to about 10% better than before for P-192 and P-224).
7330 *) New functions/macros
7332 SSL_CTX_set_msg_callback(ctx, cb)
7333 SSL_CTX_set_msg_callback_arg(ctx, arg)
7334 SSL_set_msg_callback(ssl, cb)
7335 SSL_set_msg_callback_arg(ssl, arg)
7337 to request calling a callback function
7339 void cb(int write_p, int version, int content_type,
7340 const void *buf, size_t len, SSL *ssl, void *arg)
7342 whenever a protocol message has been completely received
7343 (write_p == 0) or sent (write_p == 1). Here 'version' is the
7344 protocol version according to which the SSL library interprets
7345 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
7346 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
7347 the content type as defined in the SSL 3.0/TLS 1.0 protocol
7348 specification (change_cipher_spec(20), alert(21), handshake(22)).
7349 'buf' and 'len' point to the actual message, 'ssl' to the
7350 SSL object, and 'arg' is the application-defined value set by
7351 SSL[_CTX]_set_msg_callback_arg().
7353 'openssl s_client' and 'openssl s_server' have new '-msg' options
7354 to enable a callback that displays all protocol messages.
7357 *) Change the shared library support so shared libraries are built as
7358 soon as the corresponding static library is finished, and thereby get
7359 openssl and the test programs linked against the shared library.
7360 This still only happens when the keyword "shard" has been given to
7361 the configuration scripts.
7363 NOTE: shared library support is still an experimental thing, and
7364 backward binary compatibility is still not guaranteed.
7365 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
7367 *) Add support for Subject Information Access extension.
7368 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7370 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
7371 additional bytes when new memory had to be allocated, not just
7372 when reusing an existing buffer.
7375 *) New command line and configuration option 'utf8' for the req command.
7376 This allows field values to be specified as UTF8 strings.
7379 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
7380 runs for the former and machine-readable output for the latter.
7383 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
7384 of the e-mail address in the DN (i.e., it will go into a certificate
7385 extension only). The new configuration file option 'email_in_dn = no'
7386 has the same effect.
7387 [Massimiliano Pala madwolf@openca.org]
7389 *) Change all functions with names starting with des_ to be starting
7390 with DES_ instead. Add wrappers that are compatible with libdes,
7391 but are named _ossl_old_des_*. Finally, add macros that map the
7392 des_* symbols to the corresponding _ossl_old_des_* if libdes
7393 compatibility is desired. If OpenSSL 0.9.6c compatibility is
7394 desired, the des_* symbols will be mapped to DES_*, with one
7397 Since we provide two compatibility mappings, the user needs to
7398 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
7399 compatibility is desired. The default (i.e., when that macro
7400 isn't defined) is OpenSSL 0.9.6c compatibility.
7402 There are also macros that enable and disable the support of old
7403 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
7404 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
7405 are defined, the default will apply: to support the old des routines.
7407 In either case, one must include openssl/des.h to get the correct
7408 definitions. Do not try to just include openssl/des_old.h, that
7411 NOTE: This is a major break of an old API into a new one. Software
7412 authors are encouraged to switch to the DES_ style functions. Some
7413 time in the future, des_old.h and the libdes compatibility functions
7414 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
7415 default), and then completely removed.
7418 *) Test for certificates which contain unsupported critical extensions.
7419 If such a certificate is found during a verify operation it is
7420 rejected by default: this behaviour can be overridden by either
7421 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
7422 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
7423 X509_supported_extension() has also been added which returns 1 if a
7424 particular extension is supported.
7427 *) Modify the behaviour of EVP cipher functions in similar way to digests
7428 to retain compatibility with existing code.
7431 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
7432 compatibility with existing code. In particular the 'ctx' parameter does
7433 not have to be to be initialized before the call to EVP_DigestInit() and
7434 it is tidied up after a call to EVP_DigestFinal(). New function
7435 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
7436 EVP_MD_CTX_copy() changed to not require the destination to be
7437 initialized valid and new function EVP_MD_CTX_copy_ex() added which
7438 requires the destination to be valid.
7440 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
7441 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
7444 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
7445 so that complete 'Handshake' protocol structures are kept in memory
7446 instead of overwriting 'msg_type' and 'length' with 'body' data.
7449 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
7450 [Massimo Santin via Richard Levitte]
7452 *) Major restructuring to the underlying ENGINE code. This includes
7453 reduction of linker bloat, separation of pure "ENGINE" manipulation
7454 (initialisation, etc) from functionality dealing with implementations
7455 of specific crypto interfaces. This change also introduces integrated
7456 support for symmetric ciphers and digest implementations - so ENGINEs
7457 can now accelerate these by providing EVP_CIPHER and EVP_MD
7458 implementations of their own. This is detailed in crypto/engine/README
7459 as it couldn't be adequately described here. However, there are a few
7460 API changes worth noting - some RSA, DSA, DH, and RAND functions that
7461 were changed in the original introduction of ENGINE code have now
7462 reverted back - the hooking from this code to ENGINE is now a good
7463 deal more passive and at run-time, operations deal directly with
7464 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
7465 dereferencing through an ENGINE pointer any more. Also, the ENGINE
7466 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
7467 they were not being used by the framework as there is no concept of a
7468 BIGNUM_METHOD and they could not be generalised to the new
7469 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
7470 ENGINE_cpy() has been removed as it cannot be consistently defined in
7474 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
7477 *) Change mkdef.pl to sort symbols that get the same entry number,
7478 and make sure the automatically generated functions ERR_load_*
7479 become part of libeay.num as well.
7482 *) New function SSL_renegotiate_pending(). This returns true once
7483 renegotiation has been requested (either SSL_renegotiate() call
7484 or HelloRequest/ClientHello received from the peer) and becomes
7485 false once a handshake has been completed.
7486 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
7487 sends a HelloRequest, but does not ensure that a handshake takes
7488 place. SSL_renegotiate_pending() is useful for checking if the
7489 client has followed the request.)
7492 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7493 By default, clients may request session resumption even during
7494 renegotiation (if session ID contexts permit); with this option,
7495 session resumption is possible only in the first handshake.
7497 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7498 more bits available for options that should not be part of
7499 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7502 *) Add some demos for certificate and certificate request creation.
7505 *) Make maximum certificate chain size accepted from the peer application
7506 settable (SSL*_get/set_max_cert_list()), as proposed by
7507 "Douglas E. Engert" <deengert@anl.gov>.
7510 *) Add support for shared libraries for Unixware-7
7511 (Boyd Lynn Gerber <gerberb@zenez.com>).
7514 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7515 be done prior to destruction. Use this to unload error strings from
7516 ENGINEs that load their own error strings. NB: This adds two new API
7517 functions to "get" and "set" this destroy handler in an ENGINE.
7520 *) Alter all existing ENGINE implementations (except "openssl" and
7521 "openbsd") to dynamically instantiate their own error strings. This
7522 makes them more flexible to be built both as statically-linked ENGINEs
7523 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
7524 Also, add stub code to each that makes building them as self-contained
7525 shared-libraries easier (see README.ENGINE).
7528 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
7529 implementations into applications that are completely implemented in
7530 self-contained shared-libraries. The "dynamic" ENGINE exposes control
7531 commands that can be used to configure what shared-library to load and
7532 to control aspects of the way it is handled. Also, made an update to
7533 the README.ENGINE file that brings its information up-to-date and
7534 provides some information and instructions on the "dynamic" ENGINE
7535 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
7538 *) Make it possible to unload ranges of ERR strings with a new
7539 "ERR_unload_strings" function.
7542 *) Add a copy() function to EVP_MD.
7545 *) Make EVP_MD routines take a context pointer instead of just the
7546 md_data void pointer.
7549 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
7550 that the digest can only process a single chunk of data
7551 (typically because it is provided by a piece of
7552 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
7553 is only going to provide a single chunk of data, and hence the
7554 framework needn't accumulate the data for oneshot drivers.
7557 *) As with "ERR", make it possible to replace the underlying "ex_data"
7558 functions. This change also alters the storage and management of global
7559 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
7560 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
7561 index counters. The API functions that use this state have been changed
7562 to take a "class_index" rather than pointers to the class's local STACK
7563 and counter, and there is now an API function to dynamically create new
7564 classes. This centralisation allows us to (a) plug a lot of the
7565 thread-safety problems that existed, and (b) makes it possible to clean
7566 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
7567 such data would previously have always leaked in application code and
7568 workarounds were in place to make the memory debugging turn a blind eye
7569 to it. Application code that doesn't use this new function will still
7570 leak as before, but their memory debugging output will announce it now
7571 rather than letting it slide.
7573 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
7574 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
7575 has a return value to indicate success or failure.
7578 *) Make it possible to replace the underlying "ERR" functions such that the
7579 global state (2 LHASH tables and 2 locks) is only used by the "default"
7580 implementation. This change also adds two functions to "get" and "set"
7581 the implementation prior to it being automatically set the first time
7582 any other ERR function takes place. Ie. an application can call "get",
7583 pass the return value to a module it has just loaded, and that module
7584 can call its own "set" function using that value. This means the
7585 module's "ERR" operations will use (and modify) the error state in the
7586 application and not in its own statically linked copy of OpenSSL code.
7589 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
7590 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
7591 the operation, and provides a more encapsulated way for external code
7592 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
7593 to use these functions rather than manually incrementing the counts.
7595 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
7598 *) Add EVP test program.
7601 *) Add symmetric cipher support to ENGINE. Expect the API to change!
7604 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
7605 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
7606 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
7607 These allow a CRL to be built without having to access X509_CRL fields
7608 directly. Modify 'ca' application to use new functions.
7611 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
7612 bug workarounds. Rollback attack detection is a security feature.
7613 The problem will only arise on OpenSSL servers when TLSv1 is not
7614 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7615 Software authors not wanting to support TLSv1 will have special reasons
7616 for their choice and can explicitly enable this option.
7617 [Bodo Moeller, Lutz Jaenicke]
7619 *) Rationalise EVP so it can be extended: don't include a union of
7620 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7621 (similar to those existing for EVP_CIPHER_CTX).
7626 EVP_MD_CTX_init(&md); /* new function call */
7627 EVP_DigestInit(&md, EVP_sha1());
7628 EVP_DigestUpdate(&md, in, len);
7629 EVP_DigestFinal(&md, out, NULL);
7630 EVP_MD_CTX_cleanup(&md); /* new function call */
7634 *) Make DES key schedule conform to the usual scheme, as well as
7635 correcting its structure. This means that calls to DES functions
7636 now have to pass a pointer to a des_key_schedule instead of a
7637 plain des_key_schedule (which was actually always a pointer
7640 des_key_schedule ks;
7642 des_set_key_checked(..., &ks);
7643 des_ncbc_encrypt(..., &ks, ...);
7645 (Note that a later change renames 'des_...' into 'DES_...'.)
7648 *) Initial reduction of linker bloat: the use of some functions, such as
7649 PEM causes large amounts of unused functions to be linked in due to
7650 poor organisation. For example pem_all.c contains every PEM function
7651 which has a knock on effect of linking in large amounts of (unused)
7652 ASN1 code. Grouping together similar functions and splitting unrelated
7653 functions prevents this.
7656 *) Cleanup of EVP macros.
7659 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7660 correct _ecb suffix.
7663 *) Add initial OCSP responder support to ocsp application. The
7664 revocation information is handled using the text based index
7665 use by the ca application. The responder can either handle
7666 requests generated internally, supplied in files (for example
7667 via a CGI script) or using an internal minimal server.
7670 *) Add configuration choices to get zlib compression for TLS.
7673 *) Changes to Kerberos SSL for RFC 2712 compliance:
7674 1. Implemented real KerberosWrapper, instead of just using
7675 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7676 2. Implemented optional authenticator field of KerberosWrapper.
7678 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7679 and authenticator structs; see crypto/krb5/.
7681 Generalized Kerberos calls to support multiple Kerberos libraries.
7682 [Vern Staats <staatsvr@asc.hpc.mil>,
7683 Jeffrey Altman <jaltman@columbia.edu>
7684 via Richard Levitte]
7686 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7687 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7688 values for each of the key sizes rather than having just
7689 parameters (and 'speed' generating keys each time).
7692 *) Speed up EVP routines.
7695 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7696 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7697 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7698 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7700 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7701 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7702 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7705 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7707 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7710 *) Added the OS2-EMX target.
7711 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7713 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7714 to support NCONF routines in extension code. New function CONF_set_nconf()
7715 to allow functions which take an NCONF to also handle the old LHASH
7716 structure: this means that the old CONF compatible routines can be
7717 retained (in particular wrt extensions) without having to duplicate the
7718 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7721 *) Enhance the general user interface with mechanisms for inner control
7722 and with possibilities to have yes/no kind of prompts.
7725 *) Change all calls to low level digest routines in the library and
7726 applications to use EVP. Add missing calls to HMAC_cleanup() and
7727 don't assume HMAC_CTX can be copied using memcpy().
7728 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7730 *) Add the possibility to control engines through control names but with
7731 arbitrary arguments instead of just a string.
7732 Change the key loaders to take a UI_METHOD instead of a callback
7733 function pointer. NOTE: this breaks binary compatibility with earlier
7734 versions of OpenSSL [engine].
7735 Adapt the nCipher code for these new conditions and add a card insertion
7739 *) Enhance the general user interface with mechanisms to better support
7740 dialog box interfaces, application-defined prompts, the possibility
7741 to use defaults (for example default passwords from somewhere else)
7742 and interrupts/cancellations.
7745 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7746 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7749 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7750 tidy up some unnecessarily weird code in 'sk_new()').
7751 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
7753 *) Change the key loading routines for ENGINEs to use the same kind
7754 callback (pem_password_cb) as all other routines that need this
7758 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
7759 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
7760 than this minimum value is recommended.
7763 *) New random seeder for OpenVMS, using the system process statistics
7764 that are easily reachable.
7767 *) Windows apparently can't transparently handle global
7768 variables defined in DLLs. Initialisations such as:
7770 const ASN1_ITEM *it = &ASN1_INTEGER_it;
7772 won't compile. This is used by the any applications that need to
7773 declare their own ASN1 modules. This was fixed by adding the option
7774 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
7775 needed for static libraries under Win32.
7778 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
7779 setting of purpose and trust fields. New X509_STORE trust and
7780 purpose functions and tidy up setting in other SSL functions.
7783 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
7784 structure. These are inherited by X509_STORE_CTX when it is
7785 initialised. This allows various defaults to be set in the
7786 X509_STORE structure (such as flags for CRL checking and custom
7787 purpose or trust settings) for functions which only use X509_STORE_CTX
7788 internally such as S/MIME.
7790 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
7791 trust settings if they are not set in X509_STORE. This allows X509_STORE
7792 purposes and trust (in S/MIME for example) to override any set by default.
7794 Add command line options for CRL checking to smime, s_client and s_server
7798 *) Initial CRL based revocation checking. If the CRL checking flag(s)
7799 are set then the CRL is looked up in the X509_STORE structure and
7800 its validity and signature checked, then if the certificate is found
7801 in the CRL the verify fails with a revoked error.
7803 Various new CRL related callbacks added to X509_STORE_CTX structure.
7805 Command line options added to 'verify' application to support this.
7807 This needs some additional work, such as being able to handle multiple
7808 CRLs with different times, extension based lookup (rather than just
7809 by subject name) and ultimately more complete V2 CRL extension
7813 *) Add a general user interface API (crypto/ui/). This is designed
7814 to replace things like des_read_password and friends (backward
7815 compatibility functions using this new API are provided).
7816 The purpose is to remove prompting functions from the DES code
7817 section as well as provide for prompting through dialog boxes in
7818 a window system and the like.
7821 *) Add "ex_data" support to ENGINE so implementations can add state at a
7822 per-structure level rather than having to store it globally.
7825 *) Make it possible for ENGINE structures to be copied when retrieved by
7826 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
7827 This causes the "original" ENGINE structure to act like a template,
7828 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7829 operational state can be localised to each ENGINE structure, despite the
7830 fact they all share the same "methods". New ENGINE structures returned in
7831 this case have no functional references and the return value is the single
7832 structural reference. This matches the single structural reference returned
7833 by ENGINE_by_id() normally, when it is incremented on the pre-existing
7837 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
7838 needs to match any other type at all we need to manually clear the
7842 *) Changes to the "openssl engine" utility to include;
7843 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
7844 about an ENGINE's available control commands.
7845 - executing control commands from command line arguments using the
7846 '-pre' and '-post' switches. '-post' is only used if '-t' is
7847 specified and the ENGINE is successfully initialised. The syntax for
7848 the individual commands are colon-separated, for example;
7849 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
7852 *) New dynamic control command support for ENGINEs. ENGINEs can now
7853 declare their own commands (numbers), names (strings), descriptions,
7854 and input types for run-time discovery by calling applications. A
7855 subset of these commands are implicitly classed as "executable"
7856 depending on their input type, and only these can be invoked through
7857 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
7858 can be based on user input, config files, etc). The distinction is
7859 that "executable" commands cannot return anything other than a boolean
7860 result and can only support numeric or string input, whereas some
7861 discoverable commands may only be for direct use through
7862 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
7863 pointers, or other custom uses. The "executable" commands are to
7864 support parameterisations of ENGINE behaviour that can be
7865 unambiguously defined by ENGINEs and used consistently across any
7866 OpenSSL-based application. Commands have been added to all the
7867 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
7868 control over shared-library paths without source code alterations.
7871 *) Changed all ENGINE implementations to dynamically allocate their
7872 ENGINEs rather than declaring them statically. Apart from this being
7873 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
7874 this also allows the implementations to compile without using the
7875 internal engine_int.h header.
7878 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
7879 'const' value. Any code that should be able to modify a RAND_METHOD
7880 should already have non-const pointers to it (ie. they should only
7881 modify their own ones).
7884 *) Made a variety of little tweaks to the ENGINE code.
7885 - "atalla" and "ubsec" string definitions were moved from header files
7886 to C code. "nuron" string definitions were placed in variables
7887 rather than hard-coded - allowing parameterisation of these values
7888 later on via ctrl() commands.
7889 - Removed unused "#if 0"'d code.
7890 - Fixed engine list iteration code so it uses ENGINE_free() to release
7891 structural references.
7892 - Constified the RAND_METHOD element of ENGINE structures.
7893 - Constified various get/set functions as appropriate and added
7894 missing functions (including a catch-all ENGINE_cpy that duplicates
7895 all ENGINE values onto a new ENGINE except reference counts/state).
7896 - Removed NULL parameter checks in get/set functions. Setting a method
7897 or function to NULL is a way of cancelling out a previously set
7898 value. Passing a NULL ENGINE parameter is just plain stupid anyway
7899 and doesn't justify the extra error symbols and code.
7900 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
7901 flags from engine_int.h to engine.h.
7902 - Changed prototypes for ENGINE handler functions (init(), finish(),
7903 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
7906 *) Implement binary inversion algorithm for BN_mod_inverse in addition
7907 to the algorithm using long division. The binary algorithm can be
7908 used only if the modulus is odd. On 32-bit systems, it is faster
7909 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
7910 roughly 5-15% for 256-bit moduli), so we use it only for moduli
7911 up to 450 bits. In 64-bit environments, the binary algorithm
7912 appears to be advantageous for much longer moduli; here we use it
7913 for moduli up to 2048 bits.
7916 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
7917 could not support the combine flag in choice fields.
7920 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
7921 extensions from a certificate request to the certificate.
7924 *) Allow multiple 'certopt' and 'nameopt' options to be separated
7925 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
7926 file: this allows the display of the certificate about to be
7927 signed to be customised, to allow certain fields to be included
7928 or excluded and extension details. The old system didn't display
7929 multicharacter strings properly, omitted fields not in the policy
7930 and couldn't display additional details such as extensions.
7933 *) Function EC_POINTs_mul for multiple scalar multiplication
7934 of an arbitrary number of elliptic curve points
7935 \sum scalars[i]*points[i],
7936 optionally including the generator defined for the EC_GROUP:
7937 scalar*generator + \sum scalars[i]*points[i].
7939 EC_POINT_mul is a simple wrapper function for the typical case
7940 that the point list has just one item (besides the optional
7944 *) First EC_METHODs for curves over GF(p):
7946 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
7947 operations and provides various method functions that can also
7948 operate with faster implementations of modular arithmetic.
7950 EC_GFp_mont_method() reuses most functions that are part of
7951 EC_GFp_simple_method, but uses Montgomery arithmetic.
7953 [Bodo Moeller; point addition and point doubling
7954 implementation directly derived from source code provided by
7955 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
7957 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
7958 crypto/ec/ec_lib.c):
7960 Curves are EC_GROUP objects (with an optional group generator)
7961 based on EC_METHODs that are built into the library.
7963 Points are EC_POINT objects based on EC_GROUP objects.
7965 Most of the framework would be able to handle curves over arbitrary
7966 finite fields, but as there are no obvious types for fields other
7967 than GF(p), some functions are limited to that for now.
7970 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
7971 that the file contains a complete HTTP response.
7974 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
7975 change the def and num file printf format specifier from "%-40sXXX"
7976 to "%-39s XXX". The latter will always guarantee a space after the
7977 field while the former will cause them to run together if the field
7978 is 40 of more characters long.
7981 *) Constify the cipher and digest 'method' functions and structures
7982 and modify related functions to take constant EVP_MD and EVP_CIPHER
7986 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
7987 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
7990 *) Modify EVP_Digest*() routines so they now return values. Although the
7991 internal software routines can never fail additional hardware versions
7995 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
7997 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
7998 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
8001 ERR_R_NESTED_ASN1_ERROR
8003 ERR_R_MISSING_ASN1_EOS
8004 were 4 .. 9, conflicting with
8005 ERR_LIB_RSA (= ERR_R_RSA_LIB)
8007 ERR_LIB_PEM (= ERR_R_PEM_LIB).
8008 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
8010 Add new error code 'ERR_R_INTERNAL_ERROR'.
8013 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
8017 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
8018 sets the subject name for a new request or supersedes the
8019 subject name in a given request. Formats that can be parsed are
8020 'CN=Some Name, OU=myOU, C=IT'
8022 'CN=Some Name/OU=myOU/C=IT'.
8024 Add options '-batch' and '-verbose' to 'openssl req'.
8025 [Massimiliano Pala <madwolf@hackmasters.net>]
8027 *) Introduce the possibility to access global variables through
8028 functions on platform were that's the best way to handle exporting
8029 global variables in shared libraries. To enable this functionality,
8030 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
8031 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
8032 is normally done by Configure or something similar).
8034 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
8035 in the source file (foo.c) like this:
8037 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
8038 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
8040 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
8041 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
8043 OPENSSL_DECLARE_GLOBAL(int,foo);
8044 #define foo OPENSSL_GLOBAL_REF(foo)
8045 OPENSSL_DECLARE_GLOBAL(double,bar);
8046 #define bar OPENSSL_GLOBAL_REF(bar)
8048 The #defines are very important, and therefore so is including the
8049 header file everywhere where the defined globals are used.
8051 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
8052 of ASN.1 items, but that structure is a bit different.
8054 The largest change is in util/mkdef.pl which has been enhanced with
8055 better and easier to understand logic to choose which symbols should
8056 go into the Windows .def files as well as a number of fixes and code
8057 cleanup (among others, algorithm keywords are now sorted
8058 lexicographically to avoid constant rewrites).
8061 *) In BN_div() keep a copy of the sign of 'num' before writing the
8062 result to 'rm' because if rm==num the value will be overwritten
8063 and produce the wrong result if 'num' is negative: this caused
8064 problems with BN_mod() and BN_nnmod().
8067 *) Function OCSP_request_verify(). This checks the signature on an
8068 OCSP request and verifies the signer certificate. The signer
8069 certificate is just checked for a generic purpose and OCSP request
8073 *) Add OCSP_check_validity() function to check the validity of OCSP
8074 responses. OCSP responses are prepared in real time and may only
8075 be a few seconds old. Simply checking that the current time lies
8076 between thisUpdate and nextUpdate max reject otherwise valid responses
8077 caused by either OCSP responder or client clock inaccuracy. Instead
8078 we allow thisUpdate and nextUpdate to fall within a certain period of
8079 the current time. The age of the response can also optionally be
8080 checked. Two new options -validity_period and -status_age added to
8084 *) If signature or public key algorithm is unrecognized print out its
8085 OID rather that just UNKNOWN.
8088 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
8089 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
8090 ID to be generated from the issuer certificate alone which can then be
8091 passed to OCSP_id_issuer_cmp().
8094 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
8095 ASN1 modules to export functions returning ASN1_ITEM pointers
8096 instead of the ASN1_ITEM structures themselves. This adds several
8097 new macros which allow the underlying ASN1 function/structure to
8098 be accessed transparently. As a result code should not use ASN1_ITEM
8099 references directly (such as &X509_it) but instead use the relevant
8100 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
8101 use of the new ASN1 code on platforms where exporting structures
8102 is problematical (for example in shared libraries) but exporting
8103 functions returning pointers to structures is not.
8106 *) Add support for overriding the generation of SSL/TLS session IDs.
8107 These callbacks can be registered either in an SSL_CTX or per SSL.
8108 The purpose of this is to allow applications to control, if they wish,
8109 the arbitrary values chosen for use as session IDs, particularly as it
8110 can be useful for session caching in multiple-server environments. A
8111 command-line switch for testing this (and any client code that wishes
8112 to use such a feature) has been added to "s_server".
8113 [Geoff Thorpe, Lutz Jaenicke]
8115 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
8116 of the form '#if defined(...) || defined(...) || ...' and
8117 '#if !defined(...) && !defined(...) && ...'. This also avoids
8118 the growing number of special cases it was previously handling.
8121 *) Make all configuration macros available for application by making
8122 sure they are available in opensslconf.h, by giving them names starting
8123 with "OPENSSL_" to avoid conflicts with other packages and by making
8124 sure e_os2.h will cover all platform-specific cases together with
8126 Additionally, it is now possible to define configuration/platform-
8127 specific names (called "system identities"). In the C code, these
8128 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
8129 macro with the name beginning with "OPENSSL_SYS_", which is determined
8130 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
8134 *) New option -set_serial to 'req' and 'x509' this allows the serial
8135 number to use to be specified on the command line. Previously self
8136 signed certificates were hard coded with serial number 0 and the
8137 CA options of 'x509' had to use a serial number in a file which was
8141 *) New options to 'ca' utility to support V2 CRL entry extensions.
8142 Currently CRL reason, invalidity date and hold instruction are
8143 supported. Add new CRL extensions to V3 code and some new objects.
8146 *) New function EVP_CIPHER_CTX_set_padding() this is used to
8147 disable standard block padding (aka PKCS#5 padding) in the EVP
8148 API, which was previously mandatory. This means that the data is
8149 not padded in any way and so the total length much be a multiple
8150 of the block size, otherwise an error occurs.
8153 *) Initial (incomplete) OCSP SSL support.
8156 *) New function OCSP_parse_url(). This splits up a URL into its host,
8157 port and path components: primarily to parse OCSP URLs. New -url
8158 option to ocsp utility.
8161 *) New nonce behavior. The return value of OCSP_check_nonce() now
8162 reflects the various checks performed. Applications can decide
8163 whether to tolerate certain situations such as an absent nonce
8164 in a response when one was present in a request: the ocsp application
8165 just prints out a warning. New function OCSP_add1_basic_nonce()
8166 this is to allow responders to include a nonce in a response even if
8167 the request is nonce-less.
8170 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
8171 skipped when using openssl x509 multiple times on a single input file,
8172 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
8175 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
8176 set string type: to handle setting ASN1_TIME structures. Fix ca
8177 utility to correctly initialize revocation date of CRLs.
8180 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
8181 the clients preferred ciphersuites and rather use its own preferences.
8182 Should help to work around M$ SGC (Server Gated Cryptography) bug in
8183 Internet Explorer by ensuring unchanged hash method during stepup.
8184 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
8187 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
8188 to aes and add a new 'exist' option to print out symbols that don't
8192 *) Additional options to ocsp utility to allow flags to be set and
8193 additional certificates supplied.
8196 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8197 OCSP client a number of certificate to only verify the response
8201 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8202 handle the new API. Currently only ECB, CBC modes supported. Add new
8205 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8206 Encryption Standard (AES) Ciphersuites for Transport Layer
8207 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8208 not enabled by default and were not part of the "ALL" ciphersuite
8209 alias because they were not yet official; they could be
8210 explicitly requested by specifying the "AESdraft" ciphersuite
8211 group alias. In the final release of OpenSSL 0.9.7, the group
8212 alias is called "AES" and is part of "ALL".)
8213 [Ben Laurie, Steve Henson, Bodo Moeller]
8215 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8216 request to response.
8219 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8220 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8221 extract information from a certificate request. OCSP_response_create()
8222 creates a response and optionally adds a basic response structure.
8223 OCSP_basic_add1_status() adds a complete single response to a basic
8224 response and returns the OCSP_SINGLERESP structure just added (to allow
8225 extensions to be included for example). OCSP_basic_add1_cert() adds a
8226 certificate to a basic response and OCSP_basic_sign() signs a basic
8227 response with various flags. New helper functions ASN1_TIME_check()
8228 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8229 (converts ASN1_TIME to GeneralizedTime).
8232 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8233 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8234 structure from a certificate. X509_pubkey_digest() digests the public_key
8235 contents: this is used in various key identifiers.
8238 *) Make sk_sort() tolerate a NULL argument.
8239 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8241 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8242 passed by the function are trusted implicitly. If any of them signed the
8243 response then it is assumed to be valid and is not verified.
8246 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8247 to data. This was previously part of the PKCS7 ASN1 code. This
8248 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8249 [Steve Henson, reported by Kenneth R. Robinette
8250 <support@securenetterm.com>]
8252 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8253 routines: without these tracing memory leaks is very painful.
8254 Fix leaks in PKCS12 and PKCS7 routines.
8257 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8258 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8259 effectively meant GeneralizedTime would never be used. Now it
8260 is initialised to -1 but X509_time_adj() now has to check the value
8261 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8262 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8263 [Steve Henson, reported by Kenneth R. Robinette
8264 <support@securenetterm.com>]
8266 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8267 result in a zero length in the ASN1_INTEGER structure which was
8268 not consistent with the structure when d2i_ASN1_INTEGER() was used
8269 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8270 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8271 where it did not print out a minus for negative ASN1_INTEGER.
8274 *) Add summary printout to ocsp utility. The various functions which
8275 convert status values to strings have been renamed to:
8276 OCSP_response_status_str(), OCSP_cert_status_str() and
8277 OCSP_crl_reason_str() and are no longer static. New options
8278 to verify nonce values and to disable verification. OCSP response
8279 printout format cleaned up.
8282 *) Add additional OCSP certificate checks. These are those specified
8283 in RFC2560. This consists of two separate checks: the CA of the
8284 certificate being checked must either be the OCSP signer certificate
8285 or the issuer of the OCSP signer certificate. In the latter case the
8286 OCSP signer certificate must contain the OCSP signing extended key
8287 usage. This check is performed by attempting to match the OCSP
8288 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8289 in the OCSP_CERTID structures of the response.
8292 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8293 and related routines. This uses the standard OpenSSL certificate
8294 verify routines to perform initial checks (just CA validity) and
8295 to obtain the certificate chain. Then additional checks will be
8296 performed on the chain. Currently the root CA is checked to see
8297 if it is explicitly trusted for OCSP signing. This is used to set
8298 a root CA as a global signing root: that is any certificate that
8299 chains to that CA is an acceptable OCSP signing certificate.
8302 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8303 extensions from a separate configuration file.
8304 As when reading extensions from the main configuration file,
8305 the '-extensions ...' option may be used for specifying the
8307 [Massimiliano Pala <madwolf@comune.modena.it>]
8309 *) New OCSP utility. Allows OCSP requests to be generated or
8310 read. The request can be sent to a responder and the output
8311 parsed, outputed or printed in text form. Not complete yet:
8312 still needs to check the OCSP response validity.
8315 *) New subcommands for 'openssl ca':
8316 'openssl ca -status <serial>' prints the status of the cert with
8317 the given serial number (according to the index file).
8318 'openssl ca -updatedb' updates the expiry status of certificates
8320 [Massimiliano Pala <madwolf@comune.modena.it>]
8322 *) New '-newreq-nodes' command option to CA.pl. This is like
8323 '-newreq', but calls 'openssl req' with the '-nodes' option
8324 so that the resulting key is not encrypted.
8325 [Damien Miller <djm@mindrot.org>]
8327 *) New configuration for the GNU Hurd.
8328 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8330 *) Initial code to implement OCSP basic response verify. This
8331 is currently incomplete. Currently just finds the signer's
8332 certificate and verifies the signature on the response.
8335 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
8336 value of OPENSSLDIR. This is available via the new '-d' option
8337 to 'openssl version', and is also included in 'openssl version -a'.
8340 *) Allowing defining memory allocation callbacks that will be given
8341 file name and line number information in additional arguments
8342 (a const char* and an int). The basic functionality remains, as
8343 well as the original possibility to just replace malloc(),
8344 realloc() and free() by functions that do not know about these
8345 additional arguments. To register and find out the current
8346 settings for extended allocation functions, the following
8347 functions are provided:
8349 CRYPTO_set_mem_ex_functions
8350 CRYPTO_set_locked_mem_ex_functions
8351 CRYPTO_get_mem_ex_functions
8352 CRYPTO_get_locked_mem_ex_functions
8354 These work the same way as CRYPTO_set_mem_functions and friends.
8355 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
8356 extended allocation function is enabled.
8357 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
8358 a conventional allocation function is enabled.
8359 [Richard Levitte, Bodo Moeller]
8361 *) Finish off removing the remaining LHASH function pointer casts.
8362 There should no longer be any prototype-casting required when using
8363 the LHASH abstraction, and any casts that remain are "bugs". See
8364 the callback types and macros at the head of lhash.h for details
8365 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
8368 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
8369 If /dev/[u]random devices are not available or do not return enough
8370 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
8372 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
8373 /etc/entropy will be queried once each in this sequence, querying stops
8374 when enough entropy was collected without querying more sockets.
8377 *) Change the Unix RAND_poll() variant to be able to poll several
8378 random devices, as specified by DEVRANDOM, until a sufficient amount
8379 of data has been collected. We spend at most 10 ms on each file
8380 (select timeout) and read in non-blocking mode. DEVRANDOM now
8381 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
8382 (previously it was just the string "/dev/urandom"), so on typical
8383 platforms the 10 ms delay will never occur.
8384 Also separate out the Unix variant to its own file, rand_unix.c.
8385 For VMS, there's a currently-empty rand_vms.c.
8388 *) Move OCSP client related routines to ocsp_cl.c. These
8389 provide utility functions which an application needing
8390 to issue a request to an OCSP responder and analyse the
8391 response will typically need: as opposed to those which an
8392 OCSP responder itself would need which will be added later.
8394 OCSP_request_sign() signs an OCSP request with an API similar
8395 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
8396 response. OCSP_response_get1_basic() extracts basic response
8397 from response. OCSP_resp_find_status(): finds and extracts status
8398 information from an OCSP_CERTID structure (which will be created
8399 when the request structure is built). These are built from lower
8400 level functions which work on OCSP_SINGLERESP structures but
8401 won't normally be used unless the application wishes to examine
8402 extensions in the OCSP response for example.
8404 Replace nonce routines with a pair of functions.
8405 OCSP_request_add1_nonce() adds a nonce value and optionally
8406 generates a random value. OCSP_check_nonce() checks the
8407 validity of the nonce in an OCSP response.
8410 *) Change function OCSP_request_add() to OCSP_request_add0_id().
8411 This doesn't copy the supplied OCSP_CERTID and avoids the
8412 need to free up the newly created id. Change return type
8413 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
8414 This can then be used to add extensions to the request.
8415 Deleted OCSP_request_new(), since most of its functionality
8416 is now in OCSP_REQUEST_new() (and the case insensitive name
8417 clash) apart from the ability to set the request name which
8418 will be added elsewhere.
8421 *) Update OCSP API. Remove obsolete extensions argument from
8422 various functions. Extensions are now handled using the new
8423 OCSP extension code. New simple OCSP HTTP function which
8424 can be used to send requests and parse the response.
8427 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
8428 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
8429 uses the special reorder version of SET OF to sort the attributes
8430 and reorder them to match the encoded order. This resolves a long
8431 standing problem: a verify on a PKCS7 structure just after signing
8432 it used to fail because the attribute order did not match the
8433 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
8434 it uses the received order. This is necessary to tolerate some broken
8435 software that does not order SET OF. This is handled by encoding
8436 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
8437 to produce the required SET OF.
8440 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
8441 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8442 files to get correct declarations of the ASN.1 item variables.
8445 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
8446 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
8447 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
8448 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
8449 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
8450 ASN1_ITEM and no wrapper functions.
8453 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
8454 replace the old function pointer based I/O routines. Change most of
8455 the *_d2i_bio() and *_d2i_fp() functions to use these.
8458 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
8459 lines, recognize more "algorithms" that can be deselected, and make
8460 it complain about algorithm deselection that isn't recognised.
8463 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
8464 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
8465 to use new functions. Add NO_ASN1_OLD which can be set to remove
8466 some old style ASN1 functions: this can be used to determine if old
8467 code will still work when these eventually go away.
8470 *) New extension functions for OCSP structures, these follow the
8471 same conventions as certificates and CRLs.
8474 *) New function X509V3_add1_i2d(). This automatically encodes and
8475 adds an extension. Its behaviour can be customised with various
8476 flags to append, replace or delete. Various wrappers added for
8477 certificates and CRLs.
8480 *) Fix to avoid calling the underlying ASN1 print routine when
8481 an extension cannot be parsed. Correct a typo in the
8482 OCSP_SERVICELOC extension. Tidy up print OCSP format.
8485 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
8486 entries for variables.
8489 *) Add functionality to apps/openssl.c for detecting locking
8490 problems: As the program is single-threaded, all we have
8491 to do is register a locking callback using an array for
8492 storing which locks are currently held by the program.
8495 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8496 SSL_get_ex_data_X509_STORE_idx(), which is used in
8497 ssl_verify_cert_chain() and thus can be called at any time
8498 during TLS/SSL handshakes so that thread-safety is essential.
8499 Unfortunately, the ex_data design is not at all suited
8500 for multi-threaded use, so it probably should be abolished.
8503 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8504 [Broadcom, tweaked and integrated by Geoff Thorpe]
8506 *) Move common extension printing code to new function
8507 X509V3_print_extensions(). Reorganise OCSP print routines and
8508 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8511 *) New function X509_signature_print() to remove duplication in some
8515 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8516 set (this was treated exactly the same as SET OF previously). This
8517 is used to reorder the STACK representing the structure to match the
8518 encoding. This will be used to get round a problem where a PKCS7
8519 structure which was signed could not be verified because the STACK
8520 order did not reflect the encoded order.
8523 *) Reimplement the OCSP ASN1 module using the new code.
8526 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
8527 for its ASN1 operations. The old style function pointers still exist
8528 for now but they will eventually go away.
8531 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
8532 completely replaces the old ASN1 functionality with a table driven
8533 encoder and decoder which interprets an ASN1_ITEM structure describing
8534 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
8535 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
8536 has also been converted to the new form.
8539 *) Change BN_mod_exp_recp so that negative moduli are tolerated
8540 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
8541 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
8542 for negative moduli.
8545 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
8546 of not touching the result's sign bit.
8549 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
8553 *) Changed the LHASH code to use prototypes for callbacks, and created
8554 macros to declare and implement thin (optionally static) functions
8555 that provide type-safety and avoid function pointer casting for the
8556 type-specific callbacks.
8559 *) Added Kerberos Cipher Suites to be used with TLS, as written in
8561 [Veers Staats <staatsvr@asc.hpc.mil>,
8562 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
8564 *) Reformat the FAQ so the different questions and answers can be divided
8565 in sections depending on the subject.
8568 *) Have the zlib compression code load ZLIB.DLL dynamically under
8572 *) New function BN_mod_sqrt for computing square roots modulo a prime
8573 (using the probabilistic Tonelli-Shanks algorithm unless
8574 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
8575 be handled deterministically).
8576 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8578 *) Make BN_mod_inverse faster by explicitly handling small quotients
8579 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
8580 512 bits], about 30% for larger ones [1024 or 2048 bits].)
8583 *) New function BN_kronecker.
8586 *) Fix BN_gcd so that it works on negative inputs; the result is
8587 positive unless both parameters are zero.
8588 Previously something reasonably close to an infinite loop was
8589 possible because numbers could be growing instead of shrinking
8590 in the implementation of Euclid's algorithm.
8593 *) Fix BN_is_word() and BN_is_one() macros to take into account the
8594 sign of the number in question.
8596 Fix BN_is_word(a,w) to work correctly for w == 0.
8598 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
8599 because its test if the absolute value of 'a' equals 'w'.
8600 Note that BN_abs_is_word does *not* handle w == 0 reliably;
8601 it exists mostly for use in the implementations of BN_is_zero(),
8602 BN_is_one(), and BN_is_word().
8605 *) New function BN_swap.
8608 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
8609 the exponentiation functions are more likely to produce reasonable
8610 results on negative inputs.
8613 *) Change BN_mod_mul so that the result is always non-negative.
8614 Previously, it could be negative if one of the factors was negative;
8615 I don't think anyone really wanted that behaviour.
8618 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8619 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8620 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8621 and add new functions:
8630 BN_mod_lshift1_quick
8634 These functions always generate non-negative results.
8636 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8637 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8639 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8640 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8641 be reduced modulo m.
8642 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8645 The following entry accidentally appeared in the CHANGES file
8646 distributed with OpenSSL 0.9.7. The modifications described in
8647 it do *not* apply to OpenSSL 0.9.7.
8649 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8650 was actually never needed) and in BN_mul(). The removal in BN_mul()
8651 required a small change in bn_mul_part_recursive() and the addition
8652 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8653 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8654 bn_sub_words() and bn_add_words() except they take arrays with
8659 *) In 'openssl passwd', verify passwords read from the terminal
8660 unless the '-salt' option is used (which usually means that
8661 verification would just waste user's time since the resulting
8662 hash is going to be compared with some given password hash)
8663 or the new '-noverify' option is used.
8665 This is an incompatible change, but it does not affect
8666 non-interactive use of 'openssl passwd' (passwords on the command
8667 line, '-stdin' option, '-in ...' option) and thus should not
8671 *) Remove all references to RSAref, since there's no more need for it.
8674 *) Make DSO load along a path given through an environment variable
8675 (SHLIB_PATH) with shl_load().
8678 *) Constify the ENGINE code as a result of BIGNUM constification.
8679 Also constify the RSA code and most things related to it. In a
8680 few places, most notable in the depth of the ASN.1 code, ugly
8681 casts back to non-const were required (to be solved at a later
8685 *) Make it so the openssl application has all engines loaded by default.
8688 *) Constify the BIGNUM routines a little more.
8691 *) Add the following functions:
8693 ENGINE_load_cswift()
8695 ENGINE_load_atalla()
8697 ENGINE_load_builtin_engines()
8699 That way, an application can itself choose if external engines that
8700 are built-in in OpenSSL shall ever be used or not. The benefit is
8701 that applications won't have to be linked with libdl or other dso
8702 libraries unless it's really needed.
8704 Changed 'openssl engine' to load all engines on demand.
8705 Changed the engine header files to avoid the duplication of some
8706 declarations (they differed!).
8709 *) 'openssl engine' can now list capabilities.
8712 *) Better error reporting in 'openssl engine'.
8715 *) Never call load_dh_param(NULL) in s_server.
8718 *) Add engine application. It can currently list engines by name and
8719 identity, and test if they are actually available.
8722 *) Improve RPM specification file by forcing symbolic linking and making
8723 sure the installed documentation is also owned by root.root.
8724 [Damien Miller <djm@mindrot.org>]
8726 *) Give the OpenSSL applications more possibilities to make use of
8727 keys (public as well as private) handled by engines.
8730 *) Add OCSP code that comes from CertCo.
8733 *) Add VMS support for the Rijndael code.
8736 *) Added untested support for Nuron crypto accelerator.
8739 *) Add support for external cryptographic devices. This code was
8740 previously distributed separately as the "engine" branch.
8741 [Geoff Thorpe, Richard Levitte]
8743 *) Rework the filename-translation in the DSO code. It is now possible to
8744 have far greater control over how a "name" is turned into a filename
8745 depending on the operating environment and any oddities about the
8746 different shared library filenames on each system.
8749 *) Support threads on FreeBSD-elf in Configure.
8752 *) Fix for SHA1 assembly problem with MASM: it produces
8753 warnings about corrupt line number information when assembling
8754 with debugging information. This is caused by the overlapping
8756 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
8759 NCONF_get_number() has no error checking at all. As a replacement,
8760 NCONF_get_number_e() is defined (_e for "error checking") and is
8761 promoted strongly. The old NCONF_get_number is kept around for
8762 binary backward compatibility.
8763 Make it possible for methods to load from something other than a BIO,
8764 by providing a function pointer that is given a name instead of a BIO.
8765 For example, this could be used to load configuration data from an
8769 *) Fix for non blocking accept BIOs. Added new I/O special reason
8770 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
8771 with non blocking I/O was not possible because no retry code was
8772 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
8776 *) Added the beginnings of Rijndael support.
8779 *) Fix for bug in DirectoryString mask setting. Add support for
8780 X509_NAME_print_ex() in 'req' and X509_print_ex() function
8781 to allow certificate printing to more controllable, additional
8782 'certopt' option to 'x509' to allow new printing options to be
8786 *) Clean old EAY MD5 hack from e_os.h.
8789 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
8791 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
8792 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8793 [Joe Orton, Steve Henson]
8795 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
8797 *) Fix additional bug revealed by the NISCC test suite:
8799 Stop bug triggering large recursion when presented with
8800 certain ASN.1 tags (CVE-2003-0851)
8803 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
8805 *) Fix various bugs revealed by running the NISCC test suite:
8807 Stop out of bounds reads in the ASN1 code when presented with
8808 invalid tags (CVE-2003-0543 and CVE-2003-0544).
8810 If verify callback ignores invalid public key errors don't try to check
8811 certificate signature with the NULL public key.
8815 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
8816 if the server requested one: as stated in TLS 1.0 and SSL 3.0
8820 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
8821 extra data after the compression methods not only for TLS 1.0
8822 but also for SSL 3.0 (as required by the specification).
8823 [Bodo Moeller; problem pointed out by Matthias Loepfe]
8825 *) Change X509_certificate_type() to mark the key as exported/exportable
8826 when it's 512 *bits* long, not 512 bytes.
8829 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8831 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8832 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8833 a protocol version number mismatch like a decryption error
8834 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8837 *) Turn on RSA blinding by default in the default implementation
8838 to avoid a timing attack. Applications that don't want it can call
8839 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8840 They would be ill-advised to do so in most cases.
8841 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
8843 *) Change RSA blinding code so that it works when the PRNG is not
8844 seeded (in this case, the secret RSA exponent is abused as
8845 an unpredictable seed -- if it is not unpredictable, there
8846 is no point in blinding anyway). Make RSA blinding thread-safe
8847 by remembering the creator's thread ID in rsa->blinding and
8848 having all other threads use local one-time blinding factors
8849 (this requires more computation than sharing rsa->blinding, but
8850 avoids excessive locking; and if an RSA object is not shared
8851 between threads, blinding will still be very fast).
8854 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
8856 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8857 via timing by performing a MAC computation even if incorrect
8858 block cipher padding has been found. This is a countermeasure
8859 against active attacks where the attacker has to distinguish
8860 between bad padding and a MAC verification error. (CVE-2003-0078)
8862 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8863 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8864 Martin Vuagnoux (EPFL, Ilion)]
8866 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
8868 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
8869 memory from its contents. This is done with a counter that will
8870 place alternating values in each byte. This can be used to solve
8871 two issues: 1) the removal of calls to memset() by highly optimizing
8872 compilers, and 2) cleansing with other values than 0, since those can
8873 be read through on certain media, for example a swap space on disk.
8876 *) Bugfix: client side session caching did not work with external caching,
8877 because the session->cipher setting was not restored when reloading
8878 from the external cache. This problem was masked, when
8879 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
8880 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
8883 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
8884 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
8885 [Zeev Lieber <zeev-l@yahoo.com>]
8887 *) Undo an undocumented change introduced in 0.9.6e which caused
8888 repeated calls to OpenSSL_add_all_ciphers() and
8889 OpenSSL_add_all_digests() to be ignored, even after calling
8893 *) Change the default configuration reader to deal with last line not
8894 being properly terminated.
8897 *) Change X509_NAME_cmp() so it applies the special rules on handling
8898 DN values that are of type PrintableString, as well as RDNs of type
8899 emailAddress where the value has the type ia5String.
8900 [stefank@valicert.com via Richard Levitte]
8902 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
8903 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
8904 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
8905 the bitwise-OR of the two for use by the majority of applications
8906 wanting this behaviour, and update the docs. The documented
8907 behaviour and actual behaviour were inconsistent and had been
8908 changing anyway, so this is more a bug-fix than a behavioural
8910 [Geoff Thorpe, diagnosed by Nadav Har'El]
8912 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
8913 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
8916 *) Fix initialization code race conditions in
8917 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
8918 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
8919 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
8920 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
8921 ssl2_get_cipher_by_char(),
8922 ssl3_get_cipher_by_char().
8923 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
8925 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
8926 the cached sessions are flushed, as the remove_cb() might use ex_data
8927 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
8928 (see [openssl.org #212]).
8929 [Geoff Thorpe, Lutz Jaenicke]
8931 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
8932 length, instead of the encoding length to d2i_ASN1_OBJECT.
8935 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
8937 *) [In 0.9.6g-engine release:]
8938 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
8939 [Lynn Gazis <lgazis@rainbow.com>]
8941 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
8943 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
8944 and get fix the header length calculation.
8945 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
8946 Alon Kantor <alonk@checkpoint.com> (and others),
8949 *) Use proper error handling instead of 'assertions' in buffer
8950 overflow checks added in 0.9.6e. This prevents DoS (the
8951 assertions could call abort()).
8952 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
8954 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
8956 *) Add various sanity checks to asn1_get_length() to reject
8957 the ASN1 length bytes if they exceed sizeof(long), will appear
8958 negative or the content length exceeds the length of the
8960 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8962 *) Fix cipher selection routines: ciphers without encryption had no flags
8963 for the cipher strength set and where therefore not handled correctly
8964 by the selection routines (PR #130).
8967 *) Fix EVP_dsa_sha macro.
8971 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
8972 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
8973 that was added in OpenSSL 0.9.6d.
8975 As the countermeasure turned out to be incompatible with some
8976 broken SSL implementations, the new option is part of SSL_OP_ALL.
8977 SSL_OP_ALL is usually employed when compatibility with weird SSL
8978 implementations is desired (e.g. '-bugs' option to 's_client' and
8979 's_server'), so the new option is automatically set in many
8983 *) Changes in security patch:
8985 Changes marked "(CHATS)" were sponsored by the Defense Advanced
8986 Research Projects Agency (DARPA) and Air Force Research Laboratory,
8987 Air Force Materiel Command, USAF, under agreement number
8990 *) Add various sanity checks to asn1_get_length() to reject
8991 the ASN1 length bytes if they exceed sizeof(long), will appear
8992 negative or the content length exceeds the length of the
8993 supplied buffer. (CVE-2002-0659)
8994 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8996 *) Assertions for various potential buffer overflows, not known to
8998 [Ben Laurie (CHATS)]
9000 *) Various temporary buffers to hold ASCII versions of integers were
9001 too small for 64 bit platforms. (CVE-2002-0655)
9002 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
9004 *) Remote buffer overflow in SSL3 protocol - an attacker could
9005 supply an oversized session ID to a client. (CVE-2002-0656)
9006 [Ben Laurie (CHATS)]
9008 *) Remote buffer overflow in SSL2 protocol - an attacker could
9009 supply an oversized client master key. (CVE-2002-0656)
9010 [Ben Laurie (CHATS)]
9012 Changes between 0.9.6c and 0.9.6d [9 May 2002]
9014 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
9015 encoded as NULL) with id-dsa-with-sha1.
9016 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
9018 *) Check various X509_...() return values in apps/req.c.
9019 [Nils Larsch <nla@trustcenter.de>]
9021 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
9022 an end-of-file condition would erroneously be flagged, when the CRLF
9023 was just at the end of a processed block. The bug was discovered when
9024 processing data through a buffering memory BIO handing the data to a
9025 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
9026 <ptsekov@syntrex.com> and Nedelcho Stanev.
9029 *) Implement a countermeasure against a vulnerability recently found
9030 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
9031 before application data chunks to avoid the use of known IVs
9032 with data potentially chosen by the attacker.
9035 *) Fix length checks in ssl3_get_client_hello().
9038 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
9039 to prevent ssl3_read_internal() from incorrectly assuming that
9040 ssl3_read_bytes() found application data while handshake
9041 processing was enabled when in fact s->s3->in_read_app_data was
9042 merely automatically cleared during the initial handshake.
9043 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
9045 *) Fix object definitions for Private and Enterprise: they were not
9046 recognized in their shortname (=lowercase) representation. Extend
9047 obj_dat.pl to issue an error when using undefined keywords instead
9048 of silently ignoring the problem (Svenning Sorensen
9049 <sss@sss.dnsalias.net>).
9052 *) Fix DH_generate_parameters() so that it works for 'non-standard'
9053 generators, i.e. generators other than 2 and 5. (Previously, the
9054 code did not properly initialise the 'add' and 'rem' values to
9055 BN_generate_prime().)
9057 In the new general case, we do not insist that 'generator' is
9058 actually a primitive root: This requirement is rather pointless;
9059 a generator of the order-q subgroup is just as good, if not
9063 *) Map new X509 verification errors to alerts. Discovered and submitted by
9064 Tom Wu <tom@arcot.com>.
9067 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
9068 returning non-zero before the data has been completely received
9069 when using non-blocking I/O.
9070 [Bodo Moeller; problem pointed out by John Hughes]
9072 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
9073 [Ben Laurie, Lutz Jaenicke]
9075 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
9076 Yoram Zahavi <YoramZ@gilian.com>).
9079 *) Add information about CygWin 1.3 and on, and preserve proper
9080 configuration for the versions before that.
9081 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
9083 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
9084 check whether we deal with a copy of a session and do not delete from
9085 the cache in this case. Problem reported by "Izhar Shoshani Levi"
9086 <izhar@checkpoint.com>.
9089 *) Do not store session data into the internal session cache, if it
9090 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
9091 flag is set). Proposed by Aslam <aslam@funk.com>.
9094 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
9098 *) [In 0.9.6d-engine release:]
9099 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
9100 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
9102 *) Add the configuration target linux-s390x.
9103 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
9105 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
9106 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
9107 variable as an indication that a ClientHello message has been
9108 received. As the flag value will be lost between multiple
9109 invocations of ssl3_accept when using non-blocking I/O, the
9110 function may not be aware that a handshake has actually taken
9111 place, thus preventing a new session from being added to the
9114 To avoid this problem, we now set s->new_session to 2 instead of
9115 using a local variable.
9116 [Lutz Jaenicke, Bodo Moeller]
9118 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
9119 if the SSL_R_LENGTH_MISMATCH error is detected.
9120 [Geoff Thorpe, Bodo Moeller]
9122 *) New 'shared_ldflag' column in Configure platform table.
9125 *) Fix EVP_CIPHER_mode macro.
9126 ["Dan S. Camper" <dan@bti.net>]
9128 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
9129 type, we must throw them away by setting rr->length to 0.
9130 [D P Chang <dpc@qualys.com>]
9132 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
9134 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
9135 <Dominikus.Scherkl@biodata.com>. (The previous implementation
9136 worked incorrectly for those cases where range = 10..._2 and
9137 3*range is two bits longer than range.)
9140 *) Only add signing time to PKCS7 structures if it is not already
9144 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
9145 OBJ_ld_ce should be OBJ_id_ce.
9146 Also some ip-pda OIDs in crypto/objects/objects.txt were
9147 incorrect (cf. RFC 3039).
9148 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
9150 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
9151 returns early because it has nothing to do.
9152 [Andy Schneider <andy.schneider@bjss.co.uk>]
9154 *) [In 0.9.6c-engine release:]
9155 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
9156 [Andy Schneider <andy.schneider@bjss.co.uk>]
9158 *) [In 0.9.6c-engine release:]
9159 Add support for Cryptographic Appliance's keyserver technology.
9160 (Use engine 'keyclient')
9161 [Cryptographic Appliances and Geoff Thorpe]
9163 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
9164 is called via tools/c89.sh because arguments have to be
9165 rearranged (all '-L' options must appear before the first object
9167 [Richard Shapiro <rshapiro@abinitio.com>]
9169 *) [In 0.9.6c-engine release:]
9170 Add support for Broadcom crypto accelerator cards, backported
9172 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
9174 *) [In 0.9.6c-engine release:]
9175 Add support for SureWare crypto accelerator cards from
9176 Baltimore Technologies. (Use engine 'sureware')
9177 [Baltimore Technologies and Mark Cox]
9179 *) [In 0.9.6c-engine release:]
9180 Add support for crypto accelerator cards from Accelerated
9181 Encryption Processing, www.aep.ie. (Use engine 'aep')
9182 [AEP Inc. and Mark Cox]
9184 *) Add a configuration entry for gcc on UnixWare.
9185 [Gary Benson <gbenson@redhat.com>]
9187 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
9188 messages are stored in a single piece (fixed-length part and
9189 variable-length part combined) and fix various bugs found on the way.
9192 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9193 instead. BIO_gethostbyname() does not know what timeouts are
9194 appropriate, so entries would stay in cache even when they have
9196 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9198 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9199 faced with a pathologically small ClientHello fragment that does
9200 not contain client_version: Instead of aborting with an error,
9201 simply choose the highest available protocol version (i.e.,
9202 TLS 1.0 unless it is disabled). In practice, ClientHello
9203 messages are never sent like this, but this change gives us
9204 strictly correct behaviour at least for TLS.
9207 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9208 never resets s->method to s->ctx->method when called from within
9209 one of the SSL handshake functions.
9210 [Bodo Moeller; problem pointed out by Niko Baric]
9212 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9213 (sent using the client's version number) if client_version is
9214 smaller than the protocol version in use. Also change
9215 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9216 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9217 the client will at least see that alert.
9220 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9224 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9225 client receives HelloRequest while in a handshake.
9226 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9228 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9229 should end in 'break', not 'goto end' which circumvents various
9230 cleanups done in state SSL_ST_OK. But session related stuff
9231 must be disabled for SSL_ST_OK in the case that we just sent a
9234 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9235 before just sending a HelloRequest.
9236 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9238 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9239 reveal whether illegal block cipher padding was found or a MAC
9240 verification error occurred. (Neither SSLerr() codes nor alerts
9241 are directly visible to potential attackers, but the information
9242 may leak via logfiles.)
9244 Similar changes are not required for the SSL 2.0 implementation
9245 because the number of padding bytes is sent in clear for SSL 2.0,
9246 and the extra bytes are just ignored. However ssl/s2_pkt.c
9247 failed to verify that the purported number of padding bytes is in
9251 *) Add OpenUNIX-8 support including shared libraries
9252 (Boyd Lynn Gerber <gerberb@zenez.com>).
9255 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9256 'wristwatch attack' using huge encoding parameters (cf.
9257 James H. Manger's CRYPTO 2001 paper). Note that the
9258 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9259 encoding parameters and hence was not vulnerable.
9262 *) BN_sqr() bug fix.
9263 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9265 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9266 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9267 followed by modular reduction.
9268 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9270 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9271 equivalent based on BN_pseudo_rand() instead of BN_rand().
9274 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9275 This function was broken, as the check for a new client hello message
9276 to handle SGC did not allow these large messages.
9277 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9280 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9283 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9284 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9287 *) Rework the configuration and shared library support for Tru64 Unix.
9288 The configuration part makes use of modern compiler features and
9289 still retains old compiler behavior for those that run older versions
9290 of the OS. The shared library support part includes a variant that
9291 uses the RPATH feature, and is available through the special
9292 configuration target "alpha-cc-rpath", which will never be selected
9294 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9296 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9297 with the same message size as in ssl3_get_certificate_request().
9298 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9299 messages might inadvertently be reject as too long.
9300 [Petr Lampa <lampa@fee.vutbr.cz>]
9302 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9305 *) Modified SSL library such that the verify_callback that has been set
9306 specificly for an SSL object with SSL_set_verify() is actually being
9307 used. Before the change, a verify_callback set with this function was
9308 ignored and the verify_callback() set in the SSL_CTX at the time of
9309 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9310 to allow the necessary settings.
9313 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9314 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9315 done automatically (in contradiction to the requirements of the C
9316 standard). This made problems when used from OpenSSH.
9319 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9320 dh->length and always used
9322 BN_rand_range(priv_key, dh->p).
9324 BN_rand_range() is not necessary for Diffie-Hellman, and this
9325 specific range makes Diffie-Hellman unnecessarily inefficient if
9326 dh->length (recommended exponent length) is much smaller than the
9327 length of dh->p. We could use BN_rand_range() if the order of
9328 the subgroup was stored in the DH structure, but we only have
9333 BN_rand(priv_key, l, ...)
9335 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
9341 RSA_eay_public_encrypt
9342 RSA_eay_private_decrypt
9343 RSA_eay_private_encrypt (signing)
9344 RSA_eay_public_decrypt (signature verification)
9346 (default implementations for RSA_public_encrypt,
9347 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
9348 always reject numbers >= n.
9351 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
9352 to synchronize access to 'locking_thread'. This is necessary on
9353 systems where access to 'locking_thread' (an 'unsigned long'
9354 variable) is not atomic.
9357 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
9358 *before* setting the 'crypto_lock_rand' flag. The previous code had
9359 a race condition if 0 is a valid thread ID.
9360 [Travis Vitek <vitek@roguewave.com>]
9362 *) Add support for shared libraries under Irix.
9363 [Albert Chin-A-Young <china@thewrittenword.com>]
9365 *) Add configuration option to build on Linux on both big-endian and
9367 [Ralf Baechle <ralf@uni-koblenz.de>]
9369 *) Add the possibility to create shared libraries on HP-UX.
9372 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
9374 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
9375 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
9376 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
9377 PRNG state recovery was possible based on the output of
9378 one PRNG request appropriately sized to gain knowledge on
9379 'md' followed by enough consecutive 1-byte PRNG requests
9380 to traverse all of 'state'.
9382 1. When updating 'md_local' (the current thread's copy of 'md')
9383 during PRNG output generation, hash all of the previous
9384 'md_local' value, not just the half used for PRNG output.
9386 2. Make the number of bytes from 'state' included into the hash
9387 independent from the number of PRNG bytes requested.
9389 The first measure alone would be sufficient to avoid
9390 Markku-Juhani's attack. (Actually it had never occurred
9391 to me that the half of 'md_local' used for chaining was the
9392 half from which PRNG output bytes were taken -- I had always
9393 assumed that the secret half would be used.) The second
9394 measure makes sure that additional data from 'state' is never
9395 mixed into 'md_local' in small portions; this heuristically
9396 further strengthens the PRNG.
9399 *) Fix crypto/bn/asm/mips3.s.
9402 *) When only the key is given to "enc", the IV is undefined. Print out
9403 an error message in this case.
9406 *) Handle special case when X509_NAME is empty in X509 printing routines.
9409 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
9410 positive and less than q.
9413 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
9414 used: it isn't thread safe and the add_lock_callback should handle
9416 [Paul Rose <Paul.Rose@bridge.com>]
9418 *) Verify that incoming data obeys the block size in
9419 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
9423 [Ulf Möller, Bodo Möller]
9425 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
9426 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
9427 when fixing the server behaviour for backwards-compatible 'client
9428 hello' messages. (Note that the attack is impractical against
9429 SSL 3.0 and TLS 1.0 anyway because length and version checking
9430 means that the probability of guessing a valid ciphertext is
9431 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
9434 Before 0.9.5, the countermeasure (hide the error by generating a
9435 random 'decryption result') did not work properly because
9436 ERR_clear_error() was missing, meaning that SSL_get_error() would
9437 detect the supposedly ignored error.
9439 Both problems are now fixed.
9442 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
9443 (previously it was 1024).
9446 *) Fix for compatibility mode trust settings: ignore trust settings
9447 unless some valid trust or reject settings are present.
9450 *) Fix for blowfish EVP: its a variable length cipher.
9453 *) Fix various bugs related to DSA S/MIME verification. Handle missing
9454 parameters in DSA public key structures and return an error in the
9455 DSA routines if parameters are absent.
9458 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
9459 in the current directory if neither $RANDFILE nor $HOME was set.
9460 RAND_file_name() in 0.9.6a returned NULL in this case. This has
9461 caused some confusion to Windows users who haven't defined $HOME.
9462 Thus RAND_file_name() is changed again: e_os.h can define a
9463 DEFAULT_HOME, which will be used if $HOME is not set.
9464 For Windows, we use "C:"; on other platforms, we still require
9465 environment variables.
9467 *) Move 'if (!initialized) RAND_poll()' into regions protected by
9468 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
9469 having multiple threads call RAND_poll() concurrently.
9472 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
9473 combination of a flag and a thread ID variable.
9474 Otherwise while one thread is in ssleay_rand_bytes (which sets the
9475 flag), *other* threads can enter ssleay_add_bytes without obeying
9476 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
9477 that they do not hold after the first thread unsets add_do_not_lock).
9480 *) Change bctest again: '-x' expressions are not available in all
9484 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
9486 *) Fix a couple of memory leaks in PKCS7_dataDecode()
9487 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9489 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9490 the default extension for executables, if any. Also, make the perl
9491 scripts that use symlink() to test if it really exists and use "cp"
9492 if it doesn't. All this made OpenSSL compilable and installable in
9496 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9497 If SEQUENCE is length is indefinite just set c->slen to the total
9498 amount of data available.
9499 [Steve Henson, reported by shige@FreeBSD.org]
9500 [This change does not apply to 0.9.7.]
9502 *) Change bctest to avoid here-documents inside command substitution
9503 (workaround for FreeBSD /bin/sh bug).
9504 For compatibility with Ultrix, avoid shell functions (introduced
9505 in the bctest version that searches along $PATH).
9508 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9509 with des_encrypt() defined on some operating systems, like Solaris
9513 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9514 On the Importance of Eliminating Errors in Cryptographic
9515 Computations, J. Cryptology 14 (2001) 2, 101-119,
9516 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9519 *) MIPS assembler BIGNUM division bug fix.
9522 *) Disabled incorrect Alpha assembler code.
9525 *) Fix PKCS#7 decode routines so they correctly update the length
9526 after reading an EOC for the EXPLICIT tag.
9528 [This change does not apply to 0.9.7.]
9530 *) Fix bug in PKCS#12 key generation routines. This was triggered
9531 if a 3DES key was generated with a 0 initial byte. Include
9532 PKCS12_BROKEN_KEYGEN compilation option to retain the old
9533 (but broken) behaviour.
9536 *) Enhance bctest to search for a working bc along $PATH and print
9538 [Tim Rice <tim@multitalents.net> via Richard Levitte]
9540 *) Fix memory leaks in err.c: free err_data string if necessary;
9541 don't write to the wrong index in ERR_set_error_data.
9544 *) Implement ssl23_peek (analogous to ssl23_read), which previously
9548 *) Replace rdtsc with _emit statements for VC++ version 5.
9549 [Jeremy Cooper <jeremy@baymoo.org>]
9551 *) Make it possible to reuse SSLv2 sessions.
9554 *) In copy_email() check for >= 0 as a return value for
9555 X509_NAME_get_index_by_NID() since 0 is a valid index.
9556 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
9558 *) Avoid coredump with unsupported or invalid public keys by checking if
9559 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
9560 PKCS7_verify() fails with non detached data.
9563 *) Don't use getenv in library functions when run as setuid/setgid.
9564 New function OPENSSL_issetugid().
9567 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
9568 due to incorrect handling of multi-threading:
9570 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
9572 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
9574 3. Count how many times MemCheck_off() has been called so that
9575 nested use can be treated correctly. This also avoids
9576 inband-signalling in the previous code (which relied on the
9577 assumption that thread ID 0 is impossible).
9580 *) Add "-rand" option also to s_client and s_server.
9583 *) Fix CPU detection on Irix 6.x.
9584 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
9585 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9587 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
9590 [This change does not apply to 0.9.7.]
9592 *) Use the cached encoding of an X509_NAME structure rather than
9593 copying it. This is apparently the reason for the libsafe "errors"
9594 but the code is actually correct.
9597 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
9598 Bleichenbacher's DSA attack.
9599 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
9600 to be set and top=0 forces the highest bit to be set; top=-1 is new
9601 and leaves the highest bit random.
9602 [Ulf Moeller, Bodo Moeller]
9604 *) In the NCONF_...-based implementations for CONF_... queries
9605 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
9606 a temporary CONF structure with the data component set to NULL
9607 (which gives segmentation faults in lh_retrieve).
9608 Instead, use NULL for the CONF pointer in CONF_get_string and
9609 CONF_get_number (which may use environment variables) and directly
9610 return NULL from CONF_get_section.
9613 *) Fix potential buffer overrun for EBCDIC.
9616 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9617 keyUsage if basicConstraints absent for a CA.
9620 *) Make SMIME_write_PKCS7() write mail header values with a format that
9621 is more generally accepted (no spaces before the semicolon), since
9622 some programs can't parse those values properly otherwise. Also make
9623 sure BIO's that break lines after each write do not create invalid
9627 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9628 macros previously used would not encode an empty SEQUENCE OF
9629 and break the signature.
9631 [This change does not apply to 0.9.7.]
9633 *) Zero the premaster secret after deriving the master secret in
9637 *) Add some EVP_add_digest_alias registrations (as found in
9638 OpenSSL_add_all_digests()) to SSL_library_init()
9639 aka OpenSSL_add_ssl_algorithms(). This provides improved
9640 compatibility with peers using X.509 certificates
9641 with unconventional AlgorithmIdentifier OIDs.
9644 *) Fix for Irix with NO_ASM.
9645 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9647 *) ./config script fixes.
9648 [Ulf Moeller, Richard Levitte]
9650 *) Fix 'openssl passwd -1'.
9653 *) Change PKCS12_key_gen_asc() so it can cope with non null
9654 terminated strings whose length is passed in the passlen
9655 parameter, for example from PEM callbacks. This was done
9656 by adding an extra length parameter to asc2uni().
9657 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9659 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9660 call failed, free the DSA structure.
9663 *) Fix to uni2asc() to cope with zero length Unicode strings.
9664 These are present in some PKCS#12 files.
9667 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9668 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9669 when writing a 32767 byte record.
9670 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9672 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9673 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9675 (RSA objects have a reference count access to which is protected
9676 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9677 so they are meant to be shared between threads.)
9678 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9679 "Reddie, Steven" <Steven.Reddie@ca.com>]
9681 *) Fix a deadlock in CRYPTO_mem_leaks().
9684 *) Use better test patterns in bntest.
9687 *) rand_win.c fix for Borland C.
9690 *) BN_rshift bugfix for n == 0.
9693 *) Add a 'bctest' script that checks for some known 'bc' bugs
9694 so that 'make test' does not abort just because 'bc' is broken.
9697 *) Store verify_result within SSL_SESSION also for client side to
9698 avoid potential security hole. (Re-used sessions on the client side
9699 always resulted in verify_result==X509_V_OK, not using the original
9700 result of the server certificate verification.)
9703 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9704 SSL3_RT_APPLICATION_DATA, return 0.
9705 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9709 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9710 releases, have been re-implemented by renaming the previous
9711 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9712 and ssl3_read_internal, respectively, and adding 'peek' parameters
9713 to them. The new ssl[23]_{read,peek} functions are calls to
9714 ssl[23]_read_internal with the 'peek' flag set appropriately.
9715 A 'peek' parameter has also been added to ssl3_read_bytes, which
9716 does the actual work for ssl3_read_internal.
9719 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9720 the method-specific "init()" handler. Also clean up ex_data after
9721 calling the method-specific "finish()" handler. Previously, this was
9722 happening the other way round.
9725 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9726 The previous value, 12, was not always sufficient for BN_mod_exp().
9729 *) Make sure that shared libraries get the internal name engine with
9730 the full version number and not just 0. This should mark the
9731 shared libraries as not backward compatible. Of course, this should
9732 be changed again when we can guarantee backward binary compatibility.
9735 *) Fix typo in get_cert_by_subject() in by_dir.c
9736 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9738 *) Rework the system to generate shared libraries:
9740 - Make note of the expected extension for the shared libraries and
9741 if there is a need for symbolic links from for example libcrypto.so.0
9742 to libcrypto.so.0.9.7. There is extended info in Configure for
9745 - Make as few rebuilds of the shared libraries as possible.
9747 - Still avoid linking the OpenSSL programs with the shared libraries.
9749 - When installing, install the shared libraries separately from the
9753 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
9755 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
9756 and not in SSL_clear because the latter is also used by the
9757 accept/connect functions; previously, the settings made by
9758 SSL_set_read_ahead would be lost during the handshake.
9759 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
9761 *) Correct util/mkdef.pl to be selective about disabled algorithms.
9762 Previously, it would create entries for disabled algorithms no
9766 *) Added several new manual pages for SSL_* function.
9769 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
9771 *) In ssl23_get_client_hello, generate an error message when faced
9772 with an initial SSL 3.0/TLS record that is too small to contain the
9773 first two bytes of the ClientHello message, i.e. client_version.
9774 (Note that this is a pathologic case that probably has never happened
9775 in real life.) The previous approach was to use the version number
9776 from the record header as a substitute; but our protocol choice
9777 should not depend on that one because it is not authenticated
9778 by the Finished messages.
9781 *) More robust randomness gathering functions for Windows.
9782 [Jeffrey Altman <jaltman@columbia.edu>]
9784 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
9785 not set then we don't setup the error code for issuer check errors
9786 to avoid possibly overwriting other errors which the callback does
9787 handle. If an application does set the flag then we assume it knows
9788 what it is doing and can handle the new informational codes
9792 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
9793 a general "ANY" type, as such it should be able to decode anything
9794 including tagged types. However it didn't check the class so it would
9795 wrongly interpret tagged types in the same way as their universal
9796 counterpart and unknown types were just rejected. Changed so that the
9797 tagged and unknown types are handled in the same way as a SEQUENCE:
9798 that is the encoding is stored intact. There is also a new type
9799 "V_ASN1_OTHER" which is used when the class is not universal, in this
9800 case we have no idea what the actual type is so we just lump them all
9804 *) On VMS, stdout may very well lead to a file that is written to
9805 in a record-oriented fashion. That means that every write() will
9806 write a separate record, which will be read separately by the
9807 programs trying to read from it. This can be very confusing.
9809 The solution is to put a BIO filter in the way that will buffer
9810 text until a linefeed is reached, and then write everything a
9811 line at a time, so every record written will be an actual line,
9812 not chunks of lines and not (usually doesn't happen, but I've
9813 seen it once) several lines in one record. BIO_f_linebuffer() is
9816 Currently, it's a VMS-only method, because that's where it has
9817 been tested well enough.
9820 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
9821 it can return incorrect results.
9822 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
9823 but it was in 0.9.6-beta[12].)
9826 *) Disable the check for content being present when verifying detached
9827 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9828 include zero length content when signing messages.
9831 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9832 BIO_ctrl (for BIO pairs).
9835 *) Add DSO method for VMS.
9838 *) Bug fix: Montgomery multiplication could produce results with the
9842 *) Add RPM specification openssl.spec and modify it to build three
9843 packages. The default package contains applications, application
9844 documentation and run-time libraries. The devel package contains
9845 include files, static libraries and function documentation. The
9846 doc package contains the contents of the doc directory. The original
9847 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
9850 *) Add a large number of documentation files for many SSL routines.
9851 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9853 *) Add a configuration entry for Sony News 4.
9854 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
9856 *) Don't set the two most significant bits to one when generating a
9857 random number < q in the DSA library.
9860 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
9861 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
9862 the underlying transport is blocking) if a handshake took place.
9863 (The default behaviour is needed by applications such as s_client
9864 and s_server that use select() to determine when to use SSL_read;
9865 but for applications that know in advance when to expect data, it
9866 just makes things more complicated.)
9869 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
9873 *) Add a few more EBCDIC conditionals that make `req' and `x509'
9874 work better on such systems.
9875 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9877 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
9878 Update PKCS12_parse() so it copies the friendlyName and the
9879 keyid to the certificates aux info.
9882 *) Fix bug in PKCS7_verify() which caused an infinite loop
9883 if there was more than one signature.
9884 [Sven Uszpelkat <su@celocom.de>]
9886 *) Major change in util/mkdef.pl to include extra information
9887 about each symbol, as well as presenting variables as well
9888 as functions. This change means that there's n more need
9889 to rebuild the .num files when some algorithms are excluded.
9892 *) Allow the verify time to be set by an application,
9893 rather than always using the current time.
9896 *) Phase 2 verify code reorganisation. The certificate
9897 verify code now looks up an issuer certificate by a
9898 number of criteria: subject name, authority key id
9899 and key usage. It also verifies self signed certificates
9900 by the same criteria. The main comparison function is
9901 X509_check_issued() which performs these checks.
9903 Lot of changes were necessary in order to support this
9904 without completely rewriting the lookup code.
9906 Authority and subject key identifier are now cached.
9908 The LHASH 'certs' is X509_STORE has now been replaced
9909 by a STACK_OF(X509_OBJECT). This is mainly because an
9910 LHASH can't store or retrieve multiple objects with
9911 the same hash value.
9913 As a result various functions (which were all internal
9914 use only) have changed to handle the new X509_STORE
9915 structure. This will break anything that messed round
9916 with X509_STORE internally.
9918 The functions X509_STORE_add_cert() now checks for an
9919 exact match, rather than just subject name.
9921 The X509_STORE API doesn't directly support the retrieval
9922 of multiple certificates matching a given criteria, however
9923 this can be worked round by performing a lookup first
9924 (which will fill the cache with candidate certificates)
9925 and then examining the cache for matches. This is probably
9926 the best we can do without throwing out X509_LOOKUP
9927 entirely (maybe later...).
9929 The X509_VERIFY_CTX structure has been enhanced considerably.
9931 All certificate lookup operations now go via a get_issuer()
9932 callback. Although this currently uses an X509_STORE it
9933 can be replaced by custom lookups. This is a simple way
9934 to bypass the X509_STORE hackery necessary to make this
9935 work and makes it possible to use more efficient techniques
9936 in future. A very simple version which uses a simple
9937 STACK for its trusted certificate store is also provided
9938 using X509_STORE_CTX_trusted_stack().
9940 The verify_cb() and verify() callbacks now have equivalents
9941 in the X509_STORE_CTX structure.
9943 X509_STORE_CTX also has a 'flags' field which can be used
9944 to customise the verify behaviour.
9947 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
9948 excludes S/MIME capabilities.
9951 *) When a certificate request is read in keep a copy of the
9952 original encoding of the signed data and use it when outputting
9953 again. Signatures then use the original encoding rather than
9954 a decoded, encoded version which may cause problems if the
9955 request is improperly encoded.
9958 *) For consistency with other BIO_puts implementations, call
9959 buffer_write(b, ...) directly in buffer_puts instead of calling
9962 In BIO_puts, increment b->num_write as in BIO_write.
9963 [Peter.Sylvester@EdelWeb.fr]
9965 *) Fix BN_mul_word for the case where the word is 0. (We have to use
9966 BN_zero, we may not return a BIGNUM with an array consisting of
9970 *) Avoid calling abort() from within the library when problems are
9971 detected, except if preprocessor symbols have been defined
9972 (such as REF_CHECK, BN_DEBUG etc.).
9975 *) New openssl application 'rsautl'. This utility can be
9976 used for low level RSA operations. DER public key
9977 BIO/fp routines also added.
9980 *) New Configure entry and patches for compiling on QNX 4.
9981 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
9983 *) A demo state-machine implementation was sponsored by
9984 Nuron (http://www.nuron.com/) and is now available in
9985 demos/state_machine.
9988 *) New options added to the 'dgst' utility for signature
9989 generation and verification.
9992 *) Unrecognized PKCS#7 content types are now handled via a
9993 catch all ASN1_TYPE structure. This allows unsupported
9994 types to be stored as a "blob" and an application can
9995 encode and decode it manually.
9998 *) Fix various signed/unsigned issues to make a_strex.c
10000 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
10002 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
10003 length if passed a buffer. ASN1_INTEGER_to_BN failed
10004 if passed a NULL BN and its argument was negative.
10005 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
10007 *) Modification to PKCS#7 encoding routines to output definite
10008 length encoding. Since currently the whole structures are in
10009 memory there's not real point in using indefinite length
10010 constructed encoding. However if OpenSSL is compiled with
10011 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
10014 *) Added BIO_vprintf() and BIO_vsnprintf().
10017 *) Added more prefixes to parse for in the strings written
10018 through a logging bio, to cover all the levels that are available
10019 through syslog. The prefixes are now:
10021 PANIC, EMERG, EMR => LOG_EMERG
10022 ALERT, ALR => LOG_ALERT
10023 CRIT, CRI => LOG_CRIT
10024 ERROR, ERR => LOG_ERR
10025 WARNING, WARN, WAR => LOG_WARNING
10026 NOTICE, NOTE, NOT => LOG_NOTICE
10027 INFO, INF => LOG_INFO
10028 DEBUG, DBG => LOG_DEBUG
10030 and as before, if none of those prefixes are present at the
10031 beginning of the string, LOG_ERR is chosen.
10033 On Win32, the LOG_* levels are mapped according to this:
10035 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
10036 LOG_WARNING => EVENTLOG_WARNING_TYPE
10037 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
10041 *) Made it possible to reconfigure with just the configuration
10042 argument "reconf" or "reconfigure". The command line arguments
10043 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
10044 and are retrieved from there when reconfiguring.
10047 *) MD4 implemented.
10048 [Assar Westerlund <assar@sics.se>, Richard Levitte]
10050 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
10053 *) The obj_dat.pl script was messing up the sorting of object
10054 names. The reason was that it compared the quoted version
10055 of strings as a result "OCSP" > "OCSP Signing" because
10056 " > SPACE. Changed script to store unquoted versions of
10057 names and add quotes on output. It was also omitting some
10058 names from the lookup table if they were given a default
10059 value (that is if SN is missing it is given the same
10060 value as LN and vice versa), these are now added on the
10061 grounds that if an object has a name we should be able to
10062 look it up. Finally added warning output when duplicate
10063 short or long names are found.
10066 *) Changes needed for Tandem NSK.
10067 [Scott Uroff <scott@xypro.com>]
10069 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
10070 RSA_padding_check_SSLv23(), special padding was never detected
10071 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
10072 version rollback attacks was not effective.
10074 In s23_clnt.c, don't use special rollback-attack detection padding
10075 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
10076 client; similarly, in s23_srvr.c, don't do the rollback check if
10077 SSL 2.0 is the only protocol enabled in the server.
10080 *) Make it possible to get hexdumps of unprintable data with 'openssl
10081 asn1parse'. By implication, the functions ASN1_parse_dump() and
10082 BIO_dump_indent() are added.
10085 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
10086 these print out strings and name structures based on various
10087 flags including RFC2253 support and proper handling of
10088 multibyte characters. Added options to the 'x509' utility
10089 to allow the various flags to be set.
10092 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
10093 Also change the functions X509_cmp_current_time() and
10094 X509_gmtime_adj() work with an ASN1_TIME structure,
10095 this will enable certificates using GeneralizedTime in validity
10096 dates to be checked.
10099 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
10100 negative public key encodings) on by default,
10101 NO_NEG_PUBKEY_BUG can be set to disable it.
10104 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
10105 content octets. An i2c_ASN1_OBJECT is unnecessary because
10106 the encoding can be trivially obtained from the structure.
10109 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
10110 not read locks (CRYPTO_r_[un]lock).
10113 *) A first attempt at creating official support for shared
10114 libraries through configuration. I've kept it so the
10115 default is static libraries only, and the OpenSSL programs
10116 are always statically linked for now, but there are
10117 preparations for dynamic linking in place.
10118 This has been tested on Linux and Tru64.
10121 *) Randomness polling function for Win9x, as described in:
10122 Peter Gutmann, Software Generation of Practically Strong
10126 *) Fix so PRNG is seeded in req if using an already existing
10130 *) New options to smime application. -inform and -outform
10131 allow alternative formats for the S/MIME message including
10132 PEM and DER. The -content option allows the content to be
10133 specified separately. This should allow things like Netscape
10134 form signing output easier to verify.
10137 *) Fix the ASN1 encoding of tags using the 'long form'.
10140 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
10141 STRING types. These convert content octets to and from the
10142 underlying type. The actual tag and length octets are
10143 already assumed to have been read in and checked. These
10144 are needed because all other string types have virtually
10145 identical handling apart from the tag. By having versions
10146 of the ASN1 functions that just operate on content octets
10147 IMPLICIT tagging can be handled properly. It also allows
10148 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
10149 and ASN1_INTEGER are identical apart from the tag.
10152 *) Change the handling of OID objects as follows:
10154 - New object identifiers are inserted in objects.txt, following
10155 the syntax given in objects.README.
10156 - objects.pl is used to process obj_mac.num and create a new
10158 - obj_dat.pl is used to create a new obj_dat.h, using the data in
10161 This is currently kind of a hack, and the perl code in objects.pl
10162 isn't very elegant, but it works as I intended. The simplest way
10163 to check that it worked correctly is to look in obj_dat.h and
10164 check the array nid_objs and make sure the objects haven't moved
10165 around (this is important!). Additions are OK, as well as
10166 consistent name changes.
10169 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
10172 *) Addition of the command line parameter '-rand file' to 'openssl req'.
10173 The given file adds to whatever has already been seeded into the
10174 random pool through the RANDFILE configuration file option or
10175 environment variable, or the default random state file.
10178 *) mkstack.pl now sorts each macro group into lexical order.
10179 Previously the output order depended on the order the files
10180 appeared in the directory, resulting in needless rewriting
10184 *) Patches to make OpenSSL compile under Win32 again. Mostly
10185 work arounds for the VC++ problem that it treats func() as
10186 func(void). Also stripped out the parts of mkdef.pl that
10187 added extra typesafe functions: these no longer exist.
10190 *) Reorganisation of the stack code. The macros are now all
10191 collected in safestack.h . Each macro is defined in terms of
10192 a "stack macro" of the form SKM_<name>(type, a, b). The
10193 DEBUG_SAFESTACK is now handled in terms of function casts,
10194 this has the advantage of retaining type safety without the
10195 use of additional functions. If DEBUG_SAFESTACK is not defined
10196 then the non typesafe macros are used instead. Also modified the
10197 mkstack.pl script to handle the new form. Needs testing to see
10198 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10199 the default if no major problems. Similar behaviour for ASN1_SET_OF
10200 and PKCS12_STACK_OF.
10203 *) When some versions of IIS use the 'NET' form of private key the
10204 key derivation algorithm is different. Normally MD5(password) is
10205 used as a 128 bit RC4 key. In the modified case
10206 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10207 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10208 as the old Netscape_RSA functions except they have an additional
10209 'sgckey' parameter which uses the modified algorithm. Also added
10210 an -sgckey command line option to the rsa utility. Thanks to
10211 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10212 algorithm to openssl-dev.
10215 *) The evp_local.h macros were using 'c.##kname' which resulted in
10216 invalid expansion on some systems (SCO 5.0.5 for example).
10217 Corrected to 'c.kname'.
10218 [Phillip Porch <root@theporch.com>]
10220 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10221 a STACK of email addresses from a certificate or request, these look
10222 in the subject name and the subject alternative name extensions and
10223 omit any duplicate addresses.
10226 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10227 This makes DSA verification about 2 % faster.
10230 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10231 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10232 plus overhead for 1024 bit moduli).
10233 This makes exponentiations about 0.5 % faster for 1024 bit
10234 exponents (as measured by "openssl speed rsa2048").
10237 *) Rename memory handling macros to avoid conflicts with other
10239 Malloc => OPENSSL_malloc
10240 Malloc_locked => OPENSSL_malloc_locked
10241 Realloc => OPENSSL_realloc
10242 Free => OPENSSL_free
10245 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10246 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10249 *) CygWin32 support.
10250 [John Jarvie <jjarvie@newsguy.com>]
10252 *) The type-safe stack code has been rejigged. It is now only compiled
10253 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10254 by default all type-specific stack functions are "#define"d back to
10255 standard stack functions. This results in more streamlined output
10256 but retains the type-safety checking possibilities of the original
10260 *) The STACK code has been cleaned up, and certain type declarations
10261 that didn't make a lot of sense have been brought in line. This has
10262 also involved a cleanup of sorts in safestack.h to more correctly
10263 map type-safe stack functions onto their plain stack counterparts.
10264 This work has also resulted in a variety of "const"ifications of
10265 lots of the code, especially "_cmp" operations which should normally
10266 be prototyped with "const" parameters anyway.
10269 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10270 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10271 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10272 where all of 'md' is used each time the PRNG is used, but 'state'
10273 is used only indexed by a cyclic counter. As entropy may not be
10274 well distributed from the beginning, 'md' is important as a
10275 chaining variable. However, the output function chains only half
10276 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10277 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10278 in all of 'state' being rewritten, with the new values depending
10279 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10282 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10283 the handshake is continued after ssl_verify_cert_chain();
10284 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10285 can lead to 'unexplainable' connection aborts later.
10286 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10288 *) Major EVP API cipher revision.
10289 Add hooks for extra EVP features. This allows various cipher
10290 parameters to be set in the EVP interface. Support added for variable
10291 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10292 setting of RC2 and RC5 parameters.
10294 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10297 Remove lots of duplicated code from the EVP library. For example *every*
10298 cipher init() function handles the 'iv' in the same way according to the
10299 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10300 for CFB and OFB modes they zero ctx->num.
10302 New functionality allows removal of S/MIME code RC2 hack.
10304 Most of the routines have the same form and so can be declared in terms
10307 By shifting this to the top level EVP_CipherInit() it can be removed from
10308 all individual ciphers. If the cipher wants to handle IVs or keys
10309 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10312 Change lots of functions like EVP_EncryptUpdate() to now return a
10313 value: although software versions of the algorithms cannot fail
10314 any installed hardware versions can.
10317 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10318 this option is set, tolerate broken clients that send the negotiated
10319 protocol version number instead of the requested protocol version
10323 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10324 i.e. non-zero for export ciphersuites, zero otherwise.
10325 Previous versions had this flag inverted, inconsistent with
10326 rsa_tmp_cb (..._TMP_RSA_CB).
10327 [Bodo Moeller; problem reported by Amit Chopra]
10329 *) Add missing DSA library text string. Work around for some IIS
10330 key files with invalid SEQUENCE encoding.
10333 *) Add a document (doc/standards.txt) that list all kinds of standards
10334 and so on that are implemented in OpenSSL.
10337 *) Enhance c_rehash script. Old version would mishandle certificates
10338 with the same subject name hash and wouldn't handle CRLs at all.
10339 Added -fingerprint option to crl utility, to support new c_rehash
10343 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
10346 *) Fix for SSL server purpose checking. Server checking was
10347 rejecting certificates which had extended key usage present
10348 but no ssl client purpose.
10349 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
10351 *) Make PKCS#12 code work with no password. The PKCS#12 spec
10352 is a little unclear about how a blank password is handled.
10353 Since the password in encoded as a BMPString with terminating
10354 double NULL a zero length password would end up as just the
10355 double NULL. However no password at all is different and is
10356 handled differently in the PKCS#12 key generation code. NS
10357 treats a blank password as zero length. MSIE treats it as no
10358 password on export: but it will try both on import. We now do
10359 the same: PKCS12_parse() tries zero length and no password if
10360 the password is set to "" or NULL (NULL is now a valid password:
10361 it wasn't before) as does the pkcs12 application.
10364 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
10365 perror when PEM_read_bio_X509_REQ fails, the error message must
10366 be obtained from the error queue.
10369 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
10370 it in ERR_remove_state if appropriate, and change ERR_get_state
10371 accordingly to avoid race conditions (this is necessary because
10372 thread_hash is no longer constant once set).
10375 *) Bugfix for linux-elf makefile.one.
10378 *) RSA_get_default_method() will now cause a default
10379 RSA_METHOD to be chosen if one doesn't exist already.
10380 Previously this was only set during a call to RSA_new()
10381 or RSA_new_method(NULL) meaning it was possible for
10382 RSA_get_default_method() to return NULL.
10385 *) Added native name translation to the existing DSO code
10386 that will convert (if the flag to do so is set) filenames
10387 that are sufficiently small and have no path information
10388 into a canonical native form. Eg. "blah" converted to
10389 "libblah.so" or "blah.dll" etc.
10392 *) New function ERR_error_string_n(e, buf, len) which is like
10393 ERR_error_string(e, buf), but writes at most 'len' bytes
10394 including the 0 terminator. For ERR_error_string_n, 'buf'
10396 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
10398 *) CONF library reworked to become more general. A new CONF
10399 configuration file reader "class" is implemented as well as a
10400 new functions (NCONF_*, for "New CONF") to handle it. The now
10401 old CONF_* functions are still there, but are reimplemented to
10402 work in terms of the new functions. Also, a set of functions
10403 to handle the internal storage of the configuration data is
10404 provided to make it easier to write new configuration file
10405 reader "classes" (I can definitely see something reading a
10406 configuration file in XML format, for example), called _CONF_*,
10407 or "the configuration storage API"...
10409 The new configuration file reading functions are:
10411 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
10412 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
10414 NCONF_default, NCONF_WIN32
10416 NCONF_dump_fp, NCONF_dump_bio
10418 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
10419 NCONF_new creates a new CONF object. This works in the same way
10420 as other interfaces in OpenSSL, like the BIO interface.
10421 NCONF_dump_* dump the internal storage of the configuration file,
10422 which is useful for debugging. All other functions take the same
10423 arguments as the old CONF_* functions wth the exception of the
10424 first that must be a `CONF *' instead of a `LHASH *'.
10426 To make it easer to use the new classes with the old CONF_* functions,
10427 the function CONF_set_default_method is provided.
10430 *) Add '-tls1' option to 'openssl ciphers', which was already
10431 mentioned in the documentation but had not been implemented.
10432 (This option is not yet really useful because even the additional
10433 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
10436 *) Initial DSO code added into libcrypto for letting OpenSSL (and
10437 OpenSSL-based applications) load shared libraries and bind to
10438 them in a portable way.
10439 [Geoff Thorpe, with contributions from Richard Levitte]
10441 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
10443 *) Make sure _lrotl and _lrotr are only used with MSVC.
10445 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
10446 (the default implementation of RAND_status).
10448 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
10449 to '-clrext' (= clear extensions), as intended and documented.
10450 [Bodo Moeller; inconsistency pointed out by Michael Attili
10451 <attili@amaxo.com>]
10453 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
10454 was larger than the MD block size.
10455 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
10457 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
10458 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
10459 using the passed key: if the passed key was a private key the result
10460 of X509_print(), for example, would be to print out all the private key
10464 *) des_quad_cksum() byte order bug fix.
10465 [Ulf Möller, using the problem description in krb4-0.9.7, where
10466 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
10468 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
10470 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
10472 *) For easily testing in shell scripts whether some command
10473 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
10474 returns with exit code 0 iff no command of the given name is available.
10475 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
10476 the output goes to stdout and nothing is printed to stderr.
10477 Additional arguments are always ignored.
10479 Since for each cipher there is a command of the same name,
10480 the 'no-cipher' compilation switches can be tested this way.
10482 ('openssl no-XXX' is not able to detect pseudo-commands such
10483 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
10486 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10489 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10490 is set; it will be thrown away anyway because each handshake creates
10492 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10493 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10494 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10495 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10498 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10499 'Q' and 'R' lose their special meanings (quit/renegotiate).
10500 This is part of what -quiet does; unlike -quiet, -ign_eof
10501 does not suppress any output.
10504 *) Add compatibility options to the purpose and trust code. The
10505 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10506 accepts a certificate or CA, this was the previous behaviour,
10507 with all the associated security issues.
10509 X509_TRUST_COMPAT is the old trust behaviour: only and
10510 automatically trust self signed roots in certificate store. A
10511 new trust setting X509_TRUST_DEFAULT is used to specify that
10512 a purpose has no associated trust setting and it should instead
10513 use the value in the default purpose.
10516 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10517 and fix a memory leak.
10520 *) In util/mkerr.pl (which implements 'make errors'), preserve
10521 reason strings from the previous version of the .c file, as
10522 the default to have only downcase letters (and digits) in
10523 automatically generated reasons codes is not always appropriate.
10526 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
10527 using strerror. Previously, ERR_reason_error_string() returned
10528 library names as reason strings for SYSerr; but SYSerr is a special
10529 case where small numbers are errno values, not library numbers.
10532 *) Add '-dsaparam' option to 'openssl dhparam' application. This
10533 converts DSA parameters into DH parameters. (When creating parameters,
10534 DSA_generate_parameters is used.)
10537 *) Include 'length' (recommended exponent length) in C code generated
10538 by 'openssl dhparam -C'.
10541 *) The second argument to set_label in perlasm was already being used
10542 so couldn't be used as a "file scope" flag. Moved to third argument
10546 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
10547 instead of RAND_bytes for encryption IVs and salts.
10550 *) Include RAND_status() into RAND_METHOD instead of implementing
10551 it only for md_rand.c Otherwise replacing the PRNG by calling
10552 RAND_set_rand_method would be impossible.
10555 *) Don't let DSA_generate_key() enter an infinite loop if the random
10556 number generation fails.
10559 *) New 'rand' application for creating pseudo-random output.
10562 *) Added configuration support for Linux/IA64
10563 [Rolf Haberrecker <rolf@suse.de>]
10565 *) Assembler module support for Mingw32.
10568 *) Shared library support for HPUX (in shlib/).
10569 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
10571 *) Shared library support for Solaris gcc.
10572 [Lutz Behnke <behnke@trustcenter.de>]
10574 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
10576 *) PKCS7_encrypt() was adding text MIME headers twice because they
10577 were added manually and by SMIME_crlf_copy().
10580 *) In bntest.c don't call BN_rand with zero bits argument.
10581 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
10583 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
10584 case was implemented. This caused BN_div_recp() to fail occasionally.
10587 *) Add an optional second argument to the set_label() in the perl
10588 assembly language builder. If this argument exists and is set
10589 to 1 it signals that the assembler should use a symbol whose
10590 scope is the entire file, not just the current function. This
10591 is needed with MASM which uses the format label:: for this scope.
10592 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
10594 *) Change the ASN1 types so they are typedefs by default. Before
10595 almost all types were #define'd to ASN1_STRING which was causing
10596 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
10600 *) Change names of new functions to the new get1/get0 naming
10601 convention: After 'get1', the caller owns a reference count
10602 and has to call ..._free; 'get0' returns a pointer to some
10603 data structure without incrementing reference counters.
10604 (Some of the existing 'get' functions increment a reference
10605 counter, some don't.)
10606 Similarly, 'set1' and 'add1' functions increase reference
10607 counters or duplicate objects.
10610 *) Allow for the possibility of temp RSA key generation failure:
10611 the code used to assume it always worked and crashed on failure.
10614 *) Fix potential buffer overrun problem in BIO_printf().
10615 [Ulf Möller, using public domain code by Patrick Powell; problem
10616 pointed out by David Sacerdote <das33@cornell.edu>]
10618 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10619 RAND_egd() and RAND_status(). In the command line application,
10620 the EGD socket can be specified like a seed file using RANDFILE
10624 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10625 Some CAs (e.g. Verisign) distribute certificates in this form.
10628 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10629 list to exclude them. This means that no special compilation option
10630 is needed to use anonymous DH: it just needs to be included in the
10634 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10635 EVP_MD_type. The old functionality is available in a new macro called
10636 EVP_MD_md(). Change code that uses it and update docs.
10639 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10640 where the 'void *' argument is replaced by a function pointer argument.
10641 Previously 'void *' was abused to point to functions, which works on
10642 many platforms, but is not correct. As these functions are usually
10643 called by macros defined in OpenSSL header files, most source code
10644 should work without changes.
10647 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10648 sections with information on -D... compiler switches used for
10649 compiling the library so that applications can see them. To enable
10650 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10651 must be defined. E.g.,
10652 #define OPENSSL_ALGORITHM_DEFINES
10653 #include <openssl/opensslconf.h>
10654 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10655 [Richard Levitte, Ulf and Bodo Möller]
10657 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10661 *) Change the 'other' type in certificate aux info to a STACK_OF
10662 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10663 the required ASN1 format: arbitrary types determined by an OID.
10666 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10667 argument to 'req'. This is not because the function is newer or
10668 better than others it just uses the work 'NEW' in the certificate
10669 request header lines. Some software needs this.
10672 *) Reorganise password command line arguments: now passwords can be
10673 obtained from various sources. Delete the PEM_cb function and make
10674 it the default behaviour: i.e. if the callback is NULL and the
10675 usrdata argument is not NULL interpret it as a null terminated pass
10676 phrase. If usrdata and the callback are NULL then the pass phrase
10677 is prompted for as usual.
10680 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10681 the support is automatically enabled. The resulting binaries will
10682 autodetect the card and use it if present.
10683 [Ben Laurie and Compaq Inc.]
10685 *) Work around for Netscape hang bug. This sends certificate request
10686 and server done in one record. Since this is perfectly legal in the
10687 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10688 the bugs/SSLv3 entry for more info.
10691 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10694 *) Add -rand argument to smime and pkcs12 applications and read/write
10698 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10701 *) Add command line password options to the remaining applications.
10704 *) Bug fix for BN_div_recp() for numerators with an even number of
10708 *) More tests in bntest.c, and changed test_bn output.
10711 *) ./config recognizes MacOS X now.
10714 *) Bug fix for BN_div() when the first words of num and divisor are
10715 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10718 *) Add support for various broken PKCS#8 formats, and command line
10719 options to produce them.
10722 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10723 get temporary BIGNUMs from a BN_CTX.
10726 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10730 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10731 include a #define from the old name to the new. The original intent
10732 was that statically linked binaries could for example just call
10733 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10734 link with digests. This never worked because SSLeay_add_all_digests()
10735 and SSLeay_add_all_ciphers() were in the same source file so calling
10736 one would link with the other. They are now in separate source files.
10739 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10742 *) Use a less unusual form of the Miller-Rabin primality test (it used
10743 a binary algorithm for exponentiation integrated into the Miller-Rabin
10744 loop, our standard modexp algorithms are faster).
10747 *) Support for the EBCDIC character set completed.
10748 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10750 *) Source code cleanups: use const where appropriate, eliminate casts,
10751 use void * instead of char * in lhash.
10754 *) Bugfix: ssl3_send_server_key_exchange was not restartable
10755 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
10756 this the server could overwrite ephemeral keys that the client
10760 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
10761 using 50 iterations of the Rabin-Miller test.
10763 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
10764 iterations of the Rabin-Miller test as required by the appendix
10765 to FIPS PUB 186[-1]) instead of DSA_is_prime.
10766 As BN_is_prime_fasttest includes trial division, DSA parameter
10767 generation becomes much faster.
10769 This implies a change for the callback functions in DSA_is_prime
10770 and DSA_generate_parameters: The callback function is called once
10771 for each positive witness in the Rabin-Miller test, not just
10772 occasionally in the inner loop; and the parameters to the
10773 callback function now provide an iteration count for the outer
10774 loop rather than for the current invocation of the inner loop.
10775 DSA_generate_parameters additionally can call the callback
10776 function with an 'iteration count' of -1, meaning that a
10777 candidate has passed the trial division test (when q is generated
10778 from an application-provided seed, trial division is skipped).
10781 *) New function BN_is_prime_fasttest that optionally does trial
10782 division before starting the Rabin-Miller test and has
10783 an additional BN_CTX * argument (whereas BN_is_prime always
10784 has to allocate at least one BN_CTX).
10785 'callback(1, -1, cb_arg)' is called when a number has passed the
10786 trial division stage.
10789 *) Fix for bug in CRL encoding. The validity dates weren't being handled
10793 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
10796 *) New function BN_pseudo_rand().
10799 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
10800 bignum version of BN_from_montgomery() with the working code from
10801 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
10805 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
10806 made it impossible to use the same SSL_SESSION data structure in
10807 SSL2 clients in multiple threads.
10810 *) The return value of RAND_load_file() no longer counts bytes obtained
10811 by stat(). RAND_load_file(..., -1) is new and uses the complete file
10812 to seed the PRNG (previously an explicit byte count was required).
10813 [Ulf Möller, Bodo Möller]
10815 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
10816 used (char *) instead of (void *) and had casts all over the place.
10819 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
10822 *) Retain source code compatibility for BN_prime_checks macro:
10823 BN_is_prime(..., BN_prime_checks, ...) now uses
10824 BN_prime_checks_for_size to determine the appropriate number of
10825 Rabin-Miller iterations.
10828 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10829 DH_CHECK_P_NOT_SAFE_PRIME.
10830 (Check if this is true? OpenPGP calls them "strong".)
10833 *) Merge the functionality of "dh" and "gendh" programs into a new program
10834 "dhparam". The old programs are retained for now but will handle DH keys
10835 (instead of parameters) in future.
10838 *) Make the ciphers, s_server and s_client programs check the return values
10839 when a new cipher list is set.
10842 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
10843 ciphers. Before when the 56bit ciphers were enabled the sorting was
10846 The syntax for the cipher sorting has been extended to support sorting by
10847 cipher-strength (using the strength_bits hard coded in the tables).
10848 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
10850 Fix a bug in the cipher-command parser: when supplying a cipher command
10851 string with an "undefined" symbol (neither command nor alphanumeric
10852 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
10853 an error is flagged.
10855 Due to the strength-sorting extension, the code of the
10856 ssl_create_cipher_list() function was completely rearranged. I hope that
10857 the readability was also increased :-)
10858 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10860 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
10861 for the first serial number and places 2 in the serial number file. This
10862 avoids problems when the root CA is created with serial number zero and
10863 the first user certificate has the same issuer name and serial number
10867 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
10868 the new code. Add documentation for this stuff.
10871 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
10872 X509_*() to X509at_*() on the grounds that they don't handle X509
10873 structures and behave in an analogous way to the X509v3 functions:
10874 they shouldn't be called directly but wrapper functions should be used
10877 So we also now have some wrapper functions that call the X509at functions
10878 when passed certificate requests. (TO DO: similar things can be done with
10879 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
10880 things. Some of these need some d2i or i2d and print functionality
10881 because they handle more complex structures.)
10884 *) Add missing #ifndefs that caused missing symbols when building libssl
10885 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
10886 NO_RSA in ssl/s2*.c.
10887 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
10889 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
10890 has a return value which indicates the quality of the random data
10891 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
10892 error queue. New function RAND_pseudo_bytes() generates output that is
10893 guaranteed to be unique but not unpredictable. RAND_add is like
10894 RAND_seed, but takes an extra argument for an entropy estimate
10895 (RAND_seed always assumes full entropy).
10898 *) Do more iterations of Rabin-Miller probable prime test (specifically,
10899 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
10900 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
10901 in crypto/bn/bn_prime.c for the complete table). This guarantees a
10902 false-positive rate of at most 2^-80 for random input.
10905 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
10908 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
10909 in the 0.9.5 release), this returns the chain
10910 from an X509_CTX structure with a dup of the stack and all
10911 the X509 reference counts upped: so the stack will exist
10912 after X509_CTX_cleanup() has been called. Modify pkcs12.c
10915 Also make SSL_SESSION_print() print out the verify return
10919 *) Add manpage for the pkcs12 command. Also change the default
10920 behaviour so MAC iteration counts are used unless the new
10921 -nomaciter option is used. This improves file security and
10922 only older versions of MSIE (4.0 for example) need it.
10925 *) Honor the no-xxx Configure options when creating .DEF files.
10928 *) Add PKCS#10 attributes to field table: challengePassword,
10929 unstructuredName and unstructuredAddress. These are taken from
10930 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
10931 international characters are used.
10933 More changes to X509_ATTRIBUTE code: allow the setting of types
10934 based on strings. Remove the 'loc' parameter when adding
10935 attributes because these will be a SET OF encoding which is sorted
10939 *) Initial changes to the 'req' utility to allow request generation
10940 automation. This will allow an application to just generate a template
10941 file containing all the field values and have req construct the
10944 Initial support for X509_ATTRIBUTE handling. Stacks of these are
10945 used all over the place including certificate requests and PKCS#7
10946 structures. They are currently handled manually where necessary with
10947 some primitive wrappers for PKCS#7. The new functions behave in a
10948 manner analogous to the X509 extension functions: they allow
10949 attributes to be looked up by NID and added.
10951 Later something similar to the X509V3 code would be desirable to
10952 automatically handle the encoding, decoding and printing of the
10953 more complex types. The string types like challengePassword can
10954 be handled by the string table functions.
10956 Also modified the multi byte string table handling. Now there is
10957 a 'global mask' which masks out certain types. The table itself
10958 can use the flag STABLE_NO_MASK to ignore the mask setting: this
10959 is useful when for example there is only one permissible type
10960 (as in countryName) and using the mask might result in no valid
10964 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
10965 SSL_get_peer_finished to allow applications to obtain the latest
10966 Finished messages sent to the peer or expected from the peer,
10967 respectively. (SSL_get_peer_finished is usually the Finished message
10968 actually received from the peer, otherwise the protocol will be aborted.)
10970 As the Finished message are message digests of the complete handshake
10971 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
10972 be used for external authentication procedures when the authentication
10973 provided by SSL/TLS is not desired or is not enough.
10976 *) Enhanced support for Alpha Linux is added. Now ./config checks if
10977 the host supports BWX extension and if Compaq C is present on the
10978 $PATH. Just exploiting of the BWX extension results in 20-30%
10979 performance kick for some algorithms, e.g. DES and RC4 to mention
10980 a couple. Compaq C in turn generates ~20% faster code for MD5 and
10984 *) Add support for MS "fast SGC". This is arguably a violation of the
10985 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
10986 weak crypto and after checking the certificate is SGC a second one
10987 with strong crypto. MS SGC stops the first handshake after receiving
10988 the server certificate message and sends a second client hello. Since
10989 a server will typically do all the time consuming operations before
10990 expecting any further messages from the client (server key exchange
10991 is the most expensive) there is little difference between the two.
10993 To get OpenSSL to support MS SGC we have to permit a second client
10994 hello message after we have sent server done. In addition we have to
10995 reset the MAC if we do get this second client hello.
10998 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
10999 if a DER encoded private key is RSA or DSA traditional format. Changed
11000 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
11001 format DER encoded private key. Newer code should use PKCS#8 format which
11002 has the key type encoded in the ASN1 structure. Added DER private key
11003 support to pkcs8 application.
11006 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
11007 ciphersuites has been selected (as required by the SSL 3/TLS 1
11008 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
11009 is set, we interpret this as a request to violate the specification
11010 (the worst that can happen is a handshake failure, and 'correct'
11011 behaviour would result in a handshake failure anyway).
11014 *) In SSL_CTX_add_session, take into account that there might be multiple
11015 SSL_SESSION structures with the same session ID (e.g. when two threads
11016 concurrently obtain them from an external cache).
11017 The internal cache can handle only one SSL_SESSION with a given ID,
11018 so if there's a conflict, we now throw out the old one to achieve
11022 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
11023 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
11024 some routines that use cipher OIDs: some ciphers do not have OIDs
11025 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
11029 *) Simplify the trust setting structure and code. Now we just have
11030 two sequences of OIDs for trusted and rejected settings. These will
11031 typically have values the same as the extended key usage extension
11032 and any application specific purposes.
11034 The trust checking code now has a default behaviour: it will just
11035 check for an object with the same NID as the passed id. Functions can
11036 be provided to override either the default behaviour or the behaviour
11037 for a given id. SSL client, server and email already have functions
11038 in place for compatibility: they check the NID and also return "trusted"
11039 if the certificate is self signed.
11042 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
11043 traditional format into an EVP_PKEY structure.
11046 *) Add a password callback function PEM_cb() which either prompts for
11047 a password if usr_data is NULL or otherwise assumes it is a null
11048 terminated password. Allow passwords to be passed on command line
11049 environment or config files in a few more utilities.
11052 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
11053 keys. Add some short names for PKCS#8 PBE algorithms and allow them
11054 to be specified on the command line for the pkcs8 and pkcs12 utilities.
11055 Update documentation.
11058 *) Support for ASN1 "NULL" type. This could be handled before by using
11059 ASN1_TYPE but there wasn't any function that would try to read a NULL
11060 and produce an error if it couldn't. For compatibility we also have
11061 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
11062 don't allocate anything because they don't need to.
11065 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
11067 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
11069 *) Rebuild of the memory allocation routines used by OpenSSL code and
11070 possibly others as well. The purpose is to make an interface that
11071 provide hooks so anyone can build a separate set of allocation and
11072 deallocation routines to be used by OpenSSL, for example memory
11073 pool implementations, or something else, which was previously hard
11074 since Malloc(), Realloc() and Free() were defined as macros having
11075 the values malloc, realloc and free, respectively (except for Win32
11076 compilations). The same is provided for memory debugging code.
11077 OpenSSL already comes with functionality to find memory leaks, but
11078 this gives people a chance to debug other memory problems.
11080 With these changes, a new set of functions and macros have appeared:
11082 CRYPTO_set_mem_debug_functions() [F]
11083 CRYPTO_get_mem_debug_functions() [F]
11084 CRYPTO_dbg_set_options() [F]
11085 CRYPTO_dbg_get_options() [F]
11086 CRYPTO_malloc_debug_init() [M]
11088 The memory debug functions are NULL by default, unless the library
11089 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
11090 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
11091 gives the standard debugging functions that come with OpenSSL) or
11092 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
11093 provided by the library user) must be used. When the standard
11094 debugging functions are used, CRYPTO_dbg_set_options can be used to
11095 request additional information:
11096 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
11097 the CRYPTO_MDEBUG_xxx macro when compiling the library.
11099 Also, things like CRYPTO_set_mem_functions will always give the
11100 expected result (the new set of functions is used for allocation
11101 and deallocation) at all times, regardless of platform and compiler
11104 To finish it up, some functions that were never use in any other
11105 way than through macros have a new API and new semantic:
11107 CRYPTO_dbg_malloc()
11108 CRYPTO_dbg_realloc()
11111 All macros of value have retained their old syntax.
11112 [Richard Levitte and Bodo Moeller]
11114 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
11115 ordering of SMIMECapabilities wasn't in "strength order" and there
11116 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
11120 *) Some ASN1 types with illegal zero length encoding (INTEGER,
11121 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
11122 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
11124 *) Merge in my S/MIME library for OpenSSL. This provides a simple
11125 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
11126 functionality to handle multipart/signed properly) and a utility
11127 called 'smime' to call all this stuff. This is based on code I
11128 originally wrote for Celo who have kindly allowed it to be
11129 included in OpenSSL.
11132 *) Add variants des_set_key_checked and des_set_key_unchecked of
11133 des_set_key (aka des_key_sched). Global variable des_check_key
11134 decides which of these is called by des_set_key; this way
11135 des_check_key behaves as it always did, but applications and
11136 the library itself, which was buggy for des_check_key == 1,
11137 have a cleaner way to pick the version they need.
11140 *) New function PKCS12_newpass() which changes the password of a
11144 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
11145 dynamic mix. In both cases the ids can be used as an index into the
11146 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
11147 functions so they accept a list of the field values and the
11148 application doesn't need to directly manipulate the X509_TRUST
11152 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
11156 *) Modify the way the V3 extension code looks up extensions. This now
11157 works in a similar way to the object code: we have some "standard"
11158 extensions in a static table which is searched with OBJ_bsearch()
11159 and the application can add dynamic ones if needed. The file
11160 crypto/x509v3/ext_dat.h now has the info: this file needs to be
11161 updated whenever a new extension is added to the core code and kept
11162 in ext_nid order. There is a simple program 'tabtest.c' which checks
11163 this. New extensions are not added too often so this file can readily
11164 be maintained manually.
11166 There are two big advantages in doing things this way. The extensions
11167 can be looked up immediately and no longer need to be "added" using
11168 X509V3_add_standard_extensions(): this function now does nothing.
11169 [Side note: I get *lots* of email saying the extension code doesn't
11170 work because people forget to call this function]
11171 Also no dynamic allocation is done unless new extensions are added:
11172 so if we don't add custom extensions there is no need to call
11173 X509V3_EXT_cleanup().
11176 *) Modify enc utility's salting as follows: make salting the default. Add a
11177 magic header, so unsalted files fail gracefully instead of just decrypting
11178 to garbage. This is because not salting is a big security hole, so people
11179 should be discouraged from doing it.
11182 *) Fixes and enhancements to the 'x509' utility. It allowed a message
11183 digest to be passed on the command line but it only used this
11184 parameter when signing a certificate. Modified so all relevant
11185 operations are affected by the digest parameter including the
11186 -fingerprint and -x509toreq options. Also -x509toreq choked if a
11187 DSA key was used because it didn't fix the digest.
11190 *) Initial certificate chain verify code. Currently tests the untrusted
11191 certificates for consistency with the verify purpose (which is set
11192 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11194 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11195 this is because it will reject chains with invalid extensions whereas
11196 every previous version of OpenSSL and SSLeay made no checks at all.
11198 Trust code: checks the root CA for the relevant trust settings. Trust
11199 settings have an initial value consistent with the verify purpose: e.g.
11200 if the verify purpose is for SSL client use it expects the CA to be
11201 trusted for SSL client use. However the default value can be changed to
11202 permit custom trust settings: one example of this would be to only trust
11203 certificates from a specific "secure" set of CAs.
11205 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11206 which should be used for version portability: especially since the
11207 verify structure is likely to change more often now.
11209 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11210 to set them. If not set then assume SSL clients will verify SSL servers
11213 Two new options to the verify program: -untrusted allows a set of
11214 untrusted certificates to be passed in and -purpose which sets the
11215 intended purpose of the certificate. If a purpose is set then the
11216 new chain verify code is used to check extension consistency.
11219 *) Support for the authority information access extension.
11222 *) Modify RSA and DSA PEM read routines to transparently handle
11223 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11224 public keys in a format compatible with certificate
11225 SubjectPublicKeyInfo structures. Unfortunately there were already
11226 functions called *_PublicKey_* which used various odd formats so
11227 these are retained for compatibility: however the DSA variants were
11228 never in a public release so they have been deleted. Changed dsa/rsa
11229 utilities to handle the new format: note no releases ever handled public
11230 keys so we should be OK.
11232 The primary motivation for this change is to avoid the same fiasco
11233 that dogs private keys: there are several incompatible private key
11234 formats some of which are standard and some OpenSSL specific and
11235 require various evil hacks to allow partial transparent handling and
11236 even then it doesn't work with DER formats. Given the option anything
11237 other than PKCS#8 should be dumped: but the other formats have to
11238 stay in the name of compatibility.
11240 With public keys and the benefit of hindsight one standard format
11241 is used which works with EVP_PKEY, RSA or DSA structures: though
11242 it clearly returns an error if you try to read the wrong kind of key.
11244 Added a -pubkey option to the 'x509' utility to output the public key.
11245 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11246 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11247 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11248 that do the same as the EVP_PKEY_assign_*() except they up the
11249 reference count of the added key (they don't "swallow" the
11253 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11254 CRLs would fail if the file contained no certificates or no CRLs:
11255 added a new function to read in both types and return the number
11256 read: this means that if none are read it will be an error. The
11257 DER versions of the certificate and CRL reader would always fail
11258 because it isn't possible to mix certificates and CRLs in DER format
11259 without choking one or the other routine. Changed this to just read
11260 a certificate: this is the best we can do. Also modified the code
11261 in apps/verify.c to take notice of return codes: it was previously
11262 attempting to read in certificates from NULL pointers and ignoring
11263 any errors: this is one reason why the cert and CRL reader seemed
11264 to work. It doesn't check return codes from the default certificate
11265 routines: these may well fail if the certificates aren't installed.
11268 *) Code to support otherName option in GeneralName.
11271 *) First update to verify code. Change the verify utility
11272 so it warns if it is passed a self signed certificate:
11273 for consistency with the normal behaviour. X509_verify
11274 has been modified to it will now verify a self signed
11275 certificate if *exactly* the same certificate appears
11276 in the store: it was previously impossible to trust a
11277 single self signed certificate. This means that:
11278 openssl verify ss.pem
11279 now gives a warning about a self signed certificate but
11280 openssl verify -CAfile ss.pem ss.pem
11284 *) For servers, store verify_result in SSL_SESSION data structure
11285 (and add it to external session representation).
11286 This is needed when client certificate verifications fails,
11287 but an application-provided verification callback (set by
11288 SSL_CTX_set_cert_verify_callback) allows accepting the session
11289 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11290 but returns 1): When the session is reused, we have to set
11291 ssl->verify_result to the appropriate error code to avoid
11293 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11295 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11296 case in PKCS7_dataInit() where the signed PKCS7 structure
11297 didn't contain any existing data because it was being created.
11298 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11300 *) Add a salt to the key derivation routines in enc.c. This
11301 forms the first 8 bytes of the encrypted file. Also add a
11302 -S option to allow a salt to be input on the command line.
11305 *) New function X509_cmp(). Oddly enough there wasn't a function
11306 to compare two certificates. We do this by working out the SHA1
11307 hash and comparing that. X509_cmp() will be needed by the trust
11311 *) SSL_get1_session() is like SSL_get_session(), but increments
11312 the reference count in the SSL_SESSION returned.
11313 [Geoff Thorpe <geoff@eu.c2.net>]
11315 *) Fix for 'req': it was adding a null to request attributes.
11316 Also change the X509_LOOKUP and X509_INFO code to handle
11317 certificate auxiliary information.
11320 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11324 *) Add the possibility to add extra information to the memory leak
11325 detecting output, to form tracebacks, showing from where each
11326 allocation was originated: CRYPTO_push_info("constant string") adds
11327 the string plus current file name and line number to a per-thread
11328 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11329 is like calling CYRPTO_pop_info() until the stack is empty.
11330 Also updated memory leak detection code to be multi-thread-safe.
11333 *) Add options -text and -noout to pkcs7 utility and delete the
11334 encryption options which never did anything. Update docs.
11337 *) Add options to some of the utilities to allow the pass phrase
11338 to be included on either the command line (not recommended on
11339 OSes like Unix) or read from the environment. Update the
11340 manpages and fix a few bugs.
11343 *) Add a few manpages for some of the openssl commands.
11346 *) Fix the -revoke option in ca. It was freeing up memory twice,
11347 leaking and not finding already revoked certificates.
11350 *) Extensive changes to support certificate auxiliary information.
11351 This involves the use of X509_CERT_AUX structure and X509_AUX
11352 functions. An X509_AUX function such as PEM_read_X509_AUX()
11353 can still read in a certificate file in the usual way but it
11354 will also read in any additional "auxiliary information". By
11355 doing things this way a fair degree of compatibility can be
11356 retained: existing certificates can have this information added
11357 using the new 'x509' options.
11359 Current auxiliary information includes an "alias" and some trust
11360 settings. The trust settings will ultimately be used in enhanced
11361 certificate chain verification routines: currently a certificate
11362 can only be trusted if it is self signed and then it is trusted
11366 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
11367 The problem was that one of the replacement routines had not been working
11368 since SSLeay releases. For now the offending routine has been replaced
11369 with non-optimised assembler. Even so, this now gives around 95%
11370 performance improvement for 1024 bit RSA signs.
11373 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
11374 handling. Most clients have the effective key size in bits equal to
11375 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
11376 A few however don't do this and instead use the size of the decrypted key
11377 to determine the RC2 key length and the AlgorithmIdentifier to determine
11378 the effective key length. In this case the effective key length can still
11379 be 40 bits but the key length can be 168 bits for example. This is fixed
11380 by manually forcing an RC2 key into the EVP_PKEY structure because the
11381 EVP code can't currently handle unusual RC2 key sizes: it always assumes
11382 the key length and effective key length are equal.
11385 *) Add a bunch of functions that should simplify the creation of
11386 X509_NAME structures. Now you should be able to do:
11387 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
11388 and have it automatically work out the correct field type and fill in
11389 the structures. The more adventurous can try:
11390 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
11391 and it will (hopefully) work out the correct multibyte encoding.
11394 *) Change the 'req' utility to use the new field handling and multibyte
11395 copy routines. Before the DN field creation was handled in an ad hoc
11396 way in req, ca, and x509 which was rather broken and didn't support
11397 BMPStrings or UTF8Strings. Since some software doesn't implement
11398 BMPStrings or UTF8Strings yet, they can be enabled using the config file
11399 using the dirstring_type option. See the new comment in the default
11400 openssl.cnf for more info.
11403 *) Make crypto/rand/md_rand.c more robust:
11404 - Assure unique random numbers after fork().
11405 - Make sure that concurrent threads access the global counter and
11406 md serializably so that we never lose entropy in them
11407 or use exactly the same state in multiple threads.
11408 Access to the large state is not always serializable because
11409 the additional locking could be a performance killer, and
11410 md should be large enough anyway.
11413 *) New file apps/app_rand.c with commonly needed functionality
11414 for handling the random seed file.
11416 Use the random seed file in some applications that previously did not:
11418 dsaparam -genkey (which also ignored its '-rand' option),
11421 x509 (when signing).
11422 Except on systems with /dev/urandom, it is crucial to have a random
11423 seed file at least for key creation, DSA signing, and for DH exchanges;
11424 for RSA signatures we could do without one.
11426 gendh and gendsa (unlike genrsa) used to read only the first byte
11427 of each file listed in the '-rand' option. The function as previously
11428 found in genrsa is now in app_rand.c and is used by all programs
11429 that support '-rand'.
11432 *) In RAND_write_file, use mode 0600 for creating files;
11433 don't just chmod when it may be too late.
11436 *) Report an error from X509_STORE_load_locations
11437 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
11440 *) New function ASN1_mbstring_copy() this copies a string in either
11441 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
11442 into an ASN1_STRING type. A mask of permissible types is passed
11443 and it chooses the "minimal" type to use or an error if not type
11447 *) Add function equivalents to the various macros in asn1.h. The old
11448 macros are retained with an M_ prefix. Code inside the library can
11449 use the M_ macros. External code (including the openssl utility)
11450 should *NOT* in order to be "shared library friendly".
11453 *) Add various functions that can check a certificate's extensions
11454 to see if it usable for various purposes such as SSL client,
11455 server or S/MIME and CAs of these types. This is currently
11456 VERY EXPERIMENTAL but will ultimately be used for certificate chain
11457 verification. Also added a -purpose flag to x509 utility to
11458 print out all the purposes.
11461 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
11465 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
11466 for, obtain and decode and extension and obtain its critical flag.
11467 This allows all the necessary extension code to be handled in a
11468 single function call.
11471 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
11472 platforms. See crypto/rc4/rc4_enc.c for further details.
11475 *) New -noout option to asn1parse. This causes no output to be produced
11476 its main use is when combined with -strparse and -out to extract data
11477 from a file (which may not be in ASN.1 format).
11480 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
11481 when producing the local key id.
11482 [Richard Levitte <levitte@stacken.kth.se>]
11484 *) New option -dhparam in s_server. This allows a DH parameter file to be
11485 stated explicitly. If it is not stated then it tries the first server
11486 certificate file. The previous behaviour hard coded the filename
11490 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11491 a public key to be input or output. For example:
11492 openssl rsa -in key.pem -pubout -out pubkey.pem
11493 Also added necessary DSA public key functions to handle this.
11496 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11497 in the message. This was handled by allowing
11498 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11499 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11501 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11502 to the end of the strings whereas this didn't. This would cause problems
11503 if strings read with d2i_ASN1_bytes() were later modified.
11504 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11506 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11507 data and it contains EOF it will end up returning an error. This is
11508 caused by input 46 bytes long. The cause is due to the way base64
11509 BIOs find the start of base64 encoded data. They do this by trying a
11510 trial decode on each line until they find one that works. When they
11511 do a flag is set and it starts again knowing it can pass all the
11512 data directly through the decoder. Unfortunately it doesn't reset
11513 the context it uses. This means that if EOF is reached an attempt
11514 is made to pass two EOFs through the context and this causes the
11515 resulting error. This can also cause other problems as well. As is
11516 usual with these problems it takes *ages* to find and the fix is
11517 trivial: move one line.
11518 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
11520 *) Ugly workaround to get s_client and s_server working under Windows. The
11521 old code wouldn't work because it needed to select() on sockets and the
11522 tty (for keypresses and to see if data could be written). Win32 only
11523 supports select() on sockets so we select() with a 1s timeout on the
11524 sockets and then see if any characters are waiting to be read, if none
11525 are present then we retry, we also assume we can always write data to
11526 the tty. This isn't nice because the code then blocks until we've
11527 received a complete line of data and it is effectively polling the
11528 keyboard at 1s intervals: however it's quite a bit better than not
11529 working at all :-) A dedicated Windows application might handle this
11530 with an event loop for example.
11533 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
11534 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
11535 will be called when RSA_sign() and RSA_verify() are used. This is useful
11536 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
11537 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
11538 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
11539 This necessitated the support of an extra signature type NID_md5_sha1
11540 for SSL signatures and modifications to the SSL library to use it instead
11541 of calling RSA_public_decrypt() and RSA_private_encrypt().
11544 *) Add new -verify -CAfile and -CApath options to the crl program, these
11545 will lookup a CRL issuers certificate and verify the signature in a
11546 similar way to the verify program. Tidy up the crl program so it
11547 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
11548 less strict. It will now permit CRL extensions even if it is not
11549 a V2 CRL: this will allow it to tolerate some broken CRLs.
11552 *) Initialize all non-automatic variables each time one of the openssl
11553 sub-programs is started (this is necessary as they may be started
11554 multiple times from the "OpenSSL>" prompt).
11555 [Lennart Bang, Bodo Moeller]
11557 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
11558 removing all other RSA functionality (this is what NO_RSA does). This
11559 is so (for example) those in the US can disable those operations covered
11560 by the RSA patent while allowing storage and parsing of RSA keys and RSA
11564 *) Non-copying interface to BIO pairs.
11565 (still largely untested)
11568 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
11569 ASCII string. This was handled independently in various places before.
11572 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
11573 UTF8 strings a character at a time.
11576 *) Use client_version from client hello to select the protocol
11577 (s23_srvr.c) and for RSA client key exchange verification
11578 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
11581 *) Add various utility functions to handle SPKACs, these were previously
11582 handled by poking round in the structure internals. Added new function
11583 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
11584 print, verify and generate SPKACs. Based on an original idea from
11585 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
11588 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
11591 *) Allow the config file extension section to be overwritten on the
11592 command line. Based on an original idea from Massimiliano Pala
11593 <madwolf@comune.modena.it>. The new option is called -extensions
11594 and can be applied to ca, req and x509. Also -reqexts to override
11595 the request extensions in req and -crlexts to override the crl extensions
11599 *) Add new feature to the SPKAC handling in ca. Now you can include
11600 the same field multiple times by preceding it by "XXXX." for example:
11603 this is the same syntax as used in the req config file.
11606 *) Allow certificate extensions to be added to certificate requests. These
11607 are specified in a 'req_extensions' option of the req section of the
11608 config file. They can be printed out with the -text option to req but
11609 are otherwise ignored at present.
11612 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11613 data read consists of only the final block it would not decrypted because
11614 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11615 A misplaced 'break' also meant the decrypted final block might not be
11616 copied until the next read.
11619 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11620 a few extra parameters to the DH structure: these will be useful if
11621 for example we want the value of 'q' or implement X9.42 DH.
11624 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11625 provides hooks that allow the default DSA functions or functions on a
11626 "per key" basis to be replaced. This allows hardware acceleration and
11627 hardware key storage to be handled without major modification to the
11628 library. Also added low level modexp hooks and CRYPTO_EX structure and
11629 associated functions.
11632 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11633 as "read only": it can't be written to and the buffer it points to will
11634 not be freed. Reading from a read only BIO is much more efficient than
11635 a normal memory BIO. This was added because there are several times when
11636 an area of memory needs to be read from a BIO. The previous method was
11637 to create a memory BIO and write the data to it, this results in two
11638 copies of the data and an O(n^2) reading algorithm. There is a new
11639 function BIO_new_mem_buf() which creates a read only memory BIO from
11640 an area of memory. Also modified the PKCS#7 routines to use read only
11644 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11645 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11646 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11647 but a retry condition occurred while trying to read the rest.
11650 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11651 NID_pkcs7_encrypted by default: this was wrong since this should almost
11652 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11653 the encrypted data type: this is a more sensible place to put it and it
11654 allows the PKCS#12 code to be tidied up that duplicated this
11658 *) Changed obj_dat.pl script so it takes its input and output files on
11659 the command line. This should avoid shell escape redirection problems
11663 *) Initial support for certificate extension requests, these are included
11664 in things like Xenroll certificate requests. Included functions to allow
11665 extensions to be obtained and added.
11668 *) -crlf option to s_client and s_server for sending newlines as
11669 CRLF (as required by many protocols).
11672 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11674 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11675 [Ralf S. Engelschall]
11677 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11678 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11680 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11684 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11685 DH parameters/keys (q is lost during that conversion, but the resulting
11686 DH parameters contain its length).
11688 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11689 much faster than DH_generate_parameters (which creates parameters
11690 where p = 2*q + 1), and also the smaller q makes DH computations
11691 much more efficient (160-bit exponentiation instead of 1024-bit
11692 exponentiation); so this provides a convenient way to support DHE
11693 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11694 utter importance to use
11695 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11697 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11698 when such DH parameters are used, because otherwise small subgroup
11699 attacks may become possible!
11702 *) Avoid memory leak in i2d_DHparams.
11705 *) Allow the -k option to be used more than once in the enc program:
11706 this allows the same encrypted message to be read by multiple recipients.
11709 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11710 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11711 it will always use the numerical form of the OID, even if it has a short
11715 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11716 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11717 otherwise bn_mod_exp was called. In the case of hardware keys for example
11718 no private key components need be present and it might store extra data
11719 in the RSA structure, which cannot be accessed from bn_mod_exp.
11720 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11721 private key operations.
11724 *) Added support for SPARC Linux.
11727 *) pem_password_cb function type incompatibly changed from
11728 typedef int pem_password_cb(char *buf, int size, int rwflag);
11730 ....(char *buf, int size, int rwflag, void *userdata);
11731 so that applications can pass data to their callbacks:
11732 The PEM[_ASN1]_{read,write}... functions and macros now take an
11733 additional void * argument, which is just handed through whenever
11734 the password callback is called.
11735 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11737 New function SSL_CTX_set_default_passwd_cb_userdata.
11739 Compatibility note: As many C implementations push function arguments
11740 onto the stack in reverse order, the new library version is likely to
11741 interoperate with programs that have been compiled with the old
11742 pem_password_cb definition (PEM_whatever takes some data that
11743 happens to be on the stack as its last argument, and the callback
11744 just ignores this garbage); but there is no guarantee whatsoever that
11747 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11748 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11749 problems not only on Windows, but also on some Unix platforms.
11750 To avoid problematic command lines, these definitions are now in an
11751 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11752 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
11755 *) MIPS III/IV assembler module is reimplemented.
11758 *) More DES library cleanups: remove references to srand/rand and
11759 delete an unused file.
11762 *) Add support for the free Netwide assembler (NASM) under Win32,
11763 since not many people have MASM (ml) and it can be hard to obtain.
11764 This is currently experimental but it seems to work OK and pass all
11765 the tests. Check out INSTALL.W32 for info.
11768 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
11769 without temporary keys kept an extra copy of the server key,
11770 and connections with temporary keys did not free everything in case
11774 *) New function RSA_check_key and new openssl rsa option -check
11775 for verifying the consistency of RSA keys.
11776 [Ulf Moeller, Bodo Moeller]
11778 *) Various changes to make Win32 compile work:
11779 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
11780 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
11781 comparison" warnings.
11782 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
11785 *) Add a debugging option to PKCS#5 v2 key generation function: when
11786 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
11787 derived keys are printed to stderr.
11790 *) Copy the flags in ASN1_STRING_dup().
11791 [Roman E. Pavlov <pre@mo.msk.ru>]
11793 *) The x509 application mishandled signing requests containing DSA
11794 keys when the signing key was also DSA and the parameters didn't match.
11796 It was supposed to omit the parameters when they matched the signing key:
11797 the verifying software was then supposed to automatically use the CA's
11798 parameters if they were absent from the end user certificate.
11800 Omitting parameters is no longer recommended. The test was also
11801 the wrong way round! This was probably due to unusual behaviour in
11802 EVP_cmp_parameters() which returns 1 if the parameters match.
11803 This meant that parameters were omitted when they *didn't* match and
11804 the certificate was useless. Certificates signed with 'ca' didn't have
11806 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
11808 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
11809 The interface is as follows:
11810 Applications can use
11811 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
11812 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
11813 "off" is now the default.
11814 The library internally uses
11815 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
11816 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
11817 to disable memory-checking temporarily.
11819 Some inconsistent states that previously were possible (and were
11820 even the default) are now avoided.
11822 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
11823 with each memory chunk allocated; this is occasionally more helpful
11824 than just having a counter.
11826 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11828 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11832 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11833 which largely parallels "options", but is for changing API behaviour,
11834 whereas "options" are about protocol behaviour.
11835 Initial "mode" flags are:
11837 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
11838 a single record has been written.
11839 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
11840 retries use the same buffer location.
11841 (But all of the contents must be
11845 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
11848 *) Fix problems with no-hmac etc.
11849 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
11851 *) New functions RSA_get_default_method(), RSA_set_method() and
11852 RSA_get_method(). These allows replacement of RSA_METHODs without having
11853 to mess around with the internals of an RSA structure.
11856 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
11857 Also really enable memory leak checks in openssl.c and in some
11859 [Chad C. Mulligan, Bodo Moeller]
11861 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
11862 up the length of negative integers. This has now been simplified to just
11863 store the length when it is first determined and use it later, rather
11864 than trying to keep track of where data is copied and updating it to
11866 [Steve Henson, reported by Brien Wheeler
11867 <bwheeler@authentica-security.com>]
11869 *) Add a new function PKCS7_signatureVerify. This allows the verification
11870 of a PKCS#7 signature but with the signing certificate passed to the
11871 function itself. This contrasts with PKCS7_dataVerify which assumes the
11872 certificate is present in the PKCS#7 structure. This isn't always the
11873 case: certificates can be omitted from a PKCS#7 structure and be
11874 distributed by "out of band" means (such as a certificate database).
11877 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
11878 function prototypes in pem.h, also change util/mkdef.pl to add the
11879 necessary function names.
11882 *) mk1mf.pl (used by Windows builds) did not properly read the
11883 options set by Configure in the top level Makefile, and Configure
11884 was not even able to write more than one option correctly.
11885 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
11888 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
11889 file to be loaded from a BIO or FILE pointer. The BIO version will
11890 for example allow memory BIOs to contain config info.
11893 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
11894 Whoever hopes to achieve shared-library compatibility across versions
11895 must use this, not the compile-time macro.
11896 (Exercise 0.9.4: Which is the minimum library version required by
11898 Note: All this applies only to multi-threaded programs, others don't
11902 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
11903 through a BIO pair triggered the default case, i.e.
11904 SSLerr(...,SSL_R_UNKNOWN_STATE).
11907 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
11908 can use the SSL library even if none of the specific BIOs is
11912 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
11913 for the encoded length.
11914 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
11916 *) Add initial documentation of the X509V3 functions.
11919 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
11920 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
11921 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
11922 secure PKCS#8 private key format with a high iteration count.
11925 *) Fix determination of Perl interpreter: A perl or perl5
11926 _directory_ in $PATH was also accepted as the interpreter.
11927 [Ralf S. Engelschall]
11929 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
11930 wrong with it but it was very old and did things like calling
11931 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
11932 unusual formatting.
11935 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
11936 to use the new extension code.
11939 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
11940 with macros. This should make it easier to change their form, add extra
11941 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
11945 *) Add to configuration table a new entry that can specify an alternative
11946 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
11947 according to Mark Crispin <MRC@Panda.COM>.
11951 *) DES CBC did not update the IV. Weird.
11954 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
11955 Changing the behaviour of the former might break existing programs --
11956 where IV updating is needed, des_ncbc_encrypt can be used.
11959 *) When bntest is run from "make test" it drives bc to check its
11960 calculations, as well as internally checking them. If an internal check
11961 fails, it needs to cause bc to give a non-zero result or make test carries
11962 on without noticing the failure. Fixed.
11965 *) DES library cleanups.
11968 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
11969 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
11970 ciphers. NOTE: although the key derivation function has been verified
11971 against some published test vectors it has not been extensively tested
11972 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
11976 *) Instead of "mkdir -p", which is not fully portable, use new
11977 Perl script "util/mkdir-p.pl".
11980 *) Rewrite the way password based encryption (PBE) is handled. It used to
11981 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
11982 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
11983 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
11984 the 'parameter' field of the AlgorithmIdentifier is passed to the
11985 underlying key generation function so it must do its own ASN1 parsing.
11986 This has also changed the EVP_PBE_CipherInit() function which now has a
11987 'parameter' argument instead of literal salt and iteration count values
11988 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
11991 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
11992 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
11993 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
11994 KEY" because this clashed with PKCS#8 unencrypted string. Since this
11995 value was just used as a "magic string" and not used directly its
11996 value doesn't matter.
11999 *) Introduce some semblance of const correctness to BN. Shame C doesn't
12003 *) "linux-sparc64" configuration (ultrapenguin).
12004 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
12005 "linux-sparc" configuration.
12006 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
12008 *) config now generates no-xxx options for missing ciphers.
12011 *) Support the EBCDIC character set (work in progress).
12012 File ebcdic.c not yet included because it has a different license.
12013 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12015 *) Support BS2000/OSD-POSIX.
12016 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12018 *) Make callbacks for key generation use void * instead of char *.
12021 *) Make S/MIME samples compile (not yet tested).
12024 *) Additional typesafe stacks.
12027 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
12031 Changes between 0.9.3 and 0.9.3a [29 May 1999]
12033 *) New configuration variant "sco5-gcc".
12035 *) Updated some demos.
12036 [Sean O Riordain, Wade Scholine]
12038 *) Add missing BIO_free at exit of pkcs12 application.
12041 *) Fix memory leak in conf.c.
12044 *) Updates for Win32 to assembler version of MD5.
12047 *) Set #! path to perl in apps/der_chop to where we found it
12048 instead of using a fixed path.
12051 *) SHA library changes for irix64-mips4-cc.
12054 *) Improvements for VMS support.
12058 Changes between 0.9.2b and 0.9.3 [24 May 1999]
12060 *) Bignum library bug fix. IRIX 6 passes "make test" now!
12061 This also avoids the problems with SC4.2 and unpatched SC5.
12062 [Andy Polyakov <appro@fy.chalmers.se>]
12064 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
12065 These are required because of the typesafe stack would otherwise break
12066 existing code. If old code used a structure member which used to be STACK
12067 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
12068 sk_num or sk_value it would produce an error because the num, data members
12069 are not present in STACK_OF. Now it just produces a warning. sk_set
12070 replaces the old method of assigning a value to sk_value
12071 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
12072 that does this will no longer work (and should use sk_set instead) but
12073 this could be regarded as a "questionable" behaviour anyway.
12076 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
12077 correctly handle encrypted S/MIME data.
12080 *) Change type of various DES function arguments from des_cblock
12081 (which means, in function argument declarations, pointer to char)
12082 to des_cblock * (meaning pointer to array with 8 char elements),
12083 which allows the compiler to do more typechecking; it was like
12084 that back in SSLeay, but with lots of ugly casts.
12086 Introduce new type const_des_cblock.
12089 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
12090 problems: find RecipientInfo structure that matches recipient certificate
12091 and initialise the ASN1 structures properly based on passed cipher.
12094 *) Belatedly make the BN tests actually check the results.
12097 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
12098 to and from BNs: it was completely broken. New compilation option
12099 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
12100 key elements as negative integers.
12103 *) Reorganize and speed up MD5.
12104 [Andy Polyakov <appro@fy.chalmers.se>]
12107 [Richard Levitte <richard@levitte.org>]
12109 *) New option -out to asn1parse to allow the parsed structure to be
12110 output to a file. This is most useful when combined with the -strparse
12111 option to examine the output of things like OCTET STRINGS.
12114 *) Make SSL library a little more fool-proof by not requiring any longer
12115 that SSL_set_{accept,connect}_state be called before
12116 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
12117 in many applications because usually everything *appeared* to work as
12118 intended anyway -- now it really works as intended).
12121 *) Move openssl.cnf out of lib/.
12124 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
12125 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
12126 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
12127 [Ralf S. Engelschall]
12129 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
12130 handle PKCS#7 enveloped data properly.
12131 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
12133 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
12134 copying pointers. The cert_st handling is changed by this in
12135 various ways (and thus what used to be known as ctx->default_cert
12136 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
12137 any longer when s->cert does not give us what we need).
12138 ssl_cert_instantiate becomes obsolete by this change.
12139 As soon as we've got the new code right (possibly it already is?),
12140 we have solved a couple of bugs of the earlier code where s->cert
12141 was used as if it could not have been shared with other SSL structures.
12143 Note that using the SSL API in certain dirty ways now will result
12144 in different behaviour than observed with earlier library versions:
12145 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
12146 does not influence s as it used to.
12148 In order to clean up things more thoroughly, inside SSL_SESSION
12149 we don't use CERT any longer, but a new structure SESS_CERT
12150 that holds per-session data (if available); currently, this is
12151 the peer's certificate chain and, for clients, the server's certificate
12152 and temporary key. CERT holds only those values that can have
12153 meaningful defaults in an SSL_CTX.
12156 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
12157 from the internal representation. Various PKCS#7 fixes: remove some
12158 evil casts and set the enc_dig_alg field properly based on the signing
12162 *) Allow PKCS#12 password to be set from the command line or the
12163 environment. Let 'ca' get its config file name from the environment
12164 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
12168 *) Allow certificate policies extension to use an IA5STRING for the
12169 organization field. This is contrary to the PKIX definition but
12170 VeriSign uses it and IE5 only recognises this form. Document 'x509'
12174 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
12175 without disallowing inline assembler and the like for non-pedantic builds.
12178 *) Support Borland C++ builder.
12179 [Janez Jere <jj@void.si>, modified by Ulf Möller]
12181 *) Support Mingw32.
12184 *) SHA-1 cleanups and performance enhancements.
12185 [Andy Polyakov <appro@fy.chalmers.se>]
12187 *) Sparc v8plus assembler for the bignum library.
12188 [Andy Polyakov <appro@fy.chalmers.se>]
12190 *) Accept any -xxx and +xxx compiler options in Configure.
12193 *) Update HPUX configuration.
12196 *) Add missing sk_<type>_unshift() function to safestack.h
12197 [Ralf S. Engelschall]
12199 *) New function SSL_CTX_use_certificate_chain_file that sets the
12200 "extra_cert"s in addition to the certificate. (This makes sense
12201 only for "PEM" format files, as chains as a whole are not
12205 *) Support verify_depth from the SSL API.
12206 x509_vfy.c had what can be considered an off-by-one-error:
12207 Its depth (which was not part of the external interface)
12208 was actually counting the number of certificates in a chain;
12209 now it really counts the depth.
12212 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12213 instead of X509err, which often resulted in confusing error
12214 messages since the error codes are not globally unique
12215 (e.g. an alleged error in ssl3_accept when a certificate
12216 didn't match the private key).
12218 *) New function SSL_CTX_set_session_id_context that allows to set a default
12219 value (so that you don't need SSL_set_session_id_context for each
12220 connection using the SSL_CTX).
12223 *) OAEP decoding bug fix.
12226 *) Support INSTALL_PREFIX for package builders, as proposed by
12230 *) New Configure options "threads" and "no-threads". For systems
12231 where the proper compiler options are known (currently Solaris
12232 and Linux), "threads" is the default.
12235 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12238 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12239 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12240 such as /usr/local/bin.
12243 *) "make linux-shared" to build shared libraries.
12244 [Niels Poppe <niels@netbox.org>]
12246 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12249 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12250 extension adding in x509 utility.
12253 *) Remove NOPROTO sections and error code comments.
12256 *) Partial rewrite of the DEF file generator to now parse the ANSI
12260 *) New Configure options --prefix=DIR and --openssldir=DIR.
12263 *) Complete rewrite of the error code script(s). It is all now handled
12264 by one script at the top level which handles error code gathering,
12265 header rewriting and C source file generation. It should be much better
12266 than the old method: it now uses a modified version of Ulf's parser to
12267 read the ANSI prototypes in all header files (thus the old K&R definitions
12268 aren't needed for error creation any more) and do a better job of
12269 translating function codes into names. The old 'ASN1 error code imbedded
12270 in a comment' is no longer necessary and it doesn't use .err files which
12271 have now been deleted. Also the error code call doesn't have to appear all
12272 on one line (which resulted in some large lines...).
12275 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12278 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12279 0 (which usually indicates a closed connection), but continue reading.
12282 *) Fix some race conditions.
12285 *) Add support for CRL distribution points extension. Add Certificate
12286 Policies and CRL distribution points documentation.
12289 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12292 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12293 8 of keying material. Merlin has also confirmed interop with this fix
12294 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12295 [Merlin Hughes <merlin@baltimore.ie>]
12297 *) Fix lots of warnings.
12298 [Richard Levitte <levitte@stacken.kth.se>]
12300 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12301 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12302 [Richard Levitte <levitte@stacken.kth.se>]
12304 *) Fix problems with sizeof(long) == 8.
12305 [Andy Polyakov <appro@fy.chalmers.se>]
12307 *) Change functions to ANSI C.
12310 *) Fix typos in error codes.
12311 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12313 *) Remove defunct assembler files from Configure.
12316 *) SPARC v8 assembler BIGNUM implementation.
12317 [Andy Polyakov <appro@fy.chalmers.se>]
12319 *) Support for Certificate Policies extension: both print and set.
12320 Various additions to support the r2i method this uses.
12323 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12324 return a const string when you are expecting an allocated buffer.
12327 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12328 types DirectoryString and DisplayText.
12331 *) Add code to allow r2i extensions to access the configuration database,
12332 add an LHASH database driver and add several ctx helper functions.
12335 *) Fix an evil bug in bn_expand2() which caused various BN functions to
12336 fail when they extended the size of a BIGNUM.
12339 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
12340 support typesafe stack.
12343 *) Fix typo in SSL_[gs]et_options().
12344 [Nils Frostberg <nils@medcom.se>]
12346 *) Delete various functions and files that belonged to the (now obsolete)
12347 old X509V3 handling code.
12350 *) New Configure option "rsaref".
12353 *) Don't auto-generate pem.h.
12356 *) Introduce type-safe ASN.1 SETs.
12359 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
12360 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
12362 *) Introduce type-safe STACKs. This will almost certainly break lots of code
12363 that links with OpenSSL (well at least cause lots of warnings), but fear
12364 not: the conversion is trivial, and it eliminates loads of evil casts. A
12365 few STACKed things have been converted already. Feel free to convert more.
12366 In the fullness of time, I'll do away with the STACK type altogether.
12369 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
12370 specified in <certfile> by updating the entry in the index.txt file.
12371 This way one no longer has to edit the index.txt file manually for
12372 revoking a certificate. The -revoke option does the gory details now.
12373 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
12375 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
12376 `-text' option at all and this way the `-noout -text' combination was
12377 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
12378 [Ralf S. Engelschall]
12380 *) Make sure a corresponding plain text error message exists for the
12381 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
12382 verify callback function determined that a certificate was revoked.
12383 [Ralf S. Engelschall]
12385 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
12386 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
12387 all available ciphers including rc5, which was forgotten until now.
12388 In order to let the testing shell script know which algorithms
12389 are available, a new (up to now undocumented) command
12390 "openssl list-cipher-commands" is used.
12393 *) Bugfix: s_client occasionally would sleep in select() when
12394 it should have checked SSL_pending() first.
12397 *) New functions DSA_do_sign and DSA_do_verify to provide access to
12398 the raw DSA values prior to ASN.1 encoding.
12401 *) Tweaks to Configure
12402 [Niels Poppe <niels@netbox.org>]
12404 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
12408 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
12411 *) New config option to avoid instructions that are illegal on the 80386.
12412 The default code is faster, but requires at least a 486.
12415 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
12416 SSL2_SERVER_VERSION (not used at all) macros, which are now the
12417 same as SSL2_VERSION anyway.
12420 *) New "-showcerts" option for s_client.
12423 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
12424 application. Various cleanups and fixes.
12427 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
12428 modify error routines to work internally. Add error codes and PBE init
12429 to library startup routines.
12432 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
12433 packing functions to asn1 and evp. Changed function names and error
12434 codes along the way.
12437 *) PKCS12 integration: and so it begins... First of several patches to
12438 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
12439 objects to objects.h
12442 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
12443 and display support for Thawte strong extranet extension.
12446 *) Add LinuxPPC support.
12447 [Jeff Dubrule <igor@pobox.org>]
12449 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
12450 bn_div_words in alpha.s.
12451 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
12453 *) Make sure the RSA OAEP test is skipped under -DRSAref because
12454 OAEP isn't supported when OpenSSL is built with RSAref.
12455 [Ulf Moeller <ulf@fitug.de>]
12457 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
12458 so they no longer are missing under -DNOPROTO.
12459 [Soren S. Jorvang <soren@t.dk>]
12462 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
12464 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
12465 doesn't work when the session is reused. Coming soon!
12468 *) Fix a security hole, that allows sessions to be reused in the wrong
12469 context thus bypassing client cert protection! All software that uses
12470 client certs and session caches in multiple contexts NEEDS PATCHING to
12471 allow session reuse! A fuller solution is in the works.
12472 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
12474 *) Some more source tree cleanups (removed obsolete files
12475 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
12476 permission on "config" script to be executable) and a fix for the INSTALL
12478 [Ulf Moeller <ulf@fitug.de>]
12480 *) Remove some legacy and erroneous uses of malloc, free instead of
12482 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
12484 *) Make rsa_oaep_test return non-zero on error.
12485 [Ulf Moeller <ulf@fitug.de>]
12487 *) Add support for native Solaris shared libraries. Configure
12488 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12489 if someone would make that last step automatic.
12490 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12492 *) ctx_size was not built with the right compiler during "make links". Fixed.
12495 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12496 except NULL ciphers". This means the default cipher list will no longer
12497 enable NULL ciphers. They need to be specifically enabled e.g. with
12498 the string "DEFAULT:eNULL".
12501 *) Fix to RSA private encryption routines: if p < q then it would
12502 occasionally produce an invalid result. This will only happen with
12503 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12506 *) Be less restrictive and allow also `perl util/perlpath.pl
12507 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12508 because this way one can also use an interpreter named `perl5' (which is
12509 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12510 installed as `perl').
12511 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12513 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12514 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12516 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12517 advapi32.lib to Win32 build and change the pem test comparison
12518 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12519 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
12520 and crypto/des/ede_cbcm_enc.c.
12523 *) DES quad checksum was broken on big-endian architectures. Fixed.
12526 *) Comment out two functions in bio.h that aren't implemented. Fix up the
12527 Win32 test batch file so it (might) work again. The Win32 test batch file
12528 is horrible: I feel ill....
12531 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
12532 in e_os.h. Audit of header files to check ANSI and non ANSI
12533 sections: 10 functions were absent from non ANSI section and not exported
12534 from Windows DLLs. Fixed up libeay.num for new functions.
12537 *) Make `openssl version' output lines consistent.
12538 [Ralf S. Engelschall]
12540 *) Fix Win32 symbol export lists for BIO functions: Added
12541 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
12542 to ms/libeay{16,32}.def.
12543 [Ralf S. Engelschall]
12545 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
12546 fine under Unix and passes some trivial tests I've now added. But the
12547 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
12548 added to make sure no one expects that this stuff really works in the
12549 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
12550 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
12552 [Ralf S. Engelschall]
12554 *) Fix the generation of two part addresses in perl.
12555 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
12557 *) Add config entry for Linux on MIPS.
12558 [John Tobey <jtobey@channel1.com>]
12560 *) Make links whenever Configure is run, unless we are on Windoze.
12563 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
12564 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
12568 *) Add a useful kludge to allow package maintainers to specify compiler and
12569 other platforms details on the command line without having to patch the
12570 Configure script everytime: One now can use ``perl Configure
12571 <id>:<details>'', i.e. platform ids are allowed to have details appended
12572 to them (separated by colons). This is treated as there would be a static
12573 pre-configured entry in Configure's %table under key <id> with value
12574 <details> and ``perl Configure <id>'' is called. So, when you want to
12575 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
12576 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
12577 now, which overrides the FreeBSD-elf entry on-the-fly.
12578 [Ralf S. Engelschall]
12580 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
12583 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
12584 on the `perl Configure ...' command line. This way one can compile
12585 OpenSSL libraries with Position Independent Code (PIC) which is needed
12586 for linking it into DSOs.
12587 [Ralf S. Engelschall]
12589 *) Remarkably, export ciphers were totally broken and no-one had noticed!
12593 *) Cleaned up the LICENSE document: The official contact for any license
12594 questions now is the OpenSSL core team under openssl-core@openssl.org.
12595 And add a paragraph about the dual-license situation to make sure people
12596 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
12597 to the OpenSSL toolkit.
12598 [Ralf S. Engelschall]
12600 *) General source tree makefile cleanups: Made `making xxx in yyy...'
12601 display consistent in the source tree and replaced `/bin/rm' by `rm'.
12602 Additionally cleaned up the `make links' target: Remove unnecessary
12603 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
12604 to speed processing and no longer clutter the display with confusing
12605 stuff. Instead only the actually done links are displayed.
12606 [Ralf S. Engelschall]
12608 *) Permit null encryption ciphersuites, used for authentication only. It used
12609 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
12610 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12614 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12615 signed attributes when verifying signatures (this would break them),
12616 the detached data encoding was wrong and public keys obtained using
12617 X509_get_pubkey() weren't freed.
12620 *) Add text documentation for the BUFFER functions. Also added a work around
12621 to a Win95 console bug. This was triggered by the password read stuff: the
12622 last character typed gets carried over to the next fread(). If you were
12623 generating a new cert request using 'req' for example then the last
12624 character of the passphrase would be CR which would then enter the first
12628 *) Added the new `Includes OpenSSL Cryptography Software' button as
12629 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12630 button and can be used by applications based on OpenSSL to show the
12631 relationship to the OpenSSL project.
12632 [Ralf S. Engelschall]
12634 *) Remove confusing variables in function signatures in files
12635 ssl/ssl_lib.c and ssl/ssl.h.
12636 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12638 *) Don't install bss_file.c under PREFIX/include/
12639 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12641 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12642 functions that return function pointers and has support for NT specific
12643 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12644 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12645 unsigned to signed types: this was killing the Win32 compile.
12648 *) Add new certificate file to stack functions,
12649 SSL_add_dir_cert_subjects_to_stack() and
12650 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12651 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12652 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12653 This means that Apache-SSL and similar packages don't have to mess around
12654 to add as many CAs as they want to the preferred list.
12657 *) Experiment with doxygen documentation. Currently only partially applied to
12659 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12660 openssl.doxy as the configuration file.
12663 *) Get rid of remaining C++-style comments which strict C compilers hate.
12664 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12666 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12667 compiled in by default: it has problems with large keys.
12670 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12671 DH private keys and/or callback functions which directly correspond to
12672 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12673 is needed for applications which have to configure certificates on a
12674 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12676 For the RSA certificate situation is makes no difference, but
12677 for the DSA certificate situation this fixes the "no shared cipher"
12678 problem where the OpenSSL cipher selection procedure failed because the
12679 temporary keys were not overtaken from the context and the API provided
12680 no way to reconfigure them.
12681 The new functions now let applications reconfigure the stuff and they
12682 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12683 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12684 non-public-API function ssl_cert_instantiate() is used as a helper
12685 function and also to reduce code redundancy inside ssl_rsa.c.
12686 [Ralf S. Engelschall]
12688 *) Move s_server -dcert and -dkey options out of the undocumented feature
12689 area because they are useful for the DSA situation and should be
12690 recognized by the users.
12691 [Ralf S. Engelschall]
12693 *) Fix the cipher decision scheme for export ciphers: the export bits are
12694 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12695 SSL_EXP_MASK. So, the original variable has to be used instead of the
12696 already masked variable.
12697 [Richard Levitte <levitte@stacken.kth.se>]
12699 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12700 [Richard Levitte <levitte@stacken.kth.se>]
12702 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12703 from `int' to `unsigned int' because it's a length and initialized by
12704 EVP_DigestFinal() which expects an `unsigned int *'.
12705 [Richard Levitte <levitte@stacken.kth.se>]
12707 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12708 script. Instead use the usual Shell->Perl transition trick.
12709 [Ralf S. Engelschall]
12711 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12712 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12713 -noout -modulus' as it's already the case for `openssl rsa -noout
12714 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12715 currently the public key is printed (a decision which was already done by
12716 `openssl dsa -modulus' in the past) which serves a similar purpose.
12717 Additionally the NO_RSA no longer completely removes the whole -modulus
12718 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12720 [Ralf S. Engelschall]
12722 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12723 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12724 [Arne Ansper <arne@ats.cyber.ee>]
12726 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12727 to be added. Now both 'req' and 'ca' can use new objects defined in the
12731 *) Add cool BIO that does syslog (or event log on NT).
12732 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12734 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12735 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12736 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12737 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12740 *) Add preliminary config info for new extension code.
12743 *) Make RSA_NO_PADDING really use no padding.
12744 [Ulf Moeller <ulf@fitug.de>]
12746 *) Generate errors when private/public key check is done.
12749 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12750 for some CRL extensions and new objects added.
12753 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
12754 key usage extension and fuller support for authority key id.
12757 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
12758 padding method for RSA, which is recommended for new applications in PKCS
12759 #1 v2.0 (RFC 2437, October 1998).
12760 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
12761 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
12762 against Bleichbacher's attack on RSA.
12763 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
12766 *) Updates to the new SSL compression code
12767 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12769 *) Fix so that the version number in the master secret, when passed
12770 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
12771 (because the server will not accept higher), that the version number
12772 is 0x03,0x01, not 0x03,0x00
12773 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12775 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
12776 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
12777 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
12780 *) Support for RAW extensions where an arbitrary extension can be
12781 created by including its DER encoding. See apps/openssl.cnf for
12785 *) Make sure latest Perl versions don't interpret some generated C array
12786 code as Perl array code in the crypto/err/err_genc.pl script.
12787 [Lars Weber <3weber@informatik.uni-hamburg.de>]
12789 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
12790 not many people have the assembler. Various Win32 compilation fixes and
12791 update to the INSTALL.W32 file with (hopefully) more accurate Win32
12792 build instructions.
12795 *) Modify configure script 'Configure' to automatically create crypto/date.h
12796 file under Win32 and also build pem.h from pem.org. New script
12797 util/mkfiles.pl to create the MINFO file on environments that can't do a
12798 'make files': perl util/mkfiles.pl >MINFO should work.
12801 *) Major rework of DES function declarations, in the pursuit of correctness
12802 and purity. As a result, many evil casts evaporated, and some weirdness,
12803 too. You may find this causes warnings in your code. Zapping your evil
12804 casts will probably fix them. Mostly.
12807 *) Fix for a typo in asn1.h. Bug fix to object creation script
12808 obj_dat.pl. It considered a zero in an object definition to mean
12809 "end of object": none of the objects in objects.h have any zeros
12810 so it wasn't spotted.
12811 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
12813 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
12814 Masking (CBCM). In the absence of test vectors, the best I have been able
12815 to do is check that the decrypt undoes the encrypt, so far. Send me test
12816 vectors if you have them.
12819 *) Correct calculation of key length for export ciphers (too much space was
12820 allocated for null ciphers). This has not been tested!
12823 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
12824 message is now correct (it understands "crypto" and "ssl" on its
12825 command line). There is also now an "update" option. This will update
12826 the util/ssleay.num and util/libeay.num files with any new functions.
12828 perl util/mkdef.pl crypto ssl update
12829 it will update them.
12832 *) Overhauled the Perl interface (perl/*):
12833 - ported BN stuff to OpenSSL's different BN library
12834 - made the perl/ source tree CVS-aware
12835 - renamed the package from SSLeay to OpenSSL (the files still contain
12836 their history because I've copied them in the repository)
12837 - removed obsolete files (the test scripts will be replaced
12838 by better Test::Harness variants in the future)
12839 [Ralf S. Engelschall]
12841 *) First cut for a very conservative source tree cleanup:
12842 1. merge various obsolete readme texts into doc/ssleay.txt
12843 where we collect the old documents and readme texts.
12844 2. remove the first part of files where I'm already sure that we no
12845 longer need them because of three reasons: either they are just temporary
12846 files which were left by Eric or they are preserved original files where
12847 I've verified that the diff is also available in the CVS via "cvs diff
12848 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
12849 the crypto/md/ stuff).
12850 [Ralf S. Engelschall]
12852 *) More extension code. Incomplete support for subject and issuer alt
12853 name, issuer and authority key id. Change the i2v function parameters
12854 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
12855 what that's for :-) Fix to ASN1 macro which messed up
12856 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
12859 *) Preliminary support for ENUMERATED type. This is largely copied from the
12863 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
12864 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12866 *) Make sure `make rehash' target really finds the `openssl' program.
12867 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12869 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
12870 like to hear about it if this slows down other processors.
12873 *) Add CygWin32 platform information to Configure script.
12874 [Alan Batie <batie@aahz.jf.intel.com>]
12876 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
12877 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
12879 *) New program nseq to manipulate netscape certificate sequences
12882 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
12886 *) Fixes to BN code. Previously the default was to define BN_RECURSION
12887 but the BN code had some problems that would cause failures when
12888 doing certificate verification and some other functions.
12889 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12891 *) Add ASN1 and PEM code to support netscape certificate sequences.
12894 *) Add ASN1 and PEM code to support netscape certificate sequences.
12897 *) Add several PKIX and private extended key usage OIDs.
12900 *) Modify the 'ca' program to handle the new extension code. Modify
12901 openssl.cnf for new extension format, add comments.
12904 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
12905 and add a sample to openssl.cnf so req -x509 now adds appropriate
12909 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
12910 error code, add initial support to X509_print() and x509 application.
12913 *) Takes a deep breath and start adding X509 V3 extension support code. Add
12914 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
12915 stuff is currently isolated and isn't even compiled yet.
12918 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
12919 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
12920 Removed the versions check from X509 routines when loading extensions:
12921 this allows certain broken certificates that don't set the version
12922 properly to be processed.
12925 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
12926 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
12927 can still be regenerated with "make depend".
12930 *) Spelling mistake in C version of CAST-128.
12931 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
12933 *) Changes to the error generation code. The perl script err-code.pl
12934 now reads in the old error codes and retains the old numbers, only
12935 adding new ones if necessary. It also only changes the .err files if new
12936 codes are added. The makefiles have been modified to only insert errors
12937 when needed (to avoid needlessly modifying header files). This is done
12938 by only inserting errors if the .err file is newer than the auto generated
12939 C file. To rebuild all the error codes from scratch (the old behaviour)
12940 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
12941 or delete all the .err files.
12944 *) CAST-128 was incorrectly implemented for short keys. The C version has
12945 been fixed, but is untested. The assembler versions are also fixed, but
12946 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
12947 to regenerate it if needed.
12948 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
12949 Hagino <itojun@kame.net>]
12951 *) File was opened incorrectly in randfile.c.
12952 [Ulf Möller <ulf@fitug.de>]
12954 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
12955 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
12956 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
12957 al: it's just almost always a UTCTime. Note this patch adds new error
12958 codes so do a "make errors" if there are problems.
12961 *) Correct Linux 1 recognition in config.
12962 [Ulf Möller <ulf@fitug.de>]
12964 *) Remove pointless MD5 hash when using DSA keys in ca.
12965 [Anonymous <nobody@replay.com>]
12967 *) Generate an error if given an empty string as a cert directory. Also
12968 generate an error if handed NULL (previously returned 0 to indicate an
12969 error, but didn't set one).
12970 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
12972 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
12975 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
12976 parameters. This was causing a warning which killed off the Win32 compile.
12979 *) Remove C++ style comments from crypto/bn/bn_local.h.
12980 [Neil Costigan <neil.costigan@celocom.com>]
12982 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
12983 based on a text string, looking up short and long names and finally
12984 "dot" format. The "dot" format stuff didn't work. Added new function
12985 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
12986 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
12987 OID is not part of the table.
12990 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
12991 X509_LOOKUP_by_alias().
12994 *) Sort openssl functions by name.
12997 *) Get the gendsa program working (hopefully) and add it to app list. Remove
12998 encryption from sample DSA keys (in case anyone is interested the password
13002 *) Make _all_ *_free functions accept a NULL pointer.
13003 [Frans Heymans <fheymans@isaserver.be>]
13005 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
13007 [Anonymous <nobody@replay.com>]
13009 *) s_server should send the CAfile as acceptable CAs, not its own cert.
13010 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13012 *) Don't blow it for numeric -newkey arguments to apps/req.
13013 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13015 *) Temp key "for export" tests were wrong in s3_srvr.c.
13016 [Anonymous <nobody@replay.com>]
13018 *) Add prototype for temp key callback functions
13019 SSL_CTX_set_tmp_{rsa,dh}_callback().
13022 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
13023 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
13026 *) X509_name_add_entry() freed the wrong thing after an error.
13027 [Arne Ansper <arne@ats.cyber.ee>]
13029 *) rsa_eay.c would attempt to free a NULL context.
13030 [Arne Ansper <arne@ats.cyber.ee>]
13032 *) BIO_s_socket() had a broken should_retry() on Windoze.
13033 [Arne Ansper <arne@ats.cyber.ee>]
13035 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
13036 [Arne Ansper <arne@ats.cyber.ee>]
13038 *) Make sure the already existing X509_STORE->depth variable is initialized
13039 in X509_STORE_new(), but document the fact that this variable is still
13040 unused in the certificate verification process.
13041 [Ralf S. Engelschall]
13043 *) Fix the various library and apps files to free up pkeys obtained from
13044 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
13047 *) Fix reference counting in X509_PUBKEY_get(). This makes
13048 demos/maurice/example2.c work, amongst others, probably.
13049 [Steve Henson and Ben Laurie]
13051 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
13052 `openssl' and second, the shortcut symlinks for the `openssl <command>'
13053 are no longer created. This way we have a single and consistent command
13054 line interface `openssl <command>', similar to `cvs <command>'.
13055 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
13057 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
13058 BIT STRING wrapper always have zero unused bits.
13061 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
13064 *) Make the top-level INSTALL documentation easier to understand.
13067 *) Makefiles updated to exit if an error occurs in a sub-directory
13068 make (including if user presses ^C) [Paul Sutton]
13070 *) Make Montgomery context stuff explicit in RSA data structure.
13073 *) Fix build order of pem and err to allow for generated pem.h.
13076 *) Fix renumbering bug in X509_NAME_delete_entry().
13079 *) Enhanced the err-ins.pl script so it makes the error library number
13080 global and can add a library name. This is needed for external ASN1 and
13081 other error libraries.
13084 *) Fixed sk_insert which never worked properly.
13087 *) Fix ASN1 macros so they can handle indefinite length constructed
13088 EXPLICIT tags. Some non standard certificates use these: they can now
13092 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
13093 into a single doc/ssleay.txt bundle. This way the information is still
13094 preserved but no longer messes up this directory. Now it's new room for
13095 the new set of documentation files.
13096 [Ralf S. Engelschall]
13098 *) SETs were incorrectly DER encoded. This was a major pain, because they
13099 shared code with SEQUENCEs, which aren't coded the same. This means that
13100 almost everything to do with SETs or SEQUENCEs has either changed name or
13101 number of arguments.
13102 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
13104 *) Fix test data to work with the above.
13107 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
13108 was already fixed by Eric for 0.9.1 it seems.
13109 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
13111 *) Autodetect FreeBSD3.
13114 *) Fix various bugs in Configure. This affects the following platforms:
13118 unixware-2.0-pentium
13122 *) Eliminate generated files from CVS. Reorder tests to regenerate files
13123 before they are needed.
13126 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
13130 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
13132 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
13133 changed SSLeay to OpenSSL in version strings.
13134 [Ralf S. Engelschall]
13136 *) Some fixups to the top-level documents.
13139 *) Fixed the nasty bug where rsaref.h was not found under compile-time
13140 because the symlink to include/ was missing.
13141 [Ralf S. Engelschall]
13143 *) Incorporated the popular no-RSA/DSA-only patches
13144 which allow to compile a RSA-free SSLeay.
13145 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
13147 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
13148 when "ssleay" is still not found.
13149 [Ralf S. Engelschall]
13151 *) Added more platforms to Configure: Cray T3E, HPUX 11,
13152 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
13154 *) Updated the README file.
13155 [Ralf S. Engelschall]
13157 *) Added various .cvsignore files in the CVS repository subdirs
13158 to make a "cvs update" really silent.
13159 [Ralf S. Engelschall]
13161 *) Recompiled the error-definition header files and added
13162 missing symbols to the Win32 linker tables.
13163 [Ralf S. Engelschall]
13165 *) Cleaned up the top-level documents;
13166 o new files: CHANGES and LICENSE
13167 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
13168 o merged COPYRIGHT into LICENSE
13169 o removed obsolete TODO file
13170 o renamed MICROSOFT to INSTALL.W32
13171 [Ralf S. Engelschall]
13173 *) Removed dummy files from the 0.9.1b source tree:
13174 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
13175 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
13176 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
13177 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
13178 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
13179 [Ralf S. Engelschall]
13181 *) Added various platform portability fixes.
13184 *) The Genesis of the OpenSSL rpject:
13185 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
13186 Young and Tim J. Hudson created while they were working for C2Net until
13188 [The OpenSSL Project]
13191 Changes between 0.9.0b and 0.9.1b [not released]
13193 *) Updated a few CA certificates under certs/
13196 *) Changed some BIGNUM api stuff.
13199 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13200 DGUX x86, Linux Alpha, etc.
13203 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13204 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13208 *) Add -strparse option to asn1pars program which parses nested
13210 [Dr Stephen Henson <shenson@bigfoot.com>]
13212 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13215 *) DSA fix for "ca" program.
13218 *) Added "-genkey" option to "dsaparam" program.
13221 *) Added RIPE MD160 (rmd160) message digest.
13224 *) Added -a (all) option to "ssleay version" command.
13227 *) Added PLATFORM define which is the id given to Configure.
13230 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13233 *) Extended the ASN.1 parser routines.
13236 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13239 *) Added a BN_CTX to the BN library.
13242 *) Fixed the weak key values in DES library
13245 *) Changed API in EVP library for cipher aliases.
13248 *) Added support for RC2/64bit cipher.
13251 *) Converted the lhash library to the crypto/mem.c functions.
13254 *) Added more recognized ASN.1 object ids.
13257 *) Added more RSA padding checks for SSL/TLS.
13260 *) Added BIO proxy/filter functionality.
13263 *) Added extra_certs to SSL_CTX which can be used
13264 send extra CA certificates to the client in the CA cert chain sending
13265 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13268 *) Now Fortezza is denied in the authentication phase because
13269 this is key exchange mechanism is not supported by SSLeay at all.
13272 *) Additional PKCS1 checks.
13275 *) Support the string "TLSv1" for all TLS v1 ciphers.
13278 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13279 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13282 *) Fixed a few memory leaks.
13285 *) Fixed various code and comment typos.
13288 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13289 bytes sent in the client random.
13290 [Edward Bishop <ebishop@spyglass.com>]