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 *) Correct the extended master secret constant on EBCDIC systems. Without this
13 fix TLS connections between an EBCDIC system and a non-EBCDIC system that
14 negotiate EMS will fail. Unfortunately this also means that TLS connections
15 between EBCDIC systems with this fix, and EBCDIC systems without this
16 fix will fail if they negotiate EMS.
19 *) Use Windows installation paths in the mingw builds
21 Mingw isn't a POSIX environment per se, which means that Windows
22 paths should be used for installation.
26 *) Changed DH parameters to generate the order q subgroup instead of 2q.
27 Previously generated DH parameters are still accepted by DH_check
28 but DH_generate_key works around that by clearing bit 0 of the
29 private key for those. This avoids leaking bit 0 of the private key.
32 *) Significantly reduce secure memory usage by the randomness pools.
35 *) Revert the DEVRANDOM_WAIT feature for Linux systems
37 The DEVRANDOM_WAIT feature added a select() call to wait for the
38 /dev/random device to become readable before reading from the
41 It turned out that this change had negative side effects on
42 performance which were not acceptable. After some discussion it
43 was decided to revert this feature and leave it up to the OS
44 resp. the platform maintainer to ensure a proper initialization
45 during early boot time.
47 Changes between 1.1.1b and 1.1.1c [28 May 2019]
49 *) Add build tests for C++. These are generated files that only do one
50 thing, to include one public OpenSSL head file each. This tests that
51 the public header files can be usefully included in a C++ application.
53 This test isn't enabled by default. It can be enabled with the option
54 'enable-buildtest-c++'.
57 *) Enable SHA3 pre-hashing for ECDSA and DSA.
60 *) Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
61 This changes the size when using the genpkey app when no size is given. It
62 fixes an omission in earlier changes that changed all RSA, DSA and DH
63 generation apps to use 2048 bits by default.
66 *) Reorganize the manual pages to consistently have RETURN VALUES,
67 EXAMPLES, SEE ALSO and HISTORY come in that order, and adjust
68 util/fix-doc-nits accordingly.
69 [Paul Yang, Joshua Lock]
71 *) Add the missing accessor EVP_PKEY_get0_engine()
74 *) Have apps like 's_client' and 's_server' output the signature scheme
75 along with other cipher suite parameters when debugging.
78 *) Make OPENSSL_config() error agnostic again.
81 *) Do the error handling in RSA decryption constant time.
84 *) Prevent over long nonces in ChaCha20-Poly1305.
86 ChaCha20-Poly1305 is an AEAD cipher, and requires a unique nonce input
87 for every encryption operation. RFC 7539 specifies that the nonce value
88 (IV) should be 96 bits (12 bytes). OpenSSL allows a variable nonce length
89 and front pads the nonce with 0 bytes if it is less than 12
90 bytes. However it also incorrectly allows a nonce to be set of up to 16
91 bytes. In this case only the last 12 bytes are significant and any
92 additional leading bytes are ignored.
94 It is a requirement of using this cipher that nonce values are
95 unique. Messages encrypted using a reused nonce value are susceptible to
96 serious confidentiality and integrity attacks. If an application changes
97 the default nonce length to be longer than 12 bytes and then makes a
98 change to the leading bytes of the nonce expecting the new value to be a
99 new unique nonce then such an application could inadvertently encrypt
100 messages with a reused nonce.
102 Additionally the ignored bytes in a long nonce are not covered by the
103 integrity guarantee of this cipher. Any application that relies on the
104 integrity of these ignored leading bytes of a long nonce may be further
105 affected. Any OpenSSL internal use of this cipher, including in SSL/TLS,
106 is safe because no such use sets such a long nonce value. However user
107 applications that use this cipher directly and set a non-default nonce
108 length to be longer than 12 bytes may be vulnerable.
110 This issue was reported to OpenSSL on 16th of March 2019 by Joran Dirk
115 *) Add DEVRANDOM_WAIT feature for Linux systems
117 On older Linux systems where the getrandom() system call is not available,
118 OpenSSL normally uses the /dev/urandom device for seeding its CSPRNG.
119 Contrary to getrandom(), the /dev/urandom device will not block during
120 early boot when the kernel CSPRNG has not been seeded yet.
122 To mitigate this known weakness, use select() to wait for /dev/random to
123 become readable before reading from /dev/urandom.
125 *) Ensure that SM2 only uses SM3 as digest algorithm
128 Changes between 1.1.1a and 1.1.1b [26 Feb 2019]
130 *) Added SCA hardening for modular field inversion in EC_GROUP through
131 a new dedicated field_inv() pointer in EC_METHOD.
132 This also addresses a leakage affecting conversions from projective
133 to affine coordinates.
134 [Billy Bob Brumley, Nicola Tuveri]
136 *) Change the info callback signals for the start and end of a post-handshake
137 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
138 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
139 confused by this and assume that a TLSv1.2 renegotiation has started. This
140 can break KeyUpdate handling. Instead we no longer signal the start and end
141 of a post handshake message exchange (although the messages themselves are
142 still signalled). This could break some applications that were expecting
143 the old signals. However without this KeyUpdate is not usable for many
147 *) Fix a bug in the computation of the endpoint-pair shared secret used
148 by DTLS over SCTP. This breaks interoperability with older versions
149 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
150 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
151 interoperability with such broken implementations. However, enabling
152 this switch breaks interoperability with correct implementations.
154 *) Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
155 re-used X509_PUBKEY object if the second PUBKEY is malformed.
158 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
161 *) Remove the 'dist' target and add a tarball building script. The
162 'dist' target has fallen out of use, and it shouldn't be
163 necessary to configure just to create a source distribution.
166 Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
168 *) Timing vulnerability in DSA signature generation
170 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
171 timing side channel attack. An attacker could use variations in the signing
172 algorithm to recover the private key.
174 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
178 *) Timing vulnerability in ECDSA signature generation
180 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
181 timing side channel attack. An attacker could use variations in the signing
182 algorithm to recover the private key.
184 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
188 *) Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
189 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
190 are retained for backwards compatibility.
193 *) Fixed the issue that RAND_add()/RAND_seed() silently discards random input
194 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
195 of two gigabytes and the error handling improved.
197 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
198 categorized as a normal bug, not a security issue, because the DRBG reseeds
199 automatically and is fully functional even without additional randomness
200 provided by the application.
202 Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
204 *) Add a new ClientHello callback. Provides a callback interface that gives
205 the application the ability to adjust the nascent SSL object at the
206 earliest stage of ClientHello processing, immediately after extensions have
207 been collected but before they have been processed. In particular, this
208 callback can adjust the supported TLS versions in response to the contents
212 *) Add SM2 base algorithm support.
215 *) s390x assembly pack: add (improved) hardware-support for the following
216 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
217 aes-cfb/cfb8, aes-ecb.
220 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
221 parameter is no longer accepted, as it leads to a corrupt table. NULL
222 pem_str is reserved for alias entries only.
225 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
226 step for prime curves. The new implementation is based on formulae from
227 differential addition-and-doubling in homogeneous projective coordinates
228 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
229 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
230 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
231 to work in projective coordinates.
232 [Billy Bob Brumley, Nicola Tuveri]
234 *) Change generating and checking of primes so that the error rate of not
235 being prime depends on the intended use based on the size of the input.
236 For larger primes this will result in more rounds of Miller-Rabin.
237 The maximal error rate for primes with more than 1080 bits is lowered
239 [Kurt Roeckx, Annie Yousar]
241 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
244 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
245 moving between systems, and to avoid confusion when a Windows build is
246 done with mingw vs with MSVC. For POSIX installs, there's still a
247 symlink or copy named 'tsget' to avoid that confusion as well.
250 *) Revert blinding in ECDSA sign and instead make problematic addition
251 length-invariant. Switch even to fixed-length Montgomery multiplication.
254 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
255 step for binary curves. The new implementation is based on formulae from
256 differential addition-and-doubling in mixed Lopez-Dahab projective
257 coordinates, modified to independently blind the operands.
258 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
260 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
261 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
262 EC_METHODs to implement their own specialized "ladder step", to take
263 advantage of more favorable coordinate systems or more efficient
264 differential addition-and-doubling algorithms.
265 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
267 *) Modified the random device based seed sources to keep the relevant
268 file descriptors open rather than reopening them on each access.
269 This allows such sources to operate in a chroot() jail without
270 the associated device nodes being available. This behaviour can be
271 controlled using RAND_keep_random_devices_open().
274 *) Numerous side-channel attack mitigations have been applied. This may have
275 performance impacts for some algorithms for the benefit of improved
276 security. Specific changes are noted in this change log by their respective
280 *) AIX shared library support overhaul. Switch to AIX "natural" way of
281 handling shared libraries, which means collecting shared objects of
282 different versions and bitnesses in one common archive. This allows to
283 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
284 doesn't affect the way 3rd party applications are linked, only how
285 multi-version installation is managed.
288 *) Make ec_group_do_inverse_ord() more robust and available to other
289 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
290 mitigations are applied to the fallback BN_mod_inverse().
291 When using this function rather than BN_mod_inverse() directly, new
292 EC cryptosystem implementations are then safer-by-default.
295 *) Add coordinate blinding for EC_POINT and implement projective
296 coordinate blinding for generic prime curves as a countermeasure to
297 chosen point SCA attacks.
298 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
300 *) Add blinding to ECDSA and DSA signatures to protect against side channel
301 attacks discovered by Keegan Ryan (NCC Group).
304 *) Enforce checking in the pkeyutl command line app to ensure that the input
305 length does not exceed the maximum supported digest length when performing
306 a sign, verify or verifyrecover operation.
309 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
310 I/O in combination with something like select() or poll() will hang. This
311 can be turned off again using SSL_CTX_clear_mode().
312 Many applications do not properly handle non-application data records, and
313 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
314 around the problems in those applications, but can also break some.
315 It's recommended to read the manpages about SSL_read(), SSL_write(),
316 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
317 SSL_CTX_set_read_ahead() again.
320 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
321 now allow empty (zero character) pass phrases.
324 *) Apply blinding to binary field modular inversion and remove patent
325 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
328 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
329 binary and prime elliptic curves.
332 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
333 constant time fixed point multiplication.
336 *) Revise elliptic curve scalar multiplication with timing attack
337 defenses: ec_wNAF_mul redirects to a constant time implementation
338 when computing fixed point and variable point multiplication (which
339 in OpenSSL are mostly used with secret scalars in keygen, sign,
340 ECDH derive operations).
341 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
344 *) Updated CONTRIBUTING
347 *) Updated DRBG / RAND to request nonce and additional low entropy
348 randomness from the system.
349 [Matthias St. Pierre]
351 *) Updated 'openssl rehash' to use OpenSSL consistent default.
354 *) Moved the load of the ssl_conf module to libcrypto, which helps
355 loading engines that libssl uses before libssl is initialised.
358 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
361 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
362 [Ingo Schwarze, Rich Salz]
364 *) Added output of accepting IP address and port for 'openssl s_server'
367 *) Added a new API for TLSv1.3 ciphersuites:
368 SSL_CTX_set_ciphersuites()
369 SSL_set_ciphersuites()
372 *) Memory allocation failures consistently add an error to the error
376 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
377 in libcrypto when run as setuid/setgid.
380 *) Load any config file by default when libssl is used.
383 *) Added new public header file <openssl/rand_drbg.h> and documentation
384 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
385 [Matthias St. Pierre]
387 *) QNX support removed (cannot find contributors to get their approval
388 for the license change).
391 *) TLSv1.3 replay protection for early data has been implemented. See the
392 SSL_read_early_data() man page for further details.
395 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
396 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
397 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
398 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
399 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
400 configuration has been separated out. See the ciphers man page or the
401 SSL_CTX_set_ciphersuites() man page for more information.
404 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
405 in responder mode now supports the new "-multi" option, which
406 spawns the specified number of child processes to handle OCSP
407 requests. The "-timeout" option now also limits the OCSP
408 responder's patience to wait to receive the full client request
409 on a newly accepted connection. Child processes are respawned
410 as needed, and the CA index file is automatically reloaded
411 when changed. This makes it possible to run the "ocsp" responder
412 as a long-running service, making the OpenSSL CA somewhat more
413 feature-complete. In this mode, most diagnostic messages logged
414 after entering the event loop are logged via syslog(3) rather than
418 *) Added support for X448 and Ed448. Heavily based on original work by
422 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
423 objects loaded. This adds the functions OSSL_STORE_expect() and
424 OSSL_STORE_find() as well as needed tools to construct searches and
425 get the search data out of them.
428 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
429 version of OpenSSL should review their configuration settings to ensure
430 that they are still appropriate for TLSv1.3. For further information see:
431 https://wiki.openssl.org/index.php/TLS1.3
434 *) Grand redesign of the OpenSSL random generator
436 The default RAND method now utilizes an AES-CTR DRBG according to
437 NIST standard SP 800-90Ar1. The new random generator is essentially
438 a port of the default random generator from the OpenSSL FIPS 2.0
439 object module. It is a hybrid deterministic random bit generator
440 using an AES-CTR bit stream and which seeds and reseeds itself
441 automatically using trusted system entropy sources.
443 Some of its new features are:
444 o Support for multiple DRBG instances with seed chaining.
445 o The default RAND method makes use of a DRBG.
446 o There is a public and private DRBG instance.
447 o The DRBG instances are fork-safe.
448 o Keep all global DRBG instances on the secure heap if it is enabled.
449 o The public and private DRBG instance are per thread for lock free
451 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
453 *) Changed Configure so it only says what it does and doesn't dump
454 so much data. Instead, ./configdata.pm should be used as a script
455 to display all sorts of configuration data.
458 *) Added processing of "make variables" to Configure.
461 *) Added SHA512/224 and SHA512/256 algorithm support.
464 *) The last traces of Netware support, first removed in 1.1.0, have
468 *) Get rid of Makefile.shared, and in the process, make the processing
469 of certain files (rc.obj, or the .def/.map/.opt files produced from
470 the ordinal files) more visible and hopefully easier to trace and
471 debug (or make silent).
474 *) Make it possible to have environment variable assignments as
475 arguments to config / Configure.
478 *) Add multi-prime RSA (RFC 8017) support.
481 *) Add SM3 implemented according to GB/T 32905-2016
482 [ Jack Lloyd <jack.lloyd@ribose.com>,
483 Ronald Tse <ronald.tse@ribose.com>,
484 Erick Borsboom <erick.borsboom@ribose.com> ]
486 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
487 as documented in RFC6066.
488 Based on a patch from Tomasz Moń
489 [Filipe Raimundo da Silva]
491 *) Add SM4 implemented according to GB/T 32907-2016.
492 [ Jack Lloyd <jack.lloyd@ribose.com>,
493 Ronald Tse <ronald.tse@ribose.com>,
494 Erick Borsboom <erick.borsboom@ribose.com> ]
496 *) Reimplement -newreq-nodes and ERR_error_string_n; the
497 original author does not agree with the license change.
500 *) Add ARIA AEAD TLS support.
503 *) Some macro definitions to support VS6 have been removed. Visual
504 Studio 6 has not worked since 1.1.0
507 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
508 without clearing the errors.
511 *) Add "atfork" functions. If building on a system that without
512 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
513 requirements. The RAND facility now uses/requires this.
519 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
520 not possible to disable entirely. However, it's still possible to
521 disable the console reading UI method, UI_OpenSSL() (use UI_null()
524 To disable, configure with 'no-ui-console'. 'no-ui' is still
525 possible to use as an alias. Check at compile time with the
526 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
527 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
530 *) Add a STORE module, which implements a uniform and URI based reader of
531 stores that can contain keys, certificates, CRLs and numerous other
532 objects. The main API is loosely based on a few stdio functions,
533 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
534 OSSL_STORE_error and OSSL_STORE_close.
535 The implementation uses backends called "loaders" to implement arbitrary
536 URI schemes. There is one built in "loader" for the 'file' scheme.
539 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
540 then adjusted to work on FreeBSD 8.4 as well.
541 Enable by configuring with 'enable-devcryptoeng'. This is done by default
542 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
545 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
546 util/mkerr.pl, which is adapted to allow those prefixes, leading to
547 error code calls like this:
549 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
551 With this change, we claim the namespaces OSSL and OPENSSL in a manner
552 that can be encoded in C. For the foreseeable future, this will only
554 [Richard Levitte and Tim Hudson]
556 *) Removed BSD cryptodev engine.
559 *) Add a build target 'build_all_generated', to build all generated files
560 and only that. This can be used to prepare everything that requires
561 things like perl for a system that lacks perl and then move everything
562 to that system and do the rest of the build there.
565 *) In the UI interface, make it possible to duplicate the user data. This
566 can be used by engines that need to retain the data for a longer time
567 than just the call where this user data is passed.
570 *) Ignore the '-named_curve auto' value for compatibility of applications
572 [Tomas Mraz <tmraz@fedoraproject.org>]
574 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
575 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
576 alerts across multiple records (some of which could be empty). In practice
577 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
578 prohibts this altogether and other libraries (BoringSSL, NSS) do not
579 support this at all. Supporting it adds significant complexity to the
580 record layer, and its removal is unlikely to cause inter-operability
584 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
585 with Z. These are meant to replace LONG and ZLONG and to be size safe.
586 The use of LONG and ZLONG is discouraged and scheduled for deprecation
590 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
591 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
592 [Richard Levitte, Andy Polyakov]
594 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
598 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
599 platform rather than 'mingw'.
602 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
603 success if they are asked to add an object which already exists
604 in the store. This change cascades to other functions which load
605 certificates and CRLs.
608 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
609 facilitate stack unwinding even from assembly subroutines.
612 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
613 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
616 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
617 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
618 which is the minimum version we support.
621 *) Certificate time validation (X509_cmp_time) enforces stricter
622 compliance with RFC 5280. Fractional seconds and timezone offsets
623 are no longer allowed.
626 *) Add support for ARIA
629 *) s_client will now send the Server Name Indication (SNI) extension by
630 default unless the new "-noservername" option is used. The server name is
631 based on the host provided to the "-connect" option unless overridden by
635 *) Add support for SipHash
638 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
639 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
640 prevent issues where no progress is being made and the peer continually
641 sends unrecognised record types, using up resources processing them.
644 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
645 using the algorithm defined in
646 https://www.akkadia.org/drepper/SHA-crypt.txt
649 *) Heartbeat support has been removed; the ABI is changed for now.
650 [Richard Levitte, Rich Salz]
652 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
655 *) The RSA "null" method, which was partially supported to avoid patent
656 issues, has been replaced to always returns NULL.
660 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
662 *) Client DoS due to large DH parameter
664 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
665 malicious server can send a very large prime value to the client. This will
666 cause the client to spend an unreasonably long period of time generating a
667 key for this prime resulting in a hang until the client has finished. This
668 could be exploited in a Denial Of Service attack.
670 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
674 *) Cache timing vulnerability in RSA Key Generation
676 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
677 a cache timing side channel attack. An attacker with sufficient access to
678 mount cache timing attacks during the RSA key generation process could
679 recover the private key.
681 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
682 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
686 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
687 parameter is no longer accepted, as it leads to a corrupt table. NULL
688 pem_str is reserved for alias entries only.
691 *) Revert blinding in ECDSA sign and instead make problematic addition
692 length-invariant. Switch even to fixed-length Montgomery multiplication.
695 *) Change generating and checking of primes so that the error rate of not
696 being prime depends on the intended use based on the size of the input.
697 For larger primes this will result in more rounds of Miller-Rabin.
698 The maximal error rate for primes with more than 1080 bits is lowered
700 [Kurt Roeckx, Annie Yousar]
702 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
705 *) Add blinding to ECDSA and DSA signatures to protect against side channel
706 attacks discovered by Keegan Ryan (NCC Group).
709 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
710 now allow empty (zero character) pass phrases.
713 *) Certificate time validation (X509_cmp_time) enforces stricter
714 compliance with RFC 5280. Fractional seconds and timezone offsets
715 are no longer allowed.
718 *) Fixed a text canonicalisation bug in CMS
720 Where a CMS detached signature is used with text content the text goes
721 through a canonicalisation process first prior to signing or verifying a
722 signature. This process strips trailing space at the end of lines, converts
723 line terminators to CRLF and removes additional trailing line terminators
724 at the end of a file. A bug in the canonicalisation process meant that
725 some characters, such as form-feed, were incorrectly treated as whitespace
726 and removed. This is contrary to the specification (RFC5485). This fix
727 could mean that detached text data signed with an earlier version of
728 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
729 signed with a fixed OpenSSL may fail to verify with an earlier version of
730 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
731 and use the "-binary" flag (for the "cms" command line application) or set
732 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
735 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
737 *) Constructed ASN.1 types with a recursive definition could exceed the stack
739 Constructed ASN.1 types with a recursive definition (such as can be found
740 in PKCS7) could eventually exceed the stack given malicious input with
741 excessive recursion. This could result in a Denial Of Service attack. There
742 are no such structures used within SSL/TLS that come from untrusted sources
743 so this is considered safe.
745 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
750 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
752 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
753 effectively reduced to only comparing the least significant bit of each
754 byte. This allows an attacker to forge messages that would be considered as
755 authenticated in an amount of tries lower than that guaranteed by the
756 security claims of the scheme. The module can only be compiled by the
757 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
759 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
764 *) Add a build target 'build_all_generated', to build all generated files
765 and only that. This can be used to prepare everything that requires
766 things like perl for a system that lacks perl and then move everything
767 to that system and do the rest of the build there.
770 *) Backport SSL_OP_NO_RENGOTIATION
772 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
773 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
774 changes this is no longer possible in 1.1.0. Therefore the new
775 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
776 1.1.0 to provide equivalent functionality.
778 Note that if an application built against 1.1.0h headers (or above) is run
779 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
780 accepted but nothing will happen, i.e. renegotiation will not be prevented.
783 *) Removed the OS390-Unix config target. It relied on a script that doesn't
787 *) rsaz_1024_mul_avx2 overflow bug on x86_64
789 There is an overflow bug in the AVX2 Montgomery multiplication procedure
790 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
791 Analysis suggests that attacks against RSA and DSA as a result of this
792 defect would be very difficult to perform and are not believed likely.
793 Attacks against DH1024 are considered just feasible, because most of the
794 work necessary to deduce information about a private key may be performed
795 offline. The amount of resources required for such an attack would be
796 significant. However, for an attack on TLS to be meaningful, the server
797 would have to share the DH1024 private key among multiple clients, which is
798 no longer an option since CVE-2016-0701.
800 This only affects processors that support the AVX2 but not ADX extensions
801 like Intel Haswell (4th generation).
803 This issue was reported to OpenSSL by David Benjamin (Google). The issue
804 was originally found via the OSS-Fuzz project.
808 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
810 *) bn_sqrx8x_internal carry bug on x86_64
812 There is a carry propagating bug in the x86_64 Montgomery squaring
813 procedure. No EC algorithms are affected. Analysis suggests that attacks
814 against RSA and DSA as a result of this defect would be very difficult to
815 perform and are not believed likely. Attacks against DH are considered just
816 feasible (although very difficult) because most of the work necessary to
817 deduce information about a private key may be performed offline. The amount
818 of resources required for such an attack would be very significant and
819 likely only accessible to a limited number of attackers. An attacker would
820 additionally need online access to an unpatched system using the target
821 private key in a scenario with persistent DH parameters and a private
822 key that is shared between multiple clients.
824 This only affects processors that support the BMI1, BMI2 and ADX extensions
825 like Intel Broadwell (5th generation) and later or AMD Ryzen.
827 This issue was reported to OpenSSL by the OSS-Fuzz project.
831 *) Malformed X.509 IPAddressFamily could cause OOB read
833 If an X.509 certificate has a malformed IPAddressFamily extension,
834 OpenSSL could do a one-byte buffer overread. The most likely result
835 would be an erroneous display of the certificate in text format.
837 This issue was reported to OpenSSL by the OSS-Fuzz project.
841 Changes between 1.1.0e and 1.1.0f [25 May 2017]
843 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
844 platform rather than 'mingw'.
847 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
848 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
849 which is the minimum version we support.
852 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
854 *) Encrypt-Then-Mac renegotiation crash
856 During a renegotiation handshake if the Encrypt-Then-Mac extension is
857 negotiated where it was not in the original handshake (or vice-versa) then
858 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
859 and servers are affected.
861 This issue was reported to OpenSSL by Joe Orton (Red Hat).
865 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
867 *) Truncated packet could crash via OOB read
869 If one side of an SSL/TLS path is running on a 32-bit host and a specific
870 cipher is being used, then a truncated packet can cause that host to
871 perform an out-of-bounds read, usually resulting in a crash.
873 This issue was reported to OpenSSL by Robert Święcki of Google.
877 *) Bad (EC)DHE parameters cause a client crash
879 If a malicious server supplies bad parameters for a DHE or ECDHE key
880 exchange then this can result in the client attempting to dereference a
881 NULL pointer leading to a client crash. This could be exploited in a Denial
884 This issue was reported to OpenSSL by Guido Vranken.
888 *) BN_mod_exp may produce incorrect results on x86_64
890 There is a carry propagating bug in the x86_64 Montgomery squaring
891 procedure. No EC algorithms are affected. Analysis suggests that attacks
892 against RSA and DSA as a result of this defect would be very difficult to
893 perform and are not believed likely. Attacks against DH are considered just
894 feasible (although very difficult) because most of the work necessary to
895 deduce information about a private key may be performed offline. The amount
896 of resources required for such an attack would be very significant and
897 likely only accessible to a limited number of attackers. An attacker would
898 additionally need online access to an unpatched system using the target
899 private key in a scenario with persistent DH parameters and a private
900 key that is shared between multiple clients. For example this can occur by
901 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
902 similar to CVE-2015-3193 but must be treated as a separate problem.
904 This issue was reported to OpenSSL by the OSS-Fuzz project.
908 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
910 *) ChaCha20/Poly1305 heap-buffer-overflow
912 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
913 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
914 crash. This issue is not considered to be exploitable beyond a DoS.
916 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
920 *) CMS Null dereference
922 Applications parsing invalid CMS structures can crash with a NULL pointer
923 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
924 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
925 structure callback if an attempt is made to free certain invalid encodings.
926 Only CHOICE structures using a callback which do not handle NULL value are
929 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
933 *) Montgomery multiplication may produce incorrect results
935 There is a carry propagating bug in the Broadwell-specific Montgomery
936 multiplication procedure that handles input lengths divisible by, but
937 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
938 and DH private keys are impossible. This is because the subroutine in
939 question is not used in operations with the private key itself and an input
940 of the attacker's direct choice. Otherwise the bug can manifest itself as
941 transient authentication and key negotiation failures or reproducible
942 erroneous outcome of public-key operations with specially crafted input.
943 Among EC algorithms only Brainpool P-512 curves are affected and one
944 presumably can attack ECDH key negotiation. Impact was not analyzed in
945 detail, because pre-requisites for attack are considered unlikely. Namely
946 multiple clients have to choose the curve in question and the server has to
947 share the private key among them, neither of which is default behaviour.
948 Even then only clients that chose the curve will be affected.
950 This issue was publicly reported as transient failures and was not
951 initially recognized as a security issue. Thanks to Richard Morgan for
952 providing reproducible case.
956 *) Removed automatic addition of RPATH in shared libraries and executables,
957 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
960 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
962 *) Fix Use After Free for large message sizes
964 The patch applied to address CVE-2016-6307 resulted in an issue where if a
965 message larger than approx 16k is received then the underlying buffer to
966 store the incoming message is reallocated and moved. Unfortunately a
967 dangling pointer to the old location is left which results in an attempt to
968 write to the previously freed location. This is likely to result in a
969 crash, however it could potentially lead to execution of arbitrary code.
971 This issue only affects OpenSSL 1.1.0a.
973 This issue was reported to OpenSSL by Robert Święcki.
977 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
979 *) OCSP Status Request extension unbounded memory growth
981 A malicious client can send an excessively large OCSP Status Request
982 extension. If that client continually requests renegotiation, sending a
983 large OCSP Status Request extension each time, then there will be unbounded
984 memory growth on the server. This will eventually lead to a Denial Of
985 Service attack through memory exhaustion. Servers with a default
986 configuration are vulnerable even if they do not support OCSP. Builds using
987 the "no-ocsp" build time option are not affected.
989 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
993 *) SSL_peek() hang on empty record
995 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
996 sends an empty record. This could be exploited by a malicious peer in a
997 Denial Of Service attack.
999 This issue was reported to OpenSSL by Alex Gaynor.
1003 *) Excessive allocation of memory in tls_get_message_header() and
1004 dtls1_preprocess_fragment()
1006 A (D)TLS message includes 3 bytes for its length in the header for the
1007 message. This would allow for messages up to 16Mb in length. Messages of
1008 this length are excessive and OpenSSL includes a check to ensure that a
1009 peer is sending reasonably sized messages in order to avoid too much memory
1010 being consumed to service a connection. A flaw in the logic of version
1011 1.1.0 means that memory for the message is allocated too early, prior to
1012 the excessive message length check. Due to way memory is allocated in
1013 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
1014 to service a connection. This could lead to a Denial of Service through
1015 memory exhaustion. However, the excessive message length check still takes
1016 place, and this would cause the connection to immediately fail. Assuming
1017 that the application calls SSL_free() on the failed connection in a timely
1018 manner then the 21Mb of allocated memory will then be immediately freed
1019 again. Therefore the excessive memory allocation will be transitory in
1020 nature. This then means that there is only a security impact if:
1022 1) The application does not call SSL_free() in a timely manner in the event
1023 that the connection fails
1025 2) The application is working in a constrained environment where there is
1026 very little free memory
1028 3) The attacker initiates multiple connection attempts such that there are
1029 multiple connections in a state where memory has been allocated for the
1030 connection; SSL_free() has not yet been called; and there is insufficient
1031 memory to service the multiple requests.
1033 Except in the instance of (1) above any Denial Of Service is likely to be
1034 transitory because as soon as the connection fails the memory is
1035 subsequently freed again in the SSL_free() call. However there is an
1036 increased risk during this period of application crashes due to the lack of
1037 memory - which would then mean a more serious Denial of Service.
1039 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1040 (CVE-2016-6307 and CVE-2016-6308)
1043 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
1044 had to be removed. Primary reason is that vendor assembler can't
1045 assemble our modules with -KPIC flag. As result it, assembly
1046 support, was not even available as option. But its lack means
1047 lack of side-channel resistant code, which is incompatible with
1048 security by todays standards. Fortunately gcc is readily available
1049 prepackaged option, which we firmly point at...
1052 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
1054 *) Windows command-line tool supports UTF-8 opt-in option for arguments
1055 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
1056 (to any value) allows Windows user to access PKCS#12 file generated
1057 with Windows CryptoAPI and protected with non-ASCII password, as well
1058 as files generated under UTF-8 locale on Linux also protected with
1062 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
1063 have been disabled by default and removed from DEFAULT, just like RC4.
1064 See the RC4 item below to re-enable both.
1067 *) The method for finding the storage location for the Windows RAND seed file
1068 has changed. First we check %RANDFILE%. If that is not set then we check
1069 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
1070 all else fails we fall back to C:\.
1073 *) The EVP_EncryptUpdate() function has had its return type changed from void
1074 to int. A return of 0 indicates and error while a return of 1 indicates
1078 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
1079 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
1080 off the constant time implementation for RSA, DSA and DH have been made
1081 no-ops and deprecated.
1084 *) Windows RAND implementation was simplified to only get entropy by
1085 calling CryptGenRandom(). Various other RAND-related tickets
1087 [Joseph Wylie Yandle, Rich Salz]
1089 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
1090 and OPENSSL_LH_, respectively. The old names are available
1091 with API compatibility. They new names are now completely documented.
1094 *) Unify TYPE_up_ref(obj) methods signature.
1095 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
1096 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
1097 int (instead of void) like all others TYPE_up_ref() methods.
1098 So now these methods also check the return value of CRYPTO_atomic_add(),
1099 and the validity of object reference counter.
1100 [fdasilvayy@gmail.com]
1102 *) With Windows Visual Studio builds, the .pdb files are installed
1103 alongside the installed libraries and executables. For a static
1104 library installation, ossl_static.pdb is the associate compiler
1105 generated .pdb file to be used when linking programs.
1108 *) Remove openssl.spec. Packaging files belong with the packagers.
1111 *) Automatic Darwin/OSX configuration has had a refresh, it will now
1112 recognise x86_64 architectures automatically. You can still decide
1113 to build for a different bitness with the environment variable
1114 KERNEL_BITS (can be 32 or 64), for example:
1116 KERNEL_BITS=32 ./config
1120 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
1121 256 bit AES and HMAC with SHA256.
1124 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
1127 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
1130 *) To enable users to have their own config files and build file templates,
1131 Configure looks in the directory indicated by the environment variable
1132 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
1133 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
1134 name and is used as is.
1137 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
1138 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
1139 X509_CERT_FILE_CTX was removed.
1142 *) "shared" builds are now the default. To create only static libraries use
1143 the "no-shared" Configure option.
1146 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
1147 All of these option have not worked for some while and are fundamental
1151 *) Make various cleanup routines no-ops and mark them as deprecated. Most
1152 global cleanup functions are no longer required because they are handled
1153 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
1154 Explicitly de-initing can cause problems (e.g. where a library that uses
1155 OpenSSL de-inits, but an application is still using it). The affected
1156 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
1157 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
1158 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
1159 COMP_zlib_cleanup().
1162 *) --strict-warnings no longer enables runtime debugging options
1163 such as REF_DEBUG. Instead, debug options are automatically
1164 enabled with '--debug' builds.
1165 [Andy Polyakov, Emilia Käsper]
1167 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
1168 have been moved out of the public header files. New functions for managing
1169 these have been added.
1172 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
1173 objects have been moved out of the public header files. New
1174 functions for managing these have been added.
1177 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1178 have been moved out of the public header files. New functions for managing
1179 these have been added.
1182 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1183 moved out of the public header files. New functions for managing these
1187 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
1190 *) Removed the mk1mf build scripts.
1193 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1194 it is always safe to #include a header now.
1197 *) Removed the aged BC-32 config and all its supporting scripts
1200 *) Removed support for Ultrix, Netware, and OS/2.
1203 *) Add support for HKDF.
1204 [Alessandro Ghedini]
1206 *) Add support for blake2b and blake2s
1209 *) Added support for "pipelining". Ciphers that have the
1210 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1211 encryptions/decryptions simultaneously. There are currently no built-in
1212 ciphers with this property but the expectation is that engines will be able
1213 to offer it to significantly improve throughput. Support has been extended
1214 into libssl so that multiple records for a single connection can be
1215 processed in one go (for >=TLS 1.1).
1218 *) Added the AFALG engine. This is an async capable engine which is able to
1219 offload work to the Linux kernel. In this initial version it only supports
1220 AES128-CBC. The kernel must be version 4.1.0 or greater.
1223 *) OpenSSL now uses a new threading API. It is no longer necessary to
1224 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1225 are two supported threading models: pthreads and windows threads. It is
1226 also possible to configure OpenSSL at compile time for "no-threads". The
1227 old threading API should no longer be used. The functions have been
1228 replaced with "no-op" compatibility macros.
1229 [Alessandro Ghedini, Matt Caswell]
1231 *) Modify behavior of ALPN to invoke callback after SNI/servername
1232 callback, such that updates to the SSL_CTX affect ALPN.
1235 *) Add SSL_CIPHER queries for authentication and key-exchange.
1238 *) Changes to the DEFAULT cipherlist:
1239 - Prefer (EC)DHE handshakes over plain RSA.
1240 - Prefer AEAD ciphers over legacy ciphers.
1241 - Prefer ECDSA over RSA when both certificates are available.
1242 - Prefer TLSv1.2 ciphers/PRF.
1243 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1247 *) Change the ECC default curve list to be this, in order: x25519,
1248 secp256r1, secp521r1, secp384r1.
1251 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1252 disabled by default. They can be re-enabled using the
1253 enable-weak-ssl-ciphers option to Configure.
1256 *) If the server has ALPN configured, but supports no protocols that the
1257 client advertises, send a fatal "no_application_protocol" alert.
1258 This behaviour is SHALL in RFC 7301, though it isn't universally
1259 implemented by other servers.
1262 *) Add X25519 support.
1263 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1264 for public and private key encoding using the format documented in
1265 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1266 key generation and key derivation.
1268 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1272 *) Deprecate SRP_VBASE_get_by_user.
1273 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1274 In order to fix an unavoidable memory leak (CVE-2016-0798),
1275 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1276 seed, even if the seed is configured.
1278 Users should use SRP_VBASE_get1_by_user instead. Note that in
1279 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1280 also that even though configuring the SRP seed attempts to hide
1281 invalid usernames by continuing the handshake with fake
1282 credentials, this behaviour is not constant time and no strong
1283 guarantees are made that the handshake is indistinguishable from
1284 that of a valid user.
1287 *) Configuration change; it's now possible to build dynamic engines
1288 without having to build shared libraries and vice versa. This
1289 only applies to the engines in engines/, those in crypto/engine/
1290 will always be built into libcrypto (i.e. "static").
1292 Building dynamic engines is enabled by default; to disable, use
1293 the configuration option "disable-dynamic-engine".
1295 The only requirements for building dynamic engines are the
1296 presence of the DSO module and building with position independent
1297 code, so they will also automatically be disabled if configuring
1298 with "disable-dso" or "disable-pic".
1300 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1301 are also taken away from openssl/opensslconf.h, as they are
1305 *) Configuration change; if there is a known flag to compile
1306 position independent code, it will always be applied on the
1307 libcrypto and libssl object files, and never on the application
1308 object files. This means other libraries that use routines from
1309 libcrypto / libssl can be made into shared libraries regardless
1310 of how OpenSSL was configured.
1312 If this isn't desirable, the configuration options "disable-pic"
1313 or "no-pic" can be used to disable the use of PIC. This will
1314 also disable building shared libraries and dynamic engines.
1317 *) Removed JPAKE code. It was experimental and has no wide use.
1320 *) The INSTALL_PREFIX Makefile variable has been renamed to
1321 DESTDIR. That makes for less confusion on what this variable
1322 is for. Also, the configuration option --install_prefix is
1326 *) Heartbeat for TLS has been removed and is disabled by default
1327 for DTLS; configure with enable-heartbeats. Code that uses the
1328 old #define's might need to be updated.
1329 [Emilia Käsper, Rich Salz]
1331 *) Rename REF_CHECK to REF_DEBUG.
1334 *) New "unified" build system
1336 The "unified" build system is aimed to be a common system for all
1337 platforms we support. With it comes new support for VMS.
1339 This system builds supports building in a different directory tree
1340 than the source tree. It produces one Makefile (for unix family
1341 or lookalikes), or one descrip.mms (for VMS).
1343 The source of information to make the Makefile / descrip.mms is
1344 small files called 'build.info', holding the necessary
1345 information for each directory with source to compile, and a
1346 template in Configurations, like unix-Makefile.tmpl or
1349 With this change, the library names were also renamed on Windows
1350 and on VMS. They now have names that are closer to the standard
1351 on Unix, and include the major version number, and in certain
1352 cases, the architecture they are built for. See "Notes on shared
1353 libraries" in INSTALL.
1355 We rely heavily on the perl module Text::Template.
1358 *) Added support for auto-initialisation and de-initialisation of the library.
1359 OpenSSL no longer requires explicit init or deinit routines to be called,
1360 except in certain circumstances. See the OPENSSL_init_crypto() and
1361 OPENSSL_init_ssl() man pages for further information.
1364 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1365 "peer" argument is now expected to be a BIO_ADDR object.
1367 *) Rewrite of BIO networking library. The BIO library lacked consistent
1368 support of IPv6, and adding it required some more extensive
1369 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1370 which hold all types of addresses and chains of address information.
1371 It also introduces a new API, with functions like BIO_socket,
1372 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1373 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1374 have been adapted accordingly.
1377 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1381 *) CRIME protection: disable compression by default, even if OpenSSL is
1382 compiled with zlib enabled. Applications can still enable compression
1383 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1384 using the SSL_CONF library to configure compression.
1387 *) The signature of the session callback configured with
1388 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1389 was explicitly marked as 'const unsigned char*' instead of
1393 *) Always DPURIFY. Remove the use of uninitialized memory in the
1394 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1397 *) Removed many obsolete configuration items, including
1398 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1399 MD2_CHAR, MD2_INT, MD2_LONG
1401 IDEA_SHORT, IDEA_LONG
1402 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1403 [Rich Salz, with advice from Andy Polyakov]
1405 *) Many BN internals have been moved to an internal header file.
1406 [Rich Salz with help from Andy Polyakov]
1408 *) Configuration and writing out the results from it has changed.
1409 Files such as Makefile include/openssl/opensslconf.h and are now
1410 produced through general templates, such as Makefile.in and
1411 crypto/opensslconf.h.in and some help from the perl module
1414 Also, the center of configuration information is no longer
1415 Makefile. Instead, Configure produces a perl module in
1416 configdata.pm which holds most of the config data (in the hash
1417 table %config), the target data that comes from the target
1418 configuration in one of the Configurations/*.conf files (in
1422 *) To clarify their intended purposes, the Configure options
1423 --prefix and --openssldir change their semantics, and become more
1424 straightforward and less interdependent.
1426 --prefix shall be used exclusively to give the location INSTALLTOP
1427 where programs, scripts, libraries, include files and manuals are
1428 going to be installed. The default is now /usr/local.
1430 --openssldir shall be used exclusively to give the default
1431 location OPENSSLDIR where certificates, private keys, CRLs are
1432 managed. This is also where the default openssl.cnf gets
1434 If the directory given with this option is a relative path, the
1435 values of both the --prefix value and the --openssldir value will
1436 be combined to become OPENSSLDIR.
1437 The default for --openssldir is INSTALLTOP/ssl.
1439 Anyone who uses --openssldir to specify where OpenSSL is to be
1440 installed MUST change to use --prefix instead.
1443 *) The GOST engine was out of date and therefore it has been removed. An up
1444 to date GOST engine is now being maintained in an external repository.
1445 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1446 support for GOST ciphersuites (these are only activated if a GOST engine
1450 *) EGD is no longer supported by default; use enable-egd when
1452 [Ben Kaduk and Rich Salz]
1454 *) The distribution now has Makefile.in files, which are used to
1455 create Makefile's when Configure is run. *Configure must be run
1456 before trying to build now.*
1459 *) The return value for SSL_CIPHER_description() for error conditions
1463 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1465 Obtaining and performing DNSSEC validation of TLSA records is
1466 the application's responsibility. The application provides
1467 the TLSA records of its choice to OpenSSL, and these are then
1468 used to authenticate the peer.
1470 The TLSA records need not even come from DNS. They can, for
1471 example, be used to implement local end-entity certificate or
1472 trust-anchor "pinning", where the "pin" data takes the form
1473 of TLSA records, which can augment or replace verification
1474 based on the usual WebPKI public certification authorities.
1477 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1478 continues to support deprecated interfaces in default builds.
1479 However, applications are strongly advised to compile their
1480 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1481 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1482 or the 1.1.0 releases.
1484 In environments in which all applications have been ported to
1485 not use any deprecated interfaces OpenSSL's Configure script
1486 should be used with the --api=1.1.0 option to entirely remove
1487 support for the deprecated features from the library and
1488 unconditionally disable them in the installed headers.
1489 Essentially the same effect can be achieved with the "no-deprecated"
1490 argument to Configure, except that this will always restrict
1491 the build to just the latest API, rather than a fixed API
1494 As applications are ported to future revisions of the API,
1495 they should update their compile-time OPENSSL_API_COMPAT define
1496 accordingly, but in most cases should be able to continue to
1497 compile with later releases.
1499 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1500 0x10000000L and 0x00908000L, respectively. However those
1501 versions did not support the OPENSSL_API_COMPAT feature, and
1502 so applications are not typically tested for explicit support
1503 of just the undeprecated features of either release.
1506 *) Add support for setting the minimum and maximum supported protocol.
1507 It can bet set via the SSL_set_min_proto_version() and
1508 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1509 MaxProtocol. It's recommended to use the new APIs to disable
1510 protocols instead of disabling individual protocols using
1511 SSL_set_options() or SSL_CONF's Protocol. This change also
1512 removes support for disabling TLS 1.2 in the OpenSSL TLS
1513 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1516 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1519 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1520 and integrates ECDSA and ECDH functionality into EC. Implementations can
1521 now redirect key generation and no longer need to convert to or from
1524 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1525 include the ec.h header file instead.
1528 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1529 ciphers who are no longer supported and drops support the ephemeral RSA key
1530 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1533 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1534 opaque. For HMAC_CTX, the following constructors and destructors
1537 HMAC_CTX *HMAC_CTX_new(void);
1538 void HMAC_CTX_free(HMAC_CTX *ctx);
1540 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1541 destroy such methods has been added. See EVP_MD_meth_new(3) and
1542 EVP_CIPHER_meth_new(3) for documentation.
1545 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1546 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1547 EVP_MD_CTX_reset() should be called instead to reinitialise
1548 an already created structure.
1549 2) For consistency with the majority of our object creators and
1550 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1551 EVP_MD_CTX_(new|free). The old names are retained as macros
1552 for deprecated builds.
1555 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1556 cryptographic operations to be performed asynchronously as long as an
1557 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1558 further details. Libssl has also had this capability integrated with the
1559 introduction of the new mode SSL_MODE_ASYNC and associated error
1560 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1561 pages. This work was developed in partnership with Intel Corp.
1564 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1565 always enabled now. If you want to disable the support you should
1566 exclude it using the list of supported ciphers. This also means that the
1567 "-no_ecdhe" option has been removed from s_server.
1570 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1571 SSL_{CTX_}set1_curves() which can set a list.
1574 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1575 curve you want to support using SSL_{CTX_}set1_curves().
1578 *) State machine rewrite. The state machine code has been significantly
1579 refactored in order to remove much duplication of code and solve issues
1580 with the old code (see ssl/statem/README for further details). This change
1581 does have some associated API changes. Notably the SSL_state() function
1582 has been removed and replaced by SSL_get_state which now returns an
1583 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1584 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1588 *) All instances of the string "ssleay" in the public API were replaced
1589 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1590 Some error codes related to internal RSA_eay API's were renamed.
1593 *) The demo files in crypto/threads were moved to demo/threads.
1596 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1598 [Matt Caswell, Rich Salz]
1600 *) New ASN.1 embed macro.
1602 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1603 structure is not allocated: it is part of the parent. That is instead of
1611 This reduces memory fragmentation and make it impossible to accidentally
1612 set a mandatory field to NULL.
1614 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1615 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1616 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1620 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1623 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1624 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1625 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1626 DES and RC4 ciphersuites.
1629 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1630 This changes the decoding behaviour for some invalid messages,
1631 though the change is mostly in the more lenient direction, and
1632 legacy behaviour is preserved as much as possible.
1635 *) Fix no-stdio build.
1636 [ David Woodhouse <David.Woodhouse@intel.com> and also
1637 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1639 *) New testing framework
1640 The testing framework has been largely rewritten and is now using
1641 perl and the perl modules Test::Harness and an extended variant of
1642 Test::More called OpenSSL::Test to do its work. All test scripts in
1643 test/ have been rewritten into test recipes, and all direct calls to
1644 executables in test/Makefile have become individual recipes using the
1645 simplified testing OpenSSL::Test::Simple.
1647 For documentation on our testing modules, do:
1649 perldoc test/testlib/OpenSSL/Test/Simple.pm
1650 perldoc test/testlib/OpenSSL/Test.pm
1654 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1655 are used; the latter aborts on memory leaks (usually checked on exit).
1656 Some undocumented "set malloc, etc., hooks" functions were removed
1657 and others were changed. All are now documented.
1660 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1662 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1664 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1665 from RFC4279, RFC4785, RFC5487, RFC5489.
1667 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1668 original RSA_PSK patch.
1671 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1672 era flag was never set throughout the codebase (only read). Also removed
1673 SSL3_FLAGS_POP_BUFFER which was only used if
1674 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1677 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1678 to be "oneline" instead of "compat".
1681 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1682 not aware of clients that still exhibit this bug, and the workaround
1683 hasn't been working properly for a while.
1686 *) The return type of BIO_number_read() and BIO_number_written() as well as
1687 the corresponding num_read and num_write members in the BIO structure has
1688 changed from unsigned long to uint64_t. On platforms where an unsigned
1689 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1693 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1694 OpenSSL without support for them. It also means that maintaining
1695 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1696 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1699 *) Removed support for the two export grade static DH ciphersuites
1700 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1701 were newly added (along with a number of other static DH ciphersuites) to
1702 1.0.2. However the two export ones have *never* worked since they were
1703 introduced. It seems strange in any case to be adding new export
1704 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1707 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1708 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1709 and turned into macros which simply call the new preferred function names
1710 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1711 should use the new names instead. Also as part of this change the ssl23.h
1712 header file has been removed.
1715 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1716 code and the associated standard is no longer considered fit-for-purpose.
1719 *) RT2547 was closed. When generating a private key, try to make the
1720 output file readable only by the owner. This behavior change might
1721 be noticeable when interacting with other software.
1723 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1727 *) Added HTTP GET support to the ocsp command.
1730 *) Changed default digest for the dgst and enc commands from MD5 to
1734 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1737 *) Added support for TLS extended master secret from
1738 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1739 initial patch which was a great help during development.
1742 *) All libssl internal structures have been removed from the public header
1743 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1744 now redundant). Users should not attempt to access internal structures
1745 directly. Instead they should use the provided API functions.
1748 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1749 Access to deprecated functions can be re-enabled by running config with
1750 "enable-deprecated". In addition applications wishing to use deprecated
1751 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1752 will, by default, disable some transitive includes that previously existed
1753 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1756 *) Added support for OCB mode. OpenSSL has been granted a patent license
1757 compatible with the OpenSSL license for use of OCB. Details are available
1758 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1759 for OCB can be removed by calling config with no-ocb.
1762 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1763 compatible client hello.
1766 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1767 done while fixing the error code for the key-too-small case.
1768 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1770 *) CA.sh has been removed; use CA.pl instead.
1773 *) Removed old DES API.
1776 *) Remove various unsupported platforms:
1782 Sinix/ReliantUNIX RM400
1787 16-bit platforms such as WIN16
1790 *) Clean up OPENSSL_NO_xxx #define's
1791 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1792 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1793 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1794 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1795 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1796 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1797 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1798 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1799 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1800 Remove MS_STATIC; it's a relic from platforms <32 bits.
1803 *) Cleaned up dead code
1804 Remove all but one '#ifdef undef' which is to be looked at.
1807 *) Clean up calling of xxx_free routines.
1808 Just like free(), fix most of the xxx_free routines to accept
1809 NULL. Remove the non-null checks from callers. Save much code.
1812 *) Add secure heap for storage of private keys (when possible).
1813 Add BIO_s_secmem(), CBIGNUM, etc.
1814 Contributed by Akamai Technologies under our Corporate CLA.
1817 *) Experimental support for a new, fast, unbiased prime candidate generator,
1818 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1819 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1821 *) New output format NSS in the sess_id command line tool. This allows
1822 exporting the session id and the master key in NSS keylog format.
1823 [Martin Kaiser <martin@kaiser.cx>]
1825 *) Harmonize version and its documentation. -f flag is used to display
1827 [mancha <mancha1@zoho.com>]
1829 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1830 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1831 [mancha <mancha1@zoho.com>]
1833 *) Fix some double frees. These are not thought to be exploitable.
1834 [mancha <mancha1@zoho.com>]
1836 *) A missing bounds check in the handling of the TLS heartbeat extension
1837 can be used to reveal up to 64k of memory to a connected client or
1840 Thanks for Neel Mehta of Google Security for discovering this bug and to
1841 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1842 preparing the fix (CVE-2014-0160)
1843 [Adam Langley, Bodo Moeller]
1845 *) Fix for the attack described in the paper "Recovering OpenSSL
1846 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1847 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1848 http://eprint.iacr.org/2014/140
1850 Thanks to Yuval Yarom and Naomi Benger for discovering this
1851 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1852 [Yuval Yarom and Naomi Benger]
1854 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1855 this fixes a limitation in previous versions of OpenSSL.
1858 *) Experimental encrypt-then-mac support.
1860 Experimental support for encrypt then mac from
1861 draft-gutmann-tls-encrypt-then-mac-02.txt
1863 To enable it set the appropriate extension number (0x42 for the test
1864 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1866 For non-compliant peers (i.e. just about everything) this should have no
1869 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1873 *) Add EVP support for key wrapping algorithms, to avoid problems with
1874 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1875 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1876 algorithms and include tests cases.
1879 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1883 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1884 MGF1 digest and OAEP label.
1887 *) Make openssl verify return errors.
1888 [Chris Palmer <palmer@google.com> and Ben Laurie]
1890 *) New function ASN1_TIME_diff to calculate the difference between two
1891 ASN1_TIME structures or one structure and the current time.
1894 *) Update fips_test_suite to support multiple command line options. New
1895 test to induce all self test errors in sequence and check expected
1899 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1900 sign or verify all in one operation.
1903 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1904 test programs and fips_test_suite. Includes functionality to parse
1905 the minimal script output of fipsalgest.pl directly.
1908 *) Add authorisation parameter to FIPS_module_mode_set().
1911 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1914 *) Use separate DRBG fields for internal and external flags. New function
1915 FIPS_drbg_health_check() to perform on demand health checking. Add
1916 generation tests to fips_test_suite with reduced health check interval to
1917 demonstrate periodic health checking. Add "nodh" option to
1918 fips_test_suite to skip very slow DH test.
1921 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1925 *) More extensive health check for DRBG checking many more failure modes.
1926 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1927 combination: call this in fips_test_suite.
1930 *) Add support for canonical generation of DSA parameter 'g'. See
1933 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1934 POST to handle HMAC cases.
1937 *) Add functions FIPS_module_version() and FIPS_module_version_text()
1938 to return numerical and string versions of the FIPS module number.
1941 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
1942 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
1943 outside the validated module in the FIPS capable OpenSSL.
1946 *) Minor change to DRBG entropy callback semantics. In some cases
1947 there is no multiple of the block length between min_len and
1948 max_len. Allow the callback to return more than max_len bytes
1949 of entropy but discard any extra: it is the callback's responsibility
1950 to ensure that the extra data discarded does not impact the
1951 requested amount of entropy.
1954 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
1955 information in FIPS186-3, SP800-57 and SP800-131A.
1958 *) CCM support via EVP. Interface is very similar to GCM case except we
1959 must supply all data in one chunk (i.e. no update, final) and the
1960 message length must be supplied if AAD is used. Add algorithm test
1964 *) Initial version of POST overhaul. Add POST callback to allow the status
1965 of POST to be monitored and/or failures induced. Modify fips_test_suite
1966 to use callback. Always run all selftests even if one fails.
1969 *) XTS support including algorithm test driver in the fips_gcmtest program.
1970 Note: this does increase the maximum key length from 32 to 64 bytes but
1971 there should be no binary compatibility issues as existing applications
1972 will never use XTS mode.
1975 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
1976 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
1977 performs algorithm blocking for unapproved PRNG types. Also do not
1978 set PRNG type in FIPS_mode_set(): leave this to the application.
1979 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
1980 the standard OpenSSL PRNG: set additional data to a date time vector.
1983 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
1984 This shouldn't present any incompatibility problems because applications
1985 shouldn't be using these directly and any that are will need to rethink
1986 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
1989 *) Extensive self tests and health checking required by SP800-90 DRBG.
1990 Remove strength parameter from FIPS_drbg_instantiate and always
1991 instantiate at maximum supported strength.
1994 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
1997 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
2000 *) New function DH_compute_key_padded() to compute a DH key and pad with
2001 leading zeroes if needed: this complies with SP800-56A et al.
2004 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
2005 anything, incomplete, subject to change and largely untested at present.
2008 *) Modify fipscanisteronly build option to only build the necessary object
2009 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
2012 *) Add experimental option FIPSSYMS to give all symbols in
2013 fipscanister.o and FIPS or fips prefix. This will avoid
2014 conflicts with future versions of OpenSSL. Add perl script
2015 util/fipsas.pl to preprocess assembly language source files
2016 and rename any affected symbols.
2019 *) Add selftest checks and algorithm block of non-fips algorithms in
2020 FIPS mode. Remove DES2 from selftests.
2023 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
2024 return internal method without any ENGINE dependencies. Add new
2025 tiny fips sign and verify functions.
2028 *) New build option no-ec2m to disable characteristic 2 code.
2031 *) New build option "fipscanisteronly". This only builds fipscanister.o
2032 and (currently) associated fips utilities. Uses the file Makefile.fips
2033 instead of Makefile.org as the prototype.
2036 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
2037 Update fips_gcmtest to use IV generator.
2040 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
2041 setting output buffer to NULL. The *Final function must be
2042 called although it will not retrieve any additional data. The tag
2043 can be set or retrieved with a ctrl. The IV length is by default 12
2044 bytes (96 bits) but can be set to an alternative value. If the IV
2045 length exceeds the maximum IV length (currently 16 bytes) it cannot be
2049 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
2050 underlying do_cipher function handles all cipher semantics itself
2051 including padding and finalisation. This is useful if (for example)
2052 an ENGINE cipher handles block padding itself. The behaviour of
2053 do_cipher is subtly changed if this flag is set: the return value
2054 is the number of characters written to the output buffer (zero is
2055 no longer an error code) or a negative error code. Also if the
2056 input buffer is NULL and length 0 finalisation should be performed.
2059 *) If a candidate issuer certificate is already part of the constructed
2060 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
2063 *) Improve forward-security support: add functions
2065 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
2066 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
2068 for use by SSL/TLS servers; the callback function will be called whenever a
2069 new session is created, and gets to decide whether the session may be
2070 cached to make it resumable (return 0) or not (return 1). (As by the
2071 SSL/TLS protocol specifications, the session_id sent by the server will be
2072 empty to indicate that the session is not resumable; also, the server will
2073 not generate RFC 4507 (RFC 5077) session tickets.)
2075 A simple reasonable callback implementation is to return is_forward_secure.
2076 This parameter will be set to 1 or 0 depending on the ciphersuite selected
2077 by the SSL/TLS server library, indicating whether it can provide forward
2079 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2081 *) New -verify_name option in command line utilities to set verification
2085 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
2086 Add CMAC pkey methods.
2089 *) Experimental renegotiation in s_server -www mode. If the client
2090 browses /reneg connection is renegotiated. If /renegcert it is
2091 renegotiated requesting a certificate.
2094 *) Add an "external" session cache for debugging purposes to s_server. This
2095 should help trace issues which normally are only apparent in deployed
2096 multi-process servers.
2099 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
2100 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
2101 BIO_set_cipher() and some obscure PEM functions were changed so they
2102 can now return an error. The RAND changes required a change to the
2103 RAND_METHOD structure.
2106 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
2107 a gcc attribute to warn if the result of a function is ignored. This
2108 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
2109 whose return value is often ignored.
2112 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
2113 These allow SCTs (signed certificate timestamps) to be requested and
2114 validated when establishing a connection.
2115 [Rob Percival <robpercival@google.com>]
2117 Changes between 1.0.2g and 1.0.2h [3 May 2016]
2119 *) Prevent padding oracle in AES-NI CBC MAC check
2121 A MITM attacker can use a padding oracle attack to decrypt traffic
2122 when the connection uses an AES CBC cipher and the server support
2125 This issue was introduced as part of the fix for Lucky 13 padding
2126 attack (CVE-2013-0169). The padding check was rewritten to be in
2127 constant time by making sure that always the same bytes are read and
2128 compared against either the MAC or padding bytes. But it no longer
2129 checked that there was enough data to have both the MAC and padding
2132 This issue was reported by Juraj Somorovsky using TLS-Attacker.
2136 *) Fix EVP_EncodeUpdate overflow
2138 An overflow can occur in the EVP_EncodeUpdate() function which is used for
2139 Base64 encoding of binary data. If an attacker is able to supply very large
2140 amounts of input data then a length check can overflow resulting in a heap
2143 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
2144 the PEM_write_bio* family of functions. These are mainly used within the
2145 OpenSSL command line applications, so any application which processes data
2146 from an untrusted source and outputs it as a PEM file should be considered
2147 vulnerable to this issue. User applications that call these APIs directly
2148 with large amounts of untrusted data may also be vulnerable.
2150 This issue was reported by Guido Vranken.
2154 *) Fix EVP_EncryptUpdate overflow
2156 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
2157 is able to supply very large amounts of input data after a previous call to
2158 EVP_EncryptUpdate() with a partial block then a length check can overflow
2159 resulting in a heap corruption. Following an analysis of all OpenSSL
2160 internal usage of the EVP_EncryptUpdate() function all usage is one of two
2161 forms. The first form is where the EVP_EncryptUpdate() call is known to be
2162 the first called function after an EVP_EncryptInit(), and therefore that
2163 specific call must be safe. The second form is where the length passed to
2164 EVP_EncryptUpdate() can be seen from the code to be some small value and
2165 therefore there is no possibility of an overflow. Since all instances are
2166 one of these two forms, it is believed that there can be no overflows in
2167 internal code due to this problem. It should be noted that
2168 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
2169 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
2170 of these calls have also been analysed too and it is believed there are no
2171 instances in internal usage where an overflow could occur.
2173 This issue was reported by Guido Vranken.
2177 *) Prevent ASN.1 BIO excessive memory allocation
2179 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2180 a short invalid encoding can cause allocation of large amounts of memory
2181 potentially consuming excessive resources or exhausting memory.
2183 Any application parsing untrusted data through d2i BIO functions is
2184 affected. The memory based functions such as d2i_X509() are *not* affected.
2185 Since the memory based functions are used by the TLS library, TLS
2186 applications are not affected.
2188 This issue was reported by Brian Carpenter.
2194 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2195 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2196 in arbitrary stack data being returned in the buffer.
2198 This issue was reported by Guido Vranken.
2202 *) Modify behavior of ALPN to invoke callback after SNI/servername
2203 callback, such that updates to the SSL_CTX affect ALPN.
2206 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2210 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2211 methods are enabled and ssl2 is disabled the methods return NULL.
2214 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2216 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2217 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2218 provide any "EXPORT" or "LOW" strength ciphers.
2221 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2222 is by default disabled at build-time. Builds that are not configured with
2223 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2224 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2225 will need to explicitly call either of:
2227 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2229 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2231 as appropriate. Even if either of those is used, or the application
2232 explicitly uses the version-specific SSLv2_method() or its client and
2233 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2234 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2235 ciphers, and SSLv2 56-bit DES are no longer available.
2239 *) Fix a double-free in DSA code
2241 A double free bug was discovered when OpenSSL parses malformed DSA private
2242 keys and could lead to a DoS attack or memory corruption for applications
2243 that receive DSA private keys from untrusted sources. This scenario is
2246 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2251 *) Disable SRP fake user seed to address a server memory leak.
2253 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2255 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2256 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2257 was changed to ignore the "fake user" SRP seed, even if the seed
2260 Users should use SRP_VBASE_get1_by_user instead. Note that in
2261 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2262 also that even though configuring the SRP seed attempts to hide
2263 invalid usernames by continuing the handshake with fake
2264 credentials, this behaviour is not constant time and no strong
2265 guarantees are made that the handshake is indistinguishable from
2266 that of a valid user.
2270 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2272 In the BN_hex2bn function the number of hex digits is calculated using an
2273 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2274 large values of |i| this can result in |bn_expand| not allocating any
2275 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2276 field as NULL leading to a subsequent NULL ptr deref. For very large values
2277 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2278 In this case memory is allocated to the internal BIGNUM data field, but it
2279 is insufficiently sized leading to heap corruption. A similar issue exists
2280 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2281 is ever called by user applications with very large untrusted hex/dec data.
2282 This is anticipated to be a rare occurrence.
2284 All OpenSSL internal usage of these functions use data that is not expected
2285 to be untrusted, e.g. config file data or application command line
2286 arguments. If user developed applications generate config file data based
2287 on untrusted data then it is possible that this could also lead to security
2288 consequences. This is also anticipated to be rare.
2290 This issue was reported to OpenSSL by Guido Vranken.
2294 *) Fix memory issues in BIO_*printf functions
2296 The internal |fmtstr| function used in processing a "%s" format string in
2297 the BIO_*printf functions could overflow while calculating the length of a
2298 string and cause an OOB read when printing very long strings.
2300 Additionally the internal |doapr_outch| function can attempt to write to an
2301 OOB memory location (at an offset from the NULL pointer) in the event of a
2302 memory allocation failure. In 1.0.2 and below this could be caused where
2303 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2304 could be in processing a very long "%s" format string. Memory leaks can
2307 The first issue may mask the second issue dependent on compiler behaviour.
2308 These problems could enable attacks where large amounts of untrusted data
2309 is passed to the BIO_*printf functions. If applications use these functions
2310 in this way then they could be vulnerable. OpenSSL itself uses these
2311 functions when printing out human-readable dumps of ASN.1 data. Therefore
2312 applications that print this data could be vulnerable if the data is from
2313 untrusted sources. OpenSSL command line applications could also be
2314 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2315 as command line arguments.
2317 Libssl is not considered directly vulnerable. Additionally certificates etc
2318 received via remote connections via libssl are also unlikely to be able to
2319 trigger these issues because of message size limits enforced within libssl.
2321 This issue was reported to OpenSSL Guido Vranken.
2325 *) Side channel attack on modular exponentiation
2327 A side-channel attack was found which makes use of cache-bank conflicts on
2328 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2329 of RSA keys. The ability to exploit this issue is limited as it relies on
2330 an attacker who has control of code in a thread running on the same
2331 hyper-threaded core as the victim thread which is performing decryptions.
2333 This issue was reported to OpenSSL by Yuval Yarom, The University of
2334 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2335 Nadia Heninger, University of Pennsylvania with more information at
2336 http://cachebleed.info.
2340 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2341 if no keysize is specified with default_bits. This fixes an
2342 omission in an earlier change that changed all RSA/DSA key generation
2343 apps to use 2048 bits by default.
2346 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2347 *) DH small subgroups
2349 Historically OpenSSL only ever generated DH parameters based on "safe"
2350 primes. More recently (in version 1.0.2) support was provided for
2351 generating X9.42 style parameter files such as those required for RFC 5114
2352 support. The primes used in such files may not be "safe". Where an
2353 application is using DH configured with parameters based on primes that are
2354 not "safe" then an attacker could use this fact to find a peer's private
2355 DH exponent. This attack requires that the attacker complete multiple
2356 handshakes in which the peer uses the same private DH exponent. For example
2357 this could be used to discover a TLS server's private DH exponent if it's
2358 reusing the private DH exponent or it's using a static DH ciphersuite.
2360 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2361 TLS. It is not on by default. If the option is not set then the server
2362 reuses the same private DH exponent for the life of the server process and
2363 would be vulnerable to this attack. It is believed that many popular
2364 applications do set this option and would therefore not be at risk.
2366 The fix for this issue adds an additional check where a "q" parameter is
2367 available (as is the case in X9.42 based parameters). This detects the
2368 only known attack, and is the only possible defense for static DH
2369 ciphersuites. This could have some performance impact.
2371 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2372 default and cannot be disabled. This could have some performance impact.
2374 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2378 *) SSLv2 doesn't block disabled ciphers
2380 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2381 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2382 been disabled, provided that the SSLv2 protocol was not also disabled via
2385 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2386 and Sebastian Schinzel.
2390 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2392 *) BN_mod_exp may produce incorrect results on x86_64
2394 There is a carry propagating bug in the x86_64 Montgomery squaring
2395 procedure. No EC algorithms are affected. Analysis suggests that attacks
2396 against RSA and DSA as a result of this defect would be very difficult to
2397 perform and are not believed likely. Attacks against DH are considered just
2398 feasible (although very difficult) because most of the work necessary to
2399 deduce information about a private key may be performed offline. The amount
2400 of resources required for such an attack would be very significant and
2401 likely only accessible to a limited number of attackers. An attacker would
2402 additionally need online access to an unpatched system using the target
2403 private key in a scenario with persistent DH parameters and a private
2404 key that is shared between multiple clients. For example this can occur by
2405 default in OpenSSL DHE based SSL/TLS ciphersuites.
2407 This issue was reported to OpenSSL by Hanno Böck.
2411 *) Certificate verify crash with missing PSS parameter
2413 The signature verification routines will crash with a NULL pointer
2414 dereference if presented with an ASN.1 signature using the RSA PSS
2415 algorithm and absent mask generation function parameter. Since these
2416 routines are used to verify certificate signature algorithms this can be
2417 used to crash any certificate verification operation and exploited in a
2418 DoS attack. Any application which performs certificate verification is
2419 vulnerable including OpenSSL clients and servers which enable client
2422 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2426 *) X509_ATTRIBUTE memory leak
2428 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2429 memory. This structure is used by the PKCS#7 and CMS routines so any
2430 application which reads PKCS#7 or CMS data from untrusted sources is
2431 affected. SSL/TLS is not affected.
2433 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2438 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2439 This changes the decoding behaviour for some invalid messages,
2440 though the change is mostly in the more lenient direction, and
2441 legacy behaviour is preserved as much as possible.
2444 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2446 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2448 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2450 *) Alternate chains certificate forgery
2452 During certificate verification, OpenSSL will attempt to find an
2453 alternative certificate chain if the first attempt to build such a chain
2454 fails. An error in the implementation of this logic can mean that an
2455 attacker could cause certain checks on untrusted certificates to be
2456 bypassed, such as the CA flag, enabling them to use a valid leaf
2457 certificate to act as a CA and "issue" an invalid certificate.
2459 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2463 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2465 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2466 incompatibility in the handling of HMAC. The previous ABI has now been
2470 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2472 *) Malformed ECParameters causes infinite loop
2474 When processing an ECParameters structure OpenSSL enters an infinite loop
2475 if the curve specified is over a specially malformed binary polynomial
2478 This can be used to perform denial of service against any
2479 system which processes public keys, certificate requests or
2480 certificates. This includes TLS clients and TLS servers with
2481 client authentication enabled.
2483 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2487 *) Exploitable out-of-bounds read in X509_cmp_time
2489 X509_cmp_time does not properly check the length of the ASN1_TIME
2490 string and can read a few bytes out of bounds. In addition,
2491 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2494 An attacker can use this to craft malformed certificates and CRLs of
2495 various sizes and potentially cause a segmentation fault, resulting in
2496 a DoS on applications that verify certificates or CRLs. TLS clients
2497 that verify CRLs are affected. TLS clients and servers with client
2498 authentication enabled may be affected if they use custom verification
2501 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2502 independently by Hanno Böck.
2506 *) PKCS7 crash with missing EnvelopedContent
2508 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2509 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2510 with missing content and trigger a NULL pointer dereference on parsing.
2512 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2513 structures from untrusted sources are affected. OpenSSL clients and
2514 servers are not affected.
2516 This issue was reported to OpenSSL by Michal Zalewski (Google).
2520 *) CMS verify infinite loop with unknown hash function
2522 When verifying a signedData message the CMS code can enter an infinite loop
2523 if presented with an unknown hash function OID. This can be used to perform
2524 denial of service against any system which verifies signedData messages using
2526 This issue was reported to OpenSSL by Johannes Bauer.
2530 *) Race condition handling NewSessionTicket
2532 If a NewSessionTicket is received by a multi-threaded client when attempting to
2533 reuse a previous ticket then a race condition can occur potentially leading to
2534 a double free of the ticket data.
2538 *) Only support 256-bit or stronger elliptic curves with the
2539 'ecdh_auto' setting (server) or by default (client). Of supported
2540 curves, prefer P-256 (both).
2543 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2545 *) ClientHello sigalgs DoS fix
2547 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2548 invalid signature algorithms extension a NULL pointer dereference will
2549 occur. This can be exploited in a DoS attack against the server.
2551 This issue was was reported to OpenSSL by David Ramos of Stanford
2554 [Stephen Henson and Matt Caswell]
2556 *) Multiblock corrupted pointer fix
2558 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2559 feature only applies on 64 bit x86 architecture platforms that support AES
2560 NI instructions. A defect in the implementation of "multiblock" can cause
2561 OpenSSL's internal write buffer to become incorrectly set to NULL when
2562 using non-blocking IO. Typically, when the user application is using a
2563 socket BIO for writing, this will only result in a failed connection.
2564 However if some other BIO is used then it is likely that a segmentation
2565 fault will be triggered, thus enabling a potential DoS attack.
2567 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2571 *) Segmentation fault in DTLSv1_listen fix
2573 The DTLSv1_listen function is intended to be stateless and processes the
2574 initial ClientHello from many peers. It is common for user code to loop
2575 over the call to DTLSv1_listen until a valid ClientHello is received with
2576 an associated cookie. A defect in the implementation of DTLSv1_listen means
2577 that state is preserved in the SSL object from one invocation to the next
2578 that can lead to a segmentation fault. Errors processing the initial
2579 ClientHello can trigger this scenario. An example of such an error could be
2580 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2583 This issue was reported to OpenSSL by Per Allansson.
2587 *) Segmentation fault in ASN1_TYPE_cmp fix
2589 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2590 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2591 certificate signature algorithm consistency this can be used to crash any
2592 certificate verification operation and exploited in a DoS attack. Any
2593 application which performs certificate verification is vulnerable including
2594 OpenSSL clients and servers which enable client authentication.
2598 *) Segmentation fault for invalid PSS parameters fix
2600 The signature verification routines will crash with a NULL pointer
2601 dereference if presented with an ASN.1 signature using the RSA PSS
2602 algorithm and invalid parameters. Since these routines are used to verify
2603 certificate signature algorithms this can be used to crash any
2604 certificate verification operation and exploited in a DoS attack. Any
2605 application which performs certificate verification is vulnerable including
2606 OpenSSL clients and servers which enable client authentication.
2608 This issue was was reported to OpenSSL by Brian Carpenter.
2612 *) ASN.1 structure reuse memory corruption fix
2614 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2615 memory corruption via an invalid write. Such reuse is and has been
2616 strongly discouraged and is believed to be rare.
2618 Applications that parse structures containing CHOICE or ANY DEFINED BY
2619 components may be affected. Certificate parsing (d2i_X509 and related
2620 functions) are however not affected. OpenSSL clients and servers are
2625 *) PKCS7 NULL pointer dereferences fix
2627 The PKCS#7 parsing code does not handle missing outer ContentInfo
2628 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2629 missing content and trigger a NULL pointer dereference on parsing.
2631 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2632 otherwise parse PKCS#7 structures from untrusted sources are
2633 affected. OpenSSL clients and servers are not affected.
2635 This issue was reported to OpenSSL by Michal Zalewski (Google).
2639 *) DoS via reachable assert in SSLv2 servers fix
2641 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2642 servers that both support SSLv2 and enable export cipher suites by sending
2643 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2645 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2646 (OpenSSL development team).
2650 *) Empty CKE with client auth and DHE fix
2652 If client auth is used then a server can seg fault in the event of a DHE
2653 ciphersuite being selected and a zero length ClientKeyExchange message
2654 being sent by the client. This could be exploited in a DoS attack.
2658 *) Handshake with unseeded PRNG fix
2660 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2661 with an unseeded PRNG. The conditions are:
2662 - The client is on a platform where the PRNG has not been seeded
2663 automatically, and the user has not seeded manually
2664 - A protocol specific client method version has been used (i.e. not
2665 SSL_client_methodv23)
2666 - A ciphersuite is used that does not require additional random data from
2667 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2669 If the handshake succeeds then the client random that has been used will
2670 have been generated from a PRNG with insufficient entropy and therefore the
2671 output may be predictable.
2673 For example using the following command with an unseeded openssl will
2674 succeed on an unpatched platform:
2676 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2680 *) Use After Free following d2i_ECPrivatekey error fix
2682 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2683 could cause a use after free condition. This, in turn, could cause a double
2684 free in several private key parsing functions (such as d2i_PrivateKey
2685 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2686 for applications that receive EC private keys from untrusted
2687 sources. This scenario is considered rare.
2689 This issue was discovered by the BoringSSL project and fixed in their
2694 *) X509_to_X509_REQ NULL pointer deref fix
2696 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2697 the certificate key is invalid. This function is rarely used in practice.
2699 This issue was discovered by Brian Carpenter.
2703 *) Removed the export ciphers from the DEFAULT ciphers
2706 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2708 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2709 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2710 So far those who have to target multiple platforms would compromise
2711 and argue that binary targeting say ARMv5 would still execute on
2712 ARMv8. "Universal" build resolves this compromise by providing
2713 near-optimal performance even on newer platforms.
2716 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2717 (other platforms pending).
2718 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2720 *) Add support for the SignedCertificateTimestampList certificate and
2721 OCSP response extensions from RFC6962.
2724 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2725 for corner cases. (Certain input points at infinity could lead to
2726 bogus results, with non-infinity inputs mapped to infinity too.)
2729 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2730 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2731 common cases are optimized and there still is room for further
2732 improvements. Vector Permutation AES for Altivec is also added.
2735 *) Add support for little-endian ppc64 Linux target.
2736 [Marcelo Cerri (IBM)]
2738 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2739 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2740 are optimized and there still is room for further improvements.
2741 Both 32- and 64-bit modes are supported.
2742 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2744 *) Improved ARMv7 NEON support.
2747 *) Support for SPARC Architecture 2011 crypto extensions, first
2748 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2749 SHA256/512, MD5, GHASH and modular exponentiation.
2750 [Andy Polyakov, David Miller]
2752 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2754 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2756 *) Support for new and upcoming Intel processors, including AVX2,
2757 BMI and SHA ISA extensions. This includes additional "stitched"
2758 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2761 This work was sponsored by Intel Corp.
2764 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2765 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2766 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2769 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2770 this fixes a limitation in previous versions of OpenSSL.
2773 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2774 MGF1 digest and OAEP label.
2777 *) Add EVP support for key wrapping algorithms, to avoid problems with
2778 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2779 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2780 algorithms and include tests cases.
2783 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2785 [Douglas E. Engert, Steve Henson]
2787 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2788 difference in days and seconds between two tm or ASN1_TIME structures.
2791 *) Add -rev test option to s_server to just reverse order of characters
2792 received by client and send back to server. Also prints an abbreviated
2793 summary of the connection parameters.
2796 *) New option -brief for s_client and s_server to print out a brief summary
2797 of connection parameters.
2800 *) Add callbacks for arbitrary TLS extensions.
2801 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2803 *) New option -crl_download in several openssl utilities to download CRLs
2804 from CRLDP extension in certificates.
2807 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2810 *) New function X509_CRL_diff to generate a delta CRL from the difference
2811 of two full CRLs. Add support to "crl" utility.
2814 *) New functions to set lookup_crls function and to retrieve
2815 X509_STORE from X509_STORE_CTX.
2818 *) Print out deprecated issuer and subject unique ID fields in
2822 *) Extend OCSP I/O functions so they can be used for simple general purpose
2823 HTTP as well as OCSP. New wrapper function which can be used to download
2824 CRLs using the OCSP API.
2827 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2830 *) SSL_CONF* functions. These provide a common framework for application
2831 configuration using configuration files or command lines.
2834 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2835 message callback and prints the results. Needs compile time option
2836 "enable-ssl-trace". New options to s_client and s_server to enable
2840 *) New ctrl and macro to retrieve supported points extensions.
2841 Print out extension in s_server and s_client.
2844 *) New functions to retrieve certificate signature and signature
2848 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2852 *) New Suite B modes for TLS code. These use and enforce the requirements
2853 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2854 only use Suite B curves. The Suite B modes can be set by using the
2855 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2858 *) New chain verification flags for Suite B levels of security. Check
2859 algorithms are acceptable when flags are set in X509_verify_cert.
2862 *) Make tls1_check_chain return a set of flags indicating checks passed
2863 by a certificate chain. Add additional tests to handle client
2864 certificates: checks for matching certificate type and issuer name
2868 *) If an attempt is made to use a signature algorithm not in the peer
2869 preference list abort the handshake. If client has no suitable
2870 signature algorithms in response to a certificate request do not
2871 use the certificate.
2874 *) If server EC tmp key is not in client preference list abort handshake.
2877 *) Add support for certificate stores in CERT structure. This makes it
2878 possible to have different stores per SSL structure or one store in
2879 the parent SSL_CTX. Include distinct stores for certificate chain
2880 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2881 to build and store a certificate chain in CERT structure: returning
2882 an error if the chain cannot be built: this will allow applications
2883 to test if a chain is correctly configured.
2885 Note: if the CERT based stores are not set then the parent SSL_CTX
2886 store is used to retain compatibility with existing behaviour.
2890 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2891 mask based on the current session, check mask when sending client
2892 hello and checking the requested ciphersuite.
2895 *) New ctrls to retrieve and set certificate types in a certificate
2896 request message. Print out received values in s_client. If certificate
2897 types is not set with custom values set sensible values based on
2898 supported signature algorithms.
2901 *) Support for distinct client and server supported signature algorithms.
2904 *) Add certificate callback. If set this is called whenever a certificate
2905 is required by client or server. An application can decide which
2906 certificate chain to present based on arbitrary criteria: for example
2907 supported signature algorithms. Add very simple example to s_server.
2908 This fixes many of the problems and restrictions of the existing client
2909 certificate callback: for example you can now clear an existing
2910 certificate and specify the whole chain.
2913 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2914 the certificate can be used for (if anything). Set valid_flags field
2915 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2916 to have similar checks in it.
2918 Add new "cert_flags" field to CERT structure and include a "strict mode".
2919 This enforces some TLS certificate requirements (such as only permitting
2920 certificate signature algorithms contained in the supported algorithms
2921 extension) which some implementations ignore: this option should be used
2922 with caution as it could cause interoperability issues.
2925 *) Update and tidy signature algorithm extension processing. Work out
2926 shared signature algorithms based on preferences and peer algorithms
2927 and print them out in s_client and s_server. Abort handshake if no
2928 shared signature algorithms.
2931 *) Add new functions to allow customised supported signature algorithms
2932 for SSL and SSL_CTX structures. Add options to s_client and s_server
2936 *) New function SSL_certs_clear() to delete all references to certificates
2937 from an SSL structure. Before this once a certificate had been added
2938 it couldn't be removed.
2941 *) Integrate hostname, email address and IP address checking with certificate
2942 verification. New verify options supporting checking in openssl utility.
2945 *) Fixes and wildcard matching support to hostname and email checking
2946 functions. Add manual page.
2947 [Florian Weimer (Red Hat Product Security Team)]
2949 *) New functions to check a hostname email or IP address against a
2950 certificate. Add options x509 utility to print results of checks against
2954 *) Fix OCSP checking.
2955 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
2957 *) Initial experimental support for explicitly trusted non-root CAs.
2958 OpenSSL still tries to build a complete chain to a root but if an
2959 intermediate CA has a trust setting included that is used. The first
2960 setting is used: whether to trust (e.g., -addtrust option to the x509
2964 *) Add -trusted_first option which attempts to find certificates in the
2965 trusted store even if an untrusted chain is also supplied.
2968 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
2969 platform support for Linux and Android.
2972 *) Support for linux-x32, ILP32 environment in x86_64 framework.
2975 *) Experimental multi-implementation support for FIPS capable OpenSSL.
2976 When in FIPS mode the approved implementations are used as normal,
2977 when not in FIPS mode the internal unapproved versions are used instead.
2978 This means that the FIPS capable OpenSSL isn't forced to use the
2979 (often lower performance) FIPS implementations outside FIPS mode.
2982 *) Transparently support X9.42 DH parameters when calling
2983 PEM_read_bio_DHparameters. This means existing applications can handle
2984 the new parameter format automatically.
2987 *) Initial experimental support for X9.42 DH parameter format: mainly
2988 to support use of 'q' parameter for RFC5114 parameters.
2991 *) Add DH parameters from RFC5114 including test data to dhtest.
2994 *) Support for automatic EC temporary key parameter selection. If enabled
2995 the most preferred EC parameters are automatically used instead of
2996 hardcoded fixed parameters. Now a server just has to call:
2997 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
2998 support ECDH and use the most appropriate parameters.
3001 *) Enhance and tidy EC curve and point format TLS extension code. Use
3002 static structures instead of allocation if default values are used.
3003 New ctrls to set curves we wish to support and to retrieve shared curves.
3004 Print out shared curves in s_server. New options to s_server and s_client
3005 to set list of supported curves.
3008 *) New ctrls to retrieve supported signature algorithms and
3009 supported curve values as an array of NIDs. Extend openssl utility
3010 to print out received values.
3013 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
3014 between NIDs and the more common NIST names such as "P-256". Enhance
3015 ecparam utility and ECC method to recognise the NIST names for curves.
3018 *) Enhance SSL/TLS certificate chain handling to support different
3019 chains for each certificate instead of one chain in the parent SSL_CTX.
3022 *) Support for fixed DH ciphersuite client authentication: where both
3023 server and client use DH certificates with common parameters.
3026 *) Support for fixed DH ciphersuites: those requiring DH server
3030 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
3032 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
3033 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
3034 X509_CINF_get_signature were reverted post internal team review.
3036 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
3038 *) Build fixes for the Windows and OpenVMS platforms
3039 [Matt Caswell and Richard Levitte]
3041 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
3043 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
3044 message can cause a segmentation fault in OpenSSL due to a NULL pointer
3045 dereference. This could lead to a Denial Of Service attack. Thanks to
3046 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
3050 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
3051 dtls1_buffer_record function under certain conditions. In particular this
3052 could occur if an attacker sent repeated DTLS records with the same
3053 sequence number but for the next epoch. The memory leak could be exploited
3054 by an attacker in a Denial of Service attack through memory exhaustion.
3055 Thanks to Chris Mueller for reporting this issue.
3059 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
3060 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
3061 method would be set to NULL which could later result in a NULL pointer
3062 dereference. Thanks to Frank Schmirler for reporting this issue.
3066 *) Abort handshake if server key exchange message is omitted for ephemeral
3069 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
3070 reporting this issue.
3074 *) Remove non-export ephemeral RSA code on client and server. This code
3075 violated the TLS standard by allowing the use of temporary RSA keys in
3076 non-export ciphersuites and could be used by a server to effectively
3077 downgrade the RSA key length used to a value smaller than the server
3078 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
3079 INRIA or reporting this issue.
3083 *) Fixed issue where DH client certificates are accepted without verification.
3084 An OpenSSL server will accept a DH certificate for client authentication
3085 without the certificate verify message. This effectively allows a client to
3086 authenticate without the use of a private key. This only affects servers
3087 which trust a client certificate authority which issues certificates
3088 containing DH keys: these are extremely rare and hardly ever encountered.
3089 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
3094 *) Ensure that the session ID context of an SSL is updated when its
3095 SSL_CTX is updated via SSL_set_SSL_CTX.
3097 The session ID context is typically set from the parent SSL_CTX,
3098 and can vary with the CTX.
3101 *) Fix various certificate fingerprint issues.
3103 By using non-DER or invalid encodings outside the signed portion of a
3104 certificate the fingerprint can be changed without breaking the signature.
3105 Although no details of the signed portion of the certificate can be changed
3106 this can cause problems with some applications: e.g. those using the
3107 certificate fingerprint for blacklists.
3109 1. Reject signatures with non zero unused bits.
3111 If the BIT STRING containing the signature has non zero unused bits reject
3112 the signature. All current signature algorithms require zero unused bits.
3114 2. Check certificate algorithm consistency.
3116 Check the AlgorithmIdentifier inside TBS matches the one in the
3117 certificate signature. NB: this will result in signature failure
3118 errors for some broken certificates.
3120 Thanks to Konrad Kraszewski from Google for reporting this issue.
3122 3. Check DSA/ECDSA signatures use DER.
3124 Re-encode DSA/ECDSA signatures and compare with the original received
3125 signature. Return an error if there is a mismatch.
3127 This will reject various cases including garbage after signature
3128 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
3129 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
3130 (negative or with leading zeroes).
3132 Further analysis was conducted and fixes were developed by Stephen Henson
3133 of the OpenSSL core team.
3138 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
3139 results on some platforms, including x86_64. This bug occurs at random
3140 with a very low probability, and is not known to be exploitable in any
3141 way, though its exact impact is difficult to determine. Thanks to Pieter
3142 Wuille (Blockstream) who reported this issue and also suggested an initial
3143 fix. Further analysis was conducted by the OpenSSL development team and
3144 Adam Langley of Google. The final fix was developed by Andy Polyakov of
3145 the OpenSSL core team.
3149 *) Do not resume sessions on the server if the negotiated protocol
3150 version does not match the session's version. Resuming with a different
3151 version, while not strictly forbidden by the RFC, is of questionable
3152 sanity and breaks all known clients.
3153 [David Benjamin, Emilia Käsper]
3155 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
3156 early CCS messages during renegotiation. (Note that because
3157 renegotiation is encrypted, this early CCS was not exploitable.)
3160 *) Tighten client-side session ticket handling during renegotiation:
3161 ensure that the client only accepts a session ticket if the server sends
3162 the extension anew in the ServerHello. Previously, a TLS client would
3163 reuse the old extension state and thus accept a session ticket if one was
3164 announced in the initial ServerHello.
3166 Similarly, ensure that the client requires a session ticket if one
3167 was advertised in the ServerHello. Previously, a TLS client would
3168 ignore a missing NewSessionTicket message.
3171 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
3173 *) SRTP Memory Leak.
3175 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
3176 sends a carefully crafted handshake message, to cause OpenSSL to fail
3177 to free up to 64k of memory causing a memory leak. This could be
3178 exploited in a Denial Of Service attack. This issue affects OpenSSL
3179 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
3180 whether SRTP is used or configured. Implementations of OpenSSL that
3181 have been compiled with OPENSSL_NO_SRTP defined are not affected.
3183 The fix was developed by the OpenSSL team.
3187 *) Session Ticket Memory Leak.
3189 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
3190 integrity of that ticket is first verified. In the event of a session
3191 ticket integrity check failing, OpenSSL will fail to free memory
3192 causing a memory leak. By sending a large number of invalid session
3193 tickets an attacker could exploit this issue in a Denial Of Service
3198 *) Build option no-ssl3 is incomplete.
3200 When OpenSSL is configured with "no-ssl3" as a build option, servers
3201 could accept and complete a SSL 3.0 handshake, and clients could be
3202 configured to send them.
3204 [Akamai and the OpenSSL team]
3206 *) Add support for TLS_FALLBACK_SCSV.
3207 Client applications doing fallback retries should call
3208 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3210 [Adam Langley, Bodo Moeller]
3212 *) Add additional DigestInfo checks.
3214 Re-encode DigestInto in DER and check against the original when
3215 verifying RSA signature: this will reject any improperly encoded
3216 DigestInfo structures.
3218 Note: this is a precautionary measure and no attacks are currently known.
3222 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3224 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3225 SRP code can be overrun an internal buffer. Add sanity check that
3226 g, A, B < N to SRP code.
3228 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3229 Group for discovering this issue.
3233 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3234 TLS 1.0 instead of higher protocol versions when the ClientHello message
3235 is badly fragmented. This allows a man-in-the-middle attacker to force a
3236 downgrade to TLS 1.0 even if both the server and the client support a
3237 higher protocol version, by modifying the client's TLS records.
3239 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3240 researching this issue.
3244 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3245 to a denial of service attack. A malicious server can crash the client
3246 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3247 ciphersuite and sending carefully crafted handshake messages.
3249 Thanks to Felix Gröbert (Google) for discovering and researching this
3254 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3255 to leak memory. This can be exploited through a Denial of Service attack.
3256 Thanks to Adam Langley for discovering and researching this issue.
3260 *) An attacker can force openssl to consume large amounts of memory whilst
3261 processing DTLS handshake messages. This can be exploited through a
3262 Denial of Service attack.
3263 Thanks to Adam Langley for discovering and researching this issue.
3267 *) An attacker can force an error condition which causes openssl to crash
3268 whilst processing DTLS packets due to memory being freed twice. This
3269 can be exploited through a Denial of Service attack.
3270 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3275 *) If a multithreaded client connects to a malicious server using a resumed
3276 session and the server sends an ec point format extension it could write
3277 up to 255 bytes to freed memory.
3279 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3284 *) A malicious server can crash an OpenSSL client with a null pointer
3285 dereference (read) by specifying an SRP ciphersuite even though it was not
3286 properly negotiated with the client. This can be exploited through a
3287 Denial of Service attack.
3289 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3290 discovering and researching this issue.
3294 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3295 X509_name_oneline, X509_name_print_ex et al. to leak some information
3296 from the stack. Applications may be affected if they echo pretty printing
3297 output to the attacker.
3299 Thanks to Ivan Fratric (Google) for discovering this issue.
3301 [Emilia Käsper, and Steve Henson]
3303 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3304 for corner cases. (Certain input points at infinity could lead to
3305 bogus results, with non-infinity inputs mapped to infinity too.)
3308 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3310 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3311 handshake can force the use of weak keying material in OpenSSL
3312 SSL/TLS clients and servers.
3314 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3315 researching this issue. (CVE-2014-0224)
3316 [KIKUCHI Masashi, Steve Henson]
3318 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3319 OpenSSL DTLS client the code can be made to recurse eventually crashing
3322 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3324 [Imre Rad, Steve Henson]
3326 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3327 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3328 client or server. This is potentially exploitable to run arbitrary
3329 code on a vulnerable client or server.
3331 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3332 [Jüri Aedla, Steve Henson]
3334 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3335 are subject to a denial of service attack.
3337 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3338 this issue. (CVE-2014-3470)
3339 [Felix Gröbert, Ivan Fratric, Steve Henson]
3341 *) Harmonize version and its documentation. -f flag is used to display
3343 [mancha <mancha1@zoho.com>]
3345 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3346 in i2d_ECPrivateKey.
3347 [mancha <mancha1@zoho.com>]
3349 *) Fix some double frees. These are not thought to be exploitable.
3350 [mancha <mancha1@zoho.com>]
3352 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3354 *) A missing bounds check in the handling of the TLS heartbeat extension
3355 can be used to reveal up to 64k of memory to a connected client or
3358 Thanks for Neel Mehta of Google Security for discovering this bug and to
3359 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3360 preparing the fix (CVE-2014-0160)
3361 [Adam Langley, Bodo Moeller]
3363 *) Fix for the attack described in the paper "Recovering OpenSSL
3364 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3365 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3366 http://eprint.iacr.org/2014/140
3368 Thanks to Yuval Yarom and Naomi Benger for discovering this
3369 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3370 [Yuval Yarom and Naomi Benger]
3372 *) TLS pad extension: draft-agl-tls-padding-03
3374 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3375 TLS client Hello record length value would otherwise be > 255 and
3376 less that 512 pad with a dummy extension containing zeroes so it
3377 is at least 512 bytes long.
3379 [Adam Langley, Steve Henson]
3381 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3383 *) Fix for TLS record tampering bug. A carefully crafted invalid
3384 handshake could crash OpenSSL with a NULL pointer exception.
3385 Thanks to Anton Johansson for reporting this issues.
3388 *) Keep original DTLS digest and encryption contexts in retransmission
3389 structures so we can use the previous session parameters if they need
3390 to be resent. (CVE-2013-6450)
3393 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3394 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3395 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3396 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3397 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3398 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3399 [Rob Stradling, Adam Langley]
3401 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3403 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3404 supporting platforms or when small records were transferred.
3405 [Andy Polyakov, Steve Henson]
3407 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3409 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3411 This addresses the flaw in CBC record processing discovered by
3412 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3413 at: http://www.isg.rhul.ac.uk/tls/
3415 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3416 Security Group at Royal Holloway, University of London
3417 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3418 Emilia Käsper for the initial patch.
3420 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3422 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3423 ciphersuites which can be exploited in a denial of service attack.
3424 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3425 and detecting this bug and to Wolfgang Ettlinger
3426 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3430 *) Return an error when checking OCSP signatures when key is NULL.
3431 This fixes a DoS attack. (CVE-2013-0166)
3434 *) Make openssl verify return errors.
3435 [Chris Palmer <palmer@google.com> and Ben Laurie]
3437 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3438 the right response is stapled. Also change SSL_get_certificate()
3439 so it returns the certificate actually sent.
3440 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3441 [Rob Stradling <rob.stradling@comodo.com>]
3443 *) Fix possible deadlock when decoding public keys.
3446 *) Don't use TLS 1.0 record version number in initial client hello
3450 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3452 *) Sanity check record length before skipping explicit IV in TLS
3453 1.2, 1.1 and DTLS to fix DoS attack.
3455 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3456 fuzzing as a service testing platform.
3460 *) Initialise tkeylen properly when encrypting CMS messages.
3461 Thanks to Solar Designer of Openwall for reporting this issue.
3464 *) In FIPS mode don't try to use composite ciphers as they are not
3468 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3470 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3471 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3472 mean any application compiled against OpenSSL 1.0.0 headers setting
3473 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3474 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3475 0x10000000L Any application which was previously compiled against
3476 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3477 will need to be recompiled as a result. Letting be results in
3478 inability to disable specifically TLS 1.1 and in client context,
3479 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3482 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3483 disable just protocol X, but all protocols above X *if* there are
3484 protocols *below* X still enabled. In more practical terms it means
3485 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3486 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3487 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3491 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3493 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3494 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3495 in CRYPTO_realloc_clean.
3497 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3498 issue and to Adam Langley <agl@chromium.org> for fixing it.
3500 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3502 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3505 *) Workarounds for some broken servers that "hang" if a client hello
3506 record length exceeds 255 bytes.
3508 1. Do not use record version number > TLS 1.0 in initial client
3509 hello: some (but not all) hanging servers will now work.
3510 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3511 the number of ciphers sent in the client hello. This should be
3512 set to an even number, such as 50, for example by passing:
3513 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3514 Most broken servers should now work.
3515 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3516 TLS 1.2 client support entirely.
3519 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3522 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3524 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3525 STRING form instead of a DigestInfo.
3528 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3529 and the RSA_sign/RSA_verify functions. This was made more apparent when
3530 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3531 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3532 the correct format in RSA_verify so both forms transparently work.
3535 *) Some servers which support TLS 1.0 can choke if we initially indicate
3536 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3537 encrypted premaster secret. As a workaround use the maximum permitted
3538 client version in client hello, this should keep such servers happy
3539 and still work with previous versions of OpenSSL.
3542 *) Add support for TLS/DTLS heartbeats.
3543 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3545 *) Add support for SCTP.
3546 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3548 *) Improved PRNG seeding for VOS.
3549 [Paul Green <Paul.Green@stratus.com>]
3551 *) Extensive assembler packs updates, most notably:
3553 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3554 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3555 - x86_64: bit-sliced AES implementation;
3556 - ARM: NEON support, contemporary platforms optimizations;
3557 - s390x: z196 support;
3558 - *: GHASH and GF(2^m) multiplication implementations;
3562 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3563 (removal of unnecessary code)
3564 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3566 *) Add TLS key material exporter from RFC 5705.
3569 *) Add DTLS-SRTP negotiation from RFC 5764.
3572 *) Add Next Protocol Negotiation,
3573 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3574 disabled with a no-npn flag to config or Configure. Code donated
3576 [Adam Langley <agl@google.com> and Ben Laurie]
3578 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3579 NIST-P256, NIST-P521, with constant-time single point multiplication on
3580 typical inputs. Compiler support for the nonstandard type __uint128_t is
3581 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3582 Code made available under Apache License version 2.0.
3584 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3585 line to include this in your build of OpenSSL, and run "make depend" (or
3586 "make update"). This enables the following EC_METHODs:
3588 EC_GFp_nistp224_method()
3589 EC_GFp_nistp256_method()
3590 EC_GFp_nistp521_method()
3592 EC_GROUP_new_by_curve_name() will automatically use these (while
3593 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3595 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3597 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3598 all platforms. Move ssize_t definition from e_os.h to the public
3599 header file e_os2.h as it now appears in public header file cms.h
3602 *) New -sigopt option to the ca, req and x509 utilities. Additional
3603 signature parameters can be passed using this option and in
3607 *) Add RSA PSS signing function. This will generate and set the
3608 appropriate AlgorithmIdentifiers for PSS based on those in the
3609 corresponding EVP_MD_CTX structure. No application support yet.
3612 *) Support for companion algorithm specific ASN1 signing routines.
3613 New function ASN1_item_sign_ctx() signs a pre-initialised
3614 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3615 the appropriate parameters.
3618 *) Add new algorithm specific ASN1 verification initialisation function
3619 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3620 handling will be the same no matter what EVP_PKEY_METHOD is used.
3621 Add a PSS handler to support verification of PSS signatures: checked
3622 against a number of sample certificates.
3625 *) Add signature printing for PSS. Add PSS OIDs.
3626 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3628 *) Add algorithm specific signature printing. An individual ASN1 method
3629 can now print out signatures instead of the standard hex dump.
3631 More complex signatures (e.g. PSS) can print out more meaningful
3632 information. Include DSA version that prints out the signature
3636 *) Password based recipient info support for CMS library: implementing
3640 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3641 neatly separates the code into cipher and PBE sections and is required
3642 for some algorithms that split PBES2 into separate pieces (such as
3643 password based CMS).
3646 *) Session-handling fixes:
3647 - Fix handling of connections that are resuming with a session ID,
3648 but also support Session Tickets.
3649 - Fix a bug that suppressed issuing of a new ticket if the client
3650 presented a ticket with an expired session.
3651 - Try to set the ticket lifetime hint to something reasonable.
3652 - Make tickets shorter by excluding irrelevant information.
3653 - On the client side, don't ignore renewed tickets.
3654 [Adam Langley, Bodo Moeller (Google)]
3656 *) Fix PSK session representation.
3659 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3661 This work was sponsored by Intel.
3664 *) Add GCM support to TLS library. Some custom code is needed to split
3665 the IV between the fixed (from PRF) and explicit (from TLS record)
3666 portions. This adds all GCM ciphersuites supported by RFC5288 and
3667 RFC5289. Generalise some AES* cipherstrings to include GCM and
3668 add a special AESGCM string for GCM only.
3671 *) Expand range of ctrls for AES GCM. Permit setting invocation
3672 field on decrypt and retrieval of invocation field only on encrypt.
3675 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3676 As required by RFC5289 these ciphersuites cannot be used if for
3677 versions of TLS earlier than 1.2.
3680 *) For FIPS capable OpenSSL interpret a NULL default public key method
3681 as unset and return the appropriate default but do *not* set the default.
3682 This means we can return the appropriate method in applications that
3683 switch between FIPS and non-FIPS modes.
3686 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3687 ENGINE is used then we cannot handle that in the FIPS module so we
3688 keep original code iff non-FIPS operations are allowed.
3691 *) Add -attime option to openssl utilities.
3692 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3694 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3697 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3698 FIPS EC methods unconditionally for now.
3701 *) New build option no-ec2m to disable characteristic 2 code.
3704 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3705 all cases can be covered as some introduce binary incompatibilities.
3708 *) Redirect RSA operations to FIPS module including keygen,
3709 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3712 *) Add similar low level API blocking to ciphers.
3715 *) Low level digest APIs are not approved in FIPS mode: any attempt
3716 to use these will cause a fatal error. Applications that *really* want
3717 to use them can use the private_* version instead.
3720 *) Redirect cipher operations to FIPS module for FIPS builds.
3723 *) Redirect digest operations to FIPS module for FIPS builds.
3726 *) Update build system to add "fips" flag which will link in fipscanister.o
3727 for static and shared library builds embedding a signature if needed.
3730 *) Output TLS supported curves in preference order instead of numerical
3731 order. This is currently hardcoded for the highest order curves first.
3732 This should be configurable so applications can judge speed vs strength.
3735 *) Add TLS v1.2 server support for client authentication.
3738 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3742 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3743 FIPS modules versions.
3746 *) Add TLS v1.2 client side support for client authentication. Keep cache
3747 of handshake records longer as we don't know the hash algorithm to use
3748 until after the certificate request message is received.
3751 *) Initial TLS v1.2 client support. Add a default signature algorithms
3752 extension including all the algorithms we support. Parse new signature
3753 format in client key exchange. Relax some ECC signing restrictions for
3754 TLS v1.2 as indicated in RFC5246.
3757 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3758 to new signature format when needed using client digest preference.
3759 All server ciphersuites should now work correctly in TLS v1.2. No client
3760 support yet and no support for client certificates.
3763 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3764 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3765 ciphersuites. At present only RSA key exchange ciphersuites work with
3766 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3767 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3768 and version checking.
3771 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3772 with this defined it will not be affected by any changes to ssl internal
3773 structures. Add several utility functions to allow openssl application
3774 to work with OPENSSL_NO_SSL_INTERN defined.
3777 *) A long standing patch to add support for SRP from EdelWeb (Peter
3778 Sylvester and Christophe Renou) was integrated.
3779 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3780 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3783 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3786 *) Permit abbreviated handshakes when renegotiating using the function
3787 SSL_renegotiate_abbreviated().
3788 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3790 *) Add call to ENGINE_register_all_complete() to
3791 ENGINE_load_builtin_engines(), so some implementations get used
3792 automatically instead of needing explicit application support.
3795 *) Add support for TLS key exporter as described in RFC5705.
3796 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3798 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3799 a few changes are required:
3801 Add SSL_OP_NO_TLSv1_1 flag.
3802 Add TLSv1_1 methods.
3803 Update version checking logic to handle version 1.1.
3804 Add explicit IV handling (ported from DTLS code).
3805 Add command line options to s_client/s_server.
3808 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3810 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3811 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3812 content decryption and always return the same error. Note: this attack
3813 needs on average 2^20 messages so it only affects automated senders. The
3814 old behaviour can be re-enabled in the CMS code by setting the
3815 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3816 an MMA defence is not necessary.
3817 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3818 this issue. (CVE-2012-0884)
3821 *) Fix CVE-2011-4619: make sure we really are receiving a
3822 client hello before rejecting multiple SGC restarts. Thanks to
3823 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3826 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3828 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3829 Thanks to Antonio Martin, Enterprise Secure Access Research and
3830 Development, Cisco Systems, Inc. for discovering this bug and
3831 preparing a fix. (CVE-2012-0050)
3834 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3836 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3837 of the Vaudenay padding oracle attack on CBC mode encryption
3838 which enables an efficient plaintext recovery attack against
3839 the OpenSSL implementation of DTLS. Their attack exploits timing
3840 differences arising during decryption processing. A research
3841 paper describing this attack can be found at:
3842 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3843 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3844 Security Group at Royal Holloway, University of London
3845 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3846 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3847 for preparing the fix. (CVE-2011-4108)
3848 [Robin Seggelmann, Michael Tuexen]
3850 *) Clear bytes used for block padding of SSL 3.0 records.
3852 [Adam Langley (Google)]
3854 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3855 Kadianakis <desnacked@gmail.com> for discovering this issue and
3856 Adam Langley for preparing the fix. (CVE-2011-4619)
3857 [Adam Langley (Google)]
3859 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3860 [Andrey Kulikov <amdeich@gmail.com>]
3862 *) Prevent malformed RFC3779 data triggering an assertion failure.
3863 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3864 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3865 [Rob Austein <sra@hactrn.net>]
3867 *) Improved PRNG seeding for VOS.
3868 [Paul Green <Paul.Green@stratus.com>]
3870 *) Fix ssl_ciph.c set-up race.
3871 [Adam Langley (Google)]
3873 *) Fix spurious failures in ecdsatest.c.
3874 [Emilia Käsper (Google)]
3876 *) Fix the BIO_f_buffer() implementation (which was mixing different
3877 interpretations of the '..._len' fields).
3878 [Adam Langley (Google)]
3880 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3881 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3882 threads won't reuse the same blinding coefficients.
3884 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3885 lock to call BN_BLINDING_invert_ex, and avoids one use of
3886 BN_BLINDING_update for each BN_BLINDING structure (previously,
3887 the last update always remained unused).
3888 [Emilia Käsper (Google)]
3890 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3891 [Bob Buckholz (Google)]
3893 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3895 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3896 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3897 [Kaspar Brand <ossl@velox.ch>]
3899 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3900 for multi-threaded use of ECDH. (CVE-2011-3210)
3901 [Adam Langley (Google)]
3903 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3906 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3907 signature public key algorithm by using OID xref utilities instead.
3908 Before this you could only use some ECC ciphersuites with SHA1 only.
3911 *) Add protection against ECDSA timing attacks as mentioned in the paper
3912 by Billy Bob Brumley and Nicola Tuveri, see:
3914 http://eprint.iacr.org/2011/232.pdf
3916 [Billy Bob Brumley and Nicola Tuveri]
3918 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3920 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3921 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3923 *) Fix bug in string printing code: if *any* escaping is enabled we must
3924 escape the escape character (backslash) or the resulting string is
3928 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3930 *) Disable code workaround for ancient and obsolete Netscape browsers
3931 and servers: an attacker can use it in a ciphersuite downgrade attack.
3932 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3935 *) Fixed J-PAKE implementation error, originally discovered by
3936 Sebastien Martini, further info and confirmation from Stefan
3937 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
3940 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
3942 *) Fix extension code to avoid race conditions which can result in a buffer
3943 overrun vulnerability: resumed sessions must not be modified as they can
3944 be shared by multiple threads. CVE-2010-3864
3947 *) Fix WIN32 build system to correctly link an ENGINE directory into
3951 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
3953 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
3955 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
3957 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
3959 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
3960 context. The operation can be customised via the ctrl mechanism in
3961 case ENGINEs want to include additional functionality.
3964 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
3967 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
3968 output hashes compatible with older versions of OpenSSL.
3969 [Willy Weisz <weisz@vcpc.univie.ac.at>]
3971 *) Fix compression algorithm handling: if resuming a session use the
3972 compression algorithm of the resumed session instead of determining
3973 it from client hello again. Don't allow server to change algorithm.
3976 *) Add load_crls() function to apps tidying load_certs() too. Add option
3977 to verify utility to allow additional CRLs to be included.
3980 *) Update OCSP request code to permit adding custom headers to the request:
3981 some responders need this.
3984 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
3986 [Julia Lawall <julia@diku.dk>]
3988 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
3989 needlessly dereferenced structures, used obsolete functions and
3990 didn't handle all updated verify codes correctly.
3993 *) Disable MD2 in the default configuration.
3996 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
3997 indicate the initial BIO being pushed or popped. This makes it possible
3998 to determine whether the BIO is the one explicitly called or as a result
3999 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
4000 it handles reference counts correctly and doesn't zero out the I/O bio
4001 when it is not being explicitly popped. WARNING: applications which
4002 included workarounds for the old buggy behaviour will need to be modified
4003 or they could free up already freed BIOs.
4006 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
4007 renaming to all platforms (within the 0.9.8 branch, this was
4008 done conditionally on Netware platforms to avoid a name clash).
4009 [Guenter <lists@gknw.net>]
4011 *) Add ECDHE and PSK support to DTLS.
4012 [Michael Tuexen <tuexen@fh-muenster.de>]
4014 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
4018 *) Add "missing" function EVP_MD_flags() (without this the only way to
4019 retrieve a digest flags is by accessing the structure directly. Update
4020 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
4021 or cipher is registered as in the "from" argument. Print out all
4022 registered digests in the dgst usage message instead of manually
4023 attempting to work them out.
4026 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
4027 this allows the use of compression and extensions. Change default cipher
4028 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
4029 by default unless an application cipher string requests it.
4032 *) Alter match criteria in PKCS12_parse(). It used to try to use local
4033 key ids to find matching certificates and keys but some PKCS#12 files
4034 don't follow the (somewhat unwritten) rules and this strategy fails.
4035 Now just gather all certificates together and the first private key
4036 then look for the first certificate that matches the key.
4039 *) Support use of registered digest and cipher names for dgst and cipher
4040 commands instead of having to add each one as a special case. So now
4047 openssl dgst -sha256 foo
4049 and this works for ENGINE based algorithms too.
4053 *) Update Gost ENGINE to support parameter files.
4054 [Victor B. Wagner <vitus@cryptocom.ru>]
4056 *) Support GeneralizedTime in ca utility.
4057 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
4059 *) Enhance the hash format used for certificate directory links. The new
4060 form uses the canonical encoding (meaning equivalent names will work
4061 even if they aren't identical) and uses SHA1 instead of MD5. This form
4062 is incompatible with the older format and as a result c_rehash should
4063 be used to rebuild symbolic links.
4066 *) Make PKCS#8 the default write format for private keys, replacing the
4067 traditional format. This form is standardised, more secure and doesn't
4068 include an implicit MD5 dependency.
4071 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
4072 committed to OpenSSL should pass this lot as a minimum.
4075 *) Add session ticket override functionality for use by EAP-FAST.
4076 [Jouni Malinen <j@w1.fi>]
4078 *) Modify HMAC functions to return a value. Since these can be implemented
4079 in an ENGINE errors can occur.
4082 *) Type-checked OBJ_bsearch_ex.
4085 *) Type-checked OBJ_bsearch. Also some constification necessitated
4086 by type-checking. Still to come: TXT_DB, bsearch(?),
4087 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
4091 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
4092 seconds to a tm structure directly, instead of going through OS
4093 specific date routines. This avoids any issues with OS routines such
4094 as the year 2038 bug. New *_adj() functions for ASN1 time structures
4095 and X509_time_adj_ex() to cover the extended range. The existing
4096 X509_time_adj() is still usable and will no longer have any date issues.
4099 *) Delta CRL support. New use deltas option which will attempt to locate
4100 and search any appropriate delta CRLs available.
4102 This work was sponsored by Google.
4105 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
4106 code and add additional score elements. Validate alternate CRL paths
4107 as part of the CRL checking and indicate a new error "CRL path validation
4108 error" in this case. Applications wanting additional details can use
4109 the verify callback and check the new "parent" field. If this is not
4110 NULL CRL path validation is taking place. Existing applications won't
4111 see this because it requires extended CRL support which is off by
4114 This work was sponsored by Google.
4117 *) Support for freshest CRL extension.
4119 This work was sponsored by Google.
4122 *) Initial indirect CRL support. Currently only supported in the CRLs
4123 passed directly and not via lookup. Process certificate issuer
4124 CRL entry extension and lookup CRL entries by bother issuer name
4125 and serial number. Check and process CRL issuer entry in IDP extension.
4127 This work was sponsored by Google.
4130 *) Add support for distinct certificate and CRL paths. The CRL issuer
4131 certificate is validated separately in this case. Only enabled if
4132 an extended CRL support flag is set: this flag will enable additional
4133 CRL functionality in future.
4135 This work was sponsored by Google.
4138 *) Add support for policy mappings extension.
4140 This work was sponsored by Google.
4143 *) Fixes to pathlength constraint, self issued certificate handling,
4144 policy processing to align with RFC3280 and PKITS tests.
4146 This work was sponsored by Google.
4149 *) Support for name constraints certificate extension. DN, email, DNS
4150 and URI types are currently supported.
4152 This work was sponsored by Google.
4155 *) To cater for systems that provide a pointer-based thread ID rather
4156 than numeric, deprecate the current numeric thread ID mechanism and
4157 replace it with a structure and associated callback type. This
4158 mechanism allows a numeric "hash" to be extracted from a thread ID in
4159 either case, and on platforms where pointers are larger than 'long',
4160 mixing is done to help ensure the numeric 'hash' is usable even if it
4161 can't be guaranteed unique. The default mechanism is to use "&errno"
4162 as a pointer-based thread ID to distinguish between threads.
4164 Applications that want to provide their own thread IDs should now use
4165 CRYPTO_THREADID_set_callback() to register a callback that will call
4166 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
4168 Note that ERR_remove_state() is now deprecated, because it is tied
4169 to the assumption that thread IDs are numeric. ERR_remove_state(0)
4170 to free the current thread's error state should be replaced by
4171 ERR_remove_thread_state(NULL).
4173 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
4174 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
4175 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
4176 application was previously providing a numeric thread callback that
4177 was inappropriate for distinguishing threads, then uniqueness might
4178 have been obtained with &errno that happened immediately in the
4179 intermediate development versions of OpenSSL; this is no longer the
4180 case, the numeric thread callback will now override the automatic use
4182 [Geoff Thorpe, with help from Bodo Moeller]
4184 *) Initial support for different CRL issuing certificates. This covers a
4185 simple case where the self issued certificates in the chain exist and
4186 the real CRL issuer is higher in the existing chain.
4188 This work was sponsored by Google.
4191 *) Removed effectively defunct crypto/store from the build.
4194 *) Revamp of STACK to provide stronger type-checking. Still to come:
4195 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
4196 ASN1_STRING, CONF_VALUE.
4199 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4200 RAM on SSL connections. This option can save about 34k per idle SSL.
4203 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4204 STACK, TXT_DB, bsearch, qsort.
4207 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4208 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4209 support for data, signedData, compressedData, digestedData and
4210 encryptedData, envelopedData types included. Scripts to check against
4211 RFC4134 examples draft and interop and consistency checks of many
4212 content types and variants.
4215 *) Add options to enc utility to support use of zlib compression BIO.
4218 *) Extend mk1mf to support importing of options and assembly language
4219 files from Configure script, currently only included in VC-WIN32.
4220 The assembly language rules can now optionally generate the source
4221 files from the associated perl scripts.
4224 *) Implement remaining functionality needed to support GOST ciphersuites.
4225 Interop testing has been performed using CryptoPro implementations.
4226 [Victor B. Wagner <vitus@cryptocom.ru>]
4228 *) s390x assembler pack.
4231 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4235 *) Implement Opaque PRF Input TLS extension as specified in
4236 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4237 official specification yet and no extension type assignment by
4238 IANA exists, this extension (for now) will have to be explicitly
4239 enabled when building OpenSSL by providing the extension number
4240 to use. For example, specify an option
4242 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4244 to the "config" or "Configure" script to enable the extension,
4245 assuming extension number 0x9527 (which is a completely arbitrary
4246 and unofficial assignment based on the MD5 hash of the Internet
4247 Draft). Note that by doing so, you potentially lose
4248 interoperability with other TLS implementations since these might
4249 be using the same extension number for other purposes.
4251 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4252 opaque PRF input value to use in the handshake. This will create
4253 an internal copy of the length-'len' string at 'src', and will
4254 return non-zero for success.
4256 To get more control and flexibility, provide a callback function
4259 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4260 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4264 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4267 Callback function 'cb' will be called in handshakes, and is
4268 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4269 Argument 'arg' is for application purposes (the value as given to
4270 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4271 be provided to the callback function). The callback function
4272 has to return non-zero to report success: usually 1 to use opaque
4273 PRF input just if possible, or 2 to enforce use of the opaque PRF
4274 input. In the latter case, the library will abort the handshake
4275 if opaque PRF input is not successfully negotiated.
4277 Arguments 'peerinput' and 'len' given to the callback function
4278 will always be NULL and 0 in the case of a client. A server will
4279 see the client's opaque PRF input through these variables if
4280 available (NULL and 0 otherwise). Note that if the server
4281 provides an opaque PRF input, the length must be the same as the
4282 length of the client's opaque PRF input.
4284 Note that the callback function will only be called when creating
4285 a new session (session resumption can resume whatever was
4286 previously negotiated), and will not be called in SSL 2.0
4287 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4288 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4289 for applications that need to enforce opaque PRF input.
4293 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4296 [Victor B. Wagner <vitus@cryptocom.ru>]
4298 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4299 RFC4507bis. The encrypted ticket format is an encrypted encoded
4300 SSL_SESSION structure, that way new session features are automatically
4303 If a client application caches session in an SSL_SESSION structure
4304 support is transparent because tickets are now stored in the encoded
4307 The SSL_CTX structure automatically generates keys for ticket
4308 protection in servers so again support should be possible
4309 with no application modification.
4311 If a client or server wishes to disable RFC4507 support then the option
4312 SSL_OP_NO_TICKET can be set.
4314 Add a TLS extension debugging callback to allow the contents of any client
4315 or server extensions to be examined.
4317 This work was sponsored by Google.
4320 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4321 OpenSSL should now compile cleanly on gcc 4.2
4322 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4324 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4325 support including streaming MAC support: this is required for GOST
4326 ciphersuite support.
4327 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4329 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4330 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4331 to output in BER and PEM format.
4334 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4335 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4336 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4337 ENGINE support for HMAC keys which are unextractable. New -mac and
4338 -macopt options to dgst utility.
4341 *) New option -sigopt to dgst utility. Update dgst to use
4342 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4343 alternative signing parameters such as X9.31 or PSS in the dgst
4347 *) Change ssl_cipher_apply_rule(), the internal function that does
4348 the work each time a ciphersuite string requests enabling
4349 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4350 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4351 the order of disabled ciphersuites such that those ciphersuites
4352 that most recently went from enabled to disabled not only stay
4353 in order with respect to each other, but also have higher priority
4354 than other disabled ciphersuites the next time ciphersuites are
4357 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4358 the same ciphersuites as with "HIGH" alone, but in a specific
4359 order where the PSK ciphersuites come first (since they are the
4360 most recently disabled ciphersuites when "HIGH" is parsed).
4362 Also, change ssl_create_cipher_list() (using this new
4363 functionality) such that between otherwise identical
4364 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4368 *) Change ssl_create_cipher_list() so that it automatically
4369 arranges the ciphersuites in reasonable order before starting
4370 to process the rule string. Thus, the definition for "DEFAULT"
4371 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4372 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4373 This makes it much easier to arrive at a reasonable default order
4374 in applications for which anonymous ciphers are OK (meaning
4375 that you can't actually use DEFAULT).
4376 [Bodo Moeller; suggested by Victor Duchovni]
4378 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4379 processing) into multiple integers instead of setting
4380 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
4381 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
4382 (These masks as well as the individual bit definitions are hidden
4383 away into the non-exported interface ssl/ssl_locl.h, so this
4384 change to the definition of the SSL_CIPHER structure shouldn't
4385 affect applications.) This give us more bits for each of these
4386 categories, so there is no longer a need to coagulate AES128 and
4387 AES256 into a single algorithm bit, and to coagulate Camellia128
4388 and Camellia256 into a single algorithm bit, which has led to all
4391 Thus, among other things, the kludge introduced in 0.9.7m and
4392 0.9.8e for masking out AES256 independently of AES128 or masking
4393 out Camellia256 independently of AES256 is not needed here in 0.9.9.
4395 With the change, we also introduce new ciphersuite aliases that
4396 so far were missing: "AES128", "AES256", "CAMELLIA128", and
4400 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
4401 Use the leftmost N bytes of the signature input if the input is
4402 larger than the prime q (with N being the size in bytes of q).
4405 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
4406 it yet and it is largely untested.
4409 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
4412 *) Initial incomplete changes to avoid need for function casts in OpenSSL
4413 some compilers (gcc 4.2 and later) reject their use. Safestack is
4414 reimplemented. Update ASN1 to avoid use of legacy functions.
4417 *) Win32/64 targets are linked with Winsock2.
4420 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
4421 to external functions. This can be used to increase CRL handling
4422 efficiency especially when CRLs are very large by (for example) storing
4423 the CRL revoked certificates in a database.
4426 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
4427 new CRLs added to a directory can be used. New command line option
4428 -verify_return_error to s_client and s_server. This causes real errors
4429 to be returned by the verify callback instead of carrying on no matter
4430 what. This reflects the way a "real world" verify callback would behave.
4433 *) GOST engine, supporting several GOST algorithms and public key formats.
4434 Kindly donated by Cryptocom.
4437 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4438 partitioned by DP are handled but no indirect CRL or reason partitioning
4439 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4440 selected via a scoring technique which handles IDP and AKID in CRLs.
4443 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4444 will ultimately be used for all verify operations: this will remove the
4445 X509_STORE dependency on certificate verification and allow alternative
4446 lookup methods. X509_STORE based implementations of these two callbacks.
4449 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4450 Modify get_crl() to find a valid (unexpired) CRL if possible.
4453 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4454 this would be called X509_CRL_cmp() but that name is already used by
4455 a function that just compares CRL issuer names. Cache several CRL
4456 extensions in X509_CRL structure and cache CRLDP in X509.
4459 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4460 this maps equivalent X509_NAME structures into a consistent structure.
4461 Name comparison can then be performed rapidly using memcmp().
4464 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4468 *) Allow digests to supply their own micalg string for S/MIME type using
4469 the ctrl EVP_MD_CTRL_MICALG.
4472 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4473 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4474 ctrl. It can then customise the structure before and/or after signing
4478 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4479 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4480 to free up any added signature OIDs.
4483 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4484 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4485 digest and cipher tables. New options added to openssl utility:
4486 list-message-digest-algorithms and list-cipher-algorithms.
4489 *) Change the array representation of binary polynomials: the list
4490 of degrees of non-zero coefficients is now terminated with -1.
4491 Previously it was terminated with 0, which was also part of the
4492 value; thus, the array representation was not applicable to
4493 polynomials where t^0 has coefficient zero. This change makes
4494 the array representation useful in a more general context.
4497 *) Various modifications and fixes to SSL/TLS cipher string
4498 handling. For ECC, the code now distinguishes between fixed ECDH
4499 with RSA certificates on the one hand and with ECDSA certificates
4500 on the other hand, since these are separate ciphersuites. The
4501 unused code for Fortezza ciphersuites has been removed.
4503 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4504 (not "ECDHE"). For consistency with the code for DH
4505 certificates, use of ECDH certificates is now considered ECDH
4506 authentication, not RSA or ECDSA authentication (the latter is
4507 merely the CA's signing algorithm and not actively used in the
4510 The temporary ciphersuite alias "ECCdraft" is no longer
4511 available, and ECC ciphersuites are no longer excluded from "ALL"
4512 and "DEFAULT". The following aliases now exist for RFC 4492
4513 ciphersuites, most of these by analogy with the DH case:
4515 kECDHr - ECDH cert, signed with RSA
4516 kECDHe - ECDH cert, signed with ECDSA
4517 kECDH - ECDH cert (signed with either RSA or ECDSA)
4518 kEECDH - ephemeral ECDH
4519 ECDH - ECDH cert or ephemeral ECDH
4525 AECDH - anonymous ECDH
4526 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4530 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
4531 Use correct micalg parameters depending on digest(s) in signed message.
4534 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
4535 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
4538 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
4539 an engine to register a method. Add ENGINE lookups for methods and
4540 functional reference processing.
4543 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
4544 EVP_{Sign,Verify}* which allow an application to customise the signature
4548 *) New -resign option to smime utility. This adds one or more signers
4549 to an existing PKCS#7 signedData structure. Also -md option to use an
4550 alternative message digest algorithm for signing.
4553 *) Tidy up PKCS#7 routines and add new functions to make it easier to
4554 create PKCS7 structures containing multiple signers. Update smime
4555 application to support multiple signers.
4558 *) New -macalg option to pkcs12 utility to allow setting of an alternative
4562 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
4563 Reorganize PBE internals to lookup from a static table using NIDs,
4564 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
4565 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
4566 PRF which will be automatically used with PBES2.
4569 *) Replace the algorithm specific calls to generate keys in "req" with the
4573 *) Update PKCS#7 enveloped data routines to use new API. This is now
4574 supported by any public key method supporting the encrypt operation. A
4575 ctrl is added to allow the public key algorithm to examine or modify
4576 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
4580 *) Add a ctrl to asn1 method to allow a public key algorithm to express
4581 a default digest type to use. In most cases this will be SHA1 but some
4582 algorithms (such as GOST) need to specify an alternative digest. The
4583 return value indicates how strong the preference is 1 means optional and
4584 2 is mandatory (that is it is the only supported type). Modify
4585 ASN1_item_sign() to accept a NULL digest argument to indicate it should
4586 use the default md. Update openssl utilities to use the default digest
4587 type for signing if it is not explicitly indicated.
4590 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
4591 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
4592 signing method from the key type. This effectively removes the link
4593 between digests and public key types.
4596 *) Add an OID cross reference table and utility functions. Its purpose is to
4597 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
4598 rsaEncryption. This will allow some of the algorithm specific hackery
4599 needed to use the correct OID to be removed.
4602 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
4603 structures for PKCS7_sign(). They are now set up by the relevant public
4607 *) Add provisional EC pkey method with support for ECDSA and ECDH.
4610 *) Add support for key derivation (agreement) in the API, DH method and
4614 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
4615 public and private key formats. As a side effect these add additional
4616 command line functionality not previously available: DSA signatures can be
4617 generated and verified using pkeyutl and DH key support and generation in
4622 [Oliver Tappe <zooey@hirschkaefer.de>]
4624 *) New make target "install_html_docs" installs HTML renditions of the
4626 [Oliver Tappe <zooey@hirschkaefer.de>]
4628 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4629 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4630 support key and parameter generation and add initial key generation
4631 functionality for RSA.
4634 *) Add functions for main EVP_PKEY_method operations. The undocumented
4635 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4636 EVP_PKEY_{encrypt,decrypt}_old.
4639 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4640 key API, doesn't do much yet.
4643 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4644 public key algorithms. New option to openssl utility:
4645 "list-public-key-algorithms" to print out info.
4648 *) Implement the Supported Elliptic Curves Extension for
4649 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4652 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4653 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4656 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4657 utilities such as rsa, dsa, dsaparam etc except they process any key
4661 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4662 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4663 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4667 *) Initial support for pluggable public key ASN1.
4668 De-spaghettify the public key ASN1 handling. Move public and private
4669 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4670 algorithm specific handling to a single module within the relevant
4671 algorithm directory. Add functions to allow (near) opaque processing
4672 of public and private key structures.
4675 *) Implement the Supported Point Formats Extension for
4676 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4679 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4680 for the psk identity [hint] and the psk callback functions to the
4681 SSL_SESSION, SSL and SSL_CTX structure.
4684 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4688 SSL_CTX_use_psk_identity_hint
4689 SSL_get_psk_identity_hint
4690 SSL_get_psk_identity
4691 SSL_use_psk_identity_hint
4693 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4695 *) Add RFC 3161 compliant time stamp request creation, response generation
4696 and response verification functionality.
4697 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4699 *) Add initial support for TLS extensions, specifically for the server_name
4700 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4701 have new members for a host name. The SSL data structure has an
4702 additional member SSL_CTX *initial_ctx so that new sessions can be
4703 stored in that context to allow for session resumption, even after the
4704 SSL has been switched to a new SSL_CTX in reaction to a client's
4705 server_name extension.
4707 New functions (subject to change):
4709 SSL_get_servername()
4710 SSL_get_servername_type()
4713 New CTRL codes and macros (subject to change):
4715 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4716 - SSL_CTX_set_tlsext_servername_callback()
4717 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4718 - SSL_CTX_set_tlsext_servername_arg()
4719 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4721 openssl s_client has a new '-servername ...' option.
4723 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4724 '-key2 ...', '-servername_fatal' (subject to change). This allows
4725 testing the HostName extension for a specific single host name ('-cert'
4726 and '-key' remain fallbacks for handshakes without HostName
4727 negotiation). If the unrecognized_name alert has to be sent, this by
4728 default is a warning; it becomes fatal with the '-servername_fatal'
4731 [Peter Sylvester, Remy Allais, Christophe Renou]
4733 *) Whirlpool hash implementation is added.
4736 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4737 bn(64,32). Because of instruction set limitations it doesn't have
4738 any negative impact on performance. This was done mostly in order
4739 to make it possible to share assembler modules, such as bn_mul_mont
4740 implementations, between 32- and 64-bit builds without hassle.
4743 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4744 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4748 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4749 dedicated Montgomery multiplication procedure, is introduced.
4750 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4751 "64-bit" performance on certain 32-bit targets.
4754 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4755 in SSL structures. New SSL ctrl to set maximum send fragment size.
4756 Save memory by setting the I/O buffer sizes dynamically instead of
4757 using the maximum available value.
4760 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
4761 in addition to the text details.
4764 *) Very, very preliminary EXPERIMENTAL support for printing of general
4765 ASN1 structures. This currently produces rather ugly output and doesn't
4766 handle several customised structures at all.
4769 *) Integrated support for PVK file format and some related formats such
4770 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4771 these in the 'rsa' and 'dsa' utilities.
4774 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4777 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4778 place for the (very old) "NETSCAPE" format certificates which are now
4779 handled using new ASN1 code equivalents.
4782 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4783 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4784 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4787 *) Modify CRL distribution points extension code to print out previously
4788 unsupported fields. Enhance extension setting code to allow setting of
4792 *) Add print and set support for Issuing Distribution Point CRL extension.
4795 *) Change 'Configure' script to enable Camellia by default.
4798 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4800 *) When rejecting SSL/TLS records due to an incorrect version number, never
4801 update s->server with a new major version number. As of
4802 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
4803 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
4804 the previous behavior could result in a read attempt at NULL when
4805 receiving specific incorrect SSL/TLS records once record payload
4806 protection is active. (CVE-2010-0740)
4807 [Bodo Moeller, Adam Langley <agl@chromium.org>]
4809 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
4810 could be crashed if the relevant tables were not present (e.g. chrooted).
4811 [Tomas Hoger <thoger@redhat.com>]
4813 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
4815 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
4816 [Martin Olsson, Neel Mehta]
4818 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
4819 accommodate for stack sorting, always a write lock!).
4822 *) On some versions of WIN32 Heap32Next is very slow. This can cause
4823 excessive delays in the RAND_poll(): over a minute. As a workaround
4824 include a time check in the inner Heap32Next loop too.
4827 *) The code that handled flushing of data in SSL/TLS originally used the
4828 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
4829 the problem outlined in PR#1949. The fix suggested there however can
4830 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4831 of Apache). So instead simplify the code to flush unconditionally.
4832 This should be fine since flushing with no data to flush is a no op.
4835 *) Handle TLS versions 2.0 and later properly and correctly use the
4836 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4837 off ancient servers have a habit of sticking around for a while...
4840 *) Modify compression code so it frees up structures without using the
4841 ex_data callbacks. This works around a problem where some applications
4842 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4843 restarting) then use compression (e.g. SSL with compression) later.
4844 This results in significant per-connection memory leaks and
4845 has caused some security issues including CVE-2008-1678 and
4849 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4850 change when encrypting or decrypting.
4853 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4854 connect and renegotiate with servers which do not support RI.
4855 Until RI is more widely deployed this option is enabled by default.
4858 *) Add "missing" ssl ctrls to clear options and mode.
4861 *) If client attempts to renegotiate and doesn't support RI respond with
4862 a no_renegotiation alert as required by RFC5746. Some renegotiating
4863 TLS clients will continue a connection gracefully when they receive
4864 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4865 waiting for a server hello which it will never receive. Now we treat a
4866 received no_renegotiation alert as a fatal error. This is because
4867 applications requesting a renegotiation might well expect it to succeed
4868 and would have no code in place to handle the server denying it so the
4869 only safe thing to do is to terminate the connection.
4872 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4873 peer supports secure renegotiation and 0 otherwise. Print out peer
4874 renegotiation support in s_client/s_server.
4877 *) Replace the highly broken and deprecated SPKAC certification method with
4878 the updated NID creation version. This should correctly handle UTF8.
4881 *) Implement RFC5746. Re-enable renegotiation but require the extension
4882 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
4883 turns out to be a bad idea. It has been replaced by
4884 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
4885 SSL_CTX_set_options(). This is really not recommended unless you
4886 know what you are doing.
4887 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
4889 *) Fixes to stateless session resumption handling. Use initial_ctx when
4890 issuing and attempting to decrypt tickets in case it has changed during
4891 servername handling. Use a non-zero length session ID when attempting
4892 stateless session resumption: this makes it possible to determine if
4893 a resumption has occurred immediately after receiving server hello
4894 (several places in OpenSSL subtly assume this) instead of later in
4898 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
4899 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
4900 fixes for a few places where the return code is not checked
4902 [Julia Lawall <julia@diku.dk>]
4904 *) Add --strict-warnings option to Configure script to include devteam
4905 warnings in other configurations.
4908 *) Add support for --libdir option and LIBDIR variable in makefiles. This
4909 makes it possible to install openssl libraries in locations which
4910 have names other than "lib", for example "/usr/lib64" which some
4912 [Steve Henson, based on patch from Jeremy Utley]
4914 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
4915 X690 8.9.12 and can produce some misleading textual output of OIDs.
4916 [Steve Henson, reported by Dan Kaminsky]
4918 *) Delete MD2 from algorithm tables. This follows the recommendation in
4919 several standards that it is not used in new applications due to
4920 several cryptographic weaknesses. For binary compatibility reasons
4921 the MD2 API is still compiled in by default.
4924 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
4928 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
4929 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
4931 [Guenter <lists@gknw.net>]
4933 *) Fix the server certificate chain building code to use X509_verify_cert(),
4934 it used to have an ad-hoc builder which was unable to cope with anything
4935 other than a simple chain.
4936 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
4938 *) Don't check self signed certificate signatures in X509_verify_cert()
4939 by default (a flag can override this): it just wastes time without
4940 adding any security. As a useful side effect self signed root CAs
4941 with non-FIPS digests are now usable in FIPS mode.
4944 *) In dtls1_process_out_of_seq_message() the check if the current message
4945 is already buffered was missing. For every new message was memory
4946 allocated, allowing an attacker to perform an denial of service attack
4947 with sending out of seq handshake messages until there is no memory
4948 left. Additionally every future message was buffered, even if the
4949 sequence number made no sense and would be part of another handshake.
4950 So only messages with sequence numbers less than 10 in advance will be
4951 buffered. (CVE-2009-1378)
4952 [Robin Seggelmann, discovered by Daniel Mentz]
4954 *) Records are buffered if they arrive with a future epoch to be
4955 processed after finishing the corresponding handshake. There is
4956 currently no limitation to this buffer allowing an attacker to perform
4957 a DOS attack with sending records with future epochs until there is no
4958 memory left. This patch adds the pqueue_size() function to determine
4959 the size of a buffer and limits the record buffer to 100 entries.
4961 [Robin Seggelmann, discovered by Daniel Mentz]
4963 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
4964 parent structure is freed. (CVE-2009-1379)
4967 *) Handle non-blocking I/O properly in SSL_shutdown() call.
4968 [Darryl Miles <darryl-mailinglists@netbauds.net>]
4971 [Ilya O. <vrghost@gmail.com>]
4973 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
4975 *) Disable renegotiation completely - this fixes a severe security
4976 problem (CVE-2009-3555) at the cost of breaking all
4977 renegotiation. Renegotiation can be re-enabled by setting
4978 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
4979 run-time. This is really not recommended unless you know what
4983 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
4985 *) Don't set val to NULL when freeing up structures, it is freed up by
4986 underlying code. If sizeof(void *) > sizeof(long) this can result in
4987 zeroing past the valid field. (CVE-2009-0789)
4988 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
4990 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
4991 checked correctly. This would allow some invalid signed attributes to
4992 appear to verify correctly. (CVE-2009-0591)
4993 [Ivan Nestlerode <inestlerode@us.ibm.com>]
4995 *) Reject UniversalString and BMPString types with invalid lengths. This
4996 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
4997 a legal length. (CVE-2009-0590)
5000 *) Set S/MIME signing as the default purpose rather than setting it
5001 unconditionally. This allows applications to override it at the store
5005 *) Permit restricted recursion of ASN1 strings. This is needed in practice
5006 to handle some structures.
5009 *) Improve efficiency of mem_gets: don't search whole buffer each time
5011 [Jeremy Shapiro <jnshapir@us.ibm.com>]
5013 *) New -hex option for openssl rand.
5016 *) Print out UTF8String and NumericString when parsing ASN1.
5019 *) Support NumericString type for name components.
5022 *) Allow CC in the environment to override the automatically chosen
5023 compiler. Note that nothing is done to ensure flags work with the
5027 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
5029 *) Properly check EVP_VerifyFinal() and similar return values
5031 [Ben Laurie, Bodo Moeller, Google Security Team]
5033 *) Enable TLS extensions by default.
5036 *) Allow the CHIL engine to be loaded, whether the application is
5037 multithreaded or not. (This does not release the developer from the
5038 obligation to set up the dynamic locking callbacks.)
5039 [Sander Temme <sander@temme.net>]
5041 *) Use correct exit code if there is an error in dgst command.
5042 [Steve Henson; problem pointed out by Roland Dirlewanger]
5044 *) Tweak Configure so that you need to say "experimental-jpake" to enable
5045 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
5048 *) Add experimental JPAKE support, including demo authentication in
5049 s_client and s_server.
5052 *) Set the comparison function in v3_addr_canonize().
5053 [Rob Austein <sra@hactrn.net>]
5055 *) Add support for XMPP STARTTLS in s_client.
5056 [Philip Paeps <philip@freebsd.org>]
5058 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
5059 to ensure that even with this option, only ciphersuites in the
5060 server's preference list will be accepted. (Note that the option
5061 applies only when resuming a session, so the earlier behavior was
5062 just about the algorithm choice for symmetric cryptography.)
5065 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
5067 *) Fix NULL pointer dereference if a DTLS server received
5068 ChangeCipherSpec as first record (CVE-2009-1386).
5071 *) Fix a state transition in s3_srvr.c and d1_srvr.c
5072 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
5075 *) The fix in 0.9.8c that supposedly got rid of unsafe
5076 double-checked locking was incomplete for RSA blinding,
5077 addressing just one layer of what turns out to have been
5078 doubly unsafe triple-checked locking.
5080 So now fix this for real by retiring the MONT_HELPER macro
5081 in crypto/rsa/rsa_eay.c.
5083 [Bodo Moeller; problem pointed out by Marius Schilder]
5085 *) Various precautionary measures:
5087 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
5089 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
5090 (NB: This would require knowledge of the secret session ticket key
5091 to exploit, in which case you'd be SOL either way.)
5093 - Change bn_nist.c so that it will properly handle input BIGNUMs
5094 outside the expected range.
5096 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
5099 [Neel Mehta, Bodo Moeller]
5101 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
5102 the load fails. Useful for distros.
5103 [Ben Laurie and the FreeBSD team]
5105 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
5108 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
5111 *) Expand ENGINE to support engine supplied SSL client certificate functions.
5113 This work was sponsored by Logica.
5116 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
5117 keystores. Support for SSL/TLS client authentication too.
5118 Not compiled unless enable-capieng specified to Configure.
5120 This work was sponsored by Logica.
5123 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
5124 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
5125 attribute creation routines such as certificate requests and PKCS#12
5129 Changes between 0.9.8g and 0.9.8h [28 May 2008]
5131 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
5132 handshake which could lead to a client crash as found using the
5133 Codenomicon TLS test suite (CVE-2008-1672)
5134 [Steve Henson, Mark Cox]
5136 *) Fix double free in TLS server name extensions which could lead to
5137 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
5140 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
5142 Clear the error queue to ensure that error entries left from
5143 older function calls do not interfere with the correct operation.
5144 [Lutz Jaenicke, Erik de Castro Lopo]
5146 *) Remove root CA certificates of commercial CAs:
5148 The OpenSSL project does not recommend any specific CA and does not
5149 have any policy with respect to including or excluding any CA.
5150 Therefore it does not make any sense to ship an arbitrary selection
5151 of root CA certificates with the OpenSSL software.
5154 *) RSA OAEP patches to fix two separate invalid memory reads.
5155 The first one involves inputs when 'lzero' is greater than
5156 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
5157 before the beginning of from). The second one involves inputs where
5158 the 'db' section contains nothing but zeroes (there is a one-byte
5159 invalid read after the end of 'db').
5160 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5162 *) Partial backport from 0.9.9-dev:
5164 Introduce bn_mul_mont (dedicated Montgomery multiplication
5165 procedure) as a candidate for BIGNUM assembler implementation.
5166 While 0.9.9-dev uses assembler for various architectures, only
5167 x86_64 is available by default here in the 0.9.8 branch, and
5168 32-bit x86 is available through a compile-time setting.
5170 To try the 32-bit x86 assembler implementation, use Configure
5171 option "enable-montasm" (which exists only for this backport).
5173 As "enable-montasm" for 32-bit x86 disclaims code stability
5174 anyway, in this constellation we activate additional code
5175 backported from 0.9.9-dev for further performance improvements,
5176 namely BN_from_montgomery_word. (To enable this otherwise,
5177 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
5179 [Andy Polyakov (backport partially by Bodo Moeller)]
5181 *) Add TLS session ticket callback. This allows an application to set
5182 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
5183 values. This is useful for key rollover for example where several key
5184 sets may exist with different names.
5187 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
5188 This was broken until now in 0.9.8 releases, such that the only way
5189 a registered ENGINE could be used (assuming it initialises
5190 successfully on the host) was to explicitly set it as the default
5191 for the relevant algorithms. This is in contradiction with 0.9.7
5192 behaviour and the documentation. With this fix, when an ENGINE is
5193 registered into a given algorithm's table of implementations, the
5194 'uptodate' flag is reset so that auto-discovery will be used next
5195 time a new context for that algorithm attempts to select an
5197 [Ian Lister (tweaked by Geoff Thorpe)]
5199 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5200 implementation in the following ways:
5202 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5205 Lack of BER streaming support means one pass streaming processing is
5206 only supported if data is detached: setting the streaming flag is
5207 ignored for embedded content.
5209 CMS support is disabled by default and must be explicitly enabled
5210 with the enable-cms configuration option.
5213 *) Update the GMP engine glue to do direct copies between BIGNUM and
5214 mpz_t when openssl and GMP use the same limb size. Otherwise the
5215 existing "conversion via a text string export" trick is still used.
5216 [Paul Sheer <paulsheer@gmail.com>]
5218 *) Zlib compression BIO. This is a filter BIO which compressed and
5219 uncompresses any data passed through it.
5222 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5223 RFC3394 compatible AES key wrapping.
5226 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5227 sets string data without copying. X509_ALGOR_set0() and
5228 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5229 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5230 from an X509_ATTRIBUTE structure optionally checking it occurs only
5231 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5235 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5236 to get the expected BN_FLG_CONSTTIME behavior.
5237 [Bodo Moeller (Google)]
5241 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5242 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5243 - added some more tests to do_tests.pl
5244 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5245 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5246 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5247 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5248 - various changes to netware.pl to enable gcc-cross builds on Win32
5250 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5251 - various changes to fix missing prototype warnings
5252 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5253 - added AES, WHIRLPOOL and CPUID assembler code to build files
5254 - added missing AES assembler make rules to mk1mf.pl
5255 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5256 [Guenter Knauf <eflash@gmx.net>]
5258 *) Implement certificate status request TLS extension defined in RFC3546.
5259 A client can set the appropriate parameters and receive the encoded
5260 OCSP response via a callback. A server can query the supplied parameters
5261 and set the encoded OCSP response in the callback. Add simplified examples
5262 to s_client and s_server.
5265 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5267 *) Fix various bugs:
5268 + Binary incompatibility of ssl_ctx_st structure
5269 + DTLS interoperation with non-compliant servers
5270 + Don't call get_session_cb() without proposed session
5271 + Fix ia64 assembler code
5272 [Andy Polyakov, Steve Henson]
5274 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5276 *) DTLS Handshake overhaul. There were longstanding issues with
5277 OpenSSL DTLS implementation, which were making it impossible for
5278 RFC 4347 compliant client to communicate with OpenSSL server.
5279 Unfortunately just fixing these incompatibilities would "cut off"
5280 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5281 server keeps tolerating non RFC compliant syntax. The opposite is
5282 not true, 0.9.8f client can not communicate with earlier server.
5283 This update even addresses CVE-2007-4995.
5286 *) Changes to avoid need for function casts in OpenSSL: some compilers
5287 (gcc 4.2 and later) reject their use.
5288 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5291 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5292 RFC4507bis. The encrypted ticket format is an encrypted encoded
5293 SSL_SESSION structure, that way new session features are automatically
5296 If a client application caches session in an SSL_SESSION structure
5297 support is transparent because tickets are now stored in the encoded
5300 The SSL_CTX structure automatically generates keys for ticket
5301 protection in servers so again support should be possible
5302 with no application modification.
5304 If a client or server wishes to disable RFC4507 support then the option
5305 SSL_OP_NO_TICKET can be set.
5307 Add a TLS extension debugging callback to allow the contents of any client
5308 or server extensions to be examined.
5310 This work was sponsored by Google.
5313 *) Add initial support for TLS extensions, specifically for the server_name
5314 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5315 have new members for a host name. The SSL data structure has an
5316 additional member SSL_CTX *initial_ctx so that new sessions can be
5317 stored in that context to allow for session resumption, even after the
5318 SSL has been switched to a new SSL_CTX in reaction to a client's
5319 server_name extension.
5321 New functions (subject to change):
5323 SSL_get_servername()
5324 SSL_get_servername_type()
5327 New CTRL codes and macros (subject to change):
5329 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5330 - SSL_CTX_set_tlsext_servername_callback()
5331 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5332 - SSL_CTX_set_tlsext_servername_arg()
5333 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5335 openssl s_client has a new '-servername ...' option.
5337 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5338 '-key2 ...', '-servername_fatal' (subject to change). This allows
5339 testing the HostName extension for a specific single host name ('-cert'
5340 and '-key' remain fallbacks for handshakes without HostName
5341 negotiation). If the unrecognized_name alert has to be sent, this by
5342 default is a warning; it becomes fatal with the '-servername_fatal'
5345 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
5347 *) Add AES and SSE2 assembly language support to VC++ build.
5350 *) Mitigate attack on final subtraction in Montgomery reduction.
5353 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
5354 (which previously caused an internal error).
5357 *) Squeeze another 10% out of IGE mode when in != out.
5360 *) AES IGE mode speedup.
5361 [Dean Gaudet (Google)]
5363 *) Add the Korean symmetric 128-bit cipher SEED (see
5364 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
5365 add SEED ciphersuites from RFC 4162:
5367 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
5368 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
5369 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
5370 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
5372 To minimize changes between patchlevels in the OpenSSL 0.9.8
5373 series, SEED remains excluded from compilation unless OpenSSL
5374 is configured with 'enable-seed'.
5375 [KISA, Bodo Moeller]
5377 *) Mitigate branch prediction attacks, which can be practical if a
5378 single processor is shared, allowing a spy process to extract
5379 information. For detailed background information, see
5380 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
5381 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
5382 and Necessary Software Countermeasures"). The core of the change
5383 are new versions BN_div_no_branch() and
5384 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
5385 respectively, which are slower, but avoid the security-relevant
5386 conditional branches. These are automatically called by BN_div()
5387 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
5388 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
5389 remove a conditional branch.
5391 BN_FLG_CONSTTIME is the new name for the previous
5392 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
5393 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
5394 in the exponent causes BN_mod_exp_mont() to use the alternative
5395 implementation in BN_mod_exp_mont_consttime().) The old name
5396 remains as a deprecated alias.
5398 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
5399 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
5400 constant-time implementations for more than just exponentiation.
5401 Here too the old name is kept as a deprecated alias.
5403 BN_BLINDING_new() will now use BN_dup() for the modulus so that
5404 the BN_BLINDING structure gets an independent copy of the
5405 modulus. This means that the previous "BIGNUM *m" argument to
5406 BN_BLINDING_new() and to BN_BLINDING_create_param() now
5407 essentially becomes "const BIGNUM *m", although we can't actually
5408 change this in the header file before 0.9.9. It allows
5409 RSA_setup_blinding() to use BN_with_flags() on the modulus to
5410 enable BN_FLG_CONSTTIME.
5412 [Matthew D Wood (Intel Corp)]
5414 *) In the SSL/TLS server implementation, be strict about session ID
5415 context matching (which matters if an application uses a single
5416 external cache for different purposes). Previously,
5417 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
5418 set. This did ensure strict client verification, but meant that,
5419 with applications using a single external cache for quite
5420 different requirements, clients could circumvent ciphersuite
5421 restrictions for a given session ID context by starting a session
5422 in a different context.
5425 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5426 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5427 authentication-only ciphersuites.
5430 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
5431 not complete and could lead to a possible single byte overflow
5432 (CVE-2007-5135) [Ben Laurie]
5434 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
5436 *) Since AES128 and AES256 (and similarly Camellia128 and
5437 Camellia256) share a single mask bit in the logic of
5438 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5439 kludge to work properly if AES128 is available and AES256 isn't
5440 (or if Camellia128 is available and Camellia256 isn't).
5443 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
5444 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
5445 When a point or a seed is encoded in a BIT STRING, we need to
5446 prevent the removal of trailing zero bits to get the proper DER
5447 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
5448 of a NamedBitList, for which trailing 0 bits need to be removed.)
5451 *) Have SSL/TLS server implementation tolerate "mismatched" record
5452 protocol version while receiving ClientHello even if the
5453 ClientHello is fragmented. (The server can't insist on the
5454 particular protocol version it has chosen before the ServerHello
5455 message has informed the client about his choice.)
5458 *) Add RFC 3779 support.
5459 [Rob Austein for ARIN, Ben Laurie]
5461 *) Load error codes if they are not already present instead of using a
5462 static variable. This allows them to be cleanly unloaded and reloaded.
5463 Improve header file function name parsing.
5466 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
5467 or CAPABILITY handshake as required by RFCs.
5470 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5472 *) Introduce limits to prevent malicious keys being able to
5473 cause a denial of service. (CVE-2006-2940)
5474 [Steve Henson, Bodo Moeller]
5476 *) Fix ASN.1 parsing of certain invalid structures that can result
5477 in a denial of service. (CVE-2006-2937) [Steve Henson]
5479 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5480 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5482 *) Fix SSL client code which could crash if connecting to a
5483 malicious SSLv2 server. (CVE-2006-4343)
5484 [Tavis Ormandy and Will Drewry, Google Security Team]
5486 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
5487 match only those. Before that, "AES256-SHA" would be interpreted
5488 as a pattern and match "AES128-SHA" too (since AES128-SHA got
5489 the same strength classification in 0.9.7h) as we currently only
5490 have a single AES bit in the ciphersuite description bitmap.
5491 That change, however, also applied to ciphersuite strings such as
5492 "RC4-MD5" that intentionally matched multiple ciphersuites --
5493 namely, SSL 2.0 ciphersuites in addition to the more common ones
5494 from SSL 3.0/TLS 1.0.
5496 So we change the selection algorithm again: Naming an explicit
5497 ciphersuite selects this one ciphersuite, and any other similar
5498 ciphersuite (same bitmap) from *other* protocol versions.
5499 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5500 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5502 Since SSL 2.0 does not have any ciphersuites for which the
5503 128/256 bit distinction would be relevant, this works for now.
5504 The proper fix will be to use different bits for AES128 and
5505 AES256, which would have avoided the problems from the beginning;
5506 however, bits are scarce, so we can only do this in a new release
5507 (not just a patchlevel) when we can change the SSL_CIPHER
5508 definition to split the single 'unsigned long mask' bitmap into
5509 multiple values to extend the available space.
5513 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5515 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5516 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5518 *) Add AES IGE and biIGE modes.
5521 *) Change the Unix randomness entropy gathering to use poll() when
5522 possible instead of select(), since the latter has some
5523 undesirable limitations.
5524 [Darryl Miles via Richard Levitte and Bodo Moeller]
5526 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
5527 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
5528 cannot be implicitly activated as part of, e.g., the "AES" alias.
5529 However, please upgrade to OpenSSL 0.9.9[-dev] for
5530 non-experimental use of the ECC ciphersuites to get TLS extension
5531 support, which is required for curve and point format negotiation
5532 to avoid potential handshake problems.
5535 *) Disable rogue ciphersuites:
5537 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5538 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5539 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5541 The latter two were purportedly from
5542 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5545 Also deactivate the remaining ciphersuites from
5546 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5547 unofficial, and the ID has long expired.
5550 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5551 dual-core machines) and other potential thread-safety issues.
5554 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
5555 versions), which is now available for royalty-free use
5556 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
5557 Also, add Camellia TLS ciphersuites from RFC 4132.
5559 To minimize changes between patchlevels in the OpenSSL 0.9.8
5560 series, Camellia remains excluded from compilation unless OpenSSL
5561 is configured with 'enable-camellia'.
5564 *) Disable the padding bug check when compression is in use. The padding
5565 bug check assumes the first packet is of even length, this is not
5566 necessarily true if compression is enabled and can result in false
5567 positives causing handshake failure. The actual bug test is ancient
5568 code so it is hoped that implementations will either have fixed it by
5569 now or any which still have the bug do not support compression.
5572 Changes between 0.9.8a and 0.9.8b [04 May 2006]
5574 *) When applying a cipher rule check to see if string match is an explicit
5575 cipher suite and only match that one cipher suite if it is.
5578 *) Link in manifests for VC++ if needed.
5579 [Austin Ziegler <halostatue@gmail.com>]
5581 *) Update support for ECC-based TLS ciphersuites according to
5582 draft-ietf-tls-ecc-12.txt with proposed changes (but without
5583 TLS extensions, which are supported starting with the 0.9.9
5584 branch, not in the OpenSSL 0.9.8 branch).
5587 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
5588 opaque EVP_CIPHER_CTX handling.
5591 *) Fixes and enhancements to zlib compression code. We now only use
5592 "zlib1.dll" and use the default __cdecl calling convention on Win32
5593 to conform with the standards mentioned here:
5594 http://www.zlib.net/DLL_FAQ.txt
5595 Static zlib linking now works on Windows and the new --with-zlib-include
5596 --with-zlib-lib options to Configure can be used to supply the location
5597 of the headers and library. Gracefully handle case where zlib library
5601 *) Several fixes and enhancements to the OID generation code. The old code
5602 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
5603 handle numbers larger than ULONG_MAX, truncated printing and had a
5604 non standard OBJ_obj2txt() behaviour.
5607 *) Add support for building of engines under engine/ as shared libraries
5608 under VC++ build system.
5611 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
5612 Hopefully, we will not see any false combination of paths any more.
5615 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5617 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5618 (part of SSL_OP_ALL). This option used to disable the
5619 countermeasure against man-in-the-middle protocol-version
5620 rollback in the SSL 2.0 server implementation, which is a bad
5621 idea. (CVE-2005-2969)
5623 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5624 for Information Security, National Institute of Advanced Industrial
5625 Science and Technology [AIST], Japan)]
5627 *) Add two function to clear and return the verify parameter flags.
5630 *) Keep cipherlists sorted in the source instead of sorting them at
5631 runtime, thus removing the need for a lock.
5634 *) Avoid some small subgroup attacks in Diffie-Hellman.
5635 [Nick Mathewson and Ben Laurie]
5637 *) Add functions for well-known primes.
5640 *) Extended Windows CE support.
5641 [Satoshi Nakamura and Andy Polyakov]
5643 *) Initialize SSL_METHOD structures at compile time instead of during
5644 runtime, thus removing the need for a lock.
5647 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5648 attempting to decrypt each encrypted key in turn. Add support to
5652 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5654 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5657 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5660 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5661 key into the same file any more.
5664 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5667 *) Add -utf8 command line and config file option to 'ca'.
5668 [Stefan <stf@udoma.org]
5670 *) Removed the macro des_crypt(), as it seems to conflict with some
5671 libraries. Use DES_crypt().
5674 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5675 involves renaming the source and generated shared-libs for
5676 both. The engines will accept the corrected or legacy ids
5677 ('ncipher' and '4758_cca' respectively) when binding. NB,
5678 this only applies when building 'shared'.
5679 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5681 *) Add attribute functions to EVP_PKEY structure. Modify
5682 PKCS12_create() to recognize a CSP name attribute and
5683 use it. Make -CSP option work again in pkcs12 utility.
5686 *) Add new functionality to the bn blinding code:
5687 - automatic re-creation of the BN_BLINDING parameters after
5688 a fixed number of uses (currently 32)
5689 - add new function for parameter creation
5690 - introduce flags to control the update behaviour of the
5691 BN_BLINDING parameters
5692 - hide BN_BLINDING structure
5693 Add a second BN_BLINDING slot to the RSA structure to improve
5694 performance when a single RSA object is shared among several
5698 *) Add support for DTLS.
5699 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5701 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5702 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5705 *) Remove buggy and incomplete DH cert support from
5706 ssl/ssl_rsa.c and ssl/s3_both.c
5709 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5710 the apps/openssl applications.
5713 *) Compile clean with "-Wall -Wmissing-prototypes
5714 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5715 DEBUG_SAFESTACK must also be set.
5718 *) Change ./Configure so that certain algorithms can be disabled by default.
5719 The new counterpiece to "no-xxx" is "enable-xxx".
5721 The patented RC5 and MDC2 algorithms will now be disabled unless
5722 "enable-rc5" and "enable-mdc2", respectively, are specified.
5724 (IDEA remains enabled despite being patented. This is because IDEA
5725 is frequently required for interoperability, and there is no license
5726 fee for non-commercial use. As before, "no-idea" can be used to
5727 avoid this algorithm.)
5731 *) Add processing of proxy certificates (see RFC 3820). This work was
5732 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5733 EGEE (Enabling Grids for E-science in Europe).
5736 *) RC4 performance overhaul on modern architectures/implementations, such
5737 as Intel P4, IA-64 and AMD64.
5740 *) New utility extract-section.pl. This can be used specify an alternative
5741 section number in a pod file instead of having to treat each file as
5742 a separate case in Makefile. This can be done by adding two lines to the
5745 =for comment openssl_section:XXX
5747 The blank line is mandatory.
5751 *) New arguments -certform, -keyform and -pass for s_client and s_server
5752 to allow alternative format key and certificate files and passphrase
5756 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5757 update associated structures and add various utility functions.
5759 Add new policy related verify parameters, include policy checking in
5760 standard verify code. Enhance 'smime' application with extra parameters
5761 to support policy checking and print out.
5764 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
5765 Nehemiah processors. These extensions support AES encryption in hardware
5766 as well as RNG (though RNG support is currently disabled).
5767 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
5769 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
5772 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
5773 [Andy Polyakov and a number of other people]
5775 *) Improved PowerPC platform support. Most notably BIGNUM assembler
5776 implementation contributed by IBM.
5777 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
5779 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
5780 exponent rather than 'unsigned long'. There is a corresponding change to
5781 the new 'rsa_keygen' element of the RSA_METHOD structure.
5782 [Jelte Jansen, Geoff Thorpe]
5784 *) Functionality for creating the initial serial number file is now
5785 moved from CA.pl to the 'ca' utility with a new option -create_serial.
5787 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
5788 number file to 1, which is bound to cause problems. To avoid
5789 the problems while respecting compatibility between different 0.9.7
5790 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
5791 CA.pl for serial number initialization. With the new release 0.9.8,
5792 we can fix the problem directly in the 'ca' utility.)
5795 *) Reduced header interdependencies by declaring more opaque objects in
5796 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
5797 give fewer recursive includes, which could break lazy source code - so
5798 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
5799 developers should define this symbol when building and using openssl to
5800 ensure they track the recommended behaviour, interfaces, [etc], but
5801 backwards-compatible behaviour prevails when this isn't defined.
5804 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
5807 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
5808 This will generate a random key of the appropriate length based on the
5809 cipher context. The EVP_CIPHER can provide its own random key generation
5810 routine to support keys of a specific form. This is used in the des and
5811 3des routines to generate a key of the correct parity. Update S/MIME
5812 code to use new functions and hence generate correct parity DES keys.
5813 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
5814 valid (weak or incorrect parity).
5817 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
5818 as looking them up. This is useful when the verified structure may contain
5819 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
5820 present unless the new PKCS7_NO_CRL flag is asserted.
5823 *) Extend ASN1 oid configuration module. It now additionally accepts the
5826 shortName = some long name, 1.2.3.4
5829 *) Reimplemented the BN_CTX implementation. There is now no more static
5830 limitation on the number of variables it can handle nor the depth of the
5831 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
5832 information can now expand as required, and rather than having a single
5833 static array of bignums, BN_CTX now uses a linked-list of such arrays
5834 allowing it to expand on demand whilst maintaining the usefulness of
5835 BN_CTX's "bundling".
5838 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5839 to allow all RSA operations to function using a single BN_CTX.
5842 *) Preliminary support for certificate policy evaluation and checking. This
5843 is initially intended to pass the tests outlined in "Conformance Testing
5844 of Relying Party Client Certificate Path Processing Logic" v1.07.
5847 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
5848 remained unused and not that useful. A variety of other little bignum
5849 tweaks and fixes have also been made continuing on from the audit (see
5853 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
5854 associated ASN1, EVP and SSL functions and old ASN1 macros.
5857 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
5858 and this should never fail. So the return value from the use of
5859 BN_set_word() (which can fail due to needless expansion) is now deprecated;
5860 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
5863 *) BN_CTX_get() should return zero-valued bignums, providing the same
5864 initialised value as BN_new().
5865 [Geoff Thorpe, suggested by Ulf Möller]
5867 *) Support for inhibitAnyPolicy certificate extension.
5870 *) An audit of the BIGNUM code is underway, for which debugging code is
5871 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
5872 is considered valid when processing BIGNUMs, and causes execution to
5873 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
5874 further steps are taken to deliberately pollute unused data in BIGNUM
5875 structures to try and expose faulty code further on. For now, openssl will
5876 (in its default mode of operation) continue to tolerate the inconsistent
5877 forms that it has tolerated in the past, but authors and packagers should
5878 consider trying openssl and their own applications when compiled with
5879 these debugging symbols defined. It will help highlight potential bugs in
5880 their own code, and will improve the test coverage for OpenSSL itself. At
5881 some point, these tighter rules will become openssl's default to improve
5882 maintainability, though the assert()s and other overheads will remain only
5883 in debugging configurations. See bn.h for more details.
5884 [Geoff Thorpe, Nils Larsch, Ulf Möller]
5886 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
5887 that can only be obtained through BN_CTX_new() (which implicitly
5888 initialises it). The presence of this function only made it possible
5889 to overwrite an existing structure (and cause memory leaks).
5892 *) Because of the callback-based approach for implementing LHASH as a
5893 template type, lh_insert() adds opaque objects to hash-tables and
5894 lh_doall() or lh_doall_arg() are typically used with a destructor callback
5895 to clean up those corresponding objects before destroying the hash table
5896 (and losing the object pointers). So some over-zealous constifications in
5897 LHASH have been relaxed so that lh_insert() does not take (nor store) the
5898 objects as "const" and the lh_doall[_arg] callback wrappers are not
5899 prototyped to have "const" restrictions on the object pointers they are
5900 given (and so aren't required to cast them away any more).
5903 *) The tmdiff.h API was so ugly and minimal that our own timing utility
5904 (speed) prefers to use its own implementation. The two implementations
5905 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
5906 its object type properly exposed (MS_TM) instead of casting to/from "char
5907 *". This may still change yet if someone realises MS_TM and "ms_time_***"
5908 aren't necessarily the greatest nomenclatures - but this is what was used
5909 internally to the implementation so I've used that for now.
5912 *) Ensure that deprecated functions do not get compiled when
5913 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
5914 the self-tests were still using deprecated key-generation functions so
5915 these have been updated also.
5918 *) Reorganise PKCS#7 code to separate the digest location functionality
5919 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
5920 New function PKCS7_set_digest() to set the digest type for PKCS#7
5921 digestedData type. Add additional code to correctly generate the
5922 digestedData type and add support for this type in PKCS7 initialization
5926 *) New function PKCS7_set0_type_other() this initializes a PKCS7
5927 structure of type "other".
5930 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
5931 sure the loop does correctly stop and breaking ("division by zero")
5932 modulus operations are not performed. The (pre-generated) prime
5933 table crypto/bn/bn_prime.h was already correct, but it could not be
5934 re-generated on some platforms because of the "division by zero"
5935 situation in the script.
5936 [Ralf S. Engelschall]
5938 *) Update support for ECC-based TLS ciphersuites according to
5939 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
5940 SHA-1 now is only used for "small" curves (where the
5941 representation of a field element takes up to 24 bytes); for
5942 larger curves, the field element resulting from ECDH is directly
5943 used as premaster secret.
5944 [Douglas Stebila (Sun Microsystems Laboratories)]
5946 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
5947 curve secp160r1 to the tests.
5948 [Douglas Stebila (Sun Microsystems Laboratories)]
5950 *) Add the possibility to load symbols globally with DSO.
5951 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
5953 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
5954 control of the error stack.
5957 *) Add support for STORE in ENGINE.
5960 *) Add the STORE type. The intention is to provide a common interface
5961 to certificate and key stores, be they simple file-based stores, or
5962 HSM-type store, or LDAP stores, or...
5963 NOTE: The code is currently UNTESTED and isn't really used anywhere.
5966 *) Add a generic structure called OPENSSL_ITEM. This can be used to
5967 pass a list of arguments to any function as well as provide a way
5968 for a function to pass data back to the caller.
5971 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
5972 works like BUF_strdup() but can be used to duplicate a portion of
5973 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
5977 *) Add the function sk_find_ex() which works like sk_find(), but will
5978 return an index to an element even if an exact match couldn't be
5979 found. The index is guaranteed to point at the element where the
5980 searched-for key would be inserted to preserve sorting order.
5983 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
5984 takes an extra flags argument for optional functionality. Currently,
5985 the following flags are defined:
5987 OBJ_BSEARCH_VALUE_ON_NOMATCH
5988 This one gets OBJ_bsearch_ex() to return a pointer to the first
5989 element where the comparing function returns a negative or zero
5992 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
5993 This one gets OBJ_bsearch_ex() to return a pointer to the first
5994 element where the comparing function returns zero. This is useful
5995 if there are more than one element where the comparing function
5999 *) Make it possible to create self-signed certificates with 'openssl ca'
6000 in such a way that the self-signed certificate becomes part of the
6001 CA database and uses the same mechanisms for serial number generation
6002 as all other certificate signing. The new flag '-selfsign' enables
6003 this functionality. Adapt CA.sh and CA.pl.in.
6006 *) Add functionality to check the public key of a certificate request
6007 against a given private. This is useful to check that a certificate
6008 request can be signed by that key (self-signing).
6011 *) Make it possible to have multiple active certificates with the same
6012 subject in the CA index file. This is done only if the keyword
6013 'unique_subject' is set to 'no' in the main CA section (default
6014 if 'CA_default') of the configuration file. The value is saved
6015 with the database itself in a separate index attribute file,
6016 named like the index file with '.attr' appended to the name.
6019 *) Generate multi-valued AVAs using '+' notation in config files for
6023 *) Support for nameConstraints certificate extension.
6026 *) Support for policyConstraints certificate extension.
6029 *) Support for policyMappings certificate extension.
6032 *) Make sure the default DSA_METHOD implementation only uses its
6033 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
6034 and change its own handlers to be NULL so as to remove unnecessary
6035 indirection. This lets alternative implementations fallback to the
6036 default implementation more easily.
6039 *) Support for directoryName in GeneralName related extensions
6043 *) Make it possible to link applications using Makefile.shared.
6044 Make that possible even when linking against static libraries!
6047 *) Support for single pass processing for S/MIME signing. This now
6048 means that S/MIME signing can be done from a pipe, in addition
6049 cleartext signing (multipart/signed type) is effectively streaming
6050 and the signed data does not need to be all held in memory.
6052 This is done with a new flag PKCS7_STREAM. When this flag is set
6053 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
6054 is done after the data is output (and digests calculated) in
6055 SMIME_write_PKCS7().
6058 *) Add full support for -rpath/-R, both in shared libraries and
6059 applications, at least on the platforms where it's known how
6063 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
6064 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
6065 will now compute a table of multiples of the generator that
6066 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
6067 faster (notably in the case of a single point multiplication,
6068 scalar * generator).
6069 [Nils Larsch, Bodo Moeller]
6071 *) IPv6 support for certificate extensions. The various extensions
6072 which use the IP:a.b.c.d can now take IPv6 addresses using the
6073 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
6077 *) Added an ENGINE that implements RSA by performing private key
6078 exponentiations with the GMP library. The conversions to and from
6079 GMP's mpz_t format aren't optimised nor are any montgomery forms
6080 cached, and on x86 it appears OpenSSL's own performance has caught up.
6081 However there are likely to be other architectures where GMP could
6082 provide a boost. This ENGINE is not built in by default, but it can be
6083 specified at Configure time and should be accompanied by the necessary
6084 linker additions, eg;
6085 ./config -DOPENSSL_USE_GMP -lgmp
6088 *) "openssl engine" will not display ENGINE/DSO load failure errors when
6089 testing availability of engines with "-t" - the old behaviour is
6090 produced by increasing the feature's verbosity with "-tt".
6093 *) ECDSA routines: under certain error conditions uninitialized BN objects
6094 could be freed. Solution: make sure initialization is performed early
6095 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
6099 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
6100 and DH_METHOD (eg. by ENGINE implementations) to override the normal
6101 software implementations. For DSA and DH, parameter generation can
6102 also be overridden by providing the appropriate method callbacks.
6105 *) Change the "progress" mechanism used in key-generation and
6106 primality testing to functions that take a new BN_GENCB pointer in
6107 place of callback/argument pairs. The new API functions have "_ex"
6108 postfixes and the older functions are reimplemented as wrappers for
6109 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
6110 declarations of the old functions to help (graceful) attempts to
6111 migrate to the new functions. Also, the new key-generation API
6112 functions operate on a caller-supplied key-structure and return
6113 success/failure rather than returning a key or NULL - this is to
6114 help make "keygen" another member function of RSA_METHOD etc.
6116 Example for using the new callback interface:
6118 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
6122 BN_GENCB_set(&my_cb, my_callback, my_arg);
6124 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
6125 /* For the meaning of a, b in calls to my_callback(), see the
6126 * documentation of the function that calls the callback.
6127 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
6128 * my_callback should return 1 if it wants BN_is_prime_ex()
6129 * to continue, or 0 to stop.
6134 *) Change the ZLIB compression method to be stateful, and make it
6135 available to TLS with the number defined in
6136 draft-ietf-tls-compression-04.txt.
6139 *) Add the ASN.1 structures and functions for CertificatePair, which
6140 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
6142 CertificatePair ::= SEQUENCE {
6143 forward [0] Certificate OPTIONAL,
6144 reverse [1] Certificate OPTIONAL,
6145 -- at least one of the pair shall be present -- }
6147 Also implement the PEM functions to read and write certificate
6148 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
6150 This needed to be defined, mostly for the sake of the LDAP
6151 attribute crossCertificatePair, but may prove useful elsewhere as
6155 *) Make it possible to inhibit symlinking of shared libraries in
6156 Makefile.shared, for Cygwin's sake.
6159 *) Extend the BIGNUM API by creating a function
6160 void BN_set_negative(BIGNUM *a, int neg);
6161 and a macro that behave like
6162 int BN_is_negative(const BIGNUM *a);
6164 to avoid the need to access 'a->neg' directly in applications.
6167 *) Implement fast modular reduction for pseudo-Mersenne primes
6168 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
6169 EC_GROUP_new_curve_GFp() will now automatically use this
6171 [Nils Larsch <nla@trustcenter.de>]
6173 *) Add new lock type (CRYPTO_LOCK_BN).
6176 *) Change the ENGINE framework to automatically load engines
6177 dynamically from specific directories unless they could be
6178 found to already be built in or loaded. Move all the
6179 current engines except for the cryptodev one to a new
6181 The engines in engines/ are built as shared libraries if
6182 the "shared" options was given to ./Configure or ./config.
6183 Otherwise, they are inserted in libcrypto.a.
6184 /usr/local/ssl/engines is the default directory for dynamic
6185 engines, but that can be overridden at configure time through
6186 the usual use of --prefix and/or --openssldir, and at run
6187 time with the environment variable OPENSSL_ENGINES.
6188 [Geoff Thorpe and Richard Levitte]
6190 *) Add Makefile.shared, a helper makefile to build shared
6191 libraries. Adapt Makefile.org.
6194 *) Add version info to Win32 DLLs.
6195 [Peter 'Luna' Runestig" <peter@runestig.com>]
6197 *) Add new 'medium level' PKCS#12 API. Certificates and keys
6198 can be added using this API to created arbitrary PKCS#12
6199 files while avoiding the low level API.
6201 New options to PKCS12_create(), key or cert can be NULL and
6202 will then be omitted from the output file. The encryption
6203 algorithm NIDs can be set to -1 for no encryption, the mac
6204 iteration count can be set to 0 to omit the mac.
6206 Enhance pkcs12 utility by making the -nokeys and -nocerts
6207 options work when creating a PKCS#12 file. New option -nomac
6208 to omit the mac, NONE can be set for an encryption algorithm.
6209 New code is modified to use the enhanced PKCS12_create()
6210 instead of the low level API.
6213 *) Extend ASN1 encoder to support indefinite length constructed
6214 encoding. This can output sequences tags and octet strings in
6215 this form. Modify pk7_asn1.c to support indefinite length
6216 encoding. This is experimental and needs additional code to
6217 be useful, such as an ASN1 bio and some enhanced streaming
6220 Extend template encode functionality so that tagging is passed
6221 down to the template encoder.
6224 *) Let 'openssl req' fail if an argument to '-newkey' is not
6225 recognized instead of using RSA as a default.
6228 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6229 As these are not official, they are not included in "ALL";
6230 the "ECCdraft" ciphersuite group alias can be used to select them.
6231 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6233 *) Add ECDH engine support.
6234 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6236 *) Add ECDH in new directory crypto/ecdh/.
6237 [Douglas Stebila (Sun Microsystems Laboratories)]
6239 *) Let BN_rand_range() abort with an error after 100 iterations
6240 without success (which indicates a broken PRNG).
6243 *) Change BN_mod_sqrt() so that it verifies that the input value
6244 is really the square of the return value. (Previously,
6245 BN_mod_sqrt would show GIGO behaviour.)
6248 *) Add named elliptic curves over binary fields from X9.62, SECG,
6249 and WAP/WTLS; add OIDs that were still missing.
6251 [Sheueling Chang Shantz and Douglas Stebila
6252 (Sun Microsystems Laboratories)]
6254 *) Extend the EC library for elliptic curves over binary fields
6255 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6258 EC_GF2m_simple_method
6262 EC_GROUP_new_curve_GF2m
6263 EC_GROUP_set_curve_GF2m
6264 EC_GROUP_get_curve_GF2m
6265 EC_POINT_set_affine_coordinates_GF2m
6266 EC_POINT_get_affine_coordinates_GF2m
6267 EC_POINT_set_compressed_coordinates_GF2m
6269 Point compression for binary fields is disabled by default for
6270 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6273 As binary polynomials are represented as BIGNUMs, various members
6274 of the EC_GROUP and EC_POINT data structures can be shared
6275 between the implementations for prime fields and binary fields;
6276 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6277 are essentially identical to their ..._GFp counterparts.
6278 (For simplicity, the '..._GFp' prefix has been dropped from
6279 various internal method names.)
6281 An internal 'field_div' method (similar to 'field_mul' and
6282 'field_sqr') has been added; this is used only for binary fields.
6284 [Sheueling Chang Shantz and Douglas Stebila
6285 (Sun Microsystems Laboratories)]
6287 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6288 through methods ('mul', 'precompute_mult').
6290 The generic implementations (now internally called 'ec_wNAF_mul'
6291 and 'ec_wNAF_precomputed_mult') remain the default if these
6292 methods are undefined.
6294 [Sheueling Chang Shantz and Douglas Stebila
6295 (Sun Microsystems Laboratories)]
6297 *) New function EC_GROUP_get_degree, which is defined through
6298 EC_METHOD. For curves over prime fields, this returns the bit
6299 length of the modulus.
6301 [Sheueling Chang Shantz and Douglas Stebila
6302 (Sun Microsystems Laboratories)]
6304 *) New functions EC_GROUP_dup, EC_POINT_dup.
6305 (These simply call ..._new and ..._copy).
6307 [Sheueling Chang Shantz and Douglas Stebila
6308 (Sun Microsystems Laboratories)]
6310 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6311 Polynomials are represented as BIGNUMs (where the sign bit is not
6312 used) in the following functions [macros]:
6315 BN_GF2m_sub [= BN_GF2m_add]
6316 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6317 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6318 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6320 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6321 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6322 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6323 BN_GF2m_cmp [= BN_ucmp]
6325 (Note that only the 'mod' functions are actually for fields GF(2^m).
6326 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6328 For some functions, an the irreducible polynomial defining a
6329 field can be given as an 'unsigned int[]' with strictly
6330 decreasing elements giving the indices of those bits that are set;
6331 i.e., p[] represents the polynomial
6332 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6334 p[0] > p[1] > ... > p[k] = 0.
6335 This applies to the following functions:
6340 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
6341 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
6343 BN_GF2m_mod_sqrt_arr
6344 BN_GF2m_mod_solve_quad_arr
6348 Conversion can be performed by the following functions:
6353 bntest.c has additional tests for binary polynomial arithmetic.
6355 Two implementations for BN_GF2m_mod_div() are available.
6356 The default algorithm simply uses BN_GF2m_mod_inv() and
6357 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
6358 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
6359 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
6361 [Sheueling Chang Shantz and Douglas Stebila
6362 (Sun Microsystems Laboratories)]
6364 *) Add new error code 'ERR_R_DISABLED' that can be used when some
6365 functionality is disabled at compile-time.
6366 [Douglas Stebila <douglas.stebila@sun.com>]
6368 *) Change default behaviour of 'openssl asn1parse' so that more
6369 information is visible when viewing, e.g., a certificate:
6371 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
6372 mode the content of non-printable OCTET STRINGs is output in a
6373 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
6374 avoid the appearance of a printable string.
6375 [Nils Larsch <nla@trustcenter.de>]
6377 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
6379 EC_GROUP_set_asn1_flag()
6380 EC_GROUP_get_asn1_flag()
6381 EC_GROUP_set_point_conversion_form()
6382 EC_GROUP_get_point_conversion_form()
6383 These control ASN1 encoding details:
6384 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
6385 has been set to OPENSSL_EC_NAMED_CURVE.
6386 - Points are encoded in uncompressed form by default; options for
6387 asn1_for are as for point2oct, namely
6388 POINT_CONVERSION_COMPRESSED
6389 POINT_CONVERSION_UNCOMPRESSED
6390 POINT_CONVERSION_HYBRID
6392 Also add 'seed' and 'seed_len' members to EC_GROUP with access
6395 EC_GROUP_get0_seed()
6396 EC_GROUP_get_seed_len()
6397 This is used only for ASN1 purposes (so far).
6398 [Nils Larsch <nla@trustcenter.de>]
6400 *) Add 'field_type' member to EC_METHOD, which holds the NID
6401 of the appropriate field type OID. The new function
6402 EC_METHOD_get_field_type() returns this value.
6403 [Nils Larsch <nla@trustcenter.de>]
6408 EC_POINT_point2hex()
6409 EC_POINT_hex2point()
6410 providing useful interfaces to EC_POINT_point2oct() and
6411 EC_POINT_oct2point().
6412 [Nils Larsch <nla@trustcenter.de>]
6414 *) Change internals of the EC library so that the functions
6415 EC_GROUP_set_generator()
6416 EC_GROUP_get_generator()
6417 EC_GROUP_get_order()
6418 EC_GROUP_get_cofactor()
6419 are implemented directly in crypto/ec/ec_lib.c and not dispatched
6420 to methods, which would lead to unnecessary code duplication when
6421 adding different types of curves.
6422 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
6424 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
6425 arithmetic, and such that modified wNAFs are generated
6426 (which avoid length expansion in many cases).
6429 *) Add a function EC_GROUP_check_discriminant() (defined via
6430 EC_METHOD) that verifies that the curve discriminant is non-zero.
6432 Add a function EC_GROUP_check() that makes some sanity tests
6433 on a EC_GROUP, its generator and order. This includes
6434 EC_GROUP_check_discriminant().
6435 [Nils Larsch <nla@trustcenter.de>]
6437 *) Add ECDSA in new directory crypto/ecdsa/.
6439 Add applications 'openssl ecparam' and 'openssl ecdsa'
6440 (these are based on 'openssl dsaparam' and 'openssl dsa').
6442 ECDSA support is also included in various other files across the
6443 library. Most notably,
6444 - 'openssl req' now has a '-newkey ecdsa:file' option;
6445 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
6446 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
6447 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
6448 them suitable for ECDSA where domain parameters must be
6449 extracted before the specific public key;
6450 - ECDSA engine support has been added.
6451 [Nils Larsch <nla@trustcenter.de>]
6453 *) Include some named elliptic curves, and add OIDs from X9.62,
6454 SECG, and WAP/WTLS. Each curve can be obtained from the new
6456 EC_GROUP_new_by_curve_name(),
6457 and the list of available named curves can be obtained with
6458 EC_get_builtin_curves().
6459 Also add a 'curve_name' member to EC_GROUP objects, which can be
6461 EC_GROUP_set_curve_name()
6462 EC_GROUP_get_curve_name()
6463 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
6465 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6466 was actually never needed) and in BN_mul(). The removal in BN_mul()
6467 required a small change in bn_mul_part_recursive() and the addition
6468 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6469 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6470 bn_sub_words() and bn_add_words() except they take arrays with
6474 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
6476 *) Cleanse PEM buffers before freeing them since they may contain
6478 [Benjamin Bennett <ben@psc.edu>]
6480 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6481 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6482 authentication-only ciphersuites.
6485 *) Since AES128 and AES256 share a single mask bit in the logic of
6486 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6487 kludge to work properly if AES128 is available and AES256 isn't.
6490 *) Expand security boundary to match 1.1.1 module.
6493 *) Remove redundant features: hash file source, editing of test vectors
6494 modify fipsld to use external fips_premain.c signature.
6497 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6498 run algorithm test programs.
6501 *) Make algorithm test programs more tolerant of whitespace.
6504 *) Have SSL/TLS server implementation tolerate "mismatched" record
6505 protocol version while receiving ClientHello even if the
6506 ClientHello is fragmented. (The server can't insist on the
6507 particular protocol version it has chosen before the ServerHello
6508 message has informed the client about his choice.)
6511 *) Load error codes if they are not already present instead of using a
6512 static variable. This allows them to be cleanly unloaded and reloaded.
6515 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6517 *) Introduce limits to prevent malicious keys being able to
6518 cause a denial of service. (CVE-2006-2940)
6519 [Steve Henson, Bodo Moeller]
6521 *) Fix ASN.1 parsing of certain invalid structures that can result
6522 in a denial of service. (CVE-2006-2937) [Steve Henson]
6524 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6525 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6527 *) Fix SSL client code which could crash if connecting to a
6528 malicious SSLv2 server. (CVE-2006-4343)
6529 [Tavis Ormandy and Will Drewry, Google Security Team]
6531 *) Change ciphersuite string processing so that an explicit
6532 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
6533 will no longer include "AES128-SHA"), and any other similar
6534 ciphersuite (same bitmap) from *other* protocol versions (so that
6535 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
6536 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
6537 changes from 0.9.8b and 0.9.8d.
6540 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
6542 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6543 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6545 *) Change the Unix randomness entropy gathering to use poll() when
6546 possible instead of select(), since the latter has some
6547 undesirable limitations.
6548 [Darryl Miles via Richard Levitte and Bodo Moeller]
6550 *) Disable rogue ciphersuites:
6552 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6553 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6554 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6556 The latter two were purportedly from
6557 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6560 Also deactivate the remaining ciphersuites from
6561 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6562 unofficial, and the ID has long expired.
6565 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6566 dual-core machines) and other potential thread-safety issues.
6569 Changes between 0.9.7i and 0.9.7j [04 May 2006]
6571 *) Adapt fipsld and the build system to link against the validated FIPS
6572 module in FIPS mode.
6575 *) Fixes for VC++ 2005 build under Windows.
6578 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
6579 from a Windows bash shell such as MSYS. It is autodetected from the
6580 "config" script when run from a VC++ environment. Modify standard VC++
6581 build to use fipscanister.o from the GNU make build.
6584 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
6586 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
6587 The value now differs depending on if you build for FIPS or not.
6588 BEWARE! A program linked with a shared FIPSed libcrypto can't be
6589 safely run with a non-FIPSed libcrypto, as it may crash because of
6590 the difference induced by this change.
6593 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
6595 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6596 (part of SSL_OP_ALL). This option used to disable the
6597 countermeasure against man-in-the-middle protocol-version
6598 rollback in the SSL 2.0 server implementation, which is a bad
6599 idea. (CVE-2005-2969)
6601 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6602 for Information Security, National Institute of Advanced Industrial
6603 Science and Technology [AIST], Japan)]
6605 *) Minimal support for X9.31 signatures and PSS padding modes. This is
6606 mainly for FIPS compliance and not fully integrated at this stage.
6609 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
6610 the exponentiation using a fixed-length exponent. (Otherwise,
6611 the information leaked through timing could expose the secret key
6612 after many signatures; cf. Bleichenbacher's attack on DSA with
6616 *) Make a new fixed-window mod_exp implementation the default for
6617 RSA, DSA, and DH private-key operations so that the sequence of
6618 squares and multiplies and the memory access pattern are
6619 independent of the particular secret key. This will mitigate
6620 cache-timing and potential related attacks.
6622 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6623 and this is automatically used by BN_mod_exp_mont() if the new flag
6624 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6625 will use this BN flag for private exponents unless the flag
6626 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6627 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6629 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6631 *) Change the client implementation for SSLv23_method() and
6632 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6633 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6634 (Previously, the SSL 2.0 backwards compatible Client Hello
6635 message format would be used even with SSL_OP_NO_SSLv2.)
6638 *) Add support for smime-type MIME parameter in S/MIME messages which some
6642 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6643 a threadsafe manner. Modify rsa code to use new function and add calls
6644 to dsa and dh code (which had race conditions before).
6647 *) Include the fixed error library code in the C error file definitions
6648 instead of fixing them up at runtime. This keeps the error code
6649 structures constant.
6652 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6654 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6657 *) Fixes for newer kerberos headers. NB: the casts are needed because
6658 the 'length' field is signed on one version and unsigned on another
6659 with no (?) obvious way to tell the difference, without these VC++
6660 complains. Also the "definition" of FAR (blank) is no longer included
6661 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6662 some needed definitions.
6665 *) Undo Cygwin change.
6668 *) Added support for proxy certificates according to RFC 3820.
6669 Because they may be a security thread to unaware applications,
6670 they must be explicitly allowed in run-time. See
6671 docs/HOWTO/proxy_certificates.txt for further information.
6674 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6676 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6677 server and client random values. Previously
6678 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6679 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6681 This change has negligible security impact because:
6683 1. Server and client random values still have 24 bytes of pseudo random
6686 2. Server and client random values are sent in the clear in the initial
6689 3. The master secret is derived using the premaster secret (48 bytes in
6690 size for static RSA ciphersuites) as well as client server and random
6693 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6696 [Stephen Henson, reported by UK NISCC]
6698 *) Use Windows randomness collection on Cygwin.
6701 *) Fix hang in EGD/PRNGD query when communication socket is closed
6702 prematurely by EGD/PRNGD.
6703 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6705 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6708 *) Back-port of selected performance improvements from development
6709 branch, as well as improved support for PowerPC platforms.
6712 *) Add lots of checks for memory allocation failure, error codes to indicate
6713 failure and freeing up memory if a failure occurs.
6714 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6716 *) Add new -passin argument to dgst.
6719 *) Perform some character comparisons of different types in X509_NAME_cmp:
6720 this is needed for some certificates that re-encode DNs into UTF8Strings
6721 (in violation of RFC3280) and can't or won't issue name rollover
6725 *) Make an explicit check during certificate validation to see that
6726 the CA setting in each certificate on the chain is correct. As a
6727 side effect always do the following basic checks on extensions,
6728 not just when there's an associated purpose to the check:
6730 - if there is an unhandled critical extension (unless the user
6731 has chosen to ignore this fault)
6732 - if the path length has been exceeded (if one is set at all)
6733 - that certain extensions fit the associated purpose (if one has
6737 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6739 *) Avoid a race condition when CRLs are checked in a multi threaded
6740 environment. This would happen due to the reordering of the revoked
6741 entries during signature checking and serial number lookup. Now the
6742 encoding is cached and the serial number sort performed under a lock.
6743 Add new STACK function sk_is_sorted().
6746 *) Add Delta CRL to the extension code.
6749 *) Various fixes to s3_pkt.c so alerts are sent properly.
6750 [David Holmes <d.holmes@f5.com>]
6752 *) Reduce the chances of duplicate issuer name and serial numbers (in
6753 violation of RFC3280) using the OpenSSL certificate creation utilities.
6754 This is done by creating a random 64 bit value for the initial serial
6755 number when a serial number file is created or when a self signed
6756 certificate is created using 'openssl req -x509'. The initial serial
6757 number file is created using 'openssl x509 -next_serial' in CA.pl
6758 rather than being initialized to 1.
6761 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
6763 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6764 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6765 [Joe Orton, Steve Henson]
6767 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
6769 [Joe Orton, Steve Henson]
6771 *) Make it possible to have multiple active certificates with the same
6772 subject in the CA index file. This is done only if the keyword
6773 'unique_subject' is set to 'no' in the main CA section (default
6774 if 'CA_default') of the configuration file. The value is saved
6775 with the database itself in a separate index attribute file,
6776 named like the index file with '.attr' appended to the name.
6779 *) X509 verify fixes. Disable broken certificate workarounds when
6780 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
6781 keyUsage extension present. Don't accept CRLs with unhandled critical
6782 extensions: since verify currently doesn't process CRL extensions this
6783 rejects a CRL with *any* critical extensions. Add new verify error codes
6787 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
6788 A clarification of RFC2560 will require the use of OCTET STRINGs and
6789 some implementations cannot handle the current raw format. Since OpenSSL
6790 copies and compares OCSP nonces as opaque blobs without any attempt at
6791 parsing them this should not create any compatibility issues.
6794 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
6795 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
6796 this HMAC (and other) operations are several times slower than OpenSSL
6800 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
6801 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6803 *) Use the correct content when signing type "other".
6806 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
6808 *) Fix various bugs revealed by running the NISCC test suite:
6810 Stop out of bounds reads in the ASN1 code when presented with
6811 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6813 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
6815 If verify callback ignores invalid public key errors don't try to check
6816 certificate signature with the NULL public key.
6820 *) New -ignore_err option in ocsp application to stop the server
6821 exiting on the first error in a request.
6824 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6825 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6829 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6830 extra data after the compression methods not only for TLS 1.0
6831 but also for SSL 3.0 (as required by the specification).
6832 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6834 *) Change X509_certificate_type() to mark the key as exported/exportable
6835 when it's 512 *bits* long, not 512 bytes.
6838 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6839 blocks during encryption.
6842 *) Various fixes to base64 BIO and non blocking I/O. On write
6843 flushes were not handled properly if the BIO retried. On read
6844 data was not being buffered properly and had various logic bugs.
6845 This also affects blocking I/O when the data being decoded is a
6849 *) Various S/MIME bugfixes and compatibility changes:
6850 output correct application/pkcs7 MIME type if
6851 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
6852 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
6853 of files as .eml work). Correctly handle very long lines in MIME
6857 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
6859 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6860 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6861 a protocol version number mismatch like a decryption error
6862 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6865 *) Turn on RSA blinding by default in the default implementation
6866 to avoid a timing attack. Applications that don't want it can call
6867 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6868 They would be ill-advised to do so in most cases.
6869 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6871 *) Change RSA blinding code so that it works when the PRNG is not
6872 seeded (in this case, the secret RSA exponent is abused as
6873 an unpredictable seed -- if it is not unpredictable, there
6874 is no point in blinding anyway). Make RSA blinding thread-safe
6875 by remembering the creator's thread ID in rsa->blinding and
6876 having all other threads use local one-time blinding factors
6877 (this requires more computation than sharing rsa->blinding, but
6878 avoids excessive locking; and if an RSA object is not shared
6879 between threads, blinding will still be very fast).
6882 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
6883 ENGINE as defaults for all supported algorithms irrespective of
6884 the 'flags' parameter. 'flags' is now honoured, so applications
6885 should make sure they are passing it correctly.
6888 *) Target "mingw" now allows native Windows code to be generated in
6889 the Cygwin environment as well as with the MinGW compiler.
6892 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
6894 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6895 via timing by performing a MAC computation even if incorrect
6896 block cipher padding has been found. This is a countermeasure
6897 against active attacks where the attacker has to distinguish
6898 between bad padding and a MAC verification error. (CVE-2003-0078)
6900 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6901 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6902 Martin Vuagnoux (EPFL, Ilion)]
6904 *) Make the no-err option work as intended. The intention with no-err
6905 is not to have the whole error stack handling routines removed from
6906 libcrypto, it's only intended to remove all the function name and
6907 reason texts, thereby removing some of the footprint that may not
6908 be interesting if those errors aren't displayed anyway.
6910 NOTE: it's still possible for any application or module to have its
6911 own set of error texts inserted. The routines are there, just not
6912 used by default when no-err is given.
6915 *) Add support for FreeBSD on IA64.
6916 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
6918 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
6919 Kerberos function mit_des_cbc_cksum(). Before this change,
6920 the value returned by DES_cbc_cksum() was like the one from
6921 mit_des_cbc_cksum(), except the bytes were swapped.
6922 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
6924 *) Allow an application to disable the automatic SSL chain building.
6925 Before this a rather primitive chain build was always performed in
6926 ssl3_output_cert_chain(): an application had no way to send the
6927 correct chain if the automatic operation produced an incorrect result.
6929 Now the chain builder is disabled if either:
6931 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
6933 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
6935 The reasoning behind this is that an application would not want the
6936 auto chain building to take place if extra chain certificates are
6937 present and it might also want a means of sending no additional
6938 certificates (for example the chain has two certificates and the
6942 *) Add the possibility to build without the ENGINE framework.
6943 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
6945 *) Under Win32 gmtime() can return NULL: check return value in
6946 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
6949 *) DSA routines: under certain error conditions uninitialized BN objects
6950 could be freed. Solution: make sure initialization is performed early
6951 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
6952 Nils Larsch <nla@trustcenter.de> via PR#459)
6955 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
6956 checked on reconnect on the client side, therefore session resumption
6957 could still fail with a "ssl session id is different" error. This
6958 behaviour is masked when SSL_OP_ALL is used due to
6959 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
6960 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6961 followup to PR #377.
6964 *) IA-32 assembler support enhancements: unified ELF targets, support
6965 for SCO/Caldera platforms, fix for Cygwin shared build.
6968 *) Add support for FreeBSD on sparc64. As a consequence, support for
6969 FreeBSD on non-x86 processors is separate from x86 processors on
6970 the config script, much like the NetBSD support.
6971 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
6973 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
6975 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
6978 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
6979 code (06) was taken as the first octet of the session ID and the last
6980 octet was ignored consequently. As a result SSLv2 client side session
6981 caching could not have worked due to the session ID mismatch between
6983 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
6987 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
6988 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
6992 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
6993 seems that in spite of existing for more than a year, many application
6994 author have done nothing to provide the necessary callbacks, which
6995 means that this particular engine will not work properly anywhere.
6996 This is a very unfortunate situation which forces us, in the name
6997 of usability, to give the hw_ncipher.c a static lock, which is part
6999 NOTE: This is for the 0.9.7 series ONLY. This hack will never
7000 appear in 0.9.8 or later. We EXPECT application authors to have
7001 dealt properly with this when 0.9.8 is released (unless we actually
7002 make such changes in the libcrypto locking code that changes will
7003 have to be made anyway).
7006 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
7007 octets have been read, EOF or an error occurs. Without this change
7008 some truncated ASN1 structures will not produce an error.
7011 *) Disable Heimdal support, since it hasn't been fully implemented.
7012 Still give the possibility to force the use of Heimdal, but with
7013 warnings and a request that patches get sent to openssl-dev.
7016 *) Add the VC-CE target, introduce the WINCE sysname, and add
7017 INSTALL.WCE and appropriate conditionals to make it build.
7018 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7020 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
7021 cygssl-x.y.z.dll, where x, y and z are the major, minor and
7022 edit numbers of the version.
7023 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7025 *) Introduce safe string copy and catenation functions
7026 (BUF_strlcpy() and BUF_strlcat()).
7027 [Ben Laurie (CHATS) and Richard Levitte]
7029 *) Avoid using fixed-size buffers for one-line DNs.
7030 [Ben Laurie (CHATS)]
7032 *) Add BUF_MEM_grow_clean() to avoid information leakage when
7033 resizing buffers containing secrets, and use where appropriate.
7034 [Ben Laurie (CHATS)]
7036 *) Avoid using fixed size buffers for configuration file location.
7037 [Ben Laurie (CHATS)]
7039 *) Avoid filename truncation for various CA files.
7040 [Ben Laurie (CHATS)]
7042 *) Use sizeof in preference to magic numbers.
7043 [Ben Laurie (CHATS)]
7045 *) Avoid filename truncation in cert requests.
7046 [Ben Laurie (CHATS)]
7048 *) Add assertions to check for (supposedly impossible) buffer
7050 [Ben Laurie (CHATS)]
7052 *) Don't cache truncated DNS entries in the local cache (this could
7053 potentially lead to a spoofing attack).
7054 [Ben Laurie (CHATS)]
7056 *) Fix various buffers to be large enough for hex/decimal
7057 representations in a platform independent manner.
7058 [Ben Laurie (CHATS)]
7060 *) Add CRYPTO_realloc_clean() to avoid information leakage when
7061 resizing buffers containing secrets, and use where appropriate.
7062 [Ben Laurie (CHATS)]
7064 *) Add BIO_indent() to avoid much slightly worrying code to do
7066 [Ben Laurie (CHATS)]
7068 *) Convert sprintf()/BIO_puts() to BIO_printf().
7069 [Ben Laurie (CHATS)]
7071 *) buffer_gets() could terminate with the buffer only half
7073 [Ben Laurie (CHATS)]
7075 *) Add assertions to prevent user-supplied crypto functions from
7076 overflowing internal buffers by having large block sizes, etc.
7077 [Ben Laurie (CHATS)]
7079 *) New OPENSSL_assert() macro (similar to assert(), but enabled
7081 [Ben Laurie (CHATS)]
7083 *) Eliminate unused copy of key in RC4.
7084 [Ben Laurie (CHATS)]
7086 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
7087 [Ben Laurie (CHATS)]
7089 *) Fix off-by-one error in EGD path.
7090 [Ben Laurie (CHATS)]
7092 *) If RANDFILE path is too long, ignore instead of truncating.
7093 [Ben Laurie (CHATS)]
7095 *) Eliminate unused and incorrectly sized X.509 structure
7097 [Ben Laurie (CHATS)]
7099 *) Eliminate unused and dangerous function knumber().
7100 [Ben Laurie (CHATS)]
7102 *) Eliminate unused and dangerous structure, KSSL_ERR.
7103 [Ben Laurie (CHATS)]
7105 *) Protect against overlong session ID context length in an encoded
7106 session object. Since these are local, this does not appear to be
7108 [Ben Laurie (CHATS)]
7110 *) Change from security patch (see 0.9.6e below) that did not affect
7111 the 0.9.6 release series:
7113 Remote buffer overflow in SSL3 protocol - an attacker could
7114 supply an oversized master key in Kerberos-enabled versions.
7116 [Ben Laurie (CHATS)]
7118 *) Change the SSL kerb5 codes to match RFC 2712.
7121 *) Make -nameopt work fully for req and add -reqopt switch.
7122 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
7124 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
7125 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
7127 *) Make sure tests can be performed even if the corresponding algorithms
7128 have been removed entirely. This was also the last step to make
7129 OpenSSL compilable with DJGPP under all reasonable conditions.
7130 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
7132 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
7133 to allow version independent disabling of normally unselected ciphers,
7134 which may be activated as a side-effect of selecting a single cipher.
7136 (E.g., cipher list string "RSA" enables ciphersuites that are left
7137 out of "ALL" because they do not provide symmetric encryption.
7138 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
7139 [Lutz Jaenicke, Bodo Moeller]
7141 *) Add appropriate support for separate platform-dependent build
7142 directories. The recommended way to make a platform-dependent
7143 build directory is the following (tested on Linux), maybe with
7146 # Place yourself outside of the OpenSSL source tree. In
7147 # this example, the environment variable OPENSSL_SOURCE
7148 # is assumed to contain the absolute OpenSSL source directory.
7149 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
7150 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
7151 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
7152 mkdir -p `dirname $F`
7153 ln -s $OPENSSL_SOURCE/$F $F
7156 To be absolutely sure not to disturb the source tree, a "make clean"
7157 is a good thing. If it isn't successful, don't worry about it,
7158 it probably means the source directory is very clean.
7161 *) Make sure any ENGINE control commands make local copies of string
7162 pointers passed to them whenever necessary. Otherwise it is possible
7163 the caller may have overwritten (or deallocated) the original string
7164 data when a later ENGINE operation tries to use the stored values.
7165 [Götz Babin-Ebell <babinebell@trustcenter.de>]
7167 *) Improve diagnostics in file reading and command-line digests.
7168 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
7170 *) Add AES modes CFB and OFB to the object database. Correct an
7171 error in AES-CFB decryption.
7174 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
7175 allows existing EVP_CIPHER_CTX structures to be reused after
7176 calling EVP_*Final(). This behaviour is used by encryption
7177 BIOs and some applications. This has the side effect that
7178 applications must explicitly clean up cipher contexts with
7179 EVP_CIPHER_CTX_cleanup() or they will leak memory.
7182 *) Check the values of dna and dnb in bn_mul_recursive before calling
7183 bn_mul_comba (a non zero value means the a or b arrays do not contain
7184 n2 elements) and fallback to bn_mul_normal if either is not zero.
7187 *) Fix escaping of non-ASCII characters when using the -subj option
7188 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
7191 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
7192 form for "surname", serialNumber has no short form.
7193 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
7194 therefore remove "mail" short name for "internet 7".
7195 The OID for unique identifiers in X509 certificates is
7196 x500UniqueIdentifier, not uniqueIdentifier.
7197 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7200 *) Add an "init" command to the ENGINE config module and auto initialize
7201 ENGINEs. Without any "init" command the ENGINE will be initialized
7202 after all ctrl commands have been executed on it. If init=1 the
7203 ENGINE is initialized at that point (ctrls before that point are run
7204 on the uninitialized ENGINE and after on the initialized one). If
7205 init=0 then the ENGINE will not be initialized at all.
7208 *) Fix the 'app_verify_callback' interface so that the user-defined
7209 argument is actually passed to the callback: In the
7210 SSL_CTX_set_cert_verify_callback() prototype, the callback
7211 declaration has been changed from
7214 int (*cb)(X509_STORE_CTX *,void *);
7215 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7216 i=s->ctx->app_verify_callback(&ctx)
7217 has been changed into
7218 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7220 To update applications using SSL_CTX_set_cert_verify_callback(),
7221 a dummy argument can be added to their callback functions.
7222 [D. K. Smetters <smetters@parc.xerox.com>]
7224 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7225 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7227 *) Add and OPENSSL_LOAD_CONF define which will cause
7228 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7229 This allows older applications to transparently support certain
7230 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7231 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7232 load the config file and OPENSSL_add_all_algorithms_conf() which will
7233 always load it have also been added.
7236 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7237 Adjust NIDs and EVP layer.
7238 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7240 *) Config modules support in openssl utility.
7242 Most commands now load modules from the config file,
7243 though in a few (such as version) this isn't done
7244 because it couldn't be used for anything.
7246 In the case of ca and req the config file used is
7247 the same as the utility itself: that is the -config
7248 command line option can be used to specify an
7252 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7253 use "openssl_conf" if filename is NULL use default openssl config file.
7256 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7257 config section name. Add a new flag to tolerate a missing config file
7258 and move code to CONF_modules_load_file().
7261 *) Support for crypto accelerator cards from Accelerated Encryption
7262 Processing, www.aep.ie. (Use engine 'aep')
7263 The support was copied from 0.9.6c [engine] and adapted/corrected
7264 to work with the new engine framework.
7265 [AEP Inc. and Richard Levitte]
7267 *) Support for SureWare crypto accelerator cards from Baltimore
7268 Technologies. (Use engine 'sureware')
7269 The support was copied from 0.9.6c [engine] and adapted
7270 to work with the new engine framework.
7273 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7274 make the newer ENGINE framework commands for the CHIL engine work.
7275 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7277 *) Make it possible to produce shared libraries on ReliantUNIX.
7278 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7280 *) Add the configuration target debug-linux-ppro.
7281 Make 'openssl rsa' use the general key loading routines
7282 implemented in apps.c, and make those routines able to
7283 handle the key format FORMAT_NETSCAPE and the variant
7285 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7287 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7288 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7290 *) Add -keyform to rsautl, and document -engine.
7291 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7293 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7294 BIO_R_NO_SUCH_FILE error code rather than the generic
7295 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7298 *) Add new functions
7300 ERR_peek_last_error_line
7301 ERR_peek_last_error_line_data.
7302 These are similar to
7305 ERR_peek_error_line_data,
7306 but report on the latest error recorded rather than the first one
7307 still in the error queue.
7308 [Ben Laurie, Bodo Moeller]
7310 *) default_algorithms option in ENGINE config module. This allows things
7312 default_algorithms = ALL
7313 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7316 *) Preliminary ENGINE config module.
7319 *) New experimental application configuration code.
7322 *) Change the AES code to follow the same name structure as all other
7323 symmetric ciphers, and behave the same way. Move everything to
7324 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7325 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7327 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7328 [Ben Laurie and Theo de Raadt]
7330 *) Add option to output public keys in req command.
7331 [Massimiliano Pala madwolf@openca.org]
7333 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7334 (up to about 10% better than before for P-192 and P-224).
7337 *) New functions/macros
7339 SSL_CTX_set_msg_callback(ctx, cb)
7340 SSL_CTX_set_msg_callback_arg(ctx, arg)
7341 SSL_set_msg_callback(ssl, cb)
7342 SSL_set_msg_callback_arg(ssl, arg)
7344 to request calling a callback function
7346 void cb(int write_p, int version, int content_type,
7347 const void *buf, size_t len, SSL *ssl, void *arg)
7349 whenever a protocol message has been completely received
7350 (write_p == 0) or sent (write_p == 1). Here 'version' is the
7351 protocol version according to which the SSL library interprets
7352 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
7353 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
7354 the content type as defined in the SSL 3.0/TLS 1.0 protocol
7355 specification (change_cipher_spec(20), alert(21), handshake(22)).
7356 'buf' and 'len' point to the actual message, 'ssl' to the
7357 SSL object, and 'arg' is the application-defined value set by
7358 SSL[_CTX]_set_msg_callback_arg().
7360 'openssl s_client' and 'openssl s_server' have new '-msg' options
7361 to enable a callback that displays all protocol messages.
7364 *) Change the shared library support so shared libraries are built as
7365 soon as the corresponding static library is finished, and thereby get
7366 openssl and the test programs linked against the shared library.
7367 This still only happens when the keyword "shard" has been given to
7368 the configuration scripts.
7370 NOTE: shared library support is still an experimental thing, and
7371 backward binary compatibility is still not guaranteed.
7372 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
7374 *) Add support for Subject Information Access extension.
7375 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7377 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
7378 additional bytes when new memory had to be allocated, not just
7379 when reusing an existing buffer.
7382 *) New command line and configuration option 'utf8' for the req command.
7383 This allows field values to be specified as UTF8 strings.
7386 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
7387 runs for the former and machine-readable output for the latter.
7390 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
7391 of the e-mail address in the DN (i.e., it will go into a certificate
7392 extension only). The new configuration file option 'email_in_dn = no'
7393 has the same effect.
7394 [Massimiliano Pala madwolf@openca.org]
7396 *) Change all functions with names starting with des_ to be starting
7397 with DES_ instead. Add wrappers that are compatible with libdes,
7398 but are named _ossl_old_des_*. Finally, add macros that map the
7399 des_* symbols to the corresponding _ossl_old_des_* if libdes
7400 compatibility is desired. If OpenSSL 0.9.6c compatibility is
7401 desired, the des_* symbols will be mapped to DES_*, with one
7404 Since we provide two compatibility mappings, the user needs to
7405 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
7406 compatibility is desired. The default (i.e., when that macro
7407 isn't defined) is OpenSSL 0.9.6c compatibility.
7409 There are also macros that enable and disable the support of old
7410 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
7411 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
7412 are defined, the default will apply: to support the old des routines.
7414 In either case, one must include openssl/des.h to get the correct
7415 definitions. Do not try to just include openssl/des_old.h, that
7418 NOTE: This is a major break of an old API into a new one. Software
7419 authors are encouraged to switch to the DES_ style functions. Some
7420 time in the future, des_old.h and the libdes compatibility functions
7421 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
7422 default), and then completely removed.
7425 *) Test for certificates which contain unsupported critical extensions.
7426 If such a certificate is found during a verify operation it is
7427 rejected by default: this behaviour can be overridden by either
7428 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
7429 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
7430 X509_supported_extension() has also been added which returns 1 if a
7431 particular extension is supported.
7434 *) Modify the behaviour of EVP cipher functions in similar way to digests
7435 to retain compatibility with existing code.
7438 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
7439 compatibility with existing code. In particular the 'ctx' parameter does
7440 not have to be to be initialized before the call to EVP_DigestInit() and
7441 it is tidied up after a call to EVP_DigestFinal(). New function
7442 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
7443 EVP_MD_CTX_copy() changed to not require the destination to be
7444 initialized valid and new function EVP_MD_CTX_copy_ex() added which
7445 requires the destination to be valid.
7447 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
7448 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
7451 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
7452 so that complete 'Handshake' protocol structures are kept in memory
7453 instead of overwriting 'msg_type' and 'length' with 'body' data.
7456 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
7457 [Massimo Santin via Richard Levitte]
7459 *) Major restructuring to the underlying ENGINE code. This includes
7460 reduction of linker bloat, separation of pure "ENGINE" manipulation
7461 (initialisation, etc) from functionality dealing with implementations
7462 of specific crypto interfaces. This change also introduces integrated
7463 support for symmetric ciphers and digest implementations - so ENGINEs
7464 can now accelerate these by providing EVP_CIPHER and EVP_MD
7465 implementations of their own. This is detailed in crypto/engine/README
7466 as it couldn't be adequately described here. However, there are a few
7467 API changes worth noting - some RSA, DSA, DH, and RAND functions that
7468 were changed in the original introduction of ENGINE code have now
7469 reverted back - the hooking from this code to ENGINE is now a good
7470 deal more passive and at run-time, operations deal directly with
7471 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
7472 dereferencing through an ENGINE pointer any more. Also, the ENGINE
7473 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
7474 they were not being used by the framework as there is no concept of a
7475 BIGNUM_METHOD and they could not be generalised to the new
7476 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
7477 ENGINE_cpy() has been removed as it cannot be consistently defined in
7481 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
7484 *) Change mkdef.pl to sort symbols that get the same entry number,
7485 and make sure the automatically generated functions ERR_load_*
7486 become part of libeay.num as well.
7489 *) New function SSL_renegotiate_pending(). This returns true once
7490 renegotiation has been requested (either SSL_renegotiate() call
7491 or HelloRequest/ClientHello received from the peer) and becomes
7492 false once a handshake has been completed.
7493 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
7494 sends a HelloRequest, but does not ensure that a handshake takes
7495 place. SSL_renegotiate_pending() is useful for checking if the
7496 client has followed the request.)
7499 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7500 By default, clients may request session resumption even during
7501 renegotiation (if session ID contexts permit); with this option,
7502 session resumption is possible only in the first handshake.
7504 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7505 more bits available for options that should not be part of
7506 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7509 *) Add some demos for certificate and certificate request creation.
7512 *) Make maximum certificate chain size accepted from the peer application
7513 settable (SSL*_get/set_max_cert_list()), as proposed by
7514 "Douglas E. Engert" <deengert@anl.gov>.
7517 *) Add support for shared libraries for Unixware-7
7518 (Boyd Lynn Gerber <gerberb@zenez.com>).
7521 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7522 be done prior to destruction. Use this to unload error strings from
7523 ENGINEs that load their own error strings. NB: This adds two new API
7524 functions to "get" and "set" this destroy handler in an ENGINE.
7527 *) Alter all existing ENGINE implementations (except "openssl" and
7528 "openbsd") to dynamically instantiate their own error strings. This
7529 makes them more flexible to be built both as statically-linked ENGINEs
7530 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
7531 Also, add stub code to each that makes building them as self-contained
7532 shared-libraries easier (see README.ENGINE).
7535 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
7536 implementations into applications that are completely implemented in
7537 self-contained shared-libraries. The "dynamic" ENGINE exposes control
7538 commands that can be used to configure what shared-library to load and
7539 to control aspects of the way it is handled. Also, made an update to
7540 the README.ENGINE file that brings its information up-to-date and
7541 provides some information and instructions on the "dynamic" ENGINE
7542 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
7545 *) Make it possible to unload ranges of ERR strings with a new
7546 "ERR_unload_strings" function.
7549 *) Add a copy() function to EVP_MD.
7552 *) Make EVP_MD routines take a context pointer instead of just the
7553 md_data void pointer.
7556 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
7557 that the digest can only process a single chunk of data
7558 (typically because it is provided by a piece of
7559 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
7560 is only going to provide a single chunk of data, and hence the
7561 framework needn't accumulate the data for oneshot drivers.
7564 *) As with "ERR", make it possible to replace the underlying "ex_data"
7565 functions. This change also alters the storage and management of global
7566 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
7567 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
7568 index counters. The API functions that use this state have been changed
7569 to take a "class_index" rather than pointers to the class's local STACK
7570 and counter, and there is now an API function to dynamically create new
7571 classes. This centralisation allows us to (a) plug a lot of the
7572 thread-safety problems that existed, and (b) makes it possible to clean
7573 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
7574 such data would previously have always leaked in application code and
7575 workarounds were in place to make the memory debugging turn a blind eye
7576 to it. Application code that doesn't use this new function will still
7577 leak as before, but their memory debugging output will announce it now
7578 rather than letting it slide.
7580 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
7581 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
7582 has a return value to indicate success or failure.
7585 *) Make it possible to replace the underlying "ERR" functions such that the
7586 global state (2 LHASH tables and 2 locks) is only used by the "default"
7587 implementation. This change also adds two functions to "get" and "set"
7588 the implementation prior to it being automatically set the first time
7589 any other ERR function takes place. Ie. an application can call "get",
7590 pass the return value to a module it has just loaded, and that module
7591 can call its own "set" function using that value. This means the
7592 module's "ERR" operations will use (and modify) the error state in the
7593 application and not in its own statically linked copy of OpenSSL code.
7596 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
7597 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
7598 the operation, and provides a more encapsulated way for external code
7599 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
7600 to use these functions rather than manually incrementing the counts.
7602 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
7605 *) Add EVP test program.
7608 *) Add symmetric cipher support to ENGINE. Expect the API to change!
7611 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
7612 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
7613 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
7614 These allow a CRL to be built without having to access X509_CRL fields
7615 directly. Modify 'ca' application to use new functions.
7618 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
7619 bug workarounds. Rollback attack detection is a security feature.
7620 The problem will only arise on OpenSSL servers when TLSv1 is not
7621 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7622 Software authors not wanting to support TLSv1 will have special reasons
7623 for their choice and can explicitly enable this option.
7624 [Bodo Moeller, Lutz Jaenicke]
7626 *) Rationalise EVP so it can be extended: don't include a union of
7627 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7628 (similar to those existing for EVP_CIPHER_CTX).
7633 EVP_MD_CTX_init(&md); /* new function call */
7634 EVP_DigestInit(&md, EVP_sha1());
7635 EVP_DigestUpdate(&md, in, len);
7636 EVP_DigestFinal(&md, out, NULL);
7637 EVP_MD_CTX_cleanup(&md); /* new function call */
7641 *) Make DES key schedule conform to the usual scheme, as well as
7642 correcting its structure. This means that calls to DES functions
7643 now have to pass a pointer to a des_key_schedule instead of a
7644 plain des_key_schedule (which was actually always a pointer
7647 des_key_schedule ks;
7649 des_set_key_checked(..., &ks);
7650 des_ncbc_encrypt(..., &ks, ...);
7652 (Note that a later change renames 'des_...' into 'DES_...'.)
7655 *) Initial reduction of linker bloat: the use of some functions, such as
7656 PEM causes large amounts of unused functions to be linked in due to
7657 poor organisation. For example pem_all.c contains every PEM function
7658 which has a knock on effect of linking in large amounts of (unused)
7659 ASN1 code. Grouping together similar functions and splitting unrelated
7660 functions prevents this.
7663 *) Cleanup of EVP macros.
7666 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7667 correct _ecb suffix.
7670 *) Add initial OCSP responder support to ocsp application. The
7671 revocation information is handled using the text based index
7672 use by the ca application. The responder can either handle
7673 requests generated internally, supplied in files (for example
7674 via a CGI script) or using an internal minimal server.
7677 *) Add configuration choices to get zlib compression for TLS.
7680 *) Changes to Kerberos SSL for RFC 2712 compliance:
7681 1. Implemented real KerberosWrapper, instead of just using
7682 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7683 2. Implemented optional authenticator field of KerberosWrapper.
7685 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7686 and authenticator structs; see crypto/krb5/.
7688 Generalized Kerberos calls to support multiple Kerberos libraries.
7689 [Vern Staats <staatsvr@asc.hpc.mil>,
7690 Jeffrey Altman <jaltman@columbia.edu>
7691 via Richard Levitte]
7693 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7694 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7695 values for each of the key sizes rather than having just
7696 parameters (and 'speed' generating keys each time).
7699 *) Speed up EVP routines.
7702 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7703 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7704 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7705 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7707 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7708 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7709 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7712 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7714 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7717 *) Added the OS2-EMX target.
7718 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7720 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7721 to support NCONF routines in extension code. New function CONF_set_nconf()
7722 to allow functions which take an NCONF to also handle the old LHASH
7723 structure: this means that the old CONF compatible routines can be
7724 retained (in particular wrt extensions) without having to duplicate the
7725 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7728 *) Enhance the general user interface with mechanisms for inner control
7729 and with possibilities to have yes/no kind of prompts.
7732 *) Change all calls to low level digest routines in the library and
7733 applications to use EVP. Add missing calls to HMAC_cleanup() and
7734 don't assume HMAC_CTX can be copied using memcpy().
7735 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7737 *) Add the possibility to control engines through control names but with
7738 arbitrary arguments instead of just a string.
7739 Change the key loaders to take a UI_METHOD instead of a callback
7740 function pointer. NOTE: this breaks binary compatibility with earlier
7741 versions of OpenSSL [engine].
7742 Adapt the nCipher code for these new conditions and add a card insertion
7746 *) Enhance the general user interface with mechanisms to better support
7747 dialog box interfaces, application-defined prompts, the possibility
7748 to use defaults (for example default passwords from somewhere else)
7749 and interrupts/cancellations.
7752 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7753 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7756 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7757 tidy up some unnecessarily weird code in 'sk_new()').
7758 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
7760 *) Change the key loading routines for ENGINEs to use the same kind
7761 callback (pem_password_cb) as all other routines that need this
7765 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
7766 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
7767 than this minimum value is recommended.
7770 *) New random seeder for OpenVMS, using the system process statistics
7771 that are easily reachable.
7774 *) Windows apparently can't transparently handle global
7775 variables defined in DLLs. Initialisations such as:
7777 const ASN1_ITEM *it = &ASN1_INTEGER_it;
7779 won't compile. This is used by the any applications that need to
7780 declare their own ASN1 modules. This was fixed by adding the option
7781 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
7782 needed for static libraries under Win32.
7785 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
7786 setting of purpose and trust fields. New X509_STORE trust and
7787 purpose functions and tidy up setting in other SSL functions.
7790 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
7791 structure. These are inherited by X509_STORE_CTX when it is
7792 initialised. This allows various defaults to be set in the
7793 X509_STORE structure (such as flags for CRL checking and custom
7794 purpose or trust settings) for functions which only use X509_STORE_CTX
7795 internally such as S/MIME.
7797 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
7798 trust settings if they are not set in X509_STORE. This allows X509_STORE
7799 purposes and trust (in S/MIME for example) to override any set by default.
7801 Add command line options for CRL checking to smime, s_client and s_server
7805 *) Initial CRL based revocation checking. If the CRL checking flag(s)
7806 are set then the CRL is looked up in the X509_STORE structure and
7807 its validity and signature checked, then if the certificate is found
7808 in the CRL the verify fails with a revoked error.
7810 Various new CRL related callbacks added to X509_STORE_CTX structure.
7812 Command line options added to 'verify' application to support this.
7814 This needs some additional work, such as being able to handle multiple
7815 CRLs with different times, extension based lookup (rather than just
7816 by subject name) and ultimately more complete V2 CRL extension
7820 *) Add a general user interface API (crypto/ui/). This is designed
7821 to replace things like des_read_password and friends (backward
7822 compatibility functions using this new API are provided).
7823 The purpose is to remove prompting functions from the DES code
7824 section as well as provide for prompting through dialog boxes in
7825 a window system and the like.
7828 *) Add "ex_data" support to ENGINE so implementations can add state at a
7829 per-structure level rather than having to store it globally.
7832 *) Make it possible for ENGINE structures to be copied when retrieved by
7833 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
7834 This causes the "original" ENGINE structure to act like a template,
7835 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7836 operational state can be localised to each ENGINE structure, despite the
7837 fact they all share the same "methods". New ENGINE structures returned in
7838 this case have no functional references and the return value is the single
7839 structural reference. This matches the single structural reference returned
7840 by ENGINE_by_id() normally, when it is incremented on the pre-existing
7844 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
7845 needs to match any other type at all we need to manually clear the
7849 *) Changes to the "openssl engine" utility to include;
7850 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
7851 about an ENGINE's available control commands.
7852 - executing control commands from command line arguments using the
7853 '-pre' and '-post' switches. '-post' is only used if '-t' is
7854 specified and the ENGINE is successfully initialised. The syntax for
7855 the individual commands are colon-separated, for example;
7856 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
7859 *) New dynamic control command support for ENGINEs. ENGINEs can now
7860 declare their own commands (numbers), names (strings), descriptions,
7861 and input types for run-time discovery by calling applications. A
7862 subset of these commands are implicitly classed as "executable"
7863 depending on their input type, and only these can be invoked through
7864 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
7865 can be based on user input, config files, etc). The distinction is
7866 that "executable" commands cannot return anything other than a boolean
7867 result and can only support numeric or string input, whereas some
7868 discoverable commands may only be for direct use through
7869 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
7870 pointers, or other custom uses. The "executable" commands are to
7871 support parameterisations of ENGINE behaviour that can be
7872 unambiguously defined by ENGINEs and used consistently across any
7873 OpenSSL-based application. Commands have been added to all the
7874 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
7875 control over shared-library paths without source code alterations.
7878 *) Changed all ENGINE implementations to dynamically allocate their
7879 ENGINEs rather than declaring them statically. Apart from this being
7880 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
7881 this also allows the implementations to compile without using the
7882 internal engine_int.h header.
7885 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
7886 'const' value. Any code that should be able to modify a RAND_METHOD
7887 should already have non-const pointers to it (ie. they should only
7888 modify their own ones).
7891 *) Made a variety of little tweaks to the ENGINE code.
7892 - "atalla" and "ubsec" string definitions were moved from header files
7893 to C code. "nuron" string definitions were placed in variables
7894 rather than hard-coded - allowing parameterisation of these values
7895 later on via ctrl() commands.
7896 - Removed unused "#if 0"'d code.
7897 - Fixed engine list iteration code so it uses ENGINE_free() to release
7898 structural references.
7899 - Constified the RAND_METHOD element of ENGINE structures.
7900 - Constified various get/set functions as appropriate and added
7901 missing functions (including a catch-all ENGINE_cpy that duplicates
7902 all ENGINE values onto a new ENGINE except reference counts/state).
7903 - Removed NULL parameter checks in get/set functions. Setting a method
7904 or function to NULL is a way of cancelling out a previously set
7905 value. Passing a NULL ENGINE parameter is just plain stupid anyway
7906 and doesn't justify the extra error symbols and code.
7907 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
7908 flags from engine_int.h to engine.h.
7909 - Changed prototypes for ENGINE handler functions (init(), finish(),
7910 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
7913 *) Implement binary inversion algorithm for BN_mod_inverse in addition
7914 to the algorithm using long division. The binary algorithm can be
7915 used only if the modulus is odd. On 32-bit systems, it is faster
7916 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
7917 roughly 5-15% for 256-bit moduli), so we use it only for moduli
7918 up to 450 bits. In 64-bit environments, the binary algorithm
7919 appears to be advantageous for much longer moduli; here we use it
7920 for moduli up to 2048 bits.
7923 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
7924 could not support the combine flag in choice fields.
7927 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
7928 extensions from a certificate request to the certificate.
7931 *) Allow multiple 'certopt' and 'nameopt' options to be separated
7932 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
7933 file: this allows the display of the certificate about to be
7934 signed to be customised, to allow certain fields to be included
7935 or excluded and extension details. The old system didn't display
7936 multicharacter strings properly, omitted fields not in the policy
7937 and couldn't display additional details such as extensions.
7940 *) Function EC_POINTs_mul for multiple scalar multiplication
7941 of an arbitrary number of elliptic curve points
7942 \sum scalars[i]*points[i],
7943 optionally including the generator defined for the EC_GROUP:
7944 scalar*generator + \sum scalars[i]*points[i].
7946 EC_POINT_mul is a simple wrapper function for the typical case
7947 that the point list has just one item (besides the optional
7951 *) First EC_METHODs for curves over GF(p):
7953 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
7954 operations and provides various method functions that can also
7955 operate with faster implementations of modular arithmetic.
7957 EC_GFp_mont_method() reuses most functions that are part of
7958 EC_GFp_simple_method, but uses Montgomery arithmetic.
7960 [Bodo Moeller; point addition and point doubling
7961 implementation directly derived from source code provided by
7962 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
7964 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
7965 crypto/ec/ec_lib.c):
7967 Curves are EC_GROUP objects (with an optional group generator)
7968 based on EC_METHODs that are built into the library.
7970 Points are EC_POINT objects based on EC_GROUP objects.
7972 Most of the framework would be able to handle curves over arbitrary
7973 finite fields, but as there are no obvious types for fields other
7974 than GF(p), some functions are limited to that for now.
7977 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
7978 that the file contains a complete HTTP response.
7981 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
7982 change the def and num file printf format specifier from "%-40sXXX"
7983 to "%-39s XXX". The latter will always guarantee a space after the
7984 field while the former will cause them to run together if the field
7985 is 40 of more characters long.
7988 *) Constify the cipher and digest 'method' functions and structures
7989 and modify related functions to take constant EVP_MD and EVP_CIPHER
7993 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
7994 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
7997 *) Modify EVP_Digest*() routines so they now return values. Although the
7998 internal software routines can never fail additional hardware versions
8002 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
8004 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
8005 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
8008 ERR_R_NESTED_ASN1_ERROR
8010 ERR_R_MISSING_ASN1_EOS
8011 were 4 .. 9, conflicting with
8012 ERR_LIB_RSA (= ERR_R_RSA_LIB)
8014 ERR_LIB_PEM (= ERR_R_PEM_LIB).
8015 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
8017 Add new error code 'ERR_R_INTERNAL_ERROR'.
8020 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
8024 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
8025 sets the subject name for a new request or supersedes the
8026 subject name in a given request. Formats that can be parsed are
8027 'CN=Some Name, OU=myOU, C=IT'
8029 'CN=Some Name/OU=myOU/C=IT'.
8031 Add options '-batch' and '-verbose' to 'openssl req'.
8032 [Massimiliano Pala <madwolf@hackmasters.net>]
8034 *) Introduce the possibility to access global variables through
8035 functions on platform were that's the best way to handle exporting
8036 global variables in shared libraries. To enable this functionality,
8037 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
8038 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
8039 is normally done by Configure or something similar).
8041 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
8042 in the source file (foo.c) like this:
8044 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
8045 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
8047 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
8048 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
8050 OPENSSL_DECLARE_GLOBAL(int,foo);
8051 #define foo OPENSSL_GLOBAL_REF(foo)
8052 OPENSSL_DECLARE_GLOBAL(double,bar);
8053 #define bar OPENSSL_GLOBAL_REF(bar)
8055 The #defines are very important, and therefore so is including the
8056 header file everywhere where the defined globals are used.
8058 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
8059 of ASN.1 items, but that structure is a bit different.
8061 The largest change is in util/mkdef.pl which has been enhanced with
8062 better and easier to understand logic to choose which symbols should
8063 go into the Windows .def files as well as a number of fixes and code
8064 cleanup (among others, algorithm keywords are now sorted
8065 lexicographically to avoid constant rewrites).
8068 *) In BN_div() keep a copy of the sign of 'num' before writing the
8069 result to 'rm' because if rm==num the value will be overwritten
8070 and produce the wrong result if 'num' is negative: this caused
8071 problems with BN_mod() and BN_nnmod().
8074 *) Function OCSP_request_verify(). This checks the signature on an
8075 OCSP request and verifies the signer certificate. The signer
8076 certificate is just checked for a generic purpose and OCSP request
8080 *) Add OCSP_check_validity() function to check the validity of OCSP
8081 responses. OCSP responses are prepared in real time and may only
8082 be a few seconds old. Simply checking that the current time lies
8083 between thisUpdate and nextUpdate max reject otherwise valid responses
8084 caused by either OCSP responder or client clock inaccuracy. Instead
8085 we allow thisUpdate and nextUpdate to fall within a certain period of
8086 the current time. The age of the response can also optionally be
8087 checked. Two new options -validity_period and -status_age added to
8091 *) If signature or public key algorithm is unrecognized print out its
8092 OID rather that just UNKNOWN.
8095 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
8096 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
8097 ID to be generated from the issuer certificate alone which can then be
8098 passed to OCSP_id_issuer_cmp().
8101 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
8102 ASN1 modules to export functions returning ASN1_ITEM pointers
8103 instead of the ASN1_ITEM structures themselves. This adds several
8104 new macros which allow the underlying ASN1 function/structure to
8105 be accessed transparently. As a result code should not use ASN1_ITEM
8106 references directly (such as &X509_it) but instead use the relevant
8107 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
8108 use of the new ASN1 code on platforms where exporting structures
8109 is problematical (for example in shared libraries) but exporting
8110 functions returning pointers to structures is not.
8113 *) Add support for overriding the generation of SSL/TLS session IDs.
8114 These callbacks can be registered either in an SSL_CTX or per SSL.
8115 The purpose of this is to allow applications to control, if they wish,
8116 the arbitrary values chosen for use as session IDs, particularly as it
8117 can be useful for session caching in multiple-server environments. A
8118 command-line switch for testing this (and any client code that wishes
8119 to use such a feature) has been added to "s_server".
8120 [Geoff Thorpe, Lutz Jaenicke]
8122 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
8123 of the form '#if defined(...) || defined(...) || ...' and
8124 '#if !defined(...) && !defined(...) && ...'. This also avoids
8125 the growing number of special cases it was previously handling.
8128 *) Make all configuration macros available for application by making
8129 sure they are available in opensslconf.h, by giving them names starting
8130 with "OPENSSL_" to avoid conflicts with other packages and by making
8131 sure e_os2.h will cover all platform-specific cases together with
8133 Additionally, it is now possible to define configuration/platform-
8134 specific names (called "system identities"). In the C code, these
8135 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
8136 macro with the name beginning with "OPENSSL_SYS_", which is determined
8137 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
8141 *) New option -set_serial to 'req' and 'x509' this allows the serial
8142 number to use to be specified on the command line. Previously self
8143 signed certificates were hard coded with serial number 0 and the
8144 CA options of 'x509' had to use a serial number in a file which was
8148 *) New options to 'ca' utility to support V2 CRL entry extensions.
8149 Currently CRL reason, invalidity date and hold instruction are
8150 supported. Add new CRL extensions to V3 code and some new objects.
8153 *) New function EVP_CIPHER_CTX_set_padding() this is used to
8154 disable standard block padding (aka PKCS#5 padding) in the EVP
8155 API, which was previously mandatory. This means that the data is
8156 not padded in any way and so the total length much be a multiple
8157 of the block size, otherwise an error occurs.
8160 *) Initial (incomplete) OCSP SSL support.
8163 *) New function OCSP_parse_url(). This splits up a URL into its host,
8164 port and path components: primarily to parse OCSP URLs. New -url
8165 option to ocsp utility.
8168 *) New nonce behavior. The return value of OCSP_check_nonce() now
8169 reflects the various checks performed. Applications can decide
8170 whether to tolerate certain situations such as an absent nonce
8171 in a response when one was present in a request: the ocsp application
8172 just prints out a warning. New function OCSP_add1_basic_nonce()
8173 this is to allow responders to include a nonce in a response even if
8174 the request is nonce-less.
8177 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
8178 skipped when using openssl x509 multiple times on a single input file,
8179 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
8182 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
8183 set string type: to handle setting ASN1_TIME structures. Fix ca
8184 utility to correctly initialize revocation date of CRLs.
8187 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
8188 the clients preferred ciphersuites and rather use its own preferences.
8189 Should help to work around M$ SGC (Server Gated Cryptography) bug in
8190 Internet Explorer by ensuring unchanged hash method during stepup.
8191 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
8194 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
8195 to aes and add a new 'exist' option to print out symbols that don't
8199 *) Additional options to ocsp utility to allow flags to be set and
8200 additional certificates supplied.
8203 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8204 OCSP client a number of certificate to only verify the response
8208 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8209 handle the new API. Currently only ECB, CBC modes supported. Add new
8212 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8213 Encryption Standard (AES) Ciphersuites for Transport Layer
8214 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8215 not enabled by default and were not part of the "ALL" ciphersuite
8216 alias because they were not yet official; they could be
8217 explicitly requested by specifying the "AESdraft" ciphersuite
8218 group alias. In the final release of OpenSSL 0.9.7, the group
8219 alias is called "AES" and is part of "ALL".)
8220 [Ben Laurie, Steve Henson, Bodo Moeller]
8222 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8223 request to response.
8226 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8227 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8228 extract information from a certificate request. OCSP_response_create()
8229 creates a response and optionally adds a basic response structure.
8230 OCSP_basic_add1_status() adds a complete single response to a basic
8231 response and returns the OCSP_SINGLERESP structure just added (to allow
8232 extensions to be included for example). OCSP_basic_add1_cert() adds a
8233 certificate to a basic response and OCSP_basic_sign() signs a basic
8234 response with various flags. New helper functions ASN1_TIME_check()
8235 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8236 (converts ASN1_TIME to GeneralizedTime).
8239 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8240 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8241 structure from a certificate. X509_pubkey_digest() digests the public_key
8242 contents: this is used in various key identifiers.
8245 *) Make sk_sort() tolerate a NULL argument.
8246 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8248 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8249 passed by the function are trusted implicitly. If any of them signed the
8250 response then it is assumed to be valid and is not verified.
8253 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8254 to data. This was previously part of the PKCS7 ASN1 code. This
8255 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8256 [Steve Henson, reported by Kenneth R. Robinette
8257 <support@securenetterm.com>]
8259 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8260 routines: without these tracing memory leaks is very painful.
8261 Fix leaks in PKCS12 and PKCS7 routines.
8264 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8265 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8266 effectively meant GeneralizedTime would never be used. Now it
8267 is initialised to -1 but X509_time_adj() now has to check the value
8268 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8269 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8270 [Steve Henson, reported by Kenneth R. Robinette
8271 <support@securenetterm.com>]
8273 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8274 result in a zero length in the ASN1_INTEGER structure which was
8275 not consistent with the structure when d2i_ASN1_INTEGER() was used
8276 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8277 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8278 where it did not print out a minus for negative ASN1_INTEGER.
8281 *) Add summary printout to ocsp utility. The various functions which
8282 convert status values to strings have been renamed to:
8283 OCSP_response_status_str(), OCSP_cert_status_str() and
8284 OCSP_crl_reason_str() and are no longer static. New options
8285 to verify nonce values and to disable verification. OCSP response
8286 printout format cleaned up.
8289 *) Add additional OCSP certificate checks. These are those specified
8290 in RFC2560. This consists of two separate checks: the CA of the
8291 certificate being checked must either be the OCSP signer certificate
8292 or the issuer of the OCSP signer certificate. In the latter case the
8293 OCSP signer certificate must contain the OCSP signing extended key
8294 usage. This check is performed by attempting to match the OCSP
8295 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8296 in the OCSP_CERTID structures of the response.
8299 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8300 and related routines. This uses the standard OpenSSL certificate
8301 verify routines to perform initial checks (just CA validity) and
8302 to obtain the certificate chain. Then additional checks will be
8303 performed on the chain. Currently the root CA is checked to see
8304 if it is explicitly trusted for OCSP signing. This is used to set
8305 a root CA as a global signing root: that is any certificate that
8306 chains to that CA is an acceptable OCSP signing certificate.
8309 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8310 extensions from a separate configuration file.
8311 As when reading extensions from the main configuration file,
8312 the '-extensions ...' option may be used for specifying the
8314 [Massimiliano Pala <madwolf@comune.modena.it>]
8316 *) New OCSP utility. Allows OCSP requests to be generated or
8317 read. The request can be sent to a responder and the output
8318 parsed, outputed or printed in text form. Not complete yet:
8319 still needs to check the OCSP response validity.
8322 *) New subcommands for 'openssl ca':
8323 'openssl ca -status <serial>' prints the status of the cert with
8324 the given serial number (according to the index file).
8325 'openssl ca -updatedb' updates the expiry status of certificates
8327 [Massimiliano Pala <madwolf@comune.modena.it>]
8329 *) New '-newreq-nodes' command option to CA.pl. This is like
8330 '-newreq', but calls 'openssl req' with the '-nodes' option
8331 so that the resulting key is not encrypted.
8332 [Damien Miller <djm@mindrot.org>]
8334 *) New configuration for the GNU Hurd.
8335 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8337 *) Initial code to implement OCSP basic response verify. This
8338 is currently incomplete. Currently just finds the signer's
8339 certificate and verifies the signature on the response.
8342 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
8343 value of OPENSSLDIR. This is available via the new '-d' option
8344 to 'openssl version', and is also included in 'openssl version -a'.
8347 *) Allowing defining memory allocation callbacks that will be given
8348 file name and line number information in additional arguments
8349 (a const char* and an int). The basic functionality remains, as
8350 well as the original possibility to just replace malloc(),
8351 realloc() and free() by functions that do not know about these
8352 additional arguments. To register and find out the current
8353 settings for extended allocation functions, the following
8354 functions are provided:
8356 CRYPTO_set_mem_ex_functions
8357 CRYPTO_set_locked_mem_ex_functions
8358 CRYPTO_get_mem_ex_functions
8359 CRYPTO_get_locked_mem_ex_functions
8361 These work the same way as CRYPTO_set_mem_functions and friends.
8362 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
8363 extended allocation function is enabled.
8364 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
8365 a conventional allocation function is enabled.
8366 [Richard Levitte, Bodo Moeller]
8368 *) Finish off removing the remaining LHASH function pointer casts.
8369 There should no longer be any prototype-casting required when using
8370 the LHASH abstraction, and any casts that remain are "bugs". See
8371 the callback types and macros at the head of lhash.h for details
8372 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
8375 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
8376 If /dev/[u]random devices are not available or do not return enough
8377 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
8379 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
8380 /etc/entropy will be queried once each in this sequence, querying stops
8381 when enough entropy was collected without querying more sockets.
8384 *) Change the Unix RAND_poll() variant to be able to poll several
8385 random devices, as specified by DEVRANDOM, until a sufficient amount
8386 of data has been collected. We spend at most 10 ms on each file
8387 (select timeout) and read in non-blocking mode. DEVRANDOM now
8388 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
8389 (previously it was just the string "/dev/urandom"), so on typical
8390 platforms the 10 ms delay will never occur.
8391 Also separate out the Unix variant to its own file, rand_unix.c.
8392 For VMS, there's a currently-empty rand_vms.c.
8395 *) Move OCSP client related routines to ocsp_cl.c. These
8396 provide utility functions which an application needing
8397 to issue a request to an OCSP responder and analyse the
8398 response will typically need: as opposed to those which an
8399 OCSP responder itself would need which will be added later.
8401 OCSP_request_sign() signs an OCSP request with an API similar
8402 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
8403 response. OCSP_response_get1_basic() extracts basic response
8404 from response. OCSP_resp_find_status(): finds and extracts status
8405 information from an OCSP_CERTID structure (which will be created
8406 when the request structure is built). These are built from lower
8407 level functions which work on OCSP_SINGLERESP structures but
8408 won't normally be used unless the application wishes to examine
8409 extensions in the OCSP response for example.
8411 Replace nonce routines with a pair of functions.
8412 OCSP_request_add1_nonce() adds a nonce value and optionally
8413 generates a random value. OCSP_check_nonce() checks the
8414 validity of the nonce in an OCSP response.
8417 *) Change function OCSP_request_add() to OCSP_request_add0_id().
8418 This doesn't copy the supplied OCSP_CERTID and avoids the
8419 need to free up the newly created id. Change return type
8420 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
8421 This can then be used to add extensions to the request.
8422 Deleted OCSP_request_new(), since most of its functionality
8423 is now in OCSP_REQUEST_new() (and the case insensitive name
8424 clash) apart from the ability to set the request name which
8425 will be added elsewhere.
8428 *) Update OCSP API. Remove obsolete extensions argument from
8429 various functions. Extensions are now handled using the new
8430 OCSP extension code. New simple OCSP HTTP function which
8431 can be used to send requests and parse the response.
8434 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
8435 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
8436 uses the special reorder version of SET OF to sort the attributes
8437 and reorder them to match the encoded order. This resolves a long
8438 standing problem: a verify on a PKCS7 structure just after signing
8439 it used to fail because the attribute order did not match the
8440 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
8441 it uses the received order. This is necessary to tolerate some broken
8442 software that does not order SET OF. This is handled by encoding
8443 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
8444 to produce the required SET OF.
8447 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
8448 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8449 files to get correct declarations of the ASN.1 item variables.
8452 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
8453 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
8454 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
8455 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
8456 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
8457 ASN1_ITEM and no wrapper functions.
8460 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
8461 replace the old function pointer based I/O routines. Change most of
8462 the *_d2i_bio() and *_d2i_fp() functions to use these.
8465 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
8466 lines, recognize more "algorithms" that can be deselected, and make
8467 it complain about algorithm deselection that isn't recognised.
8470 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
8471 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
8472 to use new functions. Add NO_ASN1_OLD which can be set to remove
8473 some old style ASN1 functions: this can be used to determine if old
8474 code will still work when these eventually go away.
8477 *) New extension functions for OCSP structures, these follow the
8478 same conventions as certificates and CRLs.
8481 *) New function X509V3_add1_i2d(). This automatically encodes and
8482 adds an extension. Its behaviour can be customised with various
8483 flags to append, replace or delete. Various wrappers added for
8484 certificates and CRLs.
8487 *) Fix to avoid calling the underlying ASN1 print routine when
8488 an extension cannot be parsed. Correct a typo in the
8489 OCSP_SERVICELOC extension. Tidy up print OCSP format.
8492 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
8493 entries for variables.
8496 *) Add functionality to apps/openssl.c for detecting locking
8497 problems: As the program is single-threaded, all we have
8498 to do is register a locking callback using an array for
8499 storing which locks are currently held by the program.
8502 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8503 SSL_get_ex_data_X509_STORE_idx(), which is used in
8504 ssl_verify_cert_chain() and thus can be called at any time
8505 during TLS/SSL handshakes so that thread-safety is essential.
8506 Unfortunately, the ex_data design is not at all suited
8507 for multi-threaded use, so it probably should be abolished.
8510 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8511 [Broadcom, tweaked and integrated by Geoff Thorpe]
8513 *) Move common extension printing code to new function
8514 X509V3_print_extensions(). Reorganise OCSP print routines and
8515 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8518 *) New function X509_signature_print() to remove duplication in some
8522 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8523 set (this was treated exactly the same as SET OF previously). This
8524 is used to reorder the STACK representing the structure to match the
8525 encoding. This will be used to get round a problem where a PKCS7
8526 structure which was signed could not be verified because the STACK
8527 order did not reflect the encoded order.
8530 *) Reimplement the OCSP ASN1 module using the new code.
8533 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
8534 for its ASN1 operations. The old style function pointers still exist
8535 for now but they will eventually go away.
8538 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
8539 completely replaces the old ASN1 functionality with a table driven
8540 encoder and decoder which interprets an ASN1_ITEM structure describing
8541 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
8542 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
8543 has also been converted to the new form.
8546 *) Change BN_mod_exp_recp so that negative moduli are tolerated
8547 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
8548 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
8549 for negative moduli.
8552 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
8553 of not touching the result's sign bit.
8556 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
8560 *) Changed the LHASH code to use prototypes for callbacks, and created
8561 macros to declare and implement thin (optionally static) functions
8562 that provide type-safety and avoid function pointer casting for the
8563 type-specific callbacks.
8566 *) Added Kerberos Cipher Suites to be used with TLS, as written in
8568 [Veers Staats <staatsvr@asc.hpc.mil>,
8569 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
8571 *) Reformat the FAQ so the different questions and answers can be divided
8572 in sections depending on the subject.
8575 *) Have the zlib compression code load ZLIB.DLL dynamically under
8579 *) New function BN_mod_sqrt for computing square roots modulo a prime
8580 (using the probabilistic Tonelli-Shanks algorithm unless
8581 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
8582 be handled deterministically).
8583 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8585 *) Make BN_mod_inverse faster by explicitly handling small quotients
8586 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
8587 512 bits], about 30% for larger ones [1024 or 2048 bits].)
8590 *) New function BN_kronecker.
8593 *) Fix BN_gcd so that it works on negative inputs; the result is
8594 positive unless both parameters are zero.
8595 Previously something reasonably close to an infinite loop was
8596 possible because numbers could be growing instead of shrinking
8597 in the implementation of Euclid's algorithm.
8600 *) Fix BN_is_word() and BN_is_one() macros to take into account the
8601 sign of the number in question.
8603 Fix BN_is_word(a,w) to work correctly for w == 0.
8605 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
8606 because its test if the absolute value of 'a' equals 'w'.
8607 Note that BN_abs_is_word does *not* handle w == 0 reliably;
8608 it exists mostly for use in the implementations of BN_is_zero(),
8609 BN_is_one(), and BN_is_word().
8612 *) New function BN_swap.
8615 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
8616 the exponentiation functions are more likely to produce reasonable
8617 results on negative inputs.
8620 *) Change BN_mod_mul so that the result is always non-negative.
8621 Previously, it could be negative if one of the factors was negative;
8622 I don't think anyone really wanted that behaviour.
8625 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8626 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8627 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8628 and add new functions:
8637 BN_mod_lshift1_quick
8641 These functions always generate non-negative results.
8643 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8644 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8646 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8647 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8648 be reduced modulo m.
8649 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8652 The following entry accidentally appeared in the CHANGES file
8653 distributed with OpenSSL 0.9.7. The modifications described in
8654 it do *not* apply to OpenSSL 0.9.7.
8656 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8657 was actually never needed) and in BN_mul(). The removal in BN_mul()
8658 required a small change in bn_mul_part_recursive() and the addition
8659 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8660 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8661 bn_sub_words() and bn_add_words() except they take arrays with
8666 *) In 'openssl passwd', verify passwords read from the terminal
8667 unless the '-salt' option is used (which usually means that
8668 verification would just waste user's time since the resulting
8669 hash is going to be compared with some given password hash)
8670 or the new '-noverify' option is used.
8672 This is an incompatible change, but it does not affect
8673 non-interactive use of 'openssl passwd' (passwords on the command
8674 line, '-stdin' option, '-in ...' option) and thus should not
8678 *) Remove all references to RSAref, since there's no more need for it.
8681 *) Make DSO load along a path given through an environment variable
8682 (SHLIB_PATH) with shl_load().
8685 *) Constify the ENGINE code as a result of BIGNUM constification.
8686 Also constify the RSA code and most things related to it. In a
8687 few places, most notable in the depth of the ASN.1 code, ugly
8688 casts back to non-const were required (to be solved at a later
8692 *) Make it so the openssl application has all engines loaded by default.
8695 *) Constify the BIGNUM routines a little more.
8698 *) Add the following functions:
8700 ENGINE_load_cswift()
8702 ENGINE_load_atalla()
8704 ENGINE_load_builtin_engines()
8706 That way, an application can itself choose if external engines that
8707 are built-in in OpenSSL shall ever be used or not. The benefit is
8708 that applications won't have to be linked with libdl or other dso
8709 libraries unless it's really needed.
8711 Changed 'openssl engine' to load all engines on demand.
8712 Changed the engine header files to avoid the duplication of some
8713 declarations (they differed!).
8716 *) 'openssl engine' can now list capabilities.
8719 *) Better error reporting in 'openssl engine'.
8722 *) Never call load_dh_param(NULL) in s_server.
8725 *) Add engine application. It can currently list engines by name and
8726 identity, and test if they are actually available.
8729 *) Improve RPM specification file by forcing symbolic linking and making
8730 sure the installed documentation is also owned by root.root.
8731 [Damien Miller <djm@mindrot.org>]
8733 *) Give the OpenSSL applications more possibilities to make use of
8734 keys (public as well as private) handled by engines.
8737 *) Add OCSP code that comes from CertCo.
8740 *) Add VMS support for the Rijndael code.
8743 *) Added untested support for Nuron crypto accelerator.
8746 *) Add support for external cryptographic devices. This code was
8747 previously distributed separately as the "engine" branch.
8748 [Geoff Thorpe, Richard Levitte]
8750 *) Rework the filename-translation in the DSO code. It is now possible to
8751 have far greater control over how a "name" is turned into a filename
8752 depending on the operating environment and any oddities about the
8753 different shared library filenames on each system.
8756 *) Support threads on FreeBSD-elf in Configure.
8759 *) Fix for SHA1 assembly problem with MASM: it produces
8760 warnings about corrupt line number information when assembling
8761 with debugging information. This is caused by the overlapping
8763 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
8766 NCONF_get_number() has no error checking at all. As a replacement,
8767 NCONF_get_number_e() is defined (_e for "error checking") and is
8768 promoted strongly. The old NCONF_get_number is kept around for
8769 binary backward compatibility.
8770 Make it possible for methods to load from something other than a BIO,
8771 by providing a function pointer that is given a name instead of a BIO.
8772 For example, this could be used to load configuration data from an
8776 *) Fix for non blocking accept BIOs. Added new I/O special reason
8777 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
8778 with non blocking I/O was not possible because no retry code was
8779 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
8783 *) Added the beginnings of Rijndael support.
8786 *) Fix for bug in DirectoryString mask setting. Add support for
8787 X509_NAME_print_ex() in 'req' and X509_print_ex() function
8788 to allow certificate printing to more controllable, additional
8789 'certopt' option to 'x509' to allow new printing options to be
8793 *) Clean old EAY MD5 hack from e_os.h.
8796 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
8798 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
8799 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8800 [Joe Orton, Steve Henson]
8802 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
8804 *) Fix additional bug revealed by the NISCC test suite:
8806 Stop bug triggering large recursion when presented with
8807 certain ASN.1 tags (CVE-2003-0851)
8810 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
8812 *) Fix various bugs revealed by running the NISCC test suite:
8814 Stop out of bounds reads in the ASN1 code when presented with
8815 invalid tags (CVE-2003-0543 and CVE-2003-0544).
8817 If verify callback ignores invalid public key errors don't try to check
8818 certificate signature with the NULL public key.
8822 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
8823 if the server requested one: as stated in TLS 1.0 and SSL 3.0
8827 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
8828 extra data after the compression methods not only for TLS 1.0
8829 but also for SSL 3.0 (as required by the specification).
8830 [Bodo Moeller; problem pointed out by Matthias Loepfe]
8832 *) Change X509_certificate_type() to mark the key as exported/exportable
8833 when it's 512 *bits* long, not 512 bytes.
8836 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8838 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8839 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8840 a protocol version number mismatch like a decryption error
8841 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8844 *) Turn on RSA blinding by default in the default implementation
8845 to avoid a timing attack. Applications that don't want it can call
8846 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8847 They would be ill-advised to do so in most cases.
8848 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
8850 *) Change RSA blinding code so that it works when the PRNG is not
8851 seeded (in this case, the secret RSA exponent is abused as
8852 an unpredictable seed -- if it is not unpredictable, there
8853 is no point in blinding anyway). Make RSA blinding thread-safe
8854 by remembering the creator's thread ID in rsa->blinding and
8855 having all other threads use local one-time blinding factors
8856 (this requires more computation than sharing rsa->blinding, but
8857 avoids excessive locking; and if an RSA object is not shared
8858 between threads, blinding will still be very fast).
8861 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
8863 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8864 via timing by performing a MAC computation even if incorrect
8865 block cipher padding has been found. This is a countermeasure
8866 against active attacks where the attacker has to distinguish
8867 between bad padding and a MAC verification error. (CVE-2003-0078)
8869 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8870 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8871 Martin Vuagnoux (EPFL, Ilion)]
8873 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
8875 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
8876 memory from its contents. This is done with a counter that will
8877 place alternating values in each byte. This can be used to solve
8878 two issues: 1) the removal of calls to memset() by highly optimizing
8879 compilers, and 2) cleansing with other values than 0, since those can
8880 be read through on certain media, for example a swap space on disk.
8883 *) Bugfix: client side session caching did not work with external caching,
8884 because the session->cipher setting was not restored when reloading
8885 from the external cache. This problem was masked, when
8886 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
8887 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
8890 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
8891 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
8892 [Zeev Lieber <zeev-l@yahoo.com>]
8894 *) Undo an undocumented change introduced in 0.9.6e which caused
8895 repeated calls to OpenSSL_add_all_ciphers() and
8896 OpenSSL_add_all_digests() to be ignored, even after calling
8900 *) Change the default configuration reader to deal with last line not
8901 being properly terminated.
8904 *) Change X509_NAME_cmp() so it applies the special rules on handling
8905 DN values that are of type PrintableString, as well as RDNs of type
8906 emailAddress where the value has the type ia5String.
8907 [stefank@valicert.com via Richard Levitte]
8909 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
8910 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
8911 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
8912 the bitwise-OR of the two for use by the majority of applications
8913 wanting this behaviour, and update the docs. The documented
8914 behaviour and actual behaviour were inconsistent and had been
8915 changing anyway, so this is more a bug-fix than a behavioural
8917 [Geoff Thorpe, diagnosed by Nadav Har'El]
8919 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
8920 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
8923 *) Fix initialization code race conditions in
8924 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
8925 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
8926 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
8927 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
8928 ssl2_get_cipher_by_char(),
8929 ssl3_get_cipher_by_char().
8930 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
8932 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
8933 the cached sessions are flushed, as the remove_cb() might use ex_data
8934 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
8935 (see [openssl.org #212]).
8936 [Geoff Thorpe, Lutz Jaenicke]
8938 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
8939 length, instead of the encoding length to d2i_ASN1_OBJECT.
8942 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
8944 *) [In 0.9.6g-engine release:]
8945 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
8946 [Lynn Gazis <lgazis@rainbow.com>]
8948 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
8950 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
8951 and get fix the header length calculation.
8952 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
8953 Alon Kantor <alonk@checkpoint.com> (and others),
8956 *) Use proper error handling instead of 'assertions' in buffer
8957 overflow checks added in 0.9.6e. This prevents DoS (the
8958 assertions could call abort()).
8959 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
8961 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
8963 *) Add various sanity checks to asn1_get_length() to reject
8964 the ASN1 length bytes if they exceed sizeof(long), will appear
8965 negative or the content length exceeds the length of the
8967 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
8969 *) Fix cipher selection routines: ciphers without encryption had no flags
8970 for the cipher strength set and where therefore not handled correctly
8971 by the selection routines (PR #130).
8974 *) Fix EVP_dsa_sha macro.
8978 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
8979 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
8980 that was added in OpenSSL 0.9.6d.
8982 As the countermeasure turned out to be incompatible with some
8983 broken SSL implementations, the new option is part of SSL_OP_ALL.
8984 SSL_OP_ALL is usually employed when compatibility with weird SSL
8985 implementations is desired (e.g. '-bugs' option to 's_client' and
8986 's_server'), so the new option is automatically set in many
8990 *) Changes in security patch:
8992 Changes marked "(CHATS)" were sponsored by the Defense Advanced
8993 Research Projects Agency (DARPA) and Air Force Research Laboratory,
8994 Air Force Materiel Command, USAF, under agreement number
8997 *) Add various sanity checks to asn1_get_length() to reject
8998 the ASN1 length bytes if they exceed sizeof(long), will appear
8999 negative or the content length exceeds the length of the
9000 supplied buffer. (CVE-2002-0659)
9001 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9003 *) Assertions for various potential buffer overflows, not known to
9005 [Ben Laurie (CHATS)]
9007 *) Various temporary buffers to hold ASCII versions of integers were
9008 too small for 64 bit platforms. (CVE-2002-0655)
9009 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
9011 *) Remote buffer overflow in SSL3 protocol - an attacker could
9012 supply an oversized session ID to a client. (CVE-2002-0656)
9013 [Ben Laurie (CHATS)]
9015 *) Remote buffer overflow in SSL2 protocol - an attacker could
9016 supply an oversized client master key. (CVE-2002-0656)
9017 [Ben Laurie (CHATS)]
9019 Changes between 0.9.6c and 0.9.6d [9 May 2002]
9021 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
9022 encoded as NULL) with id-dsa-with-sha1.
9023 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
9025 *) Check various X509_...() return values in apps/req.c.
9026 [Nils Larsch <nla@trustcenter.de>]
9028 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
9029 an end-of-file condition would erroneously be flagged, when the CRLF
9030 was just at the end of a processed block. The bug was discovered when
9031 processing data through a buffering memory BIO handing the data to a
9032 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
9033 <ptsekov@syntrex.com> and Nedelcho Stanev.
9036 *) Implement a countermeasure against a vulnerability recently found
9037 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
9038 before application data chunks to avoid the use of known IVs
9039 with data potentially chosen by the attacker.
9042 *) Fix length checks in ssl3_get_client_hello().
9045 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
9046 to prevent ssl3_read_internal() from incorrectly assuming that
9047 ssl3_read_bytes() found application data while handshake
9048 processing was enabled when in fact s->s3->in_read_app_data was
9049 merely automatically cleared during the initial handshake.
9050 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
9052 *) Fix object definitions for Private and Enterprise: they were not
9053 recognized in their shortname (=lowercase) representation. Extend
9054 obj_dat.pl to issue an error when using undefined keywords instead
9055 of silently ignoring the problem (Svenning Sorensen
9056 <sss@sss.dnsalias.net>).
9059 *) Fix DH_generate_parameters() so that it works for 'non-standard'
9060 generators, i.e. generators other than 2 and 5. (Previously, the
9061 code did not properly initialise the 'add' and 'rem' values to
9062 BN_generate_prime().)
9064 In the new general case, we do not insist that 'generator' is
9065 actually a primitive root: This requirement is rather pointless;
9066 a generator of the order-q subgroup is just as good, if not
9070 *) Map new X509 verification errors to alerts. Discovered and submitted by
9071 Tom Wu <tom@arcot.com>.
9074 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
9075 returning non-zero before the data has been completely received
9076 when using non-blocking I/O.
9077 [Bodo Moeller; problem pointed out by John Hughes]
9079 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
9080 [Ben Laurie, Lutz Jaenicke]
9082 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
9083 Yoram Zahavi <YoramZ@gilian.com>).
9086 *) Add information about CygWin 1.3 and on, and preserve proper
9087 configuration for the versions before that.
9088 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
9090 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
9091 check whether we deal with a copy of a session and do not delete from
9092 the cache in this case. Problem reported by "Izhar Shoshani Levi"
9093 <izhar@checkpoint.com>.
9096 *) Do not store session data into the internal session cache, if it
9097 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
9098 flag is set). Proposed by Aslam <aslam@funk.com>.
9101 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
9105 *) [In 0.9.6d-engine release:]
9106 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
9107 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
9109 *) Add the configuration target linux-s390x.
9110 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
9112 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
9113 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
9114 variable as an indication that a ClientHello message has been
9115 received. As the flag value will be lost between multiple
9116 invocations of ssl3_accept when using non-blocking I/O, the
9117 function may not be aware that a handshake has actually taken
9118 place, thus preventing a new session from being added to the
9121 To avoid this problem, we now set s->new_session to 2 instead of
9122 using a local variable.
9123 [Lutz Jaenicke, Bodo Moeller]
9125 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
9126 if the SSL_R_LENGTH_MISMATCH error is detected.
9127 [Geoff Thorpe, Bodo Moeller]
9129 *) New 'shared_ldflag' column in Configure platform table.
9132 *) Fix EVP_CIPHER_mode macro.
9133 ["Dan S. Camper" <dan@bti.net>]
9135 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
9136 type, we must throw them away by setting rr->length to 0.
9137 [D P Chang <dpc@qualys.com>]
9139 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
9141 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
9142 <Dominikus.Scherkl@biodata.com>. (The previous implementation
9143 worked incorrectly for those cases where range = 10..._2 and
9144 3*range is two bits longer than range.)
9147 *) Only add signing time to PKCS7 structures if it is not already
9151 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
9152 OBJ_ld_ce should be OBJ_id_ce.
9153 Also some ip-pda OIDs in crypto/objects/objects.txt were
9154 incorrect (cf. RFC 3039).
9155 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
9157 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
9158 returns early because it has nothing to do.
9159 [Andy Schneider <andy.schneider@bjss.co.uk>]
9161 *) [In 0.9.6c-engine release:]
9162 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
9163 [Andy Schneider <andy.schneider@bjss.co.uk>]
9165 *) [In 0.9.6c-engine release:]
9166 Add support for Cryptographic Appliance's keyserver technology.
9167 (Use engine 'keyclient')
9168 [Cryptographic Appliances and Geoff Thorpe]
9170 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
9171 is called via tools/c89.sh because arguments have to be
9172 rearranged (all '-L' options must appear before the first object
9174 [Richard Shapiro <rshapiro@abinitio.com>]
9176 *) [In 0.9.6c-engine release:]
9177 Add support for Broadcom crypto accelerator cards, backported
9179 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
9181 *) [In 0.9.6c-engine release:]
9182 Add support for SureWare crypto accelerator cards from
9183 Baltimore Technologies. (Use engine 'sureware')
9184 [Baltimore Technologies and Mark Cox]
9186 *) [In 0.9.6c-engine release:]
9187 Add support for crypto accelerator cards from Accelerated
9188 Encryption Processing, www.aep.ie. (Use engine 'aep')
9189 [AEP Inc. and Mark Cox]
9191 *) Add a configuration entry for gcc on UnixWare.
9192 [Gary Benson <gbenson@redhat.com>]
9194 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
9195 messages are stored in a single piece (fixed-length part and
9196 variable-length part combined) and fix various bugs found on the way.
9199 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9200 instead. BIO_gethostbyname() does not know what timeouts are
9201 appropriate, so entries would stay in cache even when they have
9203 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9205 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9206 faced with a pathologically small ClientHello fragment that does
9207 not contain client_version: Instead of aborting with an error,
9208 simply choose the highest available protocol version (i.e.,
9209 TLS 1.0 unless it is disabled). In practice, ClientHello
9210 messages are never sent like this, but this change gives us
9211 strictly correct behaviour at least for TLS.
9214 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9215 never resets s->method to s->ctx->method when called from within
9216 one of the SSL handshake functions.
9217 [Bodo Moeller; problem pointed out by Niko Baric]
9219 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9220 (sent using the client's version number) if client_version is
9221 smaller than the protocol version in use. Also change
9222 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9223 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9224 the client will at least see that alert.
9227 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9231 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9232 client receives HelloRequest while in a handshake.
9233 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9235 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9236 should end in 'break', not 'goto end' which circumvents various
9237 cleanups done in state SSL_ST_OK. But session related stuff
9238 must be disabled for SSL_ST_OK in the case that we just sent a
9241 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9242 before just sending a HelloRequest.
9243 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9245 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9246 reveal whether illegal block cipher padding was found or a MAC
9247 verification error occurred. (Neither SSLerr() codes nor alerts
9248 are directly visible to potential attackers, but the information
9249 may leak via logfiles.)
9251 Similar changes are not required for the SSL 2.0 implementation
9252 because the number of padding bytes is sent in clear for SSL 2.0,
9253 and the extra bytes are just ignored. However ssl/s2_pkt.c
9254 failed to verify that the purported number of padding bytes is in
9258 *) Add OpenUNIX-8 support including shared libraries
9259 (Boyd Lynn Gerber <gerberb@zenez.com>).
9262 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9263 'wristwatch attack' using huge encoding parameters (cf.
9264 James H. Manger's CRYPTO 2001 paper). Note that the
9265 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9266 encoding parameters and hence was not vulnerable.
9269 *) BN_sqr() bug fix.
9270 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9272 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9273 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9274 followed by modular reduction.
9275 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9277 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9278 equivalent based on BN_pseudo_rand() instead of BN_rand().
9281 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9282 This function was broken, as the check for a new client hello message
9283 to handle SGC did not allow these large messages.
9284 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9287 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9290 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9291 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9294 *) Rework the configuration and shared library support for Tru64 Unix.
9295 The configuration part makes use of modern compiler features and
9296 still retains old compiler behavior for those that run older versions
9297 of the OS. The shared library support part includes a variant that
9298 uses the RPATH feature, and is available through the special
9299 configuration target "alpha-cc-rpath", which will never be selected
9301 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9303 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9304 with the same message size as in ssl3_get_certificate_request().
9305 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9306 messages might inadvertently be reject as too long.
9307 [Petr Lampa <lampa@fee.vutbr.cz>]
9309 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9312 *) Modified SSL library such that the verify_callback that has been set
9313 specificly for an SSL object with SSL_set_verify() is actually being
9314 used. Before the change, a verify_callback set with this function was
9315 ignored and the verify_callback() set in the SSL_CTX at the time of
9316 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9317 to allow the necessary settings.
9320 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9321 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9322 done automatically (in contradiction to the requirements of the C
9323 standard). This made problems when used from OpenSSH.
9326 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9327 dh->length and always used
9329 BN_rand_range(priv_key, dh->p).
9331 BN_rand_range() is not necessary for Diffie-Hellman, and this
9332 specific range makes Diffie-Hellman unnecessarily inefficient if
9333 dh->length (recommended exponent length) is much smaller than the
9334 length of dh->p. We could use BN_rand_range() if the order of
9335 the subgroup was stored in the DH structure, but we only have
9340 BN_rand(priv_key, l, ...)
9342 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
9348 RSA_eay_public_encrypt
9349 RSA_eay_private_decrypt
9350 RSA_eay_private_encrypt (signing)
9351 RSA_eay_public_decrypt (signature verification)
9353 (default implementations for RSA_public_encrypt,
9354 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
9355 always reject numbers >= n.
9358 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
9359 to synchronize access to 'locking_thread'. This is necessary on
9360 systems where access to 'locking_thread' (an 'unsigned long'
9361 variable) is not atomic.
9364 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
9365 *before* setting the 'crypto_lock_rand' flag. The previous code had
9366 a race condition if 0 is a valid thread ID.
9367 [Travis Vitek <vitek@roguewave.com>]
9369 *) Add support for shared libraries under Irix.
9370 [Albert Chin-A-Young <china@thewrittenword.com>]
9372 *) Add configuration option to build on Linux on both big-endian and
9374 [Ralf Baechle <ralf@uni-koblenz.de>]
9376 *) Add the possibility to create shared libraries on HP-UX.
9379 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
9381 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
9382 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
9383 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
9384 PRNG state recovery was possible based on the output of
9385 one PRNG request appropriately sized to gain knowledge on
9386 'md' followed by enough consecutive 1-byte PRNG requests
9387 to traverse all of 'state'.
9389 1. When updating 'md_local' (the current thread's copy of 'md')
9390 during PRNG output generation, hash all of the previous
9391 'md_local' value, not just the half used for PRNG output.
9393 2. Make the number of bytes from 'state' included into the hash
9394 independent from the number of PRNG bytes requested.
9396 The first measure alone would be sufficient to avoid
9397 Markku-Juhani's attack. (Actually it had never occurred
9398 to me that the half of 'md_local' used for chaining was the
9399 half from which PRNG output bytes were taken -- I had always
9400 assumed that the secret half would be used.) The second
9401 measure makes sure that additional data from 'state' is never
9402 mixed into 'md_local' in small portions; this heuristically
9403 further strengthens the PRNG.
9406 *) Fix crypto/bn/asm/mips3.s.
9409 *) When only the key is given to "enc", the IV is undefined. Print out
9410 an error message in this case.
9413 *) Handle special case when X509_NAME is empty in X509 printing routines.
9416 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
9417 positive and less than q.
9420 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
9421 used: it isn't thread safe and the add_lock_callback should handle
9423 [Paul Rose <Paul.Rose@bridge.com>]
9425 *) Verify that incoming data obeys the block size in
9426 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
9430 [Ulf Möller, Bodo Möller]
9432 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
9433 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
9434 when fixing the server behaviour for backwards-compatible 'client
9435 hello' messages. (Note that the attack is impractical against
9436 SSL 3.0 and TLS 1.0 anyway because length and version checking
9437 means that the probability of guessing a valid ciphertext is
9438 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
9441 Before 0.9.5, the countermeasure (hide the error by generating a
9442 random 'decryption result') did not work properly because
9443 ERR_clear_error() was missing, meaning that SSL_get_error() would
9444 detect the supposedly ignored error.
9446 Both problems are now fixed.
9449 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
9450 (previously it was 1024).
9453 *) Fix for compatibility mode trust settings: ignore trust settings
9454 unless some valid trust or reject settings are present.
9457 *) Fix for blowfish EVP: its a variable length cipher.
9460 *) Fix various bugs related to DSA S/MIME verification. Handle missing
9461 parameters in DSA public key structures and return an error in the
9462 DSA routines if parameters are absent.
9465 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
9466 in the current directory if neither $RANDFILE nor $HOME was set.
9467 RAND_file_name() in 0.9.6a returned NULL in this case. This has
9468 caused some confusion to Windows users who haven't defined $HOME.
9469 Thus RAND_file_name() is changed again: e_os.h can define a
9470 DEFAULT_HOME, which will be used if $HOME is not set.
9471 For Windows, we use "C:"; on other platforms, we still require
9472 environment variables.
9474 *) Move 'if (!initialized) RAND_poll()' into regions protected by
9475 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
9476 having multiple threads call RAND_poll() concurrently.
9479 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
9480 combination of a flag and a thread ID variable.
9481 Otherwise while one thread is in ssleay_rand_bytes (which sets the
9482 flag), *other* threads can enter ssleay_add_bytes without obeying
9483 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
9484 that they do not hold after the first thread unsets add_do_not_lock).
9487 *) Change bctest again: '-x' expressions are not available in all
9491 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
9493 *) Fix a couple of memory leaks in PKCS7_dataDecode()
9494 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9496 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9497 the default extension for executables, if any. Also, make the perl
9498 scripts that use symlink() to test if it really exists and use "cp"
9499 if it doesn't. All this made OpenSSL compilable and installable in
9503 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9504 If SEQUENCE is length is indefinite just set c->slen to the total
9505 amount of data available.
9506 [Steve Henson, reported by shige@FreeBSD.org]
9507 [This change does not apply to 0.9.7.]
9509 *) Change bctest to avoid here-documents inside command substitution
9510 (workaround for FreeBSD /bin/sh bug).
9511 For compatibility with Ultrix, avoid shell functions (introduced
9512 in the bctest version that searches along $PATH).
9515 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9516 with des_encrypt() defined on some operating systems, like Solaris
9520 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9521 On the Importance of Eliminating Errors in Cryptographic
9522 Computations, J. Cryptology 14 (2001) 2, 101-119,
9523 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9526 *) MIPS assembler BIGNUM division bug fix.
9529 *) Disabled incorrect Alpha assembler code.
9532 *) Fix PKCS#7 decode routines so they correctly update the length
9533 after reading an EOC for the EXPLICIT tag.
9535 [This change does not apply to 0.9.7.]
9537 *) Fix bug in PKCS#12 key generation routines. This was triggered
9538 if a 3DES key was generated with a 0 initial byte. Include
9539 PKCS12_BROKEN_KEYGEN compilation option to retain the old
9540 (but broken) behaviour.
9543 *) Enhance bctest to search for a working bc along $PATH and print
9545 [Tim Rice <tim@multitalents.net> via Richard Levitte]
9547 *) Fix memory leaks in err.c: free err_data string if necessary;
9548 don't write to the wrong index in ERR_set_error_data.
9551 *) Implement ssl23_peek (analogous to ssl23_read), which previously
9555 *) Replace rdtsc with _emit statements for VC++ version 5.
9556 [Jeremy Cooper <jeremy@baymoo.org>]
9558 *) Make it possible to reuse SSLv2 sessions.
9561 *) In copy_email() check for >= 0 as a return value for
9562 X509_NAME_get_index_by_NID() since 0 is a valid index.
9563 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
9565 *) Avoid coredump with unsupported or invalid public keys by checking if
9566 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
9567 PKCS7_verify() fails with non detached data.
9570 *) Don't use getenv in library functions when run as setuid/setgid.
9571 New function OPENSSL_issetugid().
9574 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
9575 due to incorrect handling of multi-threading:
9577 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
9579 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
9581 3. Count how many times MemCheck_off() has been called so that
9582 nested use can be treated correctly. This also avoids
9583 inband-signalling in the previous code (which relied on the
9584 assumption that thread ID 0 is impossible).
9587 *) Add "-rand" option also to s_client and s_server.
9590 *) Fix CPU detection on Irix 6.x.
9591 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
9592 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9594 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
9597 [This change does not apply to 0.9.7.]
9599 *) Use the cached encoding of an X509_NAME structure rather than
9600 copying it. This is apparently the reason for the libsafe "errors"
9601 but the code is actually correct.
9604 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
9605 Bleichenbacher's DSA attack.
9606 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
9607 to be set and top=0 forces the highest bit to be set; top=-1 is new
9608 and leaves the highest bit random.
9609 [Ulf Moeller, Bodo Moeller]
9611 *) In the NCONF_...-based implementations for CONF_... queries
9612 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
9613 a temporary CONF structure with the data component set to NULL
9614 (which gives segmentation faults in lh_retrieve).
9615 Instead, use NULL for the CONF pointer in CONF_get_string and
9616 CONF_get_number (which may use environment variables) and directly
9617 return NULL from CONF_get_section.
9620 *) Fix potential buffer overrun for EBCDIC.
9623 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9624 keyUsage if basicConstraints absent for a CA.
9627 *) Make SMIME_write_PKCS7() write mail header values with a format that
9628 is more generally accepted (no spaces before the semicolon), since
9629 some programs can't parse those values properly otherwise. Also make
9630 sure BIO's that break lines after each write do not create invalid
9634 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9635 macros previously used would not encode an empty SEQUENCE OF
9636 and break the signature.
9638 [This change does not apply to 0.9.7.]
9640 *) Zero the premaster secret after deriving the master secret in
9644 *) Add some EVP_add_digest_alias registrations (as found in
9645 OpenSSL_add_all_digests()) to SSL_library_init()
9646 aka OpenSSL_add_ssl_algorithms(). This provides improved
9647 compatibility with peers using X.509 certificates
9648 with unconventional AlgorithmIdentifier OIDs.
9651 *) Fix for Irix with NO_ASM.
9652 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9654 *) ./config script fixes.
9655 [Ulf Moeller, Richard Levitte]
9657 *) Fix 'openssl passwd -1'.
9660 *) Change PKCS12_key_gen_asc() so it can cope with non null
9661 terminated strings whose length is passed in the passlen
9662 parameter, for example from PEM callbacks. This was done
9663 by adding an extra length parameter to asc2uni().
9664 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9666 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9667 call failed, free the DSA structure.
9670 *) Fix to uni2asc() to cope with zero length Unicode strings.
9671 These are present in some PKCS#12 files.
9674 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9675 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9676 when writing a 32767 byte record.
9677 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9679 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9680 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9682 (RSA objects have a reference count access to which is protected
9683 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9684 so they are meant to be shared between threads.)
9685 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9686 "Reddie, Steven" <Steven.Reddie@ca.com>]
9688 *) Fix a deadlock in CRYPTO_mem_leaks().
9691 *) Use better test patterns in bntest.
9694 *) rand_win.c fix for Borland C.
9697 *) BN_rshift bugfix for n == 0.
9700 *) Add a 'bctest' script that checks for some known 'bc' bugs
9701 so that 'make test' does not abort just because 'bc' is broken.
9704 *) Store verify_result within SSL_SESSION also for client side to
9705 avoid potential security hole. (Re-used sessions on the client side
9706 always resulted in verify_result==X509_V_OK, not using the original
9707 result of the server certificate verification.)
9710 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9711 SSL3_RT_APPLICATION_DATA, return 0.
9712 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9716 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9717 releases, have been re-implemented by renaming the previous
9718 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9719 and ssl3_read_internal, respectively, and adding 'peek' parameters
9720 to them. The new ssl[23]_{read,peek} functions are calls to
9721 ssl[23]_read_internal with the 'peek' flag set appropriately.
9722 A 'peek' parameter has also been added to ssl3_read_bytes, which
9723 does the actual work for ssl3_read_internal.
9726 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9727 the method-specific "init()" handler. Also clean up ex_data after
9728 calling the method-specific "finish()" handler. Previously, this was
9729 happening the other way round.
9732 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9733 The previous value, 12, was not always sufficient for BN_mod_exp().
9736 *) Make sure that shared libraries get the internal name engine with
9737 the full version number and not just 0. This should mark the
9738 shared libraries as not backward compatible. Of course, this should
9739 be changed again when we can guarantee backward binary compatibility.
9742 *) Fix typo in get_cert_by_subject() in by_dir.c
9743 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9745 *) Rework the system to generate shared libraries:
9747 - Make note of the expected extension for the shared libraries and
9748 if there is a need for symbolic links from for example libcrypto.so.0
9749 to libcrypto.so.0.9.7. There is extended info in Configure for
9752 - Make as few rebuilds of the shared libraries as possible.
9754 - Still avoid linking the OpenSSL programs with the shared libraries.
9756 - When installing, install the shared libraries separately from the
9760 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
9762 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
9763 and not in SSL_clear because the latter is also used by the
9764 accept/connect functions; previously, the settings made by
9765 SSL_set_read_ahead would be lost during the handshake.
9766 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
9768 *) Correct util/mkdef.pl to be selective about disabled algorithms.
9769 Previously, it would create entries for disabled algorithms no
9773 *) Added several new manual pages for SSL_* function.
9776 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
9778 *) In ssl23_get_client_hello, generate an error message when faced
9779 with an initial SSL 3.0/TLS record that is too small to contain the
9780 first two bytes of the ClientHello message, i.e. client_version.
9781 (Note that this is a pathologic case that probably has never happened
9782 in real life.) The previous approach was to use the version number
9783 from the record header as a substitute; but our protocol choice
9784 should not depend on that one because it is not authenticated
9785 by the Finished messages.
9788 *) More robust randomness gathering functions for Windows.
9789 [Jeffrey Altman <jaltman@columbia.edu>]
9791 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
9792 not set then we don't setup the error code for issuer check errors
9793 to avoid possibly overwriting other errors which the callback does
9794 handle. If an application does set the flag then we assume it knows
9795 what it is doing and can handle the new informational codes
9799 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
9800 a general "ANY" type, as such it should be able to decode anything
9801 including tagged types. However it didn't check the class so it would
9802 wrongly interpret tagged types in the same way as their universal
9803 counterpart and unknown types were just rejected. Changed so that the
9804 tagged and unknown types are handled in the same way as a SEQUENCE:
9805 that is the encoding is stored intact. There is also a new type
9806 "V_ASN1_OTHER" which is used when the class is not universal, in this
9807 case we have no idea what the actual type is so we just lump them all
9811 *) On VMS, stdout may very well lead to a file that is written to
9812 in a record-oriented fashion. That means that every write() will
9813 write a separate record, which will be read separately by the
9814 programs trying to read from it. This can be very confusing.
9816 The solution is to put a BIO filter in the way that will buffer
9817 text until a linefeed is reached, and then write everything a
9818 line at a time, so every record written will be an actual line,
9819 not chunks of lines and not (usually doesn't happen, but I've
9820 seen it once) several lines in one record. BIO_f_linebuffer() is
9823 Currently, it's a VMS-only method, because that's where it has
9824 been tested well enough.
9827 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
9828 it can return incorrect results.
9829 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
9830 but it was in 0.9.6-beta[12].)
9833 *) Disable the check for content being present when verifying detached
9834 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9835 include zero length content when signing messages.
9838 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9839 BIO_ctrl (for BIO pairs).
9842 *) Add DSO method for VMS.
9845 *) Bug fix: Montgomery multiplication could produce results with the
9849 *) Add RPM specification openssl.spec and modify it to build three
9850 packages. The default package contains applications, application
9851 documentation and run-time libraries. The devel package contains
9852 include files, static libraries and function documentation. The
9853 doc package contains the contents of the doc directory. The original
9854 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
9857 *) Add a large number of documentation files for many SSL routines.
9858 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9860 *) Add a configuration entry for Sony News 4.
9861 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
9863 *) Don't set the two most significant bits to one when generating a
9864 random number < q in the DSA library.
9867 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
9868 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
9869 the underlying transport is blocking) if a handshake took place.
9870 (The default behaviour is needed by applications such as s_client
9871 and s_server that use select() to determine when to use SSL_read;
9872 but for applications that know in advance when to expect data, it
9873 just makes things more complicated.)
9876 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
9880 *) Add a few more EBCDIC conditionals that make `req' and `x509'
9881 work better on such systems.
9882 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9884 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
9885 Update PKCS12_parse() so it copies the friendlyName and the
9886 keyid to the certificates aux info.
9889 *) Fix bug in PKCS7_verify() which caused an infinite loop
9890 if there was more than one signature.
9891 [Sven Uszpelkat <su@celocom.de>]
9893 *) Major change in util/mkdef.pl to include extra information
9894 about each symbol, as well as presenting variables as well
9895 as functions. This change means that there's n more need
9896 to rebuild the .num files when some algorithms are excluded.
9899 *) Allow the verify time to be set by an application,
9900 rather than always using the current time.
9903 *) Phase 2 verify code reorganisation. The certificate
9904 verify code now looks up an issuer certificate by a
9905 number of criteria: subject name, authority key id
9906 and key usage. It also verifies self signed certificates
9907 by the same criteria. The main comparison function is
9908 X509_check_issued() which performs these checks.
9910 Lot of changes were necessary in order to support this
9911 without completely rewriting the lookup code.
9913 Authority and subject key identifier are now cached.
9915 The LHASH 'certs' is X509_STORE has now been replaced
9916 by a STACK_OF(X509_OBJECT). This is mainly because an
9917 LHASH can't store or retrieve multiple objects with
9918 the same hash value.
9920 As a result various functions (which were all internal
9921 use only) have changed to handle the new X509_STORE
9922 structure. This will break anything that messed round
9923 with X509_STORE internally.
9925 The functions X509_STORE_add_cert() now checks for an
9926 exact match, rather than just subject name.
9928 The X509_STORE API doesn't directly support the retrieval
9929 of multiple certificates matching a given criteria, however
9930 this can be worked round by performing a lookup first
9931 (which will fill the cache with candidate certificates)
9932 and then examining the cache for matches. This is probably
9933 the best we can do without throwing out X509_LOOKUP
9934 entirely (maybe later...).
9936 The X509_VERIFY_CTX structure has been enhanced considerably.
9938 All certificate lookup operations now go via a get_issuer()
9939 callback. Although this currently uses an X509_STORE it
9940 can be replaced by custom lookups. This is a simple way
9941 to bypass the X509_STORE hackery necessary to make this
9942 work and makes it possible to use more efficient techniques
9943 in future. A very simple version which uses a simple
9944 STACK for its trusted certificate store is also provided
9945 using X509_STORE_CTX_trusted_stack().
9947 The verify_cb() and verify() callbacks now have equivalents
9948 in the X509_STORE_CTX structure.
9950 X509_STORE_CTX also has a 'flags' field which can be used
9951 to customise the verify behaviour.
9954 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
9955 excludes S/MIME capabilities.
9958 *) When a certificate request is read in keep a copy of the
9959 original encoding of the signed data and use it when outputting
9960 again. Signatures then use the original encoding rather than
9961 a decoded, encoded version which may cause problems if the
9962 request is improperly encoded.
9965 *) For consistency with other BIO_puts implementations, call
9966 buffer_write(b, ...) directly in buffer_puts instead of calling
9969 In BIO_puts, increment b->num_write as in BIO_write.
9970 [Peter.Sylvester@EdelWeb.fr]
9972 *) Fix BN_mul_word for the case where the word is 0. (We have to use
9973 BN_zero, we may not return a BIGNUM with an array consisting of
9977 *) Avoid calling abort() from within the library when problems are
9978 detected, except if preprocessor symbols have been defined
9979 (such as REF_CHECK, BN_DEBUG etc.).
9982 *) New openssl application 'rsautl'. This utility can be
9983 used for low level RSA operations. DER public key
9984 BIO/fp routines also added.
9987 *) New Configure entry and patches for compiling on QNX 4.
9988 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
9990 *) A demo state-machine implementation was sponsored by
9991 Nuron (http://www.nuron.com/) and is now available in
9992 demos/state_machine.
9995 *) New options added to the 'dgst' utility for signature
9996 generation and verification.
9999 *) Unrecognized PKCS#7 content types are now handled via a
10000 catch all ASN1_TYPE structure. This allows unsupported
10001 types to be stored as a "blob" and an application can
10002 encode and decode it manually.
10005 *) Fix various signed/unsigned issues to make a_strex.c
10006 compile under VC++.
10007 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
10009 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
10010 length if passed a buffer. ASN1_INTEGER_to_BN failed
10011 if passed a NULL BN and its argument was negative.
10012 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
10014 *) Modification to PKCS#7 encoding routines to output definite
10015 length encoding. Since currently the whole structures are in
10016 memory there's not real point in using indefinite length
10017 constructed encoding. However if OpenSSL is compiled with
10018 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
10021 *) Added BIO_vprintf() and BIO_vsnprintf().
10024 *) Added more prefixes to parse for in the strings written
10025 through a logging bio, to cover all the levels that are available
10026 through syslog. The prefixes are now:
10028 PANIC, EMERG, EMR => LOG_EMERG
10029 ALERT, ALR => LOG_ALERT
10030 CRIT, CRI => LOG_CRIT
10031 ERROR, ERR => LOG_ERR
10032 WARNING, WARN, WAR => LOG_WARNING
10033 NOTICE, NOTE, NOT => LOG_NOTICE
10034 INFO, INF => LOG_INFO
10035 DEBUG, DBG => LOG_DEBUG
10037 and as before, if none of those prefixes are present at the
10038 beginning of the string, LOG_ERR is chosen.
10040 On Win32, the LOG_* levels are mapped according to this:
10042 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
10043 LOG_WARNING => EVENTLOG_WARNING_TYPE
10044 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
10048 *) Made it possible to reconfigure with just the configuration
10049 argument "reconf" or "reconfigure". The command line arguments
10050 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
10051 and are retrieved from there when reconfiguring.
10054 *) MD4 implemented.
10055 [Assar Westerlund <assar@sics.se>, Richard Levitte]
10057 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
10060 *) The obj_dat.pl script was messing up the sorting of object
10061 names. The reason was that it compared the quoted version
10062 of strings as a result "OCSP" > "OCSP Signing" because
10063 " > SPACE. Changed script to store unquoted versions of
10064 names and add quotes on output. It was also omitting some
10065 names from the lookup table if they were given a default
10066 value (that is if SN is missing it is given the same
10067 value as LN and vice versa), these are now added on the
10068 grounds that if an object has a name we should be able to
10069 look it up. Finally added warning output when duplicate
10070 short or long names are found.
10073 *) Changes needed for Tandem NSK.
10074 [Scott Uroff <scott@xypro.com>]
10076 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
10077 RSA_padding_check_SSLv23(), special padding was never detected
10078 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
10079 version rollback attacks was not effective.
10081 In s23_clnt.c, don't use special rollback-attack detection padding
10082 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
10083 client; similarly, in s23_srvr.c, don't do the rollback check if
10084 SSL 2.0 is the only protocol enabled in the server.
10087 *) Make it possible to get hexdumps of unprintable data with 'openssl
10088 asn1parse'. By implication, the functions ASN1_parse_dump() and
10089 BIO_dump_indent() are added.
10092 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
10093 these print out strings and name structures based on various
10094 flags including RFC2253 support and proper handling of
10095 multibyte characters. Added options to the 'x509' utility
10096 to allow the various flags to be set.
10099 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
10100 Also change the functions X509_cmp_current_time() and
10101 X509_gmtime_adj() work with an ASN1_TIME structure,
10102 this will enable certificates using GeneralizedTime in validity
10103 dates to be checked.
10106 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
10107 negative public key encodings) on by default,
10108 NO_NEG_PUBKEY_BUG can be set to disable it.
10111 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
10112 content octets. An i2c_ASN1_OBJECT is unnecessary because
10113 the encoding can be trivially obtained from the structure.
10116 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
10117 not read locks (CRYPTO_r_[un]lock).
10120 *) A first attempt at creating official support for shared
10121 libraries through configuration. I've kept it so the
10122 default is static libraries only, and the OpenSSL programs
10123 are always statically linked for now, but there are
10124 preparations for dynamic linking in place.
10125 This has been tested on Linux and Tru64.
10128 *) Randomness polling function for Win9x, as described in:
10129 Peter Gutmann, Software Generation of Practically Strong
10133 *) Fix so PRNG is seeded in req if using an already existing
10137 *) New options to smime application. -inform and -outform
10138 allow alternative formats for the S/MIME message including
10139 PEM and DER. The -content option allows the content to be
10140 specified separately. This should allow things like Netscape
10141 form signing output easier to verify.
10144 *) Fix the ASN1 encoding of tags using the 'long form'.
10147 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
10148 STRING types. These convert content octets to and from the
10149 underlying type. The actual tag and length octets are
10150 already assumed to have been read in and checked. These
10151 are needed because all other string types have virtually
10152 identical handling apart from the tag. By having versions
10153 of the ASN1 functions that just operate on content octets
10154 IMPLICIT tagging can be handled properly. It also allows
10155 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
10156 and ASN1_INTEGER are identical apart from the tag.
10159 *) Change the handling of OID objects as follows:
10161 - New object identifiers are inserted in objects.txt, following
10162 the syntax given in objects.README.
10163 - objects.pl is used to process obj_mac.num and create a new
10165 - obj_dat.pl is used to create a new obj_dat.h, using the data in
10168 This is currently kind of a hack, and the perl code in objects.pl
10169 isn't very elegant, but it works as I intended. The simplest way
10170 to check that it worked correctly is to look in obj_dat.h and
10171 check the array nid_objs and make sure the objects haven't moved
10172 around (this is important!). Additions are OK, as well as
10173 consistent name changes.
10176 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
10179 *) Addition of the command line parameter '-rand file' to 'openssl req'.
10180 The given file adds to whatever has already been seeded into the
10181 random pool through the RANDFILE configuration file option or
10182 environment variable, or the default random state file.
10185 *) mkstack.pl now sorts each macro group into lexical order.
10186 Previously the output order depended on the order the files
10187 appeared in the directory, resulting in needless rewriting
10191 *) Patches to make OpenSSL compile under Win32 again. Mostly
10192 work arounds for the VC++ problem that it treats func() as
10193 func(void). Also stripped out the parts of mkdef.pl that
10194 added extra typesafe functions: these no longer exist.
10197 *) Reorganisation of the stack code. The macros are now all
10198 collected in safestack.h . Each macro is defined in terms of
10199 a "stack macro" of the form SKM_<name>(type, a, b). The
10200 DEBUG_SAFESTACK is now handled in terms of function casts,
10201 this has the advantage of retaining type safety without the
10202 use of additional functions. If DEBUG_SAFESTACK is not defined
10203 then the non typesafe macros are used instead. Also modified the
10204 mkstack.pl script to handle the new form. Needs testing to see
10205 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10206 the default if no major problems. Similar behaviour for ASN1_SET_OF
10207 and PKCS12_STACK_OF.
10210 *) When some versions of IIS use the 'NET' form of private key the
10211 key derivation algorithm is different. Normally MD5(password) is
10212 used as a 128 bit RC4 key. In the modified case
10213 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10214 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10215 as the old Netscape_RSA functions except they have an additional
10216 'sgckey' parameter which uses the modified algorithm. Also added
10217 an -sgckey command line option to the rsa utility. Thanks to
10218 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10219 algorithm to openssl-dev.
10222 *) The evp_local.h macros were using 'c.##kname' which resulted in
10223 invalid expansion on some systems (SCO 5.0.5 for example).
10224 Corrected to 'c.kname'.
10225 [Phillip Porch <root@theporch.com>]
10227 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10228 a STACK of email addresses from a certificate or request, these look
10229 in the subject name and the subject alternative name extensions and
10230 omit any duplicate addresses.
10233 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10234 This makes DSA verification about 2 % faster.
10237 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10238 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10239 plus overhead for 1024 bit moduli).
10240 This makes exponentiations about 0.5 % faster for 1024 bit
10241 exponents (as measured by "openssl speed rsa2048").
10244 *) Rename memory handling macros to avoid conflicts with other
10246 Malloc => OPENSSL_malloc
10247 Malloc_locked => OPENSSL_malloc_locked
10248 Realloc => OPENSSL_realloc
10249 Free => OPENSSL_free
10252 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10253 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10256 *) CygWin32 support.
10257 [John Jarvie <jjarvie@newsguy.com>]
10259 *) The type-safe stack code has been rejigged. It is now only compiled
10260 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10261 by default all type-specific stack functions are "#define"d back to
10262 standard stack functions. This results in more streamlined output
10263 but retains the type-safety checking possibilities of the original
10267 *) The STACK code has been cleaned up, and certain type declarations
10268 that didn't make a lot of sense have been brought in line. This has
10269 also involved a cleanup of sorts in safestack.h to more correctly
10270 map type-safe stack functions onto their plain stack counterparts.
10271 This work has also resulted in a variety of "const"ifications of
10272 lots of the code, especially "_cmp" operations which should normally
10273 be prototyped with "const" parameters anyway.
10276 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10277 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10278 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10279 where all of 'md' is used each time the PRNG is used, but 'state'
10280 is used only indexed by a cyclic counter. As entropy may not be
10281 well distributed from the beginning, 'md' is important as a
10282 chaining variable. However, the output function chains only half
10283 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10284 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10285 in all of 'state' being rewritten, with the new values depending
10286 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10289 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10290 the handshake is continued after ssl_verify_cert_chain();
10291 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10292 can lead to 'unexplainable' connection aborts later.
10293 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10295 *) Major EVP API cipher revision.
10296 Add hooks for extra EVP features. This allows various cipher
10297 parameters to be set in the EVP interface. Support added for variable
10298 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10299 setting of RC2 and RC5 parameters.
10301 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10304 Remove lots of duplicated code from the EVP library. For example *every*
10305 cipher init() function handles the 'iv' in the same way according to the
10306 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10307 for CFB and OFB modes they zero ctx->num.
10309 New functionality allows removal of S/MIME code RC2 hack.
10311 Most of the routines have the same form and so can be declared in terms
10314 By shifting this to the top level EVP_CipherInit() it can be removed from
10315 all individual ciphers. If the cipher wants to handle IVs or keys
10316 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10319 Change lots of functions like EVP_EncryptUpdate() to now return a
10320 value: although software versions of the algorithms cannot fail
10321 any installed hardware versions can.
10324 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10325 this option is set, tolerate broken clients that send the negotiated
10326 protocol version number instead of the requested protocol version
10330 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10331 i.e. non-zero for export ciphersuites, zero otherwise.
10332 Previous versions had this flag inverted, inconsistent with
10333 rsa_tmp_cb (..._TMP_RSA_CB).
10334 [Bodo Moeller; problem reported by Amit Chopra]
10336 *) Add missing DSA library text string. Work around for some IIS
10337 key files with invalid SEQUENCE encoding.
10340 *) Add a document (doc/standards.txt) that list all kinds of standards
10341 and so on that are implemented in OpenSSL.
10344 *) Enhance c_rehash script. Old version would mishandle certificates
10345 with the same subject name hash and wouldn't handle CRLs at all.
10346 Added -fingerprint option to crl utility, to support new c_rehash
10350 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
10353 *) Fix for SSL server purpose checking. Server checking was
10354 rejecting certificates which had extended key usage present
10355 but no ssl client purpose.
10356 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
10358 *) Make PKCS#12 code work with no password. The PKCS#12 spec
10359 is a little unclear about how a blank password is handled.
10360 Since the password in encoded as a BMPString with terminating
10361 double NULL a zero length password would end up as just the
10362 double NULL. However no password at all is different and is
10363 handled differently in the PKCS#12 key generation code. NS
10364 treats a blank password as zero length. MSIE treats it as no
10365 password on export: but it will try both on import. We now do
10366 the same: PKCS12_parse() tries zero length and no password if
10367 the password is set to "" or NULL (NULL is now a valid password:
10368 it wasn't before) as does the pkcs12 application.
10371 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
10372 perror when PEM_read_bio_X509_REQ fails, the error message must
10373 be obtained from the error queue.
10376 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
10377 it in ERR_remove_state if appropriate, and change ERR_get_state
10378 accordingly to avoid race conditions (this is necessary because
10379 thread_hash is no longer constant once set).
10382 *) Bugfix for linux-elf makefile.one.
10385 *) RSA_get_default_method() will now cause a default
10386 RSA_METHOD to be chosen if one doesn't exist already.
10387 Previously this was only set during a call to RSA_new()
10388 or RSA_new_method(NULL) meaning it was possible for
10389 RSA_get_default_method() to return NULL.
10392 *) Added native name translation to the existing DSO code
10393 that will convert (if the flag to do so is set) filenames
10394 that are sufficiently small and have no path information
10395 into a canonical native form. Eg. "blah" converted to
10396 "libblah.so" or "blah.dll" etc.
10399 *) New function ERR_error_string_n(e, buf, len) which is like
10400 ERR_error_string(e, buf), but writes at most 'len' bytes
10401 including the 0 terminator. For ERR_error_string_n, 'buf'
10403 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
10405 *) CONF library reworked to become more general. A new CONF
10406 configuration file reader "class" is implemented as well as a
10407 new functions (NCONF_*, for "New CONF") to handle it. The now
10408 old CONF_* functions are still there, but are reimplemented to
10409 work in terms of the new functions. Also, a set of functions
10410 to handle the internal storage of the configuration data is
10411 provided to make it easier to write new configuration file
10412 reader "classes" (I can definitely see something reading a
10413 configuration file in XML format, for example), called _CONF_*,
10414 or "the configuration storage API"...
10416 The new configuration file reading functions are:
10418 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
10419 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
10421 NCONF_default, NCONF_WIN32
10423 NCONF_dump_fp, NCONF_dump_bio
10425 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
10426 NCONF_new creates a new CONF object. This works in the same way
10427 as other interfaces in OpenSSL, like the BIO interface.
10428 NCONF_dump_* dump the internal storage of the configuration file,
10429 which is useful for debugging. All other functions take the same
10430 arguments as the old CONF_* functions wth the exception of the
10431 first that must be a `CONF *' instead of a `LHASH *'.
10433 To make it easer to use the new classes with the old CONF_* functions,
10434 the function CONF_set_default_method is provided.
10437 *) Add '-tls1' option to 'openssl ciphers', which was already
10438 mentioned in the documentation but had not been implemented.
10439 (This option is not yet really useful because even the additional
10440 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
10443 *) Initial DSO code added into libcrypto for letting OpenSSL (and
10444 OpenSSL-based applications) load shared libraries and bind to
10445 them in a portable way.
10446 [Geoff Thorpe, with contributions from Richard Levitte]
10448 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
10450 *) Make sure _lrotl and _lrotr are only used with MSVC.
10452 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
10453 (the default implementation of RAND_status).
10455 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
10456 to '-clrext' (= clear extensions), as intended and documented.
10457 [Bodo Moeller; inconsistency pointed out by Michael Attili
10458 <attili@amaxo.com>]
10460 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
10461 was larger than the MD block size.
10462 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
10464 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
10465 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
10466 using the passed key: if the passed key was a private key the result
10467 of X509_print(), for example, would be to print out all the private key
10471 *) des_quad_cksum() byte order bug fix.
10472 [Ulf Möller, using the problem description in krb4-0.9.7, where
10473 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
10475 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
10477 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
10479 *) For easily testing in shell scripts whether some command
10480 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
10481 returns with exit code 0 iff no command of the given name is available.
10482 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
10483 the output goes to stdout and nothing is printed to stderr.
10484 Additional arguments are always ignored.
10486 Since for each cipher there is a command of the same name,
10487 the 'no-cipher' compilation switches can be tested this way.
10489 ('openssl no-XXX' is not able to detect pseudo-commands such
10490 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
10493 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10496 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10497 is set; it will be thrown away anyway because each handshake creates
10499 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10500 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10501 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10502 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10505 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10506 'Q' and 'R' lose their special meanings (quit/renegotiate).
10507 This is part of what -quiet does; unlike -quiet, -ign_eof
10508 does not suppress any output.
10511 *) Add compatibility options to the purpose and trust code. The
10512 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10513 accepts a certificate or CA, this was the previous behaviour,
10514 with all the associated security issues.
10516 X509_TRUST_COMPAT is the old trust behaviour: only and
10517 automatically trust self signed roots in certificate store. A
10518 new trust setting X509_TRUST_DEFAULT is used to specify that
10519 a purpose has no associated trust setting and it should instead
10520 use the value in the default purpose.
10523 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10524 and fix a memory leak.
10527 *) In util/mkerr.pl (which implements 'make errors'), preserve
10528 reason strings from the previous version of the .c file, as
10529 the default to have only downcase letters (and digits) in
10530 automatically generated reasons codes is not always appropriate.
10533 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
10534 using strerror. Previously, ERR_reason_error_string() returned
10535 library names as reason strings for SYSerr; but SYSerr is a special
10536 case where small numbers are errno values, not library numbers.
10539 *) Add '-dsaparam' option to 'openssl dhparam' application. This
10540 converts DSA parameters into DH parameters. (When creating parameters,
10541 DSA_generate_parameters is used.)
10544 *) Include 'length' (recommended exponent length) in C code generated
10545 by 'openssl dhparam -C'.
10548 *) The second argument to set_label in perlasm was already being used
10549 so couldn't be used as a "file scope" flag. Moved to third argument
10553 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
10554 instead of RAND_bytes for encryption IVs and salts.
10557 *) Include RAND_status() into RAND_METHOD instead of implementing
10558 it only for md_rand.c Otherwise replacing the PRNG by calling
10559 RAND_set_rand_method would be impossible.
10562 *) Don't let DSA_generate_key() enter an infinite loop if the random
10563 number generation fails.
10566 *) New 'rand' application for creating pseudo-random output.
10569 *) Added configuration support for Linux/IA64
10570 [Rolf Haberrecker <rolf@suse.de>]
10572 *) Assembler module support for Mingw32.
10575 *) Shared library support for HPUX (in shlib/).
10576 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
10578 *) Shared library support for Solaris gcc.
10579 [Lutz Behnke <behnke@trustcenter.de>]
10581 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
10583 *) PKCS7_encrypt() was adding text MIME headers twice because they
10584 were added manually and by SMIME_crlf_copy().
10587 *) In bntest.c don't call BN_rand with zero bits argument.
10588 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
10590 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
10591 case was implemented. This caused BN_div_recp() to fail occasionally.
10594 *) Add an optional second argument to the set_label() in the perl
10595 assembly language builder. If this argument exists and is set
10596 to 1 it signals that the assembler should use a symbol whose
10597 scope is the entire file, not just the current function. This
10598 is needed with MASM which uses the format label:: for this scope.
10599 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
10601 *) Change the ASN1 types so they are typedefs by default. Before
10602 almost all types were #define'd to ASN1_STRING which was causing
10603 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
10607 *) Change names of new functions to the new get1/get0 naming
10608 convention: After 'get1', the caller owns a reference count
10609 and has to call ..._free; 'get0' returns a pointer to some
10610 data structure without incrementing reference counters.
10611 (Some of the existing 'get' functions increment a reference
10612 counter, some don't.)
10613 Similarly, 'set1' and 'add1' functions increase reference
10614 counters or duplicate objects.
10617 *) Allow for the possibility of temp RSA key generation failure:
10618 the code used to assume it always worked and crashed on failure.
10621 *) Fix potential buffer overrun problem in BIO_printf().
10622 [Ulf Möller, using public domain code by Patrick Powell; problem
10623 pointed out by David Sacerdote <das33@cornell.edu>]
10625 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10626 RAND_egd() and RAND_status(). In the command line application,
10627 the EGD socket can be specified like a seed file using RANDFILE
10631 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10632 Some CAs (e.g. Verisign) distribute certificates in this form.
10635 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10636 list to exclude them. This means that no special compilation option
10637 is needed to use anonymous DH: it just needs to be included in the
10641 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10642 EVP_MD_type. The old functionality is available in a new macro called
10643 EVP_MD_md(). Change code that uses it and update docs.
10646 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10647 where the 'void *' argument is replaced by a function pointer argument.
10648 Previously 'void *' was abused to point to functions, which works on
10649 many platforms, but is not correct. As these functions are usually
10650 called by macros defined in OpenSSL header files, most source code
10651 should work without changes.
10654 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10655 sections with information on -D... compiler switches used for
10656 compiling the library so that applications can see them. To enable
10657 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10658 must be defined. E.g.,
10659 #define OPENSSL_ALGORITHM_DEFINES
10660 #include <openssl/opensslconf.h>
10661 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10662 [Richard Levitte, Ulf and Bodo Möller]
10664 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10668 *) Change the 'other' type in certificate aux info to a STACK_OF
10669 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10670 the required ASN1 format: arbitrary types determined by an OID.
10673 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10674 argument to 'req'. This is not because the function is newer or
10675 better than others it just uses the work 'NEW' in the certificate
10676 request header lines. Some software needs this.
10679 *) Reorganise password command line arguments: now passwords can be
10680 obtained from various sources. Delete the PEM_cb function and make
10681 it the default behaviour: i.e. if the callback is NULL and the
10682 usrdata argument is not NULL interpret it as a null terminated pass
10683 phrase. If usrdata and the callback are NULL then the pass phrase
10684 is prompted for as usual.
10687 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10688 the support is automatically enabled. The resulting binaries will
10689 autodetect the card and use it if present.
10690 [Ben Laurie and Compaq Inc.]
10692 *) Work around for Netscape hang bug. This sends certificate request
10693 and server done in one record. Since this is perfectly legal in the
10694 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10695 the bugs/SSLv3 entry for more info.
10698 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10701 *) Add -rand argument to smime and pkcs12 applications and read/write
10705 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10708 *) Add command line password options to the remaining applications.
10711 *) Bug fix for BN_div_recp() for numerators with an even number of
10715 *) More tests in bntest.c, and changed test_bn output.
10718 *) ./config recognizes MacOS X now.
10721 *) Bug fix for BN_div() when the first words of num and divisor are
10722 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10725 *) Add support for various broken PKCS#8 formats, and command line
10726 options to produce them.
10729 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10730 get temporary BIGNUMs from a BN_CTX.
10733 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10737 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10738 include a #define from the old name to the new. The original intent
10739 was that statically linked binaries could for example just call
10740 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10741 link with digests. This never worked because SSLeay_add_all_digests()
10742 and SSLeay_add_all_ciphers() were in the same source file so calling
10743 one would link with the other. They are now in separate source files.
10746 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10749 *) Use a less unusual form of the Miller-Rabin primality test (it used
10750 a binary algorithm for exponentiation integrated into the Miller-Rabin
10751 loop, our standard modexp algorithms are faster).
10754 *) Support for the EBCDIC character set completed.
10755 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10757 *) Source code cleanups: use const where appropriate, eliminate casts,
10758 use void * instead of char * in lhash.
10761 *) Bugfix: ssl3_send_server_key_exchange was not restartable
10762 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
10763 this the server could overwrite ephemeral keys that the client
10767 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
10768 using 50 iterations of the Rabin-Miller test.
10770 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
10771 iterations of the Rabin-Miller test as required by the appendix
10772 to FIPS PUB 186[-1]) instead of DSA_is_prime.
10773 As BN_is_prime_fasttest includes trial division, DSA parameter
10774 generation becomes much faster.
10776 This implies a change for the callback functions in DSA_is_prime
10777 and DSA_generate_parameters: The callback function is called once
10778 for each positive witness in the Rabin-Miller test, not just
10779 occasionally in the inner loop; and the parameters to the
10780 callback function now provide an iteration count for the outer
10781 loop rather than for the current invocation of the inner loop.
10782 DSA_generate_parameters additionally can call the callback
10783 function with an 'iteration count' of -1, meaning that a
10784 candidate has passed the trial division test (when q is generated
10785 from an application-provided seed, trial division is skipped).
10788 *) New function BN_is_prime_fasttest that optionally does trial
10789 division before starting the Rabin-Miller test and has
10790 an additional BN_CTX * argument (whereas BN_is_prime always
10791 has to allocate at least one BN_CTX).
10792 'callback(1, -1, cb_arg)' is called when a number has passed the
10793 trial division stage.
10796 *) Fix for bug in CRL encoding. The validity dates weren't being handled
10800 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
10803 *) New function BN_pseudo_rand().
10806 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
10807 bignum version of BN_from_montgomery() with the working code from
10808 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
10812 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
10813 made it impossible to use the same SSL_SESSION data structure in
10814 SSL2 clients in multiple threads.
10817 *) The return value of RAND_load_file() no longer counts bytes obtained
10818 by stat(). RAND_load_file(..., -1) is new and uses the complete file
10819 to seed the PRNG (previously an explicit byte count was required).
10820 [Ulf Möller, Bodo Möller]
10822 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
10823 used (char *) instead of (void *) and had casts all over the place.
10826 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
10829 *) Retain source code compatibility for BN_prime_checks macro:
10830 BN_is_prime(..., BN_prime_checks, ...) now uses
10831 BN_prime_checks_for_size to determine the appropriate number of
10832 Rabin-Miller iterations.
10835 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10836 DH_CHECK_P_NOT_SAFE_PRIME.
10837 (Check if this is true? OpenPGP calls them "strong".)
10840 *) Merge the functionality of "dh" and "gendh" programs into a new program
10841 "dhparam". The old programs are retained for now but will handle DH keys
10842 (instead of parameters) in future.
10845 *) Make the ciphers, s_server and s_client programs check the return values
10846 when a new cipher list is set.
10849 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
10850 ciphers. Before when the 56bit ciphers were enabled the sorting was
10853 The syntax for the cipher sorting has been extended to support sorting by
10854 cipher-strength (using the strength_bits hard coded in the tables).
10855 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
10857 Fix a bug in the cipher-command parser: when supplying a cipher command
10858 string with an "undefined" symbol (neither command nor alphanumeric
10859 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
10860 an error is flagged.
10862 Due to the strength-sorting extension, the code of the
10863 ssl_create_cipher_list() function was completely rearranged. I hope that
10864 the readability was also increased :-)
10865 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10867 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
10868 for the first serial number and places 2 in the serial number file. This
10869 avoids problems when the root CA is created with serial number zero and
10870 the first user certificate has the same issuer name and serial number
10874 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
10875 the new code. Add documentation for this stuff.
10878 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
10879 X509_*() to X509at_*() on the grounds that they don't handle X509
10880 structures and behave in an analogous way to the X509v3 functions:
10881 they shouldn't be called directly but wrapper functions should be used
10884 So we also now have some wrapper functions that call the X509at functions
10885 when passed certificate requests. (TO DO: similar things can be done with
10886 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
10887 things. Some of these need some d2i or i2d and print functionality
10888 because they handle more complex structures.)
10891 *) Add missing #ifndefs that caused missing symbols when building libssl
10892 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
10893 NO_RSA in ssl/s2*.c.
10894 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
10896 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
10897 has a return value which indicates the quality of the random data
10898 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
10899 error queue. New function RAND_pseudo_bytes() generates output that is
10900 guaranteed to be unique but not unpredictable. RAND_add is like
10901 RAND_seed, but takes an extra argument for an entropy estimate
10902 (RAND_seed always assumes full entropy).
10905 *) Do more iterations of Rabin-Miller probable prime test (specifically,
10906 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
10907 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
10908 in crypto/bn/bn_prime.c for the complete table). This guarantees a
10909 false-positive rate of at most 2^-80 for random input.
10912 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
10915 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
10916 in the 0.9.5 release), this returns the chain
10917 from an X509_CTX structure with a dup of the stack and all
10918 the X509 reference counts upped: so the stack will exist
10919 after X509_CTX_cleanup() has been called. Modify pkcs12.c
10922 Also make SSL_SESSION_print() print out the verify return
10926 *) Add manpage for the pkcs12 command. Also change the default
10927 behaviour so MAC iteration counts are used unless the new
10928 -nomaciter option is used. This improves file security and
10929 only older versions of MSIE (4.0 for example) need it.
10932 *) Honor the no-xxx Configure options when creating .DEF files.
10935 *) Add PKCS#10 attributes to field table: challengePassword,
10936 unstructuredName and unstructuredAddress. These are taken from
10937 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
10938 international characters are used.
10940 More changes to X509_ATTRIBUTE code: allow the setting of types
10941 based on strings. Remove the 'loc' parameter when adding
10942 attributes because these will be a SET OF encoding which is sorted
10946 *) Initial changes to the 'req' utility to allow request generation
10947 automation. This will allow an application to just generate a template
10948 file containing all the field values and have req construct the
10951 Initial support for X509_ATTRIBUTE handling. Stacks of these are
10952 used all over the place including certificate requests and PKCS#7
10953 structures. They are currently handled manually where necessary with
10954 some primitive wrappers for PKCS#7. The new functions behave in a
10955 manner analogous to the X509 extension functions: they allow
10956 attributes to be looked up by NID and added.
10958 Later something similar to the X509V3 code would be desirable to
10959 automatically handle the encoding, decoding and printing of the
10960 more complex types. The string types like challengePassword can
10961 be handled by the string table functions.
10963 Also modified the multi byte string table handling. Now there is
10964 a 'global mask' which masks out certain types. The table itself
10965 can use the flag STABLE_NO_MASK to ignore the mask setting: this
10966 is useful when for example there is only one permissible type
10967 (as in countryName) and using the mask might result in no valid
10971 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
10972 SSL_get_peer_finished to allow applications to obtain the latest
10973 Finished messages sent to the peer or expected from the peer,
10974 respectively. (SSL_get_peer_finished is usually the Finished message
10975 actually received from the peer, otherwise the protocol will be aborted.)
10977 As the Finished message are message digests of the complete handshake
10978 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
10979 be used for external authentication procedures when the authentication
10980 provided by SSL/TLS is not desired or is not enough.
10983 *) Enhanced support for Alpha Linux is added. Now ./config checks if
10984 the host supports BWX extension and if Compaq C is present on the
10985 $PATH. Just exploiting of the BWX extension results in 20-30%
10986 performance kick for some algorithms, e.g. DES and RC4 to mention
10987 a couple. Compaq C in turn generates ~20% faster code for MD5 and
10991 *) Add support for MS "fast SGC". This is arguably a violation of the
10992 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
10993 weak crypto and after checking the certificate is SGC a second one
10994 with strong crypto. MS SGC stops the first handshake after receiving
10995 the server certificate message and sends a second client hello. Since
10996 a server will typically do all the time consuming operations before
10997 expecting any further messages from the client (server key exchange
10998 is the most expensive) there is little difference between the two.
11000 To get OpenSSL to support MS SGC we have to permit a second client
11001 hello message after we have sent server done. In addition we have to
11002 reset the MAC if we do get this second client hello.
11005 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
11006 if a DER encoded private key is RSA or DSA traditional format. Changed
11007 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
11008 format DER encoded private key. Newer code should use PKCS#8 format which
11009 has the key type encoded in the ASN1 structure. Added DER private key
11010 support to pkcs8 application.
11013 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
11014 ciphersuites has been selected (as required by the SSL 3/TLS 1
11015 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
11016 is set, we interpret this as a request to violate the specification
11017 (the worst that can happen is a handshake failure, and 'correct'
11018 behaviour would result in a handshake failure anyway).
11021 *) In SSL_CTX_add_session, take into account that there might be multiple
11022 SSL_SESSION structures with the same session ID (e.g. when two threads
11023 concurrently obtain them from an external cache).
11024 The internal cache can handle only one SSL_SESSION with a given ID,
11025 so if there's a conflict, we now throw out the old one to achieve
11029 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
11030 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
11031 some routines that use cipher OIDs: some ciphers do not have OIDs
11032 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
11036 *) Simplify the trust setting structure and code. Now we just have
11037 two sequences of OIDs for trusted and rejected settings. These will
11038 typically have values the same as the extended key usage extension
11039 and any application specific purposes.
11041 The trust checking code now has a default behaviour: it will just
11042 check for an object with the same NID as the passed id. Functions can
11043 be provided to override either the default behaviour or the behaviour
11044 for a given id. SSL client, server and email already have functions
11045 in place for compatibility: they check the NID and also return "trusted"
11046 if the certificate is self signed.
11049 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
11050 traditional format into an EVP_PKEY structure.
11053 *) Add a password callback function PEM_cb() which either prompts for
11054 a password if usr_data is NULL or otherwise assumes it is a null
11055 terminated password. Allow passwords to be passed on command line
11056 environment or config files in a few more utilities.
11059 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
11060 keys. Add some short names for PKCS#8 PBE algorithms and allow them
11061 to be specified on the command line for the pkcs8 and pkcs12 utilities.
11062 Update documentation.
11065 *) Support for ASN1 "NULL" type. This could be handled before by using
11066 ASN1_TYPE but there wasn't any function that would try to read a NULL
11067 and produce an error if it couldn't. For compatibility we also have
11068 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
11069 don't allocate anything because they don't need to.
11072 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
11074 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
11076 *) Rebuild of the memory allocation routines used by OpenSSL code and
11077 possibly others as well. The purpose is to make an interface that
11078 provide hooks so anyone can build a separate set of allocation and
11079 deallocation routines to be used by OpenSSL, for example memory
11080 pool implementations, or something else, which was previously hard
11081 since Malloc(), Realloc() and Free() were defined as macros having
11082 the values malloc, realloc and free, respectively (except for Win32
11083 compilations). The same is provided for memory debugging code.
11084 OpenSSL already comes with functionality to find memory leaks, but
11085 this gives people a chance to debug other memory problems.
11087 With these changes, a new set of functions and macros have appeared:
11089 CRYPTO_set_mem_debug_functions() [F]
11090 CRYPTO_get_mem_debug_functions() [F]
11091 CRYPTO_dbg_set_options() [F]
11092 CRYPTO_dbg_get_options() [F]
11093 CRYPTO_malloc_debug_init() [M]
11095 The memory debug functions are NULL by default, unless the library
11096 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
11097 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
11098 gives the standard debugging functions that come with OpenSSL) or
11099 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
11100 provided by the library user) must be used. When the standard
11101 debugging functions are used, CRYPTO_dbg_set_options can be used to
11102 request additional information:
11103 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
11104 the CRYPTO_MDEBUG_xxx macro when compiling the library.
11106 Also, things like CRYPTO_set_mem_functions will always give the
11107 expected result (the new set of functions is used for allocation
11108 and deallocation) at all times, regardless of platform and compiler
11111 To finish it up, some functions that were never use in any other
11112 way than through macros have a new API and new semantic:
11114 CRYPTO_dbg_malloc()
11115 CRYPTO_dbg_realloc()
11118 All macros of value have retained their old syntax.
11119 [Richard Levitte and Bodo Moeller]
11121 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
11122 ordering of SMIMECapabilities wasn't in "strength order" and there
11123 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
11127 *) Some ASN1 types with illegal zero length encoding (INTEGER,
11128 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
11129 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
11131 *) Merge in my S/MIME library for OpenSSL. This provides a simple
11132 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
11133 functionality to handle multipart/signed properly) and a utility
11134 called 'smime' to call all this stuff. This is based on code I
11135 originally wrote for Celo who have kindly allowed it to be
11136 included in OpenSSL.
11139 *) Add variants des_set_key_checked and des_set_key_unchecked of
11140 des_set_key (aka des_key_sched). Global variable des_check_key
11141 decides which of these is called by des_set_key; this way
11142 des_check_key behaves as it always did, but applications and
11143 the library itself, which was buggy for des_check_key == 1,
11144 have a cleaner way to pick the version they need.
11147 *) New function PKCS12_newpass() which changes the password of a
11151 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
11152 dynamic mix. In both cases the ids can be used as an index into the
11153 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
11154 functions so they accept a list of the field values and the
11155 application doesn't need to directly manipulate the X509_TRUST
11159 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
11163 *) Modify the way the V3 extension code looks up extensions. This now
11164 works in a similar way to the object code: we have some "standard"
11165 extensions in a static table which is searched with OBJ_bsearch()
11166 and the application can add dynamic ones if needed. The file
11167 crypto/x509v3/ext_dat.h now has the info: this file needs to be
11168 updated whenever a new extension is added to the core code and kept
11169 in ext_nid order. There is a simple program 'tabtest.c' which checks
11170 this. New extensions are not added too often so this file can readily
11171 be maintained manually.
11173 There are two big advantages in doing things this way. The extensions
11174 can be looked up immediately and no longer need to be "added" using
11175 X509V3_add_standard_extensions(): this function now does nothing.
11176 [Side note: I get *lots* of email saying the extension code doesn't
11177 work because people forget to call this function]
11178 Also no dynamic allocation is done unless new extensions are added:
11179 so if we don't add custom extensions there is no need to call
11180 X509V3_EXT_cleanup().
11183 *) Modify enc utility's salting as follows: make salting the default. Add a
11184 magic header, so unsalted files fail gracefully instead of just decrypting
11185 to garbage. This is because not salting is a big security hole, so people
11186 should be discouraged from doing it.
11189 *) Fixes and enhancements to the 'x509' utility. It allowed a message
11190 digest to be passed on the command line but it only used this
11191 parameter when signing a certificate. Modified so all relevant
11192 operations are affected by the digest parameter including the
11193 -fingerprint and -x509toreq options. Also -x509toreq choked if a
11194 DSA key was used because it didn't fix the digest.
11197 *) Initial certificate chain verify code. Currently tests the untrusted
11198 certificates for consistency with the verify purpose (which is set
11199 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11201 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11202 this is because it will reject chains with invalid extensions whereas
11203 every previous version of OpenSSL and SSLeay made no checks at all.
11205 Trust code: checks the root CA for the relevant trust settings. Trust
11206 settings have an initial value consistent with the verify purpose: e.g.
11207 if the verify purpose is for SSL client use it expects the CA to be
11208 trusted for SSL client use. However the default value can be changed to
11209 permit custom trust settings: one example of this would be to only trust
11210 certificates from a specific "secure" set of CAs.
11212 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11213 which should be used for version portability: especially since the
11214 verify structure is likely to change more often now.
11216 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11217 to set them. If not set then assume SSL clients will verify SSL servers
11220 Two new options to the verify program: -untrusted allows a set of
11221 untrusted certificates to be passed in and -purpose which sets the
11222 intended purpose of the certificate. If a purpose is set then the
11223 new chain verify code is used to check extension consistency.
11226 *) Support for the authority information access extension.
11229 *) Modify RSA and DSA PEM read routines to transparently handle
11230 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11231 public keys in a format compatible with certificate
11232 SubjectPublicKeyInfo structures. Unfortunately there were already
11233 functions called *_PublicKey_* which used various odd formats so
11234 these are retained for compatibility: however the DSA variants were
11235 never in a public release so they have been deleted. Changed dsa/rsa
11236 utilities to handle the new format: note no releases ever handled public
11237 keys so we should be OK.
11239 The primary motivation for this change is to avoid the same fiasco
11240 that dogs private keys: there are several incompatible private key
11241 formats some of which are standard and some OpenSSL specific and
11242 require various evil hacks to allow partial transparent handling and
11243 even then it doesn't work with DER formats. Given the option anything
11244 other than PKCS#8 should be dumped: but the other formats have to
11245 stay in the name of compatibility.
11247 With public keys and the benefit of hindsight one standard format
11248 is used which works with EVP_PKEY, RSA or DSA structures: though
11249 it clearly returns an error if you try to read the wrong kind of key.
11251 Added a -pubkey option to the 'x509' utility to output the public key.
11252 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11253 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11254 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11255 that do the same as the EVP_PKEY_assign_*() except they up the
11256 reference count of the added key (they don't "swallow" the
11260 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11261 CRLs would fail if the file contained no certificates or no CRLs:
11262 added a new function to read in both types and return the number
11263 read: this means that if none are read it will be an error. The
11264 DER versions of the certificate and CRL reader would always fail
11265 because it isn't possible to mix certificates and CRLs in DER format
11266 without choking one or the other routine. Changed this to just read
11267 a certificate: this is the best we can do. Also modified the code
11268 in apps/verify.c to take notice of return codes: it was previously
11269 attempting to read in certificates from NULL pointers and ignoring
11270 any errors: this is one reason why the cert and CRL reader seemed
11271 to work. It doesn't check return codes from the default certificate
11272 routines: these may well fail if the certificates aren't installed.
11275 *) Code to support otherName option in GeneralName.
11278 *) First update to verify code. Change the verify utility
11279 so it warns if it is passed a self signed certificate:
11280 for consistency with the normal behaviour. X509_verify
11281 has been modified to it will now verify a self signed
11282 certificate if *exactly* the same certificate appears
11283 in the store: it was previously impossible to trust a
11284 single self signed certificate. This means that:
11285 openssl verify ss.pem
11286 now gives a warning about a self signed certificate but
11287 openssl verify -CAfile ss.pem ss.pem
11291 *) For servers, store verify_result in SSL_SESSION data structure
11292 (and add it to external session representation).
11293 This is needed when client certificate verifications fails,
11294 but an application-provided verification callback (set by
11295 SSL_CTX_set_cert_verify_callback) allows accepting the session
11296 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11297 but returns 1): When the session is reused, we have to set
11298 ssl->verify_result to the appropriate error code to avoid
11300 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11302 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11303 case in PKCS7_dataInit() where the signed PKCS7 structure
11304 didn't contain any existing data because it was being created.
11305 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11307 *) Add a salt to the key derivation routines in enc.c. This
11308 forms the first 8 bytes of the encrypted file. Also add a
11309 -S option to allow a salt to be input on the command line.
11312 *) New function X509_cmp(). Oddly enough there wasn't a function
11313 to compare two certificates. We do this by working out the SHA1
11314 hash and comparing that. X509_cmp() will be needed by the trust
11318 *) SSL_get1_session() is like SSL_get_session(), but increments
11319 the reference count in the SSL_SESSION returned.
11320 [Geoff Thorpe <geoff@eu.c2.net>]
11322 *) Fix for 'req': it was adding a null to request attributes.
11323 Also change the X509_LOOKUP and X509_INFO code to handle
11324 certificate auxiliary information.
11327 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11331 *) Add the possibility to add extra information to the memory leak
11332 detecting output, to form tracebacks, showing from where each
11333 allocation was originated: CRYPTO_push_info("constant string") adds
11334 the string plus current file name and line number to a per-thread
11335 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11336 is like calling CYRPTO_pop_info() until the stack is empty.
11337 Also updated memory leak detection code to be multi-thread-safe.
11340 *) Add options -text and -noout to pkcs7 utility and delete the
11341 encryption options which never did anything. Update docs.
11344 *) Add options to some of the utilities to allow the pass phrase
11345 to be included on either the command line (not recommended on
11346 OSes like Unix) or read from the environment. Update the
11347 manpages and fix a few bugs.
11350 *) Add a few manpages for some of the openssl commands.
11353 *) Fix the -revoke option in ca. It was freeing up memory twice,
11354 leaking and not finding already revoked certificates.
11357 *) Extensive changes to support certificate auxiliary information.
11358 This involves the use of X509_CERT_AUX structure and X509_AUX
11359 functions. An X509_AUX function such as PEM_read_X509_AUX()
11360 can still read in a certificate file in the usual way but it
11361 will also read in any additional "auxiliary information". By
11362 doing things this way a fair degree of compatibility can be
11363 retained: existing certificates can have this information added
11364 using the new 'x509' options.
11366 Current auxiliary information includes an "alias" and some trust
11367 settings. The trust settings will ultimately be used in enhanced
11368 certificate chain verification routines: currently a certificate
11369 can only be trusted if it is self signed and then it is trusted
11373 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
11374 The problem was that one of the replacement routines had not been working
11375 since SSLeay releases. For now the offending routine has been replaced
11376 with non-optimised assembler. Even so, this now gives around 95%
11377 performance improvement for 1024 bit RSA signs.
11380 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
11381 handling. Most clients have the effective key size in bits equal to
11382 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
11383 A few however don't do this and instead use the size of the decrypted key
11384 to determine the RC2 key length and the AlgorithmIdentifier to determine
11385 the effective key length. In this case the effective key length can still
11386 be 40 bits but the key length can be 168 bits for example. This is fixed
11387 by manually forcing an RC2 key into the EVP_PKEY structure because the
11388 EVP code can't currently handle unusual RC2 key sizes: it always assumes
11389 the key length and effective key length are equal.
11392 *) Add a bunch of functions that should simplify the creation of
11393 X509_NAME structures. Now you should be able to do:
11394 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
11395 and have it automatically work out the correct field type and fill in
11396 the structures. The more adventurous can try:
11397 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
11398 and it will (hopefully) work out the correct multibyte encoding.
11401 *) Change the 'req' utility to use the new field handling and multibyte
11402 copy routines. Before the DN field creation was handled in an ad hoc
11403 way in req, ca, and x509 which was rather broken and didn't support
11404 BMPStrings or UTF8Strings. Since some software doesn't implement
11405 BMPStrings or UTF8Strings yet, they can be enabled using the config file
11406 using the dirstring_type option. See the new comment in the default
11407 openssl.cnf for more info.
11410 *) Make crypto/rand/md_rand.c more robust:
11411 - Assure unique random numbers after fork().
11412 - Make sure that concurrent threads access the global counter and
11413 md serializably so that we never lose entropy in them
11414 or use exactly the same state in multiple threads.
11415 Access to the large state is not always serializable because
11416 the additional locking could be a performance killer, and
11417 md should be large enough anyway.
11420 *) New file apps/app_rand.c with commonly needed functionality
11421 for handling the random seed file.
11423 Use the random seed file in some applications that previously did not:
11425 dsaparam -genkey (which also ignored its '-rand' option),
11428 x509 (when signing).
11429 Except on systems with /dev/urandom, it is crucial to have a random
11430 seed file at least for key creation, DSA signing, and for DH exchanges;
11431 for RSA signatures we could do without one.
11433 gendh and gendsa (unlike genrsa) used to read only the first byte
11434 of each file listed in the '-rand' option. The function as previously
11435 found in genrsa is now in app_rand.c and is used by all programs
11436 that support '-rand'.
11439 *) In RAND_write_file, use mode 0600 for creating files;
11440 don't just chmod when it may be too late.
11443 *) Report an error from X509_STORE_load_locations
11444 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
11447 *) New function ASN1_mbstring_copy() this copies a string in either
11448 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
11449 into an ASN1_STRING type. A mask of permissible types is passed
11450 and it chooses the "minimal" type to use or an error if not type
11454 *) Add function equivalents to the various macros in asn1.h. The old
11455 macros are retained with an M_ prefix. Code inside the library can
11456 use the M_ macros. External code (including the openssl utility)
11457 should *NOT* in order to be "shared library friendly".
11460 *) Add various functions that can check a certificate's extensions
11461 to see if it usable for various purposes such as SSL client,
11462 server or S/MIME and CAs of these types. This is currently
11463 VERY EXPERIMENTAL but will ultimately be used for certificate chain
11464 verification. Also added a -purpose flag to x509 utility to
11465 print out all the purposes.
11468 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
11472 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
11473 for, obtain and decode and extension and obtain its critical flag.
11474 This allows all the necessary extension code to be handled in a
11475 single function call.
11478 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
11479 platforms. See crypto/rc4/rc4_enc.c for further details.
11482 *) New -noout option to asn1parse. This causes no output to be produced
11483 its main use is when combined with -strparse and -out to extract data
11484 from a file (which may not be in ASN.1 format).
11487 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
11488 when producing the local key id.
11489 [Richard Levitte <levitte@stacken.kth.se>]
11491 *) New option -dhparam in s_server. This allows a DH parameter file to be
11492 stated explicitly. If it is not stated then it tries the first server
11493 certificate file. The previous behaviour hard coded the filename
11497 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11498 a public key to be input or output. For example:
11499 openssl rsa -in key.pem -pubout -out pubkey.pem
11500 Also added necessary DSA public key functions to handle this.
11503 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11504 in the message. This was handled by allowing
11505 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11506 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11508 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11509 to the end of the strings whereas this didn't. This would cause problems
11510 if strings read with d2i_ASN1_bytes() were later modified.
11511 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11513 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11514 data and it contains EOF it will end up returning an error. This is
11515 caused by input 46 bytes long. The cause is due to the way base64
11516 BIOs find the start of base64 encoded data. They do this by trying a
11517 trial decode on each line until they find one that works. When they
11518 do a flag is set and it starts again knowing it can pass all the
11519 data directly through the decoder. Unfortunately it doesn't reset
11520 the context it uses. This means that if EOF is reached an attempt
11521 is made to pass two EOFs through the context and this causes the
11522 resulting error. This can also cause other problems as well. As is
11523 usual with these problems it takes *ages* to find and the fix is
11524 trivial: move one line.
11525 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
11527 *) Ugly workaround to get s_client and s_server working under Windows. The
11528 old code wouldn't work because it needed to select() on sockets and the
11529 tty (for keypresses and to see if data could be written). Win32 only
11530 supports select() on sockets so we select() with a 1s timeout on the
11531 sockets and then see if any characters are waiting to be read, if none
11532 are present then we retry, we also assume we can always write data to
11533 the tty. This isn't nice because the code then blocks until we've
11534 received a complete line of data and it is effectively polling the
11535 keyboard at 1s intervals: however it's quite a bit better than not
11536 working at all :-) A dedicated Windows application might handle this
11537 with an event loop for example.
11540 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
11541 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
11542 will be called when RSA_sign() and RSA_verify() are used. This is useful
11543 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
11544 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
11545 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
11546 This necessitated the support of an extra signature type NID_md5_sha1
11547 for SSL signatures and modifications to the SSL library to use it instead
11548 of calling RSA_public_decrypt() and RSA_private_encrypt().
11551 *) Add new -verify -CAfile and -CApath options to the crl program, these
11552 will lookup a CRL issuers certificate and verify the signature in a
11553 similar way to the verify program. Tidy up the crl program so it
11554 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
11555 less strict. It will now permit CRL extensions even if it is not
11556 a V2 CRL: this will allow it to tolerate some broken CRLs.
11559 *) Initialize all non-automatic variables each time one of the openssl
11560 sub-programs is started (this is necessary as they may be started
11561 multiple times from the "OpenSSL>" prompt).
11562 [Lennart Bang, Bodo Moeller]
11564 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
11565 removing all other RSA functionality (this is what NO_RSA does). This
11566 is so (for example) those in the US can disable those operations covered
11567 by the RSA patent while allowing storage and parsing of RSA keys and RSA
11571 *) Non-copying interface to BIO pairs.
11572 (still largely untested)
11575 *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
11576 ASCII string. This was handled independently in various places before.
11579 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
11580 UTF8 strings a character at a time.
11583 *) Use client_version from client hello to select the protocol
11584 (s23_srvr.c) and for RSA client key exchange verification
11585 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
11588 *) Add various utility functions to handle SPKACs, these were previously
11589 handled by poking round in the structure internals. Added new function
11590 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
11591 print, verify and generate SPKACs. Based on an original idea from
11592 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
11595 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
11598 *) Allow the config file extension section to be overwritten on the
11599 command line. Based on an original idea from Massimiliano Pala
11600 <madwolf@comune.modena.it>. The new option is called -extensions
11601 and can be applied to ca, req and x509. Also -reqexts to override
11602 the request extensions in req and -crlexts to override the crl extensions
11606 *) Add new feature to the SPKAC handling in ca. Now you can include
11607 the same field multiple times by preceding it by "XXXX." for example:
11610 this is the same syntax as used in the req config file.
11613 *) Allow certificate extensions to be added to certificate requests. These
11614 are specified in a 'req_extensions' option of the req section of the
11615 config file. They can be printed out with the -text option to req but
11616 are otherwise ignored at present.
11619 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11620 data read consists of only the final block it would not decrypted because
11621 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11622 A misplaced 'break' also meant the decrypted final block might not be
11623 copied until the next read.
11626 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11627 a few extra parameters to the DH structure: these will be useful if
11628 for example we want the value of 'q' or implement X9.42 DH.
11631 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11632 provides hooks that allow the default DSA functions or functions on a
11633 "per key" basis to be replaced. This allows hardware acceleration and
11634 hardware key storage to be handled without major modification to the
11635 library. Also added low level modexp hooks and CRYPTO_EX structure and
11636 associated functions.
11639 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11640 as "read only": it can't be written to and the buffer it points to will
11641 not be freed. Reading from a read only BIO is much more efficient than
11642 a normal memory BIO. This was added because there are several times when
11643 an area of memory needs to be read from a BIO. The previous method was
11644 to create a memory BIO and write the data to it, this results in two
11645 copies of the data and an O(n^2) reading algorithm. There is a new
11646 function BIO_new_mem_buf() which creates a read only memory BIO from
11647 an area of memory. Also modified the PKCS#7 routines to use read only
11651 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11652 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11653 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11654 but a retry condition occurred while trying to read the rest.
11657 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11658 NID_pkcs7_encrypted by default: this was wrong since this should almost
11659 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11660 the encrypted data type: this is a more sensible place to put it and it
11661 allows the PKCS#12 code to be tidied up that duplicated this
11665 *) Changed obj_dat.pl script so it takes its input and output files on
11666 the command line. This should avoid shell escape redirection problems
11670 *) Initial support for certificate extension requests, these are included
11671 in things like Xenroll certificate requests. Included functions to allow
11672 extensions to be obtained and added.
11675 *) -crlf option to s_client and s_server for sending newlines as
11676 CRLF (as required by many protocols).
11679 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11681 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11682 [Ralf S. Engelschall]
11684 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11685 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11687 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11691 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11692 DH parameters/keys (q is lost during that conversion, but the resulting
11693 DH parameters contain its length).
11695 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11696 much faster than DH_generate_parameters (which creates parameters
11697 where p = 2*q + 1), and also the smaller q makes DH computations
11698 much more efficient (160-bit exponentiation instead of 1024-bit
11699 exponentiation); so this provides a convenient way to support DHE
11700 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11701 utter importance to use
11702 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11704 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11705 when such DH parameters are used, because otherwise small subgroup
11706 attacks may become possible!
11709 *) Avoid memory leak in i2d_DHparams.
11712 *) Allow the -k option to be used more than once in the enc program:
11713 this allows the same encrypted message to be read by multiple recipients.
11716 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11717 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11718 it will always use the numerical form of the OID, even if it has a short
11722 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11723 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11724 otherwise bn_mod_exp was called. In the case of hardware keys for example
11725 no private key components need be present and it might store extra data
11726 in the RSA structure, which cannot be accessed from bn_mod_exp.
11727 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11728 private key operations.
11731 *) Added support for SPARC Linux.
11734 *) pem_password_cb function type incompatibly changed from
11735 typedef int pem_password_cb(char *buf, int size, int rwflag);
11737 ....(char *buf, int size, int rwflag, void *userdata);
11738 so that applications can pass data to their callbacks:
11739 The PEM[_ASN1]_{read,write}... functions and macros now take an
11740 additional void * argument, which is just handed through whenever
11741 the password callback is called.
11742 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11744 New function SSL_CTX_set_default_passwd_cb_userdata.
11746 Compatibility note: As many C implementations push function arguments
11747 onto the stack in reverse order, the new library version is likely to
11748 interoperate with programs that have been compiled with the old
11749 pem_password_cb definition (PEM_whatever takes some data that
11750 happens to be on the stack as its last argument, and the callback
11751 just ignores this garbage); but there is no guarantee whatsoever that
11754 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11755 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11756 problems not only on Windows, but also on some Unix platforms.
11757 To avoid problematic command lines, these definitions are now in an
11758 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11759 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
11762 *) MIPS III/IV assembler module is reimplemented.
11765 *) More DES library cleanups: remove references to srand/rand and
11766 delete an unused file.
11769 *) Add support for the free Netwide assembler (NASM) under Win32,
11770 since not many people have MASM (ml) and it can be hard to obtain.
11771 This is currently experimental but it seems to work OK and pass all
11772 the tests. Check out INSTALL.W32 for info.
11775 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
11776 without temporary keys kept an extra copy of the server key,
11777 and connections with temporary keys did not free everything in case
11781 *) New function RSA_check_key and new openssl rsa option -check
11782 for verifying the consistency of RSA keys.
11783 [Ulf Moeller, Bodo Moeller]
11785 *) Various changes to make Win32 compile work:
11786 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
11787 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
11788 comparison" warnings.
11789 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
11792 *) Add a debugging option to PKCS#5 v2 key generation function: when
11793 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
11794 derived keys are printed to stderr.
11797 *) Copy the flags in ASN1_STRING_dup().
11798 [Roman E. Pavlov <pre@mo.msk.ru>]
11800 *) The x509 application mishandled signing requests containing DSA
11801 keys when the signing key was also DSA and the parameters didn't match.
11803 It was supposed to omit the parameters when they matched the signing key:
11804 the verifying software was then supposed to automatically use the CA's
11805 parameters if they were absent from the end user certificate.
11807 Omitting parameters is no longer recommended. The test was also
11808 the wrong way round! This was probably due to unusual behaviour in
11809 EVP_cmp_parameters() which returns 1 if the parameters match.
11810 This meant that parameters were omitted when they *didn't* match and
11811 the certificate was useless. Certificates signed with 'ca' didn't have
11813 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
11815 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
11816 The interface is as follows:
11817 Applications can use
11818 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
11819 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
11820 "off" is now the default.
11821 The library internally uses
11822 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
11823 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
11824 to disable memory-checking temporarily.
11826 Some inconsistent states that previously were possible (and were
11827 even the default) are now avoided.
11829 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
11830 with each memory chunk allocated; this is occasionally more helpful
11831 than just having a counter.
11833 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11835 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11839 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11840 which largely parallels "options", but is for changing API behaviour,
11841 whereas "options" are about protocol behaviour.
11842 Initial "mode" flags are:
11844 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
11845 a single record has been written.
11846 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
11847 retries use the same buffer location.
11848 (But all of the contents must be
11852 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
11855 *) Fix problems with no-hmac etc.
11856 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
11858 *) New functions RSA_get_default_method(), RSA_set_method() and
11859 RSA_get_method(). These allows replacement of RSA_METHODs without having
11860 to mess around with the internals of an RSA structure.
11863 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
11864 Also really enable memory leak checks in openssl.c and in some
11866 [Chad C. Mulligan, Bodo Moeller]
11868 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
11869 up the length of negative integers. This has now been simplified to just
11870 store the length when it is first determined and use it later, rather
11871 than trying to keep track of where data is copied and updating it to
11873 [Steve Henson, reported by Brien Wheeler
11874 <bwheeler@authentica-security.com>]
11876 *) Add a new function PKCS7_signatureVerify. This allows the verification
11877 of a PKCS#7 signature but with the signing certificate passed to the
11878 function itself. This contrasts with PKCS7_dataVerify which assumes the
11879 certificate is present in the PKCS#7 structure. This isn't always the
11880 case: certificates can be omitted from a PKCS#7 structure and be
11881 distributed by "out of band" means (such as a certificate database).
11884 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
11885 function prototypes in pem.h, also change util/mkdef.pl to add the
11886 necessary function names.
11889 *) mk1mf.pl (used by Windows builds) did not properly read the
11890 options set by Configure in the top level Makefile, and Configure
11891 was not even able to write more than one option correctly.
11892 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
11895 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
11896 file to be loaded from a BIO or FILE pointer. The BIO version will
11897 for example allow memory BIOs to contain config info.
11900 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
11901 Whoever hopes to achieve shared-library compatibility across versions
11902 must use this, not the compile-time macro.
11903 (Exercise 0.9.4: Which is the minimum library version required by
11905 Note: All this applies only to multi-threaded programs, others don't
11909 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
11910 through a BIO pair triggered the default case, i.e.
11911 SSLerr(...,SSL_R_UNKNOWN_STATE).
11914 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
11915 can use the SSL library even if none of the specific BIOs is
11919 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
11920 for the encoded length.
11921 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
11923 *) Add initial documentation of the X509V3 functions.
11926 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
11927 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
11928 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
11929 secure PKCS#8 private key format with a high iteration count.
11932 *) Fix determination of Perl interpreter: A perl or perl5
11933 _directory_ in $PATH was also accepted as the interpreter.
11934 [Ralf S. Engelschall]
11936 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
11937 wrong with it but it was very old and did things like calling
11938 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
11939 unusual formatting.
11942 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
11943 to use the new extension code.
11946 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
11947 with macros. This should make it easier to change their form, add extra
11948 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
11952 *) Add to configuration table a new entry that can specify an alternative
11953 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
11954 according to Mark Crispin <MRC@Panda.COM>.
11958 *) DES CBC did not update the IV. Weird.
11961 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
11962 Changing the behaviour of the former might break existing programs --
11963 where IV updating is needed, des_ncbc_encrypt can be used.
11966 *) When bntest is run from "make test" it drives bc to check its
11967 calculations, as well as internally checking them. If an internal check
11968 fails, it needs to cause bc to give a non-zero result or make test carries
11969 on without noticing the failure. Fixed.
11972 *) DES library cleanups.
11975 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
11976 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
11977 ciphers. NOTE: although the key derivation function has been verified
11978 against some published test vectors it has not been extensively tested
11979 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
11983 *) Instead of "mkdir -p", which is not fully portable, use new
11984 Perl script "util/mkdir-p.pl".
11987 *) Rewrite the way password based encryption (PBE) is handled. It used to
11988 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
11989 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
11990 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
11991 the 'parameter' field of the AlgorithmIdentifier is passed to the
11992 underlying key generation function so it must do its own ASN1 parsing.
11993 This has also changed the EVP_PBE_CipherInit() function which now has a
11994 'parameter' argument instead of literal salt and iteration count values
11995 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
11998 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
11999 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
12000 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
12001 KEY" because this clashed with PKCS#8 unencrypted string. Since this
12002 value was just used as a "magic string" and not used directly its
12003 value doesn't matter.
12006 *) Introduce some semblance of const correctness to BN. Shame C doesn't
12010 *) "linux-sparc64" configuration (ultrapenguin).
12011 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
12012 "linux-sparc" configuration.
12013 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
12015 *) config now generates no-xxx options for missing ciphers.
12018 *) Support the EBCDIC character set (work in progress).
12019 File ebcdic.c not yet included because it has a different license.
12020 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12022 *) Support BS2000/OSD-POSIX.
12023 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12025 *) Make callbacks for key generation use void * instead of char *.
12028 *) Make S/MIME samples compile (not yet tested).
12031 *) Additional typesafe stacks.
12034 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
12038 Changes between 0.9.3 and 0.9.3a [29 May 1999]
12040 *) New configuration variant "sco5-gcc".
12042 *) Updated some demos.
12043 [Sean O Riordain, Wade Scholine]
12045 *) Add missing BIO_free at exit of pkcs12 application.
12048 *) Fix memory leak in conf.c.
12051 *) Updates for Win32 to assembler version of MD5.
12054 *) Set #! path to perl in apps/der_chop to where we found it
12055 instead of using a fixed path.
12058 *) SHA library changes for irix64-mips4-cc.
12061 *) Improvements for VMS support.
12065 Changes between 0.9.2b and 0.9.3 [24 May 1999]
12067 *) Bignum library bug fix. IRIX 6 passes "make test" now!
12068 This also avoids the problems with SC4.2 and unpatched SC5.
12069 [Andy Polyakov <appro@fy.chalmers.se>]
12071 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
12072 These are required because of the typesafe stack would otherwise break
12073 existing code. If old code used a structure member which used to be STACK
12074 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
12075 sk_num or sk_value it would produce an error because the num, data members
12076 are not present in STACK_OF. Now it just produces a warning. sk_set
12077 replaces the old method of assigning a value to sk_value
12078 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
12079 that does this will no longer work (and should use sk_set instead) but
12080 this could be regarded as a "questionable" behaviour anyway.
12083 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
12084 correctly handle encrypted S/MIME data.
12087 *) Change type of various DES function arguments from des_cblock
12088 (which means, in function argument declarations, pointer to char)
12089 to des_cblock * (meaning pointer to array with 8 char elements),
12090 which allows the compiler to do more typechecking; it was like
12091 that back in SSLeay, but with lots of ugly casts.
12093 Introduce new type const_des_cblock.
12096 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
12097 problems: find RecipientInfo structure that matches recipient certificate
12098 and initialise the ASN1 structures properly based on passed cipher.
12101 *) Belatedly make the BN tests actually check the results.
12104 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
12105 to and from BNs: it was completely broken. New compilation option
12106 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
12107 key elements as negative integers.
12110 *) Reorganize and speed up MD5.
12111 [Andy Polyakov <appro@fy.chalmers.se>]
12114 [Richard Levitte <richard@levitte.org>]
12116 *) New option -out to asn1parse to allow the parsed structure to be
12117 output to a file. This is most useful when combined with the -strparse
12118 option to examine the output of things like OCTET STRINGS.
12121 *) Make SSL library a little more fool-proof by not requiring any longer
12122 that SSL_set_{accept,connect}_state be called before
12123 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
12124 in many applications because usually everything *appeared* to work as
12125 intended anyway -- now it really works as intended).
12128 *) Move openssl.cnf out of lib/.
12131 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
12132 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
12133 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
12134 [Ralf S. Engelschall]
12136 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
12137 handle PKCS#7 enveloped data properly.
12138 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
12140 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
12141 copying pointers. The cert_st handling is changed by this in
12142 various ways (and thus what used to be known as ctx->default_cert
12143 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
12144 any longer when s->cert does not give us what we need).
12145 ssl_cert_instantiate becomes obsolete by this change.
12146 As soon as we've got the new code right (possibly it already is?),
12147 we have solved a couple of bugs of the earlier code where s->cert
12148 was used as if it could not have been shared with other SSL structures.
12150 Note that using the SSL API in certain dirty ways now will result
12151 in different behaviour than observed with earlier library versions:
12152 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
12153 does not influence s as it used to.
12155 In order to clean up things more thoroughly, inside SSL_SESSION
12156 we don't use CERT any longer, but a new structure SESS_CERT
12157 that holds per-session data (if available); currently, this is
12158 the peer's certificate chain and, for clients, the server's certificate
12159 and temporary key. CERT holds only those values that can have
12160 meaningful defaults in an SSL_CTX.
12163 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
12164 from the internal representation. Various PKCS#7 fixes: remove some
12165 evil casts and set the enc_dig_alg field properly based on the signing
12169 *) Allow PKCS#12 password to be set from the command line or the
12170 environment. Let 'ca' get its config file name from the environment
12171 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
12175 *) Allow certificate policies extension to use an IA5STRING for the
12176 organization field. This is contrary to the PKIX definition but
12177 VeriSign uses it and IE5 only recognises this form. Document 'x509'
12181 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
12182 without disallowing inline assembler and the like for non-pedantic builds.
12185 *) Support Borland C++ builder.
12186 [Janez Jere <jj@void.si>, modified by Ulf Möller]
12188 *) Support Mingw32.
12191 *) SHA-1 cleanups and performance enhancements.
12192 [Andy Polyakov <appro@fy.chalmers.se>]
12194 *) Sparc v8plus assembler for the bignum library.
12195 [Andy Polyakov <appro@fy.chalmers.se>]
12197 *) Accept any -xxx and +xxx compiler options in Configure.
12200 *) Update HPUX configuration.
12203 *) Add missing sk_<type>_unshift() function to safestack.h
12204 [Ralf S. Engelschall]
12206 *) New function SSL_CTX_use_certificate_chain_file that sets the
12207 "extra_cert"s in addition to the certificate. (This makes sense
12208 only for "PEM" format files, as chains as a whole are not
12212 *) Support verify_depth from the SSL API.
12213 x509_vfy.c had what can be considered an off-by-one-error:
12214 Its depth (which was not part of the external interface)
12215 was actually counting the number of certificates in a chain;
12216 now it really counts the depth.
12219 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12220 instead of X509err, which often resulted in confusing error
12221 messages since the error codes are not globally unique
12222 (e.g. an alleged error in ssl3_accept when a certificate
12223 didn't match the private key).
12225 *) New function SSL_CTX_set_session_id_context that allows to set a default
12226 value (so that you don't need SSL_set_session_id_context for each
12227 connection using the SSL_CTX).
12230 *) OAEP decoding bug fix.
12233 *) Support INSTALL_PREFIX for package builders, as proposed by
12237 *) New Configure options "threads" and "no-threads". For systems
12238 where the proper compiler options are known (currently Solaris
12239 and Linux), "threads" is the default.
12242 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12245 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12246 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12247 such as /usr/local/bin.
12250 *) "make linux-shared" to build shared libraries.
12251 [Niels Poppe <niels@netbox.org>]
12253 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12256 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12257 extension adding in x509 utility.
12260 *) Remove NOPROTO sections and error code comments.
12263 *) Partial rewrite of the DEF file generator to now parse the ANSI
12267 *) New Configure options --prefix=DIR and --openssldir=DIR.
12270 *) Complete rewrite of the error code script(s). It is all now handled
12271 by one script at the top level which handles error code gathering,
12272 header rewriting and C source file generation. It should be much better
12273 than the old method: it now uses a modified version of Ulf's parser to
12274 read the ANSI prototypes in all header files (thus the old K&R definitions
12275 aren't needed for error creation any more) and do a better job of
12276 translating function codes into names. The old 'ASN1 error code imbedded
12277 in a comment' is no longer necessary and it doesn't use .err files which
12278 have now been deleted. Also the error code call doesn't have to appear all
12279 on one line (which resulted in some large lines...).
12282 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12285 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12286 0 (which usually indicates a closed connection), but continue reading.
12289 *) Fix some race conditions.
12292 *) Add support for CRL distribution points extension. Add Certificate
12293 Policies and CRL distribution points documentation.
12296 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12299 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12300 8 of keying material. Merlin has also confirmed interop with this fix
12301 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12302 [Merlin Hughes <merlin@baltimore.ie>]
12304 *) Fix lots of warnings.
12305 [Richard Levitte <levitte@stacken.kth.se>]
12307 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12308 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12309 [Richard Levitte <levitte@stacken.kth.se>]
12311 *) Fix problems with sizeof(long) == 8.
12312 [Andy Polyakov <appro@fy.chalmers.se>]
12314 *) Change functions to ANSI C.
12317 *) Fix typos in error codes.
12318 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12320 *) Remove defunct assembler files from Configure.
12323 *) SPARC v8 assembler BIGNUM implementation.
12324 [Andy Polyakov <appro@fy.chalmers.se>]
12326 *) Support for Certificate Policies extension: both print and set.
12327 Various additions to support the r2i method this uses.
12330 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12331 return a const string when you are expecting an allocated buffer.
12334 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12335 types DirectoryString and DisplayText.
12338 *) Add code to allow r2i extensions to access the configuration database,
12339 add an LHASH database driver and add several ctx helper functions.
12342 *) Fix an evil bug in bn_expand2() which caused various BN functions to
12343 fail when they extended the size of a BIGNUM.
12346 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
12347 support typesafe stack.
12350 *) Fix typo in SSL_[gs]et_options().
12351 [Nils Frostberg <nils@medcom.se>]
12353 *) Delete various functions and files that belonged to the (now obsolete)
12354 old X509V3 handling code.
12357 *) New Configure option "rsaref".
12360 *) Don't auto-generate pem.h.
12363 *) Introduce type-safe ASN.1 SETs.
12366 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
12367 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
12369 *) Introduce type-safe STACKs. This will almost certainly break lots of code
12370 that links with OpenSSL (well at least cause lots of warnings), but fear
12371 not: the conversion is trivial, and it eliminates loads of evil casts. A
12372 few STACKed things have been converted already. Feel free to convert more.
12373 In the fullness of time, I'll do away with the STACK type altogether.
12376 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
12377 specified in <certfile> by updating the entry in the index.txt file.
12378 This way one no longer has to edit the index.txt file manually for
12379 revoking a certificate. The -revoke option does the gory details now.
12380 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
12382 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
12383 `-text' option at all and this way the `-noout -text' combination was
12384 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
12385 [Ralf S. Engelschall]
12387 *) Make sure a corresponding plain text error message exists for the
12388 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
12389 verify callback function determined that a certificate was revoked.
12390 [Ralf S. Engelschall]
12392 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
12393 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
12394 all available ciphers including rc5, which was forgotten until now.
12395 In order to let the testing shell script know which algorithms
12396 are available, a new (up to now undocumented) command
12397 "openssl list-cipher-commands" is used.
12400 *) Bugfix: s_client occasionally would sleep in select() when
12401 it should have checked SSL_pending() first.
12404 *) New functions DSA_do_sign and DSA_do_verify to provide access to
12405 the raw DSA values prior to ASN.1 encoding.
12408 *) Tweaks to Configure
12409 [Niels Poppe <niels@netbox.org>]
12411 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
12415 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
12418 *) New config option to avoid instructions that are illegal on the 80386.
12419 The default code is faster, but requires at least a 486.
12422 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
12423 SSL2_SERVER_VERSION (not used at all) macros, which are now the
12424 same as SSL2_VERSION anyway.
12427 *) New "-showcerts" option for s_client.
12430 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
12431 application. Various cleanups and fixes.
12434 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
12435 modify error routines to work internally. Add error codes and PBE init
12436 to library startup routines.
12439 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
12440 packing functions to asn1 and evp. Changed function names and error
12441 codes along the way.
12444 *) PKCS12 integration: and so it begins... First of several patches to
12445 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
12446 objects to objects.h
12449 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
12450 and display support for Thawte strong extranet extension.
12453 *) Add LinuxPPC support.
12454 [Jeff Dubrule <igor@pobox.org>]
12456 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
12457 bn_div_words in alpha.s.
12458 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
12460 *) Make sure the RSA OAEP test is skipped under -DRSAref because
12461 OAEP isn't supported when OpenSSL is built with RSAref.
12462 [Ulf Moeller <ulf@fitug.de>]
12464 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
12465 so they no longer are missing under -DNOPROTO.
12466 [Soren S. Jorvang <soren@t.dk>]
12469 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
12471 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
12472 doesn't work when the session is reused. Coming soon!
12475 *) Fix a security hole, that allows sessions to be reused in the wrong
12476 context thus bypassing client cert protection! All software that uses
12477 client certs and session caches in multiple contexts NEEDS PATCHING to
12478 allow session reuse! A fuller solution is in the works.
12479 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
12481 *) Some more source tree cleanups (removed obsolete files
12482 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
12483 permission on "config" script to be executable) and a fix for the INSTALL
12485 [Ulf Moeller <ulf@fitug.de>]
12487 *) Remove some legacy and erroneous uses of malloc, free instead of
12489 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
12491 *) Make rsa_oaep_test return non-zero on error.
12492 [Ulf Moeller <ulf@fitug.de>]
12494 *) Add support for native Solaris shared libraries. Configure
12495 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12496 if someone would make that last step automatic.
12497 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12499 *) ctx_size was not built with the right compiler during "make links". Fixed.
12502 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12503 except NULL ciphers". This means the default cipher list will no longer
12504 enable NULL ciphers. They need to be specifically enabled e.g. with
12505 the string "DEFAULT:eNULL".
12508 *) Fix to RSA private encryption routines: if p < q then it would
12509 occasionally produce an invalid result. This will only happen with
12510 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12513 *) Be less restrictive and allow also `perl util/perlpath.pl
12514 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12515 because this way one can also use an interpreter named `perl5' (which is
12516 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12517 installed as `perl').
12518 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12520 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12521 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12523 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12524 advapi32.lib to Win32 build and change the pem test comparison
12525 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12526 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
12527 and crypto/des/ede_cbcm_enc.c.
12530 *) DES quad checksum was broken on big-endian architectures. Fixed.
12533 *) Comment out two functions in bio.h that aren't implemented. Fix up the
12534 Win32 test batch file so it (might) work again. The Win32 test batch file
12535 is horrible: I feel ill....
12538 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
12539 in e_os.h. Audit of header files to check ANSI and non ANSI
12540 sections: 10 functions were absent from non ANSI section and not exported
12541 from Windows DLLs. Fixed up libeay.num for new functions.
12544 *) Make `openssl version' output lines consistent.
12545 [Ralf S. Engelschall]
12547 *) Fix Win32 symbol export lists for BIO functions: Added
12548 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
12549 to ms/libeay{16,32}.def.
12550 [Ralf S. Engelschall]
12552 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
12553 fine under Unix and passes some trivial tests I've now added. But the
12554 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
12555 added to make sure no one expects that this stuff really works in the
12556 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
12557 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
12559 [Ralf S. Engelschall]
12561 *) Fix the generation of two part addresses in perl.
12562 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
12564 *) Add config entry for Linux on MIPS.
12565 [John Tobey <jtobey@channel1.com>]
12567 *) Make links whenever Configure is run, unless we are on Windoze.
12570 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
12571 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
12575 *) Add a useful kludge to allow package maintainers to specify compiler and
12576 other platforms details on the command line without having to patch the
12577 Configure script everytime: One now can use ``perl Configure
12578 <id>:<details>'', i.e. platform ids are allowed to have details appended
12579 to them (separated by colons). This is treated as there would be a static
12580 pre-configured entry in Configure's %table under key <id> with value
12581 <details> and ``perl Configure <id>'' is called. So, when you want to
12582 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
12583 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
12584 now, which overrides the FreeBSD-elf entry on-the-fly.
12585 [Ralf S. Engelschall]
12587 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
12590 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
12591 on the `perl Configure ...' command line. This way one can compile
12592 OpenSSL libraries with Position Independent Code (PIC) which is needed
12593 for linking it into DSOs.
12594 [Ralf S. Engelschall]
12596 *) Remarkably, export ciphers were totally broken and no-one had noticed!
12600 *) Cleaned up the LICENSE document: The official contact for any license
12601 questions now is the OpenSSL core team under openssl-core@openssl.org.
12602 And add a paragraph about the dual-license situation to make sure people
12603 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
12604 to the OpenSSL toolkit.
12605 [Ralf S. Engelschall]
12607 *) General source tree makefile cleanups: Made `making xxx in yyy...'
12608 display consistent in the source tree and replaced `/bin/rm' by `rm'.
12609 Additionally cleaned up the `make links' target: Remove unnecessary
12610 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
12611 to speed processing and no longer clutter the display with confusing
12612 stuff. Instead only the actually done links are displayed.
12613 [Ralf S. Engelschall]
12615 *) Permit null encryption ciphersuites, used for authentication only. It used
12616 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
12617 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12621 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12622 signed attributes when verifying signatures (this would break them),
12623 the detached data encoding was wrong and public keys obtained using
12624 X509_get_pubkey() weren't freed.
12627 *) Add text documentation for the BUFFER functions. Also added a work around
12628 to a Win95 console bug. This was triggered by the password read stuff: the
12629 last character typed gets carried over to the next fread(). If you were
12630 generating a new cert request using 'req' for example then the last
12631 character of the passphrase would be CR which would then enter the first
12635 *) Added the new `Includes OpenSSL Cryptography Software' button as
12636 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12637 button and can be used by applications based on OpenSSL to show the
12638 relationship to the OpenSSL project.
12639 [Ralf S. Engelschall]
12641 *) Remove confusing variables in function signatures in files
12642 ssl/ssl_lib.c and ssl/ssl.h.
12643 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12645 *) Don't install bss_file.c under PREFIX/include/
12646 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12648 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12649 functions that return function pointers and has support for NT specific
12650 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12651 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12652 unsigned to signed types: this was killing the Win32 compile.
12655 *) Add new certificate file to stack functions,
12656 SSL_add_dir_cert_subjects_to_stack() and
12657 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12658 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12659 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12660 This means that Apache-SSL and similar packages don't have to mess around
12661 to add as many CAs as they want to the preferred list.
12664 *) Experiment with doxygen documentation. Currently only partially applied to
12666 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12667 openssl.doxy as the configuration file.
12670 *) Get rid of remaining C++-style comments which strict C compilers hate.
12671 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12673 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12674 compiled in by default: it has problems with large keys.
12677 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12678 DH private keys and/or callback functions which directly correspond to
12679 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12680 is needed for applications which have to configure certificates on a
12681 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12683 For the RSA certificate situation is makes no difference, but
12684 for the DSA certificate situation this fixes the "no shared cipher"
12685 problem where the OpenSSL cipher selection procedure failed because the
12686 temporary keys were not overtaken from the context and the API provided
12687 no way to reconfigure them.
12688 The new functions now let applications reconfigure the stuff and they
12689 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12690 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12691 non-public-API function ssl_cert_instantiate() is used as a helper
12692 function and also to reduce code redundancy inside ssl_rsa.c.
12693 [Ralf S. Engelschall]
12695 *) Move s_server -dcert and -dkey options out of the undocumented feature
12696 area because they are useful for the DSA situation and should be
12697 recognized by the users.
12698 [Ralf S. Engelschall]
12700 *) Fix the cipher decision scheme for export ciphers: the export bits are
12701 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12702 SSL_EXP_MASK. So, the original variable has to be used instead of the
12703 already masked variable.
12704 [Richard Levitte <levitte@stacken.kth.se>]
12706 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12707 [Richard Levitte <levitte@stacken.kth.se>]
12709 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12710 from `int' to `unsigned int' because it's a length and initialized by
12711 EVP_DigestFinal() which expects an `unsigned int *'.
12712 [Richard Levitte <levitte@stacken.kth.se>]
12714 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12715 script. Instead use the usual Shell->Perl transition trick.
12716 [Ralf S. Engelschall]
12718 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12719 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12720 -noout -modulus' as it's already the case for `openssl rsa -noout
12721 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12722 currently the public key is printed (a decision which was already done by
12723 `openssl dsa -modulus' in the past) which serves a similar purpose.
12724 Additionally the NO_RSA no longer completely removes the whole -modulus
12725 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12727 [Ralf S. Engelschall]
12729 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12730 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12731 [Arne Ansper <arne@ats.cyber.ee>]
12733 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12734 to be added. Now both 'req' and 'ca' can use new objects defined in the
12738 *) Add cool BIO that does syslog (or event log on NT).
12739 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12741 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12742 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12743 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12744 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12747 *) Add preliminary config info for new extension code.
12750 *) Make RSA_NO_PADDING really use no padding.
12751 [Ulf Moeller <ulf@fitug.de>]
12753 *) Generate errors when private/public key check is done.
12756 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12757 for some CRL extensions and new objects added.
12760 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
12761 key usage extension and fuller support for authority key id.
12764 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
12765 padding method for RSA, which is recommended for new applications in PKCS
12766 #1 v2.0 (RFC 2437, October 1998).
12767 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
12768 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
12769 against Bleichbacher's attack on RSA.
12770 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
12773 *) Updates to the new SSL compression code
12774 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12776 *) Fix so that the version number in the master secret, when passed
12777 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
12778 (because the server will not accept higher), that the version number
12779 is 0x03,0x01, not 0x03,0x00
12780 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12782 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
12783 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
12784 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
12787 *) Support for RAW extensions where an arbitrary extension can be
12788 created by including its DER encoding. See apps/openssl.cnf for
12792 *) Make sure latest Perl versions don't interpret some generated C array
12793 code as Perl array code in the crypto/err/err_genc.pl script.
12794 [Lars Weber <3weber@informatik.uni-hamburg.de>]
12796 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
12797 not many people have the assembler. Various Win32 compilation fixes and
12798 update to the INSTALL.W32 file with (hopefully) more accurate Win32
12799 build instructions.
12802 *) Modify configure script 'Configure' to automatically create crypto/date.h
12803 file under Win32 and also build pem.h from pem.org. New script
12804 util/mkfiles.pl to create the MINFO file on environments that can't do a
12805 'make files': perl util/mkfiles.pl >MINFO should work.
12808 *) Major rework of DES function declarations, in the pursuit of correctness
12809 and purity. As a result, many evil casts evaporated, and some weirdness,
12810 too. You may find this causes warnings in your code. Zapping your evil
12811 casts will probably fix them. Mostly.
12814 *) Fix for a typo in asn1.h. Bug fix to object creation script
12815 obj_dat.pl. It considered a zero in an object definition to mean
12816 "end of object": none of the objects in objects.h have any zeros
12817 so it wasn't spotted.
12818 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
12820 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
12821 Masking (CBCM). In the absence of test vectors, the best I have been able
12822 to do is check that the decrypt undoes the encrypt, so far. Send me test
12823 vectors if you have them.
12826 *) Correct calculation of key length for export ciphers (too much space was
12827 allocated for null ciphers). This has not been tested!
12830 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
12831 message is now correct (it understands "crypto" and "ssl" on its
12832 command line). There is also now an "update" option. This will update
12833 the util/ssleay.num and util/libeay.num files with any new functions.
12835 perl util/mkdef.pl crypto ssl update
12836 it will update them.
12839 *) Overhauled the Perl interface (perl/*):
12840 - ported BN stuff to OpenSSL's different BN library
12841 - made the perl/ source tree CVS-aware
12842 - renamed the package from SSLeay to OpenSSL (the files still contain
12843 their history because I've copied them in the repository)
12844 - removed obsolete files (the test scripts will be replaced
12845 by better Test::Harness variants in the future)
12846 [Ralf S. Engelschall]
12848 *) First cut for a very conservative source tree cleanup:
12849 1. merge various obsolete readme texts into doc/ssleay.txt
12850 where we collect the old documents and readme texts.
12851 2. remove the first part of files where I'm already sure that we no
12852 longer need them because of three reasons: either they are just temporary
12853 files which were left by Eric or they are preserved original files where
12854 I've verified that the diff is also available in the CVS via "cvs diff
12855 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
12856 the crypto/md/ stuff).
12857 [Ralf S. Engelschall]
12859 *) More extension code. Incomplete support for subject and issuer alt
12860 name, issuer and authority key id. Change the i2v function parameters
12861 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
12862 what that's for :-) Fix to ASN1 macro which messed up
12863 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
12866 *) Preliminary support for ENUMERATED type. This is largely copied from the
12870 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
12871 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12873 *) Make sure `make rehash' target really finds the `openssl' program.
12874 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12876 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
12877 like to hear about it if this slows down other processors.
12880 *) Add CygWin32 platform information to Configure script.
12881 [Alan Batie <batie@aahz.jf.intel.com>]
12883 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
12884 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
12886 *) New program nseq to manipulate netscape certificate sequences
12889 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
12893 *) Fixes to BN code. Previously the default was to define BN_RECURSION
12894 but the BN code had some problems that would cause failures when
12895 doing certificate verification and some other functions.
12896 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12898 *) Add ASN1 and PEM code to support netscape certificate sequences.
12901 *) Add ASN1 and PEM code to support netscape certificate sequences.
12904 *) Add several PKIX and private extended key usage OIDs.
12907 *) Modify the 'ca' program to handle the new extension code. Modify
12908 openssl.cnf for new extension format, add comments.
12911 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
12912 and add a sample to openssl.cnf so req -x509 now adds appropriate
12916 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
12917 error code, add initial support to X509_print() and x509 application.
12920 *) Takes a deep breath and start adding X509 V3 extension support code. Add
12921 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
12922 stuff is currently isolated and isn't even compiled yet.
12925 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
12926 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
12927 Removed the versions check from X509 routines when loading extensions:
12928 this allows certain broken certificates that don't set the version
12929 properly to be processed.
12932 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
12933 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
12934 can still be regenerated with "make depend".
12937 *) Spelling mistake in C version of CAST-128.
12938 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
12940 *) Changes to the error generation code. The perl script err-code.pl
12941 now reads in the old error codes and retains the old numbers, only
12942 adding new ones if necessary. It also only changes the .err files if new
12943 codes are added. The makefiles have been modified to only insert errors
12944 when needed (to avoid needlessly modifying header files). This is done
12945 by only inserting errors if the .err file is newer than the auto generated
12946 C file. To rebuild all the error codes from scratch (the old behaviour)
12947 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
12948 or delete all the .err files.
12951 *) CAST-128 was incorrectly implemented for short keys. The C version has
12952 been fixed, but is untested. The assembler versions are also fixed, but
12953 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
12954 to regenerate it if needed.
12955 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
12956 Hagino <itojun@kame.net>]
12958 *) File was opened incorrectly in randfile.c.
12959 [Ulf Möller <ulf@fitug.de>]
12961 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
12962 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
12963 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
12964 al: it's just almost always a UTCTime. Note this patch adds new error
12965 codes so do a "make errors" if there are problems.
12968 *) Correct Linux 1 recognition in config.
12969 [Ulf Möller <ulf@fitug.de>]
12971 *) Remove pointless MD5 hash when using DSA keys in ca.
12972 [Anonymous <nobody@replay.com>]
12974 *) Generate an error if given an empty string as a cert directory. Also
12975 generate an error if handed NULL (previously returned 0 to indicate an
12976 error, but didn't set one).
12977 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
12979 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
12982 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
12983 parameters. This was causing a warning which killed off the Win32 compile.
12986 *) Remove C++ style comments from crypto/bn/bn_local.h.
12987 [Neil Costigan <neil.costigan@celocom.com>]
12989 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
12990 based on a text string, looking up short and long names and finally
12991 "dot" format. The "dot" format stuff didn't work. Added new function
12992 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
12993 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
12994 OID is not part of the table.
12997 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
12998 X509_LOOKUP_by_alias().
13001 *) Sort openssl functions by name.
13004 *) Get the gendsa program working (hopefully) and add it to app list. Remove
13005 encryption from sample DSA keys (in case anyone is interested the password
13009 *) Make _all_ *_free functions accept a NULL pointer.
13010 [Frans Heymans <fheymans@isaserver.be>]
13012 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
13014 [Anonymous <nobody@replay.com>]
13016 *) s_server should send the CAfile as acceptable CAs, not its own cert.
13017 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13019 *) Don't blow it for numeric -newkey arguments to apps/req.
13020 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13022 *) Temp key "for export" tests were wrong in s3_srvr.c.
13023 [Anonymous <nobody@replay.com>]
13025 *) Add prototype for temp key callback functions
13026 SSL_CTX_set_tmp_{rsa,dh}_callback().
13029 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
13030 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
13033 *) X509_name_add_entry() freed the wrong thing after an error.
13034 [Arne Ansper <arne@ats.cyber.ee>]
13036 *) rsa_eay.c would attempt to free a NULL context.
13037 [Arne Ansper <arne@ats.cyber.ee>]
13039 *) BIO_s_socket() had a broken should_retry() on Windoze.
13040 [Arne Ansper <arne@ats.cyber.ee>]
13042 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
13043 [Arne Ansper <arne@ats.cyber.ee>]
13045 *) Make sure the already existing X509_STORE->depth variable is initialized
13046 in X509_STORE_new(), but document the fact that this variable is still
13047 unused in the certificate verification process.
13048 [Ralf S. Engelschall]
13050 *) Fix the various library and apps files to free up pkeys obtained from
13051 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
13054 *) Fix reference counting in X509_PUBKEY_get(). This makes
13055 demos/maurice/example2.c work, amongst others, probably.
13056 [Steve Henson and Ben Laurie]
13058 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
13059 `openssl' and second, the shortcut symlinks for the `openssl <command>'
13060 are no longer created. This way we have a single and consistent command
13061 line interface `openssl <command>', similar to `cvs <command>'.
13062 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
13064 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
13065 BIT STRING wrapper always have zero unused bits.
13068 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
13071 *) Make the top-level INSTALL documentation easier to understand.
13074 *) Makefiles updated to exit if an error occurs in a sub-directory
13075 make (including if user presses ^C) [Paul Sutton]
13077 *) Make Montgomery context stuff explicit in RSA data structure.
13080 *) Fix build order of pem and err to allow for generated pem.h.
13083 *) Fix renumbering bug in X509_NAME_delete_entry().
13086 *) Enhanced the err-ins.pl script so it makes the error library number
13087 global and can add a library name. This is needed for external ASN1 and
13088 other error libraries.
13091 *) Fixed sk_insert which never worked properly.
13094 *) Fix ASN1 macros so they can handle indefinite length constructed
13095 EXPLICIT tags. Some non standard certificates use these: they can now
13099 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
13100 into a single doc/ssleay.txt bundle. This way the information is still
13101 preserved but no longer messes up this directory. Now it's new room for
13102 the new set of documentation files.
13103 [Ralf S. Engelschall]
13105 *) SETs were incorrectly DER encoded. This was a major pain, because they
13106 shared code with SEQUENCEs, which aren't coded the same. This means that
13107 almost everything to do with SETs or SEQUENCEs has either changed name or
13108 number of arguments.
13109 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
13111 *) Fix test data to work with the above.
13114 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
13115 was already fixed by Eric for 0.9.1 it seems.
13116 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
13118 *) Autodetect FreeBSD3.
13121 *) Fix various bugs in Configure. This affects the following platforms:
13125 unixware-2.0-pentium
13129 *) Eliminate generated files from CVS. Reorder tests to regenerate files
13130 before they are needed.
13133 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
13137 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
13139 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
13140 changed SSLeay to OpenSSL in version strings.
13141 [Ralf S. Engelschall]
13143 *) Some fixups to the top-level documents.
13146 *) Fixed the nasty bug where rsaref.h was not found under compile-time
13147 because the symlink to include/ was missing.
13148 [Ralf S. Engelschall]
13150 *) Incorporated the popular no-RSA/DSA-only patches
13151 which allow to compile a RSA-free SSLeay.
13152 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
13154 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
13155 when "ssleay" is still not found.
13156 [Ralf S. Engelschall]
13158 *) Added more platforms to Configure: Cray T3E, HPUX 11,
13159 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
13161 *) Updated the README file.
13162 [Ralf S. Engelschall]
13164 *) Added various .cvsignore files in the CVS repository subdirs
13165 to make a "cvs update" really silent.
13166 [Ralf S. Engelschall]
13168 *) Recompiled the error-definition header files and added
13169 missing symbols to the Win32 linker tables.
13170 [Ralf S. Engelschall]
13172 *) Cleaned up the top-level documents;
13173 o new files: CHANGES and LICENSE
13174 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
13175 o merged COPYRIGHT into LICENSE
13176 o removed obsolete TODO file
13177 o renamed MICROSOFT to INSTALL.W32
13178 [Ralf S. Engelschall]
13180 *) Removed dummy files from the 0.9.1b source tree:
13181 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
13182 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
13183 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
13184 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
13185 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
13186 [Ralf S. Engelschall]
13188 *) Added various platform portability fixes.
13191 *) The Genesis of the OpenSSL rpject:
13192 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
13193 Young and Tim J. Hudson created while they were working for C2Net until
13195 [The OpenSSL Project]
13198 Changes between 0.9.0b and 0.9.1b [not released]
13200 *) Updated a few CA certificates under certs/
13203 *) Changed some BIGNUM api stuff.
13206 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13207 DGUX x86, Linux Alpha, etc.
13210 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13211 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13215 *) Add -strparse option to asn1pars program which parses nested
13217 [Dr Stephen Henson <shenson@bigfoot.com>]
13219 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13222 *) DSA fix for "ca" program.
13225 *) Added "-genkey" option to "dsaparam" program.
13228 *) Added RIPE MD160 (rmd160) message digest.
13231 *) Added -a (all) option to "ssleay version" command.
13234 *) Added PLATFORM define which is the id given to Configure.
13237 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13240 *) Extended the ASN.1 parser routines.
13243 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13246 *) Added a BN_CTX to the BN library.
13249 *) Fixed the weak key values in DES library
13252 *) Changed API in EVP library for cipher aliases.
13255 *) Added support for RC2/64bit cipher.
13258 *) Converted the lhash library to the crypto/mem.c functions.
13261 *) Added more recognized ASN.1 object ids.
13264 *) Added more RSA padding checks for SSL/TLS.
13267 *) Added BIO proxy/filter functionality.
13270 *) Added extra_certs to SSL_CTX which can be used
13271 send extra CA certificates to the client in the CA cert chain sending
13272 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13275 *) Now Fortezza is denied in the authentication phase because
13276 this is key exchange mechanism is not supported by SSLeay at all.
13279 *) Additional PKCS1 checks.
13282 *) Support the string "TLSv1" for all TLS v1 ciphers.
13285 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13286 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13289 *) Fixed a few memory leaks.
13292 *) Fixed various code and comment typos.
13295 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13296 bytes sent in the client random.
13297 [Edward Bishop <ebishop@spyglass.com>]