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 1.1.1 and 3.0.0 [xx XXX xxxx]
12 *) Added a new FUNCerr() macro that takes a function name.
13 The macro SYSerr() is deprecated.
16 *) {CRYPTO,OPENSSL}_mem_debug_{push,pop} are now no-ops and have been
20 *) A new type, EVP_KEYEXCH, has been introduced to represent key exchange
21 algorithms. An implementation of a key exchange algorithm can be obtained
22 by using the function EVP_KEYEXCH_fetch(). An EVP_KEYEXCH algorithm can be
23 used in a call to EVP_PKEY_derive_init_ex() which works in a similar way to
24 the older EVP_PKEY_derive_init() function. See the man pages for the new
25 functions for further details.
28 *) The EVP_PKEY_CTX_set_dh_pad() macro has now been converted to a function.
31 *) Removed the function names from error messages and deprecated the
34 *) Removed NextStep support and the macro OPENSSL_UNISTD
37 *) Removed DES_check_key. Also removed OPENSSL_IMPLEMENT_GLOBAL,
38 OPENSSL_GLOBAL_REF, OPENSSL_DECLARE_GLOBAL.
39 Also removed "export var as function" capability; we do not export
40 variables, only functions.
43 *) RC5_32_set_key has been changed to return an int type, with 0 indicating
44 an error and 1 indicating success. In previous versions of OpenSSL this
45 was a void type. If a key was set longer than the maximum possible this
49 *) Support SM2 signing and verification schemes with X509 certificate.
52 *) Use SHA256 as the default digest for TS query in the ts app.
55 *) Change PBKDF2 to conform to SP800-132 instead of the older PKCS5 RFC2898.
56 This checks that the salt length is at least 128 bits, the derived key
57 length is at least 112 bits, and that the iteration count is at least 1000.
58 For backwards compatibility these checks are disabled by default in the
59 default provider, but are enabled by default in the fips provider.
60 To enable or disable these checks use the control
61 EVP_KDF_CTRL_SET_PBKDF2_PKCS5_MODE.
64 *) Default cipher lists/suites are now available via a function, the
65 #defines are deprecated.
68 *) Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
69 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
70 for Windows Store apps easier. Also, the "no-uplink" option has been added.
73 *) Join the directories crypto/x509 and crypto/x509v3
76 *) Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
77 This changes the size when using the genpkey app when no size is given. It
78 fixes an omission in earlier changes that changed all RSA, DSA and DH
79 generation apps to use 2048 bits by default.
82 *) Added command 'openssl kdf' that uses the EVP_KDF API.
85 *) Added command 'openssl mac' that uses the EVP_MAC API.
88 *) Added OPENSSL_info() to get diverse built-in OpenSSL data, such
89 as default directories. Also added the command 'openssl info'
90 for scripting purposes.
93 *) The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
94 deprecated. These undocumented functions were never integrated into the EVP
95 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
96 Bi-directional IGE mode. These modes were never formally standardised and
97 usage of these functions is believed to be very small. In particular
98 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
99 is ever used. The security implications are believed to be minimal, but
100 this issue was never fixed for backwards compatibility reasons. New code
101 should not use these modes.
104 *) Add prediction resistance to the DRBG reseeding process.
107 *) Limit the number of blocks in a data unit for AES-XTS to 2^20 as
108 mandated by IEEE Std 1619-2018.
111 *) Added newline escaping functionality to a filename when using openssl dgst.
112 This output format is to replicate the output format found in the '*sum'
113 checksum programs. This aims to preserve backward compatibility.
114 [Matt Eaton, Richard Levitte, and Paul Dale]
116 *) Removed the heartbeat message in DTLS feature, as it has very
117 little usage and doesn't seem to fulfill a valuable purpose.
118 The configuration option is now deprecated.
121 *) Changed the output of 'openssl {digestname} < file' to display the
122 digest name in its output.
125 *) Added a new generic trace API which provides support for enabling
126 instrumentation through trace output. This feature is mainly intended
127 as an aid for developers and is disabled by default. To utilize it,
128 OpenSSL needs to be configured with the `enable-trace` option.
130 If the tracing API is enabled, the application can activate trace output
131 by registering BIOs as trace channels for a number of tracing and debugging
134 The 'openssl' application has been expanded to enable any of the types
135 available via environment variables defined by the user, and serves as
136 one possible example on how to use this functionality.
137 [Richard Levitte & Matthias St. Pierre]
139 *) Added build tests for C++. These are generated files that only do one
140 thing, to include one public OpenSSL head file each. This tests that
141 the public header files can be usefully included in a C++ application.
143 This test isn't enabled by default. It can be enabled with the option
144 'enable-buildtest-c++'.
147 *) Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
150 *) Add KMAC to EVP_MAC.
153 *) Added property based algorithm implementation selection framework to
157 *) Added SCA hardening for modular field inversion in EC_GROUP through
158 a new dedicated field_inv() pointer in EC_METHOD.
159 This also addresses a leakage affecting conversions from projective
160 to affine coordinates.
161 [Billy Bob Brumley, Nicola Tuveri]
163 *) Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
164 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
165 those algorithms that were already supported through the EVP_PKEY API
166 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
167 and scrypt are now wrappers that call EVP_KDF.
170 *) Build devcrypto engine as a dynamic engine.
173 *) Add keyed BLAKE2 to EVP_MAC.
176 *) Fix a bug in the computation of the endpoint-pair shared secret used
177 by DTLS over SCTP. This breaks interoperability with older versions
178 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
179 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
180 interoperability with such broken implementations. However, enabling
181 this switch breaks interoperability with correct implementations.
183 *) Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
184 re-used X509_PUBKEY object if the second PUBKEY is malformed.
187 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
190 *) Change the license to the Apache License v2.0.
193 *) Change the possible version information given with OPENSSL_API_COMPAT.
194 It may be a pre-3.0.0 style numerical version number as it was defined
195 in 1.1.0, and it may also simply take the major version number.
197 Because of the version numbering of pre-3.0.0 releases, the values 0,
198 1 and 2 are equivalent to 0x00908000L (0.9.8), 0x10000000L (1.0.0) and
199 0x10100000L (1.1.0), respectively.
202 *) Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
204 o Major releases (indicated by incrementing the MAJOR release number)
205 may introduce incompatible API/ABI changes.
206 o Minor releases (indicated by incrementing the MINOR release number)
207 may introduce new features but retain API/ABI compatibility.
208 o Patch releases (indicated by incrementing the PATCH number)
209 are intended for bug fixes and other improvements of existing
210 features only (like improving performance or adding documentation)
211 and retain API/ABI compatibility.
214 *) Add support for RFC5297 SIV mode (siv128), including AES-SIV.
217 *) Remove the 'dist' target and add a tarball building script. The
218 'dist' target has fallen out of use, and it shouldn't be
219 necessary to configure just to create a source distribution.
222 *) Recreate the OS390-Unix config target. It no longer relies on a
223 special script like it did for OpenSSL pre-1.1.0.
226 *) Instead of having the source directories listed in Configure, add
227 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
231 *) Add GMAC to EVP_MAC.
234 *) Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
237 *) Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
238 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
239 to facilitate the continued use of MACs through raw private keys in
240 functionality such as EVP_DigestSign* and EVP_DigestVerify*.
243 *) Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
244 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
247 *) Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
248 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
249 are retained for backwards compatibility.
252 *) AES-XTS mode now enforces that its two keys are different to mitigate
253 the attacked described in "Efficient Instantiations of Tweakable
254 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
255 Details of this attack can be obtained from:
256 http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf
259 *) Rename the object files, i.e. give them other names than in previous
260 versions. Their names now include the name of the final product, as
261 well as its type mnemonic (bin, lib, shlib).
264 *) Added new option for 'openssl list', '-objects', which will display the
265 list of built in objects, i.e. OIDs with names.
268 *) Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
269 improves application performance by removing data copies and providing
270 applications with zero-copy system calls such as sendfile and splice.
273 Changes between 1.1.1a and 1.1.1b [xx XXX xxxx]
275 *) Change the info callback signals for the start and end of a post-handshake
276 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
277 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
278 confused by this and assume that a TLSv1.2 renegotiation has started. This
279 can break KeyUpdate handling. Instead we no longer signal the start and end
280 of a post handshake message exchange (although the messages themselves are
281 still signalled). This could break some applications that were expecting
282 the old signals. However without this KeyUpdate is not usable for many
286 Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
288 *) Timing vulnerability in DSA signature generation
290 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
291 timing side channel attack. An attacker could use variations in the signing
292 algorithm to recover the private key.
294 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
298 *) Timing vulnerability in ECDSA signature generation
300 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
301 timing side channel attack. An attacker could use variations in the signing
302 algorithm to recover the private key.
304 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
308 *) Fixed the issue that RAND_add()/RAND_seed() silently discards random input
309 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
310 of two gigabytes and the error handling improved.
312 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
313 categorized as a normal bug, not a security issue, because the DRBG reseeds
314 automatically and is fully functional even without additional randomness
315 provided by the application.
317 Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
319 *) Add a new ClientHello callback. Provides a callback interface that gives
320 the application the ability to adjust the nascent SSL object at the
321 earliest stage of ClientHello processing, immediately after extensions have
322 been collected but before they have been processed. In particular, this
323 callback can adjust the supported TLS versions in response to the contents
327 *) Add SM2 base algorithm support.
330 *) s390x assembly pack: add (improved) hardware-support for the following
331 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
332 aes-cfb/cfb8, aes-ecb.
335 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
336 parameter is no longer accepted, as it leads to a corrupt table. NULL
337 pem_str is reserved for alias entries only.
340 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
341 step for prime curves. The new implementation is based on formulae from
342 differential addition-and-doubling in homogeneous projective coordinates
343 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
344 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
345 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
346 to work in projective coordinates.
347 [Billy Bob Brumley, Nicola Tuveri]
349 *) Change generating and checking of primes so that the error rate of not
350 being prime depends on the intended use based on the size of the input.
351 For larger primes this will result in more rounds of Miller-Rabin.
352 The maximal error rate for primes with more than 1080 bits is lowered
354 [Kurt Roeckx, Annie Yousar]
356 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
359 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
360 moving between systems, and to avoid confusion when a Windows build is
361 done with mingw vs with MSVC. For POSIX installs, there's still a
362 symlink or copy named 'tsget' to avoid that confusion as well.
365 *) Revert blinding in ECDSA sign and instead make problematic addition
366 length-invariant. Switch even to fixed-length Montgomery multiplication.
369 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
370 step for binary curves. The new implementation is based on formulae from
371 differential addition-and-doubling in mixed Lopez-Dahab projective
372 coordinates, modified to independently blind the operands.
373 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
375 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
376 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
377 EC_METHODs to implement their own specialized "ladder step", to take
378 advantage of more favorable coordinate systems or more efficient
379 differential addition-and-doubling algorithms.
380 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
382 *) Modified the random device based seed sources to keep the relevant
383 file descriptors open rather than reopening them on each access.
384 This allows such sources to operate in a chroot() jail without
385 the associated device nodes being available. This behaviour can be
386 controlled using RAND_keep_random_devices_open().
389 *) Numerous side-channel attack mitigations have been applied. This may have
390 performance impacts for some algorithms for the benefit of improved
391 security. Specific changes are noted in this change log by their respective
395 *) AIX shared library support overhaul. Switch to AIX "natural" way of
396 handling shared libraries, which means collecting shared objects of
397 different versions and bitnesses in one common archive. This allows to
398 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
399 doesn't affect the way 3rd party applications are linked, only how
400 multi-version installation is managed.
403 *) Make ec_group_do_inverse_ord() more robust and available to other
404 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
405 mitigations are applied to the fallback BN_mod_inverse().
406 When using this function rather than BN_mod_inverse() directly, new
407 EC cryptosystem implementations are then safer-by-default.
410 *) Add coordinate blinding for EC_POINT and implement projective
411 coordinate blinding for generic prime curves as a countermeasure to
412 chosen point SCA attacks.
413 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
415 *) Add blinding to ECDSA and DSA signatures to protect against side channel
416 attacks discovered by Keegan Ryan (NCC Group).
419 *) Enforce checking in the pkeyutl command line app to ensure that the input
420 length does not exceed the maximum supported digest length when performing
421 a sign, verify or verifyrecover operation.
424 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
425 I/O in combination with something like select() or poll() will hang. This
426 can be turned off again using SSL_CTX_clear_mode().
427 Many applications do not properly handle non-application data records, and
428 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
429 around the problems in those applications, but can also break some.
430 It's recommended to read the manpages about SSL_read(), SSL_write(),
431 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
432 SSL_CTX_set_read_ahead() again.
435 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
436 now allow empty (zero character) pass phrases.
439 *) Apply blinding to binary field modular inversion and remove patent
440 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
443 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
444 binary and prime elliptic curves.
447 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
448 constant time fixed point multiplication.
451 *) Revise elliptic curve scalar multiplication with timing attack
452 defenses: ec_wNAF_mul redirects to a constant time implementation
453 when computing fixed point and variable point multiplication (which
454 in OpenSSL are mostly used with secret scalars in keygen, sign,
455 ECDH derive operations).
456 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida García,
459 *) Updated CONTRIBUTING
462 *) Updated DRBG / RAND to request nonce and additional low entropy
463 randomness from the system.
464 [Matthias St. Pierre]
466 *) Updated 'openssl rehash' to use OpenSSL consistent default.
469 *) Moved the load of the ssl_conf module to libcrypto, which helps
470 loading engines that libssl uses before libssl is initialised.
473 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
476 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
477 [Ingo Schwarze, Rich Salz]
479 *) Added output of accepting IP address and port for 'openssl s_server'
482 *) Added a new API for TLSv1.3 ciphersuites:
483 SSL_CTX_set_ciphersuites()
484 SSL_set_ciphersuites()
487 *) Memory allocation failures consistently add an error to the error
491 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
492 in libcrypto when run as setuid/setgid.
495 *) Load any config file by default when libssl is used.
498 *) Added new public header file <openssl/rand_drbg.h> and documentation
499 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
500 [Matthias St. Pierre]
502 *) QNX support removed (cannot find contributors to get their approval
503 for the license change).
506 *) TLSv1.3 replay protection for early data has been implemented. See the
507 SSL_read_early_data() man page for further details.
510 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
511 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
512 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
513 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
514 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
515 configuration has been separated out. See the ciphers man page or the
516 SSL_CTX_set_ciphersuites() man page for more information.
519 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
520 in responder mode now supports the new "-multi" option, which
521 spawns the specified number of child processes to handle OCSP
522 requests. The "-timeout" option now also limits the OCSP
523 responder's patience to wait to receive the full client request
524 on a newly accepted connection. Child processes are respawned
525 as needed, and the CA index file is automatically reloaded
526 when changed. This makes it possible to run the "ocsp" responder
527 as a long-running service, making the OpenSSL CA somewhat more
528 feature-complete. In this mode, most diagnostic messages logged
529 after entering the event loop are logged via syslog(3) rather than
533 *) Added support for X448 and Ed448. Heavily based on original work by
537 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
538 objects loaded. This adds the functions OSSL_STORE_expect() and
539 OSSL_STORE_find() as well as needed tools to construct searches and
540 get the search data out of them.
543 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
544 version of OpenSSL should review their configuration settings to ensure
545 that they are still appropriate for TLSv1.3. For further information see:
546 https://wiki.openssl.org/index.php/TLS1.3
549 *) Grand redesign of the OpenSSL random generator
551 The default RAND method now utilizes an AES-CTR DRBG according to
552 NIST standard SP 800-90Ar1. The new random generator is essentially
553 a port of the default random generator from the OpenSSL FIPS 2.0
554 object module. It is a hybrid deterministic random bit generator
555 using an AES-CTR bit stream and which seeds and reseeds itself
556 automatically using trusted system entropy sources.
558 Some of its new features are:
559 o Support for multiple DRBG instances with seed chaining.
560 o The default RAND method makes use of a DRBG.
561 o There is a public and private DRBG instance.
562 o The DRBG instances are fork-safe.
563 o Keep all global DRBG instances on the secure heap if it is enabled.
564 o The public and private DRBG instance are per thread for lock free
566 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
568 *) Changed Configure so it only says what it does and doesn't dump
569 so much data. Instead, ./configdata.pm should be used as a script
570 to display all sorts of configuration data.
573 *) Added processing of "make variables" to Configure.
576 *) Added SHA512/224 and SHA512/256 algorithm support.
579 *) The last traces of Netware support, first removed in 1.1.0, have
583 *) Get rid of Makefile.shared, and in the process, make the processing
584 of certain files (rc.obj, or the .def/.map/.opt files produced from
585 the ordinal files) more visible and hopefully easier to trace and
586 debug (or make silent).
589 *) Make it possible to have environment variable assignments as
590 arguments to config / Configure.
593 *) Add multi-prime RSA (RFC 8017) support.
596 *) Add SM3 implemented according to GB/T 32905-2016
597 [ Jack Lloyd <jack.lloyd@ribose.com>,
598 Ronald Tse <ronald.tse@ribose.com>,
599 Erick Borsboom <erick.borsboom@ribose.com> ]
601 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
602 as documented in RFC6066.
603 Based on a patch from Tomasz Moń
604 [Filipe Raimundo da Silva]
606 *) Add SM4 implemented according to GB/T 32907-2016.
607 [ Jack Lloyd <jack.lloyd@ribose.com>,
608 Ronald Tse <ronald.tse@ribose.com>,
609 Erick Borsboom <erick.borsboom@ribose.com> ]
611 *) Reimplement -newreq-nodes and ERR_error_string_n; the
612 original author does not agree with the license change.
615 *) Add ARIA AEAD TLS support.
618 *) Some macro definitions to support VS6 have been removed. Visual
619 Studio 6 has not worked since 1.1.0
622 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
623 without clearing the errors.
626 *) Add "atfork" functions. If building on a system that without
627 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
628 requirements. The RAND facility now uses/requires this.
634 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
635 not possible to disable entirely. However, it's still possible to
636 disable the console reading UI method, UI_OpenSSL() (use UI_null()
639 To disable, configure with 'no-ui-console'. 'no-ui' is still
640 possible to use as an alias. Check at compile time with the
641 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
642 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
645 *) Add a STORE module, which implements a uniform and URI based reader of
646 stores that can contain keys, certificates, CRLs and numerous other
647 objects. The main API is loosely based on a few stdio functions,
648 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
649 OSSL_STORE_error and OSSL_STORE_close.
650 The implementation uses backends called "loaders" to implement arbitrary
651 URI schemes. There is one built in "loader" for the 'file' scheme.
654 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
655 then adjusted to work on FreeBSD 8.4 as well.
656 Enable by configuring with 'enable-devcryptoeng'. This is done by default
657 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
660 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
661 util/mkerr.pl, which is adapted to allow those prefixes, leading to
662 error code calls like this:
664 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
666 With this change, we claim the namespaces OSSL and OPENSSL in a manner
667 that can be encoded in C. For the foreseeable future, this will only
669 [Richard Levitte and Tim Hudson]
671 *) Removed BSD cryptodev engine.
674 *) Add a build target 'build_all_generated', to build all generated files
675 and only that. This can be used to prepare everything that requires
676 things like perl for a system that lacks perl and then move everything
677 to that system and do the rest of the build there.
680 *) In the UI interface, make it possible to duplicate the user data. This
681 can be used by engines that need to retain the data for a longer time
682 than just the call where this user data is passed.
685 *) Ignore the '-named_curve auto' value for compatibility of applications
687 [Tomas Mraz <tmraz@fedoraproject.org>]
689 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
690 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
691 alerts across multiple records (some of which could be empty). In practice
692 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
693 prohibts this altogether and other libraries (BoringSSL, NSS) do not
694 support this at all. Supporting it adds significant complexity to the
695 record layer, and its removal is unlikely to cause inter-operability
699 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
700 with Z. These are meant to replace LONG and ZLONG and to be size safe.
701 The use of LONG and ZLONG is discouraged and scheduled for deprecation
705 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
706 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
707 [Richard Levitte, Andy Polyakov]
709 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
713 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
714 platform rather than 'mingw'.
717 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
718 success if they are asked to add an object which already exists
719 in the store. This change cascades to other functions which load
720 certificates and CRLs.
723 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
724 facilitate stack unwinding even from assembly subroutines.
727 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
728 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
731 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
732 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
733 which is the minimum version we support.
736 *) Certificate time validation (X509_cmp_time) enforces stricter
737 compliance with RFC 5280. Fractional seconds and timezone offsets
738 are no longer allowed.
741 *) Add support for ARIA
744 *) s_client will now send the Server Name Indication (SNI) extension by
745 default unless the new "-noservername" option is used. The server name is
746 based on the host provided to the "-connect" option unless overridden by
750 *) Add support for SipHash
753 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
754 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
755 prevent issues where no progress is being made and the peer continually
756 sends unrecognised record types, using up resources processing them.
759 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
760 using the algorithm defined in
761 https://www.akkadia.org/drepper/SHA-crypt.txt
764 *) Heartbeat support has been removed; the ABI is changed for now.
765 [Richard Levitte, Rich Salz]
767 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
770 *) The RSA "null" method, which was partially supported to avoid patent
771 issues, has been replaced to always returns NULL.
775 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
777 *) Client DoS due to large DH parameter
779 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
780 malicious server can send a very large prime value to the client. This will
781 cause the client to spend an unreasonably long period of time generating a
782 key for this prime resulting in a hang until the client has finished. This
783 could be exploited in a Denial Of Service attack.
785 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
789 *) Cache timing vulnerability in RSA Key Generation
791 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
792 a cache timing side channel attack. An attacker with sufficient access to
793 mount cache timing attacks during the RSA key generation process could
794 recover the private key.
796 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
797 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
801 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
802 parameter is no longer accepted, as it leads to a corrupt table. NULL
803 pem_str is reserved for alias entries only.
806 *) Revert blinding in ECDSA sign and instead make problematic addition
807 length-invariant. Switch even to fixed-length Montgomery multiplication.
810 *) Change generating and checking of primes so that the error rate of not
811 being prime depends on the intended use based on the size of the input.
812 For larger primes this will result in more rounds of Miller-Rabin.
813 The maximal error rate for primes with more than 1080 bits is lowered
815 [Kurt Roeckx, Annie Yousar]
817 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
820 *) Add blinding to ECDSA and DSA signatures to protect against side channel
821 attacks discovered by Keegan Ryan (NCC Group).
824 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
825 now allow empty (zero character) pass phrases.
828 *) Certificate time validation (X509_cmp_time) enforces stricter
829 compliance with RFC 5280. Fractional seconds and timezone offsets
830 are no longer allowed.
833 *) Fixed a text canonicalisation bug in CMS
835 Where a CMS detached signature is used with text content the text goes
836 through a canonicalisation process first prior to signing or verifying a
837 signature. This process strips trailing space at the end of lines, converts
838 line terminators to CRLF and removes additional trailing line terminators
839 at the end of a file. A bug in the canonicalisation process meant that
840 some characters, such as form-feed, were incorrectly treated as whitespace
841 and removed. This is contrary to the specification (RFC5485). This fix
842 could mean that detached text data signed with an earlier version of
843 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
844 signed with a fixed OpenSSL may fail to verify with an earlier version of
845 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
846 and use the "-binary" flag (for the "cms" command line application) or set
847 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
850 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
852 *) Constructed ASN.1 types with a recursive definition could exceed the stack
854 Constructed ASN.1 types with a recursive definition (such as can be found
855 in PKCS7) could eventually exceed the stack given malicious input with
856 excessive recursion. This could result in a Denial Of Service attack. There
857 are no such structures used within SSL/TLS that come from untrusted sources
858 so this is considered safe.
860 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
865 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
867 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
868 effectively reduced to only comparing the least significant bit of each
869 byte. This allows an attacker to forge messages that would be considered as
870 authenticated in an amount of tries lower than that guaranteed by the
871 security claims of the scheme. The module can only be compiled by the
872 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
874 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
879 *) Add a build target 'build_all_generated', to build all generated files
880 and only that. This can be used to prepare everything that requires
881 things like perl for a system that lacks perl and then move everything
882 to that system and do the rest of the build there.
885 *) Backport SSL_OP_NO_RENGOTIATION
887 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
888 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
889 changes this is no longer possible in 1.1.0. Therefore the new
890 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
891 1.1.0 to provide equivalent functionality.
893 Note that if an application built against 1.1.0h headers (or above) is run
894 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
895 accepted but nothing will happen, i.e. renegotiation will not be prevented.
898 *) Removed the OS390-Unix config target. It relied on a script that doesn't
902 *) rsaz_1024_mul_avx2 overflow bug on x86_64
904 There is an overflow bug in the AVX2 Montgomery multiplication procedure
905 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
906 Analysis suggests that attacks against RSA and DSA as a result of this
907 defect would be very difficult to perform and are not believed likely.
908 Attacks against DH1024 are considered just feasible, because most of the
909 work necessary to deduce information about a private key may be performed
910 offline. The amount of resources required for such an attack would be
911 significant. However, for an attack on TLS to be meaningful, the server
912 would have to share the DH1024 private key among multiple clients, which is
913 no longer an option since CVE-2016-0701.
915 This only affects processors that support the AVX2 but not ADX extensions
916 like Intel Haswell (4th generation).
918 This issue was reported to OpenSSL by David Benjamin (Google). The issue
919 was originally found via the OSS-Fuzz project.
923 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
925 *) bn_sqrx8x_internal carry bug on x86_64
927 There is a carry propagating bug in the x86_64 Montgomery squaring
928 procedure. No EC algorithms are affected. Analysis suggests that attacks
929 against RSA and DSA as a result of this defect would be very difficult to
930 perform and are not believed likely. Attacks against DH are considered just
931 feasible (although very difficult) because most of the work necessary to
932 deduce information about a private key may be performed offline. The amount
933 of resources required for such an attack would be very significant and
934 likely only accessible to a limited number of attackers. An attacker would
935 additionally need online access to an unpatched system using the target
936 private key in a scenario with persistent DH parameters and a private
937 key that is shared between multiple clients.
939 This only affects processors that support the BMI1, BMI2 and ADX extensions
940 like Intel Broadwell (5th generation) and later or AMD Ryzen.
942 This issue was reported to OpenSSL by the OSS-Fuzz project.
946 *) Malformed X.509 IPAddressFamily could cause OOB read
948 If an X.509 certificate has a malformed IPAddressFamily extension,
949 OpenSSL could do a one-byte buffer overread. The most likely result
950 would be an erroneous display of the certificate in text format.
952 This issue was reported to OpenSSL by the OSS-Fuzz project.
956 Changes between 1.1.0e and 1.1.0f [25 May 2017]
958 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
959 platform rather than 'mingw'.
962 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
963 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
964 which is the minimum version we support.
967 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
969 *) Encrypt-Then-Mac renegotiation crash
971 During a renegotiation handshake if the Encrypt-Then-Mac extension is
972 negotiated where it was not in the original handshake (or vice-versa) then
973 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
974 and servers are affected.
976 This issue was reported to OpenSSL by Joe Orton (Red Hat).
980 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
982 *) Truncated packet could crash via OOB read
984 If one side of an SSL/TLS path is running on a 32-bit host and a specific
985 cipher is being used, then a truncated packet can cause that host to
986 perform an out-of-bounds read, usually resulting in a crash.
988 This issue was reported to OpenSSL by Robert Święcki of Google.
992 *) Bad (EC)DHE parameters cause a client crash
994 If a malicious server supplies bad parameters for a DHE or ECDHE key
995 exchange then this can result in the client attempting to dereference a
996 NULL pointer leading to a client crash. This could be exploited in a Denial
999 This issue was reported to OpenSSL by Guido Vranken.
1003 *) BN_mod_exp may produce incorrect results on x86_64
1005 There is a carry propagating bug in the x86_64 Montgomery squaring
1006 procedure. No EC algorithms are affected. Analysis suggests that attacks
1007 against RSA and DSA as a result of this defect would be very difficult to
1008 perform and are not believed likely. Attacks against DH are considered just
1009 feasible (although very difficult) because most of the work necessary to
1010 deduce information about a private key may be performed offline. The amount
1011 of resources required for such an attack would be very significant and
1012 likely only accessible to a limited number of attackers. An attacker would
1013 additionally need online access to an unpatched system using the target
1014 private key in a scenario with persistent DH parameters and a private
1015 key that is shared between multiple clients. For example this can occur by
1016 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
1017 similar to CVE-2015-3193 but must be treated as a separate problem.
1019 This issue was reported to OpenSSL by the OSS-Fuzz project.
1023 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
1025 *) ChaCha20/Poly1305 heap-buffer-overflow
1027 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
1028 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
1029 crash. This issue is not considered to be exploitable beyond a DoS.
1031 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
1035 *) CMS Null dereference
1037 Applications parsing invalid CMS structures can crash with a NULL pointer
1038 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
1039 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
1040 structure callback if an attempt is made to free certain invalid encodings.
1041 Only CHOICE structures using a callback which do not handle NULL value are
1044 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
1048 *) Montgomery multiplication may produce incorrect results
1050 There is a carry propagating bug in the Broadwell-specific Montgomery
1051 multiplication procedure that handles input lengths divisible by, but
1052 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
1053 and DH private keys are impossible. This is because the subroutine in
1054 question is not used in operations with the private key itself and an input
1055 of the attacker's direct choice. Otherwise the bug can manifest itself as
1056 transient authentication and key negotiation failures or reproducible
1057 erroneous outcome of public-key operations with specially crafted input.
1058 Among EC algorithms only Brainpool P-512 curves are affected and one
1059 presumably can attack ECDH key negotiation. Impact was not analyzed in
1060 detail, because pre-requisites for attack are considered unlikely. Namely
1061 multiple clients have to choose the curve in question and the server has to
1062 share the private key among them, neither of which is default behaviour.
1063 Even then only clients that chose the curve will be affected.
1065 This issue was publicly reported as transient failures and was not
1066 initially recognized as a security issue. Thanks to Richard Morgan for
1067 providing reproducible case.
1071 *) Removed automatic addition of RPATH in shared libraries and executables,
1072 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
1075 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
1077 *) Fix Use After Free for large message sizes
1079 The patch applied to address CVE-2016-6307 resulted in an issue where if a
1080 message larger than approx 16k is received then the underlying buffer to
1081 store the incoming message is reallocated and moved. Unfortunately a
1082 dangling pointer to the old location is left which results in an attempt to
1083 write to the previously freed location. This is likely to result in a
1084 crash, however it could potentially lead to execution of arbitrary code.
1086 This issue only affects OpenSSL 1.1.0a.
1088 This issue was reported to OpenSSL by Robert Święcki.
1092 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
1094 *) OCSP Status Request extension unbounded memory growth
1096 A malicious client can send an excessively large OCSP Status Request
1097 extension. If that client continually requests renegotiation, sending a
1098 large OCSP Status Request extension each time, then there will be unbounded
1099 memory growth on the server. This will eventually lead to a Denial Of
1100 Service attack through memory exhaustion. Servers with a default
1101 configuration are vulnerable even if they do not support OCSP. Builds using
1102 the "no-ocsp" build time option are not affected.
1104 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1108 *) SSL_peek() hang on empty record
1110 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
1111 sends an empty record. This could be exploited by a malicious peer in a
1112 Denial Of Service attack.
1114 This issue was reported to OpenSSL by Alex Gaynor.
1118 *) Excessive allocation of memory in tls_get_message_header() and
1119 dtls1_preprocess_fragment()
1121 A (D)TLS message includes 3 bytes for its length in the header for the
1122 message. This would allow for messages up to 16Mb in length. Messages of
1123 this length are excessive and OpenSSL includes a check to ensure that a
1124 peer is sending reasonably sized messages in order to avoid too much memory
1125 being consumed to service a connection. A flaw in the logic of version
1126 1.1.0 means that memory for the message is allocated too early, prior to
1127 the excessive message length check. Due to way memory is allocated in
1128 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
1129 to service a connection. This could lead to a Denial of Service through
1130 memory exhaustion. However, the excessive message length check still takes
1131 place, and this would cause the connection to immediately fail. Assuming
1132 that the application calls SSL_free() on the failed connection in a timely
1133 manner then the 21Mb of allocated memory will then be immediately freed
1134 again. Therefore the excessive memory allocation will be transitory in
1135 nature. This then means that there is only a security impact if:
1137 1) The application does not call SSL_free() in a timely manner in the event
1138 that the connection fails
1140 2) The application is working in a constrained environment where there is
1141 very little free memory
1143 3) The attacker initiates multiple connection attempts such that there are
1144 multiple connections in a state where memory has been allocated for the
1145 connection; SSL_free() has not yet been called; and there is insufficient
1146 memory to service the multiple requests.
1148 Except in the instance of (1) above any Denial Of Service is likely to be
1149 transitory because as soon as the connection fails the memory is
1150 subsequently freed again in the SSL_free() call. However there is an
1151 increased risk during this period of application crashes due to the lack of
1152 memory - which would then mean a more serious Denial of Service.
1154 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1155 (CVE-2016-6307 and CVE-2016-6308)
1158 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
1159 had to be removed. Primary reason is that vendor assembler can't
1160 assemble our modules with -KPIC flag. As result it, assembly
1161 support, was not even available as option. But its lack means
1162 lack of side-channel resistant code, which is incompatible with
1163 security by todays standards. Fortunately gcc is readily available
1164 prepackaged option, which we firmly point at...
1167 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
1169 *) Windows command-line tool supports UTF-8 opt-in option for arguments
1170 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
1171 (to any value) allows Windows user to access PKCS#12 file generated
1172 with Windows CryptoAPI and protected with non-ASCII password, as well
1173 as files generated under UTF-8 locale on Linux also protected with
1177 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
1178 have been disabled by default and removed from DEFAULT, just like RC4.
1179 See the RC4 item below to re-enable both.
1182 *) The method for finding the storage location for the Windows RAND seed file
1183 has changed. First we check %RANDFILE%. If that is not set then we check
1184 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
1185 all else fails we fall back to C:\.
1188 *) The EVP_EncryptUpdate() function has had its return type changed from void
1189 to int. A return of 0 indicates and error while a return of 1 indicates
1193 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
1194 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
1195 off the constant time implementation for RSA, DSA and DH have been made
1196 no-ops and deprecated.
1199 *) Windows RAND implementation was simplified to only get entropy by
1200 calling CryptGenRandom(). Various other RAND-related tickets
1202 [Joseph Wylie Yandle, Rich Salz]
1204 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
1205 and OPENSSL_LH_, respectively. The old names are available
1206 with API compatibility. They new names are now completely documented.
1209 *) Unify TYPE_up_ref(obj) methods signature.
1210 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
1211 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
1212 int (instead of void) like all others TYPE_up_ref() methods.
1213 So now these methods also check the return value of CRYPTO_atomic_add(),
1214 and the validity of object reference counter.
1215 [fdasilvayy@gmail.com]
1217 *) With Windows Visual Studio builds, the .pdb files are installed
1218 alongside the installed libraries and executables. For a static
1219 library installation, ossl_static.pdb is the associate compiler
1220 generated .pdb file to be used when linking programs.
1223 *) Remove openssl.spec. Packaging files belong with the packagers.
1226 *) Automatic Darwin/OSX configuration has had a refresh, it will now
1227 recognise x86_64 architectures automatically. You can still decide
1228 to build for a different bitness with the environment variable
1229 KERNEL_BITS (can be 32 or 64), for example:
1231 KERNEL_BITS=32 ./config
1235 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
1236 256 bit AES and HMAC with SHA256.
1239 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
1242 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
1245 *) To enable users to have their own config files and build file templates,
1246 Configure looks in the directory indicated by the environment variable
1247 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
1248 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
1249 name and is used as is.
1252 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
1253 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
1254 X509_CERT_FILE_CTX was removed.
1257 *) "shared" builds are now the default. To create only static libraries use
1258 the "no-shared" Configure option.
1261 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
1262 All of these option have not worked for some while and are fundamental
1266 *) Make various cleanup routines no-ops and mark them as deprecated. Most
1267 global cleanup functions are no longer required because they are handled
1268 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
1269 Explicitly de-initing can cause problems (e.g. where a library that uses
1270 OpenSSL de-inits, but an application is still using it). The affected
1271 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
1272 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
1273 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
1274 COMP_zlib_cleanup().
1277 *) --strict-warnings no longer enables runtime debugging options
1278 such as REF_DEBUG. Instead, debug options are automatically
1279 enabled with '--debug' builds.
1280 [Andy Polyakov, Emilia Käsper]
1282 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
1283 have been moved out of the public header files. New functions for managing
1284 these have been added.
1287 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
1288 objects have been moved out of the public header files. New
1289 functions for managing these have been added.
1292 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1293 have been moved out of the public header files. New functions for managing
1294 these have been added.
1297 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1298 moved out of the public header files. New functions for managing these
1302 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
1305 *) Removed the mk1mf build scripts.
1308 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1309 it is always safe to #include a header now.
1312 *) Removed the aged BC-32 config and all its supporting scripts
1315 *) Removed support for Ultrix, Netware, and OS/2.
1318 *) Add support for HKDF.
1319 [Alessandro Ghedini]
1321 *) Add support for blake2b and blake2s
1324 *) Added support for "pipelining". Ciphers that have the
1325 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1326 encryptions/decryptions simultaneously. There are currently no built-in
1327 ciphers with this property but the expectation is that engines will be able
1328 to offer it to significantly improve throughput. Support has been extended
1329 into libssl so that multiple records for a single connection can be
1330 processed in one go (for >=TLS 1.1).
1333 *) Added the AFALG engine. This is an async capable engine which is able to
1334 offload work to the Linux kernel. In this initial version it only supports
1335 AES128-CBC. The kernel must be version 4.1.0 or greater.
1338 *) OpenSSL now uses a new threading API. It is no longer necessary to
1339 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1340 are two supported threading models: pthreads and windows threads. It is
1341 also possible to configure OpenSSL at compile time for "no-threads". The
1342 old threading API should no longer be used. The functions have been
1343 replaced with "no-op" compatibility macros.
1344 [Alessandro Ghedini, Matt Caswell]
1346 *) Modify behavior of ALPN to invoke callback after SNI/servername
1347 callback, such that updates to the SSL_CTX affect ALPN.
1350 *) Add SSL_CIPHER queries for authentication and key-exchange.
1353 *) Changes to the DEFAULT cipherlist:
1354 - Prefer (EC)DHE handshakes over plain RSA.
1355 - Prefer AEAD ciphers over legacy ciphers.
1356 - Prefer ECDSA over RSA when both certificates are available.
1357 - Prefer TLSv1.2 ciphers/PRF.
1358 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1362 *) Change the ECC default curve list to be this, in order: x25519,
1363 secp256r1, secp521r1, secp384r1.
1366 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1367 disabled by default. They can be re-enabled using the
1368 enable-weak-ssl-ciphers option to Configure.
1371 *) If the server has ALPN configured, but supports no protocols that the
1372 client advertises, send a fatal "no_application_protocol" alert.
1373 This behaviour is SHALL in RFC 7301, though it isn't universally
1374 implemented by other servers.
1377 *) Add X25519 support.
1378 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1379 for public and private key encoding using the format documented in
1380 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1381 key generation and key derivation.
1383 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1387 *) Deprecate SRP_VBASE_get_by_user.
1388 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1389 In order to fix an unavoidable memory leak (CVE-2016-0798),
1390 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1391 seed, even if the seed is configured.
1393 Users should use SRP_VBASE_get1_by_user instead. Note that in
1394 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1395 also that even though configuring the SRP seed attempts to hide
1396 invalid usernames by continuing the handshake with fake
1397 credentials, this behaviour is not constant time and no strong
1398 guarantees are made that the handshake is indistinguishable from
1399 that of a valid user.
1402 *) Configuration change; it's now possible to build dynamic engines
1403 without having to build shared libraries and vice versa. This
1404 only applies to the engines in engines/, those in crypto/engine/
1405 will always be built into libcrypto (i.e. "static").
1407 Building dynamic engines is enabled by default; to disable, use
1408 the configuration option "disable-dynamic-engine".
1410 The only requirements for building dynamic engines are the
1411 presence of the DSO module and building with position independent
1412 code, so they will also automatically be disabled if configuring
1413 with "disable-dso" or "disable-pic".
1415 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1416 are also taken away from openssl/opensslconf.h, as they are
1420 *) Configuration change; if there is a known flag to compile
1421 position independent code, it will always be applied on the
1422 libcrypto and libssl object files, and never on the application
1423 object files. This means other libraries that use routines from
1424 libcrypto / libssl can be made into shared libraries regardless
1425 of how OpenSSL was configured.
1427 If this isn't desirable, the configuration options "disable-pic"
1428 or "no-pic" can be used to disable the use of PIC. This will
1429 also disable building shared libraries and dynamic engines.
1432 *) Removed JPAKE code. It was experimental and has no wide use.
1435 *) The INSTALL_PREFIX Makefile variable has been renamed to
1436 DESTDIR. That makes for less confusion on what this variable
1437 is for. Also, the configuration option --install_prefix is
1441 *) Heartbeat for TLS has been removed and is disabled by default
1442 for DTLS; configure with enable-heartbeats. Code that uses the
1443 old #define's might need to be updated.
1444 [Emilia Käsper, Rich Salz]
1446 *) Rename REF_CHECK to REF_DEBUG.
1449 *) New "unified" build system
1451 The "unified" build system is aimed to be a common system for all
1452 platforms we support. With it comes new support for VMS.
1454 This system builds supports building in a different directory tree
1455 than the source tree. It produces one Makefile (for unix family
1456 or lookalikes), or one descrip.mms (for VMS).
1458 The source of information to make the Makefile / descrip.mms is
1459 small files called 'build.info', holding the necessary
1460 information for each directory with source to compile, and a
1461 template in Configurations, like unix-Makefile.tmpl or
1464 With this change, the library names were also renamed on Windows
1465 and on VMS. They now have names that are closer to the standard
1466 on Unix, and include the major version number, and in certain
1467 cases, the architecture they are built for. See "Notes on shared
1468 libraries" in INSTALL.
1470 We rely heavily on the perl module Text::Template.
1473 *) Added support for auto-initialisation and de-initialisation of the library.
1474 OpenSSL no longer requires explicit init or deinit routines to be called,
1475 except in certain circumstances. See the OPENSSL_init_crypto() and
1476 OPENSSL_init_ssl() man pages for further information.
1479 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1480 "peer" argument is now expected to be a BIO_ADDR object.
1482 *) Rewrite of BIO networking library. The BIO library lacked consistent
1483 support of IPv6, and adding it required some more extensive
1484 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1485 which hold all types of addresses and chains of address information.
1486 It also introduces a new API, with functions like BIO_socket,
1487 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1488 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1489 have been adapted accordingly.
1492 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1496 *) CRIME protection: disable compression by default, even if OpenSSL is
1497 compiled with zlib enabled. Applications can still enable compression
1498 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1499 using the SSL_CONF library to configure compression.
1502 *) The signature of the session callback configured with
1503 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1504 was explicitly marked as 'const unsigned char*' instead of
1508 *) Always DPURIFY. Remove the use of uninitialized memory in the
1509 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1512 *) Removed many obsolete configuration items, including
1513 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1514 MD2_CHAR, MD2_INT, MD2_LONG
1516 IDEA_SHORT, IDEA_LONG
1517 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1518 [Rich Salz, with advice from Andy Polyakov]
1520 *) Many BN internals have been moved to an internal header file.
1521 [Rich Salz with help from Andy Polyakov]
1523 *) Configuration and writing out the results from it has changed.
1524 Files such as Makefile include/openssl/opensslconf.h and are now
1525 produced through general templates, such as Makefile.in and
1526 crypto/opensslconf.h.in and some help from the perl module
1529 Also, the center of configuration information is no longer
1530 Makefile. Instead, Configure produces a perl module in
1531 configdata.pm which holds most of the config data (in the hash
1532 table %config), the target data that comes from the target
1533 configuration in one of the Configurations/*.conf files (in
1537 *) To clarify their intended purposes, the Configure options
1538 --prefix and --openssldir change their semantics, and become more
1539 straightforward and less interdependent.
1541 --prefix shall be used exclusively to give the location INSTALLTOP
1542 where programs, scripts, libraries, include files and manuals are
1543 going to be installed. The default is now /usr/local.
1545 --openssldir shall be used exclusively to give the default
1546 location OPENSSLDIR where certificates, private keys, CRLs are
1547 managed. This is also where the default openssl.cnf gets
1549 If the directory given with this option is a relative path, the
1550 values of both the --prefix value and the --openssldir value will
1551 be combined to become OPENSSLDIR.
1552 The default for --openssldir is INSTALLTOP/ssl.
1554 Anyone who uses --openssldir to specify where OpenSSL is to be
1555 installed MUST change to use --prefix instead.
1558 *) The GOST engine was out of date and therefore it has been removed. An up
1559 to date GOST engine is now being maintained in an external repository.
1560 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1561 support for GOST ciphersuites (these are only activated if a GOST engine
1565 *) EGD is no longer supported by default; use enable-egd when
1567 [Ben Kaduk and Rich Salz]
1569 *) The distribution now has Makefile.in files, which are used to
1570 create Makefile's when Configure is run. *Configure must be run
1571 before trying to build now.*
1574 *) The return value for SSL_CIPHER_description() for error conditions
1578 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1580 Obtaining and performing DNSSEC validation of TLSA records is
1581 the application's responsibility. The application provides
1582 the TLSA records of its choice to OpenSSL, and these are then
1583 used to authenticate the peer.
1585 The TLSA records need not even come from DNS. They can, for
1586 example, be used to implement local end-entity certificate or
1587 trust-anchor "pinning", where the "pin" data takes the form
1588 of TLSA records, which can augment or replace verification
1589 based on the usual WebPKI public certification authorities.
1592 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1593 continues to support deprecated interfaces in default builds.
1594 However, applications are strongly advised to compile their
1595 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1596 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1597 or the 1.1.0 releases.
1599 In environments in which all applications have been ported to
1600 not use any deprecated interfaces OpenSSL's Configure script
1601 should be used with the --api=1.1.0 option to entirely remove
1602 support for the deprecated features from the library and
1603 unconditionally disable them in the installed headers.
1604 Essentially the same effect can be achieved with the "no-deprecated"
1605 argument to Configure, except that this will always restrict
1606 the build to just the latest API, rather than a fixed API
1609 As applications are ported to future revisions of the API,
1610 they should update their compile-time OPENSSL_API_COMPAT define
1611 accordingly, but in most cases should be able to continue to
1612 compile with later releases.
1614 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1615 0x10000000L and 0x00908000L, respectively. However those
1616 versions did not support the OPENSSL_API_COMPAT feature, and
1617 so applications are not typically tested for explicit support
1618 of just the undeprecated features of either release.
1621 *) Add support for setting the minimum and maximum supported protocol.
1622 It can bet set via the SSL_set_min_proto_version() and
1623 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1624 MaxProtocol. It's recommended to use the new APIs to disable
1625 protocols instead of disabling individual protocols using
1626 SSL_set_options() or SSL_CONF's Protocol. This change also
1627 removes support for disabling TLS 1.2 in the OpenSSL TLS
1628 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1631 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1634 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1635 and integrates ECDSA and ECDH functionality into EC. Implementations can
1636 now redirect key generation and no longer need to convert to or from
1639 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1640 include the ec.h header file instead.
1643 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1644 ciphers who are no longer supported and drops support the ephemeral RSA key
1645 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1648 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1649 opaque. For HMAC_CTX, the following constructors and destructors
1652 HMAC_CTX *HMAC_CTX_new(void);
1653 void HMAC_CTX_free(HMAC_CTX *ctx);
1655 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1656 destroy such methods has been added. See EVP_MD_meth_new(3) and
1657 EVP_CIPHER_meth_new(3) for documentation.
1660 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1661 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1662 EVP_MD_CTX_reset() should be called instead to reinitialise
1663 an already created structure.
1664 2) For consistency with the majority of our object creators and
1665 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1666 EVP_MD_CTX_(new|free). The old names are retained as macros
1667 for deprecated builds.
1670 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1671 cryptographic operations to be performed asynchronously as long as an
1672 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1673 further details. Libssl has also had this capability integrated with the
1674 introduction of the new mode SSL_MODE_ASYNC and associated error
1675 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1676 pages. This work was developed in partnership with Intel Corp.
1679 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1680 always enabled now. If you want to disable the support you should
1681 exclude it using the list of supported ciphers. This also means that the
1682 "-no_ecdhe" option has been removed from s_server.
1685 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1686 SSL_{CTX_}set1_curves() which can set a list.
1689 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1690 curve you want to support using SSL_{CTX_}set1_curves().
1693 *) State machine rewrite. The state machine code has been significantly
1694 refactored in order to remove much duplication of code and solve issues
1695 with the old code (see ssl/statem/README for further details). This change
1696 does have some associated API changes. Notably the SSL_state() function
1697 has been removed and replaced by SSL_get_state which now returns an
1698 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1699 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1703 *) All instances of the string "ssleay" in the public API were replaced
1704 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1705 Some error codes related to internal RSA_eay API's were renamed.
1708 *) The demo files in crypto/threads were moved to demo/threads.
1711 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1713 [Matt Caswell, Rich Salz]
1715 *) New ASN.1 embed macro.
1717 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1718 structure is not allocated: it is part of the parent. That is instead of
1726 This reduces memory fragmentation and make it impossible to accidentally
1727 set a mandatory field to NULL.
1729 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1730 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1731 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1735 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1738 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1739 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1740 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1741 DES and RC4 ciphersuites.
1744 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1745 This changes the decoding behaviour for some invalid messages,
1746 though the change is mostly in the more lenient direction, and
1747 legacy behaviour is preserved as much as possible.
1750 *) Fix no-stdio build.
1751 [ David Woodhouse <David.Woodhouse@intel.com> and also
1752 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1754 *) New testing framework
1755 The testing framework has been largely rewritten and is now using
1756 perl and the perl modules Test::Harness and an extended variant of
1757 Test::More called OpenSSL::Test to do its work. All test scripts in
1758 test/ have been rewritten into test recipes, and all direct calls to
1759 executables in test/Makefile have become individual recipes using the
1760 simplified testing OpenSSL::Test::Simple.
1762 For documentation on our testing modules, do:
1764 perldoc test/testlib/OpenSSL/Test/Simple.pm
1765 perldoc test/testlib/OpenSSL/Test.pm
1769 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1770 are used; the latter aborts on memory leaks (usually checked on exit).
1771 Some undocumented "set malloc, etc., hooks" functions were removed
1772 and others were changed. All are now documented.
1775 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1777 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1779 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1780 from RFC4279, RFC4785, RFC5487, RFC5489.
1782 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1783 original RSA_PSK patch.
1786 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1787 era flag was never set throughout the codebase (only read). Also removed
1788 SSL3_FLAGS_POP_BUFFER which was only used if
1789 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1792 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1793 to be "oneline" instead of "compat".
1796 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1797 not aware of clients that still exhibit this bug, and the workaround
1798 hasn't been working properly for a while.
1801 *) The return type of BIO_number_read() and BIO_number_written() as well as
1802 the corresponding num_read and num_write members in the BIO structure has
1803 changed from unsigned long to uint64_t. On platforms where an unsigned
1804 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1808 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1809 OpenSSL without support for them. It also means that maintaining
1810 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1811 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1814 *) Removed support for the two export grade static DH ciphersuites
1815 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1816 were newly added (along with a number of other static DH ciphersuites) to
1817 1.0.2. However the two export ones have *never* worked since they were
1818 introduced. It seems strange in any case to be adding new export
1819 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1822 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1823 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1824 and turned into macros which simply call the new preferred function names
1825 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1826 should use the new names instead. Also as part of this change the ssl23.h
1827 header file has been removed.
1830 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1831 code and the associated standard is no longer considered fit-for-purpose.
1834 *) RT2547 was closed. When generating a private key, try to make the
1835 output file readable only by the owner. This behavior change might
1836 be noticeable when interacting with other software.
1838 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1842 *) Added HTTP GET support to the ocsp command.
1845 *) Changed default digest for the dgst and enc commands from MD5 to
1849 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1852 *) Added support for TLS extended master secret from
1853 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1854 initial patch which was a great help during development.
1857 *) All libssl internal structures have been removed from the public header
1858 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1859 now redundant). Users should not attempt to access internal structures
1860 directly. Instead they should use the provided API functions.
1863 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1864 Access to deprecated functions can be re-enabled by running config with
1865 "enable-deprecated". In addition applications wishing to use deprecated
1866 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1867 will, by default, disable some transitive includes that previously existed
1868 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1871 *) Added support for OCB mode. OpenSSL has been granted a patent license
1872 compatible with the OpenSSL license for use of OCB. Details are available
1873 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1874 for OCB can be removed by calling config with no-ocb.
1877 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1878 compatible client hello.
1881 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1882 done while fixing the error code for the key-too-small case.
1883 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1885 *) CA.sh has been removed; use CA.pl instead.
1888 *) Removed old DES API.
1891 *) Remove various unsupported platforms:
1897 Sinix/ReliantUNIX RM400
1902 16-bit platforms such as WIN16
1905 *) Clean up OPENSSL_NO_xxx #define's
1906 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1907 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1908 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1909 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1910 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1911 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1912 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1913 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1914 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1915 Remove MS_STATIC; it's a relic from platforms <32 bits.
1918 *) Cleaned up dead code
1919 Remove all but one '#ifdef undef' which is to be looked at.
1922 *) Clean up calling of xxx_free routines.
1923 Just like free(), fix most of the xxx_free routines to accept
1924 NULL. Remove the non-null checks from callers. Save much code.
1927 *) Add secure heap for storage of private keys (when possible).
1928 Add BIO_s_secmem(), CBIGNUM, etc.
1929 Contributed by Akamai Technologies under our Corporate CLA.
1932 *) Experimental support for a new, fast, unbiased prime candidate generator,
1933 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1934 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1936 *) New output format NSS in the sess_id command line tool. This allows
1937 exporting the session id and the master key in NSS keylog format.
1938 [Martin Kaiser <martin@kaiser.cx>]
1940 *) Harmonize version and its documentation. -f flag is used to display
1942 [mancha <mancha1@zoho.com>]
1944 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1945 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1946 [mancha <mancha1@zoho.com>]
1948 *) Fix some double frees. These are not thought to be exploitable.
1949 [mancha <mancha1@zoho.com>]
1951 *) A missing bounds check in the handling of the TLS heartbeat extension
1952 can be used to reveal up to 64k of memory to a connected client or
1955 Thanks for Neel Mehta of Google Security for discovering this bug and to
1956 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1957 preparing the fix (CVE-2014-0160)
1958 [Adam Langley, Bodo Moeller]
1960 *) Fix for the attack described in the paper "Recovering OpenSSL
1961 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1962 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1963 http://eprint.iacr.org/2014/140
1965 Thanks to Yuval Yarom and Naomi Benger for discovering this
1966 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1967 [Yuval Yarom and Naomi Benger]
1969 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1970 this fixes a limitation in previous versions of OpenSSL.
1973 *) Experimental encrypt-then-mac support.
1975 Experimental support for encrypt then mac from
1976 draft-gutmann-tls-encrypt-then-mac-02.txt
1978 To enable it set the appropriate extension number (0x42 for the test
1979 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1981 For non-compliant peers (i.e. just about everything) this should have no
1984 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1988 *) Add EVP support for key wrapping algorithms, to avoid problems with
1989 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1990 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1991 algorithms and include tests cases.
1994 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1998 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1999 MGF1 digest and OAEP label.
2002 *) Make openssl verify return errors.
2003 [Chris Palmer <palmer@google.com> and Ben Laurie]
2005 *) New function ASN1_TIME_diff to calculate the difference between two
2006 ASN1_TIME structures or one structure and the current time.
2009 *) Update fips_test_suite to support multiple command line options. New
2010 test to induce all self test errors in sequence and check expected
2014 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
2015 sign or verify all in one operation.
2018 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
2019 test programs and fips_test_suite. Includes functionality to parse
2020 the minimal script output of fipsalgest.pl directly.
2023 *) Add authorisation parameter to FIPS_module_mode_set().
2026 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
2029 *) Use separate DRBG fields for internal and external flags. New function
2030 FIPS_drbg_health_check() to perform on demand health checking. Add
2031 generation tests to fips_test_suite with reduced health check interval to
2032 demonstrate periodic health checking. Add "nodh" option to
2033 fips_test_suite to skip very slow DH test.
2036 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
2040 *) More extensive health check for DRBG checking many more failure modes.
2041 New function FIPS_selftest_drbg_all() to handle every possible DRBG
2042 combination: call this in fips_test_suite.
2045 *) Add support for canonical generation of DSA parameter 'g'. See
2048 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
2049 POST to handle HMAC cases.
2052 *) Add functions FIPS_module_version() and FIPS_module_version_text()
2053 to return numerical and string versions of the FIPS module number.
2056 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
2057 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
2058 outside the validated module in the FIPS capable OpenSSL.
2061 *) Minor change to DRBG entropy callback semantics. In some cases
2062 there is no multiple of the block length between min_len and
2063 max_len. Allow the callback to return more than max_len bytes
2064 of entropy but discard any extra: it is the callback's responsibility
2065 to ensure that the extra data discarded does not impact the
2066 requested amount of entropy.
2069 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
2070 information in FIPS186-3, SP800-57 and SP800-131A.
2073 *) CCM support via EVP. Interface is very similar to GCM case except we
2074 must supply all data in one chunk (i.e. no update, final) and the
2075 message length must be supplied if AAD is used. Add algorithm test
2079 *) Initial version of POST overhaul. Add POST callback to allow the status
2080 of POST to be monitored and/or failures induced. Modify fips_test_suite
2081 to use callback. Always run all selftests even if one fails.
2084 *) XTS support including algorithm test driver in the fips_gcmtest program.
2085 Note: this does increase the maximum key length from 32 to 64 bytes but
2086 there should be no binary compatibility issues as existing applications
2087 will never use XTS mode.
2090 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
2091 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
2092 performs algorithm blocking for unapproved PRNG types. Also do not
2093 set PRNG type in FIPS_mode_set(): leave this to the application.
2094 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
2095 the standard OpenSSL PRNG: set additional data to a date time vector.
2098 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
2099 This shouldn't present any incompatibility problems because applications
2100 shouldn't be using these directly and any that are will need to rethink
2101 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
2104 *) Extensive self tests and health checking required by SP800-90 DRBG.
2105 Remove strength parameter from FIPS_drbg_instantiate and always
2106 instantiate at maximum supported strength.
2109 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
2112 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
2115 *) New function DH_compute_key_padded() to compute a DH key and pad with
2116 leading zeroes if needed: this complies with SP800-56A et al.
2119 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
2120 anything, incomplete, subject to change and largely untested at present.
2123 *) Modify fipscanisteronly build option to only build the necessary object
2124 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
2127 *) Add experimental option FIPSSYMS to give all symbols in
2128 fipscanister.o and FIPS or fips prefix. This will avoid
2129 conflicts with future versions of OpenSSL. Add perl script
2130 util/fipsas.pl to preprocess assembly language source files
2131 and rename any affected symbols.
2134 *) Add selftest checks and algorithm block of non-fips algorithms in
2135 FIPS mode. Remove DES2 from selftests.
2138 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
2139 return internal method without any ENGINE dependencies. Add new
2140 tiny fips sign and verify functions.
2143 *) New build option no-ec2m to disable characteristic 2 code.
2146 *) New build option "fipscanisteronly". This only builds fipscanister.o
2147 and (currently) associated fips utilities. Uses the file Makefile.fips
2148 instead of Makefile.org as the prototype.
2151 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
2152 Update fips_gcmtest to use IV generator.
2155 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
2156 setting output buffer to NULL. The *Final function must be
2157 called although it will not retrieve any additional data. The tag
2158 can be set or retrieved with a ctrl. The IV length is by default 12
2159 bytes (96 bits) but can be set to an alternative value. If the IV
2160 length exceeds the maximum IV length (currently 16 bytes) it cannot be
2164 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
2165 underlying do_cipher function handles all cipher semantics itself
2166 including padding and finalisation. This is useful if (for example)
2167 an ENGINE cipher handles block padding itself. The behaviour of
2168 do_cipher is subtly changed if this flag is set: the return value
2169 is the number of characters written to the output buffer (zero is
2170 no longer an error code) or a negative error code. Also if the
2171 input buffer is NULL and length 0 finalisation should be performed.
2174 *) If a candidate issuer certificate is already part of the constructed
2175 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
2178 *) Improve forward-security support: add functions
2180 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
2181 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
2183 for use by SSL/TLS servers; the callback function will be called whenever a
2184 new session is created, and gets to decide whether the session may be
2185 cached to make it resumable (return 0) or not (return 1). (As by the
2186 SSL/TLS protocol specifications, the session_id sent by the server will be
2187 empty to indicate that the session is not resumable; also, the server will
2188 not generate RFC 4507 (RFC 5077) session tickets.)
2190 A simple reasonable callback implementation is to return is_forward_secure.
2191 This parameter will be set to 1 or 0 depending on the ciphersuite selected
2192 by the SSL/TLS server library, indicating whether it can provide forward
2194 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2196 *) New -verify_name option in command line utilities to set verification
2200 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
2201 Add CMAC pkey methods.
2204 *) Experimental renegotiation in s_server -www mode. If the client
2205 browses /reneg connection is renegotiated. If /renegcert it is
2206 renegotiated requesting a certificate.
2209 *) Add an "external" session cache for debugging purposes to s_server. This
2210 should help trace issues which normally are only apparent in deployed
2211 multi-process servers.
2214 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
2215 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
2216 BIO_set_cipher() and some obscure PEM functions were changed so they
2217 can now return an error. The RAND changes required a change to the
2218 RAND_METHOD structure.
2221 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
2222 a gcc attribute to warn if the result of a function is ignored. This
2223 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
2224 whose return value is often ignored.
2227 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
2228 These allow SCTs (signed certificate timestamps) to be requested and
2229 validated when establishing a connection.
2230 [Rob Percival <robpercival@google.com>]
2232 Changes between 1.0.2g and 1.0.2h [3 May 2016]
2234 *) Prevent padding oracle in AES-NI CBC MAC check
2236 A MITM attacker can use a padding oracle attack to decrypt traffic
2237 when the connection uses an AES CBC cipher and the server support
2240 This issue was introduced as part of the fix for Lucky 13 padding
2241 attack (CVE-2013-0169). The padding check was rewritten to be in
2242 constant time by making sure that always the same bytes are read and
2243 compared against either the MAC or padding bytes. But it no longer
2244 checked that there was enough data to have both the MAC and padding
2247 This issue was reported by Juraj Somorovsky using TLS-Attacker.
2251 *) Fix EVP_EncodeUpdate overflow
2253 An overflow can occur in the EVP_EncodeUpdate() function which is used for
2254 Base64 encoding of binary data. If an attacker is able to supply very large
2255 amounts of input data then a length check can overflow resulting in a heap
2258 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
2259 the PEM_write_bio* family of functions. These are mainly used within the
2260 OpenSSL command line applications, so any application which processes data
2261 from an untrusted source and outputs it as a PEM file should be considered
2262 vulnerable to this issue. User applications that call these APIs directly
2263 with large amounts of untrusted data may also be vulnerable.
2265 This issue was reported by Guido Vranken.
2269 *) Fix EVP_EncryptUpdate overflow
2271 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
2272 is able to supply very large amounts of input data after a previous call to
2273 EVP_EncryptUpdate() with a partial block then a length check can overflow
2274 resulting in a heap corruption. Following an analysis of all OpenSSL
2275 internal usage of the EVP_EncryptUpdate() function all usage is one of two
2276 forms. The first form is where the EVP_EncryptUpdate() call is known to be
2277 the first called function after an EVP_EncryptInit(), and therefore that
2278 specific call must be safe. The second form is where the length passed to
2279 EVP_EncryptUpdate() can be seen from the code to be some small value and
2280 therefore there is no possibility of an overflow. Since all instances are
2281 one of these two forms, it is believed that there can be no overflows in
2282 internal code due to this problem. It should be noted that
2283 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
2284 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
2285 of these calls have also been analysed too and it is believed there are no
2286 instances in internal usage where an overflow could occur.
2288 This issue was reported by Guido Vranken.
2292 *) Prevent ASN.1 BIO excessive memory allocation
2294 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2295 a short invalid encoding can cause allocation of large amounts of memory
2296 potentially consuming excessive resources or exhausting memory.
2298 Any application parsing untrusted data through d2i BIO functions is
2299 affected. The memory based functions such as d2i_X509() are *not* affected.
2300 Since the memory based functions are used by the TLS library, TLS
2301 applications are not affected.
2303 This issue was reported by Brian Carpenter.
2309 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2310 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2311 in arbitrary stack data being returned in the buffer.
2313 This issue was reported by Guido Vranken.
2317 *) Modify behavior of ALPN to invoke callback after SNI/servername
2318 callback, such that updates to the SSL_CTX affect ALPN.
2321 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2325 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2326 methods are enabled and ssl2 is disabled the methods return NULL.
2329 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2331 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2332 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2333 provide any "EXPORT" or "LOW" strength ciphers.
2336 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2337 is by default disabled at build-time. Builds that are not configured with
2338 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2339 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2340 will need to explicitly call either of:
2342 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2344 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2346 as appropriate. Even if either of those is used, or the application
2347 explicitly uses the version-specific SSLv2_method() or its client and
2348 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2349 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2350 ciphers, and SSLv2 56-bit DES are no longer available.
2354 *) Fix a double-free in DSA code
2356 A double free bug was discovered when OpenSSL parses malformed DSA private
2357 keys and could lead to a DoS attack or memory corruption for applications
2358 that receive DSA private keys from untrusted sources. This scenario is
2361 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2366 *) Disable SRP fake user seed to address a server memory leak.
2368 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2370 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2371 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2372 was changed to ignore the "fake user" SRP seed, even if the seed
2375 Users should use SRP_VBASE_get1_by_user instead. Note that in
2376 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2377 also that even though configuring the SRP seed attempts to hide
2378 invalid usernames by continuing the handshake with fake
2379 credentials, this behaviour is not constant time and no strong
2380 guarantees are made that the handshake is indistinguishable from
2381 that of a valid user.
2385 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2387 In the BN_hex2bn function the number of hex digits is calculated using an
2388 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2389 large values of |i| this can result in |bn_expand| not allocating any
2390 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2391 field as NULL leading to a subsequent NULL ptr deref. For very large values
2392 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2393 In this case memory is allocated to the internal BIGNUM data field, but it
2394 is insufficiently sized leading to heap corruption. A similar issue exists
2395 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2396 is ever called by user applications with very large untrusted hex/dec data.
2397 This is anticipated to be a rare occurrence.
2399 All OpenSSL internal usage of these functions use data that is not expected
2400 to be untrusted, e.g. config file data or application command line
2401 arguments. If user developed applications generate config file data based
2402 on untrusted data then it is possible that this could also lead to security
2403 consequences. This is also anticipated to be rare.
2405 This issue was reported to OpenSSL by Guido Vranken.
2409 *) Fix memory issues in BIO_*printf functions
2411 The internal |fmtstr| function used in processing a "%s" format string in
2412 the BIO_*printf functions could overflow while calculating the length of a
2413 string and cause an OOB read when printing very long strings.
2415 Additionally the internal |doapr_outch| function can attempt to write to an
2416 OOB memory location (at an offset from the NULL pointer) in the event of a
2417 memory allocation failure. In 1.0.2 and below this could be caused where
2418 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2419 could be in processing a very long "%s" format string. Memory leaks can
2422 The first issue may mask the second issue dependent on compiler behaviour.
2423 These problems could enable attacks where large amounts of untrusted data
2424 is passed to the BIO_*printf functions. If applications use these functions
2425 in this way then they could be vulnerable. OpenSSL itself uses these
2426 functions when printing out human-readable dumps of ASN.1 data. Therefore
2427 applications that print this data could be vulnerable if the data is from
2428 untrusted sources. OpenSSL command line applications could also be
2429 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2430 as command line arguments.
2432 Libssl is not considered directly vulnerable. Additionally certificates etc
2433 received via remote connections via libssl are also unlikely to be able to
2434 trigger these issues because of message size limits enforced within libssl.
2436 This issue was reported to OpenSSL Guido Vranken.
2440 *) Side channel attack on modular exponentiation
2442 A side-channel attack was found which makes use of cache-bank conflicts on
2443 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2444 of RSA keys. The ability to exploit this issue is limited as it relies on
2445 an attacker who has control of code in a thread running on the same
2446 hyper-threaded core as the victim thread which is performing decryptions.
2448 This issue was reported to OpenSSL by Yuval Yarom, The University of
2449 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2450 Nadia Heninger, University of Pennsylvania with more information at
2451 http://cachebleed.info.
2455 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2456 if no keysize is specified with default_bits. This fixes an
2457 omission in an earlier change that changed all RSA/DSA key generation
2458 apps to use 2048 bits by default.
2461 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2462 *) DH small subgroups
2464 Historically OpenSSL only ever generated DH parameters based on "safe"
2465 primes. More recently (in version 1.0.2) support was provided for
2466 generating X9.42 style parameter files such as those required for RFC 5114
2467 support. The primes used in such files may not be "safe". Where an
2468 application is using DH configured with parameters based on primes that are
2469 not "safe" then an attacker could use this fact to find a peer's private
2470 DH exponent. This attack requires that the attacker complete multiple
2471 handshakes in which the peer uses the same private DH exponent. For example
2472 this could be used to discover a TLS server's private DH exponent if it's
2473 reusing the private DH exponent or it's using a static DH ciphersuite.
2475 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2476 TLS. It is not on by default. If the option is not set then the server
2477 reuses the same private DH exponent for the life of the server process and
2478 would be vulnerable to this attack. It is believed that many popular
2479 applications do set this option and would therefore not be at risk.
2481 The fix for this issue adds an additional check where a "q" parameter is
2482 available (as is the case in X9.42 based parameters). This detects the
2483 only known attack, and is the only possible defense for static DH
2484 ciphersuites. This could have some performance impact.
2486 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2487 default and cannot be disabled. This could have some performance impact.
2489 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2493 *) SSLv2 doesn't block disabled ciphers
2495 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2496 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2497 been disabled, provided that the SSLv2 protocol was not also disabled via
2500 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2501 and Sebastian Schinzel.
2505 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2507 *) BN_mod_exp may produce incorrect results on x86_64
2509 There is a carry propagating bug in the x86_64 Montgomery squaring
2510 procedure. No EC algorithms are affected. Analysis suggests that attacks
2511 against RSA and DSA as a result of this defect would be very difficult to
2512 perform and are not believed likely. Attacks against DH are considered just
2513 feasible (although very difficult) because most of the work necessary to
2514 deduce information about a private key may be performed offline. The amount
2515 of resources required for such an attack would be very significant and
2516 likely only accessible to a limited number of attackers. An attacker would
2517 additionally need online access to an unpatched system using the target
2518 private key in a scenario with persistent DH parameters and a private
2519 key that is shared between multiple clients. For example this can occur by
2520 default in OpenSSL DHE based SSL/TLS ciphersuites.
2522 This issue was reported to OpenSSL by Hanno Böck.
2526 *) Certificate verify crash with missing PSS parameter
2528 The signature verification routines will crash with a NULL pointer
2529 dereference if presented with an ASN.1 signature using the RSA PSS
2530 algorithm and absent mask generation function parameter. Since these
2531 routines are used to verify certificate signature algorithms this can be
2532 used to crash any certificate verification operation and exploited in a
2533 DoS attack. Any application which performs certificate verification is
2534 vulnerable including OpenSSL clients and servers which enable client
2537 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2541 *) X509_ATTRIBUTE memory leak
2543 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2544 memory. This structure is used by the PKCS#7 and CMS routines so any
2545 application which reads PKCS#7 or CMS data from untrusted sources is
2546 affected. SSL/TLS is not affected.
2548 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2553 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2554 This changes the decoding behaviour for some invalid messages,
2555 though the change is mostly in the more lenient direction, and
2556 legacy behaviour is preserved as much as possible.
2559 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2561 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2563 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2565 *) Alternate chains certificate forgery
2567 During certificate verification, OpenSSL will attempt to find an
2568 alternative certificate chain if the first attempt to build such a chain
2569 fails. An error in the implementation of this logic can mean that an
2570 attacker could cause certain checks on untrusted certificates to be
2571 bypassed, such as the CA flag, enabling them to use a valid leaf
2572 certificate to act as a CA and "issue" an invalid certificate.
2574 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2578 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2580 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2581 incompatibility in the handling of HMAC. The previous ABI has now been
2585 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2587 *) Malformed ECParameters causes infinite loop
2589 When processing an ECParameters structure OpenSSL enters an infinite loop
2590 if the curve specified is over a specially malformed binary polynomial
2593 This can be used to perform denial of service against any
2594 system which processes public keys, certificate requests or
2595 certificates. This includes TLS clients and TLS servers with
2596 client authentication enabled.
2598 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2602 *) Exploitable out-of-bounds read in X509_cmp_time
2604 X509_cmp_time does not properly check the length of the ASN1_TIME
2605 string and can read a few bytes out of bounds. In addition,
2606 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2609 An attacker can use this to craft malformed certificates and CRLs of
2610 various sizes and potentially cause a segmentation fault, resulting in
2611 a DoS on applications that verify certificates or CRLs. TLS clients
2612 that verify CRLs are affected. TLS clients and servers with client
2613 authentication enabled may be affected if they use custom verification
2616 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2617 independently by Hanno Böck.
2621 *) PKCS7 crash with missing EnvelopedContent
2623 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2624 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2625 with missing content and trigger a NULL pointer dereference on parsing.
2627 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2628 structures from untrusted sources are affected. OpenSSL clients and
2629 servers are not affected.
2631 This issue was reported to OpenSSL by Michal Zalewski (Google).
2635 *) CMS verify infinite loop with unknown hash function
2637 When verifying a signedData message the CMS code can enter an infinite loop
2638 if presented with an unknown hash function OID. This can be used to perform
2639 denial of service against any system which verifies signedData messages using
2641 This issue was reported to OpenSSL by Johannes Bauer.
2645 *) Race condition handling NewSessionTicket
2647 If a NewSessionTicket is received by a multi-threaded client when attempting to
2648 reuse a previous ticket then a race condition can occur potentially leading to
2649 a double free of the ticket data.
2653 *) Only support 256-bit or stronger elliptic curves with the
2654 'ecdh_auto' setting (server) or by default (client). Of supported
2655 curves, prefer P-256 (both).
2658 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2660 *) ClientHello sigalgs DoS fix
2662 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2663 invalid signature algorithms extension a NULL pointer dereference will
2664 occur. This can be exploited in a DoS attack against the server.
2666 This issue was was reported to OpenSSL by David Ramos of Stanford
2669 [Stephen Henson and Matt Caswell]
2671 *) Multiblock corrupted pointer fix
2673 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2674 feature only applies on 64 bit x86 architecture platforms that support AES
2675 NI instructions. A defect in the implementation of "multiblock" can cause
2676 OpenSSL's internal write buffer to become incorrectly set to NULL when
2677 using non-blocking IO. Typically, when the user application is using a
2678 socket BIO for writing, this will only result in a failed connection.
2679 However if some other BIO is used then it is likely that a segmentation
2680 fault will be triggered, thus enabling a potential DoS attack.
2682 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2686 *) Segmentation fault in DTLSv1_listen fix
2688 The DTLSv1_listen function is intended to be stateless and processes the
2689 initial ClientHello from many peers. It is common for user code to loop
2690 over the call to DTLSv1_listen until a valid ClientHello is received with
2691 an associated cookie. A defect in the implementation of DTLSv1_listen means
2692 that state is preserved in the SSL object from one invocation to the next
2693 that can lead to a segmentation fault. Errors processing the initial
2694 ClientHello can trigger this scenario. An example of such an error could be
2695 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2698 This issue was reported to OpenSSL by Per Allansson.
2702 *) Segmentation fault in ASN1_TYPE_cmp fix
2704 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2705 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2706 certificate signature algorithm consistency this can be used to crash any
2707 certificate verification operation and exploited in a DoS attack. Any
2708 application which performs certificate verification is vulnerable including
2709 OpenSSL clients and servers which enable client authentication.
2713 *) Segmentation fault for invalid PSS parameters fix
2715 The signature verification routines will crash with a NULL pointer
2716 dereference if presented with an ASN.1 signature using the RSA PSS
2717 algorithm and invalid parameters. Since these routines are used to verify
2718 certificate signature algorithms this can be used to crash any
2719 certificate verification operation and exploited in a DoS attack. Any
2720 application which performs certificate verification is vulnerable including
2721 OpenSSL clients and servers which enable client authentication.
2723 This issue was was reported to OpenSSL by Brian Carpenter.
2727 *) ASN.1 structure reuse memory corruption fix
2729 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2730 memory corruption via an invalid write. Such reuse is and has been
2731 strongly discouraged and is believed to be rare.
2733 Applications that parse structures containing CHOICE or ANY DEFINED BY
2734 components may be affected. Certificate parsing (d2i_X509 and related
2735 functions) are however not affected. OpenSSL clients and servers are
2740 *) PKCS7 NULL pointer dereferences fix
2742 The PKCS#7 parsing code does not handle missing outer ContentInfo
2743 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2744 missing content and trigger a NULL pointer dereference on parsing.
2746 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2747 otherwise parse PKCS#7 structures from untrusted sources are
2748 affected. OpenSSL clients and servers are not affected.
2750 This issue was reported to OpenSSL by Michal Zalewski (Google).
2754 *) DoS via reachable assert in SSLv2 servers fix
2756 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2757 servers that both support SSLv2 and enable export cipher suites by sending
2758 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2760 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2761 (OpenSSL development team).
2765 *) Empty CKE with client auth and DHE fix
2767 If client auth is used then a server can seg fault in the event of a DHE
2768 ciphersuite being selected and a zero length ClientKeyExchange message
2769 being sent by the client. This could be exploited in a DoS attack.
2773 *) Handshake with unseeded PRNG fix
2775 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2776 with an unseeded PRNG. The conditions are:
2777 - The client is on a platform where the PRNG has not been seeded
2778 automatically, and the user has not seeded manually
2779 - A protocol specific client method version has been used (i.e. not
2780 SSL_client_methodv23)
2781 - A ciphersuite is used that does not require additional random data from
2782 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2784 If the handshake succeeds then the client random that has been used will
2785 have been generated from a PRNG with insufficient entropy and therefore the
2786 output may be predictable.
2788 For example using the following command with an unseeded openssl will
2789 succeed on an unpatched platform:
2791 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2795 *) Use After Free following d2i_ECPrivatekey error fix
2797 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2798 could cause a use after free condition. This, in turn, could cause a double
2799 free in several private key parsing functions (such as d2i_PrivateKey
2800 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2801 for applications that receive EC private keys from untrusted
2802 sources. This scenario is considered rare.
2804 This issue was discovered by the BoringSSL project and fixed in their
2809 *) X509_to_X509_REQ NULL pointer deref fix
2811 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2812 the certificate key is invalid. This function is rarely used in practice.
2814 This issue was discovered by Brian Carpenter.
2818 *) Removed the export ciphers from the DEFAULT ciphers
2821 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2823 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2824 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2825 So far those who have to target multiple platforms would compromise
2826 and argue that binary targeting say ARMv5 would still execute on
2827 ARMv8. "Universal" build resolves this compromise by providing
2828 near-optimal performance even on newer platforms.
2831 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2832 (other platforms pending).
2833 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2835 *) Add support for the SignedCertificateTimestampList certificate and
2836 OCSP response extensions from RFC6962.
2839 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2840 for corner cases. (Certain input points at infinity could lead to
2841 bogus results, with non-infinity inputs mapped to infinity too.)
2844 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2845 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2846 common cases are optimized and there still is room for further
2847 improvements. Vector Permutation AES for Altivec is also added.
2850 *) Add support for little-endian ppc64 Linux target.
2851 [Marcelo Cerri (IBM)]
2853 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2854 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2855 are optimized and there still is room for further improvements.
2856 Both 32- and 64-bit modes are supported.
2857 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2859 *) Improved ARMv7 NEON support.
2862 *) Support for SPARC Architecture 2011 crypto extensions, first
2863 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2864 SHA256/512, MD5, GHASH and modular exponentiation.
2865 [Andy Polyakov, David Miller]
2867 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2869 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2871 *) Support for new and upcoming Intel processors, including AVX2,
2872 BMI and SHA ISA extensions. This includes additional "stitched"
2873 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2876 This work was sponsored by Intel Corp.
2879 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2880 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2881 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2884 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2885 this fixes a limitation in previous versions of OpenSSL.
2888 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2889 MGF1 digest and OAEP label.
2892 *) Add EVP support for key wrapping algorithms, to avoid problems with
2893 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2894 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2895 algorithms and include tests cases.
2898 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2900 [Douglas E. Engert, Steve Henson]
2902 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2903 difference in days and seconds between two tm or ASN1_TIME structures.
2906 *) Add -rev test option to s_server to just reverse order of characters
2907 received by client and send back to server. Also prints an abbreviated
2908 summary of the connection parameters.
2911 *) New option -brief for s_client and s_server to print out a brief summary
2912 of connection parameters.
2915 *) Add callbacks for arbitrary TLS extensions.
2916 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2918 *) New option -crl_download in several openssl utilities to download CRLs
2919 from CRLDP extension in certificates.
2922 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2925 *) New function X509_CRL_diff to generate a delta CRL from the difference
2926 of two full CRLs. Add support to "crl" utility.
2929 *) New functions to set lookup_crls function and to retrieve
2930 X509_STORE from X509_STORE_CTX.
2933 *) Print out deprecated issuer and subject unique ID fields in
2937 *) Extend OCSP I/O functions so they can be used for simple general purpose
2938 HTTP as well as OCSP. New wrapper function which can be used to download
2939 CRLs using the OCSP API.
2942 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2945 *) SSL_CONF* functions. These provide a common framework for application
2946 configuration using configuration files or command lines.
2949 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2950 message callback and prints the results. Needs compile time option
2951 "enable-ssl-trace". New options to s_client and s_server to enable
2955 *) New ctrl and macro to retrieve supported points extensions.
2956 Print out extension in s_server and s_client.
2959 *) New functions to retrieve certificate signature and signature
2963 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2967 *) New Suite B modes for TLS code. These use and enforce the requirements
2968 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2969 only use Suite B curves. The Suite B modes can be set by using the
2970 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2973 *) New chain verification flags for Suite B levels of security. Check
2974 algorithms are acceptable when flags are set in X509_verify_cert.
2977 *) Make tls1_check_chain return a set of flags indicating checks passed
2978 by a certificate chain. Add additional tests to handle client
2979 certificates: checks for matching certificate type and issuer name
2983 *) If an attempt is made to use a signature algorithm not in the peer
2984 preference list abort the handshake. If client has no suitable
2985 signature algorithms in response to a certificate request do not
2986 use the certificate.
2989 *) If server EC tmp key is not in client preference list abort handshake.
2992 *) Add support for certificate stores in CERT structure. This makes it
2993 possible to have different stores per SSL structure or one store in
2994 the parent SSL_CTX. Include distinct stores for certificate chain
2995 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2996 to build and store a certificate chain in CERT structure: returning
2997 an error if the chain cannot be built: this will allow applications
2998 to test if a chain is correctly configured.
3000 Note: if the CERT based stores are not set then the parent SSL_CTX
3001 store is used to retain compatibility with existing behaviour.
3005 *) New function ssl_set_client_disabled to set a ciphersuite disabled
3006 mask based on the current session, check mask when sending client
3007 hello and checking the requested ciphersuite.
3010 *) New ctrls to retrieve and set certificate types in a certificate
3011 request message. Print out received values in s_client. If certificate
3012 types is not set with custom values set sensible values based on
3013 supported signature algorithms.
3016 *) Support for distinct client and server supported signature algorithms.
3019 *) Add certificate callback. If set this is called whenever a certificate
3020 is required by client or server. An application can decide which
3021 certificate chain to present based on arbitrary criteria: for example
3022 supported signature algorithms. Add very simple example to s_server.
3023 This fixes many of the problems and restrictions of the existing client
3024 certificate callback: for example you can now clear an existing
3025 certificate and specify the whole chain.
3028 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
3029 the certificate can be used for (if anything). Set valid_flags field
3030 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
3031 to have similar checks in it.
3033 Add new "cert_flags" field to CERT structure and include a "strict mode".
3034 This enforces some TLS certificate requirements (such as only permitting
3035 certificate signature algorithms contained in the supported algorithms
3036 extension) which some implementations ignore: this option should be used
3037 with caution as it could cause interoperability issues.
3040 *) Update and tidy signature algorithm extension processing. Work out
3041 shared signature algorithms based on preferences and peer algorithms
3042 and print them out in s_client and s_server. Abort handshake if no
3043 shared signature algorithms.
3046 *) Add new functions to allow customised supported signature algorithms
3047 for SSL and SSL_CTX structures. Add options to s_client and s_server
3051 *) New function SSL_certs_clear() to delete all references to certificates
3052 from an SSL structure. Before this once a certificate had been added
3053 it couldn't be removed.
3056 *) Integrate hostname, email address and IP address checking with certificate
3057 verification. New verify options supporting checking in openssl utility.
3060 *) Fixes and wildcard matching support to hostname and email checking
3061 functions. Add manual page.
3062 [Florian Weimer (Red Hat Product Security Team)]
3064 *) New functions to check a hostname email or IP address against a
3065 certificate. Add options x509 utility to print results of checks against
3069 *) Fix OCSP checking.
3070 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
3072 *) Initial experimental support for explicitly trusted non-root CAs.
3073 OpenSSL still tries to build a complete chain to a root but if an
3074 intermediate CA has a trust setting included that is used. The first
3075 setting is used: whether to trust (e.g., -addtrust option to the x509
3079 *) Add -trusted_first option which attempts to find certificates in the
3080 trusted store even if an untrusted chain is also supplied.
3083 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
3084 platform support for Linux and Android.
3087 *) Support for linux-x32, ILP32 environment in x86_64 framework.
3090 *) Experimental multi-implementation support for FIPS capable OpenSSL.
3091 When in FIPS mode the approved implementations are used as normal,
3092 when not in FIPS mode the internal unapproved versions are used instead.
3093 This means that the FIPS capable OpenSSL isn't forced to use the
3094 (often lower performance) FIPS implementations outside FIPS mode.
3097 *) Transparently support X9.42 DH parameters when calling
3098 PEM_read_bio_DHparameters. This means existing applications can handle
3099 the new parameter format automatically.
3102 *) Initial experimental support for X9.42 DH parameter format: mainly
3103 to support use of 'q' parameter for RFC5114 parameters.
3106 *) Add DH parameters from RFC5114 including test data to dhtest.
3109 *) Support for automatic EC temporary key parameter selection. If enabled
3110 the most preferred EC parameters are automatically used instead of
3111 hardcoded fixed parameters. Now a server just has to call:
3112 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
3113 support ECDH and use the most appropriate parameters.
3116 *) Enhance and tidy EC curve and point format TLS extension code. Use
3117 static structures instead of allocation if default values are used.
3118 New ctrls to set curves we wish to support and to retrieve shared curves.
3119 Print out shared curves in s_server. New options to s_server and s_client
3120 to set list of supported curves.
3123 *) New ctrls to retrieve supported signature algorithms and
3124 supported curve values as an array of NIDs. Extend openssl utility
3125 to print out received values.
3128 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
3129 between NIDs and the more common NIST names such as "P-256". Enhance
3130 ecparam utility and ECC method to recognise the NIST names for curves.
3133 *) Enhance SSL/TLS certificate chain handling to support different
3134 chains for each certificate instead of one chain in the parent SSL_CTX.
3137 *) Support for fixed DH ciphersuite client authentication: where both
3138 server and client use DH certificates with common parameters.
3141 *) Support for fixed DH ciphersuites: those requiring DH server
3145 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
3147 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
3148 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
3149 X509_CINF_get_signature were reverted post internal team review.
3151 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
3153 *) Build fixes for the Windows and OpenVMS platforms
3154 [Matt Caswell and Richard Levitte]
3156 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
3158 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
3159 message can cause a segmentation fault in OpenSSL due to a NULL pointer
3160 dereference. This could lead to a Denial Of Service attack. Thanks to
3161 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
3165 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
3166 dtls1_buffer_record function under certain conditions. In particular this
3167 could occur if an attacker sent repeated DTLS records with the same
3168 sequence number but for the next epoch. The memory leak could be exploited
3169 by an attacker in a Denial of Service attack through memory exhaustion.
3170 Thanks to Chris Mueller for reporting this issue.
3174 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
3175 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
3176 method would be set to NULL which could later result in a NULL pointer
3177 dereference. Thanks to Frank Schmirler for reporting this issue.
3181 *) Abort handshake if server key exchange message is omitted for ephemeral
3184 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
3185 reporting this issue.
3189 *) Remove non-export ephemeral RSA code on client and server. This code
3190 violated the TLS standard by allowing the use of temporary RSA keys in
3191 non-export ciphersuites and could be used by a server to effectively
3192 downgrade the RSA key length used to a value smaller than the server
3193 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
3194 INRIA or reporting this issue.
3198 *) Fixed issue where DH client certificates are accepted without verification.
3199 An OpenSSL server will accept a DH certificate for client authentication
3200 without the certificate verify message. This effectively allows a client to
3201 authenticate without the use of a private key. This only affects servers
3202 which trust a client certificate authority which issues certificates
3203 containing DH keys: these are extremely rare and hardly ever encountered.
3204 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
3209 *) Ensure that the session ID context of an SSL is updated when its
3210 SSL_CTX is updated via SSL_set_SSL_CTX.
3212 The session ID context is typically set from the parent SSL_CTX,
3213 and can vary with the CTX.
3216 *) Fix various certificate fingerprint issues.
3218 By using non-DER or invalid encodings outside the signed portion of a
3219 certificate the fingerprint can be changed without breaking the signature.
3220 Although no details of the signed portion of the certificate can be changed
3221 this can cause problems with some applications: e.g. those using the
3222 certificate fingerprint for blacklists.
3224 1. Reject signatures with non zero unused bits.
3226 If the BIT STRING containing the signature has non zero unused bits reject
3227 the signature. All current signature algorithms require zero unused bits.
3229 2. Check certificate algorithm consistency.
3231 Check the AlgorithmIdentifier inside TBS matches the one in the
3232 certificate signature. NB: this will result in signature failure
3233 errors for some broken certificates.
3235 Thanks to Konrad Kraszewski from Google for reporting this issue.
3237 3. Check DSA/ECDSA signatures use DER.
3239 Re-encode DSA/ECDSA signatures and compare with the original received
3240 signature. Return an error if there is a mismatch.
3242 This will reject various cases including garbage after signature
3243 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
3244 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
3245 (negative or with leading zeroes).
3247 Further analysis was conducted and fixes were developed by Stephen Henson
3248 of the OpenSSL core team.
3253 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
3254 results on some platforms, including x86_64. This bug occurs at random
3255 with a very low probability, and is not known to be exploitable in any
3256 way, though its exact impact is difficult to determine. Thanks to Pieter
3257 Wuille (Blockstream) who reported this issue and also suggested an initial
3258 fix. Further analysis was conducted by the OpenSSL development team and
3259 Adam Langley of Google. The final fix was developed by Andy Polyakov of
3260 the OpenSSL core team.
3264 *) Do not resume sessions on the server if the negotiated protocol
3265 version does not match the session's version. Resuming with a different
3266 version, while not strictly forbidden by the RFC, is of questionable
3267 sanity and breaks all known clients.
3268 [David Benjamin, Emilia Käsper]
3270 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
3271 early CCS messages during renegotiation. (Note that because
3272 renegotiation is encrypted, this early CCS was not exploitable.)
3275 *) Tighten client-side session ticket handling during renegotiation:
3276 ensure that the client only accepts a session ticket if the server sends
3277 the extension anew in the ServerHello. Previously, a TLS client would
3278 reuse the old extension state and thus accept a session ticket if one was
3279 announced in the initial ServerHello.
3281 Similarly, ensure that the client requires a session ticket if one
3282 was advertised in the ServerHello. Previously, a TLS client would
3283 ignore a missing NewSessionTicket message.
3286 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
3288 *) SRTP Memory Leak.
3290 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
3291 sends a carefully crafted handshake message, to cause OpenSSL to fail
3292 to free up to 64k of memory causing a memory leak. This could be
3293 exploited in a Denial Of Service attack. This issue affects OpenSSL
3294 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
3295 whether SRTP is used or configured. Implementations of OpenSSL that
3296 have been compiled with OPENSSL_NO_SRTP defined are not affected.
3298 The fix was developed by the OpenSSL team.
3302 *) Session Ticket Memory Leak.
3304 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
3305 integrity of that ticket is first verified. In the event of a session
3306 ticket integrity check failing, OpenSSL will fail to free memory
3307 causing a memory leak. By sending a large number of invalid session
3308 tickets an attacker could exploit this issue in a Denial Of Service
3313 *) Build option no-ssl3 is incomplete.
3315 When OpenSSL is configured with "no-ssl3" as a build option, servers
3316 could accept and complete a SSL 3.0 handshake, and clients could be
3317 configured to send them.
3319 [Akamai and the OpenSSL team]
3321 *) Add support for TLS_FALLBACK_SCSV.
3322 Client applications doing fallback retries should call
3323 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3325 [Adam Langley, Bodo Moeller]
3327 *) Add additional DigestInfo checks.
3329 Re-encode DigestInto in DER and check against the original when
3330 verifying RSA signature: this will reject any improperly encoded
3331 DigestInfo structures.
3333 Note: this is a precautionary measure and no attacks are currently known.
3337 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3339 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3340 SRP code can be overrun an internal buffer. Add sanity check that
3341 g, A, B < N to SRP code.
3343 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3344 Group for discovering this issue.
3348 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3349 TLS 1.0 instead of higher protocol versions when the ClientHello message
3350 is badly fragmented. This allows a man-in-the-middle attacker to force a
3351 downgrade to TLS 1.0 even if both the server and the client support a
3352 higher protocol version, by modifying the client's TLS records.
3354 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3355 researching this issue.
3359 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3360 to a denial of service attack. A malicious server can crash the client
3361 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3362 ciphersuite and sending carefully crafted handshake messages.
3364 Thanks to Felix Gröbert (Google) for discovering and researching this
3369 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3370 to leak memory. This can be exploited through a Denial of Service attack.
3371 Thanks to Adam Langley for discovering and researching this issue.
3375 *) An attacker can force openssl to consume large amounts of memory whilst
3376 processing DTLS handshake messages. This can be exploited through a
3377 Denial of Service attack.
3378 Thanks to Adam Langley for discovering and researching this issue.
3382 *) An attacker can force an error condition which causes openssl to crash
3383 whilst processing DTLS packets due to memory being freed twice. This
3384 can be exploited through a Denial of Service attack.
3385 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3390 *) If a multithreaded client connects to a malicious server using a resumed
3391 session and the server sends an ec point format extension it could write
3392 up to 255 bytes to freed memory.
3394 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3399 *) A malicious server can crash an OpenSSL client with a null pointer
3400 dereference (read) by specifying an SRP ciphersuite even though it was not
3401 properly negotiated with the client. This can be exploited through a
3402 Denial of Service attack.
3404 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3405 discovering and researching this issue.
3409 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3410 X509_name_oneline, X509_name_print_ex et al. to leak some information
3411 from the stack. Applications may be affected if they echo pretty printing
3412 output to the attacker.
3414 Thanks to Ivan Fratric (Google) for discovering this issue.
3416 [Emilia Käsper, and Steve Henson]
3418 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3419 for corner cases. (Certain input points at infinity could lead to
3420 bogus results, with non-infinity inputs mapped to infinity too.)
3423 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3425 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3426 handshake can force the use of weak keying material in OpenSSL
3427 SSL/TLS clients and servers.
3429 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3430 researching this issue. (CVE-2014-0224)
3431 [KIKUCHI Masashi, Steve Henson]
3433 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3434 OpenSSL DTLS client the code can be made to recurse eventually crashing
3437 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3439 [Imre Rad, Steve Henson]
3441 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3442 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3443 client or server. This is potentially exploitable to run arbitrary
3444 code on a vulnerable client or server.
3446 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3447 [Jüri Aedla, Steve Henson]
3449 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3450 are subject to a denial of service attack.
3452 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3453 this issue. (CVE-2014-3470)
3454 [Felix Gröbert, Ivan Fratric, Steve Henson]
3456 *) Harmonize version and its documentation. -f flag is used to display
3458 [mancha <mancha1@zoho.com>]
3460 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3461 in i2d_ECPrivateKey.
3462 [mancha <mancha1@zoho.com>]
3464 *) Fix some double frees. These are not thought to be exploitable.
3465 [mancha <mancha1@zoho.com>]
3467 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3469 *) A missing bounds check in the handling of the TLS heartbeat extension
3470 can be used to reveal up to 64k of memory to a connected client or
3473 Thanks for Neel Mehta of Google Security for discovering this bug and to
3474 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3475 preparing the fix (CVE-2014-0160)
3476 [Adam Langley, Bodo Moeller]
3478 *) Fix for the attack described in the paper "Recovering OpenSSL
3479 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3480 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3481 http://eprint.iacr.org/2014/140
3483 Thanks to Yuval Yarom and Naomi Benger for discovering this
3484 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3485 [Yuval Yarom and Naomi Benger]
3487 *) TLS pad extension: draft-agl-tls-padding-03
3489 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3490 TLS client Hello record length value would otherwise be > 255 and
3491 less that 512 pad with a dummy extension containing zeroes so it
3492 is at least 512 bytes long.
3494 [Adam Langley, Steve Henson]
3496 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3498 *) Fix for TLS record tampering bug. A carefully crafted invalid
3499 handshake could crash OpenSSL with a NULL pointer exception.
3500 Thanks to Anton Johansson for reporting this issues.
3503 *) Keep original DTLS digest and encryption contexts in retransmission
3504 structures so we can use the previous session parameters if they need
3505 to be resent. (CVE-2013-6450)
3508 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3509 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3510 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3511 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3512 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3513 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3514 [Rob Stradling, Adam Langley]
3516 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3518 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3519 supporting platforms or when small records were transferred.
3520 [Andy Polyakov, Steve Henson]
3522 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3524 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3526 This addresses the flaw in CBC record processing discovered by
3527 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3528 at: http://www.isg.rhul.ac.uk/tls/
3530 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3531 Security Group at Royal Holloway, University of London
3532 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3533 Emilia Käsper for the initial patch.
3535 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3537 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3538 ciphersuites which can be exploited in a denial of service attack.
3539 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3540 and detecting this bug and to Wolfgang Ettlinger
3541 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3545 *) Return an error when checking OCSP signatures when key is NULL.
3546 This fixes a DoS attack. (CVE-2013-0166)
3549 *) Make openssl verify return errors.
3550 [Chris Palmer <palmer@google.com> and Ben Laurie]
3552 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3553 the right response is stapled. Also change SSL_get_certificate()
3554 so it returns the certificate actually sent.
3555 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3556 [Rob Stradling <rob.stradling@comodo.com>]
3558 *) Fix possible deadlock when decoding public keys.
3561 *) Don't use TLS 1.0 record version number in initial client hello
3565 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3567 *) Sanity check record length before skipping explicit IV in TLS
3568 1.2, 1.1 and DTLS to fix DoS attack.
3570 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3571 fuzzing as a service testing platform.
3575 *) Initialise tkeylen properly when encrypting CMS messages.
3576 Thanks to Solar Designer of Openwall for reporting this issue.
3579 *) In FIPS mode don't try to use composite ciphers as they are not
3583 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3585 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3586 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3587 mean any application compiled against OpenSSL 1.0.0 headers setting
3588 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3589 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3590 0x10000000L Any application which was previously compiled against
3591 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3592 will need to be recompiled as a result. Letting be results in
3593 inability to disable specifically TLS 1.1 and in client context,
3594 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3597 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3598 disable just protocol X, but all protocols above X *if* there are
3599 protocols *below* X still enabled. In more practical terms it means
3600 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3601 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3602 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3606 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3608 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3609 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3610 in CRYPTO_realloc_clean.
3612 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3613 issue and to Adam Langley <agl@chromium.org> for fixing it.
3615 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3617 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3620 *) Workarounds for some broken servers that "hang" if a client hello
3621 record length exceeds 255 bytes.
3623 1. Do not use record version number > TLS 1.0 in initial client
3624 hello: some (but not all) hanging servers will now work.
3625 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3626 the number of ciphers sent in the client hello. This should be
3627 set to an even number, such as 50, for example by passing:
3628 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3629 Most broken servers should now work.
3630 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3631 TLS 1.2 client support entirely.
3634 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3637 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3639 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3640 STRING form instead of a DigestInfo.
3643 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3644 and the RSA_sign/RSA_verify functions. This was made more apparent when
3645 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3646 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3647 the correct format in RSA_verify so both forms transparently work.
3650 *) Some servers which support TLS 1.0 can choke if we initially indicate
3651 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3652 encrypted premaster secret. As a workaround use the maximum permitted
3653 client version in client hello, this should keep such servers happy
3654 and still work with previous versions of OpenSSL.
3657 *) Add support for TLS/DTLS heartbeats.
3658 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3660 *) Add support for SCTP.
3661 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3663 *) Improved PRNG seeding for VOS.
3664 [Paul Green <Paul.Green@stratus.com>]
3666 *) Extensive assembler packs updates, most notably:
3668 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3669 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3670 - x86_64: bit-sliced AES implementation;
3671 - ARM: NEON support, contemporary platforms optimizations;
3672 - s390x: z196 support;
3673 - *: GHASH and GF(2^m) multiplication implementations;
3677 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3678 (removal of unnecessary code)
3679 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3681 *) Add TLS key material exporter from RFC 5705.
3684 *) Add DTLS-SRTP negotiation from RFC 5764.
3687 *) Add Next Protocol Negotiation,
3688 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3689 disabled with a no-npn flag to config or Configure. Code donated
3691 [Adam Langley <agl@google.com> and Ben Laurie]
3693 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3694 NIST-P256, NIST-P521, with constant-time single point multiplication on
3695 typical inputs. Compiler support for the nonstandard type __uint128_t is
3696 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3697 Code made available under Apache License version 2.0.
3699 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3700 line to include this in your build of OpenSSL, and run "make depend" (or
3701 "make update"). This enables the following EC_METHODs:
3703 EC_GFp_nistp224_method()
3704 EC_GFp_nistp256_method()
3705 EC_GFp_nistp521_method()
3707 EC_GROUP_new_by_curve_name() will automatically use these (while
3708 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3710 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3712 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3713 all platforms. Move ssize_t definition from e_os.h to the public
3714 header file e_os2.h as it now appears in public header file cms.h
3717 *) New -sigopt option to the ca, req and x509 utilities. Additional
3718 signature parameters can be passed using this option and in
3722 *) Add RSA PSS signing function. This will generate and set the
3723 appropriate AlgorithmIdentifiers for PSS based on those in the
3724 corresponding EVP_MD_CTX structure. No application support yet.
3727 *) Support for companion algorithm specific ASN1 signing routines.
3728 New function ASN1_item_sign_ctx() signs a pre-initialised
3729 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3730 the appropriate parameters.
3733 *) Add new algorithm specific ASN1 verification initialisation function
3734 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3735 handling will be the same no matter what EVP_PKEY_METHOD is used.
3736 Add a PSS handler to support verification of PSS signatures: checked
3737 against a number of sample certificates.
3740 *) Add signature printing for PSS. Add PSS OIDs.
3741 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3743 *) Add algorithm specific signature printing. An individual ASN1 method
3744 can now print out signatures instead of the standard hex dump.
3746 More complex signatures (e.g. PSS) can print out more meaningful
3747 information. Include DSA version that prints out the signature
3751 *) Password based recipient info support for CMS library: implementing
3755 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3756 neatly separates the code into cipher and PBE sections and is required
3757 for some algorithms that split PBES2 into separate pieces (such as
3758 password based CMS).
3761 *) Session-handling fixes:
3762 - Fix handling of connections that are resuming with a session ID,
3763 but also support Session Tickets.
3764 - Fix a bug that suppressed issuing of a new ticket if the client
3765 presented a ticket with an expired session.
3766 - Try to set the ticket lifetime hint to something reasonable.
3767 - Make tickets shorter by excluding irrelevant information.
3768 - On the client side, don't ignore renewed tickets.
3769 [Adam Langley, Bodo Moeller (Google)]
3771 *) Fix PSK session representation.
3774 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3776 This work was sponsored by Intel.
3779 *) Add GCM support to TLS library. Some custom code is needed to split
3780 the IV between the fixed (from PRF) and explicit (from TLS record)
3781 portions. This adds all GCM ciphersuites supported by RFC5288 and
3782 RFC5289. Generalise some AES* cipherstrings to include GCM and
3783 add a special AESGCM string for GCM only.
3786 *) Expand range of ctrls for AES GCM. Permit setting invocation
3787 field on decrypt and retrieval of invocation field only on encrypt.
3790 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3791 As required by RFC5289 these ciphersuites cannot be used if for
3792 versions of TLS earlier than 1.2.
3795 *) For FIPS capable OpenSSL interpret a NULL default public key method
3796 as unset and return the appropriate default but do *not* set the default.
3797 This means we can return the appropriate method in applications that
3798 switch between FIPS and non-FIPS modes.
3801 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3802 ENGINE is used then we cannot handle that in the FIPS module so we
3803 keep original code iff non-FIPS operations are allowed.
3806 *) Add -attime option to openssl utilities.
3807 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3809 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3812 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3813 FIPS EC methods unconditionally for now.
3816 *) New build option no-ec2m to disable characteristic 2 code.
3819 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3820 all cases can be covered as some introduce binary incompatibilities.
3823 *) Redirect RSA operations to FIPS module including keygen,
3824 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3827 *) Add similar low level API blocking to ciphers.
3830 *) Low level digest APIs are not approved in FIPS mode: any attempt
3831 to use these will cause a fatal error. Applications that *really* want
3832 to use them can use the private_* version instead.
3835 *) Redirect cipher operations to FIPS module for FIPS builds.
3838 *) Redirect digest operations to FIPS module for FIPS builds.
3841 *) Update build system to add "fips" flag which will link in fipscanister.o
3842 for static and shared library builds embedding a signature if needed.
3845 *) Output TLS supported curves in preference order instead of numerical
3846 order. This is currently hardcoded for the highest order curves first.
3847 This should be configurable so applications can judge speed vs strength.
3850 *) Add TLS v1.2 server support for client authentication.
3853 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3857 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3858 FIPS modules versions.
3861 *) Add TLS v1.2 client side support for client authentication. Keep cache
3862 of handshake records longer as we don't know the hash algorithm to use
3863 until after the certificate request message is received.
3866 *) Initial TLS v1.2 client support. Add a default signature algorithms
3867 extension including all the algorithms we support. Parse new signature
3868 format in client key exchange. Relax some ECC signing restrictions for
3869 TLS v1.2 as indicated in RFC5246.
3872 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3873 to new signature format when needed using client digest preference.
3874 All server ciphersuites should now work correctly in TLS v1.2. No client
3875 support yet and no support for client certificates.
3878 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3879 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3880 ciphersuites. At present only RSA key exchange ciphersuites work with
3881 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3882 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3883 and version checking.
3886 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3887 with this defined it will not be affected by any changes to ssl internal
3888 structures. Add several utility functions to allow openssl application
3889 to work with OPENSSL_NO_SSL_INTERN defined.
3892 *) A long standing patch to add support for SRP from EdelWeb (Peter
3893 Sylvester and Christophe Renou) was integrated.
3894 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3895 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3898 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3901 *) Permit abbreviated handshakes when renegotiating using the function
3902 SSL_renegotiate_abbreviated().
3903 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3905 *) Add call to ENGINE_register_all_complete() to
3906 ENGINE_load_builtin_engines(), so some implementations get used
3907 automatically instead of needing explicit application support.
3910 *) Add support for TLS key exporter as described in RFC5705.
3911 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3913 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3914 a few changes are required:
3916 Add SSL_OP_NO_TLSv1_1 flag.
3917 Add TLSv1_1 methods.
3918 Update version checking logic to handle version 1.1.
3919 Add explicit IV handling (ported from DTLS code).
3920 Add command line options to s_client/s_server.
3923 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3925 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3926 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3927 content decryption and always return the same error. Note: this attack
3928 needs on average 2^20 messages so it only affects automated senders. The
3929 old behaviour can be re-enabled in the CMS code by setting the
3930 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3931 an MMA defence is not necessary.
3932 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3933 this issue. (CVE-2012-0884)
3936 *) Fix CVE-2011-4619: make sure we really are receiving a
3937 client hello before rejecting multiple SGC restarts. Thanks to
3938 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3941 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3943 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3944 Thanks to Antonio Martin, Enterprise Secure Access Research and
3945 Development, Cisco Systems, Inc. for discovering this bug and
3946 preparing a fix. (CVE-2012-0050)
3949 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3951 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3952 of the Vaudenay padding oracle attack on CBC mode encryption
3953 which enables an efficient plaintext recovery attack against
3954 the OpenSSL implementation of DTLS. Their attack exploits timing
3955 differences arising during decryption processing. A research
3956 paper describing this attack can be found at:
3957 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3958 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3959 Security Group at Royal Holloway, University of London
3960 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3961 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3962 for preparing the fix. (CVE-2011-4108)
3963 [Robin Seggelmann, Michael Tuexen]
3965 *) Clear bytes used for block padding of SSL 3.0 records.
3967 [Adam Langley (Google)]
3969 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3970 Kadianakis <desnacked@gmail.com> for discovering this issue and
3971 Adam Langley for preparing the fix. (CVE-2011-4619)
3972 [Adam Langley (Google)]
3974 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3975 [Andrey Kulikov <amdeich@gmail.com>]
3977 *) Prevent malformed RFC3779 data triggering an assertion failure.
3978 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3979 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3980 [Rob Austein <sra@hactrn.net>]
3982 *) Improved PRNG seeding for VOS.
3983 [Paul Green <Paul.Green@stratus.com>]
3985 *) Fix ssl_ciph.c set-up race.
3986 [Adam Langley (Google)]
3988 *) Fix spurious failures in ecdsatest.c.
3989 [Emilia Käsper (Google)]
3991 *) Fix the BIO_f_buffer() implementation (which was mixing different
3992 interpretations of the '..._len' fields).
3993 [Adam Langley (Google)]
3995 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3996 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3997 threads won't reuse the same blinding coefficients.
3999 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
4000 lock to call BN_BLINDING_invert_ex, and avoids one use of
4001 BN_BLINDING_update for each BN_BLINDING structure (previously,
4002 the last update always remained unused).
4003 [Emilia Käsper (Google)]
4005 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
4006 [Bob Buckholz (Google)]
4008 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
4010 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
4011 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
4012 [Kaspar Brand <ossl@velox.ch>]
4014 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
4015 for multi-threaded use of ECDH. (CVE-2011-3210)
4016 [Adam Langley (Google)]
4018 *) Fix x509_name_ex_d2i memory leak on bad inputs.
4021 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
4022 signature public key algorithm by using OID xref utilities instead.
4023 Before this you could only use some ECC ciphersuites with SHA1 only.
4026 *) Add protection against ECDSA timing attacks as mentioned in the paper
4027 by Billy Bob Brumley and Nicola Tuveri, see:
4029 http://eprint.iacr.org/2011/232.pdf
4031 [Billy Bob Brumley and Nicola Tuveri]
4033 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
4035 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
4036 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
4038 *) Fix bug in string printing code: if *any* escaping is enabled we must
4039 escape the escape character (backslash) or the resulting string is
4043 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
4045 *) Disable code workaround for ancient and obsolete Netscape browsers
4046 and servers: an attacker can use it in a ciphersuite downgrade attack.
4047 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
4050 *) Fixed J-PAKE implementation error, originally discovered by
4051 Sebastien Martini, further info and confirmation from Stefan
4052 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
4055 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
4057 *) Fix extension code to avoid race conditions which can result in a buffer
4058 overrun vulnerability: resumed sessions must not be modified as they can
4059 be shared by multiple threads. CVE-2010-3864
4062 *) Fix WIN32 build system to correctly link an ENGINE directory into
4066 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
4068 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
4070 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
4072 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
4074 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
4075 context. The operation can be customised via the ctrl mechanism in
4076 case ENGINEs want to include additional functionality.
4079 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
4082 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
4083 output hashes compatible with older versions of OpenSSL.
4084 [Willy Weisz <weisz@vcpc.univie.ac.at>]
4086 *) Fix compression algorithm handling: if resuming a session use the
4087 compression algorithm of the resumed session instead of determining
4088 it from client hello again. Don't allow server to change algorithm.
4091 *) Add load_crls() function to apps tidying load_certs() too. Add option
4092 to verify utility to allow additional CRLs to be included.
4095 *) Update OCSP request code to permit adding custom headers to the request:
4096 some responders need this.
4099 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
4101 [Julia Lawall <julia@diku.dk>]
4103 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
4104 needlessly dereferenced structures, used obsolete functions and
4105 didn't handle all updated verify codes correctly.
4108 *) Disable MD2 in the default configuration.
4111 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
4112 indicate the initial BIO being pushed or popped. This makes it possible
4113 to determine whether the BIO is the one explicitly called or as a result
4114 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
4115 it handles reference counts correctly and doesn't zero out the I/O bio
4116 when it is not being explicitly popped. WARNING: applications which
4117 included workarounds for the old buggy behaviour will need to be modified
4118 or they could free up already freed BIOs.
4121 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
4122 renaming to all platforms (within the 0.9.8 branch, this was
4123 done conditionally on Netware platforms to avoid a name clash).
4124 [Guenter <lists@gknw.net>]
4126 *) Add ECDHE and PSK support to DTLS.
4127 [Michael Tuexen <tuexen@fh-muenster.de>]
4129 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
4133 *) Add "missing" function EVP_MD_flags() (without this the only way to
4134 retrieve a digest flags is by accessing the structure directly. Update
4135 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
4136 or cipher is registered as in the "from" argument. Print out all
4137 registered digests in the dgst usage message instead of manually
4138 attempting to work them out.
4141 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
4142 this allows the use of compression and extensions. Change default cipher
4143 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
4144 by default unless an application cipher string requests it.
4147 *) Alter match criteria in PKCS12_parse(). It used to try to use local
4148 key ids to find matching certificates and keys but some PKCS#12 files
4149 don't follow the (somewhat unwritten) rules and this strategy fails.
4150 Now just gather all certificates together and the first private key
4151 then look for the first certificate that matches the key.
4154 *) Support use of registered digest and cipher names for dgst and cipher
4155 commands instead of having to add each one as a special case. So now
4162 openssl dgst -sha256 foo
4164 and this works for ENGINE based algorithms too.
4168 *) Update Gost ENGINE to support parameter files.
4169 [Victor B. Wagner <vitus@cryptocom.ru>]
4171 *) Support GeneralizedTime in ca utility.
4172 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
4174 *) Enhance the hash format used for certificate directory links. The new
4175 form uses the canonical encoding (meaning equivalent names will work
4176 even if they aren't identical) and uses SHA1 instead of MD5. This form
4177 is incompatible with the older format and as a result c_rehash should
4178 be used to rebuild symbolic links.
4181 *) Make PKCS#8 the default write format for private keys, replacing the
4182 traditional format. This form is standardised, more secure and doesn't
4183 include an implicit MD5 dependency.
4186 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
4187 committed to OpenSSL should pass this lot as a minimum.
4190 *) Add session ticket override functionality for use by EAP-FAST.
4191 [Jouni Malinen <j@w1.fi>]
4193 *) Modify HMAC functions to return a value. Since these can be implemented
4194 in an ENGINE errors can occur.
4197 *) Type-checked OBJ_bsearch_ex.
4200 *) Type-checked OBJ_bsearch. Also some constification necessitated
4201 by type-checking. Still to come: TXT_DB, bsearch(?),
4202 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
4206 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
4207 seconds to a tm structure directly, instead of going through OS
4208 specific date routines. This avoids any issues with OS routines such
4209 as the year 2038 bug. New *_adj() functions for ASN1 time structures
4210 and X509_time_adj_ex() to cover the extended range. The existing
4211 X509_time_adj() is still usable and will no longer have any date issues.
4214 *) Delta CRL support. New use deltas option which will attempt to locate
4215 and search any appropriate delta CRLs available.
4217 This work was sponsored by Google.
4220 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
4221 code and add additional score elements. Validate alternate CRL paths
4222 as part of the CRL checking and indicate a new error "CRL path validation
4223 error" in this case. Applications wanting additional details can use
4224 the verify callback and check the new "parent" field. If this is not
4225 NULL CRL path validation is taking place. Existing applications won't
4226 see this because it requires extended CRL support which is off by
4229 This work was sponsored by Google.
4232 *) Support for freshest CRL extension.
4234 This work was sponsored by Google.
4237 *) Initial indirect CRL support. Currently only supported in the CRLs
4238 passed directly and not via lookup. Process certificate issuer
4239 CRL entry extension and lookup CRL entries by bother issuer name
4240 and serial number. Check and process CRL issuer entry in IDP extension.
4242 This work was sponsored by Google.
4245 *) Add support for distinct certificate and CRL paths. The CRL issuer
4246 certificate is validated separately in this case. Only enabled if
4247 an extended CRL support flag is set: this flag will enable additional
4248 CRL functionality in future.
4250 This work was sponsored by Google.
4253 *) Add support for policy mappings extension.
4255 This work was sponsored by Google.
4258 *) Fixes to pathlength constraint, self issued certificate handling,
4259 policy processing to align with RFC3280 and PKITS tests.
4261 This work was sponsored by Google.
4264 *) Support for name constraints certificate extension. DN, email, DNS
4265 and URI types are currently supported.
4267 This work was sponsored by Google.
4270 *) To cater for systems that provide a pointer-based thread ID rather
4271 than numeric, deprecate the current numeric thread ID mechanism and
4272 replace it with a structure and associated callback type. This
4273 mechanism allows a numeric "hash" to be extracted from a thread ID in
4274 either case, and on platforms where pointers are larger than 'long',
4275 mixing is done to help ensure the numeric 'hash' is usable even if it
4276 can't be guaranteed unique. The default mechanism is to use "&errno"
4277 as a pointer-based thread ID to distinguish between threads.
4279 Applications that want to provide their own thread IDs should now use
4280 CRYPTO_THREADID_set_callback() to register a callback that will call
4281 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
4283 Note that ERR_remove_state() is now deprecated, because it is tied
4284 to the assumption that thread IDs are numeric. ERR_remove_state(0)
4285 to free the current thread's error state should be replaced by
4286 ERR_remove_thread_state(NULL).
4288 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
4289 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
4290 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
4291 application was previously providing a numeric thread callback that
4292 was inappropriate for distinguishing threads, then uniqueness might
4293 have been obtained with &errno that happened immediately in the
4294 intermediate development versions of OpenSSL; this is no longer the
4295 case, the numeric thread callback will now override the automatic use
4297 [Geoff Thorpe, with help from Bodo Moeller]
4299 *) Initial support for different CRL issuing certificates. This covers a
4300 simple case where the self issued certificates in the chain exist and
4301 the real CRL issuer is higher in the existing chain.
4303 This work was sponsored by Google.
4306 *) Removed effectively defunct crypto/store from the build.
4309 *) Revamp of STACK to provide stronger type-checking. Still to come:
4310 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
4311 ASN1_STRING, CONF_VALUE.
4314 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4315 RAM on SSL connections. This option can save about 34k per idle SSL.
4318 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4319 STACK, TXT_DB, bsearch, qsort.
4322 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4323 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4324 support for data, signedData, compressedData, digestedData and
4325 encryptedData, envelopedData types included. Scripts to check against
4326 RFC4134 examples draft and interop and consistency checks of many
4327 content types and variants.
4330 *) Add options to enc utility to support use of zlib compression BIO.
4333 *) Extend mk1mf to support importing of options and assembly language
4334 files from Configure script, currently only included in VC-WIN32.
4335 The assembly language rules can now optionally generate the source
4336 files from the associated perl scripts.
4339 *) Implement remaining functionality needed to support GOST ciphersuites.
4340 Interop testing has been performed using CryptoPro implementations.
4341 [Victor B. Wagner <vitus@cryptocom.ru>]
4343 *) s390x assembler pack.
4346 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4350 *) Implement Opaque PRF Input TLS extension as specified in
4351 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4352 official specification yet and no extension type assignment by
4353 IANA exists, this extension (for now) will have to be explicitly
4354 enabled when building OpenSSL by providing the extension number
4355 to use. For example, specify an option
4357 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4359 to the "config" or "Configure" script to enable the extension,
4360 assuming extension number 0x9527 (which is a completely arbitrary
4361 and unofficial assignment based on the MD5 hash of the Internet
4362 Draft). Note that by doing so, you potentially lose
4363 interoperability with other TLS implementations since these might
4364 be using the same extension number for other purposes.
4366 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4367 opaque PRF input value to use in the handshake. This will create
4368 an internal copy of the length-'len' string at 'src', and will
4369 return non-zero for success.
4371 To get more control and flexibility, provide a callback function
4374 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4375 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4379 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4382 Callback function 'cb' will be called in handshakes, and is
4383 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4384 Argument 'arg' is for application purposes (the value as given to
4385 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4386 be provided to the callback function). The callback function
4387 has to return non-zero to report success: usually 1 to use opaque
4388 PRF input just if possible, or 2 to enforce use of the opaque PRF
4389 input. In the latter case, the library will abort the handshake
4390 if opaque PRF input is not successfully negotiated.
4392 Arguments 'peerinput' and 'len' given to the callback function
4393 will always be NULL and 0 in the case of a client. A server will
4394 see the client's opaque PRF input through these variables if
4395 available (NULL and 0 otherwise). Note that if the server
4396 provides an opaque PRF input, the length must be the same as the
4397 length of the client's opaque PRF input.
4399 Note that the callback function will only be called when creating
4400 a new session (session resumption can resume whatever was
4401 previously negotiated), and will not be called in SSL 2.0
4402 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4403 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4404 for applications that need to enforce opaque PRF input.
4408 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4411 [Victor B. Wagner <vitus@cryptocom.ru>]
4413 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4414 RFC4507bis. The encrypted ticket format is an encrypted encoded
4415 SSL_SESSION structure, that way new session features are automatically
4418 If a client application caches session in an SSL_SESSION structure
4419 support is transparent because tickets are now stored in the encoded
4422 The SSL_CTX structure automatically generates keys for ticket
4423 protection in servers so again support should be possible
4424 with no application modification.
4426 If a client or server wishes to disable RFC4507 support then the option
4427 SSL_OP_NO_TICKET can be set.
4429 Add a TLS extension debugging callback to allow the contents of any client
4430 or server extensions to be examined.
4432 This work was sponsored by Google.
4435 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4436 OpenSSL should now compile cleanly on gcc 4.2
4437 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4439 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4440 support including streaming MAC support: this is required for GOST
4441 ciphersuite support.
4442 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4444 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4445 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4446 to output in BER and PEM format.
4449 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4450 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4451 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4452 ENGINE support for HMAC keys which are unextractable. New -mac and
4453 -macopt options to dgst utility.
4456 *) New option -sigopt to dgst utility. Update dgst to use
4457 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4458 alternative signing parameters such as X9.31 or PSS in the dgst
4462 *) Change ssl_cipher_apply_rule(), the internal function that does
4463 the work each time a ciphersuite string requests enabling
4464 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4465 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4466 the order of disabled ciphersuites such that those ciphersuites
4467 that most recently went from enabled to disabled not only stay
4468 in order with respect to each other, but also have higher priority
4469 than other disabled ciphersuites the next time ciphersuites are
4472 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4473 the same ciphersuites as with "HIGH" alone, but in a specific
4474 order where the PSK ciphersuites come first (since they are the
4475 most recently disabled ciphersuites when "HIGH" is parsed).
4477 Also, change ssl_create_cipher_list() (using this new
4478 functionality) such that between otherwise identical
4479 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4483 *) Change ssl_create_cipher_list() so that it automatically
4484 arranges the ciphersuites in reasonable order before starting
4485 to process the rule string. Thus, the definition for "DEFAULT"
4486 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4487 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4488 This makes it much easier to arrive at a reasonable default order
4489 in applications for which anonymous ciphers are OK (meaning
4490 that you can't actually use DEFAULT).
4491 [Bodo Moeller; suggested by Victor Duchovni]
4493 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4494 processing) into multiple integers instead of setting
4495 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
4496 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
4497 (These masks as well as the individual bit definitions are hidden
4498 away into the non-exported interface ssl/ssl_locl.h, so this
4499 change to the definition of the SSL_CIPHER structure shouldn't
4500 affect applications.) This give us more bits for each of these
4501 categories, so there is no longer a need to coagulate AES128 and
4502 AES256 into a single algorithm bit, and to coagulate Camellia128
4503 and Camellia256 into a single algorithm bit, which has led to all
4506 Thus, among other things, the kludge introduced in 0.9.7m and
4507 0.9.8e for masking out AES256 independently of AES128 or masking
4508 out Camellia256 independently of AES256 is not needed here in 0.9.9.
4510 With the change, we also introduce new ciphersuite aliases that
4511 so far were missing: "AES128", "AES256", "CAMELLIA128", and
4515 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
4516 Use the leftmost N bytes of the signature input if the input is
4517 larger than the prime q (with N being the size in bytes of q).
4520 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
4521 it yet and it is largely untested.
4524 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
4527 *) Initial incomplete changes to avoid need for function casts in OpenSSL
4528 some compilers (gcc 4.2 and later) reject their use. Safestack is
4529 reimplemented. Update ASN1 to avoid use of legacy functions.
4532 *) Win32/64 targets are linked with Winsock2.
4535 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
4536 to external functions. This can be used to increase CRL handling
4537 efficiency especially when CRLs are very large by (for example) storing
4538 the CRL revoked certificates in a database.
4541 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
4542 new CRLs added to a directory can be used. New command line option
4543 -verify_return_error to s_client and s_server. This causes real errors
4544 to be returned by the verify callback instead of carrying on no matter
4545 what. This reflects the way a "real world" verify callback would behave.
4548 *) GOST engine, supporting several GOST algorithms and public key formats.
4549 Kindly donated by Cryptocom.
4552 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4553 partitioned by DP are handled but no indirect CRL or reason partitioning
4554 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4555 selected via a scoring technique which handles IDP and AKID in CRLs.
4558 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4559 will ultimately be used for all verify operations: this will remove the
4560 X509_STORE dependency on certificate verification and allow alternative
4561 lookup methods. X509_STORE based implementations of these two callbacks.
4564 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4565 Modify get_crl() to find a valid (unexpired) CRL if possible.
4568 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4569 this would be called X509_CRL_cmp() but that name is already used by
4570 a function that just compares CRL issuer names. Cache several CRL
4571 extensions in X509_CRL structure and cache CRLDP in X509.
4574 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4575 this maps equivalent X509_NAME structures into a consistent structure.
4576 Name comparison can then be performed rapidly using memcmp().
4579 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4583 *) Allow digests to supply their own micalg string for S/MIME type using
4584 the ctrl EVP_MD_CTRL_MICALG.
4587 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4588 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4589 ctrl. It can then customise the structure before and/or after signing
4593 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4594 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4595 to free up any added signature OIDs.
4598 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4599 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4600 digest and cipher tables. New options added to openssl utility:
4601 list-message-digest-algorithms and list-cipher-algorithms.
4604 *) Change the array representation of binary polynomials: the list
4605 of degrees of non-zero coefficients is now terminated with -1.
4606 Previously it was terminated with 0, which was also part of the
4607 value; thus, the array representation was not applicable to
4608 polynomials where t^0 has coefficient zero. This change makes
4609 the array representation useful in a more general context.
4612 *) Various modifications and fixes to SSL/TLS cipher string
4613 handling. For ECC, the code now distinguishes between fixed ECDH
4614 with RSA certificates on the one hand and with ECDSA certificates
4615 on the other hand, since these are separate ciphersuites. The
4616 unused code for Fortezza ciphersuites has been removed.
4618 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4619 (not "ECDHE"). For consistency with the code for DH
4620 certificates, use of ECDH certificates is now considered ECDH
4621 authentication, not RSA or ECDSA authentication (the latter is
4622 merely the CA's signing algorithm and not actively used in the
4625 The temporary ciphersuite alias "ECCdraft" is no longer
4626 available, and ECC ciphersuites are no longer excluded from "ALL"
4627 and "DEFAULT". The following aliases now exist for RFC 4492
4628 ciphersuites, most of these by analogy with the DH case:
4630 kECDHr - ECDH cert, signed with RSA
4631 kECDHe - ECDH cert, signed with ECDSA
4632 kECDH - ECDH cert (signed with either RSA or ECDSA)
4633 kEECDH - ephemeral ECDH
4634 ECDH - ECDH cert or ephemeral ECDH
4640 AECDH - anonymous ECDH
4641 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4645 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
4646 Use correct micalg parameters depending on digest(s) in signed message.
4649 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
4650 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
4653 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
4654 an engine to register a method. Add ENGINE lookups for methods and
4655 functional reference processing.
4658 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
4659 EVP_{Sign,Verify}* which allow an application to customise the signature
4663 *) New -resign option to smime utility. This adds one or more signers
4664 to an existing PKCS#7 signedData structure. Also -md option to use an
4665 alternative message digest algorithm for signing.
4668 *) Tidy up PKCS#7 routines and add new functions to make it easier to
4669 create PKCS7 structures containing multiple signers. Update smime
4670 application to support multiple signers.
4673 *) New -macalg option to pkcs12 utility to allow setting of an alternative
4677 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
4678 Reorganize PBE internals to lookup from a static table using NIDs,
4679 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
4680 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
4681 PRF which will be automatically used with PBES2.
4684 *) Replace the algorithm specific calls to generate keys in "req" with the
4688 *) Update PKCS#7 enveloped data routines to use new API. This is now
4689 supported by any public key method supporting the encrypt operation. A
4690 ctrl is added to allow the public key algorithm to examine or modify
4691 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
4695 *) Add a ctrl to asn1 method to allow a public key algorithm to express
4696 a default digest type to use. In most cases this will be SHA1 but some
4697 algorithms (such as GOST) need to specify an alternative digest. The
4698 return value indicates how strong the preference is 1 means optional and
4699 2 is mandatory (that is it is the only supported type). Modify
4700 ASN1_item_sign() to accept a NULL digest argument to indicate it should
4701 use the default md. Update openssl utilities to use the default digest
4702 type for signing if it is not explicitly indicated.
4705 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
4706 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
4707 signing method from the key type. This effectively removes the link
4708 between digests and public key types.
4711 *) Add an OID cross reference table and utility functions. Its purpose is to
4712 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
4713 rsaEncryption. This will allow some of the algorithm specific hackery
4714 needed to use the correct OID to be removed.
4717 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
4718 structures for PKCS7_sign(). They are now set up by the relevant public
4722 *) Add provisional EC pkey method with support for ECDSA and ECDH.
4725 *) Add support for key derivation (agreement) in the API, DH method and
4729 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
4730 public and private key formats. As a side effect these add additional
4731 command line functionality not previously available: DSA signatures can be
4732 generated and verified using pkeyutl and DH key support and generation in
4737 [Oliver Tappe <zooey@hirschkaefer.de>]
4739 *) New make target "install_html_docs" installs HTML renditions of the
4741 [Oliver Tappe <zooey@hirschkaefer.de>]
4743 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4744 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4745 support key and parameter generation and add initial key generation
4746 functionality for RSA.
4749 *) Add functions for main EVP_PKEY_method operations. The undocumented
4750 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4751 EVP_PKEY_{encrypt,decrypt}_old.
4754 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4755 key API, doesn't do much yet.
4758 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4759 public key algorithms. New option to openssl utility:
4760 "list-public-key-algorithms" to print out info.
4763 *) Implement the Supported Elliptic Curves Extension for
4764 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4767 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4768 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4771 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4772 utilities such as rsa, dsa, dsaparam etc except they process any key
4776 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4777 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4778 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4782 *) Initial support for pluggable public key ASN1.
4783 De-spaghettify the public key ASN1 handling. Move public and private
4784 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4785 algorithm specific handling to a single module within the relevant
4786 algorithm directory. Add functions to allow (near) opaque processing
4787 of public and private key structures.
4790 *) Implement the Supported Point Formats Extension for
4791 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4794 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4795 for the psk identity [hint] and the psk callback functions to the
4796 SSL_SESSION, SSL and SSL_CTX structure.
4799 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4803 SSL_CTX_use_psk_identity_hint
4804 SSL_get_psk_identity_hint
4805 SSL_get_psk_identity
4806 SSL_use_psk_identity_hint
4808 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4810 *) Add RFC 3161 compliant time stamp request creation, response generation
4811 and response verification functionality.
4812 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4814 *) Add initial support for TLS extensions, specifically for the server_name
4815 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4816 have new members for a host name. The SSL data structure has an
4817 additional member SSL_CTX *initial_ctx so that new sessions can be
4818 stored in that context to allow for session resumption, even after the
4819 SSL has been switched to a new SSL_CTX in reaction to a client's
4820 server_name extension.
4822 New functions (subject to change):
4824 SSL_get_servername()
4825 SSL_get_servername_type()
4828 New CTRL codes and macros (subject to change):
4830 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4831 - SSL_CTX_set_tlsext_servername_callback()
4832 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4833 - SSL_CTX_set_tlsext_servername_arg()
4834 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4836 openssl s_client has a new '-servername ...' option.
4838 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4839 '-key2 ...', '-servername_fatal' (subject to change). This allows
4840 testing the HostName extension for a specific single host name ('-cert'
4841 and '-key' remain fallbacks for handshakes without HostName
4842 negotiation). If the unrecognized_name alert has to be sent, this by
4843 default is a warning; it becomes fatal with the '-servername_fatal'
4846 [Peter Sylvester, Remy Allais, Christophe Renou]
4848 *) Whirlpool hash implementation is added.
4851 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4852 bn(64,32). Because of instruction set limitations it doesn't have
4853 any negative impact on performance. This was done mostly in order
4854 to make it possible to share assembler modules, such as bn_mul_mont
4855 implementations, between 32- and 64-bit builds without hassle.
4858 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4859 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4863 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4864 dedicated Montgomery multiplication procedure, is introduced.
4865 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4866 "64-bit" performance on certain 32-bit targets.
4869 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4870 in SSL structures. New SSL ctrl to set maximum send fragment size.
4871 Save memory by setting the I/O buffer sizes dynamically instead of
4872 using the maximum available value.
4875 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
4876 in addition to the text details.
4879 *) Very, very preliminary EXPERIMENTAL support for printing of general
4880 ASN1 structures. This currently produces rather ugly output and doesn't
4881 handle several customised structures at all.
4884 *) Integrated support for PVK file format and some related formats such
4885 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4886 these in the 'rsa' and 'dsa' utilities.
4889 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4892 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4893 place for the (very old) "NETSCAPE" format certificates which are now
4894 handled using new ASN1 code equivalents.
4897 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4898 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4899 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4902 *) Modify CRL distribution points extension code to print out previously
4903 unsupported fields. Enhance extension setting code to allow setting of
4907 *) Add print and set support for Issuing Distribution Point CRL extension.
4910 *) Change 'Configure' script to enable Camellia by default.
4913 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4915 *) When rejecting SSL/TLS records due to an incorrect version number, never
4916 update s->server with a new major version number. As of
4917 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
4918 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
4919 the previous behavior could result in a read attempt at NULL when
4920 receiving specific incorrect SSL/TLS records once record payload
4921 protection is active. (CVE-2010-0740)
4922 [Bodo Moeller, Adam Langley <agl@chromium.org>]
4924 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
4925 could be crashed if the relevant tables were not present (e.g. chrooted).
4926 [Tomas Hoger <thoger@redhat.com>]
4928 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
4930 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
4931 [Martin Olsson, Neel Mehta]
4933 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
4934 accommodate for stack sorting, always a write lock!).
4937 *) On some versions of WIN32 Heap32Next is very slow. This can cause
4938 excessive delays in the RAND_poll(): over a minute. As a workaround
4939 include a time check in the inner Heap32Next loop too.
4942 *) The code that handled flushing of data in SSL/TLS originally used the
4943 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
4944 the problem outlined in PR#1949. The fix suggested there however can
4945 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4946 of Apache). So instead simplify the code to flush unconditionally.
4947 This should be fine since flushing with no data to flush is a no op.
4950 *) Handle TLS versions 2.0 and later properly and correctly use the
4951 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4952 off ancient servers have a habit of sticking around for a while...
4955 *) Modify compression code so it frees up structures without using the
4956 ex_data callbacks. This works around a problem where some applications
4957 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4958 restarting) then use compression (e.g. SSL with compression) later.
4959 This results in significant per-connection memory leaks and
4960 has caused some security issues including CVE-2008-1678 and
4964 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4965 change when encrypting or decrypting.
4968 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4969 connect and renegotiate with servers which do not support RI.
4970 Until RI is more widely deployed this option is enabled by default.
4973 *) Add "missing" ssl ctrls to clear options and mode.
4976 *) If client attempts to renegotiate and doesn't support RI respond with
4977 a no_renegotiation alert as required by RFC5746. Some renegotiating
4978 TLS clients will continue a connection gracefully when they receive
4979 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4980 waiting for a server hello which it will never receive. Now we treat a
4981 received no_renegotiation alert as a fatal error. This is because
4982 applications requesting a renegotiation might well expect it to succeed
4983 and would have no code in place to handle the server denying it so the
4984 only safe thing to do is to terminate the connection.
4987 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4988 peer supports secure renegotiation and 0 otherwise. Print out peer
4989 renegotiation support in s_client/s_server.
4992 *) Replace the highly broken and deprecated SPKAC certification method with
4993 the updated NID creation version. This should correctly handle UTF8.
4996 *) Implement RFC5746. Re-enable renegotiation but require the extension
4997 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
4998 turns out to be a bad idea. It has been replaced by
4999 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
5000 SSL_CTX_set_options(). This is really not recommended unless you
5001 know what you are doing.
5002 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
5004 *) Fixes to stateless session resumption handling. Use initial_ctx when
5005 issuing and attempting to decrypt tickets in case it has changed during
5006 servername handling. Use a non-zero length session ID when attempting
5007 stateless session resumption: this makes it possible to determine if
5008 a resumption has occurred immediately after receiving server hello
5009 (several places in OpenSSL subtly assume this) instead of later in
5013 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
5014 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
5015 fixes for a few places where the return code is not checked
5017 [Julia Lawall <julia@diku.dk>]
5019 *) Add --strict-warnings option to Configure script to include devteam
5020 warnings in other configurations.
5023 *) Add support for --libdir option and LIBDIR variable in makefiles. This
5024 makes it possible to install openssl libraries in locations which
5025 have names other than "lib", for example "/usr/lib64" which some
5027 [Steve Henson, based on patch from Jeremy Utley]
5029 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
5030 X690 8.9.12 and can produce some misleading textual output of OIDs.
5031 [Steve Henson, reported by Dan Kaminsky]
5033 *) Delete MD2 from algorithm tables. This follows the recommendation in
5034 several standards that it is not used in new applications due to
5035 several cryptographic weaknesses. For binary compatibility reasons
5036 the MD2 API is still compiled in by default.
5039 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
5043 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
5044 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
5046 [Guenter <lists@gknw.net>]
5048 *) Fix the server certificate chain building code to use X509_verify_cert(),
5049 it used to have an ad-hoc builder which was unable to cope with anything
5050 other than a simple chain.
5051 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
5053 *) Don't check self signed certificate signatures in X509_verify_cert()
5054 by default (a flag can override this): it just wastes time without
5055 adding any security. As a useful side effect self signed root CAs
5056 with non-FIPS digests are now usable in FIPS mode.
5059 *) In dtls1_process_out_of_seq_message() the check if the current message
5060 is already buffered was missing. For every new message was memory
5061 allocated, allowing an attacker to perform an denial of service attack
5062 with sending out of seq handshake messages until there is no memory
5063 left. Additionally every future message was buffered, even if the
5064 sequence number made no sense and would be part of another handshake.
5065 So only messages with sequence numbers less than 10 in advance will be
5066 buffered. (CVE-2009-1378)
5067 [Robin Seggelmann, discovered by Daniel Mentz]
5069 *) Records are buffered if they arrive with a future epoch to be
5070 processed after finishing the corresponding handshake. There is
5071 currently no limitation to this buffer allowing an attacker to perform
5072 a DOS attack with sending records with future epochs until there is no
5073 memory left. This patch adds the pqueue_size() function to determine
5074 the size of a buffer and limits the record buffer to 100 entries.
5076 [Robin Seggelmann, discovered by Daniel Mentz]
5078 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
5079 parent structure is freed. (CVE-2009-1379)
5082 *) Handle non-blocking I/O properly in SSL_shutdown() call.
5083 [Darryl Miles <darryl-mailinglists@netbauds.net>]
5086 [Ilya O. <vrghost@gmail.com>]
5088 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
5090 *) Disable renegotiation completely - this fixes a severe security
5091 problem (CVE-2009-3555) at the cost of breaking all
5092 renegotiation. Renegotiation can be re-enabled by setting
5093 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
5094 run-time. This is really not recommended unless you know what
5098 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
5100 *) Don't set val to NULL when freeing up structures, it is freed up by
5101 underlying code. If sizeof(void *) > sizeof(long) this can result in
5102 zeroing past the valid field. (CVE-2009-0789)
5103 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
5105 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
5106 checked correctly. This would allow some invalid signed attributes to
5107 appear to verify correctly. (CVE-2009-0591)
5108 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5110 *) Reject UniversalString and BMPString types with invalid lengths. This
5111 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
5112 a legal length. (CVE-2009-0590)
5115 *) Set S/MIME signing as the default purpose rather than setting it
5116 unconditionally. This allows applications to override it at the store
5120 *) Permit restricted recursion of ASN1 strings. This is needed in practice
5121 to handle some structures.
5124 *) Improve efficiency of mem_gets: don't search whole buffer each time
5126 [Jeremy Shapiro <jnshapir@us.ibm.com>]
5128 *) New -hex option for openssl rand.
5131 *) Print out UTF8String and NumericString when parsing ASN1.
5134 *) Support NumericString type for name components.
5137 *) Allow CC in the environment to override the automatically chosen
5138 compiler. Note that nothing is done to ensure flags work with the
5142 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
5144 *) Properly check EVP_VerifyFinal() and similar return values
5146 [Ben Laurie, Bodo Moeller, Google Security Team]
5148 *) Enable TLS extensions by default.
5151 *) Allow the CHIL engine to be loaded, whether the application is
5152 multithreaded or not. (This does not release the developer from the
5153 obligation to set up the dynamic locking callbacks.)
5154 [Sander Temme <sander@temme.net>]
5156 *) Use correct exit code if there is an error in dgst command.
5157 [Steve Henson; problem pointed out by Roland Dirlewanger]
5159 *) Tweak Configure so that you need to say "experimental-jpake" to enable
5160 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
5163 *) Add experimental JPAKE support, including demo authentication in
5164 s_client and s_server.
5167 *) Set the comparison function in v3_addr_canonize().
5168 [Rob Austein <sra@hactrn.net>]
5170 *) Add support for XMPP STARTTLS in s_client.
5171 [Philip Paeps <philip@freebsd.org>]
5173 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
5174 to ensure that even with this option, only ciphersuites in the
5175 server's preference list will be accepted. (Note that the option
5176 applies only when resuming a session, so the earlier behavior was
5177 just about the algorithm choice for symmetric cryptography.)
5180 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
5182 *) Fix NULL pointer dereference if a DTLS server received
5183 ChangeCipherSpec as first record (CVE-2009-1386).
5186 *) Fix a state transition in s3_srvr.c and d1_srvr.c
5187 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
5190 *) The fix in 0.9.8c that supposedly got rid of unsafe
5191 double-checked locking was incomplete for RSA blinding,
5192 addressing just one layer of what turns out to have been
5193 doubly unsafe triple-checked locking.
5195 So now fix this for real by retiring the MONT_HELPER macro
5196 in crypto/rsa/rsa_eay.c.
5198 [Bodo Moeller; problem pointed out by Marius Schilder]
5200 *) Various precautionary measures:
5202 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
5204 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
5205 (NB: This would require knowledge of the secret session ticket key
5206 to exploit, in which case you'd be SOL either way.)
5208 - Change bn_nist.c so that it will properly handle input BIGNUMs
5209 outside the expected range.
5211 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
5214 [Neel Mehta, Bodo Moeller]
5216 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
5217 the load fails. Useful for distros.
5218 [Ben Laurie and the FreeBSD team]
5220 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
5223 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
5226 *) Expand ENGINE to support engine supplied SSL client certificate functions.
5228 This work was sponsored by Logica.
5231 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
5232 keystores. Support for SSL/TLS client authentication too.
5233 Not compiled unless enable-capieng specified to Configure.
5235 This work was sponsored by Logica.
5238 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
5239 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
5240 attribute creation routines such as certificate requests and PKCS#12
5244 Changes between 0.9.8g and 0.9.8h [28 May 2008]
5246 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
5247 handshake which could lead to a client crash as found using the
5248 Codenomicon TLS test suite (CVE-2008-1672)
5249 [Steve Henson, Mark Cox]
5251 *) Fix double free in TLS server name extensions which could lead to
5252 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
5255 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
5257 Clear the error queue to ensure that error entries left from
5258 older function calls do not interfere with the correct operation.
5259 [Lutz Jaenicke, Erik de Castro Lopo]
5261 *) Remove root CA certificates of commercial CAs:
5263 The OpenSSL project does not recommend any specific CA and does not
5264 have any policy with respect to including or excluding any CA.
5265 Therefore it does not make any sense to ship an arbitrary selection
5266 of root CA certificates with the OpenSSL software.
5269 *) RSA OAEP patches to fix two separate invalid memory reads.
5270 The first one involves inputs when 'lzero' is greater than
5271 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
5272 before the beginning of from). The second one involves inputs where
5273 the 'db' section contains nothing but zeroes (there is a one-byte
5274 invalid read after the end of 'db').
5275 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5277 *) Partial backport from 0.9.9-dev:
5279 Introduce bn_mul_mont (dedicated Montgomery multiplication
5280 procedure) as a candidate for BIGNUM assembler implementation.
5281 While 0.9.9-dev uses assembler for various architectures, only
5282 x86_64 is available by default here in the 0.9.8 branch, and
5283 32-bit x86 is available through a compile-time setting.
5285 To try the 32-bit x86 assembler implementation, use Configure
5286 option "enable-montasm" (which exists only for this backport).
5288 As "enable-montasm" for 32-bit x86 disclaims code stability
5289 anyway, in this constellation we activate additional code
5290 backported from 0.9.9-dev for further performance improvements,
5291 namely BN_from_montgomery_word. (To enable this otherwise,
5292 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
5294 [Andy Polyakov (backport partially by Bodo Moeller)]
5296 *) Add TLS session ticket callback. This allows an application to set
5297 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
5298 values. This is useful for key rollover for example where several key
5299 sets may exist with different names.
5302 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
5303 This was broken until now in 0.9.8 releases, such that the only way
5304 a registered ENGINE could be used (assuming it initialises
5305 successfully on the host) was to explicitly set it as the default
5306 for the relevant algorithms. This is in contradiction with 0.9.7
5307 behaviour and the documentation. With this fix, when an ENGINE is
5308 registered into a given algorithm's table of implementations, the
5309 'uptodate' flag is reset so that auto-discovery will be used next
5310 time a new context for that algorithm attempts to select an
5312 [Ian Lister (tweaked by Geoff Thorpe)]
5314 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5315 implementation in the following ways:
5317 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5320 Lack of BER streaming support means one pass streaming processing is
5321 only supported if data is detached: setting the streaming flag is
5322 ignored for embedded content.
5324 CMS support is disabled by default and must be explicitly enabled
5325 with the enable-cms configuration option.
5328 *) Update the GMP engine glue to do direct copies between BIGNUM and
5329 mpz_t when openssl and GMP use the same limb size. Otherwise the
5330 existing "conversion via a text string export" trick is still used.
5331 [Paul Sheer <paulsheer@gmail.com>]
5333 *) Zlib compression BIO. This is a filter BIO which compressed and
5334 uncompresses any data passed through it.
5337 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5338 RFC3394 compatible AES key wrapping.
5341 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5342 sets string data without copying. X509_ALGOR_set0() and
5343 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5344 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5345 from an X509_ATTRIBUTE structure optionally checking it occurs only
5346 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5350 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5351 to get the expected BN_FLG_CONSTTIME behavior.
5352 [Bodo Moeller (Google)]
5356 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5357 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5358 - added some more tests to do_tests.pl
5359 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5360 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5361 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5362 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5363 - various changes to netware.pl to enable gcc-cross builds on Win32
5365 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5366 - various changes to fix missing prototype warnings
5367 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5368 - added AES, WHIRLPOOL and CPUID assembler code to build files
5369 - added missing AES assembler make rules to mk1mf.pl
5370 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5371 [Guenter Knauf <eflash@gmx.net>]
5373 *) Implement certificate status request TLS extension defined in RFC3546.
5374 A client can set the appropriate parameters and receive the encoded
5375 OCSP response via a callback. A server can query the supplied parameters
5376 and set the encoded OCSP response in the callback. Add simplified examples
5377 to s_client and s_server.
5380 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5382 *) Fix various bugs:
5383 + Binary incompatibility of ssl_ctx_st structure
5384 + DTLS interoperation with non-compliant servers
5385 + Don't call get_session_cb() without proposed session
5386 + Fix ia64 assembler code
5387 [Andy Polyakov, Steve Henson]
5389 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5391 *) DTLS Handshake overhaul. There were longstanding issues with
5392 OpenSSL DTLS implementation, which were making it impossible for
5393 RFC 4347 compliant client to communicate with OpenSSL server.
5394 Unfortunately just fixing these incompatibilities would "cut off"
5395 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5396 server keeps tolerating non RFC compliant syntax. The opposite is
5397 not true, 0.9.8f client can not communicate with earlier server.
5398 This update even addresses CVE-2007-4995.
5401 *) Changes to avoid need for function casts in OpenSSL: some compilers
5402 (gcc 4.2 and later) reject their use.
5403 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5406 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5407 RFC4507bis. The encrypted ticket format is an encrypted encoded
5408 SSL_SESSION structure, that way new session features are automatically
5411 If a client application caches session in an SSL_SESSION structure
5412 support is transparent because tickets are now stored in the encoded
5415 The SSL_CTX structure automatically generates keys for ticket
5416 protection in servers so again support should be possible
5417 with no application modification.
5419 If a client or server wishes to disable RFC4507 support then the option
5420 SSL_OP_NO_TICKET can be set.
5422 Add a TLS extension debugging callback to allow the contents of any client
5423 or server extensions to be examined.
5425 This work was sponsored by Google.
5428 *) Add initial support for TLS extensions, specifically for the server_name
5429 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5430 have new members for a host name. The SSL data structure has an
5431 additional member SSL_CTX *initial_ctx so that new sessions can be
5432 stored in that context to allow for session resumption, even after the
5433 SSL has been switched to a new SSL_CTX in reaction to a client's
5434 server_name extension.
5436 New functions (subject to change):
5438 SSL_get_servername()
5439 SSL_get_servername_type()
5442 New CTRL codes and macros (subject to change):
5444 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5445 - SSL_CTX_set_tlsext_servername_callback()
5446 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5447 - SSL_CTX_set_tlsext_servername_arg()
5448 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5450 openssl s_client has a new '-servername ...' option.
5452 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5453 '-key2 ...', '-servername_fatal' (subject to change). This allows
5454 testing the HostName extension for a specific single host name ('-cert'
5455 and '-key' remain fallbacks for handshakes without HostName
5456 negotiation). If the unrecognized_name alert has to be sent, this by
5457 default is a warning; it becomes fatal with the '-servername_fatal'
5460 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
5462 *) Add AES and SSE2 assembly language support to VC++ build.
5465 *) Mitigate attack on final subtraction in Montgomery reduction.
5468 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
5469 (which previously caused an internal error).
5472 *) Squeeze another 10% out of IGE mode when in != out.
5475 *) AES IGE mode speedup.
5476 [Dean Gaudet (Google)]
5478 *) Add the Korean symmetric 128-bit cipher SEED (see
5479 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
5480 add SEED ciphersuites from RFC 4162:
5482 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
5483 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
5484 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
5485 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
5487 To minimize changes between patchlevels in the OpenSSL 0.9.8
5488 series, SEED remains excluded from compilation unless OpenSSL
5489 is configured with 'enable-seed'.
5490 [KISA, Bodo Moeller]
5492 *) Mitigate branch prediction attacks, which can be practical if a
5493 single processor is shared, allowing a spy process to extract
5494 information. For detailed background information, see
5495 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
5496 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
5497 and Necessary Software Countermeasures"). The core of the change
5498 are new versions BN_div_no_branch() and
5499 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
5500 respectively, which are slower, but avoid the security-relevant
5501 conditional branches. These are automatically called by BN_div()
5502 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
5503 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
5504 remove a conditional branch.
5506 BN_FLG_CONSTTIME is the new name for the previous
5507 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
5508 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
5509 in the exponent causes BN_mod_exp_mont() to use the alternative
5510 implementation in BN_mod_exp_mont_consttime().) The old name
5511 remains as a deprecated alias.
5513 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
5514 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
5515 constant-time implementations for more than just exponentiation.
5516 Here too the old name is kept as a deprecated alias.
5518 BN_BLINDING_new() will now use BN_dup() for the modulus so that
5519 the BN_BLINDING structure gets an independent copy of the
5520 modulus. This means that the previous "BIGNUM *m" argument to
5521 BN_BLINDING_new() and to BN_BLINDING_create_param() now
5522 essentially becomes "const BIGNUM *m", although we can't actually
5523 change this in the header file before 0.9.9. It allows
5524 RSA_setup_blinding() to use BN_with_flags() on the modulus to
5525 enable BN_FLG_CONSTTIME.
5527 [Matthew D Wood (Intel Corp)]
5529 *) In the SSL/TLS server implementation, be strict about session ID
5530 context matching (which matters if an application uses a single
5531 external cache for different purposes). Previously,
5532 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
5533 set. This did ensure strict client verification, but meant that,
5534 with applications using a single external cache for quite
5535 different requirements, clients could circumvent ciphersuite
5536 restrictions for a given session ID context by starting a session
5537 in a different context.
5540 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5541 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5542 authentication-only ciphersuites.
5545 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
5546 not complete and could lead to a possible single byte overflow
5547 (CVE-2007-5135) [Ben Laurie]
5549 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
5551 *) Since AES128 and AES256 (and similarly Camellia128 and
5552 Camellia256) share a single mask bit in the logic of
5553 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5554 kludge to work properly if AES128 is available and AES256 isn't
5555 (or if Camellia128 is available and Camellia256 isn't).
5558 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
5559 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
5560 When a point or a seed is encoded in a BIT STRING, we need to
5561 prevent the removal of trailing zero bits to get the proper DER
5562 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
5563 of a NamedBitList, for which trailing 0 bits need to be removed.)
5566 *) Have SSL/TLS server implementation tolerate "mismatched" record
5567 protocol version while receiving ClientHello even if the
5568 ClientHello is fragmented. (The server can't insist on the
5569 particular protocol version it has chosen before the ServerHello
5570 message has informed the client about his choice.)
5573 *) Add RFC 3779 support.
5574 [Rob Austein for ARIN, Ben Laurie]
5576 *) Load error codes if they are not already present instead of using a
5577 static variable. This allows them to be cleanly unloaded and reloaded.
5578 Improve header file function name parsing.
5581 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
5582 or CAPABILITY handshake as required by RFCs.
5585 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5587 *) Introduce limits to prevent malicious keys being able to
5588 cause a denial of service. (CVE-2006-2940)
5589 [Steve Henson, Bodo Moeller]
5591 *) Fix ASN.1 parsing of certain invalid structures that can result
5592 in a denial of service. (CVE-2006-2937) [Steve Henson]
5594 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5595 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5597 *) Fix SSL client code which could crash if connecting to a
5598 malicious SSLv2 server. (CVE-2006-4343)
5599 [Tavis Ormandy and Will Drewry, Google Security Team]
5601 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
5602 match only those. Before that, "AES256-SHA" would be interpreted
5603 as a pattern and match "AES128-SHA" too (since AES128-SHA got
5604 the same strength classification in 0.9.7h) as we currently only
5605 have a single AES bit in the ciphersuite description bitmap.
5606 That change, however, also applied to ciphersuite strings such as
5607 "RC4-MD5" that intentionally matched multiple ciphersuites --
5608 namely, SSL 2.0 ciphersuites in addition to the more common ones
5609 from SSL 3.0/TLS 1.0.
5611 So we change the selection algorithm again: Naming an explicit
5612 ciphersuite selects this one ciphersuite, and any other similar
5613 ciphersuite (same bitmap) from *other* protocol versions.
5614 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5615 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5617 Since SSL 2.0 does not have any ciphersuites for which the
5618 128/256 bit distinction would be relevant, this works for now.
5619 The proper fix will be to use different bits for AES128 and
5620 AES256, which would have avoided the problems from the beginning;
5621 however, bits are scarce, so we can only do this in a new release
5622 (not just a patchlevel) when we can change the SSL_CIPHER
5623 definition to split the single 'unsigned long mask' bitmap into
5624 multiple values to extend the available space.
5628 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5630 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5631 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5633 *) Add AES IGE and biIGE modes.
5636 *) Change the Unix randomness entropy gathering to use poll() when
5637 possible instead of select(), since the latter has some
5638 undesirable limitations.
5639 [Darryl Miles via Richard Levitte and Bodo Moeller]
5641 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
5642 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
5643 cannot be implicitly activated as part of, e.g., the "AES" alias.
5644 However, please upgrade to OpenSSL 0.9.9[-dev] for
5645 non-experimental use of the ECC ciphersuites to get TLS extension
5646 support, which is required for curve and point format negotiation
5647 to avoid potential handshake problems.
5650 *) Disable rogue ciphersuites:
5652 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5653 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5654 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5656 The latter two were purportedly from
5657 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5660 Also deactivate the remaining ciphersuites from
5661 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5662 unofficial, and the ID has long expired.
5665 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5666 dual-core machines) and other potential thread-safety issues.
5669 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
5670 versions), which is now available for royalty-free use
5671 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
5672 Also, add Camellia TLS ciphersuites from RFC 4132.
5674 To minimize changes between patchlevels in the OpenSSL 0.9.8
5675 series, Camellia remains excluded from compilation unless OpenSSL
5676 is configured with 'enable-camellia'.
5679 *) Disable the padding bug check when compression is in use. The padding
5680 bug check assumes the first packet is of even length, this is not
5681 necessarily true if compression is enabled and can result in false
5682 positives causing handshake failure. The actual bug test is ancient
5683 code so it is hoped that implementations will either have fixed it by
5684 now or any which still have the bug do not support compression.
5687 Changes between 0.9.8a and 0.9.8b [04 May 2006]
5689 *) When applying a cipher rule check to see if string match is an explicit
5690 cipher suite and only match that one cipher suite if it is.
5693 *) Link in manifests for VC++ if needed.
5694 [Austin Ziegler <halostatue@gmail.com>]
5696 *) Update support for ECC-based TLS ciphersuites according to
5697 draft-ietf-tls-ecc-12.txt with proposed changes (but without
5698 TLS extensions, which are supported starting with the 0.9.9
5699 branch, not in the OpenSSL 0.9.8 branch).
5702 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
5703 opaque EVP_CIPHER_CTX handling.
5706 *) Fixes and enhancements to zlib compression code. We now only use
5707 "zlib1.dll" and use the default __cdecl calling convention on Win32
5708 to conform with the standards mentioned here:
5709 http://www.zlib.net/DLL_FAQ.txt
5710 Static zlib linking now works on Windows and the new --with-zlib-include
5711 --with-zlib-lib options to Configure can be used to supply the location
5712 of the headers and library. Gracefully handle case where zlib library
5716 *) Several fixes and enhancements to the OID generation code. The old code
5717 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
5718 handle numbers larger than ULONG_MAX, truncated printing and had a
5719 non standard OBJ_obj2txt() behaviour.
5722 *) Add support for building of engines under engine/ as shared libraries
5723 under VC++ build system.
5726 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
5727 Hopefully, we will not see any false combination of paths any more.
5730 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5732 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5733 (part of SSL_OP_ALL). This option used to disable the
5734 countermeasure against man-in-the-middle protocol-version
5735 rollback in the SSL 2.0 server implementation, which is a bad
5736 idea. (CVE-2005-2969)
5738 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5739 for Information Security, National Institute of Advanced Industrial
5740 Science and Technology [AIST], Japan)]
5742 *) Add two function to clear and return the verify parameter flags.
5745 *) Keep cipherlists sorted in the source instead of sorting them at
5746 runtime, thus removing the need for a lock.
5749 *) Avoid some small subgroup attacks in Diffie-Hellman.
5750 [Nick Mathewson and Ben Laurie]
5752 *) Add functions for well-known primes.
5755 *) Extended Windows CE support.
5756 [Satoshi Nakamura and Andy Polyakov]
5758 *) Initialize SSL_METHOD structures at compile time instead of during
5759 runtime, thus removing the need for a lock.
5762 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5763 attempting to decrypt each encrypted key in turn. Add support to
5767 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5769 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5772 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5775 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5776 key into the same file any more.
5779 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5782 *) Add -utf8 command line and config file option to 'ca'.
5783 [Stefan <stf@udoma.org]
5785 *) Removed the macro des_crypt(), as it seems to conflict with some
5786 libraries. Use DES_crypt().
5789 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5790 involves renaming the source and generated shared-libs for
5791 both. The engines will accept the corrected or legacy ids
5792 ('ncipher' and '4758_cca' respectively) when binding. NB,
5793 this only applies when building 'shared'.
5794 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5796 *) Add attribute functions to EVP_PKEY structure. Modify
5797 PKCS12_create() to recognize a CSP name attribute and
5798 use it. Make -CSP option work again in pkcs12 utility.
5801 *) Add new functionality to the bn blinding code:
5802 - automatic re-creation of the BN_BLINDING parameters after
5803 a fixed number of uses (currently 32)
5804 - add new function for parameter creation
5805 - introduce flags to control the update behaviour of the
5806 BN_BLINDING parameters
5807 - hide BN_BLINDING structure
5808 Add a second BN_BLINDING slot to the RSA structure to improve
5809 performance when a single RSA object is shared among several
5813 *) Add support for DTLS.
5814 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5816 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5817 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5820 *) Remove buggy and incomplete DH cert support from
5821 ssl/ssl_rsa.c and ssl/s3_both.c
5824 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5825 the apps/openssl applications.
5828 *) Compile clean with "-Wall -Wmissing-prototypes
5829 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5830 DEBUG_SAFESTACK must also be set.
5833 *) Change ./Configure so that certain algorithms can be disabled by default.
5834 The new counterpiece to "no-xxx" is "enable-xxx".
5836 The patented RC5 and MDC2 algorithms will now be disabled unless
5837 "enable-rc5" and "enable-mdc2", respectively, are specified.
5839 (IDEA remains enabled despite being patented. This is because IDEA
5840 is frequently required for interoperability, and there is no license
5841 fee for non-commercial use. As before, "no-idea" can be used to
5842 avoid this algorithm.)
5846 *) Add processing of proxy certificates (see RFC 3820). This work was
5847 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5848 EGEE (Enabling Grids for E-science in Europe).
5851 *) RC4 performance overhaul on modern architectures/implementations, such
5852 as Intel P4, IA-64 and AMD64.
5855 *) New utility extract-section.pl. This can be used specify an alternative
5856 section number in a pod file instead of having to treat each file as
5857 a separate case in Makefile. This can be done by adding two lines to the
5860 =for comment openssl_section:XXX
5862 The blank line is mandatory.
5866 *) New arguments -certform, -keyform and -pass for s_client and s_server
5867 to allow alternative format key and certificate files and passphrase
5871 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5872 update associated structures and add various utility functions.
5874 Add new policy related verify parameters, include policy checking in
5875 standard verify code. Enhance 'smime' application with extra parameters
5876 to support policy checking and print out.
5879 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
5880 Nehemiah processors. These extensions support AES encryption in hardware
5881 as well as RNG (though RNG support is currently disabled).
5882 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
5884 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
5887 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
5888 [Andy Polyakov and a number of other people]
5890 *) Improved PowerPC platform support. Most notably BIGNUM assembler
5891 implementation contributed by IBM.
5892 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
5894 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
5895 exponent rather than 'unsigned long'. There is a corresponding change to
5896 the new 'rsa_keygen' element of the RSA_METHOD structure.
5897 [Jelte Jansen, Geoff Thorpe]
5899 *) Functionality for creating the initial serial number file is now
5900 moved from CA.pl to the 'ca' utility with a new option -create_serial.
5902 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
5903 number file to 1, which is bound to cause problems. To avoid
5904 the problems while respecting compatibility between different 0.9.7
5905 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
5906 CA.pl for serial number initialization. With the new release 0.9.8,
5907 we can fix the problem directly in the 'ca' utility.)
5910 *) Reduced header interdependencies by declaring more opaque objects in
5911 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
5912 give fewer recursive includes, which could break lazy source code - so
5913 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
5914 developers should define this symbol when building and using openssl to
5915 ensure they track the recommended behaviour, interfaces, [etc], but
5916 backwards-compatible behaviour prevails when this isn't defined.
5919 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
5922 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
5923 This will generate a random key of the appropriate length based on the
5924 cipher context. The EVP_CIPHER can provide its own random key generation
5925 routine to support keys of a specific form. This is used in the des and
5926 3des routines to generate a key of the correct parity. Update S/MIME
5927 code to use new functions and hence generate correct parity DES keys.
5928 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
5929 valid (weak or incorrect parity).
5932 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
5933 as looking them up. This is useful when the verified structure may contain
5934 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
5935 present unless the new PKCS7_NO_CRL flag is asserted.
5938 *) Extend ASN1 oid configuration module. It now additionally accepts the
5941 shortName = some long name, 1.2.3.4
5944 *) Reimplemented the BN_CTX implementation. There is now no more static
5945 limitation on the number of variables it can handle nor the depth of the
5946 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
5947 information can now expand as required, and rather than having a single
5948 static array of bignums, BN_CTX now uses a linked-list of such arrays
5949 allowing it to expand on demand whilst maintaining the usefulness of
5950 BN_CTX's "bundling".
5953 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5954 to allow all RSA operations to function using a single BN_CTX.
5957 *) Preliminary support for certificate policy evaluation and checking. This
5958 is initially intended to pass the tests outlined in "Conformance Testing
5959 of Relying Party Client Certificate Path Processing Logic" v1.07.
5962 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
5963 remained unused and not that useful. A variety of other little bignum
5964 tweaks and fixes have also been made continuing on from the audit (see
5968 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
5969 associated ASN1, EVP and SSL functions and old ASN1 macros.
5972 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
5973 and this should never fail. So the return value from the use of
5974 BN_set_word() (which can fail due to needless expansion) is now deprecated;
5975 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
5978 *) BN_CTX_get() should return zero-valued bignums, providing the same
5979 initialised value as BN_new().
5980 [Geoff Thorpe, suggested by Ulf Möller]
5982 *) Support for inhibitAnyPolicy certificate extension.
5985 *) An audit of the BIGNUM code is underway, for which debugging code is
5986 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
5987 is considered valid when processing BIGNUMs, and causes execution to
5988 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
5989 further steps are taken to deliberately pollute unused data in BIGNUM
5990 structures to try and expose faulty code further on. For now, openssl will
5991 (in its default mode of operation) continue to tolerate the inconsistent
5992 forms that it has tolerated in the past, but authors and packagers should
5993 consider trying openssl and their own applications when compiled with
5994 these debugging symbols defined. It will help highlight potential bugs in
5995 their own code, and will improve the test coverage for OpenSSL itself. At
5996 some point, these tighter rules will become openssl's default to improve
5997 maintainability, though the assert()s and other overheads will remain only
5998 in debugging configurations. See bn.h for more details.
5999 [Geoff Thorpe, Nils Larsch, Ulf Möller]
6001 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
6002 that can only be obtained through BN_CTX_new() (which implicitly
6003 initialises it). The presence of this function only made it possible
6004 to overwrite an existing structure (and cause memory leaks).
6007 *) Because of the callback-based approach for implementing LHASH as a
6008 template type, lh_insert() adds opaque objects to hash-tables and
6009 lh_doall() or lh_doall_arg() are typically used with a destructor callback
6010 to clean up those corresponding objects before destroying the hash table
6011 (and losing the object pointers). So some over-zealous constifications in
6012 LHASH have been relaxed so that lh_insert() does not take (nor store) the
6013 objects as "const" and the lh_doall[_arg] callback wrappers are not
6014 prototyped to have "const" restrictions on the object pointers they are
6015 given (and so aren't required to cast them away any more).
6018 *) The tmdiff.h API was so ugly and minimal that our own timing utility
6019 (speed) prefers to use its own implementation. The two implementations
6020 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
6021 its object type properly exposed (MS_TM) instead of casting to/from "char
6022 *". This may still change yet if someone realises MS_TM and "ms_time_***"
6023 aren't necessarily the greatest nomenclatures - but this is what was used
6024 internally to the implementation so I've used that for now.
6027 *) Ensure that deprecated functions do not get compiled when
6028 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
6029 the self-tests were still using deprecated key-generation functions so
6030 these have been updated also.
6033 *) Reorganise PKCS#7 code to separate the digest location functionality
6034 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
6035 New function PKCS7_set_digest() to set the digest type for PKCS#7
6036 digestedData type. Add additional code to correctly generate the
6037 digestedData type and add support for this type in PKCS7 initialization
6041 *) New function PKCS7_set0_type_other() this initializes a PKCS7
6042 structure of type "other".
6045 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
6046 sure the loop does correctly stop and breaking ("division by zero")
6047 modulus operations are not performed. The (pre-generated) prime
6048 table crypto/bn/bn_prime.h was already correct, but it could not be
6049 re-generated on some platforms because of the "division by zero"
6050 situation in the script.
6051 [Ralf S. Engelschall]
6053 *) Update support for ECC-based TLS ciphersuites according to
6054 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
6055 SHA-1 now is only used for "small" curves (where the
6056 representation of a field element takes up to 24 bytes); for
6057 larger curves, the field element resulting from ECDH is directly
6058 used as premaster secret.
6059 [Douglas Stebila (Sun Microsystems Laboratories)]
6061 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
6062 curve secp160r1 to the tests.
6063 [Douglas Stebila (Sun Microsystems Laboratories)]
6065 *) Add the possibility to load symbols globally with DSO.
6066 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
6068 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
6069 control of the error stack.
6072 *) Add support for STORE in ENGINE.
6075 *) Add the STORE type. The intention is to provide a common interface
6076 to certificate and key stores, be they simple file-based stores, or
6077 HSM-type store, or LDAP stores, or...
6078 NOTE: The code is currently UNTESTED and isn't really used anywhere.
6081 *) Add a generic structure called OPENSSL_ITEM. This can be used to
6082 pass a list of arguments to any function as well as provide a way
6083 for a function to pass data back to the caller.
6086 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
6087 works like BUF_strdup() but can be used to duplicate a portion of
6088 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
6092 *) Add the function sk_find_ex() which works like sk_find(), but will
6093 return an index to an element even if an exact match couldn't be
6094 found. The index is guaranteed to point at the element where the
6095 searched-for key would be inserted to preserve sorting order.
6098 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
6099 takes an extra flags argument for optional functionality. Currently,
6100 the following flags are defined:
6102 OBJ_BSEARCH_VALUE_ON_NOMATCH
6103 This one gets OBJ_bsearch_ex() to return a pointer to the first
6104 element where the comparing function returns a negative or zero
6107 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
6108 This one gets OBJ_bsearch_ex() to return a pointer to the first
6109 element where the comparing function returns zero. This is useful
6110 if there are more than one element where the comparing function
6114 *) Make it possible to create self-signed certificates with 'openssl ca'
6115 in such a way that the self-signed certificate becomes part of the
6116 CA database and uses the same mechanisms for serial number generation
6117 as all other certificate signing. The new flag '-selfsign' enables
6118 this functionality. Adapt CA.sh and CA.pl.in.
6121 *) Add functionality to check the public key of a certificate request
6122 against a given private. This is useful to check that a certificate
6123 request can be signed by that key (self-signing).
6126 *) Make it possible to have multiple active certificates with the same
6127 subject in the CA index file. This is done only if the keyword
6128 'unique_subject' is set to 'no' in the main CA section (default
6129 if 'CA_default') of the configuration file. The value is saved
6130 with the database itself in a separate index attribute file,
6131 named like the index file with '.attr' appended to the name.
6134 *) Generate multi-valued AVAs using '+' notation in config files for
6138 *) Support for nameConstraints certificate extension.
6141 *) Support for policyConstraints certificate extension.
6144 *) Support for policyMappings certificate extension.
6147 *) Make sure the default DSA_METHOD implementation only uses its
6148 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
6149 and change its own handlers to be NULL so as to remove unnecessary
6150 indirection. This lets alternative implementations fallback to the
6151 default implementation more easily.
6154 *) Support for directoryName in GeneralName related extensions
6158 *) Make it possible to link applications using Makefile.shared.
6159 Make that possible even when linking against static libraries!
6162 *) Support for single pass processing for S/MIME signing. This now
6163 means that S/MIME signing can be done from a pipe, in addition
6164 cleartext signing (multipart/signed type) is effectively streaming
6165 and the signed data does not need to be all held in memory.
6167 This is done with a new flag PKCS7_STREAM. When this flag is set
6168 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
6169 is done after the data is output (and digests calculated) in
6170 SMIME_write_PKCS7().
6173 *) Add full support for -rpath/-R, both in shared libraries and
6174 applications, at least on the platforms where it's known how
6178 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
6179 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
6180 will now compute a table of multiples of the generator that
6181 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
6182 faster (notably in the case of a single point multiplication,
6183 scalar * generator).
6184 [Nils Larsch, Bodo Moeller]
6186 *) IPv6 support for certificate extensions. The various extensions
6187 which use the IP:a.b.c.d can now take IPv6 addresses using the
6188 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
6192 *) Added an ENGINE that implements RSA by performing private key
6193 exponentiations with the GMP library. The conversions to and from
6194 GMP's mpz_t format aren't optimised nor are any montgomery forms
6195 cached, and on x86 it appears OpenSSL's own performance has caught up.
6196 However there are likely to be other architectures where GMP could
6197 provide a boost. This ENGINE is not built in by default, but it can be
6198 specified at Configure time and should be accompanied by the necessary
6199 linker additions, eg;
6200 ./config -DOPENSSL_USE_GMP -lgmp
6203 *) "openssl engine" will not display ENGINE/DSO load failure errors when
6204 testing availability of engines with "-t" - the old behaviour is
6205 produced by increasing the feature's verbosity with "-tt".
6208 *) ECDSA routines: under certain error conditions uninitialized BN objects
6209 could be freed. Solution: make sure initialization is performed early
6210 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
6214 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
6215 and DH_METHOD (eg. by ENGINE implementations) to override the normal
6216 software implementations. For DSA and DH, parameter generation can
6217 also be overridden by providing the appropriate method callbacks.
6220 *) Change the "progress" mechanism used in key-generation and
6221 primality testing to functions that take a new BN_GENCB pointer in
6222 place of callback/argument pairs. The new API functions have "_ex"
6223 postfixes and the older functions are reimplemented as wrappers for
6224 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
6225 declarations of the old functions to help (graceful) attempts to
6226 migrate to the new functions. Also, the new key-generation API
6227 functions operate on a caller-supplied key-structure and return
6228 success/failure rather than returning a key or NULL - this is to
6229 help make "keygen" another member function of RSA_METHOD etc.
6231 Example for using the new callback interface:
6233 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
6237 BN_GENCB_set(&my_cb, my_callback, my_arg);
6239 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
6240 /* For the meaning of a, b in calls to my_callback(), see the
6241 * documentation of the function that calls the callback.
6242 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
6243 * my_callback should return 1 if it wants BN_is_prime_ex()
6244 * to continue, or 0 to stop.
6249 *) Change the ZLIB compression method to be stateful, and make it
6250 available to TLS with the number defined in
6251 draft-ietf-tls-compression-04.txt.
6254 *) Add the ASN.1 structures and functions for CertificatePair, which
6255 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
6257 CertificatePair ::= SEQUENCE {
6258 forward [0] Certificate OPTIONAL,
6259 reverse [1] Certificate OPTIONAL,
6260 -- at least one of the pair shall be present -- }
6262 Also implement the PEM functions to read and write certificate
6263 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
6265 This needed to be defined, mostly for the sake of the LDAP
6266 attribute crossCertificatePair, but may prove useful elsewhere as
6270 *) Make it possible to inhibit symlinking of shared libraries in
6271 Makefile.shared, for Cygwin's sake.
6274 *) Extend the BIGNUM API by creating a function
6275 void BN_set_negative(BIGNUM *a, int neg);
6276 and a macro that behave like
6277 int BN_is_negative(const BIGNUM *a);
6279 to avoid the need to access 'a->neg' directly in applications.
6282 *) Implement fast modular reduction for pseudo-Mersenne primes
6283 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
6284 EC_GROUP_new_curve_GFp() will now automatically use this
6286 [Nils Larsch <nla@trustcenter.de>]
6288 *) Add new lock type (CRYPTO_LOCK_BN).
6291 *) Change the ENGINE framework to automatically load engines
6292 dynamically from specific directories unless they could be
6293 found to already be built in or loaded. Move all the
6294 current engines except for the cryptodev one to a new
6296 The engines in engines/ are built as shared libraries if
6297 the "shared" options was given to ./Configure or ./config.
6298 Otherwise, they are inserted in libcrypto.a.
6299 /usr/local/ssl/engines is the default directory for dynamic
6300 engines, but that can be overridden at configure time through
6301 the usual use of --prefix and/or --openssldir, and at run
6302 time with the environment variable OPENSSL_ENGINES.
6303 [Geoff Thorpe and Richard Levitte]
6305 *) Add Makefile.shared, a helper makefile to build shared
6306 libraries. Adapt Makefile.org.
6309 *) Add version info to Win32 DLLs.
6310 [Peter 'Luna' Runestig" <peter@runestig.com>]
6312 *) Add new 'medium level' PKCS#12 API. Certificates and keys
6313 can be added using this API to created arbitrary PKCS#12
6314 files while avoiding the low level API.
6316 New options to PKCS12_create(), key or cert can be NULL and
6317 will then be omitted from the output file. The encryption
6318 algorithm NIDs can be set to -1 for no encryption, the mac
6319 iteration count can be set to 0 to omit the mac.
6321 Enhance pkcs12 utility by making the -nokeys and -nocerts
6322 options work when creating a PKCS#12 file. New option -nomac
6323 to omit the mac, NONE can be set for an encryption algorithm.
6324 New code is modified to use the enhanced PKCS12_create()
6325 instead of the low level API.
6328 *) Extend ASN1 encoder to support indefinite length constructed
6329 encoding. This can output sequences tags and octet strings in
6330 this form. Modify pk7_asn1.c to support indefinite length
6331 encoding. This is experimental and needs additional code to
6332 be useful, such as an ASN1 bio and some enhanced streaming
6335 Extend template encode functionality so that tagging is passed
6336 down to the template encoder.
6339 *) Let 'openssl req' fail if an argument to '-newkey' is not
6340 recognized instead of using RSA as a default.
6343 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6344 As these are not official, they are not included in "ALL";
6345 the "ECCdraft" ciphersuite group alias can be used to select them.
6346 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6348 *) Add ECDH engine support.
6349 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6351 *) Add ECDH in new directory crypto/ecdh/.
6352 [Douglas Stebila (Sun Microsystems Laboratories)]
6354 *) Let BN_rand_range() abort with an error after 100 iterations
6355 without success (which indicates a broken PRNG).
6358 *) Change BN_mod_sqrt() so that it verifies that the input value
6359 is really the square of the return value. (Previously,
6360 BN_mod_sqrt would show GIGO behaviour.)
6363 *) Add named elliptic curves over binary fields from X9.62, SECG,
6364 and WAP/WTLS; add OIDs that were still missing.
6366 [Sheueling Chang Shantz and Douglas Stebila
6367 (Sun Microsystems Laboratories)]
6369 *) Extend the EC library for elliptic curves over binary fields
6370 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6373 EC_GF2m_simple_method
6377 EC_GROUP_new_curve_GF2m
6378 EC_GROUP_set_curve_GF2m
6379 EC_GROUP_get_curve_GF2m
6380 EC_POINT_set_affine_coordinates_GF2m
6381 EC_POINT_get_affine_coordinates_GF2m
6382 EC_POINT_set_compressed_coordinates_GF2m
6384 Point compression for binary fields is disabled by default for
6385 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6388 As binary polynomials are represented as BIGNUMs, various members
6389 of the EC_GROUP and EC_POINT data structures can be shared
6390 between the implementations for prime fields and binary fields;
6391 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6392 are essentially identical to their ..._GFp counterparts.
6393 (For simplicity, the '..._GFp' prefix has been dropped from
6394 various internal method names.)
6396 An internal 'field_div' method (similar to 'field_mul' and
6397 'field_sqr') has been added; this is used only for binary fields.
6399 [Sheueling Chang Shantz and Douglas Stebila
6400 (Sun Microsystems Laboratories)]
6402 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6403 through methods ('mul', 'precompute_mult').
6405 The generic implementations (now internally called 'ec_wNAF_mul'
6406 and 'ec_wNAF_precomputed_mult') remain the default if these
6407 methods are undefined.
6409 [Sheueling Chang Shantz and Douglas Stebila
6410 (Sun Microsystems Laboratories)]
6412 *) New function EC_GROUP_get_degree, which is defined through
6413 EC_METHOD. For curves over prime fields, this returns the bit
6414 length of the modulus.
6416 [Sheueling Chang Shantz and Douglas Stebila
6417 (Sun Microsystems Laboratories)]
6419 *) New functions EC_GROUP_dup, EC_POINT_dup.
6420 (These simply call ..._new and ..._copy).
6422 [Sheueling Chang Shantz and Douglas Stebila
6423 (Sun Microsystems Laboratories)]
6425 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6426 Polynomials are represented as BIGNUMs (where the sign bit is not
6427 used) in the following functions [macros]:
6430 BN_GF2m_sub [= BN_GF2m_add]
6431 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6432 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6433 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6435 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6436 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6437 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6438 BN_GF2m_cmp [= BN_ucmp]
6440 (Note that only the 'mod' functions are actually for fields GF(2^m).
6441 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6443 For some functions, an the irreducible polynomial defining a
6444 field can be given as an 'unsigned int[]' with strictly
6445 decreasing elements giving the indices of those bits that are set;
6446 i.e., p[] represents the polynomial
6447 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6449 p[0] > p[1] > ... > p[k] = 0.
6450 This applies to the following functions:
6455 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
6456 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
6458 BN_GF2m_mod_sqrt_arr
6459 BN_GF2m_mod_solve_quad_arr
6463 Conversion can be performed by the following functions:
6468 bntest.c has additional tests for binary polynomial arithmetic.
6470 Two implementations for BN_GF2m_mod_div() are available.
6471 The default algorithm simply uses BN_GF2m_mod_inv() and
6472 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
6473 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
6474 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
6476 [Sheueling Chang Shantz and Douglas Stebila
6477 (Sun Microsystems Laboratories)]
6479 *) Add new error code 'ERR_R_DISABLED' that can be used when some
6480 functionality is disabled at compile-time.
6481 [Douglas Stebila <douglas.stebila@sun.com>]
6483 *) Change default behaviour of 'openssl asn1parse' so that more
6484 information is visible when viewing, e.g., a certificate:
6486 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
6487 mode the content of non-printable OCTET STRINGs is output in a
6488 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
6489 avoid the appearance of a printable string.
6490 [Nils Larsch <nla@trustcenter.de>]
6492 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
6494 EC_GROUP_set_asn1_flag()
6495 EC_GROUP_get_asn1_flag()
6496 EC_GROUP_set_point_conversion_form()
6497 EC_GROUP_get_point_conversion_form()
6498 These control ASN1 encoding details:
6499 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
6500 has been set to OPENSSL_EC_NAMED_CURVE.
6501 - Points are encoded in uncompressed form by default; options for
6502 asn1_for are as for point2oct, namely
6503 POINT_CONVERSION_COMPRESSED
6504 POINT_CONVERSION_UNCOMPRESSED
6505 POINT_CONVERSION_HYBRID
6507 Also add 'seed' and 'seed_len' members to EC_GROUP with access
6510 EC_GROUP_get0_seed()
6511 EC_GROUP_get_seed_len()
6512 This is used only for ASN1 purposes (so far).
6513 [Nils Larsch <nla@trustcenter.de>]
6515 *) Add 'field_type' member to EC_METHOD, which holds the NID
6516 of the appropriate field type OID. The new function
6517 EC_METHOD_get_field_type() returns this value.
6518 [Nils Larsch <nla@trustcenter.de>]
6523 EC_POINT_point2hex()
6524 EC_POINT_hex2point()
6525 providing useful interfaces to EC_POINT_point2oct() and
6526 EC_POINT_oct2point().
6527 [Nils Larsch <nla@trustcenter.de>]
6529 *) Change internals of the EC library so that the functions
6530 EC_GROUP_set_generator()
6531 EC_GROUP_get_generator()
6532 EC_GROUP_get_order()
6533 EC_GROUP_get_cofactor()
6534 are implemented directly in crypto/ec/ec_lib.c and not dispatched
6535 to methods, which would lead to unnecessary code duplication when
6536 adding different types of curves.
6537 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
6539 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
6540 arithmetic, and such that modified wNAFs are generated
6541 (which avoid length expansion in many cases).
6544 *) Add a function EC_GROUP_check_discriminant() (defined via
6545 EC_METHOD) that verifies that the curve discriminant is non-zero.
6547 Add a function EC_GROUP_check() that makes some sanity tests
6548 on a EC_GROUP, its generator and order. This includes
6549 EC_GROUP_check_discriminant().
6550 [Nils Larsch <nla@trustcenter.de>]
6552 *) Add ECDSA in new directory crypto/ecdsa/.
6554 Add applications 'openssl ecparam' and 'openssl ecdsa'
6555 (these are based on 'openssl dsaparam' and 'openssl dsa').
6557 ECDSA support is also included in various other files across the
6558 library. Most notably,
6559 - 'openssl req' now has a '-newkey ecdsa:file' option;
6560 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
6561 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
6562 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
6563 them suitable for ECDSA where domain parameters must be
6564 extracted before the specific public key;
6565 - ECDSA engine support has been added.
6566 [Nils Larsch <nla@trustcenter.de>]
6568 *) Include some named elliptic curves, and add OIDs from X9.62,
6569 SECG, and WAP/WTLS. Each curve can be obtained from the new
6571 EC_GROUP_new_by_curve_name(),
6572 and the list of available named curves can be obtained with
6573 EC_get_builtin_curves().
6574 Also add a 'curve_name' member to EC_GROUP objects, which can be
6576 EC_GROUP_set_curve_name()
6577 EC_GROUP_get_curve_name()
6578 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
6580 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6581 was actually never needed) and in BN_mul(). The removal in BN_mul()
6582 required a small change in bn_mul_part_recursive() and the addition
6583 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6584 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6585 bn_sub_words() and bn_add_words() except they take arrays with
6589 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
6591 *) Cleanse PEM buffers before freeing them since they may contain
6593 [Benjamin Bennett <ben@psc.edu>]
6595 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6596 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6597 authentication-only ciphersuites.
6600 *) Since AES128 and AES256 share a single mask bit in the logic of
6601 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6602 kludge to work properly if AES128 is available and AES256 isn't.
6605 *) Expand security boundary to match 1.1.1 module.
6608 *) Remove redundant features: hash file source, editing of test vectors
6609 modify fipsld to use external fips_premain.c signature.
6612 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6613 run algorithm test programs.
6616 *) Make algorithm test programs more tolerant of whitespace.
6619 *) Have SSL/TLS server implementation tolerate "mismatched" record
6620 protocol version while receiving ClientHello even if the
6621 ClientHello is fragmented. (The server can't insist on the
6622 particular protocol version it has chosen before the ServerHello
6623 message has informed the client about his choice.)
6626 *) Load error codes if they are not already present instead of using a
6627 static variable. This allows them to be cleanly unloaded and reloaded.
6630 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6632 *) Introduce limits to prevent malicious keys being able to
6633 cause a denial of service. (CVE-2006-2940)
6634 [Steve Henson, Bodo Moeller]
6636 *) Fix ASN.1 parsing of certain invalid structures that can result
6637 in a denial of service. (CVE-2006-2937) [Steve Henson]
6639 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6640 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6642 *) Fix SSL client code which could crash if connecting to a
6643 malicious SSLv2 server. (CVE-2006-4343)
6644 [Tavis Ormandy and Will Drewry, Google Security Team]
6646 *) Change ciphersuite string processing so that an explicit
6647 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
6648 will no longer include "AES128-SHA"), and any other similar
6649 ciphersuite (same bitmap) from *other* protocol versions (so that
6650 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
6651 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
6652 changes from 0.9.8b and 0.9.8d.
6655 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
6657 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6658 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6660 *) Change the Unix randomness entropy gathering to use poll() when
6661 possible instead of select(), since the latter has some
6662 undesirable limitations.
6663 [Darryl Miles via Richard Levitte and Bodo Moeller]
6665 *) Disable rogue ciphersuites:
6667 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6668 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6669 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6671 The latter two were purportedly from
6672 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6675 Also deactivate the remaining ciphersuites from
6676 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6677 unofficial, and the ID has long expired.
6680 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6681 dual-core machines) and other potential thread-safety issues.
6684 Changes between 0.9.7i and 0.9.7j [04 May 2006]
6686 *) Adapt fipsld and the build system to link against the validated FIPS
6687 module in FIPS mode.
6690 *) Fixes for VC++ 2005 build under Windows.
6693 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
6694 from a Windows bash shell such as MSYS. It is autodetected from the
6695 "config" script when run from a VC++ environment. Modify standard VC++
6696 build to use fipscanister.o from the GNU make build.
6699 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
6701 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
6702 The value now differs depending on if you build for FIPS or not.
6703 BEWARE! A program linked with a shared FIPSed libcrypto can't be
6704 safely run with a non-FIPSed libcrypto, as it may crash because of
6705 the difference induced by this change.
6708 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
6710 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6711 (part of SSL_OP_ALL). This option used to disable the
6712 countermeasure against man-in-the-middle protocol-version
6713 rollback in the SSL 2.0 server implementation, which is a bad
6714 idea. (CVE-2005-2969)
6716 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6717 for Information Security, National Institute of Advanced Industrial
6718 Science and Technology [AIST], Japan)]
6720 *) Minimal support for X9.31 signatures and PSS padding modes. This is
6721 mainly for FIPS compliance and not fully integrated at this stage.
6724 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
6725 the exponentiation using a fixed-length exponent. (Otherwise,
6726 the information leaked through timing could expose the secret key
6727 after many signatures; cf. Bleichenbacher's attack on DSA with
6731 *) Make a new fixed-window mod_exp implementation the default for
6732 RSA, DSA, and DH private-key operations so that the sequence of
6733 squares and multiplies and the memory access pattern are
6734 independent of the particular secret key. This will mitigate
6735 cache-timing and potential related attacks.
6737 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6738 and this is automatically used by BN_mod_exp_mont() if the new flag
6739 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6740 will use this BN flag for private exponents unless the flag
6741 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6742 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6744 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6746 *) Change the client implementation for SSLv23_method() and
6747 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6748 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6749 (Previously, the SSL 2.0 backwards compatible Client Hello
6750 message format would be used even with SSL_OP_NO_SSLv2.)
6753 *) Add support for smime-type MIME parameter in S/MIME messages which some
6757 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6758 a threadsafe manner. Modify rsa code to use new function and add calls
6759 to dsa and dh code (which had race conditions before).
6762 *) Include the fixed error library code in the C error file definitions
6763 instead of fixing them up at runtime. This keeps the error code
6764 structures constant.
6767 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6769 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6772 *) Fixes for newer kerberos headers. NB: the casts are needed because
6773 the 'length' field is signed on one version and unsigned on another
6774 with no (?) obvious way to tell the difference, without these VC++
6775 complains. Also the "definition" of FAR (blank) is no longer included
6776 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6777 some needed definitions.
6780 *) Undo Cygwin change.
6783 *) Added support for proxy certificates according to RFC 3820.
6784 Because they may be a security thread to unaware applications,
6785 they must be explicitly allowed in run-time. See
6786 docs/HOWTO/proxy_certificates.txt for further information.
6789 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6791 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6792 server and client random values. Previously
6793 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6794 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6796 This change has negligible security impact because:
6798 1. Server and client random values still have 24 bytes of pseudo random
6801 2. Server and client random values are sent in the clear in the initial
6804 3. The master secret is derived using the premaster secret (48 bytes in
6805 size for static RSA ciphersuites) as well as client server and random
6808 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6811 [Stephen Henson, reported by UK NISCC]
6813 *) Use Windows randomness collection on Cygwin.
6816 *) Fix hang in EGD/PRNGD query when communication socket is closed
6817 prematurely by EGD/PRNGD.
6818 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6820 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6823 *) Back-port of selected performance improvements from development
6824 branch, as well as improved support for PowerPC platforms.
6827 *) Add lots of checks for memory allocation failure, error codes to indicate
6828 failure and freeing up memory if a failure occurs.
6829 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6831 *) Add new -passin argument to dgst.
6834 *) Perform some character comparisons of different types in X509_NAME_cmp:
6835 this is needed for some certificates that re-encode DNs into UTF8Strings
6836 (in violation of RFC3280) and can't or won't issue name rollover
6840 *) Make an explicit check during certificate validation to see that
6841 the CA setting in each certificate on the chain is correct. As a
6842 side effect always do the following basic checks on extensions,
6843 not just when there's an associated purpose to the check:
6845 - if there is an unhandled critical extension (unless the user
6846 has chosen to ignore this fault)
6847 - if the path length has been exceeded (if one is set at all)
6848 - that certain extensions fit the associated purpose (if one has
6852 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6854 *) Avoid a race condition when CRLs are checked in a multi threaded
6855 environment. This would happen due to the reordering of the revoked
6856 entries during signature checking and serial number lookup. Now the
6857 encoding is cached and the serial number sort performed under a lock.
6858 Add new STACK function sk_is_sorted().
6861 *) Add Delta CRL to the extension code.
6864 *) Various fixes to s3_pkt.c so alerts are sent properly.
6865 [David Holmes <d.holmes@f5.com>]
6867 *) Reduce the chances of duplicate issuer name and serial numbers (in
6868 violation of RFC3280) using the OpenSSL certificate creation utilities.
6869 This is done by creating a random 64 bit value for the initial serial
6870 number when a serial number file is created or when a self signed
6871 certificate is created using 'openssl req -x509'. The initial serial
6872 number file is created using 'openssl x509 -next_serial' in CA.pl
6873 rather than being initialized to 1.
6876 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
6878 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6879 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6880 [Joe Orton, Steve Henson]
6882 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
6884 [Joe Orton, Steve Henson]
6886 *) Make it possible to have multiple active certificates with the same
6887 subject in the CA index file. This is done only if the keyword
6888 'unique_subject' is set to 'no' in the main CA section (default
6889 if 'CA_default') of the configuration file. The value is saved
6890 with the database itself in a separate index attribute file,
6891 named like the index file with '.attr' appended to the name.
6894 *) X509 verify fixes. Disable broken certificate workarounds when
6895 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
6896 keyUsage extension present. Don't accept CRLs with unhandled critical
6897 extensions: since verify currently doesn't process CRL extensions this
6898 rejects a CRL with *any* critical extensions. Add new verify error codes
6902 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
6903 A clarification of RFC2560 will require the use of OCTET STRINGs and
6904 some implementations cannot handle the current raw format. Since OpenSSL
6905 copies and compares OCSP nonces as opaque blobs without any attempt at
6906 parsing them this should not create any compatibility issues.
6909 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
6910 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
6911 this HMAC (and other) operations are several times slower than OpenSSL
6915 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
6916 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6918 *) Use the correct content when signing type "other".
6921 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
6923 *) Fix various bugs revealed by running the NISCC test suite:
6925 Stop out of bounds reads in the ASN1 code when presented with
6926 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6928 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
6930 If verify callback ignores invalid public key errors don't try to check
6931 certificate signature with the NULL public key.
6935 *) New -ignore_err option in ocsp application to stop the server
6936 exiting on the first error in a request.
6939 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6940 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6944 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6945 extra data after the compression methods not only for TLS 1.0
6946 but also for SSL 3.0 (as required by the specification).
6947 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6949 *) Change X509_certificate_type() to mark the key as exported/exportable
6950 when it's 512 *bits* long, not 512 bytes.
6953 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6954 blocks during encryption.
6957 *) Various fixes to base64 BIO and non blocking I/O. On write
6958 flushes were not handled properly if the BIO retried. On read
6959 data was not being buffered properly and had various logic bugs.
6960 This also affects blocking I/O when the data being decoded is a
6964 *) Various S/MIME bugfixes and compatibility changes:
6965 output correct application/pkcs7 MIME type if
6966 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
6967 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
6968 of files as .eml work). Correctly handle very long lines in MIME
6972 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
6974 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6975 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6976 a protocol version number mismatch like a decryption error
6977 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6980 *) Turn on RSA blinding by default in the default implementation
6981 to avoid a timing attack. Applications that don't want it can call
6982 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6983 They would be ill-advised to do so in most cases.
6984 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6986 *) Change RSA blinding code so that it works when the PRNG is not
6987 seeded (in this case, the secret RSA exponent is abused as
6988 an unpredictable seed -- if it is not unpredictable, there
6989 is no point in blinding anyway). Make RSA blinding thread-safe
6990 by remembering the creator's thread ID in rsa->blinding and
6991 having all other threads use local one-time blinding factors
6992 (this requires more computation than sharing rsa->blinding, but
6993 avoids excessive locking; and if an RSA object is not shared
6994 between threads, blinding will still be very fast).
6997 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
6998 ENGINE as defaults for all supported algorithms irrespective of
6999 the 'flags' parameter. 'flags' is now honoured, so applications
7000 should make sure they are passing it correctly.
7003 *) Target "mingw" now allows native Windows code to be generated in
7004 the Cygwin environment as well as with the MinGW compiler.
7007 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
7009 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
7010 via timing by performing a MAC computation even if incorrect
7011 block cipher padding has been found. This is a countermeasure
7012 against active attacks where the attacker has to distinguish
7013 between bad padding and a MAC verification error. (CVE-2003-0078)
7015 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
7016 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
7017 Martin Vuagnoux (EPFL, Ilion)]
7019 *) Make the no-err option work as intended. The intention with no-err
7020 is not to have the whole error stack handling routines removed from
7021 libcrypto, it's only intended to remove all the function name and
7022 reason texts, thereby removing some of the footprint that may not
7023 be interesting if those errors aren't displayed anyway.
7025 NOTE: it's still possible for any application or module to have its
7026 own set of error texts inserted. The routines are there, just not
7027 used by default when no-err is given.
7030 *) Add support for FreeBSD on IA64.
7031 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
7033 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
7034 Kerberos function mit_des_cbc_cksum(). Before this change,
7035 the value returned by DES_cbc_cksum() was like the one from
7036 mit_des_cbc_cksum(), except the bytes were swapped.
7037 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
7039 *) Allow an application to disable the automatic SSL chain building.
7040 Before this a rather primitive chain build was always performed in
7041 ssl3_output_cert_chain(): an application had no way to send the
7042 correct chain if the automatic operation produced an incorrect result.
7044 Now the chain builder is disabled if either:
7046 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
7048 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
7050 The reasoning behind this is that an application would not want the
7051 auto chain building to take place if extra chain certificates are
7052 present and it might also want a means of sending no additional
7053 certificates (for example the chain has two certificates and the
7057 *) Add the possibility to build without the ENGINE framework.
7058 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7060 *) Under Win32 gmtime() can return NULL: check return value in
7061 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
7064 *) DSA routines: under certain error conditions uninitialized BN objects
7065 could be freed. Solution: make sure initialization is performed early
7066 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
7067 Nils Larsch <nla@trustcenter.de> via PR#459)
7070 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
7071 checked on reconnect on the client side, therefore session resumption
7072 could still fail with a "ssl session id is different" error. This
7073 behaviour is masked when SSL_OP_ALL is used due to
7074 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
7075 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7076 followup to PR #377.
7079 *) IA-32 assembler support enhancements: unified ELF targets, support
7080 for SCO/Caldera platforms, fix for Cygwin shared build.
7083 *) Add support for FreeBSD on sparc64. As a consequence, support for
7084 FreeBSD on non-x86 processors is separate from x86 processors on
7085 the config script, much like the NetBSD support.
7086 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
7088 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
7090 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
7093 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
7094 code (06) was taken as the first octet of the session ID and the last
7095 octet was ignored consequently. As a result SSLv2 client side session
7096 caching could not have worked due to the session ID mismatch between
7098 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7102 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
7103 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
7107 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
7108 seems that in spite of existing for more than a year, many application
7109 author have done nothing to provide the necessary callbacks, which
7110 means that this particular engine will not work properly anywhere.
7111 This is a very unfortunate situation which forces us, in the name
7112 of usability, to give the hw_ncipher.c a static lock, which is part
7114 NOTE: This is for the 0.9.7 series ONLY. This hack will never
7115 appear in 0.9.8 or later. We EXPECT application authors to have
7116 dealt properly with this when 0.9.8 is released (unless we actually
7117 make such changes in the libcrypto locking code that changes will
7118 have to be made anyway).
7121 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
7122 octets have been read, EOF or an error occurs. Without this change
7123 some truncated ASN1 structures will not produce an error.
7126 *) Disable Heimdal support, since it hasn't been fully implemented.
7127 Still give the possibility to force the use of Heimdal, but with
7128 warnings and a request that patches get sent to openssl-dev.
7131 *) Add the VC-CE target, introduce the WINCE sysname, and add
7132 INSTALL.WCE and appropriate conditionals to make it build.
7133 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7135 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
7136 cygssl-x.y.z.dll, where x, y and z are the major, minor and
7137 edit numbers of the version.
7138 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7140 *) Introduce safe string copy and catenation functions
7141 (BUF_strlcpy() and BUF_strlcat()).
7142 [Ben Laurie (CHATS) and Richard Levitte]
7144 *) Avoid using fixed-size buffers for one-line DNs.
7145 [Ben Laurie (CHATS)]
7147 *) Add BUF_MEM_grow_clean() to avoid information leakage when
7148 resizing buffers containing secrets, and use where appropriate.
7149 [Ben Laurie (CHATS)]
7151 *) Avoid using fixed size buffers for configuration file location.
7152 [Ben Laurie (CHATS)]
7154 *) Avoid filename truncation for various CA files.
7155 [Ben Laurie (CHATS)]
7157 *) Use sizeof in preference to magic numbers.
7158 [Ben Laurie (CHATS)]
7160 *) Avoid filename truncation in cert requests.
7161 [Ben Laurie (CHATS)]
7163 *) Add assertions to check for (supposedly impossible) buffer
7165 [Ben Laurie (CHATS)]
7167 *) Don't cache truncated DNS entries in the local cache (this could
7168 potentially lead to a spoofing attack).
7169 [Ben Laurie (CHATS)]
7171 *) Fix various buffers to be large enough for hex/decimal
7172 representations in a platform independent manner.
7173 [Ben Laurie (CHATS)]
7175 *) Add CRYPTO_realloc_clean() to avoid information leakage when
7176 resizing buffers containing secrets, and use where appropriate.
7177 [Ben Laurie (CHATS)]
7179 *) Add BIO_indent() to avoid much slightly worrying code to do
7181 [Ben Laurie (CHATS)]
7183 *) Convert sprintf()/BIO_puts() to BIO_printf().
7184 [Ben Laurie (CHATS)]
7186 *) buffer_gets() could terminate with the buffer only half
7188 [Ben Laurie (CHATS)]
7190 *) Add assertions to prevent user-supplied crypto functions from
7191 overflowing internal buffers by having large block sizes, etc.
7192 [Ben Laurie (CHATS)]
7194 *) New OPENSSL_assert() macro (similar to assert(), but enabled
7196 [Ben Laurie (CHATS)]
7198 *) Eliminate unused copy of key in RC4.
7199 [Ben Laurie (CHATS)]
7201 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
7202 [Ben Laurie (CHATS)]
7204 *) Fix off-by-one error in EGD path.
7205 [Ben Laurie (CHATS)]
7207 *) If RANDFILE path is too long, ignore instead of truncating.
7208 [Ben Laurie (CHATS)]
7210 *) Eliminate unused and incorrectly sized X.509 structure
7212 [Ben Laurie (CHATS)]
7214 *) Eliminate unused and dangerous function knumber().
7215 [Ben Laurie (CHATS)]
7217 *) Eliminate unused and dangerous structure, KSSL_ERR.
7218 [Ben Laurie (CHATS)]
7220 *) Protect against overlong session ID context length in an encoded
7221 session object. Since these are local, this does not appear to be
7223 [Ben Laurie (CHATS)]
7225 *) Change from security patch (see 0.9.6e below) that did not affect
7226 the 0.9.6 release series:
7228 Remote buffer overflow in SSL3 protocol - an attacker could
7229 supply an oversized master key in Kerberos-enabled versions.
7231 [Ben Laurie (CHATS)]
7233 *) Change the SSL kerb5 codes to match RFC 2712.
7236 *) Make -nameopt work fully for req and add -reqopt switch.
7237 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
7239 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
7240 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
7242 *) Make sure tests can be performed even if the corresponding algorithms
7243 have been removed entirely. This was also the last step to make
7244 OpenSSL compilable with DJGPP under all reasonable conditions.
7245 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
7247 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
7248 to allow version independent disabling of normally unselected ciphers,
7249 which may be activated as a side-effect of selecting a single cipher.
7251 (E.g., cipher list string "RSA" enables ciphersuites that are left
7252 out of "ALL" because they do not provide symmetric encryption.
7253 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
7254 [Lutz Jaenicke, Bodo Moeller]
7256 *) Add appropriate support for separate platform-dependent build
7257 directories. The recommended way to make a platform-dependent
7258 build directory is the following (tested on Linux), maybe with
7261 # Place yourself outside of the OpenSSL source tree. In
7262 # this example, the environment variable OPENSSL_SOURCE
7263 # is assumed to contain the absolute OpenSSL source directory.
7264 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
7265 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
7266 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
7267 mkdir -p `dirname $F`
7268 ln -s $OPENSSL_SOURCE/$F $F
7271 To be absolutely sure not to disturb the source tree, a "make clean"
7272 is a good thing. If it isn't successful, don't worry about it,
7273 it probably means the source directory is very clean.
7276 *) Make sure any ENGINE control commands make local copies of string
7277 pointers passed to them whenever necessary. Otherwise it is possible
7278 the caller may have overwritten (or deallocated) the original string
7279 data when a later ENGINE operation tries to use the stored values.
7280 [Götz Babin-Ebell <babinebell@trustcenter.de>]
7282 *) Improve diagnostics in file reading and command-line digests.
7283 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
7285 *) Add AES modes CFB and OFB to the object database. Correct an
7286 error in AES-CFB decryption.
7289 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
7290 allows existing EVP_CIPHER_CTX structures to be reused after
7291 calling EVP_*Final(). This behaviour is used by encryption
7292 BIOs and some applications. This has the side effect that
7293 applications must explicitly clean up cipher contexts with
7294 EVP_CIPHER_CTX_cleanup() or they will leak memory.
7297 *) Check the values of dna and dnb in bn_mul_recursive before calling
7298 bn_mul_comba (a non zero value means the a or b arrays do not contain
7299 n2 elements) and fallback to bn_mul_normal if either is not zero.
7302 *) Fix escaping of non-ASCII characters when using the -subj option
7303 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
7306 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
7307 form for "surname", serialNumber has no short form.
7308 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
7309 therefore remove "mail" short name for "internet 7".
7310 The OID for unique identifiers in X509 certificates is
7311 x500UniqueIdentifier, not uniqueIdentifier.
7312 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7315 *) Add an "init" command to the ENGINE config module and auto initialize
7316 ENGINEs. Without any "init" command the ENGINE will be initialized
7317 after all ctrl commands have been executed on it. If init=1 the
7318 ENGINE is initialized at that point (ctrls before that point are run
7319 on the uninitialized ENGINE and after on the initialized one). If
7320 init=0 then the ENGINE will not be initialized at all.
7323 *) Fix the 'app_verify_callback' interface so that the user-defined
7324 argument is actually passed to the callback: In the
7325 SSL_CTX_set_cert_verify_callback() prototype, the callback
7326 declaration has been changed from
7329 int (*cb)(X509_STORE_CTX *,void *);
7330 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7331 i=s->ctx->app_verify_callback(&ctx)
7332 has been changed into
7333 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7335 To update applications using SSL_CTX_set_cert_verify_callback(),
7336 a dummy argument can be added to their callback functions.
7337 [D. K. Smetters <smetters@parc.xerox.com>]
7339 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7340 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7342 *) Add and OPENSSL_LOAD_CONF define which will cause
7343 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7344 This allows older applications to transparently support certain
7345 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7346 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7347 load the config file and OPENSSL_add_all_algorithms_conf() which will
7348 always load it have also been added.
7351 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7352 Adjust NIDs and EVP layer.
7353 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7355 *) Config modules support in openssl utility.
7357 Most commands now load modules from the config file,
7358 though in a few (such as version) this isn't done
7359 because it couldn't be used for anything.
7361 In the case of ca and req the config file used is
7362 the same as the utility itself: that is the -config
7363 command line option can be used to specify an
7367 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7368 use "openssl_conf" if filename is NULL use default openssl config file.
7371 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7372 config section name. Add a new flag to tolerate a missing config file
7373 and move code to CONF_modules_load_file().
7376 *) Support for crypto accelerator cards from Accelerated Encryption
7377 Processing, www.aep.ie. (Use engine 'aep')
7378 The support was copied from 0.9.6c [engine] and adapted/corrected
7379 to work with the new engine framework.
7380 [AEP Inc. and Richard Levitte]
7382 *) Support for SureWare crypto accelerator cards from Baltimore
7383 Technologies. (Use engine 'sureware')
7384 The support was copied from 0.9.6c [engine] and adapted
7385 to work with the new engine framework.
7388 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7389 make the newer ENGINE framework commands for the CHIL engine work.
7390 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7392 *) Make it possible to produce shared libraries on ReliantUNIX.
7393 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7395 *) Add the configuration target debug-linux-ppro.
7396 Make 'openssl rsa' use the general key loading routines
7397 implemented in apps.c, and make those routines able to
7398 handle the key format FORMAT_NETSCAPE and the variant
7400 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7402 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7403 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7405 *) Add -keyform to rsautl, and document -engine.
7406 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7408 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7409 BIO_R_NO_SUCH_FILE error code rather than the generic
7410 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7413 *) Add new functions
7415 ERR_peek_last_error_line
7416 ERR_peek_last_error_line_data.
7417 These are similar to
7420 ERR_peek_error_line_data,
7421 but report on the latest error recorded rather than the first one
7422 still in the error queue.
7423 [Ben Laurie, Bodo Moeller]
7425 *) default_algorithms option in ENGINE config module. This allows things
7427 default_algorithms = ALL
7428 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7431 *) Preliminary ENGINE config module.
7434 *) New experimental application configuration code.
7437 *) Change the AES code to follow the same name structure as all other
7438 symmetric ciphers, and behave the same way. Move everything to
7439 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7440 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7442 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7443 [Ben Laurie and Theo de Raadt]
7445 *) Add option to output public keys in req command.
7446 [Massimiliano Pala madwolf@openca.org]
7448 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7449 (up to about 10% better than before for P-192 and P-224).
7452 *) New functions/macros
7454 SSL_CTX_set_msg_callback(ctx, cb)
7455 SSL_CTX_set_msg_callback_arg(ctx, arg)
7456 SSL_set_msg_callback(ssl, cb)
7457 SSL_set_msg_callback_arg(ssl, arg)
7459 to request calling a callback function
7461 void cb(int write_p, int version, int content_type,
7462 const void *buf, size_t len, SSL *ssl, void *arg)
7464 whenever a protocol message has been completely received
7465 (write_p == 0) or sent (write_p == 1). Here 'version' is the
7466 protocol version according to which the SSL library interprets
7467 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
7468 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
7469 the content type as defined in the SSL 3.0/TLS 1.0 protocol
7470 specification (change_cipher_spec(20), alert(21), handshake(22)).
7471 'buf' and 'len' point to the actual message, 'ssl' to the
7472 SSL object, and 'arg' is the application-defined value set by
7473 SSL[_CTX]_set_msg_callback_arg().
7475 'openssl s_client' and 'openssl s_server' have new '-msg' options
7476 to enable a callback that displays all protocol messages.
7479 *) Change the shared library support so shared libraries are built as
7480 soon as the corresponding static library is finished, and thereby get
7481 openssl and the test programs linked against the shared library.
7482 This still only happens when the keyword "shard" has been given to
7483 the configuration scripts.
7485 NOTE: shared library support is still an experimental thing, and
7486 backward binary compatibility is still not guaranteed.
7487 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
7489 *) Add support for Subject Information Access extension.
7490 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7492 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
7493 additional bytes when new memory had to be allocated, not just
7494 when reusing an existing buffer.
7497 *) New command line and configuration option 'utf8' for the req command.
7498 This allows field values to be specified as UTF8 strings.
7501 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
7502 runs for the former and machine-readable output for the latter.
7505 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
7506 of the e-mail address in the DN (i.e., it will go into a certificate
7507 extension only). The new configuration file option 'email_in_dn = no'
7508 has the same effect.
7509 [Massimiliano Pala madwolf@openca.org]
7511 *) Change all functions with names starting with des_ to be starting
7512 with DES_ instead. Add wrappers that are compatible with libdes,
7513 but are named _ossl_old_des_*. Finally, add macros that map the
7514 des_* symbols to the corresponding _ossl_old_des_* if libdes
7515 compatibility is desired. If OpenSSL 0.9.6c compatibility is
7516 desired, the des_* symbols will be mapped to DES_*, with one
7519 Since we provide two compatibility mappings, the user needs to
7520 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
7521 compatibility is desired. The default (i.e., when that macro
7522 isn't defined) is OpenSSL 0.9.6c compatibility.
7524 There are also macros that enable and disable the support of old
7525 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
7526 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
7527 are defined, the default will apply: to support the old des routines.
7529 In either case, one must include openssl/des.h to get the correct
7530 definitions. Do not try to just include openssl/des_old.h, that
7533 NOTE: This is a major break of an old API into a new one. Software
7534 authors are encouraged to switch to the DES_ style functions. Some
7535 time in the future, des_old.h and the libdes compatibility functions
7536 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
7537 default), and then completely removed.
7540 *) Test for certificates which contain unsupported critical extensions.
7541 If such a certificate is found during a verify operation it is
7542 rejected by default: this behaviour can be overridden by either
7543 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
7544 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
7545 X509_supported_extension() has also been added which returns 1 if a
7546 particular extension is supported.
7549 *) Modify the behaviour of EVP cipher functions in similar way to digests
7550 to retain compatibility with existing code.
7553 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
7554 compatibility with existing code. In particular the 'ctx' parameter does
7555 not have to be to be initialized before the call to EVP_DigestInit() and
7556 it is tidied up after a call to EVP_DigestFinal(). New function
7557 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
7558 EVP_MD_CTX_copy() changed to not require the destination to be
7559 initialized valid and new function EVP_MD_CTX_copy_ex() added which
7560 requires the destination to be valid.
7562 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
7563 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
7566 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
7567 so that complete 'Handshake' protocol structures are kept in memory
7568 instead of overwriting 'msg_type' and 'length' with 'body' data.
7571 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
7572 [Massimo Santin via Richard Levitte]
7574 *) Major restructuring to the underlying ENGINE code. This includes
7575 reduction of linker bloat, separation of pure "ENGINE" manipulation
7576 (initialisation, etc) from functionality dealing with implementations
7577 of specific crypto interfaces. This change also introduces integrated
7578 support for symmetric ciphers and digest implementations - so ENGINEs
7579 can now accelerate these by providing EVP_CIPHER and EVP_MD
7580 implementations of their own. This is detailed in crypto/engine/README
7581 as it couldn't be adequately described here. However, there are a few
7582 API changes worth noting - some RSA, DSA, DH, and RAND functions that
7583 were changed in the original introduction of ENGINE code have now
7584 reverted back - the hooking from this code to ENGINE is now a good
7585 deal more passive and at run-time, operations deal directly with
7586 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
7587 dereferencing through an ENGINE pointer any more. Also, the ENGINE
7588 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
7589 they were not being used by the framework as there is no concept of a
7590 BIGNUM_METHOD and they could not be generalised to the new
7591 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
7592 ENGINE_cpy() has been removed as it cannot be consistently defined in
7596 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
7599 *) Change mkdef.pl to sort symbols that get the same entry number,
7600 and make sure the automatically generated functions ERR_load_*
7601 become part of libeay.num as well.
7604 *) New function SSL_renegotiate_pending(). This returns true once
7605 renegotiation has been requested (either SSL_renegotiate() call
7606 or HelloRequest/ClientHello received from the peer) and becomes
7607 false once a handshake has been completed.
7608 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
7609 sends a HelloRequest, but does not ensure that a handshake takes
7610 place. SSL_renegotiate_pending() is useful for checking if the
7611 client has followed the request.)
7614 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7615 By default, clients may request session resumption even during
7616 renegotiation (if session ID contexts permit); with this option,
7617 session resumption is possible only in the first handshake.
7619 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7620 more bits available for options that should not be part of
7621 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7624 *) Add some demos for certificate and certificate request creation.
7627 *) Make maximum certificate chain size accepted from the peer application
7628 settable (SSL*_get/set_max_cert_list()), as proposed by
7629 "Douglas E. Engert" <deengert@anl.gov>.
7632 *) Add support for shared libraries for Unixware-7
7633 (Boyd Lynn Gerber <gerberb@zenez.com>).
7636 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7637 be done prior to destruction. Use this to unload error strings from
7638 ENGINEs that load their own error strings. NB: This adds two new API
7639 functions to "get" and "set" this destroy handler in an ENGINE.
7642 *) Alter all existing ENGINE implementations (except "openssl" and
7643 "openbsd") to dynamically instantiate their own error strings. This
7644 makes them more flexible to be built both as statically-linked ENGINEs
7645 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
7646 Also, add stub code to each that makes building them as self-contained
7647 shared-libraries easier (see README.ENGINE).
7650 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
7651 implementations into applications that are completely implemented in
7652 self-contained shared-libraries. The "dynamic" ENGINE exposes control
7653 commands that can be used to configure what shared-library to load and
7654 to control aspects of the way it is handled. Also, made an update to
7655 the README.ENGINE file that brings its information up-to-date and
7656 provides some information and instructions on the "dynamic" ENGINE
7657 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
7660 *) Make it possible to unload ranges of ERR strings with a new
7661 "ERR_unload_strings" function.
7664 *) Add a copy() function to EVP_MD.
7667 *) Make EVP_MD routines take a context pointer instead of just the
7668 md_data void pointer.
7671 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
7672 that the digest can only process a single chunk of data
7673 (typically because it is provided by a piece of
7674 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
7675 is only going to provide a single chunk of data, and hence the
7676 framework needn't accumulate the data for oneshot drivers.
7679 *) As with "ERR", make it possible to replace the underlying "ex_data"
7680 functions. This change also alters the storage and management of global
7681 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
7682 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
7683 index counters. The API functions that use this state have been changed
7684 to take a "class_index" rather than pointers to the class's local STACK
7685 and counter, and there is now an API function to dynamically create new
7686 classes. This centralisation allows us to (a) plug a lot of the
7687 thread-safety problems that existed, and (b) makes it possible to clean
7688 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
7689 such data would previously have always leaked in application code and
7690 workarounds were in place to make the memory debugging turn a blind eye
7691 to it. Application code that doesn't use this new function will still
7692 leak as before, but their memory debugging output will announce it now
7693 rather than letting it slide.
7695 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
7696 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
7697 has a return value to indicate success or failure.
7700 *) Make it possible to replace the underlying "ERR" functions such that the
7701 global state (2 LHASH tables and 2 locks) is only used by the "default"
7702 implementation. This change also adds two functions to "get" and "set"
7703 the implementation prior to it being automatically set the first time
7704 any other ERR function takes place. Ie. an application can call "get",
7705 pass the return value to a module it has just loaded, and that module
7706 can call its own "set" function using that value. This means the
7707 module's "ERR" operations will use (and modify) the error state in the
7708 application and not in its own statically linked copy of OpenSSL code.
7711 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
7712 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
7713 the operation, and provides a more encapsulated way for external code
7714 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
7715 to use these functions rather than manually incrementing the counts.
7717 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
7720 *) Add EVP test program.
7723 *) Add symmetric cipher support to ENGINE. Expect the API to change!
7726 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
7727 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
7728 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
7729 These allow a CRL to be built without having to access X509_CRL fields
7730 directly. Modify 'ca' application to use new functions.
7733 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
7734 bug workarounds. Rollback attack detection is a security feature.
7735 The problem will only arise on OpenSSL servers when TLSv1 is not
7736 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7737 Software authors not wanting to support TLSv1 will have special reasons
7738 for their choice and can explicitly enable this option.
7739 [Bodo Moeller, Lutz Jaenicke]
7741 *) Rationalise EVP so it can be extended: don't include a union of
7742 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7743 (similar to those existing for EVP_CIPHER_CTX).
7748 EVP_MD_CTX_init(&md); /* new function call */
7749 EVP_DigestInit(&md, EVP_sha1());
7750 EVP_DigestUpdate(&md, in, len);
7751 EVP_DigestFinal(&md, out, NULL);
7752 EVP_MD_CTX_cleanup(&md); /* new function call */
7756 *) Make DES key schedule conform to the usual scheme, as well as
7757 correcting its structure. This means that calls to DES functions
7758 now have to pass a pointer to a des_key_schedule instead of a
7759 plain des_key_schedule (which was actually always a pointer
7762 des_key_schedule ks;
7764 des_set_key_checked(..., &ks);
7765 des_ncbc_encrypt(..., &ks, ...);
7767 (Note that a later change renames 'des_...' into 'DES_...'.)
7770 *) Initial reduction of linker bloat: the use of some functions, such as
7771 PEM causes large amounts of unused functions to be linked in due to
7772 poor organisation. For example pem_all.c contains every PEM function
7773 which has a knock on effect of linking in large amounts of (unused)
7774 ASN1 code. Grouping together similar functions and splitting unrelated
7775 functions prevents this.
7778 *) Cleanup of EVP macros.
7781 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7782 correct _ecb suffix.
7785 *) Add initial OCSP responder support to ocsp application. The
7786 revocation information is handled using the text based index
7787 use by the ca application. The responder can either handle
7788 requests generated internally, supplied in files (for example
7789 via a CGI script) or using an internal minimal server.
7792 *) Add configuration choices to get zlib compression for TLS.
7795 *) Changes to Kerberos SSL for RFC 2712 compliance:
7796 1. Implemented real KerberosWrapper, instead of just using
7797 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7798 2. Implemented optional authenticator field of KerberosWrapper.
7800 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7801 and authenticator structs; see crypto/krb5/.
7803 Generalized Kerberos calls to support multiple Kerberos libraries.
7804 [Vern Staats <staatsvr@asc.hpc.mil>,
7805 Jeffrey Altman <jaltman@columbia.edu>
7806 via Richard Levitte]
7808 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7809 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7810 values for each of the key sizes rather than having just
7811 parameters (and 'speed' generating keys each time).
7814 *) Speed up EVP routines.
7817 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7818 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7819 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7820 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7822 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7823 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7824 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7827 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7829 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7832 *) Added the OS2-EMX target.
7833 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7835 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7836 to support NCONF routines in extension code. New function CONF_set_nconf()
7837 to allow functions which take an NCONF to also handle the old LHASH
7838 structure: this means that the old CONF compatible routines can be
7839 retained (in particular wrt extensions) without having to duplicate the
7840 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7843 *) Enhance the general user interface with mechanisms for inner control
7844 and with possibilities to have yes/no kind of prompts.
7847 *) Change all calls to low level digest routines in the library and
7848 applications to use EVP. Add missing calls to HMAC_cleanup() and
7849 don't assume HMAC_CTX can be copied using memcpy().
7850 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7852 *) Add the possibility to control engines through control names but with
7853 arbitrary arguments instead of just a string.
7854 Change the key loaders to take a UI_METHOD instead of a callback
7855 function pointer. NOTE: this breaks binary compatibility with earlier
7856 versions of OpenSSL [engine].
7857 Adapt the nCipher code for these new conditions and add a card insertion
7861 *) Enhance the general user interface with mechanisms to better support
7862 dialog box interfaces, application-defined prompts, the possibility
7863 to use defaults (for example default passwords from somewhere else)
7864 and interrupts/cancellations.
7867 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7868 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7871 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7872 tidy up some unnecessarily weird code in 'sk_new()').
7873 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
7875 *) Change the key loading routines for ENGINEs to use the same kind
7876 callback (pem_password_cb) as all other routines that need this
7880 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
7881 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
7882 than this minimum value is recommended.
7885 *) New random seeder for OpenVMS, using the system process statistics
7886 that are easily reachable.
7889 *) Windows apparently can't transparently handle global
7890 variables defined in DLLs. Initialisations such as:
7892 const ASN1_ITEM *it = &ASN1_INTEGER_it;
7894 won't compile. This is used by the any applications that need to
7895 declare their own ASN1 modules. This was fixed by adding the option
7896 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
7897 needed for static libraries under Win32.
7900 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
7901 setting of purpose and trust fields. New X509_STORE trust and
7902 purpose functions and tidy up setting in other SSL functions.
7905 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
7906 structure. These are inherited by X509_STORE_CTX when it is
7907 initialised. This allows various defaults to be set in the
7908 X509_STORE structure (such as flags for CRL checking and custom
7909 purpose or trust settings) for functions which only use X509_STORE_CTX
7910 internally such as S/MIME.
7912 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
7913 trust settings if they are not set in X509_STORE. This allows X509_STORE
7914 purposes and trust (in S/MIME for example) to override any set by default.
7916 Add command line options for CRL checking to smime, s_client and s_server
7920 *) Initial CRL based revocation checking. If the CRL checking flag(s)
7921 are set then the CRL is looked up in the X509_STORE structure and
7922 its validity and signature checked, then if the certificate is found
7923 in the CRL the verify fails with a revoked error.
7925 Various new CRL related callbacks added to X509_STORE_CTX structure.
7927 Command line options added to 'verify' application to support this.
7929 This needs some additional work, such as being able to handle multiple
7930 CRLs with different times, extension based lookup (rather than just
7931 by subject name) and ultimately more complete V2 CRL extension
7935 *) Add a general user interface API (crypto/ui/). This is designed
7936 to replace things like des_read_password and friends (backward
7937 compatibility functions using this new API are provided).
7938 The purpose is to remove prompting functions from the DES code
7939 section as well as provide for prompting through dialog boxes in
7940 a window system and the like.
7943 *) Add "ex_data" support to ENGINE so implementations can add state at a
7944 per-structure level rather than having to store it globally.
7947 *) Make it possible for ENGINE structures to be copied when retrieved by
7948 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
7949 This causes the "original" ENGINE structure to act like a template,
7950 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7951 operational state can be localised to each ENGINE structure, despite the
7952 fact they all share the same "methods". New ENGINE structures returned in
7953 this case have no functional references and the return value is the single
7954 structural reference. This matches the single structural reference returned
7955 by ENGINE_by_id() normally, when it is incremented on the pre-existing
7959 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
7960 needs to match any other type at all we need to manually clear the
7964 *) Changes to the "openssl engine" utility to include;
7965 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
7966 about an ENGINE's available control commands.
7967 - executing control commands from command line arguments using the
7968 '-pre' and '-post' switches. '-post' is only used if '-t' is
7969 specified and the ENGINE is successfully initialised. The syntax for
7970 the individual commands are colon-separated, for example;
7971 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
7974 *) New dynamic control command support for ENGINEs. ENGINEs can now
7975 declare their own commands (numbers), names (strings), descriptions,
7976 and input types for run-time discovery by calling applications. A
7977 subset of these commands are implicitly classed as "executable"
7978 depending on their input type, and only these can be invoked through
7979 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
7980 can be based on user input, config files, etc). The distinction is
7981 that "executable" commands cannot return anything other than a boolean
7982 result and can only support numeric or string input, whereas some
7983 discoverable commands may only be for direct use through
7984 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
7985 pointers, or other custom uses. The "executable" commands are to
7986 support parameterisations of ENGINE behaviour that can be
7987 unambiguously defined by ENGINEs and used consistently across any
7988 OpenSSL-based application. Commands have been added to all the
7989 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
7990 control over shared-library paths without source code alterations.
7993 *) Changed all ENGINE implementations to dynamically allocate their
7994 ENGINEs rather than declaring them statically. Apart from this being
7995 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
7996 this also allows the implementations to compile without using the
7997 internal engine_int.h header.
8000 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
8001 'const' value. Any code that should be able to modify a RAND_METHOD
8002 should already have non-const pointers to it (ie. they should only
8003 modify their own ones).
8006 *) Made a variety of little tweaks to the ENGINE code.
8007 - "atalla" and "ubsec" string definitions were moved from header files
8008 to C code. "nuron" string definitions were placed in variables
8009 rather than hard-coded - allowing parameterisation of these values
8010 later on via ctrl() commands.
8011 - Removed unused "#if 0"'d code.
8012 - Fixed engine list iteration code so it uses ENGINE_free() to release
8013 structural references.
8014 - Constified the RAND_METHOD element of ENGINE structures.
8015 - Constified various get/set functions as appropriate and added
8016 missing functions (including a catch-all ENGINE_cpy that duplicates
8017 all ENGINE values onto a new ENGINE except reference counts/state).
8018 - Removed NULL parameter checks in get/set functions. Setting a method
8019 or function to NULL is a way of cancelling out a previously set
8020 value. Passing a NULL ENGINE parameter is just plain stupid anyway
8021 and doesn't justify the extra error symbols and code.
8022 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
8023 flags from engine_int.h to engine.h.
8024 - Changed prototypes for ENGINE handler functions (init(), finish(),
8025 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
8028 *) Implement binary inversion algorithm for BN_mod_inverse in addition
8029 to the algorithm using long division. The binary algorithm can be
8030 used only if the modulus is odd. On 32-bit systems, it is faster
8031 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
8032 roughly 5-15% for 256-bit moduli), so we use it only for moduli
8033 up to 450 bits. In 64-bit environments, the binary algorithm
8034 appears to be advantageous for much longer moduli; here we use it
8035 for moduli up to 2048 bits.
8038 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
8039 could not support the combine flag in choice fields.
8042 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
8043 extensions from a certificate request to the certificate.
8046 *) Allow multiple 'certopt' and 'nameopt' options to be separated
8047 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
8048 file: this allows the display of the certificate about to be
8049 signed to be customised, to allow certain fields to be included
8050 or excluded and extension details. The old system didn't display
8051 multicharacter strings properly, omitted fields not in the policy
8052 and couldn't display additional details such as extensions.
8055 *) Function EC_POINTs_mul for multiple scalar multiplication
8056 of an arbitrary number of elliptic curve points
8057 \sum scalars[i]*points[i],
8058 optionally including the generator defined for the EC_GROUP:
8059 scalar*generator + \sum scalars[i]*points[i].
8061 EC_POINT_mul is a simple wrapper function for the typical case
8062 that the point list has just one item (besides the optional
8066 *) First EC_METHODs for curves over GF(p):
8068 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
8069 operations and provides various method functions that can also
8070 operate with faster implementations of modular arithmetic.
8072 EC_GFp_mont_method() reuses most functions that are part of
8073 EC_GFp_simple_method, but uses Montgomery arithmetic.
8075 [Bodo Moeller; point addition and point doubling
8076 implementation directly derived from source code provided by
8077 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
8079 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
8080 crypto/ec/ec_lib.c):
8082 Curves are EC_GROUP objects (with an optional group generator)
8083 based on EC_METHODs that are built into the library.
8085 Points are EC_POINT objects based on EC_GROUP objects.
8087 Most of the framework would be able to handle curves over arbitrary
8088 finite fields, but as there are no obvious types for fields other
8089 than GF(p), some functions are limited to that for now.
8092 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
8093 that the file contains a complete HTTP response.
8096 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
8097 change the def and num file printf format specifier from "%-40sXXX"
8098 to "%-39s XXX". The latter will always guarantee a space after the
8099 field while the former will cause them to run together if the field
8100 is 40 of more characters long.
8103 *) Constify the cipher and digest 'method' functions and structures
8104 and modify related functions to take constant EVP_MD and EVP_CIPHER
8108 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
8109 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
8112 *) Modify EVP_Digest*() routines so they now return values. Although the
8113 internal software routines can never fail additional hardware versions
8117 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
8119 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
8120 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
8123 ERR_R_NESTED_ASN1_ERROR
8125 ERR_R_MISSING_ASN1_EOS
8126 were 4 .. 9, conflicting with
8127 ERR_LIB_RSA (= ERR_R_RSA_LIB)
8129 ERR_LIB_PEM (= ERR_R_PEM_LIB).
8130 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
8132 Add new error code 'ERR_R_INTERNAL_ERROR'.
8135 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
8139 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
8140 sets the subject name for a new request or supersedes the
8141 subject name in a given request. Formats that can be parsed are
8142 'CN=Some Name, OU=myOU, C=IT'
8144 'CN=Some Name/OU=myOU/C=IT'.
8146 Add options '-batch' and '-verbose' to 'openssl req'.
8147 [Massimiliano Pala <madwolf@hackmasters.net>]
8149 *) Introduce the possibility to access global variables through
8150 functions on platform were that's the best way to handle exporting
8151 global variables in shared libraries. To enable this functionality,
8152 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
8153 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
8154 is normally done by Configure or something similar).
8156 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
8157 in the source file (foo.c) like this:
8159 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
8160 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
8162 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
8163 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
8165 OPENSSL_DECLARE_GLOBAL(int,foo);
8166 #define foo OPENSSL_GLOBAL_REF(foo)
8167 OPENSSL_DECLARE_GLOBAL(double,bar);
8168 #define bar OPENSSL_GLOBAL_REF(bar)
8170 The #defines are very important, and therefore so is including the
8171 header file everywhere where the defined globals are used.
8173 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
8174 of ASN.1 items, but that structure is a bit different.
8176 The largest change is in util/mkdef.pl which has been enhanced with
8177 better and easier to understand logic to choose which symbols should
8178 go into the Windows .def files as well as a number of fixes and code
8179 cleanup (among others, algorithm keywords are now sorted
8180 lexicographically to avoid constant rewrites).
8183 *) In BN_div() keep a copy of the sign of 'num' before writing the
8184 result to 'rm' because if rm==num the value will be overwritten
8185 and produce the wrong result if 'num' is negative: this caused
8186 problems with BN_mod() and BN_nnmod().
8189 *) Function OCSP_request_verify(). This checks the signature on an
8190 OCSP request and verifies the signer certificate. The signer
8191 certificate is just checked for a generic purpose and OCSP request
8195 *) Add OCSP_check_validity() function to check the validity of OCSP
8196 responses. OCSP responses are prepared in real time and may only
8197 be a few seconds old. Simply checking that the current time lies
8198 between thisUpdate and nextUpdate max reject otherwise valid responses
8199 caused by either OCSP responder or client clock inaccuracy. Instead
8200 we allow thisUpdate and nextUpdate to fall within a certain period of
8201 the current time. The age of the response can also optionally be
8202 checked. Two new options -validity_period and -status_age added to
8206 *) If signature or public key algorithm is unrecognized print out its
8207 OID rather that just UNKNOWN.
8210 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
8211 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
8212 ID to be generated from the issuer certificate alone which can then be
8213 passed to OCSP_id_issuer_cmp().
8216 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
8217 ASN1 modules to export functions returning ASN1_ITEM pointers
8218 instead of the ASN1_ITEM structures themselves. This adds several
8219 new macros which allow the underlying ASN1 function/structure to
8220 be accessed transparently. As a result code should not use ASN1_ITEM
8221 references directly (such as &X509_it) but instead use the relevant
8222 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
8223 use of the new ASN1 code on platforms where exporting structures
8224 is problematical (for example in shared libraries) but exporting
8225 functions returning pointers to structures is not.
8228 *) Add support for overriding the generation of SSL/TLS session IDs.
8229 These callbacks can be registered either in an SSL_CTX or per SSL.
8230 The purpose of this is to allow applications to control, if they wish,
8231 the arbitrary values chosen for use as session IDs, particularly as it
8232 can be useful for session caching in multiple-server environments. A
8233 command-line switch for testing this (and any client code that wishes
8234 to use such a feature) has been added to "s_server".
8235 [Geoff Thorpe, Lutz Jaenicke]
8237 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
8238 of the form '#if defined(...) || defined(...) || ...' and
8239 '#if !defined(...) && !defined(...) && ...'. This also avoids
8240 the growing number of special cases it was previously handling.
8243 *) Make all configuration macros available for application by making
8244 sure they are available in opensslconf.h, by giving them names starting
8245 with "OPENSSL_" to avoid conflicts with other packages and by making
8246 sure e_os2.h will cover all platform-specific cases together with
8248 Additionally, it is now possible to define configuration/platform-
8249 specific names (called "system identities"). In the C code, these
8250 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
8251 macro with the name beginning with "OPENSSL_SYS_", which is determined
8252 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
8256 *) New option -set_serial to 'req' and 'x509' this allows the serial
8257 number to use to be specified on the command line. Previously self
8258 signed certificates were hard coded with serial number 0 and the
8259 CA options of 'x509' had to use a serial number in a file which was
8263 *) New options to 'ca' utility to support V2 CRL entry extensions.
8264 Currently CRL reason, invalidity date and hold instruction are
8265 supported. Add new CRL extensions to V3 code and some new objects.
8268 *) New function EVP_CIPHER_CTX_set_padding() this is used to
8269 disable standard block padding (aka PKCS#5 padding) in the EVP
8270 API, which was previously mandatory. This means that the data is
8271 not padded in any way and so the total length much be a multiple
8272 of the block size, otherwise an error occurs.
8275 *) Initial (incomplete) OCSP SSL support.
8278 *) New function OCSP_parse_url(). This splits up a URL into its host,
8279 port and path components: primarily to parse OCSP URLs. New -url
8280 option to ocsp utility.
8283 *) New nonce behavior. The return value of OCSP_check_nonce() now
8284 reflects the various checks performed. Applications can decide
8285 whether to tolerate certain situations such as an absent nonce
8286 in a response when one was present in a request: the ocsp application
8287 just prints out a warning. New function OCSP_add1_basic_nonce()
8288 this is to allow responders to include a nonce in a response even if
8289 the request is nonce-less.
8292 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
8293 skipped when using openssl x509 multiple times on a single input file,
8294 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
8297 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
8298 set string type: to handle setting ASN1_TIME structures. Fix ca
8299 utility to correctly initialize revocation date of CRLs.
8302 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
8303 the clients preferred ciphersuites and rather use its own preferences.
8304 Should help to work around M$ SGC (Server Gated Cryptography) bug in
8305 Internet Explorer by ensuring unchanged hash method during stepup.
8306 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
8309 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
8310 to aes and add a new 'exist' option to print out symbols that don't
8314 *) Additional options to ocsp utility to allow flags to be set and
8315 additional certificates supplied.
8318 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8319 OCSP client a number of certificate to only verify the response
8323 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8324 handle the new API. Currently only ECB, CBC modes supported. Add new
8327 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8328 Encryption Standard (AES) Ciphersuites for Transport Layer
8329 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8330 not enabled by default and were not part of the "ALL" ciphersuite
8331 alias because they were not yet official; they could be
8332 explicitly requested by specifying the "AESdraft" ciphersuite
8333 group alias. In the final release of OpenSSL 0.9.7, the group
8334 alias is called "AES" and is part of "ALL".)
8335 [Ben Laurie, Steve Henson, Bodo Moeller]
8337 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8338 request to response.
8341 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8342 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8343 extract information from a certificate request. OCSP_response_create()
8344 creates a response and optionally adds a basic response structure.
8345 OCSP_basic_add1_status() adds a complete single response to a basic
8346 response and returns the OCSP_SINGLERESP structure just added (to allow
8347 extensions to be included for example). OCSP_basic_add1_cert() adds a
8348 certificate to a basic response and OCSP_basic_sign() signs a basic
8349 response with various flags. New helper functions ASN1_TIME_check()
8350 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8351 (converts ASN1_TIME to GeneralizedTime).
8354 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8355 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8356 structure from a certificate. X509_pubkey_digest() digests the public_key
8357 contents: this is used in various key identifiers.
8360 *) Make sk_sort() tolerate a NULL argument.
8361 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8363 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8364 passed by the function are trusted implicitly. If any of them signed the
8365 response then it is assumed to be valid and is not verified.
8368 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8369 to data. This was previously part of the PKCS7 ASN1 code. This
8370 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8371 [Steve Henson, reported by Kenneth R. Robinette
8372 <support@securenetterm.com>]
8374 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8375 routines: without these tracing memory leaks is very painful.
8376 Fix leaks in PKCS12 and PKCS7 routines.
8379 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8380 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8381 effectively meant GeneralizedTime would never be used. Now it
8382 is initialised to -1 but X509_time_adj() now has to check the value
8383 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8384 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8385 [Steve Henson, reported by Kenneth R. Robinette
8386 <support@securenetterm.com>]
8388 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8389 result in a zero length in the ASN1_INTEGER structure which was
8390 not consistent with the structure when d2i_ASN1_INTEGER() was used
8391 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8392 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8393 where it did not print out a minus for negative ASN1_INTEGER.
8396 *) Add summary printout to ocsp utility. The various functions which
8397 convert status values to strings have been renamed to:
8398 OCSP_response_status_str(), OCSP_cert_status_str() and
8399 OCSP_crl_reason_str() and are no longer static. New options
8400 to verify nonce values and to disable verification. OCSP response
8401 printout format cleaned up.
8404 *) Add additional OCSP certificate checks. These are those specified
8405 in RFC2560. This consists of two separate checks: the CA of the
8406 certificate being checked must either be the OCSP signer certificate
8407 or the issuer of the OCSP signer certificate. In the latter case the
8408 OCSP signer certificate must contain the OCSP signing extended key
8409 usage. This check is performed by attempting to match the OCSP
8410 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8411 in the OCSP_CERTID structures of the response.
8414 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8415 and related routines. This uses the standard OpenSSL certificate
8416 verify routines to perform initial checks (just CA validity) and
8417 to obtain the certificate chain. Then additional checks will be
8418 performed on the chain. Currently the root CA is checked to see
8419 if it is explicitly trusted for OCSP signing. This is used to set
8420 a root CA as a global signing root: that is any certificate that
8421 chains to that CA is an acceptable OCSP signing certificate.
8424 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8425 extensions from a separate configuration file.
8426 As when reading extensions from the main configuration file,
8427 the '-extensions ...' option may be used for specifying the
8429 [Massimiliano Pala <madwolf@comune.modena.it>]
8431 *) New OCSP utility. Allows OCSP requests to be generated or
8432 read. The request can be sent to a responder and the output
8433 parsed, outputed or printed in text form. Not complete yet:
8434 still needs to check the OCSP response validity.
8437 *) New subcommands for 'openssl ca':
8438 'openssl ca -status <serial>' prints the status of the cert with
8439 the given serial number (according to the index file).
8440 'openssl ca -updatedb' updates the expiry status of certificates
8442 [Massimiliano Pala <madwolf@comune.modena.it>]
8444 *) New '-newreq-nodes' command option to CA.pl. This is like
8445 '-newreq', but calls 'openssl req' with the '-nodes' option
8446 so that the resulting key is not encrypted.
8447 [Damien Miller <djm@mindrot.org>]
8449 *) New configuration for the GNU Hurd.
8450 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8452 *) Initial code to implement OCSP basic response verify. This
8453 is currently incomplete. Currently just finds the signer's
8454 certificate and verifies the signature on the response.
8457 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
8458 value of OPENSSLDIR. This is available via the new '-d' option
8459 to 'openssl version', and is also included in 'openssl version -a'.
8462 *) Allowing defining memory allocation callbacks that will be given
8463 file name and line number information in additional arguments
8464 (a const char* and an int). The basic functionality remains, as
8465 well as the original possibility to just replace malloc(),
8466 realloc() and free() by functions that do not know about these
8467 additional arguments. To register and find out the current
8468 settings for extended allocation functions, the following
8469 functions are provided:
8471 CRYPTO_set_mem_ex_functions
8472 CRYPTO_set_locked_mem_ex_functions
8473 CRYPTO_get_mem_ex_functions
8474 CRYPTO_get_locked_mem_ex_functions
8476 These work the same way as CRYPTO_set_mem_functions and friends.
8477 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
8478 extended allocation function is enabled.
8479 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
8480 a conventional allocation function is enabled.
8481 [Richard Levitte, Bodo Moeller]
8483 *) Finish off removing the remaining LHASH function pointer casts.
8484 There should no longer be any prototype-casting required when using
8485 the LHASH abstraction, and any casts that remain are "bugs". See
8486 the callback types and macros at the head of lhash.h for details
8487 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
8490 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
8491 If /dev/[u]random devices are not available or do not return enough
8492 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
8494 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
8495 /etc/entropy will be queried once each in this sequence, querying stops
8496 when enough entropy was collected without querying more sockets.
8499 *) Change the Unix RAND_poll() variant to be able to poll several
8500 random devices, as specified by DEVRANDOM, until a sufficient amount
8501 of data has been collected. We spend at most 10 ms on each file
8502 (select timeout) and read in non-blocking mode. DEVRANDOM now
8503 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
8504 (previously it was just the string "/dev/urandom"), so on typical
8505 platforms the 10 ms delay will never occur.
8506 Also separate out the Unix variant to its own file, rand_unix.c.
8507 For VMS, there's a currently-empty rand_vms.c.
8510 *) Move OCSP client related routines to ocsp_cl.c. These
8511 provide utility functions which an application needing
8512 to issue a request to an OCSP responder and analyse the
8513 response will typically need: as opposed to those which an
8514 OCSP responder itself would need which will be added later.
8516 OCSP_request_sign() signs an OCSP request with an API similar
8517 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
8518 response. OCSP_response_get1_basic() extracts basic response
8519 from response. OCSP_resp_find_status(): finds and extracts status
8520 information from an OCSP_CERTID structure (which will be created
8521 when the request structure is built). These are built from lower
8522 level functions which work on OCSP_SINGLERESP structures but
8523 won't normally be used unless the application wishes to examine
8524 extensions in the OCSP response for example.
8526 Replace nonce routines with a pair of functions.
8527 OCSP_request_add1_nonce() adds a nonce value and optionally
8528 generates a random value. OCSP_check_nonce() checks the
8529 validity of the nonce in an OCSP response.
8532 *) Change function OCSP_request_add() to OCSP_request_add0_id().
8533 This doesn't copy the supplied OCSP_CERTID and avoids the
8534 need to free up the newly created id. Change return type
8535 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
8536 This can then be used to add extensions to the request.
8537 Deleted OCSP_request_new(), since most of its functionality
8538 is now in OCSP_REQUEST_new() (and the case insensitive name
8539 clash) apart from the ability to set the request name which
8540 will be added elsewhere.
8543 *) Update OCSP API. Remove obsolete extensions argument from
8544 various functions. Extensions are now handled using the new
8545 OCSP extension code. New simple OCSP HTTP function which
8546 can be used to send requests and parse the response.
8549 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
8550 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
8551 uses the special reorder version of SET OF to sort the attributes
8552 and reorder them to match the encoded order. This resolves a long
8553 standing problem: a verify on a PKCS7 structure just after signing
8554 it used to fail because the attribute order did not match the
8555 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
8556 it uses the received order. This is necessary to tolerate some broken
8557 software that does not order SET OF. This is handled by encoding
8558 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
8559 to produce the required SET OF.
8562 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
8563 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8564 files to get correct declarations of the ASN.1 item variables.
8567 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
8568 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
8569 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
8570 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
8571 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
8572 ASN1_ITEM and no wrapper functions.
8575 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
8576 replace the old function pointer based I/O routines. Change most of
8577 the *_d2i_bio() and *_d2i_fp() functions to use these.
8580 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
8581 lines, recognize more "algorithms" that can be deselected, and make
8582 it complain about algorithm deselection that isn't recognised.
8585 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
8586 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
8587 to use new functions. Add NO_ASN1_OLD which can be set to remove
8588 some old style ASN1 functions: this can be used to determine if old
8589 code will still work when these eventually go away.
8592 *) New extension functions for OCSP structures, these follow the
8593 same conventions as certificates and CRLs.
8596 *) New function X509V3_add1_i2d(). This automatically encodes and
8597 adds an extension. Its behaviour can be customised with various
8598 flags to append, replace or delete. Various wrappers added for
8599 certificates and CRLs.
8602 *) Fix to avoid calling the underlying ASN1 print routine when
8603 an extension cannot be parsed. Correct a typo in the
8604 OCSP_SERVICELOC extension. Tidy up print OCSP format.
8607 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
8608 entries for variables.
8611 *) Add functionality to apps/openssl.c for detecting locking
8612 problems: As the program is single-threaded, all we have
8613 to do is register a locking callback using an array for
8614 storing which locks are currently held by the program.
8617 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8618 SSL_get_ex_data_X509_STORE_idx(), which is used in
8619 ssl_verify_cert_chain() and thus can be called at any time
8620 during TLS/SSL handshakes so that thread-safety is essential.
8621 Unfortunately, the ex_data design is not at all suited
8622 for multi-threaded use, so it probably should be abolished.
8625 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8626 [Broadcom, tweaked and integrated by Geoff Thorpe]
8628 *) Move common extension printing code to new function
8629 X509V3_print_extensions(). Reorganise OCSP print routines and
8630 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8633 *) New function X509_signature_print() to remove duplication in some
8637 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8638 set (this was treated exactly the same as SET OF previously). This
8639 is used to reorder the STACK representing the structure to match the
8640 encoding. This will be used to get round a problem where a PKCS7
8641 structure which was signed could not be verified because the STACK
8642 order did not reflect the encoded order.
8645 *) Reimplement the OCSP ASN1 module using the new code.
8648 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
8649 for its ASN1 operations. The old style function pointers still exist
8650 for now but they will eventually go away.
8653 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
8654 completely replaces the old ASN1 functionality with a table driven
8655 encoder and decoder which interprets an ASN1_ITEM structure describing
8656 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
8657 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
8658 has also been converted to the new form.
8661 *) Change BN_mod_exp_recp so that negative moduli are tolerated
8662 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
8663 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
8664 for negative moduli.
8667 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
8668 of not touching the result's sign bit.
8671 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
8675 *) Changed the LHASH code to use prototypes for callbacks, and created
8676 macros to declare and implement thin (optionally static) functions
8677 that provide type-safety and avoid function pointer casting for the
8678 type-specific callbacks.
8681 *) Added Kerberos Cipher Suites to be used with TLS, as written in
8683 [Veers Staats <staatsvr@asc.hpc.mil>,
8684 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
8686 *) Reformat the FAQ so the different questions and answers can be divided
8687 in sections depending on the subject.
8690 *) Have the zlib compression code load ZLIB.DLL dynamically under
8694 *) New function BN_mod_sqrt for computing square roots modulo a prime
8695 (using the probabilistic Tonelli-Shanks algorithm unless
8696 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
8697 be handled deterministically).
8698 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8700 *) Make BN_mod_inverse faster by explicitly handling small quotients
8701 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
8702 512 bits], about 30% for larger ones [1024 or 2048 bits].)
8705 *) New function BN_kronecker.
8708 *) Fix BN_gcd so that it works on negative inputs; the result is
8709 positive unless both parameters are zero.
8710 Previously something reasonably close to an infinite loop was
8711 possible because numbers could be growing instead of shrinking
8712 in the implementation of Euclid's algorithm.
8715 *) Fix BN_is_word() and BN_is_one() macros to take into account the
8716 sign of the number in question.
8718 Fix BN_is_word(a,w) to work correctly for w == 0.
8720 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
8721 because its test if the absolute value of 'a' equals 'w'.
8722 Note that BN_abs_is_word does *not* handle w == 0 reliably;
8723 it exists mostly for use in the implementations of BN_is_zero(),
8724 BN_is_one(), and BN_is_word().
8727 *) New function BN_swap.
8730 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
8731 the exponentiation functions are more likely to produce reasonable
8732 results on negative inputs.
8735 *) Change BN_mod_mul so that the result is always non-negative.
8736 Previously, it could be negative if one of the factors was negative;
8737 I don't think anyone really wanted that behaviour.
8740 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8741 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8742 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8743 and add new functions:
8752 BN_mod_lshift1_quick
8756 These functions always generate non-negative results.
8758 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8759 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8761 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8762 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8763 be reduced modulo m.
8764 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8767 The following entry accidentally appeared in the CHANGES file
8768 distributed with OpenSSL 0.9.7. The modifications described in
8769 it do *not* apply to OpenSSL 0.9.7.
8771 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8772 was actually never needed) and in BN_mul(). The removal in BN_mul()
8773 required a small change in bn_mul_part_recursive() and the addition
8774 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8775 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8776 bn_sub_words() and bn_add_words() except they take arrays with
8781 *) In 'openssl passwd', verify passwords read from the terminal
8782 unless the '-salt' option is used (which usually means that
8783 verification would just waste user's time since the resulting
8784 hash is going to be compared with some given password hash)
8785 or the new '-noverify' option is used.
8787 This is an incompatible change, but it does not affect
8788 non-interactive use of 'openssl passwd' (passwords on the command
8789 line, '-stdin' option, '-in ...' option) and thus should not
8793 *) Remove all references to RSAref, since there's no more need for it.
8796 *) Make DSO load along a path given through an environment variable
8797 (SHLIB_PATH) with shl_load().
8800 *) Constify the ENGINE code as a result of BIGNUM constification.
8801 Also constify the RSA code and most things related to it. In a
8802 few places, most notable in the depth of the ASN.1 code, ugly
8803 casts back to non-const were required (to be solved at a later
8807 *) Make it so the openssl application has all engines loaded by default.
8810 *) Constify the BIGNUM routines a little more.
8813 *) Add the following functions:
8815 ENGINE_load_cswift()
8817 ENGINE_load_atalla()
8819 ENGINE_load_builtin_engines()
8821 That way, an application can itself choose if external engines that
8822 are built-in in OpenSSL shall ever be used or not. The benefit is
8823 that applications won't have to be linked with libdl or other dso
8824 libraries unless it's really needed.
8826 Changed 'openssl engine' to load all engines on demand.
8827 Changed the engine header files to avoid the duplication of some
8828 declarations (they differed!).
8831 *) 'openssl engine' can now list capabilities.
8834 *) Better error reporting in 'openssl engine'.
8837 *) Never call load_dh_param(NULL) in s_server.
8840 *) Add engine application. It can currently list engines by name and
8841 identity, and test if they are actually available.
8844 *) Improve RPM specification file by forcing symbolic linking and making
8845 sure the installed documentation is also owned by root.root.
8846 [Damien Miller <djm@mindrot.org>]
8848 *) Give the OpenSSL applications more possibilities to make use of
8849 keys (public as well as private) handled by engines.
8852 *) Add OCSP code that comes from CertCo.
8855 *) Add VMS support for the Rijndael code.
8858 *) Added untested support for Nuron crypto accelerator.
8861 *) Add support for external cryptographic devices. This code was
8862 previously distributed separately as the "engine" branch.
8863 [Geoff Thorpe, Richard Levitte]
8865 *) Rework the filename-translation in the DSO code. It is now possible to
8866 have far greater control over how a "name" is turned into a filename
8867 depending on the operating environment and any oddities about the
8868 different shared library filenames on each system.
8871 *) Support threads on FreeBSD-elf in Configure.
8874 *) Fix for SHA1 assembly problem with MASM: it produces
8875 warnings about corrupt line number information when assembling
8876 with debugging information. This is caused by the overlapping
8878 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
8881 NCONF_get_number() has no error checking at all. As a replacement,
8882 NCONF_get_number_e() is defined (_e for "error checking") and is
8883 promoted strongly. The old NCONF_get_number is kept around for
8884 binary backward compatibility.
8885 Make it possible for methods to load from something other than a BIO,
8886 by providing a function pointer that is given a name instead of a BIO.
8887 For example, this could be used to load configuration data from an
8891 *) Fix for non blocking accept BIOs. Added new I/O special reason
8892 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
8893 with non blocking I/O was not possible because no retry code was
8894 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
8898 *) Added the beginnings of Rijndael support.
8901 *) Fix for bug in DirectoryString mask setting. Add support for
8902 X509_NAME_print_ex() in 'req' and X509_print_ex() function
8903 to allow certificate printing to more controllable, additional
8904 'certopt' option to 'x509' to allow new printing options to be
8908 *) Clean old EAY MD5 hack from e_os.h.
8911 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
8913 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
8914 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8915 [Joe Orton, Steve Henson]
8917 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
8919 *) Fix additional bug revealed by the NISCC test suite:
8921 Stop bug triggering large recursion when presented with
8922 certain ASN.1 tags (CVE-2003-0851)
8925 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
8927 *) Fix various bugs revealed by running the NISCC test suite:
8929 Stop out of bounds reads in the ASN1 code when presented with
8930 invalid tags (CVE-2003-0543 and CVE-2003-0544).
8932 If verify callback ignores invalid public key errors don't try to check
8933 certificate signature with the NULL public key.
8937 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
8938 if the server requested one: as stated in TLS 1.0 and SSL 3.0
8942 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
8943 extra data after the compression methods not only for TLS 1.0
8944 but also for SSL 3.0 (as required by the specification).
8945 [Bodo Moeller; problem pointed out by Matthias Loepfe]
8947 *) Change X509_certificate_type() to mark the key as exported/exportable
8948 when it's 512 *bits* long, not 512 bytes.
8951 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8953 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8954 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8955 a protocol version number mismatch like a decryption error
8956 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8959 *) Turn on RSA blinding by default in the default implementation
8960 to avoid a timing attack. Applications that don't want it can call
8961 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8962 They would be ill-advised to do so in most cases.
8963 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
8965 *) Change RSA blinding code so that it works when the PRNG is not
8966 seeded (in this case, the secret RSA exponent is abused as
8967 an unpredictable seed -- if it is not unpredictable, there
8968 is no point in blinding anyway). Make RSA blinding thread-safe
8969 by remembering the creator's thread ID in rsa->blinding and
8970 having all other threads use local one-time blinding factors
8971 (this requires more computation than sharing rsa->blinding, but
8972 avoids excessive locking; and if an RSA object is not shared
8973 between threads, blinding will still be very fast).
8976 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
8978 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8979 via timing by performing a MAC computation even if incorrect
8980 block cipher padding has been found. This is a countermeasure
8981 against active attacks where the attacker has to distinguish
8982 between bad padding and a MAC verification error. (CVE-2003-0078)
8984 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8985 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8986 Martin Vuagnoux (EPFL, Ilion)]
8988 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
8990 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
8991 memory from its contents. This is done with a counter that will
8992 place alternating values in each byte. This can be used to solve
8993 two issues: 1) the removal of calls to memset() by highly optimizing
8994 compilers, and 2) cleansing with other values than 0, since those can
8995 be read through on certain media, for example a swap space on disk.
8998 *) Bugfix: client side session caching did not work with external caching,
8999 because the session->cipher setting was not restored when reloading
9000 from the external cache. This problem was masked, when
9001 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
9002 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
9005 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
9006 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
9007 [Zeev Lieber <zeev-l@yahoo.com>]
9009 *) Undo an undocumented change introduced in 0.9.6e which caused
9010 repeated calls to OpenSSL_add_all_ciphers() and
9011 OpenSSL_add_all_digests() to be ignored, even after calling
9015 *) Change the default configuration reader to deal with last line not
9016 being properly terminated.
9019 *) Change X509_NAME_cmp() so it applies the special rules on handling
9020 DN values that are of type PrintableString, as well as RDNs of type
9021 emailAddress where the value has the type ia5String.
9022 [stefank@valicert.com via Richard Levitte]
9024 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
9025 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
9026 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
9027 the bitwise-OR of the two for use by the majority of applications
9028 wanting this behaviour, and update the docs. The documented
9029 behaviour and actual behaviour were inconsistent and had been
9030 changing anyway, so this is more a bug-fix than a behavioural
9032 [Geoff Thorpe, diagnosed by Nadav Har'El]
9034 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
9035 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
9038 *) Fix initialization code race conditions in
9039 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
9040 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
9041 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
9042 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
9043 ssl2_get_cipher_by_char(),
9044 ssl3_get_cipher_by_char().
9045 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
9047 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
9048 the cached sessions are flushed, as the remove_cb() might use ex_data
9049 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
9050 (see [openssl.org #212]).
9051 [Geoff Thorpe, Lutz Jaenicke]
9053 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
9054 length, instead of the encoding length to d2i_ASN1_OBJECT.
9057 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
9059 *) [In 0.9.6g-engine release:]
9060 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
9061 [Lynn Gazis <lgazis@rainbow.com>]
9063 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
9065 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
9066 and get fix the header length calculation.
9067 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
9068 Alon Kantor <alonk@checkpoint.com> (and others),
9071 *) Use proper error handling instead of 'assertions' in buffer
9072 overflow checks added in 0.9.6e. This prevents DoS (the
9073 assertions could call abort()).
9074 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
9076 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
9078 *) Add various sanity checks to asn1_get_length() to reject
9079 the ASN1 length bytes if they exceed sizeof(long), will appear
9080 negative or the content length exceeds the length of the
9082 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9084 *) Fix cipher selection routines: ciphers without encryption had no flags
9085 for the cipher strength set and where therefore not handled correctly
9086 by the selection routines (PR #130).
9089 *) Fix EVP_dsa_sha macro.
9093 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
9094 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
9095 that was added in OpenSSL 0.9.6d.
9097 As the countermeasure turned out to be incompatible with some
9098 broken SSL implementations, the new option is part of SSL_OP_ALL.
9099 SSL_OP_ALL is usually employed when compatibility with weird SSL
9100 implementations is desired (e.g. '-bugs' option to 's_client' and
9101 's_server'), so the new option is automatically set in many
9105 *) Changes in security patch:
9107 Changes marked "(CHATS)" were sponsored by the Defense Advanced
9108 Research Projects Agency (DARPA) and Air Force Research Laboratory,
9109 Air Force Materiel Command, USAF, under agreement number
9112 *) Add various sanity checks to asn1_get_length() to reject
9113 the ASN1 length bytes if they exceed sizeof(long), will appear
9114 negative or the content length exceeds the length of the
9115 supplied buffer. (CVE-2002-0659)
9116 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9118 *) Assertions for various potential buffer overflows, not known to
9120 [Ben Laurie (CHATS)]
9122 *) Various temporary buffers to hold ASCII versions of integers were
9123 too small for 64 bit platforms. (CVE-2002-0655)
9124 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
9126 *) Remote buffer overflow in SSL3 protocol - an attacker could
9127 supply an oversized session ID to a client. (CVE-2002-0656)
9128 [Ben Laurie (CHATS)]
9130 *) Remote buffer overflow in SSL2 protocol - an attacker could
9131 supply an oversized client master key. (CVE-2002-0656)
9132 [Ben Laurie (CHATS)]
9134 Changes between 0.9.6c and 0.9.6d [9 May 2002]
9136 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
9137 encoded as NULL) with id-dsa-with-sha1.
9138 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
9140 *) Check various X509_...() return values in apps/req.c.
9141 [Nils Larsch <nla@trustcenter.de>]
9143 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
9144 an end-of-file condition would erroneously be flagged, when the CRLF
9145 was just at the end of a processed block. The bug was discovered when
9146 processing data through a buffering memory BIO handing the data to a
9147 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
9148 <ptsekov@syntrex.com> and Nedelcho Stanev.
9151 *) Implement a countermeasure against a vulnerability recently found
9152 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
9153 before application data chunks to avoid the use of known IVs
9154 with data potentially chosen by the attacker.
9157 *) Fix length checks in ssl3_get_client_hello().
9160 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
9161 to prevent ssl3_read_internal() from incorrectly assuming that
9162 ssl3_read_bytes() found application data while handshake
9163 processing was enabled when in fact s->s3->in_read_app_data was
9164 merely automatically cleared during the initial handshake.
9165 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
9167 *) Fix object definitions for Private and Enterprise: they were not
9168 recognized in their shortname (=lowercase) representation. Extend
9169 obj_dat.pl to issue an error when using undefined keywords instead
9170 of silently ignoring the problem (Svenning Sorensen
9171 <sss@sss.dnsalias.net>).
9174 *) Fix DH_generate_parameters() so that it works for 'non-standard'
9175 generators, i.e. generators other than 2 and 5. (Previously, the
9176 code did not properly initialise the 'add' and 'rem' values to
9177 BN_generate_prime().)
9179 In the new general case, we do not insist that 'generator' is
9180 actually a primitive root: This requirement is rather pointless;
9181 a generator of the order-q subgroup is just as good, if not
9185 *) Map new X509 verification errors to alerts. Discovered and submitted by
9186 Tom Wu <tom@arcot.com>.
9189 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
9190 returning non-zero before the data has been completely received
9191 when using non-blocking I/O.
9192 [Bodo Moeller; problem pointed out by John Hughes]
9194 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
9195 [Ben Laurie, Lutz Jaenicke]
9197 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
9198 Yoram Zahavi <YoramZ@gilian.com>).
9201 *) Add information about CygWin 1.3 and on, and preserve proper
9202 configuration for the versions before that.
9203 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
9205 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
9206 check whether we deal with a copy of a session and do not delete from
9207 the cache in this case. Problem reported by "Izhar Shoshani Levi"
9208 <izhar@checkpoint.com>.
9211 *) Do not store session data into the internal session cache, if it
9212 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
9213 flag is set). Proposed by Aslam <aslam@funk.com>.
9216 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
9220 *) [In 0.9.6d-engine release:]
9221 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
9222 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
9224 *) Add the configuration target linux-s390x.
9225 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
9227 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
9228 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
9229 variable as an indication that a ClientHello message has been
9230 received. As the flag value will be lost between multiple
9231 invocations of ssl3_accept when using non-blocking I/O, the
9232 function may not be aware that a handshake has actually taken
9233 place, thus preventing a new session from being added to the
9236 To avoid this problem, we now set s->new_session to 2 instead of
9237 using a local variable.
9238 [Lutz Jaenicke, Bodo Moeller]
9240 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
9241 if the SSL_R_LENGTH_MISMATCH error is detected.
9242 [Geoff Thorpe, Bodo Moeller]
9244 *) New 'shared_ldflag' column in Configure platform table.
9247 *) Fix EVP_CIPHER_mode macro.
9248 ["Dan S. Camper" <dan@bti.net>]
9250 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
9251 type, we must throw them away by setting rr->length to 0.
9252 [D P Chang <dpc@qualys.com>]
9254 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
9256 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
9257 <Dominikus.Scherkl@biodata.com>. (The previous implementation
9258 worked incorrectly for those cases where range = 10..._2 and
9259 3*range is two bits longer than range.)
9262 *) Only add signing time to PKCS7 structures if it is not already
9266 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
9267 OBJ_ld_ce should be OBJ_id_ce.
9268 Also some ip-pda OIDs in crypto/objects/objects.txt were
9269 incorrect (cf. RFC 3039).
9270 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
9272 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
9273 returns early because it has nothing to do.
9274 [Andy Schneider <andy.schneider@bjss.co.uk>]
9276 *) [In 0.9.6c-engine release:]
9277 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
9278 [Andy Schneider <andy.schneider@bjss.co.uk>]
9280 *) [In 0.9.6c-engine release:]
9281 Add support for Cryptographic Appliance's keyserver technology.
9282 (Use engine 'keyclient')
9283 [Cryptographic Appliances and Geoff Thorpe]
9285 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
9286 is called via tools/c89.sh because arguments have to be
9287 rearranged (all '-L' options must appear before the first object
9289 [Richard Shapiro <rshapiro@abinitio.com>]
9291 *) [In 0.9.6c-engine release:]
9292 Add support for Broadcom crypto accelerator cards, backported
9294 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
9296 *) [In 0.9.6c-engine release:]
9297 Add support for SureWare crypto accelerator cards from
9298 Baltimore Technologies. (Use engine 'sureware')
9299 [Baltimore Technologies and Mark Cox]
9301 *) [In 0.9.6c-engine release:]
9302 Add support for crypto accelerator cards from Accelerated
9303 Encryption Processing, www.aep.ie. (Use engine 'aep')
9304 [AEP Inc. and Mark Cox]
9306 *) Add a configuration entry for gcc on UnixWare.
9307 [Gary Benson <gbenson@redhat.com>]
9309 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
9310 messages are stored in a single piece (fixed-length part and
9311 variable-length part combined) and fix various bugs found on the way.
9314 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9315 instead. BIO_gethostbyname() does not know what timeouts are
9316 appropriate, so entries would stay in cache even when they have
9318 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9320 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9321 faced with a pathologically small ClientHello fragment that does
9322 not contain client_version: Instead of aborting with an error,
9323 simply choose the highest available protocol version (i.e.,
9324 TLS 1.0 unless it is disabled). In practice, ClientHello
9325 messages are never sent like this, but this change gives us
9326 strictly correct behaviour at least for TLS.
9329 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9330 never resets s->method to s->ctx->method when called from within
9331 one of the SSL handshake functions.
9332 [Bodo Moeller; problem pointed out by Niko Baric]
9334 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9335 (sent using the client's version number) if client_version is
9336 smaller than the protocol version in use. Also change
9337 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9338 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9339 the client will at least see that alert.
9342 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9346 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9347 client receives HelloRequest while in a handshake.
9348 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9350 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9351 should end in 'break', not 'goto end' which circumvents various
9352 cleanups done in state SSL_ST_OK. But session related stuff
9353 must be disabled for SSL_ST_OK in the case that we just sent a
9356 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9357 before just sending a HelloRequest.
9358 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9360 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9361 reveal whether illegal block cipher padding was found or a MAC
9362 verification error occurred. (Neither SSLerr() codes nor alerts
9363 are directly visible to potential attackers, but the information
9364 may leak via logfiles.)
9366 Similar changes are not required for the SSL 2.0 implementation
9367 because the number of padding bytes is sent in clear for SSL 2.0,
9368 and the extra bytes are just ignored. However ssl/s2_pkt.c
9369 failed to verify that the purported number of padding bytes is in
9373 *) Add OpenUNIX-8 support including shared libraries
9374 (Boyd Lynn Gerber <gerberb@zenez.com>).
9377 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9378 'wristwatch attack' using huge encoding parameters (cf.
9379 James H. Manger's CRYPTO 2001 paper). Note that the
9380 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9381 encoding parameters and hence was not vulnerable.
9384 *) BN_sqr() bug fix.
9385 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9387 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9388 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9389 followed by modular reduction.
9390 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9392 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9393 equivalent based on BN_pseudo_rand() instead of BN_rand().
9396 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9397 This function was broken, as the check for a new client hello message
9398 to handle SGC did not allow these large messages.
9399 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9402 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9405 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9406 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9409 *) Rework the configuration and shared library support for Tru64 Unix.
9410 The configuration part makes use of modern compiler features and
9411 still retains old compiler behavior for those that run older versions
9412 of the OS. The shared library support part includes a variant that
9413 uses the RPATH feature, and is available through the special
9414 configuration target "alpha-cc-rpath", which will never be selected
9416 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9418 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9419 with the same message size as in ssl3_get_certificate_request().
9420 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9421 messages might inadvertently be reject as too long.
9422 [Petr Lampa <lampa@fee.vutbr.cz>]
9424 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9427 *) Modified SSL library such that the verify_callback that has been set
9428 specificly for an SSL object with SSL_set_verify() is actually being
9429 used. Before the change, a verify_callback set with this function was
9430 ignored and the verify_callback() set in the SSL_CTX at the time of
9431 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9432 to allow the necessary settings.
9435 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9436 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9437 done automatically (in contradiction to the requirements of the C
9438 standard). This made problems when used from OpenSSH.
9441 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9442 dh->length and always used
9444 BN_rand_range(priv_key, dh->p).
9446 BN_rand_range() is not necessary for Diffie-Hellman, and this
9447 specific range makes Diffie-Hellman unnecessarily inefficient if
9448 dh->length (recommended exponent length) is much smaller than the
9449 length of dh->p. We could use BN_rand_range() if the order of
9450 the subgroup was stored in the DH structure, but we only have
9455 BN_rand(priv_key, l, ...)
9457 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
9463 RSA_eay_public_encrypt
9464 RSA_eay_private_decrypt
9465 RSA_eay_private_encrypt (signing)
9466 RSA_eay_public_decrypt (signature verification)
9468 (default implementations for RSA_public_encrypt,
9469 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
9470 always reject numbers >= n.
9473 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
9474 to synchronize access to 'locking_thread'. This is necessary on
9475 systems where access to 'locking_thread' (an 'unsigned long'
9476 variable) is not atomic.
9479 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
9480 *before* setting the 'crypto_lock_rand' flag. The previous code had
9481 a race condition if 0 is a valid thread ID.
9482 [Travis Vitek <vitek@roguewave.com>]
9484 *) Add support for shared libraries under Irix.
9485 [Albert Chin-A-Young <china@thewrittenword.com>]
9487 *) Add configuration option to build on Linux on both big-endian and
9489 [Ralf Baechle <ralf@uni-koblenz.de>]
9491 *) Add the possibility to create shared libraries on HP-UX.
9494 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
9496 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
9497 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
9498 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
9499 PRNG state recovery was possible based on the output of
9500 one PRNG request appropriately sized to gain knowledge on
9501 'md' followed by enough consecutive 1-byte PRNG requests
9502 to traverse all of 'state'.
9504 1. When updating 'md_local' (the current thread's copy of 'md')
9505 during PRNG output generation, hash all of the previous
9506 'md_local' value, not just the half used for PRNG output.
9508 2. Make the number of bytes from 'state' included into the hash
9509 independent from the number of PRNG bytes requested.
9511 The first measure alone would be sufficient to avoid
9512 Markku-Juhani's attack. (Actually it had never occurred
9513 to me that the half of 'md_local' used for chaining was the
9514 half from which PRNG output bytes were taken -- I had always
9515 assumed that the secret half would be used.) The second
9516 measure makes sure that additional data from 'state' is never
9517 mixed into 'md_local' in small portions; this heuristically
9518 further strengthens the PRNG.
9521 *) Fix crypto/bn/asm/mips3.s.
9524 *) When only the key is given to "enc", the IV is undefined. Print out
9525 an error message in this case.
9528 *) Handle special case when X509_NAME is empty in X509 printing routines.
9531 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
9532 positive and less than q.
9535 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
9536 used: it isn't thread safe and the add_lock_callback should handle
9538 [Paul Rose <Paul.Rose@bridge.com>]
9540 *) Verify that incoming data obeys the block size in
9541 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
9545 [Ulf Möller, Bodo Möller]
9547 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
9548 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
9549 when fixing the server behaviour for backwards-compatible 'client
9550 hello' messages. (Note that the attack is impractical against
9551 SSL 3.0 and TLS 1.0 anyway because length and version checking
9552 means that the probability of guessing a valid ciphertext is
9553 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
9556 Before 0.9.5, the countermeasure (hide the error by generating a
9557 random 'decryption result') did not work properly because
9558 ERR_clear_error() was missing, meaning that SSL_get_error() would
9559 detect the supposedly ignored error.
9561 Both problems are now fixed.
9564 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
9565 (previously it was 1024).
9568 *) Fix for compatibility mode trust settings: ignore trust settings
9569 unless some valid trust or reject settings are present.
9572 *) Fix for blowfish EVP: its a variable length cipher.
9575 *) Fix various bugs related to DSA S/MIME verification. Handle missing
9576 parameters in DSA public key structures and return an error in the
9577 DSA routines if parameters are absent.
9580 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
9581 in the current directory if neither $RANDFILE nor $HOME was set.
9582 RAND_file_name() in 0.9.6a returned NULL in this case. This has
9583 caused some confusion to Windows users who haven't defined $HOME.
9584 Thus RAND_file_name() is changed again: e_os.h can define a
9585 DEFAULT_HOME, which will be used if $HOME is not set.
9586 For Windows, we use "C:"; on other platforms, we still require
9587 environment variables.
9589 *) Move 'if (!initialized) RAND_poll()' into regions protected by
9590 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
9591 having multiple threads call RAND_poll() concurrently.
9594 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
9595 combination of a flag and a thread ID variable.
9596 Otherwise while one thread is in ssleay_rand_bytes (which sets the
9597 flag), *other* threads can enter ssleay_add_bytes without obeying
9598 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
9599 that they do not hold after the first thread unsets add_do_not_lock).
9602 *) Change bctest again: '-x' expressions are not available in all
9606 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
9608 *) Fix a couple of memory leaks in PKCS7_dataDecode()
9609 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9611 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9612 the default extension for executables, if any. Also, make the perl
9613 scripts that use symlink() to test if it really exists and use "cp"
9614 if it doesn't. All this made OpenSSL compilable and installable in
9618 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9619 If SEQUENCE is length is indefinite just set c->slen to the total
9620 amount of data available.
9621 [Steve Henson, reported by shige@FreeBSD.org]
9622 [This change does not apply to 0.9.7.]
9624 *) Change bctest to avoid here-documents inside command substitution
9625 (workaround for FreeBSD /bin/sh bug).
9626 For compatibility with Ultrix, avoid shell functions (introduced
9627 in the bctest version that searches along $PATH).
9630 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9631 with des_encrypt() defined on some operating systems, like Solaris
9635 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9636 On the Importance of Eliminating Errors in Cryptographic
9637 Computations, J. Cryptology 14 (2001) 2, 101-119,
9638 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9641 *) MIPS assembler BIGNUM division bug fix.
9644 *) Disabled incorrect Alpha assembler code.
9647 *) Fix PKCS#7 decode routines so they correctly update the length
9648 after reading an EOC for the EXPLICIT tag.
9650 [This change does not apply to 0.9.7.]
9652 *) Fix bug in PKCS#12 key generation routines. This was triggered
9653 if a 3DES key was generated with a 0 initial byte. Include
9654 PKCS12_BROKEN_KEYGEN compilation option to retain the old
9655 (but broken) behaviour.
9658 *) Enhance bctest to search for a working bc along $PATH and print
9660 [Tim Rice <tim@multitalents.net> via Richard Levitte]
9662 *) Fix memory leaks in err.c: free err_data string if necessary;
9663 don't write to the wrong index in ERR_set_error_data.
9666 *) Implement ssl23_peek (analogous to ssl23_read), which previously
9670 *) Replace rdtsc with _emit statements for VC++ version 5.
9671 [Jeremy Cooper <jeremy@baymoo.org>]
9673 *) Make it possible to reuse SSLv2 sessions.
9676 *) In copy_email() check for >= 0 as a return value for
9677 X509_NAME_get_index_by_NID() since 0 is a valid index.
9678 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
9680 *) Avoid coredump with unsupported or invalid public keys by checking if
9681 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
9682 PKCS7_verify() fails with non detached data.
9685 *) Don't use getenv in library functions when run as setuid/setgid.
9686 New function OPENSSL_issetugid().
9689 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
9690 due to incorrect handling of multi-threading:
9692 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
9694 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
9696 3. Count how many times MemCheck_off() has been called so that
9697 nested use can be treated correctly. This also avoids
9698 inband-signalling in the previous code (which relied on the
9699 assumption that thread ID 0 is impossible).
9702 *) Add "-rand" option also to s_client and s_server.
9705 *) Fix CPU detection on Irix 6.x.
9706 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
9707 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9709 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
9712 [This change does not apply to 0.9.7.]
9714 *) Use the cached encoding of an X509_NAME structure rather than
9715 copying it. This is apparently the reason for the libsafe "errors"
9716 but the code is actually correct.
9719 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
9720 Bleichenbacher's DSA attack.
9721 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
9722 to be set and top=0 forces the highest bit to be set; top=-1 is new
9723 and leaves the highest bit random.
9724 [Ulf Moeller, Bodo Moeller]
9726 *) In the NCONF_...-based implementations for CONF_... queries
9727 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
9728 a temporary CONF structure with the data component set to NULL
9729 (which gives segmentation faults in lh_retrieve).
9730 Instead, use NULL for the CONF pointer in CONF_get_string and
9731 CONF_get_number (which may use environment variables) and directly
9732 return NULL from CONF_get_section.
9735 *) Fix potential buffer overrun for EBCDIC.
9738 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9739 keyUsage if basicConstraints absent for a CA.
9742 *) Make SMIME_write_PKCS7() write mail header values with a format that
9743 is more generally accepted (no spaces before the semicolon), since
9744 some programs can't parse those values properly otherwise. Also make
9745 sure BIO's that break lines after each write do not create invalid
9749 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9750 macros previously used would not encode an empty SEQUENCE OF
9751 and break the signature.
9753 [This change does not apply to 0.9.7.]
9755 *) Zero the premaster secret after deriving the master secret in
9759 *) Add some EVP_add_digest_alias registrations (as found in
9760 OpenSSL_add_all_digests()) to SSL_library_init()
9761 aka OpenSSL_add_ssl_algorithms(). This provides improved
9762 compatibility with peers using X.509 certificates
9763 with unconventional AlgorithmIdentifier OIDs.
9766 *) Fix for Irix with NO_ASM.
9767 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9769 *) ./config script fixes.
9770 [Ulf Moeller, Richard Levitte]
9772 *) Fix 'openssl passwd -1'.
9775 *) Change PKCS12_key_gen_asc() so it can cope with non null
9776 terminated strings whose length is passed in the passlen
9777 parameter, for example from PEM callbacks. This was done
9778 by adding an extra length parameter to asc2uni().
9779 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9781 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9782 call failed, free the DSA structure.
9785 *) Fix to uni2asc() to cope with zero length Unicode strings.
9786 These are present in some PKCS#12 files.
9789 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9790 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9791 when writing a 32767 byte record.
9792 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9794 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9795 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9797 (RSA objects have a reference count access to which is protected
9798 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9799 so they are meant to be shared between threads.)
9800 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9801 "Reddie, Steven" <Steven.Reddie@ca.com>]
9803 *) Fix a deadlock in CRYPTO_mem_leaks().
9806 *) Use better test patterns in bntest.
9809 *) rand_win.c fix for Borland C.
9812 *) BN_rshift bugfix for n == 0.
9815 *) Add a 'bctest' script that checks for some known 'bc' bugs
9816 so that 'make test' does not abort just because 'bc' is broken.
9819 *) Store verify_result within SSL_SESSION also for client side to
9820 avoid potential security hole. (Re-used sessions on the client side
9821 always resulted in verify_result==X509_V_OK, not using the original
9822 result of the server certificate verification.)
9825 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9826 SSL3_RT_APPLICATION_DATA, return 0.
9827 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9831 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9832 releases, have been re-implemented by renaming the previous
9833 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9834 and ssl3_read_internal, respectively, and adding 'peek' parameters
9835 to them. The new ssl[23]_{read,peek} functions are calls to
9836 ssl[23]_read_internal with the 'peek' flag set appropriately.
9837 A 'peek' parameter has also been added to ssl3_read_bytes, which
9838 does the actual work for ssl3_read_internal.
9841 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9842 the method-specific "init()" handler. Also clean up ex_data after
9843 calling the method-specific "finish()" handler. Previously, this was
9844 happening the other way round.
9847 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9848 The previous value, 12, was not always sufficient for BN_mod_exp().
9851 *) Make sure that shared libraries get the internal name engine with
9852 the full version number and not just 0. This should mark the
9853 shared libraries as not backward compatible. Of course, this should
9854 be changed again when we can guarantee backward binary compatibility.
9857 *) Fix typo in get_cert_by_subject() in by_dir.c
9858 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9860 *) Rework the system to generate shared libraries:
9862 - Make note of the expected extension for the shared libraries and
9863 if there is a need for symbolic links from for example libcrypto.so.0
9864 to libcrypto.so.0.9.7. There is extended info in Configure for
9867 - Make as few rebuilds of the shared libraries as possible.
9869 - Still avoid linking the OpenSSL programs with the shared libraries.
9871 - When installing, install the shared libraries separately from the
9875 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
9877 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
9878 and not in SSL_clear because the latter is also used by the
9879 accept/connect functions; previously, the settings made by
9880 SSL_set_read_ahead would be lost during the handshake.
9881 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
9883 *) Correct util/mkdef.pl to be selective about disabled algorithms.
9884 Previously, it would create entries for disabled algorithms no
9888 *) Added several new manual pages for SSL_* function.
9891 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
9893 *) In ssl23_get_client_hello, generate an error message when faced
9894 with an initial SSL 3.0/TLS record that is too small to contain the
9895 first two bytes of the ClientHello message, i.e. client_version.
9896 (Note that this is a pathologic case that probably has never happened
9897 in real life.) The previous approach was to use the version number
9898 from the record header as a substitute; but our protocol choice
9899 should not depend on that one because it is not authenticated
9900 by the Finished messages.
9903 *) More robust randomness gathering functions for Windows.
9904 [Jeffrey Altman <jaltman@columbia.edu>]
9906 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
9907 not set then we don't setup the error code for issuer check errors
9908 to avoid possibly overwriting other errors which the callback does
9909 handle. If an application does set the flag then we assume it knows
9910 what it is doing and can handle the new informational codes
9914 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
9915 a general "ANY" type, as such it should be able to decode anything
9916 including tagged types. However it didn't check the class so it would
9917 wrongly interpret tagged types in the same way as their universal
9918 counterpart and unknown types were just rejected. Changed so that the
9919 tagged and unknown types are handled in the same way as a SEQUENCE:
9920 that is the encoding is stored intact. There is also a new type
9921 "V_ASN1_OTHER" which is used when the class is not universal, in this
9922 case we have no idea what the actual type is so we just lump them all
9926 *) On VMS, stdout may very well lead to a file that is written to
9927 in a record-oriented fashion. That means that every write() will
9928 write a separate record, which will be read separately by the
9929 programs trying to read from it. This can be very confusing.
9931 The solution is to put a BIO filter in the way that will buffer
9932 text until a linefeed is reached, and then write everything a
9933 line at a time, so every record written will be an actual line,
9934 not chunks of lines and not (usually doesn't happen, but I've
9935 seen it once) several lines in one record. BIO_f_linebuffer() is
9938 Currently, it's a VMS-only method, because that's where it has
9939 been tested well enough.
9942 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
9943 it can return incorrect results.
9944 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
9945 but it was in 0.9.6-beta[12].)
9948 *) Disable the check for content being present when verifying detached
9949 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9950 include zero length content when signing messages.
9953 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9954 BIO_ctrl (for BIO pairs).
9957 *) Add DSO method for VMS.
9960 *) Bug fix: Montgomery multiplication could produce results with the
9964 *) Add RPM specification openssl.spec and modify it to build three
9965 packages. The default package contains applications, application
9966 documentation and run-time libraries. The devel package contains
9967 include files, static libraries and function documentation. The
9968 doc package contains the contents of the doc directory. The original
9969 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
9972 *) Add a large number of documentation files for many SSL routines.
9973 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9975 *) Add a configuration entry for Sony News 4.
9976 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
9978 *) Don't set the two most significant bits to one when generating a
9979 random number < q in the DSA library.
9982 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
9983 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
9984 the underlying transport is blocking) if a handshake took place.
9985 (The default behaviour is needed by applications such as s_client
9986 and s_server that use select() to determine when to use SSL_read;
9987 but for applications that know in advance when to expect data, it
9988 just makes things more complicated.)
9991 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
9995 *) Add a few more EBCDIC conditionals that make `req' and `x509'
9996 work better on such systems.
9997 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9999 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
10000 Update PKCS12_parse() so it copies the friendlyName and the
10001 keyid to the certificates aux info.
10004 *) Fix bug in PKCS7_verify() which caused an infinite loop
10005 if there was more than one signature.
10006 [Sven Uszpelkat <su@celocom.de>]
10008 *) Major change in util/mkdef.pl to include extra information
10009 about each symbol, as well as presenting variables as well
10010 as functions. This change means that there's n more need
10011 to rebuild the .num files when some algorithms are excluded.
10014 *) Allow the verify time to be set by an application,
10015 rather than always using the current time.
10018 *) Phase 2 verify code reorganisation. The certificate
10019 verify code now looks up an issuer certificate by a
10020 number of criteria: subject name, authority key id
10021 and key usage. It also verifies self signed certificates
10022 by the same criteria. The main comparison function is
10023 X509_check_issued() which performs these checks.
10025 Lot of changes were necessary in order to support this
10026 without completely rewriting the lookup code.
10028 Authority and subject key identifier are now cached.
10030 The LHASH 'certs' is X509_STORE has now been replaced
10031 by a STACK_OF(X509_OBJECT). This is mainly because an
10032 LHASH can't store or retrieve multiple objects with
10033 the same hash value.
10035 As a result various functions (which were all internal
10036 use only) have changed to handle the new X509_STORE
10037 structure. This will break anything that messed round
10038 with X509_STORE internally.
10040 The functions X509_STORE_add_cert() now checks for an
10041 exact match, rather than just subject name.
10043 The X509_STORE API doesn't directly support the retrieval
10044 of multiple certificates matching a given criteria, however
10045 this can be worked round by performing a lookup first
10046 (which will fill the cache with candidate certificates)
10047 and then examining the cache for matches. This is probably
10048 the best we can do without throwing out X509_LOOKUP
10049 entirely (maybe later...).
10051 The X509_VERIFY_CTX structure has been enhanced considerably.
10053 All certificate lookup operations now go via a get_issuer()
10054 callback. Although this currently uses an X509_STORE it
10055 can be replaced by custom lookups. This is a simple way
10056 to bypass the X509_STORE hackery necessary to make this
10057 work and makes it possible to use more efficient techniques
10058 in future. A very simple version which uses a simple
10059 STACK for its trusted certificate store is also provided
10060 using X509_STORE_CTX_trusted_stack().
10062 The verify_cb() and verify() callbacks now have equivalents
10063 in the X509_STORE_CTX structure.
10065 X509_STORE_CTX also has a 'flags' field which can be used
10066 to customise the verify behaviour.
10069 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
10070 excludes S/MIME capabilities.
10073 *) When a certificate request is read in keep a copy of the
10074 original encoding of the signed data and use it when outputting
10075 again. Signatures then use the original encoding rather than
10076 a decoded, encoded version which may cause problems if the
10077 request is improperly encoded.
10080 *) For consistency with other BIO_puts implementations, call
10081 buffer_write(b, ...) directly in buffer_puts instead of calling
10084 In BIO_puts, increment b->num_write as in BIO_write.
10085 [Peter.Sylvester@EdelWeb.fr]
10087 *) Fix BN_mul_word for the case where the word is 0. (We have to use
10088 BN_zero, we may not return a BIGNUM with an array consisting of
10089 words set to zero.)
10092 *) Avoid calling abort() from within the library when problems are
10093 detected, except if preprocessor symbols have been defined
10094 (such as REF_CHECK, BN_DEBUG etc.).
10097 *) New openssl application 'rsautl'. This utility can be
10098 used for low level RSA operations. DER public key
10099 BIO/fp routines also added.
10102 *) New Configure entry and patches for compiling on QNX 4.
10103 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
10105 *) A demo state-machine implementation was sponsored by
10106 Nuron (http://www.nuron.com/) and is now available in
10107 demos/state_machine.
10110 *) New options added to the 'dgst' utility for signature
10111 generation and verification.
10114 *) Unrecognized PKCS#7 content types are now handled via a
10115 catch all ASN1_TYPE structure. This allows unsupported
10116 types to be stored as a "blob" and an application can
10117 encode and decode it manually.
10120 *) Fix various signed/unsigned issues to make a_strex.c
10121 compile under VC++.
10122 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
10124 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
10125 length if passed a buffer. ASN1_INTEGER_to_BN failed
10126 if passed a NULL BN and its argument was negative.
10127 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
10129 *) Modification to PKCS#7 encoding routines to output definite
10130 length encoding. Since currently the whole structures are in
10131 memory there's not real point in using indefinite length
10132 constructed encoding. However if OpenSSL is compiled with
10133 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
10136 *) Added BIO_vprintf() and BIO_vsnprintf().
10139 *) Added more prefixes to parse for in the strings written
10140 through a logging bio, to cover all the levels that are available
10141 through syslog. The prefixes are now:
10143 PANIC, EMERG, EMR => LOG_EMERG
10144 ALERT, ALR => LOG_ALERT
10145 CRIT, CRI => LOG_CRIT
10146 ERROR, ERR => LOG_ERR
10147 WARNING, WARN, WAR => LOG_WARNING
10148 NOTICE, NOTE, NOT => LOG_NOTICE
10149 INFO, INF => LOG_INFO
10150 DEBUG, DBG => LOG_DEBUG
10152 and as before, if none of those prefixes are present at the
10153 beginning of the string, LOG_ERR is chosen.
10155 On Win32, the LOG_* levels are mapped according to this:
10157 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
10158 LOG_WARNING => EVENTLOG_WARNING_TYPE
10159 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
10163 *) Made it possible to reconfigure with just the configuration
10164 argument "reconf" or "reconfigure". The command line arguments
10165 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
10166 and are retrieved from there when reconfiguring.
10169 *) MD4 implemented.
10170 [Assar Westerlund <assar@sics.se>, Richard Levitte]
10172 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
10175 *) The obj_dat.pl script was messing up the sorting of object
10176 names. The reason was that it compared the quoted version
10177 of strings as a result "OCSP" > "OCSP Signing" because
10178 " > SPACE. Changed script to store unquoted versions of
10179 names and add quotes on output. It was also omitting some
10180 names from the lookup table if they were given a default
10181 value (that is if SN is missing it is given the same
10182 value as LN and vice versa), these are now added on the
10183 grounds that if an object has a name we should be able to
10184 look it up. Finally added warning output when duplicate
10185 short or long names are found.
10188 *) Changes needed for Tandem NSK.
10189 [Scott Uroff <scott@xypro.com>]
10191 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
10192 RSA_padding_check_SSLv23(), special padding was never detected
10193 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
10194 version rollback attacks was not effective.
10196 In s23_clnt.c, don't use special rollback-attack detection padding
10197 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
10198 client; similarly, in s23_srvr.c, don't do the rollback check if
10199 SSL 2.0 is the only protocol enabled in the server.
10202 *) Make it possible to get hexdumps of unprintable data with 'openssl
10203 asn1parse'. By implication, the functions ASN1_parse_dump() and
10204 BIO_dump_indent() are added.
10207 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
10208 these print out strings and name structures based on various
10209 flags including RFC2253 support and proper handling of
10210 multibyte characters. Added options to the 'x509' utility
10211 to allow the various flags to be set.
10214 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
10215 Also change the functions X509_cmp_current_time() and
10216 X509_gmtime_adj() work with an ASN1_TIME structure,
10217 this will enable certificates using GeneralizedTime in validity
10218 dates to be checked.
10221 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
10222 negative public key encodings) on by default,
10223 NO_NEG_PUBKEY_BUG can be set to disable it.
10226 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
10227 content octets. An i2c_ASN1_OBJECT is unnecessary because
10228 the encoding can be trivially obtained from the structure.
10231 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
10232 not read locks (CRYPTO_r_[un]lock).
10235 *) A first attempt at creating official support for shared
10236 libraries through configuration. I've kept it so the
10237 default is static libraries only, and the OpenSSL programs
10238 are always statically linked for now, but there are
10239 preparations for dynamic linking in place.
10240 This has been tested on Linux and Tru64.
10243 *) Randomness polling function for Win9x, as described in:
10244 Peter Gutmann, Software Generation of Practically Strong
10248 *) Fix so PRNG is seeded in req if using an already existing
10252 *) New options to smime application. -inform and -outform
10253 allow alternative formats for the S/MIME message including
10254 PEM and DER. The -content option allows the content to be
10255 specified separately. This should allow things like Netscape
10256 form signing output easier to verify.
10259 *) Fix the ASN1 encoding of tags using the 'long form'.
10262 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
10263 STRING types. These convert content octets to and from the
10264 underlying type. The actual tag and length octets are
10265 already assumed to have been read in and checked. These
10266 are needed because all other string types have virtually
10267 identical handling apart from the tag. By having versions
10268 of the ASN1 functions that just operate on content octets
10269 IMPLICIT tagging can be handled properly. It also allows
10270 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
10271 and ASN1_INTEGER are identical apart from the tag.
10274 *) Change the handling of OID objects as follows:
10276 - New object identifiers are inserted in objects.txt, following
10277 the syntax given in objects.README.
10278 - objects.pl is used to process obj_mac.num and create a new
10280 - obj_dat.pl is used to create a new obj_dat.h, using the data in
10283 This is currently kind of a hack, and the perl code in objects.pl
10284 isn't very elegant, but it works as I intended. The simplest way
10285 to check that it worked correctly is to look in obj_dat.h and
10286 check the array nid_objs and make sure the objects haven't moved
10287 around (this is important!). Additions are OK, as well as
10288 consistent name changes.
10291 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
10294 *) Addition of the command line parameter '-rand file' to 'openssl req'.
10295 The given file adds to whatever has already been seeded into the
10296 random pool through the RANDFILE configuration file option or
10297 environment variable, or the default random state file.
10300 *) mkstack.pl now sorts each macro group into lexical order.
10301 Previously the output order depended on the order the files
10302 appeared in the directory, resulting in needless rewriting
10306 *) Patches to make OpenSSL compile under Win32 again. Mostly
10307 work arounds for the VC++ problem that it treats func() as
10308 func(void). Also stripped out the parts of mkdef.pl that
10309 added extra typesafe functions: these no longer exist.
10312 *) Reorganisation of the stack code. The macros are now all
10313 collected in safestack.h . Each macro is defined in terms of
10314 a "stack macro" of the form SKM_<name>(type, a, b). The
10315 DEBUG_SAFESTACK is now handled in terms of function casts,
10316 this has the advantage of retaining type safety without the
10317 use of additional functions. If DEBUG_SAFESTACK is not defined
10318 then the non typesafe macros are used instead. Also modified the
10319 mkstack.pl script to handle the new form. Needs testing to see
10320 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10321 the default if no major problems. Similar behaviour for ASN1_SET_OF
10322 and PKCS12_STACK_OF.
10325 *) When some versions of IIS use the 'NET' form of private key the
10326 key derivation algorithm is different. Normally MD5(password) is
10327 used as a 128 bit RC4 key. In the modified case
10328 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10329 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10330 as the old Netscape_RSA functions except they have an additional
10331 'sgckey' parameter which uses the modified algorithm. Also added
10332 an -sgckey command line option to the rsa utility. Thanks to
10333 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10334 algorithm to openssl-dev.
10337 *) The evp_local.h macros were using 'c.##kname' which resulted in
10338 invalid expansion on some systems (SCO 5.0.5 for example).
10339 Corrected to 'c.kname'.
10340 [Phillip Porch <root@theporch.com>]
10342 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10343 a STACK of email addresses from a certificate or request, these look
10344 in the subject name and the subject alternative name extensions and
10345 omit any duplicate addresses.
10348 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10349 This makes DSA verification about 2 % faster.
10352 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10353 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10354 plus overhead for 1024 bit moduli).
10355 This makes exponentiations about 0.5 % faster for 1024 bit
10356 exponents (as measured by "openssl speed rsa2048").
10359 *) Rename memory handling macros to avoid conflicts with other
10361 Malloc => OPENSSL_malloc
10362 Malloc_locked => OPENSSL_malloc_locked
10363 Realloc => OPENSSL_realloc
10364 Free => OPENSSL_free
10367 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10368 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10371 *) CygWin32 support.
10372 [John Jarvie <jjarvie@newsguy.com>]
10374 *) The type-safe stack code has been rejigged. It is now only compiled
10375 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10376 by default all type-specific stack functions are "#define"d back to
10377 standard stack functions. This results in more streamlined output
10378 but retains the type-safety checking possibilities of the original
10382 *) The STACK code has been cleaned up, and certain type declarations
10383 that didn't make a lot of sense have been brought in line. This has
10384 also involved a cleanup of sorts in safestack.h to more correctly
10385 map type-safe stack functions onto their plain stack counterparts.
10386 This work has also resulted in a variety of "const"ifications of
10387 lots of the code, especially "_cmp" operations which should normally
10388 be prototyped with "const" parameters anyway.
10391 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10392 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10393 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10394 where all of 'md' is used each time the PRNG is used, but 'state'
10395 is used only indexed by a cyclic counter. As entropy may not be
10396 well distributed from the beginning, 'md' is important as a
10397 chaining variable. However, the output function chains only half
10398 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10399 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10400 in all of 'state' being rewritten, with the new values depending
10401 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10404 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10405 the handshake is continued after ssl_verify_cert_chain();
10406 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10407 can lead to 'unexplainable' connection aborts later.
10408 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10410 *) Major EVP API cipher revision.
10411 Add hooks for extra EVP features. This allows various cipher
10412 parameters to be set in the EVP interface. Support added for variable
10413 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10414 setting of RC2 and RC5 parameters.
10416 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10419 Remove lots of duplicated code from the EVP library. For example *every*
10420 cipher init() function handles the 'iv' in the same way according to the
10421 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10422 for CFB and OFB modes they zero ctx->num.
10424 New functionality allows removal of S/MIME code RC2 hack.
10426 Most of the routines have the same form and so can be declared in terms
10429 By shifting this to the top level EVP_CipherInit() it can be removed from
10430 all individual ciphers. If the cipher wants to handle IVs or keys
10431 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10434 Change lots of functions like EVP_EncryptUpdate() to now return a
10435 value: although software versions of the algorithms cannot fail
10436 any installed hardware versions can.
10439 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10440 this option is set, tolerate broken clients that send the negotiated
10441 protocol version number instead of the requested protocol version
10445 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10446 i.e. non-zero for export ciphersuites, zero otherwise.
10447 Previous versions had this flag inverted, inconsistent with
10448 rsa_tmp_cb (..._TMP_RSA_CB).
10449 [Bodo Moeller; problem reported by Amit Chopra]
10451 *) Add missing DSA library text string. Work around for some IIS
10452 key files with invalid SEQUENCE encoding.
10455 *) Add a document (doc/standards.txt) that list all kinds of standards
10456 and so on that are implemented in OpenSSL.
10459 *) Enhance c_rehash script. Old version would mishandle certificates
10460 with the same subject name hash and wouldn't handle CRLs at all.
10461 Added -fingerprint option to crl utility, to support new c_rehash
10465 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
10468 *) Fix for SSL server purpose checking. Server checking was
10469 rejecting certificates which had extended key usage present
10470 but no ssl client purpose.
10471 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
10473 *) Make PKCS#12 code work with no password. The PKCS#12 spec
10474 is a little unclear about how a blank password is handled.
10475 Since the password in encoded as a BMPString with terminating
10476 double NULL a zero length password would end up as just the
10477 double NULL. However no password at all is different and is
10478 handled differently in the PKCS#12 key generation code. NS
10479 treats a blank password as zero length. MSIE treats it as no
10480 password on export: but it will try both on import. We now do
10481 the same: PKCS12_parse() tries zero length and no password if
10482 the password is set to "" or NULL (NULL is now a valid password:
10483 it wasn't before) as does the pkcs12 application.
10486 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
10487 perror when PEM_read_bio_X509_REQ fails, the error message must
10488 be obtained from the error queue.
10491 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
10492 it in ERR_remove_state if appropriate, and change ERR_get_state
10493 accordingly to avoid race conditions (this is necessary because
10494 thread_hash is no longer constant once set).
10497 *) Bugfix for linux-elf makefile.one.
10500 *) RSA_get_default_method() will now cause a default
10501 RSA_METHOD to be chosen if one doesn't exist already.
10502 Previously this was only set during a call to RSA_new()
10503 or RSA_new_method(NULL) meaning it was possible for
10504 RSA_get_default_method() to return NULL.
10507 *) Added native name translation to the existing DSO code
10508 that will convert (if the flag to do so is set) filenames
10509 that are sufficiently small and have no path information
10510 into a canonical native form. Eg. "blah" converted to
10511 "libblah.so" or "blah.dll" etc.
10514 *) New function ERR_error_string_n(e, buf, len) which is like
10515 ERR_error_string(e, buf), but writes at most 'len' bytes
10516 including the 0 terminator. For ERR_error_string_n, 'buf'
10518 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
10520 *) CONF library reworked to become more general. A new CONF
10521 configuration file reader "class" is implemented as well as a
10522 new functions (NCONF_*, for "New CONF") to handle it. The now
10523 old CONF_* functions are still there, but are reimplemented to
10524 work in terms of the new functions. Also, a set of functions
10525 to handle the internal storage of the configuration data is
10526 provided to make it easier to write new configuration file
10527 reader "classes" (I can definitely see something reading a
10528 configuration file in XML format, for example), called _CONF_*,
10529 or "the configuration storage API"...
10531 The new configuration file reading functions are:
10533 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
10534 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
10536 NCONF_default, NCONF_WIN32
10538 NCONF_dump_fp, NCONF_dump_bio
10540 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
10541 NCONF_new creates a new CONF object. This works in the same way
10542 as other interfaces in OpenSSL, like the BIO interface.
10543 NCONF_dump_* dump the internal storage of the configuration file,
10544 which is useful for debugging. All other functions take the same
10545 arguments as the old CONF_* functions wth the exception of the
10546 first that must be a `CONF *' instead of a `LHASH *'.
10548 To make it easer to use the new classes with the old CONF_* functions,
10549 the function CONF_set_default_method is provided.
10552 *) Add '-tls1' option to 'openssl ciphers', which was already
10553 mentioned in the documentation but had not been implemented.
10554 (This option is not yet really useful because even the additional
10555 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
10558 *) Initial DSO code added into libcrypto for letting OpenSSL (and
10559 OpenSSL-based applications) load shared libraries and bind to
10560 them in a portable way.
10561 [Geoff Thorpe, with contributions from Richard Levitte]
10563 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
10565 *) Make sure _lrotl and _lrotr are only used with MSVC.
10567 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
10568 (the default implementation of RAND_status).
10570 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
10571 to '-clrext' (= clear extensions), as intended and documented.
10572 [Bodo Moeller; inconsistency pointed out by Michael Attili
10573 <attili@amaxo.com>]
10575 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
10576 was larger than the MD block size.
10577 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
10579 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
10580 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
10581 using the passed key: if the passed key was a private key the result
10582 of X509_print(), for example, would be to print out all the private key
10586 *) des_quad_cksum() byte order bug fix.
10587 [Ulf Möller, using the problem description in krb4-0.9.7, where
10588 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
10590 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
10592 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
10594 *) For easily testing in shell scripts whether some command
10595 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
10596 returns with exit code 0 iff no command of the given name is available.
10597 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
10598 the output goes to stdout and nothing is printed to stderr.
10599 Additional arguments are always ignored.
10601 Since for each cipher there is a command of the same name,
10602 the 'no-cipher' compilation switches can be tested this way.
10604 ('openssl no-XXX' is not able to detect pseudo-commands such
10605 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
10608 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10611 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10612 is set; it will be thrown away anyway because each handshake creates
10614 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10615 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10616 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10617 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10620 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10621 'Q' and 'R' lose their special meanings (quit/renegotiate).
10622 This is part of what -quiet does; unlike -quiet, -ign_eof
10623 does not suppress any output.
10626 *) Add compatibility options to the purpose and trust code. The
10627 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10628 accepts a certificate or CA, this was the previous behaviour,
10629 with all the associated security issues.
10631 X509_TRUST_COMPAT is the old trust behaviour: only and
10632 automatically trust self signed roots in certificate store. A
10633 new trust setting X509_TRUST_DEFAULT is used to specify that
10634 a purpose has no associated trust setting and it should instead
10635 use the value in the default purpose.
10638 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10639 and fix a memory leak.
10642 *) In util/mkerr.pl (which implements 'make errors'), preserve
10643 reason strings from the previous version of the .c file, as
10644 the default to have only downcase letters (and digits) in
10645 automatically generated reasons codes is not always appropriate.
10648 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
10649 using strerror. Previously, ERR_reason_error_string() returned
10650 library names as reason strings for SYSerr; but SYSerr is a special
10651 case where small numbers are errno values, not library numbers.
10654 *) Add '-dsaparam' option to 'openssl dhparam' application. This
10655 converts DSA parameters into DH parameters. (When creating parameters,
10656 DSA_generate_parameters is used.)
10659 *) Include 'length' (recommended exponent length) in C code generated
10660 by 'openssl dhparam -C'.
10663 *) The second argument to set_label in perlasm was already being used
10664 so couldn't be used as a "file scope" flag. Moved to third argument
10668 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
10669 instead of RAND_bytes for encryption IVs and salts.
10672 *) Include RAND_status() into RAND_METHOD instead of implementing
10673 it only for md_rand.c Otherwise replacing the PRNG by calling
10674 RAND_set_rand_method would be impossible.
10677 *) Don't let DSA_generate_key() enter an infinite loop if the random
10678 number generation fails.
10681 *) New 'rand' application for creating pseudo-random output.
10684 *) Added configuration support for Linux/IA64
10685 [Rolf Haberrecker <rolf@suse.de>]
10687 *) Assembler module support for Mingw32.
10690 *) Shared library support for HPUX (in shlib/).
10691 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
10693 *) Shared library support for Solaris gcc.
10694 [Lutz Behnke <behnke@trustcenter.de>]
10696 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
10698 *) PKCS7_encrypt() was adding text MIME headers twice because they
10699 were added manually and by SMIME_crlf_copy().
10702 *) In bntest.c don't call BN_rand with zero bits argument.
10703 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
10705 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
10706 case was implemented. This caused BN_div_recp() to fail occasionally.
10709 *) Add an optional second argument to the set_label() in the perl
10710 assembly language builder. If this argument exists and is set
10711 to 1 it signals that the assembler should use a symbol whose
10712 scope is the entire file, not just the current function. This
10713 is needed with MASM which uses the format label:: for this scope.
10714 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
10716 *) Change the ASN1 types so they are typedefs by default. Before
10717 almost all types were #define'd to ASN1_STRING which was causing
10718 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
10722 *) Change names of new functions to the new get1/get0 naming
10723 convention: After 'get1', the caller owns a reference count
10724 and has to call ..._free; 'get0' returns a pointer to some
10725 data structure without incrementing reference counters.
10726 (Some of the existing 'get' functions increment a reference
10727 counter, some don't.)
10728 Similarly, 'set1' and 'add1' functions increase reference
10729 counters or duplicate objects.
10732 *) Allow for the possibility of temp RSA key generation failure:
10733 the code used to assume it always worked and crashed on failure.
10736 *) Fix potential buffer overrun problem in BIO_printf().
10737 [Ulf Möller, using public domain code by Patrick Powell; problem
10738 pointed out by David Sacerdote <das33@cornell.edu>]
10740 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10741 RAND_egd() and RAND_status(). In the command line application,
10742 the EGD socket can be specified like a seed file using RANDFILE
10746 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10747 Some CAs (e.g. Verisign) distribute certificates in this form.
10750 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10751 list to exclude them. This means that no special compilation option
10752 is needed to use anonymous DH: it just needs to be included in the
10756 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10757 EVP_MD_type. The old functionality is available in a new macro called
10758 EVP_MD_md(). Change code that uses it and update docs.
10761 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10762 where the 'void *' argument is replaced by a function pointer argument.
10763 Previously 'void *' was abused to point to functions, which works on
10764 many platforms, but is not correct. As these functions are usually
10765 called by macros defined in OpenSSL header files, most source code
10766 should work without changes.
10769 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10770 sections with information on -D... compiler switches used for
10771 compiling the library so that applications can see them. To enable
10772 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10773 must be defined. E.g.,
10774 #define OPENSSL_ALGORITHM_DEFINES
10775 #include <openssl/opensslconf.h>
10776 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10777 [Richard Levitte, Ulf and Bodo Möller]
10779 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10783 *) Change the 'other' type in certificate aux info to a STACK_OF
10784 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10785 the required ASN1 format: arbitrary types determined by an OID.
10788 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10789 argument to 'req'. This is not because the function is newer or
10790 better than others it just uses the work 'NEW' in the certificate
10791 request header lines. Some software needs this.
10794 *) Reorganise password command line arguments: now passwords can be
10795 obtained from various sources. Delete the PEM_cb function and make
10796 it the default behaviour: i.e. if the callback is NULL and the
10797 usrdata argument is not NULL interpret it as a null terminated pass
10798 phrase. If usrdata and the callback are NULL then the pass phrase
10799 is prompted for as usual.
10802 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10803 the support is automatically enabled. The resulting binaries will
10804 autodetect the card and use it if present.
10805 [Ben Laurie and Compaq Inc.]
10807 *) Work around for Netscape hang bug. This sends certificate request
10808 and server done in one record. Since this is perfectly legal in the
10809 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10810 the bugs/SSLv3 entry for more info.
10813 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10816 *) Add -rand argument to smime and pkcs12 applications and read/write
10820 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10823 *) Add command line password options to the remaining applications.
10826 *) Bug fix for BN_div_recp() for numerators with an even number of
10830 *) More tests in bntest.c, and changed test_bn output.
10833 *) ./config recognizes MacOS X now.
10836 *) Bug fix for BN_div() when the first words of num and divisor are
10837 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10840 *) Add support for various broken PKCS#8 formats, and command line
10841 options to produce them.
10844 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10845 get temporary BIGNUMs from a BN_CTX.
10848 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10852 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10853 include a #define from the old name to the new. The original intent
10854 was that statically linked binaries could for example just call
10855 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10856 link with digests. This never worked because SSLeay_add_all_digests()
10857 and SSLeay_add_all_ciphers() were in the same source file so calling
10858 one would link with the other. They are now in separate source files.
10861 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10864 *) Use a less unusual form of the Miller-Rabin primality test (it used
10865 a binary algorithm for exponentiation integrated into the Miller-Rabin
10866 loop, our standard modexp algorithms are faster).
10869 *) Support for the EBCDIC character set completed.
10870 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10872 *) Source code cleanups: use const where appropriate, eliminate casts,
10873 use void * instead of char * in lhash.
10876 *) Bugfix: ssl3_send_server_key_exchange was not restartable
10877 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
10878 this the server could overwrite ephemeral keys that the client
10882 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
10883 using 50 iterations of the Rabin-Miller test.
10885 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
10886 iterations of the Rabin-Miller test as required by the appendix
10887 to FIPS PUB 186[-1]) instead of DSA_is_prime.
10888 As BN_is_prime_fasttest includes trial division, DSA parameter
10889 generation becomes much faster.
10891 This implies a change for the callback functions in DSA_is_prime
10892 and DSA_generate_parameters: The callback function is called once
10893 for each positive witness in the Rabin-Miller test, not just
10894 occasionally in the inner loop; and the parameters to the
10895 callback function now provide an iteration count for the outer
10896 loop rather than for the current invocation of the inner loop.
10897 DSA_generate_parameters additionally can call the callback
10898 function with an 'iteration count' of -1, meaning that a
10899 candidate has passed the trial division test (when q is generated
10900 from an application-provided seed, trial division is skipped).
10903 *) New function BN_is_prime_fasttest that optionally does trial
10904 division before starting the Rabin-Miller test and has
10905 an additional BN_CTX * argument (whereas BN_is_prime always
10906 has to allocate at least one BN_CTX).
10907 'callback(1, -1, cb_arg)' is called when a number has passed the
10908 trial division stage.
10911 *) Fix for bug in CRL encoding. The validity dates weren't being handled
10915 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
10918 *) New function BN_pseudo_rand().
10921 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
10922 bignum version of BN_from_montgomery() with the working code from
10923 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
10927 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
10928 made it impossible to use the same SSL_SESSION data structure in
10929 SSL2 clients in multiple threads.
10932 *) The return value of RAND_load_file() no longer counts bytes obtained
10933 by stat(). RAND_load_file(..., -1) is new and uses the complete file
10934 to seed the PRNG (previously an explicit byte count was required).
10935 [Ulf Möller, Bodo Möller]
10937 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
10938 used (char *) instead of (void *) and had casts all over the place.
10941 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
10944 *) Retain source code compatibility for BN_prime_checks macro:
10945 BN_is_prime(..., BN_prime_checks, ...) now uses
10946 BN_prime_checks_for_size to determine the appropriate number of
10947 Rabin-Miller iterations.
10950 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10951 DH_CHECK_P_NOT_SAFE_PRIME.
10952 (Check if this is true? OpenPGP calls them "strong".)
10955 *) Merge the functionality of "dh" and "gendh" programs into a new program
10956 "dhparam". The old programs are retained for now but will handle DH keys
10957 (instead of parameters) in future.
10960 *) Make the ciphers, s_server and s_client programs check the return values
10961 when a new cipher list is set.
10964 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
10965 ciphers. Before when the 56bit ciphers were enabled the sorting was
10968 The syntax for the cipher sorting has been extended to support sorting by
10969 cipher-strength (using the strength_bits hard coded in the tables).
10970 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
10972 Fix a bug in the cipher-command parser: when supplying a cipher command
10973 string with an "undefined" symbol (neither command nor alphanumeric
10974 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
10975 an error is flagged.
10977 Due to the strength-sorting extension, the code of the
10978 ssl_create_cipher_list() function was completely rearranged. I hope that
10979 the readability was also increased :-)
10980 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10982 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
10983 for the first serial number and places 2 in the serial number file. This
10984 avoids problems when the root CA is created with serial number zero and
10985 the first user certificate has the same issuer name and serial number
10989 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
10990 the new code. Add documentation for this stuff.
10993 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
10994 X509_*() to X509at_*() on the grounds that they don't handle X509
10995 structures and behave in an analogous way to the X509v3 functions:
10996 they shouldn't be called directly but wrapper functions should be used
10999 So we also now have some wrapper functions that call the X509at functions
11000 when passed certificate requests. (TO DO: similar things can be done with
11001 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
11002 things. Some of these need some d2i or i2d and print functionality
11003 because they handle more complex structures.)
11006 *) Add missing #ifndefs that caused missing symbols when building libssl
11007 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
11008 NO_RSA in ssl/s2*.c.
11009 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
11011 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
11012 has a return value which indicates the quality of the random data
11013 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
11014 error queue. New function RAND_pseudo_bytes() generates output that is
11015 guaranteed to be unique but not unpredictable. RAND_add is like
11016 RAND_seed, but takes an extra argument for an entropy estimate
11017 (RAND_seed always assumes full entropy).
11020 *) Do more iterations of Rabin-Miller probable prime test (specifically,
11021 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
11022 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
11023 in crypto/bn/bn_prime.c for the complete table). This guarantees a
11024 false-positive rate of at most 2^-80 for random input.
11027 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
11030 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
11031 in the 0.9.5 release), this returns the chain
11032 from an X509_CTX structure with a dup of the stack and all
11033 the X509 reference counts upped: so the stack will exist
11034 after X509_CTX_cleanup() has been called. Modify pkcs12.c
11037 Also make SSL_SESSION_print() print out the verify return
11041 *) Add manpage for the pkcs12 command. Also change the default
11042 behaviour so MAC iteration counts are used unless the new
11043 -nomaciter option is used. This improves file security and
11044 only older versions of MSIE (4.0 for example) need it.
11047 *) Honor the no-xxx Configure options when creating .DEF files.
11050 *) Add PKCS#10 attributes to field table: challengePassword,
11051 unstructuredName and unstructuredAddress. These are taken from
11052 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
11053 international characters are used.
11055 More changes to X509_ATTRIBUTE code: allow the setting of types
11056 based on strings. Remove the 'loc' parameter when adding
11057 attributes because these will be a SET OF encoding which is sorted
11061 *) Initial changes to the 'req' utility to allow request generation
11062 automation. This will allow an application to just generate a template
11063 file containing all the field values and have req construct the
11066 Initial support for X509_ATTRIBUTE handling. Stacks of these are
11067 used all over the place including certificate requests and PKCS#7
11068 structures. They are currently handled manually where necessary with
11069 some primitive wrappers for PKCS#7. The new functions behave in a
11070 manner analogous to the X509 extension functions: they allow
11071 attributes to be looked up by NID and added.
11073 Later something similar to the X509V3 code would be desirable to
11074 automatically handle the encoding, decoding and printing of the
11075 more complex types. The string types like challengePassword can
11076 be handled by the string table functions.
11078 Also modified the multi byte string table handling. Now there is
11079 a 'global mask' which masks out certain types. The table itself
11080 can use the flag STABLE_NO_MASK to ignore the mask setting: this
11081 is useful when for example there is only one permissible type
11082 (as in countryName) and using the mask might result in no valid
11086 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
11087 SSL_get_peer_finished to allow applications to obtain the latest
11088 Finished messages sent to the peer or expected from the peer,
11089 respectively. (SSL_get_peer_finished is usually the Finished message
11090 actually received from the peer, otherwise the protocol will be aborted.)
11092 As the Finished message are message digests of the complete handshake
11093 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
11094 be used for external authentication procedures when the authentication
11095 provided by SSL/TLS is not desired or is not enough.
11098 *) Enhanced support for Alpha Linux is added. Now ./config checks if
11099 the host supports BWX extension and if Compaq C is present on the
11100 $PATH. Just exploiting of the BWX extension results in 20-30%
11101 performance kick for some algorithms, e.g. DES and RC4 to mention
11102 a couple. Compaq C in turn generates ~20% faster code for MD5 and
11106 *) Add support for MS "fast SGC". This is arguably a violation of the
11107 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
11108 weak crypto and after checking the certificate is SGC a second one
11109 with strong crypto. MS SGC stops the first handshake after receiving
11110 the server certificate message and sends a second client hello. Since
11111 a server will typically do all the time consuming operations before
11112 expecting any further messages from the client (server key exchange
11113 is the most expensive) there is little difference between the two.
11115 To get OpenSSL to support MS SGC we have to permit a second client
11116 hello message after we have sent server done. In addition we have to
11117 reset the MAC if we do get this second client hello.
11120 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
11121 if a DER encoded private key is RSA or DSA traditional format. Changed
11122 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
11123 format DER encoded private key. Newer code should use PKCS#8 format which
11124 has the key type encoded in the ASN1 structure. Added DER private key
11125 support to pkcs8 application.
11128 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
11129 ciphersuites has been selected (as required by the SSL 3/TLS 1
11130 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
11131 is set, we interpret this as a request to violate the specification
11132 (the worst that can happen is a handshake failure, and 'correct'
11133 behaviour would result in a handshake failure anyway).
11136 *) In SSL_CTX_add_session, take into account that there might be multiple
11137 SSL_SESSION structures with the same session ID (e.g. when two threads
11138 concurrently obtain them from an external cache).
11139 The internal cache can handle only one SSL_SESSION with a given ID,
11140 so if there's a conflict, we now throw out the old one to achieve
11144 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
11145 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
11146 some routines that use cipher OIDs: some ciphers do not have OIDs
11147 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
11151 *) Simplify the trust setting structure and code. Now we just have
11152 two sequences of OIDs for trusted and rejected settings. These will
11153 typically have values the same as the extended key usage extension
11154 and any application specific purposes.
11156 The trust checking code now has a default behaviour: it will just
11157 check for an object with the same NID as the passed id. Functions can
11158 be provided to override either the default behaviour or the behaviour
11159 for a given id. SSL client, server and email already have functions
11160 in place for compatibility: they check the NID and also return "trusted"
11161 if the certificate is self signed.
11164 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
11165 traditional format into an EVP_PKEY structure.
11168 *) Add a password callback function PEM_cb() which either prompts for
11169 a password if usr_data is NULL or otherwise assumes it is a null
11170 terminated password. Allow passwords to be passed on command line
11171 environment or config files in a few more utilities.
11174 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
11175 keys. Add some short names for PKCS#8 PBE algorithms and allow them
11176 to be specified on the command line for the pkcs8 and pkcs12 utilities.
11177 Update documentation.
11180 *) Support for ASN1 "NULL" type. This could be handled before by using
11181 ASN1_TYPE but there wasn't any function that would try to read a NULL
11182 and produce an error if it couldn't. For compatibility we also have
11183 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
11184 don't allocate anything because they don't need to.
11187 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
11189 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
11191 *) Rebuild of the memory allocation routines used by OpenSSL code and
11192 possibly others as well. The purpose is to make an interface that
11193 provide hooks so anyone can build a separate set of allocation and
11194 deallocation routines to be used by OpenSSL, for example memory
11195 pool implementations, or something else, which was previously hard
11196 since Malloc(), Realloc() and Free() were defined as macros having
11197 the values malloc, realloc and free, respectively (except for Win32
11198 compilations). The same is provided for memory debugging code.
11199 OpenSSL already comes with functionality to find memory leaks, but
11200 this gives people a chance to debug other memory problems.
11202 With these changes, a new set of functions and macros have appeared:
11204 CRYPTO_set_mem_debug_functions() [F]
11205 CRYPTO_get_mem_debug_functions() [F]
11206 CRYPTO_dbg_set_options() [F]
11207 CRYPTO_dbg_get_options() [F]
11208 CRYPTO_malloc_debug_init() [M]
11210 The memory debug functions are NULL by default, unless the library
11211 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
11212 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
11213 gives the standard debugging functions that come with OpenSSL) or
11214 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
11215 provided by the library user) must be used. When the standard
11216 debugging functions are used, CRYPTO_dbg_set_options can be used to
11217 request additional information:
11218 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
11219 the CRYPTO_MDEBUG_xxx macro when compiling the library.
11221 Also, things like CRYPTO_set_mem_functions will always give the
11222 expected result (the new set of functions is used for allocation
11223 and deallocation) at all times, regardless of platform and compiler
11226 To finish it up, some functions that were never use in any other
11227 way than through macros have a new API and new semantic:
11229 CRYPTO_dbg_malloc()
11230 CRYPTO_dbg_realloc()
11233 All macros of value have retained their old syntax.
11234 [Richard Levitte and Bodo Moeller]
11236 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
11237 ordering of SMIMECapabilities wasn't in "strength order" and there
11238 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
11242 *) Some ASN1 types with illegal zero length encoding (INTEGER,
11243 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
11244 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
11246 *) Merge in my S/MIME library for OpenSSL. This provides a simple
11247 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
11248 functionality to handle multipart/signed properly) and a utility
11249 called 'smime' to call all this stuff. This is based on code I
11250 originally wrote for Celo who have kindly allowed it to be
11251 included in OpenSSL.
11254 *) Add variants des_set_key_checked and des_set_key_unchecked of
11255 des_set_key (aka des_key_sched). Global variable des_check_key
11256 decides which of these is called by des_set_key; this way
11257 des_check_key behaves as it always did, but applications and
11258 the library itself, which was buggy for des_check_key == 1,
11259 have a cleaner way to pick the version they need.
11262 *) New function PKCS12_newpass() which changes the password of a
11266 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
11267 dynamic mix. In both cases the ids can be used as an index into the
11268 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
11269 functions so they accept a list of the field values and the
11270 application doesn't need to directly manipulate the X509_TRUST
11274 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
11278 *) Modify the way the V3 extension code looks up extensions. This now
11279 works in a similar way to the object code: we have some "standard"
11280 extensions in a static table which is searched with OBJ_bsearch()
11281 and the application can add dynamic ones if needed. The file
11282 crypto/x509v3/ext_dat.h now has the info: this file needs to be
11283 updated whenever a new extension is added to the core code and kept
11284 in ext_nid order. There is a simple program 'tabtest.c' which checks
11285 this. New extensions are not added too often so this file can readily
11286 be maintained manually.
11288 There are two big advantages in doing things this way. The extensions
11289 can be looked up immediately and no longer need to be "added" using
11290 X509V3_add_standard_extensions(): this function now does nothing.
11291 [Side note: I get *lots* of email saying the extension code doesn't
11292 work because people forget to call this function]
11293 Also no dynamic allocation is done unless new extensions are added:
11294 so if we don't add custom extensions there is no need to call
11295 X509V3_EXT_cleanup().
11298 *) Modify enc utility's salting as follows: make salting the default. Add a
11299 magic header, so unsalted files fail gracefully instead of just decrypting
11300 to garbage. This is because not salting is a big security hole, so people
11301 should be discouraged from doing it.
11304 *) Fixes and enhancements to the 'x509' utility. It allowed a message
11305 digest to be passed on the command line but it only used this
11306 parameter when signing a certificate. Modified so all relevant
11307 operations are affected by the digest parameter including the
11308 -fingerprint and -x509toreq options. Also -x509toreq choked if a
11309 DSA key was used because it didn't fix the digest.
11312 *) Initial certificate chain verify code. Currently tests the untrusted
11313 certificates for consistency with the verify purpose (which is set
11314 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11316 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11317 this is because it will reject chains with invalid extensions whereas
11318 every previous version of OpenSSL and SSLeay made no checks at all.
11320 Trust code: checks the root CA for the relevant trust settings. Trust
11321 settings have an initial value consistent with the verify purpose: e.g.
11322 if the verify purpose is for SSL client use it expects the CA to be
11323 trusted for SSL client use. However the default value can be changed to
11324 permit custom trust settings: one example of this would be to only trust
11325 certificates from a specific "secure" set of CAs.
11327 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11328 which should be used for version portability: especially since the
11329 verify structure is likely to change more often now.
11331 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11332 to set them. If not set then assume SSL clients will verify SSL servers
11335 Two new options to the verify program: -untrusted allows a set of
11336 untrusted certificates to be passed in and -purpose which sets the
11337 intended purpose of the certificate. If a purpose is set then the
11338 new chain verify code is used to check extension consistency.
11341 *) Support for the authority information access extension.
11344 *) Modify RSA and DSA PEM read routines to transparently handle
11345 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11346 public keys in a format compatible with certificate
11347 SubjectPublicKeyInfo structures. Unfortunately there were already
11348 functions called *_PublicKey_* which used various odd formats so
11349 these are retained for compatibility: however the DSA variants were
11350 never in a public release so they have been deleted. Changed dsa/rsa
11351 utilities to handle the new format: note no releases ever handled public
11352 keys so we should be OK.
11354 The primary motivation for this change is to avoid the same fiasco
11355 that dogs private keys: there are several incompatible private key
11356 formats some of which are standard and some OpenSSL specific and
11357 require various evil hacks to allow partial transparent handling and
11358 even then it doesn't work with DER formats. Given the option anything
11359 other than PKCS#8 should be dumped: but the other formats have to
11360 stay in the name of compatibility.
11362 With public keys and the benefit of hindsight one standard format
11363 is used which works with EVP_PKEY, RSA or DSA structures: though
11364 it clearly returns an error if you try to read the wrong kind of key.
11366 Added a -pubkey option to the 'x509' utility to output the public key.
11367 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11368 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11369 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11370 that do the same as the EVP_PKEY_assign_*() except they up the
11371 reference count of the added key (they don't "swallow" the
11375 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11376 CRLs would fail if the file contained no certificates or no CRLs:
11377 added a new function to read in both types and return the number
11378 read: this means that if none are read it will be an error. The
11379 DER versions of the certificate and CRL reader would always fail
11380 because it isn't possible to mix certificates and CRLs in DER format
11381 without choking one or the other routine. Changed this to just read
11382 a certificate: this is the best we can do. Also modified the code
11383 in apps/verify.c to take notice of return codes: it was previously
11384 attempting to read in certificates from NULL pointers and ignoring
11385 any errors: this is one reason why the cert and CRL reader seemed
11386 to work. It doesn't check return codes from the default certificate
11387 routines: these may well fail if the certificates aren't installed.
11390 *) Code to support otherName option in GeneralName.
11393 *) First update to verify code. Change the verify utility
11394 so it warns if it is passed a self signed certificate:
11395 for consistency with the normal behaviour. X509_verify
11396 has been modified to it will now verify a self signed
11397 certificate if *exactly* the same certificate appears
11398 in the store: it was previously impossible to trust a
11399 single self signed certificate. This means that:
11400 openssl verify ss.pem
11401 now gives a warning about a self signed certificate but
11402 openssl verify -CAfile ss.pem ss.pem
11406 *) For servers, store verify_result in SSL_SESSION data structure
11407 (and add it to external session representation).
11408 This is needed when client certificate verifications fails,
11409 but an application-provided verification callback (set by
11410 SSL_CTX_set_cert_verify_callback) allows accepting the session
11411 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11412 but returns 1): When the session is reused, we have to set
11413 ssl->verify_result to the appropriate error code to avoid
11415 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11417 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11418 case in PKCS7_dataInit() where the signed PKCS7 structure
11419 didn't contain any existing data because it was being created.
11420 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11422 *) Add a salt to the key derivation routines in enc.c. This
11423 forms the first 8 bytes of the encrypted file. Also add a
11424 -S option to allow a salt to be input on the command line.
11427 *) New function X509_cmp(). Oddly enough there wasn't a function
11428 to compare two certificates. We do this by working out the SHA1
11429 hash and comparing that. X509_cmp() will be needed by the trust
11433 *) SSL_get1_session() is like SSL_get_session(), but increments
11434 the reference count in the SSL_SESSION returned.
11435 [Geoff Thorpe <geoff@eu.c2.net>]
11437 *) Fix for 'req': it was adding a null to request attributes.
11438 Also change the X509_LOOKUP and X509_INFO code to handle
11439 certificate auxiliary information.
11442 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11446 *) Add the possibility to add extra information to the memory leak
11447 detecting output, to form tracebacks, showing from where each
11448 allocation was originated: CRYPTO_push_info("constant string") adds
11449 the string plus current file name and line number to a per-thread
11450 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11451 is like calling CYRPTO_pop_info() until the stack is empty.
11452 Also updated memory leak detection code to be multi-thread-safe.
11455 *) Add options -text and -noout to pkcs7 utility and delete the
11456 encryption options which never did anything. Update docs.
11459 *) Add options to some of the utilities to allow the pass phrase
11460 to be included on either the command line (not recommended on
11461 OSes like Unix) or read from the environment. Update the
11462 manpages and fix a few bugs.
11465 *) Add a few manpages for some of the openssl commands.
11468 *) Fix the -revoke option in ca. It was freeing up memory twice,
11469 leaking and not finding already revoked certificates.
11472 *) Extensive changes to support certificate auxiliary information.
11473 This involves the use of X509_CERT_AUX structure and X509_AUX
11474 functions. An X509_AUX function such as PEM_read_X509_AUX()
11475 can still read in a certificate file in the usual way but it
11476 will also read in any additional "auxiliary information". By
11477 doing things this way a fair degree of compatibility can be
11478 retained: existing certificates can have this information added
11479 using the new 'x509' options.
11481 Current auxiliary information includes an "alias" and some trust
11482 settings. The trust settings will ultimately be used in enhanced
11483 certificate chain verification routines: currently a certificate
11484 can only be trusted if it is self signed and then it is trusted
11488 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
11489 The problem was that one of the replacement routines had not been working
11490 since SSLeay releases. For now the offending routine has been replaced
11491 with non-optimised assembler. Even so, this now gives around 95%
11492 performance improvement for 1024 bit RSA signs.
11495 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
11496 handling. Most clients have the effective key size in bits equal to
11497 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
11498 A few however don't do this and instead use the size of the decrypted key
11499 to determine the RC2 key length and the AlgorithmIdentifier to determine
11500 the effective key length. In this case the effective key length can still
11501 be 40 bits but the key length can be 168 bits for example. This is fixed
11502 by manually forcing an RC2 key into the EVP_PKEY structure because the
11503 EVP code can't currently handle unusual RC2 key sizes: it always assumes
11504 the key length and effective key length are equal.
11507 *) Add a bunch of functions that should simplify the creation of
11508 X509_NAME structures. Now you should be able to do:
11509 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
11510 and have it automatically work out the correct field type and fill in
11511 the structures. The more adventurous can try:
11512 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
11513 and it will (hopefully) work out the correct multibyte encoding.
11516 *) Change the 'req' utility to use the new field handling and multibyte
11517 copy routines. Before the DN field creation was handled in an ad hoc
11518 way in req, ca, and x509 which was rather broken and didn't support
11519 BMPStrings or UTF8Strings. Since some software doesn't implement
11520 BMPStrings or UTF8Strings yet, they can be enabled using the config file
11521 using the dirstring_type option. See the new comment in the default
11522 openssl.cnf for more info.
11525 *) Make crypto/rand/md_rand.c more robust:
11526 - Assure unique random numbers after fork().
11527 - Make sure that concurrent threads access the global counter and
11528 md serializably so that we never lose entropy in them
11529 or use exactly the same state in multiple threads.
11530 Access to the large state is not always serializable because
11531 the additional locking could be a performance killer, and
11532 md should be large enough anyway.
11535 *) New file apps/app_rand.c with commonly needed functionality
11536 for handling the random seed file.
11538 Use the random seed file in some applications that previously did not:
11540 dsaparam -genkey (which also ignored its '-rand' option),
11543 x509 (when signing).
11544 Except on systems with /dev/urandom, it is crucial to have a random
11545 seed file at least for key creation, DSA signing, and for DH exchanges;
11546 for RSA signatures we could do without one.
11548 gendh and gendsa (unlike genrsa) used to read only the first byte
11549 of each file listed in the '-rand' option. The function as previously
11550 found in genrsa is now in app_rand.c and is used by all programs
11551 that support '-rand'.
11554 *) In RAND_write_file, use mode 0600 for creating files;
11555 don't just chmod when it may be too late.
11558 *) Report an error from X509_STORE_load_locations
11559 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
11562 *) New function ASN1_mbstring_copy() this copies a string in either
11563 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
11564 into an ASN1_STRING type. A mask of permissible types is passed
11565 and it chooses the "minimal" type to use or an error if not type
11569 *) Add function equivalents to the various macros in asn1.h. The old
11570 macros are retained with an M_ prefix. Code inside the library can
11571 use the M_ macros. External code (including the openssl utility)
11572 should *NOT* in order to be "shared library friendly".
11575 *) Add various functions that can check a certificate's extensions
11576 to see if it usable for various purposes such as SSL client,
11577 server or S/MIME and CAs of these types. This is currently
11578 VERY EXPERIMENTAL but will ultimately be used for certificate chain
11579 verification. Also added a -purpose flag to x509 utility to
11580 print out all the purposes.
11583 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
11587 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
11588 for, obtain and decode and extension and obtain its critical flag.
11589 This allows all the necessary extension code to be handled in a
11590 single function call.
11593 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
11594 platforms. See crypto/rc4/rc4_enc.c for further details.
11597 *) New -noout option to asn1parse. This causes no output to be produced
11598 its main use is when combined with -strparse and -out to extract data
11599 from a file (which may not be in ASN.1 format).
11602 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
11603 when producing the local key id.
11604 [Richard Levitte <levitte@stacken.kth.se>]
11606 *) New option -dhparam in s_server. This allows a DH parameter file to be
11607 stated explicitly. If it is not stated then it tries the first server
11608 certificate file. The previous behaviour hard coded the filename
11612 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11613 a public key to be input or output. For example:
11614 openssl rsa -in key.pem -pubout -out pubkey.pem
11615 Also added necessary DSA public key functions to handle this.
11618 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11619 in the message. This was handled by allowing
11620 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11621 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11623 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11624 to the end of the strings whereas this didn't. This would cause problems
11625 if strings read with d2i_ASN1_bytes() were later modified.
11626 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11628 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11629 data and it contains EOF it will end up returning an error. This is
11630 caused by input 46 bytes long. The cause is due to the way base64
11631 BIOs find the start of base64 encoded data. They do this by trying a
11632 trial decode on each line until they find one that works. When they
11633 do a flag is set and it starts again knowing it can pass all the
11634 data directly through the decoder. Unfortunately it doesn't reset
11635 the context it uses. This means that if EOF is reached an attempt
11636 is made to pass two EOFs through the context and this causes the
11637 resulting error. This can also cause other problems as well. As is
11638 usual with these problems it takes *ages* to find and the fix is
11639 trivial: move one line.
11640 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
11642 *) Ugly workaround to get s_client and s_server working under Windows. The
11643 old code wouldn't work because it needed to select() on sockets and the
11644 tty (for keypresses and to see if data could be written). Win32 only
11645 supports select() on sockets so we select() with a 1s timeout on the
11646 sockets and then see if any characters are waiting to be read, if none
11647 are present then we retry, we also assume we can always write data to
11648 the tty. This isn't nice because the code then blocks until we've
11649 received a complete line of data and it is effectively polling the
11650 keyboard at 1s intervals: however it's quite a bit better than not
11651 working at all :-) A dedicated Windows application might handle this
11652 with an event loop for example.
11655 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
11656 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
11657 will be called when RSA_sign() and RSA_verify() are used. This is useful
11658 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
11659 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
11660 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
11661 This necessitated the support of an extra signature type NID_md5_sha1
11662 for SSL signatures and modifications to the SSL library to use it instead
11663 of calling RSA_public_decrypt() and RSA_private_encrypt().
11666 *) Add new -verify -CAfile and -CApath options to the crl program, these
11667 will lookup a CRL issuers certificate and verify the signature in a
11668 similar way to the verify program. Tidy up the crl program so it
11669 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
11670 less strict. It will now permit CRL extensions even if it is not
11671 a V2 CRL: this will allow it to tolerate some broken CRLs.
11674 *) Initialize all non-automatic variables each time one of the openssl
11675 sub-programs is started (this is necessary as they may be started
11676 multiple times from the "OpenSSL>" prompt).
11677 [Lennart Bang, Bodo Moeller]
11679 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
11680 removing all other RSA functionality (this is what NO_RSA does). This
11681 is so (for example) those in the US can disable those operations covered
11682 by the RSA patent while allowing storage and parsing of RSA keys and RSA
11686 *) Non-copying interface to BIO pairs.
11687 (still largely untested)
11690 *) New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
11691 ASCII string. This was handled independently in various places before.
11694 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
11695 UTF8 strings a character at a time.
11698 *) Use client_version from client hello to select the protocol
11699 (s23_srvr.c) and for RSA client key exchange verification
11700 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
11703 *) Add various utility functions to handle SPKACs, these were previously
11704 handled by poking round in the structure internals. Added new function
11705 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
11706 print, verify and generate SPKACs. Based on an original idea from
11707 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
11710 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
11713 *) Allow the config file extension section to be overwritten on the
11714 command line. Based on an original idea from Massimiliano Pala
11715 <madwolf@comune.modena.it>. The new option is called -extensions
11716 and can be applied to ca, req and x509. Also -reqexts to override
11717 the request extensions in req and -crlexts to override the crl extensions
11721 *) Add new feature to the SPKAC handling in ca. Now you can include
11722 the same field multiple times by preceding it by "XXXX." for example:
11725 this is the same syntax as used in the req config file.
11728 *) Allow certificate extensions to be added to certificate requests. These
11729 are specified in a 'req_extensions' option of the req section of the
11730 config file. They can be printed out with the -text option to req but
11731 are otherwise ignored at present.
11734 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11735 data read consists of only the final block it would not decrypted because
11736 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11737 A misplaced 'break' also meant the decrypted final block might not be
11738 copied until the next read.
11741 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11742 a few extra parameters to the DH structure: these will be useful if
11743 for example we want the value of 'q' or implement X9.42 DH.
11746 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11747 provides hooks that allow the default DSA functions or functions on a
11748 "per key" basis to be replaced. This allows hardware acceleration and
11749 hardware key storage to be handled without major modification to the
11750 library. Also added low level modexp hooks and CRYPTO_EX structure and
11751 associated functions.
11754 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11755 as "read only": it can't be written to and the buffer it points to will
11756 not be freed. Reading from a read only BIO is much more efficient than
11757 a normal memory BIO. This was added because there are several times when
11758 an area of memory needs to be read from a BIO. The previous method was
11759 to create a memory BIO and write the data to it, this results in two
11760 copies of the data and an O(n^2) reading algorithm. There is a new
11761 function BIO_new_mem_buf() which creates a read only memory BIO from
11762 an area of memory. Also modified the PKCS#7 routines to use read only
11766 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11767 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11768 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11769 but a retry condition occurred while trying to read the rest.
11772 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11773 NID_pkcs7_encrypted by default: this was wrong since this should almost
11774 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11775 the encrypted data type: this is a more sensible place to put it and it
11776 allows the PKCS#12 code to be tidied up that duplicated this
11780 *) Changed obj_dat.pl script so it takes its input and output files on
11781 the command line. This should avoid shell escape redirection problems
11785 *) Initial support for certificate extension requests, these are included
11786 in things like Xenroll certificate requests. Included functions to allow
11787 extensions to be obtained and added.
11790 *) -crlf option to s_client and s_server for sending newlines as
11791 CRLF (as required by many protocols).
11794 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11796 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11797 [Ralf S. Engelschall]
11799 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11800 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11802 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11806 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11807 DH parameters/keys (q is lost during that conversion, but the resulting
11808 DH parameters contain its length).
11810 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11811 much faster than DH_generate_parameters (which creates parameters
11812 where p = 2*q + 1), and also the smaller q makes DH computations
11813 much more efficient (160-bit exponentiation instead of 1024-bit
11814 exponentiation); so this provides a convenient way to support DHE
11815 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11816 utter importance to use
11817 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11819 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11820 when such DH parameters are used, because otherwise small subgroup
11821 attacks may become possible!
11824 *) Avoid memory leak in i2d_DHparams.
11827 *) Allow the -k option to be used more than once in the enc program:
11828 this allows the same encrypted message to be read by multiple recipients.
11831 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11832 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11833 it will always use the numerical form of the OID, even if it has a short
11837 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11838 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11839 otherwise bn_mod_exp was called. In the case of hardware keys for example
11840 no private key components need be present and it might store extra data
11841 in the RSA structure, which cannot be accessed from bn_mod_exp.
11842 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11843 private key operations.
11846 *) Added support for SPARC Linux.
11849 *) pem_password_cb function type incompatibly changed from
11850 typedef int pem_password_cb(char *buf, int size, int rwflag);
11852 ....(char *buf, int size, int rwflag, void *userdata);
11853 so that applications can pass data to their callbacks:
11854 The PEM[_ASN1]_{read,write}... functions and macros now take an
11855 additional void * argument, which is just handed through whenever
11856 the password callback is called.
11857 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11859 New function SSL_CTX_set_default_passwd_cb_userdata.
11861 Compatibility note: As many C implementations push function arguments
11862 onto the stack in reverse order, the new library version is likely to
11863 interoperate with programs that have been compiled with the old
11864 pem_password_cb definition (PEM_whatever takes some data that
11865 happens to be on the stack as its last argument, and the callback
11866 just ignores this garbage); but there is no guarantee whatsoever that
11869 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11870 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11871 problems not only on Windows, but also on some Unix platforms.
11872 To avoid problematic command lines, these definitions are now in an
11873 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11874 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
11877 *) MIPS III/IV assembler module is reimplemented.
11880 *) More DES library cleanups: remove references to srand/rand and
11881 delete an unused file.
11884 *) Add support for the free Netwide assembler (NASM) under Win32,
11885 since not many people have MASM (ml) and it can be hard to obtain.
11886 This is currently experimental but it seems to work OK and pass all
11887 the tests. Check out INSTALL.W32 for info.
11890 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
11891 without temporary keys kept an extra copy of the server key,
11892 and connections with temporary keys did not free everything in case
11896 *) New function RSA_check_key and new openssl rsa option -check
11897 for verifying the consistency of RSA keys.
11898 [Ulf Moeller, Bodo Moeller]
11900 *) Various changes to make Win32 compile work:
11901 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
11902 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
11903 comparison" warnings.
11904 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
11907 *) Add a debugging option to PKCS#5 v2 key generation function: when
11908 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
11909 derived keys are printed to stderr.
11912 *) Copy the flags in ASN1_STRING_dup().
11913 [Roman E. Pavlov <pre@mo.msk.ru>]
11915 *) The x509 application mishandled signing requests containing DSA
11916 keys when the signing key was also DSA and the parameters didn't match.
11918 It was supposed to omit the parameters when they matched the signing key:
11919 the verifying software was then supposed to automatically use the CA's
11920 parameters if they were absent from the end user certificate.
11922 Omitting parameters is no longer recommended. The test was also
11923 the wrong way round! This was probably due to unusual behaviour in
11924 EVP_cmp_parameters() which returns 1 if the parameters match.
11925 This meant that parameters were omitted when they *didn't* match and
11926 the certificate was useless. Certificates signed with 'ca' didn't have
11928 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
11930 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
11931 The interface is as follows:
11932 Applications can use
11933 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
11934 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
11935 "off" is now the default.
11936 The library internally uses
11937 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
11938 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
11939 to disable memory-checking temporarily.
11941 Some inconsistent states that previously were possible (and were
11942 even the default) are now avoided.
11944 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
11945 with each memory chunk allocated; this is occasionally more helpful
11946 than just having a counter.
11948 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11950 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11954 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11955 which largely parallels "options", but is for changing API behaviour,
11956 whereas "options" are about protocol behaviour.
11957 Initial "mode" flags are:
11959 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
11960 a single record has been written.
11961 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
11962 retries use the same buffer location.
11963 (But all of the contents must be
11967 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
11970 *) Fix problems with no-hmac etc.
11971 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
11973 *) New functions RSA_get_default_method(), RSA_set_method() and
11974 RSA_get_method(). These allows replacement of RSA_METHODs without having
11975 to mess around with the internals of an RSA structure.
11978 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
11979 Also really enable memory leak checks in openssl.c and in some
11981 [Chad C. Mulligan, Bodo Moeller]
11983 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
11984 up the length of negative integers. This has now been simplified to just
11985 store the length when it is first determined and use it later, rather
11986 than trying to keep track of where data is copied and updating it to
11988 [Steve Henson, reported by Brien Wheeler
11989 <bwheeler@authentica-security.com>]
11991 *) Add a new function PKCS7_signatureVerify. This allows the verification
11992 of a PKCS#7 signature but with the signing certificate passed to the
11993 function itself. This contrasts with PKCS7_dataVerify which assumes the
11994 certificate is present in the PKCS#7 structure. This isn't always the
11995 case: certificates can be omitted from a PKCS#7 structure and be
11996 distributed by "out of band" means (such as a certificate database).
11999 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
12000 function prototypes in pem.h, also change util/mkdef.pl to add the
12001 necessary function names.
12004 *) mk1mf.pl (used by Windows builds) did not properly read the
12005 options set by Configure in the top level Makefile, and Configure
12006 was not even able to write more than one option correctly.
12007 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
12010 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
12011 file to be loaded from a BIO or FILE pointer. The BIO version will
12012 for example allow memory BIOs to contain config info.
12015 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
12016 Whoever hopes to achieve shared-library compatibility across versions
12017 must use this, not the compile-time macro.
12018 (Exercise 0.9.4: Which is the minimum library version required by
12020 Note: All this applies only to multi-threaded programs, others don't
12024 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
12025 through a BIO pair triggered the default case, i.e.
12026 SSLerr(...,SSL_R_UNKNOWN_STATE).
12029 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
12030 can use the SSL library even if none of the specific BIOs is
12034 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
12035 for the encoded length.
12036 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
12038 *) Add initial documentation of the X509V3 functions.
12041 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
12042 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
12043 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
12044 secure PKCS#8 private key format with a high iteration count.
12047 *) Fix determination of Perl interpreter: A perl or perl5
12048 _directory_ in $PATH was also accepted as the interpreter.
12049 [Ralf S. Engelschall]
12051 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
12052 wrong with it but it was very old and did things like calling
12053 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
12054 unusual formatting.
12057 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
12058 to use the new extension code.
12061 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
12062 with macros. This should make it easier to change their form, add extra
12063 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
12067 *) Add to configuration table a new entry that can specify an alternative
12068 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
12069 according to Mark Crispin <MRC@Panda.COM>.
12073 *) DES CBC did not update the IV. Weird.
12076 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
12077 Changing the behaviour of the former might break existing programs --
12078 where IV updating is needed, des_ncbc_encrypt can be used.
12081 *) When bntest is run from "make test" it drives bc to check its
12082 calculations, as well as internally checking them. If an internal check
12083 fails, it needs to cause bc to give a non-zero result or make test carries
12084 on without noticing the failure. Fixed.
12087 *) DES library cleanups.
12090 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
12091 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
12092 ciphers. NOTE: although the key derivation function has been verified
12093 against some published test vectors it has not been extensively tested
12094 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
12098 *) Instead of "mkdir -p", which is not fully portable, use new
12099 Perl script "util/mkdir-p.pl".
12102 *) Rewrite the way password based encryption (PBE) is handled. It used to
12103 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
12104 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
12105 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
12106 the 'parameter' field of the AlgorithmIdentifier is passed to the
12107 underlying key generation function so it must do its own ASN1 parsing.
12108 This has also changed the EVP_PBE_CipherInit() function which now has a
12109 'parameter' argument instead of literal salt and iteration count values
12110 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
12113 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
12114 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
12115 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
12116 KEY" because this clashed with PKCS#8 unencrypted string. Since this
12117 value was just used as a "magic string" and not used directly its
12118 value doesn't matter.
12121 *) Introduce some semblance of const correctness to BN. Shame C doesn't
12125 *) "linux-sparc64" configuration (ultrapenguin).
12126 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
12127 "linux-sparc" configuration.
12128 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
12130 *) config now generates no-xxx options for missing ciphers.
12133 *) Support the EBCDIC character set (work in progress).
12134 File ebcdic.c not yet included because it has a different license.
12135 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12137 *) Support BS2000/OSD-POSIX.
12138 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12140 *) Make callbacks for key generation use void * instead of char *.
12143 *) Make S/MIME samples compile (not yet tested).
12146 *) Additional typesafe stacks.
12149 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
12153 Changes between 0.9.3 and 0.9.3a [29 May 1999]
12155 *) New configuration variant "sco5-gcc".
12157 *) Updated some demos.
12158 [Sean O Riordain, Wade Scholine]
12160 *) Add missing BIO_free at exit of pkcs12 application.
12163 *) Fix memory leak in conf.c.
12166 *) Updates for Win32 to assembler version of MD5.
12169 *) Set #! path to perl in apps/der_chop to where we found it
12170 instead of using a fixed path.
12173 *) SHA library changes for irix64-mips4-cc.
12176 *) Improvements for VMS support.
12180 Changes between 0.9.2b and 0.9.3 [24 May 1999]
12182 *) Bignum library bug fix. IRIX 6 passes "make test" now!
12183 This also avoids the problems with SC4.2 and unpatched SC5.
12184 [Andy Polyakov <appro@fy.chalmers.se>]
12186 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
12187 These are required because of the typesafe stack would otherwise break
12188 existing code. If old code used a structure member which used to be STACK
12189 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
12190 sk_num or sk_value it would produce an error because the num, data members
12191 are not present in STACK_OF. Now it just produces a warning. sk_set
12192 replaces the old method of assigning a value to sk_value
12193 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
12194 that does this will no longer work (and should use sk_set instead) but
12195 this could be regarded as a "questionable" behaviour anyway.
12198 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
12199 correctly handle encrypted S/MIME data.
12202 *) Change type of various DES function arguments from des_cblock
12203 (which means, in function argument declarations, pointer to char)
12204 to des_cblock * (meaning pointer to array with 8 char elements),
12205 which allows the compiler to do more typechecking; it was like
12206 that back in SSLeay, but with lots of ugly casts.
12208 Introduce new type const_des_cblock.
12211 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
12212 problems: find RecipientInfo structure that matches recipient certificate
12213 and initialise the ASN1 structures properly based on passed cipher.
12216 *) Belatedly make the BN tests actually check the results.
12219 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
12220 to and from BNs: it was completely broken. New compilation option
12221 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
12222 key elements as negative integers.
12225 *) Reorganize and speed up MD5.
12226 [Andy Polyakov <appro@fy.chalmers.se>]
12229 [Richard Levitte <richard@levitte.org>]
12231 *) New option -out to asn1parse to allow the parsed structure to be
12232 output to a file. This is most useful when combined with the -strparse
12233 option to examine the output of things like OCTET STRINGS.
12236 *) Make SSL library a little more fool-proof by not requiring any longer
12237 that SSL_set_{accept,connect}_state be called before
12238 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
12239 in many applications because usually everything *appeared* to work as
12240 intended anyway -- now it really works as intended).
12243 *) Move openssl.cnf out of lib/.
12246 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
12247 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
12248 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
12249 [Ralf S. Engelschall]
12251 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
12252 handle PKCS#7 enveloped data properly.
12253 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
12255 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
12256 copying pointers. The cert_st handling is changed by this in
12257 various ways (and thus what used to be known as ctx->default_cert
12258 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
12259 any longer when s->cert does not give us what we need).
12260 ssl_cert_instantiate becomes obsolete by this change.
12261 As soon as we've got the new code right (possibly it already is?),
12262 we have solved a couple of bugs of the earlier code where s->cert
12263 was used as if it could not have been shared with other SSL structures.
12265 Note that using the SSL API in certain dirty ways now will result
12266 in different behaviour than observed with earlier library versions:
12267 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
12268 does not influence s as it used to.
12270 In order to clean up things more thoroughly, inside SSL_SESSION
12271 we don't use CERT any longer, but a new structure SESS_CERT
12272 that holds per-session data (if available); currently, this is
12273 the peer's certificate chain and, for clients, the server's certificate
12274 and temporary key. CERT holds only those values that can have
12275 meaningful defaults in an SSL_CTX.
12278 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
12279 from the internal representation. Various PKCS#7 fixes: remove some
12280 evil casts and set the enc_dig_alg field properly based on the signing
12284 *) Allow PKCS#12 password to be set from the command line or the
12285 environment. Let 'ca' get its config file name from the environment
12286 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
12290 *) Allow certificate policies extension to use an IA5STRING for the
12291 organization field. This is contrary to the PKIX definition but
12292 VeriSign uses it and IE5 only recognises this form. Document 'x509'
12296 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
12297 without disallowing inline assembler and the like for non-pedantic builds.
12300 *) Support Borland C++ builder.
12301 [Janez Jere <jj@void.si>, modified by Ulf Möller]
12303 *) Support Mingw32.
12306 *) SHA-1 cleanups and performance enhancements.
12307 [Andy Polyakov <appro@fy.chalmers.se>]
12309 *) Sparc v8plus assembler for the bignum library.
12310 [Andy Polyakov <appro@fy.chalmers.se>]
12312 *) Accept any -xxx and +xxx compiler options in Configure.
12315 *) Update HPUX configuration.
12318 *) Add missing sk_<type>_unshift() function to safestack.h
12319 [Ralf S. Engelschall]
12321 *) New function SSL_CTX_use_certificate_chain_file that sets the
12322 "extra_cert"s in addition to the certificate. (This makes sense
12323 only for "PEM" format files, as chains as a whole are not
12327 *) Support verify_depth from the SSL API.
12328 x509_vfy.c had what can be considered an off-by-one-error:
12329 Its depth (which was not part of the external interface)
12330 was actually counting the number of certificates in a chain;
12331 now it really counts the depth.
12334 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12335 instead of X509err, which often resulted in confusing error
12336 messages since the error codes are not globally unique
12337 (e.g. an alleged error in ssl3_accept when a certificate
12338 didn't match the private key).
12340 *) New function SSL_CTX_set_session_id_context that allows to set a default
12341 value (so that you don't need SSL_set_session_id_context for each
12342 connection using the SSL_CTX).
12345 *) OAEP decoding bug fix.
12348 *) Support INSTALL_PREFIX for package builders, as proposed by
12352 *) New Configure options "threads" and "no-threads". For systems
12353 where the proper compiler options are known (currently Solaris
12354 and Linux), "threads" is the default.
12357 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12360 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12361 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12362 such as /usr/local/bin.
12365 *) "make linux-shared" to build shared libraries.
12366 [Niels Poppe <niels@netbox.org>]
12368 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12371 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12372 extension adding in x509 utility.
12375 *) Remove NOPROTO sections and error code comments.
12378 *) Partial rewrite of the DEF file generator to now parse the ANSI
12382 *) New Configure options --prefix=DIR and --openssldir=DIR.
12385 *) Complete rewrite of the error code script(s). It is all now handled
12386 by one script at the top level which handles error code gathering,
12387 header rewriting and C source file generation. It should be much better
12388 than the old method: it now uses a modified version of Ulf's parser to
12389 read the ANSI prototypes in all header files (thus the old K&R definitions
12390 aren't needed for error creation any more) and do a better job of
12391 translating function codes into names. The old 'ASN1 error code imbedded
12392 in a comment' is no longer necessary and it doesn't use .err files which
12393 have now been deleted. Also the error code call doesn't have to appear all
12394 on one line (which resulted in some large lines...).
12397 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12400 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12401 0 (which usually indicates a closed connection), but continue reading.
12404 *) Fix some race conditions.
12407 *) Add support for CRL distribution points extension. Add Certificate
12408 Policies and CRL distribution points documentation.
12411 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12414 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12415 8 of keying material. Merlin has also confirmed interop with this fix
12416 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12417 [Merlin Hughes <merlin@baltimore.ie>]
12419 *) Fix lots of warnings.
12420 [Richard Levitte <levitte@stacken.kth.se>]
12422 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12423 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12424 [Richard Levitte <levitte@stacken.kth.se>]
12426 *) Fix problems with sizeof(long) == 8.
12427 [Andy Polyakov <appro@fy.chalmers.se>]
12429 *) Change functions to ANSI C.
12432 *) Fix typos in error codes.
12433 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12435 *) Remove defunct assembler files from Configure.
12438 *) SPARC v8 assembler BIGNUM implementation.
12439 [Andy Polyakov <appro@fy.chalmers.se>]
12441 *) Support for Certificate Policies extension: both print and set.
12442 Various additions to support the r2i method this uses.
12445 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12446 return a const string when you are expecting an allocated buffer.
12449 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12450 types DirectoryString and DisplayText.
12453 *) Add code to allow r2i extensions to access the configuration database,
12454 add an LHASH database driver and add several ctx helper functions.
12457 *) Fix an evil bug in bn_expand2() which caused various BN functions to
12458 fail when they extended the size of a BIGNUM.
12461 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
12462 support typesafe stack.
12465 *) Fix typo in SSL_[gs]et_options().
12466 [Nils Frostberg <nils@medcom.se>]
12468 *) Delete various functions and files that belonged to the (now obsolete)
12469 old X509V3 handling code.
12472 *) New Configure option "rsaref".
12475 *) Don't auto-generate pem.h.
12478 *) Introduce type-safe ASN.1 SETs.
12481 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
12482 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
12484 *) Introduce type-safe STACKs. This will almost certainly break lots of code
12485 that links with OpenSSL (well at least cause lots of warnings), but fear
12486 not: the conversion is trivial, and it eliminates loads of evil casts. A
12487 few STACKed things have been converted already. Feel free to convert more.
12488 In the fullness of time, I'll do away with the STACK type altogether.
12491 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
12492 specified in <certfile> by updating the entry in the index.txt file.
12493 This way one no longer has to edit the index.txt file manually for
12494 revoking a certificate. The -revoke option does the gory details now.
12495 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
12497 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
12498 `-text' option at all and this way the `-noout -text' combination was
12499 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
12500 [Ralf S. Engelschall]
12502 *) Make sure a corresponding plain text error message exists for the
12503 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
12504 verify callback function determined that a certificate was revoked.
12505 [Ralf S. Engelschall]
12507 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
12508 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
12509 all available ciphers including rc5, which was forgotten until now.
12510 In order to let the testing shell script know which algorithms
12511 are available, a new (up to now undocumented) command
12512 "openssl list-cipher-commands" is used.
12515 *) Bugfix: s_client occasionally would sleep in select() when
12516 it should have checked SSL_pending() first.
12519 *) New functions DSA_do_sign and DSA_do_verify to provide access to
12520 the raw DSA values prior to ASN.1 encoding.
12523 *) Tweaks to Configure
12524 [Niels Poppe <niels@netbox.org>]
12526 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
12530 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
12533 *) New config option to avoid instructions that are illegal on the 80386.
12534 The default code is faster, but requires at least a 486.
12537 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
12538 SSL2_SERVER_VERSION (not used at all) macros, which are now the
12539 same as SSL2_VERSION anyway.
12542 *) New "-showcerts" option for s_client.
12545 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
12546 application. Various cleanups and fixes.
12549 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
12550 modify error routines to work internally. Add error codes and PBE init
12551 to library startup routines.
12554 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
12555 packing functions to asn1 and evp. Changed function names and error
12556 codes along the way.
12559 *) PKCS12 integration: and so it begins... First of several patches to
12560 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
12561 objects to objects.h
12564 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
12565 and display support for Thawte strong extranet extension.
12568 *) Add LinuxPPC support.
12569 [Jeff Dubrule <igor@pobox.org>]
12571 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
12572 bn_div_words in alpha.s.
12573 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
12575 *) Make sure the RSA OAEP test is skipped under -DRSAref because
12576 OAEP isn't supported when OpenSSL is built with RSAref.
12577 [Ulf Moeller <ulf@fitug.de>]
12579 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
12580 so they no longer are missing under -DNOPROTO.
12581 [Soren S. Jorvang <soren@t.dk>]
12584 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
12586 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
12587 doesn't work when the session is reused. Coming soon!
12590 *) Fix a security hole, that allows sessions to be reused in the wrong
12591 context thus bypassing client cert protection! All software that uses
12592 client certs and session caches in multiple contexts NEEDS PATCHING to
12593 allow session reuse! A fuller solution is in the works.
12594 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
12596 *) Some more source tree cleanups (removed obsolete files
12597 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
12598 permission on "config" script to be executable) and a fix for the INSTALL
12600 [Ulf Moeller <ulf@fitug.de>]
12602 *) Remove some legacy and erroneous uses of malloc, free instead of
12604 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
12606 *) Make rsa_oaep_test return non-zero on error.
12607 [Ulf Moeller <ulf@fitug.de>]
12609 *) Add support for native Solaris shared libraries. Configure
12610 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12611 if someone would make that last step automatic.
12612 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12614 *) ctx_size was not built with the right compiler during "make links". Fixed.
12617 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12618 except NULL ciphers". This means the default cipher list will no longer
12619 enable NULL ciphers. They need to be specifically enabled e.g. with
12620 the string "DEFAULT:eNULL".
12623 *) Fix to RSA private encryption routines: if p < q then it would
12624 occasionally produce an invalid result. This will only happen with
12625 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12628 *) Be less restrictive and allow also `perl util/perlpath.pl
12629 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12630 because this way one can also use an interpreter named `perl5' (which is
12631 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12632 installed as `perl').
12633 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12635 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12636 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12638 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12639 advapi32.lib to Win32 build and change the pem test comparison
12640 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12641 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
12642 and crypto/des/ede_cbcm_enc.c.
12645 *) DES quad checksum was broken on big-endian architectures. Fixed.
12648 *) Comment out two functions in bio.h that aren't implemented. Fix up the
12649 Win32 test batch file so it (might) work again. The Win32 test batch file
12650 is horrible: I feel ill....
12653 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
12654 in e_os.h. Audit of header files to check ANSI and non ANSI
12655 sections: 10 functions were absent from non ANSI section and not exported
12656 from Windows DLLs. Fixed up libeay.num for new functions.
12659 *) Make `openssl version' output lines consistent.
12660 [Ralf S. Engelschall]
12662 *) Fix Win32 symbol export lists for BIO functions: Added
12663 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
12664 to ms/libeay{16,32}.def.
12665 [Ralf S. Engelschall]
12667 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
12668 fine under Unix and passes some trivial tests I've now added. But the
12669 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
12670 added to make sure no one expects that this stuff really works in the
12671 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
12672 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
12674 [Ralf S. Engelschall]
12676 *) Fix the generation of two part addresses in perl.
12677 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
12679 *) Add config entry for Linux on MIPS.
12680 [John Tobey <jtobey@channel1.com>]
12682 *) Make links whenever Configure is run, unless we are on Windoze.
12685 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
12686 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
12690 *) Add a useful kludge to allow package maintainers to specify compiler and
12691 other platforms details on the command line without having to patch the
12692 Configure script everytime: One now can use ``perl Configure
12693 <id>:<details>'', i.e. platform ids are allowed to have details appended
12694 to them (separated by colons). This is treated as there would be a static
12695 pre-configured entry in Configure's %table under key <id> with value
12696 <details> and ``perl Configure <id>'' is called. So, when you want to
12697 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
12698 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
12699 now, which overrides the FreeBSD-elf entry on-the-fly.
12700 [Ralf S. Engelschall]
12702 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
12705 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
12706 on the `perl Configure ...' command line. This way one can compile
12707 OpenSSL libraries with Position Independent Code (PIC) which is needed
12708 for linking it into DSOs.
12709 [Ralf S. Engelschall]
12711 *) Remarkably, export ciphers were totally broken and no-one had noticed!
12715 *) Cleaned up the LICENSE document: The official contact for any license
12716 questions now is the OpenSSL core team under openssl-core@openssl.org.
12717 And add a paragraph about the dual-license situation to make sure people
12718 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
12719 to the OpenSSL toolkit.
12720 [Ralf S. Engelschall]
12722 *) General source tree makefile cleanups: Made `making xxx in yyy...'
12723 display consistent in the source tree and replaced `/bin/rm' by `rm'.
12724 Additionally cleaned up the `make links' target: Remove unnecessary
12725 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
12726 to speed processing and no longer clutter the display with confusing
12727 stuff. Instead only the actually done links are displayed.
12728 [Ralf S. Engelschall]
12730 *) Permit null encryption ciphersuites, used for authentication only. It used
12731 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
12732 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12736 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12737 signed attributes when verifying signatures (this would break them),
12738 the detached data encoding was wrong and public keys obtained using
12739 X509_get_pubkey() weren't freed.
12742 *) Add text documentation for the BUFFER functions. Also added a work around
12743 to a Win95 console bug. This was triggered by the password read stuff: the
12744 last character typed gets carried over to the next fread(). If you were
12745 generating a new cert request using 'req' for example then the last
12746 character of the passphrase would be CR which would then enter the first
12750 *) Added the new `Includes OpenSSL Cryptography Software' button as
12751 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12752 button and can be used by applications based on OpenSSL to show the
12753 relationship to the OpenSSL project.
12754 [Ralf S. Engelschall]
12756 *) Remove confusing variables in function signatures in files
12757 ssl/ssl_lib.c and ssl/ssl.h.
12758 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12760 *) Don't install bss_file.c under PREFIX/include/
12761 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12763 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12764 functions that return function pointers and has support for NT specific
12765 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12766 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12767 unsigned to signed types: this was killing the Win32 compile.
12770 *) Add new certificate file to stack functions,
12771 SSL_add_dir_cert_subjects_to_stack() and
12772 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12773 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12774 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12775 This means that Apache-SSL and similar packages don't have to mess around
12776 to add as many CAs as they want to the preferred list.
12779 *) Experiment with doxygen documentation. Currently only partially applied to
12781 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12782 openssl.doxy as the configuration file.
12785 *) Get rid of remaining C++-style comments which strict C compilers hate.
12786 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12788 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12789 compiled in by default: it has problems with large keys.
12792 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12793 DH private keys and/or callback functions which directly correspond to
12794 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12795 is needed for applications which have to configure certificates on a
12796 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12798 For the RSA certificate situation is makes no difference, but
12799 for the DSA certificate situation this fixes the "no shared cipher"
12800 problem where the OpenSSL cipher selection procedure failed because the
12801 temporary keys were not overtaken from the context and the API provided
12802 no way to reconfigure them.
12803 The new functions now let applications reconfigure the stuff and they
12804 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12805 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12806 non-public-API function ssl_cert_instantiate() is used as a helper
12807 function and also to reduce code redundancy inside ssl_rsa.c.
12808 [Ralf S. Engelschall]
12810 *) Move s_server -dcert and -dkey options out of the undocumented feature
12811 area because they are useful for the DSA situation and should be
12812 recognized by the users.
12813 [Ralf S. Engelschall]
12815 *) Fix the cipher decision scheme for export ciphers: the export bits are
12816 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12817 SSL_EXP_MASK. So, the original variable has to be used instead of the
12818 already masked variable.
12819 [Richard Levitte <levitte@stacken.kth.se>]
12821 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12822 [Richard Levitte <levitte@stacken.kth.se>]
12824 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12825 from `int' to `unsigned int' because it's a length and initialized by
12826 EVP_DigestFinal() which expects an `unsigned int *'.
12827 [Richard Levitte <levitte@stacken.kth.se>]
12829 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12830 script. Instead use the usual Shell->Perl transition trick.
12831 [Ralf S. Engelschall]
12833 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12834 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12835 -noout -modulus' as it's already the case for `openssl rsa -noout
12836 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12837 currently the public key is printed (a decision which was already done by
12838 `openssl dsa -modulus' in the past) which serves a similar purpose.
12839 Additionally the NO_RSA no longer completely removes the whole -modulus
12840 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12842 [Ralf S. Engelschall]
12844 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12845 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12846 [Arne Ansper <arne@ats.cyber.ee>]
12848 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12849 to be added. Now both 'req' and 'ca' can use new objects defined in the
12853 *) Add cool BIO that does syslog (or event log on NT).
12854 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12856 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12857 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12858 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12859 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12862 *) Add preliminary config info for new extension code.
12865 *) Make RSA_NO_PADDING really use no padding.
12866 [Ulf Moeller <ulf@fitug.de>]
12868 *) Generate errors when private/public key check is done.
12871 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12872 for some CRL extensions and new objects added.
12875 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
12876 key usage extension and fuller support for authority key id.
12879 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
12880 padding method for RSA, which is recommended for new applications in PKCS
12881 #1 v2.0 (RFC 2437, October 1998).
12882 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
12883 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
12884 against Bleichbacher's attack on RSA.
12885 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
12888 *) Updates to the new SSL compression code
12889 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12891 *) Fix so that the version number in the master secret, when passed
12892 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
12893 (because the server will not accept higher), that the version number
12894 is 0x03,0x01, not 0x03,0x00
12895 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12897 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
12898 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
12899 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
12902 *) Support for RAW extensions where an arbitrary extension can be
12903 created by including its DER encoding. See apps/openssl.cnf for
12907 *) Make sure latest Perl versions don't interpret some generated C array
12908 code as Perl array code in the crypto/err/err_genc.pl script.
12909 [Lars Weber <3weber@informatik.uni-hamburg.de>]
12911 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
12912 not many people have the assembler. Various Win32 compilation fixes and
12913 update to the INSTALL.W32 file with (hopefully) more accurate Win32
12914 build instructions.
12917 *) Modify configure script 'Configure' to automatically create crypto/date.h
12918 file under Win32 and also build pem.h from pem.org. New script
12919 util/mkfiles.pl to create the MINFO file on environments that can't do a
12920 'make files': perl util/mkfiles.pl >MINFO should work.
12923 *) Major rework of DES function declarations, in the pursuit of correctness
12924 and purity. As a result, many evil casts evaporated, and some weirdness,
12925 too. You may find this causes warnings in your code. Zapping your evil
12926 casts will probably fix them. Mostly.
12929 *) Fix for a typo in asn1.h. Bug fix to object creation script
12930 obj_dat.pl. It considered a zero in an object definition to mean
12931 "end of object": none of the objects in objects.h have any zeros
12932 so it wasn't spotted.
12933 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
12935 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
12936 Masking (CBCM). In the absence of test vectors, the best I have been able
12937 to do is check that the decrypt undoes the encrypt, so far. Send me test
12938 vectors if you have them.
12941 *) Correct calculation of key length for export ciphers (too much space was
12942 allocated for null ciphers). This has not been tested!
12945 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
12946 message is now correct (it understands "crypto" and "ssl" on its
12947 command line). There is also now an "update" option. This will update
12948 the util/ssleay.num and util/libeay.num files with any new functions.
12950 perl util/mkdef.pl crypto ssl update
12951 it will update them.
12954 *) Overhauled the Perl interface (perl/*):
12955 - ported BN stuff to OpenSSL's different BN library
12956 - made the perl/ source tree CVS-aware
12957 - renamed the package from SSLeay to OpenSSL (the files still contain
12958 their history because I've copied them in the repository)
12959 - removed obsolete files (the test scripts will be replaced
12960 by better Test::Harness variants in the future)
12961 [Ralf S. Engelschall]
12963 *) First cut for a very conservative source tree cleanup:
12964 1. merge various obsolete readme texts into doc/ssleay.txt
12965 where we collect the old documents and readme texts.
12966 2. remove the first part of files where I'm already sure that we no
12967 longer need them because of three reasons: either they are just temporary
12968 files which were left by Eric or they are preserved original files where
12969 I've verified that the diff is also available in the CVS via "cvs diff
12970 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
12971 the crypto/md/ stuff).
12972 [Ralf S. Engelschall]
12974 *) More extension code. Incomplete support for subject and issuer alt
12975 name, issuer and authority key id. Change the i2v function parameters
12976 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
12977 what that's for :-) Fix to ASN1 macro which messed up
12978 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
12981 *) Preliminary support for ENUMERATED type. This is largely copied from the
12985 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
12986 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12988 *) Make sure `make rehash' target really finds the `openssl' program.
12989 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12991 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
12992 like to hear about it if this slows down other processors.
12995 *) Add CygWin32 platform information to Configure script.
12996 [Alan Batie <batie@aahz.jf.intel.com>]
12998 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
12999 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
13001 *) New program nseq to manipulate netscape certificate sequences
13004 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
13008 *) Fixes to BN code. Previously the default was to define BN_RECURSION
13009 but the BN code had some problems that would cause failures when
13010 doing certificate verification and some other functions.
13011 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
13013 *) Add ASN1 and PEM code to support netscape certificate sequences.
13016 *) Add ASN1 and PEM code to support netscape certificate sequences.
13019 *) Add several PKIX and private extended key usage OIDs.
13022 *) Modify the 'ca' program to handle the new extension code. Modify
13023 openssl.cnf for new extension format, add comments.
13026 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
13027 and add a sample to openssl.cnf so req -x509 now adds appropriate
13031 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
13032 error code, add initial support to X509_print() and x509 application.
13035 *) Takes a deep breath and start adding X509 V3 extension support code. Add
13036 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
13037 stuff is currently isolated and isn't even compiled yet.
13040 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
13041 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
13042 Removed the versions check from X509 routines when loading extensions:
13043 this allows certain broken certificates that don't set the version
13044 properly to be processed.
13047 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
13048 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
13049 can still be regenerated with "make depend".
13052 *) Spelling mistake in C version of CAST-128.
13053 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
13055 *) Changes to the error generation code. The perl script err-code.pl
13056 now reads in the old error codes and retains the old numbers, only
13057 adding new ones if necessary. It also only changes the .err files if new
13058 codes are added. The makefiles have been modified to only insert errors
13059 when needed (to avoid needlessly modifying header files). This is done
13060 by only inserting errors if the .err file is newer than the auto generated
13061 C file. To rebuild all the error codes from scratch (the old behaviour)
13062 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
13063 or delete all the .err files.
13066 *) CAST-128 was incorrectly implemented for short keys. The C version has
13067 been fixed, but is untested. The assembler versions are also fixed, but
13068 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
13069 to regenerate it if needed.
13070 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
13071 Hagino <itojun@kame.net>]
13073 *) File was opened incorrectly in randfile.c.
13074 [Ulf Möller <ulf@fitug.de>]
13076 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
13077 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
13078 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
13079 al: it's just almost always a UTCTime. Note this patch adds new error
13080 codes so do a "make errors" if there are problems.
13083 *) Correct Linux 1 recognition in config.
13084 [Ulf Möller <ulf@fitug.de>]
13086 *) Remove pointless MD5 hash when using DSA keys in ca.
13087 [Anonymous <nobody@replay.com>]
13089 *) Generate an error if given an empty string as a cert directory. Also
13090 generate an error if handed NULL (previously returned 0 to indicate an
13091 error, but didn't set one).
13092 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
13094 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
13097 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
13098 parameters. This was causing a warning which killed off the Win32 compile.
13101 *) Remove C++ style comments from crypto/bn/bn_local.h.
13102 [Neil Costigan <neil.costigan@celocom.com>]
13104 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
13105 based on a text string, looking up short and long names and finally
13106 "dot" format. The "dot" format stuff didn't work. Added new function
13107 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
13108 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
13109 OID is not part of the table.
13112 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
13113 X509_LOOKUP_by_alias().
13116 *) Sort openssl functions by name.
13119 *) Get the gendsa program working (hopefully) and add it to app list. Remove
13120 encryption from sample DSA keys (in case anyone is interested the password
13124 *) Make _all_ *_free functions accept a NULL pointer.
13125 [Frans Heymans <fheymans@isaserver.be>]
13127 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
13129 [Anonymous <nobody@replay.com>]
13131 *) s_server should send the CAfile as acceptable CAs, not its own cert.
13132 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13134 *) Don't blow it for numeric -newkey arguments to apps/req.
13135 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13137 *) Temp key "for export" tests were wrong in s3_srvr.c.
13138 [Anonymous <nobody@replay.com>]
13140 *) Add prototype for temp key callback functions
13141 SSL_CTX_set_tmp_{rsa,dh}_callback().
13144 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
13145 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
13148 *) X509_name_add_entry() freed the wrong thing after an error.
13149 [Arne Ansper <arne@ats.cyber.ee>]
13151 *) rsa_eay.c would attempt to free a NULL context.
13152 [Arne Ansper <arne@ats.cyber.ee>]
13154 *) BIO_s_socket() had a broken should_retry() on Windoze.
13155 [Arne Ansper <arne@ats.cyber.ee>]
13157 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
13158 [Arne Ansper <arne@ats.cyber.ee>]
13160 *) Make sure the already existing X509_STORE->depth variable is initialized
13161 in X509_STORE_new(), but document the fact that this variable is still
13162 unused in the certificate verification process.
13163 [Ralf S. Engelschall]
13165 *) Fix the various library and apps files to free up pkeys obtained from
13166 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
13169 *) Fix reference counting in X509_PUBKEY_get(). This makes
13170 demos/maurice/example2.c work, amongst others, probably.
13171 [Steve Henson and Ben Laurie]
13173 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
13174 `openssl' and second, the shortcut symlinks for the `openssl <command>'
13175 are no longer created. This way we have a single and consistent command
13176 line interface `openssl <command>', similar to `cvs <command>'.
13177 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
13179 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
13180 BIT STRING wrapper always have zero unused bits.
13183 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
13186 *) Make the top-level INSTALL documentation easier to understand.
13189 *) Makefiles updated to exit if an error occurs in a sub-directory
13190 make (including if user presses ^C) [Paul Sutton]
13192 *) Make Montgomery context stuff explicit in RSA data structure.
13195 *) Fix build order of pem and err to allow for generated pem.h.
13198 *) Fix renumbering bug in X509_NAME_delete_entry().
13201 *) Enhanced the err-ins.pl script so it makes the error library number
13202 global and can add a library name. This is needed for external ASN1 and
13203 other error libraries.
13206 *) Fixed sk_insert which never worked properly.
13209 *) Fix ASN1 macros so they can handle indefinite length constructed
13210 EXPLICIT tags. Some non standard certificates use these: they can now
13214 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
13215 into a single doc/ssleay.txt bundle. This way the information is still
13216 preserved but no longer messes up this directory. Now it's new room for
13217 the new set of documentation files.
13218 [Ralf S. Engelschall]
13220 *) SETs were incorrectly DER encoded. This was a major pain, because they
13221 shared code with SEQUENCEs, which aren't coded the same. This means that
13222 almost everything to do with SETs or SEQUENCEs has either changed name or
13223 number of arguments.
13224 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
13226 *) Fix test data to work with the above.
13229 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
13230 was already fixed by Eric for 0.9.1 it seems.
13231 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
13233 *) Autodetect FreeBSD3.
13236 *) Fix various bugs in Configure. This affects the following platforms:
13240 unixware-2.0-pentium
13244 *) Eliminate generated files from CVS. Reorder tests to regenerate files
13245 before they are needed.
13248 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
13252 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
13254 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
13255 changed SSLeay to OpenSSL in version strings.
13256 [Ralf S. Engelschall]
13258 *) Some fixups to the top-level documents.
13261 *) Fixed the nasty bug where rsaref.h was not found under compile-time
13262 because the symlink to include/ was missing.
13263 [Ralf S. Engelschall]
13265 *) Incorporated the popular no-RSA/DSA-only patches
13266 which allow to compile a RSA-free SSLeay.
13267 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
13269 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
13270 when "ssleay" is still not found.
13271 [Ralf S. Engelschall]
13273 *) Added more platforms to Configure: Cray T3E, HPUX 11,
13274 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
13276 *) Updated the README file.
13277 [Ralf S. Engelschall]
13279 *) Added various .cvsignore files in the CVS repository subdirs
13280 to make a "cvs update" really silent.
13281 [Ralf S. Engelschall]
13283 *) Recompiled the error-definition header files and added
13284 missing symbols to the Win32 linker tables.
13285 [Ralf S. Engelschall]
13287 *) Cleaned up the top-level documents;
13288 o new files: CHANGES and LICENSE
13289 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
13290 o merged COPYRIGHT into LICENSE
13291 o removed obsolete TODO file
13292 o renamed MICROSOFT to INSTALL.W32
13293 [Ralf S. Engelschall]
13295 *) Removed dummy files from the 0.9.1b source tree:
13296 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
13297 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
13298 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
13299 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
13300 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
13301 [Ralf S. Engelschall]
13303 *) Added various platform portability fixes.
13306 *) The Genesis of the OpenSSL rpject:
13307 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
13308 Young and Tim J. Hudson created while they were working for C2Net until
13310 [The OpenSSL Project]
13313 Changes between 0.9.0b and 0.9.1b [not released]
13315 *) Updated a few CA certificates under certs/
13318 *) Changed some BIGNUM api stuff.
13321 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13322 DGUX x86, Linux Alpha, etc.
13325 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13326 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13330 *) Add -strparse option to asn1pars program which parses nested
13332 [Dr Stephen Henson <shenson@bigfoot.com>]
13334 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13337 *) DSA fix for "ca" program.
13340 *) Added "-genkey" option to "dsaparam" program.
13343 *) Added RIPE MD160 (rmd160) message digest.
13346 *) Added -a (all) option to "ssleay version" command.
13349 *) Added PLATFORM define which is the id given to Configure.
13352 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13355 *) Extended the ASN.1 parser routines.
13358 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13361 *) Added a BN_CTX to the BN library.
13364 *) Fixed the weak key values in DES library
13367 *) Changed API in EVP library for cipher aliases.
13370 *) Added support for RC2/64bit cipher.
13373 *) Converted the lhash library to the crypto/mem.c functions.
13376 *) Added more recognized ASN.1 object ids.
13379 *) Added more RSA padding checks for SSL/TLS.
13382 *) Added BIO proxy/filter functionality.
13385 *) Added extra_certs to SSL_CTX which can be used
13386 send extra CA certificates to the client in the CA cert chain sending
13387 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13390 *) Now Fortezza is denied in the authentication phase because
13391 this is key exchange mechanism is not supported by SSLeay at all.
13394 *) Additional PKCS1 checks.
13397 *) Support the string "TLSv1" for all TLS v1 ciphers.
13400 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13401 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13404 *) Fixed a few memory leaks.
13407 *) Fixed various code and comment typos.
13410 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13411 bytes sent in the client random.
13412 [Edward Bishop <ebishop@spyglass.com>]
projects / openssl.git / blob