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 *) Default cipher lists/suites are now avaialble via a function, the
13 #defines are deprecated.
16 *) Add target VC-WIN32-UWP, VC-WIN64A-UWP, VC-WIN32-ARM-UWP and
17 VC-WIN64-ARM-UWP in Windows OneCore target for making building libraries
18 for Windows Store apps easier. Also, the "no-uplink" option has been added.
21 *) Join the directories crypto/x509 and crypto/x509v3
24 *) Change the default RSA, DSA and DH size to 2048 bit instead of 1024.
25 This changes the size when using the genpkey app when no size is given. It
26 fixes an omission in earlier changes that changed all RSA, DSA and DH
27 generation apps to use 2048 bits by default.
30 *) Added command 'openssl kdf' that uses the EVP_KDF API.
33 *) Added command 'openssl mac' that uses the EVP_MAC API.
36 *) Added OPENSSL_info() to get diverse built-in OpenSSL data, such
37 as default directories. Also added the command 'openssl info'
38 for scripting purposes.
41 *) The functions AES_ige_encrypt() and AES_bi_ige_encrypt() have been
42 deprecated. These undocumented functions were never integrated into the EVP
43 layer and implement the AES Infinite Garble Extension (IGE) mode and AES
44 Bi-directional IGE mode. These modes were never formally standardised and
45 usage of these functions is believed to be very small. In particular
46 AES_bi_ige_encrypt() has a known bug. It accepts 2 AES keys, but only one
47 is ever used. The security implications are believed to be minimal, but
48 this issue was never fixed for backwards compatibility reasons. New code
49 should not use these modes.
52 *) Add prediction resistance to the DRBG reseeding process.
55 *) Limit the number of blocks in a data unit for AES-XTS to 2^20 as
56 mandated by IEEE Std 1619-2018.
59 *) Added newline escaping functionality to a filename when using openssl dgst.
60 This output format is to replicate the output format found in the '*sum'
61 checksum programs. This aims to preserve backward compatibility.
62 [Matt Eaton, Richard Levitte, and Paul Dale]
64 *) Removed the heartbeat message in DTLS feature, as it has very
65 little usage and doesn't seem to fulfill a valuable purpose.
66 The configuration option is now deprecated.
69 *) Changed the output of 'openssl {digestname} < file' to display the
70 digest name in its output.
73 *) Added a new generic trace API which provides support for enabling
74 instrumentation through trace output. This feature is mainly intended
75 as an aid for developers and is disabled by default. To utilize it,
76 OpenSSL needs to be configured with the `enable-trace` option.
78 If the tracing API is enabled, the application can activate trace output
79 by registering BIOs as trace channels for a number of tracing and debugging
82 The 'openssl' application has been expanded to enable any of the types
83 available via environment variables defined by the user, and serves as
84 one possible example on how to use this functionality.
85 [Richard Levitte & Matthias St. Pierre]
87 *) Added build tests for C++. These are generated files that only do one
88 thing, to include one public OpenSSL head file each. This tests that
89 the public header files can be usefully included in a C++ application.
91 This test isn't enabled by default. It can be enabled with the option
92 'enable-buildtest-c++'.
95 *) Add Single Step KDF (EVP_KDF_SS) to EVP_KDF.
98 *) Add KMAC to EVP_MAC.
101 *) Added property based algorithm implementation selection framework to
105 *) Added SCA hardening for modular field inversion in EC_GROUP through
106 a new dedicated field_inv() pointer in EC_METHOD.
107 This also addresses a leakage affecting conversions from projective
108 to affine coordinates.
109 [Billy Bob Brumley, Nicola Tuveri]
111 *) Added EVP_KDF, an EVP layer KDF API, to simplify adding KDF and PRF
112 implementations. This includes an EVP_PKEY to EVP_KDF bridge for
113 those algorithms that were already supported through the EVP_PKEY API
114 (scrypt, TLS1 PRF and HKDF). The low-level KDF functions for PBKDF2
115 and scrypt are now wrappers that call EVP_KDF.
118 *) Build devcrypto engine as a dynamic engine.
121 *) Add keyed BLAKE2 to EVP_MAC.
124 *) Fix a bug in the computation of the endpoint-pair shared secret used
125 by DTLS over SCTP. This breaks interoperability with older versions
126 of OpenSSL like OpenSSL 1.1.0 and OpenSSL 1.0.2. There is a runtime
127 switch SSL_MODE_DTLS_SCTP_LABEL_LENGTH_BUG (off by default) enabling
128 interoperability with such broken implementations. However, enabling
129 this switch breaks interoperability with correct implementations.
131 *) Fix a use after free bug in d2i_X509_PUBKEY when overwriting a
132 re-used X509_PUBKEY object if the second PUBKEY is malformed.
135 *) Move strictness check from EVP_PKEY_asn1_new() to EVP_PKEY_asn1_add0().
138 *) Change the license to the Apache License v2.0.
141 *) Change the possible version information given with OPENSSL_API_COMPAT.
142 It may be a pre-3.0.0 style numerical version number as it was defined
143 in 1.1.0, and it may also simply take the major version number.
145 Because of the version numbering of pre-3.0.0 releases, the values 0,
146 1 and 2 are equivalent to 0x00908000L (0.9.8), 0x10000000L (1.0.0) and
147 0x10100000L (1.1.0), respectively.
150 *) Switch to a new version scheme using three numbers MAJOR.MINOR.PATCH.
152 o Major releases (indicated by incrementing the MAJOR release number)
153 may introduce incompatible API/ABI changes.
154 o Minor releases (indicated by incrementing the MINOR release number)
155 may introduce new features but retain API/ABI compatibility.
156 o Patch releases (indicated by incrementing the PATCH number)
157 are intended for bug fixes and other improvements of existing
158 features only (like improving performance or adding documentation)
159 and retain API/ABI compatibility.
162 *) Add support for RFC5297 SIV mode (siv128), including AES-SIV.
165 *) Remove the 'dist' target and add a tarball building script. The
166 'dist' target has fallen out of use, and it shouldn't be
167 necessary to configure just to create a source distribution.
170 *) Recreate the OS390-Unix config target. It no longer relies on a
171 special script like it did for OpenSSL pre-1.1.0.
174 *) Instead of having the source directories listed in Configure, add
175 a 'build.info' keyword SUBDIRS to indicate what sub-directories to
179 *) Add GMAC to EVP_MAC.
182 *) Ported the HMAC, CMAC and SipHash EVP_PKEY_METHODs to EVP_MAC.
185 *) Added EVP_MAC, an EVP layer MAC API, to simplify adding MAC
186 implementations. This includes a generic EVP_PKEY to EVP_MAC bridge,
187 to facilitate the continued use of MACs through raw private keys in
188 functionality such as EVP_DigestSign* and EVP_DigestVerify*.
191 *) Deprecate ECDH_KDF_X9_62() and mark its replacement as internal. Users
192 should use the EVP interface instead (EVP_PKEY_CTX_set_ecdh_kdf_type).
195 *) Added EVP_PKEY_ECDH_KDF_X9_63 and ecdh_KDF_X9_63() as replacements for
196 the EVP_PKEY_ECDH_KDF_X9_62 KDF type and ECDH_KDF_X9_62(). The old names
197 are retained for backwards compatibility.
200 *) AES-XTS mode now enforces that its two keys are different to mitigate
201 the attacked described in "Efficient Instantiations of Tweakable
202 Blockciphers and Refinements to Modes OCB and PMAC" by Phillip Rogaway.
203 Details of this attack can be obtained from:
204 http://web.cs.ucdavis.edu/%7Erogaway/papers/offsets.pdf
207 *) Rename the object files, i.e. give them other names than in previous
208 versions. Their names now include the name of the final product, as
209 well as its type mnemonic (bin, lib, shlib).
212 *) Added new option for 'openssl list', '-objects', which will display the
213 list of built in objects, i.e. OIDs with names.
216 *) Added support for Linux Kernel TLS data-path. The Linux Kernel data-path
217 improves application performance by removing data copies and providing
218 applications with zero-copy system calls such as sendfile and splice.
221 Changes between 1.1.1a and 1.1.1b [xx XXX xxxx]
223 *) Change the info callback signals for the start and end of a post-handshake
224 message exchange in TLSv1.3. In 1.1.1/1.1.1a we used SSL_CB_HANDSHAKE_START
225 and SSL_CB_HANDSHAKE_DONE. Experience has shown that many applications get
226 confused by this and assume that a TLSv1.2 renegotiation has started. This
227 can break KeyUpdate handling. Instead we no longer signal the start and end
228 of a post handshake message exchange (although the messages themselves are
229 still signalled). This could break some applications that were expecting
230 the old signals. However without this KeyUpdate is not usable for many
234 Changes between 1.1.1 and 1.1.1a [20 Nov 2018]
236 *) Timing vulnerability in DSA signature generation
238 The OpenSSL DSA signature algorithm has been shown to be vulnerable to a
239 timing side channel attack. An attacker could use variations in the signing
240 algorithm to recover the private key.
242 This issue was reported to OpenSSL on 16th October 2018 by Samuel Weiser.
246 *) Timing vulnerability in ECDSA signature generation
248 The OpenSSL ECDSA signature algorithm has been shown to be vulnerable to a
249 timing side channel attack. An attacker could use variations in the signing
250 algorithm to recover the private key.
252 This issue was reported to OpenSSL on 25th October 2018 by Samuel Weiser.
256 *) Fixed the issue that RAND_add()/RAND_seed() silently discards random input
257 if its length exceeds 4096 bytes. The limit has been raised to a buffer size
258 of two gigabytes and the error handling improved.
260 This issue was reported to OpenSSL by Dr. Falko Strenzke. It has been
261 categorized as a normal bug, not a security issue, because the DRBG reseeds
262 automatically and is fully functional even without additional randomness
263 provided by the application.
265 Changes between 1.1.0i and 1.1.1 [11 Sep 2018]
267 *) Add a new ClientHello callback. Provides a callback interface that gives
268 the application the ability to adjust the nascent SSL object at the
269 earliest stage of ClientHello processing, immediately after extensions have
270 been collected but before they have been processed. In particular, this
271 callback can adjust the supported TLS versions in response to the contents
275 *) Add SM2 base algorithm support.
278 *) s390x assembly pack: add (improved) hardware-support for the following
279 cryptographic primitives: sha3, shake, aes-gcm, aes-ccm, aes-ctr, aes-ofb,
280 aes-cfb/cfb8, aes-ecb.
283 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
284 parameter is no longer accepted, as it leads to a corrupt table. NULL
285 pem_str is reserved for alias entries only.
288 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
289 step for prime curves. The new implementation is based on formulae from
290 differential addition-and-doubling in homogeneous projective coordinates
291 from Izu-Takagi "A fast parallel elliptic curve multiplication resistant
292 against side channel attacks" and Brier-Joye "Weierstrass Elliptic Curves
293 and Side-Channel Attacks" Eq. (8) for y-coordinate recovery, modified
294 to work in projective coordinates.
295 [Billy Bob Brumley, Nicola Tuveri]
297 *) Change generating and checking of primes so that the error rate of not
298 being prime depends on the intended use based on the size of the input.
299 For larger primes this will result in more rounds of Miller-Rabin.
300 The maximal error rate for primes with more than 1080 bits is lowered
302 [Kurt Roeckx, Annie Yousar]
304 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
307 *) The 'tsget' script is renamed to 'tsget.pl', to avoid confusion when
308 moving between systems, and to avoid confusion when a Windows build is
309 done with mingw vs with MSVC. For POSIX installs, there's still a
310 symlink or copy named 'tsget' to avoid that confusion as well.
313 *) Revert blinding in ECDSA sign and instead make problematic addition
314 length-invariant. Switch even to fixed-length Montgomery multiplication.
317 *) Use the new ec_scalar_mul_ladder scaffold to implement a specialized ladder
318 step for binary curves. The new implementation is based on formulae from
319 differential addition-and-doubling in mixed Lopez-Dahab projective
320 coordinates, modified to independently blind the operands.
321 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
323 *) Add a scaffold to optionally enhance the Montgomery ladder implementation
324 for `ec_scalar_mul_ladder` (formerly `ec_mul_consttime`) allowing
325 EC_METHODs to implement their own specialized "ladder step", to take
326 advantage of more favorable coordinate systems or more efficient
327 differential addition-and-doubling algorithms.
328 [Billy Bob Brumley, Sohaib ul Hassan, Nicola Tuveri]
330 *) Modified the random device based seed sources to keep the relevant
331 file descriptors open rather than reopening them on each access.
332 This allows such sources to operate in a chroot() jail without
333 the associated device nodes being available. This behaviour can be
334 controlled using RAND_keep_random_devices_open().
337 *) Numerous side-channel attack mitigations have been applied. This may have
338 performance impacts for some algorithms for the benefit of improved
339 security. Specific changes are noted in this change log by their respective
343 *) AIX shared library support overhaul. Switch to AIX "natural" way of
344 handling shared libraries, which means collecting shared objects of
345 different versions and bitnesses in one common archive. This allows to
346 mitigate conflict between 1.0 and 1.1 side-by-side installations. It
347 doesn't affect the way 3rd party applications are linked, only how
348 multi-version installation is managed.
351 *) Make ec_group_do_inverse_ord() more robust and available to other
352 EC cryptosystems, so that irrespective of BN_FLG_CONSTTIME, SCA
353 mitigations are applied to the fallback BN_mod_inverse().
354 When using this function rather than BN_mod_inverse() directly, new
355 EC cryptosystem implementations are then safer-by-default.
358 *) Add coordinate blinding for EC_POINT and implement projective
359 coordinate blinding for generic prime curves as a countermeasure to
360 chosen point SCA attacks.
361 [Sohaib ul Hassan, Nicola Tuveri, Billy Bob Brumley]
363 *) Add blinding to ECDSA and DSA signatures to protect against side channel
364 attacks discovered by Keegan Ryan (NCC Group).
367 *) Enforce checking in the pkeyutl command line app to ensure that the input
368 length does not exceed the maximum supported digest length when performing
369 a sign, verify or verifyrecover operation.
372 *) SSL_MODE_AUTO_RETRY is enabled by default. Applications that use blocking
373 I/O in combination with something like select() or poll() will hang. This
374 can be turned off again using SSL_CTX_clear_mode().
375 Many applications do not properly handle non-application data records, and
376 TLS 1.3 sends more of such records. Setting SSL_MODE_AUTO_RETRY works
377 around the problems in those applications, but can also break some.
378 It's recommended to read the manpages about SSL_read(), SSL_write(),
379 SSL_get_error(), SSL_shutdown(), SSL_CTX_set_mode() and
380 SSL_CTX_set_read_ahead() again.
383 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
384 now allow empty (zero character) pass phrases.
387 *) Apply blinding to binary field modular inversion and remove patent
388 pending (OPENSSL_SUN_GF2M_DIV) BN_GF2m_mod_div implementation.
391 *) Deprecate ec2_mult.c and unify scalar multiplication code paths for
392 binary and prime elliptic curves.
395 *) Remove ECDSA nonce padding: EC_POINT_mul is now responsible for
396 constant time fixed point multiplication.
399 *) Revise elliptic curve scalar multiplication with timing attack
400 defenses: ec_wNAF_mul redirects to a constant time implementation
401 when computing fixed point and variable point multiplication (which
402 in OpenSSL are mostly used with secret scalars in keygen, sign,
403 ECDH derive operations).
404 [Billy Bob Brumley, Nicola Tuveri, Cesar Pereida GarcÃa,
407 *) Updated CONTRIBUTING
410 *) Updated DRBG / RAND to request nonce and additional low entropy
411 randomness from the system.
412 [Matthias St. Pierre]
414 *) Updated 'openssl rehash' to use OpenSSL consistent default.
417 *) Moved the load of the ssl_conf module to libcrypto, which helps
418 loading engines that libssl uses before libssl is initialised.
421 *) Added EVP_PKEY_sign() and EVP_PKEY_verify() for EdDSA
424 *) Fixed X509_NAME_ENTRY_set to get multi-valued RDNs right in all cases.
425 [Ingo Schwarze, Rich Salz]
427 *) Added output of accepting IP address and port for 'openssl s_server'
430 *) Added a new API for TLSv1.3 ciphersuites:
431 SSL_CTX_set_ciphersuites()
432 SSL_set_ciphersuites()
435 *) Memory allocation failures consistenly add an error to the error
439 *) Don't use OPENSSL_ENGINES and OPENSSL_CONF environment values
440 in libcrypto when run as setuid/setgid.
443 *) Load any config file by default when libssl is used.
446 *) Added new public header file <openssl/rand_drbg.h> and documentation
447 for the RAND_DRBG API. See manual page RAND_DRBG(7) for an overview.
448 [Matthias St. Pierre]
450 *) QNX support removed (cannot find contributors to get their approval
451 for the license change).
454 *) TLSv1.3 replay protection for early data has been implemented. See the
455 SSL_read_early_data() man page for further details.
458 *) Separated TLSv1.3 ciphersuite configuration out from TLSv1.2 ciphersuite
459 configuration. TLSv1.3 ciphersuites are not compatible with TLSv1.2 and
460 below. Similarly TLSv1.2 ciphersuites are not compatible with TLSv1.3.
461 In order to avoid issues where legacy TLSv1.2 ciphersuite configuration
462 would otherwise inadvertently disable all TLSv1.3 ciphersuites the
463 configuration has been separated out. See the ciphers man page or the
464 SSL_CTX_set_ciphersuites() man page for more information.
467 *) On POSIX (BSD, Linux, ...) systems the ocsp(1) command running
468 in responder mode now supports the new "-multi" option, which
469 spawns the specified number of child processes to handle OCSP
470 requests. The "-timeout" option now also limits the OCSP
471 responder's patience to wait to receive the full client request
472 on a newly accepted connection. Child processes are respawned
473 as needed, and the CA index file is automatically reloaded
474 when changed. This makes it possible to run the "ocsp" responder
475 as a long-running service, making the OpenSSL CA somewhat more
476 feature-complete. In this mode, most diagnostic messages logged
477 after entering the event loop are logged via syslog(3) rather than
481 *) Added support for X448 and Ed448. Heavily based on original work by
485 *) Extend OSSL_STORE with capabilities to search and to narrow the set of
486 objects loaded. This adds the functions OSSL_STORE_expect() and
487 OSSL_STORE_find() as well as needed tools to construct searches and
488 get the search data out of them.
491 *) Support for TLSv1.3 added. Note that users upgrading from an earlier
492 version of OpenSSL should review their configuration settings to ensure
493 that they are still appropriate for TLSv1.3. For further information see:
494 https://wiki.openssl.org/index.php/TLS1.3
497 *) Grand redesign of the OpenSSL random generator
499 The default RAND method now utilizes an AES-CTR DRBG according to
500 NIST standard SP 800-90Ar1. The new random generator is essentially
501 a port of the default random generator from the OpenSSL FIPS 2.0
502 object module. It is a hybrid deterministic random bit generator
503 using an AES-CTR bit stream and which seeds and reseeds itself
504 automatically using trusted system entropy sources.
506 Some of its new features are:
507 o Support for multiple DRBG instances with seed chaining.
508 o The default RAND method makes use of a DRBG.
509 o There is a public and private DRBG instance.
510 o The DRBG instances are fork-safe.
511 o Keep all global DRBG instances on the secure heap if it is enabled.
512 o The public and private DRBG instance are per thread for lock free
514 [Paul Dale, Benjamin Kaduk, Kurt Roeckx, Rich Salz, Matthias St. Pierre]
516 *) Changed Configure so it only says what it does and doesn't dump
517 so much data. Instead, ./configdata.pm should be used as a script
518 to display all sorts of configuration data.
521 *) Added processing of "make variables" to Configure.
524 *) Added SHA512/224 and SHA512/256 algorithm support.
527 *) The last traces of Netware support, first removed in 1.1.0, have
531 *) Get rid of Makefile.shared, and in the process, make the processing
532 of certain files (rc.obj, or the .def/.map/.opt files produced from
533 the ordinal files) more visible and hopefully easier to trace and
534 debug (or make silent).
537 *) Make it possible to have environment variable assignments as
538 arguments to config / Configure.
541 *) Add multi-prime RSA (RFC 8017) support.
544 *) Add SM3 implemented according to GB/T 32905-2016
545 [ Jack Lloyd <jack.lloyd@ribose.com>,
546 Ronald Tse <ronald.tse@ribose.com>,
547 Erick Borsboom <erick.borsboom@ribose.com> ]
549 *) Add 'Maximum Fragment Length' TLS extension negotiation and support
550 as documented in RFC6066.
551 Based on a patch from Tomasz Moń
552 [Filipe Raimundo da Silva]
554 *) Add SM4 implemented according to GB/T 32907-2016.
555 [ Jack Lloyd <jack.lloyd@ribose.com>,
556 Ronald Tse <ronald.tse@ribose.com>,
557 Erick Borsboom <erick.borsboom@ribose.com> ]
559 *) Reimplement -newreq-nodes and ERR_error_string_n; the
560 original author does not agree with the license change.
563 *) Add ARIA AEAD TLS support.
566 *) Some macro definitions to support VS6 have been removed. Visual
567 Studio 6 has not worked since 1.1.0
570 *) Add ERR_clear_last_mark(), to allow callers to clear the last mark
571 without clearing the errors.
574 *) Add "atfork" functions. If building on a system that without
575 pthreads, see doc/man3/OPENSSL_fork_prepare.pod for application
576 requirements. The RAND facility now uses/requires this.
582 *) The UI API becomes a permanent and integral part of libcrypto, i.e.
583 not possible to disable entirely. However, it's still possible to
584 disable the console reading UI method, UI_OpenSSL() (use UI_null()
587 To disable, configure with 'no-ui-console'. 'no-ui' is still
588 possible to use as an alias. Check at compile time with the
589 macro OPENSSL_NO_UI_CONSOLE. The macro OPENSSL_NO_UI is still
590 possible to check and is an alias for OPENSSL_NO_UI_CONSOLE.
593 *) Add a STORE module, which implements a uniform and URI based reader of
594 stores that can contain keys, certificates, CRLs and numerous other
595 objects. The main API is loosely based on a few stdio functions,
596 and includes OSSL_STORE_open, OSSL_STORE_load, OSSL_STORE_eof,
597 OSSL_STORE_error and OSSL_STORE_close.
598 The implementation uses backends called "loaders" to implement arbitrary
599 URI schemes. There is one built in "loader" for the 'file' scheme.
602 *) Add devcrypto engine. This has been implemented against cryptodev-linux,
603 then adjusted to work on FreeBSD 8.4 as well.
604 Enable by configuring with 'enable-devcryptoeng'. This is done by default
605 on BSD implementations, as cryptodev.h is assumed to exist on all of them.
608 *) Module names can prefixed with OSSL_ or OPENSSL_. This affects
609 util/mkerr.pl, which is adapted to allow those prefixes, leading to
610 error code calls like this:
612 OSSL_FOOerr(OSSL_FOO_F_SOMETHING, OSSL_FOO_R_WHATEVER);
614 With this change, we claim the namespaces OSSL and OPENSSL in a manner
615 that can be encoded in C. For the foreseeable future, this will only
617 [Richard Levitte and Tim Hudson]
619 *) Removed BSD cryptodev engine.
622 *) Add a build target 'build_all_generated', to build all generated files
623 and only that. This can be used to prepare everything that requires
624 things like perl for a system that lacks perl and then move everything
625 to that system and do the rest of the build there.
628 *) In the UI interface, make it possible to duplicate the user data. This
629 can be used by engines that need to retain the data for a longer time
630 than just the call where this user data is passed.
633 *) Ignore the '-named_curve auto' value for compatibility of applications
635 [Tomas Mraz <tmraz@fedoraproject.org>]
637 *) Fragmented SSL/TLS alerts are no longer accepted. An alert message is 2
638 bytes long. In theory it is permissible in SSLv3 - TLSv1.2 to fragment such
639 alerts across multiple records (some of which could be empty). In practice
640 it make no sense to send an empty alert record, or to fragment one. TLSv1.3
641 prohibts this altogether and other libraries (BoringSSL, NSS) do not
642 support this at all. Supporting it adds significant complexity to the
643 record layer, and its removal is unlikely to cause inter-operability
647 *) Add the ASN.1 types INT32, UINT32, INT64, UINT64 and variants prefixed
648 with Z. These are meant to replace LONG and ZLONG and to be size safe.
649 The use of LONG and ZLONG is discouraged and scheduled for deprecation
653 *) Add the 'z' and 'j' modifiers to BIO_printf() et al formatting string,
654 'z' is to be used for [s]size_t, and 'j' - with [u]int64_t.
655 [Richard Levitte, Andy Polyakov]
657 *) Add EC_KEY_get0_engine(), which does for EC_KEY what RSA_get0_engine()
661 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
662 platform rather than 'mingw'.
665 *) The functions X509_STORE_add_cert and X509_STORE_add_crl return
666 success if they are asked to add an object which already exists
667 in the store. This change cascades to other functions which load
668 certificates and CRLs.
671 *) x86_64 assembly pack: annotate code with DWARF CFI directives to
672 facilitate stack unwinding even from assembly subroutines.
675 *) Remove VAX C specific definitions of OPENSSL_EXPORT, OPENSSL_EXTERN.
676 Also remove OPENSSL_GLOBAL entirely, as it became a no-op.
679 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
680 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
681 which is the minimum version we support.
684 *) Certificate time validation (X509_cmp_time) enforces stricter
685 compliance with RFC 5280. Fractional seconds and timezone offsets
686 are no longer allowed.
689 *) Add support for ARIA
692 *) s_client will now send the Server Name Indication (SNI) extension by
693 default unless the new "-noservername" option is used. The server name is
694 based on the host provided to the "-connect" option unless overridden by
698 *) Add support for SipHash
701 *) OpenSSL now fails if it receives an unrecognised record type in TLS1.0
702 or TLS1.1. Previously this only happened in SSLv3 and TLS1.2. This is to
703 prevent issues where no progress is being made and the peer continually
704 sends unrecognised record types, using up resources processing them.
707 *) 'openssl passwd' can now produce SHA256 and SHA512 based output,
708 using the algorithm defined in
709 https://www.akkadia.org/drepper/SHA-crypt.txt
712 *) Heartbeat support has been removed; the ABI is changed for now.
713 [Richard Levitte, Rich Salz]
715 *) Support for SSL_OP_NO_ENCRYPT_THEN_MAC in SSL_CONF_cmd.
718 *) The RSA "null" method, which was partially supported to avoid patent
719 issues, has been replaced to always returns NULL.
723 Changes between 1.1.0h and 1.1.0i [xx XXX xxxx]
725 *) Client DoS due to large DH parameter
727 During key agreement in a TLS handshake using a DH(E) based ciphersuite a
728 malicious server can send a very large prime value to the client. This will
729 cause the client to spend an unreasonably long period of time generating a
730 key for this prime resulting in a hang until the client has finished. This
731 could be exploited in a Denial Of Service attack.
733 This issue was reported to OpenSSL on 5th June 2018 by Guido Vranken
737 *) Cache timing vulnerability in RSA Key Generation
739 The OpenSSL RSA Key generation algorithm has been shown to be vulnerable to
740 a cache timing side channel attack. An attacker with sufficient access to
741 mount cache timing attacks during the RSA key generation process could
742 recover the private key.
744 This issue was reported to OpenSSL on 4th April 2018 by Alejandro Cabrera
745 Aldaya, Billy Brumley, Cesar Pereida Garcia and Luis Manuel Alvarez Tapia.
749 *) Make EVP_PKEY_asn1_new() a bit stricter about its input. A NULL pem_str
750 parameter is no longer accepted, as it leads to a corrupt table. NULL
751 pem_str is reserved for alias entries only.
754 *) Revert blinding in ECDSA sign and instead make problematic addition
755 length-invariant. Switch even to fixed-length Montgomery multiplication.
758 *) Change generating and checking of primes so that the error rate of not
759 being prime depends on the intended use based on the size of the input.
760 For larger primes this will result in more rounds of Miller-Rabin.
761 The maximal error rate for primes with more than 1080 bits is lowered
763 [Kurt Roeckx, Annie Yousar]
765 *) Increase the number of Miller-Rabin rounds for DSA key generating to 64.
768 *) Add blinding to ECDSA and DSA signatures to protect against side channel
769 attacks discovered by Keegan Ryan (NCC Group).
772 *) When unlocking a pass phrase protected PEM file or PKCS#8 container, we
773 now allow empty (zero character) pass phrases.
776 *) Certificate time validation (X509_cmp_time) enforces stricter
777 compliance with RFC 5280. Fractional seconds and timezone offsets
778 are no longer allowed.
781 *) Fixed a text canonicalisation bug in CMS
783 Where a CMS detached signature is used with text content the text goes
784 through a canonicalisation process first prior to signing or verifying a
785 signature. This process strips trailing space at the end of lines, converts
786 line terminators to CRLF and removes additional trailing line terminators
787 at the end of a file. A bug in the canonicalisation process meant that
788 some characters, such as form-feed, were incorrectly treated as whitespace
789 and removed. This is contrary to the specification (RFC5485). This fix
790 could mean that detached text data signed with an earlier version of
791 OpenSSL 1.1.0 may fail to verify using the fixed version, or text data
792 signed with a fixed OpenSSL may fail to verify with an earlier version of
793 OpenSSL 1.1.0. A workaround is to only verify the canonicalised text data
794 and use the "-binary" flag (for the "cms" command line application) or set
795 the SMIME_BINARY/PKCS7_BINARY/CMS_BINARY flags (if using CMS_verify()).
798 Changes between 1.1.0g and 1.1.0h [27 Mar 2018]
800 *) Constructed ASN.1 types with a recursive definition could exceed the stack
802 Constructed ASN.1 types with a recursive definition (such as can be found
803 in PKCS7) could eventually exceed the stack given malicious input with
804 excessive recursion. This could result in a Denial Of Service attack. There
805 are no such structures used within SSL/TLS that come from untrusted sources
806 so this is considered safe.
808 This issue was reported to OpenSSL on 4th January 2018 by the OSS-fuzz
813 *) Incorrect CRYPTO_memcmp on HP-UX PA-RISC
815 Because of an implementation bug the PA-RISC CRYPTO_memcmp function is
816 effectively reduced to only comparing the least significant bit of each
817 byte. This allows an attacker to forge messages that would be considered as
818 authenticated in an amount of tries lower than that guaranteed by the
819 security claims of the scheme. The module can only be compiled by the
820 HP-UX assembler, so that only HP-UX PA-RISC targets are affected.
822 This issue was reported to OpenSSL on 2nd March 2018 by Peter Waltenberg
827 *) Add a build target 'build_all_generated', to build all generated files
828 and only that. This can be used to prepare everything that requires
829 things like perl for a system that lacks perl and then move everything
830 to that system and do the rest of the build there.
833 *) Backport SSL_OP_NO_RENGOTIATION
835 OpenSSL 1.0.2 and below had the ability to disable renegotiation using the
836 (undocumented) SSL3_FLAGS_NO_RENEGOTIATE_CIPHERS flag. Due to the opacity
837 changes this is no longer possible in 1.1.0. Therefore the new
838 SSL_OP_NO_RENEGOTIATION option from 1.1.1-dev has been backported to
839 1.1.0 to provide equivalent functionality.
841 Note that if an application built against 1.1.0h headers (or above) is run
842 using an older version of 1.1.0 (prior to 1.1.0h) then the option will be
843 accepted but nothing will happen, i.e. renegotiation will not be prevented.
846 *) Removed the OS390-Unix config target. It relied on a script that doesn't
850 *) rsaz_1024_mul_avx2 overflow bug on x86_64
852 There is an overflow bug in the AVX2 Montgomery multiplication procedure
853 used in exponentiation with 1024-bit moduli. No EC algorithms are affected.
854 Analysis suggests that attacks against RSA and DSA as a result of this
855 defect would be very difficult to perform and are not believed likely.
856 Attacks against DH1024 are considered just feasible, because most of the
857 work necessary to deduce information about a private key may be performed
858 offline. The amount of resources required for such an attack would be
859 significant. However, for an attack on TLS to be meaningful, the server
860 would have to share the DH1024 private key among multiple clients, which is
861 no longer an option since CVE-2016-0701.
863 This only affects processors that support the AVX2 but not ADX extensions
864 like Intel Haswell (4th generation).
866 This issue was reported to OpenSSL by David Benjamin (Google). The issue
867 was originally found via the OSS-Fuzz project.
871 Changes between 1.1.0f and 1.1.0g [2 Nov 2017]
873 *) bn_sqrx8x_internal carry bug on x86_64
875 There is a carry propagating bug in the x86_64 Montgomery squaring
876 procedure. No EC algorithms are affected. Analysis suggests that attacks
877 against RSA and DSA as a result of this defect would be very difficult to
878 perform and are not believed likely. Attacks against DH are considered just
879 feasible (although very difficult) because most of the work necessary to
880 deduce information about a private key may be performed offline. The amount
881 of resources required for such an attack would be very significant and
882 likely only accessible to a limited number of attackers. An attacker would
883 additionally need online access to an unpatched system using the target
884 private key in a scenario with persistent DH parameters and a private
885 key that is shared between multiple clients.
887 This only affects processors that support the BMI1, BMI2 and ADX extensions
888 like Intel Broadwell (5th generation) and later or AMD Ryzen.
890 This issue was reported to OpenSSL by the OSS-Fuzz project.
894 *) Malformed X.509 IPAddressFamily could cause OOB read
896 If an X.509 certificate has a malformed IPAddressFamily extension,
897 OpenSSL could do a one-byte buffer overread. The most likely result
898 would be an erroneous display of the certificate in text format.
900 This issue was reported to OpenSSL by the OSS-Fuzz project.
904 Changes between 1.1.0e and 1.1.0f [25 May 2017]
906 *) Have 'config' recognise 64-bit mingw and choose 'mingw64' as the target
907 platform rather than 'mingw'.
910 *) Remove the VMS-specific reimplementation of gmtime from crypto/o_times.c.
911 VMS C's RTL has a fully up to date gmtime() and gmtime_r() since V7.1,
912 which is the minimum version we support.
915 Changes between 1.1.0d and 1.1.0e [16 Feb 2017]
917 *) Encrypt-Then-Mac renegotiation crash
919 During a renegotiation handshake if the Encrypt-Then-Mac extension is
920 negotiated where it was not in the original handshake (or vice-versa) then
921 this can cause OpenSSL to crash (dependant on ciphersuite). Both clients
922 and servers are affected.
924 This issue was reported to OpenSSL by Joe Orton (Red Hat).
928 Changes between 1.1.0c and 1.1.0d [26 Jan 2017]
930 *) Truncated packet could crash via OOB read
932 If one side of an SSL/TLS path is running on a 32-bit host and a specific
933 cipher is being used, then a truncated packet can cause that host to
934 perform an out-of-bounds read, usually resulting in a crash.
936 This issue was reported to OpenSSL by Robert Święcki of Google.
940 *) Bad (EC)DHE parameters cause a client crash
942 If a malicious server supplies bad parameters for a DHE or ECDHE key
943 exchange then this can result in the client attempting to dereference a
944 NULL pointer leading to a client crash. This could be exploited in a Denial
947 This issue was reported to OpenSSL by Guido Vranken.
951 *) BN_mod_exp may produce incorrect results on x86_64
953 There is a carry propagating bug in the x86_64 Montgomery squaring
954 procedure. No EC algorithms are affected. Analysis suggests that attacks
955 against RSA and DSA as a result of this defect would be very difficult to
956 perform and are not believed likely. Attacks against DH are considered just
957 feasible (although very difficult) because most of the work necessary to
958 deduce information about a private key may be performed offline. The amount
959 of resources required for such an attack would be very significant and
960 likely only accessible to a limited number of attackers. An attacker would
961 additionally need online access to an unpatched system using the target
962 private key in a scenario with persistent DH parameters and a private
963 key that is shared between multiple clients. For example this can occur by
964 default in OpenSSL DHE based SSL/TLS ciphersuites. Note: This issue is very
965 similar to CVE-2015-3193 but must be treated as a separate problem.
967 This issue was reported to OpenSSL by the OSS-Fuzz project.
971 Changes between 1.1.0b and 1.1.0c [10 Nov 2016]
973 *) ChaCha20/Poly1305 heap-buffer-overflow
975 TLS connections using *-CHACHA20-POLY1305 ciphersuites are susceptible to
976 a DoS attack by corrupting larger payloads. This can result in an OpenSSL
977 crash. This issue is not considered to be exploitable beyond a DoS.
979 This issue was reported to OpenSSL by Robert Święcki (Google Security Team)
983 *) CMS Null dereference
985 Applications parsing invalid CMS structures can crash with a NULL pointer
986 dereference. This is caused by a bug in the handling of the ASN.1 CHOICE
987 type in OpenSSL 1.1.0 which can result in a NULL value being passed to the
988 structure callback if an attempt is made to free certain invalid encodings.
989 Only CHOICE structures using a callback which do not handle NULL value are
992 This issue was reported to OpenSSL by Tyler Nighswander of ForAllSecure.
996 *) Montgomery multiplication may produce incorrect results
998 There is a carry propagating bug in the Broadwell-specific Montgomery
999 multiplication procedure that handles input lengths divisible by, but
1000 longer than 256 bits. Analysis suggests that attacks against RSA, DSA
1001 and DH private keys are impossible. This is because the subroutine in
1002 question is not used in operations with the private key itself and an input
1003 of the attacker's direct choice. Otherwise the bug can manifest itself as
1004 transient authentication and key negotiation failures or reproducible
1005 erroneous outcome of public-key operations with specially crafted input.
1006 Among EC algorithms only Brainpool P-512 curves are affected and one
1007 presumably can attack ECDH key negotiation. Impact was not analyzed in
1008 detail, because pre-requisites for attack are considered unlikely. Namely
1009 multiple clients have to choose the curve in question and the server has to
1010 share the private key among them, neither of which is default behaviour.
1011 Even then only clients that chose the curve will be affected.
1013 This issue was publicly reported as transient failures and was not
1014 initially recognized as a security issue. Thanks to Richard Morgan for
1015 providing reproducible case.
1019 *) Removed automatic addition of RPATH in shared libraries and executables,
1020 as this was a remainder from OpenSSL 1.0.x and isn't needed any more.
1023 Changes between 1.1.0a and 1.1.0b [26 Sep 2016]
1025 *) Fix Use After Free for large message sizes
1027 The patch applied to address CVE-2016-6307 resulted in an issue where if a
1028 message larger than approx 16k is received then the underlying buffer to
1029 store the incoming message is reallocated and moved. Unfortunately a
1030 dangling pointer to the old location is left which results in an attempt to
1031 write to the previously freed location. This is likely to result in a
1032 crash, however it could potentially lead to execution of arbitrary code.
1034 This issue only affects OpenSSL 1.1.0a.
1036 This issue was reported to OpenSSL by Robert Święcki.
1040 Changes between 1.1.0 and 1.1.0a [22 Sep 2016]
1042 *) OCSP Status Request extension unbounded memory growth
1044 A malicious client can send an excessively large OCSP Status Request
1045 extension. If that client continually requests renegotiation, sending a
1046 large OCSP Status Request extension each time, then there will be unbounded
1047 memory growth on the server. This will eventually lead to a Denial Of
1048 Service attack through memory exhaustion. Servers with a default
1049 configuration are vulnerable even if they do not support OCSP. Builds using
1050 the "no-ocsp" build time option are not affected.
1052 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1056 *) SSL_peek() hang on empty record
1058 OpenSSL 1.1.0 SSL/TLS will hang during a call to SSL_peek() if the peer
1059 sends an empty record. This could be exploited by a malicious peer in a
1060 Denial Of Service attack.
1062 This issue was reported to OpenSSL by Alex Gaynor.
1066 *) Excessive allocation of memory in tls_get_message_header() and
1067 dtls1_preprocess_fragment()
1069 A (D)TLS message includes 3 bytes for its length in the header for the
1070 message. This would allow for messages up to 16Mb in length. Messages of
1071 this length are excessive and OpenSSL includes a check to ensure that a
1072 peer is sending reasonably sized messages in order to avoid too much memory
1073 being consumed to service a connection. A flaw in the logic of version
1074 1.1.0 means that memory for the message is allocated too early, prior to
1075 the excessive message length check. Due to way memory is allocated in
1076 OpenSSL this could mean an attacker could force up to 21Mb to be allocated
1077 to service a connection. This could lead to a Denial of Service through
1078 memory exhaustion. However, the excessive message length check still takes
1079 place, and this would cause the connection to immediately fail. Assuming
1080 that the application calls SSL_free() on the failed connection in a timely
1081 manner then the 21Mb of allocated memory will then be immediately freed
1082 again. Therefore the excessive memory allocation will be transitory in
1083 nature. This then means that there is only a security impact if:
1085 1) The application does not call SSL_free() in a timely manner in the event
1086 that the connection fails
1088 2) The application is working in a constrained environment where there is
1089 very little free memory
1091 3) The attacker initiates multiple connection attempts such that there are
1092 multiple connections in a state where memory has been allocated for the
1093 connection; SSL_free() has not yet been called; and there is insufficient
1094 memory to service the multiple requests.
1096 Except in the instance of (1) above any Denial Of Service is likely to be
1097 transitory because as soon as the connection fails the memory is
1098 subsequently freed again in the SSL_free() call. However there is an
1099 increased risk during this period of application crashes due to the lack of
1100 memory - which would then mean a more serious Denial of Service.
1102 This issue was reported to OpenSSL by Shi Lei (Gear Team, Qihoo 360 Inc.)
1103 (CVE-2016-6307 and CVE-2016-6308)
1106 *) solaris-x86-cc, i.e. 32-bit configuration with vendor compiler,
1107 had to be removed. Primary reason is that vendor assembler can't
1108 assemble our modules with -KPIC flag. As result it, assembly
1109 support, was not even available as option. But its lack means
1110 lack of side-channel resistant code, which is incompatible with
1111 security by todays standards. Fortunately gcc is readily available
1112 prepackaged option, which we firmly point at...
1115 Changes between 1.0.2h and 1.1.0 [25 Aug 2016]
1117 *) Windows command-line tool supports UTF-8 opt-in option for arguments
1118 and console input. Setting OPENSSL_WIN32_UTF8 environment variable
1119 (to any value) allows Windows user to access PKCS#12 file generated
1120 with Windows CryptoAPI and protected with non-ASCII password, as well
1121 as files generated under UTF-8 locale on Linux also protected with
1125 *) To mitigate the SWEET32 attack (CVE-2016-2183), 3DES cipher suites
1126 have been disabled by default and removed from DEFAULT, just like RC4.
1127 See the RC4 item below to re-enable both.
1130 *) The method for finding the storage location for the Windows RAND seed file
1131 has changed. First we check %RANDFILE%. If that is not set then we check
1132 the directories %HOME%, %USERPROFILE% and %SYSTEMROOT% in that order. If
1133 all else fails we fall back to C:\.
1136 *) The EVP_EncryptUpdate() function has had its return type changed from void
1137 to int. A return of 0 indicates and error while a return of 1 indicates
1141 *) The flags RSA_FLAG_NO_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME and
1142 DH_FLAG_NO_EXP_CONSTTIME which previously provided the ability to switch
1143 off the constant time implementation for RSA, DSA and DH have been made
1144 no-ops and deprecated.
1147 *) Windows RAND implementation was simplified to only get entropy by
1148 calling CryptGenRandom(). Various other RAND-related tickets
1150 [Joseph Wylie Yandle, Rich Salz]
1152 *) The stack and lhash API's were renamed to start with OPENSSL_SK_
1153 and OPENSSL_LH_, respectively. The old names are available
1154 with API compatibility. They new names are now completely documented.
1157 *) Unify TYPE_up_ref(obj) methods signature.
1158 SSL_CTX_up_ref(), SSL_up_ref(), X509_up_ref(), EVP_PKEY_up_ref(),
1159 X509_CRL_up_ref(), X509_OBJECT_up_ref_count() methods are now returning an
1160 int (instead of void) like all others TYPE_up_ref() methods.
1161 So now these methods also check the return value of CRYPTO_atomic_add(),
1162 and the validity of object reference counter.
1163 [fdasilvayy@gmail.com]
1165 *) With Windows Visual Studio builds, the .pdb files are installed
1166 alongside the installed libraries and executables. For a static
1167 library installation, ossl_static.pdb is the associate compiler
1168 generated .pdb file to be used when linking programs.
1171 *) Remove openssl.spec. Packaging files belong with the packagers.
1174 *) Automatic Darwin/OSX configuration has had a refresh, it will now
1175 recognise x86_64 architectures automatically. You can still decide
1176 to build for a different bitness with the environment variable
1177 KERNEL_BITS (can be 32 or 64), for example:
1179 KERNEL_BITS=32 ./config
1183 *) Change default algorithms in pkcs8 utility to use PKCS#5 v2.0,
1184 256 bit AES and HMAC with SHA256.
1187 *) Remove support for MIPS o32 ABI on IRIX (and IRIX only).
1190 *) Triple-DES ciphers have been moved from HIGH to MEDIUM.
1193 *) To enable users to have their own config files and build file templates,
1194 Configure looks in the directory indicated by the environment variable
1195 OPENSSL_LOCAL_CONFIG_DIR as well as the in-source Configurations/
1196 directory. On VMS, OPENSSL_LOCAL_CONFIG_DIR is expected to be a logical
1197 name and is used as is.
1200 *) The following datatypes were made opaque: X509_OBJECT, X509_STORE_CTX,
1201 X509_STORE, X509_LOOKUP, and X509_LOOKUP_METHOD. The unused type
1202 X509_CERT_FILE_CTX was removed.
1205 *) "shared" builds are now the default. To create only static libraries use
1206 the "no-shared" Configure option.
1209 *) Remove the no-aes, no-hmac, no-rsa, no-sha and no-md5 Configure options.
1210 All of these option have not worked for some while and are fundamental
1214 *) Make various cleanup routines no-ops and mark them as deprecated. Most
1215 global cleanup functions are no longer required because they are handled
1216 via auto-deinit (see OPENSSL_init_crypto and OPENSSL_init_ssl man pages).
1217 Explicitly de-initing can cause problems (e.g. where a library that uses
1218 OpenSSL de-inits, but an application is still using it). The affected
1219 functions are CONF_modules_free(), ENGINE_cleanup(), OBJ_cleanup(),
1220 EVP_cleanup(), BIO_sock_cleanup(), CRYPTO_cleanup_all_ex_data(),
1221 RAND_cleanup(), SSL_COMP_free_compression_methods(), ERR_free_strings() and
1222 COMP_zlib_cleanup().
1225 *) --strict-warnings no longer enables runtime debugging options
1226 such as REF_DEBUG. Instead, debug options are automatically
1227 enabled with '--debug' builds.
1228 [Andy Polyakov, Emilia Käsper]
1230 *) Made DH and DH_METHOD opaque. The structures for managing DH objects
1231 have been moved out of the public header files. New functions for managing
1232 these have been added.
1235 *) Made RSA and RSA_METHOD opaque. The structures for managing RSA
1236 objects have been moved out of the public header files. New
1237 functions for managing these have been added.
1240 *) Made DSA and DSA_METHOD opaque. The structures for managing DSA objects
1241 have been moved out of the public header files. New functions for managing
1242 these have been added.
1245 *) Made BIO and BIO_METHOD opaque. The structures for managing BIOs have been
1246 moved out of the public header files. New functions for managing these
1250 *) Removed no-rijndael as a config option. Rijndael is an old name for AES.
1253 *) Removed the mk1mf build scripts.
1256 *) Headers are now wrapped, if necessary, with OPENSSL_NO_xxx, so
1257 it is always safe to #include a header now.
1260 *) Removed the aged BC-32 config and all its supporting scripts
1263 *) Removed support for Ultrix, Netware, and OS/2.
1266 *) Add support for HKDF.
1267 [Alessandro Ghedini]
1269 *) Add support for blake2b and blake2s
1272 *) Added support for "pipelining". Ciphers that have the
1273 EVP_CIPH_FLAG_PIPELINE flag set have a capability to process multiple
1274 encryptions/decryptions simultaneously. There are currently no built-in
1275 ciphers with this property but the expectation is that engines will be able
1276 to offer it to significantly improve throughput. Support has been extended
1277 into libssl so that multiple records for a single connection can be
1278 processed in one go (for >=TLS 1.1).
1281 *) Added the AFALG engine. This is an async capable engine which is able to
1282 offload work to the Linux kernel. In this initial version it only supports
1283 AES128-CBC. The kernel must be version 4.1.0 or greater.
1286 *) OpenSSL now uses a new threading API. It is no longer necessary to
1287 set locking callbacks to use OpenSSL in a multi-threaded environment. There
1288 are two supported threading models: pthreads and windows threads. It is
1289 also possible to configure OpenSSL at compile time for "no-threads". The
1290 old threading API should no longer be used. The functions have been
1291 replaced with "no-op" compatibility macros.
1292 [Alessandro Ghedini, Matt Caswell]
1294 *) Modify behavior of ALPN to invoke callback after SNI/servername
1295 callback, such that updates to the SSL_CTX affect ALPN.
1298 *) Add SSL_CIPHER queries for authentication and key-exchange.
1301 *) Changes to the DEFAULT cipherlist:
1302 - Prefer (EC)DHE handshakes over plain RSA.
1303 - Prefer AEAD ciphers over legacy ciphers.
1304 - Prefer ECDSA over RSA when both certificates are available.
1305 - Prefer TLSv1.2 ciphers/PRF.
1306 - Remove DSS, SEED, IDEA, CAMELLIA, and AES-CCM from the
1310 *) Change the ECC default curve list to be this, in order: x25519,
1311 secp256r1, secp521r1, secp384r1.
1314 *) RC4 based libssl ciphersuites are now classed as "weak" ciphers and are
1315 disabled by default. They can be re-enabled using the
1316 enable-weak-ssl-ciphers option to Configure.
1319 *) If the server has ALPN configured, but supports no protocols that the
1320 client advertises, send a fatal "no_application_protocol" alert.
1321 This behaviour is SHALL in RFC 7301, though it isn't universally
1322 implemented by other servers.
1325 *) Add X25519 support.
1326 Add ASN.1 and EVP_PKEY methods for X25519. This includes support
1327 for public and private key encoding using the format documented in
1328 draft-ietf-curdle-pkix-02. The corresponding EVP_PKEY method supports
1329 key generation and key derivation.
1331 TLS support complies with draft-ietf-tls-rfc4492bis-08 and uses
1335 *) Deprecate SRP_VBASE_get_by_user.
1336 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
1337 In order to fix an unavoidable memory leak (CVE-2016-0798),
1338 SRP_VBASE_get_by_user was changed to ignore the "fake user" SRP
1339 seed, even if the seed is configured.
1341 Users should use SRP_VBASE_get1_by_user instead. Note that in
1342 SRP_VBASE_get1_by_user, caller must free the returned value. Note
1343 also that even though configuring the SRP seed attempts to hide
1344 invalid usernames by continuing the handshake with fake
1345 credentials, this behaviour is not constant time and no strong
1346 guarantees are made that the handshake is indistinguishable from
1347 that of a valid user.
1350 *) Configuration change; it's now possible to build dynamic engines
1351 without having to build shared libraries and vice versa. This
1352 only applies to the engines in engines/, those in crypto/engine/
1353 will always be built into libcrypto (i.e. "static").
1355 Building dynamic engines is enabled by default; to disable, use
1356 the configuration option "disable-dynamic-engine".
1358 The only requirements for building dynamic engines are the
1359 presence of the DSO module and building with position independent
1360 code, so they will also automatically be disabled if configuring
1361 with "disable-dso" or "disable-pic".
1363 The macros OPENSSL_NO_STATIC_ENGINE and OPENSSL_NO_DYNAMIC_ENGINE
1364 are also taken away from openssl/opensslconf.h, as they are
1368 *) Configuration change; if there is a known flag to compile
1369 position independent code, it will always be applied on the
1370 libcrypto and libssl object files, and never on the application
1371 object files. This means other libraries that use routines from
1372 libcrypto / libssl can be made into shared libraries regardless
1373 of how OpenSSL was configured.
1375 If this isn't desirable, the configuration options "disable-pic"
1376 or "no-pic" can be used to disable the use of PIC. This will
1377 also disable building shared libraries and dynamic engines.
1380 *) Removed JPAKE code. It was experimental and has no wide use.
1383 *) The INSTALL_PREFIX Makefile variable has been renamed to
1384 DESTDIR. That makes for less confusion on what this variable
1385 is for. Also, the configuration option --install_prefix is
1389 *) Heartbeat for TLS has been removed and is disabled by default
1390 for DTLS; configure with enable-heartbeats. Code that uses the
1391 old #define's might need to be updated.
1392 [Emilia Käsper, Rich Salz]
1394 *) Rename REF_CHECK to REF_DEBUG.
1397 *) New "unified" build system
1399 The "unified" build system is aimed to be a common system for all
1400 platforms we support. With it comes new support for VMS.
1402 This system builds supports building in a different directory tree
1403 than the source tree. It produces one Makefile (for unix family
1404 or lookalikes), or one descrip.mms (for VMS).
1406 The source of information to make the Makefile / descrip.mms is
1407 small files called 'build.info', holding the necessary
1408 information for each directory with source to compile, and a
1409 template in Configurations, like unix-Makefile.tmpl or
1412 With this change, the library names were also renamed on Windows
1413 and on VMS. They now have names that are closer to the standard
1414 on Unix, and include the major version number, and in certain
1415 cases, the architecture they are built for. See "Notes on shared
1416 libraries" in INSTALL.
1418 We rely heavily on the perl module Text::Template.
1421 *) Added support for auto-initialisation and de-initialisation of the library.
1422 OpenSSL no longer requires explicit init or deinit routines to be called,
1423 except in certain circumstances. See the OPENSSL_init_crypto() and
1424 OPENSSL_init_ssl() man pages for further information.
1427 *) The arguments to the DTLSv1_listen function have changed. Specifically the
1428 "peer" argument is now expected to be a BIO_ADDR object.
1430 *) Rewrite of BIO networking library. The BIO library lacked consistent
1431 support of IPv6, and adding it required some more extensive
1432 modifications. This introduces the BIO_ADDR and BIO_ADDRINFO types,
1433 which hold all types of addresses and chains of address information.
1434 It also introduces a new API, with functions like BIO_socket,
1435 BIO_connect, BIO_listen, BIO_lookup and a rewrite of BIO_accept.
1436 The source/sink BIOs BIO_s_connect, BIO_s_accept and BIO_s_datagram
1437 have been adapted accordingly.
1440 *) RSA_padding_check_PKCS1_type_1 now accepts inputs with and without
1444 *) CRIME protection: disable compression by default, even if OpenSSL is
1445 compiled with zlib enabled. Applications can still enable compression
1446 by calling SSL_CTX_clear_options(ctx, SSL_OP_NO_COMPRESSION), or by
1447 using the SSL_CONF library to configure compression.
1450 *) The signature of the session callback configured with
1451 SSL_CTX_sess_set_get_cb was changed. The read-only input buffer
1452 was explicitly marked as 'const unsigned char*' instead of
1456 *) Always DPURIFY. Remove the use of uninitialized memory in the
1457 RNG, and other conditional uses of DPURIFY. This makes -DPURIFY a no-op.
1460 *) Removed many obsolete configuration items, including
1461 DES_PTR, DES_RISC1, DES_RISC2, DES_INT
1462 MD2_CHAR, MD2_INT, MD2_LONG
1464 IDEA_SHORT, IDEA_LONG
1465 RC2_SHORT, RC2_LONG, RC4_LONG, RC4_CHUNK, RC4_INDEX
1466 [Rich Salz, with advice from Andy Polyakov]
1468 *) Many BN internals have been moved to an internal header file.
1469 [Rich Salz with help from Andy Polyakov]
1471 *) Configuration and writing out the results from it has changed.
1472 Files such as Makefile include/openssl/opensslconf.h and are now
1473 produced through general templates, such as Makefile.in and
1474 crypto/opensslconf.h.in and some help from the perl module
1477 Also, the center of configuration information is no longer
1478 Makefile. Instead, Configure produces a perl module in
1479 configdata.pm which holds most of the config data (in the hash
1480 table %config), the target data that comes from the target
1481 configuration in one of the Configurations/*.conf files (in
1485 *) To clarify their intended purposes, the Configure options
1486 --prefix and --openssldir change their semantics, and become more
1487 straightforward and less interdependent.
1489 --prefix shall be used exclusively to give the location INSTALLTOP
1490 where programs, scripts, libraries, include files and manuals are
1491 going to be installed. The default is now /usr/local.
1493 --openssldir shall be used exclusively to give the default
1494 location OPENSSLDIR where certificates, private keys, CRLs are
1495 managed. This is also where the default openssl.cnf gets
1497 If the directory given with this option is a relative path, the
1498 values of both the --prefix value and the --openssldir value will
1499 be combined to become OPENSSLDIR.
1500 The default for --openssldir is INSTALLTOP/ssl.
1502 Anyone who uses --openssldir to specify where OpenSSL is to be
1503 installed MUST change to use --prefix instead.
1506 *) The GOST engine was out of date and therefore it has been removed. An up
1507 to date GOST engine is now being maintained in an external repository.
1508 See: https://wiki.openssl.org/index.php/Binaries. Libssl still retains
1509 support for GOST ciphersuites (these are only activated if a GOST engine
1513 *) EGD is no longer supported by default; use enable-egd when
1515 [Ben Kaduk and Rich Salz]
1517 *) The distribution now has Makefile.in files, which are used to
1518 create Makefile's when Configure is run. *Configure must be run
1519 before trying to build now.*
1522 *) The return value for SSL_CIPHER_description() for error conditions
1526 *) Support for RFC6698/RFC7671 DANE TLSA peer authentication.
1528 Obtaining and performing DNSSEC validation of TLSA records is
1529 the application's responsibility. The application provides
1530 the TLSA records of its choice to OpenSSL, and these are then
1531 used to authenticate the peer.
1533 The TLSA records need not even come from DNS. They can, for
1534 example, be used to implement local end-entity certificate or
1535 trust-anchor "pinning", where the "pin" data takes the form
1536 of TLSA records, which can augment or replace verification
1537 based on the usual WebPKI public certification authorities.
1540 *) Revert default OPENSSL_NO_DEPRECATED setting. Instead OpenSSL
1541 continues to support deprecated interfaces in default builds.
1542 However, applications are strongly advised to compile their
1543 source files with -DOPENSSL_API_COMPAT=0x10100000L, which hides
1544 the declarations of all interfaces deprecated in 0.9.8, 1.0.0
1545 or the 1.1.0 releases.
1547 In environments in which all applications have been ported to
1548 not use any deprecated interfaces OpenSSL's Configure script
1549 should be used with the --api=1.1.0 option to entirely remove
1550 support for the deprecated features from the library and
1551 unconditionally disable them in the installed headers.
1552 Essentially the same effect can be achieved with the "no-deprecated"
1553 argument to Configure, except that this will always restrict
1554 the build to just the latest API, rather than a fixed API
1557 As applications are ported to future revisions of the API,
1558 they should update their compile-time OPENSSL_API_COMPAT define
1559 accordingly, but in most cases should be able to continue to
1560 compile with later releases.
1562 The OPENSSL_API_COMPAT versions for 1.0.0, and 0.9.8 are
1563 0x10000000L and 0x00908000L, respectively. However those
1564 versions did not support the OPENSSL_API_COMPAT feature, and
1565 so applications are not typically tested for explicit support
1566 of just the undeprecated features of either release.
1569 *) Add support for setting the minimum and maximum supported protocol.
1570 It can bet set via the SSL_set_min_proto_version() and
1571 SSL_set_max_proto_version(), or via the SSL_CONF's MinProtocol and
1572 MaxProtocol. It's recommended to use the new APIs to disable
1573 protocols instead of disabling individual protocols using
1574 SSL_set_options() or SSL_CONF's Protocol. This change also
1575 removes support for disabling TLS 1.2 in the OpenSSL TLS
1576 client at compile time by defining OPENSSL_NO_TLS1_2_CLIENT.
1579 *) Support for ChaCha20 and Poly1305 added to libcrypto and libssl.
1582 *) New EC_KEY_METHOD, this replaces the older ECDSA_METHOD and ECDH_METHOD
1583 and integrates ECDSA and ECDH functionality into EC. Implementations can
1584 now redirect key generation and no longer need to convert to or from
1587 Note: the ecdsa.h and ecdh.h headers are now no longer needed and just
1588 include the ec.h header file instead.
1591 *) Remove support for all 40 and 56 bit ciphers. This includes all the export
1592 ciphers who are no longer supported and drops support the ephemeral RSA key
1593 exchange. The LOW ciphers currently doesn't have any ciphers in it.
1596 *) Made EVP_MD_CTX, EVP_MD, EVP_CIPHER_CTX, EVP_CIPHER and HMAC_CTX
1597 opaque. For HMAC_CTX, the following constructors and destructors
1600 HMAC_CTX *HMAC_CTX_new(void);
1601 void HMAC_CTX_free(HMAC_CTX *ctx);
1603 For EVP_MD and EVP_CIPHER, complete APIs to create, fill and
1604 destroy such methods has been added. See EVP_MD_meth_new(3) and
1605 EVP_CIPHER_meth_new(3) for documentation.
1608 1) EVP_MD_CTX_cleanup(), EVP_CIPHER_CTX_cleanup() and
1609 HMAC_CTX_cleanup() were removed. HMAC_CTX_reset() and
1610 EVP_MD_CTX_reset() should be called instead to reinitialise
1611 an already created structure.
1612 2) For consistency with the majority of our object creators and
1613 destructors, EVP_MD_CTX_(create|destroy) were renamed to
1614 EVP_MD_CTX_(new|free). The old names are retained as macros
1615 for deprecated builds.
1618 *) Added ASYNC support. Libcrypto now includes the async sub-library to enable
1619 cryptographic operations to be performed asynchronously as long as an
1620 asynchronous capable engine is used. See the ASYNC_start_job() man page for
1621 further details. Libssl has also had this capability integrated with the
1622 introduction of the new mode SSL_MODE_ASYNC and associated error
1623 SSL_ERROR_WANT_ASYNC. See the SSL_CTX_set_mode() and SSL_get_error() man
1624 pages. This work was developed in partnership with Intel Corp.
1627 *) SSL_{CTX_}set_ecdh_auto() has been removed and ECDH is support is
1628 always enabled now. If you want to disable the support you should
1629 exclude it using the list of supported ciphers. This also means that the
1630 "-no_ecdhe" option has been removed from s_server.
1633 *) SSL_{CTX}_set_tmp_ecdh() which can set 1 EC curve now internally calls
1634 SSL_{CTX_}set1_curves() which can set a list.
1637 *) Remove support for SSL_{CTX_}set_tmp_ecdh_callback(). You should set the
1638 curve you want to support using SSL_{CTX_}set1_curves().
1641 *) State machine rewrite. The state machine code has been significantly
1642 refactored in order to remove much duplication of code and solve issues
1643 with the old code (see ssl/statem/README for further details). This change
1644 does have some associated API changes. Notably the SSL_state() function
1645 has been removed and replaced by SSL_get_state which now returns an
1646 "OSSL_HANDSHAKE_STATE" instead of an int. SSL_set_state() has been removed
1647 altogether. The previous handshake states defined in ssl.h and ssl3.h have
1651 *) All instances of the string "ssleay" in the public API were replaced
1652 with OpenSSL (case-matching; e.g., OPENSSL_VERSION for #define's)
1653 Some error codes related to internal RSA_eay API's were renamed.
1656 *) The demo files in crypto/threads were moved to demo/threads.
1659 *) Removed obsolete engines: 4758cca, aep, atalla, cswift, nuron, gmp,
1661 [Matt Caswell, Rich Salz]
1663 *) New ASN.1 embed macro.
1665 New ASN.1 macro ASN1_EMBED. This is the same as ASN1_SIMPLE except the
1666 structure is not allocated: it is part of the parent. That is instead of
1674 This reduces memory fragmentation and make it impossible to accidentally
1675 set a mandatory field to NULL.
1677 This currently only works for some fields specifically a SEQUENCE, CHOICE,
1678 or ASN1_STRING type which is part of a parent SEQUENCE. Since it is
1679 equivalent to ASN1_SIMPLE it cannot be tagged, OPTIONAL, SET OF or
1683 *) Remove EVP_CHECK_DES_KEY, a compile-time option that never compiled.
1686 *) Removed DES and RC4 ciphersuites from DEFAULT. Also removed RC2 although
1687 in 1.0.2 EXPORT was already removed and the only RC2 ciphersuite is also
1688 an EXPORT one. COMPLEMENTOFDEFAULT has been updated accordingly to add
1689 DES and RC4 ciphersuites.
1692 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
1693 This changes the decoding behaviour for some invalid messages,
1694 though the change is mostly in the more lenient direction, and
1695 legacy behaviour is preserved as much as possible.
1698 *) Fix no-stdio build.
1699 [ David Woodhouse <David.Woodhouse@intel.com> and also
1700 Ivan Nestlerode <ivan.nestlerode@sonos.com> ]
1702 *) New testing framework
1703 The testing framework has been largely rewritten and is now using
1704 perl and the perl modules Test::Harness and an extended variant of
1705 Test::More called OpenSSL::Test to do its work. All test scripts in
1706 test/ have been rewritten into test recipes, and all direct calls to
1707 executables in test/Makefile have become individual recipes using the
1708 simplified testing OpenSSL::Test::Simple.
1710 For documentation on our testing modules, do:
1712 perldoc test/testlib/OpenSSL/Test/Simple.pm
1713 perldoc test/testlib/OpenSSL/Test.pm
1717 *) Revamped memory debug; only -DCRYPTO_MDEBUG and -DCRYPTO_MDEBUG_ABORT
1718 are used; the latter aborts on memory leaks (usually checked on exit).
1719 Some undocumented "set malloc, etc., hooks" functions were removed
1720 and others were changed. All are now documented.
1723 *) In DSA_generate_parameters_ex, if the provided seed is too short,
1725 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
1727 *) Rewrite PSK to support ECDHE_PSK, DHE_PSK and RSA_PSK. Add ciphersuites
1728 from RFC4279, RFC4785, RFC5487, RFC5489.
1730 Thanks to Christian J. Dietrich and Giuseppe D'Angelo for the
1731 original RSA_PSK patch.
1734 *) Dropped support for the SSL3_FLAGS_DELAY_CLIENT_FINISHED flag. This SSLeay
1735 era flag was never set throughout the codebase (only read). Also removed
1736 SSL3_FLAGS_POP_BUFFER which was only used if
1737 SSL3_FLAGS_DELAY_CLIENT_FINISHED was also set.
1740 *) Changed the default name options in the "ca", "crl", "req" and "x509"
1741 to be "oneline" instead of "compat".
1744 *) Remove SSL_OP_TLS_BLOCK_PADDING_BUG. This is SSLeay legacy, we're
1745 not aware of clients that still exhibit this bug, and the workaround
1746 hasn't been working properly for a while.
1749 *) The return type of BIO_number_read() and BIO_number_written() as well as
1750 the corresponding num_read and num_write members in the BIO structure has
1751 changed from unsigned long to uint64_t. On platforms where an unsigned
1752 long is 32 bits (e.g. Windows) these counters could overflow if >4Gb is
1756 *) Given the pervasive nature of TLS extensions it is inadvisable to run
1757 OpenSSL without support for them. It also means that maintaining
1758 the OPENSSL_NO_TLSEXT option within the code is very invasive (and probably
1759 not well tested). Therefore the OPENSSL_NO_TLSEXT option has been removed.
1762 *) Removed support for the two export grade static DH ciphersuites
1763 EXP-DH-RSA-DES-CBC-SHA and EXP-DH-DSS-DES-CBC-SHA. These two ciphersuites
1764 were newly added (along with a number of other static DH ciphersuites) to
1765 1.0.2. However the two export ones have *never* worked since they were
1766 introduced. It seems strange in any case to be adding new export
1767 ciphersuites, and given "logjam" it also does not seem correct to fix them.
1770 *) Version negotiation has been rewritten. In particular SSLv23_method(),
1771 SSLv23_client_method() and SSLv23_server_method() have been deprecated,
1772 and turned into macros which simply call the new preferred function names
1773 TLS_method(), TLS_client_method() and TLS_server_method(). All new code
1774 should use the new names instead. Also as part of this change the ssl23.h
1775 header file has been removed.
1778 *) Support for Kerberos ciphersuites in TLS (RFC2712) has been removed. This
1779 code and the associated standard is no longer considered fit-for-purpose.
1782 *) RT2547 was closed. When generating a private key, try to make the
1783 output file readable only by the owner. This behavior change might
1784 be noticeable when interacting with other software.
1786 *) Documented all exdata functions. Added CRYPTO_free_ex_index.
1790 *) Added HTTP GET support to the ocsp command.
1793 *) Changed default digest for the dgst and enc commands from MD5 to
1797 *) RAND_pseudo_bytes has been deprecated. Users should use RAND_bytes instead.
1800 *) Added support for TLS extended master secret from
1801 draft-ietf-tls-session-hash-03.txt. Thanks for Alfredo Pironti for an
1802 initial patch which was a great help during development.
1805 *) All libssl internal structures have been removed from the public header
1806 files, and the OPENSSL_NO_SSL_INTERN option has been removed (since it is
1807 now redundant). Users should not attempt to access internal structures
1808 directly. Instead they should use the provided API functions.
1811 *) config has been changed so that by default OPENSSL_NO_DEPRECATED is used.
1812 Access to deprecated functions can be re-enabled by running config with
1813 "enable-deprecated". In addition applications wishing to use deprecated
1814 functions must define OPENSSL_USE_DEPRECATED. Note that this new behaviour
1815 will, by default, disable some transitive includes that previously existed
1816 in the header files (e.g. ec.h will no longer, by default, include bn.h)
1819 *) Added support for OCB mode. OpenSSL has been granted a patent license
1820 compatible with the OpenSSL license for use of OCB. Details are available
1821 at https://www.openssl.org/source/OCB-patent-grant-OpenSSL.pdf. Support
1822 for OCB can be removed by calling config with no-ocb.
1825 *) SSLv2 support has been removed. It still supports receiving a SSLv2
1826 compatible client hello.
1829 *) Increased the minimal RSA keysize from 256 to 512 bits [Rich Salz],
1830 done while fixing the error code for the key-too-small case.
1831 [Annie Yousar <a.yousar@informatik.hu-berlin.de>]
1833 *) CA.sh has been removed; use CA.pl instead.
1836 *) Removed old DES API.
1839 *) Remove various unsupported platforms:
1845 Sinix/ReliantUNIX RM400
1850 16-bit platforms such as WIN16
1853 *) Clean up OPENSSL_NO_xxx #define's
1854 Use setbuf() and remove OPENSSL_NO_SETVBUF_IONBF
1855 Rename OPENSSL_SYSNAME_xxx to OPENSSL_SYS_xxx
1856 OPENSSL_NO_EC{DH,DSA} merged into OPENSSL_NO_EC
1857 OPENSSL_NO_RIPEMD160, OPENSSL_NO_RIPEMD merged into OPENSSL_NO_RMD160
1858 OPENSSL_NO_FP_API merged into OPENSSL_NO_STDIO
1859 Remove OPENSSL_NO_BIO OPENSSL_NO_BUFFER OPENSSL_NO_CHAIN_VERIFY
1860 OPENSSL_NO_EVP OPENSSL_NO_FIPS_ERR OPENSSL_NO_HASH_COMP
1861 OPENSSL_NO_LHASH OPENSSL_NO_OBJECT OPENSSL_NO_SPEED OPENSSL_NO_STACK
1862 OPENSSL_NO_X509 OPENSSL_NO_X509_VERIFY
1863 Remove MS_STATIC; it's a relic from platforms <32 bits.
1866 *) Cleaned up dead code
1867 Remove all but one '#ifdef undef' which is to be looked at.
1870 *) Clean up calling of xxx_free routines.
1871 Just like free(), fix most of the xxx_free routines to accept
1872 NULL. Remove the non-null checks from callers. Save much code.
1875 *) Add secure heap for storage of private keys (when possible).
1876 Add BIO_s_secmem(), CBIGNUM, etc.
1877 Contributed by Akamai Technologies under our Corporate CLA.
1880 *) Experimental support for a new, fast, unbiased prime candidate generator,
1881 bn_probable_prime_dh_coprime(). Not currently used by any prime generator.
1882 [Felix Laurie von Massenbach <felix@erbridge.co.uk>]
1884 *) New output format NSS in the sess_id command line tool. This allows
1885 exporting the session id and the master key in NSS keylog format.
1886 [Martin Kaiser <martin@kaiser.cx>]
1888 *) Harmonize version and its documentation. -f flag is used to display
1890 [mancha <mancha1@zoho.com>]
1892 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
1893 in i2d_ECPrivateKey. Thanks to Ted Unangst for feedback on this issue.
1894 [mancha <mancha1@zoho.com>]
1896 *) Fix some double frees. These are not thought to be exploitable.
1897 [mancha <mancha1@zoho.com>]
1899 *) A missing bounds check in the handling of the TLS heartbeat extension
1900 can be used to reveal up to 64k of memory to a connected client or
1903 Thanks for Neel Mehta of Google Security for discovering this bug and to
1904 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
1905 preparing the fix (CVE-2014-0160)
1906 [Adam Langley, Bodo Moeller]
1908 *) Fix for the attack described in the paper "Recovering OpenSSL
1909 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
1910 by Yuval Yarom and Naomi Benger. Details can be obtained from:
1911 http://eprint.iacr.org/2014/140
1913 Thanks to Yuval Yarom and Naomi Benger for discovering this
1914 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
1915 [Yuval Yarom and Naomi Benger]
1917 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
1918 this fixes a limitation in previous versions of OpenSSL.
1921 *) Experimental encrypt-then-mac support.
1923 Experimental support for encrypt then mac from
1924 draft-gutmann-tls-encrypt-then-mac-02.txt
1926 To enable it set the appropriate extension number (0x42 for the test
1927 server) using e.g. -DTLSEXT_TYPE_encrypt_then_mac=0x42
1929 For non-compliant peers (i.e. just about everything) this should have no
1932 WARNING: EXPERIMENTAL, SUBJECT TO CHANGE.
1936 *) Add EVP support for key wrapping algorithms, to avoid problems with
1937 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
1938 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
1939 algorithms and include tests cases.
1942 *) Extend CMS code to support RSA-PSS signatures and RSA-OAEP for
1946 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
1947 MGF1 digest and OAEP label.
1950 *) Make openssl verify return errors.
1951 [Chris Palmer <palmer@google.com> and Ben Laurie]
1953 *) New function ASN1_TIME_diff to calculate the difference between two
1954 ASN1_TIME structures or one structure and the current time.
1957 *) Update fips_test_suite to support multiple command line options. New
1958 test to induce all self test errors in sequence and check expected
1962 *) Add FIPS_{rsa,dsa,ecdsa}_{sign,verify} functions which digest and
1963 sign or verify all in one operation.
1966 *) Add fips_algvs: a multicall fips utility incorporating all the algorithm
1967 test programs and fips_test_suite. Includes functionality to parse
1968 the minimal script output of fipsalgest.pl directly.
1971 *) Add authorisation parameter to FIPS_module_mode_set().
1974 *) Add FIPS selftest for ECDH algorithm using P-224 and B-233 curves.
1977 *) Use separate DRBG fields for internal and external flags. New function
1978 FIPS_drbg_health_check() to perform on demand health checking. Add
1979 generation tests to fips_test_suite with reduced health check interval to
1980 demonstrate periodic health checking. Add "nodh" option to
1981 fips_test_suite to skip very slow DH test.
1984 *) New function FIPS_get_cipherbynid() to lookup FIPS supported ciphers
1988 *) More extensive health check for DRBG checking many more failure modes.
1989 New function FIPS_selftest_drbg_all() to handle every possible DRBG
1990 combination: call this in fips_test_suite.
1993 *) Add support for canonical generation of DSA parameter 'g'. See
1996 *) Add support for HMAC DRBG from SP800-90. Update DRBG algorithm test and
1997 POST to handle HMAC cases.
2000 *) Add functions FIPS_module_version() and FIPS_module_version_text()
2001 to return numerical and string versions of the FIPS module number.
2004 *) Rename FIPS_mode_set and FIPS_mode to FIPS_module_mode_set and
2005 FIPS_module_mode. FIPS_mode and FIPS_mode_set will be implemented
2006 outside the validated module in the FIPS capable OpenSSL.
2009 *) Minor change to DRBG entropy callback semantics. In some cases
2010 there is no multiple of the block length between min_len and
2011 max_len. Allow the callback to return more than max_len bytes
2012 of entropy but discard any extra: it is the callback's responsibility
2013 to ensure that the extra data discarded does not impact the
2014 requested amount of entropy.
2017 *) Add PRNG security strength checks to RSA, DSA and ECDSA using
2018 information in FIPS186-3, SP800-57 and SP800-131A.
2021 *) CCM support via EVP. Interface is very similar to GCM case except we
2022 must supply all data in one chunk (i.e. no update, final) and the
2023 message length must be supplied if AAD is used. Add algorithm test
2027 *) Initial version of POST overhaul. Add POST callback to allow the status
2028 of POST to be monitored and/or failures induced. Modify fips_test_suite
2029 to use callback. Always run all selftests even if one fails.
2032 *) XTS support including algorithm test driver in the fips_gcmtest program.
2033 Note: this does increase the maximum key length from 32 to 64 bytes but
2034 there should be no binary compatibility issues as existing applications
2035 will never use XTS mode.
2038 *) Extensive reorganisation of FIPS PRNG behaviour. Remove all dependencies
2039 to OpenSSL RAND code and replace with a tiny FIPS RAND API which also
2040 performs algorithm blocking for unapproved PRNG types. Also do not
2041 set PRNG type in FIPS_mode_set(): leave this to the application.
2042 Add default OpenSSL DRBG handling: sets up FIPS PRNG and seeds with
2043 the standard OpenSSL PRNG: set additional data to a date time vector.
2046 *) Rename old X9.31 PRNG functions of the form FIPS_rand* to FIPS_x931*.
2047 This shouldn't present any incompatibility problems because applications
2048 shouldn't be using these directly and any that are will need to rethink
2049 anyway as the X9.31 PRNG is now deprecated by FIPS 140-2
2052 *) Extensive self tests and health checking required by SP800-90 DRBG.
2053 Remove strength parameter from FIPS_drbg_instantiate and always
2054 instantiate at maximum supported strength.
2057 *) Add ECDH code to fips module and fips_ecdhvs for primitives only testing.
2060 *) New algorithm test program fips_dhvs to handle DH primitives only testing.
2063 *) New function DH_compute_key_padded() to compute a DH key and pad with
2064 leading zeroes if needed: this complies with SP800-56A et al.
2067 *) Initial implementation of SP800-90 DRBGs for Hash and CTR. Not used by
2068 anything, incomplete, subject to change and largely untested at present.
2071 *) Modify fipscanisteronly build option to only build the necessary object
2072 files by filtering FIPS_EX_OBJ through a perl script in crypto/Makefile.
2075 *) Add experimental option FIPSSYMS to give all symbols in
2076 fipscanister.o and FIPS or fips prefix. This will avoid
2077 conflicts with future versions of OpenSSL. Add perl script
2078 util/fipsas.pl to preprocess assembly language source files
2079 and rename any affected symbols.
2082 *) Add selftest checks and algorithm block of non-fips algorithms in
2083 FIPS mode. Remove DES2 from selftests.
2086 *) Add ECDSA code to fips module. Add tiny fips_ecdsa_check to just
2087 return internal method without any ENGINE dependencies. Add new
2088 tiny fips sign and verify functions.
2091 *) New build option no-ec2m to disable characteristic 2 code.
2094 *) New build option "fipscanisteronly". This only builds fipscanister.o
2095 and (currently) associated fips utilities. Uses the file Makefile.fips
2096 instead of Makefile.org as the prototype.
2099 *) Add some FIPS mode restrictions to GCM. Add internal IV generator.
2100 Update fips_gcmtest to use IV generator.
2103 *) Initial, experimental EVP support for AES-GCM. AAD can be input by
2104 setting output buffer to NULL. The *Final function must be
2105 called although it will not retrieve any additional data. The tag
2106 can be set or retrieved with a ctrl. The IV length is by default 12
2107 bytes (96 bits) but can be set to an alternative value. If the IV
2108 length exceeds the maximum IV length (currently 16 bytes) it cannot be
2112 *) New flag in ciphers: EVP_CIPH_FLAG_CUSTOM_CIPHER. This means the
2113 underlying do_cipher function handles all cipher semantics itself
2114 including padding and finalisation. This is useful if (for example)
2115 an ENGINE cipher handles block padding itself. The behaviour of
2116 do_cipher is subtly changed if this flag is set: the return value
2117 is the number of characters written to the output buffer (zero is
2118 no longer an error code) or a negative error code. Also if the
2119 input buffer is NULL and length 0 finalisation should be performed.
2122 *) If a candidate issuer certificate is already part of the constructed
2123 path ignore it: new debug notification X509_V_ERR_PATH_LOOP for this case.
2126 *) Improve forward-security support: add functions
2128 void SSL_CTX_set_not_resumable_session_callback(SSL_CTX *ctx, int (*cb)(SSL *ssl, int is_forward_secure))
2129 void SSL_set_not_resumable_session_callback(SSL *ssl, int (*cb)(SSL *ssl, int is_forward_secure))
2131 for use by SSL/TLS servers; the callback function will be called whenever a
2132 new session is created, and gets to decide whether the session may be
2133 cached to make it resumable (return 0) or not (return 1). (As by the
2134 SSL/TLS protocol specifications, the session_id sent by the server will be
2135 empty to indicate that the session is not resumable; also, the server will
2136 not generate RFC 4507 (RFC 5077) session tickets.)
2138 A simple reasonable callback implementation is to return is_forward_secure.
2139 This parameter will be set to 1 or 0 depending on the ciphersuite selected
2140 by the SSL/TLS server library, indicating whether it can provide forward
2142 [Emilia Käsper <emilia.kasper@esat.kuleuven.be> (Google)]
2144 *) New -verify_name option in command line utilities to set verification
2148 *) Initial CMAC implementation. WARNING: EXPERIMENTAL, API MAY CHANGE.
2149 Add CMAC pkey methods.
2152 *) Experimental renegotiation in s_server -www mode. If the client
2153 browses /reneg connection is renegotiated. If /renegcert it is
2154 renegotiated requesting a certificate.
2157 *) Add an "external" session cache for debugging purposes to s_server. This
2158 should help trace issues which normally are only apparent in deployed
2159 multi-process servers.
2162 *) Extensive audit of libcrypto with DEBUG_UNUSED. Fix many cases where
2163 return value is ignored. NB. The functions RAND_add(), RAND_seed(),
2164 BIO_set_cipher() and some obscure PEM functions were changed so they
2165 can now return an error. The RAND changes required a change to the
2166 RAND_METHOD structure.
2169 *) New macro __owur for "OpenSSL Warn Unused Result". This makes use of
2170 a gcc attribute to warn if the result of a function is ignored. This
2171 is enable if DEBUG_UNUSED is set. Add to several functions in evp.h
2172 whose return value is often ignored.
2175 *) New -noct, -requestct, -requirect and -ctlogfile options for s_client.
2176 These allow SCTs (signed certificate timestamps) to be requested and
2177 validated when establishing a connection.
2178 [Rob Percival <robpercival@google.com>]
2180 Changes between 1.0.2g and 1.0.2h [3 May 2016]
2182 *) Prevent padding oracle in AES-NI CBC MAC check
2184 A MITM attacker can use a padding oracle attack to decrypt traffic
2185 when the connection uses an AES CBC cipher and the server support
2188 This issue was introduced as part of the fix for Lucky 13 padding
2189 attack (CVE-2013-0169). The padding check was rewritten to be in
2190 constant time by making sure that always the same bytes are read and
2191 compared against either the MAC or padding bytes. But it no longer
2192 checked that there was enough data to have both the MAC and padding
2195 This issue was reported by Juraj Somorovsky using TLS-Attacker.
2199 *) Fix EVP_EncodeUpdate overflow
2201 An overflow can occur in the EVP_EncodeUpdate() function which is used for
2202 Base64 encoding of binary data. If an attacker is able to supply very large
2203 amounts of input data then a length check can overflow resulting in a heap
2206 Internally to OpenSSL the EVP_EncodeUpdate() function is primarily used by
2207 the PEM_write_bio* family of functions. These are mainly used within the
2208 OpenSSL command line applications, so any application which processes data
2209 from an untrusted source and outputs it as a PEM file should be considered
2210 vulnerable to this issue. User applications that call these APIs directly
2211 with large amounts of untrusted data may also be vulnerable.
2213 This issue was reported by Guido Vranken.
2217 *) Fix EVP_EncryptUpdate overflow
2219 An overflow can occur in the EVP_EncryptUpdate() function. If an attacker
2220 is able to supply very large amounts of input data after a previous call to
2221 EVP_EncryptUpdate() with a partial block then a length check can overflow
2222 resulting in a heap corruption. Following an analysis of all OpenSSL
2223 internal usage of the EVP_EncryptUpdate() function all usage is one of two
2224 forms. The first form is where the EVP_EncryptUpdate() call is known to be
2225 the first called function after an EVP_EncryptInit(), and therefore that
2226 specific call must be safe. The second form is where the length passed to
2227 EVP_EncryptUpdate() can be seen from the code to be some small value and
2228 therefore there is no possibility of an overflow. Since all instances are
2229 one of these two forms, it is believed that there can be no overflows in
2230 internal code due to this problem. It should be noted that
2231 EVP_DecryptUpdate() can call EVP_EncryptUpdate() in certain code paths.
2232 Also EVP_CipherUpdate() is a synonym for EVP_EncryptUpdate(). All instances
2233 of these calls have also been analysed too and it is believed there are no
2234 instances in internal usage where an overflow could occur.
2236 This issue was reported by Guido Vranken.
2240 *) Prevent ASN.1 BIO excessive memory allocation
2242 When ASN.1 data is read from a BIO using functions such as d2i_CMS_bio()
2243 a short invalid encoding can cause allocation of large amounts of memory
2244 potentially consuming excessive resources or exhausting memory.
2246 Any application parsing untrusted data through d2i BIO functions is
2247 affected. The memory based functions such as d2i_X509() are *not* affected.
2248 Since the memory based functions are used by the TLS library, TLS
2249 applications are not affected.
2251 This issue was reported by Brian Carpenter.
2257 ASN1 Strings that are over 1024 bytes can cause an overread in applications
2258 using the X509_NAME_oneline() function on EBCDIC systems. This could result
2259 in arbitrary stack data being returned in the buffer.
2261 This issue was reported by Guido Vranken.
2265 *) Modify behavior of ALPN to invoke callback after SNI/servername
2266 callback, such that updates to the SSL_CTX affect ALPN.
2269 *) Remove LOW from the DEFAULT cipher list. This removes singles DES from the
2273 *) Only remove the SSLv2 methods with the no-ssl2-method option. When the
2274 methods are enabled and ssl2 is disabled the methods return NULL.
2277 Changes between 1.0.2f and 1.0.2g [1 Mar 2016]
2279 * Disable weak ciphers in SSLv3 and up in default builds of OpenSSL.
2280 Builds that are not configured with "enable-weak-ssl-ciphers" will not
2281 provide any "EXPORT" or "LOW" strength ciphers.
2284 * Disable SSLv2 default build, default negotiation and weak ciphers. SSLv2
2285 is by default disabled at build-time. Builds that are not configured with
2286 "enable-ssl2" will not support SSLv2. Even if "enable-ssl2" is used,
2287 users who want to negotiate SSLv2 via the version-flexible SSLv23_method()
2288 will need to explicitly call either of:
2290 SSL_CTX_clear_options(ctx, SSL_OP_NO_SSLv2);
2292 SSL_clear_options(ssl, SSL_OP_NO_SSLv2);
2294 as appropriate. Even if either of those is used, or the application
2295 explicitly uses the version-specific SSLv2_method() or its client and
2296 server variants, SSLv2 ciphers vulnerable to exhaustive search key
2297 recovery have been removed. Specifically, the SSLv2 40-bit EXPORT
2298 ciphers, and SSLv2 56-bit DES are no longer available.
2302 *) Fix a double-free in DSA code
2304 A double free bug was discovered when OpenSSL parses malformed DSA private
2305 keys and could lead to a DoS attack or memory corruption for applications
2306 that receive DSA private keys from untrusted sources. This scenario is
2309 This issue was reported to OpenSSL by Adam Langley(Google/BoringSSL) using
2314 *) Disable SRP fake user seed to address a server memory leak.
2316 Add a new method SRP_VBASE_get1_by_user that handles the seed properly.
2318 SRP_VBASE_get_by_user had inconsistent memory management behaviour.
2319 In order to fix an unavoidable memory leak, SRP_VBASE_get_by_user
2320 was changed to ignore the "fake user" SRP seed, even if the seed
2323 Users should use SRP_VBASE_get1_by_user instead. Note that in
2324 SRP_VBASE_get1_by_user, caller must free the returned value. Note
2325 also that even though configuring the SRP seed attempts to hide
2326 invalid usernames by continuing the handshake with fake
2327 credentials, this behaviour is not constant time and no strong
2328 guarantees are made that the handshake is indistinguishable from
2329 that of a valid user.
2333 *) Fix BN_hex2bn/BN_dec2bn NULL pointer deref/heap corruption
2335 In the BN_hex2bn function the number of hex digits is calculated using an
2336 int value |i|. Later |bn_expand| is called with a value of |i * 4|. For
2337 large values of |i| this can result in |bn_expand| not allocating any
2338 memory because |i * 4| is negative. This can leave the internal BIGNUM data
2339 field as NULL leading to a subsequent NULL ptr deref. For very large values
2340 of |i|, the calculation |i * 4| could be a positive value smaller than |i|.
2341 In this case memory is allocated to the internal BIGNUM data field, but it
2342 is insufficiently sized leading to heap corruption. A similar issue exists
2343 in BN_dec2bn. This could have security consequences if BN_hex2bn/BN_dec2bn
2344 is ever called by user applications with very large untrusted hex/dec data.
2345 This is anticipated to be a rare occurrence.
2347 All OpenSSL internal usage of these functions use data that is not expected
2348 to be untrusted, e.g. config file data or application command line
2349 arguments. If user developed applications generate config file data based
2350 on untrusted data then it is possible that this could also lead to security
2351 consequences. This is also anticipated to be rare.
2353 This issue was reported to OpenSSL by Guido Vranken.
2357 *) Fix memory issues in BIO_*printf functions
2359 The internal |fmtstr| function used in processing a "%s" format string in
2360 the BIO_*printf functions could overflow while calculating the length of a
2361 string and cause an OOB read when printing very long strings.
2363 Additionally the internal |doapr_outch| function can attempt to write to an
2364 OOB memory location (at an offset from the NULL pointer) in the event of a
2365 memory allocation failure. In 1.0.2 and below this could be caused where
2366 the size of a buffer to be allocated is greater than INT_MAX. E.g. this
2367 could be in processing a very long "%s" format string. Memory leaks can
2370 The first issue may mask the second issue dependent on compiler behaviour.
2371 These problems could enable attacks where large amounts of untrusted data
2372 is passed to the BIO_*printf functions. If applications use these functions
2373 in this way then they could be vulnerable. OpenSSL itself uses these
2374 functions when printing out human-readable dumps of ASN.1 data. Therefore
2375 applications that print this data could be vulnerable if the data is from
2376 untrusted sources. OpenSSL command line applications could also be
2377 vulnerable where they print out ASN.1 data, or if untrusted data is passed
2378 as command line arguments.
2380 Libssl is not considered directly vulnerable. Additionally certificates etc
2381 received via remote connections via libssl are also unlikely to be able to
2382 trigger these issues because of message size limits enforced within libssl.
2384 This issue was reported to OpenSSL Guido Vranken.
2388 *) Side channel attack on modular exponentiation
2390 A side-channel attack was found which makes use of cache-bank conflicts on
2391 the Intel Sandy-Bridge microarchitecture which could lead to the recovery
2392 of RSA keys. The ability to exploit this issue is limited as it relies on
2393 an attacker who has control of code in a thread running on the same
2394 hyper-threaded core as the victim thread which is performing decryptions.
2396 This issue was reported to OpenSSL by Yuval Yarom, The University of
2397 Adelaide and NICTA, Daniel Genkin, Technion and Tel Aviv University, and
2398 Nadia Heninger, University of Pennsylvania with more information at
2399 http://cachebleed.info.
2403 *) Change the req app to generate a 2048-bit RSA/DSA key by default,
2404 if no keysize is specified with default_bits. This fixes an
2405 omission in an earlier change that changed all RSA/DSA key generation
2406 apps to use 2048 bits by default.
2409 Changes between 1.0.2e and 1.0.2f [28 Jan 2016]
2410 *) DH small subgroups
2412 Historically OpenSSL only ever generated DH parameters based on "safe"
2413 primes. More recently (in version 1.0.2) support was provided for
2414 generating X9.42 style parameter files such as those required for RFC 5114
2415 support. The primes used in such files may not be "safe". Where an
2416 application is using DH configured with parameters based on primes that are
2417 not "safe" then an attacker could use this fact to find a peer's private
2418 DH exponent. This attack requires that the attacker complete multiple
2419 handshakes in which the peer uses the same private DH exponent. For example
2420 this could be used to discover a TLS server's private DH exponent if it's
2421 reusing the private DH exponent or it's using a static DH ciphersuite.
2423 OpenSSL provides the option SSL_OP_SINGLE_DH_USE for ephemeral DH (DHE) in
2424 TLS. It is not on by default. If the option is not set then the server
2425 reuses the same private DH exponent for the life of the server process and
2426 would be vulnerable to this attack. It is believed that many popular
2427 applications do set this option and would therefore not be at risk.
2429 The fix for this issue adds an additional check where a "q" parameter is
2430 available (as is the case in X9.42 based parameters). This detects the
2431 only known attack, and is the only possible defense for static DH
2432 ciphersuites. This could have some performance impact.
2434 Additionally the SSL_OP_SINGLE_DH_USE option has been switched on by
2435 default and cannot be disabled. This could have some performance impact.
2437 This issue was reported to OpenSSL by Antonio Sanso (Adobe).
2441 *) SSLv2 doesn't block disabled ciphers
2443 A malicious client can negotiate SSLv2 ciphers that have been disabled on
2444 the server and complete SSLv2 handshakes even if all SSLv2 ciphers have
2445 been disabled, provided that the SSLv2 protocol was not also disabled via
2448 This issue was reported to OpenSSL on 26th December 2015 by Nimrod Aviram
2449 and Sebastian Schinzel.
2453 Changes between 1.0.2d and 1.0.2e [3 Dec 2015]
2455 *) BN_mod_exp may produce incorrect results on x86_64
2457 There is a carry propagating bug in the x86_64 Montgomery squaring
2458 procedure. No EC algorithms are affected. Analysis suggests that attacks
2459 against RSA and DSA as a result of this defect would be very difficult to
2460 perform and are not believed likely. Attacks against DH are considered just
2461 feasible (although very difficult) because most of the work necessary to
2462 deduce information about a private key may be performed offline. The amount
2463 of resources required for such an attack would be very significant and
2464 likely only accessible to a limited number of attackers. An attacker would
2465 additionally need online access to an unpatched system using the target
2466 private key in a scenario with persistent DH parameters and a private
2467 key that is shared between multiple clients. For example this can occur by
2468 default in OpenSSL DHE based SSL/TLS ciphersuites.
2470 This issue was reported to OpenSSL by Hanno Böck.
2474 *) Certificate verify crash with missing PSS parameter
2476 The signature verification routines will crash with a NULL pointer
2477 dereference if presented with an ASN.1 signature using the RSA PSS
2478 algorithm and absent mask generation function parameter. Since these
2479 routines are used to verify certificate signature algorithms this can be
2480 used to crash any certificate verification operation and exploited in a
2481 DoS attack. Any application which performs certificate verification is
2482 vulnerable including OpenSSL clients and servers which enable client
2485 This issue was reported to OpenSSL by Loïc Jonas Etienne (Qnective AG).
2489 *) X509_ATTRIBUTE memory leak
2491 When presented with a malformed X509_ATTRIBUTE structure OpenSSL will leak
2492 memory. This structure is used by the PKCS#7 and CMS routines so any
2493 application which reads PKCS#7 or CMS data from untrusted sources is
2494 affected. SSL/TLS is not affected.
2496 This issue was reported to OpenSSL by Adam Langley (Google/BoringSSL) using
2501 *) Rewrite EVP_DecodeUpdate (base64 decoding) to fix several bugs.
2502 This changes the decoding behaviour for some invalid messages,
2503 though the change is mostly in the more lenient direction, and
2504 legacy behaviour is preserved as much as possible.
2507 *) In DSA_generate_parameters_ex, if the provided seed is too short,
2509 [Rich Salz and Ismo Puustinen <ismo.puustinen@intel.com>]
2511 Changes between 1.0.2c and 1.0.2d [9 Jul 2015]
2513 *) Alternate chains certificate forgery
2515 During certificate verification, OpenSSL will attempt to find an
2516 alternative certificate chain if the first attempt to build such a chain
2517 fails. An error in the implementation of this logic can mean that an
2518 attacker could cause certain checks on untrusted certificates to be
2519 bypassed, such as the CA flag, enabling them to use a valid leaf
2520 certificate to act as a CA and "issue" an invalid certificate.
2522 This issue was reported to OpenSSL by Adam Langley/David Benjamin
2526 Changes between 1.0.2b and 1.0.2c [12 Jun 2015]
2528 *) Fix HMAC ABI incompatibility. The previous version introduced an ABI
2529 incompatibility in the handling of HMAC. The previous ABI has now been
2533 Changes between 1.0.2a and 1.0.2b [11 Jun 2015]
2535 *) Malformed ECParameters causes infinite loop
2537 When processing an ECParameters structure OpenSSL enters an infinite loop
2538 if the curve specified is over a specially malformed binary polynomial
2541 This can be used to perform denial of service against any
2542 system which processes public keys, certificate requests or
2543 certificates. This includes TLS clients and TLS servers with
2544 client authentication enabled.
2546 This issue was reported to OpenSSL by Joseph Barr-Pixton.
2550 *) Exploitable out-of-bounds read in X509_cmp_time
2552 X509_cmp_time does not properly check the length of the ASN1_TIME
2553 string and can read a few bytes out of bounds. In addition,
2554 X509_cmp_time accepts an arbitrary number of fractional seconds in the
2557 An attacker can use this to craft malformed certificates and CRLs of
2558 various sizes and potentially cause a segmentation fault, resulting in
2559 a DoS on applications that verify certificates or CRLs. TLS clients
2560 that verify CRLs are affected. TLS clients and servers with client
2561 authentication enabled may be affected if they use custom verification
2564 This issue was reported to OpenSSL by Robert Swiecki (Google), and
2565 independently by Hanno Böck.
2569 *) PKCS7 crash with missing EnvelopedContent
2571 The PKCS#7 parsing code does not handle missing inner EncryptedContent
2572 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs
2573 with missing content and trigger a NULL pointer dereference on parsing.
2575 Applications that decrypt PKCS#7 data or otherwise parse PKCS#7
2576 structures from untrusted sources are affected. OpenSSL clients and
2577 servers are not affected.
2579 This issue was reported to OpenSSL by Michal Zalewski (Google).
2583 *) CMS verify infinite loop with unknown hash function
2585 When verifying a signedData message the CMS code can enter an infinite loop
2586 if presented with an unknown hash function OID. This can be used to perform
2587 denial of service against any system which verifies signedData messages using
2589 This issue was reported to OpenSSL by Johannes Bauer.
2593 *) Race condition handling NewSessionTicket
2595 If a NewSessionTicket is received by a multi-threaded client when attempting to
2596 reuse a previous ticket then a race condition can occur potentially leading to
2597 a double free of the ticket data.
2601 *) Only support 256-bit or stronger elliptic curves with the
2602 'ecdh_auto' setting (server) or by default (client). Of supported
2603 curves, prefer P-256 (both).
2606 Changes between 1.0.2 and 1.0.2a [19 Mar 2015]
2608 *) ClientHello sigalgs DoS fix
2610 If a client connects to an OpenSSL 1.0.2 server and renegotiates with an
2611 invalid signature algorithms extension a NULL pointer dereference will
2612 occur. This can be exploited in a DoS attack against the server.
2614 This issue was was reported to OpenSSL by David Ramos of Stanford
2617 [Stephen Henson and Matt Caswell]
2619 *) Multiblock corrupted pointer fix
2621 OpenSSL 1.0.2 introduced the "multiblock" performance improvement. This
2622 feature only applies on 64 bit x86 architecture platforms that support AES
2623 NI instructions. A defect in the implementation of "multiblock" can cause
2624 OpenSSL's internal write buffer to become incorrectly set to NULL when
2625 using non-blocking IO. Typically, when the user application is using a
2626 socket BIO for writing, this will only result in a failed connection.
2627 However if some other BIO is used then it is likely that a segmentation
2628 fault will be triggered, thus enabling a potential DoS attack.
2630 This issue was reported to OpenSSL by Daniel Danner and Rainer Mueller.
2634 *) Segmentation fault in DTLSv1_listen fix
2636 The DTLSv1_listen function is intended to be stateless and processes the
2637 initial ClientHello from many peers. It is common for user code to loop
2638 over the call to DTLSv1_listen until a valid ClientHello is received with
2639 an associated cookie. A defect in the implementation of DTLSv1_listen means
2640 that state is preserved in the SSL object from one invocation to the next
2641 that can lead to a segmentation fault. Errors processing the initial
2642 ClientHello can trigger this scenario. An example of such an error could be
2643 that a DTLS1.0 only client is attempting to connect to a DTLS1.2 only
2646 This issue was reported to OpenSSL by Per Allansson.
2650 *) Segmentation fault in ASN1_TYPE_cmp fix
2652 The function ASN1_TYPE_cmp will crash with an invalid read if an attempt is
2653 made to compare ASN.1 boolean types. Since ASN1_TYPE_cmp is used to check
2654 certificate signature algorithm consistency this can be used to crash any
2655 certificate verification operation and exploited in a DoS attack. Any
2656 application which performs certificate verification is vulnerable including
2657 OpenSSL clients and servers which enable client authentication.
2661 *) Segmentation fault for invalid PSS parameters fix
2663 The signature verification routines will crash with a NULL pointer
2664 dereference if presented with an ASN.1 signature using the RSA PSS
2665 algorithm and invalid parameters. Since these routines are used to verify
2666 certificate signature algorithms this can be used to crash any
2667 certificate verification operation and exploited in a DoS attack. Any
2668 application which performs certificate verification is vulnerable including
2669 OpenSSL clients and servers which enable client authentication.
2671 This issue was was reported to OpenSSL by Brian Carpenter.
2675 *) ASN.1 structure reuse memory corruption fix
2677 Reusing a structure in ASN.1 parsing may allow an attacker to cause
2678 memory corruption via an invalid write. Such reuse is and has been
2679 strongly discouraged and is believed to be rare.
2681 Applications that parse structures containing CHOICE or ANY DEFINED BY
2682 components may be affected. Certificate parsing (d2i_X509 and related
2683 functions) are however not affected. OpenSSL clients and servers are
2688 *) PKCS7 NULL pointer dereferences fix
2690 The PKCS#7 parsing code does not handle missing outer ContentInfo
2691 correctly. An attacker can craft malformed ASN.1-encoded PKCS#7 blobs with
2692 missing content and trigger a NULL pointer dereference on parsing.
2694 Applications that verify PKCS#7 signatures, decrypt PKCS#7 data or
2695 otherwise parse PKCS#7 structures from untrusted sources are
2696 affected. OpenSSL clients and servers are not affected.
2698 This issue was reported to OpenSSL by Michal Zalewski (Google).
2702 *) DoS via reachable assert in SSLv2 servers fix
2704 A malicious client can trigger an OPENSSL_assert (i.e., an abort) in
2705 servers that both support SSLv2 and enable export cipher suites by sending
2706 a specially crafted SSLv2 CLIENT-MASTER-KEY message.
2708 This issue was discovered by Sean Burford (Google) and Emilia Käsper
2709 (OpenSSL development team).
2713 *) Empty CKE with client auth and DHE fix
2715 If client auth is used then a server can seg fault in the event of a DHE
2716 ciphersuite being selected and a zero length ClientKeyExchange message
2717 being sent by the client. This could be exploited in a DoS attack.
2721 *) Handshake with unseeded PRNG fix
2723 Under certain conditions an OpenSSL 1.0.2 client can complete a handshake
2724 with an unseeded PRNG. The conditions are:
2725 - The client is on a platform where the PRNG has not been seeded
2726 automatically, and the user has not seeded manually
2727 - A protocol specific client method version has been used (i.e. not
2728 SSL_client_methodv23)
2729 - A ciphersuite is used that does not require additional random data from
2730 the PRNG beyond the initial ClientHello client random (e.g. PSK-RC4-SHA).
2732 If the handshake succeeds then the client random that has been used will
2733 have been generated from a PRNG with insufficient entropy and therefore the
2734 output may be predictable.
2736 For example using the following command with an unseeded openssl will
2737 succeed on an unpatched platform:
2739 openssl s_client -psk 1a2b3c4d -tls1_2 -cipher PSK-RC4-SHA
2743 *) Use After Free following d2i_ECPrivatekey error fix
2745 A malformed EC private key file consumed via the d2i_ECPrivateKey function
2746 could cause a use after free condition. This, in turn, could cause a double
2747 free in several private key parsing functions (such as d2i_PrivateKey
2748 or EVP_PKCS82PKEY) and could lead to a DoS attack or memory corruption
2749 for applications that receive EC private keys from untrusted
2750 sources. This scenario is considered rare.
2752 This issue was discovered by the BoringSSL project and fixed in their
2757 *) X509_to_X509_REQ NULL pointer deref fix
2759 The function X509_to_X509_REQ will crash with a NULL pointer dereference if
2760 the certificate key is invalid. This function is rarely used in practice.
2762 This issue was discovered by Brian Carpenter.
2766 *) Removed the export ciphers from the DEFAULT ciphers
2769 Changes between 1.0.1l and 1.0.2 [22 Jan 2015]
2771 *) Facilitate "universal" ARM builds targeting range of ARM ISAs, e.g.
2772 ARMv5 through ARMv8, as opposite to "locking" it to single one.
2773 So far those who have to target multiple platforms would compromise
2774 and argue that binary targeting say ARMv5 would still execute on
2775 ARMv8. "Universal" build resolves this compromise by providing
2776 near-optimal performance even on newer platforms.
2779 *) Accelerated NIST P-256 elliptic curve implementation for x86_64
2780 (other platforms pending).
2781 [Shay Gueron & Vlad Krasnov (Intel Corp), Andy Polyakov]
2783 *) Add support for the SignedCertificateTimestampList certificate and
2784 OCSP response extensions from RFC6962.
2787 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
2788 for corner cases. (Certain input points at infinity could lead to
2789 bogus results, with non-infinity inputs mapped to infinity too.)
2792 *) Initial support for PowerISA 2.0.7, first implemented in POWER8.
2793 This covers AES, SHA256/512 and GHASH. "Initial" means that most
2794 common cases are optimized and there still is room for further
2795 improvements. Vector Permutation AES for Altivec is also added.
2798 *) Add support for little-endian ppc64 Linux target.
2799 [Marcelo Cerri (IBM)]
2801 *) Initial support for AMRv8 ISA crypto extensions. This covers AES,
2802 SHA1, SHA256 and GHASH. "Initial" means that most common cases
2803 are optimized and there still is room for further improvements.
2804 Both 32- and 64-bit modes are supported.
2805 [Andy Polyakov, Ard Biesheuvel (Linaro)]
2807 *) Improved ARMv7 NEON support.
2810 *) Support for SPARC Architecture 2011 crypto extensions, first
2811 implemented in SPARC T4. This covers AES, DES, Camellia, SHA1,
2812 SHA256/512, MD5, GHASH and modular exponentiation.
2813 [Andy Polyakov, David Miller]
2815 *) Accelerated modular exponentiation for Intel processors, a.k.a.
2817 [Shay Gueron & Vlad Krasnov (Intel Corp)]
2819 *) Support for new and upcoming Intel processors, including AVX2,
2820 BMI and SHA ISA extensions. This includes additional "stitched"
2821 implementations, AESNI-SHA256 and GCM, and multi-buffer support
2824 This work was sponsored by Intel Corp.
2827 *) Support for DTLS 1.2. This adds two sets of DTLS methods: DTLS_*_method()
2828 supports both DTLS 1.2 and 1.0 and should use whatever version the peer
2829 supports and DTLSv1_2_*_method() which supports DTLS 1.2 only.
2832 *) Use algorithm specific chains in SSL_CTX_use_certificate_chain_file():
2833 this fixes a limitation in previous versions of OpenSSL.
2836 *) Extended RSA OAEP support via EVP_PKEY API. Options to specify digest,
2837 MGF1 digest and OAEP label.
2840 *) Add EVP support for key wrapping algorithms, to avoid problems with
2841 existing code the flag EVP_CIPHER_CTX_WRAP_ALLOW has to be set in
2842 the EVP_CIPHER_CTX or an error is returned. Add AES and DES3 wrap
2843 algorithms and include tests cases.
2846 *) Add functions to allocate and set the fields of an ECDSA_METHOD
2848 [Douglas E. Engert, Steve Henson]
2850 *) New functions OPENSSL_gmtime_diff and ASN1_TIME_diff to find the
2851 difference in days and seconds between two tm or ASN1_TIME structures.
2854 *) Add -rev test option to s_server to just reverse order of characters
2855 received by client and send back to server. Also prints an abbreviated
2856 summary of the connection parameters.
2859 *) New option -brief for s_client and s_server to print out a brief summary
2860 of connection parameters.
2863 *) Add callbacks for arbitrary TLS extensions.
2864 [Trevor Perrin <trevp@trevp.net> and Ben Laurie]
2866 *) New option -crl_download in several openssl utilities to download CRLs
2867 from CRLDP extension in certificates.
2870 *) New options -CRL and -CRLform for s_client and s_server for CRLs.
2873 *) New function X509_CRL_diff to generate a delta CRL from the difference
2874 of two full CRLs. Add support to "crl" utility.
2877 *) New functions to set lookup_crls function and to retrieve
2878 X509_STORE from X509_STORE_CTX.
2881 *) Print out deprecated issuer and subject unique ID fields in
2885 *) Extend OCSP I/O functions so they can be used for simple general purpose
2886 HTTP as well as OCSP. New wrapper function which can be used to download
2887 CRLs using the OCSP API.
2890 *) Delegate command line handling in s_client/s_server to SSL_CONF APIs.
2893 *) SSL_CONF* functions. These provide a common framework for application
2894 configuration using configuration files or command lines.
2897 *) SSL/TLS tracing code. This parses out SSL/TLS records using the
2898 message callback and prints the results. Needs compile time option
2899 "enable-ssl-trace". New options to s_client and s_server to enable
2903 *) New ctrl and macro to retrieve supported points extensions.
2904 Print out extension in s_server and s_client.
2907 *) New functions to retrieve certificate signature and signature
2911 *) Add functions to retrieve and manipulate the raw cipherlist sent by a
2915 *) New Suite B modes for TLS code. These use and enforce the requirements
2916 of RFC6460: restrict ciphersuites, only permit Suite B algorithms and
2917 only use Suite B curves. The Suite B modes can be set by using the
2918 strings "SUITEB128", "SUITEB192" or "SUITEB128ONLY" for the cipherstring.
2921 *) New chain verification flags for Suite B levels of security. Check
2922 algorithms are acceptable when flags are set in X509_verify_cert.
2925 *) Make tls1_check_chain return a set of flags indicating checks passed
2926 by a certificate chain. Add additional tests to handle client
2927 certificates: checks for matching certificate type and issuer name
2931 *) If an attempt is made to use a signature algorithm not in the peer
2932 preference list abort the handshake. If client has no suitable
2933 signature algorithms in response to a certificate request do not
2934 use the certificate.
2937 *) If server EC tmp key is not in client preference list abort handshake.
2940 *) Add support for certificate stores in CERT structure. This makes it
2941 possible to have different stores per SSL structure or one store in
2942 the parent SSL_CTX. Include distinct stores for certificate chain
2943 verification and chain building. New ctrl SSL_CTRL_BUILD_CERT_CHAIN
2944 to build and store a certificate chain in CERT structure: returning
2945 an error if the chain cannot be built: this will allow applications
2946 to test if a chain is correctly configured.
2948 Note: if the CERT based stores are not set then the parent SSL_CTX
2949 store is used to retain compatibility with existing behaviour.
2953 *) New function ssl_set_client_disabled to set a ciphersuite disabled
2954 mask based on the current session, check mask when sending client
2955 hello and checking the requested ciphersuite.
2958 *) New ctrls to retrieve and set certificate types in a certificate
2959 request message. Print out received values in s_client. If certificate
2960 types is not set with custom values set sensible values based on
2961 supported signature algorithms.
2964 *) Support for distinct client and server supported signature algorithms.
2967 *) Add certificate callback. If set this is called whenever a certificate
2968 is required by client or server. An application can decide which
2969 certificate chain to present based on arbitrary criteria: for example
2970 supported signature algorithms. Add very simple example to s_server.
2971 This fixes many of the problems and restrictions of the existing client
2972 certificate callback: for example you can now clear an existing
2973 certificate and specify the whole chain.
2976 *) Add new "valid_flags" field to CERT_PKEY structure which determines what
2977 the certificate can be used for (if anything). Set valid_flags field
2978 in new tls1_check_chain function. Simplify ssl_set_cert_masks which used
2979 to have similar checks in it.
2981 Add new "cert_flags" field to CERT structure and include a "strict mode".
2982 This enforces some TLS certificate requirements (such as only permitting
2983 certificate signature algorithms contained in the supported algorithms
2984 extension) which some implementations ignore: this option should be used
2985 with caution as it could cause interoperability issues.
2988 *) Update and tidy signature algorithm extension processing. Work out
2989 shared signature algorithms based on preferences and peer algorithms
2990 and print them out in s_client and s_server. Abort handshake if no
2991 shared signature algorithms.
2994 *) Add new functions to allow customised supported signature algorithms
2995 for SSL and SSL_CTX structures. Add options to s_client and s_server
2999 *) New function SSL_certs_clear() to delete all references to certificates
3000 from an SSL structure. Before this once a certificate had been added
3001 it couldn't be removed.
3004 *) Integrate hostname, email address and IP address checking with certificate
3005 verification. New verify options supporting checking in openssl utility.
3008 *) Fixes and wildcard matching support to hostname and email checking
3009 functions. Add manual page.
3010 [Florian Weimer (Red Hat Product Security Team)]
3012 *) New functions to check a hostname email or IP address against a
3013 certificate. Add options x509 utility to print results of checks against
3017 *) Fix OCSP checking.
3018 [Rob Stradling <rob.stradling@comodo.com> and Ben Laurie]
3020 *) Initial experimental support for explicitly trusted non-root CAs.
3021 OpenSSL still tries to build a complete chain to a root but if an
3022 intermediate CA has a trust setting included that is used. The first
3023 setting is used: whether to trust (e.g., -addtrust option to the x509
3027 *) Add -trusted_first option which attempts to find certificates in the
3028 trusted store even if an untrusted chain is also supplied.
3031 *) MIPS assembly pack updates: support for MIPS32r2 and SmartMIPS ASE,
3032 platform support for Linux and Android.
3035 *) Support for linux-x32, ILP32 environment in x86_64 framework.
3038 *) Experimental multi-implementation support for FIPS capable OpenSSL.
3039 When in FIPS mode the approved implementations are used as normal,
3040 when not in FIPS mode the internal unapproved versions are used instead.
3041 This means that the FIPS capable OpenSSL isn't forced to use the
3042 (often lower performance) FIPS implementations outside FIPS mode.
3045 *) Transparently support X9.42 DH parameters when calling
3046 PEM_read_bio_DHparameters. This means existing applications can handle
3047 the new parameter format automatically.
3050 *) Initial experimental support for X9.42 DH parameter format: mainly
3051 to support use of 'q' parameter for RFC5114 parameters.
3054 *) Add DH parameters from RFC5114 including test data to dhtest.
3057 *) Support for automatic EC temporary key parameter selection. If enabled
3058 the most preferred EC parameters are automatically used instead of
3059 hardcoded fixed parameters. Now a server just has to call:
3060 SSL_CTX_set_ecdh_auto(ctx, 1) and the server will automatically
3061 support ECDH and use the most appropriate parameters.
3064 *) Enhance and tidy EC curve and point format TLS extension code. Use
3065 static structures instead of allocation if default values are used.
3066 New ctrls to set curves we wish to support and to retrieve shared curves.
3067 Print out shared curves in s_server. New options to s_server and s_client
3068 to set list of supported curves.
3071 *) New ctrls to retrieve supported signature algorithms and
3072 supported curve values as an array of NIDs. Extend openssl utility
3073 to print out received values.
3076 *) Add new APIs EC_curve_nist2nid and EC_curve_nid2nist which convert
3077 between NIDs and the more common NIST names such as "P-256". Enhance
3078 ecparam utility and ECC method to recognise the NIST names for curves.
3081 *) Enhance SSL/TLS certificate chain handling to support different
3082 chains for each certificate instead of one chain in the parent SSL_CTX.
3085 *) Support for fixed DH ciphersuite client authentication: where both
3086 server and client use DH certificates with common parameters.
3089 *) Support for fixed DH ciphersuites: those requiring DH server
3093 *) New function i2d_re_X509_tbs for re-encoding the TBS portion of
3095 Note: Related 1.0.2-beta specific macros X509_get_cert_info,
3096 X509_CINF_set_modified, X509_CINF_get_issuer, X509_CINF_get_extensions and
3097 X509_CINF_get_signature were reverted post internal team review.
3099 Changes between 1.0.1k and 1.0.1l [15 Jan 2015]
3101 *) Build fixes for the Windows and OpenVMS platforms
3102 [Matt Caswell and Richard Levitte]
3104 Changes between 1.0.1j and 1.0.1k [8 Jan 2015]
3106 *) Fix DTLS segmentation fault in dtls1_get_record. A carefully crafted DTLS
3107 message can cause a segmentation fault in OpenSSL due to a NULL pointer
3108 dereference. This could lead to a Denial Of Service attack. Thanks to
3109 Markus Stenberg of Cisco Systems, Inc. for reporting this issue.
3113 *) Fix DTLS memory leak in dtls1_buffer_record. A memory leak can occur in the
3114 dtls1_buffer_record function under certain conditions. In particular this
3115 could occur if an attacker sent repeated DTLS records with the same
3116 sequence number but for the next epoch. The memory leak could be exploited
3117 by an attacker in a Denial of Service attack through memory exhaustion.
3118 Thanks to Chris Mueller for reporting this issue.
3122 *) Fix issue where no-ssl3 configuration sets method to NULL. When openssl is
3123 built with the no-ssl3 option and a SSL v3 ClientHello is received the ssl
3124 method would be set to NULL which could later result in a NULL pointer
3125 dereference. Thanks to Frank Schmirler for reporting this issue.
3129 *) Abort handshake if server key exchange message is omitted for ephemeral
3132 Thanks to Karthikeyan Bhargavan of the PROSECCO team at INRIA for
3133 reporting this issue.
3137 *) Remove non-export ephemeral RSA code on client and server. This code
3138 violated the TLS standard by allowing the use of temporary RSA keys in
3139 non-export ciphersuites and could be used by a server to effectively
3140 downgrade the RSA key length used to a value smaller than the server
3141 certificate. Thanks for Karthikeyan Bhargavan of the PROSECCO team at
3142 INRIA or reporting this issue.
3146 *) Fixed issue where DH client certificates are accepted without verification.
3147 An OpenSSL server will accept a DH certificate for client authentication
3148 without the certificate verify message. This effectively allows a client to
3149 authenticate without the use of a private key. This only affects servers
3150 which trust a client certificate authority which issues certificates
3151 containing DH keys: these are extremely rare and hardly ever encountered.
3152 Thanks for Karthikeyan Bhargavan of the PROSECCO team at INRIA or reporting
3157 *) Ensure that the session ID context of an SSL is updated when its
3158 SSL_CTX is updated via SSL_set_SSL_CTX.
3160 The session ID context is typically set from the parent SSL_CTX,
3161 and can vary with the CTX.
3164 *) Fix various certificate fingerprint issues.
3166 By using non-DER or invalid encodings outside the signed portion of a
3167 certificate the fingerprint can be changed without breaking the signature.
3168 Although no details of the signed portion of the certificate can be changed
3169 this can cause problems with some applications: e.g. those using the
3170 certificate fingerprint for blacklists.
3172 1. Reject signatures with non zero unused bits.
3174 If the BIT STRING containing the signature has non zero unused bits reject
3175 the signature. All current signature algorithms require zero unused bits.
3177 2. Check certificate algorithm consistency.
3179 Check the AlgorithmIdentifier inside TBS matches the one in the
3180 certificate signature. NB: this will result in signature failure
3181 errors for some broken certificates.
3183 Thanks to Konrad Kraszewski from Google for reporting this issue.
3185 3. Check DSA/ECDSA signatures use DER.
3187 Re-encode DSA/ECDSA signatures and compare with the original received
3188 signature. Return an error if there is a mismatch.
3190 This will reject various cases including garbage after signature
3191 (thanks to Antti Karjalainen and Tuomo Untinen from the Codenomicon CROSS
3192 program for discovering this case) and use of BER or invalid ASN.1 INTEGERs
3193 (negative or with leading zeroes).
3195 Further analysis was conducted and fixes were developed by Stephen Henson
3196 of the OpenSSL core team.
3201 *) Correct Bignum squaring. Bignum squaring (BN_sqr) may produce incorrect
3202 results on some platforms, including x86_64. This bug occurs at random
3203 with a very low probability, and is not known to be exploitable in any
3204 way, though its exact impact is difficult to determine. Thanks to Pieter
3205 Wuille (Blockstream) who reported this issue and also suggested an initial
3206 fix. Further analysis was conducted by the OpenSSL development team and
3207 Adam Langley of Google. The final fix was developed by Andy Polyakov of
3208 the OpenSSL core team.
3212 *) Do not resume sessions on the server if the negotiated protocol
3213 version does not match the session's version. Resuming with a different
3214 version, while not strictly forbidden by the RFC, is of questionable
3215 sanity and breaks all known clients.
3216 [David Benjamin, Emilia Käsper]
3218 *) Tighten handling of the ChangeCipherSpec (CCS) message: reject
3219 early CCS messages during renegotiation. (Note that because
3220 renegotiation is encrypted, this early CCS was not exploitable.)
3223 *) Tighten client-side session ticket handling during renegotiation:
3224 ensure that the client only accepts a session ticket if the server sends
3225 the extension anew in the ServerHello. Previously, a TLS client would
3226 reuse the old extension state and thus accept a session ticket if one was
3227 announced in the initial ServerHello.
3229 Similarly, ensure that the client requires a session ticket if one
3230 was advertised in the ServerHello. Previously, a TLS client would
3231 ignore a missing NewSessionTicket message.
3234 Changes between 1.0.1i and 1.0.1j [15 Oct 2014]
3236 *) SRTP Memory Leak.
3238 A flaw in the DTLS SRTP extension parsing code allows an attacker, who
3239 sends a carefully crafted handshake message, to cause OpenSSL to fail
3240 to free up to 64k of memory causing a memory leak. This could be
3241 exploited in a Denial Of Service attack. This issue affects OpenSSL
3242 1.0.1 server implementations for both SSL/TLS and DTLS regardless of
3243 whether SRTP is used or configured. Implementations of OpenSSL that
3244 have been compiled with OPENSSL_NO_SRTP defined are not affected.
3246 The fix was developed by the OpenSSL team.
3250 *) Session Ticket Memory Leak.
3252 When an OpenSSL SSL/TLS/DTLS server receives a session ticket the
3253 integrity of that ticket is first verified. In the event of a session
3254 ticket integrity check failing, OpenSSL will fail to free memory
3255 causing a memory leak. By sending a large number of invalid session
3256 tickets an attacker could exploit this issue in a Denial Of Service
3261 *) Build option no-ssl3 is incomplete.
3263 When OpenSSL is configured with "no-ssl3" as a build option, servers
3264 could accept and complete a SSL 3.0 handshake, and clients could be
3265 configured to send them.
3267 [Akamai and the OpenSSL team]
3269 *) Add support for TLS_FALLBACK_SCSV.
3270 Client applications doing fallback retries should call
3271 SSL_set_mode(s, SSL_MODE_SEND_FALLBACK_SCSV).
3273 [Adam Langley, Bodo Moeller]
3275 *) Add additional DigestInfo checks.
3277 Re-encode DigestInto in DER and check against the original when
3278 verifying RSA signature: this will reject any improperly encoded
3279 DigestInfo structures.
3281 Note: this is a precautionary measure and no attacks are currently known.
3285 Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
3287 *) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
3288 SRP code can be overrun an internal buffer. Add sanity check that
3289 g, A, B < N to SRP code.
3291 Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
3292 Group for discovering this issue.
3296 *) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
3297 TLS 1.0 instead of higher protocol versions when the ClientHello message
3298 is badly fragmented. This allows a man-in-the-middle attacker to force a
3299 downgrade to TLS 1.0 even if both the server and the client support a
3300 higher protocol version, by modifying the client's TLS records.
3302 Thanks to David Benjamin and Adam Langley (Google) for discovering and
3303 researching this issue.
3307 *) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
3308 to a denial of service attack. A malicious server can crash the client
3309 with a null pointer dereference (read) by specifying an anonymous (EC)DH
3310 ciphersuite and sending carefully crafted handshake messages.
3312 Thanks to Felix Gröbert (Google) for discovering and researching this
3317 *) By sending carefully crafted DTLS packets an attacker could cause openssl
3318 to leak memory. This can be exploited through a Denial of Service attack.
3319 Thanks to Adam Langley for discovering and researching this issue.
3323 *) An attacker can force openssl to consume large amounts of memory whilst
3324 processing DTLS handshake messages. This can be exploited through a
3325 Denial of Service attack.
3326 Thanks to Adam Langley for discovering and researching this issue.
3330 *) An attacker can force an error condition which causes openssl to crash
3331 whilst processing DTLS packets due to memory being freed twice. This
3332 can be exploited through a Denial of Service attack.
3333 Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
3338 *) If a multithreaded client connects to a malicious server using a resumed
3339 session and the server sends an ec point format extension it could write
3340 up to 255 bytes to freed memory.
3342 Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
3347 *) A malicious server can crash an OpenSSL client with a null pointer
3348 dereference (read) by specifying an SRP ciphersuite even though it was not
3349 properly negotiated with the client. This can be exploited through a
3350 Denial of Service attack.
3352 Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
3353 discovering and researching this issue.
3357 *) A flaw in OBJ_obj2txt may cause pretty printing functions such as
3358 X509_name_oneline, X509_name_print_ex et al. to leak some information
3359 from the stack. Applications may be affected if they echo pretty printing
3360 output to the attacker.
3362 Thanks to Ivan Fratric (Google) for discovering this issue.
3364 [Emilia Käsper, and Steve Henson]
3366 *) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
3367 for corner cases. (Certain input points at infinity could lead to
3368 bogus results, with non-infinity inputs mapped to infinity too.)
3371 Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
3373 *) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
3374 handshake can force the use of weak keying material in OpenSSL
3375 SSL/TLS clients and servers.
3377 Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
3378 researching this issue. (CVE-2014-0224)
3379 [KIKUCHI Masashi, Steve Henson]
3381 *) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
3382 OpenSSL DTLS client the code can be made to recurse eventually crashing
3385 Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
3387 [Imre Rad, Steve Henson]
3389 *) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
3390 be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
3391 client or server. This is potentially exploitable to run arbitrary
3392 code on a vulnerable client or server.
3394 Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
3395 [Jüri Aedla, Steve Henson]
3397 *) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
3398 are subject to a denial of service attack.
3400 Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
3401 this issue. (CVE-2014-3470)
3402 [Felix Gröbert, Ivan Fratric, Steve Henson]
3404 *) Harmonize version and its documentation. -f flag is used to display
3406 [mancha <mancha1@zoho.com>]
3408 *) Fix eckey_priv_encode so it immediately returns an error upon a failure
3409 in i2d_ECPrivateKey.
3410 [mancha <mancha1@zoho.com>]
3412 *) Fix some double frees. These are not thought to be exploitable.
3413 [mancha <mancha1@zoho.com>]
3415 Changes between 1.0.1f and 1.0.1g [7 Apr 2014]
3417 *) A missing bounds check in the handling of the TLS heartbeat extension
3418 can be used to reveal up to 64k of memory to a connected client or
3421 Thanks for Neel Mehta of Google Security for discovering this bug and to
3422 Adam Langley <agl@chromium.org> and Bodo Moeller <bmoeller@acm.org> for
3423 preparing the fix (CVE-2014-0160)
3424 [Adam Langley, Bodo Moeller]
3426 *) Fix for the attack described in the paper "Recovering OpenSSL
3427 ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
3428 by Yuval Yarom and Naomi Benger. Details can be obtained from:
3429 http://eprint.iacr.org/2014/140
3431 Thanks to Yuval Yarom and Naomi Benger for discovering this
3432 flaw and to Yuval Yarom for supplying a fix (CVE-2014-0076)
3433 [Yuval Yarom and Naomi Benger]
3435 *) TLS pad extension: draft-agl-tls-padding-03
3437 Workaround for the "TLS hang bug" (see FAQ and PR#2771): if the
3438 TLS client Hello record length value would otherwise be > 255 and
3439 less that 512 pad with a dummy extension containing zeroes so it
3440 is at least 512 bytes long.
3442 [Adam Langley, Steve Henson]
3444 Changes between 1.0.1e and 1.0.1f [6 Jan 2014]
3446 *) Fix for TLS record tampering bug. A carefully crafted invalid
3447 handshake could crash OpenSSL with a NULL pointer exception.
3448 Thanks to Anton Johansson for reporting this issues.
3451 *) Keep original DTLS digest and encryption contexts in retransmission
3452 structures so we can use the previous session parameters if they need
3453 to be resent. (CVE-2013-6450)
3456 *) Add option SSL_OP_SAFARI_ECDHE_ECDSA_BUG (part of SSL_OP_ALL) which
3457 avoids preferring ECDHE-ECDSA ciphers when the client appears to be
3458 Safari on OS X. Safari on OS X 10.8..10.8.3 advertises support for
3459 several ECDHE-ECDSA ciphers, but fails to negotiate them. The bug
3460 is fixed in OS X 10.8.4, but Apple have ruled out both hot fixing
3461 10.8..10.8.3 and forcing users to upgrade to 10.8.4 or newer.
3462 [Rob Stradling, Adam Langley]
3464 Changes between 1.0.1d and 1.0.1e [11 Feb 2013]
3466 *) Correct fix for CVE-2013-0169. The original didn't work on AES-NI
3467 supporting platforms or when small records were transferred.
3468 [Andy Polyakov, Steve Henson]
3470 Changes between 1.0.1c and 1.0.1d [5 Feb 2013]
3472 *) Make the decoding of SSLv3, TLS and DTLS CBC records constant time.
3474 This addresses the flaw in CBC record processing discovered by
3475 Nadhem Alfardan and Kenny Paterson. Details of this attack can be found
3476 at: http://www.isg.rhul.ac.uk/tls/
3478 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3479 Security Group at Royal Holloway, University of London
3480 (www.isg.rhul.ac.uk) for discovering this flaw and Adam Langley and
3481 Emilia Käsper for the initial patch.
3483 [Emilia Käsper, Adam Langley, Ben Laurie, Andy Polyakov, Steve Henson]
3485 *) Fix flaw in AESNI handling of TLS 1.2 and 1.1 records for CBC mode
3486 ciphersuites which can be exploited in a denial of service attack.
3487 Thanks go to and to Adam Langley <agl@chromium.org> for discovering
3488 and detecting this bug and to Wolfgang Ettlinger
3489 <wolfgang.ettlinger@gmail.com> for independently discovering this issue.
3493 *) Return an error when checking OCSP signatures when key is NULL.
3494 This fixes a DoS attack. (CVE-2013-0166)
3497 *) Make openssl verify return errors.
3498 [Chris Palmer <palmer@google.com> and Ben Laurie]
3500 *) Call OCSP Stapling callback after ciphersuite has been chosen, so
3501 the right response is stapled. Also change SSL_get_certificate()
3502 so it returns the certificate actually sent.
3503 See http://rt.openssl.org/Ticket/Display.html?id=2836.
3504 [Rob Stradling <rob.stradling@comodo.com>]
3506 *) Fix possible deadlock when decoding public keys.
3509 *) Don't use TLS 1.0 record version number in initial client hello
3513 Changes between 1.0.1b and 1.0.1c [10 May 2012]
3515 *) Sanity check record length before skipping explicit IV in TLS
3516 1.2, 1.1 and DTLS to fix DoS attack.
3518 Thanks to Codenomicon for discovering this issue using Fuzz-o-Matic
3519 fuzzing as a service testing platform.
3523 *) Initialise tkeylen properly when encrypting CMS messages.
3524 Thanks to Solar Designer of Openwall for reporting this issue.
3527 *) In FIPS mode don't try to use composite ciphers as they are not
3531 Changes between 1.0.1a and 1.0.1b [26 Apr 2012]
3533 *) OpenSSL 1.0.0 sets SSL_OP_ALL to 0x80000FFFL and OpenSSL 1.0.1 and
3534 1.0.1a set SSL_OP_NO_TLSv1_1 to 0x00000400L which would unfortunately
3535 mean any application compiled against OpenSSL 1.0.0 headers setting
3536 SSL_OP_ALL would also set SSL_OP_NO_TLSv1_1, unintentionally disabling
3537 TLS 1.1 also. Fix this by changing the value of SSL_OP_NO_TLSv1_1 to
3538 0x10000000L Any application which was previously compiled against
3539 OpenSSL 1.0.1 or 1.0.1a headers and which cares about SSL_OP_NO_TLSv1_1
3540 will need to be recompiled as a result. Letting be results in
3541 inability to disable specifically TLS 1.1 and in client context,
3542 in unlike event, limit maximum offered version to TLS 1.0 [see below].
3545 *) In order to ensure interoperability SSL_OP_NO_protocolX does not
3546 disable just protocol X, but all protocols above X *if* there are
3547 protocols *below* X still enabled. In more practical terms it means
3548 that if application wants to disable TLS1.0 in favor of TLS1.1 and
3549 above, it's not sufficient to pass SSL_OP_NO_TLSv1, one has to pass
3550 SSL_OP_NO_TLSv1|SSL_OP_NO_SSLv3|SSL_OP_NO_SSLv2. This applies to
3554 Changes between 1.0.1 and 1.0.1a [19 Apr 2012]
3556 *) Check for potentially exploitable overflows in asn1_d2i_read_bio
3557 BUF_mem_grow and BUF_mem_grow_clean. Refuse attempts to shrink buffer
3558 in CRYPTO_realloc_clean.
3560 Thanks to Tavis Ormandy, Google Security Team, for discovering this
3561 issue and to Adam Langley <agl@chromium.org> for fixing it.
3563 [Adam Langley (Google), Tavis Ormandy, Google Security Team]
3565 *) Don't allow TLS 1.2 SHA-256 ciphersuites in TLS 1.0, 1.1 connections.
3568 *) Workarounds for some broken servers that "hang" if a client hello
3569 record length exceeds 255 bytes.
3571 1. Do not use record version number > TLS 1.0 in initial client
3572 hello: some (but not all) hanging servers will now work.
3573 2. If we set OPENSSL_MAX_TLS1_2_CIPHER_LENGTH this will truncate
3574 the number of ciphers sent in the client hello. This should be
3575 set to an even number, such as 50, for example by passing:
3576 -DOPENSSL_MAX_TLS1_2_CIPHER_LENGTH=50 to config or Configure.
3577 Most broken servers should now work.
3578 3. If all else fails setting OPENSSL_NO_TLS1_2_CLIENT will disable
3579 TLS 1.2 client support entirely.
3582 *) Fix SEGV in Vector Permutation AES module observed in OpenSSH.
3585 Changes between 1.0.0h and 1.0.1 [14 Mar 2012]
3587 *) Add compatibility with old MDC2 signatures which use an ASN1 OCTET
3588 STRING form instead of a DigestInfo.
3591 *) The format used for MDC2 RSA signatures is inconsistent between EVP
3592 and the RSA_sign/RSA_verify functions. This was made more apparent when
3593 OpenSSL used RSA_sign/RSA_verify for some RSA signatures in particular
3594 those which went through EVP_PKEY_METHOD in 1.0.0 and later. Detect
3595 the correct format in RSA_verify so both forms transparently work.
3598 *) Some servers which support TLS 1.0 can choke if we initially indicate
3599 support for TLS 1.2 and later renegotiate using TLS 1.0 in the RSA
3600 encrypted premaster secret. As a workaround use the maximum permitted
3601 client version in client hello, this should keep such servers happy
3602 and still work with previous versions of OpenSSL.
3605 *) Add support for TLS/DTLS heartbeats.
3606 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3608 *) Add support for SCTP.
3609 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3611 *) Improved PRNG seeding for VOS.
3612 [Paul Green <Paul.Green@stratus.com>]
3614 *) Extensive assembler packs updates, most notably:
3616 - x86[_64]: AES-NI, PCLMULQDQ, RDRAND support;
3617 - x86[_64]: SSSE3 support (SHA1, vector-permutation AES);
3618 - x86_64: bit-sliced AES implementation;
3619 - ARM: NEON support, contemporary platforms optimizations;
3620 - s390x: z196 support;
3621 - *: GHASH and GF(2^m) multiplication implementations;
3625 *) Make TLS-SRP code conformant with RFC 5054 API cleanup
3626 (removal of unnecessary code)
3627 [Peter Sylvester <peter.sylvester@edelweb.fr>]
3629 *) Add TLS key material exporter from RFC 5705.
3632 *) Add DTLS-SRTP negotiation from RFC 5764.
3635 *) Add Next Protocol Negotiation,
3636 http://tools.ietf.org/html/draft-agl-tls-nextprotoneg-00. Can be
3637 disabled with a no-npn flag to config or Configure. Code donated
3639 [Adam Langley <agl@google.com> and Ben Laurie]
3641 *) Add optional 64-bit optimized implementations of elliptic curves NIST-P224,
3642 NIST-P256, NIST-P521, with constant-time single point multiplication on
3643 typical inputs. Compiler support for the nonstandard type __uint128_t is
3644 required to use this (present in gcc 4.4 and later, for 64-bit builds).
3645 Code made available under Apache License version 2.0.
3647 Specify "enable-ec_nistp_64_gcc_128" on the Configure (or config) command
3648 line to include this in your build of OpenSSL, and run "make depend" (or
3649 "make update"). This enables the following EC_METHODs:
3651 EC_GFp_nistp224_method()
3652 EC_GFp_nistp256_method()
3653 EC_GFp_nistp521_method()
3655 EC_GROUP_new_by_curve_name() will automatically use these (while
3656 EC_GROUP_new_curve_GFp() currently prefers the more flexible
3658 [Emilia Käsper, Adam Langley, Bodo Moeller (Google)]
3660 *) Use type ossl_ssize_t instad of ssize_t which isn't available on
3661 all platforms. Move ssize_t definition from e_os.h to the public
3662 header file e_os2.h as it now appears in public header file cms.h
3665 *) New -sigopt option to the ca, req and x509 utilities. Additional
3666 signature parameters can be passed using this option and in
3670 *) Add RSA PSS signing function. This will generate and set the
3671 appropriate AlgorithmIdentifiers for PSS based on those in the
3672 corresponding EVP_MD_CTX structure. No application support yet.
3675 *) Support for companion algorithm specific ASN1 signing routines.
3676 New function ASN1_item_sign_ctx() signs a pre-initialised
3677 EVP_MD_CTX structure and sets AlgorithmIdentifiers based on
3678 the appropriate parameters.
3681 *) Add new algorithm specific ASN1 verification initialisation function
3682 to EVP_PKEY_ASN1_METHOD: this is not in EVP_PKEY_METHOD since the ASN1
3683 handling will be the same no matter what EVP_PKEY_METHOD is used.
3684 Add a PSS handler to support verification of PSS signatures: checked
3685 against a number of sample certificates.
3688 *) Add signature printing for PSS. Add PSS OIDs.
3689 [Steve Henson, Martin Kaiser <lists@kaiser.cx>]
3691 *) Add algorithm specific signature printing. An individual ASN1 method
3692 can now print out signatures instead of the standard hex dump.
3694 More complex signatures (e.g. PSS) can print out more meaningful
3695 information. Include DSA version that prints out the signature
3699 *) Password based recipient info support for CMS library: implementing
3703 *) Split password based encryption into PBES2 and PBKDF2 functions. This
3704 neatly separates the code into cipher and PBE sections and is required
3705 for some algorithms that split PBES2 into separate pieces (such as
3706 password based CMS).
3709 *) Session-handling fixes:
3710 - Fix handling of connections that are resuming with a session ID,
3711 but also support Session Tickets.
3712 - Fix a bug that suppressed issuing of a new ticket if the client
3713 presented a ticket with an expired session.
3714 - Try to set the ticket lifetime hint to something reasonable.
3715 - Make tickets shorter by excluding irrelevant information.
3716 - On the client side, don't ignore renewed tickets.
3717 [Adam Langley, Bodo Moeller (Google)]
3719 *) Fix PSK session representation.
3722 *) Add RC4-MD5 and AESNI-SHA1 "stitched" implementations.
3724 This work was sponsored by Intel.
3727 *) Add GCM support to TLS library. Some custom code is needed to split
3728 the IV between the fixed (from PRF) and explicit (from TLS record)
3729 portions. This adds all GCM ciphersuites supported by RFC5288 and
3730 RFC5289. Generalise some AES* cipherstrings to include GCM and
3731 add a special AESGCM string for GCM only.
3734 *) Expand range of ctrls for AES GCM. Permit setting invocation
3735 field on decrypt and retrieval of invocation field only on encrypt.
3738 *) Add HMAC ECC ciphersuites from RFC5289. Include SHA384 PRF support.
3739 As required by RFC5289 these ciphersuites cannot be used if for
3740 versions of TLS earlier than 1.2.
3743 *) For FIPS capable OpenSSL interpret a NULL default public key method
3744 as unset and return the appropriate default but do *not* set the default.
3745 This means we can return the appropriate method in applications that
3746 switch between FIPS and non-FIPS modes.
3749 *) Redirect HMAC and CMAC operations to FIPS module in FIPS mode. If an
3750 ENGINE is used then we cannot handle that in the FIPS module so we
3751 keep original code iff non-FIPS operations are allowed.
3754 *) Add -attime option to openssl utilities.
3755 [Peter Eckersley <pde@eff.org>, Ben Laurie and Steve Henson]
3757 *) Redirect DSA and DH operations to FIPS module in FIPS mode.
3760 *) Redirect ECDSA and ECDH operations to FIPS module in FIPS mode. Also use
3761 FIPS EC methods unconditionally for now.
3764 *) New build option no-ec2m to disable characteristic 2 code.
3767 *) Backport libcrypto audit of return value checking from 1.1.0-dev; not
3768 all cases can be covered as some introduce binary incompatibilities.
3771 *) Redirect RSA operations to FIPS module including keygen,
3772 encrypt, decrypt, sign and verify. Block use of non FIPS RSA methods.
3775 *) Add similar low level API blocking to ciphers.
3778 *) Low level digest APIs are not approved in FIPS mode: any attempt
3779 to use these will cause a fatal error. Applications that *really* want
3780 to use them can use the private_* version instead.
3783 *) Redirect cipher operations to FIPS module for FIPS builds.
3786 *) Redirect digest operations to FIPS module for FIPS builds.
3789 *) Update build system to add "fips" flag which will link in fipscanister.o
3790 for static and shared library builds embedding a signature if needed.
3793 *) Output TLS supported curves in preference order instead of numerical
3794 order. This is currently hardcoded for the highest order curves first.
3795 This should be configurable so applications can judge speed vs strength.
3798 *) Add TLS v1.2 server support for client authentication.
3801 *) Add support for FIPS mode in ssl library: disable SSLv3, non-FIPS ciphers
3805 *) Functions FIPS_mode_set() and FIPS_mode() which call the underlying
3806 FIPS modules versions.
3809 *) Add TLS v1.2 client side support for client authentication. Keep cache
3810 of handshake records longer as we don't know the hash algorithm to use
3811 until after the certificate request message is received.
3814 *) Initial TLS v1.2 client support. Add a default signature algorithms
3815 extension including all the algorithms we support. Parse new signature
3816 format in client key exchange. Relax some ECC signing restrictions for
3817 TLS v1.2 as indicated in RFC5246.
3820 *) Add server support for TLS v1.2 signature algorithms extension. Switch
3821 to new signature format when needed using client digest preference.
3822 All server ciphersuites should now work correctly in TLS v1.2. No client
3823 support yet and no support for client certificates.
3826 *) Initial TLS v1.2 support. Add new SHA256 digest to ssl code, switch
3827 to SHA256 for PRF when using TLS v1.2 and later. Add new SHA256 based
3828 ciphersuites. At present only RSA key exchange ciphersuites work with
3829 TLS v1.2. Add new option for TLS v1.2 replacing the old and obsolete
3830 SSL_OP_PKCS1_CHECK flags with SSL_OP_NO_TLSv1_2. New TLSv1.2 methods
3831 and version checking.
3834 *) New option OPENSSL_NO_SSL_INTERN. If an application can be compiled
3835 with this defined it will not be affected by any changes to ssl internal
3836 structures. Add several utility functions to allow openssl application
3837 to work with OPENSSL_NO_SSL_INTERN defined.
3840 *) A long standing patch to add support for SRP from EdelWeb (Peter
3841 Sylvester and Christophe Renou) was integrated.
3842 [Christophe Renou <christophe.renou@edelweb.fr>, Peter Sylvester
3843 <peter.sylvester@edelweb.fr>, Tom Wu <tjw@cs.stanford.edu>, and
3846 *) Add functions to copy EVP_PKEY_METHOD and retrieve flags and id.
3849 *) Permit abbreviated handshakes when renegotiating using the function
3850 SSL_renegotiate_abbreviated().
3851 [Robin Seggelmann <seggelmann@fh-muenster.de>]
3853 *) Add call to ENGINE_register_all_complete() to
3854 ENGINE_load_builtin_engines(), so some implementations get used
3855 automatically instead of needing explicit application support.
3858 *) Add support for TLS key exporter as described in RFC5705.
3859 [Robin Seggelmann <seggelmann@fh-muenster.de>, Steve Henson]
3861 *) Initial TLSv1.1 support. Since TLSv1.1 is very similar to TLS v1.0 only
3862 a few changes are required:
3864 Add SSL_OP_NO_TLSv1_1 flag.
3865 Add TLSv1_1 methods.
3866 Update version checking logic to handle version 1.1.
3867 Add explicit IV handling (ported from DTLS code).
3868 Add command line options to s_client/s_server.
3871 Changes between 1.0.0g and 1.0.0h [12 Mar 2012]
3873 *) Fix MMA (Bleichenbacher's attack on PKCS #1 v1.5 RSA padding) weakness
3874 in CMS and PKCS7 code. When RSA decryption fails use a random key for
3875 content decryption and always return the same error. Note: this attack
3876 needs on average 2^20 messages so it only affects automated senders. The
3877 old behaviour can be re-enabled in the CMS code by setting the
3878 CMS_DEBUG_DECRYPT flag: this is useful for debugging and testing where
3879 an MMA defence is not necessary.
3880 Thanks to Ivan Nestlerode <inestlerode@us.ibm.com> for discovering
3881 this issue. (CVE-2012-0884)
3884 *) Fix CVE-2011-4619: make sure we really are receiving a
3885 client hello before rejecting multiple SGC restarts. Thanks to
3886 Ivan Nestlerode <inestlerode@us.ibm.com> for discovering this bug.
3889 Changes between 1.0.0f and 1.0.0g [18 Jan 2012]
3891 *) Fix for DTLS DoS issue introduced by fix for CVE-2011-4109.
3892 Thanks to Antonio Martin, Enterprise Secure Access Research and
3893 Development, Cisco Systems, Inc. for discovering this bug and
3894 preparing a fix. (CVE-2012-0050)
3897 Changes between 1.0.0e and 1.0.0f [4 Jan 2012]
3899 *) Nadhem Alfardan and Kenny Paterson have discovered an extension
3900 of the Vaudenay padding oracle attack on CBC mode encryption
3901 which enables an efficient plaintext recovery attack against
3902 the OpenSSL implementation of DTLS. Their attack exploits timing
3903 differences arising during decryption processing. A research
3904 paper describing this attack can be found at:
3905 http://www.isg.rhul.ac.uk/~kp/dtls.pdf
3906 Thanks go to Nadhem Alfardan and Kenny Paterson of the Information
3907 Security Group at Royal Holloway, University of London
3908 (www.isg.rhul.ac.uk) for discovering this flaw and to Robin Seggelmann
3909 <seggelmann@fh-muenster.de> and Michael Tuexen <tuexen@fh-muenster.de>
3910 for preparing the fix. (CVE-2011-4108)
3911 [Robin Seggelmann, Michael Tuexen]
3913 *) Clear bytes used for block padding of SSL 3.0 records.
3915 [Adam Langley (Google)]
3917 *) Only allow one SGC handshake restart for SSL/TLS. Thanks to George
3918 Kadianakis <desnacked@gmail.com> for discovering this issue and
3919 Adam Langley for preparing the fix. (CVE-2011-4619)
3920 [Adam Langley (Google)]
3922 *) Check parameters are not NULL in GOST ENGINE. (CVE-2012-0027)
3923 [Andrey Kulikov <amdeich@gmail.com>]
3925 *) Prevent malformed RFC3779 data triggering an assertion failure.
3926 Thanks to Andrew Chi, BBN Technologies, for discovering the flaw
3927 and Rob Austein <sra@hactrn.net> for fixing it. (CVE-2011-4577)
3928 [Rob Austein <sra@hactrn.net>]
3930 *) Improved PRNG seeding for VOS.
3931 [Paul Green <Paul.Green@stratus.com>]
3933 *) Fix ssl_ciph.c set-up race.
3934 [Adam Langley (Google)]
3936 *) Fix spurious failures in ecdsatest.c.
3937 [Emilia Käsper (Google)]
3939 *) Fix the BIO_f_buffer() implementation (which was mixing different
3940 interpretations of the '..._len' fields).
3941 [Adam Langley (Google)]
3943 *) Fix handling of BN_BLINDING: now BN_BLINDING_invert_ex (rather than
3944 BN_BLINDING_invert_ex) calls BN_BLINDING_update, ensuring that concurrent
3945 threads won't reuse the same blinding coefficients.
3947 This also avoids the need to obtain the CRYPTO_LOCK_RSA_BLINDING
3948 lock to call BN_BLINDING_invert_ex, and avoids one use of
3949 BN_BLINDING_update for each BN_BLINDING structure (previously,
3950 the last update always remained unused).
3951 [Emilia Käsper (Google)]
3953 *) In ssl3_clear, preserve s3->init_extra along with s3->rbuf.
3954 [Bob Buckholz (Google)]
3956 Changes between 1.0.0d and 1.0.0e [6 Sep 2011]
3958 *) Fix bug where CRLs with nextUpdate in the past are sometimes accepted
3959 by initialising X509_STORE_CTX properly. (CVE-2011-3207)
3960 [Kaspar Brand <ossl@velox.ch>]
3962 *) Fix SSL memory handling for (EC)DH ciphersuites, in particular
3963 for multi-threaded use of ECDH. (CVE-2011-3210)
3964 [Adam Langley (Google)]
3966 *) Fix x509_name_ex_d2i memory leak on bad inputs.
3969 *) Remove hard coded ecdsaWithSHA1 signature tests in ssl code and check
3970 signature public key algorithm by using OID xref utilities instead.
3971 Before this you could only use some ECC ciphersuites with SHA1 only.
3974 *) Add protection against ECDSA timing attacks as mentioned in the paper
3975 by Billy Bob Brumley and Nicola Tuveri, see:
3977 http://eprint.iacr.org/2011/232.pdf
3979 [Billy Bob Brumley and Nicola Tuveri]
3981 Changes between 1.0.0c and 1.0.0d [8 Feb 2011]
3983 *) Fix parsing of OCSP stapling ClientHello extension. CVE-2011-0014
3984 [Neel Mehta, Adam Langley, Bodo Moeller (Google)]
3986 *) Fix bug in string printing code: if *any* escaping is enabled we must
3987 escape the escape character (backslash) or the resulting string is
3991 Changes between 1.0.0b and 1.0.0c [2 Dec 2010]
3993 *) Disable code workaround for ancient and obsolete Netscape browsers
3994 and servers: an attacker can use it in a ciphersuite downgrade attack.
3995 Thanks to Martin Rex for discovering this bug. CVE-2010-4180
3998 *) Fixed J-PAKE implementation error, originally discovered by
3999 Sebastien Martini, further info and confirmation from Stefan
4000 Arentz and Feng Hao. Note that this fix is a security fix. CVE-2010-4252
4003 Changes between 1.0.0a and 1.0.0b [16 Nov 2010]
4005 *) Fix extension code to avoid race conditions which can result in a buffer
4006 overrun vulnerability: resumed sessions must not be modified as they can
4007 be shared by multiple threads. CVE-2010-3864
4010 *) Fix WIN32 build system to correctly link an ENGINE directory into
4014 Changes between 1.0.0 and 1.0.0a [01 Jun 2010]
4016 *) Check return value of int_rsa_verify in pkey_rsa_verifyrecover
4018 [Steve Henson, Peter-Michael Hager <hager@dortmund.net>]
4020 Changes between 0.9.8n and 1.0.0 [29 Mar 2010]
4022 *) Add "missing" function EVP_CIPHER_CTX_copy(). This copies a cipher
4023 context. The operation can be customised via the ctrl mechanism in
4024 case ENGINEs want to include additional functionality.
4027 *) Tolerate yet another broken PKCS#8 key format: private key value negative.
4030 *) Add new -subject_hash_old and -issuer_hash_old options to x509 utility to
4031 output hashes compatible with older versions of OpenSSL.
4032 [Willy Weisz <weisz@vcpc.univie.ac.at>]
4034 *) Fix compression algorithm handling: if resuming a session use the
4035 compression algorithm of the resumed session instead of determining
4036 it from client hello again. Don't allow server to change algorithm.
4039 *) Add load_crls() function to apps tidying load_certs() too. Add option
4040 to verify utility to allow additional CRLs to be included.
4043 *) Update OCSP request code to permit adding custom headers to the request:
4044 some responders need this.
4047 *) The function EVP_PKEY_sign() returns <=0 on error: check return code
4049 [Julia Lawall <julia@diku.dk>]
4051 *) Update verify callback code in apps/s_cb.c and apps/verify.c, it
4052 needlessly dereferenced structures, used obsolete functions and
4053 didn't handle all updated verify codes correctly.
4056 *) Disable MD2 in the default configuration.
4059 *) In BIO_pop() and BIO_push() use the ctrl argument (which was NULL) to
4060 indicate the initial BIO being pushed or popped. This makes it possible
4061 to determine whether the BIO is the one explicitly called or as a result
4062 of the ctrl being passed down the chain. Fix BIO_pop() and SSL BIOs so
4063 it handles reference counts correctly and doesn't zero out the I/O bio
4064 when it is not being explicitly popped. WARNING: applications which
4065 included workarounds for the old buggy behaviour will need to be modified
4066 or they could free up already freed BIOs.
4069 *) Extend the uni2asc/asc2uni => OPENSSL_uni2asc/OPENSSL_asc2uni
4070 renaming to all platforms (within the 0.9.8 branch, this was
4071 done conditionally on Netware platforms to avoid a name clash).
4072 [Guenter <lists@gknw.net>]
4074 *) Add ECDHE and PSK support to DTLS.
4075 [Michael Tuexen <tuexen@fh-muenster.de>]
4077 *) Add CHECKED_STACK_OF macro to safestack.h, otherwise safestack can't
4081 *) Add "missing" function EVP_MD_flags() (without this the only way to
4082 retrieve a digest flags is by accessing the structure directly. Update
4083 EVP_MD_do_all*() and EVP_CIPHER_do_all*() to include the name a digest
4084 or cipher is registered as in the "from" argument. Print out all
4085 registered digests in the dgst usage message instead of manually
4086 attempting to work them out.
4089 *) If no SSLv2 ciphers are used don't use an SSLv2 compatible client hello:
4090 this allows the use of compression and extensions. Change default cipher
4091 string to remove SSLv2 ciphersuites. This effectively avoids ancient SSLv2
4092 by default unless an application cipher string requests it.
4095 *) Alter match criteria in PKCS12_parse(). It used to try to use local
4096 key ids to find matching certificates and keys but some PKCS#12 files
4097 don't follow the (somewhat unwritten) rules and this strategy fails.
4098 Now just gather all certificates together and the first private key
4099 then look for the first certificate that matches the key.
4102 *) Support use of registered digest and cipher names for dgst and cipher
4103 commands instead of having to add each one as a special case. So now
4110 openssl dgst -sha256 foo
4112 and this works for ENGINE based algorithms too.
4116 *) Update Gost ENGINE to support parameter files.
4117 [Victor B. Wagner <vitus@cryptocom.ru>]
4119 *) Support GeneralizedTime in ca utility.
4120 [Oliver Martin <oliver@volatilevoid.net>, Steve Henson]
4122 *) Enhance the hash format used for certificate directory links. The new
4123 form uses the canonical encoding (meaning equivalent names will work
4124 even if they aren't identical) and uses SHA1 instead of MD5. This form
4125 is incompatible with the older format and as a result c_rehash should
4126 be used to rebuild symbolic links.
4129 *) Make PKCS#8 the default write format for private keys, replacing the
4130 traditional format. This form is standardised, more secure and doesn't
4131 include an implicit MD5 dependency.
4134 *) Add a $gcc_devteam_warn option to Configure. The idea is that any code
4135 committed to OpenSSL should pass this lot as a minimum.
4138 *) Add session ticket override functionality for use by EAP-FAST.
4139 [Jouni Malinen <j@w1.fi>]
4141 *) Modify HMAC functions to return a value. Since these can be implemented
4142 in an ENGINE errors can occur.
4145 *) Type-checked OBJ_bsearch_ex.
4148 *) Type-checked OBJ_bsearch. Also some constification necessitated
4149 by type-checking. Still to come: TXT_DB, bsearch(?),
4150 OBJ_bsearch_ex, qsort, CRYPTO_EX_DATA, ASN1_VALUE, ASN1_STRING,
4154 *) New function OPENSSL_gmtime_adj() to add a specific number of days and
4155 seconds to a tm structure directly, instead of going through OS
4156 specific date routines. This avoids any issues with OS routines such
4157 as the year 2038 bug. New *_adj() functions for ASN1 time structures
4158 and X509_time_adj_ex() to cover the extended range. The existing
4159 X509_time_adj() is still usable and will no longer have any date issues.
4162 *) Delta CRL support. New use deltas option which will attempt to locate
4163 and search any appropriate delta CRLs available.
4165 This work was sponsored by Google.
4168 *) Support for CRLs partitioned by reason code. Reorganise CRL processing
4169 code and add additional score elements. Validate alternate CRL paths
4170 as part of the CRL checking and indicate a new error "CRL path validation
4171 error" in this case. Applications wanting additional details can use
4172 the verify callback and check the new "parent" field. If this is not
4173 NULL CRL path validation is taking place. Existing applications won't
4174 see this because it requires extended CRL support which is off by
4177 This work was sponsored by Google.
4180 *) Support for freshest CRL extension.
4182 This work was sponsored by Google.
4185 *) Initial indirect CRL support. Currently only supported in the CRLs
4186 passed directly and not via lookup. Process certificate issuer
4187 CRL entry extension and lookup CRL entries by bother issuer name
4188 and serial number. Check and process CRL issuer entry in IDP extension.
4190 This work was sponsored by Google.
4193 *) Add support for distinct certificate and CRL paths. The CRL issuer
4194 certificate is validated separately in this case. Only enabled if
4195 an extended CRL support flag is set: this flag will enable additional
4196 CRL functionality in future.
4198 This work was sponsored by Google.
4201 *) Add support for policy mappings extension.
4203 This work was sponsored by Google.
4206 *) Fixes to pathlength constraint, self issued certificate handling,
4207 policy processing to align with RFC3280 and PKITS tests.
4209 This work was sponsored by Google.
4212 *) Support for name constraints certificate extension. DN, email, DNS
4213 and URI types are currently supported.
4215 This work was sponsored by Google.
4218 *) To cater for systems that provide a pointer-based thread ID rather
4219 than numeric, deprecate the current numeric thread ID mechanism and
4220 replace it with a structure and associated callback type. This
4221 mechanism allows a numeric "hash" to be extracted from a thread ID in
4222 either case, and on platforms where pointers are larger than 'long',
4223 mixing is done to help ensure the numeric 'hash' is usable even if it
4224 can't be guaranteed unique. The default mechanism is to use "&errno"
4225 as a pointer-based thread ID to distinguish between threads.
4227 Applications that want to provide their own thread IDs should now use
4228 CRYPTO_THREADID_set_callback() to register a callback that will call
4229 either CRYPTO_THREADID_set_numeric() or CRYPTO_THREADID_set_pointer().
4231 Note that ERR_remove_state() is now deprecated, because it is tied
4232 to the assumption that thread IDs are numeric. ERR_remove_state(0)
4233 to free the current thread's error state should be replaced by
4234 ERR_remove_thread_state(NULL).
4236 (This new approach replaces the functions CRYPTO_set_idptr_callback(),
4237 CRYPTO_get_idptr_callback(), and CRYPTO_thread_idptr() that existed in
4238 OpenSSL 0.9.9-dev between June 2006 and August 2008. Also, if an
4239 application was previously providing a numeric thread callback that
4240 was inappropriate for distinguishing threads, then uniqueness might
4241 have been obtained with &errno that happened immediately in the
4242 intermediate development versions of OpenSSL; this is no longer the
4243 case, the numeric thread callback will now override the automatic use
4245 [Geoff Thorpe, with help from Bodo Moeller]
4247 *) Initial support for different CRL issuing certificates. This covers a
4248 simple case where the self issued certificates in the chain exist and
4249 the real CRL issuer is higher in the existing chain.
4251 This work was sponsored by Google.
4254 *) Removed effectively defunct crypto/store from the build.
4257 *) Revamp of STACK to provide stronger type-checking. Still to come:
4258 TXT_DB, bsearch(?), OBJ_bsearch, qsort, CRYPTO_EX_DATA, ASN1_VALUE,
4259 ASN1_STRING, CONF_VALUE.
4262 *) Add a new SSL_MODE_RELEASE_BUFFERS mode flag to release unused buffer
4263 RAM on SSL connections. This option can save about 34k per idle SSL.
4266 *) Revamp of LHASH to provide stronger type-checking. Still to come:
4267 STACK, TXT_DB, bsearch, qsort.
4270 *) Initial support for Cryptographic Message Syntax (aka CMS) based
4271 on RFC3850, RFC3851 and RFC3852. New cms directory and cms utility,
4272 support for data, signedData, compressedData, digestedData and
4273 encryptedData, envelopedData types included. Scripts to check against
4274 RFC4134 examples draft and interop and consistency checks of many
4275 content types and variants.
4278 *) Add options to enc utility to support use of zlib compression BIO.
4281 *) Extend mk1mf to support importing of options and assembly language
4282 files from Configure script, currently only included in VC-WIN32.
4283 The assembly language rules can now optionally generate the source
4284 files from the associated perl scripts.
4287 *) Implement remaining functionality needed to support GOST ciphersuites.
4288 Interop testing has been performed using CryptoPro implementations.
4289 [Victor B. Wagner <vitus@cryptocom.ru>]
4291 *) s390x assembler pack.
4294 *) ARMv4 assembler pack. ARMv4 refers to v4 and later ISA, not CPU
4298 *) Implement Opaque PRF Input TLS extension as specified in
4299 draft-rescorla-tls-opaque-prf-input-00.txt. Since this is not an
4300 official specification yet and no extension type assignment by
4301 IANA exists, this extension (for now) will have to be explicitly
4302 enabled when building OpenSSL by providing the extension number
4303 to use. For example, specify an option
4305 -DTLSEXT_TYPE_opaque_prf_input=0x9527
4307 to the "config" or "Configure" script to enable the extension,
4308 assuming extension number 0x9527 (which is a completely arbitrary
4309 and unofficial assignment based on the MD5 hash of the Internet
4310 Draft). Note that by doing so, you potentially lose
4311 interoperability with other TLS implementations since these might
4312 be using the same extension number for other purposes.
4314 SSL_set_tlsext_opaque_prf_input(ssl, src, len) is used to set the
4315 opaque PRF input value to use in the handshake. This will create
4316 an internal copy of the length-'len' string at 'src', and will
4317 return non-zero for success.
4319 To get more control and flexibility, provide a callback function
4322 SSL_CTX_set_tlsext_opaque_prf_input_callback(ctx, cb)
4323 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg(ctx, arg)
4327 int (*cb)(SSL *, void *peerinput, size_t len, void *arg);
4330 Callback function 'cb' will be called in handshakes, and is
4331 expected to use SSL_set_tlsext_opaque_prf_input() as appropriate.
4332 Argument 'arg' is for application purposes (the value as given to
4333 SSL_CTX_set_tlsext_opaque_prf_input_callback_arg() will directly
4334 be provided to the callback function). The callback function
4335 has to return non-zero to report success: usually 1 to use opaque
4336 PRF input just if possible, or 2 to enforce use of the opaque PRF
4337 input. In the latter case, the library will abort the handshake
4338 if opaque PRF input is not successfully negotiated.
4340 Arguments 'peerinput' and 'len' given to the callback function
4341 will always be NULL and 0 in the case of a client. A server will
4342 see the client's opaque PRF input through these variables if
4343 available (NULL and 0 otherwise). Note that if the server
4344 provides an opaque PRF input, the length must be the same as the
4345 length of the client's opaque PRF input.
4347 Note that the callback function will only be called when creating
4348 a new session (session resumption can resume whatever was
4349 previously negotiated), and will not be called in SSL 2.0
4350 handshakes; thus, SSL_CTX_set_options(ctx, SSL_OP_NO_SSLv2) or
4351 SSL_set_options(ssl, SSL_OP_NO_SSLv2) is especially recommended
4352 for applications that need to enforce opaque PRF input.
4356 *) Update ssl code to support digests other than SHA1+MD5 for handshake
4359 [Victor B. Wagner <vitus@cryptocom.ru>]
4361 *) Add RFC4507 support to OpenSSL. This includes the corrections in
4362 RFC4507bis. The encrypted ticket format is an encrypted encoded
4363 SSL_SESSION structure, that way new session features are automatically
4366 If a client application caches session in an SSL_SESSION structure
4367 support is transparent because tickets are now stored in the encoded
4370 The SSL_CTX structure automatically generates keys for ticket
4371 protection in servers so again support should be possible
4372 with no application modification.
4374 If a client or server wishes to disable RFC4507 support then the option
4375 SSL_OP_NO_TICKET can be set.
4377 Add a TLS extension debugging callback to allow the contents of any client
4378 or server extensions to be examined.
4380 This work was sponsored by Google.
4383 *) Final changes to avoid use of pointer pointer casts in OpenSSL.
4384 OpenSSL should now compile cleanly on gcc 4.2
4385 [Peter Hartley <pdh@utter.chaos.org.uk>, Steve Henson]
4387 *) Update SSL library to use new EVP_PKEY MAC API. Include generic MAC
4388 support including streaming MAC support: this is required for GOST
4389 ciphersuite support.
4390 [Victor B. Wagner <vitus@cryptocom.ru>, Steve Henson]
4392 *) Add option -stream to use PKCS#7 streaming in smime utility. New
4393 function i2d_PKCS7_bio_stream() and PEM_write_PKCS7_bio_stream()
4394 to output in BER and PEM format.
4397 *) Experimental support for use of HMAC via EVP_PKEY interface. This
4398 allows HMAC to be handled via the EVP_DigestSign*() interface. The
4399 EVP_PKEY "key" in this case is the HMAC key, potentially allowing
4400 ENGINE support for HMAC keys which are unextractable. New -mac and
4401 -macopt options to dgst utility.
4404 *) New option -sigopt to dgst utility. Update dgst to use
4405 EVP_Digest{Sign,Verify}*. These two changes make it possible to use
4406 alternative signing parameters such as X9.31 or PSS in the dgst
4410 *) Change ssl_cipher_apply_rule(), the internal function that does
4411 the work each time a ciphersuite string requests enabling
4412 ("foo+bar"), moving ("+foo+bar"), disabling ("-foo+bar", or
4413 removing ("!foo+bar") a class of ciphersuites: Now it maintains
4414 the order of disabled ciphersuites such that those ciphersuites
4415 that most recently went from enabled to disabled not only stay
4416 in order with respect to each other, but also have higher priority
4417 than other disabled ciphersuites the next time ciphersuites are
4420 This means that you can now say, e.g., "PSK:-PSK:HIGH" to enable
4421 the same ciphersuites as with "HIGH" alone, but in a specific
4422 order where the PSK ciphersuites come first (since they are the
4423 most recently disabled ciphersuites when "HIGH" is parsed).
4425 Also, change ssl_create_cipher_list() (using this new
4426 functionality) such that between otherwise identical
4427 ciphersuites, ephemeral ECDH is preferred over ephemeral DH in
4431 *) Change ssl_create_cipher_list() so that it automatically
4432 arranges the ciphersuites in reasonable order before starting
4433 to process the rule string. Thus, the definition for "DEFAULT"
4434 (SSL_DEFAULT_CIPHER_LIST) now is just "ALL:!aNULL:!eNULL", but
4435 remains equivalent to "AES:ALL:!aNULL:!eNULL:+aECDH:+kRSA:+RC4:@STRENGTH".
4436 This makes it much easier to arrive at a reasonable default order
4437 in applications for which anonymous ciphers are OK (meaning
4438 that you can't actually use DEFAULT).
4439 [Bodo Moeller; suggested by Victor Duchovni]
4441 *) Split the SSL/TLS algorithm mask (as used for ciphersuite string
4442 processing) into multiple integers instead of setting
4443 "SSL_MKEY_MASK" bits, "SSL_AUTH_MASK" bits, "SSL_ENC_MASK",
4444 "SSL_MAC_MASK", and "SSL_SSL_MASK" bits all in a single integer.
4445 (These masks as well as the individual bit definitions are hidden
4446 away into the non-exported interface ssl/ssl_locl.h, so this
4447 change to the definition of the SSL_CIPHER structure shouldn't
4448 affect applications.) This give us more bits for each of these
4449 categories, so there is no longer a need to coagulate AES128 and
4450 AES256 into a single algorithm bit, and to coagulate Camellia128
4451 and Camellia256 into a single algorithm bit, which has led to all
4454 Thus, among other things, the kludge introduced in 0.9.7m and
4455 0.9.8e for masking out AES256 independently of AES128 or masking
4456 out Camellia256 independently of AES256 is not needed here in 0.9.9.
4458 With the change, we also introduce new ciphersuite aliases that
4459 so far were missing: "AES128", "AES256", "CAMELLIA128", and
4463 *) Add support for dsa-with-SHA224 and dsa-with-SHA256.
4464 Use the leftmost N bytes of the signature input if the input is
4465 larger than the prime q (with N being the size in bytes of q).
4468 *) Very *very* experimental PKCS#7 streaming encoder support. Nothing uses
4469 it yet and it is largely untested.
4472 *) Add support for the ecdsa-with-SHA224/256/384/512 signature types.
4475 *) Initial incomplete changes to avoid need for function casts in OpenSSL
4476 some compilers (gcc 4.2 and later) reject their use. Safestack is
4477 reimplemented. Update ASN1 to avoid use of legacy functions.
4480 *) Win32/64 targets are linked with Winsock2.
4483 *) Add an X509_CRL_METHOD structure to allow CRL processing to be redirected
4484 to external functions. This can be used to increase CRL handling
4485 efficiency especially when CRLs are very large by (for example) storing
4486 the CRL revoked certificates in a database.
4489 *) Overhaul of by_dir code. Add support for dynamic loading of CRLs so
4490 new CRLs added to a directory can be used. New command line option
4491 -verify_return_error to s_client and s_server. This causes real errors
4492 to be returned by the verify callback instead of carrying on no matter
4493 what. This reflects the way a "real world" verify callback would behave.
4496 *) GOST engine, supporting several GOST algorithms and public key formats.
4497 Kindly donated by Cryptocom.
4500 *) Partial support for Issuing Distribution Point CRL extension. CRLs
4501 partitioned by DP are handled but no indirect CRL or reason partitioning
4502 (yet). Complete overhaul of CRL handling: now the most suitable CRL is
4503 selected via a scoring technique which handles IDP and AKID in CRLs.
4506 *) New X509_STORE_CTX callbacks lookup_crls() and lookup_certs() which
4507 will ultimately be used for all verify operations: this will remove the
4508 X509_STORE dependency on certificate verification and allow alternative
4509 lookup methods. X509_STORE based implementations of these two callbacks.
4512 *) Allow multiple CRLs to exist in an X509_STORE with matching issuer names.
4513 Modify get_crl() to find a valid (unexpired) CRL if possible.
4516 *) New function X509_CRL_match() to check if two CRLs are identical. Normally
4517 this would be called X509_CRL_cmp() but that name is already used by
4518 a function that just compares CRL issuer names. Cache several CRL
4519 extensions in X509_CRL structure and cache CRLDP in X509.
4522 *) Store a "canonical" representation of X509_NAME structure (ASN1 Name)
4523 this maps equivalent X509_NAME structures into a consistent structure.
4524 Name comparison can then be performed rapidly using memcmp().
4527 *) Non-blocking OCSP request processing. Add -timeout option to ocsp
4531 *) Allow digests to supply their own micalg string for S/MIME type using
4532 the ctrl EVP_MD_CTRL_MICALG.
4535 *) During PKCS7 signing pass the PKCS7 SignerInfo structure to the
4536 EVP_PKEY_METHOD before and after signing via the EVP_PKEY_CTRL_PKCS7_SIGN
4537 ctrl. It can then customise the structure before and/or after signing
4541 *) New function OBJ_add_sigid() to allow application defined signature OIDs
4542 to be added to OpenSSLs internal tables. New function OBJ_sigid_free()
4543 to free up any added signature OIDs.
4546 *) New functions EVP_CIPHER_do_all(), EVP_CIPHER_do_all_sorted(),
4547 EVP_MD_do_all() and EVP_MD_do_all_sorted() to enumerate internal
4548 digest and cipher tables. New options added to openssl utility:
4549 list-message-digest-algorithms and list-cipher-algorithms.
4552 *) Change the array representation of binary polynomials: the list
4553 of degrees of non-zero coefficients is now terminated with -1.
4554 Previously it was terminated with 0, which was also part of the
4555 value; thus, the array representation was not applicable to
4556 polynomials where t^0 has coefficient zero. This change makes
4557 the array representation useful in a more general context.
4560 *) Various modifications and fixes to SSL/TLS cipher string
4561 handling. For ECC, the code now distinguishes between fixed ECDH
4562 with RSA certificates on the one hand and with ECDSA certificates
4563 on the other hand, since these are separate ciphersuites. The
4564 unused code for Fortezza ciphersuites has been removed.
4566 For consistency with EDH, ephemeral ECDH is now called "EECDH"
4567 (not "ECDHE"). For consistency with the code for DH
4568 certificates, use of ECDH certificates is now considered ECDH
4569 authentication, not RSA or ECDSA authentication (the latter is
4570 merely the CA's signing algorithm and not actively used in the
4573 The temporary ciphersuite alias "ECCdraft" is no longer
4574 available, and ECC ciphersuites are no longer excluded from "ALL"
4575 and "DEFAULT". The following aliases now exist for RFC 4492
4576 ciphersuites, most of these by analogy with the DH case:
4578 kECDHr - ECDH cert, signed with RSA
4579 kECDHe - ECDH cert, signed with ECDSA
4580 kECDH - ECDH cert (signed with either RSA or ECDSA)
4581 kEECDH - ephemeral ECDH
4582 ECDH - ECDH cert or ephemeral ECDH
4588 AECDH - anonymous ECDH
4589 EECDH - non-anonymous ephemeral ECDH (equivalent to "kEECDH:-AECDH")
4593 *) Add additional S/MIME capabilities for AES and GOST ciphers if supported.
4594 Use correct micalg parameters depending on digest(s) in signed message.
4597 *) Add engine support for EVP_PKEY_ASN1_METHOD. Add functions to process
4598 an ENGINE asn1 method. Support ENGINE lookups in the ASN1 code.
4601 *) Initial engine support for EVP_PKEY_METHOD. New functions to permit
4602 an engine to register a method. Add ENGINE lookups for methods and
4603 functional reference processing.
4606 *) New functions EVP_Digest{Sign,Verify)*. These are enhanced versions of
4607 EVP_{Sign,Verify}* which allow an application to customise the signature
4611 *) New -resign option to smime utility. This adds one or more signers
4612 to an existing PKCS#7 signedData structure. Also -md option to use an
4613 alternative message digest algorithm for signing.
4616 *) Tidy up PKCS#7 routines and add new functions to make it easier to
4617 create PKCS7 structures containing multiple signers. Update smime
4618 application to support multiple signers.
4621 *) New -macalg option to pkcs12 utility to allow setting of an alternative
4625 *) Initial support for PKCS#5 v2.0 PRFs other than default SHA1 HMAC.
4626 Reorganize PBE internals to lookup from a static table using NIDs,
4627 add support for HMAC PBE OID translation. Add a EVP_CIPHER ctrl:
4628 EVP_CTRL_PBE_PRF_NID this allows a cipher to specify an alternative
4629 PRF which will be automatically used with PBES2.
4632 *) Replace the algorithm specific calls to generate keys in "req" with the
4636 *) Update PKCS#7 enveloped data routines to use new API. This is now
4637 supported by any public key method supporting the encrypt operation. A
4638 ctrl is added to allow the public key algorithm to examine or modify
4639 the PKCS#7 RecipientInfo structure if it needs to: for RSA this is
4643 *) Add a ctrl to asn1 method to allow a public key algorithm to express
4644 a default digest type to use. In most cases this will be SHA1 but some
4645 algorithms (such as GOST) need to specify an alternative digest. The
4646 return value indicates how strong the preference is 1 means optional and
4647 2 is mandatory (that is it is the only supported type). Modify
4648 ASN1_item_sign() to accept a NULL digest argument to indicate it should
4649 use the default md. Update openssl utilities to use the default digest
4650 type for signing if it is not explicitly indicated.
4653 *) Use OID cross reference table in ASN1_sign() and ASN1_verify(). New
4654 EVP_MD flag EVP_MD_FLAG_PKEY_METHOD_SIGNATURE. This uses the relevant
4655 signing method from the key type. This effectively removes the link
4656 between digests and public key types.
4659 *) Add an OID cross reference table and utility functions. Its purpose is to
4660 translate between signature OIDs such as SHA1WithrsaEncryption and SHA1,
4661 rsaEncryption. This will allow some of the algorithm specific hackery
4662 needed to use the correct OID to be removed.
4665 *) Remove algorithm specific dependencies when setting PKCS7_SIGNER_INFO
4666 structures for PKCS7_sign(). They are now set up by the relevant public
4670 *) Add provisional EC pkey method with support for ECDSA and ECDH.
4673 *) Add support for key derivation (agreement) in the API, DH method and
4677 *) Add DSA pkey method and DH pkey methods, extend DH ASN1 method to support
4678 public and private key formats. As a side effect these add additional
4679 command line functionality not previously available: DSA signatures can be
4680 generated and verified using pkeyutl and DH key support and generation in
4685 [Oliver Tappe <zooey@hirschkaefer.de>]
4687 *) New make target "install_html_docs" installs HTML renditions of the
4689 [Oliver Tappe <zooey@hirschkaefer.de>]
4691 *) New utility "genpkey" this is analogous to "genrsa" etc except it can
4692 generate keys for any algorithm. Extend and update EVP_PKEY_METHOD to
4693 support key and parameter generation and add initial key generation
4694 functionality for RSA.
4697 *) Add functions for main EVP_PKEY_method operations. The undocumented
4698 functions EVP_PKEY_{encrypt,decrypt} have been renamed to
4699 EVP_PKEY_{encrypt,decrypt}_old.
4702 *) Initial definitions for EVP_PKEY_METHOD. This will be a high level public
4703 key API, doesn't do much yet.
4706 *) New function EVP_PKEY_asn1_get0_info() to retrieve information about
4707 public key algorithms. New option to openssl utility:
4708 "list-public-key-algorithms" to print out info.
4711 *) Implement the Supported Elliptic Curves Extension for
4712 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4715 *) Don't free up OIDs in OBJ_cleanup() if they are in use by EVP_MD or
4716 EVP_CIPHER structures to avoid later problems in EVP_cleanup().
4719 *) New utilities pkey and pkeyparam. These are similar to algorithm specific
4720 utilities such as rsa, dsa, dsaparam etc except they process any key
4724 *) Transfer public key printing routines to EVP_PKEY_ASN1_METHOD. New
4725 functions EVP_PKEY_print_public(), EVP_PKEY_print_private(),
4726 EVP_PKEY_print_param() to print public key data from an EVP_PKEY
4730 *) Initial support for pluggable public key ASN1.
4731 De-spaghettify the public key ASN1 handling. Move public and private
4732 key ASN1 handling to a new EVP_PKEY_ASN1_METHOD structure. Relocate
4733 algorithm specific handling to a single module within the relevant
4734 algorithm directory. Add functions to allow (near) opaque processing
4735 of public and private key structures.
4738 *) Implement the Supported Point Formats Extension for
4739 ECC ciphersuites from draft-ietf-tls-ecc-12.txt.
4742 *) Add initial support for RFC 4279 PSK TLS ciphersuites. Add members
4743 for the psk identity [hint] and the psk callback functions to the
4744 SSL_SESSION, SSL and SSL_CTX structure.
4747 PSK-RC4-SHA, PSK-3DES-EDE-CBC-SHA, PSK-AES128-CBC-SHA,
4751 SSL_CTX_use_psk_identity_hint
4752 SSL_get_psk_identity_hint
4753 SSL_get_psk_identity
4754 SSL_use_psk_identity_hint
4756 [Mika Kousa and Pasi Eronen of Nokia Corporation]
4758 *) Add RFC 3161 compliant time stamp request creation, response generation
4759 and response verification functionality.
4760 [Zoltán Glózik <zglozik@opentsa.org>, The OpenTSA Project]
4762 *) Add initial support for TLS extensions, specifically for the server_name
4763 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
4764 have new members for a host name. The SSL data structure has an
4765 additional member SSL_CTX *initial_ctx so that new sessions can be
4766 stored in that context to allow for session resumption, even after the
4767 SSL has been switched to a new SSL_CTX in reaction to a client's
4768 server_name extension.
4770 New functions (subject to change):
4772 SSL_get_servername()
4773 SSL_get_servername_type()
4776 New CTRL codes and macros (subject to change):
4778 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
4779 - SSL_CTX_set_tlsext_servername_callback()
4780 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
4781 - SSL_CTX_set_tlsext_servername_arg()
4782 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
4784 openssl s_client has a new '-servername ...' option.
4786 openssl s_server has new options '-servername_host ...', '-cert2 ...',
4787 '-key2 ...', '-servername_fatal' (subject to change). This allows
4788 testing the HostName extension for a specific single host name ('-cert'
4789 and '-key' remain fallbacks for handshakes without HostName
4790 negotiation). If the unrecognized_name alert has to be sent, this by
4791 default is a warning; it becomes fatal with the '-servername_fatal'
4794 [Peter Sylvester, Remy Allais, Christophe Renou]
4796 *) Whirlpool hash implementation is added.
4799 *) BIGNUM code on 64-bit SPARCv9 targets is switched from bn(64,64) to
4800 bn(64,32). Because of instruction set limitations it doesn't have
4801 any negative impact on performance. This was done mostly in order
4802 to make it possible to share assembler modules, such as bn_mul_mont
4803 implementations, between 32- and 64-bit builds without hassle.
4806 *) Move code previously exiled into file crypto/ec/ec2_smpt.c
4807 to ec2_smpl.c, and no longer require the OPENSSL_EC_BIN_PT_COMP
4811 *) New candidate for BIGNUM assembler implementation, bn_mul_mont,
4812 dedicated Montgomery multiplication procedure, is introduced.
4813 BN_MONT_CTX is modified to allow bn_mul_mont to reach for higher
4814 "64-bit" performance on certain 32-bit targets.
4817 *) New option SSL_OP_NO_COMP to disable use of compression selectively
4818 in SSL structures. New SSL ctrl to set maximum send fragment size.
4819 Save memory by setting the I/O buffer sizes dynamically instead of
4820 using the maximum available value.
4823 *) New option -V for 'openssl ciphers'. This prints the ciphersuite code
4824 in addition to the text details.
4827 *) Very, very preliminary EXPERIMENTAL support for printing of general
4828 ASN1 structures. This currently produces rather ugly output and doesn't
4829 handle several customised structures at all.
4832 *) Integrated support for PVK file format and some related formats such
4833 as MS PUBLICKEYBLOB and PRIVATEKEYBLOB. Command line switches to support
4834 these in the 'rsa' and 'dsa' utilities.
4837 *) Support for PKCS#1 RSAPublicKey format on rsa utility command line.
4840 *) Remove the ancient ASN1_METHOD code. This was only ever used in one
4841 place for the (very old) "NETSCAPE" format certificates which are now
4842 handled using new ASN1 code equivalents.
4845 *) Let the TLSv1_method() etc. functions return a 'const' SSL_METHOD
4846 pointer and make the SSL_METHOD parameter in SSL_CTX_new,
4847 SSL_CTX_set_ssl_version and SSL_set_ssl_method 'const'.
4850 *) Modify CRL distribution points extension code to print out previously
4851 unsupported fields. Enhance extension setting code to allow setting of
4855 *) Add print and set support for Issuing Distribution Point CRL extension.
4858 *) Change 'Configure' script to enable Camellia by default.
4861 Changes between 0.9.8m and 0.9.8n [24 Mar 2010]
4863 *) When rejecting SSL/TLS records due to an incorrect version number, never
4864 update s->server with a new major version number. As of
4865 - OpenSSL 0.9.8m if 'short' is a 16-bit type,
4866 - OpenSSL 0.9.8f if 'short' is longer than 16 bits,
4867 the previous behavior could result in a read attempt at NULL when
4868 receiving specific incorrect SSL/TLS records once record payload
4869 protection is active. (CVE-2010-0740)
4870 [Bodo Moeller, Adam Langley <agl@chromium.org>]
4872 *) Fix for CVE-2010-0433 where some kerberos enabled versions of OpenSSL
4873 could be crashed if the relevant tables were not present (e.g. chrooted).
4874 [Tomas Hoger <thoger@redhat.com>]
4876 Changes between 0.9.8l and 0.9.8m [25 Feb 2010]
4878 *) Always check bn_wexpand() return values for failure. (CVE-2009-3245)
4879 [Martin Olsson, Neel Mehta]
4881 *) Fix X509_STORE locking: Every 'objs' access requires a lock (to
4882 accommodate for stack sorting, always a write lock!).
4885 *) On some versions of WIN32 Heap32Next is very slow. This can cause
4886 excessive delays in the RAND_poll(): over a minute. As a workaround
4887 include a time check in the inner Heap32Next loop too.
4890 *) The code that handled flushing of data in SSL/TLS originally used the
4891 BIO_CTRL_INFO ctrl to see if any data was pending first. This caused
4892 the problem outlined in PR#1949. The fix suggested there however can
4893 trigger problems with buggy BIO_CTRL_WPENDING (e.g. some versions
4894 of Apache). So instead simplify the code to flush unconditionally.
4895 This should be fine since flushing with no data to flush is a no op.
4898 *) Handle TLS versions 2.0 and later properly and correctly use the
4899 highest version of TLS/SSL supported. Although TLS >= 2.0 is some way
4900 off ancient servers have a habit of sticking around for a while...
4903 *) Modify compression code so it frees up structures without using the
4904 ex_data callbacks. This works around a problem where some applications
4905 call CRYPTO_cleanup_all_ex_data() before application exit (e.g. when
4906 restarting) then use compression (e.g. SSL with compression) later.
4907 This results in significant per-connection memory leaks and
4908 has caused some security issues including CVE-2008-1678 and
4912 *) Constify crypto/cast (i.e., <openssl/cast.h>): a CAST_KEY doesn't
4913 change when encrypting or decrypting.
4916 *) Add option SSL_OP_LEGACY_SERVER_CONNECT which will allow clients to
4917 connect and renegotiate with servers which do not support RI.
4918 Until RI is more widely deployed this option is enabled by default.
4921 *) Add "missing" ssl ctrls to clear options and mode.
4924 *) If client attempts to renegotiate and doesn't support RI respond with
4925 a no_renegotiation alert as required by RFC5746. Some renegotiating
4926 TLS clients will continue a connection gracefully when they receive
4927 the alert. Unfortunately OpenSSL mishandled this alert and would hang
4928 waiting for a server hello which it will never receive. Now we treat a
4929 received no_renegotiation alert as a fatal error. This is because
4930 applications requesting a renegotiation might well expect it to succeed
4931 and would have no code in place to handle the server denying it so the
4932 only safe thing to do is to terminate the connection.
4935 *) Add ctrl macro SSL_get_secure_renegotiation_support() which returns 1 if
4936 peer supports secure renegotiation and 0 otherwise. Print out peer
4937 renegotiation support in s_client/s_server.
4940 *) Replace the highly broken and deprecated SPKAC certification method with
4941 the updated NID creation version. This should correctly handle UTF8.
4944 *) Implement RFC5746. Re-enable renegotiation but require the extension
4945 as needed. Unfortunately, SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION
4946 turns out to be a bad idea. It has been replaced by
4947 SSL_OP_ALLOW_UNSAFE_LEGACY_RENEGOTIATION which can be set with
4948 SSL_CTX_set_options(). This is really not recommended unless you
4949 know what you are doing.
4950 [Eric Rescorla <ekr@networkresonance.com>, Ben Laurie, Steve Henson]
4952 *) Fixes to stateless session resumption handling. Use initial_ctx when
4953 issuing and attempting to decrypt tickets in case it has changed during
4954 servername handling. Use a non-zero length session ID when attempting
4955 stateless session resumption: this makes it possible to determine if
4956 a resumption has occurred immediately after receiving server hello
4957 (several places in OpenSSL subtly assume this) instead of later in
4961 *) The functions ENGINE_ctrl(), OPENSSL_isservice(),
4962 CMS_get1_RecipientRequest() and RAND_bytes() can return <=0 on error
4963 fixes for a few places where the return code is not checked
4965 [Julia Lawall <julia@diku.dk>]
4967 *) Add --strict-warnings option to Configure script to include devteam
4968 warnings in other configurations.
4971 *) Add support for --libdir option and LIBDIR variable in makefiles. This
4972 makes it possible to install openssl libraries in locations which
4973 have names other than "lib", for example "/usr/lib64" which some
4975 [Steve Henson, based on patch from Jeremy Utley]
4977 *) Don't allow the use of leading 0x80 in OIDs. This is a violation of
4978 X690 8.9.12 and can produce some misleading textual output of OIDs.
4979 [Steve Henson, reported by Dan Kaminsky]
4981 *) Delete MD2 from algorithm tables. This follows the recommendation in
4982 several standards that it is not used in new applications due to
4983 several cryptographic weaknesses. For binary compatibility reasons
4984 the MD2 API is still compiled in by default.
4987 *) Add compression id to {d2i,i2d}_SSL_SESSION so it is correctly saved
4991 *) Rename uni2asc and asc2uni functions to OPENSSL_uni2asc and
4992 OPENSSL_asc2uni conditionally on Netware platforms to avoid a name
4994 [Guenter <lists@gknw.net>]
4996 *) Fix the server certificate chain building code to use X509_verify_cert(),
4997 it used to have an ad-hoc builder which was unable to cope with anything
4998 other than a simple chain.
4999 [David Woodhouse <dwmw2@infradead.org>, Steve Henson]
5001 *) Don't check self signed certificate signatures in X509_verify_cert()
5002 by default (a flag can override this): it just wastes time without
5003 adding any security. As a useful side effect self signed root CAs
5004 with non-FIPS digests are now usable in FIPS mode.
5007 *) In dtls1_process_out_of_seq_message() the check if the current message
5008 is already buffered was missing. For every new message was memory
5009 allocated, allowing an attacker to perform an denial of service attack
5010 with sending out of seq handshake messages until there is no memory
5011 left. Additionally every future message was buffered, even if the
5012 sequence number made no sense and would be part of another handshake.
5013 So only messages with sequence numbers less than 10 in advance will be
5014 buffered. (CVE-2009-1378)
5015 [Robin Seggelmann, discovered by Daniel Mentz]
5017 *) Records are buffered if they arrive with a future epoch to be
5018 processed after finishing the corresponding handshake. There is
5019 currently no limitation to this buffer allowing an attacker to perform
5020 a DOS attack with sending records with future epochs until there is no
5021 memory left. This patch adds the pqueue_size() function to determine
5022 the size of a buffer and limits the record buffer to 100 entries.
5024 [Robin Seggelmann, discovered by Daniel Mentz]
5026 *) Keep a copy of frag->msg_header.frag_len so it can be used after the
5027 parent structure is freed. (CVE-2009-1379)
5030 *) Handle non-blocking I/O properly in SSL_shutdown() call.
5031 [Darryl Miles <darryl-mailinglists@netbauds.net>]
5034 [Ilya O. <vrghost@gmail.com>]
5036 Changes between 0.9.8k and 0.9.8l [5 Nov 2009]
5038 *) Disable renegotiation completely - this fixes a severe security
5039 problem (CVE-2009-3555) at the cost of breaking all
5040 renegotiation. Renegotiation can be re-enabled by setting
5041 SSL3_FLAGS_ALLOW_UNSAFE_LEGACY_RENEGOTIATION in s3->flags at
5042 run-time. This is really not recommended unless you know what
5046 Changes between 0.9.8j and 0.9.8k [25 Mar 2009]
5048 *) Don't set val to NULL when freeing up structures, it is freed up by
5049 underlying code. If sizeof(void *) > sizeof(long) this can result in
5050 zeroing past the valid field. (CVE-2009-0789)
5051 [Paolo Ganci <Paolo.Ganci@AdNovum.CH>]
5053 *) Fix bug where return value of CMS_SignerInfo_verify_content() was not
5054 checked correctly. This would allow some invalid signed attributes to
5055 appear to verify correctly. (CVE-2009-0591)
5056 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5058 *) Reject UniversalString and BMPString types with invalid lengths. This
5059 prevents a crash in ASN1_STRING_print_ex() which assumes the strings have
5060 a legal length. (CVE-2009-0590)
5063 *) Set S/MIME signing as the default purpose rather than setting it
5064 unconditionally. This allows applications to override it at the store
5068 *) Permit restricted recursion of ASN1 strings. This is needed in practice
5069 to handle some structures.
5072 *) Improve efficiency of mem_gets: don't search whole buffer each time
5074 [Jeremy Shapiro <jnshapir@us.ibm.com>]
5076 *) New -hex option for openssl rand.
5079 *) Print out UTF8String and NumericString when parsing ASN1.
5082 *) Support NumericString type for name components.
5085 *) Allow CC in the environment to override the automatically chosen
5086 compiler. Note that nothing is done to ensure flags work with the
5090 Changes between 0.9.8i and 0.9.8j [07 Jan 2009]
5092 *) Properly check EVP_VerifyFinal() and similar return values
5094 [Ben Laurie, Bodo Moeller, Google Security Team]
5096 *) Enable TLS extensions by default.
5099 *) Allow the CHIL engine to be loaded, whether the application is
5100 multithreaded or not. (This does not release the developer from the
5101 obligation to set up the dynamic locking callbacks.)
5102 [Sander Temme <sander@temme.net>]
5104 *) Use correct exit code if there is an error in dgst command.
5105 [Steve Henson; problem pointed out by Roland Dirlewanger]
5107 *) Tweak Configure so that you need to say "experimental-jpake" to enable
5108 JPAKE, and need to use -DOPENSSL_EXPERIMENTAL_JPAKE in applications.
5111 *) Add experimental JPAKE support, including demo authentication in
5112 s_client and s_server.
5115 *) Set the comparison function in v3_addr_canonize().
5116 [Rob Austein <sra@hactrn.net>]
5118 *) Add support for XMPP STARTTLS in s_client.
5119 [Philip Paeps <philip@freebsd.org>]
5121 *) Change the server-side SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG behavior
5122 to ensure that even with this option, only ciphersuites in the
5123 server's preference list will be accepted. (Note that the option
5124 applies only when resuming a session, so the earlier behavior was
5125 just about the algorithm choice for symmetric cryptography.)
5128 Changes between 0.9.8h and 0.9.8i [15 Sep 2008]
5130 *) Fix NULL pointer dereference if a DTLS server received
5131 ChangeCipherSpec as first record (CVE-2009-1386).
5134 *) Fix a state transition in s3_srvr.c and d1_srvr.c
5135 (was using SSL3_ST_CW_CLNT_HELLO_B, should be ..._ST_SW_SRVR_...).
5138 *) The fix in 0.9.8c that supposedly got rid of unsafe
5139 double-checked locking was incomplete for RSA blinding,
5140 addressing just one layer of what turns out to have been
5141 doubly unsafe triple-checked locking.
5143 So now fix this for real by retiring the MONT_HELPER macro
5144 in crypto/rsa/rsa_eay.c.
5146 [Bodo Moeller; problem pointed out by Marius Schilder]
5148 *) Various precautionary measures:
5150 - Avoid size_t integer overflow in HASH_UPDATE (md32_common.h).
5152 - Avoid a buffer overflow in d2i_SSL_SESSION() (ssl_asn1.c).
5153 (NB: This would require knowledge of the secret session ticket key
5154 to exploit, in which case you'd be SOL either way.)
5156 - Change bn_nist.c so that it will properly handle input BIGNUMs
5157 outside the expected range.
5159 - Enforce the 'num' check in BN_div() (bn_div.c) for non-BN_DEBUG
5162 [Neel Mehta, Bodo Moeller]
5164 *) Allow engines to be "soft loaded" - i.e. optionally don't die if
5165 the load fails. Useful for distros.
5166 [Ben Laurie and the FreeBSD team]
5168 *) Add support for Local Machine Keyset attribute in PKCS#12 files.
5171 *) Fix BN_GF2m_mod_arr() top-bit cleanup code.
5174 *) Expand ENGINE to support engine supplied SSL client certificate functions.
5176 This work was sponsored by Logica.
5179 *) Add CryptoAPI ENGINE to support use of RSA and DSA keys held in Windows
5180 keystores. Support for SSL/TLS client authentication too.
5181 Not compiled unless enable-capieng specified to Configure.
5183 This work was sponsored by Logica.
5186 *) Fix bug in X509_ATTRIBUTE creation: don't set attribute using
5187 ASN1_TYPE_set1 if MBSTRING flag set. This bug would crash certain
5188 attribute creation routines such as certificate requests and PKCS#12
5192 Changes between 0.9.8g and 0.9.8h [28 May 2008]
5194 *) Fix flaw if 'Server Key exchange message' is omitted from a TLS
5195 handshake which could lead to a client crash as found using the
5196 Codenomicon TLS test suite (CVE-2008-1672)
5197 [Steve Henson, Mark Cox]
5199 *) Fix double free in TLS server name extensions which could lead to
5200 a remote crash found by Codenomicon TLS test suite (CVE-2008-0891)
5203 *) Clear error queue in SSL_CTX_use_certificate_chain_file()
5205 Clear the error queue to ensure that error entries left from
5206 older function calls do not interfere with the correct operation.
5207 [Lutz Jaenicke, Erik de Castro Lopo]
5209 *) Remove root CA certificates of commercial CAs:
5211 The OpenSSL project does not recommend any specific CA and does not
5212 have any policy with respect to including or excluding any CA.
5213 Therefore it does not make any sense to ship an arbitrary selection
5214 of root CA certificates with the OpenSSL software.
5217 *) RSA OAEP patches to fix two separate invalid memory reads.
5218 The first one involves inputs when 'lzero' is greater than
5219 'SHA_DIGEST_LENGTH' (it would read about SHA_DIGEST_LENGTH bytes
5220 before the beginning of from). The second one involves inputs where
5221 the 'db' section contains nothing but zeroes (there is a one-byte
5222 invalid read after the end of 'db').
5223 [Ivan Nestlerode <inestlerode@us.ibm.com>]
5225 *) Partial backport from 0.9.9-dev:
5227 Introduce bn_mul_mont (dedicated Montgomery multiplication
5228 procedure) as a candidate for BIGNUM assembler implementation.
5229 While 0.9.9-dev uses assembler for various architectures, only
5230 x86_64 is available by default here in the 0.9.8 branch, and
5231 32-bit x86 is available through a compile-time setting.
5233 To try the 32-bit x86 assembler implementation, use Configure
5234 option "enable-montasm" (which exists only for this backport).
5236 As "enable-montasm" for 32-bit x86 disclaims code stability
5237 anyway, in this constellation we activate additional code
5238 backported from 0.9.9-dev for further performance improvements,
5239 namely BN_from_montgomery_word. (To enable this otherwise,
5240 e.g. x86_64, try "-DMONT_FROM_WORD___NON_DEFAULT_0_9_8_BUILD".)
5242 [Andy Polyakov (backport partially by Bodo Moeller)]
5244 *) Add TLS session ticket callback. This allows an application to set
5245 TLS ticket cipher and HMAC keys rather than relying on hardcoded fixed
5246 values. This is useful for key rollover for example where several key
5247 sets may exist with different names.
5250 *) Reverse ENGINE-internal logic for caching default ENGINE handles.
5251 This was broken until now in 0.9.8 releases, such that the only way
5252 a registered ENGINE could be used (assuming it initialises
5253 successfully on the host) was to explicitly set it as the default
5254 for the relevant algorithms. This is in contradiction with 0.9.7
5255 behaviour and the documentation. With this fix, when an ENGINE is
5256 registered into a given algorithm's table of implementations, the
5257 'uptodate' flag is reset so that auto-discovery will be used next
5258 time a new context for that algorithm attempts to select an
5260 [Ian Lister (tweaked by Geoff Thorpe)]
5262 *) Backport of CMS code to OpenSSL 0.9.8. This differs from the 0.9.9
5263 implementation in the following ways:
5265 Lack of EVP_PKEY_ASN1_METHOD means algorithm parameters have to be
5268 Lack of BER streaming support means one pass streaming processing is
5269 only supported if data is detached: setting the streaming flag is
5270 ignored for embedded content.
5272 CMS support is disabled by default and must be explicitly enabled
5273 with the enable-cms configuration option.
5276 *) Update the GMP engine glue to do direct copies between BIGNUM and
5277 mpz_t when openssl and GMP use the same limb size. Otherwise the
5278 existing "conversion via a text string export" trick is still used.
5279 [Paul Sheer <paulsheer@gmail.com>]
5281 *) Zlib compression BIO. This is a filter BIO which compressed and
5282 uncompresses any data passed through it.
5285 *) Add AES_wrap_key() and AES_unwrap_key() functions to implement
5286 RFC3394 compatible AES key wrapping.
5289 *) Add utility functions to handle ASN1 structures. ASN1_STRING_set0():
5290 sets string data without copying. X509_ALGOR_set0() and
5291 X509_ALGOR_get0(): set and retrieve X509_ALGOR (AlgorithmIdentifier)
5292 data. Attribute function X509at_get0_data_by_OBJ(): retrieves data
5293 from an X509_ATTRIBUTE structure optionally checking it occurs only
5294 once. ASN1_TYPE_set1(): set and ASN1_TYPE structure copying supplied
5298 *) Fix BN flag handling in RSA_eay_mod_exp() and BN_MONT_CTX_set()
5299 to get the expected BN_FLG_CONSTTIME behavior.
5300 [Bodo Moeller (Google)]
5304 - fixed wrong usage of ioctlsocket() when build for LIBC BSD sockets
5305 - fixed do_tests.pl to run the test suite with CLIB builds too (CLIB_OPT)
5306 - added some more tests to do_tests.pl
5307 - fixed RunningProcess usage so that it works with newer LIBC NDKs too
5308 - removed usage of BN_LLONG for CLIB builds to avoid runtime dependency
5309 - added new Configure targets netware-clib-bsdsock, netware-clib-gcc,
5310 netware-clib-bsdsock-gcc, netware-libc-bsdsock-gcc
5311 - various changes to netware.pl to enable gcc-cross builds on Win32
5313 - changed crypto/bio/b_sock.c to work with macro functions (CLIB BSD)
5314 - various changes to fix missing prototype warnings
5315 - fixed x86nasm.pl to create correct asm files for NASM COFF output
5316 - added AES, WHIRLPOOL and CPUID assembler code to build files
5317 - added missing AES assembler make rules to mk1mf.pl
5318 - fixed order of includes in apps/ocsp.c so that e_os.h settings apply
5319 [Guenter Knauf <eflash@gmx.net>]
5321 *) Implement certificate status request TLS extension defined in RFC3546.
5322 A client can set the appropriate parameters and receive the encoded
5323 OCSP response via a callback. A server can query the supplied parameters
5324 and set the encoded OCSP response in the callback. Add simplified examples
5325 to s_client and s_server.
5328 Changes between 0.9.8f and 0.9.8g [19 Oct 2007]
5330 *) Fix various bugs:
5331 + Binary incompatibility of ssl_ctx_st structure
5332 + DTLS interoperation with non-compliant servers
5333 + Don't call get_session_cb() without proposed session
5334 + Fix ia64 assembler code
5335 [Andy Polyakov, Steve Henson]
5337 Changes between 0.9.8e and 0.9.8f [11 Oct 2007]
5339 *) DTLS Handshake overhaul. There were longstanding issues with
5340 OpenSSL DTLS implementation, which were making it impossible for
5341 RFC 4347 compliant client to communicate with OpenSSL server.
5342 Unfortunately just fixing these incompatibilities would "cut off"
5343 pre-0.9.8f clients. To allow for hassle free upgrade post-0.9.8e
5344 server keeps tolerating non RFC compliant syntax. The opposite is
5345 not true, 0.9.8f client can not communicate with earlier server.
5346 This update even addresses CVE-2007-4995.
5349 *) Changes to avoid need for function casts in OpenSSL: some compilers
5350 (gcc 4.2 and later) reject their use.
5351 [Kurt Roeckx <kurt@roeckx.be>, Peter Hartley <pdh@utter.chaos.org.uk>,
5354 *) Add RFC4507 support to OpenSSL. This includes the corrections in
5355 RFC4507bis. The encrypted ticket format is an encrypted encoded
5356 SSL_SESSION structure, that way new session features are automatically
5359 If a client application caches session in an SSL_SESSION structure
5360 support is transparent because tickets are now stored in the encoded
5363 The SSL_CTX structure automatically generates keys for ticket
5364 protection in servers so again support should be possible
5365 with no application modification.
5367 If a client or server wishes to disable RFC4507 support then the option
5368 SSL_OP_NO_TICKET can be set.
5370 Add a TLS extension debugging callback to allow the contents of any client
5371 or server extensions to be examined.
5373 This work was sponsored by Google.
5376 *) Add initial support for TLS extensions, specifically for the server_name
5377 extension so far. The SSL_SESSION, SSL_CTX, and SSL data structures now
5378 have new members for a host name. The SSL data structure has an
5379 additional member SSL_CTX *initial_ctx so that new sessions can be
5380 stored in that context to allow for session resumption, even after the
5381 SSL has been switched to a new SSL_CTX in reaction to a client's
5382 server_name extension.
5384 New functions (subject to change):
5386 SSL_get_servername()
5387 SSL_get_servername_type()
5390 New CTRL codes and macros (subject to change):
5392 SSL_CTRL_SET_TLSEXT_SERVERNAME_CB
5393 - SSL_CTX_set_tlsext_servername_callback()
5394 SSL_CTRL_SET_TLSEXT_SERVERNAME_ARG
5395 - SSL_CTX_set_tlsext_servername_arg()
5396 SSL_CTRL_SET_TLSEXT_HOSTNAME - SSL_set_tlsext_host_name()
5398 openssl s_client has a new '-servername ...' option.
5400 openssl s_server has new options '-servername_host ...', '-cert2 ...',
5401 '-key2 ...', '-servername_fatal' (subject to change). This allows
5402 testing the HostName extension for a specific single host name ('-cert'
5403 and '-key' remain fallbacks for handshakes without HostName
5404 negotiation). If the unrecognized_name alert has to be sent, this by
5405 default is a warning; it becomes fatal with the '-servername_fatal'
5408 [Peter Sylvester, Remy Allais, Christophe Renou, Steve Henson]
5410 *) Add AES and SSE2 assembly language support to VC++ build.
5413 *) Mitigate attack on final subtraction in Montgomery reduction.
5416 *) Fix crypto/ec/ec_mult.c to work properly with scalars of value 0
5417 (which previously caused an internal error).
5420 *) Squeeze another 10% out of IGE mode when in != out.
5423 *) AES IGE mode speedup.
5424 [Dean Gaudet (Google)]
5426 *) Add the Korean symmetric 128-bit cipher SEED (see
5427 http://www.kisa.or.kr/kisa/seed/jsp/seed_eng.jsp) and
5428 add SEED ciphersuites from RFC 4162:
5430 TLS_RSA_WITH_SEED_CBC_SHA = "SEED-SHA"
5431 TLS_DHE_DSS_WITH_SEED_CBC_SHA = "DHE-DSS-SEED-SHA"
5432 TLS_DHE_RSA_WITH_SEED_CBC_SHA = "DHE-RSA-SEED-SHA"
5433 TLS_DH_anon_WITH_SEED_CBC_SHA = "ADH-SEED-SHA"
5435 To minimize changes between patchlevels in the OpenSSL 0.9.8
5436 series, SEED remains excluded from compilation unless OpenSSL
5437 is configured with 'enable-seed'.
5438 [KISA, Bodo Moeller]
5440 *) Mitigate branch prediction attacks, which can be practical if a
5441 single processor is shared, allowing a spy process to extract
5442 information. For detailed background information, see
5443 http://eprint.iacr.org/2007/039 (O. Aciicmez, S. Gueron,
5444 J.-P. Seifert, "New Branch Prediction Vulnerabilities in OpenSSL
5445 and Necessary Software Countermeasures"). The core of the change
5446 are new versions BN_div_no_branch() and
5447 BN_mod_inverse_no_branch() of BN_div() and BN_mod_inverse(),
5448 respectively, which are slower, but avoid the security-relevant
5449 conditional branches. These are automatically called by BN_div()
5450 and BN_mod_inverse() if the flag BN_FLG_CONSTTIME is set for one
5451 of the input BIGNUMs. Also, BN_is_bit_set() has been changed to
5452 remove a conditional branch.
5454 BN_FLG_CONSTTIME is the new name for the previous
5455 BN_FLG_EXP_CONSTTIME flag, since it now affects more than just
5456 modular exponentiation. (Since OpenSSL 0.9.7h, setting this flag
5457 in the exponent causes BN_mod_exp_mont() to use the alternative
5458 implementation in BN_mod_exp_mont_consttime().) The old name
5459 remains as a deprecated alias.
5461 Similarly, RSA_FLAG_NO_EXP_CONSTTIME is replaced by a more general
5462 RSA_FLAG_NO_CONSTTIME flag since the RSA implementation now uses
5463 constant-time implementations for more than just exponentiation.
5464 Here too the old name is kept as a deprecated alias.
5466 BN_BLINDING_new() will now use BN_dup() for the modulus so that
5467 the BN_BLINDING structure gets an independent copy of the
5468 modulus. This means that the previous "BIGNUM *m" argument to
5469 BN_BLINDING_new() and to BN_BLINDING_create_param() now
5470 essentially becomes "const BIGNUM *m", although we can't actually
5471 change this in the header file before 0.9.9. It allows
5472 RSA_setup_blinding() to use BN_with_flags() on the modulus to
5473 enable BN_FLG_CONSTTIME.
5475 [Matthew D Wood (Intel Corp)]
5477 *) In the SSL/TLS server implementation, be strict about session ID
5478 context matching (which matters if an application uses a single
5479 external cache for different purposes). Previously,
5480 out-of-context reuse was forbidden only if SSL_VERIFY_PEER was
5481 set. This did ensure strict client verification, but meant that,
5482 with applications using a single external cache for quite
5483 different requirements, clients could circumvent ciphersuite
5484 restrictions for a given session ID context by starting a session
5485 in a different context.
5488 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
5489 a ciphersuite string such as "DEFAULT:RSA" cannot enable
5490 authentication-only ciphersuites.
5493 *) Update the SSL_get_shared_ciphers() fix CVE-2006-3738 which was
5494 not complete and could lead to a possible single byte overflow
5495 (CVE-2007-5135) [Ben Laurie]
5497 Changes between 0.9.8d and 0.9.8e [23 Feb 2007]
5499 *) Since AES128 and AES256 (and similarly Camellia128 and
5500 Camellia256) share a single mask bit in the logic of
5501 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
5502 kludge to work properly if AES128 is available and AES256 isn't
5503 (or if Camellia128 is available and Camellia256 isn't).
5506 *) Fix the BIT STRING encoding generated by crypto/ec/ec_asn1.c
5507 (within i2d_ECPrivateKey, i2d_ECPKParameters, i2d_ECParameters):
5508 When a point or a seed is encoded in a BIT STRING, we need to
5509 prevent the removal of trailing zero bits to get the proper DER
5510 encoding. (By default, crypto/asn1/a_bitstr.c assumes the case
5511 of a NamedBitList, for which trailing 0 bits need to be removed.)
5514 *) Have SSL/TLS server implementation tolerate "mismatched" record
5515 protocol version while receiving ClientHello even if the
5516 ClientHello is fragmented. (The server can't insist on the
5517 particular protocol version it has chosen before the ServerHello
5518 message has informed the client about his choice.)
5521 *) Add RFC 3779 support.
5522 [Rob Austein for ARIN, Ben Laurie]
5524 *) Load error codes if they are not already present instead of using a
5525 static variable. This allows them to be cleanly unloaded and reloaded.
5526 Improve header file function name parsing.
5529 *) extend SMTP and IMAP protocol emulation in s_client to use EHLO
5530 or CAPABILITY handshake as required by RFCs.
5533 Changes between 0.9.8c and 0.9.8d [28 Sep 2006]
5535 *) Introduce limits to prevent malicious keys being able to
5536 cause a denial of service. (CVE-2006-2940)
5537 [Steve Henson, Bodo Moeller]
5539 *) Fix ASN.1 parsing of certain invalid structures that can result
5540 in a denial of service. (CVE-2006-2937) [Steve Henson]
5542 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
5543 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
5545 *) Fix SSL client code which could crash if connecting to a
5546 malicious SSLv2 server. (CVE-2006-4343)
5547 [Tavis Ormandy and Will Drewry, Google Security Team]
5549 *) Since 0.9.8b, ciphersuite strings naming explicit ciphersuites
5550 match only those. Before that, "AES256-SHA" would be interpreted
5551 as a pattern and match "AES128-SHA" too (since AES128-SHA got
5552 the same strength classification in 0.9.7h) as we currently only
5553 have a single AES bit in the ciphersuite description bitmap.
5554 That change, however, also applied to ciphersuite strings such as
5555 "RC4-MD5" that intentionally matched multiple ciphersuites --
5556 namely, SSL 2.0 ciphersuites in addition to the more common ones
5557 from SSL 3.0/TLS 1.0.
5559 So we change the selection algorithm again: Naming an explicit
5560 ciphersuite selects this one ciphersuite, and any other similar
5561 ciphersuite (same bitmap) from *other* protocol versions.
5562 Thus, "RC4-MD5" again will properly select both the SSL 2.0
5563 ciphersuite and the SSL 3.0/TLS 1.0 ciphersuite.
5565 Since SSL 2.0 does not have any ciphersuites for which the
5566 128/256 bit distinction would be relevant, this works for now.
5567 The proper fix will be to use different bits for AES128 and
5568 AES256, which would have avoided the problems from the beginning;
5569 however, bits are scarce, so we can only do this in a new release
5570 (not just a patchlevel) when we can change the SSL_CIPHER
5571 definition to split the single 'unsigned long mask' bitmap into
5572 multiple values to extend the available space.
5576 Changes between 0.9.8b and 0.9.8c [05 Sep 2006]
5578 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
5579 (CVE-2006-4339) [Ben Laurie and Google Security Team]
5581 *) Add AES IGE and biIGE modes.
5584 *) Change the Unix randomness entropy gathering to use poll() when
5585 possible instead of select(), since the latter has some
5586 undesirable limitations.
5587 [Darryl Miles via Richard Levitte and Bodo Moeller]
5589 *) Disable "ECCdraft" ciphersuites more thoroughly. Now special
5590 treatment in ssl/ssl_ciph.s makes sure that these ciphersuites
5591 cannot be implicitly activated as part of, e.g., the "AES" alias.
5592 However, please upgrade to OpenSSL 0.9.9[-dev] for
5593 non-experimental use of the ECC ciphersuites to get TLS extension
5594 support, which is required for curve and point format negotiation
5595 to avoid potential handshake problems.
5598 *) Disable rogue ciphersuites:
5600 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
5601 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
5602 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
5604 The latter two were purportedly from
5605 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
5608 Also deactivate the remaining ciphersuites from
5609 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
5610 unofficial, and the ID has long expired.
5613 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
5614 dual-core machines) and other potential thread-safety issues.
5617 *) Add the symmetric cipher Camellia (128-bit, 192-bit, 256-bit key
5618 versions), which is now available for royalty-free use
5619 (see http://info.isl.ntt.co.jp/crypt/eng/info/chiteki.html).
5620 Also, add Camellia TLS ciphersuites from RFC 4132.
5622 To minimize changes between patchlevels in the OpenSSL 0.9.8
5623 series, Camellia remains excluded from compilation unless OpenSSL
5624 is configured with 'enable-camellia'.
5627 *) Disable the padding bug check when compression is in use. The padding
5628 bug check assumes the first packet is of even length, this is not
5629 necessarily true if compression is enabled and can result in false
5630 positives causing handshake failure. The actual bug test is ancient
5631 code so it is hoped that implementations will either have fixed it by
5632 now or any which still have the bug do not support compression.
5635 Changes between 0.9.8a and 0.9.8b [04 May 2006]
5637 *) When applying a cipher rule check to see if string match is an explicit
5638 cipher suite and only match that one cipher suite if it is.
5641 *) Link in manifests for VC++ if needed.
5642 [Austin Ziegler <halostatue@gmail.com>]
5644 *) Update support for ECC-based TLS ciphersuites according to
5645 draft-ietf-tls-ecc-12.txt with proposed changes (but without
5646 TLS extensions, which are supported starting with the 0.9.9
5647 branch, not in the OpenSSL 0.9.8 branch).
5650 *) New functions EVP_CIPHER_CTX_new() and EVP_CIPHER_CTX_free() to support
5651 opaque EVP_CIPHER_CTX handling.
5654 *) Fixes and enhancements to zlib compression code. We now only use
5655 "zlib1.dll" and use the default __cdecl calling convention on Win32
5656 to conform with the standards mentioned here:
5657 http://www.zlib.net/DLL_FAQ.txt
5658 Static zlib linking now works on Windows and the new --with-zlib-include
5659 --with-zlib-lib options to Configure can be used to supply the location
5660 of the headers and library. Gracefully handle case where zlib library
5664 *) Several fixes and enhancements to the OID generation code. The old code
5665 sometimes allowed invalid OIDs (1.X for X >= 40 for example), couldn't
5666 handle numbers larger than ULONG_MAX, truncated printing and had a
5667 non standard OBJ_obj2txt() behaviour.
5670 *) Add support for building of engines under engine/ as shared libraries
5671 under VC++ build system.
5674 *) Corrected the numerous bugs in the Win32 path splitter in DSO.
5675 Hopefully, we will not see any false combination of paths any more.
5678 Changes between 0.9.8 and 0.9.8a [11 Oct 2005]
5680 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
5681 (part of SSL_OP_ALL). This option used to disable the
5682 countermeasure against man-in-the-middle protocol-version
5683 rollback in the SSL 2.0 server implementation, which is a bad
5684 idea. (CVE-2005-2969)
5686 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
5687 for Information Security, National Institute of Advanced Industrial
5688 Science and Technology [AIST], Japan)]
5690 *) Add two function to clear and return the verify parameter flags.
5693 *) Keep cipherlists sorted in the source instead of sorting them at
5694 runtime, thus removing the need for a lock.
5697 *) Avoid some small subgroup attacks in Diffie-Hellman.
5698 [Nick Mathewson and Ben Laurie]
5700 *) Add functions for well-known primes.
5703 *) Extended Windows CE support.
5704 [Satoshi Nakamura and Andy Polyakov]
5706 *) Initialize SSL_METHOD structures at compile time instead of during
5707 runtime, thus removing the need for a lock.
5710 *) Make PKCS7_decrypt() work even if no certificate is supplied by
5711 attempting to decrypt each encrypted key in turn. Add support to
5715 Changes between 0.9.7h and 0.9.8 [05 Jul 2005]
5717 [NB: OpenSSL 0.9.7i and later 0.9.7 patch levels were released after
5720 *) Add libcrypto.pc and libssl.pc for those who feel they need them.
5723 *) Change CA.sh and CA.pl so they don't bundle the CSR and the private
5724 key into the same file any more.
5727 *) Add initial support for Win64, both IA64 and AMD64/x64 flavors.
5730 *) Add -utf8 command line and config file option to 'ca'.
5731 [Stefan <stf@udoma.org]
5733 *) Removed the macro des_crypt(), as it seems to conflict with some
5734 libraries. Use DES_crypt().
5737 *) Correct naming of the 'chil' and '4758cca' ENGINEs. This
5738 involves renaming the source and generated shared-libs for
5739 both. The engines will accept the corrected or legacy ids
5740 ('ncipher' and '4758_cca' respectively) when binding. NB,
5741 this only applies when building 'shared'.
5742 [Corinna Vinschen <vinschen@redhat.com> and Geoff Thorpe]
5744 *) Add attribute functions to EVP_PKEY structure. Modify
5745 PKCS12_create() to recognize a CSP name attribute and
5746 use it. Make -CSP option work again in pkcs12 utility.
5749 *) Add new functionality to the bn blinding code:
5750 - automatic re-creation of the BN_BLINDING parameters after
5751 a fixed number of uses (currently 32)
5752 - add new function for parameter creation
5753 - introduce flags to control the update behaviour of the
5754 BN_BLINDING parameters
5755 - hide BN_BLINDING structure
5756 Add a second BN_BLINDING slot to the RSA structure to improve
5757 performance when a single RSA object is shared among several
5761 *) Add support for DTLS.
5762 [Nagendra Modadugu <nagendra@cs.stanford.edu> and Ben Laurie]
5764 *) Add support for DER encoded private keys (SSL_FILETYPE_ASN1)
5765 to SSL_CTX_use_PrivateKey_file() and SSL_use_PrivateKey_file()
5768 *) Remove buggy and incomplete DH cert support from
5769 ssl/ssl_rsa.c and ssl/s3_both.c
5772 *) Use SHA-1 instead of MD5 as the default digest algorithm for
5773 the apps/openssl applications.
5776 *) Compile clean with "-Wall -Wmissing-prototypes
5777 -Wstrict-prototypes -Wmissing-declarations -Werror". Currently
5778 DEBUG_SAFESTACK must also be set.
5781 *) Change ./Configure so that certain algorithms can be disabled by default.
5782 The new counterpiece to "no-xxx" is "enable-xxx".
5784 The patented RC5 and MDC2 algorithms will now be disabled unless
5785 "enable-rc5" and "enable-mdc2", respectively, are specified.
5787 (IDEA remains enabled despite being patented. This is because IDEA
5788 is frequently required for interoperability, and there is no license
5789 fee for non-commercial use. As before, "no-idea" can be used to
5790 avoid this algorithm.)
5794 *) Add processing of proxy certificates (see RFC 3820). This work was
5795 sponsored by KTH (The Royal Institute of Technology in Stockholm) and
5796 EGEE (Enabling Grids for E-science in Europe).
5799 *) RC4 performance overhaul on modern architectures/implementations, such
5800 as Intel P4, IA-64 and AMD64.
5803 *) New utility extract-section.pl. This can be used specify an alternative
5804 section number in a pod file instead of having to treat each file as
5805 a separate case in Makefile. This can be done by adding two lines to the
5808 =for comment openssl_section:XXX
5810 The blank line is mandatory.
5814 *) New arguments -certform, -keyform and -pass for s_client and s_server
5815 to allow alternative format key and certificate files and passphrase
5819 *) New structure X509_VERIFY_PARAM which combines current verify parameters,
5820 update associated structures and add various utility functions.
5822 Add new policy related verify parameters, include policy checking in
5823 standard verify code. Enhance 'smime' application with extra parameters
5824 to support policy checking and print out.
5827 *) Add a new engine to support VIA PadLock ACE extensions in the VIA C3
5828 Nehemiah processors. These extensions support AES encryption in hardware
5829 as well as RNG (though RNG support is currently disabled).
5830 [Michal Ludvig <michal@logix.cz>, with help from Andy Polyakov]
5832 *) Deprecate BN_[get|set]_params() functions (they were ignored internally).
5835 *) New FIPS 180-2 algorithms, SHA-224/-256/-384/-512 are implemented.
5836 [Andy Polyakov and a number of other people]
5838 *) Improved PowerPC platform support. Most notably BIGNUM assembler
5839 implementation contributed by IBM.
5840 [Suresh Chari, Peter Waltenberg, Andy Polyakov]
5842 *) The new 'RSA_generate_key_ex' function now takes a BIGNUM for the public
5843 exponent rather than 'unsigned long'. There is a corresponding change to
5844 the new 'rsa_keygen' element of the RSA_METHOD structure.
5845 [Jelte Jansen, Geoff Thorpe]
5847 *) Functionality for creating the initial serial number file is now
5848 moved from CA.pl to the 'ca' utility with a new option -create_serial.
5850 (Before OpenSSL 0.9.7e, CA.pl used to initialize the serial
5851 number file to 1, which is bound to cause problems. To avoid
5852 the problems while respecting compatibility between different 0.9.7
5853 patchlevels, 0.9.7e employed 'openssl x509 -next_serial' in
5854 CA.pl for serial number initialization. With the new release 0.9.8,
5855 we can fix the problem directly in the 'ca' utility.)
5858 *) Reduced header interdependencies by declaring more opaque objects in
5859 ossl_typ.h. As a consequence, including some headers (eg. engine.h) will
5860 give fewer recursive includes, which could break lazy source code - so
5861 this change is covered by the OPENSSL_NO_DEPRECATED symbol. As always,
5862 developers should define this symbol when building and using openssl to
5863 ensure they track the recommended behaviour, interfaces, [etc], but
5864 backwards-compatible behaviour prevails when this isn't defined.
5867 *) New function X509_POLICY_NODE_print() which prints out policy nodes.
5870 *) Add new EVP function EVP_CIPHER_CTX_rand_key and associated functionality.
5871 This will generate a random key of the appropriate length based on the
5872 cipher context. The EVP_CIPHER can provide its own random key generation
5873 routine to support keys of a specific form. This is used in the des and
5874 3des routines to generate a key of the correct parity. Update S/MIME
5875 code to use new functions and hence generate correct parity DES keys.
5876 Add EVP_CHECK_DES_KEY #define to return an error if the key is not
5877 valid (weak or incorrect parity).
5880 *) Add a local set of CRLs that can be used by X509_verify_cert() as well
5881 as looking them up. This is useful when the verified structure may contain
5882 CRLs, for example PKCS#7 signedData. Modify PKCS7_verify() to use any CRLs
5883 present unless the new PKCS7_NO_CRL flag is asserted.
5886 *) Extend ASN1 oid configuration module. It now additionally accepts the
5889 shortName = some long name, 1.2.3.4
5892 *) Reimplemented the BN_CTX implementation. There is now no more static
5893 limitation on the number of variables it can handle nor the depth of the
5894 "stack" handling for BN_CTX_start()/BN_CTX_end() pairs. The stack
5895 information can now expand as required, and rather than having a single
5896 static array of bignums, BN_CTX now uses a linked-list of such arrays
5897 allowing it to expand on demand whilst maintaining the usefulness of
5898 BN_CTX's "bundling".
5901 *) Add a missing BN_CTX parameter to the 'rsa_mod_exp' callback in RSA_METHOD
5902 to allow all RSA operations to function using a single BN_CTX.
5905 *) Preliminary support for certificate policy evaluation and checking. This
5906 is initially intended to pass the tests outlined in "Conformance Testing
5907 of Relying Party Client Certificate Path Processing Logic" v1.07.
5910 *) bn_dup_expand() has been deprecated, it was introduced in 0.9.7 and
5911 remained unused and not that useful. A variety of other little bignum
5912 tweaks and fixes have also been made continuing on from the audit (see
5916 *) Constify all or almost all d2i, c2i, s2i and r2i functions, along with
5917 associated ASN1, EVP and SSL functions and old ASN1 macros.
5920 *) BN_zero() only needs to set 'top' and 'neg' to zero for correct results,
5921 and this should never fail. So the return value from the use of
5922 BN_set_word() (which can fail due to needless expansion) is now deprecated;
5923 if OPENSSL_NO_DEPRECATED is defined, BN_zero() is a void macro.
5926 *) BN_CTX_get() should return zero-valued bignums, providing the same
5927 initialised value as BN_new().
5928 [Geoff Thorpe, suggested by Ulf Möller]
5930 *) Support for inhibitAnyPolicy certificate extension.
5933 *) An audit of the BIGNUM code is underway, for which debugging code is
5934 enabled when BN_DEBUG is defined. This makes stricter enforcements on what
5935 is considered valid when processing BIGNUMs, and causes execution to
5936 assert() when a problem is discovered. If BN_DEBUG_RAND is defined,
5937 further steps are taken to deliberately pollute unused data in BIGNUM
5938 structures to try and expose faulty code further on. For now, openssl will
5939 (in its default mode of operation) continue to tolerate the inconsistent
5940 forms that it has tolerated in the past, but authors and packagers should
5941 consider trying openssl and their own applications when compiled with
5942 these debugging symbols defined. It will help highlight potential bugs in
5943 their own code, and will improve the test coverage for OpenSSL itself. At
5944 some point, these tighter rules will become openssl's default to improve
5945 maintainability, though the assert()s and other overheads will remain only
5946 in debugging configurations. See bn.h for more details.
5947 [Geoff Thorpe, Nils Larsch, Ulf Möller]
5949 *) BN_CTX_init() has been deprecated, as BN_CTX is an opaque structure
5950 that can only be obtained through BN_CTX_new() (which implicitly
5951 initialises it). The presence of this function only made it possible
5952 to overwrite an existing structure (and cause memory leaks).
5955 *) Because of the callback-based approach for implementing LHASH as a
5956 template type, lh_insert() adds opaque objects to hash-tables and
5957 lh_doall() or lh_doall_arg() are typically used with a destructor callback
5958 to clean up those corresponding objects before destroying the hash table
5959 (and losing the object pointers). So some over-zealous constifications in
5960 LHASH have been relaxed so that lh_insert() does not take (nor store) the
5961 objects as "const" and the lh_doall[_arg] callback wrappers are not
5962 prototyped to have "const" restrictions on the object pointers they are
5963 given (and so aren't required to cast them away any more).
5966 *) The tmdiff.h API was so ugly and minimal that our own timing utility
5967 (speed) prefers to use its own implementation. The two implementations
5968 haven't been consolidated as yet (volunteers?) but the tmdiff API has had
5969 its object type properly exposed (MS_TM) instead of casting to/from "char
5970 *". This may still change yet if someone realises MS_TM and "ms_time_***"
5971 aren't necessarily the greatest nomenclatures - but this is what was used
5972 internally to the implementation so I've used that for now.
5975 *) Ensure that deprecated functions do not get compiled when
5976 OPENSSL_NO_DEPRECATED is defined. Some "openssl" subcommands and a few of
5977 the self-tests were still using deprecated key-generation functions so
5978 these have been updated also.
5981 *) Reorganise PKCS#7 code to separate the digest location functionality
5982 into PKCS7_find_digest(), digest addition into PKCS7_bio_add_digest().
5983 New function PKCS7_set_digest() to set the digest type for PKCS#7
5984 digestedData type. Add additional code to correctly generate the
5985 digestedData type and add support for this type in PKCS7 initialization
5989 *) New function PKCS7_set0_type_other() this initializes a PKCS7
5990 structure of type "other".
5993 *) Fix prime generation loop in crypto/bn/bn_prime.pl by making
5994 sure the loop does correctly stop and breaking ("division by zero")
5995 modulus operations are not performed. The (pre-generated) prime
5996 table crypto/bn/bn_prime.h was already correct, but it could not be
5997 re-generated on some platforms because of the "division by zero"
5998 situation in the script.
5999 [Ralf S. Engelschall]
6001 *) Update support for ECC-based TLS ciphersuites according to
6002 draft-ietf-tls-ecc-03.txt: the KDF1 key derivation function with
6003 SHA-1 now is only used for "small" curves (where the
6004 representation of a field element takes up to 24 bytes); for
6005 larger curves, the field element resulting from ECDH is directly
6006 used as premaster secret.
6007 [Douglas Stebila (Sun Microsystems Laboratories)]
6009 *) Add code for kP+lQ timings to crypto/ec/ectest.c, and add SEC2
6010 curve secp160r1 to the tests.
6011 [Douglas Stebila (Sun Microsystems Laboratories)]
6013 *) Add the possibility to load symbols globally with DSO.
6014 [Götz Babin-Ebell <babin-ebell@trustcenter.de> via Richard Levitte]
6016 *) Add the functions ERR_set_mark() and ERR_pop_to_mark() for better
6017 control of the error stack.
6020 *) Add support for STORE in ENGINE.
6023 *) Add the STORE type. The intention is to provide a common interface
6024 to certificate and key stores, be they simple file-based stores, or
6025 HSM-type store, or LDAP stores, or...
6026 NOTE: The code is currently UNTESTED and isn't really used anywhere.
6029 *) Add a generic structure called OPENSSL_ITEM. This can be used to
6030 pass a list of arguments to any function as well as provide a way
6031 for a function to pass data back to the caller.
6034 *) Add the functions BUF_strndup() and BUF_memdup(). BUF_strndup()
6035 works like BUF_strdup() but can be used to duplicate a portion of
6036 a string. The copy gets NUL-terminated. BUF_memdup() duplicates
6040 *) Add the function sk_find_ex() which works like sk_find(), but will
6041 return an index to an element even if an exact match couldn't be
6042 found. The index is guaranteed to point at the element where the
6043 searched-for key would be inserted to preserve sorting order.
6046 *) Add the function OBJ_bsearch_ex() which works like OBJ_bsearch() but
6047 takes an extra flags argument for optional functionality. Currently,
6048 the following flags are defined:
6050 OBJ_BSEARCH_VALUE_ON_NOMATCH
6051 This one gets OBJ_bsearch_ex() to return a pointer to the first
6052 element where the comparing function returns a negative or zero
6055 OBJ_BSEARCH_FIRST_VALUE_ON_MATCH
6056 This one gets OBJ_bsearch_ex() to return a pointer to the first
6057 element where the comparing function returns zero. This is useful
6058 if there are more than one element where the comparing function
6062 *) Make it possible to create self-signed certificates with 'openssl ca'
6063 in such a way that the self-signed certificate becomes part of the
6064 CA database and uses the same mechanisms for serial number generation
6065 as all other certificate signing. The new flag '-selfsign' enables
6066 this functionality. Adapt CA.sh and CA.pl.in.
6069 *) Add functionality to check the public key of a certificate request
6070 against a given private. This is useful to check that a certificate
6071 request can be signed by that key (self-signing).
6074 *) Make it possible to have multiple active certificates with the same
6075 subject in the CA index file. This is done only if the keyword
6076 'unique_subject' is set to 'no' in the main CA section (default
6077 if 'CA_default') of the configuration file. The value is saved
6078 with the database itself in a separate index attribute file,
6079 named like the index file with '.attr' appended to the name.
6082 *) Generate multi-valued AVAs using '+' notation in config files for
6086 *) Support for nameConstraints certificate extension.
6089 *) Support for policyConstraints certificate extension.
6092 *) Support for policyMappings certificate extension.
6095 *) Make sure the default DSA_METHOD implementation only uses its
6096 dsa_mod_exp() and/or bn_mod_exp() handlers if they are non-NULL,
6097 and change its own handlers to be NULL so as to remove unnecessary
6098 indirection. This lets alternative implementations fallback to the
6099 default implementation more easily.
6102 *) Support for directoryName in GeneralName related extensions
6106 *) Make it possible to link applications using Makefile.shared.
6107 Make that possible even when linking against static libraries!
6110 *) Support for single pass processing for S/MIME signing. This now
6111 means that S/MIME signing can be done from a pipe, in addition
6112 cleartext signing (multipart/signed type) is effectively streaming
6113 and the signed data does not need to be all held in memory.
6115 This is done with a new flag PKCS7_STREAM. When this flag is set
6116 PKCS7_sign() only initializes the PKCS7 structure and the actual signing
6117 is done after the data is output (and digests calculated) in
6118 SMIME_write_PKCS7().
6121 *) Add full support for -rpath/-R, both in shared libraries and
6122 applications, at least on the platforms where it's known how
6126 *) In crypto/ec/ec_mult.c, implement fast point multiplication with
6127 precomputation, based on wNAF splitting: EC_GROUP_precompute_mult()
6128 will now compute a table of multiples of the generator that
6129 makes subsequent invocations of EC_POINTs_mul() or EC_POINT_mul()
6130 faster (notably in the case of a single point multiplication,
6131 scalar * generator).
6132 [Nils Larsch, Bodo Moeller]
6134 *) IPv6 support for certificate extensions. The various extensions
6135 which use the IP:a.b.c.d can now take IPv6 addresses using the
6136 formats of RFC1884 2.2 . IPv6 addresses are now also displayed
6140 *) Added an ENGINE that implements RSA by performing private key
6141 exponentiations with the GMP library. The conversions to and from
6142 GMP's mpz_t format aren't optimised nor are any montgomery forms
6143 cached, and on x86 it appears OpenSSL's own performance has caught up.
6144 However there are likely to be other architectures where GMP could
6145 provide a boost. This ENGINE is not built in by default, but it can be
6146 specified at Configure time and should be accompanied by the necessary
6147 linker additions, eg;
6148 ./config -DOPENSSL_USE_GMP -lgmp
6151 *) "openssl engine" will not display ENGINE/DSO load failure errors when
6152 testing availability of engines with "-t" - the old behaviour is
6153 produced by increasing the feature's verbosity with "-tt".
6156 *) ECDSA routines: under certain error conditions uninitialized BN objects
6157 could be freed. Solution: make sure initialization is performed early
6158 enough. (Reported and fix supplied by Nils Larsch <nla@trustcenter.de>
6162 *) Key-generation can now be implemented in RSA_METHOD, DSA_METHOD
6163 and DH_METHOD (eg. by ENGINE implementations) to override the normal
6164 software implementations. For DSA and DH, parameter generation can
6165 also be overridden by providing the appropriate method callbacks.
6168 *) Change the "progress" mechanism used in key-generation and
6169 primality testing to functions that take a new BN_GENCB pointer in
6170 place of callback/argument pairs. The new API functions have "_ex"
6171 postfixes and the older functions are reimplemented as wrappers for
6172 the new ones. The OPENSSL_NO_DEPRECATED symbol can be used to hide
6173 declarations of the old functions to help (graceful) attempts to
6174 migrate to the new functions. Also, the new key-generation API
6175 functions operate on a caller-supplied key-structure and return
6176 success/failure rather than returning a key or NULL - this is to
6177 help make "keygen" another member function of RSA_METHOD etc.
6179 Example for using the new callback interface:
6181 int (*my_callback)(int a, int b, BN_GENCB *cb) = ...;
6185 BN_GENCB_set(&my_cb, my_callback, my_arg);
6187 return BN_is_prime_ex(some_bignum, BN_prime_checks, NULL, &cb);
6188 /* For the meaning of a, b in calls to my_callback(), see the
6189 * documentation of the function that calls the callback.
6190 * cb will point to my_cb; my_arg can be retrieved as cb->arg.
6191 * my_callback should return 1 if it wants BN_is_prime_ex()
6192 * to continue, or 0 to stop.
6197 *) Change the ZLIB compression method to be stateful, and make it
6198 available to TLS with the number defined in
6199 draft-ietf-tls-compression-04.txt.
6202 *) Add the ASN.1 structures and functions for CertificatePair, which
6203 is defined as follows (according to X.509_4thEditionDraftV6.pdf):
6205 CertificatePair ::= SEQUENCE {
6206 forward [0] Certificate OPTIONAL,
6207 reverse [1] Certificate OPTIONAL,
6208 -- at least one of the pair shall be present -- }
6210 Also implement the PEM functions to read and write certificate
6211 pairs, and defined the PEM tag as "CERTIFICATE PAIR".
6213 This needed to be defined, mostly for the sake of the LDAP
6214 attribute crossCertificatePair, but may prove useful elsewhere as
6218 *) Make it possible to inhibit symlinking of shared libraries in
6219 Makefile.shared, for Cygwin's sake.
6222 *) Extend the BIGNUM API by creating a function
6223 void BN_set_negative(BIGNUM *a, int neg);
6224 and a macro that behave like
6225 int BN_is_negative(const BIGNUM *a);
6227 to avoid the need to access 'a->neg' directly in applications.
6230 *) Implement fast modular reduction for pseudo-Mersenne primes
6231 used in NIST curves (crypto/bn/bn_nist.c, crypto/ec/ecp_nist.c).
6232 EC_GROUP_new_curve_GFp() will now automatically use this
6234 [Nils Larsch <nla@trustcenter.de>]
6236 *) Add new lock type (CRYPTO_LOCK_BN).
6239 *) Change the ENGINE framework to automatically load engines
6240 dynamically from specific directories unless they could be
6241 found to already be built in or loaded. Move all the
6242 current engines except for the cryptodev one to a new
6244 The engines in engines/ are built as shared libraries if
6245 the "shared" options was given to ./Configure or ./config.
6246 Otherwise, they are inserted in libcrypto.a.
6247 /usr/local/ssl/engines is the default directory for dynamic
6248 engines, but that can be overridden at configure time through
6249 the usual use of --prefix and/or --openssldir, and at run
6250 time with the environment variable OPENSSL_ENGINES.
6251 [Geoff Thorpe and Richard Levitte]
6253 *) Add Makefile.shared, a helper makefile to build shared
6254 libraries. Adapt Makefile.org.
6257 *) Add version info to Win32 DLLs.
6258 [Peter 'Luna' Runestig" <peter@runestig.com>]
6260 *) Add new 'medium level' PKCS#12 API. Certificates and keys
6261 can be added using this API to created arbitrary PKCS#12
6262 files while avoiding the low level API.
6264 New options to PKCS12_create(), key or cert can be NULL and
6265 will then be omitted from the output file. The encryption
6266 algorithm NIDs can be set to -1 for no encryption, the mac
6267 iteration count can be set to 0 to omit the mac.
6269 Enhance pkcs12 utility by making the -nokeys and -nocerts
6270 options work when creating a PKCS#12 file. New option -nomac
6271 to omit the mac, NONE can be set for an encryption algorithm.
6272 New code is modified to use the enhanced PKCS12_create()
6273 instead of the low level API.
6276 *) Extend ASN1 encoder to support indefinite length constructed
6277 encoding. This can output sequences tags and octet strings in
6278 this form. Modify pk7_asn1.c to support indefinite length
6279 encoding. This is experimental and needs additional code to
6280 be useful, such as an ASN1 bio and some enhanced streaming
6283 Extend template encode functionality so that tagging is passed
6284 down to the template encoder.
6287 *) Let 'openssl req' fail if an argument to '-newkey' is not
6288 recognized instead of using RSA as a default.
6291 *) Add support for ECC-based ciphersuites from draft-ietf-tls-ecc-01.txt.
6292 As these are not official, they are not included in "ALL";
6293 the "ECCdraft" ciphersuite group alias can be used to select them.
6294 [Vipul Gupta and Sumit Gupta (Sun Microsystems Laboratories)]
6296 *) Add ECDH engine support.
6297 [Nils Gura and Douglas Stebila (Sun Microsystems Laboratories)]
6299 *) Add ECDH in new directory crypto/ecdh/.
6300 [Douglas Stebila (Sun Microsystems Laboratories)]
6302 *) Let BN_rand_range() abort with an error after 100 iterations
6303 without success (which indicates a broken PRNG).
6306 *) Change BN_mod_sqrt() so that it verifies that the input value
6307 is really the square of the return value. (Previously,
6308 BN_mod_sqrt would show GIGO behaviour.)
6311 *) Add named elliptic curves over binary fields from X9.62, SECG,
6312 and WAP/WTLS; add OIDs that were still missing.
6314 [Sheueling Chang Shantz and Douglas Stebila
6315 (Sun Microsystems Laboratories)]
6317 *) Extend the EC library for elliptic curves over binary fields
6318 (new files ec2_smpl.c, ec2_smpt.c, ec2_mult.c in crypto/ec/).
6321 EC_GF2m_simple_method
6325 EC_GROUP_new_curve_GF2m
6326 EC_GROUP_set_curve_GF2m
6327 EC_GROUP_get_curve_GF2m
6328 EC_POINT_set_affine_coordinates_GF2m
6329 EC_POINT_get_affine_coordinates_GF2m
6330 EC_POINT_set_compressed_coordinates_GF2m
6332 Point compression for binary fields is disabled by default for
6333 patent reasons (compile with OPENSSL_EC_BIN_PT_COMP defined to
6336 As binary polynomials are represented as BIGNUMs, various members
6337 of the EC_GROUP and EC_POINT data structures can be shared
6338 between the implementations for prime fields and binary fields;
6339 the above ..._GF2m functions (except for EX_GROUP_new_curve_GF2m)
6340 are essentially identical to their ..._GFp counterparts.
6341 (For simplicity, the '..._GFp' prefix has been dropped from
6342 various internal method names.)
6344 An internal 'field_div' method (similar to 'field_mul' and
6345 'field_sqr') has been added; this is used only for binary fields.
6347 [Sheueling Chang Shantz and Douglas Stebila
6348 (Sun Microsystems Laboratories)]
6350 *) Optionally dispatch EC_POINT_mul(), EC_POINT_precompute_mult()
6351 through methods ('mul', 'precompute_mult').
6353 The generic implementations (now internally called 'ec_wNAF_mul'
6354 and 'ec_wNAF_precomputed_mult') remain the default if these
6355 methods are undefined.
6357 [Sheueling Chang Shantz and Douglas Stebila
6358 (Sun Microsystems Laboratories)]
6360 *) New function EC_GROUP_get_degree, which is defined through
6361 EC_METHOD. For curves over prime fields, this returns the bit
6362 length of the modulus.
6364 [Sheueling Chang Shantz and Douglas Stebila
6365 (Sun Microsystems Laboratories)]
6367 *) New functions EC_GROUP_dup, EC_POINT_dup.
6368 (These simply call ..._new and ..._copy).
6370 [Sheueling Chang Shantz and Douglas Stebila
6371 (Sun Microsystems Laboratories)]
6373 *) Add binary polynomial arithmetic software in crypto/bn/bn_gf2m.c.
6374 Polynomials are represented as BIGNUMs (where the sign bit is not
6375 used) in the following functions [macros]:
6378 BN_GF2m_sub [= BN_GF2m_add]
6379 BN_GF2m_mod [wrapper for BN_GF2m_mod_arr]
6380 BN_GF2m_mod_mul [wrapper for BN_GF2m_mod_mul_arr]
6381 BN_GF2m_mod_sqr [wrapper for BN_GF2m_mod_sqr_arr]
6383 BN_GF2m_mod_exp [wrapper for BN_GF2m_mod_exp_arr]
6384 BN_GF2m_mod_sqrt [wrapper for BN_GF2m_mod_sqrt_arr]
6385 BN_GF2m_mod_solve_quad [wrapper for BN_GF2m_mod_solve_quad_arr]
6386 BN_GF2m_cmp [= BN_ucmp]
6388 (Note that only the 'mod' functions are actually for fields GF(2^m).
6389 BN_GF2m_add() is misnomer, but this is for the sake of consistency.)
6391 For some functions, an the irreducible polynomial defining a
6392 field can be given as an 'unsigned int[]' with strictly
6393 decreasing elements giving the indices of those bits that are set;
6394 i.e., p[] represents the polynomial
6395 f(t) = t^p[0] + t^p[1] + ... + t^p[k]
6397 p[0] > p[1] > ... > p[k] = 0.
6398 This applies to the following functions:
6403 BN_GF2m_mod_inv_arr [wrapper for BN_GF2m_mod_inv]
6404 BN_GF2m_mod_div_arr [wrapper for BN_GF2m_mod_div]
6406 BN_GF2m_mod_sqrt_arr
6407 BN_GF2m_mod_solve_quad_arr
6411 Conversion can be performed by the following functions:
6416 bntest.c has additional tests for binary polynomial arithmetic.
6418 Two implementations for BN_GF2m_mod_div() are available.
6419 The default algorithm simply uses BN_GF2m_mod_inv() and
6420 BN_GF2m_mod_mul(). The alternative algorithm is compiled in only
6421 if OPENSSL_SUN_GF2M_DIV is defined (patent pending; read the
6422 copyright notice in crypto/bn/bn_gf2m.c before enabling it).
6424 [Sheueling Chang Shantz and Douglas Stebila
6425 (Sun Microsystems Laboratories)]
6427 *) Add new error code 'ERR_R_DISABLED' that can be used when some
6428 functionality is disabled at compile-time.
6429 [Douglas Stebila <douglas.stebila@sun.com>]
6431 *) Change default behaviour of 'openssl asn1parse' so that more
6432 information is visible when viewing, e.g., a certificate:
6434 Modify asn1_parse2 (crypto/asn1/asn1_par.c) so that in non-'dump'
6435 mode the content of non-printable OCTET STRINGs is output in a
6436 style similar to INTEGERs, but with '[HEX DUMP]' prepended to
6437 avoid the appearance of a printable string.
6438 [Nils Larsch <nla@trustcenter.de>]
6440 *) Add 'asn1_flag' and 'asn1_form' member to EC_GROUP with access
6442 EC_GROUP_set_asn1_flag()
6443 EC_GROUP_get_asn1_flag()
6444 EC_GROUP_set_point_conversion_form()
6445 EC_GROUP_get_point_conversion_form()
6446 These control ASN1 encoding details:
6447 - Curves (i.e., groups) are encoded explicitly unless asn1_flag
6448 has been set to OPENSSL_EC_NAMED_CURVE.
6449 - Points are encoded in uncompressed form by default; options for
6450 asn1_for are as for point2oct, namely
6451 POINT_CONVERSION_COMPRESSED
6452 POINT_CONVERSION_UNCOMPRESSED
6453 POINT_CONVERSION_HYBRID
6455 Also add 'seed' and 'seed_len' members to EC_GROUP with access
6458 EC_GROUP_get0_seed()
6459 EC_GROUP_get_seed_len()
6460 This is used only for ASN1 purposes (so far).
6461 [Nils Larsch <nla@trustcenter.de>]
6463 *) Add 'field_type' member to EC_METHOD, which holds the NID
6464 of the appropriate field type OID. The new function
6465 EC_METHOD_get_field_type() returns this value.
6466 [Nils Larsch <nla@trustcenter.de>]
6471 EC_POINT_point2hex()
6472 EC_POINT_hex2point()
6473 providing useful interfaces to EC_POINT_point2oct() and
6474 EC_POINT_oct2point().
6475 [Nils Larsch <nla@trustcenter.de>]
6477 *) Change internals of the EC library so that the functions
6478 EC_GROUP_set_generator()
6479 EC_GROUP_get_generator()
6480 EC_GROUP_get_order()
6481 EC_GROUP_get_cofactor()
6482 are implemented directly in crypto/ec/ec_lib.c and not dispatched
6483 to methods, which would lead to unnecessary code duplication when
6484 adding different types of curves.
6485 [Nils Larsch <nla@trustcenter.de> with input by Bodo Moeller]
6487 *) Implement compute_wNAF (crypto/ec/ec_mult.c) without BIGNUM
6488 arithmetic, and such that modified wNAFs are generated
6489 (which avoid length expansion in many cases).
6492 *) Add a function EC_GROUP_check_discriminant() (defined via
6493 EC_METHOD) that verifies that the curve discriminant is non-zero.
6495 Add a function EC_GROUP_check() that makes some sanity tests
6496 on a EC_GROUP, its generator and order. This includes
6497 EC_GROUP_check_discriminant().
6498 [Nils Larsch <nla@trustcenter.de>]
6500 *) Add ECDSA in new directory crypto/ecdsa/.
6502 Add applications 'openssl ecparam' and 'openssl ecdsa'
6503 (these are based on 'openssl dsaparam' and 'openssl dsa').
6505 ECDSA support is also included in various other files across the
6506 library. Most notably,
6507 - 'openssl req' now has a '-newkey ecdsa:file' option;
6508 - EVP_PKCS82PKEY (crypto/evp/evp_pkey.c) now can handle ECDSA;
6509 - X509_PUBKEY_get (crypto/asn1/x_pubkey.c) and
6510 d2i_PublicKey (crypto/asn1/d2i_pu.c) have been modified to make
6511 them suitable for ECDSA where domain parameters must be
6512 extracted before the specific public key;
6513 - ECDSA engine support has been added.
6514 [Nils Larsch <nla@trustcenter.de>]
6516 *) Include some named elliptic curves, and add OIDs from X9.62,
6517 SECG, and WAP/WTLS. Each curve can be obtained from the new
6519 EC_GROUP_new_by_curve_name(),
6520 and the list of available named curves can be obtained with
6521 EC_get_builtin_curves().
6522 Also add a 'curve_name' member to EC_GROUP objects, which can be
6524 EC_GROUP_set_curve_name()
6525 EC_GROUP_get_curve_name()
6526 [Nils Larsch <larsch@trustcenter.de, Bodo Moeller]
6528 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
6529 was actually never needed) and in BN_mul(). The removal in BN_mul()
6530 required a small change in bn_mul_part_recursive() and the addition
6531 of the functions bn_cmp_part_words(), bn_sub_part_words() and
6532 bn_add_part_words(), which do the same thing as bn_cmp_words(),
6533 bn_sub_words() and bn_add_words() except they take arrays with
6537 Changes between 0.9.7l and 0.9.7m [23 Feb 2007]
6539 *) Cleanse PEM buffers before freeing them since they may contain
6541 [Benjamin Bennett <ben@psc.edu>]
6543 *) Include "!eNULL" in SSL_DEFAULT_CIPHER_LIST to make sure that
6544 a ciphersuite string such as "DEFAULT:RSA" cannot enable
6545 authentication-only ciphersuites.
6548 *) Since AES128 and AES256 share a single mask bit in the logic of
6549 ssl/ssl_ciph.c, the code for masking out disabled ciphers needs a
6550 kludge to work properly if AES128 is available and AES256 isn't.
6553 *) Expand security boundary to match 1.1.1 module.
6556 *) Remove redundant features: hash file source, editing of test vectors
6557 modify fipsld to use external fips_premain.c signature.
6560 *) New perl script mkfipsscr.pl to create shell scripts or batch files to
6561 run algorithm test programs.
6564 *) Make algorithm test programs more tolerant of whitespace.
6567 *) Have SSL/TLS server implementation tolerate "mismatched" record
6568 protocol version while receiving ClientHello even if the
6569 ClientHello is fragmented. (The server can't insist on the
6570 particular protocol version it has chosen before the ServerHello
6571 message has informed the client about his choice.)
6574 *) Load error codes if they are not already present instead of using a
6575 static variable. This allows them to be cleanly unloaded and reloaded.
6578 Changes between 0.9.7k and 0.9.7l [28 Sep 2006]
6580 *) Introduce limits to prevent malicious keys being able to
6581 cause a denial of service. (CVE-2006-2940)
6582 [Steve Henson, Bodo Moeller]
6584 *) Fix ASN.1 parsing of certain invalid structures that can result
6585 in a denial of service. (CVE-2006-2937) [Steve Henson]
6587 *) Fix buffer overflow in SSL_get_shared_ciphers() function.
6588 (CVE-2006-3738) [Tavis Ormandy and Will Drewry, Google Security Team]
6590 *) Fix SSL client code which could crash if connecting to a
6591 malicious SSLv2 server. (CVE-2006-4343)
6592 [Tavis Ormandy and Will Drewry, Google Security Team]
6594 *) Change ciphersuite string processing so that an explicit
6595 ciphersuite selects this one ciphersuite (so that "AES256-SHA"
6596 will no longer include "AES128-SHA"), and any other similar
6597 ciphersuite (same bitmap) from *other* protocol versions (so that
6598 "RC4-MD5" will still include both the SSL 2.0 ciphersuite and the
6599 SSL 3.0/TLS 1.0 ciphersuite). This is a backport combining
6600 changes from 0.9.8b and 0.9.8d.
6603 Changes between 0.9.7j and 0.9.7k [05 Sep 2006]
6605 *) Avoid PKCS #1 v1.5 signature attack discovered by Daniel Bleichenbacher
6606 (CVE-2006-4339) [Ben Laurie and Google Security Team]
6608 *) Change the Unix randomness entropy gathering to use poll() when
6609 possible instead of select(), since the latter has some
6610 undesirable limitations.
6611 [Darryl Miles via Richard Levitte and Bodo Moeller]
6613 *) Disable rogue ciphersuites:
6615 - SSLv2 0x08 0x00 0x80 ("RC4-64-MD5")
6616 - SSLv3/TLSv1 0x00 0x61 ("EXP1024-RC2-CBC-MD5")
6617 - SSLv3/TLSv1 0x00 0x60 ("EXP1024-RC4-MD5")
6619 The latter two were purportedly from
6620 draft-ietf-tls-56-bit-ciphersuites-0[01].txt, but do not really
6623 Also deactivate the remaining ciphersuites from
6624 draft-ietf-tls-56-bit-ciphersuites-01.txt. These are just as
6625 unofficial, and the ID has long expired.
6628 *) Fix RSA blinding Heisenbug (problems sometimes occurred on
6629 dual-core machines) and other potential thread-safety issues.
6632 Changes between 0.9.7i and 0.9.7j [04 May 2006]
6634 *) Adapt fipsld and the build system to link against the validated FIPS
6635 module in FIPS mode.
6638 *) Fixes for VC++ 2005 build under Windows.
6641 *) Add new Windows build target VC-32-GMAKE for VC++. This uses GNU make
6642 from a Windows bash shell such as MSYS. It is autodetected from the
6643 "config" script when run from a VC++ environment. Modify standard VC++
6644 build to use fipscanister.o from the GNU make build.
6647 Changes between 0.9.7h and 0.9.7i [14 Oct 2005]
6649 *) Wrapped the definition of EVP_MAX_MD_SIZE in a #ifdef OPENSSL_FIPS.
6650 The value now differs depending on if you build for FIPS or not.
6651 BEWARE! A program linked with a shared FIPSed libcrypto can't be
6652 safely run with a non-FIPSed libcrypto, as it may crash because of
6653 the difference induced by this change.
6656 Changes between 0.9.7g and 0.9.7h [11 Oct 2005]
6658 *) Remove the functionality of SSL_OP_MSIE_SSLV2_RSA_PADDING
6659 (part of SSL_OP_ALL). This option used to disable the
6660 countermeasure against man-in-the-middle protocol-version
6661 rollback in the SSL 2.0 server implementation, which is a bad
6662 idea. (CVE-2005-2969)
6664 [Bodo Moeller; problem pointed out by Yutaka Oiwa (Research Center
6665 for Information Security, National Institute of Advanced Industrial
6666 Science and Technology [AIST], Japan)]
6668 *) Minimal support for X9.31 signatures and PSS padding modes. This is
6669 mainly for FIPS compliance and not fully integrated at this stage.
6672 *) For DSA signing, unless DSA_FLAG_NO_EXP_CONSTTIME is set, perform
6673 the exponentiation using a fixed-length exponent. (Otherwise,
6674 the information leaked through timing could expose the secret key
6675 after many signatures; cf. Bleichenbacher's attack on DSA with
6679 *) Make a new fixed-window mod_exp implementation the default for
6680 RSA, DSA, and DH private-key operations so that the sequence of
6681 squares and multiplies and the memory access pattern are
6682 independent of the particular secret key. This will mitigate
6683 cache-timing and potential related attacks.
6685 BN_mod_exp_mont_consttime() is the new exponentiation implementation,
6686 and this is automatically used by BN_mod_exp_mont() if the new flag
6687 BN_FLG_EXP_CONSTTIME is set for the exponent. RSA, DSA, and DH
6688 will use this BN flag for private exponents unless the flag
6689 RSA_FLAG_NO_EXP_CONSTTIME, DSA_FLAG_NO_EXP_CONSTTIME, or
6690 DH_FLAG_NO_EXP_CONSTTIME, respectively, is set.
6692 [Matthew D Wood (Intel Corp), with some changes by Bodo Moeller]
6694 *) Change the client implementation for SSLv23_method() and
6695 SSLv23_client_method() so that is uses the SSL 3.0/TLS 1.0
6696 Client Hello message format if the SSL_OP_NO_SSLv2 option is set.
6697 (Previously, the SSL 2.0 backwards compatible Client Hello
6698 message format would be used even with SSL_OP_NO_SSLv2.)
6701 *) Add support for smime-type MIME parameter in S/MIME messages which some
6705 *) New function BN_MONT_CTX_set_locked() to set montgomery parameters in
6706 a threadsafe manner. Modify rsa code to use new function and add calls
6707 to dsa and dh code (which had race conditions before).
6710 *) Include the fixed error library code in the C error file definitions
6711 instead of fixing them up at runtime. This keeps the error code
6712 structures constant.
6715 Changes between 0.9.7f and 0.9.7g [11 Apr 2005]
6717 [NB: OpenSSL 0.9.7h and later 0.9.7 patch levels were released after
6720 *) Fixes for newer kerberos headers. NB: the casts are needed because
6721 the 'length' field is signed on one version and unsigned on another
6722 with no (?) obvious way to tell the difference, without these VC++
6723 complains. Also the "definition" of FAR (blank) is no longer included
6724 nor is the error ENOMEM. KRB5_PRIVATE has to be set to 1 to pick up
6725 some needed definitions.
6728 *) Undo Cygwin change.
6731 *) Added support for proxy certificates according to RFC 3820.
6732 Because they may be a security thread to unaware applications,
6733 they must be explicitly allowed in run-time. See
6734 docs/HOWTO/proxy_certificates.txt for further information.
6737 Changes between 0.9.7e and 0.9.7f [22 Mar 2005]
6739 *) Use (SSL_RANDOM_VALUE - 4) bytes of pseudo random data when generating
6740 server and client random values. Previously
6741 (SSL_RANDOM_VALUE - sizeof(time_t)) would be used which would result in
6742 less random data when sizeof(time_t) > 4 (some 64 bit platforms).
6744 This change has negligible security impact because:
6746 1. Server and client random values still have 24 bytes of pseudo random
6749 2. Server and client random values are sent in the clear in the initial
6752 3. The master secret is derived using the premaster secret (48 bytes in
6753 size for static RSA ciphersuites) as well as client server and random
6756 The OpenSSL team would like to thank the UK NISCC for bringing this issue
6759 [Stephen Henson, reported by UK NISCC]
6761 *) Use Windows randomness collection on Cygwin.
6764 *) Fix hang in EGD/PRNGD query when communication socket is closed
6765 prematurely by EGD/PRNGD.
6766 [Darren Tucker <dtucker@zip.com.au> via Lutz Jänicke, resolves #1014]
6768 *) Prompt for pass phrases when appropriate for PKCS12 input format.
6771 *) Back-port of selected performance improvements from development
6772 branch, as well as improved support for PowerPC platforms.
6775 *) Add lots of checks for memory allocation failure, error codes to indicate
6776 failure and freeing up memory if a failure occurs.
6777 [Nauticus Networks SSL Team <openssl@nauticusnet.com>, Steve Henson]
6779 *) Add new -passin argument to dgst.
6782 *) Perform some character comparisons of different types in X509_NAME_cmp:
6783 this is needed for some certificates that re-encode DNs into UTF8Strings
6784 (in violation of RFC3280) and can't or won't issue name rollover
6788 *) Make an explicit check during certificate validation to see that
6789 the CA setting in each certificate on the chain is correct. As a
6790 side effect always do the following basic checks on extensions,
6791 not just when there's an associated purpose to the check:
6793 - if there is an unhandled critical extension (unless the user
6794 has chosen to ignore this fault)
6795 - if the path length has been exceeded (if one is set at all)
6796 - that certain extensions fit the associated purpose (if one has
6800 Changes between 0.9.7d and 0.9.7e [25 Oct 2004]
6802 *) Avoid a race condition when CRLs are checked in a multi threaded
6803 environment. This would happen due to the reordering of the revoked
6804 entries during signature checking and serial number lookup. Now the
6805 encoding is cached and the serial number sort performed under a lock.
6806 Add new STACK function sk_is_sorted().
6809 *) Add Delta CRL to the extension code.
6812 *) Various fixes to s3_pkt.c so alerts are sent properly.
6813 [David Holmes <d.holmes@f5.com>]
6815 *) Reduce the chances of duplicate issuer name and serial numbers (in
6816 violation of RFC3280) using the OpenSSL certificate creation utilities.
6817 This is done by creating a random 64 bit value for the initial serial
6818 number when a serial number file is created or when a self signed
6819 certificate is created using 'openssl req -x509'. The initial serial
6820 number file is created using 'openssl x509 -next_serial' in CA.pl
6821 rather than being initialized to 1.
6824 Changes between 0.9.7c and 0.9.7d [17 Mar 2004]
6826 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
6827 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
6828 [Joe Orton, Steve Henson]
6830 *) Fix flaw in SSL/TLS handshaking when using Kerberos ciphersuites
6832 [Joe Orton, Steve Henson]
6834 *) Make it possible to have multiple active certificates with the same
6835 subject in the CA index file. This is done only if the keyword
6836 'unique_subject' is set to 'no' in the main CA section (default
6837 if 'CA_default') of the configuration file. The value is saved
6838 with the database itself in a separate index attribute file,
6839 named like the index file with '.attr' appended to the name.
6842 *) X509 verify fixes. Disable broken certificate workarounds when
6843 X509_V_FLAGS_X509_STRICT is set. Check CRL issuer has cRLSign set if
6844 keyUsage extension present. Don't accept CRLs with unhandled critical
6845 extensions: since verify currently doesn't process CRL extensions this
6846 rejects a CRL with *any* critical extensions. Add new verify error codes
6850 *) When creating an OCSP nonce use an OCTET STRING inside the extnValue.
6851 A clarification of RFC2560 will require the use of OCTET STRINGs and
6852 some implementations cannot handle the current raw format. Since OpenSSL
6853 copies and compares OCSP nonces as opaque blobs without any attempt at
6854 parsing them this should not create any compatibility issues.
6857 *) New md flag EVP_MD_CTX_FLAG_REUSE this allows md_data to be reused when
6858 calling EVP_MD_CTX_copy_ex() to avoid calling OPENSSL_malloc(). Without
6859 this HMAC (and other) operations are several times slower than OpenSSL
6863 *) Print out GeneralizedTime and UTCTime in ASN1_STRING_print_ex().
6864 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
6866 *) Use the correct content when signing type "other".
6869 Changes between 0.9.7b and 0.9.7c [30 Sep 2003]
6871 *) Fix various bugs revealed by running the NISCC test suite:
6873 Stop out of bounds reads in the ASN1 code when presented with
6874 invalid tags (CVE-2003-0543 and CVE-2003-0544).
6876 Free up ASN1_TYPE correctly if ANY type is invalid (CVE-2003-0545).
6878 If verify callback ignores invalid public key errors don't try to check
6879 certificate signature with the NULL public key.
6883 *) New -ignore_err option in ocsp application to stop the server
6884 exiting on the first error in a request.
6887 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
6888 if the server requested one: as stated in TLS 1.0 and SSL 3.0
6892 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
6893 extra data after the compression methods not only for TLS 1.0
6894 but also for SSL 3.0 (as required by the specification).
6895 [Bodo Moeller; problem pointed out by Matthias Loepfe]
6897 *) Change X509_certificate_type() to mark the key as exported/exportable
6898 when it's 512 *bits* long, not 512 bytes.
6901 *) Change AES_cbc_encrypt() so it outputs exact multiple of
6902 blocks during encryption.
6905 *) Various fixes to base64 BIO and non blocking I/O. On write
6906 flushes were not handled properly if the BIO retried. On read
6907 data was not being buffered properly and had various logic bugs.
6908 This also affects blocking I/O when the data being decoded is a
6912 *) Various S/MIME bugfixes and compatibility changes:
6913 output correct application/pkcs7 MIME type if
6914 PKCS7_NOOLDMIMETYPE is set. Tolerate some broken signatures.
6915 Output CR+LF for EOL if PKCS7_CRLFEOL is set (this makes opening
6916 of files as .eml work). Correctly handle very long lines in MIME
6920 Changes between 0.9.7a and 0.9.7b [10 Apr 2003]
6922 *) Countermeasure against the Klima-Pokorny-Rosa extension of
6923 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
6924 a protocol version number mismatch like a decryption error
6925 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
6928 *) Turn on RSA blinding by default in the default implementation
6929 to avoid a timing attack. Applications that don't want it can call
6930 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
6931 They would be ill-advised to do so in most cases.
6932 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
6934 *) Change RSA blinding code so that it works when the PRNG is not
6935 seeded (in this case, the secret RSA exponent is abused as
6936 an unpredictable seed -- if it is not unpredictable, there
6937 is no point in blinding anyway). Make RSA blinding thread-safe
6938 by remembering the creator's thread ID in rsa->blinding and
6939 having all other threads use local one-time blinding factors
6940 (this requires more computation than sharing rsa->blinding, but
6941 avoids excessive locking; and if an RSA object is not shared
6942 between threads, blinding will still be very fast).
6945 *) Fixed a typo bug that would cause ENGINE_set_default() to set an
6946 ENGINE as defaults for all supported algorithms irrespective of
6947 the 'flags' parameter. 'flags' is now honoured, so applications
6948 should make sure they are passing it correctly.
6951 *) Target "mingw" now allows native Windows code to be generated in
6952 the Cygwin environment as well as with the MinGW compiler.
6955 Changes between 0.9.7 and 0.9.7a [19 Feb 2003]
6957 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
6958 via timing by performing a MAC computation even if incorrect
6959 block cipher padding has been found. This is a countermeasure
6960 against active attacks where the attacker has to distinguish
6961 between bad padding and a MAC verification error. (CVE-2003-0078)
6963 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
6964 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
6965 Martin Vuagnoux (EPFL, Ilion)]
6967 *) Make the no-err option work as intended. The intention with no-err
6968 is not to have the whole error stack handling routines removed from
6969 libcrypto, it's only intended to remove all the function name and
6970 reason texts, thereby removing some of the footprint that may not
6971 be interesting if those errors aren't displayed anyway.
6973 NOTE: it's still possible for any application or module to have it's
6974 own set of error texts inserted. The routines are there, just not
6975 used by default when no-err is given.
6978 *) Add support for FreeBSD on IA64.
6979 [dirk.meyer@dinoex.sub.org via Richard Levitte, resolves #454]
6981 *) Adjust DES_cbc_cksum() so it returns the same value as the MIT
6982 Kerberos function mit_des_cbc_cksum(). Before this change,
6983 the value returned by DES_cbc_cksum() was like the one from
6984 mit_des_cbc_cksum(), except the bytes were swapped.
6985 [Kevin Greaney <Kevin.Greaney@hp.com> and Richard Levitte]
6987 *) Allow an application to disable the automatic SSL chain building.
6988 Before this a rather primitive chain build was always performed in
6989 ssl3_output_cert_chain(): an application had no way to send the
6990 correct chain if the automatic operation produced an incorrect result.
6992 Now the chain builder is disabled if either:
6994 1. Extra certificates are added via SSL_CTX_add_extra_chain_cert().
6996 2. The mode flag SSL_MODE_NO_AUTO_CHAIN is set.
6998 The reasoning behind this is that an application would not want the
6999 auto chain building to take place if extra chain certificates are
7000 present and it might also want a means of sending no additional
7001 certificates (for example the chain has two certificates and the
7005 *) Add the possibility to build without the ENGINE framework.
7006 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7008 *) Under Win32 gmtime() can return NULL: check return value in
7009 OPENSSL_gmtime(). Add error code for case where gmtime() fails.
7012 *) DSA routines: under certain error conditions uninitialized BN objects
7013 could be freed. Solution: make sure initialization is performed early
7014 enough. (Reported and fix supplied by Ivan D Nestlerode <nestler@MIT.EDU>,
7015 Nils Larsch <nla@trustcenter.de> via PR#459)
7018 *) Another fix for SSLv2 session ID handling: the session ID was incorrectly
7019 checked on reconnect on the client side, therefore session resumption
7020 could still fail with a "ssl session id is different" error. This
7021 behaviour is masked when SSL_OP_ALL is used due to
7022 SSL_OP_MICROSOFT_SESS_ID_BUG being set.
7023 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7024 followup to PR #377.
7027 *) IA-32 assembler support enhancements: unified ELF targets, support
7028 for SCO/Caldera platforms, fix for Cygwin shared build.
7031 *) Add support for FreeBSD on sparc64. As a consequence, support for
7032 FreeBSD on non-x86 processors is separate from x86 processors on
7033 the config script, much like the NetBSD support.
7034 [Richard Levitte & Kris Kennaway <kris@obsecurity.org>]
7036 Changes between 0.9.6h and 0.9.7 [31 Dec 2002]
7038 [NB: OpenSSL 0.9.6i and later 0.9.6 patch levels were released after
7041 *) Fix session ID handling in SSLv2 client code: the SERVER FINISHED
7042 code (06) was taken as the first octet of the session ID and the last
7043 octet was ignored consequently. As a result SSLv2 client side session
7044 caching could not have worked due to the session ID mismatch between
7046 Behaviour observed by Crispin Flowerday <crispin@flowerday.cx> as
7050 *) Change the declaration of needed Kerberos libraries to use EX_LIBS
7051 instead of the special (and badly supported) LIBKRB5. LIBKRB5 is
7055 *) The hw_ncipher.c engine requires dynamic locks. Unfortunately, it
7056 seems that in spite of existing for more than a year, many application
7057 author have done nothing to provide the necessary callbacks, which
7058 means that this particular engine will not work properly anywhere.
7059 This is a very unfortunate situation which forces us, in the name
7060 of usability, to give the hw_ncipher.c a static lock, which is part
7062 NOTE: This is for the 0.9.7 series ONLY. This hack will never
7063 appear in 0.9.8 or later. We EXPECT application authors to have
7064 dealt properly with this when 0.9.8 is released (unless we actually
7065 make such changes in the libcrypto locking code that changes will
7066 have to be made anyway).
7069 *) In asn1_d2i_read_bio() repeatedly call BIO_read() until all content
7070 octets have been read, EOF or an error occurs. Without this change
7071 some truncated ASN1 structures will not produce an error.
7074 *) Disable Heimdal support, since it hasn't been fully implemented.
7075 Still give the possibility to force the use of Heimdal, but with
7076 warnings and a request that patches get sent to openssl-dev.
7079 *) Add the VC-CE target, introduce the WINCE sysname, and add
7080 INSTALL.WCE and appropriate conditionals to make it build.
7081 [Steven Reddie <smr@essemer.com.au> via Richard Levitte]
7083 *) Change the DLL names for Cygwin to cygcrypto-x.y.z.dll and
7084 cygssl-x.y.z.dll, where x, y and z are the major, minor and
7085 edit numbers of the version.
7086 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
7088 *) Introduce safe string copy and catenation functions
7089 (BUF_strlcpy() and BUF_strlcat()).
7090 [Ben Laurie (CHATS) and Richard Levitte]
7092 *) Avoid using fixed-size buffers for one-line DNs.
7093 [Ben Laurie (CHATS)]
7095 *) Add BUF_MEM_grow_clean() to avoid information leakage when
7096 resizing buffers containing secrets, and use where appropriate.
7097 [Ben Laurie (CHATS)]
7099 *) Avoid using fixed size buffers for configuration file location.
7100 [Ben Laurie (CHATS)]
7102 *) Avoid filename truncation for various CA files.
7103 [Ben Laurie (CHATS)]
7105 *) Use sizeof in preference to magic numbers.
7106 [Ben Laurie (CHATS)]
7108 *) Avoid filename truncation in cert requests.
7109 [Ben Laurie (CHATS)]
7111 *) Add assertions to check for (supposedly impossible) buffer
7113 [Ben Laurie (CHATS)]
7115 *) Don't cache truncated DNS entries in the local cache (this could
7116 potentially lead to a spoofing attack).
7117 [Ben Laurie (CHATS)]
7119 *) Fix various buffers to be large enough for hex/decimal
7120 representations in a platform independent manner.
7121 [Ben Laurie (CHATS)]
7123 *) Add CRYPTO_realloc_clean() to avoid information leakage when
7124 resizing buffers containing secrets, and use where appropriate.
7125 [Ben Laurie (CHATS)]
7127 *) Add BIO_indent() to avoid much slightly worrying code to do
7129 [Ben Laurie (CHATS)]
7131 *) Convert sprintf()/BIO_puts() to BIO_printf().
7132 [Ben Laurie (CHATS)]
7134 *) buffer_gets() could terminate with the buffer only half
7136 [Ben Laurie (CHATS)]
7138 *) Add assertions to prevent user-supplied crypto functions from
7139 overflowing internal buffers by having large block sizes, etc.
7140 [Ben Laurie (CHATS)]
7142 *) New OPENSSL_assert() macro (similar to assert(), but enabled
7144 [Ben Laurie (CHATS)]
7146 *) Eliminate unused copy of key in RC4.
7147 [Ben Laurie (CHATS)]
7149 *) Eliminate unused and incorrectly sized buffers for IV in pem.h.
7150 [Ben Laurie (CHATS)]
7152 *) Fix off-by-one error in EGD path.
7153 [Ben Laurie (CHATS)]
7155 *) If RANDFILE path is too long, ignore instead of truncating.
7156 [Ben Laurie (CHATS)]
7158 *) Eliminate unused and incorrectly sized X.509 structure
7160 [Ben Laurie (CHATS)]
7162 *) Eliminate unused and dangerous function knumber().
7163 [Ben Laurie (CHATS)]
7165 *) Eliminate unused and dangerous structure, KSSL_ERR.
7166 [Ben Laurie (CHATS)]
7168 *) Protect against overlong session ID context length in an encoded
7169 session object. Since these are local, this does not appear to be
7171 [Ben Laurie (CHATS)]
7173 *) Change from security patch (see 0.9.6e below) that did not affect
7174 the 0.9.6 release series:
7176 Remote buffer overflow in SSL3 protocol - an attacker could
7177 supply an oversized master key in Kerberos-enabled versions.
7179 [Ben Laurie (CHATS)]
7181 *) Change the SSL kerb5 codes to match RFC 2712.
7184 *) Make -nameopt work fully for req and add -reqopt switch.
7185 [Michael Bell <michael.bell@rz.hu-berlin.de>, Steve Henson]
7187 *) The "block size" for block ciphers in CFB and OFB mode should be 1.
7188 [Steve Henson, reported by Yngve Nysaeter Pettersen <yngve@opera.com>]
7190 *) Make sure tests can be performed even if the corresponding algorithms
7191 have been removed entirely. This was also the last step to make
7192 OpenSSL compilable with DJGPP under all reasonable conditions.
7193 [Richard Levitte, Doug Kaufman <dkaufman@rahul.net>]
7195 *) Add cipher selection rules COMPLEMENTOFALL and COMPLEMENTOFDEFAULT
7196 to allow version independent disabling of normally unselected ciphers,
7197 which may be activated as a side-effect of selecting a single cipher.
7199 (E.g., cipher list string "RSA" enables ciphersuites that are left
7200 out of "ALL" because they do not provide symmetric encryption.
7201 "RSA:!COMPLEMEMENTOFALL" avoids these unsafe ciphersuites.)
7202 [Lutz Jaenicke, Bodo Moeller]
7204 *) Add appropriate support for separate platform-dependent build
7205 directories. The recommended way to make a platform-dependent
7206 build directory is the following (tested on Linux), maybe with
7209 # Place yourself outside of the OpenSSL source tree. In
7210 # this example, the environment variable OPENSSL_SOURCE
7211 # is assumed to contain the absolute OpenSSL source directory.
7212 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
7213 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
7214 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
7215 mkdir -p `dirname $F`
7216 ln -s $OPENSSL_SOURCE/$F $F
7219 To be absolutely sure not to disturb the source tree, a "make clean"
7220 is a good thing. If it isn't successful, don't worry about it,
7221 it probably means the source directory is very clean.
7224 *) Make sure any ENGINE control commands make local copies of string
7225 pointers passed to them whenever necessary. Otherwise it is possible
7226 the caller may have overwritten (or deallocated) the original string
7227 data when a later ENGINE operation tries to use the stored values.
7228 [Götz Babin-Ebell <babinebell@trustcenter.de>]
7230 *) Improve diagnostics in file reading and command-line digests.
7231 [Ben Laurie aided and abetted by Solar Designer <solar@openwall.com>]
7233 *) Add AES modes CFB and OFB to the object database. Correct an
7234 error in AES-CFB decryption.
7237 *) Remove most calls to EVP_CIPHER_CTX_cleanup() in evp_enc.c, this
7238 allows existing EVP_CIPHER_CTX structures to be reused after
7239 calling EVP_*Final(). This behaviour is used by encryption
7240 BIOs and some applications. This has the side effect that
7241 applications must explicitly clean up cipher contexts with
7242 EVP_CIPHER_CTX_cleanup() or they will leak memory.
7245 *) Check the values of dna and dnb in bn_mul_recursive before calling
7246 bn_mul_comba (a non zero value means the a or b arrays do not contain
7247 n2 elements) and fallback to bn_mul_normal if either is not zero.
7250 *) Fix escaping of non-ASCII characters when using the -subj option
7251 of the "openssl req" command line tool. (Robert Joop <joop@fokus.gmd.de>)
7254 *) Make object definitions compliant to LDAP (RFC2256): SN is the short
7255 form for "surname", serialNumber has no short form.
7256 Use "mail" as the short name for "rfc822Mailbox" according to RFC2798;
7257 therefore remove "mail" short name for "internet 7".
7258 The OID for unique identifiers in X509 certificates is
7259 x500UniqueIdentifier, not uniqueIdentifier.
7260 Some more OID additions. (Michael Bell <michael.bell@rz.hu-berlin.de>)
7263 *) Add an "init" command to the ENGINE config module and auto initialize
7264 ENGINEs. Without any "init" command the ENGINE will be initialized
7265 after all ctrl commands have been executed on it. If init=1 the
7266 ENGINE is initialized at that point (ctrls before that point are run
7267 on the uninitialized ENGINE and after on the initialized one). If
7268 init=0 then the ENGINE will not be initialized at all.
7271 *) Fix the 'app_verify_callback' interface so that the user-defined
7272 argument is actually passed to the callback: In the
7273 SSL_CTX_set_cert_verify_callback() prototype, the callback
7274 declaration has been changed from
7277 int (*cb)(X509_STORE_CTX *,void *);
7278 in ssl_verify_cert_chain (ssl/ssl_cert.c), the call
7279 i=s->ctx->app_verify_callback(&ctx)
7280 has been changed into
7281 i=s->ctx->app_verify_callback(&ctx, s->ctx->app_verify_arg).
7283 To update applications using SSL_CTX_set_cert_verify_callback(),
7284 a dummy argument can be added to their callback functions.
7285 [D. K. Smetters <smetters@parc.xerox.com>]
7287 *) Added the '4758cca' ENGINE to support IBM 4758 cards.
7288 [Maurice Gittens <maurice@gittens.nl>, touchups by Geoff Thorpe]
7290 *) Add and OPENSSL_LOAD_CONF define which will cause
7291 OpenSSL_add_all_algorithms() to load the openssl.cnf config file.
7292 This allows older applications to transparently support certain
7293 OpenSSL features: such as crypto acceleration and dynamic ENGINE loading.
7294 Two new functions OPENSSL_add_all_algorithms_noconf() which will never
7295 load the config file and OPENSSL_add_all_algorithms_conf() which will
7296 always load it have also been added.
7299 *) Add the OFB, CFB and CTR (all with 128 bit feedback) to AES.
7300 Adjust NIDs and EVP layer.
7301 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7303 *) Config modules support in openssl utility.
7305 Most commands now load modules from the config file,
7306 though in a few (such as version) this isn't done
7307 because it couldn't be used for anything.
7309 In the case of ca and req the config file used is
7310 the same as the utility itself: that is the -config
7311 command line option can be used to specify an
7315 *) Move default behaviour from OPENSSL_config(). If appname is NULL
7316 use "openssl_conf" if filename is NULL use default openssl config file.
7319 *) Add an argument to OPENSSL_config() to allow the use of an alternative
7320 config section name. Add a new flag to tolerate a missing config file
7321 and move code to CONF_modules_load_file().
7324 *) Support for crypto accelerator cards from Accelerated Encryption
7325 Processing, www.aep.ie. (Use engine 'aep')
7326 The support was copied from 0.9.6c [engine] and adapted/corrected
7327 to work with the new engine framework.
7328 [AEP Inc. and Richard Levitte]
7330 *) Support for SureWare crypto accelerator cards from Baltimore
7331 Technologies. (Use engine 'sureware')
7332 The support was copied from 0.9.6c [engine] and adapted
7333 to work with the new engine framework.
7336 *) Have the CHIL engine fork-safe (as defined by nCipher) and actually
7337 make the newer ENGINE framework commands for the CHIL engine work.
7338 [Toomas Kiisk <vix@cyber.ee> and Richard Levitte]
7340 *) Make it possible to produce shared libraries on ReliantUNIX.
7341 [Robert Dahlem <Robert.Dahlem@ffm2.siemens.de> via Richard Levitte]
7343 *) Add the configuration target debug-linux-ppro.
7344 Make 'openssl rsa' use the general key loading routines
7345 implemented in apps.c, and make those routines able to
7346 handle the key format FORMAT_NETSCAPE and the variant
7348 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7350 *) Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
7351 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
7353 *) Add -keyform to rsautl, and document -engine.
7354 [Richard Levitte, inspired by Toomas Kiisk <vix@cyber.ee>]
7356 *) Change BIO_new_file (crypto/bio/bss_file.c) to use new
7357 BIO_R_NO_SUCH_FILE error code rather than the generic
7358 ERR_R_SYS_LIB error code if fopen() fails with ENOENT.
7361 *) Add new functions
7363 ERR_peek_last_error_line
7364 ERR_peek_last_error_line_data.
7365 These are similar to
7368 ERR_peek_error_line_data,
7369 but report on the latest error recorded rather than the first one
7370 still in the error queue.
7371 [Ben Laurie, Bodo Moeller]
7373 *) default_algorithms option in ENGINE config module. This allows things
7375 default_algorithms = ALL
7376 default_algorithms = RSA, DSA, RAND, CIPHERS, DIGESTS
7379 *) Preliminary ENGINE config module.
7382 *) New experimental application configuration code.
7385 *) Change the AES code to follow the same name structure as all other
7386 symmetric ciphers, and behave the same way. Move everything to
7387 the directory crypto/aes, thereby obsoleting crypto/rijndael.
7388 [Stephen Sprunk <stephen@sprunk.org> and Richard Levitte]
7390 *) SECURITY: remove unsafe setjmp/signal interaction from ui_openssl.c.
7391 [Ben Laurie and Theo de Raadt]
7393 *) Add option to output public keys in req command.
7394 [Massimiliano Pala madwolf@openca.org]
7396 *) Use wNAFs in EC_POINTs_mul() for improved efficiency
7397 (up to about 10% better than before for P-192 and P-224).
7400 *) New functions/macros
7402 SSL_CTX_set_msg_callback(ctx, cb)
7403 SSL_CTX_set_msg_callback_arg(ctx, arg)
7404 SSL_set_msg_callback(ssl, cb)
7405 SSL_set_msg_callback_arg(ssl, arg)
7407 to request calling a callback function
7409 void cb(int write_p, int version, int content_type,
7410 const void *buf, size_t len, SSL *ssl, void *arg)
7412 whenever a protocol message has been completely received
7413 (write_p == 0) or sent (write_p == 1). Here 'version' is the
7414 protocol version according to which the SSL library interprets
7415 the current protocol message (SSL2_VERSION, SSL3_VERSION, or
7416 TLS1_VERSION). 'content_type' is 0 in the case of SSL 2.0, or
7417 the content type as defined in the SSL 3.0/TLS 1.0 protocol
7418 specification (change_cipher_spec(20), alert(21), handshake(22)).
7419 'buf' and 'len' point to the actual message, 'ssl' to the
7420 SSL object, and 'arg' is the application-defined value set by
7421 SSL[_CTX]_set_msg_callback_arg().
7423 'openssl s_client' and 'openssl s_server' have new '-msg' options
7424 to enable a callback that displays all protocol messages.
7427 *) Change the shared library support so shared libraries are built as
7428 soon as the corresponding static library is finished, and thereby get
7429 openssl and the test programs linked against the shared library.
7430 This still only happens when the keyword "shard" has been given to
7431 the configuration scripts.
7433 NOTE: shared library support is still an experimental thing, and
7434 backward binary compatibility is still not guaranteed.
7435 ["Maciej W. Rozycki" <macro@ds2.pg.gda.pl> and Richard Levitte]
7437 *) Add support for Subject Information Access extension.
7438 [Peter Sylvester <Peter.Sylvester@EdelWeb.fr>]
7440 *) Make BUF_MEM_grow() behaviour more consistent: Initialise to zero
7441 additional bytes when new memory had to be allocated, not just
7442 when reusing an existing buffer.
7445 *) New command line and configuration option 'utf8' for the req command.
7446 This allows field values to be specified as UTF8 strings.
7449 *) Add -multi and -mr options to "openssl speed" - giving multiple parallel
7450 runs for the former and machine-readable output for the latter.
7453 *) Add '-noemailDN' option to 'openssl ca'. This prevents inclusion
7454 of the e-mail address in the DN (i.e., it will go into a certificate
7455 extension only). The new configuration file option 'email_in_dn = no'
7456 has the same effect.
7457 [Massimiliano Pala madwolf@openca.org]
7459 *) Change all functions with names starting with des_ to be starting
7460 with DES_ instead. Add wrappers that are compatible with libdes,
7461 but are named _ossl_old_des_*. Finally, add macros that map the
7462 des_* symbols to the corresponding _ossl_old_des_* if libdes
7463 compatibility is desired. If OpenSSL 0.9.6c compatibility is
7464 desired, the des_* symbols will be mapped to DES_*, with one
7467 Since we provide two compatibility mappings, the user needs to
7468 define the macro OPENSSL_DES_LIBDES_COMPATIBILITY if libdes
7469 compatibility is desired. The default (i.e., when that macro
7470 isn't defined) is OpenSSL 0.9.6c compatibility.
7472 There are also macros that enable and disable the support of old
7473 des functions altogether. Those are OPENSSL_ENABLE_OLD_DES_SUPPORT
7474 and OPENSSL_DISABLE_OLD_DES_SUPPORT. If none or both of those
7475 are defined, the default will apply: to support the old des routines.
7477 In either case, one must include openssl/des.h to get the correct
7478 definitions. Do not try to just include openssl/des_old.h, that
7481 NOTE: This is a major break of an old API into a new one. Software
7482 authors are encouraged to switch to the DES_ style functions. Some
7483 time in the future, des_old.h and the libdes compatibility functions
7484 will be disable (i.e. OPENSSL_DISABLE_OLD_DES_SUPPORT will be the
7485 default), and then completely removed.
7488 *) Test for certificates which contain unsupported critical extensions.
7489 If such a certificate is found during a verify operation it is
7490 rejected by default: this behaviour can be overridden by either
7491 handling the new error X509_V_ERR_UNHANDLED_CRITICAL_EXTENSION or
7492 by setting the verify flag X509_V_FLAG_IGNORE_CRITICAL. A new function
7493 X509_supported_extension() has also been added which returns 1 if a
7494 particular extension is supported.
7497 *) Modify the behaviour of EVP cipher functions in similar way to digests
7498 to retain compatibility with existing code.
7501 *) Modify the behaviour of EVP_DigestInit() and EVP_DigestFinal() to retain
7502 compatibility with existing code. In particular the 'ctx' parameter does
7503 not have to be to be initialized before the call to EVP_DigestInit() and
7504 it is tidied up after a call to EVP_DigestFinal(). New function
7505 EVP_DigestFinal_ex() which does not tidy up the ctx. Similarly function
7506 EVP_MD_CTX_copy() changed to not require the destination to be
7507 initialized valid and new function EVP_MD_CTX_copy_ex() added which
7508 requires the destination to be valid.
7510 Modify all the OpenSSL digest calls to use EVP_DigestInit_ex(),
7511 EVP_DigestFinal_ex() and EVP_MD_CTX_copy_ex().
7514 *) Change ssl3_get_message (ssl/s3_both.c) and the functions using it
7515 so that complete 'Handshake' protocol structures are kept in memory
7516 instead of overwriting 'msg_type' and 'length' with 'body' data.
7519 *) Add an implementation of SSL_add_dir_cert_subjects_to_stack for Win32.
7520 [Massimo Santin via Richard Levitte]
7522 *) Major restructuring to the underlying ENGINE code. This includes
7523 reduction of linker bloat, separation of pure "ENGINE" manipulation
7524 (initialisation, etc) from functionality dealing with implementations
7525 of specific crypto interfaces. This change also introduces integrated
7526 support for symmetric ciphers and digest implementations - so ENGINEs
7527 can now accelerate these by providing EVP_CIPHER and EVP_MD
7528 implementations of their own. This is detailed in crypto/engine/README
7529 as it couldn't be adequately described here. However, there are a few
7530 API changes worth noting - some RSA, DSA, DH, and RAND functions that
7531 were changed in the original introduction of ENGINE code have now
7532 reverted back - the hooking from this code to ENGINE is now a good
7533 deal more passive and at run-time, operations deal directly with
7534 RSA_METHODs, DSA_METHODs (etc) as they did before, rather than
7535 dereferencing through an ENGINE pointer any more. Also, the ENGINE
7536 functions dealing with BN_MOD_EXP[_CRT] handlers have been removed -
7537 they were not being used by the framework as there is no concept of a
7538 BIGNUM_METHOD and they could not be generalised to the new
7539 'ENGINE_TABLE' mechanism that underlies the new code. Similarly,
7540 ENGINE_cpy() has been removed as it cannot be consistently defined in
7544 *) Change ASN1_GENERALIZEDTIME_check() to allow fractional seconds.
7547 *) Change mkdef.pl to sort symbols that get the same entry number,
7548 and make sure the automatically generated functions ERR_load_*
7549 become part of libeay.num as well.
7552 *) New function SSL_renegotiate_pending(). This returns true once
7553 renegotiation has been requested (either SSL_renegotiate() call
7554 or HelloRequest/ClientHello received from the peer) and becomes
7555 false once a handshake has been completed.
7556 (For servers, SSL_renegotiate() followed by SSL_do_handshake()
7557 sends a HelloRequest, but does not ensure that a handshake takes
7558 place. SSL_renegotiate_pending() is useful for checking if the
7559 client has followed the request.)
7562 *) New SSL option SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION.
7563 By default, clients may request session resumption even during
7564 renegotiation (if session ID contexts permit); with this option,
7565 session resumption is possible only in the first handshake.
7567 SSL_OP_ALL is now 0x00000FFFL instead of 0x000FFFFFL. This makes
7568 more bits available for options that should not be part of
7569 SSL_OP_ALL (such as SSL_OP_NO_SESSION_RESUMPTION_ON_RENEGOTIATION).
7572 *) Add some demos for certificate and certificate request creation.
7575 *) Make maximum certificate chain size accepted from the peer application
7576 settable (SSL*_get/set_max_cert_list()), as proposed by
7577 "Douglas E. Engert" <deengert@anl.gov>.
7580 *) Add support for shared libraries for Unixware-7
7581 (Boyd Lynn Gerber <gerberb@zenez.com>).
7584 *) Add a "destroy" handler to ENGINEs that allows structural cleanup to
7585 be done prior to destruction. Use this to unload error strings from
7586 ENGINEs that load their own error strings. NB: This adds two new API
7587 functions to "get" and "set" this destroy handler in an ENGINE.
7590 *) Alter all existing ENGINE implementations (except "openssl" and
7591 "openbsd") to dynamically instantiate their own error strings. This
7592 makes them more flexible to be built both as statically-linked ENGINEs
7593 and self-contained shared-libraries loadable via the "dynamic" ENGINE.
7594 Also, add stub code to each that makes building them as self-contained
7595 shared-libraries easier (see README.ENGINE).
7598 *) Add a "dynamic" ENGINE that provides a mechanism for binding ENGINE
7599 implementations into applications that are completely implemented in
7600 self-contained shared-libraries. The "dynamic" ENGINE exposes control
7601 commands that can be used to configure what shared-library to load and
7602 to control aspects of the way it is handled. Also, made an update to
7603 the README.ENGINE file that brings its information up-to-date and
7604 provides some information and instructions on the "dynamic" ENGINE
7605 (ie. how to use it, how to build "dynamic"-loadable ENGINEs, etc).
7608 *) Make it possible to unload ranges of ERR strings with a new
7609 "ERR_unload_strings" function.
7612 *) Add a copy() function to EVP_MD.
7615 *) Make EVP_MD routines take a context pointer instead of just the
7616 md_data void pointer.
7619 *) Add flags to EVP_MD and EVP_MD_CTX. EVP_MD_FLAG_ONESHOT indicates
7620 that the digest can only process a single chunk of data
7621 (typically because it is provided by a piece of
7622 hardware). EVP_MD_CTX_FLAG_ONESHOT indicates that the application
7623 is only going to provide a single chunk of data, and hence the
7624 framework needn't accumulate the data for oneshot drivers.
7627 *) As with "ERR", make it possible to replace the underlying "ex_data"
7628 functions. This change also alters the storage and management of global
7629 ex_data state - it's now all inside ex_data.c and all "class" code (eg.
7630 RSA, BIO, SSL_CTX, etc) no longer stores its own STACKS and per-class
7631 index counters. The API functions that use this state have been changed
7632 to take a "class_index" rather than pointers to the class's local STACK
7633 and counter, and there is now an API function to dynamically create new
7634 classes. This centralisation allows us to (a) plug a lot of the
7635 thread-safety problems that existed, and (b) makes it possible to clean
7636 up all allocated state using "CRYPTO_cleanup_all_ex_data()". W.r.t. (b)
7637 such data would previously have always leaked in application code and
7638 workarounds were in place to make the memory debugging turn a blind eye
7639 to it. Application code that doesn't use this new function will still
7640 leak as before, but their memory debugging output will announce it now
7641 rather than letting it slide.
7643 Besides the addition of CRYPTO_cleanup_all_ex_data(), another API change
7644 induced by the "ex_data" overhaul is that X509_STORE_CTX_init() now
7645 has a return value to indicate success or failure.
7648 *) Make it possible to replace the underlying "ERR" functions such that the
7649 global state (2 LHASH tables and 2 locks) is only used by the "default"
7650 implementation. This change also adds two functions to "get" and "set"
7651 the implementation prior to it being automatically set the first time
7652 any other ERR function takes place. Ie. an application can call "get",
7653 pass the return value to a module it has just loaded, and that module
7654 can call its own "set" function using that value. This means the
7655 module's "ERR" operations will use (and modify) the error state in the
7656 application and not in its own statically linked copy of OpenSSL code.
7659 *) Give DH, DSA, and RSA types their own "**_up_ref()" function to increment
7660 reference counts. This performs normal REF_PRINT/REF_CHECK macros on
7661 the operation, and provides a more encapsulated way for external code
7662 (crypto/evp/ and ssl/) to do this. Also changed the evp and ssl code
7663 to use these functions rather than manually incrementing the counts.
7665 Also rename "DSO_up()" function to more descriptive "DSO_up_ref()".
7668 *) Add EVP test program.
7671 *) Add symmetric cipher support to ENGINE. Expect the API to change!
7674 *) New CRL functions: X509_CRL_set_version(), X509_CRL_set_issuer_name()
7675 X509_CRL_set_lastUpdate(), X509_CRL_set_nextUpdate(), X509_CRL_sort(),
7676 X509_REVOKED_set_serialNumber(), and X509_REVOKED_set_revocationDate().
7677 These allow a CRL to be built without having to access X509_CRL fields
7678 directly. Modify 'ca' application to use new functions.
7681 *) Move SSL_OP_TLS_ROLLBACK_BUG out of the SSL_OP_ALL list of recommended
7682 bug workarounds. Rollback attack detection is a security feature.
7683 The problem will only arise on OpenSSL servers when TLSv1 is not
7684 available (sslv3_server_method() or SSL_OP_NO_TLSv1).
7685 Software authors not wanting to support TLSv1 will have special reasons
7686 for their choice and can explicitly enable this option.
7687 [Bodo Moeller, Lutz Jaenicke]
7689 *) Rationalise EVP so it can be extended: don't include a union of
7690 cipher/digest structures, add init/cleanup functions for EVP_MD_CTX
7691 (similar to those existing for EVP_CIPHER_CTX).
7696 EVP_MD_CTX_init(&md); /* new function call */
7697 EVP_DigestInit(&md, EVP_sha1());
7698 EVP_DigestUpdate(&md, in, len);
7699 EVP_DigestFinal(&md, out, NULL);
7700 EVP_MD_CTX_cleanup(&md); /* new function call */
7704 *) Make DES key schedule conform to the usual scheme, as well as
7705 correcting its structure. This means that calls to DES functions
7706 now have to pass a pointer to a des_key_schedule instead of a
7707 plain des_key_schedule (which was actually always a pointer
7710 des_key_schedule ks;
7712 des_set_key_checked(..., &ks);
7713 des_ncbc_encrypt(..., &ks, ...);
7715 (Note that a later change renames 'des_...' into 'DES_...'.)
7718 *) Initial reduction of linker bloat: the use of some functions, such as
7719 PEM causes large amounts of unused functions to be linked in due to
7720 poor organisation. For example pem_all.c contains every PEM function
7721 which has a knock on effect of linking in large amounts of (unused)
7722 ASN1 code. Grouping together similar functions and splitting unrelated
7723 functions prevents this.
7726 *) Cleanup of EVP macros.
7729 *) Change historical references to {NID,SN,LN}_des_ede and ede3 to add the
7730 correct _ecb suffix.
7733 *) Add initial OCSP responder support to ocsp application. The
7734 revocation information is handled using the text based index
7735 use by the ca application. The responder can either handle
7736 requests generated internally, supplied in files (for example
7737 via a CGI script) or using an internal minimal server.
7740 *) Add configuration choices to get zlib compression for TLS.
7743 *) Changes to Kerberos SSL for RFC 2712 compliance:
7744 1. Implemented real KerberosWrapper, instead of just using
7745 KRB5 AP_REQ message. [Thanks to Simon Wilkinson <sxw@sxw.org.uk>]
7746 2. Implemented optional authenticator field of KerberosWrapper.
7748 Added openssl-style ASN.1 macros for Kerberos ticket, ap_req,
7749 and authenticator structs; see crypto/krb5/.
7751 Generalized Kerberos calls to support multiple Kerberos libraries.
7752 [Vern Staats <staatsvr@asc.hpc.mil>,
7753 Jeffrey Altman <jaltman@columbia.edu>
7754 via Richard Levitte]
7756 *) Cause 'openssl speed' to use fully hard-coded DSA keys as it
7757 already does with RSA. testdsa.h now has 'priv_key/pub_key'
7758 values for each of the key sizes rather than having just
7759 parameters (and 'speed' generating keys each time).
7762 *) Speed up EVP routines.
7765 type 8 bytes 64 bytes 256 bytes 1024 bytes 8192 bytes
7766 des-cbc 4408.85k 5560.51k 5778.46k 5862.20k 5825.16k
7767 des-cbc 4389.55k 5571.17k 5792.23k 5846.91k 5832.11k
7768 des-cbc 4394.32k 5575.92k 5807.44k 5848.37k 5841.30k
7770 des-cbc 3482.66k 5069.49k 5496.39k 5614.16k 5639.28k
7771 des-cbc 3480.74k 5068.76k 5510.34k 5609.87k 5635.52k
7772 des-cbc 3483.72k 5067.62k 5504.60k 5708.01k 5724.80k
7775 des-cbc 4660.16k 5650.19k 5807.19k 5827.13k 5783.32k
7777 des-cbc 3624.96k 5258.21k 5530.91k 5624.30k 5628.26k
7780 *) Added the OS2-EMX target.
7781 ["Brian Havard" <brianh@kheldar.apana.org.au> and Richard Levitte]
7783 *) Rewrite apps to use NCONF routines instead of the old CONF. New functions
7784 to support NCONF routines in extension code. New function CONF_set_nconf()
7785 to allow functions which take an NCONF to also handle the old LHASH
7786 structure: this means that the old CONF compatible routines can be
7787 retained (in particular wrt extensions) without having to duplicate the
7788 code. New function X509V3_add_ext_nconf_sk to add extensions to a stack.
7791 *) Enhance the general user interface with mechanisms for inner control
7792 and with possibilities to have yes/no kind of prompts.
7795 *) Change all calls to low level digest routines in the library and
7796 applications to use EVP. Add missing calls to HMAC_cleanup() and
7797 don't assume HMAC_CTX can be copied using memcpy().
7798 [Verdon Walker <VWalker@novell.com>, Steve Henson]
7800 *) Add the possibility to control engines through control names but with
7801 arbitrary arguments instead of just a string.
7802 Change the key loaders to take a UI_METHOD instead of a callback
7803 function pointer. NOTE: this breaks binary compatibility with earlier
7804 versions of OpenSSL [engine].
7805 Adapt the nCipher code for these new conditions and add a card insertion
7809 *) Enhance the general user interface with mechanisms to better support
7810 dialog box interfaces, application-defined prompts, the possibility
7811 to use defaults (for example default passwords from somewhere else)
7812 and interrupts/cancellations.
7815 *) Tidy up PKCS#12 attribute handling. Add support for the CSP name
7816 attribute in PKCS#12 files, add new -CSP option to pkcs12 utility.
7819 *) Fix a memory leak in 'sk_dup()' in the case reallocation fails. (Also
7820 tidy up some unnecessarily weird code in 'sk_new()').
7821 [Geoff, reported by Diego Tartara <dtartara@novamens.com>]
7823 *) Change the key loading routines for ENGINEs to use the same kind
7824 callback (pem_password_cb) as all other routines that need this
7828 *) Increase ENTROPY_NEEDED to 32 bytes, as Rijndael can operate with
7829 256 bit (=32 byte) keys. Of course seeding with more entropy bytes
7830 than this minimum value is recommended.
7833 *) New random seeder for OpenVMS, using the system process statistics
7834 that are easily reachable.
7837 *) Windows apparently can't transparently handle global
7838 variables defined in DLLs. Initialisations such as:
7840 const ASN1_ITEM *it = &ASN1_INTEGER_it;
7842 won't compile. This is used by the any applications that need to
7843 declare their own ASN1 modules. This was fixed by adding the option
7844 EXPORT_VAR_AS_FN to all Win32 platforms, although this isn't strictly
7845 needed for static libraries under Win32.
7848 *) New functions X509_PURPOSE_set() and X509_TRUST_set() to handle
7849 setting of purpose and trust fields. New X509_STORE trust and
7850 purpose functions and tidy up setting in other SSL functions.
7853 *) Add copies of X509_STORE_CTX fields and callbacks to X509_STORE
7854 structure. These are inherited by X509_STORE_CTX when it is
7855 initialised. This allows various defaults to be set in the
7856 X509_STORE structure (such as flags for CRL checking and custom
7857 purpose or trust settings) for functions which only use X509_STORE_CTX
7858 internally such as S/MIME.
7860 Modify X509_STORE_CTX_purpose_inherit() so it only sets purposes and
7861 trust settings if they are not set in X509_STORE. This allows X509_STORE
7862 purposes and trust (in S/MIME for example) to override any set by default.
7864 Add command line options for CRL checking to smime, s_client and s_server
7868 *) Initial CRL based revocation checking. If the CRL checking flag(s)
7869 are set then the CRL is looked up in the X509_STORE structure and
7870 its validity and signature checked, then if the certificate is found
7871 in the CRL the verify fails with a revoked error.
7873 Various new CRL related callbacks added to X509_STORE_CTX structure.
7875 Command line options added to 'verify' application to support this.
7877 This needs some additional work, such as being able to handle multiple
7878 CRLs with different times, extension based lookup (rather than just
7879 by subject name) and ultimately more complete V2 CRL extension
7883 *) Add a general user interface API (crypto/ui/). This is designed
7884 to replace things like des_read_password and friends (backward
7885 compatibility functions using this new API are provided).
7886 The purpose is to remove prompting functions from the DES code
7887 section as well as provide for prompting through dialog boxes in
7888 a window system and the like.
7891 *) Add "ex_data" support to ENGINE so implementations can add state at a
7892 per-structure level rather than having to store it globally.
7895 *) Make it possible for ENGINE structures to be copied when retrieved by
7896 ENGINE_by_id() if the ENGINE specifies a new flag: ENGINE_FLAGS_BY_ID_COPY.
7897 This causes the "original" ENGINE structure to act like a template,
7898 analogous to the RSA vs. RSA_METHOD type of separation. Because of this
7899 operational state can be localised to each ENGINE structure, despite the
7900 fact they all share the same "methods". New ENGINE structures returned in
7901 this case have no functional references and the return value is the single
7902 structural reference. This matches the single structural reference returned
7903 by ENGINE_by_id() normally, when it is incremented on the pre-existing
7907 *) Fix ASN1 decoder when decoding type ANY and V_ASN1_OTHER: since this
7908 needs to match any other type at all we need to manually clear the
7912 *) Changes to the "openssl engine" utility to include;
7913 - verbosity levels ('-v', '-vv', and '-vvv') that provide information
7914 about an ENGINE's available control commands.
7915 - executing control commands from command line arguments using the
7916 '-pre' and '-post' switches. '-post' is only used if '-t' is
7917 specified and the ENGINE is successfully initialised. The syntax for
7918 the individual commands are colon-separated, for example;
7919 openssl engine chil -pre FORK_CHECK:0 -pre SO_PATH:/lib/test.so
7922 *) New dynamic control command support for ENGINEs. ENGINEs can now
7923 declare their own commands (numbers), names (strings), descriptions,
7924 and input types for run-time discovery by calling applications. A
7925 subset of these commands are implicitly classed as "executable"
7926 depending on their input type, and only these can be invoked through
7927 the new string-based API function ENGINE_ctrl_cmd_string(). (Eg. this
7928 can be based on user input, config files, etc). The distinction is
7929 that "executable" commands cannot return anything other than a boolean
7930 result and can only support numeric or string input, whereas some
7931 discoverable commands may only be for direct use through
7932 ENGINE_ctrl(), eg. supporting the exchange of binary data, function
7933 pointers, or other custom uses. The "executable" commands are to
7934 support parameterisations of ENGINE behaviour that can be
7935 unambiguously defined by ENGINEs and used consistently across any
7936 OpenSSL-based application. Commands have been added to all the
7937 existing hardware-supporting ENGINEs, noticeably "SO_PATH" to allow
7938 control over shared-library paths without source code alterations.
7941 *) Changed all ENGINE implementations to dynamically allocate their
7942 ENGINEs rather than declaring them statically. Apart from this being
7943 necessary with the removal of the ENGINE_FLAGS_MALLOCED distinction,
7944 this also allows the implementations to compile without using the
7945 internal engine_int.h header.
7948 *) Minor adjustment to "rand" code. RAND_get_rand_method() now returns a
7949 'const' value. Any code that should be able to modify a RAND_METHOD
7950 should already have non-const pointers to it (ie. they should only
7951 modify their own ones).
7954 *) Made a variety of little tweaks to the ENGINE code.
7955 - "atalla" and "ubsec" string definitions were moved from header files
7956 to C code. "nuron" string definitions were placed in variables
7957 rather than hard-coded - allowing parameterisation of these values
7958 later on via ctrl() commands.
7959 - Removed unused "#if 0"'d code.
7960 - Fixed engine list iteration code so it uses ENGINE_free() to release
7961 structural references.
7962 - Constified the RAND_METHOD element of ENGINE structures.
7963 - Constified various get/set functions as appropriate and added
7964 missing functions (including a catch-all ENGINE_cpy that duplicates
7965 all ENGINE values onto a new ENGINE except reference counts/state).
7966 - Removed NULL parameter checks in get/set functions. Setting a method
7967 or function to NULL is a way of cancelling out a previously set
7968 value. Passing a NULL ENGINE parameter is just plain stupid anyway
7969 and doesn't justify the extra error symbols and code.
7970 - Deprecate the ENGINE_FLAGS_MALLOCED define and move the area for
7971 flags from engine_int.h to engine.h.
7972 - Changed prototypes for ENGINE handler functions (init(), finish(),
7973 ctrl(), key-load functions, etc) to take an (ENGINE*) parameter.
7976 *) Implement binary inversion algorithm for BN_mod_inverse in addition
7977 to the algorithm using long division. The binary algorithm can be
7978 used only if the modulus is odd. On 32-bit systems, it is faster
7979 only for relatively small moduli (roughly 20-30% for 128-bit moduli,
7980 roughly 5-15% for 256-bit moduli), so we use it only for moduli
7981 up to 450 bits. In 64-bit environments, the binary algorithm
7982 appears to be advantageous for much longer moduli; here we use it
7983 for moduli up to 2048 bits.
7986 *) Rewrite CHOICE field setting in ASN1_item_ex_d2i(). The old code
7987 could not support the combine flag in choice fields.
7990 *) Add a 'copy_extensions' option to the 'ca' utility. This copies
7991 extensions from a certificate request to the certificate.
7994 *) Allow multiple 'certopt' and 'nameopt' options to be separated
7995 by commas. Add 'namopt' and 'certopt' options to the 'ca' config
7996 file: this allows the display of the certificate about to be
7997 signed to be customised, to allow certain fields to be included
7998 or excluded and extension details. The old system didn't display
7999 multicharacter strings properly, omitted fields not in the policy
8000 and couldn't display additional details such as extensions.
8003 *) Function EC_POINTs_mul for multiple scalar multiplication
8004 of an arbitrary number of elliptic curve points
8005 \sum scalars[i]*points[i],
8006 optionally including the generator defined for the EC_GROUP:
8007 scalar*generator + \sum scalars[i]*points[i].
8009 EC_POINT_mul is a simple wrapper function for the typical case
8010 that the point list has just one item (besides the optional
8014 *) First EC_METHODs for curves over GF(p):
8016 EC_GFp_simple_method() uses the basic BN_mod_mul and BN_mod_sqr
8017 operations and provides various method functions that can also
8018 operate with faster implementations of modular arithmetic.
8020 EC_GFp_mont_method() reuses most functions that are part of
8021 EC_GFp_simple_method, but uses Montgomery arithmetic.
8023 [Bodo Moeller; point addition and point doubling
8024 implementation directly derived from source code provided by
8025 Lenka Fibikova <fibikova@exp-math.uni-essen.de>]
8027 *) Framework for elliptic curves (crypto/ec/ec.h, crypto/ec/ec_lcl.h,
8028 crypto/ec/ec_lib.c):
8030 Curves are EC_GROUP objects (with an optional group generator)
8031 based on EC_METHODs that are built into the library.
8033 Points are EC_POINT objects based on EC_GROUP objects.
8035 Most of the framework would be able to handle curves over arbitrary
8036 finite fields, but as there are no obvious types for fields other
8037 than GF(p), some functions are limited to that for now.
8040 *) Add the -HTTP option to s_server. It is similar to -WWW, but requires
8041 that the file contains a complete HTTP response.
8044 *) Add the ec directory to mkdef.pl and mkfiles.pl. In mkdef.pl
8045 change the def and num file printf format specifier from "%-40sXXX"
8046 to "%-39s XXX". The latter will always guarantee a space after the
8047 field while the former will cause them to run together if the field
8048 is 40 of more characters long.
8051 *) Constify the cipher and digest 'method' functions and structures
8052 and modify related functions to take constant EVP_MD and EVP_CIPHER
8056 *) Hide BN_CTX structure details in bn_lcl.h instead of publishing them
8057 in <openssl/bn.h>. Also further increase BN_CTX_NUM to 32.
8060 *) Modify EVP_Digest*() routines so they now return values. Although the
8061 internal software routines can never fail additional hardware versions
8065 *) Clean up crypto/err/err.h and change some error codes to avoid conflicts:
8067 Previously ERR_R_FATAL was too small and coincided with ERR_LIB_PKCS7
8068 (= ERR_R_PKCS7_LIB); it is now 64 instead of 32.
8071 ERR_R_NESTED_ASN1_ERROR
8073 ERR_R_MISSING_ASN1_EOS
8074 were 4 .. 9, conflicting with
8075 ERR_LIB_RSA (= ERR_R_RSA_LIB)
8077 ERR_LIB_PEM (= ERR_R_PEM_LIB).
8078 They are now 58 .. 63 (i.e., just below ERR_R_FATAL).
8080 Add new error code 'ERR_R_INTERNAL_ERROR'.
8083 *) Don't overuse locks in crypto/err/err.c: For data retrieval, CRYPTO_r_lock
8087 *) New option '-subj arg' for 'openssl req' and 'openssl ca'. This
8088 sets the subject name for a new request or supersedes the
8089 subject name in a given request. Formats that can be parsed are
8090 'CN=Some Name, OU=myOU, C=IT'
8092 'CN=Some Name/OU=myOU/C=IT'.
8094 Add options '-batch' and '-verbose' to 'openssl req'.
8095 [Massimiliano Pala <madwolf@hackmasters.net>]
8097 *) Introduce the possibility to access global variables through
8098 functions on platform were that's the best way to handle exporting
8099 global variables in shared libraries. To enable this functionality,
8100 one must configure with "EXPORT_VAR_AS_FN" or defined the C macro
8101 "OPENSSL_EXPORT_VAR_AS_FUNCTION" in crypto/opensslconf.h (the latter
8102 is normally done by Configure or something similar).
8104 To implement a global variable, use the macro OPENSSL_IMPLEMENT_GLOBAL
8105 in the source file (foo.c) like this:
8107 OPENSSL_IMPLEMENT_GLOBAL(int,foo)=1;
8108 OPENSSL_IMPLEMENT_GLOBAL(double,bar);
8110 To declare a global variable, use the macros OPENSSL_DECLARE_GLOBAL
8111 and OPENSSL_GLOBAL_REF in the header file (foo.h) like this:
8113 OPENSSL_DECLARE_GLOBAL(int,foo);
8114 #define foo OPENSSL_GLOBAL_REF(foo)
8115 OPENSSL_DECLARE_GLOBAL(double,bar);
8116 #define bar OPENSSL_GLOBAL_REF(bar)
8118 The #defines are very important, and therefore so is including the
8119 header file everywhere where the defined globals are used.
8121 The macro OPENSSL_EXPORT_VAR_AS_FUNCTION also affects the definition
8122 of ASN.1 items, but that structure is a bit different.
8124 The largest change is in util/mkdef.pl which has been enhanced with
8125 better and easier to understand logic to choose which symbols should
8126 go into the Windows .def files as well as a number of fixes and code
8127 cleanup (among others, algorithm keywords are now sorted
8128 lexicographically to avoid constant rewrites).
8131 *) In BN_div() keep a copy of the sign of 'num' before writing the
8132 result to 'rm' because if rm==num the value will be overwritten
8133 and produce the wrong result if 'num' is negative: this caused
8134 problems with BN_mod() and BN_nnmod().
8137 *) Function OCSP_request_verify(). This checks the signature on an
8138 OCSP request and verifies the signer certificate. The signer
8139 certificate is just checked for a generic purpose and OCSP request
8143 *) Add OCSP_check_validity() function to check the validity of OCSP
8144 responses. OCSP responses are prepared in real time and may only
8145 be a few seconds old. Simply checking that the current time lies
8146 between thisUpdate and nextUpdate max reject otherwise valid responses
8147 caused by either OCSP responder or client clock inaccuracy. Instead
8148 we allow thisUpdate and nextUpdate to fall within a certain period of
8149 the current time. The age of the response can also optionally be
8150 checked. Two new options -validity_period and -status_age added to
8154 *) If signature or public key algorithm is unrecognized print out its
8155 OID rather that just UNKNOWN.
8158 *) Change OCSP_cert_to_id() to tolerate a NULL subject certificate and
8159 OCSP_cert_id_new() a NULL serialNumber. This allows a partial certificate
8160 ID to be generated from the issuer certificate alone which can then be
8161 passed to OCSP_id_issuer_cmp().
8164 *) New compilation option ASN1_ITEM_FUNCTIONS. This causes the new
8165 ASN1 modules to export functions returning ASN1_ITEM pointers
8166 instead of the ASN1_ITEM structures themselves. This adds several
8167 new macros which allow the underlying ASN1 function/structure to
8168 be accessed transparently. As a result code should not use ASN1_ITEM
8169 references directly (such as &X509_it) but instead use the relevant
8170 macros (such as ASN1_ITEM_rptr(X509)). This option is to allow
8171 use of the new ASN1 code on platforms where exporting structures
8172 is problematical (for example in shared libraries) but exporting
8173 functions returning pointers to structures is not.
8176 *) Add support for overriding the generation of SSL/TLS session IDs.
8177 These callbacks can be registered either in an SSL_CTX or per SSL.
8178 The purpose of this is to allow applications to control, if they wish,
8179 the arbitrary values chosen for use as session IDs, particularly as it
8180 can be useful for session caching in multiple-server environments. A
8181 command-line switch for testing this (and any client code that wishes
8182 to use such a feature) has been added to "s_server".
8183 [Geoff Thorpe, Lutz Jaenicke]
8185 *) Modify mkdef.pl to recognise and parse preprocessor conditionals
8186 of the form '#if defined(...) || defined(...) || ...' and
8187 '#if !defined(...) && !defined(...) && ...'. This also avoids
8188 the growing number of special cases it was previously handling.
8191 *) Make all configuration macros available for application by making
8192 sure they are available in opensslconf.h, by giving them names starting
8193 with "OPENSSL_" to avoid conflicts with other packages and by making
8194 sure e_os2.h will cover all platform-specific cases together with
8196 Additionally, it is now possible to define configuration/platform-
8197 specific names (called "system identities"). In the C code, these
8198 are prefixed with "OPENSSL_SYSNAME_". e_os2.h will create another
8199 macro with the name beginning with "OPENSSL_SYS_", which is determined
8200 from "OPENSSL_SYSNAME_*" or compiler-specific macros depending on
8204 *) New option -set_serial to 'req' and 'x509' this allows the serial
8205 number to use to be specified on the command line. Previously self
8206 signed certificates were hard coded with serial number 0 and the
8207 CA options of 'x509' had to use a serial number in a file which was
8211 *) New options to 'ca' utility to support V2 CRL entry extensions.
8212 Currently CRL reason, invalidity date and hold instruction are
8213 supported. Add new CRL extensions to V3 code and some new objects.
8216 *) New function EVP_CIPHER_CTX_set_padding() this is used to
8217 disable standard block padding (aka PKCS#5 padding) in the EVP
8218 API, which was previously mandatory. This means that the data is
8219 not padded in any way and so the total length much be a multiple
8220 of the block size, otherwise an error occurs.
8223 *) Initial (incomplete) OCSP SSL support.
8226 *) New function OCSP_parse_url(). This splits up a URL into its host,
8227 port and path components: primarily to parse OCSP URLs. New -url
8228 option to ocsp utility.
8231 *) New nonce behavior. The return value of OCSP_check_nonce() now
8232 reflects the various checks performed. Applications can decide
8233 whether to tolerate certain situations such as an absent nonce
8234 in a response when one was present in a request: the ocsp application
8235 just prints out a warning. New function OCSP_add1_basic_nonce()
8236 this is to allow responders to include a nonce in a response even if
8237 the request is nonce-less.
8240 *) Disable stdin buffering in load_cert (apps/apps.c) so that no certs are
8241 skipped when using openssl x509 multiple times on a single input file,
8242 e.g. "(openssl x509 -out cert1; openssl x509 -out cert2) <certs".
8245 *) Make ASN1_UTCTIME_set_string() and ASN1_GENERALIZEDTIME_set_string()
8246 set string type: to handle setting ASN1_TIME structures. Fix ca
8247 utility to correctly initialize revocation date of CRLs.
8250 *) New option SSL_OP_CIPHER_SERVER_PREFERENCE allows the server to override
8251 the clients preferred ciphersuites and rather use its own preferences.
8252 Should help to work around M$ SGC (Server Gated Cryptography) bug in
8253 Internet Explorer by ensuring unchanged hash method during stepup.
8254 (Also replaces the broken/deactivated SSL_OP_NON_EXPORT_FIRST option.)
8257 *) Make mkdef.pl recognise all DECLARE_ASN1 macros, change rijndael
8258 to aes and add a new 'exist' option to print out symbols that don't
8262 *) Additional options to ocsp utility to allow flags to be set and
8263 additional certificates supplied.
8266 *) Add the option -VAfile to 'openssl ocsp', so the user can give the
8267 OCSP client a number of certificate to only verify the response
8271 *) Update Rijndael code to version 3.0 and change EVP AES ciphers to
8272 handle the new API. Currently only ECB, CBC modes supported. Add new
8275 Add TLS AES ciphersuites as described in RFC3268, "Advanced
8276 Encryption Standard (AES) Ciphersuites for Transport Layer
8277 Security (TLS)". (In beta versions of OpenSSL 0.9.7, these were
8278 not enabled by default and were not part of the "ALL" ciphersuite
8279 alias because they were not yet official; they could be
8280 explicitly requested by specifying the "AESdraft" ciphersuite
8281 group alias. In the final release of OpenSSL 0.9.7, the group
8282 alias is called "AES" and is part of "ALL".)
8283 [Ben Laurie, Steve Henson, Bodo Moeller]
8285 *) New function OCSP_copy_nonce() to copy nonce value (if present) from
8286 request to response.
8289 *) Functions for OCSP responders. OCSP_request_onereq_count(),
8290 OCSP_request_onereq_get0(), OCSP_onereq_get0_id() and OCSP_id_get0_info()
8291 extract information from a certificate request. OCSP_response_create()
8292 creates a response and optionally adds a basic response structure.
8293 OCSP_basic_add1_status() adds a complete single response to a basic
8294 response and returns the OCSP_SINGLERESP structure just added (to allow
8295 extensions to be included for example). OCSP_basic_add1_cert() adds a
8296 certificate to a basic response and OCSP_basic_sign() signs a basic
8297 response with various flags. New helper functions ASN1_TIME_check()
8298 (checks validity of ASN1_TIME structure) and ASN1_TIME_to_generalizedtime()
8299 (converts ASN1_TIME to GeneralizedTime).
8302 *) Various new functions. EVP_Digest() combines EVP_Digest{Init,Update,Final}()
8303 in a single operation. X509_get0_pubkey_bitstr() extracts the public_key
8304 structure from a certificate. X509_pubkey_digest() digests the public_key
8305 contents: this is used in various key identifiers.
8308 *) Make sk_sort() tolerate a NULL argument.
8309 [Steve Henson reported by Massimiliano Pala <madwolf@comune.modena.it>]
8311 *) New OCSP verify flag OCSP_TRUSTOTHER. When set the "other" certificates
8312 passed by the function are trusted implicitly. If any of them signed the
8313 response then it is assumed to be valid and is not verified.
8316 *) In PKCS7_set_type() initialise content_type in PKCS7_ENC_CONTENT
8317 to data. This was previously part of the PKCS7 ASN1 code. This
8318 was causing problems with OpenSSL created PKCS#12 and PKCS#7 structures.
8319 [Steve Henson, reported by Kenneth R. Robinette
8320 <support@securenetterm.com>]
8322 *) Add CRYPTO_push_info() and CRYPTO_pop_info() calls to new ASN1
8323 routines: without these tracing memory leaks is very painful.
8324 Fix leaks in PKCS12 and PKCS7 routines.
8327 *) Make X509_time_adj() cope with the new behaviour of ASN1_TIME_new().
8328 Previously it initialised the 'type' argument to V_ASN1_UTCTIME which
8329 effectively meant GeneralizedTime would never be used. Now it
8330 is initialised to -1 but X509_time_adj() now has to check the value
8331 and use ASN1_TIME_set() if the value is not V_ASN1_UTCTIME or
8332 V_ASN1_GENERALIZEDTIME, without this it always uses GeneralizedTime.
8333 [Steve Henson, reported by Kenneth R. Robinette
8334 <support@securenetterm.com>]
8336 *) Fixes to BN_to_ASN1_INTEGER when bn is zero. This would previously
8337 result in a zero length in the ASN1_INTEGER structure which was
8338 not consistent with the structure when d2i_ASN1_INTEGER() was used
8339 and would cause ASN1_INTEGER_cmp() to fail. Enhance s2i_ASN1_INTEGER()
8340 to cope with hex and negative integers. Fix bug in i2a_ASN1_INTEGER()
8341 where it did not print out a minus for negative ASN1_INTEGER.
8344 *) Add summary printout to ocsp utility. The various functions which
8345 convert status values to strings have been renamed to:
8346 OCSP_response_status_str(), OCSP_cert_status_str() and
8347 OCSP_crl_reason_str() and are no longer static. New options
8348 to verify nonce values and to disable verification. OCSP response
8349 printout format cleaned up.
8352 *) Add additional OCSP certificate checks. These are those specified
8353 in RFC2560. This consists of two separate checks: the CA of the
8354 certificate being checked must either be the OCSP signer certificate
8355 or the issuer of the OCSP signer certificate. In the latter case the
8356 OCSP signer certificate must contain the OCSP signing extended key
8357 usage. This check is performed by attempting to match the OCSP
8358 signer or the OCSP signer CA to the issuerNameHash and issuerKeyHash
8359 in the OCSP_CERTID structures of the response.
8362 *) Initial OCSP certificate verification added to OCSP_basic_verify()
8363 and related routines. This uses the standard OpenSSL certificate
8364 verify routines to perform initial checks (just CA validity) and
8365 to obtain the certificate chain. Then additional checks will be
8366 performed on the chain. Currently the root CA is checked to see
8367 if it is explicitly trusted for OCSP signing. This is used to set
8368 a root CA as a global signing root: that is any certificate that
8369 chains to that CA is an acceptable OCSP signing certificate.
8372 *) New '-extfile ...' option to 'openssl ca' for reading X.509v3
8373 extensions from a separate configuration file.
8374 As when reading extensions from the main configuration file,
8375 the '-extensions ...' option may be used for specifying the
8377 [Massimiliano Pala <madwolf@comune.modena.it>]
8379 *) New OCSP utility. Allows OCSP requests to be generated or
8380 read. The request can be sent to a responder and the output
8381 parsed, outputed or printed in text form. Not complete yet:
8382 still needs to check the OCSP response validity.
8385 *) New subcommands for 'openssl ca':
8386 'openssl ca -status <serial>' prints the status of the cert with
8387 the given serial number (according to the index file).
8388 'openssl ca -updatedb' updates the expiry status of certificates
8390 [Massimiliano Pala <madwolf@comune.modena.it>]
8392 *) New '-newreq-nodes' command option to CA.pl. This is like
8393 '-newreq', but calls 'openssl req' with the '-nodes' option
8394 so that the resulting key is not encrypted.
8395 [Damien Miller <djm@mindrot.org>]
8397 *) New configuration for the GNU Hurd.
8398 [Jonathan Bartlett <johnnyb@wolfram.com> via Richard Levitte]
8400 *) Initial code to implement OCSP basic response verify. This
8401 is currently incomplete. Currently just finds the signer's
8402 certificate and verifies the signature on the response.
8405 *) New SSLeay_version code SSLEAY_DIR to determine the compiled-in
8406 value of OPENSSLDIR. This is available via the new '-d' option
8407 to 'openssl version', and is also included in 'openssl version -a'.
8410 *) Allowing defining memory allocation callbacks that will be given
8411 file name and line number information in additional arguments
8412 (a const char* and an int). The basic functionality remains, as
8413 well as the original possibility to just replace malloc(),
8414 realloc() and free() by functions that do not know about these
8415 additional arguments. To register and find out the current
8416 settings for extended allocation functions, the following
8417 functions are provided:
8419 CRYPTO_set_mem_ex_functions
8420 CRYPTO_set_locked_mem_ex_functions
8421 CRYPTO_get_mem_ex_functions
8422 CRYPTO_get_locked_mem_ex_functions
8424 These work the same way as CRYPTO_set_mem_functions and friends.
8425 CRYPTO_get_[locked_]mem_functions now writes 0 where such an
8426 extended allocation function is enabled.
8427 Similarly, CRYPTO_get_[locked_]mem_ex_functions writes 0 where
8428 a conventional allocation function is enabled.
8429 [Richard Levitte, Bodo Moeller]
8431 *) Finish off removing the remaining LHASH function pointer casts.
8432 There should no longer be any prototype-casting required when using
8433 the LHASH abstraction, and any casts that remain are "bugs". See
8434 the callback types and macros at the head of lhash.h for details
8435 (and "OBJ_cleanup" in crypto/objects/obj_dat.c as an example).
8438 *) Add automatic query of EGD sockets in RAND_poll() for the unix variant.
8439 If /dev/[u]random devices are not available or do not return enough
8440 entropy, EGD style sockets (served by EGD or PRNGD) will automatically
8442 The locations /var/run/egd-pool, /dev/egd-pool, /etc/egd-pool, and
8443 /etc/entropy will be queried once each in this sequence, querying stops
8444 when enough entropy was collected without querying more sockets.
8447 *) Change the Unix RAND_poll() variant to be able to poll several
8448 random devices, as specified by DEVRANDOM, until a sufficient amount
8449 of data has been collected. We spend at most 10 ms on each file
8450 (select timeout) and read in non-blocking mode. DEVRANDOM now
8451 defaults to the list "/dev/urandom", "/dev/random", "/dev/srandom"
8452 (previously it was just the string "/dev/urandom"), so on typical
8453 platforms the 10 ms delay will never occur.
8454 Also separate out the Unix variant to its own file, rand_unix.c.
8455 For VMS, there's a currently-empty rand_vms.c.
8458 *) Move OCSP client related routines to ocsp_cl.c. These
8459 provide utility functions which an application needing
8460 to issue a request to an OCSP responder and analyse the
8461 response will typically need: as opposed to those which an
8462 OCSP responder itself would need which will be added later.
8464 OCSP_request_sign() signs an OCSP request with an API similar
8465 to PKCS7_sign(). OCSP_response_status() returns status of OCSP
8466 response. OCSP_response_get1_basic() extracts basic response
8467 from response. OCSP_resp_find_status(): finds and extracts status
8468 information from an OCSP_CERTID structure (which will be created
8469 when the request structure is built). These are built from lower
8470 level functions which work on OCSP_SINGLERESP structures but
8471 won't normally be used unless the application wishes to examine
8472 extensions in the OCSP response for example.
8474 Replace nonce routines with a pair of functions.
8475 OCSP_request_add1_nonce() adds a nonce value and optionally
8476 generates a random value. OCSP_check_nonce() checks the
8477 validity of the nonce in an OCSP response.
8480 *) Change function OCSP_request_add() to OCSP_request_add0_id().
8481 This doesn't copy the supplied OCSP_CERTID and avoids the
8482 need to free up the newly created id. Change return type
8483 to OCSP_ONEREQ to return the internal OCSP_ONEREQ structure.
8484 This can then be used to add extensions to the request.
8485 Deleted OCSP_request_new(), since most of its functionality
8486 is now in OCSP_REQUEST_new() (and the case insensitive name
8487 clash) apart from the ability to set the request name which
8488 will be added elsewhere.
8491 *) Update OCSP API. Remove obsolete extensions argument from
8492 various functions. Extensions are now handled using the new
8493 OCSP extension code. New simple OCSP HTTP function which
8494 can be used to send requests and parse the response.
8497 *) Fix the PKCS#7 (S/MIME) code to work with new ASN1. Two new
8498 ASN1_ITEM structures help with sign and verify. PKCS7_ATTR_SIGN
8499 uses the special reorder version of SET OF to sort the attributes
8500 and reorder them to match the encoded order. This resolves a long
8501 standing problem: a verify on a PKCS7 structure just after signing
8502 it used to fail because the attribute order did not match the
8503 encoded order. PKCS7_ATTR_VERIFY does not reorder the attributes:
8504 it uses the received order. This is necessary to tolerate some broken
8505 software that does not order SET OF. This is handled by encoding
8506 as a SEQUENCE OF but using implicit tagging (with UNIVERSAL class)
8507 to produce the required SET OF.
8510 *) Have mk1mf.pl generate the macros OPENSSL_BUILD_SHLIBCRYPTO and
8511 OPENSSL_BUILD_SHLIBSSL and use them appropriately in the header
8512 files to get correct declarations of the ASN.1 item variables.
8515 *) Rewrite of PKCS#12 code to use new ASN1 functionality. Replace many
8516 PKCS#12 macros with real functions. Fix two unrelated ASN1 bugs:
8517 asn1_check_tlen() would sometimes attempt to use 'ctx' when it was
8518 NULL and ASN1_TYPE was not dereferenced properly in asn1_ex_c2i().
8519 New ASN1 macro: DECLARE_ASN1_ITEM() which just declares the relevant
8520 ASN1_ITEM and no wrapper functions.
8523 *) New functions or ASN1_item_d2i_fp() and ASN1_item_d2i_bio(). These
8524 replace the old function pointer based I/O routines. Change most of
8525 the *_d2i_bio() and *_d2i_fp() functions to use these.
8528 *) Enhance mkdef.pl to be more accepting about spacing in C preprocessor
8529 lines, recognize more "algorithms" that can be deselected, and make
8530 it complain about algorithm deselection that isn't recognised.
8533 *) New ASN1 functions to handle dup, sign, verify, digest, pack and
8534 unpack operations in terms of ASN1_ITEM. Modify existing wrappers
8535 to use new functions. Add NO_ASN1_OLD which can be set to remove
8536 some old style ASN1 functions: this can be used to determine if old
8537 code will still work when these eventually go away.
8540 *) New extension functions for OCSP structures, these follow the
8541 same conventions as certificates and CRLs.
8544 *) New function X509V3_add1_i2d(). This automatically encodes and
8545 adds an extension. Its behaviour can be customised with various
8546 flags to append, replace or delete. Various wrappers added for
8547 certificates and CRLs.
8550 *) Fix to avoid calling the underlying ASN1 print routine when
8551 an extension cannot be parsed. Correct a typo in the
8552 OCSP_SERVICELOC extension. Tidy up print OCSP format.
8555 *) Make mkdef.pl parse some of the ASN1 macros and add appropriate
8556 entries for variables.
8559 *) Add functionality to apps/openssl.c for detecting locking
8560 problems: As the program is single-threaded, all we have
8561 to do is register a locking callback using an array for
8562 storing which locks are currently held by the program.
8565 *) Use a lock around the call to CRYPTO_get_ex_new_index() in
8566 SSL_get_ex_data_X509_STORE_idx(), which is used in
8567 ssl_verify_cert_chain() and thus can be called at any time
8568 during TLS/SSL handshakes so that thread-safety is essential.
8569 Unfortunately, the ex_data design is not at all suited
8570 for multi-threaded use, so it probably should be abolished.
8573 *) Added Broadcom "ubsec" ENGINE to OpenSSL.
8574 [Broadcom, tweaked and integrated by Geoff Thorpe]
8576 *) Move common extension printing code to new function
8577 X509V3_print_extensions(). Reorganise OCSP print routines and
8578 implement some needed OCSP ASN1 functions. Add OCSP extensions.
8581 *) New function X509_signature_print() to remove duplication in some
8585 *) Add a special meaning when SET OF and SEQUENCE OF flags are both
8586 set (this was treated exactly the same as SET OF previously). This
8587 is used to reorder the STACK representing the structure to match the
8588 encoding. This will be used to get round a problem where a PKCS7
8589 structure which was signed could not be verified because the STACK
8590 order did not reflect the encoded order.
8593 *) Reimplement the OCSP ASN1 module using the new code.
8596 *) Update the X509V3 code to permit the use of an ASN1_ITEM structure
8597 for its ASN1 operations. The old style function pointers still exist
8598 for now but they will eventually go away.
8601 *) Merge in replacement ASN1 code from the ASN1 branch. This almost
8602 completely replaces the old ASN1 functionality with a table driven
8603 encoder and decoder which interprets an ASN1_ITEM structure describing
8604 the ASN1 module. Compatibility with the existing ASN1 API (i2d,d2i) is
8605 largely maintained. Almost all of the old asn1_mac.h macro based ASN1
8606 has also been converted to the new form.
8609 *) Change BN_mod_exp_recp so that negative moduli are tolerated
8610 (the sign is ignored). Similarly, ignore the sign in BN_MONT_CTX_set
8611 so that BN_mod_exp_mont and BN_mod_exp_mont_word work
8612 for negative moduli.
8615 *) Fix BN_uadd and BN_usub: Always return non-negative results instead
8616 of not touching the result's sign bit.
8619 *) BN_div bugfix: If the result is 0, the sign (res->neg) must not be
8623 *) Changed the LHASH code to use prototypes for callbacks, and created
8624 macros to declare and implement thin (optionally static) functions
8625 that provide type-safety and avoid function pointer casting for the
8626 type-specific callbacks.
8629 *) Added Kerberos Cipher Suites to be used with TLS, as written in
8631 [Veers Staats <staatsvr@asc.hpc.mil>,
8632 Jeffrey Altman <jaltman@columbia.edu>, via Richard Levitte]
8634 *) Reformat the FAQ so the different questions and answers can be divided
8635 in sections depending on the subject.
8638 *) Have the zlib compression code load ZLIB.DLL dynamically under
8642 *) New function BN_mod_sqrt for computing square roots modulo a prime
8643 (using the probabilistic Tonelli-Shanks algorithm unless
8644 p == 3 (mod 4) or p == 5 (mod 8), which are cases that can
8645 be handled deterministically).
8646 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8648 *) Make BN_mod_inverse faster by explicitly handling small quotients
8649 in the Euclid loop. (Speed gain about 20% for small moduli [256 or
8650 512 bits], about 30% for larger ones [1024 or 2048 bits].)
8653 *) New function BN_kronecker.
8656 *) Fix BN_gcd so that it works on negative inputs; the result is
8657 positive unless both parameters are zero.
8658 Previously something reasonably close to an infinite loop was
8659 possible because numbers could be growing instead of shrinking
8660 in the implementation of Euclid's algorithm.
8663 *) Fix BN_is_word() and BN_is_one() macros to take into account the
8664 sign of the number in question.
8666 Fix BN_is_word(a,w) to work correctly for w == 0.
8668 The old BN_is_word(a,w) macro is now called BN_abs_is_word(a,w)
8669 because its test if the absolute value of 'a' equals 'w'.
8670 Note that BN_abs_is_word does *not* handle w == 0 reliably;
8671 it exists mostly for use in the implementations of BN_is_zero(),
8672 BN_is_one(), and BN_is_word().
8675 *) New function BN_swap.
8678 *) Use BN_nnmod instead of BN_mod in crypto/bn/bn_exp.c so that
8679 the exponentiation functions are more likely to produce reasonable
8680 results on negative inputs.
8683 *) Change BN_mod_mul so that the result is always non-negative.
8684 Previously, it could be negative if one of the factors was negative;
8685 I don't think anyone really wanted that behaviour.
8688 *) Move BN_mod_... functions into new file crypto/bn/bn_mod.c
8689 (except for exponentiation, which stays in crypto/bn/bn_exp.c,
8690 and BN_mod_mul_reciprocal, which stays in crypto/bn/bn_recp.c)
8691 and add new functions:
8700 BN_mod_lshift1_quick
8704 These functions always generate non-negative results.
8706 BN_nnmod otherwise is like BN_mod (if BN_mod computes a remainder r
8707 such that |m| < r < 0, BN_nnmod will output rem + |m| instead).
8709 BN_mod_XXX_quick(r, a, [b,] m) generates the same result as
8710 BN_mod_XXX(r, a, [b,] m, ctx), but requires that a [and b]
8711 be reduced modulo m.
8712 [Lenka Fibikova <fibikova@exp-math.uni-essen.de>, Bodo Moeller]
8715 The following entry accidentally appeared in the CHANGES file
8716 distributed with OpenSSL 0.9.7. The modifications described in
8717 it do *not* apply to OpenSSL 0.9.7.
8719 *) Remove a few calls to bn_wexpand() in BN_sqr() (the one in there
8720 was actually never needed) and in BN_mul(). The removal in BN_mul()
8721 required a small change in bn_mul_part_recursive() and the addition
8722 of the functions bn_cmp_part_words(), bn_sub_part_words() and
8723 bn_add_part_words(), which do the same thing as bn_cmp_words(),
8724 bn_sub_words() and bn_add_words() except they take arrays with
8729 *) In 'openssl passwd', verify passwords read from the terminal
8730 unless the '-salt' option is used (which usually means that
8731 verification would just waste user's time since the resulting
8732 hash is going to be compared with some given password hash)
8733 or the new '-noverify' option is used.
8735 This is an incompatible change, but it does not affect
8736 non-interactive use of 'openssl passwd' (passwords on the command
8737 line, '-stdin' option, '-in ...' option) and thus should not
8741 *) Remove all references to RSAref, since there's no more need for it.
8744 *) Make DSO load along a path given through an environment variable
8745 (SHLIB_PATH) with shl_load().
8748 *) Constify the ENGINE code as a result of BIGNUM constification.
8749 Also constify the RSA code and most things related to it. In a
8750 few places, most notable in the depth of the ASN.1 code, ugly
8751 casts back to non-const were required (to be solved at a later
8755 *) Make it so the openssl application has all engines loaded by default.
8758 *) Constify the BIGNUM routines a little more.
8761 *) Add the following functions:
8763 ENGINE_load_cswift()
8765 ENGINE_load_atalla()
8767 ENGINE_load_builtin_engines()
8769 That way, an application can itself choose if external engines that
8770 are built-in in OpenSSL shall ever be used or not. The benefit is
8771 that applications won't have to be linked with libdl or other dso
8772 libraries unless it's really needed.
8774 Changed 'openssl engine' to load all engines on demand.
8775 Changed the engine header files to avoid the duplication of some
8776 declarations (they differed!).
8779 *) 'openssl engine' can now list capabilities.
8782 *) Better error reporting in 'openssl engine'.
8785 *) Never call load_dh_param(NULL) in s_server.
8788 *) Add engine application. It can currently list engines by name and
8789 identity, and test if they are actually available.
8792 *) Improve RPM specification file by forcing symbolic linking and making
8793 sure the installed documentation is also owned by root.root.
8794 [Damien Miller <djm@mindrot.org>]
8796 *) Give the OpenSSL applications more possibilities to make use of
8797 keys (public as well as private) handled by engines.
8800 *) Add OCSP code that comes from CertCo.
8803 *) Add VMS support for the Rijndael code.
8806 *) Added untested support for Nuron crypto accelerator.
8809 *) Add support for external cryptographic devices. This code was
8810 previously distributed separately as the "engine" branch.
8811 [Geoff Thorpe, Richard Levitte]
8813 *) Rework the filename-translation in the DSO code. It is now possible to
8814 have far greater control over how a "name" is turned into a filename
8815 depending on the operating environment and any oddities about the
8816 different shared library filenames on each system.
8819 *) Support threads on FreeBSD-elf in Configure.
8822 *) Fix for SHA1 assembly problem with MASM: it produces
8823 warnings about corrupt line number information when assembling
8824 with debugging information. This is caused by the overlapping
8826 [Bernd Matthes <mainbug@celocom.de>, Steve Henson]
8829 NCONF_get_number() has no error checking at all. As a replacement,
8830 NCONF_get_number_e() is defined (_e for "error checking") and is
8831 promoted strongly. The old NCONF_get_number is kept around for
8832 binary backward compatibility.
8833 Make it possible for methods to load from something other than a BIO,
8834 by providing a function pointer that is given a name instead of a BIO.
8835 For example, this could be used to load configuration data from an
8839 *) Fix for non blocking accept BIOs. Added new I/O special reason
8840 BIO_RR_ACCEPT to cover this case. Previously use of accept BIOs
8841 with non blocking I/O was not possible because no retry code was
8842 implemented. Also added new SSL code SSL_WANT_ACCEPT to cover
8846 *) Added the beginnings of Rijndael support.
8849 *) Fix for bug in DirectoryString mask setting. Add support for
8850 X509_NAME_print_ex() in 'req' and X509_print_ex() function
8851 to allow certificate printing to more controllable, additional
8852 'certopt' option to 'x509' to allow new printing options to be
8856 *) Clean old EAY MD5 hack from e_os.h.
8859 Changes between 0.9.6l and 0.9.6m [17 Mar 2004]
8861 *) Fix null-pointer assignment in do_change_cipher_spec() revealed
8862 by using the Codenomicon TLS Test Tool (CVE-2004-0079)
8863 [Joe Orton, Steve Henson]
8865 Changes between 0.9.6k and 0.9.6l [04 Nov 2003]
8867 *) Fix additional bug revealed by the NISCC test suite:
8869 Stop bug triggering large recursion when presented with
8870 certain ASN.1 tags (CVE-2003-0851)
8873 Changes between 0.9.6j and 0.9.6k [30 Sep 2003]
8875 *) Fix various bugs revealed by running the NISCC test suite:
8877 Stop out of bounds reads in the ASN1 code when presented with
8878 invalid tags (CVE-2003-0543 and CVE-2003-0544).
8880 If verify callback ignores invalid public key errors don't try to check
8881 certificate signature with the NULL public key.
8885 *) In ssl3_accept() (ssl/s3_srvr.c) only accept a client certificate
8886 if the server requested one: as stated in TLS 1.0 and SSL 3.0
8890 *) In ssl3_get_client_hello() (ssl/s3_srvr.c), tolerate additional
8891 extra data after the compression methods not only for TLS 1.0
8892 but also for SSL 3.0 (as required by the specification).
8893 [Bodo Moeller; problem pointed out by Matthias Loepfe]
8895 *) Change X509_certificate_type() to mark the key as exported/exportable
8896 when it's 512 *bits* long, not 512 bytes.
8899 Changes between 0.9.6i and 0.9.6j [10 Apr 2003]
8901 *) Countermeasure against the Klima-Pokorny-Rosa extension of
8902 Bleichbacher's attack on PKCS #1 v1.5 padding: treat
8903 a protocol version number mismatch like a decryption error
8904 in ssl3_get_client_key_exchange (ssl/s3_srvr.c).
8907 *) Turn on RSA blinding by default in the default implementation
8908 to avoid a timing attack. Applications that don't want it can call
8909 RSA_blinding_off() or use the new flag RSA_FLAG_NO_BLINDING.
8910 They would be ill-advised to do so in most cases.
8911 [Ben Laurie, Steve Henson, Geoff Thorpe, Bodo Moeller]
8913 *) Change RSA blinding code so that it works when the PRNG is not
8914 seeded (in this case, the secret RSA exponent is abused as
8915 an unpredictable seed -- if it is not unpredictable, there
8916 is no point in blinding anyway). Make RSA blinding thread-safe
8917 by remembering the creator's thread ID in rsa->blinding and
8918 having all other threads use local one-time blinding factors
8919 (this requires more computation than sharing rsa->blinding, but
8920 avoids excessive locking; and if an RSA object is not shared
8921 between threads, blinding will still be very fast).
8924 Changes between 0.9.6h and 0.9.6i [19 Feb 2003]
8926 *) In ssl3_get_record (ssl/s3_pkt.c), minimize information leaked
8927 via timing by performing a MAC computation even if incorrect
8928 block cipher padding has been found. This is a countermeasure
8929 against active attacks where the attacker has to distinguish
8930 between bad padding and a MAC verification error. (CVE-2003-0078)
8932 [Bodo Moeller; problem pointed out by Brice Canvel (EPFL),
8933 Alain Hiltgen (UBS), Serge Vaudenay (EPFL), and
8934 Martin Vuagnoux (EPFL, Ilion)]
8936 Changes between 0.9.6g and 0.9.6h [5 Dec 2002]
8938 *) New function OPENSSL_cleanse(), which is used to cleanse a section of
8939 memory from it's contents. This is done with a counter that will
8940 place alternating values in each byte. This can be used to solve
8941 two issues: 1) the removal of calls to memset() by highly optimizing
8942 compilers, and 2) cleansing with other values than 0, since those can
8943 be read through on certain media, for example a swap space on disk.
8946 *) Bugfix: client side session caching did not work with external caching,
8947 because the session->cipher setting was not restored when reloading
8948 from the external cache. This problem was masked, when
8949 SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG (part of SSL_OP_ALL) was set.
8950 (Found by Steve Haslam <steve@araqnid.ddts.net>.)
8953 *) Fix client_certificate (ssl/s2_clnt.c): The permissible total
8954 length of the REQUEST-CERTIFICATE message is 18 .. 34, not 17 .. 33.
8955 [Zeev Lieber <zeev-l@yahoo.com>]
8957 *) Undo an undocumented change introduced in 0.9.6e which caused
8958 repeated calls to OpenSSL_add_all_ciphers() and
8959 OpenSSL_add_all_digests() to be ignored, even after calling
8963 *) Change the default configuration reader to deal with last line not
8964 being properly terminated.
8967 *) Change X509_NAME_cmp() so it applies the special rules on handling
8968 DN values that are of type PrintableString, as well as RDNs of type
8969 emailAddress where the value has the type ia5String.
8970 [stefank@valicert.com via Richard Levitte]
8972 *) Add a SSL_SESS_CACHE_NO_INTERNAL_STORE flag to take over half
8973 the job SSL_SESS_CACHE_NO_INTERNAL_LOOKUP was inconsistently
8974 doing, define a new flag (SSL_SESS_CACHE_NO_INTERNAL) to be
8975 the bitwise-OR of the two for use by the majority of applications
8976 wanting this behaviour, and update the docs. The documented
8977 behaviour and actual behaviour were inconsistent and had been
8978 changing anyway, so this is more a bug-fix than a behavioural
8980 [Geoff Thorpe, diagnosed by Nadav Har'El]
8982 *) Don't impose a 16-byte length minimum on session IDs in ssl/s3_clnt.c
8983 (the SSL 3.0 and TLS 1.0 specifications allow any length up to 32 bytes).
8986 *) Fix initialization code race conditions in
8987 SSLv23_method(), SSLv23_client_method(), SSLv23_server_method(),
8988 SSLv2_method(), SSLv2_client_method(), SSLv2_server_method(),
8989 SSLv3_method(), SSLv3_client_method(), SSLv3_server_method(),
8990 TLSv1_method(), TLSv1_client_method(), TLSv1_server_method(),
8991 ssl2_get_cipher_by_char(),
8992 ssl3_get_cipher_by_char().
8993 [Patrick McCormick <patrick@tellme.com>, Bodo Moeller]
8995 *) Reorder cleanup sequence in SSL_CTX_free(): only remove the ex_data after
8996 the cached sessions are flushed, as the remove_cb() might use ex_data
8997 contents. Bug found by Sam Varshavchik <mrsam@courier-mta.com>
8998 (see [openssl.org #212]).
8999 [Geoff Thorpe, Lutz Jaenicke]
9001 *) Fix typo in OBJ_txt2obj which incorrectly passed the content
9002 length, instead of the encoding length to d2i_ASN1_OBJECT.
9005 Changes between 0.9.6f and 0.9.6g [9 Aug 2002]
9007 *) [In 0.9.6g-engine release:]
9008 Fix crypto/engine/vendor_defns/cswift.h for WIN32 (use '_stdcall').
9009 [Lynn Gazis <lgazis@rainbow.com>]
9011 Changes between 0.9.6e and 0.9.6f [8 Aug 2002]
9013 *) Fix ASN1 checks. Check for overflow by comparing with LONG_MAX
9014 and get fix the header length calculation.
9015 [Florian Weimer <Weimer@CERT.Uni-Stuttgart.DE>,
9016 Alon Kantor <alonk@checkpoint.com> (and others),
9019 *) Use proper error handling instead of 'assertions' in buffer
9020 overflow checks added in 0.9.6e. This prevents DoS (the
9021 assertions could call abort()).
9022 [Arne Ansper <arne@ats.cyber.ee>, Bodo Moeller]
9024 Changes between 0.9.6d and 0.9.6e [30 Jul 2002]
9026 *) Add various sanity checks to asn1_get_length() to reject
9027 the ASN1 length bytes if they exceed sizeof(long), will appear
9028 negative or the content length exceeds the length of the
9030 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9032 *) Fix cipher selection routines: ciphers without encryption had no flags
9033 for the cipher strength set and where therefore not handled correctly
9034 by the selection routines (PR #130).
9037 *) Fix EVP_dsa_sha macro.
9041 SSL_OP_DONT_INSERT_EMPTY_FRAGMENTS
9042 for disabling the SSL 3.0/TLS 1.0 CBC vulnerability countermeasure
9043 that was added in OpenSSL 0.9.6d.
9045 As the countermeasure turned out to be incompatible with some
9046 broken SSL implementations, the new option is part of SSL_OP_ALL.
9047 SSL_OP_ALL is usually employed when compatibility with weird SSL
9048 implementations is desired (e.g. '-bugs' option to 's_client' and
9049 's_server'), so the new option is automatically set in many
9053 *) Changes in security patch:
9055 Changes marked "(CHATS)" were sponsored by the Defense Advanced
9056 Research Projects Agency (DARPA) and Air Force Research Laboratory,
9057 Air Force Materiel Command, USAF, under agreement number
9060 *) Add various sanity checks to asn1_get_length() to reject
9061 the ASN1 length bytes if they exceed sizeof(long), will appear
9062 negative or the content length exceeds the length of the
9063 supplied buffer. (CVE-2002-0659)
9064 [Steve Henson, Adi Stav <stav@mercury.co.il>, James Yonan <jim@ntlp.com>]
9066 *) Assertions for various potential buffer overflows, not known to
9068 [Ben Laurie (CHATS)]
9070 *) Various temporary buffers to hold ASCII versions of integers were
9071 too small for 64 bit platforms. (CVE-2002-0655)
9072 [Matthew Byng-Maddick <mbm@aldigital.co.uk> and Ben Laurie (CHATS)>
9074 *) Remote buffer overflow in SSL3 protocol - an attacker could
9075 supply an oversized session ID to a client. (CVE-2002-0656)
9076 [Ben Laurie (CHATS)]
9078 *) Remote buffer overflow in SSL2 protocol - an attacker could
9079 supply an oversized client master key. (CVE-2002-0656)
9080 [Ben Laurie (CHATS)]
9082 Changes between 0.9.6c and 0.9.6d [9 May 2002]
9084 *) Fix crypto/asn1/a_sign.c so that 'parameters' is omitted (not
9085 encoded as NULL) with id-dsa-with-sha1.
9086 [Nils Larsch <nla@trustcenter.de>; problem pointed out by Bodo Moeller]
9088 *) Check various X509_...() return values in apps/req.c.
9089 [Nils Larsch <nla@trustcenter.de>]
9091 *) Fix BASE64 decode (EVP_DecodeUpdate) for data with CR/LF ended lines:
9092 an end-of-file condition would erroneously be flagged, when the CRLF
9093 was just at the end of a processed block. The bug was discovered when
9094 processing data through a buffering memory BIO handing the data to a
9095 BASE64-decoding BIO. Bug fund and patch submitted by Pavel Tsekov
9096 <ptsekov@syntrex.com> and Nedelcho Stanev.
9099 *) Implement a countermeasure against a vulnerability recently found
9100 in CBC ciphersuites in SSL 3.0/TLS 1.0: Send an empty fragment
9101 before application data chunks to avoid the use of known IVs
9102 with data potentially chosen by the attacker.
9105 *) Fix length checks in ssl3_get_client_hello().
9108 *) TLS/SSL library bugfix: use s->s3->in_read_app_data differently
9109 to prevent ssl3_read_internal() from incorrectly assuming that
9110 ssl3_read_bytes() found application data while handshake
9111 processing was enabled when in fact s->s3->in_read_app_data was
9112 merely automatically cleared during the initial handshake.
9113 [Bodo Moeller; problem pointed out by Arne Ansper <arne@ats.cyber.ee>]
9115 *) Fix object definitions for Private and Enterprise: they were not
9116 recognized in their shortname (=lowercase) representation. Extend
9117 obj_dat.pl to issue an error when using undefined keywords instead
9118 of silently ignoring the problem (Svenning Sorensen
9119 <sss@sss.dnsalias.net>).
9122 *) Fix DH_generate_parameters() so that it works for 'non-standard'
9123 generators, i.e. generators other than 2 and 5. (Previously, the
9124 code did not properly initialise the 'add' and 'rem' values to
9125 BN_generate_prime().)
9127 In the new general case, we do not insist that 'generator' is
9128 actually a primitive root: This requirement is rather pointless;
9129 a generator of the order-q subgroup is just as good, if not
9133 *) Map new X509 verification errors to alerts. Discovered and submitted by
9134 Tom Wu <tom@arcot.com>.
9137 *) Fix ssl3_pending() (ssl/s3_lib.c) to prevent SSL_pending() from
9138 returning non-zero before the data has been completely received
9139 when using non-blocking I/O.
9140 [Bodo Moeller; problem pointed out by John Hughes]
9142 *) Some of the ciphers missed the strength entry (SSL_LOW etc).
9143 [Ben Laurie, Lutz Jaenicke]
9145 *) Fix bug in SSL_clear(): bad sessions were not removed (found by
9146 Yoram Zahavi <YoramZ@gilian.com>).
9149 *) Add information about CygWin 1.3 and on, and preserve proper
9150 configuration for the versions before that.
9151 [Corinna Vinschen <vinschen@redhat.com> and Richard Levitte]
9153 *) Make removal from session cache (SSL_CTX_remove_session()) more robust:
9154 check whether we deal with a copy of a session and do not delete from
9155 the cache in this case. Problem reported by "Izhar Shoshani Levi"
9156 <izhar@checkpoint.com>.
9159 *) Do not store session data into the internal session cache, if it
9160 is never intended to be looked up (SSL_SESS_CACHE_NO_INTERNAL_LOOKUP
9161 flag is set). Proposed by Aslam <aslam@funk.com>.
9164 *) Have ASN1_BIT_STRING_set_bit() really clear a bit when the requested
9168 *) [In 0.9.6d-engine release:]
9169 Fix a crashbug and a logic bug in hwcrhk_load_pubkey().
9170 [Toomas Kiisk <vix@cyber.ee> via Richard Levitte]
9172 *) Add the configuration target linux-s390x.
9173 [Neale Ferguson <Neale.Ferguson@SoftwareAG-USA.com> via Richard Levitte]
9175 *) The earlier bugfix for the SSL3_ST_SW_HELLO_REQ_C case of
9176 ssl3_accept (ssl/s3_srvr.c) incorrectly used a local flag
9177 variable as an indication that a ClientHello message has been
9178 received. As the flag value will be lost between multiple
9179 invocations of ssl3_accept when using non-blocking I/O, the
9180 function may not be aware that a handshake has actually taken
9181 place, thus preventing a new session from being added to the
9184 To avoid this problem, we now set s->new_session to 2 instead of
9185 using a local variable.
9186 [Lutz Jaenicke, Bodo Moeller]
9188 *) Bugfix: Return -1 from ssl3_get_server_done (ssl3/s3_clnt.c)
9189 if the SSL_R_LENGTH_MISMATCH error is detected.
9190 [Geoff Thorpe, Bodo Moeller]
9192 *) New 'shared_ldflag' column in Configure platform table.
9195 *) Fix EVP_CIPHER_mode macro.
9196 ["Dan S. Camper" <dan@bti.net>]
9198 *) Fix ssl3_read_bytes (ssl/s3_pkt.c): To ignore messages of unknown
9199 type, we must throw them away by setting rr->length to 0.
9200 [D P Chang <dpc@qualys.com>]
9202 Changes between 0.9.6b and 0.9.6c [21 dec 2001]
9204 *) Fix BN_rand_range bug pointed out by Dominikus Scherkl
9205 <Dominikus.Scherkl@biodata.com>. (The previous implementation
9206 worked incorrectly for those cases where range = 10..._2 and
9207 3*range is two bits longer than range.)
9210 *) Only add signing time to PKCS7 structures if it is not already
9214 *) Fix crypto/objects/objects.h: "ld-ce" should be "id-ce",
9215 OBJ_ld_ce should be OBJ_id_ce.
9216 Also some ip-pda OIDs in crypto/objects/objects.txt were
9217 incorrect (cf. RFC 3039).
9218 [Matt Cooper, Frederic Giudicelli, Bodo Moeller]
9220 *) Release CRYPTO_LOCK_DYNLOCK when CRYPTO_destroy_dynlockid()
9221 returns early because it has nothing to do.
9222 [Andy Schneider <andy.schneider@bjss.co.uk>]
9224 *) [In 0.9.6c-engine release:]
9225 Fix mutex callback return values in crypto/engine/hw_ncipher.c.
9226 [Andy Schneider <andy.schneider@bjss.co.uk>]
9228 *) [In 0.9.6c-engine release:]
9229 Add support for Cryptographic Appliance's keyserver technology.
9230 (Use engine 'keyclient')
9231 [Cryptographic Appliances and Geoff Thorpe]
9233 *) Add a configuration entry for OS/390 Unix. The C compiler 'c89'
9234 is called via tools/c89.sh because arguments have to be
9235 rearranged (all '-L' options must appear before the first object
9237 [Richard Shapiro <rshapiro@abinitio.com>]
9239 *) [In 0.9.6c-engine release:]
9240 Add support for Broadcom crypto accelerator cards, backported
9242 [Broadcom, Nalin Dahyabhai <nalin@redhat.com>, Mark Cox]
9244 *) [In 0.9.6c-engine release:]
9245 Add support for SureWare crypto accelerator cards from
9246 Baltimore Technologies. (Use engine 'sureware')
9247 [Baltimore Technologies and Mark Cox]
9249 *) [In 0.9.6c-engine release:]
9250 Add support for crypto accelerator cards from Accelerated
9251 Encryption Processing, www.aep.ie. (Use engine 'aep')
9252 [AEP Inc. and Mark Cox]
9254 *) Add a configuration entry for gcc on UnixWare.
9255 [Gary Benson <gbenson@redhat.com>]
9257 *) Change ssl/s2_clnt.c and ssl/s2_srvr.c so that received handshake
9258 messages are stored in a single piece (fixed-length part and
9259 variable-length part combined) and fix various bugs found on the way.
9262 *) Disable caching in BIO_gethostbyname(), directly use gethostbyname()
9263 instead. BIO_gethostbyname() does not know what timeouts are
9264 appropriate, so entries would stay in cache even when they have
9266 [Bodo Moeller; problem pointed out by Rich Salz <rsalz@zolera.com>
9268 *) Change ssl23_get_client_hello (ssl/s23_srvr.c) behaviour when
9269 faced with a pathologically small ClientHello fragment that does
9270 not contain client_version: Instead of aborting with an error,
9271 simply choose the highest available protocol version (i.e.,
9272 TLS 1.0 unless it is disabled). In practice, ClientHello
9273 messages are never sent like this, but this change gives us
9274 strictly correct behaviour at least for TLS.
9277 *) Fix SSL handshake functions and SSL_clear() such that SSL_clear()
9278 never resets s->method to s->ctx->method when called from within
9279 one of the SSL handshake functions.
9280 [Bodo Moeller; problem pointed out by Niko Baric]
9282 *) In ssl3_get_client_hello (ssl/s3_srvr.c), generate a fatal alert
9283 (sent using the client's version number) if client_version is
9284 smaller than the protocol version in use. Also change
9285 ssl23_get_client_hello (ssl/s23_srvr.c) to select TLS 1.0 if
9286 the client demanded SSL 3.0 but only TLS 1.0 is enabled; then
9287 the client will at least see that alert.
9290 *) Fix ssl3_get_message (ssl/s3_both.c) to handle message fragmentation
9294 *) Avoid infinite loop in ssl3_get_message (ssl/s3_both.c) if a
9295 client receives HelloRequest while in a handshake.
9296 [Bodo Moeller; bug noticed by Andy Schneider <andy.schneider@bjss.co.uk>]
9298 *) Bugfix in ssl3_accept (ssl/s3_srvr.c): Case SSL3_ST_SW_HELLO_REQ_C
9299 should end in 'break', not 'goto end' which circumvents various
9300 cleanups done in state SSL_ST_OK. But session related stuff
9301 must be disabled for SSL_ST_OK in the case that we just sent a
9304 Also avoid some overhead by not calling ssl_init_wbio_buffer()
9305 before just sending a HelloRequest.
9306 [Bodo Moeller, Eric Rescorla <ekr@rtfm.com>]
9308 *) Fix ssl/s3_enc.c, ssl/t1_enc.c and ssl/s3_pkt.c so that we don't
9309 reveal whether illegal block cipher padding was found or a MAC
9310 verification error occurred. (Neither SSLerr() codes nor alerts
9311 are directly visible to potential attackers, but the information
9312 may leak via logfiles.)
9314 Similar changes are not required for the SSL 2.0 implementation
9315 because the number of padding bytes is sent in clear for SSL 2.0,
9316 and the extra bytes are just ignored. However ssl/s2_pkt.c
9317 failed to verify that the purported number of padding bytes is in
9321 *) Add OpenUNIX-8 support including shared libraries
9322 (Boyd Lynn Gerber <gerberb@zenez.com>).
9325 *) Improve RSA_padding_check_PKCS1_OAEP() check again to avoid
9326 'wristwatch attack' using huge encoding parameters (cf.
9327 James H. Manger's CRYPTO 2001 paper). Note that the
9328 RSA_PKCS1_OAEP_PADDING case of RSA_private_decrypt() does not use
9329 encoding parameters and hence was not vulnerable.
9332 *) BN_sqr() bug fix.
9333 [Ulf Möller, reported by Jim Ellis <jim.ellis@cavium.com>]
9335 *) Rabin-Miller test analyses assume uniformly distributed witnesses,
9336 so use BN_pseudo_rand_range() instead of using BN_pseudo_rand()
9337 followed by modular reduction.
9338 [Bodo Moeller; pointed out by Adam Young <AYoung1@NCSUS.JNJ.COM>]
9340 *) Add BN_pseudo_rand_range() with obvious functionality: BN_rand_range()
9341 equivalent based on BN_pseudo_rand() instead of BN_rand().
9344 *) s3_srvr.c: allow sending of large client certificate lists (> 16 kB).
9345 This function was broken, as the check for a new client hello message
9346 to handle SGC did not allow these large messages.
9347 (Tracked down by "Douglas E. Engert" <deengert@anl.gov>.)
9350 *) Add alert descriptions for TLSv1 to SSL_alert_desc_string[_long]().
9353 *) Fix buggy behaviour of BIO_get_num_renegotiates() and BIO_ctrl()
9354 for BIO_C_GET_WRITE_BUF_SIZE ("Stephen Hinton" <shinton@netopia.com>).
9357 *) Rework the configuration and shared library support for Tru64 Unix.
9358 The configuration part makes use of modern compiler features and
9359 still retains old compiler behavior for those that run older versions
9360 of the OS. The shared library support part includes a variant that
9361 uses the RPATH feature, and is available through the special
9362 configuration target "alpha-cc-rpath", which will never be selected
9364 [Tim Mooney <mooney@dogbert.cc.ndsu.NoDak.edu> via Richard Levitte]
9366 *) In ssl3_get_key_exchange (ssl/s3_clnt.c), call ssl3_get_message()
9367 with the same message size as in ssl3_get_certificate_request().
9368 Otherwise, if no ServerKeyExchange message occurs, CertificateRequest
9369 messages might inadvertently be reject as too long.
9370 [Petr Lampa <lampa@fee.vutbr.cz>]
9372 *) Enhanced support for IA-64 Unix platforms (well, Linux and HP-UX).
9375 *) Modified SSL library such that the verify_callback that has been set
9376 specificly for an SSL object with SSL_set_verify() is actually being
9377 used. Before the change, a verify_callback set with this function was
9378 ignored and the verify_callback() set in the SSL_CTX at the time of
9379 the call was used. New function X509_STORE_CTX_set_verify_cb() introduced
9380 to allow the necessary settings.
9383 *) Initialize static variable in crypto/dsa/dsa_lib.c and crypto/dh/dh_lib.c
9384 explicitly to NULL, as at least on Solaris 8 this seems not always to be
9385 done automatically (in contradiction to the requirements of the C
9386 standard). This made problems when used from OpenSSH.
9389 *) In OpenSSL 0.9.6a and 0.9.6b, crypto/dh/dh_key.c ignored
9390 dh->length and always used
9392 BN_rand_range(priv_key, dh->p).
9394 BN_rand_range() is not necessary for Diffie-Hellman, and this
9395 specific range makes Diffie-Hellman unnecessarily inefficient if
9396 dh->length (recommended exponent length) is much smaller than the
9397 length of dh->p. We could use BN_rand_range() if the order of
9398 the subgroup was stored in the DH structure, but we only have
9403 BN_rand(priv_key, l, ...)
9405 where 'l' is dh->length if this is defined, or BN_num_bits(dh->p)-1
9411 RSA_eay_public_encrypt
9412 RSA_eay_private_decrypt
9413 RSA_eay_private_encrypt (signing)
9414 RSA_eay_public_decrypt (signature verification)
9416 (default implementations for RSA_public_encrypt,
9417 RSA_private_decrypt, RSA_private_encrypt, RSA_public_decrypt),
9418 always reject numbers >= n.
9421 *) In crypto/rand/md_rand.c, use a new short-time lock CRYPTO_LOCK_RAND2
9422 to synchronize access to 'locking_thread'. This is necessary on
9423 systems where access to 'locking_thread' (an 'unsigned long'
9424 variable) is not atomic.
9427 *) In crypto/rand/md_rand.c, set 'locking_thread' to current thread's ID
9428 *before* setting the 'crypto_lock_rand' flag. The previous code had
9429 a race condition if 0 is a valid thread ID.
9430 [Travis Vitek <vitek@roguewave.com>]
9432 *) Add support for shared libraries under Irix.
9433 [Albert Chin-A-Young <china@thewrittenword.com>]
9435 *) Add configuration option to build on Linux on both big-endian and
9437 [Ralf Baechle <ralf@uni-koblenz.de>]
9439 *) Add the possibility to create shared libraries on HP-UX.
9442 Changes between 0.9.6a and 0.9.6b [9 Jul 2001]
9444 *) Change ssleay_rand_bytes (crypto/rand/md_rand.c)
9445 to avoid a SSLeay/OpenSSL PRNG weakness pointed out by
9446 Markku-Juhani O. Saarinen <markku-juhani.saarinen@nokia.com>:
9447 PRNG state recovery was possible based on the output of
9448 one PRNG request appropriately sized to gain knowledge on
9449 'md' followed by enough consecutive 1-byte PRNG requests
9450 to traverse all of 'state'.
9452 1. When updating 'md_local' (the current thread's copy of 'md')
9453 during PRNG output generation, hash all of the previous
9454 'md_local' value, not just the half used for PRNG output.
9456 2. Make the number of bytes from 'state' included into the hash
9457 independent from the number of PRNG bytes requested.
9459 The first measure alone would be sufficient to avoid
9460 Markku-Juhani's attack. (Actually it had never occurred
9461 to me that the half of 'md_local' used for chaining was the
9462 half from which PRNG output bytes were taken -- I had always
9463 assumed that the secret half would be used.) The second
9464 measure makes sure that additional data from 'state' is never
9465 mixed into 'md_local' in small portions; this heuristically
9466 further strengthens the PRNG.
9469 *) Fix crypto/bn/asm/mips3.s.
9472 *) When only the key is given to "enc", the IV is undefined. Print out
9473 an error message in this case.
9476 *) Handle special case when X509_NAME is empty in X509 printing routines.
9479 *) In dsa_do_verify (crypto/dsa/dsa_ossl.c), verify that r and s are
9480 positive and less than q.
9483 *) Don't change *pointer in CRYPTO_add_lock() is add_lock_callback is
9484 used: it isn't thread safe and the add_lock_callback should handle
9486 [Paul Rose <Paul.Rose@bridge.com>]
9488 *) Verify that incoming data obeys the block size in
9489 ssl3_enc (ssl/s3_enc.c) and tls1_enc (ssl/t1_enc.c).
9493 [Ulf Möller, Bodo Möller]
9495 *) The countermeasure against Bleichbacher's attack on PKCS #1 v1.5
9496 RSA encryption was accidentally removed in s3_srvr.c in OpenSSL 0.9.5
9497 when fixing the server behaviour for backwards-compatible 'client
9498 hello' messages. (Note that the attack is impractical against
9499 SSL 3.0 and TLS 1.0 anyway because length and version checking
9500 means that the probability of guessing a valid ciphertext is
9501 around 2^-40; see section 5 in Bleichenbacher's CRYPTO '98
9504 Before 0.9.5, the countermeasure (hide the error by generating a
9505 random 'decryption result') did not work properly because
9506 ERR_clear_error() was missing, meaning that SSL_get_error() would
9507 detect the supposedly ignored error.
9509 Both problems are now fixed.
9512 *) In crypto/bio/bf_buff.c, increase DEFAULT_BUFFER_SIZE to 4096
9513 (previously it was 1024).
9516 *) Fix for compatibility mode trust settings: ignore trust settings
9517 unless some valid trust or reject settings are present.
9520 *) Fix for blowfish EVP: its a variable length cipher.
9523 *) Fix various bugs related to DSA S/MIME verification. Handle missing
9524 parameters in DSA public key structures and return an error in the
9525 DSA routines if parameters are absent.
9528 *) In versions up to 0.9.6, RAND_file_name() resorted to file ".rnd"
9529 in the current directory if neither $RANDFILE nor $HOME was set.
9530 RAND_file_name() in 0.9.6a returned NULL in this case. This has
9531 caused some confusion to Windows users who haven't defined $HOME.
9532 Thus RAND_file_name() is changed again: e_os.h can define a
9533 DEFAULT_HOME, which will be used if $HOME is not set.
9534 For Windows, we use "C:"; on other platforms, we still require
9535 environment variables.
9537 *) Move 'if (!initialized) RAND_poll()' into regions protected by
9538 CRYPTO_LOCK_RAND. This is not strictly necessary, but avoids
9539 having multiple threads call RAND_poll() concurrently.
9542 *) In crypto/rand/md_rand.c, replace 'add_do_not_lock' flag by a
9543 combination of a flag and a thread ID variable.
9544 Otherwise while one thread is in ssleay_rand_bytes (which sets the
9545 flag), *other* threads can enter ssleay_add_bytes without obeying
9546 the CRYPTO_LOCK_RAND lock (and may even illegally release the lock
9547 that they do not hold after the first thread unsets add_do_not_lock).
9550 *) Change bctest again: '-x' expressions are not available in all
9554 Changes between 0.9.6 and 0.9.6a [5 Apr 2001]
9556 *) Fix a couple of memory leaks in PKCS7_dataDecode()
9557 [Steve Henson, reported by Heyun Zheng <hzheng@atdsprint.com>]
9559 *) Change Configure and Makefiles to provide EXE_EXT, which will contain
9560 the default extension for executables, if any. Also, make the perl
9561 scripts that use symlink() to test if it really exists and use "cp"
9562 if it doesn't. All this made OpenSSL compilable and installable in
9566 *) Fix for asn1_GetSequence() for indefinite length constructed data.
9567 If SEQUENCE is length is indefinite just set c->slen to the total
9568 amount of data available.
9569 [Steve Henson, reported by shige@FreeBSD.org]
9570 [This change does not apply to 0.9.7.]
9572 *) Change bctest to avoid here-documents inside command substitution
9573 (workaround for FreeBSD /bin/sh bug).
9574 For compatibility with Ultrix, avoid shell functions (introduced
9575 in the bctest version that searches along $PATH).
9578 *) Rename 'des_encrypt' to 'des_encrypt1'. This avoids the clashes
9579 with des_encrypt() defined on some operating systems, like Solaris
9583 *) Check the result of RSA-CRT (see D. Boneh, R. DeMillo, R. Lipton:
9584 On the Importance of Eliminating Errors in Cryptographic
9585 Computations, J. Cryptology 14 (2001) 2, 101-119,
9586 http://theory.stanford.edu/~dabo/papers/faults.ps.gz).
9589 *) MIPS assembler BIGNUM division bug fix.
9592 *) Disabled incorrect Alpha assembler code.
9595 *) Fix PKCS#7 decode routines so they correctly update the length
9596 after reading an EOC for the EXPLICIT tag.
9598 [This change does not apply to 0.9.7.]
9600 *) Fix bug in PKCS#12 key generation routines. This was triggered
9601 if a 3DES key was generated with a 0 initial byte. Include
9602 PKCS12_BROKEN_KEYGEN compilation option to retain the old
9603 (but broken) behaviour.
9606 *) Enhance bctest to search for a working bc along $PATH and print
9608 [Tim Rice <tim@multitalents.net> via Richard Levitte]
9610 *) Fix memory leaks in err.c: free err_data string if necessary;
9611 don't write to the wrong index in ERR_set_error_data.
9614 *) Implement ssl23_peek (analogous to ssl23_read), which previously
9618 *) Replace rdtsc with _emit statements for VC++ version 5.
9619 [Jeremy Cooper <jeremy@baymoo.org>]
9621 *) Make it possible to reuse SSLv2 sessions.
9624 *) In copy_email() check for >= 0 as a return value for
9625 X509_NAME_get_index_by_NID() since 0 is a valid index.
9626 [Steve Henson reported by Massimiliano Pala <madwolf@opensca.org>]
9628 *) Avoid coredump with unsupported or invalid public keys by checking if
9629 X509_get_pubkey() fails in PKCS7_verify(). Fix memory leak when
9630 PKCS7_verify() fails with non detached data.
9633 *) Don't use getenv in library functions when run as setuid/setgid.
9634 New function OPENSSL_issetugid().
9637 *) Avoid false positives in memory leak detection code (crypto/mem_dbg.c)
9638 due to incorrect handling of multi-threading:
9640 1. Fix timing glitch in the MemCheck_off() portion of CRYPTO_mem_ctrl().
9642 2. Fix logical glitch in is_MemCheck_on() aka CRYPTO_is_mem_check_on().
9644 3. Count how many times MemCheck_off() has been called so that
9645 nested use can be treated correctly. This also avoids
9646 inband-signalling in the previous code (which relied on the
9647 assumption that thread ID 0 is impossible).
9650 *) Add "-rand" option also to s_client and s_server.
9653 *) Fix CPU detection on Irix 6.x.
9654 [Kurt Hockenbury <khockenb@stevens-tech.edu> and
9655 "Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9657 *) Fix X509_NAME bug which produced incorrect encoding if X509_NAME
9660 [This change does not apply to 0.9.7.]
9662 *) Use the cached encoding of an X509_NAME structure rather than
9663 copying it. This is apparently the reason for the libsafe "errors"
9664 but the code is actually correct.
9667 *) Add new function BN_rand_range(), and fix DSA_sign_setup() to prevent
9668 Bleichenbacher's DSA attack.
9669 Extend BN_[pseudo_]rand: As before, top=1 forces the highest two bits
9670 to be set and top=0 forces the highest bit to be set; top=-1 is new
9671 and leaves the highest bit random.
9672 [Ulf Moeller, Bodo Moeller]
9674 *) In the NCONF_...-based implementations for CONF_... queries
9675 (crypto/conf/conf_lib.c), if the input LHASH is NULL, avoid using
9676 a temporary CONF structure with the data component set to NULL
9677 (which gives segmentation faults in lh_retrieve).
9678 Instead, use NULL for the CONF pointer in CONF_get_string and
9679 CONF_get_number (which may use environment variables) and directly
9680 return NULL from CONF_get_section.
9683 *) Fix potential buffer overrun for EBCDIC.
9686 *) Tolerate nonRepudiation as being valid for S/MIME signing and certSign
9687 keyUsage if basicConstraints absent for a CA.
9690 *) Make SMIME_write_PKCS7() write mail header values with a format that
9691 is more generally accepted (no spaces before the semicolon), since
9692 some programs can't parse those values properly otherwise. Also make
9693 sure BIO's that break lines after each write do not create invalid
9697 *) Make the CRL encoding routines work with empty SEQUENCE OF. The
9698 macros previously used would not encode an empty SEQUENCE OF
9699 and break the signature.
9701 [This change does not apply to 0.9.7.]
9703 *) Zero the premaster secret after deriving the master secret in
9707 *) Add some EVP_add_digest_alias registrations (as found in
9708 OpenSSL_add_all_digests()) to SSL_library_init()
9709 aka OpenSSL_add_ssl_algorithms(). This provides improved
9710 compatibility with peers using X.509 certificates
9711 with unconventional AlgorithmIdentifier OIDs.
9714 *) Fix for Irix with NO_ASM.
9715 ["Bruce W. Forsberg" <bruce.forsberg@baesystems.com>]
9717 *) ./config script fixes.
9718 [Ulf Moeller, Richard Levitte]
9720 *) Fix 'openssl passwd -1'.
9723 *) Change PKCS12_key_gen_asc() so it can cope with non null
9724 terminated strings whose length is passed in the passlen
9725 parameter, for example from PEM callbacks. This was done
9726 by adding an extra length parameter to asc2uni().
9727 [Steve Henson, reported by <oddissey@samsung.co.kr>]
9729 *) Fix C code generated by 'openssl dsaparam -C': If a BN_bin2bn
9730 call failed, free the DSA structure.
9733 *) Fix to uni2asc() to cope with zero length Unicode strings.
9734 These are present in some PKCS#12 files.
9737 *) Increase s2->wbuf allocation by one byte in ssl2_new (ssl/s2_lib.c).
9738 Otherwise do_ssl_write (ssl/s2_pkt.c) will write beyond buffer limits
9739 when writing a 32767 byte record.
9740 [Bodo Moeller; problem reported by Eric Day <eday@concentric.net>]
9742 *) In RSA_eay_public_{en,ed}crypt and RSA_eay_mod_exp (rsa_eay.c),
9743 obtain lock CRYPTO_LOCK_RSA before setting rsa->_method_mod_{n,p,q}.
9745 (RSA objects have a reference count access to which is protected
9746 by CRYPTO_LOCK_RSA [see rsa_lib.c, s3_srvr.c, ssl_cert.c, ssl_rsa.c],
9747 so they are meant to be shared between threads.)
9748 [Bodo Moeller, Geoff Thorpe; original patch submitted by
9749 "Reddie, Steven" <Steven.Reddie@ca.com>]
9751 *) Fix a deadlock in CRYPTO_mem_leaks().
9754 *) Use better test patterns in bntest.
9757 *) rand_win.c fix for Borland C.
9760 *) BN_rshift bugfix for n == 0.
9763 *) Add a 'bctest' script that checks for some known 'bc' bugs
9764 so that 'make test' does not abort just because 'bc' is broken.
9767 *) Store verify_result within SSL_SESSION also for client side to
9768 avoid potential security hole. (Re-used sessions on the client side
9769 always resulted in verify_result==X509_V_OK, not using the original
9770 result of the server certificate verification.)
9773 *) Fix ssl3_pending: If the record in s->s3->rrec is not of type
9774 SSL3_RT_APPLICATION_DATA, return 0.
9775 Similarly, change ssl2_pending to return 0 if SSL_in_init(s) is true.
9779 Both ssl2_peek and ssl3_peek, which were totally broken in earlier
9780 releases, have been re-implemented by renaming the previous
9781 implementations of ssl2_read and ssl3_read to ssl2_read_internal
9782 and ssl3_read_internal, respectively, and adding 'peek' parameters
9783 to them. The new ssl[23]_{read,peek} functions are calls to
9784 ssl[23]_read_internal with the 'peek' flag set appropriately.
9785 A 'peek' parameter has also been added to ssl3_read_bytes, which
9786 does the actual work for ssl3_read_internal.
9789 *) Initialise "ex_data" member of RSA/DSA/DH structures prior to calling
9790 the method-specific "init()" handler. Also clean up ex_data after
9791 calling the method-specific "finish()" handler. Previously, this was
9792 happening the other way round.
9795 *) Increase BN_CTX_NUM (the number of BIGNUMs in a BN_CTX) to 16.
9796 The previous value, 12, was not always sufficient for BN_mod_exp().
9799 *) Make sure that shared libraries get the internal name engine with
9800 the full version number and not just 0. This should mark the
9801 shared libraries as not backward compatible. Of course, this should
9802 be changed again when we can guarantee backward binary compatibility.
9805 *) Fix typo in get_cert_by_subject() in by_dir.c
9806 [Jean-Marc Desperrier <jean-marc.desperrier@certplus.com>]
9808 *) Rework the system to generate shared libraries:
9810 - Make note of the expected extension for the shared libraries and
9811 if there is a need for symbolic links from for example libcrypto.so.0
9812 to libcrypto.so.0.9.7. There is extended info in Configure for
9815 - Make as few rebuilds of the shared libraries as possible.
9817 - Still avoid linking the OpenSSL programs with the shared libraries.
9819 - When installing, install the shared libraries separately from the
9823 *) Fix SSL_CTX_set_read_ahead macro to actually use its argument.
9825 Copy SSL_CTX's read_ahead flag to SSL object directly in SSL_new
9826 and not in SSL_clear because the latter is also used by the
9827 accept/connect functions; previously, the settings made by
9828 SSL_set_read_ahead would be lost during the handshake.
9829 [Bodo Moeller; problems reported by Anders Gertz <gertz@epact.se>]
9831 *) Correct util/mkdef.pl to be selective about disabled algorithms.
9832 Previously, it would create entries for disabled algorithms no
9836 *) Added several new manual pages for SSL_* function.
9839 Changes between 0.9.5a and 0.9.6 [24 Sep 2000]
9841 *) In ssl23_get_client_hello, generate an error message when faced
9842 with an initial SSL 3.0/TLS record that is too small to contain the
9843 first two bytes of the ClientHello message, i.e. client_version.
9844 (Note that this is a pathologic case that probably has never happened
9845 in real life.) The previous approach was to use the version number
9846 from the record header as a substitute; but our protocol choice
9847 should not depend on that one because it is not authenticated
9848 by the Finished messages.
9851 *) More robust randomness gathering functions for Windows.
9852 [Jeffrey Altman <jaltman@columbia.edu>]
9854 *) For compatibility reasons if the flag X509_V_FLAG_ISSUER_CHECK is
9855 not set then we don't setup the error code for issuer check errors
9856 to avoid possibly overwriting other errors which the callback does
9857 handle. If an application does set the flag then we assume it knows
9858 what it is doing and can handle the new informational codes
9862 *) Fix for a nasty bug in ASN1_TYPE handling. ASN1_TYPE is used for
9863 a general "ANY" type, as such it should be able to decode anything
9864 including tagged types. However it didn't check the class so it would
9865 wrongly interpret tagged types in the same way as their universal
9866 counterpart and unknown types were just rejected. Changed so that the
9867 tagged and unknown types are handled in the same way as a SEQUENCE:
9868 that is the encoding is stored intact. There is also a new type
9869 "V_ASN1_OTHER" which is used when the class is not universal, in this
9870 case we have no idea what the actual type is so we just lump them all
9874 *) On VMS, stdout may very well lead to a file that is written to
9875 in a record-oriented fashion. That means that every write() will
9876 write a separate record, which will be read separately by the
9877 programs trying to read from it. This can be very confusing.
9879 The solution is to put a BIO filter in the way that will buffer
9880 text until a linefeed is reached, and then write everything a
9881 line at a time, so every record written will be an actual line,
9882 not chunks of lines and not (usually doesn't happen, but I've
9883 seen it once) several lines in one record. BIO_f_linebuffer() is
9886 Currently, it's a VMS-only method, because that's where it has
9887 been tested well enough.
9890 *) Remove 'optimized' squaring variant in BN_mod_mul_montgomery,
9891 it can return incorrect results.
9892 (Note: The buggy variant was not enabled in OpenSSL 0.9.5a,
9893 but it was in 0.9.6-beta[12].)
9896 *) Disable the check for content being present when verifying detached
9897 signatures in pk7_smime.c. Some versions of Netscape (wrongly)
9898 include zero length content when signing messages.
9901 *) New BIO_shutdown_wr macro, which invokes the BIO_C_SHUTDOWN_WR
9902 BIO_ctrl (for BIO pairs).
9905 *) Add DSO method for VMS.
9908 *) Bug fix: Montgomery multiplication could produce results with the
9912 *) Add RPM specification openssl.spec and modify it to build three
9913 packages. The default package contains applications, application
9914 documentation and run-time libraries. The devel package contains
9915 include files, static libraries and function documentation. The
9916 doc package contains the contents of the doc directory. The original
9917 openssl.spec was provided by Damien Miller <djm@mindrot.org>.
9920 *) Add a large number of documentation files for many SSL routines.
9921 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
9923 *) Add a configuration entry for Sony News 4.
9924 [NAKAJI Hiroyuki <nakaji@tutrp.tut.ac.jp>]
9926 *) Don't set the two most significant bits to one when generating a
9927 random number < q in the DSA library.
9930 *) New SSL API mode 'SSL_MODE_AUTO_RETRY'. This disables the default
9931 behaviour that SSL_read may result in SSL_ERROR_WANT_READ (even if
9932 the underlying transport is blocking) if a handshake took place.
9933 (The default behaviour is needed by applications such as s_client
9934 and s_server that use select() to determine when to use SSL_read;
9935 but for applications that know in advance when to expect data, it
9936 just makes things more complicated.)
9939 *) Add RAND_egd_bytes(), which gives control over the number of bytes read
9943 *) Add a few more EBCDIC conditionals that make `req' and `x509'
9944 work better on such systems.
9945 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
9947 *) Add two demo programs for PKCS12_parse() and PKCS12_create().
9948 Update PKCS12_parse() so it copies the friendlyName and the
9949 keyid to the certificates aux info.
9952 *) Fix bug in PKCS7_verify() which caused an infinite loop
9953 if there was more than one signature.
9954 [Sven Uszpelkat <su@celocom.de>]
9956 *) Major change in util/mkdef.pl to include extra information
9957 about each symbol, as well as presenting variables as well
9958 as functions. This change means that there's n more need
9959 to rebuild the .num files when some algorithms are excluded.
9962 *) Allow the verify time to be set by an application,
9963 rather than always using the current time.
9966 *) Phase 2 verify code reorganisation. The certificate
9967 verify code now looks up an issuer certificate by a
9968 number of criteria: subject name, authority key id
9969 and key usage. It also verifies self signed certificates
9970 by the same criteria. The main comparison function is
9971 X509_check_issued() which performs these checks.
9973 Lot of changes were necessary in order to support this
9974 without completely rewriting the lookup code.
9976 Authority and subject key identifier are now cached.
9978 The LHASH 'certs' is X509_STORE has now been replaced
9979 by a STACK_OF(X509_OBJECT). This is mainly because an
9980 LHASH can't store or retrieve multiple objects with
9981 the same hash value.
9983 As a result various functions (which were all internal
9984 use only) have changed to handle the new X509_STORE
9985 structure. This will break anything that messed round
9986 with X509_STORE internally.
9988 The functions X509_STORE_add_cert() now checks for an
9989 exact match, rather than just subject name.
9991 The X509_STORE API doesn't directly support the retrieval
9992 of multiple certificates matching a given criteria, however
9993 this can be worked round by performing a lookup first
9994 (which will fill the cache with candidate certificates)
9995 and then examining the cache for matches. This is probably
9996 the best we can do without throwing out X509_LOOKUP
9997 entirely (maybe later...).
9999 The X509_VERIFY_CTX structure has been enhanced considerably.
10001 All certificate lookup operations now go via a get_issuer()
10002 callback. Although this currently uses an X509_STORE it
10003 can be replaced by custom lookups. This is a simple way
10004 to bypass the X509_STORE hackery necessary to make this
10005 work and makes it possible to use more efficient techniques
10006 in future. A very simple version which uses a simple
10007 STACK for its trusted certificate store is also provided
10008 using X509_STORE_CTX_trusted_stack().
10010 The verify_cb() and verify() callbacks now have equivalents
10011 in the X509_STORE_CTX structure.
10013 X509_STORE_CTX also has a 'flags' field which can be used
10014 to customise the verify behaviour.
10017 *) Add new PKCS#7 signing option PKCS7_NOSMIMECAP which
10018 excludes S/MIME capabilities.
10021 *) When a certificate request is read in keep a copy of the
10022 original encoding of the signed data and use it when outputting
10023 again. Signatures then use the original encoding rather than
10024 a decoded, encoded version which may cause problems if the
10025 request is improperly encoded.
10028 *) For consistency with other BIO_puts implementations, call
10029 buffer_write(b, ...) directly in buffer_puts instead of calling
10032 In BIO_puts, increment b->num_write as in BIO_write.
10033 [Peter.Sylvester@EdelWeb.fr]
10035 *) Fix BN_mul_word for the case where the word is 0. (We have to use
10036 BN_zero, we may not return a BIGNUM with an array consisting of
10037 words set to zero.)
10040 *) Avoid calling abort() from within the library when problems are
10041 detected, except if preprocessor symbols have been defined
10042 (such as REF_CHECK, BN_DEBUG etc.).
10045 *) New openssl application 'rsautl'. This utility can be
10046 used for low level RSA operations. DER public key
10047 BIO/fp routines also added.
10050 *) New Configure entry and patches for compiling on QNX 4.
10051 [Andreas Schneider <andreas@ds3.etech.fh-hamburg.de>]
10053 *) A demo state-machine implementation was sponsored by
10054 Nuron (http://www.nuron.com/) and is now available in
10055 demos/state_machine.
10058 *) New options added to the 'dgst' utility for signature
10059 generation and verification.
10062 *) Unrecognized PKCS#7 content types are now handled via a
10063 catch all ASN1_TYPE structure. This allows unsupported
10064 types to be stored as a "blob" and an application can
10065 encode and decode it manually.
10068 *) Fix various signed/unsigned issues to make a_strex.c
10069 compile under VC++.
10070 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
10072 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
10073 length if passed a buffer. ASN1_INTEGER_to_BN failed
10074 if passed a NULL BN and its argument was negative.
10075 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
10077 *) Modification to PKCS#7 encoding routines to output definite
10078 length encoding. Since currently the whole structures are in
10079 memory there's not real point in using indefinite length
10080 constructed encoding. However if OpenSSL is compiled with
10081 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
10084 *) Added BIO_vprintf() and BIO_vsnprintf().
10087 *) Added more prefixes to parse for in the strings written
10088 through a logging bio, to cover all the levels that are available
10089 through syslog. The prefixes are now:
10091 PANIC, EMERG, EMR => LOG_EMERG
10092 ALERT, ALR => LOG_ALERT
10093 CRIT, CRI => LOG_CRIT
10094 ERROR, ERR => LOG_ERR
10095 WARNING, WARN, WAR => LOG_WARNING
10096 NOTICE, NOTE, NOT => LOG_NOTICE
10097 INFO, INF => LOG_INFO
10098 DEBUG, DBG => LOG_DEBUG
10100 and as before, if none of those prefixes are present at the
10101 beginning of the string, LOG_ERR is chosen.
10103 On Win32, the LOG_* levels are mapped according to this:
10105 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
10106 LOG_WARNING => EVENTLOG_WARNING_TYPE
10107 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
10111 *) Made it possible to reconfigure with just the configuration
10112 argument "reconf" or "reconfigure". The command line arguments
10113 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
10114 and are retrieved from there when reconfiguring.
10117 *) MD4 implemented.
10118 [Assar Westerlund <assar@sics.se>, Richard Levitte]
10120 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
10123 *) The obj_dat.pl script was messing up the sorting of object
10124 names. The reason was that it compared the quoted version
10125 of strings as a result "OCSP" > "OCSP Signing" because
10126 " > SPACE. Changed script to store unquoted versions of
10127 names and add quotes on output. It was also omitting some
10128 names from the lookup table if they were given a default
10129 value (that is if SN is missing it is given the same
10130 value as LN and vice versa), these are now added on the
10131 grounds that if an object has a name we should be able to
10132 look it up. Finally added warning output when duplicate
10133 short or long names are found.
10136 *) Changes needed for Tandem NSK.
10137 [Scott Uroff <scott@xypro.com>]
10139 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
10140 RSA_padding_check_SSLv23(), special padding was never detected
10141 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
10142 version rollback attacks was not effective.
10144 In s23_clnt.c, don't use special rollback-attack detection padding
10145 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
10146 client; similarly, in s23_srvr.c, don't do the rollback check if
10147 SSL 2.0 is the only protocol enabled in the server.
10150 *) Make it possible to get hexdumps of unprintable data with 'openssl
10151 asn1parse'. By implication, the functions ASN1_parse_dump() and
10152 BIO_dump_indent() are added.
10155 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
10156 these print out strings and name structures based on various
10157 flags including RFC2253 support and proper handling of
10158 multibyte characters. Added options to the 'x509' utility
10159 to allow the various flags to be set.
10162 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
10163 Also change the functions X509_cmp_current_time() and
10164 X509_gmtime_adj() work with an ASN1_TIME structure,
10165 this will enable certificates using GeneralizedTime in validity
10166 dates to be checked.
10169 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
10170 negative public key encodings) on by default,
10171 NO_NEG_PUBKEY_BUG can be set to disable it.
10174 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
10175 content octets. An i2c_ASN1_OBJECT is unnecessary because
10176 the encoding can be trivially obtained from the structure.
10179 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
10180 not read locks (CRYPTO_r_[un]lock).
10183 *) A first attempt at creating official support for shared
10184 libraries through configuration. I've kept it so the
10185 default is static libraries only, and the OpenSSL programs
10186 are always statically linked for now, but there are
10187 preparations for dynamic linking in place.
10188 This has been tested on Linux and Tru64.
10191 *) Randomness polling function for Win9x, as described in:
10192 Peter Gutmann, Software Generation of Practically Strong
10196 *) Fix so PRNG is seeded in req if using an already existing
10200 *) New options to smime application. -inform and -outform
10201 allow alternative formats for the S/MIME message including
10202 PEM and DER. The -content option allows the content to be
10203 specified separately. This should allow things like Netscape
10204 form signing output easier to verify.
10207 *) Fix the ASN1 encoding of tags using the 'long form'.
10210 *) New ASN1 functions, i2c_* and c2i_* for INTEGER and BIT
10211 STRING types. These convert content octets to and from the
10212 underlying type. The actual tag and length octets are
10213 already assumed to have been read in and checked. These
10214 are needed because all other string types have virtually
10215 identical handling apart from the tag. By having versions
10216 of the ASN1 functions that just operate on content octets
10217 IMPLICIT tagging can be handled properly. It also allows
10218 the ASN1_ENUMERATED code to be cut down because ASN1_ENUMERATED
10219 and ASN1_INTEGER are identical apart from the tag.
10222 *) Change the handling of OID objects as follows:
10224 - New object identifiers are inserted in objects.txt, following
10225 the syntax given in objects.README.
10226 - objects.pl is used to process obj_mac.num and create a new
10228 - obj_dat.pl is used to create a new obj_dat.h, using the data in
10231 This is currently kind of a hack, and the perl code in objects.pl
10232 isn't very elegant, but it works as I intended. The simplest way
10233 to check that it worked correctly is to look in obj_dat.h and
10234 check the array nid_objs and make sure the objects haven't moved
10235 around (this is important!). Additions are OK, as well as
10236 consistent name changes.
10239 *) Add BSD-style MD5-based passwords to 'openssl passwd' (option '-1').
10242 *) Addition of the command line parameter '-rand file' to 'openssl req'.
10243 The given file adds to whatever has already been seeded into the
10244 random pool through the RANDFILE configuration file option or
10245 environment variable, or the default random state file.
10248 *) mkstack.pl now sorts each macro group into lexical order.
10249 Previously the output order depended on the order the files
10250 appeared in the directory, resulting in needless rewriting
10254 *) Patches to make OpenSSL compile under Win32 again. Mostly
10255 work arounds for the VC++ problem that it treats func() as
10256 func(void). Also stripped out the parts of mkdef.pl that
10257 added extra typesafe functions: these no longer exist.
10260 *) Reorganisation of the stack code. The macros are now all
10261 collected in safestack.h . Each macro is defined in terms of
10262 a "stack macro" of the form SKM_<name>(type, a, b). The
10263 DEBUG_SAFESTACK is now handled in terms of function casts,
10264 this has the advantage of retaining type safety without the
10265 use of additional functions. If DEBUG_SAFESTACK is not defined
10266 then the non typesafe macros are used instead. Also modified the
10267 mkstack.pl script to handle the new form. Needs testing to see
10268 if which (if any) compilers it chokes and maybe make DEBUG_SAFESTACK
10269 the default if no major problems. Similar behaviour for ASN1_SET_OF
10270 and PKCS12_STACK_OF.
10273 *) When some versions of IIS use the 'NET' form of private key the
10274 key derivation algorithm is different. Normally MD5(password) is
10275 used as a 128 bit RC4 key. In the modified case
10276 MD5(MD5(password) + "SGCKEYSALT") is used instead. Added some
10277 new functions i2d_RSA_NET(), d2i_RSA_NET() etc which are the same
10278 as the old Netscape_RSA functions except they have an additional
10279 'sgckey' parameter which uses the modified algorithm. Also added
10280 an -sgckey command line option to the rsa utility. Thanks to
10281 Adrian Peck <bertie@ncipher.com> for posting details of the modified
10282 algorithm to openssl-dev.
10285 *) The evp_local.h macros were using 'c.##kname' which resulted in
10286 invalid expansion on some systems (SCO 5.0.5 for example).
10287 Corrected to 'c.kname'.
10288 [Phillip Porch <root@theporch.com>]
10290 *) New X509_get1_email() and X509_REQ_get1_email() functions that return
10291 a STACK of email addresses from a certificate or request, these look
10292 in the subject name and the subject alternative name extensions and
10293 omit any duplicate addresses.
10296 *) Re-implement BN_mod_exp2_mont using independent (and larger) windows.
10297 This makes DSA verification about 2 % faster.
10300 *) Increase maximum window size in BN_mod_exp_... to 6 bits instead of 5
10301 (meaning that now 2^5 values will be precomputed, which is only 4 KB
10302 plus overhead for 1024 bit moduli).
10303 This makes exponentiations about 0.5 % faster for 1024 bit
10304 exponents (as measured by "openssl speed rsa2048").
10307 *) Rename memory handling macros to avoid conflicts with other
10309 Malloc => OPENSSL_malloc
10310 Malloc_locked => OPENSSL_malloc_locked
10311 Realloc => OPENSSL_realloc
10312 Free => OPENSSL_free
10315 *) New function BN_mod_exp_mont_word for small bases (roughly 15%
10316 faster than BN_mod_exp_mont, i.e. 7% for a full DH exchange).
10319 *) CygWin32 support.
10320 [John Jarvie <jjarvie@newsguy.com>]
10322 *) The type-safe stack code has been rejigged. It is now only compiled
10323 in when OpenSSL is configured with the DEBUG_SAFESTACK option and
10324 by default all type-specific stack functions are "#define"d back to
10325 standard stack functions. This results in more streamlined output
10326 but retains the type-safety checking possibilities of the original
10330 *) The STACK code has been cleaned up, and certain type declarations
10331 that didn't make a lot of sense have been brought in line. This has
10332 also involved a cleanup of sorts in safestack.h to more correctly
10333 map type-safe stack functions onto their plain stack counterparts.
10334 This work has also resulted in a variety of "const"ifications of
10335 lots of the code, especially "_cmp" operations which should normally
10336 be prototyped with "const" parameters anyway.
10339 *) When generating bytes for the first time in md_rand.c, 'stir the pool'
10340 by seeding with STATE_SIZE dummy bytes (with zero entropy count).
10341 (The PRNG state consists of two parts, the large pool 'state' and 'md',
10342 where all of 'md' is used each time the PRNG is used, but 'state'
10343 is used only indexed by a cyclic counter. As entropy may not be
10344 well distributed from the beginning, 'md' is important as a
10345 chaining variable. However, the output function chains only half
10346 of 'md', i.e. 80 bits. ssleay_rand_add, on the other hand, chains
10347 all of 'md', and seeding with STATE_SIZE dummy bytes will result
10348 in all of 'state' being rewritten, with the new values depending
10349 on virtually all of 'md'. This overcomes the 80 bit limitation.)
10352 *) In ssl/s2_clnt.c and ssl/s3_clnt.c, call ERR_clear_error() when
10353 the handshake is continued after ssl_verify_cert_chain();
10354 otherwise, if SSL_VERIFY_NONE is set, remaining error codes
10355 can lead to 'unexplainable' connection aborts later.
10356 [Bodo Moeller; problem tracked down by Lutz Jaenicke]
10358 *) Major EVP API cipher revision.
10359 Add hooks for extra EVP features. This allows various cipher
10360 parameters to be set in the EVP interface. Support added for variable
10361 key length ciphers via the EVP_CIPHER_CTX_set_key_length() function and
10362 setting of RC2 and RC5 parameters.
10364 Modify EVP_OpenInit() and EVP_SealInit() to cope with variable key length
10367 Remove lots of duplicated code from the EVP library. For example *every*
10368 cipher init() function handles the 'iv' in the same way according to the
10369 cipher mode. They also all do nothing if the 'key' parameter is NULL and
10370 for CFB and OFB modes they zero ctx->num.
10372 New functionality allows removal of S/MIME code RC2 hack.
10374 Most of the routines have the same form and so can be declared in terms
10377 By shifting this to the top level EVP_CipherInit() it can be removed from
10378 all individual ciphers. If the cipher wants to handle IVs or keys
10379 differently it can set the EVP_CIPH_CUSTOM_IV or EVP_CIPH_ALWAYS_CALL_INIT
10382 Change lots of functions like EVP_EncryptUpdate() to now return a
10383 value: although software versions of the algorithms cannot fail
10384 any installed hardware versions can.
10387 *) Implement SSL_OP_TLS_ROLLBACK_BUG: In ssl3_get_client_key_exchange, if
10388 this option is set, tolerate broken clients that send the negotiated
10389 protocol version number instead of the requested protocol version
10393 *) Call dh_tmp_cb (set by ..._TMP_DH_CB) with correct 'is_export' flag;
10394 i.e. non-zero for export ciphersuites, zero otherwise.
10395 Previous versions had this flag inverted, inconsistent with
10396 rsa_tmp_cb (..._TMP_RSA_CB).
10397 [Bodo Moeller; problem reported by Amit Chopra]
10399 *) Add missing DSA library text string. Work around for some IIS
10400 key files with invalid SEQUENCE encoding.
10403 *) Add a document (doc/standards.txt) that list all kinds of standards
10404 and so on that are implemented in OpenSSL.
10407 *) Enhance c_rehash script. Old version would mishandle certificates
10408 with the same subject name hash and wouldn't handle CRLs at all.
10409 Added -fingerprint option to crl utility, to support new c_rehash
10413 *) Eliminate non-ANSI declarations in crypto.h and stack.h.
10416 *) Fix for SSL server purpose checking. Server checking was
10417 rejecting certificates which had extended key usage present
10418 but no ssl client purpose.
10419 [Steve Henson, reported by Rene Grosser <grosser@hisolutions.com>]
10421 *) Make PKCS#12 code work with no password. The PKCS#12 spec
10422 is a little unclear about how a blank password is handled.
10423 Since the password in encoded as a BMPString with terminating
10424 double NULL a zero length password would end up as just the
10425 double NULL. However no password at all is different and is
10426 handled differently in the PKCS#12 key generation code. NS
10427 treats a blank password as zero length. MSIE treats it as no
10428 password on export: but it will try both on import. We now do
10429 the same: PKCS12_parse() tries zero length and no password if
10430 the password is set to "" or NULL (NULL is now a valid password:
10431 it wasn't before) as does the pkcs12 application.
10434 *) Bugfixes in apps/x509.c: Avoid a memory leak; and don't use
10435 perror when PEM_read_bio_X509_REQ fails, the error message must
10436 be obtained from the error queue.
10439 *) Avoid 'thread_hash' memory leak in crypto/err/err.c by freeing
10440 it in ERR_remove_state if appropriate, and change ERR_get_state
10441 accordingly to avoid race conditions (this is necessary because
10442 thread_hash is no longer constant once set).
10445 *) Bugfix for linux-elf makefile.one.
10448 *) RSA_get_default_method() will now cause a default
10449 RSA_METHOD to be chosen if one doesn't exist already.
10450 Previously this was only set during a call to RSA_new()
10451 or RSA_new_method(NULL) meaning it was possible for
10452 RSA_get_default_method() to return NULL.
10455 *) Added native name translation to the existing DSO code
10456 that will convert (if the flag to do so is set) filenames
10457 that are sufficiently small and have no path information
10458 into a canonical native form. Eg. "blah" converted to
10459 "libblah.so" or "blah.dll" etc.
10462 *) New function ERR_error_string_n(e, buf, len) which is like
10463 ERR_error_string(e, buf), but writes at most 'len' bytes
10464 including the 0 terminator. For ERR_error_string_n, 'buf'
10466 [Damien Miller <djm@mindrot.org>, Bodo Moeller]
10468 *) CONF library reworked to become more general. A new CONF
10469 configuration file reader "class" is implemented as well as a
10470 new functions (NCONF_*, for "New CONF") to handle it. The now
10471 old CONF_* functions are still there, but are reimplemented to
10472 work in terms of the new functions. Also, a set of functions
10473 to handle the internal storage of the configuration data is
10474 provided to make it easier to write new configuration file
10475 reader "classes" (I can definitely see something reading a
10476 configuration file in XML format, for example), called _CONF_*,
10477 or "the configuration storage API"...
10479 The new configuration file reading functions are:
10481 NCONF_new, NCONF_free, NCONF_load, NCONF_load_fp, NCONF_load_bio,
10482 NCONF_get_section, NCONF_get_string, NCONF_get_numbre
10484 NCONF_default, NCONF_WIN32
10486 NCONF_dump_fp, NCONF_dump_bio
10488 NCONF_default and NCONF_WIN32 are method (or "class") choosers,
10489 NCONF_new creates a new CONF object. This works in the same way
10490 as other interfaces in OpenSSL, like the BIO interface.
10491 NCONF_dump_* dump the internal storage of the configuration file,
10492 which is useful for debugging. All other functions take the same
10493 arguments as the old CONF_* functions wth the exception of the
10494 first that must be a `CONF *' instead of a `LHASH *'.
10496 To make it easer to use the new classes with the old CONF_* functions,
10497 the function CONF_set_default_method is provided.
10500 *) Add '-tls1' option to 'openssl ciphers', which was already
10501 mentioned in the documentation but had not been implemented.
10502 (This option is not yet really useful because even the additional
10503 experimental TLS 1.0 ciphers are currently treated as SSL 3.0 ciphers.)
10506 *) Initial DSO code added into libcrypto for letting OpenSSL (and
10507 OpenSSL-based applications) load shared libraries and bind to
10508 them in a portable way.
10509 [Geoff Thorpe, with contributions from Richard Levitte]
10511 Changes between 0.9.5 and 0.9.5a [1 Apr 2000]
10513 *) Make sure _lrotl and _lrotr are only used with MSVC.
10515 *) Use lock CRYPTO_LOCK_RAND correctly in ssleay_rand_status
10516 (the default implementation of RAND_status).
10518 *) Rename openssl x509 option '-crlext', which was added in 0.9.5,
10519 to '-clrext' (= clear extensions), as intended and documented.
10520 [Bodo Moeller; inconsistency pointed out by Michael Attili
10521 <attili@amaxo.com>]
10523 *) Fix for HMAC. It wasn't zeroing the rest of the block if the key length
10524 was larger than the MD block size.
10525 [Steve Henson, pointed out by Yost William <YostW@tce.com>]
10527 *) Modernise PKCS12_parse() so it uses STACK_OF(X509) for its ca argument
10528 fix a leak when the ca argument was passed as NULL. Stop X509_PUBKEY_set()
10529 using the passed key: if the passed key was a private key the result
10530 of X509_print(), for example, would be to print out all the private key
10534 *) des_quad_cksum() byte order bug fix.
10535 [Ulf Möller, using the problem description in krb4-0.9.7, where
10536 the solution is attributed to Derrick J Brashear <shadow@DEMENTIA.ORG>]
10538 *) Fix so V_ASN1_APP_CHOOSE works again: however its use is strongly
10540 [Steve Henson, pointed out by Brian Korver <briank@cs.stanford.edu>]
10542 *) For easily testing in shell scripts whether some command
10543 'openssl XXX' exists, the new pseudo-command 'openssl no-XXX'
10544 returns with exit code 0 iff no command of the given name is available.
10545 'no-XXX' is printed in this case, 'XXX' otherwise. In both cases,
10546 the output goes to stdout and nothing is printed to stderr.
10547 Additional arguments are always ignored.
10549 Since for each cipher there is a command of the same name,
10550 the 'no-cipher' compilation switches can be tested this way.
10552 ('openssl no-XXX' is not able to detect pseudo-commands such
10553 as 'quit', 'list-XXX-commands', or 'no-XXX' itself.)
10556 *) Update test suite so that 'make test' succeeds in 'no-rsa' configuration.
10559 *) For SSL_[CTX_]set_tmp_dh, don't create a DH key if SSL_OP_SINGLE_DH_USE
10560 is set; it will be thrown away anyway because each handshake creates
10562 ssl_cert_dup, which is used by SSL_new, now copies DH keys in addition
10563 to parameters -- in previous versions (since OpenSSL 0.9.3) the
10564 'default key' from SSL_CTX_set_tmp_dh would always be lost, meaning
10565 you effectively got SSL_OP_SINGLE_DH_USE when using this macro.
10568 *) New s_client option -ign_eof: EOF at stdin is ignored, and
10569 'Q' and 'R' lose their special meanings (quit/renegotiate).
10570 This is part of what -quiet does; unlike -quiet, -ign_eof
10571 does not suppress any output.
10574 *) Add compatibility options to the purpose and trust code. The
10575 purpose X509_PURPOSE_ANY is "any purpose" which automatically
10576 accepts a certificate or CA, this was the previous behaviour,
10577 with all the associated security issues.
10579 X509_TRUST_COMPAT is the old trust behaviour: only and
10580 automatically trust self signed roots in certificate store. A
10581 new trust setting X509_TRUST_DEFAULT is used to specify that
10582 a purpose has no associated trust setting and it should instead
10583 use the value in the default purpose.
10586 *) Fix the PKCS#8 DSA private key code so it decodes keys again
10587 and fix a memory leak.
10590 *) In util/mkerr.pl (which implements 'make errors'), preserve
10591 reason strings from the previous version of the .c file, as
10592 the default to have only downcase letters (and digits) in
10593 automatically generated reasons codes is not always appropriate.
10596 *) In ERR_load_ERR_strings(), build an ERR_LIB_SYS error reason table
10597 using strerror. Previously, ERR_reason_error_string() returned
10598 library names as reason strings for SYSerr; but SYSerr is a special
10599 case where small numbers are errno values, not library numbers.
10602 *) Add '-dsaparam' option to 'openssl dhparam' application. This
10603 converts DSA parameters into DH parameters. (When creating parameters,
10604 DSA_generate_parameters is used.)
10607 *) Include 'length' (recommended exponent length) in C code generated
10608 by 'openssl dhparam -C'.
10611 *) The second argument to set_label in perlasm was already being used
10612 so couldn't be used as a "file scope" flag. Moved to third argument
10616 *) In PEM_ASN1_write_bio and some other functions, use RAND_pseudo_bytes
10617 instead of RAND_bytes for encryption IVs and salts.
10620 *) Include RAND_status() into RAND_METHOD instead of implementing
10621 it only for md_rand.c Otherwise replacing the PRNG by calling
10622 RAND_set_rand_method would be impossible.
10625 *) Don't let DSA_generate_key() enter an infinite loop if the random
10626 number generation fails.
10629 *) New 'rand' application for creating pseudo-random output.
10632 *) Added configuration support for Linux/IA64
10633 [Rolf Haberrecker <rolf@suse.de>]
10635 *) Assembler module support for Mingw32.
10638 *) Shared library support for HPUX (in shlib/).
10639 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE> and Anonymous]
10641 *) Shared library support for Solaris gcc.
10642 [Lutz Behnke <behnke@trustcenter.de>]
10644 Changes between 0.9.4 and 0.9.5 [28 Feb 2000]
10646 *) PKCS7_encrypt() was adding text MIME headers twice because they
10647 were added manually and by SMIME_crlf_copy().
10650 *) In bntest.c don't call BN_rand with zero bits argument.
10651 [Steve Henson, pointed out by Andrew W. Gray <agray@iconsinc.com>]
10653 *) BN_mul bugfix: In bn_mul_part_recursion() only the a>a[n] && b>b[n]
10654 case was implemented. This caused BN_div_recp() to fail occasionally.
10657 *) Add an optional second argument to the set_label() in the perl
10658 assembly language builder. If this argument exists and is set
10659 to 1 it signals that the assembler should use a symbol whose
10660 scope is the entire file, not just the current function. This
10661 is needed with MASM which uses the format label:: for this scope.
10662 [Steve Henson, pointed out by Peter Runestig <peter@runestig.com>]
10664 *) Change the ASN1 types so they are typedefs by default. Before
10665 almost all types were #define'd to ASN1_STRING which was causing
10666 STACK_OF() problems: you couldn't declare STACK_OF(ASN1_UTF8STRING)
10670 *) Change names of new functions to the new get1/get0 naming
10671 convention: After 'get1', the caller owns a reference count
10672 and has to call ..._free; 'get0' returns a pointer to some
10673 data structure without incrementing reference counters.
10674 (Some of the existing 'get' functions increment a reference
10675 counter, some don't.)
10676 Similarly, 'set1' and 'add1' functions increase reference
10677 counters or duplicate objects.
10680 *) Allow for the possibility of temp RSA key generation failure:
10681 the code used to assume it always worked and crashed on failure.
10684 *) Fix potential buffer overrun problem in BIO_printf().
10685 [Ulf Möller, using public domain code by Patrick Powell; problem
10686 pointed out by David Sacerdote <das33@cornell.edu>]
10688 *) Support EGD <http://www.lothar.com/tech/crypto/>. New functions
10689 RAND_egd() and RAND_status(). In the command line application,
10690 the EGD socket can be specified like a seed file using RANDFILE
10694 *) Allow the string CERTIFICATE to be tolerated in PKCS#7 structures.
10695 Some CAs (e.g. Verisign) distribute certificates in this form.
10698 *) Remove the SSL_ALLOW_ADH compile option and set the default cipher
10699 list to exclude them. This means that no special compilation option
10700 is needed to use anonymous DH: it just needs to be included in the
10704 *) Change the EVP_MD_CTX_type macro so its meaning consistent with
10705 EVP_MD_type. The old functionality is available in a new macro called
10706 EVP_MD_md(). Change code that uses it and update docs.
10709 *) ..._ctrl functions now have corresponding ..._callback_ctrl functions
10710 where the 'void *' argument is replaced by a function pointer argument.
10711 Previously 'void *' was abused to point to functions, which works on
10712 many platforms, but is not correct. As these functions are usually
10713 called by macros defined in OpenSSL header files, most source code
10714 should work without changes.
10717 *) <openssl/opensslconf.h> (which is created by Configure) now contains
10718 sections with information on -D... compiler switches used for
10719 compiling the library so that applications can see them. To enable
10720 one of these sections, a pre-processor symbol OPENSSL_..._DEFINES
10721 must be defined. E.g.,
10722 #define OPENSSL_ALGORITHM_DEFINES
10723 #include <openssl/opensslconf.h>
10724 defines all pertinent NO_<algo> symbols, such as NO_IDEA, NO_RSA, etc.
10725 [Richard Levitte, Ulf and Bodo Möller]
10727 *) Bugfix: Tolerate fragmentation and interleaving in the SSL 3/TLS
10731 *) Change the 'other' type in certificate aux info to a STACK_OF
10732 X509_ALGOR. Although not an AlgorithmIdentifier as such it has
10733 the required ASN1 format: arbitrary types determined by an OID.
10736 *) Add some PEM_write_X509_REQ_NEW() functions and a command line
10737 argument to 'req'. This is not because the function is newer or
10738 better than others it just uses the work 'NEW' in the certificate
10739 request header lines. Some software needs this.
10742 *) Reorganise password command line arguments: now passwords can be
10743 obtained from various sources. Delete the PEM_cb function and make
10744 it the default behaviour: i.e. if the callback is NULL and the
10745 usrdata argument is not NULL interpret it as a null terminated pass
10746 phrase. If usrdata and the callback are NULL then the pass phrase
10747 is prompted for as usual.
10750 *) Add support for the Compaq Atalla crypto accelerator. If it is installed,
10751 the support is automatically enabled. The resulting binaries will
10752 autodetect the card and use it if present.
10753 [Ben Laurie and Compaq Inc.]
10755 *) Work around for Netscape hang bug. This sends certificate request
10756 and server done in one record. Since this is perfectly legal in the
10757 SSL/TLS protocol it isn't a "bug" option and is on by default. See
10758 the bugs/SSLv3 entry for more info.
10761 *) HP-UX tune-up: new unified configs, HP C compiler bug workaround.
10764 *) Add -rand argument to smime and pkcs12 applications and read/write
10768 *) New 'passwd' tool for crypt(3) and apr1 password hashes.
10771 *) Add command line password options to the remaining applications.
10774 *) Bug fix for BN_div_recp() for numerators with an even number of
10778 *) More tests in bntest.c, and changed test_bn output.
10781 *) ./config recognizes MacOS X now.
10784 *) Bug fix for BN_div() when the first words of num and divisor are
10785 equal (it gave wrong results if (rem=(n1-q*d0)&BN_MASK2) < d0).
10788 *) Add support for various broken PKCS#8 formats, and command line
10789 options to produce them.
10792 *) New functions BN_CTX_start(), BN_CTX_get() and BT_CTX_end() to
10793 get temporary BIGNUMs from a BN_CTX.
10796 *) Correct return values in BN_mod_exp_mont() and BN_mod_exp2_mont()
10800 *) Change the SSLeay_add_all_*() functions to OpenSSL_add_all_*() and
10801 include a #define from the old name to the new. The original intent
10802 was that statically linked binaries could for example just call
10803 SSLeay_add_all_ciphers() to just add ciphers to the table and not
10804 link with digests. This never worked because SSLeay_add_all_digests()
10805 and SSLeay_add_all_ciphers() were in the same source file so calling
10806 one would link with the other. They are now in separate source files.
10809 *) Add a new -notext option to 'ca' and a -pubkey option to 'spkac'.
10812 *) Use a less unusual form of the Miller-Rabin primality test (it used
10813 a binary algorithm for exponentiation integrated into the Miller-Rabin
10814 loop, our standard modexp algorithms are faster).
10817 *) Support for the EBCDIC character set completed.
10818 [Martin Kraemer <Martin.Kraemer@Mch.SNI.De>]
10820 *) Source code cleanups: use const where appropriate, eliminate casts,
10821 use void * instead of char * in lhash.
10824 *) Bugfix: ssl3_send_server_key_exchange was not restartable
10825 (the state was not changed to SSL3_ST_SW_KEY_EXCH_B, and because of
10826 this the server could overwrite ephemeral keys that the client
10830 *) Turn DSA_is_prime into a macro that calls BN_is_prime,
10831 using 50 iterations of the Rabin-Miller test.
10833 DSA_generate_parameters now uses BN_is_prime_fasttest (with 50
10834 iterations of the Rabin-Miller test as required by the appendix
10835 to FIPS PUB 186[-1]) instead of DSA_is_prime.
10836 As BN_is_prime_fasttest includes trial division, DSA parameter
10837 generation becomes much faster.
10839 This implies a change for the callback functions in DSA_is_prime
10840 and DSA_generate_parameters: The callback function is called once
10841 for each positive witness in the Rabin-Miller test, not just
10842 occasionally in the inner loop; and the parameters to the
10843 callback function now provide an iteration count for the outer
10844 loop rather than for the current invocation of the inner loop.
10845 DSA_generate_parameters additionally can call the callback
10846 function with an 'iteration count' of -1, meaning that a
10847 candidate has passed the trial division test (when q is generated
10848 from an application-provided seed, trial division is skipped).
10851 *) New function BN_is_prime_fasttest that optionally does trial
10852 division before starting the Rabin-Miller test and has
10853 an additional BN_CTX * argument (whereas BN_is_prime always
10854 has to allocate at least one BN_CTX).
10855 'callback(1, -1, cb_arg)' is called when a number has passed the
10856 trial division stage.
10859 *) Fix for bug in CRL encoding. The validity dates weren't being handled
10863 *) New -pkcs12 option to CA.pl script to write out a PKCS#12 file.
10866 *) New function BN_pseudo_rand().
10869 *) Clean up BN_mod_mul_montgomery(): replace the broken (and unreadable)
10870 bignum version of BN_from_montgomery() with the working code from
10871 SSLeay 0.9.0 (the word based version is faster anyway), and clean up
10875 *) Avoid a race condition in s2_clnt.c (function get_server_hello) that
10876 made it impossible to use the same SSL_SESSION data structure in
10877 SSL2 clients in multiple threads.
10880 *) The return value of RAND_load_file() no longer counts bytes obtained
10881 by stat(). RAND_load_file(..., -1) is new and uses the complete file
10882 to seed the PRNG (previously an explicit byte count was required).
10883 [Ulf Möller, Bodo Möller]
10885 *) Clean up CRYPTO_EX_DATA functions, some of these didn't have prototypes
10886 used (char *) instead of (void *) and had casts all over the place.
10889 *) Make BN_generate_prime() return NULL on error if ret!=NULL.
10892 *) Retain source code compatibility for BN_prime_checks macro:
10893 BN_is_prime(..., BN_prime_checks, ...) now uses
10894 BN_prime_checks_for_size to determine the appropriate number of
10895 Rabin-Miller iterations.
10898 *) Diffie-Hellman uses "safe" primes: DH_check() return code renamed to
10899 DH_CHECK_P_NOT_SAFE_PRIME.
10900 (Check if this is true? OpenPGP calls them "strong".)
10903 *) Merge the functionality of "dh" and "gendh" programs into a new program
10904 "dhparam". The old programs are retained for now but will handle DH keys
10905 (instead of parameters) in future.
10908 *) Make the ciphers, s_server and s_client programs check the return values
10909 when a new cipher list is set.
10912 *) Enhance the SSL/TLS cipher mechanism to correctly handle the TLS 56bit
10913 ciphers. Before when the 56bit ciphers were enabled the sorting was
10916 The syntax for the cipher sorting has been extended to support sorting by
10917 cipher-strength (using the strength_bits hard coded in the tables).
10918 The new command is "@STRENGTH" (see also doc/apps/ciphers.pod).
10920 Fix a bug in the cipher-command parser: when supplying a cipher command
10921 string with an "undefined" symbol (neither command nor alphanumeric
10922 [A-Za-z0-9], ssl_set_cipher_list used to hang in an endless loop. Now
10923 an error is flagged.
10925 Due to the strength-sorting extension, the code of the
10926 ssl_create_cipher_list() function was completely rearranged. I hope that
10927 the readability was also increased :-)
10928 [Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>]
10930 *) Minor change to 'x509' utility. The -CAcreateserial option now uses 1
10931 for the first serial number and places 2 in the serial number file. This
10932 avoids problems when the root CA is created with serial number zero and
10933 the first user certificate has the same issuer name and serial number
10937 *) Fixes to X509_ATTRIBUTE utilities, change the 'req' program so it uses
10938 the new code. Add documentation for this stuff.
10941 *) Changes to X509_ATTRIBUTE utilities. These have been renamed from
10942 X509_*() to X509at_*() on the grounds that they don't handle X509
10943 structures and behave in an analogous way to the X509v3 functions:
10944 they shouldn't be called directly but wrapper functions should be used
10947 So we also now have some wrapper functions that call the X509at functions
10948 when passed certificate requests. (TO DO: similar things can be done with
10949 PKCS#7 signed and unsigned attributes, PKCS#12 attributes and a few other
10950 things. Some of these need some d2i or i2d and print functionality
10951 because they handle more complex structures.)
10954 *) Add missing #ifndefs that caused missing symbols when building libssl
10955 as a shared library without RSA. Use #ifndef NO_SSL2 instead of
10956 NO_RSA in ssl/s2*.c.
10957 [Kris Kennaway <kris@hub.freebsd.org>, modified by Ulf Möller]
10959 *) Precautions against using the PRNG uninitialized: RAND_bytes() now
10960 has a return value which indicates the quality of the random data
10961 (1 = ok, 0 = not seeded). Also an error is recorded on the thread's
10962 error queue. New function RAND_pseudo_bytes() generates output that is
10963 guaranteed to be unique but not unpredictable. RAND_add is like
10964 RAND_seed, but takes an extra argument for an entropy estimate
10965 (RAND_seed always assumes full entropy).
10968 *) Do more iterations of Rabin-Miller probable prime test (specifically,
10969 3 for 1024-bit primes, 6 for 512-bit primes, 12 for 256-bit primes
10970 instead of only 2 for all lengths; see BN_prime_checks_for_size definition
10971 in crypto/bn/bn_prime.c for the complete table). This guarantees a
10972 false-positive rate of at most 2^-80 for random input.
10975 *) Rewrite ssl3_read_n (ssl/s3_pkt.c) avoiding a couple of bugs.
10978 *) New function X509_CTX_rget_chain() (renamed to X509_CTX_get1_chain
10979 in the 0.9.5 release), this returns the chain
10980 from an X509_CTX structure with a dup of the stack and all
10981 the X509 reference counts upped: so the stack will exist
10982 after X509_CTX_cleanup() has been called. Modify pkcs12.c
10985 Also make SSL_SESSION_print() print out the verify return
10989 *) Add manpage for the pkcs12 command. Also change the default
10990 behaviour so MAC iteration counts are used unless the new
10991 -nomaciter option is used. This improves file security and
10992 only older versions of MSIE (4.0 for example) need it.
10995 *) Honor the no-xxx Configure options when creating .DEF files.
10998 *) Add PKCS#10 attributes to field table: challengePassword,
10999 unstructuredName and unstructuredAddress. These are taken from
11000 draft PKCS#9 v2.0 but are compatible with v1.2 provided no
11001 international characters are used.
11003 More changes to X509_ATTRIBUTE code: allow the setting of types
11004 based on strings. Remove the 'loc' parameter when adding
11005 attributes because these will be a SET OF encoding which is sorted
11009 *) Initial changes to the 'req' utility to allow request generation
11010 automation. This will allow an application to just generate a template
11011 file containing all the field values and have req construct the
11014 Initial support for X509_ATTRIBUTE handling. Stacks of these are
11015 used all over the place including certificate requests and PKCS#7
11016 structures. They are currently handled manually where necessary with
11017 some primitive wrappers for PKCS#7. The new functions behave in a
11018 manner analogous to the X509 extension functions: they allow
11019 attributes to be looked up by NID and added.
11021 Later something similar to the X509V3 code would be desirable to
11022 automatically handle the encoding, decoding and printing of the
11023 more complex types. The string types like challengePassword can
11024 be handled by the string table functions.
11026 Also modified the multi byte string table handling. Now there is
11027 a 'global mask' which masks out certain types. The table itself
11028 can use the flag STABLE_NO_MASK to ignore the mask setting: this
11029 is useful when for example there is only one permissible type
11030 (as in countryName) and using the mask might result in no valid
11034 *) Clean up 'Finished' handling, and add functions SSL_get_finished and
11035 SSL_get_peer_finished to allow applications to obtain the latest
11036 Finished messages sent to the peer or expected from the peer,
11037 respectively. (SSL_get_peer_finished is usually the Finished message
11038 actually received from the peer, otherwise the protocol will be aborted.)
11040 As the Finished message are message digests of the complete handshake
11041 (with a total of 192 bits for TLS 1.0 and more for SSL 3.0), they can
11042 be used for external authentication procedures when the authentication
11043 provided by SSL/TLS is not desired or is not enough.
11046 *) Enhanced support for Alpha Linux is added. Now ./config checks if
11047 the host supports BWX extension and if Compaq C is present on the
11048 $PATH. Just exploiting of the BWX extension results in 20-30%
11049 performance kick for some algorithms, e.g. DES and RC4 to mention
11050 a couple. Compaq C in turn generates ~20% faster code for MD5 and
11054 *) Add support for MS "fast SGC". This is arguably a violation of the
11055 SSL3/TLS protocol. Netscape SGC does two handshakes: the first with
11056 weak crypto and after checking the certificate is SGC a second one
11057 with strong crypto. MS SGC stops the first handshake after receiving
11058 the server certificate message and sends a second client hello. Since
11059 a server will typically do all the time consuming operations before
11060 expecting any further messages from the client (server key exchange
11061 is the most expensive) there is little difference between the two.
11063 To get OpenSSL to support MS SGC we have to permit a second client
11064 hello message after we have sent server done. In addition we have to
11065 reset the MAC if we do get this second client hello.
11068 *) Add a function 'd2i_AutoPrivateKey()' this will automatically decide
11069 if a DER encoded private key is RSA or DSA traditional format. Changed
11070 d2i_PrivateKey_bio() to use it. This is only needed for the "traditional"
11071 format DER encoded private key. Newer code should use PKCS#8 format which
11072 has the key type encoded in the ASN1 structure. Added DER private key
11073 support to pkcs8 application.
11076 *) SSL 3/TLS 1 servers now don't request certificates when an anonymous
11077 ciphersuites has been selected (as required by the SSL 3/TLS 1
11078 specifications). Exception: When SSL_VERIFY_FAIL_IF_NO_PEER_CERT
11079 is set, we interpret this as a request to violate the specification
11080 (the worst that can happen is a handshake failure, and 'correct'
11081 behaviour would result in a handshake failure anyway).
11084 *) In SSL_CTX_add_session, take into account that there might be multiple
11085 SSL_SESSION structures with the same session ID (e.g. when two threads
11086 concurrently obtain them from an external cache).
11087 The internal cache can handle only one SSL_SESSION with a given ID,
11088 so if there's a conflict, we now throw out the old one to achieve
11092 *) Add OIDs for idea and blowfish in CBC mode. This will allow both
11093 to be used in PKCS#5 v2.0 and S/MIME. Also add checking to
11094 some routines that use cipher OIDs: some ciphers do not have OIDs
11095 defined and so they cannot be used for S/MIME and PKCS#5 v2.0 for
11099 *) Simplify the trust setting structure and code. Now we just have
11100 two sequences of OIDs for trusted and rejected settings. These will
11101 typically have values the same as the extended key usage extension
11102 and any application specific purposes.
11104 The trust checking code now has a default behaviour: it will just
11105 check for an object with the same NID as the passed id. Functions can
11106 be provided to override either the default behaviour or the behaviour
11107 for a given id. SSL client, server and email already have functions
11108 in place for compatibility: they check the NID and also return "trusted"
11109 if the certificate is self signed.
11112 *) Add d2i,i2d bio/fp functions for PrivateKey: these convert the
11113 traditional format into an EVP_PKEY structure.
11116 *) Add a password callback function PEM_cb() which either prompts for
11117 a password if usr_data is NULL or otherwise assumes it is a null
11118 terminated password. Allow passwords to be passed on command line
11119 environment or config files in a few more utilities.
11122 *) Add a bunch of DER and PEM functions to handle PKCS#8 format private
11123 keys. Add some short names for PKCS#8 PBE algorithms and allow them
11124 to be specified on the command line for the pkcs8 and pkcs12 utilities.
11125 Update documentation.
11128 *) Support for ASN1 "NULL" type. This could be handled before by using
11129 ASN1_TYPE but there wasn't any function that would try to read a NULL
11130 and produce an error if it couldn't. For compatibility we also have
11131 ASN1_NULL_new() and ASN1_NULL_free() functions but these are faked and
11132 don't allocate anything because they don't need to.
11135 *) Initial support for MacOS is now provided. Examine INSTALL.MacOS
11137 [Andy Polyakov, Roy Woods <roy@centicsystems.ca>]
11139 *) Rebuild of the memory allocation routines used by OpenSSL code and
11140 possibly others as well. The purpose is to make an interface that
11141 provide hooks so anyone can build a separate set of allocation and
11142 deallocation routines to be used by OpenSSL, for example memory
11143 pool implementations, or something else, which was previously hard
11144 since Malloc(), Realloc() and Free() were defined as macros having
11145 the values malloc, realloc and free, respectively (except for Win32
11146 compilations). The same is provided for memory debugging code.
11147 OpenSSL already comes with functionality to find memory leaks, but
11148 this gives people a chance to debug other memory problems.
11150 With these changes, a new set of functions and macros have appeared:
11152 CRYPTO_set_mem_debug_functions() [F]
11153 CRYPTO_get_mem_debug_functions() [F]
11154 CRYPTO_dbg_set_options() [F]
11155 CRYPTO_dbg_get_options() [F]
11156 CRYPTO_malloc_debug_init() [M]
11158 The memory debug functions are NULL by default, unless the library
11159 is compiled with CRYPTO_MDEBUG or friends is defined. If someone
11160 wants to debug memory anyway, CRYPTO_malloc_debug_init() (which
11161 gives the standard debugging functions that come with OpenSSL) or
11162 CRYPTO_set_mem_debug_functions() (tells OpenSSL to use functions
11163 provided by the library user) must be used. When the standard
11164 debugging functions are used, CRYPTO_dbg_set_options can be used to
11165 request additional information:
11166 CRYPTO_dbg_set_options(V_CYRPTO_MDEBUG_xxx) corresponds to setting
11167 the CRYPTO_MDEBUG_xxx macro when compiling the library.
11169 Also, things like CRYPTO_set_mem_functions will always give the
11170 expected result (the new set of functions is used for allocation
11171 and deallocation) at all times, regardless of platform and compiler
11174 To finish it up, some functions that were never use in any other
11175 way than through macros have a new API and new semantic:
11177 CRYPTO_dbg_malloc()
11178 CRYPTO_dbg_realloc()
11181 All macros of value have retained their old syntax.
11182 [Richard Levitte and Bodo Moeller]
11184 *) Some S/MIME fixes. The OID for SMIMECapabilities was wrong, the
11185 ordering of SMIMECapabilities wasn't in "strength order" and there
11186 was a missing NULL in the AlgorithmIdentifier for the SHA1 signature
11190 *) Some ASN1 types with illegal zero length encoding (INTEGER,
11191 ENUMERATED and OBJECT IDENTIFIER) choked the ASN1 routines.
11192 [Frans Heymans <fheymans@isaserver.be>, modified by Steve Henson]
11194 *) Merge in my S/MIME library for OpenSSL. This provides a simple
11195 S/MIME API on top of the PKCS#7 code, a MIME parser (with enough
11196 functionality to handle multipart/signed properly) and a utility
11197 called 'smime' to call all this stuff. This is based on code I
11198 originally wrote for Celo who have kindly allowed it to be
11199 included in OpenSSL.
11202 *) Add variants des_set_key_checked and des_set_key_unchecked of
11203 des_set_key (aka des_key_sched). Global variable des_check_key
11204 decides which of these is called by des_set_key; this way
11205 des_check_key behaves as it always did, but applications and
11206 the library itself, which was buggy for des_check_key == 1,
11207 have a cleaner way to pick the version they need.
11210 *) New function PKCS12_newpass() which changes the password of a
11214 *) Modify X509_TRUST and X509_PURPOSE so it also uses a static and
11215 dynamic mix. In both cases the ids can be used as an index into the
11216 table. Also modified the X509_TRUST_add() and X509_PURPOSE_add()
11217 functions so they accept a list of the field values and the
11218 application doesn't need to directly manipulate the X509_TRUST
11222 *) Modify the ASN1_STRING_TABLE stuff so it also uses bsearch and doesn't
11226 *) Modify the way the V3 extension code looks up extensions. This now
11227 works in a similar way to the object code: we have some "standard"
11228 extensions in a static table which is searched with OBJ_bsearch()
11229 and the application can add dynamic ones if needed. The file
11230 crypto/x509v3/ext_dat.h now has the info: this file needs to be
11231 updated whenever a new extension is added to the core code and kept
11232 in ext_nid order. There is a simple program 'tabtest.c' which checks
11233 this. New extensions are not added too often so this file can readily
11234 be maintained manually.
11236 There are two big advantages in doing things this way. The extensions
11237 can be looked up immediately and no longer need to be "added" using
11238 X509V3_add_standard_extensions(): this function now does nothing.
11239 [Side note: I get *lots* of email saying the extension code doesn't
11240 work because people forget to call this function]
11241 Also no dynamic allocation is done unless new extensions are added:
11242 so if we don't add custom extensions there is no need to call
11243 X509V3_EXT_cleanup().
11246 *) Modify enc utility's salting as follows: make salting the default. Add a
11247 magic header, so unsalted files fail gracefully instead of just decrypting
11248 to garbage. This is because not salting is a big security hole, so people
11249 should be discouraged from doing it.
11252 *) Fixes and enhancements to the 'x509' utility. It allowed a message
11253 digest to be passed on the command line but it only used this
11254 parameter when signing a certificate. Modified so all relevant
11255 operations are affected by the digest parameter including the
11256 -fingerprint and -x509toreq options. Also -x509toreq choked if a
11257 DSA key was used because it didn't fix the digest.
11260 *) Initial certificate chain verify code. Currently tests the untrusted
11261 certificates for consistency with the verify purpose (which is set
11262 when the X509_STORE_CTX structure is set up) and checks the pathlength.
11264 There is a NO_CHAIN_VERIFY compilation option to keep the old behaviour:
11265 this is because it will reject chains with invalid extensions whereas
11266 every previous version of OpenSSL and SSLeay made no checks at all.
11268 Trust code: checks the root CA for the relevant trust settings. Trust
11269 settings have an initial value consistent with the verify purpose: e.g.
11270 if the verify purpose is for SSL client use it expects the CA to be
11271 trusted for SSL client use. However the default value can be changed to
11272 permit custom trust settings: one example of this would be to only trust
11273 certificates from a specific "secure" set of CAs.
11275 Also added X509_STORE_CTX_new() and X509_STORE_CTX_free() functions
11276 which should be used for version portability: especially since the
11277 verify structure is likely to change more often now.
11279 SSL integration. Add purpose and trust to SSL_CTX and SSL and functions
11280 to set them. If not set then assume SSL clients will verify SSL servers
11283 Two new options to the verify program: -untrusted allows a set of
11284 untrusted certificates to be passed in and -purpose which sets the
11285 intended purpose of the certificate. If a purpose is set then the
11286 new chain verify code is used to check extension consistency.
11289 *) Support for the authority information access extension.
11292 *) Modify RSA and DSA PEM read routines to transparently handle
11293 PKCS#8 format private keys. New *_PUBKEY_* functions that handle
11294 public keys in a format compatible with certificate
11295 SubjectPublicKeyInfo structures. Unfortunately there were already
11296 functions called *_PublicKey_* which used various odd formats so
11297 these are retained for compatibility: however the DSA variants were
11298 never in a public release so they have been deleted. Changed dsa/rsa
11299 utilities to handle the new format: note no releases ever handled public
11300 keys so we should be OK.
11302 The primary motivation for this change is to avoid the same fiasco
11303 that dogs private keys: there are several incompatible private key
11304 formats some of which are standard and some OpenSSL specific and
11305 require various evil hacks to allow partial transparent handling and
11306 even then it doesn't work with DER formats. Given the option anything
11307 other than PKCS#8 should be dumped: but the other formats have to
11308 stay in the name of compatibility.
11310 With public keys and the benefit of hindsight one standard format
11311 is used which works with EVP_PKEY, RSA or DSA structures: though
11312 it clearly returns an error if you try to read the wrong kind of key.
11314 Added a -pubkey option to the 'x509' utility to output the public key.
11315 Also rename the EVP_PKEY_get_*() to EVP_PKEY_rget_*()
11316 (renamed to EVP_PKEY_get1_*() in the OpenSSL 0.9.5 release) and add
11317 EVP_PKEY_rset_*() functions (renamed to EVP_PKEY_set1_*())
11318 that do the same as the EVP_PKEY_assign_*() except they up the
11319 reference count of the added key (they don't "swallow" the
11323 *) Fixes to crypto/x509/by_file.c the code to read in certificates and
11324 CRLs would fail if the file contained no certificates or no CRLs:
11325 added a new function to read in both types and return the number
11326 read: this means that if none are read it will be an error. The
11327 DER versions of the certificate and CRL reader would always fail
11328 because it isn't possible to mix certificates and CRLs in DER format
11329 without choking one or the other routine. Changed this to just read
11330 a certificate: this is the best we can do. Also modified the code
11331 in apps/verify.c to take notice of return codes: it was previously
11332 attempting to read in certificates from NULL pointers and ignoring
11333 any errors: this is one reason why the cert and CRL reader seemed
11334 to work. It doesn't check return codes from the default certificate
11335 routines: these may well fail if the certificates aren't installed.
11338 *) Code to support otherName option in GeneralName.
11341 *) First update to verify code. Change the verify utility
11342 so it warns if it is passed a self signed certificate:
11343 for consistency with the normal behaviour. X509_verify
11344 has been modified to it will now verify a self signed
11345 certificate if *exactly* the same certificate appears
11346 in the store: it was previously impossible to trust a
11347 single self signed certificate. This means that:
11348 openssl verify ss.pem
11349 now gives a warning about a self signed certificate but
11350 openssl verify -CAfile ss.pem ss.pem
11354 *) For servers, store verify_result in SSL_SESSION data structure
11355 (and add it to external session representation).
11356 This is needed when client certificate verifications fails,
11357 but an application-provided verification callback (set by
11358 SSL_CTX_set_cert_verify_callback) allows accepting the session
11359 anyway (i.e. leaves x509_store_ctx->error != X509_V_OK
11360 but returns 1): When the session is reused, we have to set
11361 ssl->verify_result to the appropriate error code to avoid
11363 [Bodo Moeller, problem pointed out by Lutz Jaenicke]
11365 *) Fix a bug in the new PKCS#7 code: it didn't consider the
11366 case in PKCS7_dataInit() where the signed PKCS7 structure
11367 didn't contain any existing data because it was being created.
11368 [Po-Cheng Chen <pocheng@nst.com.tw>, slightly modified by Steve Henson]
11370 *) Add a salt to the key derivation routines in enc.c. This
11371 forms the first 8 bytes of the encrypted file. Also add a
11372 -S option to allow a salt to be input on the command line.
11375 *) New function X509_cmp(). Oddly enough there wasn't a function
11376 to compare two certificates. We do this by working out the SHA1
11377 hash and comparing that. X509_cmp() will be needed by the trust
11381 *) SSL_get1_session() is like SSL_get_session(), but increments
11382 the reference count in the SSL_SESSION returned.
11383 [Geoff Thorpe <geoff@eu.c2.net>]
11385 *) Fix for 'req': it was adding a null to request attributes.
11386 Also change the X509_LOOKUP and X509_INFO code to handle
11387 certificate auxiliary information.
11390 *) Add support for 40 and 64 bit RC2 and RC4 algorithms: document
11394 *) Add the possibility to add extra information to the memory leak
11395 detecting output, to form tracebacks, showing from where each
11396 allocation was originated: CRYPTO_push_info("constant string") adds
11397 the string plus current file name and line number to a per-thread
11398 stack, CRYPTO_pop_info() does the obvious, CRYPTO_remove_all_info()
11399 is like calling CYRPTO_pop_info() until the stack is empty.
11400 Also updated memory leak detection code to be multi-thread-safe.
11403 *) Add options -text and -noout to pkcs7 utility and delete the
11404 encryption options which never did anything. Update docs.
11407 *) Add options to some of the utilities to allow the pass phrase
11408 to be included on either the command line (not recommended on
11409 OSes like Unix) or read from the environment. Update the
11410 manpages and fix a few bugs.
11413 *) Add a few manpages for some of the openssl commands.
11416 *) Fix the -revoke option in ca. It was freeing up memory twice,
11417 leaking and not finding already revoked certificates.
11420 *) Extensive changes to support certificate auxiliary information.
11421 This involves the use of X509_CERT_AUX structure and X509_AUX
11422 functions. An X509_AUX function such as PEM_read_X509_AUX()
11423 can still read in a certificate file in the usual way but it
11424 will also read in any additional "auxiliary information". By
11425 doing things this way a fair degree of compatibility can be
11426 retained: existing certificates can have this information added
11427 using the new 'x509' options.
11429 Current auxiliary information includes an "alias" and some trust
11430 settings. The trust settings will ultimately be used in enhanced
11431 certificate chain verification routines: currently a certificate
11432 can only be trusted if it is self signed and then it is trusted
11436 *) Fix assembler for Alpha (tested only on DEC OSF not Linux or *BSD).
11437 The problem was that one of the replacement routines had not been working
11438 since SSLeay releases. For now the offending routine has been replaced
11439 with non-optimised assembler. Even so, this now gives around 95%
11440 performance improvement for 1024 bit RSA signs.
11443 *) Hack to fix PKCS#7 decryption when used with some unorthodox RC2
11444 handling. Most clients have the effective key size in bits equal to
11445 the key length in bits: so a 40 bit RC2 key uses a 40 bit (5 byte) key.
11446 A few however don't do this and instead use the size of the decrypted key
11447 to determine the RC2 key length and the AlgorithmIdentifier to determine
11448 the effective key length. In this case the effective key length can still
11449 be 40 bits but the key length can be 168 bits for example. This is fixed
11450 by manually forcing an RC2 key into the EVP_PKEY structure because the
11451 EVP code can't currently handle unusual RC2 key sizes: it always assumes
11452 the key length and effective key length are equal.
11455 *) Add a bunch of functions that should simplify the creation of
11456 X509_NAME structures. Now you should be able to do:
11457 X509_NAME_add_entry_by_txt(nm, "CN", MBSTRING_ASC, "Steve", -1, -1, 0);
11458 and have it automatically work out the correct field type and fill in
11459 the structures. The more adventurous can try:
11460 X509_NAME_add_entry_by_txt(nm, field, MBSTRING_UTF8, str, -1, -1, 0);
11461 and it will (hopefully) work out the correct multibyte encoding.
11464 *) Change the 'req' utility to use the new field handling and multibyte
11465 copy routines. Before the DN field creation was handled in an ad hoc
11466 way in req, ca, and x509 which was rather broken and didn't support
11467 BMPStrings or UTF8Strings. Since some software doesn't implement
11468 BMPStrings or UTF8Strings yet, they can be enabled using the config file
11469 using the dirstring_type option. See the new comment in the default
11470 openssl.cnf for more info.
11473 *) Make crypto/rand/md_rand.c more robust:
11474 - Assure unique random numbers after fork().
11475 - Make sure that concurrent threads access the global counter and
11476 md serializably so that we never lose entropy in them
11477 or use exactly the same state in multiple threads.
11478 Access to the large state is not always serializable because
11479 the additional locking could be a performance killer, and
11480 md should be large enough anyway.
11483 *) New file apps/app_rand.c with commonly needed functionality
11484 for handling the random seed file.
11486 Use the random seed file in some applications that previously did not:
11488 dsaparam -genkey (which also ignored its '-rand' option),
11491 x509 (when signing).
11492 Except on systems with /dev/urandom, it is crucial to have a random
11493 seed file at least for key creation, DSA signing, and for DH exchanges;
11494 for RSA signatures we could do without one.
11496 gendh and gendsa (unlike genrsa) used to read only the first byte
11497 of each file listed in the '-rand' option. The function as previously
11498 found in genrsa is now in app_rand.c and is used by all programs
11499 that support '-rand'.
11502 *) In RAND_write_file, use mode 0600 for creating files;
11503 don't just chmod when it may be too late.
11506 *) Report an error from X509_STORE_load_locations
11507 when X509_LOOKUP_load_file or X509_LOOKUP_add_dir failed.
11510 *) New function ASN1_mbstring_copy() this copies a string in either
11511 ASCII, Unicode, Universal (4 bytes per character) or UTF8 format
11512 into an ASN1_STRING type. A mask of permissible types is passed
11513 and it chooses the "minimal" type to use or an error if not type
11517 *) Add function equivalents to the various macros in asn1.h. The old
11518 macros are retained with an M_ prefix. Code inside the library can
11519 use the M_ macros. External code (including the openssl utility)
11520 should *NOT* in order to be "shared library friendly".
11523 *) Add various functions that can check a certificate's extensions
11524 to see if it usable for various purposes such as SSL client,
11525 server or S/MIME and CAs of these types. This is currently
11526 VERY EXPERIMENTAL but will ultimately be used for certificate chain
11527 verification. Also added a -purpose flag to x509 utility to
11528 print out all the purposes.
11531 *) Add a CRYPTO_EX_DATA to X509 certificate structure and associated
11535 *) New X509V3_{X509,CRL,REVOKED}_get_d2i() functions. These will search
11536 for, obtain and decode and extension and obtain its critical flag.
11537 This allows all the necessary extension code to be handled in a
11538 single function call.
11541 *) RC4 tune-up featuring 30-40% performance improvement on most RISC
11542 platforms. See crypto/rc4/rc4_enc.c for further details.
11545 *) New -noout option to asn1parse. This causes no output to be produced
11546 its main use is when combined with -strparse and -out to extract data
11547 from a file (which may not be in ASN.1 format).
11550 *) Fix for pkcs12 program. It was hashing an invalid certificate pointer
11551 when producing the local key id.
11552 [Richard Levitte <levitte@stacken.kth.se>]
11554 *) New option -dhparam in s_server. This allows a DH parameter file to be
11555 stated explicitly. If it is not stated then it tries the first server
11556 certificate file. The previous behaviour hard coded the filename
11560 *) Add -pubin and -pubout options to the rsa and dsa commands. These allow
11561 a public key to be input or output. For example:
11562 openssl rsa -in key.pem -pubout -out pubkey.pem
11563 Also added necessary DSA public key functions to handle this.
11566 *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
11567 in the message. This was handled by allowing
11568 X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
11569 [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]
11571 *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
11572 to the end of the strings whereas this didn't. This would cause problems
11573 if strings read with d2i_ASN1_bytes() were later modified.
11574 [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]
11576 *) Fix for base64 decode bug. When a base64 bio reads only one line of
11577 data and it contains EOF it will end up returning an error. This is
11578 caused by input 46 bytes long. The cause is due to the way base64
11579 BIOs find the start of base64 encoded data. They do this by trying a
11580 trial decode on each line until they find one that works. When they
11581 do a flag is set and it starts again knowing it can pass all the
11582 data directly through the decoder. Unfortunately it doesn't reset
11583 the context it uses. This means that if EOF is reached an attempt
11584 is made to pass two EOFs through the context and this causes the
11585 resulting error. This can also cause other problems as well. As is
11586 usual with these problems it takes *ages* to find and the fix is
11587 trivial: move one line.
11588 [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]
11590 *) Ugly workaround to get s_client and s_server working under Windows. The
11591 old code wouldn't work because it needed to select() on sockets and the
11592 tty (for keypresses and to see if data could be written). Win32 only
11593 supports select() on sockets so we select() with a 1s timeout on the
11594 sockets and then see if any characters are waiting to be read, if none
11595 are present then we retry, we also assume we can always write data to
11596 the tty. This isn't nice because the code then blocks until we've
11597 received a complete line of data and it is effectively polling the
11598 keyboard at 1s intervals: however it's quite a bit better than not
11599 working at all :-) A dedicated Windows application might handle this
11600 with an event loop for example.
11603 *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
11604 and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
11605 will be called when RSA_sign() and RSA_verify() are used. This is useful
11606 if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
11607 For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
11608 should *not* be used: RSA_sign() and RSA_verify() must be used instead.
11609 This necessitated the support of an extra signature type NID_md5_sha1
11610 for SSL signatures and modifications to the SSL library to use it instead
11611 of calling RSA_public_decrypt() and RSA_private_encrypt().
11614 *) Add new -verify -CAfile and -CApath options to the crl program, these
11615 will lookup a CRL issuers certificate and verify the signature in a
11616 similar way to the verify program. Tidy up the crl program so it
11617 no longer accesses structures directly. Make the ASN1 CRL parsing a bit
11618 less strict. It will now permit CRL extensions even if it is not
11619 a V2 CRL: this will allow it to tolerate some broken CRLs.
11622 *) Initialize all non-automatic variables each time one of the openssl
11623 sub-programs is started (this is necessary as they may be started
11624 multiple times from the "OpenSSL>" prompt).
11625 [Lennart Bang, Bodo Moeller]
11627 *) Preliminary compilation option RSA_NULL which disables RSA crypto without
11628 removing all other RSA functionality (this is what NO_RSA does). This
11629 is so (for example) those in the US can disable those operations covered
11630 by the RSA patent while allowing storage and parsing of RSA keys and RSA
11634 *) Non-copying interface to BIO pairs.
11635 (still largely untested)
11638 *) New function ASN1_tag2str() to convert an ASN1 tag to a descriptive
11639 ASCII string. This was handled independently in various places before.
11642 *) New functions UTF8_getc() and UTF8_putc() that parse and generate
11643 UTF8 strings a character at a time.
11646 *) Use client_version from client hello to select the protocol
11647 (s23_srvr.c) and for RSA client key exchange verification
11648 (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
11651 *) Add various utility functions to handle SPKACs, these were previously
11652 handled by poking round in the structure internals. Added new function
11653 NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
11654 print, verify and generate SPKACs. Based on an original idea from
11655 Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
11658 *) RIPEMD160 is operational on all platforms and is back in 'make test'.
11661 *) Allow the config file extension section to be overwritten on the
11662 command line. Based on an original idea from Massimiliano Pala
11663 <madwolf@comune.modena.it>. The new option is called -extensions
11664 and can be applied to ca, req and x509. Also -reqexts to override
11665 the request extensions in req and -crlexts to override the crl extensions
11669 *) Add new feature to the SPKAC handling in ca. Now you can include
11670 the same field multiple times by preceding it by "XXXX." for example:
11673 this is the same syntax as used in the req config file.
11676 *) Allow certificate extensions to be added to certificate requests. These
11677 are specified in a 'req_extensions' option of the req section of the
11678 config file. They can be printed out with the -text option to req but
11679 are otherwise ignored at present.
11682 *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
11683 data read consists of only the final block it would not decrypted because
11684 EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
11685 A misplaced 'break' also meant the decrypted final block might not be
11686 copied until the next read.
11689 *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
11690 a few extra parameters to the DH structure: these will be useful if
11691 for example we want the value of 'q' or implement X9.42 DH.
11694 *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
11695 provides hooks that allow the default DSA functions or functions on a
11696 "per key" basis to be replaced. This allows hardware acceleration and
11697 hardware key storage to be handled without major modification to the
11698 library. Also added low level modexp hooks and CRYPTO_EX structure and
11699 associated functions.
11702 *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
11703 as "read only": it can't be written to and the buffer it points to will
11704 not be freed. Reading from a read only BIO is much more efficient than
11705 a normal memory BIO. This was added because there are several times when
11706 an area of memory needs to be read from a BIO. The previous method was
11707 to create a memory BIO and write the data to it, this results in two
11708 copies of the data and an O(n^2) reading algorithm. There is a new
11709 function BIO_new_mem_buf() which creates a read only memory BIO from
11710 an area of memory. Also modified the PKCS#7 routines to use read only
11714 *) Bugfix: ssl23_get_client_hello did not work properly when called in
11715 state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
11716 a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
11717 but a retry condition occurred while trying to read the rest.
11720 *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
11721 NID_pkcs7_encrypted by default: this was wrong since this should almost
11722 always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
11723 the encrypted data type: this is a more sensible place to put it and it
11724 allows the PKCS#12 code to be tidied up that duplicated this
11728 *) Changed obj_dat.pl script so it takes its input and output files on
11729 the command line. This should avoid shell escape redirection problems
11733 *) Initial support for certificate extension requests, these are included
11734 in things like Xenroll certificate requests. Included functions to allow
11735 extensions to be obtained and added.
11738 *) -crlf option to s_client and s_server for sending newlines as
11739 CRLF (as required by many protocols).
11742 Changes between 0.9.3a and 0.9.4 [09 Aug 1999]
11744 *) Install libRSAglue.a when OpenSSL is built with RSAref.
11745 [Ralf S. Engelschall]
11747 *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
11748 [Andrija Antonijevic <TheAntony2@bigfoot.com>]
11750 *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
11754 *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
11755 DH parameters/keys (q is lost during that conversion, but the resulting
11756 DH parameters contain its length).
11758 For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
11759 much faster than DH_generate_parameters (which creates parameters
11760 where p = 2*q + 1), and also the smaller q makes DH computations
11761 much more efficient (160-bit exponentiation instead of 1024-bit
11762 exponentiation); so this provides a convenient way to support DHE
11763 ciphersuites in SSL/TLS servers (see ssl/ssltest.c). It is of
11764 utter importance to use
11765 SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11767 SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
11768 when such DH parameters are used, because otherwise small subgroup
11769 attacks may become possible!
11772 *) Avoid memory leak in i2d_DHparams.
11775 *) Allow the -k option to be used more than once in the enc program:
11776 this allows the same encrypted message to be read by multiple recipients.
11779 *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
11780 an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
11781 it will always use the numerical form of the OID, even if it has a short
11785 *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
11786 method only got called if p,q,dmp1,dmq1,iqmp components were present,
11787 otherwise bn_mod_exp was called. In the case of hardware keys for example
11788 no private key components need be present and it might store extra data
11789 in the RSA structure, which cannot be accessed from bn_mod_exp.
11790 By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
11791 private key operations.
11794 *) Added support for SPARC Linux.
11797 *) pem_password_cb function type incompatibly changed from
11798 typedef int pem_password_cb(char *buf, int size, int rwflag);
11800 ....(char *buf, int size, int rwflag, void *userdata);
11801 so that applications can pass data to their callbacks:
11802 The PEM[_ASN1]_{read,write}... functions and macros now take an
11803 additional void * argument, which is just handed through whenever
11804 the password callback is called.
11805 [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]
11807 New function SSL_CTX_set_default_passwd_cb_userdata.
11809 Compatibility note: As many C implementations push function arguments
11810 onto the stack in reverse order, the new library version is likely to
11811 interoperate with programs that have been compiled with the old
11812 pem_password_cb definition (PEM_whatever takes some data that
11813 happens to be on the stack as its last argument, and the callback
11814 just ignores this garbage); but there is no guarantee whatsoever that
11817 *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
11818 (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
11819 problems not only on Windows, but also on some Unix platforms.
11820 To avoid problematic command lines, these definitions are now in an
11821 auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
11822 for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
11825 *) MIPS III/IV assembler module is reimplemented.
11828 *) More DES library cleanups: remove references to srand/rand and
11829 delete an unused file.
11832 *) Add support for the free Netwide assembler (NASM) under Win32,
11833 since not many people have MASM (ml) and it can be hard to obtain.
11834 This is currently experimental but it seems to work OK and pass all
11835 the tests. Check out INSTALL.W32 for info.
11838 *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
11839 without temporary keys kept an extra copy of the server key,
11840 and connections with temporary keys did not free everything in case
11844 *) New function RSA_check_key and new openssl rsa option -check
11845 for verifying the consistency of RSA keys.
11846 [Ulf Moeller, Bodo Moeller]
11848 *) Various changes to make Win32 compile work:
11849 1. Casts to avoid "loss of data" warnings in p5_crpt2.c
11850 2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
11851 comparison" warnings.
11852 3. Add sk_<TYPE>_sort to DEF file generator and do make update.
11855 *) Add a debugging option to PKCS#5 v2 key generation function: when
11856 you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
11857 derived keys are printed to stderr.
11860 *) Copy the flags in ASN1_STRING_dup().
11861 [Roman E. Pavlov <pre@mo.msk.ru>]
11863 *) The x509 application mishandled signing requests containing DSA
11864 keys when the signing key was also DSA and the parameters didn't match.
11866 It was supposed to omit the parameters when they matched the signing key:
11867 the verifying software was then supposed to automatically use the CA's
11868 parameters if they were absent from the end user certificate.
11870 Omitting parameters is no longer recommended. The test was also
11871 the wrong way round! This was probably due to unusual behaviour in
11872 EVP_cmp_parameters() which returns 1 if the parameters match.
11873 This meant that parameters were omitted when they *didn't* match and
11874 the certificate was useless. Certificates signed with 'ca' didn't have
11876 [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]
11878 *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
11879 The interface is as follows:
11880 Applications can use
11881 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
11882 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
11883 "off" is now the default.
11884 The library internally uses
11885 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
11886 CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
11887 to disable memory-checking temporarily.
11889 Some inconsistent states that previously were possible (and were
11890 even the default) are now avoided.
11892 -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
11893 with each memory chunk allocated; this is occasionally more helpful
11894 than just having a counter.
11896 -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.
11898 -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
11902 *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
11903 which largely parallels "options", but is for changing API behaviour,
11904 whereas "options" are about protocol behaviour.
11905 Initial "mode" flags are:
11907 SSL_MODE_ENABLE_PARTIAL_WRITE Allow SSL_write to report success when
11908 a single record has been written.
11909 SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER Don't insist that SSL_write
11910 retries use the same buffer location.
11911 (But all of the contents must be
11915 *) Bugfix: SSL_set_options ignored its parameter, only SSL_CTX_set_options
11918 *) Fix problems with no-hmac etc.
11919 [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]
11921 *) New functions RSA_get_default_method(), RSA_set_method() and
11922 RSA_get_method(). These allows replacement of RSA_METHODs without having
11923 to mess around with the internals of an RSA structure.
11926 *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
11927 Also really enable memory leak checks in openssl.c and in some
11929 [Chad C. Mulligan, Bodo Moeller]
11931 *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
11932 up the length of negative integers. This has now been simplified to just
11933 store the length when it is first determined and use it later, rather
11934 than trying to keep track of where data is copied and updating it to
11936 [Steve Henson, reported by Brien Wheeler
11937 <bwheeler@authentica-security.com>]
11939 *) Add a new function PKCS7_signatureVerify. This allows the verification
11940 of a PKCS#7 signature but with the signing certificate passed to the
11941 function itself. This contrasts with PKCS7_dataVerify which assumes the
11942 certificate is present in the PKCS#7 structure. This isn't always the
11943 case: certificates can be omitted from a PKCS#7 structure and be
11944 distributed by "out of band" means (such as a certificate database).
11947 *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
11948 function prototypes in pem.h, also change util/mkdef.pl to add the
11949 necessary function names.
11952 *) mk1mf.pl (used by Windows builds) did not properly read the
11953 options set by Configure in the top level Makefile, and Configure
11954 was not even able to write more than one option correctly.
11955 Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
11958 *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
11959 file to be loaded from a BIO or FILE pointer. The BIO version will
11960 for example allow memory BIOs to contain config info.
11963 *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
11964 Whoever hopes to achieve shared-library compatibility across versions
11965 must use this, not the compile-time macro.
11966 (Exercise 0.9.4: Which is the minimum library version required by
11968 Note: All this applies only to multi-threaded programs, others don't
11972 *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
11973 through a BIO pair triggered the default case, i.e.
11974 SSLerr(...,SSL_R_UNKNOWN_STATE).
11977 *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
11978 can use the SSL library even if none of the specific BIOs is
11982 *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
11983 for the encoded length.
11984 [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]
11986 *) Add initial documentation of the X509V3 functions.
11989 *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and
11990 PEM_write_bio_PKCS8PrivateKey() that are equivalent to
11991 PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
11992 secure PKCS#8 private key format with a high iteration count.
11995 *) Fix determination of Perl interpreter: A perl or perl5
11996 _directory_ in $PATH was also accepted as the interpreter.
11997 [Ralf S. Engelschall]
11999 *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
12000 wrong with it but it was very old and did things like calling
12001 PEM_ASN1_read() directly and used MD5 for the hash not to mention some
12002 unusual formatting.
12005 *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
12006 to use the new extension code.
12009 *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
12010 with macros. This should make it easier to change their form, add extra
12011 arguments etc. Fix a few PEM prototypes which didn't have cipher as a
12015 *) Add to configuration table a new entry that can specify an alternative
12016 name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
12017 according to Mark Crispin <MRC@Panda.COM>.
12021 *) DES CBC did not update the IV. Weird.
12024 des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
12025 Changing the behaviour of the former might break existing programs --
12026 where IV updating is needed, des_ncbc_encrypt can be used.
12029 *) When bntest is run from "make test" it drives bc to check its
12030 calculations, as well as internally checking them. If an internal check
12031 fails, it needs to cause bc to give a non-zero result or make test carries
12032 on without noticing the failure. Fixed.
12035 *) DES library cleanups.
12038 *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
12039 used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
12040 ciphers. NOTE: although the key derivation function has been verified
12041 against some published test vectors it has not been extensively tested
12042 yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
12046 *) Instead of "mkdir -p", which is not fully portable, use new
12047 Perl script "util/mkdir-p.pl".
12050 *) Rewrite the way password based encryption (PBE) is handled. It used to
12051 assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
12052 structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
12053 but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
12054 the 'parameter' field of the AlgorithmIdentifier is passed to the
12055 underlying key generation function so it must do its own ASN1 parsing.
12056 This has also changed the EVP_PBE_CipherInit() function which now has a
12057 'parameter' argument instead of literal salt and iteration count values
12058 and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
12061 *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
12062 and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
12063 Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
12064 KEY" because this clashed with PKCS#8 unencrypted string. Since this
12065 value was just used as a "magic string" and not used directly its
12066 value doesn't matter.
12069 *) Introduce some semblance of const correctness to BN. Shame C doesn't
12073 *) "linux-sparc64" configuration (ultrapenguin).
12074 [Ray Miller <ray.miller@oucs.ox.ac.uk>]
12075 "linux-sparc" configuration.
12076 [Christian Forster <fo@hawo.stw.uni-erlangen.de>]
12078 *) config now generates no-xxx options for missing ciphers.
12081 *) Support the EBCDIC character set (work in progress).
12082 File ebcdic.c not yet included because it has a different license.
12083 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12085 *) Support BS2000/OSD-POSIX.
12086 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]
12088 *) Make callbacks for key generation use void * instead of char *.
12091 *) Make S/MIME samples compile (not yet tested).
12094 *) Additional typesafe stacks.
12097 *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
12101 Changes between 0.9.3 and 0.9.3a [29 May 1999]
12103 *) New configuration variant "sco5-gcc".
12105 *) Updated some demos.
12106 [Sean O Riordain, Wade Scholine]
12108 *) Add missing BIO_free at exit of pkcs12 application.
12111 *) Fix memory leak in conf.c.
12114 *) Updates for Win32 to assembler version of MD5.
12117 *) Set #! path to perl in apps/der_chop to where we found it
12118 instead of using a fixed path.
12121 *) SHA library changes for irix64-mips4-cc.
12124 *) Improvements for VMS support.
12128 Changes between 0.9.2b and 0.9.3 [24 May 1999]
12130 *) Bignum library bug fix. IRIX 6 passes "make test" now!
12131 This also avoids the problems with SC4.2 and unpatched SC5.
12132 [Andy Polyakov <appro@fy.chalmers.se>]
12134 *) New functions sk_num, sk_value and sk_set to replace the previous macros.
12135 These are required because of the typesafe stack would otherwise break
12136 existing code. If old code used a structure member which used to be STACK
12137 and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
12138 sk_num or sk_value it would produce an error because the num, data members
12139 are not present in STACK_OF. Now it just produces a warning. sk_set
12140 replaces the old method of assigning a value to sk_value
12141 (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
12142 that does this will no longer work (and should use sk_set instead) but
12143 this could be regarded as a "questionable" behaviour anyway.
12146 *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
12147 correctly handle encrypted S/MIME data.
12150 *) Change type of various DES function arguments from des_cblock
12151 (which means, in function argument declarations, pointer to char)
12152 to des_cblock * (meaning pointer to array with 8 char elements),
12153 which allows the compiler to do more typechecking; it was like
12154 that back in SSLeay, but with lots of ugly casts.
12156 Introduce new type const_des_cblock.
12159 *) Reorganise the PKCS#7 library and get rid of some of the more obvious
12160 problems: find RecipientInfo structure that matches recipient certificate
12161 and initialise the ASN1 structures properly based on passed cipher.
12164 *) Belatedly make the BN tests actually check the results.
12167 *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
12168 to and from BNs: it was completely broken. New compilation option
12169 NEG_PUBKEY_BUG to allow for some broken certificates that encode public
12170 key elements as negative integers.
12173 *) Reorganize and speed up MD5.
12174 [Andy Polyakov <appro@fy.chalmers.se>]
12177 [Richard Levitte <richard@levitte.org>]
12179 *) New option -out to asn1parse to allow the parsed structure to be
12180 output to a file. This is most useful when combined with the -strparse
12181 option to examine the output of things like OCTET STRINGS.
12184 *) Make SSL library a little more fool-proof by not requiring any longer
12185 that SSL_set_{accept,connect}_state be called before
12186 SSL_{accept,connect} may be used (SSL_set_..._state is omitted
12187 in many applications because usually everything *appeared* to work as
12188 intended anyway -- now it really works as intended).
12191 *) Move openssl.cnf out of lib/.
12194 *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
12195 -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
12196 -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+
12197 [Ralf S. Engelschall]
12199 *) Various fixes to the EVP and PKCS#7 code. It may now be able to
12200 handle PKCS#7 enveloped data properly.
12201 [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]
12203 *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
12204 copying pointers. The cert_st handling is changed by this in
12205 various ways (and thus what used to be known as ctx->default_cert
12206 is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
12207 any longer when s->cert does not give us what we need).
12208 ssl_cert_instantiate becomes obsolete by this change.
12209 As soon as we've got the new code right (possibly it already is?),
12210 we have solved a couple of bugs of the earlier code where s->cert
12211 was used as if it could not have been shared with other SSL structures.
12213 Note that using the SSL API in certain dirty ways now will result
12214 in different behaviour than observed with earlier library versions:
12215 Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
12216 does not influence s as it used to.
12218 In order to clean up things more thoroughly, inside SSL_SESSION
12219 we don't use CERT any longer, but a new structure SESS_CERT
12220 that holds per-session data (if available); currently, this is
12221 the peer's certificate chain and, for clients, the server's certificate
12222 and temporary key. CERT holds only those values that can have
12223 meaningful defaults in an SSL_CTX.
12226 *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
12227 from the internal representation. Various PKCS#7 fixes: remove some
12228 evil casts and set the enc_dig_alg field properly based on the signing
12232 *) Allow PKCS#12 password to be set from the command line or the
12233 environment. Let 'ca' get its config file name from the environment
12234 variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
12238 *) Allow certificate policies extension to use an IA5STRING for the
12239 organization field. This is contrary to the PKIX definition but
12240 VeriSign uses it and IE5 only recognises this form. Document 'x509'
12244 *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
12245 without disallowing inline assembler and the like for non-pedantic builds.
12248 *) Support Borland C++ builder.
12249 [Janez Jere <jj@void.si>, modified by Ulf Möller]
12251 *) Support Mingw32.
12254 *) SHA-1 cleanups and performance enhancements.
12255 [Andy Polyakov <appro@fy.chalmers.se>]
12257 *) Sparc v8plus assembler for the bignum library.
12258 [Andy Polyakov <appro@fy.chalmers.se>]
12260 *) Accept any -xxx and +xxx compiler options in Configure.
12263 *) Update HPUX configuration.
12266 *) Add missing sk_<type>_unshift() function to safestack.h
12267 [Ralf S. Engelschall]
12269 *) New function SSL_CTX_use_certificate_chain_file that sets the
12270 "extra_cert"s in addition to the certificate. (This makes sense
12271 only for "PEM" format files, as chains as a whole are not
12275 *) Support verify_depth from the SSL API.
12276 x509_vfy.c had what can be considered an off-by-one-error:
12277 Its depth (which was not part of the external interface)
12278 was actually counting the number of certificates in a chain;
12279 now it really counts the depth.
12282 *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
12283 instead of X509err, which often resulted in confusing error
12284 messages since the error codes are not globally unique
12285 (e.g. an alleged error in ssl3_accept when a certificate
12286 didn't match the private key).
12288 *) New function SSL_CTX_set_session_id_context that allows to set a default
12289 value (so that you don't need SSL_set_session_id_context for each
12290 connection using the SSL_CTX).
12293 *) OAEP decoding bug fix.
12296 *) Support INSTALL_PREFIX for package builders, as proposed by
12300 *) New Configure options "threads" and "no-threads". For systems
12301 where the proper compiler options are known (currently Solaris
12302 and Linux), "threads" is the default.
12305 *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
12308 *) Install various scripts to $(OPENSSLDIR)/misc, not to
12309 $(INSTALLTOP)/bin -- they shouldn't clutter directories
12310 such as /usr/local/bin.
12313 *) "make linux-shared" to build shared libraries.
12314 [Niels Poppe <niels@netbox.org>]
12316 *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
12319 *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
12320 extension adding in x509 utility.
12323 *) Remove NOPROTO sections and error code comments.
12326 *) Partial rewrite of the DEF file generator to now parse the ANSI
12330 *) New Configure options --prefix=DIR and --openssldir=DIR.
12333 *) Complete rewrite of the error code script(s). It is all now handled
12334 by one script at the top level which handles error code gathering,
12335 header rewriting and C source file generation. It should be much better
12336 than the old method: it now uses a modified version of Ulf's parser to
12337 read the ANSI prototypes in all header files (thus the old K&R definitions
12338 aren't needed for error creation any more) and do a better job of
12339 translating function codes into names. The old 'ASN1 error code imbedded
12340 in a comment' is no longer necessary and it doesn't use .err files which
12341 have now been deleted. Also the error code call doesn't have to appear all
12342 on one line (which resulted in some large lines...).
12345 *) Change #include filenames from <foo.h> to <openssl/foo.h>.
12348 *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
12349 0 (which usually indicates a closed connection), but continue reading.
12352 *) Fix some race conditions.
12355 *) Add support for CRL distribution points extension. Add Certificate
12356 Policies and CRL distribution points documentation.
12359 *) Move the autogenerated header file parts to crypto/opensslconf.h.
12362 *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
12363 8 of keying material. Merlin has also confirmed interop with this fix
12364 between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
12365 [Merlin Hughes <merlin@baltimore.ie>]
12367 *) Fix lots of warnings.
12368 [Richard Levitte <levitte@stacken.kth.se>]
12370 *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
12371 the directory spec didn't end with a LIST_SEPARATOR_CHAR.
12372 [Richard Levitte <levitte@stacken.kth.se>]
12374 *) Fix problems with sizeof(long) == 8.
12375 [Andy Polyakov <appro@fy.chalmers.se>]
12377 *) Change functions to ANSI C.
12380 *) Fix typos in error codes.
12381 [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]
12383 *) Remove defunct assembler files from Configure.
12386 *) SPARC v8 assembler BIGNUM implementation.
12387 [Andy Polyakov <appro@fy.chalmers.se>]
12389 *) Support for Certificate Policies extension: both print and set.
12390 Various additions to support the r2i method this uses.
12393 *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
12394 return a const string when you are expecting an allocated buffer.
12397 *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
12398 types DirectoryString and DisplayText.
12401 *) Add code to allow r2i extensions to access the configuration database,
12402 add an LHASH database driver and add several ctx helper functions.
12405 *) Fix an evil bug in bn_expand2() which caused various BN functions to
12406 fail when they extended the size of a BIGNUM.
12409 *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
12410 support typesafe stack.
12413 *) Fix typo in SSL_[gs]et_options().
12414 [Nils Frostberg <nils@medcom.se>]
12416 *) Delete various functions and files that belonged to the (now obsolete)
12417 old X509V3 handling code.
12420 *) New Configure option "rsaref".
12423 *) Don't auto-generate pem.h.
12426 *) Introduce type-safe ASN.1 SETs.
12429 *) Convert various additional casted stacks to type-safe STACK_OF() variants.
12430 [Ben Laurie, Ralf S. Engelschall, Steve Henson]
12432 *) Introduce type-safe STACKs. This will almost certainly break lots of code
12433 that links with OpenSSL (well at least cause lots of warnings), but fear
12434 not: the conversion is trivial, and it eliminates loads of evil casts. A
12435 few STACKed things have been converted already. Feel free to convert more.
12436 In the fullness of time, I'll do away with the STACK type altogether.
12439 *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
12440 specified in <certfile> by updating the entry in the index.txt file.
12441 This way one no longer has to edit the index.txt file manually for
12442 revoking a certificate. The -revoke option does the gory details now.
12443 [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]
12445 *) Fix `openssl crl -noout -text' combination where `-noout' killed the
12446 `-text' option at all and this way the `-noout -text' combination was
12447 inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
12448 [Ralf S. Engelschall]
12450 *) Make sure a corresponding plain text error message exists for the
12451 X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
12452 verify callback function determined that a certificate was revoked.
12453 [Ralf S. Engelschall]
12455 *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
12456 ciphers that were excluded, e.g. by -DNO_IDEA. Also, test
12457 all available ciphers including rc5, which was forgotten until now.
12458 In order to let the testing shell script know which algorithms
12459 are available, a new (up to now undocumented) command
12460 "openssl list-cipher-commands" is used.
12463 *) Bugfix: s_client occasionally would sleep in select() when
12464 it should have checked SSL_pending() first.
12467 *) New functions DSA_do_sign and DSA_do_verify to provide access to
12468 the raw DSA values prior to ASN.1 encoding.
12471 *) Tweaks to Configure
12472 [Niels Poppe <niels@netbox.org>]
12474 *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
12478 *) New variables $(RANLIB) and $(PERL) in the Makefiles.
12481 *) New config option to avoid instructions that are illegal on the 80386.
12482 The default code is faster, but requires at least a 486.
12485 *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
12486 SSL2_SERVER_VERSION (not used at all) macros, which are now the
12487 same as SSL2_VERSION anyway.
12490 *) New "-showcerts" option for s_client.
12493 *) Still more PKCS#12 integration. Add pkcs12 application to openssl
12494 application. Various cleanups and fixes.
12497 *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
12498 modify error routines to work internally. Add error codes and PBE init
12499 to library startup routines.
12502 *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
12503 packing functions to asn1 and evp. Changed function names and error
12504 codes along the way.
12507 *) PKCS12 integration: and so it begins... First of several patches to
12508 slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
12509 objects to objects.h
12512 *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
12513 and display support for Thawte strong extranet extension.
12516 *) Add LinuxPPC support.
12517 [Jeff Dubrule <igor@pobox.org>]
12519 *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
12520 bn_div_words in alpha.s.
12521 [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]
12523 *) Make sure the RSA OAEP test is skipped under -DRSAref because
12524 OAEP isn't supported when OpenSSL is built with RSAref.
12525 [Ulf Moeller <ulf@fitug.de>]
12527 *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h
12528 so they no longer are missing under -DNOPROTO.
12529 [Soren S. Jorvang <soren@t.dk>]
12532 Changes between 0.9.1c and 0.9.2b [22 Mar 1999]
12534 *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
12535 doesn't work when the session is reused. Coming soon!
12538 *) Fix a security hole, that allows sessions to be reused in the wrong
12539 context thus bypassing client cert protection! All software that uses
12540 client certs and session caches in multiple contexts NEEDS PATCHING to
12541 allow session reuse! A fuller solution is in the works.
12542 [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]
12544 *) Some more source tree cleanups (removed obsolete files
12545 crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
12546 permission on "config" script to be executable) and a fix for the INSTALL
12548 [Ulf Moeller <ulf@fitug.de>]
12550 *) Remove some legacy and erroneous uses of malloc, free instead of
12552 [Lennart Bang <lob@netstream.se>, with minor changes by Steve]
12554 *) Make rsa_oaep_test return non-zero on error.
12555 [Ulf Moeller <ulf@fitug.de>]
12557 *) Add support for native Solaris shared libraries. Configure
12558 solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
12559 if someone would make that last step automatic.
12560 [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]
12562 *) ctx_size was not built with the right compiler during "make links". Fixed.
12565 *) Change the meaning of 'ALL' in the cipher list. It now means "everything
12566 except NULL ciphers". This means the default cipher list will no longer
12567 enable NULL ciphers. They need to be specifically enabled e.g. with
12568 the string "DEFAULT:eNULL".
12571 *) Fix to RSA private encryption routines: if p < q then it would
12572 occasionally produce an invalid result. This will only happen with
12573 externally generated keys because OpenSSL (and SSLeay) ensure p > q.
12576 *) Be less restrictive and allow also `perl util/perlpath.pl
12577 /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
12578 because this way one can also use an interpreter named `perl5' (which is
12579 usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
12580 installed as `perl').
12581 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12583 *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
12584 [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12586 *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
12587 advapi32.lib to Win32 build and change the pem test comparison
12588 to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
12589 suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
12590 and crypto/des/ede_cbcm_enc.c.
12593 *) DES quad checksum was broken on big-endian architectures. Fixed.
12596 *) Comment out two functions in bio.h that aren't implemented. Fix up the
12597 Win32 test batch file so it (might) work again. The Win32 test batch file
12598 is horrible: I feel ill....
12601 *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
12602 in e_os.h. Audit of header files to check ANSI and non ANSI
12603 sections: 10 functions were absent from non ANSI section and not exported
12604 from Windows DLLs. Fixed up libeay.num for new functions.
12607 *) Make `openssl version' output lines consistent.
12608 [Ralf S. Engelschall]
12610 *) Fix Win32 symbol export lists for BIO functions: Added
12611 BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
12612 to ms/libeay{16,32}.def.
12613 [Ralf S. Engelschall]
12615 *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
12616 fine under Unix and passes some trivial tests I've now added. But the
12617 whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
12618 added to make sure no one expects that this stuff really works in the
12619 OpenSSL 0.9.2 release. Additionally I've started to clean the XS sources
12620 up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
12622 [Ralf S. Engelschall]
12624 *) Fix the generation of two part addresses in perl.
12625 [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]
12627 *) Add config entry for Linux on MIPS.
12628 [John Tobey <jtobey@channel1.com>]
12630 *) Make links whenever Configure is run, unless we are on Windoze.
12633 *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
12634 Currently only issuerAltName and AuthorityKeyIdentifier make any sense
12638 *) Add a useful kludge to allow package maintainers to specify compiler and
12639 other platforms details on the command line without having to patch the
12640 Configure script everytime: One now can use ``perl Configure
12641 <id>:<details>'', i.e. platform ids are allowed to have details appended
12642 to them (separated by colons). This is treated as there would be a static
12643 pre-configured entry in Configure's %table under key <id> with value
12644 <details> and ``perl Configure <id>'' is called. So, when you want to
12645 perform a quick test-compile under FreeBSD 3.1 with pgcc and without
12646 assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
12647 now, which overrides the FreeBSD-elf entry on-the-fly.
12648 [Ralf S. Engelschall]
12650 *) Disable new TLS1 ciphersuites by default: they aren't official yet.
12653 *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
12654 on the `perl Configure ...' command line. This way one can compile
12655 OpenSSL libraries with Position Independent Code (PIC) which is needed
12656 for linking it into DSOs.
12657 [Ralf S. Engelschall]
12659 *) Remarkably, export ciphers were totally broken and no-one had noticed!
12663 *) Cleaned up the LICENSE document: The official contact for any license
12664 questions now is the OpenSSL core team under openssl-core@openssl.org.
12665 And add a paragraph about the dual-license situation to make sure people
12666 recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
12667 to the OpenSSL toolkit.
12668 [Ralf S. Engelschall]
12670 *) General source tree makefile cleanups: Made `making xxx in yyy...'
12671 display consistent in the source tree and replaced `/bin/rm' by `rm'.
12672 Additionally cleaned up the `make links' target: Remove unnecessary
12673 semicolons, subsequent redundant removes, inline point.sh into mklink.sh
12674 to speed processing and no longer clutter the display with confusing
12675 stuff. Instead only the actually done links are displayed.
12676 [Ralf S. Engelschall]
12678 *) Permit null encryption ciphersuites, used for authentication only. It used
12679 to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
12680 It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
12684 *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
12685 signed attributes when verifying signatures (this would break them),
12686 the detached data encoding was wrong and public keys obtained using
12687 X509_get_pubkey() weren't freed.
12690 *) Add text documentation for the BUFFER functions. Also added a work around
12691 to a Win95 console bug. This was triggered by the password read stuff: the
12692 last character typed gets carried over to the next fread(). If you were
12693 generating a new cert request using 'req' for example then the last
12694 character of the passphrase would be CR which would then enter the first
12698 *) Added the new `Includes OpenSSL Cryptography Software' button as
12699 doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
12700 button and can be used by applications based on OpenSSL to show the
12701 relationship to the OpenSSL project.
12702 [Ralf S. Engelschall]
12704 *) Remove confusing variables in function signatures in files
12705 ssl/ssl_lib.c and ssl/ssl.h.
12706 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12708 *) Don't install bss_file.c under PREFIX/include/
12709 [Lennart Bong <lob@kulthea.stacken.kth.se>]
12711 *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
12712 functions that return function pointers and has support for NT specific
12713 stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
12714 #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
12715 unsigned to signed types: this was killing the Win32 compile.
12718 *) Add new certificate file to stack functions,
12719 SSL_add_dir_cert_subjects_to_stack() and
12720 SSL_add_file_cert_subjects_to_stack(). These largely supplant
12721 SSL_load_client_CA_file(), and can be used to add multiple certs easily
12722 to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
12723 This means that Apache-SSL and similar packages don't have to mess around
12724 to add as many CAs as they want to the preferred list.
12727 *) Experiment with doxygen documentation. Currently only partially applied to
12729 See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
12730 openssl.doxy as the configuration file.
12733 *) Get rid of remaining C++-style comments which strict C compilers hate.
12734 [Ralf S. Engelschall, pointed out by Carlos Amengual]
12736 *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
12737 compiled in by default: it has problems with large keys.
12740 *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
12741 DH private keys and/or callback functions which directly correspond to
12742 their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
12743 is needed for applications which have to configure certificates on a
12744 per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
12746 For the RSA certificate situation is makes no difference, but
12747 for the DSA certificate situation this fixes the "no shared cipher"
12748 problem where the OpenSSL cipher selection procedure failed because the
12749 temporary keys were not overtaken from the context and the API provided
12750 no way to reconfigure them.
12751 The new functions now let applications reconfigure the stuff and they
12752 are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
12753 SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback. Additionally a new
12754 non-public-API function ssl_cert_instantiate() is used as a helper
12755 function and also to reduce code redundancy inside ssl_rsa.c.
12756 [Ralf S. Engelschall]
12758 *) Move s_server -dcert and -dkey options out of the undocumented feature
12759 area because they are useful for the DSA situation and should be
12760 recognized by the users.
12761 [Ralf S. Engelschall]
12763 *) Fix the cipher decision scheme for export ciphers: the export bits are
12764 *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
12765 SSL_EXP_MASK. So, the original variable has to be used instead of the
12766 already masked variable.
12767 [Richard Levitte <levitte@stacken.kth.se>]
12769 *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
12770 [Richard Levitte <levitte@stacken.kth.se>]
12772 *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
12773 from `int' to `unsigned int' because it's a length and initialized by
12774 EVP_DigestFinal() which expects an `unsigned int *'.
12775 [Richard Levitte <levitte@stacken.kth.se>]
12777 *) Don't hard-code path to Perl interpreter on shebang line of Configure
12778 script. Instead use the usual Shell->Perl transition trick.
12779 [Ralf S. Engelschall]
12781 *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
12782 (in addition to RSA certificates) to match the behaviour of `openssl dsa
12783 -noout -modulus' as it's already the case for `openssl rsa -noout
12784 -modulus'. For RSA the -modulus is the real "modulus" while for DSA
12785 currently the public key is printed (a decision which was already done by
12786 `openssl dsa -modulus' in the past) which serves a similar purpose.
12787 Additionally the NO_RSA no longer completely removes the whole -modulus
12788 option; it now only avoids using the RSA stuff. Same applies to NO_DSA
12790 [Ralf S. Engelschall]
12792 *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
12793 BIO. See the source (crypto/evp/bio_ok.c) for more info.
12794 [Arne Ansper <arne@ats.cyber.ee>]
12796 *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
12797 to be added. Now both 'req' and 'ca' can use new objects defined in the
12801 *) Add cool BIO that does syslog (or event log on NT).
12802 [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]
12804 *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
12805 TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
12806 TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
12807 Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
12810 *) Add preliminary config info for new extension code.
12813 *) Make RSA_NO_PADDING really use no padding.
12814 [Ulf Moeller <ulf@fitug.de>]
12816 *) Generate errors when private/public key check is done.
12819 *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
12820 for some CRL extensions and new objects added.
12823 *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
12824 key usage extension and fuller support for authority key id.
12827 *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
12828 padding method for RSA, which is recommended for new applications in PKCS
12829 #1 v2.0 (RFC 2437, October 1998).
12830 OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
12831 foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
12832 against Bleichbacher's attack on RSA.
12833 [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
12836 *) Updates to the new SSL compression code
12837 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12839 *) Fix so that the version number in the master secret, when passed
12840 via RSA, checks that if TLS was proposed, but we roll back to SSLv3
12841 (because the server will not accept higher), that the version number
12842 is 0x03,0x01, not 0x03,0x00
12843 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12845 *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
12846 leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
12847 in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
12850 *) Support for RAW extensions where an arbitrary extension can be
12851 created by including its DER encoding. See apps/openssl.cnf for
12855 *) Make sure latest Perl versions don't interpret some generated C array
12856 code as Perl array code in the crypto/err/err_genc.pl script.
12857 [Lars Weber <3weber@informatik.uni-hamburg.de>]
12859 *) Modify ms/do_ms.bat to not generate assembly language makefiles since
12860 not many people have the assembler. Various Win32 compilation fixes and
12861 update to the INSTALL.W32 file with (hopefully) more accurate Win32
12862 build instructions.
12865 *) Modify configure script 'Configure' to automatically create crypto/date.h
12866 file under Win32 and also build pem.h from pem.org. New script
12867 util/mkfiles.pl to create the MINFO file on environments that can't do a
12868 'make files': perl util/mkfiles.pl >MINFO should work.
12871 *) Major rework of DES function declarations, in the pursuit of correctness
12872 and purity. As a result, many evil casts evaporated, and some weirdness,
12873 too. You may find this causes warnings in your code. Zapping your evil
12874 casts will probably fix them. Mostly.
12877 *) Fix for a typo in asn1.h. Bug fix to object creation script
12878 obj_dat.pl. It considered a zero in an object definition to mean
12879 "end of object": none of the objects in objects.h have any zeros
12880 so it wasn't spotted.
12881 [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]
12883 *) Add support for Triple DES Cipher Block Chaining with Output Feedback
12884 Masking (CBCM). In the absence of test vectors, the best I have been able
12885 to do is check that the decrypt undoes the encrypt, so far. Send me test
12886 vectors if you have them.
12889 *) Correct calculation of key length for export ciphers (too much space was
12890 allocated for null ciphers). This has not been tested!
12893 *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
12894 message is now correct (it understands "crypto" and "ssl" on its
12895 command line). There is also now an "update" option. This will update
12896 the util/ssleay.num and util/libeay.num files with any new functions.
12898 perl util/mkdef.pl crypto ssl update
12899 it will update them.
12902 *) Overhauled the Perl interface (perl/*):
12903 - ported BN stuff to OpenSSL's different BN library
12904 - made the perl/ source tree CVS-aware
12905 - renamed the package from SSLeay to OpenSSL (the files still contain
12906 their history because I've copied them in the repository)
12907 - removed obsolete files (the test scripts will be replaced
12908 by better Test::Harness variants in the future)
12909 [Ralf S. Engelschall]
12911 *) First cut for a very conservative source tree cleanup:
12912 1. merge various obsolete readme texts into doc/ssleay.txt
12913 where we collect the old documents and readme texts.
12914 2. remove the first part of files where I'm already sure that we no
12915 longer need them because of three reasons: either they are just temporary
12916 files which were left by Eric or they are preserved original files where
12917 I've verified that the diff is also available in the CVS via "cvs diff
12918 -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
12919 the crypto/md/ stuff).
12920 [Ralf S. Engelschall]
12922 *) More extension code. Incomplete support for subject and issuer alt
12923 name, issuer and authority key id. Change the i2v function parameters
12924 and add an extra 'crl' parameter in the X509V3_CTX structure: guess
12925 what that's for :-) Fix to ASN1 macro which messed up
12926 IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
12929 *) Preliminary support for ENUMERATED type. This is largely copied from the
12933 *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
12934 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12936 *) Make sure `make rehash' target really finds the `openssl' program.
12937 [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]
12939 *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
12940 like to hear about it if this slows down other processors.
12943 *) Add CygWin32 platform information to Configure script.
12944 [Alan Batie <batie@aahz.jf.intel.com>]
12946 *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
12947 [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
12949 *) New program nseq to manipulate netscape certificate sequences
12952 *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
12956 *) Fixes to BN code. Previously the default was to define BN_RECURSION
12957 but the BN code had some problems that would cause failures when
12958 doing certificate verification and some other functions.
12959 [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]
12961 *) Add ASN1 and PEM code to support netscape certificate sequences.
12964 *) Add ASN1 and PEM code to support netscape certificate sequences.
12967 *) Add several PKIX and private extended key usage OIDs.
12970 *) Modify the 'ca' program to handle the new extension code. Modify
12971 openssl.cnf for new extension format, add comments.
12974 *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
12975 and add a sample to openssl.cnf so req -x509 now adds appropriate
12979 *) Continued X509 V3 changes. Add to other makefiles, integrate with the
12980 error code, add initial support to X509_print() and x509 application.
12983 *) Takes a deep breath and start adding X509 V3 extension support code. Add
12984 files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
12985 stuff is currently isolated and isn't even compiled yet.
12988 *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
12989 ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
12990 Removed the versions check from X509 routines when loading extensions:
12991 this allows certain broken certificates that don't set the version
12992 properly to be processed.
12995 *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
12996 Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
12997 can still be regenerated with "make depend".
13000 *) Spelling mistake in C version of CAST-128.
13001 [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]
13003 *) Changes to the error generation code. The perl script err-code.pl
13004 now reads in the old error codes and retains the old numbers, only
13005 adding new ones if necessary. It also only changes the .err files if new
13006 codes are added. The makefiles have been modified to only insert errors
13007 when needed (to avoid needlessly modifying header files). This is done
13008 by only inserting errors if the .err file is newer than the auto generated
13009 C file. To rebuild all the error codes from scratch (the old behaviour)
13010 either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
13011 or delete all the .err files.
13014 *) CAST-128 was incorrectly implemented for short keys. The C version has
13015 been fixed, but is untested. The assembler versions are also fixed, but
13016 new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
13017 to regenerate it if needed.
13018 [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
13019 Hagino <itojun@kame.net>]
13021 *) File was opened incorrectly in randfile.c.
13022 [Ulf Möller <ulf@fitug.de>]
13024 *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
13025 functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
13026 GeneralizedTime. ASN1_TIME is the proper type used in certificates et
13027 al: it's just almost always a UTCTime. Note this patch adds new error
13028 codes so do a "make errors" if there are problems.
13031 *) Correct Linux 1 recognition in config.
13032 [Ulf Möller <ulf@fitug.de>]
13034 *) Remove pointless MD5 hash when using DSA keys in ca.
13035 [Anonymous <nobody@replay.com>]
13037 *) Generate an error if given an empty string as a cert directory. Also
13038 generate an error if handed NULL (previously returned 0 to indicate an
13039 error, but didn't set one).
13040 [Ben Laurie, reported by Anonymous <nobody@replay.com>]
13042 *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
13045 *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
13046 parameters. This was causing a warning which killed off the Win32 compile.
13049 *) Remove C++ style comments from crypto/bn/bn_local.h.
13050 [Neil Costigan <neil.costigan@celocom.com>]
13052 *) The function OBJ_txt2nid was broken. It was supposed to return a nid
13053 based on a text string, looking up short and long names and finally
13054 "dot" format. The "dot" format stuff didn't work. Added new function
13055 OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote
13056 OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
13057 OID is not part of the table.
13060 *) Add prototypes to X509 lookup/verify methods, fixing a bug in
13061 X509_LOOKUP_by_alias().
13064 *) Sort openssl functions by name.
13067 *) Get the gendsa program working (hopefully) and add it to app list. Remove
13068 encryption from sample DSA keys (in case anyone is interested the password
13072 *) Make _all_ *_free functions accept a NULL pointer.
13073 [Frans Heymans <fheymans@isaserver.be>]
13075 *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
13077 [Anonymous <nobody@replay.com>]
13079 *) s_server should send the CAfile as acceptable CAs, not its own cert.
13080 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13082 *) Don't blow it for numeric -newkey arguments to apps/req.
13083 [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]
13085 *) Temp key "for export" tests were wrong in s3_srvr.c.
13086 [Anonymous <nobody@replay.com>]
13088 *) Add prototype for temp key callback functions
13089 SSL_CTX_set_tmp_{rsa,dh}_callback().
13092 *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
13093 DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
13096 *) X509_name_add_entry() freed the wrong thing after an error.
13097 [Arne Ansper <arne@ats.cyber.ee>]
13099 *) rsa_eay.c would attempt to free a NULL context.
13100 [Arne Ansper <arne@ats.cyber.ee>]
13102 *) BIO_s_socket() had a broken should_retry() on Windoze.
13103 [Arne Ansper <arne@ats.cyber.ee>]
13105 *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
13106 [Arne Ansper <arne@ats.cyber.ee>]
13108 *) Make sure the already existing X509_STORE->depth variable is initialized
13109 in X509_STORE_new(), but document the fact that this variable is still
13110 unused in the certificate verification process.
13111 [Ralf S. Engelschall]
13113 *) Fix the various library and apps files to free up pkeys obtained from
13114 X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
13117 *) Fix reference counting in X509_PUBKEY_get(). This makes
13118 demos/maurice/example2.c work, amongst others, probably.
13119 [Steve Henson and Ben Laurie]
13121 *) First cut of a cleanup for apps/. First the `ssleay' program is now named
13122 `openssl' and second, the shortcut symlinks for the `openssl <command>'
13123 are no longer created. This way we have a single and consistent command
13124 line interface `openssl <command>', similar to `cvs <command>'.
13125 [Ralf S. Engelschall, Paul Sutton and Ben Laurie]
13127 *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
13128 BIT STRING wrapper always have zero unused bits.
13131 *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
13134 *) Make the top-level INSTALL documentation easier to understand.
13137 *) Makefiles updated to exit if an error occurs in a sub-directory
13138 make (including if user presses ^C) [Paul Sutton]
13140 *) Make Montgomery context stuff explicit in RSA data structure.
13143 *) Fix build order of pem and err to allow for generated pem.h.
13146 *) Fix renumbering bug in X509_NAME_delete_entry().
13149 *) Enhanced the err-ins.pl script so it makes the error library number
13150 global and can add a library name. This is needed for external ASN1 and
13151 other error libraries.
13154 *) Fixed sk_insert which never worked properly.
13157 *) Fix ASN1 macros so they can handle indefinite length constructed
13158 EXPLICIT tags. Some non standard certificates use these: they can now
13162 *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
13163 into a single doc/ssleay.txt bundle. This way the information is still
13164 preserved but no longer messes up this directory. Now it's new room for
13165 the new set of documentation files.
13166 [Ralf S. Engelschall]
13168 *) SETs were incorrectly DER encoded. This was a major pain, because they
13169 shared code with SEQUENCEs, which aren't coded the same. This means that
13170 almost everything to do with SETs or SEQUENCEs has either changed name or
13171 number of arguments.
13172 [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]
13174 *) Fix test data to work with the above.
13177 *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
13178 was already fixed by Eric for 0.9.1 it seems.
13179 [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]
13181 *) Autodetect FreeBSD3.
13184 *) Fix various bugs in Configure. This affects the following platforms:
13188 unixware-2.0-pentium
13192 *) Eliminate generated files from CVS. Reorder tests to regenerate files
13193 before they are needed.
13196 *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
13200 Changes between 0.9.1b and 0.9.1c [23-Dec-1998]
13202 *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and
13203 changed SSLeay to OpenSSL in version strings.
13204 [Ralf S. Engelschall]
13206 *) Some fixups to the top-level documents.
13209 *) Fixed the nasty bug where rsaref.h was not found under compile-time
13210 because the symlink to include/ was missing.
13211 [Ralf S. Engelschall]
13213 *) Incorporated the popular no-RSA/DSA-only patches
13214 which allow to compile a RSA-free SSLeay.
13215 [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]
13217 *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
13218 when "ssleay" is still not found.
13219 [Ralf S. Engelschall]
13221 *) Added more platforms to Configure: Cray T3E, HPUX 11,
13222 [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]
13224 *) Updated the README file.
13225 [Ralf S. Engelschall]
13227 *) Added various .cvsignore files in the CVS repository subdirs
13228 to make a "cvs update" really silent.
13229 [Ralf S. Engelschall]
13231 *) Recompiled the error-definition header files and added
13232 missing symbols to the Win32 linker tables.
13233 [Ralf S. Engelschall]
13235 *) Cleaned up the top-level documents;
13236 o new files: CHANGES and LICENSE
13237 o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay
13238 o merged COPYRIGHT into LICENSE
13239 o removed obsolete TODO file
13240 o renamed MICROSOFT to INSTALL.W32
13241 [Ralf S. Engelschall]
13243 *) Removed dummy files from the 0.9.1b source tree:
13244 crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
13245 crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
13246 crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
13247 crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
13248 util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
13249 [Ralf S. Engelschall]
13251 *) Added various platform portability fixes.
13254 *) The Genesis of the OpenSSL rpject:
13255 We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
13256 Young and Tim J. Hudson created while they were working for C2Net until
13258 [The OpenSSL Project]
13261 Changes between 0.9.0b and 0.9.1b [not released]
13263 *) Updated a few CA certificates under certs/
13266 *) Changed some BIGNUM api stuff.
13269 *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD,
13270 DGUX x86, Linux Alpha, etc.
13273 *) New COMP library [crypto/comp/] for SSL Record Layer Compression:
13274 RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
13278 *) Add -strparse option to asn1pars program which parses nested
13280 [Dr Stephen Henson <shenson@bigfoot.com>]
13282 *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
13285 *) DSA fix for "ca" program.
13288 *) Added "-genkey" option to "dsaparam" program.
13291 *) Added RIPE MD160 (rmd160) message digest.
13294 *) Added -a (all) option to "ssleay version" command.
13297 *) Added PLATFORM define which is the id given to Configure.
13300 *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
13303 *) Extended the ASN.1 parser routines.
13306 *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
13309 *) Added a BN_CTX to the BN library.
13312 *) Fixed the weak key values in DES library
13315 *) Changed API in EVP library for cipher aliases.
13318 *) Added support for RC2/64bit cipher.
13321 *) Converted the lhash library to the crypto/mem.c functions.
13324 *) Added more recognized ASN.1 object ids.
13327 *) Added more RSA padding checks for SSL/TLS.
13330 *) Added BIO proxy/filter functionality.
13333 *) Added extra_certs to SSL_CTX which can be used
13334 send extra CA certificates to the client in the CA cert chain sending
13335 process. It can be configured with SSL_CTX_add_extra_chain_cert().
13338 *) Now Fortezza is denied in the authentication phase because
13339 this is key exchange mechanism is not supported by SSLeay at all.
13342 *) Additional PKCS1 checks.
13345 *) Support the string "TLSv1" for all TLS v1 ciphers.
13348 *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
13349 ex_data index of the SSL context in the X509_STORE_CTX ex_data.
13352 *) Fixed a few memory leaks.
13355 *) Fixed various code and comment typos.
13358 *) A minor bug in ssl/s3_clnt.c where there would always be 4 0
13359 bytes sent in the client random.
13360 [Edward Bishop <ebishop@spyglass.com>]